The peripheral nervous system of the head arises from two embryonic sources, the neural crest and cranial ectodermal placodes. Placodes are important for the formation of several cranial ganglia as well as portions of the eye, nose, and ear. They arise from thickened ectoderm and give rise to diverse cell types, including sensory neurons, ciliated sensory receptors, neuroendocrine and endocrine cells, glia, and other support cells. One group of placodes has been designated as “neurogenic” and forms portions of trigeminal and epibranchial ganglia, whereas other placodes are “sensory,” forming the ear, nose, lens, and hypophysis (reviewed in Baker and Bronner-Fraser,2001).
Although placodes are critical for normal development of the peripheral nervous system, many aspects of the tissue interactions and molecules involved in their induction and neurogenesis have yet to be elucidated. Tissue interactions mediated by growth factors are thought to govern cell fate decisions of particular placodes. For example, induction of the otic placode occurs through interaction of the ectoderm with the hindbrain and/or surrounding tissues (reviewed in Groves,2005), and this induction appears to involve fibroblast growth factors (FGFs) and Wnts (reviewed in Barald and Kelley,2004). The trigeminal placode, on the other hand, is induced in the ectoderm by a secreted factor(s) from the dorsal neural tube (Stark et al.,1997; Baker et al.,1999), the nature of which is unknown.
To determine candidate factors that may be involved in induction of the trigeminal placode, we have undertaken an analysis of the expression of growth factor receptors in the ectoderm of uninduced presumptive trigeminal level ectoderm in the developing chicken embryo. The rationale is that, in order for ectoderm to be responsive, it must express appropriate receptors for putative inducing ligands. To determine the nature of these receptors, we performed a reverse transcriptase-polymerase chain reaction (RT-PCR) screen with viable candidate receptors to determine which of the large number of secreted factors expressed by the dorsal neural tube at the time of trigeminal placode induction may represent putative ligands. The expression of receptors was then verified using whole-mount in situ hybridization to determine their spatial and temporal expression patterns. Although the list of receptors screened herein is not exhaustive, this represents an excellent launching point for understanding potential interactions involved in induction of the trigeminal placode. Here, we show that receptors for FGFs, insulin-like growth factors (IGFs), platelet-derived growth factors (PDGFs), Sonic hedgehog (Shh), the transforming growth factor-beta (TGFβ) superfamily, and Wnts are all expressed in patterns consistent with a role in trigeminal placode formation.
RESULTS AND DISCUSSION
We have used RT-PCR to screen for growth factor receptors that may play a role in trigeminal placode formation. By asking what receptors are present in the presumptive trigeminal placode ectoderm, we can determine candidate ligands for mediating the induction of the trigeminal placode. Because PCR does not provide spatial information, we verified the expression of many of the receptors by performing whole-mount in situ hybridization on embryos just before (stage [St.] 8, 3–4 somite stage [ss]) and during (St. 10, 9–11 ss) trigeminal placode formation. We found that 6 of 12 receptor families were present by PCR and by in situ hybridization as summarized in Table 1. These families include FGFs, IGFs, PDGFs (data not shown), Shh, the TGFβ superfamily, and Wnts. PDGFRα and PDGFRβ is presented in McCabe and Bronner-Fraser (manuscript submitted for publication,2007).
Table 1. Receptors Expressed in Ectoderm at St. 8 Ectoderm by RT-PCR and by In Situ Hybridization at St. 8 and St. 10a
Reverse transcriptase-polymerase chain reaction (RT-PCR) was performed on ectoderm isolated from the presumptive midbrain at stage(St.) 8, presence of band is depicted as yes or no. Selected receptors were subjected to whole-mount in situ hybridization and sectioning to determine spatial presence at St. 8 and St. 10. Receptors in bold indicate in situ hybridizations that are presented in the study. The qualitative levels of expressions for the probes are as follows: “++” for strong expression, “+” for some expression, “+/−” for when expression is ambiguous, and “−” for no detectable expression. The designation of presumptive trigeminal placodal ectoderm (PTGP) at St. 8 and trigeminal placode (TGP) indicates ectoderm that encompasses the presumptive placode, but may also extend to more medial and ventral ectoderm (see Fig. 1B,C to distinguish between PTGP and ventral ectoderm and Fig. 1E,F to distinguish between TGP and medial and ventral ectoderm).
All primers were designed using Primer3 (Rozen and Skaletsky,2000) with the exception of Patched and Smoothened (Lewis et al.,1999) and specificity was verified by BlastN analysis. Standard PCR conditions were used with no more than 35 cycles of amplification. PCR conditions for PCR primers were tested using cDNA from St. 18 whole chicken embryo and RT-PCR was performed using 3–4 ss (St. 8) ectodermal explant cDNA. The presence of a single band at the predicted size was indicated by a “yes” in Table 1. For many of the receptors deemed positive by PCR analysis, whole-mount in situ hybridizations were performed. The results are summarized in Table 1, with qualitative levels of expression for in situ probes designated as “++” for strong expression, “+” for some expression, “+/−” for when expression is ambiguous, and “−” for no detectable expression. The designation of presumptive trigeminal placodal ectoderm (PTGP) at St. 8 and trigeminal placode (TGP) in Table 1 indicates ectoderm that encompasses the presumptive placode but may also extend to more medial and ventral ectoderm (see Fig. 1B,C to distinguish between PTGP and ventral ectoderm and Fig. 1E, F to distinguish between TGP and medial and ventral ectoderm).
An ideal candidate for a secreted factor and its receptor to be involved in the induction of the trigeminal placode must fulfill several criteria. The ligand should be expressed in the neural folds beginning no later than St. 8 and should continue to be expressed in the dorsal neural tube until at least St. 10. At St. 10–11, the ligand should be present in the dorsal neural tube at all axial levels, as trunk neural tube is capable of inducing trigeminal placode markers when juxtaposed with St. 8 presumptive midbrain level ectoderm in culture (Baker et al.,1999; McCabe et al.,2004). The receptor expression, on the other hand, should be broadly expressed in the midbrain level ectoderm at St. 8, corresponding to the area of competence to respond to trigeminal placode inducers (Fig. 1A,B). By St. 10, induction of many of the trigeminal placode cells has already taken place, but other ectodermal cells continue to be induced over time. Therefore, the receptor expression may continue to be expressed in a large number of ectodermal cells that maintain competence to respond to inducing signals (Fig. 1D,E). However, receptor expression also may become restricted to trigeminal placode cells over time.
Cells within the ectoderm are able to respond to neural tube-derived signals by expressing the appropriate receptors. One inherent assumption in this scenario is that the secreted factors are working as positive factors, such that presence of the factor promotes induction by turning on a signaling cascade that results in gene expression necessary for trigeminal placode fate, such as Pax3. However, we cannot rule out the possibility that inhibition of a signaling cascade may mediate induction, by which ectodermal fate transcription factors are down-regulated allowing placode fate transcription factors to dominate. A classic example of inhibition resulting in neural induction is the case of Noggin inhibition of the TGFβ superfamily member bone morphogenetic protein-4 (BMP4), where blocking of BMP signaling results in neural induction (Piccolo et al.,1996; Zimmerman et al.,1996). In the presence of BMP signaling, epidermal fate is promoted. Another possibility is that multiple factors may be required for trigeminal placode formation. These may be positive and/or negative factors. To begin to unravel the complexity of trigeminal placode induction, it will be important to identify as many putative players as possible.
We tested for the presence of FGF receptor-1 (FGFR1), FGFR2, FGFR3, FGFR4, and cysteine-rich fibroblast growth factor receptor (CFR). RT-PCR analysis of 3–4 ss (St. 8) presumptive trigeminal placode revealed that FGFR1, FGFR2, FGFR4, and CFR are present (Table 1). Whole-mount in situ hybridizations for FGFR1–4 have been previously published; however, sections of presumptive midbrain level at St. 8 and midbrain level at St. 10 were not shown (Lunn et al.,2007). In situ hybridization presented here confirms that FGFR2 and CFR are readily detectable in the early ectoderm. At St. 8, FGFR2 is robustly expressed in the presumptive trigeminal placodal ectoderm and neural folds (Fig. 2A–C). The expression of FGFR2 appears to be specific in the trigeminal placode St. 10 at the midbrain (Fig. 2D–F). CFR does not act as a classic FGFR, but rather has been shown to act as a sink for some FGF ligands in the endoplasmic reticulum (Burrus et al.,1992). CFR was detected in the presumptive trigeminal placode, as well as more ventral ectoderm, and neural folds at St.8 (Fig. 2G–I), and resolved to the trigeminal placode, medial ectoderm, and perhaps the neural crest by St. 10 (Fig. 2J–L).
There are at least 13 FGF ligands (FGF1–4, 7–10, 12–14, 16, 18) present in the chicken genome (reviewed in Thisse and Thisse,2005). Although expression of many (FGF1, 2, 7, 9, 10, 12, 13, 14) has not been reported in the chick in places and times relevant to our study, there are several FGFs that are expressed at or near the presumptive midbrain cranial neural folds at St. 8. For example, FGF4 is expressed at exactly the right place in the cranial neural folds albeit at very low levels (Shamim and Mason,1999). FGF8 and FGF18 are both robustly expressed in the neural folds near the presumptive olfactory placode at St. 8 and the forming midbrain/hindbrain boundary at St. 10 (Ohuchi et al.,2000; Crossley et al.,2001). On this basis, we hypothesize that FGF4 or another FGF is a likely candidate for being a secreted ligand capable of acting during the process of the trigeminal placode induction.
In vitro binding assays suggest that splice variants of FGF receptors may impart differential specificity (Johnson et al.,1991; Werner et al.,1992; Chellaiah et al.,1994). For example, FGF4 is mitogenic in the presence of FGFR1c, FGFR2c, FGFR3c, and FGFR3 in BaF3 cell lines, whereas FGF8 is mitogenic in response to FGFR3c and FGFR4 (Ornitz et al.,1996). Binding to heparan sulfate proteoglycans may further modulate specificity (Ornitz et al.,1996). Based on expression, FGF4 and FGFR2 are the most likely candidates for trigeminal placode induction.
IGF receptor-1 (IGFR1), IGFR2, and insulin receptor (IR) were all detected by PCR in 3–4 ss (St. 8) ectoderm. Robust expression of IGFR1 was found in the presumptive trigeminal placode and neural folds at St. 8 (Fig. 3A–C). At St. 10, IGFR1 is expressed in the trigeminal placode, in the medial and ventral ectoderm, but absent from the neural tube (Fig. 3D–F).
The ligands for the IGFRs include IGF1, IGF2, and insulin. In particular, IGF2 is an excellent candidate for a trigeminal placode inducer. It is strongly expressed in the cranial neural folds at St. 8, and its expression is maintained in the neural tube at St. 10 (Allan et al.,2003). IGF1 is expressed in the neural tube at St. 12 (Allan et al.,2003), but has not been analyzed at St. 8 and St. 10.
The IGFR1 receptor tyrosine kinase is heterotrimeric (also binding to the insulin receptor) and autophoshophorylates upon binding to IGF1, IGF2, or insulin (Pandini et al.,2002), activating PI3K and MAPK pathways (reviewed in Dupont and Holzenberger,2003). In contrast, IGFR2 is monomeric glycoprotein, with no intrinsic tyrosine kinase activity. It may store ligand on the surface (reviewed in Jones and Clemmons,1995) and/or target it for lysomal degradation (Oka et al.,1985). Most IGF1, IGF2, and insulin ligands are found bound to one of the six IGF binding proteins (IGFBP1–6) in extracellular spaces and circulation, thereby modulating bioavailability of the ligand (Arai et al.,1996; Delbe et al.,1991; Rajah et al.,1999). Because IGFBP5 is expressed in the notochord at St. 8–10 (Allan et al.,2003), it may prevent IGF signaling ventrally, thus restricting IGFR1 signaling in the ectoderm.
Sonic Hedgehog Receptor, Patched (Ptc)
Both Patched (Ptc) and Smoothened (Smo) are detectable in the midbrain level ectoderm by RT-PCR at St. 8 (Table 1). Smo is readily discernable by in situ hybridization and is strongly expressed in the neural folds and present in the presumptive trigeminal placode and adjacent ectoderm at St. 8 (Fig. 3G–I). At St. 10, Smo is highly expressed in the trigeminal placode, medial and ventral ectoderm and the migrating neural crest, but present at low levels in the neural tube (Fig. 2J–L).
Shh, a ligand for Ptc, is expressed in the prechordal mesendoderm and notochord at St. 8 and 10 (Marti et al.,1995; Crossley et al.,2001). Although most of Shh signaling is thought to work at short range, experiments by Gritli-Linde et al. (2001) have shown that Shh can travel over considerable distances in association with heparan sulfate proteoglycans. Desert Hedgehog (Dhh) is known to function in spermatogenesis (Bitgood et al.,1996) and is not likely to be present in the midbrain dorsal neural tube. However, Indian Hedgehog (Ihh) may be present, but its expression at St. 8 and St. 10 in the midbrain is not currently known.
Binding of the Patched (Ptc) receptor by Hedgehog ligand relieves repression of a signaling cascade. Smoothened (Smo) is a cell membrane protein that works downstream of Ptc and acts as a positive mediator of Shh signaling (reviewed in Huangfu and Anderson,2006). The most likely scenario for involvement in trigeminal placode is that Shh signaling occurs through Ptc and is relayed by Smo.
Several of the TGFβ superfamily receptor members were detectable by RT-PCR: ACTRIA (ALK2), ACTRIB (ALK4), ACTRIIA, ACTRIIB, BMPRII, and TGFβRIIB (Table 1). Of these, ACTRIIA and ACTRIIB are abundantly expressed in the presumptive trigeminal placode at St. 8, with ACTRIIA having broader ectodermal expression (Fig. 4A–C,G–I), as well as the neural folds. At St. 10, ACTRIIA and ACTRIIB are no longer expressed in the neural tube; ACTRIIA is expressed in the trigeminal placode and medial ectoderm (Fig. 4D–F) and ACTRIIB in the placode and medial and ventral ectoderm (Fig. 4J–L). There is also some expression of ACTRIIB in the migrating neural crest (Fig. 4L).
There are at least 30 different TGFβ superfamily ligands, although the expression of many of the ligands has not been reported in birds. However, several members are excellent candidates. Vg1 is present in the neural folds at St. 8 and in the dorsal neural tube at St. 10 in the chicken (Seleiro et al.,1996). BMP4 and BMP7 are found in the developing neural plate at St. 8 (Watanabe and Le Douarin,1996; Streit et al.,1998). However, BMP4 is not present in midbrain level neural tube at St. 10 (Watanabe and Le Douarin,1996). TGFβ1 (chicken TGFβ4) is expressed in the neural folds at St. 8 (Jakowlew et al.,1992). TGFβ3 is expressed at higher levels in the neural folds at St. 8 and neural tube at St. 10 than TGFβ2 (Jakowlew et al.,1994).
Several secreted factors inhibit TGFβ superfamily signaling (reviewed in Massague and Chen,2000). Follistatin, which blocks Activin (Nakamura et al.,1990), BMP4 and BMP7 (Yamashita et al.,1995; Liem et al.,1997), is abundantly expressed in the neural folds at St. 8, down-regulated in the midbrain by St. 10, and subsequently up-regulated in the hindbrain (Connolly et al.,1995). The BMP inhibitors Chordin and Noggin (Piccolo et al.,1996; Yamashita et al.,1995; Liem et al.,1997) are expressed in the notochord but not the neural folds at St. 8 (Streit et al.,1998; Tonegawa and Takahashi,1998; Capdevila and Johnson,1998).
The TGFβ superfamily of ligands signal through two receptor serine/threonine protein kinases. There are two classes of receptors, Type 1 and Type 2. Type 2 receptors bind the ligand and phosphorylate the GS region on a Type 1 receptor. The kinase region on the Type 1 receptor then goes on to phosphorylate the appropriate SMAD proteins, which can alter transcription of many genes (reviewed in Massague and Chen,2000). Activins, Nodals, and Vg1 can signal through Type 1 receptors ACTRIB and ACTRIC and Type 2 receptors ACTRIIA and ACTRIIB. ACTRIB is present (ACTRIC currently has no chicken ortholog) and both ACTRIIA and ACTRIIB are also expressed at St. 8 and St. 10 (Table 1; Fig. 4). However, the Type 1 receptor TGFβRI is not present in the presumptive midbrain ectoderm. The best candidates include Vg1 signaling through ACTRIB and ACTRIIA or ACTRIIB; BMP7 signaling through ACTR1A and BMPRII, ACTRIIA, or ACTRIIB. Keeping in mind that the Activin inhibitor Follistatin is expressed in neural folds (Connolly et al.,1995) and could be interfering with signaling of any of the ligands.
The BMPs (BMP2, 4, and 7) as well as GDF-5 can signal through several Type 1 (BMPRIA, BMPR1B, and ACTR1A) and Type 2 receptors (BMPRII, ACTRII, ACTRIIB). Of these, the Type 1 receptor ACTR1A (the expression of BMPR1B is unknown) and the Type 2 receptors BMPRII, ACTRIIA, ACTRIIB are expressed in 3–4 ss (St. 8) midbrain ectoderm.
The TGFβ ligands, TGFβ1–3, signal through Type 1 receptor TGFβRI and Type 2 receptor TGFβRII. Because TGFβR1 was not found to be expressed at 3–4 ss midbrain ectoderm, it is unlikely that the ligands TGFβ1, TGFβ2, TGFβ3 expressed in the neural folds are inducing the trigeminal placode.
Frizzleds are seven transmembrane serpentine receptors of the Wnt ligands. All of the Frizzleds (Frizzled1, 2, 3, 6, 7, 10) tested were present as assayed by RT-PCR (Table 1). The tissue expression of Frizzled1, 2, 7, 10 has been previously published (Stark et al.,2000). We find additional regions of expression for Frizzled1 and Frizzled10 in the presumptive trigeminal placode. Frizzled1 and Frizzeld10 are robustly expressed in the presumptive trigeminal placode neural folds at St. 8 (Fig. 5A–C,M–O). Frizzled10 has additional expression in the adjacent ectoderm at St. 8 (Fig. 5M–O). At St. 10, Frizzled1 and Frizzled10 are expressed in the trigeminal placode and medial and ventral ectoderm, with auxiliary expression of Frizzled1 in the neural crest (Fig. 5D–F), while Frizzled10 is maintained in the neural tube (Fig. 5P–R). Frizzled6, on the other hand, is restricted to the presumptive trigeminal placode at the presumptive midbrain level at St. 8 (Fig. 5G–I) and is lower in the trigeminal placode than the surrounding ectoderm at St. 10 (Fig. 5J–L).
Wnt 1 is highly expressed in the neural folds and dorsal neural tube at St. 8 and St. 10, respectively (Hollyday et al.,1995), making it an excellent candidate ligand. Similarly, Wnt 4 is expressed at St. 8 and St. 10 in the neural folds and dorsal neural tube (Hollyday et al.,1995). Wnt 3A is not expressed in the neural tube until St. 9, but is found in the midbrain at St. 10 (Hollyday et al.,1995). Wnt 5B, Wnt 6, and Wnt 11 are expressed in the midbrain neural tube at St. 10 (Cauthen et al.,2001; Schubert et al.,2002). The Wnt inhibitor Dkk-3 is expressed in the neural folds at St. 8 (Monaghan et al.,1999). Shisa, a dual Wnt and FGF inhibitor, is expressed in the neural folds at St. 8 (Filipe et al.,2006). It is possible that several Wnts are involved in the process of trigeminal placode induction; however, Wnt 1 and Wnt 4 make the best candidates.
Wnt, in conjunction with LRP, binds to Frizzleds, activating canonical (reviewed in Gordon and Nusse,2006) and noncanonical pathways (reviewed in Saburi and McNeill,2005). Possible combinations that may be involved in trigeminal placode induction include the ligands Wnt 1 or Wnt 4 signaling through Frizzled1, 6, or 10 with Shisa acting as an inhibitor.
Growth Factors and Placodes
Many of the growth factor families found to be expressed in presumptive trigeminal placode have been shown to play a role in induction, specification, or neurogenesis in other placodes. For example, FGFs have been implicated in otic placode induction (Maroon et al.,2002; Leger and Brand,2002; Liu et al.,2003; Wright and Mansour,2003; Phillips et al.,2004; Ladher et al.,2005; Martin and Groves,2006) and olfactory placode specification (Bailey et al.,2006), BMP7 in epibranchial placode neurogenesis (Begbie et al.,1999), and Shh in otic morphogenesis (Liu et al.,2002; Riccomagno et al.,2002; Koebernick et al.,2003; Bok et al.,2005) and Wnts in otic placode induction (Ohyama et al.,2006) and morphogenesis (Riccomagno et al.,2002; Stevens et al.,2003). IGFs and PDGFs have not yet been implicated in placode development. It is not surprising that many of the same growth factors play multiple roles in different pathways as conservation of signaling pathways in development is common.
There is increasing evidence that all placodes arise from a common area called the “preplacodal domain,” which forms as a horseshoe-shaped domain anterior to the Hensen's node shortly after neural plate formation (reviewed in Streit,2004; Bailey and Streit,2006; Schlosser,2006). Over time, the cells within the preplacodal domain disperse into identifiable areas along the neural tube and form individual placodes (D'Amico-Martel and Noden,1983; Couly and Le Douarin,1985,1987; Noden,1993). Additional evidence for shared lineage comes from the finding that many genes involved in the specification of the trigeminal placode are also used in development of other placodes (McCabe et al.,2004). Many genes up-regulated in response to trigeminal placode induction were also expressed by other placodes at different stages of their development (McCabe et al.,2004), indicating that the same genes could be used in different combinations to result in differential placodal fates.
Normal development of the peripheral nervous system is dependent on the proper formation of cranial ectodermal placodes. The molecular interactions underlying the formation of the trigeminal placode have not yet been determined. Previous studies have indicated that an unidentified secreted factor(s) from the dorsal neural tube is responsible for the formation of the trigeminal placode (see diagram, Fig. 1A,B). We have performed an efficient screen to narrow the candidates for induction of the trigeminal placode from the large number of factors present in the dorsal neural tube. Because the trigeminal placode induction is likely to be a multifactorial process, it is critical to know as many of the factors as possible to better understand the complex molecular interactions. To this end, we have determined that the following families are potential players in trigeminal placode induction: FGFs, IGFs, PDGFs, Shh, the TGFβ superfamily, and Wnts (see Fig. 1C,F, for summary). The present RT-PCR screen for candidate receptors expressed in presumptive trigeminal ectoderm is an important first step in dissecting the interactions underlying induction of the trigeminal placode. Similar interactions may also be involved in formation of other placodes as well.
Ectodermal Explants and Embryo RNA Isolation
Fertile chicken eggs (Gallus gallus domesticus) were incubated at 38°C until they reached the desired stage (described here either according to the criteria of Hamburger and Hamilton  stage [St.] or by somite stage [ss]). Ectodermal explants from the presumptive midbrain were removed from 3–4 ss (St. 8) embryos using pulled glass needles and placed in Ringer's solution on ice until RNA could be isolated. To minimize mesodermal contamination, the ectodermal explants were visually inspected for mesodermal contamination. If mesoderm was attached and could not be removed, then the explant was discarded. Once ectodermal explants were collected, they were briefly microfuged to remove excess Ringer's solution. Whole embryos from St. 18 were collected in Ringer's solution on ice until RNA could be isolated. RNA was isolated using RNAqueous kit (Ambion) completed as directed by manufacturer.
cDNA was synthesized from total RNA using random hexamers (IDT) and SuperscriptII Reverse Transcriptase (Invitrogen) as directed by the manufacturer. PCR primers were designed by Primer3 (Rozen and Skaletsky,2000) to individual receptors to result in 600- to 800-bp products. Ptc and Smo primers were designed by Lewis et al. (1999), which result in 218-bp and 223-bp products, respectively. Oligonucleotide (IDT) sequences were subjected to BlastN (NCBI) analysis to check for specificity. Standard PCR conditions were used with no more than 35 cycles of amplification. PCR conditions for individual receptors available upon request. PCR primers were first tested using cDNA from St. 18 whole chicken embryo. If a single band were present at the predicted size, then the primers were used on cDNA from 3–4 ss (St. 8) ectodermal explants. RT-PCR was performed using 3–4 ss (St. 8) ectodermal explant cDNA with the presence of a single band at the predicted size was indicated by a “yes” in Table 1.
Table 2 provides the forward and reverse primers for the receptors studied.
Table 2. Receptor Primers Used
5′ CCAAACTGTGCAACATGGAC 3′
5′ CAGCCCCAACGACTGTATTT 3′
5′ GAATGGTTGGGAATGCAACT 3′
5′ TAGTTGCCAGGTTGATGCTG 3′
5′ GTCTTCTGCCTCGTGGTGAT 3′
5′ ATCCCATTCTTCGTCCTCCT 3′
5′ GTCTCAGACGCACTCCCTTC 3′
5′ GTCAGGCTTGAACTCCTTGC 3′
5′ GTGGCAGTGAAGATGCTGAA 3′
5′ GGTGCAGTTGGCAGGTTTAT 3′
5′ TTGGCCTTGCTAGAGACGTT 3′
5′ AGGGCAGTACCCTCAGGTTT 3′
5′ CGCATGGACAAGAAGCTGTA 3′
5′ AGGGCAGTACCCTCAGGTTT 3′
5′ TTTTTGCCCAGAAGCAGACT 3′
5′ GTTGAGCACCTCCTTCTTGC 3′
5′ CTTGGATAACCCCGATGAGA 3′
5′ GTGGAAGCCGCAAGATCTAA 3′
5′ TGCAAGCATGCGTGAGAGGATAGA 3′
5′ GGCCAGGATTACTCTCTGTGTCTG 3′
5′ AGAACTGGTGCGGGTAATTG 3′
5′ CCACCACAGCTACCTCCATT 3′
5′ CTGCGGGAGCGCATTGAGTTCCTC 3′
5′ CGACCAGGGTTATTCTCAGCGTCG 3′
5′ GAGGTGATGGTGAAGGAGTG 3′
5′ TCTGGGGAAGGCGTCTGGTT 3′
5′ GCCTCCCTCTTCCTCTCTGT 3′
5′ GAGGGTGTAGTTGCCGTTGT 3′
5′ CCCAAACTGCGACTTACCAT 3′
5′ ACAGTGAATGGAATGCACCA 3′
5′ ATGCTGCTATGGCGATCTCT 3′
5′ AAATGCTTTCAGGTGCCATC 3′
5′ CAAGTGGGAGTTTCCCAGAG 3′
5′ GAGGATCTTCAGCTCGGACA 3′
5′ TGACTTGGCAGCTCGTAATG 3′
5′ CTGTTCCTTTTGCCAAGCTC 3′
5′ ACGTAGCCTGCACTGTCCTT 3′
5′ CTGCACTGATGGGTTTCTGA 3′
5′ CCTCCGGGTACCACTATGG 3′
5′ GTGACGATGTAACTCTCCG 3′
5′ ATCGCCAGGACCTTGAAGAACC 3′
5′ TGAGGTTGGTGTCAGTGGTTCG 3′
5′ CCTCGGGGTACCACTACG 3′
5′ AGCTCGGGGCTGTCGGCC 3′
5′ TGCCAGCCTATCACTACTGTG 3′
5′ CATTCGACATCCTGAAGCTC 3′
5′ ATCCCAAAGTGCCTTCCAG 3′
5′ GAGGAAGCCGTGAGCCTAC 3′
5′ ATGAGGGAGTGTTGGTACGC 3′
5′ CACCATTGGTTTTTGCAGTG 3′
5′ TGCTGTCGGACTGATTTCTG 3′
5′ TGAGGAGAGCCCTGTTGTCT 3′
5′ AAGACTAACAGCCCTGCGAA 3′
5′ ATTACCACTTCCGCGTGTTC 3′
5′ ACACGAAGGTACATGGCTCC 3′
5′ CAGCTCCAGTGAAGAGTCCC 3′
5′ TTTGCAATTTATCGCAGCAG 3′
5′ CACTCCTAGCACTGGCATGA 3′
5′ TTGGTGAAGAGCAGCATGTC 3′
5′ GTAAACTCGCTGAACGCCTC 3′
5′ CAGGAACAATGGAGGGAAGA 3′
5′ GATGGGAAAAATTTCCAGCA 3′
5′ CTGGAAAAGCTGGAAATGGA 3′
5′ TAACGTGAGCTGATGGCTTG 3′
5′ TTCATTGGCTTTCTGCTTGA 3′
5′ ACTGCCAGCCATGGTAAAGA 3′
5′ CACTGCTCCTGTAACCACCA 3′
5′ TGACCCACCATATTGTTCCTG 3′
5′ CAGAGCGACCCATCATCTTC 3′
5′ TGCGGAACAAGGACACAAAG 3′
5′ GGCCACAAAACCAGAAAGAA 3′
In Situ Hybridization
Whole chicken embryos were collected at the desired stage in Ringer's solution on ice and subsequently fixed overnight in 4% paraformaldehyde, pH 9.5. Alkaline fixation has been reported to increase signal for in situ hybridization (Basyuk et al.,2000). Antisense digoxigenin-labeled RNA probes were made according to manufacturer's directions (Roche). Whole-mount in situ hybridization was performed as previously described (Kee and Bronner-Fraser,2001) using nitroblue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate (NBT/BCIP; Roche) for color detection. Whole-mount pictures were taken using a Zeiss Stemi SVII microscope with an Olympus DP10 digital camera. Selected embryos were then processed for cyrosectioning (20 μm thick sections) as previously described (Sechrist et al.,1995). Sections were photographed using a Zeiss Axioskop2 Plus. The qualitative levels of expressions for the in situ probes are as follows: “++” indicates strong expression, “+” indicates expression, “+/−” indicates ambiguous expression, and “−” indicates no detectable expression.
The following plasmids were obtained from Geneservice Ltd.: FGFR1, FGFR2, FGFR4, CFR, IGFR1, IGFR2 (mannose-6-phosphate receptor), IR, PDGFRα, PDGFRβ, Smo, ACTR1A, ACTR1B, ACTRIIB, BMPRII, Frizzled6. The FGFR1 plasmid was generously given to us by Dr. Helen McBride, Ptc plasmid by Dr. Cliff Tabin, and ACTRIIA plasmid by Dr. Claudio Stern.
We thank Samuel Ki and Andrea Manzo for technical assistance.