The chicken talpid3 gene encodesa novel protein essentialfor Hedgehog signaling
http://www.100md.com
基因进展 2006年第10期
1 Division of Cell and Developmental Biology, Wellcome Trust Biocentre (WTB), University of Dundee, Dundee DD1 5EH, United Kingdom; 2 Department of Genetics and Genomics, Roslin Institute (Edinburgh), Midlothian EH25 9PS, United Kingdom; 3 Division of Biomaterials and Tissue Engineering, Eastman Dental Institute, London WC1X 8LD, United Kingdom; 4 Division of Developmental Neurobiology, National Institute for Medical Research, London NW7 1AA, United Kingdom; 5 School of Biological Sciences, University of East Anglia, Norwich NR4 7TJ, United Kingdom
Abstract
Talpid3 is a classical chicken mutant with abnormal limb patterning and malformations in other regions of the embryo known to depend on Hedgehog signaling. We combined the ease of manipulating chicken embryos with emerging knowledge of the chicken genome to reveal directly the basis of defective Hedgehog signal transduction in talpid3 embryos and to identify the talpid3 gene. We show in several regions of the embryo that the talpid3 phenotype is completely ligand independent and demonstrate for the first time that talpid3 is absolutely required for the function of both Gli repressor and activator in the intracellular Hedgehog pathway. We map the talpid3 locus to chromosome 5 and find a frameshift mutation in a KIAA0586 ortholog (ENSGALG00000012025), a gene not previously attributed with any known function. We show a direct causal link between KIAA0586 and the mutant phenotype by rescue experiments. KIAA0586 encodes a novel protein, apparently specific to vertebrates, that localizes to the cytoplasm. We show that Gli3 processing is abnormal in talpid3 mutant cells but that Gli3 can still translocate to the nucleus. These results suggest that the talpid3 protein operates in the cytoplasm to regulate the activity of both Gli repressor and activator proteins.
[Keywords: Gli; Hedgehog signaling; chicken; talpid3]
Received October 6, 2005; revised version accepted March 9, 2006.
The chicken embryo has served as a powerful and influential model system. The recent completion of the chicken genome (International Chicken Genome Sequencing Consortium 2004) allows new opportunities to exploit the ease of manipulating chick embryos to study vertebrate development. A number of chicken mutants have arisen spontaneously, often in agricultural flocks where they were noticed because of reduced hatchability. Talpid3 is one such mutant (Hunton 1960), and is a member of the classical developmental talpid group, so called because their paddle-shaped limbs resemble those of the mole (Talpa). In the talpid3 mutant, besides limb defects, there is an almost bewildering set of malformations including face, skeleton, and vascular defects (Ede and Kelly 1964a, b; Hinchliffe and Ede 1968). It is now clear that all these abnormalities are in regions of the embryo that depend on Hedgehog (Hh) signaling (Ingham and McMahon 2001).
An intriguing feature of talpid3 is that some aspects of the gross phenotype (e.g., polydactylous limbs) suggest a gain of Hh function whereas others (e.g., hypoteleorism in which the eyes are pulled together) suggest a loss of function. Analysis of Hh target gene expression in both limb and face shows that, even within one region, some targets are expressed at abnormally low levels while others are expressed ectopically (Izpisua-Belmonte et al. 1992; Francis-West et al. 1995; Lewis et al. 1999a; Buxton et al. 2004). These opposite changes in gene expression seen in talpid3 may reflect dual roles of the Gli transcriptional effectors of Hh signaling, which can act as either activators (A) or repressors (R) (Ruiz i Altaba et al. 2003). There are three Gli proteins, Gli1, Gli2, and Gli3. Experiments in mice show that Gli3 but not Gli1 or Gli2 is required for limb patterning (Bai et al. 2002). In normal chick limb buds, Sonic hedgehog (Shh) produced posteriorly diffuses across the bud and prevents processing of full-length Gli3 protein to a short repressor form (Gli3R). This results in low levels of full-length Gli3 protein that, together with other Gli proteins, can act as a transcriptional activator (Gli3A) and a gradient of Gli3R, highest anteriorly (Wang et al. 2000). Absence of high-level expression of some target genes in the posterior of talpid3 limb buds (e.g., Ptc1) and the inability to rescue their expression through the addition of Shh protein (Lewis et al. 1999a) suggests that GliA function is defective. Ectopic expression of other target genes in the anterior of talpid3 limb buds suggests that Gli3R levels are reduced anteriorly. The absence of high-level expression of Ptc, which encodes the Shh receptor and binds Shh, could lead to widespread diffusion of Shh ligand, reducing Gli3R levels. However, the limbs of Gli3–/– and Shh–/–, Gli3–/– mutant mice, like those of talpid3 chick embryos, are polydactylous, with many unpatterned digits, and also show ectopic expression of Hh target genes (Hui and Joyner 1993; Litingtung et al. 2002; te Welscher et al. 2002).
Here we show that ectopic gene expression in talpid3 limb buds is ligand independent and assess the distribution of Gli3A versus Gli3R. The dorsoventral pattern of the neural tube in chick has been shown to depend on a balance between GliA and GliR function (Persson et al. 2002; Stamataki et al. 2005). We explore whether the talpid3 mutation compromises both GliA and GliR function in dorsoventral patterning of the neural tube and adjacent somites. We rescue Gli function by electroporation into the talpid3 neural tube. We then combine knowledge of the chicken genome with the manipulability of chick embryos to identify the talpid3 gene and carry out functional complementation tests within the mutant neural tube to confirm identification unequivocally. We show that the talpid3 protein localizes to the cytoplasm and that although Gli3 processing is abnormal in mutant chick cells, both Gli3A and Gli3R can still translocate to the nucleus. We conclude that the talpid3 protein acts in the cytoplasm to regulate the functional activity of both the repressor and activator forms of Gli.
Results
Basis of the talpid3 limb polydactyly
The polydactylous phenotype of talpid3 limbs suggests a gain of Hh function. Comparison of stage 19 Hamburger Hamilton (HH) (Hamburger and Hamilton 1951) limb buds using immunohistochemistry showed that Shh protein is more widespread in talpid3 (Fig. 1A,B). In Western blots, more Shh protein was found in the middle third of talpid3 limb buds than in wild type (Fig. 1C). To test whether ectopic gene expression in talpid3 limb buds is dependent on this ectopic Shh protein, we attempted to rescue the mutant phenotype by inhibiting Shh signaling or removing its source; we monitored rescue by examining Hoxd13 expression, which is ectopically expressed throughout talpid3 limb buds instead of being posteriorly restricted (Izpisua-Belmonte et al. 1992). We applied cyclopamine, which inhibits Smo activity (Frank-Kamenetsky et al. 2002) to stage 17 HH wing buds. In 5/6 treated talpid3 wings, Hoxd13 expression was still strongly expressed (Fig. 1E; right wing bud) whereas, in 3/4 treated wild-type wings, expression was much reduced (Fig. 1D; right wing bud). Expression of Hoxd13 was retained in most cases in wings treated with ethanol (4/5 wild-type, 3/3 talpid3) (data not shown). We also removed future Shh expressing cells at stage 16 HH. This had no effect on Hoxd13 expression in talpid3 wing buds (3/3) (Fig. 1G; right wing bud) but, in wild type, expression was reduced or abolished (7/7) (Fig. 1F; right wing bud). These results show that, although Shh protein is more widely distributed in talpid3 limb buds, ectopic Hoxd13 expression is ligand independent.
Figure 1. Localization of Shh protein, effects of reducing Shh activity in wild-type and talpid3 limb buds, and status of Gli3 protein. (A,B) Localization of Shh protein in stage 19 HH wild-type (A) and talpid3 (B) limb buds. Dotted lines indicate limit of staining. (C) Western blot analysis of Shh protein in different regions (anterior, middle, posterior) of stage 20 HH wild-type (wt) and talpid3 (ta3) limbs. (D,E) Hoxd13 expression in wild-type (D) and talpid3 (E) right limbs treated with cyclopamine compare with control left limbs. (F,G) Hoxd13 expression in wild-type (F) and talpid3 (G) right limbs after removal of posterior mesenchyme; compare with control left limbs. (H,I) Status of Gli3 protein in different regions (anterior, a; middle, m; posterior, p) of stage 24 wild-type and talpid3 limb buds. (H) Immunoblot shows levels of Gli3A and Gli3R in wild-type and talpid3 legs (n > 3). (I) Histograms show levels of Gli3 proteins in wild-type and talpid3 wing buds; standard deviations shown, n = 3. (J) Immunoblot directly compares levels of Gli3 protein in wild-type and talpid3 limbs and trunk tissue. Numbers at top of each lane show ratios of Gli3A/Gli3R. H and J also illustrate tubulin loading controls.
Analysis of gene expression in talpid3 limb buds and the fact that these changes in expression are ligand independent suggest that both GliA and GliR functions are defective. In order to assess the levels of the different forms of Gli proteins, we extracted proteins from the anterior, middle, and posterior thirds of talpid3 buds (either wing or leg) and carried out Western blot analysis using an antibody against Gli3 that recognizes both the full-length form, Gli3A (190 kDa) and the processed short form, Gli3R (83 kDa). We then compared the levels of the two forms of Gli3 with those seen in wild-type limb buds by carrying out analysis on either wing buds or leg buds. We present the data in histograms for the wing buds and show the Western blot for the leg buds. In the wild-type limb buds, we detected both Gli3A and Gli3R forms (Fig. 1H,I) and, as previously reported (Wang et al. 2000), Gli3R is present in a gradient along the anterior–posterior axis with maximal levels of Gli3R in the anterior region of the bud (Fig. 1H,I). In contrast, in talpid3 wing buds (Fig. 1I), levels of Gli3A are strikingly increased and the polarized distribution of Gli3R is abolished. Western blots of Gli3 in talpid3 leg buds (Fig. 1H) also confirmed that Gli3R is not graded and, in addition, that Gli3A levels are strikingly elevated. We also carried out Western blot analysis on extracts from whole limb buds (both wing and leg pooled together) from wild-type and talpid3 embryos and estimated the Gli3A/Gli3R ratio (Fig. 1J). This analysis shows that this ratio is strikingly elevated in the mutant. Furthermore, we found that the ratio of Gli3A/Gli3R was also elevated in extracts from talpid3 trunk, which includes other tissues, such as neural tube and somites, that are patterned by Hh signaling. Thus there is a generalized defect in Gli3 processing in talpid3 embryos that leads to high levels of activator compared with repressor.
Patterning of talpid3 neural tube and somite
Floor plate and associated Shh signaling is markedly reduced in talpid3 neural tube of stage 20 embryos. Shh transcripts and protein are expressed in talpid3 notochord, as in wild-type embryos, but are extremely patchy or completely absent in floor plate (Fig. 2, cf. A1,B1 and A2,B2). In wild-type embryos, floor plate is also characterized by Shh-dependent HNF3/FoxA2 expression (Fig. 2D1), which is either very reduced or absent in talpid3 (Fig. 2D2). High-level expression of Ptc1 or Ptc2 genes, considered to be a direct readout of Shh signaling (Pearse et al. 2001), is lost in talpid3 (Fig. 2, cf. C1 and C2; see Lewis et al. 1999a).
We characterized dorsoventral patterning of the neural tube in talpid3 embryos, investigating markers of chick neuronal identity that are established by Hh signaling, (Persson et al. 2002). We used embryos at stage 20 when the mutant phenotype is readily observable. In talpid3 expression of ventral markers, Nkx2.2 (p3 progenitor cells) (Fig. 2E2) and Islet2 (MNv) (Fig. 2G2) is almost lost completely while expression of Islet1 (MN) (Fig. 2F2, arrows indicating residual expression) and Lim3 (MN and V2) (Fig. 2H2, arrow indicating residual expression) is either lost or expressed in only a few ventrally located midline cells. Expression of intermediate markers, Pax6 (Fig. 2I2) and Dbx2 (Fig. 2J2) extends into the ventralmost region of the neural tube, and Lim1/2-positive cells, normally located in a stereotypical pattern in the intermediate-dorsal region (Fig. 2K1), are found more ventrally and wild-type pattern is lost (Fig. 2K2). En1 expressing cells (V1) (Fig. 2L1, bracket) are also shifted ventrally and distributed more widely (Fig. 2L2, bracket). Expression of dorsal markers Gsh1 and Pax7 is expanded ventrally (Fig. 2M2, N2) with Pax7, normally expressed in dorsal third (Fig. 2N1), being expressed in dorsal half of talpid3 neural tube (Fig. 2N2) and the band of Gsh1 expression being broader (Fig. 2M2, bracket).
In summary, in talpid3 there is a loss and/or reduction of ventral genes (talpid3, Fig. 2O, regions A and B) accompanied by ventral expansion of genes normally expressed in dorsal neural tube, suggesting a loss of Hedgehog function and failure of Gli activator. Expansion of intermediate neural (talpid3, Fig. 2O, region C) tube genes more ventrally suggests a loss of Gli repressor.
We examined dorsoventral patterning of somites, which is controlled by Shh from notochord and floor plate in the chick, mouse, and zebrafish (Munsterberg et al. 1995; Borycki et al. 1999; Lewis et al. 1999b). As in neural tube, there is low-level Ptc2 expression throughout talpid3 somites (Fig. 2C2, arrow). In talpid3, Pax1 is expressed more widely than in wild type (Fig. 3A) and positive cells are found displaced dorsally between myotome and dermomyotome (Fig. 3B, arrow) instead of being clearly demarcated from myotome expressing MyoD. The dermomyotome expresses Pax3 (Fig. 3C) and double in situ hybridization showed that, in talpid3 rostral somites, epithelial cells in dermomyotome express MyoD (Fig. 3D, arrow). Expansion of MyoD into more dorsal regions can be mimicked by overexpression of Shh in wild-type somites (Fig. 3E, arrow). Thus expansion and shift of cell types into more dorsal domains in talpid3 somites seems to resemble a gain of Hh function. To establish whether the effects of the mutation are autonomous, undifferentiated somites taken from the tail region of a stage 20 HH talpid3 embryo were implanted into stage 10–12 HH wild-type hosts. Widespread expression of MyoD in talpid3-derived somites was still seen (Fig. 3F, arrow) even in a wild-type environment. This suggests that gain of function in somite, as in limb, is more likely due to a defect in Hh signaling in responding cells than to ectopic ligand.
Rescue of talpid3 neural tube pattern
Failure to produce ventral domains of progenitor neurons in talpid3 neural tube could be due to reduced Shh protein because there is no floor plate. We first tested this by implanting Shh-soaked beads or Shh-expressing cells into talpid3 neural tube and examining Islet1 expression. Twenty-four hours later, there was no rescue of Islet1 expression in talpid3 embryos (9/9) (Supplementary Fig. S6J2). Further experiments in talpid3 embryos showed that application of a Shh bead to the neural tube did not induce Nkx2.2 or high-level Ptc1 expression or alter either Pax6 or Pax7 expression (Supplementary Fig. S6I2,K2,L2,M) as is seen in wild-type embryos treated with Shh (for details, see Fig. 7, below). These data show that Shh ligand cannot rescue dorsoventral patterning of the talpid3 neural tube and that a reduction in floor plate does not account for loss of ventral neurons.
Inability to rescue neural tube patterning with Shh suggests that talpid3 cells cannot transduce the Shh signal and that GliA does not function properly. To test this hypothesis directly, we electroporated neural tubes of stage 20 HH talpid3 and wild-type embryos with an activated Gli construct that can constitutively translocate to the nucleus (pCAGGS-Gli3AHIGH) (Stamataki et al. 2005) and monitored dorsoventral gene expression. Cell-by-cell analysis showed that Nkx2.2 (Fig. 4E1, E2), Islet1 (Fig. 4F1, F2, arrows), and Lim3 expression (data not shown) was recovered in talpid3 embryos while expression of Lim1/2 (data not shown), Pax6 (Fig. 4G1, G2, arrows), and Pax7 (Fig. 4H1, H2, arrow) was reduced in cells expressing the pCAGGS-Gli3AHIGH construct (2/2). In wild-type embryos, ectopic expression of Nkx2.2 (Fig. 4A1, A2, arrow), Islet1 (Fig. 4B1, B2, top arrow), and Lim3 (data not shown) was also induced and Pax6 (Fig. 4C1, C2, arrow) and Pax7 (Fig. 4D1, D2, arrow) expression reduced. Electroporation of control pCAGGS-eGFP into neural tube of either talpid3 or wild-type embryos produced no gene expression changes (data not shown). These results demonstrate directly that supplying talpid3 cells with functional activated Gli rescues the defect.
Figure 4. Rescue of talpid3 neural tube with activated Gli constructs. (A–H) Stage 20 HH wild-type and talpid3 embryos electroporated with Gli3AHIGH (activated Gli construct). Green indicates expression of GFP in transfected cells; red indicates protein expression detected by immunohistochemistry. A1–H1 are overlays of GFP and immunostaining; A2–H2 are immunostaining alone. (A–D) Wild-type neural tube. Induction of Nkx2.2-positive cells (arrows, A1,A2); Islet1-positive cells (arrows, B1,B2). Down-regulation of Pax6 (arrow, C1,C2); Pax7 (D1,D2). (E–H) talpid3 neural tube. Induction of Nkx2.2-positive cells (E1,E2) and Islet1-positive cells (upper arrow, F1,F2). Down regulation of Pax6 (arrow, G1,G2); Pax7 (arrow, H1,H2).
Genetic mapping and identification of talpid3
Linkage analysis of 110 individual carriers mapped the TA3 locus close to markers ADL0298 and ADL0166 on chromosome 5. Mapping further markers refined the location to an interval containing COM0184-(5.1)-ADL0166-(7.3)-SIX4-(3.6)-DAAM1-(0.9)-TA3-(4.6)-OTX2(8.3)-ADL0298-(0.0)-ROS0330-(2.8)-BMP4 (genetic distances between markers in centimorgans). We constructed a detailed physical map of this region to define gene content and develop more markers (Fig. 5A). Using these markers, the location of the talpid3 mutation was reduced to an interval encoding five genes: KCNK16-DACT1-ENSGALG00000012025-TIMM9-ARID4A (Fig. 5B).
To look for sequence changes that might identify the talpid3 mutation we sequenced cDNA clones derived from wild-type and talpid3 embryos (Supplementary Table S1). Sequence analysis of talpid3 KIAA0586 cDNA revealed an insertion mutation of a single thymine residue (Fig. 5C), which was confirmed by sequencing genomic products from carrier and noncarriers (Supplementary Fig. S1). Translation from KIAA0586 in talpid3 is terminated nine residues after the thymine insertion at a stop codon (Fig. 5C); this frameshift mutation is predicted to produce a truncated protein of 366 amino acids compared with 1524 in wild type (Fig. 6A). A multispecies comparison of KIAA0586 orthologs (Supplementary Fig. S2) shows extensive homology across all 30 exons, and a number of highly conserved noncoding regions. The complete sequence of the chicken KIAA0586 mRNA is shown in Supplementary Figure S3. The translation initiation codon at position 137 was selected as the most likely based on the longest open reading frame and homology with other vertebrate orthologs (Supplementary Fig. S4). To detect KIAA0586 orthologs in other species, we combined information from conservation of sequence and gene order (Supplementary Figs. S4, S5). Complete conservation of gene order was found for DAAM1-DACT1-KIAA0586-TIMM9-ARID4A in all vertebrates examined. No homologs were detected in Ciona intestinalis, Drosophila melanogaster, or Caenorhabditis elegans. Using Codonml (PAML 3.13) we were able to estimate the KA/KS ratio for KIAA0586 (0.28), a measure of selection constraint (Yang and Nielsen 2000). For comparison, the KA/KS ratio for mammalian orthologous gene pairs (Rat Genome Sequencing Project Consortium 2004) was 0.09–0.11, most genes being under strong to moderate purifying selection. The KA/KS ratio for KIAA0596 indicates that this protein has been under reduced purifying selection and/or increased positive selection. No specific functional domains were predicted in the peptide sequence. Most highly conserved are three globular domains, coiled-coil regions, and a central region, with as yet, no known structural features, and least conserved is the C-terminal region (Fig. 6A).
Expression of talpid3 gene and protein and analysisof Gli3 translocation in mutant cells
KIAA0586 is ubiquitously expressed in wild-type embryos (stages 12–35 HH), consistent with the talpid3 mutation affecting multiple tissues (Fig. 6B). Using the program ProtComp, the KIAA0586 protein was predicted to be cytoplasmic. In order to determine cellular localization experimentally, we expressed either C-terminal-tagged KIAA0586-myc or N-terminal-tagged KIAA0586-HA in limb bud cells from wild-type chick embryos or in the chicken DFI cell line and visualized cellular localization by immunofluorescence. In all cases, labeling was confined to the cytoplasm (Fig. 6D,E), suggesting the talpid3 protein is a cytoplasmic protein.
To further investigate the role of KIAA0586 in Hh signaling, we compared intracellular localization of Gli3 in wild-type and talpid3 mutant chick limb bud cells. In one set of experiments, we isolated cytoplasmic and nuclear fractions from wild-type and mutant cells and then carried out Western blot analysis as before (Fig. 6F). We assessed the effectiveness of the fractionation by detecting tubulin and found this to be confined to the cytoplasmic fraction (Fig. 6F). Both full-length (190 kDa) and processed (83 kDa) Gli3 proteins were readily detected in nuclei isolated from wild-type chick limb bud cells but not in the cytoplasmic fraction. Both Gli3 forms were also detected in the nuclei of talpid3 cells, although full-length Gli3 could also be detected in the cytoplasm of mutant cells (Fig. 6F). Furthermore when the ratio of Gli3A/Gli3R was compared, it is evident that talpid3 cells from limb and head have a much higher ratio of Gli3A/Gli3R in the nucleus (Fig. 6G) than wild-type cells. We also expressed pGli3-myc in wild-type and talpid3 chick limb bud cells and visualized localization using immunofluorescence (Fig. 6H). Localization of Gli3-myc was rather variable but was seen in both nucleus and cytoplasm in both wild-type and talpid3 cells. These results reinforce the conclusion that the talpid3 protein is required for efficient processing of functional GliA and GliR but suggest that it is not required for nuclear translocation of these proteins.
Rescue of talpid3 neural tube by KIAA0586
To confirm that KIAA0586 is the talpid3 gene, we carried out complementation tests. In the first set of experiments, we tested whether KIAA0586 could rescue ectopic induction of dorsoventral markers by Shh in neural tube of stage 17–18 HH talpid3 embryos. Shh beads were inserted into the neural tube and a pCAGGS-KIAA0586 construct was coelectroporated with a pCAGGS-eGFP control plasmid and dorsoventral patterning examined after 24 h. As noted above, application of a Shh bead to a wild-type neural tube induces ectopic Nkx2.2 (Fig. 7A2, arrows), Islet1 (Fig. 7B2, arrows), shifts strong Pax6 expression dorsally (Fig. 7C2, arrow), induces Ptc1 throughout the neural tube, surrounding somites, and ectoderm (Fig. 7I) and reduces Pax7 expression ventrally (Fig. 7D2, arrow) while application of Shh to a talpid3 neural tube has no effect on Nkx2.2, Islet1, Pax6, or Pax7 and only weakly induces Ptc1 (Supplementary Fig. S6I2,J2, K2,L2,M). When wild-type tissue was electroporated with pCAGGS-KIAA0586 and treated with Shh, a change in gene expression of Nkx2.2, Islet1, Pax6, Pax7, or Ptc1 was observed on both sides of the neural tube, whether electroporated or not, and therefore attributable only to the effect of the Shh bead (Fig. 7A–D). However, when talpid3 embryos were provided with both Shh and KIAA0586, ectopic expression of Nkx2.2 was induced only on the electroporated side of the neural tube (2/2) (Fig. 7E1–E3, arrows) and in some cells, expression of Pax6 (2/2) (Fig. 7G1, G2, asterisk) and Pax7 was also reduced on the electroporated side (2/2) (Fig. 7H1, H2, asterisk) in a way similar to that seen with Gli3HIGH. Islet1 expression remained unchanged (2/2) (Fig. 7F1, F2). Application of Shh with KIAA0586 also induced high-level Ptc1 expression in talpid3 dorsal neural tube on the electroporated side (Fig. 7J; asterisk denotes electroporated side). In preliminary experiments, one talpid3 embryo had a reduction in Pax7 when electroporated with pCAGGS-KIAA0586 (1/1) (Supplementary Fig. S6H1,H2) but Nkx2.2, Islet1, and Pax6 remained unchanged (Supplementary Fig. S6E–G), and there was no effect on expression of Nkx2.2, Islet1, Pax6, or Pax7 in a wild-type embryo electroporated with pCAGGS-KIAA0586 at stage 17 HH (1/1) (Supplementary Fig. S6A–D).
In a second set of experiments, we tested whether whether KIAA0586 alone could rescue endogenous dorsoventral patterning of the neural tube in very early talpid3 embryos. In order to carry out these experiments, we had to electroporate batches of embryos from the talpid3 flock at stages 12–14, well before the mutant phenotype is observable. We electroporated the pCAGGS-KIAA0586 construct, together with pCAGGS-RFP and fixed embryos at stage 20HH in order to examine gene expression. At this stage, we could usually identify mutant embryos by their phenotype, and we confirmed their identity by genotyping. In talpid3 embryos, we found very striking induction of Islet1 expression (3/3) (Fig. 7K1/2, green cells) coupled with repression of Pax6 (2/2) (Fig. 7L1/2, arrow). We also have preliminary evidence that Pax7 expression was reduced in talpid3 (1/1) (data not shown). This rescue of endogenous target gene expression provides good evidence that KIAA0568 is the talpid3 gene and has an essential role in Hh signal transduction.
Discussion
Hh signaling plays crucial roles in development and disease, and therefore it is important to understand the signal transduction pathway that mediates Hh signaling. Here we identify a novel component of the signaling pathway in vertebrates and provide evidence that it is necessary for the function of both Gli repressor and Gli activator forms. By a positional cloning strategy, exploiting the recently assembled chick genome, we provide evidence that KIAA0586 is the gene responsible for the talpid3 phenotype. Analysis of KIAA0586 indicates that it contains coiled-coil domains and extensive regions of intrinsic protein disorder (Fig. 6A), characteristic of proteins that play a role in cell signaling and gene regulation (Iakoucheva et al. 2002). Moreover, a frameshift mutation in KIAA0586 is present in talpid3 mutants, and the expression of a wild-type KIAA0586 cDNA in the neural tube of talpid3 embryos rescued the Shh responsiveness of neural cells and endogenous dorsoventral patterning. We therefore propose that KIAA00586 be renamed Talpid3.
Our analysis of Hh target gene expression in limb, face, neural tube, and somite of talpid3 embryos, together with our findings that expression of target genes is ligand independent—insensitive to either presence or absence of Shh—and can be rescued by expressing activated Gli are consistent with the idea that both GliA and GliR functions are compromised in talpid3. Thus the Talpid3 gene product plays an essential role in both GliA and GliR function in the Hh pathway. The paradoxical talpid3 phenotype, an apparent mixture of gain and loss of Hh function, is explained by the fact that GliR plays a predominant role in limb and somite, whereas GliA function is more important in face and neural tube patterning.
Our data show that Gli3 processing is impaired in talpid3 mutant cells, leading to the presence of both full-length Gli3A and Gli3R. Levels of Gli3A, in particular, are markedly increased. Although changes in processing are likely to be the main cause, changes in protein longevity and transcriptional levels could also contribute. The net result is that the ratio of Gli3A/Gli3R is much greater in talpid3 mutant cells. Abnormal processing is found not only in cells from the limb but also in cells from the trunk, which includes neural tube and somite. We have also shown that, as predicted by its structure, Talpid3 protein is localized in the cytoplasm. Thus, our results are consistent with Talpid3 interacting with other members of the Hh signaling complex to regulate processing of Gli proteins in the cytoplasm. Given the high levels of full-length Gli3, it is not clear why GliA function, in addition to GliR function, is compromised in talpid3. This is even more puzzling since we have also shown that the absence of Talpid3 protein does not prevent the ability of Gli3 proteins including Gli3 activator to enter the nucleus. The most likely possibility is that the pathway is very sensitive to the precise levels of full-length and short forms of Gli proteins such that, in the talpid3 mutant, the levels of GliA and GliR counterbalance their effects, resulting in neither activation nor repression. Alternatively, the Gli3 proteins produced in the absence of Talpid3 may be unable to regulate transcription due to, as yet, unknown post-translational modifications.
Recently two other genes involved in Hh signaling, Dzip1 (Sekimizu et al. 2004; Wolff et al. 2004) (chicken ortholog ENSGALG00000016895) and MTSS1 (Callahan et al. 2004) (also called MIM/BEG4, chicken ortholog ENSGALG00000016333), have been described. Dzip1 was identified as the Zebrafish mutant gene in the iguana mutant, which shows many similarities to the talpid3 mutant, including reduction of expression of Shh-dependent genes in neural tube coupled with a gain of Hh signaling in the somite. Furthermore, as in talpid3, iguana cannot be rescued through manipulation of Shh ligand, leading to the suggestion that either both GliA and GliR functions are reduced (Sekimizu et al. 2004), similar to our proposal for talpid3, or that GliR fails because of constitutive low-level Gli1 activation (Wolff et al. 2004). There is evidence that Dzip1 can shuttle between cytoplasm and nucleus, and it has been suggested that Dzip1 may affect nuclear import of Hh pathway proteins. This contrasts with Talpid3, which appears to be confined to the cytoplasm and is not required for nuclear import of Gli3. Interestingly, constitutive activation of Shh signaling has been suggested to occur in the talpid2 chicken mutant (Chuang and McMahon 1999), in which high levels of Gli3A are also found (Wang et al. 2000). Talpid2, like talpid3, is polydactylous, and the many digits are unpatterned, but, unlike talpid3, there is high-level Ptc expression throughout the limb (Carrucio et al. 1999), suggesting that Gli activator is functional in this mutant. Thus similarities in phenotype could arise by quite different mechanisms.
A number of intraflagellar transport proteins have been shown to play an essential role in vertebrate Hh signal transduction (Huangfu et al. 2003) in mouse, and a recent report suggests a functional link between Hh signaling and cilia at the level of the transmembrane protein smoothened which is required for GliA function (Corbit et al. 2005). Analysis of hypomorphic mouse mutants of two intraflagellar transport proteins reveals striking parallels between these mutants and talpid3 (Liu et al. 2005). Not only are the alterations in gene expression patterns in limb and neural tube similar to those seen in talpid3 and insensitive to Hh ligand, but also high levels of full-length Gli3 were detected in mutant cells, giving rise to a marked increase in the Gli3A/Gli3R ratio. Like Talpid3, intraflagellar transport proteins appear to be required for both GliA and GliR functions. It will be important to determine why mutations in intraflagellar transport proteins compromise GliR function in addition to GliA function and to explore the relationship between Talpid3 and cilia. Talpid3 joins a growing list of proteins, including these intraflagellar transport proteins, that regulate Hh transduction specifically in vertebrates (Chuang and McMahon 1999; Eggenschwiler et al. 2001) and is the first component in this pathway to be discovered in chickens.
Materials and methods
Embryo manipulations
Eggs were incubated for 2.5 d at 38°C, then windowed to assess development and classified as either mutant or wild type and reincubated until the desired stage. All limb manipulations were carried out on the right wing, leaving the left wing as a control.
Cyclopamine treatment: One-hundred microliters of 5 mM cyclopamine in 95% EtOH diluted with 200 μL DMEM was sonicated for 10 min; for the control solution, 100 μL of 95% EtOH was diluted in 200 μL DMEM. Solutions were injected with a fine glass micropipette into the amniotic sac over the right wing of stage 20 HH embryos; eggs were reincubated for 24 h.
Removal of Shh-expressing cells: At stage 16 HH, tissue was removed from flank, immediately lateral to somites, from axial level somites 22 and 18/17 using a sharpened tungsten needle; eggs were then reincubated for 18 h.
RCAS virus production, microinjection, and somite grafts
RCAS virus production and microinjection were done as per Morgan and Fekete (Morgan and Fekete 1996). In ovo surgical experiments were performed as per Schmidt et al. (Schmidt et al. 2000), and eggs were reincubated for 40 h.
Tissue graft implants—neural tube
QT-6 Shh expressing cells were grown until confluent. Cell sheets were scraped from the dish surface. A small piece of sheet was folded and tucked into a slit made in the neural tube at the level of the hind limbs using a tungsten needle; eggs were reincubated for 24 h.
Bead implants—neural tube
CM Affi-Gel Blue Beads (Bio-Rad) soaked in Shh protein as per Drossopoulou et al. (Drossopoulou et al. 2000) were picked up on the point of a tungsten needle and placed into the neural tube via a dorsal slit.
Electroporation
DNA was prepared using a Qiagen Endotoxin-Free Maxiprep Kit, resuspended in endotoxin-free water, and diluted with 0.25% Fast Green. Plasmid pCAGGS-Gli3AHIGH was used at a concentration of 7 μg/μL, pCAGGS-GFP at 7.5 μg/μL, pCAGGS-RFP at 0.1 μg/μL, and pCAGGS-KIAA0586 at 1 μg/μL. DNA was injected into the neural tube and embryos were immediately electroporated at 30 V for 5 x 50 msec with 5 x 50–100-msec intervals or at 20 V for 5 x 15 msec (stages 12–14). Embryos were reincubated for 20–48 h. Successful electroporation was assessed by expression of GFP, eGFP, or RFP using a Leica Fluorescence Dissection Microscope. The GFP section of neural tube was then dissected out in cold PBS (pH 7.4) and fixed for sectioning for immunohistochemistry or whole-mount in situ hybridization.
Whole-mount RNA in situ hybridization
For details, see the Supplemental Material.
Whole-mount immunohistochemistry
For details, see the Supplemental Material.
Section immunohistochemistry
For details, see the Supplemental Material.
Western blotting
For determining levels of Shh protein, stage 20 HH embryo limb buds were dissected into three equal parts along the anterior–posterior axis in cold PBS, lysed in RIPA buffer/1:25 Proteases Inhibitor (Roche), samples ran on 12% SDS-PAGE, and protein detected with Shh antibody. Lysates were spun at 13,000 rpm for 5 min at room temperature, supernatant was removed and methanol precipitated and then dried using the Eppendorf Concentrator. The pellet was resuspended in 30 μL of Sample Buffer 1/10 -mercaptoethanol and then denatured. Gels were subsequently wet or dry blotted in 1x Tris-Glycine/20%Methanol buffer for 1–2 h at 172 mA onto nitro-cellulose, which was blocked in 5% Marvel/PBS 1:1000 Tween-20; incubated with 1:2 antibody and Marvel/PBT overnight at 4°C; washed 3 h in PBT; incubated in anti-mouse HRP-conjugated (Sigma) in Marvel/PBT 4 h; and washed 1 h in PBT, then 30 min in PBS. Detection was with Supersignal West Pico Chemiluminescent Substrate (Pierce) for 5 min and was detected with X-ray film exposed overnight.
Levels of full-length and short forms of Gli3 protein were determined in protein extracts from limb buds and trunks dissected from stage HH24 embryos. Some limb buds were dissected into three equal parts as above. All tissues were then lysed in buffer containing 1% NP-40, 150 mM NaCl, and 20 nM Tris (pH 7.4) in the presence of Complete protease inhibitor cocktail (Roche). Protein concentrations were measured using Bradford (Pierce), and equal amounts of protein were separated by SDS-PAGE using 4%–12% precast gels from Invitrogen. Immunoblot was performed using a polyclonal antibody against Gli3 (Santa Cruz Biotechnology, Inc).
Nuclear and cytosol fractionation
Limbs and heads from stage 24HH embryos were dissected out in cold PBS and cut into small pieces with scissors. Homogenization of tissue was carried out by a pestle on ice. Extraction of nuclear and cytoplasmic fractions was performed using a Compartmental Protein extraction Kit (Chemicon) as per the manufacturer’s instruction. Protein concentration was determined by the Bradford method. Nuclear and cytoplasmic fractions were stored at –80°C for Western blot analysis.
Cell cultures and immunohistochemistry
DFI chicken cells were grown in DMEM + 10% fetal calf serum; constructs containing either KIAA0586-myc or KIAA0586-HA were introduced using FuGene (Roche).
Primary cell culture and immunofluorescence staining
Individual chick limbs from HH20–24 embryos were dissected in sterile cold PBS and incubated with 1x Trypsin and EDTA for <5 min at 37°C. A 20-μL cell suspension was plated onto cover glasses coated by fibronectin. The cover glasses were set in a Petri dish and incubated at 37°C for an hour followed by flooding cells with 1:1, DMEM-HAMS F12/10% FCS, 1% PSF, 1% L-glutamine. The next day, cells were transfected with myc-hGli3 or KIAA0586-myc or KIAA0586-HA using FuGene 6 (Roche).
After 24 h transfection, cells were washed in PBS, fixed in ice-cold methanol for 10 min, then blocked in 10% normal serum in PBS for 30 min. Cells were incubated in monoclonal anti-c-myc (1:1000, Sigma) 1 h at room temperature, washed 3 x 5 min in 0.2% Tween-20 in PBS, and incubated in anti-mouse IgG conjugated with Alexa Fluor 610 (1:1000, Molecular Probes) for 1 h at room temperature. After washes (3 x 5 min in 0.2% Tween-20 in PBS), cells were mounted in VectaShield Hard*set mount medium with DAPI (Vector Laboratories, H-1500). Fluorescence was observed using a fluorescence microscope or a Leica SP2 confocal microscope.
Animals and genetic linkage analysis
Talpid3 carriers were maintained by outcrossing with ISA Brown chickens. Line 6, an inbred White Leghorn line, was kindly provided by the Institute of Animal Health (Compton). A male carrier (TA3/N) was crossed to eight Line 6 females (N/N) and offspring of this backcross were test-crossed with known carriers. A carrier was confirmed if two or more talpid3 embryos were found from a test cross. At least 20 wild-type embryos were collected before any noncarriers were confirmed. Four DNA pools were created from backcross offspring: Carrier male (n = 13), carrier female (n = 32), noncarrier male (n = 7), and noncarrier female (n = 16) were genotyped using 129 autosomal genetic markers, covering 70% of the genome. Markers ADL0298 and ADL0166 on chromosome 5 showed a significant difference (p < 0.00001) between carrier and noncarrier pools. Further markers near these loci were used to genotype 110 individuals and refine the location of the TA3 locus. Two-point and multipoint linkage analysis was performed using CRIMAP (Green et al. 1990). From this analysis, 30 recombinant offspring were identified near TA3 between markers ADL0166 and ADL0298 and typed using markers developed from EST and genomic sequences. Using known chicken genes and reference to the Ensembl (http://www.ensembl.org) and University of California at Santa Cruz (http://www.genome.ucsc.edu) genome browsers, we identified the orthologous region on human chromosome 14. Using human gene sequences and Blast we searched databases (http://www.ncbi.org) for homologous chicken EST and genomic sequences. We developed PCR-based markers from these showing size and SSCP or SNP variation. These markers were used to isolate and order individual BAC clones (Burt et al. 2003) to create a detailed physical map. BAC-end and full-length cDNA sequences were also used to create more genetic markers. Using such a high density of markers we were able to identify specific haplotypes from recombinants in this region and map the TA3 locus to KCNK16-DACT1-KIAA0586-TIMM9-ARID4A. Genotyping was performed as described by Burt et al. (2003) and primers used for genotyping are listed in Supplementary Table S2.
Sequence analysis and characterization of talpid3 mutation
For details, see the Supplemental Material.
Genotyping
Genomic DNA was prepared from extra-embryonic membranes using Quigen DNAeasy Kit. A 957-base-pair (bp) region of the talpid3 gene was then amplified via PCR (forward: CTTCTCT TGCCCTGCCTTAC; reverse: TTGAAGCTGCATTTCCA CAG), annealed at 60°C (Thermo electron) using Fast Start Taq Polymerase and GC-rich solution (Roche). The PCR product was then digested using Pst1 (Roche). A single 957-bp band indicates a wild-type embryo, two bands (721 bp and 236 bp) indicate a talpid3 mutant, and three bands indicate a heterozygous embryo.
Cloning of chicken cDNA clones and gene constructs
RNA was extracted using TRIzol reagent (Invitrogen), and a DACT1 cDNA was cloned via RT–PCR (P1: GGGATGCTG GCCTATTAAATC in exon 2; P2: GCTTCCCAAAG CAGAAACAG, in 3' UTR) from embryo-derived mRNA (HH stage 21) using the TOPO TA cloning kit (Invitrogen), then subcloned into pGEM (Invitrogen) for preparation of antisense RNA probes. Rapid amplification of 5' cDNA ends of KIAA0586 and DAAM1 was carried out using the First Choice RLM-RACE Kit (Ambion) and cloned using the TOPO TA cloning kit. Gene-specific primers are listed in Supplementary Table S2.
Bioinformatics analyses
For details, see the Supplemental Material.
Accession numbers
All sequences have been deposited in GenBank under accession numbers DQ066927 [GenBank] –DQ066935 [GenBank] and CZ550237 [GenBank] –CZ550250 [GenBank] .
Acknowledgments
We are grateful to Dr. Jun-ichi Miyazaki and Dr. Toshihiko Agura for providing pCAGGS-eGFP, Dr. Ruiz i Altaba for plasmid myc-hGli-3, Dr. R. Johnson and Dr. C. Tabin for pRCAS-Shh, Professor Groenen for chicken BAC clones, and Dr. Placzek for Shh antibody. EST clones from UK Chick EST Project were provided by ARK-Genomics (http://www.ark-genomics.org). This research was supported by the BBSRC (to M.D., I.R.P., Y.Y., D.R.M., F.K.B., D.R.M., D.B., and C.T.), British Heart Foundation (to M.D. and C.T.), MRC (UK; to T.G.S., M.T. and C.T.), and The Royal Society (to T.G.S. and C.T.). D.S. and J.B. were supported by the MRC (UK) and an EC network grant Stembridge (QLG3-CT-2002-011441); A.E.M. and M.S. were supported by The Wellcome Trust and MYORES (EU FP6 NoE).
References
Bai C.B., Auerbach W., Lee J.S., Stephen D., Joyner A.L. 2002. Gli2, but not Gli1, is required for initial Shh signaling and ectopic activation of the Shh pathway. Development 129: 4753–4761.
Borycki A.G., Brunk B., Tajbakhsh S., Buckingham M., Chiang C., Emerson C.P.J. 1999. Sonic hedgehog controls epaxial muscle determination through Myf5 activation. Development 126: 4053–4063.
Burt D.W., Morrice D.R., Lester D.H., Robertson G.W., Mohammed M.D., Simmons J., Downey L.M., Thuang C., Bridges L.R., Paton I.R. et al. 2003. Analysis of rdd locus in chicken: A model for human retinitis pigmentosa. Mol. Vis. 9: 164–170.
Buxton P., Davey M.G., Paton I.R., Morrice D.R., Francis-West P.H., Burt D.W., Tickle C. 2004. Craniofacial development in the talpid3 chicken mutant. Differentiation 72: 348–362.
Callahan C.A., Ofstad T., Horng L., Wang J.K., Zhen H.H., Coulombe P.A., Oro A.E. 2004. MIM/BEG4, a Sonic hedgehog-responsive gene that potentiates Gli-dependent transcription. Genes & Dev. 18: 2724–2729.
Carrucio N.C., Martinez-Lopez A., Harris M., Dvorak L., Bitgood J., Simandl B.K., Fallon J.F. 1999. Constitutive activation of sonic hedgehog signaling in the chicken mutant talpid2, Shh independent outgrowth and polarizing activity. Dev. Biol. 212: 137–149.
Chuang P.T. and McMahon A.P. 1999. Vertebrate Hedgehog signaling modulated by induction of a Hedgehog-binding protein. Nature 397: 617–621.
Corbit K.C., Aanstad P., Singla V., Norman A.R., Stainier D.Y., Reiter J.F. 2005. Vertebrate Smoothened functions at the primary cilium. Nature 437: 1018–1021.
Drossopoulou G., Lewis K.E., Sanz-Ezquerro J.J., Nikbakht N., McMahon A.P., Hofmann C., Tickle C. 2000. A model for anteroposterior patterning of the vertebrate limb based on sequential long- and short-range Shh signaling and Bmp signaling. Development 127: 1337–1348.
Ede D.A. and Kelly W.A. 1964a. Developmental abnormalities in the head region of the talpid mutant of the fowl. J. Embryol. Exp. Morphol. 12: 161–182.
Ede D.A. and Kelly W.A. 1964b. Developmental abnormalities in the trunk and limbs of the fowl. J. Embryol. Exp. Morphol. 12: 339–356.
Eggenschwiler J.T., Espinoza E., Anderson K.V. 2001. Rab23 is an essential negative regulator of the mouse Sonic hedgehog pathway. Nature 412: 194–198.
Francis-West P.H., Robertson K.E., Ede D.A., Rodriguez C., Izpisua-Belmonte J.C., Houston B., Burt D.W., Gribbin C., Brickell P.M., Tickle C. 1995. Expression of genes encoding bone morphogenetic proteins and sonic hedgehog in talpid (ta3) limb buds: Their relationships in the signaling cascade involved in limb patterning. Dev. Dyn. 203: 187–197.
Frank-Kamenetsky M., Zhang X.M., Bottega S., Guicherit O., Wichterle H., Dudek H., Bumcrot D., Wang F.Y., Jones S., Shulok J. et al. 2002. Small-molecular modulators of Hedgehog signaling: Identification and characterization of Smoothened agonists and antagonists. J. Biol. 1: 10.
Green P., Falls K., Crooks S. In Documentation for CRIMAP, version 2.4. . 1990. Washington School of Medicine, St. Louis, MO.
Hamburger V. and Hamilton H.L. 1951. A series of normal stages in the development of the chick embryo. J. Morphol. 88: 49–92.
Hinchliffe J.R. and Ede D.A. 1968. Abnormalities in bone and cartilage development in the talpid mutant of the fowl. J. Embryol. Exp. Morphol. 19: 327–339.
Huangfu D., Liu A., Rakeman A.S., Murcia N.S., Niswander L., Anderson K.V. 2003. Hedgehog signaling in the mouse requires intraflagellar-transport proteins. Nature 426: 83–87.
Hui C.C. and Joyner A.L. 1993. A mouse model of greig cephalopolysyndactyly syndrome: The extra-toesJ mutation contains an intragenic deletion of the Gli3 gene. Nat. Genet. 3: 241–246.
Hunton P. 1960. "A study of some factors affecting the hatchibility of chicken eggs, with special reference to genetic control." MSc. thesis. Wye College, University of London, UK.
Iakoucheva L.M., Brown C.J., Lawson J.D., Obradovic Z., Dunker A.K. 2002. Intrinsic disorder in cell-signaling and cancer-assoicated proteins. J. Mol. Biol. 323: 573–584.
Ingham P.W. and McMahon A.P. 2001. Hedgehog signaling in animal development, paradigms and principles. Genes & Dev. 15: 3059–3087.
International Chicken Genome Sequencing Consortium . 2004. Sequence and comparative analysis of the chicken genome provide unique perspectives on vertebrate evolution. Nature 432: 695–716.
Izpisua-Belmonte J.C., Ede D.A., Tickle C., Duboule D. 1992. The mis-expression of posterior Hox-4 genes in talpid (ta3) mutant wings correlates with the absence of anteroposterior polarity. Development 114: 959–963.
Lewis K.E., Drossopoulou G., Paton I.R., Morrice D.R., Robertson K.E., Burt D.W., Ingham P.W., Tickle C. 1999a. Expression of ptc and gli genes in talpid3 suggests bifurcation in Shh pathway. Development 126: 2397–2407.
Lewis K.E., Currie P.D., Roy S., Schauerte H., Haffter P., Ingham P.W. 1999b. Control of muscle cell-type specification in the zebrafish embryo by Hedgehog signaling. Dev. Biol. 216: 469–480.
Litingtung Y., Dahn R.D., Li Y., Fallon J.F., Chiang C. 2002. Shh and Gli3 are dispensable for limb skeleton formation but regulate digit number and identity. Nature 418: 979–983.
Liu A., Wang B., Niswander L.A. 2005. Mouse intraflagellar transport proteins regulate both the activator and repressor functions of Gli transcription factors. Development 132: 3103–3111.
Morgan B.A. and Fekete D.M. 1996. Manipulating gene expression with replication-competent retroviruses. Methods Cell Biol. 51: 185–218.
Munsterberg A.E., Kitajewski J., Bumncrot D.A., McMahon A.P., Lassar A.B. 1995. Combinatorial signaling by Sonic hedgehog and Wnt family members induces myogenic bHLH gene expression in the somite. Genes & Dev. 9: 2911–2922.
Pearse R.V., Vogan K.J., Tabin C.J. 2001. Ptc1 and Ptc2 transcripts provide distinct readouts of Hedgehog signaling activity during chick embryogenesis. Dev. Biol. 239: 15–29.
Persson M., Stamataki D., te Welscher P., Andersson E., Bose J., Ruther U., Ericson J., Briscoe J. 2002. Dorso-ventral patterning of the spinal cord requires Gli3 transcriptional repressor activity. Genes & Dev. 16: 2865–2878.
. Rat Genome Sequencing Project Consortium. 2004. Genome sequence of the Brown Norway rat yields insights into mammalian evolution. Nature 428: 493–521.
Ruiz i Altaba A., Nguyen V., Palma V. 2003. The emergent design of the neural tube, pre-pattern, SHH morphogen and GLI code. Curr. Opin. Genet. Dev. 13: 513–521.
Schmidt M., Tanaka M., Munsterberg A. 2000. Expression of -catenin in the developing chick myotome is regulated by myogenic signals. Development 127: 4105–4113.
Sekimizu K., Nishioka N., Sasaki H., Takeda H., Karlstrom R.O., Kawakami A. 2004. The Zebrafish iguana locus encodes Dzip1, a novel zinc-finger protein required for proper regulation of Hedgehog Signaling. Development 131: 2521–2532.
Stamataki D., Ulloa F., Tsoni S.V., Mynett A., Briscoe J. 2005. A gradient of gli activity mediates graded Sonic Hedgehog signaling in the neural tube. Genes & Dev. 19: 626–641.
te Welscher P.T., Zuniga A., Kuijper S., Drenth T., Goedemans H.J., Meijlink F., Zeller R. 2002. Progression of vertebrate limb development through SHH-mediated counteraction of GLI3. Science 298: 827–830.
Wang B., Fallon J.F., Beachy P.A. 2000. Hedgehog-regulated processing of Gli3 produces an anterior/posterior repressor gradient in the developing vertebrate limb. Cell 100: 423–434.
Wolff C., Roy S., Lewis K.E., Schauerte H., Joerg-Rauch G., Kim A., Weiler C., Geisler R., Haffter P., Ingham P. 2004. Iguana encodes a novel zinc-finger protein required for proper regulation of Hedgehog Signaling. Development 131: 2521–2532.
Yang Z. and Nielsen R. 2000. Estimating synonymous and nonsynonymous substitution rates under realistic evolutionary models. Mol. Biol. Evol. 17: 32–43.(Megan G. Davey1,2, I. Rob)
Abstract
Talpid3 is a classical chicken mutant with abnormal limb patterning and malformations in other regions of the embryo known to depend on Hedgehog signaling. We combined the ease of manipulating chicken embryos with emerging knowledge of the chicken genome to reveal directly the basis of defective Hedgehog signal transduction in talpid3 embryos and to identify the talpid3 gene. We show in several regions of the embryo that the talpid3 phenotype is completely ligand independent and demonstrate for the first time that talpid3 is absolutely required for the function of both Gli repressor and activator in the intracellular Hedgehog pathway. We map the talpid3 locus to chromosome 5 and find a frameshift mutation in a KIAA0586 ortholog (ENSGALG00000012025), a gene not previously attributed with any known function. We show a direct causal link between KIAA0586 and the mutant phenotype by rescue experiments. KIAA0586 encodes a novel protein, apparently specific to vertebrates, that localizes to the cytoplasm. We show that Gli3 processing is abnormal in talpid3 mutant cells but that Gli3 can still translocate to the nucleus. These results suggest that the talpid3 protein operates in the cytoplasm to regulate the activity of both Gli repressor and activator proteins.
[Keywords: Gli; Hedgehog signaling; chicken; talpid3]
Received October 6, 2005; revised version accepted March 9, 2006.
The chicken embryo has served as a powerful and influential model system. The recent completion of the chicken genome (International Chicken Genome Sequencing Consortium 2004) allows new opportunities to exploit the ease of manipulating chick embryos to study vertebrate development. A number of chicken mutants have arisen spontaneously, often in agricultural flocks where they were noticed because of reduced hatchability. Talpid3 is one such mutant (Hunton 1960), and is a member of the classical developmental talpid group, so called because their paddle-shaped limbs resemble those of the mole (Talpa). In the talpid3 mutant, besides limb defects, there is an almost bewildering set of malformations including face, skeleton, and vascular defects (Ede and Kelly 1964a, b; Hinchliffe and Ede 1968). It is now clear that all these abnormalities are in regions of the embryo that depend on Hedgehog (Hh) signaling (Ingham and McMahon 2001).
An intriguing feature of talpid3 is that some aspects of the gross phenotype (e.g., polydactylous limbs) suggest a gain of Hh function whereas others (e.g., hypoteleorism in which the eyes are pulled together) suggest a loss of function. Analysis of Hh target gene expression in both limb and face shows that, even within one region, some targets are expressed at abnormally low levels while others are expressed ectopically (Izpisua-Belmonte et al. 1992; Francis-West et al. 1995; Lewis et al. 1999a; Buxton et al. 2004). These opposite changes in gene expression seen in talpid3 may reflect dual roles of the Gli transcriptional effectors of Hh signaling, which can act as either activators (A) or repressors (R) (Ruiz i Altaba et al. 2003). There are three Gli proteins, Gli1, Gli2, and Gli3. Experiments in mice show that Gli3 but not Gli1 or Gli2 is required for limb patterning (Bai et al. 2002). In normal chick limb buds, Sonic hedgehog (Shh) produced posteriorly diffuses across the bud and prevents processing of full-length Gli3 protein to a short repressor form (Gli3R). This results in low levels of full-length Gli3 protein that, together with other Gli proteins, can act as a transcriptional activator (Gli3A) and a gradient of Gli3R, highest anteriorly (Wang et al. 2000). Absence of high-level expression of some target genes in the posterior of talpid3 limb buds (e.g., Ptc1) and the inability to rescue their expression through the addition of Shh protein (Lewis et al. 1999a) suggests that GliA function is defective. Ectopic expression of other target genes in the anterior of talpid3 limb buds suggests that Gli3R levels are reduced anteriorly. The absence of high-level expression of Ptc, which encodes the Shh receptor and binds Shh, could lead to widespread diffusion of Shh ligand, reducing Gli3R levels. However, the limbs of Gli3–/– and Shh–/–, Gli3–/– mutant mice, like those of talpid3 chick embryos, are polydactylous, with many unpatterned digits, and also show ectopic expression of Hh target genes (Hui and Joyner 1993; Litingtung et al. 2002; te Welscher et al. 2002).
Here we show that ectopic gene expression in talpid3 limb buds is ligand independent and assess the distribution of Gli3A versus Gli3R. The dorsoventral pattern of the neural tube in chick has been shown to depend on a balance between GliA and GliR function (Persson et al. 2002; Stamataki et al. 2005). We explore whether the talpid3 mutation compromises both GliA and GliR function in dorsoventral patterning of the neural tube and adjacent somites. We rescue Gli function by electroporation into the talpid3 neural tube. We then combine knowledge of the chicken genome with the manipulability of chick embryos to identify the talpid3 gene and carry out functional complementation tests within the mutant neural tube to confirm identification unequivocally. We show that the talpid3 protein localizes to the cytoplasm and that although Gli3 processing is abnormal in mutant chick cells, both Gli3A and Gli3R can still translocate to the nucleus. We conclude that the talpid3 protein acts in the cytoplasm to regulate the functional activity of both the repressor and activator forms of Gli.
Results
Basis of the talpid3 limb polydactyly
The polydactylous phenotype of talpid3 limbs suggests a gain of Hh function. Comparison of stage 19 Hamburger Hamilton (HH) (Hamburger and Hamilton 1951) limb buds using immunohistochemistry showed that Shh protein is more widespread in talpid3 (Fig. 1A,B). In Western blots, more Shh protein was found in the middle third of talpid3 limb buds than in wild type (Fig. 1C). To test whether ectopic gene expression in talpid3 limb buds is dependent on this ectopic Shh protein, we attempted to rescue the mutant phenotype by inhibiting Shh signaling or removing its source; we monitored rescue by examining Hoxd13 expression, which is ectopically expressed throughout talpid3 limb buds instead of being posteriorly restricted (Izpisua-Belmonte et al. 1992). We applied cyclopamine, which inhibits Smo activity (Frank-Kamenetsky et al. 2002) to stage 17 HH wing buds. In 5/6 treated talpid3 wings, Hoxd13 expression was still strongly expressed (Fig. 1E; right wing bud) whereas, in 3/4 treated wild-type wings, expression was much reduced (Fig. 1D; right wing bud). Expression of Hoxd13 was retained in most cases in wings treated with ethanol (4/5 wild-type, 3/3 talpid3) (data not shown). We also removed future Shh expressing cells at stage 16 HH. This had no effect on Hoxd13 expression in talpid3 wing buds (3/3) (Fig. 1G; right wing bud) but, in wild type, expression was reduced or abolished (7/7) (Fig. 1F; right wing bud). These results show that, although Shh protein is more widely distributed in talpid3 limb buds, ectopic Hoxd13 expression is ligand independent.
Figure 1. Localization of Shh protein, effects of reducing Shh activity in wild-type and talpid3 limb buds, and status of Gli3 protein. (A,B) Localization of Shh protein in stage 19 HH wild-type (A) and talpid3 (B) limb buds. Dotted lines indicate limit of staining. (C) Western blot analysis of Shh protein in different regions (anterior, middle, posterior) of stage 20 HH wild-type (wt) and talpid3 (ta3) limbs. (D,E) Hoxd13 expression in wild-type (D) and talpid3 (E) right limbs treated with cyclopamine compare with control left limbs. (F,G) Hoxd13 expression in wild-type (F) and talpid3 (G) right limbs after removal of posterior mesenchyme; compare with control left limbs. (H,I) Status of Gli3 protein in different regions (anterior, a; middle, m; posterior, p) of stage 24 wild-type and talpid3 limb buds. (H) Immunoblot shows levels of Gli3A and Gli3R in wild-type and talpid3 legs (n > 3). (I) Histograms show levels of Gli3 proteins in wild-type and talpid3 wing buds; standard deviations shown, n = 3. (J) Immunoblot directly compares levels of Gli3 protein in wild-type and talpid3 limbs and trunk tissue. Numbers at top of each lane show ratios of Gli3A/Gli3R. H and J also illustrate tubulin loading controls.
Analysis of gene expression in talpid3 limb buds and the fact that these changes in expression are ligand independent suggest that both GliA and GliR functions are defective. In order to assess the levels of the different forms of Gli proteins, we extracted proteins from the anterior, middle, and posterior thirds of talpid3 buds (either wing or leg) and carried out Western blot analysis using an antibody against Gli3 that recognizes both the full-length form, Gli3A (190 kDa) and the processed short form, Gli3R (83 kDa). We then compared the levels of the two forms of Gli3 with those seen in wild-type limb buds by carrying out analysis on either wing buds or leg buds. We present the data in histograms for the wing buds and show the Western blot for the leg buds. In the wild-type limb buds, we detected both Gli3A and Gli3R forms (Fig. 1H,I) and, as previously reported (Wang et al. 2000), Gli3R is present in a gradient along the anterior–posterior axis with maximal levels of Gli3R in the anterior region of the bud (Fig. 1H,I). In contrast, in talpid3 wing buds (Fig. 1I), levels of Gli3A are strikingly increased and the polarized distribution of Gli3R is abolished. Western blots of Gli3 in talpid3 leg buds (Fig. 1H) also confirmed that Gli3R is not graded and, in addition, that Gli3A levels are strikingly elevated. We also carried out Western blot analysis on extracts from whole limb buds (both wing and leg pooled together) from wild-type and talpid3 embryos and estimated the Gli3A/Gli3R ratio (Fig. 1J). This analysis shows that this ratio is strikingly elevated in the mutant. Furthermore, we found that the ratio of Gli3A/Gli3R was also elevated in extracts from talpid3 trunk, which includes other tissues, such as neural tube and somites, that are patterned by Hh signaling. Thus there is a generalized defect in Gli3 processing in talpid3 embryos that leads to high levels of activator compared with repressor.
Patterning of talpid3 neural tube and somite
Floor plate and associated Shh signaling is markedly reduced in talpid3 neural tube of stage 20 embryos. Shh transcripts and protein are expressed in talpid3 notochord, as in wild-type embryos, but are extremely patchy or completely absent in floor plate (Fig. 2, cf. A1,B1 and A2,B2). In wild-type embryos, floor plate is also characterized by Shh-dependent HNF3/FoxA2 expression (Fig. 2D1), which is either very reduced or absent in talpid3 (Fig. 2D2). High-level expression of Ptc1 or Ptc2 genes, considered to be a direct readout of Shh signaling (Pearse et al. 2001), is lost in talpid3 (Fig. 2, cf. C1 and C2; see Lewis et al. 1999a).
We characterized dorsoventral patterning of the neural tube in talpid3 embryos, investigating markers of chick neuronal identity that are established by Hh signaling, (Persson et al. 2002). We used embryos at stage 20 when the mutant phenotype is readily observable. In talpid3 expression of ventral markers, Nkx2.2 (p3 progenitor cells) (Fig. 2E2) and Islet2 (MNv) (Fig. 2G2) is almost lost completely while expression of Islet1 (MN) (Fig. 2F2, arrows indicating residual expression) and Lim3 (MN and V2) (Fig. 2H2, arrow indicating residual expression) is either lost or expressed in only a few ventrally located midline cells. Expression of intermediate markers, Pax6 (Fig. 2I2) and Dbx2 (Fig. 2J2) extends into the ventralmost region of the neural tube, and Lim1/2-positive cells, normally located in a stereotypical pattern in the intermediate-dorsal region (Fig. 2K1), are found more ventrally and wild-type pattern is lost (Fig. 2K2). En1 expressing cells (V1) (Fig. 2L1, bracket) are also shifted ventrally and distributed more widely (Fig. 2L2, bracket). Expression of dorsal markers Gsh1 and Pax7 is expanded ventrally (Fig. 2M2, N2) with Pax7, normally expressed in dorsal third (Fig. 2N1), being expressed in dorsal half of talpid3 neural tube (Fig. 2N2) and the band of Gsh1 expression being broader (Fig. 2M2, bracket).
In summary, in talpid3 there is a loss and/or reduction of ventral genes (talpid3, Fig. 2O, regions A and B) accompanied by ventral expansion of genes normally expressed in dorsal neural tube, suggesting a loss of Hedgehog function and failure of Gli activator. Expansion of intermediate neural (talpid3, Fig. 2O, region C) tube genes more ventrally suggests a loss of Gli repressor.
We examined dorsoventral patterning of somites, which is controlled by Shh from notochord and floor plate in the chick, mouse, and zebrafish (Munsterberg et al. 1995; Borycki et al. 1999; Lewis et al. 1999b). As in neural tube, there is low-level Ptc2 expression throughout talpid3 somites (Fig. 2C2, arrow). In talpid3, Pax1 is expressed more widely than in wild type (Fig. 3A) and positive cells are found displaced dorsally between myotome and dermomyotome (Fig. 3B, arrow) instead of being clearly demarcated from myotome expressing MyoD. The dermomyotome expresses Pax3 (Fig. 3C) and double in situ hybridization showed that, in talpid3 rostral somites, epithelial cells in dermomyotome express MyoD (Fig. 3D, arrow). Expansion of MyoD into more dorsal regions can be mimicked by overexpression of Shh in wild-type somites (Fig. 3E, arrow). Thus expansion and shift of cell types into more dorsal domains in talpid3 somites seems to resemble a gain of Hh function. To establish whether the effects of the mutation are autonomous, undifferentiated somites taken from the tail region of a stage 20 HH talpid3 embryo were implanted into stage 10–12 HH wild-type hosts. Widespread expression of MyoD in talpid3-derived somites was still seen (Fig. 3F, arrow) even in a wild-type environment. This suggests that gain of function in somite, as in limb, is more likely due to a defect in Hh signaling in responding cells than to ectopic ligand.
Rescue of talpid3 neural tube pattern
Failure to produce ventral domains of progenitor neurons in talpid3 neural tube could be due to reduced Shh protein because there is no floor plate. We first tested this by implanting Shh-soaked beads or Shh-expressing cells into talpid3 neural tube and examining Islet1 expression. Twenty-four hours later, there was no rescue of Islet1 expression in talpid3 embryos (9/9) (Supplementary Fig. S6J2). Further experiments in talpid3 embryos showed that application of a Shh bead to the neural tube did not induce Nkx2.2 or high-level Ptc1 expression or alter either Pax6 or Pax7 expression (Supplementary Fig. S6I2,K2,L2,M) as is seen in wild-type embryos treated with Shh (for details, see Fig. 7, below). These data show that Shh ligand cannot rescue dorsoventral patterning of the talpid3 neural tube and that a reduction in floor plate does not account for loss of ventral neurons.
Inability to rescue neural tube patterning with Shh suggests that talpid3 cells cannot transduce the Shh signal and that GliA does not function properly. To test this hypothesis directly, we electroporated neural tubes of stage 20 HH talpid3 and wild-type embryos with an activated Gli construct that can constitutively translocate to the nucleus (pCAGGS-Gli3AHIGH) (Stamataki et al. 2005) and monitored dorsoventral gene expression. Cell-by-cell analysis showed that Nkx2.2 (Fig. 4E1, E2), Islet1 (Fig. 4F1, F2, arrows), and Lim3 expression (data not shown) was recovered in talpid3 embryos while expression of Lim1/2 (data not shown), Pax6 (Fig. 4G1, G2, arrows), and Pax7 (Fig. 4H1, H2, arrow) was reduced in cells expressing the pCAGGS-Gli3AHIGH construct (2/2). In wild-type embryos, ectopic expression of Nkx2.2 (Fig. 4A1, A2, arrow), Islet1 (Fig. 4B1, B2, top arrow), and Lim3 (data not shown) was also induced and Pax6 (Fig. 4C1, C2, arrow) and Pax7 (Fig. 4D1, D2, arrow) expression reduced. Electroporation of control pCAGGS-eGFP into neural tube of either talpid3 or wild-type embryos produced no gene expression changes (data not shown). These results demonstrate directly that supplying talpid3 cells with functional activated Gli rescues the defect.
Figure 4. Rescue of talpid3 neural tube with activated Gli constructs. (A–H) Stage 20 HH wild-type and talpid3 embryos electroporated with Gli3AHIGH (activated Gli construct). Green indicates expression of GFP in transfected cells; red indicates protein expression detected by immunohistochemistry. A1–H1 are overlays of GFP and immunostaining; A2–H2 are immunostaining alone. (A–D) Wild-type neural tube. Induction of Nkx2.2-positive cells (arrows, A1,A2); Islet1-positive cells (arrows, B1,B2). Down-regulation of Pax6 (arrow, C1,C2); Pax7 (D1,D2). (E–H) talpid3 neural tube. Induction of Nkx2.2-positive cells (E1,E2) and Islet1-positive cells (upper arrow, F1,F2). Down regulation of Pax6 (arrow, G1,G2); Pax7 (arrow, H1,H2).
Genetic mapping and identification of talpid3
Linkage analysis of 110 individual carriers mapped the TA3 locus close to markers ADL0298 and ADL0166 on chromosome 5. Mapping further markers refined the location to an interval containing COM0184-(5.1)-ADL0166-(7.3)-SIX4-(3.6)-DAAM1-(0.9)-TA3-(4.6)-OTX2(8.3)-ADL0298-(0.0)-ROS0330-(2.8)-BMP4 (genetic distances between markers in centimorgans). We constructed a detailed physical map of this region to define gene content and develop more markers (Fig. 5A). Using these markers, the location of the talpid3 mutation was reduced to an interval encoding five genes: KCNK16-DACT1-ENSGALG00000012025-TIMM9-ARID4A (Fig. 5B).
To look for sequence changes that might identify the talpid3 mutation we sequenced cDNA clones derived from wild-type and talpid3 embryos (Supplementary Table S1). Sequence analysis of talpid3 KIAA0586 cDNA revealed an insertion mutation of a single thymine residue (Fig. 5C), which was confirmed by sequencing genomic products from carrier and noncarriers (Supplementary Fig. S1). Translation from KIAA0586 in talpid3 is terminated nine residues after the thymine insertion at a stop codon (Fig. 5C); this frameshift mutation is predicted to produce a truncated protein of 366 amino acids compared with 1524 in wild type (Fig. 6A). A multispecies comparison of KIAA0586 orthologs (Supplementary Fig. S2) shows extensive homology across all 30 exons, and a number of highly conserved noncoding regions. The complete sequence of the chicken KIAA0586 mRNA is shown in Supplementary Figure S3. The translation initiation codon at position 137 was selected as the most likely based on the longest open reading frame and homology with other vertebrate orthologs (Supplementary Fig. S4). To detect KIAA0586 orthologs in other species, we combined information from conservation of sequence and gene order (Supplementary Figs. S4, S5). Complete conservation of gene order was found for DAAM1-DACT1-KIAA0586-TIMM9-ARID4A in all vertebrates examined. No homologs were detected in Ciona intestinalis, Drosophila melanogaster, or Caenorhabditis elegans. Using Codonml (PAML 3.13) we were able to estimate the KA/KS ratio for KIAA0586 (0.28), a measure of selection constraint (Yang and Nielsen 2000). For comparison, the KA/KS ratio for mammalian orthologous gene pairs (Rat Genome Sequencing Project Consortium 2004) was 0.09–0.11, most genes being under strong to moderate purifying selection. The KA/KS ratio for KIAA0596 indicates that this protein has been under reduced purifying selection and/or increased positive selection. No specific functional domains were predicted in the peptide sequence. Most highly conserved are three globular domains, coiled-coil regions, and a central region, with as yet, no known structural features, and least conserved is the C-terminal region (Fig. 6A).
Expression of talpid3 gene and protein and analysisof Gli3 translocation in mutant cells
KIAA0586 is ubiquitously expressed in wild-type embryos (stages 12–35 HH), consistent with the talpid3 mutation affecting multiple tissues (Fig. 6B). Using the program ProtComp, the KIAA0586 protein was predicted to be cytoplasmic. In order to determine cellular localization experimentally, we expressed either C-terminal-tagged KIAA0586-myc or N-terminal-tagged KIAA0586-HA in limb bud cells from wild-type chick embryos or in the chicken DFI cell line and visualized cellular localization by immunofluorescence. In all cases, labeling was confined to the cytoplasm (Fig. 6D,E), suggesting the talpid3 protein is a cytoplasmic protein.
To further investigate the role of KIAA0586 in Hh signaling, we compared intracellular localization of Gli3 in wild-type and talpid3 mutant chick limb bud cells. In one set of experiments, we isolated cytoplasmic and nuclear fractions from wild-type and mutant cells and then carried out Western blot analysis as before (Fig. 6F). We assessed the effectiveness of the fractionation by detecting tubulin and found this to be confined to the cytoplasmic fraction (Fig. 6F). Both full-length (190 kDa) and processed (83 kDa) Gli3 proteins were readily detected in nuclei isolated from wild-type chick limb bud cells but not in the cytoplasmic fraction. Both Gli3 forms were also detected in the nuclei of talpid3 cells, although full-length Gli3 could also be detected in the cytoplasm of mutant cells (Fig. 6F). Furthermore when the ratio of Gli3A/Gli3R was compared, it is evident that talpid3 cells from limb and head have a much higher ratio of Gli3A/Gli3R in the nucleus (Fig. 6G) than wild-type cells. We also expressed pGli3-myc in wild-type and talpid3 chick limb bud cells and visualized localization using immunofluorescence (Fig. 6H). Localization of Gli3-myc was rather variable but was seen in both nucleus and cytoplasm in both wild-type and talpid3 cells. These results reinforce the conclusion that the talpid3 protein is required for efficient processing of functional GliA and GliR but suggest that it is not required for nuclear translocation of these proteins.
Rescue of talpid3 neural tube by KIAA0586
To confirm that KIAA0586 is the talpid3 gene, we carried out complementation tests. In the first set of experiments, we tested whether KIAA0586 could rescue ectopic induction of dorsoventral markers by Shh in neural tube of stage 17–18 HH talpid3 embryos. Shh beads were inserted into the neural tube and a pCAGGS-KIAA0586 construct was coelectroporated with a pCAGGS-eGFP control plasmid and dorsoventral patterning examined after 24 h. As noted above, application of a Shh bead to a wild-type neural tube induces ectopic Nkx2.2 (Fig. 7A2, arrows), Islet1 (Fig. 7B2, arrows), shifts strong Pax6 expression dorsally (Fig. 7C2, arrow), induces Ptc1 throughout the neural tube, surrounding somites, and ectoderm (Fig. 7I) and reduces Pax7 expression ventrally (Fig. 7D2, arrow) while application of Shh to a talpid3 neural tube has no effect on Nkx2.2, Islet1, Pax6, or Pax7 and only weakly induces Ptc1 (Supplementary Fig. S6I2,J2, K2,L2,M). When wild-type tissue was electroporated with pCAGGS-KIAA0586 and treated with Shh, a change in gene expression of Nkx2.2, Islet1, Pax6, Pax7, or Ptc1 was observed on both sides of the neural tube, whether electroporated or not, and therefore attributable only to the effect of the Shh bead (Fig. 7A–D). However, when talpid3 embryos were provided with both Shh and KIAA0586, ectopic expression of Nkx2.2 was induced only on the electroporated side of the neural tube (2/2) (Fig. 7E1–E3, arrows) and in some cells, expression of Pax6 (2/2) (Fig. 7G1, G2, asterisk) and Pax7 was also reduced on the electroporated side (2/2) (Fig. 7H1, H2, asterisk) in a way similar to that seen with Gli3HIGH. Islet1 expression remained unchanged (2/2) (Fig. 7F1, F2). Application of Shh with KIAA0586 also induced high-level Ptc1 expression in talpid3 dorsal neural tube on the electroporated side (Fig. 7J; asterisk denotes electroporated side). In preliminary experiments, one talpid3 embryo had a reduction in Pax7 when electroporated with pCAGGS-KIAA0586 (1/1) (Supplementary Fig. S6H1,H2) but Nkx2.2, Islet1, and Pax6 remained unchanged (Supplementary Fig. S6E–G), and there was no effect on expression of Nkx2.2, Islet1, Pax6, or Pax7 in a wild-type embryo electroporated with pCAGGS-KIAA0586 at stage 17 HH (1/1) (Supplementary Fig. S6A–D).
In a second set of experiments, we tested whether whether KIAA0586 alone could rescue endogenous dorsoventral patterning of the neural tube in very early talpid3 embryos. In order to carry out these experiments, we had to electroporate batches of embryos from the talpid3 flock at stages 12–14, well before the mutant phenotype is observable. We electroporated the pCAGGS-KIAA0586 construct, together with pCAGGS-RFP and fixed embryos at stage 20HH in order to examine gene expression. At this stage, we could usually identify mutant embryos by their phenotype, and we confirmed their identity by genotyping. In talpid3 embryos, we found very striking induction of Islet1 expression (3/3) (Fig. 7K1/2, green cells) coupled with repression of Pax6 (2/2) (Fig. 7L1/2, arrow). We also have preliminary evidence that Pax7 expression was reduced in talpid3 (1/1) (data not shown). This rescue of endogenous target gene expression provides good evidence that KIAA0568 is the talpid3 gene and has an essential role in Hh signal transduction.
Discussion
Hh signaling plays crucial roles in development and disease, and therefore it is important to understand the signal transduction pathway that mediates Hh signaling. Here we identify a novel component of the signaling pathway in vertebrates and provide evidence that it is necessary for the function of both Gli repressor and Gli activator forms. By a positional cloning strategy, exploiting the recently assembled chick genome, we provide evidence that KIAA0586 is the gene responsible for the talpid3 phenotype. Analysis of KIAA0586 indicates that it contains coiled-coil domains and extensive regions of intrinsic protein disorder (Fig. 6A), characteristic of proteins that play a role in cell signaling and gene regulation (Iakoucheva et al. 2002). Moreover, a frameshift mutation in KIAA0586 is present in talpid3 mutants, and the expression of a wild-type KIAA0586 cDNA in the neural tube of talpid3 embryos rescued the Shh responsiveness of neural cells and endogenous dorsoventral patterning. We therefore propose that KIAA00586 be renamed Talpid3.
Our analysis of Hh target gene expression in limb, face, neural tube, and somite of talpid3 embryos, together with our findings that expression of target genes is ligand independent—insensitive to either presence or absence of Shh—and can be rescued by expressing activated Gli are consistent with the idea that both GliA and GliR functions are compromised in talpid3. Thus the Talpid3 gene product plays an essential role in both GliA and GliR function in the Hh pathway. The paradoxical talpid3 phenotype, an apparent mixture of gain and loss of Hh function, is explained by the fact that GliR plays a predominant role in limb and somite, whereas GliA function is more important in face and neural tube patterning.
Our data show that Gli3 processing is impaired in talpid3 mutant cells, leading to the presence of both full-length Gli3A and Gli3R. Levels of Gli3A, in particular, are markedly increased. Although changes in processing are likely to be the main cause, changes in protein longevity and transcriptional levels could also contribute. The net result is that the ratio of Gli3A/Gli3R is much greater in talpid3 mutant cells. Abnormal processing is found not only in cells from the limb but also in cells from the trunk, which includes neural tube and somite. We have also shown that, as predicted by its structure, Talpid3 protein is localized in the cytoplasm. Thus, our results are consistent with Talpid3 interacting with other members of the Hh signaling complex to regulate processing of Gli proteins in the cytoplasm. Given the high levels of full-length Gli3, it is not clear why GliA function, in addition to GliR function, is compromised in talpid3. This is even more puzzling since we have also shown that the absence of Talpid3 protein does not prevent the ability of Gli3 proteins including Gli3 activator to enter the nucleus. The most likely possibility is that the pathway is very sensitive to the precise levels of full-length and short forms of Gli proteins such that, in the talpid3 mutant, the levels of GliA and GliR counterbalance their effects, resulting in neither activation nor repression. Alternatively, the Gli3 proteins produced in the absence of Talpid3 may be unable to regulate transcription due to, as yet, unknown post-translational modifications.
Recently two other genes involved in Hh signaling, Dzip1 (Sekimizu et al. 2004; Wolff et al. 2004) (chicken ortholog ENSGALG00000016895) and MTSS1 (Callahan et al. 2004) (also called MIM/BEG4, chicken ortholog ENSGALG00000016333), have been described. Dzip1 was identified as the Zebrafish mutant gene in the iguana mutant, which shows many similarities to the talpid3 mutant, including reduction of expression of Shh-dependent genes in neural tube coupled with a gain of Hh signaling in the somite. Furthermore, as in talpid3, iguana cannot be rescued through manipulation of Shh ligand, leading to the suggestion that either both GliA and GliR functions are reduced (Sekimizu et al. 2004), similar to our proposal for talpid3, or that GliR fails because of constitutive low-level Gli1 activation (Wolff et al. 2004). There is evidence that Dzip1 can shuttle between cytoplasm and nucleus, and it has been suggested that Dzip1 may affect nuclear import of Hh pathway proteins. This contrasts with Talpid3, which appears to be confined to the cytoplasm and is not required for nuclear import of Gli3. Interestingly, constitutive activation of Shh signaling has been suggested to occur in the talpid2 chicken mutant (Chuang and McMahon 1999), in which high levels of Gli3A are also found (Wang et al. 2000). Talpid2, like talpid3, is polydactylous, and the many digits are unpatterned, but, unlike talpid3, there is high-level Ptc expression throughout the limb (Carrucio et al. 1999), suggesting that Gli activator is functional in this mutant. Thus similarities in phenotype could arise by quite different mechanisms.
A number of intraflagellar transport proteins have been shown to play an essential role in vertebrate Hh signal transduction (Huangfu et al. 2003) in mouse, and a recent report suggests a functional link between Hh signaling and cilia at the level of the transmembrane protein smoothened which is required for GliA function (Corbit et al. 2005). Analysis of hypomorphic mouse mutants of two intraflagellar transport proteins reveals striking parallels between these mutants and talpid3 (Liu et al. 2005). Not only are the alterations in gene expression patterns in limb and neural tube similar to those seen in talpid3 and insensitive to Hh ligand, but also high levels of full-length Gli3 were detected in mutant cells, giving rise to a marked increase in the Gli3A/Gli3R ratio. Like Talpid3, intraflagellar transport proteins appear to be required for both GliA and GliR functions. It will be important to determine why mutations in intraflagellar transport proteins compromise GliR function in addition to GliA function and to explore the relationship between Talpid3 and cilia. Talpid3 joins a growing list of proteins, including these intraflagellar transport proteins, that regulate Hh transduction specifically in vertebrates (Chuang and McMahon 1999; Eggenschwiler et al. 2001) and is the first component in this pathway to be discovered in chickens.
Materials and methods
Embryo manipulations
Eggs were incubated for 2.5 d at 38°C, then windowed to assess development and classified as either mutant or wild type and reincubated until the desired stage. All limb manipulations were carried out on the right wing, leaving the left wing as a control.
Cyclopamine treatment: One-hundred microliters of 5 mM cyclopamine in 95% EtOH diluted with 200 μL DMEM was sonicated for 10 min; for the control solution, 100 μL of 95% EtOH was diluted in 200 μL DMEM. Solutions were injected with a fine glass micropipette into the amniotic sac over the right wing of stage 20 HH embryos; eggs were reincubated for 24 h.
Removal of Shh-expressing cells: At stage 16 HH, tissue was removed from flank, immediately lateral to somites, from axial level somites 22 and 18/17 using a sharpened tungsten needle; eggs were then reincubated for 18 h.
RCAS virus production, microinjection, and somite grafts
RCAS virus production and microinjection were done as per Morgan and Fekete (Morgan and Fekete 1996). In ovo surgical experiments were performed as per Schmidt et al. (Schmidt et al. 2000), and eggs were reincubated for 40 h.
Tissue graft implants—neural tube
QT-6 Shh expressing cells were grown until confluent. Cell sheets were scraped from the dish surface. A small piece of sheet was folded and tucked into a slit made in the neural tube at the level of the hind limbs using a tungsten needle; eggs were reincubated for 24 h.
Bead implants—neural tube
CM Affi-Gel Blue Beads (Bio-Rad) soaked in Shh protein as per Drossopoulou et al. (Drossopoulou et al. 2000) were picked up on the point of a tungsten needle and placed into the neural tube via a dorsal slit.
Electroporation
DNA was prepared using a Qiagen Endotoxin-Free Maxiprep Kit, resuspended in endotoxin-free water, and diluted with 0.25% Fast Green. Plasmid pCAGGS-Gli3AHIGH was used at a concentration of 7 μg/μL, pCAGGS-GFP at 7.5 μg/μL, pCAGGS-RFP at 0.1 μg/μL, and pCAGGS-KIAA0586 at 1 μg/μL. DNA was injected into the neural tube and embryos were immediately electroporated at 30 V for 5 x 50 msec with 5 x 50–100-msec intervals or at 20 V for 5 x 15 msec (stages 12–14). Embryos were reincubated for 20–48 h. Successful electroporation was assessed by expression of GFP, eGFP, or RFP using a Leica Fluorescence Dissection Microscope. The GFP section of neural tube was then dissected out in cold PBS (pH 7.4) and fixed for sectioning for immunohistochemistry or whole-mount in situ hybridization.
Whole-mount RNA in situ hybridization
For details, see the Supplemental Material.
Whole-mount immunohistochemistry
For details, see the Supplemental Material.
Section immunohistochemistry
For details, see the Supplemental Material.
Western blotting
For determining levels of Shh protein, stage 20 HH embryo limb buds were dissected into three equal parts along the anterior–posterior axis in cold PBS, lysed in RIPA buffer/1:25 Proteases Inhibitor (Roche), samples ran on 12% SDS-PAGE, and protein detected with Shh antibody. Lysates were spun at 13,000 rpm for 5 min at room temperature, supernatant was removed and methanol precipitated and then dried using the Eppendorf Concentrator. The pellet was resuspended in 30 μL of Sample Buffer 1/10 -mercaptoethanol and then denatured. Gels were subsequently wet or dry blotted in 1x Tris-Glycine/20%Methanol buffer for 1–2 h at 172 mA onto nitro-cellulose, which was blocked in 5% Marvel/PBS 1:1000 Tween-20; incubated with 1:2 antibody and Marvel/PBT overnight at 4°C; washed 3 h in PBT; incubated in anti-mouse HRP-conjugated (Sigma) in Marvel/PBT 4 h; and washed 1 h in PBT, then 30 min in PBS. Detection was with Supersignal West Pico Chemiluminescent Substrate (Pierce) for 5 min and was detected with X-ray film exposed overnight.
Levels of full-length and short forms of Gli3 protein were determined in protein extracts from limb buds and trunks dissected from stage HH24 embryos. Some limb buds were dissected into three equal parts as above. All tissues were then lysed in buffer containing 1% NP-40, 150 mM NaCl, and 20 nM Tris (pH 7.4) in the presence of Complete protease inhibitor cocktail (Roche). Protein concentrations were measured using Bradford (Pierce), and equal amounts of protein were separated by SDS-PAGE using 4%–12% precast gels from Invitrogen. Immunoblot was performed using a polyclonal antibody against Gli3 (Santa Cruz Biotechnology, Inc).
Nuclear and cytosol fractionation
Limbs and heads from stage 24HH embryos were dissected out in cold PBS and cut into small pieces with scissors. Homogenization of tissue was carried out by a pestle on ice. Extraction of nuclear and cytoplasmic fractions was performed using a Compartmental Protein extraction Kit (Chemicon) as per the manufacturer’s instruction. Protein concentration was determined by the Bradford method. Nuclear and cytoplasmic fractions were stored at –80°C for Western blot analysis.
Cell cultures and immunohistochemistry
DFI chicken cells were grown in DMEM + 10% fetal calf serum; constructs containing either KIAA0586-myc or KIAA0586-HA were introduced using FuGene (Roche).
Primary cell culture and immunofluorescence staining
Individual chick limbs from HH20–24 embryos were dissected in sterile cold PBS and incubated with 1x Trypsin and EDTA for <5 min at 37°C. A 20-μL cell suspension was plated onto cover glasses coated by fibronectin. The cover glasses were set in a Petri dish and incubated at 37°C for an hour followed by flooding cells with 1:1, DMEM-HAMS F12/10% FCS, 1% PSF, 1% L-glutamine. The next day, cells were transfected with myc-hGli3 or KIAA0586-myc or KIAA0586-HA using FuGene 6 (Roche).
After 24 h transfection, cells were washed in PBS, fixed in ice-cold methanol for 10 min, then blocked in 10% normal serum in PBS for 30 min. Cells were incubated in monoclonal anti-c-myc (1:1000, Sigma) 1 h at room temperature, washed 3 x 5 min in 0.2% Tween-20 in PBS, and incubated in anti-mouse IgG conjugated with Alexa Fluor 610 (1:1000, Molecular Probes) for 1 h at room temperature. After washes (3 x 5 min in 0.2% Tween-20 in PBS), cells were mounted in VectaShield Hard*set mount medium with DAPI (Vector Laboratories, H-1500). Fluorescence was observed using a fluorescence microscope or a Leica SP2 confocal microscope.
Animals and genetic linkage analysis
Talpid3 carriers were maintained by outcrossing with ISA Brown chickens. Line 6, an inbred White Leghorn line, was kindly provided by the Institute of Animal Health (Compton). A male carrier (TA3/N) was crossed to eight Line 6 females (N/N) and offspring of this backcross were test-crossed with known carriers. A carrier was confirmed if two or more talpid3 embryos were found from a test cross. At least 20 wild-type embryos were collected before any noncarriers were confirmed. Four DNA pools were created from backcross offspring: Carrier male (n = 13), carrier female (n = 32), noncarrier male (n = 7), and noncarrier female (n = 16) were genotyped using 129 autosomal genetic markers, covering 70% of the genome. Markers ADL0298 and ADL0166 on chromosome 5 showed a significant difference (p < 0.00001) between carrier and noncarrier pools. Further markers near these loci were used to genotype 110 individuals and refine the location of the TA3 locus. Two-point and multipoint linkage analysis was performed using CRIMAP (Green et al. 1990). From this analysis, 30 recombinant offspring were identified near TA3 between markers ADL0166 and ADL0298 and typed using markers developed from EST and genomic sequences. Using known chicken genes and reference to the Ensembl (http://www.ensembl.org) and University of California at Santa Cruz (http://www.genome.ucsc.edu) genome browsers, we identified the orthologous region on human chromosome 14. Using human gene sequences and Blast we searched databases (http://www.ncbi.org) for homologous chicken EST and genomic sequences. We developed PCR-based markers from these showing size and SSCP or SNP variation. These markers were used to isolate and order individual BAC clones (Burt et al. 2003) to create a detailed physical map. BAC-end and full-length cDNA sequences were also used to create more genetic markers. Using such a high density of markers we were able to identify specific haplotypes from recombinants in this region and map the TA3 locus to KCNK16-DACT1-KIAA0586-TIMM9-ARID4A. Genotyping was performed as described by Burt et al. (2003) and primers used for genotyping are listed in Supplementary Table S2.
Sequence analysis and characterization of talpid3 mutation
For details, see the Supplemental Material.
Genotyping
Genomic DNA was prepared from extra-embryonic membranes using Quigen DNAeasy Kit. A 957-base-pair (bp) region of the talpid3 gene was then amplified via PCR (forward: CTTCTCT TGCCCTGCCTTAC; reverse: TTGAAGCTGCATTTCCA CAG), annealed at 60°C (Thermo electron) using Fast Start Taq Polymerase and GC-rich solution (Roche). The PCR product was then digested using Pst1 (Roche). A single 957-bp band indicates a wild-type embryo, two bands (721 bp and 236 bp) indicate a talpid3 mutant, and three bands indicate a heterozygous embryo.
Cloning of chicken cDNA clones and gene constructs
RNA was extracted using TRIzol reagent (Invitrogen), and a DACT1 cDNA was cloned via RT–PCR (P1: GGGATGCTG GCCTATTAAATC in exon 2; P2: GCTTCCCAAAG CAGAAACAG, in 3' UTR) from embryo-derived mRNA (HH stage 21) using the TOPO TA cloning kit (Invitrogen), then subcloned into pGEM (Invitrogen) for preparation of antisense RNA probes. Rapid amplification of 5' cDNA ends of KIAA0586 and DAAM1 was carried out using the First Choice RLM-RACE Kit (Ambion) and cloned using the TOPO TA cloning kit. Gene-specific primers are listed in Supplementary Table S2.
Bioinformatics analyses
For details, see the Supplemental Material.
Accession numbers
All sequences have been deposited in GenBank under accession numbers DQ066927 [GenBank] –DQ066935 [GenBank] and CZ550237 [GenBank] –CZ550250 [GenBank] .
Acknowledgments
We are grateful to Dr. Jun-ichi Miyazaki and Dr. Toshihiko Agura for providing pCAGGS-eGFP, Dr. Ruiz i Altaba for plasmid myc-hGli-3, Dr. R. Johnson and Dr. C. Tabin for pRCAS-Shh, Professor Groenen for chicken BAC clones, and Dr. Placzek for Shh antibody. EST clones from UK Chick EST Project were provided by ARK-Genomics (http://www.ark-genomics.org). This research was supported by the BBSRC (to M.D., I.R.P., Y.Y., D.R.M., F.K.B., D.R.M., D.B., and C.T.), British Heart Foundation (to M.D. and C.T.), MRC (UK; to T.G.S., M.T. and C.T.), and The Royal Society (to T.G.S. and C.T.). D.S. and J.B. were supported by the MRC (UK) and an EC network grant Stembridge (QLG3-CT-2002-011441); A.E.M. and M.S. were supported by The Wellcome Trust and MYORES (EU FP6 NoE).
References
Bai C.B., Auerbach W., Lee J.S., Stephen D., Joyner A.L. 2002. Gli2, but not Gli1, is required for initial Shh signaling and ectopic activation of the Shh pathway. Development 129: 4753–4761.
Borycki A.G., Brunk B., Tajbakhsh S., Buckingham M., Chiang C., Emerson C.P.J. 1999. Sonic hedgehog controls epaxial muscle determination through Myf5 activation. Development 126: 4053–4063.
Burt D.W., Morrice D.R., Lester D.H., Robertson G.W., Mohammed M.D., Simmons J., Downey L.M., Thuang C., Bridges L.R., Paton I.R. et al. 2003. Analysis of rdd locus in chicken: A model for human retinitis pigmentosa. Mol. Vis. 9: 164–170.
Buxton P., Davey M.G., Paton I.R., Morrice D.R., Francis-West P.H., Burt D.W., Tickle C. 2004. Craniofacial development in the talpid3 chicken mutant. Differentiation 72: 348–362.
Callahan C.A., Ofstad T., Horng L., Wang J.K., Zhen H.H., Coulombe P.A., Oro A.E. 2004. MIM/BEG4, a Sonic hedgehog-responsive gene that potentiates Gli-dependent transcription. Genes & Dev. 18: 2724–2729.
Carrucio N.C., Martinez-Lopez A., Harris M., Dvorak L., Bitgood J., Simandl B.K., Fallon J.F. 1999. Constitutive activation of sonic hedgehog signaling in the chicken mutant talpid2, Shh independent outgrowth and polarizing activity. Dev. Biol. 212: 137–149.
Chuang P.T. and McMahon A.P. 1999. Vertebrate Hedgehog signaling modulated by induction of a Hedgehog-binding protein. Nature 397: 617–621.
Corbit K.C., Aanstad P., Singla V., Norman A.R., Stainier D.Y., Reiter J.F. 2005. Vertebrate Smoothened functions at the primary cilium. Nature 437: 1018–1021.
Drossopoulou G., Lewis K.E., Sanz-Ezquerro J.J., Nikbakht N., McMahon A.P., Hofmann C., Tickle C. 2000. A model for anteroposterior patterning of the vertebrate limb based on sequential long- and short-range Shh signaling and Bmp signaling. Development 127: 1337–1348.
Ede D.A. and Kelly W.A. 1964a. Developmental abnormalities in the head region of the talpid mutant of the fowl. J. Embryol. Exp. Morphol. 12: 161–182.
Ede D.A. and Kelly W.A. 1964b. Developmental abnormalities in the trunk and limbs of the fowl. J. Embryol. Exp. Morphol. 12: 339–356.
Eggenschwiler J.T., Espinoza E., Anderson K.V. 2001. Rab23 is an essential negative regulator of the mouse Sonic hedgehog pathway. Nature 412: 194–198.
Francis-West P.H., Robertson K.E., Ede D.A., Rodriguez C., Izpisua-Belmonte J.C., Houston B., Burt D.W., Gribbin C., Brickell P.M., Tickle C. 1995. Expression of genes encoding bone morphogenetic proteins and sonic hedgehog in talpid (ta3) limb buds: Their relationships in the signaling cascade involved in limb patterning. Dev. Dyn. 203: 187–197.
Frank-Kamenetsky M., Zhang X.M., Bottega S., Guicherit O., Wichterle H., Dudek H., Bumcrot D., Wang F.Y., Jones S., Shulok J. et al. 2002. Small-molecular modulators of Hedgehog signaling: Identification and characterization of Smoothened agonists and antagonists. J. Biol. 1: 10.
Green P., Falls K., Crooks S. In Documentation for CRIMAP, version 2.4. . 1990. Washington School of Medicine, St. Louis, MO.
Hamburger V. and Hamilton H.L. 1951. A series of normal stages in the development of the chick embryo. J. Morphol. 88: 49–92.
Hinchliffe J.R. and Ede D.A. 1968. Abnormalities in bone and cartilage development in the talpid mutant of the fowl. J. Embryol. Exp. Morphol. 19: 327–339.
Huangfu D., Liu A., Rakeman A.S., Murcia N.S., Niswander L., Anderson K.V. 2003. Hedgehog signaling in the mouse requires intraflagellar-transport proteins. Nature 426: 83–87.
Hui C.C. and Joyner A.L. 1993. A mouse model of greig cephalopolysyndactyly syndrome: The extra-toesJ mutation contains an intragenic deletion of the Gli3 gene. Nat. Genet. 3: 241–246.
Hunton P. 1960. "A study of some factors affecting the hatchibility of chicken eggs, with special reference to genetic control." MSc. thesis. Wye College, University of London, UK.
Iakoucheva L.M., Brown C.J., Lawson J.D., Obradovic Z., Dunker A.K. 2002. Intrinsic disorder in cell-signaling and cancer-assoicated proteins. J. Mol. Biol. 323: 573–584.
Ingham P.W. and McMahon A.P. 2001. Hedgehog signaling in animal development, paradigms and principles. Genes & Dev. 15: 3059–3087.
International Chicken Genome Sequencing Consortium . 2004. Sequence and comparative analysis of the chicken genome provide unique perspectives on vertebrate evolution. Nature 432: 695–716.
Izpisua-Belmonte J.C., Ede D.A., Tickle C., Duboule D. 1992. The mis-expression of posterior Hox-4 genes in talpid (ta3) mutant wings correlates with the absence of anteroposterior polarity. Development 114: 959–963.
Lewis K.E., Drossopoulou G., Paton I.R., Morrice D.R., Robertson K.E., Burt D.W., Ingham P.W., Tickle C. 1999a. Expression of ptc and gli genes in talpid3 suggests bifurcation in Shh pathway. Development 126: 2397–2407.
Lewis K.E., Currie P.D., Roy S., Schauerte H., Haffter P., Ingham P.W. 1999b. Control of muscle cell-type specification in the zebrafish embryo by Hedgehog signaling. Dev. Biol. 216: 469–480.
Litingtung Y., Dahn R.D., Li Y., Fallon J.F., Chiang C. 2002. Shh and Gli3 are dispensable for limb skeleton formation but regulate digit number and identity. Nature 418: 979–983.
Liu A., Wang B., Niswander L.A. 2005. Mouse intraflagellar transport proteins regulate both the activator and repressor functions of Gli transcription factors. Development 132: 3103–3111.
Morgan B.A. and Fekete D.M. 1996. Manipulating gene expression with replication-competent retroviruses. Methods Cell Biol. 51: 185–218.
Munsterberg A.E., Kitajewski J., Bumncrot D.A., McMahon A.P., Lassar A.B. 1995. Combinatorial signaling by Sonic hedgehog and Wnt family members induces myogenic bHLH gene expression in the somite. Genes & Dev. 9: 2911–2922.
Pearse R.V., Vogan K.J., Tabin C.J. 2001. Ptc1 and Ptc2 transcripts provide distinct readouts of Hedgehog signaling activity during chick embryogenesis. Dev. Biol. 239: 15–29.
Persson M., Stamataki D., te Welscher P., Andersson E., Bose J., Ruther U., Ericson J., Briscoe J. 2002. Dorso-ventral patterning of the spinal cord requires Gli3 transcriptional repressor activity. Genes & Dev. 16: 2865–2878.
. Rat Genome Sequencing Project Consortium. 2004. Genome sequence of the Brown Norway rat yields insights into mammalian evolution. Nature 428: 493–521.
Ruiz i Altaba A., Nguyen V., Palma V. 2003. The emergent design of the neural tube, pre-pattern, SHH morphogen and GLI code. Curr. Opin. Genet. Dev. 13: 513–521.
Schmidt M., Tanaka M., Munsterberg A. 2000. Expression of -catenin in the developing chick myotome is regulated by myogenic signals. Development 127: 4105–4113.
Sekimizu K., Nishioka N., Sasaki H., Takeda H., Karlstrom R.O., Kawakami A. 2004. The Zebrafish iguana locus encodes Dzip1, a novel zinc-finger protein required for proper regulation of Hedgehog Signaling. Development 131: 2521–2532.
Stamataki D., Ulloa F., Tsoni S.V., Mynett A., Briscoe J. 2005. A gradient of gli activity mediates graded Sonic Hedgehog signaling in the neural tube. Genes & Dev. 19: 626–641.
te Welscher P.T., Zuniga A., Kuijper S., Drenth T., Goedemans H.J., Meijlink F., Zeller R. 2002. Progression of vertebrate limb development through SHH-mediated counteraction of GLI3. Science 298: 827–830.
Wang B., Fallon J.F., Beachy P.A. 2000. Hedgehog-regulated processing of Gli3 produces an anterior/posterior repressor gradient in the developing vertebrate limb. Cell 100: 423–434.
Wolff C., Roy S., Lewis K.E., Schauerte H., Joerg-Rauch G., Kim A., Weiler C., Geisler R., Haffter P., Ingham P. 2004. Iguana encodes a novel zinc-finger protein required for proper regulation of Hedgehog Signaling. Development 131: 2521–2532.
Yang Z. and Nielsen R. 2000. Estimating synonymous and nonsynonymous substitution rates under realistic evolutionary models. Mol. Biol. Evol. 17: 32–43.(Megan G. Davey1,2, I. Rob)