A novel gene, GliH1, with homology to the Gli zinc finger domain not required for mouse development

Mechanisms of Development 119 (2002) 21–34 www.elsevier.com/locate/modo

A novel gene, GliH1, with homology to the Gli zinc finger domain not required for mouse development M. Nakashima a,b,*, N. Tanese c, M. Ito d, W. Auerbach b,1, C. Bai b,e, T. Furukawa c, T. Toyono a,2, A. Akamine a, A.L. Joyner b,e,f a

Department of Clinical Oral Molecular Biology, Division of Oral Rehabilitation, Faculty of Dental Science, Kyushu University, Fukuoka 812-8582, Japan b Howard Hughes Medical Institute and Developmental Genetics Program, Skirball Institute of Biomolecular Medicine, New York University School of Medicine, 540 First Avenue, New York, NY 10016, USA c Department of Microbiology, New York University School of Medicine, 540 First Avenue, New York, NY 10016, USA d Department of First Anatomy, National Defense Medical College, Tokorozawa, Japan e Department of Cell Biology, New York University School of Medicine, 540 First Avenue, New York, NY 10016, USA f Department of Physiology and Neuroscience, New York University School of Medicine, 540 First Avenue, New York, NY 10016, USA Received 15 June 2002; received in revised form 6 August 2002; accepted 6 August 2002

Abstract The Sonic hedgehog (Shh)–Gli signaling pathway regulates development of many organs, including teeth. We cloned a novel gene encoding a transcription factor that contains a zinc finger domain with highest homology to the Gli family of proteins (61–64% amino acid sequence identity) from incisor pulp. Consistent with this sequence conservation, gel mobility shift assays demonstrated that this new Gli homologous protein, GliH1, could bind previously characterized Gli DNA binding sites. Furthermore, transfection assays in dental pulp cells showed that whereas Gli1 induces a nearly 50-fold increase in activity of a luciferase reporter containing Gli DNA binding sites, coexpression of Gli1 with Gli3 and/or GliH1 results in inhibition of the Gli1-stimulated luciferase activity. In situ hybridization analysis of mouse embryos demonstrated that GliH1 expression is initiated later than the three Gli genes and has a more restricted expression pattern. GliH1 is first detected diffusely in the limb buds at 10.0 days post coitus and later is expressed in the branchial arches, craniofacial interface, ventral part of the tail, whisker follicles and hair, intervertebral discs, teeth, eyes and kidney. LacZ was inserted into the GliH1 allele in embryonic stem cells to produce mice lacking GliH1 function. While this produced indicator mice for GliH1-expression, analysis of mutant mice revealed no discernible phenotype or required function for GliH1. A search of the Celera Genomics and associated databases identified possible gene sequences encoding a zinc finger domain with ,90% homology to that of GliH1, indicating there is a family of GliH genes and raising the possibility of overlapping functions during development. q 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Zinc finger transcription factor; Sonic Hedgehog; Gli; GliH1; Mouse development; Dental pulp

1. Introduction The Gli genes (Gli1, Gli2 and Gli3) (Kinzler et al., 1987; Ruppert et al., 1988, 1990; Walterhouse et al., 1993; Hui et al., 1994) encode proteins containing five tandem repeats of a C2H2 type zinc finger DNA binding motif (Ruppert et al., 1988; Pavletich and Pabo, 1993) and function downstream of the Hedgehog signaling molecule. Comparative expression studies have shown that the three Gli genes are expressed in a number of different ectoderm- and meso* Corresponding author. Tel.: 181-92-642-6432; fax: 181-92-642-6366. E-mail address: [email protected] (M. Nakashima). 1 Present address: Regeneron Pharmaceuticals, Inc., 777 Old Saw Mill River Road, Tarrytown, NY 10591-6707, USA. 2 Present address: Department of Oral Anatomy and Neurobiology, Kyushu Dental College, Fukuoka 803-8580, Japan.

derm-derived tissues including the skeleton and CNS, complementary to the signaling protein Sonic hedgehog (Shh) (Hui et al., 1994; Platt et al., 1997). The function of the Shh pathway is perhaps best understood in dorsal/ventral patterning of the neural tube and anterior/posterior patterning in the limb. Shh is expressed in the notochord underlying the neural plate and required to induce floorplate cells in the ventral midline and other ventral cell types (Chiang et al., 1996; Pierani et al., 1999). In the limbs, Shh expression in the posterior mesenchymal region called the zone of polarizing activity is required for anterior/posterior patterning of the limbs (Chiang et al., 1996, 2001; Kraus et al., 2001). Ectopic expression of Shh in the dorsal neural tube or anterior limb induces ventral neural cell types or ectopic limb structures, respectively, showing that Shh is both necessary

0925-4773/02/$ - see front matter q 2002 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0925-477 3(02)00291-5

22

M. Nakashima et al. / Mechanisms of Development 119 (2002) 21–34

for, and sufficient to induce these two patterning processes (reviewed in Matise and Joyner, 1999). The biochemical functions of a Gli family protein have been determined in greatest detail in Drosophila (reviewed in Ingham and McMahon, 2001). The Gli homolog, Ci, has a centrally located zinc finger domain, an N-terminal repressor domain and a C-terminal activation domain. Ci is cleaved downstream of the zinc finger domain in the absence of Hh signaling to produce a protein that contains the zinc finger domain and N-terminal repressor domain. Furthermore, this N-terminal truncated Ci protein is required to repress some Hh target genes (Aza-Blanc et al., 1997; Robbins et al., 1997). In the presence of Hh, Ci cleavage is inhibited and the full-length protein is modified such that it becomes a potent transcriptional activator (Ohlmeyer and Kalderon, 1998; Methot and Basler, 1999). Biochemical and mis-expression studies have shown that the three Gli vertebrate genes have unique activities (reviewed in Matise and Joyner, 1999; Stone and Rosenthal, 2000; Ingham and McMahon, 2001). Studies of the vertebrate Gli proteins in transgenic flies and cells in vitro have defined some of the functions of the Gli proteins. In the absence of Shh, mouse Gli3 is cleaved into an N-terminal truncated form that contains the zinc finger domain and a repressor domain (Dai et al., 1999; Sasaki et al., 1999; Wang et al., 2000; Aza-Blanc et al., 2000). Gli3 also contains a weak C-terminal activation domain and cleavage is inhibited by Shh. Since Gli3 is primarily expressed at a distance to sources of Shh, in general Gli3 appears to act as a repressor. Gli2 is also cleaved into an N-terminal form that contains a repressor domain and has a C-terminal activator domain (Sasaki et al., 1999). Our recent genetic studies in mouse have shown that Gli2 primarily functions as a Shhdependent activator, and unlike Gli3, any repressor form of Gli2 that is present in mice is not required for normal development (Bai and Joyner, 2001). In contrast to Gli2 and 3, Gli1 is not cleaved and does not contain an N-terminal repressor domain. Furthermore, when expressed ectopically in the neural tube, Gli1 can mimic Shh and induce floorplate genes, and co-expression of Gli2 or Gli3 with Gli1 inhibits this activation (Lee et al., 1997; Hynes et al., 1997; Ruiz i Altaba, 1998; Park et al., 2000). Studies of Gli mutant mice have shown that the three genes indeed have unique and overlapping functions. Whereas Gli1 null mutant mice are viable and have no obvious phenotypes, Gli1/2 double mutant studies (Park et al., 2000; Bai et al., submitted) and replacement of Gli2 with Gli1 (Bai and Joyner, 2001) demonstrated overlapping functions of the proteins in activating Shh targets. Gli2 and Gli3 mutants each are lethal by birth and double mutants have more extreme phenotypes than single mutants indicating overlapping functions (Hui and Joyner, 1993; Mo et al., 1997; Motoyama et al., 1998; Hardcastle et al., 1998). The polydactyly and dorsal CNS defects in Gli3 mutants and phenotype of Shh/Gli3 double mutants show that Gli3 normally represses the Shh pathway (Litingtung and

Chiang, 2000). Taken together, the studies indicate that Shh is required in some cells (the floorplate) to induce activator forms of Gli proteins, whereas in other cells (motor neurons) it is necessary for inhibiting cleavage of Gli3 into a repressor form. It is interesting to note that the phenotype of Gli1/2 double mutants is much milder than Shh mutants, raising the question of whether removing the repressor activity of Gli3 is essential to development of many tissues in the mouse, or whether genes other than the three identified Gli genes act downstream of Shh. The Zic family of C2-H2 zinc finger proteins show notable homology to the Gli proteins in the DNA binding domain (Aruga et al., 1994). Mice carrying mutations in both Zic1 and Gli3 have a more extreme defect in the segmentation of the vertebral lamina than either single mutant, suggesting an interaction between Zic and Gli proteins (Aruga et al., 1999). Consistent with this, Zic1 can bind to Gli binding sequence, but with less affinity than the Gli proteins (Mizugishi et al., 2001). Furthermore, physical and functional interactions between Zic and Gli proteins through their zinc finger domains have been shown. In many cell lines Gli proteins are found primarily in the cytoplasm, however, Gli proteins become translocated to the nucleus when they are coexpressed with Zic proteins (Koyabu et al., 2001). In addition, in co-transfection assays, Gli and Zic proteins can regulate each other’s transcriptional activity (Mizugishi et al., 2001). An additional gene, Glis2/Nk1, that encodes a zinc finger domain similar to that of the Gli and Zic proteins was recently cloned from numerous vertebrates and found to be expressed in the CNS, somites and kidney (Lamar et al., 2001; Zhang et al., 2002). The protein has proline rich domains and a VP16-fusion protein mimics the wild-type protein in inducing neurogenesis, and the zinc finger domain can bind the B1 Gli consensus DNA binding sequence (Lamar et al., 2001). A structure–function study in tissue culture cells indicated that the protein may have repressor domains, in addition to activation domains (Zhang et al., 2002). To examine whether additional Gli-like zinc fingercontaining genes exist, we used degenerate reverse transcription–polymerase chain reaction (RT–PCR) based on Gli zinc-finger motif sequences. We identified a novel gene, GliH1, that encodes a zinc finger domain that is more than ,60% identical to that of the Gli proteins, and less similar to that of the Zic and Glis2/Nk1 proteins. In general GliH1 is expressed later than the three Gli genes and in a more restricted manner. The zinc finger domain of GliH1 can bind to Gli DNA binding sequences, and in a cotransfection assay in odontoblasts GliH1 can inhibit the activation activity of Gli1. Finally, gene targeting was used to disrupt GliH1 in mice, and GliH1 mutants were found to be viable and have no obvious phenotype. Furthermore, no evidence for a role of GliH1 in Shh signaling could be determined by examining GliH1/Gli1 or Gli3 double mutants. A second gene with a very similar zinc finger

M. Nakashima et al. / Mechanisms of Development 119 (2002) 21–34

domain is present in mouse, which could have overlapping functions with GliH1. These studies identify a novel family of genes that encode proteins with a DNA binding domain homologous to the Gli proteins, and that can inhibit Gli1 activation of a reporter.

2. Results 2.1. Identification of a novel gene with homology to the Gli zinc finger domain RT–PCR with degenerate primers (see Section 4) was carried out to identify novel genes encoding zinc finger domains similar to those in the Gli genes. RNA from rat incisor pulp cells was used because conditional mutant analysis (Dassule et al., 2000) and gain-of-function studies (Hardcastle et al., 1998) have shown that Shh plays a key role in tooth development. During tooth development, Shh is localized to the inner enamel epithelium and Glis to the epithelial and mesenchymal components of the developing tooth germs (Bitgood and McMahon, 1995; Iseki et al., 1996; Hardcastle et al., 1998; Hui et al., 1994). One novel cDNA was identified, in addition to four clones of Gli1 and two clones encoding Tax helper protein 2. A fulllength mouse cDNA was cloned and sequenced and designated GliH1 (Gli homologous gene; GenBank accession number AF220434) (see Section 4). Like members of the Gli family, the novel gene encodes a protein with five copies of a C2H2 type zinc finger motif. A comparison of the zinc finger domain of the novel cDNA compared to the mouse Gli, Zic and Glis2/Nk1 genes and fly ci is shown in Fig. 1A. In the zinc finger domain, the novel gene is related to Gli1, Gli2 and Gli3 with 61, 64 and 62% amino acid sequence identity, respectively. In contrast, the novel gene has approximately 50% amino acid identity with the zinc finger domains of the four mouse Zic proteins and Glis2/Nk1 (Aruga et al., 1996; Zhang et al., 2002). Comparison of the entire 612 amino acid coding sequence of GliH1 with those of the Glis, Zics and Glis2/Nk1 indicated no obvious homology outside the zinc finger region. The length of the GliH1 protein is more similar to the Zics (,530 amino acids (aa)) and Glis2/Nk1 (520 aa) than the Glis (1111–1600 aa). A phylogenic tree comparing both the full-length coding sequences and only the zinc finger domains of the mouse Gli and Zic genes and ci to GliH1 placed GliH1 between the Gli and Zic gene families (Fig. 1B and data not shown). When the full-length sequence of the novel cDNA was compared to the mouse sequences in the Celera Genomics and associated databases, two potential exon sequences with high homology to the zinc finger region were found (Fig. 1C). One sequence (mCP25772) encodes a protein region similar to zinc fingers 1 and 2 of GliH1 with 95% amino acid identity, and the other sequence (mCP2720) shares 84% amino acid identity with fingers 3–5 of GliH1. The two

23

potential exons are located near each other on chromosome 19 and therefore may represent exons from the same gene. Since GliH1 is on mouse chromosome 4, there are at least two GliH1 genes with highly conserved zinc finger domains. Furthermore, the splice donor at the end of the potential exon encoding zinc fingers 1 and 2 and splice acceptor in the other potential exon is conserved with a splice site in GliH1. Moreover, this splice site is conserved in the three Gli genes. GliH1 and the Glis share another splice site at the end of zinc finger four, and the Glis have an additional splice site that interrupts zinc finger one (see Fig. 1A). Of interest in this regard, the Zic genes have two different splice sites in the zinc finger domains and Glis2/Nk1 has yet three different splice sites (See Fig. 1A). Thus, the exon structure of GliH1 is more similar to the Gli genes, than the Zics and Glis2/Nk1. 2.2. GliH1 can bind Gli DNA binding sites Since GliH1 protein is predicted to have a similar DNA binding domain to that of the Gli proteins, we tested whether GliH1 could bind the previously defined Gli DNA binding sites, type A, B and C (Kinzler and Vogelstein, 1990). Zic proteins can also bind to these and other Gli DNA binding sites (Aruga et al., 1994), but the DNA-binding affinity of the Zic proteins is lower than that of Gli (Mizugishi et al., 2001). Using a gel shift assay (see Section 4) and zinc finger domain-GST fusion proteins, the binding activity of the zinc finger domain of GliH1 was compared to that of Gli1 and Gli3. GliH1 was found to bind to type A and B Gli DNA binding sites, whereas Gli1 primarily bound to type B and to a lesser extent to type A and C.Gli3 in contrast bound all three Gli DNA binding sites strongly (Fig. 2A). As expected, GliH1, Gli1 and Gli3 did not bind to mutant type A, B and C sequences. The GST protein alone also did not bind to any of the oligos (data not shown). In the presence of a 100 fold excess cold oligo, the DNA binding of GliH1to type A and type B sites was severely reduced (Fig.2B), suggesting that the binding properties of GliH1 protein to type A and type B is comparable. These binding studies indicate that the zinc finger domains of Gli1, Gli3 and GliH1 have different binding activities and that GliH1 can selectively bind to type A and type B oligos, but not to type C. An 800 bp XhoI–EcoRI fragment containing the Gli1 promoter region and exon 1 sequences (Dai et al., 1999) has been shown to bind Gli3 protein (Dai et al., 1999). We therefore tested whether GliH1 can also bind this fragment, which has different Gli binding sequences than the type A, B and C sites. The GliH1 zinc finger-GST fusion protein was found to bind the fragment, however, the GliH1–DNA complex was smaller than the multiple retarded bands seen with the Gli3 zinc finger-GST fusion protein, but not smaller than the smallest complex formed with the lower amount (0.1 mg) of GST–Gli3 (Fig. 2C). Dai et al. (1999) suggested that the multiple bands indicate that

24

M. Nakashima et al. / Mechanisms of Development 119 (2002) 21–34

full-length Gli3 binds multiple sites in the Gli1 promoter sequence. Our finding that GliH1 only produces one band shift might indicate that GliH binds less sites in the 800 bp fragment than Gli3. The GST-Gli1 zinc finger fusion and the GST protein alone did not bind to this fragment (Fig. 2C and data not shown). When a 10-fold excess of normal type A or type B cold oligos were added with the 800 bp Gli1 fragment, the oligos competed with the fragment and binding to GliH1 was abolished. Furthermore, mutant type A or type B oligos did not compete for binding with the fragment (Fig. 2D), indicating that the protein–DNA complex formed

contains an activity that can also bind to oligos type A and type B. 2.3. GliH1 can repress Gli1 stimulated reporter expression in dental pulp cells Since GliH1 can bind specific DNA sequences, we were interested in determining whether it has transcriptional activity in cell culture. A luciferase reporter plasmid containing six tandem copies of the Gli type B binding consensus sequence was transfected alone or in combination

Fig. 1. Comparison of amino acid sequence of GliH1 to the Glis, Zics and Glis2/Nk1 (Glis2). (A) Amino acid sequence comparison of the five zinc finger domains (ZF1-ZF5) of the indicated genes with the length of the zinc finger and percent homology to GliH1 shown at the end. The C2-H2 type, conserved cysteine and histidine residues of the zinc finger motif are in red. GliH1 has 61, 64 and 62% identities to Gli1, Gli2 and Gli3, respectively. The red bars show the splice sites. (B) Phylogenic tree of the zinc finger domains of Gli-related genes. GliH1 is located between Gli members and Zics. The Gli1, 2 and 3, Zic1, 2, 3, Ci and Glis2/Nk1 sequences are taken from GenBank. (C) Comparison of amino acid sequence of GliH1 to mCP2720 and mCP25772 generated through use of the Celera Discovery System and Celera genomic’s associated database (http://www.celera.compublicationlibrary).

M. Nakashima et al. / Mechanisms of Development 119 (2002) 21–34

with Gli1, Gli3 and/or GliH1 expression vectors into bovine dental pulp cells and reporter activity determined. Gli1 expression resulted in a nearly 50 fold increase in reporter

25

luciferase activity in pulp cells (Fig. 3). In contrast, transfection of Gli3 or GliH1 expression vectors did not result in a detectable increase in luciferase activity. When a reporter was used with mutated Gli binding sites, no detectable activity was seen with any of the expression vectors. Next we examined the influence of Gli3 and GliH1 on the stimulation of luciferase activity by Gli1 in dental pulp cells. Cotransfection of a Gli3 or GliH1 expression vector with Gli1 resulted in a reduction in the Gli1-stimulated luciferase activity in dental pulp cells, and co-expression of both Gli3 and GliH1 with Gli1 resulted in a further inhibition of the activation (Fig. 3). These results raise the possibility that GliH1 contains a repressor domain that can inhibit the stimulation of a luciferase reporter containing Gli DNA binding sites by Gli1. Alternatively, GliH1 could be acting only as a competitor and does not have an activator or repressor domain. 2.4. GliH1 expression during embryogenesis is limited to specific tissues after 10 dpc Since GliH1 can bind some of the same DNA elements as

Fig. 2. GliH1 gel mobility shift assay. (A) Three different Gli-binding sequences, type A, type B, type C, containing a 9 bp conserved sequence 5 0 -GACCACCCA-3 0 , were incubated with GST–GliH1, GST-Gli1, and GST-Gli3 fusion proteins containing the zinc finger DNA binding domain. As controls, mutant oligos were used in which CACCC was replaced by GTGGG (Am, Bm and Cm). 0.5 or 0.1 mg of the GST fusion proteins were used in the experiment. (B) In the presence of 100£ or 10£ excess cold oligo A, B or C, GST–GliH1 fusion protein was incubated with radiolabeled type A or type B oligos. (C) The Gli1 promoter, XhoI–EcoRI 800 bp fragment was incubated with 0.5 or 0.1 mg of the GST fusion proteins. (D) In the presence of 250 £ , 100 £ , 50 £ , 10£ excess cold type A, Am, B, Bm oligos, GST–GliH1 fusion protein was incubated with the Gli1 promoter fragment.

Fig. 3. Repression of Gli1 reporter activation by Gli3 and/or GliH1. Dental pulp cells were transfected with the wild-type or mutant Gli binding site reporters, 100 ng of the Gli1, Gli3 or GliH1 expression constructs and an internal control plasmid pRL-SV, and incubated for 2 days. The numbers represent the average luciferase activity from three independent experiments and normalized by dividing by the internal control.

26

M. Nakashima et al. / Mechanisms of Development 119 (2002) 21–34

the Gli proteins and potentially antagonize or enhance their function, it was of interest to determine whether GliH1 is expressed in any of the same regions of the embryo as the Gli genes. Whole-mount RNA in situ hybridization analysis of mouse embryos did not detect expression of GliH1 until 10 days post coitus (dpc) and at this stage expression was primarily in ventral regions of the mesenchyme of the limb buds and tail (Fig. 4A). Additional expression of GliH1 in the developing maxillary process, mandibular arch and hyoid arch was seen at 10.5 dpc (Fig. 4B). Overall, expression of GliH1 became stronger at 11.0 dpc (Fig. 4C). Unlike the Gli, Zic and Glis2/Nk1 genes, GliH1 expression was absent from the developing neural tube at the stages analyzed (from 7.5 to 12.5 dpc). At 11.5 dpc GliH1 expression in the craniofacial interface became broader and stronger in the mesoderm (Fig. 4E). GliH1 was also expressed in the mesodermal cells of the whisker follicles (Fig. 4E), likely adjacent to Shh expressing ectoderm. At later stages of embryogenesis, 16.5 dpc, Northern blot analysis of RNA from various tissues showed that GliH1 was expressed at highest levels in the kidney (data not shown), where Glis2/ Nk1 is also expressed at highest levels (Zhang et al., 2002), as well as in the testis. GliH1 expression was examined in Shh 2/2 mutant embryos to determine whether GliH1 expression is dependent on Shh, similar to Gli1. In Shh 2/2 mutants the craniofacial region forms a long proboscis. GliH1 expression in the mandibular arch, the craniofacial border and limb was nevertheless present in Shh 2/2 embryos at 11.0 dpc and 11.5 dpc (Fig. 4F,G). These studies show that initiation of GliH1 expression is not regulated by Shh signaling. 2.5. GliH1 expression overlaps primarily with Gli2 and 3, but is much more restricted To determine whether expression of GliH1 overlaps with, or is adjacent to the Gli genes and Shh, expression of the genes was compared on adjacent sections in several tissues, including the neural tube, tail, molar teeth, kidney and craniofacial and vertebral regions. Transverse sections of an 11.5 dpc embryo through the craniofacial region showed that GliH1 was expressed in a very limited ventral region of the mandible, overlapping with Gli2 and Gli3, which are expressed much more broadly (Fig. 5A–E). At 11.5 dpc, sagittal sections through the body in the vertebral region (Fig. 5F–J) showed that GliH1 was expressed only in a small region of the intervertebral space, overlapping with the broader domains of Gli expression. Cross sections through the body at the forelimb level showed that GliH1 is indeed not expressed in the neural tube, whereas the Gli genes are expressed in specific dorsal/ventral domains (data not shown). In the posterior tail (Fig. 5K–O), GliH1 expression was limited to the most ventral mesoderm overlapping primarily with the broadly expressed Gli2 and 3 genes. Gli1 was primarily detected more dorsally than GliH1 in the mesenchyme

Fig. 4. Whole-mount RNA in situ hybridization analysis of GliH1 embryonic expression in wild-type (A–E) and in Shh 2/2 mutant (F,G) embryos. (A) At 10.0 dpc the forelimb (fl), hindlimb (hl) and tail (t) express GliH1. (B) At 10.5 dpc the anterior region of the forelimb (arrowhead), ventral part of the body and tail (arrow), craniofacial border beneath the forebrain (cb), maxillary process (mp), mandibular arch (ma) and hyoid arch (ha) express GliH1. (C) At 11.0 dpc GliH1 expression becomes stronger. (D) Dorsal view showing somite expression at 11.0 dpc. (E) At 11.5 dpc. the whisker follicles (wf) also express GliH1. Expression of GliH1 in the anterior forelimbs and hindlimbs is shown with arrowheads. (F) Dorsal view at 11.0 dpc of Shh 2/2 mutant showing that GliH1 expression in the anterior part of forelimb is more extended to the posterior region (arrowhead) compared with wild-type embryos. Expression around the somites is broader in Shh 2/2 mutants than in wild-types. (G) GliH1 expression at 11.5 dpc in Shh 2/2 mutant. Note that GliH1 expression in the forelimb is seen from anterior to posterior. Craniofacial border is indicated by arrowheads.

M. Nakashima et al. / Mechanisms of Development 119 (2002) 21–34

and Shh is primarily expressed in the notochord and floorplate. At 15.5 dpc, GliH1 was found at highest levels in the dermal cells surrounding both the vibrissae and first wave of pelage hairs, whereas Gli1, Gli2 and Gli3 were expressed at higher levels in the ectoderm of the hair bud (Fig. 6A–D and data not shown). Shh is also expressed at the base of the ectoderm. At postnatal day 1, GliH1 was weakly expressed in the ectodermal cells at the base of the hair follicle and in the dermal papillae (Fig. 6F). At E15.5 in the eye, Gli1, Gli2 and Gli3 were detected in the neuroblast layer of the retina and Shh in the developing ganglion cell layer, whereas no GliH1 expression could be detected in the retina (Fig. 7A–E). In the lens however, GliH1 expression was found to overlap with Gli3 throughout the basal layer of the anterior lens epithelium and with Gli2 only in the transitional zone. Shh and Gli1 were not detected in the lens. GliH1 also was strongly expressed in the presumptive iris and ciliary body, overlapping with weak expression of Gli2 and 3. In the eyelids, which are fused at this stage, Gli1, Gli2, Gli3 and GliH1 were expressed in the mesenchyme surrounding the site of fusion

27

and Shh was detected in tiny symmetrical regions of the epithelium (Fig.7A–E). Section in situ analysis of the kidney showed that GliH1 could first be detected at 13.5 dpc in mesenchymal condensations and comma shaped nodes in the kidney (Fig. 8A). GliH1 expression was absent from the ureteric bud epithelium, where Glis2/Nk1 has been reported to be expressed in rat embryos (Zhang et al., 2002). Shh transcripts were present in the ureteric bud and not the mesenchymal aggregates and comma shaped nodes (data not shown). At 14 and 15.5 dpc, GliH1 also was detected in the S-shaped tubules (Fig. 8B,C), but not in the regionally differentiated glomerulus. Indian Hh (Ihh) was seen in the collecting ducts and renal tubules (Fig. 8H), whereas Shh was restricted to the collecting ducts (Fig. 8G). At 2 weeks of age postnatally, GliH1 continued to be expressed in renal tubules, but not differentiated glomeruli in the cortex and Ihh, but not Shh, was still expressed in the cortex (data not shown). GliH1 expression in the kidney did not appear to have a similar pattern to Gli1, 2 or 3 at any stages. In the developing tooth, Shh has been found to be expressed as early as the cap stage (14.5 dpc) in the enamel

Fig. 5. Section RNA in situ hybridization analysis comparing GliH1 expression to the Glis and Shh. (A–E) Sagittal sections through the craniofacial region, showing GliH1 expression in the craniofacial border beneath the forebrain (cb), maxillary process (mp), and mandibular arch (ma) at 11.5 dpc. (F–J) Sagittal sections of 11.5 dpc embryo showing GliH1 expression in intervertebral disc. (K–O) Sections through the posterior tail. Note that GliH1 expression is limited to the ventral part and not expressed in the neural tube where the Gli genes are. The domains of expression in the mesoderm are indicated with arrows.

28

M. Nakashima et al. / Mechanisms of Development 119 (2002) 21–34

generated using gene targeting in embryonic stem (ES) cells (see Section 4 for details). The mutant allele that was made contains lacZ in place of 840 bp of exon 2 that includes the start codon and the first two zinc finger domains (Fig. 10A). Two independently derived germ-line transmitting chimeras were mated with C57Bl/6 or outbred Swiss Webster mice to produce GliH1 lz/lz mutant homozygotes on both backgrounds and both were found to be fully viable and appeared normal. To address the nature of the GliH1 lz/lz mutant allele, Northern blot analysis was carried out. The normal 3 kb GliH1 transcript was not detected in RNA from adult GliH1 lz/lz homozygotes using a probe containing the last three zinc finger domains (nucleotides 1393–1714) or a 3 0 probe containing the stop codon (nucleotides 2150–2520) (data not shown). LacZ expression in heterozygous or homozygous GliH1 lz/lz embryos resembled the GliH1 expression detected using RNA in situ analysis (Fig. 10D and data not shown), providing further evidence that the GliH1 lz/lz mutant is probably a null allele. LacZ expression was most prominently seen at 10.5 dpc in the anterior

Fig. 6. GliH1 expression compared to the Glis and Shh in hair follicles. (A– E) At E15.5 dpc GliH1 is expressed at low levels in the invaginating and ectoderm and in the adjacent mesoderm (D). Note the highest expression of Gli1 (A), Gli2 (B) and Gli3 (C) in the center of the hair bud and Shh (E) at the bottom of hair bud. GliH1 is weakly expressed in the ectodermal cells at the base of the follicle and in the dermal papillae at P1 (F).

epithelium, and Gli1 in a region of the tooth germ surrounding Shh. Gli2 is expressed in the dental follicle and weakly in the dental papilla at this early stage, and Gli3 in the dental papillae and dental follicle (see Fig. 9B–E) (Bitgood and McMahon, 1995; Iseki et al., 1996; Hardcastle et al., 1998; Hui et al., 1994). GliH1 was very weakly detected at the late cap stage (15.5 dpc) in the dental papillae near the apical region (Fig. 9A). At the bell stage (P1), GliH1 expression was highest in a region of the dental papillae (Fig. 9F), overlapping primarily with the more broadly expressed Gli1 (Fig. 9F–I). Strong expression of Shh remained in the epithelium (Fig. 9J). 2.6. Generation of GliH1 mutant mice expressing lacZ To determine whether GliH1 is required during development or in the adult, a likely null mutant allele of GliH1 was

Fig. 7. GliH1 expression in the eye at 15.5 dpc. Gli1, Gli2 and Gli3 genes expressed in the neuroblast layer of retina (arrows), Shh in the developing ganglion cell layer (arrow). In the lens, GliH1 and Gli3 are expressed in the basal layer of the anterior lens epithelium and Gli2 is expressed only in the transitional zone (arrowhead). In the transitional zone of the lens fiber cells (arrowheads), GliH1, Gli2 are expressed. Note the strong expression of GliH1 in presumptive iris and ciliary body ( p ).

M. Nakashima et al. / Mechanisms of Development 119 (2002) 21–34

hindlimbs of mutant embryos and at 11.5 dpc in the metanephrogenic mesenchyme of the urogenital system that gives rise to the kidney. It was also expressed in the maxillary process, and hyoid arch. Homozygous GliH1 lz/lz mutant animals up to 6 months of age had no obvious phenotype, despite detailed observation of the animals in their cages and extensive section analysis of the kidneys and teeth. Furthermore, double homozygous mutant GliH1 lz/lz;Gli1 zfd/zfd mutants also survive to at least 3 months and appear normal, and double homozygous GliH1 lz/lz;Gli3 xt/xt embryos appear to have polydactyly and dorsal brain defects similar to Gli3 mutants and no obvious tooth defects at 16.0 dpc.

29

mutants the expression of GliH1 was not altered, other than reflecting the mutant defects in such embryos, showing that unlike Gli1, GliH1 is not regulated by Shh.

3. Discussion In this study we identified a novel gene encoding a protein with five C2H2 zinc finger domains homologous to the Gli proteins, that is expressed during development in a limited number of tissues, including the limbs, teeth and kidney. Of significance, GliH1 is more closely related to Gli1, Gli2 and Gli3 than members of the Zic family or Glis2/Nk1 in the zinc finger domain, and contains no obvious regions of homology outside this region. Furthermore, whereas Gli1, Gli2 and Gli3 range from 1111 to 1596 amino acids in length, GliH1 is only 619 amino acids in length, more similar to the length of the Zic and Glis2/Nk1 proteins. Consistent with the conserved zinc finger domain, we found that GliH1 can bind previously defined Gli DNA binding sites. Of interest, we found that Gli1, Gli3 and GliH1 have different DNA binding activities for different Gli DNA binding sites. Using a luciferase reporter containing six tandem copies of the Gli type B DNA binding consensus sequence and transfection of bovine dental pulp cells with Gli and GliH1 expression constructs we found that unlike Gli1 which had strong activator function, GliH1 had no effect on transcription. Gli3 also had not effect on the reporter construct in dental pulp cells. However, when either Gli3 or GliH1 were co-transfected with Gli1, they reduced the activator function of Gli1. Thus, GliH1 can potentially act as a transcriptional repressor. Based on the functional and structural similarities of GliH1 to the Glis, we examined the temporal and spatial patterns of expression of GliH1 to determine whether there was overlap with the Glis or Shh. GliH1 expression begins much later than the three Gli genes and has a more limited expression pattern. GliH1 expression was first detected at 10 dpc in the posterior limb mesenchyme and then in the craniofacial interface of the head mesenchyme, and in the mesenchyme of the mandibles and tail. By midgestation, GliH1 is also expressed in the kidney, lens of the eye, hair mesenchyme, and tooth mesenchyme. Unlike the Gli, Zic and Glis2/Nk2 genes, GliH1 is not expressed in the nervous system. In general, GliH1 expression overlaps most with Gli3, although Gli3 is expressed much more broadly than GliH1. In Shh 2/2

Fig. 8. GliH1 expression in the kidney compared to the Glis and Shh. (A) GliH1 expression is seen in mesenchymal condensate (mc) and commashaped body (cb), and is absent from ureteric bud (ub) at 13.5 dpc. (B) GliH1 is expressed primarily in S-shaped tubules (rt) and not in more differentiated glomeruli at 14.5 dpc. (C) At 15.5 dpc, GliH1 is expressed in proximal or distal renal tubules (rt), but not in the regionally differentiated glomerulus. (D) Gli1 is expressed in stroma (st). (E) Gli2 is also expressed in the stroma (st) and glomerulus (gl). (F) Gli3 is expressed in glomerulus (gl). (G) Shh is expressed in collecting duct (ct). (H) Ihh is expressed in collecting ducts (cd) and the renal tubules (rt).

30

M. Nakashima et al. / Mechanisms of Development 119 (2002) 21–34

that GliH1 can function as a repressor, and not an activator like Gli1. Since Gli3 is processed into an N-terminal repressor form and we found that both GliH1 and Gli3 can repress Gli1 activator function in dental pulp cells, we also made GliH1 lz/lz; Gli3 xt/xt double homozygous mutants. The double mutants, however, appear grossly similar to Gli3 mutants and histological examination of the teeth of double mutants revealed no defects, at least at 16 dpc. Further analysis of additional double and triple mutants with GliH1 lz and other genes containing Gli-like zinc finger domains should determine whether GliH1 lz has overlapping functions with other genes and reveal the function of GliH1 in development. Of note in this respect, Glis2/Nk2 is expressed at highest levels in the developing rat kidney, although apparently not in the same cell types. Furthermore, we identified potential exon sequences from the Celera mouse data base that suggest there is at least one additional gene with a highly conserved zinc finger domain with GliH1.

4. Experimental procedures 4.1. Cloning of a cDNA with a Gli-like zinc finger domain

Fig. 9. GliH1 expression during tooth development compared to the Glis and Shh. Sections in the tooth germ at the late cap stage, 15.5 dpc (A–E), and at the bell stage (F-J), postnatal day 1 (P1). At the late cap stage, GliH1 was very weakly detected in the dental papillae near the apical region (A), and expression was stronger in the dental papillae at the bell stage (F). At the late cap stage and bell stage Gli1 was expressed in a broad region of the tooth germ (B,G) surrounding Shh in the enamel epithelium (E,J). Gli2 was weakly expressed similarly to Gli1 expression except in odontoblasts at the bell stage (C,H). Gli3 was localized mainly in dental papillae and dental follicle at the late cap stage (D) and broadly at the bell stage (I). dp, dental papillae; ee, enamel epithelium; df, dental follicle; od, odontoblasts.

Similar to Gli1 mutant mice (Park et al., 2000; Bai et al., in press), GliH1 null mutants are viable and appear normal. One possible reason for the lack of phenotype is that Gli1, 2 or 3 have overlapping functions with GliH1. However, our finding that GliH1 lz/lz;Gli1 zfd/zfd are viable and have no obvious phenotype, shows that Gli1 and GliH1 do not have any critical overlapping functions. This is consistent with our finding

RNA preparation, cDNA cloning and DNA sequencing was performed as previously described (Toyono et al., 1997). In brief, the apical portion and the entire pulp of rat incisors was isolated, and first-strand cDNA was prepared from poly(A) 1RNA.One set of degenerate primers based on the conserved amino acid alignment of the five zinc finger encoding domains of mouse Gli1, Gli2 and Gli3 were used; MT-G1 (5 0 -GCGGAATTC(C/A)(G/A)ICC(C/T/A)TT(T/ C)AA(A/G)GC(T/C/A)CA(A/G) TA-3 0 ) and MT-G2 (5 0 GCGGAATTCGG(A/G)TC(G/T/A)GT(A/G)TA(G/C/T)C(G/T)(T/C)TT-3 0 ). PCR was carried out at 94 8C for 1 min, 50 8C for 2 min, and 72 8C for 2 min for 40 cycles, followed by extension at 72 8C for 10 min. RT–PCR products of approximately 290 bp were cloned into the pCRII cloning vector (TA Cloning Kit, Invitrogen, San Diego, CA). 5 0 RACE of GliH1 was done using the 5 0 -RACE system for Rapid Amplification of cDNA Ends (Invitrogen Life Technologies Inc.) according to the manufacturer’s protocol using the GliH1 specific primer MT-G7 (5 0 -TTCTGGCAGCCTGGGTGCTGGCACAGGTA-3 0 ) for a firststrand synthesis and the nested primer MT-G8 (5 0 TGTTGGGCTTCTCGCCCGAGTGTAC-3 0 ). PCR reamplification was further performed with MT-G7 and the Universal Amplification Primer (UAP). Sequence comparisons were done using ClustalW. Approximately 5 £ 105 clones from a random-primed mouse cDNA library made from adult kidney RNA were screened using a rat cDNA probe (420 bp, ApaI–RsaI) generated by 5 0 -RACE PCR amplification. Four positive clones were subcloned into pBluescript II KS(2) and sequenced. 3 0 RACE of GliH1 was further performed using the GliH1specific primer MOL 4 (5 0 -CTACCAAGGCAGTTTC-

M. Nakashima et al. / Mechanisms of Development 119 (2002) 21–34

31

Fig. 10. Strategy used to make a lacZ knock-in null allele of GliH1. (A) 850 base pairs of exon 2 including 5 0 untranslated sequences and all the N-terminal coding sequences up to the 3rd zinc finger DNA binding domain were replaced with LacZ poly(A) and a loxP-neo-loxP cassette. Homologous recombination was confirmed by Southern blot analysis using the indicated 5 0 and 3 0 external probes. After transmitting the mutation into mice, the GliH1–lacZ allele was detected using mouse tail DNA and Southern blot analysis with the 5 0 probe after KpnI digestion (B) and the 3 0 probe after BamHI digestion (C). K, KpnI; B, BamHI. Boxes indicate exons. Pink marks the 5 0 untranslated sequences, and orange marks the coding region. (D) Expression of lacZ from the GliH1 allele in E10.5 heterozygous targeted embryo. Cb, craniofacial border beneath the forebrain; mp, maxillary process; ma, mandibular arch; ha, hyoid arch. The arrow indicates the anterior region of the hindlimb.

CATTCCATCCAG-3 0 ) and the abridged universal amplification primer (AUAP) (Life Technologies Inc.). 4.2. Production of a null mutation in GliH1 by gene targeting A mouse 129/SvEv genomic library (Stratagene) was

screened using two cDNA fragments of 429 bp (an EcoRI–Pst I fragment including nucleotides 1–429,) and 964 bp (a Pst I fragment including nucleotides 429–1393) from the 5 0 -untranslated region and zinc finger coding region, respectively, of a mouse GliH1 cDNA.Overlapping GliH1 phage clones were isolated and restriction mapped that include the two 5 0 -most exons. Exon one is an untrans-

32

M. Nakashima et al. / Mechanisms of Development 119 (2002) 21–34

lated exon and exon two contains the start codon and the first two zinc finger domains. To generate a targeting vector with positive-negative selection, a 5.7kb SmaI–XhoI fragment located 3 0 to exon 2 was subcloned into the ClaI/XhoI site of pKSlox PNT, 3 0 to the neo cassette. A 3.1 kb BamHI– SalI genomic fragment was ligated to SDK LacZ polyA (Logan et al., 1993) and then subcloned into the targeting vector 5 0 to the neo cassette. The tk counter selection cassette was positioned 5 0 to the 3.1 kb genomic fragment. Following targeting, the LacZ-polyA-loxP-neo-loxP cassette would replace the coding sequences of exon 2 (see Fig. 10) such that LacZ would be inserted into the 5 0 untranslated sequences of GliH1. W4 ES cells (Auerbach et al., 2000) were electroporated with 50 mg of XhoI-linearized plasmid, plated, and selected in G418 (150 mg/ml) and Gancyclovir (2 mM) as described (Matise et al., 2000). Drug-resistant colonies were picked into 96-well plates and screened by Southern blot analysis. DNA samples were digested with KpnI and probed with a 5 0 external probe (1.0 kb, EcoRV–BamHI genomic fragment) or with BamHI and probed with a 3 0 external probe (260 bp, EagI–EcoRI genomic fragment). GliH1 targeted clones were identified by Southern blot hybridization using both 5 0 - and 3 0 -flanking probes at a frequency of 2%. Three correctly targeted ES cell clones were injected into C57Bl/6 host blastocysts to generate chimeric animals (Papaioannou and Johnson, 2000). Male chimeras were bred with C57Bl/6 females and two of the cell lines were found to generate germ-line transmission. Germ-line transmitting chimeras were mated with C57Bl/6 or outbred mice. Homozygous mutants from two independently targeted ES cell lines were analyzed and found to have the same lack of a phenotype on both genetic backgrounds. F1 progeny heterozygous for the GliH1 mutation were identified by Southern blot analysis of 10 mg of tail genomic DNA samples using the 3 0 probe. Subsequent progeny and embryos were genotyped by PCR.The PCR primers were designed to use a common 3 0 primer located in the second intron (5 0 -AGGAACCTGTCTTGGAAGTACACT-3 0 ), and specific primers for the neo gene (5-CTATCAGGACATAGCGTTGG-3 0 ) or GliH1 exon 2 (5 0 -AAGCCCTTCAATGCCCGCTACAA-3 0 ). These primers generate products of 1 kb specific for the mutant allele and 450 bp for the wild-type. 4.3. Gel mobility shift assays For probes, three different Gli–DNA binding sequences containing a 9 bp conserved sequence (5 0 -GACCACCCA3 0 ) were made using synthetic oligonucleotides: type A (5 0 CRAAGACCACCCACAATGATGGT-3 0 ), type B (5 0 GCGTGGACCACCCAAGACGAAATT-3 0 ), and type C (5 0 -AGATAGACCACCCAGCTTCAGGT-3 0 ) (Kinzler and Vogelstein, 1990). The sequences containing the zinc finger DNA binding domains from a mouse GliH1 cDNA (nucleotides 1059–1644) and a mouse Gli1 cDNA (nucleo-

tides 870–1431) were placed into a pGEX vector in frame to produce GST fusion proteins. GST fusion proteins were expressed in Escherichia coli and purified over glutathione Sepharose using standard procedures (Ausubel et al., 1996). The gel shift assays were carried out as described (Williams and Tjian, 1991). pGST-Gli3 was kindly provided by Dr. Ishii (Dai et al., 1999). The three Gli DNA binding site double stranded oligos, 0.25 pmol, were incubated with 0.5 or 1.0 mg of purified GST–GliH1, GST–Gli1, and GST–Gli3 fusion proteins and GST protein alone. As controls, mutants in which CACCC was replaced by GTGGG were used. (CACCC was conserved among the three different Gli–DNA binding sequence, and the binding of Gli proteins is abolished by changing into GTGGG.) One hundred times or ten times excess of non-radioisotope labeled type A, B, and C oligos were used as competitor oligos. To investigate the possibility that GliH1 may regulate the transcription of Gli1, direct binding of GST–Gli1, GST– Gli3 and GST–GliH1 fusion proteins to a XhoI–EcoRI 800 bp Gli1 promoter and exon fragment that had been shown to bind Gli3, was used as a substrate in the gel mobility shift assay (Dai et al., 1999). 4.4. Whole-mount and section RNA in situ hybridization Whole-mount and section RNA in situ hybridization was carried out as described previously (Platt et al., 1997). For whole-mount and section RNA in situ experiments the mouse GliH1 cDNA (2.3 kb) linearized with XhoI was used as a probe. Northern blot analysis of GliH1 expression in mouse embryonic tissues at 16.5 dpc was done with a mouse 900 bp PstI probe. 4.5. Transient co-transfection assay A mouse GliH1 cDNA–5 0 -FLAG fusion was generated by PCR and was placed downstream of the Sr-alpha promoter (Takebe et al., 1988) which is composed of the Simian virus 40 (SV40) early promoter and the R segment and part of the U5 sequence (R-U5 0 ) of the long terminal repeat of human T-cell leukemia virus type 1. A human Gli3–FLAG fusion was made by addition of the FLAG-Tag to the SRalpha-Gli3 plasmid at the 5 0 end of the Gli3 cDNA (Dai et al., 1999). A luciferase reporter plasmid (pGL3 LUC 6GBS) containing six tandem repeats of a type B Gli DNA binding site (5 0 -GCGTGGACCACCCAAGACGAAATT-3 0 ) was generated by modifying the pA10CAT6GBS reporter plasmid (Dai et al., 1999) by cutting the pA10CAT6GBS plasmid with Bgl II and inserting the fragment containing 6GBS into the Bgl II site of the pGL3-promoter vector (Promega). To make the mutant reporter plasmid (Mut pGL3 LUC 6GBS), six tandem repeats of a synthetic oligonucleotide, (5 0 -GCGTGGACGTGGGAAGACGAAATT-3 0 ) were inserted into the pGL3-promoter vector. A mixture containing 100 ng of pSRalpha-5 0 FLAG-Gli1 and pSRalpha-5 0 FLAG-Gli3 and/or pSRalpha-5 0 FLAG-GliH1,

M. Nakashima et al. / Mechanisms of Development 119 (2002) 21–34

25 ng of the internal control plasmid pRL-SV (Promega) in which the SV40 early promoter was linked to the sea pansy luciferase gene, and 500 ng of the pGL3 LUC 6GBS were transfected into bovine dental pulp cells using lipofectamine (Invitrogen). The pulp cells were isolated as previously described (Nakashima, 1991). The dual luciferase assays were done according to the manufacturer’s instructions (Promega). The total amount of plasmid DNA was adjusted to 1 mg by adding control plasmid DNA, pSRalpha0 (Takebe et al., 1988). The experiments were repeated three times.

Acknowledgements We are grateful to Dr. H.L. Park and Ms. A. Beeghly for technical assistance, and Cindy Loomis for helpful advice. This work was supported by a Monbusho International Scientific Research Program Joint Research Grant, Japan (#09044320) and a Grant-in-Aid for Scientific Research from the Ministry of Education, Science, Sports and Culture, Japan (#10470407) to M.N., and an NICHD grant to A.L.J., who is an investigator of the Howard Hughes Medical Institute.

References Aruga, J., Yokota, N., Hashimoto, M., Furuichi, T., Fukuda, M., Mikoshiba, K., 1994. A novel zinc finger protein, zic, is involved in neurogenesis, especially in the cell lineage of cerebellar granule cells. J. Neurochem. 63, 1880–1890. Aruga, J., Nagai, T., Tokuyama, T., Hayashizaki, Y., Okazaki, Y., Chapman, V.M., Mikoshiba, K., 1996. The mouse zic gene family. Homologues of the Drosophila pair-rule gene odd-paired. J. Biol. Chem. 271, 1043–1047. Aruga, J., Mizugishi, K., Koseki, H., Imai, K., Balling, R., Noda, T., Mikoshiba, K., 1999. Zic1 regulates the patterning of vertebral arches in cooperation with Gli3. Mech. Dev. 89, 141–150. Auerbach, W., Dunmore, J.H., Fairchild-Huntress, V., Fang, Q., Auerbach, A.B., Huszar, D., Joyner, A.L., 2000. Establishment and chimera analysis of 129/SvEv- and C57BL/6-derived mouse embryonic stem cell lines. Biotechniques 29, 1024–1032. Ausubel, F.H., Brent, R., Kingston, R.E., Moore, D.E., Seidman, J.G., Smith, J.A., Struhl, K., 1996. Affinity purification of proteins binding to GST fusion proteins (Section 20.2). Current Protocols in Molecular Biology, Wiley, New York. Aza-Blanc, P., RamI´rez-Weber, F.-A., Laget, M.-P., Schwartz, C., Kornberg, T.B., 1997. Proteolysis that is inhibited by hedgehog targets cubitus interruptus protein to the nucleus and converts it to a repressor. Cell 89, 1043–1053. Aza-Blanc, P., Lin, H.Y., Ruiz i Altaba, A., Kornberg, T.B., 2000. Expression of the vertebrate Gli proteins in Drosophila reveals a distribution of activator and repressor activities. Development 127, 4293–4301. 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, in press. Bai, C.B., Joyner, A.L., 2001. Gli1 can rescue the in vivo function of Gli2. Development 128, 5161–5172. Bitgood, M.J., McMahon, A.P., 1995. Hedgehog and Bmp genes are coexpressed at many diverse sites of cell-cell interaction in the mouse embryo. Dev. Biol. 172, 126–138.

33

Chiang, C., Litingtung, Y., Lee, E., Young, K.E., Corden, J.L., Westphal, H., Beachy, P.A., 1996. Cyclopia and defective axial patterning in mice lacking Sonic hedgehog gene function. Nature 383, 407–413. Chiang, C., Litingtung, Y., Harris, M.P., Simandl, B.K., Li, Y., Beachy, P.A., Fallon, J.F., 2001. Manifestation of the limb prepattern: limb development in the absence of sonic hedgehog function. Dev. Biol. 236, 421–435. Dai, P., Akimaru, H., Tanaka, Y., Maekawa, T., Nakafuku, M., Ishii, S., 1999. Sonic Hedgehog-induced activation of the Gli1 promoter is mediated by GLI3. J. Biol. Chem. 274, 8143–8152. Dassule, H.R., Lewis, P., Bei, M., Maas, R., McMahon, A.P., 2000. Sonic hedgehog regulates growth and morphogenesis of the tooth. Development 127, 4775–4785. Hardcastle, Z., Mo, R., Hui, C.C., Sharpe, P.T., 1998. The Shh signalling pathway in tooth development: defects in Gli2 and Gli3 mutants. Development 125, 2803–2811. Hui, C.C., 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. Hui, C.C., Slusarski, D., Platt, K.A., Holmgren, R., Joyner, A.L., 1994. Expression of three mouse homologs of the Drosophila segment polarity gene cubitus interruptus, Gli, Gli-2, and Gli-3, in ectoderm- and mesoderm-derived tissues suggests multiple roles during postimplantation development. Dev. Biol. 162, 402–413. Hynes, M., Stone, D.M., Dowd, M., Pitts-Meek, S., Goddard, A., Gurney, A., Rosenthal, A., 1997. Control of cell pattern in the neural tube by the zinc finger transcription factor and oncogene Gli-1. Neuron 19, 15–26. Ingham, P.W., McMahon, A.P., 2001. Hedgehog signaling in animal development: paradigms and principles. Genes Dev. 15, 3059–3087. Iseki, S., Araga, A., Ohuchi, H., Nohno, T., Yoshioka, H., Hayashi, F., Noji, S., 1996. Sonic hedgehog is expressed in epithelial cells during development of whisker, hair, and tooth. Biochem. Biophys. Res. Commun. 218, 688–693. Kinzler, K.W., Bigner, S.H., Bigner, D.D., Trent, J.M., Law, M.L., O’Brien, S.J., Wong, A.J., Vogelstein, B., 1987. Identification of an amplified, highly expressed gene in a human glioma. Science 236, 70–73. Kinzler, K.W., Vogelstein, B., 1990. The GLI gene encodes a nuclear protein which binds specific sequences in the human genome. Mol. Cell. Biol. 10, 634–642. Koyabu, Y., Nakata, K., Mizugishi, K., Aruga, J., Mikoshiba, K., 2001. Physical and functional interactions between Zic and Gli proteins. J. Biol. Chem. 276, 6889–6892. Kraus, P., Fraidenraich, D., Loomis, C.A., 2001. Some distal limb structures develop in mice lacking Sonic hedgehog signaling. Mech. Dev. 100, 45–58. Lamar, E., Kintner, C., Goulding, M., 2001. Identification of NKL, a novel Gli-Kruppel zinc-finger protein that promotes neuronal differentiation. Development 128, 1335–1346. Lee, J., Platt, K.A., Censullo, P., Ruiz i Altaba, A., 1997. Gli1 is a target of Sonic hedgehog that induces ventral neural tube development. Development 124, 2537–2552. Litingtung, Y., Chiang, C., 2000. Specification of ventral neuron types is mediated by an antagonistic interaction between Shh and Gli3. Nat. Neurosci. 3, 979–985. Logan, C., Khoo, W.K., Cado, D., Joyner, A.L., 1993. Two enhancer regions in the mouse En-2 locus direct expression to the mid/hindbrain region and mandibular myoblasts. Development 117, 905–916. Matise, M.P., Joyner, A.L., 1999. Gli genes in development and cancer. Oncogene 18, 7852–7859. Matise, M., Auberbach, W., Joyner, A.L., 2000. Production of targeted embryonic stem cell clones. In: Joyner, A.L. (Ed.). Gene Targeting: A Practical Approach, 2nd Edition, Chapter 3. New York, Oxford University Press, pp. 101–131. Methot, N., Basler, K., 1999. Hedgehog controls limb development by regulating the activities of distinct transcriptional activator and repressor forms of Cubitus interruptus. Cell 96, 819–831.

34

M. Nakashima et al. / Mechanisms of Development 119 (2002) 21–34

Mizugishi, K., Aruga, J., Nakata, K., Mikoshiba, K., 2001. Molecular properties of Zic proteins as transcriptional regulators and their relationship to GLI proteins. J. Biol. Chem. 276, 2180–2188. Mo, R., Freer, A.M., Zinyk, D.L., Crackower, M.A., Michaud, J., Heng, H.H., Chik, K.W., Shi, X.M., Tsui, L.C., Cheng, S.H., Joyner, A.L., Hui, C., 1997. Specific and redundant functions of Gli2 and Gli3 zinc finger genes in skeletal patterning and development. Development 124, 113–123. Motoyama, J., Liu, J., Mo, R., Ding, Q., Post, M., Hui, C.C., 1998. Essential function of Gli2 and Gli3 in the formation of lung, trachea and oesophagus. Nat. Genet. 20, 54–57. Nakashima, M., 1991. Establishment of primary cultures of pulp cells from bovine permanent incisors. Arch. Oral Biol. 36, 655–663. Ohlmeyer, J.T., Kalderon, D., 1998. Hedgehog stimulates maturation of Cubitus interruptus into a liable transcriptional activator. Nature 396, 749–753. Papaioannou, V., Johnson, R., 2000. Production of chimeras by blastocyst and morula injection of targeted ES cells. In: Joyner, A.L. (Ed.). Gene Targeting: A Practical Approach, 2nd Edition, Chapter 4. Oxford University Press, New York, pp. 133–175. Park, H.L., Bai, C., Platt, K.A., Matise, M.P., Beeghly, A., Hui, C.C., Nakashima, M., Joyner, A.L., 2000. Mouse Gli1 mutants are viable but have defects in SHH signaling in combination with a Gli2 mutation. Development 127, 1593–1605. Pavletich, N.P., Pabo, C.O., 1993. Crystal structure of a five-finger GLI– DNA complex: new perspectives on zinc fingers. Science 261, 1701– 1707. Pierani, A., Brenner-Morton, S., Chiang, C., Jessell, T.M., 1999. A sonic hedgehog-independent, retinoid-activated pathway of neurogenesis in the ventral spinal cord. Cell 97, 903–915. Platt, K.A., Michaud, J., Joyner, A.L., 1997. Expression of the mouse Gli and Ptc genes is adjacent to embryonic sources of hedgehog signals suggesting a conservation of pathways between flies and mice. Mech. Dev. 62, 121–135. Ruiz i Altaba, A., 1998. Combinatorial Gli gene function in floor plate and neuronal inductions by Sonic hedgehog. Development 125, 2203–2212. Robbins, D.J., Nybakken, K.E., Kobayashi, R., Sisson, J.C., Bishop, J.M., The´ rond, P.P., 1997. Hedgehog elicits signal transduction by means of a

large complex containing the kinesin-related protein Costal2. Cell 90, 225–234. Ruppert, J.M., Kinzler, K.W., Wong, A.J., Bigner, S.H., Kao, F.T., Law, M.L., Seuanez, H.N., O’Brien, S.J., Vogelstein, B., 1988. The GLIKruppel family of human genes. Mol. Cell. Biol. 8, 3104–3113. Ruppert, J.M., Vogelstein, B., Arheden, K., Kinzler, K.W., 1990. GLI3 encodes a 190-kilodalton protein with multiple regions of GLI similarity. Mol. Cell. Biol. 10, 5408–5415. Sasaki, H., Nishizaki, Y., Hui, C., Nakafuku, M., Kondoh, H., 1999. Regulation of Gli2 and Gli3 activities by an amino-terminal repression domain: implication of Gli2 and Gli3 as primary mediators of Shh signaling. Development 126, 3915–3924. Stone, D., Rosenthal, A., 2000. Achieving neuronal patterning by repression. Nat. Neurosci. 3, 967–969. Takebe, Y., Seiki, M., Fujisawa, J., Hoy, P., Yokota, K., Arai, K., Yoshida, M., Arai, N., 1988. SR alpha promoter: an efficient and versatile mammalian cDNA expression system composed of the simian virus 40 early promoter and the R-U5 segment of human T-cell leukemia virus type 1 long terminal repeat. Mol. Cell. Biol. 8, 466–472. Toyono, T., Nakashima, M., Kuhara, S., Akamine, A., 1997. Expression of TGF-b superfamily receptors in dental pulp. J. Dent. Res. 76, 1555– 1560. Walterhouse, D., Ahmed, M., Slusarski, D., Kalamaras, J., Boucher, D., Holmgren, R., Iannaccone, P., 1993. Gli, a zinc finger transcription factor and oncogene, is expressed during normal mouse development. Dev. Dyn. 196, 91–102. 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. Williams, T., Tjian, R., 1991. Analysis of the DNA-binding and activation properties of the human transcription factor AP-2. Genes Dev. 5, 670– 682. Zhang, F., Nakanishi, G., Kurebayashi, S., Yoshino, K., Perantoni, A., Kim, Y.S., Jetten, A.M., 2002. Characterization of Glis2, a novel gene encoding a Gli-related, Kruppel- like transcription factor with transactivation and repressor functions. Roles in kidney development and neurogenesis. J. Biol. Chem. 277, 10139–10149.