SEARCH

SEARCH BY CITATION

Keywords:

  • Gli3;
  • Sonic hedgehog;
  • limb patterning

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Gli3 protein processing to generate the Gli3 repressor is mediated by proteasome and inhibited by Hedgehog signaling. The Gli3 repressor concentration is graded along the anterior–posterior axis of the developing vertebrate limb due to posteriorly restricted Sonic hedgehog expression. In this study, we created a small deletion at the Gli3 locus (Gli3Δ68), which causes a half reduction in the Gli3 repressor levels and a slightly increased activity of full-length mutant protein in the limb. Mice homozygous for Gli3Δ68 develop one to two extra partial digits in the anterior of the limb, while mice carrying one copy of the Gli3Δ68 allele die soon after birth and display seven digits. These phenotypes are more severe than those found in mice lacking one wild-type Gli3 allele. The expression of dHand, Hoxd12, and Hoxd13 is anteriorly expanded in the limb, even though no up-regulation of Gli1 and Ptc RNA expression is detected. These findings suggest that a decrease in the Gli3 repressor level in combination with an increase in Gli3 full-length activity results in more severe digit patterning abnormalities than those caused by a loss of one wild-type Gli3 allele. Developmental Dynamics 236:769–776, 2007. © 2007 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

The anterior–posterior patterning of the vertebrate limb is controlled by the secreted Hedgehog signaling molecule, Sonic hedgehog (Shh). In the developing vertebrate limb bud, Shh RNA is exclusively expressed in the posterior limb margin, the zone of polarizing activity (ZPA), but the Shh protein is thought to be transported away from the ZPA and, thus, be graded along the anterior–posterior axis of the limb (Riddle et al.,1993). It is this Shh protein gradient that is proposed to specify limb digit number and identity. In agreement with this view, Shh mutant mouse embryos lack all but the first digits in their hindlimbs (Chiang et al.,1996,2001; Kraus et al.,2001), whereas either the graft of ZPA, or implantation of cells expressing Shh or a Sepharose bead soaked in Shh to the anterior region of chicken limb buds leads to the formation of mirror image duplicated limb digits (Riddle et al.,1993; Lopez-Martinez et al.,1995; Yang et al.,1997).

Shh signaling in the mouse is primarily mediated by the Gli2 and Gli3 transcriptional factors, which subsequently activate the transcription of another Gli gene, Gli1 (Bai et al.,2004). Genetic analyses of Gli1 and Gli2 mutants have demonstrated that the function of these two genes in limb patterning is minimal (Park et al.,2000; Bai et al.,2002), whereas Gli3 plays an essential role in anterior–posterior limb patterning. Several mutations in the human Gli3 gene have been found in association with various developmental defects of the limb (Vortkamp et al.,1991; Biesecker,1997). Similarly, mice homozygous for a Gli3 null mutant allele, Extratoe or Gli3Xt, are characterized by polysyndactylous and abnormal craniofacial phenotypes (Hui and Joyner,1993). Interestingly, mice lacking both Gli3 and Shh gene function exhibit limb phenotypes indistinguishable from those of mice lacking Gli3 gene function alone (Litingtung et al.,2002; te Welscher et al.,2002b). Therefore, the role of Gli3 is to restrict the pattern formation of polydactylous digits, while Shh promotes the patterning of multiple digits, and the precisely balanced counteraction between the two establishes the normal patterning of vertebrate limb digit number and identity.

The molecular explanation underlying this counteraction in limb patterning has begun to emerge. The Gli3 protein exists in two forms, as a full-length transcription activator, Gli3-190, and a C-terminally truncated repressor, Gli3-83. The repressor is derived from the Gli3-190 precursor through proteolytic processing, which is initiated by hyperphosphorylation of its C-terminal region, but is suppressed by Shh signaling (Wang et al.,2000). Thus, in the developing vertebrate limbs, Gli3-83 repressor levels are graded from the highest in the anterior to the lowest in the posterior, while Gli3-190 levels show only a slight change along the anterior–posterior axis (Wang et al.,2000). Genetic analyses of several mutant genes in both the chicken and the mouse have demonstrated that the Gli3 repressor gradient is crucial for the normal patterning of the vertebrate limb. For example, in Shh−/− mouse embryos, the Gli3-83 repressor gradient is almost completely disrupted (Litingtung et al.,2002). The polydactylous phenotypes observed in IFTs, Dnchc2, and Rab23 mutant mouse embryos are also associated with a partial perturbation of the Gli3-83 gradient (Haycraft et al.,2005; Huangfu and Anderson,2005; Liu et al.,2005; May et al.,2005; Eggenschwiler et al.,2006). In talpid2 (ta2) and talpid3 (ta3) mutant chicken limbs, the Gli3-83 repressor gradient is also disrupted (Wang et al.,2000; Davey et al.,2006). However, it should be noted that, in all these gene mutants, not only are the levels of the Gli3 repressor altered, but the Gli3 activator function is also impaired as indicated by a loss of or a significantly reduced Ptc and Gli1 expression. In this study, we engineered a Gli3 hypermorphic allele that expressed a reduced level of the Gli3 repressor gradient, yet a slightly increased Gli3 activator function. Analysis of the limb phenotypes of the mutant mice revealed that reducing the Gli3 repressor gradient while increasing Gli3 activator function leads to a more severe polydactylous phenotype than that caused simply by the loss of one wild-type allele, thus implying that the balance between Gli3 activator and repressor forms is important for the patterning of limb digits.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Generating the Mouse Gli3Δ68 Hypermorphic Allele

While we were performing structural and functional analysis of the Gli3 protein in cell culture, the behavior of one of Gli3 expression constructs drew our attention. This construct expressed a Gli3 protein with a deletion from residue 675 to 743, which included the proteolytic processing site; the protein was designated Gli3Δ68. Surprisingly, this protein was still processed, albeit at a reduced efficiency (Fig. 1A). Gli3Δ68, when overexpressed, exhibited approximately two times the activity of wild-type Gli3 in the absence of protein kinase A (PKA) stimulation, whereas its activity was inhibited to a level only slightly higher than that of wild-type Gli3, when a constitutively active PKA, PKA*, was coexpressed (Fig. 1B). These data suggest that the processed Gli3 repressor is so potent that the difference in its levels between overexpressed wild-type and mutant Gli3 in the presence of PKA* is masked in the reporter assay. Like wild-type Gli3, Gli3Δ68 was also predominantly localized in the cytoplasm (Fig. 1C), indicating that the higher transcriptional activity of the mutant protein is not due to a change in subcellular localization. Thus, the 68-residue deletion leads to a reduced processing and an increased transcriptional activity of the full-length mutant protein.

thumbnail image

Figure 1. Characterization of the Gli3Δ68 protein. A: An immunoblot shows that the Gli3Δ68 protein is processed in transfected HEK293 cells at a lower efficiency. Full-length and processed forms of wild-type and mutant Gli3 proteins are indicated by arrows. Arrowheads point to protein fragments that were nonspecifically degraded due to a high level of expression in this particular experiment. B: A Gli-dependent reporter assay showing that Gli3Δ68 is more active than Gli3. NIH3T3 cells were transfected with a Gli-dependent luciferase reporter, TK-renillar (a transfection control), and the indicated expression constructs. The firefly luciferase activities were derived from three independent experiments. Shown below are schematic diagrams of full-length wild-type and mutant Gli3 proteins. C: An immunostaining of NIH3T3 cells overexpressing either the Gli3 or Gli3Δ68 proteins. Both proteins are localized to the cytoplasm.

Download figure to PowerPoint

Based on these observations in cultured cells, we reasoned that a Gli3 mutant allele that expresses the Gli3Δ68 protein would be hypermorphic. We thus created a mouse Gli3 mutant allele, Gli3Δ68, in which nucleotide sequences encoding the 68 residues were deleted using a targeted gene knock-in approach (Fig. 2A,B). As predicted, this mutant allele expressed a full-length mutant protein that was slightly smaller than Gli3-190, its wild-type counterpart, and its processing efficiency was clearly reduced, albeit probably not as much as in cultured cells. Interestingly, the size of the processed form of Gli3Δ68 was indistinguishable from that of the naturally processed form, Gli3-83 (Fig. 2C), indicating that the processing site shifted slightly toward the C-terminal end. Since the deletion included the Gli3 processing site, the fact that the Gli3Δ68 full-length mutant protein could be still processed supports the hypothesis that Gli3 processing is mediated by proteasome (Tempe et al.,2006; Wang and Li,2006).

thumbnail image

Figure 2. The gene targeting strategy of creating Gli3Δ68 allele and embryonic stem (ES) cell screening. A: (Top) The Gli3 genomic locus and targeting construct (top). Open boxes indicate Gli3 exons. Asterisks refer to the six PKA sites. Parts of two small exons immediately preceding the last large exon are fused to generate a deletion from residue 675 to 743. The neomycin (neo) gene and diphtheria toxin A (DTA) are used as positive and negative selection markers, respectively. (Bottom) Schematic diagrams show the expected full-length and processed forms of Gli3Δ68 protein. B, BamHI; E, EcoRI; N, NdeI, and S, SacI. B: ES cell Southern blot hybridization using the 5′- and 3′-probes indicated in A following EcoRI and SacI digestion, respectively. The expected sizes of the DNA fragments for the mutant and wild-type Gli3 allele are indicated. C: An immunoblot shows that the Gli3Δ68 protein is still processed, but at a reduced efficiency. Protein lysates were prepared from embryonic day (E) 10.5 mouse embryos and immunoblotted with a Gli3 antibody.

Download figure to PowerPoint

Limb Phenotypes of Gli3Δ68 Mutant Mice

Mice heterozygous for Gli3Δ68 did not show any abnormal phenotypes (data not shown). Gli3Δ68 homozygous mice were viable and fertile, but exhibited one to two additional first digits, either complete or partial, in their forelimbs, sometimes accompanied by a nub on their second digits. The first digits in the hindlimbs were completely duplicated, but the second digits were normal (Fig. 3C,D). This phenotype was consistent with our observations that Gli3Δ68 exhibited a higher activity than wild-type Gli3 and a reduced processing (Figs. 1B, 2C). The Gli3 null allele, Gli3Xt, is semidominant since Gli3+/Xt mice develop an extra nub on the anterior of their hindlimbs; otherwise, they are normal (Hui and Joyner,1993). To further determine the gene dose effect of the Gli3Δ68 mutant on limb patterning, Gli3Δ68/Δ68mice were crossed with Gli3+/Xt heterozygous mice to obtain animals carrying only one copy of the Gli3Δ68 allele. These mouse embryos died soon after birth and developed seven digits in both their forelimbs and hindlimbs. The two extra digits were either partial or complete and resembled the biphalangeal first digit (Fig. 3E,F). In contrast, Gli3Xt/+ heterozygous mice, which are comparable with those of Gli3Δ68/+ in terms of gene copy number, develop only one extra phalange on the first digit (Hui and Joyner,1993; data not shown). Thus, our data indicate that a further reduction of Gli3Δ68 gene dose significantly promotes the development of additional first digits, and if the Gli3Δ68 protein in vivo behaves like it does in cell culture, our findings also support the notion that the Gli3 repressor restrains the formation of multiple first digits, while the Gli3 full-length activator promotes it.

thumbnail image

Figure 3. Skeletal stains of the limbs of Gli3Xt and Gli3Δ68 mutants. A–F: Skeletons were prepared from neonates. G,H: Skeletons were prepared from E16.5 embryos. The anterior is at the top and posterior at the bottom. Genotypes are indicated to the left.

Download figure to PowerPoint

The Gli3 Repressor Gradient Is Maintained but Reduced in the Gli3Δ68 Mutant Limb

The formation of the Gli3 repressor gradient in the wild-type limb is controlled by posteriorly expressed Shh (Wang et al.,2000; Litingtung et al.,2002). If Gli3Δ68 processing is still regulated by Shh signaling, one would predict that the levels of the repressor form of Gli3Δ68 protein would be graded from the anterior to the posterior of the mutant limb bud. Immunoblotting analysis of proteins from anterior and posterior mutant limb fragments revealed that this was indeed the case (Fig. 4A). Quantification analysis of the intensity of the Gli3 repressor bands showed that the Gli3 repressor gradient in the Gli3Δ68/Δ68mutant was still maintained (Fig. 4B, compare open bars), even though the levels of the repressor were reduced by about half. Consequently, the ratio of the processed form to the full-length form of Gli3Δ68 in the Gli3Δ68/Δ68mutant limb buds was less than half of that in the wild-type limb buds (Fig. 4B, black and gray bars). The similar results were also obtained for Gli3Xt/+ and Gli3Xt/Δ68 limb buds (Fig. 4A, lanes 5–8), indicating that a loss of one Gli3 allele does not affect protein expression from the remaining allele. It should be noted that the ratio of the Gli3 repressor to the full-length Gli3Δ68 in the posterior limb fragment of Gli3Δ68/Xt embryos was not available, because the full-length Gli3Δ68 level was too low to be accurately quantified in this particular experiment (Fig. 4B). Thus, the 68-residue deletion in the Gli3 protein only caused a decrease in the Gli3 repressor gradient level, but did not affect the regulation of Gli3Δ68 processing by Shh signaling.

thumbnail image

Figure 4. The level of the Gli3 repressor is reduced, but its gradient is still maintained in Gli3Δ68 mutant limb. A: An immunoblot shows the levels of the full-length and processed forms of Gli3 in the anterior (A) and posterior (P) limb fragments of wild-type, Gli3Δ68/Δ68, Gli3Xt/+, and Gli3Xt/Δ68 embryos at embryonic day (E) 10.5. Note that the Gli3Xt allele still expresses an N-terminal fragment. The right panel is the duplication of lanes 5–8 in the left panel with long exposure. The lower panel is a tubulin blot showing loading controls. B: A graph shows the ratio of the Gli3 repressor to the full-length in the anterior (black bars) or posterior (gray bars) limb fragments and the ratio of Gli3 repressor in the anterior versus posterior limb fragments (open bars) from A. The levels of full-length and processed forms of Gli3 were quantified using the NIH image software and normalized against those of tubulin. The ratio of the Gli3 repressor to Gli3Δ68 full-length in the posterior Gli3Xt/Δ68 limb fragment could not calculated, since the mutant full-length (lane 8) was barely detectable.

Download figure to PowerPoint

Effect of the Gli3Δ68 Mutation on Marker Gene Expression

To understand the polydactylous phenotypes of the Gli3Δ68 mutant mice at the molecular level, we first analyzed the expression of the Shh transcriptional targets Gli1 and Patched (Ptc); (Goodrich et al.,1996; Marigo et al.,1996). Neither Gli1 nor Ptc was ectopically expressed in both the Gli3Δ68/Δ68and Gli3Δ68/+ mouse limb buds (Fig. 5A–J).

thumbnail image

Figure 5. Expression of Gli1, Ptc, and Fgf4 RNAs in the limb buds of Gli3Xt and Gli3Δ68 mutant embryos (embryonic day [E] 10.5). Genotypes are shown above the panels. The anterior is at the top and the posterior at the bottom. Note that Gli1, Ptc, and Fgf4 were not ectopically expressed and that the slightly higher Ptc expression in G was not consistently observed.

Download figure to PowerPoint

Proximodistal limb outgrowth requires signals from fibroblast growth factor genes (Fgf), and one of Fgf genes, Fgf4, is expressed in the posterior two thirds of the apical ridge (Martin,1998). Shh maintains this expression by regulating that of Gremlin (Zuniga et al.,1999). In the Gli3Xt/Xt mutant limb buds, Fgf4 expression was expanded to the entire apical ridge, but it was not in Gli3Δ68/Δ68and Gli3Δ68/Xt mutant apical ridges (Fig. 5K–O), suggesting that a reduced Gli3 repressor level is not sufficient to derepress Fgf4 expression.

The dHand basic helix–loop–helix transcription factor is normally expressed in the posterior region of mouse limb buds, and its expression is restricted by the Gli3 repressor (te Welscher et al.,2002a). Since Gli3Δ68 is a hypermorphic allele, we wanted to know whether dHand expression was altered. Indeed, a weak ectopic expression of dHand was observed in the anterior Gli3Δ68/Δ68 limb buds, becoming stronger in Gli3Δ68/Xtlimb buds (Fig. 6K–O). Nevertheless, in either genetic background, dHand ectopic expression was weaker than that in Gli3Xt/Xt limbs (Fig. 6, compare L and M to N), consistent with the idea that dHand expression is inhibited by the Gli3 repressor.

thumbnail image

Figure 6. Ectopic expression of Hoxd12, Hoxd13, and dHand in the limb buds of Gli3Δ68 and Gli3Xt mutant embryos (embryonic day [E] 10.5). The anterior is at the top and the posterior at the bottom.

Download figure to PowerPoint

Hoxd12 and Hoxd13 are essential for autopod patterning (Zakany and Duboule,1999). The expression of Hoxd13 is posteriorly restricted in wild-type mouse limb buds, while Hoxd12 expression covers about the posterior half of the limb buds. Expression of both genes correlates to Shh signaling, since the complete loss of Gli3 gene function results in an anterior expansion of both Hoxd12 and Hoxd13 expression (Fig. 6D,J; Litingtung et al.,2002; te Welscher et al.,2002b). The Gli3Δ68 mutation also appeared to cause this, although the ectopic expression was neither as broad nor intense as that of the Gli3Xt/Xt mutant (Fig. 6, compare B and C to D; H and I to J). In contrast, loss of one wild-type Gli3 allele in the Gli3Xt/+ limb was not sufficient to cause Hoxd12 and Hoxd13 ectopic expression (Fig. 6E,K). Since the Gli3 repressor level in Gli3Xt/+ limb was lower than that in Gli3Δ68/Δ68 mutant limb (Fig. 4, compare lanes 5–6 to lanes 3–4), the anteriorly expanded expression of Hoxd12, Hoxd13, and dHand in the Gli3Δ68/Δ68limb must be mostly due to the higher level of the full-length Gli3Δ68 activator. Taken together, our in situ hybridization results indicate that the Gli3Δ68 mutation activates the ectopic expression of dHand, Hoxd12, and Hoxd13, but it is unable to induce Gli1 and Ptc ectopic expression.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

In this study, we generated a Gli3Δ68 deletion allele in mice. The 68 deleted residues include the processing site of the Gli3 full-length protein, but, surprisingly, the mutant protein is still processed, albeit at a lower efficiency. The size of its processed form is indistinguishable from that of the naturally processed form, Gli3-83, indicating that the processing site has shifted slightly toward the C-terminus. A similar deletion in Ci, the Gli3 fly homolog, also fails to inhibit Ci protein processing (Methot and Basler,1999). These in vivo data strongly support the hypothesis that Gli3/Ci processing is carried out by proteasome. In addition to the reduced processing, the mutant full-length protein exhibits a higher transcriptional activity than the full-length Gli3 protein. This finding raises the possibility that the 68 residues may be involved in a negative regulation of Gli3 full-length activator function. For example, they may be required to recruit a negative cofactor(s), or somehow to intramolecularly inhibit the Gli3 transcriptional activity itself. The deletion of these residues leads to a derepression of the full-length Gli3 protein activity.

Genetic analyses of mutations in several genes have shown that any perturbation of Gli3 processing, which in turn alters the Gli3 repressor gradient, leads to the abnormal patterning of limb digits; these genes include Shh (Litingtung et al.,2002), IFTs (Haycraft et al.,2005; Liu et al.,2005), Dnchc2 (Huangfu and Anderson,2005; May et al.,2005), Rab23 (Eggenschwiler et al.,2006), and ta3 (Davey et al.,2006). However, in the limb buds of all these gene mutants, not only is the repressor gradient lost, but the level of full-length Gli3 protein is also either reduced (in the case of the Shh mutant) or increased (in the rest of the mutants), and yet the protein is not active, as indicated by the loss of Ptc and Gli1 expression (Fig. 7). Therefore, the polydactylous limb phenotypes of these mutants are likely caused by an alteration of the Gli3 repressor gradient as well as a loss of the full-length Gli3 activity.

thumbnail image

Figure 7. A model shows the levels of the Gli3 full-length activator (Gli3A) and Gli3 repressor (Gli3R) along the anterior–posterior axis of limb buds mutant for the indicated genes. In IFTs, Dnchc2, and ta3 mutant limbs, the full-length Gli3 protein is accumulated due to the reduced Gli3 processing, but the protein is not active based on significantly decreased expression of Ptc and Gli1 genes. In Shh mutant limbs, the full-length Gli3 protein is inactive, since Ptc and Gli1 are not expressed, and the Gli3 repressor level increases without a gradient.

Download figure to PowerPoint

In contrast, the Gli3Δ68 mutation only reduces, but does not perturb the Gli3 repressor gradient, and the full-length Gli3Δ68 mutant protein in vivo should be at least as active as, if not more than, the wild-type full-length Gli3 protein, since Gli1 and Ptc expression is normal. Thus, the abnormal limb phenotypes of Gli3Δ68/Δ68 mice could result from either the reduction in the Gli3 repressor alone, or a change in the ratio of Gli3 full-length activator to the repressor. However, compared with Gli3Xt/+ limbs, the Gli3Δ68/Δ68 limbs express a similar level of Gli3 repressor, but twice the amount of the full-length Gli3 activator because of the presence of two mutant alleles (Fig. 4). If the Gli3 repressor levels alone control limb digit patterning, one would predict that the Gli3Δ68/Δ68 limb digit phenotype would be similar to that of Gli3Xt/+ mice. On the contrary, we found that Gli3Δ68/Δ68 mice developed one to two additional first digits as compared to an extra phalange on the first digits of Gli3Xt/+ mice (Fig. 3). Therefore, the abnormal limb phenotype in Gli3Δ68/Δ68 mice should result from the reduction of Gli3 repressor gradient as well as the higher levels of the full-length activator.

Analysis of genetic interaction between Gli3 and Shh has shown that an overall Gli3 function is to restrain the formation of polydactyly, whereas Shh is to promote it (Litingtung et al.,2002; te Welscher et al.,2002b), which is likely through the regulation of Gli3 processing as well as full-length Gli3 protein activity. Reducing the Gli3 gene dose by half in Shh mutant limbs results in development of one or two first digits in forelimbs and hindlimbs, respectively, and loss of both Gli3 and Shh gene functions leads to polydactyly without digit identity, suggesting that a high level of Gli3 repressor is needed to restrain the development of multiple first digits and that anterior limb patterning is most susceptible to a reduction in Gli3 repressor levels. Since Gli3Δ68 mutant expresses a reduced level of Gli3 repressor, it is not surprising that the mutation only affects digit patterning in the anterior, but not in the posterior.

Our analysis of marker gene expression indicates that only dHand, Hoxd12, and Hoxd13 genes, not Ptc and Gli1, are ectopically expressed in Gli3Δ68 mutant limb buds. One likely explanation for this finding is that these genes are activated in response to different levels of a net output of Gli3 activity. Ptc and Gli1 ectopic expression requires a complete loss of the Gli3 repressor; although the full-length Gli3Δ68 protein is more active than that of Gli3, it is not sufficient to overcome the Gli3 repressor activity; whereas dHand, Hoxd12, and Hoxd13 can be ectopically activated as long as the Gli3 repressor level is reduced to a certain degree. In support of this view, Ptc and Gli1 are ectopically expressed in Gli3Xt/Xt mutant limbs, which completely lack the Gli3 gene function (Litingtung et al.,2002; te Welscher et al.,2002b), but the normal expression of Ptc and Gli1 is dramatically reduced in IFTs, Dnchc2, and Ta3 mutant limbs where Gli3 repressor level has been reduced (Haycraft et al.,2005; Huangfu and Anderson,2005; Liu et al.,2005; May et al.,2005; Davey et al.,2006; Eggenschwiler et al.,2006). Unlike Ptc and Gli1 genes, dHand, Hoxd12, and Hoxd13 genes are ectopically expressed in limbs of all these mutants. Thus, undetectability of Ptc and Gli1 ectopic expression does not contradict the polydactylous phenotypes of Gli3Δ68 mutant limbs.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Gli3 Mutant Mice and the Generation of Gli3Δ68 Mutant Knockin Mice

A BAC clone containing mouse Gli3 genomic DNA sequences was purchased from Genome Systems, Inc., and used to create a Gli3Δ68 targeting construct in the pGKneoloxP2DTA.2 vector (Soriano,1997). A 68 amino acid deletion (residues 675–743) was generated in the targeting construct by fusing partial sequence from two related exons. The Gli3Δ68 construct was then introduced into R1 embryonic stem (ES) cells (Nagy et al.,1993) by electroporation, and neomycin-resistant clones were selected by incubating cells in ES cell growth medium containing G418 (200 μg/ml). Targeted ES cell clones were identified by restriction enzyme digestion, followed by a Southern blot analysis of ES cell DNA using a 5′- and a 3′- probe. The 5′-probe identified a 9.5-kb fragment in the targeted allele and a 7.4-kb fragment in the wild-type allele following SacI digestion. The 3′-probe identified a 7.3-kb fragment in the targeted allele and a 6.1-kb fragment in the wild-type allele following EcoRI digestion. Six targeted clones were obtained, three of which were injected into C57BL/6 blastocysts to generate chimeras. The chimeras were then bred with C57BL/6 to establish F1 heterozygotes, creating three independent germline transmitting mouse lines with indistinguishable phenotypes. The floxed neo gene was removed by breeding them with Act-Cre transgenic mice (Lewandoski et al.,1997). Polymerase chain reaction (PCR) analysis was used for routine genotyping with the following primers: BW191, 5′-GTC CTG ATA TCA TCT TCT TGT AAG-3′, and BW192, 5′-TCA GTG TCC ATA GAA TCT AGC TT-3′. The predicted size of the PCR fragments was 180 bp for the wild-type allele and 210 bp for the mutant.

Gli3Xt mutant mice were obtained from Kathryn Anderson at the Sloan Kettering Institute and genotyped according to the sequence information (Maynard et al.,2002) using te following primers: 5′-AAT GAT GCT CAC TAG TAC AGT G-3′ and 5′-AAA CCC GTG GCT CAG GAC AAG C-3′, which produced a 400-bp fragment. The wild-type allele was genotyped with the following primers: 5′-GGC CCA AAC ATC TAC CAA CAC ATA G-3′ and 5′-GTT GGC TGC TGC ATG AAG ACT GAC-3′, which yielded a 190-bp fragment.

Immunofluorescence, Immunoblotting, Reporter Assays, Whole-Mount In Situ Hybridization, and Skeleton Preparation

Immunofluorescence, immunoblotting, and reporter assays were performed as described (Wang et al.,1997,2000). Whole-mount in situ hybridization was carried out according to a published protocol (Wilkinson,1992), and the skeletons were prepared as described (Nagy et al.,2003).

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

We thank Philip Soriano for pGKneolox2DTA vector, Andras Nagy for R1 ES cells, Licia Selleri, and Rolf Zeller for in situ hybridization probes and protocols. The use of animals in this study followed the guidelines of the National Institutes of Health and was approved by Cornell's Institutional Animal Care and Use Committee. This study was supported by a Basil O'Connor Starter Scholar Research Award from the March of Dimes Foundation and a NIH grant (R01 CA111673) to B. W.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES