Research ArticleDevelopmental Biology

The Kinesin Protein Kif7 Is a Critical Regulator of Gli Transcription Factors in Mammalian Hedgehog Signaling

See allHide authors and affiliations

Science Signaling  23 Jun 2009:
Vol. 2, Issue 76, pp. ra29
DOI: 10.1126/scisignal.2000405


From insects to humans, the Hedgehog (Hh) signaling pathway has conserved roles in embryonic development and tissue homeostasis. However, it has been suggested that the lack of mammalian equivalents of Costal2 (Cos2) contributes to a divergence between the mechanism of Drosophila and mammalian Hh signal transduction. Here, we challenge this view by showing that the kinesin protein Kif7 is a critical regulator of Hh signaling in mice. Similar to Cos2, Kif7 physically interacted with Gli transcription factors and controlled their proteolysis and stability, and acted both positively and negatively in Hh signaling. Thus, Kif7 is a missing component of the mammalian Hh signaling machinery, implying a greater commonality between the Drosophila and mammalian system than the prevailing view suggests.


During embryonic development, Hedgehog (Hh) signaling is critical for tissue patterning and cell fate determination (1, 2). The secreted glycoprotein Hh alters the balance between activator and repressor forms of the Gli family of zinc finger transcription factors to generate a gradient of signaling activity essential to many biological processes. In Drosophila, the kinesin-like molecule Costal2 (Cos2) (3, 4) functions both positively and negatively in the Hh pathway by acting as a signaling hub to transduce Hh signals from two transmembrane receptors, Patched (Ptc) and Smoothened (Smo), to the Gli transcription factor Cubitus interruptus (Ci) (58). In the absence of Hh, Ptc inhibits the activity of Smo by promoting its endocytosis and degradation, and Cos2 acts as a scaffold to recruit multiple protein kinases to phosphorylate Ci, triggering its subsequent proteolytic processing into a truncated repressor form (9). Hh signals promote cell surface accumulation and phosphorylation of Smo, which is thought to expose its C-terminal tail, allowing it to interact with Cos2 (10). As a result, Cos2 no longer functions as a processing hub, resulting in the loss of Ci repressor formation (710). In addition, Cos2 also acts positively to promote the expression of Hh target genes that are induced by strong Shh signals (5, 6). In zebrafish, kif7 has been identified as a cos2 homolog and appears to act primarily as a repressor of Hh signaling in myocyte differentiation (12). In contrast, cos2 function is reported to be absent in mammalian Hh signaling (13). Two amniote (a term that refers to tetrapod vertebrates with terrestrially adapted eggs, such as birds, mammals, and reptiles) homologs of Cos2, Kif7 and Kif27, have been identified by sequence comparison. However, in vitro studies have not shown a role for these or other kinesins in Gli-dependent transcription and Hh signal transduction (13). Instead, genetic analysis has shown that kinesin motor protein Kif3a, together with other factors essential for primary ciliogenesis, is required for mammalian Hh signal transduction (1415). These observations have led to the suggestion that primary cilia substitute for the function of Drosophila Cos2 in mammalian cells, providing a rationalization for the apparent absence of Cos2 function in mammals (3, 10).


Kif7 interacts with Gli proteins and inhibits Gli-dependent transcription

We reevaluated the role of putative mammalian Cos2 homologs first by biochemical analysis. Contrary to previous findings (13), we observed that Kif7 and Gli transcription factors interacted in mammalian cells, as assessed by coimmunoprecipitation (Fig. 1, A to D, and fig. S1). Kif7 and Kif7N-Kif27C (which contains the amino acid residues 1 to 791 of Kif7 and the amino acid residues 774 to 1394 of Kif27), but not Kif27 or Kif7C, formed complexes with Gli2, suggesting that the N-terminal half of Kif7 is required for Gli2 interaction (Fig. 1A). However, both Kif7 and Kif27 interacted with Gli3 (Fig. 1B); thus, additional experiments will be needed to map the domains involved in these interactions. These results are consistent with the observation that the N-terminal half of Cos2 contains multiple regions required for interaction with Ci (8). To determine whether Kif7 forms complexes with Gli2 and Gli3 under physiological conditions, coimmunoprecipitation experiments were performed on embryo lysates. Kif7 was present in Gli2 and Gli3 immunoprecipitates from wild-type lysates, but not in immunoprecipitates from Kif7-null lysates (Fig. 1, C and D, and fig. S2). Together, these results suggest that Kif7 and Kif27 may function differently to regulate Gli proteins. To investigate their activities in Hh signaling, we examined the effects of overexpression on Gli-dependent transcription. Consistent with a negative role for Kif7 in Hh signaling, coexpression of Kif7, but not Kif27, suppressed Gli2-dependent transcriptional activation to an extent comparable to that of Su(fu), a known negative regulator of the Hh pathway (11, 1618) (Fig. 1E). These results suggest that Kif7 functions in a manner biochemically similar to that of Cos2 as a negative regulator of Ci and Gli.

Fig. 1

Kif7 interacts with Gli proteins and inhibits Gli-dependent transcription. (A) Kif7, but not Kif27, formed complexes with Gli2. The N-terminal half of Kif7 was required for the interaction. The arrow (top blot) indicates the full-length form of Gli2. White arrowheads (bottom blot) indicate the full-length form of the corresponding proteins. IP, immunoprecipitation; IB, immunoblotting. (B) Kif7 and Kif27 both interacted with Gli3, which bound to the N-terminal but not the C-terminal domain of Kif7 and the C-terminal domain of Kif27. Arrow (top blot) indicates the full-length form of Gli3. White arrowheads (bottom blot) indicate the full-length form of the corresponding proteins. (C) Endogenous Kif7-Gli2 complexes were detected by coimmunoprecipitation in E10.5 wild-type (WT) embryo lysates, but not in lysates from mutant embryos carrying the targeted null mutation in Kif7 (top blot). Endogenous Gli2 was increased in Kif7 mutant embryo lysates (bottom blot). Arrows indicate the full-length form of Kif7 (top blot) and the full-length form of Gli2 (bottom blot). (D) Endogenous Kif7-Gli3 complexes were detected by coimmunoprecipitation in E10.5 wild-type embryo lysates, but not in lysates from mutant embryos carrying the targeted null mutation in Kif7 (top blot). Endogenous Gli3 was increased in Kif7 mutant embryos (bottom blot). Arrows indicate the full-length form of Kif7 (top blot) and the full-length form of Gli3 (bottom blot). Asterisks indicate nonspecific bands. (E) Gli luciferase reporter assay showed that Kif7 repressed Gli2-dependent transcription to a similar extent as Su(fu). In contrast, Kif27 had a less pronounced inhibitory effect. The error bars represent standard deviation of the mean (n = 3 transfections). The corresponding relative abundance of each overexpressed protein is shown by Western blot with actin as a loading control. Note that Gli2 is stabilized in the cytoplasm of transfected cells by Su(fu) overexpression (16). (F) Kif7 overexpression in chick neural tubes inhibited both endogenous and SmoM2-induced Nkx2.2 induction. Arrowheads indicate Nkx2.2 immunostaining.

To investigate whether Kif7 regulates the Hh pathway activity in vivo, we examined the effects of overexpression in the chick neural tube by in ovo electroporation. Graded Shh signaling is pivotal for specifying distinct neural progenitors in the ventral neural tube through the regulation of the three Gli transcription factors (1920). Consistent with an inhibitory function, Kif7 suppressed Shh signaling, as indicated by its ability to block the induction of the ventral progenitor marker Nkx2.2 in the electroporated neural tube in ovo (Fig. 1F). Expression of a constitutively activated form of Smo, SmoM2 (21), resulted in a dorsal expansion of the Nkx2.2+ domain (Fig. 1F) and robust activation of a Gli luciferase reporter (fig. S3). Coexpression of Kif7 efficiently blocked the ability of SmoM2 to induce Nkx2.2 (Fig. 1F) and promote Gli transcriptional activity (fig. S3). These results suggest that Kif7 is a potent inhibitor of the Hh pathway that functions downstream of Smo.

Mice lacking Kif7 display a Gli3-like skeletal phenotype and show altered Gli stability and processing

To test the requirement for Kif7 in mammalian Hh signaling, we generated a mutant allele in mice by gene targeting (fig. S2). In homozygous mutant mouse embryos, Kif7 protein could not be detected by Western blotting (fig. S2), indicating that the targeted mutation generated a null allele. Although Kif7+/− mice are viable and do not exhibit any obvious defects, Kif7−/− mice die at birth with severe malformations, including exencephaly and polydactyly (Fig. 2A), a phenotype reminiscent of that of mice lacking Gli3, which encodes the major Gli transcriptional repressor of mammalian Hh signaling (22). Because Shh-Gli3 signaling controls development of the skeleton, particularly digit patterning of the limb (2325), we examined skeletal patterning in Kif7 mutants by bone staining. Kif7−/− mice exhibited preaxial polydactyly (Fig. 2B) and sternal defects (Fig. 2C), similar to those observed in Gli3−/− mice (22). These findings suggest that Kif7 is a key inhibitor of Hh signaling in multiple tissues and that Kif7 may regulate the Hh pathway in part by controlling Gli3 function or abundance.

Fig. 2

Kif7-null mice exhibit exencephaly and polydactyly and show altered Gli2 stability and Gli3 processing. (A) E14.5 whole-mount embryos. Kif7−/− and Gli3−/− mice showed edema, exencephaly, and polydactyly. (B) Both Kif7−/− and Gli3−/− mice displayed severe polydactyly (up to eight digits) in their forelimbs. (C) Skeletal staining of E18.5 embryos. Kif7−/− and Gli3−/− mice exhibited enlarged xiphoids. Kif7−/− mice also showed fusion and branching of ribs. (D and E) Western blot analysis was performed on E11.5 wild type (WT) and Kif7−/− embryos. Full-length Gli2 (Gli2-185) was increased in Kif7−/− mice, whereas the processed form of Gli3 (Gli3-83) was reduced (top blots). Asterisks indicate the nonspecific bands detected by the Gli3 antibody. Actin was used as a loading control (bottom blots).

Regulation of the processing and degradation of Gli transcription factors are critical events in both invertebrate and vertebrate Hh signal transduction (2528). In Drosophila, Cos2 is essential for Ci processing into a repressor form (910), and truncated Ci repressor is not detected in cos2−/− cells. Furthermore, elevated amounts of full-length Ci are detected in cos2−/− flies, revealing a role for Cos2 in Ci stability or degradation (10). In mice, the abundance of Gli proteins is tightly regulated by Hh pathway activity, and the 190-kD full-length Gli3 (Gli3-190) is efficiently processed into a truncated 83-kD repressor form (Gli3-83) when the Hh pathway is inactive (2528). We reasoned that the Hh-mediated patterning defects in Kif7 mutants may be due to changes in the amounts of Gli2 and Gli3 protein (Fig. 2) and performed Western blot analysis on total embryo lysates with antibodies against Gli2 or Gli3. The Kif7 mutation consistently increased the 185-kD full-length Gli2 (Gli2-185) and reduced Gli3-83 in total embryo lysates (Fig. 2, D and E; see also Fig. 1, C and D). These results suggest that Kif7 may be involved in the regulation of Gli2 abundance and that it may play a role in the processing of Gli3 into its truncated repressor form or in the stabilization of the Gli3 repressor form.

Kif7 acts downstream of Smo

Next, we tested whether Kif7 acts downstream of Smo in mammalian Hh signaling by examining the phenotype of Smo;Kif7 mutant mice (Fig. 3A). Complete absence of Hh signaling in mice lacking Smo function leads to embryonic lethality at E9.0 (embryonic day 9.0) with holoprosencephaly (midline defect resulting in closely spaced facial features and cephalic abnormalities), heart looping and embryo turning defects, and severe growth retardation (29). In contrast, E9.0 Kif7−/− embryos were exencephalic but were normal in size. Loss of Kif7 function partially suppressed the mutant phenotype of Smo−/− embryos; Smo−/−;Kif7−/− embryos were larger than Smo−/− embryos and exhibited an exencephalic brain phenotype similar to that seen in Kif7−/− embryos (Fig. 3A). Shh signaling is required for the specification of various ventral neuronal progenitor cell populations in the spinal cord, which are identified by the presence of Shh and the Foxa2 transcription factor in the floor plate cells, the Nkx2.2 transcription factor in the V3 progenitor cells, and the Olig2 transcription factor in the motor neuron progenitors (1920). Floor plate cells (Shh+ and Foxa2+), V3 progenitors (Nkx2.2+), and motor neuron progenitors (Olig2+) were absent in Smo−/− neural tubes (28) (Fig. 3, B and C). In E9.0 Kif7−/− neural tubes, these cell types formed, but there was a consistent expansion and ectopic formation of Nkx2.2+ and Olig2+ neural progenitors. In Smo−/−;Kif7−/− neural tubes, despite the absence of Shh and Foxa2, Olig2 was found throughout the prospective spinal cord and scattered Nkx2.2 could also be detected (Fig. 3, B and C). These results indicate that even in the absence of Shh signaling in the Smo mutant background, Kif7 inactivation alone can lead to robust formation of Olig2+ motor neuron progenitors. Previous studies indicate that the Gli3 repressor form is increased when Hh signaling is inactive (25) and that removal of Gli3 in Smo mutants largely rescues the developmental defects in Smo−/− embryos (30). To investigate whether the effects of the Kif7 mutation on Gli2-185 and Gli3-83 protein abundance are dependent on Shh signaling, we performed Western blot analysis on Smo−/−;Kif7−/− embryo lysates. Similar to Kif7−/− embryos, Smo−/−;Kif7−/− embryos showed elevated Gli2-185 and reduced Gli3-83 (Fig. 3, D and E). These observations support a critical role for Kif7 downstream of Smo in the control of Gli activator and repressor protein stability or processing and offer a molecular explanation as to why the Smo mutant phenotype is partially rescued by the removal of Kif7 function. Together with the in ovo results (Fig. 1F), these molecular genetic data in mice show a key regulatory function of Kif7 in the Hh signaling cascade.

Fig. 3

Kif7 acts downstream of Smo. (A to C) Loss of Kif7 function partially rescued the mutant phenotype of Smo−/− mice, suggesting that Kif7 is epistatic to Smo in Hh signaling. (A) At E9.0, Smo−/− embryos displayed growth retardation, whereas Kif7−/−;Smo−/− embryos exhibited a less severe mutant phenotype, similar to that seen in Kif7−/− embryos. (B) Absence of Shh signaling led to loss of Nkx2.2 (green) and Olig2 (red) in Smo−/− neural tubes. Loss of Kif7 function in the Smo−/− background restored Olig2 and, to a lesser extent, Nkx2.2 in Smo−/−;Kif7−/− neural tubes. White arrowheads indicate ectopic Nkx2.2 in Kif7−/− and Smo−/−;Kif7−/− neural tubes. (C) Shh (red) and Foxa2 (green) were absent in Smo−/−;Kif7−/− neural tubes. (D and E) Western blotting of E9.0 embryo lysates revealed increased Gli2-185 (D) and reduced Gli3-83 (E) in Kif7−/− and Smo−/−;Kif7−/− mice (top blots), indicating that the effects of the Kif7 null mutation on Gli protein stability and processing are not dependent on Shh signaling. Asterisks indicate nonspecific bands. Actin was used as a loading control (bottom blots).

Kif7 acts both positively and negatively in Shh-dependent ventral neural tube patterning

In Drosophila, genetic studies indicate that Cos2 is required for the expression of Hh target genes, such as en, which are only induced by the greatest amount of Hh signaling (1011). We therefore asked whether Kif7 also acts positively in mammalian Hh signaling. In mice, floor plate cells are specified by the greatest amount of Shh signaling through Gli2-mediated transcriptional events, and these cells are absent in Gli2−/− and Gli1−/−;Gli2+/− mice (3132). To investigate a potential positive role for Kif7 in mammalian Hh signaling, we analyzed floor plate induction in Kif7;Gli2 mutant mice. Floor plate induction was unaffected in Kif7−/− mice (Fig. 3C). However, floor plate expression of Shh was reduced in E10.5 Kif7−/−;Gli2+/− mice, which displayed a defect in floor plate differentiation similar to (albeit less severe than) that observed in Gli2−/− mice (Fig. 4A). At E14.5, neural tubes in both Gli2−/− and Kif7−/−;Gli2+/− mice lost the characteristic epithelial floor plate structure at the ventral midline (Fig. 4B). The floor plate defect of Kif7−/−;Gli2+/− mice was evident when other ventral neural tube markers were assayed (Fig. 4, C and D). In Gli2−/− mice, the presence of Shh and Nkx2.2 in the ventral neural tube was lost and Olig2 was found in the ventral midline. Reduction of one dose of Gli2 in Kif7−/− mice decreased Olig2 and Nkx2.2 in the ventral neural tube compared with Kif7−/− mice (Fig. 4, C and D). Consistent with the defective floor plate differentiation observed in Kif7−/−;Gli2+/− mice, Nkx2.2+ cells were found at the ventral midline and Shh was reduced. These results show a positive role for Kif7 in Shh-dependent floor plate induction and indicate that Kif7, like Drosophila Cos2, also promotes Hh signaling.

Fig. 4

Kif7 acts both positively and negatively in Shh-dependent ventral neural tube patterning. (A) RNA in situ hybridization analysis of E10.5 embryos revealed a complete loss of floor plate expression of Shh in Gli2−/− mice and reduced floor plate expression of Shh in Kif7−/−;Gli2+/− mice. Arrowheads indicate regions with broken or missing expression of Shh in the floor plate of the corresponding mutant embryos. n, notochord; fp, floor plate. (B) Histological analysis of E14.5 embryos showed loss of a floor plate structure at the ventral midline of Gli2−/− and Kif7−/−;Gli2+/− neural tubes. (C) Immunostaining of ventral neural tubes revealed loss of Nkx2.2 (green) (indicated by white arrowhead) and reduction of Olig2 (red) in Gli2−/− mice. Kif7−/−;Gli2+/− mice showed ectopic Olig2+ cells (red arrowheads). H&E, hematoxylin and eosin. (D) Immunostaining revealed Shh+ cells in the notochord of all mutants; however, Shh+ cells were absent in the prospective floor plate region of Gli2−/− mice and decreased in the floor plate region of Kif7−/−;Gli2+/− mice. White arrowheads indicate floor plate regions with reduced Shh. (E to G) Immunostaining of mutant neural tubes by (E) Nkx6.1 (green) and Pax6 (red); (F) Nkx2.2 (green) and Olig2 (red); and (G) Shh (red). Ventralized neural tubes in Kif7−/− mice were evident by the expansion of Nkx2.2, Olig2 and Nkx6.1 immunostaining. Kif7+/−;Gli3−/− mutants showed normal ventral neural tubes. Loss of Gli3 in Kif7−/− mice resulted in a dose-dependent expansion of Nkx2.2, Olig2, and Nkx6.1 immunostaining in Kif7−/−;Gli3+/− and Kif7−/−;Gli3−/− mice. Green and red arrowheads in (E) and (F) indicate ectopic Nkx6.1 and Olig2, respectively. Bars in C to G indicate 50 μm.

To further define the inhibitory effect of Kif7 on Hh signaling, we examined dorsal-ventral patterning of the spinal cord in Kif7;Gli3 embryos. In the ventral neural tube, strong Shh signal induces Nkx2.2 in the V3 domain, whereas lower Shh signals induce Nkx6.1 and Olig2 in progenitor domains dorsal to the V3 domain (1920). Similar to Gli3−/− mice, Kif7+/−;Gli3−/− mice did not exhibit any patterning defects in the ventral spinal cord (Fig. 4, E to G). In contrast, although Shh was not affected, loss of Gli3 function in Kif7−/− mice resulted in a dose-dependent dorsal expansion of Nkx6.1+, Nkx2.2+, and Olig2+ progenitor domains (Fig. 4, E to G). Loss of one dose of Gli3 in Kif7−/− mice caused additional ectopic dorsal induction of Olig2+ motor neuron progenitors, but not Nkx2.2+ V3 progenitors, whereas complete removal of Gli3 in Kif7−/− mice increased both Olig2+ motor neuron progenitors and Nkx2.2+ V3 progenitors (Fig. 4F). Because loss of Gli3 function alone had no effect on ventral neural tube patterning, the data obtained from these mutants suggest that increased Gli2 activator function due to Kif7 inactivation contributes to increased Nkx2.2, Olig2, and Nkx6.1 in Kif7;Gli3 mutant mice and support the conclusion that Kif7 is a major inhibitor of the mammalian Hh pathway. Furthermore, the analysis of Kif7;Gli2 and Kif7;Gli3 mutant mice revealed a critical role for Kif7 in graded Shh signaling in the neural tube by regulating the balance of Gli activators and repressors.


We present here biochemical and genetic evidence showing that Kif7 is a critical regulator of mammalian Hh signaling. Similar to Cos2, Kif7 interacts with Gli transcription factors, influencing their protein abundance and processing. Furthermore, like Cos2, Kif7 acts both positively and negatively in the regulation of Hh target gene expression. Despite their functional similarities, there appear to be mechanistic differences between Cos2 and Kif7. In Drosophila, Cos2 interacts directly with both Smo and the serine-threonine kinase Fused (Fu) for Hh-dependent regulation of Ci (311). Unlike Cos2, Kif7 does not interact directly with Smo (13), and Fu does not play a role in mammalian Hh signaling (3334). Moreover, Cos2 is known to bind microtubules in an Hh-dependent manner (3, 79). Additional studies will be needed to probe the molecular function of Kif7 and its interaction with microtubules. Investigating this interaction is of particular interest given the central role that primary cilia, which are microtubule-rich organelles, play in the intracellular transduction of Hh signals in mammalian cells (14, 15). As a Cos2 homolog functioning downstream of Smo, Kif7 may participate in directional transport of Gli proteins and other intracellular Hh pathway components. In addition, the finding that Kif3a indirectly mediates cilium localization of Smo through β-arrestin (15) and our observation that Kif27 interacts with Gli3 (Fig. 1) raise the possibility that other kinesins may also participate in the Hh pathway. Together, our findings provide insight into the molecular mechanism of Hh signal transduction in mammalian cells and reconcile the previous lack of correspondence between results from Drosophila and mammals.

Materials and Methods

Generation of a Kif7 mutant allele

A genomic BAC (bacterial artificial chromosome) clone was isolated from a 129/Sv genomic library with the use of a Kif7 probe. A conditional targeting vector for Kif7 was constructed with the pDELBOY vector, which contains a neomycin cassette flanked by frt sites. A 3.4-kb genomic fragment (fragment I) with exons 2 to 4 of Kif7 was inserted between two loxP sites. A 1.9-kb fragment 5′ to the start codon of Kif7 and a 4.5-kb fragment 3′ to exon 4 were inserted as the left arm and the right arm in the targeting construct, respectively. A conditional loss-of-function mutation in Kif7 was generated in W4 ES cells by standard gene targeting procedures (31, 35). ES cell clones that carry the conditional mutant allele were identified by Southern analysis and used to generate mutant mice by blastocyst injection.


To generate Kif7del/+ mutant mice, exons 2 to 4 were deleted by crossing Kif7flox/+ mice with NLS-Cre transgenic mice (provided by C. Lobe, University of Toronto). Deletion of exons 2 to 4 was confirmed by polymerase chain reaction (PCR). The following primers were used for PCR genotyping: Kif7-P1 (CACCACCATGCCTGATAAAAC), Kif7-P2 (CTATCCCCAATTCAAAGTAGAC), Kif7-P3 (TTCTCACCCAAGCTCTTATCC), and Kif7-P4 (CCAAATGTGTCAGTTTCATAGC). All analysis on Kif7 mutant mice was done on a mixed 129/Sv and CD-1 background. Smo (Jackson Laboratory), Gli2, and Gli3 mutant mice were all maintained in an outbred background of CD1 and genotyped as previously described (29, 35). The research performed with mice was approved by The Hospital for Sick Children Animal Care Committee.

Whole-mount in situ hybridization

The midday after the presence of a vaginal plug was considered E0.5. Pregnant females of the desired stage were killed by cervical dislocation and uteri were dissected out in cold phosphate-buffered saline (PBS). Embryos were microdissected and collected at the specified stages. Yolk sacs were collected for genotyping. Embryos were fixed in 4% paraformaldehyde in PBS at 4°C overnight and subjected to methanol series for dehydration. In situ hybridization was carried out with digoxigenin-dUTP (deoxyuridine triphosphate)–labeled RNA probes for Shh (35).


Immunohistochemistry was performed as described (31, 35). Embryos were fixed in 4% paraformaldehyde in PBS overnight at 4°C and subjected to an ethanol series for dehydration. Embryos embedded in paraffin were cut at 7 μm. The following primary antibodies were used: from Developmental Studies Hybridoma Bank, FoxA2 (1:10), Nkx2.2 (1:20), and Nkx6.1 (1:50); from Chemicon, Olig2 (1:600); and from Covance, Pax6 (1:400). Fluorescence visualization was performed with biotinylated secondary antibodies followed by fluorescence avidin (Vector Laboratories). Images were acquired with a Zeiss LSM510 META laser scanning confocal microscope.

Antibody generation

Kif7 antibodies were generated with the use of standard procedures from affinity-purified rabbit polyclonal antisera raised against the C terminus (amino acid residues 1250 to 1348) of Kif7 fused to glutathione S-transferase.

Skeletal staining

E18.5 embryos were fixed in 80% ethanol overnight and fixed again in 95% ethanol after removal of skin and viscera. Bone samples were incubated in Alcian Blue solution (15% Alcian Blue in 80% ethanol and 20% glacial acetic acid) for 24 hours at room temperature, followed by 95% ethanol wash. Samples were then stained with Alizarin Red solution (7.5% Alizarin Red, 1% potassium hydroxide) for 24 hours and cleared with 1% potassium hydroxide in 20% glycerol. Skeletons were stored in a 1:1 mix of glycerol and ethanol.

Western blot analysis

Tissue or embryos were isolated, snap-frozen, and sonicated in RIPA (radioimmunoprecipitation assay) buffer [50 mM tris-HCl (pH 7.4), 150 mM NaCl, 5 mM EDTA (pH 8.0), 1 mM EGTA (pH 8.0), 25 mM sodium pyrophosphate, 1 mM sodium orthovanadate, 10 mM sodium fluoride, 1 mM β-glycerol phosphate, 0.1% SDS, 1.0% NP-40, 0.5% deoxycholate] with protease inhibitor cocktail (Roche). Western blot analysis was performed using standard protocols. Immunoblotting was performed overnight at 4°C with the following primary antibodies: from Molecular Probes, green fluorescent protein (GFP); from Santa Cruz Biotechnology, c-myc, hemagglutinin, and Gli3; and from Oncogene, β-actin.


COS-7 cells were maintained in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum. For immunoprecipitation, cells were seeded at 90% confluence in 10-cm plates and transfected with Lipofectamine 2000 (Invitrogen). Cells were harvested after 2 days with lysis buffer (150 mM NaCl, 1 mM EGTA and 1% Triton X-100). The cell lysates were incubated with the indicated antibodies overnight at 4°C; immunoprecipitates were pulled down with Protein G–Sepharose beads and subjected to Western blotting analysis.

Luciferase reporter gene assays

Gli-dependent transcription reporter assays were performed by transfecting a luciferase reporter construct containing eight copies of a Gli-binding site (GBS-Luc) with a normalization β-galactosidase expression plasmid into C3H10T1/2 mouse embryonic fibroblast cells with Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions. After 48 hours, cells were processed for the luciferase assay as previously described (36).

Chick in ovo electroporation and immunohistochemistry

Full-length mouse Kif7 complementary DNA was cloned into the pCAGGS expression vector. The SmoM2 expression vector has been previously described (21). In ovo electroporation into the neural tube of Hamburger and Hamilton (HH) stage 11 to 12 chick embryos was performed as previously described, and embryos were harvested after 48 hours (19). Embryos were fixed and processed for fluorescent immunohistochemistry as previously described (19). Images were taken with a Leica TCS-SP2 confocal microscope.

Chick in ovo luciferase assay

SmoM2 and Kif7 constructs, or pCAGGS-GFP as control, were electroporated into HH stage 11 to 12 chick embryos along with a firefly luciferase reporter construct containing eight copies of a Gli-binding site (GBS-Luc) and a normalization plasmid CMV-Renilla luciferase (Promega). After 48 hours, embryos were processed for the luciferase assay as previously described (19). Briefly, embryos were homogenized in passive lysis buffer on ice and firefly and Renilla luciferase activities were measured with the Dual Luciferase Reporter Assay System (Promega).


We thank E. Dessaud and M. Lebel for guidance and helpful discussions. H.O.C. was supported by an Ontario Graduate Scholarship and Studentship from The Hospital for Sick Children Research Institute, A.R. was supported by Fundação para a Ciência e Tecnologia (Portugal), and J.B. was supported by Medical Research Council (UK). This research was funded by the Canadian Cancer Society.

Supplementary Materials

Fig. S1. Interaction of Kif7 and Kif27 with Gli transcription factors.

Fig. S2. Gene targeting of Kif7.

Fig. S3. Kif7 inhibits SmoM2-induced Gli transcriptional activity in ovo.

References and Notes

View Abstract

Navigate This Article