Regulation of Gli1 Localization by the cAMP/Protein Kinase A Signaling Axis through a Site Near the Nuclear Localization Signal*

The hedgehog (Hh) pathway plays a critical role during development of embryos and cancer. Although the molecular basis by which protein kinase A (PKA) regulates the stability of hedgehog downstream transcription factor cubitus interruptus, the Drosophila homologue of vertebrate Gli molecules, is well documented, the mechanism by which PKA inhibits the functions of Gli molecules in vertebrates remains elusive. Here, we report that activation of PKA retains Gli1 in the cytoplasm. Conversely, inhibition of PKA activity promotes nuclear accumulation of Gli1. Mutation analysis identifies Thr374 as a major PKA site determining Gli1 protein localization. In the three-dimensional structure, Thr374 resides adjacent to the basic residue cluster of the nuclear localization signal (NLS). Phosphorylation of this Thr residue is predicted to alter the local charge and consequently the NLS function. Indeed, mutation of this residue to Asp (Gli1/T374D) results in more cytoplasmic Gli1 whereas a mutation to Lys (Gli1/T374K) leads to more nuclear Gli1. Disruption of the NLS causes Gli1/T374K to be more cytoplasmic. We find that the change of Gli1 localization is correlated with the change of its transcriptional activity. These data provide evidence to support a model that PKA regulates Gli1 localization and its transcriptional activity, in part, through modulating the NLS function.

Hedgehog (Hh) 2 proteins are a group of secreted proteins whose active forms are derived from a unique protein cleavage process and at least two post-translational modifications (1,2). Secreted Hh molecules bind to the receptor patched (PTC), thereby alleviating PTC-mediated suppression of smoothened (SMO) (2,3). Expression of sonic hedgehog (Shh) appears to stabilize SMO protein possibly through post-translational modification of SMO (4). The effect of hedgehog molecules can be inhibited by hedgehoginteracting protein (HIP) through competitive association with PTC (5,6). In Drosophila, SMO stabilization triggers complex formation with Costal-2, Fused, and Gli homologue cubitus interruptus (CI), which prevents CI degradation and formation of a transcriptional repressor (7)(8)(9)(10). SMO ultimately activates transcription factors of the Gli family. Gli molecules enter nucleus through a nuclear localization signal (11,12), but little is known about the regulatory mechanism for this process. As transcriptional factors, Gli molecules can regulate target gene expression by direct association with a consen-sus binding site (5Ј-tgggtggtc-3Ј) located in the promoter region of the target genes (13,14).
Protein kinase A (PKA) was first identified as an inhibitory component of the Hh pathway in Drosophila (15)(16)(17)(18)(19). PKA fulfills its negative role by phosphorylating full-length Ci (Ci155) at several Ser/Thr residues, priming it for further phosphorylation by glycogen synthase kinase 3 (GSK3) and casein kinase I (CKI) (20 -22). Hyperphosphorylation of Ci155 targets it for proteolytic processing to generate the repressor form (Ci75) (23). Consistent with this, overexpressing a constitutively active form of PKA catalytic subunit (PKAc), mC*, blocks Ci155 accumulation and Hh target gene expression (24). In addition to its inhibitory effects, PKA phosphorylation at the C terminus of SMO in Drosophila, but not in mammals, enhances hedgehog-mediated signaling (25,26).
While the regulation of CI cleavage by PKA phosphorylation is well documented, very little is known about the role of PKA in Gli regulation. In vertebrates, there are three Gli molecules, Gli1, Gli2, and Gli3. Gli3 can be processed in a manner similar to CI, a process regulated by PKA (27,28). The fact that Gli3 expression is often not detectable in human cancer suggests that Gli3 does not play a significant role in hedgehog-driven carcinogenesis (29 -31). In contrast, Gli1 and Gli2 are expressed in tumors with activated hedgehog signaling (2, 29 -36). Here, we report that the cAMP/PKA signaling axis regulates Gli1 protein localization, in part, through phosphorylation of Gli1 at a site near the nuclear localization signal (NLS). We propose that this unique regulation is an important mechanism by which PKA inhibits transcriptional activity of Gli molecules.

MATERIALS AND METHODS
Cell Culture and Plasmids-COS7 and NIH3T3 cells were purchased from the American Type Culture Collection (Manassas, VA) and were maintained in Dulbecco's modified Essential medium supplemented with 10% fetal bovine serum (Invitrogen). Transfection was performed using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions (the ratio of plasmid (g) to lipid (l) was 1:2.5). Stable expression of a luciferase reporter under the control of Gli responsive elements in NIH3T3 cells was achieved through selection with G418 for 3 weeks after transient transfection with Lipofectamine 2000 (Invitrogen). Two clones with good responses to Gli1 expression (over 20-fold) were selected from a total of 80 clones.
Cells with expression of Gli1 (with C-MYC tag) were treated with 10 M H89 (Calbiochem) or 0.4 ng/ml leptomycin B (LMB, Sigma) for 8 h. For forskolin treatment (20 M for 8 h), cells were pretreated with phosphodiesterase inhibitor IBMX (100 M) for 30 min before addition of forskolin. Immunofluorescent detection of C-MYC-tagged Gli1 was performed as described previously with Cy3-conjugated MYC antibody 9e10 (Sigma) (1:100 dilution) (37). Gli1 localization was detected under a fluorescent microscope; the percentage of Gli1 in the nucleus or the cytoplasm was calculated for each experiment from over 200 Gli1-expressing cells, and the experiment was repeated three times.
Gli1 cDNA was generously provided by Dr. Bert Vogelstein and cloned into pCDNA3.1 with a C-MYC tag at the N terminus. Gli1-GFP construct was made by subcloning Gli1 cDNA into pEGFP-3C using BamHI (5Ј) and NotI (3Ј) sites. Point mutations of Gli1 were made by in situ mutagenesis in our DNA Recombinant Laboratory Core Facility or by PCR-based mutagenesis. All mutations were confirmed by sequencing of the entire coding region. Clones containing only the targeted mutations were used in the studies.
Immunoprecipitation and in Vitro Kinase Assay-Cells were lysed using Nonidet P-40 cell lysis buffer 48 h following Gli1 transfection. C-MYC-tagged Gli1 proteins were immunoprecipitated with an anti-MYC 9B11 antibody (Cell Signaling Inc.) for 3 h followed by incubation with A/G plus beads (Bethyl Laboratories, Inc.) for 1 h. The immunocomplexes were divided into two portions. One portion (20%) was separated by 10% SDS-PAGE and analyzed by Western blotting using an anti-MYC antibody 9B11. The remaining 80% of them were incubated with 0.6 ng of recombinant PKA catalytic subunit with ADBI buffer (20 mM MOPS, pH 7.2, 25 mM ␤-glycerol phosphate, 5 mM EGTA, 1 mM sodium orthovanadate, 1 mM dithionthreitol) containing 1 mg/ml bovine serum albumin, and 10 Ci of [␥-32 P]ATP at 30°C for 30 min. The kinase reactions were terminated by adding 4ϫ SDS sample buffer, and the samples were separated by 10% SDS-PAGE. Gels were dried on Whatman * This work was supported by NCI/National Institutes of Health Grant R01CA94160, Department of Defense Grant DOD-PC030429, and NIEHS/National Institutes of Health Grant ES06676. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. □ S The on-line version of this article (available at http://www.jbc.org) contains a supplemental figure. 1  paper, followed by autoradiography. Western blotting analysis was performed according to a previously published procedure (37). Cellular PKA Activation Assay-COS7 cells were seeded into 6-well plates the day before the experiment. Cells were serum-starved for 12 h and then treated with 20 M H89, 20 M forskolin or both for 45 min. 100 M IBMX was added for 30 min prior to treatment with forskolin. Cellular PKA activity in the above-described conditions was determined using a PKA activation assay (Upstate Biotechnology Inc.), in which Leu-Arg-Arg-Ala-Ser-Leu-Gly (Kemptide) was used as the substrate (38). In brief, cell lysates were prepared using Nonidet P-40 lysis buffer. After sonication, the cell lysates were collected at 12,000 rpm for 5 min. 10-l cell lysates were incubated with 10 M ATP containing 10Ci [␥-32 P]ATP (3,000 Ci/mmol)], 250 M Kemptide substrate in ADBI buffer at 30°C for 10 min. Background was determined from reactions without substrate, and the total PKA activity was estimated in reactions containing 20 M dibutyryl-cAMP. Aliquots were spotted onto Whatman P-81 paper, and the filters were washed in 0.75% phosphoric acid three times for 5 min per wash. 32 P incorporation was determined by liquid scintillation counting. Protein concentration of the cell lysates was determined by a kit from Bio-Rad (Bio-Rad protein assay). Data are representative of three independent experiments.
Cell Fractionation-Following transfection, the COS7 cells were maintained in 10 cm cell culture dishes for 48 h. Before harvest, the cells were rinsed twice with cold phosphate-buffered saline, harvested by scraping with 1 ml of cold phosphate-buffered saline for each 10-cm dish, and collected by centrifugation at 3,000 ϫ g for 30 s. One-fifth of the lysates were collected for detection of Gli1 expression by Western blotting. The cell pellets were incubated with 60 l of buffer A (50 mM HEPES, pH 7.4, 10 mM KCl, 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, 0.1 g of phenylmethylsulfonyl fluoride/ml, 1 g of pepstatin A/ml, 1 g of leupeptin/ml, 10 g of soybean trypsin inhibitor/ml, 10 g of aprotinin/ml, and 0.1% IGEPAL CA-630). After 10 min on ice, the lysates were centrifuged at 6,000 ϫ g for 30 s at 4°C. The supernatant fractions were saved as the cytoplasmic fraction. The pellet (containing the nuclei) was resuspended in buffer B (buffer A containing 1.0 M sucrose) and centrifuged at 15,000 ϫ g for 10 min at 4°C. The purified nuclei (pellets) were incubated in 30 l buffer C (10% glycerol, 50 mM HEPES, pH 7.4, 400 mM KCl, 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, 0.1 g of phenylmethylsulfonyl fluoride/ml, 1 g of pepstatin A/ml, 1 g of leupeptin/ml, 10 g of soybean trypsin inhibitor/ ml, 10 g of aprotinin/ml) after vigorous vortex and centrifuged at 15,000 ϫ g for 20 min at 4°C, and supernatant fractions were collected. All extracts were normalized for protein amounts determined by Bio-Rad protein assay (Bio-Rad) and separated by 10% SDS-PAGE for further analysis (39). Antibodies to ␤-tubulin (Sigma) and Lamin A/C (Santa Cruz Biotechnology Inc.) were used to detect the purity of cytoplasmic (␤-tubulin) and nuclear (Lamin A/C) fractions using Western blotting analysis.
Luciferase Reporter Gene Assay-For luciferase assay, NIH3T3 cells with stable expression of the Gli luciferase reporter were transfected with Gli1 constructs (0.5 g/well) and the TK-Renilla control plasmid (5 ng/well) using Lipofectamine 2000. After 5 h, the medium was replaced with fresh growth medium and the cells were incubated in 5% CO 2 at 37°C overnight. Following treatment with forskolin or other compounds, the cells were harvested, and luciferase activity was measured with the dual luciferase reporter assay system (Promega) according to the manufacturer's instruction. In brief, the transfected cells were lysed in the 6-well plates with 100 l of reporter lysis buffer and the lysate transferred into Eppendorf tubes. Cell debris was removed by centrifugation at top speed for 10 min in a microcentrifuge. 20 l of the supernatant was mixed with 100 l of buffer LAR II, and the absorbance was immediately measured (the first reading). After 3 s, 100 l of Stop & Glo Reagent was added to measure the Renillar luciferase activity (the second reading). The value from the first reading was divided by the value from the second reading of each sample to obtain the luciferase activity. Each experiment was repeated three times with similar results.

RESULTS
Gli1 is not only a downstream effector but also a target gene of the hedgehog pathway (40). Thus, identification of the mechanism by which PKA phosphorylation regulates Gli1 functions will help us understand signal transduction of the hedgehog pathway in cancer.
First, we tested whether Gli1 protein localization can be altered by accumulation of the cellular cAMP level in COS7 cells. After transient transfection, the protein localization of Gli1 was detected by immunofluorescent staining of the C-MYC tag at the Gli1 N terminus and by cell fractionation. In the presence of 20 M forskolin, which directly activates adenylyl cyclase and raises the cyclic AMP level (41), we observed that the percentage of cytoplasmic Gli1 was increased over 5-fold, whereas the percentage of nuclear Gli1 was reduced by 80% (Fig. 1, A and B). The effect seems to be direct because the change in Gli1 localization can be observed 20 min after forskolin treatment. Conversely, addition of PKA inhibitor H89 led to a shift of Gli1 localization to the nucleus (Fig. 1, A and B). As a consequence of forskolin treatment, the cellular PKA activity was increased 2-fold (Fig. 1C). Conversely, addition of H89 into the medium inhibited the cellular PKA activity by 70% (Fig. 1C). Thus, Gli1 localization was correlated with the cellular PKA activity. To confirm the data from the immunofluorescent staining, we performed cell fractionation analysis. As shown in Fig. 1B, more Gli1 were in the nuclear fraction following H89 treatment whereas forskolin caused an increase of cytoplasmic Gli1 (Fig. 1B). Furthermore, we monitored localization of Gli1-GFP fusion protein with a time-lapse microscope in the presence of H89 or forskolin. Forskolin retained Gli1-GFP in the cytoplasm whereas H89 promoted nuclear accumulation of this fusion protein (data not shown). All these data indicate that Gli1 localization can be regulated by modulating the cellular PKA activity.
Gli1 protein shuttling between the nucleus and the cytoplasm was interrupted by inhibition of nuclear export with LMB, a specific inhibitor for CRM1-mediated nuclear export, resulting in nuclear accumulation of all Gli1 proteins (Fig. 1, A and B), supporting that Gli1 localization is a dynamic process and is tightly regulated. Our data suggest that direct phosphorylation of Gli1 by PKA is responsible for regulation of Gli1 protein localization.
Sequence analysis predicts five putative PKA sites in Gli1. The sequence around these PKA sites is highly conserved among Gli proteins ( Fig. 2A). Several point mutations of Gli1 were made to test Gli1 regulation by PKA (see the diagram in Fig. 2B). These mutations were made by in situ mutagenesis in our DNA Recombinant Laboratory Core Facility or by PCR-based mutagenesis.
Using these mutant constructs, we assessed Gli1 localization in cultured . Both the nuclear fraction and the cytoplasmic fraction were collected for Western blotting analysis. The purity of cell fractionation was assessed using lamin A/C for the nuclear fraction and ␤-tubulin for the cytoplasmic fraction. C, the cellular PKA activity in COS7 cells was determined using a kit from Upstate Biotechnology Inc. The high PKA activity (C) was correlated with a higher level of cytoplasmic Gli1.
cells. We found that mutation at Thr 374 (Gli1/T374V) significantly affects Gli1 localization (Fig. 2C). Furthermore, the response of Gli1/T374V to forskolin treatment was nearly diminished (in response to forskolin treatment, 5-fold increase in cytoplasmic protein for wild type Gli1 but only 40% increase for Gli1/T374V) (also see the supplemental figure). In contrast, the protein localization of Gli1/S544G and Gli1/S544G/S560A was not different from the wild type Gli1 (supplemental figure). A mutant Gli1 (Gli1/T374V/S544G/S560A) with triple mutations at the PKA sites behaved like Gli1/T374V (supplemental figure), indicating that Ser 544 and Ser 560 are not involved in regulation of Gli1 localization. On the other hand, a mutation at S640A had only slight effects on Gli1 protein localization in response to forskolin (supplemental figure). These data indicate that Thr 374 is the major site responsible for Gli1 protein localization in cultured cells. Consistent with the role of Thr 374 for Gli1 protein localization, we also confirmed that Thr 374 can be phosphorylated by recombinant PKA in vitro (Fig. 2D). We performed PKA phosphorylation with Gli1 protein purified by immunoprecipitation in the presence of [␥-32 P]ATP and recombinant PKA in test tube and found that Gli1 was highly phosphorylated by PKA in vitro (Fig.  2D, left panels, first lane). Gli1 with mutations of all five PKA sites (PKA⌬) could not be phosphorylated by PKA (Fig. 2D, left panels, right lane). To test whether the Thr 374 site of Gli1 is phosophorylated by recombinant PKA, we used a Gli1 fragment containing only two PKA sites: Thr 296 and Thr 374 . We found that this Gli1 fragment (1-514 aa) with a mutation at Thr 374 was not able to be phosphorylated by recombinant PKA in vitro (Fig. 2D, center panels). Even a single point mutation at Thr 374 significantly reduced PKA-mediated phosphorylation of full-length Gli1 (Fig. 2D, right panels), indicating that the Thr 374 site is a major PKA site of Gli1 phosphorylation. In addition, our data also suggested that Ser 544 , Ser 560 , and Ser 640 can be phosphorylated by PKA in vitro (data not shown here). The above data indicate that Thr 374 is a major PKA site involved in regulation of Gli1 protein localization.
In the three-dimensional structure, Thr 374 , together with the adjacent Asp 375 , is close to the first basic residue cluster (Arg 380 /Lys 381 /His 382 ) of the bipartite motif (the classic NLS) in Gli1 (Fig. 3A) (42). It is known that protein phosphorylation at the residue next to the bipartite motif inhibits its binding affinity to importins, leading to reduced nuclear localization of the target protein (43). We predict that phosphorylation of Thr 374 will increase the local negative charge, leading to reduced NLS functions and accumulation of Gli1 in the cytoplasm. Indeed, Gli1/T374D was preferentially localized to the cytoplasm (Fig. 3B). Conversely, Gli1/T374K has a high local positive charge near the NLS, and we found that Gli1/T374K predominantly localized to the nucleus (Fig. 3B). Localization of these Gli1 proteins was further confirmed by A shows the sequence alignment of CI and human Gli1, Gli2, and Gli3 at the five putative PKA sites. B shows the Gli1 constructs used in this study. Gli1 molecules with point mutations of one or more PKA sites were expressed in COS7 cells and their localization was detected by immunofluorescent staining. Full-length Gli1 with a mutation at Thr 374 (T374V) had the most significant effect on Gli1 protein localization (C). In contrast, a mutation at Ser 544 or other sites (C and see supplemental figure) had little effects on Gli1 protein localization. A Gli1 mutant Gli1/T374V/S544G/S560A with triple mutations at PKA sites behaved like Gli1/ T374V (supplemental figure), indicating that Thr 374 is a critical PKA site for determining Gli1 protein localization. Over 200 Gli1-positive cells were counted under a fluorescent microscope for Gli1 protein localization, and the experiment was repeated three times with similar results. The data were the average result from these experiments. D shows Gli1 phosphorylation in vitro by recombinant PKA. Wild type Gli1 and its mutant forms were expressed in COS7 cells and subsequently purified through immunoprecipitation. The ability of PKA to phosphorylate immunoprecipitated Gli1 proteins were performed in vitro (see "Materials and Methods"). The full-length Gli1, but not Gli1-PKA⌬ (see B, the mutation sites), was highly phosphorylated by PKA in vitro (D, left panels). A single point mutation in a Gli1 fragment (1-514 aa, shown in B) prevented protein phosphorylation by PKA (D, center panels). This same mutation in the full-length Gli1 also dramatically reduced the level of phosphorylation (D, right panels). Localization was confirmed by cell fractionation. Immunofluorescent staining was done as described for Fig. 1A. C shows Gli1/T374K localization following an additional mutation at Lys 381 . These experiments suggest that Thr 374 is responsible for cAMP/PKA-mediated regulation of Gli1 localization, which may be achieved through affecting the NLS function. D, correlation of Gli1 localization with its transcriptional activity was examined in NIH3T3 cells with stable expression of a luciferase reporter under the control of Gliresponsive elements (see "Materials and Methods" for details). Gli1 with a Lys to Glu mutation at Lys 381 (K381E) of the NLS motif, which predominantly localizes to the cytoplasm, was unable to activate this reporter. In contrast, Gli1 with mutations in the NES motif, which predominantly localizes to the nucleus, was more active than the wild type Gli1. E, transcriptional activity of Gli1/T374V was not responsive to forskolin treatment whereas the transcriptional activity of wild type Gli1 was reduced by 60% following forskolin treatment.
cell fractionation (Fig. 3B). These data support our hypothesis that one mechanism by which the cellular PKA activity regulates Gli1 localization is through altering the local charge nearby the NLS of Gli1.
If Gli1 nuclear localization is regulated by PKA phosphorylation through a NLS-dependent mechanism, disruption of the NLS should affect localization of these mutant Gli1 molecules. As shown in Fig. 3C, we found that Gli1/ K381E, which is predicted to disrupt the NLS, was predominantly localized to the cytoplasm, confirming the role of NLS in Gli1 localization (12). Although Gli1/T374K localizes predominantly to the nucleus, additional mutation at K381 (K381E) retained Gli1/T374K to the cytoplasm (Fig. 3C), suggesting that regulation of Gli1 localization by T374 phosphorylation requires the intact NLS.
The ultimate effect of Gli1 is transcriptional activation of the downstream target genes. To assess whether Gli1 localization affects its transcriptional activity, we established stable expression of Gli1 luciferase reporter under the control of Gli responsive elements in NIH3T3 cells (13). By measuring the reporter luciferase activity, we examined the association of Gli1 localization with its transcriptional activity (Fig. 3D). If Gli1/K381E remains preferentially in the cytoplasm, the transcriptional activity was low. In contrast, Gli1 with a defective NES (Gli1/L496V/L498V), which localizes predominantly in the nucleus, was more active (Fig. 3D). Thus, the Gli1 luciferase reporter activity in these cells is very sensitive to Gli1 localization. We examined the effects of forskolin to Gli1-mediated transcriptional activity in this NIH3T3 stable cell line. Like the wild type Gli1, Gli1/T374V can activate the Gli luciferase reporter (Fig. 3E). Consistent with its cytoplasmic localization, the luciferase activity in cells expressing the wild type Gli1 was reduced after forskolin treatment (Fig.  3E). In contrast, Gli1/T374V-mediated reporter gene activity was not affected by forskolin (Fig. 3E). These data suggest that Thr 374 is an important PKA site responsible for PKA phosphorylation and for the transcriptional activity of Gli1.
Based on these data, we proposed a mechanism by which the cAMP/PKA signaling axis mediates regulation of Gli1 localization. Gli1 enters the nucleus through a nuclear localization signal. With the accumulation of cAMP in the cell, Thr 374 gets phosphorylated. Phosphorylation of Thr 374 will increase the local negative charge nearby the NLS, which results in inhibition of NLS function. Consequently, Gli1 is retained in the cytoplasm and is unable to activate the target genes. Since this Thr residue is highly conserved among Gli proteins, we anticipate that the same mechanism is applicable to PKA regulation of other Gli molecules.

DISCUSSION
In our study, we provide direct evidence to support that the cAMP/PKA signaling axis regulates Gli1 protein localization primarily through a site at Thr 374 . Our data further indicate that PKA-mediated regulation of Gli1 localization is through Thr 374 , possibly through interfering with the NLS function.
Although our studies demonstrated that Thr 374 is a major site for PKAmediated regulation of Gli1 localization (Figs. 2 and 3 and supplemental figure), mutation at this site did not completely abolish the effects of forskolin (supplemental figure). To identify an additional site required for this regulation, we made double and triple point mutations of Gli1 at the PKA sites (Fig.  2B). Our data indicate that mutations at Thr 374 and Ser 640 completely abolished the response to forskolin treatment, indicating that Ser 640 is another site involving PKA-mediated regulation of Gli1 localization (supplemental figure). However, mutation at Ser 640 alone had no effect on Gli1 protein localization and had only a slight effect in response to forskolin treatment, suggesting that Ser 640 is not a primary site for Gli1 regulation. This hypothesis was further supported by the fact that Gli1/S640E and Gli1/S640R, unlike Gli1/T374D and Gli1/T374K, did not alter Gli1 protein localization (data not shown here). Thus, we believe that additional structural information of Gli1 near the Ser 640 region is required to understand the molecular mechanism by which PKA phosphorylation at Ser 640 affects Gli1 localization.
Further studies of Gli1 phosphorylation can be facilitated by measuring the stoichiometry of phosphorylation for Gli1. Currently, a large amount of purified Gli1 protein is not available, making it difficult to calculate the stoichiometry of phosphorylation for such a large protein (150 kDa for Gli1). This issue may be addressed in the future using highly purified Gli1 fragments from bacteria.
In addition, our data indicate that Ser 544 and Ser 560 residues are not involved in regulation of Gli1 protein localization. It will be interesting to know how these two sites are involved in regulation of Gli1 functions.