Gli2 Is Targeted for Ubiquitination and Degradation by β-TrCP Ubiquitin Ligase*

The Hedgehog (Hh) signaling pathway plays a crucial role in embryogenesis and has been linked to the development of several human malignancies. The transcription factor Gli2 plays a key role in the transduction of Hh signals by modulating transcription of some Hh target genes, yet the mechanisms that control Gli2 protein expression are largely unknown. Here we report that β-transducin repeat-containing protein (β-TrCP) E3 ubiquitin ligase is required for Gli2 degradation. β-TrCP2 directly binds wild type Gli2 and promotes its ubiquitination. Single amino acid substitution in Gli2 putative binding site inhibits its interaction with β-TrCP2, its ubiquitination, and stabilizes the Gli2 protein. Stable Gli2 mutant is expressed in higher levels and is more potent in the activation of Gli-dependent transcription as compared with wild type Gli2. We also found that GLI2 protein is expressed highly in prostate cancer cell lines and primary tumors, whereas the level of GLI2 mRNA is not appreciably different in normal and neoplastic prostate. These data identify β-TrCP2 as a pivotal regulator of Gli2 expression and point to an important role for posttranslational modulation of GLI2 protein levels in Hh pathway-associated human prostate cancer.

The Hedgehog (Hh) 2 signaling pathway plays a prominent role in embryogenesis, and its deregulation is implicated in tumorigenesis (reviewed in Refs. [1][2][3][4]. Cellular responses to the Hedgehog signal are controlled by two transmembrane proteins, the tumor suppressor patched (PTCH) and the oncoprotein smoothened (SMO). In the absence of secreted Hh proteins, PTCH actively silences intracellular signaling by inactivating SMO. During physiologic signaling, Hh proteins bind and inactivate PTCH, which alleviates PTCH-mediated suppression of SMO (5). SMO activation triggers a series of intracellular events, culminating in expression of Hh target genes through the action of the Gli family of transcription factors (6,7). Gli1, Gli2, and Gli3 and their Drosophila homolog, Cubitus interruptus (Ci) are zinc finger transcription factors that are downstream effectors of Hh signaling. In the absence of Hh signaling, Ci is truncated at the carboxyl-terminal domain to form a truncated repressor protein, whereas Hh activation leads to accumulation of transcriptional active, full-length Ci. The situation with mammalian Gli proteins is more complex. Gli3 functions primarily as a C-terminally truncated repressor, but full-length Gli3 protein accumulates in cells responding to Hh. Gli1, on the other hand, appears to modulate gene expression by acting primarily as a transcriptional activator, but Gli1 mutant mice are phenotypically normal, arguing against an essential function for this protein during development or postnatal life. Gli2 appears to be the major nuclear effector of Hh signaling in vivo (8 -12) and functions primarily as a transcriptional activator. However, little is known about the molecular mechanisms regulating Gli2 expression at the protein level.
The Hh signaling pathway is deregulated in many human malignancies, including basal cell carcinoma (BCC), medulloblastoma, lung, prostate, breast, and some gastrointestinal cancers (13)(14)(15)(16)(17)(18)(19). Recent studies have stressed the importance of Hh signaling in human prostate cancer (13)(14)(15)20). Elevated Hh signaling pathway activity may distinguish metastatic from localized prostate cancer, and pathway manipulation can modulate invasiveness and metastasis (13). In contrast to BCC and medulloblastoma, which are associated with inactivating mutations in PTCH or gain-of-function mutations in SMO, aberrant Hh signaling in prostate cancers appears to be the result of constitutive overexpression of sonic hedgehog. Hence, the growth of many of the prostate cancer cells is inhibited by Hhneutralizing antibody. Furthermore, cyclopamine, a steroidal alkaloid that interacts with SMOH directly, thus inhibiting Hh signaling, was shown to induce apoptosis and inhibit proliferation of prostate cancer cells in vivo as well as in vitro.
The ubiquitin-proteasome pathway is essential for degradation of proteins regulating growth and cell cycle progression (21). Ubiquitin-activating enzyme (E1), ubiquitin-conjugating enzyme (E2), and ubiquitin ligase (E3) sequentially tag proteins for ubiquitination and proteasomal degradation. SCF E3 ubiquitin ligases are composed of Skp1, Cul1, Roc1, and F-box proteins, where F-box proteins are substrate recognizing subunits (22). ␤-Transducin repeat-containing F-box proteins (␤-TrCP) recognize substrates phosphorylated within the DSG(X) 2ϩn S destruction motifs. SCF ␤-TrCP E3 ligases ubiquitinate specifically phosphorylated substrates and play a pivotal role in the regulation of cell division and various signal transduction pathways, which in turn are essential for many aspects of tumorigenesis (reviewed in Ref. 23). Genetic data have suggested that Drosophila SLIMB protein (orthologue of mammalian ␤-TrCP) is involved in proteolytic processing of Ci155 to Ci75 (24). However, there is no biochemical evidence that SLIMB/␤-TrCP proteins are involved in ubiquitination and degradation of Ci/Gli transcription factors.
Stabilization of the transcription factor Gli2 has been suggested as a key event in the transduction of Hh signals. The potential role of Gli2 in the development of BCC has been well documented. Gli2 is overexpressed in the majority of human BCCs, and skin-targeted overexpression of Gli2 in transgenic mice leads to the development of multiple BCCs. There is growing evidence that the transcriptional regulation of some Hh target genes, including Gli1, E2F1, Bcl2, etc., is Gli2-dependent (25)(26)(27)(28). The promoter of one such gene, Bcl2, is regulated preferentially by Gli2 (27).
In this study, we have found that SCF ␤-TrCP E3 ubiquitin ligase is responsible for Gli2 degradation. ␤-TrCP2 directly binds wild type Gli2 and promotes its ubiquitination, which is inhibited by a single amino acid substitution in Gli2 putative binding site. We also found that full-length Gli2 protein is overexpressed in prostate cancer cell lines and primary tumors. The mechanisms of Gli2 stabilization are discussed.
Tissue Culture and Transfections-293T human embryo kidney cells and HeLa human cervical adenocarcinoma cells were purchased from ATCC. Cells were grown in Dulbecco's modified Eagle's medium in the presence of 10% fetal bovine serum and antibiotics at 37°C and 5% CO 2 . Transfections were performed using the calcium phosphate procedure or lipofection with Lipofectamine 2000 (Invitrogen).
Immunoprecipitation and immunoblotting procedures were performed as described elsewhere (34).
In Vivo Binding Assay-293T cells cotransfected with FLAGtagged Gli2 or Gli2 S662A and HA-tagged ␤-TrCP2 were lysed in RIPA lysis buffer. Interaction between the expressed proteins was assessed by immunoprecipitation with FLAG or HA antibodies and immunoblotting with HA or FLAG antibodies, respectively. The interaction between endogenous proteins in the protein lysates from HeLa cells was analyzed by immunoprecipitation with Gli2 antibody and immunoblotting with ␤-TrCP antibody.
In Vitro Binding Assay-Recombinant Gli2 proteins expressed in 293T cells were immunopurified with FLAG antibody and protein A/G-agarose beads, stringently washed with stripping buffer containing 20 mM Tris-HCl (pH 7.5), 1 M NaCl, 50 mM NaF, and 0.1% Nonidet P-40 and equilibrated with binding buffer (20 mM Tris-HCl, pH 7.5, 100 mM NaCl, 50 mM NaF, and 0.1% Nonidet P-40). For treatment with phosphatase , the beads were washed with the binding buffer without phosphatase inhibitors and incubated with the phosphatase for 1 h at 37°C followed by washes in stripping buffer and re-equilibration with binding buffer. FLAG-Gli2 proteins immobilized on the beads were incubated with in vitro translated, 35 S-labeled ␤-TrCP2 for 60 min at 4°C. The beads were extensively washed with binding buffer, and associated proteins were analyzed by SDS-PAGE and autoradiography. 35 S-Labeled ␤-TrCP2 was synthesized in vitro using TNT kit (Promega). Lysates of 293T cells transfected with FLAG Gli2 and Gli2 S662A were incubated with S 35 labeled ␤-TrCP2 for 3 h at 4°C. These lysates were immunoprecipitated with FLAG antibody, and associated proteins were analyzed by SDS-PAGE and autoradiography.
In Vivo Ubiquitination Assay-293T cells were cotransfected with HA-tagged ubiquitin, HA-tagged ␤-TrCP2 and FLAGtagged Gli2, or Gli2 S662A . Cells were lysed in RIPA lysis buffer and immunoprecipitated with FLAG antibody. Immunocomplexes were analyzed by SDS-PAGE and immunoblotting with HA antibody.
Degradation Assay-Pulse-chase analysis was carried out on 293T cells as described elsewhere (34). Briefly, cells were grown in 100-mm dishes and transfected with the indicated plasmids. Cells were starved in methionine-and cysteinefree Dulbecco's modified Eagle's medium followed by metabolically labeling with a [ 35 S]methionine/[ 35 S]cysteine mixture (PerkinElmer Life Sciences). Chase was performed in complete Dulbecco's modified Eagle's medium (10% fetal bovine serum) supplemented with 2 mM unlabeled methionine and cysteine, and cells were harvested at respective time points. Gli2 proteins were immunoprecipitated from RIPA lysates with FLAG antibody, separated by SDS-PAGE, and analyzed by autoradiography.
RNA Isolation and Real-time Reverse Transcription-PCR-Real-time reverse transcription-PCR for quantitative RNA measurements of GLI2 were done using SYBR Green PCR core reagents (Applied Biosystems) as described previously (41). Glyceraldehyde-3-phosphate dehydrogenase was used as a reference gene.

RESULTS AND DISCUSSION
SCF ␤-TrCP ubiquitin ligase recognizes DSG(X) 2ϩn S destruction motif to target proteins for ubiquitination and further degradation (reviewed in Ref. 23). Sequence analysis of Gli2 revealed the DSGV/MEMPGTG-PGS motif, which is conserved among various mammalian species (Fig. 1A). We analyzed whether the substrate-recognizing component of SCF ␤-TrCP ubiquitin ligase, F-box protein ␤-TrCP2, interacts with Gli2. We found that exogenously expressed Gli2 and ␤-TrCP2 (Fig.  1B) as well as endogenous proteins (Fig. 1C) interact in vivo in coimmunoprecipitation assay. Gli2 was also shown to bind in vitro translated ␤-TrCP2 protein (Fig. 1D). To determine whether the putative ␤-TrCP recognition motif is responsible for this interaction, we substituted potentially phosphorylated serine 662 to alanine in this motif of Gli2 (Fig. 1A). This single amino acid substitution is predicted to disrupt Gli2 interactions with ␤-TrCP. Indeed, in both assays, Gli2 S662A binding to ␤-TrCP2 was greatly diminished (Fig. 1, B and  D). Treatment of FLAG-Gli2 with protein phosphatase abolished the ability of Gli2 to bind ␤-TrCP2 in vitro (Fig. 1D). These results demonstrate that phosphorylation of Gli2 is necessary for its recognition by ␤-TrCP ubiquitin ligase receptor.
Interaction of ␤-TrCP2 with specific substrates results in ubiquitination of these proteins (reviewed in Ref. 23). Cotransfection of cells with ␤-TrCP2 construct accelerated the ubiquitination of wild type Gli2 (Fig. 1E, lane 3) as measured by in vivo ubiquitination assay. In contrast, the ubiquitination of Gli2 S662A  mutant that interacts poorly with ␤-TrCP2 was less efficient and was not affected by ␤-TrCP2 overexpression (Fig. 1E, lane  6). These data demonstrate that serine 662 is critical for the interaction between Gli2 and ␤-TrCP2 and that binding to ␤-TrCP is important for Gli2 ubiquitination.
To analyze the role of ␤-TrCP in the degradation of Gli2, we inhibited ␤-TrCP activity by either knocking down the expression of ␤-TrCP1 and ␤-TrCP2 proteins using shRNA or by expression of a dominant negative mutant of ␤-TrCP (␤-TrCP2 ⌬N ) (30). Our results show that inhibition of ␤-TrCP function leads to stabilization of Gli2 protein (Fig. 2, A-C). ␤-TrCP2 ⌬N extends the half-life of Gli2 from about 6 to 12 h. Interestingly, shRNA against ␤-TrCP2 appeared to be more effective in the inhibition of Gli2 turnover than in ␤-TrCP1-specific shRNA ( Fig. 2A). Importantly, inhibition of ␤-TrCP function resulted in stabilization (Fig. 2C) and accumulation (D) of endogenous Gli2 in 293T cells. These data suggest that ␤-TrCP is involved in the degradation of Gli2 protein in mammalian cells.
Although the kinase responsible for Gli2 phosphorylation within ␤-TrCP recognition motif is not known, phosphorylation of Drosophila Ci by shaggy (Drosophila homologue of GSK3␤) is demonstrated to be a necessary step in Ci proteolysis (35,36). Recently Gli2 has been shown to be phosphorylated by GSK3␤ (37). Treatment of cells with the GSK3 inhibitor LiCl substantially increased the half-life of Gli2 protein (Fig. 2E). These data suggest that GSK3 may be involved in phosphorylation-dependent degradation of Gli2.
To further confirm the role of ␤-TrCP in proteolysis of Gli2, we compared the half-life of Gli2 wt with that of Gli2 S662A mutant. Gli2 S662A poorly interacted with ␤-TrCP2 (Fig. 1, B and C) and was not ubiquitinated by ␤-TrCP2 (D). In comparison with Gli2 wt , Gli2 S662A mutant protein was more stable and exhibited a half-life of more than 9 h (Fig. 3A). Furthermore, overexpression of dominant negative mutant of ␤-TrCP (␤-TrCP2 ⌬N ) did not affect the stability of Gli2 S662A (Fig. 3B). These data demonstrate that serine 662 is critical for the interaction between Gli2 and ␤-TrCP2. Disruption of the DSG motif by mutation of the serine residue renders it poorly interactive with ␤-TrCP, hence stabilizing the Gli2 S662A mutant protein.
Stabilization of Gli2 S662A mutant translated into the higher level of protein expression as compared with the wild type Gli2 (Fig. 3C, inset). HeLa cells transfected with the same amount of appropriate plasmids expressed higher level of Gli2 S662A mutant protein as compared with the wild type protein. Fig. 3C demonstrates that Gli2 S662A mutant is significantly more effective than Gli2 wt in the activation of Gli-dependent transcriptional activity. 8 ϫ 3ЈGli BS-LucII, pGL3-Bcl2promo luciferase, or K17-driven luciferase were utilized to measure Gli2-dependent transcriptional activation driven by Gli2 wt or Gli2 S662A . Gli2 S662A is about twice as potent as Gli2 wt in the activation of transcription as measured by these three different reporter constructs. These results demonstrate that elevated Gli2-dependent transcriptional output is likely attributed to the higher levels of Gli2 S662A protein expression. These data strongly suggest that Gli2 protein turnover is an important step in the modulation of Gli2-dependent transcription. . Gli2 S662A is expressed in higher levels and is more potent in the activation of Gli-dependent transcription than Gli2 wt . A, the indicated FLAG-Gli2 proteins (wild type or S662A mutant) were expressed in 293T cells and their degradation analyzed as described in the legend to Fig. 2A. A representative of two independent experiments is shown. B, pulse-chase analysis of FLAG-Gli2 S662A protein expressed in 293T cells with or without dominant negative ␤-TrCP2 ⌬N mutant and analyzed as described in the legend to Fig. 2A. A representative of two independent experiments is shown. Inset shows the levels of HA-␤-TrCP2 ⌬N expression in 293T transfected cells analyzed by immunoblotting with HA antibody. C, HeLa cells were transfected with Gli-luciferase (8 ϫ 3ЈGli BS-LucII), pGL3-Bcl2promo luciferase, or K17 luciferase and respective Gli2 expression plasmids as indicated. Luciferase activity was estimated using luciferase reporter assay reagent (Promega). ␤-Galactosidase was used for normalization and estimated using ␤-galactosidase assay reagent (Pierce). *, p Ͻ 0.01 compared with cells transfected with empty vector (pcDNA3.1); **, p Ͻ 0.01 compared with cells transfected with Gli2 wt in Student's t test. Inset shows the levels of FLAG-Gli2 expression in HeLa cells transfected with FLAG-Gli2 wt or FLAG-Gli2 S662A analyzed by immunoblotting with Gli2 G20 antibody (protein loading was normalized by ␤-galactosidase activity).