Casein Kinase 2 Promotes Hedgehog Signaling by Regulating both Smoothened and Cubitus Interruptus*

Casein kinase 2 (CK2) is a typical serine/threonine kinase consisting of α and β subunits and has been implicated in many cellular and developmental processes. In this study, we demonstrate that CK2 is a positive regulator of the Hedgehog (Hh) signal transduction pathway. We found that inactivation of CK2 by CK2β RNAi enhances the loss-of-Hh wing phenotype induced by a dominant negative form of Smoothened (Smo). CK2β RNAi attenuates Hh-induced Smo accumulation and down-regulates Hh target gene expression, whereas increasing CK2 activity by coexpressing CK2α and CK2β increases Smo accumulation and induces ectopic Hh target gene expression. We identified the serine residues in Smo that can be phosphorylated by CK2 in vitro. Mutating these serine residues attenuates the ability of Smo to transduce high level Hh signaling activity in vivo. Furthermore, we found that CK2 plays an additional positive role downstream of Smo by regulating the stability of full-length Cubitus interruptus (Ci). CK2β RNAi promotes Ci degradation whereas coexpressing CK2α and CK2β increases the half-life of Ci. We showed that CK2 prevents Ci ubiquitination and degradation by the proteasome. Thus, CK2 promotes Hh signaling activity by regulating multiple pathway components.

The Hedgehog (Hh) 3 pathway is one of the major signaling pathways that control animal development. Hh signaling has also been implicated in stem cell maintenance and tissue regeneration and repair, and its malfunction causes several types of cancer (1,2). The Hh signal is transduced through a reception system at the plasma membrane that includes the receptor complexes Patched (Ptc)-Ihog and the signal transducer Smoothened (Smo) (3)(4)(5). Binding of Hh to Ptc-Ihog relieves the inhibition of Smo by Ptc, which allows Smo to activate the Cubitus interruptus (Ci)/Gli family of zinc finger transcription factors. In the past three decades, many Hh pathway components have been identified, including those that control sending, propagating, receiving, and transducing the Hh signal. Among these components, multiple kinases have been identified to play either positive, negative, or dual roles in transducing the Hh signal (6). In Drosophila, the absence of Hh allows the full-length Ci to undergo proteolytic processing to generate a truncated form, Ci75, which blocks the expression of Hhresponsive genes such as decapentaplegic (dpp) (7,8). Ci processing requires extensive phosphorylation by multiple kinases, including PKA, GSK3, and members of CK1 family (9), which promotes Ci binding to the SCF ubiquitin ligase containing the F-box protein Slimb (10,11). Efficient phosphorylation of Ci requires the kinesin-like protein Costal 2 that functions as a molecular scaffold to bring Ci and its kinases in close proximity (12). Hh signaling leads to the inhibition of Ci processing by dissociating the Costal 2-Ci kinase complex (12). The seven-transmembrane protein Smo belongs to the serpentine family of G protein-coupled receptors. Interestingly, Smo activation by Hh also requires phosphorylation by PKA and CK1 (13)(14)(15). Phosphorylation of Smo promotes its cell surface accumulation (14 -16) and induces dimerization of its C-terminal cytoplasmic tail (17). In addition, it has been shown that G protein-coupled receptor kinase 2 has a positive role in Hh signaling (18,19) by regulating Smo through kinase-dependent and kinase-independent mechanisms (20).
Casein kinase 2 (CK2) is a ubiquitously expressed and highly conserved Ser/Thr kinase participating in a wide variety of cellular processes and is frequently activated in human cancers. It is a stable tetrameric complex consisting of two catalytic ␣ subunits and two regulatory ␤ subunits. CK2 holoenzyme normally phosphorylates Ser/Thr at a consensus sequence of E/D/X-S/ T-D/E/X-E/D/X-E/D-E/D/X, and its ␤ subunits play a role in anchoring substrates (21). It has been shown that CK2 phosphorylates components in a variety of signaling pathways including Wnt, Akt, and NFB pathways. However, it is unknown whether CK2 is involved in Hh signal transduction.
It has been shown that that the C-terminal cytoplasmic tail of Smo has 26 Ser/Thr sites phosphorylated in response to Hh (15). Several phosphorylated sites in the Smo C-terminal cytoplasmic tail match the consensus of CK2 sites (see below); however, the link among CK2, Smo phosphorylation, and Hh signaling is lacking. In this study, we find that Smo is indeed phosphorylated by CK2 at multiple Ser residues in the Smo C-terminal cytoplasmic tail. Mutating these CK2 sites in Smo attenuates the ability of Smo to rescue the smo Ϫ/Ϫ phenotype fully. We also find that CK2 is required for Hh-induced Smo accumulation as knockdown of CK2␤ by RNAi blocked the accumulation Smo that is induced by Hh in wing discs. In contrast, overexpressing CK2␣ with CK2␤ enhances Smo phosphorylation and elevates Smo levels. We also find that CK2 has an additional positive role in Hh signaling downstream of Smo by regulating the stability of Ci. We further provide evidence that CK2 regulates Ci by preventing its ubiquitination and thus attenuating its degradation through the proteasome.
In Vitro Kinase Assay-For the in vitro kinase assay, GST-Smo and GST-Smo CK2SA fusion proteins containing the fragment of Smo amino acids 808 -899 were expressed in bacteria and purified with standard protocols. Ser-114 of GST was mutated to Ala to abolish background phosphorylation by CK2. All of the five CK2 sites in Smo (Ser-816, Ser-817, Ser-818, Ser-819, and Ser-843) were mutated into Ala in GST-Smo CK2SA . The purified fusion proteins were subjected to a kinase assay with commercial CK2 kinase (New England Biolabs) and [␥-32 P]ATP followed by autoradiography.
Immunostaining-Standard protocol was used for wing and eye imaginal disc immunostaining. Antibodies used in this study were   (14). Expressing Smo -PKA12 by C765-Gal4 resulted in a reproducible wing phenotype with partial fusion between Vein 3 and Vein 4 (arrow in Fig. 1D), a phenotype similar to that caused by weak fu mutations (30). We reasoned that mutating genes that regulate Smo activity may dominantly modify this phenotype. We found that expressing Smo -PKA12 in the background of CK2␣ heterozygotes, CK2␣ Tik / ϩ , enhanced the Smo -PKA12 phenotype (Fig. 1E), even though CK2␣ Tik / ϩ alone produced wild-type wings (not shown). Knockdown of CK2␤ by RNAi (CK2␤ RNAi) in Smo -PKA12 expressing wing discs resulted in small wings that consist of anterior-and posterior-most wing structures (Fig.  1F), a phenotype resembling that caused by severe loss of Hh signaling (31), although CK2␤ RNAi using the C765 driver in otherwise wild-type background did not exhibit loss-of-Hh phenotype (Fig. 1C). Of note, the ectopic veins between Vein 2 and 3 associated with C765-CK2␤RNAi could be due to a role of CK2 in the other pathway. Coexpressing FLAG-CK2␤ with CK2␤RNAi in Smo -PKA12 -expressing wing discs rescued the small wing phenotype (Fig. 1G), confirming that the loss-of-Hh wing phenotype induced by RNAi was due to the loss of CK2 activity. Furthermore, we found that Hh target gene expression was modified. For example, C765-Smo -PKA12 caused a reduction in ptc expression in A-compartment cells near the A/P boundary (Fig. 1K). Loss of one copy of CK2␣ further reduced whereas CK2␤RNAi nearly abolished ptc expression in C765-Smo -PKA12 background ( Fig. 1, L and M).

CK2 Is a Positive Regulator of Hh
To explore the role of CK2 in Hh signaling further, we employed CK2 RNAi and overexpression of CK2 in otherwise wild-type background to examine their effects on Hh signaling. We found that knockdown of CK2␤ by RNAi with ptc-Gal4 inhibited Hh signaling activity, as indicated by attenuated en expression in Hh-responding cells (arrows in Fig.  2B) and fused wing phenotype (arrow in Fig. 1I). Coexpressing FLAG-CK2␤ with CK2␤RNAi rescued en expression (Fig. 2C) and abolished fused wing phenotype (data not shown), indicating that the phenotype observed in Fig To investigate the consequence of elevated CK2 activity on Hh signaling in vivo, we coexpressed the catalytic ␣-subunit and the regulatory ␤-subunit of CK2 using the wing-specific Gal4 driver, MS1096, that expresses Gal4 in the whole wing pouch (with the dorsal compartment exhibiting stronger expression than the ventral compartment) (26). We found that expressing CK2␣ or CK2␤ alone had no effect on the levels of Ci (Fig. 2, H and I) and did not cause any change in Hh target gene expression (not shown). However, coexpressing CK2␣ with CK2␤ dramatically elevated the levels of Ci (Fig. 2L), highly accumulated Smo (Fig. 2LЈ), and ectopically induced dpp expression (Fig. 2LЉ). In contrast, expressing either CK2␣KM, a kinase-dead form of CK2␣ (32), or coexpressing CK2␣KM with CK2␤ did not affect Ci (Fig. 2, J and K) and Smo (data not shown) and had no effects on Hh target gene expression (data not shown), suggesting that the CK2 kinase activity is required for Smo and Ci accumulation and Hh signal transduction. In addition, we found that A-compart-    (Fig. 3B). To abolish background phosphorylation, we also mutated Ser-114 in GST to Ala, which is a putative CK2 site. As shown in Fig. 3C (Fig. 3D, lane 2, top panel) (25). In addition, we use the tubulin␣ promoter that drove gene expression at a level close to endogenous gene expression (34). We found that expressing Myc-Smo using the tubulin␣ promoter (VK5-tub-Myc-Smo) was able to fully rescue ptc and en expression in smo mutant cells (arrows in Fig. 3, H-IЉ), whereas the expression of VK5-tub-Myc-Smo CK2SA fully rescued ptc but partially rescued en expression (arrows in Fig. 3, J- CK2SD , in which SA mutation blocks phosphorylation and SD mutation mimics phosphorylation, were increased by the overexpression of CK2 and decreased by CK2␤RNAi (supplemental Fig. S1).

CK2 Has an Additional Positive Role Downstream of Smo-
We found that overexpression of CK2 caused Ci elevation (Fig.  2, L, MЈ, and N). One possibility is that Ci stabilization is due to ectopic activation of Smo by CK2. However, it is also possible that CK2 may play an additional role downstream of Smo to regulate other pathway component(s) of the Hh pathway such as Ci. To test this, we coexpressed CK2␣ with CK2␤ by MS1096 Gal4 in wing discs carrying smo mutant clones induced by FRT/ FLP-mediated mitotic recombination. We observed that CK2induced Ci elevation still persisted in smo mutant cells (arrow in Fig. 4A), indicating that CK2 acts downstream of Smo to induce Ci elevation. We also generated Myc-Smo SD123CK2SD with CK2 sites mutated to Asp (CK2SD) to mimic phosphorylation in the context of Smo SD123 in which the three PKA/CK1 phosphorylation clusters were mutated to Asp (14). VK5-Myc-Smo SD123CK2SD had constitutive signaling activity and induced ectopic ptc-lacZ and en expression (Fig. 4, B and BЈ). Knockdown of CK2␤ by RNAi attenuated the ectopic ptc-lacZ expression (arrow in Fig. 4C) and greatly reduced the ectopic en expression (arrow in Fig. 4CЈ). Similar results were achieved by using actϾCD2ϾGal4 to generate clones that express Myc-Smo SD123CK2SD with or without CK2␤RNAi (supplemental We then examined whether CK2 could regulate Ci activity. As shown in Fig. 4D, expression of Ci -3P , a mutant form of Ci in which three PKA sites were mutated to Ala to abolish PKA phosphorylation and thus Ci processing (26), induced high levels of ectopic ptc expression. Coexpression of CK2␤RNAi with Ci -3P severely reduced the ectopic ptc expression induced by Ci -3P (arrow in Fig. 4E), suggesting that CK2 positively regulates Ci -3P activity. Of note, the levels of HA-Ci -3P (indicated by HA staining) were significantly reduced by CK2␤RNAi, suggesting that CK2 promotes Ci activity by increasing its level.
CK2 Regulates Ci Turnover-One possibility that CK2 increases endogenous Ci levels is to enhance ci transcription. To test this, we monitored ci transcription using ci-lacZ, a ci enhancer trap line that expresses lacZ. We found that coexpressing CK2␣ and CK2␤ stabilized Ci without affecting ci-lacZ expression (arrow in Fig. 5A), suggesting that CK2 does not regulate ci transcription.
Another possibility is that CK2 might block Ci processing, thus leading to the accumulation of full-length Ci. However, in our in vivo function assay, in which HA-Ci was coexpressed with FLAG-CK2␣ and CK2␤ in the P-compartment wing discs  Of note, Ci, but not Ci-lacZ, was elevated by CK2. B, CK2 kinase activity required for Ci stabilization. Myc-Ci was transfected into S2 cells with either the indicated CK2 constructs or the treatment of CK2␤ dsRNA. Cell extracts were subjected to direct Western blotting (WB) with anti-Myc antibody to detect the levels of Myc-Ci, with anti-CK2␤ antibody to detect the exogenous CK2␤ expression and the endogenous CK2␤ level that indicates the efficiency of CK2␤ RNAi, with anti-FLAG antibody to detect the FLAG-tagged CK2␣, with anti-␤-tubulin antibody to detect ␤-tubulin that served as loading control. C, quantification of Myc-Ci relative levels. The level of Ci from cells transfecting Myc-Ci alone was set as 1. **, p Ͻ 0.01 (Student's t test). D, S2 cells cotransfected with Myc-Ci and GFP, or with Myc-Ci, GFP, and FLAG-CK2␣ plus FLAG-CK2␤, followed by treatment with cycloheximide for the indicated times. GFP expression served as transfection control. E, quantification of Ci levels from Western blot analysis performed in D. CHX, cycloheximide.
carrying smo mutant clones and hh-lacZ reporter gene, we found that P-compartment smo mutant cells coexpressing HA-Ci with CK2␣ and CK2␤ fully suppressed hh-lacZ expression, similar to the effect caused by expressing HA-Ci alone (supplemental Fig. S3), suggesting that CK2 does not block Ci processing. This is consistent with the finding that HA-Ci -3P was still regulated by CK2.
A third possibility is that CK2 increases Ci levels by regulating its turnover. To test this, we examined whether CK2 regulates the turnover of Ci in cultured S2 cells. We found that Myc-tagged full-length Ci (Myc-Ci) was stabilized by coexpression of FLAG-tagged CK2␣ and CK2␤ (Fig. 5B, lane 5, top  panel) whereas Myc-Ci levels remained nearly unchanged when coexpressing either FLAG-CK2␣ or CK2␤ alone (lanes 3  and 4, top panel). The ability of CK2 to induce Ci stabilization depended on its kinase activity, as coexpressing the kinase-dead FLAG-CK2␣KM with FLAG-tagged CK2␤ did not stabilize Ci (lane 6, top panel). In addition, CK2␤RNAi in cultured S2 cells reduced Myc-Ci levels (lane 2, top panel). Fig. 5C shows the quantification of these data. We further carried out a time course experiment to examine Myc-Ci half-life. S2 cells were transfected with Myc-Ci with or without FLAG-tagged CK2␣ and CK2␤, and Myc-Ci protein levels were monitored at different time points after treatment with the protein synthesis inhibitor, cycloheximide. We found that Myc-Ci was barely detected after 6 h of incubation with cycloheximide (Fig. 5D,  lane 8, top panel). In contrast, coexpression of CK2␣ with CK2␤ increased the half-life of Myc-Ci (Fig. 5D, E). Taken together, our data support the idea that CK2 regulates Ci activity by preventing its degradation.
CK2 Down-regulates Ci Ubiquitination and Prevents Its Degradation by Proteasome-We then determine whether CK2 regulates Ci stability through proteasome. Expressing CK2␤RNAi by ap-Gal4 destabilized Ci in dorsal compartment cells of wing discs (arrow in Fig. 6A). However, down-regulation of Ci by CK2␤RNAi was prevented by treatment with the proteasome-specific inhibitor MG132 (arrow in Fig. 6B), suggesting that Ci is degraded by proteasome when CK2 is inactivated. As proteasome degrades proteins after they are ubiquitinated, we asked whether CK2 regulates Ci ubiquitination. In S2 cells transfected with Myc-Ci, ubiquitinated Ci could be detectable by Western blot analysis with an anti-ubiquitin antibody following immunoprecipitation of Myc-Ci with a Myc antibody (Fig. 6C, lane 2 6D). These data indicate that CK2 regulates Ci stability through the ubiquitin/proteasome-mediated pathway.
It has been reported that multiple E3 ligases regulate Ci stability via the proteasome (35). As CK2 does not regulate Ci processing (supplemental Fig. S3) and Ci -3P , which is no longer regulated by Slimb, is still regulated by CK2 (Fig. 4, E and EЈ), it is unlikely that CK2 regulates Ci via SCF Slimb/␤-Trcp . It has also been shown that other E3 ligases, such as HIB (27,36), Hyd (37), and Debra (38), are involved in Ci degradation under different circumstances. However, we found that the stability of Cim1-6, a mutant form of Ci that could bypass its regulation by HIB (29), was up-regulated by overexpressing CK2 and down-regulated by CK2␤RNAi (Fig. 6, E and F). Furthermore, activation of CK2 did not significantly stabilize Ci in eye discs posterior to the morphogenetic furrow where Ci is degraded by HIB (supplemental Fig. S4) (27,36). These observations suggest that CK2 may regulate Ci in a manner independent of HIB. RT-PCR analysis indicated that Debra, HIB, and Hyd are expressed in cultured S2 cells (supplemental Fig. S5C). To determine whether CK2 could stabilize Ci by attenuating Ci ubiquitination by the above E3s, Myc-Ci was transfected into S2 cells treated with Debra RNAi, HIB RNAi, or Hyd RNAi, and with or without CK2␤RNAi. RNAi efficiencies were examined by Western blotting (supplemental Fig. S5, A and B). Consistent with previous findings (27,38), RNAi of HIB and Debra down-regulated Ci ubiquitination (Fig. 6G,  CK2 Regulates Gli Proteins-To determine whether CK2 plays similar roles in regulating Gli proteins, we expressed Myc-Gli1 or Gli2 in wing discs expressing either CK2␤RNAi or FLAG-CK2␣ and CK2␤. Wing imaginal discs were collected and subjected to direct Western blot analysis with anti-Myc antibody to examine the levels of Gli proteins. We found that expressing FLAG-CK2␣ and CK2␤ enhanced the levels of Gli1 and Gli2 (Fig. 7A, lane 2 and 5, compared with lane 1 and 4, respectively) whereas CK2␤ RNAi decreased both Gli1 and Gli2 (lane 3 and 6, compared with lane 1 and 4, respectively). These data are consistent with the finding of CK2 stabilizing Ci. Taken together, our observations suggest that CK2 may play a conserved role in regulating Ci/Gli stability.

DISCUSSION
The regulation of Smo and Ci/Gli is a critical event in Hh signal transduction, and Smo and Ci/Gli are two major targets for developing diagnostic and therapeutic treatments of cancers (2). A comprehensive understanding of Hh signaling at the levels of Smo and Ci/Gli will undoubtedly shed light into the mechanism of Hh in cancer progression and into identification of potential drugs for therapeutic intervention. CK2 has long been shown, in most cases, to act as a positive regulator in different signaling pathways and tumorigenesis. For example, CK2 phosphorylates many transcription factors, proto-oncoproteins, and tumor suppressor proteins, which regulates the access of these proteins to the proteasome (39). However, it is unknown whether CK2 is also involved in Hh signal transduction. In this study, we have identified CK2 as a novel component in Hh signaling. By using both the Drosophila wing imaginal disc and the cultured cells, we provide the first genetic and biochemical evidence for a physiological function of CK2 in Hh  2D). Consistently, CK2␤RNAi in cultured S2 cells severely down-regulated Smo levels (Fig. 3E). Coexpressing CK2␣ with CK2␤ accumulates Smo in both A-and P-compartment cells (Fig. 2, LЈ and NЈ). However, mutating CK2 sites in Smo did not affect its cell surface accumulation in S2 cells (data not shown), and the levels of Smo mutant forms lacking CK2 phosphorylation or mimicking CK2 phosphorylation were still regulated by CK2 (supplemental Fig. S1), suggesting that CK2 may regulate Smo accumulation indirectly through other mechanism(s).
We provide evidence that CK2 regulates Ci stability through the ubiquitination/proteasome pathway; however, the precise mechanism awaits further investigation. It is possible that CK2 regulates Ci stability by directly phosphorylating Ci as sequence analysis identified 14 Ser/Thr residues that confirm CK2 phosphorylation consensus sites (supplemental Fig. S6). However, simultaneously mutating four optimal sites (supplemental Fig. S6) neither affected Ci stability in S2 cells nor changed Ci activity in wing discs (data not shown). FIGURE 6. CK2 down-regulates Ci ubiquitination and prevents the proteasome-mediated Ci degradation. A-BЈ, wing discs expressing UAS-CK2␤RNAi by ap-Gal4 were treated with or without MG132 and immunostained for Ci. The treatment of MG132 restored Ci that was down-regulated by CK2␤ RNAi. GFP marks the RNAi cells. C, CK2 down-regulates Ci ubiquitination. S2 cells were transfected with Myc-Ci and incubated with CK2␤ dsRNA, or TBB, or cotransfected with FLAG-CK2␣ and FLAG-CK2␤, followed by the treatment with or without MG132. Cell extracts were immunoprecipitated (IP) with anti-Myc antibody and blotted (WB) with anti-Myc antibody to determine the levels of Ci, or blotted with anti-ubiquitin antibody to examine the Ci-bound ubiquitin. IgG served as loading control. D, quantification analysis shows the ratio of ubiquitinated Ci to total Ci in C. Myc-Ci in lane 1 of C was set as 1. *, p Ͻ 0.05 (Student's t test). E, Cim1-6, with HIB-interacting sites mutated, is still regulated by CK2. S2 cells were cotransfected with HA-Cim1-6 and GFP, with either CK2␣ϩCK2␤ or CK2␤ dsRNA treatment. Cell lysates were subjected to direct Western blotting with anti-HA antibody. F, quantification of HA-Cim1-6 relative levels from E is shown. The level of Cim1-6 from cells transfecting HA-Cim1-6 alone was set as 1. *, p Ͻ 0.05 (Student's t test). G, knockdown of the known E3s does not affect Ci ubiquitination that is induced by CK2 inactivation. S2 cells were transfected with Myc-Ci and treated with HIB dsRNA, Debra dsRNA, or Hyd dsRNA, with or without CK2␤ dsRNA. Cell extracts were immunoprecipitated with anti-Myc antibody and blotted with anti-Myc or anti-ubiquitin antibodies. IgG served as loading control.
CK2 could phosphorylate Ci at other sites. It is also possible that CK2 might regulate component(s) in the ubiquitination/proteasome pathway and thus prevent Ci/Gli degradation, which may account for the ability of CK2 to regulate the stability of many proteins. We found that Ci ubiquitination was up-regulated when CK2 was inactivated. However, knockdown of the known E3 ligases that regulate Ci did not affect Ci ubiquitination induced by CK2 inactivation (Fig.  6G). In addition, the misexpression of CK2 does not phenocopy loss of HIB. For example, overexpressing CK2␣ with CK2␤ did not dramatically up-regulate the levels of Ci in posterior cells of the eye discs where HIB plays a major role in degrading Ci in the presence of Hh (supplemental Fig. S4). These data suggest that CK2 may regulate Ci ubiquitination and stability through a novel E3. The identification of such E3 awaits further investigation in the future.