Casein Kinase 2 Is Activated and Essential for Wnt/β-Catenin Signaling*

Wnt/β-catenin signaling is essential to early development. Activation of Frizzled-1 by Wnts induces nuclear accumulation of β-catenin and activation of Lef/Tcf-dependent gene expression. Casein kinase 2 has been shown to affect Wnt/β-catenin signaling. How casein kinase 2 exerts an influence in Wnt signaling is not clear; casein kinase 2 has been reported to be constitutively active (i.e. not regulated). Herein we show to the contrary that casein kinase 2 activity is rapidly and transiently increased in response to Wnt3a stimulation and is essential for Wnt/β-catenin signaling. Chemical inhibition of casein kinase 2 or suppression of its expression blocks Frizzled-1 activation of Lef/Tcf-sensitive gene expression. Treatment with pertussis toxin or knock down of Gαq or Gαo blocks Wnt stimulation of casein kinase 2 activation, as does suppression of the phosphoprotein Dishevelled, demonstrating that casein kinase 2 is downstream of heterotrimeric G proteins and Dishevelled. Expression of a constitutively active mutant of either Gαq or Gαo stimulates casein kinase 2 activation and Lef/Tcf-sensitive gene expression. Thus, casein kinase 2 is shown to be regulated by Wnt3a and essential to stimulation of the Frizzled-1/β-catenin/Lef-Tcf pathway.

Wnt signaling regulates essential aspects of early development, including cell fate determination, embryonic patterning, and cell proliferation (1,2). Wnt ligands are palmitoylated glycoproteins (3) that are secreted and bind to cell surface heptihelical receptors termed "Frizzleds" (4,5). Frizzleds are members of the superfamily of G proteincoupled receptors (6,7). For the "canonical" Wnt pathway in mammalian cells, Wnt3a binds to Frizzled-1, activating heterotrimeric G proteins (e.g. G o and G q ) and the phosphoprotein Dishevelled (Dvl) (6). In Drosophila, activation of the canonical pathway or the planar cell polarity pathway likewise requires Wnt activation of Frizzled-1, activation of G o , and downstream Dishevelled (Dsh) (8). Activation of Dsh/ Dvl inhibits glycogen synthase kinase 3␤ and its ability to phosphorylate and thereby destabilize ␤-catenin. The Wnt3a-stimulated elevation of intracellular ␤-catenin leads to nuclear accumulation of ␤-catenin and to activation of Lef/Tcf-dependent gene expression (9).
Reported as "constitutively active" in cells, CK2 has been suggested to play an important, but passive, role in signaling (21). The ability of overexpression or inhibition of CK2 to affect aspects of early development suggests a pivotal role of CK2 in Wnt/␤-catenin/Lef-Tcf pathway (20) but provides no insight as to whether the effects of CK2 are the result of "activation." Our hypothesis is that the role of CK2 is not "passive" and that of a constitutively active protein kinase but rather that of a kinase regulated by upstream signaling. In support of this hypothesis are the observations that CK2 activity may be sensitive to phosphatidylinositol turnover (22) and to phosphorylation by non-receptor tyrosine kinases (23). These data piqued our interest in investigating whether or not CK2 was activated by Wnt stimulation. The results show that CK2 is activated by stimulation of Wnt3a operating via Frizzled-1, heterotrimeric G proteins, and Dishevelled-2 in the canonical Wnt/␤catenin/Lef-Tcf pathway of mammalian cells.
In Vitro CK2 Kinase Activity Assay-CK2 activity was analyzed as previously described (26). Briefly, immunoprecipitated CK2␣ or CK2␣Ј from whole-cell lysates were washed with kinase assay buffer (20 mM MOPS, pH 7.2, 25 mM ␤-glycerol phosphate, 5 mM EGTA, 1 mM sodium orthovanadate, 1 mM dithiothreitol, and 100 nM okadaic acid) and then incubated with 120 M substrate peptide (RRRDDDSDDD; Upstate Biotechnology) in kinase assay buffer containing 10 Ci of [␥-32 P]GTP, 160 M GTP, and 25 mM MgCl 2 at 30°C for 10 min. The reaction was stopped by addition of 25 l of 40% trichloroacetic acid. Samples were spotted onto a P81 Whatman filter membrane. Air-dried membranes were washed with 0.75% H 3 PO 4 three times to eliminate any unincorporated [␥-32 P]GTP. Incorporated ␥-32 P in substrates were measured by liquid scintillation spectrometry. Samples were assayed in triplicate.
␤-Catenin Stabilization Assay-F9 cells stably transfected with either Rfz1 receptor or ␤ 2 AR/Rfz1 chimeric receptor were treated with Wnt3a (10 ng/ml) or isoproterenol (10 M), respectively, for the time periods indicated. To determine cytoplasmic levels of ␤-catenin, cells were lysed in radioimmune precipitation buffer supplemented with protease inhibitors (20 mM Tris, pH 7.4, 150 mM NaCl, 1% Triton X-100, 5 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml of aprotinin, 10 g/ml of leupeptin). Whole-cell lysate was incubated with ConA-Sepharose (Amersham Biosciences) for 1 h and centrifuged to remove membraneassociated ␤-catenin. The supernatants were subjected to SDS-PAGE, and the separated proteins were transferred onto a nitrocellulose membrane. For immunoblotting the membrane was blocked with 10% bovine serum albumin for overnight and probed with anti-␤-catenin antibody. The presence of ␤-catenin was detected by use of the enhanced chemiluminescence method.
Lef/Tcf-sensitive Reporter Gene Assay-F9 clones stably transfected with pTOPFLASH and co-transfected with either Rfz1 or ␤ 2 AR/Rfz1 chimera were stimulated with Wnt3a (10 ng/ml) or isoproterenol (10 M), respectively, for the indicated periods. Cell lysates were collected in reporter lysis buffer (Promega, Madison, WI) and subjected to luciferase activity assay. Briefly, 10 l of cell lysates were incubated for 10 s with 100 l of reaction mixture containing 0.67 mM luciferin, 0.27 mM coenzyme A, 0.1 mM EDTA, 1.1 mM MgCO 3 , 4 mM MgSO 4 , and 20 mM Tricine, pH 7.8, and the intensity of luminescence was immediately measured using a luminometer (Lumat LB 9507, Berthold Technologies, Oak Ridge, TN). Samples were assayed in triplicate, and the luciferase activity was normalized based on protein concentration. For experiments needing transient transfection, F9 cells were co-transfected with 0.2 g of pcDNA3 plasmids harboring the cDNAs encoding G␣ q , G␣ o , G␣ 11 , or TAP-tagged Dvl2 and 0.15 g of reporter plasmid Super8XTOPFLASH or mutated reporter plasmid Super8X FOP-FLASH (27) for 24 h using Lipofectamine Plus reagent (Invitrogen) according to the manufacturer's protocol. Cells were harvested in reporter lysis buffer following exposure to serum-free Dulbecco's modified Eagle's medium for 8 h. Samples were assayed in triplicate, and luciferase activity was measured as described above.
Knock Down of Dvl 2 and CK2 Subunits by siRNA-At least 3 pairs of sense and antisense oligonucleotides were designed and synthesized by Ambion (Austin, TX) for each of the following molecules: mouse Dishevelled 2 (NM_007888) and mouse CK2 subunit ␣ (NM_007788), ␣Ј (NM_009974), and ␤ (NM_009975). Each pair of oligos was annealed in 20 l of annealing buffer (100 mM potassium acetate, 2 mM magnesium acetate, and 30 mM HEPES-KOH, pH 7.4) at 37°C for 1 h. F9 cells were treated with 100 nM annealed siRNA by using SiLentFect (BD Biosciences) according to the manufacturer's protocol. Briefly, siRNA were incubated with 4 l of SiLentFect for 20 min in 50 l of serum-free Dulbecco's modified Eagle's medium, and then the mixture was added into 1 ml of growth medium in a 12-well plate in which F9 cells had been grown to 80% confluency. After siRNA treatment for 48 h, cells were lysed and subjected to Lef/Tcf reporter gene assay. The efficiency of knocking down CK2␣, CK2␣Ј, CK2␤, and Dvl2 was detected by Western blot with antibody against each individual protein. The sequences of siRNA oligos, which are the most effective in knock down of targeted proteins, are listed as follows. The siRNA oligos UGGAAUAUUUCAUG-GACtt and GUCCAUGAAAUAUUCCACCtg are for targeting CK2␣. The siRNA oligos GGGUGGUUGUAAAAAUUCUtt and AGAAUUUU-UACAACCACCCtc are for targeting CK2␣Ј. The siRNA oligos GGAAG-UCUAUGGUUUUAGtt and CUAAAAACCAUAGACUUCCtg are for targeting CK2␤. The siRNA oligos GGAAGAGAUCUCCGAUGACtt and GUCAUCGGAGAUCUCUUCCtt are for targeting Dvl2. After treatment of specific CK2 siRNA for 48 h, cells stably expressing Rfz1 and TOPFLASH were stimulated with Wnt3a for 5-6 h. Cell lysates were collected, and the Lef/Tcf-sensitive luciferase reporter gene activity was determined.
Knock Down of G␣ Subunits-To knock down G␣ subunits, F9 clones expressing ␤ 2 AR/Rfz1 receptor were treated with phosphorothioate oligodeoxynucleotide (S-ODN, cell culture grade, high pressure liquid chromatography purified; Operon Technologies, Huntsville, AL) antisense to specific G protein ␣ subunits for 72 h as previously described (28).

CK2 Is Activated by Stimulation of the Wnt/␤-Catenin/Lef-Tcf
Pathway-Mouse F9 embryonic totipotent teratocarcinoma cells were stably transfected with the ␤ 2 AR/Rfz1 chimeric receptor, a chimera that can activate the Wnt/␤-catenin/Lef/Tcf canonical pathway by activation of the receptor with the ␤-adrenergic agonist ISO (10 M). The ability to quantify chimeric receptor expression and to activate the Frizzled-1 downstream pathway was the basis for selecting these clones for the initial examination of possible CK2 activation. Treating the F9 ␤ 2 AR/Rfz1 stable transfectants with isoproterenol provoked a sharp stimulation of Lef/Tcf-dependent transcription (Fig. 1A). The activation of Lef/Tcf-dependent transcription peaks at 5-6 h poststimulation of the Frizzled-1 pathway. Activation of Frizzled-1 likewise was found to stimulate CK2 activity (Fig. 1B). This stimulation by Frizzled-1 provoked an increase in CK2 activity within 15 min of treatment with ␤-agonist. CK2 activity remained elevated (ϳ50%) for 30 min and thereafter declined to baseline values within 60 min of Frizzled-1 stimulation. Immunoblotting of CK2 displays no corresponding changes in the cellular level of this protein kinase (Fig. 1B, lower panel).
Wild-type F9 cells and stably transfected F9 clones expressing either the ␤ 2 AR/Rfz1 (␤ 2 /Fz1) or the ␤ 2 AR/Rfz2 (␤ 2 /Fz2) chimeric receptors were stimulated for 30 min and the CK2 activity then measured (Fig.  1C). Only those clones expressing the ␤ 2 AR/Rfz1 chimera displayed activation of CK2 in response to stimulation of the Frizzled-1 pathway. Stimulation of the wild-type F9 cells, in contrast, failed to stimulate CK2 activity. Stimulation of the Frizzled-2 pathway that links to changes in the intracellular levels of Ca 2ϩ and cyclic GMP (29) likewise failed to stimulate CK2 activity.
The observations derived from the study of F9 clones expressing the ␤ 2 AR/Rfz1 chimera were tested further using F9 clones expressing bona fide rat Frizzled-1 (Fz1) and treated with purified, active Wnt3a (Fig.   1D). Treating the clones expressing authentic Fz1 with purified Wnt3a provokes a rapid increase in CK2 activity, ϳ60% increase within the first 30 min. The activation of CK2 activity in response to Wnt3a appears to peak at 15-30 min poststimulation, declining to near basal levels within 60 min. The ability of Wnt3a to stimulate CK2 activity was also measured in human embryonic kidney 293 and mouse NIH3T3 cells transiently transfected to express Fz1; both cell lines responded as do the F9 cells (Fig. 1, E and F). Thus, activation of Frizzled-1 with Wnt3a stimulates a rapid but transient increase in the activity of CK2, confirming and extending the data obtained in clones expressing the ␤ 2 AR/Rfz1 chimera (compare Fig. 1, B and D).
CK2 Inhibitors Block Wnt/␤-Catenin/Lef-Tcf Signaling-To test the linkage between Wnt activation of the Wnt/␤-catenin/Lef-Tcf signaling and the role of CK2, we explored the effects of chemical inhibitors of CK2 on the Lef/Tcf-dependent transcriptional response to stimulation either of the ␤ 2 AR/Rfz1 chimera or of Frizzled-1 (Fig. 2). Apigenin (4Ј,5,7-trihydroxyflavone) inhibits CK2 activity (30). Treating F9 clones with apigenin (20 M) leads to a dramatic inhibition of Lef/Tcf-dependent transcription in response to activation of the ␤ 2 AR/Rfz1 chimera by ␤-agonist ( Fig. 2A). Measured over 7 h of activation of the ␤ 2 AR/Rfz1 chimera, Lef/Tcf-dependent transcription remained at essentially base line in the presence of this CK2 inhibitor. The inhibition of ␤ 2 AR/Rfz1 chimera-stimulated Lef/Tcf-dependent transcription by apigenin was dose dependent, with 10 M concentrations of this flavone derivative capable of inhibiting the response by ϳ50% (Fig. 2B). The Frizzled-1 chimera-mediated stimulation of the Lef/Tcf-dependent transcription was nearly abolished by 15-20 M apigenin (Fig. 2B).
In F9 clones stably expressing the rat Frizzled-1, apigenin was found to suppress the ability of Wnt3a to activate Lef/Tcf-dependent transcription (Fig. 2C). Similar experiments with 2-dimethylamino-4,5,6,7-tetrabromo-1H-benzimidazole (DMAT), a highly selective and potent CK2 inhibitor (31), displayed the same ability to block Frizzled-1-dependent activation of the Lef/Tcf-dependent transcription (Fig. 2D). Chemical inhibitors for other signaling kinases (e.g. protein kinase C and protein kinase A) were without effect (data not shown), suggesting a critical role of CK2 in the Wnt/␤-catenin/ Lef-Tcf response. We tested further the effects of inhibition of CK2 activity on Wnt3a-stimulated ␤-catenin stabilization in F9 clones expressing rat Frizzled-1 (Fig. 2E). F9 clones treated with Wnt3a displayed a time-dependent stabilization and accumulation of intracellular ␤-catenin. Wnt3a stimulated a 10-fold increase in free ␤-catenin levels within 4 h of stimulation. The Wnt3a-stimulated accumulation of ␤-catenin was markedly attenuated in the clones treated with apigenin. Quantification of the levels of ␤-catenin in response to Wnt3a stimulation in the absence and presence of apigenin confirms the data obtained using Lef/Tcf-dependent transcription as a read-out (Fig. 2, compare panels A and F). These results suggest that CK2 acts upstream of ␤-catenin to activate Lef/ Tcf-dependent transcription.
Suppression of CK2␣ Subunits Attenuates Activation of Wnt/␤-Catenin/Lef-Tcf Signaling-The CK2 holoenzyme is composed of two catalytic subunits (␣␣, ␣␣Ј, or ␣Ј␣ Ј) paired with two non-catalytic ␤ subunits. CK2 catalytic subunits as well as their holoenzymes are enzymatically active. The functions of ␣ and ␣Ј are not completely redundant, because CK2␣Ј null mice are sterile and display a defect in spermatogenesis (35). In view of the ability of chemical inhibitors of CK2 to block Wnt3a signaling via the canonical pathway, we investigated the effects of suppressing expression of CK2 subunits. We adopted a knockdown (KD) strategy that makes use of unique siR-NAs to suppress expression of individual subunits. Cells were treated FIGURE 1. CK2 is activated by stimulation of the Wnt/␤-catenin/Lef-Tcf pathway. A, stimulation of ␤ 2 AR/Rfz1 chimeric receptor with isoproterenol (ISO) activates Lef/Tcfsensitive luciferase reporter gene. Confluent F9 clones stably expressing the ␤ 2 AR/Rfz1 chimera and Lef/Tcf-sensitive luciferase reporter gene or reporter gene alone were treated with ISO (10 M) for 7 h. Relative luciferase activity at each time point was compared and presented as -fold of transcription measured as described. B, activation of ␤ 2 AR/Rfz1 receptor with ISO increases CK2 activity. F9 clones stably expressing the ␤ 2 AR/Rfz1 chimera were treated with ISO for the indicated times, and the CK2␣ was immunoprecipitated and its activity was measured as described. CK2 activities were compared and presented as -fold by setting the CK2 activity in control samples as 1 (upper panel). To compare the relative amount of CK2 in each sample, immunoprecipitated CK2␣ was probed with antibody against CK2␣ (bottom panel). C, increase of CK2 activity responding to ISO in ␤ 2 AR/Rfz1-expressing cells but not in wild-type or ␤ 2 AR/ Rfz2-expressing cells. CK2 activity of wild-type F9 cells (WT), F9 clones expressing ␤ 2 AR/ Rfz1 (␤ 2 /Fz1), or clones expressing ␤ 2 AR/Rfz2 (␤ 2 /Fz2) with or without ISO treatment for 30 min prior to lysis was determined as described. D, Wnt3a stimulation increased CK2 activity. F9 clones constitutively expressing rat Frizzled-1 were stimulated with Wnt3a for the indicated periods and then lysed and their CK2 activity determined. E and F, Wnt3a stimulates CK2 activity in human embryonic kidney (HEK) 293 cells (E) and NIH3T3 cells (F) transiently transfected to express rat Frizzled-1(Fz1). Cells were transiently transfected for 24 h with an expression vector harboring the rat Frizzled-1 cDNA. Cells were then stimulated for 30 min with purified Wnt3a. At the end of the incubation, the cells were assayed for CK2 activity. All experiments were performed in triplicate and repeated at least three times.
with siRNA targeting specific subunits CK2␣, CK2␣Ј, and CK2␤, and Wnt3a-stimulated Lef/Tcf-sensitive transcription was assayed (Fig.  3, Table 1). The siRNA designed for each subunit was found to be effective in suppressing the expression of the targeted subunit (Fig.  3B, Table 1). KD produced by siRNA treatment was effective as determined by immunoblotting.
Treating cells with siRNA targeting either CK2␣ or CK2␣Ј attenuates Wnt3a-stimulated Lef/Tcf-sensitive transcription, much like the use of . The activity of reporter luciferase was monitored as described under "Experimental Procedures." B, ␤ 2 AR/Rfz1-expressing F9 clones were treated with no or increasing concentrations of the CK2 inhibitor apigenin (0 -20 M) for 20 min prior to the stimulation with the ␤-agonist isoproterenol (10 M). Lef/Tcf-sensitive transcription was determined as described under "Experimental Procedures." C, two separate clones of F9 cells stably expressing rat Frizzled-1 were pretreated with 20 M CK2 inhibitor apigenin for 20 min and then stimulated by Wnt3a (10 ng/ml). For each clone, Lef/Tcf-sensitive luciferase reporter activity was analyzed after Wnt3a stimulation for 6 h. D, confluent F9 clones expressing either ␤ 2 AR/Rfz1 (␤2/Fz1) or authentic rat Frizzled-1 (Fz1) receptors were treated with or without 10 M of the CK2 inhibitor DMAT for 20 min prior to the addition of either ISO (for chimera-expressing clones) or Wnt3a (for Fz1-expressing clones), respectively. The luciferase activity of the reporter gene construct was determined. E, the rat Frizzled 1-expressing cells were treated with 20 M of the selective CK2 inhibitor apigenin for 20 min (ϩapigenin). Clones not treated with apigenin were employed as Controls (Ϫ). After stimulation with Wnt3a for the indicated times, cells were lysed and cytoplasmic ␤-catenin accumulation was determined as described under "Experimental Procedures." A representative image is presented. F, the results from panel E were quantified and are presented as the mean values Ϯ S.E.  selective chemical inhibitors of CK2. Treating cells with siRNA targeting CK2␤, in contrast, had no effect on the ability of Wnt3a to activate Lef/Tcf-sensitive transcription. These data demonstrate that CK2␣ subunits are essential in the signaling from Wnt3a to the canonical pathway as measured by activation of Lef/Tcf-sensitive gene expression by Wnt3a. Heterotrimeric G Proteins Mediate Wnt Activation of CK2-We examined signaling downstream of Frizzled-1 in the activation of CK2. Frizzled-1 signaling has been shown previously to require activation of heterotrimeric G proteins in mammalian cells, Xenopus, and zebrafish embryos, as well as in Drosophila (32). Both G␣ q and G␣ o are required for stabilization of ␤-catenin as well as activation of Lef/Tcfdependent transcription in canonical Wnt/␤-catenin signaling pathway (6). Pertussis toxin catalyzes ADP ribosylation on the G o / i family of G proteins and inhibits their functions. To test whether G protein mediated CK2 activation in Wnt/Frizzled-1 signaling, we pretreated F9 cells with pertussis toxin. Pertussis toxin completely blocked CK2 activation in ␤ 2 AR/Rfz1-expressing cells stimulated by ␤-agonist (Fig. 4A). The result indicated that activation of one or more members of the pertussis toxin-sensitive G i family (G␣ i1-3 , G␣ o , G␣ t1,2 ) of G proteins was necessary for CK2 activation.
To test directly which G protein mediates CK2 activation in Wnt/ Frizzled-1 signaling, we used antisense oligodeoxynucleotides (ODN) to knock down specific G␣ subunit expression in F9 clones (6). F9 clones expressing ␤ 2 AR/Rfz1 chimera were treated with antisense ODN targeting G␣ q , G␣ o , or G␣ i2 for 72 h. Following the KD of individual G␣ subunits, the Frizzled-1 chimera-expressing clones were stimulated with ␤-agonist for 15 min and CK2 activity assayed. Knock down of either G␣ q or G␣ o blocks the activation of CK2 following stimulation of the Frizzled-1 pathway (Fig. 4A). KD of G␣ i2 , in contrast, did not affect the ability of the Frizzled-1 chimera to activate CK2. We tested the ability of KD of G␣ q , G␣ o , and G␣ i2 to impact Frizzled-1 stimulation of Lef/Tcf-dependent transcription in these same cells (Fig. 4B). KD of G␣ q or G␣ o , but not G␣ i2 , blocks not only CK2 activation but also the Lef/Tcf reporter gene response to stimulation (Fig. 4B). Immunoblot analysis of the G protein subunits demonstrates that antisense ODNs targeting specific G protein subunits were effective in selective suppression of the expression of G␣ q , G␣ o , and G␣ i2 (Fig. 4C). These results show that G␣ q and G␣ o are required for CK2 activation and Lef/Tcfdependent transcription in response to Wnt.
To further probe the role of G␣ q and G␣ o in CK2 activation, constitutively active mutant forms of G proteins were expressed in the F9 clones. Q209LG␣ q (QLG␣ q ), Q205LG␣ o (QLG␣ o ), and Q209LG␣ 11 (QLG␣ 11 ) mutants are deficient in the GTPase activity intrinsic to ␣ subunits of heterotrimeric G proteins that is responsible for "turning off " G␣ subunit activation. The G protein mutant versions were expressed transiently in F9 cells, and the CK2 activity was measured (Fig. 4D). Expression of Q209LG␣ q stimulates constitutive activation of CK2 in the absence of Frizzled-1 activation (Fig. 4D). Expression of Q205LG␣ o also stimulates constitutive activation of CK2. Expression of constitutively active Q209LG␣ 11 , in sharp contrast to expression of the G␣ q and G␣ o mutants, had no such effect on CK2 activity. The expression of mutant G␣ subunits, which contain internal glutamate-glutamate epitope (EE tagged), was essentially equivalent as determined by immunoblotting of whole-cell lysates by using antibody against EE tag (data not shown). Expression of Q209LG␣ q or Q205LG␣ o , but not Q209LG␣ 11 , constitutively activated not only CK2 but also Lef/Tcf-dependent transcription as determined with the luciferase reporter gene (Fig. 4E). The constitutive activation of Lef/Tcf-dependent transcription observed in response to expression of either the constitutively active Q209LG␣ q or Q205LG␣ o mutant was abolished by treatment with the CK2 inhibitor apigenin (Fig. 4E). Constitutive activation of CK2 itself, observed in response to expression of the constitutively active Q209LG␣ q or Q205LG␣ o mutants, also was blocked by treatment with apigenin (data not shown). These results demonstrate that Wnt activation of CK2 is mediated by G␣ q and G␣ o .
Role of Dvl-2 in CK2-dependent Activation of Wnt/␤-Catenin Signaling-Dvl is a downstream element of the Wnt3a/Frizzled/G protein pathway to activation of Lef/Tcf-dependent transcription (8). As observed earlier (14, 37), Dvl-2 is phosphorylated in F9 cells stimulated with Wnt3a (Fig. 5A). Treating cells with the CK2 inhibitor DMAT . Wnt activation of CK2 is sensitive to pertussis toxin and to knock down of G␣ q and G␣ o . A, F9 clones expressing ␤ 2 AR/Rfz1 receptor were treated with either pertussis toxin (100 ng/ml) for 1 h or antisense oligos against G␣ q , G␣ o , or G␣ i2 for 72 h prior to activation by treatment with the ␤-agonist isoproterenol (ISO) (10 M) for 30 min. Cells were harvested, and CK2 activity was determined. B, F9 cells expressing ␤ 2 AR/Rfz1 receptor and luciferase reporter construct pTOPFLASH were treated with either pertussis toxin or antisense oligos against G␣ q , G␣ o , or G␣ i2 as described in panel A prior to stimulation for 5 h with isoproterenol (10 M). Cell lysates were assayed for Lef/Tcf-sensitive transcription using the luciferase activity assay. C, immunoblotting of G␣ subunits and ␤-actin in control cells and cells treated with antisense ODN. ␤ 2 AR/Rfz1-expressing clones were incubated without or with G␣ subunit-selective antisense ODN for 72 h. The whole-cell lysates were subjected to SDS-gel electrophoresis and immunoblotting. The blot shown is a representative from three separate experiments. D, expression of the constitutively active Q209LG␣ q or Q205LG␣ o mutants activates CK2. 24 h post-transfection, the activity of CK2 was determined in F9 clones transfected with either empty vector or with vector harboring the constitutively active Q209LG␣ q , Q205LG␣ o , or Q209LG␣ 11 mutant G protein subunits. The results are presented as mean values Ϯ S.E. obtained from at least three separate experiments. Statistical variance was measured (***, p Ͻ0.005; **, p Ͻ0.01). E, constitutively active G␣ q activates Lef/Tcf-sensitive transcription. The activity of transcription reporter luciferase construct was determined in F9 clones transfected with empty vector or the vector harboring constitutively active Q209LG␣ q , Q205LG␣ o , or Q209LG␣ 11 mutant G protein subunits. Clones were treated with either vehicle (Ϫ) or 20 M apigenin (ϩapigenin) for 1 h prior to lysis for analysis. The data shown are mean values Ϯ S.E. reduced the phosphorylation of Dvl-2 (Fig. 5A) as well as Lef/Tcf-sensitive transcription (Fig. 2D) in response to stimulation by Wnt3a. Overexpression of Dvl was shown to activate the transcription (33,34). We wonder whether Dvl plays a similar role in CK2-dependent activation of Wnt/␤-catenin signaling. Transient expression of a TAP-tagged Dvl2 activates Lef/Tcf-dependent transcription (Fig. 5B). Knock down of CK2␣ in F9 cells by siRNA treatment resulted in a sharp decrease in Dvl2-stimulated Lef/Tcf-sensitive transcriptional activation (Fig. 5B), indicating that CK2 is downstream of Dvl and essential for Dvl to activate Lef/Tcf-dependent transcription. Knock down of Dvl-2 expression by siRNA treatment for 24 -48 h decreased Dvl-2 Ͼ90% (Fig. 5C). The Wnt3a-stimulated CK2 activation was blocked by knocking down Dvl-2 (Fig. 5C). Knock down of Dvl-1, in contrast, had no effect on the ability of Wnt3a to stimulate CK2 activity (Fig. 5D). Thus, Wnt3a activation of CK2 requires Frizzled-1, G␣ q / o , and Dvl-2 in a vectorial arrangement.

DISCUSSION
CK2 has been shown to induce dosalization and axis formation in Xenopus embryos, mimicking the effects of XWnt8, Dvl, and ␤-catenin stimulation, and is hypothesized to play a role in Wnt/␤-catenin pathway (20). CK2 can phosphorylate ␤-catenin at Thr 393 in the armadillo repeat region of ␤-catenin in mammalian cells in vitro, leading to stabilization and accumulation of ␤-catenin and to co-transcriptional activation of Lef/Tcf-sensitive gene expression (19). How CK2 exerted its role upon Wnt stimulation, however, was insufficiently understood because of its ubiquitous expression and results suggesting CK2 to be FIGURE 5. Wnt activation of CK2 requires Dishevelled-2. A, Wnt3a stimulation of Dvl-2 phosphorylation is blocked by inhibition of CK2 with DMAT. F9 cells expressing rat Frizzled-1 were pretreated with DMAT or not (control) for 30 min, followed by Wnt3a stimulation. After stimulation with Wnt3a for the indicated times, cells were lysed and immunoblots for Dvl-2 were performed on the lysates. A major phosphorylated Dvl-2 migrating slightly more slowly than Dvl-2 was seen in response to Wnt3a stimulation. Phosphorylation of Dvl-2 was abolished in the presence of DMAT. A representative image is presented. B, overexpression of Dvl-2 stimulates Lef/Tcf-dependent transcription, mediated by CK2. F9 cells were pretreated with 100 nM siRNA for 24 h to knock down CK2␣ subunit expression. These siRNA-treated clones were transiently transfected with expression vectors harboring either Dvl-2 (pTAP-Dvl-2) and Lef/Tcf-sensitive luciferase reporter (pSuper8XTOPFLASH) or the reporter pSuper8XTOPFLASH alone. The activity of luciferase reporter gene was determined 24 h after the transfection. TAP-Dvl-2, CK2␣, and actin level in the cell lysate were subjected to SDS-PAGE and immunoblotting. The blots were stained with antibodies against Dvl-2, CK2␣, or ␤-actin (lower panel). C, CK2 activation stimulated by Wnt3a is blocked by knock down of Dvl-2. F9 clones expressing rat Frizzled-1 receptor were treated with siRNA to Dvl-2 for 48 h. After Wnt3a stimulation for 15 min, CK2 activity was determined as described. The knock down of Dvl-2 expression by treatment with siRNA was determined in whole-cell lysates by SDS-PAGE and immunoblotting (lower panel). D, knock down of Dvl-2, but not Dvl-1, abolishes Wnt3astimulated activation of CK2. Cells were treated with siRNA against Dvl-1, Dvl-2, or none (control) for 48 h and assayed for the ability of Wnt3a to stimulate CK2 activity. Results are mean values from three separate experiments. The knock down of Dvl-1, Dvl-2, or none (control) by treatment with siRNA was determined in whole-cell lysates by SDS-PAGE and immunoblotting (lower panel). Immunoblotting of ␤-actin in whole-cell lysates was performed as a loading control. Wnt3a treatment of F9 embryonic stem cells activates Lef/Tcf-sensitive gene expression and the formation of primitive endoderm. Wnt ligand binds to Frizzled-1, a member of the superfamily of G protein-coupled receptors, and thereby activates G proteins G o and G q . Downstream of G proteins in mammalian cells as well as Drosophila is the phosphoprotein Dishevelled, which suppresses the constitutive activation of glycogen synthase kinase 3␤ and its phosphorylation and destabilization of ␤-catenin. CK2 activity is regulated by Wnt3a stimulation, and this activation is shown to be downstream of heterotrimeric G proteins and Dishevelled. "constitutively active", i.e. an "unregulated" enzyme (11). Our study is the first report that CK2 activity is regulated acutely in response to stimulation of Wnt3a. We have demonstrated that Wnt stimulates transient activation of CK2 and that the CK2 response is obligate for Wntstimulated accumulation of ␤-catenin and activation of Lef/Tcfdependent transcription in canonical pathway (Fig. 6). The CK2 activation is mediated by the heterotrimeric G proteins G q and G o and requires the downstream signaling phosphoprotein Dishevelled-2. Suppression of the expression of CK2␣ subunits, G␣ q or G␣ o , or of Dvl2, like chemical inhibition of this kinase, inhibited the activation of CK2 activity as well as ␤-catenin accumulation and co-activation of transcription in response to Wnt3a stimulation. Whether the change in CK2 activity reflects a change in intrinsic activity or the removal of negative regulators of CK2 activity remains an open question.
It has been proposed that ␤ subunit of CK2 confers the substrate specificity (11). The ␣ and ␤ subunits of CK2 have been reported to be necessary in tandem to induce dual axis formation in Xenopus (36). Microinjection of either CK2 subunit alone did not affect axis development (36). Our results from siRNA treatments to specifically knock down CK2 ␤ subunits indicate that these subunits do not regulate CK2 in response to Wnt stimulation. To the contrary, the catalytic ␣ and ␣Ј subunits of CK2 are necessary for transduction of the Wnt signals.
We have demonstrated herein that CK2 is, in fact, a regulated enzyme. CK2 activity is activated by stimulation of Frizzled-1, and CK2 is positioned downstream of heterotrimeric G proteins and the phosphoprotein Dishevelled-2 in Wnt/␤-catenin signaling. One might anticipate that CK2 may associate with Dishevelleds, which act as molecular "toolboxes" for Wnt signaling. Support for this notion is provided in Wnt-1-expressing mouse mammary epithelial cells, in which CK2 has been found with Dvl in a complex that catalyzes the phosphorylation of ␤-catenin, making it resistant to proteasomal degradation (37). In Dfz2expressing Drosophila S2 cells, CK2 also was shown to be associated with Dishevelled (14). In our study, Dvl-2 is necessary for CK2 activation in response to Wnt3a stimulation in mouse F9 cells. Clearly, how CK2 is activated in response to stimulation by Wnt3a, and the role of Dvl-2 in this regard, will be important questions to be answered.