Inositol Pentakisphosphate Mediates Wnt/β-Catenin Signaling*

Wnt3a stimulates lymphoid enhancer factor/T-cell factor protein-sensitive transcription, i.e. the canonical pathway, in mouse F9 embryonal tetratocarcinoma cells expressing rat Frizzled-1. We explored the potential roles for inositol polyphosphates as mediators of Wnt signaling in the canonical path-way. Wnt3a triggers G-protein-linked phosphatidylinositol signaling, transiently generating inositol polyphosphates, especially inositol pentakisphosphate (IP5) accumulation. Knock-down of Gαq abolishes, whereas expression of the Q209L constitutively active mutant of Gαq mimics, the effects of Wnt3a on IP5 generation and downstream signaling. Phospholipase Cβ-1 and Cβ-3 mediate the G protein signal to the level of phosphatidylinositol signaling. Knock-down and inhibitor studies of the enzymes responsible for generating IP5 reveal inositol 1,4,5-trisphosphate 3-kinase and inositol polyphosphate multikinase as key mediators in the production of IP5. Wnt3a stimulation of the canonical pathway requires accumulation of IP5, which acts to inhibit the activity of glycogen synthase kinase-3β, whereas stimulating casein kinase 2. Blockade of Wnt3a stimulation of IP5 generation blocks β-catenin accumulation, activation of lymphoid enhancer factor/T-cell factor protein-sensitive transcription, and promotion of primitive endoderm formation in response to Wnt3a. Phosphatidylinositol signaling mediates Wnt3a action in the canonical pathway, acting to generate inositol pentakisphosphate, a key second messenger of Wnt3a.

Recently we reported that CK2 is activated by Wnt3a, operating downstream of Dvl and upstream of Lef/Tcf-sensitive gene transcription (23). Although earlier reported to act solely as "constitutively active" in cells, CK2 has been shown to be activated by Wnt3a (23). Overexpression or inhibition of CK2 affects early development via the Wnt/␤-catenin/Lef-Tcf-sensitive gene expression (24). CK2 enhanced ␤-catenin stabilization by phosphorylating ␤-catenin at Thr 393 in the armadillo repeat region of ␤-catenin in vitro (22). Although CK2 clearly mediates an essential aspect of Wnt3a signaling, the nature of this regulation of CK2 is unknown. CK2 activity has been shown in vitro to be sensitive to inositol polyphosphates (25). Consequently we hypothesized that Wnt may regulate CK2 (and thereby the Wnt/␤-catenin or canonical pathway) by influencing phosphatidylinositol signaling. In the current report, we investigate the validity of this hypothesis and probe the generation of water-soluble inositol polyphosphates in Wnt action. Our results show that Wnt3a stimulates phosphatidylinositol breakdown and the accumulation of inositol polyphosphates, particularly IP 5 . IP 5 generation in response to Wnt3a is shown to be operating downstream of Frizzled-1, heterotrimeric G proteins (especially G q ), and PLC␤ in the canonical Wnt/ ␤-catenin/Lef-Tcf pathway of mammalian cells. Inositol 1,4,5trisphosphate 3-kinase (IP3K) and inositol polyphosphate multikinase (IPMK) are shown to be key mediators of the production of IP 5 . Inhibiting the action of either enzyme blocks the ability of Wnt3a to accumulate ␤-catenin, to activate Lef/Tcfsensitive transcription, and to promote formation of primitive endoderm (PE).
Cell Culture-The mouse F9 teratocarcinoma (F9) cells (from ATCC, Manassas, VA) were grown in Dulbecco's modified Eagle's medium supplemented with 15% fetal bovine serum (Hyclone, South Logan, UT) at 37°C in a 5% CO 2 incubator. The human embryonic kidney HEK293 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and grown at 37°C in a humidified incubator of 5% CO 2 and 95% air. F9 and HEK293 cells were co-transfected with an expression vector harboring rat Friz-zled1 (Fz1) and a luciferase reporter construct, Super8X TOPflash by using Lipofectamine Plus (Invitrogen) according to manufacturer's protocol. Stable clones were selected in the presence of 400 ng/ml G418, and maintained in the media containing 100 ng/ml G418 as previously described. For Wnt stimulation, cells were grown to monolayer and incubated in the absence of serum for 8 h prior to the treatment by either Wnt3a (15 ng/ml) or Wnt5a (60 ng/ml).
Analysis of Soluble Inositol Phosphates-Cells expressing Fz1 were plated on a 35-mm culture dish and incubated in inositolfree Dulbecco's modified Eagle's medium (Chemicon) containing myo-[ 3 H]inositol (15 Ci/well) and 15% fetal bovine serum for 48 h. Cells were cultured in Dulbecco's modified Eagle's medium (without serum) for an additional 8 h before Wnt stimulation. Soluble inositol phosphates were extracted according to the method previously described (27). Briefly, cells were lysed in 500 l of methanol and 0.5 N HCl mixture (methanol, 0.5 N HCl ϭ 2:1) and inositol phosphates were extracted with 335 l of chloroform. The aqueous phase containing inositol phosphates was neutralized by addition of 80 l of saturated NaHCO 3 . Equal counts of soluble inositol phosphates were applied to a strong anion-exchange, Partisil 10-SAX HPLC column (4.6 ϫ 250 mm, Whatman, Florham Park, NJ) and eluted first with a linear gradient of 0.01-1.7 M ammonium phosphate (pH 3.5) over 30 min and followed by a 35-min step of 1.7 M ammonium phosphate (pH 3.5) at a flow rate of 0.5 ml/min. Fractions (0.5 ml/fraction) were collected and 5 ml of scintillation fluid were added into each tube. Radioactivity was measured by a scintillation counter. All data were normalized with respect to the total radioactivity measured in the lipid fraction. The elution times of internal standards, ADP and ATP, were monitored by a spectrophotometer at 260 nm. The identity of individual inositol phosphates was assigned on the basis of elution with known standards under the same condition.
Construction of Plasmids-pTrcHisA-hIPMK was a gift from Dr. Solomon Snyder (Johns Hopkins Medical School). To subclone the hIPMK into a mammalian expression vector, PCR was used to amplify inserted hIPMK in pTrcHisA-hIPMK by using the 5Ј-primer with EcoRI site (5Ј-GGAATTCGGAT-GGCCGCCGAGCCCCCAGC-3Ј) and 3Ј-primer with BglII site (5Ј-ACAAGATCTTCAACTGTCCAAGATACTCCG-AAG-3Ј). The PCR product encoding hIPMK was digested by EcoRI and BglII and subsequently subcloned into mammalian expression vectors, pCMV-HA and pCMV-MYC (BD Biosciences). The identity of the amplified sequences was confirmed by direct DNA sequencing.
Knock-down Target Proteins by Small Interfering RNA (siRNA)-F9 cells expressing Fz1 were grown to ϳ60% confluency on 12-well plates. siRNA (100 nM, final concentration) was introduced into cells by using Lipofectamine 2000 (Invitrogen). Cells were cultured in the presence of siRNA for an additional 48 -72 h according to manufacturer's recommendation. Sources of siRNA used in the study are listed as follows. siRNAs targeting G␣ q , G␣ o , G␣ 11 , and PLC␤ 4 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). siRNAs targeting PLC␤ 1 and PLC␤ 3 were synthesized by Ambion (Austin, TX) according to the sequence published by Kim et al. (28). All following siRNAs were obtained from Ambion: siRNA to IPMK (ID code 203326), siRNA to IP3K-A (ID code 170266), siRNA to IP3K-B (ID code 203096), siRNA to IP3K-C(ID code 171717), and siRNA to Dvl2 (ID code 61090). We used either immunoblotting or semi-quantitative RT-PCR analysis to determine the efficiency of knockdown on targeted molecules.
Immunoblotting-Cells were harvested in lysis buffer containing protease and phosphatase inhibitors (20 mM Tris, pH 7.4, 150 mM NaCl, 1% Triton X-100, 200 M phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 10 g/ml leupeptin, 10 mM NaF, 1 mM Na 3 VO 4 , and 100 nM okadaic acid) and the mixture were centrifuged at 15,000 ϫ g for 20 min at 4°C. The supernatant was collected and protein concentration was determined by the Lowry method (29). Equal amounts of proteins from samples were subjected to SDS-PAGE and separated proteins were transferred to nitrocellulose blots electrophoretically. The blots were incubated with 10% bovine serum albumin for 0.5 h, rinsed with water, and then probed with antibody against targeted proteins. To determine cytoplasmic levels of ␤-catenin, cells were lysed in RIPA 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 aprotinin, 10 g/ml leupeptin) and cell lysate was incubated with ConA-Sepharose (Amersham Biosciences) for 1 h. The suspension was applied by centrifugation to remove membrane-associated ␤-catenin. The supernatants were subjected to SDS-PAGE, followed by immunoblotting with anti-␤-catenin antibody. Immune complexes were detected by the enhanced chemiluminescence method, as per the manufacturer's instructions.
Immunoprecipitation-F9 cells stably transfected with Fz1 were grown in 100-mm dishes and transiently transfected with 4 g of pCMV-HA-IPMK. Twenty-four hours later, cells were treated with Wnt3a and then lysed at the indicated time with 3 ml of lysis buffer (20 mM Tris, pH 7.4, 150 mM NaCl, 10 mM NaF, 1 mM Na 3 VO 4 , and 1% Triton X-100) containing protease inhibitors (200 M phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, and 10 g/ml leupeptin). Whole cell lysates (2 mg) were incubated with anti-HA high affinity antibody (Roche Applied Science) immobilized on protein A/G-agarose at 4°C for 4 h. Immune complexes were precipitated by centrifugation (10,000 ϫ g) and dissolved in Laemmli's solution. Protein mixture was subjected to SDS-polyacrylamide gel electrophoresis followed by electrotransblotting. HA-tagged IPMK were analyzed by immunoblotting as described.
Lef/Tcf Reporter Assay-F9 clones stably transfected with Fz1 and Super8X TOPflash were cultured in 12-well plates and stimulated with Wnt3a for 5 h. Cell lysates were collected in reporter lysis buffer (Promega). Cell lysates (10 l) 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). The intensity of luminescence was immediately measured using a luminometer (Lumat LB 9507, Berthold Technologies, Oak Ridge, TN). To study the effect of ATA, CGA, and adriamycin on Lef/Tcf-dependent transcription, inhibitors were added into culture media for 30 min prior to Wnt3a stimulation. Samples were assayed in triplicate and the luciferase activity was normalized based on protein concentration.
In Vitro CK2 Kinase Activity Assay-CK2 activity was analyzed, as previously described (30). Briefly, immunoprecipitated CK2␣ 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 (HRRRDDDSDDD) in kinase assay buffer containing 10 Ci of [␥-32 P]GTP, 160 M GTP, 25 mM MgCl 2 and with or without inositol phosphates 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 and incorporated ␥-32 P in substrates on the membrane was measured by liquid scintillation spectrometry.
In Vitro GSK3␤ Kinase Activity Assay-GSK3␤ activity was analyzed, as previously described (31). Briefly, GSK3␤ was immunoprecipitated from whole cell lysates followed by washing the pellet with kinase assay buffer (50 mM Tris, pH 7.5, 25 mM ␤-glycerol phosphate, 5 mM EGTA, 1 mM sodium orthovanadate, and 1 mM dithiothreitol). The immunoprecipitated GSK3␤ was incubated with 10 M substrate peptide (YRRAAVPPSPSLSRHSSPHQ(pS)EDEEE) in kinase assay buffer containing 10 Ci of [␥-32 P]ATP, 160 M ATP, 25 mM MgCl 2 , and with or without inositol phosphates 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 and incorporated ␥-32 P in substrates on the membrane was measured by liquid scintillation spectrometry.
Analysis of Primitive Endoderm Formation-F9 clones expressing Fz1 were propagated on 12-well plates. Cells were treated with control siRNA, or siRNA targeting IPMK in combination with siRNA targeting IP3K-B for 48 h and then challenged without or with Wnt3a. At the end of 4 days, clones were fixed with 3% paraformaldehyde for 5 min, stained with the mouse endoderm-specific marker, cytokeratin endo A, by the monoclonal antibody TROMA-1 (32). The fixed cells were examined using a Zeiss Axiovert 200M microscope for both phase-contrast and fluorescence images.

RESULTS
Inositol Metabolism in Wnt3a-treated Cells-Mouse F9 embryonic totipotent, teratocarcinoma cells stably transfected with the rat Fz1 display activation of the Wnt/␤-catenin/Lef/ Tcf canonical pathway in response to treatment with Wnt3a, but not Wnt5a (9). More recently, we showed that Wnt3a activates CK2 and that this activation is obligate for Wnt3a stimulation of Lef/Tcf-sensitive gene transcription (23). Based upon the report that inositol phosphates (e.g. IP 4 , IP 5 , and IP 6 ) may regulate the activity of CK2 in vitro (25), we investigated the nature of water-soluble inositol polyphosphates accumulated in cells treated with Wnt3a for up to 60 min (Fig. 1A). F9 clones stably expressing either rat Fz1 or rat Frizzled-2 (Fz2) were labeled with [ 3 H]inositol for 48 h and then treated with purified Wnt ligands for 0 to 60 min. F9 cells stably transfected with empty expressing vector (EV) were used as control cells. Anion exchange HPLC of either standard IP markers or water-soluble extracts from F9 cells expressing Fz1 and treated with Wnt3a (15 ng/ml) enabled analysis of the metabolism of inositol phosphates (Fig. 1A). The HPLC retention times, profiles, and the amplitude of the changes was highly reproducible under these conditions (not shown). Separation of inositol trisphosphate (IP 3 ), inositol 1,3,4,5-tetrakisphosphate (IP 4 ), inositol 1,3,4,5,6pentakisphosphate (IP 5 ), and inositol hexakisphophate (IP 6 ) by anion exchange was excellent ( Fig. 1, standards); IP 3 and IP 5 were readily identified in the cell extracts. In the absence of Wnt3a stimulation (ϩWnt3a, at "0" min), the HPLC was able to detect IP 5 . A slight elevation of IP 5 is present in the control (empty vector-transfected as well as wild-type) cells (data not shown). Thus, the baseline levels of IP 5 are not associated per se with expression of Fz1 in these cells. Within 10 min of Wnt3a treatment, the inositol phosphate profiles revealed increased IP 3 and IP 5 accumulation. Wnt3a stimulates an accumulation of IP 5 that reaches peak values within 15 min and declines thereafter. Quantification of intracellular IP 3 , IP 4 , and IP 5 in response to Wnt3a stimulation from multiple, separate preparations of [ 3 H]inositol-labeled cells was performed to ascertain if the IP 3 response to Wnt3a was significant (Fig. 1B). The increase in IP 3 that peaks within 5 min was not significant, unlike the sharp increase in IP 5 accumulation that peaks within 15 min.
The non-canonical Wnt/cyclic GMP, Ca 2ϩ /NF-AT-sensitive transcription pathway has been shown to regulate phosphatidylinositol signaling (33)(34)(35), so we explored if the activation of the non-canonical as well as the canonical pathways both regulate accumulation of inositol polyphosphates (Fig.  1C). Cells transfected with EV, or expression vector harboring either Fz1 or Fz2 were labeled with [ 3 H]inositol. Treating the EV-transfected cells with Wnt3a or the Fz2-expressing cells with Wnt5a (60 ng/ml) stimulates no accumulation of IP 5 over 60 min. IP 5 accumulation was robust for the Fz1-expressing cells treated with Wnt3a (Fig. 1C), but not for those treated with Wnt5a (not shown). The complementary study of IP 3 accumulation in response to Wnt stimulation was performed (Fig. 1D). IP 3 accumulation is prominent in the Fz2 expressing cells stimulated with Wnt5a (Fig. 1D), but not in those same cells treated with Wnt3a (not shown). Furthermore, we treated F9 cells with LiCl in an effort to block the IP 3 catabolic cascade (see supplemental data Fig. S1). Treatment with LiCl abolishes the appearance of the IP 5 peak in cells expressing Fz1 and treated with Wnt3a; a corresponding increase in the accumulation of IP 3 was not observed. LiCl treatment enhances IP 3 accumulation in cells expressing Fz2 and treated with Wnt5a.
Finally, we tested if this downstream signaling of Wnt3a to IP 5 accumulation would be observed in another cell line. Human embryonic kidney (HEK) 293 cells in culture stably transfected with either empty vector (as control) or the expression vector harboring Fz1 were tested (Fig. 1E). Whereas control cells were not affected by Wnt3a stimulation (data not shown), Fz1-expressing cells treated with Wnt3a displayed increased accumulation of IP 5 at 10-and 15-min poststimulation (Fig. 1E), as observed in the mouse F9 teratocarcinoma embryonal cells.
G␣ q Mediates the Increase in Intracellular IP 5 in Response to Wnt3a Stimulation-Fz1 has been shown to be a member of the superfamily of seven transmembrane segment receptors that operate via heterotrimeric G proteins, from the mouse (9) to flies (10). We investigated if the Wnt3a-stimulated, Fz1-mediated accumulation of IP 5 observed in the mouse F9 cells was sensitive to the suppression (knockdown) or expression of constitutively active mutants (Gln to Leu substitutions) of those G proteins shown to mediate the canonical pathway in mouse cells and flies (Fig. 2). Knockdown of the individual ␣ subunits of G q and G o (as well as G 11 as a control) was accomplished using siRNA. The extent to which the siRNA treatments yield an effective knockdown was established by subjecting a sample of the whole cell lysates to SDS-PAGE and the resolved gel to immunoblotting ( Fig. 2A). siRNA designed for each of the G␣ subunits was shown to specifically knockdown the targeted G protein subunit, but not non-targeted G protein subunits, following a 60-h treatment with siRNA. Immunoblotting of ␤-actin provides a measure of equivalent loading, being unaffected by siRNA treatment. Treatment with the control siRNA designed and provided by the commercial supplier, likewise showed no ability to knockdown either the G protein subunits or the ␤-actin.
We investigated first if the knockdown of G protein subunits influenced the ability of the Fz1-expressing F9 cells to respond to Wnt3a stimulation, measuring the activation of Lef/Tcf-sensitive transcription as the read-out (Fig. 2B). siRNA treatments of the cells with the control siRNA or siRNA targeting G␣ 11 have no influence on the ability of the cells to activate Lef/Tcfsensitive transcription in response to Wnt3a. Knockdown of G␣ q , in contrast, virtually abolishes the ability of these cells to activate the canonical Wnt pathway in response to purified Wnt3a. Similarly, knockdown of G␣ o led to suppression of the ability of Wnt3a to activate the canonical Wnt3a/␤-catenin/ Lef-Tcf pathway, in good agreement with earlier studies in these cells (9) and in Drosophila (10). Having established the biology of the response and the ability of the knockdown of specific G protein subunits to suppress the response, we investigated the effects of knockdown of the ␣-subunits of G q , G o , and G 11 on the accumulation of tritiated IP 5 from [ 3 H]inositollabeled cells (Fig. 2C). Knockdown of G␣ q abolishes the ability of Wnt3a to accumulate IP 5 in the Fz1-expressing cells. Knockdown of G␣ o , to a lesser extent than the knockdown of G␣ q , also FIGURE 2. G␣ q mediates the increase in intracellular IP 5 in response to Wnt3a. F9 clones stably expressing the Fz1 and Super8X TOPflash (as a luciferase-coupled gene reporter to measure Lef/Tcf-sensitive transcription) were either treated with siRNA for 60 h to knockdown G␣ q , G␣ o , or G␣ 11 (A-C) or transiently transfected with constitutively activated G proteins (D). A, representative immunoblots (IB) of G-protein ␣ subunits in cells treated with siRNA to individually knockdown G␣ q , G␣ o , or G␣ 11 . Cells were treated for 60 h with siRNA. The cells were lysed and subjected to SDS-PAGE and immunoblotting. The blots of resolved proteins were stained for each subunit. ␤-Actin was probed as a loading control. B, clones treated with siRNA were stimulated by Wnt3a for 6 h and Lef/Tcf-sensitive luciferase reporter gene activity was assayed. C, F9 clones stably expressing the Fz1 were metabolically labeled with [ 3 H]inositol for a total 84 h. Twenty-four h after the start of the incubation in the media containing [ 3 H]inositol, cells were treated with siRNA reagents targeting G␣ q , G␣ o , or G␣ 11 in the same media for 60 h. Cells were serumstarved for 8 h prior to stimulation for 15 min with Wnt3a (15 ng/ml). Labeled inositol polyphosphates were isolated and quantified. The IP 5 levels were determined. Results are displayed as mean Ϯ S.E., derived from at least three separate experiments. *, p Ͻ 0.01; **, p Ͻ 0.001; versus control siRNA-treated and Wnt3a-stimulated groups; setting the value of control siRNA-treated cells without stimulation as "1." D, F9 clones were metabolically labeled with suppresses the ability of Wnt3a to stimulate IP 5 accumulation. Knockdown of G␣ 11 , in contrast, has no effect on Wnt3a to stimulate the IP 5 response.
Conversely, if the knockdown of G␣ q was able to abolish the ability of Wnt3a to stimulate IP 5 accumulation, one would predict that expression of a constitutively active (CA) mutant of this G protein ␣-subunit (lacking the intrinsic GTPase activity of the wild-type molecule) might mimic the IP 5 response in the absence of Wnt3a stimulation (Fig. 2D). Individual, CA mutants of G protein ␣-subunits were co-expressed with Fz1 in the F9 cells and the accumulation of IP 5 followed in the inositollabeled cells. Expression of the Q209L, the CA mutant of G␣ q increases the ambient levels of both IP 5 as well as IP 3 in these cells, in the absence of treatment with Wnt3a. Addition of Wnt3a did not further increase the amount of intracellular labeled inositol polyphosphates (not shown). Expression of the Q205L, the CA mutant of G␣ o also was associated with an increase in the ambient levels of IP 3 , although yielding a smaller fold-increase over that of the IP 3 levels in EV-control cells than that is similarly observed for the cells expressing the Q209L G␣ q as compared with EV cells. Expression of the Q209L CA mutant of G␣ 11 , in contrast, yields no change in the ambient levels of either IP 3 or IP 5 (Fig. 2D). Relative levels of expression of these CA mutants of G-protein ␣ subunits were equivalent (Fig. 2D).
PLC␤ Mediates Wnt3a-stimulated IP 5 Accumulation-The best known downstream effectors of G␣ q are members of the PLC ␤-family of isoforms (36). To circumvent the absence of antibodies that could reliably stain the various isoforms of PLC␤ expressed in these cells, we resorted to the use of RT-PCR with primers specifically designed to amplify the mRNAs for each of the mouse PLC␤ isoforms to evaluate expression in F9 cells (Fig.  3A). PLC␤ 1 , PLC␤ 3 , and PLC␤ 4 mRNAs are detected in the F9 cells, whereas PLC␤ 2 is not. siRNA were designed to suppress each of the isoforms and cells were treated with an siRNA for 48 h prior to analysis of F9 cells transiently expressing Fz1 (Fig.  3, A and B). RT-PCR amplication of the mRNAs from the F9 cells was performed using primers for all three isoforms and RNA from cells pretreated with either control siRNA or an individual siRNA targeting a specific isoform. The results from the amplification indicate that the siRNAs designed for each isoform indeed were specific, reducing the amount of mRNA of a targeted isoform to almost below detection, whereas not significantly reducing the mRNA encoding the other isoforms or the cyclophilin A control (Fig. 3A). Assay of the ability of Wnt3a to activate Lef/Tcf-sensitive gene transcription was assayed in cells in which knockdown of one of three isoforms of PLC had been targeted (Fig. 3B). knockdown of either PLC␤ 1 or PLC␤ 3 , but not PLC␤ 4 provokes an attenuation of the ability of Wnt3a to activate gene transcription. When used in tandem, the siRNA targeting PLC␤ 1 and PLC␤ 3 in combination yield a profound suppression of Wnt3a action on the Lef/Tcf-sensitive transcription (Fig. 3B). Having established that both PLC␤ 1 and PLC␤ 3 were participating at downstream effects for Wnt3a activation of Frizzled-1, we probed if suppression of these two PLC isoforms would impact the ability of Wnt3a to stimulate the accumulation of IP 5 (Fig. 3C). Knockdown of PLC␤ 1 and PLC␤ 3 , in tandem, resulted in substantial reduction in the RT was employed followed by PCR amplification using primers specific for mRNA of each individual PLC isoform. Representative results of RT-PCR amplifications are displayed. Cyclophylin A mRNA was amplified as a control. B, Fz1-expressing cells were treated with siRNA for 48 h prior to the stimulation of Wnt3a (15 ng/ml) and luciferase reporter activity was assayed 6 h after the administration of Wnt3a. C, F9 clones stably expressing the Fz1 were metabolically labeled with [ 3 H]inositol for 72 h. Twenty-four hours after the start of the incubation in the media containing [ 3 H]inositol, cells were treated with either control siRNA or siRNA targeting both PLC␤ 1 and PLC␤ 3 in combination in the same media for 48 h. Cells were serum-starved for 8 h prior to stimulation with Wnt3a (15 ng/ml). Fifteen minutes after Wnt3a administration, inositol polyphosphates were isolated from whole cell extracts, separated, and quantified as described. The results are displayed as mean Ϯ S.E., derived from at least three separate experiments. **, p Ͻ 0.001, versus Wnt3atreated, control groups; #, p Ͻ 0.01, versus Wnt3a-treated cells pretreated with siRNA to PLC␤ 1 or PLC␤ 3 , individually. amount of IP 5 accumulated in response to Wnt3a in the F9 cells expressing Fz1. The PLC␤ inhibitor 1-octadecyl-rac-glycero-3phosphocholine (ET-18-OH) also effectively blocks accumulation of inositol polyphosphates (not shown).
IPMK and IP3K Mediate Wnt3a-induced Accumulation of Intracellular IP 5 -The metabolism of phosphatidylinositol 4,5bisphosphate to IP 5 downstream of PLC␤ is achieved largely by the action of two enzymes: the inositol polyphosphate multikinase (IPMK or IPK2 or IP 4 5-kinase) and the IP3K (Fig. 4A) (37)(38)(39)(40)(41)(42). Aurintricarboxylic acid (ATA, 2 M) has been shown in vitro to inhibit IP3K and IPMK; whereas adriamycin (20 M) has been shown in vitro to inhibit IP3K and CGA (40 M) has been shown in vitro to inhibit IPMK (43,44). We probed the effects of each of these inhibitors on the ability of Wnt3a to stimulate IP 5 accumulation in F9 cells expressing Fz1 (Fig. 4B). Based upon the schematic of phosphatidylinositol 4,5-bisphos-phate metabolism to IP 3 , IP 4 , and IP 5 (Fig. 4A), one would predict that IPMK and IP3K are jointly involved in the production of IP 5 in response to Wnt3a, all three inhibitors would suppress to some extent the ability of Wnt3a to stimulate accumulation of IP 5 . Initial analysis of inositol phosphate metabolism in serumstarved F9 cells pretreated with either ATA or CGA and then stimulated with serum revealed a loss in accumulation of IP 5 (see supplemental data Fig. S2). Based upon multiple experiments aimed at measuring IP 5 accumulation, treatment with each of the three inhibitors proved to be an effective means of abolishing the ability of Wnt3a to stimulate IP 5 accumulation. If IP3K were not participating in this metabolism, adriamycin likely would have little effect, but in this case adriamycin treatment abolishes the IP 5 response to Wnt3a. We propose that IP3K is the major enzyme to convert IP 3 to IP 4 as it was found in human (45).
Beyond the use of enzyme inhibitors, we made use of siRNA to knockdown both IP3K and IPMK (Fig. 4C). IP3K has three isoforms in mammals, designated IP3K-A, -B, and -C (46). On the basis of RT-PCR amplification studies, all three are found to be expressed in mouse F9 cells (not shown). Treating cells with siRNA targeting the IP3K-B isoform effectively suppresses the expression of this enzyme, as shown by immunoblotting. Suitable antibodies to IP3K-A and -C isoforms were not available, but the siRNAs for each isoform were found to effectively suppress mRNA levels (not shown). Similar knockdown studies were performed targeting IPMK, again using RT-PCR amplification of mRNA in the absence of a suitable antibody for detection of IPMK by immunoblotting. The levels of IPMK mRNA in cells treated with the control siRNA versus the IPMK-targeting siRNA were established. The RT-PCR amplification data demonstrate effective suppression of IPMK mRNA through the use of these siRNA reagents (Fig.  4C). Having established conditions directed toward knockdown of both enzymes, we evaluated the ability of siRNA-mediated suppression to modulate the IP 3 and IP 5 response to Wnt3a stimulation of the canonical pathway (Fig. 4D). In the absence of stimulation by Wnt3a, cells display little change in intracellular IP 3 or IP 5 levels when IPMK is suppressed by siRNA. Earlier studies report dramatic reductions in IP 5 accu- mulation in HEK293 and HeLa cells (45), as well as in Rat-1 cells (38), when IPMK expression was suppressed with siRNA. Unlike these earlier studies performed with serum-fed cells, our studies required that we eliminate serum to study the Wnt response, likely the basis for the differences in the effects observed in response to IPMK suppression.
Suppression of IP3K-B (Fig. 4D), but not that of either IP3K-A or IP3K-C (not shown), leads to an increase in IP 3 (likely a reflection of blocked utilization of IP 3 ), but not IP 5 . Suppression of both IP3K-B and IPMK in tandem yields a reduction in intracellular IP 5 , as predicted (Fig. 4D). The ability of the suppression of IP3K-B to increase the intracellular IP 3 level is amplified in cells treated with as compared without Wnt3a (Fig. 4D). To provide independent verification of the ability of the siRNA reagents for IP3K-B and IPMK to block IP 5 accumulation, we assayed the effects of these same siRNA reagents on IP 3 , IP 4 , and IP 5 levels in serum-starved F9 cells that were metabolically labeled and then stimulated with serum for 10 min (see supplemental data Fig. S3). The results demonstrate that knock-down of either enzyme blocks IP 5 accumulation in response to an independent agonist (i.e. serum). Wnt3a stimulates a marked increase in the accumulation of IP 5 in the cells treated with control siRNAs. Suppression of either IP3K or IPMK, or both enzymes, substantially reduces the ability of Wnt3a to stimulate IP 5 accumulation. These data, taken together, suggest that the source of the IP 5 generated in response to Wnt3a includes roles of both IP3K and IPMK.
IPMK and IP3K Are Essential for Wnt3a/␤-Catenin Canonical Signaling to PE Formation-It was essential to probe further downstream in the Wnt/␤-catenin pathway, to establish the linkage between IPMK/IP3K and changes in IP 5 extended to the regulation of intracellular accumulation of ␤-catenin, activation of Lef/Tcf-sensitive gene transcription, and PE formation. At the level of ␤-catenin accumulation, treatment of cells with inhibitors, either adriamycin (IP3K-selective) or ATA (inhibits both IP3K and IPMK) attenuates Wnt3a action (Fig. 5A). The Lef/Tcf-sensitive transcriptional response to Wnt3a stimulation is similarly sensitive to each of these inhibitors. A brief treatment with adriamycin, CGA, or ATA essentially abolishes the ability of Wnt3a to signal to the level of gene transcription (i.e. Lef/Tcf-sensitive gene activation) providing data in good agreement with the results from study of ␤-catenin accumulation (Fig. 5B). Treatment of cells with siRNA targeting either IP3K-B or IPMK provokes a substantial, but not complete, reduction in the ability of Wnt3a to activate Lef/Tcf-sensitive gene transcription (Fig. 5C). Pretreating the Fz1-expressing cells with siRNAs that target both IP3K and IPMK in combination achieved greater suppression of the Wnt3astimulated gene activation than does either siRNA treatment alone (Fig. 5C).
Knockdown of IPMK suppresses ␤-catenin accumulation and Lef/Tcf-sensitive gene transcription in response to Wnt3a stimulation, prompting the query, what effect might overexpression of IPMK have on the transcriptional response? Clones expressing Fz1 were transiently transfected with expression vector harboring a Myc-tagged version of IPMK, enabling assay of the level of relative expression by immunoblotting (Fig. 5D). Analysis of Wnt3a-stimulated activation of Lef/Tcf-sensitive reporter gene activity reveals increasing reporter gene activity in response to Wnt3a as the level of exogenous IMPK increases (Fig. 5D). Likewise, the knockdown of IPMK provoked by siRNA treatment targeting this enzyme blunts Wnt3a action at this level; expression of the exogenous IPMK "rescues" the Wnt response (Fig. 5E). Finally, if the sources of IP 5 involve IP3K and IPMK and the accumulation of IP 5 is essential to the operation of the Wnt/␤-catenin canonical signaling pathway, one would predict that suppression of the expression of IPMK and IP3K-B would suppress not only IP 5 accumulation and gene reporter activity, but also the ability of Wnt3a to promote the formation of PE (Fig. 5F). We tested this hypothesis, making use of the monoclonal antibody to stain for the expression of the PE marker protein TROMA-1 antigen (Cytokerotin Endo A). In the absence of Wnt3a stimulation, TROMA-1 antigen is not present, because of the embryonic character of these cells (Fig.  5F). After treatment of the Fz1-expressing cells with Wnt3a (15 ng/ml) and a 4-day period post-treatment, positive staining of the cells demonstrates Wnt3a-induced PE formation. In those cells treated with siRNA targeting both IPMK and IP3K in combination, Wnt3a stimulates little IP 5 accumulation (Fig. 4D), little Lef/Tcf-sensitive gene transcription (Fig. 5C), and virtually no PE formation (Fig. 5F).
IP 5 Enhances CK2 Activity-We sought to explore the link between Wnt3a-stimulated accumulation of IP 5 and the activation of CK2, following up earlier reports that inositol polyphosphates can regulate the activity of CK2 in vitro. We tested the effects of specific enzyme inhibitors that regulate the Wnt3a-stimulated accumulation of IP 5 on the activity of CK2 in Fz1-expressing F9 cells (Fig. 6). The following enzymes were targeted with inhibitors: PLC␤ (1-octadecylrac-glycero-3-phosphocholine, ET-18-OH), IPMK (CGA), and IP3K/IPMK (ATA). Cells were treated with the inhibitors for 30 min prior to stimulation with purified Wnt3a (Fig. 6A). As reported earlier (23), Wnt3a stimulates CK2 activity in F9 cells expressing Fz1 (Fig. 6A). Inhibition of the generation of IP 5 at the levels of PLC␤, IP3K, or IPMK abolishes the ability of Wnt3a to increase CK2 activity (Fig. 6A). In vitro experiments were performed in parallel, making use of CK2 pulled-down from whole cell lysates and assayed for the effects of various inositol polyphosphates on CK2 activity (Fig. 6B). IP 4 , IP 5 , and IP 6 , but not IP 3, stimulated increased activity of CK2 (Fig. 6B), confirming earlier studies (25). Finally, we examined the concentration dependence of the activation of CK2 by IP 5 (0 -80 M, Fig. 6C). Increasing concentrations of IP 5 stimulates a concentration-dependent increase in CK2 activity. Half-maximal stimulation of CK2 activity is observed at ϳ10 M IP 5 .
IP 5 Suppresses the Activity of GSK3␤-To complete our analysis of likely targets for regulation by inositol polyphosphates, we performed experiments parallel to those performed on CK2, only targeting instead GSK3␤. GSK3␤ plays a pivotal role in the Wnt canonical pathway (48 -50). Although there was no indication in the literature that GSK3␤ was regulated by inositol polyphosphates, we tested this possibility experimentally. Wnt3a acts to inhibit GSK3␤, the enzyme responsible for the phosphorylation of ␤-catenin, catalyzing a pathway culminating in the destruction of the phospho-␤-catenin (51). Treating the Fz1-expressing F9 cells with Wnt3a suppresses the activity of GSK3␤ (Fig. 7A). Treating the cells with inhibitors of PLC␤ (ET-18-OH), IPMK (CGA), and IP3K/ IPMK (ATA) prior to treatment with Wnt3a abolished the ability of Wnt3a to inhibit GSK3␤ activity (Fig. 7A), much as these inhibitors abolished the ability of Wnt3a to stimulate CK2 activity (Fig. 6A).
GSK3␤ pulled-down from F9 cell extracts next were treated in vitro with various inositol polyphosphates and the enzymatic activity measured (Fig. 7B). Of the four inositol polyphosphates tested, only IP 5 displayed the ability to inhibit GSK3␤ activity. The concentration dependence of the inhibitory effect of IP 5 on GSK3␤ activity in vitro was tested. IP 5 inhibits GSK3␤ activity in a concentration-dependent manner, displaying half-maximal inhibition of enzyme activity at ϳ10 M (Fig. 7C). Thus, IP 5 appears to coordinate a key pair of enzymes essential in the Wnt canonical pathway, stimulating CK2 and inhibiting GSK3␤.

DISCUSSION
Metabolism of membrane-anchored phosphatidylinositol 4,5bisphosphate by PLC␤ yields the second messengers diacylglycerol and a water-soluble 1-myo-inositol polyphosphate IP 3 . PLC␤ is an effector for many members of the superfamily of G protein-coupled receptors (36). Frizzleds display seven transmembrane segments, N-terminal sequences on the exofacial side of the cell membrane, and C-terminal sequences on the cytoplasmic face of the cell membrane where they can act as substrates for various protein kinases involved in signaling. Despite the level of homology that exists between Frizzleds and other members of the G protein-coupled receptors, it was only recently demonstrated that Frizzleds are bona fide G proteincoupled receptors themselves (7). Genetic evidence from Drosophila (10) and biochemical data from mammalian cells (9) demonstrate the essential role of heterotrimeric G proteins in coupling Frizzleds to FIGURE 5. Signaling in Wnt canonical pathway to ␤-catenin, gene transcription, and PE formation. F9 clones expressing the Fz1 and Super8X TOPflash were stimulated by Wnt3a. Lef/Tcf-dependent transcription (B-E) and accumulation of cytoplasmic ␤-catenin (A) were measured. The results are displayed as mean Ϯ S.E., obtained from at least three separate experiments. Endoderm formation (F) was analyzed at day 4 after Wnt3a treatment. A, effects of pretreating cells with either ATA or adriamycin (as described in the legend to Fig. 4) on the accumulation of cytoplasmic ␤-catenin in the absence versus the presence of Wnt3a. B, clones pre-treated with ATA, CGA, and adriaymycin were stimulated without or with Wnt3a and the Lef/Tcf-sensitive luciferase reporter activity determined. **, p Ͻ 0.001, for the difference from cells treated only with Wnt3a (no inhibitor). C, clones were treated with siRNA for 48 h to knockdown either IPMK or IP3K-B, or both prior to Wnt3a stimulation. The Lef/Tcf-sensitive luciferase reporter activity was measured. **, p Ͻ 0.001, versus cells treated with Wnt3a only, not treated with siRNAs targeting IP3K, IPMK, or both in tandem. #, p Ͻ 0.01, versus cells treated with either "siRNA to IP3K-B" alone or "siRNA to IPMK" alone. D, Fz1-expressing cells were transiently transfected for 24 h with increasing amounts of plasmid harboring Myc-tagged IPMK as indicated. Clones were stimulated by Wnt3a and Lef/Tcfsensitive luciferase transcriptional activity was measured. The expression of Myc-tagged IPMK in transfected clones was analyzed by immunoblotting (IB) with anti-Myc antibody (lower panel). *, p Ͻ 0.01; **, p Ͻ 0.001, for the difference from Wnt3a-treated control cells. E, Fz1-expressing cells were treated with siRNA for 24 h to knockdown IPMK. These knockdown cells were transiently transfected to express exogenous, Myc-tagged IPMK. Cells were stimulated without or with Wnt3a and the Lef/Tcf-sensitive luciferase reporter gene activity measured. *, p Ͻ 0.01; **, p Ͻ 0.001 versus Wnt3a-treated control cells. F, cells were treated with either control siRNA or siRNAs targeting IPMK as well as IP3K-B in combination for 48 h. Cells then were treated without or with Wnt3a (15 ng/ml) for 4 days and the formation of PE examined by immunostaining of the fixed cells with TROMA-1 antibody, a secondary fluorescently labeled antibody, in tandem with indirect immunofluoresence. PC, phase-contrast. IIF, indirect immunofluorescence. Bar, 20 m. their effectors in Wnt signaling pathways, including the Wnt/ ␤-catenin canonical pathway (9), the non-canonical Wnt/ cGMP, Ca 2ϩ pathway (35,52), and planar cell polarity (10,53,54). In the current work, we show that Wnt/␤-catenin signaling mediated by Frizzled-1 provokes the breakdown of phosphatidylinositol 4,5-bisphosphate and elevation of intracellular levels of IP 5 .
Activation of Wnt/cGMP, Ca 2ϩ non-canonical pathway by Wnt5a operating via Fz2 (not Fz1) stimulates accumulation of IP 3 (which is linked to the Ca 2ϩ mobilization stimulated by Wnt5a), whereas activation of Fz1-expressing cells by Wnt3a transiently generates IP 3 and, in turn, can be quickly metabolized further to more highly phosphorylated forms, including IP 4 , IP 5 , and IP 6 . We show that the accumulation of IP 5 , but not accumulation of IP 3 , is stimulated by Wnt3a operating via Fz1 (not Fz2) (Fig. 8). The metabolism of IP 3 to IP 5 occurs by the combined action of at least two enzymes, IPMK and IP3K in mouse F9 cells. Targeted knockdown by siRNA or chemical inhibition of these enzymes inhibits the accumulation of IP 5 in response to Wnt3a as well as the ability of Wnt3a to stimulate Fz1-mediated ␤-catenin/ Lef-Tcf signaling pathway. A recent report by Frederick et al. (37) on IPMK-null mice showed the importance of IPMK in embryogenesis. Targeted disruption of ipmk locus in both alleles in mice blocks IP 5 accumulation and it results in early embryo lethality.
It was an earlier observation from in vitro studies of CK2 activity (23,25) that prompted us to explore the effects of Wnt3a on phosphatidylinositol signaling in F9 cells, cells amenable to biochemical manipulation and analysis. We observed that Wnt3a stimulates a dose-dependent and timedependent accumulation of IP 5 in cells, a process that could be blocked by enzyme inhibitors at the level of PLC␤, IP3K, and IPMK. How then, we asked, is the accumulation of IP 5 related to the output of the Wnt canonical pathway? IP 3 has a remarkable and diverse set of functions in cell signaling and physiology (55,56); far less is known about the functions of IP 5 . Accumulation of inositol polyphosphates has been shown to function in cell proliferation, apoptosis, differentiation (57)(58)(59), and control of left-right symmetry in zebrafish development (60). The current work expands upon these earlier observations, providing a demonstration that Wnt canonical signaling includes phosphatidylinositol signaling to the level of IP 5 .
Biochemically, the basis for the role of IP 5 in Wnt canonical signaling focuses upon two target enzymes essential to FIGURE 6. IP 5 enhances CK2 activity in vitro. A, F9 clones expressing Fz1 were pre-treated with inhibitors to PLC␤ (ET-18-OH), IPMK (CGA), or IPMK and IP3K (ATA) for 20 min prior to the stimulation by Wnt3a. Following immunoprecipitation of CK2␣, CK2 activity was measured by in vitro kinase assay as described under "Experimental Procedures." The results are displayed as mean Ϯ S.E., derived from at least three separate experiments. *, p Ͻ 0.001 versus untreated control cells; #, p Ͻ 0.001 versus Wnt3a-treated control cells. B and C, the CK2␣ was immunoprecipitated from whole cell lysates prepared from wild-type F9 cells. The immune complex was resuspended in kinase reaction buffer and aliquots with equal volumes were made. Kinase reactions were conducted in the absence (Ϫ) or presence of 20 M of inositol phosphates (IP 3 , IP 4 , IP 5 , or IP 6 as indicated) (B), or in the presence of increasing concentrations of IP 5 as indicated (C). CK2 activities were compared and the CK2 activity obtained from samples in the absence of inositol phosphates was set as 1. *, p Ͻ 0.01; **, p Ͻ 0.001 versus samples untreated with inositol polyphosphates. the operation of this pathway (Fig. 8), i.e. CK2 and GSK3␤. With respect to CK2 activity, we extended the earlier studies by probing the effects of agents that disrupt the Wnt3a-stimulated accumulation of IP 5 on CK2 activity in vivo. IP 5 accumulation is stimulated by Wnt3a and this response can be blocked by chemical inhibitors of PLC␤, IP3K, and IPMK. Similarly, these inhibitors block CK2 activation in response to Wnt3a stimulation. We demonstrate in vitro that IP 5 as well as other inositol polyphosphates, including IP 4 and IP 6 are capable of stimulating CK2 activity in CK2 pulled-down from F9 cells. Additional studies performed with purified CK2 failed to demonstrate an IP 5 -dependent activation (not shown) suggesting that IP 5 is not likely acting directly on CK2 or that the activation may require additional IP 5 -binding protein(s). Our analysis extended to the study of the effects of IP 5 on GSK3␤ activity also, both in vivo and in vitro. IP 5 appears to inhibit GSK3␤ activity, because inhibition of Wnt3a-stimulated IP 5 accumulation abolishes the ability of Wnt3a to suppress GSK3␤ activity. In vitro assay of GSK3␤ activity in pull-downs of GSK3␤ displays sensitivity to IP 5 . IP 5 inhibited GSK3␤ activity in a dose-dependent manner, FIGURE 7. IP 5 suppresses GSK3␤ activity in vitro. A, F9 clones expressing Fz1 were pre-treated with inhibitors to PLC␤ (ET-18-OH), IPMK (CGA), or IPMK and IP3K (ATA) for 20 min prior to the stimulation by Wnt3a. GSK3␤ was immunoprecipitated from whole lysates and its activity was measured by in vitro kinase assay as described under "Experimental Procedures." *, p Ͻ 0.001 versus untreated, control cells; #, p Ͻ 0.001 versus Wnt3a-treated control cells. B and C, the GSK3␤ was immunoprecipitated from whole cell lysates prepared from wild-type F9 cells. The immune complex was resuspended in kinase reaction buffer and aliquots with equal volumes were made. B, kinase reactions were conducted in the absence (Ϫ) or presence of 20 M inositol phosphates (IP 3 , IP 4 , IP 5 , or IP 6 as indicated). C, kinase activity was measured in the presence of increasing concentrations of IP 5 as indicated. GSK3␤ activity obtained from untreated control cells was set as 1. **, p Ͻ 0.001, versus samples untreated with inositol polyphosphates. FIGURE 8. IP 5 : pivotal regulation of GSK3␤ and CK2 in Frizzled1/␤-catenin/Lef-Tcf signaling pathway. Wnt3a binds to Frizzled1 (FZ1), a member of the superfamily of G protein-coupled receptors, and thereby activates G␣ q , which in turn activates PLC␤. Activation of PLC␤ leads to hydrolysis of phosphatidylinositol 4,5-bisphosphate and generates IP 3 and diacylglyceride (DAG). IP 3 is quickly converted to IP 5 , the combined action of IPMK and IP3K, and does not appreciably bind to the IP 3 receptor altering intracellular Ca 2ϩ . Increase of intracellular IP 5 accumulation stimulated (indirectly) the activation of CK2 while inhibiting (indirectly) GSK3␤ activity. Wnt-stimulated ␤-catenin stabilization and Lef/Tcf-sensitive transcriptional activation by Wnt3a are dependent upon the accumulation of IP 5 . KD, knock down by siRNA; IP 3 R, IP3 receptor.
the inhibition was half-maximal at ϳ10 M. Two research articles reported the cellular concentration of IP 5 . Szwergold et al. (61) reported IP 5 bulk measurements of IP 5 levels in mammalian tissue at 5-15 M. Georg Mayr's (62) group found the IP 5 in HL-60 cells to be ϳ35 M in the basal level. In response to chemotacticpeptide (formyl-methionylleucyl-phenylalanine) stimulation, IP 5 concentrations in these cells rise to ϳ50 M. Based up these reports, our finding that in vitro half-maximum responses of CK2 and GSK3␤ occur at ϳ10 M is physiologically relevant. Like the studies performed with purified CK2, parallel studies on the effects of inositol polyphosphates on purified GSK3␤ in vitro show no direct effect by IP 5 .
This work reveals several novel aspects of Wnt canonical signaling. Our studies are the first to show the existence of a Wnt/␤-catenin "canonical" pathway regulating phosphatidylinositol signaling. Wnt5a regulation of phosphatidylinositol signaling via a Frizzled-2 mediated "non-canonical" pathway was reported earlier (34,47). IP 5 accumulates in response to Wnt3a, whereas IP 3 accumulates in response to Wnt5a signaling. Not only does IP 5 accumulation occur in response to Wnt3a, but it also is an essential component of signaling of the canonical pathway to the level of ␤-catenin accumulation, Lef/ Tcf-sensitive transcription, and PE formation in F9 cells. Finally, we show that IP 5 plays a critical role at the level of CK2 and GSK3␤, stimulating the former while inhibiting the latter. Thus, phosphatidylinositol signaling is essential to Wnt signaling via the canonical pathway.