Atypical Protein Kinase Cι Is Required for Wnt3a-dependent Neurite Outgrowth and Binds to Phosphorylated Dishevelled 2*

Background: Wnt3a-dependent neurite outgrowth is associated with Dishevelled phosphorylation. Results: PKCι is required for Wnt3a-induced neurite extension. PKCι binds wild-type Dishevelled, but not an inactive phosphorylation-deficient Dishevelled mutant. Conclusion: PKCι mediates Wnt3a-dependent neurite outgrowth; Dishevelled phosphorylation is required for PKCι interaction. Significance: PKCι has key role in Wnt3a-induced neurite outgrowth; Dvl phosphorylation at defined sites is critical for PKCι association. Previously we reported that Wnt3a-dependent neurite outgrowth in Ewing sarcoma family tumor cell lines was mediated by Frizzled3, Dishevelled (Dvl), and c-Jun N-terminal kinase (Endo, Y., Beauchamp, E., Woods, D., Taylor, W. G., Toretsky, J. A., Uren, A., and Rubin, J. S. (2008) Mol. Cell. Biol. 28, 2368–2379). Subsequently, we observed that Dvl2/3 phosphorylation correlated with neurite outgrowth and that casein kinase 1δ, one of the enzymes that mediate Wnt3a-dependent Dvl phosphorylation, was required for neurite extension (Greer, Y. E., and Rubin, J. S. (2011) J. Cell Biol. 192, 993–1004). However, the functional relevance of Dvl phosphorylation in neurite outgrowth was not established. Dvl1 has been shown by others to be important for axon specification in hippocampal neurons via an interaction with atypical PKCζ, but the role of Dvl phosphorylation was not evaluated. Here we report that Ewing sarcoma family tumor cells express PKCι but not PKCζ. Wnt3a stimulated PKCι activation and caused a punctate distribution of pPKCι in the neurites and cytoplasm, with a particularly intense signal at the centrosome. Knockdown of PKCι expression with siRNA reagents blocked neurite formation in response to Wnt3a. Aurothiomalate, a specific inhibitor of PKCι/Par6 binding, also suppressed neurite extension. Wnt3a enhanced the co-immunoprecipitation of endogenous PKCι and Dvl2. Although FLAG-tagged wild-type Dvl2 immunoprecipitated with PKCι, a phosphorylation-deficient Dvl2 derivative did not. This derivative also was unable to rescue neurite outgrowth when endogenous Dvl2/3 was suppressed by siRNA (González-Sancho, J. M., Greer, Y. E., Abrahams, C. L., Takigawa, Y., Baljinnyam, B., Lee, K. H., Lee, K. S., Rubin, J. S., and Brown, A. M. (2013) J. Biol. Chem. 288, 9428–9437). Taken together, these results suggest that site-specific Dvl2 phosphorylation is required for Dvl2 association with PKCι. This interaction is likely to be one of the mechanisms essential for Wnt3a-dependent neurite outgrowth.

that site-specific Dvl2 phosphorylation is required for Dvl2 association with PKC. This interaction is likely to be one of the mechanisms essential for Wnt3a-dependent neurite outgrowth.
The Wnts comprise a large family of secreted glycoproteins that have many activities during embryonic development and promote tissue homeostasis in the adult. Wnt signals control cell proliferation, differentiation, survival, motility, and polarity. They also have higher order effects on the developing embryo by regulating tissue patterning, organogenesis, and specification of the body plan (1,2).
Wnts are particularly important in the development of the nervous system, where they contribute to neural tube closure, formation of specific brain structures, and the induction and migration of neural crest cells (3,4). Wnt signaling stimulates axonal remodeling, pathfinding, dendritic arborization, and neuronal connectivity in the central nervous system (5)(6)(7). Many Wnt signaling components have been implicated in neurite outgrowth (7)(8)(9). For instance, Wnt7b induces dendritic morphogenesis in mouse hippocampal neurons via a noncanonical pathway including Dvl, Rac, and JNK (10). The gene encoding the Wnt receptor, Frizzled3 (Fzd3), is required for the development of major fiber tracts in the rostral central nervous system (11), and targeted disruption of Fzd3 caused axonal growth and guidance defects (12). Derailed/Ryk is another Wnt receptor that regulates axon guidance in a variety of contexts (13)(14)(15)(16).
Neurons have highly polarized structures, and establishment of cell polarity is an essential aspect of neurite outgrowth (17). It is widely accepted that the polarity complex consisting of atypical protein kinase C (aPKC, that is PKC isoforms PKC or PKC/), 3 Par3, and Par6 has a critical role in neuronal polarity (18 -21), a process that involves neurite extension followed by the differ-entiation of a neurite into an axon (22). Recent studies have indicated that the aPKC-Par3-Par6 complex participates in neurite outgrowth before and after axon differentiation, as well as in dendrite formation (23)(24)(25). Of note, Wnt4 attracts the outgrowth of commissural axons by a mechanism that relies on aPKC (24). Furthermore, Wnt5a promotes axon differentiation in cultured hippocampal neurons via an interaction of Dvl1 with PKC (25).
Previously, we used an Ewing sarcoma family of tumors (ESFT) cell line, TC-32, to study Wnt3a-dependent neurite outgrowth (26). ESFT cells are small, round, and poorly differentiated with characteristics of primitive neuroectoderm (27)(28)(29). Gene microarray data indicated that ESFT cells have expression patterns that resemble those of neuroectoderm and endothelial cells (30). A small fraction (10 -15%) of TC-32 cells have long neurites in the basal state, whereas Wnt3a treatment elicits neurite outgrowth in 50 -75% of cells within 3 h. This response is mediated by noncanonical Wnt rather than ␤-catenin signaling and involves a mechanism that requires Fzd3 and Dvl2/3 expression, as well as JNK activation (26). The centrosomal localization of casein kinase 1␦ (CK1␦), a kinase that phosphorylates Dvl proteins, was also shown to be necessary for neurite outgrowth (31), although the functional relevance of Dvl phosphorylation was not established.
The present study was undertaken to further delineate the mechanisms responsible for Wnt3a-dependent neurite outgrowth in ESFT cells. We determined that PKC, but not PKC, was expressed in TC-32 cells. Wnt3a induced the phosphorylation of PKC, and siRNA knockdown of PKC-blocked neurite extension. Furthermore, neuritogenesis was suppressed by aurothiomalate (ATM), a specific inhibitor of PKC binding to Par6 (32). Wnt3a stimulated an association of both endogenous and FLAG-tagged WT Dvl2 with PKC. However, a FLAGtagged Dvl2 mutant with alanine substitutions at a cluster of CK1 phosphorylation sites in the C-terminal domain failed to interact with PKC. In an accompanying manuscript (43), we demonstrated that this mutant was unable to support neurite outgrowth when expression of endogenous Dvl2/3 was suppressed. These results strongly suggest that Dvl phosphorylation at these sites is necessary for the interaction with PKC and reinforce the idea that the Dvl/aPKC interaction is important for Wnt-dependent neurite outgrowth.

EXPERIMENTAL PROCEDURES
Recombinant Protein and Chemicals-Recombinant Wnt3a was purchased from R&D Systems (Minneapolis, MN). ATM was from Taylor Pharmaceuticals (Decatur, IL).
Cell Culture-TC-32 cells were seeded on cell culture dishes, cluster plates, or glass coverslips that had been precoated with type I collagen solution (Sigma-Aldrich) as described (26). HEK293 cells were maintained in DMEM supplemented with 10% fetal bovine serum.
Immunofluorescent Analysis-TC-32 cells were cultured in complete growth medium on 12-mm-diameter glass coverslips (Fisher) coated with collagen. To visualize phalloidin and phospho-PKC/, formaldehyde fixation and staining were performed as previously described (31). Anti-phospho-PKC/ was used at a 1:250 dilution. Double staining of phospho-PKC and pericentrin was performed either with formaldehyde or methanol (MeOH) fixation (31). Total PKC was detected with mouse anti-PKC (1:500 dilution) when cells were fixed in formaldehyde and with rabbit anti-PKC (sc-216, 1:500), which recognizes PKC, when cells were fixed in methanol because of high background with mouse anti-PKC antibody following methanol fixation.
Cell Imaging-Fluorescent images were collected with a Zeiss 510 LSCM, using a 63ϫ objective (Carl Zeiss Inc., Thornton, NY). Zeiss LSM image browser Version 4.0.0.157 was used for image processing, and composite figures were prepared with Adobe PhotoShop CS3 version 10.0.1 (Adobe Systems Inc., San Jose, CA).
Quantitative Analysis of Neurite Outgrowth-Stimulation of neurite outgrowth was monitored as previously described (26).
Immunoblotting-To examine the effect of Wnt3a on PKC phosphorylation, 80 -90% confluent monolayers of TC-32 cells that had been seeded in 6-or 12-well cell culture plates were serum-starved overnight. After incubation with Wnt3a for the indicated time, the cells were rinsed twice with PBS, lysed with buffer, and processed for SDS-PAGE and Western blot analysis as previously described (26). For immunoblot analysis to verify siRNA knockdown of endogenous proteins, TC-32 cells that had been transfected with siRNA were seeded in 6-or 12-well cell culture plates and harvested 48 h after transfection.
Immunoprecipitation-TC-32 cells were harvested with immunoprecipitation buffer (20 mM Tris-HCl, pH 8.0, 10 mM EDTA, 1 mM EGTA, 150 mM NaCl, 0.2% Triton-X, 0.2% Nonidet P-40), and cell lysates were homogenized by passing through 23-gauge syringe needles. Cell lysates were centrifuged at 20,817 ϫ g for 15 min at 4°C. Supernatant was transferred to another tube, and protein concentration was determined. Cell lysates (1 mg) were precleared by incubation with 1 g of rabbit IgG (sc-2027; Santa Cruz Biotechnology) and 20 l of protein A/G Plus-agarose (sc-2003; Santa Cruz Biotechnology) for 2 h at 4°C, followed by centrifugation at 20,817 ϫ g for 10 min. At the same time, 1 g of rabbit Dvl2 antibody and 20 l of protein A/G-agarose were combined and incubated for 2 h at 4°C with rotation in immunoprecipitation buffer, followed by centrifugation at 956 ϫ g for 5 min. Precleared cell lysate was combined with resuspended Dvl2 antibody/protein A/G-agarose mixture and incubated overnight at 4°C with rotation. After 18 h, the samples were washed with immunoprecipitation buffer three times, combined with sample loading buffer, heated at 95°C for 10 min, and resolved by SDS-PAGE.
Recombinant DNA-pCS2ϩ FLAG-mDvl2 WT was kindly provided by Dr. Xi He (Harvard University). pCS2ϩFLAG-mDvl2 P4m (S594A, S595T, S597A, T604A) was generated by site-directed mutagenesis according to manufacturer's protocol (QuikChange II; Agilent Technologies, Inc. Santa Clara, CA). Mutation was verified by DNA sequence analysis in the DNA sequencing mini core facility at the National Cancer Institute. Lentiviral constructs, pCMV32 FLAG- tagged mouse Dvl2 WT and Dvl2 P4m were generated as described (43).
Statistical Analysis-Except where indicated, the significance of differences in data were determined with Student's t test. The differences were considered to be significant when the p value was less than 0.05.

PKC Is Expressed in ESFT Cells and Phosphorylated in
Response to Wnt3a-To investigate the potential role of aPKC isoforms in Wnt3a-dependent neurite outgrowth in ESFT cells, we initially examined the expression of these proteins using antisera intended to recognize PKC or PKC. Immunoblotting of TC-32 cell lysates with PKC mAb revealed a band of the expected size (74 kDa) (Fig. 1A). A comparable band was observed with one PKC polyclonal antiserum but not with another (Fig. 1, A and B). To resolve this ambiguity, the cells were treated with siRNA reagents specific for PKC versus PKC, and Western blot analysis of the TC-32 lysates revealed that the bands recognized by the antibodies corresponded to PKC (Fig. 1B).
Incubation of TC-32 cells with recombinant Wnt3a at 100 ng/ml stimulated the phosphorylation of PKC, which was variably detected within 5-10 min and consistently seen at 30 min through 3 h (Fig. 1, C and D). The phospho-specific PKC// antibody was directed against T412 in PKC, a residue in the activation loop that is phosphorylated in the activated form of the enzyme (33,34). These results suggested that exposure to Wnt3a activated PKC in a time frame consistent with its possible function in neurite extension (26).
Intracellular Distribution of Phosphorylated PKC-Confocal microscopy of TC-32 cells showed little evidence of phosphorylated PKC (pPKC) in control cells ( Fig. 2A, upper row). However, cells treated for 1 h with 100 ng/ml of Wnt3a displayed a punctate pattern of staining in the cell body and the neurite ( Fig. 2A, lower row). The specificity of pPKC immunofluorescent signal was confirmed by knockdown with PKC siRNA (Fig. 2B). A punctate pattern also was seen with antibody to total PKC (Fig. 2C) and was absent following treatment with PKC siRNA (Fig. 2D). Interestingly, signal for pPKC and total PKC exhibited a cytosolic and pericentrosomal distribution (Fig. 3). Wnt3a treatment increased the percentage of cells with pericentrosomal pPKC and the intensity of the stain there ( Fig.  3B and data not shown). Previously, we reported that CK1␦ staining was strongest at the centrosome, where it was required for neurite outgrowth induced by Wnt3a (31). Although CK1␦ siRNA did not block PKC phosphorylation following Wnt3a treatment, it did inhibit the pericentrosomal distribution of pPKC (data not shown). Knockdown of both CK1␦ and the  MARCH 29, 2013 • VOLUME 288 • NUMBER 13 JOURNAL OF BIOLOGICAL CHEMISTRY 9441 highly related isoform CK1⑀ increased the basal level of phospho-PKC in some experiments, but CK1⑀ siRNA had no effect on the pericentrosomal localization of PKC (data not shown).

Inhibition of PKC Expression or Interaction with Par6
Blocked Wnt3a-dependent Neurite Outgrowth-In the absence of Wnt3a, ϳ10% of TC-32 cells have a neurite extending at least one cell diameter in length, and this proportion was unaffected by a negative control siRNA directed against luciferase. When cells that had been treated with luciferase siRNA were subsequently exposed to Wnt3a (100 ng/ml) for 3 h, the percentage with a long neurite increased 4 -5-fold (Fig. 4, A and B). In contrast, the induction of neurite outgrowth by Wnt3a was completely abrogated in cells that had received PKC siRNA (Fig. 4, A and B). Efficient and specific knockdown of PKC protein was confirmed by Western blot analysis (Figs. 1B and 4C). These results indicate that expression of PKC is required for Wnt3a-dependent neurite outgrowth in TC-32 cells.
ATM is a small molecule that disrupts the interaction of PKC with Par6 and thereby blocks the activity of the PKC-Par3-Par6 polarity complex (32). Using ATM at concentrations shown to inhibit the polarity complex in other cells, we observed a marked suppression of Wnt3a-dependent neurite extension without any negative impact on the viability of the cells (Fig. 5). This reinforces the evidence from siRNA experiments that PKC activity is necessary for neurite outgrowth in ESFT cells following Wnt3a treatment.
Wnt3a Stimulated the Interaction of PKC with Dvl2-Prompted by a report that PKC interacts with Dvl1 to mediate axonal specification in hippocampal neurons (25), we explored the possibility that Wnt3a causes an association of PKC with Dvl. Although Dvl1 protein was not detected in TC-32 cells, Wnt3a promoted the co-immunoprecipitation of endogenous PKC with Dvl2 (Fig. 6). After a 3-h treatment with Wnt3a (100 ng/ml), the cells were lysed, and proteins were precipitated with Dvl2 antiserum and immunoblotted with anti-PKC. Little or no PKC was seen in the pellet from control cells, but PKC was observed in the immunoprecipitated sample from cells that had been incubated with Wnt3a. As previously observed (31),

PKC/Dvl2 in Wnt3a-dependent Neurite Outgrowth
Wnt3a treatment also elicited an upward shift in the electrophoretic mobility of Dvl2 that corresponds to Dvl2 phosphorylation (Fig. 6). Occasionally the mobility shift was accompanied by an increase in Dvl2 band intensity, although that was not a consistent finding.
Phosphorylation-deficient Dvl2 Derivative Did Not Co-immunoprecipitate with PKC-In a parallel study, we identified CK1 phosphorylation sites in the C-terminal domain of Dvl2 that were critical for its Wnt-induced electrophoretic mobility shift (43). In contrast to Dvl2 WT, a mutant containing alanine substitutions at Ser 594 , Thr 595 , Ser 597 , and Thr 604 (FLAG-Dvl2 P4m) was unable to rescue the neurite outgrowth phenotype when expression of endogenous Dvl was suppressed by siRNA (43). This implied that phosphorylation at one or more of the sites is essential for neurite outgrowth. To test the hypothesis that phosphorylation of these residues is required for Dvl2 interaction with PKC, we compared the co-immunoprecipitation of PKC with FLAG-tagged Dvl2 WT versus Dvl2 P4m. The WT protein exhibited a slower electrophoretic mobility than the mutant protein (Fig. 7), consistent with the former being phosphorylated, presumably by endogenous Wnt/Frizzled signaling (43). Despite the fact that the mutant protein was expressed at a higher level than Dvl2 WT, PKC only precipi-tated with Dvl2 WT (Fig. 7). This result strongly suggests that Dvl phosphorylation at one or more of the four mutated sites is required for interaction with PKC. Because Dvl2 P4m was unable to substitute for native protein in the neurite outgrowth assay, this finding also implied that the interaction of PKC with Dvl2 was necessary for Wnt3a-dependent neurite extension in ESFT cells.

DISCUSSION
In this report we demonstrate that PKC is required for Wnt3a-dependent neurite outgrowth in TC-32 cells. This is consistent with earlier studies showing that aPKC was essential for commissural axon guidance induced by Wnt4 and that PKC was necessary for axon specification in hippocampal neurons treated with Wnt5a (24,25). Ectopically expressed Dvl1 co-precipitated with PKC and other members of the Par complex, and their association was important for axon specification (25). We also observed the co-precipitation of epitope-tagged Dvl2 with PKC, as well as a physical association of the endogenous proteins that was stimulated by Wnt3a. Our findings   MARCH 29, 2013 • VOLUME 288 • NUMBER 13

PKC/Dvl2 in Wnt3a-dependent Neurite Outgrowth
indicate that a Wnt-induced phosphorylation-deficient Dvl2 derivative with mutated CK1 phosphorylation sites in its C-terminal domain failed to precipitate with PKC. This reveals a functional significance of Dvl phosphorylation at these sites that is reinforced by the inability of the mutant protein to mediate neurite outgrowth (43). We propose that the interaction between Dvl2 and PKC is required for Wnt3a-dependent neurite outgrowth in TC-32 cells, just as the Dvl1/PKC interaction was found to be critical for axon specification in hippocampal neurons.
The implied significance of Dvl phosphorylation in binding to PKC and in neurite outgrowth is noteworthy because there are few instances in which Dvl phosphorylation at specific sites has been shown to be functionally important. One group reported that phosphorylation of Dvl2 at Thr 206 was necessary for proper orientation of the mitotic spindle (35), and another showed that Dvl2 phosphorylation at Ser 143 and Thr 224 enables binding to polo-like kinase 1 to facilitate primary cilia disassembly (36). Dvl2 phosphorylation mediates binding to Ror2, although the sites of phosphorylation were not determined (37). A recent extensive analysis identified dozens of phosphorylated Ser/Thr residues in Dvl (or the Drosophila ortholog, Dsh). However, these modifications were dispensable for ␤-catenin and planar cell polarity signaling in the Drosophila assays (38). The present study and our accompanying report (43) suggest that Dvl2 phosphorylation, particularly at Ser 594 , Thr 595 , and Ser 597 , is relevant for signaling in a Wnt/aPKC pathway.
In addition to the work by Zhang and co-workers (25), another group identified an interaction between Dvls and Par6, although aPKC was not reported to be part of the complex (39). Wnt5a stimulated Dvl/Par6 binding to the E3-ubiquitin ligases Smurf1 and Smurf2, which facilitated the ubiquitination and proteasomal degradation of Prickle, a core component of planar cell polarity signaling. Smurf binding was dependent on Dvl2 phosphorylation via a PY motif in the Dvl2 C-terminal domain. The association of Dvl/Par6/Smurfs is intriguing because Smurf1 and Smurf2 are essential for neurite outgrowth and neuronal polarity, respectively, due to their regulation of the GTPases RhoA and Rap1B (40,41). This raises the possibility that Wnt-dependent neurite outgrowth involves protein turnover mediated by Dvl2-Par6-Smurf complexes with or without aPKC.
The pericentrosomal localization of pPKC that we observed in TC-32 cells following Wnt3a treatment may have mechanistic relevance. CK1␦ also is concentrated at the centrosome, where it functions in neurite extension (31). Although we have not detected Dvl2 at this site, a transient substrate/kinase interaction might occur there and facilitate Dvl2/PKC binding. Knockdown of CK1␦ by siRNA did not affect Wnt3a-induced PKC phosphorylation, but it did disrupt the pericentrosomal localization of pPKC. This suggests that CK1␦ might interact with pPKC through a direct or indirect mechanism. The pericentrosomal localization of pPKC is provocative because PKC also was identified at the centrosome in hippocampal neurons where it promoted neurite outgrowth in a pathway that involved Aurora A, NDEL1, and the regulation of microtubule dynamics (42).
In summary, we have shown that PKC is required for Wnt3a-dependent neurite outgrowth in the ESFT line TC-32. Combined with its utility as a cell line, our study emphasizes the relevance of the TC-32 model for the study of Wnt-dependent FIGURE 6. Co-immunoprecipitation of endogenous PKC and Dvl2 following Wnt3a treatment. TC-32 cells were incubated Ϯ Wnt3a for 3 h, and cell lysates were prepared for immunoprecipitation with Dvl2 rabbit polyclonal antiserum. Pelleted protein and cell lysates were immunoblotted as indicated. An arrow points to immunoprecipitated PKC. *, IgG heavy chain. **, note increased relative intensity of slower migrating Dvl2 band. IB, immunoblot; IP, immunoprecipitation. FIGURE 7. Co-immunoprecipitation of FLAG-tagged Dvl2 derivatives and endogenous PKC. FLAG-tagged mouse Dvl2 WT or a Wnt-induced phosphorylation mutant containing four alanine substitutions (Dvl2 P4m) was stably expressed in TC-32 using a lentiviral vector. The cells were lysed, and the proteins were precipitated with FLAG antibody. Pellets and cell lysates were immunoblotted as indicated. An arrow points to immunoprecipitated PKC. *, IgG heavy chain. IB, immunoblot; IP, immunoprecipitation. neurite extension. Wnt3a stimulated PKC phosphorylation and its distribution to the neurite and the centrosome. We believe that PKC acts at the tip of the neurite to directly mediate protrusion through established mechanisms (18 -21), although its pericentrosomal distribution might promote neurite outgrowth via effects on microtubule dynamics. We suspect that PKC activity depends on its association with Dvl2. This interaction was blocked when four Ser/Thr residues in the C-terminal domain of Dvl2 were replaced with alanine. Our results strongly suggest that Dvl phosphorylation at these sites, particularly CK1-mediated phosphorylation at Ser 594 , Thr 595 , and Ser 597 , is functionally significant and reinforces the idea that Dvl/aPKC interactions have an important role in Wnt-dependent neurite extension.