Non-canonical Wnt Signaling Induces Ubiquitination and Degradation of Syndecan4

Dynamic regulation of cell adhesion receptors is required for proper cell migration in embryogenesis, tissue repair, and cancer. Integrins and Syndecan4 (SDC4) are the main cell adhesion receptors involved in focal adhesion formation and are required for cell migration. SDC4 interacts biochemically and functionally with components of the Wnt pathway such as Frizzled7 and Dishevelled. Non-canonical Wnt signaling, particularly components of the planar cell polarity branch, controls cell adhesion and migration in embryogenesis and metastasic events. Here, we evaluate the effect of this pathway on SDC4. We have found that Wnt5a reduces cell surface levels and promotes ubiquitination and degradation of SDC4 in cell lines and dorsal mesodermal cells from Xenopus gastrulae. Gain- and loss-of-function experiments demonstrate that Dsh plays a key role in regulating SDC4 steady-state levels. Moreover, a SDC4 deletion construct that interacts inefficiently with Dsh is resistant to Wnt5a-induced degradation. Non-canonical Wnt signaling promotes monoubiquitination of the variable region of SDC4 cytoplasmic domain. Mutation of these specific residues abrogates ubiquitination and results in increased SDC4 steady-state levels. This is the first example of a cell surface protein ubiquitinated and degraded in a Wnt/Dsh-dependent manner.

Cell adhesion and migration are key processes for embryogenesis, inflammatory response, tissue repair, and cancer. Cell migration is a very dynamic process that requires the continuous assembly and disassembly of the cell surface complexes mediating cell-cell and cell-extracellular matrix adhesion (1,2). Integrins and Syndecans are the main cell adhesion receptors involved in focal adhesion (FA) 5 formation and cell-extracellular matrix communication (3). FA formation of cells plated on fibronectin requires the specific engagement and synergism between ␣5␤1 integrin and Syndecan4 (SDC4) (3,4).
SDC4 is a cell surface heparan sulfate proteoglycan that activates intracellular signaling cascades through activation of PKC␣ (4). SDC4 overexpression or knockout compromises cell migration in normal (5) and tumorigenic cells (6). More recently it was demonstrated that SDC4 is required for localization of Rac1 and membrane protrusion formation at the leading edge to allow celldirected migration (7). In normal and tumorigenic cells, endocytosis and recycling of integrins modulate FA turnover and proper cell migration (8,9). Which extracellular cues regulate the stability and turnover of FA receptors, in particular of SDC4, is a key question to understand cell migration.
Wnt proteins constitute a large family of secreted glycoproteins that can activate diverse intracellular signaling pathways in a context-dependent manner. Wnt5a is known to promote polarized cell migration by activating a non-canonical Wnt pathway through the tyrosine kinase Ror2 receptor and c-Jun N-terminal kinase (10 -12). Wnt5a increases cell migration and invasiveness of tumorigenic cells by activation of FA kinase, PKC, Rac1, and promoting paxillin turnover (10). Modulation of FA receptor turnover remains a possible mechanism for Wnt5a-induced migration.
Gastrulation, in particular studies of convergence and extension (CE) in Xenopus and zebrafish embryos, has been a fertile ground to understand how growth factor signaling regulates cell migration in vivo. CE is a process of cell-cell intercalation driven by polarized protrusive activity of dorsal mesodermal cells during Xenopus gastrulation, which allows proper narrowing and elongation of the vertebrate embryo (13). FA components, including ␣5␤1 integrin, fibronectin, and SDC4, are essential for CE in Xenopus embryos (14 -16), supporting the importance of cell-extracellular matrix interaction in embryonic morphogenesis. Non-canonical Wnt signaling, particularly components of the planar cell polarity signaling pathway, regulates gastrulation in Xenopus and zebrafish embryos (17). Gain-and loss-of-function experiments have demonstrated that Wnt5a and Wnt11 regulate CE (18,19). Non-canonical Wnt signaling controls gastrulation through modulation of cell adhesive properties (20). The first indication in this sense came from overexpression studies showing that Wnt5a blocks cell reaggregation of dissociated dorsal mesodermal cells from Xenopus gastrula stage embryos (18). More recently, studies using atomic force microscopy showed that Wnt11 modulates the attachment and cohesiveness of dorsal endomesodermal cells through E-cadherin endocytosis (21). A role for non-canonical Wnt signaling in the modulation of the FA receptor has not been demonstrated.
In addition to its function in FA, SDC4 regulates non-canonical Wnt signaling (16). SDC4 interacts biochemically and func-tionally with the Wnt receptor Frizzled7 (Fz7) and the intracellular multidomain protein Dishevelled (Dsh) and modulates this signaling pathway in a fibronectin-dependent manner (16). This prompted us to evaluate the effect of non-canonical Wnt signaling in SDC4. We have found that Wnt5a reduces cell surface levels and promotes degradation of SDC4 in a Dsh-dependent manner. We also demonstrate that SDC4 can be modified by monoubiquitin moieties that are linked to lysine residues located in SDC4 cytoplasmic domain. Importantly, Wnt5a and Dsh regulate ubiquitination of SDC4. Therefore, this is the first example of a cell surface protein ubiquitinated and degraded in a Wnt/Dsh-dependent manner. These results underscore a novel role for non-canonical Wnt signaling in regulating the stability and turnover of FA receptors.

EXPERIMENTAL PROCEDURES
Cell Surface Biotinylation Assay-Transfected HEK293T cells were treated with pure recombinant Wnt5a (R&D Systems) for different times and then incubated with a solution of 0.5 mg/ml EZ-Link TM NHS-biotin. The biotin reaction was quenched for 10 min with 50 mM NH 4 Cl followed by two PBS washes. Cells were lysed in buffer A (50 mM Hepes, pH 7.4, 150 mM NaCl, 1 mM EGTA, 2 mM MgCl 2 , 10% glycerol, 1% Triton X-100) in the presence of protease inhibitors. Homogenates were bound and precipitated with neutroavidin-agarose beads, and proteins were eluted and analyzed by ␣-FLAG Western blotting.
Pulse-Chase Experiments-HEK293T cells were transfected with FLAG-xSDC4, and 24 h later they were incubated in serum-free medium for 30 min. [ 35 S]Met and cycloheximide were added for 1 h, and then cells were incubated in the presence or absence of pure recombinant Wnt5a (100 ng/ml). Cells were lysed in buffer B (10 mM Tris, pH 7.5, 150 mM NaCl, 0.5% Triton X-100), and homogenates were immunoprecipitated with ␣-FLAG antibody and analyzed by fluorography. SDC4 half-life was estimated by fitting an exponential trendline to the experimental values obtained.
Ubiquitination Assays-HEK293T cells were transfected and after 24 h were lysed in buffer B and subjected to two different protocols. For double-immunoprecipitation (see Fig. 5a, lane 4), cell homogenates were processed for ␣-FLAG immunoprecipitation, and after three washes with buffer B the Sepharose beads containing the precipitated immunocomplexes were boiled in 1% SDS for 5 min, diluted with TNTE (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 0.1% Triton X-100), subjected to a second ␣-FLAG immunoprecipitation, and immunocomplexes eluted as below. For single-immunoprecipitation (see Fig. 5a, lanes 1-3, b-d, and supplemental Fig. S3, a and b), cell extracts were immunoprecipitated with ␣-FLAG antibody and the immunocomplexes eluted in SDS-PAGE loading buffer. Immunoprecipitated proteins were analyzed by Western blotting using an ␣-HA-HRP-conjugated antibody.
Plasmid Construction-To generate FLAG-xSDC4⌬PBM or FLAG-xSDC4-mutK, site-directed mutagenesis was performed by using the QuikChange Site-Directed Mutagenesis kit (Stratagene, La Jolla, CA), with FLAG-xSDC4 as a template DNA (16). All constructs were confirmed by DNA sequencing.
Cell Culture-Protocols previously described were used for culturing and transfecting HEK293T (16). CHX (350 M) was added for 4 h prior to cell lysis. siDvl2 (sc-35230, Santa Cruz Biotechnology) was used for Fig. 3e.
Immunoprecipitation-For co-immunoprecipitation experiments HEK293T cells were grown to 80% confluence in 6-well plates and were transiently transfected with 1 g of FLAG-xSDC4, FLAG-xSDC4⌬C, or FLAG-xSDC4⌬PBM vectors using the Lipofectamine 2000 reagent (Invitrogen). After transfection (24 h), the cells were lysed with 250 l of immunoprecipitation buffer (150 mM NaCl, 20 mM Tris, pH 7.5, 1.5 mM CaCl 2 , 1.5 mM MgCl 2 , and 0.5% Triton X-100) containing protease inhibitors. Homogenates were precleared with 30 l of protein A-Sepharose beads (Santa Cruz Biotechnology) at 4°C for 1 h with rocking. FLAG-xSDC4 proteins were immunoprecipitated with ␣-FLAG mouse antibody (Stratagene) and incubated overnight at 4°C with rocking followed by protein A-Sepharose bead pulldown at 4°C for 2 h. After three washes with the immunoprecipitation buffer, immune complexes were released in 65 l of SDS-PAGE loading buffer by boiling at 100°C for 5 min, separated by SDS-PAGE, transferred to PVDF membranes, and then subjected to immunoblot analysis, with ␣-FLAG mouse antibody (Stratagene) 1:2000 or ␣-myc (Santa Cruz Biotechnology) 1:500.
Immunofluorescence Microscopy-HEK293T cells were growth on fibronectin-coated coverslips (10 mg/ml) in DMEM containing 10% FBS in an atmosphere of 5% CO 2 at 37°C. Cells were incubated in the presence or absence of Wnt5a protein for 3 h and analyzed by immunofluorescence. Cells were washed twice with PBS and fixed with 3% paraformaldehyde for 10 min at room temperature followed by blocking with PBS containing 5% non-fat powdered milk for 1 h at room temperature. Mouse monoclonal ␣-SDC4 antibody (5G9, sc-12766; Santa Cruz Biotechnology) was diluted at 1:100 and incubated in a humid chamber at room temperature for 1 h in PBS-blotto. After three washes cells, were incubated with ␣-mouse conjugated to Alexa Fluor 488 (Molecular Probes) and incubated for 1 h at room temperature in a humid chamber. Nuclear counterstaining was performed using Hoechst 33258 (1 mg/ml) for 5 min at room temperature. Samples were washed, mounted, and analyzed, and representative cells were selected using a Nikon Diaphot inverted microscope equipped for epifluorescence.
Internalization Assay-HEK293T cells were cultured in 6well plates and transfected by the calcium phosphate method with FLAG-SDC4 or FLAG-SDC4mutK. At 24 h after transfection, cells were plated onto poly-D-lysine-coated glass coverslips. 24 h later, the cells were incubated with binding medium (DMEM, 20 mM Hepes, pH 7.5, 0.1% BSA p/v) for 30 min at 4°C, and further incubated with 500 ng/ml anti-FLAGM2 (Stratagene) at 4°C for 1 h. The cells were then washed three times with cold PBS and incubated with 250 ng/ml Wnt5a (R&D Systems) or DMEM without serum (control medium) at 37°C. Time 0 corresponds to cells that were always maintained at 4°C. After that, the cells were processed for immunofluorescense as depicted above. A blind observer did the counting of the cells.

Non-canonical Wnt Signaling Reduces Cell Surface Levels and Promotes Degradation of Syndecan4-
The effect of Wnt5a/11 in SDC4 cell surface levels was evaluated by different approaches. Xenopus embryos were microinjected with synthetic mRNA for EGFP-xSDC4 alone or together with mRNA for xWnt5a or xWnt11. Explants were isolated at blastula or gastrula stages (animal caps or DMZs containing cells undergoing CE) and visualized by confocal microscopy. Overexpression of Wnt5a or Wnt11 resulted in a strong reduction in the cell surface levels of SDC4 ( Fig. 1, a-f and supplemental Fig. S1, a-f). In addition, incubation with pure recombinant Wnt5a for 3 h resulted in a strong reduction in the cell surface levels of endogenous SDC4 in HEK293T cells (Fig. 1, g and h). Wnt5a was also able to reduce the levels of biotinylated cell surface FLAG-xSDC4 rapidly (Fig. 1i). Therefore, Wnt5a and Wnt11 reduce SDC4 cell surface levels in cultured cells and Xenopus gastrulae.
The experiments depicted above also suggested that Wnt5a/11 could have an effect on SDC4 total levels (Fig. 1i, compare SDC4 total protein in lanes 1 and 5). To study with more detail the effects of Wnt5a/11 on SDC4 steady-state levels the following experimental paradigm was set up and used along this work. FLAG-xSDC4 was overexpressed in HEK293T cells (by DNA transfection with FLAG-xSDC4 under the control of a CMV promoter) or in Xenopus explants (by RNA microinjection) followed by incubation with pure recombinant Wnt5a (100 -250 ng/ml) for different times (1-3 h), and total levels of FLAG-xSDC4 were evaluated by Western blotting. Importantly, after 3 h of incubation, the steady-state levels of FLAG-xSDC4 protein were down-regulated in cells and embryos (Fig. 2, a and b). This effect is even more remarkable considering that FLAG-xSDC4 was constantly being overproduced under these conditions. Sim- ilar results were obtained when Wnt11 was co-expressed with FLAG-xSDC4 (see supplemental Fig. S1, g and h) and also when Wnt5a was co-expressed with mouse SDC4 (see supplemental Fig. S1i).
Because SDC4 protein was produced from synthetic mRNA in microinjected embryos, the possibility that the effects of Wnt ligands occurred at the transcriptional level was ruled out. Due to the kinetics of Wnt5a effects and the experimental layout the most plausible explanation was that Wnt5a had a post-translational effect. To address this possibility, we tested the effect of Wnt5a in the presence of the protein synthesis inhibitor, CHX. Wnt5a was able to reduce the steady-state levels of SDC4 in the presence of CHX (Fig. 2c, compare lanes 1 and 2), indicating that Wnt5a regulates SDC4 at a post-translational level. Of  SEPTEMBER 17, 2010 • VOLUME 285 • NUMBER 38 note, the effect of Wnt5a was specific because the steady-state levels of other cell surface receptors were not affected (Fig. 2c,  lanes 3-8). Together, these experiments suggested that Wnt5a promotes degradation of SDC4 protein. Accordingly, a pulsechase experiment showed that in the absence of exogenous Wnt5a, the half-life of FLAG-xSDC4 was ϳ4.7 h, whereas it was reduced to 2.3 h when Wnt5a was added (Fig. 2, d and e). Altogether, these findings demonstrated that Wnt5a accelerates SDC4 degradation.

Wnt5a Induces SDC4 Ubiquitination and Degradation
To evaluate the consequences of reducing non-canonical Wnt signaling on SDC4 steady-state levels, we tested the effect of dnWnt11 or sFz7, two constructs that block this signaling pathway in HEK293T cells and dorsal mesodermal cells (19,22). In line with the results depicted above, overexpression of dnWnt11 or sFz7 in Xenopus DMZ explants or HEK293T cells increased the steady-state levels of FLAG-xSDC4 protein ( Fig.  2f and supplemental Fig. S1, j and k). These results indicate that SDC4 is actively degraded in cultured cells and dorsal mesodermal cells by endogenous levels of non-canonical Wnt signaling and blocking this activity is sufficient to counteract SDC4 degradation. In conclusion, increasing non-canonical Wnt signaling is sufficient to promote SDC4 degradation and reduce its steady-state levels.
To determine how Wnt5a induces SDC4 degradation we tested the effect of lysosome or proteasome inhibition on Wnt5a-induced degradation. We found that SDC4 degradation was insensitive to lysosome inhibition by methylamine and chloroquine ( Fig. 2g and supplemental Fig. S2a). However, SDC4 degradation was blocked after proteasome inhibition with lactacystin (Fig. 2h). The same concentration of methylamine was able to block TrkA degradation induced by NGF, corroborating the effectiveness of the drug in cell surface recep-tors degraded by the lysosome (supplemental Fig. S2b). In conclusion, Wnt5a induces degradation of SDC4 by a lysosomeindependent and proteasome-dependent pathway. A similar mechanism has been described previously for other cell surface proteins, including the opioid receptors, urea transporter, amiloride-sensitive epithelial Na ϩ channel, and the glutamate transporter (23)(24)(25)(26).
Role of Dishevelled in Syndecan4 Steady-state Levels-Because Dsh interacts with SDC4 (16) and binds to components of the endocytic pathway (27,28), we decided to evaluate the role of Dsh in regulating SDC4 steady-state levels. In gain-of-function experiments we found that Dsh reduces SDC4 steady-state levels (Fig. 3a). More importantly, suboptimal amounts of Dsh and Wnt5a that have almost no effect when used alone (Fig. 3b, compare lane 2 with lanes 3-5) have a strong cooperative effect in reducing SDC4 steady-state levels (Fig. 3b, lanes 6 and 7). Conversely, reduction of Dsh levels/activity by morpholino and siRNA and overexpression of Dsh dominant-negative constructs (xDsh-DEP ϩ , Xdd1), which specifically block non-canonical Wnt signaling, results in increased SDC4 steady-state levels (Fig. 3, c-e, and supplemental Fig. S2, c and d). The effect of Dsh knockdown was rescued by microinjection of myc-xDsh mRNA supporting the specificity of this results (Fig. 3c, lanes  4 -6). These experiments demonstrated that Dsh is sufficient and necessary to regulate SDC4 levels in cells and embryos.
Then we studied the relevance of the interaction between Dsh and SDC4 for SDC4 stability. For this, mutants of SDC4 cytoplasmic domain were used. Deletion of the last three amino acids of SDC4 cytoplasmic domain, corresponding to a putative PDZ binding motif (PBM, FLAG-xSDC4⌬PBM, Fig. 4a) diminished co-immunoprecipitation of Dsh (Fig. 4b, compare lanes 2 and 4) without decreasing SDC4 cell surface localization (Fig.  4c). This deletion construct provided an excellent tool to correlate the disruption of Dsh-SDC4 interaction in SDC4 steadystate levels and Wnt5a-induced degradation. Overexpression of equal amounts of the RNA or DNA for xSDC4⌬PBM resulted in the production of significantly higher steady-state levels of the SDC4 mutant proteins compared with wild type (Figs. 4d, compare lanes 2-4 with 5-7 and supplemental Fig. S2e). This suggests that the interaction with Dsh makes SDC4 more suitable for degradation. To investigate this directly, we evaluated the effect of Wnt5a on SDC4⌬PBM. Importantly, FLAG-xSDC4⌬PBM is completely resistant to the presence of Wnt5a (Fig. 4e, compare lanes 3 and 4 with 6 and 7). Furthermore, SDC4⌬PBM steady-state levels are insensitive to overexpression of xDsh-DEP ϩ (Fig. 4f). These results demonstrated that the interaction with Dsh is required for Wnt5ainduced degradation of SDC4.

Syndecan4 Is Ubiquitinated in a Wnt5a-and Dsh-dependent Manner-Ubiquitin (Ub) can covalently modify proteins as a monomer (monoubiquitination) or by forming Ub chains (polyubiquitination).
Monoubiquitination has been involved in a variety of cellular processes, including internalization and endocytosis of cell surface receptors, whereas polyubiquitination regulates protein degradation (29). The four conserved lysines in SDC4 cytoplasmic domain (Fig. 4a) were candidate amino acids for ubiquitination. Importantly, these lysine residues are located in the variable region of SDC4 cytoplasmic domain, which is unique for this proteoglycan among the syndecan family and is conserved along evolution (30).
TotestSDC4ubiquitination,FLAG-xSDC4 and HA-tagged ubiquitin (Ub-HA) were overexpressed in HEK293T cells and embryos. Cell homogenates were analyzed through ␣-FLAG immunoprecipitation followed by ␣-HA Western blotting. Using this assay, we have found that SDC4 was covalently modified by Ub in HEK293T cells and Xenopus embryos (Fig. 5a, lane 2, and supplemental Fig. S3a). The possibility that a ubiquitinated SDC4-interacting protein co-precipitates with SDC4 was excluded because no ubiquitination was detected when the lysine residues of SDC4 cytoplasmic domain were mutated to alanines (SDC4-mutK, see Fig. 4a  and 5a, lane 3). Moreover, the same bands were obtained when the experiments were performed under stringent denaturing conditions (Fig. 5a, lane 4).
Ub shifted SDC4 from an apparent molecular mass of 46 kDa (corresponding approximately to a SDC4 homodimer) to bands running with an apparent molecular mass between 55 and 70 kDa (see arrows in Fig. 5a), suggesting that SDC4 could receive approximately one to three Ub moieties (Ub molecular mass is ϳ8.5 kDa). This shift could result from the addition of multiple monoubiquitins moieties or a short Ub chain. To address this issue directly, we made use of a Ub mutant that only supports monoubiquitination because their seven lysines residues that allow chain formation have been mutated to arginine (UbK0-HA) (31). Importantly, the same pattern of SDC4 ubiquitination was detected when Ub-HA and UbK0-HA were used (Fig. 5b, compare lanes 2 and 4), indicating that SDC4 is monoubiquitinated. . Interaction between Syndecan4 and Dsh regulates Syndecan4 steady-state levels. a, sequences of SDC4 cytoplasmic domain and mutants generated are shown. b, Western blot analysis of co-immunoprecipitation assays of HEK293T cells co-transfected with the indicated constructs is shown. c, HEK293T cells overexpressing SDC4 constructs were incubated with membrane-impermeable biotin, and the levels of SDC4 in the cell surface were determined by neutroavidin pulldown followed by Western blot analysis. d, Western blot of animal caps (A.caps) from embryos microinjected with decreasing amounts of the indicated mRNAs is shown. The same amounts of FLAG-xSDC4 and FLAG-xSDC4⌬PBM mRNAs were microinjected. e, Wnt5a has no effect on the steady-state levels of FLAG-xSDC4⌬PBM. Animal caps overexpressing FLAG-xSDC4 or FLAG-xSDC4⌬PBM were incubated with Wnt5a for 3 h, and SDC4 levels were evaluated by Western blotting. To obtain equal amounts of SDC4 proteins, 60 and 10 pg of synthetic mRNAs for FLAG-xSDC4 and FLAG-xSDC4⌬PBM were microinjected, respectively. f, xDsh-DEP ϩ has no effect on the steady-state levels of FLAG-xSDC4⌬PBM. DMZ from embryos microinjected with the indicated mRNAs were isolated and analyzed by Western blotting.
Considering that non-canonical Wnt signaling promotes degradation of SDC4 and that SDC4 is ubiquitinated, we hypothesized that this signaling pathway could regulate SDC4 ubiquitination. To prove this, we tested the effect of modifying the levels of non-canonical Wnt signaling on SDC4 ubiquitination. To avoid the effect of Wnt5a in SDC4 steadystate levels, we performed experiments in the presence of the proteasome inhibitor lactacystin, which blocks Wnt5a-induced degradation of SDC4 (Fig. 2h). Addition of Wnt5a for 3 h resulted in increased levels of ubiquitinated SDC4 (Fig.  5c). Conversely, overexpression of xDsh-DEP ϩ resulted in a reduction in the levels of SDC4 ubiquitination in HEK293T cells and Xenopus gastrula (Fig. 5d and supplemental Fig. S3b).
To determine the relevance of monoubiquitination in SDC4 homeostasis we tested its effect at two steps: internalization and steady-state levels. To evaluate the role of ubiquitination on SDC4 internalization we measured the effect of Wnt5a in the uptake of anti-FLAG antibody on cells overexpressing FLAG-xSDC4 and FLAG-xSDC4mutk. We found that Wnt5a was able to induce internalization of the ubiquitination mutant as efficiently as the wild type (Fig. 5, e-i). To determine the effect of ubiquitination in the protein levels of SDC4 we microinjected equal amounts of synthetic mRNAs for FLAG-EGFP-xSDC4 or FLAG-EGFP-xSDC4-mutK into Xenopus embryos, and isolated explants were visualized by confocal microscopy. Higher levels of SDC4-mutK compared with wild type SDC4 were detected by immunofluorescense (Fig. 5, j-o), indicating that ubiquitination regulates SDC4 steady-state levels in Xenopus embryos. Intriguingly, we detected very low amounts of FLAG-xSDC4mutk by Western blot analysis of Xenopus explants and HEK293T cells (Fig. 5p, compare lanes 1  and 4 and supplemental Fig. S3c). As we discuss below, we propose that this paradox could be explained if the absence of ubiquitin modification in the intracellular domain favors the SDC4 tendency to form oligomers that are unable to enter SDS-PAGE gels. In these conditions Wnt5a and Wnt11 reduced the steadystate levels of wild type FLAG-xSDC4 but not of the ubiquitination mutant (Fig. 5p, compare lanes 1-3 with 4 and 5 and supplemental Fig. S3d). We conclude from these results that ubiquitination regulates Wnt-induced degradation of SDC4 but not its internalization. A similar role for monoubiquitination has been described for other cell surface receptors (32,33) although it is still controversial for the EGF receptor (34,35).

DISCUSSION
This work introduces non-canonical Wnt signaling as a new extracellular cue involved in regulation of SDC4 internalization and stability. We have demonstrated that non-canonical Wnt signaling reduces cell surface levels and promotes degradation and monoubiquitination of SDC4 in Xenopus gastrulae and cultured cells. Dsh plays an important role on regulating SDC4 steady-state levels. Increasing Dsh levels enhanced Wnt5ainduced degradation of SDC4, and blocking Dsh activity increased SDC4 steady-state levels and diminished its ubiquitination. Mutation of SDC4 cytoplasmic lysine residues abolished ubiquitination and resulted in up-regulation of SDC4 cell surface levels in Xenopus gastrulae, indicating that ubiquitination regulates SDC4 in vivo. This represents the first example of a cell surface protein that is ubiquitinated and degraded in a Wnt5a/11-dependent manner.
The key role of Dsh in ubiquitination and degradation of SDC4 is particularly relevant because Dsh also binds to important components of the endocytic and degradative cellular machinery. Dsh interacts with the clathrin AP-2 adaptor in part via its DEP domain and promotes endocytosis and degradation of Fz4 (27). Moreover, Wnt5a induces the formation of a complex between Dsh and the E3 ubiquitin ligases Smurf1 and Smurf2, promoting ubiquitination and degradation of Prickle, a cytoplasmic component of the non-canonical Wnt signaling pathway (36). We suggest that Dsh could bridge the binding of Smurf1 or 2 and the AP2 adaptor to SDC4 and regulate its ubiquitination, endocytosis, and degradation. Furthermore, SDC4 and Smurfs regulate CE in Xenopus and mouse embryos, respectively (16,36), giving further support to the possibility that they function in a common pathway.
Mono-and multiubiquitination of cell surface receptors induce its internalization and targeting to the late endosome (multiple vesicle bodies) (37,38). Based on this and our finding that SDC4 is mainly mono-and multiubiquitinated, it is plausible to propose that Wnt/Dsh can promote SDC4 degradation by favoring sorting to the endosomal degradation route. It is important to note that similar results have been described for cell surface proteins such as the opioid receptors, urea transporter, amiloride-sensitive epithelial Na ϩ channel, and the glutamate transporter (23)(24)(25)(26). Interestingly, the intracellular protein syntenin that also binds to the conserved PBM of Syndecans promotes recycling from endosomal compartments to the plasma membrane (39). We have found that Dsh promotes SDC4 degradation through binding to its PBM, suggesting a model whereby Syntenin and Dsh binding to the C-termi-FIGURE 5. Syndecan4 is ubiquitinated in a Wnt5a-and Dsh-dependent manner. a, ubiquitination of SDC4. FLAG-xSDC4 or FLAG-xSDC4mutK was cotransfected with Ub-HA in HEK293T cells, and homogenates were immunoprecipitated with ␣-FLAG antibodies and eluted with loading buffer followed by Western blot analysis (lanes 1-3). To discard co-immunoprecipitation of other ubiquitinated proteins, beads after the first ␣-FLAG pulldown were eluted with 1% SDS at 95°C for 5 min. Afterward, this eluate was processed for a second round of ␣-FLAG immunoprecipitation (lane 4). Arrows point to ubiquitinated SDC4. b, SDC4 is monoubiquitinated. FLAG-xSDC4 was co-transfected with wild type Ub (Ub-HA) or a mutant Ub that is unable to form Ub chains (UbK0-HA) and immunoprecipitated with ␣-FLAG antibodies. c and d, non-canonical Wnt signaling regulates SDC4 ubiquitination. The effect of Wnt5a addition (3 h; c) and xDsh-DEP ϩ overexpression in SDC4 ubiquitination (d) was analyzed in HEK293T cells. For the experiment with Wnt5a, cells were incubated with lactacystin 6 h prior to homogenization. e-h, Wnt5a-induced SDC4 internalization is not affected in lysine mutant. HEK293T cells expressing xSDC4FLAG or FLAG-xSDC4mutk were treated with Wnt5a for 30 min, and the levels of cell membrane and intracellular SDC4 were determined by anti-FLAG uptake as described under "Experimental Procedures." i, quantification of the results depicted in e-h. j-p, characterization of xSDC4-mutK is shown. Embryos were microinjected with equal amounts of the indicated mRNAs. j-o, animal caps were isolated at stage 9 and fixed for confocal microscopy analysis. Scale bar, 50 m. p, DMZs were isolated at stage 10ϩ and analyzed by Western blotting. nal domain of SDC4 could have different effects in its sorting within the endosomal compartments.
Of note, ubiquitination of SDC4 occurs in the variable region of its cytoplasmic domain containing the highly conserved motif KKPIYKK. This peptide binds the membrane lipid phosphatidylinositol 4,5-bisphosphate, allowing clustering of SDC4 followed by activation of PKC␣ (30,40). SDC4 clustering results in formation of oligomers that are resistant to strong denaturing conditions (30,40). Cell surface receptor clustering, such as when SDC4 is incorporated into FA, is a common mechanism involved in signaling. The addition of Ub moieties to this motif could be a mechanism to disrupt SDC4 clustering with the consequent signaling termination and degradation of this cell surface proteoglycan. In such a case, ubiquitination-deficient mutants should have a strong tendency for oligomerization. Formation of SDC4mutk oligomers that are resistant to reducing conditions used in SDS-PAGE analysis could explain the paradoxical observation that higher levels of SDC4mutk than SDC4 were visualized by immunofluorescense (Fig. 5, j-o) but the opposite was detected by Western blot analysis (Fig. 5p).
Our findings raised the possibility that Wnt5a-induced ubiquitination could be a general mechanism to modulate the stability and turnover of cell adhesion complexes. Non-canonical Wnt signaling regulates turnover, ubiquitination, and proteasome-dependent degradation of Paxillin (10,41) and Smurf1 localizes to FA, promoting ubiquitination of the intracellular FA component Talin (42). Considering that integrin-regulated endocytosis is crucial for FA disassembly and cell migration (9,43) and that together with SDC4, they are the main FA receptors, it is plausible that non-canonical Wnt signaling could also regulate integrin turnover.
Furthermore, our findings are relevant for the current understanding of the molecular and cellular mechanism underneath the metastasic behavior of tumorigenic cells. It has been reported that abnormally high levels of Wnt5a correlate with malignancy (10,44), and disruption of integrin recycling also correlates with metastasis (9,45). The finding that Wnt5a regulates ubiquitination and degradation of the FA receptor SDC4 suggests a possible mechanism to explain the increased invasiveness of tumorigenic cells when Wnt5a activity is augmented.