Summary

Wnts are glycan- and lipid-modified morphogens that are important for cellular responses, but how Wnts are secreted in polarized epithelial cells remains unclear. Although Wntless (Wls) has been shown to interact with Wnts and support their secretion, the role of Wls in the sorting of Wnts to the final destination in polarized epithelial cells have not been clarified. Glycosylation was shown to be important for the sorting of some transmembrane and secreted proteins, but glycan profiles and their roles in the polarized secretion of Wnts has not yet been demonstrated. Here we show the apical and basolateral secretion of Wnts is regulated by different mechanisms. Wnt11 and Wnt3a were secreted apically and basolaterally, respectively, in polarized epithelial cells. Wls was localized to the basolateral membrane. Mass-spectrometric analyses revealed that Wnt11 is modified with complex/hybrid(Asn40)-, high-mannose(Asn90)- and high-mannose/hybrid(Asn300)-type glycans and that Wnt3a is modified with two high-mannose-type glycans (Asn87 and Asn298). Glycosylation processing at Asn40 and galectin-3 were required for the apical secretion of Wnt11, whereas clathrin and adaptor protein-1 were required for the basolateral secretion of Wnt3a. By the fusion of the Asn40 glycosylation site of Wnt11, Wnt3a was secreted apically. The recycling of Wls by AP-2 was necessary for the basolateral secretion of Wnt3a but not for the apical secretion of Wnt11. These results suggest that Wls has different roles in the polarized secretion of Wnt11 and Wnt3a and that glycosylation processing of Wnts decides their secretory routes.

Introduction

Wnt proteins are the large family of secreted molecules that are important in developmental processes, and defective Wnt signaling in postnatal life causes disease in humans (Kikuchi et al., 2011; Logan and Nusse, 2004; Polakis, 2007). Great advances have been made in understanding the mechanisms by which Wnt activates the signaling pathways and regulates cellular functions. In contrast, how Wnt is expressed and secreted remains unclear. Recently, Evenness interrupted (Evi) and Opossum (Opm) have been shown to be required for the export of Wingless (Wg) from the endoplasmic reticulum (ER) and the Golgi apparatus, respectively, in Drosophila (Bänziger et al., 2006; Buechling et al., 2011; Port and Basler, 2010; Port et al., 2011). These molecules have been conserved evolutionarily, as TMED5 and Wntless (Wls), respectively, in mammals. Wls interacts with Wnts and supports their transport from the trans Golgi network (TGN) to the cell surface membrane in mammalian cells (Port and Basler, 2010). However, because these mammalian studies were performed using non-polarized cells such as COS-7, HEK293T and HEK293 (Bänziger et al., 2006; Coombs et al., 2010; Jin et al., 2010), the roles of Wls in apical and basolateral secretion of Wnts in polarized epithelial cells are not well understood.

The sorting process into the apical and basal regions occurs at the level of the TGN by incorporation of apical and basolateral proteins into distinct vesicles (Cao et al., 2012; Rodriguez-Boulan et al., 2005). Several sorting signals have been identified in membrane proteins. For instance, basolateral signals, such as YxxΦ (Φ representing any amino acid with hydrophobic side chains) and LL motifs, are present in the cytoplasmic region of basolateral plasma membrane proteins. Clathrin and clathrin-associated heterotetrameric adaptor protein complexes (APs), including AP-1 and AP-4, interact with basolateral signals of membrane proteins and direct proteins to different membrane domains (Boehm and Bonifacino, 2001; Deborde et al., 2008; Fölsch, 2005; Gravotta et al., 2012; Gravotta et al., 2007; Simmen et al., 2002). Apical signals, such as glycosylphosphatidylinositol anchors, N-glycans and O-glycans, have been reported in apical membrane proteins (Huet et al., 2003; Vagin et al., 2009). Galectin-3 and galactic-4 that recognize β-galactoside structures (Boscher et al., 2011; Hughes, 1999) have been implicated in the sorting of glycosylated apical membrane proteins in epithelial cells (Delacour et al., 2006; Delacour et al., 2007; Delacour et al., 2005; Stechly et al., 2009). In contrast to membrane proteins, sorting signals for secreted proteins remain unclear. N-glycans and O-glycans in some secreted proteins were suggested to be apical signals (Huet et al., 2003).

Wnt3a and Wnt 11 are representative ligands that activate the β-catenin-dependent (canonical) and -independent (non-canonical) pathways (Heisenberg et al., 2000; Kikuchi et al., 2011; MacDonald et al., 2009; Veeman et al., 2003). Immunohistochemical and in situ hybridization analyses revealed that Wnt3a and Wnt11 are present in epithelial cells (Dean et al., 2005; Hayashi et al., 2011; Lickert et al., 2001; Uysal-Onganer and Kypta, 2012; Vainio and Lin, 2002). However, it is not clear how Wnts are secreted to the apical or basolateral region in polarized epithelial cells. Here we demonstrate the glycan profiles of Wnt11 and Wnt3a by mass spectrometry and show that Wnt11 and Wnt3a are secreted apically and basolaterally, respectively, by different mechanisms.

Results

Polarized secretion of Wnt11 and Wnt3a

To examine whether Wnt11 and Wnt3a are secreted into either the apical or basolateral fraction in polarized epithelial cells, Madin–Darby canine kidney (MDCK) epithelial cells expressing Wnt11 or Wnt3a were cultured as a monolayer on a filter support (Gottardi et al., 1995) (Fig. 1A). Secreted soluble Wnts were precipitated using Blue Sepharose from culture media in the upper and lower chambers, and cell surface membrane-associated Wnts on the apical and basolateral membranes were biotinylated and precipitated using NeutrAvidin agarose (Fig. 1B). The majority of the Wnt11 and Wnt3a was detected in the apical and basolateral fractions, respectively (Fig. 1B), suggesting polarized secretion of Wnt11 and Wnt3a in MDCK cells. Wnt3a was also observed in the apical fraction, but very little Wnt11 was detected in the basolateral fraction, if any (Fig. 1B).

Fig. 1.

Polarized secretion of Wnt11 and Wnt3a. (A) MDCK cells expressing Wnt11 or Wnt3a seeded on a Transwell polycarbonate filter were fixed, and then the cells were stained using anti-ezrin (green) and anti-β-catenin (red) antibodies. Ezrin and β-catenin were used as apical and basolateral markers, respectively. Scale bar: 5 µm. (B) MDCK cells expressing Wnt11 or Wnt3a were subjected to the apical-basolateral sorting assay. Soluble Wnts were detected in the precipitates from the apical media using Blue Sepharose, and membrane-associated Wnts were observed in the precipitates from the basolateral membranes using NeutrAvidin agarose. The signals of Wnts were quantified using NIH ImageJ and the results were expressed as the ratio of Wnts in the apical (Ap) and basolateral (Bl) fractions. The sum of Wnts secreted from MDCK cells into the apical and basolateral fractions were set to 100%. Podocalyxin and transferrin receptor were used as apical and basolateral membrane protein markers, respectively. B, precipitation using Blue Sepharose; N, precipitation using NeutrAvidin agarose. (C) MDCK cells expressing Wnt11 or Wnt3a were cultured in 3D Matrigel, and the cysts with or without permeabilization treatment were stained with the indicated antibodies or observed by relief-contrast microscopy. Scale bars: 5 µm. (D) Paraffin-embedded tissue sections of mouse kidney at embryonic day (E) 13 were immunohistochemically stained with (+) or without (−) an anti-Wnt11 antibody. The boxed region is shown at a higher magnification in the image on the right. Arrowheads indicate expression of Wnt11 in ureteric bud epithelium. Scale bars: 50 µm. (E) Immunofluorescence staining of a section of mouse kidney at E13 with anti-Wnt11 (green) and anti-ezrin (red) antibodies. Scale bar: 10 µm. (F) After MDCK cells expressing Wnt11 or Wnt3a with or without resistant Wls–MycHis were transfected with control or Wls siRNA, the cells were subjected to the apical-basolateral sorting assay. The signal of Wnt11 precipitated using Blue Sepharose in the control apical fraction or the signal of Wnt3a precipitated using NeutrAvidin agarose in the control basolateral fraction was set to 100%. Heat shock protein 90 (HSP90) was used as a loading control. (G) MDCK cells expressing Wls–MycHis were subjected to the apical-basolateral sorting assay. The sum of Wls on the apical and basolateral membranes was set to 100%. The results shown are means ± s.e.m. from four independent experiments.

Fig. 1.

Polarized secretion of Wnt11 and Wnt3a. (A) MDCK cells expressing Wnt11 or Wnt3a seeded on a Transwell polycarbonate filter were fixed, and then the cells were stained using anti-ezrin (green) and anti-β-catenin (red) antibodies. Ezrin and β-catenin were used as apical and basolateral markers, respectively. Scale bar: 5 µm. (B) MDCK cells expressing Wnt11 or Wnt3a were subjected to the apical-basolateral sorting assay. Soluble Wnts were detected in the precipitates from the apical media using Blue Sepharose, and membrane-associated Wnts were observed in the precipitates from the basolateral membranes using NeutrAvidin agarose. The signals of Wnts were quantified using NIH ImageJ and the results were expressed as the ratio of Wnts in the apical (Ap) and basolateral (Bl) fractions. The sum of Wnts secreted from MDCK cells into the apical and basolateral fractions were set to 100%. Podocalyxin and transferrin receptor were used as apical and basolateral membrane protein markers, respectively. B, precipitation using Blue Sepharose; N, precipitation using NeutrAvidin agarose. (C) MDCK cells expressing Wnt11 or Wnt3a were cultured in 3D Matrigel, and the cysts with or without permeabilization treatment were stained with the indicated antibodies or observed by relief-contrast microscopy. Scale bars: 5 µm. (D) Paraffin-embedded tissue sections of mouse kidney at embryonic day (E) 13 were immunohistochemically stained with (+) or without (−) an anti-Wnt11 antibody. The boxed region is shown at a higher magnification in the image on the right. Arrowheads indicate expression of Wnt11 in ureteric bud epithelium. Scale bars: 50 µm. (E) Immunofluorescence staining of a section of mouse kidney at E13 with anti-Wnt11 (green) and anti-ezrin (red) antibodies. Scale bar: 10 µm. (F) After MDCK cells expressing Wnt11 or Wnt3a with or without resistant Wls–MycHis were transfected with control or Wls siRNA, the cells were subjected to the apical-basolateral sorting assay. The signal of Wnt11 precipitated using Blue Sepharose in the control apical fraction or the signal of Wnt3a precipitated using NeutrAvidin agarose in the control basolateral fraction was set to 100%. Heat shock protein 90 (HSP90) was used as a loading control. (G) MDCK cells expressing Wls–MycHis were subjected to the apical-basolateral sorting assay. The sum of Wls on the apical and basolateral membranes was set to 100%. The results shown are means ± s.e.m. from four independent experiments.

MDCK cells formed a cyst in three-dimensional (3D) culture using Matrigel (O'Brien et al., 2002). When the cells were permeabilized to be stained, Wnt11 and Wnt3a were detected in the immediate vicinity of apical and basal membranes, respectively (Fig. 1C). Without permeabilization, Wnt3a, but not Wnt11, was detected on the basal membranes (Fig. 1C). Endogenous Wnt11 tended to be located in the apical region of ureteric bud epithelium of mouse embryonic kidney (Fig. 1D), which is similar to the localization of ezrin that is an apical marker (Fig. 1E). These results support the polarized secretion of Wnt11 and Wnt3a.

Wls interacts with Wnts and is essential for their secretion in non-polarized cells (Bänziger et al., 2006; Port and Basler, 2010). Knockdown of Wls almost completely inhibited the secretion of both Wnt11 apically and Wnt3a basolaterally in polarized MDCK cells under conditions in which the knockdown did not affect the polarized sorting of podocalyxin and transferrin receptor (TfR; Fig. 1F; supplementary material Fig. S1A). Expression of human Wls in Wls-depleted MDCK cells rescued the polarized secretion of Wnt11 and Wnt3a (Fig. 1F; supplementary material Fig. S1A). Interestingly, Wls was observed in the basolateral membrane mainly in polarized MDCK cells (Fig. 1G). These results suggest that although Wls is essential for the secretion of Wnt11 and Wnt3a, Wls is trafficked with Wnt3a to the basolateral membranes but not associated with Wnt11 after their exit from the TGN. Therefore, Wls may have distinct roles in the polarized secretion of Wnt11 and Wnt3a.

Glycosylation and lipidation of Wnt11 and Wnt3a

N-Glycosylation was reported to be involved in the polarized sorting of some transmembrane and secreted proteins (Huet et al., 2003; Vagin et al., 2009). Wnt11 and Wnt3a have been suggested to be glycosylated and there are five and two potential consensus sequences (Asn-X-Ser/Thr), respectively, for Asn-linked glycosylation. To analyze the glycan profiles at the possible sites (N40, N90, N160, N300 and N304 for Wnt11, and N87 and N298 for Wnt3a), Wnt11 and Wnt3a were purified from conditioned medium (CM) of mouse fibroblasts (L cells) stably expressing Wnt11 (supplementary material Fig. S2A–C; Table S1) or Wnt3a (Kishida et al., 2004; Komekado et al., 2007). Purified Wnts were digested with trypsin, and the resulting tryptic peptides were subjected to nano-flow reversed-phase liquid chromatography (LC) followed by matrix-assisted laser desorption/ionization (MALDI) mass spectrometry (MS)/MS. A number of tryptic peptides, which covered about 82.1% and 76.9% of the entire amino acid sequences of Wnt11 and Wnt3a, respectively, were recovered by this procedure (supplementary material Tables S2, S3). Based on a series of mass values with the intervals of 162 and 203, which represent the residual masses of hexose (Hex) and N-acetyl hexosamine (HexNAc), respectively, N40, N90 and N300 of Wnt11 and N298 of Wnt3a turned out to be heterogeneously glycosylated, whereas N160, N304, and a part of N300 of Wnt11 were not glycosylated. The peptide containing N87 of Wnt3a was not observed in these measurements but was observed using nano-flow LC/electrospray ionization (ESI)-MS/MS.

To further clarify the glycan profiles at each glycosylation site of Wnt11 and Wnt3a, the tryptic digests were subjected to nano-flow LC/ESI-MS/MS, because the ESI method allows for the estimation of relative abundances of the glycans of a glycopeptide based on their molecular masses (Satomi et al., 2004). The three glycopeptides with N40, N90 and N300 of Wnt11 and the two glycopeptides with N87 and N298 of Wnt3a gave a series of molecular ion peaks, whose intervals were consistent with Hex, HexNAc and sialic acid (Sia) for Wnt11 (Fig. 2A–C) and with Hex for Wnt3a (Fig. 2D,E). The glycan compositions estimated from the observed mass values indicated the glycans attached to N40, N90 and N300 of Wnt11 to be complex/hybrid, high-mannose and high-mannose/hybrid types, respectively, and those attached to N87 and N298 of Wnt3a to be both high-mannose types (supplementary material Tables S4, S5). N90 and N300 in Wnt11 and N87 and N298 in Wnt3a are highly conserved in 19 mouse Wnts, whereas N40 in Wnt11 is unique in that it is not conserved in other Wnts (supplementary material Fig. S3).

Fig. 2.

Determination of the glycan and lipid profiles of Wnt11 and Wnt3a. (A–C) Nano-flow LC/ESI mass spectra of Wnt11 glycopeptides eluted at 25.7 (A), 40.9 (B) and 24.9 minutes (C). The spectra were deconvoluted to singly charged ions. (D,E) Nano-flow LC/ESI mass spectra of Wnt3a glycopeptides eluted at 40.2 (D) and 29.3 minutes (E). The m/z values (MH+) and relative abundances of the observed ions from Wnt11 and Wnt3a glycopeptides are summarized in supplementary material Tables S4, S5, respectively. (F) MALDI mass spectrum of the peptide containing Ser215. (G) MALDI-MS/MS spectrum of the ion at m/z 1483.9 in F. The sequences from the N- and C-termini and the position of the modified residue were read out based on y″l and bm ions, respectively. The signals marked with asterisks were derived from the background at m/z 1483.9.

Fig. 2.

Determination of the glycan and lipid profiles of Wnt11 and Wnt3a. (A–C) Nano-flow LC/ESI mass spectra of Wnt11 glycopeptides eluted at 25.7 (A), 40.9 (B) and 24.9 minutes (C). The spectra were deconvoluted to singly charged ions. (D,E) Nano-flow LC/ESI mass spectra of Wnt3a glycopeptides eluted at 40.2 (D) and 29.3 minutes (E). The m/z values (MH+) and relative abundances of the observed ions from Wnt11 and Wnt3a glycopeptides are summarized in supplementary material Tables S4, S5, respectively. (F) MALDI mass spectrum of the peptide containing Ser215. (G) MALDI-MS/MS spectrum of the ion at m/z 1483.9 in F. The sequences from the N- and C-termini and the position of the modified residue were read out based on y″l and bm ions, respectively. The signals marked with asterisks were derived from the background at m/z 1483.9.

Consistent with the results of MS analyses, Wnt11 showed different migration patterns on the SDS-polyacrylamide gel by the treatment with Peptide-N-glycosidase F (PNGase F), which cleaves N-linked glycans, and endoglycosidase Hf (Endo Hf), which cleaves the high-mannose and hybrid types but not the complex type (supplementary material Fig. S4A). Wnt3a exhibited the same mobility shift when it was treated with PNGase F or Endo Hf (supplementary material Fig. S4A). When Wnt11 and Wnt3a secreted into the CM of MDCK cells were treated with PNGase F or Endo Hf, they showed the same mobility shift as Wnts from CM of L cells (supplementary material Fig. S4B), suggesting that Wnt11 and Wnt3a are glycosylated in a similar manner in L and MDCK cells.

Wnt3a is modified with palmitoleic acid at Ser209 (Takada et al., 2006). Purified Wnt11 was recovered in a detergent-enriched phase, indicating that Wnt11 is hydrophobic (supplementary material Fig. S2D). When the tryptic peptides of Wnt11 were subjected to nano-flow reversed-phase LC followed by MALDI-MS/MS, a lipid-modified peptide was observed among the identified peptides (supplementary material Table S2). The peptide containing S215, observed at m/z 1483.9 (Fig. 2F), was identified as the peptide with mono-unsaturated fatty acid (C16:1), palmitoleic acid, whose double bond was determined to be at the Δ9 position by observation of the air oxidized products. The site of lipidation was clearly assigned at S215 by MALDI-MS/MS (Fig. 2G), indicating that Wnt11 is modified with palmitoleic acid at Ser215.

Involvement of glycosylation and lipidation of Wnt11 and Wnt3a in the intracellular trafficking

To investigate the role of glycosylation of Wnt11 in the ER for secretion, glycosylation sites of Wnt11 were mutated to Gln (Q) (Fig. 3A). When Wnt11N40Q, Wnt11N90Q, Wnt11N300Q/N304Q (because N304 is glycosylated when N300 is mutated) or Wnt11N90Q/N300Q/N304Q was expressed in L cells, secretion of Wnt11N40Q or Wnt11N90Q/N300Q/N304Q was inhibited compared with wild-type (WT) Wnt11 (Fig. 3B). Secretion of Wnt11N90Q and Wnt11N300Q/N304Q was similar to that of WT Wnt11 (Fig. 3B). A mutation of either N87 or N298 in Wnt3a did not affect secretion, but double mutations blocked its secretion (Fig. 3C) (Komekado et al., 2007). When post-nuclear supernatant of L cells expressing WT Wnt11, WT Wnt3a, Wnt11N40Q or Wnt3aN87Q/N298Q was fractionated on a sucrose gradient, WT Wnt11 or WT Wnt3a was observed in all of the cytosol, Golgi and ER fractions (Fig. 3D). Most of the Wnt11N40Q, as well as Wnt3aN87Q/N298Q, was detected in the ER fraction (Fig. 3D). Therefore, the ER exit of Wnt11 requires the attachment of glycan to N40 in the ER, and either of two glycans in Wnt11 (N90 and N300) and Wnt3a (N87 and N298) is also necessary for the ER exit.

Fig. 3.

Involvement of glycosylation of Wnt11 and Wnt3a in the ER exit. (A) WT Wnt11 or the indicated Wnt11 mutants were expressed in HEK293T cells and the cell lysates were probed using an anti-Wnt11 antibody. The mobility of Wnt mutants on the SDS-PAGE was compared with that of WT Wnt11 from cells treated with tunicamycin. (B,C) CM, the ECM fraction, and lysates from L cells expressing the indicated Wnt11 or Wnt3a mutants were probed using an anti-Wnt11 (B) or anti-Wnt3a (C) antibodies (left panel). The protein levels of Wnts in CM and ECM reflect the amount of Wnts secreted from L cells. The signals of Wnt were quantified and the distribution of Wnt between cytosol, the ECM fraction and lysates was expressed as percentages (right panel). (D) Post-nuclear supernatant of L cells expressing the indicated Wnt mutants were fractionated using a linear sucrose density gradient. The distribution of Wnt11 or Wnt3a in each fraction was expressed as percentages (right panel). The results shown are means ± s.e.m. from four independent experiments. *P<0.01; **P<0.05. GSK-3β, adaptin-γ and calnexin were used as markers of the cytosol, Golgi and ER fractions, respectively.

Fig. 3.

Involvement of glycosylation of Wnt11 and Wnt3a in the ER exit. (A) WT Wnt11 or the indicated Wnt11 mutants were expressed in HEK293T cells and the cell lysates were probed using an anti-Wnt11 antibody. The mobility of Wnt mutants on the SDS-PAGE was compared with that of WT Wnt11 from cells treated with tunicamycin. (B,C) CM, the ECM fraction, and lysates from L cells expressing the indicated Wnt11 or Wnt3a mutants were probed using an anti-Wnt11 (B) or anti-Wnt3a (C) antibodies (left panel). The protein levels of Wnts in CM and ECM reflect the amount of Wnts secreted from L cells. The signals of Wnt were quantified and the distribution of Wnt between cytosol, the ECM fraction and lysates was expressed as percentages (right panel). (D) Post-nuclear supernatant of L cells expressing the indicated Wnt mutants were fractionated using a linear sucrose density gradient. The distribution of Wnt11 or Wnt3a in each fraction was expressed as percentages (right panel). The results shown are means ± s.e.m. from four independent experiments. *P<0.01; **P<0.05. GSK-3β, adaptin-γ and calnexin were used as markers of the cytosol, Golgi and ER fractions, respectively.

The modification with palmitoleic acid at S209 of Wnt3a is required for its secretion and interaction with Wls (Coombs et al., 2010). A mutation of S215 in Wnt11 suppressed its secretion in L cells (supplementary material Fig. S5A) and Wnt11S215A was accumulated in the Golgi and the ER fractions (Fig. 3D). WT Wnt11 but not Wnt11S215A formed a complex with Wls (supplementary material Fig. S5B). Therefore, palmitoleic acid in Wnt11 is necessary for its functional interaction with Wls.

Requirement of the complex/hybrid-type glycan at N40 for the apical secretion of Wnt11

To examine whether polarized secretion of Wnt11 and Wnt3a depends on the glycan profiles, cells were treated with kifunensine and deoxymannojirimycin (dMM), which inhibit ER α-mannosidase I and Golgi α-mannosidase I, respectively (Bischoff and Kornfeld, 1984; Elbein et al., 1990; Fuhrmann et al., 1984). These inhibitors prevent the synthesis of hybrid- and complex-type glycans and results in the accumulation of the high-mannose-type glycans. The treatment of MDCK cells expressing Wnt11 with kifunensine or dMM, resulted in the secreted Wnt11 becoming sensitive to Endo Hf, whereas these inhibitors did not affect the glycosylation profiles of Wnt3a (Fig. 4A). Treatment with kifunensine or dMM decreased the apical secretion of Wnt11, but the inhibitors did not affect the polarized secretion of Wnt3a (Fig. 4B).

Fig. 4.

Glycosylation processing at N40 is necessary for the apical secretion of Wnt11. (A) After MDCK cells expressing Wnt11 or Wnt3a were treated with kifunensine (kif) or deoxymannojirimycin (dMM), Wnts in the CM were precipitated using Blue Sepharose. The precipitates were incubated with Endo Hf or PNGase F and were probed using anti-Wnt11 or anti-Wnt3a antibodies. (B) After MDCK cells expressing Wnt11 or Wnt3a were treated with kif or dMM, the cells were subjected to the apical-basolateral sorting assay. The signal of Wnt11 precipitated using Blue Sepharose in the control apical fraction or the signal of Wnt3a precipitated using NeutrAvidin agarose in the control basolateral fraction was set to 100%. (C) After MDCK cells expressing Wnt11 or Wnt3a were transfected with Gnt-I siRNA, Wnts in CM were precipitated using Blue Sepharose. The precipitates were treated with Endo Hf or PNGase F. (D) After MDCK cells expressing Wnt11 or Wnt3a with or without resistant Gnt-I were transfected with control or Gnt-I siRNA, the cells were subjected to the apical-basolateral sorting assay. (E) After MDCK cells expressing Wnt11 were treated with kif, the cells were subjected to the post-Golgi sorting assay. The results shown are means ± s.e.m. from four independent experiments.

Fig. 4.

Glycosylation processing at N40 is necessary for the apical secretion of Wnt11. (A) After MDCK cells expressing Wnt11 or Wnt3a were treated with kifunensine (kif) or deoxymannojirimycin (dMM), Wnts in the CM were precipitated using Blue Sepharose. The precipitates were incubated with Endo Hf or PNGase F and were probed using anti-Wnt11 or anti-Wnt3a antibodies. (B) After MDCK cells expressing Wnt11 or Wnt3a were treated with kif or dMM, the cells were subjected to the apical-basolateral sorting assay. The signal of Wnt11 precipitated using Blue Sepharose in the control apical fraction or the signal of Wnt3a precipitated using NeutrAvidin agarose in the control basolateral fraction was set to 100%. (C) After MDCK cells expressing Wnt11 or Wnt3a were transfected with Gnt-I siRNA, Wnts in CM were precipitated using Blue Sepharose. The precipitates were treated with Endo Hf or PNGase F. (D) After MDCK cells expressing Wnt11 or Wnt3a with or without resistant Gnt-I were transfected with control or Gnt-I siRNA, the cells were subjected to the apical-basolateral sorting assay. (E) After MDCK cells expressing Wnt11 were treated with kif, the cells were subjected to the post-Golgi sorting assay. The results shown are means ± s.e.m. from four independent experiments.

Gnt-I, which is one of six N-acetylglucosamine transferases in vertebrates, initiates the formation of hybrid- and/or complex-type glycan (Rademacher et al., 1988). Knockdown of Gnt-I changed the glycosylation profiles of Wnt11 but not those of Wnt3a (Fig. 4C; supplementary material Fig. S1B). The apical secretion of Wnt11 but not the polarized secretion of Wnt3a was decreased in Gnt-I-depleted cells, and expression of mouse Gnt-I in Gnt-I-depleted cells restored the apical secretion of Wnt11 (Fig. 4D). To examine whether glycosylation processing at N40 is necessary for the release of Wnt11 from the TGN, MDCK cells were incubated at 20°C to block the secretion of proteins at the TGN. Wnt11 was secreted apically 1 or 2 hours after the cells were returned to the culture conditions at 37°C, and treatment of the cells with kifunensine inhibited the release of Wnt11 from the TGN (Fig. 4E).

To examine whether the addition of the glycosylation site of Wnt11 induces the apical secretion of Wnt3a, a chimeric protein, in which Wnt11[amino acids (aa)1–52] was fused to Wnt3a(aa50–352), was generated (Wnt11/3a; Fig. 5A). Wnt11/3a had glycan profiles similar to Wnt11 and was secreted apically like Wnt11 (Fig. 5B). Kifunensine, dMM and knockdown of Gnt-I suppressed the apical secretion of Wnt11/3a (Fig. 5C). In MDCK cysts, Wnt11/3a was observed in the vicinity of the apical membranes but not the basolateral membranes, as was Wnt11 (Fig. 5D). Taken together these results show that the complex/hybrid-type glycan at N40 of Wnt11 plays crucial roles in the apical secretion.

Fig. 5.

Galectin-3 is necessary for the apical secretion of Wnt11. (A) Glycan profiles of Wnt11, Wnt3a and Wnt11/3a are shown. (B) Wnt11/3a in CM was precipitated using Blue Sepharose, and the precipitates were incubated with Endo Hf or PNGase F and were probed using an anti-Wnt3a antibody. MDCK cells expressing Wnt11/3a were subjected to the apical-basolateral sorting assay. The signal of Wnt11/3a precipitated using Blue Sepharose in the control apical fraction was set to 100%. (C) After MDCK cells expressing Wnt11/3a were treated with kif or dMM, or transfected with Gnt-I siRNA, the cells were subjected to the apical-basolateral sorting assay. (D) MDCK cells expressing Wnt11/3a were cultured in 3D Matrigel, and the cysts with or without permeabilization treatment were stained with the indicated antibodies or observed by relief-contrast microscopy. Scale bar: 5 µm. (E,F) After MDCK cells expressing Wnt11, Wnt11/3a or Wnt3a with or without resistant galectin-3 were transfected with galectin-3 (Gal-3; E) or galectin-4 (Gal-4; F) siRNA, the cells were subjected to the apical-basolateral sorting assay. (G) Wnt11 (20 nM) was incubated with GST or GST–galectin-3 in the presence of indicated concentrations of galactose or glucose and the precipitates were probed using an anti-Wnt11 antibody. The signals of Wnt11 complexed with galectin-3 were quantified and expressed as arbitrary units. The results shown are means ± s.e.m. from four independent experiments.

Fig. 5.

Galectin-3 is necessary for the apical secretion of Wnt11. (A) Glycan profiles of Wnt11, Wnt3a and Wnt11/3a are shown. (B) Wnt11/3a in CM was precipitated using Blue Sepharose, and the precipitates were incubated with Endo Hf or PNGase F and were probed using an anti-Wnt3a antibody. MDCK cells expressing Wnt11/3a were subjected to the apical-basolateral sorting assay. The signal of Wnt11/3a precipitated using Blue Sepharose in the control apical fraction was set to 100%. (C) After MDCK cells expressing Wnt11/3a were treated with kif or dMM, or transfected with Gnt-I siRNA, the cells were subjected to the apical-basolateral sorting assay. (D) MDCK cells expressing Wnt11/3a were cultured in 3D Matrigel, and the cysts with or without permeabilization treatment were stained with the indicated antibodies or observed by relief-contrast microscopy. Scale bar: 5 µm. (E,F) After MDCK cells expressing Wnt11, Wnt11/3a or Wnt3a with or without resistant galectin-3 were transfected with galectin-3 (Gal-3; E) or galectin-4 (Gal-4; F) siRNA, the cells were subjected to the apical-basolateral sorting assay. (G) Wnt11 (20 nM) was incubated with GST or GST–galectin-3 in the presence of indicated concentrations of galactose or glucose and the precipitates were probed using an anti-Wnt11 antibody. The signals of Wnt11 complexed with galectin-3 were quantified and expressed as arbitrary units. The results shown are means ± s.e.m. from four independent experiments.

Galectin-3 and galectin-4, galactose-binding lectins, have been reported as potential apical sorting receptors (Delacour et al., 2006; Delacour et al., 2007; Delacour et al., 2005; Stechly et al., 2009). Knockdown of galectin-3 inhibited the apical secretion of Wnt11 and Wnt11/3a but did not affect the polarized secretion of Wnt3a (Fig. 5E; supplementary material Fig. S1C). Expression of human galectin-3 rescued the apical secretion of Wnt11 in galectin-3-depleted cells (Fig. 5E). However, knockdown of galectin-4 did not affect the polarized secretion of Wnt11 or Wnt3a (Fig. 5F; supplementary material Fig. S1C). Wnt11 bound to galectin-3 in vitro, and galactose, but not glucose, inhibited their binding (Fig. 5G). Therefore, the hybrid/complex-type glycan at N40 could play roles in the apical secretion of Wnt11, probably through the binding to galectin-3.

Requirement of clathrin, AP-1, AP-2 and Wls for the secretion of Wnt3a but not for the apical secretion of Wnt11

Clathrin and AP-1A, AP-1B and AP-4 sort cargo proteins to the basolateral membrane in epithelial cells (Boehm and Bonifacino, 2001; Deborde et al., 2008; Fölsch, 2005; Gravotta et al., 2012; Gravotta et al., 2007; Simmen et al., 2002). μ1A-HA and μ1B-HA, medium subunits of AP-1A and AP-1B, but not μ4-HA, a medium subunit of AP-4, formed a complex with Wls–MycHis and endogenous clathrin (supplementary material Fig. S6A). Wls–MycHis interacted with clathrin slightly and expression of μ1A-HA or μ1B-HA enhanced their complex formation (supplementary material Fig. S6A), suggesting a tertiary complex between Wls, AP-1 and clathrin. Consistent with a previous observation (Gravotta et al., 2012), depletion of μ1A and μ1B, decreased the adaptin-γ protein levels (supplementary material Fig. S1D,E). The polarized membrane localization of TfR and Wls was lost in clathrin- or μ1A/μ1B-depleted cells (Fig. 6A,B; supplementary material Fig. S1D,F). Depletion of clathrin increased the protein levels of Wls–MycHis and TfR, which is consistent with the role of clathrin in mediating degradation of plasma membrane proteins (Raiborg et al., 2006). Under these conditions, the apical and basolateral secretion of Wnt3a but not the apical secretion of Wnt11 was decreased (Fig. 6A,B). However, the depletion of μ4 did not affect the polarized distribution of Wnt11, Wnt3a, Wls and TfR (Fig. 6C; supplementary material Fig. S1D). Therefore, it is conceivable that Wls is sorted to the basolateral membrane mainly with clathrin and AP-1 and that Wnt3a secretion from the TGN is regulated by the complex of Wls, clathrin and AP-1.

Fig. 6.

Requirement of clathrin, AP-1 and AP-2 in the apical and basolateral secretion of Wnt3a. After MDCK cells expressing Wnt11, Wnt3a or Wls–MycHis were transfected with clathrin (Cla) (A), μ1A and μ1B (B), μ4 (C) or μ2 (D) siRNA, the cells were subjected to the apical-basolateral sorting assay. The signal of Wnt11 precipitated using Blue Sepharose in the control apical fraction or the signal of Wnt3a or Wls precipitated using NeutrAvidin agarose in the control basolateral fraction was set to 100%. The results shown are means ± s.e.m. from four independent experiments.

Fig. 6.

Requirement of clathrin, AP-1 and AP-2 in the apical and basolateral secretion of Wnt3a. After MDCK cells expressing Wnt11, Wnt3a or Wls–MycHis were transfected with clathrin (Cla) (A), μ1A and μ1B (B), μ4 (C) or μ2 (D) siRNA, the cells were subjected to the apical-basolateral sorting assay. The signal of Wnt11 precipitated using Blue Sepharose in the control apical fraction or the signal of Wnt3a or Wls precipitated using NeutrAvidin agarose in the control basolateral fraction was set to 100%. The results shown are means ± s.e.m. from four independent experiments.

It has been suggested that AP-2 and clathrin are involved in endocytosis of Wls and required for Wg secretion in Drosophila and C. elegans (Belenkaya et al., 2008; Pan et al., 2008; Port et al., 2008; Yang et al., 2008). μ2-HA, a medium subunit of AP-2, indeed formed a complex with Wls–MycHis and endogenous clathrin and adaptin-α (supplementary material Fig. S6B). Expression of μ2-HA enhanced the formation of a complex of Wls–MycHis with clathrin and adaptin-α (supplementary material Fig. S6B). Knockdown of μ2 decreased the adaptin-α proteins levels (supplementary material Fig. S1D,E) and led to the accumulation of Wls and TfR in the apical and basolateral membranes (Fig. 6D). Under these conditions, the apical and basolateral secretion of Wnt3a was decreased, but the apical secretion of Wnt11 was not (Fig. 6D). Therefore, the recycling of Wls by AP-2 might be required for the secretion of Wnt3a, but not for that of Wnt11, from the TGN. Consistent with these results, the apical secretion of Wnt11/3a was not decreased in clathrin-, μ1A/μ1B- and μ2-depleted cells (supplementary material Fig. S7), suggesting that the addition of the hybrid/complex-type glycan prevents Wnt3a from being secreted with the Wls complex.

Discussion

Conserved and unique glycosylation sites in Wnts

In this study we showed for the first time the glycan profiles of Wnt11 and Wnt3a. N90 and N300 in Wnt11 and N87 and N298 in Wnt3a were modified with high-mannose-type glycans, although small ratios of hybrid-type glycans were included at N300 of Wnt11. A single mutation in these sites of Wnt11 and Wnt3a did not affect their secretion, suggesting that one of two glycans at these sites is sufficient for maintaining the overall folded structure. These Asn residues are highly conserved in the mouse Wnt family (supplementary material Fig. S3). Among 19 mouse Wnts, 11 Wnts have two conserved Asn residues for glycosylation (Asn-X-Ser/Thr), and 6 Wnts have one conserved Asn residue. Although Wnt8a and Wnt8b lack these conserved Asn residues, they have possible glycosylation sites (N262 in Wnt8a and N258 in Wnt8b), which are located close to conserved Asn residues in Wnt11 (N300) and Wnt3a (N298). The glycosylation sites may play a role in maintaining the structure of Wnt8a and Wnt8b.

N40 in Wnt11 is unique in that it is not conserved in other mouse Wnts and located in the most N-terminal region. Because glycosylation at N40 in the ER was essential for the ER exit of Wnt11, the attachment of glycan to this site could be important for the structural integrity of Wnt11. Because Wnt11/3a was secreted, the addition of glycan to the N-terminal region does not appear to affect the folded structure of other Wnts. Therefore, the structure of Wnt11 might be different from other Wnts.

Although N160 and N304 in Wnt11 are possible glycosylation sites, their glycosylation was not detected by mass spectrometry. Therefore, it cannot be concluded that the Asn residue in the consensus sequence is indeed glycosylated until purified Wnt is analyzed by mass spectrometry. Mutational analyses for glycosylation sites may not always provide the correct results, because N304 might have been glycosylated instead when N300 was mutated.

Roles of Wls in the balsolateral secretion of Wnt3a in polarized epithelial cells

Wls has been shown to be a key molecule for the secretion of Wnt in non-polarized cells (Bänziger et al., 2006; Coombs et al., 2010; Jin et al., 2010; Port and Basler, 2010). In this study Wnt11 was found to be modified with palmitoleic acid at Ser215 as well as Wnt3a (Takada et al., 2006). The lack of palmitoleic acid modification in Wnt11 and Wnt3a resulted in accumulation of the mutant in the Golgi and ER fractions and failure to form a complex with Wls (Coombs et al., 2010; Takada et al., 2006) (this study). Porcupine catalyzes the lipid modification of Wnts in the ER, and Wls is thought to be required for the transport of lipid-modified Wnts between the ER and the Golgi, and in the Golgi, because Wls is localized to the cell surface membrane, endosomes, the Golgi and the ER (Coombs et al., 2010; Port and Basler, 2010). Consistently, depletion of Wls suppressed the secretion of Wnt11 and Wnt3a completely in polarized MDCK cells.

It was also demonstrated that Wls promotes the transport of Wnts from the TGN to the cell surface membrane (Port and Basler, 2010). Most of Wls and Wnt3a were sorted basolaterally, whereas Wnt11 is secreted apically. Therefore, it is possible that Wnt3a-loaded vesicles destined for the plasma membrane contain Wls but Wnt11-loaded vesicles do not. TfR is trafficked to the basolateral membrane by AP-1 and clathrin (Deborde et al., 2008; Gravotta et al., 2012). Among the subunits of AP-1, β1 interacts with clathrin and μ1 binds to the tyrosine-based sorting signal of basolateral proteins, such as TfR and low-density lipoprotein receptor (Boehm and Bonifacino, 2001; Fölsch, 2005). Wls formed a complex with μ1 and clathrin, and knockdown of these molecules suppressed the apical and basolateral secretion of Wnt3a but not the apical secretion of Wnt11. Therefore, Wls is sorted to the basolateral membrane mainly with clathrin and AP-1, and the tertiary complex of Wls, AP-1 and clathrin could sort Wnt3a to the plasma membrane. Wls is also taken up from the cell surface to the endosomes and the Golgi by clathrin/AP-2-mediated endocytosis with the retromer (Pan et al., 2008; Port and Basler, 2010; Port et al., 2008; Yang et al., 2008). Our results showed that the recycling of Wls from the cell surface membrane is required for the apical and basolateral secretion of Wnt3a but not the apical secretion of Wnt11. It is intriguing to speculate that Wls synthesized in the ER is essential for the transport of Wnt3a and Wnt11 to the Golgi and that Wls recycled to the Golgi is also necessary for the transport of Wnt3a to the final destination in polarized epithelial cells but not for that of Wnt11. It would be necessary to determine the glycan profiles of other Wnts.

Possible mechanism of the apical secretion of Wnt11

We found that modification with complex/hybrid-type glycan at N40 played roles in the apical secretion of Wnt11 in polarized MDCK cells. Furthermore, Wnt11/3a was secreted apically similar to Wnt11, suggesting that apical signals, such as the complex/hybrid-type glycan of the N-terminal region, are dominant over basolateral signals. The structures specific to the complex/hybrid-type glycan might be important for the determination of the apical secretion of Wnt11. It is noted that Wnt1 and Wg have possible glycosylation sites in the N-terminal region, similar to Wnt11 and that Wg was shown to be secreted apically in the Drosophila disc epithelium (Gallet et al., 2008).

It has been reported that N-glycans are apical sorting signals in some transmembrane and secreted proteins (Cao et al., 2012; Huet et al., 2003; Rodriguez-Boulan et al., 2005; Vagin et al., 2009), but the mechanism has not yet been elucidated. One potential sorting receptor is galectin-3, which binds to galactose and utilizes alternative pathways independent of the passage through the ER and the Golgi (Boscher et al., 2011; Hughes, 1999). Galectin-3 accumulates in acidified subapical endosomal compartments but not in the TGN of COS-1 and MDCK cells (Schneider et al., 2010). It was also shown that galectin-3 is secreted apically in filter-grown MDCK cells (Lindstedt et al., 1993; Schneider et al., 2010) and that depletion of galectin-3 from MDCK cells results in missorting of apical membrane proteins, such as lactose-phlorizin hydrolase (LPH), p75 and gp114 (Delacour et al., 2006). Our results showed that depletion of galectin-3 inhibits the apical secretion of Wnt11 and Wnt11/3a but not the polarized secretion of Wnt3a. Wnt11 might interact with galectin-3 in the endosomal vesicles fused with galectin-3-containing vesicles, and galectin-3 might be involved in the transport of Wnt11 to the apical membrane.

Unlike Wnt3a, the apical secretion of Wnt11 and Wnt11/3a did not require clathrin, AP-1 and AP-2. Thus, Wnt11 and Wnt3a are trafficked to their final destinations by different mechanisms. The complex/hybrid-type glycan at N40 might prevent Wnt11 from being sorted to the basolateral membrane by inhibiting the interaction of Wnt11 with the Wls complex when Wnt11-loaded vesicles are released from the TGN. What is the physiological relevance of the apical secretion of Wnt11? Galectin-3 is detected beneath the brush border of mouse intestine, and galectin-3 knockout causes the mislocalization of LPH and perturbation of the apical brush border marker villin into the lateral membrane of enterocytes (Delacour et al., 2008; Lindstedt et al., 1993). Therefore, it is intriguing to speculate that Wnt11 plays roles in the maintenance of the apicobasal polarity of epithelial cells.

Materials and methods

Materials and chemicals

pCS2/mouse Wnt11 and pPGK-neo/mouse Wnt3a were provided by Drs D. Wu (Yale University, New Haven, CT, USA) and S. Takada (National Institutes of Natural Sciences, Okazaki, Japan), respectively. pEGFPc1/galectin-3, and pCB6/μ1A-HA and pCB6/μ1B-HA were from Drs G. Fenteany (University of Connecticut, CT, USA) and H. Fölsch (Northwestern University, Chicago, IL, USA), respectively. pcDNA3.1hygro/Gnt-I was from Drs H. Narimatsu and N. Taniguchi (Advanced Industrial Science and Technology, Tsukuba, Japan). pRK5/IgG and pGEX-αPAK-CRIB were provided by Drs J. C. Hsieh (State University of New York, Stony Brook, NY, USA) and K. Kaibuchi (Nagoya University, Nagoya, Japan), respectively.

Wnt3a was purified to homogeneity, and an anti-Wnt3a antibody (for immunostaining) was generated as described previously (Kishida et al., 2004; Komekado et al., 2007). An anti-Wnt11 antibody was generated in rabbits by immunization with a synthetic peptide corresponding to residues 141–157 of mouse Wnt11. Anti-Myc (tag) and anti-HA (for immunoprecipitation) antibodies were generated from 9E10 and 12CA5 cells, respectively. Anti-HA (16B12; for immunoblotting) and anti-FLAG antibodies were purchased from Covance (Princeton, NJ, USA) and Sigma-Aldrich (St. Louis, MO, USA), respectively. Anti-β-catenin, anti-GSK-3β, anti-heat shock protein 90 (HSP90), anti-clathrin, anti-adaptin-γ, anti-adaptin-α and anti-calnexin antibodies were from BD Biosciences (San Jose, CA, USA). Anti-Wnt3a (for immunoblotting), anti-Wnt5a/b and anti-Dvl2 antibodies were from Cell Signaling Technology (Beverly, MA, USA). Anti-ezrin, anti-transferrin receptor (TfR) and anti-podocalyxin antibodies were from Abcam (Cambridge, UK), Invitrogen (Carlsbad, MA, USA) and Santa Cruz (San Diego, CA, USA), respectively. PNGase F and Endo Hf were from Roche Diagnostics GmbH (Mannheim, Germany) and New England Biolabs (Beverly, MA, USA), respectively. EZ-Link Sulfo-NHS-biotin and NeutrAvidin agarose were purchased from Pierce Biotechnology (Rockford, IL, USA). Gnt-I, GFP-galectin-3 and Wls-MycHis cDNAs were cloned into pCSII-CMV-MCS-IRES2-Bsd to construct lentiviral vectors (Takara Bio Inc., Shiga, Japan). Human μ4 cDNA was cloned from a human cDNA library (Invitrogen), and subcloned into pCGN-HA vector. pcDNA3.1/hWls-MycHis was constructed as described previously (Komekado et al., 2007). Standard recombinant DNA techniques were used to construct pPGK-neo/Wnt3aS209A and pPGK-neo/WT Wnt11 or its mutants. The RNA duplexes, and forward and reverse primers for quantitative RT-PCR used in this study are shown in supplementary material Tables S6, S7, respectively. Other materials were obtained from commercial sources.

Cell culture

MDCK, L, X293T and NIH3T3 cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS). HEK293T cells were maintained in DMEM/Ham's F12 supplemented with 10% FBS. L or MDCK cells stably expressing WT Wnt11, WT Wnt3a or their mutants and Wls–MycHis were generated by selection with G418. MDCK cells stably co-expressing WT Wnt11 with Gnt-I or GFP–galectin-3 and co-expressing WT Wnt3a with Wls–MycHis were generated by infection with lentiviruses. Then, the cells were selected and maintained in the same medium containing 800 µg/ml G418 and 2.5 µg/ml blastcidin S. Three-independent clones of each stable cell line were used for experiments to avoid clonal variation, and similar results were obtained.

Knockdown of proteins in MDCK cells by siRNA

To transfect siRNA into MDCK cells expressing Wnt11, Wnt3a, Wnt11/3a or Wls–MycHis, trypsinized MDCK cells were suspended in Optimem (Invitrogen) at 106 cells per 100 µl, added to 160 pmol of siRNA, and electroporated using NEPA21 (NEPAGENE, Tokyo, Japan) with five square pulses of 20 V of 50 mseconds duration. After cells were consecutively transfected twice at 2-day intervals, knockdown cells (2×105 cells) were seeded on Transwell polycarbonate filters for a further 3 days. Knockdown efficiency was determined by RT-PCR or immunoblotting.

Purification of Wnt11

After L cells expressing Wnt11 were incubated in the presence of heparin (75 µg/ml) for 2 days, Wnt11 CM was collected. 150 ml 10% Triton X-100 was added to 2850 ml Wnt11 CM and the mixture was applied to Blue Sepharose HP in an XK 26/20 column [column volume (CV) 50 ml; GE Healthcare Bio-Sciences, Buckinghamshire, UK] equilibrated with binding buffer {(20 mM Tris-HCl, pH 7.5, and 1% 3-[(3-cholamidopropyl) dimethylammonio] propanesulfonic acid (CHAPS)} containing 150 mM KCl. After the column was washed with 750 ml of binding buffer, elution was performed in a stepwise manner with 250 ml of binding buffer containing 1.5 M KCl at a flow rate of 5 ml/min. Fractions of 12.5 ml each were collected. An aliquot (20 µl) of each fraction was probed using an anti-Wnt11 antibody. To examine Wnt11 activity, NIH3T3 cells were stimulated with an aliquot (20 µl) of each fraction for 1 hour and the lysates were probed using an anti-Dvl2 antibody (the Dvl2 mobility shift assay). Wnt11 appeared in fractions 2–10, and these fractions (110 ml, 41.8 mg of protein) were pooled and concentrated to 9.5 ml using a 30 kDa molecular mass cut-off membrane (Pall Corporation, Ann Arbor, MI).

The concentrate (8.5 ml, 39 mg of protein) was applied to a HiLoad Superdex 200 column (2.6×60 cm; CV 320 ml; GE Healthcare Bio-Sciences) equilibrated with 20 mM Tris-HCl, pH 7.5, and 1% CHAPS containing 1 M NaCl. Elution was performed with the same buffer at a flow rate of 2.5 ml/min. Fractions of 2.5 ml each were collected. When an aliquot (20 µl) of each fraction was probed using an anti-Wnt11 antibody and subjected to the Dvl2 mobility shift assay, there was a single broad peak from fractions 77–84 (supplementary material Fig. S2C). These fractions (20 ml, 1.12 mg of protein) were applied to a Ceramic Hydroxyapatite Type I column (CV 1 ml; Bio-Rad Laboratories, Hercules, CA, USA) equilibrated with 1 mM Na2HPO4/NaH2PO4 (pH 7.2), 1% CHAPS and 1 M NaCl. After the column was washed with 5 ml of the same buffer, elution was performed in a stepwise manner with 5 ml 0.2 M Na2HPO4/NaH2PO4 (pH 7.2), 1% CHAPS and 1 M NaCl. Fractions of 1 ml each were collected. When an aliquot (20 µl) of each fraction was probed using an anti-Wnt11 antibody, a single broad peak was seen from the flow-through fraction. The flow-through fractions (20 ml, 0.2 mg of protein) were applied to a Concanavalin A column (CV 1 ml) (GE Healthcare Bio-Sciences) equilibrated with 0.1 M boric acid buffer (pH 6.5), 1 M NaCl, and 1% CHAPS. After the column was washed with the same buffer, elution was performed in a stepwise manner with 40 ml of the buffer containing 2.5 M α-methyl-D-glucoside. Fractions of 1 ml each were collected. When an aliquot (20 µl) of each fraction was probed using an anti-Wnt11 antibody, there was a single broad peak from fractions 1–20.

These fractions (20 ml, 0.14 mg of protein) were applied to a HiTrap Chelating HP column (CV 1 ml; GE Healthcare Bio-Sciences) loaded with 1 ml of 0.1 M CuSO4 and then equilibrated with 20 mM Na2HPO4/NaH2PO4 (pH 7.4), 1 M NaCl, 1% CHAPS and 3.5 mM imidazole. After the column was washed with 10 ml of same buffer, elution was performed with a 10 ml linear gradient of imidazole (3.5–40 mM) in 20 mM Na2HPO4/NaH2PO4 (pH 7.4), 1 M NaCl and 1% CHAPS at a flow rate of 1 ml/min. Fractions of 0.5 ml each were collected, and Wnt11 appeared mainly in fractions 9–12 (2 ml, 28 µg). In these fractions Wnt11 was purified to near homogeneity by assessment with SDS-PAGE and Coomassie Brilliant Blue staining (supplementary material Fig. S2A).

Preparation of glycopeptides and acylpeptides of Wnt3a and Wnt11

Wnt3a (1 µg) or Wnt11 (0.5 µg) were separated by SDS-PAGE on a 5–20% gradient ready-gel (ATTO, Tokyo, Japan), followed by Coomassie Brilliant Blue staining. The protein bands excised from the gel were incubated with 50 nM trypsin (Promega, Madison, WI, USA) in 20 mM ammonium bicarbonate and 10% acetonitrile at 37°C for 16 hours. The tryptic peptides were extracted from the gel slice with 100 µl of 0.1% trifluoroacetic acid in 10% acetonitrile/H2O for glycosylation analysis, followed by 100 µl of 0.1% trifluoroacetic acid in 60% acetonitrile/H2O for lipidation analysis.

Nano-flow liquid chromatography/matrix-assisted laser desorption/ionization mass spectrometry

The tryptic digests of Wnt3a and Wnt11 were injected into an Ultimate nano-LC system (Dionex, Idstein, Germany), where the digested peptides were first concentrated using a C18 trapping column (0.3×1 mm; Dionex) with 0.1% trifluoroacetic acid in water at a flow rate of 20 µl/min for 3 minutes, and then separated using a C18-Pepmap column (0.075×150 mm; Dionex) by increasing the concentration of 0.1% trifluoroacetic acid in acetonitrile from 5 to 85% over a period of 60 minutes at a flow rate of 200 nl/minute. Overall peptide identification was carried out using a MALDI-TOF/TOF (4700 proteomics analyzer; Applied Biosystems, Framingham, MA, USA) as described previously (Awada et al., 2010).

Nano-flow liquid chromatography/electrospray ionization mass spectrometry

To examine the glycan chains and relative abundances of each glycoform at each glycosylation site, the above tryptic digests were applied to a PicoFrit C18 column (0.075×100 mm, 3 µm particle size, NewObjective) equilibrated with 0.3% formic acid in 2% acetonitrile/H2O. After the column was washed with the same solvent, elution was performed with a linear gradient of acetonitrile (2–49%) in 0.3% formic acid at a flow rate of 200 nl/minute. The eluates were continuously analyzed by a linear ion-trap/time-of-flight mass spectrometer equipped with a nano-flow electrospray ion source (Hitachi NanoFrontier LD, Hitachi High-Technology, Tokyo, Japan) as described preciously (Kimura et al., 2013).

Secretion of Wnt proteins into the apical or basolateral fractions in polarized MDCK cells (apical-basolateral sorting assay)

MDCK cells (2×105 cells) expressing Wnt11, Wnt3a or Wnt11/3a were seeded on Transwell polycarbonate filters (Corning Costar Quality Biological, Gaithersburg, MD, USA) for 72 hours. To detect Wnts secreted into the culture medium, media from the apical and basolateral sites were incubated with Blue Sepharose for 2 hours at 4°C, followed by centrifugation. The precipitates were probed using anti-Wnt11 or anti-Wnt3a antibodies. Wnt11/3a was recognized by anti-Wnt3a antibody. To detect cell-surface-membrane-associated Wnts, the apical or basolateral surface membranes of the cells were selectively incubated with 0.5 mg/ml sulfo-NHS-LC-biotin for 30 minutes at 4°C (Sakane et al., 2012; Yamamoto et al., 2006). Biotinylated proteins were precipitated using NeutrAvidin agarose (Pierce) and the precipitates were probed using anti-Wnt11 or anti-Wnt3a antibodies.

The signals of Wnt11 or Wnt3a were quantified using NIH ImageJ. In Fig. 1B the sum of Wnts secreted from MDCK cells into the apical and basolateral fractions was set to 100%, and the results were expressed as the ratio of Wnts in the apical or basolateral fraction. When Wls, Gnt-I, clathrin, μ1A/μ1B or μ2 were knocked down or MDCK cells were treated with 2 µg/ml kifunensine or 200 µg/ml dMM for 24 hours, the release of Wnts was reduced severely. Therefore, in these experiments the signal of Wnt11 precipitated using Blue Sepharose in the control apical fraction or the signal of Wnt3a precipitated using NeutrAvidin agarose in the control basolateral fraction was set to 100%. The polarized sorting assay of Wls, podocalyxin or TfR to the apical and basolateral membranes was performed by the same methods using NeutrAvidin agarose (Gottardi et al., 1995; Gravotta et al., 2012). Wls precipitated using NeutrAvidin agarose in the control basolateral fraction was set to 100%.

Post-Golgi sorting assay

Polarized MDCK cells expressing Wnt11 on a Transwell polycarbonate filter were pre-incubated at 20°C for 2 hours to accumulate the proteins in the Golgi (Paladino et al., 2006). In the last hour at 20°C, they were treated with 100 µg/ml cycloheximide. After replacing with fresh culture medium containing 100 µg/ml cycloheximide, the cells were warmed to 37°C for the indicated times and then Wnt11 released into the apical medium was quantified.

Three-dimensional culture of MDCK cells

MDCK cells were suspended in Matrigel at a density of 1×105 cells/ml, and 80 µl of the cell suspension was mounted on a round coverslip. After incubation at 37°C for 30 minutes to solidify the gel, the coverslip was transferred to a 24-well culture plate and 1 ml of DMEM containing FBS added to the Matrigel and further incubation was performed for 72 hours. To visualize intracellular apical or basolateral localization of Wnt11, Wnt3a or Wnt11/3a in 3D Matrigel, the cells were fixed for 30 minutes in PBS containing 4% (w/v) paraformaldehyde (PFA) and then permeabilized with PBS containing 0.5% (w/v) Triton X-100 and 4% (w/v) BSA for 30 minutes [(+)permeabilization]. The cells were stained with anti-β-catenin and anti-Wnt11 or anti-Wnt3a antibodies for 2 hours. Cells were washed with 4% BSA in PBS three times and incubated for 2 hours with goat Alexa-Fluor-488-conjugated anti-rabbit, Alexa-Fluor-546-conjugated anti-mouse IgG and Alexa-Fluor-633-conjugated phalloidin (Invitrogen) to visualize actin. To visualize basolaterally secreted Wnts in Matrigel, after the cells were incubated with anti-Wnt11 or anti-Wnt3a antibody for 2 hours at 37°C without permeabilization [(−)permeabilization], the cells were fixed in PBS containing 4% PFA for 30 minutes at room temperature. After this, non-specific sites were blocked with 4% BSA in PBS for 30 minutes and cells were incubated with goat Alexa-Fluor-488-conjugated anti-rabbit IgG for 2 hours. The cells were then washed three times, mounted on slides, and viewed using a confocal microscope (LSM510, Carl Zeiss, Jana, Germany).

Statistical analysis

The experiments were performed at least four times and the results were expressed as means ± s.e.m. Statistical analysis was performed using StatView software (SAS Institute Inc.). Differences between the data were tested for statistical significance using t-tests. P-values less than 0.01 or 0.05 were considered statistically significant.

Other methods

Immunocytochemical studies of Wnt11 in mouse embryonic kidney were performed as described previously (Yamamoto et al., 2009). The Triton X-114 phase separation assay and preparation of CM, the extracellular matrix (ECM) fraction, and lysates from Wnt-producing cells were performed as described previously (Komekado et al., 2007; Kurayoshi et al., 2007). Sucrose density gradient fractionation was performed as described (Komekado et al., 2007) except that a linear sucrose gradient (10–40%) was used. Formation of a complex between Wls-MycHis, 1A-HA, μ1B-HA, μ2-HA or μ4-HA and clathrin was examined as described previously (Matsumoto et al., 2010).

Acknowledgements

We thank Drs D. Wu, S. Takada, G. Fenteany, H. Fölsch, F. Narimatsu and N. Taniguchi, J. C. Hsieh and K. Kaibuchi for donating plasmids.

Author contributions

H.Y. designed experiments, carried out purification of Wnt proteins and apical-basolateral sorting assay, and wrote the manuscript. H.H., H.S. and I.T. carried out apical-basolateral sorting assay and immunohisotochemical assay. Y.T. performed purification of Wnt proteins and fractionation assay. C.A. and T.T. analyzed glycosylation and lipidation profiles of Wnts by mass-spectrometry and provided advice. A.K. designed experiments and wrote the manuscript.

Funding

This work was supported by the Ministry of Education, Science, and Culture of Japan through Grants-in-Aid for Scientific Research (2009-2011) [grant number 21249017 to A.K.]; and for Scientific Research on Priority Areas (2011-2012) [grant number 23112004 to A.K.].

References

Awada
C.
,
Sato
T.
,
Takao
T.
(
2010
).
Affinity-trap polyacrylamide gel electrophoresis: a novel method of capturing specific proteins by electro-transfer.
Anal. Chem.
82
,
755
761
.
Bänziger
C.
,
Soldini
D.
,
Schütt
C.
,
Zipperlen
P.
,
Hausmann
G.
,
Basler
K.
(
2006
).
Wntless, a conserved membrane protein dedicated to the secretion of Wnt proteins from signaling cells.
Cell
125
,
509
522
.
Belenkaya
T. Y.
,
Wu
Y.
,
Tang
X.
,
Zhou
B.
,
Cheng
L.
,
Sharma
Y. V.
,
Yan
D.
,
Selva
E. M.
,
Lin
X.
(
2008
).
The retromer complex influences Wnt secretion by recycling wntless from endosomes to the trans-Golgi network.
Dev. Cell
14
,
120
131
.
Bischoff
J.
,
Kornfeld
R.
(
1984
).
The effect of 1-deoxymannojirimycin on rat liver alpha-mannosidases.
Biochem. Biophys. Res. Commun.
125
,
324
331
.
Boehm
M.
,
Bonifacino
J. S.
(
2001
).
Adaptins: the final recount.
Mol. Biol. Cell
12
,
2907
2920
.
Boscher
C.
,
Dennis
J. W.
,
Nabi
I. R.
(
2011
).
Glycosylation, galectins and cellular signaling.
Curr. Opin. Cell Biol.
23
,
383
392
.
Buechling
T.
,
Chaudhary
V.
,
Spirohn
K.
,
Weiss
M.
,
Boutros
M.
(
2011
).
p24 proteins are required for secretion of Wnt ligands.
EMBO Rep.
12
,
1265
1272
.
Cao
X.
,
Surma
M. A.
,
Simons
K.
(
2012
).
Polarized sorting and trafficking in epithelial cells.
Cell Res.
22
,
793
805
.
Coombs
G. S.
,
Yu
J.
,
Canning
C. A.
,
Veltri
C. A.
,
Covey
T. M.
,
Cheong
J. K.
,
Utomo
V.
,
Banerjee
N.
,
Zhang
Z. H.
,
Jadulco
R. C.
 et al. (
2010
).
WLS-dependent secretion of WNT3A requires Ser209 acylation and vacuolar acidification.
J. Cell Sci.
123
,
3357
3367
.
Dean
C. H.
,
Miller
L. A.
,
Smith
A. N.
,
Dufort
D.
,
Lang
R. A.
,
Niswander
L. A.
(
2005
).
Canonical Wnt signaling negatively regulates branching morphogenesis of the lung and lacrimal gland.
Dev. Biol.
286
,
270
286
.
Deborde
S.
,
Perret
E.
,
Gravotta
D.
,
Deora
A.
,
Salvarezza
S.
,
Schreiner
R.
,
Rodriguez-Boulan
E.
(
2008
).
Clathrin is a key regulator of basolateral polarity.
Nature
452
,
719
723
.
Delacour
D.
,
Gouyer
V.
,
Zanetta
J. P.
,
Drobecq
H.
,
Leteurtre
E.
,
Grard
G.
,
Moreau-Hannedouche
O.
,
Maes
E.
,
Pons
A.
,
André
S.
 et al. (
2005
).
Galectin-4 and sulfatides in apical membrane trafficking in enterocyte-like cells.
J. Cell Biol.
169
,
491
501
.
Delacour
D.
,
Cramm-Behrens
C. I.
,
Drobecq
H.
,
Le Bivic
A.
,
Naim
H. Y.
,
Jacob
R.
(
2006
).
Requirement for galectin-3 in apical protein sorting.
Curr. Biol.
16
,
408
414
.
Delacour
D.
,
Greb
C.
,
Koch
A.
,
Salomonsson
E.
,
Leffler
H.
,
Le Bivic
A.
,
Jacob
R.
(
2007
).
Apical sorting by galectin-3-dependent glycoprotein clustering.
Traffic
8
,
379
388
.
Delacour
D.
,
Koch
A.
,
Ackermann
W.
,
Eude-Le Parco
I.
,
Elsässer
H. P.
,
Poirier
F.
,
Jacob
R.
(
2008
).
Loss of galectin-3 impairs membrane polarisation of mouse enterocytes in vivo.
J. Cell Sci.
121
,
458
465
.
Elbein
A. D.
,
Tropea
J. E.
,
Mitchell
M.
,
Kaushal
G. P.
(
1990
).
Kifunensine, a potent inhibitor of the glycoprotein processing mannosidase I. J. Biol.
Chem.
265
,
15599
15605
.
Fölsch
H.
(
2005
).
The building blocks for basolateral vesicles in polarized epithelial cells.
Trends Cell Biol.
15
,
222
228
.
Fuhrmann
U.
,
Bause
E.
,
Legler
G.
,
Ploegh
H.
(
1984
).
Novel mannosidase inhibitor blocking conversion of high mannose to complex oligosaccharides.
Nature
307
,
755
758
.
Gallet
A.
,
Staccini-Lavenant
L.
,
Thérond
P. P.
(
2008
).
Cellular trafficking of the glypican Dally-like is required for full-strength Hedgehog signaling and wingless transcytosis.
Dev. Cell
14
,
712
725
.
Gottardi
C. J.
,
Dunbar
L. A.
,
Caplan
M. J.
(
1995
).
Biotinylation and assessment of membrane polarity: caveats and methodological concerns.
Am. J. Physiol.
268
,
F285
F295
.
Gravotta
D.
,
Deora
A.
,
Perret
E.
,
Oyanadel
C.
,
Soza
A.
,
Schreiner
R.
,
Gonzalez
A.
,
Rodriguez-Boulan
E.
(
2007
).
AP1B sorts basolateral proteins in recycling and biosynthetic routes of MDCK cells.
Proc. Natl. Acad. Sci. USA
104
,
1564
1569
.
Gravotta
D.
,
Carvajal-Gonzalez
J. M.
,
Mattera
R.
,
Deborde
S.
,
Banfelder
J. R.
,
Bonifacino
J. S.
,
Rodriguez-Boulan
E.
(
2012
).
The clathrin adaptor AP-1A mediates basolateral polarity.
Dev. Cell
22
,
811
823
.
Hayashi
K.
,
Yoshioka
S.
,
Reardon
S. N.
,
Rucker
E. B.
 3rd
,
Spencer
T. E.
,
DeMayo
F. J.
,
Lydon
J. P.
,
MacLean
J. A. I. I.
(
2011
).
WNTs in the neonatal mouse uterus: potential regulation of endometrial gland development.
Biol. Reprod.
84
,
308
319
.
Heisenberg
C. P.
,
Tada
M.
,
Rauch
G. J.
,
Saúde
L.
,
Concha
M. L.
,
Geisler
R.
,
Stemple
D. L.
,
Smith
J. C.
,
Wilson
S. W.
(
2000
).
Silberblick/Wnt11 mediates convergent extension movements during zebrafish gastrulation.
Nature
405
,
76
81
.
Huet
G.
,
Gouyer
V.
,
Delacour
D.
,
Richet
C.
,
Zanetta
J. P.
,
Delannoy
P.
,
Degand
P.
(
2003
).
Involvement of glycosylation in the intracellular trafficking of glycoproteins in polarized epithelial cells.
Biochimie
85
,
323
330
.
Hughes
R. C.
(
1999
).
Secretion of the galectin family of mammalian carbohydrate-binding proteins.
Biochim. Biophys. Acta
1473
,
172
185
.
Jin
J.
,
Morse
M.
,
Frey
C.
,
Petko
J.
,
Levenson
R.
(
2010
).
Expression of GPR177 (Wntless/Evi/Sprinter), a highly conserved Wnt-transport protein, in rat tissues, zebrafish embryos, and cultured human cells.
Dev. Dyn.
239
,
2426
2434
.
Kikuchi
A.
,
Yamamoto
H.
,
Sato
A.
,
Matsumoto
S.
(
2011
).
New insights into the mechanism of Wnt signaling pathway activation.
Int. Rev. Cell Mol. Biol.
291
,
21
71
.
Kimura
M.
,
Kose
S.
,
Okumura
N.
,
Imai
K.
,
Furuta
M.
,
Sakiyama
N.
,
Tomii
K.
,
Horton
P.
,
Takao
T.
,
Imamoto
N.
(
2013
).
Identification of cargo proteins specific for the nucleocytoplasmic transport carrier transportin by combination of an in vitro transport system and SILAC-based quantitative proteomics.
Mol. Cell. Proteomics
12
,
145
157
.
Kishida
S.
,
Yamamoto
H.
,
Kikuchi
A.
(
2004
).
Wnt-3a and Dvl induce neurite retraction by activating Rho-associated kinase.
Mol. Cell. Biol.
24
,
4487
4501
.
Komekado
H.
,
Yamamoto
H.
,
Chiba
T.
,
Kikuchi
A.
(
2007
).
Glycosylation and palmitoylation of Wnt-3a are coupled to produce an active form of Wnt-3a.
Genes Cells
12
,
521
534
.
Kurayoshi
M.
,
Yamamoto
H.
,
Izumi
S.
,
Kikuchi
A.
(
2007
).
Post-translational palmitoylation and glycosylation of Wnt-5a are necessary for its signalling.
Biochem. J.
402
,
515
523
.
Lickert
H.
,
Kispert
A.
,
Kutsch
S.
,
Kemler
R.
(
2001
).
Expression patterns of Wnt genes in mouse gut development.
Mech. Dev.
105
,
181
184
.
Lindstedt
R.
,
Apodaca
G.
,
Barondes
S. H.
,
Mostov
K. E.
,
Leffler
H.
(
1993
).
Apical secretion of a cytosolic protein by Madin-Darby canine kidney cells. Evidence for polarized release of an endogenous lectin by a nonclassical secretory pathway.
J. Biol. Chem.
268
,
11750
11757
.
Logan
C. Y.
,
Nusse
R.
(
2004
).
The Wnt signaling pathway in development and disease.
Annu. Rev. Cell Dev. Biol.
20
,
781
810
.
MacDonald
B. T.
,
Tamai
K.
,
He
X.
(
2009
).
Wnt/β-catenin signaling: components, mechanisms, and diseases.
Dev. Cell
17
,
9
26
.
Matsumoto
S.
,
Fumoto
K.
,
Okamoto
T.
,
Kaibuchi
K.
,
Kikuchi
A.
(
2010
).
Binding of APC and dishevelled mediates Wnt5a-regulated focal adhesion dynamics in migrating cells.
EMBO J.
29
,
1192
1204
.
O'Brien
L. E.
,
Zegers
M. M.
,
Mostov
K. E.
(
2002
).
Opinion: building epithelial architecture: insights from three-dimensional culture models.
Nat. Rev. Mol. Cell Biol.
3
,
531
537
.
Paladino
S.
,
Pocard
T.
,
Catino
M. A.
,
Zurzolo
C.
(
2006
).
GPI-anchored proteins are directly targeted to the apical surface in fully polarized MDCK cells.
J. Cell Biol.
172
,
1023
1034
.
Pan
C. L.
,
Baum
P. D.
,
Gu
M.
,
Jorgensen
E. M.
,
Clark
S. G.
,
Garriga
G.
(
2008
).
C. elegans AP-2 and retromer control Wnt signaling by regulating mig-14/Wntless.
Dev. Cell
14
,
132
139
.
Polakis
P.
(
2007
).
The many ways of Wnt in cancer.
Curr. Opin. Genet. Dev.
17
,
45
51
.
Port
F.
,
Basler
K.
(
2010
).
Wnt trafficking: new insights into Wnt maturation, secretion and spreading.
Traffic
11
,
1265
1271
.
Port
F.
,
Kuster
M.
,
Herr
P.
,
Furger
E.
,
Bänziger
C.
,
Hausmann
G.
,
Basler
K.
(
2008
).
Wingless secretion promotes and requires retromer-dependent cycling of Wntless.
Nat. Cell Biol.
10
,
178
185
.
Port
F.
,
Hausmann
G.
,
Basler
K.
(
2011
).
A genome-wide RNA interference screen uncovers two p24 proteins as regulators of Wingless secretion.
EMBO Rep.
12
,
1144
1152
.
Rademacher
T. W.
,
Parekh
R. B.
,
Dwek
R. A.
(
1988
).
Glycobiology.
Annu. Rev. Biochem.
57
,
785
838
.
Raiborg
C.
,
Wesche
J.
,
Malerød
L.
,
Stenmark
H.
(
2006
).
Flat clathrin coats on endosomes mediate degradative protein sorting by scaffolding Hrs in dynamic microdomains.
J. Cell Sci.
119
,
2414
2424
.
Rodriguez-Boulan
E.
,
Kreitzer
G.
,
Müsch
A.
(
2005
).
Organization of vesicular trafficking in epithelia.
Nat. Rev. Mol. Cell Biol.
6
,
233
247
.
Sakane
H.
,
Yamamoto
H.
,
Matsumoto
S.
,
Sato
A.
,
Kikuchi
A.
(
2012
).
Localization of glypican-4 in different membrane microdomains is involved in the regulation of Wnt signaling.
J. Cell Sci.
125
,
449
460
.
Satomi
Y.
,
Shimonishi
Y.
,
Hase
T.
,
Takao
T.
(
2004
).
Site-specific carbohydrate profiling of human transferrin by nano-flow liquid chromatography/electrospray ionization mass spectrometry.
Rapid Commun. Mass Spectrom.
18
,
2983
2988
.
Schneider
D.
,
Greb
C.
,
Koch
A.
,
Straube
T.
,
Elli
A.
,
Delacour
D.
,
Jacob
R.
(
2010
).
Trafficking of galectin-3 through endosomal organelles of polarized and non-polarized cells.
Eur. J. Cell Biol.
89
,
788
798
.
Simmen
T.
,
Höning
S.
,
Icking
A.
,
Tikkanen
R.
,
Hunziker
W.
(
2002
).
AP-4 binds basolateral signals and participates in basolateral sorting in epithelial MDCK cells.
Nat. Cell Biol.
4
,
154
159
.
Stechly
L.
,
Morelle
W.
,
Dessein
A. F.
,
André
S.
,
Grard
G.
,
Trinel
D.
,
Dejonghe
M. J.
,
Leteurtre
E.
,
Drobecq
H.
,
Trugnan
G.
 et al. (
2009
).
Galectin-4-regulated delivery of glycoproteins to the brush border membrane of enterocyte-like cells.
Traffic
10
,
438
450
.
Takada
R.
,
Satomi
Y.
,
Kurata
T.
,
Ueno
N.
,
Norioka
S.
,
Kondoh
H.
,
Takao
T.
,
Takada
S.
(
2006
).
Monounsaturated fatty acid modification of Wnt protein: its role in Wnt secretion.
Dev. Cell
11
,
791
801
.
Uysal-Onganer
P.
,
Kypta
R. M.
(
2012
).
Wnt11 in 2011 – the regulation and function of a non-canonical Wnt.
Acta Physiol. (Oxf.)
204
,
52
64
.
Vagin
O.
,
Kraut
J. A.
,
Sachs
G.
(
2009
).
Role of N-glycosylation in trafficking of apical membrane proteins in epithelia.
Am. J. Physiol. Renal Physiol.
296
,
F459
F469
.
Vainio
S.
,
Lin
Y.
(
2002
).
Coordinating early kidney development: lessons from gene targeting.
Nat. Rev. Genet.
3
,
533
543
.
Veeman
M. T.
,
Axelrod
J. D.
,
Moon
R. T.
(
2003
).
A second canon. Functions and mechanisms of β-catenin-independent Wnt signaling.
Dev. Cell
5
,
367
377
.
Yamamoto
H.
,
Komekado
H.
,
Kikuchi
A.
(
2006
).
Caveolin is necessary for Wnt-3a-dependent internalization of LRP6 and accumulation of β-catenin.
Dev. Cell
11
,
213
223
.
Yamamoto
H.
,
Kitadai
Y.
,
Yamamoto
H.
,
Oue
N.
,
Ohdan
H.
,
Yasui
W.
,
Kikuchi
A.
(
2009
).
Laminin γ2 mediates Wnt5a-induced invasion of gastric cancer cells.
Gastroenterology
137
,
242
252, e1-e6
.
Yang
P. T.
,
Lorenowicz
M. J.
,
Silhankova
M.
,
Coudreuse
D. Y.
,
Betist
M. C.
,
Korswagen
H. C.
(
2008
).
Wnt signaling requires retromer-dependent recycling of MIG-14/Wntless in Wnt-producing cells.
Dev. Cell
14
,
140
147
.

Supplementary information