The Addition of Bisecting N-Acetylglucosamine Residues to E-cadherin Down-regulates the Tyrosine Phosphorylation of β-Catenin*

The enzyme GnT-III (β1,4-N-acetylglucosaminyltransferase III) catalyzes the addition of a bisecting N-acetylglucosamine (GlcNAc) residue on glycoproteins. Our previous study described that the transfection of GnT-lll into mouse melanoma cells results in the enhanced expression of E-cadherin, which in turn leads to the suppression of lung metastasis. It has recently been proposed that the phosphorylation of a tyrosine residue of β-catenin is associated with cell migration. The present study reports on the importance of bisecting GlcNAc residues by GnT-lll on tyrosine phosphorylation of β-catenin using three types of cancer cell lines. An addition of bisecting GlcNAc residues to E-cadherin leads to an alteration in cell morphology and the localization of β-catenin after epidermal growth factor stimulation. These changes are the result of a down-regulation in the tyrosine phosphorylation of β-catenin. In addition, tyrosine phosphorylation of β-catenin by transfection of constitutively active c-src was suppressed in GnT-III transfectants as well as in the case of epidermal growth factor stimulation. Treatment with tunicamycin abolished any differences in β-catenin phosphorylation for the mockvis à vis the GnT-lll transfectants. Thus, the addition of a specific N-glycan structure, the bisecting GlcNAc to E-cadherin-β-catenin complex, down-regulates the intracellular signaling pathway, suggesting its implication in cell motility and the suppression of cancer metastasis.

The remodeling of cell surface oligosaccharides by glycosyltransferase gene transfection has led to a better understanding of the biological functions of several glycoproteins (reviewed in Ref. 1). Changes in adhesion molecules and growth factor receptors by modification of their oligosaccharide structure are associated with the function and the biological behavior of cancer cells (2)(3)(4). These phenomena appear to be linked to the progression of cancer metastasis and invasion. However, the mechanism for this is complicated. One of the likely contributing molecules is the homophilic adhesion molecule, E-cadherin (5,6), a decreased expression of which is highly associated with the initial step of metastasis in certain types of cancer (7)(8)(9)(10). E-cadherin contains five consensus N-glycosylation sites. These N-glycans on E-cadherin can be modified with a bisecting N-acetylglucosamine (GlcNAc) structure, which is the product of ␤1,4-N-acetylglucosaminyltransferase lll (GnT-III, 1 EC 2.4.1.144) (11). GnT-III catalyzes the attachment of a GlcNAc residue to a ␤1-4 mannose residue in the core region of Nglycans, as shown in Fig. 1.
In an earlier study, we demonstrated that the bisecting GlcNAc on E-cadherin suppressed lung metastasis of melanoma cells (12). However, the mechanism has not been fully investigated. The function of E-cadherin is also controlled via molecules from the cytoplasm, namely ␣-, ␤-, and ␥-catenins (13,14). The deletion or the mutational inactivation of ␤-catenin leads to the abolition of E-cadherin activity and tumor invasion (15,16). Recently, the biological significance of the tyrosine phosphorylation of ␤-catenin has been examined by several groups (17)(18)(19)(20). It is possible that this phosphorylation could disturb E-cadherin/cytoskeleton interactions (17). During cell migration, ␤-catenin accumulates in the cytosol in a free, uncomplexed, and tyrosine-phosphorylated form (19). In contrast, this type of ␤-catenin is not observed during the quiescent phase, suggesting that the tyrosine phosphorylation of ␤-catenin is linked to cell migration (20). Thus, increases in tyrosine-phosphorylated ␤-catenin are thought to contribute to the malignant progression and metastasis of tumor cells.
It has been reported that EGF or v-src phosphorylates ␤-catenin in the case of breast cancer and MDCK cells (21)(22)(23). EGF receptor (EGF-R) is also a glycoprotein, and its sugar chain represents a potential substrate for GnT-III as well as E-cadherin. Therefore, the tyrosine phosphorylation of ␤-catenin could be influenced by both aberrantly glycosylated EGF-R and E-cadherin. In this study, we report on an investigation of the effects of oligosaccharide modification by GnT-III on the tyrosine phosphorylation of ␤-catenin and cell mobility using three types of cancer cell lines. Our findings show that the pattern of ␤-catenin tyrosine phosphorylation as the result of stimulation by EGF or transfection of constitutively active c-src was dramatically changed in the case of GnT-III transfectants. Possible mechanisms for this process are discussed.

EXPERIMENTAL PROCEDURES
Materials-Dulbecco's modified Eagle's medium and RPMI were obtained from Nikken Biomedical Laboratory (Kyoto, Japan). Penicillin was obtained from Banyu Corp. (Tokyo, Japan). Kanamycin, G418, and tunicamycin were obtained from Sigma. LipofectoAMINE reagent was obtained from Life Technologies, Inc. Effectene transfection reagent was obtained from Qiagen. Biotinylated erythroagglutinating phytohemagglutinin (E4-PHA) lectin was obtained from Honen Corp (Tokyo, Japan). A monoclonal antibody against human ␤-catenin was obtained from Chemicon International Inc. (Temecula, CA). A sheep polyclonal antibody against human EGF-R was obtained from Upstate Biotechnology (Lake Placid, NY). A mouse monoclonal antibody against human GnT-lll was obtained from Fujirebio (Hachiohji, Japan). A monoclonal antibody against E-cadherin and a monoclonal antibody against phosphotyrosine (PY20) were obtained from Transduction Laboratories (Lexington, KY). A peroxidase-conjugated goat antibody against mouse lgG was obtained from Promega (Madison, WI). A peroxidase-conjugated rabbit antibody against sheep IgG was obtained from Dako (Kyoto, Japan). A detection kit for horseradish peroxidase-avidin complex (ABC kit) was obtained from Vector Laboratories (Burlingame, CA). Fluorescein isothiocyanate-labeled goat antibody against mouse lgG was obtained from Dickinson Immunocytometry Systems (San Jose, CA). EGF was obtained from Wakunaga (Hiroshima, Japan). Polyvinylidene difluoride membrane was obtained from Millipore Corp. (Bedford, MA). A Zeta-probe membrane was obtained from Bio-Rad. Protein G-Sepharose 4 FF beads and enhanced chemiluminescence system (ECL) were obtained from Amersham Pharmacia Biotech. Permafluor aqueous mounting medium was obtained from Immunon. The bicinchoninic acid (BCA) kit was obtained from Pierce.
Establishment of GnT-lll-transfected B16-hm, Huh7, and WiDr Cells-A mouse melanoma cell line, B16-hm was established from B16-F1 cells as described previously (12). A human colon cancer cell line, WiDr, was provided by the American Type Culture Collection (Manassas, VA). A human hepatoma cell line, Huh7, was provided by the Japanese Health Science Foundation (Tokyo, Japan). B16-hm cells were cultured in Dulbecco's modified Eagle's medium, and Huh7 and WiDr cells were cultured in RPMI. Each medium contained 100 g/ml kanamycin, 50 units/ml penicillin, and 10% fetal calf serum. Cell lines stably expressing GnT-lll were generated as described previously (12), and vectors that were used were constructed as follows. A human GnT-lll cDNA (24) was inserted into a mammalian expression vector, pCXN, which is regulated by the ␤-actin promoter and has a neo (G418)-resistant gene (25). Twenty micrograms of this vector was transfected into Huh7 and WiDr cells with LipofectoAMINE reagent. Selection was performed via the addition of 600 g/ml G418. One positive clone (Huh7-GIII) and one negative clone of Huh7 cells (Huh7-M) and two positive clones of WiDr cells (WiDr-GT1 and WiDr-GT2) and two negative clones of WiDr cells (WiDr-M1 and WiDr-M2) were randomly selected. The GnT-lll transfectants of B16-hm cells had previously been established by Yoshimura et al. (12).
Assay of GnT-lll and Northern Blot Analysis-GnT-lll activity of various cells was assayed with high performance liquid chromatography using a fluorescence-labeled sugar chain as a substrate (26), and the expression of GnT-III mRNA was investigated by Northern blot analysis. Total RNA was prepared from various cell lines according to the method reported previously (27). Approximately 20 g of RNAs were electrophoresed on a 1% agarose gel containing 2.2 M formaldehyde and transferred onto a Zeta-probe membrane by capillary action. Northern blot analysis of GnT-lll was performed as reported previously (28).
Western Blot Analysis-Twenty micrograms of proteins that had been extracted from mock and GnT-lll transfectants of B16-hm, Huh7, and WiDr cells were electrophoresed on a 8% polyacrylamide gel and then transferred onto a polyvinylidene difluoride membrane. After blocking with phosphate-buffered saline (PBS) containing 3% bovine serum albumin overnight at 4°C, the membrane filter was incubated with anti-human E-cadherin and anti-EGF-R and diluted with Trisbuffered saline containing 0.05% Tween 20 (TBS-T, 1:1000) for 2 h at room temperature. Each filter was washed 3 times with TBS-T for 10 min each and then incubated with TBS-T containing peroxidase-conjugated goat antibody to mouse lgG for anti-E-cadherin and rabbit antibody to sheep IgG for anti-EGF-R diluted to 1: 2500 with TBS-T for 1 h at room temperature. After the membrane was washed 3 times with TBS-T for 10 min each, it was developed by an ECL according to the manufacturer's recommended protocol.
Lectin Blot Analysis of Immunoprecipitated E-cadherin-Mock and GnT-lll transfectants of WiDr cells were plated on 10-cm dishes and incubated for 72 h. After washing with PBS, cells were lysed in a lysis buffer (10 mM Tris-HCl, pH 7.8, 0.15 M NaCl, 1 mM EDTA, 0.5% SDS, 0.5% Triton X-100, 2 mM CaCl 2 , 10% (w/v) glycerol, 1 mM phenylmethanesulfonyl fluoride, 1 M pepstatin, 10 g/ml leupeptin, and 10 g/ml aprotinin) for 30 min on ice and precleared by centrifugation at 15000 rpm for 15 min at 4°C. For immunoprecipitation, 15 l of protein G-Sepharose 4FF beads were added to the above supernatant and rotated for 1 h at 4°C with end-over-end rotations. Protein G-Sepharose 4FF beads were precleared by centrifugation at 3000 rpm for 5 min at 4°C. After incubation with 1 g of anti-E-cadherin antibody for 2 h at 4°C with end-over-end rotations, immune complexes were collected with 15 l of protein G-Sepharose 4 FF beads. After washing with a lysis buffer 5 times, the complexes were released by boiling in Laemmli sampling buffer, separated by 6% SDS-polyacrylamide gel electrophoresis, and electrotransferred onto a polyvinylidene difluoride membrane. The membrane was blocked in PBS containing 3% bovine serum albumin and incubated for 1 h with E4-PHA lectin (10 g/ml). E4-PHA lectin binds preferentially to bisecting GlcNAc residues, and the binding of E4-PHA is enhanced by the presence of bisecting GlcNAc structures (29). After washing 3 times with TBS-T for 10 min each, the membrane was incubated with TBS-T containing ABC kit for 1 h. After the membrane was washed 3 times with TBS-T for 10 min each, it was developed by ECL according to the manufacturer's protocol. After submerging in stripping buffer (62.5 mM, Tris-HCl, pH 6.8, 2% SDS, 100 mM 2-mercaptoethanol) at 50°C for 1 h, the membrane was blocked with PBS containing 3% bovine serum albumin at 4°C overnight, and then it was subjected to Western blot analysis using an anti-E-cadherin monoclonal antibody as described above.
Cell Migration and Immunofluorescence-Mock and GnT-III transfectants of WiDr cells were plated at a density of 1 ϫ 10 4 cells/cm 2 on 6-cm dishes. The mobility of mock and GnT-III transfectants of WiDr cells after EGF treatment were continuously observed with an optical microscope Diaphoto 300 (Nikon, Japan). The intracellular distribution of ␤-catenin in mock and GnT-III transfectants of WiDr cells was investigated by photofluorescent immunochemical methods. Briefly, cells were plated at a density of 1 ϫ 10 4 cells/cm 2 on 8-well chamber slides, cultured for 48 h, and incubated for 24 h in RPMI or Dulbecco's modified Eagle's medium containing 2% fetal calf serum. After treatment with EGF at the indicated time intervals, cells were washed twice with PBS, fixed with 3% paraformaldehyde in PBS for 20 min at room temperature, blocked, and permeabilized with 1% saponin and 1% bovine serum albumin in PBS for 20 min at 4°C. The cells were incubated with 1/50-diluted anti-␤-catenin antibody at room temperature for 2 h with PBS containing 1% bovine serum albumin and 1% saponin. Primary antibody binding was detected with a fluorescein isothiocyanate-labeled goat antibody to mouse lgG for 1 h at room temperature. Coverslips were mounted under a Permafluor aqueous mounting medium. Stained cells were viewed with a laser scanning confocal microscope LSM410 (Carl Zeiss, Germany) and subsequently handled using Adobe Photoshop. Tyrosine Phosphorylation of ␤-Catenin and EGF-R-Mock and GnT-III transfectants of WiDr, B16-hm, and Huh7 cells were plated at a density of 4 ϫ 10 4 cells/cm 2 on 10-cm dishes and incubated in normal conditions for 72 h. Those cells, in a subconfluent state, were further cultured for 24 h in RPMI or Dulbecco's modified Eagle's medium (Nikken, Kyoto, Japan) containing 2% fetal calf serum and were then treated with 50 ng/ml EGF. After washing with an ice-cold solution of PBS for the indicated time intervals, the cells were lysed in an ice-cold lysis buffer (5 mM EDTA, 1 mM sodium orthovanadate, 1 mM sodium pyrophosphate, 100 mM sodium fluoride, 0.5% Triton X-100, 1 mM phenylmethanesulfonyl fluoride, 1 M pepstatin, 10 g/ml aprotinin, and 10 g/ml leupeptin) for 30 min on ice and precleared to remove pellets (insoluble fraction) by centrifugation at 15000 rpm for 15 min at 4°C. Immunoprecipitation was performed using anti-␤-catenin or anti-EGF-R antibody. Bound proteins were subjected to 6% SDS-polyacrylamide gel electrophoresis and directed by immunoblotting with anti-PY20. After reprobing and blocking the membrane, it was subjected to Western blot analysis using anti-␤-catenin or EGF-R as described above.
Tunicamycin Treatment-To further investigate the importance of oligosaccharides on ␤-catenin phosphorylation, cells were treated with 0.5 g/ml tunicamycin (Sigma) for 24 h before EGF stimulation.
Tyrosine Phosphorylation of ␤-Catenin by the Transfection of a Constitutively Active Form of c-src-The src expression vector, c-srcY527F, which is regulated by SV40 promoter and constitutively activated by mutation from 527 tyrosine to phenylalanine, was kindly provided by Dr. Fukuda (Ikeda City Hospital, Osaka, Japan) (30). This vector was transfected into mock and GnT-III transfectants of WiDr cells with Effectene according to the standard protocol. After 48 h, the tyrosine phosphorylation of ␤-catenin in these cells was investigated as described above.

Establishment of GnT-lll-transfected WiDr and Huh7 Cells-
Although the GnT-lll activity of WiDr and Huh7 cells was quite low, WiDr-GT1, -GT2, and Huh7-GIII showed Ͼ100 times higher levels of GnT-lll activity than their mock transfectants (Fig. 2). Northern blot analysis showed a high level of expression of GnT-lll mRNA in GnT-III-transfected WiDr and Huh7 cells, which is consistent with the observed enzyme activity ( Fig. 2A) and suggests that the high levels of GnT-lll activity in WiDr-GT1, -GT2, and Huh7-Glll were due to the high transcriptional levels. Western blot analysis showed a high level of GnT-lll protein in GnT-lll transfectants (Fig. 2B). Similar results of GnT-lll expression were observed in GnT-lll-transfected B16-hm cells (12).
Western Blot and Lectin Blot Analyses of E-cadherin and EGF-R-The expression of E-cadherin and EGF-R was investigated by Western blot analysis. Whereas the expression of E-cadherin in B16-hm cells was increased in the case of GnT-lll transfectants compared with mock transfectants (3), the level in the GnT-lll transfectants of WiDr and Huh7 cells was nearly the same compared with each mock transfectant. However, the molecular size of E-cadherin in GnT-lll-transfected WiDr and Huh7 cells was smaller than that in the mock transfectants, suggesting that an extension of N-oligosaccharide in E-cadherin had been suppressed by the action of GnT-lll (Fig. 3A,  upper panel). The level of EGF-R expression was very low in both mock and GnT-lll transfectants of B16-hm cells. In contrast, a high level of expression of EGF-R was observed in WiDr and Huh7 cells, and this level was nearly equal in both the mock and the GnT-lll transfectants. The molecular size of EGF-R in GnT-lll-transfected WiDr and Huh7 cells was smaller than that of the mock transfectants as well as in the case of E-cadherin (Fig. 3A, lower panel).
To further investigate the source of the differences in the molecular size of E-cadherin in mock and GnT-lll transfectants, lectin blot analysis on immunoprecipitated E-cadherin was performed using biotinylated E4-PHA. E-cadherin, which had been immunoprecipitated from WiDr-GT1, -GT2, and Huh7-Glll cells exhibited an increase in E4-PHA binding compared with mock transfectants (Fig. 3B, a). An increase in E4-PHA binding to EGF-R immunoprecipitated from GnT-lll transfectants was also observed (Fig. 3B, b). These results indicate that GnT-lll transfection led to an increase in the number of bisecting GlcNAc residues on the N-glycans of E-cadherin and EGF-R in WiDr and Huh7 cells.
Cell Morphology and Distribution of ␤-Catenin after EGF Treatment-When the cells were treated with EGF, the degree of cell-cell contacts became less strong in the mock cells. A typical pattern of changes in cell morphology was observed in WiDr cells (Fig. 4A). After 60 min, the morphology of the mock cells returned to the original shape. This change, however, was negligible in the case of GnT-lll transfectants. It is interesting to note that the intracellular localization of ␤-catenin was altered along with the morphological alteration of mock transfectants (Fig. 4B). The localization of ␤-catenin was changed from the cell membrane to the cytoplasm after EGF treatment. In the GnT-lll transfectants, most of the ␤-catenin continued to remain distributed along the entire cell membrane after EGF treatment.
Tyrosine Phosphorylation of ␤-Catenin after EGF Treatment-Since EGF treatment is a well known method for stimulating the tyrosine phosphorylation of ␤-catenin, GnT-llltransfected WiDr, B16-hm, and Huh7 cells were treated with 50 ng/ml EGF. As shown in Fig. 5, a dramatic difference in tyrosine phosphorylation of ␤-catenin was observed between GnT-lll and mock transfectants. When EGF was added to mock-transfected WiDr cells, the tyrosine phosphorylation of ␤-catenin increased within 15 min of EGF stimulation and persisted for at least 60 min. In contrast, the levels of tyrosine phosphorylation of ␤-catenin were very low and increased only slightly at 15 min after EGF stimulation and then decreased in the GnT-III transfectants (Fig. 5A, upper panel). In the case of B16-hm cells, the basal level of tyrosine phosphorylation of ␤-catenin before EGF treatment was somewhat higher than that in the WiDr cells. The tyrosine phosphorylation of ␤-catenin was slightly increased in mock cells after EGF treatment but was dramatically decreased in GnT-lll transfectants (Fig.  5B, upper panel). In the case of Huh7 cells, the tyrosine phosphorylation of ␤-catenin was much weaker than the B16-hm and WiDr cells. However, the difference of ␤-catenin phospho- After 48 h, they were cultured for an additional 24 h in a RPMI buffer containing 2% fetal calf serum. After EGF (50 ng/ml) was added, cells were collected and analyzed at the indicated times after EGF treatment. Immunoprecipitated (IP) ␤-catenin was separated by SDS-polyacrylamide gel electrophoresis. Tyrosine phosphorylation levels were analyzed by Western blotting using an anti-phosphotyrosine antibody, PY20 (upper panels of A, B, and C). The membrane was reprobed with a specific antibody to ␤-catenin to verify that equal amounts of precipitated proteins (lower panels of A, B, and C) were obtained. rylation between mock and GnT-lll transfectants of Huh7 cells was also observed. The degree of ␤-catenin phosphorylation at 30 min in GnT-lll transfectants was ϳ70% that of mock transfectant levels (the mean of 3 experiments; Fig. 5C, upper panel). Collectively, the tyrosine phosphorylation of ␤-catenin increased in mock transfectants of WiDr, B16-hm, and Huh7 cells, whereas it was suppressed or decreased in GnT-lll transfectants after EGF treatment.
Tyrosine Phosphorylation of EGF-R-To verify whether or not GnT-lll transfection led to an alteration in EGF-R function, tyrosine phosphorylation levels of EGF-R after EGF treatment were investigated. As shown in Fig. 6, the levels of EGF-R phosphorylation were dramatically enhanced at 5 min after EGF treatment in both mock and GnT-lll transfectants of WiDr cells. The time course pattern of EGF-R phosphorylation in GnT-lll transfectants was quite similar to that of mock transfectants, suggesting that a similar signaling via EGF-R was transduced.
The Treatment with Tunicamycin-Tunicamycin inhibits the N-glycosylation of glycoproteins in higher organisms by blocking the first step in biosynthesis of the lipid-linked oligosaccharide precursor. To determine the extent of involvement of oligosaccharides in ␤-catenin phosphorylation, mock and GnT-lll transfectants were treated with tunicamycin before the addition of EGF. As shown in Fig. 7, the pattern of tyrosine phosphorylation of ␤-catenin for mock and GnT-lll transfectants became nearly identical after tunicamycin treatment. This result suggests a direct involvement of the bisecting GlcNAc residue in terms of variations in the phosphorylation of ␤-catenin for mock vis à vis GnT-lll transfectants. Thus, the inhibition of N-glycosylation in glycoproteins led to an alteration of the tyrosine phosphorylation in ␤-catenin after EGF treatment.
The Effect of src Transfection on Tyrosine Phosphorylation of ␤-Catenin-To investigate the effect of tyrosine phosphorylation of ␤-catenin by other well known alternative pathways, a constitutively active form of src was transfected into GnT-lll and mock-transfected WiDr cells. In mock cells, the levels of tyrosine phosphorylation of ␤-catenin by src transfection was elevated by more than five times the levels detected before transfection but was blocked by treatment with tunicamycin. In contrast, in the case of GnT-III transfectants, this elevation of ␤-catenin phosphorylation by this active form of src did not occur (Fig. 8, A and B). DISCUSSION In many glycoproteins, particularly adhesion molecules and growth factor receptors on cell surfaces, N-linked oligosaccharides contribute to the folding, stability, and biological functions of the molecules (31)(32)(33). Our results indicate that the overexpression of GnT-III could lead to the modification of the tyrosine phosphorylation pattern of ␤-catenin after EGF stimulation. Previous studies from our laboratory have demonstrated that the amount of E-cadherin expression was increased in GnT-III-transfected mouse melanoma B16-hm cells, due to delayed turn-over, compared with mock transfectants (3). Increases in E-cadherin levels on the cell surface enhanced cell-cell contacts and aggregations.
However, in human cancer cells such as WiDr and Huh7, the expression of E-cadherin in GnT-III transfectants was equivalent to that of mock transfectants, although the oligosaccharide structures of E-cadherin were altered. Although morphological alteration such as compact organization was observed in GnT-III-transfected B16-hm cells, it was not observed in the cases of WiDr and Huh7 cells. These results suggest that proteases, which contribute to the cleavage of E-cadherin, are different between humans and mice. In contrast, both E4-PHA binding and the molecular size of EGF-R in these transfectants were changed, but the tyrosine phosphorylation of EGF-R on GnT-III transfectants was observed to occur normally both in WiDr and Huh7 cells. Therefore, changes in the tyrosine phosphorylation of ␤-catenin are not due to EGF-R dysfunction but, rather, its induction by other factors such as downstream signals of EGF-R.
We next examined changes in the integrity of cell-cell contacts between mock and GnT-III-transfected WiDr cells that had been treated with EGF. The GnT-III transfectants of WiDr cells barely changed in terms of cell morphology and cell-cell contact after EGF stimulation. Interestingly, changes in the cell-cell contacts of mock cells were consistent with the tyrosine phosphorylation pattern of ␤-catenin. To further explore the relationship between cell-cell adhesion and tyrosine phosphorylation of ␤-catenin, the localization of ␤-catenin in mock and GnT-III transfectants of WiDr cells was investigated via an immunofluorescence method. The localization of ␤-catenin in mock transfectants was redistributed from the lineage of cell membrane to cytoplasm after EGF treatment. In contrast, the localization of ␤-catenin in GnT-III transfectants remained along the cell-cell contact region of adjacent-adhering cells after EGF treatment. Thus, the nature of cell-cell contacts for the GnT-III-transfected WiDr cells remained unchanged after EGF stimulation.
An alteration in cadherin-mediated cell-cell adhesion is frequently associated with the tyrosine phosphorylation of ␤-catenin. This phosphorylation is thought to be caused by the Src family of tyrosine kinases and receptors with tyrosine kinase activity such as growth factors like EGF, hepatocyte growth factor, platelet-derived growth factor, and colony stimulating factor-1. In this study, EGF stimulation and src transfection were used to phosphorylate ␤-catenin. The tyrosine phosphorylation of ␤-catenin after EGF treatment was suppressed in GnT-III transfectants of three types of cancer cells. Changes in the phosphorylation pattern were a little different in B16-hm cells. A basal level of ␤-catenin phosphorylation was higher in B16-hm cells than WiDr cell and Huh7 cells. It might be due to metastatic potential of melanoma cells in terms of migration. In the case of Huh7 cells, tyrosine phosphorylation of ␤-catenin after EGF treatment was quite low and suppressed in GnT-III transfectants at 70% levels. This might be due to low potentials of cell migration in Huh7 cells. Since the modification of EGF-R oligosaccharide by GnT-III does not involve changes in the tyrosine phosphorylation of ␤-catenin, another possibility for the suppression of tyrosine phosphorylation of ␤-catenin involves complex formation between ␤-catenin and aberrantly glycosylated E-cadherin. Conformational changes of this complex might inhibit the action of kinases for ␤-catenin. To further investigate this, we examined the effect of transfection of activated src into GnT-III transfectants. As expected, tyrosine phosphorylation of ␤-catenin by the src transfectant was dramatically suppressed in GnT-III transfectants as well as in the case of EGF stimulation (Fig. 8). In addition, to ensure the involvement of changes in the N-linked oligosaccharides structure, cells were treated with tunicamycin, which inhibits the N-glycosylation of glycoproteins by blocking the first step in the biosynthesis of a lipid-linked oligosaccharide precursor. The tyrosine phosphorylation of ␤-catenin in tunicamycin-treated mock transfectants was nearly identical to that of GnT-III transfectants, suggesting that changes in N-glycosylation directly affect the tyrosine phosphorylation of ␤-catenin. Levels of E-cadherin were increased in GnT-III transfectants of B16-hm cells but remained unchanged in GnT-III transfectants of WiDr and Huh7 cells. Nevertheless, the tyrosine phosphorylation of ␤-catenin was suppressed in all GnT-III transfectants, suggesting that aberrant glycosylated E-cadherin is not involved in cell-cell contacts. In fact, it has been reported that E-cadherin oligosaccharides are not important for their function as an adhesion molecule (34). However, the present study indicates that bisecting GlcNAc residues on E-cadherin affect intracellular signaling, such as ␤-catenin phosphorylation, which might be due to an alteration in the E-cadherin-catenin complex. The up-regulation of src or EGF signaling occurs during the progression of cancer (35). If the action of src or EGF to phosphorylate ␤-catenin is inhibited by GnT-III in vivo, GnT-III contributes to the progression of cancer in multi-steps. In summary, the addition of bisecting GlcNAc residues of E-cadherin-␤-catenin by GnT-III down-regulates the tyrosine phosphorylation of ␤-catenin, which might lead to the suppression of tumor progression such as invasion or metastasis.