Overexpression of N-acetylglucosaminyltransferase III disrupts the tyrosine phosphorylation of Trk with resultant signaling dysfunction in PC12 cells treated with nerve growth factor.

beta-1,4-N-Acetylglucosaminyltransferase III (GnT-III: EC 2.4.1.144) is a pivotal glycosyltransferase which participates in branch formation by catalysis of the synthesis of a bisecting GlcNAc structure in N-glycans. These structures are thought to be one of the unique features of the N-glycans of neural tissues. To examine the intracellullar role of GnT-III expression and its product in neural cells, its gene was overexpressed in rat pheochromocytoma PC12 cells which normally express a low level of GnT-III. In the GnT-III gene-transfected cells, lectin blot analysis showed that some glycoproteins showed increased levels of bisecting GlcNAc structures. Following treatment with nerve growth factor (NGF) the control cells showed neurite outgrowth for differentiation whereas the transfectants showed no morphological response or change in the rate of cell growth. Transient tyrosine phosphorylation of the Trk/NGF receptor was detected at 5-15 min after NGF treatment in control cells, but not detected in the GnT-III gene-transfected cells despite the intact binding of NGF to the cells. Moreover the dimerization of Trk with NGF treatment was not induced in the GnT-III transfectant as compared with the dimerization seen in control cells. These results indicate that overexpression of GnT-III gene in PC12 cells affects some functions of glycoprotein receptors such as Trk by alteration of N-glycan structures, and results in changes in the intracellular signaling pathway of tyrosine phosphorylation modified by NGF.

The marked changes in the sugar chain structures of cell surface membrane occurring during ontogenesis and oncogenesis suggest that they play pivotal roles in cell differentiation and proliferation (1). In the case of glycoproteins, N-and Oglycans are expressed in the majority of cell surface and secreted proteins (2). A gene for ␤-1,2-N-acetylglucosaminyltransferase I (GnT-I), 1 the enzyme catalyzing the formation of complex type N-glycans has been obliterated in mice. The re-sultant pathology showed that complex type N-glycans are required for normal embryonic development, especially of neural tissues (3,4).
In such tissues, it is known that the level of polysialyl Nglycans decreases in the neural cell adhesion molecule as these cells mature. This indicates that the sugar chains of the neural cell adhesion molecule regulate the cell-cell interaction in neural tissues (5)(6)(7)(8). Although the polysialyl sugar chain structure is one characteristic of most neural tissues, some unique differences for the N-glycans of mouse brain have been reported (9). Among them, a bisecting GlcNAc structure, which is biosynthesized by UDP-N-acetylglucosamine:␤-D-mannoside ␤-1,4-N-acetylglucosaminyltransferase III (GnT-III: EC 2.4.1.144), has been found ( Fig. 1).
GnT-III is one of the pivotal glycosyltransferases which participates in the branching of N-glycans (10), and produces an unique sugar chain structure, a bisecting GlcNAc (11). GnT-III has been purified from rat kidney and the rat (12), human (13), and mouse (14) genes have been cloned. The tissue distribution of its mRNA showed that the GnT-III transcript was particularly high in the brain and kidney of the mouse (14). Such high levels for the expression of GnT-III mRNA seem to be compatible with the existence of unique N-glycan structures in brain including bisecting GlcNAc. Several experimental approaches have been used to elucidate the role of GnT-III in cultured cells (15,16). In mouse melanoma B16 cells, GnT-III gene induction resulted in changes in their N-glycan structures and suppressed the metastatic potential of the original cells (17). This implies that GnT-III also has important functions in neural tissues since melanoma cells are also of neural origin. In the present study, we have investigated the biological roles of GnT-III expression and its bisecting GlcNAc product in rat pheochromocytoma PC12 cells by introduction of the GnT-III gene.

MATERIALS AND METHODS
Cell Culture-Rat pheochromocytoma PC12 cells were obtained from the Japanese Cancer Research Resources Bank (Tokyo). PC12 and the GnT-III gene-transfected cells were cultured in Dulbecco's modified Eagle's medium (DMEM), supplemented with 10% fetal calf serum, 5% horse serum, and 0.1 mg/ml of kanamycin under a humidified atmosphere of 95% air and 5% CO 2 . NGF (TaKaRa, Japan) was added to a final concentration of 50 ng/ml to induce PC12 differentiation.
Recombinant DNA Constructs-Rat GnT-III cDNA clone C4 (12) was deleted at its 5Ј non-coding region and the cDNA fragment containing the entire coding sequence was inserted into the pMEP4 (Invitrogen) EcoRI site to obtain the final construct, pMEPrIII. The pMEP4 is a mammalian expression vector with the metallothionein IIa gene enhancer/promoter (18) also includes the hygromycin-resistant gene.
Gene Transfection and Selection of Cells-PC12 cells used for transfection of cDNA were plated in a 10-cm plastic culture dish coated with collagen to a density of 1 ϫ 10 6 /ml cells. After 24 h the medium was removed and the cells washed twice with cold phosphate-buffered saline (PBS), pH 7.2, and changed to serum-free DMEM. The pMEPrIII vector (20 g) was mixed with Lipofectamine (Life Technologies, Inc.) and 100 l of this solution was added to PC12 cells. After 5 h incubation the medium was changed to the original as described above. Stable transfectants were screened with 0.5 mg/ml hygromycin.
Lectin Blot Analysis-Samples of wild type and gene transfectant cell extracts containing 5 g of protein were electrophoresed on 8 -12% SDS-polyacrylamide gels under reducing conditions and then transferred to nitrocellulose membranes (Schleicher & Schuell) as described previously (21). The membrane was blocked with 3% bovine serum albumin in PBS and then incubated with biotin-conjugated lectins (2 g/ml biotinylated E-PHA or 2 g/ml biotinylated L-PHA; Honen Corp. Tokyo) using the buffer systems as described (21). Lectin reactive proteins were detected using a Vectastain ABC kit (Vector Laboratories) and the blots were developed using the ECL chemiluminescence detection kit (Amersham) according to the manufacturer's instructions.

Characterization of Trk in Cultured Cells Treated with NGF-The
GnT-III gene-transfected PC12 cells and their controls were cultured with or without 50 ng/ml NGF and then disrupted in the lysis buffer (20 mM Tris, pH 7.2, 1% Nonidet P-40, 10% glycerol, 1 mM APMSF, 5 mM aprotinin, 0.4 mM sodium orthovanadate, 10 mM sodium pyrophosphate, 10 mM sodium fluoride, 10 mM iodoacetamide). The cell-free lysates were adjusted to the same protein concentration and immunoprecipitated with the C-14 Trk antibody (Santa Cruz Biotechnology Inc.) according to the manufacturer's instructions. The precipitate was subjected to 8% SDS-PAGE, and electroblotted samples were then characterized by immunoblot or lectin blot analyses as described above. For the detection of phosphorylated tyrosine residues, a monoclonal phosphotyrosine antibody (PY20, Transduction Laboratories) was used according to the manufacturer's protocol after blocking with 5% skim milk in PBS. The blots were detected by peroxidase-conjugated rabbit anti-mouse IgG (Cappel) and developed with the ECL kit.
Phosphorylation of Trk with or without NGF was estimated by immunoprecipitation of 32 P-labeled Trk as described previously (22). In brief, cells were preincubated in phosphate-free DMEM for 1 h. [ 32 P]Orthophosphate (0.1 mCi/ml) was then added and incubated with the cells for 1 h. After 5 min of NGF (50 ng/ml) treatment, cells were lysed and Trk was immunoprecipitated as described above. The precipitates were subjected to 8% SDS-PAGE followed by autoradiography using a Fuji Film imaging plate for the BAS2000 Bioimage analyzer (Fuji Photo Film, Japan).
DNA Synthesis-Plates of 96 wells, each containing 2 ϫ 10 4 cells in 200 l, were cultured with or without 50 ng/ml NGF for 12 h and [ 3 H]thymidine (1 Ci/well) was added to each well. After 6 h of incubation at 37°C the cells were harvested and [ 3 H]thymidine incorporated into DNA was measured with the Betaplates system (Pharmacia Biotech Inc.).
NGF Binding-Cells were plated in 24-well culture plates at a density of 2 ϫ 10 4 cells/ml and cultured for 12 h. Cells were washed with ice-cold binding buffer (PBS(Ϫ) with 0.1 mM CaCl 2 and 1 mM MgCl 2 ) and incubated with 500 l of the binding buffer containing various concentrations (0.37-10 ng/ml) of 125 I-labeled NGF at 4°C for 2 h. Cells were washed four times with ice-cold PBS(Ϫ) and lysed with 1 M NaOH, and the radioactivity was counted in a ␥-counter. To estimate the specific binding of NGF to the cells, NGF binding assay using 125 Ilabeled NGF was also performed in the presence of various amounts of cold NGF (0.02-500 ng).
Chemical Cross-linking of Trk-Cells (2 ϫ 10 6 ) were harvested, pelleted, and resuspended in the binding buffer as described under the method for NGF binding. NGF was added to a final concentration of 100 ng/ml and the cells were incubated at 4°C for 1 h. The chemical cross-linker 3,3Ј-dithiobis(sulfosuccinimidylpropionate) (Pierce) was added to a final concentration 0.5 mg/ml. The reaction was incubated at room temperature for 30 min and quenched by washing with Trisbuffered saline (pH 7.4). Cross-linked cells were lysed and subjected to the immunoblot analysis of Trk as described above.

Establishment of GnT-III Gene Transfectants Expressing
High Levels of GnT-III Activity-To investigate the functional role of GnT-III gene transfection and its bisecting GlcNAc sugar chain product, during neural cell differentiation and development, a GnT-III gene expression vector, MEPrIII, was employed. This inserted rat GnT-III cDNA downstream of the metallothionein IIa promoter, which was then transfected into PC12 cells using a LipofectAMINE complex. The hygromycin resistant transfectants were screened as described under "Materials and Methods." The specific activity of GnT-III for each clone was assayed and then three high activity transfectants (PC12-III-1, -2, and -3) were randomly taken and used in succeeding experiments. As shown in Table I, GnT-III activity was elevated about 4 -6 times over the wild type or mock transfectant (PC12-hyg). GnT-V and ␤-1,4-galactosyltransferase activities assayed as the control glycosyltransferases showed no significant changes.
Effect of Overexpression of GnT-III Activity in Cell Differentiation by NGF in PC12 Cells-PC12 cell differentiates and forms neurites under NGF treatments (24,25). To test the role of overexpressed GnT-III and its bisecting GlcNAc sugar chain product on NGF induced differentiation, parental, mock transfectants, and GnT-III gene transfected PC12 cells were cultured with or without 50 ng/ml NGF for 5 days and cell morphologies were compared. Fig. 2 shows the morphology of cells cultured with NGF. After 5 days of treatment, cell proliferation was reduced and neurite formation was observed in the case of PC12 and the PC12-hyg, a mock transfectant. However, the GnT-III gene transfectants showed no neurite formation, and cell proliferation was not affected (PC12-III-1 and -2 representing the transfectants were shown).
Effect of GnT-III Gene Expression on Cell Proliferation under NGF Treatment-To test the effect of GnT-III expression on proliferation of PC12 cells, their growth rate was evaluated by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay. As shown in Fig. 3A, part a, the cell growth was suppressed in the case of control cells under NGF treatment, compared with the cells without NGF treatment. But no suppression of cell growth was observed in the case of GnT-III gene-transfected cells with NGF treatment (Fig. 3A, part b). DNA synthesis was also evaluated by measuring the incorporation of [ 3 H]thymidine into DNA. As shown in Fig. 3B, the thymidine incorporation was decreased in the case of control cells under NGF treatment compared with the cells without NGF. But the decrease was not observed in the GnT-III genetransfected cells with NGF treatment.
Lectin Blots of Glycoproteins of Control and GnT-III Genetransfected PC12 Cells-To investigate the change of sugar component of glycoproteins from cell lysates between control and GnT-III gene-transfected cells, lectin blot analyses were performed using E-and L-PHA. E-PHA preferentially binds to sugar chains containing a bisecting GlcNAc structure which is the product of GnT-III (26). In contrast L-PHA binds to structures of complex type N-glycans containing 3 or 4 branches including GlcNAc␤-1,6, which is the product of GnT-V (27). In the case of control cells, after 5 days of treatment with NGF, E-PHA reactivity showed no significant change but L-PHA reactivity was slightly increased in the glycoproteins of 80 -95 kDa (Fig. 4, lane 1 of L-PHA). In the case of GnT-III genetransfected cells, E-PHA reactivity to glycoproteins of about 98 kDa showed a marked increase compared with control cells without NGF treatment (Fig. 4, lanes 2 and 3 of E-PHA, arrow).

FIG. 3. Cell growth and thymidine incorporation of control and GnT-III gene-transfected PC12 cells.
A, cells (2 ϫ 10 4 /ml) were seeded and cultured in DMEM supplemented with 10% fetal calf serum and 5% horse serum. Cell growth of control, mock, and GnT-III genetransfected PC12 cells with or without NGF was evaluated at indicated days using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay as described under "Materials and Methods." a, control and mock gene-transfected cells; b, GnT-III gene transfectants. Each value represents the mean of three independent experiments, and the standard deviation is always within 10% of the mean. B, cells (2 ϫ 10 4 ) were seeded into 96-well culture plates and cultured with or without NGF for 12 h. After an incubation with [ 3 H]thymidine (1 Ci/well) for 6 h, [ 3 H]thymidine incorporation into DNA was evaluated as described under "Materials and Methods." Each value represents the mean Ϯ S.D. of three independent experiments.

FIG. 2. Morphological change of control and GnT-III genetransfected PC12 cells on NGF treatment.
Cells (2 ϫ 10 4 /ml) were seeded and cultured in DMEM supplemented with 10% fetal calf serum and 5% horse serum with or without 50 ng/ml NGF for 5 days. The results were reproducible in four independent experiments. However, after 5 days of treatment with NGF, the L-PHA reactivity to glycoproteins showed no significant change in contrast to the case for control cells (Fig. 4, lanes 2 and 3 of L-PHA). These data showed that some glycoproteins were modified by overexpression of GnT-III activity and E-PHA reactive glycoproteins increased. These E-PHA reactive glycoproteins produced by overexpression of GnT-III activity may suppress the appearance of L-PHA reactive glycoproteins noted in the differentiated PC12 cells after 5 days treatment with NGF. During this period, no significant change was found in the enzyme activity levels of glycosyltransferases GnT-III, -V, and ␤-1,4-galactosyltransferase as the result of treatment with NGF (data not shown).
Tyrosine Phosphorylation of NGF Receptor/Trk-To determine the effect of NGF signaling in GnT-III gene-transfected cells, the time course of tyrosine phosphorylation of Trk was investigated. In Fig. 5A, control and GnT-III gene-transfected PC12 cells that had been cultured with NGF were harvested at the indicated time (0, 5, 15, and 30 min) after treatment and lysed. From each cell lysate sample, Trk was immunoprecipitated with anti-Trk antibody and subjected to blotting analyses with anti-phosphotyrosine antibody and anti-Trk antibody. In Fig. 5A, tyrosine phosphorylation of Trk was detected after 5-15 min of NGF treatment in control cells. However, no significant tyrosine phosphorylation was detected in the case of Trk from GnT-III gene-transfected cells. In Fig. 5B, phosphorylation of Trk with NGF treatment was also estimated using [ 32 P]orthophosphate-labeled cells. Despite an apparent signal of 32 P-labeled Trk in control cells, no signal was detected in the GnT-III gene-transfected cells under NGF treatment. Similar results were obtained in other GnT-III gene transfectants (data not shown). These results suggest that the function of Trk may be abolished by changes in its sugar chain structure and then lead to disruption of tyrosine phosphorylation. Such an effect is consistent with failure to respond to NGF as found in the GnT-III gene-transfected PC12 cells (Fig. 2).
Lectin Blots of Trk from Control and GnT-III Gene-transfected PC12 Cells-Trk was immunoprecipitated and subjected to lectin blot analyses. In Fig. 6

NGF Binding to Control and GnT-III Gene-transfected PC12
Cells-Using 125 I-labeled NGF, NGF binding to control and GnT-III gene-transfected cells was investigated as described under "Materials and Methods." As shown in Fig. 7, A and B, there is no significant difference in the binding between control and the gene-transfected cells.
Effect of GnT-III Gene Expression on Dimerization of Trk under NGF Treatment-To clarify the precise mechanism of disruption of Trk function in the GnT-III gene transfectants under NGF treatment, Trk dimerization by NGF was investigated. The cell surface proteins were cross-linked with 3,3Јdithiobis(sulfosuccinimidylpropionate) under NGF treatment as described under "Materials and Methods." After cross-linking, cell lysates were subjected to 8% SDS-PAGE followed by immunoblotting using anti-Trk antibody. In Fig. 8, control cells showed a typical homodimer form of Trk (about 300 kDa) after 5 min treatment of NGF. But the dimerization of Trk was not observed in the gene-transfected cells.
In conclusion, these results indicate that the change of sugar chain structure of Trk by GnT-III causes the functional change in Trk, and disrupts the dimerization of Trk leading to nonautophosphorylation in the gene transfectants under NGF treatment, despite the intact binding of NGF to the cells.

DISCUSSION
Certain N-glycan structures of glycoproteins appear to play an important role in neural cell development (3,4). GnT-III is one of the pivotal enzymes which regulates N-glycan branch structure, and the enzyme activity is high in such mammalian tissues as brain and kidney (14,28). In the present study we have investigated some of the biological functions of GnT-III expression and its bisecting GlcNAc product in rat pheochromocytoma PC12 cells. These cells have very low GnT-III activities despite their neural origin. A GnT-III expression vector, whose expression was controlled by the metallothionein IIa promoter, was transfected into PC12 cells using a lipid complex, and the positive clones were screened by hygromycin resistance. In the case of GnT-III gene transfectants, the sugar chain structure of N-glycans was modified and especially some specific 98-kDa proteins showed increases in their level of bisecting GlcNAc as resulted by their reactivity with E-PHA. At present, it is not clear why some specific glycoproteins are susceptible to GnT-III overexpression. Conformation of N-glycosylation sites in each glycoprotein may affect the susceptibility to GnT-III action, and determine the sugar chain structure of each glycoprotein in the cells. In the case of GnT-III genetransfected mouse melanoma B16 cells, we also experienced that specific 80-and 95-kDa proteins were enhanced in E-PHA reactivity (17). But in both of PC12 and B16 transfectants, these specific target glycoproteins could not be identified, and further investigations will be required.
The PC12 cells were found to be a most useful and popular model for the study of the actions of NGF and cell differentiation (24,29). We have investigated the biological effect of GnT-III gene transfection into PC12 cells and found that they were not responsive to NGF as indicated by the rate of cell growth and morphological changes. These results indicated that such regulation and differentiation of PC12 cells were markedly modified by modulation of the sugar chain structure of several glycoproteins regulating their cellular differentiation.
In the mechanism of NGF action, it appears to relate to the phosphorylation of a number of cellular proteins (30 -32). The NGF signaling pathway is known to be initiated by the direct binding of NGF to the high affinity NGF receptor/Trk protooncogene. This receptor is a protein tyrosine kinase and its activity and autophosphorylation are activated in response to NGF (33,34). In attempts to clarify the molecular mechanism of the non-responsiveness to NGF of GnT-III gene-transfected PC12 cells, the role of Trk was investigated. The Trk receptor has 13 potential N-glycosylation sites in its molecule (35), and any of these potential sugar chain attachments might be modified. To test the function of Trk, after treatment of the cells with NGF, Trk was immunoprecipitated and then the tyrosine phosphorylation level was estimated by immunoblot analysis using a phosphotyrosine antibody. Surprisingly, in the case of GnT-III gene transfectant, the tyrosine phosphorylation level of Trk was not increased compared with the increase seen in the control PC12 cells after adding NGF (Fig. 5). The tyrosine phosphorylation of Trk is known to be accompanied with dimerization of Trk under NGF treatment (36). In the case of GnT-III gene transfectant, this dimerization did not occur even under NGF treatment as compared with the control, despite the intact binding of NGF to the cells. These data suggest that the change of sugar chain structure of Trk by overexpressed GnT-III activity affects its conformation of protein and causes disturbance of the dimerization, and disrupts its signal transduction under NGF treatment.
Laconte et al. (37) have previously shown that the N-glycans of the insulin ␤ subunit receptor were essential for transmembrane signaling in Chinese hamster ovary cells by studies in which this receptor had been modified by site-directed mutagenesis of its N-glycosylation sites. Our present study also supports the importance of N-glycans in the cellular signaling systems, and provides further mechanical evidence of N-glycan function in the cellular signaling.
It has been recently reported that the G M1 ganglioside can bind to Trk and activate its autophosphorylation (38) and prevent apoptosis of PC12 cells (39). Mutoh et al. (38) also reported that the binding of G M1 to Trk was inhibited by blocking of N-glycosylation using tunicamycin, and this suggests the importance of N-glycosylation on Trk in its biological function. The relationship of sugar structures of glycolipid G M1 and glycoprotein in the mechanism of the NGF signaling pathway is not clear, but appears to relate to the existence of a novel regulation of Trk activation. In the case of neuroblastoma GOTO cells, the GnT-III activity level is high in the subconfluent state, but drastically decreases in the confluent state of the cells. This suggests that the regulation of GnT-III activity relates to the regulational system of cell proliferation (21). Similar phenomena were also observed in the case of sialyltransferase activity during cell growth in HepG2 cells (40). If cell proliferation is partially controlled by modulation of intracellular growth signaling due to autonomous regulation of sugar chain biosynthesis, cell growth associated changes of glycosyltransferase levels could relate to a feed-back like regulatory function. Although the present study has been concerned in part with the NGF-Trk signaling pathway in PC12 cells, investigations of other pathways offer attractive areas of investigation.
Very recently Canossa et al. (41) have reported that the low affinity NGF receptor (p75 NGFR ) could accelerate Trk-mediated signaling, and activate some p75 NGFR -associated protein kinases. To assess the glycosylation change of p75 NGFR in the GnT-III gene-transfected cells, p75 NGFR was immunoprecipitated from control and GnT-III gene-transfected cells using a rabbit anti-mouse p75 NGFR polyclonal antibody (Chemicon Int. Inc.) which can also react to rat p75 NGFR , and subjected to the lectin blot analysis using E-and L-PHA as described in the case of Trk. Even in the GnT-III gene-transfected cells, no increase of E-PHA reactivity was observed (data not shown), and suggested that p75 NGFR was not affected by GnT-III overexpression. p75 NGFR has one N-glycan in its extracellular domain and the removal of the N-glycosylation site does not affect the NGF binding to the non-glycosylated p75 NGFR (42). This implies that N-glycosylation in p75 NGFR may be not important for its function. Taken together, in the NGF-associated signaling of PC12 cells, we conclude that GnT-III overexpression mainly affects Trk itself but not another known pathway of p75 NGFR . However, to clarify the precise functional correlation between Trk and N-glycosylation, further investigations of Trk function with its modified N-glycosylation should be required using various cell models expressing Trk.
Our studies demonstrate that a specific N-glycan structural change affects some receptor glycoprotein functions, and causes modulation of a cell biological function such as cell differentiation. Until now it has been believed that a sugar chain mainly functions outside of the cell but we have found that the modulation of sugar chain structures can lead to the modulation of various biological functions by affecting intracellular events such as signal transduction pathways.