Gene Transfection-mediated Overexpression of (cid:98) 1,4- N Acetylglucosamine Bisecting Oligosaccharides in Glioma Cell Line U373 MG Inhibits Epidermal Growth Factor Receptor Function*

N -linked oligosaccharides appear to be important for the function of the epidermal growth factor (EGF) receptor. In a previous study (Rebbaa, A., Yamamoto, H., Moskal, J. R., and Bremer, E. G. (1996) J. Neurochem. 67, 2265–2272), we showed that binding of the erythroagglu- tinating phytohemagglutin lectin from Phaseolus vulgaris to the bisecting structures on the EGF receptor from U373 MG glioma cells blocked EGF binding and receptor autophosphorylation. In this study we exam- ined the consequences of overexpression of the bisecting structure on the EGF receptor by gene transfection of U373 MG cells with the N -acetylglucosaminyltrans-ferase III (GnT-III). This modification leads to a signifi- cant decrease in EGF binding and EGF receptor autophosphorylation. In addition, the cellular response to EGF was found to be altered. Proliferation of U373 MG cells in serum-free medium is inhibited by EGF. In con- trast, proliferation of the GnT-III-transfected cells was stimulated by EGF. These data demonstrate that changes in EGF three Nonspecific was the addition of anti-mouse antibody to anti-EGF and was a scintillation a and night was then replaced with twice bovine serum then added presence unlabeled range . was of M the three times cells removed by the N were

Glycosylation of the epidermal growth factor (EGF) 1 receptor is essential for its function (1). Treatment of A431 cells that overexpress EGF receptor, with N-linked glycosylation inhibitors such as tunicamycin or glucosamine, reduced ligand binding by more than 50% (2). Addition of the oligosaccharide moiety to EGF receptor during its biosynthesis was found to be essential for the acquisition of EGF binding to the receptor (3). Binding of lectins to the EGF receptor also modulates the receptor function. Concanavalin A and wheat germ agglutinin lectin inhibit EGF binding to the receptor (4). More recently, concanavalin A was described to be a potent inhibitor of EGF receptor autophosphorylation and signaling, suggesting that the oligosaccharide moieties of this receptor may be involved in receptor activation.
EGF receptors expressed in the human glioma cell line, U373 MG, contain bisecting oligosaccharide structure recognized by E-PHA lectin, as well as the tri-or tetra-antennary structure recognized by leucoaglutinating lectin from Phaseolus vulgaris (5). Binding of E-PHA to the cell surface inhibited EGF binding, receptor autophosphorylation, and consequently, cell proliferation. Binding of leucoaglutinating lectin from P. vulgaris, on the other hand, had no effect on any of these phenomena, suggesting that the bisecting oligosaccharide structure may play a role in the EGF receptor function. In this study, we examined the potential role of bisecting oligosaccharide in EGF receptor function by transfection of N-acetylglucosaminyltransfease III (GnT-III) gene into a human glioma cell line, U373 MG. This enzyme catalyzes the addition of N-acetylglucosamine in ␤1-4 linkage to the ␤-linked mannose of the trimannosyl core of N-linked sugar chains to produce a bisecting GlcNac residue (6). The expression of the bisecting structures at the cell surface as well as on EGF receptors, EGF binding, the receptor autophosphorylation, and the cell proliferation were compared between nontransfected and GnT-III-transfected U373 MG cells.

EXPERIMENTAL PROCEDURES
Stable Transfection of the GnT-III Gene into U373 Cells-The human glioma cell line, U373 MG (American Type Culture Collection, Rockville, MD) was used for stable transfection with GnT-III. The 1.6-kb cDNA encoding the open reading frame for GnT-III (6) was inserted into a pcDNA3 expression vector (Invitrogen, San Diego, CA) at the EcoRI and XbaI sites. The vector, pcDNA3, alone or the pcDNA3/GnT-III construct was then transfected into U373 MG cells using a cationic liposome system, DOTAP (Boehringer Mannheim) per the manufacturer's instructions. Transfectants were then selected by antibiotic resistance in cell culture medium containing 800 g/ml G418 (Life Technologies, Inc.). All cell lines and transfectants were maintained using Dulbecco's modified Eagle's medium (DMEM, containing 4.5 g/liter glucose) supplemented with 10% heat-inactivated fetal bovine serum (Whittaker BioProducts, Walkersville, MD) at 37°C in a humidified 10% CO 2 incubator.
GnT-III-mRNA Expression-Total RNA was isolated from 5 ϫ 10 7 to 1 ϫ 10 8 cells using the guanidinium acid/isothiocyanate method (7) followed by CsCl 2 centrifugation. Total RNA (30 g/lane) was electrophoresed in a formaldehyde denaturation gel containing ethidium bromide and transferred to a Duralon nylon membrane (Stratagene, La Jolla, CA). After UV cross-linking, blots were hybridized with 32 Pradiolabeled cDNA probes synthesized from GnT-III containing plasmids using a random priming kit (Boehinger Mannheim) and QuikHyb solution (Stratagene) and then washed at 60°C. The blots were autoradiographed by using X-Omat film (Kodak).
Detection of GnT-III Products-Cells were seeded in 25-cm 2 flasks containing DMEM medium supplemented with 10% FBS, penicillin (50 units/ml), streptomycin (50 g/ml), and L-glutamine (2 mM) and incubated at 37°C in a humidified 95% air, 5% CO 2 incubator. When the flasks were confluent, the cells were washed with cold phosphate-* This work was supported in part by grants from the Illinois division of the American Cancer Society (to H. Y.) and from the Mizutani Foundation for Glycoscience (to E. G. B.) and National Institute of Health Grant NS33383 (to E. G. B.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Analysis of the change in EGF receptor glycosylation was carried out by immunoprecipitation of the receptor; total cell lysates (500 g) were incubated at 4°C overnight with 5 g of anti-EGF receptor monoclonal antibody (clone F4, Sigma) complexed with rabbit anti-mouse IgG and protein A-agarose (Sigma). After washing the pellet, 30 l of Laemmli buffer was added, the mixture was boiled for 5 min, and the proteins in the supernatant were separated by SDS-polyacrylamide gel electrophoresis. The presence of the bisecting oligosaccharide structure on the precipitated EGF receptor was detected by incubation with biotinylated E-PHA. Reactive glycoproteins were detected by strepavidin-peroxidase and ECL reagents.
Detection of EGF Receptor Protein at the Cell Surface-Cells were seeded at 10 5 cell/well in 24-well plates and incubated overnight at 37°C in DMEM containing 10% FBS. The cells were then washed three times with PBS and fixed with 10% buffered formalin for 20 min at room temperature. After three additional PBS washes, the fixed cells were incubated for 1 h at room temperature in PBS containing 10 g/ml anti-human EGF receptor antibody (clone LA1; Upstate Biotechnology, Lake Placid, NY). The cells were then washed with PBS and incubated for 30 min at room temperature with 1 Ci/ml of 125 I-labeled anti-mouse IgG (DuPont NEN) followed by three PBS washes. Nonspecific binding was determined by the addition of radiolabeled anti-mouse antibody to cells without anti-EGF receptor antibody. Cells were solubilized in 200 l of 0.2 N NaOH, and radioactivity was detected with a liquid scintillation counter.
EGF Binding Assays-For cell surface binding, the cells were seeded at a density of 5 ϫ 10 4 cells/ml in 24-well plates and incubated overnight in DMEM containing 10% FBS. The medium was then replaced with serum-free DMEM, after 15 h of incubation at 37°C, and the cells were washed twice with binding buffer (PBS containing 0.1% bovine serum albumin). 125 I-EGF (10 5 cpm) was then added in the presence of unlabeled EGF over a concentration range of 0 -100 nM. Nonspecific binding was determined by the addition of 1 M cold EGF. After incubation for 1 h at room temperature, unbound EGF was aspirated, and the cells were washed three times with binding buffer. The cells were removed by the addition of 200 l of 0.2 N NaOH to each well. Bound radioactivity was then counted in a ␥ counter (8). The K D and the total number of EGF binding sites were determined from the specific bound counts by the method of Scatchard (9).
For binding to the immunoprecipitated receptor, EGF receptors immunoprecipitated on protein A-agarose beads were then incubated with 50,000 cpm 125 I-EGF and 10 nM cold EGF for 15 min on ice. Nonspecific binding was measured by the addition of 1 M cold EGF. The unbound EGF was removed by centrifugation (12,000 ϫ g for 1 min) and washed four times with cold PBS. Radioactivity associated with the immunocomplex-agarose pellet was measured in a ␥ counter.
EGF Receptor Autophosphorylation-Cells were incubated in serumfree medium supplemented with 0 -100 ng/ml of EGF for 10 min or with 100 ng/ml of EGF for 1-15 min at 37°C. After washing twice with ice-cold PBS, the monolayer was solubilized with 200 l of lysis buffer. 50 g of soluble protein was resolved by SDS-polyacrylamide gel electrophoresis and then transferred onto an Immobilon P membrane (Millipore, Bedford, MA). Phosphotyrosine containing proteins were identified by incubating the membrane for 1 h at room temperature with a mouse monoclonal anti-phosphotyrosine antibody (PY-20; Transduction Laboratories, Lexington, KY). Complexes were detected by sequential blotting with biotinylated goat anti-mouse IgG linked to peroxidase. Phosphorylated proteins were detected with ECL reagents (5).
[ 3 H]Thymidine Incorporation-Cells were seeded in 24-well plates at 5 ϫ 10 4 cell/ml in 10% FBS containing DMEM medium and incubated at 37°C for 24 h. The medium was then replaced by DMEM without FBS, and the cells were incubated with different concentrations of EGF for another 24 h at 37°C. [ 3 H]Thymidine (1 Ci/ml) was then added to each well and incubated for 5 h at 37°C (10). The cells were then washed three times with ice-cold PBS, incubated with 5% trichoroacetic acid at 4°C for 30 min, followed by three washes with ice-cold trichoroacetic acid. The trichoroacetic acid-precipitable radioactivity was solublized with base and counted in a liquid scintillation counter.

RESULTS AND DISCUSSION
The role of the oligosaccharide portion in the modulation of the EGF receptor function has been mainly examined by inducing transient modifications in receptor glycosylation, by incubation of cells with glycosylation inhibitors (2) or lectins (4,11). With these approaches, the function of EGF receptor was abolished in cells where glycosyltransferases were inhibited as well as when the cells were treated with certain lectins. Glycosylation of the EGF receptor seems to be tissue-specific. In human carcinoma cell line A431 for example, EGF receptor was found to contain mainly high mannose oligosaccharide structure (2). In a human glioma cell line, U373 MG, EGF receptors contain the bisecting oligosaccharide structure, and interaction of this structure with E-PHA lectin was found to inhibit ligand recognition by the receptor (5).
Stable GnT-III transfectants were generated in a glioma cell line, U373 MG, to determine the consequences of overexpression of the bisecting oligosaccharide structure on EGF receptor binding and function because E-PHA had such a profound effect on binding and function (5). Transfection was verified by GnT-III mRNA expression, which was very high in transfected cells compared with parental cells (Fig. 1A). A prominent band at approximately 2.2 kb was expressed only in the GnT-IIItransfected cells (Lane 2). This band corresponds with the expected message size for the transfected GnT-III. A faint band was also detected at 4.7 kb and is consistent with the endogenous GnT-III (6).
Overexpression of GnT-III mRNA in transfected cells should result in an increase in bisecting GlcNAc products. The presence of GnT-III enzyme products was determined by measuring the increase in E-PHA lectin binding to total proteins extracted from transfected and nontransfected cells. As shown in Fig. 1C, a significant increase in E-PHA binding to the GnT-III-transfected cells was observed when compared with the parental cells or those transfected only with vector alone (pcDNA3). Most of the glycoproteins affected were in the molecular mass range of 70 -120 kDa. These data are in accordance with the findings of Miyoshi et al. (12) that GnT-III transfection of HB611 cells altered N-linked branching of proteins in a similar molecular mass range. The identity of these proteins has not yet been determined. A 67-kDa protein, however, was reported by Ross and collaborators (13) to be an excellent substrate for GnT-III upon treatment of LAN-5 human neuroblastoma cells with retinoic acid.
Glycosylation of the EGF receptor was analyzed on immunoprecipitated receptor (Fig. 1D). In lanes 1 and 2, the blot was probed with biotinylated E-PHA lectin. The 170-kDa glycoprotein, which corresponds to EGF receptor, reacted with the lectin in both parental and transfected cells. EGF receptor from GnT-III-transfected cells also showed significantly more E-PHA staining than EGF receptor from nontransfected cells. Lanes 3 and 4 of Fig. 1D show staining of the immunocomplexes from both cell types by anti-EGF receptor antibody. Reactivity of EGF receptor appeared to be equivalent in both cell types. Taken together, these data indicate that GnT-III transfection was able to increase the amount of bisecting oligosaccharide structure on EGF receptor expressed in the transfected U373 MG cells.
The binding of EGF to its cell surface receptor on U373 MG parental and GnT-III-transfected U373 MG cells was examined. The GnT-III-transfected cells were found to bind much less EGF than the nontransfected cells (Fig. 2). The apparent number of EGF receptors at the cell surface was reduced from about 7 ϫ 10 5 in the parental cells to about 3 ϫ 10 5 in the GnT-III-transfected cells ( Table I). The binding affinity, however, did not seem to be affected.
The decrease in EGF binding at the cell surface may be explained by a decrease in the number of active receptors on the cell surface. This may be due to a decrease in receptor biosynthesis or to transport to the cell membrane or as a consequence of the change in EGF receptor glycosylation. The expression of EGF receptor protein at the cell surface was determined by incubating the cells with monoclonal anti-EGF receptor antibody followed by incubation with 125 I-anti-mouse IgG. As shown in Table I, parental and GnT-III-transfected cells bound equivalent quantities of anti-EGF receptor antibody. These data suggest that similar quantities of EGF receptor protein are expressed on the cell surface of these two cell types. 125 I-EGF binding to equal amounts of immunocomplexed receptors from transfected and nontransfected cells was also determined (Fig. 3). Similar to the reduction observed in cell surface binding, EGF binding to the immunoprecipitated receptor was also reduced by about 50%. These data and those shown in Fig. 1D suggested that the biosynthesis or transport to the cell surface of EGF receptors was not affected by GnT-III transfection of U373 MG cells. Furthermore, the overexpression of bisecting oligosaccharide structures on EGF receptor appears to be responsible for the reduction of EGF binding because the inhibition was still observed on equal amounts of immunoprecipitated EGF receptor protein. A possibility suggested by these results is that there are two populations of receptors: one that binds EGF in a manner that is similar to that of the wild type receptor and a second that is incapable of binding EGF. The alteration in EGF receptor glycosylation by GnT-III transfection may be responsible for the inhibition of EGF binding to these receptors. One or more N-linked oligosaccharide chains on or near the EGF binding site when inappropriately branched may block EGF binding.
As might be expected by the decrease in EGF binding, EGF receptor autophosphorylation was also reduced in U373 MG cells transfected with GnT-III (Fig. 4). Autophosphorylation of EGF receptor (protein migrating around 170 kDa in Fig. 4A) was significantly reduced in the GnT-III-transfected cells compared with the pcDNA3-transfected or parental cells. The kinetics of EGF receptor autophosphorylation upon stimulation of the cells with 100 ng/ml of EGF is represented in Fig. 4B. Once again, there is a significant diminution of EGF receptor

FIG. 1. Expression of GnT-III and its products in parental and transfected U373 MG cells.
Total RNA was extracted from U373 MG parental and transfected cells. A represents a Northern blot on a Duralon Nylon membrane probed with 32 P-radiolabeled cDNA synthesized from GnT-III containing plasmids. The radioactive bands were autoradiographed. Lanes 1, 2, and 3 represent respectively nontransfected U373 MG cells, cells transfected with pcDNA3 containing GnT-III fragment, and cells transfected with pcDNA3 alone. In B, total RNA was stained with ethidium bromide and photographed under UV. C, cell lysates (50 g) were electrophoresed and transferred to Immobilon P membrane. Glycoproteins containing GnT-III products were detected by biotinylated E-PHA lectin. The reactive complexes were stained by incubation of the blot with peroxidase followed by ECL reagents. D, EGF receptors were immunoprecipitated from 500 g of proteins from cell lysates by using anti-EGF receptor linked to protein A-agarose. Glycosylation of EGF receptor in the immunocomplex was detected by Western blot using biotinylated E-PHA (Lanes 1 and 2), and the receptor expression was detected by anti-EGF receptor antibody (lanes 3 and  4). The reactive complexes were then detected using peroxidase and ECL reagents.  a Determined from competition binding curves and analyzed by the method of Scatchard (9). b EGF receptor protein at the cell surface was detected by monoclonal anti-EGF receptor antibody and 125 I-labeled goat anti-mouse IgG as described under "Experimental Procedures." Data are averages Ϯ S.E. of triplicate determinations. phosphorylation in the GnT-III-transfected cells compared with the parental cells. The diminution in EGF receptor activation in transfected cells agrees with our finding on EGF binding and suggests that the modulation of the oligosaccharide branching by GnT-III on the EGF receptor may disturb the activity and subsequently the signaling cascade initiated by this receptor.
One of the final outcomes of EGF signaling is cell proliferation. Addition of EGF to cultures of many tumor cell lines that overexpress the EGF receptor, however, is inhibitory for cell growth (14 -16). The human epidermoid carcinoma cell line A431 is an example of a line that overexpresses EGF receptor and is growth inhibited by EGF (14). A431 cells can be subcloned into two populations that respond differently to EGF. The A1S subclone proliferates in the presence of EGF, and the A5I subclone, like the parental A431, is growth inhibited by EGF (17,18). One of the major differences between these clones is the number of EGF receptors. A1S cells contain about onehalf the number of receptors as A5I (17). This type of data has suggested a possible inverse relationship between EGF receptor number on the cell surface and EGF-induced mitogenesis.
Human gliomas and glioma cell lines also overexpress EGF receptor (19 -21). In our study, U373 MG cells transfected with GnT-III gene were found to grow 50% slower than the parental cells (data not shown). The proliferative response to EGF, however, was found to be different between U373 MG cells and the GnT-III-transfected cells. U373 MG cells are growth inhibited by EGF. As shown in Fig. 5, addition of EGF to the cell culture medium had an inhibitory effect on thymidine incorporation into U373 MG cells. There was about 40% inhibition with less than 1 ng/ml of EGF. There was still at least 50% when 100 ng/ml of EGF was added to the medium (data not shown). The GnT-III-transfected cells, on the other hand, were stimulated in the presence of low concentrations of EGF (from 1 to about 5 ng/ml). This response was dose-dependent and was maximal at . The proteins (50 g) from these cells were electrophoresed and transferred to an Immobilon P membrane. Phosphorylated proteins were detected with anti-phosphotyrosine antibody followed by peroxidase and ECL reagents as described under "Experimental Procedures." FIG. 5. Proliferative response to EGF. Parental and GnT-IIItransfected U373 MG cells were seeded in 24-well plates in 10% FBS containing DMEM medium. After 24 h at 37°C, the medium was replaced with serum-free medium, and EGF was added at the indicated concentrations. Thymidine incorporation was measured as indicated under "Experimental Procedures." Data are averages Ϯ S.E. (bars) of triplicate determinations. Where no error bar is present, the S.E. was smaller than the symbol. about 2 ng/ml EGF. This might be explained by the overexpression of the bisecting structure on EGF receptor reducing the apparent number of EGF binding sites on the cell surface. Similar to the epidermoid carcinoma lines described above, the EGF-dependent proliferation of these cells is inversely proportional to the the number of binding sites on the cell surface. As the number of EGF binding sites decrease on these tumor cell lines there is a change from inhibition to stimulation of proliferation by EGF.
In summary, this study demonstrates that overexpression of the GnT-III products on EGF receptor in U373 MG cells induces significant changes in EGF binding and receptor autophosphorylation. The proliferative response of the cells to EGF was also changed after transfection. The mechanism by which overexpression of bisecting structure in glioma cells alters their growth behavior remains to be determined.