Epidermal growth factor receptor glycosylation is required for ...

Glycobiology vol. 11 no. 7 pp. 515–522, 2001

Epidermal growth factor receptor glycosylation is required for ganglioside GM3 binding and GM3-mediated suppresion of activation

Xiao-Qi Wang2,3, Ping Sun2,3, Maurice O’Gorman2, Tadashi Tai4, and Amy S. Paller1,2,3 2Department

of Pediatrics, Children’s Memorial Hospital, Institute for Education and Research, Northwestern University Medical School, 2300 Children’s Plaza, Chicago, IL 60614, USA, 3Department of Dermatology, Children’s Memorial Hospital, Institute for Education and Research, Northwestern University Medical School, 2300 Children’s Plaza, Chicago, IL 60615, USA, and 4Department of Tumor Immunology, Tokyo Metropolitan Institute of Medical Science, Honkomagome, Bunkyo-ku, Tokyo, Japan Received on July 20, 2000; revised on February 15, 2001; accepted on February 15, 2001

Gangliosides are able to bind to the epidermal growth factor receptor and inhibit its activation, but the mechanism of this inhibition is unknown. To address the role of receptor carbohydrates in facilitating interaction with gangliosides, we examined the ability of GM3 to bind the deglycosylated receptor and inhibit its autophosphorylation. Flow cytometry studies demonstrated that deglycosylation of the receptor did not affect its ability to be transported to the cell membrane. In contrast with the native (fully glycosylated) receptor, GM3 did not coimmunoprecipitate with the deglycosylated receptor. Using a novel colorimetric bead binding assay, GM3 was shown to bind well to the immunoprecipitated native receptor but not at all to the deglycosylated receptor. Finally, the addition of GM3 to cells with deglycosylated epidermal growth factor receptors did not result in significant further inhibition of autophosphorylation of the receptor, despite a 10-fold decrease in phosphorylation of the native epidermal growth factor receptor by 200 µM GM3. These studies suggest that ganglioside affects epidermal growth factor receptor activity through a direct interaction that requires receptor glycosylation, and contribute to our understanding of the role of gangliosides in cell membrane function. Key words: glycosylation/epidermal growth factor receptor/ epidermal growth factor/gangliosides/ keratinocytes Introduction The epidermal growth factor receptor (EGFR) is a large transmembrane glycoprotein that mediates the cellular response to epidermal growth factor (EGF), transforming growth factor α and other EGFR ligands (Carpenter, 1999). Binding of an EGFR ligand to the receptor activates the cytoplasmic tyrosine

1To

whom correspondence should be addressed

© 2001 Oxford University Press

kinase of the receptor, resulting in phosphorylation of the receptor itself (Ullrich and Schlessinger, 1990), as well as of downstream signal pathway components, such as PI3 kinase and MAP kinase, thereby triggering cell proliferation. Sequence analysis of EGFR cDNA from the human epidermoid carcinoma cell line A-431 indicates that EGFR contains 11–12 potential glycosylation sites on the extracellular domain of the receptor (Ullrich et al., 1984; Stroop et al., 2000). The EGFR of A-431 contains both complex-type and high mannosetype N-linked oligosaccharides in the approximate ratio of 2:1, but no O-linked oligosaccharides (Cummings et al., 1985). The complex-type oligosaccharides in the EGFR are predominantly tri- and/or tri- and tetraantennary chains, with terminal Neu5Ac, Fuc, and α-GalNAc. Treatment of A-431 cells with N-linked glycosylation inhibitors, such as tunicamycin or glucosamine, reduces EGF binding to its receptor by more than 50% (Soderquist and Carpenter, 1984). Molecules that interact with the oligosaccharide moieties of the EGFR, such as lectins, have also been shown to compete with EGF in binding to the EGFR and to inhibit EGF-stimulated cell proliferation (Wakshull and Wharton, 1985; Moseley and Suva, 1986; Kaplowitz and Haar, 1988). Furthermore, glycosylation defective cells show a decreased ability to bind 125I-EGF (Pratt and Pastan, 1978), and an Asn420Gln mutation in the extracellular region of the EGFR both prevents glycosylation at the mutation site and results in constitutive activation of dimerization and inability to bind EGF (Tsuda et al., 2000). These studies attest to the critical role of glycosylation in EGFR function. Gangliosides are glycosphingolipids characterized by the presence of one or more sialic acid moieties in the oligosaccharide chain (Hakomori and Igarashi, 1993). The role of gangliosides, which are localized on the outer leaflet of the plasma membrane, is largely unknown, but many studies suggest that gangliosides may be involved in regulating cellular proliferation, differentiation, and oncogenic transformation. GM3, a monosialylated ganglioside that predominates in the membrane of normal keratinocytes, keratinocyte-derived SCC12 cells, and A-431 cells, inhibits EGFR autophosphorylation, downstream signal transduction pathway component phosphorylation, and proliferation of these cells (Bremer et al., 1986; Paller et al., 1993, 1995; Rebbaa et al., 1996), at least in part by inhibiting the binding of the EGFR to its ligands (Wang et al., 2001a). Given that GM3 coimmunoprecipitates with the EGFR (Hanai et al., 1988) and is able to bind to the EGFR by enzyme-linked immunosorbent assays (Yednak and Bremer, 1994), the inhibition by ganglioside of ligand binding to the EGFR likely requires the direct binding of ganglioside to the receptor. Sitedirected mutagenesis and gene truncation techniques have 515

X.-Q. Wang et al.

elucidated the protein–protein interactions that facilitate ligand binding to the extracellular domain of the EGFR and initiate phosphorylation of the EGFR and downstream pathway components (Lax et al., 1990; Massoglia et al., 1990; Li et al., 1991; Takahashi et al., 1999; Wong et al., 1999). In contrast, the mechanism of the interaction of ganglioside GM3 with the EGFR that interferes with its activation is not well understood. By depleting the EGFR of its carbohydrate moieties through treatment with both peptide-N-glycosidase F (PNGase F) and tunicamycin, we examined the role of the receptor carbohydrates in facilitating the interaction of the receptor with GM3. Binding assays and functional assays demonstrated that the presence of these carbohydrate residues on the EGFR is critical for allowing ganglioside to bind to the receptor and to inhibit its autophosphorylation. Results Western blotting and staining with periodate–Schiff reagent were used to confirm loss of carbohydrate from the EGFR after treatment with tunicamycin and PNGase F. Western blotting analysis with anti-EGFR antibody of the deglycosylated EGFR from A-431 and SCC12 cells showed a diminution in the apparent molecular weight from ∼170 kDa for native EGFR to ∼130 kDa for the deglycosylated receptor (Figure 1A). Periodate– Schiff reagent staining showed virtually no detectable carbohydrate staining (Figure 1B). The putative interaction between the EGFR and ganglioside occurs at the cell membrane. To consider whether deglycosylation alters the ability of the EGFR to be transported to the cell surface, flow cytometric studies with anti-EGFR antibody were performed, comparing native EGFR with EGFR deglycosylated by tunicamycin alone, PNGase F alone, and by the combination of the tunicamycin and PNGase F. The level of expression of EGFR at the cell membrane did not differ significantly among all four groups of treated SCC12 cells (Figures 2A–D) and of treated A-431 cells (data not shown). Specificity of the binding of the anti-EGFR antibody was confirmed by blocking experiments using two different unlabeled anti-EGFR antibodies, which showed a diminution in the intensity of staining of greater than 90% (Figure 2E). These studies provide strong evidence that the EGFR deglycosylated by tunicamycin and PNGase F is able to be transported normally to the membrane and is thus available for interaction with ganglioside. To determine the ability of GM3 to complex with an EGFR devoid of its carbohydrate moieties, the recovery of GM3 after immunoprecipitation of glycosylated and deglycosylated receptors was compared. To increase the sensitivity of the assay, A-431 or SCC12 cells were preloaded for 48 h with 10 nM to 200 µM GM3. As negative controls, the cells were preloaded with the same concentrations of GM2, a monosialylated ganglioside that does not affect EGFR phosphorylation or cell proliferation, or with the DMEM/F12 medium with 2% fetal bovine serum alone. After EGFR immunoprecipitation, an aliquot of the precipitated EGFR at each condition was checked by Western blot analysis to ensure that equal amounts of the precipitated receptor were present in each sample and to demonstrate the lack of effect of deglycosylation with tunicamycin and PNGase F on EGFR expression (Figure 3A). 516

Fig. 1. Treatment of cells with tunicamycin and PNGase F effectively deglycosylates the EGFR. After reaching 80% confluence, A-431 cells and SCC12 cells were switched to serum-free medium with 2 µg/ml tunicamycin for 22 h, and then 2 U/ml PNGase F was added for an additional 2 h. Cells were lysed, and the EGFR was immunoprecipitated with anti-human EGFR monoclonal antibody bound to Protein A: agarose. The immunoprecipitated EGFR was detected by Western blotting with anti-human EGFR antibody and ECL. Deglycosylation decreases the molecular weight of the EGFR from 170 kDa to 130 kDa (A). Virtually no carbohydrate is detected by periodate–Schiff reagent staining (B). D-A431 and D-SCC12 are the deglycosylated forms of A-431 and SCC12, respectively. D-EGFR refers to the deglycosylated EGF receptor.

Thin-layer chromatography (TLC) immunostaining of ganglioside extracted from the immunoprecipitated EGFR demonstrated the coimmunoprecipitation of glycosylated EGFR and GM3 but not GM2, even at a concentration of supplemental GM2 as high as 200 µM. No ganglioside coimmunoprecipitated with the deglycosylated receptor (Figure 3B). To allow visualization of the direct interaction of glycosylated or deglycosylated EGFR and GM3, a novel colorimetric bead binding assay was utilized. GM3 was crosslinked to colored amine-modified FluoSphere beads, and allowed to bind to the native (glycosylated) or deglycosylated immunoprecipitated EGFR from A-431 or SCC12 cells, coated onto protein A: agarose beads. The interaction between the ganglioside-coated FluoSphere beads and the EGFR protein A: agarose beads was visualized by immunofluorescence microscopy. As shown in Figure 4, GM3 bound well to the native EGFR from A-431 cells (Figure 4A) and SCC12 cells (Figure 4B), but did not bind to deglycosylated EGFR from A-431 cells (Figure 4C) or SCC12 cells (Figure 4D). GM2 did not bind with glycosylated EGFR from either A-431 cells (Figure 4E) or SCC12 cells (Figure 4F). Beads coated with 1% bovine serum albumin

Carbohydrates required for EGFR binding to GM3

Fig. 3. GM3 coimmunoprecipitates with glycosylated EGFR, but not with the deglycosylated receptor. Cells were treated for 48 h with or without GM3 or, as an additional control, with 200 µM GM2. A-431 or SCC12 cells were treated with tunicamycin and PNGase F during the last 24 h of the incubation. After immunoprecipitation of the receptor, the immunoprecipitated EGFR was detected with anti-human EGFR antibody. Total lipids were extracted from the remaining samples for TLC. Ganglioside was identified by immunostaining on an aluminum-backed TLC plate with antibodies directed against GM3 and GM2. Binding was detected with rabbit anti-mouse IgG and enhanced chemiluminescence. D-A431 and D-SCC12 are the deglycosylated forms of A-431 and SCC12, respectively. D-EGFR refers to the deglycosylated EGF receptor.

Fig. 2. Level of expression of membrane EGFR by flow cytometry is equivalent in the native EGFR and the EGFR deglycosylated by tunicamycin and PNGase F. SCC12 cells at 85% confluence were treated with or without (A) either 24 h of tunicamycin alone (B), 2 h of PNGase F alone (C), or the combination of 24 h tunicamycin and 2 h PNGase F (D). Cells in single cell suspension were mixed with FITC-conjugated mouse anti-human EGFR IgG2b for 2 h at 4°C, then fluorescence was detected in a FACScan flow cytometer. FITC-conjugated anti-mouse IgG2b antibodies were used as a negative control, and cells were pretreated with one of two unlabeled anti-EGFR antibodies before incubation with FITC-conjugated anti-EGFR antibody was used as a control for specificity to binding to the EGFR (E). In (E), the first peak is the isotype control; the second peak is an anti-EGFR antibody that recognizes the extracellular region of the EGFR from amino acids 6 to 273; the third peak is an anti-EGFR antibody that recognizes the extracellular portion of the EGFR from amino acids 351–364; and the fourth peak is treatment with the FITC-conjugated anti-EGFR antibody (also recognizing amino acids 6–273) without any blocking antibodies. Fluorescence intensity was expressed as arbitrary units (geometric mean fluorescence channel, mfc), and values were calculated by subtracting the mean fluorescence of the isotype control-labeled cells from the mean fluorescence of the anti-EGFR antibody-labeled cells. Flow cytometric assays were performed three times, with triplicate samples at each analysis. All analyses were performed with Cellquest R software (Becton-Dickinson).

(BSA) and uncoated beads showed no binding (data not shown). In view of the likely importance of binding of GM3 to the EGFR in inhibiting receptor activation, the effect of deglycosylation on ganglioside-mediated autophosphorylation was assessed. Ten nanomolar GM3 reduced the EGF-stimulated autophosphorylation of the native receptor by twofold and 200 µM GM3 by 10-fold (Figure 5A). Deglycosylation reduced EGFR autophosphorylation in the presence of 10 nM EGF by eight- to ninefold, regardless of whether the EGFR was deglycosylated by the combination of tunicamycin and PNGase F (Figure 5B) or with PNGase F alone (Figure 5C).

GM3, even at concentrations as high as 200 µM, did not further decrease the phosphorylation of the EGFR in deglycosylated SCC12 cells (Figure 5B, C) or A-431 cells (data not shown). Phosphorylation of the deglycosylated receptor was detectable in the absence of EGF stimulation, although almost fourfold less than when stimulated with EGF (Figure 5C) and sixfold less than that of the native receptor without exposure to EGF. Neither 10 nM GM3 nor 200 µM GM3 had any effect on phosphorylation of the deglycosylated receptor, regardless of the presence or absence of supplemental EGF (Figure 5C). GM2 had no effect on the autophosphorylation of either deglycosylated or native EGFR. Discussion Increasing evidence supports the concept that the carbohydrate side chains of glycoconjugates are critical for physiologic and pathological processes, such as cell proliferation, adhesion, migration, wound healing, and tumor cell invasion and metastasis (Allan et al., 1985; Taniguchi et al., 1996; Goupille et al., 1997; Bagriacik and Miller, 1999; Meuillet et al., 1999). EGFR, a transmembrane glycoprotein, regulates cell behavior through its protein tyrosine kinase domain. Deglycosylated EGFR or EGFR in glycosylation defective cells, although it has an intact protein tyrosine kinase, is unable to become fully phosphorylated and trigger cell proliferation (Pratt and Pastan, 1978), suggesting an important role for the extracellular EGFR oligosaccharides in modulating the function of its cytoplasmic kinase domain. Glycosphingolipids, such as GM3, have been implicated as modulators of signal transduction by influencing protein kinases associated with growth factor receptors (Hakomori, 1990). Bremer and co-workers first described the inhibition by 517

X.-Q. Wang et al.

Fig. 4. GM3 binds directly to native but not deglycosylated EGFR. 50 µM gangliosides GM3 or GM2 or 1% BSA in PBS was crosslinked to colored amine-modified FluoSphere beads by mixing overnight with EDAC. Glycosylated or deglycosylated EGFR was coated onto protein A: agarose beads by immunoprecipitation with anti-EGFR monoclonal antibody, and coated or uncoated protein A: agarose beads were mixed with the ganglioside-coated or uncoated beads in a ratio of 1:10. Binding of the FluoSphere beads to the surface of the protein A: agarose beads was visualized with a Nikon Eclipse TE300 immunofluorescence microscope at 200× magnification. GM3 directly binds to native EGFR from A-431 cells (A) and SCC12 cells (B), but does not bind to deglycosylated EGFR from A-431 cells (C) or SCC12 cells (D). Native EGFR from A-431 cells (E) and SCC12 cells (F) cannot bind GM2-coated (E, F) FluoSphere beads. Bars = 12 µm.

exogenously added GM3 of cell proliferation and EGFR autophosphorylation in cultured A-431 cells (Bremer et al., 1984, 1986). Similarly, we have shown that GM3 inhibits the proliferation of keratinocytes and several keratinocyte-derived cell lines (Paller et al., 1993) through inhibiting receptor autophosphorylation (Paller et al., 1995). Recently, we have determined that this suppression of cell proliferation and EGFR autophosphorylation results, at least in part, from a GM3-induced decrease in the number of EGFRs available for binding to ligand (Wang et al., 2001a), presumably involving the direct binding of GM3 to the receptor, as demonstrated in our study and by others (Hanai et al., 1988; Yednak and Bremer, 1994). 518

Fig. 5. GM3 does not further inhibit the autophosphorylation of deglycosylated EGFR. Cells were pretreated with or without gangliosides GM3 (10 nM or 200 µM) or GM2 (200 µM) for 48 h in serum-free medium without EGF. Tunicamycin 2 µg/ml, PNGase F 2 U/ml were added at 24 h and 2 h before cell lysis, respectively, to deglycosylate the cells. Cells starved of EGF were treated with or without EGF 10 ng/ml for 10 min before cell lysis. Cells were boiled in lysis buffer, then centrifuged to remove unlysed cell fragments. 8 µg protein from the lysate of SCC12 cells (A) or 35 µg protein from cell lysates of SCC12 cells treated with tunicamycin and PNGase F (B) or by PNGase F alone (C) was then boiled for 10 min with sample buffer, and proteins were separated on an 8% SDS–PAGE mini gel. Mouse monoclonal antibody specifically directed against the phosphorylated form of EGFR was used to detect activated EGFR. Immunoreactive bands were treated with peroxidase-conjugated goat anti-mouse antibodies, then detected with an ECL kit, and developed on Kodak X-Omat film at room temperature for 10 s. D-EGFR and D-SCC12 refer to the deglycosylated forms of the EGF receptor and SCC12, respectively.

The increase in membrane GM3 that occurs after pharmacologic addition (Laine and Hakomori, 1973; Paller et al., 1993) would be expected to provide more GM3 to compete with ligand in attaching to the receptor and triggering its phosphorylation. The precise mechanism of how GM3 binds to the EGFR and suppresses ligand binding and activation is still unknown. A-431 cells have been the most common cell model for studies of the EGFR because of the high density of EGFR per cell. However, the number of receptors (180,000/cell) and ganglioside content of the keratinocyte-derived SCC12 cell line closely resemble that of the normal keratinocyte (Wang et al., 2001b); thus, the SCC12 cell may be more relevant model for studying ganglioside/receptor interaction than the A-431 cell with its 2 × 106 receptors per cell (Haigler et al., 1978) and GM3 as its sole


detectable membrane ganglioside. As a result, the effects of deglycosylation of the epidermal growth factor receptors of both A-431 cells and SCC12 cells were assessed. Previous investigations have demonstrated the importance of the carbohydrate residues of ganglioside GM3 in inhibiting the phosphorylation of the EGFR (Bremer et al., 1984, 1986; Paller et al., 1993; Meuillet et al., 1999). To determine the role of carbohydrate moieties of the EGFR in the interaction with ganglioside, the EGFR was deglycosylated using both tunicamycin and PNGase F, a combination that reduced the molecular weight of the EGFR by 28%, and markedly decreased sugar residues as detected by periodate-Schiff assays. After deglycosylation of the EGFR from A-431 and SCC12 cells in culture, the receptor could to be immunoprecipitated with antiEGFR antibody; however, in contrast with the immunoprecipitated native EGFR, the immunoprecipitated deglycosylated receptor was not complexed with GM3 and was unable to bind to GM3 in a novel fluorescent bead binding assay. Loss of the direct interaction of the deglycosylated EGFR with GM3 suggests that the relationship between GM3 and the EGFR occurs through carbohydrate–carbohydrate interactions. Soderquist and Carpenter (1984) have previously shown that deglycosylation of the EGFR by treatment for 24 h with tunicamycin alone decreases the ability of the receptor to bind to EGF by 41%; similarly, our studies with SCC12 cells using 125I-EGF showed a reduction in binding to the EGFR deglycosylated with both tunicamycin and PNGase F by 50–70% (unpublished data). Although the residual ability to bind EGF has been attributed to receptors synthesized before addition of tunicamycin, based on the 20 h half-life for the EGFR from A-431 cells and inability of tunicamycin to degrade the receptor (Soderquist and Carpenter, 1984), our finding of 30–50% binding of ligand to the EGFR in the presence of receptor deglycosylated by both tunicamycin and PNGase F suggests that, in fact, binding to the deglycosylated receptor is reduced but not eliminated. Deglycosylation of the receptor also led to an eight- to ninefold reduction in autophosphorylation, although phosphorylation was clearly detectable and increased in the presence of EGF, regardless of whether cells were deglycosylated with both tunicamycin and PNGase F or with PNGase F alone. Ganglioside GM3 was not able to further inhibit the phosphorylation of the receptor in its deglycosylated state, regardless of the presence or absence of EGF. Though these findings provide evidence that direct interaction of GM3 with the EGFR and the effect of ganglioside on receptor function require receptor glycosylation, an alternate possible explanation for the inability of GM3 to further inhibit phosphorylation and to be complexed with immunoprecipitated deglycosylated EGFR is sequestration of the deglycosylated receptor intracellularly and inability to be transported to the membrane related to its poor glycosylation. Given that exogenously added GM3 either inserts into the outer leaflet of the plasma membrane or interacts with cell surface EGFR without inserting into the membrane, sequestration of the receptor intracellularly would obviate interaction with ganglioside. In fact, treatment for 48 h with tunicamycin has been shown to decrease the recovery of membrane EGFR by immunoprecipitation (Soderquist and Carpenter, 1984), suggesting that tunicamycin affects transport of the EGFR to the membrane. In contrast, our flow cytometric studies with anti-EGFR antibodies showed transport of the deglycosylated

receptor to the membrane, with no diminution in receptor expression on the membrane of either SCC12 cells or A-431 cells treated for 24 h with tunicamycin and 2 h with PNGase F, a combination of agents chosen to inhibit glycosylation as completely as possible. Similarly, no difference in expression was demonstrated when cells were treated with 24 h of tunicamycin or PNGase F alone. The specificity of recognition by the receptor was verified by > 90% diminution in binding to the labeled antibody when blocked by unlabeled anti-EGFR antibodies. The poor binding of GM3 to the deglycosylated EGFR and lack of significant further suppression of autophosphorylation of the EGFR by ganglioside suggests that the carbohydrate moieties of the EGFR are critical for the direct interaction of the receptor with GM3 and its inhibition of receptor phosphorylation. The importance of receptor glycosylation has recently been described as well in the relationship between highly sialylated gangliosides and α5β1 integrin. Using affinity-purified α5β1 from SCC12 cells and both insect and Escherichia coli recombinant α5 and β1 integrin proteins, Wang et al. (2001b) have shown that gangliosides GD3 and GT1b bind directly to the α5 subunit of the α5β1 receptor and thus are able to block cell interaction with fibronectin. The GT1b/α5β1 interaction similarly requires the presence of the carbohydrate chains of α5β1. It is unknown whether the role of the receptor carbohydrates of both the EGFR and α5β1 is stabilization of the secondary protein structure of the receptor or direct interaction of the receptor carbohydrate moieties with both the receptor ligand (e.g., EGF for the EGFR or the RGD sequence of fibronectin for α5β1) and the interacting ganglioside. Unfortunately, no conformation-specific anti-EGFR antibodies are available to determine retention of secondary protein structure. Recently, Iwabuchi et al. (1999) found that GM3 is localized to a distinct membrane fraction now termed the “glycosphingolipid signaling domain” or “glycosignaling domain.” EGFR, in contrast, is concentrated in vesicular invaginations of the plasma membrane called caveolae (Couet et al., 1997), although it rapidly leaves this caveolar membrane domain in response to EGF (Mineo et al., 1999). The demonstration that GM3 and perhaps other gangliosides are able to bind directly to EGFR only when it is glycoyslated to inhibit receptor activation provides a hypothetical model of carbohydrate– carbohydrate interaction between the EGFR and gangliosides in the “glycosignaling domain.” It remains to be determined whether this occurs between neighboring cells or perhaps, with membrane mobility, within the membrane of the same cell. Materials and methods Cell culture SCC12F2 (SCC12) cells, a generous gift from James Rheinwald (Boston, MA), and A-431 cells (American Type Tissue Collection, Manassas, VA) were grown at 37°C in DMEM/F12 (1:1) supplemented with 10% heat-inactivated fetal bovine serum (Gibco BRL, Grand Island, NY), except as otherwise indicated. Deglycosylation of the EGFR Carbohydrate residues were removed from A-431 or SCC12 cell proteins prior to lysis with the combination of PNGase F 519

X.-Q. Wang et al.

(cleaves asparagine-linked high mannose and hybrid and complex oligosaccharides) and tunicamycin (inhibits all N-linked glycosylation), rather than with PNGase F alone, because our previous studies have demonstrated a slightly superior deglycosylation with the combination (Wang et al., unpublished data). Cells were deglycosylated by a modification of the combined methods of Allan et al. (1985) and Zheng et al. (1994). After reaching 80% confluence, SCC12 cells were switched to serum-free medium with 2 µg/ml tunicamycin for 24 h. Two U/ml PNGase F was added for the last 2 h of incubation. Immunoprecipitation of native and deglycosylated EGFR and determination of glycosylation Cells were washed in 10 mM Tris–HCl, pH 7.4, 0.15 M NaCl, 1 mM MnCl2, and 0.2 mM phenylmethylsulfonyl fluoride (PMSF), then lysed with 10 mM Tris–HCl, pH 7.4, 0.15 M NaCl, 1 mM MnCl2, 3 mM PMSF, and 0.1 M octyl glucoside. The EGFR was immunoprecipitated with anti-human EGFR monoclonal antibody that recognizes the intracellular portion of the receptor (Transduction Laboratories, Lexington, KY) bound to protein A: agarose (5 µg anti-EGFR antibody per 500 µg total protein from the cell lysate). The immunoprecipitated EGFR was then boiled in Laemmli buffer (Laemmli, 1970), and applied to a 7.5% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) mini gel. Protein was transferred to polyvinylidene fluoride (PVDF) filters, and nonspecific binding was blocked by incubation for 1 h with 5% dry milk in 50 mM Tris–HCl, pH 7.6 with 137 mM NaCl, and 0.1% Tween-20. Lanes were treated with 50 ng/ml anti-human EGFR antibody (Transduction Laboratories) for 1 h, and the immunoprecipitated EGFR was detected by enhanced chemiluminescence (ECL) using goat anti-mouse secondary antibodies and an enhanced chemiluminescence (ECL) kit (Amersham Pharmacia, Piscataway, NJ) following manufacturer’s instructions. Glycosylation of the receptor was determined by the molecular weight determination and staining of the native or deglycosylated EGFR with a modified periodate–Schiff reagent (GelCode, Pierce, Rockford, IL). The immunoprecipated purified native or deglycosylated EGFR was boiled in Laemmli buffer (Laemmli, 1970), applied to an 8% SDS–PAGE mini gel, stained with Coomassie blue or the periodate–Schiff reagent, and compared with known molecular weight standards. Flow cytometric assays SCC12 cells at 85% confluence were treated with or without either 24 h of tunicamycin alone, 2 h of PNGase F alone, or the combination of 24 h tunicamycin and 2 h PNGase F as described above. Assays were performed as previously described (Sung et al., 1998). Cells were trypsinized gently, the trypsin was neutralized, and the single cell suspension was centrifuged. After washing in phosphate buffered saline (PBS), the suspension was adjusted to 106 cells per 90 µl PBS, and mixed with 10 µl fluorescein isothiocyanate (FITC)-conjugated mouse anti-human EGFR IgG2b (Vector Labs, Burlingame, CA) that recognizes the extracellular region of the EGFR from amino acids 6 to 273 for a final 1:20 dilution. After incubation for 2 h at 4°C, cells were centrifuged, washed with PBS, and resuspended in 300 µl PBS for measurement of EGFR expression in a FACScan flow cytometer (Becton-Dickinson). FITC-conjugated 520

anti-mouse IgG2b antibodies were used as a negative control. Two different unlabeled anti-EGFR antibodies (10 µg/ml) that recognize portions of the extracellular region (one from amino acids 6 to 273, the other from amino acids 351 to 364, both from NeoMarkers [Fremont, CA]) were added to the cells at 37°C for 2 h prior to staining with FITC-conjugated antiEGFR antibody to confirm the specificity of the binding to the EGFR. The viability of cells was verified by trypan blue dye exclusion. Fluorescence intensity was expressed as arbitrary units (geometric mean fluorescence channel, mfc), and values were calculated by subtracting the mean fluorescence of the isotype control-labeled cells from the mean fluorescence of the anti-EGFR antibody-labeled cells. Flow cytometric assays were performed three times, with triplicate samples at each analysis. All analyses were performed with Cellquest R software (Becton-Dickinson). Coimmunoprecipitation of ganglioside with EGFR and detection of gangliosides SCC12 cells and A-431 cells were treated for 48 h with or without GM3 at concentrations of 10 nM, 100 nM, 1 µM, 10 µM, 50 µM, 100 µM, or 200 µM or with 200 µM GM2. Cells were treated with tunicamycin and PNGase F as described above during the last 24 h of the incubation. After immunoprecipitation of the receptor, one-eighth of the immunoprecipitated mixture was boiled in Laemmli buffer, and immunoprecipitated EGFR was detected with anti-human EGFR antibody to ensure that equal amounts of the EGFR were extracted. Total lipids were extracted from the remaining samples for TLC with chloroform:methanol 2:1 (v/v) (Paller et al., 1992). The aqueous phase was separated and desalted, and the bands separated by TLC in chloroform:methanol:water in 0.02% CaCl2, 55:45:10, v/v/v. Ganglioside was identified by resorcinol staining on silica gel plates and by immunostaining on aluminum-backed TLC plates as previously described (Arnsmeier and Paller, 1995). Briefly, after migration, the TLC plate was air-dried and fixed with 0.05% poly(isobutyl methacrylate) in N-hexane solution. Nonspecific binding was blocked by 1% BSA, and the plate was incubated with mouse antibodies directed against GM3 (courtesy of Dr. T. Tai, Tokyo) and GM2 (courtesy of Dr. P. Livingston, New York), or with 1% BSA as a negative control for 1.5 h at room temperature. Binding was detected with rabbit anti-mouse IgG and an ECL kit on X-Omat film. Band density was quantified by the Fluorescence Phosphorimage Storm 800 (Molecular Dynamics, Sunnyvale, CA ). FluoSphere bead binding assays To visualize the effects of receptor deglycosylation on the ability of GM3 to bind to the receptor, fluorescence bead binding assays were performed (Sun et al., unpublished data). Fifty micromolar gangliosides GM3 or GM2 (diluted in PBS from 5 mM in DMSO stock solution) or 1% BSA in PBS was crosslinked to 2.5 × 104 particles of 1 µM yellow-green aminemodified FluoSphere beads (1 × 105 particles/ml in PBS; F-8765, Molecular Probes, Eugene, OR) by mixing overnight at 4°C with 5 mg/ml 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide (EDAC) (Molecular Probes). The EDAC promotes attachment of a ganglioside carboxyl group to an amine group on the FluoSphere polystyrene bead through a chemical reaction. Carbodiimides have been used to fix several glycoproteins


without compromising function (Gold and Pearlstein, 1979; Ito et al., 1991; Drabick et al., 1998). The formation of a peptide bond that includes a carboxyl group of ganglioside has not been shown to disrupt function of the binding of ganglioside GT1b to α5β1 (Wang et al., 2001b), or GM1 to the cholera toxin β subunit (unpublished data). To verify ganglioside binding to the FluoSphere beads prior to mixing with the receptor, 20 µl of GM3-coated, GM2-coated, BSA-coated, or uncoated green beads were washed and loaded into each slot of a positive charged nylon membrane on a Bio-Dot SF apparatus (BioRad). The binding of gangliosides was detected with antiGM3 or anti-GM2 antibody (1:2 in PBS), followed by ECL for detection. The colored beads were easily visible as green bands. Cells pretreated with or without tunicamycin and PNGase F were lysed and the EGFR immunoprecipitated as described above. Protein A: agarose beads coated with EGFR (or uncoated as a control) were mixed with the gangliosidecoated (or uncoated) green FluoSphere beads in a v/v ratio of 1:10 in 96-well plates, and incubated for 1–2 h at room temperature before washing. The binding of the green FluoSphere beads to the surface of the protein A: agarose beads was visualized with a Nikon Eclipse TE300 immunofluorescence microscope at 200× magnification at wavelength 505–515 nm, linked to a computer with Neurolucida software (Microbright Field, Colchester, VT). EGFR expression and phosphorylation To determine the effect of depletion of EGFR carbohydrate moieties on the ability of ganglioside GM3 to inhibit phosphorylation, cells were pretreated with or without gangliosides GM3 (10 nM or 200 µM) or GM2 (200 µM) for 48 h in serum-free medium without EGF. Tunicamycin 2 µg/ml, PNGase F 2 U/ml, and EGF 10 ng/ml were added at 24 h, 2 h, and 10 min, respectively, before cell lysis. Cells were boiled in lysis buffer (1% SDS, 10 mM Tris–HCl, pH 7.4, 1.0 mM Na3VO4), then centrifuged to remove unlysed cell fragments. Eight micrograms of protein from the lysate of A-431 or SCC12 cells or 35 µg protein from the deglycosylated cell lysates of the A-431 cells or SCC12 cells was then boiled for 10 min with sample buffer (Laemmli, 1970), and proteins were separated on an 8% SDS–PAGE mini gel. Two hundred fifty ng/ml mouse monoclonal antibody specifically directed against the phosphorylated form of EGFR (Transduction Laboratories, Lexington, KY) was used to detect activated EGFR. Immunoreactive bands were treated with peroxidase-conjugated goat anti-mouse antibodies, then detected with an enhanced chemiluminescence kit, and developed on Kodak X-Omat film at room temperature for 10 s. Cells grown in medium without tunicamycin and PNGase or with PNGase alone served as a control for the action of ganglioside on the native EGFR phosphorylation. Phosphorylation in the presence of EGF was also compared with phosphorylation of cells still starved of EGF for the additional 10 min. All studies were performed at least three times. Bands were measured by densitometry to determine relative extents of phosphorylation. Statistical analysis All data were analyzed statistically by Student’s t test, with p < 0.05 considered to be significant. Data are expressed as means ± standard deviations.

Acknowledgments These studies were supported by National Institutes of Health grant R01 AR44619. The authors thank Drs. Eric Bremer and Bruce Fenderson for their advice and support. Abbreviations BSA, bovine serum albumin; D-A431, deglycosylated A-431 cells; D-EGFR, deglycosylated EGFR; D-SCC12, deglycosylated SCC12 cells; ECL, enhanced chemiluminescence; EDAC, 1-ethyl3-(3-dimethylaminopropyl)-carbodiimide; EGF, epidermal growth factor; EGFR, epidermal growth factor receptor; FITC, fluorescein isothiocyanate; PBS, phosphate buffered saline; PMSF, phenylmethylsulfonyl fluoride; PNGase F, peptide-Nglycosidase F; PVDF, polyvinylidene fluoride; TLC, thin layer chromatography. References Allan, J.S., Coligan, J.E., Barin, F., McLane, M.F., Sodroski, J.G., Rosen, C.A., Haseltine, W.A., Lee, T.H., and Essex, M. (1985) Major glycoprotein antigens that induce antibodies in AIDS patients are encoded by HTLV-III. Science, 228, 1091–1094. Arnsmeier, S.L., and Paller, A.S. (1995) Chemiluminescence detection of gangliosides by thin-layer chromatography. J. Lipid Res., 36, 911–915. Bagriacik, E., and Miller, K.S. (1999) Cell surface sialic acid and the regulation of immune cell interactions: the neuraminidase effect reconsidered. Glycobiology, 9, 267–275. Bremer, E.G., Hakomori, S.I., Bowen-Pope, D.F., Raines, E., and Ross, R. (1984) Ganglioside-mediated modulation of cell growth, growth factor binding, and receptor phosphorylation. J. Biol. Chem., 259, 6818–6825. Bremer, E.G., Schlessinger, J., and Hakomori S.I. (1986) Gangliosidemediated modulation of cell growth: specific effect of GM3 on tyrosine phosphorylation of the epidermal growth factor receptor. J. Biol. Chem., 261, 2434–2440. Carpenter, G. (1999) Employment of the epidermal growth factor receptor in growth factor-independent signaling pathways. J. Cell Biol., 146, 697–702. Couet, J., Sargiacomo, M., and Lisanti, M.P. (1997) Interaction of a receptor tyrosine kinase, EGF-R, with caveolins. Caveolin binding negatively regulates tyrosine and serine/threonine kinase activities. J. Biol. Chem., 272, 30429–30438. Cummings, R.D., Soderquist, A.M., and Carpenter, G. (1985) The oligosaccharide moieties of the epidermal growth factor receptor in A-431 cells: presence of complex-type N-linked chains that contain terminal N-acetylgalactosamine residues. J. Biol. Chem., 260, 11944–11952. Drabick, J.J., Bhattacharjee, A.K., Hoover, D.L., Siber, G.E., Morales, V.E., Young, L.D., Brown, S.L., and Cross, A.S. (1998) Covalent polymyxin B conjugate with human immunoglobulin G and an antiendotoxin reagent. Antimicrob. Agents Chemother., 42, 583–588. Gold, L.I., and Pearlstein, E. (1979) Stability of fibronectin biological activity following chemical modification. Biochim. Biophys. Acta, 581, 237–251. Goupille, C., Hallouin, F., Meflah, K., and Le Pendu, J. (1997) Increase of rat colon carcinoma cells tumorigenicity by alpha(1-2) fucosyltransferase gene transfection. Glycobiology, 7, 221–229. Haigler, H., Ash, J.F., Singer, S.J., and Cohen, S. (1978) Visualization by fluorescence of the binding and internalization of epidermal growth factor in human carcinoma cell A-431. Proc. Natl Acad. Sci. USA, 75, 3317–3321. Hakomori, S.I. (1990) Bifunctional role of glycosphingolipids: modulators for transmembrane signaling and mediators for cellular interaction. J. Biol. Chem., 265, 18713–18716. Hakomori, S.I., and Igarashi, Y. (1993) Gangliosides and glycosphingolipids as modulators of cell growth, adhesion, and transmembrane signaling. Adv. Lipid. Res., 25, 147–162. Hanai, N., Nores, G.A., MacLeod, C., Torres-Mendez. C.R., and Hakomori, S.I. (1988) Ganglioside-mediated modulation of cell growth. Specific effects of GM3 and lyso-GM3 in tyrosine phosphorylation of the epidermal growth factor receptor. J. Biol. Chem., 263, 10915–10921.

521

X.-Q. Wang et al.

Ito, Y., Kajihara, M., Imanishi, Y. (1991) Materials for enhancing cell adhesion by immobilization of cell-adhesive peptide. J. Biomed. Mater. Res., 25, 1325–37. Iwabuchi, K., Handa, K., and Hakomori, S. (1999) Separation of “glycosphingolipid signaling domain” from caveolin-containing membrane fraction in mouse melanoma B16 cells and its role in cell adhesion coupled with signaling. J. Biol. Chem., 273, 33766–33773. Kaplowitz, P.B., and Haar, J.L. (1988) Antimitogenic actions of lectins in cultured human fibroblasts. J. Cell Physiol., 136, 13–22. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 227, 680–685. Laine, R.A. and Hakomori, S.I. (1973) Incorporation of exogenous glycosphingolipids in plasma membranes of cultured hamster cells and concurrent change of growth behavior. Biochem. Biophys. Res. Commun., 54, 1039–1045. Lax, I., Bellot, F., Honegger, A.M., Schmidt, A., Ullrich, A., Givol, D., and Schlessinger, J. (1990) Domain deletion in the extracellular portion of the EGF-receptor reduces ligand binding and impairs cell surface expression. Cell Regul., 1, 173–188. Li, W., Hack, N., Margolis, B., Ullrich, A., Skorecki, K., and Schlessinger, J. (1991) Carboxy-terminal truncations of epidermal growth factor (EGF) receptor affect diverse EGF-induced cellular responses. Cell Regul., 2, 641–649. Massoglia, S., Gray, A., Dull, T.J., Munemitsu, S., Kun, H.J., Schlessinger, J., and Ullrich, A. (1990) Epidermal growth factor receptor cytoplasmic domain mutations trigger ligand-independent transformation. Mol. Cell Biol., 10, 3048–3055. Meuillet, E.J., Kroes, R., Yamamoto, H., Warner, T.G., Ferrari, J., Mania-Farnell, B., George, D., Rebbaa, A., Moskal, J.R., and Bremer, E.G. (1999) Sialidase gene transfection enhances epidermal growth factor receptor activity in an epidermoid carcinoma cell line, A431. Cancer Res., 59, 234–240. Mineo, C., Gill, G.N., and Anderson, R.G. (1999) Regulated migration of epidermal growth factor receptor from caveolae. J. Biol. Chem., 274, 30636–30643. Moseley, J.M., and Suva, L.J. (1986) Molecular characterization of the EGF receptor and involvement of glycosyl moieties in the binding of EGF to its receptor on a clonal osteosarcoma cell line, UMR 106-06. Calcif. Tissue Int., 38, 109–114. Paller, A.S., Arnsmeier, S.L., Robinson, J.K., and Bremer, E.G. (1992) Alteration in keratinocyte ganglioside content in basal cell carcinomas. J. Invest. Dermatol., 98, 226–232. Paller, A.S., Arnsmeier, S.L., Alvarez-Franco, M., and Bremer, E.G. (1993) Ganglioside GM3 inhibits the proliferation of cultured keratinocytes. J. Invest. Dermatol., 100, 841–845. Paller, A.S., Arnsmeier, S.L., Bryk, D., and Bremer, E.G. (1995) Inhibition of phosphorylation of the epidermal growth factor receptor by ganglioside GM3. J. Invest. Dermatol., 194, 592. Pratt, R.M., and Pastan, I. (1978) Decreased binding of epidermal growth factor to BALB/c 3T3 mutant cells defective in glycoprotein synthesis. Nature (Lond.), 272, 68–70.

522

Rebbaa, A., Hurh, J., Yamamoto, H., Kersey, D.S., and Bremer EG. (1996) Ganglioside GM3 inhibition of EGF receptor mediated signal transduction. Glycobiology, 6, 399–406. Soderquist, A.M., and Carpenter, G. (1984) Glycosylation of the epidermal growth factor receptor in A-431 cells. The contribution of carbohydrate to receptor function. J. Biol. Chem., 259, 12586–12594. Stroop, C.J.M., Weber, W., Gerwig, G.J., Nimtz, M., Kamerling, J.P., and Vliegenthart, J.F.G. (2000) Characterization of the carbohydrate chains of the secreted form of the human epidermal growth factor receptor. Glycobiology, 10, 901–917. Sung C., O’Toole, E.T., Lannutti, B., Hunt, G., O’Gorman, M., Woodley, D.T., and Paller, A.S. (1998) Integrin α5β1 expression is required for inhibition of keratinocyte migration by ganglioside GT1b. Exp. Cell Res., 239, 311–319. Takahashi, Y., Akanuma, Y., Yazaki, Y., and Kadowaki, T. (1999) Formation of distinct signalling complexes involving phosphatidylinositol 3-kinase activity with stimulation of epidermal growth factor or insulin-like growth factor-I in human skin fibroblasts. J. Cell Physiol., 178, 69–75. Taniguchi, N., Yoshimura, M., Miyoshi, E., Ihara, Y., Nishikawa, A., and Fujii, S. (1996) Remodeling of cell surface glycoproteins by N-acetylglucosaminyltransferase III gene transfection: modulation of metastatic potentials and down regulation of hepatitis B virus replication. Glycobiology, 6, 691–694. Tsuda, T., Ikeda, Y., and Taniguchi, N. (2000) The Asn-420-linked sugar chain in human epidermal growth factor receptor suppresses ligand-independent spontaneous oligomerization. J. Biol. Chem., 275, 21988–21994. Ullrich, A., Coussens, L., Hayflick, J.S., Dull, T.J., Gray, A., Tam, A.W., Lee, J., Yarden, Y., Libermann, T.A., Schlessinger, J., and others. (1984) Human epidermal growth factor receptor cDNA sequence and aberrant expression of the amplified gene in A431 epidermoid carcinoma cells. Nature, 309, 418–425. Ullrich, A., and Schlessinger, J. (1990) Signal transduction by receptors with tyrosine kinase activity. Cell, 61, 203–212. Wakshull, E.M., and Wharton, W. (1985) Stabilized complexes of epidermal growth factor and its receptor on the cell surface stimulate RNA synthesis but not mitogenesis. Proc. Natl Acad. Sci. USA, 82, 8513–8517. Wang, X., Rahman, Z., Sun, P., Meuillet, E., George, D., Bremer, E.G., and Paller, A.S. (2001a) Ganglioside modulates epidermal growth factor receptor action by altering ligand binding. J. Invest. Dermatol., 116, 69–76. Wang, X., Sun, P., Al-Qamari, A., Tai, T., Kawashima, I., and Paller, A.S. (2001b) Carbohydrate-carbohydrate binding of ganglioside to integrin α5 modulates α5β1 function. J. Biol. Chem., 276, 8436–8444. Wong, L., Deb, T.B., Thompson, S.A., Wells, A., and Johnson, G.R. (1999) A differential requirement for the COOH-terminal region of the epidermal growth factor (EGF) receptor in amphiregulin and EGF mitogenic signaling. J. Biol. Chem., 274, 8900–8909. Yednak, M.A., and Bremer, E.G. (1994) Preferential binding of the epidermal growth factor receptor to ganglioside GM3 coated plates. Mol. Chem. Neuropathol., 21, 369–378. Zheng, M., Fang, H., and Hakomori, S.I. (1994) Functional role of N-glycosylation in alpha 5 beta 1 integrin receptor. De-N-glycosylation induces dissociation or altered association of alpha 5 and beta 1 subunits and concomitant loss of fibronectin binding activity. J. Biol. Chem., 269, 12325–12331.