Membrane distribution of epidermal growth factor receptors in cells expressing different gangliosides Adolfo R. Zurita 1 , Pilar M. Crespo 1 , Nicola ´ s P. Koritschoner 2 and Jose ´ L. Daniotti 1 1 Centro de Investigaciones en Quı ´ mica Biolo ´ gica de Co ´ rdoba, CIQUIBIC (UNC-CONICET), Departamento de Quı ´ mica Biolo ´ gica, Facultad de Ciencias Quı ´ micas, Universidad Nacional de Co ´ rdoba, Argentina; 2 Departamento de Bioquı ´ mica Clı ´ nica, Facultad de Ciencias Quı ´ micas, Universidad Nacional de Co ´ rdoba, Argentina Gangliosides have been found to reside in glycosphingolipid- enriched microdomains (GEM) of the plasma membrane and to be involved in the regulation of epidermal growth factor receptor (EGFr or ErbB1) activity. To gain further insight into the mechanisms involved in EGFr modulation by gangliosides, we investigated the distribution of EGFr family members in the plasma membrane of CHO-K1 cells, which were genetically modified to express different gan- glioside molecules or depleted of glycolipids. Our data demonstrate that at least four different sets of endogenously expressed gangliosides, including GD3, did not have a sig- nificant effect on EGFr distribution in the plasma mem- brane. In addition, using confocal microscopy analysis we clearly demonstrated that the EGFr co-localizes only to a minor extent with GD3. We also explored the endogenous expression, in wild-type CHO-K1 cells, of the orphan receptor ErbB2 (which is the preferred heteroassociation partner of all other ErbB proteins) and the effect of GD3 expression on its membrane distribution. Our results showed that CHO-K1 cells endogenously express ErbB2 and that expression of the GD3 affected, to some extent, the membrane distribution of endogenous ErbB2. Finally, our findings support the notion that most EGFr are excluded from GEM, while an important fraction of ErbB2 is found to be associated with these microdomains in membranes from CHO-K1 cells. Keywords: gangliosides; EGFr; ErbB2; membrane lipid domains; CHO-K1 cells. 1 Gangliosides – glycolipids containing sialic acid – are ubiquitous components of mammalian cell membranes. They are involved in the regulation of cell proliferation and differentiation [1]. They have also been implicated in tumour growth and the formation of metastases. All tumours exhibit aberrant ganglioside expression. This includes overexpression of normal ganglioside constituents and expression of gangliosides not found in normal adult tissue [2]. On the basis that the bulk of gangliosides present in the cell are plasma membrane bound, it has been speculated that they participate in cell-surface events such as modula- tion of tyrosine kinase growth factor receptors. In this sense, it has been demonstrated, mainly by the exogenous addition of gangliosides or by changes of the endogenous content, that gangliosides regulate the activities of the TrkA receptor [3,4], the insulin receptor [5,6], the epidermal growth factor receptor (EGFr, ErbB1) [7] and the platelet- derived growth factor (PDGF) receptor [8,9]. In previous studies, we described the modulation of EGFr phosphorylation by endogenously expressed ganglio- sides [10]. In particular, we observed an inhibition of EGFr autophosphorylation when CHO-K1 cells (GM3 + )were induced to express the disialoganglioside, GD3, by stable transfection of CMP-NeuAc:GM3 sialyltransferase (Sial- T2). The effect of GD3 on EGFr function could not be attributed to an alteration in the affinity of EGFr to EGF (because the K d values were essentially the same as in the wild-type cells) or to a decrease of GM3 content by conversion into GD3 ganglioside. Gangliosides are constituents of the glycosphingolipid- enriched microdomains [GEM; also called DRMs (deter- gent-resistant membranes) or rafts], dynamic assemblies of cholesterol, saturated phospholipids, and sphingolipids, which are characterized by a light buoyant density and resistance to solubilization by Triton X-100 at 4 °C [11–14]. Initial evidence showed the association of gangliosides to GEM present in the plasma membrane of CHO-K1 cells lines expressing different gangliosides [15,16]. These data Correspondence to J. L. Daniotti, CIQUIBIC (UNC-CONICET), Departamento de Quı ´ mica Biolo ´ gica, Facultad de Ciencias Quı ´ micas, Universidad Nacional de Co ´ rdoba, Ciudad Universitaria, 5000 Co ´ rdoba, Argentina. Fax: + 54 351 4334074, Tel.: + 54 351 4334171, E-mail: daniotti@dqb.fcq.unc.edu.ar Abbreviations:BS 3 , bis(sulfosuccinimidil)suberato; Cer, ceramide; CFP, cyan fluorescence protein; EGF, epidermal growth factor; EGFr, EGF receptor; ErbB1, epidermal growth factor receptor 1; ErbB2, epidermal growth factor receptor 2; GEM, glycosphingolipid- enriched microdomain(s); GFP, green fluorescent protein; Gal, galactose; GalNAc-T, UDP-GalNAc:LacCer/G3/GD3 N-acetyl- galactosaminyltransferase; GlcCer, glucosylceramide; GPI, glycosyl- phosphatidylinositol; HPTLC, high-performance thin layer chromatography; LacCer, lactosylceramide; PPPP, D , L -threo-1- phenyl-2-hexadecanoylamino-3-pyrrolidino-1-propanol-HCl; Sial-T2, CMP-NeuAc:GM3 sialyltransferase; VSVG, vesicular stomatitis virus glycoprotein; YFP, yellow fluorescence protein. (Received 3 March 2004, revised 13 April 2004, accepted 16 April 2004) Eur. J. Biochem. 271, 2428–2437 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04165.x revealed that an important fraction of plasma membrane, GM3, and most GD3 and GT3, resided in GEM, while more complex gangliosides, such as GM1, GM2 and GD1a, were almost excluded from GEM. To gain insight into the mechanisms involved in EGFr modulation by gangliosides (particularly GD3), we inves- tigated the distribution of human EGFr in the plasma membrane of CHO-K1 cells that were genetically modified to express different ganglioside molecules or depleted of glycolipids by inhibiting glucosylceramide synthase activity. Our data demonstrated that at least four different sets of endogenously expressed gangliosides, including GD3, did not have a significant effect on EGFr distribution in the plasma membrane. In addition, using confocal microscopy analysis we clearly demonstrated that EGFr co-localizes only to a minor extent with GD3. We also explored the endogenous expression, in wild-type CHO-K1 cells, of the orphan receptor epidermal growth factor receptor 2 (ErbB2), which is the preferred hetero- association partner of all other ErbB proteins, and the effect of GD3 expression on ErbB2 membrane distribution. Our results show that CHO-K1 cells endogenously express the receptor ErbB2 and that expression of the GD3 affected, to some extent, the membrane distribution of endogenous ErbB2. Finally, our data support the notion that most EGFr is excluded from GEM, while an important fraction of ErbB2 is found to be associated with these microdomains in membranes from CHO-K1 cells. Materials and methods Cell lines and DNA transfections The following CHO-K1 cell clones, expressing different ganglioside glycosyltransferases, were used: wild-type CHO- K1 cells (ATCC); clone 2 (formerly clone 18 [10]), a stable Sial-T2 transfectant expressing the ganglioside GD3 [17]; clone 3 (formerly clone 7 [10]), a stable UDP-GalNAc:Lac- Cer/G3/GD3 N-acetyl-galactosaminyltransferase (Gal- NAc-T) transfectant mostly expressing gangliosides GM3 andGM2and,toalesserextent,GM1[18]. Inhibition of glycolipid synthesis with D , L -threo-1-phenyl- 2-hexadecanoylamino-3-pyrrolidino-1-propanol-HCl (PPPP; Matreya Inc. 2 , Pleasant Gap, PA, USA) was carried out as previously described [18]. Wild-type CHO-K1 cells were treated for 5 days with 2 l M PPPP. Inhibition of glycolipid synthesis was monitored through the analysis of cellular ganglioside content by HPTLC (high-performance thin layer chromatography). Cells were maintained at 37 °C/5% CO 2 in DMEM (Dulbecco’s modified Eagle’s minimal essential medium) supplemented with 10% (v/v) fetal bovine serum and antibiotics. Lipofectamine (Gibco-BRL) was used to tran- siently transfect CHO-K1 cells with the following constructs (1 lg per dish) carrying cDNAs coding for the total sequence of the yellow fluorescence protein (YFP) fused to a glycosylphosphatidylinositol (GPI)-attachment signal (GPI-YFP), or for the vesicular stomatitis virus glycopro- tein (VSVG) fused to the cyan fluorescence protein (VSVG-CFP), or human EGFr [10]. Fifteen hours after transfection, cells were washed with cold phosphate- buffered saline (NaCl/P i ) and harvested with 10 m M Tris/HCl (pH 7.2), containing 0.25 M sucrose, for Western blot assays, or treated with lysis buffer as indicated below. Membrane extraction using Triton X-100 Cells were washed with cold NaCl/P i and harvested by scraping. Samples were treated with 0.5 mL of lysis buffer containing 150 m M NaCl, 5 m M EDTA, 1% Triton X-100, 0.1 M Na 2 CO 3 ,5lgÆmL )1 aprotinin, 0.5 lgÆmL )1 leupep- tin, 0.7 lgÆmL )1 pepstatin and 25 m M Tris/HCl, pH 7.5 (TNE lysis buffer) at 4 °C for 1 h, and then centrifuged for 1 h at 100 000 g at 4 °C. The supernatant (soluble fraction) was removed, and the pellet (insoluble fraction) was resuspended in 0.5 mL of lysis buffer. Proteins from soluble and insoluble fractions were precipitated with chloroform/ methanol (1 : 4, v/v) and subjected to SDS/PAGE and Western blot. Sucrose density-gradient separation Cells were lysed in 0.2 mL of TNE lysis buffer containing various concentrations (0.25, 0.5 or 1%) of Triton-X-100, andlefttostandat4°C for 1 h. Lysates were centrifuged (10 h, 150 000 g,4°C) on continuous sucrose gradients (5–30%) in TNE buffer without Triton X-100. Twelve fractions were collected from the bottom of the tube using a fraction collector. Proteins in each fraction were precipitated with 10% (v/v) trichloroacetic acid, resuspended and subjected to SDS/PAGE and Western blot. Chemical cross-linking The procedure of cross-linking was essentially applied as described previously [19]. Cells were washed twice with cold NaCl/P i andincubatedat4°C with 0.5 m M bis(sulfosuc- cinimidil)suberato (BS 3 ; Pierce Chemical Co.) for 45 min. Cross-linking was quenched by the addition of 50 m M glycine for 15 min at 4 °C. Cells were washed with NaCl/ P i , collected and pelleted by centrifugation for 5 min at 9000 g. Proteins from pellets were resolved by electrophoresis through SDS/PAGE gels under reducing conditions. Experiments of cross-linking with BS 3 were also carried out after stimulation with 100 n M EGF at 4 °Cfor1h. Subsequent to extensive washing with NaCl/P i , the cross- linking procedure was followed as indicated above [20,21]. Western blot Electrophoresis and transfer was carried out as previously described [10]. Membranes were blocked with 5% nonfat dry milk or with 2.5% BSA/2.5% polyvinyl pyrrolidone 40, in TBS (200 m M NaCl, 100 m M Tris/HCl pH 7.5), depend- ing on the antibody. Anti-(green fluorescent protein) (GFP) polyclonal Ig (Roche Molecular Biochemicals), anti-ErbB2 (Dako), anti-EGFr and anti-p53 (all from Santa Cruz Biotechnology) were used at a dilution of 1 : 800, 1 : 500, 1 : 200 and 1 : 500, respectively. Anti-(phospho-EGFr) Ig (Cell Signaling Technology, Inc. 3 , Beverly, MA, USA), recognizing phosphorylated tyrosines (Tyr845, 992 and 1068), was used at a dilution of 1 : 1000, following the instructions of the manufacturer. Bands were detected by Ó FEBS 2004 EGF receptors distribution in cell membranes (Eur. J. Biochem. 271) 2429 protein A or appropriate secondary antibodies coupled to horseradish peroxidase combined with the chemilumines- cence detection kit (Western lightning; PerkinElmer Life Sciences) and Kodak Biomax MS films. The molecular weights were calculated based on calibrated standards (Gibco-BRL) run in every gel. The relative contribution of individual bands was calculated using the computer soft- ware SCION IMAGE on scanned films of low-exposure images. Immunofluorescence microscopy CellsgrownincoverslipswerewashedinNaCl/P i ,fixedin acetone at )20 °C for 7 min, and incubated in NaCl/P i containing 3% (w/v) BSA 4 for 1 h at 37 °Ctoblock nonspecific binding sites. Coverslips were then incubated overnight at 4 °C with primary antibodies, washed five times with NaCl/P i buffer, and exposed to secondary antibodies for 2 h at 37 °C. The primary antibodies were mouse monoclonal anti-GD3 (IgG3), clone R24 (a gift of K. Lloyd, Memorial Sloan Kettering Cancer Research Center,NewYork,NY,USA),diluted1:200andrabbit polyclonal anti-EGFr (Santa Cruz Biotechnology), diluted 1 : 150. Secondary antibodies were Alexa 488-conjugated goat anti-mouse (Santa Cruz Biotechnology), diluted 1 : 1000, or rhodamine-conjugated donkey anti-rabbit (Jackson ImmunoResearch), diluted 1 : 500. After final washes with NaCl/P i , cells were mounted in mounting fluid (Light Diagnostics 5 , Temecula, CA, USA). To explore GD3 and EGFr expression in nonpermeabilized cells, R24 antibody (mouse IgG3), and a specific anti-EGFr antibody (mouse IgG2b, R1 antibody from Santa Cruz Biotechno- logy), recognizing its extracellular domain, were used. Cells from clone 2 (GD3 + ), transiently expressing human EGFr, were fixed with 3% formaldehyde for 30 min at 4 °C. R24 antibody was used at a dilution of 1 : 50, while R1 antibody was used at a dilution of 1 : 100. After overnight incubation at 4 °C and extensive washing with NaCl/P i , cells were incubated with goat polyclonal anti-(mouse IgG3) (Sigma- Aldrich), at a dilution of 1 : 200, for 90 min at 37 °C. Finally, coverslips were incubated with Alexa 546-conju- gated donkey anti-goat Ig (Molecular Probes) and fluoresc- ein-conjugated rat anti-IgG2b (Pharmingen) at a dilution of 1 : 800 or 1 : 700, respectively, for 90 min at 37 °C. Appropriate controls were included to guarantee the specificity of all antibodies used. Confocal images were collected using a Zeiss LSM5 Pascal laser-scanning confocal microscope equipped with an argon/helium/neon laser and an X63 1.4 NA oil-immersion objective (Zeiss Plan-Apochromat). Single confocal sections of 0.3 lm were taken parallel to the coverslip (xy-sections). Images were acquired using a Zeiss CCD camera and processed with the LSM software and ADOBE PHOTOSHOP . Results CHO-K1 cell lines To closely simulate endogenous shifts in ganglioside expression, we have recently established CHO-K1 cell clones that are able to change the glycolipid composition of the plasma membrane by altering the ganglioside biosynthetic activity of the cell, while maintaining the normal process of intracellular transport and membrane insertion [10,15,22]. Using the panel of genetically engin- eered CHO-K1 cell clones, we explored the modulation of EGFr phosphorylation in the different glycolipid environ- ments [10]. A scheme of glycolipid biosynthesis is shown in Fig. 1. It is appreciable how the pathways of ganglioside synthesis are branched by transfection of Sial-T2 (GD3 synthase, clone 2) or GalNAc-T (GM2 synthase, clone 3) to the wild-type CHO-K1 cells (CHO-K1 wt). Wild-type CHO-K1 cells predominantly express the ganglioside GM3, while those Fig. 1. Glycolipid labelling of CHO-K1 cell clones. Aschematicrep- resentation of the pathway of glycolipid biosynthesis is shown at the top of the figure. It is appreciable how the pathways of ganglioside synthesis are branched following transfection of CMP-NeuAc:GM3 sialyltransferase (Sial-T2) (GD3 synthase, clone 2) or UDP-Gal- NAc:LacCer/G3/GD3 N-acetylgalactosaminyltransferase (GalNAc- T) (GM2 synthase, clone 3) to wild-type CHO-K1 cells (CHO-K1 wt) expressing only the ganglioside, GM3. Also indicated in the scheme is the enzymatic reaction affected by the glycolipid inhibitor, D , L -threo-1- phenyl-2-hexadecanoylamino-3-pyrrolidino-1-propanol-HCl (PPPP). Wild-type CHO-K1 cells (CHO-K1 wt), cells from clones 2 and 3, and PPPP-treated wild type CHO-K1 cells (CHO-K1/PPPP) were meta- bolically labelled with 2 lCi of [ 14 C]Gal for 12 h. Lipids were purified and chromatographed by HPTLC, as previously described [10]. The positions of co-chromatographed glycolipid standards are indicated. 2430 A. R. Zurita et al. (Eur. J. Biochem. 271) Ó FEBS 2004 stably expressing the Sial-T2 cDNA (clone 2) [17] synthesize mostly GD3 and GT3, accumulate LacCer and show very little accumulation of GM3 (Fig. 1). CHO-K1 cells stably expressing the GalNAc-T cDNA (clone 3) [18], synthesize the a-series ganglioside GM2 and, to a lesser extent, GM1 because of the constitutive endogenous expression in these cells of the enzyme involved in the synthesis of GM1 (Fig. 1) [23]. To reduce the content of all glycosphingolipid classes, wild-type CHO-K1 cells were treated with PPPP, a competitive inhibitor of ceramide glucosyltransferase and hence of the synthesis of complex glycolipids (Fig. 1, CHO- K1/PPPP) [24]. Exposure of cells to 2 l M PPPP in the culture medium for 5 days led to a 95% decrease of GM3 content with respect to control cells [10,18]. EGFr membrane distribution in CHO-K1 cell lines expressing different gangliosides Wild-type CHO-K1 cells, and cells from clones 2 and 3, and wild-type CHO-K1 cells with a generalized decrease in glycolipid expression (CHO-K1/PPPP) 6 ,alltransiently expressing human EGFr, were treated with 1% Triton- X-100 at 4 °C and lysates were subjected to continuous sucrose gradient ultracentrifugation, fractionation, and detection of EGFr and known protein markers by Western blotting. Under these conditions, proteins and lipids resist- ant to 1% (v/v) Triton X-100 extraction, flow at low-density fractions. As controls, we analysed the behaviour, to Triton X-100 extraction, of a GEM marker (i.e. a fusion protein containing a GPI-anchored signal, GPI-YFP), and a non- GEM marker (i.e. VSVG-CFP). As previously described, the GEM marker, GPI-YFP, was highly concentrated in low-density fractions [15,25]. In contrast, VSVG-CFP (the non-GEM marker) was distributed in higher density fractions, thus distributing essentially as Triton X-100 soluble proteins (Fig. 2A). EGFr was mainly concentrated in higher density fractions with essentially the same distribution pattern in all cell lines, co-distributing with the Triton X-100 soluble protein, VSVG-CFP (Fig. 2A). Two nonspecific lower bands were detected with the anti- EGFr Ig, even in extracts from CHO-K1 cells that were not transfected with EGFr (Fig. 3C). It has been described that integrin receptors (e.g. alpha3, alpha5) are found completely outside GEM after treatment with 1% (v/v) Triton X-100, but almost exclusively in GEM when a lower concentration (0.25–0.5%) of Triton X-100 is used [26], suggesting a relatively weak interaction with membrane components. Taking into account that growth factor receptors, including EGFr, and integrin receptors are functionally associated [27,28], we explored the behaviour of EGFr to lower Triton X-100 concentrations. Basically, and in contrast to the behaviour of integrin receptors, we found almost the same gradient distribution of EGFr at 0.25, 0.5 and 1% (v/v) Triton X-100 (results not shown). A representative pattern of protein distribution is shown in Fig. 2B. Results from this experiment strongly suggest that changes in the composition of endogenous gangliosides did not affect the distribution of EGFr on membranes of CHO-K1 cell clones. Next, we set out to confirm these results, studying the solubility/insolubility of EGFr to extraction with cold Triton X-100 by velocity sedimentation. Cell homogenates from wild-type CHO-K1 cells, transiently expressing EGFr, were extracted with Triton X-100 at 4 °Candthen Fig. 2. Continuous sucrose gradient analysis of epidermal growth factor receptor (EGFr) in CHO-K1 cell lines. (A) CHO-K1 cell clones tran- siently expressing human EGFr, glycosylphosphatidylinositol-yellow fluorescence protein (GPI-YFP), or vesicular stomatitis virus glyco- protein-cyan fluorescence protein (VSVG-CFP) were lysed in lysis buffer at 4 °C for 1 h and centrifuged (10 h, 150 000 g,4°C) on continuous sucrose gradients (5–30%). Twelve fractions were collected from the bottom of the sucrose density gradient using a fraction col- lector. Proteins were resolved by electrophoresis through 8% (for EGFr analysis) or 10% (for GPI-YFP and VSVG-CFP) SDS–PAGE and analysed by Western blot. The antibodies used were anti-EGFr andanti-GFPtorevealGPI-YFPandVSVG-CFP,respectively.The positions (molecular masses) of recombinant proteins are indicated. (B) Protein profile of the gradient, visualized by Ponceau S staining of the nitrocellulose membrane. Ó FEBS 2004 EGF receptors distribution in cell membranes (Eur. J. Biochem. 271) 2431 centrifuged (1 h, 100 000 g,4°C). Proteins from soluble and insoluble fractions were detected by Western blot with the appropriate antibody. EGFr was totally soluble to Triton X-100 extraction (Fig. 3A). As expected, the GEM marker (GPI-YFP) was 68% insoluble, whereas the non- GEM marker (VSVG-CFP) was less than 20% insoluble (Fig. 3A). Finally, we explored the grade of association of EGFr on the cell surface of wild-type CHO-K1 cells by using the chemical cross-linking agent, BS 3 , which is membrane impermeable and possesses a spacer arm of 1.14 nm. Wild-type CHO-K1 cells, transiently expressing EGFr and subjected to BS 3 , showed a low (less than 5%) cross-linking efficiency [calculated as the percentage of cross-linked molecules (dimer plus oligomer) with respect to total protein], as expected for a homogenous distribution of a non-GEM protein marker (compare with the VSVG-CFP cross-linking efficiency of 10%; Fig. 3B). On the other hand, GPI-CFP, a GEM protein marker, showed a cross- linking efficiency of > 40% 7 , represented by dimer and oligomer species [15]. As a control, and to demonstrate that BS 3 acts on active EGFr, CHO-K1 cells expressing the receptor were incubated with buffer, or with 100 n M EGF, for 1 h at 4 °C, treated with the cross-linking agent BS 3 , lysed and analysed by SDS/PAGE. It was found that EGF was able to stimulate the appearance of EGFr homodimer (Fig. 3C). Additionally, the EGFr homodimer was absent when the receptor was not stimulated with EGF. That EGFr has a homogeneous distribution (a scarce association with microdomains) in membranes from CHO-K1 cells is consistent with the behaviour of the EGFr in both sucrose gradient centrifugation and solubilization by nonionic detergent, indicated above. Although we show results from solubility/insolubility and cross-linking experiments only for wild type CHO-K1 cells, the behaviour of EGFr in all other clones was essentially the same (results not shown). Tyrosine-phosphorylated EGFr membrane distribution after stimulation with EGF Wild-type CHO-K1 cells, and cells from clone 2 (Fig. 1), were transfected to transiently express EGFr and cultured in the absence of fetal bovine serum for 12 h. Then, the cells were incubated with EGF (100 n M )for5min,lysedwith cold Triton X-100, centrifuged in a sucrose gradient and the activated EGF receptor was analysed, in all fractions, by Western blot using antibodies recognizing phosphorylated EGFr on tyrosines located at positions 845, 992 and 1068. The activated EGF receptor 8 was found to be distributed similarly in sucrose gradients from both wild-type CHO-K1 cells (GM3 + ) and cells from clone 2 (GD3 + ) (Fig. 4). It should be emphasized that in these experiments the total amount of protein and EGFr expression levels in both cell lines were not necessarily comparable. The membrane distribution of the active EGFr fits well with that of the total Fig. 3. Detergent solubility and membrane distribution of epidermal growth factor receptor (EGFr), glycosphingolipid-enriched microdomain (GEM) and non-GEM markers. (A) Homogenates from wild-type CHO-K1 cells transiently expressing human EGFr, glycosylphospha- tidylinositol-yellow fluorescence protein (GPI-YFP) (GEM marker) or vesicular stomatitis virus glycoprotein-cyan fluorescence protein (VSVG-CFP) (non-GEM marker) were extracted with Triton X-100 at 4 °C and then ultracentrifuged (1 h, 100 000 g,4°C). Proteins from soluble (S) and insoluble (I) fractions were resolved by SDS/PAGE and probed with the appropriate antibody. (B) Detection of protein clusters by chemical cross-linking with bis(sulfosuccinimidil)suberato (BS 3 ) in membranes from CHO-K1 cells. Wild-type CHO-K1 cells, transiently expressing EGFr, GPI-YFP and VSVG-CFP, were sub- jected to cross-linking with 0.5 m M BS 3 . Protein extracts were resolved in SDS/PAGE and detected by Western blot. (C) Chemical cross- linking of EGFr after EGF stimulation. To demonstrate that BS 3 acts on active EGFr, CHO-K1 cells expressing the receptor (lines 2–4) were incubated with buffer (lanes 2 and 3) or with 100 n M EGF (lane 4). Then, samples were cross-linked with BS 3 (lanes 3 and 4) or incubated with buffer alone (line 2). As a control, an extract from mock-trans- fected cells was run in lane 1. Protein extracts were resolved in SDS– PAGE and detected by Western blot. Positions of monomers (m), dimers (d) or oligomers (o) are indicated. 2432 A. R. Zurita et al. (Eur. J. Biochem. 271) Ó FEBS 2004 EGFr (Fig. 2A), suggesting that there were no changes of membrane distribution associated with its activation status. EGFr and GD3 localization in CHO-K1 cell membranes Having demonstrated a converse segregation of GD3 and EGFr, by biochemical means, in membranes from clone 2 cells (this work) [15,16], we next investigated the spatial localization of GD3 and EGFr in both its active and inactive state. Studies were carried out by confocal micros- copy immunofluorescence, using the monoclonal antibody, R24, to detect GD3 and two anti-EGFr Igs that recognize the extracellular or intracellular domains of the receptor. GD3 was mainly detected in the plasma membrane, showing a distribution in patches, while EGFr was observed both in plasma membranes and intracellular membranes (Fig. 5). A comparison between GD3 and EGFr membrane distribution revealed a minor separation (Fig. 5A–D). To better define the localization of GD3 and EGFr in the plasma membrane of CHO-K1 cells in the absence of EGF, confocal analysis was performed using formaldehyde-fixed cells (nonpermeabilized cells) labelled with an anti-EGFr Ig recognizing the extracellular domain. In addition, to ease the separation between GEM and other parts of the plasma membrane, the focal plane was mainly adjusted through the top of the cell, allowing the visualization of larger plasma membrane areas and the identification of small membrane domains (Fig. 5E–H). These data clearly confirmed that EGFr and GD3 are colocalized, to some extent, on the plasma membrane of CHO-K1 cells. Interestingly, when the cells were stimulated with EGF for 10 min, a clear endocytosis of the EGFr was observed but, under this condition, GD3 remains at the cell surface (Fig. 5I–L). Altogether, these results showed that GD3 is mainly expressed on the plasma membrane of cells from clone 2, and that GD3 and EGFr co-localized, to some extent, only in the absence of EGF stimulation, while, upon addition of EGF, a clear separation of these two membrane components was observed. Endogenous ErbB2 membrane distribution in wild-type CHO-K1 cells (GM3 + ) and cells from clone 2 (GD3 + ) Next, we studied how other members of the ErbB family would behave in terms of membrane distribution, as shown previously for EGFr (ErbB1). To achieve this, we investi- gated the endogenous expression, in CHO-K1 cells, of the orphan receptor, ErbB2, the preferred heteroassociation partner of all other ErbB proteins. ErbB2 expression was analysed by Western blot with an antibody directed to its intracellular domain, both in wild-type and human EGFr- expressing CHO-K1 cells. First, we analysed the heterolo- gous expression of EGFr in CHO-K1 cells. As expected, EGFr is expressed in CHO-K1 cells as a functional protein of 170 kDa [10]. Endogenous EGFr expression in CHO-K1 cells was below the limit of detection (Fig. 6A, upper panel). Nonspecific lower bands were also detected with the anti- EGFr Ig (see also Fig. 3C). Additionally, ErbB2 was detected as a band of 185 kDa in both wild-type and EGFr- expressing CHO-K1 cells (Fig. 6A, upper panel). As control of protein loading, we analysed the constitutive expression of p53. No substantial differences were observed in any of the lanes analysed (Fig. 6A, lower panel). Next, we inves- tigated the membrane distribution of endogenous ErbB2 by sucrose gradient both in wild-type CHO-K1 cells (GM3 + ) and in cells from clone 2 (GD3 + ). As also observed for EGFr (Fig. 2A), there was no difference in the distribution pattern of ErbB2 at high-density fractions (Triton-X-100- soluble proteins) in wild-type CHO-K1 cells (GM3 + )and cells from clone 2 (GD3 + ) (Fig. 6B,C). However, it should be noted that in both cell lines, a significant fraction of ErbB2 (27%) was associated with low-density fractions (fractions 7–12, see the distribution of GEM and non-GEM markers in Fig. 2A), which probably represent a fraction associated with GEM. On these fractions, a small shift in Fig. 4. Tyrosine-phosphorylated epidermal growth factor receptor (EGFr) membrane distribution after EGF stimulation. Wild-type CHO-K1 cells (CHO-K1 wt, GM3 + ) and cells from clone 2 (GD3 + ), transiently expressing human EGFr, were maintained in serum-free medium for 12 h. EGF (100 n M ) was added to the medium and, after 10 min, cells were lysed and centrifuged (10 h, 150 000 g,4°C) on continuous sucrose gradients (5–30%). Twelve fractions were collected from the bottom of the sucrose density gradient. Proteins were resolved by electrophoresis through 8% SDS–PAGE and probed with anti-(phospho-EGFr) (P-EGFr) Ig, which recognizes phosphorylated tyrosines (Tyr845, 992 and 1068), to detect the active receptor. Ó FEBS 2004 EGF receptors distribution in cell membranes (Eur. J. Biochem. 271) 2433 ErbB2 distribution was observed at the top of the gradient. While the ErbB2 receptor from wild-type CHO-K1 cells (CHO-K1 wt) was found mostly in fraction 12, the ErbB2 receptor from clone 2 cells was concentrated mainly in fraction 11. In line with the results of ErbB2 distribution in sucrose gradients, solubility/insolubility analysis of ErbB2 to cold Triton X-100 extraction by velocity sedi- mentation showed that 15% of the receptor was insoluble (data not shown). Discussion The main goal of this work was to investigate the possibility that changes in the expression of gangliosides could modulate the membrane distribution of EGFr members and thereby regulate their function. By studying the solubility/insolubility of EGFr to extraction using nonionic detergents, we demonstrated that changes in the composi- tion of endogenous gangliosides did not significantly affect the distribution of EGFr on the membranes of CHO-K1 cell lines. In all clones analysed, this receptor behaved as a soluble molecule to extraction with cold Triton X-100, indicating that it is mainly excluded from GEM. Interest- ingly, the behaviour of the receptor to extraction with cold Triton X-100 was independent of its activation state because binding of EGF to EGFr did not affect its membrane distribution. The analysis of EGFr clusterization on the plasma membrane of CHO-K1 cells, by using a chemical cross-linking approach, is compatible with the notion of a homogeneous membrane distribution of the growth factor receptor. Our initial evidence demonstrated the behaviour of different gangliosides expressed in CHO-K1 cells to extrac- tion with cold Triton X-100 in order to investigate their association with GEM. The results revealed that the majority of plasma membrane GD3 reside in GEM. Confirming previous work, we found that after extraction with nonionic detergent, GD3 was associated with low Fig. 5. Epidermal growth factor receptor (EGFr) and GD3 localization in CHO-K1 cell membranes. Cells from clone 2 (GD3 + )weretransiently transfected to express human EGFr and maintained in serum-free medium for 12 h before incubation for 10 min at 37 °C in the absence (A–H) or presence (I–L) of 100 n M EGF in the cell culture medium. EGFr expression was analysed by confocal microscopy immunofluorescence in acetone- fixed (A–D, I–L) or formaldehyde-fixed (E–H) cells using a polyclonal anti-(intracellular domain) Ig (A and I) or a monoclonal (mouse IgG2) anti- (extracellular domain) Ig (E) of EGFr, and rhodamine-conjugated donkey anti-rabbit IgG (A and I) or fluorescein-conjugated rat anti-mouse IgG2 (E) as secondary antibodies (pseudo-coloured red). GD3 was detected using the monoclonal antibody R24 (mouse IgG3) as primary antibody and Alexa 488-conjugated goat anti-mouse IgG as secondary antibody (B and J, green). To reveal GD3 expression in cells shown in F, coverslips were sequentially incubated with R24, goat anti-mouse IgG3 and finally with Alexa 546-conjugated donkey anti-goat Ig (pseudo-coloured green). C, G andKaremergedimagesfromAandB,EandFandIandJ,respectively.AnenlargementoftheboxedareasinC,GandKareshowninD,Hand L, respectively. Images shown in this figure are single xy confocal sections. The focal plane in E–H was adjusted through the top of the cell. Scale bars: A, 20 lm (for A–C and I–K); E, 10 lm (for E–G, D and L); H, 3 lm. 2434 A. R. Zurita et al. (Eur. J. Biochem. 271) Ó FEBS 2004 buoyant density fractions in sucrose gradients, pelleted after ultracentrifugation and expressed as detergent-resistant patches on the plasma membrane of CHO-K1 cells [15,16]. Additionally in this work, using confocal microsco- py analysis we demonstrated that EGFr co-localizes only to a minor extent with the disialoganglioside GD3, even after stimulation with EGF. Taken together, these results make it less probable that there is a direct effect of GD3 on EGFr tyrosine phosphorylation [10] and suggests an indirect effect, perhaps through its interaction with other mem- brane-associated proteins. In this regard, it was recently described that overexpression of GM3 in cells of the human keratinocyte-derived cell line, SCC12F2, inhibited EGFr tyrosine phosphorylation, while it did not affect EGFr membrane distribution but shifted caveolin-1 to the deter- gent-soluble, EGFr-containing region [29]. The authors suggested that the GM3-induced shift of caveolin-1 mem- brane distribution is critical for its EGFr-induced phos- phorylation that is associated with the suppression of EGFr activation. The lack of EGFr in low buoyant density fractions in sucrose gradients, and the complete solubilization of the receptor to Triton X-100 extraction, strongly suggested that EGFr expressed in CHO-K1 cells is mainly excluded from GEM. However, our immunofluorescence microscopy experiments showed that GD3 and EGFr co-localized, to some extent, at the plasma membrane. A possible explan- ation for the nondetectable EGFr in GEM fractions from sucrose gradients is that the GEM-associated EGFr might besensitivetoextractionwithTritonX-100,evenwhenused at different concentrations (0.25, 0.5 and 1%), suggesting a relatively weak interaction of EGFr with Triton X-100- insoluble domains. These observations are in agreement with a study in HeLa cells, where it was shown that most of the EGFr is localized in lipid rafts containing the ganglio- side GM1 and is sensitive to Triton-X-100 extraction but insensitive to extraction with a less disrupting nonionic detergent, Brij 58 [30]. The EGF receptor family comprises four members – EGFr (ErbB1), ErbB2 (HER2 or Neu), ErbB3, and ErbB4 [31]. The orphan receptor, ErbB2, is the preferred heteroassociation partner of all other ErbB proteins, enhancing signalling potency by its strong latent kinase activity [32,33]. CHO cells express endogenous ErbB2, but no other members of the ErbB family [34]. In SKBR-3 (a breast tumour cell line), ErbB2 was found to co-localize with lipid rafts, identified by the GM1-binding B subunit of cholera toxin [35]. Taking all these observations together, we attempted to explore whether the effect of GD3 on EGFr phosphorylation might be achieved through the modulation of ErbB2 membrane distribution, its potential heteroassociation partner. Clearly, our results show that ErbB2 is expressed at a similar level in wild- type CHO-K1 cells (GM3 + ) and in cells from clone 2 Fig. 6. Detergent solubility and continuous sucrose gradient analysis of epidermal growth factor receptor 2 (ErbB2). (A) Homogenates from wild-type CHO-K1 cells (CHO-K1) and CHO-K1 cells transiently expressing human EGFr (CHO-K1 EGFr) were resolved in SDS– PAGE (8%) and probed with anti-EGFr or anti-ErbB2 Ig. The positions of EGFr (170 kDa) and ErbB2 (185 kDa) with molecular masses are indicated. Then, antibodies were removed by treatment of themembranewith1 M NaOH for 5 min and p53 was determined by Western blot. (B) Wild-type CHO-K1 cells (CHO-K1 wt) and cells from clone 2 were lysed at 4 °C for 1 h and centrifuged (10 h, 150 000 g,4°C) on a continuous sucrose gradient (5–30%). Twelve fractions were collected from the bottom of the sucrose density gra- dient. Proteins were resolved by electrophoresis through 8% SDS/ PAGE, and ErbB2 was detected by Western blot with anti-ErbB2 Ig. The position and molecular mass of ErbB2 (185 kDa) is indicated. (C) A quantitative analysis of Western blots from Figs 6B and 2A (CHO- K1 wt and clone 2 cells) was carried out to compare EGFr and ErbB2 gradient distribution. The receptor level in each fraction was normal- ized to total receptor expression. White bars, EGFr; black bars, ErbB2. Ó FEBS 2004 EGF receptors distribution in cell membranes (Eur. J. Biochem. 271) 2435 (GD3 + ). In addition, no significant changes in the biochemical behaviour of ErbB2 to extraction using a nonionic detergent were observed, suggesting that the modulation of EGFr phosphorylation by endogenously expressed GD3 does not occur because of a change in the membrane distribution of ErbB2. It was also found that an important fraction of the endogenous ErbB2 was associated with low-density fractions after extraction with cold Triton X-100 and sucrose gradient analysis. This fraction of the ErbB2 receptor represents receptor mole- cules, associated with GEM, in the membranes of CHO- K1 cells. Interestingly, it was recently reported that membrane clusters with a high concentration of ErbB2, which is regulated by lipid rafts, strongly influence homoassociation and the ligand-independent activation of ErbB2 [35]. Considering that the orphan ErbB2 is the only member of the ErbB family expressed in CHO-K1 cells, its activation by homodimerization is highly likely to occur in plasma membrane clusters (GEMs) of this cell line. In conclusion, our studies demonstrate that most of the EGFrs localize outside GEM in wild-type CHO-K1 cell (GM3 + ) membranes. Contrary to results showing that addition or depletion of cholesterol (another membrane component that regulates GEM formation) can alter the membrane distribution of EGFr [36,37], qualitative and quantitative changes in ganglioside expression do not affect the membrane distribution of EGFr and ErbB2. However, we cannot entirely rule out the possibility that fine-tuning mechanisms might be operating in the membrane distribu- tion of EGFr. An interesting possibility to explain EGFr regulation by gangliosides, particularly in GD3-expressing CHO-K1 cells, is that gangliosides might regulate the activity of ganglioside-stimulated receptor tyrosine phos- phatases [38], or enhance the co-localization of EGFr with its phosphatase, as recently suggested [7]. In this sense, our work provides the basis for testing these possibilities and gaining further insight into the regulation of the ErbB family members. Acknowledgements This work was supported, in part, by Grants from the SECyT- Universidad Nacional de Co ´ rdoba, ÔRamon Carrillo-Arturo On ˜ ativiaÕ from Ministerio de Salud de la Nacio ´ n Argentina (2001 to J.L.D. and 2003 to N.P.K.), the International Society for Neurochemistry (Special ISN One-Time Fund) and Fundacio ´ n Antorchas (Grant N°14116-112 toJ.L.D.andinpartbyGrantN°14022-10 to N.P.K.). We thank F. Cerban and A. Gruppi for their donation of anti-mouse IgG subtype antibodies and C. Alvarez for valuable reagents (Departamento de Bioquı ´ mica Clı ´ nica, Facultad de Ciencias Quı ´ mica, UNC, Argentina). The authors also thank G. Schachner and S. Deza for technical assistance with the cell culture and C. Mas for excellent assistance with confocal microscopy and image analysis. A.R.Z. and P.M.C. are recipients of CONICET (Argentina) Fellowships. J.L.D. and N.P.K. are Career Investigators of CONICET (Argentina). References 1. Hakomori, S. & Igarashi, Y. (1995) Functional role of glyco- sphingolipids in cell recognition and signaling. J. Biochem. (Tokyo) 118, 1091–1103. 2. Fredman, P., Hedberg, K. & Brezicka, T. (2003) Gangliosides as therapeutic targets for cancer. Biodrugs 17, 155–167. 3. Mutoh, T., Tokuda, A., Miyadai, T., Hamaguchi, M. & Fujiki, N. (1995) Ganglioside GM1 binds to the Trk protein and regulates receptor function. Proc. Natl Acad. Sci. USA 92, 5087–5091. 4. Fukumoto, S., Mutoh, T., Hasegawa, T., Miyazaki, H., Okada, M., Goto, G., Furukawa, K. & Urano, T. (2000) GD3 synthase gene expression in PC12 cells results in the continuous activation of TrkA and ERK1/2 and enhanced proliferation. J. Biol. Chem. 275, 5832–5838. 5. Allende, M.L. & Proia, R.L. (2002) Lubricating cell signaling pathways with gangliosides. Curr. Opin. Struct. Biol. 12, 587–592. 6. Yamashita, T., Hashiramoto, A., Haluzik, M., Mizukami, H., Beck, S., Norton, A., Kono, M., Tsuji, S., Daniotti, J.L., Werth, N., Sandhoff, R., Sandhoff, K. & Proia, R.L. (2003) Enhanced insulin sensitivity in mice lacking ganglioside GM3. Proc. Natl Acad. Sci. USA 100, 3445–3449. 7. Miljan, E.A. & Bremer, E.G. (2002) Regulation of growth factor receptors by gangliosides. Sci. STKE 160,RE15 9 . 8. Yates, A.J., Saqr, H.E. & Van Brocklyn, J. (1995) Ganglioside modulation of the PDGF receptor. A model for ganglioside functions. J. Neurooncol. 24, 65–73. 9. Mitsuda, T., Furukawa, K., Fukumoto, S., Miyazaki, H. & Urano, T. (2002) Overexpression of ganglioside GM1 results in the dispersion of platelet-derived growth factor receptor from glyco- lipid-enriched microdomains and in the suppression of cell growth signals. J. Biol. Chem. 277, 11239–11246. 10. Zurita, A.R., Maccioni, H.J. & Daniotti, J.L. (2001) Modulation of epidermal growth factor receptor phosphorylation by endogenously expressed gangliosides. Biochem. J. 355, 465–472. 11. Simons, K. & Ikonen, E. (1997) Functional rafts in cell mem- branes. Nature 387, 569–572. 12. Prinetti, A., Iwabuchi, K. & Hakomori, S. (1999) Glyco- sphingolipid-enriched signaling domain in mouse neuroblastoma Neuro2a cells. Mechanism of ganglioside-dependent neurito- genesis. J. Biol. Chem. 274, 20916–20924. 13. Simons, K. & Toomre, D. (2000) Lipid rafts and signal trans- duction. Nat. Rev. Mol. Cell Biol. 1, 31–39. 14. Galbiati, F., Razani, B. & Lisanti, M.P. (2001) Emerging themes in lipid rafts and caveolae. Cell 106, 403–411. 15. Crespo, P.M., Zurita, A.R. & Daniotti, J.L. (2002) Effect of gangliosides on the distribution of a glycosylphosphatidylinositol- anchored protein in plasma membrane from Chinese hamster ovary-K1 cells. J. Biol. Chem. 277, 44731–44739. 16. Crespo, P.M., Zurita, A.R., Giraudo, C.G., Maccioni, H.J.F. & Daniotti, J.L. (2004) Ganglioside glycosyltransferases and newly synthesized gangliosides are excluded from detergent-insoluble complexes of Golgi membranes. Biochem. J. 377, 561–568. 17.Daniotti,J.L.,Martina,J.A.,Giraudo,C.G.,Zurita,A.R.& Maccioni, H.J. (2000) GM3 alpha2,8-sialyltransferase (GD3 synthase): protein characterization and sub-Golgi location in CHO-K1 cells. J. Neurochem. 74, 1711–1720. 18. Giraudo, C.G., Rosales Fritz, V.M. & Maccioni, H.J. (1999) GA2/GM2/GD2 synthase localizes to the trans-Golgi network of CHO-K1 cells. Biochem. J. 342, 633–640. 19. Friedrichson, T. & Kurzchalia, T.V. (1998) Microdomains of GPI-anchored proteins in living cells revealed by crosslinking. Nature 394, 802–805. 20.Cochet,C.,Kashles,O.,Chambaz,E.M.,Borrello,I.,King, C.R. & Schlessinger, J. (1988) Demonstration of epidermal growth factor-induced receptor dimerization in living cells using a chemical covalent cross-linking agent. J. Biol. Chem. 263, 3290– 3295. 21. Sorkina, T., Huang, F., Beguinot, L. & Sorkin, A. (2002) Effect of tyrosine kinase inhibitors on clathrin-coated pit recruitment and 2436 A. R. Zurita et al. (Eur. J. Biochem. 271) Ó FEBS 2004 internalization of epidermal growth factor receptor. J. Biol. Chem. 277, 27433–27441. 22. Daniotti, J.L., Zurita, A.R., Trindade, V.M. & Maccioni, H.J. (2002) GD3 expression in CHO-K1 cells increases growth rate, induces morphological changes, and affects cell–substrate inter- actions. Neurochem. Res. 27, 1421–1429. 23. Rosales Fritz, V.M., Daniotti, J.L. & Maccioni, H.J. (1997) Chinese hamster ovary cells lacking GM1 and GD1a synthesize gangliosidesupontransfectionwithhumanGM2synthase. Biochim. Biophys. Acta 1354, 153–158. 24. Li, R., Manela, J., Kong, Y. & Ladisch, S. (2000) Cellular gang- liosides promote growth factor-induced proliferation of fibro- blasts. J. Biol. Chem. 275, 34213–34223. 25. Brown, D.A. & Rose, J.K. (1992) Sorting of GPI-anchored pro- teins to glycolipid-enriched membrane subdomains during trans- port to the apical cell surface. Cell 68, 533–544. 26. Kazui, A., Ono, M., Handa, K. & Hakomori, S. (2000) Glyco- sylation affects translocation of integrin Src, and caveolin into or out of GEM. Biochem. Biophys. Res. Commun. 273, 159–163. 27. Wang, F., Weaver, V.M., Petersen, O.W., Larabell, C.A., Dedhar, S., Briand, P., Lupu, R. & Bissell, M.J. (1998) Reciprocal inter- actions between beta1-integrin and epidermal growth factor receptor in three-dimensional basement membrane breast cultures: a different perspective in epithelial biology. Proc. Natl Acad. Sci. USA 95, 14821–14826. 28. Yamada, K.M. & Even-Ram, S. (2002) Integrin regulation of growth factor receptors. Nat. Cell Biol. 4, E75–E76. 29. Wang, X.Q., Sun, P. & Paller, A.S. (2002) Ganglioside induces caveolin-1 redistribution and interaction with the epidermal growth factor receptor. J. Biol. Chem. 277, 47028–47034. 30. Roepstorff, K., Thomsen, P., Sandvig, K. & van Deurs, B. (2002) Sequestration of epidermal growth factor receptors in non- caveolar lipid rafts inhibits ligand binding. J. Biol. Chem. 277, 18954–18960. 31. Yarden, Y. & Sliwkowski, M.X. (2001) Untangling the ErbB signalling network. Nat. Rev. Mol. Cell Biol. 2, 127–137. 32.Karunagaran,D.,Tzahar,E.,Beerli,R.R.,Chen,X.,Graus- Porta,D.,Ratzkin,B.J.,Seger,R.,Hynes,N.E.&Yarden,Y. (1996) ErbB-2 is a common auxiliary subunit of NDF and EGF receptors: implications for breast cancer. EMBO J. 15, 254–264. 33. Schlessinger, J. (2000) Cell signaling by receptor tyrosine kinases. Cell 103, 211–225. 34. Chausovsky, A., Waterman, H., Elbaum, M., Yarden, Y., Geiger, B. & Bershadsky, A.D. (2000) Molecular requirements for the effect of neuregulin on cell spreading, motility and colony orga- nization. Oncogene 19, 878–888. 35. Nagy,P.,Vereb,G.,Sebestyen,Z.,Horvath,G.,Lockett,S.J., Damjanovich, S., Park, J.W., Jovin, T.M. & Szollosi, J. (2002) Lipid rafts and the local density of ErbB proteins influence the biological role of homo- and heteroassociations of ErbB2. J. Cell Sci. 115, 4251–4262. 36. Ringerike, T., Blystad, F.D., Levy, F.O., Madshus, I.H. & Stang, E. (2002) Cholesterol is important in control of EGF receptor kinase activity but EGF receptors are not concentrated in caveo- lae. J. Cell Sci. 115, 1331–1340. 37. Pike, L.J. & Casey, L. (2002) Cholesterol levels modulate EGF receptor-mediated signaling by altering receptor function and trafficking. Biochemistry 41, 10315–10322. 38. SuarezPestana,E.,Tenev,T.,Gross,S.,Stoyanov,B.,Ogata,M. & Bohmer, F.D. (1999) The transmembrane protein tyrosine phosphatase RPTPsigma modulates signaling of the epidermal growth factor receptor in A431 cells. Oncogene 18, 4069–4079. Ó FEBS 2004 EGF receptors distribution in cell membranes (Eur. J. Biochem. 271) 2437 . Membrane distribution of epidermal growth factor receptors in cells expressing different gangliosides Adolfo R. Zurita 1 ,. [11–14]. Initial evidence showed the association of gangliosides to GEM present in the plasma membrane of CHO-K1 cells lines expressing different gangliosides