Overexpression of Ganglioside GM1 Results in the Dispersion of Platelet-derived Growth Factor Receptor from Glycolipid-enriched Microdomains and in the Suppression of Cell Growth Signals*

To investigate the molecular mechanisms of gangliosides for the regulation of cell proliferation, Swiss 3T3 cells were transfected with GM2/GD2 synthase and GM1 synthase cDNAs, resulting in the establishment of GM1-expressing (GM1+) lines. Compared with the vector control (GM1−) cell lines, GM1+ cells exhibited reduced cell proliferation by stimulation with platelet-derived growth factor (PDGF). In accordance with the reduced cell growth, GM1+ cells showed earlier decreases in the phosphorylation levels of PDGF receptor and less activation of MAP kinases than GM1− cells. To analyze the effects of GM1 expression on the PDGF/PDGF receptor (PDGFR) signals, the glycolipid-enriched microdomain (GEM) was isolated and the following results were obtained. (i) PDGFR predominantly distributed in the non-GEM fraction in GM1+ cells, while it was present in both GEM and non-GEM fractions in GM1− cells. (ii) Activation of PDGFR as detected by anti-phosphotyrosine antibody occurred almost in parallel with existing amounts of PDGFR in each fraction. (iii) GM1 binds with PDGFR in GEM fractions. These findings suggested that GM1 regulates signals via PDGF/PDGFR by controlling the distribution of PDGFR in- and outside of GEM, and also interacting with PDGFR in the GEM fraction as a functional constituent of the microdomain.

Gangliosides, sialic acid-containing glycosphingolipids are ubiquitously expressed in embryonal and adult tissues of mammals and birds (1). In particular, they are enriched in nervous tissues, and the structures of the carbohydrate moiety are strictly regulated according to the developmental stages and tissue differentiation (1,2). Biological roles of gangliosides have been investigated in many studies, and various functions have been claimed such as receptors for bacterial toxins (3), receptors for some viruses (4), modulators for Ca 2ϩ ions (5), those for adhesion molecules (6 -8), and for growth factor receptors (9). Some of them have also been assigned as a messenger of apotopsis signals (10). These functions can be classified into two major groups; recognition molecules for exogenous soluble molecules, and modulators of cis-acting receptor mole-cules for various growth/differentiation factors (9). However, the molecular mechanisms for ganglioside functions as described above have scarcely been clarified. A clear demonstration of the interaction between gangliosides and other receptor molecules has never been reported except for the binding of nerve growth factor receptor with GM1 (11).
The platelet-derived growth factor (PDGF) 1 receptor is a member of a family of tyrosine kinases that modulate multiple cellular processes in response to ligand binding. This receptor exerts roles through multiple phosphorylation cascades, each of which begins with the phosphorylation of the receptor itself. Recently, many of the participating molecules and substrates of PDGF/PDGFR effects have been identified, and their sites of interaction were mapped (12). These findings suggested the existence of a signaling module associated with PDGFR at the cell surface, consisting of components of the tyrosine kinase mitogen-activated protein kinase (MAPK) pathway. Anderson and co-workers (13) demonstrated that caveolae fractions from unstimulated fibroblasts contained PDGFR, Ras, Raf-1, MAP kinase kinase 1, and MAPK, and PDGF stimulation activated MAP kinase in the caveolae fraction, indicating that these components are functional in vivo.
Caveolae have been thought to be specialized plasmalemmal microdomains originally studied in numerous cell types for their involvement in the transcytosis of macromolecules (14). They are enriched in glycosphingolipids (GSLs), cholesterol, sphingomyelin, and lipid-anchored membrane proteins, and they are characterized by a light buoyant density and resistance to solubilization by Triton X-100 at 4°C. GSLs are enriched in this detergent-insoluble microdomain, that is almost equal to GEM (15,16). (GEM was used in this article for this microdomain.) In particular, GM1 has been suggested to be a useful marker of GEM as well as a main structural protein, caveolin. If gangliosides play roles in the modulation of various receptor molecules as described above, it may be possible that gangliosides such as GM1 are involved in the regulation of signaling as a functional component of GEM, being more than an indicator of the microdomain.
In fact, Bremer et al. (17,18) reported that GM3 and/or GM1 added to the culture medium of cells suppressed the cell growth and/or phosphorylation of epidermal growth factor receptor and PDGFR. Exogenous GM1, in turn, enhanced the phosphorylation of nerve growth factor receptor TrkA, resulting in neurite extension of rat pheochromocytoma PC12 cells (11). In this study, we have established transfectant cells of mouse fibroblast line Swiss 3T3 highly expressing GM1 using cloned cDNAs of GM2/GD2 synthase (19) and GM1 synthase (20). Then, we analyzed the effects of GM1 expression on the cell proliferation, phosphorylation of PDGFR/MAPK after PDGF stimulation, and also on the intracellular localization of PDGFR. The neo-expression of GM1 suppressed both cell proliferation and phosphorylation levels of PDGFR and MAPK in response to PDGF as reported in the experiments with exogenous GM1 (17). Surprisingly, the majority of PDGFR has moved from GEM to the non-GEM fraction, which appeared to be a major mechanism for the reduced PDGF/PDGFR signals in GM1 ϩ cells. Moreover, it appeared that GM1 bound PDGFR in GEM, indicating that GM1 is not only a marker of the GEM microdomain, but an important functional component probably modulating both GEM structure and PDGFR activity.
Cell Cultures-Swiss 3T3 cells were maintained in Dulbecco's modified Eagle's essential medium supplemented with 7.5% fetal calf serum at 37°C in a humidified atmosphere containing 5% CO 2 . For cell growth assay, cells were cultured in 48-well plates (Falcon, Lincoln Park, NJ), and serum-starved for 24 h before PDGF (50 ng/ml) treatment. After treatment for the indicated time, MTT assay was performed.
Gene Transfection and Selection-Swiss 3T3 cells used for cDNA transfection were plated in a 60-mm plastic tissue culture plate (Falcon). Two kinds of cDNAs were transfected into cells with LipofectAMI-NE TM (Invitrogen, Rockville, MD) according to the manufacturer's instructions. Stable transfectants were selected in the presence of 250 g/ml G418 (Sigma).
MTT Assay-Two ϫ 10 4 cells were seeded in 48-well plates. After serum deprivation for 24 h, they were cultured in the presence of PDGF (50 ng/ml) and 1% fetal calf serum. At day 0, 1, and 2 of culture, MTT assay was performed. Growth of cells was quantitated by assessing the reduction of MTT to formazan, measured as the absorbance at 590 nm using an enzyme-linkd immunosorbent assay reader (System Instruments, Japan).
Western Immunoblotting-Cells were plated at a density of 2 ϫ 10 6 cells/12 ml in two 10-cm plates, and serum-starved for 6 h before PDGF treatment. After treatment, cells were washed three times with phosphate-buffered saline, and lysed in lysis buffer (25 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Nonidet P-40, 1 mM Na 3 VO 4 , 1.5 mM EGTA, 1 mM phenylmethylsulfonyl fluoride, 20 units/ml aprotinin). Lysed cells were centrifuged at 8,000 rpm for 10 min. For MAPK phosphorylation assay, lysates were subjected to 12% SDS-PAGE. Proteins were transferred electrophoretically onto a polyvinylidene difluoride membrane and immunoblotted with antibodies reactive with MAPK (ERK1/2) or phospho-MAPK. For PDGF receptor phosphorylation assay, lysates were immunoprecipitated with anti-PDGFR antibody (958) before immunoblotting. Briefly, precleared samples were incubated with antibody for 1 h, and then incubated with protein A-Sepharose CL4B beads overnight. The beads were washed 3 times with lysis buffer. The precipitates were subjected to 8% SDS-PAGE, electroblotted, and then immunoblotted with anti-PDGFR antibody or PY20. The bands were detected by peroxidase-conjugated anti-rabbit IgG and the ECL detection system (PerkinElmer Life Science) in both assays. GEM Isolation-Five ϫ 10 7 cells were plated in five 15-cm dishes, and cultured up to 90% confluency. Then the cells were serum-starved for 6 h, and treated with PDGF (50 ng/ml) for 10 min. After treatment, cells were washed with phosphate-buffered saline containing 1 mM Na 3 VO 4 , collected, suspended in 1 ml of MNE buffer (25 mM MES, pH 6.5, 150 mM NaCl, 5 mM EDTA) containing 1% Triton X-100, then Dounce homogenized 20 times, and mixed with an equal volume of 80% sucrose (w/v) in MNE buffer. Then, samples were placed on the bottom of Ultra-Clear Centrifuge Tubes (Beckman Instruments). Two ml of 30% sucrose in MNE buffer was overlaid, and 1 ml of 5% sucrose in MNE buffer was layered on the top. These samples were centrifuged for 16 h at 20,000 ϫ g. The entire procedure was performed at 4°C. Five hundred ml each was fractionated from the top. An opaque band located immediately above the 5% interface (fraction 3) was collected and designated the GEM fraction. A sample from the bottom fraction (fraction 10) was collected and designated the non-GEM fraction.
Immunoprecipitation of PDGFR in the GEM and Non-GEM Fractions-Sample fractions isolated as described above were 500 l each. Fifty l of each sample fraction was precipitated with trichloroacetic acid, washed with acetone 3 times, and subjected to SDS-PAGE and immunoblotted. The remaining samples were dialyzed twice for 4 h against the dialysis buffer (50 mM HEPES, pH 7.4, 150 mM NaCl, 50 mM NaF, 10 mM Na pyrophosphate, 0.2% Nonidet P-40, 1 mM Na 3 VO 4 , 1 mM EDTA) before immunoprecipitation with anti-PDGFR antibody. After immunoprecipitation, beads were washed 4 times with washing buffer (50 mM Tris-HCl, pH 7.5, 0.3 M NaCl, 0.5% (w/v) sodium deoxycholate, 0.5% (v/v) Nonidet P-40/0.1% SDS), and subjected to SDS-PAGE and immunoblotting.
Cross-linking of PDGFR-Cells were plated at a density of 1 ϫ 10 7 /15 ml in five 15-cm plates, and serum starved for 6 h before PDGF treatment. Then, cells were washed three times with phosphate-buffered saline containing 1 mM Na 2 VO 4 , and cross-linked at room temperature using 8.7 mM bis(sulfosuccinimidyl)suberate (BS3; Pierce, Rockford, IL) in 25 mM HEPES (pH 8.5), 120 mM NaCl, 6 mM KCl, 1 mM MgCl 2 , and 10 mM EGTA. The reaction was terminated after 30 min. Cells were then subjected to GEM isolation.
In Vitro Kinase Assay of PDGFR-GEM fraction was isolated from V1 cells and served for immunoprecipitation with anti-PDGFR antibody. The precipitates were mixed with 20 l of tyrosine kinase assay buffer (50 mM HEPES, pH 7.4, 20 mM MnCl 2 , 5 mM MgCl 2 , 1 mM dithiothreitol, 100 M Na 3 VO 4 ). Then, 5 Ci of [␥-32 P]ATP and GM1 were added before incubation at 30°C for 10 min. Then, samples were subjected to 8% SDS-PAGE, and relative kinase activity was measured with autoradiography.
Cytostaining of GM1 and PDGFR-Intracelular localization of GM1 and PDGFR was analyzed by antibody staining. Cells were cultured on a cover glass and fixed with cold acetone containing 10% phosphatebuffered saline at Ϫ20°C for 10 min. PDGFR was stained with anti-PDGFR antibody (958) and FITC anti-rabbit antibody (BIOSOURCE Int., Camarillo, CA). GM1 was stained with rhodamine-conjugated choleratoxin B (LIST Biological Laboratories Inc.). Staining pattern was analyzed with confocal microscopy (Radiance TM , Bio-Rad Microscience Lab, Tokyo).
Effects of GM1 Expression on Cell Proliferation-The proliferation of the GM1 transfectants (M3 and M6) and the vector control lines (V1 and V2) were compared by MTT assay. As shown in Fig. 2, the GM1 transfectants showed a reduced growth rate in the presence of PDGF. The absorbance (590 nm) of transfectants (M3 and M6) in MTT assay was about 30 -40% of that of vector control lines (V1 and V2). In the absence of PDGF, the growth of the transfectants was similar to that of vector control cells (data not shown). Consequently, we found endogenously expressed GM1 suppressed PDGF-dependent growth of Swiss 3T3 cells.
Phosphorylation of MAPK-To analyze the alteration of PDGF receptor (PDGFR) and its downstream signaling molecules, the activation of MAPK was examined by Western immunoblotting (Fig. 3A). After 6 h of serum deprivation, 50 ng/ml PDGF was added and incubated for 5, 10, 30, 60, or 120 min. Then cells were lysed, resolved by SDS-PAGE, and immunoblotted with anti-MAPK antibody. After that, the membrane was reblotted with anti-phosphorylated MAPK antibody. The intensity of MAPK bands was almost equivalent between transfectants (M3 and M6) and control lines (V1 and V2) and no apparent change in the intensity was observed during the incubation. For 5-10 min after PDGF addition, the phosphorylation level of MAPK was equivalent between the two groups. However, after 30 -120 min, the transfectants showed earlier reduction in the activation levels of MAPK than those of vector control lines (Fig. 3B).
Phosphorylation of PDGFR-After analysis of the phosphorylation level of MAPK, the activation of PDGFR was examined by immunoprecipitation and immunoblotting (Fig. 4). After 6 h of serum deprivation, 50 ng/ml PDGF was added and incubated for 5 or 60 min. Then, lysates were immunoprecipitated with anti-PDGFR antibody. The immunoprecipitates were subjected to SDS-PAGE and immunoblotted with anti-phosphotyrosine antibody (PY20), (Fig. 4A, upper). Subsequently, the membrane was reblotted with anti-PDGFR antibody (Fig. 4A, lower). Phosphorylation levels of PDGFR in the transfectants were generally lower than those in vector control lines (Fig. 4B).
GEM Isolation-It is well known that GM1 localizes in GEM and has been used as a marker of GEM. PDGFR was found in caveolae by Anderson and co-workers (21). Therefore, we isolated GEM to compare the phosphorylation levels of PDGFR in GEM and non-GEM. To confirm the technique for the isolation of GEM, we isolated GEM of transfectant cells (M3) and immunoblotting was performed with anti-caveolin-1 antibody and cholera toxin B subunit to identify GM1. As shown in Fig. 5, both caveolin-1 and GM1 were detected mainly in fraction 3. Both were present only in the low density fractions, and not in high density fractions such as fractions 9 and 10 in which soluble proteins were present. Therefore, we used fraction 3 as the GEM, and fraction 10 as the non-GEM fraction in the following experiments.
Alteration of PDGFR Localization in the Transfectants-We isolated the GEM fraction and the non-GEM fraction from each cell line. In all cell lines, the isolated GEM fraction contained 0.05 mg of total protein, and the non-GEM fraction contained 0.5 mg of total protein. They were subjected to immunoprecipitation with anti-PDGFR antibody. Immunoprecipitates were then immunoblotted with anti-PDGFR antibody (Fig. 6A, top). Surprisingly, PDGFR in GEM of the transfectants (M3 and M6) was clearly and significantly reduced compared with that in non-GEM, although the vector control lines (V1 and V2) showed an almost equal distribution of PDGFR between GEM and non-GEM (Fig. 6, A and B, and Table I). Then, the membrane was reblotted with anti-phosphotyrosine antibody (PY20) to determine the phosphorylation of PDGFR (Fig. 6A). The phosphorylation of PDGFR was detected only in PDGFtreated samples in both the transfectants and vector controls, and the intensity of the phosphorylated bands was almost proportional to that of PDGFR bands stained with anti-PDGFR antibody (Fig. 6, A and C, and Table I). A part of GEM and the non-GEM fractions was trichloroacetic acid precipitated, separated on SDS-PAGE, and immunoblotted with anti-caveolin-1 antibody or the cholera toxin B subunit (Fig. 6A, bottom), confirming that caveolin-1 was present in GEM in both transfectants and vector controls, and that GM1 was enriched in the GEM fraction of the transfectants.
Alteration of Dimer Formation in the Transfectants-Crosslinking of PDGFR revealed that dimer bands could be detected in GEM fractions of control cells after PDGF treatment, but not in GEM fractions of the transfectant cells. Only a faint dimer band appeared in non-GEM of M6, but not in that of M3 (Fig. 7).
Co-precipitation of GM1 with PDGFR-To analyze the complex formation of PDGFR with GM1, the GEM fraction and FIG. 5. Isolation of GEM. Cell lysates prepared with 1% Triton X-100 in MNE buffer were mixed with 80% sucrose in MNE buffer (to make 40%). Thirty percent sucrose in MNE buffer, and 5% sucrose in MNE buffer were overlaid sequentially, then served for the density gradient centrifugation as described under "Experimental Procedures." Five hundred l each was fractionated from the top, and trichloroacetic precipitated, then served for immunoblotting. A, immunoblots with an anti-caveolin 1 rabbit antibody. B, immunoblots with biotin-conjugated cholera toxin B and detection was performed using ABC kit and ECL. non-GEM fraction were subjected to immunoprecipitation with anti-PDGFR antibody. Then, immunoprecipitates were extracted with chloroform/methanol (1:1), and the extracted samples were subjected to TLC immunostaining. The polyvinylidene difluoride membrane was blotted with cholera toxin B subunit, resulting in the detection of definite GM1 bands in the GEM fraction and faint bands in the non-GEM fraction (Fig.   8A). There was no significant difference in the intensity of GM1 bands between transfectants and vector controls (Fig. 8B), suggesting that the majority of the increased GM1 in the transfectants existed in GEM, but it was difficult for them to bind to PDGFR. The ratio of immunoprecipitated GM1 to PDGFR in GEM was markedly and significantly high in GM2 ϩ cells (Fig.  8C). The immunoprecipitates were also subjected to immunoblotting with anti-caveolin-1 antibody. Caveolin-1 could not be detected (data not shown).
Effects of GM1 on Kinase Activity of PDGFR-To analyze the effects of GM1 on the PDGFR kinase activity, immunoprecipitated PDGFR served for kinase assay in the presence of various amounts of GM1. As shown in Fig. 9, a low concentration of GM1 rather enhanced the kinase activity, while relatively high concentration of GM1 (higher than 100 M) suppressed the kinase in a dose-dependent manner.
Alteration of Intracellular Localization of PDGFR in the Transfectants as Analyzed by Immunocytostaining-To confirm the dispersion of PDGFR from GEM in the transfectants, cytostaining of GM1 and PDGFR was performed. GM1 was stained mainly on the cell membrane in both vector controls and transfectants, whereas the intensity was much stronger in the transfectant cells as expected (Fig. 10, left column). As for PDGFR, it was stained in membrane and cytoplasm as a granular pattern (Fig. 10, middle column). Merge of these two staining patterns revealed contrasting results between GM1 Ϫ cells and GM1 ϩ cells. The majority of GM1 was overlapped with PDGFR showing yellow color mainly around the cell membrane in the control cells (Fig. 10, right column). The faint green color seemed to represent PDGFR in the non-GEM fraction. In contrast, the yellow color indicating co-localizing GM1 and PDGFR was scarcely detected in the transfectants, and either the single red color (GM1/GEM) or single green color (PDGFR/non-GEM) was prominently detectable. These results as shown with higher magnification at the bottom were in good accordance with the results of the biochemical fractionation experiments as summarized in Fig. 11. DISCUSSION Cellular events regulated with signals via PDGF/PDGFR include cytoskeletal rearrangement and migration (22,23), mitogenesis (24), differentiation (25), calcium mobilization (26), and apotosis (27). These phenotypic changes induced via PDGF/PDGFR need sequential activation of various signaling molecules and substrates, and the mapping of the critical sites in the cytoplasmic domain of PDGFR for the activation of individual pathways has been achieved (12, 28 -30). The mechanisms for the PDGF/PDGFR to differentially regulate its multiple effects have been investigated by considering the existence of modular intermediates (31-33) and a functional unit at FIG. 6. Localization of PDGFR in the GEM/non-GEM fractions during PDGF stimulation. A, after 6 h of serum deprivation, cells were stimulated by PDGF (50 ng/ml) for 10 min. Then cells were lysed. GEM (G) and non-GEM (N) fractions were isolated as described in the legend to Fig. 5, and PDGFR was immunoprecipitated. The immunoprecipitates were immunoblotted using an anti-phosphotyrosine antibody (PY20) or anti-PDGF receptor as described under "Experimental Procedures." Note that levels of PDGFR in GEM were decreased in cells highly expressing GM1. Localization of caveolin-1 and GM1 was also examined using a caveolin-1 antibody or cholera toxin B, respectively. Same experiment was repeated at least 3 times with similar results. B, ratio of PDGFR bands in GEM to those in non-GEM in the cells before and after PDGF treatment. Average values Ϯ S.D. from three experiments are presented. Note that the GEM/non-GEM ratios were extremely decreased in the transfectant cells. C, ratio of phosphorylated PDGFR bands in GEM to total phosphorylated PDGFR (non-GEM ϩ GEM). Relative intensities to V2 bands (GEM ϩ non-GEM:100) are presented. Three experiments showed similar results, and mean value Ϯ S.D. is presented. Statistical significance were summarized in Table I.  7. Cross-linking of PDGFR to analyze dimer formation. Cells treated or untreated with PDGF were treated with BS3 and GEM/non-GEM fractions were prepared as described under "Experimental Procedures." Immunoprecipitates from each fraction was blotted with mAb PY20-horseradish peroxidase to detect tyrosine-phosphorylated monomer and dimer PDGFR, and ECL system was used for detection of bands. Three experiments were performed with similar results, and a representative one is shown. the cell surface membrane, designated as the caveolae (13) or detergent-insoluble microdomain (34).
Liu et al. (35) reported that caveolae membrane fractions from unstimulated human fibroblasts contain all of the molecular machinery required for PDGF to stimulate MAP kinase activation. GEM or caveolae is enriched in GSLs, cholesterol, sphingomyelin, and lipid-anchored membrane proteins. Among them, cholesterol and sphingomyelin are the major components of the lipid core of caveolae, and caveolins regulate the structure and functions of caveolae by physically associating with cholesterol and various other caveolae-residing proteins (13). Although GSLs have been considered to be enriched in caveolae or GEM, and GM1, a representative species has been used as a marker to indicate the correct preparation of GEM (36 -38), no special attention for their existence in caveolae/GEM has been paid. No significant roles of GSLs in the structure maintenance or in the signal regulation have been demonstrated except for a few studies on GM3 (39,40). Findings shown in the present study demonstrated that expression levels of ganglioside GM1 largely affected the intracellular localization of PDGFR, and consequently altered the quantity of the PDGF/PDGFR signals for cell proliferation, by reducing MAP kinase activation levels in response to PDGF stimulation. The discrepancy between the marked reduction in cell growth and MAPK activation and the mild decrease in PDGFR phosphorylation levels might imply the importance of GEM as a site of efficient signal transduction, i.e. PDGFR phosphorylation outside of GEM might be less efficient. This is the first study to demonstrate the role of GM1 to regulate the localization of a growth factor receptor inside/ outside of the GEM fraction. GM1 also significantly binds to PDGFR, although the meaning of the binding is not clear at this moment.
A few examples of the association of glycosphingolipids with growth factor receptors/signal molecules have been known; GM1 and nerve growth factor receptor TrkA (11), GM3 and Src/Rho (41), GD3 and Lyn (42), globotriaosylceramide (Gb3) and Yes (43). However, they just showed the association of protein molecules with membrane residing or exogenously added GSLs, and effects of the modification of the carbohydrate structures on the signaling functions have not been analyzed. Thus, findings obtained in the present study proposed a novel insight into the roles of glycolipids as a regulator of the organization of GEM and of the magnitude of the signals.
Bremer et al. (17,18) reported suppression of cell growth and PDGFR phosphorylation in response to PDGF by exogenous GM1 as well as GM3. Yates et al. (44) analyzed the mechanisms for the inhibition of PDGFR-stimulated mitogenesis with gangliosides, showing reduced dimerization of PDGFR as a main ganglioside function (45). Inhibition of signaling events mediated by PDGFR with GM1 in glioma cells (46,47) or with a few gangliosides in neuroblastoma cells (48) was also reported. The findings obtained in the present study appear almost similar, but the suppression levels are milder than those observed in the experiments with exogenous GM1. The conditions of the present experiments with GM1 synthase gene transfectants should be more physiological, and reflect the natural functions of GM1 in situ.
Since PDGF-dependent di-or multimerization of PDGFR is considered to be essential for the signal transduction, crosslinking experiments were performed. As shown in Fig. 7, dimer   FIG. 8. Binding of GM1 with PDGFR in GEM and non-GEM fractions. A, GEM fraction and non-GEM fraction were subjected to immunoprecipitation with anti-PDGFR antibody. Then, immunoprecipitates were extracted with chloroform/methanol (1:1), and the extracts were subjected to TLC immunostaining. To detect GM1, the polyvinylidene difluoride membranes were blotted with biotin-conjugated cholera toxin B subunit as described under "Experimental Procedures." B, relative intensities of GM1 bands in A was measured and presented. C, ratios of PDGFR-bound GM1 (co-precipitated from GEM) per PDGFR in GEM fractions (immunoprecipitated from GEM) were calculated using intensities of individual bands and resulting values are presented in arbitrary units. Results in B and C are presented as mean Ϯ S.D. of three experiments (* and **, p Ͻ 0.05).
FIG. 9. Effects of GM1 on PDGFR kinase assay. A, GEM fraction from V2 was subjected to immunoprecipitation with anti-PDGFR antibody as described under "Experimental Procedures." Precipitated PDGFR was mixed with various amount of GM1 and its kinase activity was measured by adding [␥-32 P]ATP and incubating for 10 min at 30°C, followed by autoradiography. B, relative intensity of each band was measured by NIH image and presented with the sample with no GM1 as 100 (%). Similar experiment was repeated and essentially same result was obtained.
bands were found only in PDGF-treated control cells, while GM1 ϩ transfectants showed no or minimal phosphorylated dimer bands, indicating that phosphorylated PDGFR was very much reduced in GEM of the transfectants. Dispersion of PDGFR from GEM seemed to have induced the poor phosphorylation and reduced proliferation signals.
In general, anchoring of receptors/signal molecules to caveolae/GEM are determined by lipid modification, such as acylation and glycosylphosphatidylinositol linkage. The former targets the cytoplasmic signaling molecules to the inside, and the latter targets glycosylphosphatidylinositol-anchored surface proteins to the outside of the caveolae/GEM microdomain. However, PDGFR should be targeted to the microdomains mainly by the hydrophobic nature of its transmembrane domain. Therefore, the alteration in the localization of PDGFR observed in this study might be due to either the modified lipid core of GEM with the carbohydrate change, or the modified nature of the transmembrane domain of PDGFR caused by overexpressed GM1. Whatever the mechanisms, this is the first study to find that modification of glycosphingolipids resulted in the dispersion of receptor/signaling proteins from the microdomain. The molecular mechanism is now under investigation in our laboratory.
There have been a few studies in which signaling proteins moved toward or from GEM after phosphorylation with outside stimulation (43, 49 -52). In the case of c-Ret, it is concentrated into rafts after stimulation of glial cell line-derived neurotrophic factor (53). In the present study, activation of PDGFR did not appear to cause significant changes in the localization of PDGFR, although the concentration of Shc, Syp, and MAPK in caveolae increased (21). The phosphorylation levels of PDGFR appeared almost parallel to its amounts in the GEM/non-GEM fractions. Consequently, the most critical factor to determine the quantity of PDGF/PDGFR signals in Swiss 3T3 cells appears to be the ratio of PDGFR localization between GEM and non-GEM.
Recently, Liu et al. (35) demonstrated that caveolae fractions from unstimulated fibroblasts contain PDGFR, Ras, Raf-1, MAPKK, and ERK-2, and they functionally associate upon PDGF exposure (35). Activated MAP kinase was not detected in non-caveolae fractions even after PDGF stimulation, suggesting that caveolae are a cell surface domain where MAP kinase is functionally linked to the PDGFR. Provided that PDGF/ PDGFR signals are mainly transduced via GEM as described for the glial cell line-derived neurotrophic factor/GFR␣/c-Ret signals, it seems reasonable to consider that dislocation of PDGFR in the transfectant cells causes the reduction of activation levels of PDGFR and of the downstream signaling molecules.