Role of gangliosides in the association of ErbB2 with lipid rafts in mammary epithelial HC11 cells Elena Sottocornola 1 , Roberta Misasi 2 , Vincenzo Mattei 2,3 , Laura Ciarlo 2 , Roberto Gradini 2,4 , Tina Garofalo 2,3 , Bruno Berra 1 , Irma Colombo 1 and Maurizio Sorice 2,3 1 Institute of General Physiology and Biological Chemistry, University of Milan, Italy 2 Department of Experimental Medicine and Pathology, University of Rome ‘La Sapienza’, Italy 3 Laboratory of Experimental Medicine and Environmental Pathology, Rieti, Italy 4 INM Neuromed, Pozzilli, Italy Gangliosides, ubiquitous components of eukaryotic membranes, are not uniformly distributed within the outer leaflet of the plasma membrane, but segregate, together with cholesterol, glycosylphosphatidylinositol- anchored proteins and signaling-transduction mole- cules, into unique, more or less stable clusters or microdomains called as ‘glycosphingolipid-enriched microdomains’ (GEM), which contribute to membrane structure, organization and, more importantly, func- tion. Indeed, GEM are viewed as a dynamic and pref- erential association of sphingolipids and cholesterol into moving platforms, termed lipid rafts, which can selectively incorporate or exclude proteins [1] and con- tribute to lipid-mediated protein trafficking and signal transduction [2]. The growth factor receptor tyrosine kinase ErbB2 is a 185 kDa transmembrane glycoprotein intensively investigated because of its important role in normal mammary gland development and in the deregulation of growth displayed by cancer cells, including breast and ovarian tumor cells [3,4]. A ligand which binds directly and specifically to ErbB2 has not been identi- fied to date, but it can be activated in trans by ligands binding to epidermal growth factor receptor (EGFR), such as epidermal growth factor (EGF) and transform- ing growth factor a1 [3]. Indeed, in cells coexpressing both ErbB2 and EGFR, EGF preferentially stimulates the formation of ErbB2 ⁄ EGFR heterodimers in which cross-phosphorylation occurs [5,6]. Data concerning the structural and mechanistic aspects required for Keywords epidermal growth factor receptor; ErbB2; GM3; HC11 cells; lipid rafts Correspondence M. Sorice, Department of Experimental Medicine and Pathology, University of Rome ‘La Sapienza’, viale Regina Elena 324, Rome 00161, Italy Fax: +39 6 445 4820 Tel: +39 6 499 72675 E-mail: maurizio.sorice@uniroma1.it (Received 24 December 2005, revised 15 February 2006, accepted 27 February 2006) doi:10.1111/j.1742-4658.2006.05203.x We analyzed the role of gangliosides in the association of the ErbB2 recep- tor tyrosine-kinase (RTK) with lipid rafts in mammary epithelial HC11 cells. Scanning confocal microscopy experiments revealed a strict ErbB2– GM3 colocalization in wild-type cells. In addition, analysis of membrane fractions obtained using a linear sucrose gradient showed that ErbB2, epi- dermal growth factor receptor (EGFR) and Shc-p66 (proteins correlated with the ErbB2 signal transduction pathway) were preferentially enriched in lipid rafts together with gangliosides. Blocking of endogenous ganglioside synthesis by (+ ⁄ –)-threo-1-phenyl-2-decanoylamino-3-morpho- lino-1-propanol hydrochloride ([D]-PDMP) induced a drastic cell-surface redistribution of ErbB2, EGFR and Shc-p66, within the Triton-soluble fractions, as revealed by linear sucrose-gradient analysis. This redistribution was partially reverted when exogenous GM3 was added to ganglioside- depleted HC11 cells. The results point out the key role of ganglioside GM3 in retaining ErbB2 and signal-transduction-correlated proteins in lipid rafts. Abbreviations [D]-PDMP, (+ ⁄ –)-threo-1-phenyl-2-decanoylamino-3-morpholino-1-propanol hydrochloride; EGF, epidermal growth factor; EGFR, epidermal growth factor receptor; FITC, fluorescein isothiocyanate; GEM, glycosphingolipid-enriched microdomains; HRP, horseradish peroxidase; Rf, retardation factor; RTK, receptor tyrosine-kinase; TX-100, Triton X-100. FEBS Journal 273 (2006) 1821–1830 ª 2006 The Authors Journal compilation ª 2006 FEBS 1821 EGF-dependent ‘trans activation’ of EGFR and ErbB2 in ErbB2 ⁄ EGFR heterodimers, as well as data con- cerning the overall effects induced by changes in the relative expression levels of EGFR and ErbB2, are growing rapidly and several lines of evidence have shown that ErbB2 is associated with lipid microdo- mains [7–10]. By contrast, the role of cell-membrane components, such as gangliosides, for determining the plasma membrane distribution and relative densities of receptors has not yet been investigated thoroughly. Recently, we reported the first evidence that ganglio- side depletion associates with increased levels of the activated ErbB2 and EGFR, whereas increased gan- glioside GM3 content correlates with the downregula- tion of both receptors [11]. In this study, we provide evidence of ErbB2–GM3 association on the plasma membrane of mouse mam- mary epithelial HC11 cells and demonstrate that gan- gliosides, and particularly GM3, play a key role in retaining ErbB2 and proteins correlated with its signal- transduction pathway in lipid rafts. Results ErbB2–GM3 colocalization in mammary epithelial HC11 cells In order to evaluate ErbB2 distribution and its pos- sible association with GM3, we performed immunoflu- orescence labeling, followed by scanning confocal microscopy analysis. Cells were labeled with anti- ErbB2 polyclonal serum and then with anti-GM3 monoclonal serum. Analysis of GM3 expression and distribution in untreated HC11 cells (Fig. 1A) revealed that GM3 staining appeared uneven over the cell surface, similar to that seen on ErbB2 molecule fluorescence. A merged image of the two stainings clearly revealed orange areas, resulting from the overlap of green and red fluorescence, which corresponded to colocalization areas. To analyze the effect of ganglioside depletion on the association of ErbB2 with ganglioside GM3, we preliminary treated the cells with (+ ⁄ –)-threo-1-phe- nyl-2-decanoylamino-3-morpholino-1-propanol hydro- chloride ([D]-PDMP), which blocks endogenous ganglioside biosynthesis, resulting in the almost com- plete disappearance of all ganglioside species [11]. Vir- tually no staining was observed in cells labeled with anti-GM3 serum (Fig. 1B). The lack of immunolabe- ling demonstrates the effect of [D]-PDMP on the depletion of gangliosides and the specificity of the anti- GM3 serum. The distribution of ErbB2 appeared more diffuse compared with control untreated cells. No colocalization areas between GM3 and ErbB2 were detected. By contrast, overlain areas were reverted when exo- genous GM3 was added to ganglioside-depleted HC11 cells (Fig. 1C). Scatter-plot diagrams showed how the dual labels are colocalized. Figure 1D shows a colocalization area that is evident in untreated HC11 cells. In cells treated A B C D Fig. 1. Scanning confocal microscopy analysis of GM3–ErbB2 association on HC11 cells. Cells were fixed with 4% paraformalde- hyde, permeabilized with 0.5% TX-100 and then incubated with anti-ErbB2 polyclonal serum, followed by the addition of FITC-conju- gated goat anti-(rabbit IgG) serum. Cells were then labeled with anti-GM3 monoclonal serum (GMR6), followed by the addition of Texas red-conjugated anti-(mouse IgM) serum. Merge: dual immuno- labeling of anti-GM3 (red) and anti-ErbB2 (green). Colocalization areas are stained in orange. (A) Untreated HC11 cells. (B) HC11 cells treated with 30 l M [D]-PDMP for 5 days. (C) HC11 cells treat- ed with 30 l M [D]-PDMP for 5 days and then with 125 lM GM3 for 5 min. (D) Two-dimensional scatter plot analysis of the dual- labeled fluorochromes (pseudocolor) GM3–ErbB2. Diagrams show the pixel intensity distribution of a dual-channel section. The x-axis represents intensity from the red channel; the y-axis represents intensity from the green channel; a colocalization area is evident in untreated HC11 cells. In cells treated with [D]-PDMP and then with GM3 a major colocalization index is evident because the blue area is larger and more directed towards the diagonal line. Figure shows analysis of about 40 cells. ErbB2–raft association in HC11 cells E. Sottocornola et al. 1822 FEBS Journal 273 (2006) 1821–1830 ª 2006 The Authors Journal compilation ª 2006 FEBS with [D]-PDMP and then with GM3 a major colocali- zation index is evident, because the blue area is larger and more directed towards the diagonal line. Effect of ganglioside depletion on cholesterol and caveolin-1 distribution in HC11 cells In order to verify that [D]-PDMP does not interfere with the membrane distribution of known lipid-raft markers, we investigated the cholesterol and caveolin-1 content of membrane fractions obtained by a 5–30% linear sucrose gradient from HC11 cells in the absence or presence of treatment with [D]-PDMP. As expected, cholesterol was present mainly in frac- tions 4–6 (Fig. 2A), which, under our experimental conditions, correspond to lipid-raft fractions. About 90% of the cholesterol content of the total cell extract was recovered in fractions 4–6. Virtually the same cho- lesterol distribution was observed in cells treated with [D]-PDMP. Similar findings were found in western blot analysis of caveolin-1 distribution in the sucrose gradient frac- tions. The analysis was performed loading fraction samples by volume. Because the protein content of Triton X-100 (TX-100)-soluble fractions 10 and 11 was higher than that of TX-100-insoluble fractions 4–6 (not shown) [12], we can observe that caveolin-1 was consistently enriched in TX-100-insoluble fractions in untreated, as well as in [D]-PDMP-treated, cells (Fig. 2B). ErbB2 preferential association with lipid-raft fractions in HC11 cells To evaluate the distribution of ErbB2 in raft fractions of HC11 cells, treated or not with [D]-PDMP, EGF or [D]-PDMP and EGF, all fractions obtained by sucrose gradient were analyzed by western blot (Fig. 3). The results revealed that in non-EGF-stimulated cells ErbB2 was present mainly in fractions 5 and 6, but also in fractions 7–11 (Fig. 3A), indicating that ErbB2 is preferentially present in raft fractions. EGF stimulat- ion did not seem to appreciably modify this distribut- ion (Fig. 3B). Interestingly, [D]-PDMP treatment induced a drastic cell-surface redistribution of ErbB2 (Fig. 3C). Indeed, the receptor became completely Triton soluble and was present exclusively in fractions 10 and 11. Identical profile redistribution of ErbB2 was also evident in A B Fig. 2. (A) Densitometric analysis of cholesterol content in HC11 sucrose-gradient membrane fractions. HC11 cells, either untreated or treated with 30 l M [D]-PDMP for 5 days, were lyzed in lysis buf- fer and the supernatant (postnuclear fraction) was subjected to sucrose density gradient separation. After centrifugation, the gradi- ent was fractionated and free cholesterol of each fraction was anal- yzed by TLC and quantified by densitometric scanning analysis. (B) western blot analysis of caveolin-1 distribution in HC11 sucrose gradient membrane fractions. Lysates from HC11 cells, either untreated or treated with 30 l M [D]-PDMP for 5 days, were anal- yzed by western blot with anti-(caveolin-1) polyclonal serum, fol- lowed by incubation with an HRP-conjugated anti-(rabbit IgG) serum, as a secondary antibody. A B C D E Fig. 3. ErbB2 distribution in HC11 sucrose gradient membrane fract- ions. HC11 cells were lyzed in lysis buffer and the supernatant (postnuclear fraction) was subjected to sucrose density gradient separation. After centrifugation, the gradient was fractionated and each fraction was analyzed by western blotting with anti-ErbB2 polyclonal serum, followed by incubation with a HRP-conjugated anti-(rabbit IgG) serum, as a secondary antibody. (A) Untreated HC11 cells. (B) HC11 cells treated with 10 n M EGF for 15 min. (C) HC11 cells treated with 30 l M [D]-PDMP for 5 days. (D) HC11 cells treated with 30 l M [D]-PDMP for 5 days and then with 10 nM EGF for 15 min. (E) HC11 cells treated with 30 l M [D]-PDMP for 5 days and then with 125 l M GM3 for 5 min. E. Sottocornola et al. ErbB2–raft association in HC11 cells FEBS Journal 273 (2006) 1821–1830 ª 2006 The Authors Journal compilation ª 2006 FEBS 1823 HC11 cells treated with [D]-PDMP and EGF (Fig. 3D), indicating that EGF is not determinant in defining the retention of ErbB2 into the lipid rafts. The effect of [D]-PDMP treatment was partially abolished by addition of exogenous GM3 to ganglio- side-depleted HC11 cells. In fact, after [D]-PDMP incubation, followed by GM3 treatment, a significant proportion of ErbB2 returned to fractions 4 and 5 (Fig. 3E). In order to better clarify the functional role of the association between ErbB2 and lipid rafts, we analyzed the distribution of phospho-ErbB2 in sucrose-gradient fractions obtained from HC11 cells in the absence or presence of treatment with EGF, [D]-PDMP and [D]-PDMP ⁄ GM3. Although, as expected, virtually no phosphorylated ErbB2 was detected in all the fractions obtained from control cells (Fig. 4A), after triggering with EGF (Fig. 4B), phosphorylated ErbB2 was found in both the TX-100-insoluble fractions and the TX-100-soluble fractions. Interestingly, in cells treated with [D]-PDMP and EGF (Fig. 4D) a band corres- ponding to phosphorylated ErbB2 was detected in fractions 10 and 11, whereas in ganglioside-depleted cells and cells treated with [D]-PDMP and GM3 no ErbB2 phosphorylation was observed (Fig. 4C,E respectively). These findings support the view that GM3 is mainly involved in retaining ErbB2 in lipid- raft domains, but that it is not involved in ErbB2 phosphorylation. EGFR and Shc-p66 preferential association with lipid-raft fractions in HC11 cells Because in cells coexpressing both ErbB2 and EGFR, as is the case of HC11 cells, the two proteins strictly interact and EGF preferentially stimulates the forma- tion of ErbB2 ⁄ EGFR heterodimers [5,6], in the same raft fractions of HC11 cells analyzed previously, we also examined the distribution of EGFR. EGFR was present in fractions 5 and 6, but also in fractions 7–11 (Fig. 5A). In EGF-stimulated cells, movement of the receptor to TX-100-soluble fractions was observed (Fig. 5B). After [D]-PDMP treatment, the receptor became completely Triton soluble and was present exclusively in fractions 10 and 11 (Fig. 5C). After [D]-PDMP incu- bation, followed by GM3 treatment, a proportion of EGFR returned to fractions 4–6 (Fig. 5D). Because the Shc proteins are translocated into the lipid rafts of the plasma membrane after phosphoryla- tion by ErbB2 and EGFR receptors [13], we also ana- lyzed the distribution of Shc-p66 in the same fractions (Fig. 6). The results revealed that in control cells only a small amount of Shc-p66 was detectable in fractions 5 and 6, corresponding to lipid rafts; a higher amount of Shc-p66 was detected in TX-100-soluble fractions (mainly 10 and 11) (Fig. 6A). In cells treated with EGF the higher amount of Shc-p66 was detected in the Triton-insoluble fractions (Fig. 6B), indicating that A B C D E Fig. 4. Analysis of the distribution of phosphorylated ErbB2 in HC11 sucrose gradient membrane fractions. HC11 cells were lyzed in lysis buffer and the supernatant (postnuclear fraction) was sub- jected to sucrose density gradient separation. After centrifugation, the gradient was fractionated and each fraction was analyzed by western blotting with anti-(phospho-ErbB2) polyclonal serum, fol- lowed by incubation with an HRP-conjugated anti-(rabbit IgG) serum, as a secondary antibody. (A) Untreated HC11 cells. (B) HC11 cells treated with 10 n M EGF for 15 min. (C) HC11 cells treat- ed with 30 l M [D]-PDMP for 5 days. (D) HC11 cells treated with 30 l M [D]-PDMP for 5 days and then with 10 nM EGF for 15 min. (E) HC11 cells treated with 30 l M [D]-PDMP for 5 days and then with 125 l M GM3 for 5 min. A B C D Fig. 5. EGFR distribution in HC11 sucrose gradient membrane fract- ions. HC11 cells were lyzed in lysis buffer and the supernatant (postnuclear fraction) was subjected to sucrose density gradient separation. After centrifugation, the gradient was fractionated and each gradient fraction analyzed by western blotting with anti-EGFR polyclonal serum, followed by incubation with an HRP-conjugated anti-(rabbit IgG), as a secondary antibody. (A) Untreated HC11 cells. (B) HC11 cells treated with 10 n M EGF for 15 min. (C) HC11 cells treated with 30 l M [D]-PDMP for 5 days. (D) HC11 cells treated with 30 l M [D]-PDMP for 5 days and then with 125 lM GM3 for 5 min. ErbB2–raft association in HC11 cells E. Sottocornola et al. 1824 FEBS Journal 273 (2006) 1821–1830 ª 2006 The Authors Journal compilation ª 2006 FEBS EGF induced Shc-p66 recruitment to lipid rafts. Importantly, [D]-PDMP treatment caused a significant modification of its membrane distribution, inducing, as for ErbB2 and EGFR, an almost complete shift of Shc-p66 to Triton-soluble fractions 10 and 11 (Fig. 6C), which was partially reverted in ganglioside- depleted HC11 cells by the addition of GM3 and EGF (Fig. 6D). Association of ErbB2 with Shc-p66 To verify whether Shc-p66 may interact with activated ErbB2, lysates from TX-100-insoluble fractions (frac- tions 4–6 pooled) and TX-100-soluble fractions (frac- tions 10 and 11 pooled), obtained from EGF-treated and untreated cells, were immunoprecipitated with the anti-Shc-p66 Ab, followed by protein G–acrylic beads. The results in Fig. 7A show that, in control unsti- mulated cells, ErbB2 was slightly associated with Shc- p66 mainly in TX-100 soluble fractions. By contrast, after triggering with EGF, a significant proportion of Shc-p66 also became associated with ErbB2 in the TX-100-insoluble fractions, suggesting that, after EGF stimulation, Shc-p66 may associate with activated ErbB2 within lipid rafts. No bands were detected after EGF stimulation in control immunoprecipitation experiments with a rabbit IgG having irrelevant specificity. Immunoprecipitation was verified by western blot (Fig. 7B). Profile distribution of raft markers in plasma membrane fractions of HC11 cells Because gangliosides are well-known markers of lipid- raft domains, we examined the ganglioside profile of sucrose-gradient fractions from HC11 cells. Ganglio- sides were extracted in chloroform ⁄ methanol⁄ water and separated by HPTLC. Resorcinol-positive bands were identified on the basis of their HPTLC mobility, compared with standard reference molecules. Three main resorcinol positive bands, having a retardation factor (Rf) analogous to GM3, GM2 (the most prom- inent) and GD1a, respectively, were detected (Fig. 8A). The observation that the main band comi- grates with GM2 is not surprising, because this mole- cule is the main ganglioside constituent in these cells, as reported previously [11]. All the ganglioside bands were exclusively detectable in fractions 4–6, which, under our experimental conditions, correspond to lipid rafts. A B C D Fig. 6. Shc-p66 distribution in HC11 sucrose gradient membrane fractions. HC11 cells were lyzed in lysis buffer and the supernatant (postnuclear fraction) was subjected to sucrose density gradient separation. After centrifugation, the gradient was fractionated and each gradient fraction was analyzed by western blotting with anti- Shc polyclonal serum, followed by incubation with an HRP-conju- gated anti-(rabbit IgG), as a secondary antibody. (A) Untreated HC11 cells. (B) HC11 cells treated with 10 n M EGF for 15 min. (C) HC11 cells treated with 30 l M [D]-PDMP for 5 days. (D) HC11 cells treated with 30 l M [D]-PDMP for 5 days and then with 125 lM GM3 for 5 min plus 10 nM EGF for 15 min. A B Fig. 7. HC11 cells, treated or not with EGF (10 nM for 5 min at 37 °C), were lyzed in lysis buffer and the supernatant (postnuclear fraction) was subjected to sucrose density gradient separation. After centrifugation, the gradient was fractionated and coimmunoprecipitation of Shc- p66 with ErbB2 was performed in TX-100-insoluble fractions (4–6 pooled together) or in TX-100-soluble fractions (10 and 11 pooled together) with an anti-Shc-p66 specific serum. The immunoprecipitates were analyzed by western blot with anti-ErbB2 (A) and anti-Shc-p66 (B) sera. E. Sottocornola et al. ErbB2–raft association in HC11 cells FEBS Journal 273 (2006) 1821–1830 ª 2006 The Authors Journal compilation ª 2006 FEBS 1825 To further confirm the correct sucrose density gradi- ent separation of TX-100-insoluble and TX-100-soluble fractions, we also analyzed the distribution pattern of the known raft protein flotillin-2. The analysis was performed loading fraction samples by volume. Because the protein content of TX-100- soluble fractions 10 and 11 was much higher than that of TX-100-insoluble fractions 4–6 (not shown) [12], we can observe that flotillin-2 was consistently enriched in rafts (TX-100-insoluble fractions) (Fig. 8B). Discussion In this study we analyzed primarily the localization of ErbB2 in lipid rafts of mouse mammary epithelial HC11 cells. Laser scanning confocal microscopy obser- vations revealed colocalization areas between GM3, a well-known marker of lipid rafts [14], and ErbB2. This finding is in agreement with and extends previous observations about the surface distribution of ErbB2, which is mostly excluded from clathrin-coated pits on the cell plasma membrane [15], and it gives further support to the conclusions of Nagy et al. [9], who hypothesized the association of ErbB proteins (ErbB2 and ErbB3) with these microdomains by quantitative fluorescence microscopy in SKBR-3 breast cancer cells. In addition, in CHO-K1 cells, expression of GD3 affected, to some extent, the plasma membrane distri- bution of endogenous ErbB2 [16]. The preferential distribution of ErbB2 in lipid rafts was clearly demon- strated by our membrane fractionation experiments, which also revealed that EGF is not able to modify the receptor localization. However, analysis revealed that ErbB2 is not exclusively associated with the raft fractions. This finding is consistent with the observa- tions of Hommelgaard et al. [10], prompting us to hypothesize that ErbB2 is in dynamic equilibrium with lipid rafts in the membrane protrusions so that, at a single time point, only a fraction of ErbB2 is directly interacting with the raft gangliosides. This transient ErbB2–gangliosides interaction could potentially regu- late the function of ErbB2 (heterodimerization, signa- ling and metabolic fate) [10]. We therefore analyzed the role of gangliosides in the association of ErbB2 with lipid rafts in mammary epi- thelial HC11 cells. The key role played by gangliosides in defining the distribution of ErbB2 into signaling specialized plasma membrane domains was shown by treatment of HC11 cells with [D]-PDMP. Ganglioside depletion, due to the inhibition of endogenous ganglio- side synthesis, was shown to have striking effects upon the plasma membrane localization of ErbB2. Indeed, ErbB2 underwent complete redistribution within the high-density TX-100-soluble fractions of the plasma membrane, indicating, by a novel approach, that gan- gliosides play a key role in the retention of this protein in lipid rafts. These findings are strongly supported by the observation that [D]-PDMP does not destroy the organization of lipid rafts, because cholesterol as well as caveolin-1 could still be detected in TX-100-insol- uble fractions after treatment with [D]-PDMP. In addi- tion, treatment of ganglioside-depleted HC11 cells with exogenous ganglioside GM3 induced the return of a significant proportion of ErbB2 in raft fractions. How- ever, both ErbB2 localized in TX-100-soluble and TX-100-insoluble fractions were phosphorylated. These data, together with previous results [11], strengthen the view that GM3 plays an important role in ErbB2 membrane localization but not in its phosphorylation, suggesting that gangliosides might influence the signa- ling-transduction pathways after EGF stimulation by compartmentalizing the receptor in different membrane domains. Because in cells coexpressing ErbB2 and EGFR, like HC11 cells, ligand stimulation largely favors the for- mation of ErbB2 ⁄ EGFR heterodimers [5,6], EGFR distribution in the same plasma membrane fractions was also analyzed. Conflicting results have been repor- ted in the literature on the presence of EGFR within lipid rafts [7,17–21]. It may depend on the cell type, the use of different detergents [17,18] and, mainly, the ganglioside composition of the cells. The latter may also be influenced by cell cycle and ⁄ or cell density [19]. Zurita et al. [16] found EGFR mainly in TX-100- soluble fractions in CHO-K1 cells, although they observed that EGFR and GD3 colocalized on the cell surface. In the same vein, Wang et al. demonstrated 1234567891011St GM3 GM2 GM1 GD1a GD1b GT1b 1 Flotillin Fraction 23 4567 89 10 11 45 kDa A B Fig. 8. Ganglioside distribution in HC11 sucrose gradient membrane fractions. HC11 cells were lyzed in lysis buffer and the supernatant (postnuclear fraction) was subjected to sucrose density gradient separation. After centrifugation, 11 gradient fractions were recov- ered. (A) Gangliosides were extracted in chloroform ⁄ methanol ⁄ water from each fraction and analyzed by HPTLC. St: standard gan- gliosides GM3, GM2, GM1, GD1a, GD1b, GT1b. (B) Western blot- ting with anti-(flotillin-2) polyclonal serum in the same fractions. ErbB2–raft association in HC11 cells E. Sottocornola et al. 1826 FEBS Journal 273 (2006) 1821–1830 ª 2006 The Authors Journal compilation ª 2006 FEBS that endogenous overexpression of GM3 promotes co- immunoprecipitation of GM3 with EGFR [20]. How- ever, it has been shown that GM3 can specifically interact with the purified recombinant extracellular domain of EGFR [22] and that this tyrosine kinase receptor contains a structural domain with targeting information for lipid domains [23]. Our results dis- played variations in the distribution profiles rather similar to that of ErbB2. Indeed, in control cells, EGFR is mainly enriched in TX-100-insoluble fractions, whereas treatment with [D]-PDMP and [D]-PDMP ⁄ EGF shifts the receptor towards the TX- 100-soluble fractions, confirming the direct correlation between the two receptors [3,5]. However, after stimu- lation with EGF, we observed a movement of EGFR to Triton-soluble fractions, in agreement with previous studies showing that EGFR is initially concentrated in caveolae within lipid rafts, but rapidly moves out of this membrane domain in response to EGF [7]. The inefficient movement of ErbB2 out of these micro- domains may be related to its impaired internalization by clathrin-coated pits [24]. ErbB receptors, and ErbB2 in particular, are able to activate the ras ⁄ MAP kinase signaling pathway via the Shc proteins [25]. In order to elucidate whether this sig- nal transduction pathway triggered by ErbB2 may take place inside rafts in HC11 cells, we investigated the presence of Shc-p66 in raft fractions. Our findings from fractionation experiments showed a preferential associ- ation of Shc-p66 with lipid rafts after EGF stimulation. These data were also confirmed by coimmunoprecipitat- ion experiments, in which a consistent proportion of Shc-p66 coimmunoprecipitates with ErbB2 in the lipid- raft fractions, suggesting that actually Shc-p66 is recruited by ErbB2 after triggering via EGF. In conclusion, we demonstrated a key role for gan- gliosides in the association of ErbB2 with lipid rafts in mammary epithelial HC11 cells. This finding was strongly supported by the observation that addition of GM3 after [D]-PDMP treatment induced a marked redistribution of ErbB2 and proteins correlated with the its signal transduction pathway (i.e. Shc-p66) to Triton-insoluble fractions. Experimental procedures Cell culture and treatments Mouse mammary epithelial HC11 cells were a gift from E. Garattini (Institute for Pharmacological Research ‘M. Negri’, Milan, Italy). Cells were maintained in RPMI-1640 (Gibco-BRL, Life Technologies Italia srl, Italy), supplemen- ted with 10% heat-inactivated newborn bovine serum, 8mm glutamine, 50 lgÆmL )1 gentamycin and 5 lgÆmL )1 insulin from bovine pancreas (Sigma, St. Louis, MO), in a humidified 5% CO 2 atmosphere at 37 °C. As described previously [11], total ganglioside depletion was obtained by treating cells for 5 days at 37 °C with 30 lm [D]-PDMP (Sigma), a competitive inhibitor of gluco- sylceramide synthetase, resulting in ganglioside biosynthesis inhibition. Exogenous ganglioside GM3 treatment of ganglioside- depleted HC11 cells was performed by incubating the cells at 37 °C for 5 min with 125 lm GM3 (Alexis, S. Diego, CA), dissolved in routine medium without serum. EGF stimulation was carried out by incubating the cells, treated or not with [D]-PDMP, with 10 nm EGF (Sigma) for 15 min at 37 °C. Analysis of ErbB2–GM3 colocalization by scanning confocal microscopy HC11 cells, treated or not with [D]-PDMP and GM3, were fixed in situ with 4% paraformaldehyde in NaCl ⁄ P i for 30 min at room temperature and then permeabilized with 0.5% TX-100 in NaCl ⁄ P i for 30 min at room temperature. Cells were labeled with rabbit anti-ErbB2 polyclonal serum (C18, Santa Cruz Biotechnology, Santa Cruz, CA) for 1 h at 4 °C, followed by the addition (30 min at 4 °C) of fluo- rescein isothiocyanate (FITC)-conjugated anti-(rabbit IgG) serum (Calbiochem, La Jolla CA). After three washes in NaCl ⁄ P i , cells were incubated with GMR6 anti-GM3 monoclonal serum (Seikagaku Corp., Chuo-ku, Tokyo, Japan) [26] for 1 h at 4 °C, followed by three washes in NaCl ⁄ P i and the addition (30 min at 4 °C) of Texas red-conjugated goat anti-(mouse IgM) serum (Sigma). In parallel experiments, cells were stained with anti-GM3 monoclonal serum before fixing the cells. Alternatively, control experiments were performed omitting the monoclo- nal antibody from the immunolabeling procedure. After washing as above, cells were mounted upside down onto a glass slide in 5 lL of glycerol ⁄ Tris ⁄ HCl (6 : 4, v : v), pH 9.2. As a control, cells were mounted in glycerol ⁄ NaCl ⁄ P i (6 : 4, v : v), pH 7.4 and the results were virtually the same. The images were acquired using a high-resolution ·63 objective through a confocal laser scanning microscope Zeiss LSM 510 (Zeiss, Oberkochen, Germany) equipped with argon and HeNe ion lasers. The green (FITC) and red (Texas Red) fluorophores were excited simultaneously at 488 and 543 nm. Acquisition of single FITC-stained samples in dual-fluorescence scanning configuration did not show con- tribution of green signal in red. Images were collected at 512 · 512 pixels. Isolation and analysis of lipid-raft fractions GEM fractions from HC11 cells, treated or not with EGF (10 nm for 15 min at 37 °C), [D]-PDMP (30 lm for 5 days E. Sottocornola et al. ErbB2–raft association in HC11 cells FEBS Journal 273 (2006) 1821–1830 ª 2006 The Authors Journal compilation ª 2006 FEBS 1827 at 37 °C), [D]-PDMP and EGF, [D]-PDMP and GM3 (125 lm for 5 min at 37 °C), or [D]-PDMP and GM3 plus EGF, were isolated as described previously [27]. Briefly, 2 · 10 8 cells were suspended in 1 mL of lysis buffer, con- taining 1% TX-100, 10 mm Tris ⁄ HCl pH 7.5, 150 mm NaCl, 5 mm EDTA, 1 mm Na 3 VO 4 , and 75 U aprotinin, and allowed to stand for 20 min at 4 °C. The cell suspen- sion was mechanically disrupted by Dounce homogeniza- tion (10 strokes). The lysate was centrifuged for 5 min at 1300 g to remove nuclei and large cellular debris. The supernatant fraction (postnuclear fraction) was subjected to sucrose density gradient centrifugation, i.e. the fraction was mixed with an equal volume of 85% sucrose (w ⁄ v) in lysis buffer (10 mm Tris ⁄ HCl pH 7.5, 150 mm NaCl, 5 mm EDTA). The resulting diluent was placed at the bottom of a linear sucrose gradient (5–30%) in the same buffer and centrifuged at 200 000 g for 16–18 h at 4 °C in a SW41 rotor (Beckman Institute, Palo Alto, CA). After centrifuga- tion, the gradient was fractionated, and 11 fractions were collected starting from the top of the tube. All steps were performed at 0–4 °C. The amount of protein in each frac- tion was first quantified by Bio-Rad protein assay (Bio-Rad Laboratory GmbH, Munchen, Germany). Finally, fractions were subjected to cholesterol analysis, western blot, immunoprecipitation experiments or ganglio- side extraction. Analysis of cholesterol content All the fractions obtained as reported above from HC11 cells, treated or not with [D]-PDMP, were subjected to cho- lesterol analysis. The amount of cholesterol was evaluated as described previously [28]. Free cholesterol was quantified from TLC plates by densitometric scanning. The density of the bands used to quantitate cholesterol concentration fell within the linear range of compound concentration vs. absorbance. Immunoblotting analysis of plasma membrane fractions All the fractions obtained as reported above were subjected to 7.5 or 10% SDS ⁄ PAGE. Equal volumes of each fraction were loaded in SDS ⁄ PAGE, according to Parolini et al. [12]. The proteins were electrophoretically transferred to nitrocellulose membrane (Bio-Rad, Hercules, CA) and then, after blocking with NaCl ⁄ P i containing 1% albumin, probed with rabbit anti-(ErbB2 IgG) polyclonal serum, rab- bit anti-(phospho ErbB2) polyclonal serum (Sigma), rabbit anti-(EGFR IgG) polyclonal serum, clone 1005 (Santa Cruz Biotechnology), rabbit anti-(Shc IgG) polyclonal serum (Transduction Laboratories, Lexington, KY) or, as con- trols, rabbit anti-(caveolin-1) polyclonal serum (N-20, Santa Cruz Biotechnology) or goat anti-(flotillin-2) polyclonal serum (C-20, Santa Cruz Biotechnology). Bound antibodies were visualized with horseradish peroxidase (HRP)-conju- gated anti-(rabbit IgG) or anti-(goat IgG) serum (Sigma) and immunoreactivity was assessed by chemiluminescence reaction using the ECL western blotting detection system (Amersham, UK). As a control for nonspecific reactivity, parallel blots were performed as above, using an anti- (rabbit IgG) serum (Sigma). Immunoprecipitation experiments Briefly, TX-100-insoluble (fractions 4–6) or TX-100-soluble (fractions 10–11) fractions from HC11 cells, untreated or treated with 10 nm EGF (Sigma) for 15 min at 37 °C were lyzed in lysis buffer (10 mm Tris ⁄ HCl (pH 8.0), 150 mm NaCl, 1% Nonidet P-40, 1 mm phenylmethylsulfonyl fluor- ide, 10 lg of leupeptinÆmL )1 ). Cell-free lysates were mixed with protein G–acrylic beads and stirred by a rotary shaker for 2 h at 4 °C to preclear nonspecific binding. After cen- trifugation (500 g for 1 min), the supernatant was immuno- precipitated with the rabbit polyclonal anti-(Shc IgG) serum (Transduction Laboratories) plus protein G–acrylic beads. A rabbit IgG isotypic control (Sigma) was employed. Immunoprecipitates were subjected to western blot anal- ysis with the rabbit anti-ErbB2 polyclonal serum (Santa Cruz Biotechnology). Immunoreactivity was assessed by chemiluminescence reaction using the ECL western blotting detection system (Amersham). Ganglioside extraction and analysis by HPTLC Ganglioside extraction was performed according to the method of Svennerholm & Fredman [29] with minor modi- fications. Briefly, glycosphingolipids were extracted twice in chloroform ⁄ methanol ⁄ water (4 : 8 : 3 v ⁄ v ⁄ v) and subjected to Folch partition by the addition of water resulting in a final chloroform ⁄ methanol ⁄ water ratio of 1 : 2 : 1.4 (v ⁄ v ⁄ v). The upper phase, containing polar glycosphingo- lipids, was purified of salts and low molecular mass contaminants using Bond Elut-C18 columns, 3 mL (Superchrom, Harbor City, CA), according to the method of Williams & McCluer [30]. The eluted glycosphingolipids were dried and separated by HPTLC, using silica gel 60 HPTLC plates (Merck, Darmstadt, Germany). Chromato- graphy was performed in chloroform ⁄ methanol ⁄ 0.25% aqueous KCl (5 : 4 : 1 v ⁄ v ⁄ v). Plates were then air-dried and gangliosides visualized with resorcinol. References 1 Simons K & Ikonen E (1997) Functional rafts in cell membranes. Nature 387, 569–572. 2 Ikonen E (2001) Roles of lipid rafts in membrane trans- port. Curr Opin Cell Biol 13, 470–477. ErbB2–raft association in HC11 cells E. Sottocornola et al. 1828 FEBS Journal 273 (2006) 1821–1830 ª 2006 The Authors Journal compilation ª 2006 FEBS 3 Hynes NE & Stern DF (1994) The biology of erbB-2 ⁄ neu ⁄ HER-2 and its role in cancer. Biochim Biophys Acta 1198, 165–184. 4 Dougall WC, Qian X, Peterson NC, Miller MJ, Samanta A & Greene MI (1994) The neu-oncogene: signal transduction pathways, transformation mechan- isms and evolving therapies. Oncogene 9, 2109–2123. 5 Qian X, LeVea CM, Freeman JK, Dougall WC & Greene MI (1994) Heterodimerization of epidermal growth factor receptor and wild-type or kinase-deficient Neu: a mechanism of interreceptor kinase activation and transphosphorylation. Proc Natl Acad Sci USA 91, 1500–1504. 6 Gulliford TJ, Huang GC, Ouyang X & Epstein RJ (1997) Reduced ability of transforming growth factor- alpha to induce EGF receptor heterodimerization and downregulation suggests a mechanism of oncogenic synergy with ErbB2. Oncogene 15, 2219–2223. 7 Mineo C, Gill GN & Anderson RCW (1999) Regulated migration of epidermal growth factor receptor from caveolae. J Biol Chem 274, 30636–30643. 8 Zhou W & Carpenter G (2001) Heregulin-dependent translocation and hyperphosphorylation of ErbB-2. Oncogene 20, 3918–3920. 9 Nagy P, Vereb G, Sebestyen Z, Horvath G, Lockett SJ, Damjanovich S, Park JW, Jovin TM & Szollosi J (2002) Lipid rafts and the local density of ErbB proteins influ- ence the biological role of homo- and heteroassociation of ErbB2. J Cell Sci 115, 4251–4262. 10 Hommelgaard AM, Lerdrup M & van Deurs B (2004) Association with membrane protrusions makes ErbB2 an internalization-resistant receptor. Mol Biol Cell 15, 1557–1567. 11 Sottocornola E, Berra B & Colombo I (2003) GM3 con- tent modulates the EGF-activated p185c-neu levels, but not those of the constitutively activated oncoprotein p185neu. Biochim Biophys Acta 1635, 55–66. 12 Parolini I, Topa S, Sorice M, Pace A, Ceddia P, Mon- tesoro E, Pavan A, Lisanti MP, Peschle C & Sargia- como M (1999) Phorbol ester-induced disruption of the CD4–Lck complex occurs within a detergent-resistant microdomain of the plasma membrane. Involvement of the translocation of activated protein kinase C isoforms. J Biol Chem 274, 14176–14187. 13 Lotti LV, Lanfrancone L, Migliaccio E, Zompetta C, Pelicci G, Salcini AE, Falini B, Pelicci PG & Torrisi MR (1996) Shc proteins are localized on endoplasmic reticulum membranes and are redistributed after tyro- sine kinase receptor activation. Mol Cell Biol 16, 1946– 1954. 14 Garofalo T, Lenti L, Longo A, Misasi R, Mattei V, Pontieri GM, Pavan A & Sorice M (2002) Association of GM3 with Zap-70 induced by T cell activation in plasma membrane microdomains: GM3 as a marker of microdomains in human lymphocytes. J Biol Chem 277, 11233–11238. 15 Lotti LV, Di Lazzaro C, Zompetta C, Frati L & Torrisi MR (1992) Surface distribution and internalization of erbB-2 proteins. Exp Cell Res 202, 274–280. 16 Zurita AR, Crespo PM, Koritschoner NP & Daniotti JL (2004) Membrane distribution of epidermal growth factor receptors in cells expressing different gangliosides. Eur J Biochem 271, 2428–2437. 17 Roepstorff K, Thomsen P, Sandvig K & van Deurs B (2002) Sequestration of epidermal growth factor recep- tors in non-caveolar lipid rafts inhibits ligand binding. J Biol Chem 277, 18954–18960. 18 Pike L, Han X & Gross RW (2005) Epidermal growth factor receptors are localized to lipid rafts that contain a balance of inner and outer leaflet lipids. J Biol Chem 280, 26796–26804. 19 Rosner H, Greis CH & Rodemann HP (1990) Density- dependent expression of ganglioside GM3 by human skin fibroblasts in an all-or-none fashion, as a possible modulator of cell growth in vitro. Exp Cell Res 190, 161–169. 20 Wang XQ, Sun P & Paller AS (2002) Ganglioside induces caveolin)1 redistribution and interaction with the epidermal growth factor receptor. J Biol Chem 277, 47028–47034. 21 Puri C, Tosoni D, Comai R, Rabellino A, Segat D, Caneva F, Luzzi P, Di Fiore PP & Tacchetti C (2005) Relationships between EGFR signaling-competent and endocytosis-competent membrane microdomains. Mol Biol Cell 16, 2704–2718. 22 Miljan EA, Meuillet EJ, Mania-Farnell B, George D, Yamamoto H, Simon HG & Bremer EG (2002) Interac- tion of the extracellular domain of the epidermal growth factor receptor with gangliosides. J Biol Chem 277, 10108–10113. 23 Yamabhai M & Anderson RGW (2002) Second cysteine-rich region of epidermal growth factor receptor contains targeting information for caveolae ⁄ rafts. J Biol Chem 277, 24843–24846. 24 Baulida J, Kraus MH, Alimandi M, Di Fiore PP & Carpenter G (1996) All ErbB receptors other than the epidermal growth factor receptor are endocytosis impaired. J Biol Chem 271, 5251–5257. 25 Deb TB, Wong L, Salomon DS, Zhou G, Dixon JE, Gutkind JS, Thompson SA & Johnson GR (1998) A common requirement for the catalytic activity and both SH2 domains of SHP-2 in mitogen-activated protein (MAP) kinase activation by the ErbB family of recep- tors. A specific role for SHP-2 in MAP, but not c-Jun amino-terminal kinase activation. J Biol Chem 273, 16643–16646. 26 Kotani M, Ozawa H, Kawashima I, Ando S & Tai T (1992) Generation of one set of monoclonal antibodies E. Sottocornola et al. ErbB2–raft association in HC11 cells FEBS Journal 273 (2006) 1821–1830 ª 2006 The Authors Journal compilation ª 2006 FEBS 1829 specific for a-pathway ganglio-series gangliosides. Bio- chim Biophys Acta 1117, 97–103. 27 Rodgers W & Rose JK (1996) Exclusion of CD45 inhi- bits activity of p56lck associated with glycolipid- enriched membrane domains. J Cell Biol 135, 1515– 1523. 28 Huber LA, Xu Q, Jurgens G, Bock G, Buhler E, Gey KF, Schonitzer D, Trall KN & Wick G (1991) Correla- tion of lymphocyte lipid composition membrane micro- viscosity and mitogen response in the aged. Eur J Immunol 21, 2761–2765. 29 Svennerholm L & Fredman P (1980) A procedure for the quantitative isolation of brain gangliosides. Biochim Biophys Acta 617, 97–109. 30 Williams MA & McCluer RH (1980) The use of Sep- pak C18 cartridges during the isolation of gangliosides. J Neurochem 35, 266–269. ErbB2–raft association in HC11 cells E. Sottocornola et al. 1830 FEBS Journal 273 (2006) 1821–1830 ª 2006 The Authors Journal compilation ª 2006 FEBS . analyzed the role of gangliosides in the association of the ErbB2 recep- tor tyrosine-kinase (RTK) with lipid rafts in mammary epithelial HC11 cells. Scanning. Role of gangliosides in the association of ErbB2 with lipid rafts in mammary epithelial HC11 cells Elena Sottocornola 1 , Roberta Misasi 2 , Vincenzo