| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 279, Issue 35, 36778-36787, August 27, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
RII Clustering and Phosphorylation in Rafts*
From the Nencki Institute of Experimental Biology, the Department of Cell Biology, 3 Pasteur St., 02-093 Warsaw, Poland
Received for publication, February 26, 2004 , and in revised form, June 9, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
receptor II (FcRII), the mechanism governing translocation of an activated receptor toward the rafts is unknown. We show that at the onset of FcRII cross-linking acid sphingomyelinase is rapidly activated. This enzyme is extruded from intracellular compartments to the cell surface, and concomitantly, exofacially oriented ceramide is produced. Both non-raft and, to a lesser extent, raft sphingomyelin pools were hydrolyzed at the onset of FcRII cross-linking. The time course of ceramide production preceded the recruitment of FcRII to rafts and the receptor phosphorylation. Exogenous C16-ceramide facilitated clustering of FcRII and its association with rafts. In contrast, inhibition of acid sphingomyelinase diminished both the ceramide generation and clustering of cross-linked FcRII. Under these conditions, tyrosine phosphorylation of FcRII and receptor-accompanying proteins was also reduced. All the inhibitory effects were bypassed by treatment of cells with exogenous ceramide. These data provide evidence that the generation of cell surface ceramide is a prerequisite for fusion of cross-linked FcRII and rafts, which triggers the receptor tyrosine phosphorylation and signaling. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
receptors IIA and IIC (FcRII1 A/C) are single chain, 40-kDa members of the FcR family that are expressed exclusively in human cells of the immunodefense system. Upon binding multivalent immunocomplexes of IgG, two conserved tyrosine residues of the receptor signaling motif ITAM are phosphorylated, triggering downstream signaling pathways (1). It was demonstrated that activated FcRII in U937 cells undergoes clustering and is recruited from the glycerophospholipid-rich environment of the plasma membrane to membrane microdomains, rafts, composed mainly of sphingolipids and cholesterol. Within the rafts, protein-tyrosine kinases of the Src family, most likely Lyn, phosphorylate tyrosine residues of ITAM. The receptor phosphorylation triggers a cascade of tyrosine phosphorylation of several downstream proteins, which is required for a local reorganization of the actin cytoskeleton (2–4). During FcRII-mediated phagocytosis, the actin cytoskeleton drives uptake of IgG-coated particles, e.g. pathogens. The actin-based cytoskeleton can also provide the driving force for the movement of FcRII clusters, formed after receptor cross-linking by antibodies, in the plane of the plasma membrane and their accumulation into polar caps (5).
Ample data indicate that all types of ITAM-bearing receptors, including, besides FcRs, B and T cell receptors (TCR), Fc
receptor I (FcRI), and Fc
receptor, are clustered upon activation and associate with rafts to trigger signaling pathways (6, 7). Sophisticated microscopic techniques revealed that membrane rafts are dynamic entities that merge and disassemble to permit spatial organization of multimolecular signaling complexes of TCR and FcRI (8, 9). However, the mechanism governing fusion of activated immunoreceptors with the plasma membrane rafts and subsequent coalescence of the rafts remains unknown.
The interactions between activated receptors and rafts and among the rafts can be facilitated by ceramide, a product of sphingomyelinase (SMase) activity. Ceramide level was found to be elevated in response to several extracellular stimuli (10). Ceramide can fulfill a fusogenic role because of its tendency to self-aggregate in the plane of membranes (11, 12). Compared with sphingomyelin (SM), ceramide has a small, less hydrated head group. It occupies a relatively small area in the monolayers, 40 versus 50–70 Å2 for phospholipids (13). In consequence, a membrane leaflet rich in ceramide condenses, and the ceramide molecules separate laterally into a domain. As ceramide flip-flops very slowly, the sidedness of ceramide formation in the plasma membrane is likely to define its function. Accordingly, generation of ceramide in either the outer or the inner leaflet of liposomes evoked endocytosis or blebbing of their membrane, respectively (14, 15).
Despite the characterization of ceramide behavior in model membranes, neither the topology of ceramide generation in the plasma membrane of living cells nor the identity of SMase catalyzing its release is known. About 80–90% of plasma membrane SM resides in the extracellular leaflet of the membrane (16). The small fraction of SM located in the cytoplasmic leaflet is thought to be hydrolyzed by neutral SMase (NSMase) upon activation of tumor necrosis factor (TNF) receptor family members, TNF receptor and CD40 (17). The cytosolically oriented NSMase activated by the TNF receptor was found to reside in caveolae, raft-related structures. Due to interaction with the caveolin scaffolding domain, the enzyme can be kept in an inactive state in resting cells. TNF stimulation was suggested to evoke relocation of the NSMase from the caveolae, leading to SM hydrolysis in the cytoplasmic leaflet of noncaveolar compartments of the plasma membrane (18). Another SM-hydrolyzing enzyme, acid SMase (ASMase), has been shown recently to generate ceramide in the extracellular leaflet of the plasma membrane. This cell surface-oriented ASMase was found to be activated in T cells upon stimulation of the CD95/Fas receptor, another member of the TNF receptor family. The generated ceramide was required for efficient clustering of the receptor and for generation of apoptotic signals (19, 20). A similar phenomenon was described in B cells for CD40 (21). An involvement of ASMase, but not NSMase, in ceramide generation in response to interleukin 1
was also suggested by an early report coupling the enzyme activity to caveolae (22). These data were not supported, however, by recent studies (18, 23) showing an enrichment of NSMase relative to ASMase activity in these structures. In the light of the few available studies on the compartmentalization of SMases in the plasma membrane, the complex nature of their involvement in receptor signaling is even more obscure.
In our studies we focused on the ASMase activity and the role of generated ceramide in signal transduction by FcRII in plasma membrane rafts. We found that activation of ASMase and ceramide production preceded the recruitment of FcRII to rafts and the receptor tyrosine phosphorylation. Our data indicate that ceramide, released at the onset of FcRII activation, can control efficient receptor clustering and its association with rafts, required for FcRII phosphorylation and generation of signaling cascades.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
RII (BHK-FcIIA) were cultured as described (4). Prior to experiments the cells were washed in Hepes-buffered saline (HBS) containing 125 mM NaCl, 4 mM KCl, 10 mM NaHCO3, 1 mM KH2PO4, 10 mM glucose, 0.2% bovine serum albumin, 20 mM Hepes, pH 7.4, and supplemented with 1 mM CaCl2, 1 mM MgCl2 for BHK-FcIIA cells. To cross-link FcRII, cells were first exposed for 30 min at 0 °C to anti-FcRII mouse IgG, clone IV.3 (ATCC). Subsequently, goat or rabbit anti-mouse IgG, as indicated, was applied for 1–30 min at 0 °C. Warming the cells for 10 min at 20 °C led to assembly of FcRII caps. The cells were fixed with 3% formaldehyde, and the capping efficiency was estimated by using a Nikon fluorescence microscope (5). To inhibit ASMase activity, U937 cells (3 x 106/ml) and cultures of BHK-FcIIA cells (3 x 104/coverslip/sample) were pretreated for 1 h at 37 °C with 1 mM chloroquine (Sigma) or 50 µM imipramine (Sigma) or with 0.025% N,N-dimethylformamide (imipramine carrier) in HBS. In a series of experiments, after removal of the inhibitors and prior to FcRII cross-linking, exogenous 5 µM C16-ceramide (Alexis) was added to the cells for 5 min at 20 °C from dodecane/ethanol (2:98 v/v) stock solution. The level of plasma membrane cholesterol in U937 cells was depleted with the use of 5 mM -cyclodextrin (CDX, Sigma) for 1 h at 37 °C, as described (3). Cell viability was monitored by trypan blue exclusion. CD55 and transferrin receptor (TfR) were cross-linked on the surface of living U937 cells (1.2 x 107/ml HBS) with rabbit anti-CD55 IgG and anti-TfR IgG (Santa Cruz Biotechnology) applied for 30 min at 0 °C followed by goat anti-rabbit IgG (Jackson ImmunoResearch) for 1–30 min at 0 °C. Concanavalin A (ConA, Sigma) was added to cells at 100 µg/ml HBS for 1–30 min at 0 °C.
ASMase Activity Assay—ASMase activity was estimated using NBD C6-sphingomyelin as a substrate (24). For this, cells (3 x 106/sample) were lysed for 15 min at 0 °Cin120 µl of buffer consisting of 1% Triton X-100, 1 mM Na3VO4, 50 mM Tris, pH 7.4, and a protease inhibitor mixture (Sigma) and sheared by passing through an 18-gauge needle. After clarification (2 min, 4 °C, 800 x g), 100 µl of supernatant were mixed with 150 µl of ASMase assay buffer containing 1.3 mM EDTA, 0.05% Nonidet P-40, 50 µM HgCl2, 250 mM MES, pH 5.5, and liposomes of 10 µM NBD C6-sphingomyelin (Molecular Probes) and 30 µM 1,2 dioleoyl-sn-glycerol-3-phosphocholine (Sigma). The reaction was carried out for 2 h at 37 °C and stopped by adding a mixture containing 0.45 ml of 2-propanol, 1.5 ml of heptane, and 0.2 ml of H2O. A collected upper phase was mixed with 0.35 ml of H2O, centrifuged (1 min, 20 °C, 2,000 x g), and analyzed on a Spex spectrofluorimeter (Yobin-Yvon) at 460/515 nm.
Cell Surface Ceramide Determination—U937 cells (3 x 106/sample) and BHK-FcIIA cells (3 x 104/coverslip/sample) were fixed with 1% formaldehyde in HBS without albumin (15 min, 0 °C) at various points of receptor cross-linking, washed with PBS, and probed for 45 min at 20 °C with 2 µg/ml of anti-ceramide mouse IgM (clone MID 15B4, Alexis). Alternatively, the cells were exposed to 4 µg/ml probe containing the cysteine-rich domain of kinase suppressor of RAS fused with glutathione S-transferase (CRD/
KSR-GST, kindly provided by Dr. E. Gulbins). The probe was shown previously to recognize ceramide (25, 26), and its specific reactivity with ceramide among various lipids spotted onto nitrocellulose was confirmed by us. The cells were washed with PBS and treated for 20 s with ice-cold 100 mM glycine-HCl buffer, pH 3.0. The cells were pelleted (1 min, 4 °C, 2,000 x g), and the supernatants were neutralized with 1 M Tris, pH 8.0. Subsequently, 5 µl of the supernatants were spotted onto nitrocellulose sheets and air-dried. The sheets were blocked for 2 h at room temperature in 1% polyvinylpyrrolidone, 1% gelatin (Sigma) in Tris-buffered saline containing 0.05% Tween 20 and incubated for 1 h at room temperature with goat anti-mouse IgM conjugated with peroxidase (Sigma) or with rabbit anti-GST (Sigma) and anti-rabbit IgG peroxidase (Roche Applied Science). Immunoreactive spots were visualized with the SuperSignal West Pico chemiluminescent substrate (Pierce) and quantified densitometrically in a Fluor-S MultiImager using Quantity One software (Bio-Rad). For normalization, the densitometric data were expressed in relation to the ceramide level detected in resting cells and arbitrarily equalized to 1.
Estimation of the Size of FcRII Clusters—Two methods were used to measure clusters formed by cross-linked FcRII (1). Estimation of the area of clusters that were visualized by conventional fluorescence microscopy after application of mouse IV.3 anti-FcRII and goat anti-mouse IgG-FITC for FcRII cross-linking. Fluorescent cell images were taken at the same exposure time, digitized, and analyzed with Quantity One software (Bio-Rad). For this, areas of the clusters were carefully delineated, and overlapping patches at the cell margins were excluded from the analysis. At least 15 cells from two independent experiments were analyzed in each variant, and 40–50 clusters were scored in every cell (2). For size exclusion chromatography, cells (2 x 106/sample) were exposed to biotin-labeled mouse IV.3 anti-FcRII and goat anti-mouse IgG (30 min at 0 °C each) to cross-link the receptor. The cells were lysed in 250 µl of 0.5% Triton X-100, 0.5% Nonidet P-40, 100 mM NaCl, 2 mM EDTA, 2 mM EGTA, 1 mM Na3VO4, 10 mM NaF, 10 mM sodium pyrophosphate, 50 µM phenylarsine oxide, 30 mM Hepes, pH 7.4, and a protease inhibitor mixture (0 °C). After shearing through an 18-gauge needle, lysates were centrifuged (5 min, 4 °C, 5,000 x g), and 200 µl of obtained supernatants were loaded onto a Sepharose 4B column (0.9 x 10 cm, total volume 6.3 ml). The column was developed at 4 °C with the lysis buffer; 30 fractions of 200 µl were collected and analyzed for the distribution of the biotin-conjugated IV.3 antibody by 10% SDS-PAGE and immunoblotting with anti-biotin goat IgG-peroxidase (Sigma). For calibration of the column the following size markers were applied: blue dextran (2 x 106 Da), thyroglobulin (0.67 x 106 Da), and rabbit IgG (0.15 x 106 Da).
Confocal Microscopy—To study colocalization of FcRII, ASMase, and ceramide, U937 cells (3 x 106/sample) were fixed either directly after washing with HBS (unstimulated cells), after incubation with mouse IV.3 anti-FcRII IgG (30 min, 0 °C), or after consecutive exposure to anti-FcRII and anti-mouse rabbit IgG (1–30 min at 0 °C), leading to receptor cross-linking, and finally after FcRII capping for 10 min at 20 °C. Cells were fixed with 1% formaldehyde in HBS without albumin (15 min, 0 °C), washed with PBS, and in the case of cells pretreated with anti-FcRII only, incubation with rabbit anti-mouse IgG was conducted next (30 min at room temperature). The rabbit anti-mouse IgG was conjugated with FITC when colocalization of FcRII with ASMase was examined and with lissamine rhodamine during ceramide distribution studies (Jackson ImmunoResearch). To reveal ASMase on the cell surface, fixed cells were incubated for 45 min with goat anti-human ASMase antibodies (kindly provided by Dr. K. Sandhoff) followed by donkey anti-goat lissamine rhodamine-conjugated IgG (Jackson ImmunoResearch). The specificity of the anti-ASMase antibodies was demonstrated by previous studies (27, 28). In our hands, the antibodies recognized the major band of 70–72-kDa ASMase and its proteolytic derivatives of 
55 kDa, when tested on whole lysates of U937 cells. For ceramide visualization, anti-ceramide mouse IgM and goat anti-mouse IgM-FITC were applied (Sigma). Samples were mounted in Mowiol/DABCO (diazabicyclo[2.2.2]octane, Sigma). Images were collected using a Leica confocal microscope in the mode of sequential excitation of FITC and rhodamine dyes to exclude crossover of their fluorescence. Colocalization of FcRII clusters with patches of ASMase and ceramide was estimated as described (4).
Isolation of Detergent-resistant Membranes; Immunoprecipitation of FcRII—FcRII was either cross-linked in U937 cells at 0 °C with mouse IV.3-biotin anti-FcRII IgG (30 min) and rabbit anti-mouse IgG (3–30 min) or left untreated with any antibody (unstimulated cells). In a series of experiments, the distribution of cell surface ceramide or SM in density gradient fractions was examined by exposing living cells to 2 µg/ml anti-ceramide IgM and 10 µg/ml of nonlytic lysenin fused with polyhistidine (lysenin-His), respectively, for 3 min at 0 °C prior to washing with HBS and lysis. Lysenin was purified from Escherichia coli transformed with cDNA that was synthesized in our laboratory. In the nonlytic form of lysenin, tryptophan 20 was replaced by alanine, as described previously (52). It was proved by us that the mutant protein recognized SM and bound to erythrocytes but did not evoke hemolysis.2 To assess the influence of ceramide on FcRII fusion with rafts, 5–10 µM C16-ceramide was added for 5 min at 0 °C to cells pretreated with IV.3 antibody; the ceramide was also present during next 3 min of cell incubation with anti-mouse IgG (0 °C). Detergent-resistant membranes were isolated from cells essentially as described earlier (4, 29). In brief, cells (12.5 x 106 per sample) were lysed for 30 min at 0 °C in 200 µl of buffer containing 0.2% Triton X-100, 100 mM NaCl, 2 mM EDTA, 2 mM EGTA, 30 mM Hepes, pH 7.5, and the following phosphatase inhibitors: 1 mM Na3VO4, 50 µM phenylarsine oxide, 30 mM p-nitrophenyl phosphate, and a mixture of protease inhibitors. After clarification (1.5 min, 4 °C, 480 x g), the lysates were supplemented with 40% Optiprep and 10% sucrose (600 µl total volume), overlaid with 30, 25, 20 (400 µl of each), and 0% Optiprep (300 µl) in 0.2% Triton X-100 lysis buffer containing 10% sucrose and spun for 3 h at 170,000 x g, 4 °C (RCM 100 ultracentrifuge Sorvall). For protein phosphorylation studies, 4 x 107 cell were lysed in 600 µl of the 0.2% Triton X-100 lysis buffer containing 40% Optiprep and 10% sucrose. Seven fraction of 300 µl were collected from the top of the gradient.
To precipitate FcRII, 200 µl of each fraction were diluted twice with 0.2% Triton X-100 lysis buffer and supplemented with 30 µl of 10% Pansorbin (Calbiochem). In the case of gradient fractions obtained from resting cells, Pansorbin beads with covalently bound IV.3 mouse antibody were prepared with the use of an IgG orientation kit (Pierce) according to the manufacturer's instruction. Immunoprecipitation was carried out overnight at 4 °C, followed by washing of precipitates with 0.2% Triton X-100 lysis buffer and boiling in 2x SDS sample buffer.
Biotinylation of Cell Surface Proteins—At various time points of FcRII cross-linking, U937 cells (3 x 106 per sample) were resuspended in 1 ml of PBS and subjected to biotinylation of cell surface proteins as described (30). Briefly, 100 µg/ml of sulfosuccinimidyl-6-biotin-amidohexanoate (Pierce) was added to cell suspensions for 30 min (4 °C). After washing in PBS, the cells were lysed in 200 µl of 0.2% Triton X-100 buffer (30 min, 4 °C) and clarified (1.5 min, 4 °C, 480 x g). Biotinylated proteins were isolated from lysates by streptavidin-coated magnetic beads (Sigma) (1 h, 4 °C). Washed beads were resuspended in 200 µlof2x SDS sample buffer and boiled, and 15 µl of the probe were subjected to SDS-PAGE. In parallel, 20 µl of the nonpooled fraction of the Triton X-100 lysates were also analyzed by SDS-PAGE.
Immunoblotting—Proteins were separated by 10% SDS-PAGE, transferred to nitrocellulose, and probed with the appropriate antibodies as described previously (3). The following antibodies were applied: mouse anti-phosphotyrosine, clone PY66 (Santa Cruz Biotechnology), mouse anti-actin (Roche Applied Science), and goat anti-mouse IgG-peroxidase (Santa Cruz Biotechnology). Distribution of FcRII in cell fractions was revealed either by detection of biotin-labeled IV.3 anti-FcRII, bound to intact cells, by goat anti-biotin IgG-peroxidase (Sigma) or with the use of rabbit anti-FcRII (kindly provided by Dr. J.-L. Teillaud) followed by goat anti-rabbit IgG peroxidase (Santa Cruz Biotechnology). Ceramide and SM were localized in the gradient fractions by detection of anti-ceramide IgM and lysenin-His, respectively, and bound to cells prior to lysis, with the use of goat anti-mouse IgM peroxidase and rabbit anti-His (Santa Cruz Biotechnology) followed by anti-rabbit IgG peroxidase. To reveal ASMase, goat anti-ASMase and rabbit anti-goat IgG labeled with peroxidase (Sigma) were applied. Immunoreactive bands were visualized by chemiluminescence and quantified densitometrically as described above. The relative phosphorylation of FcRII was expressed in relation to the maximum phosphorylation that was arbitrarily equalized to 1.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
RII Induces Cell Surface Ceramide Production—To assess whether activation of FcRII involves the action of ASMase, the enzyme activity was determined in U937 cell lysates derived at various time points of the receptor crosslinking (Fig. 1A). After binding of mouse anti-FcRII (clone IV.3) alone (30 min, 0 °C), ASMase activity increased 1.5-fold and did not change significantly during prolong incubation of cells with the antibody (up to 50 min at 0 °C). On the contrary, cross-linking of the receptor with anti-FcRII followed by anti-mouse IgG induced further transient elevation of ASMase activity (2.5-fold), peaking at 5 min after the addition of the anti-mouse IgG. The profile of ASMase activation at the onset of FcRII cross-linking was correlated with the generation of cell surface ceramide (Fig. 1, B and C). The amounts of cell surface ceramide were quantified by immunoblot detection of an anti-ceramide antibody (clone MID 15B4) bound to intact cells and next released by acidic washing. The same pattern of ceramide appearance was found with the use of a CRD/KSR-GST probe (Fig. 1B). The ceramide visualized in the cells after 5 min of FcRII cross-linking formed abundant small dots dispersed in the plane of the plasma membrane and occasionally fused into large clusters (Fig. 1, compare D and E). After subsequent reduction of the ceramide level for the next 5 min, at 20 min of FcRII cross-linking both ceramide-targeting probes again detected an elevated level of ceramide, which formed scattered clusters on the cell surface (Fig. 1, B, C, F, and G). The second peak of ceramide appearance could result either from an increased availability of a new pool of SM to ASMase or from activity of NSMase, because at this time point of FcRII cross-linking a decrease of ASMase activity took place (Fig. 1A). The former possibility seems more likely because no significant changes of the cell surface ceramide level were detected in cells pretreated with 50 µM imipramine (Fig. 1C, open circles), considered to act as an ASMase inhibitor (see also Fig. 7). A biphasic pattern of ceramide appearance was detected also on the plasma membrane surface in BHK-FcIIA, cells transfected with FcRIIA, during the receptor cross-linking. In these cells, however, ceramide was produced more rapidly than in U937 cells, reaching a maximum after 3 min of FcRII cross-linking followed by a second peak after 10 min (Fig. 1C, stars). Pretreatment of U937 cells with 5 mM CDX, which depleted the plasma membrane cholesterol by 60–70% (3), led to a 2.2-fold elevation of cell surface ceramide generation during FcRII cross-linking (Fig. 1H, , open circles).
|
|
RII cross-linking, the ceramide level was estimated in U937 cells after cross-linking of TfR and CD55. Stimulation of TfR, which is confined to glycerophospholipid-rich environment of the plasma membrane and does not enter membrane rafts upon clustering, induced robust and prominent ceramide production within the 1st min of cross-linking (Fig. 1I, stars). In contrast, only minute amounts of ceramide were detected during cross-linking of CD55, a plasma membrane raft constituent, when compared with the ceramide appearance induced by FcRII and TfR cross-linking (Fig. 1I, squares). Cross-linking of bulk plasma membrane glycoproteins and glycolipids by ConA yielded traces of ceramide (Fig. 1I, triangles).
Colocalization of Clustered FcRII, ASMase, and Ceramide—The activation of ASMase and generation of ceramide on the cell surface induced by FcRII cross-linking prompted us to examine the relationship between the distribution of these membrane constituents. In resting cells, not exposed to any antibody before fixation, only faint labeling of ASMase on the cell surface was detected (Fig. 2A). Exposure of the cells to anti-FcRII alone led to an increase of ASMase staining; ASMase was dispersed in the plane of the plasma membrane, resembling the diffuse distribution of the receptor (Fig. 2, B1–B3). A further prominent elevation of ASMase level on the cell surface ensued from FcRII cross-linking by mouse anti-FcRII and anti-mouse IgG (Fig. 2, C1–D3). At the early stages of FcRII cross-linking (5 min, 0 °C), the enzyme was dispersed on the cell surface as well as formed small aggregates, which colocalized with FcRII clusters (Fig. 2, C1–C3). The degree of ASMase aggregation and its colocalization with FcRII increased during receptor activation as over 95% of the receptor clusters colocalized with the ASMase patches after 20 min of FcRII cross-linking at 0 °C (Fig. 2, D1–D3). Clear colocalization of FcRII and ASMase was maintained during subsequent translocation of the receptor clusters toward one pole of the cell and their assembly into caps, although a fraction of ASMase remained scattered outside the cap region (Fig. 2, E1–E3). In cells pretreated with 50 µM imipramine, an inhibitor of ASMase activity, diffuse staining of ASMase on the cell surface was visible (Fig. 2, F1–F3). Under these conditions, ceramide production was abolished (see Fig. 1C, open circles), and clustering of FcRII was impaired (see also Fig. 8).
|
|
RII and ceramide revealed that ceramide clusters exceeded in number the clusters of the receptor. Nevertheless, about 65% of FcRII patches colocalized, at least partially overlapping, with the ceramide clusters (Fig. 3, A1–A3). The colocalization of ceramide and the cross-linked receptor was transient; during the receptor capping the ceramide level decreased and the lipid was absent from the cap region (Fig. 3, B1–B3). We found that decreasing the pH during FcRII cross-linking to 5.5, the optimum pH for ASMase activity, markedly improved the colocalization of FcRII and ceramide clusters (Fig. 3, C1–C3).
|
RII Phosphorylation in Rafts—As FcRII associates with rafts upon cross-linking for signal generation, the distribution of ASMase and ceramide in Optiprep density gradient was next examined. In resting cells, small amounts of ASMase and no traces of ceramide were found in the low density raft fractions 1–2 of the gradient, where the majority of Lyn kinase, a raft marker, was recovered (Fig. 4A). As expected, unstimulated FcRII was absent in the low density fractions, accumulating in fractions 5–6, together with the main pool of ASMase (Fig. 4B, compare with A). Upon FcRII cross-linking the receptor was progressively shifted toward lower density fractions reaching a maximum of accumulation in fractions 1–2 after 10 min of activation. The time course of FcRII enrichment in the raft fractions 1–2 was correlated with increasing receptor tyrosine phosphorylation peaking between 10 and 20 min after receptor cross-linking (Fig. 4B). After 10 min of the receptor cross-linking, the tyrosine phosphorylation level of multiple accompanying proteins also rose markedly (Fig. 4C).
|
RII cross-linking evoked rapid elevation of ceramide level throughout the gradient fractions, particularly in fractions 3–4 as well as in raft fractions 1–2, already after 5 min of receptor activation (Fig. 4A). Simultaneously, an enrichment of ASMase in fractions 1–2 at the expense of the higher density fractions was detected (Fig. 4A). In addition, at the onset of FcRII cross-linking ASMase translocated from intracellular compartments to the cell surface thus becoming susceptible to biotinylation in intact cells (Fig. 4D). The distribution of ASMase in the gradient fractions did not change significantly during the later steps of FcRII cross-linking (Fig. 4A), whereas the level of cell surface ceramide underwent fluctuations reflecting the biphasic pattern of its generation as demonstrated in Fig. 1 by quantitative and microscopic data. After the first peak was found 5 min after FcRII cross-linking, the ceramide level diminished markedly 5 min later to rise again after 20 min of receptor activation. At this stage, the generated ceramide was detected mainly in fractions 3–4, and traces of the lipid were found also in fractions 1–2 of the density gradient (Fig. 4A).
Cell surface ceramide production also preceded FcRIIA phosphorylation induced in the BHK-FcIIA transfectants. FcRIIA cross-linking triggered robust ceramide appearance, which peaked within 3 min of receptor activation and was followed by gradual FcRIIA phosphorylation during 5–20 min of its activation (Fig. 4E).
Taken together, the data imply that ceramide generation facilitates FcRII phosphorylation by enabling fusion of activated FcRII with rafts. To examine this possibility further, the influence of ceramide onto partitioning of FcRII to detergent-resistant membrane fraction was analyzed. Only traces of FcRII were detected in raft fractions 1–2 of the density gradient after 3 min of FcRII cross-linking (Fig. 5A). In parallel, no clusters of the receptor were visible on the cell surface under these conditions (Fig. 5B). However, when FcRII cross-linking was conducted in the presence of an exogenous C16-ceramide, the ceramide facilitated association of the cross-linked FcRII with raft fractions in a dose-dependent manner. Concomitantly, clustering of the receptor was markedly augmented, as revealed by confocal microscopy (Fig. 5, A and B).
|
RII cross-linking was derived from raft or non-raft SM, the level and distribution of cell surface SM in density gradient fractions of Triton X-100 lysates of U937 cells were examined. For this purpose, nonlytic lysenin-His was used as a specific SM probe. A unique SM-binding property of lysenin was discovered by Kobayashi's group (see Ref. 31 and for a review see Ref. 32). Two pools of cell surface SM were detected in resting cells, with one of them occupying fractions 5–7 and the second fractions 2–3 of lower density. Cross-linking of FcRII for 5 min affected both pools of the lipid leading to its reduction in fraction 2 and a profound depletion of SM in fraction 5. Prolonged cross-linking of FcRII (10–20 min) led to a faint enrichment of SM in gradient fractions 4–5 (Fig. 6).
|
RII Signaling—To investigate the importance of ASMase activity for initiation of FcRII signaling pathways, the effect of chloroquine and imipramine, two putative ASMase inhibitors, on tyrosine phosphorylation of FcRII and accompanying proteins was next assessed. Cross-linking of FcRII evoked intensive tyrosine phosphorylation of the receptor, as shown in Fig. 7, A and B, for the receptor expressed in BHK cells (see also Fig. 4 for FcRII phosphorylation in U937 cells). In U937 a distinct subset of proteins was identified to undergo tyrosine phosphorylation aside from the receptor (Fig. 7C, arrows). Chloroquine at 1 mM and imipramine at 50 µM significantly reduced phosphorylation of FcRII and the accompanying proteins. After incorporation of 5 µM C16-ceramide into chloroquine- or imipramine-treated cells, partial reconstitution of the protein phosphorylation was observed (Fig. 7, A–C). Reversion of the inhibitory effects of chloroquine and imipramine by exogenous C16-ceramide indicated that the drugs could target ASMase. It was shown that chloroquine and imipramine do not impair the catalytic properties of ASMase directly and selectively but can exhibit inhibitory action toward the enzyme indirectly, by an influence on ASMase trafficking and processing (33, 34). To examine if chloroquine and imipramine action onto FcRII-dependent protein phosphorylation can be attributed to ASMase inhibition, the enzyme activity was estimated in lysates derived from cells pretreated with the drugs (Fig. 7D). It was found that chloroquine and imipramine inhibited total cellular ASMase activity by 50–60% after 1 h of treatment. As simultaneously cell surface ceramide production was abrogated (Fig. 1C), ASMase extruded to the cell surface seemed to be especially sensitive to chloroquine and imipramine action.
Immunofluorescence observations of FcRII in intact U937 cells showed that chloroquine and imipramine hampered the clustering of the cross-linked receptor. A quantitative analysis of the area of the FcRII clusters confirmed the reduction of their size by about 50 and 70% under the influence of chloroquine and imipramine, respectively (Fig. 8, A and B). Accordingly, assembly of FcRII caps was substantially diminished under these conditions (Fig. 8C). The effect exerted by chloroquine and imipramine on FcRII clustering was further evaluated by determination of the size of the receptor clusters isolated from Triton X-100 cell lysates by gel filtration on Sepharose 4B. In cells not treated with the drugs, FcRII cross-linked by antibodies formed complexes eluted from the column in a sharp peak comprising fractions 11–13. These fractions contained 90% of the receptor, judging from the presence of FcRII-bound IV.3 antibody. Pretreatment of the cells with 50 µM imipramine diminished the size of the FcRII clusters, which now were eluted from the column mainly in fractions 12 and 13 and in a broad tail of fractions 15–23. A similar effect on FcRII cluster size was exerted by 1 mM chloroquine (Fig. 8D).
After incorporation of exogenous C16-ceramide into cells pretreated with chloroquine and imipramine, the size of the clusters of cross-linked FcRII was found to be restored reaching about 80% of the control value, as estimated by area measurements (Fig. 8B). Similarly, the elution profile of FcRII clusters derived from cells treated with imipramine or chloroquine followed by C16 incorporation closely mirrored the receptor distribution in fractions derived from cells not exposed to the drugs or exposed to C16-ceramide only (Fig. 8D and results not shown). The restitution of FcRII clustering after incorporation of C16-ceramide led to 50% recovery of FcRII capping in the cells (Fig. 8C). Taken together, the results show that ASMase activity and ceramide production are required for efficient FcRII clustering and tyrosine phosphorylation of the receptor and accompanying proteins.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
RII with sphingolipid-cholesterol plasma membrane rafts plays a crucial role in the generation of receptor signaling cascades, the force driving the FcRII translocations from the glycerophospholipid environment to the rafts is unknown. In this study we demonstrate that at the onset of FcRII activation cell surface ceramide is generated by ASMase. The ceramide production is a prerequisite for association of cross-linked FcRII with rafts and the following receptor phosphorylation.
To reveal cell surface ceramide, two different probes were applied, an anti-ceramide antibody (clone MID 15B4) and a cysteine-rich domain of KSR. With this approach changes of ceramide level on the surface of intact cells during FcRII activation were quantified. In addition, the concomitant clustering of ceramide was visualized under a fluorescence microscope. Because of the very slow flip-flop rate of ceramide in membranes (35), the exofacially oriented lipid was likely to be produced from SM located also in the outer leaflet of the plasma membrane. The following line of evidence indicates that the cell surface ceramide was generated by ASMase: 1) rapid translocation of ASMase from intracellular compartments to the cell surface revealed by immunofluorescence and biotinylation of cell surface proteins; 2) correlation of ASMase appearance on the cell surface and exofacial ceramide production; and 3) abrogation of ceramide production under the influence of chloroquine and imipramine, two putative ASMase inhibitors. These data also suggest that the main SM pool utilized during FcRII activation is located in the outer leaflet of the plasma membrane. Accordingly, probing of the cell surface SM by a nonlytic lysenin mutant revealed a significant depletion of SM in the exoplasmic leaflet of the plasma membrane at the onset of FcRII cross-linking. The topology of the SM cycle is the subject of ongoing discussion, complicated by data showing that various stimuli activate different SMases (36). To this end, Linardic and Hannun (37) argued that the signaling pool of SM hydrolyzed during TNF and vitamin D3 treatment of U937 and HL-60 cells was located mainly in the plasma membrane, but either in its inner leaflet or in close vicinity. TNF activates both NSMase and ASMase in U937 cells (38). The assumption that SM utilized during this stimulation came from an internal source was based on the resistance of the signaling SM pool to the action of exogenous bacterial SMase (37). In our hands, pretreatment of U937 cells with bacterial SMase and subsequent cross-linking of FcRII had synergetic effects in terms of cell surface ceramide generation.3 These results imply that the signaling pool of SM hydrolyzed by ASMase during FcRII activation is primarily cell surface-oriented, being simultaneously hidden from bacterial SMase action. The identity of ASMase activated by FcRII is unknown, although a relationship of the enzyme to a secretory form of ASMase that has been cloned (39, 40) can be considered. The mechanism of ASMase extrusion is likely to belong to nonclassical export pathways, which mediate secretion of some cytokines, growth factors, and enzymes (41). Our preliminary electron microscopy data indicate that the transmembrane transport of ASMase may be triggered by ceramide produced within seconds after FcRII activation by NSMase from a minor SM pool located in the inner leaflet of the plasma membrane. Because of its conical shape, ceramide oriented intracellularly can favor membrane blebbing (15), providing gates for ASMase extrusion.
Because of the association of activated FcRII with plasma membrane rafts, the relationship of the receptor signaling pool of SM to rafts is of special interest. The aforementioned SM depletion, detected by lysenin during FcRII cross-linking, was more pronounced in the non-raft fraction 5 than in buoyant raft gradient fraction 2 (Fig. 6). The SM pool that remained after FcRII cross-linking colocalized greatly with CD55, a raft maker, as established by immunofluorescence studies (not shown). After CDX treatment, known to disintegrate rafts by cholesterol depletion (4, 42), two times higher amounts of ceramide were released in cells during FcRII cross-linking, although under these conditions activated FcRII failed to shift from the high to the low density fractions. In line with these data is the finding that clustering of TfR, a raft-excluded receptor, induced even more robust ceramide production than activated FcRII. On the other hand, only minute amounts of ceramide ensued from cross-linking of the raft-residing CD55. Therefore, it seems likely that SM located outside rafts is converted to ceramide upon receptor clustering. The final destination of the activated receptors, e.g. rafts and coated pits, is plausibly determined by intrinsic properties of the receptors, including the structure of the transmembrane region and/or interactions with adaptor proteins (43). The suggestion that during activation of the receptors mainly the non-raft SM is hydrolyzed is in agreement with the data showing that the SM pool involved in TNF receptor signaling is soluble in 1% Triton X-100 and does not pellet at 12,500 x g (37). Studies of lipid distribution in the plasma membrane gave variable estimates on SM content in rafts. Although in some reports all plasma membrane SM was attributed to rafts/caveolae (44, 45), in others only 40–60% of the plasma membrane SM was allocated there (23, 46). The latter results correspond to our estimation of SM content in rafts, based on lysenin binding and cell fractionation. The relatively abundant non-raft pool of SM in U937 cells can serve as a substrate for ASMase activated by FcRII. Nonetheless, already after 5 min of FcRII cross-linking, ceramide and ASMase accompanied the receptor found in rafts. Concomitant hydrolysis of raft SM, detected by the lysenin-based assay, should also be taken into consideration.
It is of note that in our studies Triton X-100 lysates of U937 cells were fractionated over density gradients containing the detergent (0.2% Triton X-100). As we have shown previously, under these circumstances only constituents interacting strongly within rafts, hence resistant to detergent intercalation, are isolated in buoyant gradient fractions 1–2. Membrane constituents of lower affinity to rafts are recovered in gradient fractions 3–5 of an intermediate density (29). This was reinforced by further studies showing that in the absence of Triton X-100 during gradient fractionation, distinct plasma membrane proteins and lipid, including FcRII and SM, were shifted from the intermediate fractions to buoyant density gradients fractions 1–2 (see Ref. 29 and results not shown). At the onset of FcRII cross-linking (5 min), SM from fraction 5 was avidly hydrolyzed, although generated ceramide was located in buoyant and intermediate density fractions 1–4. Similarly, an elevation of ceramide levels occurring after 20 min of receptor cross-linking was detected mainly in fraction 3–4 of the density gradients. Taking into consideration the influence of Triton X-100 on fractionation of distinct membrane constituents, like FcRII and SM, we assume that the intermediate density gradient fractions contain membrane constituents of lower affinity to rafts and are loosely associated with them, being located plausibly at the raft borders.
Ceramide generated at the onset of FcRII cross-linking enables fusion of the receptor with rafts, and the fusion, in turn, is required for FcRII phosphorylation. This suggestion is supported by our findings that exogenous C16-ceramide incorporated into living cells facilitated clustering of FcRII and its partition to detergent-resistant membrane fractions. Accordingly, inhibition of ceramide production by chloroquine and imipramine reduced significantly the clustering of FcRII and phosphorylation of the receptor and accompanying proteins. These inhibitory effects were reversed by incorporation of C16-ceramide to the drug-treated cells, indicating that chloroquine and imipramine action can be attributed to their impact on ASMase activity rather than inhibition of other molecules. This was confirmed by our data showing that 1 mM chloroquine and 50 µM imipramine inhibit ASMase activity in a time-dependent manner in U937 cells and are in agreement with the studies of Hauck et al. (47) demonstrating strong reduction of ASMase activity by imipramine in JOSK-M, another myelomonocytic cell line. Specificity of chloroquine and imipramine as ASMase inhibitors is of concern because the drugs do not block catalytic properties of the enzyme. Instead, as cationic amphiphilic agents, they cause reduction of ASMase activity in living cells only, affecting vesicle-dependent trafficking and processing of the enzyme (33). Studies on imipramine action are linked to those on ASMase processing in cells. The enzyme is synthesized as a 75-kDa precursor polypeptide that is either processed in lysosomes to the 70-kDa mature form of ASMase or processed differently and secreted in small amounts in the medium (39). The lysosomal ASMase was found susceptible to proteolytic cleavage evoked by imipramine, and similar influence of the drug on the secreted fraction of the enzyme cannot be excluded (34). Cell surface ASMase in imipramine-treated U937 cells apparently lost its activity, despite that its extrusion to the cell surface during FcRII cross-linking was not abrogated, as demonstrated by our confocal microscopy studies. The other compound used to interfere with ASMase activity, chloroquine, is a weak base known to induce secretion of lysosomal precursor proteins, to which ASMase belongs. This seems to affect the processing of ASMase, preventing an achievement of catalytic activity by the enzyme (39).
The activity of cell surface ASMase leading to ceramide release was found to be a prerequisite for clustering and signal transduction by two receptors of the TNF family, CD95 and CD40 of T and B cells, respectively (19, 20). The involvement of ASMase in CD95 and CD40 clustering resembles the mechanism that drives accumulation of FcRII in membrane rafts. Based on confocal microscopy data showing colocalization of CD95 and CD40 conglomerates with cell surface ceramide, ASMase, and cholera toxin (used as a raft marker), the authors concluded that membrane rafts were the sites of ceramide generation. It should be noted, however, that the CD95 and CD40 conglomerates were large receptor assemblies corresponding to caps, the final destination of preformed clusters of receptors that merge at one pole of the cell. In contrast, cross-linking of FcRII at 0 °C allowed us to dissect antibody-driven receptor clustering in the plane of the plasma membrane from the subsequent actin-driven translocation of these receptor clusters toward the cap region induced by cell warming. Under these conditions, dynamic changes of ASMase and ceramide distribution in the plasma membrane were revealed in U937 cells by both confocal microscopy and gradient fractionation. Therefore, examination of the late stages of receptor activation only, without analyzing the time course of the process, may lead to the misleading inference that the SM cycle is confined to rafts/caveolae (20, 48). To this end, recent studies (23) on the distribution of the specific activity of ASMase and NSMase in density gradient fractions revealed that the activity of the enzymes was high in both low and intermediate density fractions.
In summary, we assume that the primary event triggered by FcRII cross-linking is the translocation of ASMase from intracellular compartments to the cell surface. The enzyme hydrolyzes SM of the outer leaflet of the plasma membrane leading to ceramide production. This process seems to take place mainly outside rafts, which is in agreement with the non-raft location of nonactivated FcRII. Once released, ceramide can self-associate into small platforms trapping cross-linked FcRII, possibly also the cell surface ASMase. Furthermore, the poor miscibility of ceramide with glycerophospholipids such as phosphatidylethanolamine and phosphatidylcholine (11, 49) can facilitate fusion of the ceramide/receptor clusters and rafts. This process can be enhanced by ceramide generated by ASMase from SM located in the rafts. As a result, ASMase, ceramide, and FcRII are quickly compartmentalized into rafts where the receptor is phosphorylated by the raft-residing Lyn kinase. According to this scenario, FcRII phosphorylation follows ceramide production. Simultaneously, coalescence of rafts leads to formation of a multimolecular signaling complex at the activated receptor. Such merging of small rafts of varied protein composition into 200–300-nm platforms, harboring signaling molecules including Lyn kinase, and located at their margins/phosphatidylinositol 4,5-bisphosphate, was revealed by ultrastructural studies during FcRII activation (50). Similar raft coalescence has been demonstrated to accompany activation of FcRI and TCR (9, 51).
|
FOOTNOTES |
|---|
To whom correspondence should be addressed. E-mail: a.sobota{at}nencki.gov.pl.
1 The abbreviations used are: FcRII, Fc receptor II; ASMase, acid sphingomyelinase; BHK-FcIIA, BHK cells stably transfected with FcRIIA; CDX, -cyclodextrin; ConA, concanavalin A; CRD/KSR-GST, cysteine-rich domain of kinase suppressor of RAS fused with glutathione S-transferase; FcRI, Fc receptor I; HBS, Hepes-buffered saline; lysenin-His, lysenin fused with polyhistidine; NSMase, neutral sphingomyelinase; SM, sphingomyelin; SMase, sphingomyelinase; TCR, T cell receptor; TfR, transferrin receptor; TNF, tumor necrosis factor; MES, 4-morpholineethanesulfonic acid; PBS, phosphate-buffered saline; FITC, fluorescein isothiocyanate.
2 A. Sobota and A. B. Abel Shakor, manuscript in preparation.
3 A. B. A. Shakor, K. Kwiatkowska, and A. Sobota, unpublished data.
|
ACKNOWLEDGMENTS |
|---|
KSR-GST probe; Drs. Konrad Sandhoff (University of Bonn) and Erich Gulbins for anti-ASMase antibody; Dr. J-L. Teillaud (Unite INSERM 255, Center de Recherches Biomedicals des Cordeliers, Paris) for anti-FcRII antibody; and Dr. Jürgen Frey (University of Bielefeld) for BHK-FcIIA transfectants. |
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
