Engagement of Phospholipid Scramblase 1 in Activated Cells

Phosphatidylserine (PS) in quiescent cells is predominantly confined to the inner leaflet of the plasma membrane. Externalization of PS is a marker of apoptosis, exocytosis, and some nonapoptotic activation events. It has been proposed that PS externalization is regulated by the activity of PLSCR1 (phospholipid scramblase 1), a Ca2+-dependent endofacial plasma membrane protein, which is tyrosine-phosphorylated in activated cells. It is, however, unclear how the phosphorylation of PLSCR1 is related to its membrane topography, PS externalization, and exocytosis. Using rat basophilic leukemia cells as a model, we show that nonapoptotic PS externalization induced through the high affinity IgE receptor (FcϵRI) or the glycosylphosphatidylinositol-anchored protein Thy-1 does not correlate with enhanced tyrosine phosphorylation of PLSCR1. In addition, PS externalization in FcϵRI- or Thy-1-activated cells is not associated with alterations of PLSCR1 fine topography as detected by electron microscopy on isolated plasma membrane sheets. In contrast, activation by calcium ionophore A23187 induces changes in the cellular distribution of PLSCR1. We also show for the first time that in pervanadate-activated cells, exocytosis occurs even in the absence of PS externalization. Finally, we document here that tyrosine-phosphorylated PLSCR1 is preferentially located in detergent-insoluble membranes, suggesting its involvement in the formation of membrane-bound signaling assemblies. The combined data indicate that changes in the topography of PLSCR1 and its tyrosine phosphorylation, PS externalization, and exocytosis are independent phenomena that could be distinguished by employing specific conditions of activation.

plasma membrane during cell activation and degranulation. It was, however, unclear whether tyrosine phosphorylation of PLSCR1 in mast cells is confined to Fc⑀RI triggering and in what way it is related to PS externalization and exocytosis observed in activated mast cells. To address these issues, we studied the interrelationships among tyrosine phosphorylation of PLSCR1 and its plasma membrane topography, PS externalization, and degranulation in RBL cells triggered by various stimuli known to have different effects on protein tyrosine phosphorylation, [Ca 2ϩ ] i , and exocytosis. Our data indicate that the above events are independent phenomena, which can occur individually, depending on the activation methods employed.
Production of New mAb against Plasma Membrane Microdomains and Identification of Their Target Antigens-RBL cells were lysed in ice-cold 1% Brij 96-containing lysis buffer (38). The lysate was subjected to fractionation on a sucrose density gradient, as described previously (38), and low density fraction, containing detergent-insoluble membranes (DIM), were pooled and used for immunization of BALB/c mice. Hybridoma cells were obtained after fusion of SPO2 mouse myeloma cells with spleen cells of immunized mice using standard procedures (39). The target antigens of the new mAb were immunoprecipitated from the lysate of resting RBL cells and separated by two-dimensional gel electrophoresis under reducing conditions. Alignment of immunoblotted and colloid silver-stained membranes allowed identification of silver-stained proteins corresponding to the target antigens. The corresponding in-gel silver-stained proteins were excised and destained as described previously (40). The proteins were digested by trypsin and analyzed by peptide mass mapping and/or peptide sequencing (41). Data base searches of the identified peptide fragments allowed determination of the target antigens.
DNA Constructs, Recombinant Proteins, and Immunoaffinity Purification of pAb-To prepare the N-terminal fragment of recombinant PLSCR1, full-length cDNA of rat PLSCR1 without initial ATG codon was amplified and cloned into XmaI and HindIII cloning sites of pQE30 expression vector (Qiagen, Hilden, Germany) using the forward (5Ј-AAACCCGGGGAG-AAGCACGGACCACCAGAA-3Ј) and reverse (5Ј-CCCAAG-CTTGCTACCATACTCCTGACCTTTG-3Ј) primers. A DNA fragment of the cloned cDNA was excised by SpHI and BamHI restriction enzymes and cloned into the corresponding sites in pQE70 expression vector (Qiagen). To prepare the glutathione S-transferase (GST)-PLSCR1 fusion protein, full-length cDNA of rat PLSCR1 without the initial ATG codon was cloned into EcoRI site of pGEX-3X expression vector (Amersham Biosciences) using the following primers: forward, 5Ј-AAGGAATT-CCGAGAAGCACGGACCACCAGAA-3Ј; reverse, 5Ј-AAAG-AATTCCTACCATACTCCTGACCTTTG-3Ј. Recombinant proteins were expressed in bacteria Escherichia coli strain JM109. The N-terminal fragment of PLSCR1 was isolated from inclusion bodies as previously reported (42) and used to immunize rabbits to generate R␣PLS. The recombinant GST-PLSCR1 was affinity-purified on glutathione-Sepharose beads (Amersham Biosciences) according to the manufacturer's protocol. The isolated GST-PLSCR1 or recombinant fragment of rat NTAL (37) was covalently bound to CNBr-activated beads (Sigma), which were then used for immunoaffinity purification of PLSCR1-or NTAL-specific pAb (37). In some experiments, isolated GST-PLSCR1 fusion protein was used at a concentration of 50 g/ml to confirm the specificity of anti-PLSCR1 antibodies.
Cells and Their Activation-The origin of RBL cells, clone 2H3, and their culture conditions have been described (43,44). Before activation, cells were harvested and washed with buffered saline solution (BSS; 20 mM HEPES, pH 7.4, 135 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 5.6 mM glucose), supplemented with 1.8 mM CaCl 2 and 0.1% bovine serum albumin (BSA). The cells were sensitized or not with TNP-specific IgE (IGEL b4 1 ascitic fluid diluted 1:1000 in BSS/BSA), ␣Fc⑀RI, or biotin-labeled ␣Thy-1 (both at 2 g/ml), washed, and resuspended in BSS/ BSA to 20 ϫ 10 6 /ml. The cells were activated at 37°C for the indicated time intervals by adding equal amounts of twice concentrated activators in BSS/BSA. Final concentrations of the activators were as follows: TNP-BSA (1 g/ml), ␣Fc⑀RI or ␣Thy-1 (2 g/ml), streptavidin (10 g/ml), calcium ionophore A23187 (3 M), and pervanadate (0.2 mM sodium orthovanadate, 1 mM hydrogen peroxide). To induce apoptosis by UV irradiation, cells grown in medium were exposed in open Petri dishes to UV-C (predominantly 254 nm) from a germicidal lamp (Philips TUV G30T8 30-W bulb) at 60 cm distance. After a 10-min exposure, fresh medium was added, and the cells were cultured for further 6 h before analysis. Apoptosis was confirmed by proteolytic cleavage of Lyn (45) and DNA laddering as previously described (46).
Determination of Cell Degranulation and PS Externalization-Degranulation of the cells and PS externalization were assessed by the amount of released ␤-glucuronidase and bound FITClabeled annexin V, respectively, as described (9).
Immunoprecipitation, Immunoblotting, and Silver Staining-Toward the end of the activation period, cells (10 7 ) were briefly centrifuged, and the pellet was lysed in ice-cold Triton X-100 lysis buffer (20 mM Tris-HCl, pH 8.0, 100 mM NaCl, 2 mM EDTA, 1% Triton X-100, 2 mM Na 3 VO 4 , 10 mM glycerol phosphate, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin). After 20 min, the lysate was centrifuged at 16,000 ϫ g for 5 min at 4°C, and the supernatant was further processed. To immunoprecipitate antigens recognized by mouse mAb, beads with covalently bound anti-mouse IgG (Sigma) were incubated with hybridoma supernatant for 2 h at 4°C. The beads were washed with lysis buffer and incubated with cell lysate for 2 h at 4°C. The immunoprecipitated antigen was eluted with glycine buffer (50 mM glycine-HCl, pH 2.5, 100 mM NaCl). To immunoprecipitate PLSCR1, a 1-ml aliquot of the cell lysate was incubated with 15 l of serum from PLSCR1-immunized rabbits for 2 h at 4°C, followed by incubation with UltraLink-immobilized protein A (Pierce) for 2 h at 4°C. Alternatively, 1 ml of the cell lysate was incubated with 10 g of immunoaffinity-purified R␣PLS or 2 g of affinity-purified M␣PLS*. Immunoprecipitated proteins were eluted by boiling in reducing Laemmli sample buffer for 10 min. The samples were analyzed by immunoblotting of SDS-PAGE-fractionated proteins or by two-dimensional electrophoresis, as described previously (47). Immunoblotting was performed with 2 g/ml purified antibodies, or 1:1000 diluted sera or ascitic fluids. Horseradish peroxidaseconjugated ␣PY, G␣M, or G␣R IgG were diluted 1:10,000. Horseradish peroxidase-generated signal was detected by enhanced chemiluminescence (Amersham Biosciences) on x-ray film and quantified by Luminescent Image Analyzer LAS-3000 (Fuji Photo Film Co., Tokyo). SDS-PAGE or two-dimensional fractionated proteins were visualized by in-gel silver staining using silver nitrate in combination with formaldehyde developer in alkaline carbonate buffer (48,49) or by membrane colloid silver staining (50).
Confocal Microscopy-Cells were spun down and resuspended in annexin V binding buffer (10 mM HEPES, pH 7.4, 140 mM NaCl, 2.5 mM CaCl 2 ) at a concentration of 5 ϫ 10 6 cells/ml. Forty l of the cell suspension was mixed with 3 l of annexin V-FITC and incubated at room temperature in the dark for 15 min. The cells were then transferred on poly-L-lysine-coated coverslips (12 mm in diameter) placed in wells of a 24-well plate. Five min later, the cells were fixed with 5% paraformaldehyde in annexin V binding buffer for 30 min, washed with phosphate-buffered saline solution, pH 7.2 (PBS), and nonspecific binding sites were blocked for 10 min with PBS supplemented with 5% BSA (PBS/BSA; also used for dilution of antibodies). After a 5-min permeabilization with 0.1% Triton X-100 in PBS and washing, the cells were incubated for 20 min with PBS/BSA and then labeled for 1 h with immunoaffinity-purified R␣PLS (5 g/ml) or rabbit anti-LAT serum (diluted 1:1000), followed by washing with PBS, blocking for 20 min with PBS/BSA, and incubation for 1 h with D␣R IgG-Cy3 conjugate (10 g/ml). After another washing step, the cells were fixed for 10 min with 5% paraformaldehyde in PBS, washed, and mounted in p-phenylenediamine (1 mg/ml, 50% glycerol in PBS). Images were acquired with a Leica TCS NT/SP confocal system in conjunction with Leica DMR microscope (Leica Microsystems GmbH, Wetzlar, Germany) equipped with oil objective ϫ100/1.4 numerical aperture.
Electron Microscopy-Plasma membrane sheets were prepared as described (51,52). Briefly, cells in complete culture medium were grown overnight on glass coverslips in the presence or absence of DNP-specific IgE (1 g/ml). Adherent cells were activated or not by DNP-BSA (1 g/ml) or other activators in BSS/BSA for the indicated time intervals at 37°C, and the glass coverslips with cells were washed in ice-cold PBS and then inverted and briefly pressed onto pioloform-covered and poly-L-lysine-coated electron microscopy grids kept on ice; careful separation of coverslips from the grids leaves the plasma membrane sheets attached to the grids. The membranes were fixed immediately with 2% paraformaldehyde and then exposed to immunoaffinity-purified R␣PLS (10 g/ml), rabbit anti-NTAL pAb (10 g/ml), or anti-Fc⑀RI-␤ mAb (3 g/ml). Washed membranes were exposed to gold-conjugated secondary antibodies diluted 1:20 from commercial stocks. After additional fixation with 2% glutaraldehyde and staining with osmium tetroxide, tannic acid, and uranyl acetate, the samples were dried and examined under a JEOL JEM 1200EX electron microscope operating at 60 kV. To obtain membrane sheets from cells with aggregated Thy-1, the cells were incubated for 15 min at room temperature with ␣Thy-1 (1 g/ml), washed, and then incubated at 37°C for 10 min with G␣M IgG-10-nm gold conjugate. In the case of Thy-1 labeling without aggregation, the cells were first fixed with paraformaldehyde and labeled for 15 min with ␣Thy-1, followed by 10-min exposure to G␣M IgG-10-nm gold conjugate. After the extracellular labeling, plasma membrane sheets were isolated as described above. Plasma membrane sheets from nonadherent apoptotic cells were prepared as described (53).
Evaluation of the Results-Means Ϯ S.D. were computed from at least three independent experiments. Statistical significance of intergroup differences was calculated using an unpaired Student's t test; each group consisted of four independent experiments. Evaluation of gold particle co-localization was performed by bivariate Ripley's analysis (BRA) using the L(r)-r function as described (52). Pictures of plasma membrane sheets from two independent experiments were taken to cover ϳ50 m 2 of plasma membrane surface and analyzed using Matlab software algorithms (52). Significant co-localization (p Ͻ 0.01) of the proteins occurred when the position of the L value curve (solid lines) surpassed the boundaries (dashed lines) predicted for random distribution of gold particles at a corresponding distance (Fig. 3, E and F and Fig. 5, D, and E). Cluster size was determined by the program GOLD (54).

Fc⑀RI-induced Degranulation and PS Externalization Do Not
Correlate with the Extent of PLSCR1 Tyrosine Phosphorylation-Antigen-mediated aggregation of the Fc⑀RI triggers signaling pathways leading to increased tyrosine phosphorylation of numerous proteins, including PLSCR1, enhanced [Ca 2ϩ ] i , degranulation, and transient nonapoptotic PS externalization (6,7,9,22,55). To elucidate the role of PLSCR1 tyrosine phosphorylation in these processes, we first prepared and character-ized new PLSCR1-specific mAb and pAb. A panel of mAb was prepared after immunizing BALB/c mice with DIM isolated from RBL cells. One of the mAbs, clone TEC-23, reacted with two proteins of ϳ36 and 38 kDa under reducing conditions (Fig. S1A). These two bands were observed even after direct solubilization of the cells in reducing Laemmli sample buffer and immediate boiling; this makes it unlikely that the lower band is produced by proteolysis during cell solubilization. Under nonreducing conditions, only one band of 38 kDa was observed (not shown). Interestingly, mass spectroscopic analysis identified 10 peptides, with amino acid sequences identical in both proteins and corresponding to rat PLSCR1 (Fig. S1B). Furthermore, immunoblotting experiments showed that the binding of the PLSCR1-specific mAb, clone TEC-23 (M␣PLS), to its target was completely inhibited by recombinant GST-PLSCR1 (Fig. S1C); the same recombinant protein also inhibited the binding of immunoaffinity-purified R␣PLS and M␣PLS* but had no effect on binding of anti-Lyn mAb to its target (Fig. S1C). Furthermore, material immunoprecipitated with R␣PLS or M␣PLS* reacted with M␣PLS and M␣PLS* or R␣PLS, respectively (Fig. S1D). All of these data indicated that the newly prepared mAb M␣PLS and rabbit pAb R␣PLS could specifically detect PLSCR1.
Using the newly prepared antibodies, we confirmed previous data indicating that antigen-mediated Fc⑀RI triggering leads to enhanced tyrosine phosphorylation of PLSCR1 (22) and extended them by showing that this phosphorylation is transient, with a peak at about 10 min after triggering (Fig. 1A, left). To decide whether the observed phosphorylation is related to antigen-induced Fc⑀RI aggregation (56) and/or internalization of Fc⑀RI (53,57), the cells were activated by ␣Fc⑀RI, which triggers signaling events comparable with those induced by extensive aggregation of Fc⑀RI-IgE complexes by multivalent antigen but lacking Fc⑀RI clustering or rapid clearance of Fc⑀RI from the plasma membrane (9,58). As shown in Fig. 1A (right), cells activated by ␣Fc⑀RI also exhibited transient PLSCR1 tyrosine phosphorylation, indicating that extensive FIGURE 1. Tyrosine phosphorylation of PLSCR1 in Fc⑀RI-activated cells and its role in cell degranulation and PS externalization. A, cells were sensitized or not (C-) on ice with TNP-specific IgE or exposed to ␣Fc⑀RI mAb. After 30 min, the cells were washed and then activated at 37°C for the time intervals indicated by extensive Fc⑀RI aggregation (IgE/TNP-BSA) or Fc⑀RI dimerization (␣Fc⑀RI). Whole cell lysates were prepared and analyzed by immunoblotting (IB) with ␣PY-horseradish peroxidase conjugate (␣PY; upper part). PLSCR1 was immunoprecipitated (IP) with R␣PLS from cell lysates and analyzed by immunoblotting with ␣PY. The amount of immunoprecipitated PLSCR1 was determined by immunoblotting with M␣PLS (lower part). B-E, Fc⑀RI was dimerized for 10 min with ␣Fc⑀RI in the absence (Latr-) or presence (Latrϩ) of latrunculin B (0.5 M). B, PS externalization was determined by flow cytometry after staining with annexin V-FITC. Thin and thick lines indicate annexin V-FITC binding to nonactivated and ␣Fc⑀RI-activated cells, respectively. C, data obtained as described in B were normalized to maximal response in each experiment. The effect of latrunculin on ␣Fc⑀RI-induced degranulation was determined by quantification of ␤-glucuronidase released into supernatant and normalized to maximal response in each experiment. D, PLSCR1 was immunoprecipitated from cell lysates of control (LatrϪ) and latrunculin-treated (Latrϩ) nonactivated (Ϫ) or ␣Fc⑀RI-activated (ϩ) cells and analyzed by immunoblotting for PLSCR1 tyrosine phosphorylation (␣PY) and amount of PLSCR1 immunoprecipitated (M␣PLS). E, gels obtained as described in D were quantified by densitometry; tyrosine phosphorylation of PLSCR1 was normalized to the amount of PLSCR1 and maximal signal obtained in latrunculin B-treated cells. In A and D, positions of PLSCR1 and tyrosine-phosphorylated PLSCR1 (P-PLSCR1) are shown on the right. In A, molecular weight standards in kDa are also presented. In C, statistical differences between LatrϪ and Latrϩ cells are indicated (**, p Ͻ 0.01). In A, B, and D, representative experiments from at least three performed are shown.
clustering and/or internalization of Fc⑀RI is not necessary for PLSCR1 tyrosine phosphorylation.
Secretory response and internalization of Fc⑀RI are negatively regulated by filamentous actin (59 -61). To determine whether Fc⑀RI-induced PLSCR1 tyrosine phosphorylation and PS externalization are dependent on intact actin filaments, we analyzed the properties of RBL cells stimulated with ␣Fc⑀RI in the presence or absence of 0.5 M latrunculin B, an inhibitor of actin polymerization (59,62). In accordance with our previous data (9), stimulation of RBL cells with ␣Fc⑀RI in the presence of latrunculin B significantly enhanced PS externalization, determined by annexin V-FITC binding (Fig. 1, B and C), and degranulation, measured by the release of ␤-glucuronidase into supernatant (Fig. 1C). However, tyrosine phosphorylation of PLSCR1 did not significantly increase in latrunculin-treated cells (Fig. 1, D and E). Treatment of the cells with 0.5 M latrunculin B in the absence of Fc⑀RI triggering had no effect on PLSCR1 tyrosine phosphorylation (Fig. 1D), PS externalization, and ␤-glucuronidase release (data not shown). The combined data suggest that enhanced PS externalization and degranulation are independent of enhanced tyrosine phosphorylation of PLSCR1.

Externalized PS Does Not Always Co-localize with PLSCR1 in Fc⑀RI-stimulated Cells-Fc⑀RI-induced activation of RBL cells leads to rapid and transient increase in [Ca 2ϩ ]
i . It has been suggested that this results in enhanced binding of calcium to PLSCR1 and its conformational changes, leading to PLSCR1 clustering on the plasma membrane and activation of phospholipid scrambling (20). To test this hypothesis, we analyzed the distribution of externalized PS and PLSCR1 in Fc⑀RI-activated cells by confocal microscopy. In order to induce a slower but more sustained increase in [Ca 2ϩ ] i (58), we activated RBL cells by ␣Fc⑀RI and stained them with annexin V-FITC, followed by fixation, permeabilization, and staining with PLSCR1-specific pAb. Data presented in Fig. S2A show that recombinant GST-PLSCR1 inhibited the binding of R␣PLS but not anti-LAT pAb, proving that the antibody is specific for PLSCR1 in permeabilized cells. Furthermore, binding of ␣Fc⑀RI alone did not give any signal with the secondary D␣R IgG-Cy3 conjugate used for detection of R␣PLS binding (Fig. S2B). In nonactivated cells, no externalization of PS was detected by annexin V-FITC conjugate (Fig. 2, C-/Annexin V). In the same cells, PLSCR1 was dispersed over the whole plasma membrane, as detected by R␣PLS (Fig. 2, C-/PLSCR1). When the cells were activated by ␣Fc⑀RI, externalized PS was detected by annexin V-FITC in discernible clusters, whereas PLSCR1 remained mostly randomly distributed on the plasma membrane (Fig. 2, ␣Fc⑀RI). At higher magnification, PLSCR1 was distributed in numerous patches over the whole plasma membrane in both Fc⑀RI-activated cells (Fig.  2, ␣Fc⑀RI/Zoom/PLSCR1) and nonactivated cells (not shown). Externalized PS was occasionally associated with these patches but was also found in regions with no preferential localization of PLSCR1 (green spots in Fig. 2, ␣Fc⑀RI/Zoom/Overlay). These data indicate that Fc⑀RI triggering does not induce aggregation of PLSCR1 and that patches of externalized PS do not always co-localize with patches of PLSCR1.

Aggregation of Fc⑀RI Does Not Lead to Changes in the Fine
Topography of PLSCR1-To attain higher resolution in determination of PLSCR1 topography in the plasma membrane, we used electron microscopy combined with immunogold labeling on isolated plasma membrane sheets (51,63). In pilot experiments, specificity of PLSCR1 visualization was examined. Isolated plasma membrane sheets were labeled with R␣PLS or anti-NTAL pAb in the absence or presence of recombinant GST-PLSCR1, followed by application of G␣R IgG-5-nm gold conjugate. As shown in Fig. 3A, the recombinant protein completely blocked the binding of R␣PLS, whereas it had insignificant effect on anti-NTAL pAb binding, confirming thus the suitability of the reagent for electron microscopy studies. Next we evaluated the topography of PLSCR1 in cells activated by Fc⑀RI dimerization and found no significant difference in PLSCR1 cluster size (Fig. 3B, left) and particle density (Fig. 3B, right) between controls and ␣Fc⑀RI-activated cells. These findings supported previous results obtained with confocal microscopy showing that cell activation through Fc⑀RI dimerization does not affect the distribution of PLSCR1 in the plasma membrane.
We also analyzed the topography of PLSCR1 in cells with extensively aggregated Fc⑀RI by multimeric antigen-IgE complexes. In accordance with previous data (58,63), the Fc⑀RI ␤ chain was found dispersed individually or in small clusters in nonactivated cells (Fig. 3C), whereas in multimeric antigen (DNP-BSA)-activated cells, it was found in large clusters associated with osmiophilic regions (Fig. 3D). Importantly, extensive Fc⑀RI aggregation induced no obvious changes in the topography of PLSCR1. Furthermore, statistical evaluation showed no increased co-localization of PLSCR1 and Fc⑀RI in activated cells (Fig. 3, E and F). To determine whether internalization of aggregated Fc⑀RI could trigger changes in distribution of PLSCR1, we visualized PLSCR1 and Fc⑀RI on isolated membrane sheets simultaneously in the course of a 40-min activation. Data presented in Fig. 3G indicate that aggregation of Fc⑀RI enhanced the Fc⑀RI cluster size with a peak at 2 min, followed by its return to original size 40 min after triggering. Formation of Fc⑀RI clusters was accompanied by a significant decrease in number of Fc⑀RI detected (Fig. 3H). In contrast, when PLSCR1 was evaluated, no significant changes either in its clustering ( Fig. 3G) or density (Fig. 3H) were observed. Thus, neither dimerization nor extensive aggregation of Fc⑀RI has any effect on topography of PLSCR1 in the plasma membrane.
Thy-1 Dimerization Induces PS Externalization in the Absence of Tyrosine Phosphorylation or Self-aggregation of PLSCR1-Thy-1 is a glycosylphosphatidylinositol (GPI)anchored protein expressed in a large amount on the surface of RBL and rat peritoneal mast cells (44,64). Recently, we have found that dimerization of Thy-1 by Thy-1specific mAb MRCOX7 induces nonapoptotic PS externalization in the absence of an increase in [Ca 2ϩ ] i and degranulation (9). To clarify whether or not Thy-1-triggered PS externalization is accompanied by changes in tyrosine phosphorylation and/or topography of PLSCR1, we examined RBL cells activated by ␣Thy-1. In accordance with a previous report (9), dimerization of Thy-1 induced PS externalization (Fig. 4A) without any sign of degranulation (Fig. 4B). As shown in Fig.  4C, dimerization of Thy-1 did not induce PLSCR1 tyrosine phosphorylation. As expected, under the same conditions, positive control cells exhibited strong secretory response and PLSCR1 tyrosine phosphorylation when activated through dimerized Fc⑀RI (Fig. 4, B  and C). Confocal microscopy showed that dimerization of Thy-1 FIGURE 3. Topography of PLSCR1 in plasma membrane of Fc⑀RI-activated cells as detected by electron microscopy. A, specificity of the reagents used. Fixed plasma membrane sheets were labeled on the cytoplasmic side with pAb specific for NTAL (black bars) or PLSCR1 (white bars) in the absence (rec-PLSCR1Ϫ) or presence (recPLSCR1ϩ) of recombinant GST-PLSCR1. Bound antibodies were visualized with G␣R IgG-5-nm gold particle conjugate, and particle density was determined and normalized to the rec-PLSCR1Ϫ samples. B, topography of PLSCR1 in cells activated by Fc⑀RI-dimerizing ␣Fc⑀RI. Cells were stimulated for 20 min with vehicle alone (C-) or ␣Fc⑀RI. PLSCR1 was visualized as described in A. Cluster size and particle density were determined. C-H, topography of PLSCR1 and Fc⑀RI in cells activated by extensive aggregation of Fc⑀RI. Cells were sensitized with DNP-specific IgE and then stimulated with vehicle alone (control) or with DNP-BSA. Plasma membrane sheets from control (C) or DNP-BSA (2 min)stimulated cells (D) were isolated, fixed, and simultaneously labeled on the cytoplasmic face with R␣PLS (PLSCR1) and Fc⑀RI-␤ subunit-specific mAb (Fc⑀RI-␤). Bound R␣PLS and Fc⑀RI-␤ were visualized with G␣R IgG-5-nm gold particle conjugate (arrowheads) and G␣M IgG-10-nm gold particle conjugate (arrows), respectively. BRA of co-localization of PLSCR1 and Fc⑀RI-␤ subunit in control (E) and activated (F) cells was determined. G and H, the cells were sensitized with DNP-specific IgE and then stimulated with DNP-BSA for the time intervals indicated. Visualization of PLSCR1 and Fc⑀RI-␤ subunit on isolated plasma membrane sheets was carried out as described above. Cluster size (G) and particle density (H) of PLSCR1 and Fc⑀RI-␤ subunit was evaluated for each time interval after triggering (0 -40 min). Statistical differences, calculated between recPLSCR1Ϫ and recPLSCR1ϩ (A) or nonactivated and activated cells (G and H), are indicated (*, p Ͻ 0.05; **, p Ͻ 0.01). In C and D, representative experiments from at least three performed are shown. APRIL 18, 2008 • VOLUME 283 • NUMBER 16

JOURNAL OF BIOLOGICAL CHEMISTRY 10909
induced PS externalization detected by annexin V-FITC but no changes in PLSCR1 distribution (Fig. 4D, ␣Thy-1; compare with Fig. 2, C-/PLSCR1). At higher magnification, there was no preferential localization of PLSCR1 at regions of patches of externalized PS (Fig. 4D, ␣Thy-1/Zoom). Finally, electron microscopy studies on isolated membrane sheets indicated that cluster size (Fig. 4E, left) and density of PLSCR1 (Fig. 4E, right) were not significantly different between control and ␣Thy-1treated cells. Collectively, these data indicate that PS external-ization induced by dimerized Thy-1 does not require enhanced tyrosine phosphorylation of PLSCR1 and/or its topographical changes.
Thy-1 Multimerization Induces Enhanced Tyrosine Phosphorylation of PLSCR1-In contrast to Thy-1 dimerization, extensive aggregation of Thy-1 leads to enhanced [Ca 2ϩ ] i and degranulation (64,65). To find out whether this also induces changes in PLSCR1 tyrosine phosphorylation, we aggregated Thy-1 with biotin-labeled ␣Thy-1-streptavidin complexes. Such treatment resulted in enhanced tyrosine phosphorylation of several cellular proteins peaking at 1 min after triggering for most of the proteins (Fig. 5A,  top). Immunoprecipitation analyses showed that PLSCR1 was among the tyrosine-phosphorylated proteins (Fig. 5A, bottom) with a phosphorylation peak observed at 10 min after triggering. Although aggregated Thy-1 was easily detectable on isolated membrane sheets, its enhanced co-localization with PLSCR1 was detectable neither in resting nor activated cells (Fig. 5, B-E). These data indicated that increased tyrosine phosphorylation of PLSCR1 is not accompanied by enhanced co-localization with aggregated Thy-1.
Calcium Ionophore A23187 Induces Changes in the Topography of PLSCR1 and Apoptotic-like PS Externalization-Stimulation of the cells with calcium ionophores leads to a rapid elevation of [Ca 2ϩ ] i , followed by a sequence of Ca 2ϩ -dependent signaling events, including externalization of PS, degranulation, and apoptosis (6,66,67). Furthermore, exposure of IgEsensitized RBL cells for 30 min with calcium ionophore ionomycin has been reported to enhance tyrosine phosphorylation of PLSCR1 (31). To determine kinetics of calcium ionophore-induced phosphorylation, we studied tyrosine-phosphorylated proteins at various time intervals after A23187 triggering. Data in Fig. 6A show that the rate of phosphorylation of cellular proteins, including PLSCR1, is slower in A23187-activated cells than in Fc⑀RI-activated cells (compare with Fig. 1A). As to the extent of PS externalization, it was substantially higher in cells stimulated by A23187 and labeled in its presence (Ionoϩ/ϩ) than by Fc⑀RI dimerization (Fig. 6B, FIGURE 4. Properties of PLSCR1 in cells activated through dimerized Thy-1. A, cells were exposed for 20 min to ␣Thy-1 (thick line) or ␣Fc⑀RI (dashed and dotted line) or vehicle alone (C-; thin line), and PS externalization was determined by flow cytometry after staining with annexin V-FITC. B, the cells were activated as in A, and the extent of degranulation was determined by measurement of ␤-glucuronidase released into supernatant; data obtained were normalized to maximal response in each experiment. C, the cells were activated as in A and solubilized; PLSCR1 was immunoprecipitated (IP) with R␣PLS and analyzed by immunoblotting (IB) with ␣PY. The amount of immunoprecipitated PLSCR1 was determined by immunoblotting with M␣PLS. Positions of PLSCR1 and tyrosine-phosphorylated PLSCR1 are shown on the right. D, topography of PLSCR1 and externalized PS as determined by confocal microscopy in cells activated through Thy-1 dimerization by ␣Thy-1. Externalized PS and PLSCR1 were detected with annexin V-FITC (Annexin V) and R␣PLS as described in Fig. 2. The merge of both labels is shown on the right (Overlay). Bar in ␣Thy-1/Overlay, 10 m. Zoom images of the indicated regions of ␣Thy-1-activated cells are shown at the bottom. E, electron microscopic analysis of PLSCR1 in plasma membrane sheets isolated from cells activated for 20 min with dimerized Thy-1 (␣Thy-1) or vehicle alone (C-). PLSCR1 was visualized as described in the legend to Fig. 3A, and cluster size and particle density were evaluated. In B, statistical differences between ␣Fc⑀RI-activated cells and other groups are indicated (**, p Ͻ 0.01). In A, C, and D, representative experiments from at least three performed are shown.
left and middle). However, the extent of PS externalization did not correlate with degranulation, which was substantially lower in A23187-triggered cells (Fig. 6B, right). The stability of A23187-induced PS externalization was tested by removing the drug from the cells during staining for PS (Ionoϩ/Ϫ); the amount of bound annexin V-FITC decreased to levels comparable with those observed in Fc⑀RI-activated cells (Fig. 6C, left  and middle). However, secretory response remained the same (Fig. 6C, right). Thus, extensive PS externalization in A23187-stimulated cells depends on the continuous presence of A23187 and is reversible after its removal.
Next, we studied the topography of externalized PS and PLSCR1 in cells activated by A23187 under different conditions. For comparison, we also analyzed the distribution of those molecules in apoptotic cells. Confocal microscopy showed that cells activated by A23187 and labeled in its presence displayed PS almost homogeneously distributed on the cell surface (Fig. 6D, Ionoϩ/ ϩ). Interestingly, such treatment also led to a decrease in plasma membrane staining of PLSCR1 and its appearance in cytoplasm (compare with Fig. 2, C-/PLSCR1). Staining of A23187-activated cells in the absence of A23187 resulted in decreased binding of annexin V-FITC, which was found only in spots and patches (Fig. 6D, Ionoϩ/ Ϫ). This pattern was associated with enhanced localization of PLSCR1 in the plasma membrane. Distribution of PS and PLSCR1 in apoptotic cells resembled that seen in cells activated with A23187 and stained in its presence (compare Fig. 6D, Apo and Ionoϩ/ϩ). At higher magnification, the observed differences in topography of PS and PLSCR1 in cells activated and stained under various conditions were even more evident (Fig. 6D, Overlay/Zoom). Electron microscopy analysis of plasma membrane sheets from A23187stimulated cells or apoptotic cells showed no significant difference in size of PLSCR1 clusters (Fig. 6E). However, the amount of detectable PLSCR1 on plasma membrane sheets isolated from both A23187treated or apoptotic cells was significantly lower (Fig. 6F). The observed decrease was specific for PLSCR1, since NTAL, another protein with transmembrane domain (68), showed no significant decrease in immunolabeling on the same plasma membrane sheets (Fig. 6F). These findings confirm the confocal microscopy data and suggest that PLSCR1 is specifically released from plasma membrane in both A23187activated and apoptotic cells.
Stimulation with Pervanadate Leads to Extracellular Calcium-independent Tyrosine Phosphorylation of PLSCR1 and  C and E). B, cells were fixed, and Thy-1 was labeled with ␣Thy-1 and visualized with G␣M IgG-10-nm gold particle conjugate (arrow). After this step, membrane sheets were isolated, and PLSCR1 (arrowhead) was labeled on the cytoplasmic face as described in the legend to Fig. 3A. C, nonfixed cells were exposed to ␣Thy-1, and Thy-1-␣Thy-1 immunocomplexes were aggregated by G␣M IgG-10-nm gold particle conjugate (arrow). The membrane sheets were then isolated and fixed, and PLSCR1 (arrowhead) was determined as described in B. D-E, BRA of co-localization of PLSCR1 and Thy-1 in nonactivated cell (D) or cells with extensively aggregated Thy-1 (E) were evaluated. In A-C, representative experiments from at least three performed are shown.

FIGURE 6. Changes in properties of PLSCR1 and PS externalization in calcium ionophore A23187-stimulated cells.
A, tyrosine phosphorylation of PLSCR1. Cells were activated by A23187 at 37°C for the time intervals indicated. Cell lysates were prepared and analyzed by immunoblotting (IB) for tyrosine-phosphorylated proteins with ␣PY. Tyrosine phosphorylation of PLSCR1 was analyzed in R␣PLS immunoprecipitates (IP) from the same cells and analyzed by immunoblotting with ␣PY. The amount of immunoprecipitated PLSCR1 was determined by immunoblotting with M␣PLS. Positions of PLSCR1, tyrosine-phosphorylated PLSCR1, and molecular mass standards in kDa are shown on the right. B, PS externalization and degranulation. Cells were exposed for 15 min to A23187 (Ionoϩ/ϩ, thick line), ␣Fc⑀RI (dashed and dotted line), or vehicle alone (C-, thin line) and then labeled with annexin V-FITC for 15 min in the absence (␣Fc⑀RI or C-) or presence (Ionoϩ/ϩ) of A23187. The cells were immediately analyzed by flow cytometry (left), and annexin V binding was normalized to maximal response obtained in each experiment (middle). In separate experiments, cells were treated for 30 min with A23187 or ␣Fc⑀RI, and degranulation (␤-gluc. release) was determined as described in the legend to Fig. 1C. C, cells were treated and analyzed as in B except that A23187 was absent during labeling with annexin V-FITC (Ionoϩ/Ϫ, thick line), and degranulation was determined 15 min after triggering. D, topography of externalized PS and PLSCR1 in plasma membrane of A23187-stimulated or apoptotic cells as determined by confocal microscopy. A23187-stimulated and annexin V-FITC-labeled cells, as in B (Ionoϩ/ϩ) or C (Ionoϩ/Ϫ), or annexin V-FITC-labeled apoptotic cells (Apo) were fixed in the presence (Ionoϩ/ϩ) or absence (Ionoϩ/Ϫ, Apo) of A23187 and permeabilized, and PLSCR1 was visualized as described in the legend to Fig. 2. Distribution of bound annexin V (Annexin V), PLSCR1, and their corresponding merge (Overlay) is shown. Bar in Apo/Overlay, 10 m. Zoom images of the indicated regions in Overlays are shown on the right. E and F, topography of PLSCR1 on plasma membrane as determined by electron microscopy. Plasma membrane sheets from nonactivated (C-), A23187-stimulated (Ionoϩ/ϩ, 15 min), or apoptotic (Apo) cells were prepared and labeled as described in the legend to Fig. 3A with the exception that fixation of plasma membrane sheets was performed in the presence (Ionoϩ/ϩ) or absence (C-, Apo) of A23187. Cluster size of PLSCR1 (E) and particle density of PLSCR1 and control transmembrane adaptor protein NTAL (F) were normalized to nonactivated cells. Statistical differences between ␣Fc⑀RI and Ionoϩ/ϩ (B), ␣Fc⑀RI and Ionoϩ/Ϫ (C), and controls (C-) and experimental groups (F) are shown (*, p Ͻ 0.05; **, p Ͻ 0.01). In A and D, representative experiments from at least three performed are shown.

Degranulation without PS Externalization-Previous results
suggested that tyrosine phosphorylation of PLSCR1 in calcium ionophore-activated cells is regulated downstream of the increase of [Ca 2ϩ ] i (31). More information on the role of calcium in tyrosine phosphorylation of PLSCR1 was obtained in RBL cells stimulated by pervanadate in the absence or presence of extracellular Ca 2ϩ . Pervanadate is a compound triggering tyrosine phosphorylation of cellular proteins through inhibition of protein-tyrosine phosphatases (69). Depending on the concentration and time of exposure, it induces either cell activation or apoptosis (70,71). For our purposes, we used pervanadate at a concentration that had previously been found to induce degranulation comparable with that induced by Fc⑀RI triggering (72). Using total cell lysates, we found rapid tyrosine phosphorylation of many cellular proteins in pervanadatestimulated cells in the presence or absence of extracellular calcium (Fig. 7A, top). Immunoblotting on parallel membranes with M␣PLS showed enhanced molecular weight of a fraction of PLSCR1 in pervanadate-activated cell (Fig. 7A, bottom), apparently reflecting its enhanced tyrosine phosphorylation. Immunoprecipitated PLSCR1 also showed strong tyrosine phosphorylation in the presence or absence of extracellular calcium (Fig. 7B). These data indicate that under these conditions, tyrosine phosphorylation of PLSCR1 is not dependent on extracellular calcium.
Next we examined whether pervanadate-induced degranulation was accompanied by PS externalization as was observed in Fc⑀RI-activated cells. To that aim, RBL cells were exposed to pervanadate or ␣Fc⑀RI (positive control) and concomitantly tested for PS externalization (Fig. 7C, left and middle) and degranulation (Fig. 7C, right). Surprisingly, exposure of the cells to pervanadate resulted in no significant PS externalization, although degranulation occurred as expected. There was no increase in the number of dead cells, determined by propidium iodide staining (not shown), arguing against the possibility that ␤-glucuronidase was released from cells killed by pervanadate. This finding demonstrates that, under certain conditions, externalization of PS cannot be considered as a reliable marker of cell degranulation, despite the fact that there is a correlation between PS externalization and degranulation (6 -8, 73).
We also examined the topography of PLSCR1 in pervanadate-stimulated cells. Our observations from confocal microscopy (data not shown) indicated that pervanadate induced no redistribution of PLSCR1. Electron microscopy analyses of PLSCR1 topography in plasma membrane sheets isolated from controls or pervanadate-activated cells showed no significant changes in cluster size (Fig. 7D, left) or particle density (Fig. 7D, right). These data suggest that even hyperphosphorylation of PLSCR1 need not lead to its topographic changes. and degranulation (␤-gluc. release) was determined as described in Fig. 1, B and C, and normalized to maximal response attained in each experiment. D, cells were stimulated for 15 min with pervanadate or vehicle alone, and topography of PLSCR1 on plasma membrane sheets was determined by electron microscopy as described in the legend to Fig. 3B. PLSCR1 cluster size and particle density were normalized to nonstimulated cells. In C, statistical differences between maximal and other values among ␣Fc⑀RIand pervanadate-activated cells are indicated (**, p Ͻ 0.01). In A-C, representative experiments from at least three performed are shown.
Tyrosine-phosphorylated PLSCR1 Is Localized Mostly in DIM-PLSCR1, like other palmitoylated transmembrane proteins, has been reported to be enriched in low density membranes (23,74). However, it remained questionable whether tyrosine-phosphorylated PLSCR1 in RBL cells is also associated with DIM and what changes in this association take place in the course of Fc⑀RI activation. When the cells were solubilized in a lysis buffer with 0.5% Brij 58 and then the whole cell lysates fractionated by sucrose density gradient, a small amount (ϳ10%) of PLSCR1 in DIM (fractions 1-3) was detectable by immunoblotting (Fig. 8A, top and bottom, left). This was much less than the Lyn kinase (ϳ80%), which is considered to be a typical DIM-associated protein (38). Furthermore, when the cells were solubilized in lysis buffer with 1% Triton X-100, no PLSCR1 was recovered in DIM, whereas most of the Lyn still remained in DIM (not shown). In cells activated by IgE⅐TNP-BSA complexes, there was an ϳ60% decrease in the amount of PLSCR1 in DIM but only ϳ20% decrease in Lyn (Fig. 8A, top  and bottom, right). Although only a small fraction of PLSCR1 was associated with DIM (see above), most of the tyrosinephosphorylated PLSCR1 was located in DIM when PLSCR1 was immunoprecipitated from pooled low or high density fractions and analyzed by immunoblotting, and this amount further increased after Fc⑀RI triggering (Fig. 8B). The increase in phosphorylated PLSCR1 in high density fractions was even more striking. Activation of the cells with pervanadate resulted in an enhanced amount of phosphorylated PLSCR1 in both low and high density fractions (Fig. 8C). These findings support the concept that tyrosine phosphorylation of PLSCR1 is dependent on equilibrium between kinases and phosphatases.

DISCUSSION
We have previously shown that dimerization of Thy-1 or some other GPI-anchored proteins induces nonapoptotic PS externalization in several cell types, including mast cells (9). It was an unexpected finding, because GPI-anchored proteins have no transmembrane and cytoplasmic domains, and therefore their effect must be mediated through their binding to other proteins possessing such domains or through poorly characterized interactions within lipid rafts where GPI-anchored and other signaling molecules are presumably localized (75,76). Furthermore, when the cells were activated through dimerized GPI-anchored proteins, it was possible to get PS externalization even in the absence of degranulation (9). These data suggested that rational tests on cells activated under various conditions and by different activation pathways may possibly provide new information on the relationships between, on the one hand, PS externalization, enhanced [Ca 2ϩ ] i , and exocytic responses and, on the other, tyrosine phosphorylation and topography of PLSCR1, an enzyme that is phosphorylated in activated cells and thought to take part in phospholipid scrambling. For our experiments, we selected RBL cells, which are  (1)(2)(3) or high density (6 -9) fractions, and the extent of PLSCR1 tyrosine phosphorylation and the total amount of PLSCR1 were analyzed by immunoblotting with ␣PY and M␣PLS, respectively. Tyrosine phosphorylation of PLSCR1 was normalized to amount of the protein and maximal response obtained in each experiment in low or high density fractions. C, the cells were activated by vehicle alone (C-) or pervanadate (Perv) at 37°C for 15 min, lysed, and fractionated by sucrose density gradient ultracentrifugation. PLSCR1 was immunoprecipitated from pooled low density or high density fractions as in B and analyzed by immunoblotting with ␣PY or M␣PLS. Tyrosine phosphorylation of PLSCR1 was normalized to the amount of this protein and maximal response attained in each experiment in vehicle alone or pervanadate. Positions of PLSCR1, tyrosine-phosphorylated PLSCR1, and Lyn are shown on the right. Statistical differences are indicated (*, p Ͻ 0.05; **, p Ͻ 0.01).
widely used as a model object for studies on surface receptormediated exocytosis and nonapoptotic PS externalization. These cells can be activated by various means, including dimerization of Fc⑀RI or Thy-1 by corresponding mAb, extensive aggregation of Fc⑀RI through IgE-multimeric antigen complexes, or aggregation of Thy-1 by biotinylated anti-Thy-1 mAb and streptavidin, calcium ionophore A23187, or phosphatases inhibitor, pervanadate. Data presented in this study and summarized in Table 1 show that many of the signaling events can be separately approached by employing various activation methods.
First, dimerization of Thy-1 could induce PS externalization in the absence of PLSCR1 tyrosine phosphorylation. Interestingly, the extent of PS externalization in cells activated in this way was comparable with that induced by extensively aggregated Fc⑀RI or Thy-1. This fact, together with our previous finding that the proliferation rate of cells cultured in the presence of anti-Thy-1 mAb is not impaired (9), indicate that tyrosine phosphorylation of PLSCR1 is not essential for nonapoptotic PS externalization. The conclusion is further corroborated by the observation that latrunculin B, an inhibitor of actin polymerization, known to enhance secretory responses (see also Refs. 59 and 61), also increased PS externalization without any significant effect on tyrosine phosphorylation of PLSCR1. Thus, the inhibitory effect of actin on both PS externalization and secretory response is not mediated through inhibition of PLSCR1 tyrosine phosphorylation.
Second, the above conclusion is consistent with our finding that activation by calcium ionophore A23187 induced delayed tyrosine phosphorylation of PLSCR1 but strong PS externalization. At the same time, degranulation in A23187-triggered cells was lower than in Fc⑀RI-activated cells; it strengthens the conclusion that PS externalization and secretory responses are independent phenomena.
Third, when activated with pervanadate, the cells exhibited clear secretory response and strong PLSCR1 tyrosine phosphorylation but no PS externalization. These observations provide the first evidence that degranulation can occur even in the absence of PS externalization and support the concept that phosphorylation of PLSCR1, PS externalization, and degranulation are disparate events. Comparable tyrosine phosphorylation of PLSCR1 was observed in cells activated by pervanadate in Ca 2ϩ -supplemented medium or Ca 2ϩ -free medium supplemented with EGTA. This was an unexpected finding, because a previous study showed that Fc⑀RI-mediated phosphorylation of PLSCR1 required Ca 2ϩ ; in its absence, phosphorylation was markedly reduced (31). This difference could be related to the existence of preassembled complexes of PLSCR1, kinases, and phosphatases (see below) as well as to the fact that Ca 2ϩ is required for uncoupling the phosphatases from the complexes under conditions of physiological triggering, such as through Fc⑀RI.
Fourth, it had been proposed that enhanced binding of metal ions to PLSCR1 caused its conformational changes, followed by PLSCR1 clustering and activation of phospholipid scrambling (20). It was therefore of interest to find out whether cell activation and PS externalization are indeed accompanied by changes in topography of PLSCR1. Confocal microscopy did not reveal any enhanced movement of PLSCR1 into regions of externalized PS. Although clusters of externalized PS and PLSCR1 were occasionally observed, such clustering was not found in other cases, suggesting that either clustering events are highly dynamic or clustering of PLSCR1 is not essential for PS externalization. Fine topography of PLSCR1 was studied with electron microscopy on isolated and paraformaldehyde-fixed plasma membrane sheets in combination with immunogold labeling, and again, no evidence of any clustering or other topographic changes of the protein was detected in cells activated by various means, with the exception of calcium ionophore A23187. This drug did not change the extent of PLSCR1 clustering in the plasma membrane but decreased the protein density in the membrane. This could be due to the masking of target epitopes or release of PLSCR1 from the plasma membrane. The finding of an enhanced amount of PLSCR1 in the cytoplasm of A23187-activated cells, detected by confocal microscopy, supports the release hypothesis. By studying the fine topography of NTAL adaptor protein on the plasma membrane sheets isolated from control and A23187-activated cells, we found that NTAL, unlike PLSCR1, did not exhibit any changes; this excluded the possibility of generally enhanced internalization of plasma membrane proteins by the drug. The observed decrease of PLSCR1 in the plasma membrane, coupled with strong expression of surface PS in A23187-activated cells, supports again the conclusion that PS externalization is probably not dependent on enhanced PLSCR1 clustering. Our data also suggested that PS externalization in A23187-treated cells is in part directly induced by an interaction of calcium ionophore with the plasma membrane and not just by increased  (37,77). It should be noted that A23187-induced changes in PS externalization and PLSCR1 internalization were similar to those observed in apoptotic cells. However, only A23187-triggered cells showed partial reversion of the process, adding another piece of evidence that the continuous presence of A23187 is required for its effect. Finally, no simple relationship was found between the extent of PS externalization and secretory response. Strong degranulation was observed in Fc⑀RI-activated cells, whereas no degranulation took place in cells activated via dimerized Thy-1. However, both of these activation pathways resulted in comparable exposure of PS. On the other hand, A23187-activated cells showed extremely high PS externalization but weak degranulation when compared with other activation stimuli.
Using a biochemical approach, a distinct fraction of PLSCR1 was found in DIM if weak detergent (0.5% Brij 58) was employed. The more potent nonionic detergent, 1% Triton X-100, solubilized PLSCR1 completely. Under the same solubilization conditions, 80 and 60% of Lyn was recovered in 0.5% Brij 58-and 1% Triton X-100-resistant DIM, respectively, indicating that PLSCR1 and Lyn, although both are supposedly associated with lipid rafts, differ in detergent solubility. Because palmitoylation is often responsible for localization of proteins in DIM (78), it is possible that only a small fraction of membrane-associated PLSCR1 is palmitoylated.
Interestingly, two forms of PLSCR1 (36 and 38 kDa) were found in both DIM and high density fractions. It has been suggested that retarded migration of a fraction of PLSCR1 in PAGE could reflect multiple palmitoylation of PLSCR1 on clusters of cysteines that are present in the molecule (22). Palmitoylation of cysteines is often responsible for protein localization to DIM (75). However, because the two forms of PLSCR1 were present in similar ratios in both low and high density fraction (Fig. 8A), it is unlikely that multiple palmitoylation is responsible for retarded migration of the upper band. It is still possible that some other post-transcriptional modifications account for the observed size heterogeneity of PLSCR1. These poorly defined modifications could be responsible for different ratios between the 36 and 38 kDa bands observed in some experiments.
After Fc⑀RI-mediated activation, the amount of PLSCR1 in DIM was even reduced. In this respect, PLSCR1 differs from some other proteins with a transmembrane domain that exhibited enhanced association with DIM in activated cells (79). The finding that a clear decrease in amount of PLSCR1 in DIM observed in Fc⑀RI-activated cells is not accompanied by detectable changes in the distribution of PLSCR1 in plasma membrane sheets suggests that changes in biochemical parameters do not necessarily reflect topographical transpositions. This conclusion is supported by previous studies indicating that molecules assumed to be localized within and outside the lipid rafts exhibited similar topographic distribution on the plasma membrane and vice versa (52,58,80). Furthermore, these data support previous studies indicating that signaling domains formed on the plasma membrane could be very small, perhaps confined to individual protein molecules (58,81).
Although the amount of PLSCR1 in DIM isolated from resting cells was relatively low, most of the tyrosine-phosphorylated PLSCR1 was localized in DIM. Soluble PLSCR1, recovered in high density fractions, showed only weak tyrosine phosphorylation, suggesting that palmitoylation and presumable localization in lipid rafts protect PLSCR1 from dephosphorylation. After Fc⑀RI triggering, the amount of tyrosine-phosphorylated PLSCR1 increased not only in DIM but mainly in the fraction of soluble proteins. Treatment with pervanadate, which inhibits enzymatic activity of protein-tyrosine phosphatases and thereby increases the intracellular equilibrium between phosphorylation and dephosphorylation rates (82), resulted in an enhanced amount of tyrosine-phosphorylated PLSCR1 in both low density (DIM-containing) and high density fractions, suggesting that a week phosphorylation of PLSCR1 in high density fractions in resting cells is attributable to enzymatic activity of protein-tyrosine phosphatases. When this activity is inhibited, even in the absence of any detectable movement of the PLSCR1 into new membrane domains, kinases mediate the phosphorylation of PLSCR1. The observed tyrosine phosphorylation of PLSCR1 in pervanadate-activated cells supports the model that signaling assemblies containing protein-tyrosine phosphatases and kinases are preassociated before triggering and that in resting cells the kinase activity is counterbalanced by active phosphatases (83). Although human PLSCR1 has been shown to bind to c-Abl tyrosine kinase through its Src homology 3 domain and is phosphorylated most likely at Tyr 69 and Tyr 74 , the identities of kinases that phosphorylate PLSCR1 in mast cells remain to be determined.
In conclusion, our results indicate for the first time that changes in topography of PLSCR1 and its tyrosine phosphorylation, PS externalization, and secretory responses are independent phenomena that could be separated by employing various activation methods. Based on data in this study, we propose that enhanced tyrosine phosphorylation of PLSCR1 or its topographic changes in the plasma membrane are not required for PS externalization and degranulation. These data are in line with studies documenting that genetic removal of PLSCR1 does not lead to a decrease in phospholipid scrambling (27) and support the notion that PLSCR1 has other roles in cell signaling events affected through its interactions with various signaling molecules, such as epidermal growth factor receptor (23), ␤-secretase (30), and/or proteinase 3 (84).