We show that the human IgA receptor, FcαR, redistributes to plasma membrane rafts after cross-linking and that tyrosine kinases are relocated to these sites following FcαR capping. We demonstrate by confocal microscopy that FcαR caps in membrane rafts by a γ-chain-independent mechanism but that γ-chain expression is necessary for Lyn redistribution. Immunoblotting of rafts isolated by sucrose density gradient centrifugation demonstrated recruitment of γ-chain and phosphorylated tyrosine kinases Lyn and Bruton’s tyrosine kinase to membrane rafts after FcαR cross-linking. Time-dependent differences in Lyn phosphorylation and Bruton’s tyrosine kinase distribution were observed between cells expressing FcαR plus γ-chain and cells expressing FcαR only. This study defines early FcαR-triggered membrane dynamics that take place before FcαR internalization.
The human IgA receptor (FcαR)3 is a 50- to 70-kDa transmembrane glycoprotein expressed primarily by myeloid cells including neutrophils, monocytes, and macrophages (1, 2). Cross-linking FcαR with aggregates of IgA or IgA-opsonized particles triggers a variety of cellular responses including phagocytosis, oxidative burst, and degranulation in neutrophils and monocytes (3, 4). More recently, attention has turned to the role of FcαR-targeted Ag uptake and signaling following Ag uptake.4
Despite the range of functions triggered by FcαR, relatively little is known about the initial plasma membrane events that mediate association of FcαR with downstream signaling effectors. FcαR associates with the src family member tyrosine kinase p53/56-Lyn (5). Cross-linking FcαR triggers calcium release from intracellular stores in neutrophils (6), and respiratory bursts are inhibited by phosphatidylinositol (PI) 3-kinase inhibitors (7). In the monocytic cell line U937, FcαR cross-linking results in phosphorylation of the associated FcR γ-chain and phospholipase C (PLC) γ (8). FcαR on mesangial cells mediates phosphorylation of PLCγ1 that is linked to calcium mobilization through PI phosphate hydrolysis (9).
Signaling is accomplished by FcαR by its association with the FcR γ-chain to form the trimer FcαR/γγ (10). The FcR γ-chain dimer is also found in the high-affinity IgG (FcγRI) and IgE (FcεRI) receptor complexes expressed on mast cells and monocytes, respectively (11). Arguably, many of the early signaling events triggered by FcεRI and FcαR could be similar in terms of tyrosine kinase recruitment, and indeed the high-affinity IgE receptor FcεRI has been shown to activate the tyrosine kinases Lyn and Syk through FcR γ-chain (12, 13). More recently, the site of FcεRI activation in the plasma membrane has been investigated. Cross-linked FcεRI has been shown to redistribute to membrane domains rich in glycosphingolipids and cholesterol (14). These domains, termed membrane “rafts,” have estimated average sizes ranging from 70 nm (15) to 500 nm in diameter (16). Rafts are characterized by detergent insolubility and a high content of ganglioside GM-1, which is not a significant component of other plasma membrane domains (17). The redistribution of cross-linked FcεRI to membrane rafts is significant in that rafts are rich in signaling molecules such as tyrosine kinases (18, 19). Recruitment of tyrosine kinases to rafts by FcεRI has now been demonstrated in confocal microscopy studies (14).
The aim of this study was to determine whether FcαR redistributed to rafts and whether this was the site of tyrosine kinase recruitment and phosphorylation. Further, we wished to determine whether these events were dependent on FcR γ-chain expression. To circumvent the problems caused by studying FcαR function in cells expressing endogenous γ-chain, we cotransfected the cDNA for FcαR and γ-chain or FcαR alone into the B cell line A20 IIA1.6, which does not express other Fc receptors or γ-chain. We show that while cross-linked FcαR colocalizes with membrane rafts irrespective of γ-chain, recruitment of Lyn and Btk to the rafts and phosphorylation of these kinases was dependent on the presence of γ-chain.
Materials and Methods
The pCAV vector containing the human FcαR cDNA was the gift of Dr. C. Maliszewski (Immunex, Seattle, WA). Cells of the A20 IIA1.6 B cell line, which are Fc receptor negative (20), were either cotransfected with pCAV/FcαR cDNA and pNUT/γ-chain cDNA constructs or transfected with pCAV/FcαR cDNA by electroporation using a Bio-Rad electroporator (Bio-Rad, Richmond, CA) at 250 V, 960 μF. The pNUT vector allows selection using methotrexate.
B cell transfectants and culture
Transfectants expressing FcαR and γ-chain were cultured in RPMI 1640 medium supplemented with 10% FBS, 40 μg/ml gentamicin, 2 mM l-glutamine, 1 mM sodium pyruvate, and 0.9 mg/ml methotrexate. Transfectants expressing FcαR and no γ-chain were cultured similarly except that 0.8 mg/ml G418 was used as the selection agent instead of methotrexate. Levels of FcαR cell-surface expression were routinely monitered by flow cytometry using a Becton Dickinson FACScan (San Diego, CA).
Abs and fluorochromes
Anti-FcαR (My43) is a mouse IgM mAb produced in our laboratory (21). Polyclonal rabbit anti-Lyn or anti-Btk and agarose-conjugated PY20 anti-phosphotyrosine (PY) Ab were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-mouse transferrin receptor (TfR) Ab (biotin-conjugated) was purchased from PharMingen (San Diego, CA). Indocarbocyanine 3 (Cy3)-conjugated goat anti-mouse (GAM) IgM (μ) (Caltag Laboratories, Burlingam, CA) was labeled with Cy3 according to the manufacturers instructions (Molecular Probes, Eugene, OR). FITC-conjugated GAM-IgM (μ) was purchased from Caltag. Cy3-conjugated goat anti-rabbit (GAR) IgG (H+L), HRP-conjugated GAR-IgG, and FITC-conjugated streptavidin were purchased from Jackson ImmunoResearch Laboratories (West Grove, PA). FITC-conjugated cholera toxin (ChTx) (subunit B) was purchased from Sigma (St. Louis, MO).
FcαR capping and endocytosis
Cells were assessed for viability by ethidium bromide exclusion then adjusted to a density of 105 cells/ml in RPMI 1640 plus 10% FBS and adhered for 90 min at 37°C to chamber slides previously coated with 0.1 mg/ml poly-l-lysine (Sigma). Slides were chilled to 4°C and gently washed once with 4°C media to remove nonadherent cells. Cells were incubated with 20 μg/ml My43 in RPMI 1640 for 1 h at 4°C, followed by washing three times with medium. Surface-bound My43 was cross-linked with Cy3- or FITC-conjugated F(ab′)2-GAM μ-chain as appropriate at a concentration of 1 μg/ml. Samples were incubated for 45 min at 4°C before washing three times in 4°C media. Medium was removed and replaced with 37°C medium and incubated as indicated before fixation for 30 min at room temperature with 0.5% paraformaldehyde in PBS.
Cell-surface and intracellular staining of fixed cells
GM-1 in rafts was labeled for 45 min at room temperature with 0.1 μg/ml FITC-conjugated ChTx subunit B (22). Samples were then washed three times in PBS. For intracellular staining, cells were permeabilized with 0.5% saponin, 0.1% BSA, and 0.1% NaN3 in PBS (permeabilization buffer) for 15 min at room temperature. Anti-Lyn Ab was added at a concentration of 20 μg/ml in permeabilization buffer for 45 min before washing three times. Cy3-conjugated GAR-IgG (100 μl) was then added at 1.0 μg/ml and incubated for 45 min before washing three times in permeabilization buffer. Background staining was assessed by incubation of cells with fluorochrome-conjugated Ab alone. Specificity of Lyn staining was demonstrated with a Lyn-blocking peptide (Santa Cruz Biotechnology).
Laser scanning confocal fluorescence microscopy
Cover slips were mounted using Prolong Antifade (Molecular Probes), and cells were analyzed with a Bio-Rad MRC1000 laser scanning system equipped with a Kry/Arg laser and beam splitter to allow simultaneous two- and three-color imaging. Codistribution of FcαR and GM-1 or FcαR and Lyn was assessed by selecting cells with capped cell-surface FcαR. Images were then assessed for codistribution of GM-1 or Lyn At least 100 cells from random fields were imaged and counted for each time point. Images were analyzed using Adobe PhotoShop 4.0 software (Mountain View, CA).
Isolation of detergent-insoluble membrane domains
Isolation of rafts was performed by the method of Fra et al. with modifications (23). One hundred million cells were resuspended to 5 ml in culture medium and incubated at 4°C with HRP-conjugated ChTx (5 μg/ml) followed by PBS washing. Cells were lysed for 30 min in detergent extraction buffer (25 mM Tris, pH 7.6, 150 mM NaCl, 5 mM EDTA, 20 μg/ml each of chymostatin, leupeptin, antipain, pepstatin, 40 mM Na3VO4/200 mM NaF, 0.05% Triton X-100 (all from Sigma)) and adjusted to 1.5 M sucrose in 20 mM Tris, pH 7.5. Samples (3 ml) were added to 13-ml ultracentrifuge tubes and then overlaid with 7 ml of 1.2 M sucrose followed by a layer of 0.15 M sucrose. Samples were centrifuged at 38,000 rpm in a Beckman SW41 rotor for 18 h at 4°C. One-milliliter fractions were carefully withdrawn using a pipettor and assayed for peroxidase activity by luminol chemiluminescence (Amersham-Pharmacia Biotech, Piscataway, NJ) using an EG&G Berthold Microlumat 96V chemiluminometer. For FcαR cross-linking, cells were chilled to 4°C and incubated with My43 followed by F(ab′)2-GAM μ-chain before lysis. After washing, cells were warmed to 37°C and reactions were stopped by the addition of ice-cold PBS and placing tubes in an ice water bath. Cells were then pelleted and lysed by addition of 1 ml ice-cold detergent lysis buffer followed by 30 min further incubation before ultracentrifugation.
SDS-PAGE and immunoblotting
Raft fractions were adjusted to equal protein concentrations and equal amounts of protein resolved by SDS-PAGE under reducing conditions. Proteins were transferred to nitrocellulose membranes and incubated overnight at 4°C with 5% nonfat dry milk and 0.5% Tween 20 in PBS. Membranes were incubated with Abs to Lyn and Btk (0.2 μg/ml) for 2 h at room temperature and washed six times for 5 min in PBS and transferred to tubes containing 3% nonfat dry milk/0.05% Tween 20 in PBS. HRP-conjugated anti-rabbit IgG Ab was added at a 1/10,000 dilution (0.1 μg/ml) and incubated for a further 2 h at room temperature. Membranes were then washed six times for 5 min in PBS. Proteins were detected by enhanced chemiluminescence (ECL) (Amersham-Pharmacia Biotech, Piscataway, NJ).
Agarose-coupled anti-PY Ab (PY20; Santa Cruz Biotechnology) was incubated overnight at 4°C with raft or nonraft fractions (100 μg protein/sample) in 1% Nonidet P-40 then washed three times in 10 mM Tris, 2 mM Na3VO4, and 1% Nonidet P-40, pH 7.0, before washing once in 10 mM Tris, pH 7.6. Immunoprecipitated proteins were immunoblotted as described for Lyn and Btk. Where appropriate, immunoprecipitated proteins were treated with the tyrosine phosphatase LAR-D1. Briefly, beads were washed in 50 mM NaCl, 25 mM imidazole, 5 mM DTT, and 2.5 mM EDTA, pH 7.0, and resuspended in 20 μl buffer containing 5 U LAR-D1 (Calbiochem, San Diego, CA). Samples were incubated for 2 h at 37°C, and reactions were stopped with SDS-PAGE sample buffer.
FcαR caps in rafts independently of γ-chain expression
We generated FcαR+/γ-chain+ and FcαR+/γ-chain− IIA1.6 transfectants. Flow cytometric analysis indicated the levels of FcαR cell-surface expression were comparable between γ-chain+ cells and γ-chain− cells (data not shown). To examine the plasma membrane distribution of FcαR, we performed confocal microscopy analysis of γ-chain+ and γ-chain− cells. Midsection confocal images of γ-chain+ cells (Fig. 1,a1) demonstrated a punctate staining of FcαR (red) distributed evenly in the plasma membrane. Staining of the plasma membrane ganglioside GM-1 with FITC-ChTx subunit B (green) also showed a punctate distribution of GM-1 (Fig. 1,a2). Although the punctate staining of GM-1 and FcαR could potentially be caused by aggregation through clustering of FcαR at 4°C, this is unlikely because cells fixed at 4°C before binding of Ab to FcαR showed the same distribution of FcαR and GM-1 as cells in which FcαR was cross-linked at 4°C before fixation (data not shown). Additionally, 0 min cells (Fig. 1, a1, b1, and c1) were manipulated and fixed at 4°C without warming, which would minimize FcαR or GM-1 redistribution caused by cross-linking of FcαR. The plasma membrane distribution of FcαR and GM-1 in γ-chain− cells was similar to γ-chain+ cells (Fig. 1, a3 and a4). γ-Chain+ (Fig. 1,b1) and γ-chain− (Fig. 1 c1) cells costained for FcαR and GM-1 at 4°C and then fixed showed some overlap of red and green fluorescence (indicated by yellow). However, areas of plasma membrane were observed to stain for FcαR (red) but not GM-1 and conversely GM-1 (green) but not FcαR.
FcαR on γ-chain+ cells was cross-linked and incubated at 37°C for 2 and 5 min before fixation and costaining for GM-1. We consistently observed that cross-linked FcαR capped after warming cells to 37°C (Fig. 1, b2 and b3). In γ-chain+ cells, GM-1 was observed to cap and its distribution coincided with that of FcαR (indicated by yellow color). In γ-chain+ cells, cocapping of FcαR and GM-1 persisted for at least 5 min before complete receptor internalization by 10–20 min (Fig. 1,b4). We tested the dependence of FcαR capping to rafts on expression of γ-chain. In the γ-chain− transfectants, we observed capping of FcαR in membrane rafts within 2 min (Fig. 1,c2), which persisted at 5 min (Fig. 1,c3), before receptor internalization within 10–20 min (Fig. 1 c4).
The kinetics of FcαR capping, GM-1 capping, and FcαR internalization were similar in γ-chain+ and γ-chain− cells (Fig. 1, d–g). FcαR capping was observed within 30 s in 33% of γ-chain+ cells and 36% of γ-chain− cells (Fig. 1,d). Elevated levels of FcαR capping persisted for 5 min in both γ-chain+ and γ-chain− cells. GM-1 capping was observed to occur within 30 s (Fig. 1,e) in both γ-chain + (28%) and γ-chain− cells (40%). GM-1 capping persisted for 20 min and was still observed in 18% of γ-chain+ and γ-chain− cells, although FcαR capping was no longer observed. FcαR and GM-1 cocapping were observed to have similar kinetics in γ-chain+ and γ-chain− cells (Fig. 1,f). The percent of cocapped FcαR/GM-1 returned to baseline levels within 10 min in both cell types. Similar kinetics of FcαR internalization were observed in γ-chain+ and γ-chain− cells (Fig. 1 g). By 20 min, γ-chain+ and γ-chain− cells demonstrated equal (80%) internalization of cell-surface FcαR.
GM-1 is not internalized with FcαR
To determine whether there was internalization of GM-1 when FcαR internalization was observed, cells were permeabilized with saponin to permit detection of internal GM-1 and then labeled with FITC-ChTx. When cells were fixed at 4°C before permeabilization and staining for GM-1, we did not observe internal GM-1 staining (Fig. 2,a). After internalization of FcαR, there was no detectable intracellular GM-1 and no noticeable difference in cell-surface levels of GM-1 (Fig. 2 b). Thus it appears that FcαR is internalized without GM-1 internalization, suggesting that FcαR leaves rafts before endocytosis.
Capped FcαR is selective for rafts
To further demonstrate selectivity of FcαR for rafts over other plasma membrane domains, we compared the distribution of FcαR with that of the TfR. The TfR is excluded from rafts and has a mutually exclusive distribution with respect to GM-1 (17). Thus, it is a useful marker for nonraft plasma membrane. FcαR was stained with My43 plus Cy3-GAM-IgM as described. The cells were then fixed at 4°C and permeabilized before staining for TfR with biotin-conjugated anti-TfR Ab followed by FITC-conjugated streptavidin (Fig. 3,a). Two pools of TfR were observed in cells, a pool of intracellular transferrin and a pool of cell-surface TfR. Almost no colocalization of FcαR and TfR was observed in cells fixed without warming to 37°C, indicating that FcαR and TfR partition into different plasma membrane domains. After warming cells to 37°C to cap cross-linked FcαR (Fig. 3 b), TfR staining did not colocalize with FcαR. This confirms our observation that FcαR cocaps with the raft marker GM-1 and that capped FcαR is selective for membrane rafts over other plasma membrane domains.
Lyn redistributes to FcαR caps in a γ-chain-dependent manner
Temporary localization of FcαR in rafts could increase the proximity of cross-linked FcαR/γγ complexes with raft-associated tyrosine kinases (18). We costained γ-chain+ cells for FcαR and the tyrosine kinase p53/56-Lyn (Fig. 4 a). In the steady state (formaldehyde fixed at 0 min), Lyn had a diffuse, subplasma membrane distribution. After 2 min of FcαR cross-linking at 37°C, we observed a transient redistribution of Lyn. The distribution of Lyn changed such that it was largely codistributed with FcαR. By 5 min, polarization of Lyn was not evident, although FcαR was still capped. This indicates that Lyn is recruited to FcαR/γγ complexes in membrane rafts before FcαR internalization.
There is no direct evidence for binding of Lyn to γ-chain, but γ-chain immunoreceptor tyrosine-based activation motifs are substrates for Lyn (24). Therefore we hypothesized that codistribution of FcαR and Lyn was dependent on expression of γ-chain (Fig. 4,b). Although Lyn has N-terminal myristate and palmitate chains (25), our data suggest that Lyn was not completely partitioned into the rafts. We observed that Lyn was evenly distributed in regions of plasma membrane in unstimulated cells with no GM-1 staining. In γ-chain− cells, there was no change in the distribution of Lyn, although FcαR was observed in a cap at 2 min and 5 min (Fig. 4,b). Cells were assessed for codistribution of FcαR and Lyn after 2 min and 5 min at 37°C (Fig. 4,c). γ-Chain+ and γ-chain− cells with capped FcαR were analyzed for polarized Lyn staining, which codistributed with capped FcαR. At 2 min, when FcαR capping was maximal, 70% of γ-chain+ cells with capped FcαR showed clear codistribution of Lyn. In γ-chain− cells, at 2 min, 10% of cells demonstrated codistribution of polarized Lyn. By 5 min, the difference in FcαR:Lyn copolarization between γ-chain+ and γ-chain− cells was less apparent than at 2 min, with 26% of γ-chain+ cells showing FcαR:Lyn copolarization compared with 7% in γ-chain− cells. To confirm that redistribution of Lyn to FcαR caps was occurring at membrane rafts, we performed three-color confocal microscopy staining for FcαR (red), GM-1 (green), and Lyn (blue). At 0 min, cells had an even distribution of both FcαR, GM-1, and Lyn (Fig. 5). After 2 min at 37°C, FcαR, GM-1, and Lyn had each redistributed and cocapped to a pole of the cell.
γ-Chain-dependent recruitment and activation of Lyn in membrane rafts
Membrane rafts are resistant to solubilization by the detergent Triton X-100. Low levels of Triton X-100 will cause dissolution of nonraft plasma membrane, leaving intact not only membrane rafts but associated proteins (26). We labeled the cell-surface GM-1 with HRP-conjugated ChTx (subunit B) before lysis in 0.05% Triton X-100. After sucrose density gradient centrifugation of lysates, a distinct opaque band was observed at the interface between the 0.15 M and the 1.2 M sucrose layers. HRP activity as assessed by luminol chemiluminescence was detected mainly in the fraction corresponding to this band (Fig. 6,a). Overall, 76% of the HRP activity corresponding to the raft constituent GM-1 was distributed across fractions 3, 4, and 5, verifying that the detergent insoluble domains were recovered by this technique. HRP-ChTx incubated lysates were centrifuged in parallel to lysates from cells in which FcαR had been cross-linked. The ChTx-treated samples were not used in signaling experiments described in this study but were used to demonstrate the location of rafts in the sucrose gradients. Raft fractions from cells where FcαR had been cross-linked were adjusted to the same protein concentration and used for immunoblotting experiments. Each fraction from sucrose gradients of γ-chain+ and γ-chain− cells was immunoblotted for Lyn (Fig. 6, b and c). In absence of FcαR cross-linking, there was no observable difference between γ-chain+ and γ-chain− cells in the distribution of Lyn between raft and nonraft fractions. While Lyn was observed at higher levels in raft fractions, significant levels were observed in nonraft fractions also. This indicates that in unstimulated cells Lyn does partition into membrane rafts but is also found in nonraft fractions.
Significant levels of both isoforms Lyn of were detected in raft fractions of both γ-chain+ and γ-chain− unstimulated cells (Fig. 7,a). In γ-chain+ cells, after 30 s of FcαR cross-linking at 37°C, increased levels of Lyn were detected compared with the 0-min samples, and additional anti-Lyn-reactive proteins of a higher molecular mass than the p53/56 bands observed in unstimulated cells were detected. After 2 min of FcαR cross-linking, the same banding pattern was observed, which became less pronounced after 5 min. The higher molecular mass protein reactive with anti-Lyn Ab suggests that phosphorylation of Lyn was occurring in response to FcαR cross-linking in γ-chain+ cells. In the γ-chain− cells, we did not observe the appearance of anti-Lyn-reactive higher molecular mass protein. In fact, the amount of Lyn in γ-chain− cell rafts after FcαR cross-linking for 30 s, 2 min, and 5 min decreased (Fig. 7 a). The decrease in Lyn content in rafts from γ-chain− cells was not due to an increase in total protein in the samples. After FcαR cross-linking, total protein concentration in the raft fraction remained constant.
To show that the higher molecular mass proteins detected with anti-Lyn Abs represented phosphorylated species, cells were lysed in the absence of the phosphatase inhibitor orthovanadate. In the absence of phosphatase inhibition, we did not observe the higher molecular mass protein (Fig. 7 b). Additionally, the amount of Lyn in the rafts did not appear to increase in the absence of vanadate suggesting that Lyn phosphorylation is necessary for its increased levels in the rafts.
In other experiments to confirm that the higher molecular mass proteins were phosphorylated Lyn species, we used an agarose-conjugated anti-PY Ab to immunoprecipitate phospho-proteins from raft fractions. These samples were then analyzed by immunoblotting with an anti-Lyn Ab (Fig. 7,c). In the γ-chain+ cells, we detected a doublet of phosphorylated Lyn at a higher molecular mass (60 and 63 kDa) in addition to the p53/56 doublet, which was most evident in the 30-s and 2-min samples (Fig. 7,c). In the γ-chain− cells, these higher molecular mass phosphorylated proteins were not observed, although tyrosine-phosphorylated p53/56-Lyn was precipitated by the anti-PY Ab. Inactive p53/56 Lyn has a carboxyl-terminal tyrosine phosphate group, which is removed by CD45 to allow binding to target proteins via its SH2 domains (27). The p53/56 bands may represent these inactive species. A final experiment to confirm that the higher molecular mass proteins observed in Fig. 7, a and c represent tyrosine-phosphorylation of Lyn was to treat rafts from γ-chain+ cells where FcαR had been cross-linked for 30 s at 37°C with the tyrosine-specific phosphatase LAR-D1 (Fig. 7 d). Following LAR-D1 treatment, anti-Lyn Ab did not detect the higher molecular mass proteins observed in the untreated immunoprecipitate.
Detection of phosphorylated γ-chain in rafts
Raft fractions from γ-chain+ and γ-chain− cells were resolved by SDS-PAGE under nonreducing conditions and immunoblotted for γ-chain (Fig. 8). Only γ-chain+ cells showed γ-chain as expected. At 0 min, two protein species were detected with molecular masses of 22 kDa and 24 kDa as previously observed for nonreduced γ-chain dimer (8). After FcαR cross-linking for 30 s, 2 min, and 5 min, two additional proteins reactive with anti-γ-chain Ab were detected with higher molecular masses than the 22 kDa and 24 kDa proteins. There was also a substantial increase in the level of γ-chain observed in rafts after FcαR cross-linking, which was maximal at 30 s and 2 min. This data correlates with observed increases in FcαR in rafts by confocal microscopy (Fig. 1). Interestingly, by 5 min, although the same anti-γ-chain-reactive bands were observed, the amount of γ-chain had decreased despite the fact that equal amounts of protein were loaded.
γ-Chain-dependent recruitment of Btk to membrane rafts
Btk is a member of the Tec family of protein tyrosine kinases expressed primarily in hemopoietic cells. Btk is activated by the B cell Ag receptor (28) and activates src family kinases including Lyn (29). Compared with Lyn, low levels of Btk were detected in the raft fractions of unstimulated γ-chain+ and γ-chain− cells (Fig. 9,a). In γ-chain+ cells following FcαR cross-linking, there was an increase in the levels of Btk in the raft fractions over 5 min, suggesting that Btk is recruited to the rafts on FcαR ligation. Recruitment and maintenance of Btk to the rafts is dependent on γ-chain expression. In the γ-chain− cells, a low level of Btk was detected in the raft fraction of unstimulated cells, which was comparable to that observed in γ-chain+ unstimulated cells. The amount of Btk in these fractions did not increase with time, but decreased. When we stripped and reprobed the anti-Lyn blots from our immunoprecipitation experiments (Fig. 7,c), we observed that Btk in membrane rafts was phosphorylated (Fig. 9,b). In γ-chain+ cells, an increase in the level of tyrosine-phosphorylated Btk was observed after FcαR cross-linking. In γ-chain− cells, low levels of phosphorylated Btk were detected in unstimulated cells, but this decreased after FcαR cross-linking. In γ-chain+ cells, increased tyrosine-phosphorylated Btk in rafts was matched by a concomitant decrease in phosphorlyated Btk levels in the nonraft fraction (Fig. 9,c). In γ-chain- cells, phosphorylated Btk levels in the nonraft fraction did not decrease but were maintained and increased slightly on FcαR cross-linking (Fig. 9,d). The minor lower band in Fig. 9 c may represent slight degradation of Btk in the samples. This data shows that in the absence of γ-chain, phosphorylated Btk is not targeted from the nonraft fraction to the raft fraction.
We have demonstrated for the first time that, after cross-linking, FcαR caps transiently in rafts before endocytosis, and that capping is associated with γ-chain-dependent recruitment of tyrosine kinases Lyn and Btk. FcαR capping represents a large scale reorganization of the plasma membrane and also results in capping of the raft component GM-1. Within 2 min of FcαR cross-linking, 20–30% of cells showed colocalization of FcαR and GM-1, and this persisted for 5 min before FcαR internalization. The mechanism of FcαR redistribution into rafts is unclear. Redistribution of FcαR into rafts occurred independently of γ-chain expression, showing that neither γ-chain structure nor γ-chain-triggered signal transduction is required. Rafts are less fluid than surrounding plasma membrane domains because intercalation of cholesterol between the acyl chains in rafts allows tighter packing of sphingolipid molecules (30). Cross-linking could change the conformation of FcαR, decreasing its membrane solubility and lateral mobility, and could alter the equilibrium of FcαR entering and leaving rafts to trap FcαR in rafts.
The similar kinetics of GM-1 and FcαR capping suggest that capping of GM-1 is driven by FcαR entering the rafts. In studies of association of the IgE receptor (FcεRI) with rafts (14), the formation of several clusters of FcεRI was accompanied by colocalized clustering of GM-1, suggesting that receptor distribution drives raft distribution. An alternative explanation is that cross-linked FcαR and associated rafts become linked with the actin cytoskeleton. The work of Moran and Miceli (31) demonstrated that during CD48/TCR costimulation of T cells, association of ζ-chain with the actin cytoskeleton was enhanced in a raft-dependent manner. Their results suggest rafts are a site where signal transduction and cytoskeletal reorganization are integrated.
While we did not observe internalization of GM-1, FcαR was found to be completely internalized 20 min after cross-linking. Cells permeabilized then stained for GM-1 after FcαR internalization showed no internal GM-1 or colocalization with FcαR-containing vesicles, indicating that either FcαR leaves rafts before internalization or that GM-1 is excluded from rafts during the formation of the endocytic vesicles. There is evidence that the endocytic mechanism prevents raft internalization. In common with many other surface receptors, endocytosis of Fcγ receptors occurs via clathrin-coated pits (32). Studies in lymphoid cells have shown that plasma membrane structures called caveolae, which have the same lipid composition as rafts, are excluded from clathrin-coated pits (33). If FcαR associates with clathrin-coated pits for internalization, then exclusion of raft lipids (GM-1) from clathrin-coated pits would be consistent with our observation of the physical separation of GM-1 and FcαR after 10–20 min of cross-linking.
We observed that Lyn redistributed to FcαR capped in rafts in a γ-chain-dependent manner. This suggests that redistribution of Lyn could occur through increased association of Lyn with the phosphorylated γ-chain immunoreceptor tyrosine-based activation motif (8). We have observed by confocal microscopy that Lyn was evenly distributed around the cell periphery and that Lyn was found in both raft and nonraft fractions by immunoblotting This shows that although N-terminally acylated Lyn partitions into both raft and nonraft domains. Therefore, lipid chains do not result in exclusive partitioning of Lyn into rafts. Our examination of rafts isolated from γ-chain+ and γ-chain− cells after FcαR cross-linking supports this hypothesis. Equal levels of Lyn were detected in rafts of γ-chain+ and γ-chain− cells before FcαR cross-linking, but only γ-chain+ cells demonstrated an increased level of Lyn in rafts due to Lyn γ-chain association. We observed that lysis in the absence of vanadate prevented the increase in Lyn content of the rafts normally observed in γ-chain+ cells. Thus, when cells were lysed and rafts isolated under conditions in which Lyn and other components of the proximal signaling complex would become dephosphorylated, increased amounts of Lyn were no longer detected in the raft fraction. The retention in rafts of increased amounts of Lyn in the presence of vanadate suggests that interaction between phosphorylated Lyn and the other signaling molecules intimately associated with rafts helped to retain Lyn in the raft fraction.
Btk activation has been implicated in FcαR-triggered signal transduction (34). We observed that the levels of Btk in the raft fraction increased in a γ-chain-dependent manner and that, like Lyn, Btk was phosphorylated following FcαR ligation. Btk is not acylated and was thus recruited into the raft fraction from the cytosol (29). Recruitment of Btk into rafts is most likely mediated by binding of Btk to membrane PI 3,4,5-triphospate by its pleckstrin homology domain (35), indicating that PI 3-kinase is also activated in the raft-associated signaling complex. If PI 3-kinase is dependent on γ-chain for activation, then lack of signaling in the γ-chain− cells fits well with the observed lack of recruitment of Btk to the rafts. Our data suggest signaling is required not only to increase Btk levels in the rafts, but to maintain them. Collectively, our data suggest that the function of rafts is to provide an environment where phosphorylation of plasma membrane tyrosine kinases is initiated before FcαR internalization. Lyn and Btk phosphorylation occurs within 30 s and continues for 2 min before FcαR internalization, which occurs after around 5 min and is not complete till at least 10 min. This indicates that signals triggered by FcαR in rafts activate downstream events independently of delivery of FcαR to the endosomal pathway. The novel role of γ-chain in regulating Lyn and Btk localization and phosphorylation links signal transduction and membrane localization. The importance of integrating signaling and changes in the plasma membrane orientation is required for understanding how FcαR triggers downstream changes in cell physiology.
We thank the following for their technical support and advice: Mrs. Hong Gao, Mrs. Gillian Lang, Mr. Kenneth Orndorff, Dr. Alice Givan, and Dr. Grant Yeaman.
This work was supported by a grant from the National Institutes of Health (RO1AI22816) and by a Dartmouth College Hitchcock Foundation Grant (250479).
Abbreviations used in this paper: FcR, Ig Fc-domain receptor; Btk, Bruton’s tyrosine kinase; Cy3, indocarbocyanine 3; Cy5, indodicarbocyanine 5; ChTx, cholera toxin; ECL, enhanced chemiluminescence; PY, phosphotyrosine; TfR, transferrin receptor; PI, phosphatidylinositol; PLC, phospholipase C; GAM, goat anti-mouse; GAR, goat anti-rabbit.
L. Shen, M. van Egmond, K. Siemasko, M. Clark, J.G.J. van de Winkel, and W.F. Wade. Presentation of ovalbumin internalised via the IgA Fc receptor (CD89) is enhanced through FcR γ chain signaling. Submitted for publication.