Syk-mediated tyrosine phosphorylation is required for the association of hematopoietic lineage cell-specific protein 1 with lipid rafts and B cell antigen receptor signalosome complex.

Hematopoietic lineage cell-specific protein 1 (HS1) is an F-actin- and actin-related proteins 2 and 3 (Arp2/3)-binding protein that undergoes a rapid tyrosine phosphorylation upon B cell antigen receptor (BCR) activation. Density gradient centrifugation of Triton X-100 lysates from B lymphocytes demonstrated that HS1 was translocated in response to BCR cross-linking into lipid raft microdomain along with Arp2/3 complex and Wiskott-Aldrich syndrome protein. HS1-green fluorescent protein was localized in membrane patches enriched with GM1 gangliosides and BCR in the cells treated with anti-IgM antibody. Colocalization of HS1-green fluorescent protein with BCR was also correlated with tyrosine phosphorylation of HS1. Interestingly a murine HS1 mutant at the tyrosine residues Tyr388 and Tyr405 targeted by Syk failed to respond to BCR cross-linking for either translocation into lipid rafts or colocalization with BCR within cells. Furthermore HS1 was unable to translocate into lipid rafts in a chicken B cell line deficient in Syk. Reintroducing a Syk construct into the Syk knock-out cells recovered effectively both tyrosine phosphorylation and translocation of HS1 into lipid rafts. In contrast, translocation of HS1 into rafts was normal in a Lyn knock-out B cell line, and an HS1 mutant at the tyrosine residue Tyr222 targeted by Lyn maintained the ability to partition into rafts upon BCR cross-linking. These data indicate that Syk plays an important role in the translocation of HS1 into lipid rafts and may be responsible for actin assembly recruitment to rafts and subsequent antigen presentations.

Immune response in mammals is initiated by coordinated recognition of foreign antigens through cell surface receptors on lymphocytes. B cell antigen receptor (BCR), 1 the complex of membrane-bound IgM, Ig␣, and Ig␤, is the primary machinery to recognize antigens and provoke signaling cascades for proliferation and maturation of B lymphocytes into specific memory cells (1). Recent studies indicate that BCR-mediated signaling events often occur in membrane microdomains or lipid rafts that are rich in glycosphingolipid and cholesterol (2)(3)(4). Lipid rafts tend to select certain types of proteins and exclude others (4), thereby creating a specific environment wherein a protein can be phosphorylated by local tyrosine kinases and may become prone to interact with other local proteins in the same local environment. In lymphocytes lipid rafts undergo frequent clustering and form membrane patches upon receptor cross-linking. Consequently the lipid raft-associated molecules often display asymmetric distribution. Many processes of membrane clustering appear to be dependent on reorganization of the actin cytoskeleton meshwork, which is associated with lipid rafts on the inner side of the plasma membrane (5). The actin assembly is also necessary for the maintenance of the integrity of lipid raft (6), the antigen transportation into endosomes following BCR cross-linking (7), and the formation of immunological synapses (8). However, the mechanism for the recruitment of actin assembly to lipid rafts is unclear.
Upon receptor cross-linking, actin assembly is rapidly initiated in lipid rafts in a tyrosine phosphorylation-dependent manner (9). Interestingly BCR cross-linking also provokes translocation of BCR itself into lipid rafts in an actin-independent manner (2), suggesting that BCR in the rafts may trigger a signal transduction that ultimately leads to actin assembly. It is known that an early phase of the BCR-mediated signal transduction involves activation of several non-receptor protein-tyrosine kinases including Lyn and Syk, thereby resulting in tyrosine phosphorylation of multiple intracellular proteins (10). One of the prominent phosphorylated proteins upon BCR cross-linking is HS1, which is expressed exclusively in cells of hematopoietic and lymphoid origins (11). HS1 is structurally related to cortactin, a cortical actin-associated protein that is expressed in most adherent cells and implicated in the actin assembly mediated by Arp2/3 complex (12). Like cortactin, HS1 contains a characteristic repeated domain comprised of three and one-half 37-amino acid repeat units, a Src homology 3 (SH3) domain at the C terminus, and an Arp2/3 binding do-main at the N terminus (Fig. 1A). Our recent study has demonstrated that HS1 promotes actin assembly and branching by binding to both Arp2/3 complex and F-actin (13). A pathological significance of HS1 binding to F-actin has been indicated in a recent report that aberrant expression of an HS1 mutant lacking a functional domain for F-actin binding is genetically associated with systemic lupus erythematosus, an autoimmune disease characterized by the activation of autoreactive B lymphocytes (14). Studies with HS1 knock-out mice, which displayed a defect in antigen-induced clonal expansion and lymphocyte deletion (15), demonstrated that HS1 is intimately implicated in BCR and T cell receptor signalings. There is also evidence showing that HS1 is required for the apoptotic response to BCR cross-linking because B cells with low levels of HS1 expression are apparently resistant to apoptosis (16). However, the molecular mechanism for the function of HS1 in BCR signaling remains to be established.
Biochemical studies have established HS1 as a prominent substrate of protein-tyrosine kinases Syk and Src family proteins including Lyn, Fgr, Fyn, and Lck (17)(18)(19)(20). Direct association of HS1 with Lyn or Lck is evident in T and B lymphocytes and erythroid cells during the differentiation mediated by erythropoietin (21)(22)(23). A complex of HS1 and Lck with human immunodeficiency virus type-1 virions was also reported (24). Although HS1 can be targeted by multiple tyrosine kinases, phosphorylation of HS1 at a full level requires a sequential and coordinated process involving Syk and Src. Phosphorylation is first initiated by Syk at Tyr 388 /Tyr 405 in murine HS1 or Tyr 378 / Tyr 397 in human HS1 and subsequently by Src-related kinases at Tyr 222 , presumably due to an enhanced interaction between phosphorylated HS1 and the SH2 domain of Src kinases (20,21). However, the physiological role of HS1 phosphorylation has not yet been illustrated. In this study, we show that HS1 was recruited along with Arp2/3 complex and WASP into lipid rafts and associated with BCR upon BCR cross-linking in a tyrosine phosphorylation-dependent manner. We also show that the process of HS1 translocation into lipid rafts requires Syk and its mediated tyrosine phosphorylation. Thus, our data imply that tyrosine phosphorylation of HS1 mediated by Syk may be an important mechanism to induce actin assembly within lipid rafts in response to antigen stimulation.
Plasmid Constructions-Plasmid pHJ117, which encodes HS1-GFP, was prepared by PCR. Briefly a murine HS1 cDNA clone from ATCC was used as the template in PCR using CCGGAATTCATGTGGAAGT-CTGTAGTGGGGC and CGCGGATCCAGGAGCTTGACATAGTTTG-CAGG (EcoRI and BamHI sites are underlined) as primers. The resulting PCR product was inserted into the EcoRI and BamHI sites of pEGFP-N1 (Clontech). To prepare virus HS1-GFP, the EcoRI-NotI fragment of HS1-GFP from pHJ117 was inserted into the EcoRI and NotI sites of retroviral vector MGIN (a gift of Dr. Robert Hawley, Holland Laboratory), resulting in plasmid pHJ124. To prepare pHJ133 encoding HS1 Y388F/Y405F -GFP, point mutations at Y388F and Y405F were generated using the Transformer site-directed mutagenesis kit (Clontech) based on pHJ117. The resulting DNA fragment was inserted into the EcoRI and NotI sites of MGIN.
Plasmid pHJ172, which encodes human Syk tagged by red fluorescent protein (RFP) at its C terminus, was prepared by PCR using pCDNA3.1-Syk (a gift of Dr. Susette C. Mueller) as the template and CCGGAATTCATGGCCAGCAGCGGCATGGC and CGCGGATCCCGG-TTCACCACGTCATAGTAGTA (EcoRI and BamHI sites are underlined) as primers. The resulting PCR product was inserted into the EcoRI and BamHI sites of pDsRed-Express-N1 (Clontech). The EcoRI-NotI fragment encoding Syk-RFP was isolated from pHJ172 and further inserted into the EcoRI and NotI sites of retroviral vector MGIN, resulting in plasmid pHJ179.
Virus Preparation and Viral Infection-Retrovirus-packaging cells (293GPG) were the gift of Dr. Richard C. Mulligan, Harvard Medical School (Boston, MA) and were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 1 mM sodium pyruvate, 2 mM L-glutamine, antibiotics, 2 g/ml puromycin, and 1 g/ml tetracycline. Packaging cells (1 ϫ 10 6 ) were transfected with pHJ124 or pHJ133 using Superfect transfection reagent (Qiagen Inc., Valencia, CA). To harvest viruses, the transfectants were grown in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, 1 mM sodium pyruvate, and 2 mM L-glutamine. The medium of the transfectants was collected at 48, 72, and 96 h after transfection and filtered through a 0.45-m filter (Gelman Sciences, Ann Harbor, MI). The virus medium was stored at Ϫ70°C.
For viral infection, WEHI-231 cells or DT40 cells were plated at a density of 2 ϫ 10 5 in 35-mm dishes. On the next day, the medium was replaced with 1 ml of fresh medium containing 8 g/ml Polybrene and 1 ml of viral supernatant. After 24 h of incubation, the culture medium was changed to the same fresh medium containing 1 mg/ml G418 for WEHI-231 cells or 2 mg/ml G418 for DT40 cells. Expression of GFP proteins was verified by fluorescence microscopy. To increase the efficiency of infection, the cells were infected with the same viruses for two or three times. After 2 weeks of selection in the medium containing G418, the cells were sorted in a fluorescence-activated cell sorting system (BD Biosciences) according to light scatter and fluorescence intensity. To express human Syk into DT40Syk Ϫ/Ϫ cells bearing HS1-GFP, the cells were reinfected with the virus pHJ179. Expression of Syk-RFP was verified by fluorescence microscopy and Western blot.
Immunoprecipitation and Immunoblotting-Cells (1 ϫ 10 7 ) to be analyzed were washed once with phosphate-buffered saline (PBS), suspended in 1 ml of a serum-free medium, and treated with 20 g/ml anti-IgM antibodies at 37°C. After incubation for the times as indicated, the cells were mixed with 6 ml of ice-cold PBS. Following one wash with cold PBS, cells were lysed in 500 l of TNE buffer (1% Triton X-100, 10 mM Tris-HCl, pH 7.5, 150 mM NaCl, 2 mM Na 3 VO 4 , and 5 mM EDTA) containing protease inhibitor mixture (Roche Applied Science) and centrifuged briefly to remove insoluble materials. The supernatants were immunoprecipitated with protein G-Sepharose coupled with HS1 monoclonal antibody or GFP monoclonal antibody. In some experiments, immunoprecipitation was performed with protein A-Sepharose coupled with appropriate polyclonal antibodies. The immunoprecipitates and aliquots of cell lysates were resolved by 10% (v/v) SDS-PAGE, transferred to Millipore Immobilon polyvinylidene difluoride membrane, and immunoblotted using appropriate primary antibodies followed by secondary peroxidase-conjugated antibodies (Bio-Rad). Blots were visualized with enhanced chemiluminescence (ECL, Amersham Biosciences) and autoradiography. Gel quantification was performed based on analysis of scanned blot images using Scion Image software.
Lipid Raft Isolation-Lipid rafts were isolated by lysis of cells in Triton X-100 followed by sucrose density gradient centrifugation as described previously (3). In brief, WEHI-231 or DT40 cells (1 ϫ 10 8 ) in 10 ml of growth medium were stimulated with goat anti-IgM antibody (20 g/ml) at 37°C for 5 or 15 min. The treated cells were collected and lysed for 30 min on ice in TNE buffer containing protease inhibitor mixture. The lysates were homogenized with a Dounce homogenizer and centrifuged at 900 ϫ g for 10 min to remove nuclei and cellular debris. All procedures were performed on ice. The clarified supernatants were diluted 1:1 with 85% sucrose in TNE buffer, and 2 ml of the solution was layered at the bottom of a Beckman 14 ϫ 89-mm centrifuge tube. The lysate was then overlaid with 6 ml of 35% sucrose and 4 ml of 5% sucrose in TNE buffer. The samples were centrifuged at 38,500 rpm in an SW41 rotor for 18 -20 h at 4°C. Twelve fractions of 1 ml each were collected from the top of the gradient after centrifugation. Aliquots of each fraction with the same volume were resolved by 10% (v/v) SDS-PAGE and immunoblotted with appropriate antibodies. In some experiments, the blot membrane was stripped and reblotted with antibodies as indicated.
Immunofluorescence Microscopy-To analyze colocalization of HS1-GFP or HS1 Y388F/Y405F -GFP with GM1, cells grown at log phase were stained with 25 g/ml TRITC-CTB for 20 min on ice and centrifuged briefly to remove unbound reagents. The pellets were resuspended in a serum-free medium and then incubated with 20 g/ml anti-IgM antibodies at 37°C for 5 min. The stimulated cells was washed once with cold PBS, resuspended in 50 l of cold PBS, and placed on ice. Before microscopic inspection, 5 l of the cell samples was transferred to a glass slide, covered with a glass coverslip, and inspected under a confocal microscope. The images captured by the digital camera on the microscope were further processed with Adobe Photoshop software. To analyze colocalization of HS1-GFP or HS1 Y388F/Y405F -GFP with BCR, cells grown at log phase were stimulated by addition of 25 g/ml TRITC-conjugated goat anti-mouse IgM followed by incubation at 37°C for the times as indicated. For the time point at 0, cells were placed on ice for 5 min and incubated with TRITC-anti-IgM antibody on ice without shifting to 37°C. The stimulated cells were washed once with cold PBS, resuspended in 50 l of cold PBS, and examined as above.

HS1 Is Cotranslocated into Lipid Rafts with Arp2/3 Complex and WASP-
In an effort to explore the signaling pathway in which HS1 is involved, we examined the potential of HS1 to associate with lipid rafts of WEHI-231 B lymphocytes. The B cells were stimulated with anti-mouse IgM ( chain) antibody, which cross-links BCR on the cell surface and subsequently triggers a signal cascade to apoptosis (25). The treated cells were lysed in 1% Triton X-100, and the lysates were subjected to a sucrose gradient centrifugation (3). Fractions of 1 ml were collected from the top of the centrifuge tube and analyzed for the presence of HS1 by immunoblotting with HS1 antibody. As shown in Fig. 1B, left panel, most HS1 proteins in non-treated cell lysates were found in fractions 10 -12 near the bottom of the centrifugation tube. These fractions are rich in detergentsoluble materials as well as cytoskeletal proteins and thereby referred to here as high (density) fractions. Small amounts of HS1 proteins were also detected in fractions 4 -6 in which lipid rafts are normally located (4). Lipid rafts in those fractions were verified by the presence of GM1 gangliosides and proteintyrosine kinase Lyn, both of which are known to associate constitutively with lipid rafts (3) (Fig. 1B). Upon BCR crosslinking, more HS1 proteins were apparently partitioned into rafts as evidenced by a gradual shifting from fraction 12 to fraction 5. HS1 was also found in the fractions 7-9 between high density and raft fractions of the stimulated cells. In contrast, CD71, the receptor for transferrin that is known to be a non-raft-associated protein (26,27), was predominantly found in the high fractions of the cells treated with or without anti-IgM antibody. Trace amounts of CD71 detected in the raft fractions of non-treated cells may be caused by an incomplete solubility of membranes in the lysates, which might also explain the presence of some HS1 proteins in the raft fractions of non-stimulated cells. To ensure that the shifting of HS1 to the lighter fractions was not due to a nonspecific raft or membrane were incubated with or without anti-mouse IgM antibody (20 g/ml) for 5 min at 37°C and lysed in 1% Triton X-100-containing buffer. The lysates were subjected to sucrose gradient centrifugation as described under "Methods and Materials." A total of 12 1-ml fractions was collected from the top of the tube after centrifugation. Aliquots with equal volumes of the fractions were analyzed by 10% (v/v) SDS-PAGE, transferred to a polyvinylidene difluoride membrane, and immunoblotted with peroxidase-conjugated CTB to detect GM1 ganglioside or blotted with anti-HS1, -Arp3, -WASP, -Lyn, -IgM, and -CD71 antibodies on the same membrane, respectively. Fractions 4 -6 were labeled as Raft (fractions); fractions 10 -12 are labeled as High (density fractions). As a negative control, the fractions were also probed with anti-histone H1 antibody, showing its exclusive association with high fraction 12. The data shown are representative of three independent experiments. T, total cell lysates. association, the same fractions were also probed with antibody against histone H1, which is not a membrane-associated protein. This analysis demonstrated histone H1 was exclusively distributed in high fractions of either stimulated or non-stimulated cells (Fig. 1B). Thus, the protein associated with fractions 7-9 may represent a transition form during the process of the recruitment of HS1 to lipid rafts.
In addition to HS1, we also analyzed Arp2/3 complex and WASP as these molecules are also implicated in actin dynamics. Arp2/3 complex was detected by antibody against Arp3, a subunit of Arp2/3 complex (13). Like HS1, these proteins were present predominantly in the high fractions of non-treated cells but also in the raft fractions of the cells treated with anti-IgM antibody (Fig. 1B). We also immunoblotted the fractions with anti-IgM antibody ( chain-specific) to detect BCR. While there was significant BCR immunoreactivity found in lipid raft fractions in resting cells, a slightly gradual increase in the intensity of BCR bands in fractions 5-9 of the lysates derived from stimulated cells was evident, although the extent of the translocation appeared to be less significant than that for HS1. A previous report had described that translocation of BCR into rafts is more efficient in mature B lymphocyte CH27 cells than in WEHI-231 cells (28). Therefore, we analyzed the association of BCR and HS1 with raft fractions in the lysates from CH27 cells as well. Consistent with the previous report, the translocation of BCR into rafts was readily detected in CH27 cells (Fig.  1C).
Residues Tyr 388 and Tyr 405 Are Required for HS1 Translocation into Lipid Rafts upon BCR Cross-linking-To study the role of HS1 in BCR signaling in more detail, we examined tyrosine phosphorylation of HS1 in WEHI-231 cells upon BCR cross-linking. As shown in Fig. 2, HS1 underwent a rapid tyrosine phosphorylation that reached a plateau within 5 min followed by a rapid decline and eventual disappearance in 2 h. The induction appeared to be specific for BCR cross-linking because no apparent increase in HS1 phosphorylation was observed in the cells treated with anti-IgG antibody (Fig. 2,  right panels). This result is consistent with a previous report by others (17). To determine the tyrosine phosphorylation status of HS1 in rafts, we analyzed by phosphotyrosine immunoblot the HS1 associated with raft and high density fractions of the cells with and without anti-IgM antibody treatment. As shown in Fig. 2B, a significant level of phosphorylated HS1 was found in a pooled raft fraction of the stimulated cells.
Next we analyzed whether tyrosine phosphorylation was required for HS1 raft association. As a result, we prepared a GFP-tagged murine HS1 mutant where Tyr 388 and Tyr 405 were replaced by phenylalanine; these residues correspond to Tyr 378 and Tyr 397 in human HS1 that are targets for Syk proteintyrosine kinase (19). The resulting mutant, HS1 Y388F/Y405F -GFP, was introduced into WEHI-231 cells via a retroviral vector. The defect of this mutant in tyrosine phosphorylation was verified by phosphotyrosine immunoblot (Fig. 3A). While HS1-GFP was readily phosphorylated in the cells treated with anti-IgM antibody, HS1 Y388F/Y405F -GFP failed to show any increase in its phosphorylation level in response to BCR cross-linking. To visualize the distribution of HS1-GFP variants in relation to lipid rafts within cells, cells expressing HS1 Y388F/Y405F -GFP and HS1-GFP were incubated with TRITC-CTB for GM1 ganglioside staining and anti-IgM antibody for BCR cross-linking. The treated live cells were then examined by confocal microscopy. As a control, cells without IgM antibody treatment were examined in parallel. In most (ϳ75%, n ϭ 300) non-treated cells, HS1-GFP displayed a cell peripheral and punctate staining (Fig. 3B, a and aЈ), a pattern similar to what has been described for cortactin in adherent cells (29). A punctate staining was also seen with GM1 as detected by TRITC-CTB (Fig.  3B, b and bЈ). Some GM1 puncta were apparently colocalized with or in proximity to HS1-GFP puncta (Fig. 3B, c). After BCR cross-linking, both HS1-GFP and GM1 were evidently translocated into large patches in many cells (ϳ80%, n ϭ 200) (Fig. 3B,  d, e, dЈ, and eЈ) where their colocalization was more evident (Fig. 3B, f). In contrast, HS1 Y388F/Y405F -GFP displayed a quite diffuse pattern in the cytoplasm either in most (ϳ95%, n ϭ 200) non-treated cells (Fig. 3B, g and gЈ) or in the cells treated with anti-IgM antibody (Fig. 3B, j and jЈ). No colocalization of the mutant with GM1 was detected in patches under either condition (Fig. 3B, i and l).
The lipid raft association with HS1-GFP and HS1 Y388F/Y405F -GFP mutant was also examined by gradient centrifugation of Triton X-100 lysates of their overexpressors. As shown in Fig.  3C, both HS1-GFP and HS1 Y388F/Y405F -GFP proteins were mainly found in the high fraction in resting cells. Upon BCR cross-linking, only HS1-GFP but not HS1 Y388F/Y405F -GFP was shifted into the lipid raft fractions. These data indicate that lipid raft association requires functional tyrosine residues Tyr 388 and Tyr 405 .
Residues Tyr 388 and Tyr 405 Are Required for HS1 to Colocalize with BCR in Membrane Patches-Next we examined the colocalization of HS1 with BCR within B cells. As a result, cells expressing HS1-GFP were incubated with TRITC-IgM antibody on ice for 5 min and incubated at 37°C for an additional 5 min. The fluorescent IgM antibody was used because it can cross-link and label BCR simultaneously. The live cells treated with fluorescent IgM antibody were then inspected by confocal microscopy. As shown in Fig. 4, most BCR staining was found in the periphery of cells when incubated on ice, showing little colocalization with HS1-GFP (Fig. 4c). After cells were shifted to 37°C for 5 min, BCR accumulated in distinct membrane patches that were distributed asymmetrically in one side of the cell. In these patches BCR was evidently colocalized with HS1-GFP (Fig. 4, d, dЈ, e, and f). These polarized patches were morphologically reminiscent of the caps that were described for asymmetrical clustering of actin, antigen receptors, and signaling molecules in lymphocytes (30,31). In contrast, HS1 Y388F/Y405F -GFP failed to colocalize with BCR in the cell under the same conditions (Fig. 4, g-jЈ). The failure of the mutant to translocate into patches was not due to a possible dysfunction with BCR because BCR was translocated normally into the patches within the same cell expressing the HS1 mutant (Fig. 4k). These data indicate that tyrosine residues 388 and 405 are indispensable for HS1 to colocalize with BCR and to translocate into patched areas.
HS1 Translocation into Rafts Requires the Function of Syk Tyrosine Kinase-The above data implied a role of tyrosine phosphorylation at Tyr 388 and Tyr 405 residues in HS1 association with lipid rafts. Since these two residues are the targets of Syk (17), we reasoned that Syk might play a role in the translocation of HS1 into lipid rafts as well. As a result, we examined chicken B lymphatic DT40 cells and their derivatives DT40Syk Ϫ/Ϫ (32). Because of the lack of appropriate antibodies for chicken HS1, we analyzed HS1-GFP that was introduced into these chicken cells via a retroviral vector. To ensure that HS1-GFP responds to BCR cross-linking properly in chicken cells, we examined tyrosine phosphorylation of HS1-GFP in the cells upon IgM antibody treatment. As shown in Fig. 5A, BCR cross-linking induced a marked increase in tyrosine phosphorylation of HS1-GFP in DT40 cells as significantly as in WEHI-231 cells. In contrast, no apparent phosphorylation of HS1 was found in either stimulated or non-stimulated DT40Syk Ϫ/Ϫ cells, although there was apparently a doublet recognized by phosphotyrosine antibody in the cells stimulated by BCR crosslinking. The nature of the doublet is currently not known but is also frequently found in B cells (see also Figs. 1 and 2). Next lipid raft fractions were isolated from these DT40 cells and analyzed for the presence of HS1-GFP in the Triton-insoluble fractions after density gradient centrifugation. As a control, raft and high fractions were probed with histone, which showed an exclusive association with high fraction only (Fig. 5B, right  panels). Immunoblotting the fractions with GFP antibodies demonstrated that HS1-GFP was translocated into raft fractions of DT40 cells upon BCR cross-linking in a manner similar to that in WEHI-231 cells (Fig. 5B, left panel). However, the translocation was significantly impaired in either resting or anti-IgM treated DT40Syk Ϫ/Ϫ cells. To confirm the result, we reintroduced a Syk construct, which was tagged by RFP at its FIG. 3. Tyrosine phosphorylation of HS1 is required for its translocation into lipid rafts. A, phosphotyrosine analysis of HS1 variants. WEHI-231 cells expressing HS1-GFP or HS1 Y388F/Y405F -GFP (HS1 F388F405 -GFP) were treated with anti-IgM antibody for 5 min at 37°C and lysed. The GFP fusion proteins were immunoprecipitated with anti-GFP antibody from the resulting lysates. The pellets were blotted with antiphosphotyrosine antibody (top panels) and reblotted with anti-HS1 antibody (bottom panels). B, live cell imaging analysis of translocation of HS1 variants into lipid rafts. WEHI-231 cells expressing HS1-GFP (a-f) or HS1 Y388F/Y405F -GFP (g-l) were incubated with TRITC-CTB for 20 min on ice, shifted to 37°C, and incubated with or without anti-IgM antibody for 5 min. The live cells were then inspected microscopically as described under "Materials and Methods." HS1-GFP (a and d) and HS1 Y388F/Y405F -GFP stainings (g and j) are shown as green images; TRITC-CTB (b, e, h, and k) is shown as red images. Superimposed images are presented in c, f, i, and l. Non-overlapped images are also presented at a low magnification as labeled by corresponding letters with prime. C, analysis of raft association of HS1 mutant deficient in tyrosine phosphorylation. WEHI-231 cells expressing HS1-GFP or HS1 Y388F/Y405F -GFP were stimulated with anti-IgM antibody for 5 min at 37°C and then subjected to Triton X-100 lysis, gradient centrifugation, and immunoblotting with anti-HS1 antibody as described in the legend of Fig. 1. R, raft fractions; H, high density fractions; Ab, antibody.
C terminus (Syk-RFP), into DT40Syk Ϫ/Ϫ cells by a retroviral vector. The ectopic expression of Syk-RFP increased tyrosine phosphorylation of HS1 in response to BCR cross-linking (Fig.  5A) and restored effectively BCR cross-linking-mediated translocation of HS1 into lipid rafts as well (Fig. 5B). These data demonstrated that tyrosine phosphorylation mediated by Syk is essential for HS1 recruitment into lipid rafts.
We also analyzed DT40Lyn Ϫ/Ϫ cells. Interestingly no apparent defect of HS1-GFP translocation into lipid rafts was found in the cells (Fig. 5B). In fact, HS1 even partitioned into lipid rafts in the resting DT40Lyn Ϫ/Ϫ cells. Tyrosine phosphorylation analysis also revealed that HS1 was markedly phosphorylated in both resting and BCR-stimulated DT40Lyn Ϫ/Ϫ cells (Fig. 5A), implying that other types of tyrosine kinases may be activated in DT40Lyn Ϫ/Ϫ cells. Consistent with this view, BCR cross-linking induced only a slight increase in HS1 phosphorylation in DT40Lyn Ϫ/Ϫ cells (Fig. 5A). To further verify the role of Src family kinases in the translocation of HS1 into rafts, we analyzed a mutant HS1 Y222F -GFP in which the residue Tyr 222 targeted by Src kinases was mutated to phenylalanine. As shown in Fig. 5C, the mutant was still able to be efficiently translocated into lipid rafts upon BCR cross-linking, indicating that Src family-mediated phosphorylation is not required for HS1 translocation into lipid rafts. DISCUSSION In this study we provide evidence for the first time showing that HS1, an Arp2/3 complex activator for actin assembly, is recruited to lipid rafts along with WASP and Arp2/3 complex upon BCR cross-linking, indicative of a mechanism for BCR to signal the dynamics of the actin cytoskeleton in lipid rafts. Association of HS1 with lipid rafts is indicated by both biochemical and microscopic studies. Density ultracentrifugation of cell lysates of B lymphocytes demonstrated that significant amounts of HS1 proteins were partitioned to light Triton fractions upon BCR cross-linking by anti-IgM antibody, the property that has been ascribed to many raft-associated proteins (2,4). Microscopic analysis of live cells further revealed that HS1-GFP proteins are accumulated into an asymmetric area that is filled with patched membrane structures. Within the patches, HS1 proteins are apparently colocalized with raft-enriched GM1 ganglioside and BCR. The morphology of these patches appears to be similar to previously described raft-associated patches found with other types of B cell lines and splenic B cells (33). Because HS1 binds to and colocalizes with Arp2/3 complex and cortical actin, we suggest that the asymmetric patches enriched with HS1 represent what has been previously described as cap, a membrane structure that is formed upon FIG. 4. Tyrosine phosphorylation is required for HS1 to colocalize with BCR. WEHI-231 cells expressing HS1-GFP (a-f) or HS1 Y388F/Y405F -GFP (g-l) were incubated with TRITC-conjugated F(abЈ) 2 fragment anti-mouse IgM (20 g/ml) on ice for 5 min (a-c and g-i) or at 37°C for 5 min (d-f and j-l). The treated live cells were inspected under confocal microscope. Superimposed images are presented in c, f, i, and l. The images labeled with prime letters represent the corresponding images at low magnification. receptor stimulation including BCR cross-linking and that accumulates cortical actin along with stimulated receptors in B lymphocytes (9, 31, 34 -36).
Consistent with the role of HS1 in actin assembly, translocation of HS1 into lipid rafts is accompanied by WASP and Arp2/3 complex, the molecules that are also known to play important roles in the actin reorganization. Previous studies have demonstrated that WASP, a potent activator of Arp2/3 complex, is recruited to lipid rafts in activated T cells and may be implicated in the function of immunosynapses (37). Similarly involvement of Arp2/3 complex in lipid rafts has been implied in raft-enriched vesicle trafficking based on a study with WASP (38). However, our present study shows for the first time a direct association of Arp2/3 complex with lipid rafts. Thus, our data supports the view that lipid rafts provide a physical platform for specific signaling pathways (4) or signalosome wherein HS1 is a component and consolidates with other molecules to modulate the actin cytoskeleton upon receptor stimulation.
Although HS1 is co-stained with lipid rafts and BCR in activated B cells, the mechanism for the recruitment of HS1 to lipid rafts is currently unknown because direct association of HS1 with BCR was not detected by immunoprecipitation (data not shown). Our study with an HS1 variant with mutations at tyrosine residues 388 and 405, which are targeted by Syk, suggests that Syk-mediated tyrosine phosphorylation is impor-tant for HS1 translocation into lipid rafts because the mutant fails to partition into either lipid rafts or BCR-enriched patches (Fig. 3). The role of Syk in HS1 association with rafts is further strengthened by the fact that HS1 was neither translocated into rafts nor tyrosine phosphorylated in the cells lacking Syk. Moreover association of Syk itself in the rafts of B cells upon BCR cross-linking has also been recently reported (39). Despite these facts, a strong direct association between Syk and HS1 in response to antigen stimulation remains to be established. It is also possible that Syk-mediated phosphorylation triggers a subsequent event that may promote the interaction between HS1 and other cellular factor(s) that is either directly or indirectly associated with rafts. In this regard, Lyn, which is constitutively associated with rafts (3), appears to be a good candidate. Lyn and its related kinases Lck and Fgr bind to HS1 either through their SH2 domains to phosphorylated residues Tyr 378 /Tyr 397 (Tyr 388 /Tyr 405 in mouse) or their SH3 domains to the proline rich domain of HS1. While binding of Lyn to phosphorylated HS1 is necessary for the phosphorylation at Tyr 222 residue, the interaction is transient (20,21). In contrast, the interaction between the Lyn SH3 domain and HS1 proline rich domain appears to be stable and can be readily detected by co-immunoprecipitation or pull-down assay. 2 However, there is 2 J.-J. Hao and X. Zhan, unpublished data.

FIG. 5. Translocalization of HS1 into lipid rafts requires Syk kinase.
A, analysis of tyrosine phosphorylation of HS1-GFP in DT40 cells. DT40 and its derivative cells expressing HS1-GFP were treated with or without anti-IgM antibody, lysed, and subjected to immunoprecipitation with anti-GFP antibody. The pellets were dissolved by SDS-PAGE, transferred to a membrane, immunoblotted with anti-phosphotyrosine antibody (4G10), and reblotted with GFP antibody. Note that tyrosine phosphorylation of HS1 was dramatically increased in DT40Syk Ϫ/Ϫ Syk-RFP cells in response to BCR cross-linking. B, DT40 cells and derivatives expressing HS1-GFP were treated with (ϩ) or without (Ϫ) anti-chicken IgM antibody for 15 min at 37°C, lysed in 1% Triton X-100-containing buffer, and subjected to sucrose gradient ultracentrifugation. Aliquots of equal volume of each fraction were separated by 10% SDS-PAGE, transferred to a polyvinylidene difluoride membrane, and blotted with anti-GFP antibody (left panel). Note that the efficiency of translocation of HS1 into lipid rafts in DT40Syk Ϫ/Ϫ Syk-RFP cells was evidently higher than that of the rest of cells. The fractions from DT40Syk Ϫ/Ϫ Syk-RFP were blotted by histone H1 and CD71 antibodies as well (right bottom panel). For other cells, only raft (R) and high (H) fractions were probed for histone H1 and CD71 (right top panel). C, WEHI-231 cells expressing mutant HS1 Y222F -GFP (HS1 F222 -GFP) were stimulated by anti-IgM antibody and further subjected to analysis for the presence of HS1 in raft fractions as described in the legend of Fig. 1B. pY, phosphotyrosine; T, total cell lysates. no clear evidence yet whether or not the interaction between Lyn and HS1 through SH3/proline rich domain can be regulated by tyrosine phosphorylation (21). Our data presented here indicate that Lyn itself is not necessary for HS1 to translocate into rafts because HS1 goes to rafts normally in Lyndeficient cells. While it is possible that other Lyn-related kinases may substitute for the function of Lyn in these cells, the HS1 mutant at Tyr 222 , which is targeted by Lyn and its related kinases such as Fgr, Fyn, and Lck, is recruited properly to rafts of the cells stimulated by anti-IgM antibody, indicating that phosphorylation at Tyr 222 and other Lyn-related kinases may not be indispensable for the recruitment of HS1 to rafts. Interestingly there is a significant number of HS1 proteins that are tyrosine phosphorylated in the cells deficient in Lyn even without stimulation. This may occur due to the function of Syk, which could be activated in the cell line deficient in Lyn, thereby resulting in apparent phosphorylation at Tyr 388 /Tyr 405 residues. While this is consistent with the role of the Sykmediated phosphorylation in the translocation of HS1 to rafts, additional studies are required to define the raft-associated molecule(s) that brings HS1 directly to the rafts in a tyrosine phosphorylation-dependent manner.