Syk tyrosine kinase is required for immunoreceptor tyrosine activation motif-dependent actin assembly.

Clustering of several multisubunit receptors on hematopoetic cells results in a signaling cascade involving the phosphorylation of immunoreceptor tyrosine activation motifs, or “ITAMs,” and actin polymerization. Recent experiments indicate that direct clustering of the ITAM-binding protein, p72syk (Syk), is capable of transmitting a phagocytic signal in COS cells (Greenberg, S., Chang, P., Wang, D., Xavier, R., and Seed, B. (1996) Proc. Natl. Acad. Sci. U. S. A. 93, 1103-1107). However, the possibility of redundant signaling pathways makes it difficult to test the requirement for Syk in ITAM-dependent actin polymerization in hematopoetic cells. We developed a model system to study ITAM-dependent actin assembly. DT40 lymphocytes were transfected with fusion proteins encoding the transmembrane and cytosolic domains of the ITAM-containing γ subunit of Fc receptors. Clustering the γ-containing fusion proteins with IgG-coated erythrocytes triggered submembranous actin assembly. This response depended on an intact ITAM, was absent in cell lines that had been engineered to lack Syk, and was augmented in cell lines that stably overexpressed Syk. These experiments demonstrate an absolute requirement for Syk tyrosine kinase in ITAM-dependent actin assembly in transfected lymphocytes.

Clustering of several multisubunit receptors on hematopoetic cells results in a signaling cascade involving the phosphorylation of immunoreceptor tyrosine activation motifs, or "ITAMs," and actin polymerization. Recent experiments indicate that direct clustering of the ITAM-binding protein, p72 syk (Syk), is capable of transmitting a phagocytic signal in COS cells (Greenberg, S., Chang, P., Wang, D., Xavier, R., and Seed, B. (1996) Proc. Natl. Acad. Sci. U. S. A. 93, 1103-1107). However, the possibility of redundant signaling pathways makes it difficult to test the requirement for Syk in ITAM-dependent actin polymerization in hematopoetic cells. We developed a model system to study ITAM-dependent actin assembly. DT40 lymphocytes were transfected with fusion proteins encoding the transmembrane and cytosolic domains of the ITAM-containing ␥ subunit of Fc receptors. Clustering the ␥-containing fusion proteins with IgG-coated erythrocytes triggered submembranous actin assembly. This response depended on an intact ITAM, was absent in cell lines that had been engineered to lack Syk, and was augmented in cell lines that stably overexpressed Syk. These experiments demonstrate an absolute requirement for Syk tyrosine kinase in ITAM-dependent actin assembly in transfected lymphocytes.
Actin assembly is a nearly ubiquitous response to the engagement of a diverse array of cell surface receptors that signal chemotaxis, phagocytosis, cell division, or cell adhesion. In hematopoetic cells, ligation of members of the immunoglobulin gene superfamily is accompanied by actin polymerization and several functionally important actin-dependent events. For example, clustering receptors for the Fc portion of IgG (Fc ␥ receptors) 1 by IgG-opsonized targets results in phagocytosis in macrophages (1) and neutrophils (2,3), while engagement of antigen receptors by target cells results in cytochalasin-sensitive interferon-␥ secretion by T-lymphocytes (4). Since many biochemical events accompany ligation of these multisubunit receptors, it is important to identify those events among the many that are required for triggering actin assembly. A recent clue was provided by studies demonstrating a requirement for tyrosine kinases in lymphocyte effector functions (5) and phagocytosis (6,7). However, activation of multiple tyrosine kinases accompanies engagement of Fc ␥ receptors (8 -12) and antigen receptors (13), making it difficult to implicate specific tyrosine kinases in triggering actin assembly. Models for signaling by both families of receptors have been proposed which implicate ZAP-70 and Syk tyrosine kinases in several downstream signaling events (13)(14)(15)(16)(17), and both kinases are capable of triggering actin assembly when autonomously clustered in COS cell transfectants (18). However, it is not clear whether Syk is required for ITAM-mediated actin assembly in hematopoetic cells. To test a requirement for Syk in cytoskeletal alterations mediated by specific ITAM-containing subunits, such as the ␥ subunit of Fc receptors, we developed an assay of ITAM-dependent actin assembly using the DT40 lymphocyte cell line. The advantages of this cell line are that it lacks endogenous ␥ subunits, and it undergoes a high rate of homologous recombination. Thus, by gene targeting and cDNA transfection, it is possible to isolate clones of DT40 cells that lack expression of a given gene product (19), and that express specific ITAM-containing subunits. Plasmid constructs encoding the transmembrane and cytosolic domains of the ␥ subunit were expressed in wild-type lymphocyte cell lines, or in cells lines that were engineered to either lack, or stably overexpress, Syk. Using these cell lines, we tested whether Syk was required for ␥ subunit-mediated actin assembly.
Plasmid Construction and Isolation of Transfected Cell Lines-16:␥, a fusion protein consisting of the extracellular domain of human Fc ␥ R * This work was supported in part by Grants HL02641 and HL54164 from the National Institutes of Health, a Research Grant from the American Cancer Society, and a grant-in-aid from the American Heart Association, New York City affiliate. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
¶ To whom correspondence should be addressed: Columbia University, Dept. of Medicine, 630 West 168th St., New York, NY 10032. Tel.: 212-305-1586; Fax: 212-305-1146; E-mail: greenberg@cuccfa.ccc. columbia.edu. 1 The abbreviations used are: Fc ␥ receptor, receptor for the Fc portion of IgG; ITAM, immunoreceptor tyrosine activation motif; mAb, monoclonal antibody; RBCs, red blood cells; IgG-RBCs, sheep erythrocytes opsonized with anti-sheep erythrocyte IgG; Syk ϩϩ , DT40 cell lines that overexpress Syk; Syk Ϫ , DT40 cell lines that lack expression of Syk; 16:␥, fusion proteins containing CD16 ectodomains and ␥ subunit transmembrane and cytosolic domains; 16:␥ Y 3 F , a 16:␥-based construct bearing a Tyr 3 Phe mutation in tyrosine 76; WT, wild-type DT40 cells. IIIA (CD16) and the transmembrane and cytosolic domains of the murine ␥ subunit of Fc receptors, was constructed by polymerase chain reaction (21) and was inserted into the EcoRI site of pApuro (19). 16:␥ Y 3 F , a construct bearing a Tyr 3 Phe mutation in tyrosine 76, was created by polymerase chain reaction. Resultant constructs were confirmed by DNA sequencing. DT40 lymphocytes expressing either 16:␥ or 16:␥ Y 3 F were isolated following transfection by electroporation (19) and selected in the presence of 0.25 g/ml puromycin. 16:␥ was similarly expressed in Syk Ϫ DT40 cells, previously generated by gene targeting (19). To generate 16:␥-expressing cell lines that also overexpressed Syk (Syk ϩϩ /16:␥), one clone of WT/16:␥ (clone G1) was cotranfected with pSyk (19) and pRc-CMV, and individual clones were isolated following selection with 2 mg/ml G418. Between 3 and 30 individual clones were isolated from each transfection, and individual clones were analyzed for surface expression of the CD16 epitope by flow cytometry and by rosetting with IgG-RBCs. Levels of Syk expression were confirmed by immunoblotting with anti-Syk IgG and were normalized to actin content by immunoblotting with mAb C4 against actin.
Immunoblotting, Immunoprecipitation, and in Vitro Kinase Assays-5 ϫ 10 6 cells were incubated in the presence or absence of 20 g/ml mAb 3G8 for 30 min at 4°C and further incubated for varying intervals at 37°C with 40 g/ml rabbit anti-mouse IgG. Following addition of icecold lysis buffer (1% Triton X-100, 100 mM NaCl, 2 mM EDTA, 1 mM sodium orthovanadate, 10 g/ml leupeptin, 10 g/ml aprotinin, 0.5 mM phenylmethylsulfonyl fluoride, 20 mM Tris-HCl, pH 7.2), samples were processed for either immunoblotting with anti-phosphotyrosine (20) or immunoprecipitation with anti-Syk IgG pre-adsorbed to protein A-agarose. For anti-Syk immunoblotting, 2 ϫ 10 6 cells were lysed as above and immunoblotted with rabbit anti-Syk IgG. Blots were visualized using enhanced chemiluminescence. Syk protein expression was quantitated by densitometry of anti-Syk immunoblots using NIH-Image. For in vitro kinase assays, anti-Syk immunoprecipitates derived from 10 7 cells were washed and incubated with a kinase buffer containing 5 Ci of [␥-32 P]ATP, 5 mM MnCl 2 , 5 mM MgCl 2 , and 25 mM Tris-HCl, pH 7.2. After incubation at 25°C for 5 min, samples were diluted in a buffer containing 10 mM EDTA to inhibit further kinase activity, washed, and analyzed by SDS-polyacrylamide gel electrophoresis and autoradiography.
F-actin Assays and Fluorescence Microscopy-Total cellular F-actin was quantitated as described previously (1) with the following modifications: 2 ϫ 10 5 lymphocytes were allowed to adhere for 30 min at 37°C in 96-well tissue culture plates and were incubated in the presence or absence of 4 ϫ 10 6 IgG-RBCs for 10 min at 4°C, washed to remove nonadherent RBCs, and incubated further for varying times at 37°C. Cells were fixed in 3.7% formaldehyde and stained with 0.33 M rhodamine-phalloidin. Rhodamine fluorescence (excitation wavelength: 530 nm, emission wavelength: 590 nm) was measured using a fluorescence plate reader (CytoFluor II; Millipore). To normalize for cell number, nucleii were stained with 5 M YO-PRO, fluorescence was determined (excitation wavelength: 485 nm, emission wavelength: 530 nm), and F-actin content per unit cell was calculated as the ratio of rhodamine to YO-PRO fluorescence. Previous experiments established that fluorescence of rhodamine at excitation wavelength: 485 nm, emission wavelength: 530 nm was negligible and that the F-actin content of IgG-RBCs was less than 0.5% of the F-actin content of DT40 cells.
Measurement of localized cortical F-actin content was modified from the methods of Theriot et al. (22) and Segall et al. (23). 6.5 ϫ 10 5 lymphocytes were incubated on poly-L-lysine-coated glass coverslips for 30 min at 37°C and further incubated with 7.5 ϫ 10 6 IgG-RBCs for 10 min at 4°C to allow particle binding. Following washing and incubation at 37°C for varying intervals, cells were fixed and stained with fluorescein anti-rabbit IgG (to detect bound particles) and with saturating concentrations of rhodamine-phalloidin (to quantitate F-actin) and examined with a Bio-Rad MRC 600 confocal laser scanning system equipped with a Krypton/Argon laser. 1-m sections were imaged and stored on optical discs and images were analyzed using NIH-Image software. Integrated densities of rhodamine staining were measured using a 4-m diameter circle placed on regions of lymphocyte cortex beneath attached IgG-RBCs. Since cortical F-actin staining was not necessarily uniform, staining densities were determined for three separate regions of cortex without associated RBCs in the same cells and served as controls. Data were expressed as the ratios of the integrated density of rhodamine staining in cortical areas beneath attached IgG-RBCs to the average of the integrated densities of rhodamine staining of control regions. At least five fields were examined in each experiment. For a given experiment, 16 -26 individual cortical F-actin ratios were determined and averaged. To determine the F-actin density per pixel, weighted densities were calculated as the number of pixels of a given density multiplied by that pixel intensity. Data represents measurements taken from 40 -57 cortical regions from five microscopic fields of the indicated cell lines.

RESULTS
An Intact ␥ Subunit ITAM Is Required for Signaling Enhanced Protein Tyrosine Phosphorylation-All cell lines transfected with 16:␥ or 16:␥ Y 3 F expressed the CD16 epitope at the cells' surfaces as determined by flow cytometry. We characterized two individual clones, expressing 16:␥ or 16:␥ Y 3 F , in more detail (Fig. 1). WT/16:␥ and WT/16:␥ Y 3 F , but not the parental cell line, WT, bound IgG-RBCs (not shown). To determine whether 16:␥ was capable of triggering enhanced protein tyrosine phosphorylation, we incubated transfectants or controls with mAb 3G8 followed by anti-mouse IgG. Enhanced tyrosine phosphorylation of multiple proteins was apparent within 1 min of clustering surface-bound mAb 3G8 in WT/16:␥, but not untransfected controls (WT) or WT/16:␥ Y 3 F cells (Fig. 2). These data indicate a requirement for tyrosine 76 of the ␥ subunit in transmitting a signal for enhanced protein tyrosine phosphorylation and are consistent with findings of others that document a similar requirement for an intact ␥ subunit ITAM in mediating enhanced protein tyrosine phosphorylation (24,25).
WT/16:␥, but Not WT/16:␥ Y 3 F , Mediates the Submembranous Accumulation of F-actin-We tested the ability of 16:␥ to mediate cytoskeletal alterations when expressed in DT40 cells. Incubation of WT/16:␥ cells with IgG-RBCs led to the submembranous accumulation of F-actin (Fig. 3). Most, but not all, cortical regions underlying attached particles demonstrated a broad-based accumulation of F-actin that was first apparent by 1 min (not shown), appeared maximal at 4 min (Fig. 3), and was no longer detectable after 10 min of incubation (not shown). Using phase-contrast optics, rudimentary pseudopods could be seen beneath many bound IgG-RBCs (not shown). The focal accumulation of F-actin in response to the IgG-bearing ligand was blocked by the presence of 1 M cytochalasin D (not shown), indicating a requirement for barbed-end filament growth. To test the requirement for an intact ITAM in mediating the focal appearance of F-actin, we performed parallel experiments using WT/16:␥ Y 3 F cells. Despite obvious IgG-RBC binding, there were no discernible focal accumulations of F-actin present beneath the attached particles (Fig. 3). Although surface expression of 16:␥ was somewhat higher than that of WT/16:␥ Y 3 F (Fig. 1), there was considerable overlap in the extent of surface expression of the two constructs, while there were essentially no detectable focal accumulations of F-actin in cells expressing WT/16:␥ Y 3 F . Therefore, differences in surface expression could not account for the lack of focal accumulations of F-actin in cells expressing WT/16:␥ Y 3 F . These results indicate that an intact ITAM is required for the focal accumulation of F-actin mediated by clustered ␥ subunits.
Syk Is Required for 16:␥-mediated Actin Assembly-Based on earlier work (1), we developed an assay for F-actin using rhodamine-phalloidin and YO-PRO, a nuclear stain, and quantitated average F-actin content per cell using a fluorescence plate reader. This assay was linear over a wide range of cell concentrations using the same concentration of rhodaminephalloidin, indicating that all F-actin binding sites were saturated with phalloidin (Fig. 6A). Incubation of WT/16:␥ cells by IgG-RBCs led to a small time-dependent increase in F-actin, which peaked at 4 min (not shown). These changes were more easily observed in Syk ϩϩ /16:␥ cells, which produced a greater magnitude of increase in F-actin upon addition of IgG-RBCs (Fig. 6B). The level of Syk expression influenced the magnitude of accumulation of F-actin in response to the IgG-containing ligand. Absence of Syk did not support significant increases in F-actin, while its overexpression led to an augmented ITAMinduced response (Fig. 6C). Cytochalasin D blocked the ITAM- FIG. 5. Syk-dependent protein tyrosine phosphorylation in response to clustered ITAMs. A, anti-Syk immunoblots of lysates derived from the indicated transfected cell lines. B, anti-phosphotyrosine immunoblots of lysates derived from the indicated lymphocyte cell lines. Cells were incubated in the presence or absence of mAb 3G8 at 4°C for 30 min followed by further incubation with rabbit anti-mouse IgG at 37°C for the indicated times prior to detergent lysis. C, phosphotyrosine content of Syk immunoprecipitated from the indicated cell lines. Cells were incubated with or without mAb 3G8 as above, lysed either immediately or following addition of anti-mouse IgG for 3 min at 37°C, and subjected to immunoprecipitation with anti-Syk IgG followed by blotting with anti-phosphotyrosine. D, in vitro kinase reactions of anti-Syk immunoprecipitates. Cells were stimulated as in C and subjected to immunoprecipitation and kinase reactions as described under "Experimental Procedures." Molecular weight markers appear at left. mediated increase in total cellular F-actin in WT/16:␥ (not shown) and in Syk ϩϩ /16:␥ cells (Fig. 6C), indicating a requirement for barbed end actin filament growth. While the level of Syk expression influenced the basal quantity of phosphotyrosine-containing proteins (Fig. 4), it did not influence the basal F-actin content of the cells (not shown). However, Syk was required for focal F-actin staining beneath the test particles, while overexpression of Syk led to an augmented focal accumulation of F-actin (Fig. 7), which was blocked by cytochalasin D (not shown).
Regional changes in cortical F-actin have been documented in other cells using several stimulii (23,26,27). We therefore measured cortical F-actin content in transfectants challenged with IgG-RBCs. We compared the F-actin content in lymphocyte cytoplasm subjacent to attached IgG-RBCs with the Factin content in other cortical areas of the same cell. Depending on the particular clone examined, cortical F-actin beneath attached IgG-RBCs was enriched 4 -8-fold (mean 6.2 Ϯ 0.9) in WT/16:␥ cells, 10 -14-fold (mean 11.7 Ϯ 1.1) in Syk ϩϩ /16:␥ cells, and 1-1.2-fold (mean 1.1 Ϯ 0.1) in Syk Ϫ /16:␥ cells (Fig.  8A). Interestingly, the density of F-actin staining per pixel was greatest in Syk ϩϩ /16:␥ cells, followed by WT/16:␥ and Syk Ϫ / 16:␥ cells (Fig. 8B), suggesting that maximal Syk expression favors the formation of a dense meshwork of actin filaments upon ITAM clustering. DISCUSSION Given the multitude of tyrosine kinases present in eukaryotic cells, and the possibility of redundant signaling pathways, it is important to establish functional roles for individual kinases. Previous studies, including those showing that Syk is capable of triggering (18) or enhancing (28) phagocytosis in COS cells, provided evidence that this kinase can trigger phagocytosis in a cell that is not normally phagocytic. This study demonstrates that Syk tyrosine kinase is required for ITAMmediated actin assembly and Fc ␥ receptor-mediated phagocytosis and that no redundant pathways exist in these hematopoetic cells to effect these responses.
We did not address the mechanism of Syk-mediated actin assembly in this study, nor did we address the requirement for other kinases in ITAM-mediated signaling pathways. We suspect, however, that one or more members of the Src family is required for these events, since Lyn-negative DT40 cell lines which were transfected with 16:␥ were also incapable of supporting 16:␥-mediated actin assembly (not shown). Furthermore, fusion proteins bearing kinase-deficient Syk constructs do not trigger actin assembly in transfected COS cells (18), indicating that an intact Syk kinase domain is required for mediating actin assembly. It is likely that one or more Syk tyrosine kinase substrates enhance the formation of an as yet unidentified actin-nucleating activity.
While ␥ subunit-expressing macrophages are capable of Fc ␥ receptor-mediated phagocytosis, 16:␥-expressing DT40 cells did not support phagocytosis, as assessed by resistance to hypotonic lysis of cell-associated IgG-RBCs. DT40 cells transfected with human Fc ␥ R IIA, another ITAM-containing transmembrane protein, were equally potent in inducing actin assembly, but incapable of mediating phagocytosis (not shown). Why are DT40 lymphocyte transfectants capable of ITAMdirected actin assembly, yet incapable of ITAM-mediated phagocytosis? It could be argued that the levels of surface expression of 16:␥ fusion proteins were insufficient to support phagocytosis. We were unable to isolate clones of DT40 cells that expressed higher levels of 16:␥ to test this hypothesis directly. We think this explanation is unlikely, however, since macrophages are capable of some degree of particle ingestion, even when surface Fc ␥ receptor expression is decreased by nearly 90% when the cells are plated on adherent immune complexes (29). Since mammalian B-lymphocytes express Fc ␥ RIIB (30), a receptor that transmits an inhibitory signal to antigen receptors (31)(32)(33), we also considered the possibility that DT40 cells express the avian homolog of Fc ␥ RIIB. This receptor might be expected to dampen the phagocytic response by its co-ligation of IgG-RBCs. However, this explanation is unlikely, since untransfected DT40 cells were incapable of binding IgG-RBCs. Furthermore, addition to WT/16:␥ cells of E-3G8, erythrocytes that bear F(ab)Ј 2 fragments of mAb 3G8 rather than intact IgG (34) and cannot ligate endogenous Fc ␥ receptors, also resulted in binding, but not ingestion. Finally, overexpression of Syk in 16:␥-expressing cells led to an augmented local accumulation of F-actin in response to IgG-RBCs, but did not confer to the transfected lymphocytes the ability to mediate particle ingestion (not shown), suggesting that the level of Syk tyrosine expression was not the limiting factor in promoting phagocytosis. Although we did not perform a direct comparison between the quantitative increases in F-actin due to the interaction of IgG-containing ligands with DT40 cells and macrophages, it is interesting to note that regional accumulations in F-actin beneath 16:␥-expressing lymphocytes appeared even more prominent than similar regions in Fc ␥ R-expressing macrophages (cf. Figs. 3 and 7 of the current study with Fig. 2 in Ref. 1).
While actin assembly is clearly required for pseudopod extension, these data suggest that actin assembly alone is insufficient for mediating phagocytosis. Other events, such as recruitment of new membrane components to the growing pseudopod, may be required for productive pseudopod extension. Alternatively, the newly formed submembranous actin network may need to be coupled to mechanotransducing proteins in order to achieve a significant degree of pseudopod extension, as suggested in a review of neuronal growth cone motility (35). Candidates for this activity include the MARCKS family of proteins (36) and members of the myosin I superfamily (37). Recent experiments have implicated one or both families of proteins in pseudopod extension (38) and phagocytosis (38,39). We are currently testing these possibilities.