INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Stimulation of Fc receptors and T-cell and B-cell antigen receptors induces a rapid increase in tyrosine phosphorylation of cellular proteins. This biochemical signaling plays crucial roles in inflammatory functions, including phagocytosis, cytokine synthesis, and inflammatory mediator release (4, 6, 17, 59, 74, 84). The majority of Fc receptors, together with T-cell and B-cell antigen receptors, have hetero-oligomeric structures: they are composed of ligand-binding subunits and associating signal transduction subunits (17, 35, 59, 74). The high-affinity immunoglobulin E (IgE) receptor (FcRI) has a tetrameric structure composed of an IgE binding  subunit, a  subunit, and a disulfide-bonded  dimer (8, 45, 58). Aggregation of IgE is converted to protein tyrosine phosphorylation (71) by the action of the  and  subunits (4, 6, 17, 36, 57). These signaling modules have not been shown to possess catalytic activity but instead possess tyrosine-based cell activation motifs (ITAM [immunoreceptor tyrosine-based activation motif]) (30, 62, 84). Upon receptor clustering, ITAM tyrosine is rapidly phosphorylated and creates sites for the assembly of SH2 domain-containing proteins, including Syk protein tyrosine kinase (5, 75). Association of Syk with tyrosine phosphorylated  and subsequent phosphorylation of Syk on activation loop tyrosine further trigger the downstream signaling cascade leading to cell activation (29, 37, 68, 83).

The initial activation step of ITAM tyrosine phosphorylation is ascribed to the action of Src family protein tyrosine kinases. This concept is in part based on the observations that several of Src family members physically associate with Fc receptors under resting conditions and that their kinase activities are rapidly increased after receptor engagement (18, 64, 79, 81, 87). To obtain more confirmative evidence, targeted disruption of single, or multiple Src family genes were conducted (15, 40, 41, 51). Crowley et al. showed that Fc receptor-mediated phagocytosis is delayed but well preserved in macrophages derived from Lyn^/ Hck^/ Fgr^/ mice (15). One of our laboratories demonstrated that FcRI-induced calcium mobilization and degranulation is preserved in Lyn^/ murine bone marrow-derived mast cells, albeit tyrosine phosphorylation of Syk and Bruton's tyrosine kinase were reduced (51). These findings have provided the important information that Src family kinases possess complementary roles in Fc receptor functions (40), but the functional redundancy itself generates a difficulty in ascertaining the requirement or the specificity of Src family kinases. In addition, wide distribution of FcRI in monocytes, eosinophils and Langerhans cells, besides basophils and mast cells (23, 43, 66, 82), raised the possibility that FcRI may utilize different set of Src family kinases depending on cell species.

As an alternative approach, C-terminal Src kinase (Csk) (28, 49, 50, 53) has been used as a negative regulator of Src family kinases (11, 12, 27, 73). In hematopoietic cells, Src family kinases are assumed to be in an equilibrium between C-terminal tyrosine-phosphorylated (inactive) and dephosphorylated (partially active) forms (27, 74). The balance is regulated by the opposite actions of Csk and CD45 tyrosine phosphatase (13, 48, 74). When C-terminal tyrosine is phosphorylated by Csk, catalytic activity of Src family kinases are suppressed by an intramolecular interaction between the C-terminal tyrosine-based sequence and the SH2 domain (13, 14). Recently, another intramolecular interaction that negatively regulates kinase activity was found between the SH3 domain and the N-terminal linker segment of the catalytic domain (69, 86). By variously modulating the Csk activity in RBL2H3 cells through the overexpression of Csk, a gain-of-function Csk mutant possessing N-myristoylation signal (mCsk) (13) and a kinase-dead mCsk [mCsk()] (27), we previously showed that Src family kinases are upstream regulator of FcRI-mediated biochemical signaling (27).

The current study was undertaken to further evaluate the roles of individual Src family kinase and its submolecular structures in early FcRI signaling. To this end, we utilized mCsk-overexpressing RBL2H3 cells in which basal Lyn activity and FcRI signaling were suppressed. C-terminal tyrosine-deleted Src (termed a-Src) family kinases were reconstituted in the cells. We show herein that (i) FcRI signaling is transduced by selective Src family kinases, (ii) palmitoylatable Cys in SH4 domain is required for initial interaction with FcRI, and (iii) SH2 domain is required for signal progression.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

cDNA construction. Murine cDNAs for Csk, a membrane-anchored Csk mutant (mCsk) possessing N-terminal myristoylation signal of rat c-Src, and a kinase-dead mCsk [mCsk()] were described previously (27, 49, 73). c-Myc epitope-tagged rat c-Src and human p56^lyn lacking C-terminal tyrosine (termed a-Src and a-Lyn, respectively), whose corresponding C-terminal amino acids (EPQYQPGENL for c-Src and EGQYQQQP for p56^lyn, with the negative regulatory tyrosine indicated in boldface) were replaced with 9E10 c-myc epitope sequence (TSVDEQKLISEEDLN), were described previously (25, 73). C-terminal amino acids of human p59^fyn (EPQYQPGENL) were also replaced with the c-myc epitope tag through the same procedures, thus generating a-Fyn. To create deletion mutants of a-Lyn lacking the SH3 (amino acids 68 to 117 of human p56^lyn) or SH2 (amino acids 130 to 222 of human p56^lyn) domains (termed SH3 a-Lyn and SH2 a-Lyn, respectively), a pair of AatII sites and EcoRI sites flanking SH3 and SH2 domains of a-Lyn, respectively, were introduced by PCR-based techniques using QuickChange site-directed mutagenesis kit (Stratagene, La Jolla, Calif.). The mutagenesis primers used for SH3 deletion were 5'-GGA-ACA-AGG-AGA-CGT-CGT-GGT-AGC-CTT-GTA-C-3' and 5'-CAT-CCC-CAG-CAA-CGA-CGT-CGC-CAA-ACT-CAA-C-3' encompass ing nucleotides 186 to 216 and 336 to 366 of human p56^lyn coding region, respectively. Those for SH2 deletion were 5'-CAC-CTT-AGA-AAC-AGA-AGA-ATT-CTT-TTT-CAA-GGA-TAT-AAC-C-3' and 5'-GAT-GGC-TTG- TGC-AGA-GAA-TTC-GAG-AAG-GCT-TG-3' encompassing nucleotides 366 to 405 and 646 to 677 of human p56^lyn coding region, respectively. The resultant cDNAs were digested with AatII or EcoRI and then self-ligated. Ser3 to Cys mutation in a-Src (a-Src S3C) was introduced through the same procedures, by using a mismatch primer (5'-CCA-GGA-CCA-TGG-GCT-GCA-ACA-AGA-GCA-AGC-CC-3'). a-Fyn, SH3 a-Lyn, SH2 a-Lyn, and a-Src S3C cDNAs without misincorporation of nucleotides were selected and subcloned into an expression vector, pCAGGS (46, 52).

Cells. RBL2H3 cells and its sublines were cultured as a monolayer in Dulbecco modified Eagle medium (DMEM; Nissui, Tokyo, Japan) supplemented with 10% fetal calf serum (FCS) (Equitech-Bio, Ingram, Tex.). RBL2H3 clones stably expressing mCsk, or mCsk() cDNA subcloned into neomycin-resistant pCXN2 vector (46, 52) were described previously (27, 73). As a vector control, RBL2H3 clones stably expressing an unrelated cDNA (human PAF receptor [26]) cloned into pCXN2 were also created. RBL2H3 cells overexpressing mCsk were next transfected with a puromycin-resistant vector alone or the vector in combination with a-Lyn, a-Fyn, a-Src, SH3 a-Lyn, SH2 a-Lyn, or a-Src S3C cDNA subcloned into pCAGGS by electroporation, as described elsewhere (73). Clones resistant to puromycin were selected and tested for the expression of the mutated Src family kinases and mCsk by Western blotting with anti-c-myc and anti-Csk polyclonal antibodies. Cells expressing both the molecules were further subcloned by limiting dilution, and independent clones were established for each construct.

Cell stimulation and cell lysis. Cells at subconfluence in 10-cm dishes were detached by brief trypsinization as described earlier (73) and harvested into DMEM supplemented with 10% FCS. Cells were washed once with DMEM buffered with 10 mM HEPES-NaOH (pH 7.4) and supplemented with 0.1% bovine serum albumin (BSA) (HEPES-DMEM), the cell concentration was adjusted to 10⁷/ml, and the cells were sensitized with 2.5 µg of anti-trinitrophenyl (anti-TNP) mouse monoclonal IgE (Sigma, St. Louis, Mo.) per ml for 1 h on ice under constant agitation. To remove excess antibody, cells were pelleted, washed once with HEPES-DMEM, and resuspended at 10⁷/ml in the same medium. Then, 0.75 ml of the cell suspension was prewarmed for 5 min at 30°C and stimulated with 100 ng of dinitrophenylated BSA (DNP-BSA) (LSL, Tokyo, Tokyo) per ml for the indicated periods at 30°C. The reaction was terminated by a brief centrifugation (5,000 rpm, 30 s), followed by immediate lysis of the cell pellet with 500 µl of ice-cold Triton X-100 lysis buffer (20 mM Tris-HCl, pH 7.4; 150 mM NaCl; 0.5% Triton X-100; 1 mM EDTA; 1 mM vanadate; 20 mM -glycerophosphate) containing a protease inhibitor cocktail (5 µg of leupeptin, 10 µg of pepstatin, and 10 µg of aprotinin per ml and 0.2 mM phenylmethylsulfonyl fluoride). The cell lysates were centrifuged at 12,000 rpm for 10 min at 4°C, and the supernatants were subjected to immunoprecipitation.

Immunoprecipitation, immunoblotting, and in vitro kinase assay. The supernatants were incubated at 4°C for 1 h with 1 µg of anti-FcRI  subunit monoclonal antibody (JRK) (53), 20 µg of anti-Syk (LR) polyclonal antibody (Santa Cruz, Santa Cruz, Calif.) directed to linker region of human Syk, 10 µg of anti-ERK1 and -2 polyclonal antibody (Santa Cruz), or 20 µg of anti-c-myc polyclonal antibody (Santa Cruz) and then with 15 µl of 50% protein G-Sepharose slurry (Pharmacia-LKB, Bromma, Sweden) for another 30 min under constant rotation. Then, the beads were washed twice with 500 µl of ice-cold Triton X-100 lysis buffer.

Sucrose density gradient centrifugation. Protein association with detergent-resistant membranes (DRMs) was analyzed by solubilizing RBL cells with low-concentration Triton X-100 followed by ultracentrifugation of cell lysates on sucrose density gradients according to the methods of Field et al. (19). In brief, 4 × 10⁶/ml cell suspension was solubilized with 0.05% Triton X-100, cell lysate was layered onto 80 to 10% discontinuous sucrose gradients made in Hitachi 13 PA tube (1.5 × 9.6 cm) and centrifuged at 35,000 rpm at 4°C for 18 h (19). Then, 1-ml aliquots of the gradients were collected, proteins were extracted according to the method of Wessel and Flugge (85), and the sample was subjected to immunoblotting as described above.

Staining of cell surface FcRI. To analyze surface expression of FcRI, cell suspensions at 10⁶/ml in HEPES-DMEM were sensitized or not (control) with 1.0 µg of anti-TNP IgE per ml for 1 h on ice, washed twice with phosphate-buffered saline (PBS) containing 2% FCS, and stained with fluorescein isothiocyanate (FITC)-conjugated rabbit anti-mouse IgE (Southern Biotechnology Associates) in the same buffer for 30 min on ice. After two washes with the buffer, the cells were resuspended in PBS, and surface fluorescence was analyzed by EPIX-XL flow cytometer (Coulter).

Fluorometric imaging of [Ca²⁺]_i. Measurement of intracellular calcium concentration ([Ca²⁺]_i) was performed as described previously (27). In brief, cells were cultured on glass coverslips for 24 h and incubated in HEPES-DMEM containing 1 µg of anti-DNP IgE per ml and 5 µM Fura-2 AM (Dojindo, Kumamoto, Japan) for sensitization and Fura-2 loading at 37°C for 1 h. Cells were washed twice with HEPES-Tyrode buffer (25 mM HEPES-NaOH, pH 7.4; 140 mM NaCl; 2.7 mM KCl; 1.8 mM CaCl₂; 12 mM NaHCO₃; 5.6 mM D-glucose; 0.49 mM MgCl₂; 0.37 mM NaH₂PO₄) containing 0.1% BSA and incubated in the same buffer. Fluorometric images of cells (340 nm/380 nm) were recorded at every 30 s before and after the addition of 100 ng of DNP-BSA per ml under room temperature (20 to 25°C) with an iced charged-coupled device camera-image analysis system (Argus-50; Hamamatsu Photonics, Hamamatsu, Japan). To show time-dependent changes in [Ca²⁺]_i, seven cells in a field were randomly assigned, and the calculated average of [Ca²⁺]_i within the cell area was expressed as line graphs.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Creation of RBL2H3 cell lines expressing mCsk, mCsk(), or mCsk in combination with epitope-tagged Src family kinase mutants. In a previous work, we showed that overexpression of mCsk, a gain-of-function Csk mutant possessing a membrane-anchoring signal, profoundly delayed FcRI-mediated [Ca²⁺]_i elevation in RBL2H3 cells (27). In the current study, we selected a subclone exhibiting more apparent suppressive phenotypes, in which FcRI-mediated calcium mobilization was almost negligible under optimal stimulation conditions (see below and Fig. 2A). The mCsk clone was used as a background in which mutated Src family kinases were coexpressed. Of the Src family kinases, Lyn, Fyn, and c-Src were selected and tested for their abilities to restore FcRI signaling downregulated by mCsk. Lyn has been implicated as a major Src family kinase functioning in FcRI signaling (4, 18, 65). Fyn was chosen as a molecule distributed in a wide range of hematopoietic cells (13, 48). c-Src is a kinase expressed as abundantly as Lyn in RBL2H3 cells and is presumed to complement Lyn in FcRI signaling. In addition, c-Src is structurally different from Lyn and Fyn in that c-Src is devoid of palmitoylatable Cys3 conserved in a majority of Src family kinases (see below) (60, 61). Constructs created on the backbone of the Src family kinases are illustrated in Fig. 1A. a-Src, a-Lyn, and a-Fyn were derived from rat c-Src, human p56^lyn, and human p59^fyn, respectively. In these constructs, C-terminal 8 to 10 amino acids containing negative regulatory tyrosine were deleted and replaced with a c-myc tag epitope sequence (Fig. 1A).

View larger version (20K): [in this window] [in a new window]   FIG. 1.   Creation of RBL2H3 cells expressing mCsk in combination with Src family kinases whose C-terminal sequences were replaced with a c-myc epitope. (A) Schematic representation of the constructs derived from c-Src, Lyn, and Fyn. To create C-terminal tyrosine-deleted, c-myc-tagged Src family kinases, the conserved 8 to 10 amino acids at the C termini of c-Src, p56lyn, and p59fyn (see boxed amino acids in the sequence alignment) containing negative regulatory tyrosine (underlined) were replaced with the c-myc epitope sequence. "#" indicates the identical amino acid to that of wild-type kinases. The mutated kinases derived from c-Src, p56lyn, and p59fyn were named a-Src, a-Lyn, and a-Fyn, respectively. In a-Src S3C, Ser3 in a-Src was replaced with Cys. SH2 a-Lyn and SH3 a-Lyn were created by deleting the SH2 (amino acids 130 to 222) and SH3 (amino acids 68 to 117) domains from a-Lyn, respectively. (B) Representative immunoblots of cell lines expressing mCsk, mCsk(), or mCsk in combination with mutated Src family kinases. Cells were lysed with Triton X-100 solubilization buffer as described in Materials and Methods; 10 µg of protein at each lane was then separated by SDS-PAGE and subjected to immunoblots with anti-c-myc antibody (upper panel) to detect the mutated Src (a-Src) kinases and with anti-Csk antibody (lower panel). Cells expressing Neor control (lane 1), mCsk (lane 2), and a kinase-dead mCsk, mCsk() (lane 3) are shown. Cells expressing mCsk with Puror control (lane 4), a-Lyn (lane 5), a-Fyn (lane 6), a-Src (lane 7), a-Src S3C (lane 8), SH2 a-Lyn (lane 9), and SH3 a-Lyn (lane 10) are also shown. Migration positions of the expressed proteins are indicated on the left, and molecular mass markers are given on the right. (C) Analysis of surface expression of FcRI in the cell lines. Cells were sensitized or not (control) with mouse IgE and then stained with FITC-conjugated anti-mouse IgE. Surface fluorescence was analyzed by EPICS-XL flow cytometer. All of the cell lines exhibited levels of FcRI expression almost comparable to that of the neo control.

Influence of mCsk on FcRI-mediated signaling. Figure 2A shows time-dependent changes in [Ca²⁺]_i. Vector control (neo), mCsk-expressing cells, and mCsk()-expressing cells were sensitized with a saturable concentration of anti-TNP IgE (1.0 µg/ml) and stimulated with an optimal concentration of DNP-BSA (100 ng/ml) (27). Control cells and mCsk() cells clearly responded to the addition of DNP-BSA. The mCsk-overexpressing clone did not exhibit detectable [Ca²⁺]_i response within a 20-min incubation period.

View larger version (2922K): [in this window] [in a new window]   FIG. 2.   Influence of mCsk and mCsk() expression on FcRI-mediated early signaling. (A) Calcium mobilization in control cells (neo), mCsk-expressing cells (mCsk) and mCsk()-expressing cells [mCsk()]. Arrows indicate the time of antigen addition. The line graphs represent time courses of changes in [Ca2+]i in seven single cell areas randomly assigned. Insets represent pseudocolor images of [Ca2+]i at 0 min and at 3 min (2 min after antigen addition). Control neo cells rapidly responded to the antigen addition, and the signals subsided within 10 to 20 min. The mCsk subline exhibited almost negligible calcium response. Calcium signaling was preserved in mCsk() cells. (B) Time-dependent changes in tyrosine phosphorylation of FcRI  and  subunits (left panel) and in Syk tyrosine phosphorylation (right panel). Cells were stimulated and solubilized at the indicated periods as described in Materials and Methods. In the analysis of  and  tyrosine phosphorylation (left panel), FcRI was immunoprecipitated from cell lysates with anti- monoclonal antibody (JRK) (denoted as  ip) and subjected to immunoblotting with 4G10 antiphosphotyrosine antibody (pY blot), and the membrane was reprobed with JRK ( blot). Migration positions of tyrosine phosphorylated  (pY-) and g (pY-) subunit and reprobed  subunit () are indicated on the left. "(L)" represents the IgG light chain. Molecular mass markers are on the right. Films were scanned, and the intensity of the signal of tyrosine phosphorylated  (pY-) and  (pY-) were expressed as a bar graph (lower panel). In the control neo cells, pY- was detectable before FcRI clustering and  and  tyrosine phosphorylation increased, peaked within 5 min of stimulation, and then decreased. In mCsk cells, basal and clustering-induced signals were profoundly suppressed, while these were preserved in mCsk() cells. In the analysis of Syk tyrosine phosphorylation (right panel), Syk was immunoprecipitated with anti-Syk antibody (Syk ip), followed by immunoblotting with 4G10 (pY blot). The membrane was reprobed with anti-Syk antibody (Syk blot). Tyrosine-phosphorylated Syk (pY-Syk) and reprobed Syk (Syk) are indicated by arrows on the left. Molecular mass markers are on the right. Intensity of pY-Syk was shown as a bar graph (lower panel). Clustering-induced Syk tyrosine phosphorylation was clearly detectable in neo cells and was markedly suppressed in mCsk cells, whereas it was preserved in mCsk() cells. (C) ERK1 and -2 MAP kinase activation in neo, mCsk, and mCsk() cells. Cells were sensitized and lysed before () and 5 min after (+) FcRI clustering. ERK1 and -2 were immunoprecipitated and subjected to in vitro kinase assay (ERK IVK) by using Elk-1 as a substrate. Elk-1 phosphorylation was measured and visualized by using Fuji image analyzer BAS 2000. ERK1 and -2 content in total cell lysates (and mobility shift of ERK2) was analyzed by immunoblotting with anti-ERK1 and -2 antibody (ERK1, 2 blot). Positions of phosphorylated Elk-1 (p-Elk-1) and ERK1 and -2 are indicated by arrows on the left. Molecular mass markers are on the right. ERK1 and -2 activity was markedly increased after FcRI clustering in neo cells. mCsk expression suppressed it but mCsk() did not.

Differential abilities of a-Src kinases to restore FcRI-signaling in mCsk cells. We expressed each of the a-Src kinases in an mCSk background in order to compare their abilities to restore FcRI signaling. mCSk cells transfected with vector alone (Puro/mCsk), a-Lyn (a-Lyn/mCsk), a-Fyn (a-Fyn/mCsk), a-Src (a-Src/mCsk), or a-Src S3C possessing N-terminal palmitoylation signal (a-Src S3C/mCsk) were prepared (see Fig. 1B). These cells were sensitized with IgE and stimulated by receptor clustering, and tyrosine phosphorylation of  and  subunit and Syk was analyzed as described above. As seen in Fig. 3A ( ip), basal and clustering-induced  and  tyrosine phosphorylation was again low in Puro/mCsk control cells (lanes 1 and 2). Expression of a-Lyn resulted in increased basal  tyrosine phosphorylation (compare lane 1 with lane 3) and intense clustering-induced  and  tyrosine phosphorylation (compare lane 2 with lane 4). Expression of a-Fyn only modestly increased basal  tyrosine phosphorylation (compare lane 1 with lane 5) but clearly enhanced clustering-induced  and  tyrosine phosphorylation (compare lane 2 with lane 6). a-Src expression did not appreciably increase basal  or clustering-dependent  and  tyrosine phosphorylation above those in Puro/mCsk control (compare lanes 1 and 2 with lanes 7 and 8). In contrast, a-Src S3C expression induced clear increase in tyrosine phosphorylation of  and  after FcRI-clustering (compare lane 2 with lane 10), although its effects on basal  phosphorylation was marginal (lane 1 with lane 9).

View larger version (3026K): [in this window] [in a new window]   FIG. 3.   Differential abilities of a-Src kinases to restore FcRI signaling in mCsk cells. (A) Effects of the expression of a-Src kinases on FcRI  and  tyrosine phosphorylation and Syk tyrosine phosphorylation. mCsk Cells transfected with Puror vector alone (Puro/mCsk) or with mCsk Cells stably expressing a-Lyn (a-Lyn/mCsk), a-Fyn (a-Fyn/mCsk), a-Src (a-Src/mCsk), or a-Src S3C (a-Src S3C/mCsk) were sensitized with IgE. Cells were lyzed before () or 5 min after FcRI clustering (+). Tyrosine phosphorylation of  and  subunit and Syk was analyzed, and the data were expressed as described in the legends for Fig. 2B. Expression of a-Lyn, a-Fyn, and a-Src S3C enhanced tyrosine phosphorylation of  and  subunit and Syk more than the Puro/mCsk control, but a-Src expression was ineffective. See the text for details. (B) ERK1 and -2 MAP kinase activation in mCsk cells coexpressing a-Src kinases. ERK1 and -2 activities before () and 5 min after (+) FcRI clustering were measured by in vitro kinase assay using Elk-1 as a substrate, as described in the legend for Fig. 2C. Clustering-induced ERK1 and -2 activation was clearly observed in a-Lyn/mCsk cells, a-Fyn/mCsk, and a-Src S3C/mCsk cells. Effects of FcRI clustering was only marginal in Puro/mCsk control cells and in a-Src/mCsk cells. (C) Calcium mobilization in mCsk cells coexpressing a-Src kinases. Single cell [Ca2+]i recording was conducted as described in Materials and Methods. See also the legend for Fig. 2A. FcRI-mediated calcium mobilization was restored by a-Lyn, a-Fyn, and a-Src S3C but not by a-Src. Data obtained from the functional analyses were reproducible in two independent cell lines.

a-Src kinases that reconstitute FcRI signaling are physically associated with FcRI  subunit and localized at low-density DRMs. It is postulated that clustering-induced  tyrosine phosphorylation is initiated by small amount of Lyn associated with resting  subunit (80). Therefore, we next tested physical association of a-Src kinases with  by coimmunoprecipitation procedures. Cells were solubilized before and 5 min after FcRI clustering and subjected to immunoprecipitation with anti- monoclonal antibody. a-Src kinases were detected by immunoblotting with anti-c-myc antibody. As seen in Fig. 4A, a-Lyn, a-Fyn, and a-Src S3C were found to be coimmunoprecipitated with  subunit before and after receptor clustering, but a-Src was not detectably coimmunoprecipitated under either of the conditions. These findings strongly indicated that palmitoylation signal is required for Src family kinases to physically associate with  subunit.

View larger version (34K): [in this window] [in a new window]   FIG. 4.   Coimmunoprecipitation of a-Src kinases with  subunit (A) and fractionation of a-Src kinases and  subunit by sucrose density gradient centrifugation (B). (A) mCsk cells transfected with Puror vector alone (Puro/mCsk) or mCsk cells stably expressing a-Lyn (a-Lyn/mCsk), a-Fyn (a-Fyn/mCsk), a-Src (a-Src/mCsk), or a-Src S3C (a-Src S3C/mCsk) were solubilized before () and 5 min after (+) FcRI clustering, and FcRI complex was immunoprecipitated with JRK anti- monoclonal antibody. Immunoprecipitates ( ip) were subjected to immunoblotting with anti-c-myc antibody (myc blot) to detect a-Src kinases. Almost equal amounts of  subunits were recovered in the immunoprecipitates (not shown). An equal volume of aliquots of the total cell lysates (Lysate) was also analyzed by anti-c-myc immunoblotting (myc blot) to ascertain comparable solubilization of a-Src kinases. a-Lyn, a-Fyn, and a-Src S3C were detected in  immunoprecipitates before and after FcRI clustering, but a-Src was not detectable under either of the conditions. (B) Quiescent a-Src kinase-expressing mCsk cells were solubilized and subjected to discontinuous sucrose gradient ultracentrifugation as described in Materials and Methods. Protein was extracted and subjected to immunoblotting with anti-c-myc antibody (myc blot) or with JRK ( blot). a-Lyn, a-Fyn, and a-Src S3C possessing N-terminal palmitoylation signal were recovered at low-density fractions (fractions 8 to 11), and a-Src was recovered at high-density fractions (fractions 2 to 6). The  subunit was mainly found at intermediate fractions (fractions 7 to 9) and was codistributed with palmitoylatable a-Src kinases. The  immunoblotting was conducted by using a-Src/mCsk cells. Other cell lines exhibited almost identical  distributions (not shown).

Roles of SH2 and SH3 domains of Lyn in FcRI signaling. Noncatalytic SH2 and SH3 domains of Lyn potentially contribute to FcRI signaling through molecular assembly with  and  subunits or with outer signaling molecules (1-3, 34, 55). However, it has not been determined if these domains are required for FcRI signaling. To investigate the roles of SH2 and SH3 domains, a-Lyn lacking the SH2 domain (SH2 a-Lyn) and one lacking the SH3 domain (SH3 a-Lyn) were created and tested for their abilities to restore FcRI signaling in mCsk cells. As seen in Fig. 5A ( ip), expression of a-Lyn again clearly enhanced basal  and triggering-induced  and  tyrosine phosphorylation above control levels in Puro/mCsk cells (compare lanes 1 and 2 with lanes 3 and 4). Expression of SH2 a-Lyn and SH3 a-Lyn also resulted in marked increases in basal and triggering-induced signals (compare lanes 1 and 2 with lanes 5 and 6 or with lanes 7 and 8), and the extents were similar to those caused by a-Lyn expression (compare lanes 3 and 4 with lanes 5 and 6 or with lanes 7 and 8). In contrast, the abilities of SH2 a-Lyn and by SH3 a-Lyn to induce Syk tyrosine phosphorylation were considerably different. As seen in Fig. 5A (Syk ip), expression of a-Lyn again resulted in a moderate increase in basal Syk tyrosine phosphorylation and marked clustering-induced Syk tyrosine phosphorylation (compare lanes 1 and 2 with lanes 3 and 4). Expression of SH2 a-Lyn did not have an influence on the basal signal and only slightly increased clustering-induced Syk tyrosine phosphorylation (compare lanes 1 and 2 with lanes 5 and 6). SH3 a-Lyn expression clearly enhanced basal and triggering-induced signals (compare lanes 1 and 2 with lanes 7 and 8) as effectively as a-Lyn did (compare lanes 3 and 4 with lanes 7 and 8).

View larger version (2931K): [in this window] [in a new window]   FIG. 5.   Reconstruction of FcRI signaling by a-Lyn, SH2 a-Lyn, or by SH3 a-Lyn. (A) Effects of the expression of a-Lyn, SH2 a-Lyn, or SH3 a-Lyn on FcRI  and  tyrosine phosphorylation and Syk tyrosine phosphorylation. Puro/mCsk cells, a-Lyn/mCsk cells, SH2 a-Lyn/mCsk cells, and SH3 a-Lyn/mCsk cells were lyzed before () and 5 min after FcRI clustering (+). Tyrosine phosphorylation of  and  subunits and Syk was analyzed, and the data are expressed as described in the legend for Fig. 2B. Expression of SH2 a-Lyn and SH3 a-Lyn enhanced basal- and clustering-induced  and  tyrosine phosphorylation as effectively as a-Lyn did. a-Lyn and SH3 a-Lyn clearly augmented Syk tyrosine phosphorylation above control levels in Puro/mCsk cells, whereas the effects of SH2 a-Lyn were marginal. See the text for details. (B) ERK1 and -2 MAP kinase activation in mCsk cells coexpressing a-Lyn, SH2 a-Lyn, or SH3 a-Lyn. ERK1 and -2 activities before () and 5 min after (+) FcRI clustering were analyzed as described in the legend for Fig. 2C. Clustering-induced ERK1 and -2 activation was restored in a-Lyn/mCsk cells and in SH3 a-Lyn/mCsk cells. In SH2 a-Lyn/mCsk cells, FcRI-independent, basal kinase activity was increased above that in Puro/mCsk cells, but the clustering-induced increase was considerably smaller than in a-Lyn/mCsk cells and SH3 a-Lyn/mCsk cells. See the text for details. (C) Calcium mobilization in mCsk cells expressing a-Lyn, SH2 a-Lyn, or SH3 a-Lyn. Data are expressed as described the legend for Fig. 2A. FcRI-mediated calcium mobilization was restored by a-Lyn and by SH3 a-Lyn but not by SH2 a-Lyn. Data obtained from the functional analyses were reproducible in two independent cell lines.

Deletion of SH2 or SH3 domain did not abrogate a-Lyn association with FcRI  subunit or localization at DRMs. The possible association of SH2 a-Lyn and SH3 a-Lyn with FcRI complex was tested by coimmunoprecipitation experiments with JRK anti- antibody. As seen in Fig. 6A, SH2 a-Lyn and SH3 a-Lyn were coimmunoprecipitated with  subunit, before and after receptor clustering. We next tested the association of SH2 a-Lyn and SH3 a-Lyn with low-density, DRMs by sucrose density gradient centrifugation followed by immunoblotting. As seen in Fig. 6B, SH2 a-Lyn and SH3 a-Lyn were mainly localized at low-density fractions (fractions 8 to 11), and their distributions were almost identical to that of a-Lyn. The  subunit was again found mainly in slightly higher density fractions (fractions 7 to 9), and it was partially codistributed with SH2 a-Lyn and with SH3 a-Lyn (fractions 8 and 9). These findings indicated that those submolecular domains are not essential for Lyn to physically associate with  subunit or to be localized at DRMs.

View larger version (33K): [in this window] [in a new window]   FIG. 6.   Physical association of a-Lyn, SH2 a-Lyn, or SH3 a-Lyn with FcRI  subunit (A) and their distribution as analyzed by sucrose density gradient centrifugation (B). (A) Puro/mCsk cells, a-Lyn/mCsk cells, SH2 a-Lyn/mCsk cells, and SH3 a-Lyn/mCsk cells were solubilized before () and 5 min after (+) FcRI clustering and subjected to coimmunoprecipitation analysis by using JRK anti- antibody as described in the legend for Fig. 4A.  Immunoprecipitates ( ip) were subjected to immunoblotting with anti-c-myc antibody (myc blot). Almost equal amounts of  subunits were recovered in the immunoprecipitates (not shown). An equal volume of aliquots of the total cell lysates (Lysate) was also analyzed by anti-c-myc immunoblotting (myc blot) to ascertain comparable solubilization efficiency. a-Lyn, SH2 a-Lyn, or SH3 a-Lyn were detected in -immunoprecipitates before and after FcRI-clustering. (B) Quiescent a-Lyn/mCsk cells, SH2 a-Lyn/mCsk cells, and SH3 a-Lyn/mCsk cells were solubilized, and the distributions of a-Lyn-derived kinases and  subunit were analyzed by sucrose density gradient centrifugation, as described in the legend for Fig. 4B. a-Lyn, SH2 a-Lyn, and SH3 a-Lyn were found mainly at low-density fractions (fractions 7 to 11) and  subunit was mainly in the slightly higher density fractions (fractions 6 to 9) (see also Fig. 4B). The distribution of  subunit partly overlapped with those of a-Lyn-derived kinases. The  immunoblotting was conducted with a-Lyn/mCsk cells. Other cell lines exhibited almost identical  distributions (not shown).

Catalytic activity of a-Src kinases. The inability of a-Src or SH2 a-Lyn to transmit FcRI signaling could be due to impaired catalytic activity of the constructs. To exclude the possibility, a-Src kinases were immunoprecipitated with anti-c-myc antibody and subjected to in vitro autophosphorylation assay. As seen in Fig. 7, in all of the immunoprecipitates from cells expressing a-Src kinases, kinase activity was detected, and autophosphorylation signals at expected migration positions were observed. Immunoprecipitates from Puro/mCsk cells did not contain kinase activity. In addition, autophosphorylation signals of a-Src and SH2 a-Lyn were comparable to or even higher than those of a-Src S3C and a-Lyn, respectively. From these observations, it was concluded that the defective signal transduction by a-Src or SH2 a-Lyn was not ascribed to reduced catalytic activities of these constructs.

View larger version (19K): [in this window] [in a new window]   FIG. 7.   In vitro kinase assay of a-Src kinases. Quiescent Puro/mCsk cells, a-Lyn/mCsk cells, a-Fyn/mCsk cells, a-Src/mCsk cells, a-Src S3C/mCsk cells, SH2 a-Lyn/mCsk cells, and SH3 a-Lyn/mCsk cells were solubilized, a-Src kinases were immunoprecipitated with anti-c-myc antibody, and autophosphorylation activities were assayed. Samples were separated by SDS-PAGE and analyzed by using a Fuji image analyzer BAS 2000. Expected migration positions of autophosphorylated a-Src kinase are indicated by arrows. Molecular mass markers are on the right. Kinase activity was not recovered from Puro/mCsk cell lysate. Phosphorylated proteins with molecular masses identical to those of a-Src kinases and other species were detected in the samples from a-Src kinase-expressing cells.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Although the sequential biochemical signals following FcRI aggregation have been elucidated in increasing detail, the roles of individual Src family members and their submolecular domains in the initiation and progression of the signaling cascade have not been fully elucidated. To address these issues, we attempted to dissect the early signaling by reconstitution experiments. FcRI signaling in RBL2H3 cells was first suppressed by mCsk overexpression and then reconstituted with a-Src kinases lacking C-terminal regulatory tyrosine. To directly compare the expression levels of the kinases and, consequently, their relative abilities to restore the signaling, constructs were designed in which the C-terminal sequences were replaced with a c-myc tag. We utilized a-Lyn and a-Fyn as kinases containing palmitoylatable Cys3 in their SH4 domain and a-Src as one lacking the residue (61, 76). A mutated a-Src possessing Cys3 was also prepared to further ascertain the role of the palmitoylation site. To assess the roles of SH2 and SH3 domains, a-Lyn-derived constructs lacking either of the domains were created.

Through the initial comparison of control cells, mCsk- and mCsk()-overexpressing cells, it was observed that basal tyrosine phosphorylation of FcRI  subunit was suppressed by mCsk in a kinase-dependent manner. These findings supported the proposed concept that Src family kinase(s) associated with resting  subunit is in part in a C-terminal tyrosine-dephosphorylated open conformation (56, 74, 88). mCsk also potently downregulated clustering-induced tyrosine phosphorylation of  and  subunits and Syk tyrosine kinase, calcium mobilization, and ERK1 and -2 MAP kinase activation in a kinase-dependent manner. mCsk suppressed the series of FcRI signaling most probably by decreasing the ratio of C-terminal tyrosine-dephosphorylated, open-conformation kinases. These findings also supported the physiological relevance of the current reconstitution experiments with C-terminal tyrosine-deleted a-Src kinases.

One of major findings in the current study is that FcRI signaling is restored by the expression of a-Src kinases possessing N-terminal palmitoylation signal. Expression of a-Lyn and a-Fyn, both of which possess palmitoylatable Cys3, effectively reconstructed the triggering-induced tyrosine phosphorylation of , , and Syk, the Syk association with , the ERK1 and -2 activation and calcium signal; however, a-Src expression did not induce initial signal of  and  tyrosine phosphorylation. Creation of Cys3 in a-Src (a-Src S3C) resulted in effective restoration of the series of signaling. The inability of a-Src to reconstruct the signals could not be ascribed to loss of catalytic activity in a-Src construct (see reference 25 and Fig. 7) or to accelerated downregulation of surface FcRI (see Fig. 1C). Although it should be taken into account that a-Src kinases are not in normal equilibrium, these findings strongly indicated that FcRI signaling is catalyzed by selected Src family kinases possessing N-terminal palmitoylation site. As an alternative explanation, it may be possible that overexpression of a-Src kinases compete for the inhibitory action of mCsk by titrating out mCsk molecule and that FcRI signal in a-Src-kinase-expressing cells are initiated by endogenous Src family kinases. These possibilities should be examined by further studies. However, differential abilities of a-Src kinases and a-Lyn-derived constructs to convey the signaling suggest that signal restoration by a-Src kinases is not solely due to the effects of overexpression.

It was also found that a-Lyn, a-Fyn, and a-Src S3C was physically associated with resting and aggregated  subunit but that a-Src was not. Based upon these two lines of evidence, it is strongly argued that N-terminal palmitoylation is required for Src family kinases to associate with resting FcRI and that this association is critical for kinases to initiate  and  tyrosine phosphorylation after FcRI clustering. In accordance with our observations, Timson Gauen et al. have shown that N-terminal palmitoylation is required for Src family kinases to associate with ectopically expressed chimeric TCR  subunit (76). In the initial study that related Src family kinases with FcRI, c-Src was shown to be activated after FcRI clustering (18). The current data suggest that the c-Src activation is not the direct consequence of receptor aggregation but is a more downstream event.

The series of observations including mCsk-mediated suppression of basal  and clustering-dependent  and  tyrosine phosphorylation, physical association of coexpressed a-Lyn with resting FcRI, and the resultant recovery of triggering-induced  and  tyrosine phosphorylation are reminiscent of the proposed concept that the early FcRI signaling relies upon receptor-associated kinase in open conformation (80, 88). It may be of note that a-Lyn expression enhanced clustering-independent, basal tyrosine phosphorylation of  subunit and Syk. Therefore, interaction of autoactive a-Lyn with  subunit partially mimicked FcRI signaling. However, receptor clustering was still required for full signaling. How receptor clustering facilitates the signal progression in a-Lyn/mCsk cells is an intriguing issue yet to be investigated. a-Lyn might be further activated after receptor clustering by SH3 domain displacement (86) and/or by phosphorylation of activation loop tyrosine, or it may be that entirely different signaling mechanisms are required in parallel to induce full signaling. Concerning the relative abilities of a-Src kinases to initiate FcRI signaling, basal  tyrosine phosphorylation was effectively increased by a-Lyn and to comparable levels by SH2 a-Lyn and by SH3 a-Lyn, whereas only moderately by a-Fyn and by a-Src S3C (see Fig. 3A and 5A,  ip). Clustering-induced  and  tyrosine phosphorylation was also most effectively enhanced by the Lyn-derived constructs. These findings suggest the existence of structural characteristics of Lyn, apart from palmitoylation, SH2 domain, or SH3 domain, that are advantageous for FcRI signal triggering.

It has become increasingly clear that one of the roles of N-terminal palmitoylation is to sort Src family kinases into specialized low-density membrane fractions, DRMs, GEMs, or sphingolipid-cholesterol rafts (7, 31, 67, 70). Recent biochemical and fluorescence imaging analyses showed that FcRI is rapidly translocated to DRMs, in the vicinity of Src family kinases, and that FcRI subunits are then phosphorylated by DRM-associated kinases (19, 21). This concept that recruitment of FcRI to DRMs is required for FcRI to meet Src family kinases is somewhat different from our proposal that clustering-induced early signaling is catalyzed by Src family kinases already interacting with resting receptor. We thus examined the distribution of a-Src kinases and  subunit in resting RBL clones by sucrose density gradient centrifugation of the cell lysates (19, 21). Our data showed that a-Lyn, a-Fyn, and a-Src S3C are localized almost exclusively at low-density fractions that presumably correspond to DRMs and that a-Src was at clearly separated high density fractions (see Fig. 4B). Intriguingly,  subunit was found in intermediate-density fractions that are in part overlapped with those of DRM-associated a-Src family kinases. Although precise characterization of the -associating, intermediate-density membrane domains should be elucidated by future studies, codistribution of resting  subunit with N-palmitoylatable kinases, together with their physical association (see Fig. 4C) and with increased basal  tyrosine phosphorylation in a-Lyn/mCsk cells (Fig. 3A and 5A), strongly argued that the interaction occurs before receptor clustering. Montixi et al. showed that PP1, a tyrosine kinase inhibitor relatively specific to Src family kinases, suppressed TCR recruitment to the DRMs (47). Therefore, it could be postulated that clustering-induced FcRI signaling is initiated by kinases associated with resting  subunit and that FcRI association with DRMs after receptor engagement is a signal amplification process.

The next major finding is that the SH2 domain of Lyn is not essential for tyrosine phosphorylation of FcRI subunits but is required for the progression of the signal: deletion of SH2 or SH3 domain from a-Lyn did not significantly alter its ability to associate with DRMs or with resting FcRI or to augment basal  and clustering-induced  and  tyrosine phosphorylation, albeit that  tyrosine phosphorylation by SH2 a-Lyn is slightly less effective than those by a-Lyn or by SH3 a-Lyn (see Fig. 5A,  ip). In contrast, SH2 deletion from a-Lyn profoundly decreased its ability to enhance clustering-induced Syk tyrosine phosphorylation and its association with  subunit. In the TCR system, Straus et al. showed that Lck containing a mutation in the phosphotyrosine-binding pocket of the SH2 domain only marginally induced tyrosine phosphorylation of TCR  subunit, a homologue of FcRI  subunit, in JCam1 T cells and proposed that SH2 domain is indispensable for the earliest signal of  phosphorylation (72). Our findings are in part consistent with these results, but the presence of the FcRI-specific  subunit, which lies upstream of  (39) in our system, allowed more precise dissection of the early signaling. Our data indicated that SH2 domain is required to link initial  and  tyrosine phosphorylation to recruitment of Syk to  and its tyrosine phosphorylation. It may be worthy to note that the abortive signaling by SH2 a-Lyn is quite similar to the interrupted signaling by antigen and Fc receptors stimulated with partial agonists: FcRI or TCR stimulated with agonists with low affinity or valence was shown to induce the tyrosine phosphorylation of ITAM-containing subunits, whereas they were unable to recruit or phosphorylate Syk or closely related Zap 70 tyrosine kinase (33, 38, 77). The signal interruption is ascribed to insufficient time for the agonist-receptor interaction to overcome the time-consuming step of signaling complex formation (kinetic proofreading model) (44). From the above data, it could be postulated that SH2 domain of Lyn is required to pass through the first proofreading step, presumably by stabilizing signaling molecule complex. The observed decrease in  tyrosine phosphorylation could be due to different phosphorylation patterns from that in the full signaling as observed in partial agonist-stimulated TCR  subunit (32) or to the absence of -Syk association that is protective against rapid dephosphorylation of  (66). Obviously, how SH2 domain of Lyn works in the signal progression should be elucidated by further studies. Associations of Lyn's SH2 domain with tyrosine phosphorylated  subunit (34) and with Syk (2, 3) are apparent candidates of the signal progression mechanisms.

It seemed that the SH3 domain of Lyn is not essential for early FcRI signaling leading to calcium mobilization and to ERK1 and -2 activation. The SH3 domain of Lyn potentially interacts with p85 subunit of phosphatidylinositol 3-kinase or with Bruton's tyrosine kinase (1, 55), both of which are required for calcium signaling, especially in its sustained phases (9, 22). We could not detect differences in [Ca²⁺]_i elevation between a-Lyn cells and SH3 a-Lyn cells. In addition, Bruton's tyrosine kinase was almost equally tyrosine phosphorylated after FcRI clustering in these cell lines (not shown). These data suggest that phosphatidylinositol 3-kinase and Bruton's tyrosine kinase are recruited to the membrane by interacting with more physiological ligands, such as c-Cbl (24) and phosphatidylinositol trisphosphate (9, 22), respectively. Caron et al. showed that Lck 505F lacking SH3 domain fully augments TCR-mediated early signaling, including protein tyrosine phosphorylation, but that it was unable to induce interleukin-2 production (10). We also examined FcRI-mediated TNF- synthesis in RBL cell lines and found that mCsk expression did not significantly suppress TNF- synthesis, in spite of effective downregulation of the upstream signal of ERK1 and -2 activation (see Fig. 2C). One of our laboratories also noted that FcRI-mediated cytokine production is preserved in Lyn^/ murine mast cells, in which protein tyrosine phosphorylation is profoundly suppressed (51). The apparent discrepancy between early and late signaling events might indicate unexpectedly low threshold of FcRI-mediated cytokine transcription.

The current reconstitution study using the panel of mutated Src kinases finely dissected early FcRI signaling is summarized in Fig. 8. The Cys3 and SH2 domains of Src family kinases seem to function at two different steps in the pathway. First, Cys3 is required for the interaction of kinases with resting FcRI  and (designated as Src*), as assessed by their coimmunoprecipitation and codistribution after sucrose density gradient centrifugation. a-Src kinases associated with resting FcRI  were able to reconstruct triggering-induced  and  tyrosine phosphorylation (,-pY), recruitment of Syk to phosphorylated  (*Syk), Syk tyrosine phosphorylation (Syk-pY), and calcium mobilization ([Ca²⁺]_i) and ERK1 and -2 activation (ERK1 and -2). The SH2 domain of Lyn is dispensable for its association with FcRI  (Src*) or for triggering-dependent  and  tyrosine phosphorylation (,-pY), but is required to link the signal to the next step of Syk recruitment (*Syk) and its tyrosine phosphorylation (Syk-pY). These data may further the knowledge required for the understanding of the most fundamental issue of how receptor aggregation is translated into biochemical signaling.

View larger version (11K): [in this window] [in a new window]   FIG. 8.   Relative abilities of a-Src kinases to restore the sequential signaling steps. "*" and "-pY" denote physical interaction and tyrosine phosphorylation, respectively. [Ca2+]i and "ERK1, 2" indicate the increase in intracellular calcium concentration and ERK1 and -2 MAP kinase activation. See the text for details.