MIST Functions through Distinct Domains in Immunoreceptor Signaling in the Presence and Absence of LAT*

MIST (also termed Clnk) is an adaptor protein structurally related to SLP-76 and BLNK/BASH/SLP-65 hematopoietic cell-specific adaptor proteins. By using the BLNK-deficient DT40 chicken B cell system, we demonstrated MIST functions through distinct intramolecular domains in immunoreceptor signaling depending on the availability of linker for activation of T cells (LAT). MIST can partially restore the B cell antigen receptor (BCR) signaling in the BLNK-deficient cells, which requires phosphorylation of the two N-terminal tyrosine residues. Co-expression of LAT with MIST fully restored the BCR signaling and dispenses with the requirement of the two tyrosines in MIST for BCR signaling. However, some other tyrosine(s), as well as the Src homology (SH) 2 domain and the two proline-rich regions in MIST, is still required for full reconstitution of the BCR signaling, in cooperation with LAT. The C-terminal proline-rich region of MIST is dispensable for the LAT-aided full restoration of MAP kinase activation, although it is responsible for the interaction with LAT and for the localization in glycolipid-enriched microdomains. On the other hand, the N-terminal proline-rich region, which is a binding site of the SH3 domain of phospholipase Cγ, is essential for BCR signaling. These results revealed a marked plasticity of MIST function as an adaptor in the cell contexts with or without LAT.

Adaptor/linker proteins that lack enzymatic activities exert their function as a scaffold molecule to generate active signaling complexes, which are essential for transducing receptor signals to downstream effectors (1,2). Among these adaptor/ linker proteins are SLP-76 and its close relative BLNK (also known as BASH or SLP-65), which are pivotal for T and B cell development and for T cell antigen receptor (TCR) 1 and B cell antigen receptor (BCR) signaling, respectively. Common structural features of these proteins are the presence of N-terminal tyrosine phosphorylation sites, central proline-rich (PR) regions, and a C-terminal SH2 domain (3)(4)(5)(6). Upon TCR and BCR cross-linking, SLP-76 and BLNK are phosphorylated by ZAP70 and Syk protein-tyrosine kinases (PTKs), respectively, and then associate with signaling proteins, including phospholipase C␥ (PLC␥), Vav, and Grb2 (3,4,(7)(8)(9). In a SLP-76deficient Jurkat T cell line, elevation of intracellular Ca 2ϩ concentration ([Ca 2ϩ ] i ) and the activation of the Ras pathway following TCR cross-linking are severely impaired (10). Moreover, a profound block in thymocyte development in SLP-76deficient mice indicates an essential role for SLP-76 in pre-TCR signaling (11,12). Similarly, a BLNK-deficient DT40 cell line, a chicken B lymphoma, displayed severe defects in tyrosine phosphorylation of PLC␥, [Ca 2ϩ ] i increase, and c-Jun N-terminal kinase (JNK) and p38 MAP kinase activation following BCR stimulation (13). Extracellular signal-regulated kinase (ERK) 2 activation was also moderately affected by the absence of PLC␥ activation in BLNK-deficient DT40 cells. In addition, we and others have reported abnormal B cell differentiation and functions in BLNK-deficient mice (14 -17).
The third member of SLP-76/BLNK adaptor family was recently isolated and called either MIST (for mast cell immunoreceptor signal transducer) (18) or Clnk (for cytokine-dependent hemopoietic cell linker) (19). MIST/Clnk is constitutively expressed in several mast cell lines, and its expression was induced by cytokines in T cells and a variety of cytokine-dependent cell lines of myeloid and lymphoid origins (18,19). MIST shares structural features with SLP-76 and BLNK and is tyrosine-phosphorylated upon either the high affinity IgE receptor (Fc⑀RI) or TCR stimulation. MIST can associate with PLC␥, Vav, Grb2, and LAT either constitutively or inducibly. Overexpression of a phosphorylation-deficient form of MIST in the rat mast cell line RBL-2H3 inhibited Fc⑀RI-mediated mast cell degranulation, [Ca 2ϩ ] i increase, and phosphorylation of LAT (18). Furthermore, the transient expression of Clnk in Jurkat T cells augmented TCR-induced NF-AT activation (19). These findings indicate that, similarly to SLP-76 and BLNK, MIST/Clnk functions as a signaling molecule downstream of the Fc⑀RI in mast cells and the TCR in T cells.
To clarify domains/motifs required for MIST function in immunoreceptor signal transduction, we utilized BLNK-deficient DT40 cells as a signal reconstitution system. This B cell line expresses none of the SLP-76 family members and thus can be the simplest system currently available to test the function of various mutant forms of MIST. By this reconstitution system, we studied a structure-function relationship of MIST in linking immunoreceptor signal to the downstream biochemical events.

EXPERIMENTAL PROCEDURES
Cell Lines and Reagents-Wild-type and mutant DT40 cells deficient for Lyn, Syk, Btk, or BLNK (13,20,21) were cultured in RPMI 1640 supplemented with 10% fetal calf serum, 1% chicken serum, 50 M 2-mercaptoethanol, 2 mM L-glutamine, and antibiotics. COS-7 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum and antibiotics. Glutathione S-transferase (GST) protein fused to the C-terminal SH2 domain or the SH3 domain of PLC␥1 was purchased from Santa Cruz Biotechnology. The monoclonal Ab against chicken IgM (M4) was provided from Dr. C-L. H. Chen (University of Alabama at Birmingham). The rabbit Abs against chicken PLC␥2 and mouse MIST were described previously (18,22). The following antibodies were purchased: anti-ERK2 and anti-Grb2 Abs from Santa Cruz Biotechnology, anti-phosphotyrosine PY20 Ab conjugated with horseradish peroxidase from Transduction Laboratories, anti-LAT Ab from Upstate Biotechnology, anti-JNK Ab from PharMingen, and anti-T7 Ab from Novagen.
Expression Plasmids, Transfection, and Generation of Stable Transfectants-The mutant mouse MIST containing tyrosine to phenylalanine substitutions (YF) or arginine to lysine substitution at the SH2 domain (R335K) were generated using a QuickChange site-directed mutagenesis kit (Stratagene). The MIST mutants lacking PR regions ( 160 PLPPPR 165 or/and 178 PPAPP 182 ) were also generated by PCR-mediated mutagenesis as described previously (23). Their sequences were verified by automated sequence analysis. Wild-type and the mutant MIST cDNAs were subcloned into pApuro2 and pCAT7neo expression vectors (5,21). Human LAT cDNA was also inserted in pAzeo expression vector (24). Human Grb2 expression vector was provided by Dr. M. Tanaka (Hamamatsu Medical College). COS-7 cells were transfected with plasmid vectors by using a TransIT-LT1 reagent (Pan Vera) as described previously (5). The MIST or LAT expression vectors were transfected into wild-type and mutant DT40 cell lines by electroporation, and selected in the presence of 0.5 g/ml puromycin (Sigma) for pApuro-based constructs or 0.4 mg/ml Zeocin (Invitrogen) for pAzeo-LAT. The BLNK-deficient DT40 cell clone stably expressing human LAT (2H12) was further transfected with expression vectors containing various forms of MIST. Cell surface expression of BCR on each transfectants was analyzed by FACSsort (Becton Dickinson) using fluorescein isothiocyanate-conjugated anti-chicken IgM Ab (Bethyl Laboratories), and transfectants expressing similar levels of MIST/LAT proteins as well as surface IgM were used for the analysis.
Immunoprecipitations, GST Pull-down, and Immunoblot Analysis-Cells were lysed with 1% Nonidet P-40 lysis buffer containing protease and phosphatase inhibitors, and precipitated with indicated Abs or GST fusion proteins bound to glutathione-Sepharose (Amersham Pharmacia Biotech). The precipitates and aliquots of total cell lysates were resolved by SDS-polyacrylamide gel electrophoresis and transferred to polyvinylidene difluoride membranes. The membranes were immunoblotted with Abs described above, and the secondary Ab was conjugated with horseradish peroxidase and then developed with the ECL system (Amersham Pharmacia Biotech).
Measurement of Intracellular Ca 2ϩ Concentration-DT40 cell clones (5 ϫ 10 6 cells) were loaded with 5 M Fura-2/AM (Molecular Probes) at 37°C for 30 min, washed twice with RPMI 1640 medium containing 0.1% bovine serum albumin and adjusted to 10 6 cells/ml. Cells were then stimulated with 10 g/ml of anti-IgM Ab at 37°C. [Ca 2ϩ ] i was monitored at a 510 nm emission wavelength excited by 340 nm and 360 nm using a fluorescence spectrophotometer (F-2000, Hitachi).
In Vitro Kinase Assay-The assay conditions were described previously (25). In brief, lysates from 5 ϫ 10 6 cells were immunoprecipitated with 1 g of anti-ERK2 Ab or anti-JNK Ab and protein-G Sepharose. Half of the immunoprecipitates was suspended in 30 l of kinase assay buffer containing [␥-32 P]ATP, 5 M cold ATP, and 5 g of GST-Elk or GST-Jun fusion proteins as a substrate. After 20 min of incubation at 30°C, the reaction was terminated by the addition of SDS sample buffer, followed by boiling for 5 min. The samples were separated on SDS-polyacrylamide gel electrophoresis gels, and then the gels were dried and subjected to autoradiography. The other half of the immunoprecipitates were used for Western blot analysis to quantify immunoprecipitated proteins.
Luciferase Assay-BLNK-deficient DT40 cells were transfected with 10 g of a luciferase reporter plasmid driven by NF-AT and AP-1 response element from the mouse interleukin 2 gene promoter (a gift from Dr. K. Arai, Institute of Medical Science, University of Tokyo), together with 15 g of either empty vector or an expression vector harboring various forms of MIST, BLNK, or LAT, in serum-free RPMI at a density of 10 7 cells/400 l with a Gene Pulser (Bio-Rad Laborato-ries) set at 250 V and 975 F. After electroporation, the cells were transferred to complete RPMI and incubated at 40°C for 24 h. Triplicates of 5 ϫ 10 5 viable cells were then left unstimulated or stimulated with anti-IgM Ab for 6 h and subsequently assayed for luciferase activity, as described previously (5). Light emission was measured in a Lumat LB9501 luminometer (Berthold).
Preparation of GEM Fractions-GEM fractions were prepared as described previously (24). In brief, cells (2.5 ϫ 10 8 ) were either unstimulated or stimulated with anti-IgM Ab for 2 min and then lysed in 1 ml of 0.5% Triton X lysis buffer. Cell lysates were mixed with 1 ml of 80% sucrose in lysis buffer, overlaid with 6.5 ml of 30% sucrose and 2.5 ml of 5% sucrose in lysis buffer, and then subjected to ultracentrifugation at 35,000 rpm for 16 h at 4°C. Eleven aliquots (1 ml each) were collected from the top of the gradient fraction and subjected to Western blot analysis.

Tyrosines at Positions 69 and 96 Are Essential for Tyrosine Phosphorylation of MIST upon BCR Cross-linking-Mouse
MIST contains eight tyrosine residues potentially phosphorylated by PTKs, two PR regions, and a C-terminal SH2 domain (Fig. 1A). Of these tyrosine residues, five are conserved with those in human homologue (18), and two of them are located in the sequence context (DY 69 EDP and EY 96 ADT) similar to those in mouse SLP-76 (DY 113 ESP, DY 128 ESP, and DY 145 EPP) and BLNK (DY 72 ENP and EY 119 VDN). These tyrosine residues in SLP-76 are essential for TCR-induced tyrosine phosphorylation of whole molecule and for its function in TCR-induced NF-AT transcriptional activation (26). We first determined PTKs responsible for MIST phosphorylation in the context of BCR signaling by introducing MIST into wild-type and PTK-deficient DT40 cell lines. A low level of tyrosine phosphorylation of MIST was detected before stimulation, and it was markedly increased within 1 min after BCR stimulation in wild-type DT40 cells (Fig. 1B). By contrast, the level of BCR-induced tyrosine phosphorylation of MIST was profoundly decreased in Lyn-deficient DT40 cells, modestly decreased in Syk-deficient DT40 cells, and not decreased at all in Btk-deficient DT40 cells (Fig. 1A). This result indicates that both Lyn and Syk, but not Btk, mainly mediate tyrosine phosphorylation of MIST in the context of BCR signaling.
In order to determine whether the two conserved tyrosine residues of MIST are the targets for BCR-associated PTKs, we generated mutant forms of MIST that contain either single (Y69F or Y96F) or double (YF2) tyrosine to phenylalanine substitutions at positions 69 and 96 and examined their tyrosine phosphorylation in wild-type DT40 cells. As shown in Fig.  1C, Y69F and Y96F were poorly tyrosine-phosphorylated upon BCR stimulation, whereas YF2 was not, indicating that tyrosines 69 and 96 of MIST are the main targets of BCR-associated PTKs and required for the phosphorylation of MIST in the BCR-signaling context.
MIST Partially Restores BCR-mediated [Ca 2ϩ ] i Increase, PLC␥2 Phosphorylation, and ERK2 Activation in BLNK-deficient DT40 Cells-To clarify the role of MIST as a component of the immunoreceptor signaling complex and to verify the significance of the tyrosine phosphorylation of MIST, we introduced wild-type and mutant forms of MIST into BLNK-deficient DT40 cells (13). All transfectants expressed comparative levels of MIST protein ( Fig. 2A) and surface IgM (data not shown). The level of BCR-mediated tyrosine phosphorylation of Y69F was markedly lower, whereas that of Y96F was slightly lower, compared with that of wild-type MIST ( Fig. 2A). Tyrosine phosphorylation of YF2 was undetectable. As reported previously, no [Ca 2ϩ ] i increase was induced by BCR stimulation in parental BLNK-deficient cells. Remarkably, the introduction of wildtype MIST into BLNK-deficient DT40 cells partially restored the calcium response, whereas the response was not restored in BLNK-deficient cells expressing Y69F, Y96F, or YF2 (Fig. 2B). This result indicates that both Tyr-69 and Tyr-96 of MIST are required for the BCR-induced calcium response.
Because PLC␥2 is required for BCR-induced [Ca 2ϩ ] i increase in DT40 cells (22) and the activation of PLC␥2 is accompanied with its tyrosine phosphorylation (27), we next examined tyrosine phosphorylation of PLC␥2 upon BCR stimulation in BLNK-deficient DT40 cells expressing various forms of MIST. Tyrosine phosphorylation of PLC␥2 was restored in BLNKdeficient cells expressing wild-type MIST but not in transfectants expressing Y69F, Y96F, and YF2 mutants (Fig. 2C). BLNK has been reported to interact with PLC␥2 and thus recruit PLC␥2 to the membrane proximal compartment, where PLC␥2 is phosphorylated by Syk (13). To test whether MIST can play such a role, we assessed the interaction of MIST with PLC␥ by an in vitro binding assay using a GST fusion protein containing the C-terminal SH2 domain of PLC␥1 (GST-PLC␥- SH2). COS-7 cells were transfected with plasmids encoding wild-type or tyrosine mutants of MIST, in combination with Lyn, and cell lysates were precipitated with GST-PLC␥-SH2 fusion protein. As shown in Fig. 2D, GST-PLC␥-SH2 interacted with Lyn-phosphorylated wild-type MIST but not with unphosphorylated MIST. Of five single tyrosine mutants that were equivalently phosphorylated by Lyn, only Y96F mutant did not interact with GST-PLC␥-SH2. A MIST mutant (YF6) containing six tyrosine to phenylalanine substitutions at positions 69, 96, 101, 153, 174, and 188, which showed no detectable tyrosine phosphorylation by Lyn, also failed to associate with GST-PLC␥-SH2. These results indicate that MIST interacts with the SH2 domain of PLC␥ via its phosphorylated tyrosine 96, which may lead to the recruitment of PLC␥ to the site of PTKs for its activation and the subsequent [Ca 2ϩ ] i increase. It remains to be examined, however, how tyrosine 69 is involved in the PLC␥ activation. As Y69F mutation more severely affected BCRinduced tyrosine phosphorylation of MIST than Y96F, phosphotyrosine 69 may serve as a binding site for a PTK, which subsequently phosphorylates tyrosine 96. Alternatively, the PTK bound to the phosphotyrosine 69 may be responsible for the phosphorylation of PLC␥2.
BCR-induced activation of ERK2 has been shown to require both PLC␥2 and Ras pathways in DT40 cells (25), and reduced ERK2 activity in BLNK-deficient DT40 cells has principally been ascribed to a defect of the PLC␥2-mediated pathway (13). ERK activity after the BCR cross-linking was partially restored in BLNK-deficient DT40 cells expressing wild-type MIST. However Y69F, Y96F, and YF2 mutants failed to restore BCRinduced ERK2 activation in BLNK-deficient cells (Fig. 2E).
These results indicate that MIST can restore the PLC␥2-mediated pathway of BCR signaling in BLNK-deficient cells by molecular interactions through tyrosines 69 and 96. It is of note that the restoration of the PLC␥2 pathway by MIST is partial in any biochemical events examined here.
LAT Synergizes with MIST and Dispenses with the Requirement of Tyrosine 69 and 96 of MIST in Restoring BCR-induced NF-AT Activation in BLNK-deficient Cells-A possible explanation for the incomplete reconstitution of BCR signaling by MIST in the BLNK-deficient DT40 cells is that the B cells lack some signaling molecules required for full activity of MIST. One molecule that is known to be expressed in T cells and mast cells in which MIST is also expressed, but not in B cells, is LAT (28). LAT is a membrane-anchored adaptor protein that localizes to specific plasma membrane compartments known as GEMs (29) and has been demonstrated to be required for SLP-76-mediated NF-AT activation upon TCR stimulation (30,31). Because LAT is one of the signaling proteins associated with MIST in Fc⑀RI-signaling (18), LAT seems to be required for MIST to fully exhibit its function in BCR signaling. To assess this possibility, we transiently expressed various forms of MIST in combination with LAT in BLNK-deficient cells and examined BCR-mediated NF-AT activation by luciferase-reporter assay. As shown in Fig. 3, transfection of wild-type MIST or LAT alone showed weak restoration of the BCRinduced NF-AT activation as compared with chicken BLNK. Combined expression of wild-type MIST and LAT resulted in a marked synergy in reconstituting BCR-induced NF-AT activation in BLNK-deficient cells. Surprisingly, Y69F, Y96F, and YF2 mutants, which were not functional without LAT (data not shown), restored BCR-mediated NF-AT activation when coexpressed with LAT, to a similar level that was achieved by wild-type MIST with LAT (Fig. 3). Similarly, MIST co-expressed with LAT fully reconstituted BCR-mediated calcium response in BLNK-deficient cells, independently of the two tyrosines 69 and 96 (see below). These findings indicate that LAT cooperatively functions with MIST and dispenses with the requirement of two tyrosines 69 and 96 of MIST to restore BCR-mediated [Ca 2ϩ ] i increase and the following NF-AT activation in BLNK-deficient cells.
Differential Involvement of Three Structural Domains of MIST in LAT-mediated Reconstitution of BCR Signaling in BLNK-deficient Cells-To dissect the mechanism for the LAT-MIST cooperation in the immunoreceptor signaling, we generated BLNK-deficient DT40 cells stably expressing LAT together with various forms of MIST, including mutants with a deletion of PR regions (dPR1, dPR2, and dPR1/2), a mutant containing the nonfunctional SH2 domain (R335K), and the tyrosine mutants (YF2 and YF6) (Fig. 4A).
Introduction of LAT alone into BLNK-deficient DT40 cells partially restored the calcium response to a level similar to that observed in transfectants expressing wild-type MIST alone (Figs. 2B and 4B). Consistent with the restoration of BCRinduced NF-AT activation, BCR-induced [Ca 2ϩ ] i increase in cells expressing both LAT and wild-type MIST was restored to an equivalent level to that observed in wild-type DT40 cells (Figs. 2B and 4B). In the presence of LAT, YF2 restored the BCR-induced calcium response to a level comparable to that achieved by wild-type MIST. By contrast, YF6 and the SH2 domain mutant, as well as mutants lacking either of the two PR regions, rather weakly restored calcium response. The mutant lacking both PR regions failed to restore the response.
We next assessed BCR-induced activation of MAP kinases, ERK2 and JNK, in the BLNK-deficient DT40 cells reconstituted with LAT and the mutant forms of MIST using in vitro kinase assay. Wild-type DT40 cells exhibited sustained ERK2 activation until 10 min, with a peak response at 3 min following BCR stimulation, whereas in BLNK-deficient cells expressing either wild-type MIST or LAT alone, ERK2 activation was transient and decreased at 10 min (Figs. 2E and 4C). Coexpression of wild-type MIST and LAT restored both peak and sustained responses of BCR-induced ERK2 activation. The SH2 mutant, YF6, or dPR2 co-expressed with LAT showed the reconstitution activity comparable to the wild-type MIST. By contrast, dPR1 lacking the N-terminal PR region of MIST failed to augment ERK activation above the level that was achieved by LAT alone, suggesting that the N-terminal PR region of MIST is responsible for cooperation with LAT in BCR-induced ERK2 activation. As with the calcium response, the cells expressing dPR1/2 and LAT showed even weaker ERK2 response than the cells expressing LAT alone. The reason for this is currently unclear, but it is possible that the dPR1/2 lacking two PR regions behaves as dominant-negative mutant in BCR signaling through LAT.
Maximal JNK activation was observed 10 min after BCR stimulation in wild-type DT40 cells, whereas no BCR-induced JNK activation was detected in the absence of BLNK, as previously reported (13). Independent expression of either wildtype MIST or LAT in BLNK-deficient cells showed weak BCRinduced JNK activation, whereas co-expression of wild-type MIST and LAT resulted in a full restoration of BCR-induced JNK activation (Fig. 4D). YF6 and SH2 domain mutants partially restored BCR-induced JNK activation in the presence of LAT. Of two PR regions in MIST, deletion of the C-terminal PR region did not significantly affect the reconstituting activity of MIST in JNK activation, whereas the reduced restoration was observed in transfectants expressing a MIST mutant lacking the N-terminal PR region. The double PR mutant failed to reverse the defect in BCR-induced JNK activation in BLNKdeficient cells. These results indicate that tyrosine phosphorylation sites, the SH2 domain, and the N-terminal PR region of MIST are required for calcium response as well as JNK activation, whereas the C-terminal PR region is important for calcium response but not for JNK activation.
The Two PR Regions of MIST Independently Mediate the Interaction with PLC␥ and LAT-MIST-related adaptor SLP-76 has been shown to indirectly associate with LAT through the Grb2-related adaptor Gads in T cells (32)(33)(34). The observation of synergistic action of MIST and LAT prompted us to examine the interaction between MIST and LAT. Upon BCR stimulation, a tyrosine-phosphorylated 35-kDa protein, which we later identified as LAT by anti-LAT immunoblot, was coimmunoprecipitated with MIST from transfectants expressing both MIST and LAT, but not from transfectants expressing only MIST. Phosphorylated LAT was inducibly co-immunoprecipitated not only with wild-type MIST but also with YF2, YF6, or SH2 mutants (Fig. 5A). The amount of phosphorylated LAT that co-precipitated with the mutant lacking the N-terminal PR region was not altered, whereas that with the mutant lacking the C-terminal PR region was markedly reduced. Furthermore, no phosphorylated LAT was co-precipitated with the double PR mutant. The absence of LAT protein in MIST immunoprecipitates from cells expressing the double PR mutant was confirmed by anti-LAT immunoblots (Fig. 5B). These results indicate that the C-terminal PR region of MIST ( 178 PPAPP 182 ) is mainly involved in the interaction of MIST with LAT. For the interaction of SLP-76 and LAT, the SH3 and SH2 domains of Gads has been demonstrated to associate with the PR region of SLP-76 and phosphotyrosines of LAT, respectively (32). Because Gads is not expressed in DT40 cells, we examined possible involvement of Grb2 in the MIST-LAT association. As shown in Fig. 5C, in COS-7 cell transfectants, Grb2 was co-precipitated with wild-type MIST, but not with the  Fig. 2B. C, restoration of BCR-induced ERK2 activation by MIST and LAT. ERK2 activation was determined as described in Fig. 2E. In vitro kinase assay and the amount of the precipitated ERK2 protein were shown (top panel). Activation of ERK2 and the amount of ERK2 proteins were quantitated by densitometry and normalized ERK2 activity is represented graphically as the fold induction over the activity of unstimulated cells at 0 min (bottom panel). D, restoration of BCRinduced JNK activation by MIST and LAT. BLNK-deficient DT40 cells and the stable transfectants were stimulated with anti-IgM Ab for the indicated period, lysed, and immunoprecipitated with anti-JNK Ab. Half of the immunoprecipitate was used for the in vitro kinase assay, and the other half was subjected to immunoblotting with anti-JNK Ab (top panel). Graphs (bottom panel) represent the fold activation of JNK over the activity of unstimulated cells, as quantified by densitometry and normalized by the amounts of precipitated JNK (lower lanes of the top panel). Shown are representatives of two to three independent experiments using at least two different sets of clones. mutant lacking the C-terminal PR region, indicating that the PPAPP motif in MIST is a Grb2 binding site that may mediate the association with LAT. The residual association observed between dPR2 and LAT (Fig. 5A) might be mediated by Grap, another Grb2 family adaptor expressed in DT40 cells (25), which may differ from Grb2 in the binding specificity of their SH3 domains (35).
Given the evidence that the MIST mutant (dPR1) lacking the N-terminal PR region displayed a reduced reconstitution ability of BCR signaling in BLNK-deficient cells even though it can associate with LAT and Grb2, we anticipated that the N-terminal PR region of MIST may mediate the intermolecular interaction with a critical effector protein beside LAT. It has recently been reported that the PR region of SLP-76, which is distinct from the Gads/Grb2 binding site, mediates the interaction with the SH3 domain of PLC␥1 (36). Deletion of this region of SLP-76 does not abrogate inducible LAT binding to SLP-76, but it disrupts SLP-76 function in TCR-mediated activation of PLC␥1, ERK, and NF-AT. We therefore examined whether the N-terminal PR region of MIST, as similar to that of SLP-76, serves as a binding site for the SH3 domain of PLC␥ by in vitro pull-down assay using a GST fusion protein containing the SH3 domain of PLC␥1. As shown in Fig. 5D, the SH3 domain of PLC␥1 co-precipitated dPR2 as well as wild-type MIST, but not dPR1, in a tyrosine phosphorylation-independent manner, indicating that the N-terminal PR region ( 160 PLPPPR 165 ) of MIST is a binding site of the SH3 domain of PLC␥.
The PR Region-mediated Recruitment of MIST to GEMs in BLNK-deficient DT40 Cells Expressing LAT-Recruitment of immunoreceptors, as well as signaling proteins, to the GEMs has been recognized as an essential process in the initiation and propagation of the immunoreceptor signal (37,38). In T cells, as a GEM-resident adaptor, LAT recruits SLP-76, Gads, and PLC␥1 via its phosphotyrosines, and this LAT-mediated enrichment of signaling proteins in the GEM is essential for TCR-mediated signaling (29). The SLP-76 mutant lacking a Gads-binding PR region is unable to translocate to the GEMs and therefore can not participate in TCR-signaling (39). Similarly, BLNK has been demonstrated to translocate to the GEM fraction upon BCR cross-linking (24). To further gain insight into the function of the PR regions of MIST in immunoreceptor signaling, we examined whether the PR region-mediated intermolecular association of MIST is required for the translocation of MIST into the GEM. In BLNK-deficient cells expressing both LAT and wild-type MIST, dPR1, or dPR2, LAT was mainly detected in the Triton X-insoluble GEM fraction (fraction 3) (Fig. 6). In contrast, wild-type MIST and MIST mutant lacking either one of two PR regions were mainly localized in the Triton X-soluble low density fractions (fractions 10 and 11) in unstimulated cells (Fig. 6). Upon BCR stimulation, a significant amount of wild-type MIST, as well as MIST mutant lacking the N-terminal PR region, became detectable in the GEM fraction, whereas BCR-induced translocation to the GEM fraction was abrogated by the deletion of the C-terminal PR region (Fig. 6). Together with the inducible association of MIST with LAT through its C-terminal PR region, these results indicate that the inducible association of LAT with the C-terminal PR region of MIST is mainly responsible for the recruitment of MIST, likely through Grb2, to the GEM. DISCUSSION We demonstrate here by using PTK-and BLNK-deficient DT40 model system that MIST, like SLP-76 and BLNK, can couple immunoreceptor signals to the downstream biochemical events and that this function of MIST requires at least three intramolecular domains; i.e. tyrosine phosphorylation sites, PR regions, and an SH2 domain.
Although MIST is tyrosine-phosphorylated upon BCR crosslinking as rapidly as BLNK, the upstream PTKs required for the phosphorylation of MIST and BLNK differ. The absolute requirement of Syk for the tyrosine phosphorylation of BLNK has been reported using the Syk-deficient DT40 cell line (4). By contrast, tyrosine phosphorylation of MIST was attenuated but detectable in Syk-deficient DT40 cells upon BCR stimulation, suggesting that Syk is not essential for tyrosine phosphorylation of MIST. On the other hand, a more severe reduction of tyrosine phosphorylation of MIST was observed in Lyn-deficient DT40, which is consistent with our previous observation that Lyn, but not Syk alone, is capable of phosphorylating MIST in COS cell transfectants (18). Furthermore, our finding that double tyrosine to phenylalanine mutations at tyrosine 69 and 96 in MIST profoundly reduced BCR-induced tyrosine phosphorylation of MIST indicates that these two tyrosines, which are conserved between MIST and BLNK, are essential for BCR-mediated tyrosine phosphorylation of MIST. Although the sequences flanking these tyrosine residues are quite similar between MIST and BLNK, Lyn is mainly involved in phosphorylation of MIST, whereas Syk mediates BLNK phosphorylation. In this regard, a nonfunctional mutation of the SH2 domain of MIST markedly reduced the level of BCR-induced tyrosine phosphorylation (Fig. 5A), whereas an SH2 mutant of mouse BLNK was tyrosine-phosphorylated to an equivalent level as wild-type BLNK upon BCR cross-linking of WEHI231 mouse B cells. 2 These findings suggest that the SH2 domain of MIST, but not of BLNK, regulates accessibility to upstream PTKs.
Based on the finding that BLNK and SLP-76 recruit Tec family PTKs, Btk, and Itk, in addition to PLC␥ (40,41), an attractive scenario for the mechanism of BCR-and TCR-in-duced activation of PLC␥ has been proposed: phosphorylated BLNK/SLP-76 place Btk/Itk in close proximity to PLC␥, which allows full phosphorylation and activation of PLC␥ (42). According to this scenario, tyrosine phosphorylation of MIST may provide docking sites for PLC␥2 and possibly Tec family PTKs, and this interaction may result in the activation of PLC␥2 and subsequent [Ca 2ϩ ] i increase. This prediction is partly supported by our observation that the phosphorylation of tyrosines 69 and 96 of MIST is essential for MIST functions to restore PLC␥-mediated signaling in BLNK-deficient DT40 cells and that the tyrosine 96 of MIST is involved in the binding with the SH2 domain of PLC␥ in the context of the cells lacking LAT expression. On the other hand, phosphorylation of these tyrosines is dispensable for MIST function in the cells expressing LAT. This suggests that phosphorylated LAT interacts with the same or functionally equivalent signaling proteins as those interacting with phosphotyrosines 69 or 96 of MIST. In this regard, PLC␥ has been demonstrated to directly interact with LAT (43). Thus, LAT and MIST are equally capable of recruiting PLC␥ to the vicinity of PTKs and mediating its activation. Because BCR-induced calcium response was not completely restored by co-expression of phosphorylation-deficient MIST mutant (YF6) and LAT in BLNK-deficient cells (Fig. 4B), it is likely that a phosphotyrosine(s) other than 69 and 96 in MIST is required for the full activation of PLC␥. As mentioned above, one candidate that interacts with these tyrosines of MIST would be Btk, and this possibility is currently under investigation.
The interaction of SLP-76 and LAT has been demonstrated to be mediated by Grb2 family adaptor proteins, such as Gads, Grb2, and possibly Grap (43). The members of this family possess two SH3 domains flanking a SH2 domain. Upon TCR activation, Gads bridges SLP-76 and LAT through its SH3 domain and SH2 domain (32), and this trimolecular interaction has been demonstrated to be required for the translocation of SLP-76 to the GEM, where LAT is constitutively localized (29). We have previously demonstrated that MIST is also associated with Grb2 and LAT in RBL-2H3 mast cell line transfected with wild-type MIST (18). Of the two PR regions in MIST, the C-terminal PR region ( 178 PPAPP 182 ) appears to be mainly involved in Grb2 and LAT binding (Fig. 5, A and C) and the translocation to the GEM (Fig. 6). Despite its reduced ability to interact with LAT and to localize in the GEM, the MIST mutant lacking the C-terminal PR region can cooperate with LAT to restore BCR-induced ERK2 and JNK activation to the similar level to wild-type MIST. On the other hand, the MIST mutant lacking the N-terminal PR region ( 160 PLPPPR 165 ), which is fully competent for binding LAT and Grb2, nevertheless is inefficient to restore BCR-induced calcium response, as well as ERK2 and JNK activation. These apparently conflicting data may suggest the differential requirement of the physical interaction of MIST and LAT for the downstream responses to BCR stimulation. BCR-induced activation of ERK2 and JNK has been shown to require Ras and PKC pathways and Rac1 and calcium pathways in DT40 cells, respectively (25). Our data suggest the possibility that the signaling complex nucleated by the MIST-Grb2-LAT trimolecular interaction may not be necessary for BCR-induced ERK2 and JNK activation and that signals delivered by uncoupled MIST and LAT may be sufficient for Rac1 and PKC activation. Alternatively, the residual ability of the MIST mutant lacking the C-terminal PR region to interact with LAT (Fig. 5A) might be sufficient for the BCR-induced activation of ERK2 and JNK. The observation that no detectable GEM-associated dPR2 appeared upon BCR cross-linking (Fig. 6) might be explained by the weak interaction of dPR2 and LAT, which is lost under our lysis condition in the GEM fractionation. Although GEM-associated MIST-Grb2-LAT signaling complex formed through the C-terminal PR region of MIST seems critical for only calcium response as far as we examined, the intermolecular interaction of MIST through the N-terminal PR region, which is identified as a binding site of the SH3 domain of PLC␥, is necessary not only for optimal calcium response but also JNK and ERK2 activation, in collaboration with LAT. In this regard, Yablonski et al. (36) have recently reported that the PR region of SLP-76, which is distinct from the Gads/Grb2 binding site, mediates the interaction with the SH3 domain of PLC␥1 and that this interaction is essential for TCR-mediated activation of PLC␥1 and NF-AT. They also suggested from their observations that this SH3 domain-PR region-mediated interaction may induce the intramolecular conformational change of PLC ␥1, which is required for its enzymatic full activation (44,45). Thus, it is likely that such an interaction might induce the conformational alteration of PLC␥, which leads to the relief of its autoinhibitory status and subsequent activation. The complete abrogation of the reconstitution ability of MIST by the deletion of both PR regions also strongly indicates the absolute requirement of the multivalent interaction of the PR regions of MIST with LAT and PLC␥ for MIST function.
The SH2 domain mutant fully restored BCR-induced ERK2 activation and partially restored calcium response and JNK activation in the presence of LAT. In the absence of LAT, the SH2 domain mutant, but not the YF2 mutant, retained the ability to restore BCR-induced calcium response to a level equivalent to that achieved by wild-type MIST in BLNK-deficient cells (data not shown and Fig. 2B), suggesting that BCRinduced phosphorylation of tyrosines 69 and 96 in the SH2 domain mutant is intact, even though its gross phosphorylation is reduced (Fig. 5). Therefore, the observed functional defects of the SH2 domain mutant appear to result from its inability to interact with other signaling proteins through the SH2 domain. The SH2 domain of SLP-76 has been revealed to interact with SLAP-130/Fyb, an adaptor protein, the function of which in immunoreceptor signal transduction is still controversial (46,47). Although the association of MIST with SLAP-130/Fyb remains to be defined, it is likely that the SH2 domain-mediated intermolecular association is required for the full function of MIST in immunoreceptor signal transduction.
By using BLNK-deficient DT40 cells as an in vivo assay system, we demonstrated MIST functions through distinct intramolecular domains in immunoreceptor signaling depending on the availability of LAT. In cytokine-activated T cells, NK cells, and mast cells, in which MIST, SLP-76, and LAT are all endogenously expressed (18,19,48,49), MIST and SLP-76 may form distinct signaling complexes with or without LAT and mediate divergent signals. In SLP-76-deficient mice, it has been demonstrated that Fc⑀RI-mediated degranulation and cytokine production of mast cells were profoundly abrogated even in the presence of MIST (48), whereas NK cells from these mice exhibited normal NK receptor-mediated and Fc receptor-mediated cytotoxicities (49). Further studies using MIST-deficient mice will provide insights into distinct and overlapping functions between MIST and SLP-76 in immunoreceptor signaling in a physiological context.