Requirement for a Negative Charge at Threonine 60 of the FcRγ for Complete Activation of Syk*

Aggregation of FcεRI on mast cells results in the phosphorylation of the FcεRIγ chain on tyrosine and threonine residues within the immunoreceptor tyrosine-based activation motif. In the present study we sought to identify the site of threonine phosphorylation in FcεRIγ and investigate its functional importance. We found that threonine 60 was phosphorylated in vitro and in vivo. Expression of a mutated FcεRIγ (T60A), in either FcεRIγ-deficient or γ-null mast cells, resulted in a delay of FcεRI endocytosis, inhibition of TNF-α mRNA production, and inhibition of degranulation but did not affect FcεRI-induced cell adhesion. Tyrosine phosphorylation of the T60A mutant γ chain was normal, but Syk phosphorylation was dramatically reduced in these transfectants. This correlated with reduced co-immunoprecipitation of FcεRIγ with Syk. Substitution of an aspartic residue for threonine 60 of the FcεRIγ reconstituted complete activation of Syk and co-immunoprecipitation of FcεRIγ with Syk. We conclude that the negative charge provided by phosphorylation of threonine 60 of the FcεRIγ is required for the appropriate interaction and activation of Syk. This is a likely requirement for immunoreceptor tyrosine-based activation motifs involved in Syk activation.

The high affinity receptor for IgE, Fc⑀RI, 1 is an important initiator of type I allergic reactions (1) and is associated with the development of chronic disease because its expression is significantly increased in atopic individuals (2,3). Fc⑀RI is a tetramer consisting of ␣, ␤, and two disulfide-linked ␥ chains (4,5). The ␥ chain is required for Fc⑀RI signal transduction and is also involved in the signaling pathways of Fc␥ and T cell receptors (6,7). In addition, FcR␥ is required for cell surface expression of Fc⑀RI, and thus mast cells from ␥-null mice do not degranulate nor produce interleukin-4 to an IgE-mediated stimulus (8). Furthermore, the Fc receptor ␥ chain (FcR␥) is important to the pathological consequences of immune complex deposition and inflammation (9).
The importance of tyrosine phosphorylation of the FcR␥ to downstream signaling events is well established (10,11). After FcR aggregation, a Src family tyrosine kinase phosphorylates the tyrosines within the immunoreceptor tyrosine-based activation motif (ITAM) of the ␥ chain. Tyrosine phosphorylation of the ITAM tyrosines leads to recruitment of SH2-containing proteins which are involved in subsequent signaling steps (reviewed in Ref. 12). One such protein is Syk kinase (13) whose activity is critical to mast cell responses (14 -16).
In mast cells, tyrosine phosphorylation of both Fc⑀RI␤ and ␥ is observed upon Fc⑀RI engagement (10). In addition, the Fc⑀RI␤ is also phosphorylated on serine residues, whereas the Fc⑀RI␥ is phosphorylated on threonine residues (10). In a mast cell line (RBL-2H3), threonine phosphorylation of the Fc⑀RI␥ chain was demonstrated to be mediated by receptor-associated protein kinase C␦ (PKC␦) (17). The function of threonine phosphorylation of the Fc⑀RI␥ is not known but has been the subject of speculation. Bingham et al. (18) reported that threonine phosphorylation of the Fc⑀RI␥ preceded its tyrosine phosphorylation and that both threonine and tyrosine phosphorylation were delayed in a low secreting variant. This suggested the possible functional correlation of threonine phosphorylation with degranulation. In addition, we previously found a correlation between threonine phosphorylation and the rate of Fc⑀RI endocytosis as a specific PKC inhibitor inhibited both processes (17).
In the present study we sought to identify the site of threonine phosphorylation in the Fc⑀RI␥ and to assess the functional significance of this phosphorylation event by transfecting RBL-2H3-derived ␥-deficient cells (19) or ␥-null bone marrow-derived mast cells (BMMC) with either wild type or variant ␥ chains. Using this approach we found that threonine 60 was phosphorylated in vivo and that insertion of a negatively charged amino acid residue at position 60 of the Fc⑀RI␥ substituted for phosphorylation at this site, resulting in the Fc⑀RIdependent activation of Syk.

EXPERIMENTAL PROCEDURES
Immunoglobulins and Reagents-Anti-dinitrophenyl (DNP)-specific mouse monoclonal IgE (20) was purified as described (21). Dinitrophenylated human serum albumin (DNP 30 -40 -HSA) was from Sigma. The mouse monoclonal (4G10) antibody to phosphotyrosine conjugated to horseradish peroxidase and a polyclonal rabbit antibody to Shc were from Upstate Biotechnology Inc. (Lake Placid, NY). The mouse monoclonal antibody to PKC␦ was from Transduction Laboratories (Lexington, KY). The mouse monoclonal antibody to Syk was as described (22), and a rabbit polyclonal antibody to Syk was a kind gift from U. Blank (Institut Pasteur, Paris, France). A rabbit polyclonal antibody to FcR␥ was kindly provided by R. P. Siraganian (NIDCR, National Institutes of Health), and a chicken antibody to the FcR␥ was as described previously (16). FcR␥ peptides used in the in vitro studies were obtained from Quality Controlled Biochemicals (Hopkinton, MA) or were produced in-house (George Poy, NIDDK, National Institutes of Health) and analyzed by mass spectrometry and for amino acid composition by Harvard Microchemistry (Cambridge, MA). 3,3Ј-Dithiobis[sulfosuccinimidyl propionate] (DTSSP) was purchased from Pierce. Polyvinylidene difluoride membranes, Tricine and Tris-glycine sample buffers, Trisglycine, and Tricine SDS-polyacrylamide gels were from Novex (San Diego, CA).
Cells and Activation-RBL-2H3 and the RBL-2H3-derived ␥-cell lines were cultured as a monolayer in stationary flasks essentially as described (23,24). Stable transfectants were generated by electroporation of the Fc⑀RI␥ or mutants thereof as described previously (25). To obtain cell populations expressing approximately equal numbers of receptors, stable transfectants were labeled with FITC-IgE and sorted for 10% of the cells providing the highest fluorescent signal. Cloning of the stable transfectants was avoided to minimize potential clonal variation (26). Fc⑀RI␥-null BMMCs were grown in interleukin-3-containing medium essentially as described (27). Cells were used at the fourth to sixth week where cultures were greater than 90% mast cells. For degranulation assays, transfected BMMC were stimulated with 20 ng of DNP-HSA for 10 min at 37°C, and then the percent release of ␤-hexosaminidase was determined (21). For endocytosis and protein tyrosine phosphorylation, cells were stimulated with 100 -400 ng of DNP-HSA for the indicated times.
Site-directed Mutagenesis and Constructs-Individual mutations of the threonine residues at positions 52 and 60 to alanine or aspartic acid were by PCR using mismatched primers and Extend TM high fidelity polymerase (Roche Molecular Biochemicals). For the T52A mutation the rat Fc⑀RI␥ in pSVL was used as a template in a nested PCR using the internal 5Ј and 3Ј primers CTGAACGCCCGGAACCAGGAGACA and TGTCTCCTGGTTCCGGGCGTTCAG and the external 5Ј and 3Ј primers TGGGATCTCGAGATGATCCCAGCGGTGATCTTGTTC and ATACATACGCGTCTATTGGGGTGGTTTCTCATGTTTCAG that encoded XhoI and MluI restriction sites, respectively. For the T60A and T60D mutations of Fc⑀RI␥, the same 5Ј external primer shown above was used in combination with CATACGCGTCTATTGGGGTGGTTTCT-CATGTTTCAGAGCCT and CATACGCGTCTATTGGGGTGGTTTCT-CATGTTTCAGATCCTCATATGT, respectively, both of which encoded an MluI restriction site. The obtained PCR products were cloned in the XhoI/MluI sites of the mammalian expression vector ⑀MTH (28).
For competitive PCR detection of TNF-␣, we constructed a competitor template that contained a fragment of the Fc⑀RI␤ and the nucleotide sequence of the respective TNF-␣ primers at the 5Ј and 3Ј ends. The competitor was generated by PCR using the rat Fc⑀RI␤ in pSVL as the template and the following 5Ј and 3Ј primers: CAACTCGAGCAAGGA-GGAGAAGTTCCCAAACAGACTTTGACGACGAAGTGC and CTTAA-GCTTCGGACTCCGTGATGTCTAAGTCAGCACGGCACTGCAAAAG-GC. These primers encoded the following 5Ј and 3Ј sequence recognized by the TNF-␣ primers used for reverse transcriptase-PCR: CAAGGA-GGAGAAGTTCCCAA and CGGACTCCGTGATGTCTAAG.
Semliki Forest virus constructs were PCR generated for both wild type Fc⑀RI␥ and T60A mutated Fc⑀RI␥ using as templates the previously generated constructs in the ⑀MTH vector as described above and the following 5Ј and 3Ј primers: AAAATGCATGCCACCATGATCCCA-GCGGTG and AAAGGGCCCCTATTGGGGTGGTTTCTC. The primers encoded NsiI and ApaI restriction sites, respectively. The obtained PCR products were cloned into the NsiI and ApaI sites of the modified pSFV1 vector previously described (29). Fidelity of all constructs used in this study was determined by direct sequencing.
Assays for Ligand-Receptor Binding, Endocytosis, Cell Adhesion, and TNF-␣ mRNA Expression-Assays to determine cell surface expression of Fc⑀RI using 125 I-labeled IgE were essentially as described previously (30). All transfectants expressed between 2.0 and 3.0 ϫ 10 5 receptors/ cell. Endocytosis of Fc⑀RI was defined as amount of 125 I-labeled IgE not elutable by acid pH treatment of the cell surface and determined essentially as described (31,32).
Adhesion of cells to fibronectin was determined by measuring the IgE/Agn-stimulated binding to fibronectin-coated plates after loading cells with the fluorescent dye calcein AM as described previously (33,34). Fibronectin-coated plates (96-well SmartPlastic® plates precoated with ProNectin® F, Protein Polymer Technologies, San Diego, CA) were prepared by blocking with 200 l of 20 mg/ml bovine serum albumin for 2 h at 37°C and washing three times with phosphate-buffered saline. The IgE-sensitized, calcein-loaded cells were stimulated with 400 ng/ml DNP-HSA and immediately added to the washed plates in the presence or absence of the competitor peptide, RGD. Plates were incubated in the CO 2 incubator at 37°C for 60 min and processed essentially as described (34). Fluorescence was measured using a fluorescein filter set (absorbance maximum at 494 nm, emission at 517 nm) and a fluorescent enzyme-linked immunosorbent assay plate reader.
For TNF-␣ mRNA expression, Fc⑀RI␥ stably transfected and IgEsensitized cells were stimulated for 1 h with 0.1 ng of DNP-HSA. Total RNA was isolated and reverse transcribed as described previously (21). To determine TNF-␣ mRNA expression a semiquantitative competitive PCR was used. A known concentration (0.01 pg) of an internal competitive standard (described above), containing the TNF-␣ primer binding sites, was added to the PCR reaction. To maintain linearity in the PCR we used conditions in which the internal standard (348 base pairs), and TNF-␣ mRNA (501 base pairs) gave approximately equal amounts of product. Quantitation of negative images of agarose gels was by densitometry.
In Vitro Phosphorylation Assays-A peptide encoding the cytosolic region of the Fc⑀RI␥ was synthesized, and in addition other Fc⑀RI␥ peptides were synthesized in which either all of the threonines were converted to alanine or threonine-alanine substitutions were limited to defined positions. Protein kinase C␦ was prepared by P. Acs (NCI, National Institutes of Health) and used as described previously (17). After an in vitro kinase reaction using described conditions (17), phosphorylated peptide was isolated by reversed phase HPLC. We found that the background of the assay was significantly reduced if polyetheretherketone was substituted for stainless steel in the HPLC system tubing and column jacket. Radioactivity in the collected peaks was measured by Cerenkov counting, and relevant fractions were then subjected to phosphoamino acid analysis (17).
[ 32 P]Orthophosphate Labeling of Fc⑀RI␥, Phosphoamino Acid Analysis, and Peptide Maps -Cells were metabolically labeled with [ 32 P]orthophosphate essentially as described (10). Cells were then sensitized with anti-DNP-specific IgE at a concentration of 5 g/ml for 1 h. Radiolabeled, IgE-sensitized cells (2.1 ϫ 10 7 in 0.7 ml) were stimulated with 400 ng/ml DNP-HSA for 3 min at 37°C and processed to recover IgE-occupied Fc⑀RI as described previously (17). Recovered proteins were resolved, and phosphoamino acid analysis was as described previously (17).
Peptide maps were generated as described (35). Briefly, the identified Fc⑀RI␥ was excised from the polyvinylidene difluoride membrane and incubated with polyvinylpyrrolidone 360 kDa (PVP-360, Sigma) to block nonspecific binding of the protease. The excised membrane was then washed with 50 mM ammonium bicarbonate and subsequently incubated with 10 g of Glu-C V8 protease for 2 h at 37°C. This was followed by a second addition of Glu-C V8 protease and an overnight incubation at 37°C. The soluble peptides were recovered by lyophilization and were resuspended in electrophoresis buffer for two-dimensional peptide mapping as described (35).
Chemical Cross-linking, Immunoprecipitations, and Western Blots-Chemical cross-linking conditions, immunoprecipitations, and Western blots were previously described (36). In some experiments antibody to phosphotyrosine was used to determine Syk activation, PKC␦ phosphorylation (an indication of its translocation), and the presence of Fc⑀RI␥ under reducing conditions. In other experiments proteins were identified directly with their respective antibodies.
Transient Expression of Fc⑀RI␥ in BMMC from FcR␥ Null Mice and Degranulation Assay-Generation of recombinant Semliki Forest virus was described previously (29). For infection of BMMC, 4.0 to 7.0 ϫ 10 6 cells from FcR␥-null mice were infected with a 1:250 titer of Semliki Forest virus in the presence of 7.5% polyethylene glycol as described (29). After a 30-min incubation, the virus was removed, and cells were incubated for an additional 8 h. During this time one-tenth of the cells were separated and incubated with FITC-IgE (1 g), whereas the remaining cells were incubated with 5 g of unlabeled IgE. Cells were then washed with phosphate-buffered saline to remove unbound IgE and either analyzed by FACS (FITC-IgE labeled cells) or in a degranulation assay. Degranulation (␤-hexosaminidase release (21)) was measured in either 250-or 500-l volumes using 20 ng/ml of DNP-HSA for 10 min (maximal release) at 37°C in Tyrodes buffer (32).

RESULTS
Phosphorylation of Fc⑀RI␥ on Threonine 60 -To define the site of threonine phosphorylation of the Fc⑀RI␥, the in vitro studies took advantage of the previous observation that PKC␦ was uniquely able to phosphorylate the Fc⑀RI␥ (17). Purified peptides corresponding to the entire cytosolic region of the Fc⑀RI␥ (wild type or variants) were incubated with PKC␦ in an in vitro kinase reaction. After incubation, the phosphorylated peptides were isolated by HPLC. The results confirmed our previous study (17) that showed that PKC␦ phosphorylates the wild type Fc⑀RI␥ peptide mainly on threonine ( Fig. 1) with a minor incorporation of label to serine. Substitution of alanine for all threonines (T48A,T52A,T57A,T60A) in the Fc⑀RI␥ peptide resulted in a significant reduction in the overall extent of phosphorylation, and phosphoamino acid analysis indicated that only serine was phosphorylated (Fig. 1A). Fc⑀RI␥ peptides in which the threonine residues were sequentially added (T52A,T57A,T60A; T57A,T60A; and T60A) showed minimal to no threonine phosphorylation (Fig. 1, A and B). In contrast, the Fc⑀RI␥ peptide variant (T48A,T52A,T57A) that contained only threonine 60, the threonine closest to the C terminus, incorpo-rated similar amounts of radioloabel to that seen for the wild type peptide (Fig. 1B). Furthermore, the overall level of phosphorylation was far greater than that seen for the other threonine-to-alanine variants used in these experiments. In addition, as with the wild type peptide, the amount of threonine phosphorylation was far greater than the amount of serine phosphorylation. Thus, we conclude from these in vitro studies that the target of PKC␦ activity is the threonine closest to the C terminus (Thr 60 ) on the Fc⑀RI␥ chain.
Based on the aforementioned in vitro results, we limited our in vivo analysis to the mutation of T60A but also considered a T52A mutation that was proposed in a previous report by Pribluda et al. (37) as the possible site of threonine phosphorylation. Stable transfectants were generated, in the previously described RBL Fc⑀RI␥-deficient cell line (24), that expressed wild type, T52A, and T60A Fc⑀RI␥. As shown in Fig. 2A, the transfected cell populations expressed receptor numbers similar to those of RBL cells. As compared with the parental RBL ␥ cell line, receptor expression increased approximately 10-fold with all transfectants expressing between 2 and 3 ϫ 10 5 receptors/cell. Tyrosine phosphorylation of Fc⑀RI␥ was analyzed by immunoprecipitation of phosphorylated proteins with antibody to phosphotyrosine followed by immunoblotting with antibodies to Fc⑀RI␥ or to Shc (Fig. 2B). The Shc immunoblot served as a control for protein loading as Shc tyrosine phosphorylation is unaffected by Fc⑀RI aggregation if the cells were incubated in the presence of serum (38). Although no significant difference in the tyrosine phosphorylation of Fc⑀RI␥ wild type (␥WT) and the Fc⑀RI␥ T60A (␥T60A) was observed, a dramatic increase in the tyrosine phosphorylation of the Fc⑀RI␥ T52A (␥T52A) was found even in nonstimulated cells (Fig. 2B). Phosphorylation of Fc⑀RI␤ was not dramatically different among transfectants, although a 30% increase was observed in the Fc⑀RI␥ T52A (␥T52A) transfectant (data not shown). To determine what effect the mutations of the Fc⑀RI␥ had on threonine phosphorylation, we isolated the Fc⑀RI␥ from Fc⑀RI-activated transfectants that were labeled with [ 32 P]orthophosphate and phosphoamino acid analysis was performed. As shown in Table I, threonine phosphorylation of Fc⑀RI␥ was still present in the T52A mutant. In fact, an increase in the ratio of threonine to tyrosine was observed. In contrast, the T60A mutation resulted in a dramatic decrease of threonine phosphorylation, although low levels of threonine phosphorylation were detected consistent with the expression of low levels of wild type Fc⑀RI in the transfectants derived from ␥ cells ( Fig. 2A). Two-dimensional peptide mapping (using Glu-C V8 protease) of the ␥WT and ␥T60A showed a dramatic reduction in the presence of one radiolabeled peptide that co-migrated with the expected peptide standard (Fig. 2C). We did not expect, nor did we find, complete absence of this phosphorylated peptide, given that low levels of wild type Fc⑀RI␥ are present in these transfectants ( Fig. 2A). From the collective data we conclude that threonine 60, which is found in the Yϩ2 position of the Fc⑀RI␥ ITAM C-terminal YXXL motif, is the sole threonine modified by phosphorylation in these cells.
Effects of Fc⑀RI␥ T60A Mutation on Mast Cell Function-We previously reported that the specific PKC inhibitor Ro 31-7549 was able to inhibit both the Fc⑀RI-stimulated threonine phosphorylation of the Fc⑀RI␥ and the rate of Fc⑀RI endocytosis (17). We revisited this topic to determine whether the T60A mutation would mimic our previous results. Fig. 3A shows that mutation of T60A, but not T52A, had a significant effect on the rate of Fc⑀RI endocytosis. As much as 50% inhibition of endocytosis was found at early times (5 and 10 min) following Fc⑀RI aggregation. However, at later times (40 and 60 min) no significant difference in Fc⑀RI endocytosis was observed (Fig. 3A). These results were almost identical to those observed in our previous study where the use of the PKC inhibitor, Ro 31-7549, inhibited the initial rate of endocytosis by as much as 80% (17). In contrast, the Fc⑀RI endocytosis of the T52A mutant showed kinetics similar to that of the wild type transfectant, but the overall extent of endocytosis was increased when compared with the wild type transfectant. Because the ␥T52A has a greater level of tyrosine phosphorylation (Fig. 2B), these results are consistent with the importance of tyrosine phosphorylation to endocytosis of FcR␥-containing receptors (39).
Increased adhesion of RBL cells in response to Fc⑀RI stimulation has been reported (reviewed in Ref. 40). The increased adhesion can be demonstrated by binding of these cells to fibronectin-coated surfaces (41). We analyzed the adhesion based on an integrin-fibronectin interaction, because we could inhibit up to 95% of the Fc⑀RI-stimulated adhesion by the concurrent addition of the competing RGD peptide with the cells to the fibronectin-coated wells (data not shown). We utilized this approach to determine whether the T60A mutation would have an effect on Fc⑀RI-mediated adhesion. Fig. 3B shows the fibronectin-specific adhesion of Fc⑀RI-stimulated ␥-, ␥WT, ␥T52A, and ␥T60A cell lines. Results were normalized to the Fc⑀RI-stimulated response of the ␥ cells where little difference in adhesion was observed when compared with nonstimulated cells. Although reconstitution of the ␥ cells with Fc⑀RI␥ led to increased adhesion (ϳ1.5-fold), no significant difference was observed among the transfectants. This suggested that the Fc⑀RI-stimulated adhesion to fibronectin was unaffected by increased threonine and tyrosine phosphorylation (T52A) or decreased threonine phosphorylation (T60A) of the Fc⑀RI␥. Similar results were obtained when phorbol ester was used as a stimulus for adhesion (data not shown). Collectively, the results demonstrated that some Fc⑀RI-dependent events were unaffected by mutation of the threonines in Fc⑀RI␥.
We also analyzed the effect of the T60A mutation on the stimulation of TNF-␣ mRNA production (21,42). The results shown in Fig. 3C were normalized to the Fc⑀RI-stimulated response of the ␥ cells. Both the ␥WT and ␥T52A transfectants showed an additional increase of TNF-␣ mRNA levels of up to 50% following Fc⑀RI stimulation when compared with the ␥ cell response. In contrast, the ␥T60A transfectant showed no increase in response, and in most experiments TNF-␣ mRNA levels were below those of the ␥ cells (Fig. 3C). The reduced response was a consistent observation regardless of the concentration of antigen used for stimulation (data not shown). Thus, the threonine phosphorylation of the Fc⑀RI␥ is important to the Fc⑀RI-mediated induction of TNF-␣ mRNA.
We analyzed the effect of threonine mutation (T60A) of Fc⑀RI␥ on degranulation in BMMC derived from FcR␥-null mice because the stable transfectants, like the parental ␥ cells, secreted poorly, and prior studies showed that reconstitution with wild type Fc⑀RI␥ did not induce a degranulation response comparable to the RBL cell line (24). Table II shows that in four independent experiments, reconstitution of Fc⑀RI expression with either ␥WT or ␥T60A showed that expression of the latter inhibited degranulation (11-49%). Although all experiments showed inhibition, variable levels of secretion (15 to 38%) and FIG. 2. Expression, phosphorylation, and peptide map of wild type and mutant Fc⑀RI␥. A, quantitation of receptor expression of RBL, parental ␥ cell line, and wild type Fc⑀RI␥ (␥wt), Fc⑀RI␥T52A (␥T52A), and Fc⑀RI␥T60A (␥T60A) transfectants. Transfectants were FACS sorted and selected for the 10% of highest Fc⑀RI expressing cells with FITC-IgE. Recovered cells were grown in culture and receptor expression was quantitated using 125 I-labeled IgE as described (30). Data are from four individual experiments. B, tyrosine phosphorylation of wild type (␥WT) and mutant (␥T52A and ␥T60A) Fc⑀RI␥. IgE-sensitized cells (5 ϫ 10 6 ) were stimulated (ϩ) or not (Ϫ) with 100 ng of DNP-HSA (Agn) for 5 min. Cells were lysed and tyrosine phosphorylated proteins recovered by immunoprecipitation (IP) with antibody to phosphotyrosine (Anti-PY). Immunoblots were probed with a chicken antibody to Fc⑀RI␥ follow by stripping and sequential immunoblotting (IB) with antibody to Shc (p46 and p52). The latter was used as a control for loading because no change in the Fc⑀RI-induced tyrosine phosphorylation of Shc was seen in cells incubated in the presence of serum (38). C, Glu-C V8 protease peptide map of wild type (␥WT) and mutant (␥T60A) Fc⑀RI␥. Labeling of cells, protease digest, and peptide resolution were as described under "Experimental Procedures." A synthesized threonine phosphorylated standard peptide (pTLKHE) was added to the sample to determine the appropriate migration of the expected peptide.  (43), TNF-␣ production (44), and degranulation (15) have been demonstrated to be dependent on the activation of Syk, we analyzed Syk activation in Fc⑀RI␥ transfectants by measuring its tyrosine phosphorylation, because we previously demonstrated that tyrosine phosphorylation is an accurate measure of Syk activity (16). As shown in Fig. 4A, cells expressing the ␥T60A showed a significant defect in the activation of Syk. In contrast, both ␥WT and ␥T52A expressing cells effectively activated Syk with the latter showing increased Syk phosphorylation consistent with the increased levels of tyrosine phosphorylated receptors in this transfectant (Fig. 2B). In some experiments up to a 3-fold increase in Syk activation was observed in the ␥T52A expressing cells when compared with the wild type control, whereas a 60 -90% inhibition of Syk activation was observed in the ␥T60A transfectants (Figs. 4A and 5). However, the observed inhibition did not result from a generalized defect in Fc⑀RI signal transduction because, as shown in Fig. 4B, the tyrosine phosphorylation of PKC␦ in response to Fc⑀RI stimulation was unaffected by either the ␥T52A or ␥T60A mutations. Because tyrosine phosphorylation of PKC␦ requires its translocation to the plasma membrane (45,46), these results demonstrated that PKC␦ translocation is independent of threonine phosphorylation of Fc⑀RI␥. Thus, threonine phosphorylation of the Fc⑀RI␥ is primarily important for the activation of Syk and layed in the ␥T60A transfectant, but levels are enhanced in the ␥T52A transfectant. The endocytosis assay is outlined under "Experimental Procedures." The percentage of endocytosis is net. Data are from five experiments. B, adhesion assays are described under "Experimental Procedures." The stimulated increase in adhesion for the ␥-cell line was normalized to 1.0 to measure whether significant differences could be found among the transfectants by comparing all experiments. Although an approximate 50% increase in adhesion was found for all transfectants (␥wt, ␥T52A, and ␥T60A), no significant difference was observed among them. Data are from four individual experiments. C, TNF-␣ mRNA levels are reduced in the ␥T60A transfectants. Detection of TNF-␣ mRNA levels was described previously (21)   downstream effector functions.
We explored the possible mechanism by which the ␥T60A mutation inhibited the activation of Syk. Because Syk activation in mast cells was demonstrated to be dependent on its interaction with the activated Fc⑀RI (47,48), we investigated whether the ␥T52A or the ␥T60A mutations had any effect on the interaction of Fc⑀RI␥ with Syk in activated cells. In this series of experiments we immunodepleted Syk from chemically cross-linked (DTSSP) nonactivated and activated cell lysates and determined whether Fc⑀RI␥ co-immunoprecipitated by immunoblotting. This approach was used because after chemical cross-linking the antibody to Fc⑀RI␥ was not as efficient in immunoprecipitation as the antibody to Syk. As demonstrated in Fig. 5, immunoprecipitation of Syk from chemically crosslinked cell lysates showed that in cells expressing ␥T60A Syk activation was inhibited, whereas those expressing ␥T52A showed increased Syk activation. When immunoblots were probed for co-immunoprecipitated Fc⑀RI␥ its presence was detected following Fc⑀RI stimulation in both the ␥WT and ␥T52A transfectants with the latter showing greater levels of Fc⑀RI␥. In addition, in most experiments we could detect trace amounts of phosphorylated Syk and Fc⑀RI␥ in association with Syk in nonstimulated ␥T52A transfectants (Figs. 4A and 5). In contrast, only minimal amounts of Fc⑀RI␥ were found to co-immunoprecipitate with Syk isolated from the activated ␥T60A transfectants (Figs. 5 and 6A). Thus, the results suggested that the absence of threonine phosphorylation of the Fc⑀RI␥ in the ␥T60A transfectants inhibited Fc⑀RI interaction with Syk.
A Negative Charge at Fc⑀RI␥ Threonine 60 Is Required for High Affinity Syk Interaction, Efficient Syk Activation, and the TNF-␣ mRNA Response-We investigated the possible requirement for phosphorylation at threonine 60 in the interaction and activation of Syk. Sequence analysis of receptor ITAMs that interact with Syk revealed the presence of an amino acid residue, at the equivalent location in the ITAMs, that could be phosphorylated or was negatively charged (Fig. 6A). We postulated that a negative charge at threonine 60 would be the equivalent of phosphorylation of threonine and would result in complete activation of Syk in response to Fc⑀RI engagement. To test this hypothesis a ␥T60D transfectant was generated that contained a hybrid ITAM (YTGLX 7 -YEDL; throughout, dashes in sequences represent introduced gaps) with features of both the B cell receptor Ig␣/␤ (YEGLX 7 -YEDI) and the FcR␥ (YTGLX 7 -YETL). As shown in Fig. 6A, the ␥T60D transfectant showed normal levels of Syk activation, and the association of Fc⑀RI␥ with Syk was comparable with that of the ␥WT transfectant. In these chemical cross-linking experiments we used 6-fold more cells (3.0 ϫ 10 7 ) than in the experiments shown in Fig. 5. Thus, we could detect Fc⑀RI␥ associated with Syk in nonstimulated cells (Fig. 6A), and in darker exposures of the shown phosphotyrosine blot Syk phosphorylation was also detected (data not shown). These results clearly demonstrated that mutation of threonine 60 to a negatively charged aspartic acid, rather than to the more physiochemically related alanine, completely reconstituted Fc⑀RI␥ interaction with Syk leading to the activation of the latter. Thus, we would expect that the ␥T60D transfectant would show a TNF-␣ mRNA response similar to that of the ␥WT transfectant. Fig. 6B shows that the mutation T60D reconstituted the TNF-␣ mRNA response to levels comparable with that of wild type Fc⑀RI␥. This demonstrates the requirement of a negative charge at threonine 60 for high affinity interaction of Syk with the Fc⑀RI␥ ITAM and activation of Syk to levels that result in "normal" mast cell responses. DISCUSSION In this study, we present direct evidence identifying the site of threonine phosphorylation on the Fc⑀RI␥. Specifically, results from both our in vitro and in vivo experiments demonstrated that threonine 60 is the site of phosphorylation on the Fc⑀RI␥ and that phosphorylation at threonine 60 is important for complete activation of Syk. In addition, our findings provide evidence of a general mechanism for Syk activation because we found that a negative charge at the equivalent position of threonine 60 substitutes for threonine phosphorylation at this site and is in common to all ITAMs interacting with and activating Syk (Fig. 6A). This requirement is likely to extend to other cells like platelets, where collagen activation of Syk requires the complexing of glycoprotein VI (collagen receptor) with FcR␥ (49, 50). with 300 ng of DNP-HSA (Agn). Cells were lysed and Syk was immunoprecipitated (IP) with antibody to Syk (Anti-Syk). Immunoblotting (IB) of resolved proteins with antibody to phosphotyrosine (PY) was followed by stripping and reprobing with antibody to Syk (Syk). B, cells treated as in A were lysed, and PKC␦ was immunoprecipitated (IP) with antibody to PKC␦ (Anti-PKC␦). Immunoblotting (IB) of resolved proteins with antibody to phosphotyrosine (PY) was followed by stripping and reprobing with antibody to PKC␦ (PKC␦). Tyrosine phosphorylation of PKC␦ in RBL-2H3 cells was previously shown to require its translocation (45,46).
Pribluda and colleagues (37) used HPLC and peptide standards to identify products of protease reactions with in vivo labeled Fc⑀RI␥. The evidence presented suggested that threonine 52 was a site of phosphorylation. In our hands, mutation of threonine 52 to alanine resulted in a "gain of function" because more Fc⑀RI␥ was tyrosine phosphorylated, and high levels of Fc⑀RI␥ phosphorylation occurred in nonstimulated cells. This resulted in increased endocytosis (Fig. 3A), and in several experiments higher levels of nonstimulated degranulation were observed (data not shown). Furthermore, an increased ratio of threonine to tyrosine phosphorylation was observed in the ␥T52A transfectants (Table I), clearly demonstrating that a site for threonine phosphorylation was still present. A possible explanation for the observed increased phosphorylation is that a conformational change of the Fc⑀RI␥ is induced by the T52A mutation that allows receptor-associated Lyn and PKC␦ to phosphorylate both cis-and trans-molecularly (51). In contrast, mutation of threonine 60 led to a "loss of function" and ablated threonine phosphorylation with no effect on tyrosine phosphorylation. The importance of in vivo phosphorylation of threonine 60 was demonstrated by the finding that substitution of a negatively charged amino acid provided full functionality in Syk activation.
It has been reported that ITAMs are not functionally redundant and that specificity of binding to SH2 domains is based on the residues surrounding the tyrosines (52). Furthermore, tandem SH2 domains showed an increased binding specificity to their biologically relevant ITAM such that another ITAM can bind a tandem SH2 domain but with up to 10,000-fold weaker affinity (53). Because the linker region (where threonine 52 resides) between paired YXXL motifs does not significantly contribute to the binding affinity of the Syk SH2 domain tandem with the Fc⑀RI␥ ITAM (54), this suggested that residues in the paired YXXL motifs might contribute to the high affinity interaction and specificity. Binding of both Syk SH2 domains to the Fc⑀RI␥ ITAM is required for high affinity interaction (54,55), and the N-terminal SH2 domain of Syk binds the ␥ITAM C-terminal phosphotyrosine (54). Our findings that the threonine 60 localized in the C-terminal YETL motif was solely phosphorylated in response to Fc⑀RI stimulation prompted us to look at the effects of its mutation (T60A) on Syk activation and downstream events. The loss of Syk binding with the Fc⑀RI␥T60A ITAM and its recovery with the Fc⑀RI␥T60D ITAM demonstrated the importance of threonine phosphorylation in the Fc⑀RI␥ ITAM for high affinity interaction. Because the T60D mutation (YED 60 L) mimics the dual negative charge that would result from phosphorylation of Thr 60 in the YETL motif, this suggests that a dual negative charge is important for the high affinity interaction of Syk. This dual negative charge pre-exists in the equivalent motif of the B cell receptor Ig␣/␤ (YEDI).
The crystal structures of both the C-terminal Syk SH2 by itself and the tandem SH2 domains complexed with phosphotyrosine ITAM peptides are available (56,57). However, the information provided by resolving the N-terminal SH2 domain interaction with a C-terminal phosphotyrosine of the CD3⑀ ITAM does not likely reflect the interaction of Syk with the B cell receptor Ig␣/␤ or with Fc⑀RI␥ ITAMs, because binding of the latter two occurs with at least a log fold higher affinity (54,58). Nevertheless, some insights are gained from these studies with regard to the interaction of a FcR␥ ITAM with the Nterminal SH2 domain of Syk. First, it appears that the Cterminal Yϩ2 amino acid residue (the threonine equivalent, but a glycine residue in the CD3⑀ ITAM) does not make significant contacts with Syk. For the Fc⑀RI␥ ITAM, a phosphorylated threonine would likely provide contact with a highly con-FIG. 6. Mutation of Fc⑀RI␥ threonine 60 to aspartic acid reconstitutes Syk interaction, Syk activation, and the TNF-␣ mRNA response. A, sequence alignment of ITAMs that interact with Syk (dashes represent introduced gaps), introduced mutations, and reconstitution of Syk activation and interaction with Fc⑀RI␥ in the ␥T60D mutant. IgE-sensitized Fc⑀RI␥WT, ␥T60A, and ␥T60D transfectants (3.0 ϫ 10 7 cells) were stimulated (ϩ) or not (Ϫ) with 300 ng of DNP-HSA (Agn). Cells were lysed, and postnuclear lysates were treated with DTSSP. Syk was immunoprecipitated (IP: Anti-Syk), and resolved proteins were identified as Syk and Fc⑀RI␥ by immunoblotting (IB). Because of the cell numbers used, we could detect the presence of Fc⑀RI␥ with Syk in nonstimulated cells. As compared with Fig. 5, only a short exposure to film was required for detection. PY, phosphotyrosine. B, Fc⑀RI␥T60D reconstitutes the TNF-␣ mRNA response to the level of wild type Fc⑀RI␥. TNF-␣ mRNA assay is described under "Experimental Procedures." The PCR amplified products were resolved on agarose gels and normalized to the internal standard (see "Experimental Procedures"). Quantitation was by densitometry. The TNF-␣ mRNA response of the Fc⑀RI␥T60A was normalized to 1.0 to allow comparison between experiments. Data are from two experiments. served histidine that is present in both the N-and C-terminal SH2 domains of Syk and ZAP-70. In addition, provided that the interactions of the Fc⑀RI␥ ITAM with Syk might result in a slightly different conformation than that of the CD3⑀ ITAM, a highly conserved arginine residue (which is also found in the N-terminal SH2 domain of ZAP-70) could also provide potential charge interactions. Thus, these possible interactions may fulfill the requirements for a high affinity interaction. In vitro, threonine phosphorylation does not appear to be a requirement for Syk activation because tyrosine biphosphorylated ITAMs can activate Syk and in vitro measurement of affinities demonstrated a high affinity interaction (54,55). We also found (in vitro experiments) that the introduction of a phosphorylated threonine in the tyrosine biphosphorylated Fc⑀RI␥ ITAM peptide had no effect on the affinity of interaction with the Syk SH2 domain tandem (data not shown). Although at the moment we cannot explain the difference in requirements for in vivo versus in vitro activation of Syk, it is interesting to note that the mutation of T52A also did not affect the affinity of interaction of the tyrosine biphosphorylated ITAM with Syk. However, in vivo this mutation results in a gain of function (including phosphorylation of the Fc⑀RI␥ in the absence of its aggregation) presumably from a conformational change that increases phosphorylation of Fc⑀RI␥ and thus Syk interaction (Fig. 5). Although to date there is no direct evidence for an in vivo structural conformation of the Fc⑀RI␥ cytoplasmic domain, this possibility is presently raised and has been suggested in a prior study (59).
Of physiological interest is the finding that even small amounts of activated Syk are sufficient to drive low levels of mast cell responses. In particular, degranulation was not ablated in the ␥T60A transfectant although it was reduced. Furthermore, a reduced induction of TNF-␣ mRNA could still be detected. This suggests that the level of activated Syk required to promulgate downstream signals is small, and thus engagement of a few receptors may likely be sufficient for some cytokine production and degranulation. Interestingly, it has been reported (60) that aggregation of Fc⑀RI with a low affinity ligand causes the complete tyrosine phosphorylation of the receptor but only weakly activates Syk and downstream effectors. Thus, one might speculate that receptor aggregation by high affinity ligands induces additional events (such as threonine phosphorylation of Fc⑀RI␥) required for complete Syk activation and engagement of downstream effectors. Engagement of downstream effectors may also require the movement of receptor and associated components to glycosphingolipidenriched microdomains (61), 2 a process that may not occur as efficiently with low affinity ligands. 3 Finally, the present study provides a possible explanation of why the Fc⑀RI␤ functions as an amplifier of Fc⑀RI-mediated responses (62). Because PKC␦ binds to the Fc⑀RI␤ ITAM (17), absence of the Fc⑀RI␤ may ablate threonine phosphorylation of the Fc⑀RI␥, resulting in a low affinity interaction with and weak activation of Syk. Tyrosine phosphorylation of Fc⑀RI (␣␥ 2 ) receptors in the absence of the Fc⑀RI␤ and constitutively associated Lyn (63) may still occur because of the movement of these receptors to the Lyn-containing glycosphingolipid-enriched microdomains (61) where they can be phosphorylated. The present study demonstrates that maximal mast cell responses are dependent on the phosphorylation of threonine 60 of the Fc⑀RI␥ ITAM, which mediates a high affinity interaction with Syk resulting in greater amounts of active Syk.