Molecular dissection of the FcR b signaling amplifier

e RI aggregation. These findings highlight the critical role played by FcR b in cell activation after Fc e RI aggregation. Phosphorylation of the g ITAM by lyn

4 line RBL-2H3 (9). In contrast, FcRb lacks autonomous signaling capacity (11). Instead, it acts as an amplifier of signals transduced by FcRg (12,13). This was an unexpected finding, the molecular basis of which has not been elucidated so far. The b ITAM differs from the canonical ITAM in 2 ways: 1) it contains a third tyrosine between the canonical tyrosines; and 2) it has a shorter spacing between the canonical tyrosine residues. Whether these features are responsible for the unique functional characteristics of b has not been determined. Three additional features of FcRb further explain why it has been the focus of many studies since its cloning. Firstly, it possesses the unique capacity of amplifying receptor expression, as evidenced by the higher level of expression of abg 2 compared to ag 2 (14). This expression amplification results from the capacity of FcRb to promote intracellular processing of the a chain. Secondly, genetics studies have identified the gene for FcRb as a candidate gene for atopy, even though the mechanism that underlies this effect is unknown (15). Thirdly, the two FceRI isoforms, ag 2 and abg 2 , differ in their expression pattern and their functions. While abg 2 is expressed on effector cells of the allergic response, mast cells and basophiles, ag 2 is expressed on antigen presenting cells in humans (rodents do not express ag 2 ), where it participates in antigen presentation (16,17).
Protein tyrosine kinases of the src family are responsible for phosphorylating the tyrosines in the ITAMs of the antigen receptors. In the case of FceRI this kinase is lyn (18,19). Lyn is activated upon FceRI aggregation. Lyn is bound to b in resting cells by a mechanism that has not been identified. This association increases after FceRI aggregation (11,(20)(21)(22)(23)(24). It could be mediated by binding of the lyn SH2 to the tyrosine phosphorylated b, but this is debated (11,(20)(21)(22)(23)(24). Lyn is responsible for phosphorylating the b and g chains (19), and lyn activation is the earliest signal detected upon FceRI aggregation. These findings highlight the critical role played by FcRb in cell activation after FceRI aggregation. Phosphorylation of the g ITAM by lyn by guest on March 23, 2020 http://www.jbc.org/ Downloaded from creates binding sites for the two SH2 domains of syk, a protein tyrosine kinase of the syk/ZAP-70 family (19,25). Syk binding to FcRg leads to lyn-dependent tyrosine phosphorylation and activation of the kinase. This step enables the productive interaction of active Syk with its many targets (26). These targets include various adaptors and enzymes (reviewed in (27,28)). One of these is phospholipase Cg1, which catalyses the breakdown of membrane phospholipids to generate two second messengers, inositol-1,4,5-trisphosphate (Ins(1,4,5)P3) and diacylglycerol (DAG). These signaling molecules are responsible for calcium (Ca 2+ ) release from intracellular stores and activation of various protein kinase C isoforms, respectively.
To further our understanding of how b acts as a signal amplifier we investigated the role played by each of the three tyrosines of the b ITAM in controlling early signals after FceRI aggregation. We generated mutant human receptors in which each tyrosine individually or in combination were replaced with phenylalanine, and compared the signaling capacities of these mutated receptors to that of abg 2 and a g 2 . We found that the N terminal tyrosine plays a predominant role in controlling early signals after FceRI aggregation. Our data also suggest a possible inhibitory role for the middle tyrosine on Ca 2+ release.

Cell Culture
The human monocytic cell line U937 and its transfectants were maintained in RPMI 1640 as described in (11). Transfectants were selected in the presence of 0.6 mg/ml G418 (Cellgro/ Mediatech, Inc., Herndon, VA). -sense GTA AGT AGC TGA ATA TAT GTT TAA TTC TTC AAA AAC ACG ATC   CTC TGG AAC CTT GTT TCC. To mutate Y225, sense GAG GAT CGT GTT TAT GAA   GAA TTA AAC ATA TTT TCA GCT ACT TAC AGT GAG TTG GAA GAC and antisense   GTC TTC CAA CTC ACT GTA AGT AGC TGA AAA TAT GTT TAA TTC TTC ATA AAC   ACG ATC CTC. To mutated Y229, sense TTA AAC ATA TAT TCA GCT ACT TTC  Transfectants were generated with the human a and g cDNAs subcloned into the pCDLSRa296 vector and the V5 epitope-tagged wild type (WT) or mutant human b cDNAs subcloned in the pBJ1neo vector as described in (10).

Assessment of surface FceRI and total b expression by flow cytometry
To detect surface FceRI expression cells (10 6 cells/sample) were stained with 1mg of biotinylated anti-NP human IgE, followed by PE-SA (1:200 dilution), and analyzed in a FACSCalibur flow cytometer (Becton Dickinson, Franklin Lakes, NJ). To assess the level of total b expressed, cells (10 6 cells/sample) were fixed with 1.8% paraformaldehyde in PBS, permeabilized in 0.1% saponin in PBS containing 1% BSA, then stained with 0.5mg anti-V5 Ab or mouse IgG2a (as isotype control), followed by PE-conjugated goat anti-mouse Ab.
Untransfected U937 cells were included in all experiments as negative control.

Assessment of surface FceRI subunit composition
In order to rule out a possible effect of the mutations in b on receptor assembly, the subunit composition of surface receptors was assessed. The various transfectants were loaded with the anti-a mAb 15-1, washed and lysed in mild conditions (0.5% Triton X-100, 50mM Tris-HCl (pH 7.6), 150mM NaCl, 5mM EDTA, 5mM NaF, 5mM Na 4 P 2 O 7 , 1mM sodium vanadate, 10mg/ml pepstatin, 4mg/ml leupeptin, 10mg/ml aprotinin and 1mM phenylmethylsulfonyl fluoride) in order to maintain subunit association. The lysates were placed on ice for 10 min, then centrifuged at 14,000 rpm for 15 min at 4 o C. 15-1 bound receptors were precipitated with protein G Sepharose. After a 1-hour incubation at 4 o C, the beads were washed in lysis buffer and the proteins eluted in SDS-PAGE sample buffer. The proteins were resolved by SDS-PAGE, transferred to Immobilon-P membrane (Millipore Corp., Billerica, MA) and blotted with anti-V5 (for b) and anti-g Abs. Immunoreactive proteins were visualized using the ECL system (Amersham Biosciences, Piscataway, NJ). Films were scanned using the Personal Molecular Imager (BioRad, Hercules, CA) in order to obtain quantitative data (in arbitrary units) for bands on western blots. Background counts present in windows of the same size as the ones used for the bands were subtracted to obtain specific counts.

Cell activation by antigen and lysis
U937 transfectants (10 7 cells/ml per sample) were incubated with 10 mg/ml anti-NP human IgE for 1 hour at 37 o C with rocking. Cells were washed and resuspended at 2.5 x 10 6 cells/ml in calcium (Ca 2+ ) buffer (pH 7.5) containing 135mM NaCl, 5mM KCl, 1mM MgCl 2 , 1mM CaCl 2 , 5.6mM glucose, 10mM HEPES (pH 7.4) and 0.1% BSA, and triggered with 25 ng/ml NP-BSA for 2 min at 37 o C. Near the end of stimulation, cells were pelleted by brief centrifugation and lysed at 3 x 10 7 cells per ml in the same lysis buffer as above.

Immunoprecipitation, SDS PAGE, Western transfer and immunoblotting
Immunoprecipitation was performed with the Abs indicated in each case. Samples were processed as described above and blotted with the indicated Ab. In some cases blots were stripped according to the instructions provided by the manufacturer of the ECL kit and reprobed with a different Ab.

Measurement of Ca 2+ mobilization
Transfected U937 cells were incubated with anti-NP IgE (10mg per 10 7 cells/ml at 37 o C for 1 hour) and loaded with 2mM fura-2 AM and 0.2mg/ml Pluronic F-127 in Ca 2+ buffer at 37 o C for 30 min. Samples (5 x 10 6 cells in 2 ml of Ca 2+ buffer) were analyzed in a cuvette-based Deltascan spectrofluorometer (Photon Technology International Inc., South Brunswick, NJ). Intracellular Ca 2+ concentrations were recorded as the ratio of Fura-2 emission when excited at 340 and 380 nm.

Mutation of the N terminal tyrosine mimics the phenotype of ag 2 receptors on aggregationinduced receptor phosphorylation
In order to study the tyrosines of the b chain we took advantage of a reconstitution system that has been extensively characterized in our laboratory and was used to establish that the b chain is an amplifier (11,13,14). Stable transfectants were generated in the U937 monocytoid cell line, which does not express the a and b chains, using the human a and g cDNAs alone or in conjunction with the wild type (WT) or mutated b cDNAs (used as V5 tagged) (Fig. 1A). FceRI surface expression was measured after staining with biotinylated human IgE and SAPE. For each mutation clones expressing similar FceRI density at the cell surface were selected for further analysis. The flow cytometric profile of some of these clones is shown in Fig. 1B. Human FceRI is expressed as two isoforms, abg 2 and ag 2 , which differ in their expression level and signaling capacity. Therefore, it is necessary to verify that mutations of the tyrosines do not affect either expression of b or subunit association, which could result in the expression of ag 2 trimers along with abg 2 tetramers at the cell surface, and would confer a lower signaling capacity upon these cells independently of a direct effect of the mutations in b. Note that it is unlikely that the receptors expressed at the cell surface on some mutants contain a significant proportion of ag 2 complexes since we selected clones with the same level of FceRI surface expression. If a significant proportion of ag 2 complexes was present at the cell surface in certain mutants, the mean expression level of the clones of these mutants would be reduced compared to that of WT transfectants as a result of the loss of the expression amplification effect of b (14). We did not observed such an effect. We first verified that among the clones selected for analysis the mutants did not express less b than the abg 2 clones by staining permeabilized cells with anti-V5 directed against the V5 tag appended to the N terminus intracellular tail of b ( Fig. 1C). We then assessed the subunit composition of membrane receptors. Cells were saturated with the anti-a mAb 15-1, washed and lysed. The membrane receptors were immunoprecipitated with protein G Sepharose, run on SDS-PAGE, and blotted with anti-V5 (b) and anti-g (Fig. 1D). Although there was some clonal variability in the relative amounts of b and g coprecipitated with a (Fig. 1D, compare clones 1 and 2 of YFY and clones 1 and 2 of FFY), no consistent differences were observed in the capacity of particular b mutants to interact with a and g in vivo, except possibly for FFF. This indicates that there is no systematic effect of the mutations in b on receptor assembly.
FceRI aggregation induces phosphorylation of b and g on the tyrosine residues in the ITAMs. The role played by individual tyrosines in FcRb has not been defined. In RBL cells strongly stimulated via the tetrameric FceRI the C terminal tyrosine was found to be significantly more phosphorylated than the N terminal tyrosine, while phosphorylation of the middle tyrosine was the lowest (33). Whether this difference in phosphorylation intensity of the three tyrosines is relevant for the b amplification function has not been determined. We compared aggregation induced b and g tyrosine phosphorylation among abg 2 , a g 2 and b mutant transfectants. Cells were loaded with anti-NP IgE and triggered with NP-BSA for 2 min. They were lysed in conditions (Triton X-100 0.5%) that maintain subunit association, and subjected to immunoprecipitation with the anti-g Ab 934 and blotting with the anti-PY Ab 4G10 ( Fig. 2A and data not shown from 2 additional experiments). The intensity of the b and g bands on the anti-PY and anti-V5 or anti-g blots were quantified by scanning. Although we observed some clonal variability in the amount of b coprecipitated with g, this is not due to intrinsic differences in the capacity of specific mutants of b to interact with a and g in vivo, as we showed in Figure 1D. To normalize the variable amounts of coprecipitated b, the specific values calculated from the phosphorylated b bands (in arbitrary units) were divided by the ratio of total b to total g for the same clone calculated from the anti-b (V5) and anti-g blots. The values were also normalized to the same ratio for the abg 2 clones. For phosphorylated g the values were simply normalized to the value for total g, and then to the same value for the abg 2 clone (Fig. 2A). As expected given the amplifier role of b, b and g phosphorylation was strong in the abg 2 transfectants, and g phosphorylation was substantially lower in the ag 2 transfectants (12). Please, note that in the with anti-V5 tag Ab (Fig. 2B). All together these results show that the N terminal tyrosine is the most phosphorylated of the three in intact cells, and that its mutation abolishes the b amplifier function on g phosphorylation. These results suggested that Y219 may play a predominant role in b function. Y225 and Y229 may contribute to this function.

The Y219F mutation destabilizes the association of lyn with b
Lyn of the src family of protein tyrosine kinases has been shown to be the major kinase responsible for FceRI tyrosine phosphorylation. Lyn is associated with FceRI in resting conditions and this association increases after FceRI triggering. The mode of interaction between b and lyn has not been elucidated completely. We assessed a possible effect of the b ITAM mutations on lyn association with b by comparing the amount of lyn coprecipitated with b in the various transfectants after FceRI triggering. Cell lysates were immunoprecipitated with anti-V5 to precipitate b and blotted with anti-lyn ( Fig. 3 and data not shown from 2 additional experiments). Whereas b-associated lyn was easily detected in abg 2 transfectants, it was substantially decreased or abolished when Y219 was mutated (FYY, FFY, FYF, FFF), even though reblotting with anti-V5 shows that the amount of immunoprecipitated b in these mutants was equivalent to, or higher than, that in the abg 2 transfectants. The amount of lyn coprecipitated was normalized to the amount of b precipitated and quantified as a ratio between the densities of the lyn and b band (Fig. 3). The ratio in the FYY transfectant was 11% of the ratio in the abg 2 transfectant, and 0-22% in FFY, FYF and FFF. In the mutants affecting Y225 and Y229 the ratios were moderately decreased or increased. In particular the double mutation that leaves Y219 intact, YFF, resulted only in a minor decrease in lyn association (ratio = 0.83) when compared to the double mutations that affect Y219 (FFY and FYF). These results favor a predominant role for Y219 in lyn association with b, and suggest that the effect of the Y219 mutation on b and g phosphorylation is mainly due to its capacity to decrease lyn association with b.

The Y219F mutation decreases FceRI-aggregation induced syk phosphorylation
Phosphorylation of the g ITAM creates a binding site for syk, which is recruited to the aggregated receptors, and phosphorylated and activated by lyn and by itself. We assessed the effects of the b mutations on syk phosphorylation after FceRI aggregation. Syk was immunoprecipitated with an anti-syk and blotted with an anti-phosphotyrosine, or anti-syk as a control for the amount of immunoprecipitated syk. The ratio of phosphorylated syk to total syk was calculated for each clone ( Fig. 4 and data not shown from 2 additional experiments). Syk phosphorylation was decreased in the FYY transfectant compared to the abg 2 transfectant (ratio to abg 2 = 0.56). This was similar to the ratio seen with ag 2 (0.53). It was also decreased in FFY, Although less pronounced than the differences observed in receptor subunit phosphorylation and lyn association, the differences observed between the different b mutants in syk phosphorylation also support the conclusion that Y219 is the tyrosine principally responsible for b amplification function.

The Y219F mutation reduces and delays aggregation-induced Ca 2+ mobilization
A critical signal downstream of lyn and syk is Ca 2+ mobilization. We assessed the effect of the b mutations on FceRI aggregation induced rise in intracellular Ca 2+ . Cells were loaded with fura-Ca 2+ traces obtained at various concentrations of antigen for one representative clone of each type. As expected, a more rapid and intense Ca 2+ rise was observed in abg 2 than in ag 2 as a result of the amplifier effect of b. This difference was more pronounced at low doses of triggering. The Ca 2+ flux in abg 2 was as intense and as rapid at 10 ng/mL as at 25-100 ng/mL of NP-BSA, whereas, in ag 2 the Ca 2+ flux at 10 ng/mL was lower and more delayed than at 25 ng/mL. When Y219 was mutated alone or with one or two other tyrosines (FYY, FFY, FYF and FFF) the Ca 2+ response was delayed and of lower amplitude compared to abg 2 . In FYY, FFY and FFF the Ca 2+ profiles, both in terms of intensity and kinetics, resembled that in ag 2 , while in FYF, the amplitude was lower and the delay greater than in ag 2 . In contrast the single mutants with intact Y219 (YFY, YYF and YFF) had profiles closer to that of abg 2 than that of ag 2 .
In order to fully assess the impact of the b mutations on Ca 2+ mobilization we repeated the analysis using multiple clones of each type (5-11 clones per type) and a concentration of 50 ng/mL of antigen. There was no statistically significant difference in FceRI expression levels among the various clone types (data not shown). Two parameters were used to describe the Ca 2+ responses: amplitude was measured as the difference between baseline and the top of the peak (Fig. 5B); delay was measured from the time when antigen was added to the time when the top of the peak was reached (Fig. 5C). This analysis confirmed that the Y219 mutation alone or in combination (FYY, FFY and FFF) resulted in a reduced Ca 2+ peak that was similar to that observed in ag 2 clones (p>0.05 for all). In contrast, mutating the other tyrosines (YFY, YYF, YFF) did not affect peak height significantly compared to the abg 2 clones (p>0.05 for all).
Analysis of the delay in Ca 2+ mobilization provided a more subtle picture. All the mutants exhibited delays that were significantly longer than that of abg 2 (p<0.0001 for each mutant vs. abg 2 ), indicating that each tyrosine has some effects on the Ca 2+ response. The delay was the longest in the mutants containing Y219F, being significantly longer than in the ag 2 clones in FYY and FYF (p<0.001 and p<0.0005 vs. ag 2 , respectively). This confirmed the predominant role of Y219 already identified with the Ca 2+ peak analysis. In the mutants where Y219 was intact (YFY, YYF and YFF) the delays were significantly longer than in abg 2 clones (p<0.0001), but significantly shorter than in ag 2 clones (p<0.0001-0.01). These results confirmed the conclusion, reached with peak amplitude, that Y225 and Y229 may contribute to Ca 2+ mobilization, but by themselves do not play a major role. Taken together these results confirmed the predominant role of Y219 in FcRb signal amplification observed with FcRg phosphorylation and lyn association.
One additional element that needs to be taken into account is the likelihood that the function of each tyrosine is not independent of that of the others. In fact, we demonstrated this lack of independence when we could not calculate numerical values for the effect on the Ca 2+ delay of each of the tyrosines in their wild-type and mutated states. Nevertheless, this analysis of Ca 2+ flux kinetics provided an additional finding that was unexpected. When Y225 was mutated along with Y219 (FFY and FFF), the delay was shorter than when Y225 was intact (p<0.02 for FFY and FFF vs. FYY), and was not significantly different from that of a g 2 (Fig. 5C).
Conversely, when Y219 was mutated and Y225 intact (FYY and FYF), the delay was significantly longer than in ag 2 . When Y219 was intact (YFY and YFF), mutating Y225 had no effect. This suggested an inhibitory role for Y225, which was unmasked when the amplifying role of Y219 was removed. The longer delay in FYY and FYF compared to ag 2 suggested that this increase could be due to both the loss of the positive effect of Y219 and the persistence of the negative effect of Y225.

Discussion
Activation of lyn and phosphorylation of the receptor b and g chains are the first signals detected after FceRI aggregation. The b chain plays a critical role in recruiting and activating lyn, as demonstrated by comparing ag 2 and abg 2 (12). We show here that mutation of the Nterminal canonical tyrosine in the ITAM has a profound negative effect on early signaling after FceRI aggregation. Y219F substantially inhibits not only b, but also g phosphorylation (Fig.2), and lyn association with b (Fig. 3). As a result, signals downstream of lyn are inhibited, such as syk phosphorylation (Fig. 4) and Ca 2+ mobilization (Fig. 5).
It is unlikely that the effects of the tyrosine mutations that we observed here were due to conformational alterations of b. The mutated receptors were expressed at the cell surface at the same level as WT receptors. If the mutated b chains could not associate with a and/or g, receptor expression would be decreased, or the receptors at the cell surface would be ag 2 , and would be expressed at a lower level than WT abg 2 , due to the capacity of b to amplify receptor expression (14). To confirm experimentally that receptor composition was not altered in the mutants, we analyzed subunit composition after precipitating surface receptors specifically by binding an anti-a Ab on intact cells. Figure 1D shows that there was no consistent alteration in the ratio between b and g in the mutants. In addition to demonstrating the predominant and inhibitory effect of Y219F, our results suggest a mild positive effect of Y225F on Ca 2+ flux when associated with Y219F. Some authors have suggested that FcRb could play an inhibitory role. Studies with synthetic peptides showed that the b ITAM could bind the inositol 5'-phosphatase SHIP (34,35) and the protein tyrosine phosphatase SHP-2 (36). In addition, immunoprecipitation studies showed that SHP-1 and SHP-2 associated with FceRI (36), and that they and SHIP were phosphorylated after FceRI aggregation (35,36). Both of these phosphatase families have been implicated in FceRI signaling regulation ((37) and reviewed in (38,39)). Whether the apparent negative activity associated with Y225 is mediated by the tyrosine phosphatases and inositol phosphatases that have been reported to associate with FcRb in vitro remains to be determined.
The interactions between lyn and b include, among other possible mechanisms, the binding of the lyn SH2 to the phosphorylated tyrosines of the b ITAM after receptor aggregation.
The interactions between SH2 and phosphorylated tyrosines have been extensively studied ( (40)(41)(42)(43) and reviewed in (44)(45)(46)(47)(48)). In particular, lyn SH2 has been shown to exhibit a preference for the phosphorylated tyrosines of the b ITAM (43). The question of whether both, and furthermore the three, tyrosines in the b ITAM can interact with lyn does not appear to have been addressed.
Our results showing the functional importance of the N terminal tyrosine suggest that lyn SH2 interacts preferentially with this tyrosine.
In summary, we have demonstrated that the signal amplification function of FcRb is mediated principally by N terminal tyrosine through its interaction with lyn. The other two tyrosines may play an accessory role, in particular the non-canonical tyrosine, which could inhibit Ca 2+ mobilization.