Fc Receptor γ Chain Residues at the Interface of the Cytoplasmic and Transmembrane Domains Affect Association with FcαRI, Surface Expression, and Function*

The assembly of multiple subunit immunoreceptors is dependent on transmembrane interactions. The Fc receptor γ (FcR-γ) chain is a ubiquitous immune receptor tyrosine-based activation motif-containing dimeric subunit, γ2, which in humans associates with both the activating members of the leukocyte receptor cluster, including the IgA receptor FcαRI, and the classical Fc receptors, including the IgE receptor FcϵRI. This study identifies a new site in the transmembrane region of FcR-γ that affects receptor assembly and surface expression with FcαRI but not with FcϵRI. The wild type complex, FcαRI-γ2WT, remains robustly associated in both Brij-96 and Thesit detergent conditions. However, mutation of either Tyr25 or Cys26 of FcR-γ, near the interface of the transmembrane and cytoplasmic regions, resulted in impaired FcR-γ association with FcαRI. This association was disrupted in the presence of the detergent Brij-96 but was preserved in milder conditions using the detergent Thesit. Ligand-mediated cross-linking of the FcαRI-γ2Y25F mutant receptor resulted in diminished signal transduction, including an abnormal calcium response, compared with the FcαRI-γ2WT receptor. Furthermore, the FcαRI-γ2Y25F mutant receptor was expressed at the cell surface at ∼10% of that of the wild type, whereas the surface expression of FcϵRI-γ2Y25F was not significantly different from the wild type. In contrast, although the FcαRI-γ2C26S mutant was also less stably associated, it was not reduced in surface expression or function. Thus, these TM residues of FcR-γ are important for association with FcαRI and probably other activating LRC members but not with the classical FcR, FcϵRI.

Many of the receptors known to be essential to immunity, including the receptors for immunoglobulins, the antigen receptors of T and B lymphocytes, and the activating receptors of NK and myeloid cells, have separate subunits that perform ligand binding or signal transduction roles and are organized through transmembrane (TM) 1 interactions (1)(2)(3)(4). FcR-␥ is the most widespread of such signal transduction subunits, being present on macrophages, monocytes, NK cells, platelets, eosinophils, mast cells, and some T cells. It associates with a number of immunoreceptors, including the IgA receptor Fc␣RI in humans (5)(6)(7)(8). Fc␣RI associated with FcR-␥, Fc␣RI-␥ 2 , is the receptor responsible for IgA-mediated activation of leukocytes, stimulating cells for respiratory burst, phagocytosis, cytokine secretion, and antigen presentation (9 -13). Fc␣RI not associated with FcR-␥ can bind, endocytose, and recycle IgA (14). Some control of association can occur, since assembly of Fc␣RI with FcR-␥ does not occur in colostral neutrophils although FcR-␥ is present (15). FcR-␥ also associates with the high affinity IgE receptor Fc⑀RI and the IgG receptors Fc␥RI and Fc␥RIIIa (16). Furthermore, FcR-␥ associates with the collagen receptor gpVI on platelets and, in humans, the activating leukocyte Ig-like receptors, which like Fc␣RI are members of the leukocyte receptor cluster (LRC) encoded at chromosome 19q13.4 (17)(18)(19). In mice, a small family of LRC homologues, the paired Ig-like receptors, also associate with FcR-␥ (20,21).
The FcR-␥ subunit is a disulfide-linked homodimer, ␥ 2 , with each chain consisting of five extracellular residues, a putative 21-residue transmembrane region (residues 6 -26), and a cytoplasmic domain of 42 residues, containing an immune receptor tyrosine-based activation motif essential for cell activation. A potentially charged transmembrane aspartic residue, Asp 11 , is required for dimerization and is also important in the interaction with ligand binding receptor subunits. Fc␣RI (6,7), the collagen receptor gpVI (22,23), and leukocyte Ig-like receptor A2 (ILT-1) (24) each contain a transmembrane arginine residue essential for interaction with FcR-␥.
Other FcR-␥ residues besides Asp 11 are likely to be important in the assembly of FcR-␥ in immunoreceptors. In this study, we have focused on residues at the interface of the transmembrane and cytoplasmic domains. Thus, we have investigated FcR-␥ residues Tyr 25 and Cys 26 just prior to the stop transfer sequence R27L28K29. This paper addresses whether mutant FcR-␥, altered in transmembrane residue Tyr 25 or Cys 26 , signals normally in association with Fc␣RI (CD89) as a ligand binding subunit. Mutation of either residue impaired assembly of FcR-␥ with Fc␣RI, but only the Fc␣RI-␥ 2 Y25 mutant receptor was functionally defective, exhibiting reduced surface expression and signal transduction. Neither the Y25F nor C26S FcR-␥ mutants appeared to be impaired in expression with Fc⑀RI, indicating that the role of these residues in receptor assembly is possibly restricted to interaction with activating LRC receptors.

EXPERIMENTAL PROCEDURES
Antibodies-The FcR-␥ peptide CLKHEKPPQ coupled to ovalbumin using a maleimide kit (Pierce) was used to produce a polyclonal rabbit anti-FcR ␥ as described (25). Rabbit polyclonal anti-Syk was a gift from Dr. John Cambier (26). The anti-Fc␣RI mAb A59 (27) and polyclonal rabbit anti-Fc␣RI have been described previously (28). PE-conjugated A59 was purchased from Pharmingen (San Diego, CA). HRP-conjugated mAb 4G10 was from Upstate Biotechnology, Inc. (Lake Placid, NY). The anti-FLAG M2 antibody was from Sigma.
Construction and Mutagenesis of Receptor Expression Constructs-Restriction enzymes and DNA-modifying enzymes were all from New England Biolabs (Beverly, MA) except for PCR applications, which used the polymerase Pwo (Roche Applied Science). Human FcR-␥ chain was amplified from cDNA prepared from peripheral blood mononuclear cells using a cDNA synthesis kit (Amersham Biosciences) and PCR with the oligonucleotides HT79 TCGGATCCACCATGATTCCAGCAGTGGTC and HT80 TTCTCGAGCTACTGTGGTGGTTTC and cloned into the BamHI and XhoI sites of pMXI (29). The mutations Y25F and C26S were introduced by splice overlap extension using standard molecular biology techniques. The plasmid pMXpuro was constructed by exchanging the EGFP sequence in pMXI with the puromycin gene of pAPEX3p (30). PCR was used to amplify the Fc␣RI (31) and Fc⑀RI (32) cDNAs as described previously (33,34), and the PCR product was cloned into pMXpuro. The construct encoding FcR-␥ FLAG-tagged at the mature N terminus was a gift from Dr. L. Lanier (35) and was subcloned into pMXI, and the variants encoding Y25F and C26S mutant FcR-␥ were generated as described.
Transduction of IIa1.6 Cells for IgA Receptor Expression-Human FcR-␥ wild type (WT) and mutant cDNAs in pMXI-EGFP were transfected into the packaging line Phoenix (29,36) (available on the World Wide Web at www.uib.no/mbi/nolan/NL-phoenix.html). Recombinant virus was used to infect the murine B cell line IIa1.6 lacking endogenous Fc receptors (37). The transduced cells were selected for FcR-␥ expression, using flow cytometry to sort for EGFP expression. Subsequent infection with recombinant retrovirus expressing the Fc␣RI or Fc⑀RI cDNA in pMXpuro provided the ligand binding subunit for the receptor. After selection in puromycin, Fc␣RI-expressing cells were sorted using PE-conjugated anti-Fc␣RI mAb A59 (Pharmingen) (27). For the analysis of FcR-␥ surface expression, IIa1.6 cells expressing Fc␣RI were transduced with retrovirus expressing the FLAG-tagged WT FcR-␥ or the Y25F or C26S mutants.
FACS Analysis of Cells Expressing Fc Receptors-Cells expressing FcR-␥ and Fc␣RI were incubated with a 1:25 dilution of PE-conjugated mAb A59 (Pharmingen) for 40 min on ice and washed with PBS (20 mM phosphate, 150 mM NaCl, pH 7.4) containing 0.1% bovine serum albumin, and fluorescence was measured using a FACSCalibur (BD Biosciences). Fc⑀RI expression was measured by IgE binding. Briefly, cells were incubated with mouse IgE myeloma TIB142 (ATCC, Manassas, VA) supernatant for 40 min, on ice, washed with PBS containing 0.1% bovine serum albumin, and incubated with a 1:50 dilution of 0.5 mg/ml biotin-conjugated anti-mouse Ig followed by 1:200 of 0.5 mg/ml PE-conjugated streptavidin (Pharmingen). The surface expression of FLAG-tagged FcR-␥ subunits was measured by sequential incubation of cells with anti-FLAG monoclonal antibody M2 (25 g/ml), biotinylated anti-mouse IgG1 (10 g/ml), and PE-conjugated streptavidin (Pharmingen).
Immunoprecipitation of the IgA Receptor-Gamma Bind G Sepharose TM (Amersham Biosciences) (20 l) was incubated at 4°C for 1 h with 2 l of A59 ascites for each immunoprecipitation. IIa1.6 cells (1 ϫ 10 7 cells/immunoprecipitation), unlabeled or surface-biotinylated for 40 min on ice in 1 ml of PBS containing 0.5 mg/ml EZ-Link TM Sulfo-NHS-LC-Biotin (sulfosuccinimidyl 6-(biotinamido) hexonate, Pierce), were collected by centrifugation and lysed for 10 min on ice in 0.5 ml of lysis buffer containing 0.5% Brij-96 or 0.05% Thesit (synonyms: polidocanol or polyexyethylene 9 lauryl ether; Sigma). Cell lysis with Thesit can preserve easily disrupted membrane protein interactions, including the detection of oligomers of the BCR (38,39). The concentration of Thesit was titered to optimize recovery of intact IgA receptor complexes. The minimum concentration of Thesit that yielded more than 50% recovery of Fc␣RI from the cells was 0.05% (data not shown). In addition, lysis buffer consisted of 10 mM Tris, pH 7.4, 150 mM NaCl, 1 mM EDTA, 20 mM iodoacetamide (Sigma), and one mini-complete protease tablet (Roche Applied Science) per 10 ml. The lysate was clarified by centrifugation at 10,000 ϫ g for 10 min at 4°C, and the supernatant was transferred to an aliquot of A59-loaded Gamma Bind G Sepharose and incubated at 4°C for 1 h. Analysis of samples used SDS-PAGE and semidry transfer to polyvinylidene difluoride membranes according to established procedures (40). Immunodetection of biotinylated receptor was performed using streptavidin/HRP (Amersham Biosciences), diluted 1:10,000 in 10 mM Tris, 150 mM NaCl containing 1% bovine serum albumin (CSL, Melbourne, Australia) for 1 h at 25°C, followed by ECL reagents (PerkinElmer Life Sciences). Detection of total FcR-␥ chain used rabbit anti-FcR-␥ chain followed by anti-rabbit Ig/HRP (Dako, Carpinteria, CA), both at 1:10,000 dilution, and ECL reagents. Likewise, detection of total Fc␣RI used goat anti-Fc␣RI (41), followed by anti-goat Ig/HRP (Dako), both at a 1:10,000 dilution, and ECL reagents. Semiquantitative analysis of immunoblots used Scion Image for Windows Version beta 4.0.2, with uncalibrated OD (Scion Corp., Frederick, MD). Since the FcR-␥ chain has only five extracellular amino acid residues, biotinylation was only possible through the N-terminal amino group, and longer exposures were required to detect labeled FcR-␥ than were required for detection of the Fc␣RI.
Detection of Phosphorylation following Stimulation of Cells via the IgA Receptor-Transduced IIa1.6 cells were incubated with serum IgA (CSL) at a final concentration of 12 g/ml for 1 h on ice and warmed to 37°C for 1 min prior to the addition of a 1:50 dilution of sheep FabЈ 2 anti-human IgA (Silenus, Melbourne, Australia) to initiate cross-linking of the IgA receptor. Control activation of the cells used 2 l of 150 g/ml rabbit anti-mouse IgG (Dako) to cross-link the endogenous BCR. Cells (1 ϫ 10 6 ; 50 l) were lysed in an equal volume of 0.5% Brij-96 in 150 mM NaCl, 10 mM Tris pH 7.4, 1 mM EDTA, 1 mM Na 3 VO 4 containing one mini-complete protease tablet (Roche Applied Science) for 10 min on ice. The cleared lysate was analyzed by Western immunoblot as described above using the anti-phosphotyrosine antibody 4G10 conjugated with HRP. Analysis of Syk phosphorylation used 1 ϫ 10 7 cells with the stimulation, lysis, immunoprecipitation (2 l of rabbit anti-Syk), and Western analysis (10% reducing SDS-PAGE) as described above. To detect total Syk Western blots were probed with the rabbit anti-Syk antiserum (1:2000 dilution), and phosphorylated Syk was detected using 4G10-HRP. One-way analysis of variance using a Dunnet multiple comparison test was used to compare the levels of Syk phosphorylation in the mutant cells to WT cells and used the program GraphPad Instat® (GaphPad Software Inc., San Diego, CA).
Calcium Measurements in IgA Receptor-expressing Cells-Transduced IIa1.6 cells were incubated in measurement buffer (20 mM HEPES, 138 mM NaCl, 6 mM KCl, 1 mM MgCl 2 , 5.2 mM glucose, 1% bovine serum albumin, 1 mM CaCl 2 , pH 7.4) at 1 ϫ 10 7 cells/ml, and to each ml was added 2 l of 2 mM Fura2/AM (Molecular Probes, Inc., Eugene, OR) and 2 l of 25% (w/v) pluronic acid (Molecular Probes) in Me 2 SO for 20 -30 min at 37°C with shaking. Fura2-loaded cells (10 6 ) in 100-l aliquots were incubated with serum IgA (CSL) at a final concentration of 16 g/ml for 10 min at 37°C in a well of a black 96-well plate (NUNC, Naperville, IL). Excitation was at 340/380 nm, and emission (510 nm) was recorded on a FLUOstar Optima plate reader (BMG Laboratories, Melbourne, Australia) at 1.5-s intervals. At 20 s, injection of 100 l of a 1:25 dilution of sheep anti-human IgA (Silenus, Melbourne, Australia) initiated cross-linking of the IgA receptor, or injection of goat F(ab)Ј 2 anti-mouse IgG (Cappel, Cochranville, PA) initiated cross-linking of the BCR. Cytosolic free Ca 2ϩ was calculated using a K d value of 224 nM for Ca 2ϩ and Fura2 as described (42). Analysis of variance using a Dunnet multiple comparison test was used to compare the mutant data sets to WT for the times with peak [Ca 2ϩ ] i concentration after stimulation.

WT and Mutant FcR-␥ Remain Associated with Fc␣RI under Mild Lysis Conditions-Recombinant retrovirus infection of
IIa1.6 cells generated cells expressing Fc␣RI and FcR-␥ chain, WT or the Y25F or the C26S mutant. FACS analysis showed that the FACS-selected pools of transduced cells were matched for Fc␣RI expression (Fig. 1, F-H) and for expression of the enhanced green fluorescent protein (Fig. 1, B-D). EGFP expression is linked to FcR-␥ expression, since both are translated from the same bicistronic mRNA containing an internal ribosome entry site. The cell surface proteins of the WT (Fc␣RI-␥ 2 ) and mutant (Fc␣RI-␥ 2 Y25F and Fc␣RI-␥ 2 C26S) receptorexpressing cell lines were labeled with biotin, the cells were lysed, using 0.05% Thesit to minimize disruption of the recep-tor complexes, and the Fc␣RI was immunoprecipitated using mAb A59. Western blotting confirmed the expression of equivalent levels of Fc␣RI ( Fig. 2A). Furthermore, probing with polyclonal anti-FcR-␥ antiserum indicated similar association of both WT and mutant FcR-␥ chains, since they all co-immunoprecipitated with Fc␣RI in the Thesit lysates (Fig. 2, B and C). The size of the proteins (ϳ18 -25 kDa) indicated that both mutants were expressed as covalent dimers, ␥ 2 , at similar levels to the wild type protein and that neither Tyr 25 nor Cys 26 is required for dimerization of the FcR-␥, an indicator of the structural integrity of the FcR-␥ chain (Fig. 2B). Thus, both the FcR-␥ Y25F and C26S mutants associate with Fc␣RI, and this interaction is preserved during lysis in 0.05% Thesit and immunoprecipitation with A59.
FcR-␥ Y25F Has Lesser Expression at the Cell Surface-Next, the surface biotinylation, cell lysis with 0.05% Thesit, and immunoprecipitation procedures were used to measure the expression of the FcR-␥ chains. As seen previously, probing with streptavidin-conjugated HRP detected equivalent amounts of Fc␣RI at ϳ65-75 kDa in each cell line (Fig. 3A). Also, equivalent amounts of cell surface-labeled WT and C26S mutant FcR-␥ chain were associated with Fc␣RI (Fig. 3B, nonreducing conditions; Fig. 3C, reducing conditions). In marked contrast, less biotin-labeled FcR-␥ Y25F was co-immunoprecipitated with Fc␣RI (Fig. 3, B and C, lane 3), indicating low surface expression of Y25F FcR-␥ or a qualitatively altered receptor that is less well labeled with biotin at the FcR-␥ chain N terminus. Semiquantitative analysis of blots showed that the biotinylated FcR-␥ Y25F was immunoprecipitated at 14 Ϯ 4% (n ϭ 4) of the level of the WT FcR-␥. Hence, most of the Fc␣RI-␥ 2 Y25F detected with anti-FcR-␥ antiserum (Fig. 2, B and C, lane 3), may be located inside the cell or is resistant to labeling. This was confirmed by direct immunoprecipitation of the FcR-␥ chains (Fig. 3, D-F). Probing with anti-FcR-␥ antiserum showed, as previously, that total expression of Y25F FcR-␥ was only slightly less than the WT and C26S (Fig. 3F), yet detection with streptavidin confirmed that little ␥ 2 Y25F was biotin-labeled (Fig. 3E, lane 3). The possibility that altered availability of the FcR-␥ N terminus for biotin labeling was responsible for the apparent low surface expression of the FcR-␥ Y25F mutant was addressed by transducing IIa1.6 cells expressing Fc␣RI with FcR-␥ subunits with a FLAG tag epitope at the N termini. The expression of EGFP from the bicistronic constructs was equivalent (WT MFI ϭ 193; Y25F MFI ϭ 198; C26S, MFI ϭ 215), indicating the matched expression of mRNAs for FcR-␥ in the WT, Y25F, and C26S cell lines (Fig. 3, G-I, insets). FACS analysis with anti-FLAG antibody confirmed that the surface expression of the FcR-␥ Y25F mutant (MFI ϭ 13.1 and background MFI ϭ 8.4) (Fig. 3H) was considerably less than that of the WT subunit (MFI ϭ 60.3). The surface staining of the FcR-␥ C26S mutant subunit (MFI ϭ 34.4) (Fig. 3I) lay between the levels of the WT and Y25F mutant subunits. Hence, although it is formally possible that altered structure of the Fc␣RI-␥ 2 Y25F receptor is responsible for a lower biotinylation or binding of anti-FLAG antibody, these two independent approaches indicate that it is most likely that there is impaired surface expression of the Y25F mutant subunit.
FcR-␥ Y25F and FcR-␥ C26S Association with Fc␣RI Is Aberrant-It is notable that immunoprecipitation with anti-FcR-␥ antiserum co-precipitated Fc␣RI only in association with WT FcR-␥ and not with either the Y25F or C26S mutant (Fig. 3D). Thus, the interaction of both mutants with Fc␣RI is not as robust as the WT interaction and is detected with the A59 immunoprecipitation but not with the anti-FcR-␥ immunoprecipitation. The strength of association of the Fc␣RI and FcR-␥ mutants was further examined by lysis in a different detergent condition, 0.5% Brij-96 (Fig. 4). Again, immunoprecipitation of Fc␣RI with the mAb A59 demonstrated that the receptor, at ϳ65-75 kDa, was expressed in equal amounts on the WT and mutant cells (Fig. 4A). Developing the blot further to detect biotin-labeled FcR-␥ found that whereas the WT FcR-␥ was co-immunoprecipitated, neither Y25F nor C26S mutants could be detected (Fig. 4). Thus, the C26S interaction, while preserved in 0.05% Thesit (Figs. 2 and 3), is disrupted in 0.5% Brij-96. Reprobing this blot with anti-FcR-␥ showed that the largely intracellular association of Y25F mutant FcR-␥ detected in the Thesit lysate (Fig. 2, B and C) was, like the C26S FcR-␥ interaction, not preserved under the Brij-96 lysis conditions (Fig. 4C, lane 3). Thus, under these harsher detergent conditions, the complexes between Y25F FcR-␥ and Fc␣RI and that between C26S FcR-␥ and Fc␣RI are not preserved, which is in sharp contrast to the WT complex, which is readily detected (Fig. 4C, lane 1). This demonstrates that these residues, Tyr 25 and Cys 26 , at the end of the TM region contribute to the interaction of FcR-␥ with Fc␣RI.
Y25F Mutant FcR-␥ Is Deficient in Transducing IgA-mediated Phosphorylation-Since both mutant FcR-␥ subunits had less robust association with Fc␣RI and the Y25F mutant had reduced surface expression, the efficacy of these mutant receptor complexes in signal transduction was tested. Cells were treated with different stimuli, and samples were analyzed by Western blotting (Fig. 5). Cross-linking of the B cell receptor in the WT and two mutant FcR-␥ chain cell lines gave similar patterns and levels of induction of phosphoproteins, indicating that the signaling machinery of the transduced cell pools are intact (Fig. 5D). Treatment with anti-IgA reagent in the absence of IgA failed to stimulate the IgA receptor-expressing cells (Fig. 5E). When cells were incubated with human serum IgA and then cross-linked with FabЈ 2 anti-human IgA, the cells expressing Fc␣RI-␥ 2 WT induced phosphoproteins, which peaked at ϳ2 min poststimulation and then progressively declined at 4 and 7 min (Fig. 5A). Similar kinetics, pattern, and intensity of phosphoprotein induction were observed for the cells expressing the Fc␣RI-␥ 2 C26S mutant receptor (Fig. 5C). There were minor differences at ϳ28 and Ͼ150 kDa between the WT and C26S pattern of induced phosphoproteins. The cells expressing Fc␣RI-␥ 2 Y25F receptor showed reduced intensity of induced phosphoproteins across the 7-min time course (Fig. 5B). Whereas the Y25F cell line showed lower intensity of phosphoproteins, the pattern and kinetics of the major bands detected was identical to the WT, suggesting the same program of signaling events was occurring but at reduced efficiency.
An important kinase involved early in the FcR signal transduction pathway is Syk. If the diminished phosphorylation of cellular substrates occurring after the stimulation of the cells expressing Fc␣RI-␥ 2 Y25F is due to defective activity of this mutant receptor, then diminished activation of Syk would be predicted. Cells were treated with IgA and cross-linked, the Syk kinase was immunoprecipitated from lysates, and phosphorylated Syk was detected using the mAb 4G10 (Fig. 5F). Without receptor cross-linking, no phosphorylated Syk was detected. Treatment with IgA and anti-IgA resulted in the detection of more phosphorylated Syk from the cells expressing the WT IgA receptor than from the cells expressing the Y25F mutant receptor despite there being equivalent levels of Syk in these cell lines.
FcR-␥ Y25F Is Deficient in Transducing IgA-mediated Cellular Calcium Flux-Last, we assessed the ability of the mutant receptor complexes to trigger another cell activation event, calcium flux upon IgA stimulation. Fluxes of intracellular calcium were measured when Fura-2-loaded cells were stimulated with IgA and anti-IgA cross-linking (see Fig. 7A). The WT FcR-␥ cells showed a flux in [Ca 2ϩ ] i , with the maximum elevation occurring at 53 Ϯ 3 s (n ϭ 6). Despite the greatly reduced phosphorylation seen in the cells expressing Fc␣RI-␥ 2 Y25F, the peak in [Ca 2ϩ ] i was robust but significantly delayed (73 Ϯ 8 s, n ϭ 5, p Ͻ 0.01). In addition, the Y25F receptor flux showed a broader peak of elevated [Ca 2ϩ ] i that often also had lower amplitude (Fig. 6A, open circles). Whereas the kinetics of the [Ca 2ϩ ] i flux for C26S mutant (62 Ϯ 8 s, n ϭ 5) appears delayed, this was not significant compared with the WT. The mobilization of [Ca 2ϩ ] i from internal stores was determined with EGTA in the measurement buffer. Again, differences in the kinetics of the peak mobilization were observed for the cell lines (WT, 47 s; C26S, 55 s; Y25F, 60 s; n ϭ 2), with the cells expressing the Y25F mutant ␥ chain receptor having delayed mobilization of [Ca 2ϩ ] i (Fig. 6B, open circles). The calcium flux machinery of the cells was assessed by stimulation through the surface BCR of these cells. Cross-linking with anti-mouse Ig resulted in calcium fluxes from the Y25F and C26S mutant cell lines with identical kinetics. A similar, possibly slightly delayed, mobilization was obtained with the cells expressing the WT IgA receptor (Fig. 6C). Thus, there is not any large difference in the general calcium mobilization machinery of the three IgA receptor expressing cell types.
The Y25F or C26S Mutants of FcR-␥ Enable Surface Expression of Fc⑀RI-It is possible that alteration of the conformation of FcR-␥ by the Y25F or C26S mutation is responsible for the impaired interaction of the mutants of FcR-␥ with Fc␣RI. This possibility, of gross conformational disruption, was tested by measuring the co-dependent surface expression of the FcR-␥ mutants in the context of Fc⑀RI. Cell lines expressing WT, Y25F, or C26S FcR-␥ chain were transduced with recombinant retrovirus to express the human Fc⑀RI ligand binding chain. FACS analysis of IgE binding was used to evaluate the assembly and surface expression of Fc⑀RI and FcR-␥ chain, since Fc⑀RI surface expression is dependent on assembly with FcR-␥. Transduction of FcR-␥ WT or FcR-␥ mutant cells with Fc⑀RI recombinant retrovirus and selection in puromycin resulted in cells able to bind IgE. The levels of Fc⑀RI expression in the context of both the FcR-␥ Y25F mutant (normalized IgE binding ϭ 80 Ϯ 18%, n ϭ 4) and the FcR-␥ C26S mutant (106 Ϯ 6%, n ϭ 4) did not differ significantly from that for FcR-␥ WT (100%). Thus, the FcR-␥ Y25F and C26S mutants were competent for association and surface expression of Fc⑀RI, a classical FcR. In the absence of the Fc⑀RI ␤ subunit, Fc⑀RI and FcR-␥ signal relatively ineffectively (43), so this system was not readily amenable to further testing if these mutations affected signaling efficacy of the IgE receptor.

Association of FcR-␥ and Fc␣RI Subunits-Transmembrane
residues of the FcR-␥ subunit, at the interface with the cytoplasmic domain, are important in the interaction with the Fc receptor Fc␣RI. Two conservative individual mutations, Y25F or C26S, in the FcR-␥ chain resulted in destabilization of the receptor complex, such that it dissociated under detergent lysis conditions that completely preserved the interaction of WT FcR-␥ with Fc␣RI. It is difficult to ascertain the structural effects of point mutations in the TM region of proteins, but there are several reasons to conclude that the effect of these mutations on receptor complex stability is unlikely to be mediated through conformational disruption of the FcR-␥. First, the two mutations lie outside the dimerization motif important for assembly of the CD3/FcR-␥ chain dimers (44). Second, these mutations are very conservative in nature, involving the loss of a single oxygen atom in the Y25F mutant and the swap of a sulfur for an oxygen atom in the C26S mutant. Third, the Y25F mutation changes the sequence to that of the close homologue CD3-, which shares some functional activities with the FcR-␥ chain, such as assembly with Fc␥RIIIa in humans. Fourth, there was little effect on the assembly of the FcR-␥ mutants with Fc⑀RI, which, as a consequence of the unrelated TM segments of Fc⑀RI and Fc␣RI, will interact with the FcR-␥ chain TM very differently than Fc␣RI. Hence, there is no evidence for the Y25F or C26S mutations disrupting the conformation of the FcR-␥ chain.
These mutations only destabilized the Fc␣RI-FcR-␥ complex and did not completely abolish interaction with Fc␣RI, since complexes could be detected using Thesit lysis conditions optimized to preserve very weak membrane protein interactions. Single amino acid changes in the TM region have been reported in other studies to affect the assembly of receptors with FcR-␥ chain or the homologous CD3 chain. Mutation of the TM potentially charged aspartic acid in FcR-␥, CD3, and Fc␥RIII, or Fc⑀RI abrogates assembly of these subunits into receptor complexes (45)(46)(47). Likewise, among the LRC receptors, the TM arginine residue has been shown for Fc␣RI (Arg 209 ) and gpVI (Arg 272 ) to be essential for assembly with the FcR-␥ chain (6,7,22,23). Specific hydrophobic TM residues can also be important, since amino acid residue 46 of CD3, leucine in humans and isoleucine in the mouse permit assembly with Fc␥RIIIa in NK cells only in humans (48). In this study, the interaction of tyrosine 25 and cysteine 26 of FcR-␥ with Fc␣RI describes a new TM interaction for FcR-␥ with a member of the LRC family of receptors. A model of murine CD3 predicts that the tyrosine (or phenylalanine in human CD3) equivalent to Tyr 25 of FcR-␥ forms part of the boundary of a pocket that accommodates the TCR CD3␦ and ␥ subunits (44). The involvement of both Tyr 25 and Cys 26 in interactions with Fc␣RI suggests that the Fc␣RI TM helix interacts at the interface between the two FcR-␥ helices (Fig. 7B). This also is in harmony with the predicted interaction of CD3␦ and -␥ with the cavity formed at the interface between the two helices of the CD3 dimer in the TCR (44). In contrast to the aberrant expression of Fc␣RI-␥ 2 Y25F, the normal expression of Fc⑀RI in the context of this FcR-␥ mutant, Fc⑀RI-␥ 2 Y25F, exemplified that the TM of FcR-␥ makes different interactions with the activating LRC receptors than with the "classical" Fc receptors. Although Fc⑀RI binds FcR-␥ with less affinity than Fc␣RI, perhaps by making fewer significant TM interactions, both receptor complexes are likely to have the same approximate topology. Indeed, the FcR-␥ residue Tyr 25 lies close to Leu 21 (Fig. 7B), which we predict will participate in FcR-␥ binding to both Fc␥RIIIa and Fc⑀RI, from the essential role of the equivalent Leu 46 residue in the homologous CD3 interaction with Fc␥RIIIa (48).
The assembly of Fc⑀RI or Fc␥RIIIa with FcR-␥ allows co-dependent transport through the secretory pathway to the cell surface. A similar scenario also occurs with the more complex TCR, where the CD3␥⑀TCR␣␤CD3␦⑀ complex assembled in the ER must associate with the FcR-␥ homologue CD3 for successful export of the completed complex from the Golgi (49,50). In this study, whereas both the FcR-␥ Y25F and C26S mutants appeared to associate with Fc␣RI with reduced affinities, only the Fc␣RI-␥ 2 Y25F mutant receptor complex was impaired in cell surface expression as determined by surface biotinylation and reactivity of an N-terminal FLAG epitope. Hence, additional factors beyond receptor complex stability, as we could measure by co-immunoprecipitation from different detergent lysates, may be required for efficient surface expression. The low (ϳ10% by biotinylation) surface expression of the Fc␣RI-␥ 2 Y25F mutant receptor complex is probably sufficient to account for its diminished signaling activity. Nonetheless, a qualitative functional defect in the lesser amount of surface-expressed Fc␣RI-␥ 2 Y25F is not formally ruled out. In this respect, it is noteworthy that the TM/cytoplasmic interface of CD3⑀ has been suggested to play a role in the earliest signaling events of the T cell receptor via a "piston-like displacement" mechanism (51).
Some aspects of our study complement that of Morton et al.  4). B, schematic diagram of the TMs of the FcR-␥ dimer interacting with Fc␣RI. Predicted helical wheel diagrams (available on the World Wide Web at www.site.uottawa.ca/ϳturcotte/resources/HelixWheel/) of the TM regions of FcR-␥ are shown. The dimer is an approximate representation, since D11 and D11Ј are likely to be in hydrogen bonding contact with each other in the actual dimer interface. The TM disulfide is shown as a thick line linking Cys 7 and Cys 7Ј of the two FcR-␥ helices, and the putative interactions between Arg 209 of Fc␣RI and Asp 11Ј of FcR-␥ and interactions between Tyr 25 and Cys 26Ј of FcR-␥ with Fc␣RI are indicated by the dotted lines. Leu 21 is equivalent to Leu 46 of CD3 important in Fc␥RIIIa binding and is marked with an asterisk. (7). In that study, TM residue Arg 209 of Fc␣RI was mutated conservatively to a histidine residue, which presumably somewhat maintained the potential "charge" interaction with Asp 11 of the FcR-␥. The R209H Fc␣RI assembled with FcR-␥ and initiated a calcium response with delayed kinetics (7). This delayed calcium response of the R209H mutant of Fc␣RI resembles the phenotype of the Y25F mutant of FcR-␥. It is not known if the Fc␣RI R209H is similarly defective in expressing FcR-␥ at the cell surface.
Transmembrane Tyrosines in Other Immunoreceptors-TM tyrosine residues appear to have a particular role in the assembly and function of immunoreceptor complexes. Transfection of TCR ␤-deficient Jurkat cells with a cDNA encoding the human TCR ␤ with a mutated transmembrane tyrosine reconstituted a TCR with mostly intracellular expression and weakly associated CD3 subunit (52). The weak association of CD3 resulted in defective recruitment of ZAP70 (53). Similarly, mutation of the transmembrane tyrosine residues in the mouse TCR ␤ subunit showed defects in assembly with the CD3 subunit, signal transduction, and thymocyte development (54). The reduction in TCR surface expression with impaired CD3 association is similar to the effect of the FcR-␥ Y25F mutation on association and expression with Fc␣RI seen in this study.
Fc␥RIIb is another immunoreceptor wherein a tyrosine residue is required for TM functions. Tyr 235 of Fc␥RIIb is required for the ITIM-independent inhibitory activity of this receptor, which is suggested to occur by disrupting the interaction of CD19 with the BCR (55). Tyr 235 of Fc␥RIIb is most structurally pertinent to this study, since, like FcR-␥ Tyr 25 , it is the penultimate transmembrane residue. Finally, the BCR consists of surface Ig and the signal transduction subunits Ig␣ and Ig␤. The human IgM Y587V/S588V mutant in the TM domain failed to assemble with the signal transduction subunits Ig␣ and Ig␤ (56). Mice expressing Y587V/S588V transgenic IgM heavy chains failed to normally signal through the BCR and were defective in maturation of pre-B cells (57). The single mutant IgM Y587F assembled with Ig␣ and Ig␤ to form a BCR functional in signal transduction but incapable of mediating antigen presentation (58 -60).
In this study, we showed that residues Tyr 25 and Cys 26 of the FcR-␥ subunit play roles in the assembly of an IgA receptor complex with Fc␣RI. Furthermore, the Y25F mutant FcR-␥ showed reduced expression at the cell surface, and this may account for the diminished induction of phosphorylation and calcium flux triggered by this mutant receptor. Thus, FcR-␥ Tyr 25 influences the assembly and transport to the cell surface of the IgA receptor, Fc␣RI-␥ 2 . The lack of an effect of this mutation on the expression of Fc⑀RI indicates that the FcR-␥ chain makes different TM interactions with the LRC receptors and the classical Fc receptors, like Fc⑀RI.