The Two Membrane Isoforms of Human IgE Assemble into Functionally Distinct B Cell Antigen Receptors

The WEHI 231 B cell lymphoma (provided by Dr. Roberto Sitia, Dipartimento di Ricerca Biologica e Tecnologica-San Raffaele Scientific Institute, Milan, Italy) was maintained in DMEM supplemented with 10% FCS, penicillin (100 U/ml), streptomycin (100 μg/ml), and 50 μM 2-mercaptoethanol. Transfections were performed by electroporation as described elsewhere (30), except that 800 μg/ml of G418 (Geneticin; Life Technologies, Inc., Gaithersburg, MD) were used for selection of the WEHI 231 clones. Cells were cloned by limiting dilution, and positive clones were identified by immunofluorescence.

The construction of the pCIG-CεCH4M1′-M2 has been described in detail previously (22). The pCIGCεCH4-M1-M2 vector was similarly constructed by replacing the BglII/KpnI fragment of pCIG-CεCH4-M1′-M2 with a corresponding fragment containing the M1 instead of the M1′ exon (Fig. 1 c).

The expression of surface IgE was examined by flow cytometry on a FACScan^®, (Becton Dickinson Immunocytometry Sys., Mountain View, CA) using rabbit anti– human IgE (ε chain) (Dako Corp., Carpinteria, CA), and FITCconjugated swine anti–rabbit IgG (Dako Corp.).

Metabolic labeling and immunoprecipitations were performed as previously described (30). Briefly, 5 × 10⁶ cells/ml were labeled with [³⁵S]methionine (Amersham Intl., Buckinghamshire, England) at 100–250 μCi/ml (1 Ci = 37 GBq), and chased with cold methionine as indicated in the figures. Cell lysates were immunoprecipitated with rabbit Ig's to human IgE (ε chains) or rabbit Igs to mouse IgM (μ-chains) (Dako Corp.) and purified by protein A–Sepharose. The samples were analyzed by SDS-PAGE in the presence or absence of mercaptoethanol, as indicated in the figure legends. Treatments of labeled supernatants with recombinant N-glycosidase F (PNGase F) and endoglycosidase H (Endo H) were performed according to the protocols provided by the manufacturer (New England Biolabs Inc., Beverly, MA).

Twenty million cells were washed twice with PBS and incubated in 1 ml of PBS containing 0.5 mg/ml of sulfo-NHSbiotin (Pierce Chem. Co., Rockford, IL) at room temperature for 15 min. Free succinimide groups were blocked by the addition of 10 ml of nonsupplemented medium at room temperature for 10 min. Cells were washed twice with PBS and resuspended in 500 μl lysis buffer containing 1% digitonin (Sigma Chemical Co., Steinheim, Germany), 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 5 mM EDTA, and the protease inhibitors PMSF (1 mM), aprotinin (10 mg/ml), and leupeptin (10 mg/ml) (all from Sigma Chemical Co.). Lysates were incubated on ice for 30 min and centrifuged at 10,000 g for 30 min at 4°C. The supernatants were precleared three times with 30 μl of protein A agarose beads (GIBCO BRL) at 4°C for 30 min. Immunoprecipitations were performed by incubating the lysates with 6 μl of rabbit Igs to human IgE (ε chains) (Dako Corp.), rabbit Igs to mouse IgM (μ chains) (Dako Corp.), or goat Igs to human IgE (ε chains) (Kirkegaard & Perry Labs., Inc., Gaithersburg, MD), at 4°C for 60 min. 20 μl of protein A agarose were then added and incubated for a further 30 min at 4°C. The Ig-α/Ig-β heterodimer was immunoprecipitated using a polyclonal antibody against Ig-α (provided by Dr. J.C. Cambier, National Jewish Center for Immunology and Respiratory Medicine, Denver, CO). Beads were preincubated with 10 mg/ml bovine serum albumin for 20 min and washed three times with lysis buffer before use. The beads were pelleted by centrifugation (the supernatants were saved for reprecipitation), washed twice with high salt lysis buffer (50 mM Tris, pH 7.5, 500 mM NaCl, 5 mM EDTA, 0.2% digitonin, plus inhibitors), and twice with low salt buffer (50 mM Tris, pH 7.5, 150 mM NaCl, 5 mM EDTA, 0.2% digitonin, plus inhibitors). For reprecipitation, supernatants were precleared twice with 30 μl beads before adding a different primary antibody.

Immunoprecipitates from digitonin lysates were resuspended in kinase buffer (50 mM Tris, pH 7.5, 10 mM MnCl₂, 1 mM EDTA, 1 mM sodium orthovanadate [Sigma Chemical Co.], 1% digitonin, plus inhibitors) containing 10 μCi [γ³²P] (Amersham Intl.), and incubated at room temperature for 5 min. The reaction was terminated by addition of kinase buffer containing 50 mM EDTA, and the pellets were collected by centrifugation before resuspending in SDS-PAGE reducing sample buffer. The samples were boiled and electrophoresed on a 10% SDSPAGE. The gel was dried without fixing and autoradiographed at −70°C.

Tyrosine phosphorylated proteins were detected as previously described (31). Briefly, 4 × 10⁶ cells were resuspended in 0.5 ml of DMEM and stimulated with either goat anti–human IgE (ε chain) (Kirkegaard & Perry Labs., Inc.) (20 μg/ml) or anti–mouse IgM (5 μg/ml) at 37°C for the indicated periods of time. After washing twice with ice-cold PBS containing 1 mM sodium orthovanadate, cells were lysed with 100 μl of Triton X-100 lysis buffer containing 20 mM Tris-HCl (pH 8.0), 137 mM NaCl, 10% glycerol, 1% Triton X-100, 2 mM EDTA, plus inhibitors, and 1 mM sodium orthovanadate. After incubation for 10 min on ice, the supernatants were cleared by centrifugation at 10,000 g for 15 min at 4°C. Subsequently, 30 μl of supernatant was subjected to 10% SDS-PAGE and Western blotting. The PVDF membrane (Amersham Intl.) was incubated with the antiphosphotyrosine monoclonal antibody PY20 (Transduction Laboratories, Lexington, KY), followed by horseradish peroxidase–coupled anti–mouse Ig antibody (Dako Corp.). The bound antibodies were visualized by the enhanced chemiluminescence detection system (Amersham Intl.).

Two B cell transfectomas that expressed comparable levels of the long IgE isoform (m_LIgE) and the short IgE isoform (m_SIgE) were cultured in triplicate in 96well plates at a cell density of 2 × 10⁴/well. The cells were incubated for 24 h with various amounts of goat anti–human IgE (ε chain) (Kirkegaard & Perry Labs., Inc.) in 0.2 ml medium. Proliferation was determined by incorporation of [methyl-³H]thymidine (1 μCi/well, 94.0 Ci/mmol; Amersham Intl.). After 14 h of incubation, cells were harvested and levels of incorporated [³H]thymidine were measured by scintillation counting.

To investigate the properties of the two membrane IgE isoforms, we generated two chimeric mouse/human ε-chain gene constructs which contained the mouse heavy chain variable region segment from an Ab with anti–4-hydroxy5-iodo-3-nitro phenacetyl (NIP) specificity and the human Cε region with a COOH terminus corresponding to the m_LIgE or the m_SIgE isoform. A schematic diagram of the two constructs is depicted in Fig. 1 c. The two constructs were independently transfected into WEHI-231 cells, and G418 resistant transfectants were selected for the expression of mIgE by immunofluorescence.

To investigate the assembly and transport of the m_LIgE and m_SIgE isoforms, we perfomed pulse-chase studies in the WEHI cell lines. Cells were given a 15 min pulse with [³⁵S]methionine and then chased for 1, 2, 3, and 4 h. mIgE was immunoprecipitated from the WEHI-m_SIgE or WEHIm_LIgE cell extracts and analyzed on a nonreducing 6% SDS-PAGE (Fig. 2,a). Two major species of more than 240 kD were observed in both cell lines, demonstrating that both membrane isoforms are assembled into H₂L₂ molecules. During the course of the chase, the quantity of immunoprecipitated proteins remained the same, indicating that there is no intracellular degradation of either membrane isoform. The accumulation of higher molecular weight species in the course of the chase suggested that both proteins are terminally glycosylated during their transport to the cell surface. This was confirmed by treatment of the immunoprecipitated material with Endo H which cleaves N-linked high mannose oligosaccharides of glycoproteins before they arrive in the medial Golgi, but fails to cleave terminally processed carbohydrate moieties. As shown in Fig. 2,b, after Endo H treatment two ε species were observed in each case (a) a 60-(m_SIgE) or 65-(m_LIgE) kD species corresponding to Endo H–sensitive ε chains that carried unprocessed high mannose carbohydrates and (b) an 87-(m_SIgE) or 90-(m_LIgE) kD species corresponding to mature ε chains that bore processed N-linked carbohydrates which were resistant to the enzyme. In the WEHI-m_LIgE transfectoma, only 20–30% of the newly synthesized membrane ε chains were of the mature processed type after 4 h chase (Fig. 2). This was similar to the rate of transport observed for the endogenous mouse mIgM. Interestingly, three to four times more m_SIgE H chains were present in the mature form after the same period of time. This clearly demonstrated that the m_SIgE is more efficiently transported to the cell surface than the m_LIgE, or even the endogenous mIgM.

To characterize the IgE-BCR, and to determine the m_SIgE and m_LIgE associated proteins, we immunoprecipitated the IgE antigen receptor complex from digitonin lysates of surface biotinylated WEHI-m_SIgE and WEHI-m_LIgE cells. The immunoprecipitations were perfomed in duplicate using two different anti–human IgE antibodies. As a control, the endogenous IgM-BCR was subsequently immunoprecipitated from the same cell extracts. The immunoprecipitated material was resolved on a 10% SDS-PAGE under reducing conditions (Fig. 3,a). Analysis of the mIgM complex yielded the expected characteristic pattern of an 80-kD band corresponding to the μ heavy chain, a 25-kD band corresponding to the light chain, and the 32- and 40–47-kD bands corresponding to Ig-α and Ig-β polypeptides, respectively (18). A similar pattern was observed after immunoprecipitation of the m_SIgE- and m_LIgE-BCRs, indicating that Ig-α and Ig-β chains are also present in these receptors. This was confirmed by analysis of the material that was dissociated from the immunoprecipitated IgEBCRs by treatment with NP-40. Reprecipitation of this material with an anti–Ig-α specific antiserum produced only the Ig-α and Ig-β subunits (Fig. 3,b).The Ig-β subunit was of the same molecular mass in the three BCRs, whereas the Ig-α of the m_LIgE (m_LIgE-α) showed a slower mobility than the IgM-α and m_SIgE-α (Fig. 3). Slower mobility of the Ig-α chain has also been observed in the case of the IgD-BCR. This difference has been shown to be the consequence of different terminal glycosylation of the two N-glycosylation sites in the polypeptide (8, 32, 33). To establish if this is the case with the m_LIgE-α, immunopreciptated materials were treated with endoglycosidases. Fig. 3 a shows that Endo H treatment decreased the molecular weight of the IgM- and m_SIgE-associated Ig-α polypeptides. On the other hand, the m_LIgE-α remained at the same position, indicating that both glycosylation sites are terminally glycosylated. Treatment with PNGase, which cleaves all sugar residues, eliminated the difference in size between the different Ig-α polypeptides.

Two additional proteins of 37 and 41 kD were detected in the m_SIgE- and m_LIgE-BCR immunoprecipitations (Fig. 3,a). These proteins were not present in the IgM-BCR and were immunoprecipitated with either of the two anti-IgE Abs, indicating that they are specifically associated with mIgE. The relative amount of these IgE-BCR associated proteins (further referred to as εBAP37 and εBAP41) was different in the two IgE-BCRs, with significantly lower quantities of εBAP41 in the m_LIgE-BCR. These proteins were not immunoprecipitated with the anti–Ig-α antiserum, indicating that they are not covalently associated to Ig-α or antigenically related to this polypeptide (Fig. 3 b).

To further characterize the components of the IgEBCRs, two dimensional analysis of anti-IgE immunoprecipitated material was performed on digitonin lysates from biotinylated WEHI-m_SIgE and WEHI-m_LIgE cells. The analysis was done on a 10% SDS-PAGE under nonreducing conditions in the first dimension and reducing conditions in the second (Fig. 4). In addition to the bands corresponding to ε₂L₂ and εL, a complex of ∼75 kD was present in the nonreducing PAGE analysis of the m_SIgE-BCR. A slightly higher mol wt complex (80 kD) was detected in the m_LIgE-BCR. In the second dimension, these complexes dissociated into free Ig-α and Ig-β chains. This experiment also indicated that the εBAP37 and εBAP41 proteins form a disulfide-bound complex (εBAP37/41) which, in the nonreducing PAGE analysis of the m_SIgE-BCR, migrated distinctly from the Ig-α/Ig-β heterodimer. In the m_LIgE-BCR the εBAP37 apparently formed a homodimer which comigrated with the Ig-α/Ig-β complex.

To investigate if the IgE-BCRs contain protein kinases capable of phosphorylating the Ig-α/Ig-β heterodimers as shown for the IgM-, IgD- and IgG-BCRs (34), we performed in vitro [γ³²P]ATP labeling of immunoprecipitated BCRs. The endogenous IgM-BCR from both transfectomas was analyzed as an internal control. As shown in Fig. 5, bands corresponding to the Ig-α and Ig-β proteins were found to be phosphorylated when incubated with [γ³²P]ATP in all three BCRs. The identity of these bands was again confirmed by treatment with Endo H and PNGase F. The two newly identified εBAPs were not detected in this in vitro assay, suggesting that they are not substrates for the BCR-associated protein kinases.

To analyze the activation of PTKs upon engagement of the IgE-BCRs, we used transfectant cell lines (WEHI-m_LIgE and WEHI-m_SIgE) that expressed comparable levels of m_LIgE and m_SIgE (Fig. 6). Wild-type WEHI cells were analyzed in parallel as control. The three cell lines were incubated for different periods of time (1–40 min) in the presence of either anti-IgE or anti-IgM. The PTK activation in these cells was monitored by the increase in tyrosine phosphorylation of PTK substrate proteins in total Triton X-100 cell lysates. Cross-linking the IgM- or the IgE-BCRs resulted in a similar pattern of phosphorylation, indicating that all three BCRs induce the phosphorylation of the same substrate proteins. However, the kinetics of phosphorylation of the two IgE-BCRs was quite different. After cross-linking the m_LIgE-BCR, the substrate phosphorylation reached its maximum in 1 min and drastically declined within 5 min (Fig. 7). In contrast, in the m_SIgEproducing transfectant, the substrate phosphorylation still increased after 1 min of stimulation and reached its maximum after 5 min. Increased tyrosine phosphorylation was still evident after 40 min, although for several polypeptides it had declined. These data showed that the kinetics of protein tyrosine phosphorylation induced by receptor crosslinking is significantly different for the two IgE-BCRs.

In subsequent experiments we observed that the kinetics of tyrosine phosphorylation were independent of the amount of IgE expressed on the cell surface. Analysis of WEHIm_SIgE and WEHI-m_LIgE clones expressing different amounts of IgE-BCR showed a difference only in the intensity of the signal, whereas the kinetics of both receptors remained the same (data not shown). Moreover, cross-linking of the endogenous IgM-BCR in the two transfectants and in wild-type WEHI cells showed the same kinetics of tyrosine phosphorylation, and was consistent with published data (35). These experiments excluded the possibility that the differences in signal transduction between the two IgE-BCRs were due to peculiarities of the transfected cell lines rather than the type of mIgE.

Cross-linking of the endogenous IgM-BCR in WEHI cells leads to their growth inhibition (36). On the other hand, cross-linking of the IgD-BCR in WEHI cells transfected with a δ chain does not affect their proliferation (37). To investigate whether signaling through the IgE-BCRs has an effect on cellular proliferation, WEHI-m_SIgE and WEHI-m_LIgE cells were grown for 24 h in the presence of various concentrations of anti-IgE Ab and were subsequently incubated for 14 h with [³H]thymidine. A marked inhibition in proliferation was observed only in the WEHIm_SIgE transfectoma; cross-linking of the m_LIgE had no effect (Fig. 8). The above results indicate that different cellular responses can occur after IgE-BCR signaling, depending on the type of mIgE expressed on the cell surface.

The data presented in this paper are the first demonstration of the existence of two functional human IgE B cell receptors that are generated by alternative splicing. Although both receptors were found to be correctly assembled, transported to the cell surface, and capable of signal transduction upon receptor engagement, a number of differences were noted between them that could only be attributed to the different extracellular membrane proximal domains. The first difference was in the rate of their transport to the plasma membrane. The transit time of the m_LIgE was much longer, since only 20% of the molecules were found to be terminally glycosylated within 3 h of their synthesis, whereas almost all of the m_SIgE molecules were converted into the mature form during this period. Interestingly, a different rate of transport has also been shown for IgM and IgD with a shorter transit time for the former (38). In our experiments the endogenous IgM was transported at a similar rate as the m_LIgE, indicating that the transport of the m_SIgE is extremely efficient.

Both mIgE isoforms were found to associate into complete BCR complexes that contain the Ig-α/Ig-β heterodimer as well as the PTKs involved in their phosphorylation. However, a difference was also noted in the pattern of glycosylation of the associated Ig-α; when associated with m_SIgE, the Ig-α polypeptide had only one of the sites terminally glycosylated and resistant to Endo H cleavage, whereas, when associated with m_LIgE, both sites were resistant to the enzyme. This is identical to what has been reported for IgM and IgD, which associate with partially or completely processed Ig-α, respectively (8, 32). Several possibilities have been considered to explain the difference in glycosylation of the Ig-α polypeptide associated with IgD, including the length of the extracellular membrane proximal domain, the presence of interchain disulfide bonds in this region, and the transit time of the complex through the Golgi (33). These possibilities could also account for the different processing of the IgE-associated Ig-α polypeptides, since, analogous to IgD, the m_LIgE contains a much longer extracellular domain than the m_SIgE, has three extra cysteines in this region, and also has a lower rate of transport to the cell surface.

An unexpected observation in this study was the finding that both IgE receptors associate with two proteins (εBAP37 and εBAP41) that were not present in the endogenous IgM-BCR of the WEHI lymphoma. Their association with the IgE receptors appeared to be very specific, since they were coprecipitated with two different anti-IgE antisera. These proteins obviously differ from other BCR-associated proteins, such as BAP32 and BAP37 which are exclusively associated with IgM (39), and BAP29 and BAP31 which are preferentially associated with IgD molecules (40), for the following reasons: (a) they have an extracellular domain which was labeled in the cell-surface biotinylation experiment, (b) both proteins were glycosylated, and (c) they form disulfide-bonded complexes. The εBAP41 appeared to be underrepresented in the m_LIgE-BCR, suggesting that the longer extracellular membrane–proximal domain affects the association of εBAP41 to the BCR or its accessibility to biotinylation.

The most striking difference between the two IgEBCRs was related to the response elicited by their crosslinking. Both receptors appeared to activate the same PTKs, as judged by the pattern of tyrosine phosphorylation. However, a clear difference was noted in the maximal phosphorylation time point as well as in the duration of the response. More specifically, the phosphorylation induced via the m_LIgE-BCR reached its maximum after 1 min of receptor engagement and declined within the following 5 min. In contrast, the signal induced by cross-linking of the m_SIgE-BCR reached its maximum after 5 min and remained for a prolonged period. Differences in the kinetics of the response have also been observed for the IgM- and IgD-BCRs (31). Experiments with these receptors or receptor chimeras have indicated that the extracellular membrane proximal and/or transmembrane domains influence the duration of the response (31). Since in the IgE-BCRs the difference resides only in the extracellular membrane– proximal domain, it can be concluded that this region determines the kinetics of protein tyrosine phosphorylation upon receptor engagement. The mechanisms through which this could occur are unknown, but could be mediated through association and/or interaction with the different Ig-α glycoforms or other proteins of the BCR complex.

The different kinetics of signal transduction through the IgE-BCRs were associated with different cellular responses of the WEHI transfectomas. Cross-linking of the m_SIgEBCR led to growth inhibition, whereas signaling through the m_LIgE-BCR had no effect on cellular proliferation. These cellular responses are analogous to those observed after signaling through the IgM- and IgD-BCRs in WEHI cells (37). In this case the IgM-BCR, which, like m_SIgE, has a short extracellular spacer, is capable of transmitting a growth inhibitory signal.

In conclusion, we have shown that both human mIgE isoforms assemble into functional B cell receptors. When expressed on immature B cells such as the WEHI lymphoma, the two IgE-BCRs transmit qualitatively distinct signals after cross-linking. Work is in progress to determine the properties of these receptors in more mature B cells which have undergone Ig H chain class switching. However, it should be noted here that B cells coexpressing IgM and IgE have been detected in peripheral blood (41, 42). Clearly, it would be of interest to determine the type of mIgE expressed by these cells and whether the two mIgEBCRs can be coexpressed by the same cell. Such studies should further delineate the relevance of having two receptors with identical specificity for the ligand and may allow characterization of additional components of the BCR signal-transduction pathways.