Cooperativity and Segregation of Function within the Ig- (cid:97) / (cid:98) Heterodimer of the B Cell Antigen Receptor Complex*

The B cell antigen receptor complex contains het- erodimers of Ig- (cid:97) and Ig- (cid:98) . The cytoplasmic tails of each of these chains contain two conserved tyrosines, phos- phorylation of which initiates the signal transduction cascades activated by the receptor complex. Although the cytoplasmic domains of Ig- (cid:97) and Ig- (cid:98) have been expressed individually and demonstrated to be competent signal transduction units, we postulated that within the context of a heterodimer, Ig- (cid:97) and Ig- (cid:98) could have new, complementary or even synergistic functions. Therefore we developed a system to compare the signal transducing capacities of dimers of Ig- (cid:97) /Ig- (cid:97) , Ig- (cid:98) /Ig- (cid:98) , or Ig- (cid:97) / Ig- (cid:98) . This was done by fusing the extracellular and transmembrane domains of either human platelet-de- rived growth factor receptor (PDGFR) (cid:97) or (cid:98) to the cytoplasmic tail of either Ig- (cid:97) or Ig- (cid:98) . Three cell lines expressing PDGFR (cid:98) /Ig- (cid:97) , PDGFR (cid:98) /Ig- (cid:98) , or PDGFR (cid:97) / Ig- (cid:98) together with PDGFR (cid:98) /Ig- (cid:97) were established in the murine B cell line A20 IIA1.6. Biotech Inc.) coupled to a peptide corresponding to the murine Ig- (cid:97) cytoplasmic tail ITAM (amino acid residues 177–196) (34), whereas the polyclonal anti-Ig- (cid:98) antibody was purified over a column containing the immunizing Ig- (cid:98) fusion protein. The anti-phosphoty- rosine monoclonal antibodies FB2 and Ab2 were obtained from ATCC (Rockville, MD) and Oncogene Sciences (Uniondale, NY), respectively. Cell Growth and Stimulation— For most experiments, cells were grown in Iscove’s modified Dulbecco’s medium (Sigma) supplemented with 10% fetal calf serum (HyClone, Logan UT), 2 m M glutamine, 100 units/ml penicillin, and 100 (cid:109) g/ml streptomycin at 37 °C in 7.5% CO 2 . For the experiments Fig. cells were serum-starved (0.5% fetal calf serum) for 18 h before initiation of each experiment. For all the stimulation experiments described, aliquots of 10 (cid:51) 10 6 cells were suspended in 300 (cid:109) l of Iscove’s modified Dulbecco’s medium and incu- bated at 37 °C for 5 min. To stimulate cells via the endogenous BCR, cells were incubated with a rabbit polyclonal anti-IgG antibody (Jack- son Immunoresearch, West Grove, PA) at 15 (cid:109) g/ml for the times indicated in each experiment. To stimulate transfectants through the chi- mera, cells were incubated with PDGF-BB ligand (100 ng/ml, where noted) (Sigma) for 5 min, followed by anti-PDGFR (cid:98) antibody (5 (cid:109) g/ml, noted) (Genzyme) for 3 min, then anti-IgG 1 antibody (5 (cid:109) g/ml) (Jackson Immunoresearch).

The B cell antigen receptor complex contains heterodimers of Ig-␣ and Ig-␤. The cytoplasmic tails of each of these chains contain two conserved tyrosines, phosphorylation of which initiates the signal transduction cascades activated by the receptor complex. Although the cytoplasmic domains of Ig-␣ and Ig-␤ have been expressed individually and demonstrated to be competent signal transduction units, we postulated that within the context of a heterodimer, Ig-␣ and Ig-␤ could have new, complementary or even synergistic functions. Therefore we developed a system to compare the signal transducing capacities of dimers of Ig-␣/Ig-␣, Ig-␤/Ig-␤, or Ig-␣/ Ig-␤. This was done by fusing the extracellular and transmembrane domains of either human platelet-derived growth factor receptor (PDGFR) ␣ or ␤ to the cytoplasmic tail of either Ig-␣ or Ig-␤. Three cell lines expressing PDGFR␤/Ig-␣, PDGFR␤/Ig-␤, or PDGFR␣/ Ig-␤ together with PDGFR␤/Ig-␣ were established in the murine B cell line A20 IIA1.6. While aggregation of each dimer by itself could induce the tyrosine phosphorylation of cellular substrates, only aggregation of the heterodimer induced the phosphorylation of substrates similar in range and intensity to that induced by the endogenous B cell antigen receptor complex. Interestingly, Ig-␤ remarkably enhanced the rapidity (T max decreased from 5 to 1 min) and intensity (greater than 10-fold enhancement) of Ig-␣ phosphorylation. Conversely, the phosphorylation of Ig-␤ was reduced to undetectable levels when co-aggregated with Ig-␣. The enhancement of Ig-␣ phosphorylation by Ig-␤ correlated with a lowering of the stimulation threshold for tyrosine kinase activation.
A B cell's response to antigen, whether it be proliferation, differentiation, anergy, or deletion, is dependent upon recognition of that antigen by the B cell antigen receptor (BCR) 1 (1)(2)(3). The receptor is a multimeric complex consisting of the antigenrecognition substructure, membrane-bound immunoglobulin non-covalently associated with heterodimer(s) of Ig-␣ and Ig-␤ (4 -6). Present evidence indicates that the cytoplasmic tails of Ig-␣ and ␤ (7) translate antigen engagement into cytoplasmic signaling events that initiate cellular responses (8 -12). Most proximally in the signaling cascade, one or more tyrosine kinases, including Syk and members of the Src family, are activated (13)(14)(15). These in turn activate a variety of pathways whose constituents include Ras, phosphatidylinositol 3-kinase, and phospholipase C (1). Embedded within the cytoplasmic tails of both Ig-␣ and ␤ is a sequence common to other multichain immune recognition receptor (MIRR) subunits including CD3␦, CD3␥, TCR, Fc␥RIII␥, and Fc⑀RI␥, termed the immunoreceptor tyrosine-based activation motif (ITAM) (16,17). The motif contains two tyrosines, both of which are critical for initiating tyrosine kinase activation (10,18). Phosphorylation of these tyrosines facilitates the recruitment and activation of tyrosine kinases which contain SH2 domains, such as Syk and Fyn (19 -22). Substrates for these kinases may also be recruited (22). The presence of the ITAM in all MIRR chains involved in signal transduction has led some to suggest that apparently heterologous chains such as CD3⑀ and TCR are functionally redundant and the presence of multiple ITAMs within each MIRR serve to increase the strength of signal which can be generated via the receptor. Evidence for this assertion has been obtained in studies of both the B and T cell antigen receptors (8,12,23,24). In contrast, we and others have provided evidence indicating that each heterologous ITAM containing chain has a distinct function (10,11,19,25,26).
Many of the above studies utilized chimeras in which irrelevant extracellular and transmembrane domains were fused to the single cytoplasmic domain under study (18,27,28). Although this approach has yielded considerable insight into ITAM-containing chains, it assumes that functions observed in the isolated circumstance of a single chimera are reflective of the function of that cytoplasmic domain within the intact receptor complex. This might not be true since most ITAM-containing chains, such as Ig-␣ and Ig-␤, are expressed on cell surfaces as heterodimers (29 -32). Therefore, we postulated that within the context of a heterodimer, Ig-␣ and Ig-␤ would have new, complementary or even synergistic functions, not predicted from studies of single chain chimeras.
As demonstrated in this report, Ig-␣ and ␤ have new and unpredicted functions in the context of a heterodimer. Using a novel chimera system, which allowed us to form either heteroor homodimers of the cytoplasmic domains of Ig-␣ and Ig-␤, we observed that when Ig-␤ is ligated independently it is able to activate tyrosine kinases. However, when co-aggregated with Ig-␣, Ig-␤ appears to remarkably enhance Ig-␣ phosphorylation. This in turn correlates with an increase in the range and intensity of cellular substrates phosphorylated by the heterodimer and a lowering of the stimulation threshold for tyrosine kinase activation.

MATERIALS AND METHODS
Construction of PDGFR/Ig-␣/␤ Chimeras-The construction and expression of the chimeras has been described in detail elsewhere. 2 Briefly, cDNAs encoding the PDGFR␣ and ␤ chains (gift of A. Kazlauskas, National Jewish Center, Denver, CO) were mutated to introduce BamHI and EcoRI sites immediately after that portion of each cDNA which encodes the transmembrane domain. The introduction of these sites facilitated the insertion of cDNA fragments encoding the cytoplasmic domains of Ig-␣ and Ig-␤ (25). These fragments were assembled with an EcoRI/XhoI-flanked cDNA fragment containing multiple stop codons in pSK (Stratagene, La Jolla, CA), which were then subcloned into the expression vector pCB6 ϩ /muTk (gift of H. Singh, University of Chicago, Chicago, IL) which contains a neomycin resistance gene, an IgM enhancer and a thymidine kinase promoter. The cDNAs encoding the chimeric constructs PDGFR␤/Ig-␣ and PDGFR␤/Ig-␤ were transfected either separately or together into A20 IIA1.6 (33) by electroporation. Clones were derived by selection with G-418 and stained with anti-PDGFR␣ and anti-PDGFR␤ antibodies (Genzyme, Cambridge, MA) then FITC-conjugated anti-IgG 1 (Zymed, San Francisco, CA). They were analyzed by flow cytometry (FACScan, Becton Dickinson, Bedford, MA).
Reagents-Polyclonal anti-Ig-␣ and anti-Ig-␤ antibodies were made by immunizing rabbits (HTI Bioproducts, Ramona, CA) with glutathione S-transferase fusion proteins containing the cytoplasmic domains of murine Ig-␣ or Ig-␤ (25). The serum of rabbits immunized with the Ig-␣ fusion protein was purified over a column (CNBr-activated Sepharose, Pharmacia Biotech Inc.) coupled to a peptide corresponding to the murine Ig-␣ cytoplasmic tail ITAM (amino acid residues 177-196) (34), whereas the polyclonal anti-Ig-␤ antibody was purified over a column containing the immunizing Ig-␤ fusion protein. The anti-phosphotyrosine monoclonal antibodies FB2 and Ab2 were obtained from ATCC (Rockville, MD) and Oncogene Sciences (Uniondale, NY), respectively.
For the experiments described in Fig. 6, cells were serum-starved (0.5% fetal calf serum) for 18 h before initiation of each experiment. For all the stimulation experiments described, aliquots of 10 ϫ 10 6 cells were suspended in 300 l of Iscove's modified Dulbecco's medium and incubated at 37°C for 5 min. To stimulate cells via the endogenous BCR, cells were incubated with a rabbit polyclonal anti-IgG antibody (Jackson Immunoresearch, West Grove, PA) at 15 g/ml for the times indicated in each experiment. To stimulate transfectants through the chimera, cells were incubated with PDGF-BB ligand (100 ng/ml, except where noted) (Sigma) for 5 min, followed by anti-PDGFR␤ antibody (5 g/ml, except where noted) (Genzyme) for 3 min, then anti-IgG 1 antibody (5 g/ml) (Jackson Immunoresearch).

Construction, Expression, and Stimulation of PDGFR Chimeras-To examine if
Ig-␣ and Ig-␤ may function together to initiate pathways of cellular activation, we designed a system using the human PDGFRs, which allowed us to form either homo-or heterodimers of Ig-␣ and ␤. Two forms of PDGFR exist, ␣ and ␤, each of which is recognized by specific monoclonal antibodies. Furthermore, although distinct, each has an equal affinity for the naturally occurring ligand, PDGF-BB (35). Therefore, each chain can be expressed independently, yet made to form homodimers or predominantly heterodimers on singly or doubly transfected cells, respectively, by the addition of PDGF-BB. These dimers, which are representative of the resting complex (PDGF-BB does not induce tyrosine kinase activation; data not shown), can then be activated by specific antibodies (Fig. 1A).
The Cytoplasmic Domains of Both Ig-␣ and Ig-␤ Are Needed to Induce the Efficient Tyrosine Phosphorylation of Cellular Proteins-We first asked if the cytoplasmic tails of Ig-␣ or Ig-␤ alone, or the two chains together, were capable of inducing the tyrosine phosphorylation of cellular proteins in a manner similar to that induced via the endogenous BCR. Chimeras were stimulated by sequential incubation with PDGF-BB, followed by anti-PDGFR␤ antibody and rabbit anti-mouse IgG1. In parallel samples, the endogenous BCR on each transfectant was stimulated with polyclonal antibodies to IgG. After stimulation, cells were lysed and phosphotyrosine immunoprecipitates (with FB2) from each lysate were resolved by SDS-PAGE and analyzed by blotting with anti-phosphotyrosine antibodies (Ab2). As shown in Fig. 2, stimulation of the BCR on wild type and transfected cells induced a similar spectrum and intensity of tyrosine phosphorylation. In contrast, only in ␣/Ig-␤//␤/Ig-␣ cells did stimulation of chimeras induce tyrosine phosphorylation that was similar in distribution and intensity to that induced by the endogenous antigen receptor. In ␤/Ig-␣ or ␤/Ig-␤ cells, stimulation of the chimeras could only induce the strong tyrosine phosphorylation of a subset of proteins. In related experiments, truncated co-expressed versions of each chimera which lacked cytoplasmic domains, PDGFR␣/Ϫ and PDGFR␤/Ϫ, were incapable of inducing any detectable tyrosine phosphorylation. 2 However, differences were observed in the induction via the chimeras in ␣/Ig-␤//␤/Ig-␣ cells and by the endogenous BCR. Stimulation of ␣/Ig-␤//␤/Ig-␣ failed to induce the tyrosine phosphorylation of proteins of 32 and 40 kDa. Subsequent immunoblotting revealed that the 32-kDa protein was Ig-␣ (data not shown). The 40-kDa protein appears to be a novel molecule which is associated with the endogenous Ig-␣/␤ heterodimer. 3 These observations suggest that the chimeras do not utilize the BCR, or associated structures, to initiate tyrosine phosphorylation. Finally, at least one protein of 120 -130 kDa was phosphorylated strongly in ␣/Ig-␤//␤/Ig-␣, weakly in ␤/Ig-␣ and ␤/Ig-␤, and not at all by stimulation of the BCR. Since the chimeras are predicted to have molecular masses of approximately this size, we examined if this protein was a chimeric molecule.
Stimulation of Chimeric Heterodimers Induces the Tyrosine Phosphorylation of PDGFR␤/Ig-␣ but Not PDGFR␣/Ig-␤-The wild type A20 IIA1.6 and ␣/Ig-␤//␤/Ig-␣ cells were treated with PDGF-BB and then stimulated with anti-receptor antibodies as above. Anti-phosphotyrosine, anti-Ig-␣, or anti-Ig-␤ immunoprecipitates were resolved by SDS-PAGE and probed with antibodies of the same specificity in various combinations. Immunoblotting of the Ig-␣ and Ig-␤ immunoprecipitates with the same antibodies confirmed that ␣/Ig-␤//␤/Ig-␣ cells but not wild type cells expressed PDGFR␤/Ig-␣ (135 kDa) and PDGFR␣/Ig-␤ (125 kDa) (Fig. 3). Although both chimeras were expressed in readily detectable amounts, only PDGFR␤/Ig-␣ was observed to be phosphorylated following stimulation. The phosphorylation of PDGFR␤/Ig-␣ was detected in both Ig-␣ and Ig-␤ immunoprecipitations, the latter presumably being a result of co-liga-tion during stimulation. The PDGFR␤/Ig-␣ chimera co-migrated with a prominent tyrosine phosphoprotein precipitated from the lysates of stimulated cells. No tyrosine phosphoprotein corresponding to the PDGFR␣/Ig-␤ chimera was observed, even on overexposed gels (data not shown). From these results, it is readily apparent that PDGFR␤/Ig-␣, but not PDGFR␣/ Ig-␤, was inducibly phosphorylated in ␣/Ig-␤//␤/Ig-␣ cells. However, it is not clear if the differences in phosphorylation observed are due to differences intrinsic to each chain or due to interactions between the chains. To differentiate between these possibilities, we compared PDGFR␤/Ig-␣ and PDGFR/Ig-␤ phosphorylation in singly and doubly transfected cells.
The Phosphorylation of Ig-␤ Was Extinguished in the Presence of Ig-␣-We first examined the influence of Ig-␣ on Ig-␤ phosphorylation by comparing PDGFR/Ig-␤ phosphorylation in ␤/Ig-␤ and ␣/Ig-␤//␤/Ig-␣. The chimeras were precipitated from the lysates of unstimulated cells (us) or cells stimulated with FIG. 1. A, stimulation of PDGFR chimeras. Homodimers or predominantly heterodimers were formed on singly (left) or doubly transfected (middle) cells, respectively, by first treating with PDGF-BB. Complexes were then aggregated with anti-PDGFR␤ antibodies, followed by goat anti-mouse antibodies. Omission of PDGF-BB before stimulating doubly transfected cells led to the aggregation of PDGFR␤/Ig-␣ alone (right panel). B, schematic representation of PDGFR chimeras. cDNAs encoding the PDGFR␣ and ␤ chains were mutated to introduce BamHI and EcoRI sites immediately after that portion of each cDNA encoding the transmembrane domain (Tm). The introduction of these sites facilitated the insertion of cDNA fragments encoding the cytoplasmic domains of Ig-␣ and Ig-␤. Assembled cDNAs were cloned into the expression vector pCB6ϩ/muTk, which contains a neomycin resistance gene, an IgM enhancer and a thymidine kinase promoter. C, flow cytometric analysis of A20 IIA1.6 cells expressing PDGFR chimeras. Expression of either surface IgG, PDGFR␣ or PDGFR␤ on wild type, ␤/Ig-␣, ␤/Ig-␤, and ␣/Ig-␤//␤/Ig-␣ cells. To stain for surface IgG, 2 ϫ 10 5 cells/sample from each cell line were incubated with FITC-conjugated anti-IgG at 4°C. For chimera staining, 2 ϫ 10 5 cells/sample from each cell line were incubated with anti-PDGFR␣ or anti-PDGFR␤ antibodies and subsequently FITC-conjugated anti-IgG1 at 4°C. Also shown is staining of each cell line without primary antibody. As demonstrated by immunoprecipitation and immunoblotting, cells with equal staining intensity with anti-PDGFR␣ and anti-PDGFR␤ antibodies expressed approximately equal amounts of each protein (Fig. 3).  Fig. 2. Lysates from these cells were immunoprecipitated with FB2, anti-Ig-␣, or anti-Ig-␤ antibodies. Immunoprecipitates were resolved by 7.5% SDS-PAGE, transferred to nylon membrane, and then probed with Ab2, anti-Ig-␣, or anti-Ig-␤ antibodies.
PDGF-BB ligand and anti-PDGFR␤ antibody followed by rabbit anti-mouse IgG 1 for 1, 2, or 5 min. A representative experiment is shown in Fig. 4 (n ϭ 4). In the ␤/Ig-␤ cell line, stimulation of the chimera induced its own phosphorylation, which was maximal at 2 min and transient (Fig. 4, upper panel). In contrast, phosphorylation of PDGFR␣/Ig-␤ was not detected in ␣/Ig-␤//␤/Ig-␣ cells, even though more PDGFR␣/Ig-␤ protein was immunoprecipitated (Fig. 4, lower panel). These results suggest that when expressed in isolation, Ig-␤ can be phosphorylated. However, when ligated in the presence of Ig-␣, its own phosphorylation is inhibited. These observations are in accordance with the minimal inducible tyrosine phosphorylation of Ig-␤ observed following BCR ligation in A20 (Fig. 5C). Figs. 3 and 4, it is apparent that only PDGFR␤/Ig-␣ is tyrosine-phosphorylated following chimera stimulation and its phosphorylation is strong only when co-ligated with PDGFR/Ig-␤. To examine this further, we compared the inductive tyrosine phosphorylation of PDGFR␤/Ig-␣ when expressed alone (␤/Ig-␣) or with PDGFR␣/ Ig-␤ (␣/Ig-␤//␤/Ig-␣). The cell lines ␤/Ig-␣ and ␣/Ig-␤//␤/Ig-␣ were stimulated via the chimeras, lysed, and then immunoprecipitated with a combination of anti-Ig-␣ and anti-Ig-␤ antibodies. Shown in Fig. 5 are the results of a representative experiment (n ϭ 3). Immunoprecipitates were resolved by SDS-PAGE, and after transfer to membrane, probed first with Ab2 (Fig. 5A, upper panel), then stripped and reprobed with a combination of anti-Ig-␣ and Ig-␤ antibodies (lower panel). There were remarkable differences in both the degree and kinetics of PDGFR␤/Ig-␣ phosphorylation with and without PDGFR/Ig-␤ co-ligation. In the absence of PDGFR/Ig-␤, the tyrosine phosphorylation of PDGFR␤/Ig-␣ was weak at 1 min, increased to a maximum at 5 min, and was essentially absent at 30 min. In contrast, in the presence of PDGFR/Ig-␤, PDGFR␤/Ig-␣ phosphorylation was maximal at 1 min and thereafter decreased, being almost undetectable at 30 min. To analyze this data further, we densitometrically quantitated the immunoreactivity of each sample with Ab2 and anti-Ig-␣ antibodies and then plotted the ratio of these values as a function of time. As seen in Fig. 5B, co-ligation of PDGFR/Ig-␤ enhanced the phosphorylation of PDGFR␤/Ig-␣ approximately 12-fold at 1 min. This intensity and rapidity of tyrosine phosphorylation is similar to what is observed for endogenous Ig-␣ following ligation of the BCR (Fig. 5C).

Ig-␤ Enhances the Phosphorylation of Ig-␣-In
It is possible that the observed differences in Ig-␣ phosphorylation with and without co-cross-linking Ig-␤ were due to differences intrinsic to those cells in which both chains were expressed. To address this possibility, we examined if the degree of phosphorylation of PDGFR␤/Ig-␣ was influenced by PDGFR␣/Ig-␤ co-cross-linking on the same cell line. Therefore, we studied ␣/Ig-␤//␤/Ig-␣ cells under two different stimulating conditions. First, following the stimulating protocol described above, which includes PDGF-BB, heterodimers were formed and then aggregated in ␣/Ig-␤//␤/Ig-␣. In parallel, cells were Cell lysates were immunoprecipitated with combination of anti-Ig-␣ and anti-Ig-␤ antibodies. Immunoprecipitates were resolved by 7.5% SDS-PAGE, transferred to nylon membrane, and probed with Ab2. The immunoblot was then stripped and reprobed with combination of anti-Ig-␣ and anti-Ig-␤ antibodies. B, quantitation of the enhancement of Ig-␣ phosphorylation by Ig-␤. Immunoreactivities from the immunoblot in A were quantitated densiometrically. The specific tyrosine phosphorylation of Ig-␣ was calculated as: (immunoreactivity of sample to Ab2/immunoreactivity of sample to anti-Ig-␣) ϫ 100. This value was plotted as a function of time. C, the induction of Ig-␣ and Ig-␤ tyrosine phosphorylation following BCR stimulation in wild type A20 IIA1.6. 10 ϫ 10 6 cells/sample of wild type cells were stimulated with rabbit antimouse antibodies. Cells were then lysed, and the lysates were immunoprecipitated with FB2. The immunoprecipitates were resolved by 10% SDS-PAGE, transferred to nylon membrane, and probed with Ab2. The position of Ig-␣ and Ig-␤, which was determined by probing parallel samples with antibodies to these molecules (data not shown) is indicated.
Co-aggregation of PDGFR␤/Ig-␣ and PDGFR␣/Ig-␤ Lowered the Threshold for Tyrosine Kinase Activation-Given the remarkable enhancement of Ig-␣ phosphorylation by Ig-␤, we postulated that the heterodimeric complex would be more efficient in its ability to activate tyrosine kinases than a homodimeric complex. Therefore, we compared the stimulation threshold for the induction of tyrosine phosphorylation in ␣/Ig-␤//␤/Ig-␣ and ␤/Ig-␣ by titering out the primary stimulating antibody. Cells were otherwise stimulated and analyzed as in Fig. 2. As can be seen in Fig. 7, stimulating both transfectants with high concentrations of primary stimulating antibody induced the tyrosine phosphorylation of cellular proteins. However, the tyrosine phosphorylation of cellular proteins in ␤/Ig-␣ diminished significantly after the first 4-fold dilution of anti-PDGFR␤ antibody. In contrast to ␣/Ig-␤//␤/Ig-␣, the tyrosine phosphorylation was still detected after two dilutions of anti-PDGFR␤ antibody. Similar results were obtained when ␤/Ig-␤ was compared to ␣/Ig-␤//␤/Ig-␣ (data not shown). These results suggest that the heterodimeric structure of the BCR complex facilitates B cell responses to low doses of antigen. DISCUSSION Herein we report that the ITAM-containing subunits within an immune recognition receptor complex can cooperate to efficiently initiate signal transduction cascades. Using a system that allowed us to compare the signal transduction capacities and physical properties of homo-and heterodimers of Ig-␣ and Ig-␤, we observed that the Ig-␣/␤ heterodimer induced the phosphorylation of a wider range of substrates at a lower threshold of stimulation than either homodimer. This synergy correlated with the ability of Ig-␤ to enhance the tyrosine phosphorylation of Ig-␣ by more than 10-fold. Conversely, in the presence of Ig-␣, Ig-␤ phosphorylation was extinguished to undetectable levels. These data suggest that one of the major functions of Ig-␤ is to enhance the phosphorylation and, therefore, the signal transducing capability of Ig-␣. Furthermore, these data suggest that significant "cross-talk" or cross-modulation occurs between the subunits of the B cell antigen receptor.
Our results reveal a new level of complexity in the function of the BCR. Previous studies utilizing either fusion proteins, peptides or chimeras, have sought to characterize the functional capacities of individual cytoplasmic domains of the BCR complex (18,19,27,28). This reductionist approach assumes that each ITAM-containing domain is an isolated signaling unit whose capacities can simply be added to those of other domains to form an accurate picture of the whole receptor complex. Our data suggest that this assumption is not entirely valid. Rather, we would argue that while the study of each individual subunit reveals what it can do, it is only in the context of other receptor structures that one can elucidate what that subunit does do.
We have recently obtained data directly demonstrating that the coordinate activities of Ig-␣ and Ig-␤ is of biological significance. We have established clones of WEHI 231, an immature B cell sensitive to apoptosis, expressing similar combinations of the chimeras described here. When the chimeras in these transfectants were stimulated, we observed that the induction of apoptosis required the cytoplasmic tails of both Ig-␣ and Ig-␤. In those experiments, as in the experiments described in this report, only the heterodimerized chimeras induced tyrosine kinase activation efficiently. 2 Phosphorylation of the ITAM tyrosines within the BCR cytoplasmic domains is a necessary and early event in the initi-FIG. 8. Model of Ig-␣/Ig-␤-mediated signal transduction. Previously, models of antigen receptor signal transduction have assumed that the cytoplasmic domains of each receptor chain was an independent signal transduction unit (left). In contrast, our data support a model in which cooperation between receptor components leads to the enhancement of signals initiated by particular chains (right).
FIG. 6. Ig-␣ phosphorylation was enhanced when co-ligated with Ig-␤. 10 ϫ 10 6 cells/sample of ␣/Ig-␤//␤/Ig-␣ were stimulated through chimeras with and without PDGF-BB. Cells were then lysed in 1% Nonidet P-40 at different time points after stimulation. The cell lysates were immunoprecipitated with a combination of anti-Ig-␣ and anti-Ig-␤ antibodies. The immunoprecipitates were resolved by 7.5% SDS-PAGE, transferred, and probed with Ab2. The immunoblot was then stripped and reprobed with a combination of anti-Ig-␣ and anti-Ig-␤ antibodies.
FIG. 7. The Ig-␣/␤ heterodimer had a lower threshold of stimulation than the Ig-␣/␣ homodimer. 10 ϫ 10 6 cells/sample of ␤/Ig-␣ or ␣/Ig-␤//␤/Ig-␣ were stimulated through the chimeras as before except that the indicated decreasing concentrations of anti-PDGFR␤ antibody (serial 4-fold dilutions) were used. After stimulation cells were lysed in 1% Nonidet P-40 lysis buffer. Cell lysates were immunoprecipitated with FB2, and these were resolved by 10% SDS-PAGE, transferred, and probed with Ab2. ation of tyrosine kinase activation. Previously, we and others have demonstrated that members of the Src family of tyrosine kinases are constitutively associated with the resting receptor complex and that it is probably these kinases which mediate the phosphorylation of Ig-␣ and thereby initiate signaling by recruiting and activating SH2 domain containing secondary effectors (2,13,19,36). Which effectors are recruited by the receptor complex would be determined, in part, by which of the four tyrosine(s) (34,37) in the cytoplasmic domain of Ig-␣ are phosphorylated upon receptor engagement (38). From our data it is not clear if Ig-␤ merely enhances the phosphorylation at previously modified tyrosines or directs the phosphorylation of new sites. This distinction is of potential significance because if Ig-␤ directs the phosphorylation of Ig-␣, the spectrum of substrates activated by the receptor complex would be altered.
One of the mechanisms by which Ig-␤ could augment Ig-␣ phosphorylation would be to recruit novel kinases to the receptor complex. Previously, we found that phosphoproteins of 40 and 42 kDa bind in vitro to the cytoplasmic domain of Ig-␤ via a phosphotyrosine-independent QTAT sequence embedded within the Ig-␤ ITAM (19). Recently, we have demonstrated that similar molecules are inducibly tyrosine-phosphorylated by BCR engagement and are associated with the native Ig-␣/␤ heterodimer. 3 Alternatively, it is possible that kinases are constitutively associated with Ig-␤ because their SH2 domains bind a small subpopulation of phosphorylated Ig-␤ tails. While we did not observe any phosphorylation of Ig-␤ in our heterodimerized chimeras, there was a low level of Ig-␤ phosphorylation in the native BCR, which minimally increased following receptor engagement (Fig. 5C). It is possible that such phosphorylation of PDGFR␣/Ig-␤ occurred at a level too low to detect.
Another possibility is that Ig-␤ recruits SH2 domain-containing proteins, other than kinases, which bind to and protect Ig-␣ tyrosines from dephosphorylation. This is a plausible alternative since the tyrosine phosphatase CD45 is associated with the receptor complex and its function is necessary for BCR-mediated signal transduction.
Previously, models of antigen receptor-mediated signal transduction have assumed that each ITAM-containing chain within a receptor complex generates independent signals (Fig.  8, left). Some have postulated that the signals generated are redundant (12,23,24), while others have demonstrated that some chains are functionally distinct and capable of preferentially coupling to selected secondary effectors (10,11,19,25,26). However, our data support a model in which each chain within a heterodimer can cross-modulate the phosphorylative state and, by extension, the signal transducing capabilities of the other (Fig. 8, right). This leads to specialization of each chain's function within the multimeric whole. In the particular case of the BCR and tyrosine kinase activation, Ig-␤'s function appears to be to facilitate the phosphorylation of Ig-␣, which in turn allows the efficient activation of tyrosine kinases and possibly the recruitment of their substrates. The proof of this model will require an understanding of the mechanisms whereby Ig-␤ enhances Ig-␣ phosphorylation.