Surface Ig (sIg) expression is a critical checkpoint during avian B cell development. Only cells that express sIg colonize bursal follicles, clonally expand, and undergo Ig diversification by gene conversion. Expression of a heterodimer, in which the extracellular and transmembrane domains of murine CD8α or CD8β are fused to the cytoplasmic domains of chicken Igα (chIgα) or Igβ, respectively (murine CD8α (mCD8α):chIgα + mCD8β:chIgβ), or an mCD8α:chIgα homodimer supported bursal B cell development as efficiently as endogenous sIg. In this study we demonstrate that B cell development, in the absence of chIgβ, requires both the Igα ITAM and a conserved non-ITAM Igα tyrosine (Y3) that has been associated with binding to B cell linker protein (BLNK). When associated with the cytoplasmic domain of Igβ, the Igα ITAM is not required for the induction of strong calcium mobilization or BLNK phosphorylation, but is still necessary to support B cell development. In contrast, mutation of the Igα Y3 severely compromised calcium mobilization when expressed as either a homodimer or a heterodimer with the cytoplasmic domain of Igβ. However, coexpression of the cytoplasmic domain of Igβ partially complemented the Igα Y3 mutation, rescuing higher levels of BLNK phosphorylation and, more strikingly, supporting B cell development.
Both murine and avian BCR complexes include Ig H and L chains in association with the Igαβ heterodimer. Signaling downstream of the BCR complex originates from the Igα and Igβ cytoplasmic domains, both of which contain ITAMs (1, 2). Cross-linking of the BCR complex leads to Src family kinase-mediated phosphorylation of the Igαβ ITAM tyrosines, promoting recruitment and activation of the kinase Syk, thereby initiating several downstream signaling pathways leading to B cell survival, differentiation, and proliferation (3, 4). In mice expressing a BCR complex that lacks the cytoplasmic domains of Igα and Igβ, B cell development is arrested at the pre-BI cell stage, indicating that the Igαβ heterodimer is required for B cell differentiation (5).
In the chicken, B cell development occurs in the bursa of Fabricius, a gut-associated lymphoid organ (6, 7). Nevertheless as in the mouse, surface Ig (sIg)3 expression is a critical checkpoint in avian B cell development. Only those B cell precursor cells that have undergone productive rearrangement at both the H and L chain loci and express a functional BCR complex at the cell surface colonize bursal follicles, expand, and diversify their Ig loci by gene conversion (8, 9). Subsequent corticomedullary redistribution and emigration to the periphery also require maintained sIg expression, because bursal cells losing sIg expression are eliminated by apoptosis (10).
In the absence of receptor ligation, both Src family kinases and Syk associate with the Igαβ heterodimer with low affinity and may support basal signaling (11, 12). It has been suggested that such basal signaling, initiated from the BCR complex in the absence of receptor ligation, is sufficient to support the development of B cells. To address this possibility, both murine and avian truncated μ receptor complexes (Tμ), which lack the V(D)J encoded determinants, but maintain the ability to associate with the Igαβ heterodimer, have been generated and have been shown to support early B cell development (13, 14, 15). Although these results suggested that there is no requirement for BCR ligation during the early stages of B cell development in mouse or in chicken, both murine and avian Tμ complexes included extracellular domains that could have been involved in receptor-ligand interactions. In this regard, the mammalian pre-BCR complex has been shown to interact with several stromal ligands (16, 17, 18, 19).
More recently, the cytoplasmic domains of the Igα/Igβ heterodimer have been targeted to the cell surface in the absence of any of the extracellular domains associated with the BCR complex. The cytoplasmic domains of murine Igα and Igβ, targeted together to the cell surface by a palmitoylation motif, are sufficient to generate immature B cells (20). In the chick we have used the extracellular and transmembrane domains of murine CD8α and CD8β to target the cytoplasmic domains of chicken Igα (chIgα) and chIgβ to the cell surface. Using retroviral gene transfer in vivo, we have shown that the cytoplasmic domains of the Igαβ heterodimer are as efficient as endogenous μ in supporting the early stages of B cell development (21). As such, it can be concluded that basal signals generated after surface expression of the Igαβ cytoplasmic domains are sufficient to support B cell differentiation.
We have also reported that surface targeting of the cytoplasmic domain of Igα alone, in the absence of Igβ, can support B cell development in the chick embryo. Moreover, we demonstrated that both calcium mobilization and support of B cell development require the Igα ITAM tyrosines (21, 22). However, the cytoplasmic domain of Igα is highly conserved in regions outside the ITAM (23), and it has remained unclear whether such regions are also required for B cell differentiation. In particular, a third tyrosine, C-terminal to the Igα ITAM, that we have designated Y3 is conserved and has been suggested to recruit B cell linker protein (BLNK; also known as SLP-65 or BASH) to the BCR complex (24, 25). BLNK is an adaptor protein involved in the recruitment of Grb2/Sos, phospholipase Cγ2a (PLCγ2a) and Vav to the BCR complex (26, 27, 28), and as a consequence promotes viability, proliferation, and activation. Deletion of BLNK in the DT40 B cell line has a profound effect on signaling downstream of the BCR complex, including the lost ability of the BCR complex to induce calcium mobilization (29). The importance of BLNK in signaling downstream of the BCR complex directly correlates with its critical role during B cell development in mammals. BLNK knockout mice display a partial block in B cell development, resulting in a severely diminished pool of peripheral mature B cells (30, 31, 32, 33).
Phosphorylation of non-ITAM tyrosines in Igα has been implicated as crucial for BLNK recruitment to the BCR complex (24). Recruitment of BLNK to the BCR complex and its subsequent phosphorylation lead to its association with both PLCγ2a and Btk, thereby promoting PLCγ2a activation (29, 34). Activation of PLCγ2a results in the production of inositol triphosphate, which ultimately results in the release of calcium from intracellular stores and the influx of extracellular calcium (35). Not surprisingly, ablation of PLCγ2a in the DT40 bursal cell line results in a complete loss of calcium mobilization after sIg receptor ligation (36). Similarly, PLCγ2a knockout mice have a reduced number of peripheral mature B cells due to a block of pre-B cell differentiation (37), suggesting a critical role for PLCγ2a in B cell development.
To directly determine the relative importance of the Igα ITAM and conserved Y3 residue, we have assessed the ability of receptor mutants to support signaling after their ligation in vitro and to support B cell development after expression in vivo. Mutations of both the Igα ITAM and the Igα Y3 residue were found to compromise signaling downstream of the Igα cytoplasmic domain, correlating with the inability of the mutated Igα cytoplasmic domain to support B cell development in the absence of Igβ. In the context of the Igαβ heterodimer, however, the Igα ITAM is not required for calcium mobilization or BLNK phosphorylation, whereas it is absolutely required for B cell development. As such, the Igβ ITAM cannot functionally compensate for the disrupted Igα ITAM during B cell development. Unlike the Igα ITAM, the Igα Y3 residue remains crucial for a robust calcium response induced by the Igαβ heterodimer. Strikingly, however, the Igαβ heterodimer does not require the Igα Y3 residue to support B cell development. We have therefore demonstrated that although the Y3 residue may be involved in signaling downstream of the BCR complex, the function(s) of the Igα Y3 residue required for early B cell development can also be provided by the cytoplasmic domain of Igβ.
Materials and Methods
Site-directed mutagenesis of murine CD8α (mCD8α):chIgα
The mCD8α:chIgα, mCD8α:chIgβ, and mCD8β:chIgβ chimeric constructs were generated previously (21). Point mutations of the Igα cytoplasmic domain were introduced using the QuikChange Site-Directed Mutagenesis kit (Stratagene), which involves amplification of a parental vector with complimentary reverse oligonucleotides in which the mutation (underlined) is incorporated. The mCD8α:chIgαF1F2 mutant was generated by amplification of the Cla12L-mCD8α:chIgαF2 plasmid (21) with the ααF1(5′), GGAGAACCTCTTTGAGGGCCTGGATTTGG, and ααF1(3′), CCAAATCCAGGCCCTCAAAGAGGTTCTCC primers. The mCD8α:chIgαF3 mutant was generated by the amplification of Cla12L-mCD8α:chIgα plasmid (21) with ααF3(5′), CGCCCGCAGCCCACCTTTGAGGACG, and ααF3(3′), CGTCCTCAAAGGTGGGCTGCGGGCG primers. The mutations were confirmed by sequencing (York University). Mutated mCD8α:chIgα sequences were excised by ClaI digestion and cloned into the unique ClaI site of replication-competent avian leukosis virus with splice acceptor (RCAS)-Bryan polymerase (BP)A (38).
Retroviral gene transfer
The RCAS vectors are derived from the Rous sarcoma virus and are organized as follows: long terminal repeat (LTR)-gag-pol-env-splice acceptor-cloning site-LTR. Expression of mCD8:chIg genes, inserted into the cloning site, is driven by transcription from the 5′ LTR promoter and subsequent splicing of transcripts (38).
Line O chicken embryo fibroblasts (CEFs; Regional Poultry Research Laboratories) were transfected with the RCAS-mCD8:chIg vectors as previously described (13). One week after transfection, CEFs were tested for surface expression of the mCD8:chIg chimeric proteins and injected into line 22 chick embryos (Charles River Laboratories) on day 3 of embryogenesis.
The DT40 frameshift (sIg−) cell line (provided by Dr. J. M. Buerstedde, GSF, Institute for Molecular Radiobiology, Neuherberg-Munich, Germany) was infected with the various RCAS-mCD8:chIg retroviruses. Surface Ig− DT40 cells (1 × 104) were cocultured with 1 × 106 appropriately transfected CEFs in a final volume of 2 ml of IMDM/2% chicken serum containing 8 μg/ml 1,5-dimethyl-1,5-diazaundecamethylene polymethobromide, hexadimethrine bromide (Sigma-Aldrich), for 24 h. The infected nonadherent DT40 cells were allowed to expand and were then sorted based on surface expression of mCD8α or mCD8β using the FACSAria (BD Biosciences).
Abs and flow cytometry
Surface expression of the chimeric proteins in vivo and in vitro was detected using the anti-mCD8α (53-6.72) and anti-mCD8β (53.8.84) Abs (provided by Dr. P. Hugo, PROCREA BioSciences). Cell suspensions of bursal cells generated at the time of hatching were also stained for the pan B cell marker ChB6 and μ as previously described (13).
DT40 cell stimulation, immunoprecipitation, and Western blotting
Before all stimulations, samples were prewarmed to 37°C for 2 min. Total cell lysates were generated from 1 × 106 cells stimulated with either anti-mCD8β (53-5.8.84) or anti-mCD8.2α (D9; provided by Dr. P. Hugo) for 2 min. Reactions were stopped by the addition of ice-cold phosphate buffer. Cells were then pelleted and resuspended in SDS loading dye, boiled, and electrophoresed. After transfer onto nitrocellulose, membranes were probed with the biotinylated anti-phosphotyrosine Ab 4G10 (39) and developed with streptavidin-coupled HRP (Southern Biotechnologies). Membranes were then stripped and probed for actin (AC-40; Sigma-Aldrich) as described previously (21).
The mCD8:chIg homodimers and heterodimers were immunoprecipitated from cell lysates after either Ab cross-linking or pervanadate stimulation. In the case of Ab cross-linking, 10 × 106 cells were resuspended in 500 μl of IMDM and were stimulated with either anti-mCD8β (53-5.8.84) or anti-mCD8.2α (D9) for 2 min. For pervanadate simulation, 2 × 106 cells were resuspended in 100 μl of IMDM to which 100 μl of prewarmed 50 μM pervanadate (prepared as described previously (40)) was added. After stimulation, cells were washed in ice-cold phosphate buffer, pelleted, and lysed in 1% detergent Nonidet P-40 buffer (1% Nonidet P-40, 150 mM NaCl, and 10 mM Tris, pH 7.5) containing phosphatase and protease inhibitors for 1 h at 4°C. Lysates were then centrifuged, and the supernatant was added to anti-mCD8α (53-6.7.2)-coupled cyanogen bromide-activated Sepharose beads (Amersham Biosciences) or to anti-chBLNK (provided by Dr. T. Kurosaki, Institute for Liver Research, Kansai Medical University, Moriguchi, Japan)-coupled protein A-Sepharose beads (Sigma-Aldrich) and incubated overnight at 4°C. Washed beads were resuspended in SDS loading dye and boiled for 5 min. The supernatant was then electrophoresed, transferred to nitrocellulose (Amersham Biosciences), and probed with 4G10. Membranes were subsequently stripped and probed for total BLNK using the anti-chBLNK Ab. Quantification of bands was performed by scanning densitometry.
Calcium mobilization assays
All calcium fluxes were performed as previously described, but were analyzed on FACSAria (BD Biosciences) (41).
Igα ITAM is required for Igαβ-mediated B cell development
Expression of the mCD8α:chIgα chimeric protein either alone (Fig. 1,A) or together with mCD8β:chIgβ (Fig. 1,B) supports all the early stages of chicken B cell development, including colonization of bursal follicles, oligoclonal expansion of B cells within bursal follicles, and onset of repertoire diversification by gene conversion (21). Site-directed mutagenesis of both Igα ITAM tyrosine residues to phenylalanine (mCD8α:chIgαF1F2) abolished the ability of Igα to support B cell development in the absence of Igβ (22) (Fig. 1 C). Although mCD8α:chIgαF1F2 is expressed on the surface of B-lineage B cells at similar levels to the mCD8α:chIgα protein, all mCD8α:chIgαF1F2-expressing B cells coexpressed endogenous μ. This confirms previous findings with the mCD8α:chIgαF2 receptor containing a single tyrosine to phenylalanine mutation in the Igα ITAM (21), lending additional support to our conclusion that signaling downstream of the Igα ITAM is required to support chicken B cell development.
Nonetheless, the endogenous BCR complex includes both Igα and Igβ, each of which contains an ITAM motif involved in BCR signaling (1, 2). Therefore, although the expression of an mCD8α:chIgβ chimeric protein alone does not support B cell development (21, 22), we addressed the possibility that functional motifs in the cytoplasmic domain of Igβ could complement the ITAM-deficient mCD8α:chIgαF1F2 and provide sufficient signals to support B cell development. Chicks were therefore coinfected with RCAS(BP)A-mCD8α:chIgαF1F2 and RCAS(BP)B:mCD8β:chIgβ viruses. High levels of mCD8α:chIgαF1F2 and mCD8β:chIgβ coexpression were observed in 11 neonatal chicks analyzed. However, all B-lineage bursal cells in RCAS(BP)A-mCD8α:chIgαF1F2 + RCAS(BP)B:mCD8β:chIgβ-infected chicks were μ+ (Fig. 1 D). Therefore, expression of the cytoplasmic domain of Igβ does not complement the ITAM deficiency of the mCD8α:chIgαF1F2 protein, and consequently, the Igα ITAM plays a unique role in supporting B cell development.
Igα ITAM is not required for Igαβ induction of strong calcium mobilization
The mCD8α:chIgαF1F2 chimeric protein failed to support the induction of calcium mobilization when expressed in sIg− DT40 B lymphoma cells and cross-linked with anti-CD8α Abs (Fig. 2, A and B). This is consistent with our previous finding that the single tyrosine to phenylalanine mutant mCD8α:chIgαF2 chimeric protein failed to support calcium mobilization (21). We have also reported that the cytoplasmic domain of Igβ, when expressed as a mCD8α:chIgβ chimeric homodimer, fails to support calcium mobilization in DT40 cells (21).
To assess the capacity of the mCD8α:chIgαF1F2 + mCD8β:chIgβ heterodimer to support calcium mobilization in DT40 cells, the sIg− DT40 cell line was coinfected with RCAS(BP)A-mCD8α:chIgαF1F2 + RCAS(BP)B:mCD8β:chIgβ viruses and sorted based on surface expression of the heterodimer (Fig. 2,A). Cross-linking of the mCD8α:chIgαF1F2 + mCD8β:chIgβ heterodimer with anti-mCD8α (Fig. 2 B) or anti-mCD8β (data not shown) Abs led to a rapid rise in calcium, which was comparable in duration and amplitude to that with either the mCD8α:chIgα homodimer or the mCD8α:chIgα + mCD8β:chIgβ heterodimer. Therefore, although the Igα ITAM provides unique signals required for B cell development in vivo, it is not required for the induction of strong calcium mobilization in vitro when expressed in conjunction with the cytoplasmic domain of Igβ.
Compelling evidence implicates BLNK phosphorylation as critical in coupling BCR ligation to strong calcium mobilization (42). It was therefore of interest to determine the status of BLNK phosphorylation after cross-linking of either the mCD8α:chIgαF1F2 homodimer or the mCD8α:chIgαF1F2 + mCD8β:chIgβ heterodimer. After stimulation with anti-CD8 Abs, BLNK was immunoprecipitated from cell lysates, and its state of tyrosine phosphorylation was assessed by Western blotting. Whereas cross-linking the unmodified mCD8α:chIgα homodimer resulted in strong BLNK phosphorylation, this was not seen after cross-linking the mCD8α:chIgαF1F2 homodimer (Fig. 2 C). Ligation of the mCD8α:chIgαF1F2 + mCD8β:chIgβ heterodimer, however, resulted in phosphorylation of BLNK. Thus, BLNK can be phosphorylated in the absence of the Igα ITAM, thereby providing a possible means by which the mCD8α:chIgαF1F2 + mCD8β:chIgβ heterodimer can support a robust calcium mobilization.
Thus, we have demonstrated that when the cytoplasmic domain of Igα is expressed independently of Igβ, the Igα ITAM is required for BLNK phosphorylation and calcium mobilization in vitro as well as for B cell development in the chick embryo. However, when the cytoplasmic domain of Igα is expressed in association with the cytoplasmic domain of Igβ, the Igα ITAM is dispensable for BLNK phosphorylation and strong calcium mobilization, but remains required for B cell development.
Igβ cross-linking can induce BLNK phosphorylation independently of Igα
In addition to the ITAM tyrosines, the cytoplasmic domain of chicken Igα contains a third conserved tyrosine residue, which we have designated Y3. Phosphorylation of the murine equivalent, Y204, has been implicated in the association of Igα with BLNK (24, 25). Not surprisingly, cross-linking the mCD8α:chIgαF1F2 homodimer did not result in phosphorylation of the Y3 residue (Fig. 2,D), consistent with the lack of BLNK phosphorylation and subsequent calcium mobilization. Surprisingly, however, cross-linking of the mCD8α:chIgαF1F2 + mCD8β:chIgβ heterodimer also did not result in Igα Y3 phosphorylation (Fig. 2,E) despite high levels of calcium mobilization and BLNK phosphorylation. Thus, when the heterodimer was cross-linked with anti-mCD8α Abs and immunoprecipitated with anti-mCD8β Abs, no phosphorylation of mCD8α:chIgαF1F2 was observed. In contrast, the mCD8β:chIgβ was heavily phosphorylated after heterodimer cross-linking (Fig. 2 E). Equivalent results were obtained when the heterodimer was cross-linked with anti-mCD8β and immunoprecipitated with anti-mCD8α-coupled beads (data not shown). Therefore, in the absence of the Igα ITAM, the Y3 residue is not phosphorylated after cross-linking of the mCD8α:chIgαF1F2 + mCD8β:chIgβ heterodimer.
We verified that we could detect phosphorylation of the Y3 residue by stimulating mCD8α:chIgαF1F2 + mCD8β:chIgβ-expressing DT40 cells with pervanadate, a potent phosphatase inhibitor. A 2-min stimulation with 25 μM pervanadate led to tyrosine phosphorylation of both mCD8α:chIgαF1F2 and mCD8β:chIgβ (Fig. 2, D and E).
Given the lack of tyrosine phosphorylation of the Igα Y3 residue after cross-linking of the mCD8α:chIgαF1F2 + mCD8β:chIgβ heterodimer, we addressed the possibility that the cytoplasmic domain of Igβ could play a role in the phosphorylation of BLNK. Initially, DT40 cells expressing the mCD8α:chIgβ chimeric protein (Fig. 3,A) were stimulated with cross-linking Abs, after which the chimeric protein was immunoprecipitated. Western blotting for protein tyrosine phosphorylation demonstrated that Ab cross-linking results in the phosphorylation of the cytoplasmic domain of Igβ (Fig. 3 B), consistent with evidence we have previously reported demonstrating that the cytoplasmic domain of Igβ, when expressed in the absence of the cytoplasmic domain of Igα, can imitate significant protein tyrosine phosphorylation (21).
We show in this study that the cytoplasmic domain of Igβ is also sufficient to support the induction of BLNK phosphorylation (Fig. 3,C). After Ab cross-linking of the mCD8α:chIgβ chimeric protein, total BLNK was immunoprecipitated from generated cell lysates. Western blotting for the tyrosine phosphorylation clearly demonstrated that after receptor cross-linking, the cytoplasmic domains of both Igα and Igβ can independently induce BLNK phosphorylation (Fig. 3,C). These results provide a plausible rationale for the induction of BLNK phosphorylation and calcium mobilization by the mCD8α:chIgαF1F2 + mCD8β:chIgβ heterodimer in the absence of phosphorylation of the Y3 residue of mCD8α:chIgαF1F2. Nonetheless, the inability of the mCD8α:chIgβ chimeric receptor to mobilize calcium (Fig. 3,D) despite significant levels of BLNK phosphorylation (Fig. 3 C) clearly demonstrates that although BLNK phosphorylation may be necessary for calcium mobilization, it is not sufficient to support a calcium response.
F3 mutation alters signaling capacity of mCD8α:chIgα in DT40 cells
The ability of the mCD8α:chIgαF1F2 + mCD8β:chIgβ heterodimer to induce the phosphorylation of BLNK and a robust calcium response after Ab cross-linking, despite the absence of Y3 phosphorylation, led us to assess the importance of the Y3 residue in Igα-mediated signaling and B cell development. To this end, the Y3 residue was mutated to phenylalanine by site-directed mutagenesis (mCD8α:chIgαF3) to yield a domain that retains the Igα ITAM.
The sIg− DT40 bursal lymphoma was infected with RCAS(BP)A-mCD8α:chIgαF3 alone or was coinfected with both RCAS(BP)A-mCD8α:chIgαF3 and RCAS(BP)B-mCD8β:chIgβ retroviruses. The F3 mutation did not affect surface expression of the chimeric protein, because the expression levels of the mCD8α:chIgαF3 mutant were comparable to those of the mCD8α:chIgα chimeric protein when expressed singly or in association with the mCD8β:chIgβ, and cells were subsequently sorted based on surface expression of the chimeric receptors (Fig. 4 A).
Stimulation of mCD8α:chIgαF3-expressing DT40 cells with the anti-mCD8α Ab resulted in a calcium flux of markedly reduced amplitude compared with that observed after cross-linking of mCD8α:chIgα. Cross-linking the mCD8α:chIgαF3 + mCD8β:chIgβ heterodimer with anti-mCD8α (Fig. 4 B) or anti-mCD8β (data not shown) initiated a similarly rapid, but reduced, calcium flux compared with that observed after cross-linking of mCD8α:chIgα + mCD8β:chIgβ.
Cross-linking of either the mCD8α:chIgαF3 homodimer with anti-mCD8α or the mCD8α:chIgαF3 + mCD8β:chIgβ heterodimer with anti-mCD8β resulted in an increase in total protein phosphorylation (Fig. 5 A). Although qualitative and quantitative differences were observed after stimulation of cells expressing mCD8α:chIgαF3 compared with those expressing mCD8α:chIgα, this, nonetheless, supports the conclusion that substantial signaling occurs downstream of the Igα and/or Igβ ITAMs.
Given the reduction in calcium mobilization in the F3 mutants, we verified the state of BLNK phosphorylation after ligation of mCD8α:chIgαF3 either alone or in combination with mCD8β:chIgβ. When expressed in the absence of Igβ, the mCD8α:chIgαF3 protein initiated a significant increase in BLNK tyrosine phosphorylation (Fig. 5 B) albeit at reduced levels (∼50%, as detected by scanning densitometry) compared with mCD8α:chIgα. Moreover, when expressed in association with the cytoplasmic domain of Igβ, the F3 mutation did not compromise the extent of BLNK phosphorylation compared with the wild-type Igα domain. Specifically, scanning densitometry did not reveal significant differences between BLNK phosphorylation after cross-linking mCD8α:chIgα +mCD8β:chIgβ and that after cross-linking mCD8α:chIgαF3 +mCD8β:chIgβ. We can conclude, therefore, that the cytoplasmic domain of Igβ can complement the Y3 to F3 mutation in the cytoplasmic domain of Igα with respect to the level of BLNK phosphorylation, but not calcium mobilization. As a consequence, we can also conclude that the level of calcium response after receptor cross-linking is not quantitatively limited by the level of BLNK phosphorylation.
After cross-linking of either mCD8α:chIgαF3 or CD8α:chIgαF3 +mCD8β:chIgβ with the appropriate Abs, the chimeric proteins were immunoprecipitated and probed for tyrosine phosphorylation. The mCD8α:chIgαF3 is inducibly tyrosine phosphorylated whether expressed as a homodimer or a heterodimer with mCD8β:chIgβ (Fig. 5 C). Therefore, we can conclude that that the Igα F3 mutation does not inhibit Igα ITAM phosphorylation. Similarly, the levels of mCD8β:chIgβ phosphorylation indicate that the Igα F3 mutation does not alter the capacity of mCD8β:chIgβ to become inducibly tyrosine phosphorylated in the heterodimer.
Murine CD8α:chIgαF3 does not support B cell development
To determine whether the Y3 residue is required for Igα-supported B cell development, chick embryos on day 3 of embryogenesis were infected with RCAS(BP)A-mCD8α:chIgαF3. As observed in neonatal chicks infected with RCAS(BP)A-mCD8α:chIgα, bursae isolated from RCAS(BP)A-mCD8α:chIgαF3-infected chicks were of normal size and cellularity and contained normal numbers of B cells identified as ChB6+. In 13 RCAS(BP)A-mCD8α:chIgαF3 chicks analyzed, 30–80% of bursal B cells expressed mCD8α:chIgαF3, which reflects a viral penetrance comparable to that observed in RCAS(BP)A-mCD8α:chIgα-infected chicks (Fig. 6 A).
In the RCAS(BP)A-mCD8α:chIgαF3-infected chicks, bursal B cells expressed the mutant mCD8α:chIgα F3 chimeric protein at levels comparable to those observed on bursal cells expressing mCD8α:chIgα. However, despite the high levels of mCD8α:chIgαF3 expression, all mCD8α:chIgαF3-positive B cells coexpressed μ (Fig. 5 A). Therefore, in addition to the Igα ITAM, the Igα Y3 residue plays a critical role in the ability of the Igα cytoplasmic domain to support B cell development when expressed in the absence of the cytoplasmic domain of Igβ.
Murine CD8α:chIgαF3 in association with mCD8β:chIgβ supports B cell development
Having shown that the Igα ITAM is necessary for B cell development, we assessed whether the Y3 residue is also required for B cell development in the context of heterodimer expression. Chick embryos were therefore coinfected with RCAS(BP)A-mCD8α:chIgαF3 and RCAS(BP)B-mCD8β:chIgβ. Neonatal chicks contained a population of μ− cells, all of which expressed both mCD8α and mCD8β. In 12 chicks expressing high levels of both mCD8α and mCD8β, 8–35% of bursal B cells were μ−, mCD8α+ and mCD8β+ (Fig. 6 B). In such chicks, all μ− cells expressed both mCD8α:chIgαF3 and mCD8β:chIgβ, confirming the previous conclusion that the expression of mCD8α:chIgαF3 alone was not sufficient to support chicken B cell development.
Therefore, the cytoplasmic domain of Igβ can complement the Y3 to F3 mutation in the cytoplasmic domain of Igα with respect to its ability to support B cell development. B cell development, therefore, requires an intact Igα ITAM motif as well as a conserved non-Igα ITAM tyrosine that can be provided by either the Y3 residue of Igα or the cytoplasmic domain of Igβ.
We have used retroviral infection of chick embryos with RCAS-mCD8:chIg viruses to target the cytoplasmic domains of chicken Igα and Igβ, either independently or in combination, to the cell surface of B cell precursors in vivo. Because retroviral transduction of B cell precursors is not 100%, cells expressing the chimeric mCD8:chIg compete with cells expressing the endogenous BCR complex for oligoclonal colonization of bursal follicles. Therefore, within each individual chick the efficiency of chimeric mCD8:chIg receptors to support B cell development can be directly compared with the efficiency of the endogenous BCR complex (21).
Using this system, we have previously demonstrated that expression of the mCD8α:chIgα homodimeric receptor is sufficient to support bursal follicle colonization and B cell proliferation with an efficiency indistinguishable from that of the endogenous BCR complex. It was therefore concluded that surface expression of the cytoplasmic domain of Igα alone, in the absence of overt ligation, is sufficient to support B cell development (21). In direct contrast, the mCD8α:chIgβ chimeric receptor does not support B cell development and indeed, in the absence of the cytoplasmic domain of Igα, selectively inhibits B cell differentiation (22).
In this report we have demonstrated that the Igα ITAM tyrosines are uniquely required for the Igαβ heterodimer to support B cell development. In the absence of Igβ, the IgαF1F2 mutant failed to support B cell development. Moreover, the IgαF1F2 mutant failed to support B cell development even when associated with the unmodified Igβ. Therefore, the Igβ ITAM in combination with the IgαF1F2 cytoplasmic domain is not sufficient to activate signaling pathways required to support B cell development. These results lend additional support to the conclusion that Igα activates distinct signaling pathways compared with Igβ (2, 43, 44, 45) and that the Igβ ITAM is therefore not functionally equivalent to the Igα ITAM.
Our observation that mCD8α:chIgαF1F2 + mCD8β:chIgβ does not support development is based on the lack of μ−mCD8α+ bursal B cells. We have previously shown that mCD8α:chIgα mediates allelic exclusion, albeit incompletely (21). Importantly, among μ−mCD8α+ B cells that also contain V(D)J rearrangements, the majority of rearrangements were out of frame. As a consequence, if mCD8α:chIgαF1F2 + mCD8β:chIgβ failed to mediated allelic exclusion, one would also expect the majority of rearrangements to be similarly out of frame. It follows, therefore, that cells containing such rearrangements would be μ− and be revealed as a μ−mCD8α+ bursal cell population. The absence of such a population allows us to conclude that the lack of μ−mCD8α+ cells is a direct consequence of the inability of mCD8α:chIgαF1F2 + mCD8β:chIgβ to support B cell development.
Our results are in contrast to murine in vivo studies that suggest that the Igα ITAM is not uniquely required for B cell development. Specifically, in the IgαFF/FF knockin mouse, in which the IgαFF is equivalent to our mCD8α:chIgαF1F2, normal numbers of pro-, pre-, and immature B cells are observed, whereas the recirculating B cell population is only slightly reduced (46). The discrepancy between the conclusions made in mice and those made in this study in chickens may reflect a difference in BCR signaling requirements during B cell development in the mouse vs the chicken. Alternatively, it is possible that a BCR complex in which the Igα ITAM has been compromised is less efficient than the wild-type BCR complex in supporting B cell development. If such were the case, chicken B cells expressing the mCD8α:chIgαF1F2 + mCD8β:chIgβ chimeric heterodimer would be competed out by cells expressing the endogenous BCR complex. In contrast, in the IgαFF/FF knockin mouse, all B cells express the mutated Igα and as such are not subject to competition with a wild-type BCR complex.
The amplitude and duration of calcium signals discriminate between the activation of different transcription factors, such as NF-κB, NF-AT, and JNK, thereby influencing whether a cell undergoes further differentiation, proliferation, or apoptosis (47, 48). Signaling downstream of Igα and Igβ results in distinct patterns of calcium mobilization. Although Igα induces an initial release of calcium from intracellular stores, followed by the influx of extracellular calcium, signaling downstream of Igβ only initiates transient oscillatory releases of calcium from intracellular stores detected at the individual cell level (43). The mCD8α:chIgα chimera supported an influx of extracellular calcium equivalent to that seen after cross-linking of the BCR, whereas mCD8α:chIgβ did not initiate a calcium flux detectable at the population level (21). These observations suggest that a receptor competent to induce calcium mobilization when cross-linked in vitro may be required to support the development of chicken B cells in vivo.
Despite the fact that ligation of either mCD8α:chIgαF1F2 or mCD8α:chIgβ alone failed to result in detectable calcium mobilization, coexpression of mCD8α:chIgαF1F2 with mCD8β:chIgβ resulted in a receptor complex capable of supporting a strong calcium response. The membrane-proximal events that lead to calcium mobilization remain to be fully elucidated. However, phosphorylation of the adaptor protein BLNK is required for the induction of calcium mobilization by the BCR complex (42), and the presence or the absence of BLNK phosphorylation shown in this study after cross-linking of the mCD8α:chIgαF1F2 + mCD8β:chIgβ heterodimer or the mCD8α:chIgαF1F2 homodimer, respectively, is consistent with this requirement.
As would be expected, phosphorylation of the non-ITAM Y3 residue was not observed after cross-linking the mCD8α:chIgαF1F2 homodimer. In addition, cross-linking the mCD8α:chIgαF1F2 homodimer failed to induce BLNK phosphorylation. This is consistent with a model in which Syk recruitment to a phosphorylated ITAM induces phosphorylation of the Igα Y3 residue, leading to the recruitment and phosphorylation of BLNK, Also consistent with this model, cross-linking the mCD8α:chIgαF1F2 + mCD8β:chIgβ heterodimer leads to BLNK phosphorylation, and one could predict that Syk is recruited to the Igβ ITAM, leading to phosphorylation of the Igα Y3 residue initiating BLNK recruitment and phosphorylation.
However, several lines of evidence described in this study are not consistent with this simple model. Cross-linking of mCD8α:chIgαF1F2, when expressed as a heterodimer with mCD8β:chIgβ, failed to induce Y3 phosphorylation despite high levels of Igβ ITAM phosphorylation. Thus, phosphorylation of the Igβ ITAM is not sufficient to support Igα Y3 phosphorylation (Fig. 2 E). Nevertheless, stimulation of DT40 cells expressing the mCD8α:chIgαF1F2 + mCD8β:chIgβ heterodimer resulted in substantial BLNK phosphorylation and subsequent calcium mobilization. Therefore, phosphorylation of the Igα Y3 residue is not an absolute requirement for BLNK phosphorylation and calcium mobilization. It is, therefore, likely that the Igαβ heterodimer can induce BLNK phosphorylation through alternative means.
In murine Igα, both Y204 and Y176 residues have been implicated in promoting BLNK recruitment and phosphorylation (24, 25). Although the murine Y176 residue is not conserved in the chicken, where it is replaced by glycine, the Y204 residue and sequences surrounding Y204 are highly conserved (23). Given that cross-linking of the mCD8α:chIgαF1F2 + mCD8α:chIgβ heterodimer on the surface of DT40 cells resulted in calcium mobilization in the absence of phosphorylation of the Igα Y3 residue, a residue(s) within the cytoplasmic domain of the avian Igβ may contribute to BLNK phosphorylation. In this regard, we demonstrate in this study that the cytoplasmic domain of Igβ can independently induce BLNK phosphorylation after receptor cross-linking despite published evidence that the cytoplasmic domain of Igβ cannot physically recruit BLNK. At this point it is unclear whether BLNK is directly recruited to the chimeric receptors that contain the IgαF3 mutation. It is reasonable to expect that recruitment to the receptor is required for BLNK phosphorylation. However, given that BLNK itself associates with a wide variety of binding partners, it remains possible that BLNK recruitment to receptors containing the F3 mutation may be indirect. Under such circumstances, we would predict that BLNK could be recruited to either the Igα cytoplasmic domain (mCD8α:chIgαF3) or the Igβ cytoplasmic domain (mCD8α:chIgαF3 + mCD8β:chIgβ), consistent with our demonstration that the mCD8α:chIgβ chimeric receptor mediates BLNK phosphorylation.
The inability of the mCD8α:chIgαF1F2 + mCD8β:chIgβ heterodimer to support B cell development did not address the possibility that the Igα Y3 residue was also required. Therefore, we assessed the requirement for the Igα Y3 residue in B cell development. Expression of the mCD8α:chIgαF3 homodimer in chick B cell precursors did not support B cell development in the absence of endogenous μ. Thus, although expression of a receptor containing the Igα ITAM is required to support B cell development, the Igα ITAM by itself is not sufficient.
However, when expressed in conjunction with mCD8β:chIgβ, the Igα Y3 residue is not required to support B cell development. Thus, in RCAS(BP)A-mCD8α:chIgαF3 + RCAS(BP)B-mCD8β:chIgβ-infected chicks, populations of mCD8α+/mCD8β+/μ− cells were identified. Thus, a residue(s) in the cytoplasmic domain of Igβ can functionally replace the Igα Y3 residue in B cell development.
However, although the level of BLNK phosphorylation after cross-linking the mCD8α:chIgαF3 + mCD8β:chIgβ heterodimer is markedly increased compared with that seen after cross-linking the mCD8α:chIgαF3 homodimer, the calcium flux induced by cross-linking this heterodimer was indistinguishable from that seen after ligation of the mCD8α:chIgαF3 homodimer. Therefore, although calcium mobilization requires BLNK phosphorylation, it must also be limited by signals downstream of the BCR complex distinct from the BLNK phosphorylation seen in this study.
BLNK couples BCR ligation to several distinct signaling pathways. In addition to the PLCγ2a/Btk calcium pathway, BLNK interacts with the Grb2/Sos, leading to MAPK activation, and the Vav pathway, promoting the activation of the Rho family ofGTPases. Evidence indicates that this is accomplished by multiple docking sites on BLNK (42). In particular, phosphorylation on distinct sites is responsible for allowing BLNK to recruit PLCγ2a and Btk. As a consequence, it is possible that after cross-linking of the mCD8α:chIgαF3 + mCD8β:chIgβ heterodimer, although BLNK is heavily phosphorylated, it may not be phosphorylated on all residues required for efficient coupling to calcium mobilization. Nonetheless, the expression of this heterodimer is sufficient to support B cell development, suggesting that signals required for strong calcium mobilization are dispensable for B cell development in the chick embryo.
It is possible, therefore, that depending on the means by which it is recruited to the Igαβ heterodimer, BLNK becomes phosphorylated on distinct tyrosine residues and as a result potentiates the activation of distinct downstream signaling pathways. Thus, in the context of the mCD8α:chIgαF3 + mCD8β:chIgβ heterodimer, BLNK may be selectively phosphorylated, allowing coupling to pathways critical for supporting B cell development, but not efficient coupling to pathways required for strong calcium mobilization.
The difference between the signaling capacity of a receptor in vitro and its ability to support B cell development in vivo may reflect intrinsic differences between DT40 cells and primary B cells. Nevertheless, DT40, a chicken bursal B cell lymphoma, clearly shares many signaling characteristics of primary B cells, and BCR-mediated signaling in DT40 deletion mutants frequently parallels inhibited B cell development when the homologous molecule(s) is deleted in mice (3).
Through the comparison of DT40 cells and primary B cells isolated from RCAS-mCD8:chIg-infected chicks, we have shown that the ability of a receptor complex to support B cell development in vivo is independent of its ability to support strong calcium mobilization in vitro. These observations indicate that despite a crucial role for calcium mobilization in development, there is also a requirement for coupling of the Igαβ heterodimer to other signaling pathways during B cell development.
We also provide data regarding the key residues required during B cell development. The Igα ITAM is required for the Igα cytoplasmic domain to support B cell development, either independently or in combination with the cytoplasmic domain of Igβ. Such a requirement argues that the Igα ITAM has a crucial role in the activation of critical signaling pathways distinct from those activated downstream of Igβ. Nevertheless, the Igα ITAM by itself is not sufficient to support B cell development; rather, an additional residue(s) is required to activate complimentary signaling pathways. In an apparent redundancy of function, either the Igα Y3 residue or residues in the cytoplasmic domain of Igβ are sufficient to activate such signals.
We thank Gisele Knowles for expert assistance with flow cytometry, and Trista Murphy for Ab preparation. We are grateful to Dr. T. Kurosaki for providing us with the anti-chBLNK Ab, and to Dr. P. Hugo for providing the anti-CD8 Abs.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by Canadian Institutes of Health Research Grant MT10040 (to M.J.H.R.). K.A.P. was supported by the Fonds de la Recherche en Santé du Québec FRSQ-FCAR-Santé doctoral research bursary.
Abbreviations used in this paper: sIg, surface Ig; BLNK, B cell linker protein; BP, Bryan polymerase; CEF, chicken embryo fibroblast; chIg, chicken Ig; LTR, long terminal repeat; mCD8, murine CD8; PLCγ2a, phospholipase Cγ2a; RCAS, replication-competent avian leukosis virus with splice acceptor; Tμ, truncated Ig μ-chain.