To determine the function of immunoglobulin (Ig)α immunoreceptor tyrosine–based activation motif (ITAM) phosphorylation, we generated mice in which Igα ITAM tyrosines were replaced by phenylalanines (IgαFF/FF). IgαFF/FFmice had a specific reduction of B1 and marginal zone B cells, whereas B2 cell development appeared to be normal, except that λ1 light chain usage was increased. The mutants responded less efficiently to T cell–dependent antigens, whereas T cell–independent responses were unaffected. Upon B cell receptor ligation, the cells exhibited heightened calcium flux, weaker Lyn and Syk tyrosine phosphorylation, and phosphorylation of Igα non-ITAM tyrosines. Strikingly, when the Igα ITAM mutation was combined with a truncation of Igβ, B cell development was completely blocked at the pro-B cell stage, indicating a crucial role of ITAM phosphorylation in B cell development.
B cell development is a highly ordered and controlled process allowing the formation of a diverse repertoire of B cell receptor (BCR) antigen specificities. Dependent on the developmental status and the microenvironment of the B cell, the strength and quality of BCR transmitted signals control survival, differentiation, and proliferation 1,2.
The BCR is a multimeric protein complex consisting of an antigen binding module that is noncovalently associated with one signal-transducing heterodimer of Igα (CD79a, mb-1) and Igβ (CD79b, B29) 3. An early and essential feature of the signaling cascade is phosphorylation of the tyrosine residues within immunoreceptor tyrosine–based activation motif (ITAM) by Src family protein tyrosine kinases (PTKs; references 4, 5, and 6). The phosphorylated ITAMs are thought to serve as docking sites for Src homology 2 (SH2) domain–containing proteins. For instance, dual phosphorylation of the BCR ITAMs allows high affinity association of the tandem SH2 domain–containing kinase Syk. Binding of Syk to ITAM leads to its phosphorylation and activation which in turn leads to phosphorylation of downstream cellular substrates. Igα and Igβ both contain single ITAMs that potentially confer distinct signaling functions. In support of this idea, several proteins have been shown to associate differentially with either Igα or Igβ 7,8. In transfection experiments using chimeric molecules, isolated cytoplasmic domains of either Igα or Igβ were both capable of mediating the mobilization of calcium and tyrosine phosphorylation of cellular molecules. These signaling capacities were critically dependent on ITAM phosphorylation. Further data suggest that Igα is predominant in activating PTKs 9,10,11,12,13,14,15,16.
Aspects of Igα and Igβ function in B cell development were studied in transgenic and gene-targeted mice. In Igβ knockout mice, the lack of Igα/Igβ heterodimer formation presumably also abrogates the surface expression of Igα and results in a complete block of development at the pro-B cell stage 17. However, the expression of chimeric molecules with cytoplasmic domains of either Igα or Igβ was capable of supporting further steps in B cell maturation, including H chain allelic exclusion 18,19,20. Critical roles of the cytoplasmic tails of Igα and Igβ in B cell development were revealed in mb-1Δc/Δc and IgβΔc/Δc mice 21,22. In IgβΔc/Δc mice, the cytoplasmic domain of Igβ is lacking, whereas in mb-1Δc/Δc mice (hereafter termed IgαΔc/Δc) a truncated Igα protein with a cytoplasmic domain consisting of 20 instead of 61 amino acids is expressed. In both mutants, B cell development is severely impaired resulting in a drastic reduction of peripheral B lymphocyte numbers. Functionally, the remaining B cells in the mutants are unable of mounting a measurable immune response against T cell–independent (TI) antigen. Further experiments with IgαΔc/Δc mice revealed that the Igα cytoplasmic truncation causes enhanced negative selection of autoreactive receptor specificities and generally renders immature B cells more sensitive to antigenic contact. Based on these data, an additional negative signaling function for Igα was suggested 23,24.
To investigate the specific function of Igα ITAM phosphorylation in B cell development, we generated a mutant mouse strain by gene targeting in which the Igα ITAM tyrosines were replaced with phenylalanines (IgαFF/FF mice).
Materials And Methods
Generation of Targeting Vector and Targeted Embryonic Stem Cell Clones.
The mb-1 targeting vector was designed as a replacement vector 25. A loxP site was inserted into the Tth111I site of intronIII. Intronic region and exons downstream of this loxP site are exactly repeated 3′ to a floxed neor gene (see Fig. 1 A). Tyrosine to phenylalanine replacements within the ITAM were achieved by site-directed mutagenesis (Tyr182: TAT to TTC; Tyr193: TAT to TTT). An additional BamHI restriction site within the 3′ UTR, downstream of the mutated exons, was generated by an exchange of two nucleotides.
A 5.2-kb genomic BamHI fragment, containing the mb-1 gene, was recovered from a murine cosmid (C57BL/6J-pWE15 library; Stratagene; reference 26) and subcloned into the BamHI site of pGEM-7Z+ (Promega) to generate pGEM5.2kbBamHI. Subsequently, a 1.4-kb Tth111I–XhoI fragment was recovered from this vector and subcloned into the EcoRI/XhoI sites of pGEMloxP (unpublished data) to obtain pGEMloxPTth111I-XhoI. Next, the Tth111I–PstI fragment of pGEM5.2kbBamHI was replaced with the NotI–PstI fragment from pGEMloxPTth111I-XhoI to obtain pGEM5.2kbloxP. The 5.2-kb BamHI fragment was then cloned into the BamHI site of pSL1190XX to obtain pSL5.2kbloxP. pSL1190XX was generated from pSL1190 (Amersham Pharmacia Biotech) by deleting the polylinker-region between XhoI and XbaI.
The region downstream of the mb-1 polyA site was extended 0.4 kb by the following procedure. A 1.2-kb fragment was first amplified by PCR from the genomic cosmid clone with the primers 3′ region-FOR (5′-CTGCCTCCCCACTCTTCC-3′) and 3′ region-REV (5′-CCATCGATCCTGCCGCGGCCGC-TTACTCT-3′). This PCR product was subcloned into the pCR™ TA Cloning vector II (Invitrogen), and a 1.2-kb EcoRI–XhoI fragment was recovered. This fragment was used to replace the 0.8-kb SmaI–XhoI fragment in pSL5.2kbloxP to obtain pSL5.2kbloxP-3′. Next, pSL5.2kbloxP-3′ was cut with EcoRV–XbaI and religated resulting in pSL5.2kbloxP-3′-ΔClaI. Subsequently, the tk selection marker was excised as a HindIII fragment from pIC-TKΔCla1, a modified version of pIC19R/MC1-TK 26. This fragment was cloned into the HindIII site of pSL5.2kbloxP-3′-ΔClaI to obtain pSL5.2kbloxP-3′-ΔClaI-TK. For the construction of a loxP-flanked neor cassette, a SalI–XhoI fragment was recovered from pMC1neoPolyA (Stratagene) and cloned into the XhoI site of pGEMloxP to obtain pGEMneoloxP. From this vector, an XbaI–HindIII fragment was subcloned into NotI–HindIII sites of pGEMloxP, resulting in pMMneoflox-8. A 1.2-kb XbaI–NsiI fragment from pMMneoflox-8 was cloned into Tth111I–NsiI–digested pGmb1C1C2** 21 to obtain pmb-1C1C2**neoflox. Additional restriction sites were introduced by site-directed mutagenesis. The product of two consecutive PCR reactions (primer: BamHI-FOR 5′-ATGT-CTGGATCCAGCATCCC-3′; BamHI-REV 5′-GGGATGCT-GGATCCAGACAT-3′; FLAG-C1C2**-FOR 5′-AACGGTGGCAAAATGAGAAG-3′; FLAG-C1C2**-REV 5′-TTGGGCCCCGATCGATCGCATG-3′) was subcloned into the pCR™ TA Cloning vector II (Invitrogen) and retrieved by digestion with SalI and ApaI. The fragment was used to exchange the corresponding 1-kb fragment in pmb-1C1C2**neoflox to generate pmb-1C1C2**neoflox-BamHI.
Culturing and transfection of E14.1 embryonic stem (ES) cells (129/Ola) was performed according to previously published protocols 27. Screening of 596 G418/gancyclovir double resistant colonies by Southern blot analysis led to the identification of three homologous recombinant ES cell clones (Fig. 1).
ES cells carrying the IgαFF-Flox allele were injected into C57BL/6 and CB20 blastocysts which were then transferred into foster mothers to obtain chimeric mice. Mice with the IgαFF-Flox allele were crossed to the deleter-strain 28. Some analyses were performed with IgαFF/FF mice of a mixed 129/Ola–C57BL/6 genetic background. To control for a potential influence of the genetic background (e.g., in immunization studies), control groups of 129/Ola and C57BL/6 mice were included and analyzed in parallel with the IgαFF/FF mice. In anti-phosphotyrosine immunoblotting experiments, all IgαFF/FF mice used had been backcrossed for eight generations to C57BL/6. IgβΔc/Δc mice 22 are of mixed 129/Sv–C57BL/6 genetic background. IgαΔc/Δc mice 21 are of C57BL/6 genetic background. IgαFF/FFIgβΔC/ΔC mice were generated by intercrossing IgαFF/FF mice to IgβΔC/ΔC mice 22. All mice, including IgαΔC/ΔC 21 and JHT/JHT mice 29, were housed in our conventional animal facility.
Preparation of Cell Suspensions from Lymphoid Organs.
Spleens were minced through a nylon mesh (cell strainer; Falcon) to obtain single cell suspensions in DMEM, 5% FCS, and 2 mM l-glutamine. Bones were flushed with medium to extract bone marrow (BM) cells and the peritoneal cavity was flushed twice with 10 ml of medium to recover cells. Erythrocytes were lysed from spleen and BM preparations by incubating in lysis buffer (140 mM NH4Cl, 17 mM Tris-HCl, pH 7.65) for 2 min on ice. To obtain PBLs, mice were bled from the tail vein in the presence of heparin (Liquemin; Roche) and purified by gradient centrifugation with 7% Ficoll 400 (Amersham Pharmacia Biotech).
Cells were surface stained with combinations of FITC, PE, Cy-Chrome (Cyc), peridinine chlorophyll protein (PERCP), and/or allophycocyanin (APC)-conjugated monoclonal antibodies for 20 min on ice. Stainings with biotinylated monoclonal antibodies were followed by a secondary staining with either Streptavidin-Cy-Chrome (BD PharMingen) or Streptavidin-PERCP (Becton Dickinson). After staining, the samples were washed and resuspended with PBS, 1% BSA, and 0.01% N3. Stained cells were acquired on a FACScan™ or FACSCalibur™ and data were analyzed using CELLQuest™ software (Becton Dickinson). Dead cells were labeled with propidium iodide or Topro-3 (Molecular Probes) and excluded from the analysis.
For intracellular Igμ chain staining, cells were first washed in PBS then fixed in a 2% formaldehyde solution for 30 min at room temperature (RT). After three washes with PBS, the cells were stained for 20 min at RT with antibody M41 in 0.05% Saponin in PBS. Monoclonal antibodies R33–24.12 (anti-IgM), 1.3–5 (anti-IgD), RA3–6B2 (anti-B220), M41 (anti-μ chain), LS.136 (anti-λ1), Cy34.1 (anti-CD22), M5/114 (anti-MHC class II), BP-1 (anti-BP-1), 30F1 (anti-HSA), and Cfo1 (anti-Thy1.2) were prepared and conjugated in our laboratory. Monoclonal antibodies to CD5, CD19, CD21/CD35, CD23, CD25, CD43, CD69, and CD86 were purchased from BD PharMingen. Anti-493 antibody was a gift from A. Rolink (Basel Institute of Immunology, Basel, Switzerland).
Immunization and Serum Analysis.
Mice were immunized with 5 or 50 μg of T cell–dependent (TD) antigen NP-CG (4-hydroxy-3-nitrophenylacetyl chicken-γ-globulin) in alum intraperitoneally. Mice were boosted with 5 μg/200 μl of NP17-CG without adjuvant. TI immune response was analyzed by intraperitoneal injections of 5 μg/200 μl of NP27-Ficoll. Mice were bled before and after immunization from tail veins. Ig serum concentrations were determined by ELISA as described previously 30.
Affinities of IgG1 and λ NP-specific antibodies were determined by calculating the association constant as described previously 31. Briefly, relative binding of serum antibodies to NP5-BSA versus NP14-BSA was determined by ELISA in comparison to anti-NP antibody standards of known affinity.
Measurement of Ca2+ Mobilization.
Spleen cells (5 × 106/ml) were loaded with 2 μM Indo-1 AM in RPMI, 2% FCS for 30 min at 37°C (Molecular Probes). Afterwards, the cells were washed and stained with anti–B220-PE (RA3–6B2) monoclonal antibody in RPMI at RT. Fluorescence ratios of Indo-1 emission at 395/510 nm were measured by flow cytometry on a dual laser FACStar™ (Becton Dickinson). Data acquisition was conducted for 50 s without stimulus, followed by the addition of either F(ab)′2 rabbit anti-mouse IgM+G+A fragments (Zymed Laboratories) or rabbit anti–mouse IgM+G+A antibodies (Zymed Laboratories). Data was collected for 512 s and analyzed using CELLQuest™ (Becton Dickinson) and FlowJo (TriStar) software.
Analysis of Protein Tyrosine Phosphorylation.
CD19+ B cell populations were purified from splenic cell suspensions by magnetic beads (MACS; Miltenyi Biotec) according to the manufacturer's instructions. B lymphocytes were resuspended at 107 cells/ml in RPMI and stimulated with F(ab)′2 rabbit anti–mouse IgM+ G+A fragments (Zymed Laboratories) for various time periods at 37°C. The reactions were stopped by addition of ice cold lysis buffer to a final concentration of 1% NP-40, 150 mM NaCl, 0.4 mM EDTA, 10 mM Tris-HCl, pH 8, 1 μg/ml leupeptin, 1 μg/ml aprotinin, 1 μg/ml α1-antitrypsin, 1 mM PMSF, 10 mM sodium pyrophosphate, 2 mM Na3VO4, and 10 mM NaF. After incubation on ice for 10 min, lysates were cleared at 12,000 g for 5 min at 4°C. Igα was immunoprecipitated from cleared lysates with anti-Igα monoclonal antibody 172 directly coupled to Sepharose beads. For the immunoprecipitation of Lyn or Syk, cleared lysates were incubated with either rabbit polyclonal anti–mouse Lyn (sc-15; Santa Cruz Biotechnology, Inc.) or rabbit polyclonal anti–mouse Syk antibodies 13, respectively; followed by addition of protein G-Sepharose beads (Amersham Pharmacia Biotech). Proteins were resolved by 10% SDS-PAGE, transferred to polyvinylidine difluoride (PVDF) membrane (Millipore), and immunoblotted with anti-phosphotyrosine antibody Py54 (Calbiochem), followed by horseradish peroxidase (HRP)-conjugated rat anti–mouse IgG (Amersham Pharmacia Biotech) and developed by chemiluminescence (ECL; Amersham Pharmacia Biotech). To determine the amount of protein loaded in each lane, the same membranes were stripped and reprobed with appropriate antibodies.
Generation of IgαFF/FF Mice.
To allow for conditional mutation of the Igα ITAM at different stages of B cell development, two loxP sites flanking both cytoplasmic exons were targeted into the mb-1 locus (floxed allele; Fig. 1 A, 4). Downstream of the 3′ end of mb-1 gene, mutated cytoplasmic exons encoding phenylalanine instead of tyrosine within the ITAM were inserted. Transfection of E14.1 ES cells led to the identification of three homologous recombinant ES cell clones (Fig. 1 A, 3, and Fig. 1 B). One clone (3-D7) was shown by PCR analysis to have cointegrated the 5′ loxP site and the ITAM mutations (data not shown). To remove the floxed neor selection cassette from the targeted mb-1 allele, the clone was transiently transfected with circular pICcre plasmid 29. Resulting G418-sensitive ES cell clones were screened by Southern blot (Fig. 1 C) and PCR analysis (not shown) to distinguish clones that carry the IgαFF-flox floxed allele (Fig. 1 A, 4) from clones where Cre-mediated recombination led also to the loss of the unmutated cytoplasmic exons (IgαFF mutated allele; Fig. 1 A, 5). ES cells carrying the IgαFF-flox allele were injected into mouse blastocysts to generate chimeras that subsequently transmitted the floxed allele through the germline. IgαFF-flox/+ mice were bred with the deleter mouse strain 28 to obtain IgαFF/+ mice that were then bred to homozygosity. The tyrosine to phenylalanine exchanges in the Igα ITAM were verified by sequencing. DNA for this analysis was prepared from FACS®-sorted splenic B cells of IgαFF/FF and wild-type control mice (Fig. 1 E).
B Cell Development in IgαFF/FF Mice.
The requirement of Igα ITAM phosphorylation for B cell development was analyzed in IgαFF/FF mice by flow cytometry 32. In contrast to IgαΔc/Δc mice, which have reduced fractions of pre-B and immature B cells, IgαFF/FF mice had normal numbers of pro-B (B220lowCD43+IgM−), pre-B (B220lowCD43−IgM−), and immature B (B220lowIgM+IgD−) cells in the BM (Fig. 2 A, and Table). Similarly, the earliest stages of B cell development within the pro-B cell compartment, fractions A to C′, were not blocked in IgαFF/FF mice (see Fig. 5). In addition, a striking difference exists between mature B cells in IgαΔc/Δc and IgαFF/FF mutants. Whereas recirculating B cells in the BM and splenic B cells were ∼100-fold reduced in the former, B cell numbers in IgαFF/FF mice were only marginally reduced in comparison to wild-type mice (Table; Fig. 2a and Fig. B). We analyzed the peripheral B cells in IgαFF/FF mice for the expression of several surface markers to determine their maturation and activation status. As in wild-type controls, splenic B cells from ITAM mutants were CD43− and <20% of them stained brightly for immature B cell marker 493 (data not shown). Normal surface expression of activation markers CD86 (B7.2), CD69 (data not shown), and MHC class II (Table) was observed.
Like the IgαΔc/Δc mutant 21, IgαFF/FF mice displayed allelic exclusion at the IgH locus as verified by the analysis of IgH allotype heterozygous mutant animals (unpublished data).
Alterations of Surface Receptor Expression Levels in IgαFF/FF Mice.
Although the lack of Igα ITAM phosphorylation apparently did not block the generation of a large pool of mature B cells, it caused changes in the surface level expression of BCR and several coreceptors. Fig. 2 C shows a comparison of the surface IgM levels in splenic B cells of wild-type, IgβΔc/Δc, and IgαFF/FF mice. In contrast to the extremely low surface IgM levels of B cells in IgβΔc/Δc mice, the Igα ITAM mutation led to increased surface IgM levels. In general, B cells from various lymphoid organs carrying the IgαFF allele express on average more surface IgM and IgD than B cells from control mice (Fig. 2, and Table). This applies also for B1 cells in the peritoneal cavity (Fig. 2 F, and Table). Furthermore, we consistently observed slightly higher surface expression levels of CD19 (Fig. 2 C, and Table) and lower surface levels of CD22 and B220 (Table) in B cells obtained from IgαFF/FF compared with wild-type mice.
Reduction of Marginal Zone and B1 B Cell Populations.
Are the differences in receptor expression levels sufficient to compensate for the Igα ITAM phosphorylation deficiency and to allow for the normal generation of all peripheral B cell subsets? Three different main B cell subpopulations have been described for the spleen: immature, follicular, and marginal zone (MZ) B cells. MZ B cells are large cells that are CD21brightCD23low and IgMhighIgDlow. Analysis of IgαFF/FF mice revealed that this population was reduced to 5–25% of wild-type controls (Fig. 2 D).
In the peritoneal cavity, B cell numbers, mainly CD5+ B1 cells, were drastically reduced in the Igα ITAM mutant mice (Fig. 2 F). B1 cells are phenotypically characterized as IgMbrightIgDlow and express lower levels of surface B220. Most B1 cells are also CD43+ and CD23−. As shown in Fig. 2 F, this fraction of peritoneal B cells was reduced in IgαFF/FF mice and most remaining B cells resembled B2 cells in being B220bright, IgD+, CD23+, and CD43−. Corresponding to the reduction of B1 cell numbers, the T cell fraction (CD5+IgM−; Fig. 2 F) of peritoneal lymphocytes was increased.
Higher Percentage of λ1 Light Chain Expressing B Cells in Igα/Igβ Mutants.
As in the BM of IgαFF/FF mice, all B cell fractions were present in normal numbers, the lack of Igα ITAM phosphorylation does not appear to impose a block in early B cell development (Table). However, the impact of the ITAM mutation on the rearrangement and/or selection of Ig light chains is apparent from the distortion of the ratio of κ versus λ light chain expressing B cells. Compared with wild-type controls, IgαFF/FF mice possessed more λ1 light chain expressing B cells (Table, and Fig. 2 E). This phenomenon could also be seen, and appeared even more pronounced, in the cytoplasmic truncation mutants of Igα and Igβ (Table).
Reduced IgG1 Serum Levels and Impaired TD Immune Response in IgαFF/FF Mice.
Altered BCR signaling in IgαFF/FF mice may also affect immune responsiveness. To check whether the ITAM mutation influences the basal levels of Ig titers, Ig isotypes were determined by ELISA in the sera of unimmmunized mice. As shown in Fig. 3 A, the total amount of serum Ig in IgαFF/FF mice, calculated as a sum of the individual isotypes, was comparable to controls. Although B1 cells are thought to be the main contributors of serum IgM, the reduction in B1 and MZ B cells in Igα ITAM mutants did not result in reduced serum IgM concentration. However, the serum concentration of IgG1 in IgαFF/FF mice was significantly lower than in wild-type controls.
To analyze humoral immune responses, we immunized IgαFF/FF mice with the TI type II (TI-II) antigen NP-Ficoll and TD antigen NP-CG. For two reasons we expected to see an impaired TI-II response. First, in IgαΔc/Δc mice, B cells with a cytoplasmic truncation of Igα were unable to mount a measurable response toward NP-Ficoll 21, and second, recent data suggest that MZ B cells are required for a normal TI-II response 33. However, the NP-specific serum titers at different time points after the immunization revealed that Igα ITAM phosphorylation is not required for an efficient TI-II response (Fig. 3 B). IgαFF/FF mice were also able to mount primary and secondary TD responses, which had a kinetics comparable to those of wild-type controls (Fig. 3 C). These data imply that Igα ITAM phosphorylation is not a prerequisite for B cell memory formation. However, the TD response was significantly lower in IgαFF/FF mice at all time points examined (Fig. 3 C). To evaluate further the efficiency of the TD response, we determined the averaged relative affinity of the serum antibodies. Affinities of NP-specific antibodies were comparable between IgαFF/FF mice and control groups, and were equally increased in the course of the immune response, indicating effective affinity maturation (Fig. 3 D).
Stronger Calcium Response in Igα ITAM Mutant B Cells.
To analyze the impact of the ITAM mutation on an early BCR-mediated signaling event, we stimulated Indo-1–loaded B cells either with F(ab)′2 fragment of anti-Ig antibody or with anti-Ig antibodies and monitored changes in intracellular calcium by flow cytometry. Previously, we reported that a cytoplasmic truncation of Igα leads to an elevated and more sustained calcium response in immature B cells of IgαΔc/ΔcIgHEL mice compared with controls 23. We show here that this effect appears to be correlated with the phosphorylation of the Igα ITAM, as splenic B cells from IgαFF/FF mice also showed a stronger calcium response with a delayed maximum in comparison to wild-type controls (Fig. 4, and data not shown).
BCR Ligation Leads to Tyrosine Phosphorylation of Igα in IgαFF/FF B Cells.
To extend the analysis of the signaling capacity of Igα ITAM mutant BCRs, we analyzed protein tyrosine phosphorylation after BCR cross-linking. Splenic B cells were isolated from IgαFF/FF and control mice, and stimulated with F(ab)′2 fragment of anti-Ig antibody for different time durations. Total protein tyrosine phosphorylation was revealed by anti-phosphotyrosine immunoblotting. As shown in Fig. 5 A, BCRs lacking the Igα ITAM tyrosines were still capable of inducing protein tyrosine phosphorylation to levels comparable to controls. We then specifically examined the tyrosine phosphorylation of Igα and Igβ in anti-Igα immunoprecipitates before and after BCR cross-linking. As shown in Fig. 5 B, BCR ligation of IgαFF/FF B cells resulted in a rapid and persistent tyrosine phosphorylation of Igβ ITAM. Moreover, BCR-induced tyrosine phosphorylation of Igβ in IgαFF/FF mice was enhanced compared with controls (Fig. 5a and Fig. b). Whereas Igβ possesses only the two tyrosines of its ITAM, the cytoplasmic tail of Igα contains four tyrosine residues (Fig. 1 E). Interestingly, Igα phosphorylation in Igα ITAM mutant B cells was detected after BCR ligation (Fig. 5 B). These data suggest that either one or both non-ITAM Igα tyrosine residues (#176 and #204) can be phosphorylated after BCR stimulation. This strongly suggests a role for the non-ITAM tyrosines in BCR-mediated signaling, e.g., as a binding site for SH2 domain containing proteins.
Syk kinase is one of the major signaling components that binds to doubly phosphorylated BCR ITAMs and becomes activated. Ligation of BCRs with the mutant Igα ITAM were still able to induce Syk tyrosine phosphorylation, presumably via phosphorylated Igβ. However, maximum phosphorylation was reached at a later time point than in wild-type controls (Fig. 5c). Examination of Lyn phosphorylation showed that basal tyrosine phosphorylation was similar between IgαFF/FF and wild-type B cells, but BCR-induced Lyn tyrosine phosphorylation was less than in wild-type controls (Fig. 5 D).
Complete Block of B Cell Development in IgαFF/FFIgβΔc/Δc Mice.
Targeted mutations of Igα, as in truncation of the cytoplasmic tail in IgαΔc/Δc mice 21 or with specific loss of ITAM phosphorylation in IgαFF/FF mice, lead to alteration in development and responsiveness of B lymphocytes. Similarly, a cytoplasmic truncation of Igβ in IgβΔc/Δc mice compromises B cell development and causes a strong reduction in mature B cells (22; Fig. 6). However, neither the lack of Igα nor Igβ ITAM alone leads to a complete block in B cell development. A cytoplasmic domain of either Igα or Igβ is still sufficient to signal 22. Based on the phenotypic differences between IgαΔc/Δc and IgαFF/FF mice and the finding that Igα may be phosphorylated at non-ITAM tyrosines, Igα may not mediate signaling exclusively through its phosphorylated ITAM. Whether development of B cells in the Igα and Igβ truncation mutants was absolutely dependent on the phosphorylation of the remaining BCR ITAM remained unclear. To answer this question, we generated IgαFF/FFIgβΔc/Δc mice. In comparison to IgβΔc/Δc mice, the double mutants possessed no intact ITAM. As shown in Fig. 6 A, B cell development in these mice was completely blocked at the pro-B cell stage, as no B220+CD43− B cells were present in the BM. FACS® analysis of the B220+CD43+ pro-B cell fraction revealed the presence of fractions A to C, but fraction C′ (BP-1+HSAbright) was absent (Fig. 6 B). In the periphery of IgαFF/FFIgβDc/Δc mice, a few CD19+ cells were still detectable (Fig. 6 C), however, these cells did not express surface IgM nor IgD. These CD19+ cells expressed the immature B cell marker 493 and had low levels of surface CD19 and B220 (data not shown). Thus, these cells might represent peripheral B cell precursors. The lack of any functional mature B cells in IgαFF/FFIgβΔc/Δc mice was further validated by the absence of any detectable serum Igs (data not shown). Collectively, these data indicate an absolute requirement of Igα ITAM phosphorylation in IgβΔc/Δc mice for B cells to develop beyond the pro-B cell stage.
Igα ITAM Mutation Affects B Cell Development Less Dramatically than Truncation of the Igα Cytoplasmic Tail.
The B cell antigen receptor complex relies on its Igα/Igβ heterodimer for surface expression, internalization, antigen presentation, and signaling. Therefore, a targeted ablation of the cytoplasmic effector domain of either Igα or Igβ may result in impaired B cell development and function. This is indeed evident in both IgαΔc/Δc and IgβΔc/Δc mice which exhibit a drastic reduction in peripheral B cell numbers 21,22. To date, the main signaling function of the Igα/Igβ heterodimer has been ascribed to its two ITAMs. Moreover, mutational experiments in cell lines have revealed that ITAM functions require phosphorylation at its two tyrosine residues 9,10,13,14,16,34. However, the present analysis of IgαFF/FF mice revealed striking differences in comparison to IgαΔc/Δc mice. A lack of Igα ITAM phosphorylation did not affect the relative fraction sizes of pro-B, pre-B, and immature B cells in the BM, and the absolute splenic B cell numbers in IgαFF/FF mice were comparable to those of wild-type controls. Taken together, the differences in phenotypes resulting from either the loss of the Igα cytoplasmic tail or the specific lack of Igα ITAM phosphorylation, imply a role for the non-phosphorylated ITAM and/or other parts of the Igα cytoplasmic tail in Igα-mediated function. This idea is consistent with the finding that Igα ITAM phosphorylation is not a prerequisite for binding of Src family kinases Lyn and Fyn via their N-terminal regions 35 and in agreement with the model that proposes a high affinity, phosphotyrosine-dependent, and a low affinity, phosphotyrosine-independent association of Src family kinases with Igα/Igβ 36. In addition, the Igα molecules in IgαΔc/Δc mice lack the serine/threonine phosphorylation sites which are implicated in the regulation of Igα tyrosine phosphorylation 37. Thus, differences in serine/threonine phosphorylation might also be responsible for the phenotypic differences of IgαΔc/Δc and IgαFF/FF mice. Alternatively, lack of the Igα cytoplasmic domain in IgαΔc/Δc mice might influence the structure of the remaining Igβ tail and distort its ITAM function. The different phenotype of IgαFF/FF mice could thus also reflect a difference in Igβ ITAM function.
Altered Surface Antigen Expression in B Cells of IgαFF/FF Mice Due to Adaptation or Selection.
B cells from IgαFF/FF mice express higher levels of surface IgM and IgD compared with wild-type controls. Three reasons might account for the higher BCR density. First, lack of efficient protein-kinase activation via phosphorylated Igα ITAMs might be partly compensated by expressing more Igα/Igβ on the cell surface. Second, the increase in Igβ ITAM phosphorylation might partially substitute for the lack of Igα ITAM phosphorylation. Third, Igα ITAM phosphorylation might directly control BCR surface expression or internalization. The latter possibility is supported by the analysis of Igα mutant molecules in cell cultures. While in reconstitution experiments using the J558Lμm cell line, neither a cytoplasmic deletion nor Ig−ITAM tyrosine mutations interfered with the surface expression of BCR 10, constitutive receptor endocytosis appeared to be dependent on the cytoplasmic tail of Igα in a chimeric system 38. Similarly, analyzing FcγRII-Igα or Igβ chimeric molecules in the A20/IIA1.6 cell line, efficient constitutive BCR internalization was only mediated by the cytoplasmic domain of Igα and was shown to be dependent on the first Igα ITAM tyrosine 38.
Interestingly, B cells in IgαΔc/Δc mice do not exhibit a similar increase in surface BCR expression levels as in IgαFF/FF mice, and the cytoplasmic truncation of Igβ in IgβΔc/Δc mice leads to lowered BCR expression (22; Fig. 2B and Fig. C). One might speculate that surface BCR expression levels are controlled by constitutive Igα/Igβ-dependent turnover mechanisms and in addition are adjusted to the specific signaling requirements of the B cells.
In line with the idea that Igα ITAM phosphorylation is required for efficient activation of Src family kinases are the observations of an increased CD19 and a decreased CD22 coreceptor surface expression in B cells from IgαFF/FF mice. CD19 was shown to lower the threshold of BCR-mediated cell activation and to be involved in the activation of Src family kinases 39,40. In contrast, CD22 contains immunoreceptor tyrosine-based inhibitory motifs (ITIMs) which serve as docking sites for SH2 domain-containing phosphotyrosine phosphatase (SHP)-1, and thus might serve to lower protein kinase activity 41,42,43. Interestingly, B cells in CD22-deficient mice 44,45 and transgenic mice overexpressing CD19 46 possess decreased surface IgM levels. Thus, differences in BCR and coreceptor surface expression levels in IgαFF/FF mice might be the result of selection and/or adaptation of B cells, in an attempt to cope with weakened intrinsic BCR signaling. To address the issue of selection versus adaptation, we plan to analyze IgαFF-flox/FF-flox mice in combination with inducible or developmental stage specific Cre-recombinase expression 47. The floxed Igα allele (Fig. 1 A, 4) allows undisturbed B cell development with a wild-type BCR until Cre-mediated mutation of the Igα ITAM occurs. Induced mutagenesis should thus allow us to study the adaptive response of mature B cells to a lack of Igα ITAM phosphorylation.
PTK Activation by Igα ITAM and Non-ITAM Tyrosine Phosphorylation.
Previous studies with B cell lines have implied a critical requirement for Igα ITAM tyrosine phosphorylation in the activation of PTKs after BCR ligation 9,10,13,48. However, in some initial signaling studies with splenic B cells of IgαFF/FF mice we did not observe a drastic reduction in overall protein tyrosine phosphorylation after BCR cross-linking, which could be explained by the presence of a functional Igβ ITAM. Overall higher surface expression and more efficient ITAM phosphorylation of Igβ might thus be able to compensate in part for the lack of Igα ITAM phosphorylation in IgαFF/FF B cells. We tested the capability of Syk and Lyn to undergo tyrosine phosphorylation upon BCR ligation in B cells of IgαFF/FF mice. Although Igα ITAM phosphorylation is not absolutely required to induce tyrosine phosphorylation of Syk, the overall Syk tyrosine phosphorylation was lower and delayed as a result of the Igα ITAM mutation. Similarly, Lyn was inducibly tyrosine phosphorylated in IgαFF/FF B cells, but to a lesser extent than in controls. An important novel finding is the demonstration of Igα tyrosine phosphorylation apart from its ITAM. This suggests an involvement of the Igα non-ITAM tyrosines in BCR-dependent signaling. These might recruit distinct signaling proteins or alternatively enhance or inhibit the association of molecules with the Igα/Igβ heterodimer. Both non-ITAM tyrosines are indeed in the sequence context of a potential Src family SH2 domain binding site 49,50. Compared with wild-type controls, the weak tyrosine phosphorylation of Igα in IgαFF/FF B cells supports the idea that the ITAM tyrosines are the main tyrosine phosphorylation sites in Igα 10. However, in the wild-type situation, phosphorylation of the non-ITAM tyrosines could well be linked to and dependent on Igα ITAM phosphorylation. A clarification of these matters will require a biochemical analysis which is beyond the scope of this paper.
Igα ITAM Phosphorylation and Ca2+ Mobilization.
The enhanced Ca2+-flux in B cells of IgαFF/FF mice after BCR ligation seems to contradict the idea that Igα ITAM phosphorylation functions to amplify signaling. Previous transfection experiments with FcγRII-Igα or FcγRII-Igβ chimeric molecules in the IIA1.6 cell line demonstrated that only the Igα cytoplasmic tail was capable to mobilize Ca2+ from intracellular stores and influx of extracellular calcium. In contrast, the Igβ cytoplasmic domain stimulated only an oscillatory release from intracellular stores, which overall resulted in a small increase of free intracellular Ca2+ 9,51. Further analyses specified an absolute requirement of the proximal Igα ITAM tyrosine (Y182) for Ca2+ mobilization 9,13. Thus, it was surprising to find an even stronger Ca2+ mobilization in splenic B cells of IgαFF/FF mice after BCR ligation (Fig. 4). One might interpret the data such that Igα ITAM phosphorylation is required to contain and terminate Ca2+ mobilization. While the increased Ca2+ mobilization in B cells of IgαFF/FF mice is not caused by overall higher surface BCR expression (data not shown), changes of coreceptor expression might influence the outcome of BCR ligation by resetting signaling thresholds. In support of the latter, overexpression of CD19 46 or lack of CD22 44,45,52,53 in mouse B cells also leads to enhanced calcium mobilization, whereas the lack of CD19 lowers it 54.
Igα/Igβ Mutations Distort Light Chain Rearrangements.
Although we do not yet know whether the Igα ITAM mutation leads to enhanced editing of autoreactive BCRs like the Igα truncation 23, it is noteworthy that both mutations distort the κ to λ ratio in favor of λ light chain expression. One might speculate that quantitative and qualitative alterations of signals transmitted by a mutant Igα/Igβ complex might hinder the efficient termination of light chain rearrangements in B cell development, widening the time window in which sequential rearrangements in the light chain loci occur and thus driving more cells to the λ chain expression, if κ rearrangements usually preceed λ rearrangements 55. In line with this interpretation are the observations that the shift in favor of λ chain expression is more pronounced in the tail truncation mutants than in IgαFF/FF mice and that only in the former mutants there is an apparent block of B cell development at the immature B cell stage. Alternatively, altered pre-BCR signaling might influence the order of light chain gene rearrangements and cause an earlier opening of the λ locus. In an extreme scenario this could lead to the exclusive rearrangement of the λ locus in λ chain–expressing cells. We tested this hypothesis by analyzing the rearrangement status of the κ loci in λ1+ splenic B cells. However, similar to the situation in λ1+ cells from wild-type mice 56, the κ light chain loci were found to be rearranged in λ1+ cells of IgαFF/FF mice (data not shown). Finally, it might be possible that λ light chains confer a selective advantage for B cells in case of distorted BCR signaling.
Dependence of B1 and MZ B Cell Populations on Igα ITAM Phosphorylation.
In the peritoneal cavity of IgαFF/FF mice, the total number of B cells was lower than in the wild-type due to a drastic reduction in B1a cells. The majority of the remaining peritoneal B cells were phenotypically B2 cells. B1 cells often express germline-encoded self-reactive BCRs and appear to have signaling requirements that are different from B2 cells 57. Recent studies have demonstrated that the expression of an autoreactive receptor can drive the generation of B1 cells 58,59, suggesting that B1 cell development depends on signals mediated by BCR–ligand interaction. The specific reduction of B1 cells in IgαFF/FF mice could thus be explained by enhanced negative selection of autoreactive BCR specificities, but would also be consistent with the idea that Igα ITAM phosphorylation functions as a signal amplifier. In support of the latter conclusion, several mouse mutants affecting genes involved in positive regulation of BCR signaling such as CD19 60,61, Btk 62,63, and Vav 64,65 also show a reduction in the B1 cell population. An opposing phenotype is seen upon mutation of genes thought to negatively regulate BCR-related signaling. SHP-1 deficiency in mev/mev mice 66 and knockouts of CD22 53,45 and CD72 67 all lead to an expansion of the B1 cell population. This is also true for mice in which the SHP-1 gene is selectively inactivated in B cells (data not shown).
In the spleen of IgαFF/FF mice we observed a strong reduction of MZ B cells. Similar to the B1 population, MZ B cell selection also depends on BCR specificities with a bias for autoreactivity 68. Moreover, both B1 and MZ B cells are CD23lowIgMhigh and larger in size than B2 cells. Theses results demonstrate directly that appropriate BCR signaling is required for the normal generation and maintenance of both these subsets.
Igα ITAM Phosphorylation Is Required for Efficient TD Immune Responses.
B1 and MZ B cells are thought to be the frontline of an early defense against pathogens and appear to be regulated in a T cell–independent way 69. In a recent study of Pyk-2 knockout mice, which completely lack MZ B cells, an important role for MZ B cells in T cell–independent antibody responses was suggested 33. Therefore and because TI-II responses have been shown to be severely impaired in Igα and Igβ tail truncation mutants 21,22, we expected an impairment of the anti-NP-Ficoll response in IgαFF/FF mice. However, the selective elimination of the Igα ITAM does apparently not interfere with efficient B cell activation upon strong BCR cross-linking. It is unclear at this point whether the difference between the Igα ITAM and the truncation mutants in this respect reflects a difference in B cell activation or relates to the development of B cell subsets which may be disturbed in the mutants to different extents. That the Igα ITAM mutation is compatible with an intact TI-II response, but affects the efficiency of a TD antibody response is consistent with the findings in cell lines that Igα ITAM tyrosines contribute to efficient antigen presentation, but are not absolutely required for the presentation of multivalent antigens 48,70,71. Inefficient TD responses may also explain the low IgG1 levels in the sera of IgαFF/FF mice.
Absolute Requirement of BCR ITAM Tyrosines for B Cell Development Beyond the Pro-B Cell Stage.
A striking observation is the absolute block of B cell development at the pro-B cell stage in mice lacking the Igα ITAM and the Igβ cytoplasmic tail. This contrasts with an almost unimpaired early B cell development in the Igα ITAM mutant and significant pre-B cell promotion in mice expressing a truncated Igβ or Igα chain. The simplest interpretation of these results is that there is significant functional redundancy between the Igα and Igβ ITAM (similar to that exhibited by the ITAMs of the T cell receptor complex; reference 72), and that there is an absolute requirement for the presence of an ITAM in the pre-BCR at the pro- to pre-B cell transition which is known from earlier work to depend on pre-BCR expression 17,29,73,74,75. It remains to be established whether this requirement relates to pre-BCR surface expression or more likely, the signaling function of the surface bound receptor. The latter aspect is supported by previous findings showing a requirement of ITAM tyrosines for pre-B cell development and allelic exclusion in transgenic systems using IgM/Igβ and IgM/Igα chimeric receptors 18,19,20. It will be interesting to determine whether later stages of B cell development, where the BCR rather than the pre-BCR plays a critical role 21,76,77,78, exhibit a similar ITAM dependence. Inducible targeted mutagenesis will be instrumental in resolving this issue.
We thank Dr. H. Singh for the genomic cosmid clone and Dr. A. Rolink for the monoclonal antibody 493. We are grateful to A. Egert, C. Göttlinger, G. Ringeisen, A. Leinhaas, C. Uthoff-Hachenberg, and S. Willms for excellent technical help, and to Dr. Raul Torres and members of the Rajewsky lab for discussion and support.
This work was supported by the Volkswagen Foundation (grant no. I/71768), the Deutsche Forschungsgemeinschaft through SFB 243, the European Community (grant no. BIO4-CT96-0077), the Alexander von Humboldt Foundation to L. Pao, and by the Max-Planck Research Award to K. Rajewsky.
K. Rajewsky's present address is Center for Blood Research, Harvard Medical School, 200 Longwood Ave., Boston, MA 02115.
Abbreviations used in this paper: BCR, B cell receptor; BM, bone marrow; ES, embryonic stem; ITAM, immunoreceptor tyrosine–based activation motif; MZ, marginal zone; PTK, protein tyrosine kinase; RT, room temperature; SH2, Src homology 2; TD, T cell–dependent; TI, T cell–independent.