Qualitative Regulation of B Cell Antigen Receptor Signaling by CD19: Selective Requirement for PI3-Kinase Activation, Inositol-1,4,5-Trisphosphate Production and Ca²⁺ Mobilization

The murine B cell plasmacytoma J558Lμm3 was provided by M. Reth (Max-Planck Institute for Immunology, Freiburg, Germany). We transfected this cell line previously to obtain a mouse full-length (B220) CD45⁺-expressing variant (28). The human CD19 cDNA employed was provided by T. Tedder (Duke University Medical Center, Durham, NC). Rabbit polyclonal antibodies against CD19, Ig α, Syk, and Lyn, were prepared using GST fusion proteins produced in bacteria, purified by glutathione–Sepharose chromatography and cleaved with Factor Xa. Immunogens included the complete cytoplasmic tail for Igα, residues 1-131 for Lyn, the linker region for Syk (29), the SH3 domain for p85, and residues 411-547 of the CD19 cytoplasmic tail. The polyclonal antibody to the p85 subunit of PI3-kinase and mixed mouse monoclonals to PLCγ1 were purchased from UBI (Lake Placid, NY), anti-phosphotyrosine (AB-2) from Oncogene Science (Uniondale, NY), monoclonal anti-human CD19 from DAKO (Carpinteria, CA), and rabbit polyclonal to PLCγ2 from Santa Cruz Biochemicals (Santa Cruz, CA). Rabbit F(ab′)₂ antibody fragments to mouse IgG (H+L) (F(ab′)₂RAMIG) was purchased from Zymed Labs. (South San Francisco, CA). HRP-conjugated protein A (Zymed) and rat anti–mouse IgG1 (Zymed) were used for detection using the ECL detecting system (Amersham Corp., Arlington Heights, IL). R-phycoerythrin conjugate (PE-streptavidin) was from Biosource (Camarillo, CA). Indo-1 AM was from Molecular Probes Inc. (Eugene, OR). The inositol-1,4,5-trisphosphate [³H] radioreceptor assay kit and Myo-[2-³H(N)]-inositol (21.0 Ci/mmol) was purchased from DuPont-NEN (Boston, MA). Geneticin G418 was from GIBCO BRL (Gaithersburg, MD). Wortmannin was from Sigma Chem. Co. (St. Louis, MO). Nitrophenol (NP)₉BSA was prepared by coupling of BSA (10 mg/ml) in 3% NaHCO₃ to NP-CAP-OSu (Cambridge Research Biochemicals, Cambridge, UK) (40 mg/ml) in N,N-dimethyl formamide (DMF). After dialysis in NaHCO₃ and PBS the molecular ratio of NP to BSA was determined by measurement of the absorbance at OD₄₃₀ with an extinction coefficient of 4230, pH 8.5 (NP molarity) and by Bradford analysis (BSA molarity). CD19^−/− mice (27) were provided by R.C. Rickert and K. Rajewsky (University of Cologne, Cologne, Germany).

The human CD19 cDNA was isolated from pMT2 by EcoR1 digest and subcloned into pSK Bluescript (Stratagene Inc., La Jolla, CA). The CD19 cDNA was then removed from pSK with HindIII/ClaI restriction digest and cloned into pLNCX (Williams, Fred Hutchinsen Cancer Center, Seattle, WA). This construct was electroporated into GP+E/NIH3T3 cells and cell supernatant collected at 72 h. To increase the virus titer, GP+E/NIH3T3 cells were incubated for 12 h in tunicamycin (1 μg/ml), and reinfected with pLNCX/hCD19 retrovirus by incubating with the cell supernatant collected above. After G418 selection (1 mg/ml), virus expressing clones were selected. Virus collected from these cells was then used to infect J558LμmCD45⁺ myeloma cells. Polyclonal human CD19 positive J558Lμm 3CD45⁺ cells were obtained after G418 selection by cell sorting of the bulk-infected population with a mouse monoclonal to human CD19. The cells were propagated in IMDM supplemented with 5% heat inactivated FCS (Hyclone Labs. Inc., Logan, UT), 50 U/ml penicillin and 50 μg/ml streptomycin and 1 mg/ml G418 (J558Lμm3CD45⁺CD19⁺) at 37°C with 7% CO₂. Mutation of CD19 tyrosines 484 and 515 to phenylalanines was performed by PCR using the primers TCCCAGTCCTTTGAGGATATG/ CATATCCTCAAAGGACTGGGA for Y⁴⁸⁴ and GCAGACTCTTTTGAGAACATG/CATGTTCTCAAAAGAGTCTGC for Y⁵¹⁵. Mutated CD19 was inserted into pLNCX, packaged and used for infection as described above.

For stimulation cells were harvested, washed once in IMDM and resuspended at 10⁷ cells/ml in IMDM. After 5 min incubation at 37°C, cells were stimulated with NP₉BSA (2.5 μg/ml), pelleted in a picofuge, resuspended in lysis buffer (10 mM Tris, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM Na₃VO₄, 10 mM NaF, 1% NP-40, 1 mM PMSF, and 2 μg/ml each of leupeptin, aprotinin, and α-1 antitrypsin), and then incubated on ice for 10 min. Lysates were centrifuged for 5 min at 14,000 rpm. Cleared lysates were incubated with antibodies to signaling molecules and protein A–Sepharose or protein G–Sepharose at 4°C, washed four times in lysis buffer, boiled with reducing SDS sample buffer, and then fractionated on 10% SDS-PAGE. SDS-PAGE–fractionated proteins were subjected to electrophoretic transfer to PVDF membranes. Transfers were probed with antibodies and developed using the ECL detection system. Sequential immunoblotting for phosphotyrosine and effector was performed. For this purpose PVDF membranes were stripped in stripping buffer (100 mM 2-mercaptoethanol, 2% SDS, 62.5 mM Tris-HCl, pH 6.7) for 30 min at 54°C, washed and blocked before immunoblotting.

Cells were loaded with indo-1 AM and the intracellular [Ca²⁺]_i was monitored by flow cytometry (model 50H; Ortho Diagnostic Systems, Westwood, MA) as previously described (28). The mean [Ca²⁺]_i and percent of cells responding was determined with an appended data acquisition system and the MultiTIME software (Phoenix Flow Systems, San Diego, CA).

Inositol-1,4,5-trisphosphate generation was measured using a [³H] radioreceptor assay kit (DuPont-NEN) according to the manufacturer's instructions. Generation of total inositol phosphates was measured as follows: cells were labeled for 18 h in 2 μCi/ml of myo-[2-³H(N)]- inositol in inositol-free medium containing dialyzed 5% FCS before being washed in inositol-free media and incubated with inositol-free media containing 20 mM LiCl for 5 min. Cells were then stimulated for 25 min, lysed in 2 ml of methanol/HCl (99:1), and then 2 ml of chloroform was added. The aqueous phase was extracted, applied to a formate anion-exchange column, and then inositol phosphates were eluted with elution buffer (0.1 M formic acid, 1.0 M ammonium formate). Total [³H] inositol incorporation was determined by measuring the radioactivity in the organic phase.

Cells were stimulated, lysed, and then immunoprecipitated for 1 h at 4°C using anti-p85 antibodies and protein A–Sepharose. PI3-kinase activity was measured as previously described (30). Samples were dried and reaction products resolved by thin-layer chromatography on precoated TLC Silica Gel 60 plates in CHCl₃/MeOH/2.5 M NH₄OH (9:7: 2). The incorporation of ³²PO₄ into phosphatidylinositol 3-phosphate was determined using a PhosphorImager.

Splenic B cells were prepared as previously described (31). In brief, spleens were excised from mice and cells were dispersed through a 100-μm mesh in IMDM. Red blood cells were lysed using Gey's solution. Total nucleated spleen preparations were depleted of T lymphocytes by complement-mediated lysis using HO13.4 and T24 antibodies and dense cells (ρ ⩾1.066) isolated by discontinuous Percoll density gradient centrifugation.

Although tyrosine phosphorylation of CD19 after BCR ligation is well documented, the contribution of CD19 to BCR signaling is not defined. To study the role of CD19 in antigen-induced signal transduction we generated a wild-type CD19-positive and Y^484/Y515 to F CD19-positive variant of the J558Lμm3CD45⁺ plasmacytoma. The parent clone expresses a NP-specific BCR and CD45, but does not express CD19 (28, 32). Human CD19 (hCD19) was introduced into this clone by retroviral infection, and lines (polyclonal populations) were isolated by G418 selection and fluorescence activated cell sorting of cells. Flow cytometric analysis revealed similar expression levels of sIgM, CD45 and CD19 in the plasmacytoma cell lines (Fig. 1). Human CD19 was utilized since antibodies to mouse CD19 were not available at the time the studies were undertaken, and human and mouse CD19 are highly homologous (79% in nucleotide sequence, 75% in amino acid sequence) particularly in the cytoplasmic domain (3, 33). Recent studies indicate that human CD19 complements B cell function in CD19^−/− mice (34). In addition, using a recently generated rat mAb, 1D3, specific for murine CD19 (35), Krop et al. (22) have demonstrated that murine CD19 shares with human CD19 an association with complement receptor CD21 and CD81, tyrosine phosphorylation, binding of PI3-kinase, and synergistic signaling with membrane IgM.

To investigate the differences in BCR signaling in the CD19-negative and -positive contexts, we loaded the J558Lμm3 CD45⁺CD19-negative and CD19-positive variants with indo-1 AM and analyzed changes in intracellular free calcium concentration ([Ca²⁺]_i) in response to antigen stimulation. Initially, this was done under conditions where the free calcium concentration in the medium ([Ca²⁺]_o) was maintained at physiological levels of 1.3 mM. As shown in Fig. 2 (solid lines), antigen stimulation induced a rise in [Ca²⁺]_i in both cell lines; in the CD19-positive cell line the maximal mean [Ca²⁺]_i reached 430 nM while it was 260 nM in the CD19-negative cells. In the CD19-positive cell line, antigen stimulation also induced a much more prolonged increase in [Ca²⁺]_i than in the CD19-negative clone. To determine whether intracellular or extracellular calcium was responsible for the respective [Ca²⁺]_i increases, we added 3 mM EGTA at the time of initiation of analysis of [Ca²⁺]_i. This results in buffering of [Ca²⁺]_o to a concentration of 65 nM (which corresponds to the [Ca²⁺]_i in resting cells) and prevents influx of extracellular calcium into the cell. Comparison of antigen-induced responses of CD19-negative and CD19-positive cells in low [Ca²⁺]_o (Fig. 2, dotted lines) indicated that optimal receptor-mediated release of calcium from intracellular stores requires CD19. Comparison of responses in high [Ca²⁺]_o revealed that influx is also dependent on the expression of CD19 (Fig. 2, solid lines).

To study CD19 dependence of signaling events upstream from calcium mobilization, we first compared antigen stimulation of IP₃ production in CD19-positive and -negative cells. IP₃ binds to its receptors on the ER leading to release of Ca²⁺ into the cytoplasm. Emptying of the ER stores triggers entry of calcium from the extracellular space by a mechanism known as capacitative Ca²⁺ entry (36). Thus, CD19 dependence of IP₃ generation-induced response by antigen could explain the dependence of both intracellular calcium release and extracellular calcium influx on CD19 expression. Stimulation of the cell lines led to a very rapid and sustained rise in IP₃ production in the CD19-positive clone as measured using an IP₃-receptor binding inhibition assay (Fig. 3,A). Antigen stimulation of the CD19-negative clone led to rapid, but much more modest and only transient generation of IP₃. The kinetics and magnitude of IP₃ generation was consistent with those of calcium mobilization responses shown in Fig. 2 supporting the possibility that the attenuated calcium mobilization response may be caused by the attenuated BCR-mediated hydrolysis of PIP₂. We also measured the generation of [³H]-inositol mono- and polyphosphates after antigen stimulation of myo-[2-³H(N)]-inositol–labeled cell lines. As shown in Fig. 3,B the fold increase in total inositol mono- and polyphosphates generated during a 25-min stimulation was consistent with the mass measurement of IP₃ production (Fig. 3 A); stimulation of the CD19-positive clone led to an ∼250% increase in inositol phosphate generation, while stimulation of the CD19-negative cell line induced an ∼30% increase.

Antigen stimulation is known to induce recruitment of Src- and Syk-family tyrosine kinases to Ig-α and Ig-β components of the BCR (37–39). The tyrosine kinases bind via their SH2-domains to immunoreceptor tyrosine based activation motif (ITAM) phosphotyrosines and this leads to secondary recruitment, phosphorylation, and activation of downstream signaling molecules such as PLCγ1, PLCγ2, and PI3-kinase (reviewed in reference 40). To define the effects of CD19 expression on these phosphorylation events we stimulated CD19-negative and -positive cell lines, immunoprecipitated intracellular signaling molecules and conducted anti-phosphotyrosine immunoblotting analysis. The spectra of phosphoproteins in whole cell lysates from the J558Lμm3CD45⁺CD19^- and the J558Lμm3CD45⁺ CD19⁺ cell lines looked similar with the exception of the presence of a tyrosine phosphorylated CD19 in the CD19-positive cell lysate (results not shown). Induced phosphorylation of several signaling molecules was analyzed after stimulation with normalization of immunoprecipitation and lane loading by sequential antieffector immunoblotting. We studied tyrosine phosphorylation after 1 min of antigen stimulation since we have shown previously that effector enzymes under study are maximally phosphorylated at this time in the J558Lμm3CD45⁺ plasmacytoma (41). As shown in Fig. 4, CD19 expression led to modest increases in antigen-induced tyrosine phosphorylation of Ig-α, Lyn, Syk, PLCγ1, and PLCγ2. Quantitation of tyrosine phosphorylation by scanning densitometry (at least three independent experiments for each effector enzyme) and normalization to effector levels revealed increased phosphorylation of Ig-α, Lyn, and Syk by 1.5–2-fold in the CD19-positive line. Induced PLCγ1 and PLCγ2 tyrosine phosphorylation was 2.5–3-fold higher in the CD19-positive cell line than in the CD19-negative cell line. These results suggest that CD19 functions in part by increasing the magnitude of certain proximal BCR coupled tyrosine phosphorylation events.

CD19 contains two cytoplasmic YXXM motifs that mediate its association with the p85 subunit of PI3-kinase after phosphorylation of the tyrosines (9). The phosphorylation of these tyrosines may contribute to BCR-mediated PI3-kinase activation. We determined the activity of PI3-kinase in the cell lines by measuring the PI3-kinase activity associated with anti-p85 immunoprecipitates after stimulation with NP₉BSA for 2 min. The results shown in Fig. 5 demonstrate a 250% increase in PI3-kinase activity in the CD19-positive cell line after antigen stimulation, but only a 40% increase in the CD19-negative variant. The residual activation of PI3-kinase in the absence of CD19 expression is probably a result of PI3-kinase activation through interaction with SH3 domains of Src-family kinases (42). The results described above are summarized in the two first columns of Table 1. It is clear that CD19 influences multiple intracellular signaling events. Some signaling events are only slightly increased in the CD19-positive cell line, (i.e., Ig-α, Lyn, and Syk tyrosine phosphorylation) while others are almost totally dependent on expression of CD19 (i.e., IP₃ generation, calcium influx, and PI3-kinase activation).

To begin to dissect the relationship among the observed CD19 effects, we assessed the effects of the PI3-kinase inhibitor wortmannin on signaling events. Wortmannin has been shown to inhibit PI3-kinase activity by covalent modification of a lysine residue in the PI3-kinase catalytic subunit, p110 (43). We also used the competitive PI3-kinase inhibitor Ly294002 (44) and obtained similar results as in our studies with wortmannin (results not shown). The effect of wortmannin on antigen-induced activation of PI3-kinase was determined. The results shown in Fig. 6 A demonstrate the inhibitory effect of wortmannin on antigen-mediated PI3-kinase activation in the J558Lμm3CD45⁺ CD19⁺ plasmacytoma. Cells were pretreated for 30 min with DMSO (vehicle) or increasing concentrations of wortmannin, stimulated with antigen for 2 min, washed, and then lysed before PI3-kinase was immunoprecipitated and activity measured in an in vitro kinase assay with phosphatidylinositol (PI) containing vesicles as substrate. Consistent with its previously established IC₅₀, 10 nM wortmannin inhibited the antigen-mediated PI3-kinase activation by 54%, and 25 nM wortmannin inhibited 75% of antigen-induced PI3-kinase activity. Wortmannin added to the cells immediately before washing and lysis did not effect the PI3-kinase activity (data not shown) indicating that the inhibitor was acting intracellularly.

To investigate the influence of PI3-kinase on the BCR-mediated calcium mobilization seen in the CD19-positive cell line, we preincubated the cells with increasing concentrations of wortmannin and measured the antigen-induced increase in [Ca²⁺]_i. As shown in Fig. 6,B, wortmannin strongly inhibited Ca²⁺ mobilization in the J558Lμm3CD45⁺ CD19⁺ cells. Sensitivity to wortmannin was also evident in the IP₃ generation response. A wortmannin concentration of 25 nM strongly inhibited IP₃ production in the J558Lμm3 CD45⁺CD19⁺ cell line after 1 minute of antigen stimulation (Fig. 6,C). In both cases wortmannin at a concentration of 25 nM converted the phenotype of the CD19-positive cells essentially to that of the CD19-negative cells. This result was consistent with the 75% inhibition of PI3-kinase activation by 25 nM wortmannin in Fig. 6 A and indicates that the CD19 requirement for IP₃ production and calcium mobilization may be attributable to PI3-kinase activation. Interestingly, wortmannin inhibited the residual antigen-stimulated PI3-kinase activation, IP₃ production, and Ca²⁺ mobilization of the CD19-negative cell line (data not shown) consistent with the possible absolute dependence of these responses on PI3-kinase activation.

To investigate the potential role of CD19-activated PI3-kinase on early BCR-mediated protein tyrosine phosphorylation events, CD19-positive cells were preincubated for 30 min with DMSO (vehicle) or increasing concentrations of wortmannin and stimulated for 1 min with antigen before effectors were immunoprecipitated and analyzed. As shown in Fig. 7, antigen-induced tyrosine phosphorylation of Ig-α, Lyn, and Syk was insensitive to wortmannin. PLCγ1 and PLCγ2 tyrosine phosphorylation was somewhat sensitive to wortmannin preincubation although the decrease in induced tyrosine phosphorylation was modest (35% inhibition at 25 nM wortmannin, Table 2) compared with the degree of wortmannin inhibition of IP₃ generation and calcium mobilization (91.5% inhibition, Table 2). It seems very unlikely that the >90% inhibition of PLCγ activity could be caused by a 35% inhibition of PLCγ phosphorylation Thus, as summarized in Table 1, CD19 expression by the J558Lμm3 CD45⁺ plasmacytoma is required for optimal antigen stimulation of multiple intracellular signaling events, but only some of these are secondary to the nearly absolute dependence of PI3-kinase activation on CD19.

To further explore the requirement of CD19 for PI3-kinase activation and optimal antigen-mediated [Ca²⁺]_i mobilization, we introduced a double-point mutant of CD19, CD19Y⁴⁸⁴F,Y⁵¹⁵F, into the J558Lμm3CD45⁺ plasmacytoma. The expression levels of sIgM, CD45 and CD19 (Y⁴⁸⁴F,Y⁵¹⁵F) in this cell line is shown in Fig. 1. It has previously been demonstrated that the mutations Y⁴⁸⁴F and Y⁵¹⁵F abrogates the ability of CD19 to interact with the p85-subunit of PI3-kinase (9). Like the CD19-negative cell line, the CD19 mutant cell line, showed no significant antigen-mediated activation of PI3-kinase (results not shown). When we assayed the ability of the cell line expressing the mutant CD19 to mobilize [Ca²⁺]_i in response to antigen stimulation, we found that the mutant CD19 cell line behaved equivalently to the J558Lμm3CD45⁺ cell line (Fig. 8). This suggests that both CD19-mediated increase in BCR-induced PI3-kinase activation and Ca²⁺ mobilization require phosphorylation of CD19 Y⁴⁸⁴ and Y⁵¹⁵.

To investigate the role of PI3-kinase in antigen-induced signaling in a more physiological system, we purified splenic B cells (ρ ⩾1.066) from spleens of normal mice and assessed the effects of wortmannin on BCR signaling. After preincubation for 30 min with DMSO or wortmannin, splenic B cells were stimulated with 13.2 μg/10⁶ cells F(ab′)₂RAMIG. As observed in the J558L μm3CD45⁺CD19⁺ cell line, wortmannin inhibited, as a function of its concentration, the antigen-mediated calcium mobilization in these primary B cells (Fig. 9). Wortmannin also inhibited the residual antibody-mediated Ca²⁺ mobilization in CD19^−/− splenic B cells (data not shown). Thus, BCR signaling in J558Lμm3CD45⁺CD19⁺ and normal B cells is similarly sensitive to PI3-kinase inhibition and CD19 function (see below).

To extend our observations regarding CD19 dependence of BCR-mediated signaling in B lymphocytes we compared BCR-mediated PI3-kinase activation and calcium mobilization responses of B cells from CD19 knockout mice and normal mice. It has previously been reported that no distinguishable difference can be seen between CD19-deficient and normal B cells in the pattern or the degree of substrate phosphorylation after surface immunoglobulin stimulation (27). We obtained the same results when comparing whole cell lysates from CD19^−/− and CD19^+/+ cells (results not shown). However, careful investigation of several signaling events revealed that ρ ⩾1.066 resting splenic B lymphocytes from CD19^−/− mice are impaired in their response to surface immunoglobulin crosslinking when compared with CD19^+/+ littermates. F(ab′)₂RAMIG-mediated PI3-kinase activation was severely diminished in the CD19^−/− B cells (Fig. 10), reminiscent of the situation in the plasmacytoma cell lines. As shown in Fig. 11, absence of CD19 also clearly affects BCR-mediated calcium mobilization. A decrease in calcium mobilization was seen in B lymphocytes from the CD19^−/− mice in response to stimulation with an optimal dose of antibody (Fig. 11,A). At suboptimal concentrations of antibody (1.3 μg F(ab′)₂RAMIG) almost no BCR-mediated calcium mobilization was seen in cells from CD19^−/− mice, while cells from the CD19^+/+ littermates still responded well (Fig. 11 B). We obtained equivalent results in 10 independent experiments. Our results appear to conflict with those of Sato et al.(34), who observed enhancement of later phase Ca²⁺ mobilization in CD19^−/− cells. These contrasting findings may result from technical differences in the respective studies.

Recent studies of CD19^−/− mice and CD19 transgenic mice have demonstrated the importance of CD19 in B cell development, but particularly in the immune response of mature B cells. In human CD19 transgenic mice B cell development is impaired (45) as indicated by the fact that these mice have dramatically fewer B cells in the periphery. In these transgenic animals proper B cell development is very sensitive to even small increases in CD19 expression levels (46). CD19^−/− mice also have lower levels of B cells in the periphery, and mature B cells from these animals mount a poor proliferative response to anti-BCR antibodies. These animals respond only weakly to T cell–dependent antigens, germinal center formation is lacking or diminished, and the serum immunoglobulin levels are significantly reduced (26, 27, 47). Based on these findings, it has been suggested that CD19 functions by enhancing the cellular sensitivity to antigen, setting the threshold for antigen-mediated signaling (46). This may reflect a CR2-dependent coreceptor function and/or an accessory function of CD19 in BCR signal transduction.

Here we have defined an important accessory function for CD19 in antigen-induced BCR signal transduction and described a molecular basis for this accessory function. CD19 was found to influence the magnitude of activation of multiple intracellular signaling pathways after antigen stimulation of the BCR. While expression of CD19 leads to modestly enhanced antigen-induced Ig-α, Lyn, Syk, PLCγ1, and PLCγ2-inductive tyrosine phosphorylation, CD19 is absolutely required in order to obtain substantial PI3-kinase activation, IP₃ generation, and maximal calcium mobilization. Thus, CD19 modifies BCR signaling in a qualitative manner. Using the PI3-kinase inhibitor wortmannin we further determined that while the increased antigen-induced tyrosine phosphorylation of Ig-α, Lyn, and Syk is independent of PI3-kinase activation, differences in IP₃ turnover and calcium mobilization are a function of PI3-kinase activation. Results obtained in CD19-positive and -negative variants of the J558Lμm3CD45⁺ plasmacytoma cells were confirmed using primary B lymphocytes from CD19^−/− mice and wild-type littermates. Thus, CD19 functions in promoting BCR-mediated phosphoinositide hydrolysis via activation of PI3-kinase, and also supports antigen-induced phosphorylation of BCR receptor components and receptor activated kinases by an unknown mechanism. Since the src-family kinases Lyn and Lck are found associated with CD19 in unstimulated cells (5, 6), CD19 may mediate the latter effect via an increase in Ig-α phosphorylation by these kinases.

It has been demonstrated previously that wortmannin can inhibit angiotensin II–mediated IP₃ production in adrenal glomerulosa cells (48) as well as anti-DNP-IgE–mediated IP₃ production in RBL-2H3 cells (49). The mechanisms by which PI3-kinase affects inositol lipid hydrolysis is unclear. PLCγ1 and PLCγ2 are highly homologous enzymes; both contain two SH2 domains and one SH3 domain. In vitro studies suggest that tyrosine phosphorylation of the enzymes is required to increase the catalytic activity, but phosphatase treatment of phosphorylated PLCγ1 only partially inhibits PLCγ1-mediated inositol trisphosphate production (50), indicating that mechanisms other than tyrosine phosphorylation may regulate PLCγ1 activation. Our studies using the PI3-kinase inhibitor, wortmannin, indicate that antigen-stimulated inositol lipid turnover and calcium mobilization are dependent on PI3-kinase activation in B cells. Interestingly, induction of PLCγ1 and PLCγ2 tyrosine phosphorylation is relatively insensitive to wortmannin. As shown in Table 2, IP₃ production is substantially inhibited by 10 nM wortmannin, while this concentration of wortmannin has a neglectable effect on PLCγ1 and PLCγ2 tyrosine phosphorylation. PI3-kinase phosphorylates the phosphoinositol containing lipids; PI, PI(4)P and PI(4,5)P₂ at the D3 position of the inositol ring (30). Two of the products of this phosphorylation, PI(3,4)P₂ and PI(3,4,5)P₃, are important regulators of cell proliferation (51). Recently it was suggested that PI(3,4,5)P₃ may play a role in signaling by associating with SH2 domains, and thereby modulating PI3-kinase association with tyrosine phosphorylated proteins (52). PI(3,4,5)P₃ could thereby regulate the activity of SH2 containing effector enzymes either directly or indirectly by relocating them. We suggest that this may reflect PI(3,4,5)P₃, the product of PI3-kinase, association with the SH2 domain of PLCγ1 and translocation of the enzyme to the plasma membrane where its substrate resides. Indeed addition of PI(3,4,5)P₃ to substrate vesicles has been shown to increase PLCγ hydrolysis of PI(4,5)P₂ (53). It has been demonstrated in other cell types that cell stimulation results in translocation of PLCγ1 and PLCγ2 from the cytosol to the membrane (54, 55). This suggests that activation of phosphatidylinositol hydrolysis involves both a PLCγ tyrosine phosphorylation event and a translocation event. Alternatively, or additionally, interactions between PI(3,4,5)P₃ and pleckstrin homology (PH) domains could be involved in translocation of effector proteins to the membrane to obtain optimal phosphoinositide hydrolysis and Ca²⁺ mobilization. A likely candidate is Bruton's tyrosine kinase, Btk, which has been shown to selectively interact with PI(3,4,5)P₃ vesicles via its PH domain (56). Indeed recent findings suggest that Btk is required for BCR-mediated Ca²⁺ mobilization; in particular for extracellular influx (Rawlings, D., personal communication). In this regard we do see a higher antigen-mediated inductive tyrosine phosphorylation of Btk in our CD19-positive plasmacytoma cell line than in the CD19-negative line, suggesting that CD19 may influence the activation of Btk (results not shown). Recently it was shown in both T cells and B cells that under certain circumstances cell stimulation can lead to tyrosine phosphorylation of PLCγ, but no IP₃ production or calcium mobilization (41, 57–59). These examples further underscores the concept that tyrosine phosphorylation of PLCγ is not in itself sufficient to induce maximal IP₃ generation.

A critical role for CD19 phosphorylation in BCR activation of PI3-kinase, IP₃ production, and calcium mobilization is also supported by studies of FcγRIIB1 signal transduction. We and others have recently shown that coligation of BCR with FcγRIIB1 leads to rapid dephosphorylation of CD19 and failed PI3-kinase activation (60, 61). Therefore, FcγRIIB1-mediated CD19 dephosphorylation appears to prevent the normal contribution of CD19 to BCR signaling. Importantly, FcγRIIB1 cocrosslinking with BCR also results in aborted IP₃ production and reduced calcium mobilization. Thus, the calcium mobilization and IP₃ production phenotypes of the CD19-positive and -negative cell lines are clearly reminiscent of the signaling phenotypes of B cells stimulated with F(ab′)₂ or intact antibodies, respectively (62, 63). F(ab′)₂ stimulation leads to a strong increase in both intracellular and extracellular [Ca²⁺]_i mobilization, comparable to the antigen-mediated calcium mobilization in the CD19-positive cells, and a strong and sustained IP₃ production. In contrast coengagement of the FcγRIIB1 receptor by stimulation with intact antibodies only induces a transient production of IP₃, and calcium mobilization as seen in the CD19-negative B cell line (Figs. 2 and 3). These data support a role for CD19 and its effector PI3-kinase in BCR-mediated PLCγ activation and calcium mobilization.

Since our studies indicate that CD19 is required for maximal (i.e., both primary and secondary phase) Ca²⁺ mobilization, modulation of CD19 tyrosine phosphorylation, and resulting modulation of [Ca²⁺]_i mobilization pattern may be important in differentially regulating downstream signaling pathways such as activation of MAP kinase pathways, transcription factors, and proliferation/apoptosis responses (64, 65). Recently it was demonstrated that the amplitude and duration of Ca²⁺ signals contribute to transcriptional specificity (64). Such effects could at least in part be a consequence of the contribution of CD19, i.e., stimulation of B lymphocytes with F(ab′)₂RAMIG, in which case CD19 is highly tyrosine phosphorylated, leads to a Ca²⁺ spike and a constant elevated second phase Ca²⁺ response. In contrast binding of immune complexes to FcγRIIB1 and the BCR, in which case CD19 tyrosine phosphorylation is decreased, results in a Ca²⁺ spike, but no prolonged second phase Ca²⁺ response triggering an effector output different from the F(ab′)₂RAMIG stimulation. These modulations of the Ca²⁺ response may result in modulation of cell proliferation and apoptosis. Splenic B cells proliferate in response to B cell mitogens such as Staphylococcus aureus (SA) and anti-IgM. The SA-proliferation response is partially inhibited by the PI3-kinase inhibitor wortmannin (66), consistent with a role of PI3-kinase in BCR-mediated proliferation responses. CD19 activation may also be important for the regulation of apoptosis. When immune complexes are bound to the FcγRIIB1 and the BCR, in which case CD19 tyrosine phosphorylation is decreased, B cell proliferation is inhibited and cells undergo apoptosis (67). We hypothesize that this inhibition of proliferation and induction of apoptosis is a consequence of inhibition of PI3-kinase activity associated with CD19, leading to inhibition of downstream signaling pathways. Potentially FcγRIIB1-mediated inhibition of CD19 tyrosine phosphorylation and PI3-kinase activation could result in a failure to activate the serine/ threonine protein kinase Akt/PKB, which has been shown to be a target of PI3-kinase (68–70), and is known to protect cells against apoptosis (71–73).

Here we have demonstrated a potential molecular basis for the essential role played by CD19 in antigen-mediated B cell activation, germinal center formation and generation of immune responses. CD19 functions as a constitutive component of the BCR signaling apparatus and is required for antigen-mediated BCR stimulation of PI3-kinase activation and, via PI3-kinase activation, phosphoinositide hydrolysis, and calcium mobilization. In fact, this is the first formal demonstration that CD19 modifies BCR signaling directly. The findings have important biological implications since CD19 and BCR expression are differentially regulated during B cell development. Variation of the ratio of expression of these molecules may provide a mechanism to vary the output of the receptor as a function of cell differentiation stage. Of potential importance in this context is the fact that the contribution of CD19 to BCR signaling is qualitative; it does not simply change the magnitude of the signal.

We thank Michael Reth for provision of the J558Lμm3 cells, Thomas Tedder for the hCD19 cDNA, and Marc Hertz for help with use of the JMP statistical analysis program, comments and reading of the manuscript.