Src Homology Region 2 Domain-Containing Phosphatase 1 Positively Regulates B Cell Receptor-Induced Apoptosis by Modulating Association of B Cell Linker Protein with Nck and Activation of c-Jun NH₂-Terminal Kinase¹

Cross-linking of B cell receptor (BCR)⁴ rapidly activates Src family protein tyrosine kinases (PTKs) such as Lyn, Fyn, and Blk and induces phosphorylation of immunoreceptor tyrosine-based activation motifs (ITAMs) in the cytoplasmic portion of Igα and Igβ (1, 2). Phosphorylated ITAMs of Igα and Igβ then recruit and activate Syk PTK, which in turn phosphorylates multiple cellular substrates, and transduces a variety of downstream signaling events (3, 4), including activation of Ras, activation of mitogen-activated protein kinases (MAPKs), phosphoinositide turnover, increases in intracellular calcium, and other intermediary events, which ultimately leads to cell proliferation, differentiation, activation, or cell death. Transmission of BCR signals to intracellular signaling machinery is highly dependent on reversible protein tyrosine phosphorylation, which is precisely balanced between PTKs and protein tyrosine phosphatases (PTPs) (5, 6).

Src homology region 2 (SH2) domain-containing phosphatase 1 (SHP-1) is a cytoplasmic PTP containing two SH2 domains at the NH₂ terminus, which is implicated in negative regulation of several receptors, including BCR (6, 7). In the resting state, SHP-1 interacts with the BCR complex, most likely maintaining its dephosphorylated state (8). Upon BCR ligation, SHP-1 dissociates from the BCR complex and binds to BCR-inducible tyrosine-phosphorylated CD22, CD72, and paired Ig-like receptor B (PIR-B), all of which contain immunoreceptor tyrosine-based inhibition motifs (ITIMs) in their cytoplasmic portions (9, 10, 11, 12, 13). Since SHP-1 is activated when its SH2 domains interact with tyrosine-phosphorylated peptides (14), SHP-1 bound to CD22, CD72, or PIR-B might efficiently dephosphorylate signaling molecules such as Igα/Igβ and Syk, thereby acting as a negative regulator (15).

In addition, our earlier study (16) showed that SHP-1 dephosphorylates B cell linker protein (BLNK) (17), also named SH2 domain leukocyte protein of 65 kDa (SLP-65) (18) or B cell adaptor containing SH2 domain (19). Upon BCR ligation, BLNK is tyrosine-phosphorylated by Syk and translocates to the membrane fraction (17), where it acts as a scaffold for recruiting various signaling proteins, such as phospholipase Cγ, Vav, Grb2, Nck, and Btk and is therefore capable of mediating an array of distinct cellular outcomes (20). For example, B cell development is blocked at the transition from pro-B to pre-B cells in BLNK-deficient mice (21, 22). Moreover, BLNK-deficient B cells fail to enter the cell cycle upon BCR ligation due to the inability to induce the expression of cell cycle regulatory proteins and exhibit impaired activation of NF-κB (23). Our finding that BLNK is a physiological substrate of SHP-1 in B cells suggests that SHP-1 can regulate multiple signaling pathways by dephosphorylating BLNK tyrosine residues to which multiple signaling molecules are recruited (16).

By introducing a form of SHP-1 lacking PTP activity (SHP-1-C/S) into WEHI-231 cells, we were also able to show that BCR-induced c-Jun NH₂-terminal kinase (JNK) is significantly enhanced (16). JNK is a member of the MAPK family and is known to play an important role in a variety of downstream signaling pathways (for review, see Ref. 24). When cells are treated with cytokines or exposed to environmental stress, JNK is phosphorylated at tyrosine and threonine residues by SEK1 (also known as MKK4) or MKK7 and activated (24). Several lines of evidence indicate that JNK is activated by Rho family small GTPases (25, 26), Nck adaptor protein and Ste20-like protein kinase, Nck-interacting kinase (NIK) (27), TNFR-associated factor 2 (TRAF2) (28), or Ste20-like kinase, HPK1 (29, 30). However, it is not yet clear how these upstream molecules are interrelated or how JNK modulates final cellular responses. For example, one group demonstrated that activation of JNK and p38 correlates with cell death in one human B cell line (31), whereas no correlation was found between JNK activation and BCR-mediated cell death in WEHI-231 cells (32, 33). In the present study, we investigated molecular mechanisms whereby SHP-1 regulates BCR-induced JNK activation and apoptosis in B cells. Our findings indicate that SHP-1 negatively regulates JNK activation by modulating the association of BLNK with Nck, while exerting a positive effect on BCR-induced apoptosis.

WEHI-231 cells were obtained from the American Type Culture Collection (Manassas, VA) and maintained in RPMI 1640 (Sigma-Aldrich, St. Louis, MO) supplemented with 10% FBS (PAA Laboratories, Linz, Austria), 50 μM 2-ME, 100 μg/ml streptomycin, and 100 U/ml penicillin (complete medium).

Mice homozygous for the motheaten (me) mutation were obtained by mating C3HeBFeJ-me/+ breeding pairs, which were originally purchased from The Jackson Laboratory (Bar Harbor, ME) and were maintained in our animal facility in accordance with institutional policies for animal care. All mice were genotyped by using PCR amplification of tail DNA as described previously (34) and used at 2 wk of age. Splenic B cells were prepared by treating spleen cells from C3HeBFeJ-me/me and C3HeBFeJ-+/+ mice with anti-Thy-1.2 mAb and rabbit C. The resultant cells were >95% positive for surface IgM expression.

Goat anti-mouse IgM Ab was purchased from Cappel, Organon Teknika (Durham, NC). Anti-phosphotyrosine (PY) mAb (PY20) and anti-HPK1 Ab were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-Vav, anti-Nck, and anti-SHP-1 Abs were purchased from Upstate Biotechnology (Lake Placid, NY). Anti-flag epitope M2 mAb was obtained from Sigma-Aldrich. Anti-BLNK Ab was described previously (16). Anti-hemagglutinin (HA) mAb (12CA5) was purchased from Roche Diagnostics (Mannheim, Germany). Rabbit anti-mouse phospho-specific p38 and phospho-specific JNK Abs were purchased from New England Biolabs (Beverly, MA) and rabbit anti-mouse phospho-specific extracellular signal-regulated kinase (ERK) Ab was obtained from Promega (Madison, WI). Rabbit anti-mouse ERK-2, anti-JNK-2, and anti-p38 Abs were purchased from Santa Cruz Biotechnology. Alkaline phosphatase (AP)-conjugated goat anti-mouse IgG and AP-conjugated mouse anti-rabbit IgG were obtained from Bio-Rad (Richmond, CA) and Jackson ImmunoResearch Laboratories (West Grove, PA), respectively. HRP-conjugated anti-rabbit and anti-mouse IgGs were purchased from Santa Cruz Biotechnology.

Immunoprecipitation and Western blot analysis were performed as described previously (16). Briefly, cells (1 × 10⁷) were suspended in 1 ml of complete medium and stimulated for the indicated times with 20 μg/ml anti-IgM Ab. The reactions were stopped with 6 ml ice-cold PBS containing 1 mM Na₃VO₄ and 2 mM EDTA (PBS-VE). After washing twice with PBS-VE, cells were lysed in 500 μl of TNE buffer (1% Nonidet P-40, 10 mM Tris-HCl (pH 7.5), 150 mM NaCl, 2 mM Na₃VO₄, and 2 mM EDTA) and centrifuged. The supernatants were immunoprecipitated with protein G-Sepharose coupled with anti-flag mAb or Abs against BLNK and Nck. Immunoprecipitates were boiled in reducing Laemmli SDS sample buffer and subjected to 10% SDS-PAGE, followed by transfer to nitrocellulose membranes. The membranes were incubated with anti-PY mAb, anti-flag mAb, anti-BLNK Ab, or anti-Nck Ab, visualized by incubating with AP-conjugated goat anti-mouse IgG or AP-conjugated mouse anti-rabbit IgG, and developed with nitroblue tetrazolium and 5-brono-4-chloro-3-indoyl phosphate. Alternatively, the membranes were incubated with HRP-conjugated anti-mouse IgG or anti-rabbit IgG and then visualized using an ECL Western blot detection kit (Amersham Pharmacia Biotech, Piscataway, NJ).

Expression constructs for flag-tagged SHP-1-wild type (wt), SHP-1-C/S, and BLNK-wt have been described elsewhere (16). A flag-BLNK cDNA in which aa 1–190 and aa 130–312 were deleted was generated by PCR using BLNK-wt cDNA as a template and subcloned into pEF-flag/stop vector (a gift from Dr. G. Koretzky, University of Pennsylvania, Philadelphia, PA), yielding pEF-flag-BLNKΔN and pEF-flag-BLNKΔPro, respectively. To generate an expression construct for Nck-wt, a PCR fragment containing the entire open reading frame of mouse Nck was cloned downstream of the EF promoter in pEF-flag/stop vector (pEF-flag-Nck-wt). A Nck cDNA with a deletion of the SH2 domain was generated by PCR, and the amplified product, encompassing aa 1–281 of Nck, was cloned in pEF-flag/stop vector, yielding pEF-flag-NckΔSH2. To introduce an arginine to lysine substitution at position 373 in the SH2 domain of BLNK and position 289 in the SH2 domain of Nck, in vitro mutagenesis was respectively performed on full-length BLNK and Nck cDNAs cloned in pBluescript using a Gene Editor in vitro Site-directed Mutagenesis System (Promega). Mutation was verified by sequencing and the mutated cDNA fragments were then cloned in pEF-flag/stop vector, yielding pEF-flag-BLNK-RK and pEF-flag-Nck-RK. The expression plasmid encoding a HA-tagged dominant-negative SEK1 (SEK-DN) has been described (35). Transient transfection was performed as described previously (16). The expression level of transfected genes was 20–22% as revealed by flow cytometric analysis with cells transfected with pEF-flag-green fluorescence protein.

TNE-soluble supernatants from cells, either left unstimulated or stimulated with anti-IgM for 5 min, were subjected to 10% SDS-PAGE and transferred to nitrocellulose membranes. The blots were incubated first with Abs against ERK, JNK, p38, and their phosphorylated forms, then with HRP-conjugated anti-rabbit IgG, after which they were visualized using an ECL Western blot detection kit. The intensity of each band was measured with a Bio-Rad imaging densitometer. Under the conditions examined, the intensity of bands was linearly correlated to the amount of samples. The results were expressed as fold activation, with the intensity of unstimulated cells being assigned a value of 1.

Cells (1.5 × 10⁶) were suspended in 5 ml of complete medium and either left untreated or stimulated with 5 μg/ml anti-IgM Ab for 48 h. The cells were then harvested, washed twice with PBS, and subjected to DNA fragmentation assays (36) and TUNEL assays, as previously described (37).

Our previous work demonstrated that BCR-induced JNK activation is selectively enhanced in SHP-1-C/S transfectants, suggesting a negative regulatory role for SHP-1 in JNK activation (16). To confirm the physiological relevance of SHP-1 regulation of JNK activation, we used splenic B cells from SHP-1-deficient C3HeBFeJ-me/me and control C3HeBFeJ-+/+ mice. B cells were cultured with or without anti-IgM Ab for 5 min and then phosphorylation of total cellular proteins were examined by Western blot with anti-PY mAb. As shown in Fig. 1,A, tyrosine phosphorylation was constitutively enhanced in me/me mice, suggesting a negative regulatory role of SHP-1 in B cells. Activation of MAPK family members was assayed by immunoblotting with Abs against activated forms of ERK, JNK, and p38. As in WEHI-231 cells, SHP-1 regulates activation of JNK, but not ERK or p38 (Fig. 1 B). Furthermore, in me/me B cells, JNK was constitutively activated and further enhanced upon BCR ligation. The level of JNK activation, either without or with BCR ligation, was higher than that of control B cells. These results corroborate the physiological significance of negative, selective regulation of JNK by SHP-1.

Having confirmed negative regulation of JNK activation by SHP-1 in B cells, we assessed the mechanisms of SHP-1 action using WEHI-231 cells. Initially, we asked whether increased tyrosine phosphorylation of BLNK (16) was responsible for the enhanced JNK activation. WEHI-231 cells, transiently transfected with pEF-flag-tagged SHP-1-wt or SHP-1-C/S along with pEF-flag-tagged wild-type or mutant forms of BLNK, were stimulated with anti-IgM Ab for 5 min, after which activation of MAPK family members was measured by immunoblotting with anti-phospho-ERK, -JNK, and -p38 Abs (Fig. 2). The mutant forms of BLNK used in this study lacked the NH₂-terminal acidic region (BLNKΔN) or proline-rich region (BLNKΔPro) or contained an arginine to lysine substitution at the position 373 in the SH2 domain (BLNK-RK; Fig. 2,A). Since arginine 373 is the predicted phosphate binding site, BLNK-RK would be expected to lack the ability to bind phosphotyrosine-containing proteins. As shown in Fig. 2,B, activation of ERK and p38 was not significantly different among transfectants. By contrast, JNK activation was affected by various transfections. In our system, JNK2 was predominantly activated. As shown in Fig. 2,B (top panel, sample 1 vs sample 2), JNK2 was more strongly activated in transfectants with BLNK-wt and this enhancement was inhibited by cotransfection with SHP-1-wt (sample 2 vs sample 3). However, transfection with BLNK-wt and SHP-1-C/S reversed this inhibition (Fig. 2,B, sample 3 vs sample 4) and the level of JNK2 activation was higher than that observed in transfectants with BLNK-wt only (Fig. 2,B, sample 2 vs sample 4). This enhanced JNK2 activation was reduced in cells expressing SHP-1-C/S and BLNKΔN (Fig. 2 B, sample 4 vs sample 5), whereas the expression of BLNKΔPro or BLNK-RK along with SHP-1-C/S exhibited little effects on BCR-induced JNK2 activation. This result suggests that the NH₂-terminal region of BLNK containing tyrosine residues plays an essential role in JNK activation and that the enhanced JNK activation seen in SHP-1-C/S-expressing cells might be due to phosphotyrosine-based binding of signaling molecules to BLNK.

Several upstream regulators for JNK have been identified in various cellular systems, including Vav, Nck, HPK1, and TRAF2. Because tyrosine phosphorylation of BLNK is required for JNK activation, we first attempted to identify such upstream regulators in our system by identifying molecules whose association with BLNK was enhanced in SHP-1-C/S-expressing cells upon BCR ligation. WEHI-231 cells, transfected with BLNK-wt along with SHP-1-wt or SHP-1-C/S, were either left untreated or stimulated with anti-IgM Ab for 2 or 7 min, after which immunoprecipitated BLNK was blotted with Abs against candidate proteins. Although we could not detect an association between BLNK and TRAF2 (data not shown), BLNK was found to bind Nck within 2 min of BCR ligation, and this binding was significantly enhanced in SHP-1-C/S-expressing cells (Fig. 3,A). Reciprocal experiments in which anti-Nck immunoprecipitates were blotted with anti-BLNK Ab again showed the association between BLNK and Nck to be increased in SHP-1-C/S-expressing cells than in SHP-1-wt-expressing cells (Fig. 3 B).

Given that the NH₂-terminal acidic region of BLNK is required for the enhancement of BCR-induced JNK activation seen in SHP-1-C/S-expressing cells, it is likely that the SH2 domain of Nck is involved in this association. We tested this possibility using a Nck mutant possessing an arginine to lysine substitution in its SH2 domain (Nck-RK), which would eliminate the binding of the SH2 domain to phosphotyrosine-containing proteins (38). WEHI-231 cells, transfected with a combination of SHP-1-wt or SHP-1-C/S plus Nck-wt or Nck-RK, were cultured without or with anti-IgM Ab, after which binding of BLNK to Nck was examined (Fig. 3 C). We found that association of Nck-wt with endogenous BLNK was higher in SHP-1-C/S transfectants than in SHP-1-wt transfectants. In this experiment, association between Nck and BLNK was constitutively observed. One of the possibilities is that BLNK was phosphorylated, albeit at a low level, before BCR ligation. Additionally, overexpression of Nck-wt may contribute to the detectable level of association between BLNK and Nck. Significantly, even under these conditions, association between BLNK and Nck was completely abrogated when Nck-RK was introduced. It is thus very likely that Nck binds to BLNK via its SH2 domain and this association is dependent on tyrosine phosphorylation of BLNK.

The finding that Nck associates with BLNK prompted us to examine how Nck regulates JNK activation following BCR ligation. To address this question, we investigated the effect of Nck expression on BCR-induced JNK activation (Fig. 4,A). WEHI-231 cells were transiently transfected with a combination of SHP-1-wt or SHP-1-C/S and Nck-wt, Nck-RK, or a Nck mutant lacking its SH2 domain (NckΔSH2), after which BCR-induced JNK activation was compared. Both Nck-RK and NckΔSH2 were expected to act as dominant-negatives by attenuating the association of Nck with tyrosine-phosphorylated BLNK (38). We found that Nck-wt transfection strongly enhanced BCR-induced JNK2 activation (Fig. 4,B, top, sample 1 vs sample 2) and this enhancement was diminished by transfection with Nck-RK or NckΔSH2 (Fig. 4,B, sample 2 vs samples 3 and 4). Similarly in transfectants with SHP-C/S, enhanced JNK2 activation was significantly attenuated in cells expressing NckΔSH2 and Nck-RK (Fig. 4,B, sample 7 vs samples 8 and 9). The fact that inhibition was more pronounced in NckΔSH2-expressing cells than in Nck-RK-expressing cells may reflect higher expression of NckΔSH2. Again, there was no significant difference in the level of ERK and p38 activation among these transfectants (Fig. 4 B). Taken together, these results suggest that Nck recruited to tyrosine-phosphorylated BLNK via its SH2 domain acts as an upstream mediator of BCR-induced JNK activation.

WEHI-231 cells have characteristics of immature B cells and therefore are susceptible to growth arrest and apoptotic cell death upon BCR stimulation (36, 39). To clarify how SHP-1 regulates the final outcome of BCR signaling through dephosphorylation of BLNK, we examined the effect of transient SHP-1-C/S expression on BCR-induced apoptosis. WEHI-231 cells were transfected with vector alone, SHP-1-wt, or SHP-1-C/S, after which they were left unstimulated or stimulated with anti-IgM Ab for 48 h and then subjected to DNA fragmentation and TUNEL assays. As shown in Fig. 5, in contrast to augmented BCR-induced apoptosis in SHP-1-wt-expressing cells, SHP-1-C/S expression significantly inhibited BCR-induced apoptosis, implying a positive regulation of BCR-induced apoptosis by SHP-1.

The results thus far suggest that BCR-induced JNK activation is enhanced in SHP-1-C/S transfectants, while BCR-induced apoptosis is suppressed. If JNK activation is directly involved in the inhibition BCR-induced apoptosis, specific inactivation of JNK should increase cell death following BCR ligation. To test this possibility, we transfected WEHI-231 cells with SEK-DN (40) to inhibit an upstream activator of JNK (Fig. 6,A). As shown in Fig. 6,B, transfection of SEK-DN significantly reduced JNK activation, but had little effect on ERK and p38 activities (data not shown). Significantly, it partially reversed the suppression of BCR-induced apoptosis in cells overexpressing SHP-1-C/S (Fig. 6, C and D, samples 4 vs 5), suggesting that JNK activation directly regulates apoptosis in WEHI-231 cells in a negative fashion.

Finally, to assess the involvement of Nck adaptor protein in the regulation of BCR-induced apoptosis, we transiently transfected WEHI-231 cells with the aforementioned forms of Nck (Fig. 3) and performed DNA fragmentation and TUNEL assays. As shown in Fig. 7, BCR-induced apoptosis was significantly diminished in cells expressing Nck-wt, whereas the expression of Nck-RK and NckΔSH2 reversed this effect. These data, along with those in Fig. 5, suggest that there is an inverse correlation between JNK activation and apoptosis following BCR ligation, and that the SH2 domain of Nck is directly involved in SHP-1-mediated regulation of apoptosis through its interaction with BLNK.

The immunological abnormalities found in motheaten and motheaten viable mice are caused by point mutations in the gene encoding SHP-1 (34, 41), which is implicated in the regulation of AgR-mediated signaling events in lymphocytes. Biochemical and genetic analyses revealed SHP-1 to be involved in the negative regulation of BCR-induced signaling, an effect mediated mainly by its association with ITIMs found in inhibitory molecules such as CD22 (9, 10), CD72 (11, 12), and PIR-B (13). It is known that SHP-1 recruited to these molecules via its SH2 domain is activated and dephosphorylates tyrosine residues in several signaling molecules, including Igα/Igβ, Lyn (42), Syk (15), and BLNK (16), thus acting as an inhibitor of BCR-driven activation events. However, the precise molecular mechanisms by which SHP-1 regulates downstream signaling pathways are still largely unknown.

We previously demonstrated that cells expressing a form of SHP-1 lacking PTP activity (SHP-1-C/S) exhibited increased tyrosine phosphorylation of BLNK and selectively enhanced JNK activation upon BCR ligation. In this study, we first assessed the physiological relevance of the findings in a B cell line using splenic B cells from SHP-1-deficient me/me mice. In the event, JNK, but not ERK or p38, was constitutively activated and further induced upon BCR ligation in me/me B cells, corroborating a selective, negative role of SHP-1 in B cells. Phosphorylation of BLNK, however, could not be examined in splenic B cells, because BLNK was not effectively immunoprecipitated in our hands. We therefore focused, in this study, on WEHI-231 cells to investigate molecular mechanisms linking BLNK tyrosine phosphorylation to JNK activation and to apoptotic processes. We found that expression of a BLNK mutant lacking the NH₂-terminal region inhibited the enhanced JNK activation seen in SHP-1-C/S-expressing cells, indicating that this region of BLNK containing six tyrosine residues is required for the enhanced JNK activation seen in SHP-1-C/S-expressing cells. In addition, Nck adaptor protein was clearly shown to associate with BLNK in a tyrosine phosphorylation-dependent manner. Cells expressing Nck SH2 mutants that do not interact with phosphotyrosine-containing proteins did not show enhanced JNK activation upon BCR ligation, even when cotransfected with SHP-1-C/S. It is thus concluded that among several upstream activators of JNK, Nck is a candidate protein linking BLNK with JNK activation.

Nck is a ubiquitously expressed adaptor protein possessing one SH2 and two SH3 domains (43). Since it has no catalytic activity, Nck might regulate signaling processes by coupling a catalytic component, which binds to its SH3 domains, to tyrosine-phosphorylated proteins. For example, NIK is a Ste20-family serine/threonine kinase that is constitutively associated with Nck (44). As in the case of other members of the Ste20 protein kinase family, overexpression of NIK itself activates the MAPK kinase kinase 1/SEK1/JNK pathway but fails to activate ERK or p38 cascade (44). Thus, SHP-1 may regulate the amount of the Nck-NIK complex recruited to phosphorylated tyrosine residues on BLNK by controlling the level of tyrosine phosphorylation of BLNK during BCR signaling.

The MAPK module consists of three protein kinases that sequentially phosphorylate and activate downstream kinase cascade, MAP kinase kinase kinase (or MAPK kinase kinase), MAP kinase kinase (MAPKK or MEK), and MAPK (45). MAPK is phosphorylated at threonine and tyrosine residues by MAPKK to be activated. In the case of JNK, 13 MAPKKKs, 2 MAPKKs, and 3 JNKs have been identified in the mammalian cells (24). Although the kinase cascade leading to JNK activation is highly conserved, the upstream components that regulate the JNK pathway are diverse, depending on cell types and types of stimulation. Accumulating evidence indicates that the Rho family small GTPases (25, 26) and the TRAF group of adaptor proteins (28), respectively, mediate the activation of JNK in receptor-type PTK and cytokine receptor signaling pathways. Rho family GTPase is of particular interest to us because Vav, a Rho family guanine nucleotide exchange factor, is activated by tyrosine phosphorylation and is demonstrated to interact with tyrosine-phosphorylated BLNK through its SH2 domain (20). However, we could not detect differences in the amount of Vav1 recruited to BLNK in cells expressing SHP-1-wt or SHP-1-C/S (data not shown). Nevertheless, the observation that expression of the dominant-negative form of Vav1, possessing the RK mutation in its SH2 domain, inhibited enhanced JNK activation in SHP-1-C/S-expressing cells suggests that Vav1 might be involved in the pathways leading to JNK activation. The contribution of Vav to this signaling pathway is under study.

Recently, HPK1, a Ste20 family protein kinase (29, 30), was shown to be involved in BCR- and TCR-mediated signaling (46). Although HPK1 associates with BLNK and SLP-76 upon AgR stimulation in B and T cells, respectively, their association is quite different from that between BLNK and Nck or Vav. Following AgR ligation, HPK1 is phosphorylated at tyrosine 379 and subsequently interacts with the SH2 domains of BLNK and SLP-76, which in turn induces full activation (47, 48). Consistent with this model, association between BLNK and HPK1 was observed after BCR stimulation in our system, but the amount of HPK1 in the anti-BLNK immunoprecipitates was similar in lysates from SHP-1-C/S- and SHP-1-wt-expressing cells (data not shown). Interaction of BLNK with HPK1 may thus be independent of SHP-1 action.

Finally, we focused on the potential role of SHP-1 as a regulator of BCR-induced apoptotic signals. Using WEHI-231 cells, we clearly demonstrated that BCR-induced apoptosis is enhanced in SHP-1-wt transfectants but inhibited in SHP-1-C/S-expressing cells, suggesting a positive role for SHP-1 in this process. Interestingly, the inhibition of apoptosis seen in SHP-1-C/S transfectants was reversed by the expression of SEK-DN, implying that JNK plays a positive role in this pathway. Despite extensive study, the precise roles of the MAPK family members in apoptosis or cell survival remains controversial, possibly because of differences in the experimental procedures, cell types, and ligands used. The JNK pathway has been implicated in both apoptosis and cell survival signaling (49). Given that JNK is activated by exposing cells to stress, it is likely that JNK may mediate some effects of stress itself which in turn results in the induction of apoptosis. Alternatively, JNK activation may serve as a protective response to stress, facilitating antiapoptotic processes. Evidence favoring the latter comes from the observation that apoptosis is enhanced in the developing forebrain of jnk1/jnk2 double null mouse embryos (50, 51) and that integrin-mediated survival signaling is mediated by the JNK signaling pathway (52). Other studies have demonstrated that there is a positive correlation between ERK activation and BCR-induced apoptosis (32, 33) and that BCR-induced activation of JNK and p38 correlates apoptosis (31). The definitive reasons for these discrepancies are unclear at present.

In summary, we have demonstrated that SHP-1 negatively regulates BCR-induced JNK activation by dephosphorylating its physiological substrate, BLNK, thereby attenuating the association of BLNK with Nck. This negative regulation of JNK activation by SHP-1 contributes, at least in part, to the enhancement of BCR-induced apoptotic cell death. The molecular events linking JNK activation to apoptosis remains to be elucidated.

We thank Dr. G. Koretzky for providing us with valuable reagents.