Nucleocytoplasmic Trafficking of the Syk Protein Tyrosine Kinase

Abstract

The protein tyrosine kinase Syk couples the B-cell receptor (BCR) for antigen to multiple intracellular signaling pathways and also modulates cellular responses to inducers of oxidative stress in a receptor-independent fashion. In B cells, Syk is found in both the nuclear and cytoplasmic compartments but contains no recognizable nuclear localization or export signals. Through the analysis of a series of deletion mutants, we identified the presence of an unconventional shuttling sequence near the junction of the catalytic domain and the linker B region that accounts for Syk's subcellular localization. This localization is altered following prolonged engagement of the BCR, which causes Syk to be excluded from the nucleus. Nuclear exclusion requires the receptor-mediated activation of protein kinase C and new protein synthesis. Both of these processes also potentiate the activation of caspase 3 in cells in response to oxidative stress in a manner that is dependent on the localization of Syk outside of the nucleus. In contrast, restriction of Syk to the nucleus greatly diminishes the stress-induced activation of caspase 3.

Syk is a 72-kDa protein tyrosine kinase with an N-terminal, tandem pair of SH2 domains separated by a long linker (linker B) from a C-terminal catalytic domain (40). Syk plays a critical role in B cells in coupling the B-cell receptor (BCR) for antigen to multiple downstream signaling pathways, including the activation of phospholipase Cγ, mobilization of calcium from intracellular stores, and activation of the Ras/extracellular signal-regulated kinase (ERK), phosphatidylinositol 3-kinase, and NF-κB pathways (1, 4, 7, 18). Upon receptor cross-linking, the tandem SH2 domains of Syk bind two phosphorylated tyrosines in the conserved immunoreceptor tyrosine-based activation motifs (ITAMs) located within the cytoplasmic domains of the immunoglobulin-α (CD79a) and immunoglobulin-β (CD79b) components of the BCR complex (3). Subsequent phosphorylation of Syk induces its activation, leading to the binding and/or phosphorylation of adaptor proteins and downstream effectors (6, 7, 9, 13, 16, 20, 32, 33). Consequently, Syk-deficient B cells are largely nonresponsive to BCR aggregation (36) and B-cell development in Syk-deficient mice is largely blocked at the transitions from pro-B cells to pre-B cells and from immature to mature B cells (5, 39).

Increasing evidence indicates that Syk also has fundamental cellular functions that are receptor and ITAM independent. Paradoxically, in B lymphocytes, the expression of Syk is required for BCR-induced apoptosis, but its presence also protects cells from apoptosis induced by ceramide, osmotic stress, or oxidative stress (9, 25, 27, 28, 35, 37). While the activation of Syk by BCR engagement requires both intact SH2 domains, its participation in oxidative stress signaling does not, suggesting an ITAM-independent mechanism (8, 26).

In lymphoid and epithelial cells, Syk has been reported to reside in both the nucleus and cytoplasm (21, 41, 43), as has its close family member Zap-70 (34). In breast cancer cells, the expression of Syk and its localization to the nucleus have been correlated with the repression of invasive tumor growth (43). An inverse correlation between nuclear Syk and lymph node metastasis also has been observed in gastric cancer patients (44). However, little is known as to how the movement of Syk between the cytoplasm and nucleus is regulated.

In B cells, engagement of the BCR recruits Syk from both the cytoplasm and nucleus to the aggregated BCR complex (21). Syk returns rapidly to both compartments following receptor internalization. Interestingly, Syk then becomes excluded from the nucleus at longer times following BCR engagement (21). Thus, mechanisms must exist that allow Syk to readily enter and exit the nucleus despite the fact that it is too large to diffuse freely through the nuclear pore complex (45) and mechanisms must exist that regulate this trafficking. What roles the inclusion or exclusion of Syk from the nucleus play in B-cell signaling is not understood.

In this study, we have determined that sequences localized to a region of the Syk molecule near the junction of the linker B and catalytic domain are responsible for its distribution between the nucleus and cytoplasm. We show further that the distribution of Syk between the nucleus and cytoplasm is modulated by signals sent downstream from the activated BCR that require the receptor-mediated activation of protein kinase C (PKC) and the induction of new protein synthesis. We also provide evidence that this redistribution of Syk reflects or promotes its participation in a cytoplasmic signaling pathway that couples oxidative stress to the activation of caspase 3. Our data also indicate an alternative Syk-dependent pathway in the nucleus that mitigates the stress-induced activation of caspase 3.

MATERIALS AND METHODS

Plasmids and constructs.

Preparation of a plasmid for the expression of enhanced green fluorescent protein (EGFP)-tagged Syk using the pEGFP-N2 vector (Clontech) was described previously (21). Two oligonucleotides that encode the nuclear localization signal (NLS) from simian virus 40 large T antigen were synthesized, annealed, and inserted in frame into the BsrGI and NotI sites at the 3′ end of the coding sequence for Syk-EGFP to make pSyk-EGFP-NLS. cDNAs for the expression of all EGFP-tagged Syk deletion mutants as well as SykB-EGFP were constructed by a PCR strategy using the Syk-EGFP cDNA as a template. In brief, regions upstream and downstream of the sequences to be deleted were amplified by PCR. The resulting products were ligated, amplified, and inserted into the Syk-EGFP cDNA to replace the sequences coding for the complete Syk linker region.

A cDNA fragment encoding β-galactosidase was obtained by PCR using vector pYES3/CT/lacZ (Invitrogen) as a template and inserted into the pEGFP-N2 vector, resulting in a fusion between β-galactosidase and EGFP (β-Gal-EGFP). A DNA fragment encoding Syk residues 259 to 364 was obtained by PCR using Syk-EGFP cDNA as a template and inserted into the corresponding site of β-galactosidase-EGFP, resulting in β-Gal-Syk(259-364)-EGFP. Other cDNA constructs, β-Gal-Syk(259-295)-EGFP, β-Gal-Syk(296-331)-EGFP, β-Gal-Syk(332-364)-EGFP, and β-Gal-Syk(259-331)-EGFP, were made using the same strategy. All DNA constructs were confirmed by sequencing.

Cells, transfections, and stable cell lines.

DT40 B-cell lines lacking Syk (36) were cultured as described previously (13). Syk-deficient DT40 cells (10⁷) were transfected with 20 to 30 μg of each DNA construct by electroporation (16). For the generation of stable cell lines, Syk-deficient DT40 cells were electroporated with 25 μg of the various Syk-EGFP expression constructs and 2.5 μg of pBabePuro. Cells were selected by treatment with puromycin (0.5 μg/ml), cloned, and screened for Syk-EGFP expression by fluorescence microscopy and Western blotting. Where indicated, some experiments used pools of stably transfected, antibiotic-selected cells. Primary B cells were enriched from mouse splenocytes by a panning procedure described previously (30).

DT40 cells were stimulated with goat anti-chicken immunoglobulin M (IgM) antibodies (Bethyl) or treated with phorbol myristate acetate (PMA). All the stimulations were performed on 10⁶ cells in 0.5 ml chicken serum-free medium at 37°C. Leptomycin B was a gift from Minoru Yoshida (University of Tokyo, Tokyo, Japan) or was purchased from Sigma. U0126, SB203580, Gö6976, Gö6983, and wortmannin were purchased from Calbiochem.

Fluorescence microscopy.

Cells were allowed to adhere to coverslips coated with poly-l-lysine (100 μg/ml; Sigma) for 10 min. Cells were fixed in 3.7% paraformaldehyde (Sigma) in phosphate-buffered saline (PBS) for 10 min at room temperature, washed three times with PBS, stained with 4′,6′-diamidino-2-phenylindole (DAPI; Sigma), and viewed by fluorescence microscopy. For indirect immunostaining, cells were permeabilized with 1 ml 0.5% Triton X-100 in PBS for 15 min and then treated with 1 ml methanol for 10 min. Cells were incubated with blocking buffer (3% bovine serum albumin, 1% goat serum in PBS) for 1 h, followed by incubation with primary (anti-Syk N-19) (Santa Cruz) and secondary (fluorescein isothiocyanate-labeled goat anti-rabbit IgG) (Sigma) antibodies. Cells were washed and stained with DAPI as indicated above.

For quantitative measurements of protein distribution, the fluorescence intensities in the nucleus and cytoplasm of individual cells were measured using Northern Eclipse 5.0 software (Empix Imaging). The ratio between the fluorescence intensity of the nucleus and that of the cytoplasm (N/C) was calculated; 25 images of cells expressing Syk-EGFP-NLS or Syk-EGFP were measured and their average N/C values were used as standards. The localization of an unknown protein was compared to that of Syk-EGFP-NES and Syk-EGFP as follows:

The results represent the mean and standard error of measurements from a minimum of 10 randomly selected cells.

Cellular fractionation.

B cells or DG75 cells (10⁷) were washed twice with PBS and resuspended in 1 ml of an ice-cold mixture containing 10 mM HEPES, pH 7.9, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, 10 μg/ml aprotinin, 10 μg/ml leupeptin, and 1 mM sodium orthovanadate. Cells were allowed to swell on ice for 15 min. NP-40 was added to a final concentration of 0.4%. Cells were vortexed vigorously for 10 s and then centrifuged at 18,000 × g for 30 s. The supernatant was saved as the cytosol/membrane fraction. The nuclear pellet was washed twice in lysis buffer containing 0.2% NP-40 and then resuspended in 0.1 ml of a cold mixture containing 20 mM HEPES, pH 7.9, 0.4 M NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluoride. The tube was vigorously rotated at 4°C for 15 min and centrifuged for 5 min at 18,000 × g. The supernatant contained the extracted nuclear fraction. Both the cytosol/membrane and the nuclear fractions were separated by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) and analyzed by Western blotting. Antibodies to p38, SP1, and Lyn were obtained from Santa Cruz.

Measurement of NFAT activity.

Syk-deficient DT40 cells (10⁷) were transfected by electroporation with vectors containing cDNAs for the various EGFP-tagged forms of Syk and Syk mutants (20 μg) and an NFAT-luciferase reporter plasmid (10 μg). Cells were harvested 24 h following transfection, plated at a density of 10⁶ cells/ml, and activated with anti-IgM antibody (5 μg/ml) or a mixture of PMA (50 ng/ml) and ionomycin (1.0 μM) for 6 h at 37°C. Luciferase activity was determined using a luciferase assay system kit (Promega). Luciferase activity is expressed as a fraction of the activity observed after activation by PMA plus ionomycin.

Metabolic labeling.

We labeled 10⁶ cells in 500 μl chicken serum-free medium with 28 μCi of an amino acid mixture containing [³⁵S]methionine and [³⁵S]cysteine (Amersham). Cells were treated with cycloheximide (Sigma) (10 μg/ml) for 20 min, harvested, resuspended in PBS, and spotted on small squares of filter paper (Whatman 3MM). The filter papers were washed once in cold 10% trichloroacetic acid for 15 min, twice in 5% trichloroacetic acid for 15 min, and once in ethanol for 15 min. The filter papers were dried and the incorporation of ³⁵S-labeled amino acids into protein was measured using a scintillation counter.

Detection of activated caspase 3.

Caspase 3 activity in cell lysates was measured using a colorimetric assay kit (Biovision, Mountain View, CA) according to the manufacturer's instructions. For inhibitor studies, cycloheximide (Sigma) (10 μg/ml) or Gö6983 (10 μM) was added 30 min prior to the addition of PMA (50 ng/ml). Cells were treated for 60 min with PMA and then exposed to H₂O₂ (50 μM) for an additional 8 h. Antibodies to total caspase 3 and activated caspase 3 were obtained from Cell Signaling Technologies. The caspase 3 inhibitor zDEVD-FMK (benzyloxycarbonyl-Asp-Glu-Val-Asp-fluoromethyl ketone) was obtained from R&D Systems and used at a concentration of 1 μM.

RESULTS

Endogenous Syk localizes to the nucleus and cytoplasm.

A Syk-EGFP fusion protein is distributed in both the cytoplasmic and nuclear fractions when expressed in cultured B cells (21). To compare this distribution to that of the endogenous enzyme, we examined the localization of Syk in human DG75 B cells by indirect immunofluorescence. The staining pattern for Syk in DG75 cells indicated that the endogenous kinase was localized throughout the interior of the cell in both the cytoplasm and nucleus (Fig. 1A). No staining was observed in the absence of the primary antibody.

FIG. 1.

Endogenous Syk is found in the nucleus and cytoplasm of B cells. (A) Syk was detected in fixed and permeabilized DG75 B cells using the anti-Syk N-19 primary antibody followed by a fluorescein isothiocyanate (FITC)-conjugated goat anti-rabbit IgG secondary antibody (shown on the left). The immunostaining in the absence of primary antibody alone is shown on the right. The nucleus was detected by staining with DAPI. (B) DG75 B cells were fractionated into nuclear (N) and cytoplasmic/membrane (C) fractions and probed by Western blotting with anti-Syk, anti-p38, anti-SP1, and anti-Lyn antibodies. The migration positions of protein standards of known mass (kilodaltons) are indicated. (C) Murine spleen B cells were fractionated into nuclear (N) and cytoplasm/membrane (C) fractions and analyzed by Western blotting for the presence of Syk, p38, and SP1.

To confirm the localization of Syk, we separated DG75 cells into nuclear and nonnuclear (cytoplasm plus membrane) fractions and immunoblotted proteins from each with an anti-Syk antibody. As shown in Fig. 1B, Syk was present in both the cytosolic and nuclear fractions. In this experiment, we used p38 as a cytosolic protein marker (14), the transcription factor SP1 as a nuclear protein marker (24), and Lyn as a membrane marker. We observed a similar distribution of Syk in subcellular fractions prepared from resting murine spleen B cells (Fig. 1C). These studies indicate that endogenous Syk localizes to both the cytoplasm and nucleus and Syk-EGFP is a suitable marker for investigating the subcellular localization of Syk.

Linker B region is necessary and sufficient for Syk's nuclear localization.

To identify the region of Syk required for nucleocytoplasmic trafficking, we first deleted the entire linker B region (264 to 359), leaving five amino acids at each end for structural considerations (Fig. 2A), to generate Syk(linkerless)-EGFP. In Syk-deficient DT40 cells transiently transfected with a vector coding for Syk(linkerless)-EGFP, the protein was localized exclusively in the cytoplasm, as evaluated by fluorescence microscopy.

FIG. 2.

Linker B region is both necessary and sufficient for the nuclear localization of Syk. (A) Schematic diagram of Syk-EGFP and Syk(linkerless)-EGFP. (B) The localization of Syk(linkerless)-EGFP (LL) transiently expressed in Syk-deficient DT40 cells was visualized by fluorescence microscopy and analyzed by quantitative analysis of distribution in comparison to Syk-EGFP (wild type [WT]) and Syk-EGFP-NES (NE). (C) Syk-deficient DT40 cells were cotransfected with an NFAT-luciferase reporter plasmid without (Syk-) or together with plasmids coding for Syk-EGFP (wild type) or Syk(linkerless)-EGFP (LL). Transfected cells either remained unactivated (−) or were activated by anti-IgM antibody (+). Relative luciferase activity is reported as the activity observed under the experimental conditions divided by the activity produced in response to stimulation with a mixture of PMA and ionomycin. The values represent the mean and standard error of two separate analyses. The expression levels of the wild type and Syk(linkerless)-EGFP were detected by Western blotting using anti-Syk antibody (insert). (D) Schematic diagram of β-galactosidase-EGFP (ZG) and β-Gal-Syk(259-364)-EGFP (ZL). (E) The localization of ZG and ZL was visualized by fluorescence microscopy in transiently transfected, Syk-deficient DT40 cells. The locations of the EGFP fusion proteins are shown in green (GFP) and the location of nuclei is in blue (DAPI). (F) Quantitative analyses of the relative fluorescence intensities of ZG and ZL in nuclear versus cytoplasmic compartments in comparison to Syk-EGFP (wild type) and Syk-EGFP-NES (NE).

To verify a quantitative difference in the localization of Syk(linkerless)-EGFP and Syk-EGFP, we measured the intensity of fluorescence in both the nuclear and cytoplasmic compartments on a scale of 0 to 100, with 100 representing the distribution of Syk-EGFP and 0 representing the distribution of Syk-EGFP-NES, a form of Syk-EGFP to which a nuclear export signal was attached (Fig. 2B). As shown previously, Syk-EGFP-NES is excluded from the nucleus (21). To ensure that Syk(linkerless)-EGFP was active when expressed in cells, its ability to confer BCR-mediated activation of the NFAT transcription factor on Syk-deficient cells was evaluated. As shown in Fig. 2C, this mutant coupled the receptor to the expression of an NFAT-driven luciferase gene even more efficiently than Syk-EGFP even though its level of expression was somewhat lower, as determined by Western blotting (Fig. 2C). These results demonstrated that Syk(linkerless)-EGFP was a functional protein and that the linker B region was necessary for the localization of Syk to the nucleus.

To determine if the linker B region was sufficient for the localization of a heterologous protein to the nucleus, we generated a construct that fused EGFP to the C terminus of β-galactosidase with or without the linker B region (residues 259 to 364) separating the β-galactosidase and EGFP domains (Fig. 2D). DT40 cells transfected with these two constructs were examined by fluorescence microscopy (Fig. 2E). β-Gal-EGFP localized predominantly to the cytoplasm of DT40 cells. However, when the linker B region of Syk was inserted, the resulting protein distributed to both the cytoplasm and nucleus. Quantification of the protein distribution further confirmed that β-Gal-linker-EGFP (ZL) had a distinct distribution pattern compared with that of β-Gal-EGFP (ZG) (Fig. 2F). These results demonstrated that the linker B region of Syk was sufficient to bring a heterologous protein into the nucleus, resulting in a final distribution pattern similar to that of Syk-EGFP itself.

Region near the C terminus of linker B is necessary for Syk's nuclear localization.

To investigate which part of the linker was responsible for Syk's nuclear localization, we prepared a series of Syk linker B deletion mutants fused to EGFP (Fig. 3A). The subcellular localization of these deletion mutants was examined by fluorescence microscopy (Fig. 3B and D). Elimination of the first half of linker B (H1, Δ264-311) had no effect on the subcellular localization of the fusion protein. However, elimination of the second half of linker B (H2, Δ312-359) resulted in an exclusively cytoplasmic localization, indicating that the signals required for nuclear localization resided in the C-terminal half of linker B (Fig. 3B).

FIG. 3.

Region near the C terminus of linker B is necessary for Syk's nuclear localization. (A) Schematic diagram of Syk-EGFP (wild type [WT]), Syk(Δ264-311)-EGFP (H1), Syk(Δ312-359)-EGFP (H2), Syk(Δ264-295)-EGFP (TF), Syk(Δ296-331)-EGFP (TM), Syk(Δ332-359)-EGFP (TL), and Syk(Δ264-331)-EGFP (TD). (B) Syk-deficient DT40 cells transiently expressing H1 or H2 were examined by fluorescence microscopy. The locations of the EGFP fusion proteins are shown in green (GFP) and the location of nuclei is in blue (DAPI). (C) Quantification of the relative fluorescence intensities in the nucleus versus cytoplasm for cells expressing H1 and H2 compared to cells expressing Syk-EGFP (wild type) and Syk-GFP-NES (NE). (D) Syk-deficient DT40 cells transiently expressing TF, TM, TL, or TD were examined by fluorescence microscopy. (E) Quantification of the relative fluorescence intensities in the nucleus versus cytoplasm for cells expressing TF, TM, TL, and TD compared to cells expressing Syk-EGFP (wild type) and Syk-GFP-NES (NE). (F) Quantitative measurement of the nuclear/cytoplasmic fluorescence intensities of cells expressing SykB-EGFP compared to that of cells expressing Syk-EGFP (wild type [WT]) and Syk-EGFP-NES (NE).

To further narrow down the region required, linker B was divided into three parts and each was eliminated to generate mutants TF (Δ264-295), TM (Δ296-331), and TL (Δ332-359). The Syk mutants TF and TM were distributed in both the nucleus and cytoplasm, while TL was exclusively distributed in the cytoplasm. To verify that the final third of the linker was required for nuclear localization, a mutant lacking the first two-thirds of linker B, TD (Δ264-331), was prepared. TD was localized normally in both the nucleus and cytoplasm (Fig. 3D). Quantitative analyses of the distributions of the various mutants in the nucleus and cytoplasm verified that the region near the C terminus of linker B (332 to 359) was required for the nuclear localization of Syk (Fig. 3C and E).

SykB is not excluded from the nucleus.

SykB is an alternatively spliced variant of Syk lacking a stretch of 23 amino acids within the first third of the linker B region (19). This missing region has been reported to contain a potential bipartite nuclear localization signal important for the localization of Syk to the nucleus in breast epithelial cells (43). However, elimination of the first third of linker B did not alter the subcellular localization of Syk in B cells (Fig. 3D). To examine this more directly, we generated an EGFP-tagged version of SykB. A quantitative analysis indicated that the distribution of SykB-EGFP between the nuclear and cytoplasmic fractions was not significantly different from that of Syk-EGFP (Fig. 3F). These results demonstrated that the linker insert region missing from SykB was not responsible for the nuclear localization of full-length Syk in B cells.

Region near the C terminus of linker B is sufficient for Syk's nuclear localization.

To determine if the C-terminal region of linker B was sufficient to translocate a heterologous protein into the nucleus, we inserted different regions of linker B between β-galactosidase and EGFP (Fig. 4A) and expressed the proteins in DT40 cells. Mutants ZTF, ZTM, and ZTD, containing the first third, second third, and first two-thirds of linker B, respectively, were all excluded from the nucleus (Fig. 4B). The distribution of each was similar to that of β-Gal-EGFP (Fig. 4C). In contrast, mutant ZTL, which contained the final third of linker B (332 to 364), was distributed to both the nucleus and cytoplasm (Fig. 4B). This distribution was equivalent to that of Syk-EGFP (Fig. 4C). Thus, the C-terminal third of Syk linker B had the ability to bring a heterologous protein into the nucleus. Thus, this region was both necessary and sufficient for Syk's nuclear localization.

FIG. 4.

C terminus of linker B can direct a heterologous protein into the nucleus. (A) Schematic diagram of β-Gal-Syk(259-295)-EGFP (ZTF), β-Gal-Syk(296-331)-EGFP (ZTM), β-Gal-Syk(332-364)-EGFP (ZTL), and β-Gal-Syk(259-331)-EGFP (ZTD). (B) Syk-deficient DT40 cells expressing ZTF, ZTM, ZTL, or ZTD were examined by fluorescence microscopy. The locations of the EGFP fusion proteins are shown in green (GFP) and the location of nuclei is in blue (DAPI). (C) Quantification of the nuclear/cytoplasmic fluorescence intensities of ZTF, ZTM, ZTL, and ZTD relative to that of Syk-EGFP (wild type [WT]) and Syk-GFP-NES (NE).

Amino acids at the linker B catalytic domain junction are required for Syk's nuclear localization.

The experiments described above determined that amino acids 332 to 359 near the C terminus of Syk linker B were required for its nuclear localization. To further define this region, additional Syk deletion constructs were prepared in which different portions of this sequence were eliminated. Interestingly, deletion mutants A (Δ332-358), B (Δ332-351), C (Δ332-345), and D (Δ332-339) (Fig. 5A) were all distributed in both the nucleus and cytoplasm (Fig. 5B). Although the proportions of the proteins distributed into the nucleus were somewhat less than that of Syk-EGFP, particularly for mutant A (Fig. 5C), none of the deletion mutants (A to D) were excluded from the nucleus, indicating that amino acids 332 to 358 may aid Syk's nuclear localization but do not play a critical role.

FIG. 5.

Sequences near the C terminus of linker B are required for Syk's nuclear localization. (A) Schematic diagram of Syk(Δ332-358)-EGFP (A), Syk(Δ332-351)-EGFP (B), Syk(Δ332-345)-EGFP (C), and Syk(Δ332-339)-EGFP (D). (B) Syk-deficient DT40 cells transiently expressing mutants A, B, C, or D were examined by fluorescence microscopy. (C) Quantification of the relative fluorescence intensities of A, B, C, and D in the nucleus and cytoplasm compared to that of Syk-EGFP (wild type [WT]) and Syk-GFP-NES (NE). (D) Schematic diagram of Syk(Δ348-366)-EGFP (D19) and Syk(Δ356-366)-EGFP (D11). (E) Syk-deficient DT40 cells transiently expressing D19 or D11 were examined by fluorescence microscopy. The locations of the EGF -fusion proteins are shown in green (GFP) and the location of nuclei is in blue (DAPI). (F) Quantification of the fluorescence intensities of D19 and D11 in the nucleus and cytoplasm compared to that of Syk-EGFP (wild type [WT]) and Syk-GFP-NES (NE). (G) Quantification of the fluorescence intensities of Syk(linkerless)-EGFP (Δ264-359) and Syk(linkerless + L)-EGFP (Δ264-358) in the nucleus and cytoplasm. (H) Fluorescence microscopy of Syk-deficient DT40 cells transiently expressing Syk(linkerless)-EGFP and Syk(linkerless + L)-EGFP.

There were major differences between the distribution patterns of mutants TL (Δ332-359) and A (Δ332-358). While TL clearly showed nuclear exclusion, A was distributed in both the cytoplasm and nucleus. The single-amino-acid difference between TL and A is L359, which is located near the C-terminal end of Syk linker B. This suggests that the region of linker B near the junction between the linker and the catalytic domain is important for the localization of Syk.

To verify the importance of L359 and the C-terminal region of linker B, we added L359 back to the Syk(linkerless)-EGFP (Δ264-359) to generate Syk(linkerless + L)-EGFP (Δ264-358). Quantitative analyses of cells expressing the two proteins indicated that the addition of L359 increased the efficiency of nuclear import (Fig. 5G and H), but did not completely restore it to wild-type levels. These results indicated that the region near the junction between the extreme C-terminal region of linker B and the catalytic domain is important for directing the translocation of Syk into the nucleus, since Syk(linkerless +L)-EGFP lacks nearly the entire linker B region.

To confirm the importance of this region, we constructed two additional deletion mutants, D19 (Δ348-366) and D11 (Δ356-366) (Fig. 5D). These mutants retain the bulk of linker B but are missing residues at the junction of linker B and the catalytic domain. Both of these mutants were clearly excluded from the nucleus, as determined by fluorescence microscopy (Fig. 5E) and quantification studies (Fig. 5F). These results indicate that the region between amino acids 356 and 366 is a necessary part of the sequence required for Syk's nuclear localization. The region deleted in mutant D11 shares nine amino acids (EVYLDRSLL) with the C-terminal third of linker B that enabled the translocation of the fusion protein ZTL into the nucleus (Fig. 4B). Thus, this sequence is an essential part of Syk's nucleocytoplasmic trafficking signal.

This region of Syk at the junction of linker B and the catalytic domain is rich in leucines and more closely resembles a sequence required for Crm-1-mediated nuclear export than any identified nuclear localization sequence. To examine a possible role for Crm-1 in Syk's translocation, we treated cells expressing Syk-EGFP, TL, and, as a positive control, Syk-EGFP-NES with leptomycin B (LMB), a potent inhibitor of Crm-1-mediated export. Treatment of cells with LMB had no effect on the subcellular localization of Syk-EGFP, indicating that the movement of Syk into or out of the nucleus is unlikely to be mediated by Crm-1 (Fig. 6A). The exclusion of Syk-EGFP-NES from the nucleus was reversed by treatment with LMB, as expected, since the NES is recognized by Crm-1 (11, 17) (Fig. 6B). However, Syk-EGFP-NES was still distributed equally between the nucleus and cytoplasm in LMB-treated cells, indicating that an LMB-insensitive mechanism still allows shuttling of the protein into and out of the nucleus. The subcellular distribution of TL, which lacks the C-terminal third of linker B and is excluded from the nucleus, also was not altered by treatment with LMB (Fig. 6B). This result suggests that nucleus-excluded forms of Syk such as TL, which lacks amino acids 332 to 359, are not transported from the nucleus by Crm1-mediated export, but are likely unable to enter the nucleus.

FIG. 6.

Nucleocytoplasmic trafficking of Syk is not inhibited by LMB. (A) Syk-deficient DT40 cells expressing Syk-EGFP were treated with LMB for the indicated periods of time. Cells were fixed and examined by fluorescence microscopy. (B) Syk-deficient DT40 cells expressing Syk-EGFP-NES (NE) or TL were left untreated or treated with LMB (50 ng/ml) for 2 h and then examined by fluorescence microscopy.

Syk translocates from the nucleus to the cytoplasm upon prolonged treatment with PMA.

At prolonged times after treatment of cells with anti-IgM, Syk-EGFP becomes excluded from the nucleus in a large proportion of cells (21), suggesting that signals resulting from receptor engagement might control its localization. BCR engagement leads to the activation of multiple intracellular signaling pathways including the activation of phospholipase Cγ and the generation of diacylglycerol. To study a possible role for this pathway, we examined the effects of PMA, a potent diacylglycerol mimic, on the localization of Syk-EGFP. As shown in Fig. 7A, treatment of cells with PMA led to the exclusion of Syk-EGFP from the nucleus. Time course analyses indicated that PMA-induced nuclear exclusion was visible as early as 30 min and reached a peak at 90 min (Fig. 7B). The translocation of Syk was nearly maximal at a concentration of PMA of 25 ng/ml (Fig. 7C).

FIG. 7.

Syk-EGFP is excluded from the nucleus in cells treated with PMA. (A) Syk-deficient DT40 B cells expressing Syk-EGFP were treated with 100 ng/ml PMA for 0, 1, or 24 h and examined by fluorescence microscopy. (B) DT40 cells expressing Syk-EGFP were treated with 100 ng/ml PMA for the indicated periods of time. At each time point, the percentage of cells in which Syk-EGFP was excluded from the nucleus was determined. (C) DT40 cells expressing Syk-EGFP were treated with the indicated doses of PMA for 90 min. The percentage of cells in which Syk-EGFP was excluded from the nucleus was determined. For panels B and C, the reported percentages of cells showing nuclear exclusion of Syk-EGFP represent data from a typical experiment and include an examination in each case of >100 cells.

Since prolonged treatment of cells with PMA leads to the down-regulation of PKC, we treated cells with PMA for 24 h to test the possibility that the nuclear exclusion of Syk-EGFP was caused by a loss of PKC activity. However, no nuclear exclusion of Syk-EGFP was observed under these conditions (Fig. 7A). This suggested that PKC might be positively regulating the translocation of Syk from the nucleus. To investigate this further, we incubated cells expressing Syk-EGFP with two PKC inhibitors, Gö6976 and Gö6983, before treatment with PMA (Fig. 8A). Gö6976 inhibited the PMA-induced nuclear exclusion of Syk by 60%, while Gö6983 inhibited the redistribution by 85%. Treatment of cells with Gö6983 resulted in a dose-dependent decrease in PMA-induced Syk translocation (Fig. 8B). The subcellular localization of Syk was not affected when cells were treated with PKC inhibitors and not treated with PMA (data not shown).

FIG. 8.

Exclusion of Syk from the nucleus in response to PMA is mediated by PKC. (A) Syk-deficient DT40 B cells, transiently transfected with 20 μg of expression plasmid encoding Syk-EGFP, were incubated with dimethyl sulfoxide (DMSO) (solvent carrier), Gö6976 (1 μM), or Gö6983 (1 μM) for 20 min and then with PMA (25 ng/ml) for 90 min. (B) Cells transiently expressing Syk-EGFP were incubated with DMSO or Gö6983 at the indicated concentrations and then treated with PMA (50 ng/ml) for 90 min. (C) Cells transiently expressing Syk-EGFP were incubated with DMSO, U0126 (20 μM), or SB203580 (50 μM) for 30 min and then treated with PMA (25 ng/ml) for 90 min. (D) Cells transiently expressing Syk-EGFP were incubated with DMSO or wortmannin (100 nM) for 30 min and then treated with PMA (200 ng/ml) for 120 min. For all panels, cells were fixed and the localization of Syk-EGFP was examined by fluorescence microscopy. The reported percentages of cells in which Syk-EGFP was excluded from the nucleus represent data from typical experiments and include an examination in each case of >50 cells.

Besides acting as a PKC activator, PMA can also activate Ras guanyl nucleotide-releasing protein and the Ras/MEK/ERK pathway in immune cells (38). However, neither the MEK inhibitor U0126 nor the p38 inhibitor SB203580 blocked the PMA response (Fig. 8C). The ability of U0126 to block the PMA-stimulated activation of ERK was verified by Western blotting with anantibody to phosphorylated ERK (data not shown). In addition, the phosphatidylinositol 3-kinase inhibitor wortmannin had no effect on PMA-induced Syk translocation (Fig. 8D). These results indicated that the effect of PMA on the nuclear exclusion of Syk was mediated by one or more members of the PKC family.

PKC mediates the translocation of Syk in B cells treated with anti-IgM.

The role of PKC in the PMA-induced translocation of Syk suggested that PKC might also mediate the BCR-induced translocation. DT40 B cells expressing Syk-EGFP were treated with anti-IgM for 90 min and examined by fluorescence microscopy. Approximately 40% of the treated cells showed an obvious nuclear exclusion of Syk-EGFP (Fig. 9A). To determine if signals generated by engagement of the BCR were involved, we compared cells expressing Syk-EGFP to cells expressing a catalytically inactive mutant of Syk-EGFP, Syk(K396R)-EGFP. Syk(K396R)-EGFP can be recruited to the BCR but cannot couple the receptor to downstream signaling (21).

FIG. 9.

PKC mediates the translocation of Syk from the nucleus in B cells treated with anti-IgM antibodies. (A) Syk-deficient DT40 cells transiently expressing Syk-EGFP were left untreated or treated with 50 ng/ml PMA or 20 μg/ml goat anti-chicken IgM antibody for 90 min and then examined by fluorescence microscopy (upper panel). Cells transiently expressing Syk-EGFP were incubated with DMSO solvent alone or with Gö6983 (1 μM) for 20 min and then stimulated with anti-IgM (20 μg/ml) for 90 min or not treated. The percentage of cells in which Syk-EGFP was excluded from the nucleus was determined (lower panel). (B) Syk-deficient DT40 cells transiently expressing Syk(K396R)-EGFP were treated with PMA (50 ng/ml) or anti-IgM (20 μg/ml) for 90 min and examined by fluorescence microscopy (upper panel). The percentage of cells in which Syk(K396R)-EGFP was excluded from the nucleus was determined (lower panel). The reported percentages of cells showing nuclear exclusion in panels A and B represent the mean and standard errors for three separate analyses that examined >100 cells each.

Treatment of cells expressing Syk(K396R)-EGFP with anti-IgM failed to lead to the export of the mutant kinase from the nucleus (Fig. 9B). However, Syk(K396R)-EGFP did move out of the nucleus after cells were treated with PMA, which can bypass the need for receptor-mediated signaling. Preincubation of cells expressing Syk-EGFP with the PKC inhibitor Gö6983 reduced the number of cells exhibiting nucleus-excluded Syk in response to anti-IgM from 40 to 15% (Fig. 9A), consistent with an involvement of PKC in this process.

PMA-induced translocation of Syk requires new protein synthesis.

While PKC can be activated by engagement of the BCR or by treatment with PMA within seconds or minutes, the nuclear exclusion of Syk-EGFP occurred only after cells were treated with anti-IgM or PMA for longer than 30 min. Due to the timing of the response, we speculated that PMA-stimulated protein synthesis might be required. To test this, we treated cells expressing Syk-EGFP with the protein synthesis inhibitor cycloheximide prior to treatment with PMA or anti-IgM antibodies. Treatment with cycloheximide strongly inhibited the nuclear exclusion of Syk-EGFP induced by either PMA or anti-IgM (Fig. 10A). The ability of cycloheximide to inhibit protein synthesis was verified by measuring the incorporation of [³⁵S]methionine and [³⁵S]cysteine into total protein (Fig. 10B).

FIG. 10.

PMA-induced translocation of Syk from the nucleus requires protein synthesis. (A) Syk-deficient DT40 cells expressing Syk-EGFP were incubated with DMSO (solvent control) or cycloheximide (1 μg/ml) for 20 min, followed by treatment with PMA (50 ng/ml) or anti-IgM (20 μg/ml) for 90 min. The reported percentages of cells showing nuclear exclusion represent data from two separate analyses and include an examination of >100 cells. (B) Syk-deficient DT40 cells were treated with DMSO or cycloheximide (1 μg/ml) for 110 min in the presence of 28 μCi of an amino acid mixture containing [³⁵S]methionine and [³⁵S]cysteine. The incorporation of radioactivity into newly synthesized proteins was determined by scintillation spectrometry. (C) Syk-deficient DT40 cells transiently expressing ZL were left untreated or treated with PMA (200 ng/ml) for 90 min. The localization of ZL was examined by fluorescence microscopy.

To determine if PMA functioned by blocking the general translocation mechanism required for linker B-mediated import, we measured the ability of PMA to alter the localization of the heterologous β-galactosidase-EGFP fusion protein ZL. The treatment of cells expressing ZL with PMA did not change its localization (Fig. 10C). Thus, treatment of cells with PMA does not appear to block protein import or stimulate protein export in general. These results suggest instead that the activation of PKC by engagement of the BCR or application of PMA results in the production of a protein or proteins that participate selectively in the process of Syk's nuclear exclusion.

Effect of the nuclear localization of Syk on the activation of caspase 3.

To begin to explore roles for this nucleocytoplasmic shuttling of Syk, we generated pools of stably transfected, Syk-deficient DT40 B cells expressing forms of Syk-EGFP with restricted subcellular localizations. For this, cells were transfected with plasmids coding for Syk-EGFP, which is localized to both the cytoplasm and nucleus; Syk-EGFP-NES, which is cytoplasmic; and Syk-EGFP-NLS, which is nuclear (Fig. 11A). Syk-EGFP-NLS contains a C-terminal nuclear localization sequence derived from simian virus 40 large T antigen. Transfected cells were selected by antibiotic resistance. Stable pools of cells contained similar levels of each fusion protein (Fig. 11B).

FIG. 11.

Syk in the nucleus mitigates the stress-induced activation of caspase 3. (A) The localization of Syk-EGFP (Syk), Syk-EGFP-NES (NES), and Syk-EGFP-NLS (NLS) in stable pools of transfected cells was examined by fluorescence microscopy. (B) The level of expression of Syk-EGFP (Syk), Syk-EGFP-NES (NES), and Syk-EGFP-NLS (NLS) in lysates of pools of transfected cells was compared by Western blotting using an anti-Syk antibody. (C) Caspase 3 activity was measured in lysates prepared from Syk-deficient cells (Syk⁻) or Syk-deficient cells expressing Syk-EGFP (Syk), Syk-EGFP-NES (NES), or Syk-EGFP-NLS (NLS) that were grown at a density of 5 × 10⁵/ml (Exp) or 2 × 10⁶/ml (High). (D) Caspase 3 activity was measured in lysates prepared from Syk-deficient cells (Syk⁻) or Syk-deficient cells expressing Syk-EGFP (Syk), Syk-EGFP-NES (NES), or Syk-EGFP-NLS (NLS) grown at a density of 5 × 10⁵/ml and treated for 8 h with H₂O₂ (100 mM). Data illustrate the averages and standard errors of experiments performed in triplicate. (E) Lysates from Syk-deficient cells expressing Syk-EGFP-NES (NES) or Syk-EGFP-NLS (NLS) treated for the indicated periods of time with H₂O₂ were separated by SDS-PAGE and assayed by Western blotting using antibodies to procaspase 3 and activated caspase 3.

We examined roles for nuclear and cytoplasmic Syk in cellular responses to external stress stimuli using the activation of the effector caspase, caspase 3, as a readout (2). For cells in the exponential phase of growth, caspase 3 activity, as measured using an in vitro enzyme assay, was consistently highest in cells expressing Syk-EGFP-NES and lowest in cells expressing Syk-EGFP-NLS. When cells were stressed by growth at a high cell density, caspase 3 activity was elevated in all the cell lines, but markedly less so in cells expressing Syk-EGFP-NLS (Fig. 11C). The elevated caspase 3 activity that was measured in the in vitro assay could be completely eliminated by treatment of cells with zDEVD-FMK, an irreversible inhibitor of caspase 3 (data not shown), confirming the specificity of the assay.

To induce oxidative stress, we treated cells with 100 μM H₂O₂ for 8 h and then measured levels of active caspase 3 in cell lysates using the in vitro enzyme assay. Caspase 3 activity increased to some extent in each of the cell lines following treatment with H₂O₂. Again, of all the cell lines tested, caspase 3 activity was lowest in those expressing Syk-EGFP-NLS (Fig. 11D). To confirm this difference in the level of active caspase 3 by an alternative approach, we probed fractionated lysates from H₂O₂-treated cells expressing Syk-EGFP-NLS or Syk-EGFP-NES with antibodies against either procaspase 3 or activated caspase 3. Procaspase 3 levels were similar in both cell lines. Activated caspase 3 was elevated in cells expressing Syk-EGFP-NES as a function of time following treatment with H₂O₂, but not in cells expressing Syk-EGFP-NLS (Fig. 11E). Thus, the restriction of Syk to the nucleus substantially reduced the stress-induced activation of caspase 3.

Since prolonged treatment of cells with PMA alters the subcellular localization of Syk, we next examined what would happen if cells were treated with PMA and then oxidative stress was induced by the addition of H₂O₂. To test this, pools of transfected cells expressing the various forms of Syk were treated with PMA or not for 60 min prior to exposure to low levels of H₂O₂ (50 μM) and then examined for differences in their levels of active caspase 3. Interestingly, treatment with PMA greatly potentiated the activation of caspase 3 by H₂O₂ in cells expressing either Syk-EGFP or Syk-EGFP-NES (Fig. 12A). Once again, cells expressing the nucleus-localized fusion protein Syk-EGFP-NLS were refractory to the stress-induced activation of caspase 3, and this was not overcome by the treatment with PMA (Fig. 12A). These results indicate a positive role for cytoplasmic Syk in the activation of caspase 3 in cells exposed to oxidative stress that have been pretreated with PMA. This is in contrast to the negative role played by the nucleus-localized enzyme.

FIG. 12.

Cytoplasmic Syk, PKC, and protein synthesis potentiate the activation of caspase 3 by oxidative stress. (A) Syk-deficient cells (Syk⁻) or Syk-deficient cells expressing Syk-EGFP (Syk), Syk-EGFP-NES (NES), or Syk-EGFP-NLS (NLS) were incubated with PMA (50 ng/ml) for 60 min followed by treatment with H₂O₂ (50 μM) for 8 h (PMA + H₂O₂). Alternatively, cells received treatment with PMA only (PMA), H₂O₂ only (H₂O₂), or neither (control). (B) Cells expressing Syk-EGFP-NES were not treated (PMA + H₂O₂) or treated with either cycloheximide (10 μg/ml) (Cyclo.) or Gö6983 (10 μM) (Gö6983) for 30 min, incubated with PMA (50 ng/ml) for 60 min, and then treated with H₂O₂ (50 μM) for 8 h. Control cells were incubated for the same periods of time but without the activators or inhibitors (control). The data illustrated in panels A and B represent the averages and standard errors of experiments performed in triplicate. (C) Cells expressing Syk-EGFP (Syk) or Syk-EGFP-NES (NES) were treated with H₂O₂ (50 μM) for 8 h (H₂O₂), pretreated with PMA (50 ng/ml) for 60 min followed by H₂O₂ for 8 h (H + P), treated with cycloheximide (10 μg/ml) for 30 min followed by PMA for 60 min and H₂O₂ for 8 h (Cyclo.), or treated with Gö6983 (10 mM) for 30 min followed by PMA for 60 min and H₂O₂ for 8 h (Gö6983). Lysates were fractionated by SDS-PAGE and immunoblotted for procaspase 3 (PC3). The arrow indicates the migration position of one of the cleavage products of procaspase 3.

To determine if the effect of PMA on cellular responses to oxidative stress might be related to its effects on the subcellular localization of Syk, we treated cells with either Gö6983 or cycloheximide to inhibit PKC activity and protein synthesis, respectively. Cells were then treated with PMA prior to the addition of H₂O₂ and assayed for the activation of caspase 3. The ability of PMA to enhance the activation of caspase 3 in cells expressing either Syk-EGFP or Syk-EGFP-NES in response to low levels of oxidative stress was blocked by both inhibitors, as measured using either an enzymatic assay or Western blotting (Fig. 12B and C). Thus, the exclusion of Syk from the nucleus and the enhanced coupling of Syk to the activation of caspase 3 both require the activation of PKC and new protein synthesis, suggesting that these processes may be linked.

DISCUSSION

Through extensive biochemical and genetic studies, Syk has been well characterized as an essential component of the machinery required for signaling through multiple immune recognition receptors. Few studies have focused, however, on its intracellular distribution and trafficking between subcellular compartments and the role this plays in cellular functions. Syk, as well as most other protein tyrosine kinases, is known to function at the plasma membrane, where the receptors to which it is recruited are located. However, in unstimulated cells, the kinase distributes to both nuclear and cytoplasmic compartments when examined as a Syk-EGFP chimera expressed in DT40 cells (21) and as an endogenous protein detected by indirect immunofluorescence or biochemical fractionation. A similar distribution of endogenous Zap-70 in Jurkat T cells was reported previously (34). In the studies on Zap-70, the lysis of cells and washing of nuclei with relatively low concentrations of mild detergents was required for retention of the kinase in the nuclear fraction. We have found that similar procedures are necessary for the preparation of nuclear fractions containing Syk or Syk-EGFP. Thus, Syk and Zap-70 appear to move into and out of the nucleus relatively easily.

Because nuclear pore complexes restrain free passage of proteins larger than approximately 40 kDa, the translocation of Syk into and out of the nucleus must be an active process involving specific nuclear localization and export signals. A serial deletion analysis conducted to identify the region involved in Syk's nuclear import revealed a critical area of the protein near the junction of the linker B and the catalytic domain. The removal of 11 amino acids, EVYLDRSLLTL, from this region blocks nuclear shuttling and restricts the protein to the cytoplasm. This region shows no similarity to the canonical nuclear localization signal that contains a single short stretch of basic amino acids (e.g., PKKKRKV) (15) or the bipartite NLS with two basic amino acid clusters with an intervening spacer (e.g., KRPAATKKAGQAKKKK) (29). Actually, this region of Syk more closely resembles a classical leucine-rich nuclear export signal (10) than it does an NLS. However, treatment of cells with LMB, which inhibits NES-dependent, Crm1-mediated transport, has no effect on the localization of Syk-EGFP. There are examples of leucine-rich motifs that mediate protein nuclear import, including the leucine-rich Vpr fragment that harbors an NLS (31) and a leucine zipper motif from the sterol regulatory element binding protein that is able to engage importin β directly (23). It is not yet known if Syk binds directly or indirectly through an adaptor to importin family members.

A short linker insert region (TWSPGGIISRIKSYSFPKPGHKK) that is missing from SykB, the alternatively spliced variant of Syk, was reported previously to contain a bipartite NLS required for the localization of Syk to the nucleus of breast epithelial cells (43). However, when expressed in DT40 cells, SykB-EGFP distributes normally in both the nucleus and cytoplasm. We have observed a similar localization of SykB-EGFP to both compartments in MCF7 and MCF10A breast epithelial cells as well (U. Shrikhande and R. L. Geahlen, unpublished data). It may be that the shorter SykB protein is more difficult to retain in the nucleus during the biochemical fractionation of epithelial cells. A role in nuclear import for a separate region of linker B is also supported by the observation that Zap-70, which, like SykB, also lacks the linker insert region, is distributed into both the nuclear and cytoplasmic compartments much like Syk is (34).

The region required for the nuclear localization of Syk lies instead at the C-terminal end of the linker B region, extends into the beginning of the catalytic domain, and is just proximal to the ATP-binding loop. We know minimally that L359 is essential for nuclear localization (Fig. 9 and 11). With the exception of S362, this region of Syk is conserved across the human, mouse, and pig enzymes. In addition, L359 and L364 are present at the corresponding positions in Zap-70 (KLFLKRENLLV). Similar leucine-rich regions are found at the beginnings of the catalytic domains of a variety of protein tyrosine kinases and might serve similar functions. For example, the Src family kinase Lck contains the sequence LKLVERL in this region. Interestingly, we found that a chimeric protein consisting of an N-terminally deleted Lck (i.e., lacking the SH4 domain sequences for N-myristoylation and S-palmitoylation) fused to EGFP also localizes to both the nuclear and cytoplasmic compartments (J. Hu and R. L. Geahlen, unpublished data).

The Syk localization signal does not cause the kinase to localize exclusively in the nucleus. In fact, Syk is almost evenly distributed in both the nuclear and cytoplasmic compartments. In addition, when a sequence from the C-terminal region of Syk linker B (residues 332 to 364) containing the Syk localization signal is inserted into the heterologous protein β-galactosidase-EGFP, the fusion protein distributes into both cellular compartments, with only a portion translocating into the nucleus. This suggests the possibility that the transport of Syk into and out of the nucleus might be mediated by a single signal sequence.

Several RNA binding proteins have been reported to harbor nucleocytoplasmic shuttling signals that can control both nuclear import and export (22). This shuttling sequence potentially allows Syk, activated at the cell surface, to transit to the nucleus, where it could phosphorylate nuclear substrates. The appearance of phosphorylated ZAP-70 in the nucleus of activated T cells has been reported (34), and we also have observed phosphorylated Syk appearing in the nuclear fraction of activated B cells (F. Zhou and R. L. Geahlen, unpublished data).

The exact boundaries of the minimal sequence required for nucleocytoplasmic shuttling have not been defined. Clearly, a sequence containing amino acids 332 to 364 is both necessary and sufficient for the localization of either Syk-EGFP or a β-galactosidase-EGFP fusion protein to the nucleus. In the case of Syk deletion mutants that retain the first two-thirds of linker B, residues can be removed from the final third from amino acid 332 up to 358 before any measurable reduction in the extent of nuclear localization is observed, indicating that amino acids including and distal to L359 are required. However, the shuttling sequence does exhibit some context dependency, since the nuclear import of a Syk mutant lacking all linker B sequences up to residue 359, Syk(linkerless +L)-EGFP, is less efficient. Consistent with this, insertion of only either the 11 or 19 amino acids required for the shuttling of Syk (Fig. 5) between β-galactosidase and EGFP was not sufficient to allow nuclear import of the fusion protein (J. Hu and R. L. Geahlen, unpublished data). This may be due to a requirement for upstream sequences for the proper exposure of the shuttling sequence to binding proteins, which in the case of Syk can be provided by portions of the unstructured linker B region. β-Galactosidase-EGFP also may be more difficult to import than Syk due to its large size (β-galactosidase is a 470-kDa tetramer).

The distribution of Syk between the nucleus and cytoplasm is subject to modifications by signals sent from the aggregated BCR that require the activation of PKC. We think that it is unlikely that the direct phosphorylation of Syk by PKC accounts for its exclusion from the nucleus, since extensive phosphopeptide mapping and mass spectrometry experiments in our laboratory have failed to identify any PMA- or BCR-stimulated phosphorylations that occur on any serine whose replacement with alanine affects Syk's localization (J. Hu, H. Oh, and R. L. Geahlen, unpublished data). We have been unable to detect the phosphorylation of Ser-362, which lies near the C-terminal end of linker B, in response to the treatment of cells with PMA.

The timing of the exclusion event and its sensitivity to cycloheximide instead suggest that PKC-dependent changes in protein synthesis are responsible for the exclusion of Syk that occurs 30 to 90 min following receptor engagement. We speculate that one or more proteins that interact with Syk to either promote its export from the nucleus or restrain its localization to the cytoplasm are generated. We have likewise observed a decrease in the level of nuclear Syk in resting spleen B cells treated with PMA as detected by Western blotting of subcellular fractions, indicating that the phenomenon is not limited to DT40 cells (F. Zhou and R. L. Geahlen, unpublished data).

The physiological outcomes of BCR engagement can vary from clonal expansion to programmed cell death, depending on the nature of the antigen, the developmental stage of the B cell, and the presence or absence of additional stimuli (12). Syk is an essential component of the BCR signaling machinery and is required both for cell survival during B-cell maturation (5, 39) and for BCR-stimulated apoptosis (37). Thus, the mechanisms by which Syk participates in cell survival decisions are of considerable interest. Our data suggest that the subcellular localization of Syk can influence how B cells respond to external stimuli. Syk in the nucleus renders cells refractory to the stress-induced activation of caspase 3, while Syk in the cytoplasm, coupled with PMA-induced gene expression, enhances the activation of caspase 3.

The pathways involved in both of these processes are not known but are currently under investigation. The stress-induced activation of caspase 3 in cells expressing Syk-EGFP-NLS are even lower than in Syk-deficient cells. This may reflect the positive regulation of one or more signaling pathways linked to survival or the negative regulation of a proapoptotic pathway. Nuclear Syk has been reported to function as a transcriptional repressor in breast epithelial cells through its interactions with SP1 and the recruitment of histone deacetylase (42). It is possible, therefore, that the presence of Syk in the nucleus of B cells could block the expression of a protein or proteins that promote the coupling of external stress stimuli to the activation of caspase 3. Likewise, the restriction of Syk to the cytoplasm might serve to derepress specific genes that participate in stress signaling.

The enhancement of stress signaling by cytoplasmic Syk, like the nuclear exclusion of Syk, is dependent on the activation of PKC and new protein synthesis, suggesting that these two phenomena might be linked. Thus, the regulated translocation of Syk out of the nucleus may be a way to enhance its participation in a pathway that is localized to the cytoplasm and that links stress stimuli to the cleavage of procaspase 3. Alternatively, the translocation of Syk from the nucleus may be a consequence of its recruitment to a newly synthesized, cytoplasmic protein that participates in the oxidative stress signaling pathway. The exact mechanism awaits further experimentation and identification of the pathway or pathways that couple the kinase to the activation of caspase 3.

In summary, the Syk protein tyrosine kinase can traffic readily in and out of the nucleus, an ability conferred upon it by a unique shuttling sequence located at the extreme C terminus of the linker B region. This allows Syk to transit between subcellular compartments both prior to and shortly following receptor engagement. At longer times following B-cell activation, Syk is then excluded from the nucleus for a period of time by a mechanism involving the activation of PKC and stimulation of gene expression. Changes in the localization of Syk modulate the responses of cells to oxidative stress such that cells with Syk in the nucleus are resistant to the stress-induced activation of caspase 3, while cells with Syk in the cytoplasm are more susceptible.