Identification of Protein Tyrosine Kinases Required for B-Cell- Receptor-Mediated Activation of an Epstein-Barr Virus Immediate-Early Gene Promoter

Abstract

Epstein-Barr Virus (EBV) is a potentially oncogenic herpesvirus that infects >90% of the world's population. EBV exists predominantly as a latent infection in B lymphocytes, with periodic lytic-cycle reactivation essential for cellular and host transmission. Viral reactivation can be stimulated by ligand-induced activation of B-cell-receptor (BCR)-coupled signaling pathways. The critical first step in the transition from latency to the lytic cycle is the expression of the viral immediate-early gene BZLF1 through the transcription activation of its promoter, Zp. However, the BCR-coupled signal transduction cascade(s) leading to the induction of Zp and the expression of the BZLF1 gene product, Zta, is currently unclear. A major obstacle to delineating the relevant signal transduction events has been the lack of a model of EBV infection that is amenable to genetic manipulation. The use of the avian B-cell line DT40 has proven to be a powerful tool for delineating BCR-mediated signal transduction pathways that appear to be highly conserved between avian and mammalian systems. We demonstrate that the DT40 cell line is a robust and genetically tractable system for the study of BCR-mediated signaling pathways leading to transcriptional activation of BZLF1. Using this system, we demonstrate that activation of Zp requires the BCR-coupled protein tyrosine kinases Syk and Btk and that it is positively regulated by Lyn. Thus, the use of DT40 cells has allowed us to delineate the early signaling components required for BCR-dependent reactivation of latent EBV, and this system is likely to prove useful for further dissection of the downstream signaling cascades involved.

Epstein-Barr virus (EBV) is a ubiquitous human herpesvirus that infects more than 90% of the world's population (reviewed in references 17 and 28). EBV establishes a latent infection in human B lymphocytes and is a causative agent of infectious mononucleosis, a self-limiting lymphoproliferative disorder (17, 28). In addition, EBV is potentially oncogenic and has been associated with several human cancers including Burkitt's lymphoma, posttransplant lymphoma, and nasopharyngeal carcinoma, an epithelial malignancy (17, 28).

Infection of human B lymphocytes with EBV in vitro can lead to immortalization. EBV is maintained in immortalized cells as an episome and can establish a latent infection characterized by expression of a limited number of viral genes (17, 28). EBV can be switched to a lytic cycle when latently infected cells are exposed to extracellular stimuli, including anti-immunoglobulin (anti-Ig), phorbol esters, calcium ionophores, and butyrate (17, 25, 28). The critical first step in the transition from latency to the lytic cycle is the expression of the viral immediate-early genes, BZLF1 and BRLF1 (7, 29; reviewed in reference 32).

Transcription of the BZLF1 and BRLF1 genes is initiated from either a proximal promoter, Zp, or a distal promoter, Rp (14, 26, 31). Zp and Rp both exhibit low basal activity and appear to be activated simultaneously by agents that disrupt latency (31). Zp responds to numerous signaling pathways that initiate the lytic cycle by driving expression of the immediate-early protein Zta (also referred to as BRLF1 Zebra and EB1) (2, 4, 9, 12, 13, 16, 23, 30; reviewed in reference 32). The expression of Zta drives amplification of Zp activity and leads to the activation of early genes and ultimately to viral replication (6, 7, 11, 12). Therefore, control of Zta expression is critical to regulating entry into the lytic cycle.

Although it is well established that ligand-induced activation of latently infected B lymphocytes through cross-linking of the B-cell receptor (BCR) can induce the reactivation of the EBV lytic cycle, the BCR-coupled signal transduction cascade(s) leading to the induction of Zp and the expression of Zta is currently unclear (32, 33). A major obstacle to delineating the relevant signal transduction events has been the lack of a model of EBV infection that is amenable to genetic manipulation. In contrast, the early events in BCR-mediated signaling have been delineated in considerable detail. Briefly, BCR cross-linking leads to the rapid activation of cytoplasmic protein tyrosine kinases (PTK) including the src-family PTK Lyn and the tyrosine kinase Syk (18, 34, 35). The activation of PTK leads to the phosphorylation of immune receptor tyrosine-based activation motifs in the cytoplasmic domains of the Igα and Igβ chains of the BCR complex and to the recruitment and activation of downstream adaptor and effector molecules including additional cytoplasmic tyrosine kinases such as the Itk/Tec-family kinase Btk (reviewed in references 18 and 35). There then follows a diverse cascade of signaling events required for the induction of gene transcription, proliferation, differentiation, and antibody secretion, some of which are dependent on both Syk and Lyn while others are differentially regulated by Syk and Lyn (19-21).

The avian B-cell line DT40 has proven particularly useful in delineating BCR-mediated signal transduction pathways for several reasons. First, DT40 cells are subject to a high rate of homologous recombination and therefore can be genetically manipulated with high efficiency. As a result, a large number of genes have been targeted for deletion in this system. Second, DT40 cells exhibit less complexity with regard to their expression of members of the various BCR-proximal PTK than many B-cell lines and primary B cells. Specifically, DT40 cells express only Lyn of the src family of tyrosine kinases, Syk but not Zap-70, and Btk but not other Itk-family members, thus circumventing the confounding contribution of other PTK that can exhibit functions redundant to those of the targeted gene products. Another potentially important feature of DT40 cells with regard to the studies described here is that, like EBV-transformed B cells, DT40 cells constitutively express significant levels of c-myc. Understanding the signal transduction cascade leading to the induction of Zp and the expression of Zta following activation by lytic-cycle-inducing agents could provide targets for regulating EBV reactivation and the development of EBV-associated lymphoproliferative disorders and malignancies. Therefore, we investigated whether DT40 cells could be used to delineate the signal transduction pathways involved in BCR-mediated induction of Zp.

We demonstrate that BCR ligation leads to activation of Zp in DT40 cells transiently transfected with a Zp-driven luciferase reporter construct. Furthermore, the cis-acting elements required for activation of Zp in DT40 cells were similar to those described for human B cells. In addition, the remarkable amplification of Zp activity by Zta described for human cells was recapitulated in the DT40 model. Having established that DT40 cells represent a reasonably faithful model system in which to study the activation of Zp, we used this system to investigate the role of BCR-proximal tyrosine kinases in the activation of Zp. We found that Syk and Btk tyrosine kinases are required, whereas Lyn kinase positively regulates Zp but is not required for the induction of Zp.

MATERIALS AND METHODS

Cell culture.

The avian B-cell line DT40 was cultured in RPMI 1640 medium supplemented with 10% heat-inactivated fetal bovine serum, 1% heat-inactivated chicken serum, 100 μg of penicillin/ml, 100 μg of streptomycin/ml, 2 mM glutamine, and 50 μM β-mercaptoethanol. DT40 Syk⁻, Lyn⁻, and Btk⁻ cells were cultured in the same medium supplemented with 2 mg of G418/ml. Lack of expression of Syk, Lyn, and Btk, respectively, was confirmed and shown by immunoblotting to have no effect on the level of expression of either of the other PTK compared to that in wild-type DT40 cells. DT40 Syk⁻ and DT40 Lyn⁻ reconstituted cells overexpressing either wild-type or catalytically inactive mutant forms (SykK402R and LynK275R for Syk and Lyn, respectively) of human Syk or Lyn, generated as previously described (21), were cultured in the same medium supplemented with 0.4 μg of puromycin/ml or 0.7 mg of hygomycin/ml. Wild-type mouse Btk (provided by Prim Kanchanastit, University of California—Los Angeles) was subcloned into pApuro, transfected into DT40 Btk⁻ cells by electroporation using a Gene Pulser apparatus (Bio-Rad) at 330 V and 250 μF, and selected in the presence of 0.5 μg of puromycin/ml. Puromycin-resistant clones were screened by Southern blotting of genomic DNA and further selected for expression of Btk protein by immunoblotting with a rabbit anti-mouse Btk (PharMingen, San Diego, Calif.). All cells used were monitored for expression of levels of surface Ig comparable to those in the parental DT40 cells as determined by fluorescence-activated flow cytometry. The EBV-positive and EBV-negative human Akata cell lines were cultured in RPMI 1640 medium supplemented with 10% heat-inactivated fetal bovine serum, 100 μg of penicillin/ml, 100 μg of streptomycin/ml, and 2 mM glutamine.

Construction of vectors.

A DNA construct containing the region from −220 to +12 of the BZLF1 promoter (Zp) was engineered by PCR as described elsewhere (5) and was designated N292 (see Fig. 2). The BZLF1 promoter sequences were subcloned into the Nhe1/HindIII sites of the PGL3 vector (Promega, Madison, Wis.). This construct was previously determined to contain all the necessary responsive elements for maintaining low basal activity and activation by lytic-cycle-inducing agents (2, 9, 12, 30). The mutant construct N370 is identical to the mutant construct MII with the 4-bp substitution mutation in the ZII domain which contains 12-O-tetradecanoylphorbol-13-acetate (TPA)- and anti-Ig-responsive elements and has been described previously (9, 23). The mutant construct N371 is identical to the mutant MIIIA with the 4-bp substitution mutation in the ZIII domain, which is known to bind the BZLF1 gene product Zta (9, 12, 22). Finally, two deletions were made from −129 to +12 and −52 to +12; these deletion constructs were designated N375 and N376, respectively. All constructs were sequence verified and purified by Maxiprep (QIAGEN) or cesium chloride banding. Schematic representations of constructs used in this study are shown in Fig. 2. The BZLF1 gene product Zta was subcloned into the BamHI/EcoRI sites of pcDNA3, a cytomegalovirus promoter-driven expression system (Invitrogen, Carlsbad, Calif.) as described elsewhere (5).

FIG. 2.

(a) Schematic representation of the wild-type Zp construct (N292) and mutant constructs used in this study. Also indicated are the reported anti-Ig- and PDBu-responsive elements within Zp. (b) The ZI domain is absolutely required, whereas the ZII domain is only partially required, for anti-IgM induction of Zp in DT40 parental cells. A total of 10⁷ DT40 parental cells were transiently transfected with either the wild-type luciferase-driven Zp construct (N292) or a luciferase-driven mutant Zp construct; they were then stimulated with either 20 μg of anti-IgM/ml or a combination of 50 ng of PDBu/ml and 1 μM ionomycin. The induction of Zp was then determined at 16 h poststimulation. Results are expressed as the fold increase in luciferase activity in stimulated cells over that in unstimulated control cells. (c) Analysis of cis-acting elements required for anti-IgG induction of Zp in EBV-positive Akata cells. Asterisks indicate that the fold induction in stimulated cells is significant (P < 0.05); daggers indicate a significant (P < 0.05) difference in the fold induction of luciferase activity by the wild-type Zp construct versus the mutant construct in DT40 parental cells.

Transfection and luciferase assay.

A total of 10⁷ DT40 or Akata cells (>95% viable) were resuspended in 0.3 ml of culture medium and transiently transfected with 5 μg of the various luciferase reporter recombinant plasmid DNA constructs in the presence or absence of 2 μg of Zta pcDNA3 plasmid DNA by electroporation or nucleofection. The cells were then incubated on ice for 10 min prior to being cultured overnight at 37°C under 5% CO₂. The following day, cells were washed and subcultured in a 24-well plate at a density of 5 × 10⁵ cells per ml of culture medium. The viability of the cells at this point in the procedure was typically >80%. DT40 cells were then activated with 20 μg of a mouse monoclonal anti-chicken IgM (M4)/ml, and Akata cells were stimulated with 10 μg of a goat anti-human IgG (Jackson Laboratory, West Grove, Pa.)/ml. These concentrations were determined to be optimal for the respective cell lines. As a positive control, cells were stimulated with a combination of phorbol 12,13-dibutyrate (PDBu) (50 ng/ml) and ionomycin (1 μM) that has been shown previously to induce the Z promoter of the BZLF1 gene (5). Luciferase activity was determined at 8 to 24 h poststimulation by using a Perkin-Elmer Wallac Victor² luminometer. Zta expression levels were assayed by Western blot analysis with a rabbit polyclonal antibody raised against full-length Zta. For Western blotting, 2 × 10⁶ DT40 cells were washed in phosphate-buffered saline and then lysed in 2× sodium dodecyl sulfate-Laemmli buffer by heating to 100°C for 15 min.

Data analysis.

Results are expressed as fold induction of luciferase activity in stimulated cells relative to the activity observed in unstimulated control cells following normalization for protein content using a bicinchoninic acid assay (Pierce, Rockford, Ill.). Results shown are averages from multiple experiments, each performed in triplicate. Significance was determined by a t test analysis.

RESULTS AND DISCUSSION

Activation of an immediate-early gene promoter, Zp, in DT40 cells.

We studied the induction of Zp, the promoter of the immediate-early EBV gene BZLF1, following cross-linking of the BCR in the avian B-cell line DT40 in order to investigate whether DT40 cells could serve as a model in which to study the early events of EBV reactivation in human B cells. Cross-linking of the BCR on parental DT40 cells with a mouse anti-chicken IgM antibody induced a time-dependent increase in Zp activity (Fig. 1a). Optimal induction was observed at 16 to 24 h (Fig. 1a; also data not shown). The induction of Zp in DT40 cells was consistently comparable to, or even greater in magnitude than, that observed in human Akata cells, although in both cases the response was optimal at 16 h following stimulation with anti-human IgG. Both EBV-negative (Fig. 1a) and EBV-positive (Fig. 1b and 2) Akata cells were used; interestingly, we found that induction of Zp activity in response to anti-Ig was consistently more robust in the EBV-negative cells. The more robust response in EBV-negative cells is consistent with the fact that the integral membrane protein 2 (LMP2) of EBV negatively regulates BCR signaling through the regulation of BCR-coupled PTK (27). Our results indicate that DT40 cells may provide a robust model for studying BCR-mediated signaling events that lead to EBV Zp activation.

FIG. 1.

Activation of the immediate-early gene promoter Zp in DT40 and Akata cells. (a) Kinetics of Zp activation in DT40 cells. A total of 10⁷ DT40 parental cells were transiently transfected with 5 μg of a Zp-driven luciferase construct (N292). After 24 h, the cells were stimulated with either 20 μg of anti-IgM (M4)/ml (open bars) or a combination of 50 ng of PDBu/ml and 1 μM ionomycin (hatched bars) for the times indicated at 37°C, and the induction of luciferase activity was determined. Results are averages from three, six or four experiments for the 8-, 16-, and 24-h time points, respectively; each experiment was performed in triplicate. Optimal Zp activation occurs at 16 h poststimulation. For comparison, the fold induction of luciferase activity following a 16-h stimulation of EBV-negative human Akata cells with either 10 μg of anti-human IgG/ml (open bar) or a combination of 50 ng of PDBu/ml and 1 μM ionomycin (hatched bar) is also shown. Data are expressed as the fold increase in luciferase activity in stimulated cells over that in unstimulated control cells. Asterisks indicate that the fold induction in stimulated cells is significant (P < 0.05) as determined by a paired t test analysis using an Excel statistical package. (b) Zp induction following anti-Ig stimulation is enhanced in the presence of Zta. A total of 10⁷ DT40 parental cells were transiently transfected with 5 μg of a Zp-driven luciferase construct (N292) in the presence or absence of Zta. After 24 h, the cells were stimulated with 20 μg of anti-IgM (M4)/ml (open bars) or a combination of 50 ng of PDBu/ml and 1 μM ionomycin (hatched bars) for 16 h at 37°C, and the induction of luciferase activity was determined. Results are averages from five experiments, each conducted in triplicate. For comparison, the fold induction of luciferase activity following a 16-h stimulation of EBV-negative human Akata cells (n = 4) with either 10 μg of anti-human IgG/ml (open bar) or a combination of 50 ng of PDBu/ml and 1 μM ionomycin (hatched bar) is also shown. Data are expressed as the fold increase in luciferase activity in stimulated cells over that in unstimulated control cells. Asterisks indicate that the fold induction in stimulated cells is significant (P < 0.05). (c) Western blot analysis of Zta expression levels in DT40 cells transfected with a Zta expression plasmid (lanes 4 to 6) and treated with IgM (lanes 2 and 5) or PdBu and ionomycin (lanes 3 and 6).

A similar time-dependent induction of Zp activity was observed when DT40 parental cells were stimulated with a combination of PDBu and ionomycin, which bypasses early receptor-mediated signaling pathways and induces protein kinase C and calcium mobilization directly (Fig. 1a). The induction of Zp activation in Akata cells stimulated with PDBu-ionomycin was comparable. Based on the results of these time course experiments, all further experiments were performed at 16 h following stimulation with either anti-Ig or a combination of PDBu and ionomycin.

Zp induction in DT40 cells is enhanced by Zta.

Induction of Zp drives the transcription of Zta. In human B cells, expression of Zta provides an autostimulatory effect on Zp by binding to the ZIII domain of the Z promoter (12, 22). This autostimulatory effect can occur in the presence of Zta alone, or Zta can act synergistically with various stimuli including anti-Ig. We therefore examined the capacity of Zta to amplify the induction of Zp either alone or following stimulation with either anti-Ig or a combination of PDBu and ionomycin in DT40 cells compared to Akata cells. At 16 h after transfection with Zta, we observed a 19- ± 3-fold activation of Zp in unstimulated DT40 cells and a 4- ± 1-fold activation in Akata cells (Fig. 1b) compared to unstimulated cells in the absence of Zta (Fig. 1b). Stimulation with anti-Ig in the presence of Zta resulted in a 172- ± 58-fold increase in Zp transcription in DT40 cells, an even more robust synergy than that observed in Akata cells. A similar synergistic effect was also observed when DT40 cells were stimulated with a combination of PDBu and ionomycin in the presence of Zta (Fig. 1b). Synergy between PDBu-ionomycin and Zta was also evident, although somewhat less robust in EBV-positive Akata cells than in DT40 cells (Fig. 1b) or EBV-negative cells (data not shown). Zta expression levels in DT40 cells were not changed by anti-Ig but were increased by PDBu-ionomycin treatment (Fig. 1c), indicating that synergy between Zta and anti-Ig signaling arises from induced posttranslational modifications and/or interactions with Zta and cellular proteins. These data further support the usefulness of DT40 cells as a robust model for studying the induction of Zp activity and establish that the autoregulation of Zp by Zta under basal conditions, as well as the synergistic effect of Zta on anti-Ig- or PDBu-plus-ionomycin-induced Zp activity, is recapitulated in DT40 cells compared to human B-cell lines.

The ZI domain of Zp is absolutely required, while the ZII domain is only partially required, for the activation of Zp in DT40 cells.

In order to map the responsive elements required for BCR-mediated Zp induction following anti-IgM stimulation of DT40 cells, we designed constructs in which previously characterized responsive elements (24) were either mutated or deleted (Fig. 2a). We focused primarily on the ZI and ZII sites, which have been reported to contain anti-Ig- and TPA-responsive elements, and the ZIII site, which is the known binding site for Zta (the protein product of the BZLF1 gene) but contains no reported anti-Ig- or TPA responsive elements.

Stimulation of DT40 cells with anti-IgM induced a significant 19- ± 4-fold induction of activity in cells transfected with N292 (wild-type Zp −220 to +12) compared with unstimulated control cells (Fig. 2b). This induction of Zp was significantly inhibited by deletion of the distal ZIA, ZIB, and ZIC sites (N375; Zp −129 to +12) and completely abolished by the deletion of the ZI and ZII responsive sites (N376; Zp −52 to +12) (Fig. 2b). Substitution mutations in the ZII site have been reported to abolish the Zp TPA response in human B lymphocytes (12, 22). A ZII substitution mutant (N370) identical to that previously reported was tested for its response to anti-IgM induction of Zp in DT40 cells. Mutation of the ZII site reduced anti-IgM induction of Zp to 7- ± 2-fold compared with the 19- ± 4-fold observed with wild-type Zp (N292) (Fig. 2b). Substitution mutation of the ZIII domain of Zp (N371), which has been reported to abolish Zta autostimulation, had no significant effect on the induction of Zp following anti-IgM stimulation of DT40 cells (Fig. 2b).

The same Zp mutants were assayed for their response to PDBu-ionomycin stimulation of DT40 cells. PDBu-ionomycin treatment produced a 10- ± 3-fold induction of wild-type Zp (N292), which was reduced to 2- ± 1-fold and 2- ± 0-fold in cells transfected with Zp deletion mutants lacking the distal ZI sites (N375) or the ZI and ZII sites (N376) (Fig. 2b). Substitution mutation of the ZII site alone (N370) led to a 3- ± 2-fold induction of Zp when cells were stimulated with PDBu-ionomycin. In contrast, mutation of the ZIII domain had no significant effect on Zp induction when DT40 cells were stimulated with PDBu-ionomycin (10- ± 3-fold versus 12- ± 2-fold for N292 and N371, respectively). These results indicate that the ZI domain of Zp is required for induction of the BZLF1 gene in DT40 cells by either anti-IgM or PDBu-ionomycin stimulation, whereas the ZII domain is only partially required.

A similar pattern of responsiveness was seen with the various mutants in EBV-positive Akata cells stimulated with anti-Ig or PDBu-ionomycin (Fig. 2c). IgG responsiveness was largely dependent upon the ZI domain in Akata cells and was further inhibited by the combined deletion of ZI and ZII sites. These experiments were carried out with EBV-positive Akata cells, which, as indicated above, likely explains the lower activation of all Zp constructs in these cells relative to the more robust anti-IgM response in DT40 cells. However, qualitatively, the requirement for the cis-acting elements of Zp in DT40 cells appears to recapitulate that reported for human B cells, as confirmed by experiments performed with Akata cells in parallel in this study.

Syk is required for BCR-mediated activation of Zp.

One of the earliest detectable biochemical events that follows cross-linking of BCR is an increase in protein tyrosine phosphorylation through the activation of the tyrosine kinases Lyn and Syk (reviewed in references 18, 34, and 35). As previously demonstrated, a deficiency in Syk results in the loss of BCR-induced tyrosine phosphorylation of a subset of substrates detected in parental DT40 cells (21). In order to determine the role of Syk in EBV reactivation, we compared Zp induction following BCR ligation in DT40 parental and DT40 Syk⁻ cells. Anti-IgM-induced activation of Zp was abolished in DT40 Syk⁻ cells compared to DT40 parental cells (Fig. 3a). Importantly, anti-IgM induction of Zp as well as tyrosine phosphorylation (data not shown) was restored when DT40 Syk⁻ cells were reconstituted with human Syk but not when they were reconstituted with the pApuro vector alone.

FIG. 3.

(a) Syk is required for BCR-mediated induction of Zp. A total of 10⁷ DT40 parental, Syk⁻, Syk⁻.pApuro, Syk⁻.WTSyk, and Syk⁻.KRSyk cells were transiently transfected with a Zp-driven luciferase construct (N292). After 24 h, the cells were stimulated with either 20 μg of anti-IgM/ml (n = 3) or a combination of 50 ng of PDBu/ml and 1 μM ionomycin (n = 4) for 16 h at 37°C, and the induction of luciferase activity was determined. Results are means of triplicate determinations and are expressed as the fold increase in luciferase activity in stimulated cells over that in unstimulated control cells. Asterisks indicate that the fold induction in stimulated cells is significant (P < 0.05). Daggers indicate a significant (P < 0.05) difference in the fold induction in luciferase activity between DT40 parental cells and mutant DT40 cells. (b) Syk is required for the enhancing effect of Zta on Zp following anti-IgM or PDBu-ionomycin stimulation. A total of 10⁷ DT40 parental and DT40 Syk⁻ cells stably transfected with the pApuro vector alone (Syk⁻.pApuro) or with wild-type human Syk (Syk⁻.WtSyk) were transiently transfected with 5 μg of a Zp-driven luciferase construct (N292) in the presence of Zta. After 24 h, the cells were stimulated with either 20 μg of anti-IgM (M4)/ml (open bars) or a combination of 50 ng of PDBu/ml and 1 μM ionomycin (hatched bars) for 16 h at 37°C, and the induction of luciferase activity was determined. Results are averages from two experiments, each performed in triplicate. Asterisks indicate that the fold induction in stimulated cells is significant (P < 0.05). Daggers indicate a significant (P < 0.05) difference in the fold induction of luciferase activity between DT40 parental cells and Syk⁻.pApuro cells. Double daggers indicate a significant (P < 0.05) difference in the fold induction of luciferase activity in the presence versus the absence of Zta.

In addition to its role as a tyrosine kinase, Syk can also act as an adaptor molecule for other signaling molecules. In order to determine whether the kinase activity of Syk is essential for the induction of Zp following anti-IgM stimulation, we utilized DT40 Syk⁻ cells transfected with a kinase-dead human Syk (K402R) (21). Expression of kinase-dead Syk did not restore BCR-mediated Zp activation (Fig. 3a) or tyrosine phosphorylation (data not shown). In contrast, stimulation with a combination of PDBu and ionomycin, which bypasses the requirement for Syk, resulted in similar induction of Zp at 16 h poststimulation in parental, Syk⁻, and Syk-transfected Syk⁻ cells (Fig. 3a).

Following anti-Ig induction of Zp, the Zta protein binds to the promoter, inducing further enhancement of Zp induction. To date, the requirement for Syk in the Zta-dependent activation of Zp has not been investigated. We therefore utilized DT40 parental and Syk⁻ cells to determine the role of Syk in Zta enhancement of Zp either alone or following either anti-Ig or mitogen stimulation. Zta activated Zp transcription levels 18- ± 4-fold in unstimulated parental cells but only 3- ± 1-fold in unstimulated Syk⁻ cells (Fig. 3b). Following a 16-h stimulation with anti-IgM Zp, induction was enhanced to 188- ± 72-fold in the presence of Zta (Fig. 3b), compared with an 8- ± 1-fold induction observed in the absence of Zta (Fig. 3a). In the absence of Syk, the induction of Zp by Zta following anti-IgM stimulation was markedly reduced, to 4- ± 1-fold, which was not significantly different from the induction observed in the absence of Zta (Fig. 3b). In DT40 cells stimulated with a combination of PDBu and ionomycin, the induction of Zp in the presence of Zta was significantly reduced from 349- ± 118-fold in parental cells to 52- ± 34-fold in Syk⁻ cells. This induction of Zp by PDBu-ionomycin in the presence of Zta was still significantly greater than that observed in the absence of Zta (compare Fig. 3a with Fig. 3b). These results indicate that Syk plays a role in the enhancement of Zp by Zta in unstimulated and anti-Ig-stimulated DT40 cells. Our findings also indicate that Syk contributes to the PDBu-ionomycin-induced enhancement of Zp by Zta. These results were not necessarily expected and suggest that Syk may be involved in additional downstream events, such as those leading to the posttranslational modification of Zta that contribute to its autoregulatory activity, including possibly the activation of protein kinase C, which would also account for the reduced response to Zta in PDBu-ionomycin-stimulated Syk⁻ cells. The enhancement of Zp by Zta was partially restored in Syk⁻ cells reconstituted with human Syk (Fig. 3b). However, DT40 Syk⁻ cells overexpressing human Syk and transfected with Zp and Zta were considerably more fragile, and therefore we were not surprised that optimal responsiveness was not restored under these circumstances.

Btk is required for BCR-mediated activation of Zp.

The activation of Syk leads to multiple downstream signaling pathways, including the activation of Btk, which is a member of the Itk family of kinases (reviewed in reference 35). Therefore, we also examined the role of Btk in Zp induction in DT40 cells. Anti-IgM stimulation of DT40 parental cells induced a ∼10-fold induction of Zp that was significantly reduced in DT40 Btk⁻ cells (Fig. 4a). Reconstitution of Btk⁻ cells with mouse Btk but not with the pApuro vector alone restored the anti-IgM induction of Zp back to the level observed in parental DT40 cells (Fig. 4a). Stimulation with a combination of PDBu and ionomycin induced Zp to comparable levels independently of Btk expression (Fig. 4a). Btk⁻ cells were also defective with regard to anti-IgM stimulation of Zp in the presence of exogenous Zta, although PDBu-ionomycin stimulation of Zp was not affected by Btk deletion in the presence of Zta. These results suggest that Btk kinase, like Syk, is required for the induction of Zp and plays a role in regulating the autoregulatory activity of Zta transcription activation following anti-IgM stimulation of DT40 cells.

FIG. 4.

Btk is required for BCR-mediated activation of Zp. (a) A total of 10⁷ DT40 parental, Btk⁻, Btk⁻.pApuro, and Btk⁻.WTBtk cells were transiently transfected with a Zp-driven luciferase construct (N292). After 24 h, the cells were stimulated with either 20 μg of anti-IgM/ml (open bars) (n = 3) or a combination of 50 ng of PDBu/ml and 1 μM ionomycin (hatched bars) (n = 3) for 16 h at 37°C, and the induction of luciferase activity was determined. Results are means of triplicate determinations and are expressed as fold increase in luciferase activity in stimulated cells over that in unstimulated control cells. (b) Btk is required for the enhancing effect of Zta on Zp following anti-IgM but not PDBu-ionomycin stimulation. A total of 10⁷ DT40 parental or Btk⁻ cells were transiently transfected with 5 μg of a Zp-driven luciferase construct (N292) in the presence of Zta. After 24 h, the cells were stimulated with either 20 μg of anti-IgM (M4)/ml (open bars) or a combination of 50 ng of PDBu/ml and 1 μM ionomycin (hatched bars) for 16 h at 37°C, and the induction of luciferase activity was determined. Results are averages from four experiments, each performed in triplicate. Asterisks indicate that the fold induction in stimulated cells is significant (P < 0.05); daggers indicate a significant (P < 0.05) difference in the fold induction in luciferase activity between DT40 parental cells and mutant DT40 cells.

Lyn regulates the BCR-mediated activation of Zp.

As previously reported, there is a loss of tyrosine phosphorylation of a subset of substrates in Lyn-deficient DT40 cells compared to that in parental cells (21). Interestingly, the subset of tyrosine-phosphorylated species that is dependent on Lyn overlaps, but is distinct from, the subset of Syk-dependent substrates (21). The differential regulation of downstream signaling pathways by Syk and Lyn was also demonstrated functionally by comparing the activation of mitogen-activated protein kinase, p70S6 kinase, p90Rsk, and Akt in parental, Syk⁻, and Lyn⁻ DT40 cells (19-21). Akt proved particularly informative in this regard, since it was demonstrated that Syk and Lyn had opposing effects on the activation of Akt, which was Syk dependent but inhibited by Lyn (20). In agreement with these findings, we observed a less stringent though significant role for Lyn in the positive regulation of BCR-mediated activation of Zp. When DT40 parental cells were stimulated with anti-IgM, a 24- ± 3-fold induction of Zp was observed, which was reduced to 3- ± 0.5-fold in DT40 Lyn⁻ cells (Fig. 5a). Reconstitution of Lyn⁻ cells with human Lyn but not with the pApuro vector alone restored the anti-IgM induction of Zp to the level observed in parental DT40 cells, whereas Lyn⁻ DT40 cells reconstituted with human kinase-dead Lyn exhibited only a 5- ± 2-fold induction of Zp activity, confirming that the kinase activity of Lyn is essential to the activation of Zp (Fig. 5a). Stimulation with a combination of PDBu and ionomycin induced Zp to comparable levels in all cells regardless of Lyn expression (Fig. 5a).

FIG. 5.

(a) Lyn is required for optimal BCR-mediated induction of Zp. A total of 10⁷ DT40 parental, Lyn⁻, Lyn⁻.pApuro, Lyn⁻.WTLyn, and Lyn⁻.KRLyn cells were transiently transfected with a Zp-driven luciferase construct (N292). After 24 h, the cells were stimulated with either 20 μg of anti-IgM/ml or a combination of 50 ng of PDBu/ml and 1 μM ionomycin for 16 h at 37°C, and the induction of luciferase activity was determined. Results are averages from four experiments, each performed in triplicate, and are expressed as the fold increase in luciferase activity in stimulated cells over that in unstimulated control cells. Asterisks indicate that the fold induction in stimulated cells is significant (P < 0.05); daggers indicate a significant (P < 0.05) difference in the fold induction of luciferase activity between DT40 parental cells and mutant DT40 cells. (b) Lyn is required for the enhancing effect of Zta on Zp following anti-IgM stimulation. A total of 10⁷ DT40 parental or Lyn⁻ cells, or DT40 Lyn⁻ cells stably transformed with the pApuro vector alone (Lyn⁻.pApuro) or wild-type human Lyn (Lyn⁻.WTLyn), were transiently transfected with 5 μg of a Zp-driven luciferase construct (N292) in the presence of Zta. After 24 h, the cells were stimulated in triplicate with either 20 μg of anti-IgM (M4)/ml (open bars) (n = 4) or a combination of 50 ng of PDBu/ml and 1 μM ionomycin (hatched bars) (n = 5) for 16 h at 37°C, and the induction of luciferase activity was determined.

Lyn kinase was also found to positively regulate the enhancement of Zp in the presence of Zta. A 16-h incubation of DT40 parental cells transfected with both Zp and Zta induced a 12- ± 3-fold activation of Zp, which was significantly reduced to 3- ± 1-fold in the absence of Lyn kinase (Fig. 5b). Upon stimulation with anti-IgM, this induction of Zp in the presence of Zta was further enhanced in DT40 parental cells, to 139- ± 17-fold (Fig. 5b). However, this enhancement was significantly reduced, to 37- ± 26-fold, in the absence of Lyn (Fig. 5b). A similar effect was also observed when DT40 cells were stimulated with a PDBu-ionomycin combination. In parental cells, Zp induction by PDBu-ionomycin was enhanced to 319- ± 67-fold in the presence of Zta, and this induction was significantly inhibited in the absence of Lyn (Fig. 5b). Overexpression of wild-type Lyn in DT40 Lyn⁻ cells restored the synergy between Zta and anti-Ig or PDBu-ionomycin (Fig. 5b). Thus, Lyn is required for optimal enhancement of Zp activity by Zta in unstimulated as well as anti-Ig-stimulated cells.

The ZI and ZIII domains, but not the ZII domain, of Zp are required for the induction of Zp activity in DT40 Lyn⁻ cells.

The results presented above indicate that Lyn is only partially required for the induction of Zp following either anti-IgM or PDBu-ionomycin stimulation. Furthermore, it is well documented that Lyn can act as both a negative and a positive regulator of cell signaling processes (3, 15, 20). In addition, it has been suggested previously that there are both positive and negative regulatory domains within Zp itself. Therefore, in order to determine whether Lyn is acting as a positive or a negative regulator of Zp, or as both, and to map the sites in Zp that mediate the Lyn-independent activity of Zp, we compared the activities of the various mutant Zp constructs in parental and Lyn⁻ DT40 cells.

Stimulation of DT40 Lyn⁻ cells with anti-IgM induced a 4- ± 1-fold increase in Zp (N292) activity, which is significantly reduced, but not abolished, compared with the 24- ± 3-fold induction observed in DT40 parental cells (Fig. 6a). When a 6-bp mutation was introduced into the ZII domain, anti-IgM induction of Zp in Lyn⁻ cells was not further reduced compared with that of the wild-type promoter N292 (4- ± 1-fold compared with 2- ± 1-fold for N292 and N370, respectively) (Fig. 6a). However, as with the parental DT40 cells, deletion of the ZI domain either alone (N375) or together with the ZII domain (N376) totally abolished the Zp response in anti-IgM-stimulated Lyn⁻ cells (Fig. 6a). Interestingly, in contrast to the result obtained for DT40 parental cells, when the ZIII domain of Zp was mutated (N371), no induction of Zp was observed in Lyn⁻ cells (Fig. 6a).

FIG. 6.

(a) Role of the cis-acting ZI, ZII, and ZIII domains in BCR-mediated induction of Zp in DT40 Lyn⁻ cells. A total of 10⁷ DT40 Lyn⁻ cells were transiently transfected with the indicated Zp-driven luciferase constructs and then stimulated with 20 μg of anti-IgM/ml. The induction of Zp was then determined at 16 h poststimulation. (b) The ZI and ZII domains of Zp are required for the partial induction of Zp after PDBu-ionomycin stimulation of DT40 Lyn⁻ cells. A total of 10⁷ DT40 Lyn⁻ cells were transiently transfected with the indicated Zp-driven luciferase constructs and then stimulated with a combination of 100 ng of PDBu/ml and 1 μM ionomycin. The induction of Zp was then determined at 16 h poststimulation. Results are means of triplicate determinations and are expressed as fold increase in luciferase activity in stimulated cells over that in unstimulated control cells. Asterisks indicate that the fold induction in stimulated cells is significant (P < 0.05); daggers indicate a significant (P < 0.05) difference in the fold induction of luciferase activity with the wild-type Zp construct versus the mutant construct in DT40 Lyn⁻ cells.

When DT40 Lyn⁻ cells were stimulated with a combination of PDBu and ionomycin, an 8- ± 4-fold induction of Zp (N292) was observed, which was not significantly different from the 10- ± 2-fold induction observed in DT40 parental cells (Fig. 6b). However, in the absence of the ZI domain either alone (N375) or together with the ZII domain (N376), this induction of Zp activity was reduced from that for wild-type Zp (N292), although a significant induction was still observed relative to unstimulated control cells (Fig. 6b). In contrast to anti-IgM stimulation, when a 6-bp mutation was introduced into the ZII domain, PDBu-ionomycin induction of Zp in Lyn⁻ cells was further reduced from that for the wild-type promoter N292 (8- ± 4-fold versus 3- ± 2-fold for N292 and N370, respectively) (Fig. 6b). A mutation of the ZIII domain in Zp had no effect on PDBu-ionomycin induction of Zp, which contrasts with the effect we observed with anti-IgM stimulation of Lyn⁻ cells (Fig. 6b). These results suggest that in the absence of Lyn kinase, anti-IgM activation of Zp is dependent on an intact ZIII binding site in the absence of exogenous Zta. In contrast, mutation of the ZIII binding site had little or no effect on anti-IgM stimulation of Zp in wild-type DT40 cells (Fig. 2), indicating that alternative pathways for Zp activation exist and can be revealed when one arm of the BCR signaling pathway is inactivated.

Conclusions.

Our results suggest that the DT40 avian cell line is a robust and genetically tractable system for the study of signaling pathways initiating at the BCR and culminating in the transcription activation of the BZLF1 immediate-early promoter Zp. We have shown that DT40 cells lacking Syk and Btk tyrosine kinases are incapable of activating Zp transcription in response to anti-IgM stimulation of BCR. Interestingly, we also found that Syk and Btk are important for the potentiation of Zta autoactivation of Zp. Possibly, these kinases are important for the posttranslational modification of Zta or some essential cofactor required for optimal autostimulation of Zp by Zta. Several previous reports have suggested that Zta is phosphorylated in a protein kinase C-dependent manner and that this phosphorylation is essential for transcription activation (1). However, the precise signaling pathway and target of phosphorylation remains controversial (10). Earlier studies demonstrated a role for tyrosine kinase activity in EBV reactivation by using pharmacological inhibitors, such as genistein, but no specific tyrosine kinase was identified as the target of these inhibitors (8). The studies presented here provide clear genetic data showing that Syk and Btk function as essential components of the BCR signaling pathway that activates Zp, and that Lyn may further modulate this pathway. Based on other studies, the BCR signaling system appears to be highly conserved between avian and mammalian systems (reviewed in reference 18), and the results found here are likely to apply to EBV reactivation in human B cells. Furthermore, the ability to reconstitute some aspects of Zp activation with transiently transfected reporter genes allows additional dissection of promoter regions and the identification of nuclear factors associated with these sites. Although chromosomal templates may better reflect the natural latent viral genome, evidence from numerous studies suggests that histone acetylases and ATP-dependent chromatin remodeling proteins functionally activate transcription from transfected plasmid DNA (5). Taken together, these findings indicate that the DT40 cell line is a valuable tool for the genetic and biochemical analysis of BCR-dependent reactivation of latent EBV.