Epstein-Barr virus immediate-early proteins BZLF1 and BRLF1 activate the ATF2 transcription factor by increasing the levels of phosphorylated p38 and c-Jun N-terminal kinases

Abstract

Expression of either Epstein-Barr virus (EBV) immediate-early protein BZLF1 (Z) or BRLF1 (R) is sufficient to convert EBV infection from the latent to lytic form. Disruption of viral latency requires transcriptional activation of the Z and R promoters. The Z and R proteins are transcriptional activators, and each immediate-early protein activates expression of the other immediate-early protein. Z activates the R promoter through a direct binding mechanism. However, R does not bind directly to the Z promoter. In this study, we demonstrate that the ZII element (a cyclic AMP response element site) in the Z promoter is required for efficient activation by R. The ZII element has been shown to be important for induction of lytic EBV infection by tetradecanoyl phorbol acetate and surface immunoglobulin cross-linking and is activated by Z through an indirect mechanism. We demonstrate that both R and Z activate the cellular stress mitogen-activated protein (MAP) kinases, p38 and JNK, resulting in phosphorylation (and activation) of the cellular transcription factor ATF2. Furthermore, we show that the ability of R to induce lytic EBV infection in latently infected cells is significantly reduced by inhibition of either the p38 kinase or JNK pathways. In contrast, inhibition of stress MAP kinase pathways does not impair the ability of Z expression vectors to disrupt viral latency, presumably because expression of Z under the control of a strong heterologous promoter bypasses the need to activate Z transcription. Thus, both R and Z can activate the Z promoter indirectly by inducing ATF2 phosphorylation, and this activity appears to be important for R-induced disruption of viral latency.

FIG. 1

R activates the Z promoter through the ZII element. DG75 or Raji cells were transfected with 5 μg of promoter construct (−221 Zp-CAT or −221 ZpZIIm-CAT) and 1 μg of vector control DNA or R expression plasmid. CAT assays were performed. The fold activation, relative to vector alone, is shown. Error bars indicate range.

FIG. 2

ATF1, ATF2, CREB, and c-Jun bind to the ZII element of the Z promoter. EMSAs were performed with a ³²P-labeled ZII probe and B95-8 protein extracts. (Left panel) Twenty micrograms of protein extract was incubated in the absence (lane 2) or presence of a CRE consensus DNA competitor (lane 3) or an Sp1 consensus DNA competitor (lane 4); in the presence of antibodies directed against ATF1 (lane 5), ATF2 (lane 6), c-Jun (lane 7), or ATF2 plus c-Jun (lane 8); or a control antibody (lane 9). (Right panel) Twenty micrograms of protein extract was incubated in the absence (lane 1) or presence of antibodies directed against CREB (lane 2) or a control antibody (lane 3). A and B indicate the two pertinent complexes. The antibodies alone did not induce supershifts in the absence of cellular extract (data not shown).

FIG. 3

CREB, ATF1, ATF2, and c-Jun activate the Z promoter. (A) DG75 cells were transfected with 5 μg of ZpBS-CAT and 1 μg of transactivator plasmid (vector alone, CREB, ATF1, ATF2, or c-Jun), along with 1 μg of the appropriate kinase expression vectors to activate each transactivator (PKA for CREB and ATF1; p38, JNK, and cdc42 for ATF2; and JNK and cdc42 for c-Jun). (B) DG75 cells were transfected with 5 μg of −221 ZpZIIm-CAT and 1 μg of transactivator plasmid (vector alone, CREB, ATF1, ATF2, or c-Jun), along with 1 μg of the appropriate kinase expression vectors to activate each transactivator (PKA for CREB and ATF1; p38, JNK, and cdc42 for ATF2; and JNK and cdc42 for c-Jun). (C) DG75 cells were transfected with 5 μg of ZpBS-CAT, along with either 1 μg of transactivator plasmid (vector alone, ATF2 or c-Jun), 1 μg of kinase (p38 or JNK and cdc42), or a combination of these plasmids as indicated. CAT assays were performed as described. The average fold activation (and range), relative to vector alone, is presented.

FIG. 4

Z and R increase phosphorylation of ATF2. HeLa cells were infected with adenovirus (Ad)-LacZ, adenovirus-Z, or adenovirus-R as indicated. Immunoblot analysis was performed with antibodies directed against total ATF2 and phosphorylated ATF2 (A) or total CREB and phosphorylated CREB/ATF1 (B). Both phosphorylated CREB and phosphorylated ATF1 are recognized by a single antibody. At least two separate experiments are presented for each antibody. Infection of cells with adenovirus-LacZ did not affect ATF2, ATF1, or CREB phosphorylation in comparison to levels in mock-infected cells (data not shown).

FIG. 5

Z activates ATF2 and c-Jun transactivator functions. (A) DG75 cells were transfected with 5 μg of GAL4-E1B-CAT reporter construct and 1 μg of SG424 (GAL4 DNA-binding domain alone), GAL4-ATF1, GAL4-ATF2, GAL4-CREB, or GAL4-c-Jun, with or without the Z expression vector. CAT assays were performed as described in the text. The average fold activation (and range) relative to SG424 alone is presented. (B) Construction of Z deletion mutants. The location of the Z transactivator (TA) and DNA binding (DNA) and dimerization (DIM) domains is indicated. The Z311 construct is mutated at amino acid (a.a.) 185 (alanine to lysine) in the Z DNA binding domain and is defective in DNA binding. (C) DG75 cells were transfected with 5 μg of GAL4-E1B-CAT reporter construct, 1 μg of GAL4-ATF2, and 1 μg of either control vector, Z, Z311, Z-CT, or ZΔLZ expression vectors. The ATF2 activation induced by each Z construct (relative to the activation induced by wild-type Z, set at 100%) is shown.

FIG. 6

Z and R increase the levels of activated p38 kinase and JNK. (A) HeLa cells were infected with adenovirus (Ad)-LacZ, adenovirus-Z, or adenovirus-R as indicated. Immunoblot analysis was performed with antibodies directed against either total p38 or phosphorylated p38. The results from two separate experiments are presented. (B) Kinase assays were performed as described in the text with extracts from mock-, adenovirus-LacZ-, adenovirus-Z-, or adenovirus-R-infected HeLa cells. The negative and positive controls are extracts from HEL cells that were either UV irradiated (positive) or unirradiated (negative). Similar results were obtained in a second experiment (data not shown). Infection of cells with adenovirus-LacZ did not affect p38 kinase or JNK phosphorylation in comparison to levels in mock-infected cells (data not shown).

FIG. 7

p38 kinase and JNK are necessary for R to disrupt EBV viral latency. (A) DG75 cells were transfected with 5 μg of GAL4-E1B-CAT reporter construct, and 1 μg of SG424 or GAL4-ATF2, along with 1 μg of p38 or JNK expression plasmids, with or without the p38 inhibitor SB 202190 (10 μM) or 2 μg of JIP-1 expression plasmid. CAT assays were performed as described in the text. The ATF2 activation induced (relative to the activation induced by JNK alone or p38 kinase alone, set at 100%) is shown. (B) NPC-KT cells were infected with adenovirus-R or adenovirus-Z in the absence or presence of the p38 kinase inhibitor SB 202190 (10 μM). Immunoblot analysis was performed with antibodies directed against BMRF1, Z or R. Longer exposures indicated that the p38 inhibitor did not affect the level of R protein induced by adenovirus-Z. (C) D98/HE-R-1 cells were transfected with vector alone, R, Z, R plus JIP-1, or Z plus JIP-1 as indicated. Immunoblot analysis was performed with antibodies directed against BMRF1, R, or Z. (D) NIH 3T3 cells were transfected with p38 kinase or a combination of p38 kinase and the dominant-negative MKK3 or MKK6 expression plasmids (dnMKK3 or dnMKK6). Immunoblot analysis was performed with antibodies directed against phosphorylated p38 or total p38. (E) D98/HE-R-1 cells were transfected with vector alone, R, Z, R plus dnMKK3/dnMKK6, or Z plus dnMKK3/dnMKK6 as indicated. Immunoblot analysis was performed with antibodies directed against BMRF1.

FIG. 8

p38 kinase is important for the disruption of viral latency via anti-IgG cross-linking. Akata cells were treated with 100 μg of anti-IgG per ml for 24 h in the absence or presence of the p38 kinase inhibitor (inh.) SB 202190 (10 μM), and the percentage of cells positive for BZLF1 and BMRF1 expression was determined by BZLF1 and BMRF1 antibody staining and FACS analysis. Results are normalized such that the number of BZLF1- or BMRF1-positive cells for each experiment in the absence of drug is set at 100%. The average (and range) from two separate experiments is shown.