Tumor Suppressor p53 Stimulates the Expression of Epstein-Barr Virus Latent Membrane Protein 1

ABSTRACT

Epstein-Barr virus (EBV) is associated with multiple human malignancies. EBV latent membrane protein 1 (LMP1) is required for the efficient transformation of primary B lymphocytes in vitro and possibly in vivo. The tumor suppressor p53 plays a seminal role in cancer development. In some EBV-associated cancers, p53 tends to be wild type and overly expressed; however, the effects of p53 on LMP1 expression is not clear. We find LMP1 expression to be associated with p53 expression in EBV-transformed cells under physiological and DNA damaging conditions. DNA damage stimulates LMP1 expression, and p53 is required for the stimulation. Ectopic p53 stimulates endogenous LMP1 expression. Moreover, endogenous LMP1 blocks DNA damage-mediated apoptosis. Regarding the mechanism of p53-mediated LMP1 expression, we find that interferon regulatory factor 5 (IRF5), a direct target of p53, is associated with both p53 and LMP1. IRF5 binds to and activates a LMP1 promoter reporter construct. Ectopic IRF5 increases the expression of LMP1, while knockdown of IRF5 leads to reduction of LMP1. Furthermore, LMP1 blocks IRF5-mediated apoptosis in EBV-infected cells. All of the data suggest that cellular p53 stimulates viral LMP1 expression, and IRF5 may be one of the factors for p53-mediated LMP1 stimulation. LMP1 may subsequently block DNA damage- and IRF5-mediated apoptosis for the benefits of EBV. The mutual regulation between p53 and LMP1 may play an important role in EBV infection and latency and its related cancers.

IMPORTANCE The tumor suppressor p53 is a critical cellular protein in response to various stresses and dictates cells for various responses, including apoptosis. This work suggests that an Epstein-Bar virus (EBV) principal viral oncogene is activated by cellular p53. The viral oncogene blocks p53-mediated adverse effects during viral infection and transformation. Therefore, the induction of the viral oncogene by p53 provides a means for the virus to cope with infection and DNA damage-mediated cellular stresses. This seems to be the first report that p53 activates a viral oncogene; therefore, the discovery would be interesting to a broad readership from the fields of oncology to virology.

INTRODUCTION

Epstein-Barr virus (EBV) is a human gammaherpesvirus and is associated with nasopharyngeal carcinoma (NPC), Hodgkin's lymphoma (HL), Burkitt's lymphoma (BL), posttransplantation lymphoproliferative diseases (PTLD), and other malignancies (1–3). An estimated 200,000 new cases of EBV-associated cancers occur worldwide per year (4). EBV transforms adult primary B lymphocytes into continuously growing lymphoblastoid cell lines (LCLs) and concomitantly establishes latency in vitro (1). Six nuclear proteins (EBNA1, -2, -3A, -3B, -3C, and -LP), three integral latent membrane proteins (LMP1, -2A, and -2B), and viral RNAs are expressed in the latency type (3, 5).

EBV LMP1 is an integral membrane protein that acts as a constitutively active, receptor-like molecule (6). LMP1 is required for the efficient transformation of primary B cells in vitro (7–9) and transforms primary rodent fibroblasts (10–12). In addition, LMP1 appears to be a central effector of altered cell growth, survival, adhesive, invasive, and antiviral potential in EBV-transformed cells (13–19).

The cellular TP53 tumor suppressor gene encodes a transcription factor, p53, that plays a seminal role in the response of mammalian cells to physiological and environmental stress (20–23). The p53 protein has been implicated as a major mediator of cell cycle arrest and apoptosis through several mechanisms. p53 activation leads to upregulation of proapoptotic and downregulation of prosurvival genes (24, 25). The choice between growth arrest and apoptosis involves a complex interplay of numerous factors (22, 26).

Interferon regulatory factor 5 (IRF5) is a member of the IRF family and a direct target of p53 (27–31). Overexpression of IRF5 inhibits cell proliferation by inducing proapoptotic and suppressing antiapoptotic gene expression (27, 32). IRF5 is selectively involved in apoptosis but not in cell cycle arrest in vivo (33, 34). IRF5 is critically involved in B cell differentiation pathways (35) and is a factor associated with the pathogenesis of lupus, an autoimmune disease (36). Interestingly, IRF5 has been demonstrated to be a tumor-promoting factor in classical Hodgkin's lymphoma (cHL) and thyroid cancers (37, 38).

EBV infection induces p53 expression without causing mutations to TP53 (39). EBV-transformed cells are sensitive to p53-mediated apoptosis, and overexpression of p53 induces apoptosis (39). EBV infection of primary cells activates a DNA damage response (DDR) signaling pathway and inhibits cellular proliferation (40). At the same time, EBV has multiple means to counteract deleterious p53 effects. Most notably, LMP1 blocks p53-mediated apoptosis (13, 41), and EBNA3C downregulates the expression of p53 and regulates the DDR responses initiated by primary infection (40, 42, 43).

Although there has been an intensive effort to study the regulation of p53 by EBV, whether p53 regulates LMP1 expression is not clear. In this study, we found that p53 stimulates the expression of LMP1 and that IRF5 may be a mediator for p53-mediated stimulation. In addition, LMP1 blocks DNA damage or IRF5-induced apoptosis in virally infected cells. This study may suggest a novel mechanism used by EBV to respond to DNA-damaging signals to promote survival of virally infected cells.

RESULTS

LMP1 is associated with p53 protein expression.

The relation between constitutive p53 and LMP1 expression under physiological conditions was examined. TK6 (TP53^+/+), Ne72 (TP53^+/−), and NH32 (TP53^−/−) are three EBV-transformed cell lines derived from a single parental line (Wil2) (44). These cell lines have very similar and comparable genetic backgrounds with notable differences in TP53 status (45–47) (Fig. 1A). With the expression of p53, both TK6 and Ne72 cells had higher constitutive LMP1 expression than that in NH32 cells, a p53-null line (Fig. 1B). EBNA2, a viral activator of LMP1 expression (5), was detected at the similar expression levels in all three cell lines, which suggested that EBNA2 might not be the major factor responsible for the increase in LMP1 in TK6 and Ne72 cells. Additional experiments support the idea that the differences in LMP1 expression might be at the RNA level (Fig. 1C and D) but not in the stability of the LMP1 protein (Fig. 1E). The half-life of IRF1 was short, as expected (Fig. 1E) (48), which suggested that cellular protein degradation pathway was functional. Overall, the data suggested that higher physiological levels of LMP1 expression were associated with p53 expression.

FIG 1.

Association of p53 expression and LMP1. (A) Relationships among Wil2-derived cell lines used in this study. EBV-transformed cells (Wil2) were used to isolate WTK1 and TK6 cells. TK6 cells were further manipulated genetically into Ne72 and NH32 cells. The status of the TP53 gene is as shown. (B) LMP1 expression is correlated with p53 expression. Lysates from Ne72, NH32, and TK6 cells were used for Western blot analysis with the indicated antibodies. Specific proteins are identified. The membranes were stripped and subsequently reprobed with other antibodies. Images in the same boxes are from the same membranes. (C) LMP1 RNA was increased in p53-expressing cells. LMP1 RNA was examined by semiquantitative RT-PCR. Equal amounts of total RNA were used for cDNA synthesis with random hexamers with (+) or without (−) reverse transcriptase (RT). Amplification was carried out with gene-specific primers, and different PCR cycles were used to ensure that detection of each product was in the linear ranges. Images in the same boxes are from the same samples. (D) Real-time PCR analysis of LMP1 and GAPDH RNA expression in paired cell lines. The expressions of LMP1 and GAPDH RNA in TK6 and NH32 cells were examined with real-time quantitative RT-PCR analyses. Both LMP1 and GAPDH RNA transcripts were quantified, and the relative LMP1 RNA was calculated with the use of 2^−ΔΔCT methods. Error bars indicate standard deviations from five independent experiments. *, P < 0.05. A Student t test was performed with Microsoft Excel. (E) Stability of LMP1 protein in cell lines. NH32 and TK6 cells were treated with CHX at 100 μg/ml, and cell lysates were collected after the indicated periods (in hours). Cell lysates were analyzed by Western blotting for target protein. The target proteins are identified.

DNA-damaging agents stimulate the expression of LMP1.

p53 protein is stabilized in response to DNA damage as an integral part of the DDR pathway (49). Because LMP1 expression is associated with constitutive p53 expression (Fig. 1B), it was of interest to determine whether DNA damage could stimulate LMP1 expression in EBV-transformed cells. IB4, SavIII, and P2 cells were exposed to various DNA-damaging agents, such as cytosine β-d-arabinofuranoside (Ara-C), 5-fluorouracil (5-FU), and X-ray irradiation. Because the TP53 status in SavIII and P2 lines is unknown, exons 5 to 8 of TP53 in both lines were sequenced; the exon sequences in both lines were identical to wild-type (wt) TP53 exons (see Materials and Methods for details). Approximately 90% of TP53 mutations are localized between domains including exons 5 to 8 in cancers (50). Both p53 and LMP1 levels were increased in a time-dependent manner in response to the DNA-damaging treatments (Fig. 2A and B). Because those treatments cause DNA damage through different mechanisms, the data implied that DNA damage per se is capable of stimulating LMP1 expression. In addition, the data provide additional support of the idea that LMP1 is associated with p53 expression (Fig. 1). Moreover, because P2 is lacking STAT1 gene (51), the DNA damage-mediated LMP1 stimulation was not STAT1 dependent.

FIG 2.

p53 is required for DNA damage-mediated stimulation of LMP1. (A) DNA damage-stimulated LMP1 protein expression. EBV-transformed cell lines, IB4, SavIII, and P2, were treated with 5-FU (20 μM), Ara-C (20 μM), or X-ray irradiation (4 Gy) as indicated. Cell lysates were collected at increasing time periods after treatments (in hours). The expression of target proteins was detected by Western blot analysis. The intensities of the target signals were measured by a Bio-Rad ChemiDoc MP imaging system, and the relative LMP1 level (LMP1/GAPDH) is as shown. ##, the GAPDH area may have had a transfer problem, and the time point is omitted for calculation. (B) Quantitation of DNA damage-stimulated LMP1 expression. All the results from 3 to 5 h of treatment with DNA-damaging agents (drugs or X-ray treatments) were used to calculate the average LMP1 induction. Relative LMP1 protein expression levels are shown, and error bars indicate standard deviations. **, P < 0.01. A Student t test was performed with Microsoft Excel. (C) p53 is required for DNA damage-mediated stimulation of LMP1 protein. (Left) NH32 and TK6 cells were treated with 4.0 Gy of X-ray irradiation. Samples were collected at the indicated times posttreatment (in hours). (Right) NH32 and TK6 cells were treated with 5-FU (0, 5, and 20 μM) for 7 h. Western blotting was used for detection of specific proteins. (D) Quantitation of DNA damage-stimulated LMP1 RNA in p53-positive and -negative cell lines. NH32 and TK6 cells were irradiated with X-ray treatment (4 Gy) for 3 h. Total RNA was isolated and real-time RT-PCR was used to quantify both LMP1 and GAPDH transcripts. Relative LMP1 abundances were calculated with the use of 2^−ΔΔCT methods. The relative increases of LMP1 RNA after X-ray treatments (+/− X-ray) were calculated. Error bars indicate standard deviations. **, P < 0.01.

p53 is required for the DNA damage-mediated stimulation of LMP1.

Whether p53 was required for DNA damage-mediated LMP1 stimulation was examined in the TK6 (TP53^+/+) and NH32 (TP53^−/−) cell lines. Cells were treated by X-ray irradiation or 5-FU. At the desired time points, cell lysates were collected for Western blot analyses. Following DNA-damaging treatments, p53-expressing cells (TK6) had increased expression of p53 as well as LMP1; however, in p53-negative cells (NH32), LMP1 expression was not significantly changed (Fig. 2C). Furthermore, the real-time PCR assays indicated that X-ray treatment could increase the expression of LMP1 RNA in TK6 cells but not in NH32 cells (Fig. 2D). The data support the idea that p53 is required for DNA damage-mediated stimulation of LMP1 expression.

Ectopic expression of p53 increases the expression of LMP1.

Next, the role of ectopic p53 in LMP1 expression in EBV-positive lines was examined. NH32, IB4, Akata, and BRLF-1KO cells were transfected with a p53 expression plasmid. As shown in Fig. 3A, the ectopic p53 expression increased the expression of endogenous LMP1 in all lines tested. Expression of the ectopic p53 was also detected in transfected cells, although it was difficult to detect in IB4 cells (Fig. 3B and data not shown). Akata cells have a TP53 frameshift mutation rendering the absence of expression of p53 (Fig. 3B) (52). IRF5, a direct target of p53, was activated in p53-transfected cells as expected (Fig. 3A, bottom). Notably, the Akata line is a type I latency cell line with no constitutive LMP1 expression; therefore, quantification of the LMP1 increases was not done. Furthermore, a p53-mediated increase in LMP1 expression was also detected at the RNA level (Fig. 3C). These data suggested that ectopic p53 stimulated LMP1 expression.

FIG 3.

Ectopic p53 increases LMP1 expression. (A) Ectopic p53 stimulates LMP1 expression. EBV-infected cells were transfected with pcDNA3 (vector) or p53 expression plasmid (CMV-p53). For NH32 and Akata cells, CD4 expression plasmid was used for enriching the transfected cells (see Materials and Methods for details). IB4 cells were transfected with Amaxa technology. Effectene (Qiagen) was used for the transfection of BRLF1-KO cells. Cell lysates were made a day later, and the expression of target proteins was analyzed by Western blotting. The intensities of the target signals were measured, and relative LMP1level (LMP1/GAPDH) is shown. Akata cells are type I latency cells with no constitutive LMP1 expression; therefore, quantification of the increases was not done. (Bottom) Cells were transfected with vector (left lane) or p53 expression plasmid (right lane). Expression of the target genes was examined by Western blotting. (B) Expression of p53 in Akata cells. Akata cells were transfected with p53 expression plasmid or vector control, and the expression of p53 in transfected and nontransfected cells was determined. TK6 was used as a control. Specific proteins are identified. (C) Ectopic p53 expression increases LMP1 RNA. IB4 cells were transfected with vector or p53 expression plasmid with the use of Amaxa technology. One day later, total RNAs were isolated and real-time RT-PCR analyses were used to quantify LMP1 and GAPDH RNA transcripts with the use of 2^−ΔΔCT methods. Relative fold increase is shown. (D) Endogenous LMP1 blocks DNA damage-mediated apoptosis. 293EBVwt and 293EBVΔLMP1 cells were treated with increasing concentration of DNA-damaging drugs (5-FU or Ara-C). One day later, cells lysates were used for Western blotting to detect PARP. The cleavage product of PARP is shown.

EBV lytic replication also induces LMP1 expression (3, 5), and DNA damages may cause viral lytic replication (53). However, in our experimental design, the cells were collected within a few hours of DNA damage treatments (Fig. 2 and 3). Within this time frame, EBV lytic replication would not efficiently initiate LMP1 expression (54). In addition, LMP1 in BRLF-KO cells was also stimulated by p53 despite the fact that BRLF1 is a gene required for complete lytic replication (Fig. 3A) (55). Of note, both LMP1 and p53 have been demonstrated to be negative regulators for EBV lytic replication (56, 57). Therefore, EBV lytic replication might play a minor role in p53-mediated LMP1 stimulation.

Endogenous LMP1 blocks DNA damage-mediated apoptosis.

It is well established that ectopic LMP1 blocks apoptosis. However, whether endogenous LMP1 blocks DNA damage-mediated apoptosis is unknown. To explore this, two isogenic cell lines (293-EBV and 293-EBVΔLMP1) were compared because the two lines are identical except that one harbors a wild-type genome, while the other has an LMP1 deletion in the EBV genome (58). Cells were treated with DNA-damaging agents, and apoptosis induction was examined by detection of the cleavage of poly(ADP-ribose) polymerase (PARP) (59). The DNA-damaging drugs also increased LMP1 expression, as expected (data not shown). The cleavage of PARP is enhanced in the absence of LMP1 (Fig. 3D). Of note, the use of the 293-based system is not ideal because the cells are transformed with an adenoviral protein that is known to block p53 activities. Nevertheless, the data showed that endogenous LMP1 blocked DNA damage-mediated apoptosis (Fig. 3D).

IRF5 is associated with p53-mediated induction of LMP1.

The LMP1 promoter region does not contain an obvious p53 binding site based on bioinformatics analyses (data not shown). It is suspected that p53 may use one or more of its targets to regulate LMP1 indirectly. IRF5 and STAT1 are both p53 targets and LMP1 regulators. STAT1 may not play a significant role in the DNA damage-mediated LMP1 stimulation, because STAT1-null LCL expressed higher levels of LMP1 (Fig. 2A). IRF5 was identified as a candidate because it is (i) a direct p53 target (28), (ii) highly expressed in EBV-transformed cells in which LMP1 is expressed (60), and (iii) involved in the regulation of LMP1 expression (61).

An association between IRF5 and LMP1 expression is apparent among three genetically related cell lines (TK6, Ne72, and NH32) (Fig. 1A and B). Other LMP1 regulators, such as IRF7 and RBP-Jk, were not obviously altered (data not shown). WTK1 is a mutant p53 (M237I) line that is genetically related to TK6, NH32 and Ne72 (Fig. 1A) (45). Albeit with a p53 mutant, WTK1 has high expression of IRF5, and similar LMP1 expressions were observed between TK6 and WTK1 cells (Fig. 4A).

FIG 4.

IRF5 is associated with LMP1 expression and binds to its promoter. (A) IRF5 is associated with LMP1. Lysates from TK6 and WTK1 cells were used for Western blot analysis with the indicated antibodies. The same membranes were stripped and reprobed with other antibodies. Specific proteins are identified. (B) Increased IRF5 expression after DNA damaging treatments in EBV cell lines. NH32, TK6, and IB4 cells were treated with 4 Gy of X-ray irradiation and collected after treatment for the indicated periods (in hours). Western blot analyses were done with proper antibodies. Images in the same boxes are from the same membranes. Specific proteins are identified. Images in the same boxes are from the same membranes. (C) IRF5 binds to LMP1 promoter in vivo. IRF5 antibody was used for ChIP analysis to detect in vivo DNA binding to the LMP1 promoter region in various cells. Primers for the LMP1 promoter region surrounding the ISRE were used to amplify the DNA from immunoprecipitates. The identity of IRF5-bound LMP1 promoter amplification is as shown. “Input” represents the PCR amplification from input lysates with various dilutions. PCR amplification of immunoprecipitates from normal goat serum (NGS) is also shown as a control. (D) Comparison of IRF5 binding to the LMP1 promoter in different cell lines. NH32 and TK6 cells were used for semiquantitative CHIP assays. CHIP assays were used for detection of IRF5-bound LMP1 promoter in NH32 and TK6 cells. DNA markers (in base pairs) are also shown. Dilutions of inputs are shown. Blank, blank lane; control, PCR was done without any ChIP DNA as a negative control. (E) Comparison of IRF5 binding to LMP1 promoter after X-ray irradiation. IB4 and P2 cells were treated with 4 Gy of X-ray irradiation and collected at 0 and 3 h after treatment. ChIP assays were used for detection of IRF5-bound LMP1 promoter. Specific binding is shown. Images in the same boxes are from the same membranes, with different exposure times.

Because DNA damage stimulates IRF5 expression (28), the expression of IRF5 in response to X-ray irradiation in EBV-transformed lines was examined. Stimulation of IRF5 expressions was observed in TK6 and IB4 cells, but not in NH32 cells. However, IRF1 levels increased in all three cell lines (Fig. 4B), suggesting that NH32 was still responding to radiation. Because X-ray stimulates the expression of LMP1 (Fig. 2), all these data suggest that IRF5 expression is associated with LMP1 stimulation and, furthermore, p53 expression.

IRF5 binds to the LMP1 promoter.

The LMP1 promoter has a functional interferon-stimulated response element (ISRE), and IRFs could bind to this site (62, 63). Whether IRF5 could bind to the LMP1 promoter was examined by chromatin immunoprecipitation (ChIP) assay. IRF5 was able to bind to the LMP1 promoter region containing the ISRE in EBV lines (Fig. 4C). Also, more LMP1 promoter DNA could be immunoprecipitated from TK6 cells than that from NH32 line (Fig. 4D). Furthermore, X-ray irradiation seemed to increase the binding of IRF5 to LMP1 promoters (Fig. 4E). All data suggest that IRF5 binds to LMP1 promoter in vivo, and the binding may be correlated with the expression levels of IRF5 as well as stimulation of LMP1.

IRF5 activates LMP1 gene.

Whether IRF5 is able to activate LMP1 promoter reporter constructs was examined in 293T cells. IRF5 expression and LMP1 promoter reporter constructs were cotransfected into cells, and the reporter activities were measured a day later. As seen in Fig. 5A, IRF5 activated a LMP1 promoter reporter construct with the ISRE element but not a reporter construct without the ISRE. Additionally, an IRF5 mutant (IRF5P68) which is considered a dominant negative mutant for IRF5 (64) was unable to activate the LMP1 promoter reporter (Fig. 5A). Of note, at a higher dosage, IRF5-mediated activation of the LMP1 promoter reporter constructs may be less potent, which may be related to the fact that overexpression of IRF5 may cause apoptosis in cells.

FIG 5.

IRF5 increases EBV LMP1 expression. (A) IRF5 activates LMP1 promoter reporters. (Top) Schematic diagram of EBV LMP-1 promoter reporter constructs. The RNA start site is shown. Numbers represent coordinates to the RNA starting site. ISRE, interferon-stimulated response element; RBP-Jκ, recombination signal sequence binding protein RBP-Jκ; GAS, interferon gamma-activated sequence. The drawing is not to scale. (Bottom) 293T cells were transfected with various LMP1 promoter reporter constructs and IRF5 expression plasmid (0.01, 0.05, and 0.1 μg) as shown at the top. Cell lysates were used for the luciferase and β-galactosidase assays. Relative promoter reporter activities (luciferase/β-galactosidase) are shown. The results represent averages from triplicate transfections. Standard derivations are also shown. (B) IRF5 increases EBV LMP1 expression. NH32 cells were transfected via electroporation with pcDNA3 (vector) or IRF5 expression plasmids along with CD4 expression plasmid. One day after the transfection, the transfected cells were enriched using the CD4 magnetic beads (see Materials and Methods for details), and the cell lysates were used for Western blot analysis. IB4 cells were transfected with Amaxa technology with pcDNA3 or IRF5 expression plasmid. One day later, cell lysates were made for Western analysis. The identities of the proteins are shown. The intensities of the target signals were measured by a Bio-Rad ChemiDoc MP imaging system, and the relative LMP1 level (LMP1/GAPDH) is as shown. (C) Reduction of IRF5 suppresses endogenous LMP1 expression. Cells were transfected with shLuc (5 μg) or shIRF5 (mixtures of shIRF5 plasmids) with a Nucleofector device from Amaxa. Total DNA were normalized to 5 μg with vector DNA. Cells were collected 1 day later. The expression levels of LMP1, IRF5, and GAPDH were examined by Western blotting. The intensities of the target signals were measured, and relative LMP1 levels (LMP1/GAPDH) are shown.

Whether IRF5 could stimulate endogenous LMP1 expression was tested. NH32 and IB4 cells were transfected with IRF5 expression plasmids and analyzed by Western blotting. Overexpression of IRF5 in both cell lines increased LMP1 expression (Fig. 5B). To test if endogenous IRF5 is involved in the expression of LMP1, IRF5 expression was knocked down in EBV-transformed cells. IRF5 knockdown resulted in decreases of LMP1 expression in both IB4 and NH32 cells (Fig. 5C). Of note, the times of exposure of the membranes to films for the two panels were different: minimum exposure was used to show the increase, and moderate exposure was used for observation of inhibition. These data suggest that IRF5 is one of the factors involved in the regulation of LMP1 in native environments.

LMP1 inhibits IRF5-mediated apoptosis.

IRF5 is predominantly considered a tumor suppressor that potentiates apoptosis in a p53-independent manner (65), whereas LMP1 is a well-known viral oncogene that blocks apoptosis. Therefore, we suspected that LMP1 may repress apoptosis mediated by IRF5. The two isogeneic cell lines (293EBVwt and 293EBVΔLMP1) were used to examine if LMP1 affects IRF5-mediated apoptosis. Cells were transfected with IRF5 expression plasmid, and apoptosis was examined by the monitoring of the cleavage products of PARP (59). The cleavage of PARP was greatly enhanced in the 293EBVΔLMP1 line, suggesting that IRF5 was able to potentiate apoptosis in EBV-infected cells (Fig. 6A). IRF5 stimulated the expression of LMP1 in those cells, as expected (Fig. 6B). Therefore, endogenous LMP1 was capable of blocking IRF5-mediated apoptosis.

FIG 6.

Endogenous LMP1 blocks IRF5-mediated apoptosis. (A) 293EBVwt and 293EBVΔLMP1 cells were transfected with pcDNA3 (vector) or IRF5 expression plasmid. Total DNAs were normalized with pcDNA3 to the same amounts. The cells were cultured in the 5% or 2% FBS medium 4 to 6 h after transfection. The next day, the transfected cells were collected, and the cell lysates were used for Western blot analysis. Cleavage products are as shown. Images in the same boxes are from the same membranes. The identities of the proteins are shown. (B) IRF5 increases the expression of LMP1. 293EBVwt and 293EBVΔLMP1 cells were transfected with various plasmids as shown at the top. The cells were cultured in the 2% FBS medium. The cell lysates were used for Western blot analysis next day. Images in the same boxes are from the same membranes. The identities of the proteins are shown.

DISCUSSION

It is well established that p53 is highly expressed and functional during EBV infection and transformation in vitro (39). In addition, EBV infection activates DDR prior to LMP1 expression (40). In this study, we investigated the relation between p53 and LMP1. We demonstrated that (i) LMP1 expression was associated with p53 expression in EBV-transformed cells under physiological and DNA-damaging conditions (Fig. 1 and 2) and (ii) ectopic expression of p53-stimulated LMP1 and endogenous LMP1 could block p53-mediated apoptosis (Fig. 3). Therefore, LMP1 may be a target of p53 under physiological and DNA-damaging conditions. Notably, p53 is not absolutely required for LMP1 expression: LMP1 is still expressed in the absence of p53, and some BLs with p53 mutations still express LMP1.

To our knowledge, the LMP1 gene might be the first viral oncogene identified as a p53 target. Interestingly, the LMP1 promoter region does not contain an obvious binding site for p53. Therefore, it is likely that p53 activates LMP1 expression indirectly through an intermediate(s). We demonstrated the following. (i) The expression of IRF5 was correlated with both p53 and LMP1 under physiological and DNA-damaging conditions (Fig. 1, 2, and 4). (ii) IRF5 bound to the LMP1 promoter region, and the binding was correlated with IRF5 expression levels (Fig. 4C to E). (iii) IRF5 activated the LMP1 promoter reporter construct (Fig. 5A). Ectopic IRF5 increased the expression of endogenous LMP1 (Fig. 5B), and the reduction of IRF5 suppressed LMP1 expression (Fig. 5C). (iv) IRF5 potentiates apoptosis independent of p53 (32). IRF5-stimulated LMP1 expression inhibited IRF5-mediated apoptosis (Fig. 6). Finally, (v) WTK1, a mutant p53 cell line, has IRF5 expression higher than that of and LMP1 expression similar to that of the genetically similar TK6 line (Fig. 3A). However, when IRF5 expression was ablated, we could not detect obvious reduction of DNA damage-mediated LMP1 stimulation (data not shown). Because LMP1 could be induced by multiple signals, it is likely that IRF5 is just one of the mediators for p53-mediated stimulation of LMP1.

IRF5 may be activated by phosphorylation and nuclear translocation. However, IRF5 is predominantly localized in the cytoplasm even under DNA-damaging conditions in EBV-transformed cells (data not shown). Of note, IRF5 nuclear localization seemed to be difficult to detect, even in virally activated situations (66). Because overexpression of IRF5 would lead to activation of target genes and IRF5 is bound to DNA in the ChIP assay, overexpression of IRF5 per se would lead to its nuclear translocation.

IRF5 has been reported as a negative regulator of LMP1 (67). While it is hard to reconcile our data with the report, genetic differences in cell lines used, types of assays, promoter construct differences, and IRF5 dosages may collectively cause the two quite different conclusions. IRFs traditionally have dual roles: IRF2, -4, and -7 could be transcriptional repressors and activators in various contexts (68–71). IRF5 is predominantly considered a tumor suppressor; however, it may be a key factor for promoting tumor growth in classical HLs and thyroid cancers (37, 38). Therefore, it might not be very surprising that IRF5 behaves differently in various genetic backgrounds and under various environmental conditions. IRF7 is an established activator for LMP1 (62). Interestingly, previous work on IRF5's functions on LMP1 was mainly conducted with IRF7-low cell lines (67), and our work was mainly done in IRF7-high environments (type III latency cells). Because IRF7 and IRF5 can interact with one another, whether IRF7 plays a role in modulating IRF5's role in LMP1 regulation is of great interest. Of note, LMP1 itself is a dual-function protein, and high levels of LMP1 may be toxic to cells (72). LMP1 may negatively regulate its own expression through a classic feedback loop via NF-κB (73), and the higher levels of IRF5 in WTK1 cells but similar levels of LMP1 between TK6 and WTK1 cells (Fig. 4A) may also be related to the dual effects of LMP1 (72).

During EBV primary infection, LMP1 is expressed at early stages, and LMP1 may be a virion protein (74, 75). DDR is induced prior to LMP1 expression in EBV-infected cells (40). EBV-mediated DDR response and, furthermore, p53-mediated apoptosis may be nullified by LMP1 expression, to ensure the survival of the infected cell (Fig. 7). LMP1 also helps the establishment of EBV latency by inhibiting lytic replication processes (56, 57).

FIG 7.

Schematic diagram of the relation between LMP1 and p53. EBV infection or DNA damage increases p53 and, furthermore, IRF5 expression. IRF5 may contribute to the expression of LMP1, and LMP1 blocks both p53- and IRF5-mediated apoptosis for the survival of virally infected cells. An additional factor(s) may be present for p53-mediated LMP1 stimulation.

Other than EBV primary infection, EBV latency cells may experience various stresses in native environments. Therefore, it is tempting to speculate that EBV may sense various cellular stresses via p53 and adjust LMP1 expression to counteract stress-related cellular damages to ensure the survival of the EBV-infected cells (Fig. 7).

Wild-type p53 correlates with LMP1 expression in some EBV-associated cancers. Of LMP1-expressing NPC and PTLD specimens, the majority have wild-type p53 that is highly expressed (76–78). In cases of EBV-positive HLs, p53 is predominantly wild type, while high frequencies of p53 mutations are observed in EBV-negative HL cases (79). The majority of endemic BLs have p53 mutations and, interestingly, do not express LMP1, although they are EBV positive (80). Furthermore, peripheral blood mononuclear cells (PBMCs) from lupus patients have higher p53 and LMP1 expression, and both of these proteins are associated with disease severity (81, 82). Our data may explain the apparent correlation between wild-type TP53 status and LMP1 expression in some EBV-associated diseases. While additional mechanistic studies of p53-mediated LMP1 stimulation need to be conducted, the relations among p53 and LMP1 may play an important role in EBV latency and transformation in vitro and in vivo and impact EBV-associated disease severity.

MATERIALS AND METHODS

Cell lines, plasmids, antibodies, and viruses.

TK6 (TP53^+/+), NH32 (TP53^−/−), Ne72 (TP53^+/−), and WTK1 (TP53 M237I) are latent EBV-infected, genetically similar cell lines derived from a single EBV-transformed Wil2 line (Fig. 1A). They differ in TP53 gene status and are gifts from Howard Liber (45–47). IB4, SavIII, and P2 are EBV-transformed human primary B cell lines. P2 has STAT1 deletion in the genome (51). TP53 status in IB4 is wild type. TP53 sequence statuses in P2 and SavIII lines were determined by PCR amplification of the exon 5 to 8 regions and Sanger sequencing (Eurofins Genomics). Primers for PCR amplifications were as follows: for exons 5 and 6, 5′-TGTTCACTTGTGCCCTGACT-3′ and 5′-GGAGGGCCACTGACAACCA-3; for exon 7, 5′-GGCGACAGAGCGAGATTCCA-3′ and 5′-GGGTCAGCGGCAAGCAGAGG-3′; and for exon 8, 5′-GACAAGGGTGGTTGGGAGTAGATG-3′ and 5′-GCAAGGAAAGGTGATAAAAGTGGAA-3. Primers for DNA sequencing were as follows: for exons 5 and 6, 5′-TGGTTGCCCAGGGTCCCC-3′, 5′-CACTTGTGCCCTGACTTT-3′, 5′-CCACCCTTACCCCTCC-3′, and 5′-CCTGGGACCCTGGGCAA-3′; for exon 7, 5′-CTCCCCTGCTTGCCACA-3′ and 5′-TCAGCGGCAAGCAGAGG-3′; and for exon 8, 5′-ATGGGACAGGTAGGACC-3′ and 5′-CATAACTGCACCCTTGG-3′. TP53 exon sequences in both lines are identical to the published TP53 sequence in GenBank (accession number NC_018928.2; Homo sapiens chromosome 17, alternate assembly CHM1_1.1) and in the p53 database of International Agency for Research on Cancer (IARC) (http://p53.iarc.fr/TP53Sequence_NC_000017-9.aspx). Akata is a Burkitt's lymphoma (BL) cell line that expresses no endogenous p53 (one p53 allele is deleted and the other allele has a frameshift mutation) (52). All of these cell lines were cultured in RPMI medium with 10% fetal bovine serum (FBS) and 1% Penicillin-streptomycin (PS) at 37°C. The 293T cell line is a human embryonic kidney fibroblast line that is EBV negative. The 293-BRLF1-KO line harbors an EBV genome missing the BRLF1 gene (55). The 293-EBVwt line harbors the wild-type EBV genome, and 293EBVΔLMP1 (293EBVΔ) harbors an EBV genome with an LMP1 deletion (58). All these EBV lines are 293-based lines and were maintained in Dulbecco modified Eagle medium (DMEM) plus 10% FBS. Expression plasmids for IRF5 (v4IRF5), IRF5P68, and p53 were gifts from Paula Pitha, Rongtuan Lin (64), and Shannon Kenney, respectively. The small hairpin RNA expression plasmids for IRF5 (shIRF5) and its control shLuc, β-galactosidase, LMP1 promoter reporter constructs, and CD4 expression plasmid were described before (83–86). LMP-1 (CS1-4) and EBNA2 (PE-2) antibodies were purchased from Dako. The antibodies for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (0411), p53 (C11 or DO-1), IRF1 (C-20), and IRF4 (H140) were purchased from Santa Cruz. IRF5 antibody for Western blot analysis was from Protein Tech (10547-1-AP), and IRF5 antibody for ChIP assays was from Abcam (ab2932-100). Tubulin antibody was purchased from Sigma (T6557).

Western blot analysis with ECL.

Separation of proteins on SDS-PAGE was carried out following the standard protocol. After the proteins were transferred to a nitrocellulose or Immobilon membrane, the membrane was blocked with 5% nonfat dry milk in TBST (50 mM Tris-HCl [pH 7.5], 200 mM NaCl, 0.05% Tween 20) at room temperature for 30 min. It was washed briefly with TBST and incubated with the primary antibody in 5% milk in TBST for 1 h at room temperature or overnight at 4°C. After the membrane was washed with TBST three times (10 min each), it was incubated with the secondary antibody at room temperature for 1 h. The membrane was then washed three times with TBST, treated with enhanced chemiluminescence (ECL) detection reagents, and exposed to BlueBlotHS film from Life Science Products (XR-0810-100). The intensities of the target signals were measured by a Bio-Rad ChemiDoc MP imaging system.

RT-PCR.

Routine methods for semiquantitative reverse transcription-PCR (RT-PCR) were performed. Total RNAs were extracted using TRIzol reagent (Invitrogen) according to the manufacturer's instructions. The RNA samples were treated with DNase I before RT reactions. Two micrograms of total RNA was used for cDNA synthesis with random hexamers with reverse transcriptase or without RT. The same amounts of cDNAs were used for PCR amplification for LMP1 and actin. The LMP1 primers were as follows: LMP1-F, 5′-TCGTTATGAGTGACTGGACTGG-3′, and LMP1-R, 5′-CCTGTCCGTGCAAATTCCA-3′. The actin primers were as follows: Actin 1, 5′-TTCTACAATGAGCTGCGTGT-3′, and Actin 2, 5′-GCCAGACAGCACTGTGTTGG-3′. PCR was performed with Go Taq Flexi DNA polymerase (Promega) by following the manufacturer's protocol. Different PCR cycles were used to ensure that the products were in the linear ranges of the amplifications. The PCR products were separated by electrophoresis on an 8% polyacrylamide gel.

Quantitative real-time RT-PCR assays.

Abundance of target RNAs was quantified with a CFX96 real-time system (Bio-Rad) by following the manufacturer's recommendations. The LMP1 primers were 5′-AATTTGCACGGACAGGCATT-3′ and 5′-AAGGCCAAAAGCTGCCAGAT-3′. Primers for GAPDH were 5′-GAAGGTGAAGGTCGGAGTA-3′ and 5′-GAAGATGGTGATGGGATTTC-3′. The probes for LMP1 and GAPDH were 5′-TCCAGATACCTAAGACAAGTAAGCACCCGAAGAT-3′ and 5′-CAAGCTTCCCGTTCTCAGCC-3′, respectively. The probes were labeled with 6-carboxyfluorescein phosphoramidite (FAM) reporter dye at the 5′ end and 6-carboxytetramethylrhodamine (TAMRA) at the 3′ end as previously described (87). The threshold cycle (2[^minus]ΔΔCT) method was used for calculation of the relative LMP1 expression (LMP1/GAPDH) (88).

Transient transfection, drug treatment, and reporter assays.

Transfection of IB4 and NH32 cells was achieved by using a Nucleofector device from Amaxa. The Nucleofector kits for human B cells and X-001 program were employed for NH32 or IB4 cells. Electroporation was used for the transfection of IB4 (320 V; 925 μF) or NH32 cells (270 V; 975 μF) along with CD4 expression plasmid for concentration of transfected cells as described previously (12, 71, 89, 90). For transfection of 293T and 293-BRLF1-KO cells, Attractene transfection reagent (Qiagen) was used by following the manufacturer's recommendations. Cytosine β-d-arabinofuranoside (Ara-C; catalog number C1768) and 5-fluorouracil (5-FU; catalog number F6627) were purchased from Sigma. X-ray irradiation was done with an RS 2000 biological research irradiator (Rad Source Technologies) by following the manufacturer's procedures. Luciferase assays were performed using an assay kit from Promega according to the manufacturer's recommendation. β-Galactosidase assays were also performed for transfection efficiency. Data were averaged from the triplicate experiments. The luciferase and β-galactosidase assays were performed by standard methods (83, 91).

Isolation of CD4-positive cells.

Enrichment for CD4-positive cells was performed with the use of anti-CD4-antibody conjugated to magnetic beads according to the manufacturer's recommendation (Invitrogen; 11145D). Cells were transfected with CD4 expression and other plasmids. One day after the transfection, the cells were used for isolation of CD4-positive cells with the use of Dynabeads CD4 (Dynal Inc.) The transfected cells were incubated with Dynabeads CD4 at 72 μl of beads/10⁷ cells for 20 to 30 min at room temperature with gentle rotation. CD4-positive cells were isolated by placing the test tube in a magnetic separation device (Dynal magnet). The supernatant was discarded; the CD4-positive cells were attached to the wall of the test tube. The CD4-positive cells were washed 3 times in phosphate-buffered saline plus 2% FBS. The isolated cells were used to prepare cell lysates immediately.

ChIP assay.

Cells were fixed in 1% formaldehyde for 10 min, followed by quenching with 125 mM glycine for 5 min. Chromatin immunoprecipitation (ChIP) was performed using anti-IRF5 antibody (Abcam; ab2932-100) according to a protocol for the EZ ChIP chromatin immunoprecipitation kit (Millipore; catalog no. 17-371), with minor modifications. Normal goat serum (NGS) was added. Equal amounts of DNA were used for PCR with primers that flank the LMP1 ISRE region in the promoter region: LMP1Chip1, 5′-TTCCCACAACACTACGCACT-3′, and LMP1Chip2, 5′-CAGGGCAGTGTGTCAGGAGCAAG-3′. The PCR results were analyzed in an 8% polyacrylamide gel, and the images were captured in a Bio-Rad ChemiDoc MP imaging system.

Protein stability assays.

The protein biosynthesis inhibitor cycloheximide (CHX) was purchased from Sigma (C7698). The cells were treated with CHX (50 to 100 μg/ml), and cell lysates were made after the desired periods of treatment and used for Western blot analysis.