Epstein-Barr Virus (EBV) Nuclear Protein 2-Induced Disruption of EBV Latency in the Burkitt’s Lymphoma Cell Line Akata: Analysis by Tetracycline-Regulated Expression

Abstract

The Burkitt’s lymphoma (BL) cell line Akata retains the latency I program of Epstein-Barr virus (EBV) gene expression and cross-linking of its surface immunoglobulin G (IgG) by antibodies results in activation of viral replication. When EBV nuclear antigen 2 (EBNA2) was artificially expressed by a constitutive expression vector, the Cp EBNA promoter remained inactive and accordingly the latency III program was not induced. In contrast, expression of LMP2A and activity of the Fp lytic promoter were activated. Consistent with this Fp activity, the rate of spontaneous activation of the EBV replicative cycle was increased significantly, suggesting the possibility that EBNA2 can induce EBV replication. The efficiency of anti-IgG-induced activation of the viral replication was reduced in Akata cells expressing EBNA2. To obtain more direct evidence for EBNA2-induced activation of the EBV replicative cycle, this protein was next expressed by a tetracycline-regulated expression system. EBNA2 was undetectable with low doses (<0.5 μg/ml) of tetracycline, while its expression was rapidly induced after removal of the antibiotic. This induced expression of EBNA2 was immediately followed by expression of EBV replicative cycle proteins in up to 50% of the cells, as shown by indirect immunofluorescence and immunoblot analysis. These results suggest an unexpected potential of EBNA2 to disrupt EBV latency and to activate viral replication.

Epstein-Barr virus (EBV) (for reviews, see references 18 and 25) is a ubiquitous herpesvirus, endemic in human populations throughout the world. EBV has been associated with the pathogenesis of a number of malignancies, including Burkitt’s lymphoma (BL), nasopharyngeal carcinoma (NPC), Hodgkin’s disease, peripheral T-cell lymphoma, gastric carcinoma, and immunoblastic lymphoma in immunosuppressed patients. EBV is also the cause of infectious mononucleosis, a self-limiting lymphoproliferative disorder. In vitro, EBV infection of human mature B lymphocytes results in morphological transformation resembling lymphocyte activation and establishment of lymphoblastoid cell lines (LCLs) with capability of unlimited growth in culture.

Two different programs of latent EBV gene expression have been described in B cells that are latently infected with the virus. The latency I program is exemplified by BL cells in vivo and is characterized by selective expression of the EBV nuclear antigen 1 (EBNA1), BARF0, and occasionally the latent membrane protein 2A (LMP2A) (10, 27). The other program, latency III, seen in immunoblastic lymphomas in immunosuppressed patients and EBV-immortalized LCLs in vitro, is characterized by expression of six different EBNAs (EBNAs 1, 2, 3A, 3B, 3C, and LP), three LMPs (LMPs 1, 2A, and 2B), and BARF0 (reviewed in reference 18). EBNA2 is essential for the transformation of B lymphocytes (3, 12, 17) and plays a central role in latency III by up-regulating promoters for all these latent EBV genes (1, 6, 15, 16, 33, 36, 39, 44). EBNA2 exerts its transcriptional activation function by masking the transcriptional repression domain of the recombination signal-binding protein Jκ (RBP-Jκ) (11, 13, 14, 43). Although typical BL cells exhibit the latency I program in vivo, this program is not usually retained in vitro and is replaced by the latency III program after long-term culture (10, 27). In this context, the Akata BL line (34) is exceptional in that latency I has been maintained through long-term in vitro culture. Another unique property of Akata cells is that they have a tendency to lose EBV genomes spontaneously and to give rise to virus-negative sublines (30). Akata cells express surface immunoglobulin G (IgG) molecules, and their cross-linking by antibodies results in activation of EBV replication, through signal transduction pathways involving Ca²⁺ mobilization and activation of protein kinase C (5, 35). In contrast, EBV genomes in LCLs with the latency III phenotype are not significantly activated by ligation of surface immunoglobulin molecules. To examine the effects of EBNA2 on EBV gene expression and anti-IgG-induced viral replication in Akata cells, the EBNA2 gene was introduced by gene transfer experiments.

Establishment of Akata clones stably expressing EBNA2.

For stable and constitutive expression of EBNA2, the expression plasmid pOH-SGE2 was constructed. An AccII-DraI fragment (B95-8 coordinates 48472 to 50303) of EBV DNA including the entire EBNA2 coding region was cloned into the EcoRI site of the eukaryotic expression vector pSG5 (Stratagene) after ligation with an EcoRI linker, and the resulting construct was termed pSGE2. EBV Ori-P DNA fragment (SphI-SacII fragment corresponding to B95-8 coordinates 7333 to 9516) was cloned into the SmaI site of the plasmid vector pBluescript SK(−) (Stratagene) by blunt-end ligation. This construct was then opened by digestion with ClaI and SalI and ligated with a ClaI-SalI fragment containing the simian virus 40 (SV40) early promoter-driven hygromycin B phosphotransferase gene (8), and the resulting plasmid was termed pOH. A fragment containing the SV40 promoter, β-globin intron, EBNA2 gene, and poly(A) signal was excised from pSGE2 by digestion with SalI and ligated with pOH that had been digested by the same enzyme, to generate pOH-SGE2. pOH-SG2E is a control plasmid with its EBNA2 gene put in a reverse direction with respect to the SV40 promoter of pSG5. When EBNA1 is provided in trans, Ori-P is the only cis element required for episomal persistence of a plasmid (40, 41). Since Akata cells produce EBNA1, expressed from their endogenous EBV genomes, pOH-SGE2 was expected to be maintained in Akata cells as multiple copies of episomes, thereby facilitating efficient expression of EBNA2.

pOH-SGE2 or pOH-SG2E was transfected into Akata cells by electroporation, and Akata clones capable of growing with 300 μg of hygromycin B per ml were selected. These clones were further examined for expression of EBNA2 by the labeled streptavidin biotin (LSAB) method using the PE2 anti-EBNA2 monoclonal antibody (42). In total, 220 hygromycin-resistant clones were screened, and two clones were identified in which the majority of the cells were positive for EBNA2 (Fig. 1). Concurrently, two hygromycin-resistant Akata clones were isolated after transfection with pOH-SG2E and used as negative controls.

FIG. 1.

Constitutive expression of EBNA2 in Akata clones harboring pOH-SGE2. (A) Immunoblot analysis. Cellular lysates from Akata cells (lane 1), Akata clones harboring pOH-SG2E (lanes 2 and 3), Akata clones harboring pOH-SGE2 (lanes 4 and 5), B95-8 cells (lane 6), and an LCL (lane 7) were examined by immunoblot analysis with the PE2 anti-EBNA2 monoclonal antibody. Cell lysates representing 5 × 10⁵ cells were analyzed in each lane. Horseradish peroxidase-conjugated antibody to mouse immunoglobulins was used as secondary antibody, and the membrane was developed by the enhanced chemiluminescence method (Pharmacia). (B) Immunoenzymatic staining. Smears of Akata cells (a), an Akata clone harboring pOH-SG2E (b), and an Akata clone harboring pOH-SGE2 (c) were examined with the PE2 antibody by the LSAB method (DAKO), following the protocol supplied by the manufacturer.

Influence of EBNA2 expression on other latent EBV genes.

Since EBNA2 is known to up-regulate transcription from latent EBV promoters, such as Cp and LMP1 and LMP2 promoters (1, 6, 15, 16, 33, 36, 39, 44), activities of some of these promoters as well as expression of latent EBV genes characteristic to latency III were examined by the reverse transcription-PCR method as described previously (24). The results are shown in Fig. 2A and indicated that Cp was not detectably activated by expression of EBNA2, and this is consistent with the absence of detectable levels of EBNA3A and EBNA3B mRNAs. The EBNA2 mRNAs transcribed from the endogenous Akata EBV genome were not detected either, confirming that EBNA2 detected in these cells is indeed expressed from the transfected plasmid. Consistent with the latency I program, the Qp-derived EBNA1 mRNA was detected in both Akata transfectants and controls. LMP1 was detected by the S12 antibody (21) in an Akata transfectant clone expressing EBNA2, but its level was much lower than those in B95-8 and LCLs (Fig. 2B). In contrast, mRNAs transcribed from the Fp promoter and those coding for LMP2A were evident in EBNA2-positive cells and not in control cells (Fig. 2A). EBNA2-induced LMP2A expression is consistent with previous work (44), yet activation of Fp was unexpected and suggested a possibility that expression of EBNA2 is associated with activation of EBV replication, since this promoter was recently shown to be activated in lytic EBV infection (20, 28).

FIG. 2.

Analysis of latent EBV gene expression in Akata cells expressing EBNA2. (A) Reverse transcription-PCR analyses. Messenger RNAs coding for EBNA2, EBNA3A, EBNA3B, and LMP2A, Qp-derived EBNA1 mRNA, and activities of Fp and Cp were assayed with the primers listed below. Two Akata transfectant clones harboring pOH-SGE2 (lanes 5 and 6), and two control Akata clones harboring pOH-SG2E (lanes 3 and 4) were examined. As references, B95-8 cells (lane 1) and EBV-negative Ramos cells (lane 2) were also examined. The sequences and the B95-8 EBV nucleotide coordinates of PCR primers used are as follows: EBNA2 primers, 5′-AGAGGAGGTGGTAAGCGGTTC-3′ (nucleotide [nt] 14802 to 14822) and 5′-TGACGGGTTTCCAAGACTATCC-3′ (nt 48584 to 48563); EBNA3A primers, 5′-TTAGGAAGCGTTTCTTGAGC-3′ (nt 67483 to 67502) and 5′-TCTTCCATGTTGTCATCCAGGG-3′ (nt 92292 to 92271); EBNA3B primers, 5′-TTAGGAAGCGTTTCTTGAGC-3′ (nt 67483 to 67502) and 5′-CATAATCTGGTGGGTCCTCGG-3′ (nt 95431 to 95411); LMP2A primers, 5′-ATGACTCATCTCAACACATA-3′ (nt 166874 to 166893) and 5′-gacgaattcTTTCCAGTGTAAGGCAGTAG-3′ (nt 1639 to 1620); primers for Qp-initiated EBNA1 mRNA, 5′-GTGCGCTACCGGATGGCG-3′ (nt 62440 to 62457) and 5′-CATTTCCAGGTCCTGTACCT-3′ (nt 107986 to 107967); primers for Fp-initiated mRNAs, 5′-ACCCTCCTGTCACCACCTCC-3′ (nt 62284 to 62303) and 5′-ATGCCCTGAGACTACTCTCT-3′ (nt 67563 to 67544); and primers for Cp-initiated mRNAs, 5′-CATCTAAACCGACTGAAGAA-3′ (nt 11470 to 11479 and nt 11626 to 11635) and 5′-CCCTGAAGGTGAACCGCTTA-3′ (nt 14832 to 14813). The sequences and the B95-8 nucleotide coordinates of the probes used are as follows: EBNA2 probe, 5′-GAGAGTGGCTGCTACGCATT-3′ (nt 47885 to 47904); probe for EBNA3A and EBNA3C, 5′-AGAGAGTAGTCTCAGGGCAT-3′ (nt 67544 to 67563); LMP2A probe, 5′-gacggatccATGCTTGTGCTCCTGATACT-3′ (nt 548 to 567); probe for Qp-initiated EBNA1 mRNA, 5′-AGAGAGTAGTCTCAGGGCAT-3′ (nt 67544 to 67563); probe for Fp-initiated mRNAs, 5′-TTAGGAAGCGTTTCTTGAGC-3′ (nt 67483 to 67502); and probe for Cp-initiated mRNAs, 5′-TGGGCGACCGGTGCCTTCTT-3′ (nt 14740 to 14721). Lowercase letters represent non-EBV sequences attached for cloning purposes. (B) Immunoblot analysis. Two Akata transfectant clones harboring pOH-SGE2 (lanes 5 and 6) and control Akata clones harboring pOH-SG2E (lanes 3 and 4) were examined for expression of LMP1 by the S12 monoclonal antibody. As references, EBV-negative BJAB cells (lane 1), Akata cells (lane 2), B95-8 cells (lane 7), and an LCL (lane 8) were also examined. The arrowhead indicates the truncated form of LMP1 characteristic of lytic infection.

Spontaneous disruption of EBV latency in Akata cells expressing EBNA2.

To test if expression of EBNA2 has any influence on spontaneous and anti-IgG-induced disruption of EBV latency, the two Akata transfectant clones expressing EBNA2 and two control clones were cultured with or without goat affinity-purified antibodies to human IgG (whole molecule) (Cappel) and activation of viral cycle was assessed by indirect immunofluorescence (Table 1). Without anti-IgG, less than 0.1% of the control transfectant cells expressed the early antigens (EA). Upon stimulation with anti-IgG, EA was induced in 15 to 29% of these control cells. In contrast, an unexpectedly large fraction (1.1 to 4.8%) of the EBNA2-expressing cells were shown already positive with EA even without treatment with anti-IgG. After addition of anti-IgG, however, the increase in the number of EA-positive cells was significantly smaller than that of control cells (1.4 to 8.1%).

TABLE 1.

Spontaneous and anti-IgG-induced expression of EA in Akata cells expressing EBNA2^a

Cell line	% EA-positive cells
	Expt 1		Expt 2		Expt 3
	anti-IgG(−)	anti-IgG(+)	anti-IgG(−)	anti-IgG(+)	anti-IgG(−)	anti-IgG(+)
Ak/E2/1	1.3	2.9	1.1	1.4	2.5	4.3
Ak/E2/2	2.6	8.1	4.8	7.7	4.1	7.5
Ak/2E/1	<0.1	25	<0.1	27	<0.1	29
Ak/2E/2	<0.1	24	<0.1	15	<0.1	26

^aPercentages of EA-positive cells were counted by indirect immunofluorescence at 24 h after addition of anti-IgG antibody [anti-IgG(+)] and compared with percentages present before addition of anti-IgG antibody [anti-IgG(−)]. More than a thousand cells were examined to estimate the percentages of immunofluorescence-positive cells. 

Tetracycline-regulated expression of EBNA2 in Akata cells.

The results described above suggested that EBNA2 has a potential to induce EBV cycle, and EBV replication is generally presumed to exert deleterious effects on cells. It was therefore suspected that significant numbers of EBNA2-positive cells were lost after transfection with pOH-SGE2. The unexpectedly low rate (2 of 220) of EBNA2-expressing clones among hygromycin-resistant clones (see above), despite the presence of both the EBNA2 gene and the hygromycin resistance gene on the same plasmid, supported this notion. To obtain more direct evidence of EBNA2-induced activation of EBV replication in Akata cells, a tetracycline-regulated expression system was used (9, 31). Two plasmids, pTet-SGE2 and pTAk-Hyg, were constructed according to the scheme described by Floettmann and others (7). pSGE2 was digested with ClaI and SalI and the fragment containing the β-globin intron, EBNA2 gene, and poly(A) signal was isolated. This fragment was then inserted by blunt-end ligation in the EcoRV site of pTet-Splice (GIBCO-BRL), which contains the tetracycline-responsive promoter, and the resulting plasmid was termed pTet-SGE2. A hygromycin resistance gene (ClaI-SalI fragment) (8) was inserted by blunt-end ligation in the NotI site of pTet-tTAk (GIBCO-BRL), which encodes the tetracycline-regulated transactivator, and the resulting construct was designated pTAk-Hyg. Akata cells cotransfected with pTet-SGE2 and pTAk-Hyg were kept in the presence of 0.5 μg of tetracycline per ml and were selected for resistance to hygromycin. Hygromycin-resistant clones were further examined for EBNA2 expression after tetracycline was removed from the culture medium. A number of Akata clones that expressed EBNA2 in a tetracycline-regulated manner were isolated, and three such clones are shown in Fig. 3A. In the presence of 0.5 μg of tetracycline per ml, EBNA2 was not detected either by immunoblot analysis or immunoenzymatic staining. Upon removal of tetracycline, EBNA2 expression was efficiently induced and the level of its expression was higher than the average among EBV-immortalized LCLs (Fig. 3A). Dose response analysis indicated that a threshold level of tetracycline concentration exists around 10 to 20 ng/ml and that two- to fourfold dilution spanning this range gave almost full induction (data not shown). When the cells were examined at various times after removal of tetracycline, EBNA2 was first detected at 12 h and reached the plateau by 48 h (Fig. 3C).

FIG. 3.

Tetracycline-regulated expression of EBNA2 in Akata cells. (A) Immunoblot analysis. Cells of three Akata transfectant clones (lanes 2 through 7) were maintained for 48 h with (lanes 2, 4, and 6) or without (lanes 3, 5, and 7) tetracycline, and cell lysates were analyzed by immunoblotting with the PE2 monoclonal antibody. As references, untransfected Akata cells (lane 1), B95-8 cells (lane 8), and two LCLs (lanes 9 and 10) were also analyzed. (B) Immunoenzymatic staining. Cell smears of an Akata transfectant clone harboring pTet-SGE2 and pTAk-Hyg were cultured for 48 h with (a) or without (b) tetracycline and examined with the PE2 antibody by the LSAB method. (C) Time course of EBNA2 expression after removal of tetracycline. Cells of an Akata transfectant clone harboring pTet-SGE2 and pTAk-Hyg were washed twice with fresh tetracycline-free culture medium and then resuspended in the same medium (0 h). Thereafter, cell lysates were prepared at the indicated times after the start of culture and analyzed by immunoblot analysis.

Disruption of EBV latency by inducible EBNA2 expression.

Smears of an Akata transfectant clone were prepared at 72 h after removal of tetracycline and examined for expression of EA and viral capsid antigens (VCA) by indirect immunofluorescence. The results are shown in Fig. 4A and indicate that EA was induced in more than 50% of the cells and that VCA was induced in around 25% of the cells. Similar levels of EA and VCA were detected in three other Akata transfectant clones after the removal of tetracycline (data not shown). Proliferation of these cells was significantly retarded, and their viability declined in several days after induction of EBNA2 (data not shown). Expression of EBV lytic-cycle proteins similar to those induced by anti-IgG antibodies was also detected by immunoblot analysis of three Akata transfectant clones (Fig. 4B). As expected, no such replication cycle proteins were induced in control Akata transfectants harboring pTAk-Hyg alone (Fig. 4A and B). These lytic-cycle proteins were first detected at a low level at 18 h after the removal of tetracycline, increased until 36 h, and remained at a plateau until 96 h (Fig. 4C). The BZLF1 protein, an immediate-early protein critical to activation of EBV replication from the latent state, was also shown to be induced (Fig. 4D). A separate immunoblot analysis using a monoclonal antibody to the BFRF3 protein, an immunodominant component of the EBV capsid, indicated that the 18-kDa band seen in Fig. 4B and C corresponds to this protein (data not shown).

FIG. 4.

Expression of EBV replicative cycle proteins following induced expression of EBNA2. (A) Immunofluorescence. An Akata transfectant clone harboring both pTet-SGE2 and pTAk-Hyg were maintained in culture medium with (a) or without (b and c) tetracycline and examined by indirect immunofluorescence. Serum from an NPC patient was used to detect EA (a and b) and that from a healthy EBV carrier was used to detect VCA (c). As a reference, a control Akata transfectant clone harboring pTAk-Hyg alone was maintained in tetracycline-free medium and examined with serum from a patient with NPC (d). (B) Immunoblot analysis. Three Akata transfectant clones harboring pTet-SGE2 and PTAk-Hyg (lanes 3 through 8) and a control Akata clone harboring pTAk-Hyg alone (lanes 9 and 10) were washed twice with tetracycline-free fresh medium and cultured for 48 h in the same medium (lanes 3, 5, 7, and 9) or medium containing tetracycline (lanes 4, 6, 8, and 10) and examined by immunoblot analysis with pooled sera from patients with NPC. As a reference, Akata cells before (lane 1) and after (lane 2) treatment with anti-IgG antibodies were also examined. (C) Time course of induction of EBV replicative cycle proteins. Protein samples prepared at the indicated times after removal of tetracycline from culture medium were probed with pooled sera from patients with NPC. The protein samples are identical to those shown in Fig. 3C. (D) Time course of synthesis of the BZLF1 protein. Protein samples prepared at the indicated times after removal of tetracycline from culture medium were probed with the BZ.1 monoclonal antibody. The protein samples are identical to those shown in Fig. 3C.

In this study, EBNA2 was artificially expressed in Akata cells first by the vector pOH-SGE2 that achieved its constitutive expression. The result of this experiment indicated that in Akata cells expressing EBNA2 (i) the rate of spontaneous activation of the EBV replicative cycle was increased significantly, (ii) the efficiency of the anti-IgG-induced viral cycle was decreased, and (iii) the Cp promoter was not activated and therefore the latency III program was not induced. The decreased rate of anti-IgG-induced viral cycle is not due to down-regulation of surface IgG expression, since flow cytometrical analysis with fluorescein isothiocyanate-conjugated anti-human IgG indicated that EBNA2 did not alter its level on the cell surface (data not shown). Instead, it is better explained by the EBNA2-induced up-regulation of LMP2A. LMP2A has been shown to block Ca²⁺ mobilization and thereby to impede activation of the EBV cycle triggered by cross-linking of surface immunoglobulins (22). The Cp promoter, a hallmark of latency III, was not induced by EBNA2, and this is consistent with the recent finding that DNA methylation around Cp is a decisive factor in the maintenance and possibly establishment of latency I (23, 26, 29).

Inducible EBNA2 expression by the tetracycline-regulated system confirmed and gave more decisive evidence for the disruption of EBV latency induced by the protein. The EBV replicative cycle was more efficiently induced (>50%) as compared with the experiments with the noninducible vector pOH-SGE2 (1.1 to 4.8%). Considering the deleterious effects of the EBV replicative cycle on cells, clones with lower rates of EBV activation should have an advantage after transfection with pOH-SGE2. The lower rates of EBV activation in the pOH-SGE2 transfectant clones are thus likely to be a result of selection. The relationship between the EBNA2 dose and activation of EBV cycle remains to be elucidated.

Akata is a rare exception among BL cell lines in that its latency I phenotype was not replaced by latency III after long-term in vitro culture. The EBNA2-induced disruption of the EBV cycle may provide an explanation for this unique property of the cell line. Similar to other BL-derived cell lines, occasional Akata cells may express EBNA2 spontaneously, yet instead of inducing the latency III phenotype, EBNA2 expression will result in activation of the EBV replicative cycle and cell death. This scenario may be also relevant to the mechanism of spontaneous loss of EBV genomes from Akata cells. If the rate of spontaneous EBNA2 expression and consequent viral replication is beyond a certain level, cells that have lost EBV genomes should have an advantage for survival. It will be interesting to test if EBNA2 activates viral cycles in the other few BL cell lines that retain the latency I phenotype.

The effect of EBNA2 on the expression of other EBV and cellular genes has been analyzed mainly with gene transfer experiments with constitutive expression vectors. These experiments provided evidence that EBNA2 plays a central role in the latency III program by transactivating the Cp EBNA promoter and promoters for LMP1 and LMP2 (1, 6, 15, 16, 33, 36, 39, 44). It was also demonstrated that, in cooperation with LMP1, EBNA2 is responsible for the induction of the activated B-cell phenotype (2–4, 12, 17, 19, 32, 37, 38). In these previous experiments, the EBNA2 gene was transferred mainly to EBV-negative cell lines, and, to our knowledge, its effects on replicative-cycle genes have not been investigated. The present study, in contrast, employed an inducible expression vector and chose EBV-positive BL cells with the latency I phenotype as recipients. This unconventional approach has been essential for the unexpected finding of the EBNA2-induced EBV cycle. This EBV-activating potential of EBNA2 may be dependent on cellular phenotype, because the protein is expressed in LCLs of the latency III phenotype without apparently inducing EBV replication. It is also plausible that EBNA2 acts differently on gene regulation depending on its level of expression or the presence of other latent EBV proteins or both.

The molecular mechanism involved in the EBNA2-induced EBV cycle remains an open question. Since the RBP-Jκ-binding motif (GTGGGAA) is not found close to immediate-early EBV genes such as BZLF1 and BRLF1, it does not appear likely that EBNA2 directly transactivates them. A more likely mechanism is that EBNA2 primarily transactivates certain cellular genes and that their products change the cellular environment to enhance EBV activation. Although the significance of EBNA2-induced activation of viral cycle in the physiology of EBV is not known now, it may shed light on an unknown area of the EBNA2 function.