The Epstein-Barr Virus Rta Protein Activates Lytic Cycle Genes and Can Disrupt Latency in B Lymphocytes

Abstract

The transition of Epstein-Barr virus (EBV) from latency into the lytic cycle is associated with the expression of two immediate-early viral genes, BZLF1 and BRLF1. Overexpression of ZEBRA, the product of BZLF1, is sufficient to disrupt latency in B lymphocytes and epithelial cells by stimulating expression of lytic cycle genes, including BRLF1. The BRLF1 product Rta functions as a transcriptional activator in both B lymphocytes and epithelial cells. However, Rta has recently been reported to disrupt latency in an epithelial specific manner (S. Zalani, E. Holley-Guthrie, and S. Kenney, Proc. Natl. Acad. Sci. USA 93:9194–9199, 1996). Here we demonstrate that expression of Rta is also sufficient for disruption of latency in a permissive B-cell line. In HH514-16 cells, transfection of Rta leads to synthesis of ZEBRA, viral DNA replication, and late gene expression. However, Rta by itself is less potent than ZEBRA in the ability to activate most early and late lytic cycle genes. In light of previous work implicating ZEBRA in the activation of Rta, we suggest a cooperative model for EBV entry into the lytic cycle. Expression of either BZLF1 or BRLF1 triggers expression of the other immediate-early factor, and together these activators act individually or in synergy on downstream targets to activate the viral lytic cycle.

Latency, a state of limited viral gene expression observed for many viruses, is a prominent feature of infection by gammaherpesviruses. Epstein-Barr virus (EBV), a human gammaherpesvirus associated with both lymphoid and epithelial cell-derived human cancers, is latent in human B lymphocytes (for a review see reference 28). EBV can reactivate from B-lymphoid cells in vivo, an event that can be simulated in B-lymphoid cell culture by addition of a variety of inducing agents that include phorbol esters, n-butyrate, calcium ionophores, and cross-linking of surface immunoglobulin (3, 32, 34, 61). Induction of the switch between latency and lytic cycle gene expression by such stimuli is associated with expression of two immediate-early (IE) proteins that function as transcriptional activators, ZEBRA (Z EB replication activator), the product of the EBV BZLF1 gene, and Rta (R transactivator), the product of the BRLF1 gene. BZLF1 and BRLF1 are expressed simultaneously, usually within 2 h after application of an inducing stimulus (34, 37, 50, 52). ZEBRA and Rta then act on downstream promoters to activate a cascade of lytic gene expression, including in some permissive cell backgrounds, viral DNA replication, late gene expression, and the assembly of infectious virus (5, 8, 9, 11, 12, 22, 23, 27). ZEBRA also plays an essential role in lytic DNA replication (14, 46, 47).

The BRLF1 and BZLF1 genes are expressed from the virus in an overlapping transcription unit (33). The BRLF1 gene is expressed as 4.0- and 3.0-kb mRNAs driven from its upstream promoter, Rp; the two mRNAs are related through alternative splicing. The BZLF1 gene is contained as a bicistronic unit in these two messages. BZLF1 is also expressed as a monocistronic 1.0-kb mRNA that is controlled from a promoter immediately upstream the BZLF1 open reading frame (ORF), Zp (15, 16, 50). It is unclear, however, whether ZEBRA is efficiently translated from the bicistronic message in vivo (30).

A key unanswered question that is addressed in this report is whether ZEBRA or Rta is the principal controlling element of the latency to lytic cycle switch in B lymphocytes. An answer to this question is important to the basic understanding of the mechanisms of maintenance of latency and reactivation into the productive cycle. All current models assume that latency and reactivation are regulated through cellular control of Zp, the promoter that regulates BZLF1, and Rp, the promoter controlling BRLF1. Until now, considerable evidence has favored the dominant role of ZEBRA. Transfection of plasmids expressing ZEBRA into B-cell lines latently infected with EBV efficiently initiates the entire lytic cascade leading to production of virions (11, 19, 44, 53). As an early event, ZEBRA stimulates expression of Rta by acting on Rp, which contains binding sites for ZEBRA (29, 30, 50). If the lytic cascade is linear, with ZEBRA directing synthesis of Rta, entry into the lytic cycle should be accelerated by heterologous expression of BRLF1. Rta is a potent transcriptional activator that drives EBV gene expression by directly binding to responsive promoters that contain Rta response elements; Rta also activates other promoters that do not contain identifiable Rta response elements, presumably by an indirect mechanism (31, 41, 50, 58). In addition, Rta is known to synergize with ZEBRA in the activation of many viral genes (7, 8, 12, 23, 27, 42).

Available evidence about the capacity of Rta to activate EBV lytic gene expression in B cells is conflicting (4, 12, 57). Early reports indicated that transfection of Rta by itself could not disrupt latency (12). One study suggested that Rta might induce low-level lytic gene expression in lymphoblastoid cells but not in Burkitt’s lymphoma (BL)-derived cell lines (4). A recent study indicated that Rta disrupted latency in a cell-specific manner, exerting a positive effect in epithelial cells but not in lymphoid cells (57).

Several observations provoked us to reexamine the role of EBV Rta in control of lytic cycle gene expression in B cells. Since Rta is expressed in the same temporal class as ZEBRA, Rta should exert an essential regulatory function during the initiation of the lytic cycle. Homologues of EBV Rta are found in all members of the gammaherpesvirus family, while ZEBRA is often poorly conserved (36, 54). The homologue of EBV Rta, encoded in ORF 50 of human herpesvirus 8, a gammaherpesvirus associated with Kaposi’s sarcoma and primary effusion lymphoma, is capable of activating human herpesvirus 8 lytic gene expression in B cells derived from primary effusion lymphoma (51a). Our results indicate that EBV Rta can likewise activate lytic gene expression in B lymphocytes.

MATERIALS AND METHODS

Cell lines.

Cells were maintained in 5% CO₂ at 37°C in RPMI 1640 supplemented with 8% fetal calf serum. HH514-16 is a clonal derivative of the P3J-HR-1 B-cell line derived from an EBV-positive BL that is permissive for viral replication (43); Raji is a human B-cell line derived from a BL containing an EBV strain that is defective for DNA replication and late gene expression (40). BJAB is a human B-cell line originating from an EBV-negative B-cell lymphoma (35); BJAB-B1 is the same cell line converted with the P3J-HR-1 EBV strain (18).

Antisera.

Anti-R is a polyclonal rabbit antiserum raised to a purified 320-amino-acid N-terminal fragment of R (R320) that was generated by using the pET expression system (Novagen). Anti-ZEBRA and anti-BLRF2 are polyclonal rabbit antisera raised to TrpE-BZLF1 and TrpE-BLRF2 fusion proteins (26, 48). R3 is a monoclonal antibody to BMRF1 (38), and SJ is a human serum with reactivity to EBNA1 and BFRF3 (48).

Chemical induction.

Cells were subcultured 3 days prior to drug treatment or transfection. Drug treatment consisted of the addition of either 10 ng of tetradecanoylphorbol-13-acetate (TPA) per ml or 3 mM sodium butyrate, or both, to induce the lytic cycle.

Expression vectors and reporter plasmids.

The ZEBRA expression vectors pBXG1-genomic Z (pBXG1/gZ) and CMV-genomic Z (CMV/gZ; CMV denotes cytomegalovirus) and their parent vectors pBXG1 and pHD1013 have been described previously (13, 17). CMV-RIE (CMV/Rta) contains the BRLF1 gene in the vector pHD1013 (23). The Rta expression vector RTS15 (pRTS/Rta), a kind gift of Diane Hayward, contains the BRLF1 gene linked to the simian virus 40 early promoter and is followed by a polyadenylation signal in plasmid pRTS2. RTS15-ΔHIII (pRTS) was generated by excising the EBV BRLF1 gene with HindIII and XbaI, filling in the sites, and religating the vector. The chloramphenicol acetyltransferase (CAT) construct ZpCAT contains the BamHI-NaeI fragment of the BZLF1 promoter cloned into pCAT basic (Promega); RpCAT was generated by cloning the PCR-amplified region spanning nucleotides (nt) 106123 to 107143 (1) of the BamHI R fragment of the EBV genome into the XbaI site in pCAT basic. RpΔTATACAT (equivalent to RpCAT with a 7-bp TATA box deletion) was constructed by a two-step PCR method. The first set of amplification reactions were carried out with primers 5′-GCTCTAGAAAAGGCCCTGTCGTCGGG (I) and 5′-GCCATTGGCATGGGCG (II) for the 5′ fragment and primers 5′-CCATGCCAATGGCTGACCAGTAATCCATGT (III) and 5′-GCTCTAGACCTGCGTCTGTTTGTG (IV) for the 3′ fragment. The second-step reaction contained the products of the first round as templates and primers I and IV. The reporter plasmid pGL2 basic+HMP has been described elsewhere (49). The bacterial R320 expression vector pET22-R320 was generated by cloning the BRLF1 ORF spanning residues 1 to 320 into the NdeI and SalI sites of pET22b (Novagen) in frame with the C-terminal His tag.

Transfections.

Transfections were carried out by electroporation (49). For Western analysis, 10 μg of plasmid DNA (5 μg of expression vector plus 5 μg of empty vector, either pBXG1 or pRTS) was used. Reporter assays were carried out with 10 μg of CAT construct plus 5 μg of expression vector pBXG1 or pRTS and 1 μg of pGL2 basic+HMP. For Southern analysis, the total amount of DNA per transfection was 20 μg (10 μg of each expression vector or one expression vector and pHD1013). For Northern analysis, cells were transfected with 5 μg of each expression vector.

Reporter assays.

CAT and luciferase assays were performed as described previously (49). CAT results represent the averages of at least two separate transfections and are standardized for transfection efficiency with luciferase.

Protein extracts and Western blots.

Cells were collected by centrifugation, washed once in phosphate-buffered saline, and resuspended in sodium dodecyl sulfate (SDS) sample buffer at 10⁶ cells/10 μl. Prior to separation on SDS–12% polyacrylamide gels, samples (20 μl) were heated to 100°C for 5 min. Following electrophoresis, the proteins were transferred to nitrocellulose membranes by electroblotting and blocked in 5% nonfat dry milk overnight at 4°C. The blots were incubated with antiserum, diluted in 5% nonfat dry milk at 25°C for 2 h, then washed (three times for 10 min each) in TS (10 mM Tris [pH 7.5], 200 mM NaCl, 5% Tween 20), incubated with ¹²⁵I-protein A for 1 h, and washed again. The membranes were exposed overnight with intensifying screens to Kodak XAR-5 film at −70°C.

RNA isolation and Northern blots.

Samples of 2.5 × 10⁶ cells were harvested 30 h following transfection. Cellular RNA prepared with an RNeasy Mini Kit (Qiagen) was electrophoresed through a 1% agarose–6% formaldehyde gel in 20 mM MOPS (morpholinepropanesulfonic acid [pH 7.0]), the RNA transferred to a Nytran membrane, and probed with ³²P-radiolabeled oligonucleotides. The BZLF1 oligonucleotide used spanned nt 93 to 128 in exon I of the BZLF1 gene. A radiolabeled 370-bp NcoI-PstI fragment of H1 RNA of RNase P was used as a probe to control for RNA loading (2).

DNA replication assay.

Three days following subculturing, triplicate aliquots of 10⁷ cells were transfected and pooled into 24 ml of RPMI 1640–8% fetal calf serum. After a 3-day incubation period at 37°C, the cells were harvested and total DNA was extracted (49). After resuspension in TE (10 mM Tris [pH 8.0], 1 mM EDTA [pH 8.0]), 5 μg of each DNA sample was digested with BamHI for 5 h at 37°C. The fragments were separated on a 0.8% agarose–1× Tris-borate-EDTA gel and transferred to nitrocellulose in 20× SSC (3 M NaCl, 0.3 M sodium citrate). The blot was hybridized with a ³²P-labeled 1.9-kb XhoI fragment covering the unique sequences adjacent to the terminal repeat region of the EBV genome (Xho 1.9 [45]) or with a 32-base oligonucleotide in this region (5′-CAGCTGTTTTCGTGGACTTTTATACAGTAAGG).

RESULTS

EBV Rta activates early lytic gene expression in Raji cells.

Zalani et al. reported that transfection of the BRLF1 gene under control of the CMV IE promoter into Raji cells, a BL B-lymphocyte cell line, failed to activate the EBV lytic cycle, as measured by the appearance of the EBV diffuse early antigen (EA-D), a lytic cycle product (57). In their experiments, however, Zalani et al. provided no data on the level of expression of the Rta protein. To monitor the expression of Rta while assessing its biologic activity, we used an antiserum raised in rabbits to the N-terminal 320 amino acids of Rta. Two constructs were used to express Rta in Raji cells. One, under control of the CMV promoter but lacking mRNA 3′-end processing signals (CMV/Rta), was identical to the construct used by Zalani et al. (57). The other, driven by the simian virus 40 promoter (pRTS/Rta), contained signals for mRNA 3′-end processing. Transfection of BZLF1 served as a positive control for activation of lytic cycle gene expression in B cells. As expected, transfection of BZLF1 activated EA-D and Rta (Fig. 1A, lane 3). High levels of Rta protein were expressed following transfection of pRTS/Rta but not of CMV/Rta (compare lanes 4 and 7). Transfection of plasmid pRTS/Rta also activated EA-D expression in Raji cells, while CMV/Rta did not. Since the same amount of EBNA1 was detected in all samples, the absence of Rta and EA-D signals in CMV/Rta-transfected cells was not due to a loading artifact. The level of EA-D expressed following transfection of pRTS/Rta was somewhat less than following transfection of BZLF1 (compare lanes 3 and 7). The failure of Zalani et al. to observe lytic cycle activation in lymphoid cells is likely to be due to a low level of expression of the EBV Rta protein from the CMV/Rta vector. In related experiments, we found that transfection of pRTS/Rta activated expression of mRNAs of several lytic cycle genes, including BaRF1 and BMRF1, in Raji cells (data not shown).

FIG. 1.

Activation of early and late viral genes following transfection of BRLF1 into B-cell lines. (A) Raji cells were untreated (lane 1) or were transfected with 5 μg of the indicated plasmids (lanes 2 to 7). Immunoblots prepared 72 h following transfection were probed sequentially with polyclonal rabbit antibody to Rta, with murine monoclonal antibody to EA-D, and with a human antibody to EBNA1. (B) HH514-16 cells were untreated (lane 1), chemically induced with TPA–n-butyrate (lane 2), or transfected with 10 μg of plasmid DNA (lanes 3 to 7). In lanes 4 and 6, the cells received 5 μg of activator and 5 μg of empty vector. In lanes 3 and 5, cells received only vector pBXG1 (lane 3) or pRTS (lane 5). In lane 7, both BZLF1 and BRLF1 were transfected. Immunoblots were probed sequentially with the indicated antisera. Anti-BZLF1 detects both the endogenous (top arrow) and transfected (bottom arrow) ZEBRA.

EBV Rta leads to the activation of late gene expression and lytic viral DNA replication in permissive B lymphocytes.

Due to a mutation in the EBV major DNA binding protein, Raji cells are not permissive for lytic viral DNA replication or production of virions. However, in cell lines such as HH514-16 (Fig. 1B), the latent EBV genome can be activated by inducing chemicals or transfection of BZLF1 to high-level viral lytic gene expression that proceeds to expression of late genes and production of progeny virus. Transfection of pRTS/Rta into HH514-16 cells led not only to the appearance of EA-D but also to the synthesis of low levels of ZEBRA from the endogenous virus (Fig. 1B, lane 6). Endogenous ZEBRA was readily distinguished from ZEBRA from the transfected expression vector due to a convenient polymorphism within BZLF1 (10, 25, 37). Moreover, transfection of pRTS/Rta also induced the expression of a 21-kDa virus structural protein encoded in BFRF3, a late viral gene (49). Transfection of HH514-16 cells with comparable amounts of vector induced neither early nor late genes. As in Raji cells (Fig. 1A), transfection of HH514-16 cells with BZLF1 was more potent than transfection of BRLF1 in activation of expression of EA-D; BFRF3 was also induced to a higher level by ZEBRA than by Rta (compare lanes 4 and 6). In B95-8 marmoset cells, another line that is completely permissive for EBV replication, transfection of BRLF1 also induced expression of EA-D and BFRF3 (not shown).

The appearance of a late viral polypeptide in HH514-16 cells following transfection of BRLF1 suggested that the EBV genome in these cells was induced into lytic replication. Lytic EBV replication is accompanied by amplification of the circular EBV episome and a complex DNA processing reaction that generates linear viral genomes with a heterogeneous collection of terminally repeated DNA fragments (45). Transfection of HH514-16 cells with BRLF1 or BZLF1 led to amplification of the genome and to the appearance of the ladder of terminal repeats (Fig. 2). The probe used in this assay also detected the transfected plasmids; however, an oligonucleotide probe that does not detect the transfected DNA also demonstrated plasmid amplification and terminal repeat processing following transfection of Rta (Fig. 2B, lane 3).

FIG. 2.

Activation of EBV DNA replication following transfection of BRLF1 into HH514-16 cells. Cells were transfected with 20 μg of total plasmid DNA containing vector alone (lane 1), vector plus genomic ZEBRA (lane 2), vector plus Rta (lane 3), or both activators (lane 4). Untreated cells (lane 5 and lane 6) or butyrate-treated cells (lane 7) were used as negative and positive controls. DNA digested with BamHI was analyzed by Southern blotting using either the Xho 1.9 probe (45) (A) or an oligonucleotide probe (B). Episomal EBV DNA (E) and input vector DNA (•) are indicated. L, linear DNA.

Rta strongly activates Rp linked to the CAT reporter in EBV-positive cells.

Since Rta activates ZEBRA expression and ultimately the entire lytic viral cascade in HH514-16 cells, it is likely to do so by acting on one or both promoters of the two IE genes, namely, Rp and Zp. This possibility was examined in the HH514-16 cell background in two ways: by the use of reporter-based transient transfection assays (Fig. 3) and by means of analysis of the viral transcripts originating from Rp and Zp (Fig. 4).

FIG. 3.

Comparison of transcriptional activation of CAT reporters containing Rp and Zp by Rta and ZEBRA. (A) Autostimulation of RpCAT by Rta. HH514-16 cells were transfected with 16 μg of total plasmid DNA with the indicated amount of activator, 10 μg of CAT reporter, and 1 μg of luciferase reporter. CAT and luciferase assays were performed 72 h following transfection. CAT activity was standardized for transfection efficiency by using the luciferase data. Fold activation is the ratio of CAT activity generated by a reporter in the presence of Rta (R) divided by the CAT activity in the presence of empty pRTS vector (U). The data for RpΔTATACAT are standardized together with those for RpCAT, where RpCAT transfected with empty vector equals 1. pCAT, reporter lacking any promoter or enhancer elements; RpΔTATACAT, 7-nt TATA box deletion in Rp. (B and C) Effects of different cell backgrounds on transcriptional activation of RpCAT and ZpCAT by Rta (B) and ZEBRA (C). For both panels B and C, fold activation for each cell line represents the CAT activity generated by the reporter pCAT, RpCAT, or ZpCAT in the presence of Rta or ZEBRA divided by the activity in the presence of empty vector. cl16, HH514-16.

FIG. 4.

Activation of Rp- and Zp-driven transcripts from the latent EBV genome following transfection of Rta and ZEBRA. HH514-16 cells were untreated (lane 1), chemically induced (lane 2), or transfected with 5 μg of total plasmid DNA (5 μg of pRTS [lane 3] or pRTS/Rta [lane 4], 5 μg of pBXG1 [lane 5], or 4 μg of pBXG1 plus 1 μg of pBXG1/gZ [lane 6]). Total cell RNA prepared 28 h following transfection was analyzed by Northern blotting using a 33-nt oligonucleotide in BZLF1 as a probe. Transcripts initiating from Rp, Zp, and vector (V) are indicated.

In transient assays, Rta stimulated its own promoter powerfully, up to 65-fold (Fig. 3A). Stimulation of RpCAT was linearly related to the input dose of Rta expression plasmid and was dependent on the presence of an intact core promoter. Removal of the TATA element in plasmid RpΔTATACAT abolished any Rta-mediated activity.

The extent of Rta stimulation of RpCAT was markedly augmented in the presence of an endogenous EBV genome. In EBV-negative BJAB cells, Rta activated RpCAT weakly, about sixfold. To distinguish between an effect of cell background and the effect of the presence or absence of EBV, we performed the same assays with BJAB-B1 cells, which are BJAB cells that have been permanently converted to carriage of EBV (18). In BJAB-B1 cells, Rta stimulated RpCAT 70-fold, at least 10-fold more strongly than in the parental EBV-negative BJAB cells (Fig. 3B). Thus, RpCAT is strongly activated by Rta in an EBV-positive cell background.

Although ZpCAT was activated by Rta, this promoter was less responsive to Rta than was RpCAT in EBV-positive cells. In HH514-16 cells, Rta stimulated ZpCAT 12-fold; direct comparison of the responses of ZpCAT and RpCAT to Rta in HH514-16 cells revealed that RpCAT was about sevenfold more responsive than ZpCAT (Fig. 3B). In BJAB-B1 cells, Rta did not stimulate ZpCAT above the background level of stimulation of a pCAT basic vector that contained no promoter.

ZEBRA strongly activates the BRLF1 promoter in EBV-negative cells.

ZEBRA behaved very differently from Rta in the transient transfection assays (Fig. 3C). By contrast to Rta, ZEBRA was considerably more active in an EBV-negative than in an EBV-positive cell background. For example, ZEBRA stimulated RpCAT 80-fold in BJAB cells and only 4-fold in BJAB-B1 cells. ZEBRA stimulated ZpCAT 30-fold in EBV-negative BJAB cells but not at all in EBV-positive BJAB-B1 cells. The relatively weak activity of ZEBRA on RpCAT and ZpCAT reporters in an EBV-positive background was also seen in HH514-16 cells, where RpCAT was stimulated only 5-fold and ZpCAT was not stimulated at all. The differences observed in the activities of Rta and ZEBRA in the three different cell lines could not be accounted for by differences in levels of expression of the two proteins. For example, Rta was expressed at comparable levels in HH514-16 cells, where it was highly active, and in BJAB cells, where its activity was low. ZEBRA was expressed at lower levels in BJAB cells, where it was highly active, than in HH514-16 cells, where its activity was low. In BJAB-B1 cells, where ZEBRA showed little activity, its expression was comparable to that in BJAB cells. In summary, the transient transfection assays showed that Rta was a more potent transcriptional activator of RpCAT and ZpCAT in EBV-containing cell lines than was ZEBRA (data not shown).

Rta activates Rp and Zp from the virus.

A probe for BZLF1 detects the 4.0- and 3.0-kb mRNAs originating from Rp and the 1.0-kb mRNA originating from Zp. The 4.0-kb mRNA is characteristically less abundant. Figure 5 shows that chemical induction of the lytic cycle in HH514-16 cells by TPA and n-butyrate leads to the appearance of two abundant 3.0- and 1.0-kb mRNAs detected by a probe for BZLF1; this finding indicates that both Rp and Zp are active in HH514-16 cells. Transfection of ZEBRA also leads to the appearance of these two mRNAs as well as a third 1.3-kb mRNA that represents a transcript from the ZEBRA expression plasmid (29). Transfection of BRLF1 also leads to the appearance of the 3.0- and 1.0-kb mRNAs. Thus, overexpression of Rta following transfection of HH514-16 cells leads to activation of Rp and Zp, the two promoters that control the EBV lytic cascade.

FIG. 5.

Proposed model for disruption of latency in HH514-16 cells. The promoter regions of the BZLF1 and BRLF1 genes are illustrated. See Discussion for an explication of the model. H, repressive host factor (possibly histone); A, host cell activator; ZRE, ZEBRA response element; TRE, TPA response element, also known as ZII (16); Z, ZEBRA; R, Rta; X, a factor which mediates R activity on Rp and Zp.

DISCUSSION

Here we present three novel lines of evidence that emphasize the importance of the EBV Rta protein in events leading to activation of the viral lytic cascade in human B lymphocytes. First, transfection of Rta-expressing plasmids leads to the appearance of early and late viral lytic cycle polypeptides and, in permissive B cells, lytic viral DNA replication (Fig. 1 and 2). Second, in transient transfection assays, Rta powerfully stimulates its own promoter, Rp, in a manner that is dependent on the presence of an EBV genome (Fig. 3). Third, overexpression of Rta in permissive B cells leads to transcription from Rp and Zp, the two distinct promoters that control the IE genes BRLF1 and BZLF1 (Fig. 4).

Previous assumptions about the inability of Rta to drive lytic EBV gene expression in lymphoid cells may have been the result of the use of expression vectors that performed unequally in different cell types and yielded low or nondetectable levels of Rta protein in lymphocytes (Fig. 1A). When an appropriate expression system was used, transfection of Rta resulted in activation of both early and late genes. In HH514-16 cells, this property can be attributed in part to the ability of Rta to activate Zp and thus promote ZEBRA expression from the endogenous virus (Fig. 1B and 4). Once ZEBRA was induced in HH514-16 cells, lytic DNA replication and progression of the cycle to late gene expression were inevitable. In other cell lines, such as Raji, that are not fully permissive, transfection of BRLF1 also stimulated EA-D expression and several other lytic cycle genes (Fig. 1A and data not shown). Whether examined in fully permissive cells, such as HH514-16, or in Raji, Rta by itself was always less efficient at inducing lytic gene expression than was ZEBRA (Fig. 1B and data not shown). This finding suggests that maximal lytic EBV gene expression in B cells requires the joint and interactive contributions of Rta and ZEBRA.

In reporter-based transient assays Rta specifically and powerfully activated its own promoter, an autostimulatory activity that was maximal in EBV-positive cells (Fig. 3B). The EBV genome could contribute to Rp activation in one of several ways. Rp does not contain known binding sites for Rta (20); instead, it contains binding sites for cellular proteins including YY1, Sp1, and Zif268 (55, 56, 58). An EBV gene might induce or modify one or more cellular proteins enabling it to bind to Rp by an indirect mechanism. Alternatively, a latent or lytic EBV gene product that is induced by Rta may mediate the interaction between Rta and Rp. The identification of proteins that mediate Rta/Rp interactions remains a key unanswered problem.

By contrast to Rta, ZEBRA was relatively inefficient at activating Rp reporter constructs in EBV-positive cells and totally inactive at stimulating Zp reporters in this background (Fig. 3C). These findings raise the possibility that viral factors or virally induced cellular factors act at the posttranslational level to repress ZEBRA-mediated activity. ZEBRA is known to interact with a number of cellular factors that could provide this function (21, 24, 39, 51, 59, 60). In agreement with previous reports, ZEBRA was capable of stimulating these two promoters in an EBV-negative background (15). Nonetheless, transfection of ZEBRA into EBV-positive HH514-16 cells led to appearance of mRNAs initiated from Rp and Zp. These results suggest that in the context of latent EBV, ZEBRA may not autostimulate its own promoter directly but may do so indirectly either by activation of Rta or by contacting the Zp promoter through cellular factors.

In accordance with this idea, previous studies have shown that transfection of ZEBRA into Raji cells leads to the appearance of the 3.0-kb bicistronic mRNA driven from Rp but not the 1.0-kb monocistronic mRNA from Zp (29, 30). These previous studies using Raji cells suggested that ZEBRA’s capacity to stimulate Rp, rather than its ability to autostimulate Zp, was a pivotal event in disruption of latency. However, the pathway of activation of the endogenous virus following transfection of ZEBRA and Rta appears to differ among cell lines. The present experiments in HH514-16 cells show that Rta and ZEBRA each stimulate both Zp and Rp (Fig. 4). The experiments do not allow us to conclude whether each overexpressed protein stimulates both promoters, its own promoter, or the promoter of its partner IE gene. Nonetheless, the experiments clearly demonstrate that in permissive cells, both Zp and Rp become active as the result of transfection of either IE gene and suggest a mutual cross-stimulatory mechanism.

The experiments lead to the model proposed in Fig. 5, which should be valid for permissive cells such as HH514-16. During latency, both Rp and Zp are tightly repressed by cellular factors, by higher-order chromatin structure, and/or by the absence of specific activators. Following an inducing stimulus, some cellular factors or structures are displaced or modified by newly synthesized or modified cellular activators. On Zp, these may include AP-1 family members as well as other, unidentified proteins (6, 16, 29, 30). On Rp, Zif268, Sp1, or other proteins may be involved (56, 58). During an amplification or autostimulation event, Rta and ZEBRA activate each other’s promoter. Each of these proteins may also stimulate its own promoter; however, this capacity may be cell specific since in Raji cells Zp is not activated by ZEBRA (29, 30). In permissive cells, Rta activates both Rp and Zp in conjunction with an unidentified protein whose activity is enhanced in the presence of an EBV genome.