Viral Genome Methylation Differentially Affects the Ability of BZLF1 versus BRLF1 To Activate Epstein-Barr Virus Lytic Gene Expression and Viral Replication

Cell lines and culture.

HEK 293T, HeLa, and D98/HR-1 cells were maintained in Dulbecco modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS), and 1% penicillin-streptomycin (pen-strep). HONE-1 (a gift from Ron Glaser, Ohio State University) is an early-passage EBV-negative NPC cell line and was maintained in RPMI 1640 supplemented with 10% FBS and 1% pen-strep. The NOKs cell line (a gift from Karl Munger, Harvard University) is a telomerase-immortalized normal oral epithelial keratinocyte cell line that was derived as previously described (51) and was maintained in keratinocyte-SFM (Life Technologies, Inc.) supplemented with epidermal growth factor, bovine pituitary extract, and 1% pen-strep (K-SFM). Akata-GFP Burkitt lymphoma (BL) cells (a gift from Kenzo Takada [received from Bill Sugden]) were maintained in RPMI 1640 supplemented with 10% FBS, 1% pen-strep, and 500 μg/ml G418. Akata-GFP BL cells are derived from Akata BL, a type I latency Burkitt lymphoma line, that lost the endogenous EBV genome and then was superinfected with an Akata EBV containing inserted green fluorescent protein (GFP) and G418 resistance genes as previously described (52). HONE-Akata cells are derived from HONE cells stably infected with EBV produced from the Akata-GFP BL line and were maintained in RPMI 1640 supplemented with 10% FBS, 1% pen-strep, and G418 (400 μg/ml). NOKs-Akata cells were derived from the EBV-negative NOKs cell line. Briefly, 10⁵ NOKs cells were plated in a 6-well dish and cocultured with 2 × 10⁶ Akata-GFP BL cells in 2 ml of K-SFM for 24 h. Akata-GFP BL cells were removed by washing with phosphate-buffered saline (PBS), and EBV-positive NOKs cells were selected by adding 50 μg/ml of G418 to the medium starting 1 week after infection. The NOKs-Akata cell line used in these studies had been infected with EBV approximately 6 months prior to experimentation and maintained in G418 since the time of infection.

Plasmids and cloning.

Plasmid DNA was prepared using the Qiagen Midi/Maxiprep kit according to the manufacturer's instructions. pSG5 was obtained from Stratagene. The SG5-BRLF1 expression vector (a gift from S. D. Hayward, Johns Hopkins University) was constructed as previously described (17) and carries the BRLF1 open reading frame under the control of the simian virus 40 (SV40) early promoter. SG5-BRLF1 aa1-550 (R550) was constructed using the Stratagene QuikChange II XL site-directed mutagenesis kit and the following primer set: BRLF1(aa1-550) forward (5′-CCCCTCGTGGCCATTTGTAGGAACTGACCACAACACTAGAGTCC-3′) and reverse (5′-GGACTCTAGTGTTGTGGTCAGTTCCTACAAATGGCCACGAGGGG-3′). SG5-BZLF1 was a gift from Diane Hayward, Johns Hopkins University, and contains the BZLF1 genomic sequence under the control of the SV40 promoter (47). Flag-tagged-BZLF1 contains BZLF1 cDNA inserted into a p3XFLAG-myc-CMV24 vector (Sigma) for mammalian cell expression (a gift from Paul Lieberman, Wistar Institute). The promoterless luciferase reporter gene construct pCpGL-basic (a gift from Micheal Rehli, Universitätsklinikum Regensburg) was constructed as previously described (53) and contains no CpG motifs in the entire vector. Various EBV promoters (Table 1) were PCR amplified from the EBV B95.8 genome and cloned upstream of the luciferase gene in pCpGL-basic using the SpeI and BglII restriction sites. The following promoters were cloned into pCpGL-basic (the position in the EBV genome is in parentheses): BALF2p (164776 to 165375), BARF1p (164825 to 165503), BFLF2p (56948 to 57548), BGLF4p (123619 to 124322), BGLF5p (122355 to 122966), BLRF2p (88203 to 88895), BMLF1p (84311 to 84922), BMRF1p (79317 to 79886), BRLF1p (106144 to 107250), and BRRF1p (104447 to 105161). The BHLF1p and BHRF1p-luciferase reporter gene constructs were constructed by PCR amplifying the divergent BHLF1 and BHRF1 promoter sequences (52781 to 53797) with the primer set 5′-CCCCAGATCTCGACGCTGGCGAGCCGGGCC-3′ and 5′-CCCCAGATCTGTGATGAAACAGGCAACTCC-3′ within the oriLyt region of EBV B95.8 genomic DNA and were inserted upstream of the luciferase gene in pCpGL-basic using the BglII restriction site.

EBV bacmid preparation.

The B95.8 bacmid (p2089) contains the EBV B95.8 genome as well as the hygromycin resistance and GFP genes on an Escherichia coli F-factor-based plasmid as previously described (a gift from Henri-Jacques Delecluse) (54). The Akata bacmid (AK-BAC) contains the EBV Akata genome in addition to the GFP gene and chloramphenicol resistance gene, as previously described (52) (a gift from Kenzo Takada, Hokkaido University, via Clare Sample at Pennsylvania State University College of Medicine). EBV bacmid DNA was isolated from 2.5-liter bacterial cultures by alkaline lysis and purified with CsCl₂-ethidium bromide gradients (55).

In vitro methylation of reporter gene constructs and bacmid DNA.

Reporter gene constructs and EBV bacmids were methylated in vitro using CpG methyltransferase M.SssI (NEB) according the manufacturer's instructions. EBV bacmid DNA was methylated using the large-scale methylation protocol. After completion of the methylation reaction, the DNA was cleaned by phenol chloroform extraction and salt precipitation. Successful methylation was confirmed by enzymatic digestion with two restriction enzymes (NEB) that recognize the same cut site: HpaII (digestion is blocked by methylation) and MspI (cuts regardless of methylation status).

Transient transfection.

HONE-1, HEK 293T, NOKs-Akata, and D98/HR-1 cells were transfected with Lipofectamine 2000 transfection reagent (Invitrogen) according to the manufacturer's instructions. HeLa cells were transfected using FuGENE6 transfection reagent (Roche) according to the manufacturer's instructions.

Reporter gene assays.

HONE-1 cells were transfected in 12-well dishes with 50 ng of pCpGL-basic promoter constructs, 10 ng of BZLF1 alone, 10 ng of BRLF1 alone, 5 ng of both BZLF1 and BRLF1 (synergy studies), and up to 500 ng of SG5 control expression vectors. The cells were washed with PBS and harvested in 1× Reporter lysis buffer (Promega) at 48 h posttransfection. Lysates were subjected to one freeze-thaw cycle, and relative luciferase units were quantified with a BD Monolight 3010 luminometer (BD Biosciences) using Promega luciferase assay reagent. For each condition, at least 3 independent experiments were performed in duplicate.

Immunoblotting.

Immunoblotting was performed as previously described (27). Cells lysates were harvested in Sumo lysis buffer including proteasome inhibitor cocktail (Roche), and the protein concentration was determined using the Sumo protein assay (Bio-Rad). Equal amounts of protein were resolved with 10% or 4 to 20% gradient (Bio-Rad) sodium dodecyl sulfate (SDS)-polyacrylamide gels and transferred to nitrocellulose membranes. Membranes were first blocked in a phosphate-buffered saline solution containing 5% milk and 0.1% Tween 20 and then incubated with primary antibody. The following antibodies were used: anti-β-actin (Sigma; 1:5,000), anti-BMRF1 (Vector; 1:250), anti-BRLF1 (Argene; 1:250), anti-BZLF1 (Santa Cruz, sc-53904; 1:250), and anti-tubulin (Sigma; 1:2000). The murine monoclonal antibody against BALF2 (1:250) was a gift from Jaap Middeldorp (VU University medical center). The secondary antibody used was horseradish peroxidase (HRP)–goat anti-mouse (Fisher Scientific; 1:5,000).

Reverse transcription-PCR (RT-PCR).

HEK 293T cells were transfected in 6-well dishes with 550 ng of methylated or mock-treated EBV bacmid DNA and cotransfected with either 225 ng of BZLF1, 100 ng of BRLF1, or SG5 control vector. NOKs-Akata cells were transfected in 6-well dishes with SG5 control vector, 100 ng BZLF1, 100 ng BRLF1, or 20 ng of BZLF1 plus 20 ng BRLF1 (synergy studies). RNA was isolated at 2 days posttransfection using the RNeasy Minikit (Qiagen). The RNA concentration was determined, equivalent amounts of RNA were DNase treated, and cDNA was made using the Improm-II reverse transcription system (Promega) according to the manufacturer's instructions. PCR using the cDNA was performed to quantify relative transcript levels of multiple EBV lytic genes with the following primers: BALF2, 5′-TCAATGTCAAGGCTCTGCACAGGA-3′ and 5′-ACCATATGGGCATTGTGGAACACG-3′; BLRF2, 5′-TGTCAGCTCCACGCAAAGTCAGAT-3′ and 5′-AGGACCTGTTGCTTCAGAGCCTTA-3′; BHLF1, 5′-ATGAGCTCCAGGACCAGGCAA-3′ and 5′-TAGGGTTCGAATGGGCGTGGT-3′; BMLF1, 5′-TCTCCCGAACTAGCAGCATTTCCT-3′ and 5′-ATCGCAGTCTGTGTTGGTGTCTGA-3′; BMRF1, 5′ 5′-GCCGCCGTGTCATTTAGAAACCTT-3′ and 5′-TGTGGTGGCTCTTGGACACCTTAT-3′; BRLF1, 5′-TGGCTTGGAAGACTTTCTGAGGCT-3′ and 5′-AATCTCCACACTCCCGGCTGTAAA-3′; BZLF1, 5′-AATGCCGGGCCAAGTTTAAGCAAC-3′ and 5′-TTGGGCACATCTGCTTCAACAGGA-3′; and beta-2 microglobulin (β2 M) cellular gene, 5′-TTCTGGCCTGGAGGGCATCC-3′ and 5′-ATCTTCAAACCTCCATGATG-3′.

EMSAs.

Electrophoretic mobility shift assays (EMSAs) were performed as previously described (33). Methylated and unmethylated probes were prepared as previously described (29) or commercially obtained from IDT. Whole-cell extracts containing BRLF1 protein (missing the inhibitory carboxy-terminal domain) were created by transfection of HeLa cells with the R550 deletion construct. Cells were harvested at 48 h posttransfection in lysis buffer (0.42 M NaCl, 20 mM HEPES [pH 7.5], 25% glycerol, 1.5 mM MgCl₂, 0.2 mM EDTA, 1 mM dithiothreitol [DTT], 1 mM phenylmethylsulfonyl fluoride [PMSF], and 1× proteasome inhibitor cocktail [Roche]) and spun down for 15 min at 10,000 rpm at 4°C. Protein concentrations were determined by the Bradford assay (Bio-Rad), and supernatants were stored at −80°C. The whole-cell extracts were used to perform EMSAs with known and novel BRLF1-responsive elements (RREs). The BMLF1 promoter probe spanned positions 84691 to 84728 (5′-GGCCCAGATGTCCCTCTATCATGGCGCAGACATTCTC-3′). BALF2 probe 1 spanned positions 165056 to 165093 (5′-GCACAGCACCACCCTGAGCCGCGACCAGTAGTCGTAG-3′), and BALF2 probe 2 spanned positions 165266 to 165303 (5′-CCGGGTGAACACCGCGTACATGGCCCTGAACATGAGG-3′). The BLRF2 promoter probe spanned positions 88638 to 88675 (5′-GCGCTTCCAGTCCCACAAACGCGGCGGCGGCTTCCCT-3′). The underlined portion of the probes contains the site of the RRE, and the bold nucleotides are CpGs that were methylated or not. Double-stranded, annealed DNA oligonucleotides were labeled with [γ-³²P]ATP (Perkin-Elmer) using T4 polynucleotide kinase (NEB) and desalted with G-25 Sephadex columns (GE Healthcare). Whole-cell extracts (1 μg for BMLF1 reactions and 15 μg for BALF2 and BLRF2 reactions) were incubated for 5 min in binding buffer containing 10 mM HEPES, (pH 7.5), 50 mM NaCl, 2 mM MgCl₂, 2.5 μM ZnSO₄, 0.5 M EDTA, 1 mM DTT, 15% glycerol, and 0.5 μg poly(dI-dC), and 30,000 cpm of labeled oligonucleotide was added to the reaction mixture and allowed to incubate for 20 min at room temperature. For supershift reactions, 1 μl of anti-BRLF1 (Argene) was added, and reactions were allowed to incubate for an additional 20 min at room temperature. The reaction mixtures were loaded onto a 4% polyacrylamide gel in 0.5× Tris-borate-EDTA buffer and electrophoresed at 35 mA. Gels were dried on Whatman paper under a vacuum and exposed to autoradiography film for 12 to 40 h at −80°C.

Chromatin immunoprecipitation (ChIP) assays.

HEK 293T cells were transfected in 10-cm dishes with 200 ng methylated or mock-treated EBV bacmid DNA, 1 μg BZLF1, 1 μg BRLF1, or SG5 control vector (up to 6 μg). Cells were cross-linked for 10 min at room temperature with fresh 1% paraformaldehyde at 48 h posttransfection. The cross-linking reaction was quenched using 125 mM glycine, and the cells were lysed. The lysate was sonicated to yield approximately 500-bp DNA fragments. DNA-protein complexes were immunoprecipitated with the following antibodies: anti-BRLF1 (Argene), anti-BZLF1 (Argene), anti-FLAG (Sigma; F1804), anti-acetyl H3K9 (Abcam), mouse isotype anti-IgG control (Santa Cruz), and rabbit isotype anti-IgG control (Santa Cruz). Immunoprecipitated DNA-protein complexes were washed with low-salt, high-salt, lithium chloride, and Tris-EDTA (TE) wash buffers. The protein-DNA cross-links were reversed at 65°C overnight, and the DNA was purified using the Qiagen gel extraction kit. PCR was used to determine the presence and relative amount of specific DNA fragments that were immunoprecipitated. Primers used for amplifying the for the BALF2 promoter were 5′-AAACACCACTGTGTAGCACAGCAC-3′ and 5′-TGAGTCCAGCTACCTCATGTTCAG-3′, those for the BHLF1 promoter were 5′-CTCTTTTTGGGGTCTCTGTG-3′ and 5′-CCTCCTCCTCTCGTTATCC-3′ (56), those for BLRF2 were 5′-ACTGAAGCCCAGGACCAGTTCTA-3′ and 5′-TAAGACAAGCGTCAGAAGTGCCCA-3′, those for BMLF1 were 5′-CGTGACATGGAGAAACTGGGGG-3′ and 5′-CCTCTTACATCACTCACTGCACG-3′, those for the BMRF1 promoter were 5′-ATGCCCAGAAACCTGAGCAAGTAGCC-3′ and 5′-CCTTGGTGGATGTGCGAGCCATAAAG-3′, those for BRLF1 were 5′-CTCTTACCTGCGTCTGTTTGTG-3′ and 5′-CTCTCTGCTGCCCACTCATACT-3′, those for BZLF1 were 5′-GGTGTAAATTTTACATCTTC-3′ and 5′-GCTAATGTACCTCATAGACACACC-3′, and those for β-globin were 5′-AGGGCTGGGCATAAAAGTCA-3′ and 5′-GCCTCACCACCAACTTCATC-3′.

qPCR.

Quantification of ChIP samples was performed by quantitative PCR (qPCR) analysis using SYBR green (Bio-Rad) according to the manufacturer's protocol. Samples were measured with an ABI Prism 7900 real-time PCR system (Applied Biosystems). BMRF1 was amplified with primers 5′-CACTGCGGTGGAGGTAGAG-3′ and 5′-GGTGGTGTGCCATACAAGG-3′ (56). Input samples were diluted to 5%, 1%, and 0.2% into H₂O with 100 μg/ml sonicated salmon sperm DNA (Agilent). A standard curve was calculated from the threshold cycle (C_T) of the input sample dilution series and used to calculate percent input bound in the tested samples. Each condition and input dilution was loaded in triplicate.

OriLyt plasmid-based replication assays.

OriLyt plasmid replication assays were performed as previously described (44). Latently infected, EBV-positive D98/HR-1 cells were transfected with 500 ng of the oriLyt-containing plasmid p588 (57) (a gift from Bill Sugden, University of Wisconsin-Madison), 1 μg of BZLF1, or SG5 control vector (up to 6 μg) in 10-cm dishes. Cells were harvested at various time points posttransfection, and nuclei were isolated using a modified REAP method (58). Briefly, cells were resuspended twice in a solution of cold PBS plus 0.1% NP-40. DNA was isolated from the nuclear pellets using the DNeasy blood and tissue kit (Qiagen), concentrated with salt precipitation, and quantified using spectrophotometry. Equivalent amounts (4 to 6 μg) of DNA were digested overnight with 2 μl of restriction enzymes (BamHI and DpnI) and spiked with an additional 1 μl of restriction enzymes. The DNA was separated on 0.8% agarose gels at 25 V for 16 h. The gel was prepared for transfer by incubation with 0.25 N HCl for 30 min, denaturing buffer (0.5 M NaOH, 1.5 M NaCl) for 30 min, neutralizing buffer (0.5 M Tris-HCl, 1.5 M NaCl, [pH 7.0]) for 30 min, and 20× SSC (3 M NaCl, 0.3 M sodium citrate [pH 7.0]) for 30 min. The DNA was transferred to nylon membranes overnight using the Turboblotter rapid downward-transfer system (Whatman) and cross-linked with UV irradiation. Membranes were prehybridized in Church hybridization buffer (0.5 M Na₂HPO₄ [pH 7.2], 1% bovine serum albumin, 7% sodium dodecyl sulfate [SDS], 5 mM EDTA [pH 8.0]) for 1 h at 65°C. Membranes were then hybridized at 65°C overnight with a DNA probe directed against the hygromycin resistance gene labeled with [γ-³²P]ATP using the random primer labeling system (GE Healthcare). After hybridization, membranes were washed with Church wash buffer (1% SDS, 20 mM Na₂HPO₄ [pH 7.2], 1 mM EDTA) one time at 65°C (15 min) and three times at 45°C (10 min for each wash). The membrane was exposed to film at −80°C overnight, and films were developed.

Methylation status of selected EBV promoters.

The methylation status of various EBV promoters in HONE-Akata cells and NOKs-Akata cells was determined. Cells were treated for 3 days with 100 μg/ml of acyclovir (Sigma) prior to DNA extraction. Genomic DNA was prepared using the Qiagen DNeasy blood and tissue kit. Two hundred nanograms of genomic DNA and 20 ng of methylated or mock-treated bacterial artificial chromosome (BAC) DNA (control) were digested with HpaII and then assayed by PCR amplification using the following primers: BALF2, 5′-GCGACTAGTTGTTTGTGAGGACCCCGGTCGAGGCGT-3′ and 5′-CTGAGATCTCCAAGGTATCGCCCCGGCCTCCCAGT-3′; BHLF1, 5′-GAGACTAGTGGAGACCTGCATCTGCACACC-3′ and 5′-CTGTGTAATACTTTAAGGTTTGCTCAGGAG-3′; BLRF2, 5′-GCAACTAGTCGCTGATTCTGGAGGATTAGCC-3′ and 5′-GACAGATCTCAAACAGCCGAGATTGCTGCC-3′; BMLF1, 5′-GCGACTAGTTGCGCCTCTTTGTCTGTCATCCGGAA-3′ and 5′-CAGAGATCTTAGCTGGGATGTAGTGCTGTCTTGACTGGC-3′; BRLF1, 5′-AATAGATCTTGAGGTGTTGTGTCCTGTATGGTATTC-3′ and 5′-CTGACTAGTCCCAACACCATGGGTGATAACGTC-3′; and BZLF1, 5′-GCGACTAGTAGGTGTGTCAGCCAAAGAGGATCA-3′ and 5′-GCGAGATCTCCGGCAAGGTGCAATGTTTAGTGA-3′. The BMRF1 promoter does not contain an HpaII site and hence could not be assessed by this assay.

Virus titration assay.

Virus titration assays were performed in NOKs-Akata cells as previously described (59). NOKs-Akata cells were plated onto a 12-well dish and then transfected with control SG5 vector, 50 ng of BZLF1, 50 ng of BRLF1, or 10 ng of BZLF1 plus 10 ng of BRLF1 expression vectors (for synergy studies). Supernatant was harvested at 48 h posttransfection and filtered through a 0.8-μm-pore-size filter. Raji cells (2 × 10⁵ cells/infection) were infected with 100 μl of supernatant and incubated at 37°C. Phorbol-12-myristate-13-acetate (TPA) (20 ng/ml) and sodium butyrate (3 mM final concentration) were added 24 h after infection. GFP-positive Raji cells were counted at 48 h postinfection to determine the viral titer.

Methylation enhances BZLF1-mediated activation of many, but not all, early lytic promoters.

To examine the effect of promoter methylation on the ability of BZLF1 to activate various different early lytic EBV promoters, we cloned multiple different lytic promoters (Table 1) upstream of the luciferase gene in a CpG-free vector (53) and then methylated or mock treated the various promoter constructs in vitro as previously described using the CpG methyltransferase M.SssI (29). The CpG-free luciferase vector prevents nonspecific inhibitory effects of total plasmid DNA methylation on luciferase gene activity by ensuring that only the inserted EBV promoter sequences can be methylated. EBV-negative HONE-1 NPC cells were transfected with the methylated or the mock-treated promoter constructs in the presence or absence of a limiting amount of cotransfected BZLF1 expression vector (10 ng/12-well dish), and the amount of luciferase activity for each condition was quantitated 2 days later.

As shown in Fig. 1A, methylation of promoter DNA increased the ability of BZLF1 to activate 9 out of 11 early lytic promoters tested (BALF2, BARF1, BFLF2, BGLF4, BGLF5, BMLF1, BMRF1, BRLF1, and BRRF1). We documented that similar levels of transfected BZLF1 were expressed under each condition (Fig. 1C and data not shown). Of note, while we previously reported that the BRLF1 promoter is more efficiently activated by BZLF1 in the methylated form (28), the positive effect of methylation on BZLF1 activation of this promoter is even more apparent in the current study, likely reflecting the use of the CpG-free luciferase vector, as well as the limiting amount of BZLF1 used in the current (but not the former) study. Interestingly, using this lower level of transfected BZLF1 (which we found to be similar to that expressed in transforming growth factor β [TGF-β]-treated Mutu 1 Burkitt cells [data not shown]), we did not observe autoactivation of the BZLF1 promoter in either the methylated or unmethylated form (data not shown).

The oriLyt early lytic promoters BHLF1 and BHRF1 are more efficiently activated by BZLF1 in the unmethylated form.

In contrast to the case for the majority of early lytic promoters, we also identified two early lytic EBV promoters that are activated by BZLF1 more efficiently in the unmethylated form (Fig. 1B). Interestingly, both of these promoters, BHRF1 and BHLF1, are located within the EBV lytic origin of replication (oriLyt), and in contrast to many early lytic promoters, the previously identified ZREs located upstream of the BHRF1 and BHLF1 promoters do not contain CpG motifs (10, 25, 45). These results suggest that while promoter methylation generally enhances BZLF1-mediated activation of early lytic promoters, the two oriLyt promoters are potentially important exceptions to this rule.

BRLF1 activation of early lytic promoters is inhibited by DNA methylation.

Although BRLF1 can bind directly to, and activate, many of the same early lytic EBV promoters that are activated by BZLF1, the effect of DNA methylation on BRLF1-mediated activation has not yet been explored. We therefore examined the ability of limiting levels of BRLF1 (10 ng/12-well dish, which produced a level of BRLF1 similar to that in TGF-β-treated Mutu 1 cells [data not shown]) to activate a series of methylated and mock-methylated early and late lytic EBV promoters. As shown in Fig. 2A, we found that BRLF1 activated five different early lytic promoters (BALF2, BARF1, BFLF2, BMRF1, and BRRF1) much more efficiently in the unmethylated form than in the methylated form. Four other promoters (BGLF4, BHLF1, BHRF1, and BMLF1) were also activated more efficiently in the unmethylated forms, although the inhibitory effect of methylation was not as dramatic. We documented that similar levels of transfected BRLF1 were expressed under each condition (Fig. 2C and data not shown). At the low level of transfected BRLF1 used in these studies, we did not observe BRLF1 activation of the other early lytic promoters listed in Table 1 (including the BZLF1 and BRLF1 IE promoters) in either the methylated or unmethylated form (data not shown). Thus, all BRLF1-responsive early lytic promoters tested were more efficiently activated by the BRLF1 protein in the unmethylated form than in the methylated form, although the effect of methylation was more dramatic for some early lytic promoters (such as the BALF2 promoter) than for others (such as the BMLF1 promoter).

BRLF1 activation of the BLRF2 late lytic viral promoter is also inhibited by promoter DNA methylation.

BRLF1 has also been reported to activate certain late gene viral promoters in a replication-independent manner in reporter gene assays, and it binds directly to at least two of these promoters (BLRF2 and BFRF3) (33, 37). Interestingly, the previously identified RRE in the BLRF2 promoter (GTCCCACAAACGCGGCG) contains several CpG motifs (33). Therefore, we examined how promoter DNA methylation affects the ability of BRLF1 to activate the BLRF2 promoter in the methylated versus the unmethylated form. We found that promoter DNA methylation greatly inhibits the ability of BRLF1 to turn on the BLRF2 promoter (Fig. 2B). However, we did not observe BRLF1 activation of several other late lytic viral promoters tested (including the BcLF1, BDLF3, and BLLF1 promoters) in either the unmethylated or methylated form under the conditions used in our studies (data not shown). Thus, the ability of BRLF1 to activate at least one late lytic EBV promoter, BLRF2, requires that the viral promoter be in the unmethylated form.

Viral genome methylation differentially affects the ability of BZLF1 versus BRLF1 to induce early lytic gene expression in the context of the intact viral genome.

To examine how methylation affects lytic gene expression in the context of the intact viral genome, purified EBV bacmid DNA was methylated or mock treated in vitro and then transfected into HEK 293T cells in the presence or absence of cotransfected BZLF1 or BRLF1. Immunoblotting was performed 3 days later to compare the ability of cotransfected BZLF1 versus BRLF1 to activate expression of the BMRF1 and BALF2 early lytic genes from the EBV bacmid genome in the methylated versus unmethylated state. As indicated in Fig. 3A, methylation of the EBV bacmid genome enhances BZLF1-induced BMRF1 and BALF2 protein expression, in agreement with the results of the reporter gene assays (Fig. 1). In contrast, methylation of the EBV genome decreases BRLF1-induced BMRF1 and BALF2 protein expression (Fig. 3A), as also predicted by the reporter gene assays (Fig. 2). Similar results were obtained using either the B95.8 or Akata bacmid DNA (data not shown). These results confirm that methylation differentially affects the ability of BZLF1 versus BRLF1 to induce early lytic viral protein expression in the context of the intact viral genome.

To determine if the differential effect of EBV bacmid methylation on the ability of BZLF1 versus BRLF1 to activate lytic protein expression is associated with differences in lytic viral gene transcription, we harvested RNA from 293T cells transfected with methylated or mock-methylated EBV bacmids (in the presence or absence of cotransfected BZLF1 or BRLF1) and performed RT-PCR to detect various EBV transcripts. As shown in Fig. 3B, BRLF1 preferentially activates expression of the BMRF1, BALF2, and BLRF2 transcripts from the unmethylated EBV bacmid, whereas BZLF1 preferentially activates expression of the BMRF1 and BALF2 early lytic EBV transcripts from the methylated EBV bacmid. Activation of the BLRF2 late transcript occurs only in response to BRLF1 expression, and this activation requires an unmethylated viral genome. Of note, the ability of transfected BZLF1 to induce BRLF1 gene transcription from the cotransfected EBV bacmid construct and, vice versa, the ability of BRLF1 to induce BZLF1 transcription from the cotransfected bacmid construct were very limited, for unclear reasons.

The combination of BZLF1 and BRLF1 synergistically activates early lytic BMRF1 protein expression from either the methylated or unmethylated EBV bacmid genome.

Deletion of either the BZLF1 or BRLF1 protein severely inhibits expression of the early lytic BMRF1 protein in stably EBV-infected 293 cells (16), which contain a highly methylated viral genome (29). However, the mechanism(s) by which BZLF1 and BRLF1 cooperate to synergistically activate expression of early lytic viral proteins, and in particular whether this effect is dependent upon the DNA methylation state of the lytic viral promoters, is not well understood. To examine whether the combination of BZLF1 and BRLF1 can synergistically induce early lytic BMRF1 protein expression from either the methylated or unmethylated forms of EBV bacmids, we transfected methylated or mock-treated EBV bacmid DNA into 293T cells in the presence of BZLF1 alone, BRLF1 alone, or the combination of both BZLF1 and BRLF1 and compared the amounts of BMRF1 protein expression derived from the transfected EBV bacmid DNA 3 days later. As shown in Fig. 4A, the combination of BZLF1 and BRLF1 synergistically activated expression of the BMRF1 protein from either the unmethylated (left) or methylated (right) form of the EBV genome; similar results were obtained using either B95.8 or Akata bacmid DNA.

To determine if the synergistic effect of the BZLF1/BRLF1 combination on BMRF1 protein expression is associated with an increase in BMRF1 transcription, we harvested RNA from 293T cells transfected with methylated or mock-methylated EBV bacmids (in the presence or absence of cotransfected BZLF1, BRLF1, or BZLF1 and BRLF1 together) and performed RT-PCR to examine the level of BMRF1 transcript (Fig. 4B). Somewhat surprisingly, for both the methylated and unmethylated forms of the EBV bacmid, the combination of BZLF1 and BRLF1 together resulted in only a relatively modest increase in the level of BMRF1 transcript relative to the effect of BZLF1 or BRLF1 alone, in contrast to the large effect observed at the BMRF1 protein level. These results suggest that BZLF1 and BRLF1 may cooperate to enhance BMRF1 protein expression through at least a partially (as-yet-unknown) posttranscriptional mechanism(s).

DNA methylation does not affect BRLF1 binding to RREs in vitro.

Given our finding that promoter DNA methylation decreases the ability of BRLF1 alone to activate lytic gene expression, we next asked if methylation of CpG-containing RREs inhibits BRLF1 binding. To examine the effect of RRE methylation on BRLF1 binding in vitro, we prepared extracts from HeLa cells transfected with a BRLF1 expression vector containing amino acids 1 to 550 (since it is difficult to detect BRLF1 binding activity by EMSA in cells transfected with the intact BRLF1 protein [33]) and performed EMSAs using unmethylated versus methylated RRE probes. As shown in Fig. 5A, BRLF1 binds similarly to the methylated and unmethylated forms of a CpG-containing RRE in the BMLF1 promoter. Likewise, the methylated versus unmethylated forms of two different CpG-containing RREs within the BALF2 promoter were bound similarly in vitro (Fig. 5B), even though BRLF1 activation of this promoter in vivo is much more efficient for the unmethylated form of the promoter (Fig. 2 and 3). BRLF1 binding to the methylated and unmethylated forms of a CpG-containing RRE in the late BLRF2 promoter was also similar (by EMSA) (Fig. 5C), even though BRLF1 activates this promoter much more efficiently in the unmethylated form (Fig. 2 and 3).

DNA methylation does not affect BRLF1 binding to RREs in vivo but enhances BZLF1 binding to most ZRE-containing promoters.

We next performed ChIP assays to examine the effect of viral genome methylation on BRLF1 (full length) versus BZLF1 DNA binding in vivo in 293T cells transfected with the methylated or unmethylated forms of the EBV bacmid DNA (Fig. 6). BRLF1 bound similarly to the methylated and unmethylated forms of the BALF2, BMLF1, and BMRF1 promoters in vivo (Fig. 6A), similar to the results of the EMSA studies (Fig. 5); quantitative PCR analysis of the BMRF1 promoter ChIP results (Fig. 6B) confirmed that BRLF1 binding to the methylated and unmethylated forms of this promoter is similar. Although BRLF1 clearly activates the unmethylated forms of the BMRF1 and BALF2 promoters more efficiently than the methylated forms in reporter gene assays (Fig. 2A) and in the bacmid studies (Fig. 3), these results suggest that the inhibitory effect of DNA methylation on BRLF1-mediated activation of the BALF2 and BMRF1 promoters is not due to a decreased ability of BRLF1 to bind to methylated BALF2 or BMRF1 promoter DNA.

As shown in Fig. 6C, methylation of EBV bacmid DNA promotes BZLF1 binding to multiple CpG-containing early lytic viral promoters (BMRF1, BRLF1, and BMLF1), consistent with the enhanced BZLF1-mediated activation of the methylated forms of these promoters in reporter gene assays (Fig. 1A) and bacmid studies (Fig. 3). Increased BZLF1 binding to the methylated versus unmethylated form of the BMRF1 promoter was confirmed by quantitative PCR analysis (Fig. 6D). Interestingly, although BZLF1 activates the unmethylated form of the BHLF1 promoter (which has CpG-free ZREs) more efficiently than the methylated form in reporter gene assays (Fig. 1B), it bound at least as well to the methylated, versus unmethylated, form of this promoter in vivo. Thus, the inhibitory effect of DNA methylation on BZLF1 activation of the BHLF1 promoter is not associated with reduced BZLF1 DNA binding to this promoter.

We also performed ChIP assays to examine whether BRLF1 and BZLF1 increase one another's ability to bind to several different lytic EBV promoters. As shown in Fig. 6E and F, BZLF1 and BRLF1 did not significantly increase each other's ability to bind to any of the six different lytic promoters examined in the methylated EBV bacmid. Similar results were obtained in studies examining BZLF1 and BRLF1 binding to lytic promoters in the unmethylated EBV bacmid (data not shown).

BRLF1 induces an activating histone modification (H3K9 acetylation) more efficiently on unmethylated viral promoters.

Since BRLF1 interacts directly with the histone acetyltransferase CBP (60), we hypothesized that BRLF1 binding to promoter DNA may induce the activating histone modification H3K9 acetylation. To determine if promoter DNA methylation affects the ability of BRLF1 to induce H3K9 acetylation, we performed ChIP assays (using an antibody that recognizes acetylated H3K9) in 293T cells transfected with the methylated or unmethylated forms of EBV bacmid DNA in the presence or absence of cotransfected BRLF1 or BZLF1. The results of these experiments confirmed that BRLF1 can induce H3K9 acetylation on numerous different early lytic EBV promoters (Fig. 7). Importantly, however, BRLF1 induced much more H3K9 acetylation on the unmethylated form of the EBV bacmid DNA than on the methylated form. These results indicate that while BRLF1 binds similarly to both the unmethylated and methylated forms of viral promoters, it preferentially confers H3K9 acetylation to the unmethylated forms of the promoters. Of note, BRLF1 also conferred H3K9 acetylation to the unmethylated (but not methylated) form of the BZLF1 promoter, even though it is not known to bind directly to this promoter. Interestingly, in comparison to BRLF1, binding by the BZLF1 protein to the unmethylated and methylated forms of EBV bacmid DNA induced relatively little H3K9 acetylation, even though we and others have shown that BZLF1 interacts directly with CBP and p300 (61, 62).

Methylation does not affect lytic replication of an oriLyt-containing vector in cis.

Since the BHLF1 transcript has been shown to be required in cis for efficient lytic replication and BZLF1 binding to oriLyt is essential for efficient lytic replication independent of the transcriptional function of BZLF1 (45, 47), we also studied the effect of methylation in cis on oriLyt replication. A vector containing the EBV BamHI fragment (which contains the entire EBV oriLyt), in addition to a hygromycin resistance gene, was methylated or mock treated in vitro and transfected into EBV-positive D98/HR-1 cells in the presence or absence of a BZLF1 expression vector. Nuclear DNA was harvested at various time points after transfection and digested with DpnI, and a Southern blot assay was performed using a probe directed against the hygromycin resistance gene (to avoid detection of the replicated endogenous D98/HR-1 viral genome). As previously described (44), the unreplicated oriLyt plasmid is sensitive to DpnI-mediated cutting (since plasmid DNA replicated in bacteria is dam methylated at the adenine in the GATC motif), whereas oriLyt plasmid DNA replicated by the viral DNA polymerase in human cells is not methylated at this site and is thus resistant to DpnI cutting.

As shown in Fig. 8A, the methylated and unmethylated oriLyt-containing vectors replicated similarly at all time points, in a BZLF1-dependent manner. Note that transfection of BZLF1 into D98/HR-1 cells results in strong expression of the BRLF1 protein (derived from the endogenous viral genome) (Fig. 8B), and hence both BZLF1 and BRLF1 are available in this replication assay. Since the trans-acting BZLF1-induced viral replication proteins in this experiment were all derived from the endogenous viral genome of D98/HR-1 cells, this oriLyt plasmid replication assay result indicates that DNA methylation of oriLyt does not alter the efficiency of lytic replication in cis.

BRLF1, but not BZLF1, expression results in lytic viral reactivation and release of infectious viral particles in a cell line infected with a highly unmethylated form of the EBV genome.

We have recently identified a telomerase-immortalized oral keratinocyte cell line (NOKs) that can be stably infected with EBV in a latent form and maintains the lytic viral promoters on the EBV genome in a highly unmethylated state. To examine the methylation status of the various lytic viral promoters in NOKs-Akata cells, DNA was purified from the cells and cut or mock cut with the HpaII restriction enzyme (which can cut the unmethylated, but not methylated, form of the CCGG recognition sequence), and lytic EBV promoter sequences were then PCR amplified using primers located on either side of the HpaII restriction site(s). As shown in Fig. 9A, the methylated EBV bacmid DNA was resistant to HpaII cutting (and hence could be PCR amplified when exposed to the restriction enzyme), while the unmethylated form of the bacmid was sensitive to cutting (and hence could not be PCR amplified), as expected. The EBV DNA purified from NOKs-Akata cells could not be PCR amplified following HpaII cutting at any of a variety of different lytic EBV promoters tested (including the BZLF1, BRLF1, BALF2, BHLF1, BLRF2, and BMLF1 promoters), indicating that the CpG-containing HpaII sites present in each of these promoters are not methylated. In contrast, with the exception of the BZLF1 and BHLF1 promoters, each of the lytic viral promoters in EBV DNA purified from the HONE-Akata line was partially protected from HpaII digestion, suggesting that this cell line contains a mixture of methylated and unmethylated viral genomes. Interestingly the BZLF1 promoter was recently shown to be generally unmethylated in various EBV-positive tumors, even when other lytic viral promoters were methylated (5).

We next compared the ability of transfected BRLF1 versus BZLF1 expression vectors to induce early lytic protein expression in NOKs-Akata versus HONE-Akata cells. As shown in Fig. 9B, although the NOKs-Akata cells expressed at least as much transfected BZLF1 as the HONE-Akata cells, BZLF1 activated BRLF1 and BMRF1 expression from the endogenous EBV genome in HONE-Akata cells but had no effect whatsoever in the NOKs-Akata cells. In contrast, BRLF1 activated BZLF1 and BMRF1 expression from the endogenous viral genomes in both NOKs-Akata and HONE-Akata cells (Fig. 9B and data not shown). Similar results were obtained when lytic viral gene expression was examined using RT-PCR analysis (Fig. 9C).

To determine if the combination of BRLF1 and BZLF1 induces synergistic early lytic BMRF1 protein expression in NOKs-Akata cells (as was observed using the unmethylated as well as methylated EBV bacmids), cells were transfected with control vector, BZLF1 alone, BRLF1 alone, or the combination of BZLF1 and BRLF1. As shown in Fig. 9D, the combination of BZLF1 and BRLF1 together induced much more BMRF1 protein expression than either BZLF1 or BRLF1 alone. Interestingly, NOKs-Akata cells did not show a significant increase in lytic gene transcript levels (including the BMRF1 transcript) in cells transfected with BRLF1 alone versus the combination of BRLF1 and BZLF1 (Fig. 9C). This result is similar to that obtained using EBV bacmids (Fig. 4) and again suggests that the BRLF1/BZLF1 combination synergistically enhances BMRF1 protein expression (and perhaps other lytic viral proteins as well) through an at least partially posttranscriptional mechanism.

Finally, we also examined the amount of infectious viral particles released (using the Green Raji cell assay) from NOKs-Akata cells transfected with control vector, BZLF1 alone, BRLF1 alone, or the combination of BZLF1 and BRLF1. As shown in Fig. 9E, BZLF1 alone did not result in release of infectious viral particles (in comparison to cells transfected with a control vector), while BRLF1 alone induced release of infectious viral particles. However, the combination of BZLF1 and BRLF1 together resulted in the greatest number of infectious viral particles, consistent with the ability of this combination to increase expression of the essential viral replication protein BMRF1 (the viral DNA polymerase processivity factor). These results confirm that BRLF1 plays a critical and primary role in initiating lytic gene expression in cells containing the unmethylated form of the EBV genome and show that cells infected with a highly unmethylated form of the EBV genome are capable of undergoing the lytic form of viral replication in response to BRLF1 but not BZLF1 expression.