Methylation Status of theEpstein-Barr Virus (EBV) BamHI W Latent Cycle Promoter and Promoter Activity: Analysis with Novel EBV-Positive Burkitt and Lymphoblastoid Cell Lines

0022-538X/06/$08.00⫹0 doi:10.1128/JVI.01204-06

Copyright © 2006, American Society for Microbiology. All Rights Reserved.

Methylation Status of the Epstein-Barr Virus (EBV) BamHI W Latent

Cycle Promoter and Promoter Activity: Analysis with Novel

EBV-Positive Burkitt and Lymphoblastoid Cell Lines

䌤

Isabel A. Hutchings, Rosemary J. Tierney, Gemma L. Kelly, Julianna Stylianou,

Alan B. Rickinson, and Andrew I. Bell*

Cancer Research UK Institute for Cancer Studies, The University of Birmingham, Edgbaston, Birmingham B15 2TT, United Kingdom

Received 9 June 2006/Accepted 8 August 2006

The Epstein-Barr virus (EBV) latent cycle promoter Wp, present in each tandemly arrayed copy of the BamHI W region in the EBV genome, drives expression of the EB viral nuclear antigens (EBNAs) at the initiation of virus-induced B-cell transformation. Thereafter, an alternative EBNA promoter, Cp, becomes dominant, Wp activity declines dramatically, and bisulfite sequencing of EBV-transformed lymphoblastoid cell lines (LCLs) shows extensive Wp methylation. Despite this, Wp is never completely silenced in LCLs. Here, using a combination of bisulfite sequencing and methylation-specific PCR, we show that in standard LCLs transformed with wild-type EBV isolates, some Wp copies always remain unmethylated, and in LCLs trans-formed with a recombinant EBV carrying just two BamHI W copies, Wp is completely unmethylated. Further-more, we have analyzed rare LCLs, recently established using wild-type EBV isolates, and rare Burkitt lymphoma (BL) cell clones, recently established from tumors carrying EBNA2-deleted EBV genomes, which express EBNAs exclusively from Wp-initiated transcripts. Here, in sharp contrast to standard LCL and BL lines, all resident copies of Wp appear to be predominantly hypomethylated. Thus, studies of B cells with atypical patterns of Wp usage emphasize the strong correlation between the presence of unmethylated Wp sequences and promoter activity.

Epstein-Barr virus (EBV), a B-lymphotropic herpesvirus im-plicated in the pathogenesis of several human malignancies, efficiently transforms resting B cells in vitro into permanently proliferating lymphoblastoid cell lines (LCLs). Such LCLs all display a latency III form of infection characterized by the constitutive expression of six EBV-encoded nuclear antigens, EBNA1, -2, -3A, -3B, -3C, and -LP, and three latent membrane proteins, LMP1, -2A, and -2B (38). While the LMP genes are transcribed from their own individual promoters, all six EBNA mRNAs are generated by the splicing of long primary tran-scripts which initiate from one of two alternative promoters, Wp or Cp (9, 10, 41, 47). Wp, present in each BamHI W repeat of the EBV genome, is selectively activated immediately postinfection (3, 56). While these Wp-initiated transcripts can potentially encode all six EBNAs, at these early time points there appears to be a preferential expression of EBNA2 and EBNA-LP; subsequently, these two antigens activate the alter-native EBNA promoter, Cp (39, 48, 54), leading to the broad-ening of virus antigen expression to all six EBNAs, and up-regulate the expression of the LMP promoters (1, 17, 53, 62). The early stages of B-cell transformation are, therefore, char-acterized by a marked switch in EBNA promoter usage, with Cp becoming dominant over Wp, leading to the outgrowth of Cp-using LCLs (43, 55, 56).

While there has been some progress in identifying the

cel-lular factors important for the initial activation of Wp in rest-ing B cells (8, 28, 50), how Wp is subsequently repressed remains poorly understood. This is an important question, however, since the flexibility of latent promoter usage is central to the virus’ strategy for persistence in vivo (49), and the Wp-to-Cp switch provides a rare opportunity in which changes in promoter usage can be followed in vitro in real time. The first clue that DNA methylation may play a role in the initia-tion or in the subsequent maintenance of Wp silencing came from earlier studies of EBV-positive Burkitt lymphoma (BL) cell lines displaying a restricted latency I form of infection (5, 16, 24, 30–32). In such cells, the Wp, Cp, and LMP promoters are all silent and hypermethylated, and only a single latent protein, the genome maintenance protein EBNA1, is ex-pressed from an alternative promoter, Qp (33, 42). However, to what extent Wp methylation status and Wp activity are linked remains a subject of debate.

One of the constraints in this regard is the lack of cell culture models available for analysis and, in particular, the absence of well-characterized lines in which Wp is the exclusive EBNA promoter. Here we attempt to overcome this limitation by (i) identifying rare Wp-using LCLs in which Cp, though present in the resident EBV episomes, is silent; and (ii) studying recently isolated Wp-restricted BL cell lines in which Wp, rather than Qp, is active and leads to the expression of EBNA1, -3A, -3B, -3C, and -LP in the continued absence of EBNA2, LMP1, and LMP2 (25). A second constraint is the difficulty of analyzing Wp sequence methylation exhaustively by the usual methods of PCR analysis and bisulfite sequencing (20). Very large num-bers of sequences need to be analyzed in this way in order to gain a representative picture of promoter methylation since, * Corresponding author. Mailing address: Cancer Research UK

In-stitute for Cancer Studies, The University of Birmingham, Edgbaston, Birmingham B15 2TT, United Kingdom. Phone: 44 121 414 4495. Fax: 44 121 414 4486. E-mail: a.i.bell@bham.ac.uk.

䌤_{Published ahead of print on 18 August 2006.}

10700

on November 8, 2019 by guest

http://jvi.asm.org/

within most latently infected cells, there are multiple copies of the viral genome and multiple copies of Wp within each ge-nome (4, 7). Here we attempt to overcome this limitation by (i) establishing LCLs using a recombinant virus with only 2 copies of Wp per genome, (ii) monitoring EBV genome load in all the lines analyzed, and (iii) developing a sensitive methylation-specific PCR assay (22) that more accurately reflects the full range of methylated and unmethylated Wp sequences present in any cell line.

MATERIALS AND METHODS

Cell lines.The panel of LCLs carrying natural EBV isolates included lines established by spontaneous transformation from EBV-infected donors (IM53.1 to IM107.1 and EBH41.2) or by virus infection of EBV-naı¨ve donor B cells in vitro (CD⫹Oku). LCLs transformed by recombinant EBVs are described below. All LCLs were maintained in standard medium (RPMI 1640 medium [Gibco] containing 10% [vol/vol] selected fetal calf serum, 2 mM glutamine, and 100 mg/liter gentamicin). The panel of BL lines and clones included Akata-BL, Rael-BL, and Awia-BL clones 9 and 20 (latency I); Ava-BL clone 1, Oku-BL clone 1, Sal-BL clone 1, and Awia-BL clones 3 and 4 (Wp restricted); and Awia-BL clones 1 and 2 (EBNA2⫹LMP1⫺) (25, 26). All BL cells were main-tained in standard medium supplemented with 1 mM pyruvate, 50␮M␣ -thio-glycerol, and 20 nM bathocupronine disulfonic acid. Before cells were harvested for EBV genome load determination and methylation analysis, all lines were grown for at least 2 weeks in the presence of 200␮M acyclovir to prevent lytic EBV DNA replication.

Preparation and use of recombinant EBVs. The recombinant EBV B95.8 strain genome from which the immediate-early BZLF1 gene was deleted has been described previously (18). Recombinant EBVs carrying different numbers of BamHI W repeats were made using the same technique (R. Tierney, unpub-lished data). Briefly, a vector was designed that contained a BamHI C-derived 5⬘ flanking region and a BamHI Y-derived 3⬘flanking region into which preligated BamHI W fragments were inserted and then introduced into the EBV bacterial artificial chromosome 2089 (13) by homologous recombination. Clones were screened to determine the numbers of BamHI W repeats present, and recom-binant genomes with 2, 4, 6, 8, and 11 Wp copies were selected. Genomes were transfected into 293 producer cells, virus preparations were generated, and the EBV genome content was assayed as described previously (44).

Peripheral blood mononuclear cells were prepared from buffy coat samples (Blood Transfusion Service, Birmingham, United Kingdom), and B cells were isolated by positive selection using CD19 Dynabeads (Dynal). Resting B cells were exposed to virus overnight at 37°C at a multiplicity of infection of 50. Following infection, B cells were cultured in standard medium.

Quantitative PCR assays for EBV gene expression and genome load.Total RNA was extracted from 2⫻106

to 5⫻106

cells using a Nucleospin RNA extraction kit (Macherey-Nagel) according to the manufacturer’s instructions. Four hundred nanograms of RNA was reverse transcribed into cDNA by using a mix of primers specific for numerous EBV transcripts, as described previously (8a). Quantitative reverse transcription (RT)-PCR assays to detect Wp- and Cp-initiated transcripts, EBNA2 transcripts, and BamHI Q-U-K-spliced EBNA1 and BamHI Y3-U-K-spliced EBNA1 transcripts were performed. EBV tran-scripts were normalized to cellular GAPDH (glyceraldehyde-3-phosphate dehy-drogenase) transcripts and expressed relative to an appropriate reference cell line, assigned an arbitrary value of 1. Reference cell lines included X50-7 (Wp-initiated transcripts), CD⫹Oku (Cp-initiated, EBNA2, and BamHI Y3 -U-K-spliced EBNA1 transcripts), and Rael-BL (BamHI Q-U-K--U-K-spliced EBNA1 tran-scripts). To determine EBV genome load, genomic DNA was extracted using standard methods and assayed by quantitative PCR amplification of the EBV DNA polymerase (BALF5) gene in parallel with the cellular beta 2-microglobu-lin gene, as described previously (26).

Western blot analysis of EBV latent proteins.Immunoblotting was performed using monoclonal antibodies 1H4 (anti-EBNA1) (21), JF186 (anti-EBNA-LP) (19), PE2 (anti-EBNA2) (61), and CS1 to CS4 (anti-LMP1) (40).

Bisulfite genomic sequencing.Genomic DNA was treated with sodium bisul-fite, and for each sample, 2- to 5-␮l aliquots of bisulfite-modified and unmodified DNA were amplified in strand-specific PCRs using primers specific for the regulatory regions of Cp as described previously (51) or for the regulatory region of Wp, as follows. Unmodified Wp DNA was amplified in nested PCRs using the following primers and conditions: Wp outer1 (5⬘-CCCCCAAACTTTGTCCAG ATG-3⬘; B95.8 coordinates 13796 to 13816) and Wp outer2 (5⬘-TGGAGTGTT

GGGCTTAGCAG-3⬘; B95.8 coordinates 14660 to 14641) amplified for 30 cycles of 95°C for 30 s, 59°C for 60 s, and 72°C for 90 s; followed by Wp inner1 (5⬘-CCTGTCACCAGGCCTGCCA-3⬘; B95.8 coordinates 13918 to 13936) and Wp inner2 (5⬘-GGGGAAAAGTTAGAAACT-3⬘; B95.8 coordinates 14485 to 14469) amplified for 30 cycles of 95°C for 30 s, 42°C for 60 s, and 72°C for 60 s. Bisulfite-treated Wp DNA was amplified in nested PCRs using the following primers and conditions: Wp outer3 (5⬘-TTTTTAAATTTTGTTTAGATG-3⬘; B95.8 coordinates 13796 to 13816) and Wp outer4 (5⬘-TAAAATATTAAACTT AACAA-3⬘; B95.8 coordinates 14660 to 14641) amplified for 40 cycles of 95°C for 30 s, 45°C for 60 s, and 72°C for 90 s; followed by Wp inner3 (5⬘-TTTGTT ATTAGGTTTGTTA-3⬘; B95.8 coordinates 13918 to 13936) and Wp inner4 (5⬘-AAAAAAAAATTAAAAACT-3⬘; B95.8 coordinates 14485 to 14469) am-plified for 40 cycles of 95°C for 30 s, 38°C for 60 s, and 72°C for 60 s. PCR products were gel purified, cloned, and sequenced as described previously (51). MSP.Methylation-specific PCR (MSP) was used to determine Wp promoter methylation status from bisulfite-treated DNA (22). PCR primer pairs were designed with regions with several CpG sites specific for either methylated or unmethylated Wp DNA. Unmethylated bisulfite-treated DNA was amplified using the following primers and conditions: Wpu1 (5⬘-TATGTGTGTATAATG GTGGAT-3⬘; B95.8 coordinates 14100 to 14120) and Wpu2 (5⬘-TAACTTACA TAAACACACTAAACT-3⬘; B95.8 coordinates 14305 to 14282) amplified through 30 cycles of 95°C for 30 s, 58°C for 15 s and 72°C for 30 s. Methylated bisulfite-treated DNA was amplified using the following primers and conditions: Wpm1 (5⬘-TTTACGCGCGTATAATGGCGGATTT-3⬘; B95.8 coordinates 14098 to 14122) and Wpm2 (5⬘ -TAACTTACGTAAACGCGCTAAACTAAA-3⬘; B95.8 coordinates 14305 to 14278) amplified through 30 cycles of 95°C for 30 s, 60°C for 30 s and 72°C for 30 s. In each case, 100 ng bisulfite-treated DNA was added to the reaction mixture in a 25-␮l volume. Akata-BL was routinely used as a positive control for methylated Wp DNA, while B cells harvested 1 day after infection with EBV were used as the source for unmethylated Wp se-quences. Five to ten microliters of PCR product was analyzed on a nondenatur-ing 8% polyacrylamide gel, stained with ethidium bromide, and directly visual-ized under UV illumination.

RESULTS

Analysis of Wp and Cp during B-cell transformation.The Wp and Cp promoters are shown in their relative genomic positions in Fig. 1A, along with the downstream exon struc-tures of the individual EBNA mRNAs expressed from both promoters. Note that Wp (and the first two exons of the EBNA-LP mRNA) lie within a BamHI W fragment that is tandemly repeated in the viral genome; thus, all natural EBV isolates contain multiple copies of Wp and express EBNA-LP species with multiple copies of a repeat domain.

The first set of experiments sought to reexamine the kinetics of Wp and Cp activation over time in freshly infected B-lym-phocyte cultures, taking advantage of newly developed quan-titative RT-PCR assays for Wp- and Cp-initiated EBNA tran-scripts. Here we used a recombinant B95.8 virus (with 11 Wp copies), rendered incapable of lytic virus replication by dele-tion of the BZLF1 immediate-early gene (18), in order to ensure that all viral DNA analyzed in emerging LCLs was episomal and not contaminated with newly replicated progeny virions. In repeated experiments, we found that Wp was acti-vated to high levels within 24 to 48 h of infection and then gradually declined over the following 14 days, after which time it remained stable at a low level. However, as illustrated by the data from one such experiment (Fig. 1B), residual Wp activity was still detectable up to day 75, by which time the resultant LCL had undergone more than 20 population doublings in vitro. In the same experiment, Cp was not detectable until 60 h postinfection but then rose quickly as Wp activity declined.

Aliquots of the infected cells were harvested at regular intervals throughout this same experiment, and the methylation status of both Wp and Cp promoters was analyzed by bisulfite sequencing.

on November 8, 2019 by guest

http://jvi.asm.org/

This analysis focuses on a 570-bp region of Wp encompassing two regulatory regions, the promoter-proximal “B-cell-specific” up-stream activation sequence 1 (UAS1⫺87 to⫺264 relative to the transcription start site) and the promoter-distal “lineage-indepen-dent” UAS2 (⫺264 to⫺352) (8). As illustrated in Fig. 1C, UAS1 contains binding sites for CREB and RFX transcription factors as well as two sites for the B-lineage-restricted BSAP protein (28, 50), whereas UAS2 contains two binding sites for the YY1 tran-scription factor (8; A. Hutchings, unpublished data). This entire region contains 20 CpG dinucleotides (Fig. 1C). Of these, two CpGs (l and m [Fig. 1]) lie within the RFX binding site but, from the evidence of in vitro binding assays, do not affect the RFX interaction when methylated (51). By contrast, there is one CpG within the CREB site (p) and two CpGs within each BSAP site (j

and k; n and o) which, if methylated, block CREB and BSAP binding, respectively, and abolish Wp activity in reporter assays (51). We confirmed that all amplifiable Wp sequences in virus preparations used to infect fresh B cells were unmethylated (data not shown). Thereafter, the methylation of Wp sequences did occur during the course of the transformation process, affecting all CpGs with the exception of site c (upstream of UAS2), sites q, r, and s (close to the transcription start site), and, in some exper-iments, sites j and k. However, such methylation was not wide-spread until day 28 (Fig. 1D) and therefore lagged significantly behind the decline in Wp transcription from its initial peak. It is important to note that Cp sequences remained entirely unmeth-ylated throughout the course of such experiments (data not shown).

FIG. 1. (A) Diagrammatic representation of the Cp- and Wp-initiated EBNA transcripts expressed in latency III LCLs. (B) Analysis of Wp and Cp activity in recently infected B cells. The graphs show the results of quantitative RT-PCR assays specific for Wp-initiated and Cp-initiated transcripts, expressed relative to an appropriate reference cell line. Results from one representative experiment for B cells harvested at 12 h and 2, 5, 8, 11, 14, 21, and 75 days postinfection are shown. Error bars indicate standard deviations for results of duplicate assays. (C) Diagram illustrating the main regulatory elements of Wp and the relative positions of the CpG dinucleotides analyzed. Shown are previously identified upstream activating sequences UAS2 and UAS1 which include binding sites for YY1, BSAP, RFX, and CREB, together with a recently identified second YY1 site between⫺270 and⫺276 relative to the transcription start site. Also marked are 20 CpG dinucleotides (black lollipop-shaped symbols, a to t) which represent potential methylated cytosines (B95.8 coordinates 13956, 13976, 14015, 14077, 14085, 14101, 14103, 14105, 14115, 14143, 14161, 14259, 14261, 14288, 14290, 14296, 14381, 14391, 14445, and 14462). CpG sites j, k, and n to p (boxed) have been shown previously to abrogate factor binding when methylated. (D) Results of bisulfite sequencing analysis of Wp in B cells 8 to 28 days postinfection. The Wp regulatory region was PCR amplified, cloned, and sequenced. Bisulfite-treated DNA was amplified with primers specific for Wp, and several PCR clones were sequenced for each sample. Individual CpG dinucleotides are identified as either methylated (⫹, shaded) or unmethylated (⫺).

on November 8, 2019 by guest

http://jvi.asm.org/

Analysis of Wp and Cp in established LCLs.We then ex-amined Wp and Cp usage in a panel of 25 LCLs carrying natural EBV isolates which had been established either by spontaneous or experimentally induced transformation in vitro. The quantitative RT-PCR data from a representative set of lines (Fig. 2) show that all but two LCLs expressed typical levels of Cp-initiated transcripts accompanied by some Wp transcription, while the two exceptions (EBH41.2 and IM53.1) showed relatively high levels of Wp but no detectable Cp activity. These two “Wp-only” LCLs were not obviously differ-ent from the Cp/Wp-using LCLs in terms of viral genome load (determined by quantitative PCR assay of acyclovir-treated cells) (Fig. 2) and showed similar levels of viral latent antigen expression and similar cell growth phenotypes (data not shown). Sequencing showed that Cp was, nevertheless, intact in these lines, at least up to 1 kb upstream of the transcription start site.

The two “Wp-only” LCLs were then compared with several standard Cp/Wp-using LCLs in bisulfite sequencing assays. The standard Cp/Wp-using LCLs all showed extensive Wp methylation by this criterion, typical for results illustrated by the IM100.1 and CD⫹Oku lines (Fig. 3). Interestingly, as seen in the earlier transformation experiments, we consistently noted the sparing of CpGs q and r immediately adjacent to the Wp transcription start site and of CpGs b and c upstream of UAS2. In addition, some, but not all, standard LCLs showed the sparing of CpGs j and k within one of the two BSAP sites in UAS1. However, the bulk of the Wp regulatory region, including CpGs n, o, and p in the BSAP/CREB site in UAS1, was clearly hypermethylated in Cp/Wp-using LCLs. By con-trast, the two “Wp-only” LCLs gave a markedly different pat-tern. In both cases, there was extensive (but not total) meth-ylation in the promoter-distal UAS2 “lineage-independent” region. However the promoter-proximal UAS1

“B-cell-spe-cific” region, containing the methylation-sensitive BSAP and CREB binding sites, was largely unmethylated (Fig. 3). One of these two “Wp-only” LCLs (IM53.1) also showed partial meth-ylation of Cp sequences, whereas the other (EBH41.2) resem-FIG. 2. Analysis of Wp and Cp transcription in established LCLs.

Histograms show the results of quantitative RT-PCR assays to mea-sure Wp-initiated and Cp-initiated transcripts. Error bars indicate standard deviations for duplicate assays. Also shown are mean EBV genome loads for the corresponding acyclovir-treated cell lines deter-mined by quantitative DNA PCR using a primer-probe combination specific for the EBV BALF5 gene.

FIG. 3. Bisulfite sequencing analysis of Wp in Wp-only LCLs and standard Cp/Wp-using LCLs. Bisulfite-treated DNA was amplified with primers specific for Wp, and several PCR clones were sequenced for each sample. Individual CpG dinucleotides are identified as either methylated (⫹, shaded) or unmethylated (⫺). EBH41.2 and IM100.1 share a se-quence polymorphism (x) such that CpG site a is not present in Wp. Shown at the top is a diagram illustrating the main regulatory elements of Wp and the relative positions of the CpG dinucleotides analyzed.

on November 8, 2019 by guest

http://jvi.asm.org/

bled all standard LCLs in that Cp was completely unmethyl-ated (data not shown).

The presence of multiple EBV episomes in all these LCLs, and the fact that all such episomes carry multiple Wp copies, means that bisulfite sequencing analysis will be biased toward the numerically dominant methylation pattern and therefore that this approach may miss functionally important minor pat-terns. For this reason, we designed an MSP assay for Wp in which bisulfite-treated DNA was amplified with two alternative sets of primers specific for methylated or unmethylated se-quences. As shown in Fig. 4A, these primers were located over CpG-containing regions (CpGs f to i in UAS2 and CpGs n to p within the BSAP/CREB site in UAS1), already known from bisulfite sequencing data (Fig. 3) as sensitive indicators of Wp methylation status. Figure 4B shows the results of such MSP assays as ethidium bromide-stained gels of PCR products from methylated- and unmethylated-sequence-specific amplifica-tions. In such assays, B cells harvested 1 day postinfection provide the positive control for unmethylated Wp sequences, and the Akata-BL cell line (using Qp, with both Wp and Cp silent) provides the positive control for methylated Wp se-quences. The assay confirms that “Wp-only” LCLs contain both methylated and significant levels of unmethylated Wp sequences. Importantly, the same assay also shows that, al-though not detected by the bisulfite sequencing assay,

un-methylated Wp sequences are present in the Cp/Wp-using LCLs.

Analysis of LCLs with reduced Wp copies.Although there is no way to reduce the EBV episomal copy number in in vitro-transformed LCLs, it is nevertheless possible to reduce the numbers of Wp copies per episome by manipulating the cloned EBV episome as a bacterial artificial chromosome. By cloning from the original 2089 B95.8 recombinant (containing 11 Wp copies), we selected recombinant EBV genomes with 2, 4, 6, 8, and 11 Wp copies, transferred them into 293 producer cells, and rescued the corresponding infectious viruses. Panels of LCLs with different numbers of Wp copies per episome were then generated from the same initial B-cell population, and the LCLs were analyzed for Wp and Cp usage. Figure 5B shows data from cell lines transformed with 2-Wp-copy (2W LCL) virus or with 11-Wp-copy (11W LCL) virus; these data are representative of six lines of each type that were studied. Both sets of LCLs showed roughly equivalent levels of Wp activity, whereas Cp activity was consistently stronger in the 2W LCLs. The overall levels of EBNA mRNA transcription, assayed here using the BamHI Y3-U-K-spliced EBNA1 mRNA that is

ex-pressed from Wp and Cp, were not markedly different between the sets of cells. Furthermore, there were no major differences in the steady-state levels of EBNA1, EBNA2, or LMP1 be-tween the LCLs detected by immunoblotting with specific monoclonal antibodies (Fig. 5C). Note that immunoblotting for EBNA-LP, a protein whose size is determined by the num-ber of BamHI W-encoded repeat domains, confirmed that the 2W LCLs did indeed express a low-molecular-weight EBNA-LP consistent with the presence of only two repeats, whereas the 11W LCLs expressed a ladder of species with one dominant isoform.

These same LCLs were then analyzed for Wp methylation by bisulfite sequencing (Fig. 6A). The 11W LCLs showed the pattern typical of extensive Wp methylation but with particular CpGs spared, as seen earlier for standard LCLs (Fig. 3). By contrast, the 2W LCLs were almost entirely unmethylated at Wp, a pattern that we have never observed before with estab-lished LCLs. Subsequently, these same LCLs and additional LCLs carrying 4-, 6-, and 8-Wp-copy virus strains were ana-lyzed using the Wp-specific MSP assay (Fig. 6B). This con-firmed that the 2W LCLs were almost entirely devoid of meth-ylated Wp sequences, whereas LCLs with 4 to 11 Wp copies per episome carried both methylated and unmethylated Wp sequences.

Analysis of BL cell lines with different patterns of EBNA promoter usage.We then turned our attention to examining the possible relationship between Wp usage and Wp methyl-ation status in a unique series of BL cell lines and derived cell clones, recently established in this laboratory (25, 26), and to displaying three different programs of restricted EBV latent gene expression (Fig. 7A). The transcriptional profiles of these BL lines, as determined by quantitative RT-PCR assays, are presented in Fig. 7B, which also includes for comparison two standard Cp/Wp-using LCLs to show how each of the BL programs described above differs from the standard latency III (LCL-like) form of infection.

First, we compared the Akata-BL and Rael-BL lines, dis-playing the classical latency I form of infection in which Wp and Cp are silent and EBNA1 is expressed selectively from the FIG. 4.(A) Design of the MSP assay used to analyze Wp

methyl-ation status. Shown is a diagram of the main regulatory elements of Wp, together with positions of the primers used in MSP analysis. (B) Results of MSP analysis of Wp methylation status in established LCLs. Bisulfite-treated DNA was amplified using primers specific for methylated (M) and unmethylated (U) Wp sequences, and the results were visualized on ethidium bromide-stained agarose gels. DNA from B cells 1 day postinfection served as a positive control for unmethylated sequences, while DNA from Akata-BL served as a positive control for methylated sequences.

on November 8, 2019 by guest

http://jvi.asm.org/

Qp promoter, with the recently described Ava-BL, Oku-BL, and Sal-BL lines displaying a “Wp-restricted” form of latency (25). These latter lines carry EBNA2-deleted genomes and show exclusive use of Wp, in the absence of Cp or Qp activity, and express EBNA1, -3A, -3B, -3C, and -LP in the absence of EBNA2 or the LMPs. Because the parental Ava-BL, Oku-BL, and Sal-BL lines also carry a silent wild-type genome in many cells, here we used clones of these lines which retain the Wp-restricted pattern of virus antigen expression but carry only EBNA2-deleted genomes (26). All of these la-tency I and Wp-restricted BL cell populations carried mul-tiple episomes (Fig. 7B).

Results from bisulfite sequencing analysis of Wp in these same lines (Fig. 8A) showed that Akata-BL (and Rael-BL [data not shown]) adheres to the previously published pattern for Qp-using latency I cells, where Wp is almost entirely meth-ylated, except for the upstream CpG c and the CpGs q and r near the transcription start site. Interestingly, in the Wp-re-stricted BL cells, bisulfite sequencing showed that the

domi-nant Wp species was again heavily methylated. However, we note that both the Ava-BL and Oku-BL clones nevertheless contained minor Wp copies that were almost entirely non-methylated. Subsequent MSP analysis (Fig. 8B) showed that there were indeed unmethylated Wp copies in all three Wp-restricted BL clones, whereas no such unmethylated sequences could be detected in the latency I Akata-BL clone.

We next analyzed a series of clones recently established from early passages of the Awia-BL tumor line. This unique line is heterogeneous at the single-cell level and yielded clones which, though all derived from the same malignant BL population and all showed the same single-cell pattern of BL growth in vitro, nevertheless display three different forms of restricted virus latency (26a). These three different latency programs are illus-trated in Fig. 7A, and the corresponding transcriptional data are shown in Fig. 7B. Some clones (9 and 20) carry multiple wild-type EBV episomes and display the classic latency I form of infection like Akata-BL and Rael-BL. Other clones (3 and 4) resemble Ava-BL, Oku-BL, and Sal-BL and show Wp-re-FIG. 5. (A) Schematic diagram of the recombinant B95.8 genome carrying either 11 BamHI W repeats (11W EBV) or 2 BamHI W repeats (2W EBV). The recombinant EBV genome also contains genes encoding hygromycin resistance (HygR_{) and green fluorescent protein (GFP). Also}

marked are the origin of plasmid replication (oriP) and terminal repeats (TR). (B) Analysis of Wp, Cp, and EBNA1 transcription in 2W and 11W LCLs. The histograms show the results of quantitative RT-PCR assays used to measure Wp-initiated, Cp-initiated, and BamHI Y3-U-K-spliced

EBNA1 transcripts. Error bars indicate standard deviations of duplicate assays. (C) Western blot analysis for expression of EBV latent antigens EBNA1, EBNA-LP, EBNA2, and LMP1 in 2W and 11W LCLs.

on November 8, 2019 by guest

http://jvi.asm.org/

FIG. 6. (A) Bisulfite sequencing analysis of Wp in 2W and 11W LCLs. Data are presented as described in the Fig. 3 legend. (B) Results of MSP analysis of Wp methylation status in 2W, 4W, 6W, 8W, and 11W LCLs. Data are presented as described in the Fig. 4 legend.

FIG. 7. (A) Diagrammatic representation of three programs of EBV latent gene expression found in different Awia-BL clones. Conventional latency I clones express EBNA1 alone from the BamHI Q promoter Qp. Atypical Wp-restricted clones carrying only the EBNA2-deleted form of the genome express EBNA1, -3A, -3B, -3C, and -LP from the BamHI W promoter Wp. Novel EBNA2⫹ LMP1⫺ clones express all six EBNAs from an unidentified promoter in the absence of the LMPs. (B) Analysis of EBV latent gene expression in BL lines and Awia-BL clones. The histograms show the results of quantitative RT-PCR assays used to measure BamHI Q-U-K-spliced EBNA1, Wp-initiated, Cp-initiated, and EBNA2 transcripts. Error bars indicate standard deviations of results of duplicate assays. Also shown are mean EBV genome loads for the corresponding acyclovir-treated cell lines determined by quantitative DNA PCR using a primer-probe combination specific for the EBV BALF5 gene. Included as controls were the standard Cp/Wp-using LCLs IM100.1 and CD_⫹Oku.

on November 8, 2019 by guest

http://jvi.asm.org/

stricted latency but in this case carry just a single EBNA2-deleted genome. A third set of Awia-BL clones (1 and 2) are unique in that they express all six EBNA proteins (in the absence of LMP1) but show no detectable Wp, Cp, or Qp usage in the quantitative RT-PCR assays of latent gene tran-scription (Fig. 7B); these EBNA2⫹LMP1⫺clones were also unusual in that they carry just a single copy of a wild-type EBV genome (26a).

Bisulfite sequencing analysis of these different sets of Awia-BL clones revealed that the latency I Awia-BL clone 9 (and clone 20 [data not shown]) resembled Akata-BL in show-ing extensive Wp methylation at all sites except for CpGs c, q, and r (Fig. 9A). Interestingly, the Wp-restricted Awia-BL clones 3 and 4, where the EBNA2-deleted genome load per cell was much lower than in the Wp-restricted Ava-BL, Oku-BL, and Sal-BL clones studied earlier, gave a distinct pattern (Fig. 9A); although there was extensive methylation in the UAS2 region of Wp, CpGs in the B-cell-specific UAS1 region were only partially methylated, a pattern similar to that seen with the Wp-only LCLs (Fig. 3). By contrast, the EBNA2-positive, LMP1-negative Awia-BL clones 1 and 2, also carrying low genome loads but where Wp was silent, showed the same extensive levels of methylation throughout Wp UAS1 and UAS2, as typically seen in latency I BL clones. These patterns were subsequently confirmed by MSP analysis (Fig. 9B). Thus, latency I BL clones and also the EBNA2-positive, LMP1-neg-ative clones showed almost no unmethylated Wp sequences, whereas in the Wp-restricted clones, it was clear that Wp usage was associated with the presence of some unmethylated Wp copies.

Summary of Wp methylation status in different LCL and BL lines.Table 1 presents a summary of the overall results. For each cell line studied, the pattern of EBNA promoter usage (Wp, Cp, or Qp) is shown alongside the methylation status of Wp as determined by MSP assay and by bisulfite sequencing. In the case of the bisulfite sequencing data, the results are expressed as the percentage of methylation of all 20 CpGs analyzed or, specifically, of those methylation-sensitive CpGs lying within the upstream BSAP sites (j and k) or the adjacent BSAP/CREB sites (n, o, and p) in UAS1. By focusing on this bisulfite sequencing data, it can be seen that (with the excep-tion of Ava-BL, Oku-BL, and Sal-BL clones with high genome loads) all cell lines showing Wp transcription tend to have low levels of methylation of CpGs j and k, and cell lines using Wp selectively also have relatively low levels of methylation of CpGs n, o, and p and of overall CpGs in the Wp region. Most importantly, Table 1 emphasizes the absolute correlation be-tween detectable Wp activity in a cell line and the presence of unmethylated Wp sequences, as revealed by MSP analysis.

FIG. 8. (A) Bisulfite sequencing analysis of Wp in latency I Akata-BL and Wp-restricted clones of Ava-BL, Oku-BL, and Sal-BL. Data are presented as described in the Fig. 3 legend. (B) Results of MSP analysis of Wp methylation status in latency I Akata-BL and Wp-restricted clones of Ava-BL, Oku-BL, and Sal-BL. Data are pre-sented as described in the Fig. 4 legend.

on November 8, 2019 by guest

http://jvi.asm.org/

DISCUSSION

Wp is the first viral promoter to be activated during the in vitro transformation of primary resting B cells, but thereafter, levels of Wp-initiated transcripts decline and Cp becomes the dominant EBNA promoter in most established LCLs. While this Wp-to-Cp switch is well documented (3, 43, 56, 58), the mechanism of promoter switching remains unknown. The high levels of Wp-initiated transcripts seen in the early stages of infection might reflect the fact that the incoming viral genomes are unmethylated and nucleosome-free and, thus, are readily accessible to the transcriptional machinery. By contrast, the viral genome in established LCLs is known to adopt a structure similar to that of host chromatin (14, 45). If this structural change occurs early postinfection, perhaps linked to genome circularization which is detectable within the first 24 h (23), then it may contribute to the rapid decline in Wp activity. Indeed, a similar mechanism may be involved in silencing other regions of the EBV genome such as the BHRF1 and BALF1 lytic cycle genes which are transiently expressed following EBV infection (6). An alternative hypothesis is that the activation of the distal EBNA promoter Cp blocks the activity of the down-stream copies of Wp through a transcriptional interference mechanism (35, 36, 59).

In previous work, we suggested that DNA methylation might be implicated in the downregulation of Wp, since we observed that Wp sequences were progressively methylated between 7 and 21 days postinfection, in experiments where peak Wp activity was not reached until day 7 (51). By contrast, in the present study, using newly developed quantitative RT-PCR assays to monitor virus promoter usage, we noted kinetics of Wp methylation that were similar to those described above but in circumstances where Wp activity clearly peaked much ear-lier. These discrepancies may be due in part to differences in RT-PCR methods and/or the dose of transforming virus used in the two studies. Importantly, the present findings make it clear that for Wp, the kinetics of promoter methylation lags significantly behind down-regulation, implying that Wp meth-ylation does not initiate promoter silencing but may serve to maintain promoter sequences in an inactive state (34).

In line with recent findings from other groups (15, 59), the present analysis of a panel of LCLs carrying different EBV strains revealed that Cp was the dominant EBNA promoter in most lines but that Wp was never completely silenced (Fig. 2). This is in contrast to early studies, often based on long-estab-lished LCLs, that propose that Wp and Cp are mutually ex-clusive in their usage (55). The persistence of Wp activity in LCLs also calls into question the relevance of promoter meth-ylation as a regulatory factor, since our bisulfite sequencing data, like those already in the literature (34, 51), showed ex-tensive Wp methylation in all amplified sequences. However, bisulfite sequencing itself gives limited information, since,

un-FIG. 9. (A) Bisulfite sequencing analysis of Wp in latency I, Wp-restricted, and EBNA2⫹LMP1⫺Awia-BL clones. Data are presented as described in the Fig. 3 legend. (B) Results of MSP analysis of Wp methylation status in latency I Akata-BL and Wp-restricted clones of Ava-BL, Oku-BL, and Sal-BL. Data are presented as described in the Fig. 4 legend.

on November 8, 2019 by guest

http://jvi.asm.org/

less very large numbers of amplified products are analyzed, the pattern obtained reflects only the most abundant Wp species. Therefore, we used our understanding of the organization of Wp transcription factor binding sites to design an MSP assay that would detect Wp sequences that had not been methylated in the critical UAS1 regulatory region that is sensitive to meth-ylation. The MSP assay revealed that there was heterogeneity within the standard LCLs such that some unmethylated Wp sequences did exist as a minority species, even though they were never seen by bisulfite sequencing (Fig. 3 and 4). These unmethylated Wp species could therefore account for the low level of Wp transcription observed in standard LCLs.

Importantly, we found two LCLs (EBH41.2 and IM53.1) where Wp was the only active EBNA promoter yet where Cp was apparently intact. These lines are therefore quite distinct from the long-established Wp-only LCLs X50-7 and IB4 which have deletions in BamHI C encompassing Cp (56, 57). Inter-estingly, our recently established Wp-only LCLs, in contrast to conventional Wp/Cp users, were substantially hypomethylated in the promoter-proximal UAS1 region (Fig. 3), and the fact that this difference was apparent even in bisulfite sequencing assays suggests that it affects the majority of Wp copies in the resident EBV episomes. Further studies of these unusual LCLs, both of which arose by spontaneous transformation in peripheral blood mononuclear cells from EBV-infected do-nors, could help identify the controls governing the interrela-tionship between Cp and Wp activities.

Studies of Wp methylation are further complicated by the

presence of multiple Wp copies in each EBV episome. We attempted to overcome this problem experimentally by reduc-ing the Wp copy number in a recombinant EBV genome con-text. We therefore specifically generated a recombinant with only two BamHI W repeats, thought to be the minimum re-quired for transformation (27), and compared this construct with recombinants generated on the same B95.8 background but containing 4, 6, 8, and 11 BamHI W repeats. This work showed that the overall Wp methylation status was critically affected by the Wp repeat number. Thus, Wp sequences were almost entirely unmethylated in 2W LCLs, based on both bisul-fite sequencing and the more-sensitive MSP analysis, whereas 4W, 6W, 8W, and 11W LCLs contained both unmethylated and methylated sequences. Interestingly, Elliott et al. (15) re-cently reported the hypomethylation of Wp in LCLs trans-formed by a recombinant virus that was fortuitously low in BamHI W repeats, but the present work clearly shows the significance of this finding in an experiment with internal high-Wp-copy-number control viruses. The presence of only un-methylated Wp sequences in 2W LCLs strongly suggests that Wp methylation in standard LCLs preferentially targets the downstream copies of Wp and supports the hypothesis that only the most-5⬘copies remain active and unmethylated (59). Analysis of the same 2W LCLs showed the Cp activity to be consistently higher than that in the corresponding 11W LCLs. This may reflect a compensatory mechanism whereby Cp tran-scription is increased to ensure that overall EBNA expression is maintained at the optimal levels required for B-cell trans-TABLE 1. Summary of Wp methylation status in different LCL and BL lines

Cell linec _Latency Promoter usage

Methylation status of Wp determined by:

Bisulfite sequencing (%)a _MSPb

Specific CpG sites _{Total CpG}

a to t U M

Wp Cp Qp j, k n, o, p

EBH41.2 LCL III _⫹ 0 54 35 _{⫹ ⫹}

IM53.1 LCL III ⫹ 8 26 28 ⫹ ⫹

CD_⫹Oku LCL III _{⫹ ⫹} 10 97 66 _{⫹ ⫹}

IM100.1 LCL III ⫹ ⫹ 9 94 64 ⫹ ⫹

11W LCL1 III ⫹ ⫹ 6 81 59 ⫹ ⫹

11W LCL2 III _{⫹ ⫹} 25 100 59 _{⫹ ⫹}

11W LCL3 III ⫹ ⫹ 11 100 65 ⫹ ⫹

2W LCL1 III _{⫹ ⫹} 0 18 15 _⫹

2W LCL2 III ⫹ ⫹ 0 0 4 ⫹

2W LCL3 III _{⫹ ⫹} 0 15 9 _⫹

Akata-BL I _⫹ 92 100 82 _⫹

Rael-BL I ⫹ 88 97 82 ⫹

Ava-BL c.1 Wp restricted _⫹ 71 72 63 _{⫹ ⫹}

Oku-BL c.1 Wp restricted ⫹ 61 76 67 ⫹ ⫹

Sal-BL c.1 Wp restricted _⫹ 65 100 75 _{⫹ ⫹}

Awia-BL c.9 I _⫹ 90 100 79 _⫹

Awia-BL c.20 I ⫹ 86 100 81 ⫹

Awia-BL c.1 EBNA2⫹LMP⫺ 68 82 81 _⫹

Awia-BL c.2 EBNA2⫹LMPI⫺ 85 97 85 ⫹

Awia-BL c.3 Wp restricted _⫹ 30 47 45 _{⫹ ⫹}

Awia-BL c.4 Wp restricted ⫹ 25 64 55 ⫹ ⫹

a_{The methylation status at CpG sites a to t was determined by bisulfite sequence analysis, and the fraction of methylated CpG sites is shown as a percentage of the}

total sites analyzed.

b_{The methylation status was determined by MSP analysis using primers specific for unmethylated (U) or methylated (M) DNA sequences.} c_{c, clone.}

on November 8, 2019 by guest

http://jvi.asm.org/

formation. However, another interesting possibility is that in standard LCLs, the methylation of downstream Wp sequences has a negative effect on Cp transcription. Such a scenario, first proposed by Paulson and Speck (34), is further supported by earlier findings that sequences in the Wp regulatory region are also important for Cp activity (35, 52, 60).

A complementary approach to studying the relationship be-tween Wp activity and methylation status was provided by the recent identification of BL cell lines with an unusual Wp-restricted form of latency (25). These lines are phenotypically very similar to conventional latency I BL lines (expressing EBNA1 from Qp) yet have a quite different pattern of viral transcription in which Qp is silent and EBNA1, -3A, -3B, -3C, and -LP mRNAs are expressed exclusively from Wp. Thus, such lines provide the opportunity to look at an active Wp in the unusual context of a cell with a germinal-center-like (BL) phenotype, rather than a lymphoblastoid (LCL) phenotype. This work revealed interesting parallels with the data from LCLs but also some notable differences. In the Wp-restricted Oku-BL, Sal-BL, and Ava-BL lines carrying multiple viral epi-somes, we found by bisulfite sequencing that the dominant Wp species were hypermethylated, just as they are in latency I BLs. However MSP analysis revealed the presence of unmethylated Wp sequences in these Wp-restricted BLs, a situation not seen in conventional latency I lines. This suggests that in these Wp-restricted BLs, as in standard LCLs, Wp is active in only a minority of the resident Wp copies. A more interesting picture emerged from the analysis of the Wp-restricted clones of the Awia-BL line which carried only a single EBV genome. Here the dominant Wp species were hypomethylated in the critical UAS1 region, extending up to the CpG site i near the UAS2-UAS1 boundary, a pattern similar to that seen in the Wp-only LCLs. Thus, a pattern which appears to be imposed on every resident copy of the virus genome in Wp-only LCLs is only seen in Wp-restricted BLs when the genome copy number is low; this implies that in Wp-restricted BLs with multiple copies of the genome, there may be heterogeneity in methylation patterns between individual genomes.

It is important to note, however, that the Wp-restricted clones of Awia-BL carry a single integrated EBV genome rather than an episomal copy of the EBV genome (26a), and therefore, care must be taken in interpreting the general rel-evance of these particular findings. It is nonetheless interesting to note the contrast between these Wp-restricted Awia-BL clones and clones derived from the same tumor, again with a single integrated virus genome, in which Wp (and Cp) is silent and the cells display an EBNA2⫹LMP1⫺form of infection. Unlike the Wp-restricted clones, where Wp is hypomethylated, in the EBNA2⫹ LMP1⫺ clones, Wp is hypermethylated. Therefore, even in these unusual circumstances, a correlation is maintained between Wp activity and hypomethylation of promoter sequences.

The broader significance of these findings stems from the relationship between DNA methylation and other epigenetic regulatory controls (29). Thus, methylated DNA, through its interaction with methyl CpG binding factors, can recruit his-tone deacetylases and chromatin remodeling complexes that can alter chromatin structure and interfere with access to the transcription machinery. It is very likely that the EBV episome can be remodeled in the same way as cellular chromatin (11,

12, 46). Indeed, this would be consistent with the finding that several EBV-encoded transcription factors, notably EBNA2 and EBNA3C (2, 37), exploit interactions with chromatin re-modeling complexes to regulate viral latent gene expression. The novel cellular models described here may be useful in the dissection of these epigenetic processes.

ACKNOWLEDGMENTS

This work was supported by Cancer Research UK.

We thank the Functional Genomics Laboratory, School of Bio-sciences, University of Birmingham, for help with DNA sequencing.

REFERENCES

1.Abbot, S. D., M. Rowe, K. Cadwallader, A. Ricksten, J. Gordon, F. Wang, L. Rymo, and A. B. Rickinson. 1990. Epstein-Barr virus nuclear antigen 2 induces expression of the virus-encoded latent membrane protein. J. Virol. 64:2126–2134.

2.Alazard, N., H. Gruffat, E. Hiriart, A. Sergeant, and E. Manet.2003. Dif-ferential hyperacetylation of histones H3 and H4 upon promoter-specific recruitment of EBNA2 in Epstein-Barr virus chromatin. J. Virol.77:8166– 8172.

3.Alfieri, C., M. Birkenbach, and E. Kieff.1991. Early events in Epstein-Barr virus infection of human B lymphocytes. Virology181:595–608.

4.Allan, G. J., and D. T. Rowe.1989. Size and stability of the Epstein-Barr virus major internal repeat (IR-1) in Burkitt’s lymphoma and lymphoblastoid cell lines. Virology173:489–498.

5.Allday, M. J., D. Kundu, S. Finerty, and B. E. Griffin.1990. CpG methylation of viral DNA in EBV-associated tumours. Int. J. Cancer45:1125–1130. 6.Altmann, M., and W. Hammerschmidt.2005. Epstein-Barr virus provides a

new paradigm: a requirement for the immediate inhibition of apoptosis. PLoS Biol.3:e404.

7.Baer, R., A. T. Bankier, M. D. Biggin, P. L. Deininger, P. J. Farrell, T. J. Gibson, G. Hatfull, G. S. Hudson, S. C. Satchwell, and C. Seguin.1984. DNA sequence and expression of the B95-8 Epstein-Barr virus genome. Nature 310:207–211.

8.Bell, A., J. Skinner, H. Kirby, and A. Rickinson.1998. Characterisation of regulatory sequences at the Epstein-Barr virus BamHI W promoter. Virol-ogy252:149–161.

8a.Bell, A. I., K. Groves, G. L. Kelly, D. Croom-Carter, E. Hui, A. T. C. Chan, and A. B. Rickinson.Analysis of Epstein-Barr virus latent gene expression in endemic Burkitt’s lymphoma and nasopharyngeal carcinoma tumor cells by using quantitative real-time PCR assays. J. Gen. Virol., in press. 9.Bodescot, M., and M. Perricaudet.1986. Epstein-Barr virus mRNAs

pro-duced by alternative splicing. Nucleic Acids Res.14:7103–7114.

10.Bodescot, M., M. Perricaudet, and P. J. Farrell.1987. A promoter for the highly spliced EBNA family of RNAs of Epstein-Barr virus. J. Virol.61: 3424–3430.

11.Chau, C. M., and P. M. Lieberman.2004. Dynamic chromatin boundaries delineate a latency control region of Epstein-Barr virus. J. Virol.78:12308– 12319.

12.Chau, C. M., X.-Y. Zhang, S. B. McMahon, and P. M. Lieberman.2006. Regulation of Epstein-Barr virus latency type by the chromatin boundary factor CTCF. J. Virol.80:5723–5732.

13.Delecluse, H. J., T. Hilsendegen, D. Pich, R. Zeidler, and W. Hammer-schmidt.1998. Propagation and recovery of intact, infectious Epstein-Barr virus from prokaryotic to human cells. Proc. Natl. Acad. Sci. USA95:8245– 8250.

14.Dyson, P. J., and P. J. Farrell.1985. Chromatin structure of Epstein-Barr virus. J. Gen. Virol.66:1931–1940.

15.Elliott, J., E. B. Goodhew, L. T. Krug, N. Shakhnovsky, L. Yoo, and S. H. Speck.2004. Variable methylation of the Epstein-Barr virus Wp EBNA gene promoter in B-lymphoblastoid cell lines. J. Virol.78:14062–14065. 16.Ernberg, I., K. Falk, J. Minarovits, P. Busson, T. Tursz, M. G. Masucci, and

G. Klein.1989. The role of methylation in the phenotype-dependent mod-ulation of Epstein-Barr nuclear antigen 2 and latent membrane protein genes in cells latently infected with Epstein-Barr virus. J. Gen. Virol.70: 2989–3002.

17.Fahraeus, R., A. Jansson, A. Ricksten, A. Sjo¨blom, and L. Rymo.1990. Epstein-Barr virus-encoded nuclear antigen 2 activates the viral latent mem-brane protein promoter by modulating the activity of a negative regulatory element. Proc. Natl. Acad. Sci. USA87:7390–7394.

18.Feederle, R., M. Kost, M. Baumann, A. Janz, E. Drouet, W. Hammer-schmidt, and H. J. Delecluse.2000. The Epstein-Barr virus lytic program is controlled by the co-operative functions of two transactivators. EMBO J. 19:3080–3089.

19.Finke, J., M. Rowe, B. Kallin, I. Ernberg, A. Rose´n, J. Dillner, and G. Klein. 1987. Monoclonal and polyclonal antibodies against Epstein-Barr virus nu-clear antigen 5 (EBNA-5) detect multiple protein species in Burkitt’s lym-phoma and lymphoblastoid cell lines. J. Virol.61:3870–3878.

on November 8, 2019 by guest

http://jvi.asm.org/

20.Frommer, M., L. E. McDonald, D. S. Millar, C. M. Collis, F. Watt, G. W. Grigg, P. L. Molloy, and C. L. Paul.1992. A genomic sequencing protocol that yields a positive display of 5-methylcytosine residues in individual DNA strands. Proc. Natl. Acad. Sci. USA89:1827–1831.

21.Grasser, F. A., P. G. Murray, E. Kremmer, K. Klein, K. Remberger, W. Feiden, G. Reynolds, G. Niedobitek, L. S. Young, and N. Mueller-Lantzsch. 1994. Monoclonal antibodies directed against the Epstein-Barr virus-en-coded nuclear antigen 1 (EBNA1): immunohistologic detection of EBNA1 in the malignant cells of Hodgkin’s disease. Blood84:3792–3798. 22.Herman, J. G., J. R. Graff, S. Myohanen, B. D. Nelkin, and S. B. Baylin.

1996. Methylation-specific PCR: a novel PCR assay for methylation status of CpG islands. Proc. Natl. Acad. Sci. USA93:9821–9826.

23.Hurley, E. A., and D. A. Thorley-Lawson.1988. B cell activation and the establishment of Epstein-Barr virus latency. J. Exp. Med.168:2059–2075. 24.Jansson, A., M. Masucci, and L. Rymo.1992. Methylation of discrete sites

within the enhancer region regulates the activity of the Epstein-Barr virus BamHI W promoter in Burkitt lymphoma lines. J. Virol.66:62–69. 25.Kelly, G., A. Bell, and A. Rickinson.2002. Epstein-Barr virus-associated

Burkitt lymphomagenesis selects for downregulation of the nuclear antigen EBNA2. Nat. Med.8:1098–1104.

26.Kelly, G. L., A. E. Milner, R. J. Tierney, D. S. Croom-Carter, M. Altmann, W. Hammerschmidt, A. I. Bell, and A. B. Rickinson.2005. Epstein-Barr virus nuclear antigen 2 (EBNA2) gene deletion is consistently linked with EBNA3A, -3B, and -3C expression in Burkitt’s lymphoma cells and with increased resistance to apoptosis. J. Virol.79:10709–10717.

26a.Kelly, G. L., A. E. Milner, G. S. Baldwin, A. I. Bell, and A. B. Rickinson. Three restricted forms of Epstein-Barr virus latency counteracting apoptosis in c-myc expressing Burkitt’s lymphoma cells. Proc. Natl. Acad. Sci. USA, in press.

27.Kempkes, B., D. Pich, R. Zeidler, B. Sugden, and W. Hammerschmidt.1995. Immortalization of human B lymphocytes by a plasmid containing 71 kilo-base pairs of Epstein-Barr virus DNA. J. Virol.69:231–238.

28.Kirby, H., A. Rickinson, and A. Bell.2000. The activity of the Epstein-Barr virus BamHI W promoter in B cells is dependent on the binding of CREB/ ATF factors. J. Gen. Virol.81:1057–1066.

29.Klose, R. J., and A. P. Bird.2006. Genomic DNA methylation: the mark and its mediators. Trends Biochem. Sci.31:89–97.

30.Masucci, M. G., B. Contreras-Salazar, E. Ragnar, K. Falk, J. Minarovits, I. Ernberg, and G. Klein.1989. 5-Azacytidine up regulates the expression of Epstein-Barr virus nuclear antigen 2 (EBNA-2) through EBNA-6 and latent membrane protein in the Burkitt’s lymphoma line Rael. J. Virol.63:3135– 3141.

31.Minarovits, J., L. F. Hu, S. Minarovits-Kormuta, G. Klein, and I. Ernberg. 1994. Sequence-specific methylation inhibits the activity of the Epstein-Barr virus LMP 1 and BCR2 enhancer-promoter regions. Virology200:661–667. 32.Minarovits, J., S. Minarovits-Kormuta, B. Ehlin-Henriksson, K. Falk, G. Klein, and I. Ernberg.1991. Host cell phenotype-dependent methylation patterns of Epstein-Barr virus DNA. J. Gen. Virol.72:1591–1599. 33.Nonkwelo, C., J. Skinner, A. Bell, A. Rickinson, and J. Sample.1996.

Tran-scription start sites downstream of the Epstein-Barr virus (EBV) Fp pro-moter in early-passage Burkitt lymphoma cells define a fourth propro-moter for expression of the EBV EBNA-1 protein. J. Virol.70:623–627.

34.Paulson, E. J., and S. H. Speck.1999. Differential methylation of Epstein-Barr virus latency promoters facilitates viral persistence in healthy seropos-itive individuals. J. Virol.73:9959–9968.

35.Puglielli, M. T., N. Desai, and S. H. Speck.1997. Regulation of EBNA gene transcription in lymphoblastoid cell lines: characterization of sequences downstream of BCR2 (Cp). J. Virol.71:120–128.

36.Puglielli, M. T., M. Woisetschlaeger, and S. H. Speck.1996.oriPis essential for EBNA gene promoter activity in Epstein-Barr virus-immortalized lym-phoblastoid cell lines. J. Virol.70:5758–5768.

37.Radkov, S. A., R. Touitou, A. Brehm, M. Rowe, M. West, T. Kouzarides, and M. J. Allday.1999. Epstein-Barr virus nuclear antigen 3C interacts with histone deacetylase to repress transcription. J. Virol.73:5688–5697. 38.Rickinson, A. B., and E. Kieff.2001. Epstein-Barr virus, p. 2575–2627.In

D. M. Knipe and P. M. Howley (ed.), Fields virology, 4th ed., vol. II. Lippincott Williams & Wilkins, Philadelphia, Pa.

39.Rooney, C. M., M. Brimmell, M. Buschle, G. Allan, P. J. Farrell, and J. L. Kolman.1992. Host cell and EBNA-2 regulation of Epstein-Barr virus la-tent-cycle promoter activity in B lymphocytes. J. Virol.66:496–504. 40.Rowe, M., H. S. Evans, L. S. Young, K. Hennessy, E. Kieff, and A. B.

Rickinson.1987. Monoclonal antibodies to the latent membrane protein of Epstein-Barr virus reveal heterogeneity of the protein and inducible expres-sion in virus-transformed cells. J. Gen. Virol.68:1575–1586.

41.Sample, J., M. Hummel, D. Braun, M. Birkenbach, and E. Kieff.1986. Nucleotide sequences of mRNAs encoding Epstein-Barr virus nuclear pro-teins: a probable transcriptional initiation site. Proc. Natl. Acad. Sci. USA 83:5096–5100.

42.Schaefer, B. C., J. L. Strominger, and S. H. Speck.1995. Redefining the Epstein-Barr virus-encoded nuclear antigen EBNA-1 gene promoter and transcription initiation site in group I Burkitt lymphoma cell lines. Proc. Natl. Acad. Sci. USA92:10565–10569.

43.Schlager, S., S. H. Speck, and M. Woisetschlager.1996. Transcription of the Epstein-Barr virus nuclear antigen 1 (EBNA1) gene occurs before induction of the BCR2 (Cp) EBNA gene promoter during the initial stages of infection in B cells. J. Virol.70:3561–3570.

44.Shannon-Lowe, C., G. Baldwin, R. Feederle, A. Bell, A. Rickinson, and H. J. Delecluse.2005. Epstein-Barr virus-induced B-cell transformation: quanti-tating events from virus binding to cell outgrowth. J. Gen. Virol.86:3009– 3019.

45.Shaw, J. E., L. F. Levinger, and C. W. Carter, Jr.1979. Nucleosomal structure of Epstein-Barr virus DNA in transformed cell lines. J. Virol. 29:657–665.

46.Sjo¨blom-Halle´n, A., W. Yang, A. Jansson, and L. Rymo.1999. Silencing of the Epstein-Barr virus latent membrane protein 1 gene by the Max-Mad1-mSin3A modulator of chromatin structure. J. Virol.73:2983–2993. 47.Speck, S. H., and J. L. Strominger.1985. Analysis of the transcript encoding

the latent Epstein-Barr virus nuclear antigen I: a potentially polycistronic message generated by long-range splicing of several exons. Proc. Natl. Acad. Sci. USA82:8305–8309.

48.Sung, N. S., S. Kenney, D. Gutsch, and J. S. Pagano.1991. EBNA-2 trans-activates a lymphoid-specific enhancer in the BamHI C promoter of Epstein-Barr virus. J. Virol.65:2164–2169.

49.Thorley-Lawson, D. A.2001. Epstein-Barr virus: exploiting the immune system. Nat. Rev. Immunol.1:75–82.

50.Tierney, R., H. Kirby, J. Nagra, A. Rickinson, and A. Bell.2000. The Epstein-Barr virus promoter initiating B-cell transformation is activated by RFX proteins and the B-cell-specific activator protein BSAP/Pax5. J. Virol.74: 10458–10467.

51.Tierney, R. J., H. E. Kirby, J. K. Nagra, J. Desmond, A. I. Bell, and A. B. Rickinson.2000. Methylation of transcription factor binding sites in the Epstein-Barr virus latent cycle promoter Wp coincides with promoter down-regulation during virus-induced B-cell transformation. J. Virol.74:10468– 10479.

52.Walls, D., and M. Perricaudet.1991. Novel downstream elements up-regu-late transcription initiated from an Epstein-Barr-Virus up-regu-latent promoter. EMBO J.10:143–151.

53.Wang, F., S.-F. Tsang, M. G. Kurilla, J. I. Cohen, and E. Kieff.1990. Epstein-Barr virus nuclear antigen 2 transactivates latent membrane protein LMP1. J. Virol.64:3407–3416.

54.Woisetschlaeger, M., X. W. Jin, C. N. Yandava, L. A. Furmanski, J. L. Strominger, and S. H. Speck.1991. Role for the Epstein-Barr virus nuclear antigen 2 in viral promoter switching during initial stages of infection. Proc. Natl. Acad. Sci. USA88:3942–3946.

55.Woisetschlaeger, M., J. L. Strominger, and S. H. Speck.1989. Mutually exclusive use of viral promoters in Epstein-Barr virus latently infected lym-phocytes. Proc. Natl. Acad. Sci. USA86:6498–6502.

56.Woisetschlaeger, M., C. N. Yandava, L. A. Furmanski, J. L. Strominger, and S. H. Speck.1990. Promoter switching in Epstein-Barr virus during the initial stages of infection of B lymphocytes. Proc. Natl. Acad. Sci. USA87:1725– 1729.

57.Yandava, C. N., and S. H. Speck.1992. Characterization of the deletion and rearrangement in the BamHI C region of the X50-7 Epstein-Barr virus genome, a mutant viral strain which exhibits constitutive BamHI W pro-moter activity. J. Virol.66:5646–5650.

58.Yoo, L., and S. H. Speck.2000. Determining the role of the Epstein-Barr virus Cp EBNA2-dependent enhancer during the establishment of latency by using mutant and wild-type viruses recovered from cottontop marmoset lymphoblastoid cell lines. J. Virol.74:11115–11120.

59.Yoo, L. I., M. Mooney, M. T. Puglielli, and S. H. Speck.1997. B-cell lines immortalized with an Epstein-Barr virus mutant lacking the Cp EBNA2 enhancer are biased toward utilization of theoriP-proximal EBNA gene promoter Wp1. J. Virol.71:9134–9142.

60.Yoo, L. I., J. Woloszynek, S. Templeton, and S. H. Speck.2002. Deletion of Epstein-Barr virus regulatory sequences upstream of the EBNA gene pro-moter Wp1 is unfavorable for B-cell immortalization. J. Virol.76:11763– 11769.

61.Young, L., C. Alfieri, K. Hennessy, H. Evans, C. O’Hara, K. C. Anderson, J. Ritz, R. S. Shapiro, A. Rickinson, E. Kieff, et al.1989. Expression of Epstein-Barr virus transformation-associated genes in tissues of patients with EBV lymphoproliferative disease. N. Engl. J. Med.321:1080–1085.

62.Zimber-Strobl, U., K.-O. Suentzenich, G. Laux, D. Eick, M. Cordier, A. Calender, M. Billaud, G. M. Lenoir, and G. W. Bornkamm.1991. Epstein-Barr virus nuclear antigen 2 activates transcription of the terminal protein gene. J. Virol.65:415–423.

on November 8, 2019 by guest

http://jvi.asm.org/