Epstein-Barr virus exploits intrinsic B-lymphocyte transcription programs to achieve immortal cell growth

Abstract

Epstein-Barr virus nuclear antigen 2 (EBNA2) regulation of transcription through the cell transcription factor RBPJ is essential for resting B-lymphocyte (RBL) conversion to immortal lymphoblast cell lines (LCLs). ChIP-seq of EBNA2 and RBPJ sites in LCL DNA found EBNA2 at 5,151 and RBPJ at 10,529 sites. EBNA2 sites were enriched for RBPJ (78%), early B-cell factor (EBF, 39%), RUNX (43%), ETS (39%), NFκB (22%), and PU.1 (22%) motifs. These motif associations were confirmed by LCL RBPJ ChIP-seq finding 72% RBPJ occupancy and Encyclopedia Of DNA Elements LCL ChIP-seq finding EBF, NFκB RELA, and PU.1 at 54%, 31%, and 17% of EBNA2 sites. EBNA2 and RBPJ were predominantly at intergene and intron sites and only 14% at promoter sites. K-means clustering of EBNA2 site transcription factors identified RELA-ETS, EBF-RUNX, EBF, ETS, RBPJ, and repressive RUNX clusters, which ranked from highest to lowest in H3K4me1 signals and nucleosome depletion, indicative of active chromatin. Surprisingly, although quantitatively less, the same genome sites in RBLs exhibited similar high-level H3K4me1 signals and nucleosome depletion. The EBV genome also had an LMP1 promoter EBF site, which proved critical for EBNA2 activation. LCL HiC data mapped intergenic EBNA2 sites to EBNA2 up-regulated genes. FISH and chromatin conformation capture linked EBNA2/RBPJ enhancers 428 kb 5' of MYC to MYC. These data indicate that EBNA2 evolved to target RBL H3K4me1 modified, nucleosome-depleted, nonpromoter sites to drive B-lymphocyte proliferation in primary human infection. The primed RBL program likely supports antigen-induced proliferation.

Fig. 1.

EBNA2 binding site enriched TF motifs, K-means TF clusters, and associated H3K4me1 signals in LCLs and at the same genome sites in RBLs. (A) EBNA2 binding site (±250 bp) enriched TF motifs are ranked by enrichment over random sequences controlling for GC content from top to bottom. All had corrected P values <10⁻⁵⁰. (B) EBNA2 binding sites (5,151) were K-means clustered based on the presence of RBPJ, EBF, PU.1, and NFκB subunit RELA binding by LCL ChIPseq and Homer-imputed ETS and RUNX motifs. EBNA2 site clusters are named by the principal TF components. (C) LCL EBNA2 site clusters have distinct H3K4me1 signal distributions (Left). H3K4me1 distributions at the same sites in RBLs (Right). The numbers in the upper left corners of the panels are the total signals (reads per kilobase per million mapped reads) under each curve.

Fig. 2.

EBF is important for EBNA2 activation of the EBV LMP1 promoter. (A) Schematic of the LMP1 promoter EBNA2 response element (−227/−137), with RBPJ, EBF, and PU.1 binding sites underlined in the continuous sequence. Null EBF or PU.1 point mutations are indicated below the continuous sequence. (B) EBNA2 activated the LMP1 promoter-luciferase reporter eightfold, the EBF mutant site twofold, and the PU.1 mutant site onefold in BJAB cells. Luciferase activities were normalized to cotransfected EBNA2-independent β-galactosidase activity. EBNA2 was expressed at similar levels in all experiments as shown in the EBNA2 immune blot at the bottom of B. (C) In vitro translated EBF bound to a ³²P labeled oligonucleotide that included the EBF binding site. Binding was completed by a 500-fold excess cold wild-type oligonucleotide (w), but not by 500-fold cold mutant nucleotide (m). Protein-DNA complexes were separated by PAGE and visualized by phosphoimager.

Fig. 3.

Nonpromoter EBNA2 binding sites are enriched for long-range interactions with EBNA2 regulated or LCL >fivefold up-regulated. HiC detected 3.2 long-range interactions between intergenic EBNA2 binding sites and EBNA2 conditionally regulated genes versus control genes (corrected P < 0.0001) and two long-range interactions between EBV >fivefold up-regulated genes (corrected P < 0.05). Control genes had similar RNA levels and distance to EBNA2 binding sites as the 81 conditionally regulated or >fivefold up-regulated genes.

Fig. 4.

EBNA2 activates MYC via long-range enhancer and promoter interactions. (A) CTCF, EBNA2, RBPJ, EBF, RELA, H3K9ac, H3K4me1, PolII, and p300 signals and MYC promoter (arrow) are shown. Normalized signals are at the left end of each track. The asterisk indicates unidirectional EBNA2 reads. The three small dots at −168 kb to −186 kb indicate three significant EBNA2/RBPJ peaks. Major EBNA2/RBPJ peaks are −428 kb to −556 kb. For enlarged view, see Fig. S5B. (B) FISH for IB4 LCL and RBLs. Fixed LCLs or RBLs were hybridized with fluorescent labeled BACmid probes for DNA 428 to 556 kb upstream of MYC, (green) and 150 kb centered around MYC (orange). Cell DNA is stained blue with DAPI. Mycp indicates myc promoter and myce indicates myc enhancer. (C) One-hundred LCLs or RBLs were scored for colocalization of EBNA2 up-stream sites and MYC. Blue indicates the percentage of cells with one allele colocalized and one allele separate. (D) Chromatin conformation capture qPCR assay using conditional EBNA2 LCLs grown under permissive or nonpermissive conditions for 4 d. Cells were cross-linked with formaldehyde, DNA digested with Csp6I, diluted, ligated, and quantitated by Taqman qPCR. Fold differences were determined using the ΔΔCt method with EBNA2-expressing cells set at 1.