Mechanism of activation of the BNLF2a immune evasion gene of Epstein-Barr virus by Zta

Abstract

The human gamma herpes virus Epstein-Barr virus (EBV) exploits multiple routes to evade the cellular immune response. During the EBV lytic replication cycle, viral proteins are expressed that provide excellent targets for recognition by cytotoxic T cells. This is countered by the viral BNLF2a gene. In B cells during latency, where BNLF2a is not expressed, we show that its regulatory region is embedded in repressive chromatin. The expression of BNLF2a mirrors the expression of a viral lytic cycle transcriptional regulator, Zta (BZLF1, EB1, ZEBRA), in B cells and we propose that Zta plays a role in up-regulating BNLF2a. In cells undergoing EBV lytic replication, we identified two distinct regions of interaction of Zta with the chromatin-associated BNLF2a promoter. We identify five potential Zta-response elements (ZREs) in the promoter that are highly conserved between virus isolates. Zta binds to these elements in vitro and activates the expression of the BNLF2a promoter in both epithelial and B cells. We also found redundancy amongst the ZREs. The EBV genome undergoes a biphasic DNA methylation cycle during its infection cycle. One of the ZREs contains an integral CpG motif. We show that this can be DNA methylated during EBV latency and that both Zta binding and promoter activation are enhanced by its methylation. In summary, we find that the BNLF2a promoter is directly targeted by Zta and that DNA methylation within the proximal ZRE aids activation. The implications for regulation of this key viral gene during the reactivation of EBV from latency are discussed.

Keywords: BNLF2a; BZLF1; Epstein-Barr virus; epigenetics; gene expression; immune evasion.

Fig. 1.

Chromatin organization at the BNLF2a promoter is associated with repressive H3K27me3 and H3K9me3 modifications during latency. Chromatin was isolated from cells harbouring latent EBV, an LCL (a, b), Akata BL (c, d) and Raji BL (e, f) cells. Chromatin precipitation was undertaken with antibodies specific for the modified histones (H3K27me3 (a, c, e) and H3K9me3 (b, d, f) and their relevant species-specific controls. DNA was eluted from the precipitate and the relative amounts of each of the indicated loci analysed by Q-PCR relative to the input genomes, and is expressed as a percentage of input binding. In each case the standard deviation is shown (triplicate measurements). The significance of the difference in binding is shown as **P≤0.01; ***P 0.001).

Fig. 2.

Zta binding to the BNLF2a promoter region. ChIP-Seq data from Akata BL cells undergoing lytic cycle with an antibody that recognizes chromatin-bound Zta was aligned the EBV genome. The nucleotide position of the EBV genome is shown on the x-axis. The locations of gene coding regions are shown as arrows, with amplicons used to identify the specific regions by ChIP shown as boxes. (a). The BNLF2a Locus. (b). The OriLyt region. (c-d). ChIP coupled to Q-PCR analysis of Zta binding was undertaken from chromatin from cells harbouring lytic EBV: induced Akata BL cells (48 h) (c) and spontaneously lytic LCL#3 cells (d). The Zta ChIP is shown as a black bar and the control antibody as an open bar. Q-PCR amplification of the indicated loci was undertaken in triplicate. The axes show percentage binding relative to input chromatin and the error bars indicate the standard deviation. The significance of the difference in Zta binding to OriLyt flank and Zta is shown (**P≤0.01).

Fig. 3.

Interaction between Zta and the BNLF2a ZREs in vitro. (a). A His-tagged GST-Zta expression vector encoding the DNA binding and dimerization region of Zta was generated. (b). Protein was produced in E. coli and purified using cobalt ion affinity purification. 160 ng of the resulting protein was fractionated on SDS-PAGE together with 160 ng of a his-tagged GST protein and visualized with Simply Blue stain. (c). 80 ng of His GST and His-GST-Zta proteins were incubated with IR-labelled double-strand oligonucleotide probes corresponding to BNLF2a ZRE2 or a mutant version of the sequence, and the reactions were then separated on native polyacrylamide gels using EMSA. The migration of free DNA and bound DNA complexes is shown. (d). Equivalent EMSAs were undertaken with each of the individual BNLF2a ZREs and the mutant version. The bar graph shows quantitation of binding (% of probe bound) with the standard deviation from triplicate assays. The statistical significance of binding was compared to the signal generated with His-GST protein (*P≤0.05; **P≥0.01).

Fig. 4.

Contribution of Zta-binding regions to BNLF2a promoter activation by Zta. (a). Schematic diagrams of the BNLF2a promoter–luciferase reporter system. The ZREs are represented by 1–5, with the TATA box and start of the luciferase gene shown. The mutation of ZREs is represented by an X. (b). The indicated plasmids were introduced into DG75 cells with the his-Zta expression vector (black box) or a control plasmid (open box). Cells were incubated for 48 h and luciferase activity determined. Zta and actin protein expression were determined following Western blotting. The statistical significance of the difference in Zta-driven promoter activity between BNLF2a 1–5 and the mutants is shown **P≤0.01). (c). The indicated plasmids were introduced into HeLa cells with either the his-Zta expression vector or a control plasmid. Cells were incubated for 48 h and Zta-driven promoter activity determined as in (b).

Fig. 5.

Contribution of individual proximal Zta-binding elements to BNLF2a promoter activation by Zta. (a). Schematic diagrams of the mutations introduced into the BNLF2a promoter–luciferase reporter system. (b). The indicated plasmids were introduced into DG75 cells with or without the his-Zta expression vector. Cells were incubated for 48 h and Zta-driven promoter activity determined. Zta and actin protein expression were determined following Western blotting. The significance of the difference in promoter activity between BNLF2a ZRE1, 3 and 5 and 1–5 is shown, **P≤0.01). (c). The indicated plasmids were introduced into HeLa cells with or without the his-Zta expression vector. Cells were incubated for 48 h and Zta-driven promoter activity determined as in (b).

Fig. 6.

Methylation at BNLF2a ZRE1 and impact on Zta DNA binding (a). The DNA sequence of BNLF2 ZRE1 is shown with the C of each CpG motif underlined. (b). DNA methylation analysis of the 193 bp region surrounding the BNLF2a ZRE1 region of EBV is shown for Akata cell latency. The location and orientation of the BNLF2a coding region is indicated. Following bisulphite treatment of genomic EBV DNA, and amplification of the 193 base-pair region, the DNA was sub-cloned and individual colonies were subject to DNA sequence analysis. The methylation status of the six CpG motifs within the amplicon was determined and is shown. Black represents 100 % methylation, white represents non-methylation and grey represents a mixture of methylated and non-methylated templates. The locations of the TATA box and ZRE1 are shown for reference. (c). His-GST or His-GST-Zta was incubated with IR-labelled double-strand oligonucleotide probes corresponding to either BNLF2a ZRE1 or methylated ZRE1, and then the reactions were separated on native polyacrylamide gels using EMSA. The bar graph shows quantitation of binding (% of probe bound), together with the standard deviation from triplicate assays.

Fig. 7.

Impact of methylation of ZRE1 on BNLF2a promoter activity (a). Schematic diagram of mutations that leave only the proximal ZRE within the BNLF2a promoter. (b). The indicated plasmids were subject to methylation or not using CpG methyltransferase (M.SssI). These were introduced into 293T cells with either the his-Zta expression vector or its control vector. Cells were incubated for 24 h and promoter activity determined using luciferase assays. The black bars denote the change in promoter activity when his-Zta is expressed. The statistical significance of the difference in Zta-driven promoter activity between non- and methylated BNLF2a 1 is shown (**P≤0.01). (c). Zta and actin protein expression was determined following Western blotting.