Fig 1.
Repression of ADAM28, ADAMDEC1 and COBLL1 after infection of primary B cells with EBV.
Infection of primary B cells with wild type (wtI6), recombinant EBNA3A, EBNA3B or EBNA3C knockout (KO) or revertant (Rev) recombinant viruses. Gene expression for ADAM28 (A), ADAMDEC1 (B) and COBLL1 (C) was normalised to GAPDH and is shown relative to uninfected primary B cells. Normalisation to other internal control genes (e.g. GNB2L1) showed very similar results.
Fig 2.
Repression of ADAM28, ADAMDEC1 and COBLL1 in EBNA3C conditional cell line.
(A-C) Time-course using EBNA3C-conditional LCL 3CHT A13. Cells were grown over 60 days either in absence of HT (-HT), presence of HT (+HT) or with HT removed after 30 days +HT (washed). Gene expression for ADAM28 (A), ADAMDEC1 (B) and COBLL1 (C) was normalised to GAPDH and is shown relative to -HT on day 0. The dashed line indicates gene expression levels of the repressed state, which equals the limit of detection for COBLL1 –values below this line become extremely unreliable. (D) COBLL1 and γ-tubulin protein expression during 3CHT A13 time-course in the absence of HT (-HT), presence of HT (+HT) or HT washed away on day 30 (washed).
Fig 3.
Epigenetic changes and factor accumulation at sites within the COBLL1 locus during the 3CHT A13 time-course.
(A) UCSC genome browser overview of COBLL1 genomic locus showing ChIP-seq tracks for EBNA3A-TAP, EBNA3C-TAP, the TSS (horizontal arrow) and CpG islands. Primer locations for ChIP-qPCR (COBLL1 peak, control and TSS) are indicated (vertical arrows below UCSC genes track). (B-H) ChIP for H3K9ac (B), H3K27ac (C), H3K4me3 (D), H3K27me3 (E), SUZ12 (F), BMI1 (G) and RBPJ (H) on samples from 3CHT A13 time-course at locations within the COBLL1 locus, at GAPDH or myoglobin as indicated. Cells were grown in the absence (-HT) or presence of HT (+HT) and numbers indicate the day of harvest. ChIP values represent enrichment relative to input ± standard deviations of triplicate qPCR reactions for ChIP and input of each sample.
Fig 4.
Epigenetic changes and factor accumulation at sites within the ADAM28-ADAMDEC1 locus during the 3CHT A13 time-course.
(A) UCSC genome browser overview of ADAM28-ADAMDEC1 genomic locus showing ChIP-seq tracks for EBNA3A-TAP, EBNA3C-TAP, the TSSs (horizontal arrows) and CpG islands. Primer locations for ChIP-qPCR (TSS ADAM28, control, ADAM peak and TSS ADAMDEC1) are indicated (vertical arrows below UCSC genes track). (B-H) ChIP for H3K9ac (B), H3K27ac (C), H3K4me3 (D), H3K27me3 (E), SUZ12 (F), BMI1 (G) and RBPJ (H) on samples from 3CHT A13 time-course at locations across the ADAM28-ADAMDEC1 locus, at GAPDH or myoglobin as indicated. Cells were grown in the absence (-HT) or presence of HT (+HT) and numbers indicate the day of harvest. ChIP values represent enrichment relative to input ± standard deviations of triplicate qPCR reactions for ChIP and input of each sample.
Fig 5.
Correlation between changes of epigenetic marks over time at the TSS of ADAM28, ADAMDEC1 and COBLL1 with gene expression.
ChIP-qPCR values (top) at the TSS of ADAM28 (A), ADAMDEC1 (B) and COBLL1 (C) from 3CHT A13 and 3CHT C19 time-courses were normalised to day 30 +HT for the repressive histone mark H3K27me3 and to day three -HT for the activation-associated histone marks H3K4me3, H3K9ac and H3K27ac, which were then set as day 0. Mean values ± standard deviation from both replicate time-courses are shown. For better visual clarity, error bars are colour-matching the corresponding histone mark and only upper bars are shown for activation-associated marks and lower bars for H3K27me3. mRNA expression data for each gene (bottom) are shown as mean values ± standard deviation from three replicate time courses (3CHT A13 rep 1+2 and 3CHT C19).
Fig 6.
Recruitment of SUZ12, BMI1 and RBPJ to EBNA3C-binding peaks in ADAM28-ADAMDEC1 and COBLL1 loci.
ChIP-qPCR values for SUZ12 at the TSS of COBLL1 (A), BMI1 (B) and RBPJ (C) at the EBNA3C-binding peaks at ADAM28-ADAMDEC1 (ADAM peak) and COBLL1 (COBLL1 peak) from 3CHT A13 and 3CHT C19 time-courses were normalised by setting the maximal ChIP enrichment level during each time course to one and by using -HT day three as representative value for day 0. For BMI1, day nine of the C19 time course was treated as an outlier and not taken into account. Mean values ± standard deviation from both replicate time-courses are shown. For better visual clarity, error bars are colour-matched to either ADAM peak (only lower bars displayed) or COBLL1 peak (only upper bars displayed).
Fig 7.
Repression of COBLL1 transient reporter construct by EBNA3C is dependent on the COBLL1 peak and RBPJ.
(A) Schematic of luciferase vectors used in transient reporter assays. Black arrow represents 1 kb promoter region upstream of COBLL1 TSS and black box 1.5 kb around EBNA3C-binding peak at COBLL1 cloned into pGL3-basic vector. (B) Luciferase reporter assay in Raji using 1 μg of luciferase vector as indicated and expressed relative to pGL3-basic vector (n = 3). (C) Luciferase reporter assay in Raji using 1 μg of pGL3-COBLL1-peak luciferase vector and co-transfecting increasing amounts of EBNA3 expression plasmids as indicated (n = 4). (D) Luciferase reporter assay as in C but co-transfecting increasing amounts of RBPJ binding mutant (BM) of EBNA3C (n = 3). Immunoblots show corresponding EBNA3 protein and γ-tubulin as loading control. All luciferase units were normalised to beta-galactosidase units of the same transfection. Results are shown as mean ± standard deviation.
Fig 8.
Repression of ADAM28 transient reporter construct by EBNA3A and EBNA3C is dependent on the ADAM peak and RBPJ.
(A) Schematic of luciferase vectors used in transient reporter assays. Black arrow represents 1 kb promoter region upstream of ADAM28 TSS and black box 1 kb around EBNA3C-binding peak at ADAM28-ADAMDEC1 locus cloned into pGL3-basic vector. (B) Luciferase reporter assay in DG75 using 1 μg of luciferase vector as indicated and expressed relative to pGL3-basic vector (n = 6). (C) Luciferase reporter assay in DG75 using 1 μg of pGL3-ADAM28-peak luciferase vector and co-transfecting increasing amounts of EBNA3 expression plasmids as indicated (n = 3 for EBNA3A and EBNA3B, n = 6 for EBNA3C). (D) Luciferase reporter assays as in C for EBNA3A (n = 3) and EBNA3C (n = 4) but in RBPJ-null DG75. (E) Luciferase reporter assay as in C but co-transfecting increasing amounts of expression plasmid encoding for RBPJ binding mutant (BM) of EBNA3C (n = 3). Immunoblots show corresponding EBNA3 protein and γ-tubulin as loading control. All luciferase units were normalised to beta-galactosidase units of the same transfection. Results are shown as mean ± standard deviation.
Fig 9.
Construction of recombinant EBV with RBPJ binding mutant EBNA3C and validation of established LCL.
(A) Schematic representation of the construction of the EBV recombinant encoding the EBNA3C binding mutant (EBNA3C BM) incapable of binding RBPJ. The N-terminal XbaI/BglII fragment of EBNA3C was excised from the B95.8 EBV bacterial artificial chromosome (BAC) and used as template for In-Fusion based mutagenesis. The two RBPJ binding sites, 209TFGC and 227WTP, of EBNA3C were mutated to 209AAAA and W227S based on previous studies [11,42]. The nucleotide sequence of the wild type EBV BAC is shown in red and the mutated sequence in blue. Mutations were introduced in a two-step In-Fusion based mutagenesis process, with the 209AAAA mutation (homology domain (HD) mutant) generated first, introducing a NotI restriction site, followed by the W227S mutation, introducing a SalI restriction site. The resulting RBPJ binding mutant EBNA3C was reintroduced into the B95.8 EBV BAC by RecA mediated homologous recombination. (B) Cell proliferation assay at day 36 after infection of primary B cells with wild type, EBNA3C knockout (3CKO), EBNA3C revertant (3CRev) or RBPJ binding mutant EBNA3C (RBPJ BM) recombinant viruses. Live cells were analysed for proliferation by measuring EdU incorporation and DNA content by FxCycle Far red DNA stain. Gates show populations of cells in sub-G1, G1, S-phase or G2/M phase. (C) Immunoblot for EBNA3A, EBNA3B, EBNA3C, EBNA2, EBNALP, LMP1, RBPJ and γ-tubulin on LCLs established from primary B cell infection with wild-type (WT), EBNA3C Revertant (3CRev), EBNA3C knockout (3CKO) and RBPJ binding mutant EBNA3C (RBPJ BM) EBV. (D) Immunoprecipitation (IP) of RBPJ or antibody isotype control (IgG) in WT LCL or RBPJ BM LCL and western blotting (WB) for EBNA3C or RBPJ as indicated. Input represents 10% of lysate used in IPs and arrows indicate bands of EBNA3C or RBPJ.
Fig 10.
Ability of EBNA3C to bind to RBPJ appears to be essential for repression of ADAM28, ADAMDEC1 and COBLL1.
(A-C) Infection of primary B cell with either wild type (wtA5), EBNA3C knockout (3CKO), EBNA3C revertant (3CRev) or RBPJ binding mutant EBNA3C recombinant EBV. Gene expression of ADAM28 (A), ADAMDEC1 (B) and COBLL1 (C) was normalised to GNB2L1 and is shown relative to uninfected primary B cells.