Fig 1.
Overview of Integrated Omics Analysis of EBV immortalization of B-lymphocytes.
(A) Schematic representation of multi-omics experimental design across the time course of EBV-mediated B-cell immortalization. (B) Independent principal component analyses for ATAC-seq, RNA-seq, and metabolomics datasets.
Fig 2.
ATAC-seq analysis of accessible chromatin across EBV immortalization.
Chromatin accessibility was determined over the infection time course using ATAC-seq. (A) Number of ATAC-seq sites at different stages of data filtering and analysis of differentially accessible chromatin. (B) Distribution of ATAC-seq sites relative to genes. (C) Heatmap showing distribution of ATAC-seq signal across time and unsupervised clustering for all differentially accessible ATAC-seq peaks at any time point. ATAC-seq sites within 1kb from any transcription start site (TSS) are indicated. Mean signal profile for each cluster was plotted to visualize overall pattern of changes. (D) Enrichment of ATAC-seq sites within each cluster at defined genomic locations, calculated as ratio between percent of sites within the region from the cluster vs all ATAC-seq sites. (E) DNA binding motif analysis of peaks within each ATAC-seq cluster. Logos and factor names for the top 3 motifs for all ATAC-seq sites and each cluster are shown. Additional transcription factors with significantly enriched motifs are also listed.
Fig 3.
Transcription factor-mediated regulation of chromatin accessibility.
Publicly available LCL ChIP-seq datasets were used to derive average binding signal of various factors and histones within high confidence ATAC-seq regions. Heatmaps visualize distribution of mean ChIP-seq signal across clusters. Line plots show average signal profiles for each cluster (c1-c5) as well as non-significantly (ns) changed regions for select specific examples. Examples of factors with higher binding within a single cluster (A) and associated line plots (B) are shown, as well as examples of factors with increased binding in multiple clusters (C, D). (E-F) Distribution of ChIP-seq signal across ATAC-seq clusters for EBNA1, EBNA2, and EBNA3C.
Fig 4.
Transcriptomic analysis of EBV-mediated B-cell reprogramming.
Analysis of gene expression changes over the time course of infection was performed by RNA-seq and differentially expressed genes at any time point were identified. (A) Unsupervised clustering of differentially expressed genes with representative mean expression pattern is shown for each cluster. The top 5 most changed genes along with their respective relative expression levels are also shown for each cluster. (B) Pathways significantly enriched among all differentially expressed genes. Pathways significantly enriched within specific gene clusters are indicated. (C) Relative expression level of genes associated with the nucleotide metabolism pathway. (D) Volcano plot showing differentially expressed genes at Day 21 post infection with examples of genes continuously differentially expressed across the time course highlighted.
Fig 5.
Integration of ATAC-seq and RNA-seq and EBNA ChIP-seq datasets.
(A) Transcriptomic data was integrated with ATAC-seq data by generating sequential subsets of genes based on their association with differentially accessibly chromatin, and the associated gene expression pattern. Genes were considered directly correlated if there was correlation (r > 0.5) between changes in gene expression and changes in chromatin accessibility. (B) Pathway analysis was done on the subset of directly correlated genes, with the Top 20 associated pathways shown. (C) Venn diagram indicating the association between directly correlated genes and EBNA1, EBNA2, and EBNA3C DNA binding locations. Genes within the grey cluster show no associated EBNA binding, while those in the colored circles have ChIP-seq peaks for the indicated EBNA protein at the TSS and/or enhancer regions. (D-E) EBNA1-associated directly correlated genes were ordered based on the strength of the associated EBNA1 peak (D), and plotted based on the strength of associated EBNA1 peak versus the strength of the correlation between the RNA expression and ATAC-seq signal patterns (E).
Fig 6.
Metabolomic analysis of EBV-mediated B-cell reprogramming.
(A-C) Significantly altered metabolites and genes (|FC| >2, FDR < 5%) were used in an integrated pathway analysis to identify altered metabolic pathways at Day 2 (A), Day 7 (B) and Day 21 (C) post-infection. Green marker was used to highlight purine metabolism. (D) Bubble chart showing significant disruption of purine metabolism compared to the average significantly altered pathway. Bubble size represents the ratio of significantly altered metabolites and genes in the pathway compared to the number expected due to chance. (E) Heatmap showing relative levels of metabolites in the purine metabolism pathway. Metabolites are grouped based on general trend of metabolite levels across the complete time course of infection.
Fig 7.
Role of EBNA1 in transcription activation of ADA and AK4 during primary infection.
(A-B) Alignment of ATAC-seq data and EBNA1, EBNA2 and EBNA3C ChIP-seq data for ADA (A) and AK4 (B). ATAC-seq data is overlaid for each patient showing Day 0 (black), Day 2 (green), Day 7 (yellow), and Day 21 (red). (C-D) Normalized RNA-seq reads showing expression levels at each time point for ADA (C) and AK4 (D). (E-F) RT-qPCR analysis of ADA levels during primary B-cell infection with EBV (E) or in EBV+ cell lines. (G) EBV M81 infection of NP-TERT cells at day 3, 4, or 7 analyzed by RT-PCR for ADA expression relative to GAPDH. (H) NP-TERT cells with Dox-inducible EBNA1 were treated without (0) or with Dox for 1, 2, 3, or 5 days and assayed by Western blot for EBNA1 or loading control β-Actin. (I) iEBNA1 NP-TERT cells were induced with Dox for 0, 2, 4, or 7 days and assayed by RT-qPCR for ADA expression relative to GAPDH. *, p<0.05; ***, p<0.001;****, p < .0001 p values determined by two-tailed t-test; data represents a minimum of 3 independent experiments.
Fig 8.
Requirement for ADA activation during EBV immortalization.
(A) ADA expression was knocked down by transduction with lentivirus shRNA in EBV- BJAB or LCL cells. Relative proliferation was determined by CellTiter-Glo assay and normalized to shControl-transduced cells. Western blot for ADA and loading control GAPDH (lower panel). (B) Reseeding efficiency was compared between LCLs from ADA-associated SCID patients and their phenotypically normal parents (*, p<0.05; **, p<0.01; ***, p<0.001; ns, not significant; p values determined by two-tailed t-test; data represents a minimum of 3 independent experiments). (C) EBV+ and EBV- cell lines were treated with serial dilutions of the ADA inhibitor, EHNA, for 72hrs and the IC50 value for each cell line was determined. Data represents two independent experiments with 3 technical replicates per experiment. (D) Same as in panel C, except using Pentostatin as ADA inhibitor. (E) Primary B-cells infected with EBV in the absence (blue bar) or presence of increasing concentrations of EHNA and assayed at day 2 for ATP levels by CellTiterGlo. In parallel, primary B-cells were treated without or with IL-4/CD40L and assayed with CellTiterGlo. (F) Primary B-cells treated with IL-4/CD40L in the absence or presence of increasing concentrations of EHNA and assayed for ATP levels by CellTiterGlo. *, p<0.05; ***, p<0.001; ****, p < .0001; p values determined by two-tailed t-test; data represents a minimum of 3 independent experiments.