Fig 1.
ChIP-seq analysis of MHV-68 and KSHV epigenomes in tumor-derived B-cell lines.
ChIP-seq coverage for H3K4-me3 and H3K27-me3, or corresponding input coverage across the KSHV genome in BCBL1 (A) or the MHV-68 genome in S11E cells (B) for the indicated antibodies. Regions on the MHV-68 genome enriched for H3K4-me3 as detected using MACS14 are indicated by black bars. Asterisks indicate likely false positives also present in the IgG control (S1A Fig). Hashed boxes indicate repetitive regions (including the terminal repeats) that were excluded from the analysis as they do not allow unique read mapping.
Fig 2.
RNA-seq analysis of S11E and de novo infected MLE12 cells.
(A) RNA-seq analysis of persistently infected S11E cells (upper panel), de novo MHV-68 infected MLE12 cells at 12 hours post infection (center panel) or GFP-sorted MLE12 cells which had been infected with MHV-68Δ50 for more than 3 weeks (lower panel). RNA sequencing was performed using a strand-specific sequencing protocol and resulting paired-end RNA-seq reads were mapped to the MHV-68 reference sequence (NC_001826) using the splice-sensitive STAR pipeline (see Material and methods for details). Coverage tracks depict mean coverage across 100 bp binning windows. Forward and reverse strand coverage is shown in the upper and lower plots of each panel. (B) Heatmaps depicting normalized read coverage across individual MHV-68 ORFs annotated in the NC_001826 GenBank entry for the experiments shown in A.
Fig 3.
ChIP-seq analysis of KSHV and MHV-68Δ50 epigenomes in SLKp or MLE12 cells.
ChIP-seq results from (A) KSHV infected SLKP cells or (B) long-term MHV-68Δ50 infected MLE12 cells and (C) MLE12 cells at day 5 post infection with MHV-68Δ50. Solid bars underneath H3K4-me3 plots indicate peak regions as detected by MACS14. Regions that were uniquely enriched at either day five or in long-term latency are marked with an asterisk (*) in each panel. Hashed boxes above the MHV-68 and KSHV genomes map indicate repetitive regions excluded from the analysis, and the dotted box above MHV-68 indicates the deletion of ORF50 coding sequences in MHV-68Δ50.
Fig 4.
Statistical analysis of H3K27-me3 levels acquired by viral episomes.
Shown is an analysis of two biological replicates collected from MHV-68Δ50-infected MLE12 cells (left panels) or KSHV-infected SLKp cells (right panels) each. The graphs depict average enrichment of H3K27-me3 levels in positive and negative control regions of the host genomes, relative to average enrichment across the viral genome. H3K27-me3 positive (host pos) and negative (host neg) host regions (200 each) were detected by SICER/EPIC2, and enrichment across viral sequences was calculated using a 10kb sliding window (see Methods section for details). For each region we individually calculated the H3K27-me3 to input read count ratio and normalized all groups to the average of the respective negative control. Data are shown as box-whisker-plots with 5th-95th percentile and median (+). Significance was calculated by 1way ANOVA testing (MLE-12: F = 64.36, df = 817; SLKP: F = 341, df = 770). Significance is indicated by asterisks or ns (not significant).
Fig 5.
ChIP-seq analysis of KSHV and MHV-68Δ50 epigenomes in superinfected MLE12 cells.
(A) H3K4-me3 and H3K27 ChIP-seq coverage across MHV-68Δ50 (top) and KSHV (bottom) genomes in long-term MHV-68Δ50-infected MLE12 cells superinfected with KSHV for five days. Read counts in all samples were normalized to input DNA to correct for differences among KSHV and MHV-68Δ50 episome copy numbers per cell. (B) Statistical analysis of H3K27-me3 enrichment in two replicates of MHV-68Δ50-positive MLE12 cells superinfected with KSHV. Analysis was performed analogous to that shown in Fig 4. Data are shown as box-whisker-plots with 5th-95th percentile and median (+). Significance was calculated by 1way ANOVA testing (F = 270.8, df = 889). Significance is indicated by asterisks or ns (not significant). (C) Confirmatory ChIP-qPCR using KSHV-, MHV-68- and mouse-specific primers as indicated (n = 3). Positive controls were as follows: H3K4-me3: GAPDH, RPL30; H3K27-me3: MYT1; H3K4-me3/H3K27-me3 (bivalent chromatin): PITX1. Data are represented as mean ± SEM.
Fig 6.
kLANA-expressing MHV-68Δ50 genomes do not gain the ability to rapidly recruit H3K27-me3 marks.
(A) H3K27-me3 ChIP-seq coverage across the MHV-68Δ50 genome in MLE12 cells sorted for GFP expression at day 5 post infection with MHV-68Δ50-kLANA (top) or MHV-68Δ50 (bottom). Due to the low titers of the MHV-68Δ50-kLANA virus, ChIP-seq was performed for both cultures using a low-cell ChIP protocol as described in the material and methods section. (B) Relative H3K27-me3 enrichment analysis for the data shown in panel A, performed as described in the legend to Fig 4. Significance was calculated by 1way ANOVA testing (F = 640.7, df = 781) and is indicated by asterisks or ns (not significant). (C) Confirmatory ChIP-qPCR analysis of MHV-68Δ50-kLANA or MHV-68Δ50-infected MLE12 cells at 5 days post infection (n = 1, top panel), or after a 35 day period during which cells were repeatedly sorted to achieve a population of 100% GPF positive cells (n = 2, bottom panel). Data are represented as mean ± SEM.(D) Western blot analysis of mLANA and kLANA expression in cells infected with MHV-68Δ50-kLANA or MHV-68Δ50, respectively. Unspecific bands are indicated by asterisks.
Fig 7.
KSHV-BAC16Δ73 genomes trans-complemented by mLANA do not lose the ability to rapidly recruit H3K27-me3 marks.
(A) ChIP-qPCR analysis of H3K4-me3 and H3K27-me3 in KSHV-BAC16Δ73-infected MLE12 cells that had been stably transduced with lentiviral mLANA or kLANA expression constructs. Cells were cultured in the presence of hygromycin to enrich for KSHV-BAC16Δ73-infected cells and ChIP was performed at day 5 (n> = 2, top panel) or day 33 post infection (n> = 2, bottom panel). Data are represented as mean ± SEM. (B) H3K27-me3 and H3K4-me3 ChIP-Seq analysis of material harvested after 33 days from KSHV-BAC16Δ73-infected kLANA (top panels) or mLANA (bottom panels)-expressing MLE12 cultures. Lower input coverage across the left half of the KSHV genome in MLE12-mLANA cells (see bottom panel) suggests loss of sub-genomic material from a fraction of episomes, similar to what has recently been observed with KSHV mutants expressing an oligomerization-deficient kLANA protein [26]. For visualization purposes ChIP-seq tracks were therefore normalized for using mean input coverage values. Non-normalized coverage data as used for statistical analysis is given in S1 Dataset. (C) Statistical analysis of H3K27-me3 enrichment on BAC16Δ73 episomes calculated as described in the legend to Fig 4. Significance was calculated by 1way ANOVA testing (F = 431.2, df = 838). The difference between BAC16ΔLANA in kLANA vs. mLANA expressing cells was not significant (ns). (D) Western blot analysis of mLANA and kLANA expression in stably transduced MLE12 cells.
Fig 8.
Histone-modification patterns of latent MHV-68 genomes in vivo.
(A) ChIP-qPCR analysis of H3K27-me3 (top), H3K4-me3 (center) or IgG (negative control, bottom) in splenocytes harvested from two mice (#1 and #2) that had been intranasally infected with wildtype MHV-68 for 17 days, using MHV-68 or endogenous control primers as indicated. Data are represented as mean ± SEM of three independent ChIP replicates performed with the isolated chromatin from each mouse. (B) H3K27-me3 ChIP-seq coverage across the MHV-68 genome in two pools of splenocytes isolated from a total of six mice that had been intranasally infected with MHV-68-H2BYFP for 17 days. Splenocytes were FACS sorted for YFP expression prior to analysis. Due do the low number of positive cells, YFP-positive cells from three mice were pooled to generate pools #1 and #2. Approximately 5000 cells were subjected to low cell ChIP-seq using an H3K27-me3 specific antibody. Hashed boxes above the MHV-68 map indicate the position of repetitive regions (left and right internal repeat regions, as well as terminal repeat sequences) which had to be masked since they do not allow unique read mapping. (C) Relative enrichment of H3K27-me3 at viral episomes in splenocyte pools #1 and #2 was assessed using the same statistical method as described in the legend to Fig 4. (D) Normalized mean H3K27-m3 and input coverage from all MHV-68 ChIP-seq experiments performed in our study. Quantile normalized mean values and standard deviation (colored area) of H3K27-me3 and input tracks were generated from all MHV-68-specific coverage tracks as given in S1 Dataset (ChIP MHV-68).
Fig 9.
KDM2B recruitment and acquisition of PRC1-associated histone modifications by KSHV and MHV-68Δ50 genomes.
(A) KDM2B ChIP-seq coverage (top) or input (bottom) profiles across the KSHV genome in SLK cells at 24 hours post-infection. (B) Relative enrichment of KDM2B on de novo infecting KSHV episomes measured at 24 hours post infection. Enrichment was quantified similar to the method described in the legend to Fig 4. KDM2B enriched human positive control regions were detected by MACS peak calling. Significance was calculated by 1way ANOVA testing (F = 372.5, df = 508). (C) Heatmap of KDM2B enrichment at transcriptional start sites (TSS) of human genes generated from KDM2B ChIP-seq data at 24 hours post infection with KSHV infection. The Y-axis represents 70,474 individual length normalized annotated transcripts (from TSS to TTS) of human genes, and the X-axis represents the 200% surrounding of the length-normalized genes. Data was sorted according to decreasing KDM2B signal at the TSS. (D) Global anti-correlation of cellular H3K27-me3 and H3K36-me2 patterns across the human genome. The 2D histogram shows the number of genomic 10000 bp windows with the combination of the number of reads described on the axes in SLKp cells. The colored bar to the right side illustrates the number of windows with each combination of reads in the two datasets. (E+F) ChIP-seq coverage for H3K36-me2 (top) and H2AK119-ub (center) or input (bottom in each panel) from (E) stably KSHV-infected SLKP cells or (F) long-term MHV-68Δ50-infected MLE12 cells.
Fig 10.
Acquisition of H2AK119-ub, H3K27-me3 and H3K36-me2 by KSHV and MHV-68Δ50 genomes.
Analysis of significant enrichment of H2AK119ub, H3K27me3 and H3K36me3 across the KSHV genome in (left panel) SLK cells after 24 hours of infection, (2nd panel from left) SLK cells after 5 days of infection or (2nd panel from right) SLKp cells, or (right panel) the MHV68 genome in long-term MHV68-Δ50-infected MLE12 cells. Based on the statistical method described in the legend to Fig 4, normalized enrichment was calculated relative to the median value of the host loci that showed strongest and weakest enrichment (set to 100% and 0%, respectively) for each modification. Significance indicators (asterisks) are shown for those modifications for which median enrichment along the viral genome was significantly above levels observed for the cellular background. ‘nd’ (not detectable) demarks modifications in which enrichment was not significantly different from (or significantly lower) than in the negative control regions. Data from MLE12 and SLKp cells correspond to those shown in Fig 9E and 9F, or those in Fig 4 for H3K27-me3. Coverage tracks for SLK cells at 24 h.p.i and 5 d.p.i. are provided in S7 Fig. Non-normalized statistical data for each of the modifications and cell lines shown in this figure are provided in S8 Fig.
Fig 11.
Model of PRC1 and -2 recruitment by KSHV and MHV-68 genomes.
Immediately after nuclear entry, sequence-specific transcription factors bind to KSHV as well as MHV-68 genomes and lead to deposition of activation-associated H3K4-me3 marks. In KSHV (left panel), high density of unmethylated CpG motifs mediates rapid acquisition of the non-canonical PRC1.1 complex, followed by PRC2 recruitment as a secondary event (alternatively, PRC2 might also be directly recruited to CpG-rich DNA). Owing to the lower CpG frequency of MHV-68 genomes (right panel), only relatively short sequence segments which exhibit characteristics of CpG islands (such as the internal repeat regions) initially acquire PRC1.1 complexes in a delayed (depicted) or stochastic manner. Once established, canonical PRC occupancy and associated histone modifications may slowly spread from these sites.
Table 1.
Primers used in this study.