ATM, KAP1 and the Epstein-Barr virus polymerase processivity factor direct traffic at the intersection of transcription and replication

Abstract

The timing of transcription and replication must be carefully regulated for heavily-transcribed genomes of double-stranded DNA viruses: transcription of immediate early/early genes must decline as replication ramps up from the same genome-ensuring efficient and timely replication of viral genomes followed by their packaging by structural proteins. To understand how the prototypic DNA virus Epstein-Barr virus tackles the logistical challenge of switching from transcription to DNA replication, we examined the proteome at viral replication forks. Specifically, to transition from transcription, the viral DNA polymerase-processivity factor EA-D is SUMOylated by the epigenetic regulator and E3 SUMO-ligase KAP1/TRIM28. KAP1's SUMO2-ligase function is triggered by phosphorylation via the PI3K-related kinase ATM and the RNA polymerase II-associated helicase RECQ5 at the transcription machinery. SUMO2-EA-D then recruits the histone loader CAF1 and the methyltransferase SETDB1 to silence the parental genome via H3K9 methylation, prioritizing replication. Thus, a key viral protein and host DNA repair, epigenetic and transcription-replication interference pathways orchestrate the handover from transcription-to-replication, a fundamental feature of DNA viruses.

Graphical Abstract

Figure 1.

EBV proteins are enriched at viral DNA forks. (A) Schematic of the EBV replication fork. EBV genome replication during the lytic phase originates at one of two lytic origins of replication and requires a primase helicase complex comprised of a primase (BSLF1), primase-associated factors BBLF2 and BBLF3, and a helicase (BBLF4). BSLF1 synthesizes RNA primers that accumulate on the lagging strand for the initiation of DNA synthesis. The viral major DNA binding protein (BALF2) further unwinds DNA and stabilizes ssDNA within the replication fork while the DNA polymerase (BALF5) catalyzes the synthesis of new viral DNA in association with early antigen-diffuse (EA-D, BMRF1) which enhances the processivity of the polymerase complex. (B–E) EBV proteins identified by iPOND-mass spectrometry (MS). B shows a volcano plot of log2 fold change (log₂FC) versus –log₁₀ (P-value) for each protein enriched by iPOND. Green indicates proteins enriched in iPOND with log₂FC > 0.5 over background. Components of the cellular pre-replication complex (MCM2-7) and EBV DNA replication fork proteins are labeled. (C) Heat map of normalized spectral abundance factors (NSAF) for all viral proteins meeting the above cut-off are shown as the average of two technical replicates for each of six biological replicate samples as well as the overall mean (last column). (D) Volcano plot of log₂FC versus –log10 (P-value) for proteins enriched by iPOND in thymidine chase samples compared to background (red). (E) Volcano plot of the ratios of all significant proteins (P-value < 0.05) in viral iPOND samples to thymidine chase samples expressed as log2FC showing greater enrichment in viral iPOND samples (green) versus thymidine chase samples (red). (F) EBV proteins at replication forks using iPOND-immunoblotting; 293-EBV cells were transfected with HA-BZLF1, HA-BRLF1 and FLAG-tagged EBV ORF plasmids for 36 h and pulsed with EdU for 20 min prior to harvest and iPOND for active forks (EdU). For isolating mature newly replicated viral DNA (Thymidine), EdU-labeled cells were washed with thymidine containing medium and then incubated with thymidine for 30 min prior to performing iPOND and immunoblotting. (G, H) 293-EBV cells were transfected with HA-BZLF1 (Z) and HA-BRLF1 (R) for 36 h and analyzed by immunoblotting (G) or qPCR to measure EBV genome replication; error bars, SEM (H). Experiments were performed three times.

Figure 2.

Pulse and pulse chase analysis showing cellular proteins at active viral forks and newly replicated linear viral genomes, respectively. (A) Normalized iPOND (231 proteins) and iPOND Thymidine chase (229 proteins) with P-values <0.05 and enriched over background (no click and PAA treated) by log₂ fold change >0.5. Many active replication fork proteins (green, 231) overlapped with iPOND Thymidine chase proteins (168 proteins), while proteins associated with newly replicated linear viral DNA appeared exclusively in the iPOND Thymidine chase set (pink, 61 proteins). Statistically significant biological processes for Active Forks (B) versus New Viral DNA (C) were determined by overrepresentation analysis (ORA) and plotted according to the Enrichment ratio score for each gene set and biological process using WebGestalt 2019; FDR, false discovery rate. (D) Putative functional and physical interactomes of EBV iPOND enriched cellular proteins in the active forks dataset. Statistically significant EBV iPOND cellular proteins were subjected to STRING analysis to identify high confidence interactions (depicted). Functional pathways (circled) include DNA replication & repair, chromatin modification & transcriptional repression, RNA processing and proteasome. (E) Cross-analysis depicting putative functional and physical interactomes of 21 cellular proteins in common at EBV replication forks (from D) and the previously published nascent HSV-1 genome in an infection model (20). Edges (lines) in (D) and (E) indicate functional links or experimentally observed interactions (light green, text mining; cyan, database links; magenta, experimental evidence; purple, homology).

Figure 3.

KAP1 conjugates SUMO2 to polymerase processivity factor EA-D at K135 during DNA replication. (A, B) BL cells (A) and LCL (B) were treated with NaB with or without PAA for 36 h and analyzed using iPOND for active forks (EdU) and mature newly replicated DNA (using thymidine) and immunoblotted with indicated antibodies. Click group was processed with biotin azide. (C) BL cells exposed to NaB for 36 h were then labeled with BrdU for 2 h. Harvested KAP1^hi and KAP1^−/lo cells (left two panels) were assayed for BrdU uptake (right two panels) using flow cytometry. (D) HEK-293T cells were transfected with FLAG-BMRF1 (EA-D), FLAG-KAP1, FLAG-SENP1 and HA-SUMO2. After 48 h, lysates were immunoprecipitated with anti-EA-D antibody prior to immunoblot analyses. (E) 293-EBV cells were co-transfected with siRNA targeting KAP1 alongside FLAG-BZLF1 and FLAG-BRLF1 to reactivate EBV. After 36 h, lysates were immunoprecipitated with anti-SUMO2 antibody or control IgG and then immunoblotted. (F) HEK-293T cells were transfected with wild type FLAG-BMRF1 (WT) or mutants (K135R, K212R) together with FLAG-KAP1 and HA-SUMO2 for 48 h. Lysates were immunoprecipitated with anti-EA-D antibody or control IgG and immunoblotted as indicated. (G, H) BL cells were treated with NaB for 24 h and lysates were immunoprecipitated with anti-EA-D antibody versus control IgG (G) or anti-SUMO2 antibody versus control IgG (H), and then immunoblotted as indicated. (I–K) To assess proximity by proximity ligation assay (PLA), LCL were exposed to NaB without (I, J) or with (K) the addition of PAA for 36 h; lytic cells were marked using an anti-ZEBRA antibody. (L) LCL were exposed to NaB with or without the addition of PAA for 36 h prior to immunoprecipitation using an anti-KAP1 antibody or control IgG followed by immunoblotting. (M) BL cells were exposed to NaB for different lengths of time prior to immunoprecipitation with anti-EA-D antibody versus control IgG and then immunoblotted. Experiments were performed three times.

Figure 4.

Recruitment of KAP1 to EA-D at replication forks requires transcription. (A) BL cells were exposed to NaB for 30h with α-amanitin added for the last 6h. Cells were then harvested for iPOND followed by immunoblotting. YC, yes click group, processed with biotin azide; NC, no click control group, processed without biotin azide. (B, C, E, F) EBV in BL cells was reactivated using NaB for 24 h while blocking transcription using α-amanitin for the last 6 h. Cells were harvested for immunoprecipitation with anti-EA-D antibody (versus IgG as control; (B), immunoblotting (C), quantitation of viral genomes by qPCR (E) and quantitation of immediate early and early gene transcripts via RT-qPCR (F). In panel D, BL cells were exposed to NaB for 30 h with α-amanitin added for the last 6h. DNA was precipitated using antibody to γH2AX and exposed to MboI to digest newly replicated non-methylated DNA. Primer pairs spanning MboI sites were used to amplify indicated loci on episomal genomes using qPCR. Efficiency of MboI digestion was assayed by qPCR using a pair of primers spanning an MboI site on the BALF5 gene (middle and right-hand side panels). Error bars, SEM of three technical replicates. Experiments were performed twice.

Figure 5.

KAP1 mediated SUMOylation of EA-D stabilizes viral DNA replication forks. (A) EA-D/BMRF1-deleted EBV bacmid-carrying 293-EBV (B072-1) cells were transfected with wild type FLAG-BMRF1 (WT) or mutants (K135R, K212R) together with FLAG-BZLF1 and FLAG-BRLF1 to reactivate EBV. After 36 h, viral genomes were quantified via qPCR. (B–E) 293-EBV cells were transfected with siKAP1 (versus siCtrl) and FLAG-BZLF1 and FLAG-BRLF1 for 36 h and extracted DNA subjected to DNA fiber fluorography (B–D) or lysates immunoblotted with indicated antibodies (E). Representative images from DNA fiber analysis (B), fiber counts (C) and cumulative data including median value of track/fiber length from at least 150 tracks per experimental condition (D) are shown. *P < 0.05, ***P < 0.001. (F) BMRF1-KO 293-EBV cells were transfected with wild type FLAG-BMRF1 (WT) or mutants (K135R, K212R) together with FLAG-BZLF1 and FLAG-BRLF1 plasmids for 36 h prior to performing immunoblotting. Experiments were performed at least twice. Error bars in (A), SEM of three technical replicates.

Figure 6.

The PI3KK ATM recruits and phosphorylates KAP1 at the RPOII-RECQ5 complex. (A, B) BL cells were treated with NaB with or without the ATM inhibitor KU55933 for 24 h. Cells were harvested for immunoblotting (A) or immunoprecipitation with anti-EA-D antibody or control IgG (B). (C, D) BL cells were transfected with siATM or scrambled siRNA (siCtrl) for 20 h followed by exposure to NaB for 24 h. Lysates were immunoprecipitated with anti-RECQ5 antibody (C) or anti-KAP1 antibody (D) and then immunoblotted using indicated antibodies. Experiments were performed twice.

Figure 7.

RECQ5 facilitates EA-D-mediated recruitment of silencing histones to the episomal EBV genome. (A–D) BL cells were transfected with siRECQ5 or scrambled siRNA (siCtrl) for 20h followed by treatment with NaB for 24h. Nuclei were subjected to chromatin immunoprecipitation with anti-CAF1A antibody (A), anti-Histone H3.1 antibody (B) and anti-H3K9me3 antibody (C). Precipitated DNA was exposed to MboI to digest newly replicated non-methylated DNA. Primer pairs spanning MboI sites were used to amplify indicated loci on episomal genomes using qPCR in A-C. Fold enrichment of CAF1A, H3.1 and H3K9me3 on DNA was analyzed after normalizing to control IgG pull down. Efficiency of MboI was assayed by qPCR using a pair of primers spanning an MboI site on the BALF5 gene (D). (E) EA-D/BMRF1-deleted EBV bacmid-carrying 293-EBV (B072-1) cells were transfected with wild type (WT) or mutant (K135R) FLAG-BMRF1 together with HA-BZLF1 and HA-BRLF1 for 36 h prior to immunoprecipitation with an anti-FLAG antibody. (F) BL cells were transfected with siRECQ5 or scrambled siRNA (siCtrl) for 20 h followed by treatment with NaB for 24h prior to immunoprecipitation with an anti-EA-D antibody. (G, I) BL cells were transfected with siRECQ5 or scrambled siRNA (siCtrl) for 20 h followed by exposure to NaB for 24 h and subjected to BrU-RT-qPCR analysis to measure nascent BZLF1 and BMRF1 transcripts (G) or immunoblotting (I). (H) BL cells were transfected with empty vector (EV), FLAG-ATM plasmid, siRECQ5, or scrambled siRNA (siCtrl) for 20 h followed by treatment with NaB for 24 h before immunoblotting with indicated antibodies. Error bars, SEM of three technical replicates. These experiments were performed at least twice.