EBNA1-mediated recruitment of a histone H2B deubiquitylating complex to the Epstein-Barr virus latent origin of DNA replication

Abstract

The EBNA1 protein of Epstein-Barr virus (EBV) plays essential roles in enabling the replication and persistence of EBV genomes in latently infected cells and activating EBV latent gene expression, in all cases by binding to specific recognition sites in the latent origin of replication, oriP. Here we show that EBNA1 binding to its recognition sites in vitro is greatly stimulated by binding to the cellular deubiquitylating enzyme, USP7, and that USP7 can form a ternary complex with DNA-bound EBNA1. Consistent with the in vitro effects, the assembly of EBNA1 on oriP elements in human cells was decreased by USP7 silencing, whereas assembly of an EBNA1 mutant defective in USP7 binding was unaffected. USP7 affinity column profiling identified a complex between USP7 and human GMP synthetase (GMPS), which was shown to stimulate the ability of USP7 to cleave monoubiquitin from histone H2B in vitro. Accordingly, silencing of USP7 in human cells resulted in a consistent increase in the level of monoubquitylated H2B. The USP7-GMPS complex formed a quaternary complex with DNA-bound EBNA1 in vitro and, in EBV infected cells, was preferentially detected at the oriP functional element, FR, along with EBNA1. Down-regulation of USP7 reduced the level of GMPS at the FR, increased the level of monoubiquitylated H2B in this region of the origin and decreased the ability of EBNA1, but not an EBNA1 USP7-binding mutant, to activate transcription from the FR. The results indicate that USP7 can stimulate EBNA1-DNA interactions and that EBNA1 can alter histone modification at oriP through recruitment of USP7.

Figure 1. EBNA1 binding to DNA is stimulated by USP7.

A. Schematic representation of EBNA1 and the EBNA1 mutants used in this study. Shown are the Gly-Ala repeat (GA), the large Gly-Arg repeat (GR), the USP7 binding site (USP7) and the flanking and core DNA binding domains. B and C. EMSAs showing titrations of EBNA1 (B) or EBNA1_452–641 (C) with a fixed amount of DNA recognition site in the presence or absence of 10 pmols of USP7 or in the presence or absence of 10 pmols BSA as a negative control (B, right panel).

Figure 2. Analyses of the USP7 effect on DNA interactions of EBNA_395–641.

A. EMSAs showing titrations of EBNA_395–641 with a fixed amount of DNA recognition site in the presence or absence of 10 pmols of USP7 (left panel) or USP7-NTD (right panel). B. EMSAs performed with a fixed amount of EBNA_395–641 and DNA and the indicated amounts of USP7 (left panel). Titrations of USP7 and the USP7-NTD with DNA in the absence of EBNA1 are shown in the right panel. C. Complexes of EBNA_395–641 and DNA were preformed then incubated with the indicated increasing amounts of USP7. Complexes formed as in lanes 1,2 and 7 were then incubated with anti-EBNA1 antibody (R4) prior to polyacrylamide gel electrophoresis.

Figure 3. Effects of USP7 silencing on EBNA1-DNA interactions in vivo.

A. CNE2Z cells were treated with siRNA against USP7 or GFP then co-transfected with pLacZ and with an oriP plasmid expressing EBNA1 or Δ395–450 as indicated or empty oriP plasmid (oriP). Equal amounts of cell lysates were analysed for protein expression by Western blotting (left panel) and ChIP assays were performed with EBNA1 and nonspecific antibodies for the DS and FR elements of oriP and for the lacZ gene. Results are shown after normalization to nonspecific IgG and input DNA. B. D98/Raji cells were transfected with siRNA against USP7 or GFP then ChIP assays were performed with EBNA1 antibodies and nonspecific antibodies (IgG) and a primer set near region III. Changes with P values less than 0.01 (**) and less than 0.05 (*) relative to siGFP samples are indicated.

Figure 4. Chromatin IP assays for USP7, GMPS and Ub-H2B in EBV genomes.

A. ChIP experiments were performed in Raji cells using antibodies against EBNA1 (left panel), USP7 (middle panel), GMPS (right panel) and nonspecific rabbit IgG as a negative control. Recovered DNA fragments were quantified by real-time PCR using primer sets for the oriP DS and FR regions or the BZLF1 promoter region. B. D98/Raji cells were treated with siRNA against USP7 or GFP (negative control), then ChIP experiments were performed as in A using antibodies against GMPS (right panel). Down-regulation of USP7 by siUSP7 treatment was confirmed by Western blotting, while GMPS levels were unaffected by this treatment (left panel). C. D98/Raji cells were treated with siRNA against USP7 or GFP and ChIP assays were performed using antibodies against histone H2B and monoubiquitylated histone H2B (Ub-H2B) and primer sets for the indicated region of the EBV genome (LMP = LMP1 promoter region). Relative ratios of Ub-H2B to total H2B were determined for each treatment and the average fold increase in Ub-H2B after siUSP7 treatment (as compared to siGFP treatment) from multiple experiments is shown.

Figure 5. USP7 forms a complex with GMPS.

A. Purified USP7 was coupled to a fixed amount of resin at the indicated concentrations to generate a series of affinity columns. A constant amount of HeLa whole cell lysate was applied to each, followed by washing then elution of the bound proteins with 1 M NaCl then with 1% SDS. A silver-stained gel is shown in which the band marked by the arrow was excised and identified as GMPS by MALDI-ToF mass spectrometry. The band at 120 kDa in the 1% SDS elution is USP7 itself. B. Purified USP7 and GMPS were analysed by glycerol gradient sedimentation individually (top and middle panels) and after mixing the two proteins (bottom panel). Equal volume fractions were collected from the top of each gradient and analysed by SDS-PAGE and colloidal Coomassie staining. The positions of 158 kDa (aldolase) and 232 kDa (catalase) molecular weight markers are indicated at the top of the gels.

Figure 6. GMPS stimulates histone H2B deubiquitylation by USP7.

A. Total histones isolated from HeLa cells were incubated with USP7 (1∶1000 ratio of USP7∶histones) for 0, 1, 5 or 30 minutes then analysed by Western blotting using antibodies against histones H2B (left panel) or H2A (right panel). The positions of the unmodified (H2B/H2A) and monoubiquitylated (mUb) histones are indicated. B. Total histones were incubated with USP7 as in A for the indicated number of minutes, with (USP7+GMPS) or without (USP7) GMPS, at a ratio of 1∶1, 1∶10 or 10∶1 USP7∶GMPS as indicated. Western blot analysis was then performed using anti-ubiquitin antibody and the band corresponding to monoubiquitylated H2B in part A is shown. C. Polyubiquitylated p53 was incubated with USP7 for the indicated number of minutes with (USP7+GMPS) or without (USP7) a 10-fold excess of GMPS. Samples were analysed by Western blotting using p53 antibody. The positions of unmodified (p53) and ubiquitylated p53 (Ub-p53) are indicated. D. HeLa cells were transfected with siRNA against GFP or USP7 and USP7 silencing was confirmed by Western blotting of whole cell extracts as compared to an actin loading control (top two gel panels). Total histones were prepared from the siRNA treated cells and Western blots were performed using antibodies against histones H2B or H2A (bottom two gel panels). The ratio of the monoubiquitylated to unmodified forms was determined for H2A and H2B and the results from multiple experiments are shown in the histogram, in relationship to the ratio observed with siGFP treatment (set to 1).

Figure 7. GMPS can form a quaternary complex with USP7, EBNA1 and DNA.

The indicated combinations of EBNA1_395–641 , USP7 and GMPS were preincubated then combined with the DNA containing the EBNA1 recognition site and EMSAs were performed as in Figure 2C. Excess amounts of USP7 alone or USP7 and GMPS were used relative to EBNA1_395–641. In lane 6, the USP7-EBNA complex was preformed prior to the addition of GMPS then DNA. The positions of complexes formed by EBNA1 alone, EBNA1+USP7 and EBNA1+USP7+GMPS are indicated by arrowheads 1, 2 and 3 respectively. DNA incubated with the same amounts of GMPS, USP7 or GMPS+USP7 but in the absence of EBNA1 are also shown (lanes 8–10).

Figure 8. Effect of USP7 silencing on EBNA1-mediated transcriptional activation.

CNE2Z cells were treated with siRNA against USP7 or GFP then were co-transfected with an FR-CAT reporter plasmid and an oriP plasmid expressing EBNA1, Δ395–450 or no EBNA1 (oriP). CAT assays were then performed on equal amounts of cell lysates and the percent of acetylated substrate was determined as a measure of transcriptional activation. Changes with P values less than 0.01 (**) and less than 0.05 (*) relative to siGFP samples are indicated.