The Ubiquitin-Specific Protease USP7 Modulates the Replication of Kaposi's Sarcoma-Associated Herpesvirus Latent Episomal DNA

Expression plasmids.

All full-length LANA constructs were in a pcDNA 3.1 background. The numbering of amino acids of LANA corresponds to the KSHV BC-1 sequences (38). An expression vector for full-length LANA was described previously (50). The sequences of PCR primers used in this study, and their numbered designations, are listed in Table 1. The mutant with deletion of amino acids (aa) 971 to 986 was cloned by overlapping PCR with primers 1, 2, 3, and 4. For the cLANA Δ 971-986 construct, primer 5 was used for overlapping PCR instead of primer 1. LANA GST fusion proteins used in glutathione S-transferase (GST) pulldown experiments represented different segments of the C-terminal domain (aa 951 to 1162) of LANA and were described previously (35). Additionally, the construct C6a was generated by PCR using primers 6 and 7. To obtain the cLANA Δ 971-986 mutant in the GST vector backbone, another overlapping PCR was performed with primers 5, 3, 4, and 8.

Mammalian LANA constructs harboring only the N- or C-terminal domain of LANA were cloned with the following primers (in myc epitope-containing pcDNA3myc provided by S. Mittnacht, London, United Kingdom, or HA-containing pVR1255 provided by J. Stewart, Liverpool, United Kingdom): primers 9, 10, 11, 12, 2, and 13. For the cLANA 977-980/AAAA mutant, a potential USP7 binding motif in the C-terminal domain of LANA (aa 977 to 980) was changed to four alanines by site-directed mutagenesis using primers 14 and 15.

For electrophoretic mobility shift assays (EMSAs), cLANA was cloned in pGEX-6p-1 by PCR with primers 13 and 16, with full-length LANA as the template. Mutation of the PYG motif (aa 1065 to 1067) to alanine was introduced by site-directed mutagenesis using primers 17 and 18.

Full-length RRV Orf73 (1,347 bp) of RRV isolate 17577 tagged N terminally with a myc epitope was provided by R. Searles (Portland, OR). The MHV68 Orf73 construct cloned from murine herpesvirus 4 strain 68-infected cells in the HA-tag-containing pVR1255 vector was provided by J. Stewart.

The USP7 expression plasmid expressing GFP-tagged USP7 was kindly provided by Roger Everett, Glasgow, United Kingdom. An N-terminal USP7 construct was cloned based on the full-length construct by PCR with primers 19 and 20 into a GST fusion vector (pGEX-6p-1).

The plasmid containing the TR region of the KSHV genome was described previously (50). This plasmid was used to generate a 1× TR luciferase reporter vector (51).

The cyclin E luciferase reporter vector consisting of the human cyclin E promoter region cloned in front of a luciferase gene (in pGL2Basic) was from R. Weinberg's laboratory (Cambridge, MA) (16). The p53 luciferase reporter construct containing the p53-responsive promoter region of the human mdm2 gene (in pGL2Basic) was provided by M. Dobbelstein (Göttingen, Germany), as was an expression construct for HA-tagged human p53 in a pcDNA backbone.

Cells and cell culture.

HEK 293, HEK 293-T, and p53-deficient H1299 cells were grown in Dulbecco's modified Eagle medium (DMEM) supplemented with 10% fetal calf serum (FCS), 50 IU/ml penicillin, and 50 μg/ml streptomycin at 37°C in a 5% CO₂ incubator.

Suspension cells such as BJAB and BJAB stably infected with KSHV (S. Kati, S. Hübner, T. Rothämel, and T.F. Schulz, unpublished data), as well as the PEL cell lines BC1 and BC3, were grown in RPMI 1640 (containing l-glutamine) supplemented with 20% fetal calf serum (FCS), 50 IU/ml penicillin, and 50 μg/ml streptomycin at 37°C in a 5% CO₂ incubator.

For transfection, adherent cells were grown to subconfluence in six-well plates and transfected with the indicated expression constructs with FuGENE transfection reagent (Roche) according to the manufacturer's instructions.

SDS-PAGE and immunoblotting.

Forty-eight hours after transfection with the indicated expression plasmids, cells were lysed in 300 μl TBST lysis buffer (20 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100) per well. Cleared lysates were subjected to SDS-PAGE, and proteins were detected by Ponceau S staining or immunoblotting.

The following antibodies were used: mouse or rabbit α-GFP (Clontech), mouse α-actin (Millipore), rat α-LANA (ABI), mouse or rat α-HA (Roche), human α-KS serum from KSHV-seropositive patients, and polyclonal rabbit serum detecting GST fusion proteins.

GST pulldown assay.

For GST pulldown experiments, Escherichia coli Rosetta cultures transformed with GST expression plasmids or GST only were grown at 37°C in LB medium plus ampicillin and chloramphenicol. Cultures were induced at an optical density at 600 nm of 0.4 to 0.6 with 1 mM isopropyl-β-d-thiogalactopyranoside (IPTG) and harvested by centrifugation 5 h after induction. Pelleted cells were resuspended in 500 μl phosphate-buffered saline (PBS) with protease inhibitors, sonicated for 1 min on ice, supplemented with 1% Triton X-100, and incubated for 10 min at 4°C. Following centrifugation, the supernatant was incubated with 80 μl glutathione-Sepharose beads (Amersham Biosciences) overnight at 4°C. Beads were washed three times with PBS plus protease inhibitors and analyzed on a Coomassie-stained gel to calculate equal amounts of GST fusion proteins. HEK 293-T cells were transfected and lysed as mentioned above. Cleared lysate (300 μl) was incubated with either GST fusion proteins or GST control protein for 1 h at room temperature (RT). Beads were washed six times with PBS and analyzed by SDS-PAGE and Western blotting.

Coimmunoprecipitation.

HEK 293-T cells were cotransfected with plasmids encoding the interaction partners of interest; the amounts used are indicated in the figure legends. Forty-eight hours after transfection, cells were washed once in cold PBS and lysed in 300 μl of TBST buffer containing protease inhibitors per well of a six-well plate. For precipitation of endogenous proteins, suspension cell lines were also lysed in 300 μl TBST buffer containing protease inhibitors.

The monoclonal antibodies α-GFP (Clontech), α-HA (Roche), and α-LANA (ABI) were immobilized on protein A or G Sepharose beads (Amersham). For precipitation of myc-tagged proteins, c-myc monoclonal antibody-agarose beads (Clontech) were used.

An aliquot of 40 μl cleared cell lysates was taken (input control), and the remaining 260 μl were incubated overnight on a roller at 4°C with 12 μl of beads coupled with antibody to precipitate the protein of interest. Afterwards, samples were centrifuged (1,100 × g, 30 s) and washed extensively with 700 μl TBST (supplemented with 1% sodium deoxycholate, if higher stringency was necessary). Finally, beads were boiled with 30 μl of 5× loading buffer (62.5 mM Tris-HCl, pH 6.8, 2% [wt/vol] SDS, 10% glycerol, 50 mM dithiothreitol [DTT], 0.01% [wt/vol] bromophenol blue) for 3 min and 20 μl of the samples was analyzed by SDS-PAGE. As input control, 1/10 of the input lysate was loaded.

Transient replication assay.

Short-term replication assays were performed with HEK 293 cells cotransfected with 500 ng expression vector containing 1× TR and 250 ng pBluescript (input control) plus 500 ng of LANA expression constructs or the empty vector pcDNA3. Thirty hours after transfection, the cells were split 1:2 and allowed to grow for a further 48 h. Cells were then harvested in 150 μl of lysis buffer per well of a six-well plate (10 mM Tris-HCl, 10 mM EDTA, 0.6% SDS), and duplicate wells, obtained by splitting, were pooled. Chromosomal DNA was precipitated overnight with 0.85 M NaCl. An aliquot was collected from the supernatant to determine the expression levels of the different LANA constructs by immunoblotting. Plasmid DNA was purified from the supernatant using phenol-chloroform extraction and precipitated afterwards with ethanol. Finally, plasmid DNA was resuspended in H₂O. A total of 90% of the DNA was digested with 40 U KpnI and 40 U of DpnI for 48 h, while the remaining 10% was digested with 10 U KpnI to measure input DNA. The digested DNA was separated on a 0.8% agarose gel and analyzed by Southern blotting with a 962-bp fragment of pBluescript used as an alkaline phosphatase-labeled probe (AlkPhos Direct Labeling and Detection System with CDP-Star; GE Healthcare).

EMSA.

IRD700-labeled oligonucleotides (IBA GmbH, Göttingen, Germany) were prepared as follows: 5 μl of each oligonucleotide (LANA BS top, 5′-GAG GCG GCG CGC GGC CCC ATG CCC GGG CGG GAG GCG CCG CAG GCC CCG GCG GCG TCC CCG GC-3′; LANA BS bottom, 5′-GCC GGG GAC GCC GCC GGG GCC TGC GGC GCC TCC CGC CCG GGC ATG GGG CCG CGC GCC GCC TC-3′) was mixed with 2 μl of 10× annealing buffer (200 mM Tris-HCl, pH 7.6, 500 mM NaCl, 100 mM MgCl₂), and the volume was adjusted to 20 μl with H₂O. This annealing reaction mixture was incubated for 5 min at 95°C, 10 min at 65°C, 10 min at 37°C, and 10 min at 20°C and cooled to 4°C. The annealed oligonucleotides were diluted 1:5 for the binding reaction.

For binding, 0.5 to 2 μl of LANA GST fusion proteins was added to 1.5 μl binding buffer (300 mM Tris-HCl, pH 7.9, 500 mM KCl, 100 mM MgCl₂, 10 mM DTT, 10 mM EDTA) as well as 1.5 μl 100% glycerol, 1.5 μl of a 2.5% Tween 20 solution, 0.75 μl poly(dI-dC) (1 mg/ml), 0.75 μl of a 10-mg/ml solution of bovine serum albumin (BSA), and 1 μl of the annealed oligonucleotides. Reactions were made up with H₂O to a 15-μl final volume and incubated for 30 min at RT in the dark. For competition, a 100× to 200× excess of unlabeled probe was added to the band shift mixture. Supershifts were carried out with 4 μl of a polyclonal LANA rabbit serum or a rabbit antibody not reacting with LANA for a control. For supershift, binding reaction mixtures were incubated for another 30 min after addition of the antibodies.

A 4% native acrylamide gel was prerun for 1 h at 160 V in 1× TBE. Afterwards, the band shift reaction mixtures were loaded onto the gel and separated for 2 h at 160 V. After the run, gels were immediately analyzed with the Odyssey infrared imaging system from Licor Biosciences.

Luciferase reporter assay.

For luciferase reporter assays, HEK 293, HEK 293-T, or H1299 cells were transiently cotransfected in duplicates with 50 ng of reporter plasmid and expression constructs as indicated in the figure legends. Forty-eight hours after transfection, cells were washed once with PBS and lysed in 300 μl reporter lysis buffer (Promega) per well of a six-well plate. Luciferase activity was measured in cleared lysates with a luciferase system in accordance with the manufacturer's instructions (Promega).

LANA interacts with USP7.

USP7 was previously reported to interact with two herpesviral proteins, HSV-1 ICP0 and EBV EBNA-1, and to be important for the life cycle of both viruses. Since LANA is a functional homologue of EBNA-1, we investigated whether LANA could also interact with USP7. We transfected HEK 293-T cells with a GFP-tagged expression construct for USP7 as well as an expression vector for LANA and precipitated protein complexes with α-LANA antibody coupled to protein G-Sepharose beads. Following washing of the immunocomplexes, the interaction was analyzed via immunoblotting using an α-GFP antibody. Transfection of empty vector served as a negative control. We found that LANA is capable of interacting with USP7 (Fig. 1A). This observation could be confirmed in a reciprocal coimmunoprecipitation experiment, in which LANA was found in a complex with precipitated USP7 (Fig. 1B). Additionally, the interaction of LANA and USP7 could be verified with coimmunoprecipitation experiments using naturally KSHV-infected PEL cell lines, as well as BJAB cells stably infected with recombinant KSHV virus (Fig. 1C and D). Endogenously expressed (untransfected) USP7 interacted with LANA in KSHV-infected cells, but not in the uninfected control BJAB cells (Fig. 1C), and in all PEL cell lines tested (BC1, BC3) (Fig. 1D).

Mapping of the USP7 binding site of LANA.

Following the observation that LANA interacts with USP7, we investigated which domains of both proteins are involved in the interaction. We started by mapping the USP7 binding site in LANA to the N-terminal or C-terminal domain. USP7 coimmunoprecipitates with the C-terminal domain of LANA but not with the N-terminal domain (Fig. 2A). To further determine the residues of LANA critical for USP7 binding, GST pulldown experiments were performed. Truncated versions of the C-terminal domain of LANA fused to GST (Fig. 2B) were expressed in bacteria, isolated, and purified by binding to GST beads, and their interaction with eukaryotically expressed plasmids of GFP-tagged USP7 or the appropriate empty vector was investigated. In the GST pulldown experiments, USP7 interacted with the constructs C4, C5, and C6 (Fig. 2B and C). In addition, the small LANA fragment C6a, comprising a region of 16 aa, interacted with USP7 (Fig. 2B and D). This region, which is located between aa 971 and 986, contains a PGPS motif (Fig. 3) that corresponds to the consensus sequence P/A-X-X-S for USP7 binding found by Sheng and coworkers (43).

An alignment of the USP7 binding site of LANA with the previously described USP7 binding site in EBNA-1 (19) showed that the consensus motif, as well as further residues of the identified LANA region involved in USP7 binding, are conserved in EBNA-1. Within the fragment C6 (aa 951 to 1006), the sequence of aa 971 to 986 (C6a) is the only one to show significant sequence homology between LANA and EBNA-1 (Fig. 3). These sequence similarities emphasize the importance of these residues for the interaction with USP7.

As depicted in Fig. 4, removal of aa 971 to 986 from the C-terminal LANA domain abolished binding of LANA to USP7 in GST pulldown experiments (Fig. 4A) as well as in coimmunoprecipitation experiments (Fig. 4B). In addition, the substitution of the PGPS motif in LANA (aa 977 to 980) with four alanine residues eliminated the binding of USP7 to the LANA C-terminal domain in coimmunoprecipitation assays (Fig. 4B). This finding confirms the importance of LANA aa 971 to 986, and in particular aa 977 to 980, for binding to USP7.

In the next step, the deletion of the region aa 971 to 986 was introduced into a full-length LANA construct by overlapping PCR. The full-length LANA construct harboring the deletion of the potential USP7 binding site (LANA Δ 971–986) showed a strongly reduced binding to USP7 (Fig. 4C). The signal intensities of the immunoprecipitated USP7 bands in Fig. 4C (second panel from bottom) were quantified with ImageJ and normalized to the upper LANA band in the bottom panel of Fig. 4C, indicating a residual interaction of 29% (data not shown). This residual binding cannot be attributed to an additional USP7 binding site in the N-terminal domain of LANA, as the N-terminal domain of LANA was not found to interact with USP7 (Fig. 2A).

LANA binds to the N-terminal domain of USP7.

Most USP7 substrates and interaction partners like Mdm2, Mdmx, p53, and EBNA-1 were shown to bind to the N-terminal TRAF domain of USP7 (18, 21, 39, 40, 43). In contrast, the HSV-1 protein ICP0 interacts with the C-terminal part of USP7, a domain of unknown function (18). We investigated whether LANA is able to interact with the N-terminal domain of USP7 in GST pulldown experiments. The N-terminal domain of USP7 was expressed as a GST fusion protein in bacteria and purified via the GST tag. Lysates of cells transiently transfected with expression constructs of LANA or LANA Δ 971–986 were incubated with adjusted amounts of GST-USP7 protein coupled to beads. After washing, protein interaction was analyzed by immunoblotting. These experiments revealed that LANA can interact with the N-terminal TRAF domain whereas the LANA mutant lacking the USP7 binding site (aa 971 to 986) is not able to bind to the N-terminal domain of USP7 (Fig. 4D). This result suggests that LANA resembles EBNA-1, p53, Mdm2, and MdmX in targeting the N-terminal TRAF domain of USP7 and underlines the importance for the LANA region (aa 971 to 986) for interaction with the TRAF domain of USP7.

Deletion of aa 971 to 986 increases the ability of LANA to replicate latent viral DNA.

As binding of USP7 to a LANA mutant lacking aa 971 to 986 is strongly reduced (in coimmunoprecipitation assays) or absent (in GST pulldown assays), this LANA mutant was used to study the functional importance of the USP7-LANA interaction. LANA mediates episome persistence by tethering the viral episome to human chromatin and thus is important for latent replication. LANA binds via its C-terminal domain to two regions in the terminal repeats (TRs) of the KSHV genome, which are named LANA binding sites 1 and 2 (LBS1 and LBS2) (20). LANA is also sufficient to replicate a TR-containing plasmid in cells in a short-term replication assay (17, 20, 29, 51). Therefore, the mutant lacking the USP7 binding site (aa 971 to 986) was tested in a transient replication assay to analyze its replication ability.

The LANA mutant lacking aa 971 to 986 replicated a plasmid containing the KSHV latent origin of replication more efficiently than the wild-type LANA (LANA WT) (Fig. 5A). Quantification of the relative band intensity observed in three independent experiments suggests an approximately four times stronger replication by LANA Δ 971–986 in comparison to LANA WT (Fig. 5B).

Increased replication ability of LANA Δ 971–986 is not due to altered DNA binding capacity.

To investigate if the increased replication activity of the LANA mutant lacking the USP7 binding site (aa 971 to 986) could be explained by differences in the ability of both LANA proteins to bind DNA, EMSAs (electrophoretic mobility shift assays) were performed. Therefore, a fluorescently labeled (IRD700) oligonucleotide harboring both LBS-1 and LBS-2 was used. The oligonucleotide was incubated with previously expressed and purified GST-tagged cLANA proteins in the EMSA binding reaction. Protein complexes were analyzed on a native polyacrylamide gel, and bands were detected via fluorescence.

The C-terminal LANA domain lacking aa 971 to 986 bound to DNA to the same extent as cLANA WT (Fig. 5C). For both LANA proteins, a pattern of shifted bands was observed, of which those indicated by arrows in Fig. 5C are considered to represent specific shifts. Application of unlabeled oligonucleotide in great excess efficiently competed with binding of cLANA to DNA, and the bands seen in the presence of WT cLANA could be “supershifted” using an antibody to LANA. These results indicate the specific nature of the observed bands. As an additional control, a LANA mutant with a PYG motif mutated to three alanine residues was included; this mutant had previously been reported to abolish binding to the LBS regions (23). As expected, this mutant did not interact with DNA (Fig. 5C).

USP7 does not influence the transactivation function of LANA.

LANA acts as a transcriptional repressor or activator of several cellular and viral genes. Therefore, we analyzed the possible influence of USP7 binding on the role of LANA as a transactivator or repressor. Since LANA plays a role in E2F-dependent transcription, we examined the effect of USP7 binding on the activation of the cyclin E promoter, as previously shown (36). The mutant LANA Δ 971–986 was able to activate the cyclin E promoter to levels comparable to those achieved by LANA WT (Fig. 6A). Second, the full-length LANA mutant lacking the identified USP7 binding site (LANA Δ 971–986) (Fig. 6B) and the corresponding C-terminal mutant (cLANA Δ 971–986) (data not shown) were analyzed for their ability to modulate the p53-dependent activation of a p53-responsive promoter, known to be repressed by LANA (13), in p53-deficient H1299 cells. Neither of the LANA mutants was compromised in its ability to repress p53-dependent transcription (Fig. 6B and data not shown) in comparison to LANA WT.

LANA also represses the promoter/enhancer activity of the TR promoter region in the KSHV genome (51). We analyzed the impact of the LANA Δ 971–986 mutant on a TR reporter vector in HEK 293 cells. Both the LANA WT protein and the mutant lacking the USP7 binding site (aa 971 to 986) inhibited the TR reporter vector (Fig. 6C). Taken together, these findings indicate that recruitment of USP7 to LANA does not influence LANA transcriptional activation or repression and does not impact on the regulation of p53-dependent transcription by LANA.

USP7 interacts with the Orf73 proteins of MHV68 and RRV.

We next explored whether homologous Orf73 proteins from KSHV-related gammaherpesviruses could also interact with USP7. We focused on the Orf73 proteins from murine herpesvirus 68 (MHV68) as well as rhesus rhadinovirus (RRV). Expression constructs for both proteins were cotransfected with a USP7 expression construct or the corresponding empty vector. Protein complexes were precipitated via tags fused to the Orf73 proteins and analyzed on Western blots. In each experiment, a C-terminal LANA construct harboring the same tag served as a positive control. Both proteins, MHV68 Orf73 (Fig. 7A) and RRV Orf73 (Fig. 7B), interacted reproducibly with USP7 in coimmunoprecipitation experiments. Together with the previously investigated role of USP7 in the function of EBV EBNA-1, these findings suggest that the interaction of USP7 with Orf73 proteins is conserved among different gammaherpesviruses and underlines the possible importance of this interaction for the viral latent life cycle.