A Survey of the Interactome of Kaposi's Sarcoma-Associated Herpesvirus ORF45 Revealed Its Binding to Viral ORF33 and Cellular USP7, Resulting in Stabilization of ORF33 That Is Required for Production of Progeny Viruses

ABSTRACT

The ORF45 protein of Kaposi's sarcoma-associated herpesvirus (KSHV) is a gammaherpesvirus-specific immediate-early tegument protein. Our previous studies have revealed its crucial roles in both early and late stages of KSHV infection. In this study, we surveyed the interactome of ORF45 using a panel of monoclonal antibodies. In addition to the previously identified extracellular regulated kinase (ERK) and p90 ribosomal S6 kinase (RSK) proteins, we found several other copurified proteins, including prominent ones of ∼38 kDa and ∼130 kDa. Mass spectrometry revealed that the 38-kDa protein is viral ORF33 and the 130-kDa protein is cellular USP7 (ubiquitin-specific protease 7). We mapped the ORF33-binding domain to the highly conserved carboxyl-terminal 19 amino acids (aa) of ORF45 and the USP7-binding domain to the reported consensus motif in the central region of ORF45. Using immunofluorescence staining, we observed colocalization of ORF45 with ORF33 or USP7 both under transfected conditions and in KSHV-infected cells. Moreover, we noticed ORF45-dependent relocalization of a portion of ORF33/USP7 from the nucleus to the cytoplasm. We found that ORF45 caused an increase in ORF33 protein accumulation that was abolished if either the ORF33- or USP7-binding domain in ORF45 was deleted. Furthermore, deletion of the conserved carboxyl terminus of ORF45 in the KSHV genome drastically reduced the level of ORF33 protein in KSHV-infected cells and abolished production of progeny virions. Collectively, our results not only reveal new components of the ORF45 interactome, but also demonstrate that the interactions among these proteins are crucial for KSHV lytic replication.

IMPORTANCE Kaposi's sarcoma-associated herpesvirus (KSHV) is the causative agent of several human cancers. KSHV ORF45 is a multifunctional protein that is required for KSHV lytic replication, but the exact mechanisms by which ORF45 performs its critical functions are unclear. Our previous studies revealed that all ORF45 protein in cells exists in high-molecular-weight complexes. We therefore sought to characterize the interactome of ORF45 to provide insights into its roles during lytic replication. Using a panel of monoclonal antibodies, we surveyed the ORF45 interactome in KSHV-infected cells. We identified two new binding partners of ORF45: the viral protein ORF33 and cellular ubiquitin-specific protease 7 (USP7). We further demonstrate that the interaction between ORF45 and ORF33 is crucial for the efficient production of KSHV viral particles, suggesting that the targeted interference with this interaction may represent a novel strategy to inhibit KSHV lytic replication.

INTRODUCTION

Kaposi's sarcoma-associated herpesvirus (KSHV) is the etiological agent of Kaposi's sarcoma, the most common malignancy in HIV/AIDS patients. It is also associated with two lymphoproliferative disorders: primary effusion lymphoma and multicentric Castleman's disease (1–3). Like other herpesviruses, KSHV exhibits two alternative life cycles, a quiescent latent stage and a productive lytic stage. KSHV adopts primarily latent infection both in vitro and in vivo. During latency, only a few genes are expressed and no progeny are produced. The latent episomal viral genome can be reactivated to initiate lytic replication, during which most viral genes are expressed in a temporally regulated cascade, ultimately resulting in assembly and release of progeny virions and de novo infection of naive cells. Although lytic replication of herpesviruses ultimately results in death of the infected cells, spontaneous lytic replication of KSHV is believed to play critical roles in viral pathogenesis by disseminating the virus and providing paracrine regulation to the tumor microenvironment (4, 5). Therefore, elucidating the roles of viral proteins that are crucial for lytic replication will improve our understanding of KSHV pathobiology.

KSHV open reading frame 45 protein (ORF45) is expressed during the lytic cycle and is known to have multiple functions throughout the viral life cycle. ORF45 was originally identified as an immediate-early gene product (6) and later as a component of the tegument in KSHV virions (7, 8). KSHV ORF45 is involved in evasion of the host innate antiviral responses by inhibiting interferon regulatory factor 7 (IRF7) (9–11). It also plays a role in the intracellular transport of newly formed viral particles by association with the kinesin-2 motor protein KIF3A (12). More importantly, KSHV ORF45 has also been shown to cause persistent activation of the extracellular regulated kinase (ERK) and p90 ribosomal S6 kinases (RSKs) (13, 14). This activity is important not only for virus-host interactions (15, 16), but also for virus-virus interaction between KSHV and HIV (17, 18).

Although ORF45 is conserved among gammaherpesviruses (no homologue exists in alpha- or betaherpesviruses), the homology is limited and resides mostly at the amino- and carboxyl-terminal ends. ORF45 homologues also differ dramatically in protein length. KSHV ORF45 is the longest, at 407 amino acids (aa), while rhesus rhadinovirus (RRV), herpesvirus saimiri (HVS), Epstein-Barr virus (EBV), and murine herpesvirus 68 (MHV-68) have homologous proteins of 353, 257, 217, and 206 aa, respectively. Despite the differences, ORF45 homologues have been identified as virion protein components in all gammaherpesviruses examined so far, including RRV, MHV-68, and EBV, suggesting that certain tegument functions of ORF45 are conserved (19–24). However, the mechanism by which ORF45 is assembled into viral particles, how its multiple functions are regulated, and the evolutionary advantage of increased protein size remain to be determined.

We previously noticed that all ORF45 protein in cells exists in high-molecular-weight complexes and sought to characterize its associated viral and cellular proteins (13, 14). In this study, we isolated several such proteins, including the viral protein ORF33 and the cellular protein USP7. We mapped the ORF33-binding domain to the highly conserved carboxyl-terminal 19 aa of ORF45 and the USP7-binding domain to the reported consensus motif in the central region of ORF45. We further demonstrated that ORF45 binding to ORF33 through the conserved carboxyl terminus is required for stabilization of ORF33 protein and production of progeny virions.

MATERIALS AND METHODS

Cell culture and reagents.

BCBL-1 cells latently infected with KSHV were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS) and antibiotics. HEK293T cells were cultured in Dulbecco's modified Eagle's medium with 10% FBS and antibiotics. LS88 cells, provided by Madelon Maurice at the University Medical Center, Utrecht, the Netherlands, were cultured in RPMI 1640 medium supplemented with 10% FBS, antibiotics, 10 μg/ml blasticidin, and 500 μg/ml zeocin. iSLK cells, obtained from Don Ganem and Jinjong Myoung at the Novartis Institutes for Biomedical Research, were cultured in Dulbecco's modified Eagle's medium with 10% FBS, antibiotics, 450 μg/ml G418, and 1 μg/ml puromycin. iSLK.BAC16 and iSLK.BAC16-Stop45 cells were described previously (25). These cells were cultured similarly to iSLK cells with the addition of 400 μg/ml hygromycin B. BAC16.FLAG-ORF33 was generated by bacterial artificial chromosome (BAC) mutagenesis by inserting a FLAG tag preceding the N terminus of ORF33. Anti-hemagglutinin (HA), anti-FLAG M2, and anti-β-actin antibodies; 3×FLAG peptide; 12-O-tetradecanoylphorbol-13-acetate (TPA); ionomycin; isopropyl β-d-1-thiogalactopyranoside (IPTG); cycloheximide; leptomycin B; and EZview red anti-FLAG M2 affinity resin were purchased from Sigma-Aldrich (St. Louis, MO). Anti-USP7 antibody (for immunofluorescence assays [IFAs]) was purchased from Cell Signaling Technology (Danvers, MA). Anti-USP7 antibody (for Western blotting) was purchased from Bethyl Laboratories (Montgomery, TX). Antibodies detecting ORF45 and ORF33 were generated as previously described (8, 9). Epoxy-activated magnetic Dynabeads were purchased from Invitrogen (Carlsbad, CA). Protein G beads were purchased from ThermoFisher Scientific (Waltham, MA). Easytag [³⁵S]cysteine-methionine mixture was purchased from PerkinElmer (Waltham, MA). EZ-Run prestained protein ladder was purchased from Fisher Scientific (Pittsburgh, PA).

Generation, purification, and epitope mapping of monoclonal antibodies (MAbs).

Hybridomas against ORF45 were generated by the Florida State University hybridoma facility using a standard protocol. Briefly, polyhistidine-tagged ORF45 protein produced in a baculovirus expression system was purified and used as an antigen. Immunization, fusion, and cloning were performed by the facility using standard protocols. The hybridoma clones were screened by enzyme-linked immunosorbent assay (ELISA) against purified ORF45 proteins and by Western blotting to lysates of TPA-induced BCBL-1 cells and purified viruses. IgGs of different clones were purified using protein G columns from GE Life Sciences (Piscataway, NJ) according to the manufacturer's protocol. The purified IgGs were used for epitope mapping, using a custom library of peptides encompassing the entire ORF45 coding region from Mimotopes (Melbourne, Australia). The library consists of 66 peptides of 17 aa with an 11-aa overlap. Each peptide is tagged with biotin at its N terminus. Ninety-six-well plates were coated with 100 μl of 5-μg/ml streptavidin (Sigma-Aldrich) at 37°C overnight and then blocked using phosphate-buffered saline (PBS) with 0.1% bovine serum albumin (BSA) for 1 h at room temperature. The wells were washed with PBS with 0.1% Tween 20, and then, 100 μl of 8.5-μg/ml biotin-tagged peptides was added to each well. The wells were washed with PBS with 0.1% Tween 20, and the peptide-coated plate was incubated with 1 μg/ml purified IgG of each anti-ORF45 hybridoma clone for 1 h at room temperature. Horseradish peroxidase-conjugated anti-mouse IgG (Pierce, Rockford, IL) was used as a secondary antibody, and an enhanced colorimetric system (Pierce) was used for ELISA detection. The results were read using a microplate reader at 480 nm.

Plasmid constructs.

Plasmids pCR3.1-ORF45 and pKH3-Ubiquitin and derivatives have been described previously (9). Plasmids encoding the ORF45 Δ2–19, Δ19–77, Δ77–90, Δ90–115, Δ115–175, Δ175–219, Δ219–229, Δ229–294, Δ300–332, Δ332–383, and Δ384–407 internal-deletion mutants were generated by using the QuikChange mutagenesis kit (Stratagene). Plasmid pCMV-ORF33 expressing a 3×FLAG-tagged protein was generated by cloning a PCR fragment of the ORF33 coding sequence into the pCMV-3Tag1 vector (Stratagene). Plasmids encoding glutathione S-transferase (GST) fusion proteins of ORF45 aa 1 to 115, aa 115 to 237, aa 238 to 332, and aa 332 to 407 have been described previously (13). Plasmids encoding GST-ORF45 aa 115 to 172, aa 168 to 214, and aa 210 to 237 were generated similarly by cloning corresponding PCR fragments into the pGEX-5X vector (Pharmacia). pGEX-ORF45-C19 was constructed by inserting a pair of annealed synthesized oligonucleotides into pGEX-5X. The sequences are as follows: ORF45-C19 sense, 5′-gatcCTGGCTTCCACTCCGCCGCTGTGCGGTAACGGTGCTTACAACTGGCCGTGGCTGGAC-3′, and ORF45-C19 antisense, 5′-ggccTAGTCCAGCCACGGCCAGTTGTAAGCACCGTTACCGCACAGCGGCGGAGTGGAAGCCAG-3′ (lowercase letters represent the overhanging sticky ends required for ligating the sequence into the pGEX plasmid). FLAG-tagged pCI-USP7 was provided by B. Vogelstein at the Johns Hopkins School of Medicine (26).

Radiolabeling and immunoprecipitation (IP).

Twenty milliliters of BCBL-1 cells at 0.5 × 10⁶ cells/ml were grown overnight in RPMI 1640 medium and then induced by the addition of 20 ng/ml TPA and 250 nM ionomycin. After a 48-h induction, the cells were pelleted at 400 × g for 5 min; washed with RPMI 1640 medium lacking glutamine, methionine, and cysteine; pelleted again at 400 × g for 5 min; and resuspended in 2 ml of glutamine-, methionine-, and cysteine-deficient RPMI medium supplemented with 1× glutamine, 10% dialyzed FBS, 20 ng/ml TPA, 250 nM ionomycin, and 1× antibiotics. After starving for 15 min, the cells were labeled with 200 μCi of ³⁵S for 4 h at 37°C by adding 20 μl of EasyTag [³⁵S]cysteine-methionine mixture (PerkinElmer) directly to the medium. The labeled cells were collected by centrifugation at 500 × g for 2 min, washed with PBS, and lysed in 1 ml of whole-cell lysis buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Nonidet P-40, 1 mM sodium orthovanadate [Na₃VO₄], 40 mM glycerophosphate, 30 mM sodium fluoride, 10% glycerol, 5 mM EDTA, 1× protease inhibitor mixture [Roche, Indianapolis, IN], and 1 mM phenylmethylsulfonyl fluoride). The insoluble material was removed by centrifugation at 10,000 × g for 10 min, and 500 μl of the supernatant was precleared by incubation with 25 μl of protein G beads for 5 min at 4°C. The cleared supernatants were incubated with ORF45 monoclonal antibodies or control IgG with gentle rotation at 4°C for 1 h. To each tube, 50 μl of prewashed protein G beads was added, and the tubes were rotated at 4°C for 1 h. The beads were then washed with lysis buffer twice and with PBS three times before mixing with an equal volume of 2× Laemmli loading buffer and boiling for 10 min. The eluted proteins were run on an SDS-12% PAGE gel, dried, and analyzed using a phosphorimager.

Expression and preparation of GST fusion proteins.

Escherichia coli BL21 transformed with plasmids encoding GST or GST fusion proteins were induced with 1 mM IPTG for 3 h at room temperature. The cells were pelleted, washed once with PBS, and resuspended in PBS containing lysozyme and phenylmethylsulfonyl fluoride (PMSF). After sonicating twice, Triton X-100 was added to a final concentration of 1%, and the mixture was rotated at 4°C for 1 h. After pelleting at 10,000 × g for 10 min to remove cellular debris, the supernatant was incubated with glutathione-agarose beads (Sigma-Aldrich) at 4°C overnight. After washing the beads 5 times with PBS, the bound proteins were eluted using glutathione elution buffer (10 mM gluthathione, 50 mM Tris-HCl, pH 8.5) and then dialyzed against PBS overnight. Concentrations were determined with a bicinchoninic acid (BCA) protein kit (Pierce) according to the manufacturer's instructions. The purity was assessed by SDS-PAGE, followed by Coomassie brilliant blue staining.

Pulldown assays with GST fusion proteins.

HEK293T cells were transfected with FLAG-ORF33, FLAG-USP7, or empty vector. After 24 h, cells were harvested, and the lysates were mixed with 20 μg of purified GST or GST fusion protein for 1 h at 4°C. The mixture was then incubated with glutathione-agarose beads and rotated for an additional hour at 4°C. The beads were pelleted by centrifugation at 8,000 × g for 1 min and washed 3 times using whole-cell lysis buffer, followed by 2 times using PBS. Finally, the beads were mixed with an equal volume of 2× Laemmli loading buffer and boiled for 10 min before being run on an SDS-12% PAGE gel. The gels were then stained using Coomassie brilliant blue or used for Western blotting.

Immunoaffinity purification, mass spectrometry, and Western blot analysis.

Conjugations of mouse monoclonal antibodies to epoxy-activated Dynabeads were performed as previously described (27). Briefly, 10 mg of Dynabeads was suspended in 1 ml of 0.1 M sodium phosphate (SP) buffer, pH 7.4, and then washed in fresh SP buffer. The beads were resuspended in 200 μl of SP buffer containing 100 μg of protein G purified antibodies/bead and 100 μl of 3 M ammonium sulfate. After binding overnight at 30°C, the beads were washed sequentially with 1 ml each of SP buffer, 100 mM glycine-HCl, pH 2.7, 10 mM Tris-HCl, pH 8.8, 100 mM triethylamine, PBS, and PBS with 0.5% Triton X-100. Finally, the beads were resuspended in 200 μl of PBS plus 0.02% NaN₃ and stored on ice until use.

Immunoaffinity purification was performed as previously described (13, 27). Briefly, for immunoprecipitation with anti-ORF45 antibodies, 200 ml of BCBL-1 cells that had been induced for 48 h with 20 ng/ml TPA and 250 nM ionomycin were pelleted at 5,000 × g, washed with PBS twice, and then concentrated in a 1.5-ml tube. The cell pellet was lysed with 1.5 ml of Dynabead lysis buffer (20 mM HEPES-KOH, pH 7.4, 110 mM potassium phosphate, 1% NP-40, 1 mM Na₃VO₄, 40 mM glycerophosphate, 30 mM NaF, 10% glycerol, 5 mM EDTA, 1× protease inhibitor mixture [Roche], and 1 mM PMSF). The lysates were sonicated and centrifuged at 10,000 × g for 10 min, and 0.5 ml of the supernatants was incubated with 65 μl of monoclonal-antibody-bound epoxy-activated Dynabeads for 1 h at 4°C. After washing with the lysis buffer and PBS, proteins were eluted by incubation with 0.5 N NH₄OH, 0.5 mM EDTA for 20 min at room temperature. The resulting immunocomplexes were concentrated by lyophilization, resuspended in 15 μl of 1× Laemmli loading buffer, and then boiled and resolved on an SDS-PAGE gel. The gels were stained with colloidal blue (Invitrogen) according to the manufacturer's protocol. Prominent bands were excised and analyzed by liquid chromatography-tandem mass spectrometry (LC–MS-MS) at the University of Florida Proteomics Facility.

For immunoprecipitation of FLAG-tagged proteins, we used EZview red anti-FLAG M2 affinity resin from Sigma. The lysates of cells expressing FLAG-tagged proteins were incubated with 50 μl of M2 affinity resins for 1 h or overnight at 4°C. After washing with the lysis buffer and Tris-buffered saline (TBS) (50 mM Tris-HCl, pH 7.4, 150 mM NaCl), proteins were eluted by incubation with 150 ng/μl 3× FLAG peptide in TBS for 1 h at 4°C. The immunocomplexes were then analyzed by Western blotting.

For Western blotting, protein lysates were resolved by SDS-PAGE and transferred to nitrocellulose membranes. The membranes were blocked in 5% dried milk in 1× phosphate-buffered saline plus 0.2% Tween 20 and then incubated with diluted primary antibodies for 2 h at room temperature or overnight at 4°C. Anti-rabbit, anti-rat, or anti-mouse IgG antibodies conjugated to horseradish peroxidase (Pierce) were used as the secondary antibodies. An enhanced-chemiluminescence system (Pierce) was used for detection. For silver staining, protein samples were resolved by SDS-PAGE and stained using a GE silver-staining kit (GE Life Sciences, Piscataway, NJ).

Ubiquitination assay.

Plasmids expressing FLAG-ORF33 (7 μg), HA-ubiquitin (7 μg), and ORF45 (7 μg) were cotransfected into HEK293T cells seeded in 100-mm dishes. Forty-eight hours after transfection, the cells were washed with PBS, harvested, and lysed with 150 μl SDS lysis solution (150 mM Tris-HCl, pH 6.8, 5% SDS, 30% glycerol). After brief sonication, the cell lysates were diluted 1:10 with dilution buffer (PBS with 0.5% NP-40, 1× complete protease inhibitor, and 20 mM freshly dissolved N-ethylmaleimide). The diluted cell lysates were immunoprecipitated with anti-FLAG affinity resins. The immunocomplexes were then analyzed by Western blotting with anti-HA and anti-FLAG antibodies.

Indirect immunofluorescence staining.

HeLa cells cultured on coverslips in 12-well plates were transfected with ORF45 using Fugene 6 (Roche). After 24 h, the transfected cells were fixed with 4% paraformaldehyde in PBS for 10 min at room temperature, permeabilized with 0.2% Triton X-100 for 15 min on ice, and then incubated with primary antibodies (10 μg/ml) for 1 h at room temperature or overnight at 4°C. After four washes with PBS with 0.1% Tween 20, the cells were incubated with Alexa 488-labeled and Alexa 543-labeled secondary antibodies (Invitrogen) for 1 h. The cells were then mounted in 50% glycerol solution in double-distilled H₂O (ddH₂O) and visualized with a Zeiss LSM 510 laser-scanning confocal microscope using the 63× objective.

Genetic manipulation of the KSHV BAC genome.

Mutagenesis of BAC16 was performed by using a recombineering system as described by Tischer et al. (28, 29). In brief, the Kan/I-SceI cassettes were amplified from plasmid pEPKan-S by PCR with primers as follows: KS45-ΔC19aa-5′ (5′-AACTACCGCGGGGTCCAGGTGCAGCGACCCAACCCGCATCAAGCTTTGATTAATTGACTGGCCTCCACGCCACCCCTAGGATGACGACGATAAGTAGGG-3′) and KS45-ΔC19aa-3′ (5′-TATATGCACCGTTTCCACACAGGGGTGGCGTGGAGGCCAGTCAATTAATCAAAGCTTGATGCGGGTTGGGTCGCTGCGCCAGTGTTACAACCAATTAACC-3′) for the BAC16-ORF45 ΔC19 mutant. The purified PCR fragment was electroporated into BAC16-containing E. coli GS1783 cells that had been induced at 42°C for 15 min. The recombinant clones were selected at 32°C on LB plates containing 34 μg/ml chloramphenicol and 50 μg/ml kanamycin and then characterized by restriction fragment length polymorphism (RFLP). Positive clones were induced again at 42°C and plated on LB plates containing 1% l-arabinose for secondary recombination. Then, we picked replicas of the clones from the l-arabinose plates onto plates with 34 μg/ml chloramphenicol alone or plates with 34 μg/ml chloramphenicol plus 50 μg/ml kanamycin. The kanamycin-sensitive clones were second-recombinant clones and confirmed by RFLP and sequencing.

To make a revertant mutant, we replaced the mutant ORF45 with a wild-type ORF45 sequence by a homologous-recombination strategy similar to that described above. The Kan/I-SceI cassettes were amplified from plasmid pEPKan-S by PCR with primers as follows: KS45-ΔC19aa-R-5′ (5′-AACTACCGCGGGGTCCAGGTGCAGCGACCCAACCCGCATCCTGGCCTCCACGCCACCCCT AGGATGACGACGATAAGTAGGG-3′) and KS45-ΔC19aa-R-3′ (5′-TATATGCACCGTTTCCACACAGGGGTGGCGTGGAGGCCAGGATGCGGGTTGGGTCGCTGCGCCAGTGTTACAACCAATTAACC-3′) for the BAC16-ORF45 ΔC19R mutant.

Quantification of extracellular virion genomic DNA by real-time quantitative PCR (qPCR).

Viral DNA was isolated from the supernatant medium of induced iSLK.BAC16 and derivative cells as previously described (30, 31). Briefly, medium from the infected cells was centrifuged to remove any cellular debris and treated with TurboDNase (Ambion) to remove any unprotected DNA. The viral particles were lysed with buffer AL (Qiagen), and the proteins were degraded with protease K (Qiagen). The DNA was then extracted using phenol-chloroform extraction and analyzed by SYBR green real-time PCRs using KSHV-specific primers: ORF73-LCN (5′-CGCGAATACCGCTATGTACTCA-3′) and OFF73-LCC (5′-GGAACGCGCCTCATACGA-3′) with Bio-Rad C1000TM thermal cycler and the CFX96TM real-time detection system. Viral DNA copy numbers were calculated with external standards of known concentrations of serially diluted BAC16 DNA ranging from 1 to 10⁷ genome copies per reaction.

RESULTS

Identification of viral and cellular proteins associated with ORF45 by immunoaffinity purification and mass spectrometric analysis.

Because all ORF45 protein in cells exists in high-molecular-weight complexes (13, 14), we sought to isolate additional associated viral and cellular proteins under native conditions by immunoaffinity purification. To this end, we generated a panel of anti-ORF45 mouse MAbs and mapped their epitopes (Fig. 1A and data not shown). To identify the most suitable one for immunoaffinity purification, we screened the MAb clones by immunoprecipitating ORF45 from ³⁵S-labeled TPA/ionomycin-induced BCBL-1 cell lysates. These MAbs isolated similar but distinct profiles of proteins. A representative result is shown in Fig. 1B. As expected, ORF45, which represents a diffuse band of around 70 kDa, and apparently RSK1 and RSK2 of 90 kDa, were immunoprecipitated by all MAbs with various efficiencies (13). Among all the clones, 8B8 was the most efficient, whereas 2D7 was the least (Fig. 1B, bottom). In addition to ORF45 and RSKs, several other proteins were also copurified by these MAbs. In particular, a prominent ∼38-kDa protein (p38) was immunoprecipitated by MAb clones 2D4, 8B8, and 3A6, but not by 2D7, whereas an ∼130-kDa protein (p130) was purified by all MAbs except 2D4. These results confirmed the interactions between ORF45 and the RSKs, attested to the specificities of these MAbs, and also suggested that ORF45 exists in different complexes or conformations in cells during KSHV lytic replication (13, 14).

FIG 1.

Survey of the ORF45 interactome in KSHV-infected cells. (A) Schematic representation of the unique epitopes of anti-ORF45 monoclonal antibodies. The black arrowheads indicate epitopes of antibodies used in this study. (B) Distinct profiles of ORF45-associated proteins were precipitated by different anti-ORF45 monoclonal antibodies. BCBL-1 cells were induced for 48 h with TPA-ionomycin (Ind) and then labeled for 4 h with [³⁵S]methionine-cysteine. The cell lysates were immunoprecipitated with the indicated monoclonal anti-ORF45 IgGs or control IgG. ORF45 recovery was confirmed by Western blot analysis. (C) The ORF45 interactome changes during the course of KSHV lytic replication. ORF45 and associated proteins were immunoaffinity purified from BCBL-1 cells at different times postinduction using Dynabeads conjugated to 8B8 IgG. The proteins were eluted with NH₄OH buffer, and the eluates were resolved by SDS-PAGE. The prominent bands were excised and identified by mass spectrometry (marked on the right), and proteins of interest were confirmed by Western blot analysis (bottom). dpi, days postinduction.

We next conjugated purified IgGs of 8B8 to magnetic Dynabeads and purified ORF45 and its associated proteins from lysates of BCBL-1 cells induced with TPA/ionomycin for different lengths of time. The proteins eluted from the immunoaffinity beads were resolved by SDS-PAGE and revealed by silver (Fig. 1C) or Coomassie (data not shown) staining. Distinct profiles of proteins were observed, and the patterns were similar to the one shown in Fig. 1B, in which ³⁵S-labeled proteins were detected by autoradiography; bands of putative ORF45, RSKs, p130, and p38 were all visible (Fig. 1C). Importantly, the time course analysis revealed that the complement of ORF45-interacting proteins changed over the course of lytic replication. In particular, association with p38 increased over time whereas associations with RSK and ERK were apparent immediately following expression of ORF45 and decreased as the lytic cycle progressed, and association with p130 peaked 3 days after induction (Fig. 1C). These results suggest that ORF45 forms dynamic, temporally regulated interactions with other proteins during the KSHV lytic cycle.

The prominent protein bands were excised and subjected to LC–MS-MS to reveal protein identities, as we previously described (7, 8, 13). In addition to the expected ORF45, RSKs, ERKs, and cytoskeleton proteins reported (13), p38 was identified as viral protein ORF33 and p130 was identified as the cellular protein ubiquitin-specific protease 7 (USP7). USP7 is also known as herpesvirus-associated ubiquitin-specific protease (HAUSP) and was originally identified as a binding partner of ICP0 of herpes simplex virus 1 (HSV-1) (32). The identities of these proteins were confirmed by Western blotting with specific antibodies (Fig. 1C, bottom). We found that interactions between ORF45 and the cellular proteins RSK, ERK, and USP7 do not require other viral factors, because the proteins were coimmunoprecipitated with ectopically expressed ORF45 in KSHV-free HEK293 and BJAB cells. In contrast, the p38 band was absent in the IP complexes under the same conditions, consistent with the viral origin of ORF33 (data not shown).

Binding of ORF33 to the conserved carboxyl terminus of ORF45.

Because p38 was immunoprecipitated by most MAbs except 2D7, which recognizes an epitope at the conserved carboxyl end of ORF45, we speculated that ORF33 binds the same region and thus prevents its access by MAb 2D7. To test this hypothesis, we coexpressed ORF33 with full-length ORF45 (ORF45 aa 1 to 407, designated ORF45-FL) or a deletion mutant in which the last 24 aa were deleted (ORF45 aa 1 to 383, designated ORF45-ΔC24) and then performed IP assays with MAbs 8B8 and 2D7. As expected, ORF45-FL was precipitated by both MAbs 2D7 and 8B8, whereas ORF45-ΔC24 was pulled down by only 8B8 and not 2D7 (Fig. 2A, second gel from bottom, compare lanes 9 to 12 to lanes 3 and 4), confirming recognition of the C-terminal epitope by 2D7. When expressed together, ORF33 was efficiently coimmunoprecipitated with ORF45-FL by 8B8 but not with ORF45-ΔC24 (Fig. 2A, compare lane 6 to lane 12), confirming interaction between ORF33 and ORF45-FL and indicating the last 24 aa of ORF45 is required for the interaction. In contrast, ORF33 was not coprecipitated by MAb 2D7 under the same conditions, although the antibody precipitated a substantial amount of ORF45-FL (lane 4), suggesting that binding of ORF33 to ORF45 blocked epitope access by MAb 2D7. Taken together, these experiments suggest that ORF33 binds to the conserved C terminus of ORF45 that overlaps the epitope recognized by MAb 2D7.

FIG 2.

ORF33 binds to the conserved C terminus of ORF45. (A) ORF33 and the antibody 2D7 bind mutually exclusively to the C terminus of ORF45. ORF33 was coexpressed with ORF45-FL or ORF45 aa 1 to 382 (ORF45-ΔC24) in HEK293T cells, and the lysates were immunoprecipitated using antibody 8B8 or 2D7. The immunocomplexes were analyzed by Western blotting. (B) The ORF45 C terminus is sufficient for binding to ORF33. Lysates of ORF33-expressing HEK293T cells were pulled down by the indicated GST-tagged fragments of ORF45. The eluates were analyzed by Western blotting. -C, negative control. (C) Representation of the ORF45 fragments used in panel B and their abilities to bind to ORF33.

To survey which region of ORF45 is sufficient for binding to ORF33, we performed pulldown assays with four GST-tagged fragments of ORF45 that collectively encompass the entire coding region. Only the C-terminal fragment (aa 333 to 407) was able to bind to ORF33 (Fig. 2B, lane 6). Furthermore, the 19 aa in the C terminus of KSHV ORF45 was sufficient for binding to ORF33 (Fig. 2B, lane 7). These results are depicted in Fig. 2C. The C terminus of ORF45 is the only region in its entire coding sequence that is highly conserved among all gammaherpesviruses (Fig. 2C). Altogether, these results led us to conclude that the conserved carboxyl terminus of ORF45 is essential and sufficient for binding to ORF33.

Colocalization of ORF45 and ORF33.

KSHV ORF45 was observed primarily in the cytoplasm, and we and others recently showed that it actually shuttles between the nucleus and cytoplasm (7, 9, 31, 33). In contrast, KSHV ORF33 has been seen mostly in the nucleus (33). Because of this discrepancy, we examined the subcellular localization of transfected ORF45 and FLAG-ORF33 in HeLa cells. As previously reported, ORF33 by itself was mostly localized in the nucleus (Fig. 3I and L), whereas ORF45 was seen in both the nucleus and cytoplasm (Fig. 3F and H). When ORF33 was coexpressed with ORF45, a portion of ORF33 was seen in the cytoplasm. Interestingly, in about a third of the cells, a significant portion of both ORF33 and ORF45 accumulated in the nuclear periphery (Fig. 3A and B). Significantly, the signals of both ORF33 and ORF45 overlapped perfectly, confirming their physical association in cells (Fig. 3D and D′). Relocation of ORF33 by ORF45 seems to depend on their interaction, because an ORF45 truncation mutant lacking the C terminus (ORF45-ΔC24) failed to relocate ORF33 (Fig. 3M and P). To provide further evidence that relocation of ORF33 to the cytoplasm depends on ORF45, we asked whether altering ORF45 localization affects that of ORF33. We therefore treated cells with leptomycin B (LMB), which we have previously shown to inhibit nuclear export of ORF45 (31). As shown in Fig. 3Q and R, the cytoplasmic accumulation of ORF45 was abolished by LMB treatment. Consequently, the cytoplasmic presence of ORF33 was also abrogated. Similar results were seen in transfected SLK, HEK293, and Vero cells (data not shown). Furthermore, the colocalization of ORF45 and ORF33 was also apparent in KSHV-infected iSLK cells (a derivative of SLK cells) (Fig. 3X). These results demonstrate that ORF45 colocalizes with ORF33 in cells and substantially alters ORF33 subcellular localization through their direct interactions.

FIG 3.

ORF45 colocalizes with and relocalizes ORF33. (A to T) FLAG-ORF33 was expressed in HeLa cells with and without ORF45 or ORF45-ΔC24. Primes indicate additional localization patterns. (Q to T), cells were treated with LMB (10 ng/ml) for 1 h before staining. The cells were stained with mouse anti-FLAG and rabbit anti-ORF45 antibodies and visualized by confocal microscopy. (U to X) iSLK.BAC16.FLAG-ORF33 cells were induced for 72 h by adding doxycycline and butyrate and then stained with mouse anti-FLAG and rabbit anti-ORF45 antibodies and visualized by fluorescence microscopy. The dashed lines mark the edges of the nuclei. hpi, hours postinduction; DAPI, 4′,6-diamidino-2-phenylindole.

Binding of USP7 to a central region of ORF45 via a consensus USP7-binding motif.

We next used the same four GST-tagged ORF45 fragments to map the domain for binding to USP7 and found that only the aa 116 to 237 fragment pulled down USP7 (Fig. 4A and C). We divided the region into three overlapping fragments (aa 116 to 172, aa 168 to 214, and aa 210 to 237) and found that only the aa 210 to 237 fragment retained the binding (Fig. 4B and C). Inspection of this region revealed an amino acid sequence (²²³EGPS²²⁶) matching the consensus USP7-binding motif P/A/E-G-X-S (34, 35). To determine whether this sequence is required for binding between ORF45 and USP7, we mutated the residues and found that a G224E/S226A mutation abolished coimmunoprecipitation of USP7 with ORF45 by MAb 8B8, indicating that the ²²³EGPS²²⁶ motif is required for binding between USP7 and ORF45. Because the epitope recognized by 2D4 resides in this region (Fig. 1A) and the p130 band was not precipitated by 2D4 (Fig. 1B), we speculated that the USP7-binding site overlaps with the 2D4 recognition epitope. Indeed, while MAb 8B8 immunoprecipitated ORF45 and USP7, 2D4 pulled down ORF45 but not USP7 (Fig. 4D, compare lane 4 to lane 2), suggesting that binding of USP7 to ORF45 excluded accession of its epitope by 2D4. Furthermore, the G224E/S226A mutation abolished recognition of ORF45 by 2D4, because the antibody efficiently pulled down the wild type but not the mutant ORF45 protein (Fig. 4D, second gel from bottom, compare lanes 9 and 10 to lanes 3 and 4). In sum, these results suggest that USP7 binds to ORF45 through the ²²³EGPS²²⁶ motif, which resides within the epitope recognized by MAb 2D4.

FIG 4.

USP7 binds to ORF45 through a consensus USP7-binding motif. (A and B) GST pulldown assays identified an internal region of ORF45 (aa 210 to 237) capable of binding to USP7. Lysates of FLAG-USP7-expressing HEK293T cells were pulled down with purified GST-tagged fragments as indicated, and the eluates were analyzed by Western blotting. -C, negative control. (C) Representation of the ORF45 fragments used in panels A and B and their abilities to bind to USP7. (D) USP7 and antibody 2D4 bind mutually exclusively to a central region of ORF45. USP7 was coexpressed with ORF45 wild type (WT) or ORF45 G224E/S226A in HEK293T cells. The lysates were immunoprecipitated using antibody 8B8 or 2D4, and the immunocomplexes were analyzed by Western blotting. (E) ORF45 colocalizes with and relocalizes USP7. In the top two rows, HeLa cells were transfected with ORF45, left untreated or treated with LMB, and then stained using rabbit anti-USP7 and mouse anti-ORF45 antibodies and visualized by confocal microscopy. Cells lacking ORF45 signal are marked by arrowheads, and cells with ORF45 signal are marked by arrows. In the bottom row, iSLK.BAC16 cells were induced for 60 h by adding doxycycline and butyrate and then stained using rabbit anti-USP7 and mouse anti-ORF45 antibodies and visualized by fluorescence microscopy.

Having confirmed the association between ORF45 and USP7, we next examined their localization in cells. In cells expressing no ORF45, USP7 was localized exclusively in the nucleus, for example, the cell in the lower left in the images in the top row of Fig. 4E (marked with an arrowhead), as reported previously (36). However, in cells expressing KSHV ORF45, for example, the cell in the upper right in the images in the top row of Fig. 4E (marked with an arrow), USP7 was also found in the cytoplasm, and their signals showed considerable overlap, which is consistent with an association between the two proteins in cells (Fig. 4E, D). Upon treatment with LMB, localization of both USP7 and ORF45 was restricted to the nucleus (Fig. 4E, H). These results suggest that nuclear export of USP7 depends on ORF45. Furthermore, colocalization of ORF45 and USP7 in both the nucleus and cytoplasm was also observed in KSHV-infected iSLK cells (Fig. 4E, bottom row).

Increase of ORF33 stability by ORF45.

We noticed that the ORF33 protein level was increased by coexpression with the full-length ORF45 (Fig. 5A). The increase was dependent on the interaction between the two proteins, because the effect was compromised by deletion of the ORF33-binding domain in ORF45 (Fig. 2A, top). The effect was also clear in the IFA images (Fig. 3A, I, and M). As shown in Fig. 5A, coexpression of ORF45 significantly increased the protein level of ORF33 (compare lane 4 to lane 1). The effect of ORF45 was more dramatic than treatment with MG132, a proteasome inhibitor (Fig. 5A).

FIG 5.

ORF45 stabilizes ORF33 protein. (A) ORF45 inhibits ORF33 degradation. ORF33 was expressed in HEK293T cells with and without ORF45. The cells were treated with cycloheximide (CHX) for 12 h in the presence or absence of MG132, and the lysates were analyzed by Western blotting. (B) ORF45 extends the half-life of ORF33. ORF33 was expressed in HEK293T cells with and without ORF45 WT or ORF45-ΔC24. The cells were treated with CHX, and the lysates were analyzed by Western blotting. (C) Two regions of ORF45 are required for ORF33 stabilization. ORF33 was expressed in HEK293T cells with and without ORF45 WT or deletion mutants. The cells were treated with CHX for 12 h, and the lysates were analyzed by Western blotting. The error bars indicate standard deviations.

To determine whether ORF45 affects ORF33 stability, we treated the transfected cells with cycloheximide to inhibit protein synthesis and then measured the ORF33 protein level over time to estimate its half-life. When expressed alone, ORF33 had a half-life shorter than 8 h and became undetectable by 24 h following cycloheximide treatment. When coexpressed with ORF45, the ORF33 protein level was significantly increased and remained high after 24 h despite an initial decrease at 4 h after cycloheximide treatment. The initial decrease in the ORF33 level is presumably a result of degradation of unbound ORF33, while the bound population is stabilized. The increase of ORF33 stability depends on binding to ORF45, because the half-life of ORF33 was not extended by ORF45-ΔC24, in which the binding domain was deleted (Fig. 5B). In fact, coexpression of ORF45-ΔC24 appeared to cause a slight decrease in ORF33 accumulation (Fig. 5B, compare lane 11 to lane 1), suggesting a possible dominant-negative mutant.

Increase of ORF33 stability by ORF45 requires its association with USP7.

Although homologues of ORF45 in other gammaherpesviruses all bind to ORF33 through the conserved C terminus, none of them extended the ORF33 half-life to the extent of KSHV ORF45 (data not shown), suggesting an additional region(s) in KSHV ORF45 also contributes to the stabilization of ORF33. To locate this region(s), we used a series of ORF45 internal-deletion mutants and determined their effects on accumulation of ORF33 in the presence of cycloheximide. The results revealed that deletion of either aa 219 to 229 or aa 383 to 407 of ORF45 compromised its ability to extend the half-life of ORF33 (Fig. 5C). Not unexpectedly, these two regions are known to bind to USP7 and ORF33, respectively, suggesting association with USP7 is necessary for the increase of ORF33 stability. To confirm a role of USP7 in the increase of ORF33 accumulation by ORF45, we used a cell line (LS88) in which USP7 can be knocked down by doxycycline-inducible small interfering RNA (siRNA) (37). In the presence of ORF45, the level of ORF33 protein was reduced by USP7 knockdown, but in the absence of ORF45, it was not affected (Fig. 6A). This result demonstrates that the ORF45-dependent stabilization of ORF33 requires USP7.

FIG 6.

USP7 is required for ORF45-induced stabilization of ORF33. (A) Knockdown of USP7 reduces ORF45-dependent accumulation of ORF33. ORF33 was expressed with and without ORF45 in LS88 cells (which contain a doxycycline [Dox]-inducible siRNA targeting USP7 [siUSP7]). The cells were treated with doxycycline, and the lysates were analyzed by Western blotting. Short and long exposures are indicated by S and L, respectively. (B) ORF45 reduces ubiquitination of ORF33. FLAG-ORF33 was coexpressed in HEK293T cells with HA-ubiquitin (ub) and ORF45 wild type (WT), ORF45 Δ219–229, or ORF45-ΔC24. ORF33 was precipitated with anti-FLAG-agarose beads, and the eluates were analyzed by Western blotting. (C) Knockdown of USP7 increases ORF33 ubiquitination in the presence of ORF45. FLAG-ORF33 was expressed with and without ORF45 in LS88 cells, which were then treated with doxycycline. FLAG-ORF33 from the lysates was precipitated with anti-FLAG-agarose beads under denatured conditions, and the eluates were analyzed by Western blotting. WCL, whole-cell lysate. (D) Model of ORF45-dependent deubiquitination of ORF33 by USP7.

Decrease of ORF33 ubiquitination by ORF45.

Because ubiquitination is a prerequisite for proteasome-dependent protein degradation and USP7 is an enzyme that cleaves off ubiquitin chains from target proteins, we examined whether ORF45 affects the ubiquitination of ORF33. We expressed FLAG-ORF33 and HA-ubiquitin in the presence of the wild-type ORF45 or the binding-deficient mutants in cells. We lysed cells under denaturing conditions, immunoprecipitated ORF33 with anti-FLAG affinity beads, and probed the eluted immunocomplexes by Western blotting to detect ubiquitinated ORF33. As shown in Fig. 6B, ORF33 was ubiquitinated in the absence of ORF45 (lane 2). Ubiquitination of ORF33 was greatly reduced by coexpression of the wild-type ORF45, suggesting that ORF45 reduces ubiquitination of ORF33 (Fig. 6B, lane 3). The reduction apparently depended on ORF45 binding to ORF33 and to USP7, because deletion of either aa 383 to 407 or aa 219 to 229 in ORF45 strongly inhibited its ability to reduce ORF33 ubiquitination (Fig. 6B, lanes 4 and 5). We also found similar results following USP7 knockdown (Fig. 6C). Although ORF45 significantly reduced ubiquitination of ORF33, ORF45 was found to have no effect on USP7 enzymatic activity (data not shown). Because the interactions with both USP7 and ORF33 are required, reduction of ORF33 ubiquitination by ORF45 appeared to be a result of recruitment of both enzyme and substrate into close proximity (Fig. 6D).

Abolishment of ORF33 accumulation and production of progeny virions by deletion of 19 aa from the C terminus of ORF45.

The conserved C terminus of ORF45 was shown to be essential for MHV-68, but the underlying mechanism was unknown (38). Our data demonstrated that the same region in KSHV ORF45 is involved in interaction with ORF33 and is critical for accumulation of ORF33 protein in cells. To determine the roles of this conserved region of ORF45 in KSHV lytic replication, we deleted the C-terminal 19 aa by introducing stop codons following leucine-389 of ORF45 in BAC16, an infectious bacterial artificial chromosome clone of the entire KSHV genome (39). Similarly, we repaired the mutations and created a revertant version of BAC16 (Fig. 7A). The mutant and revertant BAC DNAs were reconstituted in iSLK cells, and stable cell lines carrying latent viral genomes were established as described in Materials and Methods. The cells were treated with doxycycline and sodium butyrate to induce lytic replication. Medium containing extracellular viruses and cells was collected daily following induction. Analysis of the cell lysates by Western blotting revealed that ORF33 protein levels were drastically reduced by deletion of the last C-terminal 19 aa (C19) from ORF45 (Fig. 7B). We measured extracellular viral genome copy numbers by real-time qPCR and found that this deletion also abolished virion production to an extent comparable to that with the ORF45-null mutant Stop45 (Fig. 7C). To determine whether C19 deletion affects packaging of ORF45 and/or ORF33 into virions, we concentrated extracellular viral particles and analyzed them by Western blotting. These results suggested that the C19 region of ORF45 is required for packaging of ORF45 and ORF33 into viral particles (Fig. 7D). Altogether, our data reveal the essential roles of the conserved C terminus of ORF45 by interacting with and stabilizing ORF33 in KSHV-infected cells.

FIG 7.

The ORF45 C terminus is required for ORF33 protein accumulation and progeny virus production. (A) Diagram of KSHV BAC16 mutant BAC16-ORF45 ΔC19 (deletion of the C-terminal 19 amino acids via insertion of a premature stop codon) and its revertant, BAC16-ORF45 ΔC19R. (B) Deletion of the ORF45 C terminus abolishes accumulation of ORF33 protein. iSLK.BAC16-ORF45 ΔC19 or ΔC19R cells were treated with doxycycline, and the lysates were analyzed by Western blotting at the indicated times postinduction. Antibody 2D7 recognizes the C terminus of ORF45. (C) Deletion of the ORF45 C terminus dramatically reduces progeny virion production. iSLK.BAC16, ORF45 ΔC19, ORF45 ΔC19R, or Stop45 (deletion of ORF45 through insertion of a premature stop codon) cells were treated with doxycycline, and extracellular viral genome copy numbers were assessed by qPCR. The error bars indicate standard deviations. (D) Progeny viral particles from BAC16-ORF45 ΔC19 lack the tegument proteins ORF33 and ORF45. One million concentrated viral particles, as measured by qPCR, were analyzed by Western blotting for the presence of the viral capsid ORF65 and tegument proteins ORF33 and ORF45.

DISCUSSION

Our earlier observation that all ORF45 protein in cells exists in high-molecular-weight complexes motivated us to characterize the native ORF45-containing complexes in KSHV-infected cells (14). To this end, we generated a panel of MAbs that recognize different epitopes of ORF45. Immunoprecipitation with these MAbs isolated distinct profiles of proteins, including the expected ORF45 itself and RSKs from KSHV-infected cells. Although the differences between these MAbs in the precipitation of ORF45 and its associated proteins could be partially explained by differences in their affinities and avidities, we favor an interpretation in which ORF45 can assume different conformations or exist in different complexes in cells. Indeed, time course analysis by immunoprecipitation with the most efficient MAb, 8B8, confirmed that the ORF45 interactome changes as lytic replication progresses. These results suggest that ORF45 forms dynamic interactions with other proteins during KSHV lytic replication. Future characterization of these ORF45-containing complexes may provide insights into its multiple roles throughout the KSHV life cycle.

We have identified two prominent components of the ORF45 interactome, in addition to the previously identified RSK, ERK, and cytoskeleton proteins. These components are the viral protein ORF33, which binds to the conserved carboxyl terminus of ORF45, and the cellular protein USP7, which binds to ORF45 through the consensus sequence in the central region. Interestingly, we found that the protein level of ORF33 is increased by its association with the C terminus of ORF45 and by the ability of ORF45 to recruit USP7 via the consensus motif. Furthermore, we found that deletion of the conserved C terminus of ORF45 abolished production of progeny virions. In support of a role of ORF45 in stabilization of ORF33 through their interaction, the accumulation of ORF33 protein in cells was also abolished upon this deletion. Although high conservation of the carboxyl terminus implies its importance, which was first confirmed in MHV-68 when deletion of the region abolished production of progeny virions, the mechanism was not revealed (38). Our results suggest that association with ORF33 is a crucial function encoded by the conserved carboxyl terminus of ORF45.

Unlike ORF45, which is found only in gammaherpesviruses, ORF33 is conserved in all herpesviruses. Its homologues, HSV-1 UL16, EBV BGLF2, and human cytomegalovirus (HCMV) UL94, are all present in the tegument layer of mature virions (24, 40–44), but the roles of the ORF33 homologues in herpesviral replication remain elusive. Although deletion of UL16 reduces the viral yield of HSV-1 (an alphaherpesvirus) only moderately (45), deletion of UL94 abolishes progeny virion production by HCMV (a betaherpesvirus) (46). In gammaherpesviruses, ORF33 of MHV-68 was found to be essential, initially by genome-wide signature-tagged transposon mutagenesis studies (47, 48). Guo et al. further demonstrated that an ORF33-null mutation does not affect viral DNA replication, viral gene expression, or capsid assembly but abolishes release of infectious virions, resulting in accumulation of partially tegumented viral particles in the cytoplasm (40). Interestingly, although these particles contain capsid proteins and another gammaherpesvirus-specific tegument protein, ORF52, they contain virtually no ORF45. This suggests that ORF33 is required for selective packaging of certain tegument proteins, such as ORF45, into virions in the cytoplasm. Such remarkable selectivity can now be explained by the specific binding of ORF33 to the conserved carboxyl terminus of ORF45. This binding between ORF33 and ORF45 can also be observed between their homologues in MHV-68 and in other gammaherpesviruses (data not shown). Although the detailed mechanisms remain to be determined, our results suggest a role for the association of ORF45 with ORF33 in viral assembly, particularly tegumentation processes of KSHV. Because ORF33 is conserved among all herpesvirus whereas ORF45 is unique to gammaherpesviruses, further study of the ORF45-ORF33 interaction will provide significant insights into the mechanisms by which gammaherpesviruses acquire tegument proteins during viral assembly and maturation.

USP7, a ubiquitin-specific protease, was originally identified as a binding partner of HSV-1 ICP0 that cooperatively facilitates viral replication (32, 49). Later, USP7 was found to regulate the delicate MDM2-p53 equilibrium by controlling the p53 protein level in cells, which is crucial for normal cellular homeostasis and diverse stress responses (50). In addition to the p53 pathway, USP7 has been shown to interact with various other substrates, such as histone 2B, many of which are involved in epigenetic and transcriptional regulation (51). We have demonstrated that USP7 binds to ORF45 through the EGPS sequence located in the middle region. A significant interaction between USP7 and ORF45 was recently revealed by a global mapping of KSHV-host protein interactions (52). Although USP7 had no apparent effect on the stability of ORF45 itself, we have found that USP7 is required for the increase in stability of ORF33 by ORF45, correlating with the level of ubiquitination of ORF33 in cells. Because ORF45 has no apparent effect on USP7 enzymatic activity, we speculate that this is likely achieved by recruitment of USP7 into close proximity of the ORF45-associated ORF33.

USP7 is known to be hijacked by herpesviruses to counteract the host innate antiviral responses. The disruption of PML bodies and degradation of PML proteins by HSV-1 have been shown to require ICP0 and its ability to bind to USP7 (53). USP7 itself has been recently shown to disrupt PML bodies (54). Similarly to ICP0, binding of USP7 by EBNA1 also interferes with PML body formation and increases PML protein turnover (55). Besides disruption of PML bodies, ICP0 directs deubiquitination of TRAF6 and IKKγ through USP7 and thus deactivates their responses to viral infection (56). Like ICP0, ORF45 is an immediate-early protein and is also incorporated in virions. In addition, ORF45 also increases USP7 cytoplasmic localization, similar to ICP0. Because of these similarities, we speculate that ORF45 can adopt the strategies utilized by ICP0 to hijack USP7 for similar goals. Given the known role of ORF45 in evasion of innate immunity through targeting IRF7, it is possible that ORF45 is also involved in evasion of other aspects of innate antiviral responses, including PML bodies and Toll-like receptor (TLR) responses. However, we also should be aware that differences exist between ORF45 and ICP0. While ICP0 binds to the C-terminal domain of USP7, the consensus USP7-binding motif within ORF45 is known to bind to the TRAF domain in the N-terminal region of USP7 (57). In addition, ICP0 itself has E3 ligase activity, but ORF45 does not. While the analogy between ORF45 and ICP0 is not a perfect one, their shared modulation of USP7 activity during herpesviral lytic replication warrants further studies regarding the functional consequences of ORF45-USP7 interaction.

USP7 is also exploited by herpesviruses to modulate p53 pathways. The latent protein EBNA1 of EBV has been shown to bind with USP7 through the same region as p53 (58). This binding excludes USP7 from interacting with p53 and consequently prevents USP7 from deubiquitinating p53. Therefore, ubiquitinated p53 is rapidly degraded, resulting in increased survival of latently EBV-infected cells (34, 59). Recently, the homologue of EBNA1 in KSHV, LANA, has also been found to bind to USP7. In addition, a lytic protein of KSHV, vIRF4, has been shown to bind to USP7 and inhibit its enzymatic activity, leading to increased degradation of p53 (60, 61). A role of ORF45-USP7 in regulation of p53 is also possible and awaits future studies. USP7 is also involved in epigenetic and transcriptional regulation through deubiquitination of H2B and other substrates (51). Recent studies have revealed that ORF45 and its associated RSK are involved in transcriptional regulation (25). Whether USP7 is involved in this function is still unknown.

The distinct profile of proteins coimmunoprecipitated by different anti-ORF45 MAbs suggests that ORF45 protein in cells exists in different complexes or assumes different conformations, resulting in exposure of different epitopes accessible to different MAbs. The differences are most obvious when we compare the profiles of the immunocomplexes precipitated by 8B8 and 2D7. 2D7, which recognizes the C terminus of ORF45, precipitated no ORF33 and much less ORF45 than 8B8, but similar levels of RSKs and ERKs. Further experiments revealed that ORF33 and 2D7 bound to the same C-terminal region, explaining why binding of ORF33 excluded binding of the 2D7 antibody. Importantly, these results clearly suggest that there are at least two populations of ORF45 protein in lytically KSHV-infected cells: one bound to ORF33 and another unbound to ORF33. Because 2D7 precipitated only about half as much ORF45 protein as other antibodies, we can infer that a significant portion of ORF45 protein in cells is bound to ORF33. However, it is currently unknown whether ORF33-bound ORF45 proteins exist as a heterodimer or as part of even larger complexes with other proteins, for example, partially assembled capsid-tegument complexes. Because binding to ORF33 is a critical function performed by the conserved C terminus of ORF45, determining the ultimate fates of these two populations of ORF45 will facilitate elucidation of the functional consequences of ORF45-ORF33 interactions.

Distinct immunoprecipitation profiles of proteins can also be observed for antibody 2D4, which recognizes an epitope that overlaps with the USP7-binding motif, explaining why USP7 was not coimmunoprecipitated by this antibody. Although 2D4 failed to coimmunoprecipitate USP7, it precipitated RSKs, ERKs, and ORF33 efficiently, and it was actually used to identify RSKs as ORF45-interacting proteins in our previous work (13). These results suggest that two populations of ORF45, one bound to USP7 and the other not bound to USP7, exist in cells. Because the amount of ORF45 isolated by this antibody is only slightly less than that isolated by other antibodies, it appears that only a relatively small fraction of ORF45 protein in cells is bound to USP7 and/or that ORF45 binds to USP7 with low affinity and can be competitively excluded by MAb 2D4. It is currently unknown whether these subpopulations represent a single homogeneous complex or multiple heterogeneous complexes. Though the latter seems more likely, future studies using epitope peptides to elute the native complexes will enable more detailed analyses of the ORF45-containing native protein complexes in KSHV-infected cells.

Because of their capacities to purify different populations of ORF45, these MAbs can be used to deplete or enrich subpopulations of ORF45. The MAbs provide us with unique tools for dissecting tegumentation processes, which will ultimately improve our understanding of how gammaherpesvirus-specific tegument proteins are assembled into viral particles. Differences between the antibodies in isolation of distinct subpopulations of ORF45 may allow isolation of immature or partially assembled (tegumented) viral particles at various assembly stages. Biochemical analyses of these immature virions and structural studies with electron microscopy (EM) and cryo-EM will facilitate delineation of the mechanisms of assembly of ORF45 into virions and the overall tegumentation processes of herpesviruses. Furthermore, these MAbs can also be used to decorate or label ORF45 protein on the immature virions for study by cryo-EM, which could reveal the structural organization of tegument protein in virions. Similarly, mature extracellular virions can be stripped of envelope to expose tegument proteins that become accessible for labeling with epitope-specific MAbs, thus allowing structural probing by high-resolution cryo-EM. Although artificial tags have been used extensively in affinity purification and proteomic studies, they cannot be easily adapted to these studies. Therefore, our collection of MAbs recognizing different epitopes of ORF45 offers great advantages that cannot be matched by the artificial affinity tags.

As evidenced by the dynamic nature of the ORF45 interactome over the course of KSHV lytic replication, the ability of ORF45 to interact with multiple partners in a spatially/temporally regulated manner is conceivably crucial for its diverse functions. Curiously, the length of ORF45 seems to have expanded significantly during herpesvirus evolution; the ORF45 of KSHV is almost double the size of its homologues in EBV and MHV-68. Despite the length difference and lower overall homology, except the highly conserved C-terminal 19 aa, all ORF45 homologues are present in virions, suggesting that packaging of ORF45 into virions is a conserved feature that is presumably determined by the interaction of its conserved C terminus with ORF33. This is supported by our finding that C19 deletion abolishes the incorporation of ORF45 and ORF33 into viral particles. However, since deletion of C19 dramatically reduces ORF33 accumulation, it is unclear whether the defect in assembly can be attributed to ORF45-ORF33 interaction or loss of ORF33 protein. Therefore, further investigation of the roles of ORF33 in KSHV lytic replication is warranted.

What has driven the expansion of ORF45 protein length during evolution remains mysterious. The increased length ostensibly would allow KSHV ORF45 to acquire unique functions through association with additional partners. For example, the USP7-binding motif is not detected in all its homologues, and whether USP7 is also involved in the increase in stability of ORF33 by other ORF45 homologues remains to be determined. Because ORF45 can associate with partners cooperatively, it appears to function as a scaffold or adaptor protein that recruits different proteins into close proximity to accomplish its many diverse and important functions, including manipulation of the MAPK-ERK pathway, inhibition of the IRF7 response, and formation of functional virions. In the present and previous studies, we focused only on the prominent components of the ORF45 interactome. Further analyses of the entire interactome and comparison to those of ORF45 homologues in other viruses will lead to a comprehensive understanding of the multifaceted roles of ORF45 in herpesviral life cycles and may shed light on the unique functions acquired by the significant expansion of ORF45 coding capacity during evolution.