The herpes simplex virus type 1 (HSV-1) regulatory protein ICP0 interacts with and Ubiquitinates p53.

Herpes simplex virus type 1 regulatory protein ICP0 contains a zinc-binding RING finger and has been shown to induce the proteasome-dependent degradation of a number of cellular proteins in a RING finger-dependent manner during infection. This domain of ICP0 is also required to induce the formation of unanchored polyubiquitin chains in vitro in the presence of ubiquitin-conjugating enzymes UbcH5a and UbcH6. These data indicate that ICP0 has the potential to act as a RING finger ubiquitin ubiquitin-protein isopeptide ligase (E3) and to induce the degradation of certain cellular proteins through ubiquitination and proteasome-mediated degradation. Here we demonstrate that ICP0 is a genuine RING finger ubiquitin E3 ligase that can interact with and mediate the ubiquitination of the major oncoprotein p53 both in vitro and in vivo. Ubiquitination of p53 requires ICP0 to have an intact RING finger domain and occurs independently of its ability to bind to the ubiquitin-specific protease USP7.

Herpes simplex virus type 1 regulatory protein ICP0 contains a zinc-binding RING finger and has been shown to induce the proteasome-dependent degradation of a number of cellular proteins in a RING fingerdependent manner during infection. This domain of ICP0 is also required to induce the formation of unanchored polyubiquitin chains in vitro in the presence of ubiquitin-conjugating enzymes UbcH5a and UbcH6. These data indicate that ICP0 has the potential to act as a RING finger ubiquitin ubiquitin-protein isopeptide ligase (E3) and to induce the degradation of certain cellular proteins through ubiquitination and proteasomemediated degradation. Here we demonstrate that ICP0 is a genuine RING finger ubiquitin E3 ligase that can interact with and mediate the ubiquitination of the major oncoprotein p53 both in vitro and in vivo. Ubiquitination of p53 requires ICP0 to have an intact RING finger domain and occurs independently of its ability to bind to the ubiquitin-specific protease USP7.
Herpes simplex virus type 1 (HSV-1) 1 is a common human pathogen that is capable of establishing a quiescent state of infection within sensory neurons following primary lytic infection of epithelial cells. HSV-1 gene expression during lytic infection occurs in a regulated temporal cascade with three groups of genes: immediate early (IE), early (E), and late (L) (reviewed in Ref. 1). Although the majority of IE genes have been associated with the regulation of viral gene expression, only the IE protein ICP0 (also known as Vmw110) is capable of trans-activating all three classes of viral genes (reviewed in Ref. 2). Virus mutants that do not express ICP0 are severely impaired in their ability to replicate in limited passage human fibroblasts at low doses of input virus and are more likely to establish a quiescent infection. ICP0 is sufficient to reactivate quiescent viral genomes in both cultured cells and mouse models (2)(3)(4)(5). ICP0 has been reported to interact with various cellular proteins, including cyclin D3 (6), elongation factor EF-1␦ (7), and the transcription factor BMAL1 (8), and it forms a strong and specific interaction with ubiquitin-specific prote-ase USP7 (also known as HAUSP) (9). Although the exact mechanism by which ICP0 functions has yet to be determined, one crucial feature is its zinc-binding RING finger domain.
Recently, it has been demonstrated that the RING finger domain of ICP0 has ubiquitin E3 ligase activity (10,11). E3 ligases provide the substrate specificity to mediate the transfer of ubiquitin from ubiquitin-conjugating enzymes (E2s) to substrate proteins targeted for ubiquitin-dependent proteasomemediated degradation (reviewed in Ref. 12). ICP0 has also been shown to induce the accumulation of co-localizing conjugated ubiquitin in vivo (13) and to induce the proteasome-mediated degradation of certain cellular proteins, including the major ND10 (PML nuclear body) constituent proteins PML and Sp100 (14 -17), the centromeric proteins CENP-A and CENP-C (18,19), and the catalytic subunit of DNA-dependent protein kinase (20). The degradation of all these proteins requires that ICP0 has an intact RING finger domain and provides strong evidence that ICP0 acts as an E3 ligase in vivo. However, to date, no specific substrate has been shown to be directly ubiquitinated by full-length ICP0.
A number of DNA viruses have been shown to affect the stability of the major oncoprotein p53. The E6 protein of human papillomaviruses 16 and 18, in association with E6-AP, and the adenovirus E1B-55K/E4-orf6 complex, have both been shown to induce the degradation of p53 through ubiquitination (21)(22)(23). Conversely, the simian virus 40 (SV40) large T antigen and the adenovirus E1B protein have been shown to inhibit specific p53 transcriptional functions through its sequestration (24,25). There is limited information on the fate of p53 during HSV-1 infection. A number of cellular proteins, including p53, have been shown to be recruited into HSV-1 DNA replication compartments (26), although the significance of this recruitment remains unclear. The available evidence suggests that the overall p53 levels are not greatly affected in the HSV-1 infection systems so far examined. Analysis of HSV-1 mutants restricted in their ability to express IE proteins has shown that cell cycle arrest occurs independently of p53 (27) and that the ICP0-induced mitotic block is a direct result of its ability to induce the degradation of CENP-A and CENP-C (19,18).
Cellular levels of p53 are typically maintained at a low level where p53 turnover is tightly regulated by ubiquitination and proteasome-mediated degradation. Mdm2, the RING finger ubiquitin E3 ligase responsible for p53 ubiquitination, binds to the N terminus of p53, and in conjunction with UbcH5, mediates the ubiquitination of C-terminal lysine residues, resulting in its eventual degradation by the 26 S proteasome (28 -30). Stabilization of p53 has been shown to occur by several mechanisms including phosphorylation of p53 and ADP ribosylation factor-mediated sequestration of mdm2. These mechanisms result in the inhibition of mdm2 interacting with and therefore ubiquitinating p53 (reviewed in Ref. 31). Recently, Li et al. (32) demonstrated an alternative mechanism by which p53 could be stabilized. They showed that USP7 could bind directly to p53 and that the ubiquitin protease activity of USP7 was sufficient to deubiquitinate p53 targeted for degradation by mdm2 (32). The authors proposed that this deubiquitination and consequent stabilization of p53 by USP7 provided an additional mechanism by which cells could regulate their p53 transcriptional response.
Due to the strong and specific interaction formed between ICP0 and USP7 and the ability of ICP0 to act as a RING finger ubiquitin E3 ligase, we wanted to determine whether ICP0 had any effect on p53. Our results show that ICP0 can recruit p53 in vivo independently of other viral proteins and during viral infection and that it can interact directly with p53 in vitro, independent of its USP7-binding domain. This interaction, in conjunction with the E3 ligase activity, of ICP0 is sufficient to allow ICP0 to mediate the ubiquitination of p53 both in vitro and in vivo in a RING finger-dependent manner. We also demonstrate that U2OS cells infected with an HSV-1 mutant that fails to express functional ICP0 are more susceptible to UV-induced apoptosis then those infected with wild type HSV-1. These results demonstrate that ICP0 is a genuine ubiquitin E3 ligase and suggest that one of its functions is to influence p53-mediated pathways.

MATERIALS AND METHODS
Plasmids-pCDNA-p53 and pCDNA-mdm2 were kind gifts from Ron Hay (University of St. Andrews). pGex2T-p53 was a kind gift from Nick La Thangne (Glasgow University). Plasmids expressing the E2-conjugating enzymes UbcH5a, UbcH6, and cdc34 were gifts from Seth Sadis (Millennium Pharmaceuticals); the open reading frames were amplified from their respective pC-CMVFLAG backbone vectors by PCR and cloned into the pET28a vector (Novagen) as NdeI/NotI fragments allowing the insertion of sequences encoding a polyhistidine tag. pET3aUbcH7 and pTriexUbcH10 were kind gifts from Phil Robinson (University of Leeds). Plasmids pCI-110, pCI-FXE, pCI-M1, and pGex241 have been described previously (19,10). pCW7 (expressing polyhistidine-tagged ubiquitin) was a kind gift from Ron Kopito (University of Standford).
Construction of Baculoviruses-Viruses Ac.ICP0His, Ac.FXEHis, Ac.110, Ac.FXE, and Ac.E52X have been described previously (10,33). Baculoviruses Ac.CMV.EYFP and Ac.CMV.EYFP-ICP0 were constructed to express the relevant proteins in mammalian cells as follows. Plasmid pEYFP-ICP0 contains the NcoI-HpaI genomic fragment containing the ICP0 coding region inserted into the SmaI site of pEYFP-C1 (Clontech) to express an enhanced yellow fluorescent protein (EYFP)-ICP0 fusion protein. The BglII-NdeI fragment of pCIneo that contains the 5Ј part of the HCMV IE promoter region, was linked to the NdeI-HindIII fragment of pEYFP-ICP0 (containing the 3Ј part of the HCMV promoter and the EYFP-ICP0 fusion fragment) and inserted into vector pFastBacHTa (Invitrogen) between its BamHI and HindIII sites. The resultant plasmid was used to generate recombinant baculovirus Ac.C-MV.EYFP-ICP0 by following the Bac-toBac protocol (Invitrogen). Similarly, the NdeI-XhoI EYFP coding region of pEYFP-C1 was inserted with the BglII-NdeI fragment of pCI-neo into the BamHI-SalI sites of pFastBacHTa, and the resultant plasmid was used to generate Ac.CMV.EYFP.
Cells and Co-Transfection Assays-U2OS and mouse p53: mdm2(Ϫ/Ϫ) cells were grown in Dulbecco's modified Eagles' medium supplemented with 10% fetal calf serum. Cells were seeded into 35-mm dishes at a cell density of 1.5 ϫ 10 5 cells/dish and transfected using LipofectAMINE Plus reagent (Invitrogen) with 10 ng of pCDNA-p53 plasmid, 50 ng of pCW7, and 100 ng of pCDNA-mdm2, pCI-110, pCI-FXE, or pCI-M1 according to the manufacturer's instructions. Promoter competition was balanced by including appropriate amounts of pCI-neo vector, and empty pUC-9 was used to bring the total DNA amounts to 500 ng. At 16 h after transfection, the cells were treated with MG132 (final concentration 10 M) and incubated for an additional 8 h. Cell monolayers were subsequently washed twice in ice-cold phosphatebuffered saline before being harvested in either 1ϫ SDS-PAGE boiling mix buffer containing 3 M urea and 25 mM dithiothreitol or 1 ml of guanidine HCl buffer A (phosphate-buffered saline plus 6 M guanidine HCl, 0.1% Nonidet P-40, 10 mM ␤-mercaptoethanol, and 5% glycerol), pH 8.0. His-tagged ubiquitinated proteins were isolated by nickel affinity chromatography using 35 l of equilibrated nickel-nitrilotriacetic acid beads (Qiagen) per sample as described in Ref. 34. Samples were resolved by 10% SDS-PAGE and Western blotted, and membranes were probed with either anti-p53 (Oncogene Ab-6) or anti-ICP0 (11060) monoclonal antibodies.
Immunofluorescence and Confocal Microscopy-Aliquots of 1 ϫ 10 5 U2OS cells were seeded onto 12-mm coverslips in 24-well dishes and infected with 50 plaque-forming units/cell of either Ac.CMV.EYFP or Ac.CMV.EYFP-ICP0. The cells were stained for immunofluorescence 16 h after infection using an anti-p53 monoclonal antibody (Oncogene Ab-6, 1 ⁄500) and a secondary Cy5-conjugated goat anti-rabbit IgG antibody (Amersham Biosciences, 1 ⁄500) and examined by confocal microscopy as described previously (10). Human fetal foreskin fibroblast cells (HFFF-2; European Collection of Cell Cultures) were grown in Dulbecco's modified Eagles' medium supplemented with 10% fetal calf serum, seeded onto coverslips at 1 ϫ 10 5 cells/well, and then infected with HSV-1 strain 17ϩ at a multiplicity of 10 plaque-forming units/cell or with HSV-1 virus FXE (35) that expresses the FXE RING finger mutant ICP0 protein, at a multiplicity of 0.1 plaque-forming units/cell. The strain 17ϩ and FXE-infected cells were fixed at 2 or 24 h after infection respectively and then stained for ICP0 (monoclonal antibody 11060) and p53 (rabbit FL-393, Santa Cruz Biotechnology). Fluorescein isothiocyanate-and Cy5-labeled secondary antibodies were used to prevent channel overlap.
Expression and Purification of Proteins-Human ubiquitin-activating enzyme (E1), full-length ICP0, RING finger deletion mutant FXE, and GST-241 were purified as described (10). Ubiquitin was purchased from Sigma. Clones expressing recombinant E2-conjugating enzymes and pGex2T-p53 were transformed into BL21 (DE3) pLysS bacteria, and single colonies were used to inoculate 100 ml YTB broth. Cultures were grown to mid-log phase at 37°C before being induced with 0.1 mM isopropyl-1-thio-␤-D-galactopyranoside for 4 h at 25°C. Bacterial pellets (equivalent to 10 ml of bacterial culture) were resuspended in 10 ml of the appropriate buffer, lysed by probe sonication, and clarified by centrifugation at 13,000 rpm for 10 min. UbcH5a, UbcH6, UbcH10, and cdc34 were purified by nickel affinity chromatography using 150 l of equilibrated nickel-nitrilotriacetic acid beads (Qiagen) in buffer A (50 mM Hepes, pH 7.2, 150 mM NaCl, 10% glycerol, 0.1% Nonidet P-40, and 10 mM ␤-mercaptoethanol). Supernatants were tumbled at 4°C for 1 h before being washed with 3 ϫ 1 ml of buffer A plus 20 mM imidazole (pH 7.5). Recombinant protein was eluted from the beads in 300 l of buffer A plus 250 mM imidazole (pH 7.5) and dialyzed against 50 mM Tris, pH 7.5, 50 mM NaCl, 2.5 mM ␤-mercaptoethanol, aliquoted, and stored at Ϫ70°C. UbcH7 was purified by a combination of cation and anion exchange in buffer B (50 mM Tris (pH 7.5), 25 mM NaCl, 1 mM EDTA, 2 mM benzamidine, 5 mM ␤-mercaptoethanol). The soluble supernatant fraction was bound to 1 ml of Q-Sepharose (Sigma) for 1 h at 4°C, and the flow-through was collected and bound to 1 ml of SP-Sepharose (Sigma) for 2 h at 4°C. The column was extensively washed in buffer B before proteins were eluted using an NaCl gradient (0 -0.6 M NaCl) in 50 mM Tris (pH 7.5) and 5 mM ␤-mercaptoethanol. Fractions containing UbcH7 were identified by SDS-PAGE and Coomassie Brilliant Blue staining. GST-p53 was purified in buffer C (100 mM Tris-HCl (pH 8.0), 150 mM NaCl, 5% glycerol, 0.1% Nonidet P-40, 5 mM ␤-mercaptoethanol) using 150 l of equilibrated GST beads (50% w/v) for 1 h at 4°C. The beads were subsequently washed three times with 1 ml of buffer C. GST-p53 was eluted from the beads in 200 l of buffer C plus 50 mM reduced glutathione (pH 7.5) and dialyzed against buffer C to remove glutathione.
GST-p53 Pull-down Assays-Frozen SF21 cell pellets (equivalent to 4 ϫ 10 6 cells) infected previously for 72 h at a multiplicity of infection of 2 with recombinant baculoviruses expressing ICP0, E52X, FXE, or mock-infected were resuspended in 500 l of buffer C plus protease inhibitors (Roche Applied Science). Cells were lysed by gentle bath sonication, and the extracts were clarified by ultracentrifugation at 30,000 rpm for 25 min at 4°C. Extracts were precleared using 50 l (50% w/v) of equilibrated GST beads in buffer C end-over-end for 45 min at 4°C. 20 l of GST beads prebound to either GST or GST-p53 were mixed with precleared supernatants for 2 h at 4°C. The beads were washed five times with 200 l in ice-cold buffer C. Soluble protein complexes were eluted from the beads in 30 l of ice-cold buffer C containing 50 mM reduced glutathione (pH 7.5). The samples and 1 ⁄100 input material were denatured by adding 10 l of 1ϫ SDS-PAGE boiling mix (as described above) and resolved by 7.5% SDS-PAGE and analyzed by Western blotting with an anti-ICP0 monoclonal antibody 11060.
In Vitro Transcription/Translation and Ubiquitination Assays-In vitro transcription/translation was performed using the rabbit reticulocyte lysate system (Promega) in the presence of [ 35 S]methionine fol-lowing the manufacturer's instructions. In vitro ubiquitination assays were performed either in the presence of 0.5 l of [ 35 S]methioninelabeled substrate or in the presence of 50 ng of purified substrate protein. Reactions were carried out in buffer D (50 mM Tris-HCl, pH 7.5, 50 mM NaCl, 5 mM ATP) containing an ATP regenerating system (10 mM creatine phosphate, 3.5 units/ml creatine kinase, 0.6 units/ml inorganic pyrophosphatase) containing 20 ng of E1, 50 ng of E2, 5 g of ubiquitin, and 50 ng of purified full-length ICP0, RING finger mutant (FXE), or GST-241. Reactions were carried out in a final volume of 10 l for 3 h at 28°C and terminated by the addition of 5 l of 3ϫ SDS-PAGE boil mix buffer containing 9 M urea and 100 mM dithiothreitol. Samples were resolved either by 7.5% SDS-PAGE or by 4 -12% NuPAGE gels (Invitrogen). Gels were either Western blotted and probed with antiubiquitin (Santa Cruz Biotechnology P4D1) or with anti-p53 (Oncogene Ab-6) monoclonal antibodies or stained with Coomassie Brilliant Blue, destained, dried, and analyzed by phosphorimaging.
FACS Analysis-Aliquots of 4 ϫ 10 5 U2OS cells were seeded onto 35-mm dishes and were either mock-infected or infected with five plaque-forming units/cell with wild type HSV-1 syn17ϩ or dl1403 (⌬ICP0) (35), an HSV-1 mutant that fails to express ICP0. Following absorption, the cells were overlaid with medium containing 10 M acycloguanosine to prevent the initiation of viral DNA replication. At 3 h after infection, the medium was removed, and then the cells were washed in phosphate-buffered saline and subjected to a UV dose of 50 J/m 2 (UV cross-linker, Stratagene). The cells were subsequently incubated in normal medium again containing acycloguanosine, and then detached, and adherent cells were harvested 24 h after UV irradiation using cell dissociation buffer (Sigma). The cells were pelleted by centrifugation and washed twice with 1 ml in phosphate-buffered saline. Cell pellets were either stained directly for apoptotic markers using annexin V-PE apoptosis detection kit (BD PharMingen) following the manufacturer's instructions or fixed and permeabilized using FIX & PERM cell permeabilization kit (Caltag Laboratories) and stained for the presence of ICP4 using the monoclonal antibody 58 S (36) and anti-mouse fluorescein isothiocyanate (Sigma) following the manufacturer's instructions. Cells were subsequently analyzed by flow cytometry (FACS Calibur, BD Biosciences).

ICP0
Ubiquitinates p53 in Vitro-We have reported previously that full-length ICP0 has ubiquitin E3 ligase activity in vitro and generates unanchored polyubiquitin chains in the presence of E2-conjugating enzymes UbcH5a and UbcH6 (10). This activity requires ICP0 to have an intact RING finger domain. To test whether p53 could be ubiquitinated by ICP0, radiolabeled p53 was incubated in the presence or absence of a number of purified ubiquitin-conjugating enzymes (Fig. 1). Ubiquitinated p53 was readily detected in reactions containing ICP0 and a full complement of ubiquitin conjugation enzymes; ubiquitin, E1, and an E2, either UbcH5a or UbcH6. No significant p53 ubiquitination could be detected if either ICP0 or ubiquitin were omitted from the reaction mixture. The RING finger deletion mutant of ICP0, termed FXE (deletion of amino acids 106 -146), was inactive in this assay.
To determine whether the RING finger domain of ICP0 was sufficient to induce the ubiquitination of p53, reactions were carried out either in the presence of full-length ICP0 or in the presence of an N-terminal fragment of ICP0 containing the RING finger (amino acids 1-241) that has efficient E3 ligase activity in vitro (10). Ubiquitinated p53 was detected in complete reaction mixtures containing full-length ICP0 but not in those containing the RING finger fragment ( Fig. 2A), despite the fact that this fragment is able to produce unanchored polyubiquitin chains with activity comparable with that of fulllength ICP0 (Fig. 2B). This indicates that ubiquitination of p53 by ICP0 occurs in a substrate-specific manner and that ICP0 residues downstream of amino acid 241 are required for this activity.
To confirm that ICP0 acted directly on p53 and not through additional proteins found within the rabbit reticulolysate used for the production of the p53 substrate, GST-tagged p53 was purified from bacteria and was found to be ubiquitinated only in the presence of a full complement of ubiquitin-conjugating enzymes and ICP0 (Fig. 3). Therefore, ICP0 acts directly on p53 independently of other cellular proteins apart from those associated with the basic ubiquitination machinery.
ICP0 has been reported previously by others to contain an additional E3 ligase domain within the C-terminal region of the FIG. 1. ICP0 ubiquitinates p53 in vitro in a RING finger-dependent manner. Radiolabeled rabbit reticulocyte p53 was incubated in the presence of a combination of purified His-tagged E1, ubiquitin, His-tagged UbcH5a or UbcH6, and full-length His-tagged ICP0 or RING finger deletion mutant FXE as indicated. Reaction mixtures were incubated for 3 h prior to analysis by SDS-PAGE and autoradiography. The position of unmodified p53 is as indicated, and ubiquitinated (Ub) forms of p53 are indicated by a bar.

FIG. 2. ICP0 requires additional sequences downstream of its RING finger domain to ubiquitinate (Ub) p53 in vitro.
Ubiquitination reactions were carried out in the presence of full-length ICP0 or its isolated RING finger domain (GST-241). A, autoradiograph of the nitrocellulose membrane. B, Western blot of the same nitrocellulose membrane probed with a monoclonal anti-ubiquitin antibody. poly-Ub, polyubiquitin.
protein (amino acid residues 616 -680) (11,37). The authors reported that this domain of ICP0 stimulated the production of autoubiquitinated forms of the E2-conjugating enzyme cdc34 (UbcH3). However, in our original E2 screening assay using full-length ICP0, we could not detect the formation of any polyubiquitinated proteins in the presence of cdc34 (10). To test whether ICP0 stimulated the activity of cdc34 in a substratedependent manner, p53 ubiquitination reactions were performed utilizing a variety of different E2-conjugating enzymes. Ubiquitination of p53 was only detected in the presence of UbcH5a and UbcH6 and not with UbcH7, UbcH10, or cdc34 (Fig. 4), despite all the E2 preparations being able to form thiol-ester intermediates with ubiquitin (data not shown).
ICP0 Interacts with p53 Independent of Its USP7-binding Domain-ICP0 forms a strong and specific interaction with USP7, resulting in increased amounts of USP7 co-localizing with ICP0 in ND10 domains during the early stages of infection (9). This interaction has been mapped to a region containing two lysine residues within the C-terminal region of ICP0 (33). To test whether the USP7-binding domain is required for ICP0 to interact with p53, GST-p53 pull-down assays were performed using soluble extracts containing ICP0 or ICP0 mutant proteins (Fig. 5). GST-p53, but not GST, formed soluble complexes with full-length ICP0, the RING finger deletion mutant FXE (deletion of amino acids 106 -146), and the USP7 negative binding mutant E52X (deletion of amino acids 594 -775) (Fig.  5B). These data indicate that p53 binds to ICP0 via sequences that do not include its RING finger or USP7-binding domains. As p53 binds to E52X but is not ubiquitinated by the N-terminal RING finger domain of ICP0 (Fig. 2), the p53 interaction domain appears to be located between amino acids 241 and 594 of ICP0. Consistent with this interpretation, we found that the E52X protein ubiquitinated p53 with activity comparable with that of full-length ICP0 (data not shown), demonstrating that ICP0 does not require its USP7-binding domain, nor an interaction with USP7, to bind to and ubiquitinate p53.
ICP0 Ubiquitinates and Recruits p53 in Vivo-To determine whether ICP0 could ubiquitinate p53 in vivo, p53(Ϫ/Ϫ): mdm2(Ϫ/Ϫ) mouse cells were transfected with plasmids expressing p53, polyhistidine-tagged ubiquitin, mdm2, wild type ICP0, or various ICP0 mutants. The cells were treated with the proteasome inhibitor MG132 at 16 h after transfection and incubated for a further 8 h before analysis (Fig. 6). Ubiquiti-nated p53 in whole cell extracts could readily be detected in the presence of p53 and mdm2 (Fig. 6A) but not in the presence of any of the ICP0 proteins tested (Fig. 6B). As p53 undergoes numerous post-translational modifications, including phosphorylation, sumoylation, and acetylation, we investigated whether a minor subpopulation of p53 molecules could be ubiquitinated in vivo in response to ICP0. Ubiquitinated proteins were isolated by affinity chromatography from cells transfected as above and analyzed by Western blotting (Fig. 6C). Ubiquitinated p53 was readily detected in the presence of mdm2 and was reproducibly detected in the presence of ICP0 and its USP7 negative binding mutant M1, although the amount of ubiquitinated p53 produced by ICP0 was greatly reduced as compared with that induced by mdm2. Ubiquitination of p53 in vivo was not detected in the presence of the ICP0 RING finger deletion mutant FXE. Therefore, ICP0 can ubiquitinate a proportion of p53 molecules in vivo in a RING finger-dependent manner. Consistent with the in vitro ubiquitination data, ICP0-induced p53 ubiquitination in vivo does not occur through indirect binding via USP7 as ubiquitinated p53 was equally detectable in the presence of ICP0 mutant M1. Furthermore, the C-terminal region of ICP0 containing the entire USP7-binding domain (and putative cdc34 associated E3 ligase domain) was not required for ubiquitination of p53 in vivo as the mutant E52X also ubiquitinated p53 with activity comparable with that of wild type ICP0 in this assay (data not shown).
Partial Co-localization of p53 with ICP0 in Vivo-As p53 has been reported to be recruited to viral DNA replication compartments during HSV-1 infection (26), we determined the effects of ICP0 on the localization of endogenous p53. U2OS cells were infected with baculoviruses expressing either EYFP or EYFP-ICP0 driven from the HCMV IE promoter, and the cellular localization of p53 was analyzed by confocal microscopy (Fig. 7). Baculoviruses have proved to be an efficient method for the expression of recombinant proteins within certain mammalian cells (38). Expression of EYFP in U2OS cells had no effect on the localization of p53, which typically showed a microgranular nuclear staining pattern with EYFP expressed from a control baculovirus being diffuse (Fig. 7, A-C). Conversely, the expression of EYFP-ICP0 resulted in the recruitment of p53 to tight nuclear foci that co-localized with ICP0 (Fig. 7, D-F). As extended high level expression of ICP0 has been shown to induce both p53 and a number of p53-responsive genes (27), these data suggest that the induction, sequestration, and binding of p53 by ICP0 exceeded the rate of its ubiquitination and proteasome mediated degradation, at least in this system.
In a normal HSV-1 infection of HFFF-2 human fibroblasts, although in many cells there was no obvious co-localization of ICP0 with p53 (Fig. 7, G-I), in a subset of cells, p53 formed distinct nuclear foci (Fig. 7J). These foci were commonly associated with or juxtaposed to ICP0 foci, and in some instances, there was significant co-localization, as shown in Fig. 7L, insets. Thus at the early times of a normal HSV-1 infection, there is no gross co-localization of p53 with ICP0, although instances of co-localizing foci can be observed. After overnight infection of HFFF-2 with HSV-1 virus FXE at low multiplicity, however, extensive co-localization of a proportion of p53 with the RING finger deletion mutant ICP0 protein was observed (Fig. 7, M-O). Under these conditions, the majority of cells enter a non-productive infection characterized by expression of ICP0 but not later classes of viral gene products. 2 Thus ICP0 and p53 have the potential to exhibit extensive co-localization in infected cells, although in the 2 R. D. Everett, C. Boutell, and A. Orr, submitted for publication. course of a normal lytic infection, only a small proportion of the p53 and ICP0 signals coincide.
ICP0 Inhibits UV-induced Apoptosis in U2OS Cells-We next explored whether the ability of ICP0 to ubiquitinate and interact with a proportion of p53 molecules affected p53-mediated pathways. It was not possible to assess the effect of ICP0 on p53-activated transcription in model reporter assays as ICP0 strongly transactivates gene expression in transfected cells in a promoter-independent manner (2). In addition, ICP0 also inhibits the G 1 -S and G 2 -M stages of the cell cycle (27,39), precluding the analysis of the effect of ICP0 on p53-induced cell cycle arrest. Therefore, we investigated whether ICP0 could inhibit a p53 response to DNA damage. U2OS cells, in which ICP0 is not required for normally efficient HSV-1 infection (40) but which express wild type p53, were analyzed for their ability to undergo UV-induced apoptosis following infection. High levels of UV irradiation have been shown to induce the stabilization of p53, resulting in the induction of cycle arrest or apoptosis in response to DNA damage (reviewed in Ref. 41). Cells were either mock-infected or infected with wild type HSV-1 or dl1403, an HSV-1 ICP0-null mutant, in the presence of an inhibitor of viral DNA replication to stop the infection progressing beyond the early phase. At 3 h after infection, the cells were subjected to UV irradiation (50 J/m 2 ) and analyzed 24 h after irradiation for those undergoing apoptosis by annexin-V staining ( Fig. 8) (42)(43)(44). Unirradiated cells infected with dl1403 showed a small increase in the number of cells undergoing apoptosis as compared with either mock-or 17ϩ-infected cells (Fig. 8, A and B). This difference became significantly greater after cells had been subjected to UV irradiation 3 h after infection. Mock-and dl1403-infected cells showed a similar percentage increase in the number of cells undergoing apoptosis 24 h after irradiation, whereas wild type HSV-1-infected cells showed no such increase over the number of apoptotic cells in the unirradiated controls. FACS analysis of parallel cultures by staining for the IE3 gene product ICP4 showed that the degree of infection was equivalent in the dl1403 and wild type HSV-1-infected cultures (Fig. 8C). These results indicate that ICP0 expression is able to inhibit the apoptotic response to DNA damage in irradiated U2OS cells. DISCUSSION ICP0 has been shown to induce the degradation of a number of cellular proteins in vivo, including the major ND10 proteins PML (14) and Sp100 (17), the centromeric proteins CENP-A (18) and CENP-C (19), and the catalytic subunit of DNA-PK (20). The degradation of all of these proteins can be inhibited by the proteasome inhibitor MG132 and requires that ICP0 has an intact RING finger domain. Consistent with these observations, we have recently shown that purified full-length ICP0 has ubiquitin E3 ligase activity in vitro in the presence of the E2-conjugating enzymes UbcH5a and UbcH6 (10). These data imply that ICP0 acts as an ubiquitin E3 ligase that targets specific proteins for ubiquitination and proteasome-mediated degradation, although evidence for direct ubiquitination of specific substrates by ICP0 had been lacking. In this report, we demonstrate that ICP0 is a genuine RING finger ubiquitin E3 ligase that can mediate the ubiquitination of p53 both in vitro and in vivo.
Several DNA viruses have been shown to induce the ubiquitination of p53, for example the papillomavirus E6 protein in association with its cellular partner E6-AP (21). Here we provide the first example of a viral RING finger protein that interacts with and directly ubiquitinates p53 in a RING fingerdependent manner. Transfection analysis of p53/mdm2(Ϫ/Ϫ) cells that were subsequently treated with MG132 resulted in substantial accumulation of ubiquitinated p53 in response to co-expression of mdm2. Although ubiquitinated p53 could be isolated from cells co-transfected with plasmids expressing p53, ICP0, and His-tagged ubiquitin, ubiquitinated p53 was below the level of detection in whole cell extracts. These data imply that although ICP0 can ubiquitinate p53, its activity in this situation is low when compared with that of mdm2. Consistent with these observations, comparison of the levels of p53 in mock-infected or HSV-1-infected HFFF-2 cells, which express wild type p53, showed no decrease in the overall levels of p53 during infection (data not shown and Ref. 27). Indeed, normal lytic HSV-1 infection of HFFF-2 cells resulted in the phosphorylation and stabilization of cellular p53 in an ICP0independent manner (data not shown). Therefore, HSV-1 must express or induce factors other than ICP0 that result in the biochemical modulation of p53, and in the course of a normal lytic infection, these stabilizing factors may be dominant over any sequestration and ubiquitination that can be induced by ICP0. The absence of a strong p53 ubiquitination profile in the presence of ICP0 in vivo suggests that ICP0 might target a specific subpopulation of p53 molecules for degradation. For example, ICP0 might ubiquitinate a specific phosphorylated, acetylated, or sumoylated modified form of p53. Indeed, ICP0 has been shown to induce the efficient degradation of other post-translationally modified proteins, for example sumoylated PML and Sp100 (14 -17). Alternatively, ICP0 may only target p53 molecules at those specific cellular foci to which it localizes and where it induces the degradation of other cellular proteins. For example, p53 has been shown to associate with PML in ND10 and with DNA-PK on double-stranded DNA breaks (45)(46)(47). As ICP0 induces the degradation of both of these proteins, p53 may also become ubiquitinated as part of a complex of proteins targeted for degradation. The degradation of such a subpopulation of p53 molecules and associated proteins could prevent the initiation of p53-induced transcriptional responses that may otherwise be detrimental to viral replication, for example, the induction of p53-dependent cellular senescence or apoptosis. Indeed, we show here that ICP0 is required to inhibit apoptosis induced by irradiation of infected cells (Fig. 8).
Much interest has arisen from the recent observation that USP7 could interact with and deubiquitinate p53 (32,48). Our study was initiated because ICP0 also binds strongly to USP7. However, we found that the interaction formed between ICP0 and p53 occurs independently of its ability to bind to USP7. Although an explanation as to why ICP0 should bind so strongly to USP7 has yet to be identified, it is clear that in the absence of USP7 binding, ICP0 can still ubiquitinate p53 and mediate the degradation of other cellular proteins (17,20). Therefore, there is no evidence that ICP0 binds to USP7 to modulate its effects on p53.
Although our data clearly show that ICP0 ubiquitinates p53 quite efficiently in vitro and to a lesser extent in vivo, the significance of the effect is less clear. As cited above, several viruses have been shown to inactivate p53 by a number of different mechanisms, presumably to protect infected cells from p53-mediated responses such as apoptosis. Although HSV-1 has been reported to encode a number of proteins that inhibit infected cells from initiating an apoptotic response (reviewed in Ref. 1), we have found that ICP0 expression can protect HSV-1-infected U2OS cells from UV-induced apoptosis (Fig. 8). These results indicate a potential connection between ICP0, p53, and the DNA damage response pathway. Indeed, there are several potential connections between ICP0 and p53. For example, a number of proteins that regulate p53 (including PML, DNA-PK, and USP7) also interact with or are destabilized by ICP0 and have also been shown to be associated with p53-mediated apoptotic responses (32,45,47). In addition, the observation that ICP0 causes global decreases in SUMO-1 modified species (14,49) could also impinge on p53 transcriptional activity. Our observations that ICP0 can ubiquitinate a proportion of p53 molecules in vivo and protect cells against UV-induced apoptosis suggest that there may well be circumstances during the natural HSV-1 life cycle in which the effects of ICP0 on p53 become biologically significant.