Deubiquitination of Tip60 by USP7 Determines the Activity of the p53-Dependent Apoptotic Pathway

Abstract

Tip60 is an essential acetyltransferase required for acetylation of nucleosomal histones and other nonhistone proteins. Tip60 acetylates the p53 tumor suppressor at lysine 120 (K120), a modification essential for p53-dependent induction of PUMA and apoptosis. It is known that Tip60 is turned over in cells by the ubiquitin-proteasome system. However, the deubiquitinase activity for stabilizing Tip60 is unknown. Here we show that USP7 interacts with and deubiquitinates Tip60 both in vitro and in vivo. USP7 deubiquitinase activity is required for the stabilization of Tip60 in order to operate an effective p53-dependent apoptotic pathway in response to genotoxic stress. Inhibiting USP7 with the small-molecule inhibitor P22077 attenuates the p53-dependent apoptotic pathway by destabilizing Tip60. P22077, however, is still cytotoxic, and this is partly due to destabilization of Tip60.

INTRODUCTION

Tip60 is a crucial acetyltransferase required at multiple levels of gene transcription, DNA repair, and growth control by acetylating histones and nonhistone proteins (1). As a part of the NuA4 complex, Tip60 acetylates nucleosomal histones and thereby plays an important role in chromatin remodelling during different DNA transactions (2–4). Tip60 is ubiquitously expressed and is required at all levels of DNA damage response, i.e., sensing, signaling, and repair. In response to the double-strand breaks, Tip60 is directly recruited to break sites where it acetylates ATM and activates it (5). ATM phosphorylates H2AX (γ-H2AX) at the damaged site and thereby provides a platform for recruitment of repair proteins (1, 6, 7). At the final stage of repair, Tip60 acetylates H4 and γ-H2AX, an event required for the exchange of γ-H2AX-containing nucleosomes at the repair site with nucleosomes containing unphosphorylated H2AX (8, 9). Tip60 has also been shown to recruit ribonucleotide reductase at the damaged sites to enhance the pool of deoxynucleoside triphosphates (dNTPs) available locally during damage repair. This recruitment of ribonucleotide reductase is independent of Tip60’s acetyltransferase activity (10).

Tip60 plays a role in transcription activation by acetylation of histones on the target promoters (1, 4). Alternatively, some transcription regulatory factors such as p53, FOXP3, Myc, and androgen receptor (AR) are themselves substrates for Tip60’s acetyltransferase activity (11–14). The most important nonhistone target of Tip60 acetyltransferase activity is the p53 tumor suppressor. Following irreparable DNA damage, p53 induces apoptosis through the transcriptional activation of a BCL-2 family member, PUMA (15, 16). Tip60 acetylates lysine 120 (K120) in the DNA binding domain of p53, and this K120 acetylation is crucial for p53 to activate the transcription of PUMA (14, 17). PUMA is regarded as the main inducer of the p53-dependent apoptotic response (18, 19). Tip60 is also involved in UV-induced apoptosis in a manner independent of the p53 pathway (20).

Tip60’s role in cancer is complex, and it is deregulated in various cancers. Tip60 has been identified as a haploinsufficient tumor suppressor gene required for oncogene-induced DNA damage response (21). Mouse models which are haploid for Tip60 are prone to tumors. Consistent with a role as a haploinsufficient tumor suppressor, Tip60 is highly downregulated in breast (21) and colorectal (22) cancers. In contrast, Tip60 transcript and protein levels are upregulated in prostate cancer, where Tip60 promotes cell proliferation by translocating androgen receptor into the nucleus (23, 24).

Tip60 is turned over in cells by ubiquitin-mediated proteasomal degradation (25). Previously, Tip60 has been shown to be polybiquitinated by overexpressing exogenous Mdm2, which acts as a signal for its proteasomal degradation (26). However, Mdm2 depletion with small interfering RNA (siRNA) or inhibition of Mdm2 E3 ligase activity with Nutlin3 does not increase Tip60 levels under basal conditions (20), suggesting that some other protein, most likely a deubiquitinase (DUB), normally keeps Tip60 stable so that one can decrease the E3 ligase without significantly stabilizing the Tip60 any further. Besides the normal turnover activity, Tip60 is targeted by viral oncoproteins for proteasomal degradation in an Mdm2-independent manner to prevent the apoptotic death of the host cell. The human immunodeficiency virus (HIV) tat protein recruits CBP/p300 to induce the polyubiquitination and degradation of Tip60 in the proteasome (27). Similarly, human papillomavirus (HPV) E6 or adenovirus E1b55k and E4orf6 proteins promote Tip60 degradation via the proteasome (28, 29).

Many cellular proteins are stabilized posttranslationally by deubiquitination carried out by a class of enzymes called deubiquitinases (DUBs). DUBs remove the polyubiquitin chains from their substrates and thereby increase their cellular pool. In human cells, there are ∼100 DUBs (30). However, the cellular deubiquitinase activity for Tip60 is unknown. Using a biochemical and siRNA approach, we have identified ubiquitin-specific protease 7 (USP7 [or “HAUSP”]) as one of the deubiquitinases for Tip60. USP7 stabilizes Tip60 by deubiquitination and increases its half-life in cells. USP7 deubiquitinates and stabilizes Mdm2, and Mdm2 in turn ubiquitinates and degrades the p53 tumor suppressor (31, 32). Recent efforts have been focused on the discovery of small-molecule inhibitors of USP7 with the aim to stabilize p53 (33, 34). Interestingly, we found that inhibiting USP7 deubiquitinase activity with a small-molecule inhibitor, P22077, attenuates the p53-dependent apoptotic pathway by destabilization of Tip60, suggesting that this strategy of restoring the tumor suppressor function of p53 is not likely to work. We explain that USP7 acts as a master protein regulating all members of the complex circuit (Tip60, Mdm2, and p53) involved in the p53-mediated apoptotic pathway. In a surprising twist, however, we show that although the p53-dependent apoptotic pathway is attenuated, P22077 is still cytotoxic, in part through the destabilization of Tip60.

MATERIALS AND METHODS

Cell culture and transfection.

U2OS and 293T cells were cultured in Dulbecco’s modified Eagle's medium (DMEM) containing 10% donor calf serum. H1299 cells were grown in DMEM containing 10% fetal bovine serum (FBS), whereas HCT116 cells were maintained in McCoy's 5A medium supplemented with 10% FBS. siRNA reverse transfection (1 to 10 nM) into cells (25 × 10⁵ cells in 10-cm-diameter dishes) was carried out by using RNAiMax reagent (Invitrogen) following the manufacturer's instructions. The sequences of the different double-stranded siRNAs used are given in Table S2 in the supplemental material. Plasmids were transfected in 293T cells by using Lipofectamine 2000 reagent (Invitrogen) and in H1299 cells by using Lipofectamine LTX and Plus reagents (Invitrogen).

Plasmid constructions.

For in vivo studies, Tip60 was cloned downstream from a myc tag at BamHI and XhoI sites of modified pcDNA 3.1 vector. For bacterial expression, Tip60 was amplified by PCR using a forward primer with a BamHI restriction site and a reverse primer with an XhoI site. The PCR product was cloned downstream of a MBP tag in Malc2x vector at the BamHI and SalI sites. USP7 and USP47 were amplified with PCR and cloned into LPCX destination vector by a gateway recombination method. For bacterial purification of His-tagged USP7 and its various domains, the corresponding DNA fragments were PCR amplified and cloned into the NheI and XhoI sites of pET21c vector. Site-directed mutagenesis was used to convert wild-type USP7 into a catalytic mutant form by mutating cysteine 223 into serine.

Preparation of cytoplasmic and nuclear extracts.

Cells were cultured and harvested when in log phase. After harvesting, cells were washed two times with phosphate-buffered saline (PBS). To prepare the cytoplasmic fraction, 1 volume of packed cells was gently resuspended into 10 volumes of cold cytoplasmic extraction buffer (10 mM HEPES [pH 7.9], 1.5 mM MgCl₂, 10 mM KCl, 0.5 mM dithiothreitol [DTT], and 1× protease inhibitor cocktail [Sigma]) and left on ice for 5 min. NP-40 was added to achieve a final concentration of 0.3% and mixed with the lysate immediately by pipetting the mixture in and out 3 times. The lysate was further incubated on ice for 3 min and again pipetted in and out 3 times followed by centrifugation at 16,000 × g for 5 min. The supernatant was isolated and labeled as the cytoplasmic extract. The remaining pellet was resuspended in 5 volumes of nuclear extraction buffer (20 mM HEPES [pH 7.9], 25% glycerol, 420 mM KCl, 1.5 mM MgCl₂, 0.2 mM EDTA, 1 mM DTT, and 1× protease inhibitor cocktail [Sigma]). The lysate was left on ice for 40 min with 10 s of intermittent vortexing every 10 min. The nuclear extract was separated by centrifugation at 16,000 × g for 10 min. Both the cytoplasmic and the nuclear extracts isolated were dialyzed against 25 mM Tris-Cl (pH 7.5), 5% glycerol, 150 mM KCl, 1 mM EDTA, 1 mM DTT, 0.1% NP-40, and phenylmethylsulfonyl fluoride (PMSF) and stored at −80°C.

Coimmunoprecipitation.

Cells were lysed in lysis buffer (50 mM Tris-Cl [pH 8.0], 10% glycerol, 1 mM EDTA, 0.1% NP-40, 1 mM DTT, protease inhibitors) by incubation on ice for 15 min followed by centrifugation at 15,000 rpm for 10 min. The supernatant recovered was immunoprecipitated with the antibodies used.

Purification of recombinant proteins from bacteria.

Escherichia coli BL21 cells were transformed with MALc2X or MALc2x-Tip60, pET21c-USP7, and other pET21c-USP7 variants. Cells were cultured at 37°C until an optical density at 595 nm (OD₅₉₅) of 0.4 to 0.5 was reached. Cells were induced with 0.25 mM isopropyl-β-d-thiogalactopyranoside (IPTG) for 14 to 16 h at 22°C and harvested. All proteins were purified under native conditions.

In vitro MBP pulldown assay.

Bacterially purified MBP-Tip60 or MBP coupled to amylose beads was incubated with recombinant His-tagged USP7 and its different domains in the pulldown buffer (25 mM HEPES [pH 7.5], 150 mM NaCl, 5 mM MgCl₂, 1 mM DTT, 0.5 mM EDTA, 5% glycerol, and 0.05% NP-40) for 2 h at 4°C. The beads were washed 3 times with a buffer (25 mM HEPES [pH 7.5], 150 mM NaCl, 5 mM MgCl₂, 0.5 mM EDTA, and 0.5% NP-40) and boiled in 2× SDS sample buffer. The samples were analyzed with Western blotting using anti-His antibody to detect USP7 binding to MBP-Tip60.

Deubiquitination assay in vivo and in vitro.

For the in vivo deubiquitination assay, the plasmid-transfected 293T cells were treated with MG132 (40 μM) for 1 h before harvesting. Cells were harvested in denaturing ubiquitination buffer (50 mM Tris-Cl [pH 8.0], 5 mM DTT, and 1% SDS) and immediately boiled for 10 min at 95°C followed by cooling on ice for 10 min. The lysate was sonicated and supernatant recovered after centrifugation at 15,000 rpm for 20 min. The supernatant was diluted with 9 volumes of buffer containing 50 mM Tris-HCl (pH 8.0), 150 mM KCl, 5% glycerol, 0.4% NP-40, and protease inhibitors and subjected to immunoprecipitation followed by Western blotting.

For the in vitro deubiquitination assay, cells were lysed in radioimmunoprecipitation assay (RIPA) buffer followed by immunoprecipitation. The protein G-coupled antibody-bound ubiquitinated myc-Tip60 was used as a substrate for the in vitro deubiquitination assay. Briefly, ubiquitinated myc-Tip60 was incubated in a 10-μl reaction mixture with nuclear or cytoplasmic extracts (6 μg) or 1 μg of bacterially purified His-USP7/His-USP7(C223S) in deubiquitination buffer (50 mM Tris-Cl [pH 8.0], 50 mM NaCl, 1 mM EDTA, 5% glycerol, 10 mM DTT)at 37°C for 3 h. The beads were washed 3 times with wash buffer (50 mM Tris-Cl [pH 8.0], 150 mM NaCl, 0.4% NP-40) and resuspended in 2× SDS sample buffer. The samples were boiled and analyzed for ubiquitinated myc-Tip60 by Western blotting.

Western blotting and antibodies.

For Western blotting of cell lysates, cells extracts were prepared by lysing cells in RIPA buffer containing 1× protease inhibitor cocktail (Sigma). For detection of K120ac on endogenous p53, cells were treated with 1.0 μM trichostatin A (TSA) and 5.0 mM nicotinamide for 6 h before harvesting. RIPA-extracted lysate from control siRNA (GL2)-treated cells (900 μg) and 450 μg of lysate from USP7 siRNA-treated cells were used. For detection of K120ac on hemagglutinin (HA)-p53, cells were treated with TSA and nicotinamide (as described above) followed by immunoprecipitation using equal amounts of RIPA-extracted lysates.

The antibodies and sources were as follows: Tip60 (rabbit) (29), Mdm2 (Santa Cruz), actin (Santa Cruz), tubulin (Santa Cruz), USP7 (Cell Signaling), USP47 (Bethyl), p53/DO-1 (Santa Cruz), p53K120ac (Abcam), PUMA (Sigma or Cell Signaling), caspase 3 (Cell Signaling), H2AK5ac (Cell Signaling), H2A (Cell Signaling), H4 and H4K5ac (Cell Signaling), His6 (Qiagen), and HA (Santa Cruz).

Stable cell lines.

U2OS cells were transduced with retrovirus to generate stably expressing Flag-Tip60 and control cell lines.

Real-time PCR.

Total RNA was isolated by the use of TRIzol reagent (Invitrogen) according to the manufacturer’s instructions and used for cDNA synthesis with Superscript III (Invitrogen). The cDNAs were used as the templates for real-time PCR using SYBR green PCR master mix (Applied Biosystems). Sequences of primers used for RT-PCR analysis are given (in the 5′→3′ direction) in Table S1 in the supplemental material.

Chemicals.

Doxorubicin was purchased from Sigma and used at 0.1 μM, and P22077 was purchased from Axon MedChem and used at 20 to 30 μM.

Cell viability.

Cell viability was monitored by the trypan blue exclusion method using 0.4% trypan blue solution (Invitrogen) and a Countess automated cell counter (Invitrogen).

RESULTS

Identification of deubiquitinating activity for Tip60.

Tip60 is turned over in cells in a proteasome-dependent manner (25). We treated cells with MG132 (proteasome inhibitor) to confirm that blockage of proteasome increased the level of endogenous Tip60 (Fig. 1A). Indeed, coexpressing myc-Tip60 and HA-ubiquitin in 293T cells showed that Tip60 was polyubiquitinated in vivo (Fig. 1B). Myc-Ub-Tip60 (myc-tagged ubiquitinated Tip60) was purified from the cell extracts under denaturing heat conditions (Fig. 1B, left panel) or native conditions (with RIPA buffer) (Fig. 1B, right panel). The natively purified antibody-bound myc-Ub-Tip60 immobilized on protein G-Sepharose beads was later used as the substrate for deubiquitination reactions in vitro.

Fig 1.

Identification of a deubiquitinase that acts on Tip60. (A) 293T cells were treated with dimethyl sulfoxide (DMSO) or MG132 for 2 h and total cell lysates immunoblotted. (B) In vivo ubiquitination of Tip60 in the 293T cells. Polyubiquitinated myc-Tip60 was immunopurified with anti-myc antibody from 293T cells in either denaturing ubiquitination buffer (left panel) or RIPA buffer (right panel). The blots were first probed with anti-HA antibody to detect HA-ubiquitinated myc-Tip60 (top) and later reprobed with anti-myc antibody to detect the immunoprecipitated (IP) myc-Tip60 (bottom). (C) Immunoblots of cell fractionation. Tip60 is exclusively present in nuclear fraction of HCT116 cells. Cyt., cytoplasm; Nuc., nucleus. (D) Scheme for preparation of cytoplasmic and nuclear fractions. Ex, extract. (E) Dialyzed cytoplasmic and nuclear fractions (6 μg each) were used in the in vitro deubiquitination assay. The immunopurified polyubiquitinated myc-Tip60 (with RIPA buffer) was used as the substrate for the deubiquitination reaction in vitro. The blot was probed with anti-HA and anti-myc antibodies as described in the panel B legend. (F) Venn diagram showing the presence of different ubiquitin-specific proteases (USPs) in nuclear or cytoplasmic compartments. Some of the USPs are localized in both the nucleus and the cytoplasm. (G) U2OS cells stably expressing Flag-Tip60 were treated with control GL2 siRNA (si) or siRNA against the indicated USPs for 36 h followed by immunoblotting using two-thirds of each cell pellet. (H) One-third of each cell pellet obtained from the experiment described in the panel G legend was used for measurement of mRNA of the indicated USPs using qRT-PCR. Actin was used as the control. Data represent the means ± standard deviations (SD) of three measurements.

Fractionation of cells into nuclear and cytoplasmic fractions revealed that Tip60 was exclusively present in the nucleus (Fig. 1C). Nuclear and cytoplasmic fractions prepared as shown in the scheme in Fig. 1D were used in an in vitro deubiquitination assay on antibody-bound myc-Ub-Tip60 prepared by coexpressing myc-Tip60 and HA-Ub in 293T cells (Fig. 1B). The results show that most of the deubiquitinase activity that acted on Tip60 was localized to the nucleus (Fig. 1E).

Ubiquitin-specific proteases (USPs) are a class of DUBs (30). Since the nuclear fraction contained most of the deubiquitination activity for Tip60 (Fig. 1E), we focused on nuclear USPs. siRNAs specific to 10 different nuclear USPs were tested to discover that USP7 knockdown specifically decreased the levels of exogenous Flag-Tip60 (Fig. 1F and G). Knockdown of USPs was confirmed by quantitative reverse transcription-PCR (qRT-PCR) analysis of their respective transcripts after transfection of cells with control (GL2) and USP-specific (Fig. 1H) siRNAs. These results suggest that USP7 is a deubiquitinase for Tip60, although we cannot rule out the possibility that there are other deubiquitinases that act on Tip60 under other conditions.

USP7 stabilizes and increases the half-life of Tip60.

To test the effect of USP7 on the stability of Tip60, we transiently transfected 293T cells with plasmids expressing myc-Tip60 and USP7. USP7 coexpression significantly increased the steady-state levels of Tip60 (Fig. 2A and B). In contrast, coexpression of USP47, which shares considerable homology with USP7 (33), had no obvious effect on the Tip60 protein level (Fig. 2C). Next, we measured the half-life of ectopically expressed Tip60 using cycloheximide, an inhibitor of protein translation. Exogenous myc-Tip60 has a half-life of ∼30 min when expressed along with empty vector. The half-life was significantly increased when Tip60 was coexpressed with USP7 (Fig. 2D and E). To test whether the increase in half-life by USP7 was specific to Tip60, we expressed myc-Set8 along with myc-Tip60 and USP7 and treated the cells with cycloheximide for 30 min. USP7 increased the basal level and stabilized myc-Tip60 but had no effect on the basal level or stability of myc-Set8 (Fig. 2F).

Fig 2.

USP7 stabilizes Tip60. (A) USP7 increases myc-Tip60 levels in cells. The results of Western blot analysis of 293T cells transfected with plasmids expressing myc-Tip60 alone or myc-Tip60 plus increasing amounts of USP7 are shown. (B) Western blot analysis of 293T cells transfected with plasmids expressing myc-Tip60 alone or myc-Tip60 along with increasing amounts of USP7-expressing plasmid. Green fluorescent protein (GFP)-expressing plasmid was used to determine that the cells received equal amounts of DNA during transfection. (C) Western blot analysis of 293T cells transfected with plasmids expressing myc-Tip60 alone or myc-Tip60 along with increasing amounts of USP47. The remainder of the conditions were as described in the panel B legend. (D) USP7 increases the half-life of Tip60. 293T cells were transfected with either myc-Tip60 plus empty vector or myc-Tip60 plus USP7 vector. After 48 h of transfection, cells were treated with cycloheximide (CHX; 80 μg/ml) and harvested at different time points as indicated. Immunoblots of the indicated proteins are shown. (E) The myc-Tip60 signal described in the panel D legend was quantitated and plotted relative to the actin signal to measure the half-life of Tip60 with or without coexpressed USP7. (F) USP7 specifically increases and stabilizes Tip60. 293T cells were transfected with myc-Tip60, myc-Set8, and USP7 vectors as indicated. After 48 h of transfection, cells were either mock treated or treated with cycloheximide for 30 min, harvested, and immunoblotted. (G) Two different siRNA oligonucleotides specific to USP7 were transfected into U2OS cells followed by Western blotting as indicated. (H) USP7 knockdown decreases endogenous Tip60. Different cell types as indicated were treated with control siRNA (GL2) or siRNA against USP7 for 36 h and immunoblotted with the indicated antibodies. (I) USP7 knockdown does not decrease Tip60 mRNA levels. Cells were treated with control (GL2) or USP7-specific siRNA for 36 h. RNA transcripts were measured by qRT-PCR. A glyceraldehyde-3-phosphate dehydrogenase (GAPDH) transcript was used a control.

Two different USP7-specific siRNA oligonucleotides were used to transiently knockdown USP7 in U2OS osteosarcoma cells (Fig. 2G). Transient knockdown of USP7 decreased endogenous Tip60 levels in U2OS osteosarcoma cells and HCT116 wild-type and p53-null colon cancer cells (Fig. 2H; see also Fig. 5A). The Tip60 mRNA level was not decreased by USP7 knockdown (Fig. 2I). These results show that USP7 stabilizes Tip60 protein in a variety of cells and in a manner independent of the presence of p53.

Fig 5.

Regulation of Tip60 function by USP7. (A) USP7 knockdown decreases Tip60 and associated functions. U2OS cells were treated with control (GL2) or USP7-specific siRNA. The lysates were immunoblotted with the indicated antibodies. (B) USP7 knockdown decreases K120ac levels on p53. U2OS cells were treated with control (GL2) or USP7-specific siRNA for 48 h and immunoprecipitated with p53-specific DO-1 antibody. The blot was first probed with K120ac-specific antibody and later reprobed with p53 antibody. Equal amounts of lysates were loaded in input lanes to show the decrease of Tip60 levels by USP7 knockdown. (C) H1299 (p53^−/−) cells were transfected with the indicated combinations of plasmids expressing HA-p53, myc-Tip60, USP7, or USP7(C223S) for 48 h. The anti-HA immunoprecipitate was probed with K120ac antibody and p53-specific DO-1 antibody. (D) H1299 cells were transfected with indicated plasmids for 60 h. After 36 h of plasmid transfection, cells were treated with DMSO (DXR−) or 0.1 μM doxorubicin (DXR+) for 24 h. The PUMA mRNA was measured by qRT-PCR. Actin was used as internal control. (E) H1299 cells were transfected with indicated plasmids and processed for PUMA/actin transcript measurement as described in the panel D legend.

USP7 interacts with Tip60.

To evaluate if USP7 interacts with Tip60 in vivo, we transfected 293T cells with myc-Tip60 and USP7-expressing plasmids. Immunoprecipitation of myc-Tip60 with anti-myc antibodies from cell lysates readily coprecipitated USP7 (Fig. 3A). We next tested whether endogenous Tip60 interacts with endogenous USP7. Immunoprecipitates of Tip60 from cell extracts coimmunoprecipitated endogenous USP7 (Fig. 3B).

Fig 3.

USP7 interacts directly with Tip60. (A) 293T cells were transfected with plasmids expressing myc-Tip60 and USP7 (without tag)-expressing plasmids. Cell lysates were immunoprecipitated with anti-myc or control IgG antibodies. The immunoprecipitates were blotted with the indicated antibodies. Input, 3% of the lysate used in the IP reaction. (B) Coimmunoprecipitation of endogenous Tip60 and USP7. Cell extracts from U2OS cells were immunoprecipitated with anti-Tip60 or control IgG antibodies and the precipitates immunoblotted with anti-Tip60 or anti-USP7 antibodies. The lysate (0.48%) used for immunoprecipitation was loaded in the input lane. (C) Domain organization of USP7. (D) In vitro MBP pulldown assay. Purified bacterially produced His-tagged USP7 (full length or different domains) was incubated with MBP-Tip60 or MBP alone immobilized on amylose beads. After washing, proteins bound to beads were resolved on a 10% SDS gel and the blot was probed with anti-His antibody (upper blot). The lower blot is the Ponceau S-stained membrane showing the loading of MBP-Tip60 and MBP proteins used in the pulldown assay. IB, immunoblot.

To evaluate whether the USP7 interaction with Tip60 is direct, we generated different recombinant constructs and purified MBP-tagged Tip60 and His-tagged USP7 (full length and different domains) from E. coli. Purified MBP-Tip60 pulled down full-length His-USP7 whereas MBP alone did not interact with USP7 (Fig. 3D). When different domains of USP7 (Fig. 3C) were tested in the pulldown assay, the N-terminal TRAF domain of USP7 (amino acids 1 to 206) was necessary and sufficient for interaction with Tip60 (Fig. 3D), with very weak interaction with the USP7 catalytic domain (CAT) (Fig. 3C and D).

USP7 deubiquitinates Tip60.

To determine whether the catalytic activity of USP7 is required to stabilize Tip60, we mutated the catalytic cysteine residue (C223S) in USP7 followed by coexpression of myc-Tip60 with either wild-type or catalytically dead USP7. Wild-type USP7 stabilized Tip60 whereas catalytic mutant (C223S) decreased the Tip60 levels, possibly because of a dominant-negative effect of catalytic mutant USP7 on the endogenous wild-type USP7 (Fig. 4A). Therefore, the stabilization of Tip60 by USP7 might require the deubiquitination activity of USP7. Consistent with this, wild-type USP7 abrogated the polyubiquitination of myc-Tip60 in cells, whereas the catalytic mutant of USP7 had no effect on the polyubiquitination of Tip60 (Fig. 4B). Notably, another closely related ubiquitin-specific protease, USP47, did not decrease the polyubiquitination of Tip60 (Fig. 4C). Therefore, USP7 decreases Tip60 polyubiquitination in cells.

Fig 4.

USP7 deubiquitinates Tip60 in vivo and in vitro. (A) Catalytic activity of USP7 required to stabilize Tip60. myc-Tip60 was coexpressed with either wild-type USP7 or catalytically dead mutant USP7(C223S) in 293T cells for 48 h followed by immunoblotting of lysates with the indicated antibodies. Three times more lysate was loaded in the two lanes on the right than in the two lanes on left as indicated by actin levels. (B) USP7 deubiquitinates Tip60 in vivo. 293T cells were transfected with indicated plasmids for 48 h and treated with MG132 for an hour before harvesting. Myc-Tip60 was immunoprecipitated, and the precipitates were immunoblotted with anti-HA antibody to detect ubiquitinated forms of myc-Tip60. Reprobing with anti-myc antibody showed the nonubiquitinated myc-Tip60. The bottom three panels show the input (0.5%) of lysates used in immunoprecipitation. (C) 293T cells were transfected with different plasmid combinations expressing myc-Tip60 plus HA-Ub, myc-Tip60 plus HA-Ub plus USP7, or myc-Tip60 plus HA-Ub plus USP47 for an in vivo deubiquitination assay as described in the panel B legend. (D) Wild-type His-tagged USP7 and catalytic mutant USP7(C223S) purified from E. coli BL21. An SDS-PAGE gel stained with Coomassie brilliant blue is shown. (E) In vitro deubiquitination assay for Tip60. Polyubiquitinated myc-Tip60 was immunoprecipitated from 293T cells using RIPA extraction and myc antibodies and incubated with purified His-USP7 or catalytic mutant His-USP7(C223S). The blot was probed with HA (HA-ubiquitin) antibody to detect HA-ubiquitinated myc-Tip60 followed by reprobing with myc antibody to show immunoprecipitated myc-Tip60.

In order to directly examine the deubiquitination activity of USP7 toward Tip60, we utilized a cell-free system. We first purified recombinant His-USP7 and its catalytically dead form (C223S) from E. coli (Fig. 4D) and used antibody-bound bead-immobilized myc-Ub-Tip60 as the substrate in the deubiquitination assay in vitro. Purified wild-type USP7 but not catalytic mutant USP7 deubiquitinated Tip60 in vitro (Fig. 4E).

USP7 regulates Tip60 function.

We next evaluated whether knockdown of USP7 and the attendant decrease of Tip60 decreases the acetylation of key substrates of Tip60 in U2OS cells. USP7 knockdown with siRNA decreased Tip60 levels, which in turn decreased the acetylation of histones H2A (K5) and H4 (K5), whereas total histone levels remained unchanged under these conditions (Fig. 5A).

The acetylation of p53 on K120 by Tip60 is important for proapoptotic function of the p53 (14). Since USP7 deubiquitinates Tip60, we hypothesized that USP7 should have an effect on p53 K120 acetylation levels via Tip60. USP7 knockdown with siRNA decreased Tip60 levels, which in turn decreased the endogenous K120 acetylation of p53 (Fig. 5B). We also transfected H1299 cells (p53^−/−) with plasmids expressing HA-p53, myc-Tip60, USP7, and USP7(C223S) as indicated in Fig. 5C. Immunoblotting of HA-p53 immunoprecipitates with DO-1 anti-p53 antibody showed equal levels of p53 in all lanes. Cells expressing p53 and Tip60 together showed an increase in p53 K120 acetylation. The acetylation level was significantly increased upon coexpression of wild-type USP7 but not of the catalytic mutant USP7(C223S). These results show that deubiquitinase activity of USP7 enhances the Tip60-dependent K120 acetylation of p53.

USP7 enhances p53-dependent PUMA induction in a Tip60-dependent manner.

The acetylation of p53 at lysine 120 increases the proapoptotic transcriptional activity of p53 (14). H1299 (p53^−/−) cells treated with a DNA-damaging agent, doxorubicin (DXR +), for 24 h induced puma when p53 was present (Fig. 5D, column 4). The expression of puma was further increased in cells expressing p53 along with either Tip60 or USP7 (Fig. 5D, columns 6 and 8). The mild stimulation of puma expression with USP7 alone most likely occurred via the stabilization of endogenous Tip60. The highest levels of puma transcripts were observed in cells expressing p53 and Tip60 along with wild-type USP7 (Fig. 5D, column 10) but not with catalytically mutant USP7(C223S) (Fig. 5D, column 12). To determine whether the induction of puma by USP7 is mediated by the enzymatic activity of the Tip60 acetyltransferase on K120 of p53, we transfected H1299 cells with the combination of plasmids indicated in Fig. 5E. The highest stimulation in puma expression by p53 with Tip60 and USP7 (Fig. 5E, column 5) was compromised when wild-type Tip60 was replaced with the Tip60 catalytic mutant (K327R) (Fig. 5E, column 7) or when wild-type p53 was replaced with the p53-K120R mutant (Fig. 5E, column 8). These results indicate that USP7 stimulates p53-mediated puma expression through Tip60 acetyltransferase activity on p53 at K120.

A USP7 inhibitor (P22077) promotes Tip60 degradation and attenuates the p53-dependent apoptotic pathway.

USP7 stabilizes Mdm2, the E3 ligase for p53 (31). siRNA against USP7 or P22077, a recently identified USP7 inhibitor (33), decreases Mdm2 levels and increases p53 levels in cells without exogenous DNA damage, because the primary function of Mdm2 is to polyubiquitinate and target p53 for degradation (Fig. 5A and 6A). P22077 also reduced Tip60 protein in a manner consistent with the role of USP7 in stabilizing Tip60 (Fig. 6B). Although treatment of undamaged cells with P22077 increased p53 levels, PUMA protein and mRNA levels were decreased due to degradation of Tip60 acetyltransferase (Fig. 6B and C). Thus, inhibition of USP7 deubiquitination activity with P22077 and subsequent destabilization of Tip60 compromise the basal transcription of PUMA activated by basal levels of p53. To confirm this result, we turned to the cells where USP7 is specifically knocked down by siRNA. Although USP7 knockdown decreased Mdm2 and increased p53 protein levels (Fig. 5A), and although Mdm2 knockdown increased PUMA levels in undamaged cells (Fig. 6D), USP7 knockdown failed to induce PUMA (Fig. 6D). We believe that this is due to the concurrent decrease of Tip60 levels that is seen when USP7 is knocked down. Indeed, the induction of PUMA seen when Mdm2 was knocked down was lost when Tip60 was concurrently knocked down (Fig. 6D).

Fig 6.

P22077 destabilizes Tip60 and inhibits p53-dependent PUMA synthesis. (A) U2OS cells were treated with either DMSO or P22077 (25 μM) for 24 h, and the lysates were probed with antibodies as indicated. (B) U2OS cells were treated with P22077 for two different time periods as indicated. The lysates were probed with the indicated antibodies. (C) U2OS cells treated with DMSO or P22077 as indicated. The PUMA and actin mRNA levels were measured by qRT-PCR. (D) U2OS cells were transfected with control (GL2) and other siRNAs as indicated for 48 h, and the lysates were blotted for PUMA and tubulin (loading control). (E) U2OS cells were treated with P22077 (25 μM) or doxorubicin (0.1 μM) or both drugs together for 24 h. The lysates were analyzed with the indicated antibodies. (F) U2OS cells are treated with P22077 and doxorubicin in the same way as described in the panel E legend, and the PUMA and actin mRNA levels were measured by qRT-PCR. (G) U2OS cells were treated with control or USP7-specific RNA for 60 h. Before harvesting, cells were treated with DMSO or doxorubicin (0.1 μM) for 24 h. Two-thirds of each cell pellet was used for immunoblotting with the indicated antibodies. (H) One-third of each of the cell pellets obtained as described in the panel G legend was used for puma/actin transcript measurement by using qRT-PCR. (I) U2OS cells were treated with P22077 (30 μM) or doxorubicin (0.1 μM) or both the drugs together for 10 h. The cells were also treated with TSA (1.0 μM) and nicotinamide (5.0 mM). RIPA buffer-extracted cell lysates were immunoprecipitated with p53-specific (DO-1) antibody and the blots probed with K120ac-specific antibody and with DO-1 antibody. Twice as much lysate was used for immunoprecipitation in lane 1 as was used in the other three lanes so as to visualize the p53 and the K120ac under basal conditions. The lower panels show the immunoblots for the whole-cell extract (WCE), loaded in equal amounts in all four lanes. (J) U2OS cells stably transfected with control vector or Flag-Tip60 were treated with P22077 or doxorubicin or both as described in the panel E legend. The cell lysates were probed with the indicated antibodies. (K) The experimental setup was the same as that described in the panel J legend. The puma and actin levels were measured by qRT-PCR. (L) Schematic showing that USP7 regulates the p53-dependent apoptotic pathway in response to DNA damage. USP7 is responsible for Tip60 stability, and in response to genotoxic stress, USP7 increases p53. Tip60 carries out K120 acetylation of p53, directing it to proapoptotic PUMA transcription. Under basal conditions, USP7 stabilizes Mdm2, which in turn downregulates p53 by polyubiquitination. There is a positive feedback loop between p53 and Mdm2 in DNA-damaged cells, where Mdm2 becomes an activator of p53 translation.

A recent paper highlighted a switch of Mdm2 after DNA damage, when Mdm2 becomes a promoter of p53 translation and thus increases p53 levels instead of decreasing them by polyubiquitination (35, 36). We were, therefore, interested in testing whether USP7 inhibition with P22077 would still be antiapoptotic in cells where DNA damage had been induced with doxorubicin. In the presence of doxorubicin, p53 was induced along with the Mdm2. Under these conditions, Tip60 was intact and the increase in the p53 level resulted in induction of PUMA and subsequent apoptosis as indicated by the presence of cleaved caspase 3 (Fig. 6E, lane 3). Treating cells with P22077 inhibited USP7 deubiquitinase activity, resulting in destabilization of Mdm2 and Tip60 (Fig. 6E, lane 4). In support of the work of Gajjar et al. (35), although P22077 decreased Mdm2 when doxorubicin was present, it did not further increase p53 levels, most likely because Mdm2 was now not a repressor but an inducer of p53. The p53-dependent apoptotic pathway was not induced, as indicated by failure of PUMA induction and very low cleavage of caspase 3 (Fig. 6E, lane 4). The PUMA transcript was not induced by doxorubicin in the presence of P22077 (Fig. 6F). Thus, even in the presence of DNA damage, P22077 is antiapoptotic. To rule out off-target effects of P22077 on the p53-dependent apoptotic pathway, we used USP7-specific siRNA in U2OS cells and measured the effect of USP7 knockdown on downstream target proteins in the presence and absence of doxorubicin. As seen with P22077, siRNA against USP7 decreased PUMA at the protein level (Fig. 6G) and the transcript level (Fig. 6H) by decreasing the Tip60 level (Fig. 6G).

The failure to induce apoptosis when cells were treated with doxorubicin and P22077 could be attributed to the failure to superinduce p53 (because Mdm2 was now an activator of p53) or to the degradation of Tip60, the acetyltransferase required for K120 acetylation of p53 to target the latter to proapoptotic genes. The doxorubicin-mediated increase of K120 acetylation on p53 was compromised when P22077 inhibited USP7 and downregulated Tip60 (Fig. 6I), suggesting that the decrease of the Tip60 level was responsible for the failure to induce apoptosis.

We therefore asked whether overexpression of Tip60 can rescue PUMA synthesis in cells treated with doxorubicin and P22077. In a U2OS cell line stably expressing Flag-Tip60, blocking USP7 with P22077 drastically decreased endogenous Tip60 levels. Although the overexpressed Flag-Tip60 was also degraded by the inhibition of USP7, enough Flag-Tip60 remained to rescue PUMA induction (Fig. 6J, lane 8 compared to lane 4). The rescue of PUMA mRNA expression was confirmed by quantitative RT-PCR (Fig. 6K). The induction of PUMA is achieved without any increase in the total level of p53, suggesting that P22077 blocked apoptosis not because of a lack of p53 but because of a lack of Tip60 to acetylate the p53 on K120.

Together, these data suggest that USP7 deubiquitination activity is required for the stability of Tip60 and for the functioning of the p53-dependent apoptotic pathway in the absence or presence of DNA damage (Fig. 6L).

Decrease of Tip60 levels contributes to the cytotoxicity of USP7 inhibitor (P22077).

Inhibitors of USP7 are being developed as anticancer agents at least partly because, as in the case of USP7 knockouts, it is expected that p53 would be induced by USP7 inhibition and that this would lead to apoptosis (33, 34). However, our results suggest that this is unlikely to be the case because of the complication that USP7 is required to stabilize Tip60. Despite attenuating the p53-dependent apoptotic pathway through the destabilization of Tip60, however, P22077 induced cell death in U2OS cells (Fig. 7A). Very recently, P5091, another USP7 inhibitor, has been shown to induce cell death in multiple myeloma cells in a p53-independent manner but HCT116 p21^−/− cells were shown to be resistant to cytotoxic effects of P5091 (37). In our hands, however, P22077 was equally toxic to HCT 116 p53^+/+, HCT 116 p53^−/−, and HCT116 p21^−/− cells (Fig. 7B). Thus, the toxicity of P22077 was not mediated by p53 or p21. Tip60 levels were decreased by P22077 in a manner independent of the cell’s p53 or p21 status (Fig. 7C). A decrease in cell viability was also seen after transfection of two different siRNAs against USP7 (Fig. 7D and E), supporting the suggestion that at least part of the toxicity of P22077 could be from inhibition of USP7.

Fig 7.

Destabilization of Tip60 contributes to P22077-mediated cytoxicity. (A) U2OS cells were treated with DMSO or P22077 as indicated for 48 h, and the percentage of viable cells was determined by trypan blue exclusion and counting. The graph represents the averages of the results determined for viable cells in three different experiments. (B) HCT116 p53^+/+, HCT116 p53^−/−, and HCT116 p21^−/− cells were treated with DMSO or P22077 as indicated. Cell viabilities at 48 h are plotted as the averages of the results of three independent experiments. (C) Cells were treated with DMSO or P22077 as described in the panel B legend and immunoblotted for indicated proteins. (D) HCT116 p53^+/+, HCT116 p53^−/−, and HCT116 p21^−/− cells were treated after every 24 h with control GL2 or two different USP7-specific siRNAs for 5 days. Cell viabilities are plotted as the averages of the results of three independent experiments. (E) HCT116 p53^+/+, HCT116 p53^−/−, and HCT116 p21^−/− cells were treated with control GL2 or two different USP7-specific siRNAs for 48 h. The lysates were immunoblotted with anti-USP7 antibody to test the efficiency of USP7 knockdown with two different USP7-specific siRNA oligonucleotides as described in the panel D legend. (F) Control or Flag-Tip60-expressing U2OS cells were treated with DMSO or P22077 (20 μM) for 48 h. The cell viability (with DMSO-treated cells at 100%) was determined by the trypan blue exclusion method. (G) Immunoblots of lysates from a experiment similar to that described for panel F.

Since Tip60 is an essential acetyltransferase involved in a number of cellular functions such as chromatin remodelling, transcription, DNA repair, cell cycle progression, regulation of metabolic enzymes, etc., we wondered whether the toxicity of P22077 was at least partly due to the decrease of Tip60 levels. Indeed, stable overexpression of Flag-Tip60 increased the viability of U2OS cells treated with P22077 (Fig. 7F). Although P22077 decreased the level of overexpressed Flag-Tip60, a significant amount of residual Flag-Tip60 remained (Fig. 6J). The residual Flag-Tip60 maintained the acetylation of nucleosomal H4 (H4K5ac) to a significant extent in the presence of P22077 (Fig. 7G). Thus, the increased viability of Flag-Tip60-expressing cells is likely due to residual Flag-Tip60, suggesting that at least part of the toxicity following USP7 inhibition is due to Tip60 destabilization.

DISCUSSION

The results presented in this report support the idea of a new role of USP7 in regulating the p53-dependent apoptotic pathway. The Mdm2-p53 pathway is complex, and the regulation of this pathway by USP7 makes it more complex (38). Under basal conditions, USP7 stabilizes Mdm2 and Mdm2 in turn ubiquitinates and degrades p53 in the proteasome. However, under conditions of genotoxic stress, both USP7 and Mdm2 become positive regulators of p53 (31, 36). In response to DNA damage, USP7 deubiquitinase activity is necessary to increase p53 levels (31). In turn, p53 induces the transcriptional synthesis of Mdm2. Mdm2 transactivates p53 synthesis by interacting with p53 mRNA (35). Acetylation of p53 at K120 by Tip60 is necessary for p53 to transactivate PUMA and thus promote apoptosis in response to irreparable DNA damage (14). Here, we establish a critical role of USP7 deubiquitinase activity in stabilizing Tip60 acetyltransferase (Fig. 2A and D and 4B). Thus, the USP7 deubiquitinase activity is important not only for stabilizing Mdm2 (and thus increasing p53 levels in cells with DNA damage) but also for directing the p53 to proapoptotic promoters and driving apoptosis (Fig. 5A to E and 6E and G). This appears to explain why USP7 knockdown in undamaged or damaged cells fails to induce apoptosis, even though p53 is induced or present. Homozygous knockout of USP7 in mice is lethal, and the mice die during early embryonic development. Though the knockout embryos show p53 induction, there is no apparent apoptosis (39). Concurrent destabilization of Tip60 in the USP7^−/− mice may explain why there is no activation of the p53-dependent apoptotic pathway.

Tip60 interacts with the TRAF domain of USP7 (Fig. 3D). Mdm2 and p53 also interact with USP7 at its TRAF domain (40), though we do not know whether the exact same residues of TRAF domain are involved in the interactions with all three substrates of USP7. DNA damage abrogates the interaction between Mdm2 and USP7 (41). It will be interesting to investigate whether DNA damage abrogates the interaction of USP7 with Tip60 because of changes in USP7 that prevent interaction with any of its three substrates or whether the interaction is enhanced because the displacement of Mdm2 makes more of the TRAF domain of USP7 available to bind Tip60. If different residues of the TRAF domain are involved in interacting with p53 and Tip60, another interesting possibility would be that of formation of a tripartite complex of Tip60-USP7-p53. In that case, the USP7 would not only stabilize Tip60 and p53 but also promote the proapoptotic K120 acetylation of p53.

Recently, several small-molecule inhibitors have been developed to inhibit USP7 deubiquitinase activity with the primary aim to destabilize Mdm2 and stabilize p53 in the Mdm2-p53 pathway (33). One complication of the strategy is that in DNA-damaged cells, Mdm2 switches to become an activator of p53, so that P22077 actually inhibits p53 levels in DNA-damaged cells (Fig. 6E, G, and J). Our findings add the complication that Tip60 is destabilized when USP7 is inhibited, so that there is no acetylation of p53 at K120 (Fig. 6I) and p53 is unable to induce PUMA and the apoptotic pathway (Fig. 6E to H). These two complications suggest that P22077 is unlikely to become cytotoxic through the activation of p53. This conclusion is consistent with results from others (37) and us (Fig. 7B) showing that p53-null cancer cells are still rendered inviable by P22077.

P22077 is cytotoxic even though it attenuates the p53-dependent apoptotic pathway (Fig. 7A). Very recently, another USP7 inhibitor, P5091, was shown to cause cell death in multiple myeloma cells in a p53-independent but p21-dependent manner (37). However, we observed that P22077 induces death in cells irrespective of the p53 or p21 status in HCT116 cells (Fig. 7B). In all three cell types used (HCT 116 p53^+/+, HCT 116 p53^−/−, and HCT116 p21^−/−), the level of Tip60 decreased with P22077 treatment and this decrease contributed to the cytoxicity (Fig. 7C and F). We could not completely rescue the toxicity of P22077 with Flag-Tip60 overexpression. This could have been because the level of overexpressed Flag-Tip60 was still decreased by P22077 (Fig. 6J) or because the residual Flag-Tip60 was not as active as endogenous Tip60 (Fig. 7G). Alternatively, the deubiquitination activity of USP7 is also required for the stability of other proteins such as PTEN, Claspin, FOXO4, UHRF1, and NF-κB (42–46). It is therefore also possible that the residual cytotoxicity of P22077 in cells overexpressing Flag-Tip60 is due to a decrease of the levels of these other cellular targets of USP7. The definitive way to discover the critical targets of USP7 responsible for the toxicity of P22077 is to generate cells where USP7 substrates, such as Tip60, are stable in the absence of USP7, either because of mutations in the ubiquitin-acceptor sites, or because of inactivation of the relevant E3 ubiquitin ligases. P22077 is not toxic on cells where the critical target is thus stabilized.

Tip60 has several demonstrated roles in DNA repair (1). Tip60 acetylates ATM in response to DNA damage and thereby enhances its kinase activity to phosphorylate p53 and Chk2 (5, 47). Tip60 also acetylates histones, particularly H4, at the damage sites to relax the chromatin for the repair machinery (48). The destabilization of Tip60 by the USP7 inhibitor P22077 is likely to block the ATM’s kinase activity and deregulate the cell's response to DNA damage and contribute to cytotoxicity.

Deubiquitinase inhibitors, particularly USP7 inhibitors, are being developed for anticancer therapy (37). Our results suggest that it will be important to identify which programmed cell death pathways, other than apoptosis, such as necrosis or autophagy are initiated by inhibiting USP7 with P22077.

New therapeutic possibilities of USP7 inhibitors appear with our discovery that such inhibitors decrease Tip60 levels. For example, our results suggest that castration-resistant prostate cancer (CRPC) could be responsive to USP7 inhibitors because they decrease Tip60 levels. Many forms of CRPC are marked by hypersensitivity of the prostate cancer cells to androgen such that they are viable even in the presence of low levels of androgens that persist in patients on antiandrogen therapy. Both USP7 and Tip60 are overexpressed in CRPC (24, 45), and it is known that acetylation of the nuclear localization signal of the androgen receptor (AR) by Tip60 promotes the nuclear import of AR (24). Thus, decreasing Tip60 levels in CRPC with USP7 inhibitors is expected to prevent the nuclear localization of AR and blunt the hypersensitivity to androgen that is a distinguishing feature of many forms of CRPC.

Another therapeutic possibility emerges from the observation that Tip60 is required for homologous recombination (HR) because it promotes the recruitment of BRCA1 to damaged chromatin (49). Hereditary breast and ovarian cancers arising from mutations in BRCA1 are deficient in HR, and poly(ADP-ribose) polymerase (PARP) inhibitors (e.g., Olaparib) were developed to exploit the susceptibility of tumors with HR deficiency to PARP inhibition (50). Tip60 knockdown with siRNA, by inhibiting HR, makes U2OS osteosarcoma cells highly susceptible to Olaparib (49). Our findings now suggest that USP7 inhibitors, by decreasing Tip60 levels and thus inhibiting BRCA1 recruitment and HR, sensitize tumors without BRCA1 mutations to PARP inhibitors.