Haoran Guo a, 1, Siyu Shen a, 1, Yan Li a, Ran Bi a, Nannan Zhang a, Wenwen Zheng a, Yuyou Deng a, Ying Yang b, Xiao-Fang Yu a, c, Chunxi Wang a, Wei Wei a,
ABSTRACT
The human adenovirus oncoprotein E4orf6 hijacks intracellular Cullin 5-based E3 ubiquitin ligases (CRL5s) to induce the degradation of host proteins, including p53, that impede eficient viral replication. The complex also relies on another viral protein, E1B55K, to recruit substrates for ubiquitination. However, the determinants of adenoviral E4orf6-CRL5 E3 ligase-mediated p53 degradation in the scaf- folding protein Cullin5 remain rarely investigated. Here, we demonstrated that the viral protein E4orf6 triggered relocalization of the Cullin5 protein from the cytoplasm to the nucleus and induced activation of the CRL5 E3 ligase via facilitating neddylation. The expression of the deneddylase SENP8/Den1 was signiicantly downregulated by E4orf6. We then identiied SENP8 as a natural restriction factor for E4orf6-induced p53 degradation. Furthermore, our results indicated that the NEDD8-conjugating E2 enzyme UBE2M was essential for E4orf6-mediated p53 degradation and that its dominant negative mutant UBE2M C111S dramatically blocked E4orf6 functions. The Nedd8-activating enzyme inhibitor MLN4924 decreased E4orf6-induced neddylation of the cullin5 protein and subsequently suppressed p53 degradation. Collectively, our indings illuminate the strategy by which this viral oncoprotein speciically utilizes the neddylation pathway to activate host CRL E3 ligases to degrade host restriction factors. Disrupting this post-translational modiication is an attractive pharmacological intervention against human adenoviruses.
Keywords:Adenovirus;Viral oncoprotein;E4orf6;p53 degradation;Neddylation
1.Introduction
The tumour suppressor p53 has an important role in mediating responses to cellular stress and maintaining genomic integrity [1,2]. p53 is mutated in over half of human cancers, and p53 inactivation in mice renders them highly predisposed to cancer [3,4]. In addi- tion, accumulating evidence supports the view that p53 signalling is involved in host innate immunity against viral infection. Tran- scription of the p53 gene is stimulated by type I interferon [5]. Increased p53 protein expression enhances type I interferon- dependent antiviral activity[6]. Most DNA viruses frequently elicit DNA damage responses and cell cycle deregulation, resulting in p53 stabilization and activation [7e9]. To counteract this effect, DNA viruses exploit diverse strategies to perturb the functions of p53 [2,9e11]. Human adenovirus (AdV)proteins, namely, the 34 kDa product of open reading frame 6 of early region 4 (E4orf6) and the 55 kDa product of early region 1B (E1B55K), induce p53 polyubiquitination and proteolysis through the recruitment of the ElonginB/C-Cullin5/2-SOCS-box protein (ECS) E3 ligase [12e14]. Preventing p53 degradation by viral proteins is a host-targeted strategy for the development of antiviral therapies.Cullin-based E3 ubiquitin ligases (CRLs), the most prominent class of E3 ubiquitin ligases, act as scaffolds to bring a speciic substrate into close proximity to the E2 ubiquitin-conjugating enzyme, thereby facilitating ubiquitination and subsequent pro- teasomal degradation [15]. The ubiquitin-like protein NEDD8 has a dedicated E1-activating enzyme (NAE1) and E2-conjugating enzymes (UBE2M and UBE2F) and plays an important role in the enzymatic activity of CRL E3 ligase family members through direct conjugation to the Cullin scaffold. However, the effect of neddyla- tion on AdV protein E4orf6-recruited CRL5 E3 ligase activity re- mains uncertain.
In the present study, we demonstrate that E4orf6-mediated p53 degradation requires neddylation. E4orf6 can trigger the trans- location of Cullin5 proteins to assemble the E4orf6-CRL5 E3 ligase complex in the nucleus to recognize the nuclear substrate p53. Importantly, E4orf6 also enhanced Cullin5 neddylation, which is a classic marker for the activation of CRL5-based E3 ligases. By screening the effects of E4orf6 on the expression of neddylation- related proteins, we found that E4orf6 suppressed the expression of the host deneddylase sentrin-speciic protease 8 (SENP8) at the transcriptional level and demonstrated that SENP8 was a novel restriction factor of E4orf6-induced p53 degradation.Both a dominant negative mutant of the E2 enzyme, UBE2M C111S, and RNA interference-mediated UEB2M reduction Affinity biosensors inhibited E4orf6- induced p53 degradation. In addition, neddylation inhibition by a irst-in-class inhibitor MLN4924, which targets NAE1, disrupted p53 turnover by the E4orf6-CRL5 E3 ligase complex. Therefore, neddylation pathway overactivation by E4orf6 is an important target for the development of antiviral drugs and vaccines against human adenoviruses.
2.Materials and methods
2.1.Plasmids and reagents
The expression vector pHA-p53 was kindly provided by Dr. Y. Yang. The AdV Myc-E4orf6 expression vector was previously described [12]. The HA-Cullin5, Myc-SENP8, Flag-UBE2M C111S, and Flag-UBE2F C116S expression vectors were purchased from Generay Biotech Co. Ltd. (Shanghai, CN). Briefly, fragments con- taining the coding sequences flanked by 50 SalI and 3’ BglII/Xbal sites were inserted into the HA-VR1012 vector [16]. The shRNA plasmids targeting UBE2M or UBE2F (psh-UBE2M or psh-UBE2F) were gifts from Dr. J. Gross [17]. The following reagents were used: MLN4924 (MedChemExpress, HY-70062), anti-HA mono- clonal antibody (mAb) (Covance, MMS-101R), anti-Myc tag anti- body (Abcam, ab32), anti-FLAG M2 monoclonal antibody (Sigma, F1804), anti-Cullin5/CUL5 antibody (Abcam, ab184177), anti-alpha Tubulin antibody (Abcam, ab7291), and anti-Histone H3 antibody (Abcam, ab176842).
2.2. Cell lines
HEK293T cells (AIDS Research Reagents Program) were main- tained in Dulbecco’s modiied Eagle’s medium supplemented with 10% foetal bovine serum and penicillin/streptomycin. The cultured cell line was maintained at 37。C in a humidiied atmosphere con- taining 5% CO2. To generate UBE2M or UBE2F stable knockdown HEK293T cell lines, we co-transfected pshRNA-UBE2M or pshRNA- UBE2F, respectively, with pRSVRev (Addgene, 12253), pMDLg/pRRE (Addgene,12251), and pCMV-VSV-G (Addgene, 8454) into HEK293T cells. Viruses in the supernatants of infected cells were harvested 48 h post transfection. Viral particles were pelleted through a 20% sucrose cushion at 28,000 rpm for 90 min. Puriied virions were incubated with HEK293T cells for 48 h and were then selected with puromycin (3 mg/mL). Downregulation of UBE2M or UBE2F in the collected cells was veriied by RT-PCR.
2.3. Transfection, co-immunoprecipitation, and immunoblotting
DNA transfection was carried out using Lipofectamine 2000(Invitrogen) according to the manufacturer’s instructions. HEK293T
cells were harvested 48 h after transfection, washed twice with cold PBS, lysed in lysis buffer (150 mM Tris [pH 7.5] with 150 mM NaCl, 1% Triton X-100, and complete protease inhibitor cocktail tablets [Roche]) at 4 。C for 30 min, and centrifuged at 10,000xg for 30 min. For Myc tag immunoprecipitation, precleared cell lysates were mixed with anti-Myc antibody-conjugated agarose beads (Abmart, M20012) and incubated at 4 。C overnight. Samples were then washed eight times with wash buffer (20 mM Tris [pH 7.5] with 100 mM NaCl, 0.1 mM EDTA, and 0.05% Tween 20). Beads were eluted with elution buffer (0.1 M glycine-HCl, pH 2.0). The eluted materials were then analysed by SDS-PAGE and immunoblotting with the appropriate antibodies as previously described [18].
2.4.Immunostaining and confocal microscopy
HEK293T cells were transfected with HA-Cullin5, HA-p53 or Myc-E4ORF6 expression plasmids, passaged onto coverslips 24 h after transfection, and allowed to reattach to the coverslips for another 24 h. Selleckchem Defactinib Cells were then ixed for 15 min with 4% para- formaldehyde in PBS, permeabilized for 10 minin 0.1% Triton X-100 in PBS, and blocked using 5% BSA for 1 h. Then, cells were incubated with anti-HA.11 epitope tag antibody (BioLegend) and anti-Myc tag antibody (Cell Signalling Technology) at 4 。C for 16 h. Cells were then washed twice with PBS, followed by incubation with Alexa Fluor 594-conjugated goat anti-mouse IgG (H þ L) or Alexa Fluor 488-conjugated goat anti-rabbit IgG (H þ L) cross-adsorbed sec- ondary antibody (Invitrogen) at 4 。C for 1 h. Nuclei were counter- stained with 4’,6-diamidino-2-phenylindole (DAPI). Images were captured using a Zeiss laser scanning confocal microscope.
2.5.Quantitative real-time PCR (qRT-PCR)
Total RNA from cells was isolated using Trizol (Life Technologies) according to the manufacturer’s instructions, including the DNase I digestion step. Samples were incubated in 10 ml of diethyl pyro- carbonate (DEPC)-treated water with 1x RQ1 RNase-Free DNase buffer, l ml of RQ1 RNase-free DNase (Promega), and 4 U of RNase inhibitor (New England Biolabs) for 30 min at 37 。C. The DNase activity was inactivated by the addition of 1 ml RQ1 DNase stop solution and incubation at 65。C for 10 min. The RNA was reverse- transcribed by using random primers and the Multiscribe reverse transcriptase from the High-Capacity cDNA Archive Kit (Applied Biosystems) according to the manufacturer’s instructions. The cDNA was either used undiluted or serially diluted in DEPC-treated water before the real-time PCR reaction to ensure that the ampli- ication was within the linear range of detection. The StepOne Real- Time PCR system (Applied Biosystems, Carlsbad, CA) was used for the qRT-PCR ampliications.
2.6.Statistical analysis
Differences among test groups were analysed by ANOVA (Stata Corp, College Station, TX). A value of p < 0.05 was considered signiicant.
3.Results
3.1. AdVE4orf6 induces nuclear localization of Cullin5 and enhances Cullin5 neddylation
To date, the mechanism by which E4orf6 recruits cytoplasmic CRL5 E3 ligaseto induce the degradation of nuclear proteins such as p53 and Mre11 remains poorly deined. By using an immunofluo- rescence assay, we conirmed that in transfected HEK293T cells,HA-tagged Cullin5 was localized in the cytoplasm, while HA-tagged p53 was mainly localized in the nucleus (Fig. 1A and B). However, ectopic expression of AdVE4orf6 clearly changed the localization of Cullin5 proteins, which translocated from the cytoplasm to the nucleus (Fig. 1A). However, E4orf6 did not signiicantly influence p53 localization (Fig. 1B).Thus, these indings suggest that the viral oncoprotein E4orf6 induces nuclear import of the scaffolding pro- tein Cullin5, thus subsequently facilitating the upload of the nu- clear substrate p53 onto the E4orf6-CRL5 E3 ligase complex.A decisive factor of CRL E3 ligase function is that Cullin proteins must be covalently modiied with the ubiquitin-like modiication neddylation to retain the CRL complex in an active state [19,20]. The neddylated form of Cullin proteins migrates more slowly than the unmodiied form during SDS/PAGE, which provides an optimal method to measure protein neddylation. However, our immuno- blotting data showed that neddylation of both exogenous and endogenous Cullin5 proteins occurred at undetectably low levels in HEK293T cells (Fig. 1C, lane 1; 1D, lane 1). We also used the same assay to investigate theneddylation levels of the other Cullin family proteins Cullin 3, Cullin4A and Cullin4B, which were much higher than that of Cullin5 (data not shown). Considering the importance of neddylation in the activation of CRL5 E3 ligase functions [12,21], we investigated Cullin5 neddylation in the presence or absence of E4orf6 (Fig. 1C and D). HEK293T cells were co-transfected with pMyc-E4orf6 or HA-Cullin5 for two days. The treated cells were harvested for immunoblotting with the indicated antibodies. Strikingly, ectopic expression of E4orf6 elicited the expression of a slower-migrating exogenous or endogenous Cullin5 species readily detectable in cell lysates prepared from the transfected cells (Fig. 1C, lanes 2 and 3; 1D, lane 2). Quantiication of the immuno- blotting data by ImageJ software further supported the conclusion that these viral oncoproteins directly increase the levels of ned- dylated Cullin5 to enable CRL5 E3 ligase activation (Fig. 1C and D).
3.2. The deneddylase SENP8 can inhibit E4orf6-mediated p53 degradation and is downregulated by E4orf6
The identiication of E4orf6-mediated Cullin5 neddylation led us to investigate the effects of E4orf6 on the intracellular neddy- lation pathway. We screened the mRNA levels of various genes involved in neddylation, including NAE1, UBE2M, UBE2F, RBX1, RBX2 and SENP8, in the presence or absence of E4orf6 expression.
Fig. 1. AdV E4orf6 induces the nuclear localization of Cullin5 and enhances Cullin5 neddylation. (A) Cellular localization of HA-Cullin5 and Myc-E4orf6 using confocal immuno- fluorescence microscopy imaging. (B) Cellular localization of HA-p53 and Myc-E4orf6 using confocal immunofluorescence microscopy imaging. (C) Western blotting for HA-Cullin5, Myc-E4orf6 and β-actin in E4orf6-expressing HEK293T cells. (D) Western blotting for endogenous Cullin5, Myc-E4orf6 and β-actin.We found that E4orf6 expression signiicantly decreased the tran- scription of the cellular deneddylase SENP8 (Fig. 2A). Our immu- noblotting data also suggested that E4orf6 downregulated the expression levels of SENP8 in a dose-dependent manner (Fig. 2B), supporting the hypothesis that the viral oncoprotein E4orf6 selec- tively decreases the expression of the host deneddylation protein SENP8.Subsequently, we measured the effects of SENP8 on E4orf6- induced p53 degradation. pHA-p53 and pMyc-E4orf6 were co- transfected with increasing concentrations of vectors expressing Myc-tagged SENP8 into HEK293T cells. Cells were harvested 48 h post transfection for immunoblotting. Interestingly, overexpression of SENP8 impaired p53 degradation by E4orf6 in a dose-dependent manner (Fig. 2C). Our indings revealed that SENP8 is a novel host restriction factor for E4orf6-CRL5 E3 ligase-mediated p53 degra- dation but is downregulated by E4orf6.
3.3. The neddylation E2 enzyme UBE2M is required for E4orf6- mediated p53 degradation
To determine the effects of the E2 conjugating enzymes UBE2M and UBE2F on AdV oncogene-mediated p53 degradation, we used shRNA targeting ube2m or ube2f to generate UBE2M or UBE2F stable knockdown HEK293T cells.HA-p53 and Myc-E4orf6 expression vectors were co-transfected into sh-Control, sh- UBE2M or sh-UBE2F HEK293T cells. Forty-eight hours later, cells were harvested for immunoblotting, and the results indicated that E4orf6-induced p53 degradation was blocked by downregulation of only UBE2M rather than UBE2F (Fig. 3A).
Next, we investigated the effects of the UBE2M dominant negative mutant UBE2M C111S (UBE2M DN) and the UBE2F dominant negative mutant UBE2F C116S (UBE2F DN) on E4orf6-CRL5 E3 ligase functions [22,23]. Both pHA-p53 and pMyc-E4orf6 were co-transfected into HEK293T cells with or without pFlag- UBE2M DN or pFlag-UBE2F DN. After 48 h, cells were harvested for immunoblotting with the indicated antibodies. The data revealed that overexpression of UBE2M DN but not UBE2F DN potently inhibited p53 degradation by E4orf6 (Fig. 3B and C). Hence, theneddylation E2 enzyme UBE2M is important for E4orf6- CRL5 E3 ligase-mediated p53 degradation.
3.4. The NAE1 inhibitor MLN4924 suppresses E4orf6-dependent Cullin5 neddylation and p53 degradation
To subvert viral oncoprotein-dependent neddylation activation, we used the irst-in-class antitumour drug MLN4924 as a tool to block cellular neddylation [24]. MLN4924 speciically inhibits NAE1 with potent antineoplastic activity (Fig. 4A). HEK293T cells were transfected with pMyc-E4orf6 and incubated with 300 nM MLN4924 at 24 h post transfection. After another 24 h, cells were harvested for immunoblotting. MLN4924 strongly impaired the neddylation of Cullin5 in the presence of E4orf6 (Fig. 4A, right panel).Then, we tested the effects of MLN4924 on AdV E4orf6-mediated p53 degradation. HEK293T cells were transfected with p53-HA and/ or pMyc-Adv5 E4orf6, and 24 h later, MLN4924 (300 nM) was added to the cell culture media. After another 24 h, cells were harvested for immunoblotting. The immunoblotting data indicated that MLN4924 dramatically impaired the ability of E4orf6 to induce p53 degrada- tion, with a potency equivalent to that of the proteasome inhibitor MG132 (Fig. 4B). Treatment with MLN4924 did not affect the viability of HEK293T cells (Fig. 4C), indicating that this agent did not cause signiicant cytotoxicity under our experimental conditions. Furthermore, we performed a co-immunoprecipitation assay
Fig. 2.SENP8 inhibits E4orf6-mediated p53 degradation and is downregulated by E4orf6. (A) Relative mRNA levels of NAE1, UBE2M, UBE2F, RBX1, RBX2, SENP8 in Myc-E4ORF6- or empty vector (EV)- transfected HEK293Tcells. (B) Western blotting for endogenous SENP8, Myc-E4orf6 and a-tubulinin E4orf6-expressing HEK293T cells (top panel). The bar graph shows the percentages of the relative band intensity representing SENP8 expression (bottom panel). (C) Western blotting for HA-p53, Myc-E4orf6, Myc-SENP8, and a-tubulin in E4orf6-expressing HEK293T cells in the presence of increased SENP8 expression (top panel). The bar graph shows the percentages of the relative band intensity representing p53 expression (bottom panel).
Fig. 3. UBE2M is required for E4orf6-mediated p53 degradation. (A) Western blotting for HA-p53, Myc-E4orf6 and β-actin in HEK293T cells stably expressing sh-Control, sh-UBE2M or sh-UBE2F. (B) Western blotting for HA-p53, Myc-E4orf6, Flag-UBE2M DN and β-actin in p53, E4orf6 and UBE2M DN-expressing HEK293T cells. (C) Western blotting for HA-p53, Myc-E4orf6, Flag-UBE2F DN and β-actin in HEK293T cells expressing p53, E4orf6 and UBE2F DN.
Fig. 4. MLN4924 suppresses E4orf6-dependent Cullin5 neddylation and p53 degradation. (A) Schematic indicating MLN4924 inhibition of NAE1 at the initial neddylation step (left). Western blotting for Cullin5, Myc-E4orf6, and a-tubulinin E4orf6-expressing HEK293T cells in the presence or absence of 300 nM MLN4924 (right). (B) Western blotting for HA-p53, Myc-E4orf6 and β-actin in HEK293T cells expressing p53 and E4orf6 in the presence or absence of MG132 or MLN4924. (C) MLN4924 treatment was minimally cytotoxic to HEK293T cells. (D) Western blot results showing the interaction among E4orf6, p53 and Cullin5 in the presence or absence of MLN4924 investigate the effects of MLN4924 on the assembly of the E4orf6- CRL5/p53 protein complex. MLN4924 did not interrupt E4orf6 recruitment to Cullin5 and p53, but strongly blocked E4orf6- mediated Cullin5 neddylation (Fig. 4D). Collectively, these data demonstrate that the enhancement of neddylation by the viral oncoprotein E4orf6 is a determining factor of E4orf6-CRL5 E3 ligase activation to induce p53 degradation.
4.Discussion
p53 signalling-dependent apoptosis or the Non-specific immunity innate immune response is commonly activated upon DNA virus replication, thus compromising an important host defence against viral infection [7]. The AdV E4orf6-recruited CRL5 E3 ligase triggers proteasome- dependent degradation of p53 and contributes to maintaining an environment supporting viral infection; however, this process is accompanied by an increased risk of cancer development. In the present study, we revealed the strategy by which the viral oncoprotein E4orf6 manipulates neddylation signalling, which contributes to p53 degradation by the E4orf6-CRL5 E3 ligase complex.We found that the neddylation of both exogenous and endog- enous Cullin5 proteins is maintained at a much lower level than that of other Cullin proteins, implying that Cullin5-based E3 ligase activity is strictly controlled during cell proliferation. The pAdV E4orf6 that we used in this study was derived from human adenovirus type 5 (species C), which has been demonstrated to induce p53 degradation dependent on the Cul5-based E3 ligase [12]. Importantly, we found that E4orf6 can activate the CRL5 E3 ligase by enhancing Cullin5 neddylation. By screening the effects of E4orf6 on the expression of neddylation-related proteins, we found that E4orf6 suppressed the expression of the host deneddylase SENP8 at the transcriptional level (Fig. 2A and B). The identiication of SENP8 as a novel restriction factor of E4orf6-induced p53 degradation reveals that the viral protein E4orf6 negatively regu- lates intracellular deneddylation processes to create the optimal conditions for CRL5 complex activation.
Subcellular colocalization of substrates with the CRLs E3 ligase is a determining factor of the ubiquitination process. We found that E4orf6 can trigger the nuclear accumulation of cytoplasmic Cullin5, which could enable the E4orf6-Cullin5 E3 ligase complex to assemble in the nucleus, with the subsequent upload of nuclear p53 onto this protein complex for ubiquitination. AdVE4orf6 acts as the substrate adaptor and transport protein for the host CRL5 E3 ligase to induce the degradation of p53 and other nuclear proteins.Interestingly, E4orf6-mediated Cullin5 translocation also alters the signalling pathway for Cullin5 neddylation. Only two known E2-conjugating enzymes, UBE2M andUBE2F, are involved in Nedd8 conjugation. UBE2M is responsible for Cul1-4 neddylation, whereas UBE2F contributes to Cul5 neddylation [22]. However, in our study, we found that UBE2M but notUBE2F is required forE4orf6-CRL5 E3 ligase functions (Fig. 3), which challenged our understanding of Cullin protein neddylation. Recent studies have demonstrated the flexible interaction between Cullin5 and the neddylation E3 en- zymes RBX1 and RBX2 [20]. One possible explanation for the essential role of UBE2M on E4orf6-CRL5 E3 ligase functions is that the viral oncoprotein E4orf6 utilizes the flexible features of the interaction of Cullin5 with RBX1/RBX2. The UBE2M protein is mainly localized in the nucleus, while UBE2F is localized in the cytoplasm (www.proteinatlas.org). Thus, the transport of Cullin5 proteins into the nucleus byE4orf6 would replace the normal CRL5- RBX2-UBE2F module in the cytoplasm with the nuclear E4orf6- CRL5-RBX1-UBE2M module. This study has further expanded our understanding of Cullin5-based E3 ligases, and additional in- vestigations on the normal cellular functions of the Cullin5-RBX1- UBE2M module are necessary.
Inhibition of neddylation by the irst-in-class antitumour in- hibitor MLN4924 also potently abolished E4orf6-induced p53 degradation but did not influence the assembly of the E4orf6-CRL5/ p53 complex. Our indings illuminate the mechanisms by which the AdV oncoprotein E4orf6 usurps and activates the host CRL5 E3 ligase to trigger p53 degradation via manipulating neddylation signalling and are anticipated to help provide promising thera- peutic targets against human adenoviruses.