Ubiquitination at the interface of tumor viruses and DNA damage responses

Abstract

Viruses exploit cellular ubiquitination machinery to shape the host proteome and promote productive infection. Among the cellular processes influenced by viral manipulation of ubiquitination is the DNA damage response, a network of cellular signaling pathways that sense and respond to genomic damage. This host-pathogen interaction is particularly important during virus replication and transformation by DNA tumor viruses. Manipulating DDR pathways can promote virus replication but also impacts host genomic instability, potentially leading to cellular transformation and tumor formation. We review ways in which viruses are known to hijack the cellular ubiquitin system to reshape host DDR pathways.

Introduction

Viruses have evolved ways to manipulate the host proteome in order to evade antagonistic antiviral host responses and to generate cellular environments conducive to infection. One strategy employed by viruses is to harness cellular ubiquitin machinery to direct ubiquitin modification of cellular proteins [1–3]. Ubiquitination is a protein post-translational modification that either directs substrates toward proteasomal degradation or mediates non-degradative outcomes such as controlling protein interactions, location or function [4–6]. Dynamic regulation of ubiquitination by ubiquitin ligase and deubiquitinase (DUB) enzymes allows ubiquitin to act as a molecular switch for rapid cellular responses. Viruses redirect ubiquitination of target substrates through four main approaches: (1) viral-encoded E3 ligases, (2) integration into cellular cullin E3 ligase complexes, (3) manipulation of cellular E3 ligases or (4) control of deubiquitinase activity (Figure 1).

Figure 1: Viral strategies to hijack the cellular ubiquitin system.

Viruses encode proteins with E3 ubiquitin activity that directly ubiquitinate cellular substrates (I). Viral proteins integrate into cellular cullin-RING E3 ligases, assuming the substrate adaptor function to dictate cellular substrates that are ubiquitinated (II). Viral proteins interact with cellular E3 ligases to direct ubiquitination of cellular substrates or alter cellular E3 ligase activity (III). Viruses can impart the opposite function of E3 ligases by directing deubiquitination of cellular substrates either by encoding viral proteins with DUB activity or by interacting with cellular DUBs (IV). Studying the repertoire of mechanisms that viruses use to exploit the host ubiquitin system will broaden our understanding of host-pathogen interactions and functions of ubiquitin in cellular processes such as the DDR.

Entry and amplification of exogenous viral DNA activates components of the host DNA damage response (DDR) signaling network, a cellular system integral to maintaining genomic stability by sensing and responding to compromised DNA [7,8]. In response to activation of DDR pathways, viruses manipulate these processes to promote productive infection [2,9–11]. Viral manipulation of cellular DDR pathways can also affect cellular transformation and oncogenesis [11]. The human “tumor viruses” include Human Papilloma Virus (HPV), Merkel Cell Polyomavirus (MCV), Kaposi’s Sarcoma-associated Herpes Virus (KSHV), Epstein-Barr Virus (EBV), Human T-Lymphotropic Virus 1 (HTLV-1) and Hepatitis B and C Viruses (HBV/HCV). Additionally, Adenovirus (AdV) and Simian Virus 40 (SV40) Polyomavirus can transform cells in vitro [12,13] although they are not known to cause human cancers. Human tumor virus infections are associated with over 20 different types of cancers and approximately 10% of diagnosed cancer cases worldwide [14].

Ubiquitination plays a critical role at the interface between tumor viruses and host DDR pathways. Ubiquitin is an integral component of DNA damage recognition and repair, exhibiting important functions in protein recruitment and removal at sites of DNA damage, as well as regulation of downstream effectors [15–18]. Viruses direct ubiquitination to affect cellular processes and facilitate viral DNA replication and progeny production [1–3]. This review highlights specific examples of ways that tumor viruses hijack host ubiquitin machinery to promote viral replication and manipulate host DDR responses. Finally, viral mediated ubiquitination that does not result in substrate degradation by the proteasome (non-degradative ubiquitination) is proposed as a potential strategy to alter the DDR.

Ubiquitin as a post-translation modification

Ubiquitination is a protein post-translational process whereby a ubiquitin molecule is covalently attached to a lysine residue of a protein substrate through a multi-step enzymatic reaction. Ubiquitin is first activated by an E1 ubiquitin activating enzyme. Next, the activated ubiquitin molecule is transferred to an E2 ubiquitin conjugating enzyme. Finally, ubiquitin is covalently attached to the protein substrate in a reaction catalyzed by an E3 ubiquitin ligase enzyme. The process of ubiquitination is highly dynamic and can be reversed by DUBs that catalyze removal or editing of ubiquitin on substrates. Ubiquitin molecules have seven lysines that can be ubiquitinated, which confers the ability to form polyubiquitin chains on protein substrates. The chain type, as defined by the amino acid position of the modified ubiquitin lysine, is thought to dictate the functional outcome of ubiquitination.

The most well-understood outcome of ubiquitination is proteasome-mediated degradation of substrates where polyubiquitin chains linked by lysine at position 48 of the ubiquitin molecule (K48-chains) direct substrates to the 26S proteasome [5]. The importance of non-degradative ubiquitination is becoming increasingly appreciated. Monoubiquitination, multi-monoubiquitination and non-degradative polyubiquitin chains (K6, K11, K27, K29, K33, K63) can regulate protein-protein interactions, alter protein functions or direct protein localization [4–6].

The DNA damage response preserves genome integrity

The cellular genome is under constant threat of damage by endogenous events and exogenous sources. The DDR signaling network is responsible for recognition and repair of DNA damage, driving the cell toward cell-cycle arrest, senescence, or DNA repair [7,8]. DDR proteins are broadly categorized into “sensors” of DNA damage, “transducers” of DDR signaling and downstream “effectors” of DNA repair and cell cycle control (Figure 2). The DDR network is mediated by three PIKKs (phosphatidylinositol 3-kinase-like protein kinase); ATM (ataxia telangiectasia mutated), ATR (ataxia telangiectasia and Rad3 related) and DNA-PK (DNA-dependent protein kinase) [19]. The MRN complex (MRE11, RAD50 and NBS1 proteins) senses double-strand breaks (DSBs) and activates ATM [20] through TIP60 [21]. ATM phosphorylates substrates involved in DNA repair (e.g. BRCA1 and RAD51), as well as cell cycle control and apoptosis (e.g. CHK2 and p53). DSBs are also sensed by the Ku70/Ku80 heterodimer, which activates DNA-PK, and leads to end-joining by DNA ligase IV [22]. The RPA protein coats single-strand DNA and recruits ATR through ATRIP (ATR-interacting protein) [23] and the 9–1-1 complex (RAD9, HUS1, RAD1), which binds the ATR-activating protein TOPBP1 [24]. ATR substrates control cell cycle and apoptosis (e.g. CHK1 and p53). Overall, there are hundreds of proteins associated with pathways mediating the DDR [25,26], and this extensive network of proteins and interactions provide an array of opportunities for viral manipulation.

Figure 2: Viruses target DDR pathways using multiple ubiquitination strategies.

Viruses use ubiquitination to target the PIKK and PCNA DDR pathways at the levels of the “sensors”, “mediators” and “effectors” of DDR signaling. Viruses also target downstream responses of the DDR, especially by degradation of tumor suppressors such as p53, which regulates cell cycle checkpoints and apoptosis. Each of the mechanisms of viral-mediated ubiquitination (Figure 1) can be identified in viral manipulation of DDR signaling.

Post-translational modifications coordinate complex cellular responses to DNA damage. Ubiquitin plays an integral role in orchestrating and regulating both assembly and disassembly of repair proteins at DNA damage sites [15–18]. Ubiquitination of many cellular proteins increases upon DNA damage [26]. K63-linked ubiquitination plays important roles in signaling for recruitment of DDR-associated proteins to DSBs, while protein turnover through K48-linked modification regulates levels of DDR proteins. Overall, there is a complex landscape of ubiquitination at DNA lesions that collectively serve to remodel chromatin and direct repair. In addition to roles at damage sites, ubiquitin functions to modulate downstream effectors of the DDR network that lead to cell cycle arrest and apoptosis.

Impacts of DDR signaling on viral infection

DDR signaling pathways can be induced as host cells react to viral genomes and/or by viral manipulation of DDR components. Activation of cellular DDR signaling has outcomes that are pro-viral or anti-viral. Signaling through tumor suppressors such as p53 leads to activation of cell cycle checkpoints that must be overcome by viruses. Processing of viral genomes by host DNA repair factors can limit virus replication. In contrast, many viruses utilize subsets of DNA repair factors during viral genome replication. Many previous reviews provide detailed descriptions of interactions between viruses and the DDR, especially viral oncoproteins that control cell cycle checkpoints [9–11,27–29].

Viruses harness ubiquitination to manipulate the host proteome

There are multiple ways in which viruses hijack host ubiquitin machinery (Figure 1). The mechanisms employed by viruses to harness cellular ubiquitination include: encoding viral proteins with E3 ligase activity (Figure 1, I), integrating into cellular protein complexes to form host-viral cullin E3 ligases (Figure 1, II) or redirecting cellular E3 ligase activity (Figure 1, III). Viruses also exert control by removing ubiquitin through viral or cellular DUBs (Figure 1, IV). All these strategies are employed by viruses to modify key regulators of the cellular DDR machinery (Figure 2).

Viral proteins with E3 ligase activity alter DDR pathways

Components of the DDR network can be directly targeted by viral-encoded E3 ligases (Figure 1, I). One strategy is to manipulate recruitment of DNA repair proteins by targeting key steps in ubiquitination. RNF8-mediated histone ubiquitination is one of the first steps in DDR responses to DSBs [30–32]. ATM-dependent phosphorylation of MDC1 prompts E3 ligase RNF8 recruitment to breaks. RNF8 mediates K63-linked ubiquitination of H1-linker histones, which promotes recruitment of a second ligase RNF168, which ubiquitinates histone H2A and other cellular factors to generate assemblies of repair proteins. The HSV ICP0 protein is a RING-type E3 ligase that targets RNF8 and RNF168 for ubiquitination and degradation, to prevent localization of DDR factors at sites of incoming viral genomes [33,34]. ICP0 also ubiquitinates and degrades DNA-PK [35,36], which may influence circularization of viral genomes [37]. ICP0 targeting RNF8 and RNF168 is an intriguing example of viral-mediated ubiquitination used to directly target and suppress cellular ubiquitination. The DDR-associated E3 ligase RNF8 is also harnessed by the HTLV-1 protein Tax to redirect ubiquitination and suppress DDR responses and activate IKK/NF-κB pathways [38].

Viral proteins integrate into cellular cullin complexes to target DDR pathways

Since E3 ubiquitin ligases confer substrate specificity, viruses predominantly interface with this step of the ubiquitin conjugation process. Cullin proteins act as scaffolds that interact with E2 binding proteins and substrate adaptors in cullin-RING E3 ligases (CRLs), which constitute the largest known family of E3 ligases [39]. Viruses commonly integrate into CRL complexes to act as substrate adaptors/receptors to redirect ubiquitination specificity (Figure 1, II).

The Ad5 E1B55K and E4orf6 proteins complex with cellular Cullin5 to form a CRL complex [40] that redirects ubiquitination to multiple DDR substrates. The Ad5 serotype E1B55K/E4orf6 CRL induces degradation of the MRN complex [41] and TIP60 [42] and prevents ATM activation [43]. The E1B55K/E4orf6 complex also promotes degradation of DNA repair proteins such as the BLM helicase [44] and DNA ligase IV [45]. Additionally, the Ad12 serotype degrades TOPBP1 to prevent ATR activation [46]. In the absence of viral proteins that target these repair proteins, AdV infection results in activation of DDR signaling, MRN inhibits viral replication, and viral genomes are joined into concatemers [41,47–50]. The MRN complex is also targeted by the SV40 large T antigen which binds Cullin7 to prevent ATM activation [51].

Downstream of DDR signaling, tumor suppressor p53 is phosphorylated to regulate cell cycle checkpoints and apoptosis in response to DNA damage. Phosphorylation results in p53 stabilization by preventing association with the MDM2 E3 ubiquitin ligase, thereby suppressing ubiquitin-mediated degradation. Many viruses counter p53 and other tumor suppressors by inducing degradation, which ultimately contributes to genomic instability, cellular transformation and oncogenesis [52]. For example, the E1B55K/E4orf6 CRL complex ubiquitinates and degrades p53 [40]. The KSHV LANA protein also harnesses the same cellular Cullin5-complex, via a SOCS-box-like motif on the viral protein, and degrades p53 as well as the tumor suppressor von Hippel-Lindau [53]. The HPV E7 protein complexes with a Cullin2 CRL ligase to degrade Rb [54,55]. EBV protein BZLF1 functions as a substrate receptor in a Cullin2/5 complex to degrade p53 [56,57], and the EBNA3C viral protein forms a complex with SCF^Skp2 to degrade Rb [58].

A recently published example of manipulating genome maintenance pathways comes from HBV where the HBx protein complexes with Cullin4-DDB1-ROC1 to form a CRL E3 ligase [59] that degrades DDR-associated proteins SMC5/6 [60,61], which promotes disassociation from viral episomal DNA and facilitates viral gene expression. It was shown that SMC5/6 knockdown increases HBV replication during infection with HBx deficient virus but had negligible effects during wildtype infection, indicating that SMC5/6 is targeted because it acts as a restriction factor against HBV.

Viral proteins regulate cellular E3 ligases that maintain genome integrity

Viral proteins can indirectly mediate ubiquitination by interacting with cellular E3 ligases (Figure 1, III). A quintessential example is the HPV E6 oncoprotein, which interacts with cellular E3 ligase E6AP (E6 associated protein) to redirect ubiquitination and degradation towards cellular substrates such as p53 [62–64]. HPV E6 also facilitates ubiquitin-mediated degradation of the DDR protein TIP60 through cellular E3 ligase EDD1/UBR5 [65] although the exact mechanism is unclear. The HPV E7 oncogene promotes proteasome-dependent degradation of tumor suppressor PTPN14 through the cellular E3 ligase UBR4 [66]. The Rb protein is degraded during HCV infection through an interaction of HCV NS5B with E6AP [67]. An alternative viral strategy is to suppress ubiquitination by cellular E3 ligases. For example, EBV LMP1 protein both disrupts interaction of MDM2 with p53 to suppress K48-polyubiquitin-mediated degradation, and additionally induces K63-polyubiquitin by TRAF2 to stabilize p53 [68]. Similarly, MCV viral sT protein inhibits Fbw7 of the SCF^Fbw7 E3 ligase by binding via the LT-stabilization domain, thereby stabilizing SCF^Fbw7 substrates such as LT, c-myc, cyclin E, PLK1 and YAP [69,70]. The combination of these strategies, which target tumor suppressors and genome guardians, facilitate viral propagation while also contributing to host genome instability during tumorigenesis.

Viral strategies to remove ubiquitin within the DDR

Deubiquitination can be employed by viruses to modify ubiquitin signaling pathways by either blocking or redirecting cellular DUBs, or through virally-encoded DUBs (Figure 1, IV). In response to DNA damage, PCNA (proliferating cell nuclear antigen) is monoubiquitinated by RAD18 to activate the translesion synthesis pathway of post-replication repair. PCNA accumulates at viral replication compartments of many DNA viruses, although its function in viral replication is unclear. The EBV late protein BPLF1 is a DUB that deubiquitinates PCNA, preventing recruitment of cellular proteins and disrupting DNA repair [71]. BPLF1 can also promote viral replication by acting as a deneddylase to modify the activity of CRLs [72], which induces accumulation of the cellular replication licensing factor CDT1 and deregulates S-phase DNA synthesis. The BPLF1 protein may therefore have dual functions during infection which inhibit cellular DNA replication/repair and enhance viral DNA replication.

The herpesvirus-associated ubiquitin-specific protease (HAUSP), also known as USP7, was originally discovered through its interaction with the ICP0 E3 ligase of HSV-1 [73] and was subsequently implicated in several cellular processes including oncogenesis, tumor suppression, and epigenetic modifications [74,75]. The first substrate identified for USP7/HAUSP deubiquitination was tumor suppressor p53 whose half-life is extended by the DUB activity [76]. In addition to p53, its own regulator MDM2 is also a substrate for USP7/HAUSP which protects it from auto-ubiquitination [77]. USP7/HAUSP also appears to have direct roles in cellular DNA replication and DNA damage. Among its substrates are CHK1, CLASPIN and TIP60. USP7/HAUSP can regulate stability of cellular E3 ligases including RNF168 [78] and RNF169 [79] which have roles in ubiquitin-dependent DDR signaling. Several viral proteins interact with USP7/HAUSP, including EBNA1 of EBV [80], LANA of KSHV [81] and E1B55K of AdV [82]. In the case AdV, USP7 is required for efficient viral growth and cellular transformation by E1A/E1B55K oncoproteins [82]. USP7 also affects stability of E1B55K but there is no direct evidence that it is a substrate for deubiquitination. In addition to ICP0, other herpesvirus proteins bind USP7/HAUSP. The EBNA1 protein of EBV competes with p53 and MDM2 for the same binding pocket in USP7/HAUSP and thus lead to p53 instability [83,84]. In this way EBNA1 may modulate p53 in EBV-infected cells to promote survival by decreasing DNA damage-induced apoptosis [84]. There are likely to be other DUBs that are harnessed in similar ways to USP7 to regulate cellular processes such as DDR signaling.

Can viruses direct non-degradative ubiquitination?

The best-characterized function of ubiquitin is to direct substrate degradation, and this is also the most well-understood function of ubiquitination during host-pathogen interactions. However, ubiquitination that has non-degradative outcomes is clearly widespread and is likely to be similarly targeted and harnessed by viruses. There are currently very limited known examples of viruses directing non-degradative ubiquitin, although a few viral ligases have been shown to harness K63 ubiquitin linkages to counter host responses. These include EBV protein LMP1, which was shown to associate with K63-polyubiquitination of RIPK1/3 to suppress necroptosis, although did not directly ubiquitinate the substrate [85]. KSHV encodes K3 and K5, which demonstrate E3 ligase activity and are able to employ K63 poly-ubiquitin linkages to promote internalization and endosomal degradation of surface MHC molecules [86].

Harnessing non-degradative ubiquitination would vastly increase the potential for viruses to modulate the cellular proteome and could avoid collateral effects of inducing degradation of cellular regulators. Viral-mediated non-degradative ubiquitination could be employed to recruit cellular proteins onto viral genomes, to sequester antiviral functions, to modulate cellular signaling pathways or to control protein complex formation. There is an emerging body of literature suggesting that CRLs, which are generally known to direct degradative ubiquitination, can in fact promote K63-linked non-degradative ubiquitination [87,88]. Ubiquitin linkages other than K48 could be conjugated by viral proteins that integrate into CRLs. It would also be expected that viral-encoded E3 ligases have evolved the ability to interact with appropriate cellular E2 enzymes to form non-degradative poly-ubiquitin linkages.

Non-degradative ubiquitination has been shown to be particularly relevant within the DDR [15–18] and could therefore be employed during viral manipulation. It was recently shown that FBXW7, which forms the CRL SCF^FBXW7, is a substrate of ATM and is recruited to sites of DNA damage. SCF^FBXW7 interacts with XRCC4 and promotes K63-linked non-degradative polyubiquitination, which ultimately facilitates activation of NHEJ [87]. This example illustrates CRL-mediated non-degradative ubiquitination within the DDR and could be broadly exploited by viruses that recruit CRLs (Figure 1, II).

Despite the preponderance of examples of viral-mediated degradative ubiquitin, there is a relative dearth of known non-degradative events. This likely results from limited availability of experimental approaches but is poised to change. Advanced proteomics technologies now facilitate identification of non-degradative ubiquitination events. Whole cell proteomics and diglycine remnant profiling [89,90] allow global investigation of modifications by ubiquitin and ubiquitin-like proteins and resulting impacts on protein abundance. Bioinformatic analysis and integration of relevant datasets can generate a comprehensive view of the ubiquitin landscape and effects of ubiquitination on the proteome. Thus, classification of both degradative and non-degradative ubiquitination will likely drive future studies of ubiquitin in host-pathogen interactions, especially within DDR signaling pathways.

Conclusions

Cellular E3 ligases and DUBs are regulated dynamically in supramolecular complexes. These molecular machines are often manipulated by viruses to tip the balance in favor of viral processes and disruption of antiviral defenses. In some cases, modulation may be subtle enough to retain cellular functions, while in other cases, host processes may be completely disrupted or redirected. In this review we focused specifically on ubiquitin modifications but ubiquitin-like proteins are important in both DDR signaling and virus-host interactions. Another area that has not been extensively investigated is ubiquitin binding domains in viral proteins, although it is likely that viral proteins will exploit the power of these “readers” to recognize specific ubiquitin chain linkages.

Viruses have evolved a variety of strategies to harness host ubiquitin machinery to redirect ubiquitination of cellular substrates and shape the host proteome (Figure 1). Manipulating the DDR pathways, either by stimulating or blunting DDR responses, can ultimately promote productive viral infection. Furthermore, viral modulation of DDR responses likely contributes to cellular transformation and oncogenesis. Ubiquitin-mediated processes represent targets for antiviral therapeutics, with both E3 ligases and DUBs providing compelling targets for small molecule inhibitors. Understanding how viruses use ubiquitin to target the DDR could offer insights into therapeutic means to control virus infection and combat viral-mediated oncogenesis.