Role of SUMOylation in Human Oncogenic Herpesvirus Infection

Highlights

• Regulation of SUMO signaling in the EBV/KSHV latent infection.

• Role of SUMOylation in reactivation of EBV/KSHV lytic replication.

• Cross-talk between SUMO and Ubiquitin controls EBV/KSHV latent and lytic replication.

Abstract

Post-translational modification of target proteins by the Small Ubiquitin-like Modifier (SUMO) plays a critical role in regulation of many cellular processes including transcription, RNA processing, protein trafficking, DNA repair, and chromosome segregation, and is also often hijacked by viral infections. Epstein-Barr Virus (EBV) and Kaposi’s sarcoma-associated Herpesvirus (KSHV), two human oncogenic herpesviruses with a typical life cycle of latent and lytic replication, have been shown to be associated with many human cancers. In the past decade, intensive studies have investigated the interplay between tumor virus infection and SUMO-modification. In this review, we summarize the current knowledge as to how SUMOylation can regulate latent and lytic replication of EBV and KSHV, and the strategies by which these oncogenic herpesviruses usurp the SUMO pathways to establish a favorable microenvironment to promote host cell survival and proliferation in latency, and reactivate virion production during lytic replication, which are critical contributors to the development of EBV/KSHV-associated human malignancies.

Introduction

The Small Ubiquitin-like Modifiers (SUMOs) are small peptides that consist of more than 100 amino acid residues, and have a molecular weight of approximately 11 KDa (Bayer et al., 1998; Hay, 2001; Isogai and Shirakawa, 2007)(Mossessova and Lima, 2000). The three-dimension structure of the SUMO molecule is quite similar to ubiquitin protein, while they exert different functions. It has been demonstrated that there are at least four isoforms of SUMO proteins, namely SUMO1, SUMO2, SUMO3, and SUMO4, encoded by human genome (Guo et al., 2015; Melchior, 2000). Among them, SUMO1 has 50% sequence homology with SUMO2, while SUMO2 and SUMO3 are highly similar, with only 3 amino acid differences (a homology of 95%), which is often referred as SUMO2/3 (Matic et al., 2008). Although both SUMO1 and SUMO2/3 play a major role in many cellular processes and are expressed in various tissues (Vertegaal et al., 2006), they differ greatly in function, and the substrate proteins bound to them in vivo are also different (Rosas-Acosta et al., 2005). SUMO1 mainly modifies protein molecules under physiological conditions, whereas SUMO2/3 modifies protein molecules that are in response to stress, and can even compensate for loss of SUMO1 in knockout mice (Saitoh and Chemistry, 2000). SUMO4 is encoded by a sequence within the intron of the TAB2 gene and is expressed mainly in the kidney tissue (Zhong et al., 2009). Even though the function of SUMO4 remains unclear, a few studies have linked it to the susceptibility of type 1 diabetes (Song et al., 2012; Zhong et al., 2009). Interestingly, recent studies have also identified that another isoform SUMO5, as a novel and controversial small ubiquitin-related modifier, is considered to be a nucleosome regulator with restricted tissue expression. However, it needs to be further studies are required to confirm whether SUMO5 is an endogenous protein required for translation (Liang et al., 2016).

Increasing studies have shown that SUMO modifications can act as adapters to regulate the interaction between target proteins and other proteins through a specific conserved cis-acting DNA sequence referred to as the SUMO-interaction motif (SIM)) or SUMO-binding motif (SBM) which is non-covalently bound to SUMO. Once the target protein is modified by SUMO, it can promote the interaction between the target protein and the protein containing the SIM/SBM (Geiss-Friedlander and Melchior, 2007; Kerscher, 2007). Similar to ubiquitylation, SUMOylation is also involved in multiple cellular processes, such as protein trafficking, protein stability, the ability of proteins to interact with substrates including proteins or DNA, and proteins function (Kerscher, 2007; Kerscher et al., 2006). Likewise, SUMOylation also requires a cascade of enzymatic steps catalyzed by the E1-activating enzyme, E2-conjugating enzyme, E3-ligating enzyme (i.e. the protein inhibitor of activated STAT [PIAS] family, RanBP2, MMS21, Pc2, and Topors) (Wimmer et al., 2012), and is removed by SUMO proteases (i.e., the Ulp and SUMO/sentrin-specific peptidase [SENP] family, DESI1, DESI2, and USPL1) (Hickey et al., 2012; Müller et al., 2001). However, in contrast to ubiquitylation, each SUMO isoform is initially synthesized as an immature precursor protein with a C-terminal tail which contains a short peptide of about 2 ∼ 11 residues (Huang et al., 2004). Therefore, it must undergo maturation when a SUMO-specific protease (Ulp) removes the C-terminal tail of the short pro-peptide and exposes the iconic C-terminal Gly-Gly module sequence (Wimmer et al., 2012), so that SUMO can then be transferred to the mature functional proteins to play its reversible role through modification (Geiss-Friedlander and Melchior, 2007).

In the process of SUMOylation, the first step is the activation of a mature protein by the SUMO activating enzyme (E1) (a heterodimer composed of Uba2 and Aos1), mediated through an ATP-dependent SUMO-adenylate intermediate, and is activated by a thioester bond to the cysteinyl residues of Uba2 (Geiss-Friedlander and Melchior, 2007). Next, the activated SUMO is transferred to the cystine residue of SUMO specific binding enzyme (E2) Ubc9 by a transester reaction, which forms a thioester linkage between the catalytic Cys residue of Ubc9 and the C-terminal carboxyl group of SUMO (Johnson and Blobel, 1997; Lee et al., 2000). Finally, Ubc9 transfers SUMO to the substrates by forming an isopeptide bond between the C-terminal Gly residue of SUMO and a Lys side chain of the target proteins (Müller et al., 2001; Saitoh et al., 1998). This process can be promoted by SUMO E3 ligase, which catalyzes the transfer of SUMO from Ubc9 to a substrate (Geiss-Friedlander and Melchior, 2007). Although experiments in vitro have shown that E1 and E2 are sufficient to regulate SUMOylation of various substrates, it has been demonstrated that the process of most SUMO conjugation to the target molecule still requires participation of E3-ligating enzyme in vivo.

Herpesviruses is a large class of DNA viruses that are morphologically distinct from other viruses, and have a linear, double-stranded DNA genome (encoding 100-200 genes) contained within a 16-icosahedral capsid (Davison et al., 2009; M. et al., 2013). According to their physicochemical properties, they are categorized into three subfamilies of α, β, and γ (ME and EM, 2017; Roizmann et al., 1992​). In comparison to other members of the herpesvirus family, γ-herpesvirus is the only lymphotropic viruses that can induce tumors from natural or laboratory host cells (Geng and Wang, 2015; Roizmann et al., 1992). Among these, Epstein-Barr virus (EBV, also referred to as human herpesvirus-4, HHV-4), and Kaposi’s sarcoma-associated herpesvirus (KSHV, also named human herpesvirus-8, HHV-8), are members of the human γ-herpesvirus subfamily, which have been shown to infect different cell types including B cells, endothelial cells, and epithelial cells, and are closely associated with many human malignancies (Table 1). EBV, the first human tumor virus isolated from Burkitt’s Lymphoma Cells, was identified for the first time by Epstein, Achong and Barr in 1964 (Martel et al., 2012), and causes more than 200,000 cases of cancer every year (Ko and Medicine, 2015). This virus could primarily cause various lymphomas, including Burkitt lymphoma (BL), Hodgkins lymphoma (HL), NK/T-cell lymphoma and immunodeficiency-associated lymphoma. In addition, it can also induce some carcinomas, such as nasopharyngeal carcinoma (NPC) and EBV-associated gastric carcinoma (EBV-GC) (Young et al., 2016). KSHV was discovered from Kaposi's sarcoma (KS) patients infected with human immunodeficiency virus by Chang and Moore in 1994 (Chang et al., 1994), and it is also associated with various lymphomas, including primary effusion lymphoma (PEL), Multicentric Castleman’s disease (MCD) and KSHV-associated germinotropic lymphoproliferative disorder (Reid, 2011). Because these cancers often appears in the condition of immunosuppression, it has made KSHV-associated malignancies an increasing global health concern with the persistence of the AIDS epidemic (Giffin and Damania, 2014).

Like other viruses in the herpesviruses family, EBV and KSHV are capable of establishing latent infection in host cells and reactivating for lytic replication under the certain condition (Longnecker and Neipel, 2016). It has been assumed that the transforming capacities of EBV and KSHV are predominantly mediated by viral latent proteins, along with low level of reactivation of infectious virus production from latently infected cells (Okano, 1998). During latency, EBV expresses all eight latency-associated proteins, namely nuclear antigens EBNA1, 2, 3A–3C, LP, and the three latent membrane proteins LMP1 and LMP2A and 2B (Murray et al., 1992; Styles et al., 2017)(Jha et al., 2015). According to the expression patterns of the EBV latent proteins in different tumors, the EBV latency programs are classified into four main types of 0, I, II and III. The latency III program is detected predominantly in EBV-transformed lymphoblastoid cell lines and lymphoproliferative diseases in immunocompromised individuals with all the EBNAs and LMPs expressed (Rea et al., 1994). Latency II is found in the most Hodgkin’s lymphoma and nasopharyngeal carcinoma with only expression of EBNA1, LMP1, LMP2-A and LMP2-B (Thorley-Lawson, 2001). In latency I, there are two EBER genes expressed as in proliferating memory B cells and EBV-associated BL, the BART transcripts, and EBNA1 (EBV nuclear antigen 1) (Dirmeier et al., 2003). In latency 0, the healthy individuals carrying EBV with expression of non-translated viral RNAs (EBERs and miRNAs) but no viral protein (Gregory et al., 1998; Hochberg et al., 2004). In contrast to EBV, KSHV expresses a limited number of proteins including LANA, vFLIP, vCyclin during latent infection (Dittmer et al., 1998; Jennera and Boshoffb, 2002), which play critical roles in cell proliferation, apoptosis, and escape of the host immune surveillance. The majority of the KSHV genome remains silent during latency. However, expression of the lytic master regulator RTA, encoded by ORF50 (Gradoville et al., 2000; Lukac et al., 1999), is necessary and sufficient to switch from latent to lytic replication (Martin et al., 2000). Similarly, the expression of EBV lytic genes Zta and Rta, also play essential roles in switching from latent to lytic replication.

Emerging evidence has shown that many viral proteins can also be a SUMO-binding substrate, and alternation of SUMO signaling can promote virus infection in host cells [The progression about global effect of SUMOylation on virus could refer to recent reviews (Isaacson and Ploegh, 2009)(Zhang et al., 2009)(Peter and Viruses, 2015)(Everett et al., 2013)]. The interaction between the human oncogenic herpesviruses and SUMO modifying system in host cells has also been comprehensively investigated in the past decade. In this review, we mainly summarize the role of SUMOylation in EBV/KSHV infection, particularly as to how the SUMO system alters latent and lytic replication of the EBV/KSHV life cycles, and how these viruses interfere with cell cycle progression of the host by hijacking the SUMO modification pathway.