Tegument proteins of Epstein-Barr virus: Diverse functions, complex networks, and oncogenesis

The tegument is the structure between the envelope and nucleocapsid of herpesvirus particles. Viral (and cellular) proteins accumulate to create the layers of the tegument. Some Epstein-Barr virus (EBV) tegument proteins are conserved widely in Herpesviridae, but others are shared only by members of the gamma-herpesvirus subfamily. As the interface to envelope and nucleocapsid, the tegument functions in virion morphogenesis and budding of the nucleocapsid during progeny production. When a virus particle enters a cell, enzymes such as kinase and deubiquitinase, and transcriptional activators are released from the virion to promote virus infection. Moreover, some EBV tegument proteins are involved in oncogenesis. Here, we summarize the roles of EBV tegument proteins, in comparison to those of other herpesviruses.


Introduction
Herpesviruses have double-stranded linear DNA genomes in icosahedral containers composed of viral capsid proteins. The nucleocapsid is wrapped in a lipid bilayer envelope, from which glycoproteins protrude. The space and/or the components between the nucleocapsid and envelope is termed the tegument [1]. The tegument of herpesviruses is mainly composed of viral proteins, with some cellular proteins. Tegument proteins do not comprise a solid structure like the nucleocapsid and, thus, a portion of the tegument structure can be released into the cytoplasm when the viral envelope fuses with the plasma membrane or endosomal membrane [2]. Some types of tegument proteins form interfaces with the exterior of the nucleocapsid (the inner tegument), and some others with membrane proteins on the interior surface of the envelope (outer tegument). Other tegument proteins form a complex network that connects the inner and outer tegument proteins.
Herpesvirus infections comprise latent and lytic phases [3,4]. In the latent phase, the virus exists as a circular double-stranded DNA molecule in the nucleus of a host cell, expressing only a small number of viral genes. This is an endurant type of infection, where the virus minimizes viral antigen presentation to avoid host immunity. By contrast, the lytic phase is an active type of infection, in which all viral genes are expressed, viral DNA is replicated, and progeny virus particles are produced. The lytic cycle starts with expression of a handful of viral genes, named immediate-early (IE) genes. IE genes are activators of viral gene expression and induce the expression of early viral genes. The proteins encoded by early genes, including those needed for nucleotide metabolism and DNA replication, amplify the viral genome in replication compartments in the host cell nucleus. The remaining viral lytic genes are categorized as late, and most encode viral structural proteins i.e., capsid, tegument, and glycoproteins. Capsid proteins assemble into an icosahedral architecture, into which a viral genome is incorporated in the nucleus. Because the resultant nucleocapsid is too large to pass through nuclear pores, it must traverse the nuclear double membrane [5,6]. First, the nucleocapsid buds into the inner nuclear membrane (primary/initial envelopment). Some tegument proteins are attached to the exterior of the nucleocapsid in the nucleus before the initial envelopment. The temporal envelope fuses with the outer nuclear membrane (de-envelopment) to release nucleocapsid into the cytoplasm. The nucleocapsid buds into a cytoplasmic membrane structure (possibly derived from the TGN or endosome) [7]. The other tegument proteins are incorporated during this secondary envelopment. The cytoplasmic membrane structure finally fuses with the plasma membrane, and a mature progeny virion is released. The released virion attaches to a receptor on the plasma membrane and enters a cell. Two routes have been indicated for herpesvirus entry [8]. Route 1 involves fusion of the viral envelope with the plasma membrane and release of nucleocapsid and tegument proteins into the cytoplasm. In route 2, the virion is engulfed by an endosome, and the viral envelope fuses with the endosomal membrane to release nucleocapsid and tegument proteins. Released tegument proteins play a regulatory role in newly infected cells and promote virus infectivity. Other tegument proteins attached to the incoming nucleocapsid support nuclear transportation of the nucleocapsid. The linear viral DNA genome enters the nucleus, where the genome is circularized and gains histones to become an episome.
Herpesviruses are categorized as alpha-, beta-, and gammaherpesviruses. To date, nine human herpesviruses have been identified. Herpes simplex virus type 1 (HSV-1), HSV-2, and varicella zoster virus (VZV) belong to the alpha-herpesvirus group. HSV-1 and 2 cause herpes labialis, keratitis, encephalitis, and genital herpes. Chicken pox in children is attributable to initial VZV infection, and the same virus later causes shingles. The beta-herpesviruses include human cytomegalovirus (HCMV), human herpesvirus type 6A (HHV-6A), HHV-6B, and HHV-7. HCMV causes interstitial pneumonia, retinitis, and congenital HCMV infection, and HHV-6B and 7 are agents of exanthema subitem. Epstein-Barr virus (EBV) and Kaposi's sarcoma-associated herpesvirus (KSHV) are gamma-herpesviruses. Initial EBV infection in children is associated with no obvious symptoms but can lead to infectious mononucleosis if initially infected during/after adolescence. Furthermore, EBV is associated with several cancers, such as Burkitt lymphoma, Hodgkin lymphoma, post-transplant lymphoproliferative disorder, T/ NK cell lymphoma, chronic active EBV infection, nasopharyngeal carcinoma, and gastric carcinoma. KSHV is also an oncogenic virus, being the cause of Kaposi sarcoma, primary effusion lymphoma, and multicentric Castleman disease.
Some EBV tegument proteins are conserved across the Herpesviridae, but others-including BKRF4, BLRF2, and BNRF1-are unique to gamma-herpesviruses [9]. In this review, we summarize the structures and functions of EBV tegument proteins by comparison with other herpesviruses. Detailed electron microscopic analysis of the tegument structure of the whole EBV virion has not been performed, but protein-protein interactions (PPIs) analyses showed a plausible meshwork structure. Such complex PPIs hamper functional analysis of tegument proteins; following knockout of a tegument gene, loss of the encoded protein may affect incorporation of other tegument proteins. The multiple overlapping PPIs of tegument proteins may compensate for the loss of a tegument gene and obscure the phenotype of the mutant virus. Nevertheless, we and others have performed phenotypic analyses of EBV knockouts to assess the functions of tegument genes. We here summarize the findings of those analyses.

PPI network of EBV tegument proteins, and with nucleocapsid and envelope proteins
Tegument proteins form a complex PPI network that attaches the viral envelope to the nucleocapsid. We prepared a simplified diagram of the EBV PPI network based on prior reports [20][21][22][23] (Fig. 1). For comparison, a PPI network centering on HSV tegument proteins was also generated (Fig. 2) [2,7,[24][25][26][27]. Most PPIs have been detected in lysates (by immunoprecipitation) or cells (yeast two-hybrid or complementation assay of split marker), but not in the tegument. Nevertheless, these interactions are important because layers of tegument proteins are formed during envelopment in cells. Unfortunately, no similar schema for HCMV is available possibly because of the large number of tegument components [28][29][30].

Roles of EBV tegument proteins in the virus lifecycle
The roles of EBV BGLF1 and BVRF1 are unclear, but can be inferred from other herpesviruses, such as HSV-1. HSV UL17 (homolog of BGLF1) is essential for viral genome DNA cleavage and packaging [64]. UL25 (homolog of BVRF1) is required for packaging but not for cleavage [65,66]. Interestingly, the UL25 is also involved in viral genome uncoating immediately after infection [67] and nuclear egress of the nucleocapsid via interaction with the nuclear egress complex [36,68,69]. So, EBV BGLF1 and BVRF1 may be involved in viral genome cleavage and packaging, and possibly also viral genome containment, nuclear export, and nucleocapsid uncoating.

BOLF1 (homolog of HSV UL37, HCMV UL47, and KSHV ORF63)
To date, only one study has focused on the role of EBV BOLF1 [41]. Disruption of BOLF1 had little or no effect on viral gene expression or DNA replication. The knockout virus produced progeny at a level similar to the wild-type, but the progeny had reduced infectivity.

BSRF1 (homolog of HSV UL51, HCMV UL71, and KSHV ORF55)
BSRF1 is post-translationally modified by palmitoylation and associates with viral envelope or membrane. The phenotype of BSRF1knockout EBV was indistinguishable from wild-type EBV in HEK293 cells, but knockdown of BSRF1 in B95-8 cells significantly decreased infectious progeny production [14]. Expression of BSRF1 promoted BBRF2 transport from the nucleus to the cytoplasm [48].
HSV UL51 is also modified by palmitoylation, which is required for its membrane localization [46]. UL51 is implicated in viral maturation, secondary envelopment, egress, and intercellular spread [50,98]. UL51 is needed for proper localization of UL7 to cytoplasmic membranes [45]. It also associates with gE and UL14, which promote proper localization and envelopment, respectively [49,50]. In addition, UL51 phosphorylation affects viral nuclear egress and intercellular spread [99].

BBRF2 (homolog of HSV UL7, HCMV UL103, and KSHV ORF42)
Knockout of the EBV BBRF2 gene reduced progeny production without affecting viral gene expression and DNA replication [48]. The number of progeny produced was not reduced, but their infectivity was decreased. Co-expression of BBRF2 protected its interacting partner, BSRF1, from proteasome-dependent degradation, and co-expression of BSRF1 is needed for proper localization of BBRF2 [48], indicating the Network formed by protein-protein interactions (PPIs) of all tegument proteins (oval), nucleocapsid (dark red hexagon at the bottom), and envelope proteins (green box). Purple and blue ovals, conserved and non-conserved tegument proteins, respectively. Orange ovals show possible tegument components not included in Fig. 1. EBV gene names are provided in ovals. Saw-toothed line, post-translational lipid modifications-palmitoylation and myristoylation. A copy of BGLF1, two copies of BVRF1, and two copies of BPLF1 comprise the capsid-associated tegument complex (CATC). Red lines indicate interactions conserved across herpesviruses. Thick lines indicate interactions that were previously reported in at least one species of human herpesviruses in addition to EBV, or that were found only in EBV but validated by two or more methods. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) importance of the PPI.
Loss of the UL7 gene, a homolog of BBRF2, in HSV results in reduced virus replication and smaller plaques [100][101][102]. A UL7 deletion mutant of pseudorabies virus, an animal alpha-herpesvirus, had defects in secondary envelopment and egress [103].

BBLF1 (homolog of HSV UL11, HCMV UL99, and KSHV ORF38)
BBLF1 is a membrane-associated protein modified by palmitoylation and myristoylation [54]. Myristoylation of BBLF1 is crucial for its stabilization and localization to the TGN [54]. BBLF1 interacts with BGLF2, thereby mediating proper cytoplasmic localization of BGLF2 to the TGN [51]. BBLF1 knockdown reduced EBV progeny production [54]. We found that, rather than being secreted extracellularly, BBLF1-null mutant virus progeny remained associated with the host cell. Because BBLF1 knockout and wild-type viruses produced similar numbers of infectious progeny in total, BBLF1 is likely to promote extracellular egress without affecting processes at virion maturation [104].
All EBV BBLF1 homologs in other herpesviruses undergo palmitoylation and myristoylation. HSV UL11 is involved in secondary envelopment [105] and so knockout of the encoding gene reduces progeny production and causes cytoplasmic accumulation of unenveloped nucleocapsids. Its associations with UL16 and gE are implicated in secondary envelopment [58,106,107].

BGLF2 (homolog of HSV UL16, HCMV UL94, and KSHV ORF33)
EBV BGLF2 arrests the cell cycle at G 1 /S phase by inducing p21 [108]. It also induced AP-1 signaling by activating the P38 MAPK and JNK signaling pathways, thereby promoting EBV lytic replication [109,110]. In addition, BGLF2 inhibits NF-κB [111] and interferon-induced JAK/STAT pathway [112,113] signaling, to subvert antiviral innate immunity and induce the viral lytic cycle. It was also reported to de-regulate miRNA functions by targeting AGO2 [114]. BGLF2 is incorporated not only into virions but also exosomes, and exosomal BGLF2 facilitates de novo EBV infection [115]. Disruption of the BGLF2 gene had little effect on viral gene expression and DNA synthesis, but extracellular production of infectious progeny was suppressed by about 10-fold and the cell-associated progeny level was also decreased [51,109]. So, the BGLF2 gene improves not only virus egress but also the infectivity of secreted virions [109]. BGLF2 interacts with BBLF1 and BKRF4 [51,57]. When expressed alone, BGLF2 localizes to the nucleus and cytoplasm, but its co-expression with BBLF1 recruited BGLF2 to the TGN [51], and addition of BKRF4 re-localized BGLF2 to the nucleus and perinuclear region [57]. MHV-68 ORF33 is associated not only with cytoplasmic but also with nuclear nucleocapsids [116], suggesting that BGLF2 associates with nucleocapsids in the nucleus before primary envelopment. Other reports on KSHV ORF33 suggest roles for BGLF2 in the cytoplasm [117,118].
UL16-null mutant of HSV-1 exhibits about a 10-fold reduction in virus yield [119], and that of HSV-2 has 50-100-fold progeny loss [120]. Disruption of the UL16 gene in HSV-1 is associated with cytoplasmic accumulation of nucleocapsids [60]. HSV-1 UL16 is localized to the nucleus and cytoplasm, and interacts with nucleocapsid [121,122] and membrane proteins, including UL11 [123] and gE [59]. These findings, in conjunction with information from other herpesviruses, implicate UL16 in secondary envelopment and intercellular spread [124]. However, it must also be noted that HSV-2 UL16 plays an important role in the nuclear egress of nucleocapsids [125,126], suggesting that the roles of UL16 homologs are different depending on the virus species. In addition, modification of signaling pathways, e.g., MAPK, NF-κB, and JAK/STAT, has not been reported for HSV UL16 or KSHV ORF33.

BNRF1 (homolog of KSHV ORF75)
The major tegument protein BNRF1 is conserved only among gamma-herpesviruses. It is expressed in latent cells, too, and is targeted by CD4 + and CD8 + T cells [152,153]. Disruption of the BNRF1 gene had little effect on EBV replication and progeny production. However, upon infection to primary B cells, viral gene expression was significantly repressed by the disruption, thereby suppressing B-cell transformation [62]. The group therefore speculated that nuclear transport of the nucleocapsid was inhibited by disruption of the BNRF1 gene. Later, however, other group showed that BNRF1 is involved in transcriptional activation of viral genes upon infection; BNRF1 targets PML-NB (also known as ND10) and disrupts the antiviral histone chaperone complex Daxx-ATRX, thereby increasing the expression of viral genes immediately after infection [154,155]. In addition, BNRF1 destabilizes the SMC5/6 cohesin complex to increase viral DNA replication [156] and induces centrosome amplification and thus chromosomal instability [157].

BKRF4 (homolog of KSHV ORF45)
BKRF4 is an EBV late phosphoprotein [57] conserved among gamma-herpesviruses, albeit at low similarity. Knockout of the BKRF4 gene decreased progeny levels, possibly by reducing infectivity, but had no effect on viral gene expression and DNA replication [57]. Upon lytic induction of infected cells, BKRF4 localizes to the nucleus and peri-nuclear region. When expressed alone, BKRF4 and BGLF2 localize to the nucleus and cytoplasm, respectively, and co-expression of BKRF4 with BGLF2 re-localizes BGLF2 to the nucleus [57]. Interestingly, BGLF2 activates AP-1 signaling, an effect repressed by co-expression of BKRF4 [109]. BKRF4 associates with many other tegument proteins (Fig. 3), suggesting it to be a hub in the tegument meshwork [22]. BKRF4 inhibits the DNA damage response, suggesting involvement in oncogenesis [159]. Mutagenesis and structural analyses revealed that BKRF4 inhibits the DNA damage response by binding to partially unfolded nucleosomes [160]. In addition, BKRF4 has histone chaperone activity [161].

BLRF2 (homolog of KSHV ORF52)
The EBV BLRF2 gene encodes a gamma-herpesvirus-specific tegument protein expressed with late kinetics. The protein localizes to the nucleus and nuclear rim in transfected or infected cells [22]. Viral gene expression and viral DNA replication were not affected by the BLRF2 knockout, but production of infectious progeny into the culture medium was mildly but significantly dropped, possibly because of decreased infectivity [22]. A dimerization motif in the middle of BLRF2 mediates its self-association [177] and formation of this homodimer enhances protein stability [22]. The C-terminus of BLRF2 has a motif for phosphorylation by SRPK2, which is linked to progeny production [178]. Interestingly, BLRF2 also associates with many other tegument proteins and, thus, is likely to be a hub protein [22].

BGLF3.5 (homolog of HSV UL14 and KSHV ORF35)
EBV BGLF3.5 is conserved in alpha-and gamma-, but not in beta-, herpesviruses [187]. BGLF3.5 associates with other tegument proteins, such as BSRF1 and BLRF2. BGLF3.5 is likely to be a tegument gene, although its product has not been identified in EBV virion, possibly because of its small size. Knockout of BGLF3.5 did not affect virus replication or progeny production in HEK293 cells [188].
HSV UL14 is a homologous tegument protein [189] that increases the progeny titer by about one order of magnitude [16,190]. It has multiple functions, such as chaperone-like activity [191], protection from apoptosis [192], and aiding nuclear transport of VP16 and nucleocapsids [190,193]. KSHV ORF35 was detected in purified virions [18]. Knockout of ORF35 in KSHV slightly decreased viral gene expression, viral DNA synthesis, and progeny titer compared to the wild-type [194]. In MHV-68, viral gene expression and viral DNA synthesis were unaffected, but production of infectious progeny was decreased by the knockout [195].

BALF1 (possibly related to KSHVORF16)
EBV encodes two homologs of the host anti-apoptotic protein BCL2: BHRF1 and BALF1 [196,197]. Knockout of either or their simultaneous disruption had little effect on virus multiplication and progeny production. However, knockout of either moderately decreased B cell transformation efficiency, and their simultaneous disruption almost abolished B cell transformation [198], indicating that both vBCL2 proteins are required for B cell transformation. BALF1 reportedly stimulates autophagy [199]. We reported that it interacts with BSRF1 and is incorporated into the tegument [14].
KSHV has one BCL2 homolog, ORF16, the absence of which decreases viral gene expression, DNA synthesis, and progeny production [200]. Intriguingly, KSHV ORF16 protein interacts with ORF55, a homolog of the EBV tegument protein BSRF1, and mediates the incorporation of tegument proteins into virions [15].

BXLF1 (homolog of HSV UL23 and KSHV ORF21)
BXLF1 encodes a thymidine kinase (TK), which is a component of the thymidine salvage pathway of nucleotide biosynthesis [210]. Disruption of the gene had little effect on virus multiplication in cell culture [211]. The protein has been detected in EBV particles [10], but its importance is unknown.
The alpha-and gamma-herpesviruses, but not the betaherpesviruses, have TK genes. HSV TK (UL23) and KSHV TK (ORF21) have been detected in the virion tegument fraction [18,189,204]. Disruption of HSV UL23 moderately decreased virus replication [212]. Knockout of KSHV ORF21 had no effect on viral gene expression and DNA synthesis, but markedly suppressed infectious progeny production, presumably by decreasing infectivity [213]. Upon infection, ORF21 protein in the tegument is released into the cytoplasm, where it stimulates MEK signaling to enhance infectivity.

BORF2 (homolog of UL39, HCMV UL45, and KSHV ORF61)
The BORF2 gene of EBV encodes the large subunit of ribonucleotide reductase (RR), which is involved in nucleotide biosynthesis. A BORF2null mutant virus exhibited lower progeny production than the wildtype [214]. Although the physiological role of BORF2 in the tegument is unknown, expression of the BORF2 gene increases the proportion of cells at G 1 /S phase by increasing the P53 protein level [108]. In addition, BORF2 associates with, and thereby suppresses the activity of, APOBEC3B, to maintain genomic integrity [214,215], which increases the number of cells in G 1 /S phase [216].
The large subunit of the RR gene is preserved in HSV, CMV, and KSHV, but HCMV UL45 lacks enzymatic activity [217]. HSV UL39 is non-essential for multiplication in cell lines but required for efficient replication in non-dividing cells [218]. EBV BORF2, HSV UL39, and KSHV ORF61 inhibit the restriction factor, APOBEC3B [219]. HCMV UL45 [220] and HSV UL39, but not KSHV ORF61 protein, have been detected in virions [204], together with the RR small subunit [189]. Two-hybrid analysis indicated that HCMV UL45 protein serves as a tegument network hub, and has many interactions [221].

BRLF1 (homolog of KSHV ORF50)
BRLF1 is one of the two IE genes encoded by EBV implicated in viral reactivation. This transcriptional activator is required for lytic initiation, particularly in differentiated epithelial cells [222]. BRLF1 protein is incorporated into the tegument [19] and binds a capsid triplex protein, BORF1, thereby preventing its ubiquitin-dependent degradation in transfected cells [19].
Homologs of EBV BRLF1 are found only in gamma-herpesviruses. KSHV ORF50, also known as K-Rta, is a key lytic activator [223], which, unlike EBV BRLF1, has not been detected in purified virions [204].

BRRF2 (homolog of KSHV ORF48)
The BRRF2 gene product is a late phosphoprotein. BRRF2 localizes to the cytoplasm and its knockout decreased progeny production [224]. Its homologs are present only in gamma-herpesviruses. KSHV ORF48, a homolog of BRRF2, was detected in virions [18], but its function is unclear.

Roles of EBV tegument proteins in oncogenesis
There are at least 10 hallmarks of cancer-evasion of apoptosis, selfsufficiency in growth signals, insensitivity to anti-growth signals, tissue invasion and metastasis, limitless replicative potential, sustained angiogenesis, dysregulation of cellular energetics, avoidance of immune destruction, tumor-promoting inflammation, and genome instability and mutation [225]. Latent genes of EBV, such as LMP1, LMP2A, EBNA1, EBNA2, EBNA3A, EBNA3C, and some viral microRNAs, have oncogenic activity, but accumulating pieces of evidence indicate that lytic genes are also involved in oncogenesis [226][227][228][229]. Below we summarize the EBV genes linked to oncogenesis.

Conclusion
We summarized the interactions and roles of EBV tegument proteins. The tegument proteins of the Herpesviridae have several conserved properties and functions whereas others are unique to EBV. The phenotypes of EBV knockout mutants are typically analyzed in HEK293 cells, but some processes of the EBV lifecycle are difficult to evaluate in the cell line, e.g., secondary envelopment. This may be because EBV replication is less efficient than HSV, and nucleocapsids cannot be easily detected in HEK293 cells. Alternatively, HEK293 cells might be conducive to secondary envelopment even in the absence of one or more tegument genes. BSRF1 knockout virus had a phenotype similar to wildtype virus in HEK293 cells, but knockdown of the BSRF1 gene reduced progeny virus production in B95-8 cells [14]. Therefore, EBV gene functions should be analyzed in the virus's natural host cell type, such as B cells, for further evaluation.
EBV tegument genes are good candidates for viral attenuation, for development of live vaccines, because their disruption does not cause complete inactivation and so does not prevent the production of other viral antigens, e.g., glycoproteins. Such development requires detailed analyses of knockout viruses, including in animal models.

Funding sources
This work was supported by Japan Agency for Medical Research and Development (JP21wm0325042) and the Takeda Science Foundation.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability
No data was used for the research described in the article.