The ability of viruses to co-opt cell signalling pathways has, over millions of years of co-evolution, come to pervade nearly every facet of cellular functions. Recognition of the extent to which the ubiquitin–proteasome system can be directed or subverted by viruses is relatively recent. Viral products interact with, and adjust, the ubiquitin–proteasome machinery precisely and at many levels, and they do so at distinct stages of viral life-cycles. The implications for both cells and viruses are fundamental, and understanding viral strategies in this context opens up fascinating new areas for research that span from basic cell biology to therapeutic interventions against both viruses and malignancies.
Many aspects of the life cycle of viruses rely on specific interactions between viral products and regulatory systems of the target cells. Over millions of years of co-evolution with their hosts, viruses have developed highly effective strategies that disrupt integrated functions of the cell and transmute cellular functions into viral functions that serve propagation of the virus. As the UPS (ubiquitin–proteasome system) is the nexus for many fundamental cellular processes, it is not surprising that viruses reprogramme this machinery according to viral needs.
Ubiquitin-dependent proteasomal degradation is one of the essential modes whereby the lifetime of cell-signalling participants is controlled, and viruses lose no opportunity to interfere with the UPS on every possible level. As will become apparent in this chapter, viral strategies are based on specific modulations rather than on extreme changes in ubiquitin–proteasome activity. This is because normally functioning proteasomes are too important for the survival of infected cells, and a serious imbalance in protein processing by the UPS would be detrimental to virus propagation. Moreover, the UPS offers an enzymatic way for a viral protein of low abundance to deal with a cellular target present in stoichiometrically greater amounts.
Knowledge of how viruses can affect host ubiquitin-dependent machinery grows with our understanding of the purpose of this system in general. We have to keep in mind that the term UPS embraces today much more than the simple concept that ‘the protein is ubiquitinated, therefore it will be destroyed’. Participation of ubiquitination in cell signalling is not limited to protein destruction by the proteasome. Ubiquitin and ubiquitin-like modifications are an integral part of the regulatory machinery of many cell processes, and through modulation of the UPS, viruses can successfully interrupt different aspects of ‘normal’ cell function.
In this chapter we discuss some known aspects of the relationships between animal viruses and the UPS. We place the emphases on three areas of viral manipulation of host-cell machinery with respect to ubiquitination: interruption of signal transduction pathways, inhibition of cellular immune responses to viral infection, and virion assembly and release from the host cell.
Dysregulation of cell-signalling pathways by tumour viruses: ubiquitination and deubiquitination
In 1911, Peyton Rous described a sarcoma that could be transmitted in chickens by inoculating them with a cell-free filtrate, later identified as containing RSV (Rous sarcoma virus). The era of oncogenic viruses had begun. Over the last 40–50 years, much knowledge of the molecular biology of cell transformation by oncogenic DNA and RNA viruses has accumulated. For example, the discovery of reverse transcriptase, which synthesizes DNA using the information in the viral RNA, and the identification of important transcriptional factors, such as AP-1 (activator protein 1) or tumour-suppressor proteins, such as p53, came from studying animal viruses .
A number of human viruses from different families are implicated in carcinogenesis [1,2]. Several HPVs (human papillomaviruses) are detectable in virtually all cervical cancers worldwide and in some other cancers (including some head and neck cancers). The contribution of HBV (hepatitis B virus) and HCV (hepatitis C virus) to HCC (hepatocellular carcinoma) has long been recognized. Other viruses that present a substantial cancer risk in certain populations include EBV (Epstein–Barr virus; associated with various lymphoid malignancies and nasopharyngeal cancer), HTLV-1 (human T-cell lymphotropic virus type-1; associated with adult T-cell leukaemia/lymphoma), KSHV [Kaposi's sarcoma-associated herpesvirus (human herpesvirus-8)], polyoma viruses [e.g. JCV (JC virus); associated with brain tumours], and SV40 (simian virus 40; associated with mesothelioma) (Table 1).
In spite of differences among human oncogenic viruses, the general strategy is the same: to protract cell-cycle progression and protect cells from apoptosis, resulting in the perpetuation of the virus genome. Therefore the common cellular targets for tumour virus oncoproteins are the most important transcriptional factors involved in oncogenesis, such as c-Myc, NF-κB (nuclear factor κB), AP-1 and p53 . Because there are innumerable cellular pathways that can regulate the transcriptional machinery of the cell, there are many opportunities for tumour viruses to dysregulate them, to the benefit of the virus. As a part of the basic functional machinery of the cell, the UPS is inevitably involved in viral oncogenesis.
Perhaps one of the most striking examples of the ability of oncogenic viruses to redirect the UPS to the advantage of the virus is provided by HPVs that are associated with malignant lesions of the anogenital tract . More than 99% of cervical carcinomas harbour at least one copy of a high-risk HPV genome. The viral E6 and E7 genes are the only viral genes that are generally retained and expressed, and are essential for malignant progression and maintenance of the transformed phenotype. Both viral proteins interact with important negative cell-regulatory proteins, namely E6 with the tumour suppressor protein p53 and E7 with the pRbs (retinoblastoma proteins). Both E6 and E7 utilize the UPS to target these proteins for degradation and, thus, inactivation  (Figure 1).
One of the major biochemical functions of the oncogenic E7 protein is induction of cellular DNA replication in differentiated epithelial cells. In differentiated cells, pRb and the related protein p130 bind E2F transcription factors to repress the expression of the replication enzyme genes [4,5]. E7 disrupts the interaction between pRb and E2F, resulting in the release of the E2F factors in their transcriptionally active forms. E7 induces degradation of pRb through the ubiquitin–26 S proteasome, but the mechanism is still unknown. The ubiquitin-dependent proteolysis of pRb involves both N- and C-terminal regions of E7 that are also critical for the transforming function of E7, suggesting that the proteolysis of pRb is linked to the transforming function of E7 . One of the possibilities is that, similar to the E6–p53 interaction (see below), E7 recruits a cellular E3 (ubiquitin ligase) to target pRb for ubiquitination and subsequent degradation. Another suggestion is that E7 may function as an adaptor between pRb and the proteasome, thereby targeting pRb directly to the proteasome without prior ubiquitination. No evidence has yet been provided to support any of these hypotheses. More recent studies showed that the E7 oncoprotein is itself regulated by the ubiquitin–proteasome pathway. The components of this regulatory ubiquitin-dependent cascade have been identified. In the presence of UbcH7 [E2 (ubiquitin-conjugating enzyme) H7], the SCF (Skp1–Cdc53/Cullin 1–F-box protein) E3 complex was shown to be responsible for E7 ubiquitination in vitro .
As a conclusion we have to admit that in spite of progress in the last few years, many questions remain unanswered about the relationship between E7 proteins and the UPS.
Much more has been achieved in understanding how another HPV oncogenic product, E6, dysregulates the cellular UPS. First, E6 recruits the E3 E6-AP (E6-associated protein). The dimeric E6–E6-AP complex then binds to p53, resulting in the E6-AP-mediated ubiquitination of p53 in the presence of certain E2s (UbcH5, UbcH7, or UbcH8) that interact functionally with E6-AP. Finally, polyubiquitinated p53 is recognized and degraded by the proteasome in HPV-infected cells. Although each of these steps can be reconstituted in vitro with highly purified proteins, it is still not clear how polyubiquitinated p53 is transferred from the E6–E6-AP complex to the proteasome in vivo.
E6-AP was originally isolated and identified as a cellular protein that binds to high-risk E6 proteins and is required by HPV E6 protein to bind efficiently to p53 . E6-AP contains the C-terminal E3 HECT (homologous to E6-AP C-terminus) domain that interacts with a number of E2 enzymes, including UbcH5, UbcH6, UbcH7 and UbcH8. Unlike other known E3s, which presumably function as bridging proteins between substrate proteins and E2s, HECT domain E3s have been proposed to catalyse directly the final attachment of ubiquitin to their substrate proteins . From the perspective of cell biology this is an interesting situation because the virus product preferentially alters endogenous substrate specificity: normally, p53 is a target for Mdm2 (murine double minute clone 2 oncoprotein) E3 [8,9], which belongs to a different, RING (really interesting new gene)-finger-containing class of E3s. What is the physiological relevance of this finding? It is well established that Mdm2-induced degradation of p53 is inhibited upon treatment of cells with appropriate stress stimuli, including genotoxic stress. Therefore one of the possibilities is that E6 targets p53 for degradation under conditions when the normal pathway for p53 degradation is not functional. This hypothesis is supported by results obtained in experimental cell-culture systems, where E6 was ectopically expressed from heterologous promoters. In contrast to parental cells (that do not express E6), p53 levels did not increase in E6-expressing cells in response to genotoxic stress [3,10]. Moreover, expression of E6 in various cell types results in an increased rate of mutagenesis and genetic instability. Since p53 plays an important role in preserving genome integrity, this connection suggests that expression of E6 interferes with p53 activity not only under conditions of stress, but also under apparently normal growth conditions.
Several E6-independent substrates of E6-AP have been reported, including HHR23A and HHR23B (the human homologues of Saccharomyces cerevisiae RAD23), Blk (a member of the Src-family of non-receptor tyrosine kinases), Bak (a human pro-apoptotic protein), and Mcm7 (which is involved in DNA replication). With respect to cervical carcinogenesis, an attractive but purely speculative possibility is that E6 not only tricks E6-AP into ubiquitinating p53, but that it also influences the turnover rate of E6-AP substrates in general. This hypothesis is supported by the observation that binding of E6 targets E6-AP for self-ubiquitination and degradation . Thus even if E6 does not directly influence the substrate specificity of E6-AP, an E6-induced decrease of intracellular E6-AP levels should have profound effects on the stability of E6-AP substrates. However, if this is indeed the case, future investigation is needed to prove it.
As well as p53, numerous cellular proteins have been reported to interact with the HPV E6 proteins and to be targeted for proteasomal degradation in an E6-AP-dependent (e.g. Bak, c-Myc, Mcm7, hScrib) or E6-AP-independent [e.g. hDlg (human discs large), MAGI-1 (membrane-associated guanylate kinase with inverted domain structure 1), MUPP-1 (multiple PDZ domain-containing protein 1)] manner . This apparent difference may seem surprising but opens up the exciting possibility that E6 has the capacity to interact specifically with at least two different E3s. To resolve this issue, it will be important to reconstitute E6-induced ubiquitination of E6-AP-independent targets of E6 in vitro and to identify all the components involved . Similarly, it will be interesting to see whether, in cells derived from E6-AP null mice, the respective murine homologues of hDlg and hScrib are targeted for degradation by E6 or not.
In the viral world, the ability to target p53 for degradation is not restricted to papillomavirus products. It has been demonstrated that a complex consisting of the E1B 55 kDa and the E4ORF6 proteins of adenoviruses (that have oncogenic properties in cell-culture systems and in animal models) binds to and induces the degradation of p53 via the UPS. In this case, the adenoviral oncoproteins are part of an E3 complex whose structure is reminiscent of the class of CBC (Cullin-2 elongin B/C complex) E3s. CBC E3s were first described in the context of the pVHL (von Hippel–Lindau tumour suppressor protein 1). In analogy to the pVHL CBC, the E1B/E4ORF6-containing complex consists of the elongins B and C, the RING-finger protein Rbx1 (RING box protein 1)/Roc that interacts with its respective E2 enzyme, and Cullin-5, which serves as a platform to bring these different proteins into a multi-subunit complex (Figure 2). Whether, like E6, the E1B/E4ORF6-containing complex can circumvent the normal stability regulation of p53 remains to be determined [1,12,13].
HBX (the X product of HBV) can activate the transcription of a variety of viral and cellular genes [14,15] and induce liver cancer in certain transgenic-mouse models. Since HBX does not bind to DNA directly, its activity is thought to be mediated via protein–protein interactions. HBX has been shown to enhance transcription through AP-1 and AP-2 and to activate various signal transduction pathways. Several studies have also identified possible cellular targets of HBX, including members of the CREB (cAMP-response-element-binding protein)/ATF (activating transcription factor) family, the TATA-box-binding protein, RNA polymerase subunit RPB5, the UV-damage DNA-binding protein, and the replicative senescence p55sen protein. HBX also interacts with p53 and inhibits its function. HBX binds to the proteasome complex in vitro and in vivo, and this interaction leads to inhibition of the chymotryptic peptidase and protease activities of the proteasome [16–18]. Some authors propose an intriguing hypothesis that HBX functions to counteract the increased proteolytic function of the cells.
Herpes simplex virus
Although HSV (herpes simplex virus) is not on the list of viruses showing oncogenic potential, this does not mean that problems caused by HSV infection are limited to annoying cold sores. HSV infections have a worldwide distribution and have been described in the medical literature for centuries .
Among other HSV gene products, ICP0 (infected cell protein 0) holds a special place in the ubiquitin–proteasome pathway. ICP0 is a unique example of a viral E3 protein with two independent E3 sites. The N-terminus of ICP0 contains a RING-finger domain that is found in the largest known class of E3s. The RING domain of ICP0 can induce the accumulation of polyubiquitin chains in the presence of the E2s UbcH5a and UbcH6 in vitro, and is required for ubiquitin-dependent degradation of ICP0 substrates in vivo. The targets of the RING domain of ICP0 include PML (promyelocytic leukaemia) antigen and Sp100, constituents of nuclear structures known as ND10 (nuclear domain 10). In addition to PML antigen and ND10 components, ICP0 has been reported to cause the degradation of the catalytic subunit of the DNA-dependent protein kinase and centromeric proteins C and A [19,20].
The C-terminal region of ICP0 contains a different E3 domain that does not have a RING finger and binds the E2 enzyme UbcH3. UbcH3 is the major E2 in the E1 (ubiquitin-activating enzyme)/E2/E3 complex that promotes ubiquitination and degradation of cyclin D1. Together with the evidence that ICP0 can stabilize cyclin D1 without binding to it, these results lead to the attractive hypothesis that the C-terminal domain of ICP0 acts as a pseudo-E3, competitively inhibiting proteasomal degradation of cyclin D1 .
In general, despite intensive research and enormous interest, it is still not clear how the functions encoded in ICP0 account for its phenotype in either infected or transfected cells. So far all evidence leads to the conclusion that ICP0 is a multifunctional protein and that its role in viral infection reflects the sum of its multiple and diverse functions. It is more than likely that the number and diversity of cellular proteins known to be targeted for destruction by ICP0 will increase, and therefore more information regarding its functions will be forthcoming.
Knowledge of viral modulation of the host cell signalling through the UPS has been augmented by new information. The product of KSHV [the IE (immediate-early) nuclear transcription factor RTA (replication transcription activator), a DNA-binding nuclear transcription factor that can act as the trigger for the entire KSHV lytic cycle] encodes E3 activity . IRF7 (interferon regulatory factor 7), a key mediator of type I interferon induction, is targeted by RTA for ubiquitin-dependent proteasomal degradation. RTA promotes polyubiquitination of IRF7 in an in vitro cell-free assay, demonstrating that RTA itself has E3 activity. Interestingly, RTA has an unconventional intrinsic E3 activity; therefore the authors suggest that the RTA non-canonical domain associated with its E3 activity may represent as-yet-unidentified variants of Cys- plus His-rich E3s.
Increasingly, EBV is proving to be remarkably versatile in its use of the UPS. In contrast to KSHV, the EBV oncoprotein LMP-1 (latent membrane protein 1) appears to regulate IRF7 by ubiquitination at ubiquitin Lys63, which leads to activation, not degradation, of the protein (L. Huye and J. Pagano, unpublished work). Quite possibly, EBV might also induce proteasomal disposal of IRF7 in later interactions.
For several years, Wnt signalling has been the object of intense attention in diverse biological areas. A central effector of the Wnt pathway is β-catenin, a multifunctional protein that was first identified as a component of the cadherin cell-adhesion complex. Normally the level of ‘free’ β-catenin in the cytoplasm is tightly regulated by rapid degradation through the UPS. To control levels of β-catenin, EBV can manipulate two distinct degradation pathways: the classical GSK3β (glycogen synthase kinase 3β)-dependent destruction machinery and the Siah-1 E3, which does not require phosphorylation of β-catenin. Both phenomena can operate in the same cells, possibly governed by different EBV gene products. Interestingly LMP-1, which can activate HIF (hypoxia-inducible factor) 1a, the oxygen-sensing transcription factor, appears to do so in part by affecting the level of Siah-1 (S. Kondo, K.L. Jang, J. Shackelford and J. Pagano, unpublished work).
In conclusion, it is pretty clear that the targeting of cellular proteins for proteasomal degradation through ubiquitination by way of viral or cellular E3s is an important aspect of infection and cell transformation by tumour viruses (Figure 3).
Although the ability of tumour viruses to manipulate the process of ubiquitination is now appreciated, whether viral manipulation of the opposite branch of these linked systems — deubiquitination — also holds true remains largely unstudied. If oncogenic viruses target signalling pathways in a ubiquitin-dependent manner, and DUBs (deubiquitinating enzymes) are also important regulators of these pathways, it would be logical to suggest that tumour viruses should affect cellular deubiquitinating processes as well. Consistent with this hypothesis, ICP0 binds HAUSP (herpes virus-associated ubiquitin-specific protease), a DUB that is important for p53 stabilization, thus adding another route for a viral product to inactivate p53. Interestingly, EBNA1 (EBV nuclear antigen 1) also interacts with HAUSP, and this interaction influences EBNA1 transcriptional activity .
Recently, two cytokine-inducible DUBs (DUB-1 and DUB-2) have been described. These haematopoietic-specific genes with unclear function are rapidly induced after cytokine stimulation. IL-2 (interleukin-2)-inducible DUB-2 is constitutively expressed in HTLV-1-transformed cells. This DUB prolongs cytokine-induced activation of STATs (signal transduction and activators of transcription) and suppresses apoptosis following cytokine withdrawal. Since IL-2 is constitutively expressed in HTLV-1-infected cells, this may be an example of where an oncogenic virus regulates DUB expression indirectly, through the activation of another gene .
That a human oncogenic virus might direct synthesis of its own DUB is an intriguing scenario that has recently been examined. Adenovirus infection increases deubiquitinating activity in infected cells via Avp (adenovirus proteinase), which can function as a DUB in vitro and in vivo . Compared with classical DUBs, Avp seems to act as an enzyme of low specificity, which suggests that this viral DUB might deubiquitinate different viral and cellular ubiquitinated substrates, although none have yet been identified.
In EBV latently infected B-lymphocytes, β-catenin is stabilized, cytoplasmic β-catenin is associated with active DUBs, and the Wnt pathway is activated . This observation supports the concept that, as a human tumour virus, EBV might affect signalling pathways through dysregulation of cellular deubiquitinating machinery .
It has become clear that a number of proteins regulating cellular mechanisms for homoeostasis in all eukaryotes may be controlled by both ubiquitination and deubiquitination, and that oncogenic viruses play a certain role in dysregulation of cell-signalling pathways that intervene in this system. In normal cells, the balance between the two processes is probably determined by a dynamic equilibrium and is highly regulated. Tumour viruses may affect ubiquitination directly by using their own ubiquitinating enzymes, or indirectly, by use of endogenous cellular components of the UPS. The recent studies of DUBs encoded by oncogenic viruses, as well as the evidence that a tumour virus can regulate cell-signalling pathways through deubiquitination, suggest the same possibility for the deubiquitinating system, and this has begun to open this new area of viral functionality (Figure 4).
Viral infection and the immune response: down-regulation of antigen presentation
Most multicellular organisms are capable of defending themselves from infectious intruders by mounting an immune response. The specificity of the T-lymphocyte of the immune system of higher vertebrates determines antigenic peptides bound to cell-surface MHC molecules. There are two classes of MHCs: MHC class I presents endogenous antigenic fragments to CD8+ T-cells, and MHC class II presents exogenous antigen peptides to CD4+ T-cells. The system provides continual surveillance against potential danger by monitoring whether cells are synthesizing ‘foreign’ proteins. In this process, MHC class I molecules bind oligopeptide fragments derived from a cell's expressed proteins and display them on the cell surface. Under normal physiological conditions, all of the class-I-presented peptides are derived from normal autologous sequences to which the immune system is non-reactive owing to self-tolerance. However, if the cell is infected by viruses or is expressing mutant gene products, then non-native peptides will be displayed and will stimulate CTLs (cytotoxic T-lymphocytes) to kill the affected cell.
By generating peptides from intracellular antigens, which are then presented to T-cells, the UPS plays a central role in this type of cellular immune response. The generation of peptide-loaded MHC class I molecules mostly requires the proteolytic generation of peptides with a preferred length of 8–10 amino acids in the cytosol and the transport of peptides via TAP (transporter associated with antigen processing) proteins into the ER (endoplasmic reticulum). There, peptides bind to MHC class I proteins, followed by transport of the peptide-loaded MHC molecules to the cell surface. One of the characteristics of the MHC class I antigen-presentation pathway is that several of its components are induced by the cytokine IFNγ (interferon-γ). These include the MHC class I heavy-chain, the TAP proteins, several of the 20 S proteasome subunits [LMP2 (low-molecular mass polypeptide 2), LMP7 and MECL-1 (multicatalytic endopeptidase complex-like-1)] and the proteasome activator PA28.
Although the immune system has evolved machinery to eradicate virally infected cells, many viruses can persist inside cells to cause latent or chronic infection of the host. One mechanism used by the virus to avoid recognition by immune surveillance is the down-regulation of MHC class I antigen presentation [26,27].
HCMV (human cytomegalovirus)
The degradation of MHC class I heavy-chains by HCMV is an example of this strategy. HCMV encodes four immunomodulatory proteins [gp (glycoprotein) US2, gpUS3, gpUS6 and gpUS11], which decrease cell-surface expression of MHC class I proteins. These are single-transmembrane-spanning, immunoglobulin-domain-family proteins that probably arose by duplication of a single ancestral gene. In the presence of any one of these viral proteins, MHC class I proteins fail to translocate to the infected cell surface; importantly, cell-surface MHC expression is not reduced by infection with a US2-US11 mutant virus. The US3 gene product, gpUS3, is an IE gene product that prevents egress of MHC class I proteins from the ER to the Golgi apparatus. US3 encodes three differentially spliced transcripts, only one of which encodes an MHC class I modulatory protein. The product of US6, gpUS6, complexes with TAP protein in the lumen of the ER, locking the conformation of TAP to prevent ATP-dependent peptide loading. The products of US2 and US11, gpUS2 and gpUS11 respectively, target newly synthesized MHC class I heavy chains for destruction via a pathway that involves ubiquitin-dependent retrograde transport, or ‘dislocation’, of the heavy chains from the ER to the cytosol, where the proteins are degraded by proteasomes. An Ig-like ER-luminal domain of gpUS2 is essential for this process and engages the α3 domain of the MHC class I heavy chain.
The UPS is involved in the sorting of proteins within compartments of the endocytic pathway. An increasing number of examples show that the down-regulation of antigen-presenting MHC class I molecules can also be achieved after membrane delivery by targeting into the endocytic pathway. The proteins required for this process are identified for KSHV.
KSHV K3 and K5 proteins [also known as MIR1 (modulator of immune recognition 1) and MIR2] have E3 activity that induces ubiquitin conjugation to MHC class I in transfected cells, and a critical cysteine residue in the PHD (plant homeodomain) of K5 protein is required for its self-ubiquitination in vitro. Similarly, the K3 protein of murine γ-herpesvirus also down-modulates the MHC class I in a PHD-dependent manner. The C-terminal domain of K3 and the cytoplasmic tail of MHC class I are essential for the association of K3 with newly synthesized MHC class I molecules in the ER membrane and their subsequent degradation. .
An interesting example of how a virus avoids immune-system responses comes from EBNA1, which is required for EBV to maintain latency. CTLs specific for EBNA1 can be readily isolated from EBV-infected subjects, but they fail to recognize cells expressing endogenous EBNA1. This viral protein contains Gly-Ala residues that prevent EBNA1 degradation within the proteasome and additionally sequester the cleaved viral products in a cytoplasmic compartment, rendering them inaccessible for presentation by MHC class I molecules. Although EBNA1 is not the only viral protein expressed during EBV latency, its inaccessibility to proteasomes makes a perfect camouflage to prevent recognition by the immune system .
Contribution of the UPS to retrovirus release
Interestingly, some enveloped RNA viruses have evolved sophisticated strategies to exploit the UPS at late stages in their replication. Critical steps in the retrovirus life cycle are virion assembly and fusion with the plasma membrane of infected cells and the final release or budding from the host cell . Studies by several independent groups showed that proteasome inhibitors interfere with the budding process of several viruses. The retroviral Gag protein is essential for the budding process, and it contains conserved motifs (late budding or L domains) like PPXY (Pro-Pro-Xaa-Tyr) or PTAP (Pro-Thr-Ala-Pro) sequences. Using the PPXY motif as bait in a two-hybrid system, several groups subsequently identified WW domain-containing HECT E3s such as Nedd4 (neural precursor cell expressed developmentally down-regulated 4) or Nedd4-like proteins. A functional E3 is required for viral budding, as expression of the WW domains alone inhibits the release of mature viral particles from the host cell. In the case of Ebola virus, the late domain-containing protein VP40 (viral protein 40) is a target for ubiquitin conjugation by the Nedd4 yeast homologue Rsp5. Thus E3s (HECT-type) are involved in viral budding through their recruitment to the PPXY motifs in the late budding domains .
Another exciting discovery that involves the UPS involves HIV budding. Unlike other retroviruses, HIV does not contain the PPXY motif in the Gag protein, but has a PTAP motif, which has been implicated in viral budding. By using the yeast two-hybrid screen with HIV Gag as bait, two groups identified Tsg101 (tumour susceptibility gene 101) as a binding partner. The PATP motif in the HIV Gag protein mediates direct binding with the UEV (ubiquitin E2 variant) domain in the N-terminus of Tsg101. More importantly, depletion of Tsg101 by use of small interfering RNAs significantly reduced budding of HIV from infected cells. Structural studies of the Tsg101 UEV domain in the complex with the HIV PTAP motif demonstrated that the UEV domain forms a binding groove that makes close contacts with the PTAP residues. The importance of Tsg101 in HIV budding is further supported by the finding that overexpression of the N-terminal UEV domain of Tsg101 inhibits HIV budding, suggesting that the full-length Tsg101 is required for facilitating virus release. As described above, Tsg101 belongs to the UEV family that contains E2 homologous sequences but not the active cysteine, and it functions in late endosomal trafficking for the activated epidermal growth factor receptor in mammalian cells. In yeast, the Tsg101 homologue Vps23 (vacuolar-protein sorting 23) forms a multimolecular complex with other Vps proteins, called ESCRT (endosomal sorting complex required for transport)-1, which recognizes monoubiquitinated cargo and helps sort it into MVBs (multi-vesicular bodies). Interestingly, the MVB pathway is quite similar to the viral budding process, both of which involve invagination of lipid membrane. In addition, a recent study showed that Tsg101 is recruited, together with other ESCRT-1 components, to the late budding domain . One distinctive difference is that Tsg101 binds directly to the PTAP motif in HIV instead of the ubiquitinated cargo protein, as in the MVB pathway in the yeast.
Studies of the role of UPS in virus budding are of more than academic interest, as virus-release machinery may constitute a novel target for the development of anti-retroviral therapies. Currently, two E3s, in addition to one E3-like protein, have been identified as regulators of HIV budding. These ligases might represent interesting targets for therapeutic intervention [36,37].
Despite considerable recent progress in our understanding of the role of the UPS in retroviral budding, fundamental questions remain to be answered. The search for answers is likely to keep this an exciting field of study for the foreseeable future.
Viruses and ubiquitin-like modifications
Understanding of the importance of post-transcriptional protein–protein modifications is no longer limited to ubiquitin-dependent alterations. Ubiquitin-like proteins also participate in signal transduction in general and virus–host-cell relations in particular.
SUMO (small ubiquitin modifier) belongs to a growing number of ubiquitin-like proteins that covalently modify their target proteins. Although some evidence supports a role of SUMO modification in regulating protein stability, most examples still support a model by which SUMO alters the interaction properties of its targets, often affecting their subcellular localization behaviour. The PML protein localizes both in the nucleoplasm and in matrix-associated multi-protein complexes known as nuclear bodies. PML is essential for the proper formation and the integrity of the nuclear bodies. Modification of PML by SUMO was shown to be required for its localization in nuclear bodies. The finding that early gene products from several DNA viruses alter or disrupt the PML nuclear bodies  provided much impetus to investigate the effects of viral proteins on SUMO conjugation.
It could thus be shown that the disruption of nuclear bodies by HSV-ICPO, CMV-IE1 and EBV IE protein BZLF1 correlates with abrogation of sumoylation of PML and SP100. In contrast, the adenoviral E4ORF3 product, while also altering but not entirely disrupting the PML nuclear bodies, causes no such effect. Later work extended these findings by showing that some IE viral gene products are themselves substrates for sumoylation. These include CMV-IE1, and -IE2 proteins, BPV (bovine papillomavirus) and HPV E1 proteins and the EBV-BZFL1 protein. The significance of sumoylation in this context remains poorly understood, but for BPV-E1 it appears necessary for normal nuclear localization, whereas for CMV-IE2, sumoylation enhances IE2-mediated transactivation. These results provide additional evidence that SUMO modification plays a critical role in nuclear body dynamics, although it remains to be determined whether viral-induced PML and SP100 de-sumoylation causes nuclear body dispersal or vice versa .
From the first steps of infection to the last moment of release, viruses establish specific alliances within the cell, selectively altering its functional systems, including the UPS. The ability of tumour viruses to transform cells includes redirection of signalling pathways. Since ubiquitin-dependent degradation is the main way to control the lifetime of a plethora of cell signalling regulators, viruses successfully intervene in the ‘normal’ work of the UPS and, as a result, alter the activity of transduction pathways.
The UPS is also centrally positioned within the panoply of immune and inflammatory responses. Ubiquitin-dependent sorting is the source of antigenic peptides that are presented to the immune system, and many viruses have developed escape mechanisms that manipulate the UPS in order to persist in the infected host.
Viral budding is another example whereby the UPS serves the needs of viruses, namely, the process of their release from host cells. Even though there are still many questions, findings in this area may open up approaches for therapeutic antiviral applications.
In general, although there is certainly progress in understanding the relationships between viruses and the UPS, this is just the beginning. Recounting a few puzzles that have recently yielded to investigation only begins to sketch how viruses target host-cell ubiquitin and ubiquitin-like pathways.
The UPS is integral to the cell machinery that today we call post-transcriptional protein–protein covalent modifications. The essential role of such modifications in different aspects of cell signalling directed by viruses is becoming increasingly obvious and will provide rich fields for discovery.
Human tumour viruses dysregulate UPS-dependent protein turnover of many central transcriptional factors, such as p53, pRb and β-catenin.
Viral intervention in ubiquitin-dependent pathways can include modulation of both of the processes that precede the destruction of proteins (ubiquitination and deubiquitination).
Oncogenic viruses can express their own E3s, triggering ubiquitination of host protein targets, or utilize viral products to interrupt host cell E3 ubiquitinating complexes.
Many viruses have developed mechanisms to escape the cellular immune response by inhibiting ubiquitin-dependent antigen presentation of MHC molecules.
RNA viruses use host ubiquitin–proteasome pathways for viral release from host cells.
The relationship between viruses and host ubiquitin-dependent pathways is not limited to exploitation of the UPS, but applies also to ubiquitin-like modifications.
Owing to space restrictions, references to certain works have not been included in the list below. A complete set of references is available from the corresponding author.
- © 2005 The Biochemical Society