key: cord-0035688-a1lb5e15
authors: Prosch, Susanna; Kaspari, Marion
title: Proteasome inhibitors as complementary or alternative antiviral therapeutics
date: 2006
journal: New Concepts of Antiviral Therapy
DOI: 10.1007/978-0-387-31047-3_17
sha: 9120fcd34cceefd75dd5a98487bef979ab88a511
doc_id: 35688
cord_uid: a1lb5e15

nan

Viruses are obligate intracellular parasites and thus widely depend on the cellular machinery for replication and transcription as well as protein and energy metabolism. Options for antiviral therapy are limited to drugs targeting virus-encoded enzymes like DNA-or RNA-polymerases which control replication of the viral genome (e.g. inhibitors of herpesvirus polymerase, of HIV reverse transcriptase or RNA-dependent RNA polymerases), proteases which catalyse processing of viral (poly)proteins (e.g inhibitors of HIV protease) or viral proteins which are involved in uncoating and viral egress (e.g. inhibitors of Influenza virus neuraminidase or M2/hemagglutinin fusion). These drugs are highly specific for certain viruses or virus families and show a satisfying overall in vivo tolerance. However, prophylaxis or even broad therapeutic treatment over a long period of time favours evolvement of drug resistant mutants, particularly when the drug dose or the drug itself is suboptimal and reduces virus replication only marginally. The molecular basis for this phenomenon is a high mutation rate of viral genomes (especially of RNA viruses) combined with a high reproduction rate leading to millions of progeny viruses in one generation. Consequently, drug-mediated modulation of specific cellular pathways essentially involved in virus replication represents an attractive complementary or even alternative strategy to block virus multiplication. Cellular genes have much lower mutation rates so that the problem of drug-resistance should be limited. However, drugs modulating cellular pathways may have a higher toxicity and thus undesirable side effects could limit their in vivo utility.

The ubiquitin-proteasome system (UPS) is the central cellular mediator of regulated protein turnover in living cells. Apart from degradation of misfolded proteins, the proteasome also functions in defined processing of inactive precursor proteins (Ciechanover and Schwartz, 1998) . Furthermore, by catalysing generation of immunogenic peptides for MHC-I presentation, the UPS plays an important role for the MHC-I-restricted cellular immune response against viruses (Kloetzel, 2004) . Moreover, the UPS is essentially involved in regulation of transcription, signalling, cell-cycle, apoptosis and protein sorting which represent crucial processes for viral replication. Therefore, intervention into these elementary cellular processes may have more or less severe consequences for replication of certain viruses. Therefore, it seemed legitimate to investigate whether proteasome inhibitors (PI) are useful complementary or alternative therapeutics for treatment of viral infections for which no or only suboptimal drugs are currently available. A new impetus for these studies was the development of pharmacologically potent PI like PS-341 (Bortezomib) which recently has been introduced in clinical use as an anti-cancer drug in relapsed myeloma patients that are refractory to conventional chemotherapy (Adams, 2002) . The clinical data support that partial and time-limited inhibition of the UPS is a calculable option for therapeutic intervention, even though certain adverse and toxic effects depending on the dose and the individual patient situation might occur (Richardson et al., 2003) .

The following article will give a short overview concerning the structure of the UPS, its function in cellular metabolism and in replication of certain viruses. Furthermore, currently available PI for research and clinical practice as well as their influence on viral replication will be addressed.

The ubiquitin/adenosine triphosphate (ATP)-dependent proteolytic pathway comprises a cascade of enzymes that first catalyse polyubiquitinylation of substrates -ubiquitin-activating enzyme (E1), ubiquitinconjugating enzyme (E2), ubiquitin ligases (E3) and ubiquitin chainassembly factor (E4). The poly-ubiquitinylated proteins are finally recognised by the proteasome for degradation or processing (for details see Fig. 1 , Ciechanova and Schwartz, 1998) . Additionally, the UPS catalyses mono-ubiquitinylation of certain proteins. Mono-ubiquitinylation has been shown to regulate protein function and serves as a signal in protein sorting and membrane receptor recycling but does not target the protein for proteasomal degradation (Fig.1) . Interestingly, processes depending on mono-ubiquitinylation of proteins can be suppressed by PI indicating that they directly or indirectly require proteasome activity (Dupre et al., are two functionally distinct proteasome types -the 20S proteasome and the 26S proteasome. The barrel-shaped 20S proteasome comprises four stacked rings each composed of 7 α or β homodimeric subunits that enclose a central chamber. The two central β-rings harbour three main proteolytic activities, a chymotrypsin-like (CT-L) activity that cleaves after large hydrophobic residues, a trypsin-like (T-L) activity that cleaves after acidic residues and a peptidylglutamyl-peptide hydrolysing (PGPH) or caspase-like activity that also cleaves after acidic residues. There are two further activities that cleave after branched-chain residues or small neutral amino acids. The primary sequence of these proteolytically active β-subunits and their mechanisms of action differ from those of other known cellular proteases. 2001) . present in the cytoplasm and the nucleus of all eukaryotic cells. There Figure 1 : The ubiquitin-proteasome system (UPS). The ubiquitin-proteasome system involves a cascade of enzymes which catalyse mono-or poly-ubiquitinylation of substrates starting with ubiquitin-activating enzyme (E1), followed by ubiquitin-conjugating enzymes (E2) and ubiquitin ligases (E3) which finally mediate covalent conjugation of ubiquitin to the target protein. Repeated cycles of ubiquitin conjugation result in generation of insoluble polyubiquitinylated proteins which are then recognised by the proteasome as mis-folded and damaged proteins and are degraded or processed. Eukaryotic cells contain one E1, several E2 and more than 100 different E3 enzymes. The E3 enzymes are suggested to determine the specificity of ubiquitinylation. Mono-ubiquitinylation of proteins is associated with changes in protein function and plays a role in protein transport and sorting, as well as transcription regulation.

Proteasomes are highly conserved multimeric peptidase complexes that are The β -subunits are flanked by two α-rings that guide the entrance of unfolded polypeptide chains into the 20S core (Arendt and Hochstrasser, 1997; Baumeister et al., 1998; Bochtler et al., 1999) .

The 26S proteasome is composed of the 20S core particle and two additional 19S regulatory complexes that associate with the outer α-rings and contain the substrate recognition and binding sites. The 19S complex harbours six ATPases and catalyses de-polymerisation of poly-ubiquitin chains as well as unfolding of the protein substrate thus facilitating its entry into the 20S proteasome (Braun et al., 1999; Ferrell et al., 2000) . A minimum of four ubiquitin molecules is required for recognition by the 26S proteasome. The ubiquitin chains are then removed from the substrate by de-ubiquitinylating enzymes and re-enter the pool of free ubiquitin.

The 20S proteasome can also associate with the proteasome activator 28/11S regulator (PA28) in an ATP-independent manner thus allowing degradation of non-ubiquitinylated proteins (Baumeister et al., 1998; Rechsteiner et al., 2000; Hoffman et al., 1992) . PA28 activates the 20S proteasome by induction of conformational changes which presumably open the core maximally and thus facilitate substrate entry and cleavage product exit without affecting the active sites of the proteases (Stohwasser et al., 2000; Hill et al., 2002) . In vivo, the 20S proteasome also forms so-called hybrid proteasomes with one 19S and one PA28 regulatory subunit. However, these hybrid proteasomes are as active as the 19S-20S-19S

In the presence of interferon (IFN)-γ (or other cytokines like TNFα and IFN-β) the catalytic β-subunits of the proteasome are replaced by the subunits LMP-2, LMP-7 and MECL-1 thus generating the immunoproteasome. The immuno-proteasome plays a central role for the proteolytic generation of antigenic peptides for presentation via MHC class I molecules (Kloetzel, 2004) . These IFN-inducible subunits modify the cleavage specificity profile of the proteasome by altered cleavage site preference and enhanced generation of specific peptides.

Originally, the proteasome was discovered due to its function in degradation of aged or misfolded proteins into small peptides of 3-20 amino acid residues. However, the proteasome also controls specific degradation of regulatory proteins and processing of inactive precursor proteins.

The UPS is involved in regulation of key cellular processes like cellcycle progression, transcription, translation, signal transduction, stress response, apoptosis, receptor function and protein sorting by catalysing degradation, processing or translocation of several proteins involved in regulation of these processes, e.g. cyclins and cyclin-dependent kinases, p53, c myc, bcl-2, Bax, Nuclear factor (NF)-κB precursor p100 and p105, I-κB (Inhibitor of NF-κB), JNK and c-jun (Rolfe et al., 1997; Salghetti et al., 1999; Breitschopf et al., 1999; Strous and Govers, 1999; Longva et al., 2002; Baldwin, 1996; Karin and Ben-Neriah, 2000; Dupre et al., 2001; Adams, 2002; Stahl and Barbieri, 2002; Beinke and Ley, 2004) . Additionally, the proteasome plays an important role in the cellular immune

proteasome (Kopp et al 2001; Tanahashi et al., 2000) .

., response by generating peptides for presentation via MHC class I molecules (Kloetzel, 2004) . The most intensively studied proteasome substrate is the transcription factor NF-κB which is crucial for transcription regulation of genes involved in pro-inflammatory processes (e.g. TNF-α, IL-1 and -6, MIP-1α, RANTES, E-selectin and VCAM-1), immune response (e.g. T-cell activation by upregulation of MHC molecules and CD80/86 on antigen presenting cells), anti-apoptosis (e.g. c.IAP-1/2, AI, Bcl-2 and BCl-XL) and cell cycle progression (cell-cycle regulator cyclin D1) (Zhang and Ghosh, 2001, Li and Verma, 2002; Karin and Li, 2002; . Inhibition of NF-κB activation therefore represents the main mechanism by which PI induce apoptosis and block cell cycle progression or inhibit inflammatory processes.

The most dominant form of NF-κB is a heterodimer consisting of p50 and p65. The p50/p65 heterodimer remains in an inactive state by association with its inhibitors IκBα or IkBβ in the cytoplasm. Various stimuli are able to trigger activation of NF-κB by induction of proteasomedependent IκB degradation following its phosphorylation by the IKK complex and ubiquitinylation by the SCF (Skp1/Cu11/F-box) E3 ubiquitin ligase complex (Baldwin, 1996; Karin and Ben-Neriah, 2000) . The active p50/p65 heterodimer is then translocated into the nucleus where it binds to the promoters of responsive genes. Apart from IκB degradation, the 26S proteasome also controls processing of the NF-κB p105 and p100 precursors to generate the NF-κB p52 and p50 subunits. Furthermore, p50 can be retained in the cytoplasm by binding to its precursor p105. Following stimulation of the cells p105 can be completely degraded by the UPS allowing translocation of p50 into the nucleus (Beinke and Ley, 2004 ).

In the last few years, various low-molecular-weight compounds have been identified that more or less selectively inhibit the UPS through interaction with the active site subunits of the 20S proteasome. These compounds are either synthetic products or naturally occurring substances; the most relevant compounds are listed in Tab. 1. Synthetic PI consist of linear di-or tripeptides linked to different pharmacophores such as benzamides, αketoamides, aldehydes, α-ketoaldehydes, vinyl sulfones or boronic acids (Fig. 2) . Natural product PI display a broad structural heterogeneity containing both linear as well as non-linear scaffold structures (Fig. 3) .

In the early 1980s, serine/cysteine protease and calpain inhibitors belonging to the family of peptide aldehydes were initially discovered to interfere with T-L and CT-L activity of the 20S proteasome, respectively (Wilk et al., 1983; Figueiredo-Pereira et al., 1994) . Today, a variety of peptide aldehydes with higher potency and increased selectivity towards the 20S proteasome like MG132, MG115 and PSI have been developed. Binding of peptide aldehydes to the active site β-subunits results in formation of a reversible hemiacetal adduct between the aldehyde group and the hydroxyl group of the N-terminal threonine residue. Due to the highly reactive functional aldehyde group, cross-reactivity with cellular proteases remains a major drawback of this class of inhibitors that limits their in vivo utility. As a strategy to overcome these limitations, boronic acid derivatives of peptide aldehydes with improved selectivity and higher affinity were developed (Adams et al., 1998; Gardner et al., 2000) . Studies on their mode of action revealed that the empty p-orbital of boron accepts the oxygen lone pair of the N-terminal threonine residue to form a stable tetrahedral intermediate. Boronic acid-based inhibitors can be truncated to dipeptides, thus improving important therapeutic characteristics like solubility and membrane permeability. Apart from that, slower dissociation rates of boronic acid peptides lead to a prolonged inhibition of the proteasome. The most common representatives of this class in the context of in vivo studies and clinical applications are MG262 and PS-341 (see below).

Peptide vinyl sulfones irreversibly modify the hydroxyl group of the active site N-terminal threonine (Bogyo et al., 1997) . However, since they inhibit both the 20S proteasome and cysteine proteases, these compounds are not regarded as suitable for in vivo applications.

The first naturally occurring proteasome inhibitor identified was the Streptomyces lactacystinaeus metabolite lactacystin (Omura et al., 1991) . In aqueous solutions, lactacystin is spontaneously converted into clastolactacystin β-lactone that reacts with the hydroxyl group of the β-subunit N-terminal threonine to form a covalent ester adduct (Fenteany et al., 1995) . Thus, lactacystin irreversibly inhibits the CT-L and T-L activity of the proteasome. Although it was originally believed to exclusively target the 20S proteasome, there is increasing experimental evidence that lactacystin also inhibits cathepsin A (Ostrowska et al., 2000) . With the aim to reduce nonspecific interactions, a synthetic analogue of lactacystin termed PS-519 was designed that has already demonstrated effectiveness against inflammatory diseases in clinical studies (Elliott et al., 1999; Kondagunta et al., 2004) .

Screening for anti-tumour agents in mice led to the discovery of peptide α', β'-epoxyketones, another class of natural product PI (Hanada et al., 1992) . Epoxomicin and eponemycin, the most common representatives of this class, were isolated from Actinomycetes strain No. Q996-17 and Streptomyces hygroscopicus, respectively. They inhibit the 20S proteasome by forming a morpholino ring complex with the active site threonine residue. Cross-reactions with other cellular targets have not been reported for this class of inhibitors. However, due to the inhibition of all catalytic activities of the 20S proteasome, α', β' epoxyketones might be highly cell toxic and thus of limited therapeutic potential.

A family of non-linear PI was isolated from the fermentation broth of Apiospora montagnei Sacc. TC 1093 (Koguchi et al., 2000) . Compounds of this class like the TMC-95 stereoisomers are characterised by a macrocyclic scaffold structure and bind non covalently to the proteasome.

Another non-linear PI class termed epipolythiopiperazin toxins was discovered in Aspergillus fumigatus. Representatives of this class such as gliotoxin consist of a heterobicyclic core containing bisulfide bridges. Proteasome inhibition by gliotoxin can be reversed by dithiothreitol (Kroll et al., 1999) .

Generally, PI target the catalytic activities of the proteasome, thus preventing degradation and cleavage of proteins. Consequently, polyubiquitinylated proteins accumulate in the cell, precursor proteins remain in their inactive configuration and several proteins are mis-sorted. Signalling pathways are blocked (e.g. the NF-κB pathway, see above) or constitutively activated (e.g. the JNK/c-jun pathway, Nakayama et al., 2001) causing either a decreased or increased expression of responsive genes. The most extensively studied inhibitors are PS-341 (Bortezomib, Velcade) and PS-519. Both compounds have been tested as potential therapeutics for oncological and inflammatory conditions. Bortezomib has demonstrated activity in phase II trials as a treatment option for renal cell cancer, lung cancer, sarcoma and lymphoma (Richardson et al., 2003; Davis et al., 2004; Kondagunta et al., 2004; Maki et al., 2005) . Since 2003, bortezomib has been approved for clinical use in patients with multiple myeloma. PS-519 is the most advanced inhibitor in the context of inflammatory diseases like psoriasis, rheumatoid arthritis, asthma, multiple sclerosis, sepsis and vascular restenosis and has entered phase II clinical studies for the indications of acute stroke and myocardial infarction (Phillips et al., 2000; Palombella et al., 1998; Elliott et al., 1999; Kondagunta et al., 2004) .

The anti-cancer effect of bortezomib and other PI is mainly based on their ability to stabilise cell cycle inhibitory proteins. Inhibition of NF-κB transcriptional activity furthermore down-regulates expression of various growth, survival and angiogenic factors (Sunwoo et al., 2001) . The ability of PI to increase tumour suppressor p53 levels and to activate c-Jun N-terminal kinase (JNK) leading to Fas up-regulation and caspase activation also contributes to the anti-neoplastic, pro-apoptotic effect (Hideshima et al., 2001; Mitsiades et al., 2002; Drexler, 1997 , Shinohara et al., 1996 . Additionally, it has also been shown that proteasome inhibition enhances the sensitivity of cancer cells for traditional anti-cancer agents (Bold et al., 2001) . The anti-inflammatory effect of PS-519 and other PIs is probably based on inhibition of NF-κB activation that plays an important role in the pathogenesis of many inflammatory diseases by controlling expression of cytokines, adhesion molecules and pro-inflammatory molecules.

Given the fundamental metabolic role of the proteasome for all eukaryotic cells, undesirable adverse effects of proteasome inhibition might occur. However, most somatic cells are in a quiescent state and are thus less susceptible for proteasome inhibitors than actively dividing cells like cancer or immune cells. In accordance with this, bortezomib displayed a satisfying overall tolerance in clinical testing. The reported side-effects include fatigue and/or low fever, thrombocytopenia, gastrointestinal symptoms, electrolyte disturbances, headache, anemia, arthralgia, low diarrhea, frequent skin rashes and peripheral neuropathy (Richardson et al., 2003; Davis et al., 2004) . Autoradiography experiments in rats showed that bortezomib is present in all organs except CNS, spinal cord, testes and eyes thus sparing these organs adverse effects of proteasome inhibition. Apart from that, proteasome inhibition by bortezomib is reversible and it could be shown that the inhibitor is rapidly metabolised (Adams, 2002) . Similarly, PS-519 showed a good bioavailability in all organs except brain tissue. Side-effects have not been reported yet.

Interactions between viruses and the UPS are highly complex and variable. Certain viruses essentially require the proteasome for different steps of their replication cycle. For other viruses, however, the proteasome represents a barrier and a degradative factor. In vitro studies also revealed that the interaction between the UPS and the virus presumably depends on the cell type. Most interestingly, the UPS may play different and even opposing roles during different stages of virus replication. Consequently, virus replication is either PI-sensitive or -insensitive and proteasome inhibition either suppresses or even supports replication of the virus depending on the specific situation. So far, antiviral effects of PI have been demonstrated in vitro for members of seven virus families -retroviridae, herpesviridae, parvoviridae, picornaviridae, orthomyxoviridae, paramyxoviridae and coronaviridae. For other viruses it has been shown that certain steps in virus replication are controlled by the UPS, however, an antiviral effect of PI has not been demonstrated yet. Involvement of the UPS in the interaction between the virus and cellular metabolism does not necessarily imply that proteasome inhibition blocks virus replication as shown for Adenovirus transcription/translation which depend on the cellular gene expression machinery. Data arise exclusively from in vitro experiments using virusinfected cell cultures. However, preliminary data from animal experiments performed in our laboratory raise the hope that PI do not only suppress virus replication in vitro but also in vivo (Gille, Prösch., unpublished observation, see below).

Several studies indicate that proteasome inhibition may block entry of viruses which essentially involve the endocytotic pathway and in particular the late endosome, as has been demonstrated for Influenza virus and a strain of Coronaviruses (Khor et al., 2003; Sieczkarski and Whittaker, 2003; Yu and Lai, 2005) . This observation is in accordance with the finding that sorting of proteins into the late endosome via "inwards" budding of membranes is coupled with the protein trafficking and ubiquitin-vacuolar sorting system and requires the functional UPS (Stahl and Barbieri, 2002; Strous and Govers, 1999; Longva et al., 2002) .

Productive infection by influenza viruses essentially depends on late endosomes and is thus highly sensitive to proteasome inhibition. Treatment of cells with PI caused a mis-targeting of the virus into so-called sorting and recycling endosomes which are different from early and late endosomes. Accordingly, the virus did not enter into the late endosomes and into the nucleus where replication occurs (Khor et al., 2003; Sieczkarski and 

of PI target entry/egress of the virus or interfere with viral (Corbin-Lickfett and Bridge, 2003) . The so far described antiviral effects Yu and Lai, 2005) . Since PI do not influence the entry of Vesicular Stomatitis Virus (VSV) and Semliki Forest Virus (SFV) which only require the early and not the late endosome, the target for PI might be localised beyond the early endosome. The exact mechanism of PI-dependent inhibition of influenza virus entry, however, remains unclear up to date. A cellular target for proteasomal degradation seems more likely because viral proteins have not been found to be ubiquitinylated. Similar to influenza virus, PI also suppress entry and nuclear translocation of Mouse Hepatitis Virus (MHV) strain JHM, a murine Coronavirus, into mouse astrocytoma cells (DBT cell line) (Ros and Kempf, 2004) . Treatment of MHV strain JHM infected cells with PI caused a strong accumulation of incoming virus in the early endosome and led to a missorting into the lysosome. MHV may enter its target cells either by plasma membrane fusion or by endocytosis depending on the virus strain and the cell type (Nash and Buchmeier, 1997) . Correspondingly, replication of MHV strain JHM which enters DBT cells via the endocytotic pathway was highly sensitive to proteasome inhibition, while MHV strain A59, which enters the cells by fusion, was less efficiently inhibited by PI.

Interestingly, the UPS has also been found to be involved in the pH dependent but endosome-independent entry and nuclear translocation of certain parvoviridae. Replication of at least two members of this virus family, the Minute Virus of Mice (MVM) and the Canine Parvovirus (CPV), but not the distantly related Bovine Parvovirus (BPV), were inhibited by PI. In the presence of PI virus particles aggregated in ring-like structures around the nuclei but were unable to enter into the nucleus (Ros et al., 2002, Ros and Kempf, 2004) . The underlying mechanism remains unclear because the proteasome does not seem to be involved in cleavage/processing of capsid protein VP and the virus did not co-localise with the early or late endosomal or the lysosomal compartment. Furthermore, uncoating of parvoviruses occurs in the nuclear core or in the nucleus but not in the cytosol.

Generally, virus budding at the outer plasma membrane follows the same strategy of "inwards" budding of membranes into so-called multivesicular bodies and association with the protein trafficking and ubiquitin-vacuolar sorting system. Accordingly, budding of certain retroviruses including Human Immunodeficiency Virus (HIV) type 1 and 2, Rous Sarkoma Virus (RSV), Murine Leukemia Virus (MuLV) and Mazon Pfizer Monkey Virus (MPMV) as well as the rhabdovirus VSV and the paramyxovirus SV5 was strongly inhibited by PI (Patniak et al., 2000; Schubert, et al., 2000; Strack et al., 2000; Harty et al., 2001; Ott et al., 2003, recently reviewed in Klinger and Schubert, 2005; Schmitt et al., 2005) . Intensive studies on HIV-1 revealed in the absence of a functional UPS retarded budding structures, which accumulated and the released HIV-1 particles were not infectious. It was also shown that processing of the Gag-polyprotein but not the envprecursor, was strongly impaired. As a consequence Gag processing intermediates accumulated in the cell and the intracellular and virusassociated levels of the capsid protein p24 were strongly reduced. Interestingly, the viral protease which normally cleaves the Gag-polyprotein was not targeted by PI Klinger and Schubert, 2005) . Despite intensive investigations, the mechanism of PI-dependent inhibition of HIV budding is not completely understood. Several studies demonstrated that the C-terminal portion of Gag proteins of different retroviruses (e.g. p6 of HIV-1 and SIV Gag, p12 of Mo-MLV Gag, p9 of EIAV Gag) is monoubiquitinylated. The ubiquitin residue was suggested to play a role in function of the L-domains that are present in these proteins and are required for budding (Ott et al., 1998 (Ott et al., , 2002a (Ott et al., and b and 2003 Pornillos et al., 2002) . Since treatment with PI causes an accumulation of polyubiquitinylated proteins, limititation of the endogenous pool of free ubiquitin may prevent Gag ubiquitinylation. However, HIV-1 p6 Gag mutants deficient for ubiquitinylation showed an unchanged phenotype, indicating that ubiquitinylation alone is not sufficient for virus budding . Moreover, mono-ubiquitinylation of Gag alone does not provide an explanation why budding of some retroviruses like Equine Infectious Anemia Virus (EIAV) and Mouse Mamma Tumour Virus (MMTV) are insensitive to PI (Patniak et al., 2002; Ott et al., 2003) . The structure of the Gag L-domains and their interaction with different components of the UPSdependent vacuolar protein-sorting system seems to be more crucial for virus budding. Only viruses with proteins containing an L-domain of the PPXYtype (e.g. MuLV, RSV, Ebola virus) or the PTAP type (e.g. HIV-1 and-2) but not of the YXDL-type (e.g. EIAV) are sensitive to PI (Shehu-Xhilaga et al., 2004; Klinger and Schubert, 2005) . By binding to different "bridging molecules", L-domains mediate interaction between the viruses and the vacuolar protein-sorting system required for budding. L-domains of the PPXY-type are able to bind different HECT ubiquitin ligases like WWP1, WWP2 or Itch, while L-domains of the PTAP-motif bind Tsg 101 (e.g. HIV, Ebola virus VP40) which then recruits other components of the endosomal sorting system like ESCRT-I,-II,-III and thus connects the Gag protein to the vacuolar protein sorting system (Garrus et al., 2001; Amit et al., 2004; Martin-Serrano et al., 2001 .

Since VP40 of Ebola Virus and Human T-cell Lymphotropic virus (HTLV-1) Gag also contain L-domain motifs of the PPXY type it has been speculated that Ebola virus and HTLV-1 are also sensitive to proteasome inhibition. However, this still has to be proved in an infection model (Harty et al., 2000; Bouamr et al., 2003; Heidecker et al., 2004; Martin-Serrano et al. 2005) .

In the paramyxovirus SV5 matrix protein, very recently, a new type of Ldomains (FPIV) has been idenitfied which may functionally substitute the PTAP motif in HIV-1 Gag and cause the PI-sensitivity of this virus (Schmitt et al., 2005) . The precise role of ubiquitinylation of either viral or cellular targets or of the ubiquitin ligases in virus budding remains unclear, however. Most interestingly, HIV protease inhibitors like nelfinavir and saquinavir display proteasome inhibitory activity (Piccinini et al., 2002 and . This unexpectected activity may be responsible for some side effects of these drugs, whether they contribute to the anti-HIV activity, e.g. by inhibiting NF-κB activation which is involved in positive regulation of HIV gene expression or by inhibiting the budding -is unclear.

An alternative model to explain the PI-sensitivity of HIV suggests that PI facilitate the production of Gag-related defective ribosomal products (DriPs) which are encapsidated together with the normal Gag polyprotein and which may act as competitive non-cleavable substrate inhibitors of the viral protease during Gag-polyprotein processing . This hypothesis, however, contradicts the finding that PI are selective only for certain retroviruses.

The UPS is not only involved in entry and budding of viruses but also in regulation of viral gene expression. Very recently, this effect was demonstrated for Coxsackievirus B3, the most prevalent virus associated with the pathogenesis of myocarditis and myocardiopathy (McManus et al., 1991; Martino et al., 1994) . In vitro treatment of virus-infected murine cardiomyocytes with different PIs caused a strong reduction in virus replication. Inhibition occurred at the level of viral RNA transcription and protein synthesis but did not affect virus entry or proteolytic activity of the virus-encoded protease (Luo et al., 2003) . However, the mechanism by which PI inhibit coxsackievirus transcription/translation remains unclear so far.

As a further example, Human Cytomegalovirus (HCMV) as well as Rat Cytomegalovirus (RCMV) in vitro-replication in fibroblasts is efficiently suppressed by different reversible as well as irreversible PI (Prösch et al., , 2003; Kaspari, Prösch unpublished data) . Inhibition of HCMV replication occurs at the immediate early (IE) stage of infection by suppression of the IE gene expression. Virus adsorption as well as entry and trafficking of the capsid/tegument complex into the nucleus are not influenced. Several lines of evidence support the hypothesis that PI specifically block generation of the IE2 messenger while generation of the IE1 messenger is not affected. Both IE1 and IE2 messengers are generated by different splicing of the IE locus, suggesting that the proteasome may be involved in regulation of IE2 mRNA splicing. In accordance with this it was found that expression and localisation of several splice factors are altered in the absence of a functional UPS (Rockel and Mikecz, 2002) . Since the IE2 protein is essential for induction of early and late viral proteins and for virus-dependent cellular gene expression, PI are detrimental to HCMV replication. The structure of the IE gene locus of RCMV is very similar to that of HCMV regarding its modular composition and splicing processes, thus, it might be assumed that PI inhibit replication of RCMV and HCMV by the same mechanism.

Remarkably, in a first experiment in which we treated RCMV-infected immunosuppressed rats with MG262, the viral load in all tested organs was clearly reduced (Gille, Prösch, unpublished data). These results raise the hope that the in vitro observed anti-CMV effect of PI is also present in vivo and that PI represent an alternative treatment option against GCV-resistant strains or a complementary drug to ganciclovir/valganciclovir which block virus replication only at the late stage (Prösch et al., 2003) .

A third example for involvement of the proteasome in regulation of viral gene expression are Adenoviruses (Corbin-Lickfett and Bridge, 2003) . Adenoviruses require the proteasome for initiation of late gene expression by the E4-34kDa protein, subsequently, PI block expression of late genes. However, in vitro replication of Adenovirus was only marginally reduced in the presence of PI, indicating that the UPS function is not essentially required for efficient replication.

Furthermore, transfection experiments suggest that the UPS influences translation of hepatitis C virus, however, it has not been proved whether PI inhibit HCV replication (Kruger et al., 2001) . 

The use of PI as antiviral drugs may be contraindicated by the fact that under certain circumstances some viruses may benefit from proteasome inhibition. One example is the Herpes Simplex Virus (HSV) type 1. Establishment and maintenance of latency in neuronal cells has been found to depend on the proteasomal degradation of the viral UL9 protein which may facilitate reactivation of the virus by binding to and unwinding of the origin 3. UL9 interacts with the neurone specific NFB24 protein and is thus recruited to an E3 ubiquitin ligase complex facilitating its ubiquitinylation and proteasomal degradation. Proteasome inhibition allows accumulation of UL9 and thus induce viral replication/reactivation (Eom and Lehmann, 2003) . Similarly, reactivation of latent integrated HIV may benefit from proteasome inhibition (Krishnan and Zeichner, 2004; Schwartz et al., 1998) . As a reason it has been speculated that key factors required to effectively activate the HIV-l LTR are degraded or repressed in latently infected cells and that their stabilisation by PI then promotes transcription. Accordingly, latently infected cells show increased expression of various proteasome subunits and inhibition of the UPS may restore this process.

Additionally, for some HIV strains and the Adeno-associated Virus (AAV), a parvovirus, it has been shown that the proteasome may limit their infectivity. Consequently, PI increased their infectivity by preventing viral degradation during entry. This effect is especially relevant for HIV strains entering the cell by fusion, while strains using the endocytotic pathway are less successible (Wei et al., 2005; Schwartz et al., 2001) . One explanation might be that the proteasome degrades Gag proteins within the cytoplasm. Similarly, PI support transduction of the recombinant AAV (rAAV). Normally, in less efficiently transduced cells incoming virus is effectively degraded by the proteasome (Douar et al., 2001; Hansen et al., 2000; Duan et al., 2000) . Inhibition of the proteasome enhances nuclear accumulation of the virus and viral gene expression (Duan et al., 2000; Jennings et al., 2004) . Since rAAV is of great interest as a vector for gene therapy, proteasome inhibition has demonstrated to promote transgene expression (Yan et al., 2002) .

During the last years it became more clearly that the UPS is essentially involved in different steps of virus replication. For some viruses, inhibition of the UPS by PI has been shown fatal at least under certain conditions. This makes PI highly interesting candidates for alternative or complementary treatment of infections with these viruses. All studies so far have been carried out with PI developed as research tools only which are not suitable for clinical use. To evaluate PI as antiviral therapeutics, these experiments have to be confirmed with clinically approved PI like PS-341 and PS-519 and have to be verified in animal settings. Our first in vivo data raise the hope that PI display their antiviral activity not only in vitro but also in vivo.

The therapeutic index of PI as observed in our lab for HCMV and RCMV is low compared to classical antiviral drugs which target virus-specific genes. This emphasises the need to develop PI with increased specificity and reduced toxicity. Alternatively, drugs which target the more specific E3 ubiquitin ligases involved in virus UPS interaction should be more specific and less toxic. Development of such compounds, however, requires the knowledge of viral and cellular factors involved in the interaction between virus and UPS. A further question which has to be addressed is the influence of PI on (re)activation of highly prevalent latent or persistent viruses to exclude that treatment of an acute infection with one virus does not trigger reactivation of another virus. This question is also of substantial interest for the use of PI in anti-cancer therapy because persistent viruses like EBV or HCMV are co-factors for tumour genesis.

Reduced resistance should be one advantage of PI compared to classical antiviral drugs, however, it can not be guaranteed as already shown for other drugs targeting cellular factors. Apart from that, PI are an ideal alternative option for the treatment of viruses resistant to classical antiviral drugs (e.g. GCV-resistant HCMV, Prösch et al., 2003) since they target a cellular factor thus excluding or minimising cross-resistance.

In summary, the present data raise the hope that inhibitors of the UPS support successful treatment or even prevention of severe viral infections as caused e.g. by HIV-1, HCMV or Coxsackievirus because PI are detrimental to replication of these viruses.

Development of the proteasome inhibitor PS-341

Potent and selective inhibitors of the proteasome: dipeptidyl boronic acids

Identification of the yeast 20S proteasome catalytic centers and subunit interactions required for active-site formation

The NF-κB and IκB proteins: new discoveries and insights

The proteasome: paradigm of a self-compartmentalizing protease

Function of NF-κB1 and NF-κB2 in immune cell biology

The proteasome

Covalent modification of the active site threonine of proteasomal beta subunits and the Escherichia coli homolog HslV by a new class of inhibitors

Chemosensitization of pancreatic cancer by inhibition of the 26S proteasome

PPPYEPTAP motif is the late domain of human T-cell leukemia virus type 1 Gag and mediates its functional interaction with cellular proteins Nedd4 and Tsg101

The base of the proteasome regulatory particle exhibits chaperone-like activity

Dephosphorylation targets Bcl-2 for ubiquitin-dependent degradation: a link between the apoptosome and the proteasome pathway

The ubiquitin-proteasome pathway: the complexity and myriad functions of protein death

Adenovirus E4-34kDA requires active

Phase II trial of PS-341 in patients with renal cell cancer: a University of Chicago Phase II consortium study

Intracellular trafficking of adeno-associated virus vectors. Routing to the late endosomal compartment and proteasome degradation

Activation of the cell death program by inhibition of proteasome function

proteasomes to promote late gene expression

Endosomal processing limits gene transfer to polarized airway epithelia by adeno-associated virus

Membrane transport: ubiquitinylation in endosomal sorting

Proteasome inhibition: a novel mechanism to combat asthma

Replication-initiator protein (UL9) of the herpes simplex virus 1 binds NFB42 and is degraded via the ubiquitin-proteasome pathway

Inhibition of proteasome activities and subunit-specific amino-terminal threonine modification by lactacystin

Regulatory subunit interactions of the 26S proteasome, a complex problem

Comparison of the effect of calpain inhibitors on two extralysosomal proteinases: the multicatalytic proteinase complex and m-calpain

Characterisation of peptidyl boronic acid inhibitors of mammalian 20S and 26S proteasomes and their inhibition of proteasomes in cultured cells

Tsg101 and the vacuolar protein sorting pathway are essential for HIV-1 budding

Epoxomicin, a new antitumor agent of microbial origin

Impaired intracellular trafficking of adeno-associated virus vectors limits efficient transduction of murine fibroblasts

A PPxY motif within the VP40 protein of Ebola virus interacts physically and functionally with a ubiquitin ligase: implication for filovirus budding

Rhabdoviruses and the cellular ubiquitin-proteasome system: a budding interaction

Late assembly motifs of human T-cell leukemia virus Type 1 and their relative roles in particle release

The proteasome inhibitor PS-341 inhibits growth, induces apoptosis and overcomes drug resistance in human multiple myeloma cells

The 11S regulators of 20S proteasome activity

Multiple forms of the 20 S multicatalytic and the 26 S ubiquitin/ATP-dependent proteases from rabbit reticulocyte lysate

Proteasome inhibition enhances AAVmediated transgene expression in human synoviocytes in vitroo and in vivo

Phosphorylation meets ubiquitinylation: the control of NF-κkappa B activity

NF-B in cancer: from innocent bystander to major aelprill

NF-κB at the crossroads of life and death

The ubiquitin-vacuolar protein sorting system is selectively required during entry of influenza virus into host cells

The ubiquitin-proteasome system in HIV replication: potential targets for antiretroviral therapy

The proteasome and MHC class I antigen processing

novel proteasome inhibitors produced by Apiospora montagnei Sacc. TC 1093. Taxonomy, production, isolation and biological activities

Phase II trial of bortezomib for patients with advanced renal cell carcinoma

Reconstitution of hybrid proteasomes from purified PA700-20S complexes and PA28 alpha beta activator: ultrastructure and peptidase activities

Host cell expression during human immunodeficiency virus type 1 latency and reactivation and effects of targeting genes that are differentially expressed in viral latency

Involvement of proteasome α-subunit PSMA7 in hepatitis C virus internal ribosome entry site-mediated translation

The secondary fungal metabolite gliotoxin targets proteolytic activities of the proteasome

NF-B regulation in the immune system

Ubiquitinylation and proteasomal activity is required for transport of the EGF receptor to inner membranes of multivesicular bodies

Proteasome inhibition reduces coxsackievirus B3 replication in murine cardiomyocytes

Progress and challenges in the pathological diagnosis of myocarditis

A multicenter Phase II study of bortezomib in recurrent or metastatic sarcomas

Viral infection and the pathogenesis of dilated cardiomyopathy

HECT ubiquitin ligases link viral and cellular PPXY motifs to the protein-sorting pathway

Role of ESCRT-I in retroviral budding

HIV-1 and Ebola virus encode small peptide motifs that recruit Tsg101 to sites of particle assembly to facilitate egress

Molecular sequelae of proteasome inhibition in human multiple myeloma cells

Unexpected transcriptional induction of monocyte chemoattractant protein 1 by proteasome inhibition: involvement of the c-Jun N-terminal kinase-activator protein 1 pathway

Entry of mouse hepatitis virus into cells by endosomal and nonendosomal pathways

Lactacystin, a novel microbial metabolite, induces neuritogenesis of neuroblastoma cells

Separation of cathepsin Alike enzyme and the proteasome: evidence that lactacystin / beta-lactone is not a specific inhibitor of the proteasome

Equine infectious anemia virus and the ubiquitin-proteasome system

Ubiquitin is covalently attached to the p6 Gag proteins of human immunodeficiency virus type 1 and simian immunodeficiency virus and the p12Gag protein of Moloney murine leukemia virus

Retroviruses have differing requirements for proteasome function in the budding process

Role of the proteasome in and NF-kappaB in streptococcal cell wall-induced polyarthritis

Potential roles of cellular proteins in HIV-1

Budding of equine infectious anemia virus is insensitive to proteasome inhibitors

Proteasome inhibitor PS-519 reduces infarction and attenuates leukocyte infiltration in a rat model of focal cerebral ischemia

The HIV protease inhibitors nelfinavir and saquinavir, but not a variety of HIV reverse transcriptase inhibitors, adversely affect human proteasome function

The human 26S proteasome is a target of antiretroviral agents

Structure of the Tsg101 UEV domain in complex with the PTAP motif of the HIV-1 p6 protein

Proteasome inhibition: a novel tool to suppress human cytomegalovirus replication and virus-induced immune modulation

The proteasome activator 11S REG (PA28) and class I antigen presentation

A Phase 2 study of bortezomib in relapsed, refractory myeloma

Proteasome-dependent processing of nuclear proteins is correlated with their subnuclear localization

The ubiquitin-mediated proteolytic pathway as a therapeutic area

The ubiquitin-proteasome machinery is essential for nuclear translocation of incoming minute virus of mice

Cytoplasmic trafficking of minute virus of mice: low pH requirement, routing to late endosomes, and proteasome interaction

Destruction of Myc by ubiquitinmediated proteolysis: cancer-associated and transforming mutations stabilize Myc

Evidence for a new viral late-domain core sequence, FPIV, necessary for budding of a paramyxovirus

Proteasome inhibition interferes with Gag polyprotein processing, release, and maturation of HIV-1 and HIV-2

Antiviral activity of the proteasome on incoming human immunodeficiency virus type 1

Late domain-dependent inhibition of equine infectious anemia virus budding

Apoptosis induction resulting from proteasome inhibition

Differential requirement of Rab5 and Rab7 for endocytosis of influenza and other enveloped viruses

Multivesicular bodies and multivesicular endosomes: the "in and outs

Kinetic evidences for facilitation of peptide channelling by the proteasome activator PA28

A role for ubiquitin ligase recruitment in retrovirus release

The ubiquitin-proteasome system and endocytosis

Novel proteasome inhibitor PS-341 inhibits activation of nuclear factor-kappaB, cell survival, tumor growth and angiogenesis in squamous cell carcinoma

Hybrid proteasomes: induction by interferon-γ and contribution to ATP-dependent proteolysis

Inhibition of lysosomal and proteasome function enhances human immunodeficiency virus type 1 infection

Evidence that pituitary cation-sensitive neutral endopeptidase is a multicatalytic protease complex

Ubiquitination of both adeno-associated virus type 2 and 5 capsid proteins affects the transduction efficiency of recombinant vectors

The ubiquitin-proteasome system facilitates the transfer of murine coronavirus from endosome to cytoplasm during virus entry

Toll-like receptor-mediated NF-κB activation: a phylogenetically conserved paradigm inEdwards, L., 1990, PhD Thesis Cancer in Rats

The authors would like to thank Min-Hi Lee for her support in preparing the Figures and M. Raftery for critical reading the paper.