OVERVIEW OF VV LIFE CYCLE

VV produces two forms of infectious progeny, intracellular mature virus (IMV) and extracellular enveloped virus (EEV). IMV represents the majority of infectious progeny and remains inside the infected cell until cell lysis (Appleyard, Hapel & Boulter, 1971; Ichihashi, Matsumoto & Dales, 1971). Although EEV is only a small proportion of the progeny virus, it is important biologically. It is released into the extracellular space before cell death (Appleyard et al. 1971). The EEV contains an additional lipid membrane, which contains cellular proteins and at least 5 VV-encoded polypeptides that are absent from IMV (Smith & Vanderplasschen, 1998; Vanderplasschen, Hollinshead & Smith, 1998; Smith, Vanderplasschen & Law, 2002).

Due to their large size (approximately 350×270 nm) VV virions are discernible by light microscopy in infected cells and were the first viruses to be observed microscopically. In traditional electron microscopy analyses of infected cells, the virions appear brick-shaped and consist of an internal electron-dense, biconcave core (containing the virus genome, major structural proteins and virion enzymes and two lateral bodies on each of its sides) (Moss, 2001).

The surface of the core contains closely packed, regularly arranged spikes, which comprise the palisade (Dubochet et al. 1994). The core structure is wrapped by a lipid membrane and additional virus proteins, which collectively constitute the IMV membrane. Morphologically, EEV particles are similar to IMV, but contain an additional lipid membrane, which is extremely loose and fragile (Stokes, 1976; Roos et al. 1996). The presence of this extra lipid membrane and associated proteins makes EEV structurally and biologically distinct from IMV. EEV mediates the dissemination of virus within the infected host and protection is better achieved by immunity to EEV proteins than IMV, although immunity to IMV can still be helpful (Boulter, 1969; Appleyard et al. 1971; Boulter et al. 1971; Boulter & Appleyard, 1973; Payne, 1980; Law, Putz & Smith, 2005).

Like all the other members of the Orthopoxvirus genus, the VV genome is a dsDNA molecule of approximately 190 kb (length varies according to virus strain). Generally, genes from the central region of the genome are well conserved, are mostly essential for virus replication and are transcribed from either DNA strand. In contrast, genes located near the termini are more diverse, non-essential for virus replication and are generally transcribed towards the termini.

EFFECTS OF POXVIRUS INFECTION ON THE HOST CELL

Infection with VV results in profound changes in cell function, morphology and metabolism, which are termed cytopathic effect (CPE). Visible effects in vitro include cell rounding and detachment from neighbouring cells (Bablanian, 1970; Bablanian et al. 1978), alterations to the actin cytoskeleton (Hiller et al. 1979; Cudmore et al. 1995), microtubules (Ploubidou et al. 2000) and membrane permeability (Carrasco & Esteban, 1982). This is followed by a phase of enhanced cell migration involving the sequential elongation and condensation of lamellopodia from the cell body (Sanderson, Way & Smith, 1998). Other distinctive changes in cell morphology late during infection include the formation of virus-tipped microvilli (Stokes, 1976; Cudmore et al. 1995) as a result of actin tail extension. VV-induced CPE was traditionally viewed as a gradual degeneration of cell function, morphology and viability. Although this ultimately happens, CPE should not be viewed as a simple shut off of cell functions and random cell degeneration. Some aspects of CPE are a consequence of direct action of viral genes and may constitute a manipulation of cellular mechanisms to the advantage of the virus.

Cellular DNA, mRNA and protein synthesis are inhibited following infection with VV (Buller & Palumbo, 1991). Inhibition of host DNA replication may occur by hydrolysis of nascent single stranded DNA, by a viral endonuclease present in the incoming virion (des Gouttes Olgiati, Pogo & Dales, 1976; Dales, 1990). Concurrent with the orderly expression of VV polypeptides, host protein synthesis decreases gradually after VV infection and is completely shut off around 6 h post-infection (p.i.) (Buller & Palumbo, 1991). In the absence of VV early gene expression, IMV surface tubules (Mbuy, Morris & Bubel, 1982), the B1R protein kinase (Beaud et al. 1994) and F17R phosphoprotein (Person-Fernandez & Beaud, 1986) have been shown to inhibit host protein translation. Following expression of VV early genes, translation of cellular mRNA is inhibited selectively by small non-translated polyadenylated mRNAs (Bablanian et al. 1991; Lu & Bablanian, 1996). Additionally, the overall population of cellular mRNAs is reduced progressively, being replaced by VV mRNAs (Rice & Roberts, 1983), which contributes to the predominant synthesis of VV polypeptides especially at late times p.i. (Moss, 2001). The reduction in the population of host mRNAs, could be due to a decrease in RNA polymerase II activity (and thus cellular mRNA synthesis) and to higher degradation rate of all mRNAs that occurs during VV infection (Rice & Roberts, 1983; Pedley & Cooper, 1984; Buller & Palumbo, 1991).

Although VV has a wide host range, replicating in many cell types of different mammalian and avian species, there are VV host range mutants that fail to replicate in a few specific cell lines (Buller & Palumbo, 1991). VV causes CPE in mouse peritoneal cells or in human leukocytes in vitro, but is unable to produce virus progeny from these cells (Nishmi & Bernkopf, 1958). Productive replication of VV, however, occurs if human leukocytes are pre-treated with a mitogen, suggesting that VV requires replicating target cells for efficient replication (Miller & Enders, 1968). More recently it was shown that later stages of VV infection are blocked in primary human macrophages and murine dendritic cells (Broder et al. 1994; Bronte et al. 1997). Unlike cowpox (CPV), VV is unable to replicate in Chinese hamster ovary (CHO) cells, due to a block in the translation of intermediate mRNAs (Perkus et al. 1990). This restriction is caused by a disruption in VV of a host range gene that is present in CPV, encoding the CP77 protein (Spehner et al. 1988). In VV WR, two genes, K1L and C7L encode proteins that enable VV replication in human cell lines. The K1L gene product is also necessary for replication in RK-13 cells (rabbit kidney epithelioid cells) (Perkus et al. 1990). In a VV K1L deletion mutant, the translation of early VV mRNAs in human or RK-13 cell lines is blocked, and thus there is no VV DNA replication or further gene expression (Ramsey-Ewing & Moss, 1995). Interestingly, the insertion of CP77 into a VV WR K1L deletion mutant partially replaces K1L function, allowing VV to replicate in RK-13 cells (Ramsey-Ewing & Moss, 1996). Therefore, despite the absence of amino acid similarity, CP77, K1L and C7L all influence host range in one cell type or another.

HOST ANTI-VIRAL INNATE IMMUNE RESPONSE

The host immune response to viral infection is biphasic, with innate effectors such as interferons (IFNs), natural killer (NK) cells and macrophages being critical in the early phase, and with adaptative antigen-specific T and B cell responses, which are often essential for clearance of the pathogen and establishment of immunity, developing later. Here, we focus on the innate immune response to VV infection, firstly highlighting its main aspects and then concentrating on the viral strategies to subvert it.

Inflammation is characterized by pain, redness, heat and swelling at the site of infection. The complement system is a crucial innate response that may destroy pathogens directly by lysis or indirectly by opsonising pathogens for phagocytosis by macrophages and neutrophils (Janeway, 2004). Macrophages have an important function as antigen presenting cells (APC) for the activation of T cells and the initiation of a specific immune response. Mice depleted of macrophages are unable to control VV infections due to impaired virus clearance and antigen presentation (Karupiah et al. 1996). NK cells are attracted to the site of infection as part of the inflammatory response and kill virus-infected cells, especially cells with reduced levels of MHC class I on their surface (Maudsley & Pound, 1991; See et al. 1997). NK cells have a direct cytotoxic activity against VV-infected cells in vitro (Brutkiewicz, Klaus & Welsh, 1992) and their in vivo depletion is associated with enhanced VV virulence (Bukowski et al. 1983).

IFNs are a group of secreted proteins that induce an antiviral state in infected or uninfected cells (Samuel, 1991; Johnson et al. 1994). There are two groups of IFNs: type I IFNs are secreted from leukocytes and fibroblasts and include IFN-α and IFN-β. Type I IFNs offer resistance to virus infection, stimulate the expression of MHC class I molecules on the surface of the cell and inhibit cell proliferation (Biron, 1998). Type II IFN, or IFN-γ is secreted from macrophages, NK cells and T lymphocytes. Type II IFN is important in the activation of immune and inflammatory responses and for cell-mediated immunity (Farrar & Schreiber, 1993; Boehm et al. 1997). Both type I and type II IFNs are crucial for the restriction of poxviral infections (Schellekens et al. 1981; Werenne et al. 1985; Rodriguez, Rodriguez & Esteban, 1991; Huang et al. 1993; Muller et al. 1994; van den Broek et al. 1995; Ramshaw et al. 1997; Deonarain et al. 2000).

Mice deficient in IFN receptors are abnormally susceptible to VV infection (van den Broek et al. 1995), showing the importance of these molecules in controlling VV infection, and it is therefore not surprising that VV encodes proteins that counteract their activity. The innate immune response also includes the initial inflammatory response that consists of the infiltration of the site of infection with plasma fluids and large numbers of leukocytes. The recruitment of leukocytes to the site of infection is orchestrated by several cytokines and chemokines secreted by the infected cells or resident macrophages. Amongst the inflammatory cells are monocytes that mature into macrophages and which together with neutrophils, the major cell type of early inflammatory infiltrates, may participate in the phagocytosis of infected cells and therefore contain virus infection (Buller & Palumbo, 1991). Importantly, cells present in the inflammatory infiltrate at the site of infection secrete more antiviral and inflammatory cytokines such as tumour necrosis factor (TNF) and IFNs. The complement system of proteins present in the plasma is able to destroy enveloped virions or infected-cells via the classic (Ab-dependent) or alternative (Ab-independent) pathways, which may also have a role in controlling VV infections given that VV has a strategy to avoid both complement pathways (see Table 1).

Apoptosis, programmed cell death (Kerr, Wyllie & Currie, 1972), is a natural and complex process that occurs in response to a variety of stimuli (Osborne, 1996; Jacobson, Weil & Raff, 1997). Apoptosis can be considered as part of the innate immune response because virus infection can induce apoptosis (Everett & McFadden, 1999). This process can also be targeted by poxviruses, as reviewed in Barry & McFadden (2000).

REPRESENTATIVE VV STRATEGIES THAT EVADE THE INNATE IMMUNE RESPONSE

Table 1 shows representative mechanisms utilised by VV to evade the host immune response, which include both secreted and intracellular proteins. Viral immune evasion strategies often target key aspects of anti-viral immunity and are tuned for maximum benefit to the virus. Therefore by studying such strategies one can predict and probe host anti-viral mechanisms. In this section, we focus on some of the strategies used by VV to evade the host innate immunity. However, it should be noted that VV possesses numerous mechanisms to counteract the host immune responses that will not be discussed here. One example is the product of the H1L gene (Fig. 1) (Najarro, Traktman & Lewis, 2001), shown to bind and dephosphorylate STAT1.

Fig. 1. VV proteins involved in blocking type I IFN and PKR. Please see text for details about E3L, K3L and B18R. STAT1 was shown to be a substrate for H1L in experiments with IFNγ. Although experiments to determine its effect on type I IFN signalling were not carried out, it is possible that H1L might also block this signalling via dephosphorylation of STAT1. Abbreviations: dsRNA, double stranded RNA; PKR, dsRNA-dependent protein kinase; IRF, interferon-regulatory factor; STAT, signal transducer and activator of transcription; ISRE, interferon-stimulated response element; IFN, interferon; IFNAR, interferon α/β receptor; Jak, Janus kinase; Tyk, tyrosine kinase; eIF-2α, eukaryotic translation initiation factor 2α; NFκB, nuclear factor kappa B.

Binding soluble cytokines and interferons and their receptors

The first viral cytokine-receptor to be identified was a viral soluble TNF receptor (vTNFR) (encoded by A53R, see Table 1) (Smith et al. 1991), soon followed by viral soluble receptors for IL-1β (vIL-1βR) (encoded by B15R, Fig. 1) (Alcami & Smith, 1992) and IFN-γ (vIFN-γR) (encoded by B8R) (Upton, Mossman & McFadden, 1992). These viral genes encode proteins with sequence similarity to the extracellular binding domains of host cytokine receptors, but they lack the transmembrane and signalling domains. They were shown to bind cytokines with high affinity and to neutralise their activity (Smith, 2000).

This viral strategy mimics the expression of host soluble versions of cytokine receptors. These receptors are either cleaved from the cell surface or generated by alternative splicing. Cytokine-binding proteins, which modulate cytokine activity, are also present. However, the viral counterparts have unique properties, possibly because they have been optimised during virus-host evolution to be potent cytokine inhibitors (Alcami, 2003). Some viral proteins have binding specificities that are different from those of the cellular receptors. The mammalian type I IL-1R binds IL-1α, IL-1β and IL-1 receptor antagonist (IL-1ra), but VV B15R is specific for IL-1β (Alcami & Smith, 1992). Further studies on B15R demonstrated that IL-1β, and not other cytokines, is the major endogenous pyrogen in a poxvirus infection (Alcami & Smith, 1996).

Other viral soluble cytokine-binding proteins have limited or no sequence similarity to cellular counterparts (Symons, Alcami & Smith, 1995). A common folding of the immunoglobulin domains in the viral IFN-α/β receptors or binding protein (vIFN-α/β-BP) (encoded by B18R) and of the fibronectin type III domains of the cellular counterpart might account for their interaction with the same ligands. VV vIFN-α/β-BP, like vIFN-γR, binds IFNs from several species (Symons et al. 1995). It can also bind to the cell surface after secretion as a mechanism to prevent IFN from binding to cellular IFN-α/βRs and so protect cells from IFN in solution and on the cell surface (Alcami, Symons & Smith, 2000).

Subsequently, a family of secreted proteins encoded by a number of poxviruses (VV (encoded by C12L, Fig. 2), CPV, molluscum contagiosum virus (MCV), ectromelia virus (ECV) and variola virus) that bind IL-18 (vIL-18BPs) and block IL-18 activity (Xiang & Moss, 1999; Born et al. 2000; Smith, Bryant & Alcami, 2000) have been identified. The vIL-18-BPs have no sequence similarity to the mammalian IL-18Rs that mediate the biological effects of IL-18, but the viral proteins are related to human and mouse IL-18BPs that inhibit IL-18 activity (Aizawa et al. 1999; Novick et al. 1999). The vIL-18BP that is encoded by MCV has an extended carboxy-terminal region that is not found in the IL-18BPs that are encoded by VV, CPV, ECV, humans or mice (Smith et al. 2000).

Fig. 2. VV proteins involved in blocking IL-1 and IL-18. Please see text for details about B13R, B22R, B15R, C12L, A46R and A52R. Abbreviations: IL, interleukin; IRAK, IL-1R associated protein kinase; TRAF, tumour necrosis factor receptor-associated factor; IL-1R, IL-1 receptor; IL-1RacP, IL-1 receptor associated protein; IL-18R, IL-18 receptor; NFκB, nuclear factor kappa B; MyD88, myeloid differentiation factor 88.

VV STRATEGIES TO BLOCK TLR-MEDIATED SIGNALLING

TLRs are members of the IL-1R/TLR superfamily, which is defined by the presence of a Toll/Interleukin-1 receptor/Resistance (TIR) domain (responsible for mediating downstream signalling). TLRs are PRRs that recognise and respond to specific pathogen-associated molecules (PAMs), such as bacterial lipoproteins and peptidoglycan (TLR2), dsRNA (TLR3), LPS (TLR4), flagellin (TLR5), ssRNA (TLR7 and TLR8) and CpG DNA (TLR9) (reviewed in Takeda, Kaisho & Akira, 2003), and drive the innate and adaptive immune responses, depending on stimuli and cell type. TLRs are required for cellular responses to some viruses and trigger antiviral signalling pathways leading to IFN responses (Rassa & Ross, 2003; Vaidya & Cheng, 2003; Bowie & Haga, 2005).

In mammals, the host defence against microorganisms mainly relies on pathways that originate from the common TIR domain of the TLRs. The TLR family signalling pathway is highly similar to that of the IL-1R family. Both IL-1R and TLRs (with the exception of TLR3) utilise the adaptor protein myeloid differentiation factor-88 (MyD88) for signalling. MyD88 has a TIR domain in its C-terminal portion but a death domain in its N-terminal portion. Apart from MyD88, there are at least other four mammalian TIR domain-containing adaptors: MyD88-adaptor like (Mal), TIR domain-containing adaptor inducing IFN (TRIF), TRIF-related adaptor molecule (TRAM) and sterile and HEAT–Armadillo motifs (SARM) (reviewed in O'Neill, Fitzgerald & Bowie, 2003).

MyD88 associates with the receptors via homotypic TIR domain interactions. Upon stimulation, MyD88 recruits a death domain-containing serine/threonine kinase, the IL-1R-associated kinase (IRAK) (Martin et al. 1994). There are four IRAKs in the human and murine genomes: two with demonstrated kinase activity, IRAK1 and IRAK4, and two inactive kinases, IRAK2 and IRAKM (Janssens & Beyaert, 2003). The exact physiological role of IRAK2 has not been clarified. IRAK1 and IRAK4 are activated by phosphorylation, leading to the dissociation of IRAK1 from the receptor complex. IRAK1 in turn interacts with the tumour necrosis factor receptor-associated factor (TRAF) family member TRAF6, eventually leading to activation of the transcription factor nuclear factor kappa B (NFκB). TRAF6 was initially identified as a signal tranducer for IL-1 (Cao et al. 1996).

NFκB is a transcription factor that has crucial roles in inflammation, immunity, cell proliferation and apoptosis. The activity of NFκB is regulated by association with IκB, which sequesters NFκB in the cytoplasm until phosphorylated on serine residues by the IκB kinase (IKK) complex (Viatour et al. 2005). This phosphorylation targets IκB for degradation, allowing nuclear translocation of NFκB. The IKK complex contains two catalytic subunits, IKKα and IKKβ, as well as a scaffolding protein, IKKγ. A mammalian protein complex that activates IKK was purified and analysed and found to be composed of two subunits: TAK1 and a ubiquitin conjugating enzyme complex (Deng et al. 2000). TAK1 is consequently activated via its association with the ubiquitinated TRAF6. Once activated, TAK1 mediates phosphorylation of the IKK complex (Wang et al. 2001).

NFκB is not the only transcription factor activated by TLRs. Members of the interferon regulatory factor (IRF) family of transcription factors (Taniguchi et al. 2001), namely IRF3, IRF5 and IRF7, are also activated by TLRs. Specifically, TLR3 and TLR4, via TRIF, activate IRF3, while TLR7 and TLR8, via MyD88, activate IRF5 and IRF7 (Wietek et al. 2003; Takaoka et al. 2005; Uematsu et al. 2005). IRFs are transcriptional activators for IFNα, IFNβ and IFN-stimulated genes and can act as direct transducers of virus-mediated signaling pathways activating IFNα and IFNβ in infected cells (Taniguchi et al. 2001).

We have identified two VV proteins that can antagonise IL-1R and TLR signalling (Bowie et al. 2000). A database search for VV-encoded proteins containing putative TIR domains initially identified A46R. A52R was then identified as the highest scoring sequence with significant similarity (A46R and A52R share 16·3% amino acid identity and 32·9% amino acid similarity). Neither A46R nor A52R have signal sequences, or potential N- and O-glycosylation sites.

Preliminary studies showed that overexpression of both proteins could block IL-1-mediated NFκB activation (representing a third VV mechanism of counteracting IL-1, Fig. 1), and A52R could also block NFκB activation triggered by TLR4 (Bowie et al. 2000). Further experiments revealed that A52R actually blocks NFκB activation by multiple TLRs and associates with both IRAK2 and TRAF6, two molecules directly involved in TLR signalling. These associations disrupt signalling complexes containing these proteins (Harte et al. 2003). A mutant virus in which the A52R gene was deleted was attenuated in vivo when compared to control wild type and revertant viruses. The role of A52R in VV virulence is at least partially explained by the inhibition of multiple pathways leading to NFκB activation. Interestingly, a peptide derived from A52R was shown to inhibit in vitro TLR-induced cytokine secretion and reduced bacterial-induced inflammation in a mouse model of otitis media with effusion (McCoy et al. 2005).

Recently, we have demonstrated that A46R can target multiple TLR adaptors (namely MyD88, Mal, TRIF and TRAM) and interfere with downstream activation of MAP kinases and NFκB. A46R inhibited TRIF-dependent signalling, including IRF3 activation and TRIF-dependent gene induction (Stack et al. 2005). Considering the potential role of TRIF in controlling VV replication in macrophages (VV replicated to a higher titre in peritoneal macrophages from TRIF −/− knockout mice when compared to heterozygous and wild type mice, suggesting a role for TRIF in inhibiting VV replication) (Hoebe et al. 2003) and the attenuation caused by the deletion of A46R from the VV genome, the targeting of this molecule by A46R may represent a novel and particularly important viral immune evasion mechanism.

Furthermore, A46R and A52R were shown to be non-redundant. A52R potently inhibited poly(I:C)/TLR3-induced NFκB activation, whereas A46R had little effect. This contrasts with the finding that A46R, but not A52R, is a strong inhibitor of poly(I:C)/TLR3-induced IRF3 activation. Also, only A46R was able to inhibit TLR-induced MAPK activation (Stack et al. 2005).

Another VV protein recently shown to inhibit NFκB is N1L. Originally identified in concentrated supernatant of VV-infected cell cultures, the product of the N1L gene is actually an intracellular protein and exists as an homodimer in physiological conditions (Bartlett et al. 2002). The N1L gene is not essential for virus replication in vitro and its deletion does not affect growth properties in cell culture. Viruses lacking N1L are strongly attenuated in vivo, both in the intradermal and in the intranasal models of infection (Bartlett et al. 2002). This is in contrast to a variety of other VV immunomodulators, associated with virulence only in one of these two routes (Tscharke et al. 2002).

Further studies with the overexpression of N1L in mammalian cells revealed that N1L inhibits NFκB activation by a variety of stimuli, including IL-1β, TNFα and TLR2, TLR3 and TLR4 agonists (DiPerna et al. 2004). NFκB activation by different TIR adaptors was also shown to be inhibited by N1L, pointing to a putative N1L target downstream in the signalling pathway. This was confirmed by experimental evidence showing that N1L associates with several proteins of the IKK complex (the point of convergence for all pathways inhibited by N1L), with TBK1 being the most likely target of N1L.

A final identified VV inhibitor of NFκB is K1L. Following the finding that MVA infection induces NFκB activation, subsequent experiments revealed that the insertion of the K1L gene into the MVA genome was sufficient for inhibiting degradation of host IκBα (Shisler & Jin, 2004). Transcription of both the NFκB-regulated cellular TNF gene (regulated by a natural promoter) and a luciferase gene (regulated by an artificial promoter) was repressed by infection of MVA expressing K1L. K1L is necessary for inhibiting NFκB activation in RK-13 cells (Shisler & Jin, 2004), as well as being a host range factor in this cell line (see above). The authors did not imply a link between host range and NFκB activation nor did they investigate whether MVA-mediated NFκB activation is TLR-dependent. K1L may inhibit IκBα degradation by interacting with IKK to prevent phosphorylation or by hampering kinases that act upstream of IKK. The targets of K1L have not yet been identified.

FINAL REMARKS

For several reasons, the NFκB pathway provides an attractive target for viral pathogens. Activation of NFκB is a rapid, immediate-early event that occurs within minutes after exposure to a relevant inducer, does not require de novo protein synthesis, and results in a strong transcriptional stimulation of several early viral genes as well as cellular genes (Hiscott, Kwon & Genin, 2001). Many viruses have evolved mechanisms to target the NFκB pathway to facilitate their replication, cell survival, and evasion of immune responses. In addition, some viruses use the NFκB pathway either for its anti-apoptotic properties to evade the host defence mechanisms or to trigger apoptosis as a mechanism of virus spread (Hiscott et al. 2001).

NFκB is known to be activated by multiple viruses, including HIV-1, HTLV-1, hepatitis B virus (HBV), hepatitis C virus (HCV), Epstein-Barr virus (EBV), influenza virus (Hiscott et al. 2001) and respiratory syncytial virus (RSV) (Kurt-Jones et al. 2000). Poxviruses, however, seem to favour the inhibition of NFκB. At least four different VV-encoded proteins are capable of inhibiting NFκB activation – and three of those (A46R, A52R and N1L) specifically block multiple TLR-induced signals, suggesting that TLRs might have a role in sensing VV. Recently, many TLRs have been linked to sensing different viruses (for reviews, see (Rassa & Ross, 2003; Vaidya & Cheng, 2003; Bowie & Haga, 2005)), and the ligands for the TLRs in question can be either viral proteins or nucleic acids, we are faced with important questions: does any TLR actually sense VV and, if so, which TLR recognise which viral component? Considering the characteristics of poxviruses, we could speculate that the most likely TLRs to be involved would be TLR3 (which senses dsRNA) and TLR9 (which can sense viral dsDNA genomes) (Lund et al. 2003). TLR3, TLR7, TLR8 and TLR9 are all intracellular receptors that recognise nucleic acids, with TLR7 and TLR8 linked to ssRNA (Diebold et al. 2004; Heil et al. 2004; Lund et al. 2004). However, other TLRs could not be excluded. For instance, VV also targets TLR4 signalling (Stack et al. 2005) and this receptor is known to be activated by proteins from other viruses (Kurt-Jones et al. 2000; Burzyn et al. 2004; Triantafilou & Triantafilou, 2004). Fig. 3 shows a hypothetical model for the sensing of VV by TLR3 and the known virus-encoded proteins able to block the triggered signalling pathways. Apart from the fact that VV is able to block every single one of the pathways depicted, it should be noted that blocking of NFκB activation by different proteins occurs at different points in the signalling cascade.

Fig. 3. Proposed model for the inhibition of TLR3/TRIF-mediated signalling by VV. Please see text for details about A46R, A52R and N1L. Abbreviations, TLR, Toll-like receptor; TRIF, TIR domain-containing adaptor-inducing IFN; IKK, IκB kinase; TBK1, TANK-binding kinase 1; NFκB, nuclear factor kappa B; IRF, interferon-regulatory factor.

In conclusion, VV possesses many strategies to evade the host innate immune response. Recent work with TLRs support the hypothesis that these receptors are involved in sensing and responding to VV infection.