Transcriptional Activation of Alpha/Beta Interferon Genes: Interference by Nonsegmented Negative-Strand RNA Viruses

Alpha/beta interferons (IFN-α/β), including the single IFN-β and a large group of IFN-α isotypes, represent an essential element of host defense against viruses and other pathogens, as both IFN-α and IFN-β excite immediate innate antiviral and antiproliferative activities in cells and alert the adaptive immune system to mount an adequate Th1-biased immune response (30, 48, 68). Production of IFN is induced by substances containing conserved molecular patterns that identify them as nonself, or “strange” (41, 83). Once secreted, IFN-α/β works in an auto- and paracrine fashion to exert its antiviral and biological effects. All members of the IFN-α/β family have a single IFN-α receptor (IFNAR), which is associated with Jak1 and Tyk2 tyrosine kinases. Binding of IFN to the IFNAR results in the activation of the latent transcription factors STAT1 and STAT2, which then dimerize and associate with p48 (IFN regulatory factor [IRF] 9). This complex, known as IFN-stimulated gene factor 3, binds to DNA sequences (IFN-stimulated response elements) present in the promoters of hundreds of genes and promotes their transcription (1, 30, 32, 68).

The function of the powerful IFN response relies on both rapid production of IFN and controls that regulate and limit the extent of response. Transcription of the IFN-α/β genes is primarily controlled by proteins of the IRF family, in particular, IRF-3 and IRF-7 (7, 56). In most body cells, activation of the latent IRF-3 triggers expression of only a small subset of IFN genes, in particular, IFN-β. This early IFN acts in an auto- or paracrine manner by JAK/STAT signaling to stimulate the synthesis of IRF-7, which controls transcription of many additional members of the IFN-α gene family. Like IRF-3, IRF-7 must be activated by danger signals to ensure that an extensive IFN response is limited to endangered cells (8, 50, 57).

However, among hematopoietic cells there are task forces that instantaneously fire with all arms. In the plasmacytoid dendritic cells (PDC), specialized hematopoietic cells also known as natural IFN-producing cells, IRF-7 is constitutively expressed, allowing a rapid and comprehensive IFN-α response to danger signals (6, 33, 40). Notably, the small population of PDC produces the bulk of IFN in infected hosts. Because of their enormous capacity to produce IFN-α, PDC are regarded as the main sentinels for triggering a general response to viruses and represent a key interface between the innate and adaptive arms of the immune system (4, 18, 26).

Tremendous progress is currently being made in the elucidation of the signaling pathways leading from the recognition of danger signals to the activation of IRFs. The sensors triggering IFN production include members of the Toll-like receptor (TLR) family present at the cell surface or in endosomal compartments, namely, TLR3, TLR4, TLR7, TLR8, and TLR9 (2). In addition, two RNA helicases, retinoic acid-inducible gene I protein (RIG-I) and melanoma differentiation-associated gene product MDA-5, have recently been identified as cytosolic receptors for viruses that may transmit signals downstream to activate IRFs (3, 94). Whereas TLR3/4 and the RNA helicases can activate both IRF-3 and IRF-7, the pathway triggered by TLR7/8/9 appears to be specific for the activation of IRF-7 (Fig. 1). Since TLR7 and TLR9 are present on PDC, this pathway is of special impact in the activation of these cells. An important common downstream component of the former pathways appears to be TANK-binding kinase-1 (TBK1) or its equivalent, the inducible IKKi, which are the first, and so far only, kinases identified as being able to phosphorylate IRF-3 and IRF-7 (24, 79).

FIG. 1.

Schematic representation of the positive IFN feedback loop. Latent IRF-3 is activated through TLR3 and TLR4 signaling by dsRNA and LPS, respectively, or by virus infection. The early IFN-β (and early IFN-α subtypes) induces the expression of ISGs, including IRF-7. IRF-7 is activated again by dsRNA or LPS signaling or by virus infection. In addition, ligands of TLR7 and TLR8, including single-stranded RNA (ssRNA), and TLR9, including CpG DNA, can activate IRF-7. α-prom., IFN-α promoter; β-prom., IFN-β promoter.

Natural viruses must have the means to attenuate the IFN system in order to establish an infection. Indeed, viral IFN antagonists can interfere with any aspect of the IFN system, including IFN gene induction, IFN JAK/STAT signaling, and antiviral effects of IFN-stimulated genes (ISGs) (for reviews, see references 9, 28, 30, 42, and 90). The establishment of reverse genetics systems for negative-strand RNA viruses (for recent reviews, see references 19 and 62) has more recently also allowed the characterization of IFN antagonists from this virus group and the study of their relevance in the virus context. The first IFN antagonist of a negative-strand RNA virus identified was the NS1 protein of influenza A virus (29). NS1 is a double-stranded RNA (dsRNA) binding protein which may prevent the recognition of viral RNAs and the activation of the antiviral and antiproliferative protein kinase R (PKR) (for a comprehensive review on the functions of NS1, see reference 28). More recently, IFN antagonists of nonsegmented negative-strand RNA viruses (NNSV), which constitute the Mononegavirales order and comprise the Rhabdoviridae, Paramyxoviridae, Filoviridae, and Bornaviridae families, have been identified. Interestingly, these antagonists follow quite distinct strategies. It becomes evident that members of all NNSV families encode IFN antagonists that actively interfere with pathways activating IRF-3 and IRF-7. The first viruses that are able to block the IFN response of PDC and which therefore make these important cells blind to pathogens and immunomodulatory substances like CpG oligodeoxynucleotides (ODN) (73) have now been identified. This review briefly outlines the latest findings on IRF-activating pathways and points out how NNSV may interfere with IRF activation.

ACTIVATION OF IRF-3 AND IRF-7: THE TBK1 PATHWAY

Although the essential role of IRF-3 in the transcription of IFN genes in response to viruses has been known for some time, the mechanism of IRF-3 activation by upstream signals had remained unclear. A major breakthrough came a few years ago when Yoneyama et al. identified phosphorylation of IRF-3 as the key downstream event triggered by virus infection to induce IFN transcription (95). After infection, IRF-3, which is normally in the cytoplasm, is immediately phosphorylated at specific serine residues in its C-terminal regulatory domain. It then dimerizes and can interact with the transcription coactivator p300 or CBP in the nucleus, resulting in specific DNA-binding activity (53, 70, 88, 89, 95).

Another breakthrough was the identification of two noncanonic IkB kinases, IKKi (or IKKɛ) and TBK1, as kinases that can phosphorylate IRF-3 and IRF-7 (24, 79). A major role in the initial activation of IRF-3 has been attributed to the ubiquitous and constitutively expressed TBK1, since transcriptional activation of IRF-3-controlled genes, including IFN-β and RANTES, is abolished in TBK1-deficient mice in response to Sendai virus infection (58). The expression of IKKi is inducible and limited to hematopoietic cells. Notably, TBK1 (and IKKi) was identified as an essential component of TLR3- and TLR4-mediated IFN induction by dsRNA and bacterial lipopolysaccharides (LPS), respectively (24, 37, 71), suggesting that the responses to bacterial and viral pathogens are highly similar (65).

After the binding of ligands, individual TLRs trigger distinct responses by recruiting a different combination of adapter molecules, such as MyD88, TIRAP/MAL, TRIF/TICAM-1, or TRAM, to their C-terminal Toll-interleukin-1 receptor (TIR) domains (for a review, see reference 2). Signaling of TLR3 and TLR4 to stimulate IFN induction requires the adapter TRIF, which binds to TLR3 directly and to TLR4 via TRAM (25, 92, 93). A direct interaction has been shown for TRIF and TBK1 (25, 71), and formation of a complex containing TRIF, TBK1, and IRF-3 was found to be dependent on the kinase activity of TBK1 and on TRIF phosphorylation. Notably, binding of TBK1 to TRIF and binding of TRAF6 (which is involved in transmitting TRIF signals from TLR3/4 to produce NF-κB activation) to TRIF were mutually exclusive (71).

TBK1 has been shown to be necessary for the activation of IRF-3 and IRF-7 by virus infection and by LPS- and dsRNA-triggered TLR signaling in cells. Overexpression of TBK1 or of TRIF in cells activates IRF-3, and, as shown by in vitro experiments, TBK1 phosphorylates IRF-3 at the C terminus. A major question to be addressed now is how the activity of TBK1 is controlled or regulated. An interesting finding in this regard has been published most recently by Sarkar et al. (69). These authors studied TLR3 signaling in response to dsRNA and report that full phosphorylation and transcriptional activation of IRF-3 required an additional signaling pathway triggered by recruitment of PI3K to the TIR and involving activation of PI3K and the downstream kinase Akt. This finding implies that this second pathway is needed for full activation of TBK1 to phosphorylate IRF-3 or provides another kinase that phosphorylates the critical residues of IRF-3 upon priming of IRF-3 by TBK1-mediated phosphorylation.

Evidence is accumulating that the status of the Ser-386 residue critically determines the transcriptional activity of IRF-3. IRF-3 phosphorylated at Ser-386 is observed exclusively with dimers, and mutation of Ser-386 abolishes the dimerization potential of IRF-3 (59). TBK1 can phosphorylate Ser-386 and IRF-3 in vitro, although the primary targets are Ser-396, -398, -402, and -405 (58). Moreover, an IRF-3 mutant with a phosphomimetic substitution of Ser-396 (IRF-3 5D) is constitutively active in the transcription of IRF-3-responsive genes (16, 52, 77). A model that could accommodate these findings is one in which phosphorylation of IRF-3 at Ser-396 makes the molecule a more suitable substrate for TBK1 or for other kinases, allowing effective phosphorylation of Ser-386.

SHORTCUT TO IRF-7 ACTIVATION: THE MYD88/TRAF6/IRF-7 CONNECTION

PDC, which constitutively express high levels of IRF-7, can produce huge amounts of IFN-α in an immediate response (6, 33, 40, 44). The TLR repertoire of human PDC is restricted to TLR7 and TLR9, which are both located in endosomal membranes (87). As shown recently, TLR7 and TLR8 recognize viral single-strand RNA (21, 36) as well as imidazoquinolines such as imiquimod or resiquimod (R848) and guanosine analogs (reviewed in references 2 and 85). In contrast, TLR9 recognizes bacterial or viral DNA, including synthetic CpG oligodeoxynucleotides (ODN) (34, 45). In addition, incubation with a variety of inactivated or live DNA and RNA viruses, including herpes simplex virus types 1 and 2 (40, 47, 54), murine cytomegalovirus (20), human immunodeficiency virus type 1 (96), influenza A virus (21, 55), Sendai virus (39, 40), and vesicular stomatitis virus (VSV) (6, 55), was shown to trigger potent IFN-α production in PDC. Notably, IFN-α induction in PDC depends on the canonical TIR adapter MyD88, which is otherwise involved in the activation of NF-κB. As for the synthetic TLR ligands, stimulation of IFN-α by inactivated herpes simplex virus, influenza A virus, and VSV was found to critically depend on MyD88, suggesting that viral nucleic acids or other virus components are somehow recognized by TLR7 and TLR9 (21, 47, 54, 55).

Most recently, work by Kawai and colleagues has disclosed the first details of MyD88-dependent IFN induction (43). In a yeast two-hybrid screen, IRF-7 was identified as a potential interactor with MyD88. By further analyses including colocalization studies and immune precipitations, these authors were able to confirm that in mammalian cells MyD88 forms a complex with IRF-7 but not with IRF-3. For transcriptional activation of IRF-7, further association with TRAF6 was found to be required, as was the ubiquitin ligase activity of TRAF6, which is essential also for NF-κB activation downstream of TLR signaling via kinases like TAK1 (2, 82). The first experiments with bone marrow cells from TBK1- or IKKi-deficient mice suggested that TBK1 and IKKi is not involved in MyD88- and TRAF6-dependent IFN-α induction (43). Thus, the IRF-7 kinase activated by TLR9 signaling remains to be identified.

RECOGNITION OF INTRACELLULAR VIRUS BY RNA HELICASES

PKR was one of the first dsRNA binding and antiviral proteins identified (for a review, see reference 91). PKR is activated by cytosolic dsRNA and by virus infection and stimulates the activation of NF-κB. PKR is therefore synergistic in IFN-β induction, but it is not sufficient to activate IRFs (80). Another breakthrough, by Yoneyama et al., was the recent identification of RIG-I as an important component of the detection system upstream of IRF-3 phosphorylation (94). RIG-I is a DEXD/H box RNA helicase that contains two caspase-recruiting domains (CARDs) at its N terminus. CARDs may interact with CARDs from other proteins and are platforms for nucleating signaling events. RIG-I was found to be required to trigger efficient IFN-β induction in cells infected with a paramyxovirus, Newcastle disease virus. Ectopic expression of the RIG-I CARD alone was sufficient for IFN induction. Both the CARDs and ATPase activity were found to be required for IFN induction in virus-infected cells by RIG-I, suggesting that not only binding to viral RNA but also some processing of virus components is necessary for downstream signaling by the CARD. These findings suggest a direct downstream adapter to the CARD of active RIG-I that relays to IRF kinases or their upstream activators. TRIF was found not to be required, distinguishing the RIG-I-triggered pathway from the TLR3/4-TRIF pathway.

As indicated in the Yoneyama et al. paper, expression of the CARD of the closest relative to RIG-I, MDA-5, was found to similarly stimulate IRF3-dependent reporter genes, suggesting functional similarities between these two RNA helicases (94). Indeed, the role of MDA-5 in triggering IFN gene activation in response to virus infection has very recently been confirmed (3). Moreover, MDA-5, but not RIG-I, was identified as a direct target of the paramyxovirus V proteins (3) which interfere with IFN-β induction (see below). Both RIG-I and MDA-5 are IFN inducible in most cells, such that an initial IFN encounter would potentiate the expression of these dsRNA receptors.

IRF ANTAGONISTS OF RHABDOVIRUSES: MATRIX AND PHOSPHOPROTEIN

NNSV are enveloped viruses with a typical helical ribonucleoprotein complex in which the viral RNA is tightly enclosed. The prototypic rhabdoviruses, like VSV and rabies virus (RV), are the “minimal” NNSV, encoding only five structural proteins, the nucleoprotein (N), the phosphoprotein (P), the inner matrix protein (M), the transmembrane glycoprotein (G), and the RNA polymerase (L), in five monocistronic genes. The N, P, and L proteins along with the RNA make up the RNP, which is enwrapped during virus budding at the cell surface into an envelope containing the M and G proteins. VSV is the typical IFN-sensitive virus and has been utilized as an indicator of the biological activity of IFNs. As shown recently for VSV, rhabdovirus infection does not go unrecognized and IRF-3 is activated (84). Therefore, rhabdoviruses must have the means to limit IFN expression.

VSV is a fast and highly cytopathic virus which can shut down host cell gene expression. Production of IFN by VSV isolates or mutants is inversely correlated with the ability of the viruses to inhibit cellular protein expression (27). VSV cell shutdown and inhibition of IFN are largely due to the M protein, which is a potent inhibitor of cellular mRNA transcription and transport (13, 22, 38). VSV functions that more specifically interfere with IRF-activating pathways have not been demonstrated so far, suggesting that the general shutdown of host gene expression obviates the need for specific IFN or IRF antagonists in fast and cytopathic viruses. Similar strategies to avoid regulated host responses have been observed for the segmented NNSV, such as bunyaviruses which interfere with host gene transcription by targeting essential components of the transcription machinery. For example, the small nonstructural protein (NSs) of Rift Valley fever virus interacts with and blocks components of TFIIH (12, 49).

Compared to VSV, RV and other lyssaviruses are more slowly growing, and a general cell shutoff has not been observed. Since the expression of host cell genes is possible throughout infection, IFN expression must be specifically controlled. We have recently identified the RV P protein as a protein that interferes with the activation of IRF-3 (17). Like in other NNSV, the RV P is an essential virus protein that is required for RNA synthesis, in which it acts as a polymerase cofactor, as well as for the assembly of RNPs, in which it functions as a chaperone for specific and proper encapsidation of viral RNA by the N protein. A first hint of the IFN-inhibitory function of RV P was obtained with a recombinant virus in which P was modified by fusion to a fluorescent protein, eGFP (23), and which induced IFN-β. Similarly, a virus expressing very low amounts of P, just sufficient to support replication of the virus, was a strong IFN inducer. In contrast to the case for wild-type RV, infection with these IFN-inducing viruses led to strong activation of IRF-3-controlled reporter genes. Notably, NF-κB and AP1 activities were similar in IFN-inducing and noninducing viruses, suggesting specific interference of RV P with IRF-3 activation. Ectopic expression of the P protein or infection with wild-type RV inhibited IRF-3 activation by overexpression of TBK1. In the presence of P, the critical phosphorylation of the IRF-3 Ser-386 residue was abolished, as were formation of IRF-3 dimers, nuclear import, and transcriptional activity of IRF-3 (17). Thus, P interferes with the function of TBK1, and probably also IKKi, in activation of IRF-3 (Fig. 2). A direct interaction of P with IRF-3 or TBK1 could not be shown in preliminary coprecipitation experiments, suggesting either weak interactions or indirect effects. Further experiments will be aimed at identification of the domains of P required for inhibition of IRF-3 activity. It will also be interesting to see whether functions of P in viral RNA synthesis, N chaperoning, and IRF inhibition can be dissociated. It has been shown that the above-mentioned IFN-inducing RVs are not viable in IFN competent cells, since they have lost the ability to counteract IFN. Such rhabdoviruses are particularly interesting with regard to the development of live vaccines and of oncolytic virus vectors (for a review, see reference 51).

FIG. 2.

Activation of IRF-3 and IRF-7 by TLR-dependent pathways and intracellular virus. TLR3 and TLR4 activate the IRF kinases TBK1 and IKKi (not shown) through the TLR adapter TRIF, which has been shown to associate with TBK1 and IRF-3 (for details, see the text). Intracellular virus is recognized by RNA helicases RIG-I and MDA-5, which transmit signals downstream, probably to TBK1, through CARD interactions. IFN induction by TLR7 and TLR9 ligands depends on the TLR adapter MyD88 and involves a complex of MyD88, IRF-7, and TRAF6. TRAF6 is thought to activate an IRF-7 kinase (X) that awaits identification. The targets of some NNSV are indicated. Influenza virus NS1 protein prevents viral RNA from being recognized. The V proteins of paramyxoviruses, including SV5, mumps, and Hendra viruses, bind to MDA-5 and block downstream signaling. Ebola virus, RSV, and rabies virus interfere with activation of IRF-3 in virus-infected cells. For Ebola virus VP35, an inhibitory effect on TRIF-mediated TLR signaling has been observed. Rabies virus P was shown to inhibit TBK1-mediated phosphorylation of IRF-3. Measles virus and RSV can effectively block MyD88-dependent IFN induction by TLR7 and TLR9 ligands (for details, see the text).

THE V AND C PROTEINS OF PARAMYXOVIRINAE

In addition to the essential phosphoprotein P, the P genes of most members of the Paramyxovirinae subfamily encode additional, nonessential proteins. These proteins include the V proteins, which are translated from “edited” P gene transcripts. Paramyxovirinae RNA editing, also known as “pseudotemplated transcription,” involves the cotranscriptional introduction of one or more G residues at defined sites by the “stuttering” viral RNA polymerase (46). The V and P proteins therefore have identical N-terminal moieties and specific C-terminal domains. Most interestingly, the V-specific C-terminal region is highly conserved throughout the paramyxoviruses and contains several cysteine residues similar to zinc finger and RING finger domains. As revealed in the past, the V proteins of most paramyxoviruses are responsible for the pronounced IFN resistance of these viruses. The V proteins associate with STAT1/STAT2 complexes and abolish IFN JAK/STAT signaling by targeting either STAT1 or STAT2 for proteasomal degradation or by interfering with STAT phosphorylation. In most cases, both N-terminal and C-terminal domains of V proteins contain regions important for JAK/STAT signal inhibition. In addition to V proteins, members of the respiroviruses, including the prototypic Sendai virus, and members of the morbilliviruses, including, for example, measles virus, may encode one or more C proteins which are expressed from the P mRNAs by alternative translation initiation. These proteins do not share any sequences with the P or V proteins. Notably, in Sendai virus the C proteins (C′, C, Y1, and Y1) rather than the V protein are responsible for abolishing IFN STAT signaling in a complex manner. In the case of measles virus, the blocking of JAK/STAT signaling has been attributed to both C (78) and V proteins (63, 64). For recent comprehensive reviews on the inhibition of JAK/STAT signaling by paramyxovirus V and C proteins, see references 28, 31, and 61.

In view of the IFNAR-dependent induction of IFN-α isotypes in most cell types, it is obvious that paramyxoviruses capable of abolishing IFN JAK/STAT signaling are poor inducers of IFN-α. However, paramyxoviruses can in addition directly block IFN-β induction. A simian virus 5 (SV5) in which expression of the V protein-specific C-terminal domain was abolished was found to have lost not only the ability to degrade STAT1 but also the ability to suppress nuclear import and activation of IRF-3 and activation of NF-κB, leading to the induction of IFN-β expression (35). Ectopic expression of the SV5 V protein and of V proteins from parainfluenza virus type 2 and Sendai virus counteracted induction of IFN-β by dsRNA and virus infection. Moreover, the C-terminal Cys-rich domain of V alone was found to be sufficient to counteract IFN-β promoter-driven luciferase reporter genes, whereas this domain is not sufficient to induce STAT degradation (66).

Work by Andrejeva and colleagues has now identified the RNA CARD helicase MDA-5 as a direct target of the SV5 V protein and of the V proteins from human parainfluenza virus type 2, Sendai virus, mumps virus, and Hendra virus (3). A 150-kDa protein that was coprecipitated with SV5 V from IFN-treated cells was identified as the IFN-induced MDA-5 (Fig. 2). Overexpression of MDA-5 or of the MDA-5 CARD alone stimulated IRF-3 and NF-κB activity and IFN expression, particularly in the presence of dsRNA, as was previously reported for RIG-I (94). Coexpression of full-length V or of the V-specific C-terminal domain containing the cysteine cluster alone greatly reduced this MDA-5 activity. Since it appears that SV5 V does not bind RIG-I and that V cannot block IFN induction by dsRNA completely (3), it is suggested that RIG-I and MDA-5 independently contribute to dsRNA recognition and downstream signaling.

THE NS PROTEINS OF PNEUMOVIRUSES

The Pneumovirinae subfamily of paramyxoviruses comprises the Pneumovirus and Metapneumovirus genera, represented by respiratory syncytial virus (RSV) and human metapneumovirus, respectively. In contrast to the Paramyxovirinae, these viruses do not edit RNA and lack a protein corresponding to V. However, pneumoviruses possess extra genes not present in other paramyxoviruses, such as two genes encoding nonstructural (NS) proteins, NS1 and NS2.

Human RSV (HRSV) and its bovine counterpart (BRSV) are highly resistant to treatment with IFN-α/β (5, 97). By analysis of gene deletion mutants of BSRV, we were able to map the IFN resistance phenotype to the NS genes (72). Surprisingly, viruses lacking either NS1 or NS2 lost their ability to replicate in IFN-treated cells, suggesting a cooperative function of the two proteins in counteracting antiviral IFN effects. This function was confirmed by experiments in which NS proteins from BRSV or from HRSV were expressed from recombinant RV. Only coexpression of NS1 and NS2 conferred an increase in IFN resistance to RV. The exchange of the BRSV NS genes with those of HRSV further showed that the IFN resistance function mediated by NS proteins is better adapted to the respective homologous host species (14). In contrast to V proteins, the NS proteins do not significantly interfere with IFN JAK/STAT signaling (67, 72, 97), suggesting activities in counteracting the functions of IFN-induced antiviral proteins.

Further analyses of BRSV and HRSV NS gene deletion virus mutants revealed that the NS proteins are also critical for preventing induction of IFN-β in RSV-infected cells (15, 81, 86). Recombinant viruses lacking either NS1 or NS2 caused a significant increase in the transcriptional activity of IRF-3 compared to that observed for wild-type RSV. This increase in transcriptional activity was correlated with highly increased levels of phosphorylated IRF-3. Notably, NF-κB activities remained similarly high in cells infected with wild-type and NS deletion mutant viruses. Therefore, the RSV NS proteins seem to specifically target the IRF-3 activation branch of the IFN-β-activating pathways (15), in contrast to the V proteins of Paramyxovirinae, which inhibit both IRF-3 and NF-κB activation by blocking functions of MDA-5 (see above).

EBOLA VIRUS VP35

The Ebola virus VP35 protein, which corresponds to the P protein of other NNSV, was identified as an IFN antagonist on the basis of its ability to rescue the growth of an NS1-deficient influenza A virus in IFN-competent cells. Ectopic expression of VP35 enhanced influenza delNS1 virus growth in Madin Darby canine kidney cells more than 100-fold. VP35 subsequently was shown to block dsRNA- and virus-mediated induction of the IFN-β promoter and of other IRF-3-dependent promoters, including those of IFN-α4, ISG54, and ISG56. It was also found that VP35 did not interfere with IFN JAK/STAT signaling (11). Ectopic expression of Ebola virus VP35 blocked virus-induced IRF-3 phosphorylation, IRF-3 dimerization, and nuclear translocation. Consistent with these observations, Ebola virus infection of Vero cells activated neither transcription from the ISG54 promoter nor nuclear accumulation of IRF-3, suggesting that in Ebola virus-infected cells VP35 inhibits the induction of IFNs by blocking IRF-3 activation (10).

Notably, and in contrast to the RV P protein, which interferes with TBK1-mediated phosphorylation of IRF-3 (see above) (17), it appears that VP35 does not greatly affect the kinases upstream of IRF-3, TBK1, and IKKi, suggesting that the VP35 target may lie upstream of these kinases (C. Basler, personal communication). Interestingly, the expression of VP35 has the ability to interfere with TRIF-mediated activation of IFN-β. Overexpression of either TRIF or TRAM in cells leads to IRF-3 activation. VP35 inhibits TRIF-mediated activation at low TRIF concentrations and also inhibits TRAM-induced activation of IRF-3 over a wider range of TRAM concentrations. Moreover, coimmunoprecipitation data also suggest that VP35 can interact with TRAM. This finding fits well with the observation that the expression of VP35 in U373 cells inhibits signaling through LPS, the ligand for TLR4. In addition, RIG-I-induced activation of IRF-3, which is TRIF independent, is prevented by VP35 (C. Basler, personal communication), suggesting multiple targets of VP35.

INHIBITORS OF TLR7 AND TLR9 SIGNALING IN HUMAN PDC

PDC are equipped with TLR7 and TLR9 and constitutively express high levels of IRF-7. Viruses recognized by TLR7 or TLR9 (see above) or synthetic TLR agonists like R848 or CpG ODN can therefore rapidly induce IFN-α production in PDC through MyD88 and TRAF6 (Fig. 2) (43). In addition, PDC are able to sense virus replicating in the cytosol, as we were able to show recently. The infection of human PDC with a particular laboratory strain of RSV (subtype A, strain Long) led to potent IFN-α induction in a TLR-independent manner. Entry into the cytosol was necessary for IFN induction, since UV-inactivated virus or virus constructs with inactive fusion proteins did not activate PDC (39).

In view of the key role of PDC in scenting viruses and in setting off a general IFN alert, the IFN-inducing pathways of PDC should be a major target for viruses. Indeed, we have recently identified the first viruses able to block IFN production by PDC, namely, measles virus and HRSV (73). Infection of human PDC with clinical HRSV isolates from hospitalized children and with the RSV laboratory strain A2 as well as with the measles vaccine virus Schwarz did not lead to considerable IFN-α production, in striking contrast to infection with the above-mentioned RSV strain Long, which is an IFN-inducing virus in other cells also. Given that HRSV and measles virus are not recognized by the PDC TLRs, this observation suggests that at least the pathways triggered by recognition of cytosolic replicating virus were successfully counteracted. Treatment of RSV A2 or measles virus-infected PDC with the TLR7 agonist R848 (resiquimod) or with a CpG ODN otherwise activating TLR9-mediated IFN-α production revealed that TLR-dependent IFN-α production of PDC was also abolished in virus-infected cells. Notably, PDC activated by pretreatment with the TLR agonists were still permissive to subsequent infection with RSV A2 and measles virus, and IFN production triggered by the TLR agonists could be switched off by virus infection. Thus, both measles virus and RSV have the capacity to block MyD88- and TRAF6-mediated activation of IRF-7 and to blindfold PDC to TLR ligands.

In vivo, RSV and measles virus should therefore be able not only to avoid an immediate IFN alert caused by themselves but also to prevent or diminish IFN responses triggered by coinfecting unrelated pathogens, including single-stranded RNA or DNA viruses and bacteria. Consequently, infection by RSV and measles virus should make hosts more permissive to other pathogens. Indeed, measles is characterized by a typical Th2-biased immune pathology, and deaths from measles are largely due to an increased susceptibility to secondary bacterial and viral infections, which is attributed to a general immune suppression induced by measles virus (for reviews, see references 60 and 76). One aspect contributing to immune suppression and which is shared by measles virus and RSV is contact-mediated inhibition of T-cell proliferation (74-76). It now remains to be determined to what extent the ability of measles virus and RSV to counteract PDC functions contributes to the pathology of measles and RSV infection.

CONCLUDING REMARKS

There is little doubt that the balance between host responses and viral evasion mechanisms is a key determinant of the outcome of viral infection and that the IFN system is a major bridge between the innate and adaptive arms of the immune system. It is therefore not surprising that even viruses with a limited coding capacity, like the Mononegavirales, encode antagonists that specifically interfere with the activation of IRFs and the production of IFN. As new molecular mechanisms of the regulation of IFN induction by IRFs in different cell types are discovered, the mechanisms by which these viruses may interfere with the development of immunity are being discovered. Detailed knowledge of these molecular mechanisms is important for understanding viral pathogenicity and may lead to improved virus vaccines and antiviral agents. Some members of the Mononegavirales have long been known to have oncolytic properties, which are probably related to preferential replication in IFN-deficient tumors (51). More-detailed knowledge of the molecular mechanisms involved and the possibility of generating viruses with accordingly modified IFN antagonists may help in the development of viral vectors for oncolytic approaches.