The Herpes Simplex Virus ICP0 RING Finger Domain Inhibits IRF3- and IRF7-Mediated Activation of Interferon-Stimulated Genes

Virus infection induces a rapid cellular response in cells characterizedby the induction of interferon. While interferon itself doesnot induce an antiviral response, it activates a number of interferon-stimulatedgenes that collectively function to inhibit virus replicationand spread. Previously, we and others reported that herpes simplexvirus type 1 (HSV-1) induces an interferon -independent antiviralresponse in the absence of virus replication. Here, we reportthat the HSV-1 proteins ICP0 and vhs function in concert todisable the host antiviral response. In particular, we showthat ICP0 blocks interferon regulatory factor IRF3- and IRF7-mediatedactivation of interferon-stimulated genes and that the RINGfinger domain of ICP0 is essential for this activity. Furthermore,we demonstrate that HSV-1 modifies the IRF3 pathway in a mannerdifferent from that of the small RNA viruses most commonly studied.

INTRODUCTION

Top

Introduction

Infection of cells by viruses elicits a potent and immediatecellular response aimed at limiting virus replication and spread.The hallmark of this response is the production and secretionof interferons (IFNs), a family of pleiotropic cytokines withantiviral, antiproliferative, and immune modulatory functions(37, 45, 69, 79, 80, 86, 92). Once produced, IFNs ${alpha}$ and ßare secreted from virally infected cells, bind to their cognatereceptors on surrounding cells, and initiate a cascade of phosphorylationevents of specific Jak/Stat molecules, culminating in translocationof a phosphorylated Stat-containing complex into the nucleus.This complex binds the IFN-stimulated response element (ISRE)within the promoter region of interferon-stimulated genes (ISGs)and initiates gene transcription (44, 46, 98).

While IFNs ${alpha}$ and ß possess multiple activities, theywere initially discovered and are best known for their antiviralproperties. The classical IFN-induced antiviral pathways includethe double-stranded RNA-dependent protein kinase R, 2',5'-oligoadenylatesynthetase, and Mx pathways, which lead to protein synthesisinhibition, RNA degradation, and inhibition of viral replication,respectively (31, 47, 77, 79). The interferon response can alsoinduce apoptosis (see reference 31 for a review).

It has become clear over the last several years that ISGs canalso be activated by viral infection or double-stranded RNA,a by-product of viral infection, in the absence of IFN production(2, 90, 94-96, 100). For example ISG56, an IFN-stimulated proteininvolved in translational regulation via binding to eukaryoticinitiation factor 3 (33), has been shown to be upregulated byIFN, double-stranded RNA, or virus through independent pathways(34). Recent work has shown that IRF3, a member of the interferonregulatory factor (IRF) family of transcription factors (36,53, 78, 88), can bind directly to ISRE elements and induce ISGsin the absence of IFN production (49, 96, 97, 99).

The current model of virus-induced IRF3 activation consistsof five sequential steps: (i) phosphorylation by a cellular"virus-activated" kinase(s), (ii) a conformational change resultingin protein dimerization, (iii) translocation to the nucleus,(iv) association with CBP and/or p300 coactivators, and (v)ISRE binding (41, 49, 82). Virus infection activates the cellularDNA-dependent protein kinase DNA-PK, leading to the phosphorylationof IRF3 on Thr135 followed by IRF3 nuclear localization (38).However, the role of N-terminal IRF3 phosphorylation is unclear,as other studies have found that virus-mediated IRF3 dimerization,nuclear translocation, and cofactor binding require C-terminalphosphorylation (49, 51, 99). The cellular kinases TBK-1 andIKK ${varepsilon}$ were found to phosphorylate IRF3 and IRF7 following infectionwith Sendai virus (29, 83). While multiple residues within theC-terminal region of IRF3 are phosphorylated (49, 99), to dateonly Ser396 has been found to be required for virus-mediatedIRF3 activation in vivo (81).

Previously, we and others reported that herpes simplex virustype 1 (HSV-1), a large enveloped DNA virus, triggers a hostcellular response characterized by the induction of a specificset of genes, primarily ISGs, resulting in the activation ofan antiviral state in an IFN-independent fashion (60, 67). Thiscellular response is mediated by HSV-1 virion particles andis inhibited by virus replication, suggesting that a newly synthesizedviral protein(s) inhibits the response. The related herpesvirushuman cytomegalovirus (HCMV) induces expression of ISGs in boththe presence and absence of viral gene expression, althoughsignificantly more ISGs are induced when viral gene expressionis inhibited, again suggesting that a newly made virus productsuppresses the host response (7, 100, 101).

Recently, the HCMV virion protein pp65 was shown to inhibitantiviral gene expression by antagonizing NF- ${kappa}$ B and IRF1 pathways(6). With HSV-1, virus binding and penetration are essentialfor antiviral state induction (60, 71), whereas with HCMV, solubleglycoprotein B appears sufficient to trigger the response (5,84). The IFN-independent induction of ISGs by HSV-1 and HCMVdoes not require the Jak-Stat IFN pathway and appears to involveIRF3 (5, 60, 66, 67, 71).

Recently, the HSV-1 immediate-early protein ICP0 was shown toblock ISG induction in a proteasome-dependent fashion (14).The mechanism by which ICP0 blocks ISG induction remains tobe elucidated and is the focus of this study. ICP0 is expressedvery early in infection and performs a variety of functions(21). ICP0 is a promiscuous activator of both viral and cellularpromoters and appears to function synergistically with anotherimmediate-early protein, ICP4 (28, 74, 93). During viral infection,ICP0 has been found to associate with a number of cellular proteins,including elongation factor 1 ${delta}$ (39), cyclin D3 (40), kinetochoreprotein CENP-C (23), ubiquitin-specific protease 7 (USP7, alsoknown as HAUSP) (26, 59), and PML, the prototypic member ofnuclear domains known as ND10, PML bodies, or PODs (25, 55,56).

Considering that many ND10 constituents, including PML, areupregulated by IFN and that a number of viruses, including HSV-1,localize to and subsequently disrupt ND10, one hypothesis hasbeen that ND10 functions as "nuclear defense" sites (54, 57).Of particular interest is that ICP0 affects the SUMO-1 modificationof ND10 components and their subsequent proteasome-mediateddegradation (10, 24, 25, 55, 65). While a number of functionalregions of ICP0 have been identified, an N-terminal cysteine/histidinemotif termed the RING finger appears to be important for thebiological activities of ICP0 (21). Although the RING fingerdomain resembles a zinc finger DNA binding domain, it does notappear to bind DNA (22, 27) and instead acts as a ubiquitinE3 ligase that is required for the degradation of several cellularproteins, including PML (4).

While ICP0 may function to counteract host defense mechanismsthrough disruption of ND10, we and others have shown that ICP0renders HSV-1 relatively resistant to the effects of IFN, inpart by counteracting an IFN-induced block to virus transcription(35, 61, 63). In the current study, we set out to address themechanism by which HSV-1 blocks the IFN-independent antiviralresponse initiated following entry of virus particles. We havefound that ICP0 and vhs, the virion host shut-off protein, counteractthe cellular IFN-independent antiviral response at early andlate times postinfection, respectively. ICP0 specifically blocksIRF3- and IRF7-mediated cellular responses and requires theRING finger for this activity. Since HSV-1 does not appear toactivate the IRF3 pathway in the same fashion as the small RNAviruses most commonly studied, the precise mechanism of ICP0inhibition of IRF3 activity remains to be determined. Thesestudies, however, expand our knowledge on both cellular antiviralresponses to virus infection and the countermeasures adoptedby viruses to ensure their successful replication and spread.

MATERIALS AND METHODS

Top

Materials and Methods

Cells and viruses. Human embryonic lung (HEL) fibroblasts and U2OS osteosarcomacells were maintained in Dulbecco's modified Eagle's mediumsupplemented with 10% fetal bovine serum, while Vero, E5 (13),V27 (73), and A549 cells were maintained in Dulbecco's modifiedEagle's medium supplemented with 5% fetal bovine serum. 293HEK cells were maintained in ${alpha}$ minimal essential medium supplementedwith 10% fetal bovine serum. HSV-1 strains KOS (wild type),d22lacZ (ICP22^-) (52), N38 (ICP47^-) (91), ${Delta}$ ICP6 (ICP6^-) (30),and ${Delta}$ sma (vhs^-) (72) were grown in Vero cells, n212 and 7134(ICP0^-) (8), KM110 (ICP0^-VP16^-) (64), and FXE (ICP0 RING fingermutant) (18) were grown in U2OS cells, d120 (ICP4^-) was grownon E5 cells (13), and 5dl1.2 (ICP27^-) was grown on V27 cells(73). Newcastle disease virus (NDV, strain La Sota) and Sendaivirus (strain Cantell) were kindly provided by E. Nagy and J.Hiscott, respectively. All HSV-1 infections utilized a multiplicityof infection of 10 PFU per cell (as determined in the cell lineused for propagation) unless otherwise specified. NDV and Sendaivirus were used at 40 hemagglutination activity units per 10⁶cells. UV inactivation was performed with a UV Stratalinker2400 (Stratagene) for the length of time required to reduceviral titers by a factor of 10⁵ (data not shown).

Creation and characterization of 7134 ${Delta}$ sma (ICP0^- vhs^-) double mutant virus. U2OS cells were coinfected with 5 PFU per cell of both 7134and ${Delta}$ sma viruses. The ICP0 coding region in 7134 is replacedwith the lacZ gene, allowing color selection with the additionof 5-bromo-4-chloro-3-indolyl-ß-D-galactopyranoside(X-Gal) (100 µg per ml) in the overlay (1% agarose inDulbecco's modified Eagle's medium plus 10% fetal bovine serum).Individual blue plaques were isolated, and following amplificationDNA was isolated from virally infected cells as previously described(64) for use in PCR and Southern blot analyses. PCR was usedto detect an internal SmaI-SmaI deletion within the ${Delta}$ sma vhsgene as previously described (62).

Southern blot analysis to ensure that both copies of ICP0 retainedthe 7134 mutation was performed as previously described (8,64). Briefly, an ICP0 fragment spanning the entire open readingframe was hybridized to DNA digested with PstI and SstI. ICP0resides in a genomic PstI fragment and is not cleaved by SstI.lacZ, however, contains an internal SstI site, and thus recombinant7134 ICP0 is cleaved into two fragments, whereas wild-type ICP0is not. Positive clones were plaque purified three times onU2OS cells, followed by a final round of PCR and Southern blotanalysis.

RT-PCR and Western blot analysis. Total cellular RNA was harvested with Trizol (Gibco-BRL) accordingto the manufacturer's instructions. For reverse transcription(RT)-PCR, aliquots (2 µg) were reverse transcribed with200 ng of random hexamer primer and Superscript II reverse transcriptase(Invitrogen). PCR was subsequently performed as per the manufacturer'sspecifications with the following primers: human ISG56 forwardand reverse (5'-ACGGCTGCCTAATTTACAGC-3' and 5'-AGTGGCTGATATCTGGGTGC-3')and human glyceraldehyde-3-phosphate dehydrogenase forward andreverse (5'-CGGAGTCAACGGATTTGGTCGTA-3' and 5'-AGCCTTCTCCATGGTGGTGAAGAC-3').

Preparation of whole-cell extracts and immunoblot analyses. For preparation of whole-cell extracts, cells were washed twiceand then harvested in cold phosphate-buffered saline, followedby centrifugation at 200 x g for 3 min at 4°C. Cell pelletswere resuspended in whole-cell extract buffer [20 mM HEPES,pH 7.4, 100 mM NaCl, 10 mM ß-glycerophosphate, 0.2%Triton X-100, 50 mM NaF, 1 mM sodium orthovanadate, 1 mM phenylmethylsulfonylfluoride, 2 mM dithiothreitol, and 1X protease inhibitor cocktail(Sigma)], lysed on ice for 15 min, and then centrifuged for10 min at 13,000 x g at 4°C. Extract concentrations weredetermined with a Bradford assay kit (Bio-Rad), and the indicatedamounts of extracts were run on 9% polyacrylamide gels, transferredto nitrocellulose, and probed with antibodies specific to IRF3(Santa Cruz SC-9082), IRF7 (Santa Cruz SC-9083), ß-actin(Santa Cruz SC-1616), CBP (Santa Cruz SC-369), TBK1 (Santa CruzM-375), ICP0 (Goodwin Institute), or ISG56 (provided by G. Sen)at a dilution of 1:1,000. To visualize DNA-PK, nuclear extractswere prepared as previously described (1), and 25 µg ofextracts was run on gels containing 8% acrylamide and 0.07%bisacrylamide. Western blot analysis was performed with a 1:2,000dilution of an anti-DNA-PK monoclonal antibody (provided byS. Lees-Miller).

Luciferase assays. HEK 293 cells were transfected with pRLTK, RANTES-pGL3, or IFNB-pGL3reporter constructs along with expression constructs encodingconstitutively activated IRF3 (IRF3-5D) or IRF7 (IRF7- ${Delta}$ 247-467)(48-50). At 8 h posttransfection, cells infected with 40 hemagglutinationactivity units of Sendai virus per 10⁶ cells or left uninfected,as indicated. Luciferase activity was analyzed 24 h posttransfectionby the dual-luciferase reporter assay (Promega) as per the manufacturer'sinstructions. Relative luciferase activity was measured as theactivation relative to the basal level of reporter gene in thepresence of pFLAG-CMV2 vector after normalization with cotransfectedactivity. Values represent the mean ± standard deviationfor three experiments.

RESULTS

Top

Results

ISG induction in response to virus particles is abolished by ICP0 and vhs. Previously, we used microarray analysis to show that HSV-1 particlesinduce ISGs in an IFN-independent fashion (60). In these studieswe found that ISG induction was masked following virus replication,yet we failed to identify a potential viral inhibitor usingHSV-1 mutants individually deleted for an immediate-early geneproduct. Subsequently, DeLuca and colleagues determined thatICP0 was capable of inhibiting ISG induction through the useof HSV-1 mutants deleted for several immediate-early gene productsalong with an adenovirus expressing ICP0 (14). In our originalstudy, we assayed mRNA accumulation by Northern blot analysis18 to 24 h postinfection, because ISG mRNA was readily detectableat this time with either genetically inactivated virus (KM110)or UV-inactivated wild-type HSV-1.

To reconcile our results with those of DeLuca and colleagues,we assayed ISG56 mRNA and protein levels at various times postinfectionwith a panel of single immediate-early mutant viruses in humanembryonic lung (HEL) fibroblasts. As shown in Fig. 1A, asidefrom UV-inactivated wild-type virus, only the virus deletedfor ICP0 demonstrated ISG56 protein accumulation at 12 h postinfection,in accordance with the work of DeLuca and colleagues. Interestingly,RT-PCR analysis with the ICP0 mutant n212 consistently yieldedISG56 message at early but not late times postinfection (Fig.1B), in accordance with our previous results. Similar resultswere obtained with two additional ICP0 mutants, 7134 (referto Fig. 2) and dlX3.1 (75) (data not shown), where ISG56 mRNAaccumulation peaked at 6 h postinfection and then declined rapidly.In the absence of virus gene expression, using either UV-inactivatedHSV-1 or genetically inactivated KM110, we observed sustainedISG56 mRNA levels at late times postinfection. These data suggestthat in the absence of ICP0, ISG56 mRNA accumulates at earlytimes postinfection but subsequently disappears due to the presenceof an additional virus-induced activity.

View larger version (57K):
[in this window]
[in a new window]
FIG. 1. Expression of ISG56 is only observed in the absence of ICP0 expression. (A) Western blot analysis of whole-cell extracts (40 µg) of HEL fibroblasts harvested 12 h postinfection with the indicated viruses. The ISG56 protein is only visible following infection with UV-inactivated wild-type virus (strain KOS, WT-UV) or an ICP0 null mutant (n212). (B) RT-PCR analysis of total RNA harvested from HEL fibroblasts at various times postinfection with the indicated viruses. In the absence of ICP0, transient ISG56 mRNA accumulation is seen, whereas in the absence of HSV-1 replication (UV-inactivated KOS or genetically inactivated KM110), sustained ISG56 mRNA accumulation is observed. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

View larger version (84K):
[in this window]
[in a new window]
FIG. 2. Deletion of vhs and ICP0 results in sustained ISG56 mRNA accumulation. (A) Construction of a double mutant virus deleted for vhs and ICP0. U2OS cells were coinfected with ${Delta}$ sma and 7134 to create a double mutant virus as outlined in Materials and Methods. (Left panel) PCR analysis was used to confirm the internal SmaI-SmaI fragment deletion that defines the vhs ${Delta}$ sma mutation. BstEII-cut lambda DNA was used as a marker. (Right panel) Southern blot analysis with a radiolabeled ICP0 probe was used to distinguish wild-type ICP0 and 7134 ICP0 sequences based on the presence of a unique SstI site within the lacZ gene that replaces the ICP0 sequence in the 7134 mutation. (B) RT-PCR analysis of total RNA harvested from HEL fibroblasts at various times postinfection with the indicated viruses. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

HSV-1 particles contain within their tegument the virion hostshut-off protein (vhs) that either is, or is part of, an RNasethat causes the degradation of both viral and cellular mRNAs(15, 72, 87). Thus, we assayed whether this protein was responsiblefor the decline in ISG56 mRNA levels at later times postinfection.For this experiment, we created a double mutant virus by recombiningthe 7134 ICP0 null virus (8) with the ${Delta}$ sma vhs-deleted virus(72) (Fig. 2). Due to an internal deletion of a SmaI-SmaI fragment,the ${Delta}$ sma and wild-type vhs alleles are easily distinguishableby PCR analysis, while Southern blot analysis distinguisheswild-type and 7134 ICP0 alleles based on the introduction ofan SstI restriction site within the lacZ gene that replacesICP0 coding sequences in the 7134 mutant (8) (Fig. 2A).

An RT-PCR time course analysis of HEL fibroblasts infected with ${Delta}$ sma, 7134, or 7134 ${Delta}$ sma was performed in order to determine thecontributions of ICP0 and vhs activities in the accumulationof ISG message. As demonstrated above with n212 (Fig. 1B), ISG56mRNA accumulation following infection with 7134 was transient,peaking at approximately 6 h postinfection (Fig. 2B). In contrast,sustained ISG56 mRNA was observed upon deletion of both ICP0and vhs, indicating that vhs activity contributes to the lossof ISG56 mRNA at late times postinfection. This vhs-inducedactivity was only observed in the absence of ICP0, consistentwith the observation that ICP0 prevents ISG induction followingvirus entry (14). Accordingly, we failed to detect ISG56 mRNAaccumulation in fibroblasts infected with ${Delta}$ sma, as this viralmutant contains wild-type ICP0. Similar results were found byRT-PCR with primers specific for other ISGs originally identifiedby microarray analysis (data not shown).

ICP0 RING finger inhibits IRF3- and IRF7-mediated ISG induction. IFN-independent ISG induction following virus infection mayoccur through the activation of the constitutively expressedtranscription factor IRF3 due to its ability to bind to theISRE element within ISG promoters (32, 36). In addition, Prestonand colleagues showed that in the absence of virus gene expression,HSV-1 induces the DNA binding activity of IRF3 (71). In a seriesof transient-transfection assays with a luciferase reporterunder the control of two different ISRE-containing ISG promoters,RANTES and IFN-ß, we found that expression of ICP0significantly dampened ISG promoter activation in response toSendai virus infection and expression of constitutively activatedforms of both IRF3 [IRF-3(5D)] and IRF7 [IRF-7( ${Delta}$ 247-467)] (48-50)(Fig. 3). In all transient transfection assays, a dose-responsecurve was performed with the ICP0 plasmid (data not shown).

View larger version (25K):
[in this window]
[in a new window]
FIG. 3. ICP0 blocks virus-, IRF3-, and IRF7-mediated activation of ISG-specific promoters. Luciferase assays in 293 cells were used to measure activation of the RANTES (A) or IFN-ß (B) promoters in response to infection with the paramyxovirus Sendai virus or expression of constitutively activated forms of IRF3 [IRF-3(5D)] and IRF7 [IRF-7( ${Delta}$ 247-467)] in the presence or absence of full-length ICP0. Results represent the mean ± standard deviation of three independent experiments.

To determine which region of ICP0 was responsible for this inhibition,constructs containing ICP0 deletion, truncation, and point mutationswere utilized (Fig. 4). We first assayed a series of ICP0 truncationmutants for their ability to inhibit the activation of the IFN-ßpromoter by Sendai virus or constitutively activated forms ofIRF3 and IRF7. As shown in Fig. 5A, successive truncation ofthe C terminus to residue 312 had no effect on the inhibitoryactivity of ICP0. However, the truncation mutant 1-241 failedto inhibit activation of the IFN-ß promoter by virus,IRF3, or IRF7. Along with the observation that mutant 105-519effectively repressed IFN-ß promoter activation, thesedata suggest that the region between residues 105 and 312 isimportant for ICP0-mediated inhibition.

View larger version (22K):
[in this window]
[in a new window]
FIG. 4. Schematic representation of ICP0 functional domains and mutations used in this study. The uppermost line depicts the 775-amino-acid ICP0 open reading frame (thick line) with the positions of the introns (thin angled lines) and the amino acid coordinates at the intron-exon junctions marked. These sequences are contained within the ICP0 expression plasmid p111 (19). Also marked are the positions of the first and last cysteine residues of the RING finger, the nuclear localization signal (nls) (residues 501 to 510) (17), the USP7 interaction domain (residues 594 to 633) (58), and the sequences required for self-multimerization and efficient localization at ND10 (residues 633 to 720) (11, 55, 58). Deletion mutants 1-519, 1-388, and 1-312 express C-terminally truncated proteins and were constructed by utilizing the MluI, XmaI, and NruI restriction sites at the positions of these 3' truncations. Mutant 105-519 expresses the coding region between the XhoI and MluI sites. Mutant ins150 contains a 4-amino-acid insertion in codon 150 (p110F1) (19), mutant K144E has a lysine to glutamic acid substitution in the helix region of the RING finger (16), and deletion mutants ${Delta}$ 594-632, ${Delta}$ 106-149, ${Delta}$ 162-188, and ${Delta}$ 9-76 have the stated deletions (p110D12, p110FXE, p110D22, and p110D1, respectively) (17). Truncation mutant 1-593 has stop codons inserted into codon 594 (p110E52X) (58). Finally, mutant 1-241 contains only the sequences encoded in exons 1 and 2 (p110262) (25). Mutants ${Delta}$ 106-149, ${Delta}$ 162-188, and ins150 have lost ubiquitin E3 ligase activity both in vitro and in vivo, while mutant K144E retains in vitro activity but appears to have reduced activity in vivo (4, 20).

View larger version (43K):
[in this window]
[in a new window]
FIG. 5. ICP0 RING finger is both necessary and sufficient to block virus-, IRF7-, and IRF3-mediated ISG promoter activation. Luciferase assays in 293 cells were used to measure the activation of the IFN-ß promoter following Sendai virus infection or expression of constitutively activated IRF3 [IRF-3(5D)] and IRF7 [IRF-7( ${Delta}$ 247-467)] in the presence or absence of full-length and mutant forms of ICP0 (refer to Fig. 4). Results represent the mean ± standard deviation of three independent experiments.

Within this region of ICP0 is the RING finger motif, which hasbeen shown to be responsible for many of the biological activitiesof ICP0. The RING finger is a zinc-stabilized fold that requiresa histidine and several cysteine residues (of which the firstand last are at residues 116 and 156) to form a structure thathas catalytic E3 ligase activity (3, 4, 20). To assess the roleof the RING finger motif in inhibiting IRF3- and IRF7-mediatedIFN-ß promoter activation, we next tested a seriesof ICP0 mutants harboring deletions, insertions, and point mutationsboth within and exterior to the RING finger domain (Fig. 5B).

Deletion mutant ${Delta}$ 106-149 (the lesion in virus FXE) removes severalof the essential metal-coordinating residues and thus abolishesthe RING finger activity, while the mutation in mutant ins150affects a critical alpha helix within the RING fold with similarfunctional consequences (4, 20). Mutation K144E lies withinthe same helix and causes reduced catalytic activity in transfectedand infected cells (20). Deletion ${Delta}$ 162-188 removes sequencesthat are part of the RING domain fold (3) and produces a phenotypethat is indistinguishable in all available assays from thatof a core RING deletion (20). These four ICP0 mutants failedto inhibit virus-, IRF3-, and IRF7-mediated activation of theIFN-ß promoter. Consistent with the data in Fig. 5A,deletions ( ${Delta}$ 9-76 and ${Delta}$ 594-632) and point mutations (R623L/K624I)outside of the RING finger region had no adverse affect on theinhibitory activity of ICP0. Similar results were obtained withthe RANTES promoter construct (data not shown).

ICP0 does not induce the degradation of known IRF3 pathway components. DeLuca and colleagues demonstrated that the proteasome inhibitorMG132 prevented ICP0 from inhibiting ISG induction (14), suggestingthat the function of ICP0 is to induce the proteasomal degradationof a member(s) of the ISG induction pathway. To determine whetherICP0 blocks ISG induction by inducing the degradation of knownIRF3 pathway components, we determined the levels of endogenousTBK-1, IRF3, IRF7, and CBP protein following infection of HELfibroblasts for 10 h with wild-type HSV-1 (strain KOS), n212,and the RING finger mutant FXE (Fig. 6). Due to the ${Delta}$ 106-149deletion of ICP0 within the FXE virus, the ICP0 protein migratesslightly faster than its wild-type counterpart (top panel).ISG56 protein was readily observed following infection withn212 and FXE, consistent with previous results (Fig. 1A and5B, respectively). Actin served as a control for protein loadingand nonspecific protein degradation, while DNA-PK served asa positive control for ICP0-mediated degradation, as this nuclearprotein was shown to be specifically degraded by wild-type ICP0(43).

View larger version (59K):
[in this window]
[in a new window]
FIG. 6. ICP0 fails to degrade known components of the IRF3 pathway. Western blot analysis of whole-cell extracts (40 µg) or nuclear extracts (25 µg, for DNA-PK) of HEL fibroblasts harvested 10 h postinfection with the indicated viruses. Wild-type ICP0 fails to induce the degradation of IRF3, CBP, or TBK-1, components of the cellular antiviral pathway, whereas DNA-PK is specifically degraded by wild-type but not mutant ICP0. In the absence of wild-type ICP0, ISG56 protein is observed, whereas expression of IRF7 is only detected with a recombinant adenovirus expressing human IRF7 (AdIRF7). Actin was used as an internal loading control.

We failed to detect a significant decrease in the level of endogenousTBK-1, one of two kinases shown to phosphorylate IRF3 followingvirus infection (29, 83). As expression of the second kinase,IKK ${varepsilon}$ , is thought to be restricted to activated leukocytes (70),we did not test for endogenous levels of this kinase in HELfibroblasts. However, we did observe that expression of wild-typeICP0 along with IKK ${varepsilon}$ in a transient transfection assay failedto decrease levels of the IKK ${varepsilon}$ protein (data not shown). We alsofailed to detect a decrease in the endogenous levels of IRF3and CBP. IRF7 is not constitutively expressed in cells, andwe failed to detect significant induction of this protein atthe time of harvest (10 h postinfection). However, in subsequenttime course analyses, we frequently detected low levels of IRF7at late times (12 to 24 h) postinfection with wild-type HSV-1(data not shown). Infection with an adenovirus overexpressingIRF7 was used as a positive control for IRF7 detection. Thus,ICP0 does not appear to block ISG induction by inducing thedegradation of known IRF3 pathway constituents. Further workis required to identify the mechanism of ICP0 action with respectto the proteasome and ISG inhibition.

HSV-1 does not induce the hyperphosphorylation of IRF3. In the current model of virus-mediated IRF3 activation, viralinfection induces the phosphorylation on multiple C-terminalserine and threonine residues of IRF3, followed by nuclear translocationand proteasome-mediated degradation (49, 82, 99). These studieshave primarily utilized small, related RNA viruses such as Sendaivirus and Newcastle disease virus, members of the paramyxovirusfamily of negative, single-stranded RNA viruses, in transformedcell lines such as A549 lung epithelial cells. In the previousexperiment, we failed to detect significant degradation of IRF3following HSV-1 infection.

To determine if replicating HSV-1 induces the hyperphosphorylationof IRF3, we performed Western blot analysis on whole-cell extractsharvested from A549 cells and HEL fibroblasts infected withKOS, UV-inactivated KOS, n212, and NDV. Results from A549 cellsare presented in Fig. 7. Newcastle disease virus induced thehyperphosphorylation of IRF3 between 6 and 12 h postinfection,illustrated by the appearance of slower-migrating species ofIRF3. We also noted the accompanying degradation of IRF3 followinghyperphosphorylation in response to Newcastle disease virusinfection, as previously reported in the literature (49). FollowingHSV-1 infection, however, we failed to detect a similar levelof IRF3 hyperphosphorylation. Instead, we routinely observeda shift from the lower to the upper form of IRF3. While thisshift was routinely observed in both A549 cells and HEL fibroblasts,the shift was consistently easier to detect in A549 cells andwas primarily observed at later times (18 to 24 h postinfection)in HEL fibroblasts (data not shown).

View larger version (37K):
[in this window]
[in a new window]
FIG. 7. HSV-1 fails to hyperphosphorylate IRF3. Western blot analysis of whole-cell extracts (40 µg) of A549 cells harvested 6 and 12 h postinfection with the indicated viruses. Endogenous IRF3 is present as two species and becomes hyperphosphorylated and subsequently degraded following infection with the paramyxovirus Newcastle disease virus (NDV). A shift from the lower to the upper form of IRF3 is readily observed following infection with wild-type and ICP0 null viruses but not in the absence of virus gene expression.

The apparent lack of hyperphosphorylation characteristic ofsmall RNA viruses was not due to the action of ICP0, since n212failed to demonstrate a similar level of IRF3 phosphorylation.Of interest, although UV-inactivated KOS consistently failedto elicit a similar shift from the lower to the upper form ofIRF3, n212 mediated this posttranslational modification withkinetics similar to the wild-type virus. A similar shift wasobserved with three additional ICP0 mutant viruses (data notshown). Thus, HSV-1 replication induces a cellular functionthat is distinct from that mediated by RNA viruses but doesnot rely on the activity of ICP0. Studies are currently underway to delineate the mechanism by which HSV-1 modifies IRF3following productive infection.

DISCUSSION

Top

Discussion

Our current working hypothesis of HSV-1 pathogenesis is thatincoming virus particles stimulate an antiviral response thatinduces the direct induction of ISGs through the activationof IRF3 (12). Following virus replication, HSV-1 abolishes thiscellular response through the activities of ICP0 and vhs. Atearly times postinfection, ICP0 appears to specifically inhibitthe IRF3 pathway via the RING finger domain, while vhs presumablyfunctions in a nonspecific fashion later in infection. Sincevhs mutants that produce ICP0 fail to accumulate detectablelevels of ISG mRNA, while ICP0 mutants that produce vhs transientlyaccumulate ISG mRNA, it is likely that ICP0 represents the primarymechanism utilized by HSV-1 to counteract the cellular antiviralresponse to incoming virus particles. A related herpesvirus,HCMV, appears to function in a similar fashion. Using microarrayanalysis, Shenk and colleagues demonstrated that while bothwild-type and UV-inactivated HCMV induce ISGs, significantlymore ISGs were induced following infection with UV-inactivatedvirus, indicating that virus replication produces a proteinthat functions to suppress the host antiviral response (100,101). Furthermore, both HSV-1 and HCMV particles were foundto induce the DNA binding activity of IRF3 (66, 71).

ICP0 is a multifunctional protein to which diverse activitieshave been ascribed (21). With respect to virus pathogenesisand the antiviral response, ICP0 appears to play an importantrole in both IFN-dependent and -independent pathways. Classicalantiviral pathways involve the production of IFNs, primarilyIFN- ${alpha}$ and IFN-ß. Along with ICP34.5, ICP0 renders HSV-1relatively resistant to the effects of these IFNs, in part byovercoming an IFN-induced block to virus transcription (35,61, 63). The cellular antiviral response to incoming particles,however, was shown to be IFN independent (60, 67).

Recent studies with the IRF family of transcription factorshave found that IRF3 and IRF7 are involved not only in virus-inducedproduction of type I IFNs but also in the direct induction ofISGs through binding to ISRE elements in ISG promoters (36,88). Given the observations that HSV-1 particles induce IRF3DNA-binding activity and that ICP0 blocks ISG induction, itwas not surprising that ICP0 was found to specifically blockIRF3- and IRF7-mediated activation of two ISRE-containing ISGpromoters. The ICP0 inhibitory effect on activated IRF7 wasnot as great as that observed following virus infection or withactivated IRF3. Since IRF3 is constitutively expressed whileIRF7 induction is dependent on IFN signaling (88), the preferentialability of ICP0 to inhibit IRF3 would be predicted given itsapparent role in overcoming early antiviral responses, includingthose not mediated by IFN.

However, it remains to be determined exactly how ICP0 blocksIRF3-mediated antiviral responses. The initial work of DeLucaand colleagues found that ICP0 functions via a proteasome-dependentpathway (14). Known components of the virus-induced IRF3 pathway,including TBK-1, IKK ${varepsilon}$ , IRF3, and CBP, were not degraded followingthe production of ICP0. During these studies, however, we discoveredthat HSV-1 does not appear to activate the IRF3 pathway in thesame fashion as the RNA viruses most commonly studied. First,several groups studying paramyxovirus infection have reportedthat virus replication is required for IRF3 and/or IRF7 C-terminalhyperphosphorylation and subsequent activation (85, 89). Notonly does HSV-1 replication mask ISG induction, we failed toobserve IRF3 hyperphosphorylation following infection with eitherwild-type or ICP0 null viruses. In contrast, we routinely observeda shift from the lower band to the upper band of the endogenousIRF3 doublet, which has been observed previously in the contextof N-terminal phosphorylation of IRF3.

Work from the laboratories of Hiscott and Howley demonstratedthat N-terminal phosphorylation occurs in response to stressstimuli and involves DNA-PK (38, 82), a cellular kinase involvedin DNA double-strand break repair and V(D)J recombination (42).However, the role of N-terminal IRF3 phosphorylation duringvirus infection remains to be determined, as the majority ofpublished data suggest that C-terminal IRF3 phosphorylationis essential (49, 51, 81, 99).

Interestingly, ICP0, through its RING finger domain, has beenfound to induce the proteasome-dependent degradation of DNA-PK(43, 68). It is unclear at this time, however, how ICP0, DNA-PK,and IRF3 may be related, particularly with respect to the IFN-independentantiviral state. Also unclear are the differences in the activationof the IRF3 pathway between large DNA and small RNA viruses.Specifically, although we failed to detect IRF3 hyperphosphorylationcharacteristic of infection with paramyxoviruses, replicatingHSV-1 appeared to posttranslationally modify IRF3 in some fashionin an ICP0-independent manner. Additionally, in the transienttransfection assays, ICP0 was able to block activation of twoISRE-containing ISG promoters in response to paramyxovirus infection.Thus, although HSV-1 and paramyxovirus infection may differentiallymodify components of the IRF3 pathway, a common event in thispathway is inhibited by ICP0.

While the precise mechanism of blocking IRF3-mediated activationremains to be elucidated, it is clear that the RING finger motifof ICP0 is critical. While sequences in the C-terminal regionof ICP0, including the nuclear localization signal, the USP7-bindingdomain and the ND10 localization region, are important for ICP0function, the N-terminal RING finger domain appears to be importantfor most biological activities of ICP0 (21). In this study,we found that the region containing the RING finger domain appearsto be both necessary and sufficient for the ICP0-mediated suppressionof the IRF3 pathway, at least within our experimental design.

Although similar to zinc finger DNA binding motifs, the ICP0RING finger fails to bind DNA in solution, either alone or inthe context of full-length ICP0 (22, 27) and it is now clearthat the RING finger of ICP0 constitutes a typical ubiquitinE3 ligase domain (4). The RING finger is essential, however,for localization to and the subsequent disruption of ND10 (25,55). Since many ND10 constituents are upregulated by IFN, onehypothesis is that ND10 serve as an intracellular defense mechanismand that ICP0 functions to overcome this defense. Recent workwith mouse fibroblasts that are deficient in PML has indicatedthat, in contrast to the PML-positive control cells, pretreatmentof PML^-/- cells with IFN had little effect on the replicationof ICP0-negative HSV-1 (9). Of interest, CBP has been foundto localize to ND10 via its interaction with PML (76).

Although ICP0 induces the proteasome-dependent degradation ofboth SUMO-1 modified and unmodified forms of PML, the levelsof CBP appear to remain constant throughout infection. We havefound, however, that following infection with wild-type virusor transfection with full-length ICP0, the punctate nuclearstaining of CBP, characteristic of ND10 constituents, changesto a diffuse nuclear staining (K. Laryea and K. Mossman, unpublisheddata). This difference in nuclear localization is not observed,however, with ICP0 null or RING finger mutant virus infection.How these observations collectively fit in to the mechanismof ICP0 suppression of the cellular antiviral response is unclearat the moment, but warrants further investigation.

In conclusion, we have shown that HSV-1 responds to the IRF3-mediatedcellular response to incoming virus particles by subvertingthe IRF3 pathway through an ICP0-specific mechanism. This mechanisminvolves the ICP0 RING finger and the proteasome but does notappear to induce the degradation of known signaling components.In addition, the RNase activity associated with vhs furtherensures that ISG mRNA fails to accumulate, due to its abilityto nondiscriminately degrade both viral and cellular mRNA. Thus,HSV-1 elicits at least two mechanisms to ensure successful virusreplication in light of the rapid induction of an immune response,characterized by the direct upregulation of both intracellularand secreted immune modulatory functions. These studies serveto increase our knowledge of both the immune response to pathogeninfection and the strategies used by viruses to overcome thesebarriers and will ultimately permit the design of effectivetherapeutics to suppress viral infection.

ACKNOWLEDGMENTS

We thank E. Nagy, J. Hiscott, S. Lees-Miller, and G. Sen forreagents and D. Cummings and K. Laryea for technical assistance.

The Canadian Institutes of Health Research sponsored this work.R.L. was supported by a Fonds de la Recherche en sante du Quebec(FRSQ) Chercheur Boursier. The Medical Research Council supportsthe work in the laboratory of R.D.E.

FOOTNOTES

* Corresponding author. Mailing address: Department of Pathology & Molecular Medicine and Centre for Gene Therapeutics, McMaster University, 1200 Main St. West, Hamilton, Ontario, Canada L8N 3Z5. Phone: (905) 525-9140. Fax: (905) 522-6750. E-mail: mossk{at}mcmaster.ca.

REFERENCES

Top