The Pestivirus N Terminal Protease Npro Redistributes to Mitochondria and Peroxisomes Suggesting New Sites for Regulation of IRF3 by Npro

The N-terminal protease of pestiviruses, Npro is a unique viral protein, both because it is a distinct autoprotease that cleaves itself from the following polyprotein chain, and also because it binds and inactivates IRF3, a central regulator of interferon production. An important question remains the role of Npro in the inhibition of apoptosis. In this study, apoptotic signals induced by staurosporine, interferon, double stranded RNA, sodium arsenate and hydrogen peroxide were inhibited by expression of wild type Npro, but not by mutant protein Npro C112R, which we show is less efficient at promoting degradation of IRF3, and led to the conclusion that Npro inhibits the stress-induced intrinsic mitochondrial pathway through inhibition of IRF3-dependent Bax activation. Both expression of Npro and infection with Bovine Viral Diarrhea Virus (BVDV) prevented Bax redistribution and mitochondrial fragmentation. Given the role played by signaling platforms during IRF3 activation, we have studied the subcellular distribution of Npro and we show that, in common with many other viral proteins, Npro targets mitochondria to inhibit apoptosis in response to cell stress. Npro itself not only relocated to mitochondria but in addition, both Npro and IRF3 associated with peroxisomes, with over 85% of Npro puncta co-distributing with PMP70, a marker for peroxisomes. In addition, peroxisomes containing Npro and IRF3 associated with ubiquitin. IRF3 was degraded, whereas Npro accumulated in response to cell stress. These results implicate mitochondria and peroxisomes as new sites for IRF3 regulation by Npro, and highlight the role of these organelles in the anti-viral pathway.


Introduction
Viruses employ many strategies to establish persistent infection of their host. Most involve inhibition of innate immune responses, including inhibition of Type I interferon and inhibition of apoptosis [1,2]. Pestiviruses, which include classical swine fever virus (CSFV) and bovine viral diarrhea virus (BVDV) generate persistent infections as both viruses block interferon synthesis and apoptosis [3]. CSFV and BVDV are members of the Flavivirus family, with a single positive-stranded RNA genome of 12 kb which encodes a single large polyprotein. N pro is a unique autoprotease, located at the N-terminus which cleaves itself from the polyprotein soon after translation on ribosomes [4]. Our previous work has shown that this single viral protein expressed alone in cells can block Type 1 interferon production by promoting the loss of IRF3 [5], to which it binds directly through a zinc binding TRASH motif [6], leading to IRF3 degradation in proteasomes [7]. Both these properties of N pro have generated useful biological tools; the autoprotease activity has been used to enhance refolding and expression of fused peptides and proteins during bacterial protein expression [8] and the direct antagonism of IRF3 has been used to generate cell lines deficient in IRF3 to block antiviral responses to study viral replication [9].
Recent work has linked the central role played by IRF3 in signaling both interferon production and apoptosis [10], however little is known about the IRF3 signaling pathways targeted by N pro to inhibit the apoptotic pathway, important in generating persistent viral infections. IRF3 is activated by dsRNA interaction with Toll-like receptors and the RNA helicases family (RLRs), including Rig-I and mda5. Binding of viral dsRNA to RNA helicases results in the activation of a CARD domain in Rig-I and mda5, allowing them to associate with the mitochondrial adaptor protein MAVS/IPS-1 [11,12]. MAVS acts as a platform to recruit several cytosolic protein kinase complexes to mitochondria, including I kappa B kinase and tank binding kinase-1, which phosphorylate IRF3, resulting in IRF3 translocation to the nucleus to bind to interferon-sensitive response elements (ISRE) and turn on interferon stimulated genes (ISGs) [13]. In addition, IRF3 translocates to mitochondria through its interaction with the proapoptotic protein Bax to activate the intrinsic apoptotic pathway [14]. This and the fact that MAVS is located predominantly on the outer mitochondrial membrane have highlighted the importance of mitochondria in host innate immune signaling [15]. Interestingly, MAVS has also been found to localize to peroxisomes, [16] a cytosolic organelle important in lipid oxidation, which under normal physiological conditions proliferates by fission [17]. However, recent evidence has described peroxisomes originating from the endoplasmic reticulum and exchanging proteins with mitochondria, forming an ER sub-compartment called the pre-peroxisomal reticulum [18], but this is a matter of considerable debate [17]. The peroxisome pool of MAVS was shown to induce an immediate antiviral response of interferon stimulated genes, such as viperin, ISG15 and ISG56, but not type I interferon itself [16].
Given that the mechanism and site of action of N pro is poorly understood, we investigated its cellular distribution when expressed alone in cells following induction of cellular stress, since virus entry is known to activate the cellular stress pathway to induce the innate response [19]. Initially, N pro relocated to tubular mitochondria, and the pro-apoptotic protein Bax remained cytoplasmic with the subsequent inhibition of apoptosis. Mutations in N pro that prevented binding and loss of IRF3 did not inhibit stress-induced apoptosis, and Bax redistributed to spherical, fragmented mitochondria, revealing the inhibition involved the inactivation of Bax by the loss of IRF3. Interestingly, both N pro and IRF3 were recruited to mitochondria and also to spherical structures identified as peroxisomes, and led to an increased accumulation of ubiquitin associated with peroxisomes. The results show the prevention of apoptosis by N pro depends on its capacity to bind and eliminate IRF3, supporting a direct role for IRF3 in initiating the intrinsic mitochondrial pathway of apoptosis. N pro redistributed to cellular platforms important for IRF3 signaling, implicating mitochondria and peroxisomes as new sites for IRF3 regulation by N pro .

Cells and Reagents
Mouse embryonic fibroblast (MEFs), HeLa and Mardin-Darby Bovine Kidney (MDBK) cells were obtained from European Collection of Cell Cultures (ECACC). MEF and HeLa cells were maintained in DMEM with GlutaMAX with 10% fetal calf serum. MDBK cells were grown in MEM and 10% BVDV-free media with Glutamax (Invitrogen). Cells were seeded on coverslips in 24well plates and infected with BVDV Kyle non-cytopathic strain of BVDV (from J. Brownlie, Royal Vet College, South Mimms UK). After absorption of virus for 1 hour, cells were fed with fresh media and incubated overnight at 37 C before treatment or transfection with DNA. Reagents were from Sigma unless other stated. Cells were treated with sodium arsenate (NaA) at 100 uM, staurosporine at 20 nM, hydrogen peroxide at 200 mM. Interferon alpha (PBL assay science) was used at 100 U/ml, and poly I:C (Roche) was transfected into cells at 100 mg/ml using lipofectamine (Invitrogen).

Plasmids and Transfections
N pro cherry was generated from pcDNA3 N pro [5] and cloned in-frame into pcDNA3mCherry (Invitrogen). N pro mCherry point mutants C112R and D136N were created using the Quick Change kit (Stratagene). Plasmids were transfected into cells with jetPRIME (Polyplus) and cell lines stably expressing N pro mCherry or mutant proteins were selected with G418 (1 mg/ml, Invitrogen). Plasmids encoding tBid-GFP [20] and BAX -GFP (from Stephen Tait, Beatson Institute of Cancer Sciences, University of Glasgow) and IRF3-GFP (from John Hiscott, MacGill University, Canada) were transfected into cells using Fugene HD (Roche). MDBK cells were transfected with plasmids encoding Bax-GFP using jetPRIME (Polyplus) according to the manufacturer's instructions.

Protein Extraction and Western Blotting
Protein was extracted with M-PER (Pierce) for 30 minutes on ice in the presence of HALT protease inhibitor cocktail (Pierce).
Protein concentrations were determined using the BCA protein assay kit (Pierce) according to the manufacturer's protocols. Equal amount of proteins were separated on 4-12% SDS-PAGE gradient gels, and analyzed by blotting onto Immobilon PVDF membranes using antibodies to IRF3 (Abcam Ab25950) and antibodies to N pro , produced in rabbits immunized with synthetic peptides for N pro conjugated to OVA and KLH as described previously [5]. Equal protein loading was confirmed by staining with anti-beta-actin monoclonal antibody AC-15 (Sigma A5441). Primary antibodies were detected using IRDye-labelled secondary antibodies (Li-Cor biosciences) at 1:5000 dilution or HRP conjugated secondary antibodies at 1:500 (Jackson Immunoresearch, Europe). Proteins were visualised either by Odyssey infrared system, or enhanced chemi-luminescence activity. Membranes were stripped and re blotted using Re-blot plus strong solution (106) (Millipore, 2504). Bands were quantitated using the Odyssey software or Scion image software http://scion-image. software.informer.com.

Immunofluorescence and Image Analysis
Cells were fixed in 4% paraformaldehyde and permeabilised with Triton-X100. Cells were stained in 30% goat-gelatin buffer with anti-ubiquitin monoclonal antibody FK2 (Enzo PW8810) or anti-rabbit peroxisome antibody PMP70 (Sigma P0497). Alexa fluorescent conjugated anti-rabbit and anti-mouse secondary antibodies were from Invitrogen and donkey anti-mouse Cy5 was from Jackson Immunoresearch, Europe. For BVDV staining, MDBK cells were fixed in ice cold 80% acetone and BVDV glycoproteins were stained with bovine hyperimmune serum V182 against BVDV (a gift from J. Brownlie, Royal Vet College, South Mimms) and visualized with a secondary anti-cow Cy 3 (Jackson Immunoresearch, Europe). MitoTracker CMXRos (Molecular Probes, Invitrogen) was used at a final concentration of 400 nM and applied to live cells 30 minutes before fixation. DNA was stained with DAPI (Sigma D9542) and mounted with Fluoromount G (Southern Biotech). The fixed cell images were obtained at 663 magnification on a Zeiss Axioplan 2 or a Zeiss laserscanning confocal microscope LSM510. Images were analysed using the Axioplan software version 4.8. For pixel density analysis, images were analyzed using Imaris 664 7.2.3 software (Bitplane Software Incorporated). Punctate structures within the images were identified using the ''spots'' function, and larger structures, such as nuclei, were identified using the ''surfaces'' function within the software. These object layers were used to create the rendered images and to assimilate data concerning vesicle number.

Caspase 3/7 Apoptosis Assay
Following cell treatment, caspase activity was measured in cell lysates using the Apo-ONE Caspase-3/7 Assay (Promega) in 96well plates with the addition of the fluorescent substrate Z-DEVD. Fluorescence was measured using Envision plate-reader (Perki-nElmer). Fold induction of enzyme activity was obtained by dividing the activity in the treated samples by the activity obtained in the untreated cells.

Inhibition of Caspase-dependent Apoptosis by N pro Correlates with Degradation of IRF3
As we have previously shown that virus or N pro expressed alone in cells can promote the degradation of IRF3 [5], we analyzed the levels of IRF3 in cell lines stably expressing wild type N pro fused to m-cherry, or cells expressing N pro carrying mutations in cysteine 112 or aspartic acid 136 residues. We tested these point mutations because previous work has implicated these individual mutations in the loss of IRF3 when introduced into the virus, abolishing IRF3 binding and interferon inhibition [21]. A representative Western blot in Figure 1A shows that IRF3 was undetectable in cells expressing wild type N pro and detectable, but reduced compared to normal, in cells expressing N pro mutations C112R and D136N. Quantitation of IRF3 bands against actin intensity showed expression of N pro C112R or Npro D136N led to approximately 50% of the IRF3 levels seen in control cells ( Figure 1A graph), indicating that there was less efficient degradation of IRF3 compared to cells expressing wild type N pro but the mutants were not completely inactive. The two residues at 112 and 136 have been shown to be part of a site for co-ordination of zinc; however proteins where these residues were mutated were found to still contain up to 26% of control levels of zinc [6].
The ability of N pro , or mutated N pro proteins to protect cells from apoptosis after challenge for 4 hours with inducers of caspase-dependent apoptosis is shown in Figure 1B. Each chemical stimulates a different pathway, but each leads to the activation of caspase 3. Caspase 3/7 activity was measured after 4 hours using luminogenic substrate and results were compared to untreated cells which were normalized to 1. Cells expressing wild type N pro had reduced caspase 3/7 activation by staurosporine, interferon, poly IC, sodium arsenate, and H 2 O 2 compared to control cells not expressing N pro . Cells expressing N pro mutated at positions C112R significantly reversed inhibition of caspase activation when compared to cells expressing wild type N pro . The significance values for caspase activation by each drug treatment for cells expressing viral protein compared to the drug treatment of control cells are shown in the panel below the graph ( Figure 1B). Treatment with sodium arsenate and hydrogen peroxide gave the most significant values for inhibition of apoptosis by N pro . Both chemicals trigger apoptosis via a reduction in mitochondrial membrane potential, cytochrome c release, and caspase activation. Sodium arsenate is also an inducer of the cell stress pathway and we chose this chemical to look at the role of N pro in inhibition of apoptosis, both because mitochondria play a central role in IRF3 signalling and because virus infection activates the cell stress pathway [19].

N pro Prevents Fragmentation of Mitochondria following Cell Stress
It has been shown that apoptosis may be induced during viral infection by IRF3-mediated activation of Bax [14]. We asked whether the mechanism of apoptosis inhibition by N pro involved the intrinsic mitochondrial pathway which is dependent on IRF3 interaction with Bax, by investigating the location of Bax-GFP, tBid-GFP and N pro before and after treatment with sodium arsenate, an inducer of cell stress and apoptosis as shown in Figure 1B above. In normal cells, Bax generally distributed to cytoplasm and a fraction to mitochondria (Figure 2A: a-c), loosely bound to the outer mitochondrial membrane [22,23]. tBid and N pro localised to the cytoplasm and nucleus (Figure 2A: g and k). Following treatment with sodium arsenate, Bax and tBid redistributed to small, punctate perinuclear structures that colocalised with mitochondria identified with Mitotracker (MT) (Figure 2A: d-f and h-j respectively). This mitochondrial fragmentation has been previously shown after cell stress, and it demonstrates that Bax and Bid remain associated following fragmentation, and cells die by apoptosis [23]. In cells expressing N pro alone following treatment with sodium arsenate, N pro also colocalised with mitochondria ( Fig. 2A: l-n), but in contrast to the spherical, perinuclear puncta seen in cells expressing Bax or tBid alone, the mitochondria remained tubular and spread throughout the cytoplasm. N pro expression stabilized the tubular mitochondrial network and the cells were protected from apoptosis. In addition, this shows that since levels of IRF3 are depleted ( Figure 1A), N pro association with mitochondria is independent of its binding to IRF3. In the next experiment, when both Bax-GFP and N pro -cherry were expressed together, Bax distribution remained unchanged after addition of sodium arsenate in the cytoplasm and a fraction on mitochondria ( Fig. 2B: a-b), demonstrating that when cells are stressed, in the presence of N pro and the absence of IRF3, cytoplasmic Bax does not redistribute to spherical fragmented mitochondria and cells are protected from apoptosis. However, in cells expressing the N pro C112R mutant unable to bind and target IRF3 for degradation or protect cells completely against apoptosis (Figure 1), Bax located to spherical perinuclear punctae (Fig. 2B: c), and the cells died. N pro redistributed to mitochondria and cytoplasmic puncta ( Figure 2B: b and d), and the nature of these puncta were further investigated in the next series of experiments.
To investigate Bax localization during virus infection, the experiment was repeated using primary Madin-Darby bovine kidney cells (MDBK cells) infected with bovine viral diarrhea virus (BVDV) Kyle, a non-cytopathic strain. In control MDBK cells treated with sodium arsenate for 4 hours, Bax GFP relocated to spherical perinuclear punctae and cells died by apoptosis, seen by chromatin condensation (Fig. 2B:e-f). In contrast, in MDBK cells infected with BVDV for 24 hours, Bax showed a cytoplasmic and partly mitochondrial distribution ( Fig. 2B: g-h) and cells were protected from apoptosis.
Taken together, the results suggest that following induction of cell stress, as occurs during viral infection, N pro redistributes and associates with mitochondria, not only to manipulate the innate immune pathway but also to promote the survival of cells. Both of these effects are through its interaction with IRF3. That the loss of the pro-apoptotic effect of Bax is an indirect effect of N pro , due to the loss of IRF3, is supported by the fact that following expression of mutant N pro , which does not cause degradation of IRF3, Bax relocates to spherical fragmented mitochondria. It has been found that IRF3 binding to Bax is required for a conformational change in Bax to take place to bind to mitochondria and regulate the permeability of the outer mitochondrial membrane (14). Thus both virus and the single viral protein, N pro inhibit apoptosis through targeting of the Bax mitochondrial apoptotic pathway.

N pro Rapidly Redistributes with IRF3 to Mitochondria and Peroxisomes
To investigate the cellular localisation of N pro expressed together with IRF3 before the disappearance of IRF3, (as would be the case immediately after infection), cells stably expressing N pro -cherry were transfected with IRF3-GFP and visualized by confocal microscopy over a time course of treatment (Figure 3). At 30 minutes post-stimulation, N pro was associated with tubular structures and round punctate structures co-localizing with Mitotracker ( Figure 3A). Mitochondria were seen as tubular and punctate structures. IRF3 was seen in discrete punctate structures associated with N pro (Figure 3A). After 60 mins, N pro remained associated with tubular mitochondria but the size of N pro cytoplasmic puncta increased. IRF3 was almost exclusively colocalised with N pro in larger ovoid puncta in the cytoplasm, or in the nucleus ( Figure 3B).
As a central regulator of interferon signaling, IRF3 and MAVS have been reported to locate to both mitochondria and to peroxisomes in response to cell stress [16] and since this implicates peroxisomes in anti-viral signaling, we investigated whether the IRF3 and N pro puncta detected in Figure 3 were related to peroxisomes. In control cells, peroxisomes were seen by immunostaining with PMP70 as abundant small punctate cytoplasmic spheres and when N pro cherry was co-expressed there was no obvious effect on numbers or sizes of peroxisomes (Figure 4Ai). Following treatment with sodium arsenate, N pro accumulated in cytoplasmic puncta as before and many of these puncta codistributed with peroxisomes (Figure 4Aii). A statistical analysis was performed using Imaris software and the digital rendering of pixel densities in fluorescent punctae for the same cells is shown below (Figure 4Aiii) Figure 4B shows the average number of N pro granules per cell treated with sodium arsenate for 4 hours, and the average number of N pro granules associated with PMP70 per cell. The box plot in Figure 4C is a statistical analysis of the number of puncta containing PMP 70 which also contained Npro ( Figure 4C; box plot) and shows a distribution of 75% in the upper end of the box give a ratio of 0.85 puncta with both N pro and PMP70 compared to the total number of N pro granules. This shows that most of the N pro granules associate with peroxisomes 4 hours after treatment with sodium arsenate.

N pro and IRF3 Co-distribute to Cytoplasmic Peroxisomes Associated with Ubiquitin
Since binding to N pro results in increased degradation of IRF3 by the proteasome [7], we looked for the site of ubiquitin accumulation in cells expressing N pro following stress. In control cells co-expressing N pro and IRF3-GFP together and immunostained for polyubiquitin (Figure 5Ai), IRF3 and N pro were seen throughout the cytoplasm and at a perinuclear site which contained a small amount of ubiquitin. Sodium arsenate rapidly induced the formation of small (block arrow) and large (arrow) N pro puncta positive for IRF3 and both forms localized with ubiquitin (Fig. 5Aii). The small punctate structures were confirmed as peroxisomes using an anti-peroxisome antibody PMP70 (Fig. 5Aiii). The large puncta remain unidentified, but are not peroxisomes. Mutant N pro also redistributed to these puncta (not shown), indicating that redistribution was independent of binding to IRF3, however we cannot exclude that the redistribution of IRF3 to peroxisomes may depend on its binding to N pro . These results indicate that IRF3 and N pro redistribute together to peroxisomes and larger cytoplasmic stress granules where ubiquitin accumulates. Although IRF3 and N pro movement to the peroxisomes correlated with accumulation of ubiquitin, we were unable to show by Western blot whether IRF3-specific ubiquitination occurred in this subcellular organelle or if the rate of IRF3 degradation was increased following sodium arsenate treatment ( Figure 5B, middle panel). Interestingly, however, when we looked at total levels of N pro in cells following cell stress compared to unstressed cells, we found that N pro protein levels increased  ( Fig. 5B: N pro , lanes 3 and 4). Also levels of mutant N pro C112R increased, showing this accumulation was independent of binding to IRF3 (Fig. 5B lanes 5-6). These results demonstrate that peroxisomes are one possible site of IRF3 ubiquitination.

Discussion
Anti-viral signaling mediated by IRF3 leads to apoptosis and increased interferon production. Our previous work has shown that N pro inhibits interferon signaling by accelerating the degradation of IRF3 during virus infection or when expressed alone in porcine cell lines [5]. We show here that these intrinsic properties of the protein are transferable to other species since they are also exhibited when expressed in mouse and human cells. The mutation of residues involved in the co-ordination of zinc led to a partial but not complete reversal of IRF3 degradation. N pro contains a C-terminal TRASH motif [24] consisting of the amino acids C112-C134-D136-C138 which act together to co-ordinate a zinc atom. Mutation of a single amino acid may not completely abolish zinc and therefore IRF3 binding, because D136N mutation still contained 26% of the wild type zinc levels [6], and mutation of two or more residues is necessary to completely inactivate the motif. In this study, various apoptotic signals induced by staurosporine, poly I:C, interferon, sodium arsenate and hydrogen peroxide were inhibited by wild type N pro , but not by the mutant protein, suggesting that the mechanism of apoptosis inhibition is also through IRF3 and that levels of IRF3 in cells expressing mutant N pro were sufficient to promote apoptosis. The subcellular localization of N pro following induction of cellular stress showed that N pro associated with tubular mitochondria, and its expression prevented mitochondrial fragmentation and Bax redistribution, thus inhibiting the Bax-dependent intrinsic apoptotic pathway. Since mutant N pro unable to bind IRF3 did not protect against apoptosis and mitochondria fragmention, it suggests that inhibition of apoptosis is a consequence of reduced IRF3, and not a direct action of N pro , although it does not rule out that this mutation interferes with a direct action of N pro on mitochondrial fission. Our work supports the recent finding that IRF3-dependent activation of Bax is an important part of the intrinsic mitochondrial apoptotic pathway [14], where the BH3 domain of IRF3 binds to cytosolic Bax and induces translocation of Bax to mitochondria to activate apoptosis in response to Sendai virus and double stranded RNA. Here we show that sodium arsenate induction of apoptosis and stress is also dependent on IRF3 interaction with Bax since the pathway is compromised when IRF3 is degraded by N pro . However, the situation may be more complex since transcription of genes regulated by IRF3 and involved in apoptosis will also be decreased. Our results show that Bax redistribution was inhibited in the same manner during BVDV infection. Many large DNA viruses target the mitochondrial pathway, encoding bcl2 homologues to manipulate apoptosis [25], and our work shows smaller RNA virus with limited coding sequence, manipulate both apoptosis and innate immune pathways through targeting mitochondria as an important sites of IRF3 signaling. N pro was localized both to the mitochondrial tubular network, as well as to punctate structures containing the peroxisome marker PMP70 that were in close proximity to mitochondria, but excluded the mitochondrial marker Mitotracker. The peroxisome is emerging as an important platform for regulation of innate immune responses to viral infection. Several other important host anti-viral proteins which localize to peroxisomes include MAVS [26], LSm14A, a sensor for both viral RNA and DNA [27] and also a fraction of Rig-I [27]. The pool of MAVS located to peroxisomes activates the early rapid and transient expression of ISGs that is independent of interferon production [16]. In addition, MAVS recuitment to peroxisomes modulates the NF-kB signaling pathway [28], and notably, we have previously shown that N pro interacts with I-kBa, the inhibitor of NF-kB [29]. Taken together, the peroxisome platform is an important site of N pro function in inhibiting the anti-viral pathway. Viral proteins that have been shown to localize to peroxisomes [30] include VP4 spike protein of rotovirus [31], and interestingly, NS1 the non-structural protein of influenza virus, which binds to a multifunctional peroxisomal enzyme [32]. In common with N pro when transfected into cells, NS-1 showed a punctate immunofluorescence pattern characteristic of peroxisomes and it can also inhibit interferon production [32]. Significantly, there was a rapid relocation of N pro to peroxisomes containing ubiquitin. The colocalisation of IRF3 and N pro with ubiquitin in peroxisomes suggests a link between IRF3 binding to N pro and rapid IRF3 degradation, but the mechanisms remain still unclear. An E3 ligase that has been implicated in specific IRF3 ubiquitination is Ro52 (TRIM21) [33] which, interestingly has been localized to cytoplasmic bodies or speckles. The recruitment of ubiquitin to peroxisomes has been previously shown, for example peroxisome PTS-receptors are ubiquitinated, which is crucial for their function [17] and peroxisomes themselves can be degraded by autophagy, a process termed pexophagy, a process may be involved here. In summary our work shows that the pestivirus protein N pro disrupts the host innate immune response by associating with IRF3 on mitochondria before relocating together to peroxisomes. There is co-localisation of ubiquitin with peroxisomes containing IRF3, which we speculate may lead to enhanced degradation of IRF3. In targeting IRF3 for enhanced degradation, N pro simultaneously inactivates two crucial host defense pathways of apoptosis and interferon, providing an environment for virus persistence.