Foot-and-Mouth Disease Virus 3C Protease Cleaves NEMO To Impair Innate Immune Signaling

Abstract

Foot-and-mouth disease is a highly contagious viral illness of wild and domestic cloven-hoofed animals. The causative agent, foot-and-mouth disease virus (FMDV), replicates rapidly, efficiently disseminating within the infected host and being passed on to susceptible animals via direct contact or the aerosol route. To survive in the host, FMDV has evolved to block the host interferon (IFN) response. Previously, we and others demonstrated that the leader proteinase (L^pro) of FMDV is an IFN antagonist. Here, we report that another FMDV-encoded proteinase, 3C^pro, also inhibits IFN-α/β response and the expression of IFN-stimulated genes. Acting in a proteasome- and caspase-independent manner, the 3C^pro of FMDV proteolytically cleaved nuclear transcription factor kappa B (NF-κB) essential modulator (NEMO), a bridging adaptor protein essential for activating both NF-κB and interferon-regulatory factor signaling pathways. 3C^pro specifically targeted NEMO at the Gln 383 residue, cleaving off the C-terminal zinc finger domain from the protein. This cleavage impaired the ability of NEMO to activate downstream IFN production and to act as a signaling adaptor of the RIG-I/MDA5 pathway. Mutations specifically disrupting the cysteine protease activity of 3C^pro abrogated NEMO cleavage and the inhibition of IFN induction. Collectively, our data identify NEMO as a substrate for FMDV 3C^pro and reveal a novel mechanism evolved by a picornavirus to counteract innate immune signaling.

INTRODUCTION

Foot-and-mouth disease (FMD) severely compromises livestock production, resulting in high economic losses and international restrictions on the export of animals and animal products (22). The etiological agent, FMD virus (FMDV) (genus Aphthovirus, family Picornaviridae), possesses a positive-polarity, single-stranded RNA genome of 8.5 kb that encodes a single polyprotein. Upon cleavage by three virus-encoded proteinases, leader (L^pro), 2A, and 3C^pro (35), the polyprotein is processed into intermediate precursors and fully cleaved individual structural and nonstructural proteins that execute distinct functions in the viral life cycle. To propagate rapidly and efficiently at the initial site of infection, FMDV has developed the ability to counteract the host innate immune response, which is the first line of defense combating viral invasion or replication before more specific protection by the adaptive immune system is established (21, 23, 46).

Innate immune responses are activated through host pattern recognition receptors (PRRs), which recognize molecular structures called pathogen-associated molecular patterns (PAMPs) that are structurally conserved within large groups of pathogens. Upon engagement of PAMPs, PRRs initiate signaling pathways, ultimately triggering the production of type I interferons (IFNs) and inflammatory cytokines and the expression of IFN-stimulated genes (ISGs). These effector proteins collectively suppress the replication of invading pathogens and facilitate the development of adaptive immune responses (28). Sensing of picornavirus is mediated primarily by melanoma differentiation-associated gene 5 (MDA5) (20, 25, 27), but recent evidence has suggested a role for retinoic acid-inducible gene I (RIG-I), as well (39). After sensing cytoplasmic viral RNAs, RIG-I and/or MDA5 interacts with IFN-beta promoter stimulator 1 (IPS-1, also known as MAVS/VISA/Cardif) via the caspase activation and recruitment domain (CARD)-like domains. Subsequently, through a process facilitated by an essential adaptor, nuclear transcription factor kappa B (NF-κB) essential modulator (NEMO) and inhibitors of κB kinase (IKK)-related kinases, such as IKK-α, -β, and -ε, as well as TANK-binding kinase 1 (TBK1), are activated, leading to the activation of various latent transcription factors, including interferon regulatory factor 3 (IRF-3) and NF-κB, through phosphorylation-dependent mechanisms. These transcription factors migrate into the nucleus and directly activate promoters of type I IFNs and inflammatory cytokines/chemokines (26, 58). During coevolution with their hosts, many viruses have acquired mechanisms to circumvent these host innate immune responses. Of note, a number of cysteine proteases encoded by different RNA viruses whose primary functions are to process viral polyprotein have been found to antagonize the induction of type I IFNs (8, 10, 14, 30, 32, 33, 38, 40, 44, 45, 47, 56, 61). Prime examples are the coronaviral papain-like proteases (PLpro), such as severe acute respiratory syndrome coronavirus (SARS-CoV) PLpro (14), human coronavirus NL63 (HCoV-NL63) PLP2 (10), and mouse hepatitis virus A59 (MHV-A59) PLP2 (61); the arteriviral cysteine proteases, nonstructural protein 1 (nsp1) and nonstructural protein 2 (nps2) of the porcine reproductive and respiratory syndrome virus (PRRSV) (30, 44, 45, 47); and the 3C^pro cysteine proteases of several picornaviruses, such as coxsackievirus B3 (CVB3) (38), enterovirus 71 (EV71) (32, 33), and hepatitis A virus (HAV) (40, 56). It remains unclear whether the 3C^pro of FMDV antagonizes the innate immune signaling pathways.

Here, we provide evidence that FMDV 3C^pro impairs RIG-I/MDA5 signaling by cleaving a downstream essential adaptor protein, NEMO. We found that infection of cells with FMDV led to 3C^pro-mediated proteolysis of NEMO. We identified the 3C^pro cleavage site within NEMO and revealed that the cleavage products were severely impaired for signaling to either IRF-3 or NF-κB. These data uncover a novel mechanism evolved by FMDV to disarm host innate immune responses.

MATERIALS AND METHODS

Cells, virus, and chemicals.

Porcine kidney (PK-15) cells were maintained in Dulbecco's modified Eagle medium (DMEM) (Invitrogen) supplemented with 10% heat-inactivated fetal calf serum (FCS), 100 U/ml penicillin, and 10 μg/ml streptomycin sulfate at 37°C with 5% CO₂ in a humidified incubator. Porcine kidney (IBRS-2) cells were cultured in Eagle's minimal essential medium (MEM) supplemented with 10% heat-inactivated FCS, 100 U/ml penicillin, and 10 μg/ml streptomycin sulfates. FMDV (strain O/ES/2001) was propagated in IBRS-2 cells using standard virology techniques, and the supernatants of infected cells were clarified and stored at −80°C. Poly(I·C) (Sigma-Aldrich), a surrogate for viral double-stranded RNA (dsRNA), was used to stimulate the cells. The broad caspase inhibitor z-VAD-FMK and the proteasome inhibitor MG132 were obtained from Beyotime (China).

Plasmids.

The luciferase reporter plasmids IFN-α1-Luc, IFN-β-Luc, 4× PRDIII/I-Luc, and 4× PRDII-Luc have been described previously (50, 52). The luciferase reporter plasmids 4× PRDIII/I-Luc (referred to as IRF-3-Luc) and 4× PRDII-Luc (referred to as NF-κB-Luc) contained four copies of the IRF- or NF-κB-binding motif of the porcine IFN-β promoter, respectively, upstream of the firefly luciferase reporter gene. cDNA expression constructs encoding porcine IPS-1, MDA5, RIG-I, and a constitutively activated form of RIG-I (RIG-I-N) have been described previously (50–52). cDNA encoding porcine NEMO was amplified by standard reverse transcription (RT)-PCR from total RNA extracted from porcine peripheral blood mononuclear cells (PBMCs) and cloned into pCMV-Tag 2B (Stratagene). NEMO-K277A, a constitutively activated porcine NEMO mutant, was generated in a manner analogous to that described for mouse NEMO (7) and cloned into pCMV-Tag 2B. The cDNA expression constructs encoding various FMDV proteins and intermediate precursors used in this study have been described previously (51). Mutagenesis of NEMO (K277A, Q348R, E349R, E354R, E362R, Q365R, Q383R, E390R, E391R, and Q401R) and FMDV 3C^pro (H46Y, C163G, and H181Y) constructs was carried out by overlap extension PCR using specific mutagenic primers (available upon request). All constructs were validated by DNA sequencing.

Luciferase reporter gene assay.

PK-15 cells grown in 24-well plates were cotransfected with 0.1 μg/well of a reporter plasmid, 0.05 μg/well of pRL-TK plasmid (Promega) (as an internal control for normalization of transfection efficiency), and the indicated expression plasmids or an empty control plasmid. Where indicated, cells were further transfected with poly(I·C) (1 μg/well) 24 h after the initial cotransfection. The cells were lysed 12 h later, and firefly luciferase and Renilla luciferase activities were determined using the Dual-Luciferase reporter assay system (Promega) according to the manufacturer's protocol. The data represent relative firefly luciferase activity normalized to Renilla luciferase activity and are representative of three independently conducted experiments. Data are presented as means ± standard deviations (SD). Statistical P values of <0.05 were considered significant, and P values of <0.01 were considered highly significant.

RNA extraction and quantitative real-time RT-PCR.

To determine the effect of 3C^pro on the expression of 2′,5′-oligoadenylate synthetase (2′,5′-OAS); ISG54; IFN-inducible protein 10 (IP-10, also known as CXCL10); regulated upon activation, normal T-cell-expressed and -secreted (RANTES, also known as CCL5); tumor necrosis factor alpha (TNF-α); interleukin-6 (IL-6); and interleukin-1 beta (IL-1β), PK-15 cells in 24-well plates were transfected with 1 μg of empty vector or a plasmid encoding 3C^pro. Twenty-fours later, the cells were mock transfected or transfected with 1 μg of poly(I·C) for 12 h. Total RNA was extracted from the cells using TRIzol reagent (Invitrogen). One microgram of this total RNA was reverse transcribed to cDNA using avian myeloblastosis virus (AMV) reverse transcriptase (Toyobo, Japan), which (1 μl of 20 μl cDNA) was subsequently used in a SYBR green PCR assay (Applied Biosystems). The abundance of individual mRNA transcript in each sample was assayed three times and normalized to that of porcine glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA (as an internal control). The primers were designed with Primer Express software v.3.0 (Applied Biosystems). The sequences of the primers are listed in Table S1 in the supplemental material.

Western blot analysis.

Briefly, PK-15 cells cultured in 60-mm dishes were transfected with the indicated plasmids. After 48 h, the cells were harvested by adding lysis buffer, and protein concentrations were measured in whole cellular extracts. Equal amounts of samples were then subjected to SDS-PAGE and analyzed for expression of RIG-I, MDA5, IPS-1, or NEMO protein by Western blotting using an anti-Flag antibody (Macgene, China). To confirm the expression levels of hemagglutinin (HA)-tagged wild-type (WT) and mutant 3C^pro, an anti-HA antibody (MBL, Japan) was used for immunoblotting. Expression of GAPDH was detected with an anti-GAPDH mouse monoclonal antibody (MAb) (Beyotime, China) to demonstrate equal protein sample loading.

Nucleotide sequence accession numbers.

GenBank accession numbers are as follows: FMDV strain O/ES/2001, AY686687; RIG-I, EU126659; MDA5, EU006039; IPS-1, EU082069; and NEMO, EU258760.

RESULTS

Identification of FMDV 3C^pro as a viral inhibitor of IFN-α/β induction.

To investigate how FMDV regulates type I IFN expression in porcine kidney cells, IBRS-2 cells were cotransfected with a luciferase reporter plasmid under the control of the porcine IFN-β promoter (IFN-β-Luc) and an internal control reporter plasmid (pRL-TK), followed by either mock infection or infection with FMDV at 0.1 PFU/cell. Transfection with poly(I·C), a widely used dsRNA surrogate, strongly activated the IFN-β promoter, indicating IBRS-2 cells harbor intact cytoplasmic pathways to respond to viral dsRNA. In contrast, IFN-β promoter-driven luciferase activity was barely detectable in FMDV-infected cells at any time points examined (Fig. 1A). This suggests that FMDV infection does not induce IFN-β expression and/or it actively inhibits IFN-β induction in IBRS-2 cells. This result was also consistent with those observed in other cell types (11). We and others have previously reported that L^pro (Lb^pro or Lab^pro) of FMDV is an IFN antagonist (11–13, 49, 51, 52). To determine if other FMDV proteins or protein precursors have the ability to inhibit IFN-β expression, we constructed expression vectors encoding the 15 known FMDV proteins or processing intermediates and determined the impact of their ectopic expression on poly(I·C) activation of the IFN-β promoter in transiently transfected PK-15 cells (Fig. 1B). Overexpression of L^pro (Lb^pro or Lab^pro) strongly suppressed poly(I·C)-stimulated IFN-β promoter activity compared to cells transfected with the empty vector (HA), as previously reported (52). Although their effects were less pronounced than that of L^pro, overexpression of 3C^pro or its intermediate precursors (3ABC, 3BC, and 3CD) substantially reduced IFN-β expression (Fig. 1B). In contrast, such an inhibitory effect was not observed in cells expressing 2AB, 2C, 3A, 3AB, 3B, or any of the 4 structural proteins (VP1 to -4). This implies that certain properties of 3C^pro may contribute to the ability of FMDV to inhibit induction of type I IFN transcription. Supporting this notion, we found that FMDV 3C^pro acted in a dose-dependent fashion to inhibit the activation of IFN-β (Fig. 1C) and IFN-α1 (Fig. 1D) promoters by poly(I·C). Furthermore, expression of FMDV 3C^pro also dose-dependently inhibited the upregulation of endogenous transcript for porcine IFN-β (Fig. 1E) and IFN-α1 (Fig. 1F) mRNAs. Collectively, these data demonstrate that FMDV 3C^pro is an antagonist of type I IFNs.

Fig 1.

FMDV 3C^pro independently inhibits type I IFN promoter activation. (A) IBRS-2 cells were cotransfected with IFN-β-Luc and pRL-TK, followed by FMDV infection at 0.1 PFU/cell. As positive controls, 24 h later, the cells were transfected with poly(I·C). Cells were harvested 12 h later and subjected to a dual-luciferase assay. (B) PK-15 cells were transfected with IFN-β-Luc, along with pRL-TK plasmid and the indicated vectors expressing the FMDV proteins (0.8 μg), using Lipofectamine 2000. Twenty-four hours after the initial transfection, the cells were transfected with poly(I·C). Luciferase assays were performed 12 h after the second transfection. (C and D) PK-15 cells were transfected with IFN-β-Luc or IFN-α1-Luc, along with pRL-TK plasmid and increasing quantities (0, 0.125, 0.25, or 0.5 μg) of plasmid encoding 3C^pro, using Lipofectamine 2000. Twenty-four hours after the initial transfection, the cells were transfected with poly(I·C). Luciferase assays and real-time RT-PCR were performed 12 h after the second transfection. The results represent the means and standard deviations of three independent experiments. The firefly luciferase activity was normalized to the Renilla reniformis luciferase, and the untreated empty-vector control value was set to 1. (E and F) RNAs were extracted to measure porcine IFN-β (E) and IFN-α1 (F) mRNA by real-time RT-PCR.

FMDV 3C^pro substantially attenuates the expression of IFN-stimulated genes and inflammatory cytokines/chemokines.

Since FMDV 3C^pro inhibited poly(I·C)-induced IFN-α1/β transcription, we asked whether 3C^pro also attenuates expression of dsRNA-induced ISGs and cytokines/chemokines. To this end, PK-15 cells were transfected with a plasmid encoding FMDV 3C^pro (HA-3C^pro) or the empty control vector. Twenty-four hours posttransfection, the cells were further transfected with poly(I·C) or mock transfected. Total RNA was extracted from the cells and analyzed for the abundance of 2′,5′-OAS, ISG54, IP-10, RANTES, TNF-α, IL-6, and IL-1β mRNAs by SYBR green real-time RT-PCR. Poly(I·C) robustly induced the expression of 2′,5′-OAS (Fig. 2A), ISG54 (Fig. 2B), IP-10 (Fig. 2C), RANTES(Fig. 2D), TNF-α (Fig. 2E), IL-6 (Fig. 2F), and IL-1β (Fig. 2G) in cells transfected with control vector but was substantially less effective in cells expressing 3C^pro. These results show that 3C^pro efficiently impairs intracellular dsRNA-induced IFN-stimulated genes and inflammatory cytokines/chemokines.

Fig 2.

3C^pro significantly reduced the transcription of ISGs and inflammatory cytokines/chemokines. PK-15 cells were transfected with 1 μg of plasmid encoding 3C^pro (HA-3C^pro) or an empty vector (pCMV-HA), and 24 h later, the cells were transfected with 1 μg of poly(I·C). Twelve hours after the second transfection, total RNA was extracted and the expression of 2′,5′-OAS (A), ISG54 (B), IP-10 (C), RANTES (D), TNF-α (E), IL-6 (F), and IL-1β (G) and GAPDH genes was evaluated by quantitative real-time RT-PCR. The results are expressed as increases in mRNA levels relative to those in cells transfected in the absence of poly(I·C) and were normalized by using GAPDH housekeeping gene expression. The results are representative of those from three independent experiments. The error bars indicate standard deviations.

Protease activity governs the ability of FMDV 3C^pro to suppress dsRNA-induced IFN response.

As with the 3C^pro of other picornaviruses, FMDV 3C^pro is a cysteine proteinase responsible for most cleavages within the viral polyprotein (31, 48). For FMDV 3C^pro, residues His-46 and Cys-163 are part of the catalytic triad and His-181 is part of the binding pocket (6, 24). Mutations at any of these residues (Tyr for His-46 [H46Y], Gly for Cys-163 [C163G], or Tyr for His-181 [H181Y]) eliminated the capacity of FMDV 3C^pro to catalytically cleave P1, P2, or P3 substrates in trans (24). We tested the abilities of these 3C^pro mutants to inhibit IFN induction in transiently transfected PK-15 cells. As shown in Fig. 3, compared with control vector-expressing cells, WT 3C^pro strongly suppressed poly(I·C)-induced activation of the IFN-β promoter. In contrast, all three 3C^pro mutants (H46Y, C163G, and H181) were without effect. Thus, 3C^pro depends on its protease activity to disrupt the IFN antiviral response.

Fig 3.

The protease activity of 3C^pro is required to suppress dsRNA-mediated IFN-β induction. PK-15 cells were cotransfected with IFN-β-Luc, along with pRL-TK plasmid, and the designated 3C^pro expression plasmids (1 μg). An empty vector (pCMV-HA) was used as a control. Twenty-four hours after the initial transfection, the cells were transfected with poly(I·C). Luciferase assays were performed 12 h after the second transfection. The error bars indicate standard deviations.

FMDV 3C^pro disrupts RIG-I/MDA5 signaling for activation of both IRFs and NF-κB.

Picornaviruses are sensed primarily by MDA5, but RIG-I may also be involved under some circumstances (20, 25, 27, 39). Given the pivotal role of RIG-I/MDA5 in mediating viral induction of IFN-α/β (26, 58), we further investigated whether overexpression of 3C^pro inhibits RIG-I/MDA5-mediated signaling. Overexpression of a constitutively activated RIG-I mutant (RIG-I-N) or MDA5 significantly activated the IFN-β and IFN-α1 promoters compared with that of the empty-plasmid control (Fig. 4A and B). However, activation of both promoters by RIG-I-N or MDA5 was significantly inhibited in the presence of 3C^pro. Similar results were obtained when these promoters were activated by overexpression of IPS-1, the downstream adaptor for RIG-I/MDA5 (29, 36, 43, 54). Transcription of type I IFNs is dependent on coordinated activation of the latent transcription factors IRF-3 and NF-κB (28). We found that activation of the IRF-3-dependent 4× PRDIII/I promoter and the NF-κB-dependent 4× PRDII promoters by overexpression of RIG-I-N, MDA5, or IPS-1 was significantly reduced by 3C^pro (Fig. 4C and D), suggesting that 3C^pro likely targets a step in the RIG-I/MDA5 pathway prior to the signaling bifurcation into IRFs and NF-κB. Alternatively, 3C^pro may act on two or more signaling proteins in the IRF and NF-κB branches, respectively.

Fig 4.

The FMDV 3C^pro protein disrupts RIG-I/MDA5 signaling. (A) PK-15 cells were cotransfected with IFN-β-Luc, pRL-TK plasmid, and 0.5 μg of plasmid encoding 3C^pro, together with the RIG-I-N, MDA5, or IPS-1 expression vector (0.5 μg). Twenty-four hours after the initial transfection, the cells were transfected with 1 μg of poly(I·C). Luciferase assays were performed 12 h after the second transfection. (B to D) The experiments were performed similarly to those in panel A, except that the IFN-α1-Luc (B), 4× PRDIII/I-Luc (C), or 4× PRDII-Luc (D) promoter reporter plasmid was used. **, P < 0.01. The error bars indicate standard deviations.

FMDV 3C^pro cleaves NEMO.

Our earlier finding that the protease activity was critical for FMDV 3C^pro-mediated inhibition of IFN response (Fig. 3) raised the possibility that 3C^pro may target a signaling component(s) of the IFN-inducing pathway(s) for cleavage. RIG-I/MDA5 signaling to IRFs and NF-κB diverges at NEMO in human and mouse cells (60). In agreement with this, we found that porcine NEMO is also required for NF-κB and IRF activation in PK-15 cells by small interfering RNA (siRNA) knockdown assays (see Table S2 and Fig. S1 in the supplemental material). Because 3C^pro disrupted activation of both IRFs and NF-κB via RIG-I/MDA5 (Fig. 4C and D), we investigated whether NEMO and its upstream signaling molecules, IPS-1, RIG-I, and MDA5, serve as substrates for 3C^pro. Since no commercial antibodies against porcine RIG-I, MDA5, IPS-1, and NEMO were available, we determined whether 3C^pro would cleave the ectopically expressed Flag–RIG-I, Flag-MDA5, Flag–IPS-1, or Flag-NEMO. The protein abundance of full-length Flag-NEMO was reduced in PK-15 cells coexpressing 3C^pro compared with those coexpressing the control vector (Fig. 5A, compare lanes 7 and 8, and B, compare lanes 2 and 3). This was accompanied by the appearance of a faster-migrating anti-Flag-reactive band of ∼48 kDa, presumably a NEMO cleavage product (Fig. 5A, lane 8, and B, lane 2). Importantly, the degree of NEMO cleavage was positively correlated with the expression level of 3C^pro (Fig. 5C). Furthermore, we found that 3C^pro formed a complex with NEMO in coimmunoprecipitation (co-IP) experiments (see Fig. S2 in the supplemental material), suggesting that FMDV 3C^pro associates with NEMO and induces NEMO cleavage. Interestingly, overexpression of 3C^pro also reduced the abundance of constitutively coexpressed Flag-RIG-I, Flag-MDA5, and Flag-IPS-1 (Fig. 5A, compare lanes 2, 4, and 6 to lanes 1, 3, and 5, respectively); however, no cleavage products were detected by an anti-Flag antibody. We also did not find evidence that 3C^pro interacted with RIG-I or MDA5 (see Fig. S2 in the supplemental material). The 3C^pro-mediated reduction in expression of Flag-RIG-I, Flag-MDA5, and Flag-IPS-1 was not investigated further in this study.

Fig 5.

FMDV 3C^pro cleaves NEMO. (A) PK-15 cells were transfected with Flag-tagged RIG-I, MDA5, IPS-1, or NEMO expression plasmid (2 μg), along with 1 μg of plasmid encoding 3C^pro, using Lipofectamine 2000. Cell lysates were prepared 48 h posttransfection and analyzed by Western blotting. (B) PK-15 cells were transfected with Flag-tagged NEMO expression plasmid (2 μg), along with 1 μg of plasmid encoding 3C^pro. Cell lysates were prepared 48 h posttransfection and analyzed by Western blotting. The lane with protein markers (Bio-Rad; catalog no. 161-0376) includes 250-, 150-, 100-, 75-, 50-, 37-, 25-, and 20-kDa molecular-mass bands. (C) PK-15 cells were transfected with Flag-tagged NEMO expression plasmid (2 μg), along with increasing quantities (0, 0.25, 0.5, 1, and 2 μg) of plasmid encoding 3C^pro. Cell lysates were prepared at 48 h posttransfection and analyzed by Western blotting. (D) IBRS-2 cells cultured in 60-mm dishes were transfected with Flag-tagged NEMO (Flag-NEMO) and infected with FMDV (1 PFU/cell) 24 h posttransfection. The cells were lysed 12 h postinfection and analyzed by Western blotting. Flag-NEMO-conjugated protein was verified with mouse anti-FLAG antibody. Rabbit anti-HA was used to confirm the expression of 3C^pro, and mouse anti-GAPDH antibody was used to detect GAPDH, which served as a protein-loading control.

NEMO is cleaved in FMDV-infected cells.

To exclude the possibility that the observed 3C^pro-mediated NEMO cleavage was an artificial effect of plasmid overexpression in cell culture, we analyzed NEMO cleavage in the context of FMDV infection. To this end, we transfected IBRS-2 cells with the Flag-NEMO-encoding vector for 24 h and then infected them with FMDV or, as a control, UV light-inactivated FMDV (UV-FMDV). Compared with uninfected or UV-FMDV-infected cells, the expression level of full-length Flag-NEMO in FMDV-infected cells was significantly reduced, and concomitantly, a NEMO cleavage product appeared (Fig. 5D). These phenomena closely mirrored those observed in cells ectopically expressing 3C^pro (Fig. 5A, B, and C).

3C^pro inhibits NEMO-mediated type I IFN signaling by disrupting activation of IRFs and NF-κB.

To better understand the effect of 3C^pro-mediated NEMO cleavage in type I IFN signaling, we assessed whether expression of 3C^pro disrupts NEMO-dependent type I IFN signaling. However, expression of wild-type NEMO did not significantly activate IRFs and NF-κB to induce type I IFN (Fig. 6A). In previous studies, constitutively activated forms of components in the type I IFN signaling pathway, such as RIG-I (RIG-I-N) and IRF-3 [IRF-3(5D)], were extensively used to study the regulation of IFN signaling by viral protein (4, 42, 59). Thus, we transfected cells with a constitutively activated form of NEMO (NEMO-K277A) to assess whether 3C^pro disrupts NEMO-dependent type I IFN signaling. Compared with WT NEMO, NEMO-K277A was substantially more potent in activating the IFN-β promoter and synthetic IRF- and NF-κB-dependent promoters (Fig. 6A). The K277A substitution did not alter the susceptibility of NEMO to 3C^pro cleavage (Fig. 6B). When transfected into PK-15 cells, NEMO-K277A resulted in constitutive activation of the IFN-β promoter. However, this was inhibited by coexpression of 3C^pro in a dose-dependent manner (Fig. 6C). Similar results were obtained when the cells were transfected with 4× PRDII-Luc or 4× PRDIII/I-Luc in place of IFN-β-Luc (see Fig. S3 in the supplemental material). Taken together, these data show that 3C^pro-mediated cleavage of NEMO-K277A likely impairs the ability of the latter to induce type I IFN expression by disrupting activation of IRFs and NF-κB.

Fig 6.

FMDV 3C^pro disrupts NEMO-mediated type I IFN signaling. (A) PK-15 cells were cotransfected with the indicated promoter reporter plasmid, pRL-TK plasmid, and the wild-type NEMO, NEMO-K277A, or empty expression vector (1 μg). Luciferase assays were performed 36 h after the transfection. (B) PK-15 cells were transfected with Flag-tagged NEMO-K277A expression plasmid (2 μg), along with 1 μg of plasmid encoding 3C^pro. Cell lysates were prepared 48 h posttransfection and analyzed for Flag-NEMO-K277A- and HA-3C^pro-conjugated proteins by Western blotting. Anti-GAPDH antibody was used to detect GAPDH, which served as a protein-loading control. (C) PK-15 cells were transfected with IFN-β-Luc, pRL-TK plasmid, and Flag-tagged NEMO-K277A expression plasmid (0.5 μg), along with increasing quantities (0, 0.02, 0.1, and 0.5 μg) of plasmid encoding 3C^pro. Luciferase assays were performed at 36 h after the transfection. (D) PK-15 cells cultured in 60-mm dishes were transfected with Flag-tagged NEMO expression plasmid (2 μg), along with the indicated 3C^pro expression plasmids (1 μg). Cell lysates were prepared 48 h posttransfection and analyzed for Flag-NEMO- and HA-3C^pro-conjugated proteins by Western blotting. Anti-GAPDH antibody was used to detect beta actin, which served as a protein-loading control. (E) PK-15 cells were cotransfected with Flag-tagged NEMO expression plasmid (2 μg) and 1 μg of plasmids encoding 3C^pro or empty vector. Forty hours after transfection, MG132 or zVAD-FMK was added to a final concentration of 20 nM. Cell lysates were prepared 8 h after treatment and analyzed by Western blotting. (F) PK-15 cells were cotransfected with IFN-β-Luc, pRL-TK plasmid, and Flag-tagged NEMO-K277A expression plasmid (0.5 μg), along with the indicated 3C^pro expression plasmids (0.5 μg). An empty vector (pCMV-HA) was used as a control. Cell extracts were collected 36 h after transfection and analyzed for firefly and Renilla luciferase expression. **, P < 0.01 compared with empty vector plus NEMO-K277A. The error bars indicate standard deviations.

The protease activity of 3C^pro is responsible for NEMO cleavage and the subversion of NEMO signaling.

We next determined whether 3C^pro cleaves NEMO by means of its protease activity. To this end, the abilities of the WT and various catalytically inactive mutants (H46Y, C163G, and H181Y) of 3C^pro to cleave ectopically coexpressed Flag-NEMO-K277A were compared (Fig. 6D). While WT 3C^pro induced NEMO cleavage, none of the 3C^pro mutants was able to do so. Thus, NEMO is a proteolytic substrate for 3C^pro.

A previous study showed that MDA5 is cleaved in poliovirus-infected cells in a way that is dependent on caspases, as well as functional proteasome (2). We considered such a possibility for FMDV 3C^pro-induced cleavage of NEMO. However, neither the broad caspase inhibitor z-VAD-FMK nor the proteasome inhibitor MG132 inhibited 3C^pro-mediated NEMO cleavage (Fig. 6E), indicating that the cleavage of NEMO observed in 3C^pro-expressing cells was independent of the apoptosis and proteasome pathways.

Importantly, all three 3C^pro mutants (H46Y, C163G, and H181Y) no longer inhibited signaling for activation of the IFN-β promoter by ectopically expressed NEMO-K277A (Fig. 6F), nor were they capable of suppressing activation of IRF- and NF-κB-dependent promoters (see Fig. S4 in the supplemental material). This result was consistent with the inability of these 3C^pro mutants to cleave NEMO (Fig. 6D), which may explain our earlier observation that the catalytically inactive 3C^pro mutants lost their capacity to inhibit dsRNA activation of IFN-β transcription (Fig. 3).

3C^pro cleaves NEMO at glutamine 383.

We next examined the sequence of porcine NEMO for potential 3C^pro cleavage sites. Previous studies on 3C^pro substrate specificity documented a preference for glutamine (Gln, Q) or glutamic acid (Glu, E) at the P1 position (6). As the cleavage of N-terminally Flag-tagged NEMO by 3C^pro yielded a slightly shorter than full-length, ∼48-kDa product that could be detected by anti-Flag (Fig. 5B), we focused on a cluster of potential 3C^pro cleavage sites in the carboxyl-terminal region of NEMO (Fig. 7A). We constructed a series of NEMO mutants in which the invariant Gln or Glu at each potential P1 position was replaced with arginine (Arg, R) and examined their cleavage by 3C^pro. A specific mutation, Q383R, blocked the cleavage event, giving rise to only full-length NEMO protein in the presence of 3C^pro (Fig. 7B, lane 8). Escarmis et al. showed that a mutation in the FMDV capsid precursor may affect 3C^pro-mediated proteolytic cleavage at a distant site (17). Therefore, glutamine 383 may be the 3C^pro cleavage site per se, or mutation at that site may affect potential 3C^pro cleavage sites in NEMO. However, 3C^pro-mediated NEMO cleavage was not affected by the Q348R, E349R, E354R, E362R, Q365R, E390R, E391R, or Q401R mutation (Fig. 7B and C), supporting the notion that glutamine 383 is a unique cleavage site targeted by 3C^pro in NEMO. Further corroborating this notion, NEMO-K277A bearing the Q383R mutation also became resistant to 3C^pro cleavage (Fig. 7D).

Fig 7.

FMDV 3C^pro-mediated NEMO cleavage is involved in the inhibition of type I IFN induction. (A) Schematic representation of wild-type porcine NEMO and its derivatives. The domains of porcine NEMO were annotated analogously to those of human NEMO. (B to D) PK-15 cells were transfected with Flag-tagged wild-type NEMO or NEMO mutants as indicated, along with HA-3C^pro or empty vector. Cell lysates were prepared 48 h posttransfection and analyzed for NEMO by Western blotting with an anti-Flag antibody. Western blots with anti-HA antibody show expression of 3C^pro, and Western blots for GAPDH served as a protein-loading control. (E) Schematic representation of constitutively activated NEMO-K277A and its derivatives. (F) PK-15 cells were cotransfected with IFN-β-Luc, pRL-TK plasmid, and either 0.5 μg of plasmid encoding Flag-fused NEMO-K277A (Full), putative 3C^pro-induced cleavage fragments of NEMO-K277A, or empty vector. Cell extracts were collected at 36 h after transfection and analyzed for firefly and Renilla luciferase expression. *, P < 0.05; **, P < 0.01. The error bars indicate standard deviations.

3C^pro-induced NEMO cleavage fragments exhibit impaired type I IFN signaling.

Previous studies have indicated that the carboxyl-terminal zinc finger domain of NEMO is required for full NEMO activation (34, 55, 60). Because 3C^pro targeted NEMO at a specific residue (Q383) between the leucine zipper region and the zinc finger domain (Fig. 7E), we next determined whether either of the 3C^pro-generated cleavage fragments of NEMO remained active. PK-15 cells were transfected with IFN-β-Luc, together with DNA constructs encoding full-length NEMO-K277A or either of the two putative fragments that would result from 3C^pro cleavage (fragments encompassing residues 1 to 383 and 384 to 419, respectively). The N-terminal fragment (1 to 383, bearing the K277A substitution) induced IFN-β promoter activation at a significantly lower level than full-length NEMO-K277A, whereas the C-terminal fragment (384 to 419) was completely defective for such activity (Fig. 7F). Similar results were obtained when the cells were transfected with 4× PRDII-Luc or 4× PRDIII/I-Luc in lieu of IFN-β-Luc (see Fig. S5 in the supplemental material). These data demonstrate that 3C^pro-mediated proteolysis of NEMO results in disruption of innate immune signaling by reducing the availability of full-length NEMO and by generating fragments severely impaired in or deficient for signaling.

DISCUSSION

FMD is one of the most contagious diseases of cloven-hoofed animals. Like many other positive-strand RNA viruses, FMDV is very sensitive to the antiviral actions of IFNs (9, 15, 16, 37, 53, 57). To survive in the host, FMDV has evolved strategies to block IFN expression. Previous studies have revealed that FMDV L^pro, a papain-like proteinase (22), antagonizes the innate immune responses by inhibiting type I interferon production (11–13, 49, 52). L^pro performs this function by acting as a viral deubiquitinase (49), by degrading the NF-κB and IRF-3/7 proteins (12, 52), and by cleaving eIF-4G to shut off host protein synthesis (19). The L^pro-mediated immune evasion is thought to play an important role in FMDV pathogenesis and virulence. In this study, we present evidence that 3C^pro is another FMDV-encoded IFN antagonist. Furthermore, we demonstrate that 3C^pro adopts a mechanism distinct from that of L^pro, i.e., by proteolytically cleaving NEMO, an adaptor protein essential for viral activation of IRFs and NF-κB and, ultimately, IFN synthesis. Not only do our data highlight the multifaceted control of host innate immunity by FMDV, but they uncover a novel mechanism by which FMDV antagonizes innate immune signaling.

Host cells have developed highly specialized mechanisms to detect and combat invading viruses. Among these, the cytoplasmic RIG-I-like receptors (RLR) and membrane-associated Toll-like receptors (TLRs) represent two major classes of PRRs that sense virus infection and initiate signaling cascades culminating in the activation of IRFs and NF-κB and subsequent IFN antiviral response. Not surprisingly, many viruses have evolved countermeasures to disrupt signaling through these pathways, and picornaviruses are no exception. Among the known viral evasion strategies, cleavage of or interaction with receptors or adaptor molecules serves as a powerful means of inactivating antiviral signaling, which has the effect of suppressing common downstream targets of key innate immune signaling pathways. For example, poliovirus, rhinovirus type 1a, and encephalomyocarditis virus induce MDA5 degradation, which in the case of poliovirus is accomplished in a proteasome- and caspase-dependent manner (2). RIG-I is cleaved in cells infected with poliovirus, rhinoviruses, echovirus, and encephalomyocarditis virus (3, 39). Coxsackievirus B and HAV cleave the RIG-I/MDA5 adaptor protein IPS-1, as well as the Toll-like receptor 3 (TLR3) adaptor, Toll–IL-1 receptor domain-containing adaptor-inducing IFN-β (TRIF) (38, 40, 56). TRIF is also reported to be cleaved in EV71-infected cells (33). Most of these events are associated with the 3C^pro proteinase of picornaviruses (3, 33, 38, 40, 56). However, viral cleavage of a signaling adaptor proximal to the IKK complex has not been reported to date.

Results from the current study demonstrate that the 3C^pro of FMDV inhibits innate antiviral signaling by proteolytically cleaving NEMO, an adaptor downstream of the RLR and TLR pathways bridging the activation of IRFs (by the IRF-activating kinases TBK1 and IKKε) and NF-κB (by the classical IKKs, IKKα and IKKβ) (60). We found NEMO cleavage was specifically attributable to the protease activity of 3C^pro and not a result of activation of cellular caspases or proteasome. Importantly, we demonstrated that porcine NEMO was cleaved by 3C^pro in FMDV-infected porcine (the natural host for FMDV) cells, confirming the biological relevance of this finding. Of note, evidence for NEMO as a prime target for microbial pathogens to disrupt host innate immunity is just emerging. The M45 protein of mouse cytomegalovirus interacts with NEMO and redirects it to autophagosomes for subsequent lysosomal degradation (18). The lpaH9.8 protein of Shigella bacteria acts as an E3 ubiquitin ligase to promote polyubiquitination and proteosomal degradation of NEMO (1). However, FMDV 3C^pro-mediated proteolysis of NEMO (revealed in the current study) represents the first example of a viral immune evasion mechanism that involves direct cleavage of NEMO by a viral protease.

The porcine NEMO protein exhibits a high degree (89%) of sequence conservation with its well-characterized human counterpart (41). The most prominent functional domains, including the two coiled-coil motifs (CC1 and CC2), a leucine zipper (LZ) motif, and a zinc finger (ZF) domain (41), are recognized in the porcine counterpart and arranged in the same order as in human NEMO (Fig. 7E). This evolutionary conservation is a strong indicator that human and porcine NEMO proteins likely employ common modes of type I IFN regulation. We found that FMDV 3C^pro specifically targeted the Gln 383 residue (Fig. 7B and D), which lies between the C-terminal LZ motif and ZF domain of NEMO. Thus, NEMO proteolysis by 3C^pro cleaves off the ZF domain, which plays a crucial role in full activation of NF-κB and IRFs that orchestrate immune and inflammatory responses (34, 55, 60). Indeed, we found the NEMO fragments resulting from 3C^pro cleavage were either severely impaired for or completely deficient in activation of NF-κB, IRF, and IFN induction (Fig. 7F; see Fig. S5 in the supplemental material).

Interestingly, we also observed that the abundances of ectopically expressed Flag–RIG-I, Flag-MDA5, and Flag–IPS-1 proteins were reduced in cells coexpressing 3C^pro (Fig. 5A). However, no cleavage products were observed, suggesting these proteins are unlikely to be proteolytic substrates for FMDV 3C^pro. While the nature of this observation remains to be studied, we speculate that it may result from inhibition of cellular translational machinery following 3C^pro-mediated cleavage of the translation initiation factor eIF-4A (5). Regardless, 3C^pro-mediated suppression of cellular protein synthesis, especially that of the antiviral proteins (RIG-I, MDA5, IPS-1, etc.), may contribute to the FMDV disruption of innate immunity. Unfortunately, it is impossible to dissect the activity of 3C^pro toward NEMO cleavage from that toward inhibiting host protein translation, as mutations disrupting the catalytic activity of 3C^pro abrogated not only NEMO cleavage, but also eIF-4A proteolysis (data not shown).

In summary, our data identify 3C^pro as a novel FMDV-encoded IFN antagonist and demonstrate that the 3C^pro-mediated proteolytic cleavage of NEMO is directly involved in the inhibition of type I IFN transcription by FMDV. This adds another layer of complexity to the immune evasion strategies evolved by this economically important viral pathogen. It should be noted, however, that apart from this novel viral countermeasure, 3C^pro may dampen host innate immunity by cleaving the translation initiation factor eIF-4A (5) and thereby shutting off host protein synthesis, resulting in reduced expression of IFN and antiviral ISG proteins. Furthermore, it will be interesting to investigate whether NEMO cleavage is a shared attribute among 3C proteinases of different picornaviruses, which may facilitate the survival of this family of viruses.