Disruption of MDA5-Mediated Innate Immune Responses by the 3C Proteins of Coxsackievirus A16, Coxsackievirus A6, and Enterovirus D68

ABSTRACT

Coxsackievirus A16 (CV-A16), CV-A6, and enterovirus D68 (EV-D68) belong to the Picornaviridae family and are major causes of hand, foot, and mouth disease (HFMD) and pediatric respiratory disease worldwide. The biological characteristics of these viruses, especially their interplay with the host innate immune system, have not been well investigated. In this study, we discovered that the 3C^pro proteins from CV-A16, CV-A6, and EV-D68 bind melanoma differentiation-associated gene 5 (MDA5) and inhibit its interaction with MAVS. Consequently, MDA5-triggered type I interferon (IFN) signaling in the retinoic acid-inducible gene I-like receptor (RLR) pathway was blocked by the CV-A16, CV-A6, and EV-D68 3C^pro proteins. Furthermore, the CV-A16, CV-A6, and EV-D68 3C^pro proteins all cleave transforming growth factor β-activated kinase 1 (TAK1), resulting in the inhibition of NF-κB activation, a host response also critical for Toll-like receptor (TLR)-mediated signaling. Thus, our data demonstrate that circulating HFMD-associated CV-A16 and CV-A6, as well as severe respiratory disease-associated EV-D68, have developed novel mechanisms to subvert host innate immune responses by targeting key factors in the RLR and TLR pathways. Blocking the ability of 3C^pro proteins from diverse enteroviruses and coxsackieviruses to interfere with type I IFN induction should restore IFN antiviral function, offering a potential novel antiviral strategy.

IMPORTANCE CV-A16, CV-A6, and EV-D68 are emerging pathogens associated with hand, foot, and mouth disease and pediatric respiratory disease worldwide. The pathogenic mechanisms of these viruses are largely unknown. Here we demonstrate that the CV-A16, CV-A6, and EV-D68 3C^pro proteins block MDA5-triggered type I IFN induction. The 3C^pro proteins of these viruses bind MDA5 and inhibit its interaction with MAVS. In addition, the CV-A16, CV-A6, and EV-D68 3C^pro proteins cleave TAK1 to inhibit the NF-κB response. Thus, our data demonstrate that circulating HFMD-associated CV-A16 and CV-A6, as well as severe respiratory disease-associated EV-D68, have developed a mechanism to subvert host innate immune responses by simultaneously targeting key factors in the RLR and TLR pathways. These findings indicate the potential merit of targeting the CV-A16, CV-A6, and EV-D68 3C^pro proteins as an antiviral strategy.

INTRODUCTION

Hand, foot, and mouth disease (HFMD) is a common infectious disease among young children. Enterovirus 71 (EV-A71 or EV71) and coxsackievirus A16 (CV-A16 or CA16) are most frequently associated with HFMD worldwide. Extensive studies have focused on EV-A71 because of its association with severe complications involving the central nervous system and significant mortality (1, 2). However, since the CV-A16-related outbreak of HFMD occurred in England in 1994 (3), accumulating evidence has demonstrated that CV-A16 infection can also cause severe neurological complications (4) and death (5, 6). Recently, CV-A6 infection has also been increasingly associated with HFMD outbreaks around the world (7–17). Furthermore, EV-D68 has been linked to severe respiratory disease worldwide in recent years. To date, no effective vaccines or treatments for enterovirus or coxsackievirus infection are available.

The interferon (IFN) induction pathway is commonly targeted by viruses. Host cells orchestrate the production of type I IFNs upon the detection of invading viral pathogens. To antagonize viral invasion, pathogen-associated molecular patterns (PAMPs) are sensed by cellular pattern recognition receptors (PRRs) to activate the type I IFN induction signaling pathway (18, 19). The retinoic acid-inducible gene I (RIG-I)-like receptors (RLRs) are PRRs that play a pivotal role in the innate immune system. The members of the RLR family, including RIG-I, melanoma differentiation-associated gene 5 (MDA5), and Laboratory of Genetics and Physiology 2 (LGP2), are located in the cytoplasm and monitor for the presence of viral RNA (20). RIG-I and MDA5 belong to the family of DExD/H box RNA helicases that contain two caspase activation and recruiting domains (CARDs) at their N termini and a single DExD/H box RNA helicase domain at their C terminal domains (CTDs). Upon viral RNA virus infection, the CTDs of RIG-I and MDA5 sense viral RNA that bears a 5′-triphosphate group and long kilobase-scale genomic RNA, respectively, both of which are lacking in host mRNA (21–26). After binding viral RNA, RIG-I and MDA5 undergo conformational changes and transduce signals to the downstream adaptor MAVS (virus-induced signaling adaptor) (also named VISA, Cardif, or IPS-1) (27–29) through a CARD-CARD interaction. As a result, tumor necrosis factor receptor-associated factors (TRAFs), IκB kinase ε (IKKε), and TANK-binding kinase 1 (TBK1) are recruited by the MAVS signaling complex. IKKε and TBK1 then phosphorylate MAVS, which results in the recruitment of interferon regulatory factor 3 (IRF3) for its phosphorylation. Furthermore, phosphorylated IRF3 (p-IRF3) subsequently forms a dimer and translocates into the nucleus to activate the IFN promoter (30).

For the nuclear factor kappa light chain enhancer of activated B cell (NF-κB) activation pathway, transforming growth factor β-activated kinase 1 (TAK1) is a key player. In mammalian cells, TAK1 forms a complex with TAK1 binding protein 1 (TAB1), TAB2, and TAB3, which can recruit adaptor proteins, such as TRAF6, to activate TAK1. TAK1 activation phosphorylates and activates the IKK complex (IKKβ, IKKα, and NEMO), leading to the activation of NF-κB (31).

CV-A16, along with EV-A71, EV-D68, and CV-A6, is a positive-stranded RNA virus belonging to the Picornaviridae family. Like other members of this family, CV-A16 encodes only one single open reading frame (ORF), which can be processed into four structural (VP1, VP2, VP3, and VP4) and seven nonstructural (2A, 2B, 2C, 3A, 3B, 3C, and 3D) proteins upon viral infection (1). Among them, the EV-A71 3C protease (3C^pro) is one of the most versatile functional proteins. It possesses both proteolytic and RNA binding activities (32, 33), which enable the protease to perform multiple tasks in systems involving viral replication and pathogen-host interactions. Increasing evidence suggests that EV-A71 3C^pro targets innate immune factors, such as RIG-I, TRIF, IRF7/9, the TAK1/TAB1/TAB2/TAB3 complex, and NLRP3, to modulate type I IFN and cytokine responses (34–38). Recently, EV-D68 3C^pro has been reported to target IRF7 and TRIF to disable innate sensing responses (39, 40). However, little is known about the function of the CV-A16 and CV-A6 3C^pro proteins in innate immunity.

Here we demonstrate that the CV-A16, CV-A6, and EV-D68 3C^pro proteins block MDA5-triggered type I IFN induction. CV-A16 3C^pro binds MDA5 and inhibits its interaction with MAVS. In addition, the CVA-16, CV-A6, and EV-D68 3C^pro proteins cleave TAK1 to inhibit the NF-κB response. Thus, our data demonstrate that circulating HFMD-associated CV-A16 and CV-A6, as well as severe respiratory disease-associated EV-D68, have developed a mechanism to subvert host innate immune responses by simultaneously targeting key factors in the RLR and TLR pathways. These findings indicate the potential merit of targeting CV-A16, CV-A6, or EV-D68 3C^pro as an antiviral therapy.

RESULTS

Type I interferon induces potent CV-A16 inhibition.

Type I IFNs are early host responses to virus infection and can induce broad antiviral effects. Whether CV-A16 replication could be influenced by type I IFNs is not fully known. To address this question, we first treated susceptible rhabdomyosarcoma (RD) cells with 1,000 U/ml of IFN-α2 for 12 h. IFN-α2-treated or untreated RD cells were then infected with equal amounts of CV-A16. Viral replication was monitored by the appearance of a cytopathic effect (CPE) in the cell culture. In untreated RD cells, a significant CPE was observed 72 h after infection. As expected, no CPE was observed in uninfected RD cells (Fig. 1A). IFN-α2 treatment resulted in significant viral inhibition, as indicated by the reduced appearance of CPEs in RD cells infected with CV-A16 (Fig. 1A). Reduced CV-A16 replication due to IFN-α2 treatment was also evaluated by monitoring viral RNA levels in infected RD cells. In IFN-α2-treated RD cells, the level of CV-A16 RNA was reduced by >90% compared to that of virus-infected RD cells in the absence of IFN-α2 treatment (Fig. 1B).

FIG 1.

Type I interferon induces potent CV-A16 inhibition. (A) RD cells were untreated (columns 1 and 3) or treated with 1,000 U/ml of IFN-α2 (columns 2 and 4). After 12 h, IFN-α2-treated or untreated RD cells were infected with equal amounts (MOI of 0.5) of CV-A16 (columns 3 and 4) or left uninfected (columns 1 and 2). Seventy-two hours later, cells were observed for morphological changes and photographed by light microscopy at a ×400 magnification. (B) Total viral RNA was extracted from the treated RD cells described above (A). The viral RNA levels of CV-A16 were evaluated by real-time PCR using SYBR green. Primers targeted CV-A16 VP1 to monitor viral replication. GAPDH expression was used as a control. Data represent the averages of results from three independent experiments. The error bars indicate the standard deviations of data from three independent experiments. P values of <0.05 were considered significant. (C) CV-A16 from RD cells was collected at days 0, 1, 2, 3, and 4 after infection with or without IFN-α2 treatment. The virus titer was determined by a CCID₅₀ assay.

Reduced CV-A16 virus production due to IFN-α2 treatment was also evaluated by monitoring the viral titers in infected RD cells at different time points. In IFN-α2-treated RD cells, the viral titer was reduced significantly compared to that of virus-infected RD cells in the absence of IFN-α2 treatment at different time points (Fig. 1C). These data establish that type I IFN can induce potent inhibition of CV-A16 replication.

Type I IFN treatment also resulted in significant viral inhibition of CV-A6 (Fig. 2A) and EV-D68 (Fig. 2B), as indicated by the reduced appearance of CPEs in RD cells infected with these viruses. Reduced CV-A6 (Fig. 2C) or EV-D68 (Fig. 2D) replication due to IFN-α2 treatment was also evaluated by monitoring viral titers in infected RD cells. In IFN-α2-treated RD cells, the CV-A6 or EV-D68 viral titer was reduced significantly at different time points compared to that in virus-infected RD cells in the absence of IFN-α2 treatment (Fig. 2). These data establish that type I IFNs can induce potent inhibition of CV-A16, CV-A6, and EV-D68.

FIG 2.

Type I interferon induces potent inhibition of CV-A6 and EV-D68. (A) RD cells were untreated (left) or treated with 1,000 U/ml of IFN-α2 (right). After 12 h, IFN-α2-treated and untreated RD cells were infected with equal amounts (MOI of 0.5) of CV-A6 (A) or EV-D68 (B). Ninety-six hours later, cells were observed for morphological changes and photographed by light microscopy at a ×400 magnification. (C and D) CV-A16 (C) and EV-D68 (D) from RD cells were collected at days 0, 1, 2, 3, and 4 after infection with or without IFN-α2 treatment. The virus titer was determined by a CCID₅₀ assay.

CV-A16, CV-A6, and EV-D68 suppress the type I IFN response triggered by RNA virus infection.

It has been established that infection with the RNA virus Sendai virus (SeV) can stimulate the expression of interferons in mammalian cell lines (41, 42). Whether CV-A16 infection activates the innate immune system has not been widely explored. Since type I IFNs can induce potent viral inhibition, CV-A16 may have to avoid type I IFN induction during viral infection. To identify how CV-A16 affects the innate immune pathway, we first determined whether it can stimulate a type I IFN response. The IFN-β luciferase (IFN-β–luc) reporter (34) and an internal control pRL-TK renilla reporter (34) were cotransfected into human embryonic kidney 293T (HEK293T) cells with CV-A16 or EV-A71 infectious clones or a control vector plus pT7 RNA polymerase. At 24 h posttransfection, cells were infected with SeV or left uninfected for 18 h. IFN-β promoter activity (as monitored by luciferase activity) was then measured. CV-A16 replication did not trigger IFN-β production (Fig. 3). As shown in Fig. 3A, SeV infection stimulated IFN-β promoter activity (lane 2) almost 50-fold compared to that in uninfected cells (lane 1). In contrast, CV-A16 replication in HEK293T cells showed no stimulation of the IFN-β promoter (Fig. 3A, lane 3). HEK293T cells were sensitive to CV-A16 infection, resulting in the appearance of a CPE (Fig. 3C). At the same time, CV-A16 replication in HEK293T cells inhibited SeV infection-induced IFN-β production (Fig. 3A, lane 4). Consistent with data from previous reports (35), EV-A71 replication also suppressed SeV-triggered IFN-β induction (Fig. 3A, lane 6).

FIG 3.

CV-A16, CV-A6, and EV-D68 suppress the type I IFN response triggered by RNA virus infection. (A and B) HEK293T cells were transfected with the pRL-TK plasmid and the IFN-β (A) or NF-κB (B) promoter plus control plasmids or CV-A16 or EV-A71 infectious clone vectors for 24 h. The cells were subsequently infected with SeV (20 hemagglutination [HA] units/ml) or left uninfected for 20 h, and the cell lysates were then assayed for luciferase activity. Data represent the averages of results from three independent experiments. The error bars indicate the standard deviations of data from three replicates within one experiment. P values of <0.05 were considered significant. Rel.Lucif.Act., relative luciferase activity. (C) HEK293T cells were infected with CV-A16 (MOI of 0.5) or left uninfected. At the indicated time points, cells were observed for CPEs. (D and E) HEK293T cells were transfected with the pRL-TK plasmid and the IFN-β (D) or NF-κB (E) promoter for 12 h and then infected with CV-A6 and EV-D68 (MOI of 0.5) for 12 h. The cells were subsequently infected with SeV (20 HA units/ml) or left uninfected for 20 h, and the cell lysates were then assayed for luciferase activity. Data represent the averages of results from three independent experiments. The error bars indicate the standard deviations of data from three replicates within one experiment. P values of <0.05 were considered significant.

In addition to activating type I IFN pathways, RNA virus infection can also trigger NF-κB activation. We examined NF-κB promoter activity after SeV infection in the absence or presence of CV-A16 or EV-A71. EV-A71 barely activated the NF-κB promoter, which is consistent with data from previous reports (35, 36, 43). Moreover, we revealed that CV-A16 did not trigger NF-κB activation. Similar to their effects on the suppression of IFN-β production, SeV-mediated NF-κB promoter activation was significantly suppressed by CV-A16 and EV-A71 (Fig. 3B, compare lanes 2, 4, and 6). These results indicate that host immune defense systems are impaired during CV-A16 propagation. We also observed that CV-A6 or EV-D68 infection did not trigger type I IFN (Fig. 3D) or NF-κB (Fig. 3E) production. Furthermore, SeV-mediated type I IFN (Fig. 3D) and NF-κB (Fig. 3E) production was significantly suppressed by CV-A6 or EV-D68 infection. Enterovirus 2A^pro cleaves eukaryotic initiation factor 4GI (eIF4GI) and eIF4GII within virus-infected cells (44–46). The CV-A16 2A protease inhibited SeV replication by inhibiting cap-dependent mRNA translation and inhibiting the SeV-induced activation of MAVS signaling. It is thus important to address whether CV-A16 3C^pro expression alone could suppress SeV-triggered immune activation.

3C^pro of CV-A16 suppresses the SeV-induced type I IFN response.

In addition to its role in viral polyprotein processing, 3C^pro proteins of certain enteroviruses have been shown to play a pivotal role in suppressing IFN-β production (35–38, 47–51). However, the role of the 3C^pro proteins of CV-A16, CV-A6, and EV-D68 in suppressing IFN-β activation has not been well characterized. To determine the capacity of CV-A16 3C^pro to block the type I IFN response, IFN-β promoter activity in the absence or presence of CV-A16 3C^pro was assessed by using a luciferase assay. As shown in Fig. 4A, the expression of CV-A16 3C^pro inhibited SeV-induced IFN-β promoter activity significantly. As expected, a similar activity for EV-A71 3C^pro was observed (Fig. 4A) when both proteins were expressed at similar levels (Fig. 4B). Thus, our results indicate that CV-A16 utilizes 3C^pro to antagonize type I IFN induction. Consistent with the IFN-β promoter assay results, endogenous IFN-β mRNA expression was induced by SeV infection and was then reduced in a dose-dependent manner by the addition of CV-A16 3C^pro (Fig. 4C). Type I IFN-stimulated gene 56 (ISG56) mRNA expression was also induced by SeV infection and was then similarly inhibited by the addition of CV-A16 3C^pro (Fig. 4C).

FIG 4.

The 3C^pro proteins of CV-A16, CV-A6, and EV-D68 suppress the SeV-induced type I IFN response. (A and F) HEK293T cells were transfected with an empty vector, CV-A16 3C, or EV-A71 3C (A) or with an empty vector, CV-A6 3C, or EV-D68 3C (F), along with the pRL-TK plasmid and the IFN-β reporter. After 24 h, cells were challenged with SeV (20 HA units/ml) or left uninfected for 20 h and were monitored for luciferase activity. (B) Protein expression was detected by Western blotting. Data represent the averages of results from three independent experiments. The error bars indicate the standard deviations of data from three replicates within one experiment. P values of <0.05 were considered significant. (C) HEK293T cells were transfected with an empty vector, 150 ng CV-A16 3C, and 450 ng CV-A16 3C. After 24 h, cells were challenged with SeV (20 HA units/ml) or left uninfected for 20 h. The cells were then harvested, and RNA was extracted to determine IFNB and ISG56 mRNA levels by quantitative real-time PCR. (D) HEK293T cells were transfected with the Flag-IRF3 plasmid alone (lane 1), Flag–CV-A16 3C alone (lane 2), Flag–EV-A71 3C alone (lane 3), Flag-IRF3 plus Flag–CV-A16 3C (lane 5), or Flag-IRF3 plus Flag–EV-A71 3C (lane 6). Twenty-four hours after transfection, cells were treated with SeV (+) (20 HA units/ml) or were left untreated (−). Twenty-four hours later, cells were analyzed by Western blotting with antibodies against Flag-IRF3, p-IRF3, Flag-3C, or the tubulin control. (E) HEK293T cells were transfected with 200 ng CV-A16 2A (lane 3), CV-A16 3C at two doses (150 ng and 450 ng) (lanes 4 and 5), or 200 ng 2A with 3C at two doses (150 ng and 450 ng) (lanes 6 and 7). Twenty-four hours after transfection, cells were treated with SeV (20 HA units/ml) or were left untreated. Twenty-four hours later, cell lysates were assayed for luciferase activity. (G) HEK293T cells were transfected with the Flag-IRF3 plasmid alone (lanes 1 and 4) or in combination with HA–EV-D68 3C (lanes 2 and 5) or HA–CV-A6 3C (lanes 3 and 6). Twenty-four hours after transfection, cells were treated with SeV (20 HA units/ml) or were left untreated. Twenty-four hours later, cells were analyzed by Western blotting with antibodies against Flag-IRF3, p-IRF3, HA-3C, or the tubulin control.

In response to RNA virus infection, the cytosolic RNA sensors RIG-I and MDA5 initiate antiviral signaling, and consequently, a number of downstream molecules are recruited and activated (18). IFN-β transcription and production require IRF3 phosphorylation and dimerization after virus stimulation. To further confirm the inhibitory effect of CV-A16 3C^pro on type I IFN responses, we assayed for IRF3 phosphorylation. Plasmids encoding CV-A16 3C^pro or EV-A71 3C^pro were cotransfected into HEK293T cells with an IRF3 expression vector for 24 h. Transfected cells were then treated with SeV for another 24 h, followed by Western blot (WB) analysis for phosphorylated IRF3. IRF3 phosphorylation was triggered by SeV infection (Fig. 4D, lane 4). However, IRF3 phosphorylation was inhibited significantly in cells expressing CV-A16 3C^pro or EV-A71 3C^pro (Fig. 4D, lanes 5 and 6). 2A^pro of EV-A71 has also been reported to inhibit innate immune responses (38, 43, 52–56). We observed that CV-A16 2A^pro enhanced the inhibition of CV-A16 3C^pro-mediated type I IFN activation induced after SeV infection (Fig. 4E).

We also observed that CV-A6 or EV-D68 3C^pro inhibited SeV-induced IFN-β promoter activation (Fig. 4F). SeV infection-triggered IRF3 phosphorylation was also inhibited significantly in cells expressing CV-A6 3C^pro or EV-D68 3C^pro (Fig. 4G). Collectively, our results demonstrate that the 3C^pro proteins of CV-A16, CV-A6, and EV-D68 are all able to inhibit type I IFN responses.

3C^pro of CV-A16, CV-A6, or EV-D68 inhibits type I IFN responses upstream of IRF3 activation.

Since CV-A16 3C^pro expression inhibited IRF3 phosphorylation, we first determined whether CV-A16 3C^pro can directly influence the function of IRF3. To address this question, a constitutively active form of IRF3 (IRF3-5D) (42) was used to induce the activation of the IFN-β promoter in HEK293T cells transfected with IRF3-5D alone or cotransfected with CV-A16 3C^pro. Consistent with data from a previous report (37), the ectopic expression of IRF3-5D stimulated IFN-β induction (Fig. 5A). IFN-β activation stimulated by IRF3-5D was not inhibited by CV-A16 3C^pro (Fig. 5A). CV-A16 3C^pro protein expression was confirmed by Western blot analysis (Fig. 5B). IFN-β promoter activation stimulated by IRF3-5D was also not inhibited by CV-A6 or EV-D68 3C^pro (Fig. 5A). IKKε (Fig. 5C) and TBK1 (Fig. 5D) also stimulated IFN-β expression, which was consistent with data from previous reports (35, 57, 58). None of the CV-A16, CV-A6, and EV-D68 3C^pro proteins had an inhibitory effect on IFN-β promoter activation induced by IKKε (Fig. 5C) or TBK1 (Fig. 5C). Collectively, our results demonstrate that 3C^pro of CV-A16, CV-A6, or EV-D68 inhibits the induction of type I IFN responses upstream of IRF3, TBK1, and IKKε.

FIG 5.

The 3C^pro proteins of CV-A16, CV-A6, and EV-D68 inhibit type I interferon responses upstream of IRF3 activation. (A, C, and D) HEK293T cells were transfected with IFN-β–luc and IRF3-5D (A), IKKε (C), or TBK1 (D) together with control (pRL-TK), CV-A16 3C, CV-A6 3C, or EV-D68 3C expression vectors. Thirty-six hours after transfection, cells were assayed for luciferase activity. Data are representative of results from three independent experiments. The error bars indicate the standard deviations of data from three different experiments. P values of <0.05 were considered significant. (B) CV-A16, CV-A6, and EV-D68 3C protein expression was analyzed by Western blotting with HA antibody. Tubulin antibody was used as a control.

CA16 3C^pro neutralizes MDA5-mediated type I IFN activation.

Upon virus infection, MDA5 binds viral RNA and interacts with the adaptor protein MAVS. This interaction leads to the activation of downstream factors (TRAFs and TBK1, etc.). These activated factors cause IRF3 to dimerize and then translocate into the nucleus, inducing an IFN-β response. Therefore, using previously established experimental strategies (59), we asked whether CV-A16 3C^pro could target MDA5 and impair its immune activation functionalities. In contrast to type I IFN activation by IRF3, TBK1, and IKKε, which is resistant to CV-A16 3C^pro, we observed that type I IFN activation induced by MDA5-N (35, 60) was inhibited by CV-A16 3C^pro (Fig. 6A). A CV-A16 3C^pro catalytic-site mutant (H40D) could still inhibit MDA5-N-induced IFN activation (Fig. 6A). The MDA5-N protein level was not downregulated in the presence of CV-A16 3C^pro (Fig. 6B). Endogenous MDA5 (Fig. 6C) and MAVS (Fig. 6D) expression was also not affected by CV-A16 3C^pro. It was reported previously that the EV-A71 and EV-D68 3C^pro proteins induce the cleavage of IRF7 (37, 39). We did not detect a cleaved IRF7 product in the presence of CV-A16 3C^pro (Fig. 6E).

FIG 6.

CV-A16 3C^pro neutralizes MDA5-mediated IFN-β activation. (A) HEK293T cells were transfected with the pRL-TK plasmid and the IFN-β–luc promoter reporter plus a control plasmid (lane 1), Flag–MDA5-N alone (lane 2), Flag–MDA5-N plus three increasing doses of CV-A16 3C (lanes 3 to 5), or Flag–MDA5-N plus three increasing doses of the CV-A16 3C H40D mutant (lanes 6 to 8). (B) After a 36-h transfection, cells were assayed for luciferase activity and then analyzed by Western blotting. Proteins were detected with the indicated antibodies. Tubulin antibody was used as a control. Data are representative of results from three independent experiments with triplicate samples. The error bars indicate the standard deviations of data from three different experiments. P values of <0.05 were considered significant. WT, wild type. (C and D) HEK293T cells were transfected with two increasing doses of CV-A16 3C plasmids (lanes 2 and 3) or the control vector (lane 1). Thirty-six hours later, cells were harvested, and 3C-HA along with endogenous MDA5 (C) or MAVS (D) proteins were analyzed by Western blotting. Tubulin antibody was used as a control. (E) HEK293T cells were transfected with IRF7-Flag and two increasing doses of CV-A16 3C or a negative control. Thirty-six hours later, proteins were detected with the indicated antibodies by Western blotting. (F) HEK293T cells were cotransfected with Myc-MDA5 (lanes 2 and 3) or Myc–MDA5-N (lanes 4 and 5) and with the control vector (lane 1), HA–CV-A16 3C (lanes 2 and 4), or the HA–CV-A16 3C H40D mutant (lanes 3 and 5). Forty-eight hours after transfection, cell lysates were immunoprecipitated with anti-Myc beads. Western blotting was used to analyze protein expression.

We next examined whether CV-A16 3C^pro has the ability to interact with MDA5. A coimmunoprecipitation (co-IP) assay was performed, as previously described (61), to evaluate whether MDA5 associates with CV-A16 3C^pro. As shown in Fig. 6F, CV-A16 3C^pro (hemagglutinin [HA] tagged) was immunoprecipitated with both Myc-tagged MDA5 (lane 7) and Myc-tagged MDA5-N (lane 9). In the absence of Myc-tagged MDA5 or MDA5-N, CV-A16 3C^pro was not detected (Fig. 6F, lane 6), indicating the specificity of the assay system. An interaction of CV-A16 3C^pro (H40D) with MDA5 (Fig. 6F, lane 8) or MDA5-N (lane 10) was also observed.

Although the CV-A16 3C^pro catalytic-site mutant (H40D) could still interact with MDA5, mutations of surface-exposed residues (Fig. 7A) of the N-terminal region of CV-A16 3C^pro (CV-A16 3C^proM) disrupted its interaction with MDA5 (Fig. 7B). CV-A16 3C^proM also had an impaired ability to inhibit MDA5-N-induced type I IFN activation (Fig. 7C). The interaction of MDA5 with CV-A16 3C^pro was not disrupted by RNase treatment (Fig. 7D). The CV-A16 3C^pro V154S mutant, containing a mutation of the C-terminal residue involved in RNA binding (33), also maintained the ability to inhibit MDA5-N-induced type I IFN activation (Fig. 7E).

FIG 7.

The N-terminal domain of CV-A16 is critical for its function. (A) Schematic of CV-A16 3C^pro N-terminal amino acids and the designed mutant. (B) HEK293T cells were transfected with MDA5-N–Myc and HA–CV-A16 3C (lane 2) or the HA–CV-A16 3C-M mutant (lane 3). Forty-eight hours after transfection, cells were harvested, and cell lysates were incubated with anti-HA beads overnight. Protein expression was detected with the indicated antibodies by Western blotting. Tubulin antibody was used as a control. (C) HEK293T cells were transfected with the pRL-TK plasmid and the IFN-β–luc promoter reporter plus a control plasmid (lane 1), MDA5-N–Myc (lanes 2 to 7), or different amounts of HA-3C (lanes 3 to 5) or HA–3C-M (lanes 6 and 7) expression vectors. Thirty-six hours after transfection, cell lysates were assayed for luciferase activity. Data are representative of results from three independent experiments. (D) HEK293T cells were transfected with MDA5-Myc and HA–CV-A16 3C (lanes 2 and 3) or HA–CV-A16 3C alone (lane 1). Forty-eight hours after transfection, cells were harvested, and cell lysates were treated with RNase (lane 3) or left untreated (lanes 1 and 2) and incubated with anti-Myc beads overnight. Protein expression was detected by Western blotting using the indicated antibodies. Tubulin antibody was used as a control. (E) HEK293T cells were transfected with the pRL-TK plasmid and the IFN-β–luc promoter reporter plus a control plasmid (lane 1) or MDA5-N–Myc alone (lane 2) or cotransfected with different amounts of HA-3C (lanes 3 to 5) or HA-3C V154S (lanes 6 to 8) expression vectors. Thirty-six hours after transfection, cell lysates were assayed for luciferase activity. Data are representative of results from three independent experiments. The error bars indicate the standard deviations of data from three different experiments. P values of <0.05 were considered significant.

MDA5 binds viral RNA and interacts with the adaptor protein MAVS to trigger immune activation. Next, we asked whether CV-A16 3C^pro can impair the association of MAVS with MDA5. To answer this question, we carried out a co-IP assay. HEK293T cells were transfected with vectors expressing Flag-MAVS along with Myc–MDA5-N, CV-A16 3C-HA, or a control vector. Forty-eight hours after transfection, cells were lysed and incubated with anti-Flag beads, followed by analysis by Western blotting. As expected, MAVS interacted with MDA5-N (Fig. 8A, lane 8). In the absence of MAVS, MDA5-N was not detected (Fig. 8A, lane 6), indicating the specificity of the assay. The results additionally showed that the association of MAVS and MDA5-N was disrupted by 3C^pro in a dose-dependent manner (Fig. 8A, lanes 9 and 10, and B). Immune activation triggered by MDA5-N was also disrupted by CV-A16 3C^pro in a dose-dependent manner (Fig. 8C). CV-A16 infection inhibited the interaction of MDA5 with endogenous MAVS (Fig. 8D). Interestingly, MDA5 dimerization was also inhibited by CV-A16 3C^pro (Fig. 8E). Dimerization of MDA5 has been reported to be important for the interaction of MDA5 with MAVS (60, 62, 63).

FIG 8.

CV-A16 3C disrupts the interaction of MAVS and MDA5-N. (A) HEK293T cells were transfected with plasmids encoding MDA5-N–Myc (lanes 1 and 3 to 5), MAVS-Flag (lanes 2 to 5), or HA-3C (lanes 1, 2, 4, and 5) or a control vector (lanes 1 to 3). Forty-eight hours after transfection, cells were harvested, and cell lysates were incubated with anti-Flag beads overnight. Protein expression was detected by Western blotting utilizing the indicated antibodies. Tubulin antibody was used as a control. (B) ImageJ software (NIH) was used to quantitate protein band intensities, and the value of MDA5-N that coprecipitated with MAVS in the absence of 3C was set to 1. The data shown are representative of results from three independent experiments. (C) HEK293T cells were transfected with the pRL-TK plasmid and the IFN-β–luc promoter reporter plus a control plasmid (lane 1), MDA5-N–Myc (lanes 3 to 6), or different amounts (50 ng, 150 ng, and 450 ng) of the HA-3C (lanes 2 and 4 to 6) expression vector. Thirty-six hours after transfection, cell lysates were assayed for luciferase activity. Data are representative of results from three independent experiments. The error bars indicate the standard deviations of data from three different experiments. P values of <0.05 were considered significant. (D) HEK293T cells were transfected with MDA5-N–Flag or left untreated for 12 h and then infected with increasing MOIs (2 and 3) of CV-A16 or left untreated. Twenty hours later, cells were harvested, and cell lysates were incubated with anti-Flag beads overnight. Protein expression was detected by Western blotting using the indicated antibodies. Tubulin antibody was used as a control. (E) HEK293T cells were transfected with plasmids encoding MDA5-N–Myc (lanes 1, 3, and 4), MDA5-N–Flag (lanes 2 to 4), or HA-3C (lanes 1, 2, and 4) or a control vector (lanes 1 to 3). Forty-eight hours after transfection, cells were harvested, and cell lysates were incubated with anti-Flag beads overnight. Protein expression was detected by Western blotting using the indicated antibodies. Tubulin antibody was used as a control.

Upon RNA virus infection, the RIG-I-like receptors RIG-I and MDA5 detect viral RNA through their CTDs. The CARDs of RIG-I and MDA5 are responsible for transducing signals to the downstream adaptor MAVS through CARD-CARD interactions (59, 60, 64, 65).To investigate whether CV-A16 3C^pro also antagonizes RIG-I-induced MAVS activation in the RLR pathway, we performed an IFN-β luciferase assay in HEK293T cells. A RIG-IN expression vector, which includes the caspase recruitment domain, was employed (64, 66). As illustrated in Fig. 9A, CV-A16 3C^pro significantly suppressed RIG-IN-mediated IFN-β activation in a dose-dependent manner. The expression of 3C^pro impaired the interaction between MAVS and RIG-IN (Fig. 9B). In addition, the RIG-IN protein level was not downregulated in the presence of 3C^pro. The association of MAVS and RIG-IN was disrupted by CV-A16 3C^pro in a dose-dependent manner (Fig. 9B and C).

FIG 9.

CV-A16 3C^pro disrupts the MAVS–RIG-IN complex. (A) HEK293T cells were transfected with the pRL-TK plasmid and the IFN-β–luc promoter reporter plus a control plasmid (lane 1), RIG-IN–Myc (lanes 3 to 6), or different amounts of HA-3C (lanes 2 and 4 to 6) expression vectors. Thirty-six hours after transfection, cell lysates were assayed for luciferase activity. Data are representative of results from three independent experiments. The error bars indicate the standard deviations of data from three different experiments. P values of <0.05 were considered significant. (B) HEK293T cells were transfected with plasmids encoding RIG-IN–Myc (lanes 1 and 3 to 5), MAVS-Flag (lanes 2 to 5), HA-3C (lanes 1, 2, 4, and 5), or a control vector (lanes 1 to 3). Forty-eight hours after transfection, cells were harvested, lysed, and then incubated with anti-Flag beads overnight. Protein expression was analyzed by Western blotting using the indicated antibodies. Tubulin antibody was used as a control. (C) ImageJ software (NIH) was used to quantitate protein band intensities, and the value of RIG-IN coprecipitated with MAVS in the absence of 3C was set to 1. The data shown are representative of results from three independent experiments.

CV-A16 3C^pro interacted with both RIG-I (Fig. 10B) and RIG-IN (Fig. 10B). Its interaction with RIG-I was not impaired by RNase treatment (Fig. 10C). Like wild-type CV-A16 3C^pro, the CV-A16 3C^pro H40D mutant maintained the ability to inhibit RIG-IN-triggered immune activation (Fig. 10A) and could still interact with RIG-I (Fig. 10B, lane 8) and RIG-IN (Fig. 10B, lane 10). Also, the H40D mutant could still inhibit the interaction of RIG-IN with MAVS (Fig. 10D), similar to its ability to inhibit the interaction of MDA5-N with MAVS (Fig. 10E). Together, these results indicate that CV-A16 3C^pro targets MDA5 and RIG-I, culminating in the inactivation of the IFN-β induction pathway. 3C^pro failed to block IFN-β activation mediated by TBK1 and downstream factors (Fig. 5C). This finding indicates that 3C^pro probably acts upstream of TBK1 in the RIG-I–RLR pathway and that RIG-I, like MDA5, is a target of 3C^pro.

FIG 10.

CV-A16 3C^pro disruption of the MAVS–RIG-IN complex is independent of protease activity. (A) HEK293T cells were transfected with the pRL-TK plasmid and the IFN-β–luc promoter reporter plus a control plasmid (lane 1) and RIG-IN–Myc (lane 2) alone or cotransfected with different amounts of HA-3C (lanes 3 to 5) or HA-3C H40D (lanes 6 to 8) expression vectors. Thirty-six hours after transfection, cell lysates were assayed for luciferase activity. Protein expression was detected by Western blotting using the indicated antibodies. Data are representative of results from three independent experiments. The error bars indicate the standard deviations of data from three different experiments. P values of <0.05 were considered significant. (B) HEK293T cells were transfected with RIG-I–Flag (lanes 2 and 3) or RIG-IN–Flag (lanes 4 and 5) in combination with a control vector, HA–CV-A16 3C (lanes 2 and 4), or the HA–CV-A16 3C H40D mutant (lanes 3 and 5). Forty-eight hours after transfection, cell lysates were immunoprecipitated with anti-Flag beads. Western blotting was used to analyze protein expression. (C) HEK293T cells were transfected with RIG-I–Flag and HA–CV-A16 3C (lanes 2 and 3) or HA–CV-A16 3C alone (lane 1). Forty-eight hours after transfection, cells were harvested, and cell lysates were treated with RNase (lane 3) or were left untreated (lane 2). Cell lysates were then incubated with anti-Flag beads overnight. Protein expression was detected by Western blotting using the indicated antibodies. Tubulin antibody was used as a control. (D) HEK293T cells were transfected with plasmids encoding RIG-IN–Myc with MAVS-Flag (lanes 2 to 5), the HA-3C H40D mutant (lanes 1 and 3 to 5), or a control vector (lane 2). Forty-eight hours after transfection, cells were harvested, lysed, and then incubated with anti-Flag beads overnight. Protein expression was analyzed by Western blotting using the indicated antibodies. Tubulin antibody was used as a control. (E) HEK293T cells were transfected with plasmids encoding MDA5-N–Myc with MAVS-Flag (lanes 2 to 5), the HA-3C H40D mutant (lanes 1 and 3 to 5), or a control vector (lane 2). Forty-eight hours after transfection, cells were harvested, lysed, and then incubated with anti-Flag beads overnight. Protein expression was analyzed by Western blotting using the indicated antibodies. Tubulin antibody was used as a control.

The 3C^pro proteins of EV-D68 and CV-A6 neutralize MDA5-mediated type I IFN activation.

Direct targeting of MDA5 by the Picornaviridae family of viruses has not been widely reported previously. To further investigate whether 3C proteases from other enterovirus species also possess the ability to interact with MDA5 and affect its function, the expression vectors for the EV-A71, EV-D68, CV-A6, and CV-B3 3C^pro proteins were constructed and compared. The interaction of 3C^pro with MDA5-N was first examined by co-IP experiments. The co-IP results (Fig. 11A) indicated that, similarly to CV-A16 3C^pro, the 3C^pro proteins of EV-D68 and CV-A6 also interacted with MDA5-N. At the same time, the EV-A71 and CV-B3 3C^pro proteins interacted with MDA5-N poorly (Fig. 11A).

FIG 11.

The 3C^pro proteins of CV-A6 and EV-D68 neutralize MDA5-mediated IFN-β activation, and enterovirus 3C proteases selectively associate with MDA5-N. (A) HEK293T cells were transfected with plasmids encoding MDA5-N–Myc and HA-3C (lanes 2 to 6) of different species of enterovirus or a control vector (lane 1). Forty-eight hours after transfection, cells were harvested, lysed, and then incubated with anti-HA beads overnight. Protein expression was analyzed by Western blotting using the indicated antibodies. Tubulin antibody was used as a control. (B) HEK293T cells were transfected with green fluorescent protein (GFP); MDA5-N–Flag; and a control vector (lane 1), CV-A6 3C (lane 2), or the CV-A6 3C H40D mutant (lane 3). Thirty-six hours later, cells were harvested and analyzed by Western blotting. (C) HEK293T cells were transfected with GFP; MDA5-N–Flag; and a control vector (lane 1), EV-D68 3C (lane 2), or the EV-D68 3C H40D mutant (lane 3). Thirty-six hours later, cells were harvested and analyzed by Western blotting. (D) HEK293T cells were transfected with the pRL-TK plasmid and the IFN-β–luc promoter reporter plus a control plasmid (lane 1) or the MDA5-N–Myc (lanes 2 and 3) or CV-A6 3C (lane 3) expression vector. Thirty-six hours after transfection, cell lysates were assayed for luciferase activity. Data are representative of results from three independent experiments. The error bars indicate the standard deviations of data from three different experiments. (E) The pRL-TK plasmid and the IFN-β–luc promoter reporter were transfected with a control plasmid (lane 1) or the MDA5-N–Myc (lanes 2 and 3) or EV-D68 3C (lane 3) expression vector into HEK293T cells. Thirty-six hours after transfection, cell lysates were assayed for luciferase activity. Data are representative of results from three independent experiments. The error bars indicate the standard deviations of data from three different experiments.

The level of MDA5-N was lower in the presence of the EV-D68 and CV-A6 3C^pro proteins (Fig. 11A). To determine whether this was due to the protease activity of these 3C^pro proteins, we constructed CV-A6 and EV-D68 3C protease active-site mutants (H40D). The reduced levels of MDA5-N were still noted with these mutants compared to the controls (Fig. 11B and C). Thus, it is unlikely that the reduced levels of MDA5-N in the presence of the CV-A6 and EV-D68 3C^pro proteins are due to protease digestion. Future studies will be required to determine the mechanism of the reduction of the MDA5 level by the CV-A6 and EV-D68 3C^pro proteins. However, despite lower levels of MDA5-N in the presence of the CV-A6 and EV-D68 3C^pro proteins (Fig. 11A, lanes 5 and 6) than in the control sample (Fig. 11A, lane 1), co-IP of MDA5-N with the CV-A6 and EV-D68 3C^pro proteins was still detected (Fig. 11A, lanes 11 and 12) but not in the control sample (lane 7).

To address the question of whether the EV-D68 and CV-A6 3C^pro proteins could also antagonize MDA5-mediated IFN-β activation, we performed an IFN-β luciferase assay in HEK293T cells. The CV-A6 (Fig. 11D) and EV-D68 (Fig. 11E) 3C^pro proteins efficiently suppressed MDA5-N-induced IFN-β promoter activity. Collectively, our data suggest that disabling the interaction of MDA5 and MAVS through the activity of 3C proteases may be a common strategy used by a variety of enterovirus species.

The 3C^pro proteins of CV-A16, CV-A6, and EV-D68 target TAK1 to impair the SeV-triggered NF-κB response.

As indicated in Fig. 2B, CV-A16 can also block NF-κB activation after SeV infection. In fact, CV-A16 3C^pro alone can potently suppress NF-κB activation after SeV infection (Fig. 12A, lanes 3 and 4). Endogenous NF-κB-activated IκB (Fig. 12B) and interleukin-8 (IL-8) (Fig. 12C) mRNA levels were also suppressed by CV-A16 3C^pro after SeV infection. Several studies have demonstrated that TAK1 is a key player in NF-κB activation (34, 67). In response to interleukin-1, tumor necrosis factor alpha, and Toll-like receptor agonists, it mediates the activation of NF-κB, c-Jun N-terminal kinase (JNK), and p38 pathways (67). In order to assess whether TAK1 is the target of CV-A16 3C^pro in the NF-κB pathway, we examined TAK1 expression and processing in the presence or absence of 3C^pro in HEK293T cells. TAK1 was transfected with different amounts of the 3C^pro expression vector, and protein expression was detected by Western blotting. Interestingly, TAK1 cleavage products were induced by CV-A16 3C^pro expression (Fig. 12D, lanes 2 and 3) compared with the control (Fig. 12D, lane 1). A TAK1-related smaller protein band at around 45 kDa was detected in the presence of CV-A16 3C^pro. To further evaluate whether 3C^pro-induced TAK1 cleavage is protease activity dependent, we examined the CV-A16 3C^pro H40D mutant versus wild-type 3C^pro. As shown in Fig. 12D (lanes 4 and 5), the cleaved TAK1 fragment was not detected in the presence of the CV-A16 3C^pro H40D mutant. Thus, the protease activity of CV-A16 3C^pro is required for TAK1 cleavage. At the same time, the ability of the 3C^pro H40D mutant to suppress NF-κB activation after SeV infection was also impaired (Fig. 12A, lanes 5 and 6). The ability to cleave TAK1 was also observed for the 3C^pro proteins of EV-D68 (Fig. 12F) and CV-A6 (Fig. 12G). Similarly, the 3C^pro proteins of EV-D68 and CV-A6 suppressed NF-κB activation after SeV infection (Fig. 12E).

FIG 12.

The 3C^pro proteins of CV-A16, CV-A6, and EV-D68 target TAK1 to impair the NF-κB response. (A) HEK293T cells were transfected with the pRL-TK plasmid and the NF-κB–luc promoter reporter plus a control plasmid (lane 1), HA–CV-A16 3C (lanes 3 and 4), or the HA-3C H40D mutant (lanes 5 and 6). Twenty-four hours after transfection, cells were challenged with SeV (20 HA units/ml). Twenty-four hours later, cell lysates were assayed for luciferase activity. Data are representative of results from three independent experiments. The error bars indicate the standard deviations of data from three different experiments. P values of <0.05 were considered significant. (B) HEK293T cells were transfected with an empty vector or increasing amounts of the CV-A16 3C expression vector. After 24 h, cells were challenged with SeV (20 HA units/ml) or left uninfected for 20 h. The cells were then harvested, and RNA was extracted to determine IKBA (B) and IL-8 (C) mRNA levels by quantitative real-time PCR. (D) HEK293T cells were transfected with plasmids encoding TAK1-Flag and a control vector (lane 1) and different amounts (150 ng and 450 ng) of HA-3C (lanes 2 and 3) or the HA-3C H40D mutant (lanes 4 and 5). Thirty-six hours later, lysates of cells were analyzed by Western blotting with antibodies against Flag and tubulin. Tubulin antibody was used as a control. (E) HEK293T cells were transfected with the pRL-TK plasmid and the NF-κB–luc promoter reporter plus a control plasmid (lane 1), CV-A6 3C (lane 3), or EV-D68 3C (lane 4). Twenty-four hours after transfection, cells were challenged with SeV (20 HA units/ml). Twenty-four hours later, cell lysates were assayed for luciferase activity. Data are representative of results from three independent experiments. The error bars indicate the standard deviations of data from three different experiments. P values of <0.05 were considered significant. (F and G) HEK293T cells were transfected with plasmids encoding TAK1-Flag and a control vector (lane 1), EV-D68 3C (lane 2) (F), or CV-A6 3C (lane 2) (G). Thirty-six hours later, lysates of cells were analyzed by Western blotting with antibodies against Flag and tubulin. Tubulin antibody was used as a control.

The 3C^pro proteins of CV-A16, CV-A6, and EV-D68 inhibit the TLR3-mediated NF-κB response.

In addition to the RLR pathway, TLR3 detection of the RNA component of RNA viruses also triggers NF-κB activation. The ectopic expression of TLR3 in HEK293T cells led to NF-κB activation in the presence of poly(I·C) (Fig. 13A, lane 2). CV-A16 3C^pro suppressed TLR3-triggered NF-κB activation (Fig. 13A, lane 3). In contrast, the ability of the CV-A16 3C^pro H40D mutant to suppress TLR3-triggered NF-κB activation was impaired (Fig. 13A, lane 4). Cleavage of endogenous TAK1 was detected in the presence of CV-A16 3C^pro (Fig. 13B). CV-A16 infection also triggered TAK cleavage (Fig. 13C). The 3C^pro proteins of EV-D68 and CV-A6 also suppressed TLR3-triggered NF-κB activation (Fig. 13D). The CV-A6 3C^pro H40D mutant (Fig. 13E) or the EV-D68 3C^pro H40D mutant (Fig. 13F) was not able to cleave TAK1, unlike wild-type 3C^pro. These mutants also had an impaired ability to suppress TLR3-triggered NF-κB activation (Fig. 13D).

FIG 13.

CV-A16, CV-A6, or EV-D68 3C^pro inhibits the TLR3-mediated NF-κB response. (A) HEK293T cells were transfected with the pRL-TK plasmid, the NF-κB–luc promoter reporter, and the Flag-TLR3 expression vector alone or with CV-A16 3C or the CV-A16 3C H40D mutant. Twenty-four hours later, cells were treated with poly(I·C) for 12 h, and cell lysates were then analyzed for luciferase activity. Data are representative of results from three independent experiments. The error bars indicate the standard deviations of data from three different experiments. (B) HEK293T cells were transfected with the TAK1-Flag expression vector. Twelve hours later, cells were treated with increasing MOIs (2 and 3) of CV-A16 for 20 h. The cell lysates were then analyzed by Western blotting. (C) HEK293T cells were transfected with different amounts of HA–CV-A16 3C plasmids (lanes 2 and 3) or a control vector (lane 1). Thirty-six hours later, cells were harvested, and endogenous TAK1 and 3C-HA proteins were analyzed by Western blotting. Tubulin antibody was used as a control. (D) HEK293T cells were transfected with the pRL-TK plasmid; the NF-κB–luc promoter reporter; and the Flag-TLR3 expression vector alone or along with CV-A6 3C, EV-D68 3C, or their H40D protease mutants. Twenty-four hours later, cells were treated with poly(I·C) for 12 h, and cell lysates were then analyzed for luciferase activity. Data are representative of results from three independent experiments. The error bars indicate the standard deviations of data from three different experiments. (E and F) TAK1-Flag was transfected with CV-A6 3C or its H40D mutant (E) or with EV-D68 3C or its H40D mutant (F). Thirty-six hours later, cells were harvested, and proteins were analyzed by Western blotting. Tubulin antibody was used as a control.

DISCUSSION

In this study, we evaluated the role of 3C^pro in viral evasion of host innate immune responses for several poorly studied enteroviruses and coxsackieviruses, including CV-A16, CV-A6, and EV-D68. In a systematic evaluation, we observed that the CV-A16, CV-A6, and EV-D68 3C^pro proteins blocked immune activation triggered by the RNA virus SeV (Fig. 4). Furthermore, the CV-A16, CV-A6, and EV-D68 3C^pro proteins inhibited IRF3 phosphorylation after SeV infection (Fig. 4). At the same time, the CV-A16, CV-A6, and EV-D68 3C^pro proteins did not block type I IFN activation induced by the TBK1 or IKKε downstream factors (Fig. 5). Instead, the CV-A16 (Fig. 6), CV-A6, and EV-D68 (Fig. 10) 3C^pro proteins directly interfere with MDA5-MAVS functional signaling.

CV-A16 3C^pro impaired MDA5-CARD-induced IFN-β activation in a dose-dependent manner (Fig. 6). CV-A16 3C^pro binds to MDA5 (Fig. 6) and disrupts its interaction with MAVS (Fig. 8). In a manner distinct from that of other RNA viruses (49, 52), CV-A16 3C^pro suppressed MDA5-triggered type I IFN activation without reducing MDA5 (Fig. 6C) or MAVS (Fig. 6D) expression. The interaction between CV-A16 3C^pro and MDA5 was not affected by RNase treatment (Fig. 7). Mutation of the protease active site of CV-A16 3C^pro also did not abolish its interaction with MDA5 (Fig. 6). On the other hand, mutations of N-terminal surface-exposed amino acids inhibited its interaction with MDA5 (Fig. 7). To our knowledge, this is the first example showing that the 3C proteins of the Picornaviridae family of viruses interfere with the MDA5-MAVS interaction to abrogate type I IFN induction.

The ability to interact with MDA5 was also evaluated for various enteroviruses and coxsackieviruses linked to HFMD and pediatric respiratory disease. We observed the interaction between MDA5 and the 3C^pro proteins of CV-A6 and EV-D68 (Fig. 11). Like CV-A16, the 3C^pro proteins of CV-A6 and EV-D68 efficiently antagonized MDA5-induced IFN-β activation (Fig. 11). On the other hand, we did not detect an interaction between 3C^pro of CV-B3 or EV-A71 and MDA5 (Fig. 11).Whether a disruption of MDA5 function by 3C^pro is seen among other diverse Picornaviridae is a question that requires further investigation.

MDA5 is a critical component of the host RLR pathway. Several studies using MDA5^−/− mice demonstrated that MDA5 signaling is critical for type I IFN induction after virus infection (68, 69). MDA5-deficient mice have an increased susceptibility to lethal infection induced by poliovirus (68, 69). It is interesting to note that EV-A71 3C^pro has been reported not to interfere with MDA5-mediated IFN restriction (35). These results are consistent with our observation that EV-A71 3C^pro did not interact with MDA5 (Fig. 11). It is possible that the EV-A71 and CV-B3 3C proteases use different strategies to neutralize IFN-β activation mediated by MDA5. For instance, the CV-B3 3C protein targets MAVS (47) and the EV-A71 3C protein targets IRF7 (35) in the RLR pathway to impair MDA5-induced IFN-β activation.

In addition to type I IFN activation, SeV infection also triggers NF-κB activation through TAK1. We demonstrated that TAK1 was also targeted by CV-A16 3C^pro (Fig. 12). CV-A16 3C^pro expression resulted in TAK1 cleavage, and this function was blocked by a CV-A16 3C^pro active-site (H40D) mutation (Fig. 12). Consequently, NF-κB activation triggered by SeV infection was inhibited by CV-A16 3C^pro but not by the 3C^pro H40D mutant (Fig. 12). We also observed that both the CV-A6 and EV-D68 3C^pro proteins harbor the same ability to diminish the NF-κB signal by targeting TAK1 (Fig. 12). The protease active site was also important for the CV-A6 and EV-D68 3C^pro proteins to diminish the NF-κB signal by cleaving TAK1 (Fig. 12).

Apart from the RLR pathway, TLRs are also critical for detecting the RNA component of RNA viruses. Among them, TLR3 is one of the essential TLRs for sensing invading viral RNA. TLR3 can activate TAK1 and subsequently stimulate p38, JNK, and NF-κB signals, leading to cytokine production. TLR3-triggered NF-κB activation was inhibited by CV-A16, CV-A6, or EV-D68 3C^pro but not the H40D mutants (Fig. 13). The cleaved fragment of TAK1 was around 45 kDa. As a result, the integrity of TAK1 functional domains was destroyed. The N-terminal fragment of TAK1 includes the TAB1 binding and kinase domains, which are critical for TAK1 activation. The C-terminal fragment of TAK1 includes the TAB2/3 binding domain, which is necessary for the activation of downstream IKK kinases (IKKβ and IKKα). Thus, the separation of the N-terminal and C-terminal domains of TAK1 by CV-A16, CV-A6, or EV-D68 3C^pro will lead to the disruption of host RLR and TLR pathways. In addition to innate immune responses, TAK1 plays a central role in host adaptive immunity. T cell receptor (TCR) ligation leads to the TAK1-mediated activation of both NF-κB and JNK, which is required for the development and maturation of T cells. CV-A16 has been shown to infect T cells (70). In light of the impaired NF-κB activation by 3C^pro targeting TAK1, CV-A16 may interfere with T cell activation and maturation in vivo. The interplay between CV-A16 infection and adaptive immunity should be studied in the future.

During virus infection, host cells utilize different strategies to combat virus invasion. Viruses have also evolved sophisticated mechanisms to subvert host immune responses. Overall, we showed in this study that the 3C^pro proteins of CV-A16, CV-A6, and EV-D68 utilize novel tactics to subvert host innate immune responses by targeting key factors in the RLR and TLR pathways (Fig. 14). This information should contribute to our understanding of the pathogenesis of CV-A16, CV-A6, and EV-D68 infections. Furthermore, the recognition of the critical role of this protease in the pathogenesis of CV-A16, CV-A6, and EV-D68 infections and the success of protease inhibitors in the treatment of other viral infections provide a foundation for the development of new therapeutic agents to treat HFMD and pediatric respiratory disease.

FIG 14.

Summary of the sites of CV-A16 3C^pro activity in suppressing type I IFN and NF-κB pathways. The RNA virus enters target cells through the endocytic pathway, and TLRs detect the viral genomes in endosomes. TLRs then engage the cytosolic adaptor to activate the E3 ubiquitin (Ubi) ligase TRAF6, which ubiquitinates itself and recruits TAB proteins. TAB proteins activate TAK1. As a result, the IKK complex (IKKβ, IKKα, and NEMO) is activated, which phosphorylates IκB proteins. The phosphorylation of IκB leads to its ubiquitination and proteasomal degradation, freeing NF-κB. Active NF-κB translocates to the nucleus to induce inflammatory cytokine production. Once in the cytoplasm, viral RNA can be detected by RIG-I/MDA5, which activates the IFN-β induction pathway through MAVS, located on mitochondria. In our model, the CV-A16, CV-A6, and EV-D68 3C^pro proteins all target TAK1 and also disrupt the association of MDA5 and MAVS.

MATERIALS AND METHODS

Cell lines and viruses.

HEK293T cells (ATCC CRL-11268) and human rhabdomyosarcoma RD cells (ATCC CCL-136) were purchased from the ATCC and maintained in Dulbecco's modified Eagle's medium (DMEM) containing 10% heat-inactivated fetal bovine serum (FBS) (HyClone, Logan, UT), 100 U/ml penicillin, and 100 μg/ml streptomycin at 37°C in a 5% CO₂ humidified atmosphere. SeV was kindly provided by Tao Wang (Tianjin University) and Junliang Chang (Jilin University).

Antibodies and reagents.

The following antibodies were used for Western blotting in this study: anti-Flag monoclonal M2 antibody (catalog number F1804; Sigma), anti-Myc (9E10) monoclonal antibody (catalog number MMS-150P; Covance), anti-HA monoclonal antibody (catalog number MMS-101R; Covance), antitubulin monoclonal antibody (catalog number MMS-410P; Covance), anti-V5 antibody (catalog number V8012; Sigma), VP1 antiserum against CA16 obtained from rabbits immunized with CV-A16 CC024, RIG-I (D14g6) rabbit antibody (catalog number 3743; Cell Signal), Cardif (human) antibody (AT107) (catalog number ALX-210-929-C100; Enzo Life Sciences), MDA5 (human) polyclonal antibody (AT113) (catalog number ALX-210-935-C100; Enzo Life Sciences), TAK1 antibody (catalog number 2E10; Novus), and anti-p-IRF3 monoclonal antibody (catalog number AB76493; Abcam). Poly(I·C) was purchased from Guandong South China Pharmaceutical Co. Ltd.

Plasmids.

Flag-MAVS, V5-TBK1, Flag-IKKε, Flag-IRF3, Flag-IRF7, and firefly luciferase reporter plasmids for IFN-β were kindly provided by Tao Wang (Tianjin University). The NF-κB luciferase reporter plasmid (catalog number E8491; Promega) was kindly provided by Yong Cai (College of Life Science, Jilin University, China). Flag-TAK1 was generously gifted by Mingyu Lv (Jilin University, China). Flag–RIG-IN (CARD) was generously gifted by Jinhua Yang (Baylor College of Medicine, Houston, TX, USA) (71). Flag–MDA5-N (amino acids 1 to 213) was amplified from a template (purchased from Fitgene Company, Shanghai, China) and cloned into the SalI and BamHI sites of the VR1012 vector. Flag–IRF3-5D was constructed by Generay Biotech Co. Ltd. (Shanghai, China). Flag-TLR3 was purchased from Addgene. 3C ORFs of CV-A16 strain CC024 (GenBank accession number KF055238.1), EV-A71 strain 063 (GenBank accession number HQ647172.1), CV-A6 Changchun046/CHN/2013 7434 (GenBank accession number KT779410), and EV-D68 Fermon (GenBank accession number AY426531.1) were cloned into the VR1012-HA-Flag vector using the SalI and BamHI sites. The CV-B3 (GenBank accession number JX312064.1) 3C expression vector was obtained from Generay Biotech Co. Ltd., and the 3C-containing DNA fragment was cloned into the VR1012 vector using the SalI and BamHI sites. CV-A16, EV-D68, and CV-A6 3C H40D plasmids were obtained by single-site mutation.

Luciferase reporter assay.

HEK293T cells were plated into 24-well dishes and transfected the following day. One hundred nanograms of the reporter plasmid for IFN-β, NF-κB promoters, 1 ng the Renilla luciferase control plasmid (pRL-TK), and the indicated amounts of the expression plasmids were used per well. At 24 h posttransfection, cells were infected with SeV (20 hemagglutination [HA] units/ml) or left uninfected for 18 h. Luciferase activities were then measured by using a Dual-Luciferase reporter assay system (Promega, Madison, WI) according to the manufacturer's instructions. Firefly luciferase activity was normalized to Renilla luciferase activity. Finally, the relative luciferase activities were expressed as fold changes over the empty-plasmid-transfected or SeV-noninfected controls.

CPE observation.

For observing CPEs, RD cells were grown on a culture dish; infected with CV-A16, CV-A6, or EV-D68 (multiplicity of infection [MOI] of 0.5); and treated with IFN-α2 (1,000 U/ml; Changchun Institute of Biological Products) separately or together at the indicated time points. Morphological changes were observed and photographed under a light microscope (IX51; Olympus, Tokyo, Japan).

RNA extraction and quantitative real-time PCR.

Total RNA was extracted from cells by using TRIzol (Invitrogen, Carlsbad, CA). Reverse transcription was carried out with a 20-μl volume by using a Superscript cDNA synthesis kit (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. Quantitative real-time PCR (qPCR) was carried out on an Mx3005P instrument (Agilent Technologies, Stratagene, USA) by using a master mix (SYBR green) kit (Bio-Rad) and the following primers designed by using the VP1 conserved region sequences of CV-A16: CA16-F1 (CATGCAGCGCTTGTGCTT), CA16-F2 (CATGCAACGACTGTGCTTTC), CA16-R1 (CACACAATTCCCCCGTCTTAC), CA16-R2 (CATAATTCGCCCGTTTTGCT), GAPDH-F (CCCATCACCATCTTCCAGG) and GAPDH-R (TTCTCCATGGTGGTGAAGAC) for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as a control, IFNB-F (CACTGGCTGGAATGAGACT), IFNB-R (TTTCGGAGGTAACCTGTAAG), IKBA-F (CGGCCTGGACTCCATGAAAG), IKBA-R (CCTTCACCTGGCGGATCACT), IL-8-F (CGGAAGGAACCATCTCACTGTG), and IL-8-R (AGAAATCAGGAAGGCTGCCAAG). The qPCR assay was carried out with a 20-μl volume consisting of 10 μl of a 2× SYBR green mix solution, 0.4 μl of 5 μmol/liter of each oligonucleotide primer, and 2 μl of the cDNA template. Target fragment amplification was carried out as follows: 50°C for 2 min and then 95°C for 10 min, followed by 50 cycles consisting of 95°C for 15 s and 60°C for 1 min. Melting-curve analysis was carried out at 90°C for 1 min and then at 55°C for 30 s and 95°C for 30 s.

Determination of viral titers.

Virus titers were determined by using the median endpoint of the cell culture infectious dose (CCID₅₀). Serially diluted viruses were added to RD cells grown in 96-well plates, and 8 replicate samples were used for each dilution. CCID₅₀ values were measured by counting infected RD cell culture wells with obvious CPEs and calculated by using the Reed-Muench method (72).

Immunoprecipitation.

After 48 h of transfection, cells were harvested and lysed with lysis buffer (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 0.5% NP-40) containing a protease inhibitor cocktail (Roche, Indianapolis, IN). Lysates of cells were incubated with an anti-Flag affinity matrix (catalog number A-2220; Sigma) or an anti-HA affinity matrix (Roche) at 4°C overnight on a rotator. After washing with wash buffer (20 mM Tris-Cl [pH 7.5], 100 mM NaCl, 0.05% Tween 20, 0.1 mM EDTA) six times, 50 μl of elution buffer (100 mM glycine-HCl [pH 2.5]) was added to resuspend the beads, and the eluted proteins were obtained by centrifugation, followed by SDS-PAGE and WB analysis.

Western blot analysis.

Cells were harvested and lysed in radioimmunoprecipitation assay (RIPA) lysis buffer (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 1% NP-40), and the lysate was cleared by centrifugation at 16,000 × g at 4°C for 5 min. Total cell extracts were subjected to SDS-PAGE and transferred onto nitrocellulose membranes (catalog number 10401196; Whatman). After blocking with 5% nonfat dry milk in Tris-buffered saline–Tween (TBST) for 1 h at room temperature (RT), membranes were incubated with the indicated primary antibodies at 4°C overnight and then with the corresponding alkaline phosphatase (AP)-conjugated secondary antibodies (Sigma) for 1 h at RT. After three washes with TBST, the blots were reacted with nitroblue tetrazolium (NBT) and 5-bromo-4-chloro-3′-indolylphosphate (BCIP) (Sigma). Around endogenous TAK1, MDA5 and MAVS proteins were detected by using a horseradish peroxidase (HRP)-conjugated secondary antibody (Cell Signal) for 1 h at RT. After three washes with TBST, the membranes were reacted with the ECL sensitive kit (catalog number B500023; Proteintech) and developed by using the c500 azure system.

Statistical analysis.

Data from the luciferase reporter assay and reverse transcription-qPCR (qRT-RCR) results are presented as means and standard errors. Viral titers were analyzed by using GraphPad Prism 6 software (GraphPad Software, La Jolla, CA, USA). Differences among groups were analyzed by analysis of variance (ANOVA) (Stata Corp., College Station, TX). P values of <0.05 were considered significant.