|Home | About | Journals | Submit | Contact Us | Français|
The innate immune system senses RNA virus infections through membrane-bound Toll-like receptors or the cytoplasmic proteins RIG-I and MDA-5. RIG-I is believed to recognize the 5′-triphosphate present on many viral RNAs, and hence is important for sensing infections by paramyxoviruses, influenza viruses, rhabdoviruses, and flaviviruses. MDA-5 recognizes dsRNA, and senses infection with picornaviruses, whose RNA 5′-ends are linked to a viral protein, VPg, not a 5′-triphosphate. We previously showed that MDA-5 is degraded in cells infected with different picornaviruses, and suggested that such cleavage might be a mechanism to antagonize production of type I IFN in response to viral infection. Here we examined the state of RIG-I during picornavirus infection. RIG-I is degraded in cells infected with poliovirus, rhinoviruses, echovirus, and encephalomyocarditis virus. In contrast to MDA-5, cleavage of RIG-I is not accomplished by cellular caspases or the proteasome. Rather, the viral proteinase 3Cpro cleaves RIG-I, both in vitro and in cells. Cleavage of RIG-I during picornavirus infection may constitute another mechanism for attenuating the innate response to viral infection.
When viruses infect cells, intrinsic defensive actions are initiated almost immediately. These defenses include the innate immune system, which provides cytokines to halt virus infection, and modulate the adaptive immune response should the infection proceed unchecked (Janeway and Medzhitov, 2002). The innate immune system is activated when microbial products, such as lipopolysaccharide or viral nucleic acids, are detected. RNA viruses are recognized as foreign by cellular sensors that are activated by viral proteins or nucleic acids, leading to the production of the critical antiviral type I interferons.
Sensing of RNA virus infection by the innate immune system is carried out by membrane-bound Toll-like receptors, or by cytoplasmic sensors such as PKR, RIG-I, and MDA-5 (reviewed in (Kato et al., 2005; Yoneyama and Fujita, 2007; Yoneyama et al., 2004)). RIG-I and MDA-5 proteins comprise an amino-terminal caspase recruitment domain (CARD) and an RNA helicase domain (Kang et al., 2002). Results of a recent study on the evolution of RIG-I and MDA-5 indicate that the unique protein domain arrangement evolved independently by domain grafting and not by a simple gene duplication event of the entire four-domain arrangement, which may have been initiated by differential sensitivity of these proteins to viral infection (Sarkar et al., 2008). Additionally, MDA-5, but not RIG-I, orthologs are found in fish indicating that MDA-5 might have evolved before RIG-I (Sarkar et al., 2008). After binding viral RNA, these sensors interact with a CARD-containing adaptor protein, IPS-1, located in the outer membrane of mitochondria (Kawai et al., 2005; Meylan et al., 2005; Seth et al., 2005; Xu et al., 2005). This interaction mediates recruitment and activation of protein kinases that phosphorylate the transcription protein IFN-regulatory factor 3, leading to synthesis of type I IFN.
An important question is how RIG-I and MDA-5 distinguish viral from cellular RNAs. It was originally believed that these proteins recognize dsRNA, which is rarely found in the cytoplasm of cells but is abundant in virus-infected cells (Yoneyama et al., 2004). More recently it has become clear that RIG-I recognizes RNA with a 5′-triphosphate (Hornung et al., 2006; Pichlmair et al., 2006). Because most cellular cytoplasmic RNAs bear a 5′-cap structure, this observation seems to explain the ability of RIG-I to discriminate between host and viral RNA. This substrate specificity is supported by observations that suggest that RIG-I and MDA-5 specialize in recognition of different viruses. Infection of mice lacking the gene encoding either protein reveals that RIG-I is essential for detecting infection by rhabdoviruses, influenza viruses, paramyxoviruses, and flaviviruses (Kato et al., 2006). Replication of these viruses leads to production of RNAs with a 5′-triphosphate. In contrast, MDA-5 senses infection with picornaviruses, whose RNA 5′-ends are linked to a viral protein, VPg, not a 5′-triphosphate (Gitlin et al., 2006; Kato et al., 2006). It has been suggested that dsRNAs produced during picornavirus replication are the substrates for MDA-5 recognition.
Despite these elegant innate mechanisms, virus infections still occur because their genomes encode proteins that antagonize this and every other step of host defense. Examples include inhibition of RIG-I function by binding of the influenza virus NS1 protein (Guo et al., 2007; Mibayashi et al., 2007; Opitz et al., 2007), and cleavage of IPS-1 by proteases encoded in the genomes of picornaviruses and hepatitis C virus (Li et al., 2005; Lin et al., 2006) (J. Drahos and V. Racaniello, unpublished data). We previously showed that MDA-5 is degraded in cells infected with different picornaviruses, and suggested that such cleavage might be a mechanism to antagonize production of type I IFN in response to viral infection (Barral et al., 2007). Here we examine the state of RIG-I during picornavirus infection. We found that RIG-I is degraded in cells infected with poliovirus, rhinoviruses, echovirus, and encephalomyocarditis virus. In contrast to MDA-5, cleavage of RIG-I is not accomplished by cellular caspases or the proteasome (Barral et al., 2007). Rather, the viral proteinase 3Cpro cleaves RIG-I, both in vitro and in cells. Cleavage of RIG-I during picornavirus infection may constitute another mechanism for attenuating the innate response to viral infection.
Consistent with the suggestion that picornavirus infections are detected by MDA-5 (Kato et al., 2006) is the observation that this protein is degraded during infection with poliovirus, rhinovirus type 1a, and EMCV (Barral et al., 2007). It was therefore of interest to determine the state of RIG-I during picornavirus infection. HeLa cells were infected with poliovirus, and at different times after infection, RIG-I protein was examined by western blot analysis. Beginning at 4 h post-infection, levels of RIG-I protein declined, and a protein of ~70 kDa appeared which might be a cleavage product (Figure 1A). Cleavage of RIG-I protein was also observed during poliovirus infection of the neuroblastoma cell line SH-SY-5Y (Figure 1B).
Cleavage of RIG-I was also observed in cells infected with other picornaviruses. In cells infected with echovirus type 1 (Figure 1C) or EMCV (Figure 1F), a ~70 kDa putative cleavage product was first detected at 6 h post-infection. When cells were infected with rhinovirus type 16 at 33°C, a temperature at which viral replication is more efficient, only the ~70 kDa protein was observed at 14 hr post-infection and later (Figure 1D). Slight cleavage of RIG-I was detected in cells infected with rhinovirus type 1A at 37°C (Figure 1E).
Poliovirus-induced cleavage of MDA-5 is carried out by the proteasome and caspases (Barral et al., 2007). The effect of inhibitors of the proteasome (MG132) and caspases (Z-VAD) on RIG-I cleavage was therefore determined. HeLa cells were infected with poliovirus, and culture medium was added with or without inhibitor. The level of RIG-I at different times after infection was determined by western blot analysis. In the absence of inhibitor, cleavage of RIG-I was observed beginning at 4 hr post-infection and was complete by 8 hr (Figure 2A). In the presence of Z-VAD or MG132, degradation of RIG-I was first observed at 6 hr post-infection, and was complete by 8 hr post-infection (Figure 2B, C). In contrast, poliovirus-induced degradation of MDA-5 is completely inhibited by MG132 and Z-VAD, even though neither drug impairs viral yields (Barral et al., 2007). These results indicate that, in contrast to poliovirus-induced degradation of MDA-5, cleavage of RIG-I during poliovirus infection occurs by a process that is independent of caspases and the cellular proteasome.
The two polioviral proteinases, 2Apro and 3Cpro, not only process the viral polyprotein to produce the functional viral proteins, but also degrade cellular proteins such as eIF4G (Krausslich et al., 1987) and cellular transcription proteins (Weidman et al., 2003). Polioviruses with single amino acid changes in either 2Apro or 3Cpro have been isolated which prevent cleavage of cellular proteins. These viral mutants were used to determine whether either viral proteinase plays a role in poliovirus-induced cleavage of RIG-I. Poliovirus mutant 2Apro Y88L contains a single amino acid change in 2Apro that abolishes cleavage of eIF4G, but does not affect cleavage of the viral polyprotein (Yu et al., 1995). Poliovirus mutant Se1–3C-02 contains a single amino acid change in 3Cpro that has been reported to block cleavage of host cell transcription proteins, but not processing of the viral polyprotein (Clark et al., 1991; Dewalt and Semler, 1987). RIG-I cleavage was observed starting at 6 hr post-infection in cells infected with either 2Apro Y88L or Se1–3C-02 (Figure 3B, 3C). The delay in cleavage is likely a consequence of the slower replication kinetics of theses viruses (unpublished data).
To further explore the identity of the viral proteinase that cleaves RIG-I, we determined whether RIG-I could be directly cleaved in vitro by either viral proteinase. Purified poliovirus 3CDpro or coxsackievirus B3 2Apro was added to a cytoplasmic extract produced from HeLa cells, and after incubation, RIG-I protein was detected by western blot analysis. Coxsackievirus B3 2Apro was used in these experiments because it has not been possible to purify poliovirus 2Apro. Poliovirus 3CDpro is the precursor to 3Cpro and is believed to carry out the majority of protein processing during infection. Degradation of RIG-I and production of the putative ~70 kDa cleavage product was observed after incubation with 3CDpro, but not with 2Apro (Figure 4A) or incubation without enzyme (unpublished data). Proteinase activity of 2Apro was confirmed by demonstrating cleavage of the known substrates PABP (Figure 4B) and eIF4GI (Figure 4C). Although PABP is known to be cleaved by 3CDpro, for unknown reasons the protein remained intact in this assay (Figure 5B).
To determine whether either poliovirus proteinase cleaves RIG-I in cells, plasmids encoding these proteins linked to a FLAG epitope were introduced into 293A cells by DNA-mediated transformation. Twenty-four hours later, cell extracts were prepared and RIG-I was detected by western blot analysis. Cleavage of RIG-I and production of the ~70 kDa protein was observed in cells transformed with plasmids encoding 3Cpro, but not 2Apro (Figure 5). Synthesis of both proteinases in 293A cells was verified by western blot analysis using anti-FLAG antibody. Furthermore, activity of 2Apro was confirmed by observation of cleavage of its known substrate, eIF4G1. We conclude that 3Cpro is the poliovirus proteinase responsible for cleavage of RIG-I during infection.
Cleavage of RIG-I was observed in cells infected with all picornaviruses examined, including poliovirus, rhinovirus types 1a and 16, echovirus type 1, and encephalomyocarditis virus. We previously showed that another cytoplasmic RNA sensor, MDA-5, is degraded during picornavirus infection in a proteasome- and caspase-dependent manner (Barral et al., 2007). However, inhibitors of these cellular proteases had no effect on poliovirus-induced cleavage of RIG-I. The results of in vitro cleavage assays, and expression of DNAs encoding viral proteinases in cultured cells, showed that RIG-I is cleaved by poliovirus 3Cpro. The 3Cpro proteinase of the other picornaviruses examined is also likely to cleave RIG-I. The second enterovirus proteinase, 2Apro, is not encoded by the genome of encephalomyocarditis virus and therefore could not explain the cleavage of RIG-I observed in these experiments.
Although the kinetics of cleavage induced by different picornaviruses varied, a ~70 kDa putative cleavage product was always observed. This cleavage product represents the carboxy-terminal portion of RIG-I, because the antibody used to detect it by western blot analysis is directed against a peptide from the last 17 amino acids of the protein. It was therefore possible to predict 3Cpro cleavage sites in RIG-I that would yield the ~70 kDa cleavage product. We introduced amino acid changes at 12 of these cleavage sites, but none altered processing of RIG-I during poliovirus infection (unpublished data). Therefore either the correct 3Cpro cleavage site has not yet been identified, or the amino acid changes made were not sufficient to block cleavage.
It is believed that MDA-5, not RIG-I, is crucial for sensing infections with picornaviruses. Mice lacking the gene encoding MDA-5 are more susceptible to infection with encephalomyocarditis virus, and produce less IFN after infection compared with wild type littermates (Gitlin et al., 2006; Kato et al., 2006). Mice lacking the gene encoding RIG-I were no more susceptible to infection with encephalomyocarditis virus and showed no difference in IFN production (Kato et al., 2006). The cleavage of MDA-5 during picornavirus infection is consistent with a role for this protein in detecting infection with members of this virus family (Barral et al., 2007). It is not clear why RIG-I would be cleaved during picornavirus infection if this sensor plays no role in innate responses against these viruses. RIG-I is known to be activated by short (~1 kb) stretches of dsRNA (Hornung et al., 2006; Pichlmair et al., 2006) that are certainly found in picornavirus infected cells. A U-rich sequence in the genome of hepatitis C virus has been shown to activate RIG-I (Saito et al., 2008). Similar sequences are present in the genomes of picornaviruses and might serve as substrates for RIG-I. Perhaps the results obtained by infecting rig-I−/− mice with encephalomyocarditis virus are not representative of all picornaviruses. Understanding the role of RIG-I cleavage during enterovirus infection will require synthesis in cell cultures and in mice of non-cleavable forms of the protein.
IPS-1 is also cleaved during infection with poliovirus and rhinovirus (J. Drahos and V. Racaniello, unpublished data) as well as hepatitis A virus (Yang et al., 2007). Therefore, infection with certain picornaviruses leads to cleavage of not only both cytoplasmic RNA sensors, but also the mitochondrial membrane protein that is crucial in transmitting the signal from RIG-I and MDA-5 that leads to induction of IFN transcription. It seems unlikely that the cleavage of three members of this sensing pathway is coincidental. It is possible that these cleavages target unknown functions of RIG-I, MDA-5, and IPS-1 unrelated to sensing RNA. Further experiments are clearly required to understand why these components of the innate RNA sensing pathway are cleaved during picornavirus infection.
S3 HeLa and SH-SY5Y cells were grown in Dulbecco’s modified Eagle medium (Invitrogen, Carlsbad, California USA), 10% bovine calf serum (Hyclone, Logan, Utah USA), and 1% penicillin/streptomycin (Invitrogen). For plaque assays HeLa cells were grown in Dulbecco’s modified Eagle medium (Specialty Media, Philipsburg, New Jersey, USA), 0.2% NaHCO3, 5% bovine calf serum, 1% penicillin/streptomycin, and 0.9% bacto-agar (Difco, Franklin Lakes, New Jersey, USA).
Stocks of poliovirus strain P1/Mahoney, rhinovirus type 16, and encephalomyocarditis virus (EMCV) were produced by transfecting HeLa cells with RNA transcripts derived by in vitro transcription of plasmids harboring complete DNA copies of the viral genomes (Duke and Palmenberg, 1989; Lee, Wang, and Rueckert, 1995; Racaniello and Baltimore, 1981a). Stocks of echovirus type 1 and rhinovirus type 1a were obtained from the American Type Culture Collection, Manassas, VA, and were propagated in HeLa cells. Poliovirus mutant Se1–3C-02, which contains the single amino acid change V54A in 3Cpro (Dewalt and Semler, 1987), was obtained from B. Semler, University of California, Irvine. A poliovirus mutant with a single amino acid change, Y88L (Yu et al., 1995), in 2Apro was constructed by site-directed mutagenesis of a full length DNA copy of the genome of poliovirus strain P1/Mahoney.
A cytoplasmic extract was prepared from HeLa cells as described (Todd, Towner, and Semler, 1997). Briefly, 1 × 108 HeLa cells were centrifuged, washed with phosphate-buffered saline, resuspended in hypotonic buffer (20 mM Hepes KOH pH 7.4, 10 mM KCl, 1.5 mM MgOAc, 1 mM dithiothreitol), and lysed with a Dounce homogenizer. The extract was then centrifuged at 10,000 × g and the supernatant was stored. For proteinase cleavage, 25 μl of cytoplasmic extract was incubated for 6 hr with 0.5 μg purified 2Apro or 3CDpro (gift of Bert Semler, University of California, Irvine) in a 100 μl reaction containing 50 mM NaCl, 5 mM MgCl2. Cleavage of RIG-I was assessed by western blot analysis as described below.
Primers were designed to place a FLAG epitope at the amino terminus of 2Apro and 3Cpro, using a DNA copy of the genome of poliovirus type 1 Mahoney as the template (Racaniello and Baltimore, 1981b). DNAs encoding the proteinases were cloned in the vector pcDNA3 (Invitrogen). DNA-mediated transformation of 293A cells was done using Lipofectamine reagent (Invitrogen) according to the manufacturer’s protocol. Cell extracts were prepared 24 hr later as described below for western blot analysis.
Rabbit antibody against a peptide comprising amino acids 909–925 of RIG-I was purchased from Abgent, Inc., San Diego, CA. Mouse monoclonal anti-EF1α was purchased from Upstate USA Inc., Chicago, IL. Mouse monoclonal anti-PABP was purchased from Abcam, Cambridge, MA. The proteasome inhibitor MG132 was purchased from Calbiochem, San Diego, CA. The general caspase inhibitor benzyloxycarbonyl-Val-Ala-Asp-(OMe) fluoromethylketone (Z-VAD-FMK) was purchased from R&D Systems, Minneapolis, MN. Purified poliovirus proteinase 3Cpro and purified coxsackievirus B3 2Apro were gifts of Richard Lloyd, Baylor College of Medicine (Joachims, Van Breugel, and Lloyd, 1999).
Cells were harvested into the culture medium with a plastic scraper, collected by centrifugation, washed with phosphate-buffered saline (PBS, 136.9 mM NaCl, 2.68 mM KCl, 8.1 mM Na2HPO4, 1.47 mM KH2PO4), and lysed in radio-immunoprecipitation buffer (RIPA, PBS containing 1% NP40, 0.5% sodium deoxycholate, and 0.1% sodium dodecyl sulfate). Aliquots containing 50 μg total protein were fractionated by 10% SDS-PAGE. Proteins were transferred electrophoretically to a nitrocellulose membrane, which was then incubated with antibody in PBS containing 5% nonfat milk for 2 h at room temperature. The membrane was washed in PBS containing 0.1% Tween, followed by addition of the appropriate secondary antibody. Proteins were visualized using the ECL chemiluminescence system (Amersham BioSciences, Piscataway, NJ).
This study was supported in part by supported in part by Public Health Service Grants AI50754 and T32AI07161 from the National Institute of Allergy and Infectious Diseases, and by NIGMS grant GM068448. DS is the Harrison Endowed Scholar in Cancer Research at the Massey Cancer Center. PBF holds the Thelma Newmeyer Corman Chair in Cancer Research at the Massey Cancer Center and is a Samuel Waxman Cancer Research Foundation (SWCRF) Investigator. We thank Richard Lloyd and Bert Semler for gifts of 2Apro and 3CDpro viral proteinases, respectively, and Bert Semler for the gift of poliovirus mutant Se1–3C-02.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.