Adenovirus Virus-Associated RNAs Induce Type I Interferon Expression through a RIG-I-Mediated Pathway

Abstract

The current study demonstrates that adenovirus virus-associated RNA (VA) is recognized by retinoic acid-inducible gene I (RIG-I), a cytosolic pattern recognition receptor, and activates RIG-I downstream signaling, leading to the induction of type I interferons (IFNs), similarly to Epstein-Barr virus-encoded small RNA. Further analysis revealed that adenovirus infection leads to biphasic type I IFN induction at 12 to 24 h and 48 to 60 h postinfection. The later induction coincided with VA expression and was reduced by virus UV inactivation or RIG-I silencing. These results suggest that VA-mediated RIG-I activation is involved in activating innate immune responses during adenovirus infection.

Adenoviruses, nonenveloped, double-stranded DNA (dsDNA) viruses of the Adenoviridae family, activate innate immunity through viral capsid or genomic DNA, leading to production of interferons (IFNs) and inflammatory cytokines that play key roles in antiviral responses (4, 7, 14, 18, 25, 26, 31).

Adenovirus virus-associated RNA I (VAI) and VAII, 157 and 158 nucleotides in length, respectively, are short noncoding RNAs transcribed by RNA polymerase III and expected to form dsRNA-like secondary structures (3). Retinoic acid-inducible gene I (RIG-I) is a cytosolic pattern recognition receptor (PRR) that senses pathogen-associated molecular patterns on viral short dsRNA and 5′-triphosphorylated RNA (8, 20, 30). The binding of such RNAs to RIG-I is a trigger for antiviral immune responses, such as induction of type I IFNs and inflammatory cytokines through activation of IFN regulatory factor 3 (IRF-3) and nuclear factor κB (NF-κB) (29). Epstein-Barr virus (EBV)-encoded small RNA (EBER) has many similarities to VA. EBERs are also short noncoding RNAs transcribed by RNA polymerase III and expected to form dsRNA-like structures (6, 21). In addition, VA and EBER have been reported to competitively block the antiviral effect of IFN through inhibition of dsRNA-activated protein kinase R (PKR) in the same manner (11, 16, 17, 19, 23). Recently, we reported that EBER activates RIG-I, leading to type I IFN induction, in Burkitt's lymphoma cell lines (22). In this study, we assessed whether, like EBERs, VAs induce type I IFN through activation of RIG-I signaling in adenovirus-infected cells.

EBER1 and EBER2 used in this study were synthesized by an in vitro transcription method as described previously (22). To synthesize VAI and VAII, T7 promoter-tagged VAI and VAII were amplified by PCR from a pAdVAntage vector containing VAI and VAII genes (Promega, Madison, WI) (primer pairs: for VAI, 5′-GGGGGTAATACGACTCACTATAGGGGGGGGCACTCTTCCGTG-3′ and 5′-AGGAGCGCTCCCCCGTTGTC-3′, and for VAII, 5′-GGGGGTAATACGACTCACTATAGGGGGGGCTCGCTCCCTGTA-3′ and 5′-AGGGGCTCGTCCCTGTTTCCG-3′). The PCR products were used as a template, and in vitro transcription was carried out according to the manufacturer's protocol (Epicentre Biotechnologies, Madison, WI). Figure 1A shows these synthesized RNAs (50 ng each) after electrophoresis in a 5% denaturing polyacrylamide gel (7 M urea). Synthesized RNA or poly(I:C) (0.1 μg) was transfected into human gastric carcinoma-derived NU-GC-3 cells (2) in 24-well plates using Lipofectamine 2000 reagent (Invitrogen, Carlsbad, CA). After the times of culture indicated in Fig. 1B, the expression of type I IFN in these cells was analyzed by reverse transcriptase PCR (RT-PCR). For RT-PCR, reverse transcription was performed using Moloney murine leukemia virus (MMLV) reverse transcriptase (Invitrogen), a random 6-mer primer (Takara, Otsu, Japan), and gene-specific 3′-end primers for VAs and EBERs. The primers (and conditions) used for PCR were as follows: for IFN-α, 5′-ATCCAGCAGATCTTCAATCT-3′ and 5′-AAGAAAAAGATCTCATGATT-3′ (40 cycles); for IFN-β, 5′-GATTCATCGAGCACTGGCTGG-3′ and 5′-CTTCAGGTAATGCAGAATCC-3′ (38 cycles); for IFN-stimulated gene 15 (ISG15), 5′-GGTGGACAAATGCGACGAAC-3′ and 5′-ATGCTGGTGGAGGCCCTTAG-3′ (26 cycles); for ISG56, 5′-TAGCCAACATGTCCTCACAGAC-3′ and 5′-TCTTCTACCACTGGTTTCATGC-3′ (30 cycles); for EBER1, 5′-AGGACCTACGCTGCCCTAGA-3′ and 5′-AAAACATGCGGACCAGC-3′ (26 cycles); for EBER2, 5′-AGGACAGCCGTTGCCCTAGT-3′ and 5′-AAAAACAGCGGACAAGCCGA-3′ (26 cycles); for RIG-I, 5′-GCATATTGACTGGACGTGGCA-3′ and 5′-CAGTCATGGCTGCAGTTCTGTC-3′ (30 cycles); for VAI, 5′-GGGCACTCTTCCGTGGTCTG-3′ and 5′-AGGAGCGCTCCCCCGTTGTC-3′ (26 cycles); for VAII, 5′-GGCTCGCTCCCTGTAGCCGG-3′ and 5′-AGGGGCTCGTCCCTGTTTCC-3′ (26 cycles); and for glyceraldehyde-3-phosphate dehydrogenase (GAPDH), 5′-CCATGCCATCACTGCCACCC-3′ and 5′-GCCAGTGAGCTTCCCGTTCAG-3′ (25 cycles). Single or combined VAI and VAII transfection, as well as EBER1/EBER2 cotransfection, induced IFN-β, IFN-α, ISG56, ISG15, and RIG-I (Fig. 1B). Next, we assessed the level of VAs needed to stimulate expression of IFN. As shown in Fig. 1C, VAI/II clearly induced IFN-β expression in a dose-dependent manner, and 0.001 μg VAI/II was sufficient for induction.

Fig. 1.

VAI and VAII induce type I IFNs in a RIG-I-dependent manner. (A) Detection of in vitro-synthesized VAI, VAII, EBER1, and EBER2 in a 5% denaturing polyacrylamide gel (7 M urea). (B) Effect of in vitro-synthesized VAI and VAII on the induction of type I IFNs. NU-GC-3 cells in 24-well plates were transfected with 0.1 μg of poly(I:C), EBER1, EBER2, EBER1 and EBER2 (1:1), VAI, VAII, or VAI and VAII (1:1). The expression of IFNs, ISGs, and RIG-I was examined at 6, 12, and 24 h posttransfection by RT-PCR. Control indicates the treatment with only Lipofectamine 2000. Three or more independent experiments were performed. (C) Dose-dependent effect of VAI and VAII transfection on IFN-β expression. NU-GC-3 cells in 24-well plates were transfected with 0.001, 0.01, or 0.1 μg of VAI and VAII (1:1). The expression of IFN-β and RIG-I was examined at 24 h posttransfection by RT-PCR. “Non” indicates the nontransfected cells, and “control” indicates the Lipofectamine 2000-treated cells. Three or more independent experiments were performed. (D) Effect of RIG-I knockdown on VAI- and VAII-induced IFN-β expression. NU-GC-3 cells in 24-well plates were transfected with control siRNA (cont) or RIG-I siRNA (RIG). After 24 h, the cells were transfected with 0.1 μg of poly(I:C), EBER1 and EBER2 (1:1), VAI, VAII, or VAI and VAII (1:1). The expression of IFN-β and RIG-I was examined at 24 h posttransfection by RT-PCR. Control indicates the treatment with only Lipofectamine 2000. Three or more independent experiments were performed. (E) Effect of RIG-I knockdown on VAI- and VAII-induced IFN-β production. NU-GC-3 cells in 24-well plates were transfected with control siRNA or RIG-I siRNA. After 24 h, the cells were transfected with 0.1 μg of VAI and VAII (1:1). The amount of IFN-β in culture supernatants was examined at 24 h posttransfection by ELISA. Error bars indicate the standard deviations (SD) for duplicate wells. The data presented are representative of three independent experiments. (F) Effect of stable VAI and VAII expression on type I IFN expression. NU-GC-3 cells were stably cotransfected with the pAdVAntage vector and an SV40 promoter-driven hygromycin-resistant gene vector. The expression of IFNs, ISGs, and RIG-I in a control cell line and two independent VAI/II stably transfected cell lines was examined by RT-PCR (left). The expression level of VAI and VAII in two independent VAI/II stably transfected cell lines (cell line) was compared to that in cells transiently transfected with 0.001 or 0.01 μg of VAI and VAII (1:1) by RT-PCR. “Non” indicates the nontransfected cells (right). Three or more independent experiments were performed. (G) Effect of RIG-I knockdown on IFN-β expression in VAI/II stably transfected cell lines. NU-GC-3 cells stably transfected with VAI and VAII were transfected with control siRNA (cont) or RIG-I siRNA (RIG). After 48 h, the expression of IFN-β and RIG-I was examined by RT-PCR. Three or more independent experiments were performed. (H) Effect of RIG-I knockdown on IFN-β production of VAI/II stably transfected cell lines. NU-GC-3 cells stably transfected with VAI and VAII were transfected with control siRNA or RIG-I siRNA. After 48 h, the amount of IFN-β in culture supernatants was examined by ELISA. Error bars indicate the SD for duplicate wells. The data presented are representative of three independent experiments.

To examine whether RIG-I is involved in the induction of IFNs by VAs, NU-GC-3 cells in 24-well plates were transfected with RIG-I small interfering RNA (siRNA) (sense, 5′-UAAGGUUGUUCACAAGAAUCUGUGG-3′, and antisense, 5′-CCACAGAUUCUUGUGAACAACCUUA-3′) or control siRNA (Stealth RNAi negative control; Invitrogen) using Lipofectamine RNAiMAX reagent (Invitrogen). After a 24-h transfection, the cells were transfected with 0.1 μg of VAs, EBERs, or poly(I:C) and cultured for 24 h. RIG-I knockdown reduced the induction of IFN-β by VA or EBER transfection at the mRNA level (Fig. 1D). Under the same conditions, IFN-β in culture supernatants was quantified by enzyme-linked immunosorbent assay (ELISA) using a human IFN-β ELISA kit (PBL Interferon Source, Piscataway, NJ) according to the manufacturer's protocol. As shown in Fig. 1E, RIG-I knockdown reduced the IFN-β induction by VA transfection at the protein level, indicating that RIG-I is required for induction of type I IFN by VAs.

Subsequently, to assess the effect of VAs synthesized in cultured cells on IFN induction, VAI/II stably transfected cell lines were established. A control or pAdVAntage plasmid and a simian virus 40 (SV40) promoter-driven hygromycin-resistant gene plasmid were cotransfected into NU-GC-3 cells. The cells were cultured in medium containing 300 μg/ml hygromycin B (Merck, Frankfurt, Germany) to select hygromycin-resistant cell clones. As shown in Fig. 1F, stable expression of VAs resulted in upregulation of type I IFNs, RIG-I, and ISGs. Furthermore, knockdown of RIG-I resulted in reduced expression of IFN-β in the VAI/II stably transfected cells at both the mRNA and the protein level (Fig. 1G and H). These results suggest that VAs synthesized in NU-GC-3 cells also upregulate type I IFN expression through RIG-I. The level of VAs expressed by transfection with 0.01 μg of RNA was much higher than that in cells stably expressing VAs (Fig. 1F). This is why IFN-β expression in cells stably transfected with VAs is much lower than that in cells transfected with 0.1 μg of RNA (Fig. 1E and H).

Next, to analyze whether VAs bind with RIG-I, a FLAG-tagged RIG-I-expressing plasmid was constructed by subcloning human RIG-I cDNA into the pCMV-Tag2A vector (Stratagene, La Jolla, CA). This, or a control plasmid, was transiently transfected into NU-GC-3 cells stably transfected with VAs. After a 48-h transfection, the cells were subjected to coimmunoprecipitation using anti-FLAG antibody (Sigma-Aldrich, St. Louis, MO) as described previously (17, 22). As shown in Fig. 2, full-length VAs were strongly detected in immunoprecipitates of RIG-I-transfected cells compared to those of control plasmid-transfected cells, indicating that VAs bind with RIG-I in NU-GC-3 cells.

Fig. 2.

Binding of VAI and VAII with RIG-I. VAI/II stably transfected cells were transfected with FLAG-tagged RIG-I. After UV irradiation, the cell lysates were immunoprecipitated with anti-FLAG antibody. RNA was extracted from the immunoprecipitates and subjected to RT-PCR to detect full-length VAI and VAII, 157 and 158 nucleotides, respectively. The results were reproducible in four independent experiments.

Next, it was determined whether IRF-3 and NF-κB, which function in downstream RIG-I-mediated signaling, were activated by VAs. NU-GC-3 cells were transfected with VAI/II, EBER1/2, or poly(I:C) and, after the times of culture indicated in Fig. 3A, subjected to immunoblot analysis. For immunoblot analysis, we used primary antibodies directed against phospho-IRF-3 (Ser396), phospho-NF-κB p65 (Ser536), β-actin (Cell Signaling Technology, Beverly, MA), and IRF-3 (Abcam, Cambridge, United Kingdom). As shown in Fig. 3A, IRF-3 was phosphorylated in VAI/II- and EBER1/2-transfected cells. In contrast, NF-κB was highly phosphorylated in NU-GC-3 cells regardless of transfection with VAI/II or EBER1/2, and therefore the effect of NF-κB could not be evaluated (Fig. 3A).

Fig. 3.

VAs induce IFN-β through IRF-3 activation. (A) Effect of VAI and VAII on the activation of IRF-3 and NF-κB. NU-GC-3 cells were transfected with poly(I:C), EBER1 and EBER2 (1:1), or VAI and VAII (1:1). Phospho-IRF-3 (pIRF-3), IRF-3, and phospho-NF-κB (pNF-κB) were detected by immunoblotting using cell lysates harvested at the indicated time points. The data presented are representative of three independent experiments. (B) Effect of IRF-3 knockdown on VA-induced IFN-β expression. NU-GC-3 cells were transfected with control siRNA (cont) or three IRF-3 siRNAs (IRF). After 24 h, the cells were transfected with poly(I:C), EBER1 and EBER2 (1:1), or VAI and VAII (1:1). The expression of IFN-β and IRF-3 was examined at 24 h posttransfection by RT-PCR. Control indicates the treatment with only Lipofectamine 2000. Three or more independent experiments were performed. (C) Effect of IRF-3 knockdown on VA-induced IFN-β production. NU-GC-3 cells in 24-well plates were transfected with control siRNA or three IRF-3 siRNAs. After 24 h, the cells were transfected with 0.1 μg of VAI and VAII (1:1). The amount of IFN-β in culture supernatants was examined at 24 h posttransfection by ELISA. Error bars indicate the SD for duplicate wells. The data presented are representative of three independent experiments. (D) Effect of IRF-3 knockdown on IFN-β expression in VAI/II stably transfected cell lines. NU-GC-3 cells stably transfected with VAI and VAII were transfected with control siRNA (cont) or three IRF-3 siRNAs (IRF). After 48 h, the expression of IFN-β and IRF-3 was examined by RT-PCR. Three or more independent experiments were performed. (E) Effect of IRF-3 knockdown on IFN-β production of VAI/II stably transfected cell lines. NU-GC-3 cells stably transfected with VAI and VAII were transfected with control siRNA or IRF-3 siRNA. After 48 h, the amount of IFN-β in culture supernatants was examined by ELISA. Error bars indicate the SD for duplicate wells. The data presented are representative of three independent experiments.

To confirm whether VAs induce IFN through IRF-3 activation, NU-GC-3 cells were transfected with three IRF-3 siRNAs (5′-UAAACGCAACCCUUCUUUGCGGUUG-3′ [sense] and 5′-CAACCGCAAAGAAGGGUUGCGUUUA-3′ [antisense], 5′-AUCAGAAGUACUGCCUCCACCAUUG-3′ [sense] and 5′-CAAUGGUGGAGGCAGUACUUCUGAU-3′ [antisense], and 5′-AACUCAUCCAGAAUGUCUUCCUGGG-3′ [sense] and 5′-CCCAGGAAGACAUUCUGGAUGAGUU-3′ [antisense]) or control siRNA. After a 24-h transfection, the cells were transfected with VAI/II, EBER1/2, or poly(I:C) and cultured for 24 h. The following primers were used for amplification of IRF-3 mRNA: 5′-CACAGCAGGAGGATTTCGG-3′ and 5′-CCTGGGTATCAGAAGTAC-3′ (28 cycles). IRF-3 knockdown reduced the induction of IFN-β by VAI/II or EBER1/2 transfection at the mRNA level (Fig. 3B). At the protein level, IRF-3 knockdown reduced the IFN-β induction by VAI/II transfection (Fig. 3C). Furthermore, in VAI/II stably transfected cells, IRF-3 knockdown resulted in reduced expression of IFN-β at both the mRNA and the protein level (Fig. 3D and E). These results suggest that VAs upregulate type I IFN expression in NU-GC-3 cells through IRF-3 activation.

Subsequently, to test whether adenovirus infection induces IFN, NU-GC-3 cells were infected with β-galactosidase-expressing E1-deleted type 5 adenovirus (29) at a multiplicity of infection (MOI) of 1,000, and induction of type I IFN expression was analyzed by sequential RT-PCR. At this MOI, more than 90% of cells were successfully infected, as confirmed by X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside) staining (data not shown). VAs were first detected at 12 h postinfection (hpi) and maximally detected at 48 to 60 hpi in adenovirus-infected cells. Biphasic induction of IFN-β and ISGs was observed in adenovirus-infected cells at 12 to 24 hpi and 48 to 60 hpi (Fig. 4A). The level of VAs induced by adenovirus infection was comparable with that expressed by transfection with 0.001 μg of RNA, which was sufficient for IFN induction (Fig. 4A and 1C). Furthermore, to assess whether viral gene transcription is important for induction of IFN, adenovirus inactivated by 10 cycles of 1 J/cm² UV irradiation in a UV cross-linker was used to treat NU-GC-3 cells. UV inactivation had no effect on early induction of IFN-β (12 to 24 hpi) by adenovirus infection. In contrast, UV-inactivated adenovirus did not induce IFN-β in the late phase (48 to 60 hpi) (Fig. 4B). These results indicate that viral gene transcription is required for the late-phase induction of IFN-β, but not early-phase induction, in adenovirus-infected cells.

Fig. 4.

Adenovirus infection leads to biphasic induction of IFN-β, and the late induction depends on RIG-I and IRF-3. (A) IFN-β induction by adenovirus infection. NU-GC-3 cells were infected with β-galactosidase-expressing E1-deleted type 5 adenovirus (MOI of 1,000). The expression of IFN-β, ISGs, and RIG-I was examined by RT-PCR at the indicated time points postinfection (left). The expression level of VAI and VAII in adenovirus-infected cells at 60 hpi was compared to that in cells transfected with 0.001 or 0.01 μg of VAI and VAII (1:1) by RT-PCR. “Non” indicates the nontransfected cells (right). Three or more independent experiments were performed. (B) IFN-β induction by UV-inactivated adenovirus treatment. NU-GC-3 cells were treated with adenovirus or UV-inactivated adenovirus. The expression of IFN-β and RIG-I was examined by RT-PCR at the indicated time points posttreatment. Three or more independent experiments were performed. (C) Effect of RIG-I knockdown on IFN-β induction by adenovirus infection. NU-GC-3 cells were transfected with control siRNA or RIG-I siRNA. After 12 h, the cells were infected with adenovirus (MOI of 1,000). The expression of IFN-β and RIG-I was examined by RT-PCR at the indicated time points postinfection. Three or more independent experiments were performed. (D) Effect of IRF-3 knockdown on IFN-β induction by adenovirus infection. NU-GC-3 cells were transfected with control siRNA or three IRF-3 siRNAs. After 12 h, the cells were infected with adenovirus (MOI of 1,000). The expression of IFN-β and IRF-3 was examined by RT-PCR at the indicated time points postinfection. Three or more independent experiments were performed. (E) Effect of RIG-I and IRF-3 knockdown on IFN-β production of adenovirus-infected cells. NU-GC-3 cells were transfected with control siRNA, RIG-I siRNA, or three IRF-3 siRNAs. After 12 h, the cells were infected with adenovirus (MOI of 1,000). The amount of IFN-β in culture supernatants was examined by ELISA at the indicated time points postinfection. Error bars indicate the SD for duplicate wells. The data presented are representative of three independent experiments.

Finally, we tested whether the late induction of IFN-β by adenovirus infection depends on RIG-I and IRF-3. RIG-I siRNA, IRF-3 siRNA, or control siRNA was transfected into NU-GC-3 cells, and 12 h after transfection, the cells were infected with adenovirus. Knockdown of RIG-I or IRF-3 reduced the late induction of IFN-β in adenovirus-infected cells at both the mRNA and the protein level (Fig. 4C, D, and E), indicating that the late induction of type I IFN in adenovirus-infected cells depends on both RIG-I and IRF-3. These results suggest that VAs induce type I IFN through the RIG-I-mediated pathway during adenovirus infection.

In this study, we have demonstrated that VA induces type I IFN through activation of RIG-I, as EBER does. EBER and VA are similar in that they are expected to form dsRNA-like secondary structures consisting of many short stem-loops (3, 6, 21). Short dsRNA and 5′-triphosphorylated RNA have been reported to activate RIG-I (8, 20, 30). Since VAI contains triphosphorylated 5′ termini (27), both short stems and triphosphorylated 5′ termini on VA may conceivably cooperate in the activation of RIG-I. More recently, Yamaguchi et al. showed that VA induces type I IFN in a RIG-I-independent manner in adenovirus-infected mouse embryonic fibroblasts (28), suggesting that the different PRRs according to the cell types recognize VA.

Adenovirus infection was further demonstrated to activate RIG-I. Induction of inflammatory genes after adenovirus infection occurred biphasically in vivo, within 24 h and in 3 to 4 days, as reported previously (13). The early induction occurred even after inoculation with UV-inactivated viruses, indicating that virus adsorption or penetration was sufficient to induce the inflammatory genes. In contrast, the late induction required intact viruses, was dependent on RIG-I, and coincided with the appearance of VA. Therefore, although further studies using VA-deficient adenovirus are needed, it is most probable that VA is responsible for the late induction of IFN.

Many single-stranded RNA viruses have been reported to activate RIG-I (9, 10, 15, 24). The present findings have demonstrated that DNA viruses also activate RIG-I signaling, when they carry noncoding RNAs (22). Although Cheng et al. (5) reported that herpes simplex virus type 1 (HSV-1) and adenovirus dsDNA activated RIG-I signaling, Ablasser et al. (1) concluded that RIG-I activation by these DNAs was induced by RNA polymerase III-transcribed RNA intermediates. It remains to be determined whether, in addition to VA, other RNA intermediates contribute to RIG-I activation during adenovirus infection.

Adenovirus vectors have been used in human gene therapy as a tool to transduce foreign genes into target cells (12). Since VA activates innate immune responses, as shown in this study, adenovirus vectors in which the VA gene is deleted may be more suitable tools for gene therapy.