The cDNAs coding for human H-Ras and RIG-I was obtained by PCR from HEK293T cells. H-RasV12 and H-RasN17 mutants were then generated using HRAS cDNA as a template via the PCR site-directed mutagenesis method (26). All wild-type and mutant HRAS cDNAs were cloned into the pcDNA3 expression vector. The RIG-I cDNA was cloned downstream of and in-frame with the Flag sequences of the pCMV2-Flag expression vector, whereas pcDNA3-Flag-MAVS was provided by Dr. Takashi Fujita (Department of Tumor Cell Biology, Tokyo Metropolitan Institute of Medical Sciences, Japan), pcDNA3.1-Flag-TBK1 was provided by Dr. John Hiscott (Lady Davis Institute for Medical Research, QC, Canada), and pIRF3-5D was provided by Dr. Yi-Ling Lin (Institute of Biomedical Sciences, Academia Sinica, Taiwan). The pIFN-β-Luc (containing the entire IFN-β promoter upstream of firefly luciferase gene) and the pIRF3-Luc (containing the IRF3 responsive element upstream of firefly luciferase gene) reporter plasmids were provided by Dr. Takashi Fujita, and the pNF-κB-Luc reporter plasmid was provided by Dr. Yi-Ling Lin (Institute of Biomedical Sciences, Academia Sinica, Taiwan). The pIFN-stimulated responsive element (ISRE)-Luc plasmid was purchased from Stratagene. Lentiviruses containing control shRNA or shRNAs targeting HRAS (#64: CCACCAGTACAGGGAGCAGAT, and #66: GTGTGTGTTTGCCATCAACAA) were purchased from the National RNAi Core Facility, Taiwan. To generate HRAS knockdown cells, the cells were infected with shRNA-containing lentiviruses at a multiplicity of infection (MOI) of 5 in the presence of polybrene (8 µg/ml). The infected cells were selected with puromycin (2.5 µg/ml) for 4 days, and the resulting cells were used for subsequent experiments.

The antibodies used for Western blot include the following: mouse anti-Flag M2 antibody (F3164) and mouse anti-β-actin monoclonal antibody (A5441) from Sigma-Aldrich, rabbit anti-TRAF3 antibody (C-20, H-20) and rabbit anti-IRF3 antibody (sc-9082) from Santa Cruz Biotechnology, rabbit anti-phospho-IRF3 (Ser 386) (ab76493) from Abcam, mouse anti-MAVS (ALX-804-847) from Enzo, and rabbit anti-H-Ras (GTX61164) from Genetex. Antibodies used for immunoprecipitation include the following: mouse anti-Flag M2 antibody (F3164) from Sigma-Aldrich, and rabbit anti-TRAF3 (H-20) from Santa Cruz.

Sendai virus and VSV were purchased from ATCC and amplified in the laboratory using embryonated chicken eggs and Vero cells, respectively. HuS-E/2 is an immortalized primary human hepatocyte cell line (27) (kindly provided by Dr. Kunitada Shimotohno, Research Center for Hepatitis and Immunology, National Center for Global Health and Medicine, Chiba, Japan) and was cultivated in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 20 mM Hepes, 15 μg/ml l-proline, 0.25 µg/ml insulin, 5 × 10⁻⁸ M dexamethasone, 44 mM NaHCO₃, 10 mM nicotinamide, 5 ng/ml EGF, 0.1 mM Asc-2P, and 1 µg/ml Plasmocin, whereas HEK293T, HeLa, and Vero cells were cultured in DMEM; all cultures were supplemented with 10% fetal bovine serum (FBS) and penicillin (100 U/ml), streptomycin (0.1 mg/ml), and amphotericin B (0. 25 µg/ml). The cells were cultivated at 37°C with 5% CO₂.

Plasmid DNAs were transfected into HEK293T cells using the calcium phosphate precipitation method. However, Lipofectamine 2000 (Invitrogen) was used to perform transfections in 96-well plates for reporter assays. To carry out luciferase activity assays from H-Ras knockdown or overexpressed cells, the cells (1 × 10⁴) were transfected in 96-well plates with pIFN-β-Luc, pIRF3-Luc, pNF-κB-Luc, or pISRE-Luc reporter DNA together with the indicated DNAs and the pTK-RL (Renilla luciferase) reporter plasmid, which was used as a transfection control. A control vector DNA was added, wherever it was needed, to make the total amount of transfected DNA equivalent in each transfection experiment. Twenty-four hours after transfection, cells were infected with SeV (for transfection with pIFN-β-Luc, pIRF3-Luc, or pNF-κB-Luc) or treated with 1,000 IU/ml IFN-α (for transfection with pISRE-Luc) for 24 h. Cells were assayed using Alamar Blue (AbD Serotec) to measure cell viability, and cell lysates were harvested for luciferase activity using a Dual-Glo Luciferase Assay System (Promega) according to the manufacturer’s instructions. The firefly luciferase activities were first normalized to Renilla luciferase activity and to cell viability, and the induction fold was calculated by dividing the normalized Luc values of the SeV-infected (or IFN-α-treated) cells with those of the uninfected (or untreated) cells in each group.

Total cellular RNA was extracted using Trizol reagent, and 1 µg of RNA was reverse transcribed with a random hexamer primer using HiScript I reverse transcriptase (Bionovas Biotechnology Co. Ltd.). One-twentieth of the volume of the cDNA product was then subjected to qPCR with gene specific primers. qPCR was conducted using Fast SYBR Green Master Mix (Thermo Fisher Scientific) according to the manufacturer’s instructions. The induction folds of the investigated genes were calculated as ΔΔCt, by normalizing the values to those of uninfected (or untreated) cells and to glyceraldehyde-3-phosphate dehydrogenase.

Activated GTP-bound H-Ras was detected using a Ras activation kit (Abcam, ab128504) according to the manufacturer’s instructions. In brief, cells (6 × 10⁵) transfected with different amounts of the HRAS-, HRASV12-, or HRASN17-expressing plasmid were lysed in 1 ml of ice-cold kit-provided lysis buffer containing protease inhibitors. Fifty microliters of lysate were removed, and 10 µl of 6× Laemmli sample buffer were added; this sample was referred to as the “total Ras load.” The remaining lysates were incubated with 40 µl of GST–RBD fusion protein-linked glutathione Sepharose beads, which had been pre-equilibrated with lysis buffer, at 4°C for 30 min. After centrifugation, the beads were washed once with 1 ml of ice-cold lysis buffer. Fifty microliters of 1× Laemmli sample buffer were added to the samples, which were referred to as “Ras-GTP pulldown” samples. The total Ras load and Ras-GTP pulldown samples were resolved on a 12.5% SDS-PAGE gel. The upper portion of the gel (protein sizes greater than 30 kDa) was stained with Coomassie blue to reveal the amounts of GST–RBD pulled down, whereas the lower part of the gel was processed for Western blot using an anti-H-Ras antibody to detect GTP-bound H-Ras.

Cells were lysed on ice for 20 min in lysis buffer containing 1% NP-40, 10 mM Tris–HCl (pH 7.4), 120 mM NaCl, and cocktail protease inhibitor (Roche). The lysates were centrifuged at 13,000 rpm at 4°C for 20 min, and supernatants were recovered for protein quantification with a BCA protein assay kit (Pierce). For immunoprecipitation, 500 µg of lysates were incubated with the indicated antibody at 4°C overnight. Thereafter, 10–20 µl of protein G Sepharose beads (GE Healthcare) were added, and samples were further incubated at 4°C for 2 h. The Sepharose beads were washed twice with ice-cold wash buffer containing 0.2% NP-40, 10 mM Tris–HCl (pH 7.4), and 120 mM NaCl, and heated at 95°C for 5 min in Laemmli sample buffer. After centrifugation, the supernatants were subjected to Western blot analysis.

HEK293T cells (3 × 10⁵) treated with lentiviruses containing control shRNA or shRNAs targeting HRAS were infected with VSV at an MOI of 0.001 in serum-free DMEM for 2 h. The medium was then replaced with complete medium containing 10% FBS. Twenty-four hours later, the cell supernatants were harvested, and the VSV titer in the supernatant was determined by plaque assays on Vero cells.

An SBV replicon containing a luciferase gene was constructed, and recombinant SBV-Luc virus was produced as previously described (23). HEK293T cells treated with control shRNA or shRNAs targeting HRAS were infected with recombinant SBV-Luc virus at an MOI of 30. Forty-eight hours after infection, Luc activity and cell viability were measured using the Bright-Glo Luciferase Assay System (Promega) and an Alamar Blue assay kit, respectively. The luciferase activity was normalized to the Alamar Blue data to reflect SBV-Luc replication activity.

The data were analyzed by one-way ANOVA followed by Dunnett’s post hoc test (Graphpad Prism 5 Software). *** indicates P < 0.001, ** indicates P < 0.01, and * indicates P < 0.05. P < 0.05 was regarded as statistically significant.

We have previously demonstrated that cells exhibiting low susceptibility to infection with SBV, a virus that is extremely sensitive to IFN activity, normally possess intact IFN responses (23). Thus, we rationalized that silencing the expression of molecules that positively regulate IFN responses should render cells more susceptible to SBV infection. By contrast, silencing negative regulators may further inhibit SBV infection in cells. Therefore, we have conducted a high-throughput shRNA screening using a recombinant SBV replicon expressing a luciferase (Luc) gene (SBV-Luc) to screen for novel molecules involved in IFN responses. A mini shRNA library, which targets 123 human oncogenes and tumor suppressor genes, was used for screening in this study. As illustrated in Figure S1A in Supplementary Material, the primary screen identified 57 genes, each of which had at least 2 shRNA clones simultaneously showing the z-scores greater than 1 or smaller than −1. All of the shRNAs of these candidate genes were proceeded to the secondary screen using the IFN-β-Luc reporter assay stimulated by N-RIG, the N-terminal 2CARD domain of RIG-I. The secondary screen identified 11 genes, each of which had at least 2 shRNA clones simultaneously exhibiting a twofold or more increase or reduction in the induction fold of IFN-β-Luc activity when compared to the empty vector-transduced cells. Among them, H-Ras was identified (Table S1 in Supplementary Material). Given that the z-scores of 7 out of 8 shHRAS clones were positive, albeit some of them were smaller than 1 (Figure S1B in Supplementary Material), we strongly speculated that H-Ras might play a positive role in regulating IFN-I responses. To confirm this thought, we conducted SeV-induced IFN-stimulated responsive element (ISRE)-Luc reporter assay by using the six most effective shHRAS clones based on the z-scores. As anticipated, depletion of H-Ras by these shRNA clones significantly reduced the SeV-stimulated ISRE activity (Figure S1C in Supplementary Material). As a result, H-Ras was chosen as a candidate and shHRAS#64 and #66 were used for the following studies.

First, we confirmed previous screening results by demonstrating that depletion of H-Ras by shRNA #64 or #66 (Figure 1A) led to a robust increase in Luc activity from the SBV replicon (Figure 1B). SBV is a positive-strand RNA virus and belongs to the Togaviridae family. To expand these findings, we further examined the effects of H-Ras depletion on infection with VSV, which is a negative-strand RNA virus and is a member of the Rhabdoviridae family. Similar to SBV infection, H-Ras depletion increased VSV replication in HEK293T cells as well (Figure 1C). Notably, increased Luc activity from the SBV replicon and increased VSV titers correlated with the decrease in H-Ras protein levels. Therefore, these data demonstrate that H-Ras exerts a positive regulatory role in antiviral activity.

To investigate the mechanism underlying the antiviral activity of H-Ras, we used these two shRNAs to further conduct reporter assays in HEK293T cells to understand whether H-Ras regulated the RLR signaling induced by SeV infection. The results showed that shRNA-mediated silencing of H-Ras significantly decreased the SeV-induced activities of IFN-β-Luc (Figure 2A), IRF-3-Luc (Figure 2B), and NF-κB-Luc (Figure 2C), suggesting that H-Ras indeed positively regulated RLR signaling. Consistently, H-Ras silencing reduced the SeV-induced IRF3 phosphorylation, leading to reduced RNA expression of downstream genes such as IFNB1 and IFIT1 (Figure 2D). To understand whether this phenomenon is specific to HEK293T cell only, we also examined the effect of HRAS knockdown on RLR signaling in two other cell lines. Similarly, in HuS-E/2 cells, an immortalized primary human hepatocyte cell line (Figure 2E) or in HeLa cells (Figure 2F), HRAS knockdown also led to reduced IRF3 phosphorylation albeit a little less in HuS-E/2 cells and reduced IFNB1 and IFIT1 gene expression upon SeV infection. Collectively, these data confirm a positive regulatory role of H-Ras in IFN-I responses and demonstrate that the stimulatory effect is not limited to a particular cell type.

Next, we examined the effects of ectopic expression of HRAS on SeV-induced RLR signaling using the IFN-β-Luc reporter assay. Surprisingly, we found that H-Ras protein influenced SeV-induced IFN-β promoter activity in a biphasic manner rather than strictly enhancing promoter activity as we originally anticipated (Figure 3A). The ectopically expressed H-Ras enhanced IFN-β promoter activity in a dose-dependent manner at low transfection doses, e.g., 10 and 20 ng/ml. However, IFN-β promoter activity was gradually inhibited by increasing the transfection dose to 100 or 400 ng/ml. Using the GST-fused RBD of Raf as an activation-specific probe for the detection of Ras-GTP (28) to reveal the activation status of ectopically expressed H-Ras in the cells, we found that H-Ras was activated to a certain extent when it was expressed to an extreme high level (Figure 3A).

As reported previously, oncogenic Ras exerts negative effects on IFN-I responses (24, 25). Thus, we speculated that the inhibition of SeV-induced IFN-β promoter activity at high levels of wild-type H-Ras might be caused by the negative effects of H-Ras activation on IFN responses. To confirm this, we examined SeV-stimulated IFN-β-Luc activity in the presence of ectopically expressed H-RasV12, which is the constitutively GTP-bound form of Ras, or H-RasN17, which is the constitutively GDP-bound form. The results demonstrated that, indeed, H-RasV12 inhibited SeV-induced IFN-β-Luc activity even at a low transfection dose such as 20 ng/ml, and that the inhibition was strongly correlated with the levels of activated Ras (Figure 3B). Conversely, H-RasN17 increased SeV-induced IFN-β-Luc activity in a dose-dependent manner (Figure 3C). Consistent with its constitutive negative activity, the activation-specific probe revealed minimal levels of GTP-bound Ras even when HRASN17 DNA was transfected at a high dose, e.g., 400 ng/ml. These results thus exclude the possibility that the inhibitory effects of wild-type H-Ras transfected at 400 ng/ml dose was due to a competitive expression between the transfected reporter DNA and HRAS DNA, but support for the argument that the H-Ras protein itself rather than its activity exerts a positive role, whereas activated H-Ras exerts a negative role in regulating RLR signaling. As a result, wild-type H-Ras, when expressed at low levels, functions in a manner similar to H-RasN17, which enhances RLR signaling. However, H-Ras becomes partially activated when artificially expressed to high levels, which resembles H-RasV12 and inhibits RLR signaling.

To dissect the step at which H-Ras acts to regulate the RLR signaling pathway, we expressed the Flag-tagged individual molecules in the signaling pathway, such as RIG-I, MAVS, TBK1, or IRF3-5D (a constitutively active form of IRF3), to induce the signaling and IFN-β-Luc activity under HRAS knockdown condition or in the presence of ectopically expressed wild-type H-Ras, H-RasV12, or H-RasN17 was determined. The constitutively active form of IRF3 was used because IRF3 per se would not be activated in the absence of upstream activated TBK1. The reporter assay data showed that H-Ras depletion significantly reduced IFN-β promoter activity induced by all inducers except IRF3-5D (Figure 4A). Ectopically expressed wild-type H-Ras exhibited a similar biphasic mode of regulation on the IFN-β-Luc activity mediated by all inducers (Figure 4B), whereas ectopically expressed H-RasV12 significantly inhibited IFN-β-Luc activity mediated by all inducers (Figure 4C). H-RasN17 demonstrated a lower effect and only upregulated MAVS- and TBK1-induced IFN-β-Luc activity (Figure 4D). Under all of the induction conditions, the expression levels of each inducer were more or less comparable within each experimental set, arguing against the possibility that the stimulatory or inhibitory effects were due to differential expression levels of the inducers. These data suggest that the H-Ras protein may function at a step upstream of IRF3, likely involving the mitochondria-associated platform. Conversely, the oncogenic H-Ras may affect not only the RLR pathway but also the signaling downstream of the RLR pathway, e.g., IFN signaling as previously reported (24, 25, 29).

Since H-Ras functions at a step upstream of IRF3, we hypothesized that the H-Ras protein might participate in the MAVS-associated platform, as a subpopulation of H-Ras proteins localize to the membranes of the ER, Golgi, endosome, and mitochondria (30–32). To understand whether H-Ras could possibly associate with any components within the MAVS-associated signalosome, we expressed H-Ras together with Flag-tagged RIG-I, MAVS, TBK1, or TRAF3 in HEK293T cells, and co-immunoprecipitation experiments were performed. However, none of these proteins were found to interact with H-Ras (data not shown). We further examined whether H-Ras played a role in the formation of the MAVS-associated signalosome. The formation of MAVS–TRAF3 complex stimulated by SeV infection was detected by co-immunoprecipitation experiments under H-Ras depletion conditions. As shown in Figure 5A, we were able to detect significant amounts of MAVS interacting with TRAF3 in the cells treated with control shRNA, upon SeV infection but not in the absence of viral infection. However, MAVS–TRAF3 interaction was reduced with the depletion of H-Ras by shHRAS#64 or #66. It was noteworthy that the total levels of MAVS proteins in the whole cell lysate were all discernibly reduced upon SeV infection, a phenomenon that has also been reported previously by Castanier et al. (33). In their study, they have found that the activation of RLR signaling induces the formation of MAVS-associated signalosome, which in turn leads to a proteasomal degradation of MAVS induced by TRIM25-mediated ubiquitination, thus freeing TBK1 to phosphorylate IRF3 (33). Hence, we preferred to use the ratios of immunoprecipitated MAVS/total MAVS, shown at the bottom of each knockdown group, to more accurately reflect the efficiency of the MAVS–TRAF3 complex formation. The results indicated that the depletion of H-Ras indeed reduced the formation of the MAVS–TRAF3 complex.

We also investigated the formation of the MAVS–TRAF3 complex under ectopic expression of H-RasV12 or H-RasN17. However, since the doses of RasV12 and RasN17 to exhibit their maximal effects on RLR signaling were different, the DNA amounts used for transfection of each molecule were titrated individually and different amounts of DNA were transfected here. As shown in Figure 5B, at the optimal effective dose, H-RasV12 was found to significantly suppress the MAVS–TRAF3 complex formation. H-RasN17 also seemed to reduce the apparent levels of MAVS–TRAF3 complex when compared to the control vector, but the total MAVS levels in the whole cell lysates were also significantly reduced, indicating an active transduction of the signal. Thus, the ratio of immunoprecipitated MAVS/total MAVS was actually increased, suggesting an increased efficiency of the MAVS–TRAF3 complex formation, by ectopic expression of H-RasN17. The results were consistent with those of reporter assays shown in Figures 3 and 4.

On the other hand, we would also like to confirm that oncogenic H-Ras downregulated IFN signaling pathway. We first carried out an IFN-stimulated responsive element (ISRE)-Luc reporter assay under IFN-α treatment conditions in the presence of H-RasV12 or H-RasN17 (used as a control). The results showed that H-RasV12 significantly inhibited, whereas H-RasN17 slightly upregulated, the IFN-α-induced ISRE promoter activity (Figure 6A). RT-qPCR analysis also indicated that H-RasV12 inhibited the expression of several ISGs downstream of IFN signaling, such as IFIT2, MX1, and OAS2, whereas H-RasN17 had minimal effects on or even enhanced the expression of these ISGs (e.g., OAS2) (Figure 6B).

We tried to investigate the mechanisms underlying the inhibitory activity of H-RasV12 on IFN-I signaling by examining whether H-RasV12 affected the activation levels of STAT1 and STAT2. However, no inhibition of IFN-α-stimulated STAT1 or STAT2 phosphorylation by H-RasV12 or by H-RasN17 was observed (Figure S2A in Supplementary Material). The total amounts of STAT1 or STAT2 proteins were also not reduced, either at 30 min or at 24 h after IFN-α treatment (Figure S2A in Supplementary Material). Thus, our data did not support the mechanism reported previously, in which H-RasV12 downregulated IFN signaling by reducing the total amounts of STAT2 (25). We further examined whether H-RasV12 influenced the translocation of the activated STAT1/STAT2 complex into the nucleus. The cytosolic and nuclear fractions of HRASV12- or HRASN17-transfected cells were separated at 1 h post IFN-α treatment, and phospho-STAT1 levels were determined by Western blot. However, we still did not observe significant differences in the translocation of phospho-STAT1 between control cells and cells expressing H-RasV12 or H-RasN17 (Figure S2B in Supplementary Material). Thus, although our data demonstrate that oncogenic H-Ras indeed impairs IFN-I signaling, the mechanism still remains elusive.

In addition to the enhancement of RIG-I-mediated signaling by H-Ras, we sought to verify whether H-Ras also enhanced IFN responses induced by MDA5- or STING (stimulator of interferon genes)-mediated signaling. For studying MDA5 signaling, we first transfected the HEK293T cells that were treated with control shRNA or shRNAs targeting HRAS with MDA5 cDNA since HEK293T cells express extremely low levels of MDA5 (23), and subsequently transfected them with long-chain poly(I:C) to induce MDA5 signaling. To induce STING signaling, the shRNA-treated HEK293T cells were transfected with poly(dA:dT). Expression of IFNB1 and IFIT1 RNA in response to ligand stimulation was analyzed by RT-qPCR. The results showed that depletion of H-Ras significantly downregulated the transcription of these two genes induced by poly(I:C) (Figure 7A) or by poly(dA:dT) (Figure 7B), indicating that H-Ras also positively regulates the IFN-I responses induced by MDA5 or STING signaling. On the other hand, depletion of H-Ras did not affect the TNF-α-induced CXCL8 expression (Figure 7C), or the IL-6-induced SOCS3 expression (Figure 7D), indicating that HRAS knockdown did not influence TNF-α or IL-6 signaling. Alternatively, luciferase reporter assays also demonstrated that depletion of H-Ras had no effects on the TNF-α-induced NF-κB-Luc activity (Figure 7E) or the IFN-γ-induced IRF1-Luc activity (Figure 7F), further indicating that H-Ras also did not influence IFN-γ signaling. Taken together, these results strongly suggest that H-Ras exhibits specific effects on IFN-I responses, and ruled out the possibility that the observed suppressive effects on IFN-I responses by H-Ras knockdown was due to a reduced cell proliferation effect.