Suppression of the IFN-α and -β Induction through Sequestering IRF7 into Viral Inclusion Bodies by Nonstructural Protein NSs in Severe Fever with Thrombocytopenia Syndrome Bunyavirus Infection

The innate immunity is the first line of defense against invading pathogens (1). Host cells are induced to produce IFNs in response to infection by viruses, which subsequently induce a variety of antiviral components in neighboring cells to suppress virus replication with different mechanisms (2). IFNs are divided into at least three distinct groups, type I, II, and III. Type I IFNs are composed of the members including IFN-α, -β, and the others, such as IFN-ε, -ω, and -κ (3, 4), which are primarily regulated at the gene transcription level. Transcription factors IRF 3 and IRF7 have been considered as key regulators of type I IFN expression during viral infections. Although IRF3 and IRF7 are highly homologous, IRF3 is constitutively expressed and resides in the cytosol in a latent form, which is phosphorylated and activated by tank-binding kinase 1 (TBK1) in response to viral infection. In contrast, IRF7 is expressed at a low amount or undetectable in most cells but could be vigorously induced by IFN-β and activated by inhibitor of nuclear factor kappa B kinase subunit epsilon (IKKε) (5).

Upon viral infection, both IRF3 and IRF7 undergo phosphorylation, dimerization, and nuclear translocation to activate IFN gene transcription (6, 7). IRF3 and IRF7 may have different roles in the induction of type I IFNs. In irf7^−/− mice, IFN-α induction was completely abolished, and the induction of IFN-β was markedly inhibited in plasmacytoid dendritic cells (pDC) in response to viral infections. In contrast, in irf3^−/− mice, the induction of IFN-α was not affected, whereas IFN-β was moderately inhibited, indicating that IFN-α was dependent on IRF7, whereas IRF3 is dispensable (8). Type I IFNs differ in their protection against certain types of viral infections, as well. In irf7^−/− mice infected with West Nile virus (WNV), IFN-α induction was severely compromised, which correlated to earlier and elevated viral burdens in both peripheral and CNS tissues and increased viral titers in primary macrophages, fibroblasts, dendritic cells, and neurons. However, IFN-β response remained intact in WNV-infected irf7^−/− mice (9), indicating that IFN-α, but not IFN-β, was protective in WNV-infected mice.

IRF7 is not constitutively expressed in many cell types, unlike IRF3. IFN-α induction would, therefore, rely on the activation of IRF3, which subsequently stimulates the transcriptional induction of IFN-β and further induces the production of IRF7 as a positive feedback loop, critical for a full-scale launch of a sustained antiviral innate immunity at the late stage of infection, in particular for viral clearance. In this scenario, IRF-3 is primarily responsible for the induction of IFN-β genes in the early phase, and IRF-7 is involved in the induction of both IFN-α and -β at a later phase of infection (4, 10). In infections in which IFN-α is essential to bring the virus under control, such as WNV (9), induction of IRF7 and its proper function would be essential in host defense. With IRF7 in deficiency, IFN-α would not be produced, and a lethal infection would follow by WNV even though IRF3 was not affected, and the IFN-β production remained intact (9). This does not merely occur in animal models but also happens in human. It was found that a severe infection occurred in a patient who has a dysfunctional IRF7 as a result of a point mutation in the IRF7 gene while contracting influenza virus (11). Viruses have evolved and developed numerous strategies to counteract host defense. However, little has been known about how viruses target IRF7 to evade host immunity.

The severe fever with thrombocytopenia syndrome virus (SFTSV) is a tick-borne emerging pathogen, causing high fever and drastic loss of leukocytes and platelets and multiorgan failure with high mortality in severe patients (12). There have been over 3000 cases reported in China since 2009 (13), and the virus has also been isolated recently in Japan and Korea and from patients and ticks (14, 15). SFTSV possesses a genome with three ssRNA segments of negative sense, L, M, and S. Whereas the L segment encodes an RNA-dependent RNA polymerase, and M encodes two envelope proteins, Gn and Gc, the S segment, is an ambisense RNA with two opposite open reading frames encoding a nucleoprotein (NP) and a nonstructural protein (NSs) (12). Previous studies have shown that NSs was inhibitory to antiviral IFN induction and activation (16–20) and could interact with TBK1, IKKε, and IRF3 directly or indirectly (16, 17), which has been implicated in viral evasion of innate immunity. Interestingly, our data indicated that NSs could form unique cytoplasmic inclusion bodies, in which viral RNA and NP as well as cellular TBK1, IKKε, and IRF3 could be identified (21). This suggests that SFTSV may sequester components of the IFN signaling pathways in the inclusion bodies to suppress IFN induction, which was corroborated by studies from other laboratories (17, 18). A recent study showed that the IFN response was affected by a similar mechanism in which NSs brought STAT 1/2 into the inclusion bodies (18).

We realized that IFN-α and -β may play different roles in antiviral infection, which could be differentially regulated by IRF3 and IRF7 in different tissue types. In infection with SFTSV, we understand how IRF3 interacts with NSs. However, no study has been performed to examine how IRF7 is regulated, which might affect the induction of IFN-α as well as IFN-β. In this study, we report in several cell types that IRF7 was induced by SFTSV infection, which promoted IFN-β and, in particular, IFN-α2 and -α4 induction. Unlike IRF3, IRF7 directly interacted with and sequestered NSs to the inclusion bodies, demonstrating a view in which host innate immunity is suppressed by SFTSV through host–virus interaction, a characteristic strategy by this novel bunyavirus.

Human embryonic kidney (HEK) 293T cells, HeLa cells, and African green monkey kidney cells (Vero) were grown in DMEM (Life Technologies/Invitrogen, Carlsbad, CA) supplemented with 10% FBS (HyClone, Logan, UT), 1 mM sodium pyruvate (HyClone), and 1% antibiotic-antimycotic solution (Life Technologies). Human THP-1 cells were grown in RPMI Medium 1640 (Life Technologies) supplemented with 10% FBS, 1 mM sodium pyruvate, and 1% antibiotic-antimycotic solution. Cells were cultured at 37°C with 5% CO₂. SFTSV strain JS-2010-014, as described previously (22), and an avian influenza virus strain subtype H9N2 (hereafter referred to as AIV H9N2), also described before, were used in this study. All viral aliquots were stored at −80°C.

An anti-IRF7 Ab was purchased from Abcam (Cambridge, MA), secondary Abs with fluorescence Alex Fluor-488–labeled donkey anti-mouse and Alex Fluor-594–labeled donkey anti-rabbit IgG were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-Flag M2 Ab, fluorescence Nile red dye, and DAPI were obtained from Sigma-Aldrich (St. Louis, MO). Anti-GAPDH and anti-HA Abs were purchased from Santa Cruz Biotechnology. Other secondary Abs used for immunoblots, confocal microscopy, and immunoprecipitation were obtained from Cell Signaling Technology (Danvers, MA). Rabbit and mouse antiviral NSs Abs were obtained by immunizing rabbits or mice with purified recombinant NSs of SFTSV, which were further purified with IgG affinity columns, also used in the previous studies (16). Protein A+G agarose beads were purchased from Beyotime Biotechnology (Nanjing, China). A PrimeScript (R036A) kit for reverse transcription was obtained from Takara (Shiga, Japan) and used for cDNA synthesis. TRIzol reagent and Lipofectamine 2000 reagents were purchased from Invitrogen.

A total of 1.0–1.5 × 10⁵ HEK 293T cells were seeded on six-well plates. Twenty-four hours later, each well was transfected with 200 pmol short interfering RNA (siRNA; GenePharma) and 5 μl of Lipofectamine 2000 in 200 μl of Opti-MEM (Life Technologies). The siRNA sequences used in this study for knockdown of IRF7 gene was obtained from a previously published study (23). The siRNA sequence targeting IRF7 and scrambled siRNA sequence as a negative control were as follows: IRF7 sense 5′-CGAGCUGCACGUUCCUAUATT-3′ and antisense 5′-UAUAGGAACGUGCAGCUCGTT-3′ and scramble control sense 5′-UUCUCCGAACGUGUCACGUTT-3′ and antisense 5′-ACGUGACACGUUCGGAGAATT-3′. Efficiency of mRNA or protein knockdown was measured by real-time RT-PCR or Western blot analysis, respectively. Forty-eight hours after siRNA transfection, HEK 293T cells were infected with 1 multiplicity of infection (MOI) of SFTSV. For studies with additional IFN-β pretreatment, HEK 293T cells were incubated with 100 U/ml IFN-β (PBL Assay Science, Piscataway, NJ) for 1 h. Next, RNA and protein were collected at indicated time points, and real-time RT-PCR and Western blot analysis were performed as previously described (23). The supernatants were harvested from cell cultures, and IFN-β concentrations were detected using commercial ELISA kits (Cusabio Biotech, Wuhan, China).

Cell lysates of transfected or SFTSV-infected cells were incubated with specific or control Abs at 4°C overnight and then precipitated with protein A+G agarose beads. After 2 h of incubation, the beads were washed four times with a lysis buffer containing 25 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1 mM EDTA, 1% Nonidet P-40, 5% glycerol, 1 mM DTT, 1 mM PMSF, 2 mM NaF, 1 mM Na₃VO₄, and 1 μg of aprotinin/ml. The immunoprecipitates were electrophoresed on SDS-PAGE and then transferred onto Immunoblot PVDF membranes (MilliporeSigma, Billerica, MA) for a primary Ab incubation at 4°C overnight. HRP-conjugated secondary Abs were used for further incubation with the membranes for 120 min, and signals on blots were developed by the ECL reagents (MilliporeSigma).

To prepare soluble and insoluble cytoplasmic fractions of cell lysates, we used a Keygen Nuclear and Cytoplasmic Protein Extraction Kit (KeyGen Biotech, China) to obtain the cytoplasmic soluble fraction. Procedures were performed as the manufacturer’s instructions indicated. Cell pellets were resuspended in a buffer containing 1% SDS and 10 mM Tris-EDTA (pH 7.5), which became the cytoplasmic insoluble fraction. Proteins from cytoplasmic soluble and insoluble fractions were quantified using a BCA Protein Assay Kit (Beyotime, Nanjing, China) before loading for SDS-PAGE.

HeLa cells, transfected with plasmids encoding the cDNA of IRF7, TBK1, and NSs or infected with SFTSV were washed twice with PBS at various time points after transfection or infection, fixed with 4% paraformaldehyde for 30 min, and permeabilized with 0.1% Triton X-100 for 10 min, followed by three washes with PBS. The coverslips were blocked with 5% BSA in PBS for 30 min at 37°C and then incubated with the indicated Abs at ∼1:50–1:200 dilution at 4°C overnight. After four washes with PBS-Tween 20, the cells were incubated with a 1:200 dilution of an indicated secondary Ab (conjugated with a fluorescein) at 37°C for 1 h. After the incubation, the cells were further washed four times with PBS-Tween 20 and stained for 10 min with DAPI (1 μg/ml). The coverslips were subsequently analyzed under an Olympus confocal fluorescence microscope. In some experiments, the cells were transfected with plasmids expressing NSs 12 h prior to the infection with AIV H9N2 at an MOI of 1. The cells were then treated as described above for fluorescence Ab staining and confocal microscopy.

Total RNA was extracted from cells plated in 12-well plates using TRIzol reagent, according to the manufacturer’s protocol. The concentration of total RNA in each sample was measured using a NanoDrop 2000 Spectrophotometer (Thermo Fisher Scientific, MA). Viral RNA was extracted from cell cultures using Ambion MagMax-96 Viral RNA Isolation Kit (Life Technologies) following the manufacturer's instruction. Quantitated RNA samples were aliquoted and stored at −80°C until use.

Two hundred nanograms of total RNA were prepared from infected or noninfected HEK 293T, HepG2, and Thp-1 cells and were used for reverse transcription with the Primescript reagent kit. Real-time RT-PCR was performed with 1 μl of cDNA in a total volume of 10 μl with SYBR Premix Ex Taq II (Takara), following the manufacturer’s instructions. Relative expression values were standardized by an internal GAPDH control. The fold change of indicated genes was calculated following the formula: 2^{(△Ct of gene−△Ct of GAPDH)}. The assays were performed in duplicates and repeated three times for each reaction, and the mean values and SDs were calculated.

A one-step RT-PCR using a minor-groove-binding probe was applied for the measurement of virus RNA copy numbers as previously described (24). Briefly, a 737-bp fragment of the SFTSV S segment (JS-2010-014 strain) was transcribed, amplified, and purified, which was treated as the SFTSV standard RNA. A standard curve of the viral RNA copy numbers was generated using a serially diluted purified viral RNA, with concentrations ranging from 1.0 × 10⁷ to 1.0 × 10¹ copies/μl obtained by 10-fold serial dilutions. The RT-PCR was performed using the QuantiTect Probe One Step RT-PCR Kit (Qiagen) with the conditions as recommended by the manufacturer: 12.5 μl of 2 × Master Mix, 0.25 μl of RT-PCR enzyme mix, 0.5 μl of each primer (10 mM), 0.5 μl of the probe (10 mM), 5 μl of extracted RNA or standard RNA, and 5.75 μl of diethyl pyrocarbonate–treated water. Amplification and detection were performed with an Applied Biosystems 7500 Real-Time PCR System (Applied Biosystems, Foster City, CA) under the following conditions: 50°C for 20 min, followed by PCR activation at 95°C for 10 min and 45 cycles of amplification (15 s at 94°C and 45 s at 60°C). The data were analyzed using an SDS software provided by Applied Biosystems. Primer and probe sequences were available but not shown in the report.

HEK 293T cells were grown in 24-well tissue culture plates in DMEM with 10% FBS until 60–70% confluence was reached. The cells were transfected with a total of 505 ng of plasmid DNA consisting of 100 ng of the reporter plasmid pIFN-Luc, 5 ng of pRL-SV8 for normalization, and 400 ng of blank control pRK5-Flag, pRK5-TBK1, or pRK5-NSs. IFN-Luc activation was assessed using a Dual-luciferase reporter system as follows. Twenty-four hours posttransfection, the cells were further stimulated with 50 μg/ml poly(I:C) (Sigma-Aldrich) for another 8 h before the cells were washed in PBS and lysed in 0.1 ml of a reporter lysis buffer (Promega, Madison, MI). Firefly and Renilla luciferase activities were measured by luciferase assay (Promega). Results were expressed as fold changes over the nontransfected controls.

SFTSV stock or the culture cell medium from the cells infected with SFTSV was diluted in 10-fold serial dilutions, and each dilution (1 ml) was inoculated into a confluent monolayer of Vero cells in six-well plates. After 2 h of incubation at 37°C, the suspension was removed, and 3 ml of DMEM containing 1% methylcellulose and 1% FBS was added to each well. After 7 d, the overlay suspension was removed and washed with PBS twice. The cells in each well were then fixed with 4% paraformaldehyde for 20 min and later washed with PBS twice. After that, the cell layer was blocked with 5% milk for 1 h at room temperature, followed by incubation with a rabbit anti-SFTSV serum collected from a rabbit at a 1:100 dilution at 4°C overnight. After several washes with PBS, the cells were incubated with an alkaline phosphatase–conjugated goat anti-rabbit secondary Ab at a 1:2000 dilution at room temperature for 2 h, followed by PBS washes for three times. Virus foci were developed by BCIP/NBT reagents at room temperature (MilliporeSigma) and observed under a light microscope.

For statistical analysis, a two-tailed Student t test was used to evaluate the data by SPSS software (IBM SPSS, Armonk, NY). An analysis was used to determine significant differences of the data in two or more groups, and a p value < 0.05 was considered statistically significant.

Transcription factors IRF3 and IRF7 are highly homologous in their sequences. Both are considered important in the induction of antiviral immunity (25), but little has been known about the role of IRF7 during the SFTSV infection. To explore the potential function of IRF7, we first examined the expression of IRF7, IRF3, and other proteins involved in innate immunity in infected cells. We infected HEK 293T cells with SFTSV at an MOI of 1. The cell lysates were further analyzed with SDS-PAGE and Western blot analysis. As shown in Fig. 1A, whereas IRF3 was constitutively expressed and could even be detected in noninfected cells, the expression of IRF7 was essentially induced in response to the SFTSV infection. IRF7 appeared with three or four isoforms produced by alternative splicing and reached the peak for isoform A in particular around 24 h postinfection (p.i.) (Fig. 1A). Quantitative analyses of the IRF3/p-IRF3 and IRF7/p-IRF7 induction were presented in Supplemental Fig. 1A–D. It has been reported that IRF7 has isoforms A (503 aa, 54 kDa), B (474 aa, 51 kDa), C (164 aa, 18 kDa), and D (516 aa, 56 kDa), with the isoform A sequence to be chosen as the “canonical” one (26), which is, therefore, the focus of our study. Sequence alignment of these isoforms is shown in Supplemental Fig. 1.

Both IRF3 and IRF7 were phosphorylated and activated along with their increased protein expression levels during the course of infection (Fig. 1A). Induction of other proteins involved in innate immunity was also apparent, as shown in Fig. 1B, in which RIG-I and MAVS were induced and increased significantly during the infection. TRAF3, a component required for activating kinases essential for IRF3 and IRF7 activation (27), increased in response to the SFTSV infection. TBK1, the serine/threonine kinase downstream of TRAF3, was transiently induced and responsible for the phosphorylation of IRF3 (28). IKKε is considered to phosphorylate and activate IRF7 as well as IRF3 (28, 29).

We performed real-time RT-PCR to quantify mRNA copy numbers of IRF3 and IRF7 induced in four cell types during the SFTSV induction. The number of the IRF3 transcripts remained largely unchanged in HEK 293T, HeLa, and Thp-1 cells, whereas a weak and transient increase up to 10-fold was observed at the later stage of infection in HepG2 cells (Fig. 1C). However, the fold change of the IRF7 transcript numbers increased steadily and reached the greatest increase in most cell types at 36 or 48 h p.i. The increase of the IRF7 transcripts in HepG2 cells was most significant, with 11.4-, 26.9-, 61.3-, and 366.8-fold increases at 12, 24, 36, and 48 h p.i., respectively (Fig. 1D). All four cells types were sensitive to SFTSV infection as we also quantified the viral S gene copy numbers in the cells infected with SFTSV at various time points p.i. (Fig. 1E). We conclude that the induction of the IRF3 and IRF7 transcripts varied greatly among the tested cell types, with the HepG2 cells to be the most potent for induction. HepG2, originated from human liver epithelial cells, could be unique in response to SFTSV infection as we know that the liver is one of the severely affected organs during SFTSV infection (30).

Previous studies have shown that IRF3 and IRF7 are key regulators in the transcriptional induction of the type I IFN gene triggered by virus infections (31–33), and we have demonstrated the function of IRF3 in inducing innate immunity in SFTSV-infected cells (16). To characterize the effect of IRF7 signaling on the SFTSV infection, we used the IRF7-specific siRNA to knockdown its expression in HEK 293T cells. Expression of IRF7, especially the isoform A, was effectively suppressed in siRNA-treated cells stimulated by IFN-β (Fig. 2A). The cells, treated or not treated with IRF7-siRNA, were infected with SFTSV at an MOI of 1. After the infection, the induction of IRF7 transcripts was significantly inhibited in IRF7-knockdown cells, and the IRF7 isoform A failed to be induced, in contrast to that in the scramble siRNA–treated cells (Fig. 2B). Quantitative analyses of the inhibition on the IRF7 transcript induction in specific or control siRNA knockdown cells were presented in Supplemental Fig. 1E.

We next extracted either total RNA or viral RNA from the cells and the medium, respectively, at different time points p.i. Using real-time RT-PCR, we were able to quantify the copy numbers of the viral S gene with specific primers in infected cells, treated with either IRF7 or scramble siRNA. Our data showed that in IRF7-knockdown cultures, the copy numbers of the virus were significantly higher than those in scramble siRNA–treated cells 24 h p.i. (Fig. 2C, 2D). We also performed viral focus forming assay to titrate infectious viral titers in the culture medium. Our data showed that the number of the viral foci was significantly higher at 36 h p.i. in IRF7-siRNA knockdown cultures than in the scramble siRNA–treated cells (Fig. 2E), suggesting that IRF7 plays an important role in inhibiting SFTSV replication.

We noted that the knockdown of IRF7 appeared to be complete, and the effect on viral replication, although significant, was not dramatic. We consider that this was probably due to the additional role of another IRFs, such as IRF3, which was still functional in the induction of IFNs.

To confirm the role of IRF7 in IFN-α/β induction during SFTSV infection, we knocked down IRF7 in HEK 293T cells with IRF7-specific siRNA. We noted that the IRF7 knockdown occurred, and the induction of IRF7 was severely compromised after SFTSV infection (Fig. 3A). Real-time RT-PCR and ELISA results showed that the knockdown of IRF7 led to a decrease of IFN-α and IFN-β induction in SFTSV-infected HEK 293T cells (Fig. 3B, 3C). This result indicates that IRF7 was responsible for upregulation of both IFN-α and -β during SFTSV infection.

To characterize the mechanism as to how IRF7 was counteracted in cells infected with SFTSV, we subcloned the cDNA of IRF3, IRF7, and viral NSs into expression vector pRK5 for expression in HEK 293T cells together with the reporter plasmids for IFN-β or IFN-α2/α4 promoter activities. We found that, first of all, overexpression of IRF3 had no effect on the activities or induction of IFN-α2, IFN-α4, and IFN-β (Fig. 4A). However, overexpression of IRF7 predominantly activated the promoter activities of IFN-α2 and -α4 while having only mild effect on activating IFN-β (Fig. 4B).

Our studies also indicate how NSs of SFTSV impacted the induction of IFNs by IRF7. As shown in Fig. 4C–E, the IRF7-stimulated promoter activities of IFN-β, IFN-α2, and IFN-α4 were suppressed by the expression of NSs in a dual-luciferase reporter assay, and the suppression was the most significant on IFN-α2 and IFN-α4 in a dose-dependent manner (Fig. 4D, 4E). We concluded that NSs appeared to compete or block the role of IRF7 for its activation of the IFN-α promoter activities or induction of IFN-α2/α4.

We further examined how NSs interacts with IRF7 in cells. First, we transfected HEK 293T cells with plasmids expressing Flag-tagged IRF7 and HA-tagged NSs. The cell lysates were prepared for coimmunoprecipitation with anti-Flag or anti-HA Abs, respectively. Followed by a Western blot analysis, we were able to show that NSs and IRF7 were associated in the transfected cells (Fig. 5A). Second, we harvested the cell lysates from SFTSV-infected HEK 293T cells, which were immunoprecipitated with either anti-IRF7 or anti-NSs Abs, and the immunoprecipitates were subjected to a Western blot analysis with anti-NSs or anti-IRF7 Abs. Viral NP and NSs proteins were detected in the input controls as the lysates were from the infected cells. Our data showed that IRF7 and NSs interacted with each other in the infected cells, as well (Fig. 5B).

IRF7 is rarely expressed in normal cells. In IFN-stimulated cells, IRF7 was found to be induced. In SFTSV-infected cells, we were able to demonstrate that a major portion of IRF7 was distributed in the insoluble fraction of the cell lysates, which increased at later times p.i. (Fig. 5C). It is noted that the p-IRF7 was also in the insoluble form, indicating that in SFTSV-infected cells, IRF7 was distributed in insoluble structures likely with or without being phosphorylated or activated. It appears that the phosphorylation of IRF7 occurred after the induction because still no p-IRF was detectable at 6 h p.i. when IRF7 was present. Because this insoluble fraction was from the cytosol, free of histone proteins, the activated form of IRF7, or p-IRF7, appeared to be retained in the cytoplasm (Fig. 5C), indicating that a large portion of p-IRF7 failed to be translocated into the nucleus during the infection.

We tried to confirm the distribution of IRF7 and p-IRF7 in the infected cells. IRF7 and p-IRF7 in either soluble fraction or insoluble fraction were quantified and presented in Fig. 5D. The results indicate that most of IRF7 and p-IRF7 were distributed in the insoluble fraction of the cytosol or in retention of the cytoplasm, instead of translocation into the nucleus, in SFTSV-infected cells.

We intended to further examine the distribution of both IRF7 and p-IRF7, together with other immune components important in the activation of IFN induction, in SFTSV-infected cells. Cell lysates were prepared from the cells infected with the virus at different time points p.i., and both soluble and insoluble fractions were prepared for Western blot analyses with the respective Abs. We also infected the cells with AIV H9N2 and prepared the cell lysates as a control. In the soluble fractions from the SFTSV-infected cells (Fig. 6A), MAVS and IRF3 were weakly detected without a sign to be induced p.i.; IRF7 was only briefly detected at 6 h p.i. before it vanished afterward. Almost no signal of p-IRF7 was detected during the SFTSV infection. On the contrary, a strong induction of MAVS, IRF7, and p-IRF7 was detected, along with IRF3 and p-IRF3, in H9N2 AIV–infected cells (Fig. 6C).

When we examined the insoluble fractions of the cells, increased amounts of MAVS, IRF7, p-IRF7, IRF3, and p-IRF3 were detected in SFTSV-infected cells (Fig. 6B), whereas none of the above was detectable in the insoluble fraction in H9N2-infected cells (Fig. 6D). p-TBK1 was shown to increase in both the soluble and insoluble fractions in SFTSV-infected cells (Fig. 6A, 6B). Quantitative analyses of the IRF7/p-IRF7 and IRF3/p-IRF3 distributions in the soluble and insoluble fractions at different time points p.i. were presented in Supplemental Fig. 2. Taken together with the results shown in Fig. 5, we conclude that IRF7 was significantly induced (also phosphorylated and activated), especially in the late stage of infection with SFTSV; both IRF7 and p-IRF7 were translocated into the insoluble fraction in the cytosol of the infected cells, along with other key molecules, including IRF3, p-IRF3, TBK1, p-TBK1, and probably MAVS, as well. We hypothesized that the NSs-formed inclusion bodies were the sanctuaries for IRF7 and p-IRF7 in the cytosol, similar to the process for other molecules demonstrated in the previous studies (16–18) as a result of IRF7 interaction with NSs, as shown in Fig. 5.

We previously showed that IRF3 did not directly interact with NSs in SFTSV-infected cells. In the presence of TBK1, however, NSs and IRF3 were associated, indicating that NSs and IRF3 interacted indirectly (16). Because TBK1 is constitutively expressed, we cannot exclude the possibility that the interaction of IRF7 and NSs is also through TBK1. To clarify this point, we performed an experiment with cell lysates prepared from HEK 293T cells transfected with plasmids expressing HA-NSs, Flag-IRF7, or Flag-TBK1. As shown in Fig. 7A, our coimmunoprecipitation results demonstrated that, although Flag-TBK1 interacted with HA-NSs, Flag-IRF7 interacted with HA-NSs as well in the absence of Flag-TBK1, indicating that NSs interacts with IRF7 independent of TBK1 (Fig. 7A). In fact, the interaction of Flag-IRF7 and Flag-NSs did not increase in the presence of TBK1, also suggesting that TBK1 does not affect the IRF7-NSs interaction (Fig. 7A).

In addition, we failed to detect the interaction of TBK1 and IRF7 in cells transfected with the plasmids expressing Flag-TBK1 and Flag-IRF7 in another coimmunoprecipitation assay, indicating that IRF7, unlike IRF3, does not interact with TBK1 but rather interacts directly with NSs in the cells (Fig. 7B). Failure of IRF7 in interaction with either TBK1 or IKKε was also detected in the cells infected with SFTSV (Fig. 7C, 7D) so that a direct and steady interaction between the kinases and IRF7 might not occur, but a transient interaction could still happen.

We did observe that although TBK1 did not interact with IRF7, they colocalized in the structures where NSs was present, as shown in a confocal imaging experiment in HeLa cells cotransfected with Flag-IRF7, Flag-TBK1, and enhanced GFP (EGFP)-NSs (Fig. 7E). Likely, TBK1 was colocalized with IRF7 through its interaction with NSs. Because IRF7 is supposed to be mainly phosphorylated by IKKε, IKKε could also be brought into the NSs-formed inclusion bodies, which was in fact proved in the previous study (16).

We hypothesized that a large portion of IRF7 was brought into the insoluble fraction of the cytosol, suggesting that IRF7 may be sequestered to the NSs-formed viral inclusion bodies, as well. To confirm whether this happens in the cells, we carried out more confocal imaging with HeLa cells transfected with plasmids expressing IRF7 and/or NSs. As shown in Fig. 8, IRF7 was dispersedly distributed in the cytoplasm (Fig. 8A), whereas NSs emerged to form characteristic viral inclusion bodies, which increased in size with time after the initial wide distribution in the cytoplasm (Fig. 8B) in the cells expressing only Flag-IRF7 (Fig. 8A) or HA-NSs (Fig. 8B), respectively.

When the cells were cotransfected with plasmids expressing both proteins, however, IRF7 did not spread dispersedly but rather distributed exclusively in the unique structures colocalized with NSs from the first examined time point at 12 h posttransfection (Fig. 8C). The structures varied with increased size with time posttransfection in the cytoplasm, clearly indicating that IRF7 was translocated into the NSs-formed inclusion bodies, which could turn to be the characteristic ring-like structures (Fig. 8D) in transfected cells.

We further confirmed that the colocalization of NSs and IRF7 occurred in SFTSV-infected cells, as well. HeLa cells were infected with the virus, and the cells were stained with anti-NSs or/and anti-IRF7 Abs at various time points p.i. for confocal imaging. IRF7 was barely detectable in noninfected cells (Fig. 9A, top). In infected cells, however, IRF7 appeared to increase in quantity and be colocalized instantly in the unique structures with NSs in the cytoplasm (Fig. 9A). During the infection, the structures, colocalized by IRF7 and NSs, increased in number and size, and the colocalization of IRF7 and NSs in the structures remained consistent in the cytoplasm, indicating that IRF7 was induced, and induced IRF7 was intensively sequestered into the viral inclusion bodies during the infection (Fig. 9A).

We also examined the localization of the activated IRF7 after viral infection. p-IRF7 appeared to be detectable at 12 h p.i. and increased in quantity through 36 h p.i. and was sequestered and translocated, as well, into the NSs-formed viral inclusion bodies after the initial dispersed distribution (Fig. 9B). We prestimulated the HeLa cells with IFN-β, and the induced IRF7 was found to be colocalized with NSs in SFTSV-infected cells, as well (Supplemental Fig. 3). We also have data to show that both IRF7 and p-IRF7 were colocalized with NSs in HepG2 cells infected with SFTSV (Supplemental Fig. 4), with identical results shown in HeLa cells. Apparently, the interaction of IRF7 and NSs did not interfere with IRF7 being phosphorylated and activated by upstream kinases, such as IKKε in SFTSV-infected cells.

We tried to understand whether NSs directly blocked the translocation of IRF7 into the nucleus after being activated during the infection. In HeLa cells infected with AIV H9N2, IRF7 was induced in expression, phosphorylated, and subsequently translocated into the nucleus, whereas IRF7 was absent in noninfected cells (Fig. 10A–C). We transfected the HeLa cells with plasmids expressing HA-NSs and Flag-IRF7 prior to the infection. After the AIV H9N2 infection, IRF7 was still phosphorylated, but its translocation into the nucleus was completely blocked in the presence of NSs. Instead of translocation into the nucleus, p-IRF7 was in retention in the cytoplasm and sequestered into the NSs-formed inclusion bodies in AIV-infected cells (Fig. 10D).

We also prestimulated the HeLa cells with IFN-β, followed by infection with SFTSV, to see how the translocation of the activated IRF7 would be affected by viral NSs. As shown in Fig. 11, p-IRF7 was detected in the nuclear regions of the uninfected cells after IFN-β stimulation. However, in the SFTSV-infected cells, the p-IRF7 was retained in the viral inclusion bodies (colocalized with NSs) outside the nucleus, as shown in the results mentioned above (Fig. 10).

The type I IFN system is essential for antiviral immune response, a primary target of immune evasion strategies developed by viruses. Viruses have evolved and adopted various mechanisms to evade or subvert the innate immunity. The VP35 proteins of Ebola and Marburg viruses bind to viral dsRNA to prevent their recognition by RIG-I and MDA5 (34, 35). Paramyxoviruses encode V proteins to bind to MDA5 and LGP2, resulting in the disruption of downstream signaling responsible for IFN induction (36–38). NS1 of influenza virus targets MAVS as well as RIG-I and MDA5 (39). The NSs NS1 and NS2 of respiratory syncytial virus directly interfere with the interaction between RIG-I and MAVS (40). Viral proteins possessing protease activities, such as 3C of picornavirus EV71 (41) and NS3/4A of flavivirus hepatitis C virus (42), have the potential to target and degrade the adaptor protein TRIF and disrupt the TLR signaling, resulting in compromised IFN induction. Rabies virus P proteins, filoviral VP35 protein, and paramyxoviral V protein inhibit phosphorylation of IRF3 and IRF7 by IKKε /TBK1 kinases, and, consequently, the translocation of IRF3 and IRF7 into the nucleus for IFN transcriptional activation is inhibited (43–46). Middle East respiratory syndrome coronavirus encodes at least five proteins to confront IFN with a variety of mechanisms (47).

Previous studies have shown that SFTSV not only suppressed the IFN induction through sequestration of the IKK complex, including TBK1, IKKε, and IRF3, into the viral inclusion bodies (IBs) formed by NSs (16, 21) but also disrupted the signaling of the IFN response, leading to the induction of ISGs through hijacking STAT1 and STAT2 into IBs (18). In this report, we demonstrated that SFTSV could inhibit the induction of both IFN-α and -β through NSs-IRF7 interaction and sequestration of IRF7 into the viral inclusion bodies.

IRFs are a family of transcription factors composed of nine members, IRF1 through IRF9, in mammalian cells. IRF1, 3, 5, 7, and 8 are positive regulators of type I IFNs in different cell types (48). Although IRF3 and IRF7 are highly homologs, they function distinctly. IRF3 is crucial for the initial induction of IFN-β and IFN-α1 (49), whereas IRF7 has a major role in subsequent feedback amplification of both IFN-α and IFN-β after IRF7 is newly synthesized in replace of degraded IRF3 (50).

It is believed that IFN-α and IFN-β differ in their distribution in various tissues or organs and are regulated differentially, as well, by the upstream transcription factors or regulators. Transcription of IFN-β is considered to be activated by IRF7 as well as IRF3, and IRF7 plays a key role in the induction of IFNs at the later stage of infection because it is barely present in naive cells of many types but only induced upon viral infection. In contrast, IFN-α is thought to be exclusively regulated or induced by IRF7, which makes IRF7 essential for innate antiviral defense in organs or tissues in which only IFN-α is present or IFN-α plays a dominant role in host defense (9, 11). Studies have found that IRF7 has a preferential ability to activate the IFN-α promoters (51, 52).

We did observe that IRF7 was regulated or induced differentially in the cell types tested in this study. Although IRF7 responded most vigorously in HepG2, the liver epithelial cells, it showed almost no response in Thp-1, the monocytic cells of hemopoietic origin. Our finding in this study demonstrated that IRF7 was significantly upregulated in HepG2 (Figs. 1, 2). IRF7 could be targeted by NSs of SFTSV and compromised through sequestration into the viral inclusion bodies, which may have a significant consequence, apparently, for particular organs or tissues in which IFN-α may play an exclusive role in innate immunity. In certain cell types, such as pDC, IFN-α could be induced by the MyD88-dependent pathway, in which MyD88 interacts directly with IRF7 as well as TRAF6 to form a cytoplasmic complex for activating and stimulating IFN-α promoters (53). Direct interaction of viral proteins, such as the NSs of SFTSV, and IRF7 would apparently hinder the formation of the complex, leading to failure of IFN-α induction.

The timing effect on suppressing further induction of IFN-β by NSs in SFTSV infection is important. Constitutively expressed IRF3 would activate the induction of IFN-β upon viral infection in cells. Because IRF7 is predominately induced by IFN-β, it fits into a positive feedback (5, 10), and more IFN-β will produced and sustained for a full scope of host defense. This positive loop could be interrupted when IRF7 was compromised, including sequestration by viral NSs. Moreover, IRF7 is responsible for the induction of IFN-α, which would be uniquely induced in tissues or organs for antiviral activities. This response could be vital to the host in the scenario in which mice are infected with WNV (9) and would be apparently compromised when IRF7 is not functionally present in the cytosol in which IFN-α plays a dominate role in antiviral immunity. In this case, interaction of NSs with IRF7 would impede inevitable viral clearance. In fact, viral infection could be fatal in the patient who was inheritably deficient in IRF7 (11), suggesting that IRF7 is not simply an ancillary factor to IRF3 in the induction of IFNs. Indeed, IRF7 may function as the “master” regulator of type I IFN production on certain occasions.

Regulation of IRF3 and IRF7 by SFTSV appears to differ. Our previous study showed that IRF3 was sequestered into the inclusion bodies through the interaction of TBK1 and IRF3, and viral NSs interacted with TBK1. In another study, IRF3 interacted with NSs indirectly and was brought into IBs because of the interaction of NSs and TBK1. In this study, our data indicate clearly that NSs directly interacted with IRF7, with or without TBK1. IRF7 is also phosphorylated and activated by kinase IKKε, which is sequestered into IBs, as well, as shown in our study (16). The difference between IRF3 and IRF7 in their interactions with NSs may be owed to different C termini of IRFs. The IRF family members share significant homology within the conserved N termini, which have the DNA-binding domain (54). However, the C termini of IRFs are different, which may confer on each member distinct functions (48, 55) or unique interactions with cellular components. So far, we have been unable to identify the exact motif or domain in IRF7 that interacted with NSs. If we could characterize the responsive amino acids, we would be able to generate mutant IRF7 deficient of interaction with NSs for better demonstration of functions in this virus–cell interaction.

Although we showed that IRF3 was sequestered into the viral IBs, resulting in significant inhibition of antiviral IFNs (16), the sequestration of IRF7 as shown in this study was meaningful for two reasons. First, IRF3 was expressed constitutively in the cells and was phosphorylated and activated at the early stage of infection and fully functional in transactivating the transcription of the downstream genes, including IRF7, before a sufficient amount of viral NSs was expressed. Second, sequestration of IRF3 or p-IRF3 into the viral inclusion bodies would likely not be complete, and evasion of IRF3 could lead to induction of IRF7. Therefore, IRF7 was present during the course of SFTSV infection, especially at the later stages, and the sequestration of IRF7 into the insoluble fraction as demonstrated in this study is significant for its suppression of the host’s antiviral immunity.

The role of IRF7 in anti-SFTSV replication could also be shown in in vivo animal experiments. SFSTV could cause pathogenicity and fatal infection in the ifnar1^−/− mouse model in which IFN-α1 receptor is deficient (56). IFN signal-dependent IRF7 induction, which fails in the ifnar1^−/− mice infected with SFTSV, may have played an essential role in sustaining IFN gene transcription activation, including robust production of IFN-α. Interaction of viral proteins with IRF7 may even have an impact on adaptive immunity in infected hosts. In irf7^−/− mice infected with HSV-1, encephalomyocarditis virus, or vesicular stomatitis virus, the induction of IFN-α and -β transcripts was impaired, and IFN-α production was abolished in pDC, whereas both IFN-α and -β induction was normal in irf3^−/− mice, indicating that IRF7 is essential, but IRF3 is dispensable for the induction of IFNs in a particular APC and functional stimulation of CD8 T cells (8). No study has been carried out yet to evaluate how the interaction of IRF7 with SFTSV NSs and the sequestration of IRF7 into the viral inclusion bodies occur in pDC and, if it happens, what the consequences would be to patients’ protective Ab or cytotoxic T cell production p.i.

In sum, viruses exploit numerous strategies in counteracting functions of IRF7 and IRF3, and viral NSs are essential to viral evasion of host immunity. Rotavirus NSs 1 mediates the degradation of IRF3, 5, and 7 (57). The NS1 and NS2 proteins from respiratory syncytial virus induces proteasomal degradation of IRF3 and IRF7 (58). Zaire Ebola virus VP35 promotes sumoylation of both IRF3 and IRF7 and therefore blocks their induction of type I IFN (59). Our findings indicated that sequestration of IRF7 directly or IRF3 indirectly through the interaction of NSs into viral inclusion bodies is a novel mechanism for a bunyavirus to suppress IFN-α and -β induction during viral infection.

We thank Dr. Fuping You at Peking University for kindly providing us with the IFN-α2-Luc and IFN-α4-Luc reporter plasmids.

The authors have no financial conflicts of interest.