Relative Contributions of the cGAS-STING and TLR3 Signaling Pathways to Attenuation of Herpes Simplex Virus 1 Replication

The magnitude of fatal infections caused by all different viruses in human and animal populations justifies a better understanding of the host innate immune response process that attenuates virus replication. In particular, the relative contributions of different signaling pathways which are responsible for the generation of the innate immune response are still largely unknown. In this study, we used STING-noninteracting TRIF mutants to decipher the relative contributions of the TLR3 and cGAS-STING signaling pathways to the attenuation of HSV-1 infection. We show that the relative contributions of the two pathways to the attenuation of viral infection are different in mouse versus human cell lines. Together, our results provide new insights into the relative contributions of two different signaling pathways in the attenuation of viral infection and may lead to the development of new antiviral strategies aimed at blocking viral infection at very early stages.

ABSTRACT

The innate immune response is crucial for defense against viral infections. Cells recognize virus infection through pattern recognition receptors and induce type I interferons as well as proinflammatory cytokines to orchestrate an innate immune response. Herpes simplex virus 1 (HSV-1) triggers both the cyclic GMP-AMP synthase (cGAS)–stimulator of interferon genes (STING) and Toll-like receptor 3 (TLR3) pathways. It is well known that TLR3 uses the adaptor protein Toll/interleukin-1 receptor (IL-1R) domain-containing adaptor-inducing beta interferon (TRIF) for signaling, but we recently reported that STING signaling also requires TRIF. Because STING directly binds to TRIF, we identified the STING-interacting domain of TRIF and generated STING-noninteracting mutants of human and mouse TRIFs. The mutant TRIFs were unable to support STING signaling, although they were fully functional in the TLR3 pathway. These mutants were used to assess the relative contributions of the TLR3 and STING pathways to the attenuation of HSV-1 replication in mouse and human cell lines. For this purpose, the mouse L929 and NB41A3 cell lines and the human HT1080 and HeLa-M cell lines, in which both the TLR3 and the STING pathways are operational, were used. The TRIF gene was disrupted in these lines by CRISPR/Cas9, before reconstituting them with mutant and wild-type TRIF expression vectors. Infection of the reconstituted cells with HSV-1 revealed that both the cGAS-STING and the TLR3 signaling pathways are required for the attenuation of virus replication, but their relative contributions in attenuating HSV-1 replication were found to be different in mouse versus human cell lines. Thus, our study suggests that the relative contributions of the cGAS-STING and the TLR3 pathways in the attenuation of viral infection may be species specific.

IMPORTANCE The magnitude of fatal infections caused by all different viruses in human and animal populations justifies a better understanding of the host innate immune response process that attenuates virus replication. In particular, the relative contributions of different signaling pathways which are responsible for the generation of the innate immune response are still largely unknown. In this study, we used STING-noninteracting TRIF mutants to decipher the relative contributions of the TLR3 and cGAS-STING signaling pathways to the attenuation of HSV-1 infection. We show that the relative contributions of the two pathways to the attenuation of viral infection are different in mouse versus human cell lines. Together, our results provide new insights into the relative contributions of two different signaling pathways in the attenuation of viral infection and may lead to the development of new antiviral strategies aimed at blocking viral infection at very early stages.

INTRODUCTION

The innate immune response is the first line of defense against a wide range of viruses (1–3). The innate immune response is triggered upon sensing of virus or its components by pattern recognition receptors (PRRs), which detect pathogen-associated molecular patterns (PAMPs) (4, 5). In the case of a viral infection, the main targets of innate immune recognition by the cellular PRRs are the viral nucleic acids: DNA as well as RNA (4, 6–9). Thus, viral infection activates several PRRs, including Toll-like receptors (TLRs), cyclic AMP-GMP synthase (cGAS)–stimulator of interferon genes (STING), and/or RIG-I-like receptors (RLRs) (3, 7, 10, 11). Upon activation, these PRRs initiate robust downstream signaling pathways through their respective adaptor proteins to activate the critical transcription factors: interferon regulatory factors (IRFs) and nuclear factor κB (NF-κB) (1, 3, 9, 12–14). The cooperative transcriptional activity of these transcription factors results in the induction of interferon (IFN) genes, IFN-stimulated genes (ISGs), and other proinflammatory cytokines to mount a rapid innate immune response (1, 9, 15). The innate immune response against virus infection is multifaceted; however, induction of type I IFNs is the major restriction factor for virus replication and its course of infection (3). Type I IFNs synthesized and released from virus-infected cells act on neighboring infected, as well as noninfected, cells to activate the production of antiviral proteins which restrict virus replication (3).

Herpes simplex virus 1 (HSV-1) is a DNA virus that belongs to the family Herpesviridae (16, 17). HSV-1 causes a range of diseases, including herpes keratitis, herpes labialis, and herpes encephalitis, which can be fatal (18). HSV-1 establishes latency in the peripheral nervous system, hence causing lifelong infections, and is highly prevalent in both North America and worldwide (16). Like other viral infections, HSV-1 triggers more than one signaling pathway to induce type I IFNs and other cytokines (4). HSV-1 infection is mainly sensed and reciprocated by the TLR3 and cGAS-STING pathways for the anti-HSV-1 response (19–22). TLR3 is a double-stranded RNA sensor, and cGAS-STING is a DNA sensor (23–25). All PRRs use specific adaptor proteins to initiate downstream signaling (24, 26). An adaptor protein is a protein that is the accessory to the main receptor or sensor proteins in a signal transduction pathway (27, 28). It contains protein-binding motifs, which facilitate interactions and activation of protein-binding partners, including protein kinases, ubiquitin ligases, and transcription factors (28). Activated transcription factors then translocate into the nucleus and induce the transcription of genes encoding inflammatory cytokines, chemokines, and type I IFNs (13, 14).

Toll/interleukin-1 receptor (IL-1R) domain-containing adaptor-inducing beta interferon (TRIF) is an adaptor protein (28). It is the sole adaptor for the TLR3 signaling pathway and is also required for the endosomal branch of the TLR4 signaling pathway (28–30). Our lab has recently reported that TRIF is also required for STING signaling (31). TRIF interacts directly with STING through its C terminus and promotes STING-mediated signaling (31). TRIF-TLR3 and TRIF-STING interactions show that the TRL3 pathway and the STING pathway are connected with each other through TRIF. Both the TLR3 and the STING signaling pathways are required to inhibit HSV-1 replication (31). To study the relative contributions of the TLR3 and the STING pathways in the attenuation of HSV-1 infection, we mapped the STING-binding domain in TRIF, identified the critical residues in that domain which are required for STING binding, and generated STING-noninteracting mutants. Using these mutants, we demonstrated that the relative contributions of the cGAS-STING and the TLR3 pathways to the attenuation of HSV-1 infection are different in mouse versus human cell lines.

RESULTS

Identification of TRIF residues needed for STING interaction.

We previously showed that TRIF interacts directly with STING. This TRIF-STING interaction is mediated through their C termini. For the interaction with STING, both mouse and human TRIFs require amino acids 542 to 732 (TRIF 542–732) and 542 to 712 (TRIF 542–712), respectively (31). We decided to find and characterize the STING-binding domain in the C terminus of TRIF by a coimmunoprecipitation technique. For this, we used several N-terminal as well as C-terminal truncated forms of mouse TRIF and looked for their interaction with STING (Table 1). It is worth noting that, when quantitated, the interactions with STING of many truncated TRIF mutants were stronger than that of wild-type (wt) TRIF. We found that all N-terminal truncated forms except the TRIF 628–732 truncated form interacted strongly with STING (Fig. 1A). Similarly, all of the C-terminal truncated forms except TRIF 1–585 interacted with STING (Fig. 1A). However, TRIF 1–595 strongly interacted with STING, suggesting that the STING-binding domain in TRIF lies between amino acids 586 and 595 of TRIF. Using the mutant in which these 10 amino acids were deleted from TRIF (TRIF Δ586–595), we found that the TRIF-STING interaction is lost with the deletion of amino acids 586 to 595 of TRIF (Fig. 1B). We analyzed further to find the amino acid among these 10 amino acids that is most crucial for the TRIF-STING interaction. Substitution mutations were introduced for these required amino acids, and the mutants were checked for the interaction with STING. We found that the mutant in which leucine 589 was replaced by alanine (TRIF-L/A) and the mutant in which leucine 589, glycine 590, and valine 591 were each replaced by alanine (TRIF-LGV/AAA) did not interact with STING (Fig. 1C). We could distinguish between the ectopic tagged TRIF and STING proteins from their endogenous counterparts by their molecular weight differences, and we found that the tagged proteins were expressed at higher levels than the endogenous STING and TRIF proteins (Fig. 1D). Besides that, we observed that the ectopic tagged TRIF and endogenous TRIF appear as a doublet on the blots, and this finding is consistent with the findings of previous studies (32, 33). Schematic presentations of mouse TRIF with the newly found STING-binding domain and its mutants, the TRIF deletion mutant (TRIFΔ) and TRIF-L/A, are shown in Fig. 1E. To determine if this domain is conserved between the mouse and human TRIFs, we aligned both TRIFs and found that 6 out of these 10 amino acids are highly conserved in both proteins. In human TRIF, the location of these 10 amino acids was amino acids 584 to 593; we introduced the same mutations into human TRIF to find if these 10 amino acids (amino acids 584 to 593) and the leucine 587 in human TRIF are required for the human TRIF-STING interaction. We found that the human TRIF mutant with a deletion of amino acids 584 to 593 (TRIF Δ584–593) and the human TRIF mutant in which leucine 587 was replaced by alanine (TRIF-L/A) did not bind with human STING (Fig. 2A). The quantitation of the interaction is presented in Fig. 2B. For human STING and TRIF mutants, too, we measured the ectopic and endogenous protein expression levels in 293T cells. Because 293T cells lack STING expression, we could detect only the tagged STING (34); we found that the tagged TRIF protein was expressed at a higher level than the endogenous protein (Fig. 2C). A schematic presentation of human TRIF with the newly found STING-binding domain and its mutants, TRIFΔ and TRIF-L/A, is shown in Fig. 2D.

TABLE 1.

Interaction of different truncated forms of mouse TRIF with mouse STING^a

Truncated form of mouse TRIF	Interaction with mouse STING	Quantitation of interaction (%)
TRIF FL	+	100
TRIF 478–732	+	361
TRIF 488–732	+	136
TRIF 498–732	+	23
TRIF 508–732	+	78
TRIF 518–732	+	240
TRIF 528–732	+	45
TRIF 578–732	+	282
TRIF 628–732	−	0
TRIF 1–585	−	0
TRIF 1–595	+	246
TRIF 1–605	+	502
TRIF 1–615	+	249
TRIF 1–625	+	235
TRIF 1–650	+	255
TRIF 1–700	+	253

^aMouse TRIF and its various C-terminal and N-terminal truncated forms, their interactions with mouse STING, and quantitation of the interactions are presented. The numbers after TRIF indicate the amino acid residues of full-length (FL) TRIF (amino acids 1 to 732) that were retained in the mutant. The interaction value, measured by determination of the OD value using ImageJ software, for the bait (full-length TRIF) in the immunoprecipitated samples for the full-length TRIF-STING interaction was taken to be 100 to normalize the OD values for the TRIF mutants. For the target (STING), the OD value in the immunoprecipitated sample for the full-length TRIF-STING interaction was taken to be 100 to normalize the OD values for the target in TRIF mutant-STING samples. Target/bait values (in percent) were then calculated from the normalized values for the target and bait. An interaction that quantitated at less than 10% was considered no interaction and is indicated with a minus sign.

FIG 1.

Mapping of the STING-binding domain of mouse TRIF. (A) (Left) Mouse HA-STING was coexpressed with full-length (FL) or various truncated forms of mouse TRIF-3xF in L929 cells. The empty vector (EV) was set as a negative control. From the lysates, 3xF-tagged TRIF was immunoprecipitated (IP) by an M2 anti-Flag affinity gel, and the levels of STING and TRIF bound to the beads were determined by Western blotting (immunoblotting [IB]). (Right) Schematic diagram of mouse TRIF and its mutants and their interaction with mouse STING. (B and C) Results of an interaction experiment similar to that described in the legend to panel A with deletion and substitution mutants of mouse TRIF, TRIF Δ586–595, TRIF-LGV/AAA, and TRIF-L/A. The nontransfected control (NTC) and the empty vector were set as the negative controls. (D) For comparison of ectopic and endogenous protein expression, mouse HA-STING (mSTING-HA) was coexpressed with full-length TRIF-3xF or mutants of mouse TRIF-3xF (mTRIF) in L929 cells. An empty vector-transfected control was set as a negative control. From the lysates, protein levels for endogenous as well as tagged STING and TRIF proteins were determined by Western blotting with anti-STING and anti-TRIF antibodies. Actin was used as a loading control. (E) Schematic presentation of mouse TRIF and its different interaction domains. The STING-binding domain is found to be between amino acids 586 and 595. The numbers to the left of the gels in panels A and D are molecular masses (in kilodaltons).

FIG 2.

Mapping of the STING-binding domain of human TRIF. (A) Human HA-STING was coexpressed with full length (FL) or with deletion and substitution mutants of human TRIF-V5 in 293T cells. A nontransfected control (NTC) and empty vector (EV) was set as negative controls. From the lysates, V5-tagged TRIF was immunoprecipitated (IP) by use of an anti-V5 agarose affinity gel, and the levels of STING and TRIF bound to the beads were determined by Western blotting (immunoblotting [IB]). (B) Quantitation of the interaction of human HA-STING with full-length or mutant human TRIF-V5. (C) For comparison of ectopic and endogenous protein expression, human HA-STING was coexpressed with full-length or mutant mouse TRIF-V5 in 293T cells. An empty vector-transfected control was set as the negative control. From the lysates, protein levels for endogenous as well as tagged STING and TRIF proteins were determined by Western blotting with anti-STING and anti-TRIF antibodies. Actin was used as a loading control. (D) Schematic presentation of human TRIF and its different interaction domains. The STING-binding domain was found to be between amino acids 584 and 593. The numbers to the left of the gels in panels A and C are molecular masses (in kilodaltons).

Reconstitution of TRIF^−/− cells with STING-noninteracting TRIF mutants.

To determine the functional properties of the newly generated TRIF mutants, we initially chose the mouse L929 and the human HT1080 cell lines because both of these cell lines express STING and TLR3 and respond to the respective ligands. TRIF-nonexpressing mutants of the two cell lines were generated by the CRISPR/Cas9 technique without affecting the expression of STING. DNA from the TRIF^−/− cells was sequenced for confirmation of the deletion of the target nucleotide residues in the TRIF gene, and the abolished expression of TRIF in these cells was verified by immunoblotting. L929 TRIF^−/− cells were then transfected with plasmids expressing wt mouse TRIF or its STING-noninteracting mutants, TRIFΔ and TRIF-L/A. Antibiotic-resistant cell clones were isolated and screened for the stable expression of TRIF. The clones, which showed expression of TRIF or its mutants close to the physiological level of TRIF expression in wt L929 cells, were selected for further experiments. Stable expression of TRIF and its mutants in the reconstituted cells, L929-TRIF, L929-TRIFΔ, and L929-TRIF-L/A cells, was verified by Western blotting (Fig. 3A). Similarly, HT1080 TRIF^−/− cells were reconstituted with human TRIF or its STING-noninteracting TRIF mutants, TRIFΔ and TRIF-L/A. The stable expression of TRIF and its mutants in the reconstituted cells, HT1080-TRIF, HT1080-TRIFΔ, and HT1080-TRIF-L/A cells, was verified by Western blotting (Fig. 3B). TRIF^−/− cells were also generated by the CRISPR/Cas9 method from the mouse neuronal cell line NB41A3, and NB41A3 TRIF^−/− cells were transfected with mouse TRIF or its STING-noninteracting TRIF mutants, TRIFΔ and TRIF-L/A, and antibiotic selected for five passages. Expression of TRIF and its mutants in the transfected and selected cells, NB41A3-TRIF, NB41A3-TRIFΔ, and NB41A3-TRIF-L/A cells, was verified by Western blotting (Fig. 3C). Similarly, human HeLa-M TRIF^−/− cells, generated by the CRISPR/Cas9 method, were transfected with human TRIF or its STING-noninteracting TRIF mutants, TRIFΔ and TRIF-L/A, and antibiotic selected for five passages. Expression of TRIF and its mutants in the transfected and selected cells, HeLa-M-TRIF, HeLa-M-TRIFΔ, and HeLa-M-TRIF-L/A cells, was verified by Western blotting (Fig. 3D).

FIG 3.

Reconstitution of TRIF^−/− cells with TRIF or its STING-noninteracting mutants. (A) L929 TRIF^−/− cells reconstituted with mouse wt TRIF and its STING-noninteracting mutants for stable expression of TRIF and mutants. Their expression in stable cells was validated by Western blotting. (B) As described in the legend to panel A, HT1080 TRIF^−/− cells reconstituted with human wt TRIF and STING-noninteracting mutants for stable expression of TRIF and mutants. (C) For reconstitution of NB41A3 TRIF^−/− cells with mouse wt TRIF and its STING-noninteracting mutants, cells were transfected and selected for comparable equal expression of TRIF and its mutants. (D) For reconstitution of HeLa-M TRIF^−/− cells with human wt TRIF and its STING-noninteracting mutants, cells were transfected and selected for comparable equal expression of TRIF and its mutants.

Gene induction by the cGAS-STING signaling pathway in cells expressing TRIF mutants.

TRIF is required for gene induction in cGAS-STING signaling. It interacts with STING and promotes its dimerization and activation. In the absence of TRIF, STING signaling is abolished (31). After finding the STING-binding domain in TRIF and generating STING-noninteracting TRIF mutants, we hypothesized that these mutants would not support gene induction by STING signaling. To test our hypothesis, we measured gene induction by measuring the levels of the Ifit1 (p56) and Ifit2 (p54) proteins in the reconstituted cell lines in response to 2′,3′-cGAMP, a direct activator of STING. 2′,3′-cGAMP binds and activates STING through a series of structural changes to initiate STING signaling, which leads to the induction of interferons, interferon-stimulated genes (ISGs; such as IFN-induced proteins with tetratricopeptide repeats [IFITs]), and other proinflammatory cytokines to trigger the innate immune response. Upon 2′,3′-cGAMP stimulation, STING signaling failed to induce Ifit2 in mouse L929 TRIF mutant-reconstituted cells, L929-TRIFΔ and L929-TRIF-L/A cells, similar to the findings for L929 TRIF^−/− cells (Fig. 4A), whereas the same cells reconstituted with wt TRIF, L929-TRIF cells, showed a strong induction of Ifit2 comparable to that in L929 cells (Fig. 4A). Similar to the findings for reconstituted L929 cells, STING signaling failed to induce Ifit1 in HT1080 TRIF mutant-reconstituted cells, HT1080-TRIFΔ and HT1080-TRIF-L/A cells, similar to the findings for HT1080 TRIF^−/− cells, whereas the same cells reconstituted with wt TRIF, HT1080-TRIF cells, showed a strong induction of Ifit1, as did the HT1080 cells (Fig. 4B).

FIG 4.

STING-noninteracting TRIF mutants do not support the cGAS-STING signaling pathway in reconstituted cells. (A) TRIF^−/− or wt TRIF- and mutant-reconstituted (for stable expression of wt TRIF and its mutants in TRIF^−/− cells) L929 cells were transfected with the STING ligand 2′,3′-cGAMP for 6 h, and mouse Ifit2 induction was determined by Western blotting. (B) TRIF^−/− or wt TRIF- and mutant-reconstituted HT1080 cells were transfected with the STING ligand 2′,3′-cGAMP for 6 h, and human Ifit1 induction was determined by Western blotting.

Activation of the components of the STING signaling pathway in cells expressing STING-noninteracting TRIF mutants.

Activated STING recruits TBK1 to phosphorylate IRF3 and activate IκB kinase to phosphorylate IκBα, leading to its degradation and the release of NF-κB. Phosphorylated IRF3 dimerizes and translocates to the nucleus to induce type I IFNs and ISGs, such as IFITs, and other cytokines to trigger the innate immune response. Because we did not observe IFIT induction upon 2′,3′-cGAMP stimulation in cells expressing the STING-noninteracting TRIF mutants, we measured the activation of the components (the transcription factors and the kinases) of the cGAS-STING signaling pathway. Activation of the transcription factors and the protein kinases can be determined by their phosphorylation. Activation of the NF-κB branch of the STING signaling pathway was measured by IκBα phosphorylation and degradation in response to 2′,3′-cGAMP stimulation, and it was severely impaired in L929 TRIF mutant-reconstituted cells, L929-TRIFΔ and L929-TRIF-L/A cells, similar to the findings for L929 TRIF^−/− cells. In contrast, wt TRIF-reconstituted cells, L929-TRIF cells, showed the strong phosphorylation of IκBα (Fig. 5A). In the IRF3 branch of the STING signaling pathway, the phosphorylation of IRF3, TBK1, and Akt in response to 2′,3′-cGAMP stimulation was also impaired in L929 TRIF mutant-expressing cells, whereas wt TRIF-expressing cells showed the strong phosphorylation of these components of the STING signaling pathway (Fig. 5B). Similarly when we measured the phosphorylation of the same transcription factors and kinases in human HT1080 reconstituted cells, we found that the phosphorylation and degradation of IκBα in response to 2′,3′-cGAMP stimulation were impaired in HT1080 TRIF mutant-reconstituted cells, HT1080-TRIFΔ and HT1080-TRIF-L/A cells, whereas wt TRIF-reconstituted cells showed a strong phosphorylation of IκBα (Fig. 5C). In HT1080 reconstituted cells, for the IRF3 branch of the STING signaling pathway, only IRF3 phosphorylation was measured. The phosphorylation of IRF3 in response to 2′,3′-cGAMP stimulation was found to be equally impaired in HT1080 TRIF mutant-reconstituted cells to a level comparable to that in HT1080 TRIF^−/− cells, whereas wt TRIF-reconstituted cells showed a strong phosphorylation of IRF3 (Fig. 5D).

FIG 5.

STING-noninteracting mutants of TRIF do not support the activation of IκBα, IRF3, TBK1, or Akt. (A) TRIF^−/− or wt TRIF- and mutant-reconstituted L929 cells were transfected with the STING ligand 2′,3′-cGAMP for the indicated times. The phosphorylated and unphosphorylated levels of mouse IκBα were determined by Western blotting. (B) Cells were transfected as described in the legend to panel A. The phosphorylated and unphosphorylated levels of other mouse kinases and transcription factors (IRF3, TBK1, and Akt) were determined by Western blotting. Actin was used as a loading control. (C) TRIF^−/− or wt TRIF- and mutant-reconstituted HT1080 cells were transfected with the STING ligand 2′,3′-cGAMP for the indicated times. The phosphorylated and unphosphorylated levels of human IκBα were determined by Western blotting. (D) Cells were transfected as described in the legend to panel C. The phosphorylated and unphosphorylated levels of human IRF3 were determined by Western blotting.

Activation of the TLR3 pathway in cells expressing the TRIF mutants.

TRIF is the sole adaptor protein for TLR3, and we wanted to examine whether the STING-noninteracting TRIF mutants support the TLR3 pathway. For this purpose, we measured TLR3 signaling in the reconstituted cells after treating them with poly(I·C), a direct activator of TLR3. Poly(I·C) binds with TLR3 in the endosome and initiates TLR3 signaling, leading to the induction of type I IFNs and ISGs, including IFITs. An early step of the activation of the TLR3 pathway is the ligand-induced binding of TRIF to TLR3. We found that the mutations introduced in mouse TRIF did not affect the TRIF-TLR3 interaction, as it was evident from the data that the mutants coimmunoprecipitated with TLR3 equally as well as wt TRIF (Fig. 6A). Moreover, upon poly(I·C) treatment, TLR3 signaling induced Ifit1 in mouse L929 TRIF mutant-reconstituted cells (Fig. 6B). Like mouse TRIF mutants, human TRIF mutants interacted as strongly with TLR3 as wt TRIF (Fig. 6C), and upon poly(I·C) treatment, TLR3 signaling induced Ifit1 in human HT1080 TRIF mutant-reconstituted cells (Fig. 6D). From these experiments, we concluded that the STING-noninteracting mutants of TRIF cannot support STING signaling but that they can bind TLR3 and mediate its ability to induce genes.

FIG 6.

TRIF mutants bind with TLR3 and mediate gene induction in the TLR3 pathway. (A) Mouse TLR3-Flag was coexpressed with mouse wt TRIF or its STING-noninteracting mutants in L929 cells. From the lysates, Flag-tagged TLR3 was immunoprecipitated (IP) by use of an M2 anti-Flag affinity gel, and the levels of wt TRIF, the TRIF mutants, and TLR3 bound to the beads were determined by Western blotting. (B) TRIF^−/− or wt TRIF- and mutant-reconstituted L929 cells were treated with the TLR3 ligand poly(I·C) for 6 h, and mouse Ifit1 induction was determined by Western blotting. (C) Human TLR3-Flag was coexpressed with human wt TRIF or its STING-noninteracting mutants in 293T cells. From the lysates, V5-tagged wt TRIF and mutants were immunoprecipitated by use of an anti-V5 agarose affinity gel, and the levels of TLR3, wt TRIF, and TRIF mutants bound to the beads were determined by Western blotting. (D) TRIF^−/− or wt TRIF- and mutant-reconstituted HT1080 cells were treated with the TLR3 ligand poly(I·C) for 6 h, and human Ifit1 induction was determined by Western blotting.

Effects of activating the STING and TLR3 signaling pathways on HSV-1 replication in the reconstituted cell lines.

Activation of the cGAS-STING and TLR3 signaling pathways leads to an antiviral response and the attenuation of HSV-1 replication. To demonstrate that the activated cGAS-STING pathway restricts HSV-1 replication in the wt TRIF-reconstituted cells but not in the TRIF mutant-reconstituted cells, we treated the cells with either poly(I·C) or 2′,3′-cGAMP overnight prior to infection (Fig. 7). Virus replication in wt TRIF-reconstituted cells (4.2 × 10⁷ PFU/ml) was inhibited by pretreatment of the cells with poly(I·C) (1.5 × 10⁶ PFU/ml) or 2′,3′-cGAMP (4.9 × 10⁶ PFU/ml), indicating that both pathways efficiently exerted an antiviral response. In contrast, in TRIF-L/A-reconstituted cells, virus replication (6.5 × 10⁷ PFU/ml) was found to be inhibited only in poly(I·C)-treated cells (8.1 × 10⁶ PFU/ml). Virus replication in the 2′,3′-cGAMP-treated HT1080-TRIF-L/A cells (5.6 × 10⁷ PFU/ml) was similar to that in cells reconstituted with wt TRIF, indicating that only the TLR3 pathway restricted virus replication in HT1080-TRIF-L/A cells.

FIG 7.

HSV-1 replication in the reconstituted cell lines treated with poly(I·C) and 2′,3′-cGAMP prior to infection. (A) Wild-type TRIF- and TRIF-L/A mutant-reconstituted HT1080 cells were treated with poly(I·C) and 2′,3′-cGAMP overnight prior to infection with HSV-1. HSV-1 replication was determined by a TCID₅₀ assay at 16 h postinfection. ns, not significant (P > 0.05); *, P < 0.05; **, P < 0.01.

Differential contribution of the cGAS-STING and the TLR3 signaling pathways in triggering the innate immune response to HSV-1 infection of different cell types.

STING as well as TLR3 signaling is crucial for mounting an innate immune defense against viruses, such as HSV-1. HSV-1 infection activates the STING and the TLR3 signaling pathways, leading to the induction of type I IFNs, which, in turn, attenuate virus replication. Once we established a system in which the mutual connection between STING and TLR3 signaling, through TRIF, was broken, we wanted to study the relative contributions of these two signaling pathways to the attenuation of HSV-1 replication. For this purpose, we measured HSV-1 replication in wt cells and compared that with virus replication in cells expressing the wt TRIF and its mutants, which support TLR3, but not STING, signaling. We measured HSV-1 replication by two methods; by measuring the replicating viral DNA load and by quantitating infectious virus yields. HSV-1 replication in L929 TRIF^−/− cells (8.8 × 10⁵ PFU/ml) was higher than that in wt L929 cells (1.6 × 10⁵ PFU/ml) and L929-TRIF cells (1.2 × 10⁵ PFU/ml) (Fig. 8A and B) because in L929 TRIF^−/− cells, both the TLR3 and the STING signaling pathways were nonfunctional in the absence of TRIF. In L929-TRIF-L/A cells, expressing the TRIF-L/A mutant, which supports TLR3 signaling but not STING signaling, HSV-1 replication (7.4 × 10⁵ PFU/ml) was found to be similar to its replication in L929 TRIF^−/− cells. This indicated that the STING pathway is the major attenuator of HSV-1 replication in these cells. In the case of HSV-1 replication in HT1080 reconstituted cells, we made an entirely different observation. Similar to the findings for L929 TRIF^−/− cells, HSV-1 replication in HT1080 TRIF^−/− cells (6.0 × 10⁸ PFU/ml) was found to be higher than that in wt HT1080 cells (1.9 × 10⁷ PFU/ml) and HT1080-TRIF cells (2.2 × 10⁷ PFU/ml) (Fig. 8C and D). This was due to the fact that HT1080 TRIF^−/− cells lack TRIF, thus rendering the TLR3 and STING signaling pathways nonfunctional. However, in HT1080-TRIF-L/A cells, expressing the TRIF-L/A mutant, which supports TLR3 signaling but not STING signaling, the HSV-1 replication level (7.4 × 10⁷ PFU/ml) was found to be similar to the HSV-1 replication level in HT1080-TRIF cells. This result indicates that in HT1080 cells the TLR3 pathway is the major contributor to the attenuation of virus replication and blocking only the cGAS-STING pathway does not promote virus replication much. To extend our observation of the differential contributions of the TLR3 and the STING pathways to HSV-1 replication in the mouse L929 and human HT1080 cell lines, we tested the phenomenon in mouse neuronal NB41A3 cells. As expected, in these cells, HSV-1 replicated extremely well and the absence of TRIF enhanced virus replication further (Fig. 8E). In NB41A3 cells expressing the TRIF mutant, virus replication (1.5 × 10¹¹ PFU/ml) was high and similar to that in TRIF^−/− cells (2.1 × 10¹¹ PFU/ml); in contrast, virus titers for wt NB41A3 cells and for NB41A3-TRIF cells were found to be lower, 2.3 × 10¹⁰ PFU/ml and 3.6 × 10¹⁰ PFU/ml, respectively, indicating that, like in L929 cells, the STING pathway was the major contributor to the antiviral effect in NB41A3 cells. In cells of another human cell line, HeLa-M, the HSV-1 replication pattern was found to be similar to that in the HT1080 cell line. In HeLa-M cells expressing the TRIF mutant, the level of virus replication (2.1 × 10⁷ PFU/ml) was similar to the level of virus replication in wt HeLa-M cells (1.6 × 10⁷ PFU/ml) and HeLa-M-TRIF cells (1.2 × 10⁷ PFU/ml), whereas the virus titers for HeLa-M TRIF^−/− cells were found to be higher (9.6 × 10⁷ PFU/ml) (Fig. 8F). Thus, the relative contributions of the cGAS-STING and the TLR3 pathways to the attenuation of HSV-1 replication are different in the HT1080, HeLa-M, L929, and NB41A3 cell lines (Fig. 8G).

FIG 8.

HSV-1 replication in the reconstituted cell lines. (A) HSV-1 DNA levels in TRIF^−/− or wt TRIF- and TRIF-L/A mutant-reconstituted L929 cells were quantified by qPCR at 16 h postinfection. (B) HSV-1 replication in TRIF^−/− or wt TRIF- and mutant-reconstituted cells along with wt L929 parent cells was determined by a TCID₅₀ assay at 16 h postinfection. (C) HSV-1 DNA levels in TRIF^−/− or wt TRIF- and mutant-reconstituted HT1080 cells were quantified by qPCR at 16 h postinfection. (D) HSV-1 replication in TRIF^−/− or wt TRIF- and mutant-reconstituted cells along with wt HT1080 parent cells was determined by a TCID₅₀ assay at 16 h postinfection. (E) HSV-1 replication in TRIF^−/− or wt TRIF- and mutant-expressing cells along with wt NB41A3 parent cells was determined by a TCID₅₀ assay at 16 h postinfection. (F) HSV-1 replication in TRIF^−/− or wt TRIF- and mutant-expressing cells along with wt HeLa-M parent cells was determined by a TCID₅₀ assay at 16 h postinfection. (G) Relative contributions of the cGAS-STING and TLR3 signaling pathways to attenuate HSV-1 replication in HT1080, HeLa-M, L929, and NB41A3 cells. The difference between HSV-1 replication in TRIF^−/− and in wt TRIF-reconstituted cells was attributed to the collective contribution of the cGAS-STING and TLR3 pathways, and the difference value was taken to be 100%. The difference between HSV-1 replication in TRIF-L/A and in wt TRIF-reconstituted cells was attributed to the contribution of only the cGAS-STING pathway, and the difference value was normalized to the collective contribution value of 100% of the cGAS-STING and TLR3 pathways. Similarly, the difference between HSV-1 replication in TRIF^−/− and in TRIF-L/A-reconstituted cells was attributed to the contribution of only the TLR3 pathway, and the difference value was normalized on a scale of 100% as explained above. ns, not significant (P > 0.05); *, P < 0.05; **, P < 0.01; ***, P < 0.001.

DISCUSSION

TRIF is a cytoplasmic protein that serves as the platform for signaling by TLR3, the endosomal double-stranded RNA (dsRNA)-recognizing TLR; it has the same role in the endosomal branch of TLR4 signaling, which is activated by lipopolysaccharide (28). Ligand binding to these receptors changes their conformation and exposes the TRIF-binding sites. For TLR3, TRIF binding is preceded by epidermal growth factor receptor- and Src-mediated phosphorylation of two specific tyrosine residues in its cytoplasmic domain; receptor-bound TRIF assembles different protein kinases and transcription factors to form the signaling complex (35). The TLR-TRIF interaction is mediated by the TIR domains of the partner proteins, whereas the RHIM domain of TRIF mediates apoptosis through its interaction with RIP1 and RIP3 (27, 36, 37). We made the unexpected observation that TRIF is also essential for STING signaling (31). STING is an endoplasmic reticulum-bound protein whose long cytoplasmic tail binds its ligands as well as assembles the signaling complex. Microbial or cellular cytoplasmic DNA binds and activates the enzyme cGAS, which synthesizes the STING ligand 2′,3′-cGAMP. We demonstrated that the STING and TRIF interaction leads to multimerization of STING, its intermembrane translocation, and transcriptional signaling, but unlike its interaction with TLRs, TRIF binds to STING directly through their carboxyl-terminal and not TIR domains, and the binding is constitutive and not ligand dependent (31). In the current study, we have characterized the STING-TRIF interaction further and identified the specific amino acid residues in TRIF that are required for STING binding.

We extensively documented the functional need of TRIF for STING-mediated signaling, as evidenced by the lack of STING dimerization, the activation of protein kinases and transcription factors, gene induction, and antiviral action in STING ligand-stimulated TRIF^−/− cells (31). On the other hand, Takashima et al. (38) did not observe a need for TRIF (TICAM-1) for STING signaling in myeloid cells. The reasons for the noted discrepancy are not apparent; however, different TRIF^−/− mouse strains and cells derived from them were used in the two studies (31, 38). To determine the generality of our observation, we tested additional murine and human cell lines in which the TRIF gene was disrupted by CRISPR/Cas9. In the current study, the TRIF gene was knocked out in three more cell lines, and, consequently, STING signaling was disrupted in all of them. The STING-interacting domain of both mouse and human TRIFs was mapped to a region between the TIR and RHIM domains. The amino acid residues in this domain are highly conserved between the two species, and the replacement of a single leucine residue by alanine eliminated TRIF’s ability to bind STING. However, the mutant TRIFs could still bind TLR3, presumably through their uninterrupted TIR domains. For functional testing of the mutant TRIFs, we used a mouse cell line, L929, and a human cell line, HT1080. These cell lines were chosen because they both express TLR3 and STING and can respond to their respective ligands. The TRIF gene was disrupted in the two lines by CRISPR/Cas9, and the TRIF-null cells could not respond to either poly(I·C), which is the TLR3 ligand, or 2′,3′-cGAMP, which is the STING ligand. The null cells were reconstituted with wt or mutant TRIFs, and cell clones that expressed TRIF at a level comparable to that in which it is expressed in the parental cells were selected for functional testing. In both cell lines, mutant TRIFs failed to support kinase activation, transcription factor activation, or gene induction in response to 2′,3′-cGAMP, the STING ligand, but they were competent to support poly(I·C)-stimulated TLR3 signaling.

Host-virus interactions are complex and multifaceted, with the type I IFN system playing a major role. Viruses induce IFN synthesis by triggering different intracellular pattern recognition receptors that recognize viral DNA or RNA as foreign objects (2, 7). In the case of the cytoplasmic DNA receptor cGAS-STING, cellular nuclear or mitochondrial DNA leaking into the cytoplasm of a virus-infected cell can serve as the activating ligand as well (39, 40). Similarly, not only viral dsRNA but also cellular dsRNA released from dead infected cells can be endocytosed and can activate TLR3 in neighboring cells (41, 42). In HSV-1-infected cells, both the cGAS-STING and the TLR3 pathways are activated and contribute to IFN induction and the resultant attenuation of virus replication (43). In this study, we took advantage of the newly generated TRIF mutants to assess the relative contributions of the two pathways in controlling virus replication. Our results indicate that in human HT1080 and HeLa-M cells, the TLR3 pathway is the major player. In contrast, in mouse L929 cells, the STING pathway is the dominant contributor. Because neurons are a primary target of HSV-1 replication in vivo, we tested HSV-1 replication in the mouse neuronal line NB41A3, which can respond to both STING and TLR3 ligands. The virus replicated to very high titers, and both the STING and the TLR3 pathways contributed significantly to virus attenuation in these cells. Therefore, we conclude that different cell types employ the two pattern recognition receptors differentially to attenuate HSV-1 replication, and it may be species specific.

MATERIALS AND METHODS

Cell lines.

The L929, HEK293T (293T), HT1080, HeLa-M, and Vero cell lines were maintained in Dulbecco’s modified Eagle’s medium (DMEM; LRI Cell Service-Media Core) supplemented with 10% fetal bovine serum (FBS), penicillin-streptomycin (100 U/ml), and glutamine (300 mg/ml). The NB41A3 cell line was maintained in Kaighn’s modification of Ham’s F-12 (F-12K) medium (LRI Cell Service-Media Core) supplemented with, 15% horse serum, 2.5% fetal bovine serum (FBS), penicillin-streptomycin (100 U/ml), and glutamine (300 mg/ml).

Plasmids.

Mouse wt TRIF and its N-terminal as well as C-terminal truncated forms were cloned into the pcDNA3.1(+) mammalian expression vector with a 3xFlag (3xF) epitope on their C termini. The scheme of the cloned inserts is BamHI-native Kozak sequence-TRIF without a stop codon-3xFlag-double stop codons-XbaI. Cloning of the mouse wt TRIF and its mutants was validated by sequencing and protein expression. Human wt TRIF with V5 and His epitopes on its C terminus was cloned into pcDNA3.1(+) as described previously (31). pcDNA3.1(+) TRIFΔ and TRIF-L/A STING-noninteracting mutants were generated using a Q5 site-directed mutagenesis kit (New England Biolabs) and validated by sequencing. Mouse wt TRIF and its N-terminal as well as C-terminal truncated forms were amplified using the mouse TRIF (mTRIF) BamHI forward and XbaI reverse primers shown in the Table 2.

TABLE 2.

Primers used to amplify TRIF and its truncated forms or for generation of mutations in TRIF^a

Primer name	Sequence (5′-3′)
mTRIF BamHI forward	AGACGGATCCTCTCCCCCATCCATGGATAACCCAGGGCCTTCGC
mTRIF XbaI reverse	GGATTCTAGATTATCACTTGTCATCGTCATCCTTGTAGTCGATGTCATGATCTTTATAATCACCGTCATGGTCTTTGTAGTCCTCTGGAGTCTCAAGAAGGGGTTCGC
478–732 BamHI forward	AGACGGATCCTCTCCCCCATCCATGCTTACACAGTCTGGGAGGCA
488–732 BamHI forward	AGACGGATCCTCTCCCCCATCCATGATCCCCCTCCTCCCACTTGAGT
498–732 BamHI forward	AGACGGATCCTCTCCCCCATCCATGGCCCAGCTCAGCCCAGATACA
508–732 BamHI forward	AGACGGATCCTCTCCCCCATCCATGCTCCACAGCATTGTGTGGCTG
518–732 BamHI forward	AGACGGATCCTCTCCCCCATCCATGTCCCCAATCTTCGCCAGAAAGGTG
528–732 BamHI forward	AGACGGATCCTCTCCCCCATCCATGACCTTCAAGACACAGAAGCTC
578–732 BamHI forward	AGACGGATCCTCTCCCCCATCCATGTATAGGGCCTGGCAAGCAGAG
628–732 BamHI forward	AGACGGATCCTCTCCCCCATCCATGCAGCCTCCATCCTTCCCTCAG
1–585 XbaI reverse	GGATTCTAGATTATCACTTGTCATCGTCATCCTTGTAGTCGATGTCATGATCTTTATAATCACCGTCATGGTCTTTGTAGTCCTCTGCTTGCCAGGCCCTATAG
1–595 XbaI reverse	GGATTCTAGATTATCACTTGTCATCGTCATCCTTGTAGTCGATGTCATGATCTTTATAATCACCGTCATGGTCTTTGTAGTCCTTCCCAAAAGCCACCCCAAG
1–605 XbaI reverse	GGATTCTAGATTATCACTTGTCATCGTCATCCTTGTAGTCGATGTCATGATCTTTATAATCACCGTCATGGTCTTTGTAGTCGCTGGGTGTTGGAGTCCCCAG
1–615 XbaI reverse	GGATTCTAGATTATCACTTGTCATCGTCATCCTTGTAGTCGATGTCATGATCTTTATAATCACCGTCATGGTCTTTGTAGTCAGAAGGTATTGGCTGTGGAC
1–625 XbaI reverse	GGATTCTAGATTATCACTTGTCATCGTCATCCTTGTAGTCGATGTCATGATCTTTATAATCACCGTCATGGTCTTTGTAGTCGGGGAAAACTGGAGTACCACCCTG
1–650 XbaI reverse	GGATTCTAGATTATCACTTGTCATCGTCATCCTTGTAGTCGATGTCATGATCTTTATAATCACCGTCATGGTCTTTGTAGTCGGATGGAGGCTGAGGGAAGGATG
1–700 XbaI reverse	GGATTCTAGATTATCACTTGTCATCGTCATCCTTGTAGTCGATGTCATGATCTTTATAATCACCGTCATGGTCTTTGTAGTCTGTGTGGCCCCACATGTGATTG
mTRIFΔ (586–95) forward	AACTTGTCACTGGGGACTCC
mTRIFΔ (586–95) reverse	CTCTGCTTGCCAGGCCCT
mTRIF-LGV/AAA forward	TGCTGCTTTTGGGAAGAACTTG
mTRIF-LGV/AAA reverse	GCAGCTTTGTTCATCTCTGCTTG
mTRIF-L/A (589) forward	GATGAACAAAGCTGGGGTGGCTTTTGGGAAGAAC
mTRIF-L/A (589) reverse	TCTGCTTGCCAGGCCCTA
hTRIFΔ (584–93) forward	CACATGTCATTTGGGACTG
hTRIFΔ (584–93) reverse	CTGTGCCTGGTAGGACAA
hTRIF-L/A (587) forward	GATGGAGCAGGCTCAGGTGGCTTTTG
hTRIF-L/A (587) reverse	TGTGCCTGGTAGGACAAG

^aThe primers used to amplify the mouse TRIF and its truncated forms or for generation of mutations in the mouse as well as human TRIF are presented.

Immunoprecipitation and immunoblotting.

For immunoprecipitation, cells were lysed in alternative lysis buffer as previously described (31). Lysates were precleared with mouse IgG-agarose (catalog number A0919; Sigma-Aldrich) for 1 h and incubated overnight with an anti-Flag M2 affinity gel (catalog number A2220; Millipore) or with an anti-V5 agarose affinity gel (catalog number A7345; Millipore). After incubation, the beads were washed with lysis buffer and boiled with 4× Laemmli sample buffer. Samples were boiled in 2× (vol/vol) Laemmli sample buffer. For Western blotting, cellular lysates were boiled in a 1× final concentration of Laemmli sample buffer, fractionated in an SDS-PAGE gel (Bio-Rad TGX gels), and processed for immunoblotting using standard techniques.

Quantitation of the protein-protein interaction.

Protein-protein interactions were quantified by measuring the optical density (OD) values by the use of ImageJ (version 1.44) software.

Sequence alignment.

Protein sequences were aligned using the global pairwise alignment tool EMBOSS Needle.

Generation of TRIF^−/− cell lines.

L929 TRIF^−/−, NB41A3 TRIF^−/−, HT1080 TRIF^−/−, and HeLa-M TRIF^−/− cells were generated using the CRISPR/Cas9 method. Introduction of a CRISPR-induced genomic deletion was performed by the overnight transduction of subconfluent L929 (mouse TRIF), NB41A3 (murine TRIF), HT1080 (human TRIF), or HeLa-M (human TRIF) cells with lentivirus (lentiCRISPRv2 expressing Cas9 endonuclease and the appropriate guide RNA sequence) using the method of Cong et al. (44). The mouse TRIF small guide RNA (sgRNA) sequence was CGTGAACCCCGAGTGATCGA, and the human TRIF sgRNA sequence was TGGCCCCGTCGGGCACGCCA. Posttransduction, virus-containing medium was replaced with complete medium to allow cellular growth for 48 h, at which time the cells were selected under puromycin. After drug selection, single cell clones were screened for maximum TRIF deficiency using genomic DNA sequencing and Western blotting.

Reconstitution of TRIF^−/− cells with STING-noninteracting TRIF mutants.

L929 TRIF^−/− and HT1080 TRIF^−/− cells were transfected with wt TRIF and the mutants. L929 TRIF^−/− and HT1080 TRIF^−/− cells were transfected with the Lipofectamine 2000 transfection reagent. At 24 h posttransfection, growth medium containing G418 (Life Technologies) was added onto the cells for clonal selection. For the L929 TRIF^−/− cells, 200 μg/ml of G418 was used for selection, and for HT1080 TRIF^−/− cells, 400 μg/ml of G418 was used. The medium was changed every 2 days. Individual colonies were selected, expanded, and validated for expression of wt TRIF and the mutants by Western blotting. NB41A3 TRIF^−/− and HeLa-M cells were transfected with wt TRIF and the mutants and selected with G418 (200 μg/ml) for a few passages to get a pool of cells with transient expression of TRIF and mutants at a similar level.

Transfection and ligand treatment.

Cells were transfected with 2′,3′-cGAMP (8 μg/ml; InvivoGen) by use of the Lipofectamine 2000 transfection reagent (Invitrogen) according to the manufacturer’s recommended protocol in incomplete DMEM. After 30 min of transfection, the medium was removed, the transfected cells were washed with phosphate-buffered saline, and fresh medium was added onto the cells. Poly(I·C) was directly added into the culture medium to a final concentration of 50 μg/ml. At 6 h posttransfection and/or posttreatment, the cells were lysed in alternative lysis buffer as previously described (31). The lysates were processed for immunoblotting. For HSV-1 replication in the cells pretreated with poly(I·C) and 2′,3′-cGAMP, the cells were treated as explained above but were treated overnight.

HSV-1 infection and replication.

Subconfluent monolayers of L929, NB41A3, HT1080, and HeLa-M reconstituted cells (1 × 10⁶ cells/well in 6-well plates) were inoculated at a multiplicity of infection (MOI) of 5 for 1 h at 37°C in incomplete DMEM. The inoculum was then removed, fresh Dulbecco’s modified Eagle’s medium (DMEM; LRI Cell Service-Media Core) supplemented with 10% fetal bovine serum (FBS), penicillin-streptomycin (100 U/ml), and glutamine (300 mg/ml) was added, and the cells were incubated at 37°C in 5% CO₂ for 16 h. Virus titers were determined by a 50% tissue culture infective dose (TCID₅₀) assay in Vero cells. The viral genome load in the infected cells was measured by quantitative PCR (qPCR) by determining the relative expression levels of the HSV-1 protein ICP50, expressed as the ratios of target ICP50 over 18S rRNA, which were graphed by using GraphPad Prism (version 5.0) software.

Antibodies.

The M anti-Flag MAb (MAb F1804), Rb anti-Flag MAb (MAb F7425), Rb anti-V5 (MAb V8137), Rb anti-hemagglutinin (anti-HA; MAb H6908), and M anti-β-actin (MAb A5441) were from Millipore. The M anti-V5 MAb (MAb R960-25) was from Invitrogen. The M anti-HA MAb (MAb 18181) was from Abcam. The rabbit anti-STING MAb (MAb 13647), rabbit anti-TRIF polyclonal antibody (antibody 4596S), Rb phospho-IκBα (MAb Ser32) MAb (MAb 2859), Rb anti-IκBα MAb (MAb 9242), Rb anti-IRF-3 MAb (MAb 4302), Rb anti-phospho-IRF-3 (Ser396) MAb (MAb 4947), Rb TBK1/NAK MAb (MAb 3013), Rb anti-phospho-TBK1/NAK (Ser172) MAb (MAb 5483), Rb anti-Akt (MAb 9272), and Rb anti-phospho-Akt (Ser473) (MAb 9271) were from Cell Signaling Technology. The M phospho-IκBα (Ser32) MAb (MAb sc-8404) was from Santa Cruz Biotechnology. The goat anti-mouse (MAb 610-1102) and goat anti-rabbit (MAb 611-1302) horseradish peroxidase-conjugated secondary antibodies were from Rockland Immunochemicals. Rb anti-mouse Ifit1 and anti-mouse Ifit2 and Rb anti-human Ifit1 were from our lab.

Statistics.

All experiments were repeated at least three times. Data are presented as the standard error of the mean (SEM), and P values between the groups were calculated by one-way analysis of variance using GraphPad Prism (version 5.0) software.