Deficiency in Either 4E-BP1 or 4E-BP2 Augments Innate Antiviral Immune Responses

Genetic deletion of both 4E-BP1 and 4E-BP2 was found to protect cells against viral infections. Here we demonstrate that the individual loss of either 4E-BP1 or 4E-BP2 in mouse embryonic fibroblasts (MEFs) is sufficient to confer viral resistance. shRNA-mediated silencing of 4E-BP1 or 4E-BP2 renders MEFs resistant to viruses, and compared to wild type cells, MEFs knockout for either 4E-BP1 or 4E-BP2 exhibit enhanced translation of Irf-7 and consequently increased innate immune response to viruses. Accordingly, the replication of vesicular stomatitis virus, encephalomyocarditis virus, influenza virus and Sindbis virus is markedly suppressed in these cells. Importantly, expression of either 4E-BP1 or 4E-BP2 in double knockout or respective single knockout cells diminishes their resistance to viral infection. Our data show that loss of 4E-BP1 or 4E-BP2 potentiates innate antiviral immunity. These results provide further evidence for translational control of innate immunity and support targeting translational effectors as an antiviral strategy.


Introduction
The 59 end of all nuclear-transcribed mRNAs possesses a cap structure that is recognized by the eukaryotic translation initiation factor 4E (eIF4E) [1][2][3]. eIF4E binds to the cap as a subunit of a complex (termed eIF4F) containing the large scaffolding protein eIF4G and the RNA helicase eIF4A [3][4][5]. The eIF4F complex facilitates 40S ribosome recruitment and canonical initiation factors, eIF4E and eIF4G, stimulate this reaction [6]. Almost all of the factors involved in the recruitment of the ribosome, including eIF4E, eIF4B, and eIF4G are phosphoproteins whose phosphorylation state correlates with translation efficiency and cellular growth rate [7]. The interaction between eIF4E and eIF4G is regulated by members of the eIF4E-binding proteins (4E-BPs), a family of translational repressors [8][9][10]. The mammalian family consists of three low molecular weight proteins: 4E-BP1, 4E-BP2, and 4E-BP3 [10]. The 4E-BPs compete with eIF4G for a shared binding site on eIF4E [11,12], in such a way that the binding of 4E-BPs and eIF4G is mutually exclusive [13]. The activity of 4E-BPs is controlled by the mammalian target of rapamycin (mTOR) kinase complex I (mTORC1), which consists of the protein kinase mTOR, RAPTOR (regulatory associated protein of mTOR), GbL (GTPase b-like protein) DEPTOR (disheveled, Egl-10, pleckstrin domain containing mTOR interacting protein) and PRAS40 (proline-rich Akt substrate of 40 kDa) [14][15][16]. Hypophosphorylated 4E-BPs bind with high affinity to eIF4E and repress translation. mTORC1-mediated 4E-BP hyperphosphorylation causes the dissociation of the 4E-BP/eIF4E inhibitory complex and thus stimulates cap-dependent translation [17,18]. A large body of evidence indicates that the mTOR pathway is an integral part of innate immunity through its critical roles in signaling and translational control of interferon stimulated genes (ISGs) [19][20][21][22].
Innate immunity constitutes the first line of defence against viral infection and type-I IFN is critical in this process [23][24][25][26][27][28]. Type-I IFNs are synthesized upon the activation of IRF-3 and IRF-7, which act as ''master'' transcription factors for IFN-a/b mRNAs [22,29,30]. Secreted IFN-a/b then activate the Janus kinase (JAK)/signal transducer and activator of transcription (STAT) pathway leading to the transcription of more than one hundred ISGs [31][32][33]. It is well documented that the mTOR pathway is also activated by type-I IFN [34][35][36][37][38][39] and is essential for type-I IFN production and innate immunity [21,40]. The critical role of the mTOR signaling pathway in innate immunity is based on several findings using MEFs harboring genetic ablation of mTOR-upstream and downstream components. For instance, the lack of the mTOR negative regulator TSC2 in MEFs enhances type-I IFN production [41]. In addition, MEFs knockout for mTOR regulators and effectors (such as AKT, PI3K, 4E-BPs, and S6Ks), have reduced or enhanced (for the 4E-BPs) type-I IFN production [20-22, 40, 42]. The mechanism by which lack of both 4E-BP1 and 4E-BP2 leads to the activation of IFN signaling was previously described in 4E-BP1/2 double knockout (DKO) MEFs, and involves translational derepression of IRF-7 mRNA (a potent transcription factor for type-I IFN genes) [22], Consistent with these results, in PI3K-depleted cells, IRF-7 expression is diminished [20]. These data are further supported by the observation that inhibition of PI3K or mTOR suppresses type-I IFN induction in plasmacytoid dendritic cells (pDCs) and in mice [19,21,42].
Based on the knowledge that 4E-BP1/2 DKO in MEFs confers resistance to viral infection, it is conceivable that inhibitors of 4E-BPs could be used as antiviral drugs. As proof of principle, we first asked whether depletion of 4E-BPs by shRNA would confer resistance to virus infection. We thus initiated studies using lentiviruses directed against 4E-BP1 or 4E-BP2 in MEFs. Surprisingly, the single knockdown for either 4E-BP1 or 4E-BP2 alone was sufficient to render MEFs resistant to infection by different viruses. Furthermore, we found that knockout cells for either 4E-BP1 or 4E-BP2 have similar phenotypes, and that reintroduction of one of the missing translational repressors restored susceptibility to virus infection. These results suggest that silencing either 4E-BP1 or 4E-BP2 can protect cells against viral infection and is sufficient to contribute to the translational regulation of innate antiviral responses.

Knockdown of 4E-BP1 or 4E-BP2 inhibits VSV replication
Lentiviruses encoding shRNAs directed against mouse 4E-BP1, 4E-BP2, or a scrambled sequence, were used to lower the expression of 4E-BP1 or 4E-BP2 in WT MEFs. After lentiviral transduction and cell selection, Western blotting analysis was performed to assess the depletion efficiency and specificity of the shRNAs. WT MEFs transduced with lentiviruses expressing shRNA against 4E-BP1 showed little to no detectable 4E-BP1 protein, as compared to MEFs transduced with a scrambled shRNA (Fig. 1A). Lentiviral depletion of 4E-BP1 was specific, since the expression of 4E-BP2 remained unaltered. Similarly, in MEFs transduced with lentiviral particles expressing a shRNA against 4E-BP2, 4E-BP2 was dramatically reduced but 4E-BP1 expression remained unchanged (Fig. 1A). Stable 4E-BP1 or 4E-BP2 knockdown MEFs were then infected with VSV-GFP at a multiplicity of infection (MOI) of 1. WT MEFs transduced with scrambled shRNA showed strong GFP expression and cytopathic effects (CPE) 12 hours post-infection (hpi). Strikingly, MEFs transduced with shRNA against 4E-BP1 or 4E-BP2 showed little to no GFP fluorescence or CPE upon infection (Fig. 1B). Western blotting analysis of the kinetics of viral protein production demonstrated that silencing 4E-BP1 or 4E-BP2 remarkably delayed the appearance of viral proteins (Fig. 1C) and decreased viral yield by up to 50 fold (Fig. 1D). Knockdown of 4E-BP1 or 4E-BP2 in human cells (HeLa) using shRNAs directed against the human proteins led to similar results albeit less pronounced as compared to MEFs (S1 Fig.). Therefore the single silencing of either 4E-BP1 or 4E-BP2 effectively prevents VSV infection/replication.
To determine if the VSV-resistant phenotype in 4E-BP1 2/2 and 4E-BP2 2/2 MEFs is applicable to other RNA viruses, we infected the cells with encephalomyocarditis virus (EMCV), Sindbis virus and influenza virus (FLU). Similar to VSV, cytopathic effects induced by these viruses were severely impaired in 4E-BP1 2/2 and 4E-BP2 MEFs as compared to their WT counterpart (Fig. 2E). To assess virus-induced cell death, we performed a MTT assay demonstrating higher cell viability for 4E-BP1 2/2 and 4E-BP2 2/2 MEFs than for WT MEFs (Fig. 2F). Taken together, these data show that depletion of either 4E-BP1 or 4E-BP2 is sufficient to impair the propagation of a broad spectrum of viruses.

Cells lacking 4E-BP1 or 4E-BP2 have elevated antiviral cytokine production
To determine whether deletion of either 4E-BP1 or 4E-BP2 increases type-I IFN production, we performed an interferon protection assay (Fig. 3A). Briefly, WT, 4E-BP1 2/2 or 4E-BP2 2/2 MEFs were treated for 6 hours with poly(I:C), a synthetic double-stranded RNA and a potent inducer of type-I IFN. The media were then harvested and used to condition WT MEFs. Following overnight incubation with the conditioned media, MEFs were challenged with VSV-GFP (1 MOI) and viral infections were analyzed by fluorescent microscopy and plaque assay (Figs. 3B and C). Culture medium from poly(I:C)-treated WT MEFs was insufficient to protect cells from VSV-GFP infection. In contrast, almost no GFP fluorescence or CPE was observed in WT MEFs incubated with the conditioned media from 4E-BP1 2/2 or 4E-BP2 2/2 MEFs (Fig. 3B). Accordingly, virus titers were significantly reduced in cells conditioned with media from MEFs lacking 4E-BP1 (,2 logs), or 4E-BP2 (,3 logs) (Fig. 3C). Altogether, these data show that lack of either 4E-BP1 or 4E-BP2 results in cells that secrete cytokines leading to a protective innate immune response.
We next determined the levels of Ifn-a, Ifn-b and Irf-7 mRNAs in WT, 4E-BP1 2/2 and 4E-BP2 2/2 MEFs by RT-PCR after either mock stimulation, or VSV infection. No Irf-7 mRNA was observed in mock stimulated cells, whereas the levels of Ifn-a and Ifn-b mRNAs were higher in MEFs lacking 4E-BP1 or 4E-BP2 as compared to WT MEFs (Fig. 3D). Upon, VSV infection, the induction of Irf-7, Ifn-a, and Ifn-b mRNAs in 4E-BP1 2/2 and 4E-BP2 2/2 MEFs was enhanced as compared to WT cells (Fig. 3D). The 59UTR of Irf-7 mRNA is highly structured [22], which can be sensitive to changes in the 4E-BPs/eIF4E ratio [8,43,44]. To assess the effect of the absence of 4E-BP1 or 4E-BP2 on Irf-7 mRNA translation, we used a luciferase reporter construct containing the IRF-7 59UTR [22]. The translation of this mRNA was elevated in MEFs lacking either 4E-BP1 or 4E-BP2 when compared to WT cells (Fig. 3E). Furthermore, to monitor the expression of ISRE and IFN-b genes in 4E-BP1 2/2 and 4E-BP2 2/2 MEFs, we used two constructs in which the expression of luciferase is driven by either the ISRE or IFN-b promoter. Luciferase expression from both constructs was increased in 4E-BP1 2/2 and 4E-BP2 2/2 MEFs as compared to WT MEFs (Fig. 3E). These results demonstrate that the production of type-I IFN following VSV infection, and the translational regulation of the Irf-7 mRNA, is enhanced in MEFs lacking either 4E-BP1 or 4E-BP2.

Re-introduction of 4E-BP1 or 4E-BP2 in double or single knockout MEFs restores the susceptibility to VSV infection
To support the findings that lack of either 4E-BP1 or 4E-BP2 is sufficient to confer resistance to viral infection, we reintroduced 4E-BP1 or 4E-BP2 in 4E-BP1 2/2 4E-BP2 2/2 DKO MEFs. DKO MEFs were transduced with a control retrovirus, or a retrovirus expressing either recombinant 4E-BP1 or 4E-BP2. Expression of 4E-BP1 or 4E-BP2 was confirmed by Western blotting (Fig. 4A). Cells were then infected with VSV-GFP at a MOI of 1 PFU/cell and infection was analyzed by fluorescence microscopy at 20 hpi. Re-expression of either 4E-BP1 or 4E-BP2 augmented the susceptibility of DKO MEFs to VSV infection (Fig. 4B). Viral proteins appeared after 14 hpi in DKO MEFs expressing recombinant 4E-BP1, and at 16 hpi in DKO MEFs expressing recombinant 4E-BP2. By contrast, in  control cells transduced with empty vector, viral proteins were not detected (Fig. 4C). Using the same strategy, 4E-BP1 or 4E-BP2 were reintroduced into 4E-BP1 2/2 and 4E-BP2 2/2 MEFs, respectively After cell selection, recombinant 4E-BP expression was assessed via Western blotting (Fig. 5A) and cells were infected with VSV-GFP at a MOI of 1 PFU per cell. At 16 hpi, an increase in infection was found rescued 4E-BP1 2/2 MEFs compared to control cells; however, little differences in terms of fluorescence were detected between control and rescued 4E-BP2 2/2 MEFs at this time or later time points (Fig. 5B). Western blotting analysis of whole extracts prepared from infected cells (16 hpi) confirmed the difference in infection observed by fluorescence between 4E-BP1 2/2 MEFs and cells rescued with 4E-BP1 (Fig. 5C). There was also an increase in viral protein accumulation in rescued 4E-BP2 2/2 MEFs compared to control cells when assessed by Western blotting although less dramatic than in 4E-BP1 2/2 rescued cells (Fig. 5C). Altogether, these results indicate that the reintroduction of either 4E-BP1 or 4E-BP2 in cells lacking these translation repressors enhance their susceptibility to VSV infection. The data from knockdown on 4E-BP1 or 4E-BP2 expression in WT MEFs, and the restored susceptibility of 4E-BP1/2 DKO MEFs following reintroduction of either 4E-BP1 or 4E-BP2, provide genetic evidence that the virus resistant phenotype of 4E-BP1 2/2 and 4E-BP2 2/2 MEFs is associated with 4E-BP1 or 4E-BP2 expression.

Discussion
Our results show that MEFs lacking 4E-BP1 or 4E-BP2 have enhanced type-I IFN production, higher resistance to viral infection, and increased IRF-7 59UTRreporter gene translation as compared to WT cells. Furthermore, re-introduction of one of the missing translational repressors in DKO (4E-BP1 2/2 4E-BP2 2/2 ) MEFs, or single knockout MEFs, renders cells increasingly susceptible to viral infection. Our data indicate that the loss of one 4E-BP is sufficient to alleviate a negative regulatory effect on the translation of the Irf-7 mRNA, potentiating a more efficient innate immune response. Pathogens, including viruses, induce an antimicrobial and pro-inflammatory cytokine production that serves as the initial barrier to infection [22,[26][27][28]45]. However, continued cytokine expression can have detrimental effects; therefore, the balance between activators and repressors of innate immune responses needs to be tightly regulated [46][47][48][49]. Many activators of the type-I IFN have been described, and recently repressors of this signaling pathway have been reported. The SOCS (suppressors of cytokine signaling) are classical examples [50]. Our study strengthens the findings that 4E-BP1 and 4E-BP2 act as translation repressors of innate immunity by negatively regulating mRNA translation of the master regulator of type-I IFN, IRF-7 [22].
In mammalian cells, there are three 4E-BPs (4E-BP1, 4E-BP2, and 4E-BP3), with 4E-BP1 being the best characterized [9,10]. In MEFs, 4E-BP3 is expressed at undetectable levels [22,51]. MEFs lacking both 4E-BP1 and 4E-BP2 are extremely resistant to viral infections due to increased type-I IFN production and high Irf-7 mRNA translation [22]. Kaur and colleagues previously reported that cells lacking only 4E-BP1 have increased type-I IFN production [41]. Here we show that silencing either 4E-BP1 or 4E-BP2 augments the cell resistance to viral infections in MEFs. We found that loss of 4E-BP1 or 4E-BP2 results in reduced translational repression of IRF-7 and an enhancement of innate antiviral responses. While we observed variations in viral susceptibility between cells expressing either 4E-BP1 or 4E-BP2, the significance of these differences is currently being investigated to determine whether distinct ISGs could be selectively regulated at the translation level by either 4E-BP1 or 4E-BP2. Interestingly, 4E-BP expression levels vary among different tissues [52]. Such a definite expression of 4E-BPs may suggest potentially distinct biological functions with regard to tissue-specific gene expression, cell differentiation, and perhaps innate immune responses [53,54].
In conclusion, we report that cells lacking 4E-BP1 or 4E-BP2 have an augmented type-I IFN production and are resistant to various viral infections. These results confirm that 4E-BP1 and 4E-BP2 play critical roles in the translation regulation of innate immune responses and support the basis for targeting these translational repressors as an antiviral strategy.

Western blotting analysis
MEFs were homogenized in buffer A (50 mM Tris-HCl, pH 7.4, 100 mM NaCl, 1% Triton X-100, 1 mM EDTA, 1 mM dithiothreitol, protease inhibitor cocktail (Roche), 20 mM b-glycerophosphate, 0.25 mM Na 3 VO 4 , 10 mM NaF, 10 nM okadaic acid, 1 mM PMSF) and incubated for 30 min at 4˚C. Cell debris was removed by centrifugation at 10,000 g for 10 min at 4˚C and total protein content was determined using a Bio-Rad assay. Laemmli sample buffer was added to the supernatant, which was then subjected to SDS-PAGE (15%). Proteins were transferred onto a nitrocellulose membrane, which was blocked for 2 hours at room temperature with 5% skim milk in PBS containing 0.2% Tween-20 (PBS-T) and washed twice with PBS-T. The membrane was incubated overnight at 4˚C with primary antibodies followed by three 10-min washes in PBS-T and further incubated with peroxidase-coupled secondary antibody for 30 min at room temperature, and washed three times. Detection of peroxidase-coupled secondary antibody was performed with ECL (GE-Healthcare). 4E-BP1 and 4E-BP2 antibodies were purchased from Cell Signaling and the b-actin antibody was purchased from Santa Cruz Biotechnology. VSV antibody was a gift from J. Bell.

Virus titration by plaque assay
The day prior to the assay, HEK (human embryonic kidney) cells were seeded onto six-well plates at 3610 5 cells/well in 2 ml of growth media. Cells were maintained at 37˚C with 5% CO 2 . On the day of assay, serial viral dilutions were made in the inoculation medium. Growth medium was aspirated from cells, and 0.5 ml of each dilution was added per well and allowed to adsorb for 2 hours at 37˚C in a humidified incubator containing 5% CO 2 . The agarose overlay was prepared by mixing one part of 26McCoy's 5A containing 30% heat-inactivated FBS and 2% DMSO with one part of pre-warmed 1.5% Sea Plaque low melting agarose. The mixture was kept warm at 42˚C. The inoculum was removed and 5 ml of the pre-warmed agarose overlay was added to the cells. The agarose was allowed to set for at least 30 min at room temperature prior to a 35 or 37˚C incubation with 5% CO 2 . Plaques were counted on day 5 as follows: 3 ml of 10% formaldehyde was added to each well, the plates were allowed to sit at room temperature for 1 h, and then the formaldehyde was aspirated and agarose overlay was carefully removed. The cells were stained with 2 ml of 0.4% crystal violet in an aqueous alcohol solution (Becton Dickinson Microbiology Systems, Cockeysville, MD) for 2 min. The staining solution was decanted and the plaques were washed with water before inverting the plates and allowing them to dry. The plaques were counted and expressed as plaque forming units (PFU). Results are the mean ¡SD of three independent experiments, the P values were calculated using the Student's T-test where: (*) P,0.05; (**) P,0.01; (***) P,0.001.

Plasmid constructs and luciferase assay
MEFs were seeded at 3610 4 cells per 24-well then were transiently co-transfected with 500 ng of 59 UTR-IRF-7-Fluc vector and 50 ng of transfection control, Renilla luciferase vector (Rluc; Promega), which is under the control of the CMV promoter, using Lipofectamine 2000 as describedhttp://www.nature.com/nature/ journal/v452/n7185/full/nature06730.html -B36 in [22]. For the promoter reporter assay, we used the same condition as described previously; MEFs were co-transfected with a plasmid containing either the IFN-a promoter or the IFN-b promoter followed by firefly luciferase (FL). In this last assay we used as a transfection-control, a plasmid containing the thymidine kinase (TK)-promoter followed by Renilla luciferase RL (kind gift from D. Muruve). Twenty hours after transfection, cell extracts were prepared in passive lysis buffer (Promega) and assayed for Rluc and Fluc activity in a Lumat LB9507 bioluminometer (EG&G Bertold) using a dual-luciferase reporter assay system (Promega), according to the manufacturer's instructions. Fluc activity was normalized against Rluc activity, which was used as a transfection control. Data represent the mean ¡SD of firefly luciferase value of three to six samples, normalized by Renilla luciferase activity. The P values were calculated using the Student's T-test where: (*) P,0.05; (**) P,0.01; (***) P,0.001.

Protection assay
The VSV protection assay was performed as followed: WT, 4E-BP1 2/2 and 4E-BP2 2/2 MEFs, were either mock-treated, or transfected with 1 mg/ml of poly(I:C) (Sigma) using FuGENE 6 transfection reagent (Roche) according to the manufacturer's protocol. Supernatants were collected at 6 hours from mock or stimulated cells and were subsequently added to the cells overnight. On the next day, cells were infected with VSV WT or VSV-GFP at a MOI of 1 for a 24-hour period. Viral infection was assessed by fluorescence microscopy and cell viability was measured at 12 or 24 hours post-infection by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazoium bromide (MTT) assay.

MTT cell viability assay
MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide) (Sigma, St. Louis, MO, USA) was used to assess cell viability after viral infection. MTT was diluted to 5 mg/ml in sterile phosphate buffered saline (PBS) and filtered once through a 0.45 mm filter. The diluted MTT was added to the wells in a final volume of 25 ml/well. Four hours later, 100 ml of DMSO were added to each well. The microtiter plates were incubated at 37˚C overnight to dissolve reduced MTT crystals. Absorbance values were determined by ELISA reader (Multiskan Plus, Lab Systems, Helsinki, Finland) at dual wavelengths of 570 and 690 nm. Results are the mean ¡SD of three independent experiments, the P values were calculated using the Student's T-test where: (*) P,0.05; (**) P,0.01; (***) P,0.001.