ABSTRACT
The RNA-dependent protein kinase (PKR) has broad antiviral activity inducing translational shutdown of viral and cellular genes and is therefore targeted by various viral proteins to facilitate pathogen propagation. The pleiotropic NS1 protein of influenza A virus acts as silencer of PKR activation and ensures high-level viral replication and virulence. However, the exact manner of this inhibition remains controversial. To elucidate the structural requirements within the NS1 protein for PKR inhibition, we generated a set of mutant viruses, identifying highly conserved arginine residues 35 and 46 within the NS1 N terminus as being most critical not only for binding to and blocking activation of PKR but also for efficient virus propagation. Biochemical and Förster resonance energy transfer (FRET)-based interaction studies showed that mutation of R35 or R46 allowed formation of NS1 dimers but eliminated any detectable binding to PKR as well as to double-stranded RNA (dsRNA). Using in vitro and in vivo approaches to phenotypic restoration, we demonstrated the essential role of the NS1 N terminus for blocking PKR. The strong attenuation conferred by NS1 mutation R35A or R46A was substantially alleviated by stable knockdown of PKR in human cells. Intriguingly, both NS1 mutant viruses did not trigger any signs of disease in PKR+/+ mice, but replicated to high titers in lungs of PKR−/− mice and caused lethal infections. These data not only establish the NS1 N terminus as highly critical for neutralization of PKR's antiviral activity but also identify this blockade as an indispensable contribution of NS1 to the viral life cycle.
IMPORTANCE Influenza A virus inhibits activation of the RNA-dependent protein kinase (PKR) by means of its nonstructural NS1 protein, but the underlying mode of inhibition is debated. Using mutational analysis, we identified arginine residues 35 and 46 within the N-terminal NS1 domain as highly critical for binding to and functional silencing of PKR. In addition, our data show that this is a main activity of amino acids 35 and 46, as the strong attenuation of corresponding mutant viruses in human cells was rescued to a large extent by lowering of PKR expression levels. Significantly, this corresponded with restoration of viral virulence for NS1 R35A and R46A mutant viruses in PKR−/− mice. Therefore, our data establish a model in which the NS1 N-terminal domain engages in a binding interaction to inhibit activation of PKR and ensure efficient viral propagation and virulence.
INTRODUCTION
Influenza A viruses (IAVs) are widespread human pathogens causing epidemics and pandemics of respiratory disease and also having large host reservoirs in animals such as pigs and birds (1). IAVs have a segmented single-stranded RNA genome of negative polarity, which is replicated by the viral polymerase in the cell nucleus and packaged into progeny virions at the plasma membrane following export of replicated viral genome segments into the cytosol in the late phase of infection (2). The influenza virus RNA genome triggers innate antiviral responses in the infected cell, which include prominent transcriptional upregulation of type I and III interferon (IFN) genes (3–5). Released IFNs can induce an antiviral state in cells in the tissue by autocrine or paracrine binding to type I or III IFN receptors resulting, in the activation of the JAK/STAT pathway and transcription of IFN-stimulated genes (ISGs) involved in the innate antiviral defense (reviewed in reference 6).
An important factor among the several hundred ISGs is the pleiotropic RNA-dependent protein kinase (PKR), as mice with engineered genetic deletions showed increased susceptibility toward viruses (7–9). PKR is a latent serine-threonine kinase that consists of two successive N-terminal double-stranded RNA (dsRNA) binding motifs that are separated by a flexible linker region from the C-terminal kinase domain. Activation of the cytosolic monomeric PKR in virus-infected cells is believed to be mediated by recognition of viral double-stranded or structured single-stranded RNAs with a 5′-triphosphate cap, which triggers dimerization and autophosphorylation of the kinase establishing the catalytically active enzyme (10, 11). Activated PKR phosphorylates the eukaryotic initiation factor 2α (eIF2α), resulting in translation inhibition that is deleterious to efficient viral growth (12). Furthermore, PKR is involved in a set of additional pathways, such as enhancement of IFN-β levels following virus infections (13, 14), activation of NF-κB (15), inflammatory reactions (16), metabolic responses (17), and regulation of cellular growth (18).
Given its potential to severely restrict virus propagation, it is not surprising that several virus families have evolved mechanisms to avoid or to block activation of PKR (19). These include expression of an eIF2α pseudosubstrate such as the vaccinia virus K3L protein (20, 21), induced degradation of PKR (22–24), formation of an inhibitory complex with PKR (25, 26), or expression of PKR-inhibitory small noncoding RNA (27). PKR inhibition by the vaccinia virus dsRNA binding protein E3L had originally been attributed to a sequestration mechanism (28), but this view has recently been challenged by mutational analysis suggesting that this E3L activity is not strictly depending on dsRNA binding (29).
In case of influenza viruses, PKR is silenced by the highly expressed NS1 protein (30–32). The IAV NS1 protein consists of about 230 amino acids (aa), depending on the strain (33, 34), and forms homodimers or multimers (35, 36). The protein comprises two separate domains, an N-terminal RNA binding domain (RBD) (aa 1 to 72) and a C-terminal effector domain (ED) (aa 85 to 220), joined by a short linker domain (LD) (aa 73 to 84) (37, 38). The RBD binds to a variety of RNA species with variable affinity (39–42). Crystal structure and NMR analyses, in combination with biochemical studies, showed that the RBD exists as a dimer consisting of two antiparallel α-helices that form a conserved surface on which R35 and R46′ as well as R38 and R38′ build hydrogen bonds fastening the backbone of a dsRNA track (36, 43–45). In addition, amino acids K41, S42, and T49 within the RNA binding pocket were identified as being important for RNA binding (43, 44). Positions R35 and R46 were also described to be critical for dimerization of the protein in vitro or in transfected cells (36, 43, 46). The functional importance of this conserved structure is highlighted by the finding that R35, R38, and K41 are also part of a nuclear import signal (47, 48). Recent analyses showed that the NS1 effector domain also contains inherent dimerization activity which allows NS1 to take on different quaternary conformations (49–52).
In addition to blocking PKR, the IAV NS1 protein has additional functions in infected cells; these include inhibition of other innate immune reactions, in particular type I IFN induction via targeting RIG-I-dependent signals (53, 54), the 2′,5′-oligoadenylate synthetase (2′,5′-OAS)/RNase L pathway (55), and production of proinflammatory interleukin-1β (56, 57). Moreover, the NS1 proteins of many IAVs bind CPSF30 (the 30-kDa subunit of the cleavage and adenylation specificity factor) and thereby inhibit 3′-end processing of cellular pre-mRNAs, which also affects maturation of IFN-β mRNA (summarized in reference 58). NS1 also promotes efficient expression of late viral gene products (59, 60) and has been found to prevent early apoptosis in infected cells involving activation of the phosphatidylinositol 3-kinase (PI3K)/Akt pathway (61, 62). These activities likely form the basis for NS1's role in controlling viral pathogenicity and tropism (63, 64).
Despite the detailed structural information, the precise mechanism by which IAV NS1 inhibits activation of PKR in virus-infected cells is far from being well understood. Previous studies employed various biochemical and genetic approaches and resulted in conflicting models. Some investigators suggested an important role of RNA binding, possibly enabling a sequestration mechanism, since constitutive or conditional mutations within the RBD strongly reduced the ability of NS1 to inhibit PKR (65, 66). Another study provided evidence that NS1 binds PKR through its RBD and speculated this interaction to be mediated by RNA (67). Finally, two reports suggested formation of a complex between NS1 and PKR preventing dsRNA-mediated stimulation of PKR autophosphorylation and postulated that NS1 amino acids at positions 123 to 127 outside the RBD were critical for this to occur (68, 69). These conclusions were supported by binding assays and increased phosphorylation of PKR in cells infected with two recombinant influenza A/Udorn/72 viruses expressing NS1 mutant proteins in which amino acids 123/124 or 126/127 had been changed to alanines. However, the exact manner of inhibition has remained uncertain.
To clarify the basis for NS1's inhibitory action on PKR, we systematically analyzed a set of recombinant viruses expressing NS1 proteins with mutations in key amino acid positions in the RBD and ED. Investigations conducted in human cells with physiological or strongly reduced levels of PKR revealed a pivotal role of amino acids R35 and R46 within the RBD for the inhibition of PKR and efficient propagation in human cells. These findings were strongly supported by a murine model of influenza, in which the corresponding NS1 R35 and R46 mutant viruses were benign in PKR+/+ mice but propagated in and killed PKR −/− mice like wild-type (WT) virus. Unlike all other tested mutants, the NS1 R35A and R46A proteins associated with neither RNA nor PKR in biochemical assays and intracellular Förster resonance energy transfer (FRET)-based studies in living cells. An alanine replacement of R38 strongly reduced dsRNA binding of the NS1 protein but did not affect its capacity to inhibit PKR activation. These findings not only support a model in which NS1 silences activation of PKR by an RNA-independent inhibitory binding interaction but also highlight the versatile activities of the N-terminal domain of the NS1 protein.
RESULTS
Recombinant influenza A viruses with mutations in the RNA binding and effector domains of the NS1 protein.To assess the roles of distinct NS1 amino acids in binding to PKR and regulating its activity, we generated a panel of recombinant influenza A/PR/8/34-based viruses encoding mutations in the RBD and/or ED of NS1 (Fig. 1A). Targeted amino acid positions were selected based on previous reports analyzing interactions of influenza virus NS1 proteins with PKR and dsRNA by means of biochemical, biophysical, and molecular dynamics simulation tools (19, 36, 43–45, 66, 69). Mutations in the NS1 RBD consisted of single exchanges of arginine or lysine residues (R35, R38, K41, and R46) to alanine, whereas double alanine exchanges were constructed in the ED at amino acid positions I123/M124 and K126/N127, which were previously implicated in PKR binding (69). We also produced the so-called NS1 Triple mutant in which amino acids R38, I123, and M124 were replaced by alanine, which was predicted to bind to neither PKR nor to dsRNA. All viruses were rescued from cells, and stocks were produced in 7-day-old embryonated chicken eggs to minimize selection of revertant viruses, as the IFN system is immature in these hosts (70). Sequence analyses confirmed the desired NS1 mutations in the absence of unwanted changes. In parallel, we constructed plasmids facilitating the expression of corresponding NS1 proteins in cells and in vitro.
Generation and PKR binding analysis of recombinant influenza A virus NS1 proteins with mutations in the RNA binding domain (RBD) and effector domain. (A) The schematic diagram indicates the amino acid positions mutated to alanine within the NS1 protein. The asterisks designate the mutated amino acid positions altered in the triple mutant R38A/I123A/M124A [NS1(Triple)]. (B) In vitro-translated [35S]methionine-labeled NS1 proteins were incubated with GST-PKR(K296R)- or control GST-PKR(aa266-551)-Sepharose. Precipitated proteins were separated by SDS-PAGE. Coomassie blue-stained gels depicting GST-PKR(K296R) and GST-PKR(aa266-551) were dried, and precipitated NS1 proteins were detected by autoradiography (upper panel) and quantified by densitometry of bands with normalization to PKR 266-551 (lower panel) (n = 4, mean + standard error of the mean [SEM]). (C) Model for FRET measurement shown in panel D. Top, the acceptor (eGFP) is farther than 10 nm away, and Förster resonance energy transfer (FRET) cannot occur. Therefore, τD is not influenced by the presence of the acceptor and is defined in this case as “donor-only lifetime” (∼3.6 ns). Bottom, the eGFP acceptor is closer (≤10 nm), and mTurquoise will thus transfer energy to the acceptor in a nonradiative way (i.e., FRET occurs). The donor lifetime will accordingly decrease (i.e., τDA is “donor-plus-acceptor lifetime,” ∼3.1 ns), as shown in the pseudocolor FLIM image. The average FRET efficiency (EFRET) was determined by calculating the ratio of donor-plus-acceptor to donor-only lifetime. (D) HEK293T cells were transfected with pmTurquoise-NS1 constructs and peGFP-PKR in the presence of poly(I·C)-rhodamine. Calculation of the FRET efficiency between mTurquoise and eGFP was used to measure an interaction between the NS1 proteins and PKR. Data are shown as box plots, and outliers are represented individually as dots (Tukey style) (n ≥ 11 spots/cells from at least two independent experiments). (E) Immunoblot analysis of eGFP-PKR and mTurquoise-NS1 WT and mutant proteins. HEK293T cells were transfected as described for panel D, lysed at 16 h posttransfection, and subjected to immunoblot detection for PKR, NS1, and tubulin.
NS1 mutant proteins R35A and R46A are unable to bind PKR.As previous studies suggested that NS1-mediated inhibition of PKR involves a direct protein-protein interaction, we first studied the binding of NS1 proteins to a glutathione S-transferase (GST)-PKR(K296R) fusion protein in an established coprecipitation assay (64) (Fig. 1B). Wild-type NS1 as well as the mutant proteins NS1-I123A/M124A and NS1-K126A/N127A were efficiently precipitated by full-length PKR, whereas mutants NS1-R38A and -K41A as well as -Triple presented a 30 to 40% reduced interaction. In contrast, the mutants NS1-R35A and -R46A were not captured by PKR (Fig. 1B). A fusion protein carrying only the PKR catalytic domain [GST-PKR(266-551)] was used as a control unable to interact with any NS1 variant (68).
In a second approach, we analyzed the NS1-PKR interaction by documenting FRET in transfected cells expressing the fluorescent fusion proteins mTurquoise-NS1 and enhanced green fluorescent protein (eGFP)-PKR in the presence of rhodamine-labeled poly(I·C), a synthetic dsRNA stimulating PKR. Wild-type NS1 and NS1 mutant proteins R38A, K41A, I123A/M124A, K126A/N127A, and Triple colocalized with eGFP-PKR and poly(I·C)-rhodamine in expected spot-like structures (71), in which FRET from mTurquoise to eGFP was measured (Fig. 1C and D and data not shown). In contrast, NS1 mutant proteins R35A and R46A did not vary in intensity between PKR-reactive spots or regions outside those. Measurements were performed on the same location as the PKR reactive spot and presented only background levels of FRET (EFRET), confirming that these mutants do not bind to PKR (Fig. 1D). Comparable expression of eGFP-PKR and all mutant mTurquoise-NS1 proteins was confirmed (Fig. 1E). Hence, we concluded that NS1 can form an intracellular complex with PKR that critically requires amino acid R35 and R46 in the RBD but none of the other mutagenized positions (R38, K41, I123/M124, and K126/N127).
Mutant NS1 proteins R35A, R38A, and R46A are unable to bind to dsRNA.Previous reports indicated that NS1 amino acids R35, R38, and R46 are essential for dsRNA binding of a purified recombinant NS1 RBD (residues 1 to 70 or 1 to 73) (36, 43, 44). Therefore, we assessed the impact of alanine replacements in the context of a full-length NS1 protein. HEK293T cells were transfected with plasmids expressing WT NS1 and mutant proteins. Cell lysates were prepared at 48 h posttransfection, followed by precipitation with synthetic dsRNA [poly(I·C)] coupled to agarose beads and subsequent immunoblot analysis of bound proteins using NS1-specific antiserum. The NS1 WT and mutant proteins K41A, I123A/M124A, and K126A/N127A were found to bind to synthetic dsRNA, whereas the mutants R35A, R38A, and R46A and the Triple variant had in essence lost this ability (Fig. 2). We therefore concluded that residues R35, R38, and R46 are crucial for RNA binding of full-length NS1.
dsRNA binding of mutant NS1 proteins. HEK293T cells were transfected with pHW2000 plasmids expressing WT or mutant NS1 for 48 h. Cell lysates were used for precipitation with poly(I·C)-coupled agarose. Bound proteins were separated by SDS-PAGE and analyzed by immunoblotting for NS1. Input, 2% of whole-cell lysate. Results from representative experiment of 3 are shown.
Mutations in the RBD do not eliminate NS1 dimerization in cells.NS1 protein function is believed to depend on formation of dimers or multimers (35, 43). We therefore examined dimerization of mutant NS1 proteins, first by coimmunoprecipitation. Cells were transfected with peYFP-NS1 or empty vector, followed by infection with recombinant WT or mutant virus (Fig. 3A). Lysates were prepared at 16 h postinfection (hpi) and used for precipitation by a GFP-Trap, a matrix with covalently bound anti-GFP antibody recognizing enhanced yellow fluorescent protein (eYFP). Precipitated proteins were analyzed by immunoblotting for NS1. As shown in Fig. 3A, all viral NS1 (vNS1) mutant proteins bound to eYFP-NS1 wild-type protein. It was unexpected to detect signals of dimer formation for the R35A and R46A mutants, as previous studies had indicated critical roles of these positions for NS1-NS1 interaction (36, 43). Interestingly, dimer formation was also maintained for the NS1 R35A, R38A, R46A, and Triple mutants when the identical mutation was present in the transfected eYFP-NS1 (data not shown). Figure 3B demonstrates that transfected NS1 R35A and R46A mutants also dimerized with eYFP-NS1 in the absence of other viral proteins. Moreover, we verified that the experimental conditions used were suitable to characterize a dimerization-defective NS1 mutant (M4) containing a simultaneous deletion within the N-terminal domain and alanine replacement of Trp187, the latter of which was previously shown to be important for effector domain dimerization (49, 52). These findings suggested that dimers of full-length NS1 containing mutations in the N-terminal domain could be stabilized by the described homotypic interaction of the effector domain (49, 52).
Dimerization of mutant NS1 proteins. (A) HEK293T cells were transfected with peYFP-NS1 wild-type (eYFP-NS1 WT) and subsequently infected with the indicated recombinant viruses (vNS1). The control was transfected with peYFP empty vector and infected with WT virus. Cell lysates were used for pulldown experiments using GFP-Trap. Precipitated proteins were separated by SDS-PAGE followed by immunoblotting for NS1. Results from a representative experiment of 3 are shown. (B) Dimerization assay for plasmid-expressed NS1 mutants. Lysates of HEK293T cells transfected with peYFP-NS1 WT (eYFP-NS1) together with plasmids expressing NS1 WT, NS1-M4, NS1-R35A, NS1-R46A, or empty vector were used for precipitation by GFP-Trap as described for panel A. (C) Model for homo-FRET measurement. Top, fluorophores with an excitation dipole parallel to the excitation light will be excited. Fluorescence emission will occur before the molecule has had the chance to rotate, and therefore, the emitted photons will be detected mostly by the parallel detector. This results in high anisotropy values. Bottom, a randomly orientated mixture of homodimers is excited by polarized light. In this case, the excited fluorophores will nonradiatively transfer the energy to the next fluorophore of the same species but different orientation. Radiative emission will originate from the acceptor fluorophore. Due to the formation of homodimers/homo-FRET, there will be a loss of correlation between the polarization of the excitation and emission light, and the measured anisotropy will decrease. (D) Anisotropy calculation indicating dimerization ability of NS1 constructs. HEK293T cells were transfected with peYFP-NS1 WT or mutant plasmids. For a control, eYFP-NS1 WT was cotransfected with nontagged NS1 WT and measured. Data are shown as box plots (Tukey style) (n ≥ 10 spots/cells from at least two independent experiments).
To evaluate those findings in a cell-based assay, we assessed NS1 complex formation likely representing dimerization by fluorescence anisotropy imaging, in which we transfected peYFP-NS1 fusion constructs and calculated anisotropy histograms from parallel and perpendicular fluorescence signals (Fig. 3C and D). In this assay, a decrease of anisotropy indicates dimerization of a protein due to homo-FRET (see Materials and Methods). This analysis showed a decrease in anisotropy for the WT as well as for all mutant NS1 proteins, including R35A and R46A, compared to the control setting in which an eYFP-NS1 fusion was paired with an unlabeled NS1 protein (Fig. 3D). In conclusion, these data suggest that the introduced mutations, including the ones at positions 35 and 46, do not eliminate intracellular dimerization of full-length NS1 proteins.
PKR is strongly activated by NS1 mutant viruses R35A and R46A.To study the impact of the introduced NS1 mutations on the suppression of PKR activation, we infected A549 cells with the recombinant viruses or mock treated the cells. Whole-cell lysates were prepared at 4, 8, 16, and 24 hpi and separated by SDS-PAGE, and expression of PKR and its phosphorylated active form (P-PKR), the viral nucleoprotein (NP), NS1, and β-actin was determined by immunoblotting (Fig. 4). We detected strong PKR phosphorylation in cells infected with the NS1(R35A) and ΔNS1 mutant viruses and an intermediate stimulation with NS1(R46A) virus. In contrast, only background phospho-PKR levels were induced by the WT and other mutant viruses, including NS1(R38A), which expresses an NS1 protein with strongly reduced dsRNA binding activity. Activation of PKR by the NS1(R35A) and NS1(R46A) mutant viruses was not due to a lack of NS1 accumulation (Fig. 4). Thus, amino acid positions R35 and R46 are crucial for inhibition of PKR.
Activation of PKR by recombinant viruses NS1(R35A), NS1(R46A), and ΔNS1. A549 cells were infected with the indicated recombinant mutant and WT viruses at an MOI of 1 or were mock treated for the indicated times. Cell lysates were subjected to SDS-PAGE and immunoblotting to detect phospho-PKR (P-PKR), PKR, NP, NS1, and β-actin.
PKR knockdown largely rescues growth impairment of NS1 mutant viruses R35A and R46A.Next, we analyzed the effect of the introduced mutations in the NS1 protein on viral replication and the role of PKR played therein. For this purpose, we employed a human A549 lung epithelial cell-based cell line with a stably integrated PKR-specific short hairpin RNA (shRNA) expression vector (A549 PKR KD) and control cells expressing a nontarget shRNA (A549 NT). The A549 NT and PKR KD cells were infected in parallel with the set of recombinant viruses at a multiplicity of infection (MOI) of 0.1, and growth curve analysis was done with standard plaque assays (Fig. 5A to D). The NS1(R35A) (gray line), NS1(R46A) (red line), and ΔNS1 (green line) viruses were strongly attenuated for replication on A549 NT cells, by approximately 2.5 (R46A), 4 (ΔNS1), and 5 (R35A) orders of magnitude, compared to the WT at 72 hpi (Fig. 5A). Significantly, the same viruses gained substantial replicative activity when grown on A549 PKR KD cells, with increases in viral titers at 24 hpi in the range of 5,000-fold [NS1(R35A)], 450-fold (ΔNS1), and 30-fold [NS1(R46A)] (Fig. 5B and E). In contrast, the NS1 mutations K41A, K126A/N127A, and Triple affected the kinetics and extent of viral propagation only mildly in control and PKR KD cells, suggesting that the corresponding positions are not critically involved in blocking PKR activation (Fig. 5C to E). Replication of the NS1(R38A) mutant virus was significantly affected by more than one order of magnitude late in infection in both PKR KD and control cells, indicating that attenuation caused by the R38A mutation was independent of PKR (Fig. 5C). Finally, replication of the NS1(I123A/M124A) mutant virus was delayed in control and PKR KD cells but approached levels of the WT virus at 72 hpi, also indicating a PKR-independent mode of attenuation (Fig. 5C to E). Knockdown of PKR expression was controlled by immunoblotting (Fig. 5F). Hence, we concluded that amino acids R35 and R46 of NS1 are essential for efficient viral replication in the presence of physiological levels of PKR.
Multicyclic growth of recombinant viruses on A549 NT and PKR KD cells. (A and B) Replication of recombinant viruses NS1(R35A), NS1(R46A), ΔNS1, and WT. Viruses were grouped based on an increase of replication on PKR KD cells of more or less than 30-fold (see panel E). (C and D) Replication of recombinant viruses NS1(R38A), NS1(K41A), NS1(I123A/M124A), NS1(K126A/N127A), NS1(Triple), and WT. A549 NT cells (A and C) or A549 PKR KD cells (B and D) were infected with viruses at an MOI of 0.1, and replication was observed for 72 h. Viral titers were assessed by plaque assay (n ≥ 4, mean ± SEM). Significances in panels A to D are given in comparison to WT replication in the assigned cells. *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001 (Mann-Whitney U test). (E) Fold increase of replication of viruses on PKR KD cells compared to NT cells. Significances are given for virus growth on PKR KD to compared to NT cells. *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001 (Mann-Whitney U test). (F) Immunoblot detection of PKR in A549 NT and PKR KD cells.
NS1-dependent suppression of PKR activation and inhibition of viral IFN-β induction can be segregated by mutation.Another primary function of the viral NS1 protein is the inhibition of type I IFN induction via RIG-I-dependent signaling, which has been associated with amino acid positions R38 and K41 (53, 54, 72). Accordingly, we asked how the distinct changes in the set of NS1 mutant viruses affected viral control of type I IFN induction in the presence of physiological or strongly reduced PKR expression levels. A549 control or PKR KD cells were infected with the recombinant viruses, and IFN-β levels were determined by enzyme-linked immunosorbent assay (ELISA) at 24 hpi (Fig. 6). The WT and the mutant viruses expressing NS1 with alanine exchanges in the ED (I123A/M124A and K126A/N127A) induced less than 105 IU/ml IFN-β in the A549 NT cell supernatants. As expected, cells infected with the mutant viruses NS1(R38A), NS1(K41A), and NS1(Triple) secreted unusually large amounts (>1,000 IU/ml) of IFN-β that exceeded the values for the ΔNS1 virus in A549 NT cells (Fig. 6) (53). Interestingly, the NS1(R35A) and NS1(R46A) mutant viruses stimulated comparatively weak IFN-β secretion not only in the controls but also in the A549 PKR KD cells (Fig. 6), although at least the NS1(R46A) virus replicated in these hosts to similar titers as NS1(R38A) and NS1(K41A) (Fig. 5B and D). Consequently, these findings suggest that the corresponding NS1(R35A) and NS1(R46A) viruses are able to control IFN-β induction to a certain extent even when viral replication is enhanced by PKR knockdown.
The NS1(R35A) and NS1(R46A) mutant viruses are weak IFN-β inducers. A549 NT (left) and PKR KD (right) cells were infected at an MOI of 1 with the indicated recombinant viruses or were mock treated for 24 h. Cell supernatants were collected and used for detection of active IFN-β by ELISA (n ≥ 2, mean + SEM). **, P ≤ 0.01 (Mann-Whitney U test, in comparison to WT).
PKR deficiency restores viral virulence and replication of NS1 mutant viruses R35A and R46A in vivo.Many aspects of PKR's antiviral activity, such as the translational shutoff, act on the level of the infected cell. However, PKR also has established roles in activating innate immune pathways directed against infected epithelial cells elicited by leukocytes such as mononuclear phagocytes (73). To assess if the effects of NS1 mutants on PKR inhibition and viral fitness validated in vitro were relevant as well for the course of disease in an infected animal in vivo, we examined replication and virulence of NS1 WT and mutant viruses in mice with normal or ablated PKR expression (74) (Fig. 7). If silencing of PKR was a major contribution of distinct NS1 amino acids to virulence in WT mice, we would expect attenuation of a corresponding mutant virus in such animals but an increase in morbidity and mortality in PKR-deficient mice. Intratracheal infection with 500 PFU WT virus lead to a drastic loss of body weight starting at day 3 and produced a lethality of >80% both in wild-type (PKR+/+) and PKR−/− mice (Fig. 7A to D). Interestingly, the NS1 mutant viruses R35A, R38A, K41A, and R46A did not cause any weight loss or lethality in PKR+/+ mice (Fig. 7A and C). However, only the NS1(R35A) and NS1(R46A) mutants were virulent in PKR-deficient mice, causing lethal infections similar to that caused by the WT virus (Fig. 7B and D). This phenotypic restoration strongly suggests that a main activity of the respective amino acid positions is to block the antiviral function of PKR. In contrast, no sign of pathogenicity was seen in PKR−/− mice infected with the NS1(R38A) or NS1(K41A) virus, suggesting that their mode of attenuation did not depend on PKR but might rather involve their strong IFN-β-inducing phenotype (Fig. 6). Lethal infections correlated with strong virus replication as determined by quantification of viral titers in lung bronchoalveolar lavage fluids of infected PKR+/+ and PKR−/− mice (Fig. 7E and F). Correspondingly, histological sections of PKR+/+ and PKR−/− mice infected with WT virus and PKR-deficient mice infected with NS1 mutant viruses R35A and R46A, but not with R38A and K41A viruses, revealed an increased alveolar inflammatory response at day 7 after infection (Fig. 7G and data not shown). Thus, our data lead to the conclusion that amino acids R35 and R46 in the viral NS1 protein are essential not only to prevent activation of PKR but also to enable efficient viral replication and to cause lethal viral infections.
Virulence of recombinant influenza A viruses in wild-type (PKR+/+) and PKR−/− mice. (A to D) Groups of seven [for NS1(K41A)] or eight (for WT, R35A, R38A, and R46A viruses) PKR+/+ (A and C) or PKR−/− (B and D) mice were infected intratracheally with ∼5 × 102 PFU of the WT and indicated mutant NS1 viruses (R35A, R38A, K41A, or R46A). Survival of mice (A and B) and changes in body weight (C and D) were recorded for 21 days. Mice were euthanized according to morbidity scores, including clinical parameters such as opaque eyes/enophthalmus, ruffled fur, hypothermia, reduced activity, or laborious breathing (each given 5 to 10 points according to severity), summed to a value of 20 or when weight loss within 2 days exceeded 20%. (E and F) Groups of three to five PKR+/+ (E) and PKR−/− mice (F) were infected with the indicated viruses for 3, 5, and 7 days before bronchoalveolar lavage fluid was obtained and used to determine viral titers by plaque titration (#, no virus detected). (G) Representative paraffin sections were obtained from PKR+/+ and PKR−/− mice infected with the indicated viruses or a PBS control (mock) at day 7 after infection and stained with hematoxylin and eosin. Graphs represent means ± SEM. *, P < 0.05; **, P < 0.01, ***, P < 0.005.
DISCUSSION
Neutralization of PKR is of paramount importance for many viruses (19, 75). The findings presented in this study provided novel insights into the mode of inhibition by influenza A virus. We previously showed that the N-terminal RBD of the NS1 protein of influenza B virus (B/NS1), which adopts a structure similar to that of its type A counterpart (45), inhibits PKR most likely via an RNA-mediated bridging mechanism, as three dsRNA binding-deficient NS1 mutants failed to block PKR (32). This mechanism differed fundamentally from the previously suggested mode of action for the IAV NS1 protein involving a direct protein interaction mediated by amino acids in the ED (68, 69). This discrepancy prompted us to systematically compare the relative contributions of amino acids previously implicated in IAV NS1 RNA binding (R35, R38, K41, and R46), dimer formation (R35 and R46), and PKR binding (I123/M124 and K126/N127) in blocking activation of the kinase. Additional recombinant viruses with mutations in the NS1 RBD (K20A/R21A, K41A/S42G, R44A, R59A/K62A, and R67A/K70A) did not show any phenotype in suppressing PKR activation or supporting efficient replication (data not shown). Our results indicated that NS1 in fact does not require RNA binding activity to associate with and silence PKR, as the R38A mutant, a potent suppressor of PKR activation, was not captured by immobilized synthetic dsRNA, which is in agreement with previous reports (36, 68, 69).
Among the tested mutants, we found the most striking phenotype for alanine exchanges at NS1 positions R35 and R46, which strongly reduced PKR inhibition of the corresponding viruses in human cells. These two amino acids form multiple intermolecular salt bridges and hydrogen bonds that stabilize the dimeric antiparallel α-helices of the NS1 RBD and contribute to dsRNA binding (38, 43, 45). We confirmed that single alanine exchanges of those residues eliminated association with dsRNA, but we documented in coimmunoprecipitation and homo-FRET analyses that NS1 mutants R35A and R46A dimerize in cells. The latter finding was unexpected, as those mutations had previously been shown to eliminate dimer formation (36). However, that analysis investigated a recombinant purified NS1 RBD in vitro and did not take into account contributions of the C-terminal domain to NS1 dimer formation (35, 49, 52, 76, 77). Moreover, Lalime and Pekosz interpreted a weak bimolecular fluorescence complementation of the NS1(R35A) mutation as a complete loss of dimerization, but this signal was above the background level and may have reflected residual activity (46). Together with the comparatively efficient suppression of IFN-β secretion by the NS1(R35A) and NS1(R46A) mutant viruses, these data suggest that the corresponding mutations do not fundamentally alter the normal structure of the RBD. The strong PKR activation phenotype correlated directly with a loss of complex formation with the kinase that was not observed for any of the other NS1 mutants (R38A, K41A, I123A/M124A, K126A/N127A, and Triple). Collectively, these data suggest a mechanism in which inhibition of PKR is mediated by a binding interaction with NS1 requiring amino acids R35 and R46.
The high relevance of the NS1 R35 and R46 positions for virus propagation became apparent by the significant and strong attenuation of the corresponding mutant viruses on human cells, with a reduction in final titers of 5 and 2.5 orders of magnitude, respectively. Significantly, our data demonstrate a causal relationship of this attenuation with PKR activity, as viral growth was strongly enhanced in corresponding PKR knockdown cells (up to 5,000-fold), although the rescue was not complete. The comparatively stronger attenuation of the NS1(R35A) virus toward R46A may be explained by a more complex role of arginine 35 within the dimeric structure of the RBD involving monomer-monomer interactions, direct contact with dsRNA, and its contribution to nuclear targeting (46). The NS1(R35A) and (R46A) mutant viruses were still partially restricted in PKR knockdown cells, most likely because their dsRNA binding-defective NS1 proteins were unable to block other antiviral dsRNA-dependent mechanisms such as activation of the 2′,5′-OAS (55). This interpretation is supported by our observation that the NS1(R38A) mutant virus was attenuated on cells with either physiological or reduced PKR levels. Two other groups previously attempted to characterize influenza A/WSN/33-derived mutants expressing an NS1 R35A mutant protein and confirmed the importance of this position for viral propagation. Lalime and Pekosz recovered only revertant viruses from transfected 293T cell cultures encoding amino acids changes at position D39 in addition to R35A (46). In contrast, Mok and colleagues observed strong attenuation of a rescued NS1 R35A mutant virus on human cells but did not define the underlying reasons (78).
Most NS1 functions can be recapitulated in vitro, such as activation of PKR and inhibition of IFN-β. However, experiments in cell culture cannot mimic the more complex interplay between the virus and the host's cellular and humoral immune responses. Therefore, we performed experiments in PKR+/+ and PKR−/− mice to reveal the effects of the introduced mutations in NS1 on viral replication and disease progression in vivo. As expected, mutant viruses NS1(R35A) and NS1(R46A) presented a strong defect in replication and virulence in wild-type mice, which was rescued completely in mice with PKR deficiency. This further underlines the importance for NS1 to inhibit PKR as a major antagonist in infection in vivo and provides strong evidence for the pivotal roles of the NS1 R35-R46 pair in this blockade. Moreover, PKR−/− mice infected with the mutant viruses presented a strong increase in infiltration of lymphocytes compared to that in WT virus infection, suggesting an increase in pathogenicity. Interestingly, the NS1(R38A) and (K41A) mutant viruses behaved as benignly in wild-type mice as the NS1(R35A) and NS1(R46A) mutants, but this phenotype was not rescued by PKR knockout. This effect contrasts with our in vitro data for these mutants, which were only slightly attenuated for replication in human cells. We suggest that the strong attenuation of the NS1(R38A) and NS1(K41A) mutants in vivo, which was independent of the PKR expression status, was associated with or even caused by a vigorous IFN-β response, as observed on human cells, which would upregulate additional antiviral ISGs such as 2′,5′-OAS, plasminogen activator inhibitor 1 (79), and ISG15 (80). These observations are in full agreement with earlier findings presenting the importance of positions R38 and K41 in NS1 of the PR8 strain in inhibiting IFN-β expression in vitro and supporting viral replication in vivo (53).
In contrast to the severe phenotypes of the R35A and R46A mutants, we did not identify a similar impact of I123A/M124A or K126A/N127A mutations, although those positions were previously found to be essential for the NS1 of an H3N2 subtype virus (Udorn/72) to bind to and/or prevent activation of PKR (68, 69). How can this apparent discrepancy be possibly explained? We cannot exclude that the differing sequence requirements of the PR8 and Udorn NS1 homologs for PKR inhibition reflect different modes of action. However, recent knowledge of the structural heterogeneity and plasticity of NS1 proteins from different strains may provide some clues for an alternative interpretation (49–51). The conserved N-terminal RBD forms a stable dimeric structure, whereas the attached ED can engage in dimeric interactions with alternative interface structures to other NS1 dimers, resulting in the formation of different higher-order structures (reviewed in reference 51). An important structural element in ED dimer formation is a long α-helix (positions 170 to 188) including the highly conserved tryptophan 187 (49, 52). This residue can be either exposed or buried, indicating alternative orientations of the ED that have been dubbed “open,” “semiopen,” or “closed” and may be found in the same cell depending on the specific context and protein concentration, allowing NS1 to engage in various protein-protein and RNA-protein interactions (50). Interestingly, it has been shown that the ED also contributes to RNA binding of the full-length NS1 protein, since mutation of tryptophan 187 strongly reduced dsRNA binding, possibly by disturbing cooperative nucleic acid binding (49, 52). Although positions 123 to 127 and tryptophan 187 are conserved, the EDs of the Udorn/72 and PR8 NS1 homologs differ by 12 amino acids and in their dimerization affinities (52) and were reported to have slightly different spatial structures (77). Therefore, it is tempting to speculate that mutations of positions 123/124 or 126/127 within the two NS1 proteins could strongly affect the activity of the RBD to associate with PKR in case of the Udorn strain but that this is less pronounced for PR8. The Udorn NS1 protein also differs from the PR8 homolog in that its ED binds to CPSF30 with involvement of amino acid positions 121 and 125 (81, 82), which may further directly or indirectly influence its interplay with PKR. A testable prediction of this model is that mutations of R35 and R46 within the NS1 protein of Udorn/72 and possibly other strains will cause a phenotype on the control of PKR activation similar to that described here for the PR8 strain.
In conclusion, this study identified the highly conserved amino acids R35 and R46 within the RBD of NS1 as being important for interaction with as well as inhibition of PKR and could show that the observed effects are independent of RNA binding and IFN-β control. These data not only define the NS1 N-terminal domain as a highly critical element to neutralize PKR's antiviral activity but also demonstrate that this blockade is an essential and major contribution of NS1 to the viral life cycle.
MATERIALS AND METHODS
Cells, viruses, and virus infection.Human A549 and HEK293T cells were grown in Dulbecco's modified Eagle medium supplemented with 10% fetal bovine serum and 2 mM l-glutamine (complete DMEM). Constructions of the A549 PKR KD and A549 NT cell lines, which contain a stable integrate of the PKR-specific shRNA (PKR KD) or nontarget shRNA (NT) expression cassette, have been described elsewhere (83). Madin-Darby canine kidney type II (MDCKII) cells were grown in minimal essential medium supplemented with 10% fetal bovine serum and 2 mM l-glutamine (complete MEM). All cells were maintained at 37°C and 5% CO2. Influenza A virus stocks were grown in the allantoic cavities of 10-day-old (A/Puerto Rico/8/34 wild type) or 7-day-old (A/Puerto Rico/8/34-ΔNS1 [84] and recombinant NS1 mutant viruses) embryonated chicken eggs for 2 days at 37°C. To analyze viral replication, A549 PKR KD and A549 NT cells were infected at the indicated multiplicity of infection (MOI) and incubated for 3 days at 37°C in DMEM containing 0.2% bovine albumin and 2 mM l-glutamine (inf-DMEM) supplemented with 1 μg/ml puromycin and 0.25 μg/ml tosylsulfonyl phenylalanyl chloromethyl ketone (TPCK)-treated trypsin. Virus titers were determined on MDCKII cells by standard plaque assay.
Construction of plasmids.The plasmids pPolI-PR8-NS-R35A, pPolI-PR8-NS-R38A, pPolI-PR8-NS-K41A, pPolI-PR8-NS-R46A, pPolI-PR8-NS-I123A/M124A, pPolI-PR8-NS-K126A/N127A, and pPolI-PR8-NS-R38A/I123A/M124A are derivatives of pPolI-PR8-NS (85) and were constructed by site-directed mutagenesis using the QuikChange II site-directed mutagenesis kit (Stratagene). All targeted amino acid codons encoded alanine at the indicated positions. The plasmids pHW2000-PR8-NS-R35A, pHW2000-PR8-NS-R38A, pHW2000-PR8-NS-K41A, pHW2000-PR8-NS-R46A, pHW2000-PR8-NS-I123A/M124A, pHW2000-PR8-NS-K126A/N127A, pHW2000-PR8-NS-R38A/I123A/M124A, and pHW2000-PR8-NS-WT were generated by transferring NS cDNA of the corresponding pPolI-PR8-NS plasmids into pHW2000 (86). Plasmid pcDNA3-NS1-M4 is a derivative of pcDNA3-NS1 (87) that was engineered to encode a PR8-NS1 protein lacking amino acids 19 to 38 in the RBD and a W187A point mutation. The integrity of the expected sequences was verified by DNA cycle sequencing using an ABI Prism 3100 genetic analyzer (Applied Biosystems). For generation of pmTurquoise-NS1 constructs, NS1 cDNAs were amplified from the corresponding pPolI-PR8-NS plasmids using NS-specific primers and integrated between the EcoRI and KpnI restriction sites of pmTurquoise-C1 (88), which disturbs nuclear export protein (NEP) expression. peYFP constructs were produced by exchanging the mTurquoise fluorophore with eYFP using its NheI and SacI restriction sites.
Transfection-mediated recovery of recombinant influenza A viruses.An established plasmid-based rescue system was employed to generate recombinant influenza A/PR/8/34-derived viruses with mutated NS1 genes (89). Briefly, 3 × 106 HEK293T cells were transfected in suspension with the plasmids pCAGGS-PR8-PB1, pCAGGS-PR8-PB2, pCAGGS-PR8-PA, pCAGGS-PR8-NP, pPolI-PR8-PB1, pPolI-PR8-PB2, pPolI-PR8-PA, pPolI-PR8-NP, pPolI-PR8-M, pPolI-PR8-HA, and pPolI-PR8-NA and a pPolI-PR8-NS construct (0.5 μg each) using Lipofectamine 2000 (Invitrogen). The medium was replaced with complete DMEM supplemented with 1 μg/ml TPCK-treated trypsin at 4 to 6 h posttransfection. After 48 h, the supernatants of transfected cells were used to inoculate 7-day-old chicken eggs. Recombinant viruses were passaged a second time in 7-day-old chicken eggs for generation of virus stocks. To confirm the rescue of recombinant viruses, viral RNA was extracted from allantoic fluid using the QIAampMinElute virus spin kit (Qiagen) followed by reverse transcription-PCR (RT-PCR) amplification of the NS segment using NS-specific primers with the OneStep RT-PCR kit (Qiagen). The cDNA was purified using the QIAquick PCR purification kit (Qiagen) and used for DNA cycle sequencing on an ABI Prism 3100 genetic analyzer (Applied Biosystems).
PKR binding assay.Assay conditions to assess NS1 binding to PKR have been described elsewhere (64). Briefly, GST-PKR(K296R) and GST-PKR(aa266-551) fusion proteins were expressed in Escherichia coli BL26 from pGEX-PKR(K296R) and pGEX-PKR(aa266-551) and adsorbed to glutathione-Sepharose 4B beads (GE Healthcare). The catalytically inactive PKR(K296R) variant was used due to its more stable expression compared to that of PKR WT (90). NS1 WT and mutant proteins were translated and labeled with [35S]methionine in vitro using pHW2000-derived constructs together with the TNT-coupled reticulocyte lysate system (Promega). Loaded glutathione beads were rotated with translated NS1 proteins in GST binding buffer (20 mM Tris-HCl [pH 7.5], 50 mM NaCl, 0.8% Triton X-100, 20% glycerol, 1 mM Pefabloc, and 2 mM sodium vanadate) at 4°C for 2 h. The beads were washed three times with GST binding buffer, and precipitated proteins were dissolved in SDS sample buffer, separated by SDS-PAGE, and stained using Coomassie blue (Roth). The gels were dried, and NS1 proteins were detected by autoradiography. Quantification of bands was performed using the LabImage1D software (Kapelan Bio-Imaging GmbH, Leipzig, Germany).
Analysis of NS1 dimerization, dsRNA binding, and immunoblotting.NS1 dimerization was analyzed by a coimmunoprecipitation assay using transfected eYFP-NS1 proteins expressed in virus-infected cells or cotransfected with untagged NS1 proteins. To test NS1 dimerization in infected cells, HEK293T cells were transfected in suspension with 2.5 μg of peYFP-NS1 WT plasmid using Lipofectamine 2000 (Invitrogen). At 24 h posttransfection, cells were infected with a corresponding recombinant virus at an MOI of 5 and incubated for 16 h in inf-DMEM at 37°C and 5% CO2. Cells were collected and lysed in GFP-Trap lysis buffer (10 mM Tris-HCl [pH 7.5], 150 mM NaCl, 0.5 mM EDTA, 0.5% NP-40, 1 mM Pefabloc, and 2 mM sodium vanadate) and further processed following the manufacturer's protocol for the GFP-Trap antibody matrix (Chromotek, Munich, Germany). Precipitated proteins were dissolved in SDS sample buffer, separated by SDS-PAGE, and detected by immunoblotting. NS1 dimerization was analyzed in lysates of HEK293T cells transfected in suspension with peYFP-NS1-WT plasmid together with either pcDNA3-NS1-WT, pcDNA3-NS1-M4, the pHW2000-NS derivative pHW2000-NS-R35A or pHW2000-NS-R46A, or empty vector, using Lipofectamine 2000 (Invitrogen). At 24 h posttransfection cells were collected and lysed in GFP-Trap lysis buffer and processed as described above.
dsRNA binding of NS1 proteins was determined by using extracts of 293T cells transfected with the indicated pHW2000 plasmids. Cells were lysed in poly(I·C) lysis buffer (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 1 mM EDTA, 1% NP-40, 1 mM Pefabloc, and 2 mM Na vanadate) for 30 min on ice. Cleared lysates were rotated overnight at 4°C with poly(I·C)-coupled beads. For generation of poly(I·C)-coupled beads, poly-cytidylic acid-coupled agarose (Sigma-Aldrich) was incubated with an equal amount of poly-inosinic acid 5′-potassium (2 mg/ml; Sigma-Aldrich) in resuspension buffer (50 mM Tris-HCl, 50 mM NaCl, pH 7.5) overnight at 4°C. The lysate-bead mixture was washed once in washing buffer (50 mM Tris-HCl, 150 mM NaCl, pH 7.5), followed by five washes with poly(I·C) lysis buffer. Precipitated proteins were dissolved in SDS sample buffer, separated by SDS-PAGE, and detected by immunoblotting using an enhanced chemiluminescence protocol (SuperSignal West Dura; Thermo Scientific). Primary antibodies used to detect viral and cellular proteins included rabbit anti-phospho-PKR (pThr446) (Epitomics), rabbit anti-PKR (Epitomics), mouse anti-A/NP (Serotec), mouse anti-β-actin (Santa Cruz), and rabbit anti-A/NS1 (91).
FLIM for FRET.HEK293T cells were seeded 1 day before transfection in a four-chamber glass-bottom dish (Greiner Bio-One). Cells were transfected with a total amount of 2 μg of plasmid DNA and 0.5 μg of poly(I·C)-rhodamine (InvivoGen, Toulouse, France) using TurboFect (Life Technologies, Darmstadt, Germany) following the manufacturer's protocol. Cells were incubated for 16 h at 37°C and 5% CO2 prior to microscopy. Fluorescence lifetime imaging microscopy (FLIM) measurements were performed with an inverted FluoView 1000 microscope (Olympus, Tokyo, Japan) modified with an LSM upgrade kit (PicoQuant, Berlin, Germany). For imaging, a 60× water immersion objective (numerical aperture, 1.2) was used. Images of a size of 512 by 512 pixels were obtained with a scanning speed of 4 μs/pixel. FLIM measurements were generated using a pulsed laser diode with a 20-MHz repetition frequency in combination with a τ single-photon avalanche photodiode (τ-SPAD) (PicoQuant). Fluorophores were excited and detected with the following combinations: mTurquoise, 405-nm laser excitation and 470/30-nm band-pass filter; eGFP/eYFP, 483-nm laser excitation and 540/40-nm band-pass filter. SPAD signals were processes with a TimeHarp 300 (PicoQuant) photon counting board. Data were analyzed using the SymPhoTime 64 bit (PicoQuant) software. Taking the instrument response function into account allows the analysis of short lifetime components. FLIM images were recorded for 63 s with an average count rate of 0.5 × 105 to 1 × 105 counts per second for measurements in the absence of poly(I·C)-rhodamine and 0.5 × 105 to 5 × 104 counts per second in the presence of poly(I·C)-rhodamine. The excitation power in the presence of poly(I·C)-rhodamine had to be reduced due to the presence of very bright spots in the sample. Photon arrival time histograms for the selected region of interest (ROI) within the FLIM images were fitted by a nonlinear least-squares iterative procedure as the sum of two exponential terms. Decays were judged by the χ2 values and the distribution of the residuals. For every single cell as well as for the region of interest in which all three components [PKR, NS1 and poly(I·C)] are located, the amplitude weighted average lifetime was calculated and used to calculate the Förster resonance energy transfer (FRET) efficiency: EFRET = 1 − (τDA/τD), where EFRET is the relative FRET efficiency and τD and τDA represent the lifetime of the donor in the absence or presence of an acceptor, respectively.
Fluorescence anisotropy imaging.Details for the expanded setup used for polarization measurements are given on the manufacturer's website (92). The anisotropy value was calculated as the ratio between parallel and perpendicular detected fluorescence light and reflects the oligomerization state of a fluorophore due to homo-FRET (FRET between two identical fluorophores) (93). Thus, a higher anisotropy value is indicative of a monomeric state, whereas a lower value signals oligomerization or dimerization of the fluorophore. Images were obtained under the same conditions as described above for FLIM-FRET imaging, but with the differential interference contrast (DIC) depolarization filter being removed. A polarization beam splitter was used to split the polarized emitted light. Parallel and perpendicular fluorescence signals were detected using a 540/30-nm emission filter for the perpendicular polarized light and a 540/40-nm emission filter for the parallel polarized light prior to a τ-SPAD and PerkinElmer SPAD. For this microscope setup, the G factor was calculated from point measurements in an Alexa 488 (Life Technologies) solution. A value of 0.81 was found for the setup described above. The G factor could be approximated by the ratio between parallel and perpendicular fluorescence intensity, since the rotational diffusion time of this dye is significantly higher than its average fluorescence lifetime. Images were analyzed using the SymPhoTime 64 bit software after selection of the ROI, and obtained pixel intensity weighted values were summed into anisotropy histograms.
Detection of IFN-β by ELISA.A549 cell lines were infected with the indicated viruses or were mock treated and incubated with inf-DMEM supplemented with 1 μg/ml puromycin at 37°C and 5% CO2 for the indicated times. Supernatants were collected, cleared of cell debris, and immediately frozen at −80°C. Active IFN-β was assessed using an IFN-β enzyme-linked immunosorbent assay (ELISA) (Fujirebio) following the manufacturer's protocol.
In vivo treatment protocols.Wild-type C57BL/6 mice (PKR+/+) were purchased from Charles River Laboratories. PKR-deficient mice (PKR−/−) were a kind gift from C. Reis e Sousa (The Francis Crick Institute, London, UK). Mice were housed under pathogen-free conditions at the University of Giessen in accordance with federal and university guidelines. For infection, 12- to 16-week-old PKR+/+ and PKR−/− mice were anesthetized (100 mg/kg ketamine, 16 mg/kg xylazine, premedication 0.05 mg/kg atropine) and inoculated intratracheally with 70 μl of sterile phosphate-buffered saline (PBS) containing 500 PFU of the indicated viruses (WT, R35A, R38A, K41A, and R46A). Infected mice were monitored 1 to 3 times daily and euthanized when total morbidity scores, including clinical parameters such as opaque eyes/enophthalmus, ruffled fur, hypothermia, reduced activity, or laborious breathing (each scored as 5 to 10 points according to severity), summed to a value of 20 or when weight loss within 2 days exceeded 20%. For viral lung titration, mice were sacrificed on days 3, 5, and 7, and titers were determined from bronchoalveolar lavage samples by plaque titration on MDCKII cells as described previously (73, 94). Histological sections were obtained at days 3, 5, and 7 from paraformaldehyde-perfused and -fixed and paraffin-embedded (Leica ASP200S) lungs. Paraffin lung sections were stained by hematoxylin and eosin. In brief, slides were immersed in hematoxylin solution (5%) for 30 s, rinsed with water for 1 min, and immersed in 1% eosin Y solution for 30 s. After dehydrating the sections in 70%, 96%, and 100% ethanol for 30 s each and extracting ethanol with xylene for 2 min, samples were mounted and visualized with an Evos FL Auto microscope.
Ethics statement.All animal experiments were conducted according to the legal regulations of the German Animal Welfare Act (Tierschutzgesetz) and were approved by the regional authorities of the State of Hesse (Regierungspräsidium Giessen; reference number 67-2013).
ACKNOWLEDGMENTS
We thank Gudrun Heins for excellent technical support, Peter Palese for provision of the recombinant plasmid system for PR8-derived viruses, and Caetano Reis e Sousa for the gift of PKR−/− mice.
This study was supported by grants from the German Research Foundation (TransRegio 84, project B2, to T.W., S.H., and S.P.; WO554/4-1 to T.W.; SFB1021 project C05 to S.H.; and KFO309 P2/P8 to C.P. and S.H.), the German Center for Lung Research (DZL) (to S.H.), the German Center for Infection Research (DZIF) (to S.H.), and the BMBF (project ViroSign to T.W. and A.H.). K.L.S. and F.J. acknowledge funding and support from the ZIBI graduate school (GRK1121, IMPRS-IDI).
The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.
FOOTNOTES
- Received 3 February 2017.
- Accepted 20 February 2017.
- Accepted manuscript posted online 1 March 2017.
- Address correspondence to Thorsten Wolff, wolfft{at}rki.de.
↵* Present address: Kristina L. Schierhorn, Department of Infectious Diseases, King's College, Guy's Hospital, London, UK; Julia Dzieciolowski, Department of Biochemistry and Molecular Biology, Nutritional Science, Justus Liebig University Giessen, Giessen, Germany.
Citation Schierhorn KL, Jolmes F, Bespalowa J, Saenger S, Peteranderl C, Dzieciolowski J, Mielka M, Budt M, Pleschka S, Herrmann A, Herold S, Wolff T. 2017. Influenza A virus virulence depends on two amino acids in the N-terminal domain of its NS1 protein to facilitate inhibition of the RNA-dependent protein kinase PKR. J Virol 91:e00198-17. https://doi.org/10.1128/JVI.00198-17.
REFERENCES
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.↵
- 85.↵
- 86.↵
- 87.↵
- 88.↵
- 89.↵
- 90.↵
- 91.↵
- 92.↵
- 93.↵
- 94.↵
- Copyright © 2017 American Society for Microbiology.
