Grail attenuates influenza A virus infection and pathogenesis by inhibiting viral nucleoprotein

Grail is a well-characterized mediator of metabolic disease, tumour progression, and immune response. However, its role in influenza A virus (IAV) infection remains poorly understood. In this study, we demonstrated that Grail knockdown potentiates IAV infection, whereas Grail overexpression blocks IAV replication. The intranasal administration of IAV to Grail KO mice led to a lower survival rate than in similarly infected wild-type mice. Additionally, IAV-infected Grail KO mice had higher viral titres, greater immune cell infiltration, and increased expression of inflammatory cytokines in the lungs. Mechanistically, we showed that Grail interacts with viral nucleoprotein (NP), targeting it for degradation and inhibiting IAV replication. NP expression was increased in Grail knockdown cells and reduced in cells overexpressing Grail. Collectively, our results demonstrate that Grail acts as a negative regulator of IAV infection and replication by degrading viral NP. These data increase our understanding of the host antiviral response to infection with IAV.

Influenza A virus (IAV) is a highly feared pathogen that poses a significant threat to public health and holds the potential for worldwide outbreaks (pandemics). IAV belongs to the Orthomyxoviridae family of RNA viruses and contains a negative-sense, single-stranded RNA genome of 8 segments that encode 16 viral proteins on eight segments 1 . Replication of the IAV genome requires viral ribonucleoprotein (vRNP), which consists of the heterotrimeric, RNA-dependent RNA polymerase complex (PB1, PB2, and PA), oligomeric nucleoproteins (NPs), and viral RNA (vRNA). IAV vRNP plays a vital role in viral mRNA synthesis in the early stage of infection and genomic vRNA production later in infection 2 . Recent genome-wide RNAi screens and supporting experimental evidence have suggested that cellular host proteins interact with IAV at every stage of the viral life cycle [3][4][5][6] . Moreover, a number of host factors and cellular processes have been identified as potential regulators of vRNP function and thus may be implicated in host adaptation and viral pathogenicity [7][8][9][10][11][12][13][14][15] .
Increasing evidence has suggested that the host ubiquitin-proteasome system regulates key stages of the IAV life cycle. E3 ubiquitin ligases, of which there are over 600 types in human, predominantly control substrate specificity during ubiquitination 16 . Two of these, Itch and Nedd4, are involved in IAV uncoating from the endosome 6,17 , where the ubiquitinated viral capsids then activate and exploit the aggresome for efficient uncoating 18 . Several members of the tripartite motif (TRIM) superfamily have been reported to contribute to the host anti-influenza response by mediating proteasomal degradation of viral proteins; for example, TRIM22 and TRIM41 have been shown to ubiquitinate viral NP 19,20 , while TRIM32 conjugates polyubiquitin at the polymerase basic protein 1 protein 21 . However, the E3 ligase Ccr4-Not transcription complex subunit 4 (CNOT4) is another key host-derived mediator of NP ubiquitination that positively regulates viral RNA replication and does not lead to proteasomal degradation of NP 22 .
Gene related to anergy in lymphocytes (Grail) encodes a transmembrane protein that is involved in the expression of cytokines related to T cell activation. The deletion of Grail in mice leads to reduced T cell responsiveness under TCR stimulation 23,24 . Additionally, the mammalian target of rapamycin pathway has been shown to mediate the cell cycle progression and proliferation of naïve T cells through the regulation of Grail expression 25  evidence implies that Grail may play an important role in cell cycle arrest and proliferation. Our previous work also shows that Grail can regulate p53-mediated cell cycle arrest and apoptosis in response to DNA damage 26 .
There is also evidence that Grail plays a crucial role in adipogenesis and diet induced obesity 27 . The dysregulation of Grail has been linked to ulcerative colitis, a chronic inflammatory disease of the gastrointestinal tract 28 . Recent work also indicates that Grail can regulate host innate immune responses to vesicular stomatitis virus (VSV) and herpes simplex virus type 1 (HSV-1) infections 29 . However, despite this body of research, the involvement of Grail during IAV infection and replication has not yet been well characterized. In the present study, we demonstrated that Grail aids in the control of IAV replication and infection by enhancing the degradation of viral NP in vitro and in vivo. Given this, we suggest that Grail is a potential drug target for the further prevention and treatment of IAV infections.

Results
Grail expression is upregulated after IAV infection in vitro and in vivo. Grail holds multiple biological roles in cell growth, adipogenesis, and immunity. However, the effect of Grail on the regulation of IAV infection is unclear. To address this, we first determined the expression level of Grail during IAV infection. Grail expression was shown to be induced in A549 cells infected with virus at an MOI of 1 ( Fig. 1A and Supplementary  Fig. 1) as well as in low titre infections using an MOI of 0.01 (Fig. 1B). Furthermore, mouse data showed that Grail expression was significantly higher in the lung tissue of animals infected with 1000 PFU of WSN than in control animals (Fig. 1C). These results show that IAV infection induces Grail expression.
Grail knockout (KO) mice succumb to IAV infection. Recently, it was shown that Grail improved innate antiviral immune responses to VSV (an RNA virus) and HSV-1 (a DNA virus) by enhancing TANK-binding kinase 1 (TBK1) activity 29 . To determine whether this antiviral property of Grail was also active against IAV infection in vivo, we employed mice carrying a deletion of the Grail gene. Wild-type (WT) and Grail KO mice were intranasally infected with 1000 PFU of mouse-adapted influenza A/WSN/33 H1N1 (WSN) virus. Both groups lost weight from day 2; however, Grail KO mice revealed a rapid period of weight loss at 3 dpi when compared to WT mice ( Fig. 2A). In WT mice, WSN infection resulted in 66.6% mortality, while all Grail KO mice died by day 7 post-infection (Fig. 2B). The viral load in the lungs at 3 dpi, as determined by plaque assay, was 2 logs higher in the Grail KO mice than in the WT mice (Fig. 2C). Furthermore, we used the substrate-free IAV-iRFP reporter virus and strategy described by Fukuyama et al. 30 to visualize the viral dynamics in vivo. We observed that iRFP fluorescence was detectable in mice infected with IAV-iRFP from 1 to 6 dpi. Grail KO mice demonstrated a wider distribution of this signal in the infected lung tissues than did WT mice (Fig. 2D). The iRFP signal was generally consistent with the viral load in the lungs of WSN-infected mice at 3 dpi (Fig. 2C,D). These in vivo data suggest that Grail may contribute to the regulation of influenza infection and replication, as these processes appear bolstered by the absence of functional Grail.
Grail KO mice show more severe lung damage. To further assess the causes of the increased susceptibility of Grail KO mice to WSN infection, histopathological examination of lung tissues collected at 3 and 6 dpi was performed. This revealed that both strains of mice developed viral pneumonia (Fig. 3B,D,H,J), in contrast to sham-infected mice (Fig. 3F,L). However, Grail KO mice had higher levels of multifocal acute alveolitis, with inflammatory cells infiltrating throughout the tissue, intra-alveolar oedema, red blood cell extravasation, epithelial necrosis (Fig. 3H,J), and higher histology scores than WT mice. Immunohistochemistry further revealed differing levels of IAV antigen expression near the focal inflammation sites in lung sections ( Fig. 3A,C,G,I). WSN-infected KO mice had higher levels and more widespread expression of viral antigen in the alveolar epithelium at 3 dpi (Fig. 3G). In contrast, WT mice displayed viral antigen expression most prominently in the bronchioles and had more limited expression in the alveolar epithelium ( Fig. 3A).

IAV infection induces increased inflammatory cytokine expression in Grail KO mice. To assess
whether pulmonary responses to IAV infection in Grail KO mice differ from those of WT mice, we also determined the levels of inflammatory cytokines in the lung tissue of both types of infected mice. The mRNA and protein levels of TNF-α, IL-6, and IL-1β in the lung tissue of IAV-infected Grail KO mice were found to be significantly higher than those in similarly infected WT mice at 3 and 6 dpi (Fig. 4). Importantly, the expression of these cytokines did not differ between uninfected WT and Grail KO mice (Fig. 4).
Grail suppresses viral replication. The data described thus far indicated that Grail may significantly enhance IAV-induced lung injury and lethality in mice. This prompted us to further investigate the functional relevance of Grail to IAV replication in human cells. We evaluated the growth kinetics of both WSN (an H1N1 influenza A virus) and 3446 (an H3N2 influenza A virus) in A549 cells under the conditions of either Grail overexpression or knockdown. We used shRNA to downregulate endogenous Grail and the retroviral transduction system for its stable overexpression, and examined the silencing or overexpressing efficacy ( Supplementary  Fig. 3). Upon downregulation of endogenous Grail, infected A549 cells had a significantly higher WSN titre at 24 hpi. Conversely, Grail overexpression led to a significantly reduced WSN titre at 36 hpi (Fig. 5A). While the growth curves of the 3446 virus showed no significant influence of Grail silencing in infected cells, the viral yield was significantly decreased at 36 hpi in A549 cells that overexpressed Grail, similar to the effect seen in WSN or pandemic/09 H1N1-infected cells ( Fig. 5B and Supplementary Fig. 2). Furthermore, the expression of IAV NP protein correlated with the viral growth curves in the instances of both Grail knockdown or overexpression in cells infected with WSN at an MOI of 0.01 for 24 h, similar to the effect seen in 3446 virus-infected cells (Fig. 5C and Supplementary Fig. 4). These results clearly demonstrate that viral replication is suppressed by Grail. Grail regulates NP protein levels. Grail is an E3 ligase that functions by interacting with and degrading its targets. To elucidate the possible mechanism of Grail-mediated repression of IAV replication, we analysed whether Grail interacts with viral proteins (NP, PA, PB1, PB2, HA, NA, M1, M2 and NS1) and targets them for degradation. Lysates from IAV-infected A549 cells were treated with anti-Grail antibody and IP proceeded with NP, PA, PB1, PB2, HA, NA, M1, M2 or NS1 antibody. As shown in Fig. 6A, Grail interacted with NP but not PA, PB1, PB2, HA, NA, M1, M2 or NS1. In order to determine whether Grail mediates the ubiquitination of NP, we co-transfected Flag-NP and Grail into HEK293 cells and performed further IP experiments. These data showed that NP ubiquitination was significantly increased in the presence of Grail when compared with controls (Fig. 6B). To identify whether Grail targets NP for proteasome-mediated degradation, we transfected NP plasmids into A549/Vector and A549/Grail cells with or without MG132 treatment. In these cells, Grail overexpression reduced the NP protein level when compared to controls. However, NP expression was rescued in the presence of MG132 (Fig. 6C). These data provide evidence that Grail can interact with viral NP and subsequently target it for proteasome-mediated degradation.

Discussion
The cellular ubiquitin-proteasome system impacts important steps of the IAV life cycle. Ubiquitination plays important roles in innate antiviral immunity by directly targeting viral proteins for degradation and activating antiviral signalling cascades 31 . On the other hand, recent studies have shown that IAV-associated ubiquitination promotes viral polymerase function independently of the proteasomal degradation of viral replication machinery (RNP complexes) 15,22 . Therefore, ubiquitination can be considered a double-edged sword in viral pathogenesis because it has both proviral and antiviral effects 32 . Influenza NP is abundant in both virions 33 and infected cells 34 . NP plays key roles in the transcription and replication of the viral RNA genome as well as in the nuclear transport of vRNP during the life cycle of IAV 35 . NP is also one of main viral proteins to be targeted by host factors for posttranslational modifications such as phosphorylation, sumoylation, and ubiquitination 15 . Several recent studies on IAV-associated ubiquitin modifications have shown that NP ubiquitination can be catalysed by different E3 ligases, such as TRIM22, TRIM41, and CNOT4 19,22,36 . We found that the E3 ligase Grail is a host immune factor that restricts IAV infection directly through the ubiquitination of NP and its subsequent proteasomal degradation. Interestingly, it remains unclear how different E3 ligases can have the same substrate specificity for NP.
In this study, we demonstrated that Grail interacts with viral NP and targets it for degradation. As viral NP is required for the replication of viral RNA and thus also virus amplification, it follows that a loss of Grail would increase IAV replication by increasing the levels of NP, as has also been supported by experiments within this study. Conversely, the overexpression of Grail was shown to reduce the rate of virus replication by reducing the levels of NP. Our in vitro data were further supported by the in vivo observations that Grail KO mice had more immune cell infiltration, higher viral titres, and a greater level of inflammation in their lung tissue than WT mice following infection with IAV.
In a previous study, Grail expression was induced in peritoneal macrophages infected with Sev or HSV-1 virus and subsequently correlated with increases in TBK1 activation, thereby upregulating IRF3 activation and INF-β expression, ultimately leading to a more robust antiviral response 29 . Our data reveals an alternate antiviral mechanism of action by showing that Grail protects lung epithelial cells from IAV infection by enhancing the degradation of viral NP. Combining our data with the literature, we suggest that Grail exerts various antiviral effects through different mechanisms in a cell type-dependent manner. In future studies, it will be interesting to decipher the possible function of Grail in the cross-talk between macrophages and epithelial cells of the lung during IAV invasion.
In summary, we have found that Grail inhibits IAV replication through the promotion of NP degradation. However, it remains possible that, in addition to its influence of NP stability, Grail may interact with host proteins in ways that could also influence the viral infection process.
In conclusion, our findings provide evidence that Grail plays an important regulatory role during IAV infection and suggest that Grail may be a suitable target for the development of anti-influenza drugs.

Materials and Methods
Cell lines and viruses. HEK293, GP2-293, 293 T, A549, and MDCK cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% foetal bovine serum (FBS). The influenza viruses A/ WSN/1933 (WSN, H1N1), pandemic/09 H1N1 and A/Taiwan/3446/2003 (3446, H3N2) were kindly provided by Professor Shin-Ru Shih (Chang Gung University, Taiwan). All viruses were propagated in MDCK cells in DMEM medium containing 1 µg/ml of L-1-tosyl phenylalanyl chloromethyl ketone (TPCK)-treated trypsin at 37 °C. IAV containing a near-infrared fluorescent protein (iRFP) reporter was generated by A/Puerto Rico/8/1934 (PR8, H1N1) using backbone 8-plasmid reverse genetics as described previously 37 . This strain (subsequently referred to as IAV-iRFP) was designed with the complete NS1 and NEP/NS2 genes flanking the iRFP reporter gene (iRFP702, Addgene). A Gly-Ser-Gly-Gly (GSGG) linker was inserted between the NS1 and iRFP genes, while a GSG linker and the 19 amino acid porcine teschovirus-1 (PTV-1) 2 A proteolytic sequence were inserted between the iRFP and NEP/NS2 coding sequences. In addition, two silent point mutations (525-CCCGGG-530) were introduced into the NS1 ORF in order to disrupt the endogenous splice acceptor site.  Immunoprecipitation and immunoblotting. Cells were harvested in lysis buffer (50 mM Tris pH 8.0, 5 mM NaCl, 0.5% NP-40, and 1X protease inhibitor), frozen and thawed three times, and then the proteins were recovered. Immunoprecipitation (IP) proceeded overnight at 4 °C in IP buffer containing antibodies against Grail or Flag. The IP mixture was then incubated with Dynabeads Protein G (Invitrogen) for 1 h prior to isolation using a DynaMag magnet and washing three times with SNNTE buffer (5% sucrose, 1% NP-40, 0.5 M NaCl, 50 mM Tris pH 7.4, and 5 mM EDTA). The immunoprecipitates were resuspended in SDS-PAGE sample buffer, boiled, and loaded onto a gel. Following separation, the proteins were transferred to a nitrocellulose membrane and the blot was probed with antibodies diluted in PBS/Tween 20 with 5% non-fat milk. Antibody detection was carried out using enhanced chemiluminescence reagents (GE Healthcare), as described by the manufacturer. The primary antibodies used for immunoblotting were: anti-PA (GeneTex), anti-PB1 (GeneTex), anti-PB2 (GeneTex), anti-HA (GeneTex), anti-NA (GeneTex), anti-NP (GeneTex), anti-M1 (GeneTex), anti-M2   The infected mice were weighed and observed daily for 14 days, noting any signs of illness or death. In accordance with institutional guidelines, mice that lost more than 25% of their initial body weight were considered moribund and thus scored as dead and euthanised. For the determination of lung titre, three mice from each group were euthanized at 3 days post-infection (dpi) and their lungs were collected for virological and pathological examination. Right lung samples were homogenized in 1.0 ml of unsupplemented DMEM using a Precellys 24 Tissue Homogenizer (Bertin). The homogenates were spun for 5 min at 2000 × g to remove cellular debris. Virus titres were determined by plaque assays in MDCK cells.
Histopathology and immunohistochemistry. Three mice from each group were euthanised on days 4 and 8 post infection. Their left lungs were collected, fixed in 10% neutral-buffered formalin, and embedded in paraffin. Histopathological examination of sequential sections stained with haematoxylin and was performed. The MultiVision Polymer Detection System (Thermo) was used for antigen staining in serial sections. The primary antibody used for immunohistochemistry was a goat anti-influenza A antiserum (Chemicon AB1074, Millipore) that recognizes both the surface glycoproteins and internal proteins of the virus.
In vivo imaging. Mice were anesthetized with 2% isoflurane and shaved before imaging. Mice infected with the iRFP reporter virus were imaged using the Fluorescence Molecular Tomography (FMT) 2000 Quantitative Tomography System (PerkinElmer, Waltham, MA, USA). Mice were placed in a biplanar FMT imaging cassette and carefully adjusted to ensure precise identification and repeat observation of the regions of interest (ROIs). Mice were imaged for 3-5 minutes using a 680 nm laser. Isoflurane anaesthesia was maintained for the duration of the imaging. Fluorescence intensity data were analysed and 3D images reconstructed through the use of the TrueQuant software package provided with the FMT2000.
Real-time PCR. Total RNA from cells and tissues was isolated using TRIzol reagent (Sigma-Aldrich).
Complementary DNA was synthesized using MMLV Reverse Transcriptase (Epicentre). Gene expression was determined using a Roche LightCycler 480. The primers used are listed in Supplementary Table 1.
Cytokine assay. The expression of IL-1β, IL-6, and TNF-α in lung tissue lysates was determined using the Bio-Plex Multiplex Immunoassay kit (Bio-Rad) according to the manufacturer's instructions. All assays were performed at room temperature in 96-well round-bottomed microtiter plates protected from light. Measurements and data analyses were performed with the Bio-Plex system in combination with Bio-Plex Manager software.

Statistical analysis.
Graphing and statistical analysis of data were performed using GraphPad Prism 7 (GraphPad Software). Kaplan-Meier survival curves were analysed using the Mantel-Cox log rank test for statistical significance. For comparison of multiple data sets, one-way analysis of variance (ANOVA) with Tukey's multiple comparison was used. For analysis of two data sets, an unpaired two-tailed Student's t-test was used. P values of ≤0.05 were considered statistically significant. (*P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant).