Inhibition of the IFN-α JAK/STAT Pathway by MERS-CoV and SARS-CoV-1 Proteins in Human Epithelial Cells

Coronaviruses (CoVs) have caused several global outbreaks with relatively high mortality rates, including Middle East Respiratory Syndrome coronavirus (MERS)-CoV, which emerged in 2012, and Severe Acute Respiratory Syndrome (SARS)-CoV-1, which appeared in 2002. The recent emergence of SARS-CoV-2 highlights the need for immediate and greater understanding of the immune evasion mechanisms used by CoVs. Interferon (IFN)-α is the body’s natural antiviral agent, but its Janus kinase/signal transducer and activators of transcription (JAK/STAT) signalling pathway is often antagonized by viruses, thereby preventing the upregulation of essential IFN stimulated genes (ISGs). Therapeutic IFN-α has disappointingly weak clinical responses in MERS-CoV and SARS-CoV-1 infected patients, indicating that these CoVs inhibit the IFN-α JAK/STAT pathway. Here we show that in lung alveolar A549 epithelial cells expression of MERS-CoV-nsp2 and SARS-CoV-1-nsp14, but not MERS-CoV-nsp5, increased basal levels of total and phosphorylated STAT1 & STAT2 protein, but reduced IFN-α-mediated phosphorylation of STAT1-3 and induction of MxA. While MERS-CoV-nsp2 and SARS-CoV-1-nsp14 similarly increased basal levels of STAT1 and STAT2 in bronchial BEAS-2B epithelial cells, unlike in A549 cells, they did not enhance basal pSTAT1 nor pSTAT2. However, both viral proteins reduced IFN-α-mediated induction of pSTAT1-3 and ISGs (MxA, ISG15 and PKR) in BEAS-2B cells. Furthermore, even though IFN-α-mediated induction of pSTAT1-3 was not affected by MERS-CoV-nsp5 expression in BEAS-2B cells, downstream ISG induction was reduced, revealing that MERS-CoV-nsp5 may use an alternative mechanism to reduce antiviral ISG induction in this cell line. Indeed, we subsequently discovered that all three viral proteins inhibited STAT1 nuclear translocation in BEAS-2B cells, unveiling another layer of inhibition by which these viral proteins suppress responses to Type 1 IFNs. While these observations highlight cell line-specific differences in the immune evasion effects of MERS-CoV and SARS-CoV-1 proteins, they also demonstrate the broad spectrum of immune evasion strategies these deadly coronaviruses use to stunt antiviral responses to Type IFN.


Introduction
Coronaviruses (CoVs) pose a major global threat to humanity. They are a large class of zoonotic viruses that exist widely in nature and infect vertebrates [1]. They are enveloped, single-stranded, positive-sense RNA viruses. Coronaviruses are the largest RNA genomic virus, with a genome of around 30 kilobases (kb) in length [2]. In 2002, Severe Acute Respiratory Syndrome-Coronavirus-1 (SARS-CoV-1) emerged in Guangdong, China and spread across the world. The epidemic lasted 8 months, resulting in 8089 infections cantly suppress IFN induction and IFN responses [40,41]. ISRE promoter activity was also reduced by the presence of SARS-CoV-2-nsp2 [41,42].
Even though SARS-CoV-1 and MERS-CoV broke out 20 and 10 years ago, respectively, research progression has failed to fully elucidate the function of all their proteins, with little published data specifically on MERS-CoV-nsp2, MERS-CoV-nsp5, and SARS-CoV-1-nsp14. Moreover, the effect of these three viral proteins on the IFN-α JAK/STAT pathway has not been examined. Therefore, this study sought to examine the impact of MERS-CoV-nsp2, MERS-CoV-nsp5, and SARS-CoV-1-nsp14 upon IFN-α JAK/STAT pathway in two respiratory tract epithelial cell lines and HEK293T cells that are widely used for in vitro molecular analysis.
Here, we report that IFN-α-induced STAT1-3 phosphorylation was stunted by expression of MERS-CoV-nsp2 or SARS-CoV-1-nsp14 in both A549 and BEAS-2B cells. Meanwhile, ISG induction was reduced in the presence of MERS-CoV-nsp2 and SARS-CoV-1-nsp14 in A549 and BEAS-2B cells. In contrast, MERS-CoV-nsp5 did not inhibit pSTAT1-3 nor ISG induction in A549 cells. However, even though MERS-CoV-nsp5 did not inhibit pSTAT1-3 in BEAS-2B cells, it did suppress their ISG induction, suggesting an alternative immune evasion mechanism. Further examination revealed that all three viral proteins inhibited IFN-α-mediated nuclear translocation of STAT1 in BEAS-2B cells. These findings suggest that MERS-CoV-nsp2 and SARS-CoV-1-nsp14 suppress Type 1 IFN responses in a range of epithelial cells, whereby MERS-CoV-nsp5 selectively suppresses these antiviral responses in bronchial BEAS-2B cells. Collectively these discoveries reveal the ability of both of these CoVs to block responses to exogenous IFN-α in human epithelial cells, possibly providing evidence for the ineffectiveness of exogenous IFN-α treatment during CoV infection.

Cell Culture
The alveolar basal epithelial A549 cell line (a kind gift from Dr Kim Roberts, Trinity College Dublin), the bronchial epithelial BEAS-2B cell line (a kind gift from Prof Ultan Power, Queen's University Belfast), and the human embryonic kidney (HEK)293T cell line, were all cultured in DMEM containing 10% FBS and 1% Penicillin/Streptomycin.

qRT-PCR
RNA was isolated from cells using TRI Reagent following the manufacturer's instructions (Sigma). RT-PCR was performed using Sensi-FAST reverse transcriptase (Bioline, Cardiff, UK). qRT-PCR was performed using SYBR-green (Bio-rad) at the following parameters: 95 • C for 15 min, 39 cycles of 95 • C for 30 s, 59 • C for 1 min, and 72 • C for 30 s, using primers specific for the following human genes obtained from Sigma. All genes names, along with forward and reverse primers were summarised in Table 1. Data analysis was carried out using the 2 −∆∆ct method. The relative expression of each result was calculated based on expression of the constitutively expressed housekeeping reference gene ribosomal protein 15 (RPS15).
The expression of all three viral proteins was confirmed by immunoblotting for their HA tags (Figure 2a). Similarly to A549 cells, we observed a significant increase in the expression of STAT1 and STAT2, but not STAT3, in BEAS-2B cells, after 24 h transfection of MERS-CoV-nsp2 (  After 24 h, cells were treated with IFN-α (1000 U/mL) for 15 min. Lysates were generated and subjected into immunoblotting with antibodies for (a) HA, (b) pSTAT1 and STAT1, (e) pSTAT2 and STAT2, or (h) pSTAT3 and STAT3. All blots were also probed with β-actin antibody. (N.B pSTAT1 and pSTAT2 were probed in one membrane and therefore share the same β-actin, and the STAT2 and STAT1 were probed in one membrane and therefore share the same β-actin). Densitometric analysis of (c) pSTAT1, (d) STAT1, (f) pSTAT2, (g) STAT2, (i) pSTAT3 and (j) STAT3 was performed using Image Lab software, and values for STATs or phosphorylated STATs were calculated relative to β-actin and compared to the EV transfected UT (untreated) control, which was normalised to 1. All graphs are the mean ± SEM of three independent experiments. * p < 0.05, ** p < 0.01, *** p < 0.001 (Two-way ANOVA).

MERS-CoV-nsp2, MERS-CoV-nsp5 and SARS-CoV-1-nsp14 Inhibit Nuclear Translocation of STAT1 in BEAS-2B Cells
Since MERS-CoV-nsp5 expression did not suppress the phosphorylation of STAT1-3 (Figure 2b,c,e,f,h,i), but did reduce IFN-α-mediated ISG induction of bronchial BEAS-2B epithelial cells (Figure 5a-c), we next explored if it could suppress ISG induction via a blockade of STAT1 nuclear translocation. To investigate this, we next transfected MERS-CoV-nsp5, MERS-CoV-nsp2, SARS-CoV-1-nsp14, or EV control into BEAS-2B cells for 24 h and stimulated with IFN-α for 30 min, before analysing the localisation of STAT1 by confocal microscopy. Expressions of all three HA-tagged viral protein (green) and STAT1 protein (red) were monitored with the described primary and fluorescent antibodies. Nuclei was stained using DAPI (blue). To present an individual cell, a section of the 1× image was magnified 16 times (Figure 6a). For immunofluorescence analysis, the mean fluorescence intensity of STAT1 for the single cell was measured and then the ratio of cytoplasmic to nuclear STAT1 was quantified in EV, MERS-CoV-nsp2, SARS-CoV-1-nsp14, and MERS-CoV-nsp5-expressing cells. The level of STAT1 nuclear translocation in IFNα-treated EV-transfected cells was normalised to 100% and all relative STAT1 nuclear translocation percentages for MERS-CoV-nsp2, SARS-CoV-1-nsp14, and MERS-CoV-nsp5 were plotted on graphs (Figure 6b). Our immunofluorescence analysis showed that "normal" IFN-α-induced STAT1 nuclear translocation of EV-transfected cells (Figure 6a,b) and was significantly reduced upon expression of MERS-CoV-nsp5, MERS-CoV-nsp2 or SARS-CoV-1-nsp14 (Figure 6a,b). While our previous results suggest that MERS-CoV-nsp2 and SARS-CoV-1-nsp14 reduce ISG induction via suppression of STAT1-3 phosphorylation (Figure 2b,c,e,f,h,i), these confocal microscopy results suggest that all three viral proteins inhibit IFN-α-mediated ISG induction by suppressing STAT1 nuclear translocation of bronchial BEAS-2B epithelial cells.    and anti-Alexa Fluor 647 were probed as secondary antibodies to monitor HA expression and STAT1 subcellular localisation. Nuclei were stained by DAPI. Each full scaled image was then zoomed in for presenting single transfected cells. (b) Quantification of nuclear translocation of STAT1 in control, MERS-CoV-nsp2, SARS-CoV-1-nsp14, and MERS-CoV-nsp5-positive cells from three fields of view on each coverslip collected from three independent experiments. The level of STAT1 translocation to nucleus for IFN-α-treated EV was normalised to 100%. Scale bar, 20 µm. All graphs are the mean ± SEM of three independent experiments * p < 0.05, *** p < 0.01, **** p < 0.0001 (Student's t test).

Discussion
IFN-α signalling is essential for early anti-viral innate immune responses and its induction leads to the up-regulation of hundreds of ISGs. However, many viruses have evolved mechanisms to suppress the IFN-α JAK/STAT pathway. Here we investigated the effect of several non-structural proteins (from MERS-CoV and SARS-CoV-1), upon the IFN-α pathway in epithelial cells. We found that nsp2 from MERS-CoV and nsp14 from SARS-CoV-1 enhanced protein expression of STAT1 and STAT2, but not STAT3, in both A549 and BEAS-2B epithelial cells. In the absence of exogenous Type 1 IFN, MERS-CoV-nsp2 and SARS-CoV-1-nsp14 also induced basal STAT1 and STAT2 tyrosine phosphorylation in A549 cells (but not BEAS-2B cells), suggesting that these viral proteins induce IFNs and/or other cytokines, which may act in a paracrine/autocrine fashion and induce and activate STATs. Indeed, increased cytokine expression is a common feature of both MERS-CoV and SARS-CoV-1 infection [50] and differences between these two epithelial cell lines have previously been observed in response to other viruses. In fact, when compared to BEAS-2B cells, A549 cells produced higher cytokine levels upon respiratory syncytial virus or influenza A virus infection, further supporting our hypothesis that upon expression of MERS-CoV-nsp2 and SARS-CoV-1-nsp14, basal STAT and pSTATs observed in A549 cells, might be as a result of increased cytokine secretion [51,52].
The JAK/STAT pathway plays an essential role in eliminating viruses [6,53], therefore, it is not surprising that STAT protein expression and activation are targeted by these pathogens. Indeed, the Hepatitis B virus (HBV) HBX protein has been shown to interact with JAK1 and activate downstream phosphorylation of STATs [54]. Similarly, v-abl from Abelsom Murine-leukaemia Virus (A-MuLV) interacts with JAK1 and JAK3, also leading to enhanced STAT phosphorylation [55]. The EBV SM protein also enhanced expression of STAT1 mRNA and protein and upregulated ISG expression in B cells [56]. Hepatitis C Virus (HCV) core, NS5B, and p7 proteins have all been shown to regulate STAT3 phosphorylation, which, in the case of HCV-p7, induces SOCS3, thus suppressing proinflammatory TNF-α responses [57][58][59]. Porcine reproductive and respiratory syndrome virus (PRRSV) and EBV have also been shown to induce serine phosphorylation of STAT1 [60,61], with constitutive activation of STAT1 restricting IFN-stimulated STAT1-DNA binding [61]. Recently, STAT2 was also shown to play a dual role in SARS-CoV-2 hamster model infection; whereby STAT2 drove severe lung injury, due to elevated proinflammatory cytokine expression, but also restricted systemic virus dissemination, clearly showing a significant effect of another CoV upon the JAK/STAT pathway [62].
In our current study, we found that the exogenous addition of IFN-α to A549 and BEAS-2B epithelial cells, expressing MERS-CoV-nsp2 and SARS-CoV-1-nsp14, did not induce normal levels of STAT1-3 phosphorylation. This contrasted with HEK293T embryonic kidney cells, where IFN-α-induced STAT1-3 phosphorylation was not affected by either viral protein, identifying differential responses to MERS-CoV and SARS-CoV-1 proteins between human respiratory epithelial and kidney cells. Furthermore, while our over-expression studies did not specifically require CoV cell receptors, DDP4 and ACE2 expression levels are important for infection studies. A549 and HEK293T cells both lack ACE2 and DPP4 expression [63][64][65][66][67], whereby BEAS-2B cells express ACE2 and can thus be readily infected with either SARS-CoV-1 or SARS-CoV-2 [68,69]. While there is little known about DDP4 expression in BEAS-2B cells, a protein atlas database shows DPP4 is moderately expressed in the human bronchus [70], suggesting that BEAS-2B cells could also be susceptible to MERS-CoV infection. Since ACE2 has been revealed as an ISG [71], dysregulation of the IFN-α JAK/STAT pathway might also reduce ACE2 levels, thus inhibiting CoV infection. Type-I IFN signalling was also dysregulated in SARS-CoV-2-infected-human epithelial Calu-3 and Caco-2 cell lines, but as with our findings, this dysregulation was not observed in HEK293T cells [72], again identifying cell-type-specific differences.
The decrease of IFN-α-induced pSTAT1-3 by MERS-CoV-nsp2 and SARS-CoV-1-nsp14 in A549 cells and BEAS-2B cells clearly underlined the ability of these viral proteins to inhibit IFN-α JAK/STAT signalling. The anti-viral activity of STAT1 & STAT2 makes them primary targets for viruses such as RSV, HIV, and HCV [43,73,74]. Even though the anti-viral activity of STAT3 is less clearly defined, its expression has been shown to be indispensable for the IFN-α-mediated upregulation of a specific ISG subset, including MxA and ISG15 [28]. However, several viruses, including mumps, HCV, and HIV, have been shown to target STAT3 and block its anti-viral activity [43,74,75], indicating that our observed CoV-mediated reduction in pSTAT3 may also be responsible for reduced ISG induction. A recent genomics-based study identified differential regulation of STAT3 in two highly pathogenic strains of MERS-CoV (MERS-CoV Eng 1 and MERS-CoV SA 1); as a result, changes in STAT3 activity were speculated to result in altered pro-inflammatory gene expression in human airway Calu-3 cells [76], suggesting that STAT3 is closely related to MERS-CoV pathogenesis. Having shown that MERS-CoV-nsp2 inhibited IFN-α-induced phosphorylation of STAT3 in A549 and BEAS-2B cells, our results further support a role for STAT3 in the anti-viral Type I IFN response during MERS-CoV infection. Our findings are also in agreement with a previous study showing decreased STAT3 phosphorylation in SARS-CoV-1-infected Vero E6 cells [77]; but we may have revealed a specific role for SARS-CoV-1-nsp14, as a suppressor in the pSTAT3 in A549 and BEAS-2B cells.
Type I IFNs stimulate cells to upregulate the expression of ISGs, which inhibit viral infection [53]. In our study we investigated the effect of MERS-CoV-nsp2, MERS-CoV-nsp5, and SARS-CoV-1-nsp14 on several well-characterised anti-viral ISGs (MxA, PKR and ISG15). While these ISGs have several actions, MxA is well known to prevent the endocytic trafficking of viral particles into the cell [78]; PKR prevents viral protein translation [79] and ISG15 inhibits viral transcription, translation, and egress from the cell [29]. This spectrum of anti-viral effects highlights the importance of limiting ISG expression and effectiveness by viruses; for example, HIV-Vif protein has been shown to block IFN-αinduced ISG15 induction and HCV to attenuate IFN-α-induced OAS1 upregulation [43,80]. In this study, we found that significant IFN-α-induced upregulation of MxA was lost in A549 cells expressing either SARS-CoV-1-nsp14 or MERS-CoV-nsp2. However, since we did not observe significant induction of ISG15 nor PKR mRNA in EV control A549 cells, it was impossible to measure any statistical changes between EV and viral protein-expressing cells. One of our previous studies, analysing the effect of IFN-α and Ribavirin upon ISG induction, also showed an expression difference between MxA and other ISGs. We observed an increase in IFN-α & ribavirin-mediated MxA in Huh7 cells, whereas, PKR, OAS, and CXCL10 were not induced, which not only highlights the complexity of specific ISG expression levels, but, combined with our current results, possibly reveals the importance of MxA in combating viral infection [81]. The significant IFN-α-mediated upregulation of all three ISGs in EV-control BEAS-2B cells was lost upon SARS-CoV-1-nsp14 and MERS-CoV-nsp5 expression. While MERS-CoV-nsp2 expression reduced the induction of all three ISGs, PKR was the only ISG to see a statically significant loss. As IFN-α-induced ISG15 induction has been shown to correlate with STAT3 expression [28], the attenuated ISG15 induction in MERS-CoV-nsp2 and SARS-CoV-1-nsp14-expressing BEAS-2B cells may be linked to their decrease in STAT3 phosphorylation. Taken together, these results provide clues to how MERS-CoV and SARS-CoV-1 individual proteins blocked anti-viral ISG induction in epithelial cells, making them an optimal environment for CoV infection and replication.
We found that MERS-CoV-nsp5 expression had no effect on total-nor phopho-STAT1-3 protein levels, but still attenuated IFN-α-mediated ISGs induction, indicating that it regulated the JAK/STAT pathway through a different mechanism than MERS-CoV-nsp2 and SARS-CoV-1-nsp14. Since SARS-CoV-2-nsp5 has been reported to impair the IFNα-mediated induction of anti-viral ISGs through blocking the nuclear translocation of STAT1 [40], we investigated if MERS-CoV-nsp5 could also be using this immune evasion mechanism. Indeed, consistent with this study, we found that MERS-CoV-nsp5 also reduced the nuclear translocation of STAT1, possibly accounting for the inhibition of IFNα-mediated ISG induction. Meanwhile, we also found MERS-CoV-nsp2 and SARS-CoV-1-nsp14 inhibited STAT1 nuclear translocation in BEAS-2B cells; together these results reveal that MERS-CoV-nsp2 and SARS-CoV-1-nsp14 target the IFN-JAK/STAT pathway by suppressing both IFN-α-mediated STAT1-3 phosphorylation and STAT1 nuclear translocation, which may result in our observed reduction of downstream ISGs. Similarly, SARS-CoV-2-N and SARS-CoV-2-nsp5 proteins antagonize type I IFN signalling, by both suppressing phosphorylation and nuclear translocation of STAT1 and STAT2 [40,82], highlighting that CoVs may have developed multiple strategies to restrict the anti-viral IFN-α JAK/STAT pathway.
Having observed three deadly CoVs to emerge over the past 20 years, it is clear that we need novel therapeutic approaches for their treatment. IFN-alphacon-1 (synthetic IFN-α), in combination with steroids, has shown partially protective effects in the treatment of SARS-CoV-1 patients [83]. However, other studies failed to see the benefit of IFN-α treatment [36], suggesting that the JAK/STAT pathway is blocked. Clinical trials investigating lopinavirritonavir and IFN-β1b combination therapy for MERS-CoV [84] and IFN-β (in combination with other anti-viral treatments), for treatment of SARS-CoV-2 [85,86], are ongoing. Indeed, pre-treatment with IFN-α has shown some protective effects against SARS-CoV-2 infection, and SARS-CoV-2 has been shown to be more sensitive to IFN-α, when compared with SARS-CoV-1 in Vero cells [42,87]. Overall, the results of our study indicate that MERS-CoV-nsp2, MERS-CoV-nsp5, and SARS-CoV-1-nsp14 inhibit the IFN-α JAK/STAT pathway, which may also provide insight into the functional effects of homologous SARS-CoV-2 proteins and even future emergent CoVs. Indeed, the suppressive effects of MERS-CoV-nsp2, MERS-CoV-nsp5, and SARS-CoV-nsp14 upon the IFN-α JAK/STAT pathway, may also explain the previously observed therapeutic failure of Type I IFN. Therefore, it may be beneficial for future studies to focus on restoring IFN-α responsiveness, as a therapeutic approach to existing and future CoV infections.