MERS-CoV-nsp5 expression in human epithelial BEAS 2b cells attenuates type I interferon production by inhibiting IRF3 nuclear translocation

Abstract

Middle East Respiratory Syndrome Coronavirus (MERS-CoV) is an enveloped, positive-sense RNA virus that emerged in 2012, causing sporadic cases and localized outbreaks of severe respiratory illness with high fatality rates. A characteristic feature of the immune response to MERS-CoV infection is low type I IFN induction, despite its importance in viral clearance. The non-structural proteins (nsps) of other coronaviruses have been shown to block IFN production. However, the role of nsp5 from MERS-CoV in IFN induction of human respiratory cells is unclear. In this study, we elucidated the role of MERS-CoV-nsp5, the viral main protease, in modulating the host’s antiviral responses in human bronchial epithelial BEAS 2b cells. We found that overexpression of MERS-CoV-nsp5 had a dose-dependent inhibitory effect on IFN-β promoter activation and cytokine production induced by HMW-poly(I:C). It also suppressed IFN-β promoter activation triggered by overexpression of key components in the RIG-I-like receptor (RLR) pathway, including RIG-I, MAVS, IKK-ε and IRF3. Moreover, the overexpression of MERS-CoV-nsp5 did not impair expression or phosphorylation of IRF3, but suppressed the nuclear translocation of IRF3. Further investigation revealed that MERS-CoV-nsp5 specifically interacted with IRF3. Using docking and molecular dynamic (MD) simulations, we also found that amino acids on MERS-CoV-nsp5, IRF3, and KPNA4 may participate in protein-protein interactions. Additionally, we uncovered protein conformations that mask the nuclear localization signal (NLS) regions of IRF3 and KPNA4 when interacting with MERS-CoV-nsp5, suggesting a mechanism by which this viral protein blocks IRF3 nuclear translocation. Of note, the IFN-β expression was restored after administration of protease inhibitors targeting nsp5, indicating this suppression of IFN-β production was dependent on the enzyme activity of nsp5. Collectively, our findings elucidate a mechanism by which MERS-CoV-nsp5 disrupts the host’s innate antiviral immunity and thus provides insights into viral pathogenesis.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00018-024-05458-y.

Introduction

Among seven human CoVs, SARS-CoV-2, SARS-CoV, and MERS-CoV are considered more pathogenic, while the other four CoVs usually cause mild infections [1]. MERS-CoV is a highly contagious viral respiratory illness that was first identified in 2012 in Saudi Arabia. According to the WHO, the overall mortality rate for MERS-CoV is estimated to be around 35% (WHO 2023) which is higher than that of SARS-CoV-1, 9.6% [2] and SARS-CoV-2, 0.9% [3], but this rate can be higher in certain populations such as those with underlying health conditions or who are immunocompromised. There is no specific treatment or vaccine available for MERS-CoV, and public health officials continue to closely monitor the virus for any new developments [4]. As such, development of therapeutics against MERS-CoV are of great importance.

A thorough understanding of the interplay of MERS-CoV and the host immune response will likely provide insight into potential therapeutic targets. Upon viral entry, the first stage of the innate immune response is viral detection and the cytosolic RIG-I-like receptors signalling pathway is a vital mediator of this. The RLRs, including RIG-I and MDA5, recognize viral dsRNA in the cytoplasm and activate a cascade of signalling events leading to the production of type I IFN and other antiviral cytokines [5]. Upon detection of viral RNA, RIG-I and MDA5 undergo a conformational change that allows them to interact with the mitochondrial adapter protein MAVS, leading to the recruitment and activation of downstream signalling molecules. This ultimately results in the activation of the transcription factors IRF3 and NF-κB, which induce the expression of type I IFN and other antiviral genes [5]. The type I IFN is the first line of defense against viral infections in the innate immune system [6]. Dysregulation of these pathways can result in viral replication, persistence, and exacerbation of disease. Many viruses, including Coronaviruses, have evolved strategies to limit detection through the RLR signalling pathway and to evade or suppress type I IFN responses, allowing them to establish infection and increase replication. Despite their small genome sizes, the multifunctional roles of viral proteins serve to antagonise elements of these pathways.

MERS-CoV is a single-stranded, positive-sense RNA virus belonging to the Betacoronavirus genus of the Coronaviridae family [1]. The viral genome of MERS-CoV is approximately 30 kb in length which encodes structural proteins including the S, E, M, and N proteins. The S protein plays a crucial role in viral entry into host cells by binding to the cellular receptor DPP4 [7]. The other structural proteins are involved in the assembly and release of new viral particles. The MERS-CoV genome also encodes 5 accessory proteins including protein 3 [p3], p4a, p4b, p5, and p8b and 16 non-structural proteins (nsps) [8]. While the accessory proteins of CoVs are not crucial for viral replication and virion assembly, they do impact virulence by affecting the release, stability and pathogenesis of the virus [9]. Therefore, understanding the MERS-CoV genome and the roles of its encoded proteins is essential to understand how the virus modulates the host immune response to enhance viral pathogenesis.

A characteristic feature of the immune response to MERS-CoV infection is delayed type I IFN induction [10], despite its importance in viral clearance. Several MERS-CoV proteins have been reported to interfere with the type I IFN response at different levels. For example, the N protein has been shown to inhibit the production of IFNs by interacting with TRIM25, therefore impeding RIG-I activation and inhibiting the phosphorylation of IRF3, that is known to be important for IFN gene activation [11]. Additionally, the M protein has been identified to suppress type I IFN expression at the level of TANK binding kinase 1 (TBK1)-dependent phosphorylation and activation of IRF3 [12]. The ORF4a could act as an efficient type I IFN antagonist inhibiting MDA5-dependent IFN induction and interaction with dsRNA [13]. The ORF4b protein has also been found to inhibit type I IFN production through a direct interaction with TANK binding kinase 1 (TBK1) and IκB kinase epsilon (IKKε), resulting in reduced molecular interaction between MAVS and IKKε [14]. It’s been reported that ORF8b suppressed type I IFN expression by competing with IKKε for interaction with HSP70 which is required for the activation of IRF3 [15]. The papain-like protease (PLpro), which is a protease and can recognize and process the polyproteins encoded by the MERS-CoV genomic RNA has been characterized an IFN antagonist through blocking the phosphorylation and nuclear translocation of IRF3 [16]. Interestingly, a recent study revealed that three PLpros of SARS-CoV-2, SARS-CoV-1 and MERS-CoV have similar overall structures and all bound to the stimulator of interferon genes (STING) and inhibited the activation of IFN-β and ISRE luciferase reporters [17].

The CoV non-structural protein 5 (nsp5), also known as 3 C-like protease (3CLpro), is an indispensable enzyme involved in the replication of CoV [18]. It is responsible for the cleavage of the viral polyprotein into functional individual proteins required for viral replication [18]. Due to its essential role in viral replication, inhibitors of nsp5 have been identified to have promising antiviral activity against CoVs in vitro, including SARS-CoV-2 and MERS-CoV [19]. In addition to its role in viral replication, the nsp5 has also been implicated in modulation of host innate immune responses. The nsp5 proteins encoded by SARS-CoV-1 and SARS-CoV-2 were reported to antagonize IFN production by retaining phosphorylated IRF3 in the cytoplasm [20]. Meanwhile, SARS-CoV-2-nsp5 has been shown to cleave RIG-I in a protease dependent manner and promote MAVS ubiquitination and degradation [21]. However, another study revealed that SARS-CoV-2-nsp5 had no effect upon MAVS, but specifically restricted IFN induction by reducing K63-linked ubiquitination of RIG-I [22]. Collectively, these findings suggest that coronavirus nsp5 is a significant IFN antagonist. However, there is limited research on the specific effects of MERS-CoV-nsp5 upon the IFN response. As MERS-CoV and SARS-CoV-1 & 2 share many similarities in their genome structure and pathogenesis [23], it is possible that MERS-CoV-nsp5 may also modulate the host immune response by similar mechanisms. Here, we identified that MERS-CoV-nsp5 attenuation of type I IFN induction is mediated by specific binding to and inhibition of IRF3 nuclear translocation in human bronchial epithelial BEAS 2b cells. By analysing the docking and MD calculations in protein-protein interactions, we have also identified protein conformations that involve masking of the nuclear localization signal (NLS) region of IRF3 [24] and KPNA4 (co-ordinates from NCBI Reference Sequence NP_002259) with MERS-CoV-nsp5. Together this data provides the first evidence demonstrating that MERS-CoV-nsp5 can antagonize type I IFN responses.

Materials and methods

Cells culture

BEAS 2b cells were cultured at 37 °C and 5% CO₂ using Dulbecco’s minimal essential medium (DMEM) containing 10% foetal bovine serum (FBS), 1 µg/ml penicillin, streptomycin (P/S).

Transfection

BEAS 2b cells were seeded into 6-well plates at a density of 2.5 × 10⁶ cells per well or 12-well plates at density of 1.25 × 10⁶ and grown in DMEM with 10% FBS and 1 µg/ml P/S. The following day cells were transfected with 200 or 400 ng/ml plasmid DNA Empty Vector (EV) control plasmid or HA-MERS-CoV-nsp5 (full length gene) are kind gifts from Prof. Matthew Frieman, University of Maryland; the plasmids Flag-RIG-I, Flag-MAVS, Flag-IKK-ε and Flag-IRF3-5D are kind gifts from Dr Yongxu Lu, University of Cambridge) using lipofectamine 2000 following the manufactures instructions. After 24 h cells were treated as described and harvested for protein, RNA, or prepared for imaging.

Treatment

BEAS 2b cells were stimulated with 4ug/ml HMW-poly(I:C) (Invivogen) for 24, 4–2 h as indicated. Calpain inhibitor II (Sigma), Rupintrivir inhibitor (Sigma) and MG132 inhibitor (Sigma) were applied with 10 μm for 24 h.

qRT-PCR

Total RNA was extracted from cells using TRIreagent (Sigma, USA) following manufacture instructions. All RNA samples were diluted to 250ng in 18.5 µl (13.5ng/l) in RNase free water (Sigma). The RNA was converted to cDNA using the SensiFAST cDNA Synthesis kit (Bioline, UK). qRT-PCR was performed using SYBR-green (Bio-rad) at the following parameters: 95 °C for 15 min, 40 cycles of 92 °C for 30 s, 63.5 °C for 1 min, and 72 °C for 30 s, using primers specific for the following human genes obtained from Sigma. 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). Primer sequences: IFN-β forward CTAGCACTGGCTGGAATGAGA, reverse CTGACTATGGTCCAGGCACA, IL-6 forward TCCACAAGCGCCTTCGGTCC, IL-6 reverse GTGGCTGTCTGTGTGGGGCG, IL-8 forward GCAGAGGGTTGTGGAGAAGTTT, IL-8 reverse ACCCTACAACAGACCCACACAAT, KPNA3 forward CTTGGAGAACCACCGCATCA, reverse GCGGGGGATCCTTATTCCTG, KPNA4 forward TACCTCCCACCAGAGGACTG, reverse GCTGAACAGGACAACCCTGA, RPS15 forward CGGACCAAAGCGATCTCTTC, reverse CGCACTGTACAGCTGCATCA.

ELISA

BEAS 2b cells cultured in 6-well plate were transfected with empty vector control or plasmid expressing MERS-CoV-nsp5. After 24 h transfection, the cells were transfected with 4ug/ml HMW-poly(I:C) (tlrl-pic, invivogen) for another 24 h. Supernatants were harvested and cytokine concentrations of IFN-β were quantified by ELISA (DY814-05, R&D Systems) according to the manufacturer’s guideline.

Luciferase reporter assay

BEAS 2b cells cultured in 12-well plates were transfected with IFN-β firefly luciferase reporter and pRL-TK Renilla luciferase reporter, together with plasmids expressing indicated proteins. After 48 h cells were harvested and lysed using 1X Passive lysis buffer (Promega). Firefly and Renilla luciferase signals were quantified using Dual Luciferase Reporter Assay System. The firefly luciferase activity levels were normalized to the Renilla luciferase activity levels.

Western blotting

Total protein was extracted from cells using RIPA buffer supplemented with phosphatase and protease inhibitors (Phenylmethylsulfonyl fluoride (PMSF), Na₃VO₄, Leupeptin, Dithiothreitol (DTT)) immediately prior to use. Protein lysates were run through 10–15% acrylamide gels and then transferred onto Polyvinylidene diflouride (PVDF) membrane. The PVDF membrane was incubated with primary antibody (HA, 1:1000, 3724, Cell Signalling Technology; RIG-I, 1:1000, 3743T, Cell Signalling Technology; MAVS, 1:1000, 3993T, Cell Signalling Technology; IKK-ε, 1:1000, 2905T, Cell Signalling Technology; NEMO, 1:1000, 2695T, Cell Signalling Technology; FLAG, 1:1000, F7425-.2MG, Merck; IRF3, 1:1000, 11904T, Cell Signalling Technology; pIRF3, 1:1000, 37829, Cell Signalling Technology; Vinculin, 1:1000, 13901T, Cell Signalling Technology; Lamin B2, 1:500, 871502, Biolegend; KPNA3, 1:5000, 67892-1-IG, Proteintech; KPNA4, 1:5000, 12463-1-AP, Proteintech; β-actin, 1:2000, A5441-.2ML, Sigma-Aldrich) overnight at 4 ℃. The membranes were then washed and incubated in the appropriate secondary antibody (anti-Rabbit, 1:2000, 11859140, Fisher Scientific or anti-Mouse, 1:2000, 10158113, Fisher Scientific) for 1 h before imaging (BioRad Imager). The protein bands were viewed using a Gel Doc EZ imager (BioRad, USA). The densitometry for each band was carried out using Bio-Rad Image Lab software (BioRad, USA). The data produced from this software was exported to GraphPad Prism 9 where it was graphed and analysed.

Confocal microscopy

BEAS 2b cells were seeded onto glass slides and transfected for 48 h as described above. The cells were then transfected with 4ug/ml HMW-poly(I:C) for 2 h and fixed with 4% paraformaldehyde (PFA) for 20 min. Cells were washed with PBS and permeabilised with 0.2% Triton X-100 for 30 min and blocked in 0.5% BSA for 1 h at room temperature. The slides were then treated with diluted primary antibodies (IRF3, 11904T, Cell Signalling Technology; HA, 2367 S, Cell Signalling Technology) and incubated overnight at 4 °C. These were then washed and incubated in secondary antibodies (anti-Rabbit, SAB4600084, Merck or anti-Mouse, 405322, MSC) for 1 h in the dark at RT. The slides were mounted using DAPI ProGold Mounting media (P36941, Thermo Fisher) and imaged using a Lecia SP8 scanning confocal microscope. Quantitative analysis was performed using IMARIS software (Oxford Instruments).

Cytoplasmic and nuclear fractions

BEAS 2b cells cultured in 6-well plates were transfected with empty vector control or plasmid expressing MERS-CoV-nsp5. After 48 h, cells were transfected with 4ug/ml HMW-poly(I:C) for 2 h and harvested for cytoplasmic and nuclear lysate though using NE-PER Nuclear and Cytoplasmic Extraction Reagents (78833, Thermo Fisher Scientific) according to the manufacturer’s guideline.

Immunoprecipitation

All samples for immunoprecipitation were lysed in HEPES lysis buffer (50 mM HEPES, 150 mM NaCl, 2 mM EDTA, 1% NP40 and 0.5% sodium deoxycholate), which was supplemented with phosphatase and protease inhibitors immediately prior to use. For 2 mL buffer: 20 µL Phenylmethylsulfonyl fluoride (PMSF, 1 mM), 20 µL Na₃VO₄ (1 mM), 1 µL Leupeptin (5 µg/mL) and 2 µL Dithiothreitol (DTT, 1 mM) was added to 1957 µL RIPA buffer. For cell lysis, sample cells were resuspended in 200 µL lysis buffer and left on ice for 30 min, agitating every 10 min. Samples were then centrifuged at 18,000 RCF for 10 min and the supernatant were collected. 40 µL of lysate was transferred to new tubes and store at -20 °C. The remainder (150 µL) of lysate was transferred to another tube and immunoprecipitated with FLAG antibodies (F7425-.2MG, Merck) a on rotator overnight at 4 °C. 20 µL of resuspended protein A/G agarose beads (Santa Cruz Biotechnologies) were added to immunoprecipitated lysate and incubated on the rotator at room temperature overnight, before centrifuging at 18,000 RCF for 5 min at 4 °C. Supernatant was removed and the pellet was washed 4 times with 500 µL HEPES buffer. After the final wash, the supernatant was discarded, and the pellet was resuspended in 40 µl of 1x sample buffer. Samples were boiled for 10 min and then stored at -20 °C.

3D protein-protein modelling

Protein-protein interactions were predicted by generating docked poses using the molecular docking protocol implemented in MOE 2022.02 [25]. The X-ray crystal structure of IRF3 (PDB: 3QU6) had a resolution of 2.3 Å [26], providing sufficient detail to resolve amino acids within the NLS region. For MERS-CoV-nsp5, the complete X-ray crystal structure (PDB: 8E6B) with a resolution of 1.55 Å was utilized [27]. The X-ray crystal structure of KPNA4 (PDB: 7LFC) with a resolution of 2.1 Å was selected [28]. To assess the stability of the predicted docked poses, molecular dynamics (MD) simulations were conducted using GROMACS 2020.7 (GNU General Public License http://www.gromacs.org) [29].

Structure preparation and molecular docking

The protein structures were prepared using the MOE Quick Prep option, utilizing the default Amber10: EHT force field to account for explicit hydrogen atoms, tautomeric states, and potential breaks in the protein structure before conducting restrained all-atom molecular mechanics minimization and electrostatics calculations. Subsequently, these prepared protein structures were employed for MOE protein-protein docking, with patch analysis configured to utilize a hydrophobic patch potential. The number of poses was set to 10,000 for pre-placement and 100 for refinement. Docking was performed in triplicate to verify the reproducibility of the docking algorithm implemented in MOE. Additionally, to compare the most stable docked pose predicted by MOE, docking was also performed using the same prepared protein files on the ClusPro 2.0 server [30] and ZDOCK 3.0.2 server [31] with default parameters. Although AlphaFold3 [32] was utilized, the docked output structures exhibited numerous regions with undefined secondary structure and substantial differences compared to the known X-ray structures, rendering them unsuitable for further consideration.

Molecular dynamics (MD) simulations

The input files for GROMACS 2020.7 MD simulations were generated using the CHARMM-GUI web server [33]. The CHARMM36m force field was used for parameterisation of the protein chains. The protein complexes were centred in a rectangular waterbox with an edge distance of 10.0 Å. The box was solvated using the TIP3P water model with physiological pH as 7. The systems were then neutralized with K + and Cl − ions added via the Monte-Carlo ion-placing method and simulated within periodic boundary conditions. Energy minimisation was performed using steepest descent for 5,000 steps. The minimised system was then equilibrated with constant number of particles, volume, and temperature (NVT) ensemble, subjected to 125 picoseconds (ps) run followed by dynamics run with constant number of particles, pressure (1 atm), and temperature (NPT) ensemble at a constant temperature of 303.15 K. The MD production runs for each pose was set to 100 nanoseconds (ns) and all restraints were removed. The GROMACS in-built tools were used to calculate the number of hydrogen bonds, root-mean-square deviations (RMSD) between the chains and clustering analysis to output the most representative frame from the MD simulation. The output files were visualised using XMGRACE software [34] and the most representative frame was structurally overlaid on the predicted docked pose and the interaction surface amino acids RMSD was calculated with MOE.

Contact analysis

MOE contact analysis was used to analyse the interactions of the most stable docked protein-protein interaction poses from MOE. In MOE contact analysis six types of contacts can be identified: Hydrogen bonds (Hbond), Metal, Ionic, Arene, Covalent and van der Waals distance interactions (Distance). All of these options were selected, and the calculations were performed by setting the display and within option to all.

Statistical analysis

Statistical comparisons between groups were performed using GraphPad Prism statistical analysis software (version 9). Data is represented as the mean ± SEM unless otherwise stated.

Results

MERS-CoV-nsp5 suppresses HMW-poly(I:C)-induced type I IFN promoter activation and production

To identify if MERS-CoV-nsp5 could regulate innate immune signalling, the effect of MERS-CoV-nsp5 on IFN-β promoter activation was firstly investigated. Human bronchial epithelial BEAS 2b cells were transfected with increasing concentrations (200 & 400 ng/ml) of EV control or plasmid expressing MERS-CoV-nsp5 along with an IFN-β promoter-driven luciferase reporter plasmid (IFN-β-Luc) and a control RL-TK plasmid. After 24 h, the cells were transfected with high molecular weight poly(I:C) (HMW-poly(I:C)), which activates RIG-I/MDA5 signalling pathways. Subsequently, the luciferase activity was measured after 24 h. It was found that MERS-CoV-nsp5 inhibited HMW-poly(I:C)-induced IFN-β promoter activation in a dose-dependent manner (Fig. 1A). To investigate if IFN-β protein was affected by MERS-CoV-nsp5, BEAS 2b cells were transfected with increasing concentrations (200 & 400 ng/ml) of EV control or plasmid expressing MERS-CoV-nsp5. After 24 h, the cells were transfected with HMW-poly(I:C) for another 24 h and protein levels of supernatant IFN-β were measured. We found that higher concentrations of MERS-CoV-nsp5 significantly reduced HMW-poly(I:C)-induced IFN-β protein expression (Fig. 1B). The protein expression of MERS-CoV-nsp5 was determined by immunoblotting for its HA tag and densitometric analysis (Fig. 1A and B). Since high concentration of MERS-CoV-nsp5 demonstrated a heightened inhibitory impact on IFN-β expression, we then assessed IFN-β gene expression level by RT-qPCR in BEAS 2b cells expressing high concentrations of EV control or MERS-CoV-nsp5 in response to intracellular treatment of HMW-poly(I:C) for 4 h. The IFN-β mRNA expression was significantly reduced by MERS-CoV-nsp5 (Fig. 1C). In addition to IFN-β, it was also observed that IL-6 mRNA expression was reduced (Fig. 1D), while IL-8 mRNA levels were not significantly affected (Fig. 1E), revealing that MERS-CoV-nsp5 inhibits HMW-poly(I:C) triggered IFN-β and specific proinflammatory cytokine production.

Fig. 1.

MERS-CoV-nsp5 suppresses HMW-poly(I:C)-induced type I IFN production. (A) BEAS 2b cells were co-transfected with IFN-β reporter plasmid, TK Renilla plasmid and increasing concentrations (200 & 400 ng/ml) of HA-tagged MERS-CoV-nsp5 or EV control plasmids. At 24 h post transfection, cells were transfected with 4 ug/ml HMW-poly(I:C) for another 24 h and assayed for luciferase activity. The levels of indicated proteins (HA and β-actin) were assessed by western blotting. (B) BEAS 2b cells were transfected with increasing concentrations (200 & 400 ng/ml) of MERS-CoV-nsp5 or EV control plasmids. At 24 h post transfection, cells were transfected with or without 4 ug/ml HMW-poly(I:C) for another 24 h and supernatants were harvested and assayed for IFN-β production. The levels of indicated proteins (HA-MERS-CoV-nsp5 and β-actin) were assessed by western blotting. Densitometry was performed using Image Lab software and values of HA-MERS-CoV-nsp5 were calculated relative to β-actin and compared to the lower concentrations of EV control. BEAS 2b cells were transfected with high concentration (400 ng/ml) of MERS-CoV-nsp5 or EV control plasmid. After 48 h, cells were transfected with 4 ug/ml HMW-poly(I:C) for 4 h. Total RNA was extracted, and the expression of (C) IFN-β, (D) IL-6 and (E) IL-8 was detected by RT-qPCR. Gene expression was normalised to house-keeping gene RSP15 and compared to the EV HMW-poly(I:C) untransfected control (UT). All experiments were performed at least three times and all data is shown as mean ± SEM. *P < 0.05. **P < 0.01 (Figure A was analysed by Student’s t-test and Figure B, C, D & E were analysed by Two-way ANOVA)

MERS-CoV-nsp5 inhibits IRF3-triggered IFN-β activation

To determine the specific host targets affected by the inhibitory action of MERS-CoV-nsp5, key components of the RIG-I-IFN pathway, including RIG-I, MAVS, IKK-ε and IRF3-5D (constitutive active form of IRF3) were overexpressed. Then the IFN induction was examined in the presence of MERS-CoV-nsp5. BEAS 2b cells were co-transfected with increasing concentrations (200 & 400 ng/ml) of EV control or MERS-CoV-nsp5 along with an IFN-β promoter-driven luciferase reporter plasmid (IFN-β-Luc), control RL-TK plasmid and RIG-I, MAVS, IKK-ε or IRF3-5D plasmids for 48 h. As shown by IFN-β luciferase reporter assay, it was found that overexpression of MERS-CoV-nsp5 inhibited RIG-I-, MAVS-, IKK-ε- and IRF3-5D-triggered IFN-β promoter activation in a dose-dependent manner (Fig. 2A, B, C & D). These results further suggest that MERS-CoV-nsp5 might target the RIG-I signalling pathway through IRF3 or downstream of IRF3 activation. However, it remains possible that MERS-CoV-nsp5 may affect other components alongside IRF3 within the RIG-I signalling cascades.

Fig. 2.

MERS-CoV-nsp5 inhibits IFN-β production at the level of IRF3 activation or downstream of it. BEAS 2b cells were co-transfected with IFN-β reporter plasmid, TK Renilla plasmid and increasing concentrations (200 & 400 ng/ml) of HA-tagged MERS-CoV-nsp5 or EV control together with plasmids expressing (A) RIG-I, (B) MAVS, (C) IKK-ε or (D) FLAG-IRF3-5D. At 48 h post transfection, cells were harvested and assayed for luciferase activity. The expression levels of the indicated proteins were analysed by western blotting. All luciferase assay experiments were performed at least three times and all data is shown as mean ± SEM. *P < 0.05, **P < 0.01 (Student’s t-test)

Endogenous RIG-I, MAVS nor NEMO protein expression are unaffected by presence of MERS-CoV-nsp5

After verifying where MERS-CoV-nsp5 inhibits the signalling cascades through overexpression of signalling components, we then evaluated the ‘real’ effect of MERS-CoV-nsp5 on endogenous proteins. Given that other CoV nsp5 proteins have been reported to regulate them and inhibit IFN-β production, we assessed RIG-I, MAVS and NEMO (IKK-ε) [35, 36]. BEAS 2b cells were transfected with increasing concentrations (200 & 400 ng/ml) of EV control or plasmid expressing nsp5 for 48 h. The western blotting showed that the endogenous RIG-I, MAVS and NEMO expression were not affected by high concentrations of MERS-CoV-nsp5 (Fig. 3A, B, C, D, E & F). While it appeared that RIG-I protein expression was visibly reduced by low concentrations of MERS-CoV-nsp5 in the blots, upon analysis of the densitometric values, this was found to not be statistically significant (Fig. 3A & B). Additionally, no decrease was observed for MAVS and NEMO (Fig. 3C, D, E & F). These results suggest MERS-CoV-nsp5 does not target RIG-1, MAVS nor NEMO of RLR signalling pathway.

Fig. 3.

MERS-CoV-nsp5 does not affect RIG-I, MAVS nor NEMO protein expression. BEAS 2b cells were transfected with increasing concentrations (200 & 400 ng/ml) of EV control or MERS-CoV-nsp5 plasmid. At 48 h post transfection, cells were harvested for whole cell lysates and expression levels of endogenous (A) RIG-I, (C) MAVS and (E) NEMO were assessed by western blotting. Densitometry of (B) RIG, (D) MAVS and (F) NEMO was performed using Image Lab software and values for RIG, MAVS and NEMO were calculated relative to β-actin and compared to the lower concentrations of EV control. All experiments were performed at least three times and all data is shown as mean ± SEM. (Student’s t-test)

MERS-CoV-nsp5 prevents IRF3 nuclear translocation

Given previous results suggested that nsp5 antagonized IFN-β production through IRF3 or a component downstream of IRF3, we next determined the effect of nsp5 on basal IRF3 protein and its phosphorylation. BEAS 2b cells were transfected with high concentrations of EV control or MERS-CoV-nsp5 for 48 h, followed by intracellular HMW-poly(I:C) stimulation for 2 h. It was found that IRF3 was phosphorylated at comparable levels in the absence or presence of MERS-CoV-nsp5 (Fig. 4A & B). Moreover, MERS-CoV-nsp5 did not affect basal IRF3 expression (Fig. 4A & C). Collectively, this indicates that MERS-CoV-nsp5 does not affect IRF3 protein levels or HMW-poly(I:C)-mediated activation. Since MERS-CoV-nsp5 had no effect on IRF3 protein expression, nor HMW-poly(I:C)-mediated phosphorylation, we hypothesised that downstream nuclear translocation of IRF3 could be targeted by MERS-CoV-nsp5. BEAS 2b cells were transfected with EV control or MERS-CoV-nsp5 for 48 h, followed with intracellular HMW-poly(I:C) stimulation. To further test our hypothesis, we performed immunofluorescence assays. Our immunofluorescence assay showed that IRF3 translocated to the nucleus 2 h after intracellular HMW-poly(I:C) stimulation (Supplementary Fig. 1), indicating the effectiveness of the 2 h time point for visualising nuclear translocation of activated IRF3. We next found that BEAS 2b cells transfected with EV and treated with intracellular HMW-poly(I:C) had an increase in nuclear IRF3 and a decrease in cytoplasmic IRF3 compared to untreated controls at 2 h (Fig. 4D). However, in cells expressing MERS-CoV-nsp5, intracellular HMW-poly(I:C) treatment did not increase nuclear IRF3 levels (Fig. 4D). To quantify these observations, the ratio of nuclear to cytoplasmic IRF3 was determined using IMARIS software. Immunofluorescence analysis confirmed that this ratio was significantly reduced upon expression of MERS-CoV-nsp5 with intracellular HMW-poly(I:C) treatment, compared to EV transfected cells (Fig. 4E). The percentage of IRF3 translocation to the nucleus after intracellular HMW-poly(I:C) treatment was also quantified, revealing that the nuclear translocation of IRF3 was significantly inhibited by MERS-CoV-nsp5. These results suggest that MERS-CoV-nsp5 prevents IRF3 nuclear translocation. This observation was further confirmed by cytoplasm and nuclear fractionation assays in which IRF3 nuclear accumulation was significantly reduced by MERS-CoV-nsp5 expression (Fig. 4F & G). Altogether, these data indicate that MERS-CoV-nsp5 prevents HMW-poly(I:C)-mediated IRF3 nuclear translocation, which may explain the previously observed reduction of IFN-β expression.

Fig. 4.

MERS-CoV-nsp5 prevents the nuclear translocation of IRF3. (A) BEAS 2b cells were transfected with high concentration (400 ng/ml) of EV control or MERS-CoV-nsp5 plasmid. After 48 h, cells were transfected with or without 4 ug/ml HMW-poly(I:C) for 2 h. Cells were harvested for whole cell lysates and expression levels of pIRF3, IRF3 and HA were assessed by western blotting. Densitometry of (B) pIRF3 and (C) IRF3 was performed using Image Lab software and values for IRF3 and pIRF3 were calculated relative to β-actin and compared to the EV transfected without HMW-poly(I:C) control. (D) BEAS 2b cells were transfected with EV control or MERS-CoV-nsp5 plasmid. After 48 h, cells were transfected with or without 4 ug/ml HMW-poly(I:C) for 2 h (annotated as “+PIC” or “Ctrl” respectively). Cells were stained for HA, IRF3 and DAPI, and visualised using confocal microscopy. Images are representative of three independent experiments. (E) Quantification of mean IRF3 intensity in the nucleus and cytoplasm was determined using IMARIS software and the ratio of nuclear to cytoplasmic IRF3 intensity and the % of IRF3 nuclear translocation was determined. (Over the three replicate experiments there was a mean total of 36 cells in the EV control group and 6 cells in the MERS-CoV-nsp5 transfected group analysed). (F) BEAS 2b cells were transfected with higher concentration of EV control or MERS-CoV-nsp5. After 48 h, cells were transfected with or without 4 ug/ml HMW-poly(I:C) for 2 h and fractionated into cytoplasmic and nuclear fractions. The fractions were analysed by western blotting for IRF3 and HA detection. Vinculin and Lamin-B2 were used as a cytoplasmic and a nuclear marker, respectively. (G) Densitometry of nuclear IRF3 for EV or MERS-CoV-nsp5 transfected with or without HMW-poly(I:C) was performed using Image Lab software. Values for IRF3 were calculated relative to Lamin-B2 and compared to the EV transfected without HMW-poly(I:C) control. All experiments were performed at least three times and all data is shown as mean ± SEM. *P < 0.05, **P < 0.01 (Figure B, C, E left (Ratio Nuclear/Cyto IRF3) & G were analysed by Two-way ANOVA and Figure E right (%IRF3 nuclear translocation) was analysed by Student’s t-test)

MERS-CoV-nsp5 interacts with IRF3

Having observed that MERS-CoV-nsp5 impaired nuclear translocation of IRF3, without affecting protein expression nor phosphorylation of IRF3, we then investigated whether there was a physical interaction between IRF3 and MERS-CoV-nsp5, which could possibly result in the observed inhibition of IRF3 nuclear translocation. BEAS 2b cells were co-transfected with EV control or MERS-CoV-nsp5, along with or without FLAG-tagged IRF3 for 48 h. Co-immunoprecipitation assays showed that MERS-CoV-nsp5 interacted with IRF3 (Fig. 5A). Interestingly, one study has verified the interaction of SARS-CoV-2-nsp5 with IRF3 as well, but this interaction is not involved in nsp5-mediated inhibition of IFN-β expression [37]. Given MERS-CoV-nsp5 blocks IFN production, interacts with IRF3 and prevents IRF3 nuclear translocation, we wonder if it maybe targeting nuclear transport importins. Karyopherins, a class of nucleocytoplasmic transport receptors, play a pivotal role in mediating the bidirectional movement of macromolecules between the nucleus and cytoplasm [38]. Karyopherin α 1–6 (KPNA1–6) have been characterized as importing factors for nuclear translocation of cargos, including IRFs and STATs. Given that KPNA3 and KPNA4 were shown to be dominant importins involved in IRF3 and NF-κB nuclear translocation [39, 40], we hypothesised that MERS-CoV-nsp5 might target them to block IRF3 nuclear transport. Therefore, to investigate this, RT-qPCR and immunoblotting assays were conducted to measure their expression. BEAS 2b cells were transfected with high concentrations (400 ng/ml) or increasing concentrations (200 & 400 ng/ml) of EV control or MERS-CoV-nsp5 for 48 h. Results showed that MERS-CoV-nsp5 had no effect upon KPNA3 nor KPNA4 mRNA or protein expression, respectively, at either concentration (Fig. 5B, C, D, E & F). Our findings indicate that MERS-CoV-nsp5 inhibition of IRF3 nuclear translocation and IFN production is not mediated via KPNA3 and KPNA4, suggesting alternative mechanisms contribute to its interference with host immune responses.

Fig. 5.

MERS-CoV-nsp5 interacts with IRF3. (A) BEAS 2b cells were transfected with EV control or MERS-CoV-nsp5 along with or without FLAG-IRF3. After 48 h, cells were lysed and immunoprecipitated using FLAG antibody. IP and WCL were subject to Western blotting using HA, FLAG and β-actin antibodies. (B) BEAS 2b cells were transfected with higher concentration (400 ng/ml) of EV control or MERS-CoV-nsp5 plasmids. After 48 h, total RNA was extracted and the expression of KPNA3 and KPNA4 was detected by real-time RT-qPCR. Gene expression was normalised to house-keeping gene RSP15 and compared to the EV control. (C) BEAS 2b cells were transfected with increasing concentrations (200 & 400 ng/ml) of EV control or MERS-CoV-nsp5. At 48 h post transfection, cells were harvested for whole cell lysates and expression levels of endogenous KPNA4 were assessed by western blotting. (E) BEAS 2b cells were transfected with high concentrations (400 ng/ml) of EV control or MERS-CoV-nsp5. At 48 h post transfection, cells were harvested for whole cell lysates and expression levels of endogenous KPNA3 were assessed by western blotting. Densitometry of (D) KPNA4 and (F) KPNA3 was performed using Image Lab software and values for KPNA4 and KPNA3 were calculated relative to β-actin and compared to the EV control (NB in Figure D they were compared to the 200 ng/ml of EV control. All experiments were performed at least three times and all data is shown as mean ± SEM. ns = not significant (Student’s t-test)

MERS-CoV-nsp5 inhibits IFN induction depending on its enzymatic activity

Since we saw that IFN-β was reduced by MERS-CoV-nsp5 expression in BEAS 2b cells without affecting KPNA3 and KPNA4, we hypothesised the involvement of alternative mechanisms. As nsp5 proteins of CoV are responsible for cleaving viral polyproteins during viral replication, we subsequently investigated the potential involvement of the enzymatic activity of MERS-CoV-nsp5 in IFN regulation. To examine this, we tested two 3 C like protease inhibitors, Calpain inhibitor II and Rupintrivir inhibitor, which have been proven to inhibit the cysteine protease activity. Calpain inhibitor II and Rupintrivir inhibitor have been reported to efficiently inhibit SARS-CoV-2-nsp5 main protease activity and viral replication [15, 41, 42]. BEAS 2b cells were co-transfected with EV control or MERS-CoV-nsp5 along with IFN-β promoter-driven luciferase reporter plasmid (IFN-β-Luc) and control RL-TK plasmid. After 24 h, cells were transfected with HMW-poly(I:C) for 24 h and treated with 10 μm Calpain inhibitor II or Rupintrivir inhibitor for 18 h. Since the proteasome was not thought to be involved in this protein-driven immune evasion mechanism, MG132, a well-characterized proteasome inhibitor that prevents the degradation of ubiquitin-conjugated proteins [43], was included as a negative control to elucidate the mechanism of IFN-β suppression by MERS-CoV-nsp5. Inhibiting the enzyme activity showed that Calpain inhibitor II and Rupintrivir inhibitor efficiently blocked reduction of IFN-β promoter activity by MERS-CoV-nsp5 (Fig. 6A & B). This result was not mirrored in the negative control MG132, which failed to block antagonistic activity of MERS-CoV-nsp5 on IFN-β promoter activity triggered by HMW-poly(I:C) (Fig. 6C). The use of MG132 allowed us to determine that the observed suppression of IFN-β was not attributable to proteasome-mediated protein degradation. These data suggests that MERS-CoV-nsp5 can suppress IFN-β activation in an enzymatic activity dependent manner. After checking promoter activation, we then examined whether the IFN-β cytokine production levels were restored as well. BEAS 2b cells were co-transfected with EV control or MERS-CoV-nsp5. After 24 h, cells were transfected with HMW-poly(I:C) for 24 h and treated with 10 μm Calpain inhibitor II, Rupintrivir inhibitor or MG132 negative control for 18 h. Similarly, it is found that Calpain inhibitor II and Rupintrivir inhibitor, but not MG132, significantly restored the IFN-β production suppressed by MERS-CoV-nsp5 (Fig. 6D, E & F). Collectively, these data show that MERS-CoV-nsp5 can suppress IFN-β activation in an enzymatic activity-dependent manner.

Fig. 6.

MERS-CoV-nsp5 inhibits IFN induction depending on its enzymatic activity. (A-C) BEAS 2b cells were co-transfected with IFN-β reporter plasmid, TK Renilla plasmid and MERS-CoV-nsp5 or EV control plasmid. At 24 h post transfection, cells were transfected with 4ug/ml HMW-poly(I:C) for another 24 h and treated with 10 μm Calpain II inhibitor, 10 μm Rupintrivir or 10 μm MG132 for a further 18 h and subsequently assayed for luciferase activity. (D-F) BEAS 2b cells were transfected with EV control or MERS-CoV-nsp5 plasmid. At 24 h post transfection, cells were transfected with or without 4 ug/ml HMW-poly(I:C) for another 24 h and treated with 10 μm Calpain inhibitor II, 10 μm Rupintrivir or MG132 for another 18 h and supernatants were harvested and assayed for IFN-β production. The levels of indicated proteins (HA and β-actin) were assessed by western blotting. All experiments were performed at least three times and all data is shown as mean ± SEM. **P < 0.01, ***P < 0.001 ****P < 0.0001 (Two-way ANOVA)

3D protein-protein modelling shows interaction between MERS-CoV-nsp5 and IRF3

To further investigate the interaction seen between MERS-CoV-nsp5 and IRF3, protein-protein interaction analyses were carried out by taking into account molecular docking predictions, followed by detailed MD simulations to check the stability of the predicted poses. We used three docking protocols including Molecular Operating Environment (MOE), ClusPro and ZDOCK to identify interaction sites between the two proteins. Different interaction sites were identified on each protein with a high degree of overlap between the sites. However, there was no consensus between the docked poses identified from the three docking protocols. But interestingly there was a common region on MERS-CoV-nsp5 (Fig. 7A), where IRF3 was predicted to bind by MOE and ClusPro protocols. We then investigated the stability of the top 6 docked poses from MOE and the best pose from ClusPro by performing MD simulations. From the MD simulations, we calculated hydrogen bonds at the interface of MERS-CoV-nsp5 and IRF3, RMSD of the chains from their initial position and the RMSD obtained by structurally overlaying the most representative frame from MD simulations over the predicted docked poses. We observed that the most stable conformations were pose 4 and pose 6 (Fig. 7B and C) from the MOE docking protocol considering the greater number of hydrogen bonds, low RMSD fluctuation between the chains and better structural overlays between MD and docking output structures.

Fig. 7.

MERS-CoV-nsp5 suggested binding to IRF3. (A) Common region on MERS-CoV nsp5 highlighted in orange where IRF3 is predicted to bind by MOE and ClusPro docking protocols. (PDB: 8E6B, dark green). Most stable MOE predicted docked poses (B) pose 4 and (C) pose 6 highlighting the NLS region amino acids in red between MERS-CoV-nsp5 (PDB: 8E6B dark green) and IRF3 (PDB: 3QU6 yellow). (D) Number of hydrogen bonds maintained throughout the MD run between MERS-CoV nsp5 and IRF3 on pose 4 in MD simulation analysis. (E) RMSD fluctuation graph for pose 4 (distance: Angstrom Å, time: nanoseconds ns) (F) Structural overall between the representative frame from MD (MERS-CoV-nsp5: orange, IRF3: pink) and docked pose 4 (MERS-CoV-nsp5: dark green, IRF3: yellow). (G) Number of hydrogen bonds maintained throughout the MD run between MERS-CoV nsp5 and IRF3 on pose 6 in MD simulation analysis. (H) RMSD fluctuation graph for pose 6 (distance: Angstrom Å, time: nanoseconds ns). (I) Structural overall similarity between the representative frame from MD (MERS-CoV-nsp5: orange, IRF3: pink) and docked pose 6 (MERS-CoV-nsp5: dark green, IRF3: yellow). Contact analysis of (J) pose 4 showing presence of hydrogen bonds (inset) between MERS-CoV-nsp5 amino acids Glu21, Leu151, Lys147, Glu295, Met312, Gln313 and IRF3 amino acids Lys5, Arg7, Asp17, Leu18, Ala60, Ser112. (K) pose 6 showing presence of hydrogen bonds (inset) between MERS-CoV-nsp5 amino acids Met308, Gln313 and IRF3 Arg43, Arg78. Protein-protein interface contacts are highlighted in red. MERS-CoV-nsp5 (PDB: 8E6B) is shown in dark green and IRF3 (PDB: 3QU6) in yellow. The NLS region amino acids are highlighted in red (inset)

The two poses showed different conformations of IRF3 binding, in pose 4 we observed that the NLS region was not interacting with MERS-CoV-nsp5 whereas, in pose 6 the NLS region was masked by nsp5. Moreover, from the MD simulations of pose 4 it was observed that throughout the 100 ns MD production run there was an average of 8 hydrogen bonds between the MERS-CoV-nsp5 and IRF3 interface (Fig. 7D). The RMSD fluctuations of the two protein chains from their initial co-ordinates were very low, suggesting that the pose is a stable conformation (Fig. 7E). Lastly, the structural overlay between the most representative frame from the MD run and the MOE predicted docked pose gave an RMSD value of 4.25 Å (Fig. 7F). Similarly, from the MD simulations of pose 6 we observed that throughout the MD production run there was an average of 7 hydrogen bonds between the MERS-CoV-nsp5 and IRF3 interface (Fig. 7G). The RMSD fluctuations of the two protein chains from their initial coordinates was very low suggesting that the pose is a stable conformation (Fig. 7H). Lastly, the structural overlay between the most representative frame from the MD run and the MOE predicted docked pose gave an RMSD value of 4.24 Å (Fig. 7I).

The MOE contact analysis tool was used to identify the interactions between the MERS-CoV-nsp5 and IRF3 interface. The representative frame from the MD runs of the two poses was used to identify different types of interactions (pose 4: 52 contacts, pose 6: 43 contacts) at the protein-protein interface (Supplementary Tables 1 and 2). The contact analysis revealed the presence of possible PPI hot spot hydrogen bonds [44] that are highlighted in Fig. 7J and K.

3D protein-protein modelling shows interaction between MERS-CoV-nsp5 and KPNA4

Having identified the interaction sites between MERS-CoV-nsp5 and IRF3, we next investigated the potential interaction between MERS-CoV-nsp5 and KPNA3/4. However, KPNA3 was not included in the molecular modelling studies as there were no crystal structures available and the AlphaFold [45] generated structure did not appear to be suitable for inclusion in our workflow. The contacts made between MERS-CoV-nsp5 and KPNA4 were investigated using protein-protein interaction predictions made from the MOE, ClusPro and ZDOCK outputs. We observed that ClusPro and ZDOCK docking protocols predicted a similar docked pose (Fig. 8A), indicating a preferred conformation whereas, MOE predicted a different docked pose. Surprisingly, the docked poses indicated that KPNA4 binds at the same region on MERS-CoV-nsp5 as the IRF3. In order to confirm the stability of the docked pose predicted by ClusPro and ZDOCK, we ran 100 ns MD simulation. It was observed that throughout the MD production run there was an average of 4 hydrogen bonds between MERS-CoV-nsp5 and KPNA4 interface (Fig. 8B). The RMSD fluctuations of the two protein chains from their initial co-ordinates was very low suggesting that the pose is a stable conformation (Fig. 8C). Lastly, the structural overlay between the most representative frame from the MD run and the ClusPro/ZDOCK docked pose gave an RMSD value of 3.19 Å (Fig. 8D). The MOE contact analysis tool was used to identify the interactions between the MERS-CoV-nsp5 and KPNA4 interface. The representative frame from the MD run was used to identify different types of interactions (42 contacts) at the protein-protein interface (Supplementary Table 3). The contact analysis revealed the presence of possible PPI hot spot hydrogen bonds that are highlighted in Fig. 8E. Collectively, our analysis identified protein conformations that involve masking of the NLS region of IRF3 and KPNA4 with MERS-CoV-nsp5.

Fig. 8.

MERS-CoV-nsp5 suggested binding to KPNA4. (A) Best protein-protein docked pose of MERS-CoV-nsp5 (PDB: 8E6B dark green) and KPNPA4 (PDB: 7LFC light green) predicted by ClusPro and ZDOCK docking protocols and the NLS binding site amino acids highlighted in red. (B) Number of hydrogen bonds maintained throughout the MD run between MERS-CoV-nsp5 and KPNA4 in MD simulation analysis on ClusPro/ZDOCK predicted docked pose. (C) RMSD fluctuation graph in MD simulation analysis on ClusPro/ZDOCK predicted docked pose. (distance: Angstrom Å, time: nanoseconds ns) (D) Structural overall similarity between the representative frame from MD (MERS-CoV-nsp5: orange, KPNA4: pink) and docked pose (MERS-CoV-nsp5: dark green, KPNA4: light green). (E) Contact analysis of ClusPro/ZDOCK predicted docked pose showing presence of hydrogen bonds (inset) between MERS-CoV-nsp5 amino acids Gly148, Leu223, Gln306 and KPNA4 amino acids Arg96, Arg103, Pro213, Ile214. Contacts highlighted in red at the protein-protein interface between MERS-CoV-nsp5 (PDB: 8E6B dark green) and KPNA4 (PDB: 7LFC, light green) and NLS binding site amino acids highlighted in red

Discussion

The emergence of three deadly coronaviruses in as many decades and their subsequent global spread has highlighted the urgent need for a comprehensive understanding of molecular mechanisms employed by coronaviruses to evade the host immune response. Although several virus–host protein interactome analyses have been performed with MERS-CoV factors [46, 47], none of them have identified the components of the type I IFN signalling pathway as targets of MERS-CoV-nsp5. In this study, MERS-CoV-nsp5 was identified as a novel IFN antagonist that specifically targeted the IFN-inducing IRF3. It was found that MERS-CoV-nsp5 had no effect on IRF3 levels nor its phosphorylation, but instead specifically interacted with IRF3 and hinders the movement of IRF3 into the nucleus. Moreover, our results revealed that MERS-CoV-nsp5 potently inhibited IFN induction by double-stranded RNA (dsRNA) in an enzyme dependent manner.

Similar immune evasion tactics were previously demonstrated for other CoV encoded nsp5 proteins. Studies have implicated SARS-CoV-2-nsp5 to inhibit the host IFN response via multiple mechanisms, including SARS-CoV-2-nsp5 been shown to cleave and inactivate key components of the host immune response, such as the RIG-I and MAVS, which are crucial for inducing IFN production [35]. Other studies have demonstrated that SARS-CoV-2-nsp5 antagonizes IFN production by retaining phosphorylated IRF3 in the cytoplasm and interacting with IRF3 [20, 37]. Due to the high similarity between MERS-CoV-nsp5 and SARS-CoV-2-nsp5, we postulated that MERS-CoV-nsp5 may antagonize the IFN response similarly. To investigate the role of MERS-CoV-nsp5 in the IFN response, HMW-poly(I:C) was used to stimulate the RIG-I and MDA5 signalling pathways in human bronchial epithelial BEAS 2b cells and it is found that HMW-poly(I:C) potently activated the RIG-I and MDA5 pathway, as evidenced by IFN-β induction. However, with the presence of increasing concentrations of MERS-CoV-nsp5 we saw significantly reduced IFN-β induction. Interestingly, our study observed that high concentrations of EV significantly enhanced the HMW-poly(I:C)-induced expression of IFN-β (Fig. 1B). This increased IFN-β induction may be as a result of EV-DNA detection by cellular sensors, such as cyclic GMP-AMP synthase (cGAS). cGAS recognizes cytosolic DNA and activates the cGAS-STING pathway, which subsequently triggers the activation of transcription factors IRF3 and NF-κB [48]. Therefore, it is possible that high EV concentrations trigger this pathway, leading to an enhanced induction of IFN-β. However, we observed that IFN-β induction was significantly reduced when high concentrations of MERS-CoV-nsp5 were present, highlighting the inhibitory effect of MERS-CoV-nsp5 on IFN-β induction. While the low concentrations of MERS-CoV-nsp5 did not show inhibition in IFN-β induction, this may be attributed to a threshold effect, where the amount of nsp5 is insufficient to reach the critical level required to inhibit IRF3 nuclear translocation. Indeed, it is commonly observed that viral proteins modulate innate immune signalling cascades in a dose-dependent manner [20, 49, 50].

Collectively our observations suggest that MERS-CoV-nsp5 subverts dsRNA-induced innate immune activation. Further biochemical assays showed MERS-CoV-nsp5 targeted signalling at IRF3 or downstream levels. Specifically, the IRF3 nuclear translocation was impaired in response to HMW-poly(I:C) in MERS-CoV-nsp5 expressing cells. It was further observed that MERS-CoV-nsp5 had no effect upon the endogenous protein expression of RIG-I, MAVS and NEMO (IKK-ε). In contrast to our present finding, a distinct observation was reported regarding SARS-CoV-2-nsp5, wherein it was demonstrated to induce degradation of MAVS protein expression in HEK293T cells [35]. This phenomenon was not observed in our investigation of MERS-CoV-nsp5 in BEAS 2b cells, which may be due to different cell types and distinctions between the two viral nsp sequences. Indeed, recent research found that SARS-CoV-2-ORF6 did not antagonize IRF3 translocation and IFN production in respiratory Calu-3 cells [51], whereas a previous study showed ORF6 blocked IRF3 translocation and IFN induction in HEK293T cells [52]. This highlights the impact of the cellular context on the observed results. Indeed, respiratory cells play a pivotal role in CoV infections, acting as early targets for viral entry and replication. They constitute a primary locus of infection within the respiratory tract [53]. While HEK293T cells, which are embryonic kidney cells, are widely used for in vitro molecular analysis in SARS-CoV-2 studies [20, 35, 54], BEAS 2b cells, derived from human bronchial epithelium, provide a physiologically relevant model for investigating respiratory infectious diseases and innate immune responses of the respiratory tract. Additionally, BEAS 2b cells have demonstrated a superior intrinsic antiviral state due to high-level expression of antiviral ISGs, pattern recognition receptors and other signaling intermediaries, such as RIG-I and IRF9 [55], further highlighting their usefulness in studying anti-viral molecular immune responses.

We further checked if MERS-CoV-nsp5 affected the KPNA3 and KPNA4 expression which belong to the karyopherin alpha protein family, that are essential for mediating transcription factors, including IRF3 and NF-κB, into the nucleus [39, 40, 56]. Here our results showed that KPNA3 and KPNA4 expression were not affected by MERS-CoV-nsp5. We also investigated if MERS-CoV-nsp5 interacts with IRF3. Intriguingly, it was found that MERS-CoV-nsp5 specifically associated with IRF3. However, it remains unclear how this interaction inhibited IRF3 nuclear translocation, and the underlying mechanism warrants further investigation. It is possible that MERS-CoV-nsp5 directly binds to phosphorylated IRF3, hindering its dimerization or recognition by nuclear transport receptors [57, 58]. This mechanism could be similar to how Pestivirus N-terminal protease (Npro) directly interacts with both the monomeric and dimeric forms of IRF3 [59]. Consequently, this binding could impede the essential conformational changes needed for nuclear import, resulting in the sequestration of IRF3 in the cytoplasm [59]. Notably, one study verified the interaction of SARS-CoV-2-nsp5 with IRF3, but this interaction was not involved in nsp5-mediated inhibition of IFN-β expression [37]. As described above, another possibility is that nsp5 interferes with the function of nuclear transport proteins, such as KPNA3/4. Although their expression is unaffected, nsp5 may disrupt their function or their ability to interact with IRF3. This interference could involve post-translational modifications of KPNA3/4 or competitive binding by nsp5. Indeed, such a mechanism is used by the Japanese Encephalitis Virus (JEV), whereby its NS5 protein competitively blocks the interaction of KPNA3/4 with IRF3 by directly interacting with KPNA3/4 [39]. Moreover, given that nsp5 is a protease, it might cleave specific cellular proteins other than KPNA3/4, which are essential for IRF3’s nuclear translocation. In fact, it has been found that USP22, a cytoplasmic and nuclear deubiquitinating enzyme, promotes nuclear translocation of IRF3 by deubiquitianting and stabilizing KPNA2 after viral infection, suggesting that additional proteins could be involved in the nuclear translocation process [60]. This proteolytic activity could selectively cleave components of the nuclear import pathway, thereby preventing IRF3 from reaching the nucleus. Therefore, further studies will be required to determine the impact of this interaction.

To further determine plausible binding modes between MERS-CoV-nsp5-IRF3 and MERS-CoV-nsp5-KPNA4, we performed protein-protein docking analysis. Proteins exist in numerous conformations, but the docking protocol only extracts a low energy favoured rigid structure. To look at the dynamic behaviour of the protein-protein interactions we ran 100 ns MD runs. In our molecular modelling protocol, we observed the interactions focussed on most stable protein conformations. The stable conformations of protein-protein interactions were confirmed by observing the RMSD fluctuations from the initial co-ordinates, the number of hydrogen bonds at the interface and, lastly, structurally overlaying the most representative frame from the MD run with the predicted docked pose. From the molecular modelling protocol, we predicted that MERS-CoV-nsp5 might interact with two different stable conformations of IRF3. One of the conformations (pose 4) does not involve the amino acids of IRF3 NLS region, whereas the other conformation (pose 6) results in masking of the NLS region by nsp5, thus potentially impacting its nuclear translocation. The study also indicated the possibility of MERS-CoV-nsp5 interacting with the NLS binding site on KPNA4. Therefore, we hypothesise that the translocation of IRF3 might be impacted by the masking of the NLS region either on IRF3 (pose 6) or on KPNA4, by nsp5. Another possibility is the interaction of IRF3 (pose 4) with MERS-CoV-nsp5, which may block the subsequent interactions of IRF3 with the importins. However, if we compare the average number of hydrogen bonds present between these interactions, the preferred interaction is MERS-CoV-nsp5 with IRF3. The presence of hydrogen bonds, arene and ionic interactions between nsp5-IRF3 and nsp5-KPNA4 could be key hot spot PPIs, which should be explored in future mutagenesis studies to confirm their involvement.

The enzyme activity of CoV nsp5 proteins not only influences viral replication but also modulates the host immune response. CoV nsp5 proteins subvert the host’s antiviral defense by breaking down innate immune signalling factors like MAVS and NEMO, as seen with SARS-CoV-2 and Porcine Deltacoronavirus, respectively [35, 36]. This immunomodulatory role contributes to viral persistence and evasion of host immune surveillance. Since the SARS-CoV-2 outbreak, a series of inhibitors has been reported against the main protease of SARS-CoV-2 that prevent viral replication [41, 61]. Rupintrivir is a 3CL protease inhibitor that prevents CoV and human rhinovirus replication through binding to the protease enzyme [62]. Calpain inhibitor II is capable of inhibiting human cathepsin L which plays an essential role in SARS-CoV-2 cell entry by activating the viral spike protein, and was later characterized to bind and inhibit 3CL protease [61, 63]. In this study, we tested two main protease inhibitors of nsp5, Calpain inhibitor II and Rupintrivir inhibitor, to investigate if the inhibition of IFN production was dependent on the enzyme activity of MERS-CoV-nsp5. Intriguingly, the enzyme activity of MERS-CoV-nsp5 was required to suppress type I IFN production. Previous studies have shown that coronavirus nsp5 possesses catalytic activity to cleave host immune-related proteins. For example, PEDV (Alphacoronavirus) nsp5 and PDCoV (Deltacoronavirus) nsp5 cleaved NEMO to inhibit type I IFN production [36, 64]. Taken together, these findings indicate that the MERS-CoV main protease, nsp5, plays catalytic roles in the host immune response. In addition, this study identified that two broad-spectrum main protease inhibitors exhibited antiviral activity against MERS-CoV-nsp5, suggesting them as potential therapeutics against MERS-CoV infection. Future work should aim to evaluate the mechanism underlying nsp5-mediated inhibition of IRF3 nuclear translocation. Reconstitution assays using known inhibitors of MERS-CoV-nsp5, in combination with HMW-poly(I:C) stimulation, could also be conducted to confirm whether inhibiting the enzymatic activity of MERS-CoV-nsp5 can restore IRF3 nuclear translocation, thereby further elucidating the mechanistic pathway involved.

It should be noted that our study utilized overexpression approaches, which are common experimental methods, but come with certain limitations. Transformed cell lines often exhibit genetic and phenotypic changes that affect their immune responses, such as defects in signalling pathways [65], altered immune receptor expression, or dysregulated cytokine production [66, 67]. However, despite these drawbacks, overexpression systems are invaluable in biological research. They enable the functional characterization of genes, the study of protein-protein interactions, and the analysis of signaling pathways [68]. They also facilitate drug screening and therapeutic target validation, providing crucial insights that drive scientific and medical advancements [69]. That said, it will be essential for future research to complement these overexpression approaches using models that include MERS-CoV infection studies, human bronchial primary cells and in vivo models [70].

In summary, this study demonstrated that MERS-CoV-nsp5 attenuates type I IFN activation by inhibiting IRF3 nuclear translocation. This inhibition on IFN activity is dependent on nsp5 enzyme activity. Elucidating the mechanisms by which nsp5 affects the IFN response contributes to our understanding of the pathogenesis of CoV infections and identification of potential therapeutic targets.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Yamie Zhang and Shubhangi Kandwal. The first draft of the manuscript was written by Yamie Zhang and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Funding

This work was supported by Chinese Scholarship Council (Award No. 201908300032), and Science Foundation Ireland, grant numbers SFI 20/SPP/3685 and SFI 19/FFP/6483. Part of the research conducted in this publication was funded by the Irish Research Council under grant number GOIPG/2021/954.

Data availability

The datasets generated during and/or analysed during the current study are not publicly available, but are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors have no relevant financial or non-financial interests to disclose.