Proteolytic Activities of Enterovirus 2A Do Not Depend on Its Interaction with SETD3

Enterovirus 2Apro is a protease that proteolytically processes the viral polyprotein and cleaves several host proteins to antagonize host responses during enteroviral infection. Recently, the host protein actin histidine methyltransferase SET domain containing 3 (SETD3) was identified to interact with 2Apro and to be essential for virus replication. The role of SETD3 and its interaction with 2Apro remain unclear. In this study, we investigated the potential involvement of SETD3 in several functions of 2Apro. For this, we introduced the 2Apro from coxsackievirus B3 (CVB3) in a mutant of encephalomyocarditis virus (EMCV) containing an inactivated Leader protein (EMCV-Lzn) that is unable to shut down host mRNA translation, to trigger nucleocytoplasmic transport disorder (NCTD), and to suppress stress granule (SG) formation and type I interferon (IFN) induction. Both in wt HeLa cells and in HeLa SETD3 knockout (SETD3KO) cells, the virus containing active 2Apro (EMCV-2Apro) efficiently cleaved eukaryotic translation initiation factor 4 gamma (eIF4G) to shut off host mRNA translation, cleaved nucleoporins to trigger NCTD, and actively suppressed SG formation and IFN gene transcription, arguing against a role of SETD3 in these 2Apro-mediated functions. Surprisingly, we observed that the catalytic activity of enteroviral 2A is not crucial for triggering NCTD, as a virus containing an inactive 2Apro (EMCV-2Am) induced NCTD in both wt and SETD3KO cells, albeit delayed, challenging the idea that the NCTD critically depends on nucleoporin cleavage by this protease. Taken together, our results do not support a role of SETD3 in the proteolytic activities of enterovirus 2Apro.


Introduction
The family Picornaviridae comprises a large group of small non-enveloped positivesense single-stranded RNA viruses. The picornaviral genome is about 7-8.5 kb in length, which is composed of a single open reading frame and two untranslated regions (UTRs) at the 5 and 3 ends. During infection, the viral genome is rapidly translated into a single polyprotein harboring the P1 region capsid proteins and the P2 and P3 region non-structural proteins. This polyprotein is proteolytically processed into mature viral proteins by viral proteases [1].
The picornaviruses family include important pathogens, causing a wide range of disease in humans and animals. The genus Enterovirus is one of the many genera of this family. This genus can be further divided into 15 species, Enterovirus A-L and Rhinovirus A-C. Among these species, numerous serotypes have been identified [2]. Well-known enteroviruses are poliovirus, coxsackievirus, and echovirus. Enteroviruses produce three viral proteases 2A pro , 3CD pro , and 3C pro , which process the viral polyprotein into mature tease [28]. We previously showed that heterologous expression of CVB3 2A pro in an EMCV recombinant (EMCV-2A pro ) ( Figure 1) [18], could functionally substitute for an inactive L protein containing mutations in its zinc-finger domain (L zn ) [29,30]. Using EMCV-2A pro , we here show that the functions of 2A pro to cleave eIF4G, trigger NCTD, and suppress SG formation and IFN induction do not rely on SETD3. Notably, during our studies we observed that catalytically inactive 2A pro could still induce NCTD, albeit it was delayed, indicating that the virus-induced NCTD does not critically rely on the proteolytic activity of 2A pro .
Viruses 2022, 14, x FOR PEER REVIEW 3 of 15 formation and IFN induction [27], although the underlying mechanisms differ as L is not a protease [28]. We previously showed that heterologous expression of CVB3 2A pro in an EMCV recombinant (EMCV-2A pro ) ( Figure 1) [18], could functionally substitute for an inactive L protein containing mutations in its zinc-finger domain (L zn ) [29,30]. Using EMCV-2A pro , we here show that the functions of 2A pro to cleave eIF4G, trigger NCTD, and suppress SG formation and IFN induction do not rely on SETD3. Notably, during our studies we observed that catalytically inactive 2A pro could still induce NCTD, albeit it was delayed, indicating that the virus-induced NCTD does not critically rely on the proteolytic activity of 2A pro .

Cells and Viruses
H1 Hela and H1 SETD3 KO cell lines were kind gifts from Jan E. Carette's group. All Hela and Hela-derived cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serums and a mixture of penicillin and streptomycin. Recombinant EMCV viruses were derived from the pM16.1 infectious clone [31]. In EMCV-L zn , the stress and IFN antagonist (i.e., the Leader protein) is inactivated by the insertion of point mutations in its Zn finger domain (C19A/C22A). The genes encoding CVB3 2A pro or 2A m (mutation C109A) were introduced at the 5′ end of the EMCV [18]. The recombinant EMCV viruses mentioned in this study were recovered by transfection of RNA transcripts into BHK21 cells. The viruses were passaged on BHK21 cells. Once total CPE, cells were subjected to three freeze-thaw cycles and debris was pelleted at 4000× g for 15 min. Subsequently, the viruses were concentrated by ultracentrifugation through 30% sucrose cushion at 250,000× g for 16 h in an SW32Ti rotor and stored at −80 °C [18].

Cells and Viruses
H1 Hela and H1 SETD3 KO cell lines were kind gifts from Jan E. Carette's group. All Hela and Hela-derived cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serums and a mixture of penicillin and streptomycin. Recombinant EMCV viruses were derived from the pM16.1 infectious clone [31]. In EMCV-L zn , the stress and IFN antagonist (i.e., the Leader protein) is inactivated by the insertion of point mutations in its Zn finger domain (C19A/C22A). The genes encoding CVB3 2A pro or 2A m (mutation C109A) were introduced at the 5 end of the EMCV [18]. The recombinant EMCV viruses mentioned in this study were recovered by transfection of RNA transcripts into BHK21 cells. The viruses were passaged on BHK21 cells. Once total CPE, cells were subjected to three freeze-thaw cycles and debris was pelleted at 4000× g for 15 min. Subsequently, the viruses were concentrated by ultracentrifugation through 30% sucrose cushion at 250,000× g for 16 h in an SW32Ti rotor and stored at −80 • C [18].

Quantitative Real-Time PCR (qRT-PCR Analysis)
Cells were seeded in 24-well plates and the next day infected with the indicated viruses at an MOI of 10. At the indicated time points, the cells were lysed and total RNA from cells was isolated using a total RNA isolation kit (Macherey-Nagel, Düren, Germany) according to the manufacturer's instructions. Reverse transcription was set up using TaqMan reverse transcription reagents (Applied Biosystems, California, USA) before performing qPCR analysis with SYBR green (Roche, Basel, Switzerland). Data shown are the relative abundance of the indicated mRNA normalized to actin. The primers for IFN detection are: Fw_5 -ATGACCAACAAGTGTCTCCTCC-3 ; Rv_5 -GCTCATGGAAAGAGCTGTAGTG-3 . The primers for actin detection: Fw_5 -CCTTCCTGGGCATGGAGTCCTG-3 ; Rv_5 -GGAGCAATGATCTTGATCTTC-3 .

Western Blot Analysis
The WB assay was performed as described previously [18]. Briefly, cells were seeded in 6-well plates and the next day infected with the EMCV viruses indicated above at MOI of 10. At the indicated time points, cells were released using trypsin, washed and cell lysates were prepared. 100 micrograms of protein from the cleared cell lysates were resolved using SDS-PAGE (Bio-Rad, Irvine, CA, USA), and the proteins were transferred to 0.2 µm-poresize nitrocellulose membranes. The membranes were incubated overnight with primary antibody and for 30 min with the secondary antibody. The membranes were scanned using an Odyssey imager (LiCOR, Lincoln, NE, USA).

Immunofluorescence Analysis
IFA was performed as described previously [18]. Briefly, cells were grown on 12-mm glass coverslips and the next day infected with the viruses indicated above (MOI = 10 or 50). At the indicated time points, cells were fixed in 4% paraformaldehyde for 30 min and then permeabilized using 0.1% Triton X-100. Samples were incubated with primary antibody for 1 h and incubated with secondary antibodies and DAPI for 30 min. Coverslips were mounted on microscopy slides with FluorSafe TM reagent (Merck Millipore, Massachusetts, USA). Finally, cells were examined by confocal microscopy (Leica SPE-II, Wetzlar, Germany) and Leica Application Suite Advanced Fluorescence (LAS-AF, Wetzlar, Germany) software.

Quantification of Cells with SGs
After staining of the cells with the antibodies against proteins of interest, 250 cells were counted for quantifying the ratio of cells with SGs in each sample. GraphPad Prism 7 software was used for statistical analysis. Error bars represent standard deviations. One-way analysis of variance (ANOVA) analysis was used to calculate p values.

Replication of EMCV-2A pro in WT and SETD KO Cells
To verify that CVB3 is unable to infect SETD3 KO cells, we infected both WT Hela cells and SETD3 KO cells with CVB3-Rluc viruses at an MOI of 0.1. Then, Renilla luciferase activity was tested at 2, 4, 6, and 8 hpi. CVB3 infection was severely impaired in SETD3 KO cells ( Figure S1), as shown previously [26]. Therefore, EMCV recombinant viruses carrying 2A pro were used to study the importance of SETD3 for 2A pro function. The construction and characterization of these viruses have been described previously [18]. Here, we set out to ensure that the recombinant EMCV-2A pro viruses (wt 2A pro and catalytically inactive mutant 2A m ) could replicate well in the absence of SETD3. For this, single-cycle growth curves were determined in HeLa wt and SETD KO cells. Both cell lines were infected with EMCV-wt, EMCV-L zn , EMCV-2A pro , and EMCV-2A m , a virus that contains inactivated 2A protein due to mutation C109A at an MOI of 5. Virus titers were determined at 2, 4, 6, and 8 h post infection (hpi), respectively, by endpoint titration (Figure 2). In both cell lines, EMCV-wt showed most efficient replication, whereas EMCV-L zn showed delayed replication. EMCV-2A pro also showed reduced replication but replicated slightly faster than EMCV-L zn and EMCV-2A m . Similar virus replication kinetics were observed in WT and SETD3 KO cells. Therefore, these viruses represent a good model to investigate the function(s) of CVB3 2A pro and the dependence on SETD3.  Figure 2). In both cell lines, EMCV-wt showed most efficient replication, whereas EMCV-L zn showed delayed replication. EMCV-2A pro also showed reduced replication but replicated slightly faster than EMCV-L zn and EMCV-2A m . Similar virus replication kinetics were observed in WT and SETD3 KO cells. Therefore, these viruses represent a good model to investigate the function(s) of CVB3 2A pro and the dependence on SETD3.

CVB3 2A pro Expression and eIF4GI Cleavage in SETD3 KO Hela Cells
Next, we tested whether the cleavage of eIF4GI by enteroviral 2A pro , which occurs early during infection, is dependent on SETD3. WT and SETD3 KO cells were infected with EMCV-2A pro at an MOI of 10 and total cellular proteins were collected at 2, 4, 6, and 8 hpi. Additionally, cells infected with EMCV-2A m at 6 hpi were included as a control. Western blot analysis was performed to detect eIF4GI cleavage and 2A expression (Figure 3). The molecular weight of intact eIF4GI is 220 kDa, and that of cleavage products are about 100 and 120 kDa, respectively [32]. In both cell lines, the infection of EMCV-2A m did not result in the cleavage of eIF4GI (upper part). On the contrary, the cleavage of eIF4GI was evident already at 2 hpi in EMCV-2A pro -infected cells regardless of cell type. With time increasing, more eIF4GI cleavage products were observed, but little if any difference was observed between WT and SETD3 KO cells. The absence of SETD3 had no evident effect on the expression and stability of CVB3 2A pro (lower part) as comparable amounts of 2A pro were detected at 6 and 8 hpi in WT and SETD3 KO cells. Thus, SETD3 is not required for the CVB3 2A pro -mediated cleavage of eIF4GI during enteroviral infection.

CVB3 2A pro Expression and eIF4GI Cleavage in SETD3 KO Hela Cells
Next, we tested whether the cleavage of eIF4GI by enteroviral 2A pro , which occurs early during infection, is dependent on SETD3. WT and SETD3 KO cells were infected with EMCV-2A pro at an MOI of 10 and total cellular proteins were collected at 2, 4, 6, and 8 hpi. Additionally, cells infected with EMCV-2A m at 6 hpi were included as a control. Western blot analysis was performed to detect eIF4GI cleavage and 2A expression (Figure 3). The molecular weight of intact eIF4GI is 220 kDa, and that of cleavage products are about 100 and 120 kDa, respectively [32]. In both cell lines, the infection of EMCV-2A m did not result in the cleavage of eIF4GI (upper part). On the contrary, the cleavage of eIF4GI was evident already at 2 hpi in EMCV-2A pro -infected cells regardless of cell type. With time increasing, more eIF4GI cleavage products were observed, but little if any difference was observed between WT and SETD3 KO cells. The absence of SETD3 had no evident effect on the expression and stability of CVB3 2A pro (lower part) as comparable amounts of 2A pro were detected at 6 and 8 hpi in WT and SETD3 KO cells. Thus, SETD3 is not required for the CVB3 2A pro -mediated cleavage of eIF4GI during enteroviral infection.

SETD3 Is Not Required for the CVB3 2A pro -Triggered Nucleocytoplasmic Traffic Disorder
Another important event, which happens early during enteroviral infection, is the 2A pro -mediated cleavage of nucleoporin proteins [5]. This leads to an NCTD that is characterized by the efflux of nuclear proteins and influx of cytosolic proteins into the nucleus. To investigate whether SETD3 is required for the CVB3 2A pro -induced NCTD, WT and SETD3 KO cells were infected with EMCV-2A pro , EMCV-L zn , and EMCV-wt at an MOI of 50 and the localization of hnRNP K was monitored by IFA at 2, 4, 6, and 8 hpi. hnRNP K is a scaffolding protein that constitutes the ribonucleoprotein (RNP) complexes of heterogeneous nuclear RNA in mammalian cells [33] and contains a classic nucleus localization signal (NLS) motif [34]. Additionally, an antibody against dsRNA was used as an indication of viral replication. EMCV-wt, which was included as a positive control, induced NCTD very early within 2 hpi in both cell lines, and from 6 hpi-a time point at which significant viral replication was detected-the hnRNP K proteins were mainly located in the cytosol (Figure 4). In cells infected with EMCV-L zn virus, which was included as a negative control, the hnRNP K proteins always remained in the nucleus even in cells containing comparable amounts of dsRNA as observed in EMCV-wt-infected cells. Similar to EMCV-wt, EMCV-2A pro rapidly induced NCTD within 2hpi. Efflux of hnRNP K protein was observed over time and the hnRNP K proteins localized almost exclusively in the cytoplasm from 6 hpi. No difference was detected between WT and SETD3 KO cells. Therefore, the SETD3 is also not important for CVB3 2A pro to induce NCTD.

SETD3 Is Not Required for the CVB3 2A pro -Triggered Nucleocytoplasmic Traffic Disorder
Another important event, which happens early during enteroviral infection, is the 2A pro -mediated cleavage of nucleoporin proteins [5]. This leads to an NCTD that is characterized by the efflux of nuclear proteins and influx of cytosolic proteins into the nucleus. To investigate whether SETD3 is required for the CVB3 2A pro -induced NCTD, WT and SETD3 KO cells were infected with EMCV-2A pro , EMCV-L zn , and EMCV-wt at an MOI of 50 and the localization of hnRNP K was monitored by IFA at 2, 4, 6, and 8 hpi. hnRNP K is a scaffolding protein that constitutes the ribonucleoprotein (RNP) complexes of heterogeneous nuclear RNA in mammalian cells [33] and contains a classic nucleus localization signal (NLS) motif [34]. Additionally, an antibody against dsRNA was used as an indication of viral replication. EMCV-wt, which was included as a positive control, induced NCTD very early within 2 hpi in both cell lines, and from 6 hpi-a time point at which significant viral replication was detected-the hnRNP K proteins were mainly located in the cytosol (Figure 4). In cells infected with EMCV-L zn virus, which was included as a negative control, the hnRNP K proteins always remained in the nucleus even in cells containing comparable amounts of dsRNA as observed in EMCV-wt-infected cells. Similar to EMCV-wt, EMCV-2A pro rapidly induced NCTD within 2hpi. Efflux of hnRNP K protein   Subsequently, IFA assays were performed with antibodies against hnRNP K, as a marker for NCTD, and against dsRNA, as a marker for viral RNA replication. The uninfected mock was collected from 6 hpi. Scale bar represents 10 µm, all images are in the same scale. Three independent experiments were performed.

SETD3 Is Not Required for the CVB3 2A pro -Mediated Suppression of SG Formation
Enterovirus 2A pro suppresses SG formation through a mechanism that is poorly understood. Previously, we showed that infection of cells with EMCV-2A pro leads to the appearance of atypical SGs (aSGs) early in infection due to cleavage of eIF4G and the ensuing inhibition of host cell mRNA translation. These aSGs disappear through a poorly understood function of 2A pro and no typical SGs (tSGs) are observed in the mid and late phases of infection [18]. To investigate a possible involvement of SETD3 in this 2A pro activity, we analyzed SG formation in WT and SETD3 KO cells infected with EMCV-2A pro . As a control, we included EMCV-L zn , which-unlike EMCV-2A pro -is unable to suppress SG formation and triggers the formation of typical SGs (tSGs) [18]. WT and SETD3 KO cells were infected with EMCV viruses at an MOI of 10, and SG formation was monitored at 2, 4, 6, and 8 hpi by IFA (Figure 5a). Antibodies against eucaryotic initiation factor 3 (eIF3) and RasGAP SH3-domain binding protein 2 (G3BP2) were used to detect SGs, and a dsRNA antibody was used to detect viral RNA replication.
The results show that EMCV-2A pro infection triggered formation of aSGs at 4 hpi but no tSGs at later time points in either WT cells or SETD3 KO cells, whereas tSGs were observed in cells infected with EMCV-L zn , which is in line with the previous studies [18,35]. To accurately compare SG formation between WT and SETD3 KO cells, the occurrence of SGs was counted for 250 cells for each sample (Figure 5b). tSGs were formed in about 80% of cells at 6 hpi and almost all cells at 8 hpi during EMCV-L zn infection, but EMCV-2A pro effectively suppressed tSG formation in both WT and SETD3 KO cells. Therefore, the SETD3 protein is not required for the suppression of typical SGs by 2A pro . Figure 5. SETD3 protein is not required for the ability of CVB3 2A pro to suppress SG formation. (a) WT and SETD3 KO Hela cells were infected with indicated viruses at an MOI of 10. At 2, 4, 6, and 8 hpi, cells were fixed. Subsequently, IFA assays were performed with antibodies against dsRNA, as a marker for viral RNA replication, and against eIF3 and G3BP2 as markers for SGs. Three independent experiments were performed. Scale bar represents 10 µm, all images are in the same scale. (b) Quantification of (a). In total, 250 cells were quantified in each sample. Graphs depict the ratio of SG-positive cells. The solid columns represent the ratio from WT cells and the diagonal columns represent the ratio from SETD3 KO cells.

CVB3 2A pro Does Not Require SETD3 to Suppress Type I Interferon Induction
IFN induction is another key aspect of antiviral innate immune response [36]. Previously, we showed that enteroviruses strongly suppressed type I interferon (IFN) induction, most likely through the activity of 2A pro to cleave MDA5 and MAVS [19]. Consistently, whereas EMCV-L zn infection triggered a strong IFN response, this response was efficiently Viruses 2022, 14, 1360 9 of 14 suppressed in cells infected with EMCV-2A pro [18]. To investigate a possible role of SETD3 in the 2A pro -mediated suppression of IFN induction, WT and SETD3 KO cells were infected with EMCV-L zn and EMCV-2A pro at an MOI of 10. Total RNA was collected at 2, 4, 6, and 8 hpi and RT-qPCR analysis was performed to determine the IFN induction ( Figure 6). The results showed that EMCV-L zn infection induced significant IFN induction in both cell lines, but that EMCV-2A pro did not trigger IFN induction in either WT or SETD3 KO cells. Therefore, SETD3 does not seem to be required by CVB3 2A pro to suppress IFN induction.
protein is not required for the suppression of typical SGs by 2A pro .

CVB3 2A pro Does Not Require SETD3 to Suppress Type I Interferon Induction
IFN induction is another key aspect of antiviral innate immune response [36]. Previously, we showed that enteroviruses strongly suppressed type I interferon (IFN) induction, most likely through the activity of 2A pro to cleave MDA5 and MAVS [19]. Consistently, whereas EMCV-L zn infection triggered a strong IFN response, this response was efficiently suppressed in cells infected with EMCV-2A pro [18]. To investigate a possible role of SETD3 in the 2A pro -mediated suppression of IFN induction, WT and SETD3 KO cells were infected with EMCV-L zn and EMCV-2A pro at an MOI of 10. Total RNA was collected at 2, 4, 6, and 8 hpi and RT-qPCR analysis was performed to determine the IFN induction ( Figure 6). The results showed that EMCV-L zn infection induced significant IFN induction in both cell lines, but that EMCV-2A pro did not trigger IFN induction in either WT or SETD3 KO cells. Therefore, SETD3 does not seem to be required by CVB3 2A pro to suppress IFN induction.

CVB3 2A m Protein Still Induces NCTD
We also tested the effect of EMCV-2A m infection on NCTD in WT and SETD3 KO cells. We expected this proteolytically inactivate mutant not to induce NCTD, since NCTD was suggested to be highly dependent on the catalytic activity of 2A pro [9,37]. Surprisingly, we found that CVB3 2A m could still induce NCTD in both cell lines, albeit delayed ( Figure 7). EMCV-2A m infection induced hnRNP K protein efflux from 4 hpi onwards. At 8 hpi, hnRNP K protein efflux was detected in nearly all cells with no difference in WT and

CVB3 2A m Protein Still Induces NCTD
We also tested the effect of EMCV-2A m infection on NCTD in WT and SETD3 KO cells. We expected this proteolytically inactivate mutant not to induce NCTD, since NCTD was suggested to be highly dependent on the catalytic activity of 2A pro [9,37]. Surprisingly, we found that CVB3 2A m could still induce NCTD in both cell lines, albeit delayed ( Figure 7). EMCV-2A m infection induced hnRNP K protein efflux from 4 hpi onwards. At 8 hpi, hnRNP K protein efflux was detected in nearly all cells with no difference in WT and SETD3 KO cells. Consistent with our previous results (Figure 4), no hnRNP K efflux was observed in cells infected with EMCV-L zn in this experiment (data not shown).
To further substantiate this unexpected finding, we performed similar experiments in Hela-GFP-3NLS cells stably expressing a GFP fused to three tandem nucleus localization signals (NLS). Hela-GFP-3NLS cells were infected with EMCV-wt, EMCV-L zn , EMCV-2A pro , and EMCV-2A m at an MOI of 50 and the localization of both GFP-3NLS and hnRNP K was evaluated ( Figure 8). All viruses replicated well in Hela-GFP-3NLS cells as indicated by similar levels of dsRNA. Upon infection of cells with EMCV-L zn , both GFP and hnRNP K protein remained in the nucleus, resembling mock-infected cells, whereas increasing efflux of hnRNP K and GFP-3NLS was observed starting from 2 hpi in cells infected with EMCV-2A pro . In cells infected with the EMCV-2A m , GFP-3NLS and hnRNP K located in the nucleus at an early time point, i.e., at 2 hpi, but from 4 hpi both proteins started to appear in the cytosol, in line with the results above in normal Hela cells.  IFA assays were performed with corresponding antibodies. The uninfected mock was collected from 6 hpi. DsRNA staining was performed as a viral replication marker. GFP fluorescence was observed and hnRNP K staining was performed as the NCTD marker. Three independent experiments were performed.
Thus, for two different nuclear proteins, we showed for the first time that an enteroviral 2A catalytic-dead mutant was found to be able to induce NCTD, albeit with a delay. This finding challenges the previous notion that NCTD induced by enteroviral 2A protein is strictly dependent on its catalytic activity.

Discussion
Enterovirus 2A pro is a protease that has several important functions in the viral life cycle. It cleaves at the P1-P2 junction of the polyprotein, it cleaves several host factors, such as eIF4G and nup proteins [5], and its proteolytic activity is important for suppressing type I IFN and antiviral stress responses [18][19][20]. Recently, SETD3 was identified as a novel essential host factor for enterovirus RNA replication. SETD3 was found to interact with 2A pro [26] and mutations in 2A pro that disrupt the interaction with SETD3 abolished viral RNA replication. However, the precise role of SETD3 during enterovirus replication remains unknown. SETD3 is an actin-specific histidine N-methyltransferase [24], but this activity is not required for its interaction with 2A pro or virus replication. It is difficult to study the importance of SETD3-2A pro interaction for any 2A pro function as enteroviruses cannot replicate in SETD3 KO cells. Unlike enteroviruses, EMCV can replicate in SETD3 KO cells [26]. We previously generated an EMCV-2A pro virus by inserting CVB3 2A pro into the genome of a mutant virus in which the function of the L protein was disrupted, allowing us to demonstrate that CVB3 2A pro suppresses SG formation and IFN induction through its proteolytic activity [18]. In this study, we infected SETD3 KO cells with this virus to gain insight into the importance of the SETD3-2A pro interaction for the function(s) of 2A pro . Our results show that SETD3 is not required for the CVB3 2A pro -mediated cleavage of eIF4GI, the cleavage of nups and the ensuing NCTD, and the suppression of SG formation and IFN induction. Previously, we showed that the ability of 2A pro to suppress SG formation and IFN induction is conserved among members of all human enterovirus species [18]. Therefore, we speculate that it is likely that SETD3 is dispensable for the function of other enteroviruses as well.
Having shown that SETD3 is not required for the known proteolytic functions of 2A pro , the importance of the SETD3-2A pro interaction remains to be established. Previously, it has been shown that insertion of an IRES between P1 and P2 in the poliovirus genome yielded a viable recombinant virus [38]. This virus no longer relies on 2A pro for cleaving at this junction, which is essential for virus replication. RNA transcripts containing mutation of C109A, an amino acid of the catalytic triad of 2A pro , failed to yield viable poliovirus, but these transcripts showed some RNA replication, albeit with reduced efficiency compared to RNA transcripts containing wt 2A pro . Recently, a viable EV-A71 recombinant virus was described containing an IRES between P1 and P2 and a catalytic inactive 2A pro , lending support that its proteolytic activity is dispensable for viral RNA replication [39]. In contrast, deletion of 2A pro from the above-mentioned poliovirus RNAs containing an IRES between P1 and P2 rendered the RNA transcripts unable to replicate [38]. Together, these findings hint to the importance of a physical role of enterovirus 2A pro in a yet unknown step in viral RNA replication, possibly involving an interaction with SETD3. The functional importance of the interaction of 2A pro with SETD3 for viral RNA replication awaits further studies.
While performing our experiments, we made an unexpected observation with EMCV-2A m , a virus expressing a catalytic inactivated 2A pro . As expected, infection of cells with this virus did not result in eIF4G cleavage and this virus failed to suppress the IFN induction and SG formation. Strikingly, despite its inability to cleave host factors, infection with EMCV-2A m , but not with EMCV with an inactivated L protein lacking inactive 2A pro (EMCV-L zn ), triggered NCTD, albeit delayed. We cannot rule out that this unexpected finding is due to some functional interaction with an EMCV protein, although this possibility seems unlikely as we never observed any effect on NCTD by the EMCV-L zn control virus. Together, these findings show that the proteolytic activity of 2A pro is not absolutely required for NCTD induction. Notably, SETD3 is not important for 2A pro -triggered NCTD, as the same observations were made in SETD3 KO cells. These findings suggest that 2A pro activity affects NCTD by two distinct mechanisms; one is through the cleavage of Nups, the other is through an unknown proteolytic activityindependent mechanism. Cardiovirus Leader proteins, which are devoid of any enzymatic activity, also trigger NCTD [40,41]. L protein binds directly to Ran, which regulates both nuclear import and export pathways [42]. This complex subsequently recruits activated kinase cargos leading to the hyperphosphorylation of Phe/Gly-containing Nups, thereby perturbing the integrity and functioning of NPCs and resulting in NCTD [28,43]. The Nup phosphorylation triggered by L protein is mediated by mitogen-activated protein kinases [44]. Recently, the Leader protein was shown to interact with mitogen-activated protein kinase p90-ribosomal S6-kinases (RSKs) and prevent their dephosphorylation, thereby maintaining the activity of kinases [45]. Whether 2A pro also affects the activity of some kinases and thereby triggers Nup phosphorylation or, alternatively, whether it triggers NCTD through another mechanism requires further investigation.

Data Availability Statement:
The data presented in this study are available in the insert article.