Foot-and-mouth disease virus induces PERK-mediated autophagy to suppress the antiviral interferon response

Foot-and-mouth disease virus (FMDV), is a positive-sense non-enveloped RNA virus of the Picornaviridae family. It is a major cause of contagious acute viral infection in cloven-hoofed animals, though asymptomatic persistent infection is also reported (Knowles and Samuel, 2003). FMDV proliferates rapidly and causes vesicular lesions on oro-nasal mucosa and the interdigital cleft. Worldwide, there exist seven serotypes – O, A, C, Asia1, South African Territories (SAT) 1, SAT2 and SAT3. In addition, numerous variants and subtypes have also been reported (Bachrach, 1968). The infection causes high morbidity in adult animals and high mortality in young animals, leading to reduced animal productivity and economic loss (Grubman and Baxt, 2004). Also, FMD is considered as the most important constraint for international trade of animals and animal products (Leforban, 1999).

FMDV derives membranes from the endoplasmic reticulum (ER) and pre-Golgi membranes of the early secretory pathway for its replication (Midgley et al., 2013; Moffat et al., 2005). FMDV 2B protein localizes mainly in the ER exerting viroporin-like activity and increases Ca²⁺ levels in the cytosol. Also, the 2BC protein inhibits host protein secretion (Moffat et al., 2005), therefore it is likely that FMDV infection induces ER stress. The induction of ER stress results in the activation of the unfolded protein response (UPR). During this process, BiP (also known as HSPA5 and GRP78), a major ER chaperone is released from ER transmembrane signal transducers, including PKR-like ER kinase (PERK; also known as EIF2AK3), inositol-requiring enzyme 1 (IRE1; also known as ERN1), and activating transcription factor 6 (ATF6), leading to their activation (Lee, 2005). The UPR promotes cell survival either by attenuating translation mediated by phosphorylation of eIF2α to reduce the load of unfolded proteins (Malhotra and Kaufman, 2007; Ron and Walter, 2007) or by enhancing the ER protein folding capacity via ER chaperone expression through the PERK–ATF6–IRE1α pathway (Harding et al., 2000b; Hetz, 2012). The UPR also targets misfolded proteins for degradation (Harding et al., 2002). UPR activation consequently induces an adaptive cellular homeostasis response – autophagy (Jheng et al., 2014; Senft and Ronai, 2015). Autophagy is responsible for sequestering damaged organelles and cytoplasmic protein aggregates within double-membrane vesicles for degradation and recycling (Klionsky and Emr, 2000). While some viruses suppress the autophagy pathway, a few others appear to utilize this pathway to benefit their replication by avoiding the host immune response. The study of the intrinsic link between ER stress and autophagy is gaining research importance in RNA virus infections for the purpose of developing antiviral strategies (Jheng et al., 2014).

Although the involvement of autophagy in FMDV infection has been reported earlier, the mechanism of its induction and implication in FMDV multiplication is not clear. There exist ambiguous reports on the role of autophagy during FMDV infection (Medina et al., 2018; Rodriguez Pulido and Saiz, 2017). Earlier studies showed that FMDV induces and utilizes autophagosomes at an early stage of the replication (Berryman et al., 2012; O'Donnell et al., 2011). It has also been reported that the 2B protein of FMDV disturbs cellular Ca²⁺ homeostasis and induces autophagy (Ao et al., 2015), in contrast to an earlier report showing that the interaction of FMDV 2C protein with Beclin1 prevented the fusion of lysosomes to autophagosomes, allowing virus survival (Gladue et al., 2012). Another study showed that FMDV induces autophagy only at the early stage of infection and viral 3C protease degrades the Atg5–Atg12 complex, suppressing autophagy at a later time point of infection (Fan et al., 2017). While our study was underway, a recent work reported that FMDV induces complete autophagy, with the fusion of autophagosomes and lysosomes for subsequent degradation (Sun et al., 2018). Although a very recent study showed the activation of PERK and ATF6 mediated UPR during FMDV infection, however, they reported that these pathways did not influence FMDV replication (Han et al., 2019). Even more recently, Yang et al. (2020) showed that FMDV 3A protein upregulates the autophagy-related protein leucine-rich repeat containing protein 25 (LRRC25), which negatively regulates RLR-mediated type I interferon (IFN) signaling by interacting with Ras-GAP SH3-binding protein 1 (G3BP1). However, the role of ER stress-mediated UPR and autophagy, and their function in modulating innate immunity during FMDV infection mostly remains unclear. Therefore in this study, we aimed to systematically investigate FMDV-induced ER stress, UPR and autophagy, and their role in innate immune responses.

Our results demonstrate that FMDV infection induces ER stress and the UPR response through the PERK-mediated pathway. The induction of the autophagy marker lipidated LC3B (LC3B-II; LC3B is also known as MAP1LC3B) is influenced by PERK activation during the virus infection. This work also provides insights into the fact that the FMDV replication in cells is dependent on the expression of PERK and the autophagy marker LC3B-II. Furthermore, the inhibition of PERK pathway and autophagy significantly restricted FMDV replication by enhancing the levels of antiviral IFN-β and IFN-λ3 levels. Overall, these results help in the understanding of FMDV pathogenesis and host cellular antiviral mechanisms for the future development of antiviral strategies.

To study the impact of viral infection on cellular ER, we observed for ER morphology at 4 h post infection (hpi) [1 multiplicity of infection (MOI)] by transmission electron microscopy (TEM) in LFBK cells. FMDV-induced ER stress caused a dilation of the ER lumen (Fig. 1A, right panel). SYBR green based quantitative real-time RT-PCR (RT-qPCR) in relation to that of the housekeeping β-actin gene was carried out to monitor the levels of ER stress and UPR marker genes. The genes encoding BiP, PERK, ATF4 and IRE1 were upregulated in the range of 1.8–2.5-fold at 6 hpi (1 MOI). However, there was no significant upregulation of the genes encoding ATF6 and CHOP (C/EBP homologous protein; also known as DDIT3) (Fig. 1B). ATF6 transcript level remained unchanged, suggesting that the FMDV-induced UPR signaling may not utilize the ATF6-mediated pathway. Although IRE1 was significantly (P<0.05) upregulated at 6 hpi, there was no apparent splicing of its downstream XBP1 mRNA up to 12 hpi with FMDV, while the positive control, treatment with tunicamycin, led to splicing of XBP1 mRNA with the appearance of both spliced (269 bp) and unspliced (295 bp) forms (Fig. 1C), suggesting that the FMDV-induced UPR response is also independent of the IRE1–XBP1 pathway. Western blot analysis of the cells infected with FMDV (1 MOI) showed distinct enhancement in the BiP protein level and the PERK protein expression from 4 hpi (Fig. 1D), correlating with the data of real-time quantification of increased mRNA levels of the genes encoding BiP and PERK. To further investigate whether activation of the PERK pathway leads to phosphorylation of the downstream eIF2α transcription factor, immunofluorescence was carried out using an anti-p-eIF2α antibody which showed an increase in the fluorescence of phosphorylated eIF2α (p-eIF2α) at 6 hpi (1 MOI) (Fig. 1E). Furthermore, the time course of phosphorylation of the PERK substrate eIF2α was assessed by immunofluorescence following FMDV infection. A time-dependent virus cytopathic effect associated with the p-eIF2α signal was evident (Fig. S1). In line with this, the mRNA level of ATF4 was increased at 6 hpi (Fig. 1B). Taken together, these results indicate that FMDV infection triggers ER stress and induces the UPR through the PERK-mediated pathway.

To understand the relationship between ER stress-induced UPR and autophagy, VER-155008 (20 μM) an inhibitor of BiP (Macias et al., 2011), which is a target of both the ER stress response and UPR, was used to treat LFBK cells 1 h prior to infection. Our results indicated a significant reduction in the viral titer at 12 hpi (1 MOI) upon VER-155008 treatment (Fig. 2A). Similarly, the PERK pathway was inhibited by treatment with ISRIB (Rabouw et al., 2019), which reverses the effects of eIF2α phosphorylation. ISRIB (200 nM) treated and infected cells showed a decrease in autophagy marker protein LC3B-II, an absence of SQSTM1 degradation and, correspondingly, there was a reduction in viral protein levels and the viral titer (Fig. 2B–D). The concentrations of the VER-155008 and ISRIB drugs used in this study did not affect cell viability (Fig. S2).

To preclude any non-specific effect caused by chemical inhibition, knockdown of the PERK gene was carried out. Two PERK-specific artificial microRNAs (amiRNAs) with different targets in 5′-UTR (miR-PERK-T1 and miR-PERK-T2) along with negative control (miR-Neg) were expressed in LFBK cells. The reduction in expression of PERK protein was more significant in cells expressing miR-PERK-T2 (data not shown); hence, further work was done using miR-PERK-T2. Upon FMDV infection (1 MOI) of PERK-specific amiRNA-expressing cells, expectedly PERK expression was reduced, resulting in decreased LC3B-II conversion and reduced SQSTM1 degradation. Also, the virus replication was affected, as seen by decreased viral protein expression (Fig. 3A). The knockdown of PERK also resulted in the reduction of extracellular progeny virus titer (P<0.05) at 6 and 12 hpi (Fig. 3B). Furthermore, the levels of IFN-β and IFN-λ3 proteins in the infected culture supernatant of cells expressing miR-PERK-T2, as determined by ELISA, showed a significant increase at 8 and 12 hpi (Fig. 3C,D). Therefore, these results suggest that PERK knockdown reduces FMDV replication by enhancing the levels of anti-viral IFNs. The data indicate that the ER stress is induced during FMDV infection, which in turn activates the PERK pathway of UPR-mediated autophagy.

To confirm the induction of autophagy during FMDV infection, we investigated the cellular changes associated with the virus infection, using TEM. Double-membrane vesicles resembling autophagic vesicles were observed in FMDV-infected (1 MOI) LFBK cells at 6 hpi (Fig. 4B). However, similar vesicles were absent in the uninfected LFBK cells (Fig. 4A). A magnified view of the part highlighted in FMDV infected cell shows the presence of double-membrane vesicles, characteristic of autophagosomes (Fig. 4C). A quantification showed that the number of autophagosome vesicles per cell was significantly increased in FMDV-infected cells (Fig. 4D).

To further investigate the effect of FMDV infection on autophagy, the expression of candidate genes involved in autophagy activation was studied. There was clear evidence of upregulation of ATG9A and LC3B by 2.6-fold, and ULK1 by 3.3-fold in 6 h infected (1 MOI) samples (P<0.05), but there was no significant change in SQSTM1 level (Fig. 5A). These data suggest that there is an upregulation of autophagy initiation genes during FMDV infection.

The Atg12–Atg5–Atg16L complex assists in the conjugation of LC3B-I with phosphatidylethanolamine (PE), which is essential for the formation of the autophagosomal membrane. The conversion of cytosolic LC3B-I into PE-conjugated LC3B-II is indicative of increased autophagic activity (Kabeya et al., 2000; Klionsky et al., 2008). Western blotting results showed an increase in the levels of Atg5–Atg12 conjugate together with an increase in LC3B-II in cells infected with FMDV (1 MOI) from 4 hpi onwards and that further maintained for more than 9 hpi (data not shown). Rapamycin-treated cells were used as autophagy-positive control (Fig. 5B,C). The adaptor protein SQSTM1, mediating selective autophagy, gets degraded by the autophagolysosome pathway. It accumulates when autophagy is inhibited and its level decreases when functional autophagy is induced, thus serving as a marker for autophagic flux (Klionsky et al., 2008). Western blot analysis showed a gradual decline in the levels of SQSTM1 protein from 4 hpi (Fig. 5B,D), suggesting the activation of functional autophagy with the fusion of autophagosomes with lysosomes for its degradation.

The induction of autophagy brings about the association of LC3B-II to the autophagosomal membrane, revealing a characteristic punctate pattern (Klionsky et al., 2008). The widefield immunofluorescence microscopy revealed a significant increase (P<0.01) in characteristic endogenous LC3B puncta in FMDV-infected LFBK cells (1 MOI at 6 hpi) (Fig. 5E,F). These puncta were similar to those seen in cells treated with the autophagy inducer rapamycin (5 µM) (Fig. S3). This confirmed the activation of autophagy.

The inhibition of autophagic flux by bafilomycin A1, a specific inhibitor of the V-ATPase (Yamamoto et al., 1998) prevented the degradation of LC3B-II (Fig. 6A, right panel) compared to no BafA1 treatment (Fig. 6A, left panel). Consequently increased accumulation of SQSTM1 was observed in BafA1-treated cells (Fig. 6A, right panel, B). This resulted in reduced virus replication, as evidenced by considerably low viral protein expression in BafA1-treated cells (Fig. 6A, right panel, C), confirming the positive correlation between LC3B, SQSTM1 turnover and FMDV replication.

In addition, the extracellular viral titer was significantly reduced in cells treated with autophagic flux inhibitor BafA1 at 6 and 12 hpi (P<0.01). Furthermore, treatment with spautin-1, which acts by blocking the initiation of autophagy (Liu et al., 2011), also caused a significant reduction (P<0.01) in the viral titer at 12 hpi. However, the virus titer increased significantly in FMDV-infected cells treated with the autophagy inducer rapamycin compared to control FMDV-infected cells at 12 h (P<0.05) (Fig. 6D). To understand the antiviral mechanism of autophagy inhibition, RT-qPCR for IFN-β and IFN-λ3 transcript levels were analyzed in the BafA1- and spautin-1-treated and FMDV-infected (1 MOI) LFBK cells. The IFN-β and IFN-λ3 transcripts were upregulated in the range of 4.5–6-fold at 8 hpi in the presence of autophagy inhibitors (Fig. 6E).

Two of LC3B-specific amiRNAs (miR-LC3B-ORF and miR-LC3B-5′UTR) along with negative control (miR-Neg) were expressed in LFBK cells. Western blotting showed that there was a reduction in the expression of LC3B in cells expressing miR-LC3B-ORF and miR-LC3B-5′UTR, but it was more significant in the latter (data not shown). Therefore, miR-LC3B-5′UTR was used for our study. Upon FMDV infection (1 MOI) of LC3B-specific amiRNA-expressing cells, expectedly LC3B expression was decreased and SQSTM1 turnover was impeded, and the virus replication was affected as seen by decreased viral protein expression (Fig. 7A). The knockdown of LC3B also resulted in reduction of extracellular progeny-virus titer (P<0.05) at 6 and 12 hpi (Fig. 7B).

In addition, cells expressing miR-LC3B-5′UTR leading to knockdown of LC3B had enhanced levels of IFN-β and IFN-λ3 transcript levels of 8- and ∼11-fold respectively at 8 hpi (1 MOI) as determined by RT-qPCR (Fig. 7C). Furthermore, the levels of IFN-β and IFN-λ3 proteins in the infected culture supernatant of cells expressing miR-LC3B-5′UTR, as determined by ELISA, showed a significant increase at 8 and 12 hpi (Fig. 7D,E). These data reveal that suppression of autophagy affects the replication ability of FMDV by enhancing IFN levels.

During infection by cytoplasmic viruses, host organelles like ER, are utilized for virus replication. It is known that FMDV derives replication site membranes from ER and pre-Golgi membranes (Midgley et al., 2013; Moffat et al., 2005). In the process, induced ER stress initiates the UPR to restore cellular homeostasis (Pincus et al., 2010). Here, we have investigated the role of ER stress, the UPR and autophagy in FMDV replication. We demonstrated by TEM that the FMDV infection induces volume expansion and dilation of the ER due to stress. During ER stress, pronounced dilation of the ER lumen occurs in pancreatic β cells and also in yeast cells under UPR-inducing conditions (Bernales et al., 2006; Despa, 2009). This suggests that FMDV infection-associated stress causes a dilation of ER. This, in turn, activates UPR signaling for cell survival. Analysis of mRNA transcripts of ER stress-associated UPR genes during FMDV infection showed that the levels of BiP, PERK, ATF4 and IRE1α were significantly upregulated at 6 hpi. However, the levels of CHOP and ATF6 remained unchanged (P>0.05). Furthermore, the protein levels of BiP and PERK were enhanced from 4 h post-FMDV infection and the phosphorylation of eIF2α was also observed from 6 hpi, indicating the involvement of ER stress-associated UPR.

Among the three arms of the UPR, the level of ATF6 mRNA did not alter. Therefore, we then studied the downstream effect of IRE1α and PERK activation. Activation of IRE1 is generally determined by measuring the splicing of mRNA encoding the XBP1 (Chen et al., 2014; Hetz et al., 2011). We observed that during FMDV infection, the splicing event was not apparent, suggesting that FMDV did not activate the IRE1–XBP1 pathway. However, an enhanced level of PERK protein and Ser51 phosphorylation of eIF2α were evident. The transcript level of ATF4 was increased significantly (P<0.01) at 6 hpi, the time point at which enhanced phosphorylation of eIF2α was observed. This is in line with the fact that eIF2α phosphorylation at Ser51 results in the repression of global protein synthesis and preferential translation of selected genes, such as ATF4 to reduce ER stress (Rajesh et al., 2015). It was reported previously that ATF4 mRNA transcript levels are elevated in response to ER stress (Harding et al., 2000a; Dey et al., 2010), as well as in response to amino acid starvation (Siu et al., 2002). This could possibly be because the expression of ATF4 is subject to both transcriptional regulation and translational control. On the other hand, the mRNA level of CHOP, which plays a role in the induction of apoptosis (Nishitoh, 2012; Oyadomari and Mori, 2004), was not significantly altered during FMDV replication. Therefore, upregulated ATF4 likely influences the induction of autophagy (Matsumoto et al., 2013; Rzymski et al., 2010). Furthermore, FMDV-infected cells treated with VER-155008, a novel small-molecule inhibitor of BiP, a master regulator of UPR (Macias et al., 2011), caused a reduction in virus titer (P<0.05), suggesting that mitigation of UPR negatively influenced FMDV replication. Our results are partly in agreement with the very recent study that showed activation of PERK–eIF2α and ATF6 branches of UPR signaling while inhibiting the IRE1α pathway during FMDV infection (Han et al., 2019). However, they reported that either overexpression or knockdown of PERK or ATF6 did not affect FMDV multiplication (Han et al., 2019). By contrast, we found that treatment with ISRIB, a potent inhibitor of the PERK pathway (Rabouw et al., 2019), and knockdown of PERK, achieved by transient expression of specific amiRNA, significantly reduced the viral protein expression. Thus, we find that the PERK pathway is important for FMDV multiplication. The UPR and autophagy occur simultaneously following ER stress (Schröder, 2008). Furthermore, ER stress-mediated PERK and eIF2α phosphorylation is known to induce LC3 conversion (Kouroku et al., 2007). Therefore, we investigated the role of the PERK–p-eIF2α–ATF4 pathway on autophagy. We found that the PERK pathway inhibitor ISRIB, as well as gene silencing of PERK by amiRNA significantly reduced the FMDV induced LC3B-II levels and associated degradation of SQSTM1, thus suggesting that PERK and p-eIF2α play an important role in FMDV replication by prompting activation of functional autophagy. Since there are ambiguous reports on the role of autophagy in FMDV replication, we examined the FMDV-infected cells under a transmission electron microscope for autophagosome-like vesicles, as this is one of the valid methodologies to monitor autophagy (Klionsky et al., 2008). We observed that FMDV infection induced an increased accumulation of double-membrane structures resembling autophagosomes. These structures were similar to the autophagy vesicles induced by rapamycin, a positive control for autophagy (Lee et al., 2013). Thus, our data suggest that induction of ER stress via the PERK pathway of UPR prompts autophagy during FMDV infection.

Furthermore, widefield immunofluorescence microscopy of FMDV-infected LFBK cells revealed an increase in the level of endogenous LC3B. The colocalization of FMDV with LC3B is indicated by yellow puncta, which shows the association of FMDV with autophagosomes in LFBK cells. A western blot analysis of FMDV-infected cell lysates showed a gradual increase in the conversion of endogenous LC3B-I to LC3B-II and associated degradation of SQSTM1, suggesting a positive correlation between virus replication, LC3B lipidation and SQSTM1 turnover. This also suggested the induction of autophagic flux (Klionsky et al., 2008), involving a fusion of the autophagosome with lysosome, during the course of FMDV infection. A previous study had shown that Atg5–Atg12 conjugate, a ubiquitin-protein ligase (E3)-like enzyme for LC3–PE conjugation, decreased through degradation mediated by the viral protease 3Cpro (Fan et al., 2017). However, in contrast, our study clearly showed that the level of Atg5–Atg12 complex increased during the course of FMDV infection, indicating that the Atg5–Atg12 conjugate helps in the continuous formation of autophagosomes. Upon assessing the effect of autophagy-modulating drugs on FMDV replication, rapamycin (a known autophagy inducer) treatment caused an increase in the viral titer in the infected cells. The viral titer considerably reduced in the presence of autophagy-initiation blocker spautin-1 (Liu et al., 2011). Furthermore, bafilomycin A1, an inhibitor of fusion between autophagosomes and lysosomes, also reduced the FMDV replication, which is in contrast to what was shown in a previous report where no effect of BafA1 was observed in MCF-10A cells (Gladue et al., 2012). We also observed that BafA1 treatment significantly increased LC3B levels and decreased SQSTM1 turnover together with the reduction in virus multiplication. Our data suggest that FMDV infection induces functional autophagy. The virus relies on the autophagic flux for its multiplication, as BafA1 treatment reduced the virus replication. Furthermore, to study the dependence of FMDV infection on autophagy, LFBK cells with knockdown of LC3B, a known autophagy marker, were infected with the virus. This affected SQSTM1 degradation and significantly reduced FMDV protein expression and the titer, suggesting that FMDV replication is dependent on autophagy.

Both type I and III IFNs are well recognized as anti-FMDV molecules (Ma et al., 2018; Perez-Martin et al., 2012). We, therefore, estimated their levels by blocking PERK pathway and autophagy. PERK knockdown by means of amiRNA enhanced the level of IFN-β and IFN-λ3 proteins and reduced viral replication. Earlier studies have shown that activation of PERK promotes ligand- and Jak-independent phosphorylation of IFNAR1, leading to its ubiquitylation and degradation. This enables efficient viral replication (Liu et al., 2009). However, our study indicated that PERK may as well be negatively controlling the production of antiviral IFNs. Furthermore, we noticed that inhibition of FMDV-associated autophagy upon knockdown of LC3B or chemical inhibition resulted in enhancement of the IFN-β- and IFN-λ3-encoding genes, over and above the levels induced by the virus infection alone. It was noted previously that the Atg5–Atg12 conjugate directly associates with the CARD domain of the RIG-I and IFN-β promoter impairing their interaction (Jounai et al., 2007). Also, it has been reported that autophagy diminishes the early IFN-β response to influenza A virus (Perot et al., 2018) and knockdown of autophagy-related genes increased the expression of IFNs and IFN signaling pathway in hepatitis C virus-infected hepatocytes (Shrivastava et al., 2011). These results, taken together, suggest that autophagy may interfere with the RIG-mediated IFN production during FMDV infection, and therefore inhibition of autophagy enhances the levels of the antiviral IFNs.

In summary, this study substantiated that FMDV infection induces ER stress and PERK-mediated UPR, and for the first time, demonstrated the link between PERK-mediated UPR and the activation of autophagic flux. The autophagy in turn, affected IFN production to promote virus multiplication (Fig. 8). Therefore, blocking of the PERK pathway and autophagic flux reduced FMDV multiplication, with a concomitant increase in the levels of antiviral molecules IFN-β and IFN-λ3. We, therefore, propose that this pathway offers potential targets to develop therapy against FMDV infection. These findings could form the basis for further investigations to explore the interactions of non-enveloped RNA viruses with the host cell and may help elucidate molecular targets for the development of novel antiviral strategies in the future.

LFBK, a porcine origin cell line (Swaney, 1988), provided by PIADC (ARS), and the baby hamster kidney 21 (BHK-21) cell line (clone 13, ATCC) were used in this study. The cells were maintained in Glasgow's minimum essential medium (GMEM) (Himedia, India) supplemented with 10% fetal bovine serum (FBS; USA origin), 60 μg/ml penicillin (Sigma, P3032), 100 μg/ml streptomycin (Sigma, S9137) and 100 μg/ml kanamycin (Sigma, K1377), in a humidified incubator with 5% CO₂ at 37°C.

FMDV serotype O/IND/R2/75, a vaccine strain (passage level 5–8), was used in this study. The virus titer was determined by a median tissue culture infective dose (TCID₅₀) assay (Reed and Muench, 1938) in BHK-21 cells. Uniformly, 1 MOI was used in the study.

Oligonucleotide primers for PCR (Table S1) were designed using Primer 3 plus software, and were commercially synthesized (Shrimpex Biotech, India). Bafilomycin A1 (B1793), Rapamycin (R0395), Spautin-1 (SML0440), Tunicamycin (T7765), ISRIB (SML0843) and VER-155008 (SML 0271) were obtained from Sigma-Aldrich. Anti-LC3B antibody (3868S), anti-SQSTM1 antibody (5114S), anti-Atg5 antibody (A0731), anti-PERK antibody (3192S), anti-phospho-eIF2α, Ser51 (3398S) and anti-BiP antibody (3183S) were purchased from Cell Signaling Technology. Anti-GAPDH antibody (SC-59540) was from Santa Cruz Biotechnology. The monoclonal antibody against FMDV 3AB protein (10H9D8) and rabbit hyperimmune serum raised against 146S mature virus antigen of FMDV serotype O were from our laboratory stocks. Atto-633 anti-rabbit-IgG (Sigma, 41176), Alexa-Fluor-488 anti-mouse-IgG (ThermoFisher Scientific, A-11001), Alexa-Fluor 488 anti-rabbit-IgG (ThermoFisher Scientific, A-11008), anti-mouse-IgG HRPO (Dako, P0260) and anti-rabbit-IgG HRPO (Dako, P0448) conjugates were also used. Porcine IFN-β ELISA kit (E0086Po) and porcine IL28B (IFN-λ3) ELISA kit (E0471Po) were procured from Bioassay Technology Laboratory.

LFBK cells seeded either in six-well plates (1.5×10⁶ cells/well) or in 24-well plates (2.0×10⁵ cells/well) were infected with FMDV serotype O, at a multiplicity of infection (MOI) of 1. After 1 h adsorption, the cells were washed twice in GMEM (without FBS), pH 6.5, and cultured in GMEM, pH 7.4 (with 5% FBS).

Pre-miRNA sequences were designed using the guidelines of Block-it RNAi designer tool (Invitrogen Inc.), for the porcine PERK and LC3B gene. For PERK, two targets (T1 and T2) within the 5′UTR were analyzed, while for LC3B, one targeting 5′UTR and other targeting ORF regions were analyzed. The effective artificial miRNAs that gave better knockdown were chosen for the study. The sequences of these miRNAs are given in Table S2. The required sense and antisense oligonucleotides for desired miRNAs were synthesized (Eurofins Genomics), annealed and were cloned into pcDNA™6.2-GW/miR vector, as described previously (Basagoudanavar et al., 2018) to generate pcDNA6.2-miR-PERK-T1, pcDNA6.2-miR-PERK-T2, pcDNA6.2-miR-LC3B-5′UTR and pcDNA6.2-miR-LC3B-ORF. As a negative control, pcDNA™6.2-GW/miR-neg control plasmid (Invitrogen) was used.

LFBK cells were transfected with recombinant plasmids expressing pre-miRNAs targeting PERK and LC3B and negative control miRNA. After 24 h of plasmid delivery, the selection was carried out using blasticidin 6 µg/ml. The cells were maintained under the selection pressure.

To evaluate the effect of the amiRNAs targeting PERK and LC3B on FMDV replication, LFBK cell lines expressing PERK- and LC3B-specific amiRNA were infected with FMDV at a MOI of 1. The infected cell lysate was analyzed by western blotting. Extracellular progeny-virus titer was assayed by using the TCID₅₀ method (Reed and Muench, 1938). The protein levels in the supernatant of infected cells were determined by using the porcine IFN-β and IFN-λ3 ELISA kit as per the manufacturer's protocol and analyzed using four parameter logistic curve-fit software (www.elisaanalysis.com).

LFBK cells (1.5×10⁶ cells/well) were seeded in a six-well dish and incubated overnight. The cells were pre-treated with drugs – bafilomycin A1 (100 nM), rapamycin (5 μM), spautin-1 (20 μM), tunicamycin (2.5 μg/ml), VER-155008 (20 μM) or ISRIB (200 nM) – for 1 h prior to FMDV infection and also added after adsorption.

Cells were fixed in 3% glutaraldehyde (TAAB UK, G002) in phosphate buffer (pH 7.2) for 24 h and post fixed in 1% osmium tetroxide for 1 h, followed by dehydration in grades of ethyl alcohol. Later, samples were cleared in propylene oxide, embedded in Araldite CY212 resin (TAAB UK, E009) and polymerized at 60°C for 48 h. The blocks were cut using a Leica EM UC7 ultramicrotome (Leica Mikrosysteme, Austria) and stained by using uranyl acetate and lead citrate (Frasca and Parks, 1965). The stained sections were scanned under JEM 1400 plus TEM (JEOL Japan) at 80 KVA and images captured using Gatan SC 1000B camera.

Total RNA from the samples was extracted using RNeasy mini kit (Qiagen, 74104). Reverse transcription was carried out with Genesure™ H Minus strand cDNA synthesis kit using oligo(Dt) primers (PureGene, PGK163). RT-qPCR was performed using the Maxima SYBR Green/ROX qPCR Master Mix (Thermo Scientific, K0221) in Applied Biosystems 7500 real-time thermal cycler (Applied Biosystems, CA, USA). Each sample was run in duplicate. The fold change in expression levels of the genes was calculated by 2^−ΔΔCt method (Livak and Schmittgen, 2001), relative to that of the housekeeping β-actin (ACTB) gene.

An X-box-binding protein 1 (XBP1) fragment was PCR amplified from the cDNA isolated from FMDV-infected or tunicamycin-treated cells, using the specific primers pXBP1F, 5′-GGATGCCTTAGTTACTGAAG-3′, and pXBP1R, 5′-GTCCTTCTGGGTCGACTTCT-3′. The PCR amplicons were checked in 2% agarose gel and visualized using a UV transilluminator.

Lysates from the cells following drug treatment or infection or knockdown of LC3B, were prepared using RIPA lysis buffer (Amresco, N653) containing a protease inhibitor cocktail (Sigma, P2714) and 0.1 mM PMSF (SIGMA, P7626) for 10 min. Protein samples (10 µg per lane) were subjected to SDS-PAGE and transferred to a polyvinylidene fluoride (PVDF) membrane of either 0.22 or 0.45 µm pore size (Merck, ISEQ08100, BM7JA0926A). The membrane was incubated overnight at 4°C with a primary antibody (1:500) and for 1 h with species-specific HRP-conjugated secondary antibody (1:1000). Gels were visualized using SuperSignal West Pico Chemiluminescent Substrate (Thermo Scientific, RB231022) in a chemiluminescent imager (UVITEC Mini HD9). The protein band intensity was measured by densitometry (myImage Analysis, Thermo Scientific).

LFBK cells (0.5×10⁵ cells/well) were seeded on a coverslip (Genetix, 20012) in a 24-well-dish and incubated overnight. After infection with FMDV (1 MOI), for a required time, the cells were fixed with 4% paraformaldehyde (HiMedia, TCL119) for 15 min at 4°C and permeabilized using 0.1% Triton-X100 (Amresco, 0694). The cells were blocked with Tris-buffered saline (TBS) containing 3% bovine serum albumin (Alfa Aesar, J65788) and 0.02% Tween 20 (Merck, 8.22184) for 30 min at room temperature. Then the cells were incubated for 1.5 h with rabbit anti-LC3B (1:500), anti-peIF2α (1:100) or anti-3AB (1:100) mouse monoclonal antibody at 37°C, followed by probing with anti-species fluorescent dye conjugated secondary antibody for 1 h (1:500). The fluorescence signals were visualized using a widefield Delta Vision microscope (API, GE) with an Olympus 60×/1.42 NA objective. Post-acquisition, the images were deconvolved using DV softWoRX software. The number of LC3B-II punctae in the cells were quantified by using the projected image (collapsed Z-stacks). The colocalization of proteins was observed by analyzing the individual Z-stacks using the ‘Colocalization’ plugin in ImageJ (NIH).

All values are expressed as mean±s.d. Statistical analyses were performed using Student's t-test to identify statistical significance between groups. P<0.05 was considered significant.

We acknowledge the Director, ICAR-Indian Veterinary Research Institute (IVRI) Izatnagar, for facilitating this work, and also PIADC, ARS-USDA, USA, for providing LFBK cell line used in this study. H.B.R. acknowledges the fellowship received from IVRI, Izatnagar. We acknowledge the use of widefield Delta Vision microscope procured under the Wellcome Trust/DBT India Alliance Intermediate Fellowship (509159-Z-09-Z) and LSRB-DRDO grant (LSRB-310/BTB/2017) awarded to R.M. We thank Mr S. Ram Prasad for help in language editing.

The peer review history is available online at https://jcs.biologists.org/lookup/doi/10.1242/jcs.240622.reviewer-comments.pdf