The Foot-and-Mouth Disease Virus Lb Protease Cleaves Intracellular Transcription Factors STAT1 and STAT2 to Antagonize IFN-β–Induced Signaling

The antiviral innate immune response, largely dependent on IFN-stimulated gene (ISG) transcription, induced JAK-STAT signaling pathway activation through IFN production. Secreted IFN proteins play a critical role in both host antiviral innate and adaptive immunity (1–5). To date, three types of IFNs have been identified to constitute the whole IFN family (6–11), of which type I IFN (IFN-α/β) is the most widely deployed to mediate the host defense responses through JAK-STAT signaling transduction. IFN-α and IFN-β interact with their cognate receptors, IFNARs, inducing the activation of JAK1 and tyrosine kinase 2 (Tyk2), which lead to STAT1/2 phosphorylation. The phosphorylated STAT1-STAT2 dimer can then translocate into the nucleus and interact with IFN regulatory factor (IRF)9, which forms an ISG factor 3 (ISGF3) complex. Similar to other transcription factors, ISGF3 binds to the promoter of IFN-sensitive response element (STAT1/2 binding promoter), which results in the expression of a large number of antiviral ISGs (12, 13).

To counter antiviral ISG expression, viruses have evolved numerous elegant strategies to disrupt signaling of the JAK-STAT pathway. For instance, Japanese encephalitis virus NS5 has been demonstrated to block Tyk2 phosphorylation to hinder JAK-STAT signal transduction (14), whereas respirovirus C protein inhibits JAK1 and Tyk2 activation to antagonize ISG expression (15). Moreover, enterovirus 71 infection blocks JAK-STAT signaling through downregulating the nuclear localization signal receptor karyopherin α1 (KPNA1) of p-STAT1. Although this effect is not mediated via the 2A and 3C proteases of enterovirus 71, the underlying mechanism remains elusive (16).

FMDV is the causative agent of foot-and-mouth disease, which belongs to the Aphthovirus genus of the Picornaviridae family and is one of the most highly infectious animal viruses in the world. FMDV contains a positive-stranded RNA genome, which encodes four structural proteins (VP1, VP2, VP3, and VP4) and eight nonstructural proteins (leader protease [L^pro], 2A, 2B, 2C, 3A, 3B, 3C, and 3D). The FMDV enzyme proteins have been demonstrated to be crucial for immune regulation and FMDV replication (e.g., FMDV 3C^pro targeted and cleaved NF-κB essential modulator [NEMO] at the Gln³⁸³ residue to inhibit innate immune signaling [17]). In particular, FMDV 3C^pro blocks STAT1/2 nuclear translocation to antagonize the IFN signaling pathway (18), 3D is a polymerase used to catalyze FMDV replication (19), and L^pro cleaves the polyprotein of FMDV (20, 21). The L gene, located at the N terminus of the FMDV genome, is translated in two protein forms, termed Lab protein and Lb protein (Lb leader protease [Lb^pro]). The N termini of the two proteins differ since viral protein synthesis begins at two different initiation sites, separated by 84 nt. The Lb protein (Lb^pro) is more favorably synthetized than the Lab protein (22). In addition, Lb^pro functions as a papain-like cysteine proteinase (23, 24) and targets the host proteins (25–32) to play an essential role in viral replication and immunosuppression. To date, the Lb^pro of FMDV has been shown to be crucial for immune regulation by cleaving the essential components of the critical signal transduction pathway; however, this mechanism has only been investigated upstream of IFN-β production (30, 33, 34). For instance, L^pro decreases the level of IRF3 and IRF7 proteins to inhibit dsRNA-induced type I IFN (IFN-I) production (34). Moreover, L^pro also interacts with activity-dependent neuroprotective protein to decrease the expression of IFN-β and ISGs (32). Although previous studies have shown that the FMDV 3C protein specifically induces KPNA1 degradation to block STAT1/2 nuclear translocation (18), the effect of Lb^pro of FMDV on the IFN-β–induced JAK-STAT signaling transduction at the mechanistic level remains unclear.

To elucidate whether FMDV may have evolved diverse functions to evade IFN-β–induced JAK-STAT signaling, we found that FMDV Lb^pro targeted STAT1/2 to hinder JAK-STAT signaling transduction, as well as STAT1/2 translocation. In this study, we demonstrate that Lb^pro decreased the level of STAT1/2 protein expression to inhibit IFN-β–induced JAK-STAT signaling. Lb^pro was found to use its protease activity to cleave STAT1/2. STAT1 and STAT2 protein cleavage was also demonstrated during FMDV infection both in vitro and in vivo. Infection with the rFMDV Lb gene knockout (LbKO) attenuated the reduction in the level of the STAT1 and STAT2 proteins. This study uncovered a new mechanism through which FMDV Lb^pro can antagonize IFN-β–induced JAK-STAT–mediated antiviral immune responses.

Baby hamster kidney 21 (BHK21) cells and porcine kidney 15 (PK15) cells were cultured in MEM (Life Technologies). Human embryonic kidney 293 (HEK293) cells were cultured in DMEM (Life Technologies) and then supplemented with 10% FBS, 100 U/ml penicillin, and 100 μg/ml streptomycin at 37°C with 5% CO₂. These cells are all maintained in our laboratory. The FMDV (FMDV/O/BY/CHA/2010) Lb knockout recombinant mutant was constructed in our laboratory. Wild-type (WT) FMDV (FMDV/O/BY/CHA/2010), vesicular stomatitis virus (VSV) with a GFP tag (VSV-GFP), and Sendai virus were maintained in our laboratory. Replication and titration of FMDV were performed on BHK21 cells. Abs against hemagglutinin (HA) (catalog no. H3663), myc (catalog no. M4439), STAT1 (catalog no. HPA000982), STAT2 (SAB1404413), and Flag (catalog no. F1804) were purchased from Sigma-Aldrich (St. Louis, MO). Abs of STAT2 (catalog no. 72640), β-actin (catalog no. 3700), HSP90 (catalog no. 4877), p-STAT1 (catalog no. 8826), p-STAT2 (catalog no. 88410), and lamin B1 (catalog no. 17416) were obtained from Cell Signaling Technology (Danvers, MA). TRIzol was purchased from Invitrogen (Thermo Fisher Scientific, Waltham, MA). Lipofectamine 2000 was purchased from InvivoGen (San Diego, CA).

Lb gene cDNA was amplified from the total RNA extract from FMDV/O/BY/CHA/2010 and L gene as a template for L_C51A (Lb^pro with cysteine substituted with alanine at amino acid residue 51) gene construction. DNA fragments of STAT1 or STAT2 were generated by corresponding PCR amplification from total RNA of HEK293 or PK15 cells. Mutant expression plasmids were constructed by cloning the indicated sequences into pcDNA3.1(+) or pCAGGS (our laboratory collection). Using Lipofectamine 2000 (Invitrogen) for transfection, STAT1/2 plasmids and Lb^pro were cotransfected into the indicated cell lines.

We successfully constructed rFMDV using the reverse genetic system established in our laboratory, with the deletion of the Lb gene of the FMDV/O/BY/CHA/2010 strain. The resultant rFMDV lacks the Lb coding region while containing the entire structural proteins. The engineered plasmids were transfected into BHK21 cells, and rFMDV LKO was obtained after 48–60 h.

MAVS^−/−, IRF3^−/−, STAT1^−/−, and STAT2^−/− knockout cells were constructed using the CRISPR/Cas9 technique. The single-guide RNA sequences were as follows: MAVS^−/−, 5′-TCCGCCGAAGGCTGTCGAAG-3′; IRF3^−/−, 5′-GTGTGGCGGACTACGTGCGC-3′; STAT1^−/−, 5′-AAGGGGCCATCACGTTCACG-3′; STAT2^−/−, 5′-GGGCAGCAATAGCTCGATTA-3′. The single-guide RNA was inserted into the lentiviral plasmid pHBLV-U6-gRNA-EF1-CAS9-PURO (purchased from Hanbio Biotechnology, Shanghai, China), which contained Cas9, and the lentivirus was packaged to infect PK15 cells. After puromycin screening, the positive cells were sequenced and verified by Western blotting detection.

STAT1/2 (100 ng/well) binding promoter-luc along with various plasmids encoding the indicated target genes, an empty vector (EV), or pRL-TK plasmid (for normalization of transfection efficiency) was cotransfected into HEK293 cells (1 × 10⁵/well). IFN-β was further stimulated or unstimulated with cells at 24 h posttransfection, harvested after 6 or 12 h, and the Dual-Luciferase reporter assay system (Promega) was used to determine both Renilla luciferase and firefly luciferase activities.

RNA was collected from whole cells pretreated with various stimuli using an RNA extraction kit (Promega). RNA was reverse transcribed into cDNA using the Moloney murine leukemia virus reverse transcription system (Promega). Gene RNA expression was determined by SYBR Green real-time PCR or absolute quantification module using the Bio-Rad CFX Connect system. Data of quantitative PCR (qPCR) were normalized to the expression of GAPDH. Real-time qPCR primers are listed in Table I.

The indicated plasmids were transiently expressed into HEK293 cells for 24 h. Then cells were harvested after IFN-β stimulation with or without virus infection and lysed with lysis buffer containing complete protease inhibitor mixture (Roche). Whole-cell lysates were sonicated at 4°C and cleared by centrifugation. For immunoprecipitation, the supernatant was collected and incubated with the specific Abs overnight at 4°C, and then the mixed supernatant was incubated with protein A+G–agarose (Roche) for 2 h at 4°C. The Sepharose beads were washed five times with 1 ml of lysis buffer, with centrifugation at 13,800 × g for 30 s. The resulting immunoprecipitates were resolved by 6–18% SDS-PAGE. Proteins were transferred from SDS-PAGE gel onto a polyvinylidene difluoride membrane (Millipore). The membranes were incubated with the indicated primary Abs after blocking with 5% nonfat milk in 0.05% Tween 20 with TBS, followed by HRP-labeled secondary Abs. The membrane was detected by an ECL substrate (Western Lightning Plus chemiluminescent substrate kit, PerkinElmer), and a protein imaging system (ChemiDoc XRS+, Bio-Rad) was used for the visualization.

Recombinant VSV-GFP was used to evaluate antiviral responses. Briefly, the indicated plasmids were expressed in HEK293 cells for 24 h and pretreated with IFN-β for 6 or 12 h. The cells were washed three times with Opti-MEM medium, and the recombinant VSV-GFP (multiplicity of infection [MOI] of 1) was permitted to infect the cells for 1 h at 37°C, then maintained in DMEM medium without FBS. After 10 h postinfection (hpi), HEK293 cells were observed by fluorescence microscopy and VSV-GFP titer was determined.

The indicated plasmids were transfected into PK15 cells for 24 h and pretreated with IFN-β for 6 h. Then cells were washed three times with Opti-MEM medium, and FMDV LbWT or FMDV LbKO (MOI of 1) was permitted to infect the cells for 1 h at 37°C, then maintained in MEM medium. After 12 h postinfection (hpi), the viruses were collected after freeze-thawing three times. Furthermore, BHK21 cells were cultured in a 96-well plate to produce cell monolayers after 24 h. The cells were infected by incubation with serial 10-fold dilutions (made in MEM without FBS) of FMDV LbWT or FMDV LbKO virus for 1 h at 37°C. After the indicated infection period, the virus cytopathic effect (CPE) was observed under an inverted microscope, and virus titer was determined.

PK15 cells were cultured in 60-mm plastic dishes and maintained in MEM medium containing 10% FBS. The cells were treated with IFN-β or left untreated for 6 h after EV or Lb^pro transfection for 24 h. The cells were then infected with FMDV (MOI of 1) or left uninfected for 1 h. After 12 h, FMDV was collected. The PK15 cells were cultured in 12-well plates, and the collected FMDV was serially diluted by a 10-fold gradient and added to the 12-well plates in an equal volume. After 1 h, the cells were washed three times with ice-cold PBS and then cultured with a medium containing 1.5% methylcellulose for 3–5 d. After crystal violet staining, statistical analysis was performed.

The indicated plasmids were expressed in PK15 or HEK293 cells; after 24 h, the cells were washed with cold PBS three times. Subsequently, the cells were fixed with 4% paraformaldehyde solution for 30 min at room temperature and then washed again with PBS three times and permeabilized in 1% Triton X-100/PBS containing 10% sheep serum for 20 min at room temperature. Cells were then washed with PBS three times. After that, cells for nonspecific binding were blocked in 5% BSA and 0.1% Tween 20 in PBS for 1 h, gently washed three times with ice-cold PBS for 5 min, and then incubated overnight at 4°C with specific primary Abs, followed by fluorescence-labeled secondary Ab incubation for 1 h. Subsequently, cell nuclei were stained with DAPI (Sigma-Aldrich). A Zeiss LSM 710 confocal microscope was used for image acquisition.

Animal experiments with viral infection were carried out in the biosafety level 3 bio-containment facility of Lanzhou Veterinary Research Institution of the Chinese Academy of Agricultural Sciences, following the guidelines approved by the Lanzhou Veterinary Research Institution Animal Ethics Committee.

WT (type O) FMDV, its mutant LbKO, or MEM medium was used to inoculate pigs. Forty-two pigs (20–30 kg of 2-mo-old piglets) were randomly assigned to three groups, with five animals each for the FMDV WT and LbKO groups, and five for the control group. Pigs were inoculated via i.m. injection in the neck with 3 × 10¹¹ copy numbers/animal. Clinical signs, such as rectal temperatures, vesicular lesions of the mouth, and lesions of the foot bearing, were observed daily.

Samples were collected at 1, 2, 3, 5, and 10 d postinfection (dpi). The fresh spleen and lymph node were collected, snap-frozen with liquid nitrogen, and kept at −80°C for subsequent Western blotting and RNA expression determination.

The tested serum was diluted from 1:4 to 512. The dilution method used for the positive or negative controls was the same as the test samples. The diluted serum in 50 μl was added to the ELISA microplate followed by 50 μl of enzyme Ab to each well and mixed. The plate was covered and incubated for 30 min at 37°C. The plate was washed five times and 50 μl of substrate was distributed to each well. The plate was covered and left for 10–15 min at 37°C in the dark. Finally, 50 μl of stop solution was added to each well. The absorbance at 450 nm for each well was measured with a microplate reader (FLUOstar OPTIMA, BMG Labtech, Ortenberg, Germany).

Sample serum and MEM were inactivated at 56°C for 30 min after mixing. The samples, as well as positive and negative serum, were serially diluted from 1:4 to 512. The virus titer was then diluted to a 50% tissue culture infective dose of 200. The diluted-serum was mixed with virus and incubated for 1 h at 37°C. After that, the mixture was added to BHK21 cells, which were cultured in a 96-well plate for 24 h. The cells were incubated in a 37°C incubator and observed once on the third day. The cells were monitored for the appearance of CPEs by light microscopy on the fourth or fifth day. The neutralizing Ab titers were assessed.

Each independent experiment was performed in triplicate and subjected to two-way ANOVA or a Student t test using GraphPad Prism 8 (GraphPad Software, San Diego, CA). A p value of <0.05 was considered statistically significant.

To explore whether FMDV affected the IFN-β–induced signal transduction blockade, PK15 cells were first infected with FMDV followed by the addition of porcine IFN-β, which can induce ISGF3 binding promoter (STAT1/2-luc) activity and consequently trigger antiviral gene (e.g., ISGs) expression. The luciferase reporter assays showed that FMDV infection inhibited IFN-β–induced STAT1/2-luc activity (Fig. 1A). Next, we investigated whether Lb^pro, 3C, or 3D influenced IFN-β–induced STAT1/2-luc activity. We found that both 3C and Lb^pro inhibited STAT1/2 binding promoter activities, which were enhanced by 3D (Fig. 1B). Consistent with the findings of a previous study, the FMDV 3C protein inhibited JAK-STAT signaling (18). Therefore, we narrowed down Lb^pro as the candidate to inhibit STAT1/2 binding promoter activity. Furthermore, we found that Lb^pro suppressed IFN-β–induced STAT1/2 binding promoter activation in a dose-dependent manner (Fig. 1C). This observation prompted us to determine whether Lb^pro could impact IFN-β–induced antiviral activity and thus facilitate viral replication using a FMDV or VSV-GFP infection assay. The virus plaque, titer, and CPE data demonstrated that Lb^pro facilitated FMDV replication following IFN-β treatment comparing with control (EV) (Fig. 1D, 1E). Consistently, virus GFP fluorescence intensity and VSV-GFP titer results in (Fig. 1F showed that Lbpro increased the replication of VSV-GFP during IFN-β stimulation as compared with EV. Next, we assessed the IFN-β–induced expression of ISGs after L gene transfection. The data revealed that Lb^pro could significantly inhibit transcription of the antiviral genes ISG20, ISG54, and ISG56 (Fig. 1G).

IFN-β combines with its receptors and induces the phosphorylation of STAT1 and STAT2. Phosphorylated STAT1 and STAT2 then interact with IRF9 to form a complex of ISGF3, which binds with the IFN-sensitive response element, finally inducing the antiviral ISG expression. Thus, we next investigated the role of Lb^pro on STAT1/2 phosphorylation after IFN-β stimulation. The results demonstrated that after IFN-β treatment, Lb^pro inhibited both STAT1 and STAT2 phosphorylation (Fig. 2A). Moreover, we found that Lb^pro also reduced the level of STAT1/2 protein (Fig. 2A). Upon IFN-β stimulation, phosphorylated STAT1 and STAT2 translocate from the cytoplasm to the nucleus. We examined the cellular localization of Lb^pro and STAT1 and STAT2 proteins after IFN-β stimulation with an immunofluorescence assay. The confocal images shown in (Fig. 2B suggest that Lb^pro blocked IFN-β–induced STAT1/2 shuttling from the cytoplasm to the nucleus. This phenomenon was further corroborated by nuclear and cytoplasmic fractionation experiments. As shown in (Fig. 2C, greater levels of p-STAT1 and p-STAT2 were accumulated in both fractions in the EV-expressed compared with the Lb^pro-expressed cells after IFN-β treatment. Moreover, the levels of STAT1 and STAT2 protein expression were decreased after Lb^pro expression. Taken together, these data suggest that FMDV Lb^pro inhibits STAT-dependent signal transduction and decreases STAT1/2 stability and translocation.

To further determine whether FMDV Lb^pro targeting to STAT1/2 was a contributing factor to the decreased STAT expression and translocation interference following IFN-β stimulation, Lb^pro was coexpressed with myc-porSTAT1 (porcine STAT1) or myc-porSTAT2, followed by coimmunoprecipitation (co-IP) detection with tag Abs. (Fig. 3A and 3B show that Lb^pro interacted with STAT1/2 directly, whereas no L-STAT complex was detected when IgG was used the IP Ab. Next, we examined the interaction of Lb^pro and endogenous STAT1 or STAT2. As expected, Lb^pro also associated with endogenous STAT1/2 (Fig. 3C). These data suggest that Lb^pro may target the STAT1/2 proteins to interfere with IFN-β–induced signal transduction.

As shown in (Fig. 2, the level of STAT1 and STAT2 proteins was decreased after Lb^pro transfection. To further explore the mechanism behind this observation, we first characterized the effect of Lb^pro on STAT1 and STAT2. As shown in (Fig. 4A and 4C, Lb^pro overexpression reduced the level of myc-STAT1 (porSTAT1 or human STAT1) and STAT2-Flag (porSTAT2) proteins. In addition, we found that Lb^pro decreased the level of STAT1 and STAT2 proteins in a dose-dependent manner (Fig. 4B, 4D). Interestingly, we observed a fragment of STAT2 following Lb^pro expression (Fig. 4D).

Lb^pro is a well-characterized papain-like proteinase and its protease activity site is located at amino acid residue 51 (35). To elucidate the effect of papain-like activity on IFN-β–induced signaling, a mutant of L_C51A was evaluated in IFN-β–stimulated HEK293 cells coexpressing the STAT1 and STAT2 proteins. The results showed that L_C51A restored the STAT1/2 phosphorylation and protein levels in HEK293 cells stimulated with IFN-β (Fig. 4E, 4F). The grayscale analysis results also showed that Lb^pro reduced the STAT1/2 proteins and their phosphorylation level compared with L_C51A or EV (Supplemental Fig. 1).

Previous studies have demonstrated that L cleaves the translation initiation factor of eIF-4G to affect host cell translation (36, 37). To eliminate the effect of cleavage of eIF-4G by Lb^pro on the reduction of STAT proteins, treatment with actinomycin D (transcription inhibitor) or cycloheximide (translation inhibitor) was added after Lb^pro or L_C51A expression. As shown in (Fig. 4G and 4H, we found that treatment with protein translation or transcription inhibitors did not restore the STAT1 and STAT2 protein levels during Lb^pro expression, and Lb^pro still decreased the level of STAT1 and STAT2 protein. This suggests that the reduction in the level of the STAT1 and STAT2 proteins was not dependent on eIF-4G during Lb^pro expression.

Moreover, to determine whether the Lb^pro-mediated reduction in STAT protein levels was dependent on eIF-4G cleavage, we constructed D163 and D164 (D163N/D164N) amino acids mutagenesis of L^pro, which can partially abolish the ability of eIF-4G cleavage (36, 37). A D163N/D164N mutant (L_D163N/D164N) of L^pro, and WT L^pro, as well as STAT1 and STAT2 proteins, were purified, and L_D163N/D164N or L^pro was incubated together with STAT1 or STAT2. As observed in (Fig. 4I, L_D163N/D164N had the ability to cleave STAT1 and STAT2. This finding suggests that the decrease in STAT1 and STAT2 proteins may not depend on eIF-4G cleavage. We also assessed the effect of L_D163N/D164N on IFN-β–induced STAT1/2 promoter activation following IFN-β stimulation. As shown in (Fig. 4J, it was found that although L_D163N/D164N could partially lose the function of eIF-4G cleavage, it could still inhibit activation of the IFN-β–induced STAT1/2 promoter. These results suggest that eIF-4G cleavage is not the reason for the decrease in STAT1 and STAT2 proteins.

To further determine whether STAT1/2 cleavage was related to Lb^pro papain-like protease activity, the papain protease inhibitor rupintrivir was also added to the medium during STAT1/2 expression. After Lb^pro expression, the decrease in STAT1 and STAT2 protein was inhibited by rupintrivir treatment (Fig. 4K, 4L). The STAT2 cleavage fragment that belongs to the full-length protein disappeared after rupintrivir treatment compared with DMSO treatment (Fig. 4L). Furthermore, to confirm whether L protein overexpression induces caspase activation to cleave STAT proteins, we used different inhibitors, as shown in Supplemental Fig. 2. It was found that the caspase inhibitor Z-VAD-FMK could not alter the reduction in the level of STAT1/2 protein during Lb^pro expression. These data suggest that Lb^pro can cleave STAT1/2 using protease activity.

Next, we assessed the effects of Lb^pro or L_C51A on IFN-induced antiviral activity. As shown in (Fig. 4M–O, Lb^pro- or L_C51A-expressing PK15 cells were stimulated with IFN-β, followed by an infection with FMDV. The virus titer, CPE, and plaque of FMDV assay results indicated that L_C51A failed to promote FMDV replication compared with Lb^pro. Consistently, Lb^pro- or L_C51A-expressing HEK293 cells were stimulated with IFN-β, followed by an infection with VSV-GFP. GFP fluorescence intensity and virus titer of VSV-GFP results in (Fig. 4P showed that Lb^pro increased the replication of VSV-GFP during IFN-β stimulation as compared with EV, and L_C51A failed to promote VSV-GFP replication compared with Lb^pro. The effects of a series of Lb^pro mutants on STAT1/2 binding promoter activity were evaluated in IFN-β–stimulated HEK293 cells coexpressing the STAT1/2-luc reporter. As seen in (Fig. 4Q, L_C51A, the mutants of Lb (△1–120 bp), and Lb (△1–315 bp) abolished the inhibition of Lb^pro on STAT1/2 binding promoter reporter gene activation after IFN-β stimulation. In contrast, Lb (△316–630 bp) continued to exhibit a suppressive function to some extent. These data indirectly suggest that the N terminal of Lb^pro, which contains the protease active site, is important for IFN-β–induced ISG production.

Following Lb or L_C51A transfection, PK15 cells were subjected to IFN-β stimulation and FMDV infection as indicated in (Fig. 4R. The level of porcine ISG mRNA levels and FMDV titers were measured. Consistent with the above results, the mRNA expression of ISGs in the Lb-transfected PK15 cells was significantly lower than that in L_C51A or EV-expressing cells following IFN-β treatment. In contrast, the FMDV titer was drastically increased in the cells transfected with Lb versus L_C51A. Altogether, the results demonstrated that Lb^pro could block the innate immune response by cleaving the level of STAT1 and STAT2 proteins. In addition, the L_C51A mutant lost the ability to suppress IFN-β–induced signal transduction.

We also attempted to further evaluate the role of FMDV Lb^pro on the STAT-protein stability at the virus level. To this end, PK15 cells were transfected with STAT1-Glag or STAT2-Flag, followed by the Lb knockout of FMDV (rFMDV LbKO) or the WT of FMDV (FMDV LbWT) infection at different times. The results shown in (Fig. 5A and 5B indicate that the STAT1 and STAT2 levels decreased upon rFMDV LbKO or FMDV LbWT infection. It was clear that STAT1/2 were substantially decreased at 12 hpi for either FMDV WT or FMDV LbKO, respectively; however, the STAT1/2 protein levels were significantly decreased after WT FMDV infection for 4 h, which is ahead of LbKO infection (Fig. 5A, 5B). Moreover, we observed that the decrease in endogenous STAT1/2 occurred earlier during FMDV LbWT than FMDV LbKO infection (Fig. 5C). Interestingly, the cleavage fragment of STAT2 disappeared following FMDV LbKO infection compared with FMDV LbWT infection.

As shown in (Fig. 5D, the mRNA expression of ISGs, including OAS1, ISG15, ISG54, STAT1, and STAT2, increased significantly at 6 h after rFMDV LbKO infection compared with FMDV WT infection. Therefore, these data suggest that rFMDV LbKO infection delays the cleavage of STAT1/2 proteins and promotes IFN-β–induced signal transduction.

Moreover, we assessed the growth changes of rFMDV LbKO on cells lacking innate immune activity. First, BHK21 (IFN-β deficiency cells), MAVS-, IRF3-, STAT1-, and STAT2-knockout PK15 cells were infected with rFMDV LbKO or FMDV LbWT. As shown in (Fig. 5E, the titer of LbKO FMDV remained unchanged compared with that of WT FMDV in BHK21 cells; however, rFMDV LbKO growth was lower than that of FMDV LbWT after IFN-β stimulation. Furthermore, we found that in MAVS^−/− or IRF3^−/− cells, FMDV LbKO growth was similar compared with WT PK15 cells (Fig. 5F). In addition, rFMDV LbKO and FMDV LbWT replications also were assessed in STAT1^−/− or STAT2^−/− cells during IFN-β treatment or without treatment. (Fig. 5G shows that FMDV LbKO growth was lower compared with that of FMDV LbWT in STAT1 or STAT2 WT PK15 cells after IFN-β treatment. However, the titers were similar between FMDV LbKO and FMDV LbWT in STAT1^−/− or STAT2^−/− cells following IFN-β stimulation. These results indicate that Lb promotes FMDV replication related with innate immune activity after IFN-β treatment.

Because Lb^pro reduced the level of STAT1/2 expression, we were curious about the Lb^pro cleavage sites in STAT1/2. Although Lb^pro targets several proteins for cleavage, only a small number of substrate sequences have been experimentally identified thus far. Moreover, the amino acid sequences of these proteins are not conserved motifs. To better understand the cleavage function of Lb^pro, we first expressed Lb^pro at increasing concentrations together with STAT1/2 and checked the molecular mass of the cleavage fragment protein. The results in (Fig. 6A show that Lb^pro induced a marked reduction in the full-length STAT1 levels in a dose-dependent manner. Interestingly, a STAT1-derived small product >20 kDa was specifically detected with Abs for STAT1. To identify the cleavage site in STAT1, we examined the truncated STAT1 mutants, which divided the intact STAT1 protein into two or three segments on average (Fig. 6F). We did not pinpoint the exact cleavage site in STAT1; however, as seen in (Fig. 6B, the range was limited to aa 252–502.

From (Fig. 4C and 4D above, a Flag tag fused in the C-terminal of STAT2 protein was used, and a fragment at ?80 kDa was detected with a Flag Ab for STAT2. This suggested that the site of STAT2 cleavage by Lb protease was located at the N-terminal of the STAT2 protein. To confirm this hypothesis, we first constructed four STAT2 deletion mutants, using co-IP detection, as shown in (Fig. 6C. We found that L^pro could interact with STAT2 (1–316 aa), for which aa 316–865 were deleted. The results also indicated that the Lb^pro-induced cleavage site of STAT2 was located at the N-terminal.

The STAT2 full protein molecular mass is 98 kDa. Based on the size of the STAT2 cleavage fragment (?80 kDa), we first constructed mutants that contained deletions at every 50 aa at the N-terminal of the STAT2 protein (Fig. 6F). We found that when the STAT2 (△1–150 aa) mutant (N-terminal of STAT2 mutant was deleted to 150 aa) was coexpressed with Lb^pro, no cleavage fragment was found. The result also showed that the molecular mass of the STAT2 (△1–150 aa) mutant was similar to that of the large cleavage fragment of the intact STAT2 protein cleaved by Lb^pro. Therefore, we speculated that the site of STAT2 cleavage by Lb^pro was in the 100–150 aa at the STAT2 N-terminal of (Fig. 6D).

Next, we constructed the mutants that contained a deletion every 10 aa within the 100–150 aa at the STAT2 N-terminal. As shown in (Fig. 6E, the STAT2 mutant with a deletion in residues 140–150 (QQHEIESRIL) abolished Lb^pro cleavage. Furthermore, we also constructed an Lb^pro-cleaved site mutation (EERQTQLCTS) of STAT2. The mutant together with Lb^pro was transfected into HEK293T cells. (Fig. 6G shows that no cleaved fragment of STAT2 was observed during Lb and STAT2 mutant coexpression. This finding suggested that the QQHEIESRIL region of STAT2 was essential for the Lb^pro-mediated cleavage.

Taken together, we speculated that the cleavage site of STAT1 was in the range of aa 252–502 after Lb^pro transfection. Moreover, we also demonstrated that the STAT2 protein cleavage site was located at residues 140–150 (QQHEIESRIL) after Lb^pro expression.

Given that FMDV Lb^pro was associated with JAK-STAT signaling, to directly assess the effect of Lb^pro on the IFN-induced signal transduction in vivo, we conducted a pig study to further corroborate the response of FMDV infection on the IFNs induced by signal transduction. Pigs were inoculated with rFMDV LbKO and FMDV WT to assess Lb^pro-mediated cleavage of STAT1 and STAT2 in vivo. (Fig. 7A and 7B show that the level of STAT1 and STAT2 protein expression in the lymph nodes and spleen was decreased in the FMDV WT group compared with rFMDV LbKO before 5 dpi. Moreover, the mRNA level of ISGs, including OAS1, ISG15, ISG54, STAT1, and STAT2, from the spleen increased over time after rFMDV LbKO inoculation compared with FMDV WT (Fig. 7C). In addition, consistent with the above results, we observed a decrease in STAT1 and STAT2 proteins by immunofluorescence staining in FMDV LbWT–infected pigs compared with rFMDV LbKO (Fig. 7D). These results indicate that Lb^pro inhibits JAK-STAT signal transduction through decreasing the stability of the STAT1 and STAT2 proteins.

A comprehensive neutralizing titer represents a key marker of the protective immune response against viral infection. Using the mAb competition-based ELISA method and cell-based comprehensive neutralizing effect detection, we determined the anti-FMDV WT-neutralizing effect in pig sera following inoculation with FMDV WT or rFMDV LbKO. The results of the mAb competition-based ELISA detection showed that all pigs inoculated with both FMDV Ags developed detectable FMDV Ab responses on a similar level after 5 dpi, whereas the Ab titers were lower in the rFMDV LbKO group immediately after inoculation (Fig. 7E). Interestingly, in cell-based comprehensive neutralizing effect detection, higher and different titers of the anti-FMDV WT-neutralizing effect were detectable in BHK21 cells on days 1, 3, and 5 postimmunization compared with FMDV WT. On day 10 postimmunization, a similar trend occurred (Fig. 7E).

Taken together, our results suggest that FMDV WT infection hindered the JAK-STAT signaling responses in pigs, supporting the function of Lb^pro-dependent STAT cleavage, and suppressed the host defense upon FMDV infection.

ISGs are important components of innate immune antiviral response. Viruses have evolved different methods of evading the innate immune response by interfering with various steps involved in IFN-induced JAK-STAT signaling. To survive in the host, FMDV utilizes viral proteins to heavily suppress the cellular innate immune response to create a “friendly” environment for successful viral replication. With regard to IFN-induced JAK-STAT signaling, a previous study reported that the FMDV 3C protein induced KPNA1 degradation through proteasome- and caspase-independent methods to disrupt the nuclear localization signal receptor for tyrosine-phosphorylated STAT1, thereby blocking JAK-STAT signaling (18). However, in this study, we found that FMDV Lb^pro represented another candidate through which IFN-induced JAK-STAT signaling can be inhibited. Lb^pro is a well-known protein that plays a regulatory role in virus replication (38, 39), host protein synthesis suppression (40), and host defense disruption (28, 33, 34, 41, 42). Although the functions of Lb^pro have been studied more recently, the effect of Lb^pro on the JAK-STAT signaling pathway has not been reported.

STAT1 and STAT2 proteins are also the target for other viruses, including the bluetongue virus NS3 protein, which is ubiquitinated on lysine 13 and 15 and recruits an E3 ligase for STAT2 degradation (43), and the Zika virus NS2A targets STAT1 and STAT2 to promote their degradation (44).

In this study, to evaluate the role of the IFN-β–induced JAK-STAT pathway during FMDV infection, we studied the interplay between STAT1/2 and the viral Lb^pro complex. Our model revealed that FMDV infection is accompanied by the production of Lb^pro, which targets and cleaves STAT1/2 to antagonize IFN-β–induced signaling, and such cleavage occurred prior to STAT1/2 phosphorylation. We also demonstrated that the cleavage these proteins were dependent on protease catalytic activity. Previous studies have shown that the L protease cleaves eIF-4G to inhibit host cell translation. Our results also determined that Lb^pro suppresses FMDV-induced ISGs of STAT1 and STAT2 mRNA expression. The caspase signaling, transcription, and translation inhibition results indicated that the effects of Lb^pro on STAT protein stability are independent of caspase, eIF-4G, and mRNA transcription.

Evidence of STAT1 and STAT2 cleavage was also observed during infection with FMDV LbWT and rFMDV LbKO. We also found that rFMDV LbKO growth was lower compared with FMDV LbWT. Because both the genome of FMDV LbWT and rFMDV LbKO can be recognized by the RIG-I–like receptor and eventually induce IFN production, we speculate that the reduction of rFMDV LbKO replication may due to: 1) the Lb of FMDV LbWT inhibits the production of IFN, in which L protease cleaves TBK1 or MAVS to inhibit IFN production, resulting in a decrease in the expression of antiviral ISGs, and ultimately increased FMDV LbWT replication; and/or 2) the Lb of FMDV LbWT inhibits the IFN-induced JAK-STAT pathway activation, which in turn antagonizes the expression of antiviral ISGs and increases the replication of FMDV LbWT. Using the IFN-deficient cells (BHK21; MAVS^−/−, IRF3^−/−, STAT1^−/−, and STAT2^−/−), we demonstrated that rFMDV LbKO replication was reduced because Lb affects both the upstream signaling of IFN-β production and downstream IFN-induced JAK-STAT signaling.

The levels of STAT1 and STAT2 protein expression were noticeably decreased during Lb expression and FMDV WT infection in vitro and in vivo. When STAT1 or STAT2 was coexpressed with Lb^pro, a small protein fragment of >20 kDa of STAT1 or a fragment of 80 kDa of STAT2 was detected. Lb^pro-mediated cleavage in a protease-dependent manner was found in MDA5, TBK1, G3BP1, eIF-4G, LGP2, and Gemin5, thereby blocking their respective downstream signaling (25, 29–31, 42, 45); however, only a very small number of substrate sequences were experimentally identified (25, 42, 45). Rodríguez Pulido et al. (42) analyzed the LGP2 sequence with Gemin5 and Daxx, and identified a conserved motif of (R)(R/K)(L/A)(R). However, we did not find any conserved motif for the STAT1 and STAT2 proteins when compared with LGP2, Germin5, and Daxx. In this study, we were unable to identify the exact cleavage site of STAT1 after Lb^pro expression; however, we found that the cleavage site was located in the 252–502 aa range. For the cleavage site of STAT2 after Lb^pro expression, as shown in (Fig. 4C and 4D, using STAT2 protein, which contains a Flag tag at the C terminus, we found a high-molecular-mass cleavage fragment of STAT2 (∼80 kDa) after Lb^pro expression. This finding suggests that the STAT2 cleavage site is in the N terminus of the protein. When we detected the interaction between the STAT2 deletion mutants and Lb^pro, the results showed that Lb^pro interacts with 1–316 aa of STAT2. These results directly indicate that Lb^pro targets the N terminus of STAT2. Finally, as shown in (Fig. 6D, there was no band within the 140- to 150-aa-deleted STAT2 and Lb^pro coexpression. Therefore, we speculated that Lb^pro recognizes the motif QQHEIESRIL, which spans 140–150 aa of STAT2. Moreover, we replaced the 140–150 aa of STAT2 with different characteristic amino acids and found that compared with WT STAT2, no cleavage fragment appeared in the mutant following Lb expression. Taken together, these results suggest that Lb cleaved both STAT1 and STAT2. According to the STAT1 and STAT2 domains, the cleavage position localized in the coiled-coil domain (CCD) and DNA binding domain of STAT1, and the CCD of STAT2. The CCD is also targeted by Zika virus, and the NS5 protein interacted with CCD of STAT2 to induce STAT2 degradation (46). Moreover, the CCD is essential for the IRF9 interaction. Thus, we speculated that the Lb^pro cleavage function would likely abolish STAT1/2 and IRF9 complex formation.

The primary characteristic of Lb^pro is a papain-like protease, and the cysteine at 51 aa is the catalytic residue (35). STAT1/2 cleavage was inhibited by the papain-like protease inhibitor, rupintrivir, and the L_C51A mutant. This suggests that the protease activity of Lb^pro is essential for STAT1/2 cleavage. The Lb deletion mutants that contain protease activity sites (△1–120 bp and △1–315 bp) lost the function to inhibit STAT1/2 binding promoter activity following IFN-β treatment. This finding also indirectly proved that the Lb^pro protease activity could impact the IFN-I–induced antiviral effect.

Additionally, given that the STAT1 and STAT2 proteins are cleaved by Lb^pro transfection, we also analyzed the fate of STAT1 and STAT2 during FMDV infection. The time for STAT1/2 cleavage was delayed with rFMDV LbKO compared with FMDV LbWT. Moreover, a moderate decline in the level of STAT1/2 was also observed at 9 h after rFMDV LbKO infection. The study by Du et al. (18) demonstrated that since FMDV 3C did not decrease the level of STAT1 and STAT2 expression of proteins, we excluded the effect of 3C on STAT1 and STAT2. Evidence of STAT1 and STAT2 protein cleavage was also observed during FMDV infection, and a decrease in STAT1 and STAT2 protein expression was observed from different tissues following FMDV infection. Thus, the STAT1 and STAT2 proteins were processed by Lb^pro after FMDV infection. STAT1 and STAT2 protein cleavage was more serious in FMDV WT infected pig tissues, which compared with rFMDV LbKO infected tissues, suggesting that Lb^pro plays an important role in the inhibition of JAK-STAT signaling. Somewhat confusingly, because the level of STAT1 and STAT2 protein expression was not cleaved significantly on 5 dpi, we speculate that these results may be due to the degradation of FMDV L^pro, which restored the level of STAT1 and STAT2 proteins. To confirm this suspicion, we also found that KPNA1 can induce L^pro degradation (data not shown).

In line with the FMDV copy numbers observed in the serum postinfection, the results from the anti-FMDV monoclonal competition-based ELISA testing showed that the Ab titers began increasing slowly at the earlier stage of infection. However, from BHK21 cell–based comprehensive neutralizing effect detection, the neutralizing effects were different between the inoculation with both viruses. We surmise that the FMDV might not effectively infect BHK21 cells due to the presence of antiviral cytokines, such as IFN or ISGs in the serum of rFMDV LbKO-infected cells.

Taken together, we propose a model depicting the mechanism through which Lb^pro hijacks the IFN-induced activated JAK-STAT signaling pathway via decreasing the component stability of the pathway (Fig. 8). In this model, IFN-induced signaling transduction was hindered by the early viral proteins of Lb^pro. Moreover, Lb^pro targeted and cleaved STAT1/2 to inhibit IFN-induced signal transduction. The cleavage was dependent on eIF4G protein level. Therefore, this study expands the horizon for FMDV Lb^pro to perform different functions.

We thank Xuan Guo for modifying the manuscript.

The authors have no financial conflicts of interest.