Nucleolin Promotes IRES-Driven Translation of Foot-and-Mouth Disease Virus by Supporting the Assembly of Translation Initiation Complexes

ABSTRACT

Nucleolin (NCL), a stress-responsive RNA-binding protein, has been implicated in the translation of internal ribosome entry site (IRES)-containing mRNAs, which encode proteins involved in cell proliferation, carcinogenesis, and viral infection (type I IRESs). However, the details of the mechanisms by which NCL participates in IRES-driven translation have not hitherto been described. Here, we identified NCL as a protein that interacts with the IRES of foot-and-mouth disease virus (FMDV), which is a type II IRES. We also mapped the interactive regions within FMDV IRES and NCL in vitro. We found that NCL serves as a substantial regulator of FMDV IRES-driven translation but not of bulk cellular or vesicular stomatitis virus cap-dependent translation. NCL also modulates the translation of and infection by Seneca Valley virus (type III-like IRES) and classical swine fever virus (type III IRES), which suggests that its function is conserved in unrelated IRES-containing viruses. We also show that NCL affects viral replication by directly regulating the production of viral proteins and indirectly regulating FMDV RNA synthesis. Importantly, we observed that the cytoplasmic relocalization of NCL during FMDV infection is a substantial step for viral IRES-driven translation and that NCL specifically promotes the initiation phase of the translation process by recruiting translation initiation complexes to viral IRES. Finally, the functional importance of NCL in FMDV pathogenicity was confirmed in vivo. Taken together, our findings demonstrate a specific function for NCL in selective mRNA translation and identify a target for the development of a broad-spectrum class of antiviral interventions.

IMPORTANCE FMDV usurps the cellular translation machinery to initiate viral protein synthesis via a mechanism driven by IRES elements. It allows the virus to shut down bulk cellular translation, while providing an advantage for its own gene expression. With limited coding capacity in its own genome, FMDV has evolved a mechanism to hijack host proteins to promote the recruitment of the host translation machinery, a process that is still not well understood. Here, we identified nucleolin (NCL) as a positive regulator of the IRES-driven translation of FMDV. Our study supports a model in which NCL relocalizes from the nucleus to the cytoplasm during the course of FMDV infection, where the cytoplasmic NCL promotes FMDV IRES-driven translation by bridging the translation initiation complexes with viral IRES. Our study demonstrates a previously uncharacterized role of NCL in the translation initiation of IRES-containing viruses, with important implications for the development of broad antiviral interventions.

INTRODUCTION

In eukaryotic cells, translation initiation occurs through two alternative mechanisms, a classical cap-dependent mechanism for most eukaryotic mRNAs and a 5′-end-independent mechanism driven by internal ribosome entry site (IRES) elements (1, 2). During canonical eukaryotic translation initiation, 43S preinitiation complexes (PICs) containing 40S subunits and the translation initiation factors eIF1, eIF1A, eIF3, eIF2/GTP/Met-tRNAi^Met, and probably eIF5 are first loaded onto the cap-proximal region of the mRNA by eIF4F, eIF4A, and eIF4B. After attachment, the 43S complex scans the mRNA downstream from the cap to the AUG start codon, where the 48S initiation complex is assembled. The subsequent displacement of eIF1 results in the eIF5-mediated hydrolysis of eIF2-bound GTP, the joining of 60S subunit and concomitant dissociation of other eIFs are mediated by eIF5B, and GTP hydrolysis by eIF5B triggers the release of eIF5B-GDP and eIF1A to yield the final 80S initiation complex (2, 3).

Unlike cap-dependent translation, the translation initiation driven by the IRES can occur in the absence of canonical cap recognition or binding and conventional ribosome scanning. IRES elements are characterized by complicated secondary and tertiary structures, assembled with stem-loops and pseudoknots, which mediate the internal binding of the ribosomes. This, in turn, initiates translation through multiple RNA-RNA and RNA-protein interactions (1–4). Numerous IRESs have been identified in the mRNAs of viruses, as well as in a subset of cellular mRNAs involved in stress responses, the cell cycle, and apoptosis (5, 6). Notably, IRES-driven translation can occur when the dominant cap-dependent translation route is shut down, and viruses can take advantage of this to usurp the host translation machinery and direct the synthesis of viral proteins (6–8).

Viral IRESs have been classified into four major groups on the basis of how they recruit the ribosome, the eIFs that they require to function, and similarities in their secondary structures. Type I and II IRESs, which are found exclusively in picornaviruses, have large and flexible architectures and depend on extensive cellular eIFs and IRES trans-acting factors (ITAFs) for efficient ribosome assembly, whereas type III IRESs (hepatitis C virus-like) require only a limited set of eIFs and ITAFs (9, 10) and type IV IRESs promote translation without any eIFs (6–8, 11).

Foot-and-mouth disease virus (FMDV), the prototypical member of the genus Aphthovirus within the family Picornaviridae, is a cytoplasmic RNA virus containing a positive-sense single-stranded genome of about 8,500 nucleotides (nt) in length, which is composed of the 5′ untranslated region (5′ UTR), an open reading frame, and the 3′ UTR (12, 13). The genomic RNA contains all the information required to subvert the cellular machinery and to shut down host protein synthesis in infected cells, where the viral products are translated in an IRES-driven manner. During FMDV infection, viral proteases L^pro and 3C^pro cleave several cellular proteins, including eIF4G, eIF4A, polyadenylate-binding protein (PABP), and pyrimidine tract binding protein (PTB) (14–16). These cleavages cause the rapid shutdown of host cap-dependent translation, thereby eliminating any competition for resources with viral protein synthesis via IRES-mediated translation.

The type II IRES of FMDV is one of the strongest IRESs described to date, and the function of FMDV IRES in mediating translation initiation has been extensively studied. Canonical initiation factors (such as eIF4A and eIF4G) and ITAFs (PTB and ITAF₄₅) that bind to the IRES are required to reconstitute the translation initiation process in vitro (17–19). A number of host RNA-binding proteins (RBPs) have also been suggested to act as ITAFs for picornavirus translation, and the roles of several novel ITAFs in the activities of specific IRESs have been confirmed by luciferase reporter analyses (for reviews of this work, see references 1, 6, 20, and 21). For example, heterogeneous nuclear riboprotein K (hnRNP K), hnRNP A1, and FBP1 enhance the IRES activity of enterovirus 71 (EV71) (22–24), whereas Gemin5, AUF1, and FBP2 are negative regulators of FMDV or EV71 translation (25–27). However, the exact mechanisms by which ITAFs regulate viral IRES-driven translation initiation have not yet been well defined.

In the present study, by using the RNA affinity and mass spectrometry (MS), we screened and determined that nucleolin (NCL) interacts with the FMDV IRES, which is consistent with previously published work (28). NCL is a multifunctional cellular RBP that has been heavily implicated in DNA metabolism and RNA regulatory mechanisms, including transcription, ribosome assembly, mRNA stability, and translation (29, 30). NCL is involved in the nuclear egress of viral proteins, viral DNA and RNA synthesis, and viral morphogenesis (31–36), as well as the entry of multiple viruses (37–39). Previous reports showed that NCL associates with the 5′ UTR or 3′ UTR of a virus genome, such as poliovirus (PV) and rhinovirus, to enhance the production of viral proteins (40–42). The details, however, remain unclear. Here, we show that NCL positively regulates the translation and replication of FMDV but is not required for cellular or vesicular stomatitis virus (VSV) cap-dependent translation. We also report that the translocalization of NCL to the cytoplasm is an important event for viral IRES-driven translation, specifically at the initiation stage of the translation process, and that NCL knockdown reduces the binding of the components of the translation initiation complexes to the FMDV IRES. The effect of NCL knockdown on FMDV infectivity in vivo was further evaluated. This study demonstrates a previously uncharacterized role of NCL in the translation of IRES-containing viruses.

RESULTS

NCL was screened for interaction with the FMDV IRES.

To identify the host RBPs involved in FMDV replication, a biotinylated RNA pulldown assay, followed by liquid chromatography-electrospray ionization-tandem mass spectrometry (LC-ESI-MS/MS), was used to isolate and identify cellular proteins that might associate with the FMDV 5′ UTR or IRES. As shown in Fig. 1A, several unique bands were detected in the mixtures of biotinylated FMDV 5′ UTR (lane 5), IRES (lane 9), and extracts of FMDV-infected PK-15 cells in the pulldown assay, relative to the control reactions, which contained no RNA (lane 2), biotin-16-UTP (lane 3), or nonbiotinylated FMDV RNA (lanes 4 and 8) with the cell lysates. The protein bands that were specifically associated with the biotinylated FMDV 5′ UTR or IRES were excised, digested with trypsin, and subjected to LC-ESI-MS/MS analysis. This procedure identified a substantial number of cellular proteins that interact with the FMDV 5′ UTR or IRES (see Tables S1 and S2 in the supplemental material), including several members of the hnRNP family (PTB, PCBP1, PCBP2, AUF1, hnRNP A1, and hnRNP K), which are ITAFs known to interact with picornavirus IRESs (6). The presence of these RBPs supports the validity of our experimental approach. Notably, NCL had 29 peptide matches, 29.4% coverage in the 5′ UTR, 27.6% coverage in the IRES, and an Unused ProtScore of 323.31 in the MS/MS data (Tables S1 and S2).

FIG 1.

NCL positively regulates the translation and replication of FMDV. (A) The FMDV 5′ UTR and IRES were transcribed in vitro, biotinylated, and incubated with FMDV-infected PK-15 cell lysate for 60 min at 4°C. After the associated cellular proteins were pulled down with streptavidin beads, they were eluted, boiled, and subjected to SDS-PAGE. Silver staining was used for visualization. The specific associations of NCL with the FMDV 5′ UTR (lanes 6 and 7) and IRES (lanes 10 and 11) were then confirmed by Western blotting. (B) PK-15 cells were transfected with the indicated siRNAs for 48 h or transfected with the empty vector or FLAG-NCL for 24 h, and then dead cells were estimated with the LIVE/DEAD viability/cytotoxicity kit (column) and cell viability was assessed with an MTS assay (scatter). eIF3a, an essential host factor for cell viability and proliferation, was used as the positive control. (C and D) PK-15 cells transfected with either the indicated siRNA or FLAG-tagged plasmid were infected with FMDV (MOI of 1), and cell lysates and supernatants were harvested at the indicated times. Viral titers were determined with a TCID₅₀ assay. (E and F) PK-15 cells transfected with either the indicated siRNA or FLAG-tagged plasmid were challenged with FMDV (MOI of 1). At various times postinfection, cell lysates were analyzed for the expression of NCL and viral proteins (VP0, VP1, VP2, and VP3) by Western blotting. (G) BSR-T7 cells treated with either the indicated siRNA or FLAG-tagged plasmid were transfected with the replicon rFMDV-EGFP. At 24 h posttransfection, the cells were subjected to fluorescence analysis. (H) PK-15 cells or BHK-21 cells were pretreated with medium alone (No Ab), control antibody (IgG), anti-NCL antibody (anti-NCL), or anti-integrin αvβ6 antibody (anti-αvβ6) for 1 h at 37°C. The cells were then exposed to FMDV (MOI of 0.5) and incubated at 37°C to allow infection to proceed. The cell supernatants and lysates were collected at 8 hpi, and the viral titers were determined by plaque assays. Data in panels B to D, G, and H are the means of results of three independent experiments, and error bars indicate standard deviations (SD). **, P < 0.01.

NCL is required for efficient viral replication of FMDV.

To determine the functional role of NCL in the FMDV replication cycle, we investigated the effect of NCL overexpression or knockdown on FMDV replication. In PK-15 and BHK-21 cells, the ectopic expression or knockdown of NCL had negligible effects on cell death or cell proliferation when viability assays were performed immediately before infection, unlike a small interfering RNA (siRNA) targeting eIF3a, which caused significant cell death (Fig. 1B). The viral titers were then measured to assess the impact of NCL on FMDV replication. NCL knockdown reduced the viral yields ∼13-, ∼21-, and ∼15-fold in PK-15 cells (Fig. 1C) and ∼18-, ∼28-, and ∼17-fold in BHK-21 cells (data not shown) at 4, 6, and 8 h postinfection (hpi), respectively. In contrast, the viral yield increased ∼9.5-, ∼5.8-, and ∼5.1-fold at 4, 6, and 8 hpi, respectively, in PK-15 cells ectopically expressing NCL (Fig. 1D). Similarly, the viral loads increased ∼9.9-, ∼4.9-, and ∼4.9-fold at 4, 6, and 8 hpi, respectively, in BHK-21 cells expressing NCL (data not shown). A Western blot analysis of viral protein production revealed a dramatic reduction in the expression of viral proteins VP0, VP1, VP2, and VP3 in the NCL knockdown cells (Fig. 1E). However, the production of viral proteins showed a clear increase when NCL was ectopically expressed (Fig. 1F). Collectively, these results suggest that NCL plays a proviral role during FMDV replication in both cell types.

Cell surface NCL facilitates the attachment and internalization of multiple viruses, including EV71 in the family Picornaviridae (37–39). To investigate whether NCL mediates FMDV entry, we bypassed viral entry by directly transfecting cells with an FMDV subgenomic replicon that expresses the enhanced green fluorescent protein (EGFP) reporter gene (rFMDV-EGFP). As shown in Fig. 1G, the expression level of specific fluorescence decreased dramatically in NCL-depleted BSR-T7 cells, whereas the replicon activity was elevated when NCL was ectopically expressed. Furthermore, the preincubation of cells with an anti-αvβ6 antibody significantly reduced the viral yields (∼14-fold) (Fig. 1H) and protein expression (data not shown) but had no inhibitory effect when cells were treated with anti-NCL antibody. Integrin αvβ6 is known to be a functional receptor for FMDV. Overall, our data indicate that NCL is specifically implicated in the translation and replication of FMDV.

NCL positively regulates IRES-driven translation of FMDV and indirectly affects viral RNA synthesis.

To determine whether NCL is an essential factor for FMDV IRES-driven translation, a bicistronic luciferase construct psiCHECK-FMDV, containing a cap-dependent Renilla luciferase (RLuc) gene and an FMDV IRES-dependent firefly luciferase (FLuc) gene, was used (Fig. 2A, top). We found that the RLuc value, which reflects cellular cap-dependent activity, was not affected when NCL was knocked down or ectopically expressed. However, FLuc activity, which reflects viral IRES-driven translation, was significantly reduced by the depletion of NCL and greatly elevated by the ectopic expression of NCL (Fig. 2A, bottom). We also demonstrated that the effect of NCL on FMDV IRES activity was comparable to the effect of hnRNP K, which is recognized as a positive ITAF for picornavirus translation (data not shown). Moreover, the reduction or increase in NCL expression did not affect the amount of bicistronic reporter RNA in the cells, indicating that NCL specifically affects translation, rather than transcription or RNA stability (data not shown). These results indicate that NCL functions as a positive regulator of FMDV IRES-driven translation.

FIG 2.

NCL positively regulates IRES-driven translation of FMDV and indirectly affects viral RNA synthesis. (A) Schematic illustration of bicistronic FMDV IRES construct (psiCHECK-FMDV) (top). PK-15 cells treated with either the indicated siRNA or FLAG-tagged plasmid were transfected with the bicistronic construct psiCHECK-FMDV. At 24 h posttransfection, the RLuc and FLuc activities were determined (bottom). (B) PK-15 cells transfected with siCtrl or siNCL were challenged with FMDV at an MOI of 1. At various times postinfection, total RNA was extracted and subjected to RT-qPCR analysis to determine the levels of total viral RNA and negative-strand RNA. (C and D) PK-15 cells transfected with siCtrl or siNCL were infected with FMDV (MOI of 1), and CHX (100 μg/ml) was added to the culture medium at 2 h or 3 h postinfection. RNA was extracted at the indicated times, and the levels of total viral RNA and negative-strand RNA were determined by RT-qPCR. Data are the means ± SD of results of three independent experiments. **, P < 0.01.

During FMDV infection, genomic RNA acts as the template for both translation and RNA replication, closely coupling these two processes (43). Therefore, it is tempting to speculate that restriction of translation contributes to the decrease in viral RNA. As expected, after infection with FMDV, a significant reduction in the levels of total and negative-strand viral RNAs was observed in the NCL knockdown cells at 4 to 8 hpi (Fig. 2B). To further distinguish between a direct and an indirect effect on viral RNA synthesis, the protein synthesis inhibitor cycloheximide (CHX) was used to separate the translation and RNA replication processes. As shown in Fig. 2C, when viral translation was inhibited at 2 hpi, there was no measurable difference in the viral RNA levels regardless of NCL depletion. However, when translation was blocked at 3 hpi, an obvious reduction in viral RNA synthesis was observed in the NCL-depleted cells, which may be attributable to an initial difference in the relative amounts of viral replication proteins and progeny RNAs present (Fig. 2D). Our results demonstrate that the effect of NCL on viral RNA synthesis is indirect, predominantly because the levels of viral proteins required for replication are reduced in NCL knockdown cells.

NCL specifically associates with FMDV IRES.

Because NCL is involved in viral IRES-driven translation, we next investigated the interaction between NCL and FMDV IRES in infected cells. After cells were infected with FMDV at a multiplicity of infection (MOI) of 1 for 5 h, they were lysed and their RNA-protein complexes were immunoprecipitated with an antibody specific to NCL or with isotype anti-immunoglobulin G (IgG). The total RNAs isolated from these immunoprecipitates were subjected to reverse transcription (RT)-PCR analysis to identify the FMDV IRES, coding regions, and 3′ UTR or the mRNAs for ribosomal proteins S16 (RPS16) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH; control). We identified the FMDV IRES, 3D gene, and the 3′ UTR in the samples immunoprecipitated with the antibody specific to NCL (Fig. 3A, lanes 3 and 8). However, RT-PCR did not detect RPS16 or GAPDH in the precipitated samples (Fig. 3B, lanes 3 and 8). No specific RT-PCR band was detected with the isotype IgG antibody (lanes 4 and 9), with no antibody (lanes 5 and 10), or when H₂O was used as the template (lanes 6 and 11). These results demonstrate that NCL specifically binds the FMDV genome in FMDV-infected cells.

FIG 3.

Interaction between NCL and FMDV IRES and their interactive regions. (A and B) Cell lysates were generated from PK-15 cells infected with FMDV (MOI of 1) at 5 h postinfection and subjected to an immunoprecipitation assay with an anti-NCL antibody or isotype anti-IgG antibody. After washing and dissociation, RNA was extracted and subjected to RT-PCR using primers specific to the FMDV IRES or 3D/3′ UTR (A) and RPS16 or GAPDH (B). (C) Schematic representation of FMDV genome (top). Lysates from FMDV-infected PK-15 cells were incubated separately with the biotin-labeled 5′ UTR, S fragment, cre, IRES, or 3′ UTR of FMDV. Nonbiotinylated RNA probes were used as controls. After the protein complexes were pulled down with streptavidin beads, they were dissolved for Western blotting (bottom). (D) Various amounts of unlabeled FMDV IRES (lanes 3 to 5) and VP1 RNA (lanes 8 to 10) probes were added to compete with the biotinylated FMDV IRES probe that interacted with NCL. Lanes 1 and 6 contain cell lysate only. (E) Schematic representation of the FMDV IRES structure (left). Various truncated forms of the FMDV IRES, as indicated, were transcribed in vitro and biotinylated. Lysates from FMDV-infected cells expressing or not expressing FLAG-NCL were used to perform the biotinylated RNA pulldown assay. (F) Schematic representation of NCL and various truncated mutants. Cell extracts from PK-15 cells expressing the indicated proteins were collected at 36 h posttransfection and incubated with the biotinylated FMDV IRES (lanes 3, 6, 9, 12, and 15) or nonbiotinylated IRES (lanes 2, 5, 8, 11, and 14). The input (lanes 1, 4, 7, 10, and 13) was used to examine protein expression. (G) RNA-protein pulldown assays were performed with nonbiotinylated or biotinylated IRES and recombinant NCL-RC (lanes 1 and 2), NCL-R (lanes 3 and 4), or NCL-C (lanes 5 and 6). The bound complexes were analyzed by immunoblotting, and the input was detected by Coomassie brilliant blue staining.

To further assess the reciprocal interaction between NCL and the FMDV genome and to define the functional elements in the FMDV genome underlying this interaction, we used RNA affinity assays with biotinylated RNA transcripts corresponding to the FMDV 5′ UTR, S fragment, cre, IRES, or 3′ UTR and extracts from FMDV-infected cells. The RNA-protein complexes were pulled down with streptavidin beads and analyzed by immunoblotting. NCL was pulled down by biotinylated 5′ UTR or IRES (Fig. 3C). Furthermore, the interaction between NCL and the biotinylated FMDV IRES was outcompeted by the nonbiotinylated FMDV IRES (lanes 2 to 5) but not noticeably by VP1 RNA (lanes 7 to 10) (Fig. 3D). Collectively, these results confirm the association between NCL and the IRES in the FMDV 5′ UTR.

Interactive regions between truncated IRES and NCL protein.

To confirm and clarify the interaction between NCL and the FMDV IRES in greater detail, the binding regions in both the viral RNA and NCL were mapped. A series of truncated forms of the viral IRES were transcribed and biotinylated in vitro. Biotin-RNA pulldown assays were then performed with extracts from FMDV-infected cells ectopically expressing NCL. A Western blot analysis showed that biotinylated probes d3-5, d4-5, and d5 pulled down ectopically expressed or endogenous NCL, whereas the other probes did not (Fig. 3E). These results suggest that domain 5 of the FMDV IRES is required for its association with NCL, consistent with previous reports (28).

NCL is composed of three functional domains: an N-terminal domain with multiple phosphorylation sites, a central domain containing four RNA recognition motifs, and a C-terminal domain rich in arginine and glycine residues (29, 30). To determine which of these domains is involved in its interaction with the FMDV IRES, PK-15 cells were transfected with plasmids expressing different truncated forms of NCL fused to FLAG, and the cell lysates were used for RNA-protein pulldown assays. It is shown that the biotinylated FMDV IRES captured full-length NCL and two truncated forms, NCL-RC and NCL-R, but not NCL-N or NCL-C (Fig. 3F). To determine whether NCL binds the FMDV IRES directly, we performed RNA affinity assays using biotinylated FMDV IRES and recombinant NCL. Because it was not possible to express full-length NCL, probably because it contains a highly acidic N-terminal domain (41, 44), the truncated forms of NCL (NCL-RC, NCL-R, and NCL-C) were expressed and incubated with the biotinylated IRES. As shown in Fig. 3G, the biotinylated IRES specifically interacted with recombinant NCL-RC or NCL-R. These results demonstrate that the RNA-binding domain of NCL is sufficient for it to interact directly with the FMDV IRES.

Cytoplasmic translocalization of NCL is required for efficient IRES-driven translation of FMDV.

Although NCL is a shuttling protein, in uninfected cells it localizes predominantly in the nucleolus. Upon PV infection, NCL undergoes dramatic nucleolar-cytoplasmic relocalization (40). During FMDV replication, we found that NCL is gradually redistributed from the nucleus to the cytoplasm, as expected, especially after 4 hpi and 6 hpi (Fig. 4A). The translocalization of NCL was also evaluated by nuclear-cytoplasmic fractionation, and we found that the accumulation of NCL in the cytoplasm of infected cells increased at 4 hpi and then continually increased over the course of infection (Fig. 4B). These results demonstrate the cytoplasmic translocalization of NCL during FMDV infection.

FIG 4.

Translocalization of NCL from the nucleus to the cytoplasm during FMDV infection supports viral IRES-driven translation. (A) PK-15 cells were mock or FMDV infected (MOI of 5) and then fixed at 4 or 6 hpi. The cells were permeabilized and then probed with indirect immunofluorescence for FMDV (green) and NCL (red). Nuclei are indicated by DAPI staining (blue). Cells were observed by confocal microscopy. (B) Nuclear and cytoplasmic fractions from PK-15 cells infected with FMDV for the indicated times were subjected to Western blot analysis with the indicated antibodies. (C) PK-15 cells were transfected with various plasmids expressing FLAG-tagged 3C^pro-WT or variants 3C^pro-H46Y, -D84N, -C163G, and -H205R or the FLAG vector. At 24 h posttransfection, the cells were fixed and incubated with anti-FLAG and anti-NCL antibodies and then with secondary antibodies conjugated with FITC (green) and TRITC (red), respectively. Nuclei were counterstained with DAPI (blue), and localization was determined using confocal microscopy. (D) Schematic diagram of capped bicistronic mRNAs including viral IRES. (E and F) Schematic representation of the transfection of bicistronic mRNAs, followed by infection. Control and NCL-depleted PK-15 cells were transfected with capped bicistronic mRNA siCHECK-FMDV. Immediately after transfection, the cells were mock infected or infected with FMDV (MOI of 1) and the RLuc and FLuc activities were determined at 6 h postinfection. In panel F, the data are the means ± SD of results of three independent experiments. **, P < 0.01.

To determine whether NCL translocalization is related to the regulation of viral protein expression, we examined the subcellular location of NCL in cells ectopically expressing FLAG-tagged viral proteins. The ectopic expression of wild-type (WT) 3C resulted in the partial redistribution of NCL compared to expression of vector (Fig. 4C), whereas the expression of other viral proteins did not (data not shown). To determine the requirement for proteinase activity in the redistribution of NCL induced by 3C^pro, several previously described mutated forms of 3C^pro were expressed (45), and the localization of NCL was examined. When the proteinase activity of 3C^pro was abolished (with mutations H46Y, D84N, or C163G), no relocalization of NCL was detected. In contrast, the 3C-H205R mutant, which retains its enzyme activity, induced NCL translocalization (Fig. 4C). These data suggest that FMDV 3C^pro induces cytoplasmic translocalization of NCL, which requires the proteinase activity of 3C^pro.

Because NCL binds to the FMDV IRES and promotes viral translation in the cytoplasm, we predicted that the nucleocytoplasmic trafficking of NCL in infected cells plays a role in FMDV IRES-driven translation. To confirm this possibility, cells were transfected with the bicistronic reporter mRNA (Fig. 4D), which was transcribed in vitro in the presence of cap analog, before infection (or not) with FMDV. The luciferase activity was measured at 6 hpi, when NCL had relocalized to the cytoplasm (Fig. 4E). When uninfected cells were transfected with bicistronic mRNA siCHECK-FMDV, there was no difference in cap-driven translation between the control and NCL-depleted cells, whereas a significant reduction in FMDV IRES-driven translation was detected in the NCL-depleted cells (Fig. 4F), consistent with the observations made with the bicistronic luciferase plasmid psiCHECK-FMDV. After infection, a significant increase (∼4-fold) in FMDV IRES-driven translation was measured in the control cells, whereas a small increase (∼2.4-fold) was observed in the NCL-depleted cells (Fig. 4F), suggesting that NCL had a more profound effect on FMDV IRES-driven translation in infected cells than in uninfected cells. These results indicate that relocalization of NCL from the nucleus to the cytoplasm during FMDV infection is required for efficient viral translation.

NCL is a positive modulator for translation of IRES-containing picornaviruses and flaviviruses.

To determine whether NCL knockdown affects the translation from other viral IRESs, the IRES activities of Seneca Valley virus (SVV) (type III-like IRES) and classical swine fever virus (CSFV) (type III IRES) were measured in NCL-depleted cells. The results of the bicistronic plasmid experiment showed that the IRES activities of SVV and CSFV decreased significantly, to 59% and 51%, respectively, after NCL knockdown (Fig. 5A). Furthermore, NCL depletion greatly depressed the viral yields of SVV (Fig. 5B) and CFSV (Fig. 5C), by 11-fold at 12 hpi and 5.5-fold at 48 hpi, respectively. The RNA levels of SVV and CFSV were also reduced (data not shown). In contrast, the depletion of NCL did not affect the viral yields (Fig. 5D) or the RNA levels of VSV (data not shown), a virus that depends solely on cap-dependent translation (46). The direct transfection of dicistronic reporter mRNAs was also performed. As shown in Fig. 5E and F, a strong reduction in both SVV and CSFV IRES-driven translation was observed in NCL-depleted cells, and following FMDV infection, there is a stronger effect of NCL depletion on viral IRES activity. Overall, these results suggest that NCL is a conserved determinant of the translation and infection of IRES-containing picornaviruses and flaviviruses and that the cytoplasmic relocalization of NCL contributes to efficient viral translation.

FIG 5.

NCL knockdown suppresses translation and infection of SVV and CSFV. (A) PK-15 cells treated with either control siRNA (siCtrl) or NCL-targeting siRNA (siNCL) were transfected with the indicated bicistronic plasmids (psiCHECK-SVV or psiCHECK-CSFV), and the RLuc and FLuc activities were determined at 24 h posttransfection. (B and C) Control and NCL-depleted IBRS-2 cells or ST cells were infected with SVV (MOI of 0.5) or CSFV (MOI of 0.05), respectively. At the indicated times, the supernatants and cell lysates were collected and viral yields were determined by the TCID₅₀ assay. (D) Control and NCL-depleted Vero cells were challenged with VSV-EGFP (MOI of 0.5). At 8 h postinfection, cells were monitored by fluorescence microscopy (left), and viral yields at the indicated time points were determined by the TCID₅₀ assay (right). (E and F) Control and NCL-depleted PK-15 cells were transfected with capped bicistronic mRNA siCHECK-SVV and siCHECK-CSFV, respectively. Immediately after transfection, cells were mock infected or infected with FMDV (MOI of 1) and the RLuc and FLuc activities were determined at 6 h postinfection. Data are the means ± SD of results of three independent experiments. **, P < 0.01.

NCL positively regulates translation initiation of FMDV mRNA.

To investigate the distribution of NCL on ribosomal subunits, we subjected extracts of mock- or FMDV-infected PK-15 cells to sucrose gradient fractionation to separate the ribosomal subunits (40S and 60S), monosomes (80S), and polysomes. As shown in Fig. 6, the peaks in the polysome profiles were detected by measuring the OD₂₅₄ profiles in the gradient fractions and by probing an immunoblot for RPS6 (40S component) and RPLP0 (60S component). As shown in Fig. 6A to C, probing for NCL revealed that most of the protein sedimented in the upper and 40S gradient fractions and that a small amount tailed into the 60S and 80S fractions in both the mock-infected and FMDV-infected samples. Interestingly, much more NCL appeared to be in 80S/monosome fractions during infection. However, NCL was not detected in the polysome fractions, indicating that the protein does not remain associated with actively translating ribosomes. Moreover, the sedimentation of NCL somewhat resembled the sedimentation of eIF2α, a canonical translation initiation factor, suggesting that NCL participates in the initiation phase rather than the elongation phase of translation.

FIG 6.

NCL positively regulates translation initiation of FMDV mRNA but not that of cellular GAPDH mRNA. (A to C) Extracts were generated from mock-infected or FMDV-infected (MOI of 1) PK-15 cells (3 or 6 h postinfection), sedimented through 10%-to-50% sucrose gradients, and fractionated. Polysome profiles were generated by measuring the OD₂₅₄ in the gradient fractions (left). Collected fractions were also subjected to Western blotting to detect cosedimentation of NCL with ribosomal subunits (40S and 60S), monosomes (80S), or polysomes (right). (D to K) PK-15 cells were transfected with either control siRNA or siRNA targeting NCL for 48 h (D to G) or were transfected with the FLAG vector or FLAG-NCL for 24 h (H to K) and challenged with FMDV (MOI of 1). At 4 h postinfection, the cell lysates were resolved and fractionated through sucrose gradients. Polysome profiles were generated by measuring the OD₂₅₄ of the gradient fractions (D and H). The fractions were then subjected to RT-qPCR analysis of the FMDV 3D and GAPDH mRNA levels. The transcript number is expressed as the percentage of total mRNA transcripts recovered and is plotted against the fraction number (E, F, I, and J). The fractions were also subjected to Western blotting to determine the distribution of NCL on the ribosomal subunits (G and K).

To determine the mechanism by which NCL controls translation, we compared the formation of ribosomal complexes on FMDV mRNA with their formation on cellular transcripts of GAPDH. Cells ectopically expressing or depleted of NCL were infected with FMDV, and the lysates were fractionated through sucrose gradients. Following fractionation, the association of the ribosomal subunits with specific mRNAs was determined with a reverse transcription-quantitative PCR (RT-qPCR) analysis, and the polysomal distribution of NCL was probed by immunoblotting. FMDV mRNA levels peaked in fraction 14 of the control cells, but the distribution of FMDV mRNA shifted toward lighter fractions when NCL was knocked down, peaking in fraction 11. The distribution of FMDV mRNA clearly shifted toward heavier fractions after NCL overexpression. Probing for NCL showed that the protein was no longer present in the 40S, 60S, and 80S fractions in the NCL-depleted cells, whereas much more NCL appeared in these fractions when NCL was overexpressed, which is consistent with the RT-qPCR results (Fig. 6D to K). These results indicate that NCL supports FMDV mRNA in forming polysomes for effective translation. In contrast, the distribution of GAPDH mRNA was not affected by a reduction or increase in NCL protein levels. Furthermore, the polysome profiles in the mock-infected cells showed that canonical global translation was not compromised by a reduction or increase in NCL (data not shown). Taken together, these results demonstrate that NCL promotes the initiation of translation of FMDV mRNAs but not that of canonical cellular transcripts.

NCL is involved in binding of translation initiation complex to FMDV IRES.

FMDV shuts down cap-dependent translation, allowing the initiation of IRES-driven translation to promote viral protein synthesis by recruiting ribosomes, eIFs, and TAFs (14–16). Because NCL appears to promote the initial steps of ribosome binding to the FMDV IRES (Fig. 6), we reasoned that it must specifically interact with components of the translation initiation complexes during IRES-mediated translation. To explore this hypothesis, we immunoprecipitated NCL from lysates of mock- or FMDV-infected PK-15 cells and analyzed the precipitates by Western blotting to detect any associated translation initiation factors. In this study, we demonstrated that eIF4G and eIF5B are cleaved in FMDV-infected PK-15 cells (Fig. 7A, lanes 1 and 2). A Western blot analysis of the immunoprecipitation complex showed that NCL specifically coprecipitated with the majority of eIFs, including eIF4G (cleaved eIF4G), eIF4A, eIF3a, eIF3e, eIF3j, and eIF2α, but not with eIF5B. NCL was also associated with PABP and the 40S component RPS6 (Fig. 7A, lanes 3 and 4). These results support the proposition that NCL specifically interacts with the translation initiation complexes, consistent with previous reports that identified nucleolin in hepatitis C virus (HCV) IRES initiation complexes (47).

FIG 7.

NCL supports the binding of translation initiation machinery to FMDV IRES. (A) Mock-infected and FMDV-infected PK-15 cells were immunoprecipitated at 4 hpi with an anti-NCL antibody (lanes 3 and 4), and the associated proteins were analyzed by Western blotting. The input (lanes 1 and 2) corresponds to 1/10 of the total cell extract. (B) PK-15 cells transfected with either control siRNA or siRNA targeting NCL for 48 h were challenged with FMDV (MOI of 1). At 4 h postinfection, the cell lysates from mock- or FMDV-infected cells were generated and subjected to an RNA pulldown assay with the biotinylated FMDV IRES. The input (lanes 1 to 4) and the precipitates (lanes 5 to 8) were analyzed by Western blotting with the indicated antibodies. (C) Translational extracts prepared from FMDV-infected BHK-21 cells were preincubated with 0, 0.1, or 0.2 μg of control IgG, anti-NCL antibody (α-NCL), or anti-PTB antibody (α-PTB) and then incubated with capped bicistronic mRNA siCHECK-FMDV for 90 min at 30°C. The in vitro translation products were then analyzed by luciferase activity assays. Data are the means ± SD of results of three independent experiments. **, P < 0.01.

To further clarify the specific function of NCL in the initiation of FMDV IRES-driven translation, a biotin-RNA affinity assay was performed with extracts of mock- or FMDV-infected PK-15 cells. As shown in Fig. 7B, the depletion of NCL did not affect the expression levels of the indicated eIFs (lanes 1 to 4). The specific binding of eIF4G, eIF4A, eIF3a, eIF3e, eIF3j, eIF2, PABP, or RPS6, but not eIF5B, to the biotinylated FMDV IRES was detected (lanes 5 and 7). However, these specific interactions were greatly impaired when NCL was depleted (lanes 6 and 8). These observations suggest the importance of NCL in the initiation of FMDV IRES-driven translation.

We also performed an in vitro translation assay using bicistronic mRNA containing the FMDV IRES (siCHECK-FMDV) in the presence or absence of an anti-NCL antibody. As shown in Fig. 7C, the addition of the anti-NCL antibody, but not the control IgG, reduced IRES-driven translation to 54% to 62%. An anti-PTB antibody was included in a parallel assay as a positive control. However, the addition of anti-NCL antibody did not affect cap-dependent translation. These results suggest a direct role for NCL in IRES-mediated translation.

Taken together, our results demonstrate that NCL supports the binding of translation initiation complexes to the FMDV IRES and therefore promotes viral translation and infection.

Silencing NCL reduces the susceptibility of suckling mice to FMDV infection.

To test the functional relevance of NCL in FMDV infection in vivo, we established NCL-depleted suckling mice by subcutaneous inoculation of the mice with pAdM-shNCL, an adenovirus expressing four specific short-hairpin RNAs (shRNAs) targeting NCL. At 72 h after inoculation, the expression levels of NCL mRNA and protein were significantly reduced in the different tissues of the pAdM-shNCL-treated mice compared with those of pAdM-shCtrl-treated mice (Fig. 8A and B). The suckling mice were then challenged with 20 median lethal doses (LD₅₀) of FMDV at the inoculation site, and the survival of the infected mice was recorded (Fig. 8C). All the phosphate-buffered saline (PBS)- and pAdM-shCtrl-treated mice died within 6 days of infection, whereas 90% of the NCL-depleted mice remained alive on day 6 postinfection and 70% of the mice survived to the end of the experiment (Fig. 8D). Consistent with these observations, the viral titers in the muscle tissues of the NCL knockdown mice were approximately 10^3.2-fold lower than those in the control mice (Fig. 8E). Tissue samples collected from the lungs, hearts, livers, and kidneys of the suckling mice at 5 days postinfection were subjected to pathological analysis by hematoxylin and eosin (H&E) and immunohistochemical (IHC) staining (Fig. 8F). In the tissues of the infected pAdM-shCtrl-treated mice, extensive edema and inflammatory cell infiltration were observed around the vessels in the lungs (Fig. 8F, a1), severe myocardial deformation was detected in the hearts (b1), hepatocytes displayed severe vacuolar degeneration and spotted and focal necrosis in the livers (c1), and denaturalization, necrosis, and abscission were detected in the renal tubule epithelial cells in the kidneys (d1). The IHC analysis showed that FMDV antigen was diffusely distributed in these tissues (a2 to d2). By contrast, no obvious pathological changes (a3 to d3) or viral antigen (a4 to d4) were detected in the tissue samples from the NCL knockdown mice. These results suggest that the depletion of NCL depressed the production of FMDV in vivo, rendering the mice less susceptible to FMDV infection.

FIG 8.

Suckling mice with NCL knockdown are less susceptible to FMDV. (A and B) Suckling mice inoculated with 1.0 × 10⁸ PFU of purified pAdM-shCtrl or pAdM-shNCL for 72 h were killed, and their muscles, hearts, livers, lungs, and kidneys were homogenized. RT-qPCR and Western blotting were used to determine the mRNA and protein levels of NCL in the tissues. (C) Schematic representation of the experimental procedure. Suckling mice were inoculated with 1.0 × 10⁸ PFU of purified pAdM-shCtrl or pAdM-shNCL in 100 μl of PBS. At 72 h after inoculation, the mice were challenged with 20 LD₅₀ of FMDV strain O/BY/CHA/2010. (D) Survival rates of the PBS-, pAdM-shCtrl-, and pAdM-shNCL-treated suckling mice (n = 10 each) infected with FMDV. Suckling mice were monitored for up to 14 days. (E) Viral challenge of the suckling mice showed viral loads in the muscle tissues of the pAdM-shCtrl- and pAdM-shNCL-treated mice. Each symbol represents an individual mouse. (F) Histopathological and IHC analyses of tissues from pAdM-shCtrl- and pAdM-shNCL-treated suckling mice infected with FMDV for 5 days. Pathological changes in the lungs (a1), hearts (b1), livers (c1), and kidneys (d1) were more serious in the pAdM-shCtrl-treated mice than in their pAdM-shNCL-treated counterparts (a3 to d3). IHC analysis indicated that the viral antigen was diffusely distributed in the tissues of the pAdM-shCtrl-treated mice (a2 to d2), whereas less antigen was detected in the pAdM-shNCL-treated mice (a4 to d4). Magnification for H&E and IHC, ×400. In panels A, B, and E, the data are the means ± SD of results of three independent experiments, **, P < 0.01.

DISCUSSION

FMDV has limited coding capacity in its own genome and therefore employs various strategies to coopt key steps in the cellular gene expression pathway; appropriation of the host translation machinery is a critical step in viral propagation (6). In the current study, we found that NCL interacts with FMDV IRES both in infected cells and in vitro but lacks direct or indirect interaction with other elements in the FMDV 5′ UTR or 3′ UTR. Functional analysis underscored the proviral function of NCL during FMDV infection and identified its specific functions on IRES-driven translation of FMDV, SVV, and CSFV. Our findings also support a model in which NCL relocalizes from the nucleus to the cytoplasm during FMDV infection and then cytoplasmic NCL promotes the initiation of viral IRES-driven translation by supporting the binding of translation initiation complexes to FMDV IRES (Fig. 9). The present work defines a previously uncharacterized role for NCL in the initiation of translation by IRES-containing viruses and identifies a target for the development of broad-spectrum antiviral interventions.

FIG 9.

Proposed model of positive NCL regulation of FMDV infection. Following FMDV infection, NCL translocates from the nucleus to the cytoplasm by an as yet incompletely defined mechanism, possibly linked to the proteinase activity of 3C^pro. Once in the cytoplasm, NCL promotes the initiation of FMDV IRES-driven translation by supporting the binding of translation initiation complexes to viral IRES. The promotion of viral IRES-driven translation may also contribute indirectly to the positive effect of NCL on viral RNA synthesis. However, NCL is not involved in FMDV binding or entry.

Most ITAFs are cellular RBPs, and their nucleolar-cytoplasmic redistribution often occurs when the shuttling ITAFs are trapped by viral IRESs, or through their proteolytic cleavage, which either removes the nuclear localization signal (NLS) of nuclear-resident proteins or degrades several nucleoporin proteins in the nuclear pore complex (NPC) during picornavirus infection (6, 48, 49). In the present study, we found that the redistribution of NCL from the nucleus to the cytoplasm is probably dependent on the proteinase activity of FMDV 3C^pro, and this protease does not lead to NCL cleavage. This suggests that the redistribution of NCL is a consequence of the dysregulation of the barrier function of the NPC caused by FMDV 3C^pro. Several members of the hnRNP family, such as PTB and hnRNP K, have also been shown to relocalize to the cytoplasm of FMDV-infected cells (data not shown). However, another hnRNP member, RALY, remains within the nucleus throughout FMDV infection (data not shown). Together, these data suggest some specificity and selectivity in the redistribution of nuclear proteins. NCL phosphorylation events are also linked to its cytoplasmic localization and enhanced RNA-binding activity (29, 50, 51). The effects of these processes on FMDV infection remain to be determined.

Given the length and flexibility of type II FMDV IRES, many ITAFs function as RNA chaperones to reconstruct or maintain the proper IRES conformation for ribosome assembly. ITAF45 binds specifically to a central domain (domain 3) of the FMDV IRES, and PTB specifically binds to domain 5 of the IRES. These two ITAFs individually and synergistically cause localized conformational changes in the adjacent regions, promote the stable binding of eIF4G/4A to the IRES J-K domain (domain 4), and possibly support efficient ribosomal recruitment (17, 19, 52). In contrast, Gemin5 has been identified as a negative regulator of FMDV IRES-driven translation (26, 53, 54). Surprisingly, hnRNP L interacts with domains 4-5 of the FMDV IRES to inhibit viral RNA synthesis, without affecting IRES-dependent translation (55). In the present study, we have shown that the central domain of NCL, which has RNA-binding capacity, is responsible for the direct interaction of NCL with domain 5 of the FMDV IRES. Domain 5 contains an Yn-Xm-AUG motif at its 3′ border, in which the Yn pyrimidine tract (n = 8 to 10 nt) is separated by a spacer (n = 18 to 20 nt) from an AUG triplet. Domain 5 is also recognized by eIF3, eIF4B, PTB, and Gemin5, and the AUG triplet is located at the 3′ border of the IRES to initiate translation, indicating that this region plays an essential role in the initiation of translation involving the FMDV IRES (52, 53, 56, 57). Interestingly, truncated forms of NCL, which lack the N-terminal domain, the C-terminal domain, or both, can act as dominant negative regulators and effectively inhibit viral protein synthesis (data not shown). These results further confirm the IRES-binding ability of the central domain of NCL and suggest the involvement of both the N and C domains in the protein-protein interactions that support IRES-mediated internal initiation. Importantly, the depletion of NCL greatly reduced the binding of eIF4G, eIF4A, and the core of the 43S PIC to the FMDV IRES, and blocking NCL with an NCL-specific antibody strongly reduced IRES-driven translation in vitro. These findings support the proposition that NCL cooperates with or outcompetes other factors to promote the relative reorientation of domain 5 of the FMDV IRES and thus enhances the stable binding of eIF4G/4A to the IRES and induces the conformational changes in the IRES that are required for the productive attachment of the 43S complex and translational activation. The present study provides evidence for a model in which NCL acts as a bridge that supports the binding of the translation initiation apparatus to viral IRESs to promote IRES-driven translation.

A polysome profile analysis showed that NCL is crucial for the formation of elongation-competent ribosomes on FMDV mRNAs. NCL was present mainly in the 40S fractions, only a small amount of the protein was found at 60S and 80S, and NCL did not cofractionate with polysomes. The sedimentation of NCL partly resembled that of the canonical translation initiation factor eIF2α. Given these data, it is tempting to speculate that NCL contributes to the initiation of translation but not to translational elongation during FMDV IRES-mediated translation. This hypothesis is consistent with the findings that NCL coprecipitated with most components of the 43S PIC (eIF3a, eIF3e, eIF3j, and eIF2α, and the 40S ribosomal subunits), eIF4G, and eIF4A and that specific interactions are required for the recruitment of the 43S PIC to the FMDV IRES. In addition, NCL was able to interact with components of the 43S PIC regardless of the presence of viral infection. This supports the assumption that NCL acts as a component of the 43S PIC and is a part of the 48S translation initiation complex and that FMDV takes advantage of this to initiate the translation of the viral mRNAs. Furthermore, the level of NCL expression was increased in the 40S, 60S, and 80S fractions after FMDV infection, which is consistent with the relocalization of NCL in response to viral infection.

It is not yet clear whether NCL supports the formation of the 80S translation apparatus on viral mRNAs through its direct or indirect interaction with the 60S ribosomal subunits. There are growing lines of evidence arguing for this possibility. First, it appears that NCL is enriched in 80S/monosome fractions during FMDV infection. Second, given its well-known affinity for ribosomal proteins, NCL could help in the assembly of the ribosomal subunits and probably participates in the transport of these newly assembled ribosomal subunits into the cytoplasm (29, 30, 58). Third, previous studies have demonstrated that NCL and RPL26 interact with each other and that this interaction plays a critical role in the regulation of translation and the control of TP53 mRNA after DNA damage (59, 60). Finally, our previous study showed that RPL13, an essential factor for FMDV IRES-dependent translation initiation (61), may specifically interact with NCL during FMDV infection (data not shown). A comprehensive study of the NCL interactome might extend our understanding of this challenging issue. In particular, a fractionation and mass spectrometry analysis of the 40S, 60S, and 80S gradient fractions from FMDV-infected cells is required to identify the NCL-interacting proteins involved in the formation of the translation initiation complexes.

MATERIALS AND METHODS

Ethics statement.

The animal experiments were performed at the Biosafety Level 3 Laboratory of Lanzhou Veterinary Research Institute, Chinese Academy of Agricultural Sciences (permission number SYXK-GAN-2018-0005). All animals were handled in strict accordance with good animal practice, according to the Animal Ethics Procedures and Guidelines of the Ministry of Science and Technology of the People’s Republic of China. The study was approved by the Animal Ethics Committee of Lanzhou Veterinary Research Institute, Chinese Academy of Agricultural Sciences (no. LVRIAEC2018-005).

Cells and viruses.

BHK-21 (baby hamster kidney; American Type Culture Collection [ATCC] CCL-10), PK-15 (porcine kidney; ATCC CCL-33), IBRS-2 (porcine kidney; ATCC CRL-1835), Vero (African green monkey kidney; ATCC CCL-81), and ST (swine testicular; ATCC CRL-1746) cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco, CA, USA) supplemented with 10% fetal bovine serum (FBS) (Gibco), penicillin (100 U/ml), and streptomycin (100 μg/ml) (Gibco) at 37°C under 5% CO₂. BSR-T7 cells (a kind gift from Karl-Klaus Conzelmann [62]), which stably express T7 RNA polymerase, were grown in DMEM supplemented with 10% FBS, 100 U/ml penicillin, 100 μg/ml streptomycin, and 1 mg/ml G418 (Life Technologies, CA, USA).

FMDV serotype O strain O/BY/CHA/2010 (GenBank accession no. JN998085.1) was maintained by the OIE/National Foot-and-Mouth Disease Reference Laboratory (Lanzhou, China). FMDV was propagated in BHK-21 cells, and the viral titers were determined with a 50% tissue culture infective dose (TCID₅₀) assay in BHK-21 cells. SVV (accession no. KY747511.1), propagated in IBRS-2 cells, and CSFV (accession no. AY805221.1), propagated in ST cells, are stored in our laboratory. A reporter plasmid expressing VSV-EGFP was generously provided by Zhifang Zhang.

Antibodies.

Anti-NCL, anti-RPS6, anti-RPLP0, anti-eIF2α, anti-eIF3a, anti-eIF3e, anti-eIF3j, anti-eIF4GI, and anti-PTB antibodies were purchased from Abcam (Cambridge, MA, USA). Anti-eIF4A, anti-eIF5B, anti-FLAG, anti-glutathione S-transferase (anti-GST), anti-α-tubulin, anti-lamin B1, and anti-β-actin antibodies were purchased from Santa Cruz Biotechnology (CA, USA). Anti-integrin αvβ6 antibody was purchased from EMD Millipore (Temecula, CA, USA). The secondary antibodies, conjugated with horseradish peroxidase (HRP), fluorescein isothiocyanate (FITC), or tetramethylrhodamine isothiocyanate (TRITC), were purchased from Sigma-Aldrich (St. Louis, MO, USA). Polyclonal pig antiserum directed against FMDV was prepared in our laboratory. Anti-FMDV 3B and anti-CSFV E2 monoclonal antibodies were kindly provided by Zengjun Lu and Hanchun Yang, respectively.

RNA interference (RNAi).

Small interfering RNAs (siRNAs) targeting candidate genes and negative-control siRNA were synthesized by GenePharma (Shanghai, China). The sequences of the siRNAs were: porcine NCL, 5′-GCUUCGUUUGAAGAUGCUA-3′; hamster NCL, 5′-GCGAAAGCAUUGGUAGCAA-3′; Chlorocebus sabaeus NCL, 5′-CCUUGGAACUCACUGGUUU-3′; porcine eIF3a, 5′-GCAGAUGGUCUUAGACAUATT-3′; hamster eIF3a, 5′-GGAACUGUGUGUGGAUCUU-3′; and negative-control siRNA, 5′-UUCUCCGAACGUGUCACGU-3′. Cells grown to 70% confluence were transfected with siRNAs using Lipofectamine RNAi MAX (Invitrogen, CA, USA), according to the manufacturer’s instructions.

Plasmid constructs.

To generate full-length and truncated forms of NCL, the corresponding cDNAs were amplified from PK-15 or BHK-21 cells by conventional RT-PCR. These cDNAs were then cloned into the EcoRI and XhoI sites of the pCMV-N-FLAG vector (Beyotime Biotechnology, Shanghai, China) to generate FLAG-tagged constructs or were inserted into the BamHI and XhoI sites of the pGEX-4T-1 vector (Amersham Biosciences, NJ, USA) to generate GST-tagged constructs. The genes of the FMDV structural and nonstructural proteins were amplified from the cDNA of FMDV strain O/BY/CHA/2010 and subcloned into the pXJ41-FLAG vector (Invitrogen). Site-directed mutagenesis of FLAG-3C^pro (H46Y, D84N, C163G, and H205R) was performed with the QuikChange II site-directed mutagenesis kit (Agilent Technologies, CA, USA). The FMDV-EGFP replicon, rFMDV-EGFP, was described in a previous study. All DNA constructs were verified with DNA sequencing.

Cell viability assay.

After cells were grown to 70% confluence in 96-well plates, they were transfected with specific siRNAs for 48 h or transfected with specific FLAG-tagged plasmids for 24 h. To determine the percentage of dead cells, the cells were stained for membrane integrity and intracellular esterase activity using the LIVE/DEAD viability/cytotoxicity kit (Invitrogen). After the cells were washed twice with PBS, they were incubated with the staining reagents calcein AM and ethidium homodimer at room temperature for 30 min. Images of the stained cells were obtained with an Olympus IX73 inverted microscope. For the 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt (MTS) assay, 10 μl of CellTiter 96 AQ_ueous one solution cell proliferation assay reagent (Promega, WI, USA) was directly added to the cells, which were then incubated for 4 h. The absorbance at 490 nm was recorded.

Immunofluorescence assay.

Cells grown on coverslips were infected with FMDV at an MOI of 5, and at specific times after infection, the cells were fixed with 4% paraformaldehyde for 15 min. To visualize the colocalization of endogenous NCL and overexpressed viral proteins, PK-15 cells seeded on coverslips were transfected with plasmids expressing FLAG-tagged viral proteins and fixed in 4% paraformaldehyde at 24 h posttransfection. After two washes with PBS, the fixed cells were permeabilized with 0.1% Triton X-100 in PBS for 15 min, washed with PBS, and blocked in 5% bovine serum albumin in PBS for 1 h. The cells were then incubated with a pig anti-FMDV polyclonal antibody, anti-NCL rabbit monoclonal antibody, or anti-FLAG mouse monoclonal antibody for 1 h and then with an FITC-conjugated rabbit anti-pig IgG secondary antibody, TRITC-conjugated goat anti-rabbit IgG secondary antibody, or FITC-conjugated goat anti-mouse IgG secondary antibody, respectively, for 1 h. The nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI; Beyotime) for 10 min and visualized. Images were captured with a laser-scanning confocal microscope (Leica SP8, Solms, Germany).

Nucleus-cytosol fractionation assay.

Cells infected with FMDV (MOI of 1) were harvested at different times postinfection and subjected to nuclear and cytosolic fractionation with a nuclear/cytosol fractionation kit (BioVision, CA, USA), according to the manufacturer’s instructions. The cytoplasmic and nuclear fractions were then subjected to Western blotting with the indicated antibodies. α-Tubulin and lamin B1 were used as the cytoplasmic and nuclear protein markers, respectively.

Dual-luciferase reporter assay.

The bicistronic reporter plasmids used in this study have been described in a previous study (61). Cells in which NCL was depleted or overexpressed were transfected with the indicated bicistronic constructs. At 24 h posttransfection, the cell extracts were harvested in passive buffer and examined for RLuc and FLuc activities in a Lumat LB9507 bioluminometer with the dual-luciferase reporter assay (Promega). For the bicistronic mRNA reporter assays, capped mRNAs were synthesized by incorporating the Ribo m7G cap analog (Promega) into the transcript during the RiboMAX transcription reaction and then used to transfect pretreated cells. Immediately after transfection, the cells were infected with FMDV (MOI of 1) or mock infected, and the luciferase activities were measured with the dual-luciferase reporter assay system at 6 h postinfection.

RNA-protein immunoprecipitation and RT-PCR.

Lysates from FMDV-infected (MOI of 1) cells were collected at 5 h postinfection and preincubated with protein A-agarose (GE Healthcare) on ice for 1 h. Nonspecific complexes were pelleted by centrifugation at 1,000 × g at 4°C for 10 min. An equal quantity of supernatant was mixed with rabbit anti-NCL antibody, control rabbit IgG antibody, or buffer containing no antibody and incubated on ice for 2 h. The prewashed protein A-agarose was then added to each sample, and the samples were incubated on ice for 1 h. The immunoprecipitated RNA-protein complexes were pelleted by centrifugation at 1,000 × g at 4°C for 5 min and washed three times with lysis buffer. RNA was extracted from the complexes with TRIzol reagent (Invitrogen) and then amplified by RT-PCR. RT-PCR was performed with a PrimeScript one-step RT-PCR kit (TaKaRa, Dalian, China) and primers specific for the FMDV IRES (5′-CACAGGTTCCCACAACCGACAC-3′ and 5′-GCAGTGATAGTTAAGGAAAGGC-3′) and primers specific for FMDV 3D and 3′ UTR (5′-GTTGCTAGTGATTATGACTTGGAC-3′ and 5′-CTTACGGCGTCGCTCGCCTCAGAG-3′), for porcine RPS16 mRNA (5′-CTGCAGCCATGCCTTCCAAGGGT-3′ and 5′-TCATCACGATGGGCTTATCGGT-3′), for porcine GAPDH mRNA (5′-GTCCATGCCATCACTGCCACCCAG-3′ and 5′-GCTGTTGAAGTCACAGGACACAAC-3′), for hamster RPS16 mRNA (5′-TCGCAGCCATGCCGTCCAAGGGT-3′ and 5′-TCATTAAGATGGGCTCATCGGT-3′), or for hamster GAPDH mRNA (5′-GTCCATGCCATCACGGCCACCCAG-3′ and 5′-ACTCTTGAAGTCGCAGGAGACAAC-3′). The PCR products were resolved in 1% agarose gel prestained with GelRed nucleic acid gel stain (Biotium, CA, USA).

In vitro transcription and synthesis of biotinylated RNA.

Viral cDNAs corresponding to the 5′ UTR (nucleotides 1 to 1112), the S fragment (1 to 370), the cis-acting replication elements (cre) (371 to 653), the IRES (654 to 1112) and its various deletion constructs, and the 3′ UTR (8112 to 8237) of the FMDV genome were amplified from the cDNA of FMDV strain O/BY/CHA/2010 (GenBank accession no. JN998085.1) and inserted into the pcDNA3.1 vector (Invitrogen). Before in vitro transcription, these cDNAs were linearized with BamHI, and the RNA transcripts were synthesized with RiboMAX large-scale RNA production systems (Promega). Biotinylated RNAs were synthesized with a Pierce RNA 3′ end desthiobiotinylation kit (Pierce Biotechnology, Rockford, USA), according to the manufacturer’s instructions. Briefly, T4 RNA ligase was used to attach a single biotinylated nucleotide to the 3′ terminus of the RNA transcript.

Biotinylated RNA pulldown assay.

The biotinylated RNA pulldown assay was performed essentially as previously described (61). Briefly, the biotinylated RNAs (50 pmol) were incubated with prewashed streptavidin-conjugated magnetic beads (50 μl) in an RNA capture buffer for 30 min at room temperature with agitation, to allow binding. A reaction mixture containing 200 μg of cell extract or a recombinant purified protein, 30 μl of 50% glycerol, RNA-bound beads, and protein-RNA binding buffer was prepared. The mixture (with a final volume of 100 μl) was incubated for 60 min at 4°C with agitation, and then the protein-RNA complexes were washed three times with wash buffer. After the last wash, 50 μl of elution buffer was added to the beads and the mixture was incubated for 15 min at 37°C with agitation to dissociate the complexes from the beads. Then, 6× SDS-PAGE sample buffer (10 μl) was added to the eluted samples. Finally, the samples were boiled and subjected to Western blotting.

Polysome profile analysis.

Mock- or FMDV-infected (MOI of 1) PK-15 cells were incubated with 0.1 mg/ml CHX (Sigma-Aldrich) for 5 min at 37°C to arrest the ribosome at specific times after infection. The cells were then lysed in polysomal extraction buffer (20 mM Tris-HCl [pH 7.5], 5 mM MgCl₂, 100 mM KCl, 1% Triton X-100, 0.1 mg/ml CHX, 1× protease inhibitor cocktail [EDTA free], and 50 U/ml RNase inhibitor). The cell lysates were centrifuged at 15,000 × g for 10 min at 4°C, and the supernatants were resolved on a linear 10%-to-50% sucrose gradient (composed of 20 mM Tris-HCl [pH 7.5], 5 mM MgCl₂, and 100 mM KCl) by centrifugation at 35,000 rpm at 4°C for 3 h in a Beckman SW41 Ti rotor. After centrifugation, the fractions were collected, and the optical density was measured at a wavelength of 254 nm (OD₂₅₄) with a density gradient fractionation system (Teledyne, USA). The total RNAs in these fractions were extracted with phenol-chloroform, precipitated with ethanol, and dissolved in RNase-free H₂O for RT-PCR analysis. The proteins in the fractions were precipitated with trichloroacetic acid (TCA) and analyzed by Western blotting.

Quantitative real-time PCR.

TRIzol (Invitrogen) was added to siRNA-transfected, virus-infected cells at the indicated times postinfection, followed by RNA extraction. cDNA was generated from 1 μg of total RNA using either oligo(dT)18 or an FMDV-specific forward primer listed below and Moloney murine leukemia virus (MMLV) reverse transcriptase (Promega), according to the manufacturer's instructions. The cDNAs were then subjected to qPCR with an Applied Biosystems 7500 real-time PCR system. The reaction mixtures contained 1× TB green premix Ex Taq II (TaKaRa) and 0.2 μM (each) forward and reverse primers. The primer sequences are as follows: FMDV 3D F, 5′-CAAACCTGTGATGGCTTCGA-3′, and R, 5′-CCGGTACTCGTCAGGTCCA-3′; mouse NCL F, 5′-TCCTGCTGCTCCTGCCTCAG-3′, and R, 5′-ATCATCCTCCTCCTCAGCCACAC-3′; SVV F, 5′-CACCGACAACGCCGAGAC-3′, and R, 5′-GAGATCGGTCAAACAGGAATTTGAC-3′; CSFV F, 5′-AGCCCACCTCGAGATGCTA-3′, and R, 5′-CTATCAGGTCGTACTCCCATCAC-3′; VSV F, 5′-AGGGAACTGTGGGATGACTG-3′, and R, 5′-GAACACCTGAGCCTTTGAGC-3′; porcine GAPDH F, 5′-ACATGGCCTCCAAGGAGTAAGA-3′, and R, 5′-GATCGAGTTGGGGCTGTGACT-3′; mouse GAPDH F, 5′-AAGAAGGTGGTGAAGCAGGCATC-3′, and R, 5′-CGGCATCGAAGGTGGAAGAGTG-3′; Chlorocebus sabaeus GAPDH F, 5′-TGGGCTACACTGAGCACCAG-3′, and R, 5′-AAGTGGTCGTTGAGGGCAAT-3′; GAPDH was used as a reference gene for internal standardization.

In vitro translation assay.

The in vitro translation assay was performed as previously described (63). Briefly, cytoplasmic extracts of FMDV-infected BHK-21 cells (MOI of 1) were generated at 4 h postinfection using hypotonic lysis buffer {10 mM HEPES-KOH [pH 7.6], 10 mM potassium acetate [KOAc], 0.5 mM magnesium acetate [Mg(OAc)₂], 2 mM dithiothreitol [DTT], and 1× protease inhibitor cocktail [EDTA free]}. The in vitro translation assays were performed in a final volume of 25 μl, containing 250 ng of bicistronic mRNA siCHECK-FMDV, 100 μg of cytoplasmic extract, 10 mM creatine phosphate, 50 μg/ml creatine phosphokinase, 79 mM KOAc, 0.5 mM Mg(OAc)₂, 2 mM DTT, 0.02 mM hemin, 0.5 mM spermidine, 20 mM HEPES-KOH (pH 7.6), 20 μM amino acid mixture (Promega), 0.4 mM ATP, and RNase inhibitor. Translation was allowed to proceed for 90 min at 30°C, and the luciferase activities were measured with the luciferase assay system (Promega). For the blocking experiments, the cytoplasmic extracts were preincubated overnight with the indicated concentrations of control IgG, anti-NCL antibody, or anti-PTB antibody at 4°C with rotation. The translation mixture was then added.

Production of adenovirus expressing short hairpin NCL (pAdM-shNCL).

Four shRNAs targeting the mouse NCL mRNA and a nontargeting control shRNA (shCtrl) were designed and synthesized by Vigene Biosciences (Jinan, China). Four sequences targeting NCL mRNA, i.e., 5′-GCCTTTCCTACAGTGCAACAAACTCGAGTTTGTTGCACTGTAGGAAAGGTTTTT-3′, 5′-GCAAGAACACTTCTAGCCAAACTCGAGTTTGGCTAGAAGTGTTCTTGCTTTTT-3′, 5′-GCAAGGAAAGAAGACGAAGTTTCTCGAGAAACTTCGTCTTCTTTCCTTGTTTTT-3′, 5′-GAACGGTAAGAATGCCAAGAACTCGAGTTCTTGGCATTCTTACCGTTCTTTTT-3′, were cloned into the vector pAdM-shRNA-GFP, and adenoviruses carrying the recombinant vector were generated as previously described (64). The final viral yield was 4.0 × 10¹⁰ PFU/ml.

Viral challenge in suckling mice.

Groups of 2-day-old BALB/c mice (suckling mice; n = 14 per group) were purchased from the Laboratory Animal Center of Lanzhou Veterinary Research Institute, Chinese Academy of Agricultural Sciences (Lanzhou, China). The 50% lethal dose (LD₅₀) of FMDV strain O/BY/CHA/2010 for suckling mice was estimated by the Reed and Muench method. The suckling mice in several groups were subcutaneously inoculated in the neck with 1.0 × 10⁸ PFU of purified pAdM-shCtrl or pAdM-shNCL in 100 μl of PBS. At 72 h after inoculation, the mice were challenged with a subcutaneous injection of 20 LD₅₀ of O/BY/CHA/2010 in 200 μl of PBS at the inoculation sites. The survival rate of the suckling mice was monitored for 14 days after challenge.

Histopathological and IHC staining.

Mouse tissues (hearts, lungs, livers, and kidneys) were collected and fixed in 4% paraformaldehyde for at least 2 days. The tissues were then dissected and embedded in paraffin. For the histopathological analysis, the tissues were sectioned and then stained with H&E. For IHC examination, the tissue sections were dewaxed, dehydrated, and microwaved in citrate buffer. The sections were blocked with 5% new bovine serum (NBS) for 1 h, incubated with a mouse anti-FMDV 3B monoclonal antibody at 4°C overnight, and then incubated with an HRP-conjugated anti-mouse IgG antibody for 30 min at 37°C. 3,3′-Diaminobenzidine tetrahydrochloride (DAB) was added for color development, and the sections were finally counterstained with hematoxylin. These sections were observed under an Olympus BH-2 microscope (Tokyo, Japan).

Statistical analysis.

All data were collected from at least three independent experiments and are presented as the means ± standard deviations (SD) of the results of triplicate experiments. Statistical significance was determined with Student's t test in SPSS Statistics (IBM Corporation, USA) and assessed based on the P values: **, P < 0.01.