Feline Calicivirus Proteinase-Polymerase Protein Degrades mRNAs To Inhibit Host Gene Expression

ABSTRACT

To replicate efficiently and evade the antiviral immune response of the host, some viruses degrade host mRNA to induce host gene shutoff via encoding shutoff factors. In this study, we found that feline calicivirus (FCV) infection promotes the degradation of endogenous and exogenous mRNAs and induces host gene shutoff, which results in global inhibition of host protein synthesis. Screening assays revealed that proteinase-polymerase (PP) is a most effective factor in reducing mRNA expression. Moreover, PP from differently virulent strains of FCV could induce mRNA degradation. Further, we found that the key sites of the PP protein required for its proteinase activity are also essential for its shutoff activity but also required for viral replication. The mechanism analysis showed that PP mainly targets Pol II-transcribed RNA in a ribosome-, 5′ cap-, and 3′ poly(A) tail-independent manner. Moreover, purified glutathione S-transferase (GST)-PP fusion protein exhibits RNase activity in vitro in assays using green fluorescent protein (GFP) RNA transcribed in vitro as a substrate in the absence of other viral or cellular proteins. Finally, PP-induced shutoff requires host Xrn1 to complete further RNA degradation. This study provides a newly discovered strategy in which FCV PP protein induces host gene shutoff by promoting the degradation of host mRNAs.

IMPORTANCE Virus infection-induced shutoff is the result of targeted or global manipulation of cellular gene expression and leads to efficient viral replication and immune evasion. FCV is a highly contagious pathogen that persistently infects cats. It is unknown how FCV blocks the host immune response and persistently exists in cats. In this study, we found that FCV infection promotes the degradation of host mRNAs and induces host gene shutoff via a common strategy. Further, PP protein for different FCV strains is a key factor that enhances mRNA degradation. An in vitro assay showed that the GST-PP fusion protein possesses RNase activity in the absence of other viral or cellular proteins. This study demonstrates that FCV induces host gene shutoff by promoting the degradation of host mRNAs, thereby introducing a potential mechanism by which FCV infection inhibits the immune response.

INTRODUCTION

Viruses in the family Caliciviridae have a linear, positive-sense, single-stranded RNA genome of 7.7 kb. They are classified into 11 recognized genera, and seven of them (Lagovirus, Norovirus, Nebovirus, Recovirus, Sapovirus, Valovirus, and Vesivirus) infect mammals. Noroviruses often cause human viral gastrointestinal disease worldwide. Most caliciviruses are resistant to in vitro cultivation, which severely restricts the efficient development of vaccines and viral pathogenesis research. Due to their culturability, murine norovirus (MNV) (genus Norovirus) and feline calicivirus (FCV) (genus Vesivirus) both have been widely used as models to analyze the characteristics of the calicivirus life cycle (1).

After entry into cells, FCV replicates quickly using unknown mechanisms to combat the immune response to infection (2). Our previous study reported that some FCV isolates cannot activate the interferon response in vitro or that the response was initiated only at the late phase of infection when activated (3). Yumiketa et al. reported that FCV p39 suppresses type 1 IFN production by preventing IRF-3 activation, but the precise mechanism for this action was not investigated (4).

Many viruses have evolved diverse strategies to inhibit host antiviral response. Among these, global suppression of host gene expression, termed host shutoff, is an important feature shared by various sets of viruses (5). Virus infection-induced shutoff is the result of manipulating targeted or global cellular gene expression and leads to efficient viral replication and immune evasion (6). Three different viral subfamilies, alphaherpesviruses, gammaherpesviruses, and betacoronaviruses, induce global host mRNA degradation to abrogate host gene expression (5). Herpes simplex virus 1 (HSV-1) uses the virion host shutoff (vhs) protein to function (7, 8). Kaposi’s sarcoma-associated herpesvirus (KSHV)-encoded SOX protein induces host gene shutoff (9). Epstein-Barr virus (EBV) uses BGLF5 (10) while murine herpesvirus 68 (MHV68) utilizes muSOX to induce host gene shutoff (11). The betacoronavirus severe acute respiratory syndrome coronavirus (SARS-CoV) induces mRNA degradation through nsp1 (12, 13). Moreover, these viruses use a common strategy to promote host RNA degradation (5). Virus-encoded host shutoff factors bind to mRNAs either through the ribosome (SARS-CoV nsp1) or the cap-binding complex (HSV-1 vhs) or through yet-unknown factors (KSHV SOX, MHV68 muSOX, and EBV BGLF5). Some shutoff factors are endoribonucleases (14, 15) and perform endonucleolytic mRNA cleavage near the cap (vhs) or at one or more internal positions (SOX, muSOX, and BGLF5). After primary cleavage, host Xrn1 is recruited to complete the degradation of the cleaved intermediates.

In this study, we found that FCV infection promotes degradation of host mRNAs and induces host gene shutoff. Further, proteinase-polymerase (PP) expression enhances mRNA degradation via a common strategy, and the key sites of the PP protein required for its proteinase activity are also essential for its shutoff activity. An in vitro assay showed that the glutathione S-transferase (GST)–PP fusion protein could act as an RNase in the absence of other viral or cellular proteins. This study explains a newly discovered strategy by which FCV induces host gene shutoff by promoting degradation of host mRNAs.

RESULTS

FCV infection contributes to the shutoff of host genes.

Since many viral infections cause host gene shutoff, we sought to determine whether infection with FCV resulted in inhibited host gene expression. Crandall-Rees feline kidney (CRFK) cells were infected with FCV strain 2280 at a multiplicity of infection (MOI) of 1 for 6 h, 8 h, or 10 h and then pulse-labeled for 20 min with [³⁵S]methionine ([³⁵S]Met). Cell extracts were subjected to SDS-PAGE, and the global efficiency of the cellular protein synthesis was monitored. Colloid Coomassie blue staining of the gel confirmed that the quantities of the host proteins in these samples were similar (Fig. 1A, left). From 6 h to 10 h postinfection, there was a clear reduction in overall cellular protein synthesis (Fig. 1A, right), which was indicative of host shutoff. After infection, viral proteins accumulate, and only a few host proteins that benefit viral replication are continually produced (11). At the early stage of FCV infection, host protein production is reduced but the synthesis of viral proteins, such as VP1, is increased (Fig. 1A).

FIG 1.

FCV infection triggers host shutoff. (A) CRFK cells were either mock infected or infected with FCV at an MOI of 1. At the indicated time points, the cells were pulse-labeled with [³⁵S]Met, and the lysates were analyzed by SDS-PAGE. The gel was then dried and visualized by autoradiography. The expression of VP1 was evaluated by Western blotting using anti-VP1 antibody. (B) The CRFK cells were transfected with plasmid pEGFP-N1 for 12 h and then infected with FCV at an MOI of 1. At the indicated times, RNA was extracted and subjected to Northern blotting with GAPDH and GFP probes. The expression of VP1was evaluated by Western blotting using anti-VP1 antibody. (C and D) CRFK cells were transfected with plasmid pEGFP-N1 for 24 h (0 h) and then treated with ActD (2 μg/ml) for 8 h. RNAs were extracted, and the amounts of GFP RNA (C) and GAPDH mRNA (D) were determined by qRT-PCR analysis. (E) CRFK cells were transfected with plasmid pEGFP-N1 for 24 h and then infected with FCV at an MOI of 1. At 1 h postinfection, RNAs were extracted (0 h) or ActD (2 μg/ml) was added to the culture. RNAs were extracted 10 h after ActD addition (10 h). The loading control was 18S rRNA. The expression of VP1 was evaluated by Western blotting using anti-VP1 antibody. (F) CRFK cells were infected with FCV at an MOI of 1. After 1 h, the cells were treated with ActD (2 μg/ml) or medium only. At 12 h postinfection, cellular supernatant was harvested and viral titers were determined as 50% tissue culture infective doses (TCID₅₀). Blots in panel B and E were repeated three times, and the intensity relative to 18S RNA is shown as means and standard deviations (SD). Data in panels C, D, and F are means and SD from three independent experiments. Differences (*, P < 0.05; **, P < 0.01; ns, not significant) between the experimental and control groups are noted.

Many viruses, such as KSHV and EBV, promote host shutoff by degrading host mRNA (10, 16). CRFK cells were inoculated with FCV strain 2280 at an MOI of 0.1 for 12 h, and then the cells were collected for transcriptome sequencing. A total of 1,335 downregulated genes were identified upon FCV infection (17). Among these downregulated genes, multiple genes from one family, such as the DDX gene family, including DDX3X, DDX5, DDX6, DDX17, DDX18, DDX20, DDX23, DDX24, DDX27, DDX31, DDX39A, DDX42, DDX47, DDX50, DDX51, and DDX52 (see Table S1 in the supplemental material), are sensitive to FCV infection-induced mRNA decay.

To further explore the mechanism of shutoff induced by FCV infection, we evaluated the mRNA levels of the host endogenous gene encoding glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and the exogenous reporter gene encoding green fluorescent protein (GFP) during FCV infection by Northern blotting assay (Fig. 1B). At 8 h postinfection, the levels of GFP and GAPDH mRNAs were clearly reduced in the FCV-infected cells compared to the level in the mock-infected samples (Fig. 1B). The data indicate that FCV infection inhibits host protein synthesis and host mRNA accumulation.

To determine whether the reduction of expressed GFP RNA and endogenous GAPDH mRNAs in FCV-infected cells was attributed to degradation of these mRNAs, we first demonstrated that actinomycin D (ActD) treatment blocked the synthesis of expressed GFP RNA (Fig. 1C) and endogenous GAPDH mRNAs (Fig. 1D) in cells. Next, cells were transfected with pEGFP-N1. At 24 h after transfection, cells were mock infected or infected with FCV at an MOI of 1. After virus adsorption for 1 h, intracellular RNAs were extracted from half of the samples (Fig. 1E, 0 h). The remaining cells were incubated in the presence of ActD, and cellular RNAs were extracted at 10 h after ActD addition. FCV replication was not affected by ActD treatment (Fig. 1F). Northern blot analysis showed that the amounts of GFP RNA and GAPDH mRNAs in the 10 h samples were lower in FCV-infected cells than in mock-infected cells (Fig. 1E), demonstrating that FCV infection indeed promoted degradation of both mRNAs.

These data suggested that FCV infection promotes the decay of host mRNAs and induces host gene shutoff.

FCV PP expression contributes to host mRNA degradation.

To determine whether viral protein could promote host mRNA decay, CRFK cells were cotransfected with an empty vector or a plasmid expressing each protein of FCV strain 2280 and a GFP reporter plasmid, and then the mRNA level of GFP was measured by Northern blotting. As predicted, the results indicated a significant PP-induced reduction in reporter mRNA levels (Fig. 2A). To verify that the PP-induced mRNA decrease was not specific to the exogenously expressed GFP reporter, we also measured the mRNA level of endogenous GAPDH in the presence of PP. Indeed, GAPDH mRNA was also reduced (Fig. 2A). Moreover, the effect depended on the concentration of transfected plasmid encoding the PP protein (Fig. 2B). Further, PP-mediated reduction of GFP and GAPDH mRNA was also tested on other cells, including F81 and 293T cells (Fig. 2C). To examine whether the PP proteins of other FCV strains also display the shutoff activity, we investigated the ability of the PP protein of FCV strain F9 to induce mRNA decrease. The expression of strain F9 PP also contributed to the reduction of GFP and GAPDH mRNA in CRFK cells in a manner similar to that of 2280 PP (Fig. 2D).

FIG 2.

FCV PP protein degrades mRNA. (A) Plasmids encoding FCV p5.6, p30, p32, p39, Vpg, PP, VP1, or VP2 or empty vector (Vec) as well as plasmid pEGFP-N1 (0.5 μg/ml) at a ratio of 3:1 were cotransfected into 293T cells for 24 h. The mRNA expression levels of GFP were determined by Northern blotting (NB), and 18S rRNA was used as a loading control. The expression of plasmids encoding each FCV protein was evaluated by Western blotting (WB) using anti-Flag antibody. (B) CRFK cells were cotransfected with pEGFP-N1 (0.5 μg/ml) and different doses of p3×Flag-PP for 24 h. Then, the RNA was harvested and subjected to Northern blotting with GAPDH and GFP probes. The loading control was 18S rRNA. (C) F81 or 293T cells were cotransfected with pEGFP-N1 (0.5 μg/ml) and p3×Flag-PP (1.5 μg/ml) for 24 h. Then, the RNA was harvested and subjected to Northern blotting with GAPDH and GFP probes. The loading control was 18S rRNA. (D) 293T cells were cotransfected with pEGFP-N1 (0.5 μg/ml) and a plasmid encoding 2280 PP or F9 PP (0.5 to 1.5 μg/ml) for 24 h. Then, the RNA was harvested and subjected to Northern blotting with GAPDH and GFP probes. The loading control was 18S rRNA. (E and F) CRFK cells were cotransfected as described above, and the GFP (E) and GAPDH (F) mRNA half-life was determined by qPCR at different time points after addition of ActD (2 μg/ml). (G) CRFK cells were transfected with an empty vector (Vec) or with pEGFP-N1 alone or with p3×Flag-PP at a ratio of 1:3. The samples were treated with or without nuclease prior to cell lysis to remove extracellular nucleic acids and were then divided in half, and either total cellular DNA or RNA was harvested. GFP DNA and mRNA levels were then calculated via qPCR. Blots in panels A, B, C, and D were repeated three times, and the intensity relative to 18S RNA is shown as means and SD. Data in panels E, F, and G are from three independent experiments, and data are means and SD. Differences (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, not significant) between the experimental and control groups are noted.

Next, an RNA half-life assay was performed in the presence of ActD to confirm that the decreased mRNA levels in PP-expressing cells were a result of increased mRNA degradation. As shown in Fig. 2E, the half-life of the GFP mRNA decreased from 45 h to 7 h or 6.4 h in the presence of 2280 PP or F9 PP. Moreover, the half-life of the GAPDH mRNA decreased from 78 h to 14 h or 13.8 h in the presence of 2280 PP or F9 PP (Fig. 2F). To exclude the possibility that PP expression degraded the plasmid DNA, which would lead to the decreased transcription of GFP gene, we simultaneously measured the levels of the GFP reporter plasmid DNA and mRNA in the presence or absence of PP by quantitative PCR (qPCR) (Fig. 2G). The cells were treated with the nuclease just prior to lysis to remove any extracellular nucleic acid. Importantly, PP expression resulted in a significant decrease in GFP mRNA levels but not in DNA levels (Fig. 2G).

We conclude that FCV PP expression enhances host mRNA degradation.

FCV PP amino acids 39, 60, 122, and 137 are key sites required for PP shutoff activity and viral replication.

The N terminus of the PP protein is required for its proteinase activity, and the amino acid residues H^1110(39), E^1131(60), C^1193(122), and H^1208(137) are important for the FCV F4 3C-like protease activity (18). To determine whether proteinase and shutoff activities share common sites, we constructed four 2280 PP and F9 PP mutants with a single amino acid mutation according to the proteinase activity sites, and each mutant together with the EGFP reporter plasmid was cotransfected into 293T cells. The fluorescence signal was observed (Fig. 3A and D) 24 h posttransfection. PP expression decreased the magnitude of the fluorescence, but a single amino acid mutation of 2280 PP or F9 PP located at H^1110(39), E^1131(60), C^1193(122), or H^1208(137) recovered the fluorescence.

FIG 3.

Identification of the key sites of FCV PP that are required for its shutoff activity. (A and B) Plasmids encoding FCV 2280 PP protein as well as its mutants or an empty vector (Vec) were cotransfected into the 293T cells with the GFP reporter plasmid for 24 h. The green fluorescence signal from GFP expression was detected (A), and the mRNA expression levels of the GFP were determined by Northern blotting (NB) (B). The expression of plasmids was evaluated by Western blotting (WB) using an anti-Flag antibody. (C) CRFK cells were transfected with plasmids encoding FCV 2280 PP protein or PP mutants for 24 h and then subjected to immunofluorescence assay (IFA) with anti-Flag antibodies. (D and E) Plasmids encoding the FCV F9 PP protein as well as its mutants or an empty vector (Vec) were cotransfected into the 293T cells with the GFP reporter plasmid for 24 h. The green fluorescence signal from GFP expression was imaged (D), and the mRNA expression levels of the GFP were determined by Northern blotting (E). The expression of plasmids was evaluated by Western blotting using an anti-Flag antibody. (F) CRFK cells were transfected with plasmids encoding the FCV F9 PP protein or PP mutants for 24 h and then subjected to IFA with anti-Flag antibodies. (G) A total of 26 amino acid sequences from FCV PP protein ranging from aa 1 to 137 were retrieved from GenBank and aligned using MEGA5.10. Blots in panels B and E were repeated three times, and the intensity relative to 18S RNA is shown as means and SD. Differences (*, P < 0.05; **, P < 0.01) between the experimental and control groups are noted.

Next, enhanced GFP (EGFP) and GAPDH mRNAs were measured 24 h posttransfection using the Northern blotting method. As shown in Fig. 3B and E, all four kinds of mutations attenuated the PP-induced shutoff activity. Among these, the shutoff activities of the PP^H39A, PP^C122A, and PP^H137A mutants were fully suppressed and did not reduce GFP mRNA expression. The PP^E60A mutant degraded only a portion of the GFP mRNA such that the shutoff activity of the mutant was only attenuated (Fig. 3B and E).

Cytoplasmic localization plays an important role in host shutoff (11). To exclude the possibility that the loss or attenuation of the shutoff activity caused by the PPs with single amino acid mutations was attributable to a change in their cellular locations, we analyzed the cytoplasmic localization of the mutants. As shown in Fig. 3C and F, both wild-type PP and the PP mutants had the same locations.

These results revealed that the amino acid residues H^1110(39), E^1131(60), C^1193(122), and H^1208(137) are important for PP-induced shutoff activity. Moreover, the four sites are highly conserved in FCVs (Fig. 3G). To explore the role of the four amino acids in FCV replication, we constructed four infectious clones, including rFCV-2280 H39A, E60A, C122A, and H137A, but the four recombinant viruses could not be recovered, suggesting that the four amino acids in FCV replication are necessary.

FCV PP selectively targets mRNA.

Next, we wanted to determine which type of cytoplasmic RNA is the target for the PP. GFP RNA reporters were transcribed by RNA polymerase (Pol) I, II, and III in the presence of different doses of PP, and then the RNA levels of the GFP were measured by Northern blotting. Neither 2280 nor F9 PP expression resulted in the degradation of the Pol I- or Pol III-driven transcripts (Fig. 4A and B). In contrast, the Pol II-driven GFP was reduced by PP in a dose-independent manner. These data revealed that Pol II-driven mRNA is targeted in cells by PP.

FIG 4.

FCV PP protein targets Pol II-transcribed RNAs. (A and B) The 293T cells were cotransfected with 250 ng/ml Pol II-driven GFP together with either 1 μg/ml Pol I-GFP (A) or Pol III-GFP (B) constructs as well as Vec or increasing amounts of plasmid encoding FCV 2280 PP or F9 PP (0.25 to 1 μg/ml). Reporter RNA levels were determined by Northern blotting with probes against GFP. The loading control was 18S rRNA. (C and D) The 293T cells were cotransfected with 250 ng/ml Pol II-driven GFP together with either 1 μg/ml Pol I-GFP (C) or Pol III-GFP (D) constructs as well as Vec or plasmid encoding PP or PP H39A (1 μg/ml). Reporter RNA levels were determined by Northern blotting with probes against GFP. The loading control was 18S rRNA. (E) Analysis of the location for the Pol I/II/III-driven GFP RNAs. The 293T cells were transfected with 1 μg/ml Pol I/II/III-driven GFP constructs for 24 h, the cytoplasmic and nuclear RNAs were extracted, and the ratio of expression was determined by RT-qPCR. GAPDH mRNA and U6 RNA were respectively used as cytoplasmic and nuclear control. Data are percentage of input (means; n = 3). Blots in panels A, B, C, and D were repeated three times, and the intensity relative to 18S RNA is shown as means and SD. Differences (**, P < 0.01) between the experimental and control groups are noted.

Remarkably, PP preferentially targets mRNA over other cytoplasmic RNA species. To further demonstrate the effect of PP on mRNAs, 293T cells were transfected with Pol I GFP (Fig. 4C) or Pol III GFP (Fig. 4D) plasmid together with wild-type PP or mutated PP H39A; then, the RNA levels of the GFP were measured by Northern blotting. Both 2280 and F9 PP preferentially targeted mRNA, and PP-mediated mRNA decay was blocked by the H39A mutation. However, neither wild-type PP nor mutated PP H39A directed decay of the Pol I- or Pol III-driven GFP RNA.

Nuclear localization for the Pol I/III-driven GFP RNAs may partly explain why both RNAs are not susceptible to degradation by PP protein. To exclude this possibility, the ratio of expression for the Pol I/II/III-driven GFP RNAs in the cytoplasm and nucleus was determined. The RNAs in the cytoplasm and nucleus were separated, and the levels were determined by qPCR. As shown in Fig. 4E, the three types of RNAs are located in both the cytoplasm and nucleus. Therefore, the fact that Pol I/III-driven GFP RNAs are resistant to degradation by PP protein is not due to their location in the nucleus.

These data revealed that nontranslatable RNAs are not targets for PP protein within cells.

PP-mediated mRNA decay does not need ribosomes.

In reticulocyte lysates, SARS-CoV nsp1 has been shown to bind directly to the 40S ribosomal subunit to mediate mRNA cleavage, whereas ribosomes are not required for vhs, SOX, or BGLF5 activity (5). Since PP preferentially targets mRNA over other cytoplasmic RNA species, we investigated whether ribosomes recruit PP to substrate mRNA. To determine whether ribosomes are involved in PP-mediated RNA degradation, we used a plasmid in which a strong hairpin (hp7) was integrated into the 5′ nontranslated region (UTR) (3 nucleotides downstream of the predicted transcription start site) of the pd2-eGFP-N1 reporter mRNA (hp-GFP) (5). The hairpin completely inhibited the translation of the hp-GFP mRNA through examining the expression of GFP by Western blot (WB) assay (Fig. 5A). Next, 293T cells were transfected with hp-GFP plasmid together with different doses of PP, and then the RNA levels of the GFP were measured by Northern blotting. As shown in Fig. 5B, both 2280 and F9 PP expression reduced the levels of hp-GFP mRNA in a dose-dependent manner. Moreover, PP-mediated degradation of hp-GFP mRNA could be blocked by the H39A mutation (Fig. 5C). Therefore, ribosomes are not required for selective PP targeting of mRNA.

FIG 5.

The 40S ribosomal subunit is nor required for PP-mediated mRNA degradation. (A) Strategy for constructing hp-GFP. Different doses of plasmid pEGFPN1 or hp-GFP were transfected into 293T cells for 24 h, the green fluorescence from GFP expression was imaged, and the protein expression of GFP was determined by WB. (B) The 293T cells were cotransfected with 250 ng/ml plasmid encoding the hp-GFP bearing a strong hairpin and Vec or increasing amounts of plasmid encoding FCV 2280 PP or F9 PP (0.25 to 1 μg/ml). Reporter RNA levels were determined by Northern blotting with probes against GFP. The loading control was 18S rRNA. (C) The 293T cells were cotransfected with 250 ng/ml plasmid encoding the hp-GFP bearing a strong hairpin and Vec or plasmid encoding PP or PP H39A (1 μg/ml). Reporter RNA levels were determined by Northern blotting with probes against GFP. The loading control was 18S rRNA. Blots in panels B and C were repeated three times, and the intensity relative to 18S RNA is shown as means and SD. Differences (**, P < 0.01) between the experimental and control groups are noted.

PP-mediated mRNA decay does not need a 5′ cap and a 3′ poly(A) tail in the RNA substrate.

Next, a 5′ cap or a 3′ poly(A) tail is exclusive in the mRNA and may direct PP-mediated mRNA decay. To test this possibility, we first purified the GST-PP fusion protein as well as one mutant, PP^H39A (Fig. 6A and C). GFP RNA without a 5′ cap and a 3′ poly(A) tail was obtained using a T7 in vitro transcription system. Equal amounts of the reporter RNA were incubated at 30°C with 2280 PP protein or with GST alone. The samples were harvested at the indicated time points and analyzed in agarose-formaldehyde gels. The RNA substrate mixed with GST protein alone remained stable through the incubation time (Fig. 6B), but the RNA incubated with the 2280 PP protein was cleaved into smaller fragments (Fig. 6B). To further verify this result in vitro, a GST-2280 PP mutant protein (GST-2280 PP H39A) was purified. As shown in Fig. 6B, no cleaved fragment was detected when the RNA substrate was incubated with GST-PP H39A. A similar result was also obtained with the GST-F9 PP and its mutant protein (Fig. 6D). The reporter RNA was degraded into smaller fragments by F9 PP with an increase in incubation time but was stable when incubated with F9 PP H39A (Fig. 6D). Both results indicated that the absence of a 5′ cap and a 3′ poly(A) tail does not influence the cleavage activity of PP and also indicated that the PP protein could act as an RNase in the absence of other viral or cellular proteins.

FIG 6.

PP-induced RNA degradation is not influenced by a 5′ cap and 3′ poly(A) tail. (A and C) Proteins are from E. coli expressing GST-PP, GST-PP H39A, and GST tags. After recombinant proteins bind to glutathione-Sepharose, they were eluted with 25 mM glutathione. A Coomassie blue-stained gel of recombinant proteins is shown. (B and D) GST (10 μg), GST-PP fusion proteins (10 μg), and PP H39A fusion protein (10 μg) were incubated at 30°C with GFP RNA (4 μg) transcribed in vitro. At the indicated times, the RNA was purified and resolved on an agarose-formaldehyde gel. (E and F) The GST-PP and PP H39A fusion proteins (10 μg) were incubated at 30°C with DDX50 or TRAF1 RNA (4 μg) transcribed in vitro. At the indicated times, the RNA was purified and resolved on an agarose-formaldehyde gel. (G) GST (10 μg) and GST-PP fusion protein (10 μg) and RNase A (10 μg) were incubated with FCV genomic RNA or GFP RNA (4 μg) transcribed in vitro at 30°C for 90 min, respectively. The RNA was purified and resolved on an agarose-formaldehyde gel. (H) GST-2280 PP (5 μg) with different doses of RRI or PMSF was incubated at 37°C or 4°C, respectively, for 30 min. Following treatment, RNA (4 μg) was added and incubated at 30°C for 90 min. The RNA was purified and resolved on an agarose-formaldehyde gel. Blots in panels B, D, E, and F were repeated three times, and the intensity relative to the input is shown as means and SD. Differences (*, P < 0.05; **, P < 0.01) between the experimental and control groups are noted.

Next, we used two downregulated genes, DDX50 and TRAF1 (see Table S2), upon FCV infection to further verify that in vitro RNase activity of PP does not require a 5′ cap and a 3′ poly(A) tail. DDX50 and TRAF1 RNA lacking a 5′ cap and a 3′ poly(A) tail but containing 5′ and 3′ untranslated regions (UTR) was prepared using T7 in vitro transcription system. Equal amounts of the RNA were incubated at 30°C with 2280 PP, F9 PP, or GST alone. At 90 min postincubation, the intact DDX50 RNA (Fig. 6E) and TRAF1 RNA (Fig. 6F) were degraded by 2280 PP and F9 PP. Thus, in the absence of other viral or cellular proteins, the PP protein could act as an RNase to degrade host mRNAs in a 5′ cap- and a 3′ poly(A) tail-independent manner.

Next, we tested whether FCV genomic RNA is susceptible to degradation by purified PP protein. As shown in Fig. 6G, both 2280 and F9 RNAs were degraded by RNase A but were not degraded by PP protein; GFP RNAs were degraded by PP protein. Therefore, FCV RNA may avoid degradation by PP protein.

Since the N terminus of PP has both RNase and protease activities, we investigated the effects of RNase and protease inhibitors on PP RNase activity. As shown in Fig. 6H, compared to GST, 2280 PP was able to degrade GFP RNA, but in the presence of recombinant RNase inhibitor (RRI) and phenylmethylsulfonyl fluoride (PMSF), the degradation by PP was blocked, suggesting that both RNase and protease inhibitors could inhibit RNase activity of PP.

PP requires the host factor Xrn1 for completion of mRNA degradation.

As shown in Fig. 6B and D, the GFP RNA incubated with the PP protein was cleaved into smaller fragments, but no visible intermediates were observed within cells transfected with PP (Fig. 7A), suggesting that host factors containing exonuclease activity may be involved in further degradation. Moreover, KSHV SOX, HSV vhs, and EBV BGLF5 need host Xrn1 nuclease to complete the degradation of the endonucleolytically cleaved mRNAs (5). A flavivirus-derived Xrn1 blocking element (SLII) (19) was cloned into the 3′ UTR of the GFP mRNA (Fig. 7B). If Xrn1 is required to complete the degradation of the cleaved mRNA, the sequences following the SLII element should be protected. A probe against the GFP 3′ UTR was used during Northern blotting. A visible fragment was detected in the GFP-3′SLII and PP cotransfection group (Fig. 7B). To further confirm the result, the expression of Xrn1 was downregulated via RNA interference (RNAi) (Fig. 7C), and the extent of the GFP mRNA degradation was analyzed by Northern blotting using a probe against the GFP 3′ UTR. In agreement with the GFP-3′SLII results, knocking down Xrn1 changed the mRNA degradation profile in the cells transfected with PP, and an accumulation of degradation intermediates was detected (Fig. 7D). Both results demonstrate that PP requires host Xrn1 nuclease to complete the degradation of the mRNAs after primary cleavage.

FIG 7.

Cellular Xrn1 nuclease is required for PP-induced mRNA degradation. 293T cells were cotransfected with pEGFP-N1 (0.5 μg/ml) and plasmids encoding 2280 PP or F9 PP (0.5 to 1.5 μg/ml) for 24 h. Then, the RNA was harvested and subjected to Northern blotting with probes against the 3′ UTR of the GFP. The loading control was 18S rRNA. (B) 293T cells were cotransfected with 500 ng/ml GFP-3′SLII construct and 1,500 ng/ml p3×Flag-PP or Vec. Reporter mRNA levels were determined by Northern blotting with probes against the 3′ UTR of the GFP mRNA. The loading control was 18S rRNA. (C) 293T cells were transfected with 50 nM si-Ctrl (control) and si-Xrn1 for 36 h, and then total RNA and protein were extracted for the qRT-PCR and Western blotting assay, respectively. (D) 293T cells were transfected with 50 nM si-Ctrl (control) or si-Xrn1 for 24 h and then transfected with 2.5 μg/ml p3×Flag-PP and 2 μg/ml pEGFP-N1 for 48 h. Total RNA was extracted for Northern blot assay using probes against the 3′ UTR of the GFP. Blots in panels B and D were repeated three times, and the intensity relative to 18S RNA is shown as means and SD. Differences (**, P < 0.01) between the experimental and control groups are noted.

DISCUSSION

Blocking host gene expression in a global manner is beneficial for viral replication and immune evasion, which is an important feature of a diverse set of viruses. To maintain viral persistence in the host, viral genomes have evolved to encode shutoff viral factors. These factors target preexisting and newly produced mRNA during the first few hours after infection. FCV is a highly contagious pathogen and can persistently infect cats. These features may be attributed to shutoff viral factors. In this study, we demonstrated that FCV infection induces a significant host shutoff feature in cells and identified the PP protein as a shutoff factor. Pol II-transcribed RNAs are degraded by PP with the help of the host exonuclease Xrn1. A purified GST-PP fusion protein exhibits RNase activity in an in vitro assay. Our study explains a newly discovered strategy by which FCV manipulates host gene expression to evade the host immune response and efficiently establish infection.

Mouse norovirus (MNV) NS6 displays protease activity and could diminish host mRNA translation, which activates the cellular caspases and leads to the establishment of an apoptotic environment (20). No studies had reported that NS6 displayed RNase activity. FCV infection inhibited host protein synthesis via the cleavage of eukaryotic initiation factor eIF4G and poly(A)-binding protein (PABP) (21, 22), but no further study was performed to evaluate which viral protein was required for the result. In this study, for the first time, we determined that proteinase-polymerase (PP) exhibits RNase activity and could degrade the host mRNA, revealing a new strategy by which FCV inhibits host gene expression.

FCV PP has proteinase and polymerase activity and is necessary factor for viral replication. No report has indicated that PP is responsible for virulence. Our previous study showed that strain F9 displays lower virulence than strain 2280 (23). However, F9 PP-induced reduction of GFP and GAPDH mRNA in CRFK cells was comparable to 2280 PP-induced reduction (Fig. 2D). In the meantime, we constructed a chimeric virus, r2280-F9 PP. The backbone of r2280-F9 PP is strain 2280 and the virus expresses F9 PP. The in vitro growth kinetics showed that FCV 2280, r2280, and r2280-F9 PP shared a similar trend (data not shown).

Virus-encoded shutoff factors, including KSHV SOX, MHV68 muSOX, EBV BGLF5, HSV-1 vhs, SCoV nsp1 (5), and FCV PP, first identified in this study, target host transcripts driven by RNA polymerase II (RNAP II) (24). However, not all Pol II-transcribed RNAs are sensitive to shutoff factor-mediated RNA degradation. An examination of the relative rates of degradation of several cellular mRNAs upon HSV-1 infection indicated that at least three classes of cellular RNAs constitute the main targets, including constitutively expressed mRNAs, some inducible mRNAs containing AU-rich elements (AREs) in the 3′ UTR, and some inducible mRNAs, as exemplified by GADD45β and tristetraprolin (15). During HSV-1 lytic infection, vhs regulates the sequential expression of different classes of viral genes through its shutoff activity (15). However, some viral RNAs are not sensitive to these shutoff factors. In this study, the nascent host proteins were significantly reduced at 8 h postinfection, whereas the viral VP1 protein was produced in large quantities (Fig. 1A). During FCV infection, viral RNA genomes avoid degradation by PP protein (Fig. 6G). In a previous study, we found that the 5′ UTR sequence of IFNAR1 mRNA contributes to it sensitive to FCV p30-mediated decay (17). Hutin et al. reported that the interleukin 6 (IL-6) mRNA contains a dominant, cis-acting ∼100-nucleotide element within its 3′ UTR that renders it directly refractory to cleavage by KSHV SOX (25). Therefore, PP-resistant elements in the viral genome RNA may protect viral RNA from being cleaved by PP.

Binding of the 40S ribosomal subunit is selectively required for some shutoff factors. SCoV nsp1 could not degrade hp-GFP, which may be associated with recruitment to mRNA in a ribosome-dependent manner (5). Other viral shutoff factors were able to degrade hp-GFP (5), including FCV PP. All these host shutoff factors required Xrn1, in particular, to complete the mRNA degradation (5). The FCV PP was also dependent on the host Xrn1 to complete the mRNA degradation (Fig. 7D). Therefore, FCV PP uses a mechanism similar to those of other shutoff factors to induce host shutoff.

The N terminus of the PP protein is required for its proteinase activity, and the amino acid residues H^1110(39), E^1131(60), C^1193(122), and H^1208(137) are important for FCV F4 3C-like protease activity (18). Our results showed that the four amino acid residues are essential for PP shutoff activity. The crystal structure of the norovirus (NoV) 3C-like protease (Protein Data Bank [PDB] code 1WQS) (19) was used as the template for the molecular modeling of the FCV 3C-like proteases (18). Three-dimensional models of the FCV 3C-like proteases, ranging from amino acids (aa) 1 to 154 (18), were constructed by the homology modeling technique and revealed that the N- and C-terminal subdomains of the PP proteinase domain are separated by a large cleft. H and E are located along the inner surface of the N-terminal subdomain, whereas C and H are located along the inner surface of the C-terminal subdomain. The four amino acids are highly conserved among 3C-like FCV, Sapporo virus (SaV), rabbit hemorrhagic disease virus (RHDV), and NoV proteases and are considered conserved catalytic surfaces. It is important to determine the role of these four amino acids in viral replication and virulence. A reverse genetics system for FCV 2280 has been constructed in our lab, but we could not recover the virus from an infectious full-length cDNA clone with each of the single amino acid mutants derived from the four key amino acids (data not shown), which may be attributable to the loss of proteinase activity.

In conclusion, we have shown that FCV infection promotes the degradation of host mRNAs and induces host gene shutoff via a common strategy. Further, the PP protein of different FCV strains is a key factor that acts as a RNase and directly cleaves host mRNA. Our study explores a potential mechanism by which FCV infection blocks the host immune response.

MATERIALS AND METHODS

Cells, viruses, and reagents.

Crandell-Rees feline kidney (CRFK) cells and HEK 293T cells were maintained in Dulbecco’s modified minimal essential medium (DMEM) (Gibco) containing 10% fetal bovine serum (FBS, Gibco), 100 U/ml penicillin, and 100 μg/ml streptomycin at 37°C under 5% CO₂. FCV strains 2280 and F9 were purchased from the ATCC and propagated in CRFK cells. Actinomycin D (HY-17559) was purchased from MedChem Express (MCE, Shanghai, China).

Plasmid construction.

The plasmids Pol I-GFP, Pol II-GFP, Pol III-GFP, hp-GFP, and GFP-3′SLII (5) were gifts from Britt A. Glaunsinger. For plasmid Pol I-GFP, Pol II-GFP, and Pol III-GFP, GFP transcription was driven by Pol I, Pol II, and Pol III promoters, respectively. Plasmid hp-GFP contains a strong hairpin at the 5′ UTR, which completely blocked translation. Plasmid GFP-3′SLII contains the flaviviral SLII element at the 3′ UTR. The plasmids encoding the individual FCV proteins were constructed by cloning each gene into the p3×Flag-CMV-10 vector (Sigma-Aldrich) as previously described (26), and plasmids encoding PP mutants were constructed by site-directed mutagenesis using a kit (New England Biolabs [NEB]). pET-32a-EGFP was constructed by cloning the GFP open reading frame (ORF) into the pET-32a vector (Novagen). Feline DDX50 and TRAF1 genes containing 5′ and 3′ UTR were cloned into the pJet1.2 vector. pEGFP-N1 vector (Clontech), which encodes the GFP driven by the cytomegalovirus (CMV) promoter.

Northern blotting.

For Northern blotting, total RNA was harvested using TRIzol (Life Technologies) and analyzed by 1.2% agarose-formaldehyde gel electrophoresis. RNAs were transferred to a 0.45-μm nylon membrane and probed with biotin-labeled DNA probes (Table 1) generated using a North2South biotin random prime DNA labeling kit (Thermo Scientific). The membrane was imaged on an Odyssey CLx infrared imaging system (Li-Cor Biosciences) using IRDye 800-conjugated streptavidin (Li-Cor Biosciences). The specific method and procedure were described in our recent study (17).

TABLE 1.

Primers for production of Northern blotting probes

Primer name	Primer sequence (5′–3′)
GFP-F	AAGTTCATCTGCACCACCGGCAAG
GFP-R	ACCATGTGATCGCGCTTCTCGTTG
GFP-3′SLII-3′UTR-F	CCCCTGAACCTGAAACATAAAA
GFP-3′SLII-3′UTR-R	AAAATGCTTTATTTGTGAAATT
GFP-3′UTR-F	GCCGCGACTCTAGATCATAATC
GFP-3′UTR-R	TTCATTTTATGTTTCAGGTTCAG
GAPDH	ACGGCACAGTCAAGGCTGAGAACGGG
GAPDH	GGGCCGTCCACGGTCTTCTGGGT

Radiolabeling of cells.

CRFK cells cultured in 6-well plates were infected with FCV at an MOI of 1 at the indicated time points. Then, the cells were incubated in methionine- and cysteine-free medium (Life Technologies) for 30 min and then incubated in medium containing 20 μCi/ml of [³⁵S]Met (PerkinElmer) for 20 min. Cell extracts were analyzed after 10% SDS-PAGE.

RNAi.

The small interfering RNA (siRNA) si-Xrn1 (AGATGAGACTGAAGAT) was used. The siRNAs were produced by RiboBIO (Guangzhou, China). HEK293T cells were transfected with 10 pmol of siRNA diluted in 100 μl of Opti-MEM medium (Gibco) containing 3 μl of Lipofectamine RNAiMAX transfection reagent (Invitrogen) for 36 h. Knockdown efficiency was determined using quantitative real-time PCR (qRT-PCR) analysis.

Quantitative real-time PCR.

Total cellular RNA was prepared with an RNeasy minikit (Qiagen, Valencia, CA, USA), and cDNA was produced using Fast King RT SuperMix containing DNase (Tiangen, China) according to the manufacturer’s protocol. RNA (100 ng) was added to 8 μl of 5× FastKing RT SuperMix. The reaction mixture was incubated at 42°C for 15 min and then at 95°C for 5 min to inactive DNase. DNA was extracted using a TIANamp genomic DNA kit (Tiangen, China). qPCR was based on the SYBR green (TaKaRa) assay and performed using QuantStudio 5 (Applied Biosystems). The relative mRNA and DNA expression levels were calculated by the 2^−ΔΔCT method using 18S rRNA or GAPDH DNA as an internal control for normalization. The primer sequences used are shown in Table 2.

TABLE 2.

Primers for qRT-PCR

Primer name	Primer sequence (5′–3′)
18S rRNA-F	CGGCTACCACATCCAAGGAA
18S rRNA-R	GCTGGAATTACCGCGGCT
GFP-F	ACCACATGAAGCAGCACGACT
GFP-R	TCAGCTCGATGCGGTTCACCA
GAPDH-F	TGACCACAGTCCATGCCATC
GAPDH-F	GCCAGTGAGCTTCCCGTTCA
Xrn1-F	TAGTTGGTCGGTATGTAG
Xrn1-R	CAGAATGAAATAGGCAGT
U6-F	CTCGCTTCGGCAGCACA
U6-R	AACGCTTCACGAATTTGCGT

Sensitivity of the plasmid DNA to PP-mediated degradation.

After CRFK cells were transfected with an empty vector (Vec) or with pEGFP-N1 alone or together with p3×Flag-PP at a ratio of 1:3, the cells were treated with nuclease (2,500 U/ml; NEB) at 37°C for 10 min prior to cell lysis to remove extracellular nucleic acids. After that, the sample was divided in half and either total cellular DNA or RNA was extracted. The RNA contamination in DNA samples was eliminated using RNase A (100 μg/ml), and the DNA contamination in RNA templates was eliminated during reverse transcription using Fast King RT SuperMix containing DNase. GFP DNA and mRNA levels were determined using a SYBR green-based qPCR.

Western blot analysis.

Western blot analysis was performed as described previously (27). Mouse monoclonal antibody (MAb) to FCV VP1 (ab33990), rabbit MAb to GFP (ab32146), rabbit MAb to GAPDH (ab22555), rabbit polyclonal antibody (PAb) to Flag tag (ab1162) and rabbit PAb to XrnI (ab70259) were purchased from Abcam.

Subcellular localization.

Subcellular localization for proteins was determined as previously described (28, 29). At 24 h posttransfection, the CRFK cells were washed three times with phosphate-buffered saline with Tween 20 (PBST) and then fixed with 4% paraformaldehyde for 30 min. Following permeabilization with 0.2% Triton X-100 for 20 min at room temperature, the cells were blocked with 5% bovine serum albumin (BSA) for 1 h at room temperature and then incubated with mouse anti-Flag diluted in PBS for 2 h. The cells were then washed and incubated with secondary antibodies (Alexa Fluor 488 goat anti-mouse IgG [heavy plus light chain {H+L}]) (Abcam; ab150077) for 1 h. Finally, the cells were washed three times and stained using 1 μM DAPI (4′,6-diamidino-2-phenylindole; Sigma-Aldrich). A confocal laser scanning microscope (Leica) was used to detect fluorescence.

The subcellular localization of RNAs was determined as previously described (30). Briefly, the cytoplasmic and nuclear fractions were separated, and the RNA was extracted using a cytoplasmic and nuclear RNA purification kit (Norgen) according to the manufacturer’s instructions. cDNA was produced using Fast King RT SuperMix containing DNase (Tiangen, China) according to the manufacturer’s protocol, and quantitative PCR was performed to determine the ratio of RNA expression in the cytoplasmic and nuclear fractions.

Expression and purification of GST-PP fusion proteins.

The PP gene and its mutant PP H39A ORF were cloned into the pGEX-6P-1 vector, and then the recombination plasmid was transformed into Escherichia coli BL21(DE3). The expression and purification procedures were performed according to the method described previously (15). The purified proteins were then divided and stored at −80°C.

In vitro assay of RNase activity.

GFP, DDX50, and TRAF1 RNAs were synthesized by in vitro transcription of linearized pET-32a-EGFP with T7 RNA polymerase in the absence of a 5′ cap analog (Promega) according to the manual for HiScribe T7 High Yield RNA Synthesis Kit (NEB). Construction of the full-length cDNAs of 2280 and F9 and production of the full genomic RNAs of FCV 2280 and F9 are described in our previous study (17). An in vitro assay of RNase activity was performed according to the method described previously (15). Briefly, GST-PP or GST (10 μg) was incubated at 30°C with RNA (4 μg) in a 100-μl reaction mixture containing 25 mM Tris-HCl (pH 8.0), 80 mM potassium acetate, 1.5 mM magnesium acetate, 2 mM dithiothreitol (DTT), and 0.1 mM EDTA (15). After the reaction was finished, the samples were extracted with phenol-chloroform-isoamyl alcohol (pH 8.0; Ambion), precipitated with ethanol, and analyzed by 1.2% agarose-formaldehyde gel electrophoresis.

Assay of RNase activity of 2280 PP in the presence of protease or RNase inhibitors.

Briefly, a 100-μl reaction mixture containing GST-2280 PP (5 μg) and different doses of RRI (TaKaRa) or PMSF (Thermo Scientific) was incubated at 37°C for 30 min for RRI treatment or at 4°C for 30 min for PMSF treatment. After treatment, RNA (4 μg) was added into the reaction and incubated at 30°C for 90 min. After the reaction was finished, the samples were extracted with phenol-chloroform-isoamyl alcohol (pH 8.0; Ambion), precipitated with ethanol, and analyzed by 1.2% agarose-formaldehyde gel electrophoresis.

Statistics.

Statistical significance was determined using unpaired t tests in Prism 5.0 software (GraphPad Software). For all tests, a P value of <0.05 was considered to indicate a significant difference.