Host and Viral Translational Mechanisms during Cricket Paralysis Virus Infection

Abstract

The dicistrovirus is a positive-strand single-stranded RNA virus that possesses two internal ribosome entry sites (IRES) that direct translation of distinct open reading frames encoding the viral structural and nonstructural proteins. Through an unusual mechanism, the intergenic region (IGR) IRES responsible for viral structural protein expression mimics a tRNA to directly recruit the ribosome and set the ribosome into translational elongation. In this study, we explored the mechanism of host translational shutoff in Drosophila S2 cells infected by the dicistrovirus, cricket paralysis virus (CrPV). CrPV infection of S2 cells results in host translational shutoff concomitant with an increase in viral protein synthesis. CrPV infection resulted in the dissociation of eukaryotic translation initiation factor 4G (eIF4G) and eIF4E early in infection and the induction of deIF2α phosphorylation at 3 h postinfection, which lags after the initial inhibition of host translation. Forced dephosphorylation of deIF2α by overexpression of dGADD34, which activates protein phosphatase I, did not prevent translational shutoff nor alter virus production, demonstrating that deIF2α phosphorylation is dispensable for host translational shutoff. However, premature induction of deIF2α phosphorylation by thapsigargin treatment early in infection reduced viral protein synthesis and replication. Finally, translation mediated by the 5′ untranslated region (5′UTR) and the IGR IRES were resistant to impairment of eIF4F or eIF2 in translation extracts. These results support a model by which the alteration of the deIF4F complex contribute to the shutoff of host translation during CrPV infection, thereby promoting viral protein synthesis via the CrPV 5′UTR and IGR IRES.

For productive viral protein expression, viruses have to compete for and hijack the host translational machinery (45). Some viruses such as poliovirus, vesicular stomatitis virus (VSV), and influenza virus selectively antagonize the translation apparatus to shut off host translation, resulting in the release of ribosomes from host mRNAs and the inhibition of antiviral responses. On the other hand, the host cell can counteract through antiviral mechanisms to shutdown viral translation. For instance, viral RNA replication intermediates can trigger PKR, leading to an inhibition of overall translation. To bypass the block in translation, viruses have evolved unique mechanisms to preferentially recruit the ribosome for viral protein synthesis. Thus, the control of the translational machinery during infection is a major focal point in the battle between the host and the virus and often, elucidation of these viral translational shutoff strategies reveals key targets of translational regulation.

The majority of cellular mRNAs initiate translation through the recruitment of the cap-binding complex, eukaryotic translation initiation factor 4F (eIF4F), to the 5′ cap of the mRNA (56). eIF4F consists of the cap-binding protein eIF4E, the RNA helicase, eIF4A, and the adaptor protein eIF4G. eIF4G acts as a bridge to join eIF4E and the 40S subunit via eIF3. With the ternary eIF2-Met-tRNA_i-GTP complex bound, the 40S subunit scans in a 5′-to-3′ direction until an AUG start codon is encountered. Here, eIF5 mediates GTP hydrolysis on the ternary complex, releasing the eIFs and subsequently leading to 60S subunit joining to assemble an elongation-competent 80S ribosome. The ternary eIF2-Met-tRNA_i-GTP complex is reactivated for another round of translation by exchange of GDP for GTP, which is mediated by the guanine nucleotide exchange factor, eIF2B. The 3′ poly(A) tail of the mRNA also stimulates translational initiation by binding to the poly(A) binding protein (PABP), which in turn interacts with eIF4G at the 5′end, resulting in a circularized mRNA. PABP has been proposed to enhance eIF4E affinity for the 5′cap and promote 60S joining, indicating that PABP functions at multiple steps of translational initiation (33).

A common tactic viruses use to inhibit host translation is to selectively target eIFs. One of the best studied is the cleavage of eIF4G by viral proteases during picornavirus infection. In humans, two isoforms, eIF4GI and eIF4GII, are cleaved early in poliovirus infection by the viral protease 2A, where cleavage of eIFGII correlates more precisely with host translation shutoff (20). Cleavage of eIF4G produces an amino-terminal fragment that binds to eIF4E and a C-terminal fragment that binds to eIF4A and eIF3 (26, 39, 42). PABP is also cleaved by the viral protease 3C during poliovirus infection, thus contributing to shutoff of both host and viral translation and thereby enabling the switch from viral translation to replication (3, 31, 38). Another major target is the availability of the cap-binding protein eIF4E, which is regulated by binding to the repressor protein 4E-BP (21, 41). 4E-BP and eIF4G compete for an overlapping site on eIF4E (42). In its hypophosphorylated state, 4E-BP binds to and sequesters eIF4E, preventing eIF4G recruitment. Dephosphorylation and activation of 4E-BP has been observed during poliovirus, encephalomyocarditis (EMCV), and VSV infections (7, 18).

During virus infection, host antiviral responses are triggered that also inhibit translation to counteract viral protein synthesis. An integral antiviral response is phosphorylation at Ser⁵¹ of eIF2α, which reduces the pool of the ternary complex by blocking the eIF2B-dependent exchange of GDP to GTP. In mammals, four known eIF2α kinases exist including the endoplasmic reticulum (ER)-stress-inducible PERK, GCN2, which senses the accumulation of deacylated tRNAs during amino acid starvation conditions; the heme-regulated kinase HRI; and the interferon-inducible double-stranded RNA-binding PKR (64). In mammalian cells, PKR is activated by binding to double-stranded viral RNA replication intermediates, leading to eIF2α phosphorylation and inhibition of overall host and viral translation. PERK and GCN2 have also been shown to be activated during virus infections by VSV and members of the alphavirus family (2, 6, 43, 65, 79). Often, viruses rely on the ER for synthesis and proper folding of viral proteins. The large burden on the ER activates PERK to phosphorylate eIF2α, thereby inhibiting global protein synthesis to reduce the load on the ER (23). Some viruses such as HCV and herpes simplex viruses have adapted to responses that induce eIF2α phosphorylation by producing viral proteins that counteract PKR or modulate the ER stress response (27, 76). Thus, virus infection can trigger several eIF2α kinases that lead to translational shutoff to counteract viral protein synthesis.

To circumvent these translation blocks, viruses such as poliovirus and hepatitis C virus utilize internal ribosome entry sites (IRES), which are RNA elements that directly recruit ribosomes in a cap-independent manner and require only a subset of canonical eIFs (15, 25). It is generally thought that IRES-containing viral mRNAs can be translated under conditions when specific eIFs are compromised during infection. Except for a few cases, the specific mechanisms and factors that lead to IRES stimulation is poorly understood. For example, poliovirus and the related EMCV possess an IRES that allows viral translation despite cleavage of eIF4G during infection or inhibiting eIF4E by 4E-BP binding. This type of IRES can still bind to the central domain of eIF4G and mediate 40S subunit recruitment (11, 37, 57).

One of the most unique and simplest IRES is found within the intergenic region (IGR) of the Dicistroviridae family (for extensive reviews, see references 28, 36, and 49). Members of this family include the cricket paralysis virus (CrPV), drosophila C virus (DCV), taura syndrome virus, the Plautia stali intestine virus (PSIV), the Rhopalosiphum padi virus (RhPV), and several bee viruses such as the black queen cell virus and the Israeli acute paralysis virus, which has been recently linked to colony collapse disorder (10). The dicistroviruses encode a positive-strand 8- to 10-kb single-stranded RNA genome, which contains two main open reading frames, ORF1 and ORF2, encoding the nonstructural and structural proteins, respectively, separated by an IGR (see Fig. 1A). The 5′ end of the CrPV RNA is linked to the viral protein VpG and the 3′ end contains a poly(A) tail (16). Radiolabeling of intracellular RNA in infected cells reveals no subgenomic RNA species smaller than the full-length genomic RNA, and this has been supported by Northern blot analysis (16, 81). Translation of ORF2 is directed by the IGR IRES, whereas ORF1 expression is mediated by an IRES within the 5′ untranslated region (5′UTR) (35, 67, 81, 82). Remarkably, the IGR IRES element can directly recruit the ribosome independently of eIFs or the initiator Met-tRNA_i (29, 30, 54, 80). Furthermore, the IRES occupies the P-site of the ribosome to initiate translation from the ribosomal A-site encoding non-AUG codon (35, 81). Extensive biochemical and structural analyses from several groups have revealed that the IGR IRES mimics a tRNA that occupies the mRNA cleft of the ribosome and sets the ribosome into an elongation state (9, 29, 30, 34, 51, 55, 58, 68, 72, 83). Using reporter constructs, it has also been demonstrated that CrPV IGR IRES-mediated translation is active under a number of cellular conditions when the activity of the ternary complex eIF2-Met-tRNA_i-GTP is compromised (17, 63, 78, 80). Because IGR IRES-mediated translation does not require initiation factors, the IRES can direct translation under a number of cellular conditions when the activity of multiple eIFs is compromised (12). Although the majority of studies have focused on the IGR IRES of CrPV, PSIV, and TSV, it is predicted that the IGRs within this viral family all function similarly based on the predicted conserved RNA structures (28, 36, 49). In contrast, only the 5′UTR IRES mechanism of RhPV has been studied in detail (77). Despite the wealth of studies on the mechanics of these IRES, the mechanisms that lead to translational shutoff during dicistrovirus infection and the interaction of dicistrovirus with the host machinery to allow virus production have been relatively unexplored.

FIG. 1.

Kinetics of host protein synthesis and viral protein expression in CrPV-infected Drosophila S2 cells. (A) Genomic arrangement of the CrPV RNA. The viral open reading frames, ORF1 and ORF2, that encode nonstructural (NS) and structural (S) proteins, respectively, are shown, which are separated by the intergenic internal ribosome entry site (IGR IRES). Translation of ORF1 and ORF2 is directed by the 5′UTR IRES and the IGR IRES, respectively. The first amino acid of ORF2 directed by the IGR IRES is encoded by a GCU alanine codon. (B) Autoradiography of protein lysates resolved on a SDS-12% PAGE gel. The protein lysates were collected from S2 cells that were untreated (U), mock infected (M), CrPV infected (5 FFU/cell), or thapsigargin treated (Tg; 0.4 μM) for the indicated times (h p.i.) and metabolically labeled with [³⁵S]methionine for 30 min at each time point. The migration of proteins with known molecular masses is shown on the left. The expression of detectable nonstructural (NS) and structural (S) proteins is denoted. (C) Quantitation of host protein synthesis during CrPV infection. To calculate the host translation at each time point, the amount of radioactivity of the bands between 55 and 70 kDa in panel A was quantitated by using ImageQuant, and the percent translation was calculated at each time point of virus infection or thapsigargin treatment compared to the mock infection. Shown are averages (± the standard deviation) from at least three independent experiments. (D) Immunoblots of viral ORF1 and ORF2 during CrPV infection at various times postinfection (h p.i.). Antibodies were raised against peptides within ORF1 and ORF2. The expression of ORF1 and ORF2 was quantitated by a LI-COR Odyssey system, plotted against time of infection, and normalized to the amount of ORF1 or ORF2 expression at 6 h p.i. (100%). As a comparison, viral RNA synthesis as detected by Northern blot analysis (see Fig. 2B) is plotted on the same graph.

Previous studies have shown that the CrPV and the related DCV can infect a wide range of insect hosts, including the Drosophila melanogaster S2 cell line (60, 69). In the present study, we have explored how CrPV infection leads to host translational shutoff in S2 cells. Two steps of translational initiation are targeted during CrPV infection. First, the interaction of deIF4G with deIF4E is disrupted early in infection and remains dissociated during the course of infection. Second, deIF2α is phosphorylated at a time that lags after the initial host translational shutoff during infection. Premature phosphorylation of deIF2α early in infection inhibited translation directed by the 5′UTR IRES, but IGR IRES-mediated translation remained relatively resistant. These results support the model that multiple mechanisms, including impairment of deIF4F complex formation and induction of deIF2α phosphorylation, contribute to the host translational shutoff during CrPV infection. The inhibition of host translation and the release of ribosomes from host mRNAs ensures that translation mediated by the 5′UTR and IGR IRES is optimal to produce sufficient viral nonstructural and structural proteins for proper CrPV maturation and assembly.

MATERIALS AND METHODS

Cell culture and transfection.

Drosophila Schneider line 2 (S2) cells were passaged in M3+BPYE medium supplemented with 10% fetal bovine serum. Cells were treated with 0.4 μM thapsigargin or infected with five to six fluorescence-forming units (FFU) of CrPV/cell as specified. Transfections were performed with Lipofectamine 2000 (Invitrogen) using 2 μg of plasmid DNA into S2 cells according to the manufacturer's directions. For stable cells lines, we cotransfected a blasticidin selection plasmid (pIB/His/lacZ) at 0.2 μg and selected for stably expressing cells in the presence of 25 μg of blasticidin (Invitrogen)/ml.

Virus production and UV-irradiated virus.

Virus production in Drosophila S2 cells has been previously described (69, 70). A total of 3 × 10⁹ S2 cells were infected with 1 ml of virus suspension for 30 min, and then 60 ml of media was added back and the cells were incubated at 25°C for 14 h. The remaining cells were pelleted, resuspended in phosphate-buffered saline (PBS), and subjected to four freeze-thaws. CrPV was UV-irradiated by spotting in a thin layer and exposed to UV for 1 h on ice.

Metabolic radiolabeling and Western blots.

After treatment of cells with thapsigargin or infection with CrPV for the indicated times, 250 μCi of Easy-Tag Express [³⁵S]protein labeling mix (Perkin-Elmer)/ml was directly added for 30 min. Cells were washed in cold 1× PBS and lysed in lysis buffer (20 mM HEPES, 150 mM NaCl, 1% Triton X-100, 10% glycerol, 1 mM EDTA, 10 mM tetrapyrophosphate, 100 mM NaF, 17.5 mM β-glycerophosphate, and a protease inhibitor cocktail [Roche]). The concentration of protein was determined by Bradford assay (Bio-Rad), and equal amounts of lysates were separated by SDS-PAGE. For autoradiography, gels were dried, imaged by using a Typhoon imager, and quantified by using ImageQuant software. For immunoblots, protein from the gel was transferred to a polyvinylidene difluoride Immobilon-FL membrane (Millipore). Blots were blocked for 1 h in 5% skim milk and TBST (20 mM Tris, 150 mM NaCl, 0.1% Tween) and probed at 4°C overnight with one of the following antibodies: phospho-eIF2α rabbit polyclonal (Cell Signaling) (1:1,000), deIF2α rabbit polyclonal (1:500), d4E-BP rabbit polyclonal (1:2,000), PABP rabbit polyclonal (1:5,000), deIF4G rabbit polyclonal (1:500), deIF4A rabbit polyclonal antibody (1:2,000), deIF4E rabbit polyclonal (1:1,000), FLAG mouse monoclonal (Sigma) (1:2,000), CrPV ORF1 (raised against the CrPV RDRP peptide sequence C-REIVYHGKSEYQKLR-NH₂) rabbit polyclonal (1:10,000), or CrPV ORF2 (raised against CrPV structural protein VP2 peptide sequences C-ATFQDKQENSHIENE-NH₂ and C-KLWIHKTYLKRPAR-NH2) rabbit polyclonal (PL Laboratories, Burnaby, British Columbia, Canada) (1:10,000). The blots were washed three times in TBST, incubated with either IRDye 800CW goat anti-rabbit IgG (LI-COR biosciences) at1:25,000 or IRDye 680CW goat anti-mouse IgG (LI-COR biosciences) at 1:2000 at room temperature for 1 h, washed four times in TBST for 10 min, and scanned on an Odyssey imager (LI-COR Biosciences). Alternatively, a 1:30,000 dilution of donkey anti-rabbit IgG-horseradish peroxidase (Amersham) was used to detect proteins by enhanced chemiluminescence (Millipore).

Isolation of proteins using m⁷GTP-Sepharose.

Protein lysates (0.2 mg) were incubated with m⁷GTP Sepharose beads (GE Healthcare) at 4°C for 1 h and then washed twice with buffer (20 mM HEPES, 50 mM β-glycerophosphate, 0.5 mM EGTA, 0.5 mM EDTA, 1% Triton X-100, and a protease inhibitor cocktail [Roche] at pH 7.4). Proteins were eluted from the beads by boiling in SDS-PAGE loading buffer, and the supernatant was separated on an SDS-10% PAGE gel and subsequently analyzed by Western blot analysis.

Polysome analysis and RNA extractions.

Sucrose gradient centrifugation and polysome analysis were prepared as described previously (32). Briefly, for each gradient, 3 × 10⁷ of stably transfected S2 cells (pAct-GFP) were either mock infected or infected with CrPV. At the indicated time points, cells were treated with 0.1 mg of cycloheximide/ml and placed on ice for 5 min. Cell pellets were harvested in 500 μl of polysome lysis buffer (15 mM Tris-Cl [pH 7.5], 15 mM MgCl₂, 300 mM NaCl, 1% Triton X-100. 0.1 mg of cycloheximide/ml, 1 mg of heparin/ml, and 1 U of Ribolock [Fermentas]/μl). After incubation on ice for 5 min and centrifugation at 12,000 × g at 4°C for 5 min, the supernatants were layered onto an 11-ml 10 to 50% sucrose gradient containing polysome lysis buffer that has 0.1 U of Ribolock/μl and no Triton X-100) and centrifuged at 35,000 rpm using a SW41 rotor for 195 min at 4°C. Fractions were collected from the top by using an ISCO fraction collector and a Brandel syringe pump system. To collect RNA, 3 ml of 8 M guanidine HCl and 5 ml of 95% ethanol were added to each fraction, followed by precipitation and resuspension in water. Equal volumes of fractionated RNA were subjected to Northern blot analysis.

Northern blots.

For total RNA, RNA was isolated from cells using TRIzol reagent (Invitrogen). RNA was separated on a denaturing agarose gel and transferred to Zeta-probe blotting membrane (Bio-Rad). Radiolabeled DNA hybridization probes were generated by using a RadPrime kit (Invitrogen). The amount of radiolabeled probe hybridized to the blot was quantitated by phosphorimager analysis (Typhoon; Amersham Biosciences).

Cloning of Drosophila FLAG-GADD34.

The protein sequence of human GADD34 or phosphatase 1 regulatory subunit 15A (accession number NP_055145) was used to search the Drosophila database for the GADD34 homolog using BLAST. A single protein (NP_611863) corresponding to a single gene (NM_138019, CG3825) with 42% identity in the C-terminal region (human GADD34 amino acids 546 to 616 and Drosophila version amino acids 241 to 309) was identified. The C-terminal region of the human GADD34 is responsible for the activity that dephosphorylates eIF2α (52, 53). Using this sequence, primers were designed to PCR amplify the full-length coding region of GADD34 or 180 amino acids of the most C-terminal region of GADD34. The primers were designed such that a FLAG tag was fused in-frame upstream of the GADD34 coding region: full-length FLAG GADD34 forward primer, 5′-GGTACCATGGATTACAAGGATGACGATGACAAGTCGAAATTTCACACCCTTATGGGC-3′; FLAG C-term GADD34 forward primer, 5′-GGTACCATGGATTACAAGGATGACGATGACAAGCAGCGTAGTATTTCCGAGTGCAG-3′; and GADD34 reverse primer, 5′-TCTAGACTACTAGTCCTCGTGCAAGAATCGAG-3′. The italicized nucleotides of the primer correspond to the FLAG-tag and the underlined nucleotides denote either a KpnI or XbaI restriction site. PCR products were amplified from a Drosophila cDNA library, digested with KpnI and XbaI, and subsequently ligated into the same sites of the pAc5 plasmid containing the Drosophila Act5C promoter to produce pAc5-FLAG FL-GADD34 and pAc5-FLAG C-term GADD34. A similar strategy was used to clone the reporter eGFP using KpnI and XbaI sites to produce pAc5-eGFP.

Viral titer.

S2 cells were infected with 6 FFU/cell for 12 h. Cells were harvested and lysed in PBS by using four freeze-thaw cycles. Subsequently, the cell debris was pelleted by centrifugation. A total of 2 × 10⁵ S2 cells were incubated with serial dilutions of supernatants for 6 h and then plated in wells coated with 0.5 mg of concanavalin A (Calbiochem)/ml for 1 h. The cells were washed once with 1× PBS and then fixed for 20 min with 3% paraformaldehyde, followed by methanol for 5 min. After blocking overnight (1× PBS, 10 mM glycine, 0.2% Triton X-100, 1% bovine serum albumin), a 1:125 dilution of anti-ORF2 antibody in blocking solution was added, followed by incubation for 1 h. The cells were washed twice for 10 min each time in blocking solution and incubated in the dark with a 1:125 dilution of goat anti-rabbit Texas Red IgG (Invitrogen). Cells were washed twice for 10 min each time in blocking solution, and nuclei were stained with 0.5 μg of Hoechst stain/ml in blocking solution. A Cellomics Arrayscan HCS instrument for automated microscopic analysis was used to image and quantify the amount of infected cells showing fluorescence over background. Through serial dilutions of CrPV, the FFU/ml can be calculated. Virus titer calculations were from at least three replicates for mock and CrPV infections.

dXbp-1 splicing.

cDNA was synthesized from total RNA using random primers by Superscript III (Invitrogen) according to the manufacturer's directions. Primers for reverse transcription-PCR (RT-PCR) were designed that flank the 23-nucleotide intron splice sites (nucleotides 898 to 920) of dXbp-1 (accession number NM_079983): dXbp-1 forward primer, 5′-TGCTGCGCCAAGAACTCGCCG-3′ (nucleotides 781 to 801); and dXbp-1 reverse primer, 5′-GCCACAACTTTCCAGAGTGAG-3′ (nucleotides 1021 to 1001). Unspliced and spliced PCR products yield 240-bp and 217-bp fragments, respectively. Primers that PCR amplify dAct5C were designed as controls (dAct5C fwd primer, 5′-TTCTTGGGAATGGAGGCTTGCG-3′; dAct5c rev primer, 5′-ACAG CACGGTGTTGGCATACAG-3′). The PCR products were separated on a 5% acrylamide gel and detected with ethidium bromide.

In vitro translation.

Plasmids containing dicistronic luciferase plasmids were linearized with XbaI. Dicistronic RNAs were in vitro transcribed using a bacteriophage T7 RNA polymerase reaction, and RNA was purified by using an RNeasy kit (Qiagen). The integrity and purity of the RNAs were confirmed by gel analysis. Uncapped RNAs were incubated in rabbit reticulocyte lysate (RRL) with 154 mM (final) potassium acetate. Luciferase activities were measured by using the Promega dual luciferase assay. We also analyzed the expression of luciferase protein by measuring the incorporation of [³⁵S]methionine and by SDS-PAGE. Gels were dried and exposed to film. Where indicated, the rhinovirus 2A^pro or 4E-BP1 was expressed in RRL. For 2A^pro (a generous gift from Kurt Gustin), the pCITE-2A plasmid, containing the rhinovirus type 14 2A protease open reading frame, was linearized with BamHI. 2A^pro was in vitro transcribed in a T7 polymerase reaction. Purified RNA encoding 2A^pro was incubated in RRL for 30 min at 30°C prior to the addition of reporter RNAs. For 4E-BP1, the 4E-BP1-expressing plasmid was incubated in a TNT RRL (Promega) for 30 min at 30°C prior to the addition of the reporter RNAs.

RESULTS

Host translation shutoff during CrPV infection.

CrPV infection of S2 cells resulted in the expected rapid shutoff of host protein synthesis (Fig. 1B and C) (48, 81). Cells were pulse-labeled with [³⁵S]methionine for 30 min at each time point and assayed by SDS-PAGE and phosphorimager analysis. In mock-infected cells, labeling of proteins was detected throughout the lane, indicating active translation (M, mock lanes, Fig. 1B). In contrast, protein synthesis in CrPV-infected cells was inhibited by 50% as early as 2 h postinfection (h p.i.), and a maximal inhibition was observed by 4 h p.i. (80 to 90% inhibition of translation) (Fig. 1B and C). A quantitation of host protein synthesis was calculated by measuring the radioactive bands between 55 and 70 kDa at each time point of infection compared to the corresponding mock-infected lysates. The extent of host translational shutoff was similar to that observed in cells treated with thapsigargin (Fig. 1B), which induces ER stress by inhibiting the ER-resident Ca²⁺ transport ATPase. Despite an overall inhibition in host protein synthesis in CrPV-infected cells, several viral proteins were readily detected (Fig. 1B). Most notable, the ORF2 structural proteins—VP1, VP2, and VP3—that migrate between 30 and 45 kDa were detected as early as 2 h p.i. and increased dramatically thereafter (Fig. 1B) (48, 81). In contrast, the viral nonstructural proteins were barely detectable. These results are in agreement with previous reports that the ORF2 structural proteins are produced in supramolar excess over the ORF1 proteins and is likely reflected in the distinct IRES mechanisms that direct translation of each ORF (46, 48, 81). To confirm that these bands corresponded to CrPV proteins, we used antibodies that were raised against peptides within the RNA-dependent-RNA polymerase (RdRP) and the viral capsid protein VP2 encoded by ORF1 and ORF2, respectively (Fig. 1D). The ORF1 and ORF2 antibodies recognize proteins of the expected sizes, i.e., ∼140 and 35 kDa, of RdRP and VP2, respectively. In some experiments, the ORF1 antibody detected a minor higher molecular weight protein (>200 kDa), which likely represents the precursor ORF1 polyprotein. Measurements of the amount of ORF1 and ORF2 expression over time normalized to the amount of each viral protein produced at 6 h p.i. (100%) showed that the overall rates of accumulation of ORF1 and ORF2 proteins were similar (Fig. 1D). To determine whether the rate of viral RNA synthesis correlated with viral protein expression, we monitored CrPV RNA levels by Northern blot analysis. (Fig. 1D and 2B). CrPV RNA production mirrored the increasing expression of CrPV ORF1 and ORF2 protein, indicating that CrPV viral mRNAs are translated during infection when host protein synthesis is shutoff.

FIG. 2.

UV-irradiated CrPV infection in Drosophila S2 cells. (A) S2 Cells were infected with CrPV (10 FFU/cell) or the equivalent of UV-irradiated CrPV for 0, 2, 4, or 6 h and then directly radiolabeled with [³⁵S]methionine for 30 min at each time point. The protein lysates were visualized by autoradiography of a SDS-12% PAGE gel with the amount of radioactivity in each lane normalized to that of mock-infected cells (100%) (below each lane). (B) In a parallel experiment, RNA was purified at each h p.i. and subjected to Northern blot analysis by probing for CrPV RNA.

Host translation shutoff requires ongoing viral translation and/or replication.

A previous report showed that the related dicistrovirus, DCV, binds to and enters S2 cells via endocytosis (5). The host translational shutoff in CrPV-infected cells could be due to an initial host response to virus entry. To determine whether CrPV entry into S2 cells specifically leads to the rapid shutoff of protein synthesis, S2 cells were infected with UV-irradiated CrPV. Incubation of S2 cells with UV-irradiated CrPV resulted in a nearly undetectable amplification of viral RNA, indicating that the majority of CrPV RNA was not functional (Fig. 2B). In contrast to that observed in CrPV-infected cells, host proteins synthesis was not inhibited in cells incubated with UV-irradiated CrPV (Fig. 2A), suggesting that host translational shutoff during CrPV infection is not in response to viral entry per se but likely due to viral translation and/or replication.

Translation initiation is blocked in CrPV-infected S2 cells.

To determine whether host translation shutoff during CrPV infection is due to a block at the translational initiation step, we performed sucrose gradient centrifugation analysis. In this assay, lysates from mock- or CrPV-infected cells were layered on a 10 to 50% sucrose gradient, centrifuged, and fractionated. If translational initiation is blocked, ribosomes will dissociate from mRNAs leading to a decrease in polysomes. In these experiments, we used a stable S2 cell line that expresses a reporter green fluorescent protein (GFP) mRNA to monitor an abundantly expressed mRNA that is under cap-dependent translational control. In mock-infected cells, the majority of ribosomes sedimented to higher-molecular-weight fractions, indicating that most ribosomes are actively engaged in translation (polysomes) (Fig. 3). In contrast, CrPV-infected cells at 2, 4, and 6 h p.i. resulted in a progressive loss in polysomes and an accompanying increase in free 40S and 60S ribosomal subunits during the course of infection (Fig. 3). The decrease of polysomes during infection correlated with the shutoff in protein synthesis (Fig. 1B). To confirm that translation was inhibited, RNA from each fraction was subjected to Northern blot analysis. As expected, in mock-infected GFP-expressing cells, eIF1A, GAPDH, and GFP mRNAs associated mostly with higher-molecular-weight fractions (fractions 11 to 15), indicating that these mRNAs were actively translated. In contrast, by 4 h p.i., the majority of mRNAs shifted to lower-molecular-weight fractions in CrPV-infected cells, demonstrating that translation of host mRNAs was blocked at the initiation step. As expected, the CrPV genomic RNA was detected in high-molecular-weight fractions associated with polysomes, which is probably due to the ability of the CrPV RNA to recruit the ribosomes via its 5′UTR and IGR IRES.

FIG. 3.

Polysome analysis of CrPV-infected S2 cells. Lysates from CrPV-infected S2 cells (6 FFU/cell) that stably express the GFP reporter mRNA at 2, 4, and 6 h p.i. were subjected to sucrose gradient centrifugation sedimentation. Lysates were layered on a 10 to 50% sucrose gradient, centrifuged, and fractionated. Shown at the top are absorbances at 254 nm, and the sedimentation of 40S, 60S, 80S, and polysomal peaks are indicated. The top to the bottom of the gradient is represented from left to right, respectively. Below each spectrograph is a Northern blot analysis of RNA from each fraction probing for deIF1A, dGAPDH, GFP, and CrPV RNAs. Note that deIF1A, dGAPDH, and GFP mRNAs shift from high-molecular-weight fractions in mock-infected cells to lower-molecular-weight fractions in CrPV-infected cells. Note also the extra peak in CrPV-infected cells at 6 h p.i. representing packaged virus.

deIF4F complex is altered during CrPV infection.

To begin exploring how host translation initiation is shutoff during CrPV infection, we first assayed the status of selective eIFs. d4E-BP, deIF4G, and dPABP protein levels in CrPV-infected cells were analyzed by Western blot analysis. Unlike that observed in poliovirus-infected cells, deIF4G and dPABP proteins were not cleaved during CrPV infection, demonstrating that these proteins are not targeted by proteases (Fig. 4A). To determine whether d4E-BP may have a role during CrPV infection, we monitored the phosphorylation status of d4E-BP. Hyper- and hypophosphorylated d4E-BP proteins can be resolved on a high percentage SDS-PAGE gel and assayed by Western blot analysis with antibodies directed against 4E-BP (44). After 1 h p.i., an increase in the hypophosphorylated form of d4E-BP was reproducibly observed (indicated with an asterisk, Fig. 4A). From 2 h p.i. and onward, expression of both hyper- and hypophosphorylated forms of d4E-BP was reduced, suggesting that deIF4E is not regulated by d4E-BP at later times of infection (Fig. 4A). Interestingly, the mechanism by which overall d4E-BP levels are reduced during infection remains to be examined.

FIG. 4.

Characterization of translation initiation factors during CrPV infection. (A) Immunoblots of d4E-BP, deIF4A, dPABP, deIF4G, and deIF4E from equal amounts of lysates (shown in Fig. 1B) of mock (M)- and CrPV-infected S2 cells at the indicated time postinfection. (B) Equal amounts of lysates from mock (M)-, UV-irradiated (UV)-, and CrPV-infected S2 cells (5 FFU) at each time point were subjected to m⁷G-Sepharose bead pulldowns. “1 M” and “6 M” denote mock infections at 1 and 6 h, respectively. Proteins retained in the cap resin were resolved by SDS-PAGE analysis and immunoblotted with antisera against deIF4G, deIF4E, and d4E-BP. (C) Immunoblots of phospho-eIF2α (P-deIF2α) and deIF2α from equal amounts of lysates (shown in Fig. 1B) of mock (M)- and CrPV-infected S2 cells at the indicated times postinfection (below). Quantitation of the fraction of P-deIF2α in CrPV-infected S2 cells compared to that in mock-infected cells (given as 1.0) was carried out. (D) Immunoblots of phospho-eIF2α (P-deIF2α) and total deIF2α from lysates of CrPV and UV-irradiated infected cells.

To further investigate whether the deIF4F complex was affected, we assayed for proteins that interact with deIF4E by using 7-methyl-GTP (m⁷G) Sepharose-bead pulldown assays at different times postinfection. Candidate proteins that interact with the cap resin were then analyzed by Western blot analysis. deIF4E protein bound to the cap resin throughout the course of CrPV infection (Fig. 4B), suggesting that the activity of deIF4E was not impaired. In contrast, CrPV infection resulted in the dissociation of deIF4G and deIF4E, as indicated by the reduced levels of deIF4G in the cap resins at all time points of infection compared to that in mock-infected lysates at 1 and 6 h p.i. (Fig. 4B). At 1 h p.i. but not at later times postinfection, d4E-BP was reproducibly detected in the cap resin, suggesting that d4E-BP interacted with and sequestered deIF4E and in agreement with the transient increase in the hypophosphorylated form of d4E-BP early in infection (Fig. 4A and B, left panel). However, during the course of experiments, we found that S2 cells were sensitive to mechanical stress. For instance, our infection protocol requires a quick centrifugation and incubation of cells with virus at the beginning of the infection. This resulted in a stress response that increases d4E-BP association with deIF4E, as observed in the 1 h mock- and UV-irradiated-CrPV-infected lysates (Fig. 4B, right panel). As the cell recovers from the stress, d4E-BP no longer associated with the cap resin at 6 h p.i. (Fig. 4B, right panel, lane 2). Therefore, the transient increase in d4E-BP binding to deIF4E in CrPV-infected cells is due to the stress imposed on the cells in the experimental protocol. Despite an increase in 4E-BP association with deIF4E, similar amounts of deIF4G bound to the cap resin in mock-infected and UV-irradiated-CrPV-infected lysates, whereas deIF4G is not bound in CrPV-infected lysates (Fig. 4B, right panel, compare lanes 1, 3, and 5), suggesting that the amount of d4E-BP pulled down in the cap resin is not sufficient to prevent deIF4G-deIF4E interactions. This is consistent with our findings that overall translation is not affected in mock- and UV-irradiated-CrPV-infected cells (Fig. 1B and 2A). Thus, it appears that host translational shutoff in CrPV-infected cells correlates with the dissociation of deIF4G from the cap resin and not d4E-BP binding.

CrPV infection induces deIF2α phosphorylation.

We next examined the phosphorylation status of deIF2α during CrPV infection, a key component of the heterotrimeric factor deIF2 (Fig. 4C). Using phospho-specific deIF2α, phosphorylated deIF2α is first detected at 3 h p.i. and becomes elevated at later time points postinfection (Fig. 4C). The extent of deIF2α phosphorylation compared to that in mock-infected cells is quantitated and shown below in Fig. 4C. The slight increase in deIF2α phosphorylation is consistently observed at 3 h p.i. and reproducibly occurred after the initial shutoff of protein synthesis, as shown in Fig. 1C (∼2 h p.i., 50% inhibition). In contrast, deIF2α was not phosphorylated in cells incubated with UV-irradiated CrPV at all time points (Fig. 4D), indicating that ongoing viral replication is required for deIF2α phosphorylation.

CrPV infection does not affect the UPR pathway.

Because host translational shutoff precedes deIF2α phosphorylation in CrPV-infected cells, we investigated whether inducing deIF2α phosphorylation early in infection had an effect on CrPV infection. To test this, we treated mock- and CrPV-infected S2 cells with 400 nM thapsigargin at 0 and 3 h p.i. and then assayed for protein synthesis at 6 h p.i. by ³⁵S pulse-labeling for 30 min. In mammalian cells, it has been shown that thapsigargin treatment leads to a transient inhibition, followed by a recovery in protein synthesis, which is due to an acute increase in eIF2α phosphorylation, followed by dephosphorylation of eIF2α at later time points (52). As expected, thapsigargin treatment of S2 cells for 3 and 6 h resulted in a 60% decrease in protein synthesis concomitant with an increase in eIF2α phosphorylation (Fig. 5A, lanes 2 and 3). Unlike that observed in mammalian cells, protein synthesis did not recover, nor did deIF2α become dephosphorylated at the later time points during this treatment in S2 cells. As shown in Fig. 1B, CrPV infection of S2 cells for 6 h resulted in a shutoff in host protein synthesis (∼80% inhibition) (Fig. 5A, lane 4). When thapsigargin was added to CrPV-infected cells at 0 and 3 h p.i., host translation was shut off to a greater extent than CrPV infection alone to ∼90% inhibition, as assayed by radiolabeling at 6 h p.i. (Fig. 5A, ∼90% inhibition, lanes 5 and 6). In agreement, thapsigargin treatment at earlier times resulted in a greater degree of deIF2α phosphorylation compared to thapsigargin treatment alone (Fig. 5A). Furthermore, thapsigargin incubation at these early time points (0 and 3 h p.i) of infection also led to decreased radiolabel incorporation in viral proteins, suggesting that translation of ORF1 and ORF2 was impaired (Fig. 5A). To monitor viral expression more closely, we assayed for ORF1 and ORF2 expression by Western blot analysis (Fig. 5B). The addition of thapsigargin to CrPV-infected cells at 0 and 3 h p.i. consistently resulted in a decrease in ORF1 and ORF2 protein expression (Fig. 5A). The quantitations of ORF1 and ORF2 expression are shown in Fig. 5B. Interestingly, viral ORF1 protein expression was inhibited more than ORF2 protein expression, suggesting that translation of ORF1 mediated by the 5′UTR IRES was more sensitive to the effects of thapsigargin treatment than ORF2 translation mediated by the IGR IRES (Fig. 5B). This is consistent with previous reports that IGR IRES-mediated translation is relatively resistant to eIF2α phosphorylation (17, 78, 80). Because the expression of viral RdRP encoded within ORF1 was significantly inhibited, it would be predicted that viral replication would also be inhibited in thapsigargin-treated CrPV-infected cells. Although the addition of thapsigargin to CrPV-infected cells at 3 h p.i. moderately inhibited (20% decrease) viral synthesis, the addition of thapsigargin at the beginning of infection led to a dramatic 75% decrease in the amount of viral RNA, as measured by Northern blot analysis (Fig. 5B). These results suggest that premature induction of eIF2α phosphorylation by thapsigargin treatment inhibits CrPV replication and translation.

FIG. 5.

CrPV infection in S2 cells does not induce ER stress. (A) Mock-infected (lanes 1 to 3) or CrPV-infected (lanes 4 to 6) S2 cells (10 FFU/cell) were left untreated (zero time with Tg) or treated with 0.4 μM thapsigargin (Tg) for 3 or 6 h. For CrPV-infected cells, thapsigargin was added at 0 or 3 h p.i. and subsequently incubated for 6 and 3 h, respectively (denoted as time with Tg [in hours]). After CrPV infection for 6 h p.i., the cells were metabolically labeled with [³⁵S]methionine for 30 min. The protein lysates were visualized by autoradiography of a SDS-12% PAGE gel, and the percent radioactivity in each lane is shown below each lane. To calculate the host translation at each time point, the amount of radioactivity of the bands between 50 and 70 kDa was quantitated by using ImageQuant, and the percent translation was calculated. Shown below are immunoblots of ORF1, ORF2, P-deIF2α, and deIF2α from the lysates in panel A. The gel and immunoblots shown are representative of at least two independent experiments, which produced similar results. (B) Quantitation of ORF1 and ORF2 expression normalized to that in CrPV-infected cells. In a parallel experiment, RNA was purified from cells treated as in panel A and subjected to Northern blot analysis using a probe for CrPV RNA. The quantitation of the CrPV RNA is shown. (C) dXbp-1 mRNA splicing during CrPV infection. Total RNA was purified from mock- or CrPV-infected S2 cells that were left untreated (lane 1) or treated with 0.4 μM thapsigargin for the indicated times. For CrPV-infected cells, RNA was purified at 6 h p.i. The purified RNA was subjected to RT-PCR analysis using specific dXbp-1 primers that flank the spliced intron (see Materials and Methods). Upon thapsigargin-induced ER stress, the 23-nucleotide intron of dXbp-1 mRNA is spliced to produce the shorter spliced PCR product. Note that CrPV infection for 6 h does not lead to the spliced version of dXbp-1 mRNA. Moreover, CrPV infection does not prevent thapsigargin-induced splicing of dXbp-1 mRNA.

Phosphorylation of deIF2α during CrPV infection could be induced by either of the two known eIF2α kinases in D. melanogaster, dGCN2 or dPERK (1, 61, 66). In amino acid-starved cells, the accumulation of deacylated-tRNAs triggers GCN2 to induce deIF2α phosphorylation, whereas in response to the accumulation of unfolded proteins in the ER, the ER-resident protein, PERK, is activated, resulting in the phosphorylation of deIF2α (13). The activation of PERK is one arm of three main signaling pathways that responds to ER stress, collectively called the unfolded protein response (UPR) (64). The UPR activates another ER-resident protein called IRE1, which possesses endonuclease activity and when activated during ER stress, leads to the splicing of a 23 nucleotide intron of dXbp-1 mRNA (4, 84). The spliced version of Xbp-1 mRNA produces a frameshift in the translational reading frame resulting in the synthesis of a potent transcription factor that activates downstream ER stress-response genes (4, 84). To begin distinguishing which eIF2α kinase may be responsible for eIF2α phosphorylation in CrPV-infected cells, we reasoned that if dPERK was activated via ER stress, then dIRE1 would also be activated and thereby lead to splicing of the 23 nucleotide intron of dXbp-1 mRNA (59, 71). To detect the spliced dXbp-1 mRNA, we designed primers that flank the intron, from which we could distinguish the spliced and unspliced versions by RT-PCR amplification. As expected, the treatment of cells with the UPR-inducer, thapsigargin, for 3 and 6 h led to the expected shorter spliced product of dXbp-1, as detected by RT-PCR amplification (Fig. 5C). In contrast, CrPV infection did not lead to the spliced product of dXbp-1 mRNA, and only the unspliced version was detected (Fig. 5C, lane 4), indicating that the UPR response and thus dIRE1 and dPERK are not activated during CrPV infection. However, it is possible that CrPV infection may suppress the UPR response. To address this possibility, we treated CrPV-infected cells with thapsigargin at 3 h p.i. and assayed for the spliced dXbp-1 mRNA at 6 h p.i. As in the thapsigargin-treated cells, thapsigargin treatment of CrPV-infected S2 cells produced the spliced dXbp-1 mRNA (Fig. 5C, lane 5), demonstrating that the UPR signaling pathway via dIRE1 is intact during infection. Given that the IRE1 pathway of the UPR is not induced, the results suggest that CrPV infection in S2 cells does not lead to ER stress.

deIF2α phosphorylation is dispensable for CrPV replication.

The phosphorylation of eIF2α during CrPV infection is intriguing since it has been demonstrated by several groups that the CrPV IGR IRES can direct translation under various cellular conditions that induce eIF2α phosphorylation (17, 40, 62, 63, 80). Thus, the finding that deIF2α is phosphorylated during CrPV infection may provide a mechanistic explanation of how host translation is shutoff in CrPV-infected cells without effecting translation of ORF2. To determine whether deIF2α phosphorylation has a direct role in these properties, we suppressed deIF2α phosphorylation during CrPV infection by overexpressing either the full-length or the C-terminal fragment of GADD34, both of which can interact with and activate protein phosphatase I (PPI) to dephosphorylate eIF2α (52, 53). In mammalian cells, overexpression of the full-length or a truncated C-terminal version of GADD34 can suppress deIF2α phosphorylation and prevent the inhibition of protein synthesis during ER stress (8, 52, 53). The induction of GADD34 is critical for translational recovery during ER stress (53). Because the Drosophila GADD34 had not been identified, we searched the D. melanogaster database for the mammalian homolog of GADD34 and identified a single protein, CG3825, with homology (42% amino acid identity) within the C-terminal region (see Materials and Methods), which is the critical domain that interacts with PP1 to dephosphorylate eIF2α (8, 52). Similar to what we observed in mammalian cells, transfection of plasmids containing a FLAG-tagged full-length or the C-terminal fragment of GADD34 in S2 cells prevented the phosphorylation of eIF2α during CrPV infection and thapsigargin- and dithiothreitol (DTT)-induced ER stress (Fig. 6A and B, data not shown). DTT, like thapsigargin, induces the accumulation of unfolded proteins in the ER to activate the UPR. Suppression of deIF2α phosphorylation was specific since transfection of a control plasmid that expresses the reporter GFP did not prevent deIF2α phosphorylation (Fig. 6A). The extent of deIF2α phosphorylation of a representative experiment is quantitated and shown in Fig. 6B. Although deIF2α remained dephophosphorylated, overexpression of GADD34 did not prevent the shutoff of host protein synthesis during CrPV infection (Fig. 6A). The extent of translational shutoff during CrPV infection was similar in control- and GADD34-transfected cells (quantitation in Fig. 6C). To determine whether viral protein synthesis was affected, we monitored ORF1 and ORF2 expression by Western blot analysis in control- and GADD34-transfected CrPV-infected cells. ORF1 and OF2 expression were not significantly altered (Fig. 6A), suggesting that the translation of ORF1 and ORF2 was not affected when eIF2α was dephosphorylated during CrPV infection.

FIG. 6.

eIF2α phosphorylation is dispensable for host translation shutoff in CrPV-infected S2 cells. (A) S2 cells were transfected with plasmids that contain the dAct5C promoter and either the reporter GFP (control) or the FLAG-tagged C-term GADD34 (C-term GADD34). After a 60-h transfection, cells were either mock (M) or CrPV (V) infected (6 FFU/cell; at 2, 4, and 6 h p.i.) or treated with 2 mM DTT (D) or 0.4 μM thapsigargin (T) and then metabolically labeled with [³⁵S]methionine for 30 min at the indicated times. The lysates were subjected to SDS-PAGE, and the radioactivity detected by phosphorimager analysis. Below the gel are immunoblots of phospho-eIF2α, total eIF2α, ORF1, ORF2, and the FLAG-tagged C-term GADD34 from the same lysates. Shown are representative gels from three independent experiments. (B) Quantitation of the fraction of P-eIF2α/total eIF2α immunoblots shown in panel A. (C) The percent radioactivity of each lane shown in panel A, given that the mock-infected value is 100%.

To determine whether CrPV virus production was affected, virus titers were determined by detection of ORF2 expression in infected cells. Briefly, control- and GADD34-transfected cells were infected with CrPV for 72 h. The lysates from the infected cells were then used to infect a fresh batch of S2 cells. After 6 h of incubation, the cells were fixed, and the number of cells expressing ORF2 protein was detected by immunofluorescence. Expression of the C-terminal fragment of GADD34 did not alter CrPV virus titers compared to cells transfected with the control plasmid (Fig. 7). In summary, these results demonstrate that deIF2α phosphorylation is dispensable for host translational shutoff, and viral production in CrPV-infected S2 cells.

FIG. 7.

Viral production in GADD34-transfected CrPV-infected S2 cells. After a 60-h transfection with the control GFP or FLAG-tagged GADD34 vector, lysates from mock- or CrPV-infected S2 cells were plated on fresh S2 cells. After 6 h p.i., the cells were fixed and stained for DNA with Hoechst (blue) and for ORF2 expression using antisera against the VP2 capsid protein (red). Representative immunostains are shown, and the percentage of cells staining for ORF2 expression is also shown. Shown are the averages (± the standard deviation) from at least three independent experiments.

Analysis of CrPV IRES translation in vitro.

Our results suggest that the impairment of eIF4F and induction of eIF2α phosphorylation during CrPV infection may lead to viral IRES translation. To test this hypothesis, we turned to the established RRL translation system where specific initiation factors can be altered and IRES translation can be directly assayed by incubating reporter dicistronic RNAs. Within the dicistronic RNAs, the CrPV 5′UTR IRES or IGR IRES mediate the translation of the downstream cistron (firefly luciferase), whereas the upstream cistron, Renilla luciferase, monitors scanning-mediated translation (Fig. 8A). As shown previously, translation mediated by the IGR IRES was much more active than that mediated by the 5′UTR IRES (Fig. 8A) (81). 5′UTR IRES translation was relatively weak but reproducibly above background compared to translation of the dicistronic RNA containing no IRES (Fig. 8A). To impair eIF4F formation, we preincubated RRL with a plasmid that expresses 4E-BP1, which dissociates eIF4G and eIF4E. Using [³⁵S]methionine to monitor protein expression, 4E-BP1 protein (∼20 kDa) was readily detected after a 30-min incubation in RRL (zero time point in Fig. 8B). Subsequently, reporter dicistronic RNAs were added and lysates collected after 0, 30 and 60 min of incubation. Addition of the 4E-BP resulted in a significant decrease (∼80 to 90% inhibition) in scanning-mediated translation of Renilla luciferase (Fig. 8B, lanes 2, 3, 8, 9, 14, and 15). Because the luciferase bands were difficult to detect, we quantitated the fold Renilla (scanning-mediated) and firefly (IRES-mediated) luciferase activities by using an enzymatic assay and are shown in Fig. 8C. In contrast to Renilla luciferase expression, translation of the firefly luciferase mediated by the IGR IRES and the 5′UTR IRES was still active under these conditions. In fact, translation by the IGR IRES was approximately two- to fourfold more active in RRL expressing 4E-BP, indicating that IRES activity was stimulated (Fig. 8C). The 5′UTR IRES was slightly but reproducibly stimulated (∼1.3-fold) when 4E-BP was expressed (Fig. 8C). To determine whether impairing eIF4F through another mechanism also stimulated IRES activity, we expressed the rhinovirus 2A protease (2A^pro), which cleaves eIF4G (19, 22). The coding region of 2A^pro was engineered downstream of the EMCV IRES, which ensures expression of 2A^pro during the reaction when eIF4G is cleaved (data not shown) (2A^pro expression plasmid generous gift by Kurt Gustin). Preincubating an in vitro transcribed RNA encoding the EMCV IRES-2A^pro in RRL resulted in a detectable 2A^pro protein by [³⁵S]methionine labeling after a 30-min incubation (data not shown). 2A^pro expression decreased Renilla luciferase activity, indicating that scanning-mediated translation is reduced (Fig. 8D). However, IGR IRES-mediated translation was resistant to this treatment, whereas 5′UTR IRES-mediated translation was barely detectable (Fig. 8D). In summary, translation of the 5′UTR IRES and IGR IRES activities was still active in RRL when 2A^pro was expressed (Fig. 8D). However, unlike under conditions when 4E-BP was expressed, 2A^pro expression did not result in higher IRES activity.

FIG. 8.

IRES activities in vitro. (A) IRES translational activities in RRL. Dicistronic RNAs containing no IRES, CrPV IGR IRES, or the CrPV 5′UTR IRES were incubated in RRL at 30°C for 60 min. The first cistron, encoding the Renilla luciferase, measures scanning-mediated translation, and the second cistron, firefly luciferase, measures the IGR IRES-mediated translation. The ratios of firefly to Renilla luciferase, normalized to the ratio of the IGR IRES dicistronic RNA (100%), are shown. (B) 4E-BP expression stimulates CrPV 5′UTR and IGR IRES translation in RRL. Autoradiogram of 15% acrylamide PAGE showing translation products. Transcription and translation (TNT) RRLs were incubated alone (lanes 4 to 6, lanes 10 to 12, and lanes 16 to 18) or with a plasmid that expresses 4E-BP1 (lanes 1 to 3, lanes 7 to 9, and lanes 13 to 15) at 30°C for 30 min before dicistronic RNAs were added for 0, 30, or 60 min thereafter as indicated. Firefly and Renilla luciferase and 4E-BP1 are marked (arrows). (C and D) Quantitation of the fold luciferase activities. Shown are the fold Renilla luciferase (black), IGR IRES (white), and 5′UTR IRES (black) translation activities when 4E-BP (C) or 2A^pro (D) is expressed in RRL normalized to that without 4E-BP or 2A^pro expression which is given as 1. Note that Renilla luciferase activity is ∼10 to 20% of full activity when 4E-BP or 2A^pro is expressed. (E) IRES translation in RRL treated with NSC119889. Dicistronic RNAs containing the IGR IRES (left) or 5′UTR IRES (right) were incubated in RRL with 10, 25, or 50 μM NSC119889 at 30°C for 60 min. Shown is the percent luciferase activities in RRL normalized to the activity without drug treatment. Protein standards in kilodaltons are indicated in all gels. For all graphs, luciferase activities were measured by using the Promega dual luciferase assay kit. Shown in panel B is a representative gel from at least three independent experiments, and quantitations are averages are from at least three independent experiments ± the standard deviation.

Because eIF2α is phosphorylated during CrPV infection, we examined more closely whether the CrPV IRES was active under these conditions. The initial approach was to induce eIF2α phosphorylation in RRL; however, in our hands, this proved to be inconsistent and not reproducible. As an alternative, we impaired eIF2 function by using the compound, NSC119889, which leads to dissociation of eIF2 and the initiator Met-tRNA_i (63). As shown previously, addition of NSC119889 to RRL resulted in a significant reduction in scanning-mediated translation (first cistron) but not IGR IRES-mediated translation (second cistron) (Fig. 8E) (63). IGR IRES-mediated translation remained close to 100% of activity at 10 and 25 μM NSC119889 compared to that of the no-drug treatment. Similarly, 5′UTR IRES-mediated translation was also resistant to drug treatment compared to scanning-mediated translation. In summary, translation mediated by the CrPV 5′UTR and IGR IRES was relatively resistant to conditions that impair eIF2 and eIF4F formation.

DISCUSSION

The dicistrovirus RNA genome possesses two distinct IRES that direct translation of the nonstructural ORF1 and structural ORF2 proteins. Most notable, the structural proteins are produced in supramolar excess over the nonstructural proteins in CrPV-infected cells, suggesting that the IRES are differentially regulated. Although much is known about the biochemical properties of the IGR IRES that directs translation of ORF2, the mechanisms that lead to translational shutoff during CrPV infection has not been investigated. The data presented here demonstrate that CrPV infection in S2 cells results in the modification of at least two major translational control targets: the deIF4F and deIF2 complexes. Starting as early as 1 h p.i., deIF4G dissociates from the cap-binding protein deIF4E, which correlates with the initial shutoff of host translation (Fig. 4B). Beginning at 3 h p.i., deIF2α becomes phosphorylated, however, this modification appears to be dispensable for viral production (Fig. 4C) (7). Our data are consistent with a model that impairment of deIF4F contributes to the translational shutoff in CrPV-infected S2 cells, thereby promoting the synthesis of viral proteins.

The eIF4F complex is a common target in many virus infections. Here, we show that the components of the eIF4F complex, eIF4G, eIF4A, and eIF4E, and PABP remain intact (Fig. 4A), demonstrating that CrPV does not use this strategy to shutoff host translation. Moreover, eIF4E activity is not altered as shown by its association with the cap resin throughout the course of infection (Fig. 4B). The activation of 4E-BP has been observed under poliovirus, VSV, and EMCV infections, contributing to translational shutoff in cells infected with these viruses (7, 18, 21). Although hypophosphorylated 4E-BP is transiently upregulated early during CrPV infection which may explain how eIF4G is dissociated from eIF4E at 1 h p.i., our results indicate that 4E-BP is not the main factor responsible for the initial translational shutoff. First, 4E-BP associates with the cap resin in mock- and UV-irradiated CrPV-infected lysates, which do not result in host translational shutoff (Fig. 2 and 4B). Second, eIF4G still associates with the cap resin in these lysates (Fig. 4B). These results suggest that association of 4E-BP in CrPV-infected cells is not sufficient to dissociate eIF4G from eIF4E and likely plays a minor role in the shutoff of translation. Given that eIF4G remains dissociated from the cap resin at later times of infection, we propose that another factor may be preventing the deIF4G-deIF4E interaction. Other than 4E-BP, several proteins can interact with eIF4E including Drosophila Cup, which contains an eIF4E binding domain and has been shown to inhibit eIF4G recruitment (50). Cup does not bind to the cap resin during CrPV infection (J. Garrey and E. Jan, unpublished data). Current efforts are ongoing into the mechanism of deIF4E-deIF4G dissociation during CrPV infection.

Although deIF2α is phosphorylated, our results indicate that this modification is dispensable for host translational shutoff in CrPV-infected S2 cells. Suppression of deIF2α phosphorylation did not prevent host translation shutoff during CrPV infection, suggesting that other mechanisms, such as the dissociation of deIF4G and deIF4E, likely play a more significant role (Fig. 6A). Surprisingly, preventing eIF2α phosphorylation also did not have an effect on overall translation during thapsigargin or DTT treatments (Fig. 6A). This result is surprising since it has been shown that overexpressing GADD34 in mammalian cells prevents translational shutoff in DTT-treated cells (53). It is possible that in S2 cells the PERK-eIF2α signaling arm of the UPR plays a minor role in inhibiting translation. Supporting this, a previous report suggested that eIF2B activity is less sensitive to increases eIF2α phosphorylation in the insect cell line SF-9 (14). Whether this is a general phenomenon to all insect cells remains to be investigated.

The CrPV ORF2 proteins are expressed in supramolar excess over ORF1 (46, 81). This suggests that translation via the IGR IRES is more active than the 5′UTR IRES. Alternatively, it is possible that the timing of IRES translation may be regulated during CrPV infection. For example, the 5′UTR IRES may be active at earlier times of infection to produce the viral nonstructural proteins such as the RdRP and viral protease. Western blot analysis indicates that the rates of ORF1 and ORF2 expression are similar, suggesting that the IRES translation activities do not change during the course of infection. Thus, the differential expression of ORF1 and ORF2 is likely due to their distinct abilities to recruit the ribosome during infection.

Our results indicate that CrPV 5′UTR and IGR IRES are resistant to alterations in the translational machinery that mimic those that occur during CrPV infection. Specifically, impairing eIF4F formation by expressing 4E-BP in RRL stimulated translation of both IRES (Fig. 8B and C). This result correlates nicely with our findings that eIF4E and eIF4G are dissociated during CrPV infection and provides a mechanistic model by which viral IRES translation is induced when host translation is shutoff. Previously, it has been shown that 4E-BP expression in nuclease-treated RRL does not stimulate EMCV IRES translation (74). However, in untreated RRL or when competing mRNAs are present, the EMCV IRES, which competes with capped cellular mRNAs for eIF4G, can be stimulated when 4E-BP is expressed (74). In contrast, we demonstrate here that CrPV 5′UTR and IGR IRES translation is stimulated in nuclease-treated RRL, suggesting that CrPV IRES translation is inhibited when the eIF4F complex is intact and enhanced when specific translation factors are impaired (Fig. 8). Our data support this model and are consistent with previous findings that CrPV IRES translation is stimulated when overall translation is reduced, presumably under conditions that increase the available pool of ribosomes for IRES translation (17, 80). Indeed, a recent report showed that compromising several initiation factors can stimulate IGR IRES translation in yeast (12). However, impairing eIF4F by expressing 2A^pro did not stimulate CrPV IRES translation in RRL (Fig. 8D). This suggests that a decrease in overall translation may not be the only prerequisite for CrPV IRES stimulation but instead may involve specific modifications of the translational machinery. It will also be of interest to see whether the CrPV IRES are stimulated when competing mRNAs are present, a situation similar to physiological conditions during infection.

It has been well established that the IGR IRES can mediate translation under stresses that induce eIF2α phosphorylation or impair eIF2 function (Fig. 8E) (17, 40, 62, 63, 80). Our findings are in agreement with these results and suggest that IGR IRES-mediated translation of the ORF2 structural proteins is resistant during CrPV infection when eIF2α phosphorylation is induced. Although the induction of eIF2α phosphorylation may play a minor role in host translational shutoff, eIF2α phosphorylation may in part promote IGR IRES-mediated translation of ORF2 and ensure that high levels of structural proteins are produced for viral assembly. However, this may not be the full story since there is a significant amount of ORF2 protein still expressed when eIF2α phosphorylation is suppressed, indicating that IGR IRES-mediated translation is relatively active. This is consistent with the finding that the IGR IRES is active under conditions when initiation factors other than eIF2 are compromised (12). Therefore, it is probable that multiple mechanisms contribute to the release of ribosomes from host mRNAs, from which the IGR IRES can efficiently recruit from. Finally, it may not be too surprising that suppressing eIF2α phosphorylation did not have an effect on CrPV titers (Fig. 7), given that the IGR IRES is active enough to produce sufficient structural proteins for viral assembly.

An interesting finding is that the timing of deIF2α phosphorylation during CrPV infection may be important. Premature phosphorylation of eIF2α by thapsigargin treatment resulted in a decrease in viral protein and RNA levels (Fig. 5B). These effects could be attributed to translational inhibition of ORF1, which encode key proteins for CrPV maturation and assembly including the RdRP and viral 3C protease. ORF1 translation is mediated by the 5′UTR IRES (81). Currently, the factors required for CrPV 5′UTR IRES translation are not known. However, it has been shown that the initiation factors eIF1, eIF3, and eIF2 are necessary for translation by the 5′UTR IRES of the related dicistrovirus, RhPV (77). It will be interesting to determine whether CrPV 5′UTR IRES translation requires similar factors. The interpretation described above is based on the assumption that the major effect of thapsigargin treatment is phosphorylation of eIF2α. However, thapsigargin affects other processes in the cell, including activation of ER-stress response pathway and release of calcium stores from the ER. Thus, we cannot rule out the inhibition of CrPV infection may be due to these other effects by thapsigargin treatment.

An outstanding question is how eIF2α phosphorylation is induced in CrPV-infected cells? During some virus infections, the translation of viral proteins may pose as a burden and result in a load on the ER, thus triggering PERK activation. Members of the flaviviruses and HCMV cause ER stress in infected cells, activating the UPR (27, 73, 75). Besides the large translation output during infection, viral replication complexes that are associated with the ER can also add to the burden, thus magnifying ER stress and leading to downstream UPR events such as host translational shutoff. In this report, we demonstrate that CrPV infection does not lead to splicing of dXbp-1 mRNA, which is an indicator of ER stress (Fig. 5C). This suggests that ER stress is not induced during CrPV infection and that, by extension, the ER-stress-inducible eIF2α kinase dPERK is not activated. However, without direct analysis of dPERK, these results still leave the possibility that dPERK may still be induced during CrPV infection. To counter this point, we show that both dPERK and dIRE1 pathways of the UPR are intact since thapsigargin treatment of CrPV-infected cells still induced eIF2α phosphorylation and splicing of dXbp-1 mRNA (Fig. 5C). Thus, we argue that CrPV infection in S2 cells does not induce ER stress or activate dPERK. Because D. melanogaster contains only two known eIF2α kinases, dPERK and dGCN2, these results suggest that dGCN2 may be responsible for phosphorylation of eIF2α during CrPV infection. It has been reported that GCN2 is activated in Sindbis virus-infected cells (2). Although GCN2 is activated in response to amino acid starvation, that study proposed that the GCN2 could also act as a host antiviral factor that senses SV RNA to inhibit translation (2). It is of considerable interest to determine whether dGCN2 has a role in CrPV infection.

Several viruses have evolved mechanisms to counteract eIF2α phosphorylation. Most strategies either inhibit eIF2α kinases or promote dephosphorylation. For instance, the vaccinia virus K3L protein acts as a pseudosubstrate that binds PKR and prevents it from binding to eIF2. In herpes simplex virus-infected cells, the viral protein γ34.5 actively dephosphorylates eIF2α (24). Unlike these viruses, the CrPV is part of a growing list of viruses that induce eIF2α phosphorylation during infection. Infection by VSV and members of the Alphavirus family results in eIF2α phosphorylation and, similar to CrPV, these viruses have evolved mechanisms to counteract this translational block (2, 6, 43, 65, 79). In Sindbis virus-infected cells, the viral RNA utilizes the initiation factor eIF2Α instead of eIF2 to recruit the initiator Met-tRNA (79). In another example, HCV not only produces viral proteins that counteract eIF2α phosphorylation but also initiates translation via its IRES that is resistant to eIF2α phosphorylation (62, 63). Thus, it appears that these viruses, like CrPV, have adapted multiple, alternative strategies to bypass the inhibitory effects on translation by eIF2α phosphorylation.

Translational shutoff is a common strategy utilized by viruses. Inhibition of translation results in the release of ribosomes from host mRNAs and effectively increases the pool of ribosomes for viral translation. Also, translational shutoff counteracts and inhibits host antiviral responses. In the present study, we have shown that the infection of S2 cells by CrPV, a member of the Dicistroviridae family, induces host translational shutoff by a strategy that impairs eIF4F complex formation and induces eIF2α phosphorylation. We propose that these alterations in turn promote translation mediated by the CrPV 5′UTR and IGR IRES to ensure production of viral proteins. Although we have focused on these specific alterations during CrPV infection, it is likely that other translational control mechanisms also contribute to host translation shutoff. Finally, it remains to be investigated whether the translational inhibition mechanisms observed during CrPV infection in S2 cells are utilized by other dicistroviruses. It is known that the related dicistrovirus, DCV, does not lead to a rapid shutoff of host translation (47). This suggests that within the Dicistroviridae family distinct viral strategies may be utilized to manipulate the host translational machinery for ribosome recruitment.