Novel Function of African Swine Fever Virus pE66L in Inhibition of Host Translation by the PKR/eIF2α Pathway

African swine fever virus (ASFV) is a member of the nucleocytoplasmic large DNA virus superfamily that predominantly replicates in the cytoplasm of infected cells. The ASFV double-stranded DNA genome varies in length from approximately 170 to 193 kbp depending on the isolate and contains between 150 and 167 open reading frames (ORFs), of which half the encoded proteins have not been explored.

ABSTRACT

African swine fever virus (ASFV) is one of the most contagious and lethal viruses infecting pigs. This virus is endemic in many countries and has very recently spread to China, but no licensed vaccines or treatments are currently available. Despite extensive research, the basic question remains of how ASFV-encoded proteins inhibit host translation. Here, we examined how ASFV interferes with host translation and optimizes viral gene expression. We found that 14 ASFV proteins inhibited Renilla luciferase (Rluc) activity greater than 5-fold, and the protein with the strongest inhibitory effect was pE66L, which was not previously reported. Combined with bioinformatic analysis and biochemical experiments, we determined that the transmembrane (TM) domain (amino acids 13 to 34) of pE66L was required for the inhibition of host gene expression. Notably, we constructed a recombinant plasmid with the TM domain linked to enhanced green fluorescent protein (EGFP) and further demonstrated that this domain broadly inhibited protein synthesis. Confocal and biochemical analyses indicated the TM domain might help proteins locate to the endoplasmic reticulum (ER) to suppress translation though the PKR/eIF2α pathway. Deletion of the E66L gene had little effect on virus replication in macrophages, but significantly recovered host gene expression. Taken together, our findings complement studies on the host translation of ASFV proteins and suggest that ASFV pE66L induces host translation shutoff, which is dependent on activation of the PKR/eIF2α pathway.

IMPORTANCE African swine fever virus (ASFV) is a member of the nucleocytoplasmic large DNA virus superfamily that predominantly replicates in the cytoplasm of infected cells. The ASFV double-stranded DNA genome varies in length from approximately 170 to 193 kbp depending on the isolate and contains between 150 and 167 open reading frames (ORFs), of which half the encoded proteins have not been explored. Our study showed that 14 proteins had an obvious inhibitory effect on Renilla luciferase (Rluc) protein synthesis, with pE66L showing the most significant effect. Furthermore, the transmembrane (TM) domain of pE66L broadly inhibited host protein synthesis in a PKR/eIF2α pathway-dependent manner. Loss of pE66L during ASFV infection had little effect on virus replication, but significantly recovered host protein synthetic. Based on the above results, our findings expand our view of ASFV in determining the fate of host-pathogen interactions.

INTRODUCTION

African swine fever (ASF) is an acute and highly contagious disease in both domestic pigs and wild boars, in which it causes pathological characteristics of high fever, hemorrhage, ataxia, and severe depression (1). Since ASF was first discovered in Kenya in the 1920s, it rapidly spread to many areas in sub-Saharan Africa, the Caucasus, the Russian Federation, and Eastern Europe (2). This disease was of particular concern in China, the largest pork producer in the world, where it was first reported in 2018 (3). After a short period of one year, the disease killed more than 1,192,000 pigs in China, with an estimated cost of over 300 million US dollars, and threatened the pig industry and food security (4, 5). Unfortunately, no available African swine fever virus (ASFV) vaccine or other useful treatments against this complicated virus have been developed.

ASFV is the only member of the Asfarviridae family and is categorized as a nucleocytoplasmic large DNA virus (NCLDV) (6, 7). The ASFV genome is a linear double-stranded DNA (dsDNA) molecule of 170 to 190 kbp that contains 151 to 167 open reading frames (ORFs), which are primarily divided into structural proteins, viral DNA replication proteins, and immune evasion proteins (8–11). Notably, the structural proteins are generally composed of envelope proteins, capsid proteins, nucleocapsid proteins, and DNA-binding proteins. The envelope protein is the main structural protein of viral particles, and is closely associated with host cell tropism, pathogenicity, and immunogenicity (12). The proteins of pB646L, pB438L, and pE120R are located on the viral capsule membrane, which ensures viral genome integrity by protecting the enclosed nucleic acids and participating in viral infection (13–17). The proteins of pCP2475L and pCP530R are nucleocapsid proteins that may be assembled together into nucleosome-like structures (18, 19). Immunoelectron microscopy showed that two DNA-binding proteins, pK78R and pA104R, were located in the nucleoids of mature virions and might play a role in the assembly of viral nucleoids (12, 20). In addition to the structural proteins, ASFV virus particles contain many genomic copies, which play roles in viral DNA replication and immune evasion. The proteins of pMGF360 and pMGF505/530 determine cytotropism, which is closely related to viral replication in macrophages (21, 22). A recent study showed that pDP96R mediates the cGAS-STING pathway via inhibition of TBK1 and IKKβ activation (23). In addition, the proteins of pA224L, pA179L, and pEP153R inhibit the premature apoptosis of host cells to promote replication (24–26).

Despite extensive research, effective antiviral therapies or vaccines are lacking, and the detailed molecular mechanisms underlying ASFV inhibition of host translation remain unclear. There is very little research on this topic, but some studies have shown that pA238L can inhibit the activation of the host nuclear transcription factors NF-kb and NFAT to regulate host gene expression, and that pDP71L can promote dephosphorylation of eukaryotic translation initiation factor 2 alpha (eIF2α) to prevent protein synthesis inhibition (27, 28). The ASFV g5R protein (g5Rp) is a viral decapping enzyme that is involved in the regulation of mRNA metabolism (29, 30). However, other ASFV proteins that inhibit the synthesis of host genes are not well characterized. Our study found that 14 proteins significantly inhibited Renilla luciferase (Rluc) gene expression, and an unknown functional protein, pE66L, had the most significant inhibitory effect. We also determined that the transmembrane (TM) domain of pE66L was required for the inhibition of host gene expression. Interestingly, we added this region to enhanced green fluorescent protein (EGFP) and further demonstrated that this domain exerted inhibitory effects. This region may help proteins locate to the endoplasmic reticulum (ER) to induce translational suppression, which is dependent on the phosphorylation of eIF2α and protein kinase R (PKR). Using homologous recombination technology to generate specific knockouts, we demonstrated that the loss of pE66L during ASFV infection had little effect on virus replication. Taken together, these results complement the existing knowledge on the effects of ASFV proteins on host translation and suggest that ASFV pE66L induces host translation shutoff, which is dependent on activation of the PKR/eIF2α pathway.

RESULTS

Preliminary screening of ASFV genome-mediated inhibition of host gene expression by Rluc assays.

To identify the ASFV genes that inhibit host gene translation, we cloned the ASFV genes (strain ASFV-SY18) into a PCAGGS vector containing a hemagglutinin (HA) tag. Our study successfully constructed 148 plasmids with ASFV genes involved in replication, immune evasion, structure, intracellular virulence, and other unknown functions (Fig. 1A). Unfortunately, 15 ASFV genes were not able to be cloned into a vector (F1055L, EP1242L, EP364R, M1249L, C717R, C962R, B602L, G1340L, G1211R, CP2475L, NP1450L, DP79L, P1192R, QP383R, and DP60R). Based on our previous studies (31, 32), we chose the Rluc gene as a model to preliminarily screen ASFV genes for the ability to inhibit host gene expression. The results showed that many ASFV genes affected Rluc gene expression to different degrees (Fig. 1B), which suggested that the biological effects displayed in ASFV infection might be the result of comprehensive regulation.

FIG 1.

Preliminary screening of African swine fever virus (ASFV) genes for inhibition of Renilla luciferase (Rluc) gene expression. (A) Genome organization of the ASFV genome. The organization of open reading frames (ORFs) on the genome of the virulent ASFV-SY18 is shown. ORFs are shown as arrows to indicate their size and direction. (B) Determination of the ASFV genes that inhibit Rluc gene expression. HEK-293T cells were cotransfected with pRL-SV40 and ASFV genes. At 24 h posttransfection, the cells were lysed and subjected to Rluc assays. The error bars show the standard deviation (SD) of the results from three independent experiments.

Identification of ASFV proteins that inhibit Rluc activity by more than 5-fold.

Because we aimed to identify the ASFV genes that inhibited host protein translation, we examined the genes that were responsible for this function. The screening results described above identified that many ASFV genes inhibited the expression of Rluc, but some genes showed low efficacy. For example, the inhibitory effects of MGF_110-13L, MGF_110-12L, MGF_110-14L, and MGF_110-11L were less than 10%. However, some genes (MGF_505-3R, MGF_505-2R, and MGF_360-14L) showed inhibitory effects of approximately 50% (Fig. 1B).

To identify the ASFV genes that play a dominant role in host protein translation, we established the criterion that the inhibitory effect of the sample gene compared to that of the mock control must be greater than 5-fold. Based on this criterion, we focused on 14 genes: MGF_360-9L, MGF_360-10L, MGF_360-11L, A238L, EP152R, EP402R, C122R, C257L, CP312R, H171R, QP509L, E66L, I243L, and I329L (Fig. 2A and C). Other functions of these genes were reported previously (Fig. 2B). Deletion of the MGF_360-9L, MGF_360-10L, and MGF_360-11L genes reduced the virulence of ASFV (22). The protein encoded by EP402R inhibits lymphocyte proliferation because it interacts with the cellular AP-1 protein to participate in intracellular viral transport (33, 34). The A238L protein regulates inflammation via inhibition of NF-κb activation (27). The I329R protein inhibits the antiviral effect of type I interferon (IFN) expression (35, 36). The proteins of EP152R, C122R, C257L, CP312R, and H171R are structural proteins involved in the stabilization of virus particles (9). The QP509L and I243L proteins can regulate viral replication upon the entry of ASFV into host cells (37, 38).Our findings expand our knowledge of the biological functions of these genes and help elucidate their importance in viral reproduction. Notably, pE66L was not studied previously, and it showed the strongest inhibitory effect of the selected genes, with greater than 15-fold inhibition. These results supported further investigation into the mechanism of pE66L in mediating host gene translation.

FIG 2.

Identification of the ASFV proteins that inhibit Rluc activity at levels greater than 5-fold. (A and B) MGF_360-9L, MGF_360-10L, MGF_360-11L, A238L, EP152R, EP402R, C122R, C257L, CP312R, H171R, QP509L, E66L, I243L, and I329L, all of which inhibited the Rluc gene expression at levels greater than 5-fold, were selected from the screen and further characterized. (C) The 14 selected ASFV proteins were expressed in HEK-293T cells. At 24 h posttransfection, the cells were fixed, incubated with a monoclonal antibody against the HA protein, and visualized by indirect immunofluorescence (IF). Original magnification 200× (scale bars 100 μm).

ASFV pE66L is a membrane protein.

Because of the unknown function of pE66L, we first analyzed its primary structure. The E66L protein was approximately 6 kDa, which raised the question of how this small protein might inhibit host gene expression. We tried to predict its tertiary structure using SwissModel, but the complete structure could not be modeled, most likely because no homologous structure has been determined, which increased the difficulty of pE66L research. Surprisingly, we found that pE66L was a membrane protein using bioinformatics analysis. We added enhance green fluorescent protein (EGFP) to pE66L and performed the fluorescence confocal assay. The results showed that pE66L could indeed affect the positioning of EGFP and made it surround the nucleus. Membrane proteins must adopt their proper topologies within biological membranes, and the topology of pE66L was predicted using InterProScan and TMpred. The E66L protein encodes a highly conserved single-pass type I TM protein that consists of 50 amino acids with a 12-amino acid intracellular N-terminal domain, a 22-amino acid TM domain (amino acid positions 13 to 34), and a 16-amino acid extracellular C-terminal domain. The TM domain anchors the protein to the membrane via its hydrophobic surface, and it exhibits a high frequency of hydrophobic amino acids (e.g., isoleucine and leucine), which are necessary for its membrane anchorage.

TM domain (amino acids 13 to 34) is critical for ASFV pE66L inhibition of Rluc gene expression.

We determined which domain in ASFV pE66L was essential for its inhibitory activity. Because pE66L is a membrane protein that includes an extramembrane domain, a TM domain, and an intramembrane domain, we used random combinations to design a series of truncated pE66L plasmids: E66L(1–12 aa), E66L(13–34 aa), E66L(35–50 aa), E66L(1–34 aa), and E66L(13–50 aa) (Fig. 3A). Based on the plasmid expression and virus susceptibility in various cells, we selected human embryonic kidney (HEK-293T) and porcine ileum epithelial (IPI-2I) cells for further functional examination using the Rluc assay. Comparisons of the effects of the extramembrane domain (35 to 50 amino acids [aa]), the TM domain (13 to 34 aa), and the intramembrane domain (1 to 12 aa) on the expression of reporter gene found the TM domain significantly reduced the inhibition of Rluc gene expression in both cell lines and exhibited no significant difference from the effects of the full-length E66L protein (Fig. 3B). However, pE66L(1–12) and pE66L(35–50) had no inhibitory effect on Rluc gene expression compared with that of the mock control because these proteins were not expressed in cells (Fig. 3B and C). To further explore the functions of pE66L(1–12) and pE66L(35–50), we linked them into EGFP and then confirmed that EGFP-E66L(1–12) and EGFP-E66L(35–50) had no effect on reporter gene synthesis (Fig. 3D and E). Furthermore, we found that the abilities of pE66L(1–34) and pE66L(13–50) to suppress the reporter gene were indistinguishable from that of the wild-type protein (Fig. 3B). Once the TM domain (13 to 34 aa) was retained, it exerted an obvious ability to inhibit host protein synthesis, which meant the TM domain (13 to 34 aa) was critical for the ASFV pE66L-mediated inhibition of Rluc gene expression.

FIG 3.

The TM domain (amino acids 13 to 34) is critical for the ASFV E66L inhibition of Rluc gene expression. (A) Based on the genomic structure of E66L, we constructed a series of truncated plasmids. The recombinant plasmids were as follows: E66L(1–12), E66L(13–34), E66L(35–50), E66L(1–34), and E66L(13–50). (B) HEK-293 T and IPI-2I cells were cotransfected with pRL-SV40 (0.1 μg) and one of the above plasmids (0.5 μg). After 24 h, the cell lysates were prepared and subjected to luciferase assays. Data are presented as the mean ± SD of the results from three independent experiments; ns, not significant; ***, P < 0.001. (C) The above-described plasmids were transfected into HEK-293T and IPI-2I cells. At 24 h, the cells were fixed and IF was performed with a monoclonal antibody against the HA protein. Original magnification 200× (scale bars 100 μm). (D) EGFP was linked to the HA tag and then E66L(1–12) or E66L(35–50), and the resulting construct was named EGFP-E66L(1–12) or EGFP-E66L(35–50). (E) HEK-293 T cells were cotransfected with pRL-SV40 (0.1 μg) and one of the above plasmids (0.5 μg). After 24 h, the cell lysates were prepared and subjected to luciferase assays. Data are presented as the mean ± SD of the results from three independent experiments; ns, not significant.

Broad inhibition of host protein synthesis by the TM domain.

To further examine whether ASFV pE66L inhibited overall gene expression in HEK-293T or IPI-2I cells, we examined the host proteins using ribopuromycylation assays. The data showed that both pE66L and pE66L(13–34) inhibited host protein synthesis in both cell lines (Fig. 4A and B). We also examined whether the TM domain (13 to 34 aa) played an important role in inhibiting host protein synthesis in the context of other proteins. To evaluate the role of the TM domain in inhibiting host gene expression, we linked EGFP to an HA tag and E66L(13–34 aa) (Fig. 4C). The recombinant plasmid, named EGFP-E66L(13–34), was verified via sequencing of the altered regions. Notably, the Rluc and Western blot assays showed that although the expression of EGFP in the cells was much higher than that of EGFP-E66L(13–34), its inhibitory effect on the Rluc gene was significantly lower than that of EGFP-E66L(13–34) (Fig. 4D and E). Taken together, these results indicated that the TM domain had a broad-spectrum inhibitory effect on host protein synthesis.

FIG 4.

Broad inhibition of host protein synthesis by the TM domain. (A and B) HEK-293T and IPI-2I cells were transfected with one of the following plasmids: PCAGGS, E66L, E66L(13–34), E66L (1–12), or E66L(35–50) (2.5 μg). The cells were pulsed with 3 μM puromycin for 1 h at 24 h posttransfection and then subjected to Western blot analysis or an IFA. Original magnification 200× (scale bars 100 μm). (C) EGFP was linked to the HA tag and E66L(13–34), and the resulting construct was named EGFP-E66L(13–34). (D and E) HEK-293T and IPI-2I cells were cotransfected with pRL-SV40 encoding the Rluc reporter gene downstream of the SV40 promoter and one of the following plasmids: PCAGGS, EGFP-HA, or EGFP-E66L(13–34)-HA. At 24 h posttransfection, the cells were lysed and subjected to Western blot analysis. The error bars show the SD of the results from three independent experiments; **, P < 0.01; ***, P < 0.001.

ASFV pE66L promotes cell retention in the G0/G1 phase.

To assess whether pE66L and pE66L(13–34) manipulated the cell cycle of host cells, the cell cycle distribution was analyzed in IPI-2I cells using flow cytometry at 24 h posttransfection. An obvious increase in the G0/G1 phase was observed using ModFit analysis, with an increase from 45.25% for PCAGGS-transfected to 51.15% for E66L-transfected and 48.83% for E66L(13–34 aa)-transfected. A decreased proportion in the G2/M phase was observed when pE66L and pE66L(13–34) were present, with a decrease from 11.88% for PCAGGS-transfected to 9.45% for E66L-transfected and 8.37% for E66L(13–34 aa)-transfected (Fig. 5B). These data suggested that pE66L and pE66L(13–34) induced the accumulation of G0/G1 phase and the decrease of G2/M phase. To determine whether the G0/G1 phase arrest was exclusive to the IPI-2I cell line, HEK-293T cells were selected for further screening with pE66L and pE66L(13–34). HEK-293T cells in G0/G1 phase were increased from 45.06% for PCAGGS-transfected to 51.48% for E66L-transfected and 51.82% for E66L(13–34 aa)-transfected (Fig. 5A) at 24 h posttransfection. These results indicated that the effects of pE66L and pE66L(13–34) on G0/G1 phase arrest were broadly applicable and demonstrated that the TM domain (13 to 34 aa) was the critical region for pE66L function. This phenomenon is also widely observed in other viruses and viral proteins. Influenza A virus NS1 induces G0/G1 cell cycle arrest via inhibition of the expression of the RhoA protein (39). Newcastle disease virus induces G0/G1 cell cycle arrest in asynchronously growing cells (40). These findings suggest that pE66L promotes cell retention in the G0/G1 phase.

FIG 5.

ASFV pE66L promotes cell retention in the G0/G1 phase. (A and B) HEK-293T or IPI-2I cells were transfected with an empty vector, E66L vector, or E66L(13–34 aa) vector for 24 h. The cells were fixed, stained with propidium iodide (PI) in the presence of RNase A, and analyzed using flow cytometry to determine the DNA content and the cell population distributions in the various cell cycle phases. Error bars show the SD of the results from three independent experiments. The asterisks indicate the statistical significance calculated using Student's t test; ns, not significant; *, P < 0.05; **, P < 0.01; ***, P < 0.001.

ASFV pE66L inhibits host protein translation through the PKR/eIF2α signaling pathway.

To investigate which step of the viral cycle was affected by pE66L, RNA analysis in the presence of cycloheximide (CHX) was performed using quantitative real-time PCR (qRT-PCR). The results showed that pE66L had no effect on the degradation of Rluc mRNA, which indicated that pE66L did not affect gene transcription (Fig. 6A). Therefore, we determined whether pE66L affected host protein synthesis at the translational level. Through an ER isolation kit, we found that pE66L located in the ER (Fig. 6B). To further determine the subcellular localization of pE66L and pE66L(13–34), we used a plasmid with pDsRed2ER (which showed red fluorescence in the ER) for labeling. The fluorescence confocal assay confirmed that the location of pE66L and pE66L(13–34) in cells had no difference with pDsRed2ER (Fig. 6C and D). Taken together, the results indicated that pE66L and pE66L(13–34) inhibited protein synthesis by affecting host protein translation.

FIG 6.

ASFV pE66L inhibits host protein synthesis at the translation stage. (A) The effect of overexpressed pE66L on Rluc mRNA production was determined using qRT-PCR when the cells were treated with or without cycloheximide (CHX) at 12 and 24 h. Error bars show the SD of the results from three independent experiments; ns, not significant. (B) HEK-293T cells were transfected with an empty vector or E66L vector for 24 h and then ER were isolated. Proteins were separated by SDS-PAGE and evaluated by Western blotting. (C and D) Subcellular localization of E66L and E66L(13–34). HEK-293T and IPI-2I cells were seeded on slides in 24-well plates and cotransfected with pDsRed2-ER and one of the following plasmids: PCAGGS, PCAGGS-E66L-HA, or PCAGGS-E66L(13–34)-HA (none of which encoded a protein), E66L, or E66L(13–34). At 24 h posttransfection, the cells were fixed and permeabilized with Triton X-100. The cells were then incubated with a mouse anti-HA MAb for 1 h, followed by incubation with an Alexa Fluor 488-conjugated goat anti-mouse (green) secondary antibody to visualize E66L or E66L(13–34). The ER protein was fused with an ER targeting sequence (Clontech) and directly visualized (red). Nuclei (blue) were stained with DAPI. The images were collected using a Zeiss LSM-510META confocal laser scanning microscope and processed with the LSM image browser (Zeiss). The white arrows indicate the cosubcellular localization between pE66L or pE66L(13–34) and pDsRed2-ER in the cells. Scale bars 5 μm.

We examined the specific mechanism by which pE66L and pE66L(13–34) regulated translation. The phosphorylation of eIF2α is a key mechanism of translational control (41). This event is vital to the regulation of global protein levels, and is essential for the maintenance of cellular homeostasis. PKR is involved in cell cycle regulation, and the highest PKR activity is observed in the early stages of G0/G1 (42). Moreover, pE66L and pE66L(13–34) promoted cell retention in the G0/G1 phase. Given these findings, we examined whether pE66L colocalized with the PKR/eIF2α pathway using fluorescence confocal and immunoprecipitation assays. Interestingly, the results showed that pE66L could interact with PKR to regulate the PKR/eIF2α pathway (Fig. 7A to C). To test whether PKR/eIF2α plays a role in pE66L-induced translational arrest, we measured the PKR and eIF2α phosphorylation levels in HEK-293T cells overexpressing E66L and found that pE66L could increase PKR and eIF2α phosphorylation in a dose-dependent manner (Fig. 7D). Meanwhile, we examined the effect in other cell types and saw that E66L(13–34) could also strongly increase PKR and eIF2α phosphorylation in different cell lines (Fig. 7E to H). These results support the conclusion that both E66L and E66L(13–34) trigger PKR/eIF2α phosphorylation to induce translational arrest.

FIG 7.

ASFV pE66L inhibits host protein translation through the PKR/eIF2α signaling pathway. (A and B) PKR and eIF2α colocalize with pE66L. HEK-293T cells were transfected with the plasmid of E66L and analyzed at 24 h posttransfection by confocal microscopy using antibodies against HA (green), PKR (red) (A), or eIF2α (red) (B). Individual channels and merged images are shown. The white arrows indicate the cosubcellular localization between PKR or eIF2α and pE66L in the cells. Scale bars 5 μm. (C) The immunoprecipitated proteins (PKR and eIF2α) and pE66L were examined by Western blotting using anti-PKR, anti-eIF2α, and anti-HA antibody. The expression of GAPDH was detected with an anti-GAPDH monoclonal Ab to confirm equal protein loading. (D) HEK-293T cells were transfected with different doses of the E66L plasmid (0 to 2.0 μg) for 24 h. The cells were subjected to Western blot analysis (left) and immunoblotted with antibodies against phosphorylated PKR (p-PKR), PKR, phosphorylated eIF2α (p-eIF2α), eIF2α, and GAPDH. The grayscale values of the protein bands were analyzed by ImageJ (right). Data are presented as the means ± SD of three independent experiments; ns, not significant; *, P < 0.05; ***, P < 0.001. (E and F) HEK-293T (E) or IPI-2I (F) cell lysates were harvested at 24 h posttransfection, and the proteins were separated using Western blot analysis (left) and immunoblotted with antibodies against p-PKR, PKR, p-eIF2α, eIF2α, and GAPDH. The gray scale values of the protein bands were analyzed using ImageJ (right). Data are presented as the means ± SD of three independent experiments; ns, not significant; *, P < 0.05; **, P < 0.01; ***, P < 0.001. (G and H) Expression of E66L and E66L(13–34) in the cells of panels (E) and (F). At 24 h posttransfection, the cells were fixed and incubated with a monoclonal antibody against the HA protein and then visualized by IF. Original magnification 200× (scale bars 100 μm).

Characterization of the ASFV pE66L deletion mutant.

Fragments of the left and right flanking regions of the E66L gene from ASFV-SY18, with sizes of 1,202 bp and 1,198 bp, respectively, were subcloned into a plasmid upstream and downstream of the EGFP gene under the control of a p72 promoter (Fig. 8A). Following transfection of this plasmid into porcine bone macrophages infected with virulent ASFV-SY18, cells infected with recombinant viruses were identified by expression of the EGFP gene and purified from wild-type virus. The expected deletion was confirmed using PCR analysis and primers that amplified either the E66L gene or the EGFP gene and by nucleotide sequencing of fragments from the sites of insertion. The ASFV mutation, which was subsequently named ASFVΔE66L, was confirmed using immunofluorescence assay (IFA) (Fig. 8B). After the desired mutant was successfully prepared, we tested whether the knockout of the indicated gene influenced the replication of this mutant in vitro. Porcine alveolar macrophage (PAM) cells were infected with the ASFVΔE66L virus or the parental ASFV-SY18 virus at a multiplicity of infection (MOI) of 0.1 to determine whether deletion of the E66L gene affected the ability of the virus to replicate in these cells in culture. At different times postinfection (0, 6, 12, 24, 36, 48, and 72 h), cell culture supernatants were harvested and the virus in these supernatants was titrated (Fig. 8C). Both viruses reached a plateau of approximately 10⁸ 50% tissue culture infective doses (TCID₅₀)/ml at 72 h postinfection. The viral growth and peak titers of ASFVΔE66L in susceptible cells showed no significant difference from those of the wild-type virus, indicating that the E66L gene was not critical for viral growth in cell culture.

FIG 8.

Effects of ASFVΔE66L at the cellular level. (A) Schematic representation of the ASFVΔE66L genome organization. (B) Detection of ASFVΔE66L using immunofluorescence assay (IFA). (C) Replication kinetics of knockout viruses. PAM cells were infected with wild-type ASFV or the knockout mutant ASFVΔE66L at an MOI of 0.1. The infected cells were collected at 0, 6, 12, 24, 48, and 72 h postinfection. The viral titers were recorded at different times as the TCID₅₀. The results represent the means ± SD of three independent experiments.

We performed cell proliferation analyses on ASFV- or ASFVΔE66L-infected PAM cells. The results showed that ASFVΔE66L induced more cells into the S phases of the cell cycle at the expense of the G0/G1 phases (Fig. 9A). We demonstrated that ASFV significantly inhibited host protein synthesis, which was consistent with a previous report (43). Notably, loss of the E66L gene significantly recovered host gene expression at the viral level (Fig. 9B). We compared the differential expression of RNAs in PAM cells infected with parental and ASFVΔE66L using DESeq (Fig. 9C), and verified the relevance of the sequencing results using qRT-PCR. The results showed that many genes (IFN-ALPHAOMEGA, IFNB1, TNF, MAP3K8, CCNL1, CCNT2, and so on) could be upregulated when ASFV lacked pE66L (Fig. 9D). Taken together, these results show that pE66L is an indispensable element for ASFV inhibition of host protein translation.

FIG 9.

In vitro analysis of ASFVΔE66L. (A) ASFV or ASFVΔE66L infection induced the subversion of cell cycle in asynchronously growing cells. PAM cells were infected with ASFV or ASFVΔE66L at an MOI of 0.1. At 48 h, cells were collected and stained with PI for cell cycle analyses using flow cytometry. The data are from one of three experiments. The histograms were analyzed to determine the percentage of cells in each phase of the cell cycle. The results are shown as mean ± SD of three independent experiments; ns, not significant; **, P < 0.01. (B) Cellular protein synthesis during ASFV and ASFVΔE66L infection. Cultures of PAM cells (1 × 10⁶) were mock infected (Mock), infected with ASFV (0.1 MOI), or infected with ASFVΔE66L (0.1 MOI), and labeled at 24 h after infection with 3 μM puromycin for 1 h. Samples were analyzed using Western blot analysis. (C) Genes upregulated by infection with wild-type ASFV or the mutant ASFVΔE66L are displayed as a heat map. Colors represent the log₂ (normalized counts) values of the expression levels for the genes in the samples. (D) Quantitative real-time PCR analysis of selected RNAs (IFN-ALPHAOMEGA, IFNB1, TNF, MAP3K8, CCNL1, and CCNT2) in PAM cells. The results are shown as means ± SD of three independent experiments; *, P < 0.05; ***, P < 0.001.

DISCUSSION

Many studies demonstrate that viruses strictly depend on host cell translation for the production of new progeny, but infected cells also synthesize antiviral proteins to limit viral infection (44–46). Therefore, viruses must manipulate cellular mRNA translation, impair host protein synthesis, and liberate cellular resources to ensure and optimize their own replication and spread. RNA viruses rely on the host translational machinery for the efficient synthesis of their proteins. Dengue virus (DENV) and Zika virus (ZIKV) infections suppress host cell stress responses, and the synthesis of viral proteins remains efficient (47). At the level of translational initiation, selective suppression of host cell translation is mediated by the nonstructural protein NS1 of influenza A viruses (48, 49). Similarly, the DNA virus ASFV stimulates cap-dependent translation to increase the initiation of viral mRNA translation via activation of the eIF4F complex (43). However, the mechanisms by which ASFV interferes with host translation and optimizes viral gene expression are not clear.

ASFV is a large, cytoplasmic, double-stranded DNA virus that is successively wrapped by a thick protein core shell, an inner lipid membrane, an icosahedral protein capsid, and an outer lipid membrane (50). Similar to other nucleocytoplasmic large DNA viruses, ASFV encodes many proteins dedicated to virus assembly, DNA replication, and gene expression (51). The ASFV genome also encodes many proteins involved in the evasion of host defense, including type I interferons and proteins in cell death pathways (10). However, approximately half of ASFV genes lack any known or predicted function. A previous study reported that ASFV infection inhibited host protein synthesis (43). However, the viral genes involved in regulating host translation were not known. Our screening analysis herein identified many ASFV genes that were likely involved in this process. We identified 14 protein genes that had a strong regulatory effect: MGF_360-9L, MGF_360-10L, MGF_360-11L, A238L, EP152R, EP402R, C122R, C257L, CP312R, H171R, QP509L, E66L, I243L, and I329L. Some of these genes have been reported to have other functions in the virus life cycle. The MGF_360-9L, MGF_360-10L, and MGF_360-11L proteins are multigene family members that enhance virulence (22). The EP402R, A238L, and I329R proteins are involved in the immune response via which the virus evades its host (27, 33, 35). EP152R, C122R, C257L, CP312R, and H171R are structural proteins (9), and QP509L and I243L regulate viral replication (37, 38). Among the 148 genes screened, pE66L was the most effective at inhibiting Rluc gene expression, and it was not reported previously. Therefore, the role of pE66L in inhibiting host gene expression was further examined.

The genomic sequence contains the life plan of an organism, and determining the functions of many individual proteins with known sequences and structures is challenging. To further characterize pE66L, we examined its tertiary structure, but, unfortunately, the protein could not be modeled. To our surprise, we found that pE66L was a membrane protein. Using the ribopuromycylation and Rluc assays, we showed the TM domain of pE66L was indispensable for its inhibitory function. Furthermore, we linked this TM domain to an unrelated fragment (EGFP) and found that the recombinant plasmid inhibited protein synthesis. Similarly, that the TM domain is the critical functional domain has been demonstrated for many receptors. FcγRIIB, the only inhibitory IgG Fc receptor, suppresses the hyperactivation of immune cells via the TM domain (52). The TM domain mediates the tetramerization of α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors (53). Taken together, we conclude the TM domain of pE66L acts as a functional peptide in regulating the expression of host genes.

Our studies utilizing fluorescence confocal microscopy and an associated biochemical experiment showed that pE66L and pE66L(13–34) located in ER. Furthermore, we demonstrated that pE66L and pE66L(13–34) did not affect the transcription level of the reporter gene. Therefore, pE66L appeared to inhibit protein synthesis by regulating host translation. As has been well established, eIF2α is a key target for the antiviral response to regulate cellular and viral protein synthesis (41). PERK, GCN2, HRI, and PKR phosphorylate eIF2α in response to various environmental stress conditions (54, 55). Notably, our findings indicated increased proportions of cells in the G0/G1 phase of the cell cycle when pE66L was present. PKR activity was previously shown to be increased when cells were in the G0/G1 phase (42). Furthermore, pE66L was determined to interact with PKR by immunoprecipitation and confocal experiments. Thus, pE66L plays the function of translation suppression via activating the PKR. The ASFV DP71L protein binds the catalytic subunit of protein phosphatase 1 (PP1) and induces eIF2α dephosphorylation (28). However, deletion of this gene from the virulent ASFV strain Malawi LIL 20/1 had no effect on virulence, which suggested that the virus encoded other proteins able to compensate for the loss of the DP71L gene (56). Our study found that pE66L could activate the phosphorylation of eIF2α, which demonstrated that ASFV regulation of the host cell was a comprehensive and complex process. As a previous study on human herpesvirus 6A (HHV-6A) has shown (57), virus limits its own replication due to the inability to bypass the eIF2α phosphorylation. Although deletion of the E66L gene from wild-type virus significantly enhances the transcription of interferon and proinflammatory cytokine genes, the ASFV can escape the host immune response by targeting different intracellular signaling intermediates, thereby maintaining viral replication at a stable level. The MGF360-12L protein blocks NF-κB nuclear translocation (58). The A238L protein inhibits tumor necrosis factor alpha (TNF-α) expression by modulating NF-κB (59). The I329L protein inhibits Toll-like receptor 3 signaling pathway (60). The A179L protein binds to Bcl-2 proteins in preventing premature cell apoptosis (25). Similarly, although coronavirus nsp1 could significantly inhibit the expression of host genes in vitro, the deletion of nsp1 had no effect on the ability of the virus to replicate (61, 62).

In summary, 148 ASFV genes were preliminarily screened for their ability to inhibit host gene expression using the Rluc assay. We obtained the first evidence that pE66L inhibited host gene translation and found that the TM domain of pE66L was a broadly functional region involved in the inhibition of host gene expression via the PKR/eIF2α pathway. The loss of pE66L during ASFV infection had little effect on virus replication, but it significantly recovered host gene expression. Overall, our results help elucidate the molecular mechanism of ASFV in inhibiting host gene expression.

MATERIALS AND METHODS

Cells and viruses.

HEK-293T and IPI-2I cells were obtained from the China Center for Type Culture Collection (Wuhan, China) and cultured at 37°C with 5% CO₂ in Gibco Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS) (Invitrogen, USA), 100 U/ml penicillin, and 10 μg/ml streptomycin sulfates. PAM cells for ASFV infection were acquired from bronchoalveolar lavage and cultured under the same conditions described above. The ASFV strain ASFV-SY18 (GenBank accession number MH766894.1) was isolated from a piglet with severe diarrhea in China in 2018. ASFVΔE66L was derived from ASFV-SY18 via substitution of the E66L ORF from EGFP with the ASFV p72 late gene promoter.

Plasmid construction.

The genes of the ASFV strain ASFV-SY18 (GenBank accession number MH766894.1) were amplified and cloned into PCAGGS with an N-terminal HA tag. The forward and reverse primers contained the EcoRI and XhoI restriction sites, respectively. Truncated E66L was also cloned into PCAGGS. All constructed plasmids were confirmed by sequencing, and no unexpected mutations occurred.

Construction of the recombinant ASFVΔE66L.

Recombinant ASFVs were generated by homologous recombination between the parental ASFV genome and a recombination transfer vector following infection and transfection of swine macrophage cell cultures. The recombinant transfer vector (p72EGFPΔE66L) contained flanking genomic regions that included portions of E66L mapping to the left (1,202 kbp) and right (1,198 kbp) of the gene together with a reporter gene cassette containing the EGFP gene with the ASFV p72 late gene promoter p72EGFP. The recombinant transfer vector p72EGFPΔE66L was obtained via DNA synthesis (GENEWIZ). Macrophage cell cultures were infected with ASFV-SY18 and transfected with p72EGFPΔE66L. Recombinant viruses representing independent primary plaques were purified to homogeneity by successive rounds of plaque assay purification.

Virus inoculation and growth curve.

To obtain the multistep growth kinetics curves, we incubated PAM cells in 24-well plates with ASFV or ASFVΔE66L (MOI = 0.1). Whole-cell samples were collected at 0, 6, 12, 24, 48, or 72 h postinfection (hpi), followed by freezing and thawing three times and centrifugation at 3,000 rpm for 10 min to collect the supernatant. Viral titers were determined by a TCID₅₀ assay.

Purification of endoplasmic reticulum.

ER fractions were isolated following the Endoplasmic Reticulum isolation kit protocol (Sigma-Aldrich, cat. number ER0100). The purity of cell fractions was determined by immunoblotting against Bip/GRP78 (an ER marker).

Immunofluorescence assay and Western blot analysis.

For evaluation of plasmid transfection in the cells, cell samples were collected for treatment. Briefly, the cells were first gently washed twice with precooled phosphate-buffered saline (PBS) and then centrifuged at 4,000 rpm for 5 min. The supernatant was discarded, radioimmunoprecipitation assay (RIPA) lysis buffer (Beyotime) was added, and the samples were incubated at 4°C for 25 min while rotating. The extracts were prepared in sodium salt-polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer, and protein expression was analyzed via Western blotting. The proteins were visualized using an anti-ASFV P30 antibody (from our laboratory), an anti-HA antibody (Proteintech), an anti-puromycin antibody (Millipore), an anti-Bip/GRP79 antibody (ABclonal), an anti-PKR antibody (ABclonal), an anti-p-PKR antibody (ABclonal), an anti-eif2α antibody (ABclonal), and an anti-p-eif2α antibody (ABclonal). GAPDH expression was detected using an anti-GAPDH monoclonal antibody (Mab) (Proteintech) to confirm equal protein loading. Horseradish peroxidase (HRP)-linked secondary antibodies (Proteintech) and enhanced chemiluminescence (Beyotime) were utilized to visualize the specific binding of the HRP-linked secondary antibody to the primary antibody-protein complexes under a UVP BioSpectrum 500 imaging system (GE, USA). The corresponding grayscale value of each protein band was analyzed using ImageJ software.

HEK-293T and IPI-2I cells in 24-well plates were transfected with PCAGGS, E66L, or truncated E66L. At 24 h posttransfection, the cells were washed three times with PBS, fixed with 4% paraformaldehyde for 15 min, and permeabilized with 0.2% Triton X-100 for 10 min at room temperature. After three washes with PBS, the cells were blocked with PBS containing 2% bovine serum albumin for 1 h and then incubated with MAbs against HA protein for 1 h. The cells were rinsed with PBS and then treated with an Alexa Fluor 488-conjugated anti-mouse secondary antibody (Thermo Fisher Scientific, USA) for 1 h. The fluorescence images were visualized using an inverted fluorescence microscope (Olympus IX73).

Reporter assay and ribopuromycylation assay.

The luciferase reporter plasmid pRL-SV40 used in this study was described previously (32). The functional and reporter gene plasmids were transfected into HEK-293T cells using Lipofectamine 3000 according to the manufacturer’s instructions. For luciferase reporter assays, HEK-293T and IPI-2I cells were transfected with the reporter plasmid pRL-SV40. Twenty-four hours after transfection, the cells were prepared, and a luciferase reporter assay system (Promega) was used to determine the luciferase activities of the lysed cells. A ribopuromycylation assay was performed as described previously (32). Briefly, cultured HEK-293T and IPI-2I cells were transfected with an equal dose of the different plasmids. At 24 h posttransfection, the cells were pulse-labeled with 3 μM puromycin and then incubated for an additional hour at 37°C with 5% CO₂.

DNA content analysis.

HEK-293T and IPI-2I cells starved from the logarithmic growth phase were seeded in a 6-well plate (2 ml in each well at 1 × 10⁶ cells/ml) and then transfected with 2.5 μg of empty vector, E66L expression plasmid, or E66L(13–34) expression plasmid for 24 h. To verify the effect of the virus on the susceptible cell cycle, we infected PAM cells with 0.1 MOI ASFV and ASFVΔE66L for 48 h. The cells were collected, washed twice with precooled PBS, mixed with precooled 75% ethanol, and fixed overnight at 4°C. According to the manufacturer’s instructions for the Cell Cycle Detection kit (Salarbio, 100 kit), the fixed cells were incubated with 50 ng/ml propidium iodide (PI) staining solution and 0.1 mg/ml RNase A for 30 min in the dark at 37°C. Flow cytometry analysis of the processed samples was performed using standard procedures. Generally, 20,000 cells were counted. Finally, the results were analyzed using ModFit software (BD Biosciences). Cell cycle data sets are presented as the means ± standard deviation (SD) of at least three independent experiments.

RNA extraction and qRT-PCR.

To determine the effects of ASFV and ASFVΔE66L on the expression of the host genes, including IFN-ALPHAOMEGA, IFNB1, TNF, MAP3K8, CCNL1, and CCNT2, PAM cells in 6-well plates were infected with 0.1 MOI ASFV and ASFVΔE66L. After 48 h, the total cellular RNA was extracted from the infected cells with TRIzol reagent (Invitrogen). RNA was then reverse transcribed into cDNA by avian myeloblastosis virus reverse transcriptase (TaKaRa, Japan). Then, the cDNA was used as the template in a SYBR green qPCR assay. The mRNA expression levels were normalized to those of GAPDH. All qPCR primers used in this study are listed in Table 1.

TABLE 1.

Sequences of primers used for qRT-PCR

Gene	Forward primer sequence (5′ to 3′)	Reverse primer sequence (5′ to 3′)
IFN-ALPHAOMEGA	GCTCCAGAATCGCTTGCTCT	CAGCCACGGCTTGTGTTTTC
IFNB1	GTTGCCTGGGACTCCTCAAT	ACGGTTTCATTCCAGCCAGT
TNF	CCAGACCAAGGTCAACCTCC	TCCCAGGTAGATGGGTTCGT
MAP3K8	TGATGTTCTCCTCGTCCCCT	ATTTCCACGTCCGATGGCTT
CCNL1	ATCGACCACTCGCTGATTCC	AGCCCAGAATGCGTAAGTCC
CCNT2	TGCTAGCAACAGGAATGCCA	ATGGTCCCTGCTGGAAGTTG
GAPDH	TCGGAGTGAACGGATTTGGC	TGACAAGCTTCCCGTTCTCC
Rluc	ATAACTGGTCCGCAGTGGTG	TAAGAAGAGGCCGCGTTACC

Treatment of transfected cells with cycloheximide.

HEK-293T cells were transfected with PCAGGS or PCAGGS-E66L for 24 h and then transfected with pRL-SV40. Cells were incubated with cycloheximide (CHX) (Sigma) at a final concentration of 100 µg/ml for 1 h before pRL-SV40 was transfected as described previously (63). After 12 and 24 h, cells were harvested and subjected to qPCR.

RNA sequence analysis.

The differential expression of RNAs in the samples of PAM cells infected with 0.1 MOI ASFV or ASFVΔE66L was determined using DESeq. In the screening results, transcripts with log₂ (FoldChange) greater than 1 or less than −1, and the corrected padj value less than 0.05, were considered to be the differential RNA in these two groups. Many genes were significantly upregulated in the ASFVΔE66L groups. Six of the RNAs (IFN-ALPHAOMEGA, IFNB1, TNF, MAP3K8, CCNL1, and CCNT2) were selected for validation using qRT-PCR.

Confocal fluorescence microscopy.

The cotransfection experiments were performed using a mixture of either full-length plasmids (E66L), deletion plasmids (E66L[13–34]), or mock plasmids (PCAGGS) and the marker vectors (pDsRed2ER) targeting specific subcellular sites. The transfected cells were incubated at 37°C for 48 h, washed twice with PBS (pH 7.5), fixed with 4% paraformaldehyde, and mounted with Mowiol 4–88 mounting medium for microscopy. The cells were examined (with sequential excitation at 488 nm for EGFP, 568 nm for DsRed, and 600 nm for DAPI) using laser scanning confocal fluorescence microscopy (Leica TCS-NT Laser Scanning) with a 63× oil lens objective. Each cotransfection experiment was repeated at least four times. All pictures were taken using a video imaging system mounted on the microscope, which consisted of a Hitachi CCD video camera operated by the Diskus software package. In certain cases, a 2× or 4× zoom was used for improved clarity. The images were exported as TIFF files and processed in Adobe Photoshop CS2.

Statistical analysis.

GraphPad Prism 6.01 software was used to perform all statistical testing, as detailed in the figure legends. Data are presented as the means ± SD, * indicates a significant difference (P < 0.05), ** indicates a highly significant difference (P < 0.01), and *** indicates an extremely significant difference (P < 0.001).

Ethics statement.

The ethics committee of Institute of Military Veterinary Medicine, Academy of Military Medical Sciences approved the protocol for this study. The viruses were inactivated in a BSL-3 laboratory, and the inactivated samples were transferred to a BSL-2 laboratory for genomic DNA extraction and detection.

Data availability.

The RNA sequence data were submitted to NCBI and assigned GEO accession number GSE158335.