Phospho-proteomics identifies a critical role of ATF2 in pseudorabies virus replication.

Pseudorabies virus (PRV), an etiological agent of pseudorabies in livestock, has negatively affected the porcine industry all over the world. Epithelial cells are reported as the first site of PRV infection. However, the role of host proteins and its related signaling pathways in PRV replication is largely unclear. In this study, we performed a quantitative phosphoproteomics screening on PRV-infected porcine kidney (PK-15) epithelial cells. Totally 5723 phosphopeptides, corresponding to 2180 proteins, were obtained, and the phosphorylated states of 810 proteins were significantly different in PRV-infected cells compared with mock-infected cells (P ​< ​0.05). GO and KEGG analysis revealed that these differentially expressed phosphorylated proteins were predominantly related to RNA transport and MAPK signaling pathways. Further functional studies of NF-κB, transcription activator factor-2 (ATF2), MAX and SOS genes in MAPK signaling pathway were analyzed using RNA interference (RNAi) knockdown. It showed that only ATF2-knockdown reduces both PRV titer and viral genome copy number. JNK pathway inhibition and CRISPR/Cas9 gene knockout showed that ATF2 was required for the effective replication of PRV, especially during the biogenesis of viral genome DNA. Subsequently, by overexpression of the ATF2 gene and point mutation of the amino acid positions 69/71 of ATF2, it was further demonstrated that the phosphorylation of ATF2 promoted PRV replication. These findings suggest that ATF2 may provide potential therapeutic target for inhibiting PRV infection.

Introduction

Pseudorabies virus (PRV), an etiological agent for Aujeszky's disease, is a swine pathogen of the Alphaherpesvirinae subfamily which is associated with some devastating neurological and respiratory disorders in piglets or abortion in pregnant sows (Mettenleiter, 2000). The geographical widespread of pseudorabies caused by PRV has resulted in huge economic losses in the porcine industry, although in recent decades, several countries have been declared PRV-free in domestic pig populations (Moynagh, 1997; Pensaert and Morrison, 2000). The establishment of lifelong latency in peripheral ganglia is the hallmark of PRV and other alpha herpesviruses infections, resulting in the extremely difficult eradication of PRV between domestic pigs and wild boar. The most effective method for achieving a PRV-free pig herd is vaccination (Freuling et al., 2017). Unexpectedly, in 2011, a new PRV variant emerged in some pig farms in China where the Bartha-K61 vaccine was used (An et al., 2013; Yu et al., 2014). This means that there is the need to develop new prevention and treatment strategies to deal with future PRV mutations.

Effective prevention and treatment strategies rely on a complete understanding of PRV pathogenesis. As a strict intracellular parasitic pathogen, the replication cycle of PRV depends on host cells. The understanding of interactions between virus and host, and the identification of the indispensable host genes for virus proliferation are crucial for the development of novel vaccines and the exploration of potential therapeutic targets (Li et al., 2017). However, the interactions between PRV and host cellular factors are not well known. Host-pathogen interactions are complex and dynamic, and represent an area of focused research (Söderholm et al., 2016). There is increasing evidence showing that hijacking host post-translational mechanisms including protein phosphorylation is a vital strategy for viruses to effectively utilize host signal transduction (Wojcechowskyj et al., 2013; Ravikumar et al., 2015). High-throughput phosphoproteomics based on mass spectrometry is a powerful tool for analyzing the changes in host protein phosphorylation during virus infections (Macek et al., 2009). The proteomics data analysis of PRV infection on PK-15 ​cells has been reported (Flori et al., 2008; Wu et al., 2012). They found that many biological processes were altered during PRV infection and PRV infected epithelial cell line generates a diverse set of host miRNAs and a special cluster of viral miRNAs, which might facilitate PRV replication in cells. However, there are still few research reports on the high-throughput data analysis of PRV-dependent host protein phosphorylation transduction pathway (Yin et al., 2021).

Activating transcription factor 2 (ATF2) is a member of the activating protein 1 (AP1) family. In addition to being involved in gene regulation as a transcription factor, ATF2 also plays a role in the control of DNA damage response and the activity of the heteromeric amino acid transporters (HATs) complex (Bhoumik and Ronai, 2008). ATF2 also has a nuclear localization signal, which plays an important role in transporting extracellular stimulus signals to the nucleus (Yuan et al., 2009). Its activation is mediated by the phosphorylation of the threonine 69/71, which is activated by stress-activated protein kinases (SAPKs), p38, JNK and other kinases. Studies have shown that after PRV infection of PK-15 ​cells, the expression level of ATF2 does not change significantly, but its phosphorylation level is significantly increased (Yeh et al., 2008). However, whether the phosphorylation level of ATF2 has an effect on PRV infection has not been reported yet.

Here, we performed isobaric tags for relative and absolute quantitation (iTRAQ) and LC-MS/MS quantitative phosphoproteomics on PRV-infected PK-15 ​cells, which identified 5,723 phospho-peptides, corresponding to 2,180 proteins, including 810 proteins showing significant changes in phosphorylation level during early PRV infection. Kyoto Encyclopedia of Genes and Genomes (KEGG) differential enrichment analysis showed that PRV infection influences the MAPK signaling pathway, including several kinases and cellular factors. Functional studies showed that ATF2 is required for efficient PRV replication, and is activated by PRV through the c-Jun N-terminal kinase (JNK) signaling pathway. Epigenetic expression of the ATF2 wildtype and mutants at 69/71 phosphorylation sites further demonstrated that the increased phosphorylation level of ATF2 promotes PRV replication.

Materials and methods

Cell culture and virus propagation

PK-15 ​cells (ATCC CCL-33) and Madin-Darby bovine kidney (MDBK) cells (ATCC CCL-22) were purchased from the China Center for Type Culture Collection, the Cas9 stable-expressed PK-15 ​cell line (Cas9 PK-15) was a gift from Professor Shuhong Zhao (Key Lab of Agricultural Animal Genetics, Breeding, and Reproduction of Ministry of Education, Huazhong Agricultural University). These cell lines were cultured and maintained in Dulbecco's modified Eagle's medium (DMEM; Invitrogen) supplemented with 10% (v/v) FBS (Gibco, catalog no.10099-141), 100 U/mL of penicillin and 10 ​μg/mL streptomycin sulfates with 5% CO2 at 37 ​°C. The FBS level in the medium was reduced to 2% during virus infection. Wild-type PRV (WT-PRV) was isolated from an aborted pig fetus, and was propagated in PK-15 ​cells as described previously (Wu et al., 2012).

Virus infection and sample preparation for phosphoproteome analysis

PK-15 ​cells were grown to 80% confluence in 175 ​cm2 flasks (or 6-well plates, it depends on the specific experimental requirements), then the medium was discarded and cells were washed three times with ice-cold serum-free DMEM. Cells were precooled in 4 ​°C for 30 ​min, and infected with WT-PRV at a multiplicity of infection (MOI) of 10. Residual input viruses were then removed at 1.5 ​h post-attachment at 4 ​°C. Cells were washed three times with ice-cold serum-free DMEM and resuspended in fresh medium before further incubation in 5% CO2 at 37 ​°C for 2.5 ​h. Mock infected cells were treated with 0.5% DMEM instead of PRV infection using the same procedures. After 2.5 ​h post infection (hpi) (The time to collect the virus depends on the specific experiment), the medium was discarded and cells were washed three times with ice-cold PBS before lysing of cells in SDT lysis buffer [4% (w/v) SDS, 100 ​mmol/L Tris/HCl, 1 ​mmol/L DTT, pH 7.6]. Lysis samples were boiled for 15 ​min followed by ultrasonic processing (work 10 ​s, interval 15 ​s, 10 cycles). The supernatant was centrifuged at 13,200 ​×g for 30 ​min, and then quantified with the BAC Protein Assay Kit (Bio-Rad, USA). Protein quality was confirmed with SDS-PAGE (Schagger, 2006). Each group of samples was labeled with an equivalent amount of peptide (200 ​g) according to the iTRAQ reagent-8 plex Multiplex Kit (AB SCIEX) instructions.

The mixed phosphopeptides were enriched using Titanium dioxide (TiO2). The solution was vacuum-freeze-dried and diluted with 1 ​× ​DHB buffer [3% 2,5-dihydroxybenzoic acid (DHB), 80% acetonitrile (CAN), 0.1% Trifluoroacetate (TFA)]. TiO2 beads were added into the solution and oscillated for 40 ​min. The solution was then centrifuged and the supernatant was discarded. Retained beads were washed three times with washing buffer 1 (30% CAN, 1%TFA) followed by washing buffer 2 (80% CAN, 0.1% TFA) three times. The phosphorylated peptides were eluted with elution buffer, vacuumized, and dissolved in 30 ​μL 0.1% FA (Formic acid) and then 20 ​μL was taken for liquid chromatography tandem mass spectrometry (LC-MS/MS) analysis (Larsen et al., 2005). The mass spectrometer was operated in positive ion mode. MS data was acquired using a data-dependent top 10 method dynamically choosing the most abundant precursor ions from the survey scan (300–1800 m/z) for high energy collision dissociation (HCD) fragmentation. Automatic gain control (AGC) target was set to 3e6, and maximum inject time to 10 ​ms. Dynamic exclusion duration was 40.0 ​s. Survey scans were acquired at a resolution of 70,000 ​at ​m/z 200 and resolution for HCD spectra was set to 17,500 ​at ​m/z 200, and isolation width was 2 ​m/z. Normalized collision energy was 30 ​eV and the underfill ratio, which specifies the minimum percentage of the target value likely to be reached at maximum fill time, was defined as 0.1%. The instrument was run with peptide recognition mode enabled.

Mass spectrometry and data processing

The phosphopeptides were separated using a high-performance liquid chromatography (HPLC) liquid phase system Easy nLC1000 (flow rate, nL), and the samples were injected three times. Buffer solutions were as follows: buffer A, 0.1% formic acid aqueous solution; buffer B, 0.1% formic acid acetonitrile aqueous solution (84% acetonitrile). The column was balanced at 95% buffer A. The samples were added to a C18 chromatographic column by an automatic sampler [100 ​μm (diameter) ​× ​2 ​cm (length), 5 ​μm (particle size)] and separated by a C18 analytical column (75 ​μm ​× ​250 ​mm, 3 ​μm) at a flow rate of 250 ​nL/min. The gradient of the related liquid phase was as follows: 0–220 ​min, buffer B ranged linearly from 0% to 55%; 220–228 ​min, buffer B ranged linearly from 55% to 100%; 228–240 ​min, buffer B was maintained at 100%.

The samples were separated by capillary HPLC and analyzed by Q-Exactive mass spectrometer (Thermo Finnigan). Analysis was performed using the following settings: analysis time, 240 ​min; detection method, positive ion; parent ion scanning range, 350–1800 m/z; first-level mass spectral resolution, 70,000 ​at ​m/z200; AGC target, 3e6; first-level maximum Ion Trap (IT), 20 ​ms; number of scan ranges, 1; dynamic exclusion, 30 ​s. Mass charge ratios of peptides and peptide fragments were collected by the following methods: 10 fragments (MS2 scan) after each full scan; MS2 activation type, HCD; isolation window, 2 ​m/z; secondary-mass spectral resolution, 17,500 ​at m/z 200; microscans, 1; secondary-level maximum IT, 60 ​ms; normalized collision energy, 29 eV; underfill ratio, 0.1%.

The MS data were saved as raw files, then identified and quantitatively analyzed by Mascot v2.2 and Proteome Discoverer 1.4 (Thermo Scientific) software. The database used in this study was UniProt_Sus_scrofa_35472_20161104.fasta (35,472 sequences were included and the download date was 2016-11-04). When checking the library, the the raw files were submitted to the Mascot server through Proteome Discoverer, and then the database was searchable (Beausoleil et al., 2006; Olsen et al., 2006).

Western blot analysis

Cell pellets were washed for three times with ice-cold PBS, and were lysed in 1 ​× ​SDS-PAGE loading buffer [250 ​mmol/L Tris-HCL, 10% (w/v) SDS, 0.5% (w/v) bromophenol blue, 50% (v/v) glycerin, and 5% (v/v) 2-mercapto-ethanol] supplemented with a protease inhibitor cocktail (Catalog no. P8340, Sigma-Aldrich) and a phosphatase inhibitor cocktail (Catalog no. ZD416-1, Zomanbio) on ice for 5 ​min. Samples were then denatured at 100 ​°C for 10 ​min, and the subsequent steps were performed as described previously (Wang et al. 2017; Wang et al. 2018). The primary antibodies used in this study included anti-MAX (Catalog no. ab72997, Abcam; 1:2,000); anti-SOS1 (Catalog no. GTX133774, GeneTex; 1:1,000); anti-NF-κB (Catalog no. E381, Abcam; 1:2000); anti-RSK2 (Catalog no. NBP1-19295, Novus; 1:500); anti-phospho-RSK1/RSK2 (Catalog no. AF892, Biorbyt; 1:2,000); anti-ATF2 (Catalog no. SC242, Santa Cruz; 1:1,000); anti-phospho-ATF2 (Catalog no. 9221, GST; 1:1,000); anti-MAPK (Erk 1/2) (Catalog no. 9102, GST; 1:1,000); anti-GAPDH (Catalog no. 60004-1-lg, Proteintech; 1:5,000); anti-FLAG (Catalog no. AE005, Abcam; 1:10,000). To confirm equal sample loading the membranes were stripped and labeled with anti-GAPDH.

ATF2 transfection and RNA interference

ATF2 transfection was conducted following Lipofectamine 2000 (Invitrogen). The suspended PK-15 ​cells (ATF2-knockout), which were transfected with EGFP-tag recombinant plasmid (pcDNA3.1-EGFP, constructed and preserved by our laboratory), were detected by flow cytometry. Fluorescence of transfected PK-15 ​cells (ATF2-knockout) was analyzed by CytExport 2.0 CellQuest Pro software. The small interfering RNA (siRNA) used in this study (including the nonspecific control siRNA) were purchased from Gene Pharm (http://www.genepharma.com/productslist.), and three siRNA peer genes were designed and the most optimal one was chosen (Supplementary Table S1). siRNA was used at a concentration of 50 ​nmol/L.

Measurement of plaque formation unit (PFU) for viral titers

After PK-15 ​cells were grown in T-175 flasks until about 90% confluence, an MOI of 0.1 of PRV was added. At about 72 ​h post infection, cells were harvested and put into the refrigerator at −80 ​°C, and freeze-thawed for twice. After MDBK cells were grown in 24-well plates until about 60% confluence, the virus was sequentially diluted with DMEM containing 2% FBS, 100 U/mL of penicillin and 10 ​μg/mL streptomycin, and 200 ​μL per well of the 10−5, 10−6 and 10−7 dilutions were added in triplicate to 24-well plates, followed by gentle shaking. Plates were incubated at 37 ​°C and 5% CO2 for 2 ​h. During this period, the plates were shaked horizontally for 3–5 times every 15 ​min. The wells were then covered with 1 ​mL DMEM containing 4% sodium carboxymethyl cellulose, 2% FBS, 100 U/mL penicillin and 10 ​μg/mL streptomycin sulfate and incubated at 37 ​°C with 5% CO2 for 48 ​h. Cells were fixed with 10% neutral formaldehyde overnight at room temperature, washed with flowing water, and treated with 0.2 ​mL crystal violet dye solution. Plates were incubated at room temperature of 37 ​°C for 20 ​min, following which the crystal violet dye solution was discarded, and plates were washed and plaques were counted using a Stereo-Microscope.

RNA extraction and quantitative real-time reverse transcription (qRT-PCR)

Total RNA was extracted from treated cells with TRIzol® reagent (Invitrogen) following the manufacturer's instructions. One microgram of total RNA was reversely transcribed to complementary DNA (cDNA) with Easy-Script One-Step gDNA Removal and cDNA Synthesis Super-Mix (TRAN, cat. L21011), the reaction system is prepared as follows: Total RNA (DNA free) 50 ng–5 μg; Primer (10 ​μmol/L) 1 ​μL; EasyScript®RT/RI Enzyme Mix 1 ​μL; 2 ​× ​ES Resction Mix 10 ​μL; gDNA Remover 1 ​μL; RNase-free Water up to 20 ​μL, following which 1 ​μL cDNA was used in a SYBR Green real-time PCR system (Applied Biosystems, USA). qRT-PCR analysis was performed using Quantstudio™ Real-Time PCR Software (ViiA™-7 System, Life Technologies). Each sample was assayed three times and the data were normalized to porcine glyceraldehye-3-phosphate dehydrogenase (GAPDH) mRNA.

Inhibitor treatment and MTT cytotoxicity assay

PK-15 ​cells were treated with 20 ​μmol/L SP600125 inhibitor (Selleck, cat. S1460S) and 20 ​μmol/L SB202190 inhibitor (Selleck, cat. S1077), which were dissolved in DMSO. Cells were exposed to the inhibitors for 4 ​h and then infected with PRV at an MOI of 0.05. The negative control cells were treated with an equal volume of DMSO. Inhibitor cytotoxicity was detected by MTT assay (Mosmann, 1983). Cells were treated with different concentrations of pharmacological compounds for 24 ​h, then formazan was added to the cells for 2 ​h, followed by addition of isopropanol and finally, absorbance was detected with a PerkinElmer Multimode Plate Reader EnVision.

Plasmid construction

The lentivirus packaging plasmids, pMD2.G and pSPAX2 were purchased from Addgene (Watertown, MA), and the single guide (sg) RNA expression plasmid plenti-sgRNA-GFP was a gift from Professor Xingxu Huang (School of Life Science and Technology, Shanghai Tech University). Plasmid pcDNA3.1 (+) (catalog no. V709-20) was purchased from TaKaRa (Dalian, China). sgRNAs were designed to target the ATF2 gene by using 20-bp guide sequences immediately followed by a protospacer adjacent motif (PAM) sequence. The 20-bp guide sequences were analyzed according to the ATF2 gene with the help of the bioinformatics tool sgRNACas9 (Xie et al., 2014). The GC content of the guide sequences was set within 40%–60%. To construct the sgRNA expression vectors targeting ATF2, we synthesized primers followed by annealing with a PCR machine, then cloned the annealed product into the plenti-sgRNA-GFP vector. To construct the plasmid pATF2 and pmATF2 (T69A, T71A), we amplified the ATF2 coding regions from cDNA derived from PK-15 ​cells by PCR and the purified product was used as a template to amplify the mutant fragment by point mutation (T69A, T71A), then fragments were cloned into pcDNA3.1 to yield pATF2, pmATF2 (T69A, T71A), respectively. The primer sequences are described in Supplementary Table S2. The plasmid pFLAG was used as a negative control.

Construction of gene knockout cell lines

Through the third-generation lentivirus packaging system (pMD2. G: pSPAX2: plenti-sgRNA-GFP ​= ​1:2:3), we obtained an sgRNA (GTGAGTTGTTTCTACAACAGAGG) lentivirus that expressed the target ATF2 gene sequence in HEK293T cells, then Cas9 stable expressing PK-15 ​cell line was infected with lentivirus. Positive clonal cells (ATF2 knock-out PK-15 ​cells) were screened by limited dilution with fluorescence, and whether the cells expressed the ATF2 gene was determined by Sanger sequencing and Western blotting.

Restoration of ATF2 expression in knockout cell lines

Transfection was conducted using Lipofectamine 2000 (Invitrogen). The plasmid pATF2 was transfected into ATF2-knockout cell line and the restored expression of ATF2 was detected by Western blotting.

Statistical analysis

The statistical analyses of data were performed with GraphPad Software using two-way ANOVA, followed by Duncan's multiple range test (DMRT) to evaluate the significant differences, the significance level is indicated as ∗∗∗P ​< ​0.001, ∗∗P ​< ​0.01, ∗P ​< ​0.05; P ​> ​0.05 not significant (n.s.).

Results

Phosphoproteomic profiling of PRV-infected PK-15 ​cells by iTRAQ/LC-MS/MS

Recently, proteomic study of PRV infection of PK-15 ​cells was reported (Yang et al., 2017), however, global signaling pathways and host factors involved in PRV replication have not been well elucidated. To analyze the changes of early PRV infection on the phosphorylation of host proteins, 2.5 hpi was selected as the sample collection time point, (Flori et al., 2008; Tombacz et al., 2011). Additionally, cells were infected with PRV, an MOI of 10 as previously described (Yan et al., 2019). To quantify relative fold changes in host protein phosphorylation of PK-15 ​cells by PRV infection, we used TiO2-enriched phosphopeptide technology combined with iTRAQ technology and LC-MS/MS-based phosphoproteome analysis, coupled with bioinformatics analysis and functional assays. The workflow of phosphoproteomic profiling was described in Fig. 1A.

Fig. 1

Phosphoproteomic profiling of pseudorabies virus (PRV)-infected PK-15 ​cells by iTRAQ/LC-MS/MS. The omics experiment was divided into two groups, each with three biological replications. The mock group, also comprising PK-15 ​cells, was used as a negative control and treated in the same manner as the PRV-infection group. A Workflow for the phosphoproteomics experiment in PRV-infected PK-15 ​cells. B Confirmation of differentially expressed proteins and phosphoproteins of ATF2/RSK2 by Western blot. Total protein was extracted and electrophoresed by SDS-PAGE. ATF2, phospho-ATF2, RSK2 and phospho-RSK2 antibodies were used to detect the expression of ATF2 and RSK2, respectively. Protein levels were normalized to GAPDH.

In total, 5723 phosphopeptides originating from 2180 proteins were identified. Among the 2180 proteins, 810 of these proteins was significantly expressed (P ​< ​0.05). There were 678 phosphopeptides with phosphorylation level increased by 1.2-fold, corresponding to 484 proteins, and 603 phosphopeptides with phosphorylation level decreased by 0.83-fold, corresponding to 446 proteins in the PRV-infected group compared with the control group according to the FC ​> ​1.2 (Fold change), P ​< ​0.05 (Supplementary Tables S3, S4, S5).

To further confirm the phosphoproteomics data, we used phosphospecific antibodies to validate the iTRAQ phosphorylation protein ratios. Based on a previous study of PRV-induced upregulation of ATF2 (Yeh et al., 2008), we chose ATF2 (pT69 and pT71) and another ribosomal kinase RSK2 (pS378) from the phosphoproteomics panel. Immunoblot results showed that the ATF2 and RSK2 phosphorylation levels were enhanced upon the addition of PRV for 2.5 ​h (Fig. 1B), which was consistent with the phosphoproteomics results.

A recent study showed that very short incubations with low concentrations of adenosine diphosphate (ADP) in cell lysate triggers a sustained stimulation of intracellular signaling cascades using phosphoproteomics (Beck et al., 2017). To rule out the potential influence of the lysate, a control experiment with virus-free lysates was carried out. The result showed that equal amount of ATF2 appeared in PRV stock (containing PRV and cell lysates), mock (DMEM) and normal PK-15 ​cell lysates; phosphorylated ATF2 was more abundant in PRV infected PK-15 ​cells compared with mock (DMEM) and normal PK-15 ​cell lysates, while there is almost no difference between mock (DMEM) and normal PK-15 ​cell lysates (Supplementary Fig. S1). The result suggests that PRV infection induces up regulation of ATF2 phosphorylation.

Functional enrichment analysis of phosphorylated proteins differentially expressed upon PRV infection

The protein relative expression data was used for hierarchical clustering analysis (Fig. 2A). In the result of clustering grouping, the similarity of data patterns within groups is higher, while the similarity of data patterns between groups is lower, indicating that the rationale and accuracy of the selected differential proteins were reliable.

Fig. 2

Bioinformatic analysis of differentially phosphorylated proteins. A Clustering of phosphoproteomics data. M1, M2 and M3 represent three biological replicates of the mock infected group; P1, P2 and P3 represent three biological replicates of the PRV infected group. B Enriched GO terms analysis of the differentially phosphorylated proteins. The top 20 phosphorylated proteins were divided in three ontologies. (biological processes (BP), molecular functions (MF), cellular components (CC)). C Statistics for significantly enriched pathways during PRV infection with KEGG analysis. Statistical significance (P value) differentially expressed proteins was calculated with Fisher's Exact Test.

To further explore the impact of 810 differentially expressed proteins in cell physiological process and discover internal relations between these proteins, enrichment analysis was performed. We annotated the differentially expressed proteins, into three ontologies (biological processes, molecular functions, cellular components) with the Gene Ontology (GO) and UniProt databases (Ashburner et al., 2000) (Supplementary Table S6, S7, S8, S9), and the enriched GO Terms are shown (Fig. 2B, Supplementary Table S10). Among the three gene ontologies, the frequency of each GO term appears almost at the same level, but the proportion of binding in the molecular function is significantly higher than that of other terms in the same group.

To understand the enrichment of differentially expressed host proteins in the signaling pathways post PRV infection, we analyzed the data using the KEGG pathway database (Kanehisa and Goto, 2000). The statistical analysis of significant enrichment KEGG pathway revealed multiple pathways are regulated to some extent by PRV infection but the MAPK signaling pathway contained the most significant and enriched differentially expressed proteins (Fig. 2B and C, Supplementary Table S11, S12, S13). These results showed that the MAPK pathway is significantly stimulated after PRV infection.

RNAi-mediated knockdown of PRV-responsive phosphoproteins in the MAPK signaling pathway

To address whether PRV-responsive phosphoproteins in the MAPK pathway are important for PRV replication, we selected NF-κB, ATF2, MAX and SOS genes which are located in the different sub-pathways of the greater MAPK signaling pathway and have been observed to undergo changes in phosphorylation level following PRV infection according to the MAPK pathway map (Map04010 by CytoScape software (Version number: 3.2.1), Supplementary Table S11). We designed siRNAs to target the four host genes (three siRNAs per gene), and measured the siRNA efficiency at the mRNA and protein levels using RT-qPCR and Western blot, respectively, showing that these siRNAs can silence target gene sequences (Fig. 3A and B).

Fig. 3

siRNA-mediated knockdown of PRV-responsive phosphoproteins in the MAPK signaling pathway. A PK-15 ​cells were transfected with specific siRNAs against NF-κB, ATF2, MAX, and SOS, and gene transcription was detected by real-time PCR. Individual gene expression was normalized to a non targeted control siRNA (negative control). B Cell lysate was loaded onto 1 ​× ​SDS-PAGE gel and the interference efficiency of siRNA was analyzed by Western blot. Negative control siRNA was set as the loading control (100%). The normalized relative protein amount was quantified by densitometry scanning (ImageJ2x) and GAPDH was used as the internal control. C The effect of siRNA on PRV titer was quantified by plaque assay. D The effect of siRNA on PRV genome copy number was quantified by qRT-PCR. ∗P ​< ​0.05; ∗∗P ​< ​0.01; ∗∗∗P ​< ​0.001; n.s., not significant.

To further confirm the effect of these genes, PK-15 ​cells were transfected with effective siRNAs against NF-κB, ATF2, MAX and SOS genes and infected with WT-PRV at an MOI of 0.05 ​at 12 ​h post-transfection. At 12 hpi, we monitored PRV titer by plaque assay and viral genome copy number by RT-qPCR (Fig. 3C and D). The results showed that, ATF2-knockdown reduces both PRV titer (P ​< ​0.001) and viral genome copy number (P ​< ​0.05), while NF-κB-knockdown only reduces viral genome copy number (P ​< ​0.05), MAX-knockdown and SOS-knockdown only reduce PRV titer (MAX: P ​< ​0.01, SOS: P ​< ​0.05). These results suggest that ATF2-knockdown inhibits virus replication.

Effect of inhibitor-mediated interference of ATF2 gene signal transduction on PRV replication

ATF2 is a member of the ATF-CREB family of transcription factors, which are involved in cellular responses to stress (Hai et al., 1989; van Dam et al., 1995). Previous studies have indicated that phosphorylation of ATF2 is triggered by stress-activated protein kinases p38 (Gueorguiev et al., 2006), JNK (Gupta et al., 1995), protein kinase C (PKC) (Lim et al., 2005), or phosphatidyl inositol 3 kinase (PI3K) (Recio and Merlino, 2002). To investigate whether p38 or JNK cellular pathway activates ATF2 phosphorylation following PRV infection (Map04010, Supplementary Table S11), we selected inhibitors that specifically targeted the two signaling pathways: SP600125-JNK (Shen et al., 2017) and SB202190-p38 (Duzgun et al., 2017). Firstly, we detected the cytotoxicity of the different inhibitors by MTT assay. The results indicated that PK-15 ​cell growth rate was more than 85% for both inhibitors at concentrations lower than 40 ​μmol/L (Fig. 4A). Then, to determine the optimal concentration of inhibitors, PK-15 ​cells treated with two inhibitors at different concentrations were infected with WT-PRV and then, phosphorylation of ATF2 was detected at 2.5 hpi by Western blot. The results showed that both inhibitors can inhibit ATF2 phosphorylation and 20 μmol/L was the optimal concentration for inhibition (Fig. 4B).

Fig. 4

Inhibitor-mediated interference of ATF2 gene signal transduction on PRV replication. A Cytotoxicity of JNK (SP600125) and p38 (SB202190) inhibitors in PK-15 ​cells was analyzed by MTT assay. B Effect of JNK (SP600125) and p38 (SB202190) inhibitors on ATF2 phosphorylation. PK-15 ​cells were treated with 0.1% DMSO or increasing amounts (20, 30, 40 ​μmol/L) of SP600125 and SB202190. After incubation for 4 ​h, all cells were infected with PRV at an MOI of 10. Phosphorylation of ATF2 was analyzed by Western blot at 2.5 hpi. PRV replication titer was quantified by plaque assay in the presence of SP600125 (C) and SB202190 (D), respectively. The concentrations of SP600125 and SB202190 were both 20 μmol/L. ∗P ​< ​0.05; ∗∗P ​< ​0.01; ∗∗∗P ​< ​0.001; n.s., not significant.

To determine which pathways (JNK or p38) contribute to PRV replication, we infected cells with WT-PRV and detected viral titers in inhibitor-treated or untreated cells at 0, 3 and 12 hpi by plaque assay. The results showed that PRV titers were significantly reduced when the JNK signaling pathway was blocked, while inhibition of the p38 signaling pathway had no significant effect on virus proliferation (Fig. 4C and D). These data indicate that the JNK pathway regulates ATF2 phosphorylation and PRV replication.

Knockout of ATF2 gene inhibits PRV replication in PK-15 ​cells

To further explore the role of the ATF2 gene in virus proliferation, the ATF2 knock-out PK-15 ​cells were constructed by CRISPR/Cas9 lentivirus (Fig. 5A). Subsequently, Sanger sequencing and Western blot showed that ATF2-KO-#1 and ATF2-KO-#4 did not express ATF2 protein at all (Fig. 5B and C). The transfection efficiency of PK-15 was measured with Flow Cytometry. The results showed that about 48.4% of PK-15 ​cells expressed fluorescent EGFP plasmid (Supplementary Fig. S2). Then the pATF2 plasmid was transfected into cells, and the expression level of ATF2 was restored to some extent in ATF2-KO-#1 and ATF2-KO-#4 (Fig. 5C).

Fig. 5

PRV replication was inhibited by ATF2 gene knockout in PK-15 ​cells. A Schematic diagram of the generation of ATF2-null Cas9 PK-15 ​cells. sgRNA (GTGAGTTGTTTCTACAACAGAGG) was designed against exon 2 of ATF2. Prime-F (ACCCCAACACCAACACGATT) and Prime-R (TCTCCCCAATAATTTTTCTGTGTCT) were used to amplify the ATF2-exon sequence to detect the correct knockout. B Sequencing peak diagram of control and ATF2-KO-#1 and ATF2-KO-#4 Cas9-expressing PK-15 ​cells. C Verification of ATF2-null and ATF2-supplemented expression in ATF2-KO-#1 and ATF2-KO-#4. Cas9-expressing PK-15 by Western blot. Total protein was extracted and electrophoresed by SDS-PAGE. ATF2 antibody was used to detect the expression of ATF2. D The effect of ATF2 gene on PRV replication in PK-15 ​cells. ATF2-KO-#1 and ATF2-KO-#4 ​cells were directly infected with PRV or transfected with pATF2 plasmid and then infected with PRV, and viral titers were detected by plaque assay at 0, 3 and 12 hpi. ∗∗P ​< ​0.01; ∗∗∗P ​< ​0.001.

We used WT-PRV at an MOI of 0.05 to infect ATF2-KO-#1, ATF2-KO-#4 ​cells and control cells. Furthermore, ATF2-KO-#1 and ATF2-KO-#4 ​cells respectively transfected with pATF2 plasmid before WT-PRV infection. Then PRV titers were monitored at 0, 3 and 12 hpi by plaque assay. The results showed that PRV proliferation was suppressed in ATF2 knock-out cells at the middle and late infection time points, and ATF2 plasmid transfection could restore PRV production (Fig. 5D).

Phosphorylation of ATF2 promotes PRV replication

Furthermore, to investigate whether the effect of ATF2 on PRV replication through translational or post-translational phosphorylation, we overexpressed complete ATF2 and 69/71 (T69A, T71A) mutated ATF2 into PK-15 ​cells. The results showed that the phosphorylation level of wildtype ATF2 protein was increased, but the phosphorylated mutated ATF2 (T69A, T71A) was decreased (Fig. 6A).

Fig. 6

Phosphorylation and activation of ATF2 promote PRV proliferation. A Western blot analysis to detect the expression of complete or mutant ATF2 (T69A and T71A) in PK-15 ​cells. Plasmids pEGFP, pATF2, or pmATF2 (T69A and T71A) were transfected into PK-15 ​cells for 24 ​h. Total protein was extracted from cell lysate and electrophoresed by SDS-PAGE. Antibodies of ATF2 and phospho-ATF2 were used to detect the expression of ATF2 and phosphorylated ATF2. p-ATF2 represents phosphorylated ATF2. B PK-15 ​cells were tansfected with complete ATF2 or mutant ATF2 (T69A and T71A) gene for 24 ​h, EGFP plasmid was used as a negative control, and then the cells were infected by PRV at an MOI of 1. Virus titers were determined at 0 and 12 hpi. ∗∗P ​< ​0.01.

Then effects of ATF2 mutation on PRV replication were compared. Complete ATF2 or 69/71 (T69A, T71A) mutated ATF2 was transfected into PK-15 ​cells, followed by WT-PRV infection at 24 ​h post-transfection, and virus titer was measured by PFU assay at 0 and 12 hpi. The results showed that transfection of complete ATF2 significantly increased the viral titers of PRV at 12 hpi, while mutated ATF2 (T69A, T71A) had no significantly reduced effect on the viral titers compared with the control group (Fig. 6B). These data suggest that expression of ATF2 increases PRV replication and the phosphorylation level of ATF2 plays a positive role in PRV growth.

Discussion

PRV is an important veterinary virus that causes huge economic losses in livestock. PK-15 ​cells are an appropriate model for the study of the interactions between virus and host (Wu et al., 2012; Liu et al., 2016; Yang et al., 2017). Recent report documented the proteomic profile of PK-15 ​cells during PRV infection (Yang et al., 2017). In our study, we applied the isobaric tags for relative and absolute quantification (iTRAQ) labeling coupled with liquid chromatography-tandem mass spectrometry in PRV-infected PK15 ​cells, and identified 4333 proteins. About 466 proteins were expressed differently, including 234 up-expressed proteins and 232 down-expressed proteins. We found significant differences in phosphorylated state of about 810 proteins. Two types (phosphorylated and unphosphorylated) of representative proteins were verified by Western blotting. According to the results of GO Analysis, the differentially expressed proteins were assigned to the groups of biological process, molecular function or cellular components. The analysis of pathway (KEGG) indicated that there was an enhanced activation of the MAPK signaling pathway in PK-15 ​cells during PRV infection, including the variation of 25 proteins in phosphorylation level, which is in accordance with the previous report (Yang et al., 2017). Coincidentally, some reports showed that the viral infections may be bounded to the MAPK pathways, such as Japanese encephalitis virus (Xing et al., 2021), enterovirus (Zhu et al., 2020), coxsackievirus B3 (He et al., 2019), Human immunodeficiency virus type 1 (Kumar et al., 1998), HSV-1 (McLean and Bachenheimer, 1999; Zachos et al., 1999), and human cytomegalovirus (Rodems and Spector, 1998). A recent study showed that PRV triggers proteasomal degradation of the Janus kinases Jak 1 to inhibit both type I and type III IFN-induced STAT1 phosphorylation, suppressing IFN-induced ISGs expression (Yin et al., 2021). However, the phosphorylation level of the MAPK signaling pathway affects the life cycle of PRV remains unclear.

ATF2, an activating transcription factor, is a member of the activator protein 1 (AP1) family. ATF2 forms homologous or heterologous dimers with other AP1 family members (CREB, Fos, Maf and Jun etc.), and regulates gene expression in the nucleus (Yuan et al., 2009; Inoue et al., 2018; Ma et al., 2018). Our results indicated that the interference of ATF2 gene effectively inhibit the replication of virus. We also found that the overexpression of ATF2 facilitated PRV proliferation, which was consistent with the findings of Yeh et al. (Yeh et al., 2008). These results suggested that ATF2 participates in the replication process of PRV.

The phosphoproteomic data indicated that PRV infection upregulated the phosphorylation level of ATF2 in the PK-15 ​cells. And the mutation of ATF2 phosphorylation reduced PRV titer.

Previous studies showed that the transcriptional activation of ATF2 is mediated by p38 and JNK (Gupta et al., 1995; Ouwens et al., 2002; Wang et al., 2011). We found that the inhibition of JNK signal transduction decreased ATF2 phosphorylated level. These results indicated that the proliferation of PRV depends on the activation of the ATF2, which was mediated by JNK. Previous report demonstrated inhibition of JNK and p38 reduced cell apoptosis during PRV infection (Yeh et al., 2008). The inhibition of c-Jun N-terminal kinase induces the blocking of ATF2 phosphorylation and reduces ganglion cell death. However, in our study there were no dramatically effects while inhibiting the activity of p38. These results indicated that the p38 plays almost no effects on the phosphorylated ATF2 level.

The establishment of lifelong latent infection in the peripheral nervous system is a distinct character of alpha-herpesvirus, which restricts the treatment of viral infection diseases. The latency-associated transcript (LAT) was expressed to inhibit neuronal apoptosis, and promotes the establishment of latent infection (Stevens et al., 1987). The hypokalemia-induced neuronal apoptosis induced the upregulation of phosphorylation and the transcriptional activity of ATF2 was enhanced (Yuan et al., 2009). The inhibition of ATF2 facilitated neurons avoiding from cell apoptosis, and formation of heterodimers between ATF2 and c-Jun promoted neuronal apoptosis. Furthermore, ATF2 is involved in the maintenance of nerve-injury mediated tactile pain and hyperthermalgesia (Salinas-Abarca et al., 2018). The transcription and phosphorylation levels of ATF2 are also associated with the replication of multiple viruses, including Epstein-Barr virus (EBV), HCMV and Kaposi's sarcoma-associated herpesvirus (Sharma-Walia et al., 2005; Yuan et al., 2009; Murata et al., 2011). Together, there might be a correlation between the ATF2 and neuron survival, which stimulates reactivation of PRV from latent infection.

Conclusions

In summary, our study revealed the interaction between the phosphorylation level of the host protein ATF2 and the proliferation of PRV: PRV can activate ATF2 through JNK Pathway, the proliferation of PRV depends on the phosphorylation activation of ATF2 protein, up-regulation of the phosphorylation level of ATF2 will promote PRV replication. It will be important to extend this study to the field of antiviral drug research and development, as the findings presented here suggest that ATF2 may provide potential new therapeutic targets for inhibiting PRV infection.

Data availability

The mass spectrometry proteomics data was deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD014685. MS data in PRIDE is available online (URL: https://www.ebi.ac.uk/pride/archive/projects/PXD014685/priva).

Ethics statement

The study was reviewed and approved by Research Ethics Committee of the College of Veterinary Medicine, Huazhong Agricultural University, Hubei, Wuhan, China.

Author contribution

Fang-Fang Jiang: conceptualization, methodology, investigation, data curation, formal analysis, visualization, writing-original draft. Ren-Qi Wang: methodology, investigation, formal analysis, data curation. Chao-Yue Guo: writing-review & editing. Ke Zheng: conceptualization, methodology. Hai-Long Liu: formal analysis, visualization. Le Su: resources, investigation. Sheng-Song Xie: resources. Huan-Chun Chen: supervision. Zheng-Fei Liu: conceptualization, methodology, project administration, writing-review & editing, funding acquisition.

Conflict of interest

The authors declared that no competing interest exists.