4  Discussion
The interactions between host cells and viruses are highly complex, which usually involves numerous alterations in the expression of diverse genes, mRNAs, and proteins.36,37 Deciphering the laws behind these changes over the course of viral infection plays a vital role in elucidating the pathogenic mechanisms and in developing efficacious antiviral strategies.37 Over the past decade, MS-based proteomic techniques have contributed significantly to uncovering more factors and mechanisms related to viral infections and the corresponding host cellular pathophysiological processes.18 However, no research to date has focused on differential proteomic analysis of global protein profiles in host cells upon infection by PDCoV. In this study, iTRAQ combined with LC-MS/MS was used to identify the DEPs in PDCoV-infected IPEC-J2 cells. Although various cell lines including ST, LLC-PK1, PK-15, and IPI-2I cells have been shown to be highly permissive to PDCoV infection,17,34,38 considering porcine enterocytes are the natural targets for PDCoV infection in vivo,8,9 we chose to use IPEC-J2 cells for the proteomic analysis with the goal of obtaining experimental data that could better reflect the physiological state of pigs and the true state of PDCoV infection in vivo. Ultimately, a total of 78 DEPs were identified in PDCoV-infected IPEC-J2 cells, among which 23 proteins were significantly upregulated and 55 proteins were significantly downregulated. Moreover, we also identified six viral proteins including the spike, membrane, nucleocapsid, NS6, NS7, and Nsp5 proteins.
To ensure the reliability of the DEPs identified in the present study, we used qPCR and Western blot to validate two randomly selected DEPs, ANAPC7 and IFIT1, at the transcription and protein expression levels, respectively. In brief, ANAPC7 is an important constituent of the anaphase promoting complex/cyclosome, which is an E3 ubiquitin ligase that regulates the temporal progression of eukaryotic cells by mediating ubiquitination and subsequent 26S proteasome-mediated degradation of key cell cycle regulators.39 The downregulation of ANAPC7 is indicative of dysfunction of IPEC-J2 cells caused by PDCoV infection. IFIT1 is an innate immune effector molecule that can directly recognize viral single-stranded RNAs carrying a 5′-triphosphate group, thereby inhibiting the expression of viral mRNA.40 The upregulation of IFIT1 reveals that innate immune responses were activated in PDCoV-infected IPEC-J2 cells to combat with the invading virus. Through the validation of these two DEPs, we found that the obtained proteomics data were reliable and valid, and thus suitable for the subsequent bioinformatics analysis.
First, we made an attempt to assign possible functions to the 78 identified DEPs using GO functional annotation. This tool is an internationally standardized system for gene function classification, which provides a dynamic, updated, controlled vocabularies or ontologies and can well interpret the characteristics of target genes and gene products in various organisms.41 The proteins of our interest involving in different biological functions are listed as follows. Five type I interferon (IFN)-inducible proteins, ISG15, OAS1, Mx1, IFIT1, and IFIT3, were clustered into the BP category associated with the immune system process. Another five cell cycle-regulating proteins, ERCC6L, NME7, NFYB, FOSL1, and CTDP1, were also classified into the BP category but associated with the reproductive process. The nucleolar remodeling complex-constituting protein BAZ2A and the calcium ion binding protein secretagogin were clustered into the same CC category related to synapse.42,43 The apolipoprotein binding protein VLDLR and the cell cycle control and progression-related protein CDK9 were also divided into the CC category but associated with the macromolecular complex. In addition, three proteins, FOSL1, NUP37, and DDX55, were annotated to be associated with structural molecule activity, transporter activity and catalytic activity, respectively, within the MF category. For further functional annotation of the DEPs, we took a step further by performing COG function classification, which is a widely used tool for analyzing the function and evolution of proteins at the genome scale.44,45 We discovered that most of the DEPs were assigned new biological functions. For example, the immune system process-related proteins ISG15, OAS1, IFIT1, and IFIT3 categorized by GO functional annotation, were assigned to be involved in signal transduction mechanisms within the COG function classification; the proteins DDX55, SMARCA2, and ERCC6L belonging to three different GO categories were categorized into the same COG category—chromatin structure and dynamics; and the proteins NFYB, FOSL1, and CTDP1 belonging to the same GO category were still in the same COG category—transcription. The aforementioned findings suggested that there exist slight discrepancies in the classification of some proteins between the GO and COG functional annotations. This phenomenon is not unanticipated because the two functional annotation tools are based on different classification criteria.41,44
Next, we made an effort to identify the potential signaling pathways that might exist among the DEPs using KEGG pathway analysis, an extensively used method for the integration and interpretation of high-throughput proteomic and genomic data.35,46,47 We found that the DEPs were related to multitudinous signaling pathways. The top five pathways containing ≥4 DEPs included the viral infectious disease pathway, which is involved in eight proteins AKT2, IFIT1, RSAD2, NFYB, ANAPC7, COL4A1, Mx1, and ISG15; the signal transduction pathway, which is involved in six proteins AKT2, BNIP3, PDK1, WDR24, TMEM55B, and COL4A1; the immune system pathway, which is involved in four proteins ISG15, IFIT1, OAS1 and IFIT3; the digestive system pathway, which is involved in four proteins AKT2, MT-2A, COL4A1, and cystatin C; and the cancer pathway, which is involved in four proteins PDK1, AKT2, COL4A1, and CDK9. Once again, we found that the same protein can participate in different signaling pathways. For instance, ISG15, a IFN-α-inducible protein that is paramount to the host antiviral innate immunity,48 was simultaneously involved in two signaling pathways—viral infectious diseases and immune system. IFIT1 (also named p56/ISG56), an innate nucleic acid immune-sensing receptor that can recognize single-stranded viral RNA lacking 2′-O-methylation at the 5′-terminus and thus confers antiviral defense function by disrupting the machinery of host translation initiation,49,50 was also simultaneously related to viral infectious diseases and immune system signaling pathways. By contrast, COL4A1, also known as type IV collagen alpha 1 chain, was simultaneously involved in four of the top five signaling pathways, including viral infectious disease, signal transduction, digestive system, and cancers. This protein is an integral component of basement membranes, which can inhibit the migration, proliferation and tube formation by endothelial cells via binding to α-1/β-1 integrin, and thus becomes a potential therapeutic candidate for targeting tumor angiogenesis.51,52
Other interesting signaling pathways included RIG-I-like receptor, PI3K-AKT, mTOR, and autophagy signaling pathways. As an important family of cytosolic pattern recognition receptors, RIG-I is responsible for sensing of the invading viral RNA by recognizing its pathogen-associated molecular patterns to activate downstream signaling cascades, and thereby produce type I IFN.53,54 The generated IFN molecules then bind to IFN receptors and activate numerous ISGs, which exert critical antiviral innate immune functions either directly or indirectly.55 In our study, five upregulated proteins encoded by ISGs, including ISG15, OAS1, Mx1, IFIT1, and IFIT3, were identified in PDCoV-infected IPEC-J2 cells. Although these proteins related to type I IFN induction have been reported to participate in diverse viral infections,22,56 most of them were identified for the first time to be associated with PDCoV infection. These data suggest that the canonical IFN signaling pathways were activated in IPEC-J2 cells upon infection by PDCoV. This was further confirmed by a recent transcriptome-level study which also demonstrated that the RIG-I-like receptor signaling pathway was activated in PDCoV-infected cells, even though a different type of cells, PK-15, was used.17 Nevertheless, there is increasing evidence suggesting that PDCoV has evolved multiple escape strategies to interfere with the host’s innate immunity. For example, the nsp15 of PDCoV was found to be able to antagonize IFN-β production in LLC-PK1 cells by disrupting the phosphorylation and nuclear translocation of nuclear factor-κB p65 subunit, which is independent of its endoribonuclease activity.57 The Nsp5 of PDCoV was demonstrated to suppress the production of type I IFNs by cleaving the signal transducer and activator of transcription 2, depending on its protease activity.58 Moreover, the PDCoV accessory protein NS6 was shown to antagonize IFN-β production by disrupting the binding of double-stranded RNA to RIG-I/MDA5 receptors.59
Except for the RIG-I-like signaling pathway, two multifunctional signaling pathways, PI3K-AKT and mTOR, were also activated in PDCoV-infected IPEC-J2 cells, both of which participate in regulating autophagy.60 The PI3K-AKT signaling pathway could be triggered by numerous factors and regulates a variety of fundamental cellular functions, for instance, proliferation and survival.61 Activated AKT subsequently modulates numerous cellular processes, including cellular autophagy, cell cycle progression and cellular survival.62 The mTOR signaling pathway is involved in regulating diverse basic biological processes, including lipid biogenesis, protein synthesis, regulation of autophagy, cytoskeletal organization and so on,63,64 whose dysfunction has been associated with the pathophysiology of many diseases like diabetes and cancer.65 Although the mTOR and PI3K-AKT signaling pathways are able to negatively and positively regulate autophagy, respectively, in an independent manner,60,66−68 they usually coregulate autophagy via merging into a single PI3K-AKT-mTOR signaling pathway, which serves as one of the classical pathways for negatively regulating autophagy.69,70 From the KEGG analysis, we noticed that the PI3K-AKT and mTOR signaling pathways were respectively upregulated and downregulated in PDCoV-infected IPEC-J2 cells, which are the two important indicators of autophagic activation through the PI3K-AKT-mTOR signaling pathway. Accordingly, we speculate that PDCoV infection successfully activated autophagy in IPEC-J2 cells. Our speculation was in agreement with a recent research which revealed that PDCoV infection triggered autophagy in LLC-PK1 cells,71 despite using a different cell line. Undoubtedly, further studies are definitely needed to confirm the autophagy induced by PDCoV infection, to analyze the impact of autophagy on viral replication and to explore the underlying molecular mechanisms of PDCoV-induced autophagy. It should be noted that the PI3K-AKT-mTOR signaling pathway also plays an important role in many physiological and pathological conditions,69 except for regulating autophagy.
Finally, we sought to uncover the hidden interaction networks among the DEPs using STRING analysis, and discovered three major functional networks consisting of the RIG-I-like receptor, PDK1-AKT2 and cell cycle signaling pathways. These findings are in accordance with the results of KEGG pathway analysis, thus further consolidating the credibility of these putative signaling pathways involved in PDCoV-infected IPEC-J2 cells. Given the former two signaling pathways have already been discussed earlier, we next focus on the cell cycle signaling pathway. As can be seen from the network diagram (Figure 6), three small networks with NUP37, SMARCA2, and CDK9 as the respective hub proteins together constitute the cell cycle signaling pathway. NUP37 is an important constituent of the nuclear pore Nup107–160 subcomplex, which controls the bidirectional trafficking of macromolecules that traverse the nuclear envelope.72 SMARCA2 is an important member of the SWI/SNF family with helicase and ATPase activities, and has been shown to regulate numerous biological processes such as cell proliferation and DNA repair.73 CDK9 paired with cyclin T1 forms the positive transcription elongation factor b (p-TEFb) complex and induces transcriptional activation by hyperphosphorylating RNA polymerase II, and thereby regulating numerous vital cellular functions including proliferation, differentiation, DNA repair and apoptosis.74,75 However, growing studies suggest that CDK9 is also related to many pathologic processes, such as cancer, cardiovascular diseases and viral replication.74 Currently, CDK9 has been demonstrated to be involved in the replication of multiple viruses, such as influenza A virus, dengue virus, human adenovirus, and human immunodeficiency virus.75 For example, CDK9 was found to interact with the viral RNA-dependent RNA polymerases of influenza A virus and facilitate its association with cellular RNA polymerase II, thereby promoting viral transcription.76 CDK9 was also shown to be critical for the transcription of viral early genes and the replication of human adenovirus. Treatment of host cells with the CDK9 inhibitor, FIT-039, which functions by suppressing mRNA transcription, can efficiently inhibits the replication of human adenovirus.77 In the present study, CDK9 was significantly downregulated in IPEC-J2 cells upon PDCoV infection; however, the biological functions hidden behind this change warrant further investigation.
Of note, it is of great significance to compare our proteomics data with those of the newly emergent severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2),78,79 which caused the Coronavirus Disease 2019 (COVID-19) Pandemic and has posed a serious global public health emergency. Recently, Appelberg et al. conducted an integrative proteo-transcriptomics analysis of Huh7 cells responding to SARS-CoV-2 infection, and identified ErbB, HIF-1, mTOR, and TNF signaling pathways that were significantly regulated during SARS-CoV-2 infection in vitro.78 They further demonstrated that the Akt inhibitor MK-2206, targeting the mTOR signaling pathway, was able to significantly reduce the replication of SARS-CoV-2. Moreover, another recent study investigated the translatome and proteome of Caco-2 cells in response to SARS-CoV-2 infection in vitro, and discovered that several cellular pathways linked to translation, proteostasis, splicing, carbon metabolism, and nucleotide metabolism were reshaped during viral infection.79 On this basis, the authors tested two translation inhibitors, cycloheximide and emetine, for their ability to suppress SARS-CoV-2 replication, and found that both pharmaceuticals significantly decreased the replication of SARS-CoV-2 in Caco-2 cells. Undoubtedly, by comparing the similarities and differences of cellular proteomes between PDCoV- and SARS-CoV-2-infected host cells, it is possible to find some common signaling pathways and key adaptor molecules that function to inhibit viral replication, and thus provide valuable clues for screening of therapeutic drugs for PDCoV and designing novel antiviral strategies.