The IRF family has been reported to have 10 members, namely IRF1-IRF9 and virus IRF (v-IRF). IRF3, a principal transcription factor, is significant in the antiviral immune response (34). IRF3 is highly homologous to IRF7. Both of them regulate the type I IFN synthesis, but play different roles in the innate immune response. IRF3 is critical for early induction of IFN expression in most cells post-viral infection; IRF7, which induces both IFNα and IFNβ expression, has functions in the antiviral activity of IFN in a later stage. In contrast, IRF3 can induce IFNβ gene expression, but not other IFNα expression except IFNα4 (35). Upon viral infection, a series of cellular pathways are activated subsequently to promote the translocation of phosphorylated IRF3 or IRF7 into the nucleus and initiate the transcription of type I interferon genes by attaching to IFN-α/β promoters (36).

By luciferase reporter gene experiment, La Rocca et al. found that CSFV-infected cells could inhibit IRF3 gene transcription. The use of cell lines expressing CSFV N^pro confirmed that the N^pro protein reduced the expression of IRF3, suggesting that this single viral protein specific to the pestiviruses can inhibit interferon production in the innate immune response to the virus (37) ( Figure 1 ). Hilton et al. reported that NCP-BVDV (pe515) infection could induce the translocation of a small amount of IRF-3 from the cytoplasm to the nucleus at the early stage of infection. In addition, most IRF3 in the cytoplasm was degraded by the ubiquitination-proteasome pathway mediated by the N^pro protein (27). Similar to the NCP-BVDV virus, CP-BVDV (NADL strain) does not induce interferon response after infection and blocks interferon-stimulating genes induced by paramyxovirus infection, resulting in a significant decrease in IRF3 expression. However, the IRF3 repression activity is considered independent of the protease activity of N^pro. Further studies revealed that N^pro could interact with IRF3 before its phosphorylation-induced activation, leading to the ubiquitination and proteasomal degradation of IRF3 (38).

The N^pro protein of CSFV can also mediate the degradation of IRF3 after interaction with it. However, different from the research results of La Rocca et al., There are other studies reported that CSFV infection does not inhibit a cytomegalovirus (CMV) promoter-driven IRF3 expression. Furthermore, CSFV neither reduces the transcriptional activity of the IRF3 promoter nor affects the stability of IRF3 mRNA (39, 40).

Ubiquitin contains 76 amino acid residues and is highly conserved in all eukaryotes (41). Selective binding of linear ubiquitin to a protein is the initial signal for target protein degradation. Ubiquitin chains can be conjugated to the specific protein substrate through an isopeptide bond between the ubiquitin C-terminal glycine residues and the lysine residues in the substrate. Proteasome-dependent degradation of proteins mediated by ubiquitination regulates a variety of biological reactions in the body, including cell cycle, signaling, DNA repair, and apoptosis (42, 43).

The ubiquitin modification process requires successive events associated with three enzymes: an E1 ubiquitin-activating enzyme, an E2 ubiquitin-conjugating enzyme, and an E3 ubiquitin ligase. Ubiquitin is first connected to the catalytic cysteine residues of endogenous E1 protein by an intermediate thioester bond. It is then conveyed to an E2 conjugating enzyme through a trans-esterification reaction. Subsequently, E3 ubiquitin ligase transfers the activated ubiquitin from an E2 to substrates. Finally, the substrate labeled with ubiquitin is delivered to the 26S proteasome for degradation (44). Once the E1 ubiquitin-activating enzyme was thermal inactivated, the degradation activity of IRF3 mediated by the N^pro protein of CP-BVDV was lost (39), indicating that the E1 ubiquitin-activating enzyme is involved in the N^pro protein-mediated ubiquitin modification and degradation of IRF3.

Besides contributing to the antiviral immune response, IRF3 participates in the control of the cell cycle, apoptosis, and tumor suppression as well (45). When cells are infected with Sendai virus or stimulated by double-stranded RNA, IRF3 can bind to cytoplasmic Bax via its BH3 region and transfer Bax to mitochondria to activate apoptosis. It has been shown that wild-type N^pro could restrain apoptosis signals induced by astrosporin, interferon, double-stranded RNA, sodium arsenate, and hydrogen peroxide, which was mainly achieved by the degradation of IRF3 protein. In addition, the N^pro protein can inhibit the Bax-dependent apoptosis pathway by inhibiting mitochondrial lysis and Bax redistribution through interaction with the mitochondrial lumen. Moreover, the N^pro protein could rapidly localize to ubiquitin-containing peroxisome. Thus, the N^pro protein may bind to IRF3 and degrade it by transporting it to ubiquitin-containing peroxisome (46).

Peptide-prolyl isomerase (Pin1), Ro52 (TRIM21), RBCC protein interacting with PKC1 (RBCK1), RTA-associated ubiquitin ligase (RAUL), Tripartite motif 21 (TRIM21), Forkhead transcription factors of the O class (FOXO1), Casitas B-lineage lymphoma (c-Cbl) have been suggested as E3 ligases to invoke ubiquitination and subsequent degradation of IRF3 in the cytoplasm, thus inhibiting the production of type I IFN (47–51). However, it is unclear which E3 ubiquitin ligase contributes to the ubiquitination degradation of IRF3 mediated by the N^pro protein. It is worthy of further exploration whether N^pro has crosstalk with these E3 ligases or molecules to regulate the ubiquitination and subsequent degradation of IRF3.

Currently, the N^pro binding site on IRF3 is unclear. It has been revealed that N^pro can interact with IRF3 directly and forms a soluble 1:1 complex by utilizing recombinant N^pro and IRF3 proteins. This interaction requires the complete IRF3 protein rather than any of the single domains, the DBD in N-terminal or the IAD in C-terminal (52). It has been also shown in a previous mammalian two-hybrid analysis that the association of N^pro with IRF3 needs both the DBD and the IAD of IRF3 (53). However, the exact arrangement of DBD and the IAD in the IRF3 monomer and dimer is still unknown (35, 54, 55). As the ~60-amino-acid linker region between the two domains is somewhat helical but not structured in the absence of either DBD or IAD (56). The intact linker is thus suggested to be involved in N^pro binding (52). N^pro has been shown to interact with the IRF3 monomer and phosphomimetic dimer, indicating that the N^pro binding site on IRF3 contains areas not affected by the phosphorylation and subsequent activation status of IRF3 (52). N^pro can also interact with IRF3 in the complex with its transcriptional cofactor, the CREB-binding protein (CBP). Therefore, the contact surface in the IRF3 dimer and CBP binding site is not required for N^pro binding (52).

BVDV N^pro protein has been shown to degrade IRF3 in the cytoplasm, whereas IRF3 in the nucleus is resistant to this degradation (26). The influence of cellular localization of N^pro on IRF3 degradation is unclear. A recombinant virus vSMS-IRF3 was constructed by inserting the IRF3 gene sequence between the 13th and 14th amino acid sites of the N^pro protein of the highly virulent CSFV Shimen strain by Li et, al (57). The fusion protein of IRF3-N^pro expressed by the recombinant virus only located in the cytoplasm and vSMS-IRF3 was significantly attenuated. Pigs inoculated with the recombinant virus were all resistant to the lethal CSFV challenge, but the parent virus showed a typical virulent phenotype (57). Therefore, it was suggested that the nuclear localization of N^pro is essential to the replication and virulence of CSFV (57).

However, a previous study showed that any mutants of L8P, E22L, and H49V in N^pro could abolish its IFN-α/β antagonistic activity, revealing that the 49 amino acids in the N-terminal of N^pro protein are necessary to type-I IFN suppression (33). Based on this, we speculate that the insertion of the IRF3 gene into the N^pro gene may eliminate its function of IRF3 degradation. Thus, the attenuation of vSM-IRF3 may be caused by the loss of the IFN-α/β antagonistic activity of IRF3-N^pro rather than its cytoplasmic localization. Moreover, N^pro was observed in the nucleus in a diffuse manner (58, 59), and could bind to IRF3 dimer or the IRF3 dimer in the complex with CBP (52). Therefore, the insertion of IRF3 into N^pro could also affect the cellular diffusion of the IRF3-N^pro fusion protein, leading to its accumulation in the cytoplasm. The effects of cellular localization of N^pro protein on CSFV virulence need further study.

Whether N^pro protein is the main determinant of the virulence of pestiviruses is still a controversial topic. Continuous passage of CSFV attenuated vaccine strain GPE- in pigs restored its virulence, but did not regain the ability of its N^pro to degrade IRF3 (60). However, strains containing the N136D mutation in N^pro restored the IRF3 degradation activity and IFN-α/β antagonistic ability in vitro as well as pathogenicity in vivo. These results demonstrate that the N^pro protein makes a decisive contribution to the virulence of pestiviruses, but there are other factors that can regulate the virulence of pestiviruses.

The N^pro protein of CSFV can also interact with IRF7 in plasmacytoid dendritic cells, down-regulating the expression level of IRF7 protein and further inhibiting the IFN-α expression. Whereas, the molecular mechanism of N^pro inhibiting the expression of IRF7 protein is still unclear. It is certain, however, that this antagonism does not involve either polyubiquitination or protease degradation pathways (53).

Analysis by sequence alignment revealed that the C-terminal half of the N^pro protein contains a conserved metal binding TRASH motif composed of Cys-X21-Cys-X3-Cys (where X is any amino acid). TRASH motif commonly exists in proteins associated with heavy metal recognizing, resistance, transcription regulation, cation transportation, and hydrogenase. Inductively coupled plasma–mass spectrometry (ICP-MS) assay indicated that each N^pro protein molecule could coordinate a single zinc atom. Site-directed mutagenesis studies revealed that the zinc-binding sites of N^pro protein include Cys112, Cys134, Cys138, and probably Asp136. These zinc-binding site mutations lead to the deficit of N^pro protein-mediated IRF3 degradation in cells inoculated with CSFV, suggesting that the zinc-binding capacity of N^pro protein is critical for virus-mediated IRF3 degradation (61). In addition, the zinc-binding domain of N^pro is critical for its protein stability and its interaction with IRF7 (62).

Recently, pull-down combined mass spectrometry showed that N^pro binds to more than 55 kinds of proteins, mainly RNA helicase A (DHX9), Y-box binding protein (YBX1), DDX3, DDX5, IGF2BP1, eIF3, and other ribosomal proteins, multiple myeloma tumor protein 2, interleukin enhancer binding factor 3 (IEBP3) guanine nucleotide-binding protein 3, and polyadenylate-binding protein 1 (PABP-1). Many of the interacting proteins are components in cytoplasmic ribonucleoprotein particles (RNPs). They play roles in regulating the translation of mRNA and could be recruited into stress granules to regulate the translational initiation rate or mRNA degradation (63). The assembled stress granules might control the proliferation of viruses and some viruses could in turn hinder their formation or even disassemble them (64). It has been suggested that N^pro could redistribute to stress granules after interaction with YBX1 through its TRASH domain. When exposed to oxidative stress, cells expressing N^pro alone assembled stress granules and N^pro colocalized with stress granule proteins. In contrast, the formation of stress granules in NCP-BVDV-infected cells was inhibited, indicating that this inhibition was not caused by N^pro binding to ribonucleoproteins (63). Thereby, N^pro may not influence the function and location of ribonucleoproteins although it could be localized to stress granules by interacting with these host proteins. As some interacting proteins of N^pro are also involved in RISC function during RNA silencing, further studies were conducted to determine whether N^pro affected RNA interference (RNAi). However, the outcomes suggested the expression of N^pro had no influence on RNAi silencing activity (65).

S100A9, one of the danger-associated molecular patterns (DAMPs) proteins, is vital in the innate immune system and always accumulates in large amounts in ectocytic space during inflammation responses (66). Additionally, the S100A9 could effectively trigger inflammatory responses through Toll-like receptor 4 (TLR4) as a homodimer (67–73). In the heterodimer with S100A8, the S100A9 exerts antimicrobial activity by inhibiting microbes from acquiring nutrients (74–79). Yet, the function of S100A9 in virus infection is unclear. It has been shown that the amount of S100A9 was increased after treatment with poly (I: C), which is an analog of viral dsRNA (80). The expression of S100A9 is also highly enhanced in human papillomavirus-associated dysplastic tissues (81) and BKV-infected recipients post-kidney transplantation (82). Likewise, high levels of S100A9 expression were observed in lungs and livers in autopsied subjects with COVID-19 and pre-existing chronic liver disease (83). Darweesh et al. reported that NCP-BVDV2a 1373 N^pro protein has a strong interaction with cellular S100A9 protein. Furthermore, the N^pro protein enhances the replication of BVDV in infected cells by inhibiting S100A9 activity in epidermal cells (84).

Currently, the TRIM family consists of more than 60 members and could be divided into 11 subfamilies (85). Although their exact functions are still unclear, the TRIM proteins contribute to a large variety of biological activities, such as cell proliferation, development, differentiation, immunity, apoptosis, and innate immune response to pathogens (86–90). As a protein of the fifth subfamily of TRIM, TRIM56 is expressed in the cytoplasm after type I interferon stimulation (91, 92). A previous study has identified TRIM56 as a cellular protein that associates with BVDV N^pro through its C-terminal portion. Although TRIM56 has RING-dependent E3 ubiquitin ligase function, it is not involved in N^pro-mediated IRF3 degradation nor degrade N^pro. Furthermore, it was suggested that both ectopically and endogenously expressed TRIM56 contribute to impaired replication of BVDV due to its E3 ligase activity. In contrast, the downregulation of TRIM56 expression largely improved BVDV proliferation. Moreover, it is the integrity of the TRIM56 C-terminal, rather than the TRIM56-N^pro association that contributes to TRIM56’s antiviral activity (91).

To discover host proteins that could bind to N^pro of CSFV, Doceul, et al. (58) conducted a yeast two-hybrid assay of a human library. It was revealed that N^pro had a direct association with IκBα, which is responsible for apoptosis regulation, the immune reaction, and IFN expression. As an inhibitor of NF-κB, IκBα is also a prime target for immune evasion strategies developed by many viruses (93–95). Further studies suggested the interaction of N^pro with aa 213-317 of the C-terminus of pIκBα (pig IκBα) (58), which also contact with NF-κB through the domain between aa214-280 (96, 97). This suggests that N^pro competes with NF-κB for unbound pIκBα (58).

Before stimulation, NF-κB remains in an inactive state in the cytoplasm due to its interaction with IκBα, which covers the nuclear localization signals of NF-κB. It has been reported that NF-κB/IκBα complex could be triggered by phosphorylation upon various stimulation, such as viruses and bacteria (98–101). In this case, IκBα is phosphorylated at Ser32 and Ser36 by the IKKβ subunit following the activation of the IKK complex (IKKα/IKKβ/IKKγ). Then, the E3 ubiquitin ligase complex, SCFβ−TRCP, ubiquitinates IκBα and targets it for degradation by the 26S proteasome, resulting in the release of NF-κB for nuclear translocation (102–105). However, the activated NF-κB initiates regeneration of IκBα, which detaches NF-κB from DNA after its translocation to the nucleus, and conveys NF-κB to the cytoplasm in a nuclear export sequence-dependent process (106–108). Tumor necrosis factor-alpha receptor (TNFR) activated by binding with TNF-α is one of the principal receptors that mediate NF-κB activation (109).

It’s reasonable to speculate that under stimulation of TNF-α, new synthesized cytoplasmic IκBα induced by NF-κB activation could bind to N^pro in cells expressing N^pro proteins. Therefore, limited unbound IκBα translocate into the nucleus, and thus the suppression of NF-κB DNA-binding activity by IκBα should be restricted ( Figure 2 ). It has been suggested that HIV-1 tat transactivator could activate NF-κB by interacting with IκBα and by inhibiting the repressor from binding to the NF-κB complex (110). However, the ability of the p65 subunit of NF-κB to bind the promoter sequence in CSFV-infected PK15 cells was not affected by functional analysis (111). Furthermore, after TNF-α stimulation of N^pro stable expression PK15 cells, a high concentration of pIκBα was observed in the nucleus, but the function and expression of NF-κB did not change significantly (58). Therefore, TNF-α may stimulate the rapid resynthesis and massive nuclear translocation of pIκBα, many of which are bound to N^pro and does not affect the action of NF-κB thus resulting in the accumulation of a large amount of pIκBα in the nucleus. However, the effect of N^pro binding to pIκBα on the activity of NF-κB in the nucleus and cytoplasm is worth further investigation.

HS-1-associated protein X-1 (HAX-1), a protein of approximately 35-kDa, is universally synthesized in murine and human tissues (112, 113), especially in the mitochondria (114). Primarily, it was shown to play a role in the control of apoptosis or programmed cell death (114). A recent study has indicated that HAX-1 may also contribute to the control of calcium homeostasis and cell survival in cardiac tissue (115). It was found that the protein could bind to the N^pro protein of CSFV by yeast two-hybrid, and the interaction between the two proteins was further confirmed by co-immunoprecipitation assay (116). During CSFV infection, the expression level of HAX-1 did not change significantly, indicating that N^pro interacting with HAX-1 could not degrade it. However, in the cells co-transfected with HAX-1 and N^pro, the two proteins were significantly transported to the endoplasmic reticulum, and the N^pro protein in the nucleus was significantly reduced (116). Significantly, the relocation of HAX-1 to the ER in the presence of phospholamban (PLN), a crucial regulator of Ca²⁺ homeostasis and contractility in the heart, correlated with stronger resistance to apoptosis (117). Therefore, it could be hypothesized that during CSFV infection, N^pro protein may bind to HAX-1 in the cytoplasm and transport it to ER to enhance the tolerance of infected cells to apoptosis. Thus, less N^pro protein synthesized in the cytoplasm diffuses into the nucleus. However, further research is needed to confirm this hypothesis.

As members of the K homology (KH) domain superfamily, Poly(C)-binding proteins (PCBPs) are known for their interaction with both RNA and DNA specifically. It has been suggested that the KH domain superfamily proteins associate with the stability of cellular mRNAs (118–120), regulate their translation (121–125), and also involve in the host antiviral reaction (126, 127). Among the PCBPs, PCBP1 is an ssDNA-binding protein that contributes to the transcription of the neuronal μ-opioid receptor gene (122). Cytoplasmic and nuclear expression of CBP1 been demonstrated (128). Li et al. found that PCBP1 protein could interact with N^pro protein. Knocking down the expression of PCBP1 could inhibit the replication of CSFV, while overexpression of PCBP1 could promote the reproduction of CSFV. PCBP1 inhibits the IFN signaling pathway by degrading MAVS and enhances the replication of CSFV genomic RNA, thus promoting the proliferation of CSFV (129). However, whether the interaction between PCBP1 and N^pro has any effect on their cellular localization or CSFV replication needs further exploration.