The African swine fever virus I10L protein inhibits the NF-κB signaling pathway by targeting IKKβ

ABSTRACT Proinflammatory factors play important roles in the pathogenesis of African swine fever virus (ASFV), which is the causative agent of African swine fever (ASF), a highly contagious and severe hemorrhagic disease. Efforts in the prevention and treatment of ASF have been severely hindered by knowledge gaps in viral proteins responsible for modulating host antiviral responses. In this study, we identified the I10L protein (pI10L) of ASFV as a potential inhibitor of the TNF-α- and IL-1β-triggered NF-κB signaling pathway, the most canonical and important part of host inflammatory responses. The ectopically expressed pI10L remarkably suppressed the activation of NF-κB signaling in HEK293T and PK-15 cells. The ASFV mutant lacking the I10L gene (ASFVΔI10L) induced higher levels of proinflammatory cytokines production in primary porcine alveolar macrophages (PAMs) compared with its parental ASFV HLJ/2018 strain (ASFVWT). Mechanistic studies suggest that pI10L inhibits IKKβ phosphorylation by reducing the K63-linked ubiquitination of NEMO, which is necessary for the activation of IKKβ. Morever, pI10L interacts with the kinase domain of IKKβ through its N-terminus, and consequently blocks the association of IKKβ with its substrates IκBα and p65, leading to reduced phosphorylation. In addition, the nuclear translocation efficiency of p65 was also altered by pI10L. Further biochemical evidence supported that the amino acids 1–102 on pI10L were essential for the pI10L-mediated suppression of the NF-κB signaling pathway. The present study clarifies the immunosuppressive activity of pI10L, and provides novel insights into the understanding of ASFV pathobiology and the development of vaccines against ASF. IMPORTANCE African swine fever (ASF), caused by the African swine fever virus (ASFV), is now widespread in many countries and severely affects the commercial rearing of swine. To date, few safe and effective vaccines or antiviral strategies have been marketed due to large gaps in knowledge regarding ASFV pathobiology and immune evasion mechanisms. In this study, we deciphered the important role of the ASFV-encoded I10L protein in the TNF-α-/IL-1β-triggered NF-κB signaling pathway. This study provides novel insights into the pathogenesis of ASFV and thus contributes to the development of vaccines against ASF.

ubiquitination of the NF-κB essential modulator (NEMO) is dramatically reduced by pI10L, followed by the inhibition of the phosphorylation of IKKβ. Besides, pI10L inhibits the association of IKKβ with IκBα and p65, which in turn inhibits the phosphorylation and degradation of IκBα, as well as the phosphorylation and nuclear translocation of p65, leading to reduced expression of proinflammatory cytokines. Taken together, our findings reveal the immunomodulatory activity of pI10L, which will help better illustrate the immune evasion mechanisms and pathogenesis of ASFV. These efforts will contrib ute to the development of vaccines against ASF as well as therapeutics to treat the disease.

The ASFV pI10L inhibits the TNF-α-and IL-1β-triggered activation of the NF-κB signaling pathway
To identify candidate molecules involved in virus-induced inflammatory responses, we screened 179 ASFV proteins for their ability to regulate the TNF-α-triggered activation of the NF-κB signaling in reporter assays. These efforts led to the identification of the ASFV pI10L (Fig. 1A). Further experiments in HEK293T cells indicated that pI10L inhibited the TNF-α-and IL-1β-triggered activation of the NF-κB promoter in a dose-dependent manner (Fig. 1B). To investigate whether pI10L is involved in the regulation of endoge nous NF-κB signaling, we measured the transcription levels of several proinflammatory cytokines following treatment with TNF-α and IL-1β in the pI10L-expressing HEK293T cells. The ectopically expressed pI10L remarkably inhibited the mRNA transcription levels of the CCL20, IL-8, and TNF-α genes induced by TNF-α and IL-1β, as demonstrated by quantitative RT-PCR (RT-qPCR) (Fig. 1C). Consistently, the phosphorylation of IκBα and p65 induced by TNF-α and IL-1β, which are the hallmarks of the activation of downstream signaling components, was remarkably inhibited in the pI10L-expressing HEK293T cells compared with the empty vector (Vec)-transfected cells (Fig. 1D). These results suggest that pI10L is involved in regulating the expression of proinflammatory cytokines triggered by TNF-α and IL-1β.
Considering that pigs are the only mammalian species known to be susceptible to ASFV, a PK-15 cell line stably expressing pI10L was established to further identify its function. Immunoblotting analysis demonstrated that pI10L was expressed in PK-15 cells ( Fig. 2A). Confocal microscopy indicated that pI10L was expressed in both the cytoplasm and nucleus (Fig. 2B). In agreement with results from HEK293T cells, RT-qPCR results showed that the ectopically expressed pI10L dramatically inhibited the TNF-α-and IL-1βinduced transcription of the CCL2, IL-8, and TNF-α genes in PK-15 cells (Fig. 2C). Moreover, pI10L reduced the phosphorylation of IκBα and p65, followed by delayed degradation of IκBα in PK-15 cells (Fig. 2D). Taken together, the results suggest that pI10L inhibits the TNF-α-and IL-1β-triggered NF-κB signaling pathway.

Deletion of the I10L gene from ASFV results in enhanced activation of the NF-κB signaling pathway
To characterize the functional role of pI10L in the regulation of the NF-κB signaling pathway, recombinant ASFV ΔI10L was generated from highly virulent ASFV WT by homologous recombination. The I10L gene was replaced with a fluorescent gene EGFPcontaining cassette under the control of the ASFV p72 promoter (Fig. 3A). ASFV ΔI10L was generated and purified after 10 rounds of fluorescence screening (Fig. 3B). The recombi nant ASFV without parental ASFV contamination was confirmed by diagnostic PCR (Fig.  3C). The results of hemadsorption assay revealed "rosettes" of red blood cells on the ASFV ΔI10L -or ASFV WT -infected PAMs, indicating that the deletion of I10L did not affect the hemadsorption property of ASFV (Fig. 3D). Furthermore, the growth kinetics of ASFV ΔI10L were similar to those of ASFV WT depending on the time point considered (Fig.  3E), indicating that the I10L gene deletion did not affect the replication of ASFV in PAMs.
To verify the function of pI10L, we next examined the abilities of ASFV ΔI10L in regulating the activation of the NF-κB signaling pathway. ASFV ΔH240R , a recombinant   Fig. 3F, the deletion of the I10L or H240R gene resulted in enhanced transcription of proinflammatory cytokines compared with ASFV WT upon treatment with TNF-α in PAMs, and the data of pH240R were consistent with the previous studies (23,24,30). Consistently, the ASFV ΔI10Linfected PAMs showed elevated phosphorylation levels of IκBα and p65 (Fig. 3G). These results suggest that pI10L plays an important role in the evasion of antiviral responses to ASFV.

The ASFV pI10L is associated with IKKβ, NEMO, and NF-κB
It has been previously reported that pI10L functions in the TNF-α-and IL-1β-triggered NF-κB signaling, suggesting that this protein probably works at TAK1/TABs complex or its downstream level, which is the convergence of two pathways (19). To identify the targets involved in pI10L function, various components involved in NF-κB signaling (including  TRADD, MYD88, TRAF6, TAK1, TAB1, IKKβ, and p65) were co-expressed with pI10L. As shown in Fig. 4A, pI10L inhibited the activation of the NF-κB promoter mediated by all tested molecules upstream of p65, indicating that pI10L may function at the p65 level. In addition, transient transfection and co-immunoprecipitation (co-IP) experiments confirmed that pI10L interacted with IKKβ, NEMO, p65, and p50, but not with TAK1, TAB2, TAB3, or IKKα ( Fig. 4B and C). The results of GST pulldown assays further confirmed that pI10L was directly associated with IKKβ, NEMO, p65, and p50 directly in vitro (Fig. 4D). In addition, endogenous co-IP experiments indicated that pI10L was constitutively associated with IKKβ, NEMO, p65, and p50 in PK-15 cells. Moreover, this interaction was not affected by TNF-α (Fig. 4E). Consistently, confocal microscopy showed that pI10L colocalized with IKKβ, NEMO, p65, and p50 mainly in the cytoplasm (Fig. 4F). These results suggest that pI10L is associated with IKKs and NF-κB, and may function at the p65 level.

The ASFV I10L protein regulates the activation and the association of IKKβ with IκBα and p65
IKKβ is the principal kinase responsible for phosphorylating IκBα (31) and p65 (32), and IKKβ phosphorylation plays a central role in IKK complex activation. This study indicated that the TNF-α-triggered phosphorylation of IKKβ was significantly prevented by the ectopically expressed pI10L (Fig. 5A). It has been reported that NEMO, the regulatory subunit of IKK complex, can be ubiquitinated by TRAF6 and TRAF2/5, and this K63-linked polyubiquitination is essential for the activation of IKKβ (33). To determine how pI10L regulates IKKβ activation, the K63-linked ubiquitination (Ub) of NEMO was examined.
Western blot analysis showed that Ub-K63 linked conjugation to NEMO was decreased in the presence of pI10L (  ASFV pI10L inhibits the TNF-α-and IL-1β-triggered activation of the NF-κB promoter in a dose-dependent manner. HEK293T cells were transfected with the pI10L-expressing plasmid at different concentrations (0, 0.01, 0.02, or 0.04 µg) along with pNF-κB-Fluc (0.1 µg) and pRL-TK (0.01 µg). Twenty hours later, the luciferase assay and immunoblotting analysis were performed after the cells were treated with TNF-α (10 ng/mL) or IL-1β (10 ng/mL) for 10 hours. (C) The ASFV pI10L inhibits the TNF-α-and IL-1β-triggered the transcription of proinflammatory cytokines. HEK293T cells were transfected with either the pRK (Vec) or the pFlag-pI10L (0.2 µg) for 20 hours, and RT-qPCR was performed following treatment with TNF-α (10 ng/mL) or IL-1β (10 ng/mL) for 6 hours. (D) The ASFV pI10L inhibits the TNF-α-and IL-1β-triggered phosphorylation of IκBα and p65. HEK293T cells were transfected with either the Vec or the pFlag-pI10L (0.2 µg) for 20 hours. The cells were then treated with TNF-α (20 ng/mL) or IL-1β (20 ng/mL) for 5, 10, or 20 minutes, followed by immunoblotting analysis. Densitometric analysis of protein expression level was performed with ImageJ software. The data shown are the mean ± SD from one representative experiment performed in triplicate (A to C). ***P < 0.001 (unpaired t test). It has been shown that the assembly of the IKK complex is indispensable for IKKβ activation (34). The results demonstrated that pI10L had no effect on the expression or dimerization of NEMO ( Fig. 6A and B), nor did it affect the association between IKKβ and NEMO (Fig. 6C). Thus, pI10L does not regulate the assembly of the IKK complex assembly. Competitive binding experiments were subsequently carried out to explore whether pI10L affects the assembly of NF-κB complexes, as well as the association between IKKβ and its substrates IκBα and p65. As shown in Fig. 6D, pI10L is not involved in the binding of p65 and p50. Interestingly, pI10L remarkably impaired the interaction between IKKβ and IκBα or p65 ( Fig. 6E and F). In addition, upon treatment with increasing concentra tions of pI10L, the IKKβ-mediated phosphorylation of IκBα and p65 was decreased in a dose-dependent manner (Fig. 6G). To investigate whether pI10L directly regulates the kinase activity of IKKβ, an in vitro kinase assay was performed. Immunoblotting analysis showed that IκBα and p65 were phosphorylated by IKKβ in the presence of ATP, whereas the addition of pI10L dramatically impaired this phosphorylation process (Fig. 6H). Taken together, our findings show that pI10L functions by regulating K63-linked polyubiquiti nation of NEMO to suppress the activation of IKKβ. Furthermore, pI10L is associated with IKKβ, which in turn suppresses its kinase activity towards IκBα and p65, thereby inhibiting the activation of the NF-κB signaling pathway.

The ASFV I10L protein inhibits the nuclear translocation of p65
It has been shown that the nuclear translocation of p65, which is released after IκBα is degraded upon phosphorylation by IKKβ, is an indicator of the NF-κB signaling activation (20). Therefore, we examined whether p65 is translocated to the nucleus in the presence of pI10L. The results of subcellular fractionation assay and confocal microscopy analysis showed that the TNF-α-induced nuclear translocation of p65 was inhibited in the pI10Lexpressing cells ( Fig. 7A and B). Furthermore, the results indicated that the deficiency of pI10L attenuated the ability of ASFV to antagonize the TNF-α-induced nuclear transloca tion of p65 (Fig. 7C). The results indicated that the deletion of pI10L weakened the ability of ASFV to block the activation of the NF-κB signaling pathway.

Amino acids 1-102 on pI10L are essential for suppressing NF-κB activation
To investigate the crucial regions in IKKβ and pI10L responsible for their interaction, as well as the suppression of NF-κB activation, we constructed a series of plasmids express ing the complete or truncated IKKβ (Fig. 8A) and pI10L (Fig. 8B). Domain mapping analysis revealed that the kinase domain of IKKβ was responsible for its interaction with pI10L ( Fig. 8C). Consistently, the truncated pI10L containing aa 1-102 was sufficient to bind to IKKβ (Fig. 8D). These results further confirm that the binding of pI10L to IKKβ ultimately suppresses the catalytic activity of IKKβ.
We next evaluated whether the aa 1-102 of pI10L are sufficient to inhibit the TNF-αtriggered activation of the NF-κB signaling. In reporter assay, all the mutants containing aa 1-102 were found to remarkably inhibit the activation of the NF-κB promoter (Fig. 8E). RT-qPCR experiments demonstrated that the ectopic expression of pI10L or pI10L(aa1-102) dramatically reduced the transcript level of the CCL2, IL-8, and TNF-α protein or harboring the empty vector were lysed and the expression of pI10L was measured through immunoblotting. (B) The ASFV pI10L is localized in both the cytoplasm and the nucleus of the infected cells. The PK-15 cell lines stably expressing the ASFV pI10L or harboring Vec were fixed with 4% paraformaldehyde and subjected for confocal microscopy. (C) The ASFV pI10L inhibits the TNF-α-and IL-1β-triggered transcription of proinflammatory cytokines in PK-15 cells. The PK-15 cell lines stably expressing the ASFV pI10L or harboring Vec were treated with TNF-α (10 ng/mL) or IL-1β (10 ng/mL) for 10 hours, and then total RNA was prepared for RT-qPCR assay. (D) The ASFV pI10L inhibits the TNF-α-and IL-1β-triggered phosphorylation of IκBα and p65 in PK-15 cells. The PK-15 cell lines stably expressing the ASFV pI10L or harboring Vec were treated with TNF-α (20 ng/mL) or IL-1β (20 ng/mL) for 5 or 10 minutes, after which the cells were lysed and immunoblotting analysis was performed. Densitometric analysis of protein expression level was performed with ImageJ software. Data shown are the mean ± SD from one representative experiment performed in triplicates. ***P < 0.001 (unpaired t test).

Full-Length Text
Journal of Virology genes in PK-15 cells (Fig. 8F). Consistently, the phosphorylation of IκBα and p65 (Fig. 8G), as well as the nuclear translocation of p65 (Fig. 8H), were remarkably impaired in the pI10L-or pI10L(aa1-102)-expressing PK-15 cells. Taken together, our results suggest that the domain defined by aa 1-102 on pI10L is essential for the suppression of NF-κB activation.

DISCUSSION
The innate immune response is the first line of host defense against viral infection and is initiated upon sensing conserved viral structural components called pathogen-associ ated molecular patterns (PAMPs) by pattern recognition receptors (PRRs) in host cells (35)(36)(37). The sensing of viral PAMPs by PRRs activates a series of signaling events, leading to the expression of downstream antiviral effector proteins including type I interferons and proinflammatory cytokines (37)(38)(39). ASFV is a giant, complex DNA virus that encodes more than 160 proteins required for successful infection, thereby enabling replication and immune evasion in vivo. Previous studies have demonstrated that ASFV infection activates antiviral signaling pathways with increased expression levels of interferonstimulated genes and proinflammatory cytokines (15), especially the NF-κB signaling (40). However, the mechanisms of ASFV immune evasion remain unclear. In this study, we identified the ASFV pI10L as an inhibitor of virus-induced inflammatory responses following viral infection and treatment with TNF-α or IL-1β. We found that transient transfection of I10L in HEK293T cells remarkably inhibited the activation of the NF-κB promoter, transcription of proinflammatory cytokines, and the phosphorylation of IκBα and p65 induced by TNF-α or IL-1β. We obtained the same results as in the PK-15 cells stably expressing pI10L. These findings suggest that pI10L plays an important role in the immune escape of ASFV in different cell types.
Considering the unbiased effects on the TNF-α-and IL-1β-triggered NF-κB signaling, we hypothesized that pI10L functions at the level of or downstream of the TAK1-TABs complex. Interestingly, pI10L inhibited p65 and all the tested proteins upstream of the p65-mediated activation of the NF-κB signaling pathway. In addition, pI10L was found to be associated with IKKβ, NEMO, and NF-κB. There are several possible mechanisms underlying how pI10L functions. Firstly, pI10L inhibits the activation of IKKβ by impairing the assembly of the IKK complex or the polyubiquitination of NEMO. Secondly, it is possible that pI10L suppresses the binding of p65 to p50. Thirdly, pI10L could block the activation and nuclear translocation of p65. The results demonstrated that pI10L inhibits the phosphorylation of IKKβ by regulating the K63-linked ubiquitination of NEMO, as well as reduced the catalytic activity of towards IκBα and p65 by interacting with IKKβ. However, pI10L only slightly interfered with the assembly of the NF-κB or IKK complex.
It has been reported that IκBα and p65 are phosphorylated by IKKβ in the TNF-α-and IL-1β-triggered NF-κB signaling pathway (41). Therefore, we performed kinase assays in vitro. The results confirmed that IKKβ could phosphorylate IκBα and p65, but this effect was remarkably inhibited by pI10L, owing to the fact that pI10L interrupted the binding of IκBα and p65 to IKKβ. Using subcellular fractionation assay and confocal microscopy analysis, we found that ectopically expressed pI10L dramatically inhibited the translocation of p65 to the nuclear. Domain mapping analysis indicated that pI10L interacts with the kinase domain of IKKβ through its N-terminal (aa 1-102). Together, these findings demonstrate that pI10L inhibits IKKβ kinase activity through direct association with IKKβ, obstructing its catalytic center, or by competing with IκBα and p65 as a substrate. Confirming the exact mechanism will be a very interesting direction for future investigation.
Furthermore, the effects of pI10L on the expression of proinflammatory cytokines were explored using the I10L gene-deleted ASFV. The results showed that the hemad sorption and replication of ASFV were slightly affected by pI10L in PAMs. However, ASFV ΔI10L induced higher activation of the NF-κB signaling pathway than did ASFV WT . To date, several other ASFV-encoded proteins that regulate the activation of NF-κB have been identified, such as the pL83L (26), pMGF505-7R (22,42), and the H240R protein (pH240R) (23,30). ASFV lacking these genes induces decreased expression of various proinflammatory cytokines and increased viral replication, whereas the relationship among these regulators remains unclear. Single gene deletion hardly affects the virulence of ASFV (except for pMGF505-7R and pH240R), which may be because these proteins function in concert during viral infection and pathogenesis. For example, the proteins may participate in the inflammatory response during different stages of viral infection. To hasten the development of an effective vaccine against ASF, further indepth studies are needed to explore how these proteins work together.
In conclusion, as illustrated in Fig. 9, our findings demonstrated a critical role of pI10L in regulating the TNF-α-and IL-1β-triggered NF-κB signaling by targeting IKKβ. As the inflammatory response plays an important role in the pathogenesis of ASF (15, 16) and host antiviral innate immunity, the identification of pI10L helps clarify the complicated immune evasion mechanisms of ASFV and may lead to the discovery of potential targets for the development of novel ASF vaccines or antivirals.
Full-Length Text 48 hours, the pseudovirus-containing culture supernatant was harvested to infect PK-15 cells in the presence of polybrene (8 µg/mL). The infected cells were screened with puromycin (3 µg/mL) for 6 days to establish stable cell lines. The primers used in this study are listed in Table 1. Densitometric analysis of protein expression levels were performed with ImageJ software. Data shown are the mean ± SD from one representative experiment performed in triplicates. ***P < 0.001 (unpaired t test).

Transfection and reporter assays
HEK293T and PK-15 cells were transfected using the standard polyetherimide method. In the reporter assays, an empty control plasmid was added separately to ensure the same HEK293T cells were transfected with various concentrations of the pI10L-expressing plasmid followed by immunoblotting analysis. (B-F) The ASFV pI10L inhibits the interaction between IKKβ and IκBα or p65, but does not regulate the assembly of the NF-κB or IKK complex. HEK293T cells were transfected with the plasmids: pHA-NEMO, -p65, or -p50, pHis-pI10L, pMyc-NEMO, pFlag-IKKβ, or -p65 for 24 hours. Myc-NEMO, Flag-IKKβ or Flag-p65 was then used as bait protein to perform co-IP and immunoblotting analysis. (G) The ASFV pI10L inhibits the IKKβ-mediated phosphorylation of IκBα and p65 in a dose-dependent manner.

FIG 9
Schematic diagram for the mechanism by which the ASFV I10L protein suppresses the host cell proinflammatory responses. The ASFV pI10L can inhibit the TNF-α-and IL-1β-triggered inflammatory responses by targeting IKKβ. pI10L inhibits the phosphorylation of IKKβ by reducing the K63-linked ubiquitination of NEMO and hinders the association of IKKβ with its substrates IκBα and p65, leading to reduced phosphorylation of IκBα and p65, as well as the nuclear translocation of p65, and subsequently the expression of proinflammatory cytokines.
Full-Length Text amount of total DNA was present in a simple sample. Luciferase assays were performed using a dual-specific luciferase assay kit (catalog no. E1910; Promega). The data shown are the firefly luciferase activity levels of the indicated samples normalized to Renilla luciferase activity.

RT-qPCR
Total RNA was extracted using the TRIzol Reagent (catalog no. 9108; TaKaRa) and reverse-transcribed to cDNA using the HiScript III 1st Strand cDNA Synthesis Kit (catalog no. R312-01; Vazyme) according to the manufacturer's protocol. qPCR was performed in triplicate using HiScript II Q RT SuperMix (catalog no. R223-01; Vazyme) according to the manufacturer's protocol. Data shown are the relative abundance of the indicated mRNAs normalized to those of GAPDH. The primers used for RT-qPCR are listed in Table 2.

GST pulldown assay
To prepare Escherichia coli for pI10L expression, the recombinant plasmid pGEX-6p-1-pI10L was transformed into E. coli (BL21) cells. The cells in the logarithmic phase were then incubated with IPTG (0.8 mmol/L) for 12 hours at 20°C. After centrifugation, enriched bacterial cultures were resuspended and lysed using a high-pressure homoge neous sterilizer. The supernatants containing recombinant GST or GST-pI10L protein were purified using ChromoTek GST-Trap Agarose beads (catalog no. sta; Proteintech). GST and GST-pI10L were incubated with cell lysates containing the ectopically expressed IKKβ, NEMO, p65, or p50 at 4°C for 12 hours. Subsequently, the bound proteins were separated by SDS-PAGE, followed by immunoblotting analysis.
(Continued on next page)

Subcellular fractionation
To separate the nuclear fraction (Nuc) and cytoplasmic fraction (Cyt), PK-15 cell lines or PAMs under the treatment of TNF-α were washed three times with PBS and once in hypotonic buffer (10 mM Tris-HCl pH 7.4, 10 mM KCl, 1.5 mM MgCl 2 ) (47) supplemented with the protease inhibitor PMSF, resuspended in hypotonic buffer, and lysed by leaching homogenization. The lysates were centrifuged at 4°C for 10 minutes at 500× g to obtain Nuc pellets, and the supernatants were collected as Cyt supernatant. Nuc pellets were lysed in RIPA buffer with an equal volume of Cyt supernatant, and subjected to western blot analysis with mouse anti-Flag or anti-β-actin (catalog no. CL594-66009; Proteintech) MAb, rabbit anti-lamin B1 PAb (catalog no. 12987-1-AP; Proteintech), or rabbit anti-NF-κB p65 MAb.

Generation and identification of the I10L-deleted ASFV mutant
The recombinant transfer vector pOK12-p72EGFP-ΔI10L, which harbors genomic sequences flanking the targeted gene mapping approximately 1.2 kb upstream and downstream homologous arms, and a reporter gene cassette containing the EGFP reporter under the control of the ASFV p72 gene promoter, was constructed. The left and right arms flanking the target gene were located in the ASFV WT genome at positions 180919-182119 and 182632-183832, respectively. The nucleotides in the genome at positions 182120-182632 were replaced with an expression cassette containing the EGFP reporter. Briefly, the left and right arms were amplified using PCR and assembled to contain an EGFP reporter harboring restriction enzyme sites at both termini using overlapping PCR. The cassette was cloned into the linearized pOK12 vector to generate the recombinant transfer vector pOK12-p72EGFP-ΔI10L using the Vazyme ClonExpress II one step cloning kit (Vazyme Biotech Co., Ltd., China). The I10L-deleted ASFV mutant ASFV ΔI10L was generated by homologous recombina tion between the ASFV WT genome and the recombination transfer vector, using infection and transfection procedures in PAMs. This construct generated the expected deletion of the I10L gene. The virus resulting from homologous recombination was purified using successive limiting dilutions of PAMs. The purified ASFV ΔI10L was amplified in PAMs to produce a viral stock. To ensure the absence of the desired deletion in each recombinant genome, viral DNA was extracted from ASFV ΔI10L -infected PAMs and identified by PCR using specific primers targeting these genes. The primers used in this study are listed in Table 2.

Virus growth curve
PAM monolayers were prepared in 24-well plates and infected with ASFV ΔI10L or ASFV WT at an MOI of 0.01. After 1 hour of adsorption, the cells were rinsed twice with PBS. The monolayers were incubated in the medium for 2, 12, 24, 48, 72, 96, or 120 hours. At different time points, the ASFV-infected cultures were stored at −70°C. Subsequently, the thawed lysates were used to determine the viral titers in HAD 50 /mL in PAMs.

Hemadsorption assay
Approximately 5 × 10 4 PAMs were seeded in 96-well plates and infected with ASFV ΔI10L or ASFV WT for 48 hours. The cells were then incubated with 5 × 10 5 porcine red blood cells diluted in PBS, and hemadsorption was observed on the fifth day.

Statistics analysis
GraphPad Prism and SPSS Statistics were used for statistical analysis. Quantitative data in histograms are shown as means ± SD. The data were analyzed using the Log-rank (Mantel-Cox) test or the unpaired Student's t test. The number of asterisks represents the degree of significance with respect to P values. Statistical significance was set at P < 0.05. P values are indicated by asterisks in the figures as follows: *P < 0.05; **P < 0.01; ***P < 0.001.

ACKNOWLEDGMENTS
We thank Prof. Hong-Bing Shu for providing the following essential materials: luciferase reporters, pRK vector, and HEK293T cells. We thank Prof. Youbao Zhao for valuable advice on writing the manuscript. This work was supported by grants from the National Natural Science Foundation of China (grant numbers 32102655, 32072855, 32272987, 31941001, and 32002278). The authors declare no competing financial interests.