The MGF300-2R protein of African swine fever virus is associated with viral pathogenicity by promoting the autophagic degradation of IKKα and IKKβ through the recruitment of TOLLIP

The multigene family genes (MGFs) in the left variable region (LVR) of the African swine fever virus (ASFV) genome have been reported to be involved in viral replication in primary porcine alveolar macrophages (PAMs) and virulence in pigs. However, the exact functions of key MGFs in the LVR that regulate the replication and virulence of ASFV remain unclear. In this study, we identified the MGF300-2R gene to be critical for viral replication in PAMs by deleting different sets of MGFs in the LVR from the highly virulent strain ASFV HLJ/18 (ASFV-WT). The ASFV mutant lacking the MGF300-2R gene (Del2R) showed a 1-log reduction in viral titer, and induced higher IL-1β and TNF-α production in PAMs than did ASFV-WT. Mechanistically, the MGF300-2R protein was found to interact with and degrade IKKα and IKKβ via the selective autophagy pathway. Furthermore, we showed that MGF300-2R promoted the K27-linked polyubiquitination of IKKα and IKKβ, which subsequently served as a recognition signal for the cargo receptor TOLLIP-mediated selective autophagic degradation. Importantly, Del2R exhibited a significant reduction in both replication and virulence compared with ASFV-WT in pigs, likely due to the increased IL-1β and TNF-α, indicating that MGF300-2R is a virulence determinant. These findings reveal that MGF300-2R suppresses host innate immune responses by mediating the degradation of IKKα and IKKβ, which provides clues to paving the way for the rational design of live attenuated vaccines to control ASF.

Introduction African swine fever virus (ASFV) is a large, double-stranded DNA virus that belongs to the genus Asfivirus in the family Asfarviridae. ASFV is the causal agent of African swine fever (ASF), a hemorrhagic and often lethal disease of domestic pigs and wild boars [1]. ASF has become endemic or epidemic in nearly 50 countries in Africa, Europe, Asia, and the Caribbean, causing huge economic losses that are difficult to estimate and posing a critical threat to the global pig industry and related industries [1][2][3]. Currently, only one ASF live attenuated vaccine (LAV) was marketed in Vietnam, but its use is restricted [4]. Importantly, previous experimental studies have revealed that LAV candidates have shown higher levels of protection than other types of ASF vaccines [5]. Hence, it is essential to discover and characterize the virulence-associated genes of ASFV for effective control ASF and future development of strategies for LAVs [5].
The ASFV genome varies in length between approximately 170 and 194 kilobase-pairs (kb), mainly due to variable copies of the multigene family genes (MGFs) from different isolates [6][7][8]. MGFs are mainly located at both ends of the ASFV genome and are referred to as left and right variable regions (LVR and RVR), which are divided into five clusters (MGF100, MGF110, MGF300, MGF360, and MGF505) based on the average amino acid lengths. The MGFs encode proteins that have been reported to be associated with host antiviral immunity modulation, viral virulence, cell tropism, and host range [9,10]. ASFV virulence has been associated with MGF360 and MGF505 genes which are non-essential for virus replication in primary porcine alveolar macrophages (PAMs). The deletion of six genes of MGF360 (12L, 13L, and 14L) and MGF505 (1R, 2R, and 3R) together led to decreased virulence in pigs, but the individual contribution of each MGF to attenuation remains unknown [11]. Recent studies have shown that the MGF505-7R protein can antagonize the production of interferon (IFN) and inflammatory cytokines in PAMs and pigs [12]. More importantly, the deletion of the MGF505-7R gene from a virulent strain partially attenuated its virulence in pigs [13][14][15][16]. The pandemic of ASF highlights the importance of studying the molecular mechanism of individual MGFs involved in pathogenesis, which is still largely unknown [17,18].
In mammals, the nuclear factor kappa B (NF-κB) family comprises a group of related transcription factors that control the expression of hundreds of genes associated with diverse cellular host responses, such as inflammation, programmed cell death, cell growth, and survival [19]. In most cells, the NF-κB complex remains inactive predominantly in the cytoplasm through binding to the inhibitor kappa-B alpha (IκBα). Toll-like receptors (TLRs) and IL-1 receptors transmit signals to the IκB kinase (IKK) complex, which consists of NEMO and two kinases, IKKα and IKKβ, via IRAK1 and TRAF6 [20]. In the canonical NF-κB pathway, various stimuli trigger the phosphorylation of IκBα via IKKβ, leading to IκBα ubiquitination and subsequent proteasomal degradation [21]. The released NF-κB-p65 is translocated to the nucleus, binds κB-responsive DNA sequences, recruits transcription co-regulators, and modulates hundreds of gene expressions. NF-κB plays a critical role in many cellular responses, and it is not surprising that most viruses have evolved strategies to subvert the NF-κB signaling pathway [20,21]. For example, vaccinia virus (VACV) encodes at least eleven NF-κB antagonists, including A46R, A49, A52, B14, C4, F14, E3, K1L, K7, M2L, and N1L, and can inhibit the NF-κB activation at different levels [21][22][23]. Similarly, ASFV encodes several proteins, including H240R, A238L, F317L, MGF505-7R, and D345L that can inhibit the production of interleukin 1beta (IL-1β) and IFNs by modulating NF-κB activation [24][25][26][27][28]. However, the exact impact of these proteins on the pathogenesis of ASFV is still not fully understood.
We demonstrated previously that the cell-adapted ASFV with a 24.5-kb genes deletion in the LVR lost the ability to replicate in PAMs [34]. Recently, we found that the deleted LVR genes of the cell-adapted ASFV may contain critical MGFs for ASFV replication in PAMs [35]. In this study, we reveal that the MGF300 genes (1L, 2R, and 4L) encoding proteins are critical for the replication of ASFV in PAMs. Additionally, we have demonstrated that MGF300-2R functions as an inhibitor of the NF-κB signaling pathway by specifically targeting IKKα and IKKβ for autophagic degradation, thereby preventing the production of IL-1β and TNF-α in vitro and in vivo. Deletion of the MGF300-2R gene resulted in partial attenuation of the ASFV in pigs, as evidenced by reduced viral replication and increased production of IL-1β and TNFα. Our findings demonstrate an important mechanism of MGF300-2R hijacks the NF-κB signaling pathway to facilitate ASFV pathogenesis.

Results
The MGF300 genes (1L, 2R, and 4L) are critical to ASFV replication in PAMs To identify the individual MGFs deleted from ASFV-WT that are critical to ASFV replication in PAMs, we generated six ASFV mutants with different combinations of MGFs deletion in LVR using homologous recombination (Fig 1A). The MGFs-deleted recombinant ASFVs were purified via successive limited dilution by selecting EGFP-positive PAMs. The purity of these MGFs-deleted recombinant ASFVs was confirmed by PCR (S1A Fig). Furthermore, the wholegenome sequence of DelLVR was determined by NGS as described previously [35]. Next, the replication kinetics of the MGFs-deleted recombinant ASFVs were compared with that of ASFV-WT in PAMs, and it was found that ASFV mutants DelNo1, DelNo2, DelNo4, and DelNo5 exhibited similar growth kinetics to ASFV-WT. However, lower replication levels were observed for DelNo3 and DelLVR between 3 to 7 dpi compared with ASFV-WT (Fig 1B). These findings suggested that the deletion of the MGF300 genes (1L, 2R, and 4L) significantly decreased viral replication, indicating that these genes are critical for viral replication in PAMs.
To further identify the role of the individual viral genes of MGF300 that regulate viral replication in PAMs, we constructed three recombinant viruses with single deletion of the MGF300-1L (Del1L), MGF300-2R (Del2R), or MGF300-4L (Del4L) gene. The whole-genome sequencing and analysis of the Del1L, Del2R, and Del4L confirmed that all the final recombinant viruses harbored the correct sequences (S1B-S1D Fig and S2 Table). The replication kinetics of Del1L, Del2R, and Del4L were similar but significantly lower than that of ASFV-WT after 3 dpi (Fig 1C). These data confirm that the MGF300 genes are critical to ASFV replication in PAMs. In this study, we proceed to further investigate MGF300-2R as it is a highly conserved protein localized in the cytoplasm and nucleus with nuclear localization signal (S2A-S2D Fig), which may be important for ASFV infection.

The MGF300-2R-deleted ASFV mutant induces higher production of inflammatory cytokines than does the wild-type ASFV in PAMs
To investigate the role of MGF300-2R protein in the ASFV life cycle, we compared the binding, internalization, and the release of virions of Del2R and ASFV-WT, but no significant

PLOS PATHOGENS
MGF300-2R is a virulence factor of ASFV PLOS Pathogens | https://doi.org/10.1371/journal.ppat.1011580 August 11, 2023 differences were observed between the two viruses (S3A and S3B Fig). As MGF300-2R is highly conserved and localized in both the cytoplasm and nucleus (S2C Fig.), it may play a role in ASFV morphogenesis as a structural protein. However, TEM analysis revealed no differences in the ultrastructure of viral factories or virion morphogenesis between Del2R and ASFV-WT (S3C and S3D Fig). Therefore, we conclude that MGF300-2R is not essential for ASFV replication in PAMs.
Next, to further dissect the role of MGF300-2R in ASFV replication, RNA-seq analysis was used to investigate the host immune responses induced by infection with Del2R in PAMs. Our results revealed a total of 4159 differentially expressed genes (DEGs) for Del2R vs. ASFV-WT at 12 hpi (S3 Table), including 1326 upregulated and 2833 downregulated genes, and a total of 4525 DEGs for Del2R vs. ASFV-WT at 20 hpi (S3 Table), including 1007 upregulated and 3518 downregulated genes (Fig 2A). It is of significance to observe that, in comparison with ASFV-WT, the expression level of IL-1β was markedly increased by approximately 9-fold at 12 and 20 hpi (Fig 2A). In comparison to PAMs infected with ASFV-WT, the differentially expressed mRNAs during Del2R infection were mainly involved in innate immunity signaling pathways such as NF-κB, TNF, MAPK, and PI3K-AKT signaling pathways, as well as receptor signaling, metabolic networks, and other pathways associated with viral infection (Fig 2B). The altered cellular mRNAs from NF-κB signaling pathways are presented in Fig 2C, wherein

PLOS PATHOGENS
MGF300-2R is a virulence factor of ASFV IL-1β, TNF-α, and IL-6 were found to be highly transcribed during Del2R infection compared with ASFV-WT at 12, and 20 hpi. To further validate the RNA-seq results, the levels of cytokines IL-1β, TNF-α, and IFN-α were measured by RT-qPCR. Del2R infection significantly increased the expression of IL-1β and TNF-α mRNA by approximately 20-fold at 12 and 20 hpi, compared with ASFV-WT (Fig 2D). Likewise, the secretion of IL-1β and TNF-α induced by Del2R was approximately 20-fold higher than that induced by ASFV-WT at 12 and 20 hpi in PAMs (Fig 2E). However, no differences in IFN-α expression level was observed between the two viruses (S4A and S4B Fig). Altogether, these findings suggest that Del2R activates the inflammatory signaling pathway in PAMs.

MGF300-2R interacts with IKKα and IKKβ and inhibits the NF-κB activation
It is well-known that ASFV has evolved various mechanisms to evade the host antiviral innate immune signaling and promote its replication and pathogenesis [36]. In light of this, we hypothesized that MGF300-2R may facilitate viral replication by counteracting the inflammatory signaling pathways mediated by the NF-κB. To test this hypothesis, we conducted experiments to assess the effects of MGF300-2R on NF-κB activation. We found that the ectopically expressed MGF300-2R inhibited the TNF-α-triggered activation of the NF-κB promoter in a dose-dependent manner (Fig 3A), while not significantly affecting other reporters such as IFN-β (S4C Fig), IRF3 (S4D Fig), and ISRE (S4E and S4F Fig).
Additionally, subcellular fractionation experiments showed that the ectopically expressed MGF300-2R inhibited the nuclear translocation of p65 ( Fig 3B). Next, we investigated the nuclear localization pattern of p65 in the infected PAMs. As shown in Fig 3C, p65 in the nuclear fraction of the Del2R-infected PAMs was significantly increased compared with that of the ASFV-WT-infected PAMs. To further confirm the inhibitory effects of MGF300-2R on the NF-κB signaling, the PAMs infected with Del2R or ASFV-WT were assessed for the phosphorylation of p65 and IκBα, two hallmarks of NF-κB activation. The results showed that infection with Del2R led to more phosphorylation of p65 and IκBα than did ASFV-WT infection at 2 and 6 hpi in PAMs (Fig 3D). Taken together, these results corroborate the conclusion that MGF300-2R negatively regulates the NF-κB-mediated innate immune response.
We then tried to determine the potential mechanism(s) by which MGF300-2R inhibits the NF-κB signaling pathway. Our results showed that MGF300-2R functions upstream of IκBα by inhibiting the phosphorylation of p65 and IκBα, and the production of IL-1β and TNF-α. To investigate whether MGF300-2R can interact with any protein in the IKK complex, we performed co-immunoprecipitation (co-IP) experiments in HEK293T cells. We found that MGF300-2R specifically co-precipitated with both IKKα and IKKβ, but not NEMO (Fig 3E). This interaction was further validated through a GST pulldown assay, in which GST-MGF300-2R pulled down Flag-tagged IKKα and IKKβ (Fig 3F). Confocal microscopy analysis confirmed that MGF300-2R and IKKα or IKKβ were colocalized in the cytoplasm of HEK293T cells (Fig 3G and 3H). Taken together, these results showed that MGF300-2R interferes with the IKK complex to counteract the activation of the NF-κB signaling pathway.

Autophagic degradation of IKKα and IKKβ can be induced by the ASFV MGF300-2R protein
To investigate whether MGF300-2R affects the protein levels of IKKα and IKKβ, we cotransfected HEK293T cells with pHA-MGF300-2R together with pFlag-NEMO, pFlag-IKKα, or pFlag-IKKβ. Our results showed that the protein levels of IKKα (Fig 4A) and IKKβ (Fig 4B) were markedly reduced in a dose-dependent manner, whereas the protein level of NEMO Interestingly, we observed that endogenous IKKα (Fig 4D) and IKKβ (Fig 4E) were decreased in the PAMs infected with ASFV-WT at 12 and 20 hpi, whereas NEMO levels were not affected ( Fig 4F). However, compared with ASFV-WT, the degradation of endogenous IKKα and IKKβ was diminished in the Del2R-infected PAMs at 12 and 20 hpi (Fig 4D and 4E). Additionally, the transcriptional levels of IKKα (S5D

The MGF300-2R protein interacts with IKKα and IKKβ and inhibits the NF-κB activation. (A) HEK293T cells were transfected with increasing
amounts of the expressing plasmid pHA-MGF300-2R (0.5, 1, and 1.5 μg) for 24 h and then mock-treated or treated with TNF-α (10 ng/mL) for 8 h. The protein expression of MGF300-2R was examined by western blotting and the luciferase activity was measured at 24 hours post-transfection (hpt). (B) HEK293T cells were transfected with the pFlag-MGF300-2R or empty plasmid and then treated with or without TNF-α (10 ng/mL) for 1 h. At 24 hpt, the cells were fractionated into cytoplasmic and nuclear fractions and analyzed by immunoblotting with the indicated antibodies. The p65 in the nuclear and cytoplasmic compartments was checked by western blotting. Lamin B1 and tubulin were used as nuclear and cytosolic markers, respectively. (C) PAMs were mock-infected or infected with ASFV-WT or Del2R (MOI = 5). At 20 hpi, PAMs were treated with or without TNF-α (10 ng/mL) for 1 h, the separation of cellular fractions of ASFV-infected PAMs was performed as described above. (D) Analysis of the phosphorylation levels of IκBα and p65 in PAMs mock-infected or infected with ASFV-WT, or Del2R (MOI = 5) by western blotting at 2 and 6 hpi. (E) HEK293T cells were transfected with a plasmid encoding HA-MGF300-2R along with a plasmid encoding Flag-IKKα, Flag-IKKβ, or Flag-NEMO as indicated. The cells were lysed and whole cell lysates (WCL) were immunoprecipitated with anti-HA mAb at 36 hpt. The immunoprecipitates were examined by western blotting with the indicated antibodies. (F) HEK293T cells were transfected with expressing plasmids pFlag-IKKα or pFlag-IKKβ for 36 h and lysed with NP-40 buffer. The purified GST or GST-MGF300-2R protein was used to pull down the IKKα or IKKβ in the lysates and analyzed by western blotting with the indicated antibodies. (G and H) HEK293T cells were transfected with pHA-MGF300-2R alone and cotransfected with pFlag-IKKα or pFlag-IKKβ in combination with pHA-MGF300-2R. IKKα, IKKβ, and MGF300-2R were analyzed by laser confocal microscopy. Scale bar, 10 μm. The colocalization of MGF300-2R and IKKα or IKKβ was analyzed by the Coloc2 tool of ImageJ/FIJI and shown as Pearson's correlation coefficients. https://doi.org/10.1371/journal.ppat.1011580.g003

MGF300-2R is a virulence factor of ASFV
The ubiquitin-proteasome system and the autophagy-lysosome pathway are distinct mechanisms for intracellular protein degradation in eukaryotic cells [32]. Therefore, we further investigated which pathway is utilized by MGF300-2R to mediate the degradation of IKKα and IKKβ. The HEK293T cells were transfected with MGF300-2R and IKKα or IKKβ expression plasmids and treated with protease inhibitors Lac and MG132 or autophagy inhibitors, BafA1, 3-MA, or LY294002. The results showed that treatment with 3-MA, BafA1, and LY294002, but not MG132 and Lac, restored the degradation of IKKα (Fig 4G) and IKKβ (Fig 4H). Moreover, the knockout of ATG5, an essential gene for autophagy activation [37,38], abrogated the MGF300-2R-mediated degradation of IKKα (Fig 4I) and IKKβ (Fig 4J). Collectively, these results demonstrate that MGF300-2R promotes the degradation of IKKα and IKKβ through the lysosomal degradation pathway.

The MGF300-2R protein promotes the autophagic degradation of IKKα and IKKβ by the cargo receptor TOLLIP
Previous experimental studies revealed that SARs play an essential role in defending against virus infection [30]. To investigate the mechanism of MGF300-2R-mediated autophagyinduced degradation of IKKα and IKKβ, we hypothesized that MGF300-2R might act as a bridge between these proteins to a cargo receptor for autophagic degradation. To identify the specific cargo receptor responsible for MGF300-2R-induced degradation of IKKα and IKKβ, we examined the interactions between MGF300-2R, cargo receptors (including p62, NDP52, TOLLIP, and OPTN), IKKα and IKKβ. Co-IP assays revealed that MGF300-2R interacted with the cargo receptors p62 and TOLLIP (Fig 5A). Interestingly, we found that both p62 and TOL-LIP interacted with IKKα (Fig 5B) or IKKβ (Fig 5C). To investigate which receptor was required for the degradation of IKKα and IKKβ mediated by MGF300-2R, we performed knockout experiments. Hence, HEK293T cells with TOLLIP knocked out (TOLLIP −/− cells) were constructed using the CRISPR/Cas9 system (S6A Fig). We found that the knockout of p62 had a marginal effect on the degradation of IKKα ( Fig 5D) and IKKβ (Fig 5E). In contrast, the degradation of IKKα ( Fig 5F) and IKKβ (Fig 5G) by MGF300-2R was significantly restored after the knockout of TOLLIP compared with the wild-type HEK293T cells. These data confirm that TOLLIP acts as a cargo receptor involved in the MGF300-2R-mediated autophagic degradation of IKKα and IKKβ.

The ASFV MGF300-2R protein promotes the K27-linked ubiquitination of IKKα and IKKβ
The selective autophagic pathway is involved the recognition of ubiquitinated substrates by SARs for delivery to autophagosomes for degradation via the selective autophagic pathway [29]. To examine whether MGF300-2R can promote the ubiquitination of IKKα and IKKβ, we conducted a co-IP experiment. Cotransfection of pHis-Ub and pHA-MGF300-2R into HEK293T cells resulted in the induction of ubiquitination of IKKα and IKKβ by MGF300-2R as shown in Fig 6A and 6B. To determine which type of polyubiquitination was responsible for the ubiquitination of IKKα and IKKβ, we investigated the Ub chain linkage on each protein. By co-expression of IKKα or IKKβ, MGF300-2R, and each of the seven Ub mutants,

PLOS PATHOGENS
MGF300-2R is a virulence factor of ASFV including K6, K11, K27, K29, K33, K48, and K63, that only harbor one of the seven lysine residues, we found that MGF300-2R promoted the polyubiquitination of IKKα (Fig 6C) and IKKβ (Fig 6D) only in the presence of Ub and the mutant that retained K27, but not in the presence of mutants that retained K6, K11, K29, K33, K48, or K63. To confirm the role of K27 in IKKα and IKKβ polyubiquitination, we generated seven Ub mutants, each with one of the seven lysine residues (K6, K11, K27, K29, K33, K48, and K63) mutated to arginine (R). We also generated a mutant (Ub-KO) in which all seven lysine residues were mutated to arginine. Co-IP analysis of MGF300-2R expression with different ubiquitin mutants showed that for both IKKα (Fig 6E) and IKKβ (Fig 6F), K6R, K11R, K29R, K33R, K48R, and K63R lead to ubiquitination, while K27R did not show any significant difference compared with the Ub-KO. These results indicate that the K27-linked ubiquitination is responsible for the MGF300-2Rmediated autophagic degradation of IKKα and IKKβ.

MGF300-2R deletion attenuates ASFV virulence in pigs
To evaluate the viral pathogenicity of Del2R, we conducted challenge trials in pigs, a natural host of ASFV. The pigs were inoculated intramuscularly (i.m.) with Del2R at a dose of 10 2.0 TCID 50 or 10 3.0 TCID 50 , ASFV-WT at a dose of 10 3.0 TCID 50, or RPMI 1640 (Mock group)

PLOS PATHOGENS
MGF300-2R is a virulence factor of ASFV (Fig 7A). The disease progression was monitored for 21 d. Following the viral challenge, all three pigs inoculated with ASFV-WT developed fever at 3 dpi, which then gradually decreased to 41.5˚C until died at 9 dpi (Fig 7B). In contrast, all the pigs in the 10 2.0 TCID 50 Del2R-inoculated and mock groups presented no obvious ASF-specific clinical signs or symptoms and had normal body temperatures during the 21-d observation period (Fig 7B). However, two pigs in the 10 3.0 TCID 50 Del2R-inoculated group developed a fever at 6 dpi and died at 11 dpi (Fig 7B), while the other two pigs had either a normal body temperature or a transient low fever before returning to normal. The pigs inoculated i.m. with 10 3.0 TCID 50 of ASFV-WT or Del2R showed a mortality of 100% (3/3) or 50% (2/4), respectively. By contrast, all the pigs infected with 10 2.0 TCID 50 of Del2R survived (Fig 7C). The viral genomic DNA in the pigs inoculated with ASFV-WT were detected as early as 3 dpi (10 2.5 copies/mL), and gradually increased (10 5.2 copies/mL) until all pigs died at 9 dpi. In contrast, all the pigs infected with Del2R had remarkably lower viral genomic DNA in the blood compared with those infected with ASFV-WT (Fig 7D). The tissue samples from the heart, liver, spleen, lung, kidney, submaxillary lymph node, and inguinal lymph node of the pigs infected with ASFV-WT showed high viral loads (10 4.3 to 10 5.3 copies/mL) (Fig 7E). In contrast, the results showed that tissue

PLOS PATHOGENS
MGF300-2R is a virulence factor of ASFV samples from the 10 2.0 or 10 3.0 TCID 50 Del2R-inoculated pigs had lower (100 to 1000 folds) copies of viral DNA compared with the ASFV-WT-inoculated group (Fig 7E). Overall, these findings suggest that replication of Del2R in pigs is substantially attenuated compared with ASFV-WT, supporting our conclusion that MGF300-2R is a virulence factor and the virulence of Del2R is partially attenuated.

Inflammatory response is induced by the MGF300-2R-deleted ASFV mutant in pigs
Our study has shown that infection with the attenuated recombinant Del2R induced high levels of IL-1β and TNF-α production in PAMs compared with infection with ASFV-WT (Fig 2D  and 2E). Hence, we attempted to explore whether the Del2R infection enhances IL-1β and TNF-α production in vivo. We collected serum samples at 0, 1, 3, 5, and 7 dpi to evaluate the levels of IL-1β, IFN-α, and TNF-α. The pigs infected with Del2R produced significantly higher levels of IL-1β and TNF-α than did those infected with ASFV-WT at 5 and 7 dpi (Fig 8A and  8B). In contrast, the ASFV-WT inoculated pigs had significantly increased levels of IFN-α at 5 and 7 dpi (Fig 8C). These results demonstrated that the attenuated Del2R induced a higher inflammatory response in pigs compared with ASFV-WT.
Previous studies indicate that the levels of circulating antibodies, such as the antibodies against p30 and p72, are important parameters consistently associated with humoral immune responses and protection [4]. To gain an additional understanding of immune responses in the Del2R-infected pigs, we tried to detect ASFV-specific antibody response in the sera by Serum antibodies against p72 (D) or p30 (E) in the inoculated pigs were detected by ELISA. The results were shown as a blocking percentage. For the anti-p72 antibodies, the blocking rate above 50% was considered positive, while below 40% was considered negative; for the anti-p30 antibodies, the blocking rate below 40% was considered positive, while above 50% was considered negative. Samples with blocking between 40 and 50% were considered doubtful. https://doi.org/10.1371/journal.ppat.1011580.g008

PLOS PATHOGENS
MGF300-2R is a virulence factor of ASFV ELISA. The p72 (a blocking rate above 50% is considered positive) and p30 (a blocking rate below 40% is considered positive) antibodies levels of the Del2R-infected pigs appear to increase from 7 dpi and reached peak titers at 21 dpi (Fig 8D and 8E), indicating that the partially attenuated Del2R induced better antibody responses. In contrast, no p72 and p30 antibodies were detected in the sera obtained from the mock and ASFV-WT inoculated pigs (Fig  8D and 8E). Taken together, these results suggest that MGF300-2R plays an important role in the evasion of antiviral innate immune responses and contributes to the virulence of ASFV.

Discussion
Various combination of MGF360 (9L, 10L, 11L, 12L, 13L, and 14L) and MGF505 (1R, 2R, 3R, and 7R) genes deletion has been associated with reduced virulence and induction of different levels of protection in pigs against lethal homologous virulent ASFV challenge [10,39]. However, it remains unclear which critical MGFs genes have a greater effect on viral replication and pathogenicity than others. It has been shown that the cell-adapted strain ASFV-P121, with a 24.5-kb deletion in the LVR, lost the ability to replicate in PAMs [35]. Here, we further explored this deletion and found that it contains critical MGFs that affect ASFV replication in PAMs. By constructing a mutant virus (ASFV-DelLVR) with the same LVR deletion as ASFV-P121 based on the ASFV-WT, we were able to confirm this observation [35]. In this study, we identified a novel virulence factor, MGF300-2R, of ASFV and analyzed its contribution to viral pathogenesis in pigs. Significant differences were observed in the viral replication of ASFV-WT and the MGF300-2R-deleted ASFV mutant (Del2R) in pigs, indicating that attenuation of Del2R is correlated with reduced viral growth. Therefore, we conclude that MGF300-2R is a key factor for ASFV replication and virus pathogenesis.
To identify which MGFs located in the LVR have a greater effect on viral replication in PAMs, we generated six mutants with different combinations of MGFs deletions by homologous recombination. By comparing the replication kinetics of these mutant viruses with that of ASFV-WT in PAMs, we found that MGF300 family genes (1L, 2R, and 4L) significantly affect ASFV replication in PAMs (Fig 1B). This study provided the first evidence of the role of the MGF300-2R gene, for which the function was unknown. We analyzed whether there are differences in binding, internalization, virus particle release, and viral factory formation between the Del2R and ASFV-WT, but no significant differences were observed (S3 Fig). Recently, it has been found that the MGFs (MGF110-9L, MGF360-9L, and MGF505-7R) are involved in ASFV virulence and immune evasion [13][14][15][16]. We observed an increase in the expression levels of the NF-κB mediated proinflammatory cytokines IL-1β and TNF-α in the Del2R-infected PAMs, as indicated using RNA-seq, RT-qPCR, and ELISA analysis (Fig 2). Additionally, we detected a remarkable decrease in NF-κB-regulated luciferase activity in cells expressing MGF300-2R. The signal transduction pathway(s) mediated by MGF300-2R to inhibitory NF-κB was confirmed by inhibiting phosphorylation and nuclear translocation of p65. Therefore, we speculate that MGF300-2R may target the IKK complex and thereby inhibit the activation of the NF-κB signaling pathway. Furthermore, we found that MGF300-2R interacts with IKKα and IKKβ by co-IP, GST pulldown, and confocal microscopy. In addition, our findings suggest that MGF300-2R exerts its inhibitory effect on IL-1β and TNF-α production by selective autophagy to degrade IKKα and IKKβ, which primarily occurs in the cytoplasm. The virusencoded immunoregulatory proteins that are localized in the nucleus can interfere with the NF-κB-mediated immune response either by outcompeting NF-κB for KPNA binding and translocation into the nucleus or molecular mimicry [21]. Therefore, MGF300-2R may be a multifunctional protein that may employ a similar mechanism to inhibit NF-κB activity.

MGF300-2R is a virulence factor of ASFV
While further research is needed to explore this possibility, the functional data presented here provide an important clue for such investigations.
The presence of the NF-κB antagonists A238L, F317L, and H240R further supports the importance of viral inhibition of NF-κB activation in ASFV infection [24,25,27,[40][41][42][43][44][45]. The ASFV-encoded protein A238L has been described to participate in regulating the inflammatory and immune response, mainly by controlling the NF-κB signaling pathway [41]. The A238L protein is a homolog of the NF-κB inhibitor IκB, which can inhibit the production of TNF-α by inhibiting the transcriptional coactivator p300 [42][43][44]. Recent studies have shown that deletion of the A238L and CD2v genes attenuates the virulent ASFV strain Arm/07/CBM/ c2 and protects pigs against virulent challenge [45]. The F317L protein inhibits the NF-κB signaling by preventing activation of the IKK complex through interaction with IKKβ and suppression of its phosphorylation [27]. Recently, the ASFV H240R protein has been shown to interact with NEMO and NLRP3, inhibiting subsequent the activation of the IKK complex and NLRP3 inflammasome [24,25,40]. In addition, the deletion of the H240R gene reduces the viral pathogenicity in pigs compared with its parental ASFV strain [40]. Similarly, MGF300-2R inhibits the NF-κB signal transduction by preventing IKK complex activation via interaction with IKKα and IKKβ.
IFNs and pro-inflammatory cytokines depending on NF-κB regulation are essential for the antiviral immune response of the host [46]. As a counter-defense strategy, many viruses, such as VACV, ORFV, and HSV, have evolved multiple strategies to regulate the NF-κB signaling, thereby creating a favorable environment for their replication. Some viral infections activate the NF-κB pathway and promote viral replication. For example, HHV-6A-encoded U14 protein activates the NF-κB signaling by interacting with the p65 thus promoting viral gene expression [47]. However, more often, viruses inhibit the NF-κB-regulated inflammatory and immune responses at multiple levels by encoding various proteins during different stages of the viral life cycle. Large DNA viruses such as ASFV, Pseudorabies virus (PRV), VACV, ORFV, and HSV often encode multiple redundant proteins to regulate the NF-κB signaling pathway and are involved in viral replication and pathogenicity [23,41,48]. A viral protein can play a regulatory role at different levels of the NF-κB and IFN signaling pathways. For example, the virulence-associated protein MGF505-7R of ASFV can interact with IKKα and multiple adaptors of the IFN signaling pathway, including STING, IRF9, IRF3, JAK1, and JAK2 to inhibit IL-1β and IFN production, respectively [13][14][15][16]49]. Recently, numerous studies have demonstrated that the IFN inhibitors of ASFV are likely to influence viral replication and pathogenicity [50]. However, the actual role of the NF-κB inhibitors of ASFV in viral pathogenicity remains poorly defined. Consistent with the results of in vitro studies, Del2R infection induced higher levels of IL-1β and TNF-α production with significantly reduced replication in pigs compared with ASFV-WT infection. The pigs infected with ASFV-WT all died at 9 dpi, while 50% of pigs challenged with Del2R survived (Fig 7C) at a dose of 10 3.0 TCID 50 , suggesting that Del2R was dramatically attenuated. However, two pigs inoculated with 10 3.0 TCID 50 Del2R died at 11 dpi, which may have been due to excessive secretion of inflammatory cytokines. In addition, ASFV-WT inoculated pigs had significantly increased levels of IFN-α compared with the Del2R-infected pigs. This could be due to the higher level of replication of ASFV-WT than that of Del2R. Interestingly, the survival of the Del2R-infected pigs produced anti-p72 antibodies at 14 dpi (Fig 8D), indicating the induction of a robust humoral immune response. These results indicate that the MGF300-2R gene may be a potential target for developing LAV candidates. Remarkably, single-gene deletion of the ASFV NF-κB inhibitors affects viral pathogenesis and virulence variably. The diverse NF-κB inhibitors encoded by ASFV, along with the possibility of overlapping or complementary functions, may explain this observation [13][14][15][16]40].
Selective autophagy is a powerful strategy in antiviral immunity of the host [30,51,52]. Conversely, viruses have evolved various mechanisms to antagonize selective autophagy for successful infections. The L83L protein of ASFV promotes autophagic degradation of STING by recruiting TOLLIP, which inhibits type I IFN production and promotes viral replication [53]. Similarly, ASFV-encoded proteins I215L and A137R inhibit the production of type I IFN by targeting IRF9 and TBK1 for autophagic degradation, respectively [54,55]. Different from these reports, we discovered that MGF300-2R promotes the K27-linked ubiquitination of IKKα and IKKβ and is degraded via the autophagic pathway, consequently inhibiting the production of inflammatory cytokines IL-1β and TNF-α (Fig 9). TOLLIP and p62 have been reported to act as SAR and to be involved in degrading protein aggregates [30]. Here, we demonstrated that MGF300-2R protein induced the degradation of IKKα and IKKβ through TOL-LIP-mediated autophagy, which in turn negatively modulated the NF-κB signaling and promoted ASFV replication. Further, to investigate whether MGF300-2R and TOLLIP are direct interactions, we performed GST pulldown experiments. The results showed that MGF300-2R and TOLLIP were indirect interactions (S6B Fig). This suggests that MGF300-2R and IKKα, IKKβ, and TOLLIP are likely to be degraded as a complex. This study is the first to report an association between the IKK complex and TOLLIP and provides evidence that TOL-LIP plays a role in the innate immune response during ASFV infection.

PLOS PATHOGENS
MGF300-2R is a virulence factor of ASFV Overall, in this study, we found that Del2R infection induces high levels of IL-1β and TNFα in vitro and in vivo, and the MGF300-2R plays a critical role in inhibiting the production of inflammatory cytokines. Pathogenesis studies have been conducted on ASFV in pigs using a modified strain called Del2R, which lacks a single NF-κB inhibitor gene known as MGF300-2R. These studies have shown a significant reduction in pathogenicity and a marked increase in the production of inflammatory cytokines. Mechanistically, the MGF300-2R protein interacted with IKKα and IKKβ and mediated the degradation of IKKα and IKKβ through the cargo receptor TOLLIP-mediated selective autophagy pathway (Fig 9). Our findings help us understand the functions of MGF300-2R and its role in viral pathogenesis, which might help the rational design of LAVs to control ASF.

Ethics statements
All the experiments involving the ASFV HLJ/18 and Del2R in this study were performed in the enhanced biosafety level 3 facilities (BSL3) in the Harbin Veterinary Research Institute (HVRI) of the Chinese Academy of Agricultural Sciences (CAAS) approved by the Ministry of Agriculture and Rural Affairs, China. This study was conducted in compliance with the Animal Welfare Act and Guide for the Care and Use of Laboratory Animals, approved by the Laboratory Animal Welfare Committee of HVRI.

Generation of the MGFs-deleted ASFV mutant
Nine recombinant MGFs-deleted ASFV mutants (Fig 1A) were generated by homologous recombination between transfer vectors containing left and right flanking regions of the gene (s) to be deleted and EGFP reporter genes under the control of the ASFV p72 promoter and the parental virus. The transfer vectors were constructed as described previously [35]. To confirm that the MGFs had been deleted accurately, the genomes of the MGFs-deleted ASFV mutants (DelLVR, Del1L, Del2R, and Del4L) were extracted from the infected PAMs using QIAamp blood mini kit (QIAGEN, 51104). The Illumina libraries and sequencing were performed by Novogene (Beijing, China) using the NEBNext ultra DNA library prep kit (USA) following the manufacturer's recommendations. We first trimmed the adapter sequences and removed low-quality reads from the raw data using Trimmomatic v0.36 [57]. Then, the clean reads were mapped with the reference ASFV genome sequence (MK333180.1) using Bowtie2 v2.3.5.1 with default parameters [58]. Subsequently, alignment files were used for the variant calling process with Samtools 1.3.1 and Varscan 2.4.3 [59,60]. Single nucleotide polymorphisms (SNPs), insertion and deletion (INDEL) were analyzed and compared with the reference ASFV genome sequence. The functional effects of the SNPs and INDELs were identified using SnpEff v4.0 [61]. All viruses were analyzed by PCR to identify the recombinants by distinguishing wild-type and MGFs-deleted ASFVs. To verify that all the final recombinant viruses harbored the correct sequences, the PCR fragments spanning the deleted MGFs region were sequenced. The primers for the construction of the transfer vectors and PCR were listed in S1 Table.

One-step viral growth kinetics
Comparative growth curve assays between the MGFs-deleted ASFV mutants and ASFV-WT were performed in PAMs. The PAMs were cultured in 24-well cell culture plates and infected with the above ASFV mutants at a multiplicity of infection (MOI) of 5 and harvested at 2, 12,

Transmission electron microscopy (TEM), and virus binding and entry assays
PAMs were seeded in 12-well cell culture plates and allowed to grow for 12 h before being then infected with either ASFV-WT or the MGFs-deleted ASFV mutants at a MOI of 5. At 24 hpi, the cells were fixed with 2% glutaraldehyde and prepared for TEM observation following established protocols [25]. The ASFV binding and entry assays were conducted in PAMs described previously [25].

RNA-seq analysis and quantitative real-time PCR (qPCR)
For the RNA-seq, PAMs were cultured in 12-well cell culture plates, infected with the MGF300-2R-deleted ASFV mutant (Del2R) or ASFV-WT in triplicate at a MOI of 5 for 12 or 20 h, and collected in TRIzol reagent (Invitrogen, 15596026CN). RNA-seq analysis was performed as described previously [55]. Total RNAs from the PAMs infected with Del2R or ASFV-WT was isolated using a cell total RNA isolation kit (BioFlux, BSC52M1), and cDNA was synthesized using FastKing gDNA Dispelling RT SuperMix (Tiangen, KR118-02). The resulting cDNAs were used as templates for RT-qPCR using SYBR Premix Ex Taq (TaKaRa, RR390B) to assess the expression levels of IL-1β, IFN-β, TNF-α, IKKα, IKKβ, and NEMO genes. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a reference gene for normalization. The qPCR primers used in this study are listed in S1 Table. The relative abundance of each gene was calculated using the comparative cycle threshold (2 -ΔΔCT ) method.

Detection of inflammatory cytokines
PAMs were seeded in 24-well cell culture plates for 12 h, then infected with ASFV-WT or Del2R for 1 h and rinsed twice with PBS. The infected cells were cultured in the fresh RPMI 1640 medium for 12 and 20 h. The concentrations of IL-1β (Ray Biotech, ELP-IL1b-1), TNF-α (Ray Biotech, ELP-TNFa-1), and IFN-α (Ray Biotech, ELP-IFNa-1) in the cell supernatants were measured by enzyme-linked immunosorbent assay (ELISA) kits according to the manufacturer's instructions. The cells were collected for RT-qPCR analysis as described above, and the mock-infected cells were included as a negative control.

Transfection
HEK293T and HeLa cells were transfected using linear 25-kDa polyethyleneimine (PEI) (Polysciences, 23966-2). Briefly, a transfection mixture consisting of 3 μg of the indicated plasmid and 9 μL of PEI (stock solution at 1 mg/mL) in 300 μL of Opti-MEM (Invitrogen, 31985070) was used to transfect cells of 80% confluence in 6-well cell culture plates. After a 20-min incubation at room temperature, the transfection mixture was added to the cells and incubated for 24-36 h. PAMs were transfected using X-tremeGENE HP (Roche, 6366546001) according to the manufacturer's instructions.

Laser confocal microscopy
HEK293T cells were seeded on a poly-L-lysine-coated glass-bottom cell culture dish (NEST, 801001) and transfected with the indicated expressing plasmids using PEI. At 24 hours posttransfection (hpt), the cells were fixed with 4% paraformaldehyde for 20 min at room temperature, followed by permeabilized with 0.1% triton X-100 (Sigma-Aldrich, V900502) in PBS for 20 min, and then blocked with 5% bovine serum albumin (BSA). Specific primary antibodies were added and incubated for 2 h. After washing with PBS for five times, the cells were stained with the Alexa Fluor-conjugated goat anti-mouse IgG or goat anti-rabbit IgG secondary antibodies for 1 h at room temperature. The cells were washed with PBS for three times and then incubated with DAPI for 5 min for nuclear staining. After three washing steps with PBS, the cells were observed under a confocal laser scanning microscopy with Airyscan (Zeiss, LSM800), and the most representative images were selected for presentation. Colocalization analysis between MGF300-2R and IKKα or IKKβ was performed by calculating the Pearson's correlation coefficients using the coloc2 tool of ImageJ/FIJI.

Western blotting, co-immunoprecipitation (co-IP) and GST pull-down assays
HEK293T cells were harvested and lysed in ice-cold RIPA lysis buffer (Sigma-Aldrich, R0278) supplemented with a protease inhibitor cocktail (Roche, 4693116001). The cell lysates were boiled in SDS-polyacrylamide gel electrophoresis (SDS-PAGE) loading buffer (Solarbio, P1040) and resolved by SDS-PAGE. Separated proteins were transferred onto polyvinylidene fluoride membranes, followed by blocking with 5% skim milk in Tris-buffered saline with 0.05% tween 20 (TBST) for 1 h at room temperature. The membranes were then probed by appropriate primary and secondary antibodies. The signals were detected using an Odyssey imaging system.
The transfected HEK293T cells were lysed in RIPA lysis buffer for 30 min on ice. After removal of nuclei through low-speed centrifugation and collecting 100 μL of the sample as input, the remaining 800-μL lysates were incubated with the appropriate antibody, followed by the addition of anti-Flag M2 magnetic beads or anti-HA agarose and rotated at 4˚C overnight. After incubation, the beads were collected using a magnetic separator and washed three times with PBS. The bound proteins were subjected to western blotting using appropriate antibodies. Densitometric analysis of the blots was performed with ImageJ/FIJI as previously described [32]. Levels of target proteins such as MGF300-2R were normalized to the loading control βtubulin.

Dual-luciferase reporter assay
HEK293T cells were cotransfected with the reporter plasmids (pIFN-β-Fluc, pISRE-Fluc, pIRF3, and pNF-κB-Fluc) and an internal control vector pTK-Rluc along with or without indicated expressing plasmids using X-tremeGENE HP transfection reagent. At 24 hpt, the cells were washed with PBS for three times. The passive lysis buffer was added for 20 min at 4˚C with gentle shaking. The luciferase activities were then determined by a dual-luciferase reporter assay system (DLR) (Promega, E1910), according to the manufacturer's instructions. To analyze the effects of the inhibitors, the transfected cells were left untreated or treated with 10 ng of TNF-α as an inducer of NF-κB. The luciferase activity (Fluc/Rluc) was calculated by dividing the firefly luciferase activity by the Renilla luciferase activity.

NF-κB-p65 nuclear translocation assay
HEK293T cells were transfected with expressing plasmid pFlag-MGF300-2R for 24 h, followed by being washed twice with PBS and cultured in indicated medium. The cells were stimulated PLOS PATHOGENS MGF300-2R is a virulence factor of ASFV with TNF-α (20 ng/mL) for 1 h. The separation of cellular fractions was performed by a Minute plasma membrane protein isolation and cell fractionation kit (Invent Biotechnologies, SM-005). The resulting supernatants were collected as a nuclear extract. Tubulin and lamin B1 were used as cytoplasm and nucleus markers, respectively. The nuclear and cytoplasmic compartments were examined by western blotting. To investigate the effect of MGF300-2R on nuclear translocation of the NF-κB-p65 following ASFV infection. PAMs were mock-infected or infected with ASFV-WT or Del2R (MOI = 5). At 24 hpi, the separation of cellular fractions of ASFV-infected PAMs was performed as described above.

Pig inoculation experiments
Fourteen healthy specific-pathogen-free piglets, aged 4 weeks and weighing between 11.5 and 13.5 kg, were obtained from the Laboratory Animal Center of the HVRI. Back titrations were performed to confirm the titers of the Del2R or ASFV-WT inoculum. The piglets were randomly divided into four groups and inoculated i.m. with ASFV to investigate the pathogenesis of the Del2R, with four piglets inoculated with Del2R (10 2.0 TCID 50 /piglet); 4 piglets inoculated with Del2R (10 3.0 TCID 50 /piglet); three piglets inoculated with ASFV-WT (10 3.0 TCID 50 / piglet); three piglets inoculated with RPMI 1640 (1 mL/piglet). Daily monitoring was conducted for each pig, recording rectal temperatures and the clinical signs for each pig, including lethargy, anorexia, depression, vomiting, fever, skin hemorrhages, bloody diarrhea, and joint swelling. EDTA-anticoagulated blood and sera were collected from all pigs at 0 1, 3, 5, 7, 10, 14, and 21 days post-infection (dpi) for inflammatory cytokines detection by ELISA kits as described above and virus load detection by qPCR, respectively. Anti-p72 (INGENASA, 11. PPA.K.3/5) and -p30 (ID.VET, ASFC-5P) antibodies were measured using commercial ELISA kits following the manufacturer's instructions. At 21 dpi, all surviving piglets were euthanized and seven different tissues and organs, including the heart, liver, spleen, lung, kidney, submandibular lymph nodes, and mesenteric lymph nodes were obtained from each necropsied pig, for viral load detection by qPCR as described previously [34].

Statistical analysis
An unpaired two-tailed t test was used to determine statistical significance and performed in GraphPad Prism version 8.0.0 (GraphPad Software, San Diego, California, USA). Data are presented as the mean ± SD of three independent experiments. P � 0.05 was considered statistically non-significant (ns); P < 0.05 was considered statistically significant (* P < 0.05; ** P < 0.01; *** P < 0.001).