African Swine Fever Virus L83L Negatively Regulates the cGAS-STING-Mediated IFN-I Pathway by Recruiting Tollip To Promote STING Autophagic Degradation

ABSTRACT

African swine fever (ASF) is a devastating infectious disease of pigs caused by the African swine fever virus (ASFV), which poses a great danger to the global pig industry. Many viral proteins can suppress with interferon signaling to evade the host's innate immune responses. Therefore, the development of an effective vaccine against ASFV has been dampened. Recent studies have suggested that the L83L gene may be integrated into the host genome, weakening the host immune system, but the underlying mechanism is unknown. Our study found that L83L negatively regulates the cGAS-STING-mediated type I interferon (IFN-I) signaling pathway. Overexpression of L83L inhibited IFN-β promoter and ISRE activity, and knockdown of L83L induced higher transcriptional levels of interferon-stimulated genes (ISGs) and phosphorylation levels of IRF3 in primary porcine alveolar macrophages. Mechanistically, L83L interacted with cGAS and STING to promote autophagy-lysosomal degradation of STING by recruiting Tollip, thereby blocking the phosphorylation of the downstream signaling molecules TBK1, IRF3, and IκBα and reducing IFN-I production. Altogether, our study reveals a negative regulatory mechanism involving the L83L-cGAS-STING-IFN-I axis and provides insights into an evasion strategy involving autophagy and innate signaling pathways employed by ASFV.

IMPORTANCE African swine fever virus (ASFV) is a large double-stranded DNA virus that primarily infects porcine macrophages. The ASFV genome encodes a large number of immunosuppressive proteins. Current options for the prevention and control of this pathogen remain pretty limited. Our study showed that overexpression of L83L inhibited the cGAS-STING-mediated type I interferon (IFN-I) signaling pathway. In contrast, the knockdown of L83L during ASFV infection enhanced IFN-I production in porcine alveolar macrophages. Additional analysis revealed that L83L protein downregulated IFN-I signaling by recruiting Tollip to promote STING autophagic degradation. Although L83L deletion has been reported to have little effect on viral replication, its immune evade mechanism has not been elucidated. The present study extends our understanding of the functions of ASFV-encoded pL83L and its immune evasion strategy, which may provide a new basis for developing a live attenuated vaccine for ASF.

INTRODUCTION

African swine fever (ASF) is a highly acute, contact infectious disease caused by the African swine fever virus (ASFV), and the mortality rate of infection by virulent strains can reach 100% (1). ASFV was first discovered in Kenya and Africa in the 20th century, spread rapidly in several European countries, and gradually entered the border between China and Russia (2). In 2018, the ASF epidemic broke out in Shenyang, China, spreading widely throughout the country. It has caused great harm to the pig industry and veterinary public safety in China (3). Due to the knowledge gap regarding ASFV, no effective vaccine or treatment has been developed thus far. ASFV is the only large double-stranded DNA arbovirus known to exist and is the only member of the Asfarviridae family, the genus Asfivirus (4). ASFV is composed of four concentric layers with an icosahedral nucleocapsid. Its genome is genetically large and complex, with a length of 170 to 193 kb, containing 150 to 167 open reading frames (ORFs) and encoding approximately 200 proteins (5). Due to the complex role played by these proteins in evasive defense, the immune evasion mechanisms by ASFV have yet to be defined (6). Therefore, it is urgent to define the functions of ASFV-encoded proteins to evade innate immunity.

Innate immunity is the first line of defense that has evolved gradually due to the evolutionary development of the species/lineage. Because most pathogens release nucleic acids during the replication cycle, sensory recognition has evolved as the primary strategy for host innate immunity (7). Pathogen-associated molecular patterns (PAMPs) are specifically recognized by pattern recognition receptors (PRRs), which induce the host to depart a signaling cascade that produces proinflammatory factors and interferons (IFNs), thereby clearing pathogenic infections (8). PRRs include TLRs, RLRs, and the DNA sensor cyclic GMP-AMP synthase (cGAS) (9). The cGAS-STING signaling pathway is essential in the first phase to control viral infection in the cytoplasm, including receptors such as cGAS capable of recognizing pathogenic DNA and initiating the innate immune program, and most of these highly conserved sequences are highly expressed in immune cells (10). cGAS recognition of viral DNA is catalyzed by ATP-GTP to form the second messenger cGAMP, which recruits a portion of the junction protein stimulator of interferon genes (STING) on the endoplasmic reticulum (ER), triggering oligomerization (11). Upon passage through the endoplasmic reticulum-Golgi intermediate compartment (ERGIC) and Golgi apparatus, recruitment of TANK binding kinase 1 (TBK1) self-activation and interferon regulatory factor 3 (IRF3) phosphorylation and promotion of IRF3 translocation into the nucleus triggers the IFN-I response (12).

In addition to interferon production, the cGAS-STING signaling axis can induce autophagic events as an additional immune strategy against cellular damage caused by viral infection. Pathogenic DNA is bound by cGAS to form cGAMP, which triggers STING to interact with SEC24C. In particular, ERGIC acts as a membrane source for WIPI2 recruitment and LC3 lipidation to promote autophagic vesicle formation and drag target DNA into lysosomes for degradation (13).

Numerous studies have demonstrated that some of the encoded proteins of ASFV are required for the negative regulation of the IFN-I immune response and inflammatory response (14). For instance, ASFV DP96R inhibited IFN-β production by inhibiting TBK1 phosphorylation (15), pS273R antagonized the cGAS-STING-mediated IFN-I signaling pathway by targeting IKKε, and F317L blocked NF-κB activation by decreasing IκBα phosphorylation and ubiquitination (16). In addition, viral proteins have been reported to manipulate the cGAS-STING-IFN-I signaling axis by mediating the autolysosome pathway. ASFV MGF505-7R promoted the autophagy-associated protein ULK1 to degrade STING and inhibit IFN-I production (17), and pA137R promoted autophagy-mediated lysosomal degradation of TBK1, blocking the nuclear translocation of IRF3 (18).

The L83L protein is a nonessential protein encoded by ASFV with a protein length of approximately 81 to 83 amino acid residues and is involved in early viral replication. Previous studies have shown that ASFV Georgia L83L deletion strains do not alter their pathogenicity in vitro or in vivo (19). Recent studies have found that L83L resides in the tissues of asymptomatic pigs and likely integrates into the host genome (20). Although L83L may associate with IL-1β to control inflammatory responses, how L83L regulates host innate immunity remains unclear.

We determined that L83L inhibited type I IFN production through the cGAS-STING signaling pathway in the present study. Knockdown of L83L promoted interferon-stimulated genes transcription and IRF3 phosphorylation levels. Further studies indicated that L83L interacted with STING and delivered it to autophagosomes for degradation by inducing the selective autophagy receptor Tollip. Our findings reveal a new strategy for ASFV to evade innate immunity and provide a new theoretical basis for developing recombinant vaccines.

RESULTS

L83L inhibits cGAS-STING-mediated IFN-β and ISRE reporter gene activity.

To elucidate how ASFV-encoded proteins evade the control of innate immunity at the molecular level, we screened 12 proteins with immunosuppressive potential (Fig. S1). We found that the L83L protein downregulated cGAS-STING-mediated IFN-β and interferon-stimulated response element (ISRE) reporter gene activity. L83L also strongly inhibited cGAS and STING activity at the protein level (Fig. 1A and F). To further explain which site of L83L might achieve immunosuppression, we continued to assay separately and gradually increased the dose of L83L. The results showed that L83L was able to maintain dose-dependent inhibition of IFN-β and ISRE activity (Fig. 1B and C) (Fig. 1G and H), while interference with protein expression was consistent with our expected results, especially for the inhibition of cGAS and STING (Fig. 1A and F). In contrast, no interference with the most downstream signaling molecule (TBK1 and IRF3-5D) could be detected by either assay (Fig. 1D and E) (Fig. 1I and J). Moreover, L83L manipulated against host defense by inhibiting cGAS-STING-mediated IFN-I activation but not MyD88-mediated NF-κB activation (Fig. S2A), indicating that L83L is preferentially required for controlling cGAS-STING signaling responses. These results suggest that L83L negatively regulates host innate immune responses, possibly by targeting cGAS or STING activity, inhibiting the IFN-I pathway.

FIG 1.

L83L inhibited the cGAS-STING-mediated IFN-β promoter and ISRE activity. (A-J) HEK293T cells were cultured overnight in 24-well plates. IFN-β promoter, ISRE reporter gene plasmids (100 ng), pRL-TK internal reference plasmid (50 ng), and the indicated Flag-tagged signaling molecule plasmids (cGAS, STING, TBK1, and IRF3-5D) (200 ng) were cotransfected into cells with the indicated doses of HA-tagged L83L expression plasmid (50, 100, 200 ng) for 24 h. Cells were collected for dual-luciferase reporter gene and Western blotting assays. + represents transfected, and – represents untransfected. Data are presented as the mean ± SD of 3 experiments. P-values smaller than 0.05 were considered statistically significant (*, P < 0.05; **, P < 0.01; and ***, P < 0.001).

L83L restricts cGAS-STING-mediated transcription of ISGs.

Interferon-stimulated genes (ISGs) are a class of antiviral factors induced by IFNs that play a crucial role in host resistance to viral infection. To investigate the effect of L83L on cGAS-STING-mediated transcription levels of ISGs downstream of IFN-β, cGAS and STING expression plasmids were cotransfected with the L83L expression plasmid in HEK293T cells. Quantitative PCR (qPCR) analysis indicated that L83L downregulated the transcription levels of IFNB1, IFIT2, ISG56 (Fig. 2A), and RANTES (Fig. S2B). HSV-60 and VACV-70 are double-stranded 60 or 70 bp oligonucleotide containing viral DNA motifs derived from the herpes simplex virus 1 (HSV-1) and vaccinia virus (VACV). It has been shown in multiple studies to be agonists of multiple cytosolic DNA sensors (CDSs) that trigger IFN-I production and the induction of ISGs (21). We found that L83L inhibited the gene transcription levels of IFNB1, ISG15, and ISG56 activated by mimic DNA virus infection in PAMs (Fig. 2B to D). This finding demonstrates that L83L is required for inhibiting ISGs expression.

FIG 2.

L83L suppresses the mRNA levels of ISGs. (A) HEK293T cells were cultured overnight in 6-well plates, and the indicated Flag-tagged signaling molecule plasmids (cGAS and STING) (500 ng) were cotransfected into cells with the HA-tagged L83L expression plasmid (1 μg) for 24 h. The mRNA expression levels of IFNB1, IFIT2, ISG56, and RANTES were determined by qPCR assay. (B-D) PAMs were cultured overnight in 6-well plates, and HA-tagged L83L (1 μg) was transfected into cells for 24 h and then mock left untreated or treated with HSV-60 (B), VACV-70 (C), and 2'3'-cGAMP (D) (5 μg/mL) for 16 h. The mRNA expression levels of IFNB1, ISG15, and ISG56 were determined by qPCR assay. Data are presented as the mean ± SD of 3 experiments. *, P < 0.05; **, P < 0.01; ***, P < 0.001 (unpaired t test).

L83L downregulates the phosphorylation of the downstream signaling molecules TBK1 and IRF3.

To further validate that L83L manipulates antiviral innate immunity, we examined the effect of L83L on downstream signaling molecule activity. Immunoblot analysis revealed that L83L downregulated the phosphorylation levels of TBK1 and IRF3 activated by overexpression of cGAS and STING in HEK293T cells (Fig. 3A). PAMs were considered more tempting targets for ASFV infection. Previous studies showed that HSV60, VACV-70, and 2'3'-cGAMP are essential for the activating the DNA-mediated phosphorylation of signaling molecules (Fig. 2). We then stimulated those PAMs for 16h with VACV-70, HSV-60, and 2'3'-cGAMP (a ligand of STING). The results showed that L83L reduced HSV60-, VACV-70-, and 2'3'-cGAMP-induced TBK1 and IRF3 phosphorylation levels (Fig. 3B to D). Altogether, we determined that L83L likely contributes to ASFV invasion by inhibiting TBK1 and IRF3 phosphorylation, thereby limiting the IFN-I pathway.

FIG 3.

L83L inhibits the phosphorylation of TBK1 and IRF3. (A) HEK293T cells were cultured overnight in 6-well plates, and the indicated Flag-tagged signaling molecule plasmids (cGAS and STING) (500 ng) were cotransfected into cells with the HA-tagged L83L expression plasmid (1 μg) for 24 h. The expression of cGAS, STING, pTBK1, TBK1, pIRF3, IRF3, and L83L protein was detected by Western blotting assay. (B-D) PAMs were cultured overnight in 6-well plates, and HA-tagged L83L (1 μg) was transfected into cells for 24 h, and then mock left untreated or treated with HSV-60 (B), VACV-70 (C), and 2'3'-cGAMP (D) (5 μg/mL) for 16 h. The expression of pTBK1, TBK1, pIRF3, IRF3, and L83L protein was detected by Western blotting assay. All experiments were repeated at least three times independently.

L83L degrades cGAS and STING.

We noted that overexpression of L83L downregulated the expression of cGAS and STING (Fig. 1A and F). We hypothesized that L83L might regulate the stability of cGAS and STING. To test this assumption, we cotransfected HA-tagged L83L with all Flag-tagged molecules involved in the cGAS-STING signaling pathway and performed immunoblot analysis. cGAS and STING degradation were promoted by L83L in a dose-dependent manner (Fig. 4A and B), which is consistent with our initial determination (Fig. 1B and C) (Fig. 1G and H). However, L83L did not interfere with the expression of TBK1, IKKε, and IRF3 (Fig. 4C to E). These data suggest that L83L specifically destabilizes the cGAS and STING proteins.

FIG 4.

L83L degrades cGAS and STING. (A-E) HEK293T cells were cultured overnight in 24-well plates, and the indicated Flag-tagged signaling molecule plasmids (cGAS, STING, TBK1, IKKε, and IRF3) (200 ng) were cotransfected into cells with the indicated doses of HA-tagged L83L expression plasmids (50, 100, 200 ng) for 24 h. The expression of cGAS, STING, TBK1, IKKε, IRF3, and L83L protein was detected by Western blotting assay. Immunoblots were quantified using ImageJ software and normalized to the GAPDH loading control. All experiments were repeated at least three times independently. *, P < 0.05; **, P < 0.01; ***, P < 0.001 (unpaired t test).

L83L interacts with cGAS and STING.

Next, we further screened the signaling molecule proteins that interacted with L83L by coimmunoprecipitation (Co-IP) to determine whether the degradation events derived from the molecules' interactions. It was found that Flag-tagged cGAS and STING, but not other tested molecules (TBK1, IKKε, and IRF3), specifically interacted with HA-tagged L83L (Fig. 5A). STING is an essential adaptor molecule with antiviral activity in the DNA-sensing pathway. To investigate this in more detail, STING was selected for further analysis. Confocal microscopy confirmed the colocalization of L83L and exogenous or endogenous STING in the cytoplasm (Fig. 5B and D). A semi-endogenous IP analysis in PK-15 cells confirmed the interactions between L83L and STING (Fig. 5C). We are convinced that L83L targets cGAS and STING by interacting to degrade their activities, thus limiting the production of IFN-I.

FIG 5.

L83L interacts with cGAS and STING. (A) HEK293T cells were cultured overnight in 6-well plates and cotransfected with HA-tagged L83L expression plasmid and Flag-tagged molecule expression plasmids (cGAS, STING, TBK1, and IRF3) (1 μg) for 24 h. (B) HEK293T cells were cotransfected with HA-L83L and Flag-STING plasmids. Confocal microscopy of HA (green), Flag (red), and DAPI (blue) in treated cells. Scale bars, 5 μm. (C) PK-15 cells were cultured overnight and transfected with HA-tagged L83L expression plasmid (1 μg) for 24 h. For Co-IP experiments, immunoblot analysis was performed using the antibodies shown to assess whole-cell lysates and immunoprecipitated complexes. (D) PK-15 cells were transfected with HA-L83L expression plasmid. Confocal microscopy of HA (green), STING (red), and DAPI (blue) in treated cells. Scale bars, 5 μm.

L83L promotes Tollip-mediated selective autophagic degradation of STING.

Since we repeatedly observed a strong inhibition of STING by L83L, we were interested in how L83L degrades STING through interaction with STING. We found that L83L-mediated STING degradation could be mostly restored by treatment with the autophagy inhibitor 3-MA or proteasome inhibitor MG132 but not the caspase inhibitor ZVAD (Fig. 6A). Selective autophagy requires specific autophagy receptors to link invasion proteins and autophagosomes (22), and we hypothesized that specific cargo receptors might deliver STING to autophagosomes for degradation. Co-IP experiments indicated that L83L interacted with Tollip but not with NDP52, p62, OPTN, or NBR1 (Fig. 6B). Blockade of autophagy by 3-MA also restored STING degradation (Fig. 6C). Interestingly, overexpression of L83L potentiated STING degradation mediated by Tollip (Fig. 6D).

FIG 6.

L83L recruits Tollip to promote STING autophagic degradation. (A) HEK293T cells were cultured overnight in 12-well plates, and Flag-tagged cGAS or STING expression plasmids were cotransfected with HA-tagged L83L expression plasmids for 24 h. Cells were either left untreated or treated with PBS, 3-MA (10 mM), DSMO, MG-132 (10 μM), and ZVAD (20 μM) for 6 h. (B) HEK293T cells were cotransfected with Flag-tagged autophagy receptor expression plasmids (NBR1, OPTN, p62, NDP52, and Tollip) and HA-tagged L83L expression plasmids for 24 h. For Co-IP experiments, immunoblot analysis was performed using the antibodies indicated to assess whole-cell lysates and immunoprecipitation complexes. (C) HEK293T cells were cotransfected with Flag-tagged Tollip expression plasmid and Myc-tagged STING expression plasmid for 24 h. Cells were either left untreated or treated with PBS and 3-MA (10 mM) for 6 h. (D) HEK293T cells were cotransfected with Flag-tagged Tollip expression plasmid, Myc-tagged STING expression plasmid, and HA-tagged L83L expression plasmid for 24 h. Cells were collected at the indicated times and detected using a Western blotting assay. (E) PAMs were transfected with HA-tagged L83L (1 μg) for 24 h, while mock were left untreated or treated with VACV-70 (5 μg/mL) for 16 h. The expression of pTBK1, TBK1, pIRF3, IRF3, pIκBα, IκBα, endogenous STING, LC3B, p62, and L83L protein was detected by Western blotting assay. (F) PAMs were transfected with HA-tagged L83L (1 μg) for 24 h. Then, cells were infected with HSV-1 (multiplicity of infection = 0.1) and treated with PBS, 3-MA (10 mM), CQ (50 μM), DSMO, MG-132 (10 μM), and ZVAD (20 μM) for 6 h. (G) PAMs were transfected with siNC, siTollip, or HA-tagged L83L and treated with VACV-70 (5 μg/mL) as indicated for 16 h. STING, Tollip, HA, and β-actin were assessed by immunoblot analysis. All experiments were repeated at least three times independently.

To further reveal the L83L-mediated autophagic program, we infected PAMs with HSV-1 or added VACV-70 to activate the cGAS-STING signaling pathway. Similar to previous findings, an increased dose of L83L restricted the phosphorylation of TBK1, IRF3, and IκBα induced by VACV-70 (Fig. 3C). As expected, the expression of endogenous STING was restricted. Specifically, an additional decrease in LC3-II and p62 levels in L83L-overexpressing VACV-70-treated PAMs compared to nontreated VACV-70-treated cells was observed that L83L promotes autophagic stream activation (Fig. 6E). Furthermore, the L83L-induced destabilization of STING was reversed by the autophagy inhibitor 3-MA, CQ and proteasome inhibitor MG132 (Fig. 6F), suggesting an important role of the autophagy-lysosome in the depletion of STING protein by L83L. We also found that the knockdown of Tollip slowed the degradation of endogenous STING (Fig. 6G). These results suggest that L83L promotes Tollip-mediated selective autophagic degradation of STING.

Knockdown of L83L upregulates IFN-I activation in PAMs.

ASFV infection has a very complex molecular mechanism. However, we cautiously postulate that the increased IRF3 phosphorylation and ISGs transcription resulting from L83L deficiency may reflect a loss of gene repression function. We used small interfering RNA (siRNA) to knockdown the expression of ASFV L83L protein (Fig. S2C) and found that L83L knockdown significantly increased the expression of ISGs in PAMs both at 12 and 24 h postinfection, such as IFNB1, ISG15, and ISG56. Corresponding to a gradual decrease in the levels of viral structural proteins P72 and P30 mRNA levels gradually decreased (Fig. 7A). An ELISA analysis indicated that knockdown of L83L potentiated ASFV-triggered production of IFN-β in PAMs (Fig. 7B). Meanwhile, supernatants were monitored for GFP signal, and the results showed that the supernatant could effectively inhibit the replication of GFP-expressing vesicular stomatitis virus (VSV-GFP) in HEK293T cells or PAMs, which suggested that L83L dampened the secretion of antiviral factors induced by ASFV (Fig. 7D) (Fig. S3). Additionally, the increase in cellular IRF3 and IκBα phosphorylation levels and the viral structural proteins p72 and p54 involved in early replication were repressed (Fig. 7C).

FIG 7.

Knockdown of L83L promotes IFN-I pathway activation. (A) PAMs were cultured overnight in 6-well plates, and si-L83L-86 and si-NC (nontargeting control) were transfected for 24 h. Subsequently, PAMs were infected without or with ASFV SY18 (multiplicity of infection = 1) for 12 h or 24 h. Cells were collected at the indicated times and examined for host ISGs, including IFNB1, ISG15, and ISG56 genes transcript levels and viral P72 and P30 genes transcript levels, using qPCR. (B) Cytokine (IFN-β) levels in the supernatant were determined by ELISA. (C) IRF3 and IκBα phosphorylation levels and viral protein p72 and p54 protein expression levels were examined using a Western blotting assay. (D) Inactivated cell supernatants were collected to treat fresh HEK293T cells for another 24 h. The cells were then infected with VSV-GFP (multiplicity of infection = 0.01) for 12 h. The cells were observed microscopically and then assessed by flow cytometry. Data are presented as the mean ± SD of 3 experiments. *, P < 0.05; **, P < 0.01; ***, P < 0.001 (unpaired t test).

DISCUSSION

ASFV is a virulent infectious disease with a high mortality rate that has caused great harm to the global pig industry. ASFV has evolved various strategies to counteract the cGAS-STING-IFN-I signaling axis using different viral proteins (23). Several studies have now demonstrated that the induction of higher levels of IFN-I in piglets by ASFV-Δ7R compared to its parental counterpart leads to reduced virulence (24). In addition, I226R may be involved in regulating NEMO ubiquitination, thereby inhibiting the cGAS-STING-dependent NF-κB activation response (25). In a recent study, Zheng et al. systematically elucidated the mechanisms linking key virulence genes of ASFV infection with host genes through single-cell transcriptome landscapes (26). In recent years, there have been numerous studies on ASFV. However, strategies for antagonizing the innate immune mechanism are still very limited, and the functions of some proteins have not been clearly described. This study found that L83L inhibited cGAS-STING-mediated IFN-β promoter and ISRE activity through a dual-luciferase reporter gene system screen.

Furthermore, L83L suppressed the transcription of IFN-β and ISGs. Further studies revealed that L83L interacted with cGAS and STING and inhibited the phosphorylation of TBK1, IRF3, and IκBα. Mechanistically, L83L may be involved in regulating the autophagic degradation of STING to inhibit the IFN-I pathway (Fig. 8). In conclusion, these data reveal the function of L83L and its mechanism of inhibiting the innate antiviral response, which may provide a new theoretical basis for developing new vaccines.

FIG 8.

Schematic representation of ASFV L83L targeting STING to inhibit IFN-I activity. L83L interacts with and degrades STING via the autophagy pathway by inducing the selective autophagy receptor Tollip, followed by the inhibition of TBK1 and IRF3 phosphorylation to suppress the IFN-I response.

The cGAS-STING pathway is a fundamental strategy of host innate immunity with multiple roles in physiological and pathological processes, including antiviral immune responses and inflammation (27). cGAS is a receptor that receives DNA viral signals, and its enzymatic function is achieved mainly through its C-terminal structural domain and DNA-binding sites. However, in some phagocytes, cGAS is the second DNA-stimulating signal rather than the first initiating signal of the anti-DNA viral immune response. V-ATPase-SYK-mediated tyrosine phosphorylation of cGAS was first activated by viral endocytosis, and then the innate immune response was initiated (28). At the same time, cGAS is also subject to extensive host regulation (29). For example, β-inhibin 2 positively promoted IFN-β production by targeting cGAS (30). The cGAS-DNA phase separation can constrain nuclease TREX1 activity for efficient DNA sensing (31).

In contrast, viruses have evolved multiple strategies to evade immunity. A recent study showed that herpesviruses alter the innate immune response and facilitate viral replication by disrupting cGAS-DNA phase separation (32). In the present study, we found that L83L likely promoted cGAS destabilization via autophagy or the proteasome pathway (Fig. S2D and E), which has not been reported in detail for ASFV studies thus far. Whether the L83L protein can inhibit endocytosis-mediated activation of cGAS in porcine macrophages or compete with cGAS for DNA binding and thereby disrupt cGAS-DNA phase separation remains to be further investigated.

STING is a cGAS downstream junction protein essential for antiviral immunity (27). STING acts as a platform for a signaling cascade dependent on TBK1 downward transmission and promotes IRF3 phosphorylation and activation, thereby initiating an antiviral immune response through the upregulation of IFN-I transcription (33). In addition, IKK kinase can also be activated to induce inflammatory factor production. Li et al. previously reported that ASFV GN/GS/2018 MGF505-7R protein could target STING for autophagic degradation, leading to the suppression of IFN-I production (17). Subsequently, other ASFV protein-mediated immunomodulation has also been extensively reported. Amino acids 1 to 191 and 182 to 360 of MGF505-11R and 167 to 353 of MGF360-11L have been reported to inhibit cGAS-STING-mediated activation of the IFN-β promoter (34). Moreover, MGF360-9L degraded the IFN-β downstream transcriptional activators STAT1 and STAT2 via the apoptosis and ubiquitin-proteasome pathways (35). In this study, we revealed inhibition of the host antiviral immune response by the ASFV nonessential protein L83L by targeting STING for degradation.

Posttranslational modification of proteins is a dynamic process that affects the intracellular localization of proteins, and the formation of polyubiquitin chains depends on the continued ubiquitination of ubiquitin lysine residues of the substrate protein (36). STING receives modifications from different types of polyubiquitin chains to maintain host immune homeostasis after DNA virus infection. For instance, CYLD stabilized STING activity by removing K48-linked ubiquitination and positively regulated the innate immune response against herpes simplex virus 1 (HSV-1) infection (37). TBK1, a key pivotal molecule in the immune signaling cascade (38), is similarly regulated by multiple posttranslational modifications, including NEDD4 promoting TBK1 K344 site K27-linked ubiquitination to advance TBK1 autophagic degradation (39). Conversely, the arginine methyltransferase PRMT1 enhanced TBK1 asymmetric dimethylation to activate downstream signaling against the virus (40). Viruses have evolved immune evasion strategies in response to host complexed and conserved innate immunity (41). Meq, the major oncoprotein of Marek's disease virus (MDV), blocked the recruitment of TBK1 and IRF7 to the STING complex (42). ASFV pI215L promoted RNF138 degradation of RNF128, leading to reduced TBK1 K63-linked ubiquitination and inhibition of IFN-I production (43). MGF505-7R degraded JAK1 by upregulating RNF125 expression and thus restricted IFN-II signaling (44), and MGF360-14L mediated IRF3 K63-linked ubiquitination by promoting TRIM21, leading to IRF3 instability (45).

Cellular autophagy is a crucial process for maintaining cellular homeostasis. Autophagy can be classified into nonselective and selective autophagy depending on its selectivity in degrading substrates (46). From an evolutionary point of view, STING-induced autophagy is a primitive function preceding the induction of interferon function and is independent of TBK1-IRF3 and classical autophagic signaling (13). Another study also demonstrated that STING undergoes oligomerization rather than phosphorylation, which is required for FMDV-induced autophagy (47). Resting-state STING is subject to competitive regulation by Tollip and IRE1a-lysosomes to maintain tissue immune homeostasis. It is bound by the endoplasmic reticulum-resident protein UNC93B1 and facilitates its degradation via the autophagy-lysosome pathway (48). Our data suggested that L83L mediates STING autophagic degradation through the induction of the selective autophagy receptor Tollip to evade host defense via a novel mechanism. However, further studies are needed to determine whether the balance between IRE1a-lysosomes and Tollip is disrupted.

In the present study, we identified a role for L83L in antagonizing innate immunity by controlling the cGAS-STING-mediated IFN-I signaling pathway. Indeed, L83L mediates the autophagic degradation of STING via Tollip and is essential to help ASFV evade immunity. This result provides a novel mechanism explaining how ASFV L83L manipulates innate immunity to promote the molecular immune mechanism of infection.

MATERIALS AND METHODS

Cells and viruses.

Human embryonic kidney cells (HEK293T), porcine kidney epithelial cells (PK-15), and Cercopithecus aethiops epithelial kidney cells (Vero) were purchased from ATCC and maintained in our laboratory. HEK293T cells, PK-15 cells, and Vero cells were cultured in a complete medium (DMEM, 10% FBS, 1% penicillin) (Gibco). Porcine alveolar macrophages (PAMs) were isolated aseptically from the lungs of 3-week-old specific pathogen-free (SPF) piglets according to standard preparation methods, and the cells were collected by flushing the lungs three times with precooled PBS (2% penicillin). PAMs were washed, counted, and resuspended using a complete medium (RPMI 1640, 10% FBS, 2% penicillin, 2 mM l-glutamine) and labeled using flow cytometry with antibodies against CD172 with CD11b. All cells were cultured and maintained at 37°C with 5% CO₂. HSV-1 was propagated in Vero cells and titrated as described previously (49). Recombinant vesicular stomatitis virus expressing green fluorescent protein (VSV-GFP) was generated as described previously (50).

Virus experiments.

ASFV SY18 was provided by the Changchun Veterinary Research Institute, Chinese Academy of Agricultural Sciences. ASFV SY18 is denoted by ASFV in the text, and quantification of the virus was determined using the 50% blood adsorbent dose (HAD₅₀) assay. ASFV cultivation and cell experiments were conducted in an animal biosafety level 3 lab (ABSL-3).

Construction and transfection of plasmids.

Gene synthesis was performed based on the L83L sequence in ASFV SY18 (GenBank: MH766894) and cloned into the EcoRI/XhoI site of the pCMV-N-HA vector to create the pCMV-N-HA-L83L expression plasmid. The promoter-reporter gene plasmids expressing firefly luciferase (IFN-β-Luc, ISRE-Luc, and NF-κB-Luc), the internal reference reporter gene plasmid expressing Renilla (pRL-TK), and the expression plasmids of signaling molecules with Flag- or Myc-tagged (cGAS, STING, TBK1, IKKε, IRF3, NBR1, OPTN, p62, NDP52, and Tollip) have been described previously (51). All constructed plasmids were confirmed by DNA sequencing.

Antibodies and reagents.

Anti-Flag horseradish peroxidase (HRP), anti-Flag agarose affinity gels, anti-HA agarose affinity gels, and anti-Myc agarose affinity gels were purchased from Sigma, anti-HA-HRP and anti-Myc-HRP were purchased from Roche. Anti-pTBK1/NAK (Ser172), anti-pIRF-3 (Ser396), anti-pIκBα (Ser32), anti-TBK1/NAK, and anti-SQSTM1/p62 were purchased from Cell Signaling Technology. Anti-IRF3 and anti-IκBα were purchased from Beyotime Biotechnology. Anti-GAPDH, anti-beta-actin, anti-Tollip, and anti-TMEM173/STING were purchased from Proteintech. Anti-ASFV/p72 and ASFV/p54 were purchased from Zoonogen, and anti-LC3B was purchased from MBL. The poly(dG:dC) (tlrl-pgcn), 2'3'-cGAMP (tlrl-nacga23-02), HSV-60 (tlrl-hsv60n), and VACV-70 (tlrl-vav70n) were purchased from Invitrogen. DMSO, the autophagy inhibitor 3-methyladenine (3-MA), and NH₄Cl were purchased from Sigma. The autophagy inhibitor Chloroquine phosphate (CQ), proteasome inhibitor MG-132, and the caspase inhibitor Z-VAD-FMK (ZVAD) were purchased from MCE. Lipofectamine 3000 and porcine IFN-β commercial ELISA kit were supplied by Thermo Fisher. The jetPRIME-Macrophage kit was purchased from Polyplus Transfection.

Dual-luciferase reporter assays.

HEK293T cells were uniformly spread in 24-well plates (1 × 10⁵ cells/well) overnight, and plasmid transfection was performed when cells reached 70% confluence. The pCMV-N-HA and pcDNA3.1-Flag plasmids were used as negative-control plasmids for the experiments. IFN-β-Luc, ISRE-Luc, NF-κB-Luc, pRL-TK plasmid, and the indicated dose of L83L expression plasmid were cotransfected cells for 24 h. The Dual-Luciferase Reporter (DLR) Assay System was purchased from Promega. Cell samples were collected at the indicated times and lysed on ice for 10 min using lysis solution (equilibrated at room temperature) according to the manufacturer's instructions, followed by detecting relevant reporter gene activity using firefly dual luciferase and Renilla luciferase reagents.

qPCR assay.

HEK293T cells or PAMs were evenly spread in 6-well plates (1 × 10⁶ cells/well) overnight, and plasmid transfection was performed when cells reached 70% confluence. pCMV-N-HA and pcDNA3.1-Flag were used as negative-control plasmids for the experiments. Cells were cotransfected with the indicated signaling molecule plasmids and pCMV-HA-L83L plasmid for 24 h. For the infection or addition of DNA analog experiments, ASFV or the DNA analogs indicated for transfection were added after transfection, and incubation was continued for 12 h or 24 h. Cells were collected at the indicated times and lysed by adding 1 mL TRIzol (TaKaRa). Total RNA extraction was performed according to the manufacturer's instructions, and 1 μg of RNA was reverse transcribed into cDNA using reverse transcription Moloney mouse leukemia virus (M-MLV) reverse transcriptase (Promega). qPCR was performed using SYBR green Mix (TaKaRa) in an Applied Biosystems 7500 real-time PCR system. The mean mRNA fold changes relative to the control were calculated using the 2^−△△CT method. The primers are listed in Table 1.

TABLE 1.

The primer sequences for q-PCR

Primers	Sequence (5′→3′)
Hu IFNB1-F	CAGCAATTTTCAGTGTCAGCAAGCT
Hu IFNB1-R	TCATCCTGTCCTTGAGGCAGTAT
Hu Rantes-F	GGCAGCCCTCGCTGTCATCC
Hu Rantes-R	GCAGCAGGGTGTGGTGTCCG
Hu ISG56-F	ACGGCTGCCTAATTTACAGC
Hu ISG56-R	AGTGGCTGATATCTGGGTGC
Hu IFIT2-F	AAGCACCTCAAAGGGCAAAAC
Hu IFIT2-R	TCGGCCCATGTGATAGTAGAC
Hu GAPDH-F	AAAATCAAGTGGGGCGATGCT
Hu GAPDH-R	GGGCAGAGATGATGACCCTTT
Sus IFNB1-F	GCTAACAAGTGCATCCTCCAAA
Sus IFNB1-R	AGCACATCATAGCTCATGGAAAGA
Sus ISG15-F	GATCGGTGTGCCTGCCTTC
Sus ISG15-R	CGTTGCTGCGACCCTTGT
Sus ISG56-F	AAATGAATGAAGCCCTGGAGTATT
Sus ISG56-R	AGGGATCAAGTCCCACAGATTTT
Sus GAPDH-F	ACATGGCCTCCAAGGAGTAAGA
Sus GAPDH-R	GATCGAGTTGGGGCTGTGACT
ASFV P72-F	CCCAGGRGATAAAATGACTG
ASFV P72-R	CACTRGTTCCCTCCACCGATA
ASFV P30-F	ATTCTTCTTGAGCCTGATG
ASFV P30-R	GGTAGCCTGTATAATTGGTT
ASFV L83L-F	GATTATAGAGCTGAGCCTGAT
ASFV L83L-R	TTGGTCCTTCTGGAACATC

Western blotting.

HEK293T cells or PAMs were spread uniformly in 6- or 24-well plates and incubated overnight until cells reached 70% confluence for plasmid transfection. Cells were cotransfected with the indicated plasmids for 24 h. For experiments with ASFV infection or DNA analog stimulation, the corresponding stimuli were added after transfection, and incubation was continued for 12 h or 24 h. The relevant inhibitor was added 6 h before sample collection. Cells were collected at the indicated times and then lysed using RIPA buffer (Thermo Fisher) to release proteins. The total protein was boiled for 10 min in 5 × SDS buffer according to the ratio, and electrophoresis was performed. Protein blots were transferred using NC membranes (Merck Millipore), which were then incubated at room temperature in 5% skimmed milk for 1 h. Afterward, the membranes were incubated overnight at 4°C with the relevant antibodies mentioned above. Immunoblots were observed using the Amey Imager 600RGB and quantified using ImageJ.

Coimmunoprecipitation (Co-IP).

HEK293T or PK-15 cells were evenly spread in 6-well plates and incubated overnight, and plasmid transfection was performed when the cell confluence reached 70%. The indicated signaling molecule plasmids, including cGAS, STING, TBK1, and IRF3, or cargo receptor plasmids, including NBR1, OPTN, p62, NDP52, and Tollip, were cotransfected with the pCMV-HA-L83L plasmid for 24 h. Cell samples were collected at the indicated times and lysed on ice for 20 min using NP-40 lysis buffer (20 mM Tris-HCl, 150 mM NaCl, 1% NP-40, and 1 mM EDTA) to release proteins, followed by centrifugation to collect the supernatant (12,000 rpm, 20 min) and incubation of the supernatant with anti-Flag agarose affinity gels at 4°C overnight. The IP complexes and whole-cell lysates were subjected to protein transfer by the conventional plot method, and protein development was detected using the above-mentioned antibodies.

RNA interference.

Predesigned siRNA oligomers were obtained from the Sigma-Aldrich website based on the ASFV SY18 L83L and porcine Tollip mRNA sequence published in GenBank. Specific siL83L or siTollip with unrelated control siNC was transfected with the jetPRIME-Macrophage kit for 24 h, followed by ASFV (multiplicity of infection = 1) infection or no infection treatment for 12 or 24 h. Knockdown efficiency was then verified by reverse transcription qPCR or immunoblotting. The siRNA primers are listed in Table 2.

TABLE 2.

siRNA sequences used in this study

Primers	Sequence (5′→3′)
siL83L-86-F	CAAACGAUGUAUUAGAUGUdTdT
siL83L-86-R	ACAUCUAAUACAAUCGUUUGdTdT
siL83L-97-F	UUAGAUGUUACUAAAUAUAdTdT
siL83L-97-R	UAUAUUUAGUAACAUCUAAdTdT
siTollip-284-F	CCAAGAACCCGCGCUGGAAdTdT
siTollip-284-R	UUCCAGCGCGGGUUCUUGGdTdT
siTollip-388-F	CGCAUCGCCUGGACGCAUGdTdT
siTollip-388-R	CAUGCGUCCAGGCGAUGCGdTdT
siTollip-502-F	CUGGUCAUGUCCUACACGUdTdT
siTollip-502-R	ACGUGUAGGACAUGACCAGdTdT
siNC-F	UUCUCCGAACGUGUCACGUTT
siNC-R	ACGUGACACGUUCGGAGAATT

VSV-GFP bioassay.

Antiviral cytokine secretion bioassays were as previously described (22). Briefly, HEK293T or PAMs were cultured in 12-well plates overnight, and supernatants from cells infected with ASFV for 12 or 24 h and inactivated were added to fresh confluent cells incubated for 24 h. Cells were infected with VSV-GFP at a multiplicity of infection of 0.01. At 12 h postinfection, VSV-GFP replication was monitored by fluorescence microscopy or flow cytometry to examine GFP expression levels.

Confocal microscopy.

Confocal microscopy was performed as previously described (22). HEK293T or PAMs were seeded in 12-well plates (5 × 10⁵ cells/well). After transfection for 24 h, cells were fixed with 4% paraformaldehyde for 15 min at room temperature and then gently washed three times with PBS. Cells were blocked and permeabilized with 0.5% Triton X-100 5% skimmed milk at 4°C overnight. Finally, cells were incubated with the indicated primary and secondary antibodies and DAPI. Images were acquired with a Zeiss microscope (LSM 710) and 20 × objectives to visualize stained cells. Fluorescence intensity profiles of the indicated proteins were measured using the Zen Blue program.

Statistical analysis.

Data are presented as the mean ± SD unless otherwise indicated. The Student's t test was used for all statistical analyses with GraphPad Prism 8 software. Differences between groups were considered significant when the P-value was < 0.05 (*), <0.01 (**), and < 0.001 (***).