African Swine Fever Virus pI215L Negatively Regulates cGAS-STING Signaling Pathway through Recruiting RNF138 to Inhibit K63-Linked Ubiquitination of TBK1 Free

This article is featured in Top Reads, p.

African swine fever (ASF), a lethal hemorrhagic disease in domestic pigs and wild boar, has spread into many areas of the world, including the Caucasus region, Eastern Europe, the Russian Federation, South America, and the Caribbean. Currently, ASF is a severe plague in Asian countries and threatens the world with huge economic and ecological consequences (1, 2). To date, there is no safe, effective vaccine. Therefore, curative treatment and disease control strategies preventing its spread are desperately needed. African swine fever virus (ASFV), the causative agent of ASF, is a large, enveloped dsDNA virus. More than 30 complete genome sequences are currently available, and the genomes of these ASFVs range in length from ∼170–190 kb, varying among the virus strains (3–6). Normally, the ASFV genome contains between 151 and 167 open reading frames (7), which encode more than 150 proteins (8, 9). Besides the structural proteins dedicated to viral assembly and viral entry, the ASFV genome also encodes many nonessential proteins that have important roles in evading host antiviral immune responses (10), including inhibition of type I IFN production (11, 12) and apoptosis (13). Until now, about half of ASFV genes lack any known or predictable functions.

It has been reported that pigs recovered from infection with less virulent ASFV isolates had an acute and persistent ASFV in the serum, suggesting that ASFV has effective mechanisms to evade host antiviral defense systems. Previous studies showed that virulent ASFV Armenia/07 infection inhibits the cGMP-AMP (cGAMP) synthase (cGAS)-STING pathway by impairing STING activation, whereas the NH/P68 attenuated strain NH/P68 infection efficiently activates the cGAS-STING pathway and induces the production of high amounts of IFN-β (14). However, the ASFV-encoded proteins, except for the multigene family (MGF) proteins involved in regulation of the cGAS-STING pathway, are not fully understood.

As a general cytosolic DNA sensor, cGAS recognizes viral dsDNA upon DNA virus infection and uses ATP and GTP to synthesize the second messenger cGAMP (15, 16). cGAMP binds to the endoplasmic reticulum (ER)-localized adapter protein STING (17, 18), which then activates TANK-binding kinase 1 (TBK1) to phosphorylate IFN regulatory factor 3 (IRF3). The active IRF3 translocates to the nucleus, where it initiates transcription of type I IFN and other antiviral effector genes (19). As a critical kinase, TBK1 is required for the phosphorylation of IRF3, STING, and TIR domain-containing adapter-inducing IFN-β (20). Therefore, TBK1 activity is fine-tuned during viral infection. Several E3 ubiquitin ligases have been reported to regulate TBK1 activity by regulating its polyubiquitination. For example, ubiquitin E3 ligase TNFR-associated factor–interacting protein (21), DTX4 (22), and TRIM27 (23) negatively regulate innate antiviral responses by promoting the K48-linked polyubiquitination and subsequent degradation of TBK1, whereas E3 ubiquitin ligase RING finger protein 128 (RNF128) (24) and MIB2 (25) enhance host antiviral immune responses through promoting the K63-linked polyubiquitination of TBK1 for functional activation.

ASFV pI215L is encoded by ASFV early gene I215L, which is expressed in viral factories and throughout the nucleus and cytoplasm of infected cells (26). ASFV pI215L consists of 212 aa with m.w. of 24 kDa, which shares significant homology with a family of ubiquitin-conjugating enzymes (E2s) (27). The residue cysteine-85 (C85) is essential for the E2 activity of pI215L. It has been demonstrated that ASFV pI215L plays a critical role in the transcription of late viral genes and viral DNA replication (9). Previous studies reported that several ubiquitin-conjugating enzymes (E2s), together with E3 ligases, regulate type I IFN signaling pathways. For example, the UBE2D3 (Ubc5c) and UBE2N (Ubc13), together with ubiquitin E3 ligase Riplet, are involved in polyubiquitin-mediated activation of RIG-I and MAVS for triggering innate immune signaling in response to viral infection (28). To date, the functions of pI215L in regulating type I IFN production have not been investigated.

In this study, we report that ASFV pI215L was one of the strongest inhibitors in modulating the type I IFN production by antagonizing cGAS-STING pathway. ASFV pI215L inhibited K63-linked polyubiquitination of TBK1, which was independent of its E2 enzyme activity. Mechanistically, pI215L recruited E3 ubiquitin ligase RNF138 to degrade RNF128 to synergistically suppress RNF128-mediated K63-linked polyubiquitination of TBK1. Of note, knockdown of pI215L expression enhanced type I IFN production and inhibited ASFV replication. Our findings provide a new clue to explain how ASFV escapes host antiviral immune responses.

The protein agarose A/G used for coimmunoprecipitation (Co-IP) was purchased from Santa Cruz Biotechnology (20397; Dallas, TX). The cGAMP (SML1232) was purchased from Sigma-Aldrich (St. Louis, MO). The following Abs were purchased from Cell Signaling Technology (Danvers, MA): anti-Flag (D6W5B, 14793), anti-HA (C29F4, 3724), anti-Myc (71D10, 2278), and anti–p-TBK1 (D52C2, 5483S). Anti-RNF138 polyclonal Ab (pAb) was purchased from Abclonal (A10304). Anti–HSV type 1 (anti–HSV-1) gB (ab6506) was purchased from Abcam (Cambridge, UK). Anti-swine TBK1 pAb was generated by GenScript (Nanjing, China). Anti-pI215L pAb was produced in rabbit and mouse by immunization with recombinant pI215L. Anti-mouse IgG (H+L), DyLight 800-Labeled (042-07-18-06), was purchased from Sera Care (Milford, MA). The IRDye 800CW goat anti-rabbit IgG secondary Ab (926-32211) was purchased from LI-COR Biosciences (Lincoln, NE). IPKine HRP mouse anti-rabbit IgG LCS (A25022) and IPKine HRP goat anti-mouse IgG HCS (A25112) were purchased from Abbkine (Wuhan, China).

The IFN-β reporter, IFN stimulation response element (ISRE) reporter, and thymidine kinase (TK)-Renilla reporter were obtained from Professor Hong Tang. The 102 cDNAs corresponding to ASFV-encoded proteins were synthesized on the basis of the genome of the ASFV HLJ/18 isolate (29) and cloned into pCAGGS-Flag (pFlag) vector from GenScript. To construct plasmids expressing Flag-tagged or HA-tagged RNF128, RNF138, cGAS, STING, IRF3, and TBK1, the cDNAs corresponding to these swine genes were amplified by standard RT-PCR using total RNA extracted from porcine alveolar macrophages (PAMs) as templates, and the cDNAs were then cloned into the pFlag or pCAGGS-HA (pHA) vector, respectively. All constructs were validated by DNA sequencing. The primers used in this study are listed in Table III.

HEK293T cells were purchased and cultured in DMEM, and PAMs isolated from specific pathogen-free piglets (without ASFV, porcine reproductive and respiratory syndrome virus, pseudorabies virus, porcine circovirus type 2, and 28 other pathogens) were cultured in RPMI 1640 supplemented with 10% FBS, 100 U/ml penicillin, and 100 μg/ml streptomycin at 37°C in 5% CO₂. The ASFV HLJ/18 strain was isolated from a pig sample from a farm in northeastern China during the ASF outbreak in 2018 (2). HSV-1 was kindly provided by Prof. Hongbin Shu (Wuhan University, Wuhan, China).

The HAD₅₀ assay was performed as previously described (2). Briefly, PAMs grown in a 96-well plate were infected with 0.1 ml/well of 10-fold serially diluted supernatant in quintuplicate. After incubation for 90 min at 37°C, unattached virus was removed, and DMEM supplemented with 2% FBS was added to the PAMs. At 5 days after infection, the HAD₅₀ was determined by the Reed-Muench method. All data are shown as the means of three independent experiments.

Cell viability was assessed by the MTT assay (30). Briefly, PAMs cultured in 96-well plates were infected with or without ASFV at a multiplicity of infection (MOI) of 1 and then mock stimulated or stimulated with cGAMP for 4 h, 8 h, 12 h, and 24 h, respectively, and then 200 μl DMEM containing 20 μl MTT (5 mg/ml) was added to each well and incubated for 4 h at 37°C. The culture medium was then removed, and 100 μl DMSO was added to each well, and the plate was agitated on an orbital shaker for 15 min. The light absorbance in each sample was measured at 570 nm with an ELISA reader.

CRISPR/Cas9 genomic editing for gene deletion was used as previously described (31). To create the mammalian Rnf138 gene knockout cell line (Rnf138^−/−), one CRISPR single-guide RNA (sgRNA) sequence targeting the Rnf138 locus in the genome was chosen on the basis of specificity scores (http://crispr.mit.edu/). The sgRNA sequence used was as follows: 5'-GGGGCAGTAGAAATCATCTT-3' and 5'-GTAGAAATCATCTTCGGTGT-3'. The sgRNA sequence was cloned into the pSpCas9 (BB)-2A-GFP plasmid (pX458; Addgene, Watertown, MA). The construct was then independently transfected into HEK293T cells. The cells expressing GFP were isolated by flow cytometry, and single cells were seeded into separate wells of 96-well plates by the flow cytometer. After clonal expansion, RNF138 protein expression levels in different clones were analyzed by immunoblotting, and the genomic DNA from those clones that had undetectable RNF138 protein expression was extracted and amplified by PCR for Rnf138 gene sequencing.

The target sequences of small interfering RNAs (siRNAs) are listed in Table II. The transfection of siRNA was performed with HiPerFect Transfection Reagent (QIAGEN, Germantown, MD) following the manufacturer’s instructions. Forty-eight hours after siRNA transfection, cells were infected with ASFV at an MOI of 1.0 for 24 h. The siRNA knockdown efficiency of the target protein was assessed by quantitative real-time PCR (qPCR).

Co-IP and immunoblot analysis were performed as previously described (32). Briefly, for Co-IP, whole-cell extracts were lysed in lysis buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 5 mM MgCl₂, 1 mM EDTA, 1% Triton X-100, and 10% glycerol) containing 1 mM PMSF and a 1× protease inhibitor mixture (Roche, Basel, Switzerland). Then, cell lysates were incubated with anti-Flag (M2) beads or with Protein G PLUS-Agarose immunoprecipitation reagent (Santa Cruz Biotechnology) together with 1 μg of the corresponding Abs at 4°C overnight on a roller. The precipitated beads were washed five times with cell lysis buffer. For immunoblot analysis, equal amounts of cell lysates and immunoprecipitants were resolved by 10–12% NaDodSO₄ PAGE (SDS-PAGE) and then transferred to a polyvinylidene difluoride membrane (MilliporeSigma, Burlington, MA). After incubation with primary and secondary Abs, the membranes were visualized by ECL (Thermo Fisher Scientific, Waltham, MA) or an Odyssey two-color infrared fluorescence imaging system (LI-COR Biosciences).

BL21 (DE3) or Rosetta Escherichia coli were transformed with either a pGEX-6P-1 empty vector or a plasmid pGEX-6P-1-I215L, and the recombinant bacteria were induced to express GST and GST-pI215L by adding 0.25 mM isopropyl β-d-thiogalactoside for 20 h at 16°C. Bacterial pellets were resuspended in 1× PBS (pH 7.4) with 1 mM PMSF and homogenized by sonication. Bacterial lysates were centrifuged at 12,000 × g for 30 min at 4°C, and the supernatants were immobilized on glutathione beads (GE Healthcare Life Sciences, Chicago, IL) that were prepared by washing with binding buffer. The conjugated beads were blocked for 1 h with 3% BSA. The Flag-RNF138 protein expressing in HEK293T cells was then incubated with either GST or GST-pI215L and immobilized on glutathione beads for 2 h at room temperature. The beads were washed five times, and the proteins were treated with 1× SDS-PAGE loading buffer and detected by Western blotting with an anti-GST or anti-Flag pAb.

Luciferase activities were measured with a Dual-Luciferase Reporter Assay System (Promega, Madison, WI) according to the manufacturer’s instructions. The data were normalized to the transfection efficiency by dividing the firefly luciferase activity by the Renilla luciferase activity.

To detect IFN mRNA levels, total RNA was extracted using TRIzol reagent (Invitrogen, Waltham, MA), and reverse transcription was performed with a PrimeScript RT Reagent Kit (Takara, Shiga, Japan). Reverse transcription products were amplified using a Stratagene Mx Real-Time qPCR system (Agilent, Santa Clara, CA) with SYBR Premix Ex Taq II (Takara) according to the manufacturer’s instructions. Data were normalized to the level of β-actin expression in each individual sample. For ASFV genomic DNA detection, ASFV genomic DNA was extracted by using QIAamp DNA Mini Kit (QIAGEN). qPCR was carried out on a QuantStudio5 system (Applied Biosystems, Foster City, CA) according to the World Organization for Animal Health–recommended procedure described by King et al. (33). All the qPCR primers are listed in Tables I.

The concentrations of IFN-α and IFN-β in the cell culture supernatants were measured by commercial ELISA kits specifically detecting IFN-α (RayBiotech, Peachtree Corners, GA) and IFN-β (Biorbyt, Cambridge, UK) according to the manufacturers’ instructions.

All statistical analyses were performed using one-way ANOVA via the SPSS 16.0 software package (SPSS, Chicago, IL). Data were expressed as the mean ± SD. A p value < 0.05 was considered statistically significant.

To detect whether the ASFV HLJ/18 strain induces type I IFN production, PAMs were infected with ASFV at MOIs of 0.01, 0.1, and 1 for 24 h, respectively, and then the cells were collected to detect the mRNA levels of IFN-α, IFN-β, IFN-stimulated gene 56 (ISG56), and ASFV genomic DNA copy number using qPCR, and the cell supernatants were collected to detect the levels of IFN-α and IFN-β using ELISA, respectively (Table I). As shown in (Fig. 1A, 1B, 1D, and 1E, ASFV infection induced low levels of mRNA and protein expression of both IFN-α and IFN-β, despite the increasingly high infection doses (Fig. 1F), compared with that induced by HSV-1 infection. The same result was found for ISG56, one of a large number of genes triggered by type I IFNs (Fig. 1C). To further study the effect of ASFV on type I IFN production, the PAMs were first stimulated with cGAMP, a potent inducer of type I IFN for signal amplification, and then infected with ASFV at an MOI of 1 for 4 h, 8 h, 12 h, and 24 h, respectively, and finally collected to detect the mRNA levels of IFN-β and ISG56. The results showed that cGAMP treatment alone induced high levels of IFN-β (Fig. 1G) and ISG56 (Fig. 1H) mRNA expression; however, ASFV infection dramatically blocked the induction of IFN-β and ISG56 mRNA expression triggered by cGAMP. The inhibitory effect of ASFV infection on cGAMP induction of IFN-β and ISG56 mRNA expression was unlikely due to impaired cell viability, because there were no significant differences in cell viability seen between the different treatments (Fig. 1I). Taken together, these results indicate that ASFV suppresses type I IFN production following infection.

The cGAS-STING signaling pathway functions to sense cytosolic DNA and triggers type I IFN-mediated immune responses against DNA virus infections (14). Therefore, we focused our efforts to find which ASFV-encoded protein inhibits type I IFN production boosted by activation of the cGAS-STING signaling pathway. We performed an unbiased screening of the role of 102 ASFV-encoded proteins involved in inhibition of type I IFN production. We found that pI215L, pE301R, pD345L, and pMGF505-7R showed the most significant inhibitory role in cGAS-STING–mediated type I IFN production (Supplemental Fig. 1A). To confirm the inhibitory function of the four ASFV proteins, HEK293T cells were transfected with an IFN-β luciferase (Luc) reporter, an internal control Renilla-TK Luc, and two plasmids individually expressing cGAS and STING, together with different doses of a plasmid expressing Flag-tagged pI215L, pE301R, pD345L, and pMGF505-7R. As shown in (Fig. 2A and Supplemental Fig. 1B–1D, the ectopic expression of pI215L, pE301R, pD345L, and pMGF505-7R all inhibited cGAS- and STING-induced IFN-β promoter activities in a dose-dependent manner. In addition, we also noticed that pI215L inhibited cGAS- and STING-induced ISRE (Fig. 2B), NF-κB (Fig. 2C), and ISG56 (Fig. 2D) promoter activities in a dose-dependent manner. Because pI215L showed the strongest inhibitory effect on IFN-β among the 102 ASFV proteins and also suppressed cGAS- and STING-induced ISRE, NF-κB, and ISG56 promoter activities, we chose to focus on pI215L to explore its inhibitory role and underlying mechanisms in type I IFN production for the remainder of the research.

To further confirm the inhibitory effects of pI215L on type I IFN signaling, we transfected HEK293T cells with plasmids expressing cGAS and STING and together with different doses of a plasmid expressing Flag-pI215L, and then we collected the cells for detection of the mRNA levels of IFN-β and ISG56 by qPCR. The results showed that ectopically expressed pI215L significantly decreased the mRNA levels of IFN-β (Fig. 2E) and ISG56 (Fig. 2F) induced by cGAS and STING. To test the effect of pI215L on type I IFN production during ASFV infection, we tried to modify ASFV for deletion of the I215L gene (ASFV-ΔI215L) using CRISPR/Cas9. Unfortunately, we failed to generate the ASFV-ΔI215L, indicating that the I215L gene is necessary for ASFV replication. Therefore, we turned to siRNAs to decrease the I215L levels during viral infection and synthesized five siRNAs targeting I215L, with three of them being able to efficiently reduce pI215L expression (Fig. 2G, Table II). Knockdown of pI215L expression during ASFV infection significantly reduced the virions in the infected PAMs (Fig. 2H), further supporting a critical role of the I215L gene in ASFV replication while significantly enhancing the mRNA level of IFN-β and ISG56 (Fig. 2I and 2J). Taken together, these results indicate that ASFV pI215L can significantly inhibit type I IFN production during ASFV infection.

To elucidate the underlying molecular mechanisms by which ASFV pI215L negatively regulates type І IFN production, we first assessed the effect of ASFV pI215L on IFN-β promoter activation induced by the key molecules in the cGAS-STING signaling pathway in HEK293T cells. As shown in (Fig. 3A and 3B, ectopically expressed ASFV pI215L significantly decreased the IFN-β promoter activation induced by STING or TBK1 in a dose-dependent manner. However, the IFN-β promoter activation mediated by IRF3-5D, a constitutively active IRF3 variant in which Ser396, Ser398, Ser402, Ser404, and Ser405 were replaced by phosphomimic aspartate, was not impaired by ASFV pI215L (Fig. 3C). Consistent with these results, the mRNA levels of IFN-β were also attenuated by ASFV pI215L in the HEK293T cells induced by STING (Fig. 3D) and TBK1 (Fig. 3E), but not by IRF3-5D (Fig. 3F). These results suggest that ASFV pI215L may target TBK1 or IRF3 to inhibit type І IFN production. Therefore, we further tested whether ASFV pI215L affects TBK1 phosphorylation, a key step for activation of TBK1. We first overexpressed the ASFV pI215L in the HEK293T cells and then infected them with HSV-1 to boost the type I IFN signaling. The results in (Fig. 3G and 3H clearly showed that ASFV pI215L dramatically inhibited HSV-1–induced phosphorylation of TBK1. It is reported that ubiquitination of TBK1 was required for its phosphorylation and activation (24). Therefore, we then tested whether pI215L affects the ubiquitination of TBK1. We transfected HEK293 cells with plasmids encoding Myc-tagged ubiquitin, Flag-tagged TBK1, and HA-tagged pI215L and then pulled down TBK1 using anti-Flag (M2) beads for analysis of its ubiquitin level. As shown in (Fig. 3I, pI215L almost completely blocked the ubiquitination of TBK1. To explore whether IRF3 was also targeted by pI215L for deubiquitylation, we performed the same experiment and found that pI215L had no obvious effect on IRF3 phosphorylation (Fig. 3J), suggesting that pI215L targets TBK1 for its deubiquitylation and dephosphorylation. TBK1 has been reported to be able to be ubiquitinated by either K48- or K63-linked ubiquitin chains, and K48-linked ubiquitination of TBK1 is destined for degradation, whereas K63-linked ubiquitination enhances TBK1 activation (21, 22, 24). We then tested which type of ubiquitination of TBK1 was blocked by pI215L and found that pI215L significantly inhibited K63-linked ubiquitination of TBK1 (Fig. 3K) but not K48-linked ubiquitination of TBK1 (Fig. 3L). However, we could not detect any physical interaction between ASFV pI215L and TBK1 (data not shown), indicating that pI215L functions through other proteins to target TBK1. Taken together, these data suggest that ASFV pI215L indirectly targets TBK1 for K63-linked ubiquitination and dephosphorylation.

It has been reported that pI215L is an E2 ubiquitin-conjugating enzyme. To detect whether the negative regulation of type І IFN production by ASFV pI215L is dependent on its E2 enzyme activity, we generated a mutant of ASFV pI215L (pI215L-C85A), which impaired its E2 enzyme activity (26) (Table III). The inhibitory effects of pI215L and pI215L-C85A on IFN-β and ISRE promoters were analyzed. As shown in (Fig. 4A and 4B, like pI215L, pI215L-C85A still strongly inhibited both IFN-β (Fig. 4A) and ISRE (Fig. 4B) promoter activation in a dose-dependent manner, indicating that inhibition of type І IFN production by pI215L is independent of its E2 enzyme activity. To identify which region of pI215L is necessary for its inhibition of type I IFN production, a series of plasmids expressing the truncated pI215L mutants were constructed (Fig. 4C, Table III). The results showed that overexpressed pI215L-2 (1–107 aa) and pI215L-3 (54–212 aa) significantly inhibited the IFN-β promoter activity but not that of other deleted mutants (Fig. 4D).

Because the data in (Fig. 3 suggest that pI215L functions through other partners to deubiquitinate TBK1, we speculated that pI215L may recruit deubiquitinase enzyme or ubiquitin E3 ligase to regulate the K63-linked ubiquitination of TBK1. To screen the potential effectors mediating pI215L’s deubiquitylation of TBK1, we first infected the PAMs with ASFV, pulled down the pI215L and its binding partners with anti-I215L Ab, and analyzed the binding partners with mass spectrometry. We found that a deubiquitinase, OTUD4, and two E3 ubiquitin ligases, RNF41 and RNF138, were coprecipitated with pI215L (Supplemental Fig. 2A). To confirm whether pI215L physically interacts with any of the three identified candidates, we performed the Co-IP assay between pI215L and OTUD4, RNF138, or RNF41. The results showed that pI215L coprecipitated with both E3 ubiquitin ligases RNF41 and RNF138 (Supplemental Fig. 2B and 2C), but not with the deubiquitinase OTUD4 (Supplemental Fig. 2B). Because RNF41 and RNF138 both can be coprecipitated with pI215L, we further evaluated their roles in mediating type I IFN production. Similarly, we activated the type I IFN signaling by overexpressing cGAS and STING and then tested whether RNF41 or RNF138 has any effect on type I IFN signaling. We found that RNF41 did not show obvious effect on IFN-β (Supplemental Fig. 2D) and ISRE (Supplemental Fig. 2E) promoter activities induced by cGAS-STING. However, RNF138 could dramatically suppress both IFN-β (Fig. 5C) and ISRE (Fig. 5D) promoter activities, indicating that RNF138 may be the effector mediating the functions of pI215L. Therefore, we focused on RNF138 and further confirmed their interactions. We found that Flag-tagged pI215L and HA-tagged RNF138 could coprecipitate when they were coexpressed in HEK293T cells (Fig. 5A). To further prove that pI215L and RNF138 associate with each other, we expressed and purified GST and GST-tagged pI215L in E. coli and then incubated the purified proteins with cell lysates from HEK293T cells overexpressing HA-tagged RNF138. The results showed that HA-tagged RNF138 could be pulled down by GST-tagged pI215L (Fig. 5B), suggesting that pI215L and RNF138 interact with each other. Because RNF138 could dramatically suppress both IFN-β and ISRE promoter activities triggered by cGAS-STING, we further explored whether RNF138 can block the type I IFN signaling induced by TBK1 and IRF3-5D. Interestingly, similar to pI215L, RNF138 dose-dependently inhibited TBK1-induced IFN-β promoter activity, but not IRF3-5D (Fig. 5E and 5F). These results strongly suggest that RNF138 may be the effector protein targeting TBK1 to inhibit type I IFN production. The RING domain is critical for the E3 ligase activity, which prompted us to explore whether the RING domain of RNF138 is essential for its suppression of type I IFN signaling. We constructed a series of RNF138 truncated mutants (Supplemental Fig. 3A) and found that without the RING domain, RNF138 totally lost its inhibitory effect on IFN-β promoter activity induced by cGAS and STING (Supplemental Fig. 3B).

Because pI215L and RNF138 both showed strong suppression of type I IFN signaling, we next examined the effect of pI215L and RNF138 coexpression of type I IFN signaling. We individually expressed or coexpressed pI215L and RNF138 with cGAS-STING or TBK1 and found that overexpression of pI215L and RNF138 had stronger inhibitory effects on the cGAS-STING or TBK1-mediated IFN-β promoter activities than that seen with either one of the two proteins alone (Fig. 5G and 5H), indicating that pI215L and RNF138 act synergistically to suppress IFN-β promoter activity. All the data support our hypothesis that RNF138 may be the effector aiding pI215L to suppress type I IFN signaling. To further confirm the essential role of RNF138 in mediating pI215L’s inhibitory effect on type I IFN production, we generated an Rnf138 knockout HEK293T cell line (HEK293T-Rnf138^−/−). As we expected, loss of RNF138 completely blocked the inhibitory effect of pI215L on IFN-β promoter activity induced by cGAS-STING and TBK1, whereas reexpression of RNF138 in the HEK293T-Rnf138^−/− cells rescued pI215L’s inhibitory effect on IFN-β promoter activity (Fig. 5I and 5J), suggesting that the inhibition of type I IFN production by pI215L is dependent on RNF138.

Taken together, these results demonstrate that pI215L interacts with and functions through RNF138 to inhibit type I IFN signaling.

RNF138 is a RING-type E3 ubiquitin ligase that has been reported to promote the ubiquitination of several substrates (34, 35). To detect whether RNF138 regulates the ubiquitination of TBK1, HA-tagged RNF138 was coexpressed with Flag-tagged TBK1 and Myc-tagged ubiquitin or ubiquitin-K63 (all lysines in ubiquitin were mutated except the 63rd lysine) or ubiquitin-K48 (all lysines in ubiquitin were mutated except the 48th lysine) with or without HA-tagged pI215L, and TBK1 was then pulled down by anti-Flag Ab to evaluate its ubiquitination. The results showed that either one of RNF138 and pI215L strongly suppressed TBK1 ubiquitination, and coexpression of both showed a much stronger inhibitory effect that totally blocked TBK1 ubiquitination (Fig. 6A). Similarly, coexpression of RNF138 and pI215L had a much stronger inhibitory effect on TBK1 K63-linked ubiquitination (Fig. 6B), but no obvious effect on K48-linked ubiquitination of TBK1 (Fig. 6C). These results suggest that both RNF138 and pI215L can strongly suppress and have synergistic effects on TBK1 K63-linked ubiquitination. Previous results showed that pI215L and RNF138 could synergistically suppress cGAS-STING– and TBK1-induced IFN-β promoter activity (Fig. 5G and 5H), and without RNF138, pI215L totally lost its inhibitory effect on cGAS-STING– and TBK1-induced IFN-β promoter activity (Fig. 5I and 5J). To explore whether suppression of TBK1 K63-linked ubiquitination by pI215L is similarly dependent on RNF138, we performed the same experiment in the HEK293T-Rnf138^−/− cells and found that pI215L totally lost its inhibitory effect on K63-linked polyubiquitination of TBK1 in HEK293T-Rnf138^−/− cells (Fig. 6D, lane 4) and regained the suppression when RNF138 was reexpressed in HEK293T-Rnf138^−/− cells (Fig. 6D, lane 6), suggesting that pI215L inhibits the K63-linked polyubiquitination of TBK1 in an RNF138-dependent manner (Fig. 6D).

Currently, only E3 ubiquitin ligase RNF128 was reported to induce the K63-linked ubiquitination of TBK1 upon DNA virus infection (24). Therefore, we detected whether RNF128 was involved in pI215L’s inhibition of the K63-linked polyubiquitination of TBK1. Similarly, HA-tagged RNF128 was coexpressed with Flag-tagged TBK1 and Myc-tagged ubiquitin or ubiquitin-K63 with or without HA-tagged pI215L, and TBK1 was then pulled down by anti-Flag Ab to evaluate its ubiquitination. The results showed that RNF128 promoted the overall polyubiquitination and K63-linked polyubiquitination of TBK1 and that pI215L inhibited RNF128-mediated K63-linked polyubiquitination of TBK1 (Fig. 6E and 6F), suggesting that pI215L and RNF128 have opposite effects on TBK1 ubiquitination and that pI215L can totally block RNF128’s effect when both are coexpressed.

Taken together, these data demonstrate that pI215L functions through RNF138 to inhibit TBK1 K63-linked ubiquitination.

To clarify the role of RNF128 in ASFV pI215L-mediated inhibition of the type I IFN production, we examined the interaction between ASFV pI215L protein and RNF128. When the two proteins were both expressed in HEK293T cells, HA-RNF128 could be coimmunoprecipitated by Flag-I215L (Fig. 7A). In addition, overexpressed Flag-I215L coimmunoprecipitated with endogenous RNF128 in HEK293T cells (Fig. 7B), whereas endogenous RNF128 interacted with ASFV pI215L in PAMs upon ASFV infection (Fig. 7C). These results suggested that there existed a constitutive interaction between the two proteins. To elucidate the mechanisms that both RNF138 and RNF128 were involved in ASFV pI215L inhibiting the K63-linked polyubiquitination of TBK1, Co-IP was performed to analyze the interaction among the three proteins, and the results showed that RNF138 interacted with the RING domain of RNF128 (Fig. 7D and 7E and Supplemental Fig. 4A and 4B), and overexpressed pI215L enhanced the interaction between RNF138 and RNF128 (Fig. 7F and 7G). To further confirm these results, PAMs were infected with ASFV or not, and Co-IP was performed at 24 h postinfection. The results showed that endogenous RNF128 interacted with RNF138, and ASFV infection enhanced the interaction (Fig. 7H and 7I). Knockout of Rnf138 inhibited the interaction between Flag-pI215L and HA-RNF128 (Fig. 7J and 7K). All these results indicated that pI215L enhanced the interaction between RNF138 and RNF128.

To further clarify the mechanism of ASFV I215L inhibiting RNF128-mediated K63-linked polyubiquitination of TBK1, plasmids encoding HA-I215L and Flag-RNF128 were cotransfected into HEK293T cells, and we found that pI215L decreased RNF128 expression in a dose-dependent manner (Fig. 8A and 8B). ASFV pI215L is an E2 ubiquitin-conjugating enzyme and it has no E3 ubiquitin ligase enzyme activity. Therefore, we speculated that pI215L may recruit RNF138 to degrade RNF128. As expected, RNF138 decreased RNF128 expression in a dose-dependent manner (Fig. 8C and 8D), and MG132 rescued the decrease (Fig. 8E and 8F). These results demonstrated that RNF138 degraded RNF128 through a ubiquitin proteasome pathway. Additionally, we found that Flag-pI215L promoted the degradation of RNF128 mediated by RNF138 (Fig. 8G and 8H). To further verify the results, a plasmid encoding Flag-pI215L was transfected alone or together with HA-RNF128 into HEK293T cells or HEK293T-Rnf138^−/− cells, and Western blot analysis showed that Flag-pI215L decreased both overexpressed HA-RNF128 and endogenous RNF128 in HEK293T cells, but not in HEK293T-Rnf138^−/− cells (Fig. 8I–8L), which suggested that the degradation of RNF28 by pI215L was dependent on RNF138. Consistent with these results, we also noticed that ASFV infection decreased RNF128 expression. However, the inhibitory effect was weakened when RNF138 was knocked down (Fig. 8M, 8N, Table I). We conclude that, taken together, our results show that ASFV pI215L recruits RNF138 to degrade RNF128.

In summary, the E3 ligase RNF138 degrades RNF128, which is required for K63-linked polyubiquitination of TBK1. ASFV pI215L enhances the interaction between RNF138 and RNF128 and promotes RNF138 to degrade RNF128, which results in decreased K63-linked polyubiquitination of TBK1 and type I IFN production (Fig. 9).

An ASF case was first reported in China in 2018. ASF has caused huge economic and ecological consequences. The ASFV HLJ/18 strain is the first ASFV isolate in China and has been proved to be highly virulent in domestic pigs (2). A previous study reported that virulent Armenia/07 virus infection blocks the synthesis of IFN-β through impairing the cGAS-STING signaling pathway (14). Consistent with this result, we found that ASFV HLJ/18 infection induced low-level type І IFN production in PAMs compared with HSV-1 infection (Fig. 1). An unbiased method was used to screen the candidate ASFV-encoded proteins involved in inhibiting type І IFN production via antagonizing the cGAS-STING signaling pathway, and pI215L was identified as one of the strongest inhibitors. Overexpression of pI215L significantly inhibited the activities of the IFN-β, ISRE, NF-κB, and ISG56 promoters and the mRNA levels of the IFN-β and ISG56 induced by cGAS-STING (Fig. 2). Further study showed that ectopic expression of pI215L significantly inhibited STING- and TBK1- but not IRF3/5D-mediated activation of the IFN-β promoter.

Recently, Barrado-Gil et al. showed that pI215L interacts with components of the host translation machinery, which is involved in shutoff of host protein synthesis (36). However, HEK293T cells were transfected with plasmids encoding Flag-pI215L and HA-tagged immune molecules in the cGAS-STING pathway, and Flag-pI215L had no inhibitory effect on the expression of these immune molecules, suggesting that inhibition of type I IFN production by pI215L was independent of its role of shutting off protein synthesis. Mechanistically, pI215L significantly inhibited the K63-linked polyubiquitination of TBK1 by recruiting ubiquitin E3 ligase RNF138 (Fig. 9), which resulted in inhibiting TBK1 phosphorylation and type I IFN production.

It has been widely accepted that the cGAS-STING signaling pathway plays a crucial role in host antiviral defense upon DNA virus infection (37). DNA viruses have evolved various mechanisms to antagonize the cGAS-STING signaling pathway for efficient infection and replication (38). For example, HSV-1 tegument protein UL41 directly degrades cGAS mRNA (39). Kaposi sarcoma herpesvirus protein ORF52 and its cytoplasmic isoforms of latency-associated nuclear antigen counteract the cGAS-STING pathway through binding to cGAS (40, 41). Human CMV protein UL31 disassociates DNA from cGAS and then inhibits cGAS enzymatic functions, which reduces cGAMP production. Of note, human CMV UL82 not only inhibits the translocation of STING from the ER to perinuclear microsomes by disrupting the STING-iRhom2-TRAPβ translocation complex but also impairs the recruitment of TBK1 and IRF3 to the STING complex (42). ASFV is a large dsDNA virus, and the main natural target cells of ASFV replication are macrophages and monocytes (9). To replicate in these immune cells, ASFV has developed counteracting measures against its host antiviral innate immune responses (10). Several ASFV genes were observed to modulate different steps in the cGAS-STING signaling pathway. For example, pA276R inhibits IFN-β induction via both the TLR3 and the cytosolic pathways by targeting IRF3, but not IRF7 or NF-κB (10). MGF360-12L inhibits type I IFN production by inhibiting the NF-κB signaling pathway (43). The pMGF505-7R/pA528R is able to negatively regulate the cGAS-STING signaling pathway and the JAK-STAT pathway (10, 44). As a functional viral TLR3 homolog, the pI329L suppresses type I IFN production by targeting the TIR domain-containing adapter-inducing IFN-β. Several MGF360 and MGF505/530 were also reported to inhibit the induction of type I IFNs by unknown mechanisms (10, 11, 45). However, the potential viral candidate genes, except for MGF members and their respective molecular mechanisms involved in blocking of the cGAS-STING pathway to inhibit IFN-β production, are barely studied (12). In this study, we screened and verified that several ASFV-encoded proteins, including pI215L, pE301R, pD345L and pMGF505-7R, have an inhibitory effect on cGAS-STING–mediated type I IFN production (Supplemental Fig. 1).

ASFV-encoded pI215L is a ubiquitin-conjugating E2 enzyme. However, we found that pI215L inhibition of the cGAS-STING pathway is independent of its E2 enzymic activity (Fig. 4). Interestingly, 1–107 aa and 54–212 aa of pI215L had inhibitory effects on IFN promoter activity, but not 1–160 aa of pI215L. We speculated that different truncated mutants of pI215L have different spatial structures, which needs to be verified by subsequent structural biology experiments with the truncated mutants. Recently, accumulated evidence demonstrated that recruitment of ubiquitin E3 ligases or deubiquitinases by E2s to TBK1 to modify its ubiquitination is a common strategy for the host to execute its antiviral immune responses (28). Therefore, we proposed that pI215L may recruit other effectors to exert its function. To study the molecular mechanism of pI215L inhibiting type І IFN production, immunoprecipitation–mass spectrometry was performed to identify pI215L-binding partners. We found that pI215L interacted with a ubiquitin E3 ligase, RNF138, that has been reported to inhibit IFN-β production (46). In this study, we found that RNF138 inhibited cGAS-STING–mediated type I IFN production through its RING domain. ASFV pI215L inhibited type I IFN production, and K63-linked ubiquitination of TBK1 was dependent on RNF138 (Fig. 5 and (Fig. 6). The RING domain of RNF138, which facilitates the ubiquitination of substrates, exerts a major effect on inhibition of IFN-β promoter activity induced by cGAS-STING (Supplemental Fig. 4). Subsequently, we found that RNF138 degraded RNF128, which was the unique ubiquitin E3 ligase mediating K63-linked ubiquitination of TBK1 (24) (Fig. 8E). Whether RNF138 promotes RNF128 degradation by the K48-linked polyubiquitination of RNF128 should be further studied.

In this study, we tried to generate pI215L gene deficiency ASFV (ASFV-ΔI215L) using the CRISPR/Cas9 technique to further investigate the role of pI215L on the cGAS-STING signaling pathway. Unfortunately, we could not obtain ASFV-ΔI215L, suggesting that I215L is a necessary gene for ASFV replication (data not shown). It has been reported that knockdown of pI215L expression by siRNA impairs ASFV replication (26). Consistent with these results, we also found that knockdown of pI215L expression inhibited ASFV replication. Interestingly, knockdown of pI215L expression enhanced IFN-β promoter activity, which may result in inhibiting ASFV replication.

In summary, we have shown that ASFV pI215L counteracts the cGAS-STING–mediated type І IFN production by recruiting RNF138 to degrade RNF128, which inhibits the K63-linked ubiquitination of TBK1. Our findings reveal a new strategy of immune evasion of ASFV by interfering with the cGAS-STING signaling pathway.

We thank Prof. Dongming Zhao for providing the ASFV HLJ/18 strain.

The authors have no financial conflicts of interest.