Mpox virus poxin-schlafen fusion protein suppresses innate antiviral response by sequestering STAT2

Abstract

Mpox virus (MPXV) has to establish efficient interferon (IFN) antagonism for effective replication. MPXV-encoded IFN antagonists have not been fully elucidated. In this study, the IFN antagonism of poxin-schlafen (PoxS) fusion gene of MPXV was characterized. MPXV PoxS was capable of decreasing cGAS-produced 2'3'-cGAMP, like its ortholog poxin of vaccinia virus, which is the first known cytosolic nuclease that hydrolyses the 3'-5' bond of 2'3'-cyclic GMP-AMP (cGAMP). However, MPXV PoxS did not suppress cGAS-STING-mediated type I IFN production. Instead, MPXV PoxS antagonized basal and type I IFN-induced expression of IFN-stimulated genes such as OAS1, SAMD9, SAMD9L, ISG15, ISG56 and IFIT3. Consistently, MPXV PoxS inhibited both basal and type I IFN-stimulated activity of interferon-stimulated response elements, but did not affect activation of IFN-γ-activated sites. Mechanistically, MPXV PoxS interacted with STAT2 and sequestered it in the cytoplasm. Both the viral schlafen fusion and the active site of 2'3'-cGAMP nuclease were required for STAT2 sequestration and consequent suppression of IFN-stimulated gene expression. MPXV PoxS conferred resistance to the suppression of MPXV replication by type I IFN. Taken together, our findings suggested that MPXV PoxS counteracts host antiviral response by sequestering STAT2 to circumvent basal and type I IFN-induced expression of antiviral genes.

Keywords: 2′3′-cGAMP hydrolase; STAT2; STING; cGAS; innate antiviral response; monkeypox virus; mpox virus; poxin.

Figure 1.

Diminution of cGAS-produced 2′3′-cGAMP when MPXV PoxS was expressed. (A) A diagram showing the details of MPXV PoxS gene (OPG188) of MPXV MA001. (B) Multiple sequence alignment of MPXV PoxS of clade I (Zaire-96-I-16, NC_003310), clade IIA (Nigeria-156, KJ642615.1) and clade IIB (MA001, ON563414 and hMpxV/Hong Kong/HKU-220914-001/2022, GISAID accession no.: EPI_ISL_14945299). Arrows indicated the HYK catalytic triad. Asterisks indicated 2′3′-cGAMP binding sites. (C) HEK293T cells were transfected with pcDNA™3.1/V5-His A empty vector or pcDNA™3.1/V5-His A-MPXV-PoxS. Cellular protein was extracted at 48 hours post-transfection. MPXV-PoxS-V5-6xHis fusion protein was detected by Western blotting using anti-V5 antibody. GAPDH served as loading control. (D) Cellular 2′3′-cGAMP of HEK293T cells transfected without (vector control) or with cGAS and/or MPXV PoxS expression constructs was extracted with M-PER lysis buffer. 2′3′-cGAMP was detected by ELISA. Experiments were performed in triplicates. (E) HEK293T cells were transfected with pcDNA™3.1/V5-His A empty vector or increasing dose of pcDNA™3.1/V5-His A-MPXV-PoxS together with cGAS expression construct. cGAS, MPXV PoxS and GAPDH proteins were detected by Western blotting. An unpaired student t-test was performed to compare sample groups for statistical significance. P value less than 0.05, 0.01 or 0.001 was defined as statistically significant (*), very significant (**) or extremely significant (***). Otherwise, the difference between sample groups was statistically not significant (ns).

← Figure 2.

Impact of MPXV PoxS on cGAS-STING-mediated type I IFN production. (A) Impact on cGAS-STING-induced IFNβ promoter activity. IFNβ promoter activity was stimulated by either cGAS + STING (left) or STING + cGAMP (right) in HEK293T cells transfected with vector control (Vec) or with MPXV PoxS expression construct. IFNβ promoter-driven firefly luciferase plasmid and the SV40 early constitutively active promoter-driven Renilla luciferase plasmid served as reporter. Cells were collected at 48 hours for dual luciferase assay. At 24 hours after transfection, cells were further stimulated with lipofectamine3000 reagent alone or with 1 µg/mL 2′3′-cGAMP (lipofectamine3000-2′3′-cGAMP complex). At 48 hours, cells were harvested for luciferase assay through measurement of LAR-II and S&G substrates using luminometer. (B) HEK293T cells were transfected with vector control (Vec) or with cGAS + STING plasmids together with vector control (Vec) or MPXV PoxS expression construct. At 48 hours, cellular RNA was extracted for RT-qPCR analysis of the mRNA expression levels of IFNβ, IFNα4, IFNλ and OAS-1 genes. (C) NuLi-1 cells were transfected with vector control (Vec) or with MPXV PoxS construct. At 24 hours, cells were infected with MVA-BN (VR-1508). At another 24 hours, cellular RNA was extracted for RT-qPCR analysis of the mRNA expression levels of IFNβ, OAS1 and SAMD9 gene. Human β-tubulin mRNA was used as internal control for RT-qPCR analysis. Moreover, cellular DNA was collected for qPCR quantification of MVA viral DNA at I4L gene. Experiments were performed in biological triplicates. Statistical analysis was performed as in Figure 1.

← Figure 3.

Suppression of type I IFN signalling by MPXV PoxS and its interaction with STAT2. (A) Dual luciferase reporter assay was performed in HEK293T cells to assess the ISRE- or GAS-driven promoter activity of different treatment groups as indicated. HEK293T cells were transfected with vector control (Vec) or with MPXV PoxS expression construct together with the reporter constructs. Renilla luciferase plasmid under the control of constitutively active SV40 early promoter was used as internal control. At 24 hours, cells were either mock treated or treated with 1000 U/mL IFNβ or 100 ng/mL IFNγ recombinant protein. At 48 hours, cells were harvested for detection of the luciferase activity through LAR-II and S&G substrates on luminometer. (B) HEK293T cells were transfected and treated with or without IFNβ as in (A) but without reporter constructs. At 48 hours, cellular RNA was extracted for RT-qPCR analysis of the mRNA expression levels of OAS1, SAMD9 and SAMD9L genes. Human β-tubulin mRNA served as the internal control. U.D.: undetected. (C) HEK293T cells were transfected and treated with IFNβ as in (A) but without reporter constructs. After 24 hours of treatment with IFNβ, cellular protein was extracted in RIPA lysis buffer and assayed through SDS-PAGE and Western blotting with anti-pSTAT2 (Y690), anti-STAT2, anti-pSTAT1 (S727), anti-STAT1, anti-V5 (PoxS) and GAPDH. Experiments were performed in biological triplicates. (D) Densitometry of p-STAT1, p-STAT2, STAT1 and STAT2 bands in (C) were quantified with ImageJ software. Relative band intensity of p-STAT1/STAT1 and p-STAT2/STAT2 was calculated. For the STAT1 doublet, the upper band should be doubly phosphorylated form of STAT1 at p-S727 and p-Y701 [54]. The lower band should be STAT1 with single p-S727. The combination of both represented the total level of p-S727 STAT1. The quantification in the bar chart was based on total p-S727 STAT1. Statistical analysis was performed as in Figure 1. (E) Co-immunoprecipitation. HEK293T cells were transfected with PoxS-V5 and STAT2-Flag or STAT-1-Myc expression plasmids. Cell lysates were collected by 1% NP40 buffer and immunoprecipitation was performed with the indicated antibodies. Mouse IgG (Santa Cruz sc-2025) was used as negative control.

Figure 4.

Colocalization of MPXV PoxS with STAT2. A549 cells were transfected with expression construct for eGFP-tagged MPXV PoxS. At 24 hours, cells were fixed with PBS-buffered 4% paraformaldehyde and permeabilized with PBS-buffered 0.2% Triton X100. (A) Cells were stained with anti-STAT1, anti-STAT2 or anti-JAK1 followed by secondary antibody with rhodamine conjugate and DAPI staining. Cells were visualized on confocal microscope LSM980 with Airyscan function. (B) Cells expressing MPXV PoxS and stained with STAT2 were subjected to Z-stack confocal imaging through LSM980. Z-stack images were 0.129 µm apart with 50% layer-overlap. A total of 210–211 slices were captured. 3D images were generated by Zen 3.3 software. Three bird views of the 3D images were shown.

Figure 5.

Requirement of vSlfn fusion and the active site of 2′3′-cGAMP nuclease of MPXV PoxS for STAT2 sequestration. A549 cells were transfected with expression constructs for eGFP-tagged MPXV-PoxS (A), eGFP-tagged PoxS-4A (B) and eGFP-tagged PoxS-Poxin (C). Cells were fixed and stained as in Figure 4. Three representative views were shown. Arrows indicated nuclear localization of STAT2 in eGFP-positive cells. (D) eGFP signal of A549 cells expressing eGFP-tagged MPXV-PoxS (n = 10), eGFP-tagged PoxS-4A (n = 10) or eGFP-tagged PoxS-Poxin (n = 9) were analysed for colocalization with STAT2. The degree of colocalization was expressed as the value of Pearson’s colocalization correlation (PCC) of pixels of eGFP and STAT2. (E) Similar to (D), the colocalization of DAPI and STAT2 was quantified and expressed as PCC. One-way ANOVA was used to evaluate statistical significance between independent sample groups. Adjusted P value less than 0.05 was defined as being statistically significant.

← Figure 6.

Requirement of vSlfn fusion and the active site of 2′3′-cGAMP nuclease of MPXV PoxS for suppression of ISG expression in transfected cells and type I IFN antiviral response in MPXV-infected cells. (A) NuLi-1 cells were transfected with the expression construct for eGFP (vec), MPXV PoxS-eGFP (PoxS-WT), PoxS-4A-eGFP or PoxS-Poxin-eGFP. At 24 hours, cells were either mock treated or treated with 1000 U/mL IFNβ. At 48 hours, cellular RNA was extracted for RT-qPCR analysis of the mRNA expression levels of OAS1, SAMD9 and SAMD9L genes. Human β-tubulin mRNA was used as the internal control. (B, C) Vero-E6 cells were transfected with the expression construct for eGFP (Vec), MPXV PoxS-eGFP (PoxS-WT), PoxS-4A-eGFP or PoxS-Poxin-eGFP. At 24 hours, cells were infected with hMpxV/Hong Kong/HKU-220914-001/2022 at MOI = 0.00035. At 48 hours, cells were either mock treated or treated with 1000 U/mL IFNβ. At 96 hours, infectious MPXV was obtained by three freeze-thaw cycles. (B) Viral DNA was extracted from the freeze-thaw lysate and quantitated with qPCR targeting viral TK, A10L, A27L and I4L genes. (C) Infectious viral quantities of the freeze-thaw lysates were detected and estimated by standard plaque assay. Experiments were performed in biological triplicates. Statistical analysis was performed with one-way Anova. For (B), One-way ANOVA was performed to compare all samples with Vec + IFNβ group. Adjusted P value less than 0.05 was defined as being statistically significant. Otherwise, the difference between sample groups and Vec + IFNβ group was statistically not significant (ns).

← Figure 7.

Schematic diagram of the proposed model for MPXV PoxS-mediated IFN antagonism. (A) The cGAS-STING pathway and type-I IFN signalling. Briefly, cGAS senses and binds dsDNA in the cytoplasm, which facilitates the production of 2′3′-cGAMP. In turn, 2′3′-cGAMP activates STING which triggers phosphorylation and dimerization of IRF3 through TBK1. The activated IRF3 homodimer translocates into the nucleus and stimulates type I IFN production. Secreted type I IFNs then bind to the type I IFN receptor complex on the cell membrane. This activates the JAK-STAT pathway. STAT1 is phosphorylated, self-dimerizes, binds and transactivates GAS promoter for ISG expression. STAT2 is phosphorylated and forms heterodimers with STAT1. The activated STAT1/2 complex further recruits IRF9 to form ISGF3, that finally binds and transactivates ISRE promoter for ISG expression. (B) MPXV PoxS-mediated perturbation of 2′3′-cGAMP and type I IFN signalling. Upon MPXV infection, viral DNA activates cGAS. With 2′3′-cGAMP hydrolyase activity, PoxS counteracts 2′3′-cGAMP synthesized by cGAS. Whether PoxS-mediated 2′3′-cGAMP suppression would affect type I IFN production through STING might be context-specific. On the other hand, PoxS downregulates interferon-stimulated gene expression by sequestering and inhibiting STAT2. Y690 phosphorylation of STAT2 is downregulated by PoxS. Cytosolic clustering of PoxS with STAT2 is necessary for this process which requires vSlfn domain and the intact 2′3′-cGAMP hydrolase active site. PoxS without vSlfn domain is unable to form cytosolic clusters. Catalytically dead mutant, PoxS-4A, that had alanine substitutions in its 2′3′-cGAMP hydrolase active site still forms cytosolic clusters but cannot interact with STAT2. In contrast, Poxin binds well with STAT2. Both Poxin and PoxS-4A fail to suppress type I IFN signalling compared with the WT PoxS. This suggests that intact PoxS of MPXV is required for the suppression of STAT2 and IFN signalling. Together, these findings suggest: (i) PoxS downregulates 2’3’-cGAMP level and may affect cGAS-STING-mediated antiviral response depending on the context (e.g. the resulting 2′3′-cGAMP concentration and the strength of STING protein response). (ii) PoxS suppresses STAT2-mediated IFN signalling. (iii) The potential role of vSfln for mediating cytosolic clustering of the PoxS protein. and (iv) The integrity of the active site of 2′3′-cGAMP hydrolase is necessary for PoxS to interact with STAT2. The ancestral conformational structure of the Poxin domain might be required for STAT2 interaction.