Research ArticleInnate Immunity

PKR-dependent cytosolic cGAS foci are necessary for intracellular DNA sensing

See allHide authors and affiliations

Science Signaling  26 Nov 2019:
Vol. 12, Issue 609, eaav7934
DOI: 10.1126/scisignal.aav7934

Stronger together

Aberrant intracellular recognition of self-DNA is controlled by the spatial restriction of the DNA sensor cGAS. To define the interactions that may control cGAS subcellular localization, Hu et al. used mass spectrometry and immunoprecipitation to confirm that endogenous cGAS bound to the nucleotide helicase G3BP1, which is involved in stress granule formation. After DNA stimulation of human cells, cGAS associated in an RNA-dependent manner with G3BP1 and was found in cytoplasmic foci that also contained mRNA and the RNA-dependent kinase PKR. Formation of cytoplasmic cGAS condensates necessary for DNA-stimulated type I interferon production required G3BP1 and PKR activity. Together, these data suggest that G3BP1 cytoplasmic foci act as hubs that may coordinate DNA and RNA nucleic acid sensing.

Abstract

Cyclic GMP-AMP synthase (cGAS) is a major sensor of cytosolic DNA from invading pathogens and damaged cellular organelles. Activation of cGAS promotes liquid-like phase separation and formation of membraneless cytoplasmic structures. Here, we found that cGAS bound G3BP1, a double-stranded nucleic acid helicase involved in the formation of stress granules. Loss of G3BP1 blocked subcellular cGAS condensation and suppressed the interferon response to intracellular DNA and DNA virus particles in cells. Furthermore, an RNA-dependent association with PKR promoted G3BP1 foci formation and cGAS-dependent interferon responses. Together, these results indicate that PKR promotes the formation of G3BP1-dependent, membraneless cytoplasmic structures necessary for the DNA-sensing function of cGAS in human cells. These data suggest that there is a previously unappreciated link between nucleic acid sensing pathways, which requires the formation of specialized subcellular structures.

INTRODUCTION

Innate immune response begins with recognition of pathogen-associated molecular patterns (PAMPs) (1). This important task is the function of pattern recognition receptors (PRRs) (14), including Toll-like receptors (5), Nod-like receptors (6), retinoic acid–inducible gene I (RIG-I)–like RNA receptors (7), and a group of cytosolic DNA receptors (8). Among the PRRs that sense cytosolic DNA, the cyclic guanosine 5′-monophosphate–adenosine 5′-monophosphate (cGAMP) synthase (cGAS) plays a key role in detecting cytosolic DNA and RNA:DNA hybrids (911). Upon binding to DNA, cGAS catalyzes the production of cGAMP from adenosine 5′-triphosphate (ATP) and guanosine 5′-triphosphate (GTP) (9, 10). The cGAMP binds to the adaptor, stimulator of interferon genes (STING), which induces phosphorylation of tumor necrosis factor receptor–associated factor (TRAF)–associated nuclear factor κB (NF-κB) activator (TANK)–binding kinase 1 (TBK1) and interferon regulatory factor 3 (IRF3) (10, 1214). The phosphorylated IRF3 dimerizes and translocates into the nucleus to induce the production of type I interferon (13, 15).

G3BP1 [Ras–guanosine triphosphatase activating protein (RasGAP) Src homology 3 domain–binding protein 1] binds to RasGAP in a Ras-GTP–dependent complex (16) and is required for stress granule (SG) assembly (17). An RNA/DNA and RNA/RNA helicase, G3BP1 has ATP- and magnesium-dependent activity that is not stimulated by replication protein A (18). G3BP1 also regulates several pathways including Ras signaling (16, 19), c-Myc mRNA turnover (20, 21), NF-κB signaling (22) and human epidermal growth factor receptor 2 signaling (23), tumor progression, and metastasis (2327). In addition, G3BP1 inhibits p53 expression by interacting with the deubiquitinase ubiquitin-specific peptidase 10, thus playing an important role in promoting oncogenesis (28).

G3BP1 is also involved in organization of SG, a subcellular structure tied to RNA localization, translation, and degradation (29). SG-associated proteins, including G3BP1, bind innate immune sensors that detect pathogen RNA (30). The RIG-I–like receptors such as RIG-I, melanoma differentiation-associated protein 5 (MDA5), laboratory of genetics and physiology 2, protein kinase R (PKR), ribonuclease L (RNase L), and 2′-5′-oligoadenylate synthetase are recruited to SGs in response to RNA stimulation, suggesting that these cytoplasmic foci serve as a platform for RNA sensing (31). In this study, we confirmed that cGAS bound the SG RNA helicase G3BP1 (32). In human cells, exogenous DNA stimulation promoted cGAS localization to G3BP1-containing cytoplasmic deposits. The exogenous DNA recognition also activated PKR. Depletion of either G3BP1 or PKR impaired cGAS sensing of vaccinia virus (VACV) DNA. Our data suggest that cooperation of these innate immune sensors requires formation of G3BP1 cytoplasmic foci that act as hubs of nucleic acid sensing.

RESULTS

cGAS binds and colocalizes with the RNA binding proteins G3BP1 and G3BP2

We identified cGAS-associated cellular proteins by coimmunoprecipitation (co-IP) followed by mass spectrometry and proteomic analysis. Using the PANTHER classification system (33), we found that many cGAS-associated candidate proteins were RNA binding proteins (Fig. 1A), including G3BP2, poly(A) binding protein cytoplasmic 1 (PABPC1); and eukaryotic initiation factor-4A (eIF4A) (data file S1). We confirmed the association of G3BP2, also G3BP1, with cGAS by Western blot analysis of the materials that were coimmunoprecipitated with the ectopically expressed cGAS in 293T cells (Fig. 1, B and C). Interaction of cGAS-Flag with Myc-G3BP1 and interaction of cGAS-Flag with Myc-G3BP2 were confirmed by reciprocal pulldowns using anti-Myc agarose, respectively (Fig. 1, D and E). Results of our co-IP demonstrated the association of G3BP1 with cGAS in an RNase-sensitive manner (Fig. 1D). We further observed the association of endogenous cGAS with endogenous G3BP1 in HeLa cells in co-IP experiments, and this association was also sensitive to RNase A treatment (Fig. 1F). In agreement with its association with G3BP1, cGAS was recruited into SGs that formed in response to arsenite treatment in both HeLa cells and differentiated THP-1 cells (Fig. 2A) but not by heat shock (fig. S1).

Fig. 1 cGAS binds G3BP1 and G3BP2.

(A) Gene Ontology (GO) analysis of cGAS binding proteins identified by mass spectrometry in lysates of 293T cells transfected with plasmids encoding cGAS-Flag and immunoprecipitated for Flag tag. ATPase, adenosine triphosphatase; GTPase, guanosine triphosphatase. (B) Co-IP analysis of G3BP2 interaction with cGAS in lysate of 293T cells transfected with cGAS-Flag DNA. (C) Co-IP analysis of G3BP1 interaction with cGAS in lysates of 293T cells transfected with cGAS-Flag DNA. (D) Co-IP analysis of G3BP1 interaction with cGAS in lysates of 293T cells transfected with cGAS-Flag and Myc-G3BP1 DNA. (E) Co-IP analysis of G3BP2 interaction with cGAS in lysates of 293T cells transfected with cGAS-Flag and Myc-G3BP2 DNA. (F) Co-IP of G3BP1 interaction with endogenous cGAS in lysates from HeLa cells. All blots (B to F, top) are representative of three independent experiments. All quantified band intensity values (B to F, bottom) are means ± SEM pooled from all experiments. *P < 0.05 and **P < 0.01, by Kruskal-Wallis test.

Fig. 2 cGAS colocalizes with and binds G3BP1.

(A) Immunofluorescence microscopy analysis of cGAS and G3BP1 localization in HeLa (top) and differentiated THP-1 (bottom) cells that were treated with arsenite (As) for 0.5 hours. (B) Immunofluorescence microscopy analysis of cGAS and G3BP1 localization in HeLa (top) and differentiated THP-1 (bottom) cells that were transfected with HT-DNA for 4 hours. (C) Immunofluorescence microscopy analysis of cGAS and G3BP1 localization in HeLa (top) and differentiated THP-1 (bottom) cells that were transfected with 488-ISD for 4 hours. (D) Fluorescence in situ hybridization analysis of poly(A) mRNA localization in HeLa cells transfected with ISD for 4 hours and stained for cGAS and G3BP1. (E) Co-IP analysis of G3BP1 interaction with endogenous cGAS in lysates of HeLa cells. All images with colocalization indicated by arrows are representative of five (A to C) or three (D) independent experiments. Blots (E) are representative of five independent experiments. Quantified data are means ± SEM pooled from all experiments. *P < 0.05 and **P < 0.01, by chi-square test (A to D) or Kruskal-Wallis test (E). DAPI, 4′,6-diamidino-2-phenylindole.

We further observed that activation of endogenous cGAS with the herring testis DNA (HT-DNA) or interferon stimulatory DNA (ISD) led to formation of G3BP1 punctate that also contained cGAS in HeLa, THP-1, and primary human foreskin fibroblast (HFF) cells (Fig. 2, B and C, and fig. S2A). We also conducted fluorescence in situ hybridization for polyadenylate [poly(A)] mRNA in ISD-transfected cells and found that poly(A) mRNA colocalized with G3BP1 and cGAS (Fig. 2D and fig. S2B). These data indicate that ISD detection promotes the assembly of mRNA containing cytoplasmic G3BP1 foci in human cells. In addition, we found that DNA treatment increased the association of the cGAS and G3BP1 (Fig. 2E). Together, we conclude that DNA stimulation promotes cGAS localization to G3BP1 cytoplasmic deposits.

cGAS-G3BP1 cytoplasmic foci contain PKR

In our experiments, we noted that transient transfection of cGAS DNA and, to a lesser extent, the enhanced green fluorescent protein (EGFP) DNA unexpectedly increased the amount of phosphorylated eIF2α and phosphorylated PKR in HeLa cells (Fig. 3A). When we coimmunoprecipitated Flag-tagged cGAS, we found that it associated with endogenous PKR in 293T cells (Fig. 3B). We stained HeLa cells transfected with cGAS-EGFP for endogenous PKR and G3BP1 and found that all three proteins colocalized (Fig. 3C). Knockdown of PKR prevented eIF2α phosphorylation in response to the transiently expressed cGAS (Fig. 3D). Together, these data suggested that intracellular DNA recognition may activate PKR (34).

Fig. 3 G3BP1-cGAS foci contain PKR.

(A) Western blot analysis for p-eIF2α and p-PKR in lysates of HeLa cells transfected with constructs for cGAS-EGFP. Blots are representative of three independent experiments. (B) Co-IP analysis of PKR interaction with cGAS in lysates of 293T cells transfected with cGAS-Flag DNA. (C) Immunofluorescence microscopy analysis of G3BP1 and PKR localization in HeLa cells transfected with cGAS-EGFP DNA. (D) Western blot for p-eIF2α in lysates of HeLa cells after transfection with siRNA against endogenous PKR and cGAS-EGFP DNA. All blots and images with arrows indicating colocalization are representative of three independent experiments. Quantified data are means ± SEM from all experiments. *P < 0.05 and **P < 0.01, by Kruskal-Wallis test (A, B, and D) or chi-square test (C).

We further found that depletion of cGAS in HeLa cells prevented both G3BP1 puncta formation and phosphorylation of eIF2α and PKR (Fig. 4, A and B). Furthermore, ISD transfection slightly reduced HeLa cell mRNA translation (Fig. 4C), as determined by puromycylation assay to determine de novo protein synthesis. This effect of ISD transfection was dependent on cGAS because increase in eIF2α and PKR phosphorylation was lost in cGAS knockout (KO) cells (Fig. 4B). In addition, we found that depletion of PKR with small interfering RNA (siRNA) prevented the stimulation of eIF2α phosphorylation by ISD transfection of HeLa cells (Fig. 4D). Together, these data suggest that ISD binding to cGAS can stimulate phosphorylation of PKR and eIF2α.

Fig. 4 cGAS recognition of DNA promotes G3BP1 association and PKR activation.

(A) Immunofluorescence microscopy analysis of cGAS and G3BP1 colocalization in cGAS knockout (KO) HeLa cells transfected with 488-ISD for 4 hours. Images with arrows indicating colocalization (left) are representative of three independent experiments. Quantification of G3BP1+ cGAS+ foci (right) are means ± SEM pooled from all experiments. (B) Western blot analysis for p-eIF2α and p-PKR and global protein translation in lysates of HeLa cells transfected with ISD for 4 hours. (C) Western blot analysis for p-eIF2α and p-PKR in cGAS-KO HeLa cells that were transfected with ISD for 4 hours. (D) Western blot analysis for p-eIF2α and PKR in lysates of HeLa cells transfected with siRNA against PKR and ISD. All blots (B to D) are representative of three independent experiments. Quantified band intensity values are means ± SEM pooled from all experiments. *P < 0.05 and **P < 0.01, by chi-square test (A) or Kruskal-Wallis test (B to D).

DNA recognition by cGAS is required for G3BP1 condensation and PKR activation

Transfection of cells with ISD stimulated formation of G3BP1 cytoplasmic punctate that contained cGAS, and loss of cGAS prevented this response (Fig. 4A). To determine whether G3BP1 condensation required DNA binding by cGAS, we overexpressed the cGAS C396A/C397A mutant, which does not interact with double-stranded DNA (dsDNA) in HeLa cells depleted of endogenous cGAS (35). In these cells, ISD transfection did not stimulate G3BP1 puncta formation (fig. S3). However, loss of cGAS did not affect arsenite-stimulated SG formation (fig. S4), indicating that cGAS is not an integral component of canonical SGs. Together, these data suggest that ISD binding to cGAS promotes intracellular G3BP1 assembly.

The N-terminal domain of cGAS binds dsDNA and promotes its activation (36, 37). The cGAS N-terminal region enhances cGAS-DNA phase separation by increasing the valencies of DNA binding (38). We therefore mapped the interaction of cGAS with G3BP1 by expressing cGAS wild type (WT), cGAS(1-154) (only the N-terminal 154 residues), and cGAS(155-522) (cGAS without the 154 N-terminal residues) in 293T cells and tested their interaction with G3BP1 by co-IP. We found that cGAS WT and cGAS(1-154), but not cGAS(155-522), interacted with G3BP1 (fig. S5A). In addition, overexpression of both cGAS WT and cGAS(1-154) stimulated cytoplasmic assembly of G3BP1 (fig. S5B). Collectively, these results indicate that the N-terminal region of cGAS is critical for G3BP1 binding and formation of G3BP1 puncta.

We also investigated whether cGAMP synthesis by cGAS and the downstream STING signaling pathway are required for DNA activation of PKR. The cGAS mutants E225A/D227A and G212A/S213A, inactive in cGAMP synthesis (35), still enhanced phosphorylation of PKR and eIF2α and stimulated the formation of G3BP1-containing foci as efficiently as the WT cGAS (fig. S6, A and B). In support of these data, exogenous cGAMP alone did not cause SGs to form (fig. S6, C and D). In addition, transient expression of cGAS or treatment with ISD promoted G3BP1 assembly in Vero cells in which interferon signaling pathway is defective (fig. S6, E and F). Together, these data suggest that DNA binding to cGAS alone, without the need of synthesizing cGAMP or triggering the downstream interferon signaling cascade, is sufficient to activate PKR.

Intracellular DNA binding by cGAS requires G3BP1 and PKR

Does cGAS association with G3BP1 and PKR alter DNA sensing? When we used biotin-labeled ISD to precipitate cGAS, we observed that a significant amount of cGAS and G3BP1 were precipitated together with ISD in WT cells (Fig. 5, A to C). However, ISD binding to cGAS was lost in cells that lacked either PKR or G3BP1 (Fig. 5, A and B). In the absence of either G3BP1 or PKR, cGAS puncta formation stimulated by ISD greatly diminished compared to WT cells (Fig. 5, D and E). These data suggest that cGAS association with PKR and G3BP1 promotes cGAS DNA recognition. In addition, we found that loss of PKR and G3BP1 in THP-1 cells greatly diminished both ISD- and HT-DNA–stimulated induction of p-TBK1 and p-IRF3 (Fig. 6, A and B, and fig. S7, A and B). Similarly, neither interferon-beta symbol (IFN-β) nor ISGs were produced in cells lacking PKR, G3BP1, or cGAS, either by CRISPR editing (Fig. 6, C to F) or by siRNA depletion (figs. S7, C to F, and S8). When we indirectly assayed cGAMP activity, we found that loss of PKR or G3BP1 in THP-1 cells reduced IFNb transcription stimulated by lysates of ISD transfected cells (Fig. 6, G and H, and fig. S7, G and H). We also found that disruption of another critical SG-associated protein, T cell–restricted intracellular antigen-1 (TIA1) (39), did not affect DNA-induced p-IRF3 (fig. S9). In contrast, when we exposed cells to exogenous cGAMP, we detected p-IRF3 and IFN-β production in both control cells and cells depleted of PKR or G3BP1 but not in cells depleted of cGAMP effector STING (fig. S10). These data indicate that exogenous cGAMP can restore IFN-β production in PKR- or G3BP1-depleted cells, which are incapable of producing interferon in response to ISD. Collectively, these results demonstrate that functional sensing of DNA by cGAS depends on expression of G3BP1 and PKR in THP-1 cells.

Fig. 5 Intracellular DNA binding by cGAS requires G3BP1 and PKR.

(A) Co-IP analysis of ISD interaction with cGAS-EGFP in lysates of PKR or G3BP1 knockdown HeLa cells stably expressing cGAS-EGFP. (B) Co-IP analysis of ISD interaction with cGAS in lysates of PKR or G3BP1 KO HeLa cells transfected with cGAS and biotin-ISD. (C) Co-IP analysis of ISD interaction with G3BP1 in lysates of 293T cells transfected with biotin-ISD. (D and E) Immunofluorescence microscopy analysis of ISD, cGAS, and G3BP1 localization in G3BP1-KO (D) or PKR-KO (E) HeLa cells transfected with 488-ISD for 4 hours. All blots and images with arrows indicating colocalization are representative of three independent experiments. Quantified data are means ± SEM pooled from all experiments. **P < 0.01, by Kruskal-Wallis test (A to C) or chi-square test (D and E).

Fig. 6 Loss of PKR or G3BP1 prevents DNA-stimulated interferon production.

(A and B) Western blot analysis of p-IRF3 and p-TBK1 in lysates of cGAS-KO, PKR-KO, or G3BP1-KO THP-1 cells 4 hours after transfection with ISD (A) or HT-DNA (B). Blots are representative of three independent experiments. Quantified band intensity values are means ± SEM pooled from all experiments. (C and D) qRT-PCR analysis of Ifnβ, Isg54, and Isg56 mRNA expression in cGAS-KO, PKR-KO, or G3BP1-KO THP-1 cells 4 hours after transfection with ISD (C) or HT-DNA (D). (E and F) ELISA analysis of IFN-β production by cGAS-KO, PKR-KO, or G3BP1-KO THP-1 cells 4 hours after transfection with ISD (E) or HT-DNA (F). (G and H) qRT-PCR analysis of Ifnβ mRNA expression in digitonin-permeabilized THP-1 cells that were incubated with the cytosolic extracts from cGAS-KO, PKR-KO, or G3BP1-KO THP-1 cells transfected with ISD (G) or HT-DNA (H) for 4 hours. All qRT-PCR and ELISA data are means ± SEM from three independent experiments. *P < 0.05 and **P < 0.01, by Kruskal-Wallis test.

We also observed that the exposure of cells to nonreplicating heat-inactivated VACV (Tiantan strain) particles, which contained dsDNA viral genome, led to formation of G3BP1 puncta containing cGAS in THP-1 and HFF cells. We detected VACV genomic DNA in foci containing cGAS and G3BP1, in both THP-1 and HFF cells after exposure to VACV (Fig. 7, A and B). Similarly, when we exposed these cells to heat-inactivated VACV particles, we also detected strong activation of p-TBK1 and p-IRF3 (Fig. 7C and fig. S11A) and expression of ISG mRNAs and interferon (Fig. 7, D and E, and fig. S11, B and C). This anti-VACV response required cGAS because loss of cGAS inhibited interferon production in response to VACV stimulation (Fig. 7, C to E, and fig. S11, A to C). Similarly, in the absence of either PKR or G3BP1, VACV stimulation could not activate TBK1 and IRF3 and did not trigger interferon production (Fig. 7, C to E, and fig. S11, A to C). These data demonstrate that PKR and G3BP1 play critical roles in cGAS-initiated interferon response to VACV DNA. The impaired sensing of VACV DNA by cGAS in PKR- or G3BP1-KO cells also correlated with reduced IFNb mRNA induction (Fig. 7F and fig. S11D). Together, our data suggest that the cooperative action of cGAS, G3BP1, and PKR is required for viral DNA detection and the antiviral interferon response.

Fig. 7 DNA virus particles promote cGAS-G3BP1 association required for antiviral responses.

(A and B) Immunofluorescence microscopy analysis of cGAS and G3BP1 localization in differentiated THP-1 cells (A) and HFF cells (B) exposed to heat-inactivated VACV for 4 hours. Images with arrows indicating colocalization (left) and fluorescence intensity analysis (right) are representative of three independent experiments. (C) Western blot analysis of pIRF3 and pTBK1 in lysates of cGAS-KO, PKR-KO, or G3BP1-KO THP-1 cells after exposure to heat-inactivated VACV for 4 hours. Blots (left) are representative of three independent experiments. Quantified band intensity values are means ± SEM pooled from all experiments. (D) qRT-PCR analysis of Ifnβ, Isg54, and Isg56 mRNA expression in cGAS-KO, PKR-KO, or G3BP1-KO THP-1 cells 0.5 hours after exposure to heat-inactivated VACV. (E) ELISA analysis of IFN-β production by cGAS-KO, PKR-KO, or G3BP1-KO THP-1 cells 0.5 hours after exposure to heat-inactivated VACV. (F) qRT-PCR analysis of Ifnβ mRNA expression in digitonin-permeabilized THP-1 cells that were incubated with the cytosolic extracts from cGAS-KO, PKR-KO, or G3BP1-KO THP-1 cells 0.5 hours after exposure to heat-inactivated VACV. All qRT-PCR and ELISA data (D to F) are means ± SEM pooled from three independent experiments. *P < 0.05 and **P < 0.01, by Kruskal-Wallis test.

DISCUSSION

SGs are nonmembranous cytoplasmic foci in mammalian cells, which aggregate in response to adverse environmental conditions, such as heat shock, oxidative stress, hyperosmolarity, viral infection, and ultraviolet irradiation (40). These stress signals selectively activate the protein kinases PKR, general control nonderepressible 2 (GCN2);heme-regulated inhibitor (HRI); PKR-like ER kinase (PERK) that phosphorylates translation initiation factor eIF2α to induce translational arrest that promotes SG formation (30, 40). Collectively, our results demonstrated that cGAS associated with the SG initiating protein G3BP1. However, because of the incomplete translational arrest after DNA transfection, it remains unclear whether these subcellular structures that contain cGAS, G3BP1, and mRNA are canonical SGs (32). These DNA-induced G3BP1 puncta may differ from canonical SGs in their specific components. Stress-specific differences between SG constituents have already been reported (41). Similarly, we observed that sodium arsenite treatment promoted formation of SGs that contained cGAS, whereas cGAS was not recruited to SGs induced by heat shock. Sodium arsenite causes DNA damage and formation of micronuclei in human cells (42), which can be detected by cGAS and leads to the presence of cGAS in these SGs. It is envisioned that other DNA damage–promoting stimuli may allow recruitment of cGAS into SGs. An additional difference between the DNA-induced cGAS assembly and the canonical SGs is the sizes of the foci; the activation of PKR and eIF2α-mediated translational arrest correlates with SG size (43). We found that arsenite-induced SGs were larger than ISD-induced foci (fig. S12). More work is warranted to determine how these ISD-stimulated foci differ from bona fide SGs.

In our study, we found that cGAS association with G3BP1 was necessary for cGAS DNA binding and subsequent interferon responses. Although cGAS can directly bind DNA and generate cGAMP in vitro without any cofactors (35, 44), association with G3BP1 appears to be essential for cGAS function in cells (23). We further showed that G3BP1 interacted with the N-terminal region of cGAS, which is critical for cGAS-mediated interferon response (32). cGAS depends on other cellular factors to be an effective DNA sensor. For example, cGAS sensing of reverse-transcribed HIV-1 DNA requires the polyglutamine binding protein 1 (45), an mRNA splicing factor that also promotes the assembly of SGs (46). Similarly, direct binding of the nucleic acid binding protein ZCCHC3, which is found within some SGs (47), to dsDNA enhances the binding of cGAS to dsDNA, subsequent cGAS activation, and induction of downstream effector genes (48). In addition to the possibility of being a cofactor of cGAS, G3BP1 assembles a structure to assist cGAS sensing cytosolic DNA. These cytoplasmic membraneless structures disrupt nucleocytoplasmic transport (49), increase local cGAS concentration, and enhance binding to DNA. Sequestration of cGAS in these cytoplasmic foci may also prevent its traffic into the nucleus where cGAS has much lower activity of cGAMP synthesis (50).

Our results agree with a recent report showing that G3BP1 depletion diminishes DNA binding of cGAS and inhibits interferon induction as a result of DNA transfection or DNA virus stimulation (32). One discrepancy is the effect of DNA transfection on the formation of G3BP1-containing foci in the cytoplasm and the association of cGAS with G3BP1. The potential causes of the different observations might include the type of DNA used in transfection and the cell types tested. Regardless, it is worth pursuing how G3BP1 stimulates the DNA sensor activity of cGAS, either as a direct cofactor and/or through assisting cGAS condensation.

Our data unexpectedly indicated a role for the double-stranded RNA (dsRNA)–dependent protein kinase PKR, a ubiquitously expressed serine/threonine kinase, in ISD-stimulated cGAS redistribution and activity. PKR is activated through autophosphorylation not only after dsRNA recognition but also after treatment with other stimuli, including inflammatory cytokines, Toll-like receptor ligands, and cellular stress (5153). PKR can also be activated by protein activator of interferon-induced protein kinase (54). Activated PKR stimulates mitogen-activated protein kinases (55, 56), inhibitor of κB kinase (57, 58), and interferon-β promoter stimulator 1 (IPS-1) signaling (59). The activities of the stress-activated protein kinases p38 and c-Jun N-terminal kinase are also regulated by PKR, which leads to the production of proinflammatory cytokines (60). One other study suggests that PKR can be activated by dsDNA (34). Given the exclusive RNA binding property of the dsRNA binding domains in PKR (61), it is difficult to envision how DNA could directly activate PKR. Our data provides one answer by showing that both cGAS and PKR were recruited to cytoplasmic G3BP1 foci that contain nucleic acids. Our data indicated that the DNA binding activity of cGAS was required for its interaction with G3BP1, and G3BP1 was required for PKR activation. Similar to other types of cellular stress, we found that PKR was required for formation of G3BP1 puncta in response to cytoplasmic DNA (43). One possibility is that after interacting with DNA-bound cGAS, G3BP1 activates PKR, which further reinforces the assembly of cGAS-G3BP1.

The involvement of SGs (or SG-like assemblies) in innate immune sensing of PAMPs has also been exemplified by host detection of viral RNA by the cytosolic RIG-I–like receptors (31, 62, 63). It is advantageous for RNA sensors including RIG-I and MDA5 to reside in the stress-induced RNA granules where viral RNA is also sequestered and thus more readily detected (31, 63). It is not immediately clear why the DNA sensor cGAS is concentrated in similar cytoplasmic foci where DNA is often not enriched. One possibility is that cGAS recruitment to G3BP1 signaling platforms serves as a mechanism to locally concentrate cGAS and augments the synthesis of cGAMP, which subsequently binds to STING and triggers interferon production. Cells seem to have formed elaborate mechanisms to promote and regulate cGAS sensing DNA to effectively detect invading pathogens and, in the meantime, to avoid triggering harmful inflammatory responses. For example, cGAS is not effective in sensing DNA shorter than 20 base pairs. Nucleation of cGAS on long pathogenic DNA is thus assisted by a number of cellular factors including high-mobility group box 1 protein that upon binding to DNA, causes U-turns and bends in DNA, which are preferentially recognized and bound by cGAS (64). Results of our study suggest that G3BP1 may also provide such a mechanism to regulate the DNA sensing function of cGAS.

MATERIALS AND METHODS

Reagents, DNA constructs, antibodies, viruses, cell culture, and transfection

2′3′-cGAMP (tlrl-nacga23) was purchased from InvivoGen. All oligonucleotides were synthesized by Thermo Fisher Scientific. ISD, 488-ISD, and biotin-ISD were prepared from equal molar amounts of the sense and antisense DNA oligonucleotides (sense strand sequence, 5′-TACAGATCTACTAGTGATCTATGACTGATCTGTACATGATCTACA-3′). The oligonucleotides were heated to 95°C for 10 min and were then cooled to 16°C at a cooling rate of 2°C/min. Dynabeads Protein G (10003D) and Dynabeads M-280 Streptavidin (11205D) were purchased from Thermo Fisher Scientific. Penicillin-streptomycin solution (P1400) was purchased from Solarbio, and Benzonase (E1014) was purchased from Sigma-Aldrich.

The cGAS-Flag and Myc-G3BP1 cDNA sequences were cloned into the pcDNA4 expression vector (Invitrogen). The WT and mutated cGAS-EGFP DNA sequences were cloned into pEGFP-N1 (Clontech). The cGAS E225A/D227A, G212A/S213A, and C396A/C397A mutants were generated using the polymerase chain reaction (PCR)–based mutagenesis method. The cDNA of cGAS was inserted into the pRetroX-Tight-Pur retroviral vector (Clontech) to create DNA constructs Tet-cGAS. The EGFP sequence was attached to the N terminus of cGAS.

Antibodies against Flag [dilution, 1:4000; mouse F3165 (M2); rabbit F7425], EGFP (1:4000, G1544), and β-actin [1:5000, A1978 (AC-15)] were purchased from Sigma-Aldrich. Antibodies against PKR pTyr446 [1:2000, ab32036 (E120)] and IRF3 pSer386 [1:2000, ab76493 (EPR2346)] were purchased from Abcam. Antibodies against G3BP1 (1:2000, 13057-2-AP), STING (1:2000, 19851-1-AP), TIA1 (1:2000, 12133-2-AP), and IRF3 (1:2000, 11312-1-AP) were purchased from Proteintech Group. Antibodies against PKR [1:2000, rabbit 12297 (D7F7)], eIF2α (1:2000, 9722), eIF2α pSer51 [1:2000, 3597 (119A11)], cGAS [1:2000, 15102 (D1D3G)], TBK1 (1:1000, 3013), and TBK1/NAK pSer172 [1:2000, 5483 (D52C2)] were purchased from Cell Signaling Technology. Antibody against PKR [mouse, 610764 (13/PKR)] was from BD Biosciences, and normal rabbit immunoglobulin G (IgG) (sc-2027) was from Santa Cruz Biotechnology. Antibodies against G3BP1 [1:2000, mouse 05-1938 (14E5-G9)] were purchased from Millipore, and STING [1:2000, mouse TA505023 (OTI4H1)] was purchased from OriGene. Alexa Fluor–conjugated secondary antibodies were purchased from Thermo Fisher Scientific. Anti-Flag M2 affinity gel (A2220) and Anti-Myc Agarose Affinity Gel (A7470) were purchased from Sigma-Aldrich.

VACV (Tiantan strain) was propagated in Vero cells and used at a multiplicity of infection of 10 after heat inactivation (56°C for 30 min). Human embryonic kidney 293T cells, human cervix carcinoma (HeLa) cells, primary HFF, and African green monkey kidney (Vero) cells were maintained in Dulbecco’s Modified Eagle Medium (DMEM) (Thermo Fisher). The human monocytic cell line (THP-1) was maintained in RPMI 1640 (Thermo Fisher). All cells were supplemented with 10% fetal bovine serum (Thermo Fisher), penicillin (100 U/ml) and streptomycin (100 mg/ml), maintained at 37°C with 5% CO2. Cells were transfected with polyethylenimine (PEI) (Sigma-Aldrich) or Lipofectamine (Invitrogen) according to the manufacturer’s instructions. For ISD transfection, ISD (1 μg/ml) was transfected into cells using Lipofectamine 2000 (THP-1 cell) or PEI (other cells). To create doxycycline-inducible cGAS cell line, Tet-cGAS and reverse tetracycline tranional activator (rtTA) (Clontech) were cotransfected into cells. Stable cell lines were selected in DMEM supplemented with puromycin (1 μg/ml), G418 (0.75 mg/ml), and tetracycline-free serum (Clontech). To check the expression of the cGAS protein, cells were treated with doxycycline (500 ng/ml) for 16 hours before the cell lysates were collected and examined by Western blotting (65).

Co-IP and mass spectroscopy

293T cells were transfected with cGAS-Flag DNA. Twenty-four hours after transfection, cells were incubated for 20 min on ice in lysis buffer [50 mM tris-HCl (pH 7.4), with 150 mM NaCl, 1 mM EDTA, and 1% Triton X-100]. Supernatants were incubated with anti-Flag gel (A2220) on a rocking platform at 4°C overnight. cGAS-Flag was eluted under native condition using 3X FLAG peptide (F4799). Each sample was digested with filter-aided sample preparation method (66). Digested peptides from samples were desalted on C18 columns (1 cm3, 30 mg, Oasis). The desalted peptides were lyophilized by vacuum centrifugation and redissolved in 1‰ formic acid. The desalted peptides were analyzed with a reverse-phase C18 self-packed capillary liquid chromatography (LC) column (75 μm by 100 mm). The eluted gradient was 5 to 30% buffer B2 (0.1% formic acid and 99.9% acetonitrile (ACN); flow rate, 0.3 μl/min) for 40 min. A TripleTOF 5600 mass spectrometer was used to analyze the eluted peptides from LC. The mass spectrometry data were acquired using high-sensitivity mode with the following parameters: 30 data-dependent tandem mass spectrometry scans per full scan, full scans acquired at a resolution of 40,000 and tandem mass spectrometry scans at a resolution of 20,000, rolling collision energy, charge state screening (including precursors with +2 to +4 charge state), dynamic exclusion (exclusion duration 15 s), tandem mass spectrometry scan range of 100 to 1800 mass/charge ratio, and scan time of 100 ms. ProteinPilot (V1.0) was used for database searching of all samples. In ProteinPilot, the database was set to UniProt human database, and the digestion enzyme was set to trypsin. The parent mass tolerance was 0.05 Da, and fragment ion was 0.05 Da. Carbamidomethyl of cysteine was set as a fixed modification, and a maximum of two miscleavage sites were allowed. The false discovery rate was set as <1% on both peptide and protein levels. Protein identification was accepted with at least two unique peptides (67).

siRNA knockdown

The stealth siRNA duplexes targeting cGAS (HSS132956 and HSS132957), G3BP1 (HSS115444 and HSS115445), STING (HSS139156 and HSS139157), PKR (HSS108571 and HSS183403), and siRNA negative control (12935300) were purchased from Thermo Fisher. Cells were seeded into six-well plates 1 day before two sequential transfections with siRNA oligos (30 nM) using Lipofectamine RNAiMAX (Thermo Fisher). The knockdown efficiency of each gene was examined by Western blotting.

Immunofluorescence microscopy

Cells were grown on glass coverslips before transfection. After transfection, cells were fixed with 4% paraformaldehyde (PFA) [in 1× phosphate-buffered saline (PBS)] for 10 min at room temperature, followed by a 10-min permeabilization with 0.5% Triton X-100 at room temperature. Cells were then incubated overnight at 4°C with indicated antibodies. Alexa Fluor–conjugated antibodies (1:1000; Thermo Fisher) were used as secondary antibodies as described before (68). Poly(A) mRNAs were detected by in situ hybridization assay. Briefly, cells were fixed with 4% PFA, permeabilized with 0.3% Triton X-100, washed with PBS, and hybridized overnight at 43°C in 200 μl of a mixture containing 50% formamide, 2× SSC, 10% dextran sulfate, 0.2% bovine serum albumin. 20 mM vanadyl ribonucleoside complex, and 0.1 μM 488- or 555-conjugated oligo(dT) probe. Confocal images were acquired at room temperature using a Leica TCS SP5 (Leica Microsystems) mounted on an inverted microscope (DMI6000; Leica Microsystems) with an oil immersion 63×/numerical aperture 1.4 objective lens (HCX PL APO CS; Leica Microsystems). All subsequent analysis and processing of images were performed using the LAS AF software (Leica Microsystems).

Western blotting

Cells or viruses were lysed in radioimmunoprecipitation assay (RIPA) buffer [0.1% SDS, 1% Triton X-100, 1% sodium deoxycholate, 150 mM NaCl, 10 mM tris (pH 7.5), and 1 mM EDTA]. Equal amounts of cell or viral lysates were separated in SDS–polyacrylamide gel electrophoresis (12%) (WB1103, Beijing Biotides Biotechnology Co. Ltd.). Proteins were transferred onto nitrocellulose membranes (Whatman). The membranes were probed with indicated antibodies, followed by incubation with IRDye secondary antibodies (1:20,000; LI-COR Biotechnology). Protein bands were visualized on a LI-COR Odyssey instrument (LI-COR Biotechnology). Intensities of protein bands in Western blots were determined with ImageJ (National Institutes of Health).

Immunoprecipitation

Cells were lysed in RIPA buffer and clarified by centrifugation. One milligram of cell lysate was incubated with 50 μl of anti-Flag M2 affinity gel (A2220, Sigma-Aldrich) or anti-Myc agarose (A7470, Sigma-Aldrich) overnight at 4°C. For biotin pulldown assay, cells were transfected with biotin-ISD for 24 hours, then lysed, and incubated with Dynabeads M-280 Streptavidin magnetic beads. Beads were washed at 4°C three times with the RIPA buffer. To immunoprecipitate endogenous cGAS and G3BP1, cell lysates (500 μl) were incubated overnight at 4°C with the corresponding antibody or the control IgG (Santa Cruz Biotechnology) as a negative control. Immunocomplexes were incubated with protein G beads for 4 hours at 4°C. To investigate whether protein association is RNA dependent, cell lysates were first treated with RNase A (100 μg/ml) (QIAGEN) at 37°C for 30 min before being subjected to immunoprecipitation. After centrifugation, the unbound supernatant was collected as “flow through.” Beads were washed at 4°C three times with RIPA buffer. Immunocomplexes were boiled for 5 min with 50 μl of 2× Laemmli sample buffer; 10 μl of sample was analyzed by Western blotting.

Measurement of protein synthesis

Protein synthesis was measured by the incorporation of puromycin into peptide chains as described (69). Briefly, transfected cells were incubated with puromycin (10 μg/ml) for 10 min before cell lysis. Cell extracts were blotted with anti-puromycin antibody (MABE343, Millipore), and puromycin incorporation was assessed by Western blot (70).

cGAMP activity assay

After transfection with ISD DNA or VACV stimulation, cells were lysed in hypotonic buffer [10 mM tris-HCl (pH 7.4), 10 mM KCl, and 1.5 mM MgCl2] (71) and went through three cycles of freezing and thawing in liquid nitrogen. After incubation with Benzonase at 4°C for 45 min and at 95°C for 15 min, samples were centrifuged at 12,000 rpm for 10 min to remove the denatured proteins. The heat-resistant supernatant was mixed with buffer at a ratio of 1:1 [40 mM Hepes (pH 7.2), 10 mM MgCl2, 4 mM ATP, 4 mM GTP, and digitonin (20 μg/ml)] and then used to treat with THP-1 cells for 30 min at 37°C. Cells were then incubated in fresh medium for 4 hours before measurement of type I interferon and ISGs.

Real-time quantitative reverse transcription PCR

Total RNA was isolated from cells using the TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the protocol provided by the manufacturer. RNA was dissolved in diethyl pyrocarbonate–treated water and reverse-transcribed using the SuperScript III First Strand Synthesis Super Mix Kit (Thermo Fisher) according to the manufacturer’s instruction. cDNA was quantitated with quantitative reverse transcription PCR (qRT-PCR) using the Luna Universal qPCR Master Mix (M3003L, NEB). The fold change was calculated on the basis of the unstimulated control. The qPCR primers for each gene are as follows: IFN-β, 5′-GGAGGACGCCGCATTGAC-3′/5′-TGATAGACATTAGCCAGGAGGTTC-3′; ISG54, 5′-CTGCAACCATGAGTGAGAA-3′/5′-CCTTTGAGGTGCTTTAGATAG-3′; ISG56, 5′-TACAGCAACCATGAGTACAA-3′/5′-TCAGGTGTTTCACATAGGC-3′; and GAPDH, 5′-CGGAGTCAACGGATTTGGTCGTA-3′/5′-AGCCTTCTCCATGGTGGTGAAGAC-3′.

Enzyme-linked immunosorbent assay

THP-1 cells were transfected with HT-DNA or ISD for 16 hours or were exposed with heat-inactivated VACV for 16 hours. The supernatant was collected and used to determine cytokine concentration. Enzyme-linked immunosorbent assay (ELISA) kit was used to determine the concentration of human IFN-β (EHC026, Neobioscience) according to the manufacturer’s instructions.

cGAS, PKR, G3BP1, and STING KO cells

KO cells were generated using the CRISPR-Cas9 system (72). Cells were transfected with lentiCRISPRv2 (catalog number 52961, Addgene) (73) carrying single guide RNAs (sgRNAs) that target cGAS, PKR, G3BP1, or STING. The following sgRNA sequences were used: cGAS gRNA1, AAGGCGGGAAAGTTCGGCCC; cGAS gRNA2, GAACTTTCCCGCCTTAGGCA (11); PKR gRNA1 for HeLa, ATGGCTGGTGATCTTTCAGC; PKR gRNA2 for HeLa, TAATACATACCGTCAGAAGC; PKR gRNA1 for THP-1, TCTCTTCCATTGTAGGATA (74); PKR gRNA2 for THP-1, ATTCAGGACCTCCACATGAT (75); G3BP1 gRNA1, TAGTCCCCTGCTGGTCGGGC; G3BP1 gRNA2, GCTGGTCGGGCGGGAATTTG; STING gRNA1, AGAGCACACTCTCCGGTACC; and STING gRNA2, TCCATCCATCCCGTGTCCCA (11).

After treating the transfected cells with puromycin (to select for the lentiCRISPR-transduced cells), positive cells were diluted under limiting conditions and plated in 96-well plates to obtain single-cell clones. The genotype of KO clones was analyzed by sequencing (fig. S13). To rescue the expression of cGAS in cGAS KO cells, cGAS KO cells were transfected with pcDNA 3.1/V5-His A-cGAS (WT or with the indicated mutation) and selected in G418 (750 μg/ml) for 2 weeks.

SUPPLEMENTARY MATERIALS

stke.sciencemag.org/cgi/content/full/12/609/eaav7934/DC1

Fig. S1. cGAS does not associate with G3BP1 and is not required for SG formation after heat shock.

Fig. S2. cGAS colocalizes with G3BP1 in HFF cells.

Fig. S3. cGAS Cys396 and Cys397 are required for G3BP1 association.

Fig. S4. cGAS KO did not affect arsenite stimulation of SG formation.

Fig. S5. G3BP1 binds to the N-terminal region of cGAS.

Fig. S6. The cGAS activity is not required for G3BP1 association.

Fig. S7. Loss of PKR or G3BP1 prevents DNA-stimulated interferon production.

Fig. S8. Loss of PKR or G3BP1 prevents ISD-stimulated Ifnβ transcription in HFF cells.

Fig. S9. Knockdown of TIA1 does not affect DNA-induced p-IRF3 expression.

Fig. S10. Knockdown of cGAS, PKR, or G3BP1 does not impair induction of interferon production by cGAMP.

Fig. S11. Loss of PKR or G3BP1 prevents DNA virus particle–stimulated interferon production.

Fig. S12. The sizes of arsenite- and ISD-induced G3BP1-containing foci.

Fig. S13. Schematic presentation of the coding regions of cGAS, PKR, G3BP1, and STING that are targeted by CRISPR-Cas9.

Data file S1. cGAS-associated cellular proteins through performing co-IP followed by mass spectrometry and proteomic analysis in 293T cells.

REFERENCES AND NOTES

Acknowledgments: We thank Y. Xiao and L. Li (IPB, CAMS) for technical assistance in performing confocal microscopy and L. Gao (IPB, CAMS) for technical assistance in statistical analysis. Funding: This study was supported by funds from the Ministry of Science and Technology of China (2018ZX10301408-003 and 2018ZX10731101-001-018 to F.G.), from the National Key Plan for Scientific Research and Development of China (2016YFD0500307 to F.G.), from the National Natural Science Foundation of China (81601771 to J.L., 81371808 and 81528012 to F.G., 81702451 to S.M., and 81401673 to F.X.), from CAMS Innovation Fund for Medical Sciences (CIFMS 2018-I2M-3-004 to F.G. and CIFMS 2016-I2M-1-014 to J.W.), from the Canadian Institutes of Health Research (CCI-132561 to C.L.), from the CAMS general fund (2019-RC-HL-012), and the PUMC Youth Fund/Fundamental Research Funds for the Central Universities (3332018202 to S.H.). Author contributions: S.H., J.W., C.L., S.C., and F.G. conceived the project. S.H., H.S., L.Y., J.L., S.M., F.X., C.W., X.L., F.Z., D.Z., Y.H., and L.R. performed the experiments. All authors contributed to experimental design and data analysis. S.H., C.L., and F.G. composed the manuscript. All authors reviewed the manuscript and discussed the work. Competing interests: The authors declare that they have no competing interests. Data and materials availability: The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the iProX partner repository with the dataset identifier PXD012847. All other data needed to evaluate the conclusions in the paper are present in the paper or the Supplementary Materials.

Stay Connected to Science Signaling

Navigate This Article