The stress granule protein G3BP1 promotes pre‐condensation of cGAS to allow rapid responses to DNA

Article15 November 2021free access Source DataTransparent process The stress granule protein G3BP1 promotes pre-condensation of cGAS to allow rapid responses to DNA Ming Zhao Ming Zhao orcid.org/0000-0002-7401-9332 State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China These authors contributed equally to this work Search for more papers by this author Tian Xia Tian Xia State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China These authors contributed equally to this work Search for more papers by this author Jia-Qing Xing Jia-Qing Xing State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China These authors contributed equally to this work Search for more papers by this author Le-Hua Yin Le-Hua Yin State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China These authors contributed equally to this work Search for more papers by this author Xiao-Wei Li Xiao-Wei Li State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Search for more papers by this author Jie Pan Jie Pan State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Search for more papers by this author Jia-Yu Liu Jia-Yu Liu State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Search for more papers by this author Li-Ming Sun Li-Ming Sun State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Search for more papers by this author Miao Wang Miao Wang State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Search for more papers by this author Tingting Li Tingting Li State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Nanhu Laboratory, Jiaxing, China Search for more papers by this author Jie Mao Jie Mao State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Search for more papers by this author Qiu-Ying Han Qiu-Ying Han State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Nanhu Laboratory, Jiaxing, China Search for more papers by this author Wen Xue Wen Xue State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Nanhu Laboratory, Jiaxing, China Search for more papers by this author Hong Cai Hong Cai State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Search for more papers by this author Kai Wang Kai Wang State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Search for more papers by this author Xin Xu Xin Xu State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Search for more papers by this author Teng Li Teng Li State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Search for more papers by this author Kun He Kun He State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Search for more papers by this author Na Wang Na Wang State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Search for more papers by this author Ai-Ling Li Ai-Ling Li State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Nanhu Laboratory, Jiaxing, China Search for more papers by this author Tao Zhou Tao Zhou State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Nanhu Laboratory, Jiaxing, China Search for more papers by this author Xue-Min Zhang Xue-Min Zhang State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Nanhu Laboratory, Jiaxing, China School of Basic Medical Sciences, Fudan University, Shanghai, China Search for more papers by this author Wei-Hua Li Corresponding Author Wei-Hua Li [email protected] orcid.org/0000-0001-8030-9988 State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Search for more papers by this author Tao Li Corresponding Author Tao Li [email protected] orcid.org/0000-0002-0746-8322 State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Nanhu Laboratory, Jiaxing, China School of Basic Medical Sciences, Fudan University, Shanghai, China Search for more papers by this author Ming Zhao Ming Zhao orcid.org/0000-0002-7401-9332 State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China These authors contributed equally to this work Search for more papers by this author Tian Xia Tian Xia State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China These authors contributed equally to this work Search for more papers by this author Jia-Qing Xing Jia-Qing Xing State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China These authors contributed equally to this work Search for more papers by this author Le-Hua Yin Le-Hua Yin State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China These authors contributed equally to this work Search for more papers by this author Xiao-Wei Li Xiao-Wei Li State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Search for more papers by this author Jie Pan Jie Pan State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Search for more papers by this author Jia-Yu Liu Jia-Yu Liu State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Search for more papers by this author Li-Ming Sun Li-Ming Sun State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Search for more papers by this author Miao Wang Miao Wang State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Search for more papers by this author Tingting Li Tingting Li State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Nanhu Laboratory, Jiaxing, China Search for more papers by this author Jie Mao Jie Mao State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Search for more papers by this author Qiu-Ying Han Qiu-Ying Han State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Nanhu Laboratory, Jiaxing, China Search for more papers by this author Wen Xue Wen Xue State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Nanhu Laboratory, Jiaxing, China Search for more papers by this author Hong Cai Hong Cai State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Search for more papers by this author Kai Wang Kai Wang State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Search for more papers by this author Xin Xu Xin Xu State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Search for more papers by this author Teng Li Teng Li State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Search for more papers by this author Kun He Kun He State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Search for more papers by this author Na Wang Na Wang State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Search for more papers by this author Ai-Ling Li Ai-Ling Li State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Nanhu Laboratory, Jiaxing, China Search for more papers by this author Tao Zhou Tao Zhou State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Nanhu Laboratory, Jiaxing, China Search for more papers by this author Xue-Min Zhang Xue-Min Zhang State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Nanhu Laboratory, Jiaxing, China School of Basic Medical Sciences, Fudan University, Shanghai, China Search for more papers by this author Wei-Hua Li Corresponding Author Wei-Hua Li [email protected] orcid.org/0000-0001-8030-9988 State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Search for more papers by this author Tao Li Corresponding Author Tao Li [email protected] orcid.org/0000-0002-0746-8322 State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China Nanhu Laboratory, Jiaxing, China School of Basic Medical Sciences, Fudan University, Shanghai, China Search for more papers by this author Author Information Ming Zhao1, Tian Xia1, Jia-Qing Xing1, Le-Hua Yin1, Xiao-Wei Li1, Jie Pan1, Jia-Yu Liu1, Li-Ming Sun1, Miao Wang1, Tingting Li1,2, Jie Mao1, Qiu-Ying Han1,2, Wen Xue1,2, Hong Cai1, Kai Wang1, Xin Xu1, Teng Li1, Kun He1, Na Wang1, Ai-Ling Li1,2, Tao Zhou1,2, Xue-Min Zhang1,2,3, Wei-Hua Li *,1 and Tao Li *,1,2,3 1State Key Laboratory of Proteomics, National Center of Biomedical Analysis, Beijing, China 2Nanhu Laboratory, Jiaxing, China 3School of Basic Medical Sciences, Fudan University, Shanghai, China *Corresponding author. Tel: +86 13681137903; E-mail: [email protected] *Corresponding author (lead contact). Tel: +86 15810033778; E-mail: [email protected] EMBO Reports (2022)23:e53166https://doi.org/10.15252/embr.202153166 See also: MP Gantier (January 2022) PDFDownload PDF of article text and main figures. Peer ReviewDownload a summary of the editorial decision process including editorial decision letters, reviewer comments and author responses to feedback. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions Figures & Info Abstract Cyclic GMP-AMP synthase (cGAS) functions as a key sensor for microbial invasion and cellular damage by detecting emerging cytosolic DNA. Here, we report that GTPase-activating protein-(SH3 domain)–binding protein 1 (G3BP1) primes cGAS for its prompt activation by engaging cGAS in a primary liquid-phase condensation state. Using high-resolution microscopy, we show that in resting cells, cGAS exhibits particle-like morphological characteristics, which are markedly weakened when G3BP1 is deleted. Upon DNA challenge, the pre-condensed cGAS undergoes liquid–liquid phase separation (LLPS) more efficiently. Importantly, G3BP1 deficiency or its inhibition dramatically diminishes DNA-induced LLPS and the subsequent activation of cGAS. Interestingly, RNA, previously reported to form condensates with cGAS, does not activate cGAS. Accordingly, we find that DNA – but not RNA – treatment leads to the dissociation of G3BP1 from cGAS. Taken together, our study shows that the primary condensation state of cGAS is critical for its rapid response to DNA. Synopsis Host cell encoded cGAS is a critical DNA sensor to detect invading pathogens. The stress-granule protein G3BP1 engages cGAS in a primary condensation state to enable a rapid response to free DNA. G3BP1 primes cGAS for its prompt activation. G3BP1 engages cGAS in a primary condensation state. DNA- but not RNA-interaction leads to the dissociation of G3BP1 from cGAS. Green tea compound epigallocatechin gallate (EGCG) inhibits G3BP1-promoted cGAS phase condensation and activation. Introduction A common event of pathogen infection is the introduction of foreign nucleic acids, such as DNA, into the cytoplasm of the host cell (Barbalat et al, 2011; Gurtler & Bowie, 2013). Cytosolic DNA detecting is one of the fundamental mechanisms for the host to sense invading pathogens (Stetson & Medzhitov, 2006; Takeuchi & Akira, 2010; Gurtler & Bowie, 2013; Wu & Chen, 2014). Among the identified intracellular DNA sensors, cGAS plays a pivotal role in innate immunity (Li et al, 2013b). When it binds to DNA, cGAS is activated and, in turn, catalyzes the synthesis of the second messenger molecule cyclic GMP-AMP (cGAMP) (Ablasser et al, 2013; Gao et al, 2013a; Li et al, 2013b; Sun et al, 2013). As a ligand of the endoplasmic reticulum (ER)–associated adaptor protein STING (also known as ERIS; MITA; MPYS) (Ishikawa & Barber, 2008; Zhong et al, 2008; Sun et al, 2009), cGAMP binds to STING and elicits a series of downstream events that lead to the extensive production of type I interferons (IFNs) and other cytokines (Bowie, 2012; Gao et al, 2013a; Gao et al, 2013b). cGAS thus executes a crucial function in transducing signals by sensing invading pathogens to trigger the activation of anti-infection immune responses. Besides microbial infections, reverse transcription of endogenous retroviruses or cellular damage can also result in the presence of free DNA in the cytoplasm (O'Neill, 2013; Ahn & Barber, 2014; Kassiotis & Stoye, 2016). The chronical activation of cGAS by such self-DNA is a major cause for several autoimmune diseases, such as lupus and Aicardi–Goutières syndrome (AGS; Aicardi & Goutieres, 1984; Ahn & Barber, 2014; Lisnevskaia et al, 2014; Crow & Manel, 2015). Therefore, it is critical to understand how cGAS activation is precisely regulated. It was recently reported that upon activation, DNA robustly induces the formation of liquid droplets containing cGAS (Du & Chen, 2018; Xie et al, 2019). These cGAS-DNA condensates are formed via LLPS. LLPS is a physicochemical process that allows macromolecules, such as proteins and nucleic acids, to condensate into a dense phase (Jiang et al, 2015; Boeynaems et al, 2018; Fox et al, 2018; Martin et al, 2019). Accumulating evidence demonstrates that LLPS is a key mechanism underlying the formation of membrane-less organelles (Feric et al, 2016; Lee et al, 2016; Fei et al, 2017; Mitrea et al, 2018; Fujioka et al, 2020; Yasuda et al, 2020). These temporarily assembled cellular compartments confer an important capacity for the cells to dynamically and efficiently organize functional units or reactions in response to different stresses (Kroschwald et al, 2015; Riback et al, 2017; Franzmann et al, 2018; Franzmann & Alberti, 2019). In the case of DNA detection by cGAS, the liquid droplets provide microreactors, in which cGAS, the enzyme, and the reactants, DNA, ATP and GTP, are highly concentrated (Du & Chen, 2018). This process allows the efficient activation of cGAS. Another well-studied membrane-less organelle is the stress granule (SG) (Molliex et al, 2015; Protter & Parker, 2016; Boeynaems et al, 2017; Youn et al, 2019). SGs are cytoplasmic puncta, mainly containing RNAs and proteins, assembled through LLPS unpon different types of stresses, such as oxidative stress and RNA virus infection (Buchan & Parker, 2009; Onomoto et al, 2014; McCormick & Khaperskyy, 2017). Among the known proteins in SG, G3BP1 is believed to be a critical organizer of SG assembly and regulates properties and composition of SGs (Tourriere et al, 2003; Yang et al, 2020). As a core protein for SG regulation, G3BP1 contains a special amino acid sequence feature called the intrinsically disordered region (IDR) (Guillen-Boixet et al, 2020; Sanders et al, 2020; Yang et al, 2020). For instance, the arginine–glycine–glycine (RGG) motif of G3BP1 is a IDR (Chong et al, 2018). The IDR in G3BP1 is critical for mediating weak multivalent protein–protein or protein–nucleic acid interactions and facilitating the LLPS process (Guillen-Boixet et al, 2020; Sanders et al, 2020; Yang et al, 2020). We recently found that G3BP1 is critical for DNA binding and activation of cGAS (Liu et al, 2019). In the current study, we further show that G3BP1 primes cGAS activation by forming primary condensates with cGAS in the resting state, while DNA engagement leads to the dissociation of G3BP1 from cGAS. Multiple regions of G3BP1, including the nuclear transporter factor 2 (NTF2)–like domain, the RNA recognition module (RRM) and the RGG domain, were required for cGAS-G3BP1 binding and the formation of cGAS-G3BP1 condensates. In the presence of G3BP1, DNA was much more efficient at inducing LLPS and activation of cGAS. In addition, zinc ions that are known to enhance cGAS activity (Du & Chen, 2018), also induced the formation of cGAS condensates both in vitro and in vivo. Thus, our study shows that the G3BP1-mediated formation of the primary condensation state of cGAS is a critical step for an expeditious response to cytosolic DNA. Results G3BP1 engages cGAS in a condensed state We previously reported that G3BP1 promotes the DNA binding and activation of cGAS (Liu et al, 2019). Interestingly, when recombinant cGAS was incubated with G3BP1, but not when it was incubated with bovine serum albumin (BSA), the control protein, the solution became turbid (Fig 1A). We then observed the solution under a microscope and found that many droplets appeared when cGAS and G3BP1 were mixed, and the formation of these droplets was further increased in the presence of DNA (Fig 1A). cGAS activity can be examined using an in vitro cGAMP synthesis assay, in which recombinant cGAS protein is incubated with DNA, ATP, and GTP (Li et al, 2013a; Dai et al, 2019). Using this assay, we showed that with G3BP1, cGAS produced much higher amounts of cGAMP (Fig 1B). To further analyze the cGAS-G3BP1 droplets, we incubated cGAS with G3BP1 at different concentrations and found that G3BP1 promoted cGAS condensation in a dosage-dependent manner (Fig 1C). Because cGAS is known to undergo LLPS upon DNA engagement, we further purified recombinant cGAS and G3BP1 proteins with heparin affinity chromatography followed by Ni-agarose purification to remove potential nucleic acid contaminations. With the purified proteins, we obtained similar results showing that G3BP1 alone can form condensates with cGAS (Fig EV1A). Figure 1. G3BP1 engages cGAS in a condensed state Turbidity and bright-field microscope images of indicated groups (left) and the quantitative analysis of turbidity of samples were measured by absorbance at 600 nm (right). 10 μM cGAS, 5 μM G3BP1, 5 μM BSA and 100 nM dsDNA (60 bp) were used in these assays. n = 3 biological replicates. cGAMP production assays. Recombinant cGAS (10 μM) and G3BP1 (10 μM) were incubated with 100 nM dsDNA. The produced cGAMP was quantitatively analyzed by liquid chromatography–mass spectrometry/multiple reaction monitoring (LC-MS/MRM). ND, not detected. n = 3 technical replicates. Bright-field microscope images of cGAS (10 μM) with indicated concentrations of BSA or G3BP1 (left) and a quantitative analysis of the total area of condensates (right). n = 3 biological replicates. Time-lapse imaging of liquid droplets fusion after mixing 10 μM cGAS-mCherry with 1 μM FAM-labeled dsDNA. Time-lapse imaging of liquid droplets fusion after mixing 40 μM cGAS-mCherry with 10 μM G3BP1-mEGFP. FRAP analysis of DNA-induced droplets of cGAS (1 μM dsDNA, 10 μM cGAS-mCherry), the yellow dotted circle indicated the region of photobleaching. FRAP curve of (F). n = 3 biological replicates. FRAP analysis of cGAS-G3BP1 droplets. 10 μM cGAS-mCherry and 5 μM G3BP1 were used. The yellow dotted circle indicated the region of photobleaching. FRAP curve of (H). n = 3 biological replicates. Data information: Representative images are shown (A, C–F and H). Error bars, mean with s.d. (A–C, G and I), *P < 0.05, ****P < 0.0001, two-tailed t-test. NS, non-significant; FRAP, fluorescence recovery after photobleaching. Scale bars, 10 μm (A and C), 5 μm (D), 3 μm (E), 2 μm (F and H). See also Fig EV1. Download figure Download PowerPoint Click here to expand this figure. Figure EV1. G3BP1 engages cGAS in a condensed state A. Bright-field microscope images of indicated groups (left) and a quantitative analysis of total area of the condensates. n = 3 biological replicates. 10 μM cGAS, 5 μM G3BP1, 5 μM BSA and 500 nM dsDNA were used in these assays. B. Coomassie blue staining of purified recombinant cGAS-mCherry protein. C. Fluorescent images of liquid droplets after mixing 10 μM cGAS-mCherry with 1 μM FAM-labeled dsDNA. D. Immunofluorescent staining of cGAS in WT and cGAS−/− cells. Images were acquired with Leica TCS SP8 Confocal Microscopy. E, F. Immunoblotting of WT and G3BP1−/− cells with indicated antibodies. G. Immunofluorescent staining of cGAS and G3BP1 in both WT and G3BP1−/− HeLa cells. Images were acquired with Leica TCS SP8 Confocal microscopy. The white arrows indicate the cGAS-G3BP1 condensates. H. Immunofluorescent staining of cGAS and G3BP1 in both WT and G3BP1−/− HeLa cells. 3D images were reconstituted by Leica LAS X software. I. Quantitative analysis of total cGAS puncta number (left) and volume (right) per cell of (H). n = 21 cells. J. Immunoblotting of fractionated lysates from WT and G3BP1−/− HeLa cells. GAPDH and Lamin B1 blots were used as controls for cytoplasmic and nuclear fractions, respectively. Data information: Representative images are shown (A, C–H, and J). Error bars, mean with s.d. (A and I). Scale bars, 10 μm (A), 14 μm (C), 5 μm (D, G and H). WT, wild type; IB, immunoblotting; NS, non-significant. Hoechst (blue), nuclear staining. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, two-tailed t-test. Source data are available online for this figure. Download figure Download PowerPoint We then examined if the G3BP1-mediated formation of cGAS droplets was a similar process to DNA-triggered cGAS LLPS. We first purified mCherry-tagged recombinant cGAS (Fig EV1B) and induced cGAS LLPS with a 60 base-pair (bp) FAM (6-carboxy-fluorescein)–labeled double-stranded DNA (dsDNA). As shown in Fig EV1C, mCherry-tagged cGAS robustly formed droplets with dsDNA. By recording the dynamic formation process of these droplets, we found that the smaller cGAS-DNA droplets can fuse into bigger ones (Fig 1D, Movie EV1). In contrast, the fusion of cGAS-G3BP1 droplets was barely detected (Fig 1E, Movie EV2). Next, using fluorescence recovery after photobleaching (FRAP), we showed that the fluorescence intensity of cGAS-DNA droplets recovered soon after bleaching (Fig 1F and G, Movie EV3). These data indicate that the cGAS-DNA droplets are dynamically exchanging molecules with the environment, which is a hallmark of LLPS. The cGAS-G3BP1 droplets, however, exhibited much lower fluorescence recovery rate (Fig 1H and I, Movie EV4). Thus, the G3BP1-mediated formation of cGAS droplets was not as dynamic as for the droplets induced by DNA. These observations suggested that cGAS may undergo gel-like transitions without the engagement of DNA. We termed this G3BP1-engaged gel-like transition of the cGAS ‘primary condensation state’. G3BP1 engages cGAS in a condensed state in vivo To confirm the above findings in cells, we analyzed cGAS in both wild-type (WT) and G3BP1-deficient U937 cells, a monocytic cell line that is widely used to study cGAS activation (Watson et al, 2015). To do so, we first confirmed the specificity of our anti-cGAS antibodies in immunofluorescence experiments using cGAS-deleted U937 and HeLa cells (Fig EV1D). We then generated G3BP1-null U937 and HeLa cells (Fig EV1E and F). Using high-resolution microscopy, we found that cGAS showed particle-like morphological characteristics in U937 cells (Fig 2A). Consistent with our previous findings (Liu et al, 2019), cGAS and G3BP1 were colocalized in these cells (Fig 2A). Importantly, the particle-like characteristics of cGAS was markedly weakened when G3BP1 was deleted (Fig 2A–C). This phenomenon was further confirmed in G3BP1-deficient HeLa cells (Fig EV1G–I). The above data showed that in G3BP1-deficient cells, cGAS appeared to predominantly localize to the nucleus. To examine the cGAS subcellular distribution upon G3BP1 deletion, we detected cGAS expression in the cytosolic and nuclear fractions from both WT and G3BP1−/− cells and found that G3BP1 deficiency did not obviously affect the subcellular distribution of cGAS (Figs 2D and Fig EV1J). Thus, the ablation of G3BP1 resulted in the disorganization of cGAS primary condensation in cytoplasm. The G3BP1-engaged primary condensation state of cGAS seemed to be important, because the deficiency of G3BP1 resulted in dampened cGAS activation (Fig 2E–G). Figure 2. G3BP1 engages cGAS in a condensed state in vivo Immunofluorescent staining of cGAS in both wild-type (WT) and G3BP1−/− U937 cells. Leica TCS SP8 Confocal Microscopy was used to acquire images. The white arrows indicate the cGAS-G3BP1 condensates. Immunofluorescent staining of cGAS in both WT and G3BP1−/− U937 cells. 3D images were reconstituted with Leica LAS X software. Quantitative analysis of total cGAS puncta number (left) and volume (right) per cell of (B). n = 51 cells. Immunoblotting of fractionated lysates from WT and G3BP1−/− U937 cells. GAPDH and Lamin B1 blots were used as controls for cytoplasmic and nuclear fractions, respectively. qPCR analysis of IFNB mRNA levels in HT-DNA–treated WT and the indicated rescued cells (U937) (top). Immunoblotting analysis of indicated proteins (bottom). n = 3 biological replicates. ELISA analysis of the secreted IFN-β in HT-DNA-treated WT or G3BP1−/− U937 cells. n = 3 biological replicates. cGAMP production in ISD-treated WT or G3BP1−/− U937 cells were analyzed by LC-MS/MRM. ND, not detected. n = 3 technical replicates. Data information: Representative images are shown (A and B). Scale bars, 5 μm. Hoechst (blue), nuclear staining. Error bars, mean with s.d (C, E–G), *P < 0.05, **P < 0.01, ****P < 0.0001, two-tailed t-test. NS, non-significant; WT, wild type; IB, immunoblotting; HT-DNA, Herring testes DNA; ISD, interferon stimulatory DNA. See also Figs EV1 and EV2. Source data are available online for this figure. Source Data for Figure 2 [embr202153166-sup-0011-SDataFig2.pdf] Download figure Download PowerPoint It has been reported that cancer cells exhibit basal levels of cytoplasmic DNA (Shen et al, 2015). To rule out the possibility that the cGAS condensation we observed in U937 and HeLa cells may be attributed to existing cytoplasmic DNA in these cells, we measured the cytoplasmic dsDNA levels in HeLa cells and U937 cells using anti-dsDNA antibodies. We also included a human fibroblast cell line, Hs27, in this experiment. Using immunofluorescence staining, we detected cytosolic dsDNA in both U937 and HeLa cells, but not in Hs27 cells (Fig EV2A). Using Hs27, we found that, consistently, the deficiency of G3BP1 significantly dampened cytosolic cGAS condensation (Fig EV2B–D). Click here to expand this figure. Figure EV2. G3BP1 engages cGAS in a condensed state in human fibroblast cell line Immunofluorescent staining of cytosolic dsDNA in indicated cells. Immunofluorescent staining of cGAS in both WT and G3BP1-knockdown Hs27 cells. Quantitative analysis of total cGAS puncta number per cell in (B). n = 22 cells. Error bar, mean with s.d. ****P < 0.0001, two-tailed t-test. Knockdown effect of G3BP1 was analyzed by immunoblotting. Data information: Representative images are shown (A, B and D). Scale bars, 5 µm (A and B). Hoechst (blue), nuclear staining. WT, wild type; IB, immunoblotting; NC, negative control. Source data are available online for this figure. Download figure Download PowerPoint Together, our data indicate that G3BP1 engages cGAS in a primary condensation state in cells. The engagement with DNA, but not RNA, leads to the dissociation of G3BP1 from cGAS Our previous study suggested that DNA treatment gradually disrupted the interaction of G3BP1 with cGAS (Liu et al, 2019). Here, we further investigated whether G3BP1 participated in DNA-triggered LLPS of cGAS. We first incubated mEGFP-tagged G3BP1 with mCherry-tagged cGAS and analyzed cGAS-G3BP1 condensation at early time points. Our data showed that G3BP1 formed condensates with cGAS immediately after the two proteins were mixed (Fig 3A, Movie EV5). By measuring the fluorescence intensity distribution in individual condensate (Fig 3B) and fluorescence intensity changes over time (Fig EV3A), we further confirmed this finding. Interestingly, when dsDNA was added, cGAS quickly condensed with DNA and dissociated from G3BP1 (Figs 3C and EV3B, Movie EV6). Further analysis on the fluorescence intensity distributions and fluorescence changes of cGAS-DNA droplets yielded consistent data (Figs 3D and EV3C). Given that RNA was recently reported to induce cGAS LLPS without activating cGAS (Du & Chen, 2018), we then examined the effect of RNAs on cGAS-G3BP1 primary condensates. We found that both dsRNA and single-stranded RNA (ssRNA) condensed with cGAS and G3BP1, but the stimulation with RNAs did not trigger the dissociation of G3BP1 from cGAS (Figs 3E and F, and EV3D–H, Movie EV7). These data indicated that the engagement with DNA, but not with RNA, leads to the dissociation of G3BP1 from cGAS. Figure 3. The engagement with DNA, but not with RNA, leads to the dissociation of G3BP1 from cGAS Time-lapse imaging of cGAS-G3BP1 condensates. Fluorescence intensity distribution on cGAS-G3BP1 droplets. A quantitative analysis of a representative cGAS-G3BP1 droplet is shown. Along the white line on the merged image, the fluorescence intensity of both G3BP1 and cGAS channels were recorded. Time-lapse imaging of cGAS-DNA droplets in the presence of G3BP1-mEGFP. cGAS, G3BP1, and Cy5-dsDNA were incubated. A quantitative analysis of a representative cGAS-DNA droplet is shown. Along the white line on the merged image, the fluorescence intensity of G3BP1, cGAS, and dsDNA channels were recorded. Time-lapse imaging of condensates formed by cGAS, G3BP1, and dsRNA. cGAS, G3BP1, and Cy5-dsRNA were incubated. Similar quantification analysis as in D was performed. Data information: Representative images are shown (A–F). Scale bars, 10 μm (A–E), 5 μm (F). 45 μM cGAS-mCherry, 10 μM G3BP1-mEGFP, 2 μM dsDNA and 2 μM dsRNA were used in the assays.