DNA-Mediated Cyclic GMP−AMP Synthase− Dependent and −Independent Regulation of Innate Immune Responses

This information is current as Kou Motani, Shinji Ito and Shigekazu Nagata of September 28, 2021. J Immunol 2015; 194:4914-4923; Prepublished online 8 April 2015; doi: 10.4049/jimmunol.1402705 http://www.jimmunol.org/content/194/10/4914 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2015/04/08/jimmunol.140270 Material 5.DCSupplemental

References This article cites 57 articles, 19 of which you can access for free at: http://www.jimmunol.org/ http://www.jimmunol.org/content/194/10/4914.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists by guest on September 28, 2021

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2015 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

DNA-Mediated Cyclic GMP–AMP Synthase–Dependent and –Independent Regulation of Innate Immune Responses

Kou Motani,*,† Shinji Ito,‡ and Shigekazu Nagata*

Cytoplasmic DNA activates cyclic GMP–AMP synthase (cGAS) to produce cyclic 29-5939-59GMP–AMP dinucleotide (295 9cGAMP). The binding of 2959cGAMP to an adaptor protein, stimulator of IFN genes (STING), activates a transcription factor, IFN regulatory factor 3, leading to the induction of IFN and chemokine gene expression. In this study, we found that the 2959cGAMP-dependent STING activation induced highly upregulated CXCL10 gene expression. Formation of a distinct STING dimer, which was detected by native PAGE, was induced by 2959cGAMP, but not 39-5939-59cGAMP. Analysis of DNase II2/2 mice, which constitutively produce IFN-b and CXCL10, showed the accumulation of 2959cGAMP in their fetal livers and spleens, suggesting that the undigested DNA accumulating in DNase II2/2 cells may have leaked from the lysosomes into the cytoplasm. 2/2 9 9 The DNase II mouse embryonic fibroblasts produced 2 5 cGAMP in a cGAS-dependent manner during apoptotic cell engulf- Downloaded from ment. However, cGAS deficiency did not impair the STING-dependent upregulation of CXCL10 in DNase II2/2 mouse embryonic fibroblasts that was induced by apoptotic cell engulfment or DNA lipofection. These results suggest the involvement of a cGAS- independent additional DNA sensor(s) that induces the STING-dependent activation of innate immunity. The Journal of Immu- nology, 2015, 194: 4914–4923.

he innate immune system in vertebrates is activated by their DNA is degraded by the lysosomal DNase, DNase II. We http://www.jimmunol.org/ pathogens via the recognition of pathogen-associated mo- previously reported that the impaired degradation of apoptotic or T lecular patterns (1). The genomic DNAs of various patho- pyrenocyte DNA in DNase II2/2 macrophages causes a strong gens function as pathogen-associated molecular patterns (2, 3) that inflammatory response that leads to lethal anemia or polyarthritis are recognized by DNA sensors in endosomes/lysosomes or the (7–10). Similarly, the impaired degradation of endogenous cyto- cytoplasm. Recognition by the DNA sensors activates the expression plasmic retrotransposon DNA by a mutant form of three prime of various cytokines, such as IFN-b, CXCL10, and TNF-a,which repair exonuclease (also known as DNase III), causes Aicardi- leads to the induction of acquired immunity. Normally, endogenous Goutieres syndrome, a systemic lupus erythematosus type of au- self-DNA is localized to the nucleus or mitochondria and does not toimmune disease (11). elicit an immune response. However, mislocalized DNA can activate DNA-induced signaling that leads to innate immune activation by guest on September 28, 2021 innate immunity, leading to sterile inflammation accompanied by has been extensively studied (2–4, 12). TLR9 recognizes viral and autoimmunity (4). bacterial nonmethylated CpG DNA in endosomes and transduces Many cells undergo apoptosis during animal development and signals through the Myd88 adaptor protein to activate two tran- tissue turnover and are engulfed by macrophages (5). During de- scription factors: IFN response regulatory factor (IRF)7, which finitive erythropoiesis, pyrenocytes, nuclei surrounded by plasma induces the upregulation of IFN-b and CXCL10, and NF-kB, which membrane, are released from erythroblasts at erythroblastic is- upregulates inflammatory cytokines (13). The accumulation of en- lands and engulfed by macrophages (6). After their engulfment, dogenous self-DNA in the lysosomes of DNase II2/2 cells or in the apoptotic cells and pyrenocytes are taken up by lysosomes, and cytoplasm of three prime repair exonuclease 12/2 cells leads to IFN-b and CXCL10 gene upregulation via IRF3 or IRF7, but this process is TLR9-independent (14, 15), suggesting the presence of *Department of Medical Chemistry, Graduate School of Medicine, Kyoto University, additional DNA sensor(s). † Sakyo-ku, Kyoto 606-8501, Japan; Division of Cell Signaling, Fujii Memorial In- Stimulator of IFN genes (STING), originally identified as an stitute of Medical Sciences, University of Tokushima, Tokushima 770-8503, Japan; and ‡Medical Research Support Center, Graduate School of Medicine, Kyoto Uni- adaptor protein located in the endoplasmic reticulum that activates versity, Sakyo-ku, Kyoto 606-8501, Japan IFN genes (16), was recently shown to bind cyclic dinucleotides Received for publication October 28, 2014. Accepted for publication March 11, such as cyclic di-GMP (c-di-GMP) and cyclic di-AMP produced 2015. in bacteria (17, 18). Subsequently, a series of reports showed that This work was supported in part by Grants-in-Aid for Specially Promoted Research mammalian cells produce a similar, but distinct, cyclic dinucleo- from the Japan Society for the Promotion of Science (to S.N.), and by a Grant-in-Aid for Research Activity Start-up from the Japan Society for the Promotion of Science tide (or cyclic GMP–AMP [cGAMP]) in response to cytoplasmic (to K.M.). DNA (3). The binding of DNA to cGAMP synthase (cGAS) Address correspondence and reprint requests to Prof. Shigekazu Nagata, Department activates the enzyme to synthesize 29-59,39-59-cGAMP (2959- of Medical Chemistry, Graduate School of Medicine, Kyoto University, Yoshida- cGAMP) from ATP and GTP (19–22). The 2959-cGAMP then Konoe, Sakyo-ku, Kyoto 606-8501, Japan. E-mail address: [email protected]. kyoto-u.ac.jp binds to STING and induces its dimerization, which leads to The online version of this article contains supplemental material. TBK1 activation and IRF3 phosphorylation, and then to the in- b Abbreviations used in this article: c-di-GMP, cyclic di-GMP; cGAMP, cyclic GMP–AMP; duction of IFN- gene expression (23). The innate immune re- cGAS, cGAMP synthase; CRISPR, clustered regularly interspaced short palindromic sponse to DNA viruses as well as reverse-transcribed retroviral repeats; FasL, Fas ligand; IRF, IFN response regulatory factor; MEF, mouse embryonic DNA is dependent on STING (24, 25) as well as cGAS (26, 27). fibroblast; poly(I:C), polyinosinic-polycytidylic acid; STING, stimulator of IFN genes. Surprisingly, the activation of innate immunity by the undigested 2/2 Copyright Ó 2015 by The American Association of Immunologists, Inc. 0022-1767/15/$25.00 DNA of apoptotic cells engulfed by DNase II cells is also www.jimmunol.org/cgi/doi/10.4049/jimmunol.1402705 The Journal of Immunology 4915

STING-dependent (28). However, the mechanism by which DNA CTGCGGCCCGCAAAGGT-39 and 59-TAAAACCTTTGCGGGCCGCA- is recognized in DNase II2/2 cells has not yet been elucidated. GCTTTCCGCGTGGGCC-39, and 59-CACCGAAAGCTGCGGCCCGC- 9 9 9 In this study, we found that DNase II2/2 mouse embryonic AAAG-3 and 5 -AAACCTTTGCGGGCCGCAGCTTTC-3 ) were de- 9 9 signed using the CRISPR Design Tool at Dr. Zhang’s Laboratory (http:// fibroblasts (MEFs) produced 2 5 -cGAMP in a cGAS-dependent www.genome-engineering.org/crispr/?page_id=41). They were inserted manner, upon apoptotic cell engulfment. However, the apoptotic into BbsI-digested pX260 or pX330 (Addgene, Cambridge, MA) and used 2 2 cell engulfment-mediated induction of CXCL10 expression in these to transfect DNase II / MEFs by electroporation using the NEPA21 cells was independent of cGAS. DNA introduced into DNase II2/2 system (135 V, 10 ms; Nepagene, Chiba, Japan). After 3-d incubation at 37˚C, the cells were subjected to limiting dilution, and the clones con- MEFs by lipofection also activated CXCL10 gene expression in taining mutated STING or cGAS gene were identified by sequencing the a cGAS-independent manner. These results suggest the presence of PCR products flanking the CRSPR-target site. Primers for PCR were as additional DNA sensor(s) that recognizes lysosome-localized DNA follows: STING, 59-CTCAGTGCTGAGACTCAGAC-39 and 59-AGAG- and activates the STING-mediated innate immunity. GTCCTACGTTCAATTC-39; cGAS, 59-ATACTGACCGGCTACGTTCC- 39 and 59-CAACTTTATTCACCGTCTCG-39. Materials and Methods Introduction of DNA, RNA, and cGAMP into MEFs Mice, cells, and reagents To introduce DNA or dsRNA into MEFs, plasmid DNA or poly(I:C) was C57BL/6J mice were purchased from Japan SLC. CAD2/2 and DNase II2/2 incubated for 20 min with Lipofectamine 2000 at a ratio of 1:2 (w/v) in IFNIR2/2 mice were described previously (8, 9). Mice were housed in DMEM, and added to cells at a final concentration of 0.1–1 mg/ml. In some 5 a specific pathogen-free facility at the Kyoto University, Graduate School cases, MEFs (3 3 10 in 0.1 ml Opti-MEM containing 1.0 mgDNAor of Medicine, and all animal experiments were carried out in accordance RNA) were subjected to electroporation at 125 V for 2.5 ms using the NEPA21 system. After electroporation, the cells were cultured at 37˚C in with protocols approved by the Animal Care and Use Committee of the Downloaded from Kyoto University Graduate School of Medicine. 6 ml fresh DMEM containing 10% FCS (resulting in a final DNA con- DNase II2/2 MEFs were described previously (14). MEFs, HEK293T centration of 0.15 mg/ml). To deliver cyclic dinucleotides into cells, the (293T), and Plat-E cells (29) were cultured in DMEM containing cells were incubated with them for 30 min at 37˚C in digitonin per- 10% FCS. To establish the STING transformants, mouse STING cDNA meabilization solution (50 mM PIPES buffer [pH 7.0], 100 mM KCl, (GenBank accession number NM_028261; http://www.ncbi.nlm.nih.gov/ 3 mM MgCl2, 85 mM sucrose, 0.2% endotoxin-free BSA [Nacalai Tesque, nuccore/NM_028261) was prepared by RT-PCR from DNase II2/2 fetal Kyoto, Japan], 0.1 mM DTT, 1 mM ATP, 0.1 mM GTP, and 10 mg/ml liver macrophages (14), Flag tagged at the C terminus, and introduced digitonin), as described (17). The cells were then washed with medium and into the mouse retrovirus vector, pMXs-puro (30). The STING mutant cultured in fresh medium at 37˚C. http://www.jimmunol.org/ (R231A), in which arginine at position 231 was changed to alanine, was constructed with recombinant PCR using primers of 59-CACCGGC- Engulfment of apoptotic cells 9 9 CAGTGTGGTGGAAGATGCCATACTCCAACCT-3 (BstXI-Start) and 5 - Phagocytosis of apoptotic cells was quantified by labeling the apoptotic 9 CGATTCTTGATGCCAGCAGCGTCGATGTTTTGCTGGG-3 (R231A), and cells with pHrodo, as described (36, 37). In brief, mouse thymocytes were 9 9 5 - CCCAGCAAAACATCGACGCTGCTGGCATCAAGAATCG-3 (R231A) treated at 37˚C for 2 h with 50 U/ml FasL, labeled with 0.1 mg/ml pHrodo, 9 and 5 - GCTAACCACTGTGCTGGCTACTTATCGTCGTCATCCTTGT- cultured with MEFs for 2 h at a ratio of 50:1, and analyzed by flow 9 AATCGATGAGGTCAGTGCGGAGTG-3 (Flag-Stop-BstXI). Ecotropic cytometry using the FACSCanto II (BD Biosciences). To examine gene retrovirus was produced by transfecting Plat-E cells with the pMXs vector expression during apoptotic cell engulfment, the MEFs were coincubated using Fugene 6 (Promega, Madison, WI). The virus was diluted 3-fold with for 12 h with the FasL-treated apoptotic CAD2/2 thymocytes in the fresh media and used to infect MEFs. In some cases, stable transformants m presence or absence of 1 g/ml rDNase II, washed with PBS, and further by guest on September 28, 2021 m were selected by culturing in medium containing 1 g/ml puromycin. The cultured in DMEM containing 10% FCS and 10 mM chloroquine (Nacalai DNase II2/2 MEFs expressing TIM4 and MER TK were established by 2/2 Tesque). In all experiments, the FCS was pretreated at 70˚C for 30 min to infecting DNase II MEFs with retro- and lentiviruses expressing TIM4 inactivate nucleases in the serum (38). (31) and MER TK cDNAs (32), respectively. High TIM4- and MER TK- expressing transformants were obtained by sorting with an anti-TIM4 mAb Real-time RT-PCR (31) and a biotin-conjugated goat anti-MER TK Ab (R&D Systems, Minneapolis, MN) using the FACSAria II system (BD Biosciences, Cell line RNA was isolated using the RNeasy kit with the RNase-Free Franklin Lakes, NJ). DNase Set (Qiagen, Venlo, The Netherlands). Fetal livers were homoge- Leucine zipper-tagged human Fas ligand (FasL) (33) was produced in nized in ISOGEN (Nippon Gene, Tokyo, Japan) using the TissueLyser COS7 cells. The Flag-tagged STING-expressing construct was introduced (Qiagen), and RNAwas isolated using the RNeasy kit. The RNAwas reverse into pEF-BOS (34) and then expressed in 293T cells. The cells were lysed transcribed using the High Capacity RNA-to-cDNA kit (Life Technologies), by suspension in buffer A (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, and real-time PCR was performed with the Light-Cycler 480 (Roche b 9 1 mM EDTA, 1 mM EGTA, 10% glycerol, and a mixture of protease Diagnostics). The primers used for RT-PCR were as follows: IFN- ,5- 9 9 inhibitors [Roche Diagnostics, Basel, Switzerland]) containing 1% Non- CCACCACAGCCCTCTCCATCAACTAT-3 and 5 -CAAGTGGAGAG- 9 9 idet P40. After the insoluble materials were removed by centrifugation at CAGTTGAGGACATC-3 ; CXCL10, 5 -CCATCAGCACCATGAACCC- 9 9 9 20,000 3 g for 15 min at 4˚C, the supernatant was incubated with anti-Flag AAGT-3 and 5 -CACTCCAGTTAAGGAGCCCTTTTAGACC-3 ; STING, 9 9 9 (M2)-Sepharose (Sigma-Aldrich, St. Louis, MO) and eluted with buffer A 5 -ACCTAGCCTCGCACGAACT-3 and 5 -CCAACTGAGGTATATGT- 9 9 9 9 containing 0.1% Nonidet P40 and 200 mg/ml 33Flag peptide (Sigma- CAGCAG-3 ; cGAS, 5 -AGAAGGACTACCTATTCAAGGCT-3 and 5 - 9 9 Aldrich). Analysis of the purified STING by SDS-PAGE showed that it GGGTACGAGATAAAACGGCTC-3 ; Tlr9, 5 -ACGGGAACTGCTACT- 9 9 9 b was nearly homogenous. ACAAGA-3 and 5 -CCCAGCTTGACAATGAGGTTAT-3 ; -actin, 9 9 9 HRP-conjugated mouse anti-Flag mAb (M2) and mouse anti–a-tubulin 5 -TGTGATGGTGGGAATGGGTCAG-3 and 5 -TTTGATGTCACGC- 9 mAb (DM1A) were from Sigma-Aldrich. Rabbit anti-STING mAb, anti- ACGATTTCC-3 . phosphoTBK1 (Ser172) mAb, and anti-TBK1 mAb were purchased from Cell Signaling (Danvers, MA). Lipofectamine 2000 was from Life Tech- Preparation of dicyclic nucleotides from MEFs and mouse nologies (Carlsbad, CA). Polyinosinic-polycytidylic acid [poly(I:C)] (low tissues m.w.) was from Invivogen (San Diego, CA). Cyclic dinucleotides (cyclic 9 9 9 9 9 9 9 9 Dicyclic nucleotides were prepared from mouse tissues and cell lines by di-GMP, 3 -5 3 -5 -cGAMP, and 2 -5 3 -5 -cGAMP) were from Biolog the phenol-extraction method (39). In brief, the fetal livers and spleens (Hayward, CA). Human DNase II was provided by Chugai Pharmaceutical were homogenized at 4˚C using the TissueLyser and Polytron homoge- (Tokyo, Japan). nizer, respectively, in a phenol extraction solution (20 mM Tris-HCl Gene editing using the Crispr/Cas system buffer [pH 7.5] and 1 mM EDTA-saturated phenol/chloroform/H2O [6:2:2]; 1 ml per 30 mg tissue). The MEFs (2.5 3 105 resuspended in The clustered regularly interspaced short palindromic repeats (CRISPR)- 200 ml water) were extracted by vigorous shaking with 1 ml phenol- Cas (CRISPR-associated) system (35) was used to establish MEFs lack- extraction solution. After centrifugation at 10,000 3 g for 5 min at 4˚C, ing the STING or cGAS gene. A pair of oligonucleotides for STING the upper aqueous phase was recovered and shaken with chloroform, (59-AAACCTTTTCTTCACAGACTGCAGAGACTTCCGCGT-39 and 59- followed by ether extraction. If necessary, the samples were filtrated TAAAACGCGGAAGTCTCTGCAGTCTGTGAAGAAAAG-39) and two through an Amicon ultra 3K filter, and then applied to a C18 column pairs of oligonucleotides for cGAS (59-AAACGGCCCACGCGGAAAG- (COSMOSIL 5C18-PAQ, 4.6 mm 3 250 mm; Nacalai Tesque) that had 4916 cGAS-DEPENDENT AND -INDEPENDENT INNATE IMMUNE RESPONSES been equilibrated with 20 mM triethylammonium bicarbonate in water. monomer, confirming that the introduction of DNA into the cells The nucleotides were then separated by a linear gradient of 0–100% induces STING dimerization (43). acetonitrile using HPLC (PLC2020; Gilson, Middleton, WI). To prepare The analysis of cell lysates prepared from untransfected DNase the samples for measuring STING-dimerizing activity, the solvent in the 2/2 fractions was evaporated, and the sample was dissolved in water. II MEFs using native PAGE, followed by Western blotting with anti-Flag mAb, also revealed the presence of two STING- STING dimerization assay Flag bands that were consistent with the monomeric and dimeric Cell lysates prepared from MEF transformants expressing STING-Flag or forms (Fig. 1D), whereas cell lysates from the DNA-transfected 2 2 the purified STING-Flag were used to assay STING dimerization activity. In DNase II / MEFs showed an additional STING-Flag band that 2/2 brief, DNase II MEFs expressing STING-Flag were lysed on ice for moved slightly slower than the dimer. The intensity of the addi- 15 min in buffer B (10 mM PIPES-KOH buffer [pH 7.0], 50 mM NaCl, tional band increased dose dependently with the concentration of 5 mM MgCl2, 5 mM EGTA, 10% glycerol, and a mixture of protease inhibitors and phosphatase inhibitors [Roche Diagnostics]) containing 1% the transfected DNA, whereas the intensities of the STING dimer Nonidet P40 and centrifuged at 20,000 3 g for 15 min. The cell lysates and monomer showed a concomitant reduction. These results (1 mg protein) were incubated with nucleotide samples on ice for 30 min in suggested that the transfected DNA induced the formation of m a 10- l reaction mixture (20 mM PIPES-KOH [pH 7.0] and 5 mM MgCl2). a specific STING dimer, the conformation of which may differ For the assay using purified STING-Flag, the nucleotide samples were incubated on ice for 30 min with 20 fmol STING-Flag in a 10 ml reaction from that of the STING dimer formed without DNA stimulation. mixture (10 mM PIPES-KOH [pH 7.0], 5 mM MgCl2, 50 mM NaCl, We tentatively designated this band as the “specific DNA-induced 0.1 mM EGTA, 10% glycerol, 5 mM DTT, and 1 mg/ml BSA). Samples STING dimer” or “DI-STING.” were then analyzed by native PAGE essentially as described (40). In brief, a 7.5% polyacrylamide gel was prerun at 25 mA for 30 min with 25 mM Cyclic dinucleotide-induced dimerization of STING Downloaded from Tris-HCl buffer (pH 8.4) containing 192 mM glycine. The cathode 9 9 9 9 9 9 chamber contained 0.2% deoxycholate. The samples were mixed with 23 Cyclic dinucleotides such as 2 5 cGAMP, 3 -5 ,3 -5 -cGAMP native gel sample buffer (125 mM Tris-HCl [pH 6.8], 30% glycerol, 2% (3959cGAMP), and c-di-GMP induce STING dimerization, resulting deoxycholate, and 0.02% bromphenol blue), and separated by electro- in the induction of innate immunity (18, 44). In particular, phoresis at 25 mA for 90 min at 4˚C. After electrophoresis, the gel was 2959cGAMP functions as a second messenger in DNA-induced soaked for 30 min at room temperature in 25 mM Tris-HCl buffer (pH 8.3) containing 192 mM glycine and 0.1% SDS, and transferred to a polyvi- innate immunity (23, 44). To examine whether DI-STING is 2/2 nylidene difluoride membrane. STING-Flag was detected with anti–Flag- generated in response to dicyclic nucleotides, DNase II MEFs http://www.jimmunol.org/ M2-HRP mAb. were permeabilized with digitonin and then incubated at 37˚C for 30 min with increasing concentrations of 2959cGAMP, SDS-PAGE and Western blotting 3959cGAMP, or c-di-GMP. As shown in Fig. 2A, 2959cGAMP was The cell lysates were mixed with 53 SDS sample buffer (200 mM Tris-HCl ∼2-fold more effective than 3959cGAMP in inducing CXCL10 buffer [pH 6.8], 10% SDS, 25% glycerol, and 0.05% bromphenol blue) gene expression, and treatment with 10 nM 2959cGAMP was with 5% 2-ME, and heated at 95˚C for 5 min. For nondenaturing analysis, the cell lysates were mixed with 53 SDS buffer without 2-ME and in- sufficient to fully activate CXCL10 gene. In contrast, the treat- cubated at room temperature for 60 min. Samples were separated by ment of cells with concentrations of c-di-GMP as high as 10 mM electrophoresis on a 7.5% or 10% polyacrylamide gel (Biocraft, Tokyo, resulted in minimal activation of CXCL10 expression. Japan), transferred onto a polyvinylidene difluoride membrane, and ana- To determine whether the cyclic dinucleotides can induce by guest on September 28, 2021 lyzed by Western blotting with anti–Flag-M2-HRP mAb or anti-STING STING dimer formation, cell lysates were prepared from DNase mAb. II2/2 MEFs that had been permeabilized with digitonin and in- cubated with 1 mM2959cGAMP, 1 mM3959cGAMP, or 10 mM Results c-di-GMP. Analysis by SDS-PAGE under nonreducing conditions DNA-induced formation of a distinct STING dimer (Fig. 2B) or by native PAGE (Fig. 2C), followed by Western The introduction of DNA into cells activates IFN-b and CXCL10 blotting with anti-Flag mAb, indicated that DI-STING was formed gene expression by inducing STING dimerization (3). To confirm only in response to 2959cGAMP. DI-STING formation was also the involvement of STING dimer formation in DNA-induced acti- observed when the cell lysates from MEF transformants ex- 2 2 vation of cytokine gene expression, DNase II / MEFs, which are pressing Flag-STING were incubated in a cell-free system with more efficiently transfected by lipofection than wild-type MEFs 2959cGAMP, but not when the lysates were incubated with (41), were transformed with retrovirus carrying Flag-tagged mouse 3959cGAMP or c-di-GMP (Fig. 2D). These results suggested that STING (STING-Flag). The transformants expressed high levels of DI-STING formation in DNase II2/2 cells requires the interaction STING protein, as assessed by Western blotting (Fig. 1A), but did of STING with 2959cGAMP and that this dimer is conforma- not exhibit constitutive CXCL10 gene expression (Fig. 1B). How- tionally distinct from those formed in response to 3959cGAMP or ever, the lipofection-mediated introduction of plasmid DNA into the c-di-GMP. transformed cells resulted in a robust, dose-dependent upregulation The STING gene in DNase II2/2 cells was then mutated using of CXCL10 expression (Fig. 1B). the CRISPR/Cas system (35), to obtain a DNase II2/2STINGGT/GT The STING homodimer is detected by SDS-PAGE under non- cell line. The STING null mutation in DNase II2/2STINGGT/GT reducing conditions (42) or by native PAGE (43). To detect the cells completely blocked the CXCL10 gene expression induced by DNA-induced STING dimer, cell lysates were prepared from DNA- 2959cGAMP or 3959cGAMP, which could be rescued by the ectopic 2 2 transfected DNase II / MEFs expressing STING-Flag and ana- expression of the wild-type STING (Fig. 2E). A replacement of the lyzed by SDS-PAGE and native PAGE, followed by Western blot- arginine at amino acid position 232 of human STING with alanine ting with anti-Flag mAb. As shown in Fig. 1C, STING behaved as (R232A) inactivates the STING’s ability to induce the IFN-b gene a 37-kDa monomer when analyzed by SDS-PAGE under reducing in response to 2959cGAMP (44). Similarly, the corresponding mu- conditions, whereas analysis under nonreducing conditions revealed tant of mouse STING (R231A) hardly supported the CXCL10 gene that ∼70% of STING migrated as a 75-kDa protein, and the residual expression in response to 2959cGAMP or 3959cGAMP. Accordingly, monomeric form migrated slightly faster than under reducing the 2959cGAMP-induced DI-STING formation was observed with conditions. Notably, when a lower concentration of DNA was used the wild-type but not R231A mutant STING (Fig. 2F), confirming for lipofection, the intensity of the STING dimer band reduced dose that DI-STING is the active STING dimer generated with dependently, with a concomitant increase in the intensity of the 2959cGAMP. The Journal of Immunology 4917 Downloaded from http://www.jimmunol.org/

2 2 2 2 FIGURE 1. DNA-induced STING dimer formation. (A) Cell lysates from DNase II / MEFs (Parent) or DNase II / MEFs expressing STING-Flag by guest on September 28, 2021 (STING transformants) were heated at 95˚C for 5 min in SDS sample buffer containing 5% 2-ME. The samples were then analyzed by 10% SDS-PAGE, followed by Western blotting with anti-mouse STING (upper band, exogenous STING-flag; lower band, endogenous STING), anti-Flag mAb, or anti–a- tubulin. (B) DNase II2/2 MEFs or DNase II2/2 MEFs expressing STING-Flag were transfected with plasmid DNA at the indicated concentrations using Lipofectamine 2000 and incubated at 37˚C. Sixteen hours later, CXCL10 mRNA levels were determined by real-time RT-PCR, in triplicate. The data are relative expression values, compared with b-actin mRNA. Average values with the SD are shown. (C) At 12 h after the transfection, cell lysates (2 mg protein) were heated at 95˚C for 5 min in SDS sample buffer containing 5% 2-ME, or incubated at room temperature for 60 min (nonreducing) in SDS sample buffer without 2-ME. The samples were then analyzed by 10% SDS-PAGE, followed by Western blotting with anti-Flag mAb. (D) For analysis under nondenaturing conditions, the cell lysates (2 mg protein) in native gel sample buffer were separated by electrophoresis on a 7.5% native poly- acrylamide gel and analyzed by Western blotting with anti-Flag mAb. In (C) and (D), arrows indicate STING monomers and dimers. A specific DNA- induced STING dimer (DI-STING) is indicated by an arrowhead in (D).

Accumulation of 2959cGAMP in the fetal livers of DNase II2/2 To examine the possibility that the increased expression of IFN- mice b and CXCL10 genes in DNase II2/2 fetal livers was due to the 2 2 9 9 DNase II / mice accumulate undigested DNA in the lysosomes activation of STING by 2 5 cGAMP, nucleotides were extracted of fetal liver macrophages and die in utero, due to IFN-b pro- from the fetal livers with phenol and then separated by C18- duced by the macrophages (7). As shown in Fig. 3A, DNase II2/2 HPLC. The STING-dimerizing activity in each fraction was fetal livers expressed .10-fold higher levels of IFN-b and assayed in a cell-free system using purified recombinant Flag- CXCL10 mRNA than wild-type fetal livers. The induction of STING, followed by native PAGE and Western blotting with 2/2 IFN-b and CXCL10 gene expression in DNase II2/2 fetal livers anti-Flag mAb. As shown in Fig. 3B, the sample from DNase II , was minimally affected by the lack of IFN type I receptor gene but not wild-type fetal liver, showed STING-dimerizing activity in (IFN-IR) in the DNase II2/2 IFN-IR2/2 double-knockout mice. fraction 19, where the authentic 2959cGAMP eluted. By referring Like IFN-b gene promoter, the CXCL10 promoter carries IRF to a standard curve for STING dimerization activity in response to 2 2 and NF-kB binding sites, and seems to be directly activated by authentic 2959cGAMP (Fig. 3C), we estimated that the DNase II / DNA in the macrophages, as reported for its induction during fetal liver (∼17 mg) at E14.5 contained an average of 17 fmol viral infection (45). Both STING and cGAS mRNAs were 2959cGAMP (∼1.0 pmol/g) (Fig. 3D). expressed in the fetal livers of DNase II2/2 mice, supporting this Macrophages carrying undigested DNA are also present in the 2 2 idea. Notably, the expression of these genes, previously shown to spleens and thymi of DNase II / mice. When the nucleic acid 2 2 2 2 be regulated by IFN (46), was upregulated in DNase II2/2 fetal fractionsfromthespleensofDNaseII / IFN-IR / mice were livers and returned to normal in the fetal livers of DNase II2/2 subjected to HPLC analysis, fraction 19 exhibited STING- IFN-IR2/2 mice. dimerizing activity (Fig. 3E), suggesting that macrophages 4918 cGAS-DEPENDENT AND -INDEPENDENT INNATE IMMUNE RESPONSES Downloaded from http://www.jimmunol.org/

FIGURE 2. STING dimer formation mediated by dicyclic nucleotides. (A) The DNase II2/2 MEFs (5 3 104 cells) expressing STING-Flag were in- cubated with the indicated concentrations of 2959cGAMP, 3959cGAMP, or c-di-GMP for 30 min at 37˚C in digitonin permeabilization solution. The cells by guest on September 28, 2021 were then cultured in medium for 4 h, and the CXCL10 mRNA levels were determined in triplicate by real-time RT-PCR. The data are relative expression values, compared with b-actin mRNA. Average values with the SD (bars) are shown. (B and C) The DNase II2/2 MEFs (5 3 104 cells) were treated with 1 3 1026 M2959cGAMP or 3959cGAMP, or 1 3 1025 M c-diGMP for 30 min at 37˚C in digitonin permeabilization solution, followed by cell lysate preparation. In (B), the samples were placed at room temperature for 60 min in SDS sample buffer without 2-ME and analyzed by 10% SDS-PAGE, followed by Western blotting with anti-STING mAb. In (C), the samples in native gel sample buffer were analyzed by native PAGE, followed by Western blotting with anti-STING mAb. (D) Cell lysates (1 mg protein) from DNase II2/2 MEFs expressing STING-Flag were incubated for 30 min on ice with the indicated concentration of 2959cGAMP, 3959cGAMP, or c-di-GMP, and analyzed by native PAGE, followed by Western blotting with anti-Flag mAb. Arrows indicate STING monomers and dimers, and an arrowhead indicates DI-STING. (E) The DNase II2/2 MEFs (Parent), DNase II2/2 STINGGT/GT MEFs transformed by the vector (+Vec), the wild-type STING (+STING WT), or R231A mutant STING (+STING R231A) were incubated with 1.0 mM 2959cGAMP or 3959cGAMP for 30 min at 37˚C in digitonin permeabilization solution and cultured for 4 h. The CXCL10 mRNA levels were then de- termined by real-time RT-PCR in triplicate, and the average values were plotted. (F) DNase II2/2 STINGGT/GT MEFs expressing the wild-type STING-Flag (STING WT) or R231A mutant STING-Flag (STING R231A) were incubated without or with 1.0 mM2959cGAMP or 3959cGAMP for 30 min at 37˚C in digitonin permeabilization solution, followed by cell lysate preparation. The samples were analyzed by native PAGE, followed by Western blotting with anti-Flag mAb. An arrowhead indicates DI-STING. present in the spleens of DNase II2/2 mice also produce .70% of the MEFs carried engulfed apoptotic cells after a 2-h 2959cGAMP. coincubation (Fig. 4A). Hoechst-positive materials (DNA) were present in the cytoplasm of DNase II2/2 MEFs even 36 h after the CXCL10 expression by DNA from engulfed apoptotic cells or engulfment of apoptotic cells (Fig. 4B), confirming that little introduced by lipofection DNase was present in these cells. When lysosomal enzymes are In the fetal livers, spleens, and thymi, pyrenocytes and apoptotic addedtoculturemedium,theyoften gain access to lysosomes cells are engulfed by macrophages, and their DNA is degraded by (47). Indeed, when human rDNase II was added to the culture macrophage lysosomal DNase II. The presence of 2959cGAMP in medium during DNase II2/2 MEF apoptotic cell engulfment, the fetal livers and spleens of DNase II2/2 mice suggested that the Hoechst-positive materials did not accumulate in the DNase II2/2 undigested DNA of pyrenocytes or apoptotic cells activates the MEFs (Fig. 4B), confirming that the accumulation of DNA in cGAS–STING pathway. To examine this possibility, DNase II2/2 these cells was due to the lack of lysosomal DNase II. MEFs were transfected with expression constructs encoding TIM4 We next used these cells to confirm the effect of apoptotic and MER TK, cell surface receptors that are required for the ef- engulfment on cytokine expression in the DNase II2/2 MEFs. In ficient engulfment of apoptotic cells (32). Accordingly, the ability accordance with a previous report (14), the DNase II2/2 MEFs of DNase II2/2 MEFs to engulf apoptotic cells was strongly en- coincubated with apoptotic cells showed upregulated CXCL10 hanced in transformants expressing TIM4 and MER TK, and gene expression (Fig. 4C). This upregulation was suppressed by The Journal of Immunology 4919 Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021

FIGURE 3. The 2959cGAMP production in DNase II-null mice. (A) Upregulation of IFN-b,CXCL10,STING,andcGASgeneexpressioninDNase II2/2 fetal liver. RNAs were prepared from the fetal livers of E14.5 DNase II+/+IFNIR+/+, DNase II2/2IFNIR+/+, and DNase II2/2FNIR2/2 mouse embryos (n = 3). IFN-b, CXCL10, STING, and cGAS mRNA levels were determined by real-time RT-PCR and are expressed as relative values compared with that of b-actin mRNA. (B) Identification of the STING-dimerizing factor in DNase II2/2 fetal liver. Fetal livers from E14.5 DNase II+/2 IFN-IR2/2 and DNase II2/2 IFN-IR2/2 mice were extracted with phenol-based reagents, as described in Materials and Methods. The aqueous phase was loaded onto a C18 column, and the nucleotides were separated with a linear gradient of acetonitrile, followed by the analysis of each fraction for STING-dimerizing activity. The results of samples eluted at retention times of 15–22 min are shown. The analysis of an authentic 2959cGAMP sample using a C18 column and UV absorption for detection is shown to the right.(C) Quantitative STING dimerization assay for 2959cGAMP. The purified STING-Flag (20 fmol) was in- cubated with the indicated amounts of 2959cGAMP on ice for 30 min and analyzed by native PAGE, followed by Western blotting with anti-Flag mAb. In the right panel, the chemiluminescence of DI-STING (indicated by arrowhead) detected by the LAS4000 imager was quantified using the NIH Image program (National Institutes of Health) and plotted. (D) Referring to the standard curve in (C), the amount of 2959cGAMP in the fraction was determined based on the intensity of DI-STING and plotted. The average value is indicated by a horizontal bar. n.d., Not detected. (E) Analysis of the STING- dimerizing factor in DNase II2/2 spleens. Spleens from 8-wk-old DNase II+/2 IFN-IR2/2 (n = 1) and DNase II2/2 IFN-IR2/2 (n = 2) mice were extracted with phenol, and the aqueous phase was separated on a C18 column. The cGAMP fraction (fraction 19) was incubated with STING-Flag, followed by native PAGE and Western blotting with anti-Flag mAb. exogenously added DNase II, confirming that the DNA accumu- tation in DNase II2/2 cGASGT/GT cells had no effect on CXCL10 lating in the lysosomes was responsible for activating the CXCL10 gene induction (Fig. 4C). gene. As reported previously (28), the STING null mutation in To analyze the involvement of 2959cGAMP, the nucleic acid DNase II2/2 STINGGT/GT cells completely blocked the CXCL10 extracts were prepared from DNase II2/2 MEFs, apoptotic cells, gene expression induced by apoptotic cell engulfment, which was and DNase II2/2 MEFs coincubated with apoptotic cells, and rescued by the ectopic expression of STING (Fig. 4C). To deter- subjected to the STING dimerization assay. A weak STING- mine whether cGAS is involved in the CXCL10 gene induction in dimerizing activity, most likely 2959cGAMP, was detected in the DNase II2/2 MEFs, the cGAS gene in DNase II2/2 cells was extract prepared from DNase II2/2 MEFs engulfing apoptotic mutated using the CRISPR/Cas system, to obtain DNase II2/2 cells (Fig. 4D), but not in the extracts from apoptotic cells cGASGT/GT cell lines. In contrast to STING, the cGAS-null mu- (Supplemental Fig. 1). The null mutation of cGAS in DNase II2/2 4920 cGAS-DEPENDENT AND -INDEPENDENT INNATE IMMUNE RESPONSES Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021

FIGURE 4. cGAS-independent and STING-dependent CXCL10 gene expression induced by apoptotic cell engulfment. (A) DNase II2/2 MEFs (Parent) and DNase II2/2 MEFs expressing TIM4 and MER TK (TIM4/MER transformants) (5 3 104 cells) were incubated with pHrodo-labeled apoptotic thy- mocytes (2.5 3 106 cells) for 2 h and analyzed by flow cytometry. (B) Apoptotic CAD2/2 thymocytes (2.5 3 106 cells) and DNase II2/2 MEFs expressing TIM4 and MER TK (5 3 104 cells) were coincubated at 37˚C for 36 h, stained with Hoechst, and examined by fluorescence microscopy. Scale bar, 50 mm. (C–E) DNase II2/2cGAS+/+STING+/+ (WT), DNase II2/2cGAS+/+STINGGT/GT, DNase II2/2cGAS+/+STINGGT/GT-STING, and (Figure legend continues) The Journal of Immunology 4921 cGASGT/GT MEFs blocked its production (Fig. 4D). These results c-diGMP (44). Accordingly, tertiary structure analysis showed suggested that the accumulation of apoptotic cell DNA in lyso- that the two STING molecules in the 2959cGAMP–STING somes can activate cGAS to produce cGAMP, leading to STING complex are more tightly packed than the STING dimer in the activation and the induction of CXCL10 gene expression. How- c-diGMP–STING complex. ever, DNA from apoptotic cells also activates the CXCL10 gene In this work, we showed that the 2959cGAMP-mediated STING via a cGAS-independent pathway. dimer, but not the 3959cGAMP- or c-diGMP–mediated STING Next, we introduced naked DNA or dsRNA (polyI:C) into cells via dimer, is resistant to mild SDS treatment under nonreducing electroporation or lipofection. Electroporation delivers DNA or RNA conditions and behaves differently from other isomer-induced directly into cells, whereas lipofection delivers DNA into the cy- STING dimers on native PAGE. Although our analysis using the toplasm by endocytosis via lysosomes (41). As shown in Fig. 4E, full-length STING protein may not be directly comparable to the DNA or poly(I:C) introduced by either electroporation or lipofection structural analysis carried out with the C-terminal domain of led to strong activation of CXCL10 gene expression in DNase II2/2 STING, our results appear to support the model proposed by MEFs. CXCL10 gene expression induced by electroporated DNA Zhang et al. (44). Despite the different affinities of 2959cGAMP required cGAS, but CXCL10 gene expression induced by the lip- and 3959cGAMP for STING, the dose-response effect of these ofected DNA was cGAS-independent. The poly(I:C)-induced compounds on IFN-b gene induction was comparable, indicating CXCL10 gene induction required neither STING nor cGAS, con- that 2959cGAMP was only 2-fold more active than 3959cGAMP in firming the cGAS–STING pathway is specific for DNA (3). cells (44). Our finding that 2959cGAMP and 3959cGAMP exhibit In contrast from the CXCL10 gene induction, the DNA trans- similar effects on CXCL10 gene induction in permeabilized 2 2 fected by lipofection and electroporation methods resulted DNase II / NIH3T3 cells supports this previous finding. Al- Downloaded from in a high level of the STING-dimerizing activity, which was though the different STING dimers formed by 2959cGAMP and cGAS-dependent (Fig. 4D). These results indicated that cells 3959cGAMP may impact downstream signals similarly, it is pos- use cGAS-dependent and -independent pathways to activate sible that 2959cGAMP is quickly degraded in the cells by a specific STING-dependent CXCL10 gene expression in response to 2959 phosphodiesterase, as reported for similar compounds, such DNA. To confirm the DNA-induced cGAS-dependent and as the 2959 oligoadenylates (48). This potential for increased

-independent innate immune activation, the wild-type and turnover of 2959cGAMP, compared with 3959cGAMP, could ac- http://www.jimmunol.org/ cGASGT/GT MEFs were transformed with STING and trans- count for the disconnect between their differential affinities for fected with DNA. In the downstream of STING, TBK1 is ac- STING and their similar cellular efficacies. In this regard, it is also tivated or phosphorylated for the IFN-b gene expression (48). noteworthy that the concentrations of 2959cGAMP required for When DNA was introduced into MEFs, either by electro- immune cell activation appear to be quite different (ranging from poration or lipofection, TBK1 was phosphorylated in the wild- nM to mM) among various cell lines (23, 44). type, but not STING-null MEFs (STINGGT/GT MEFs). This To date, 2959cGAMP has been detected by time-consuming DNA-induced phosphorylation of TBK1 was strongly enhanced procedures such as tandem mass spectrometry coupled with when the wild-type or STINGGT/GT MEFs were transformed liquid chromatography–mass spectrometry analysis (44, 52). In with STING (Fig. 4F). In contrast, overexpression of STING this work, we developed a sensitive assay method to detect by guest on September 28, 2021 enhanced the phosphorylation of TBK1 in cGAS-null MEFs 2959cGAMP in cells, based on its ability to induce STING di- (cGASGT/GT MEFs) only when DNA was introduced via lip- merization. Nucleic acid samples containing 2959cGAMP were ofection, but not via electroporation. These results confirmed prepared from cells or tissues by using phenol, followed by that the DNA introduced via electroporation requires cGAS to HPLC fractionation. The individual fractions were then incu- activate STNG–TBK1 pathway, whereas the DNA introduced bated with Flag-STING, followed by analysis using native PAGE by lipofection has the ability to activate this pathway in a and Western blotting. The sensitivity of the assay was 1.5 fmol cGAS-independent manner. 2959cGAMP, which was 100-fold more sensitive than the method using liquid chromatography–mass spectrometry that requires Discussion a tandem mass spectrometry fragmentation at positive mode. The role of the cGAS–STING signaling axis in DNA sensing This method is specific for 2959cGAMP because DI-STING was and immune activation is well established (3, 49), although detected only in response to 2959cGAMP, but not other isomers. there are some conflicting findings on the mechanistic details Using this system, we detected 1-2 pmol 2959cGAMP in 2.5 3 that remain to be clarified. For example, several groups re- 105 cells, in which DNA was introduced by electroporation or ported that 2959cGAMP and its linkage isomers bind STING lipofection. Approximately 17 and 120 fmol 2959cGAMP were with similar affinities and induce its dimerization, without detected in the fetal liver (∼17 mg) and spleen (∼250 mg) of causing obvious conformational changes in STING (23, 50, DNase II2/2 mice. This method may be used to detect 2959cGAMP 51). In contrast, another group reported that 2959cGAMP binds and to monitor intracellular DNA in virus-infected or damaged to STING with an affinity that is 200-fold higher than that of tissues.

DNase II2/2cGASGT/GTSTING+/+ MEFs (2.5 3 105 cells) expressing TIM4 and MER TK were incubated with 1.25 3 107 CAD2/2 apoptotic thymocytes for 36 h (C and D), transfected with plasmid DNA by electroporation (EP; 0.15 mg/ml) or lipofection (LF; 0.1 mg/ml), or transfected with poly(I:C) (RNA), followed by incubation at 37˚C for 12 h (D and E). In (C) and (E), CXCL10 mRNA was quantified by real-time RT-PCR in triplicate, and the average values, expressed as relative values compared with b-actin mRNA, are shown with the SD (bar). The data were statistically analyzed using one-way ANOVA, followed by Tukey’s post hoc test (C), and using Student t test (E), respectively. *Denotes p , 0.05, and ns denotes p . 0.05. In (D), the cells were extracted with phenol reagents, and the aqueous phase was analyzed for STING-dimerizing activity using purified STING-Flag. An arrowhead indicates DI-STING. (F) DNase II2/2cGAS+/+STING+/+ (WT), DNase II2/2cGASGT/GTSTING+/+, and DNase II2/2cGAS+/+STINGGT/GT MEFs expressing TIM4 and MER TK or their STING transformants (2.5 3 105 cells) were transfected with plasmid DNA by electroporation (EP) or lipofection (LF), and incubated for 12 h. The cell lysates (10 mg) were separated by 10% SDS-PAGE and analyzed by Western blotting with anti-phospho TBK1, anti-TBK1, anti-STING, or anti–a- Tubulin. 4922 cGAS-DEPENDENT AND -INDEPENDENT INNATE IMMUNE RESPONSES

Several groups have reported that the DNA transfected into cells 16. Ishikawa, H., and G. N. Barber. 2008. STING is an endoplasmic reticulum adaptor that facilitates innate immune signalling. Nature 455: 674–678. by lipofection activates cGAS, thereby triggering cytokine gene 17. Woodward, J. J., A. T. Iavarone, and D. A. Portnoy. 2010. c-di-AMP secreted by expression (44, 52, 53). Accordingly, the transfection of DNA by intracellular Listeria monocytogenes activates a host type I interferon response. lipofection as well as apoptotic cell engulfment by DNase II2/2 Science 328: 1703–1705. 9 9 18. Burdette, D. L., K. M. Monroe, K. Sotelo-Troha, J. S. Iwig, B. Eckert, MEFs produced 2 5 cGAMP in a cGAS-dependent manner, sup- M. Hyodo, Y. Hayakawa, and R. E. Vance. 2011. STING is a direct innate porting the idea that DNA accumulating in lysosomes leaks into immune sensor of cyclic di-GMP. Nature 478: 515–518. the cytoplasm to activate cGAS (14). However, in contrast to 19. Sun, L., J. Wu, F. Du, X. Chen, and Z. J. Chen. 2013. Cyclic GMP-AMP syn- thase is a cytosolic DNA sensor that activates the type I interferon pathway. previous findings with wild-type fibroblasts (44), the introduction Science 339: 786–791. 2 2 of the cGAS-null mutation into DNase II / MEFs did not ab- 20. Wu, J., L. Sun, X. Chen, F. Du, H. Shi, C. Chen, and Z. J. Chen. 2013. Cyclic rogate the CXCL10 gene expression induced by DNA from apo- GMP-AMP is an endogenous second messenger in innate immune signaling by cytosolic DNA. Science 339: 826–830. ptotic cells or introduced by lipofection. This could be explained if 21. Diner, E. J., D. L. Burdette, S. C. Wilson, K. M. Monroe, C. A. Kellenberger, the lysosomal DNA sensors in wild-type cells were not activated M. Hyodo, Y. Hayakawa, M. C. Hammond, and R. E. Vance. 2013. The innate because the DNA is rapidly digested by DNase II, so that only the immune DNA sensor cGAS produces a noncanonical cyclic dinucleotide that activates human STING. Cell Reports 3: 1355–1361. DNA that escaped lysosomal degradation by leaking into the cy- 22. Kranzusch, P. J., A. S.-Y. Lee, J. M. Berger, and J. A. Doudna. 2013. Structure of toplasm was recognized, by cGAS. In contrast, because DNA human cGAS reveals a conserved family of second-messenger enzymes in innate 2/2 immunity. Cell Reports 3: 1362–1368. cannot be digested in the lysosomes in DNase II cells, lyso- 23. Gao, P., M. Ascano, T. Zillinger, W. Wang, P. Dai, A. A. Serganov, somal DNA sensors may have recognized the undegraded DNA, B. L. Gaffney, S. Shuman, R. A. Jones, L. Deng, et al. 2013. Structure-function resulting in the downstream activation of CXCL10. An obvious analysis of STING activation by c[G(29,59)pA(39,59)p] and targeting by antiviral DMXAA. Cell 154: 748–762. candidate for such a lysosomal DNA sensor, TLR9, is not Downloaded from 2 2 24. Ishikawa, H., Z. Ma, and G. N. Barber. 2009. STING regulates intracellular expressed in DNase II / MEFs (Supplemental Fig. 2). The gene DNA-mediated, type I interferon-dependent innate immunity. Nature 461: 788– induction in DNase II2/2 MEFs was STING-dependent, sug- 792. 25. Gall, A., P. Treuting, K. B. Elkon, Y.-M. Loo, M. Gale, Jr., G. N. Barber, and gesting that an unknown lysosomal DNA sensor that triggers D. B. Stetson. 2012. Autoimmunity initiates in nonhematopoietic cells and STING activation through a cGAS-independent mechanism may progresses via lymphocytes in an interferon-dependent autoimmune disease. exist in DNase II2/2 MEFs. Immunity 36: 120–131. 26. Gao, D., J. Wu, Y.-T. Wu, F. Du, C. Aroh, N. Yan, L. Sun, and Z. J. Chen. 2013.

Some groups of pathogens are known to infect cells via phag- Cyclic GMP-AMP synthase is an innate immune sensor of HIV and other ret- http://www.jimmunol.org/ osomes and activate innate immunity via a TLR9-independent roviruses. Science 341: 903–906. 27. Li, X. D., J. Wu, D. Gao, H. Wang, L. Sun, and Z. J. Chen. 2013. Pivotal roles of mechanism (54–57). It will be interesting to determine whether cGAS-cGAMP signaling in antiviral defense and immune adjuvant effects. the host cells of these pathogens use a cGAS-dependent or -in- Science 341: 1390–1394. dependent mechanism to activate innate immunity in response. 28. Ahn, J., D. Gutman, S. Saijo, and G. N. Barber. 2012. STING manifests self DNA-dependent inflammatory disease. Proc. Natl. Acad. Sci. USA 109: 19386– 19391. 29. Morita, S., T. Kojima, and T. Kitamura. 2000. Plat-E: an efficient and stable Acknowledgments system for transient packaging of retroviruses. Gene Ther. 7: 1063–1066. We thank M. Fujii for secretarial assistance. 30. Kitamura, T., Y. Koshino, F. Shibata, T. Oki, H. Nakajima, T. Nosaka, and H. Kumagai. 2003. Retrovirus-mediated gene transfer and expression cloning: powerful tools in functional genomics. Exp. Hematol. 31: 1007–1014. by guest on September 28, 2021 Disclosures 31. Miyanishi, M., K. Tada, M. Koike, Y. Uchiyama, T. Kitamura, and S. Nagata. 2007. Identification of Tim4 as a phosphatidylserine receptor. Nature 450: 435– The authors have no financial conflicts of interest. 439. 32. Nishi, C., S. Toda, K. Segawa, and S. Nagata. 2014. Tim4- and MerTK-mediated engulfment of apoptotic cells by mouse resident peritoneal macrophages. Mol. References Cell. Biol. 34: 1512–1520. 1. Kawai, T., and S. Akira. 2011. Toll-like receptors and their crosstalk with other 33. Shiraishi, T., K. Suzuyama, H. Okamoto, T. Mineta, K. Tabuchi, K. Nakayama, innate receptors in infection and immunity. Immunity 34: 637–650. Y. Shimizu, J. Tohma, T. Ogihara, H. Naba, et al. 2004. Increased cytotoxicity of 2. Desmet, C. J., and K. J. Ishii. 2012. Nucleic acid sensing at the interface between soluble Fas ligand by fusing isoleucine zipper motif. Biochem. Biophys. Res. innate and adaptive immunity in vaccination. Nat. Rev. Immunol. 12: 479–491. Commun. 322: 197–202. 3. Wu, J., and Z. J. Chen. 2014. Innate immune sensing and signaling of cytosolic 34. Mizushima, S., and S. Nagata. 1990. pEF-BOS, a powerful mammalian ex- nucleic acids. Annu. Rev. Immunol. 32: 461–488. pression vector. Nucleic Acids Res. 18: 5322. 4. Ablasser, A., C. Hertrich, R. Waßermann, and V. Hornung. 2013. Nucleic acid 35. Cong, L., F. A. Ran, D. Cox, S. Lin, R. Barretto, N. Habib, P. D. Hsu, X. Wu, driven sterile inflammation. Clin. Immunol. 147: 207–215. W. Jiang, L. A. Marraffini, and F. Zhang. 2013. Multiplex genome engineering 5. Nagata, S., R. Hanayama, and K. Kawane. 2010. Autoimmunity and the clear- using CRISPR/Cas systems. Science 339: 819–823. ance of dead cells. Cell 140: 619–630. 36. Miksa, M., H. Komura, R. Wu, K. G. Shah, and P. Wang. 2009. A novel method 6. Chasis, J. A., and N. Mohandas. 2008. Erythroblastic islands: niches for eryth- to determine the engulfment of apoptotic cells by macrophages using pHrodo ropoiesis. Blood 112: 470–478. succinimidyl ester. J. Immunol. Methods 342: 71–77. 7. Yoshida, H., Y. Okabe, K. Kawane, H. Fukuyama, and S. Nagata. 2005. Lethal 37. Toda, S., R. Hanayama, and S. Nagata. 2012. Two-step engulfment of apoptotic anemia caused by interferon-beta produced in mouse embryos carrying undi- cells. Mol. Cell. Biol. 32: 118–125. gested DNA. Nat. Immunol. 6: 49–56. 38. von Ko¨ckritz-Blickwede, M., O. A. Chow, and V. Nizet. 2009. Fetal calf serum 8. Kawane, K., H. Fukuyama, G. Kondoh, J. Takeda, Y. Ohsawa, Y. Uchiyama, and contains heat-stable nucleases that degrade neutrophil extracellular traps. Blood S. Nagata. 2001. Requirement of DNase II for definitive erythropoiesis in the 114: 5245–5246. mouse fetal liver. Science 292: 1546–1549. 39. Chida, J., and H. Kido. 2014. Extraction and quantification of adenosine 9. Kawane, K., M. Ohtani, K. Miwa, T. Kizawa, Y. Kanbara, Y. Yoshioka, triphosphate in mammalian tissues and cells. Methods Mol. Biol. 1098: 21– H. Yoshikawa, and S. Nagata. 2006. Chronic polyarthritis caused by mammalian 32. DNA that escapes from degradation in macrophages. Nature 443: 998–1002. 40. Iwamura, T., M. Yoneyama, N. Koizumi, Y. Okabe, H. Namiki, C. E. Samuel, 10. Kawane, K., H. Tanaka, Y. Kitahara, S. Shimaoka, and S. Nagata. 2010. and T. Fujita. 2001. PACT, a double-stranded RNA binding protein acts as Cytokine-dependent but acquired immunity-independent arthritis caused by a positive regulator for type I interferon gene induced by Newcastle disease DNA escaped from degradation. Proc. Natl. Acad. Sci. USA 107: 19432–19437. virus. Biochem. Biophys. Res. Commun. 282: 515–523. 11. Stetson, D. B., J. S. Ko, T. Heidmann, and R. Medzhitov. 2008. Trex1 prevents 41. Howell, D. P.-G., R. J. Krieser, A. Eastman, and M. A. Barry. 2003. Deoxyri- cell-intrinsic initiation of autoimmunity. Cell 134: 587–598. bonuclease II is a lysosomal barrier to transfection. Mol. Ther. 8: 957–963. 12. Barber, G. N. 2011. Cytoplasmic DNA innate immune pathways. Immunol. Rev. 42. Jin, L., K. K. Hill, H. Filak, J. Mogan, H. Knowles, B. Zhang, A.-L. Perraud, 243: 99–108. J. C. Cambier, and L. L. Lenz. 2011. MPYS is required for IFN response factor 3 13. Kawai, T., and S. Akira. 2010. The role of pattern-recognition receptors in innate activation and type I IFN production in the response of cultured phagocytes to immunity: update on Toll-like receptors. Nat. Immunol. 11: 373–384. bacterial second messengers cyclic-di-AMP and cyclic-di-GMP. J. Immunol. 14. Okabe, Y., K. Kawane, S. Akira, T. Taniguchi, and S. Nagata. 2005. Toll-like 187: 2595–2601. receptor-independent gene induction program activated by mammalian DNA 43. Sun, W., Y. Li, L. Chen, H. Chen, F. You, X. Zhou, Y. Zhou, Z. Zhai, D. Chen, escaped from apoptotic DNA degradation. J. Exp. Med. 202: 1333–1339. and Z. Jiang. 2009. ERIS, an endoplasmic reticulum IFN stimulator, activates 15. Stetson, D. B., and R. Medzhitov. 2006. Recognition of cytosolic DNA activates innate immune signaling through dimerization. Proc. Natl. Acad. Sci. USA 106: an IRF3-dependent innate immune response. Immunity 24: 93–103. 8653–8658. The Journal of Immunology 4923

44. Zhang, X., H. Shi, J. Wu, X. Zhang, L. Sun, C. Chen, and Z. J. Chen. 2013. 52. Ablasser, A., M. Goldeck, T. Cavlar, T. Deimling, G. Witte, I. Ro¨hl, K.-P. Hopfner, Cyclic GMP-AMP containing mixed phosphodiester linkages is an endogenous J. Ludwig, and V. Hornung. 2013. cGAS produces a 29-59-linked cyclic dinucle- high-affinity ligand for STING. Mol. Cell 51: 226–235. otide second messenger that activates STING. Nature 498: 380–384. 45. Brownell, J., J. Bruckner, J. Wagoner, E. Thomas, Y. M. Loo, M. Gale, Jr., 53. Civril, F., T. Deimling, C. C. de Oliveira Mann, A. Ablasser, M. Moldt, G. Witte, T. J. Liang, and S. J. Polyak. 2014. Direct, interferon-independent activation of V. Hornung, and K.-P. Hopfner. 2013. Structural mechanism of cytosolic DNA the CXCL10 promoter by NF-kB and interferon regulatory factor 3 during sensing by cGAS. Nature 498: 332–337. hepatitis C virus infection. J. Virol. 88: 1582–1590. 54. Chessler, A. D. C., L. R. P. Ferreira, T. H. Chang, K. A. Fitzgerald, and 46. Schoggins, J. W., S. J. Wilson, M. Panis, M. Y. Murphy, C. T. Jones, P. Bieniasz, B. A. Burleigh. 2008. A novel IFN regulatory factor 3-dependent pathway ac- and C. M. Rice. 2011. A diverse range of gene products are effectors of the type I tivated by trypanosomes triggers IFN-beta in macrophages and fibroblasts. interferon antiviral response. Nature 472: 481–485. J. Immunol. 181: 7917–7924. 47. Desnick, R. J., and E. H. Schuchman. 2012. Enzyme replacement therapy for 55. Koppe, U., K. Ho¨gner, J. M. Doehn, H. C. Muller,€ M. Witzenrath, B. Gutbier, lysosomal diseases: lessons from 20 years of experience and remaining chal- S. Bauer, T. Pribyl, S. Hammerschmidt, J. Lohmeyer, et al. 2012. Streptococcus lenges. Annu. Rev. Genomics Hum. Genet. 13: 307–335. pneumoniae stimulates a STING- and IFN regulatory factor 3-dependent type I 48. Hornung, V., R. Hartmann, A. Ablasser, and K.-P. Hopfner. 2014. OAS proteins IFN production in macrophages, which regulates RANTES production in mac- and cGAS: unifying concepts in sensing and responding to cytosolic nucleic rophages, cocultured alveolar epithelial cells, and mouse lungs. J. Immunol. 188: acids. Nat. Rev. Immunol. 14: 521–528. 811–817. 49. Ablasser, A., and V. Hornung. 2013. DNA sensing unchained. Cell Res. 23: 585–587. 56. Lippmann, J., H. C. Muller,€ J. Naujoks, C. Tabeling, S. Shin, M. Witzenrath, 50. Ouyang, S., X. Song, Y. Wang, H. Ru, N. Shaw, Y. Jiang, F. Niu, Y. Zhu, W. Qiu, K. Hellwig, C. J. Kirschning, G. A. Taylor, W. Barchet, et al. 2011. Dissection of K. Parvatiyar, et al. 2012. Structural analysis of the STING adaptor protein a type I interferon pathway in controlling bacterial intracellular infection in reveals a hydrophobic dimer interface and mode of cyclic di-GMP binding. mice. Cell. Microbiol. 13: 1668–1682. Immunity 36: 1073–1086. 57. Sharma, S., R. B. DeOliveira, P. Kalantari, P. Parroche, N. Goutagny, Z. Jiang, 51. Shu, C., G. Yi, T. Watts, C. C. Kao, and P. Li. 2012. Structure of STING bound to J. Chan, D. C. Bartholomeu, F. Lauw, J. P. Hall, et al. 2011. Innate immune cyclic di-GMP reveals the mechanism of cyclic dinucleotide recognition by the recognition of an AT-rich stem-loop DNA motif in the Plasmodium falciparum immune system. Nat. Struct. Mol. Biol. 19: 722–724. genome. Immunity 35: 194–207. Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021 Supplementary Figures

Figure S1. The STING-dimerizing activity in DNase II-/- MEFs engulfing apoptotic cells. Apoptotic CAD-/- thymocytes (1.25×107 cells) and DNase II-/- MEFs expressing TIM4 and MER TK (2.5×105 cells) were co-incubated at 37°C for 36 h. The cells were extracted with phenol-reagents, and the aqueous phase was analyzed for STING dimerizing activity using the cell lysates containing STING-Flag. The nucleic acid extracts were also prepared from apoptotic thymocytes and DNase II-/- MEF without co-incubating with apoptotic cells, and subjected to the STING-dimerizing assay. An arrowhead indicates DI-STING.

Figure S2. No expression of TLR9 mRNA in MEFs. RNA (3 ng) from cGASGT/GT MEF and mouse fetal liver was analyzed by RT-PCR for Tlr9 and β-actin mRNAs, and separated by electrophoresis on 2 % agarose gel.