2394.full.pdf - The Journal of Immunology

of August 31, 2022.This information is current as

InfectionChlamydia trachomatisduring βSynthase, Is Essential for Induction of IFN-

AMP−The DNA Sensor, Cyclic GMP

M. NagarajanPrantner, Priscilla B. Wyrick, Vladimir Lupashin and Uma Yugen Zhang, Laxmi Yeruva, Anthony Marinov, Daniel

http://www.jimmunol.org/content/193/5/2394doi: 10.4049/jimmunol.1302718July 2014;

2014; 193:2394-2404; Prepublished online 28J Immunol 

MaterialSupplementary

8.DCSupplementalhttp://www.jimmunol.org/content/suppl/2014/07/26/jimmunol.130271

Referenceshttp://www.jimmunol.org/content/193/5/2394.full#ref-list-1

, 19 of which you can access for free at: cites 48 articlesThis article

        average*  

4 weeks from acceptance to publicationFast Publication! •    

Every submission reviewed by practicing scientistsNo Triage! •    

from submission to initial decisionRapid Reviews! 30 days* •    

Submit online. ?The JIWhy

Subscriptionhttp://jimmunol.org/subscription

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/About/Publications/JI/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/alertsReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists, Inc. All rights reserved.Copyright © 2014 by The American Association of1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on August 31, 2022

http://ww

w.jim

munol.org/

Dow

nloaded from

by guest on August 31, 2022

http://ww

w.jim

munol.org/

Dow

nloaded from

The Journal of Immunology

The DNA Sensor, Cyclic GMP–AMP Synthase, Is Essential forInduction of IFN-b during Chlamydia trachomatis Infection

Yugen Zhang,* Laxmi Yeruva,† Anthony Marinov,‡ Daniel Prantner,x Priscilla B. Wyrick,*

Vladimir Lupashin,{ and Uma M. Nagarajan*

IFN-b has been implicated as an effector of oviduct pathology resulting from genital chlamydial infection in the mouse model. In this

study, we investigated the role of cytosolic DNA and engagement of DNA sensors in IFN-b expression during chlamydial infection.

We determined that three-prime repair exonuclease-1, a host 39 to 59 exonuclease, reduced IFN-b expression significantly during

chlamydial infection using small interfering RNA and gene knockout fibroblasts, implicating cytosolic DNA as a ligand for this

response. The DNA sensor cyclic GMP–AMP synthase (cGAS) has been shown to bind cytosolic DNA to generate cyclic GMP–AMP,

which binds to the signaling adaptor stimulator of IFN genes (STING) to induce IFN-b expression. We determined that cGAS is

required for IFN-b expression during chlamydial infection in multiple cell types. Interestingly, although infected cells deficient for

STING or cGAS alone failed to induce IFN-b, coculture of cells depleted for either STING or cGAS rescued IFN-b expression. These

data demonstrate that cyclic GMP–AMP produced in infected cGAS+STING2 cells can migrate into adjacent cells via gap junctions

to function in trans in cGAS2STING+ cells. Furthermore, we observed cGAS localized in punctate regions on the cytosolic side of the

chlamydial inclusion membrane in association with STING, indicating that chlamydial DNA is most likely recognized outside the

inclusion as infection progresses. These novel findings provide evidence that cGAS-mediated DNA sensing directs IFN-b expression

during Chlamydia trachomatis infection and suggest that effectors from infected cells can directly upregulate IFN-b expression in

adjacent uninfected cells during in vivo infection, contributing to pathogenesis. The Journal of Immunology, 2014, 193: 2394–2404.

Chlamydia trachomatis is the most common sexuallytransmitted bacterial pathogen in the world, and infectioncan lead to pelvic inflammatory disease and infertility in

women. Chlamydial infection of epithelial cells upregulates proin-flammatory cytokines, chemokines, type I IFNs, and IFN stimula-tory genes (1–3). We and others have shown that type I IFN (IFN-aand IFN-b) signaling exacerbates host pathology during the courseof genital (4) or pulmonary (5) Chlamydia muridarum infection inthe mouse model. Furthermore, IFN-b depletion protects mice fromoviduct pathology during genital chlamydial infection (6), demon-strating a significant contribution of IFN-b to host pathology. Asimilar detrimental effect of IFN-b signaling has been reportedduring other bacterial infections as well (reviewed in Ref. 7).

A consensus mechanism for IFN-b induction during intracellularbacterial infection is yet to be defined. However, multiple host

pathogen recognition receptors that can induce IFN-b expression

during viral infection (reviewed in Ref. 8) have been identified.

These include the RNA sensors, retinoic acid–inducible gene I and

melanoma differentiation-associated protein 5 (9, 10), which rec-

ognize viral RNA and signal via the adaptor mitochondrial antiviral

signaling to induce IFN-b expression (11). In addition, several

DNA sensors have been identified that recognize cytosolic DNA

and induce IFN-b expression. These include RNA polymerase III

(9, 10), DNA-dependent activator of IFN regulatory factors (12),

IFN-g–inducible protein 16 (13), leucine-rich repeat protein FLII-

interacting protein (14), DEAD box polypeptide 41 (DDX41) (15),

meiotic recombination 11 homolog (16), LSm14A (member of

LSm protein family) (17), and DNA-protein kinase catalytic sub-

unit (18). The large number of DNA sensors identified in the host

suggests that they may play redundant roles during infection. In

contrast, stimulator of IFN genes (STING), an endoplasmic retic-

ulum (ER)–resident transmembrane protein, has been reported to

be crucial for independent recognition of cytosolic DNA during

viral infection and induction of IFN-b (19). STING is not a di-

rect sensor of DNA, but functions as an integral adaptor molecule

in DNA recognition. STING binds to a novel second-messenger,

cyclic GMP–AMP (cGAMP), generated by a host DNA sensor,

cGAMP synthase (cGAS) (20), upon DNA binding in the cytosol

(21). This interaction of cGAMP with STING activates the sig-

naling events that lead to IFN-b expression. Additionally, STING

also directly binds bacterial second messengers, cyclic di-GMP and

di-AMP, to induce IFN-b (22), suggesting that it can also function

as a direct sensor of intracellular pathogens. Indeed, cyclic di-AMP

has been shown to be produced by Listeria monocytogenes (23)

and C. trachomatis (24). However, the direct contribution of

STING relative to its cooperation with DNA sensors in IFN-b

expression during bacterial infection remains unclear.

*Department of Pediatrics, University of North Carolina at Chapel Hill, Chapel Hill,NC 27599; †Department of Pediatrics, University of Arkansas for Medical Sciences,Little Rock, AR 72202; ‡Department of Immunology, University of Pittsburgh, Pitts-burgh, PA 15261; xDepartment of Microbiology and Immunology, University ofMaryland, Baltimore, MD 21201; and {Department of Physiology and Biophysics,University of Arkansas for Medical Sciences, Little Rock, AR 72205

Received for publication October 7, 2013. Accepted for publication June 20, 2014.

This work was supported largely by Public Health Service (National Institute ofAllergy and Infectious Diseases) Grant AI067678 (to U.M.N.) and partly by GrantGM083144 (to V.L.).

Address correspondence and reprint requests to Dr. Uma M. Nagarajan, Departmentof Pediatrics, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599.E-mail address: [email protected]

The online version of this article contains supplemental material.

Abbreviations used in this article: AF, Alexa Fluor; cGAMP, cyclic GMP–AMP;cGAS, cGAMP synthase; DDX41, DEAD box polypeptide 41; EB, elementary body;ER, endoplasmic reticulum; IRF3, IFN regulatory factor 3; ISD, immunostimulatoryDNA; KD, knock down; KO, knockout; MEF, mouse embryonic fibroblast; MOI,multiplicity of infection; NT, nontargeting; p.i, postinfection; poly I:C, polyinosinic-polycytidylic acid; qRT-PCR, quantitative RT-PCR; RB, reticulate body; siRNA,small interfering RNA; STING, stimulator of IFN gene; TREX-1, three-prime repairexonuclease-1; T3SS, type III secretion system; WT, wild-type.

Copyright� 2014 by TheAmericanAssociation of Immunologists, Inc. 0022-1767/14/$16.00

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1302718

by guest on August 31, 2022

http://ww

w.jim

munol.org/

Dow

nloaded from

We have shown that STING is required for IFN-b induction duringchlamydial infection in HeLa cells and murine oviduct epithelialcells, whereas the cytosolic RNA sensing pathway is dispensable forthis response (25). In this study, to our knowledge, we demonstratefor the first time that cytosolic DNA is a trigger for IFN-b expressionduring C. trachomatis infection and that the DNA sensor cGASplays an integral role in sensing this DNA to induce IFN-b expres-sion. cGAS localized in close proximity to the chlamydial inclu-sion membrane and colocalized with STING, suggesting thatchlamydial DNA is most likely recognized on membrane com-partments outside the chlamydial inclusion. We also provide in-direct evidence for cGAS-mediated generation of cGAMP duringinfection, by demonstrating rescue of IFN-b expression duringcoculture of cGAS+STING2 cells with cGAS2STING+ cells,suggesting that cGAMP from infected cGAS+ cells can migrateinto adjacent cells to induce STING-dependent IFN-b expression.

Materials and MethodsCell culture, primary cells, reagents, cDNA constructs,chlamydial stocks, and infection

HeLa cells, wild-type (WT) and three-prime repair exonuclease-1 (TREX-1) knockout (KO) mouse embryonic fibroblasts (MEFs), and HEK293Twere cultured in supplemented DMEM, and mouse J774 cells were culturedin complete media for macrophages, as described earlier (25). MouseBM1.11 cells (26) and human OE-E6/E7 (oviduct epithelial cells) (27)were cultured in supplemented F12-DMEM, as described (25). Poly dA:dT/LyoVec (100 mg/ml), polyinosinic-polycytidylic acid (poly I:C)/LyoVec(50 mg/ml), and 2939 cGAMP (1 mg/ml) were purchased from Invivogen.Carbenoxolone (working concentration 0.2 mM) was purchased fromSigma-Aldrich. pBluscript vector (1–2.5 mg) was used as immunostimu-latory DNA (ISD). cDNA constructs for human cGAS (pCMV-cGAS) andSTING (pCMV-STING) were purchased from Origene. C. muridarum,C. trachomatis D and L2 were propagated in McCoy cells, and infectionswere performed at 1 multiplicity of infection (MOI) or as indicated, aspreviously described (25).

Small interfering RNA

Small interfering RNA (siRNA) targeting human cGAS (MB21D1) (s41746and s453378), human DDX41 (s28120), human TRIM56 (s37816), humanLSM14A (s25051), human STING (s50645), or corresponding nontargeting(NT; control, catalogue 4390843) were obtained from Ambion Life Tech-nologies and were used for HeLa and OE cells. For BM1.11 cells, AccellSMART pool siRNA duplexes targeting mouse STING (E-055528), mousecGAS (E-055608), and corresponding Accell NT siRNA (D-001910-10)from Dharmacon were used. For J774 cells, two mouse cGAS (MB21D1)siRNA oligos (catalogue 2675; Sigma-Aldrich), described earlier (20), wereused in parallel with the corresponding NT siRNA from Sigma-Aldrich(catalogue SIC001). For TREX-1 knockdown siGenome SMART pool mouseTREX-1 (M-042223-00) were used in BM1.11 cells, whereas siGenomeSMART pool human TREX-1 (M-013239-02) were used in HeLa cellswith corresponding NT pools.

siRNA, ISD transfection, and cGAMP treatment

HeLa cells were plated at 1 3 105 cells/well in 24-well plates for 18–24 hbefore transfection. A total of 30 pmol siRNA was transfected usingLipofectamine RNAiMAX reagent (Invitrogen). Forty-eight hours post-transfection, siRNA-transfected cells were split into 4 wells in a 24-welldish for multiple treatments. siRNA (10 pmol) transfections were repeatedon each well the next day to achieve maximal knockdown in expression.Twenty-four hours after second siRNA transfection, one set of cells wasinfected with Chlamydia and harvested 18–24 h postinfection (p.i). Inparallel, cells were transfected with ISD (positive control) or with poly I:C(negative control), 6–8 h prior to harvest, so that all cells, including un-treated controls, were harvested at the same time. BM1.11 cells weretransfected directly using Accell siRNA (100 pmol/well) in a 24-well dishfor multiple treatments. OE-E6/E7 cells and J774 cells were transfectedwith siRNA (100 pmol/well and 30 mM, respectively) in a 24-well dish formultiple treatments using INTERFERin (Polyplus transfection). Seventy-two hours after siRNA transfection, cells were infected with C. tracho-matis or transfected with ISD (positive control) 8 h before harvest, asdescribed previously. The effect of specific siRNAs on target gene mRNAwas assessed by quantitative RT-PCR (qRT-PCR). For ISD transfection,

1 mg ISD was transfected into HeLa, and 2.5 mg ISD was transfectedinto OE-E6/E7 or BM1.11 cells using FuGene HD transfection reagent(Promega). For cGAMP treatment, cGAMP (1 mg/ml) was added to cellsat 37˚C in a permeabilization buffer (50 mM HEPES, 100 mM KCl, 3 mMMgCl2, 85 mM sucrose, 0.1 mM DTT, 0.2% BSA, 1 mM ATP, 10 mg/mldigitonin) described earlier (23) and replaced with media 30 min aftertreatment. Cells were harvested 6 h posttreatment.

RNA extraction, qRT-PCR analysis, and ELISA

The ISD-transfected cells, infected cells, or untreated cells were processed atthe same time for RNA extraction using the RNeasy kit (Qiagen). The RNAwere then processed for reverse transcription and quantitative PCR usinga SsoAdvanced SYBR mix (Bio-Rad) using a CFX iCycler (Bio-Rad), asdescribed previously (25). Primer sequences for IFN-b, IL-8, and 16S rRNAwere described previously (25). Additional primers include mcGAS-F, 59-TAGCGGTCTCAACTCAAG-39; mcGAS-R, 59-TGGTGTCTGTTCATA-GCA-39; hcGAS-F, 59-CCTGCTGTAACACTTCTTAT-39; hcGAS-R, 59-TTAGTCGTAGTTGCTTCCT-39; hTREX-1-F, 59-TGCCTTCTGTGTG-GATAG-39; hTREX-1-R, 59-AGTGTAGATGCTGCCTAG-39; mTREX-1-F, 59-CAATAGCCACTCTGTATG-39; and mTREX-1-R, 59-TGACC-GCTATGACTTTCC-39. All primers were designed using Beacon Designsoftware (Bio-Rad). Culture supernatants were collected at 24 h p.i forIFN-b and CXCL10 by ELISA (R&D Systems).

HeLa cell coculture following siRNA transfection todemonstrate cGAMP transfer

Twenty-four hours after second individual siRNA transfection, the cellswere trypsinized and counted. cGAS siRNA and STING siRNA cells werereplated individually or mixed at different ratio (1:3, 1:1, and 3:1). Equalnumbers of the mixed or individual siRNA-transfected cells were plated intoa new 24-well plate (23 105cell/well) and were infected with C. muridarum(3 MOI) 18 h later. ISD transfection (1 mg/well) was done on parallel wells6 h before harvest, so that all cells were harvested at the same time for RNAextraction. In an independent experiment, cells with knock down (KD) forcGAS or STING expression were cocultured in a Transwell separated bypermeable support, to determine the requirement of cell–cell contact.

HEK293T transfection and immunoblots

HEK293T cells were transfected with pCMV-cGAS (Origene), pCMV-STING (Origene), or pcDNA3.1 vector (60 ng DNA). Cells were trypsi-nized 24 posttransfection and replated individually or mixed at 1:1 ratio,such that each well contained a total of 43 105 cells/well in a 24-well dish.Cells were infected or permeabilized with positive control ligands cGAMP(1 mg/ml) or transfected with ISD (1 mg) 6 h after plating, harvested atindicated times, and processed for RNA and qRT-PCR. Immunoblots forcGAS or STING protein in HEK293, HEK293T, and HeLa cells werecarried out using cell lysate prepared using radioimmunoprecipitation as-say buffer (Pierce-Thermo Scientific) and protease inhibitor cocktails(Sigma-Aldrich). Anti-cGAS Ab (catalogue AP10510c; Abgent), anti-STINGpolyclonal Ab (catalogue PA5-26751; Thermo Scientific), and anti-actinAb (catalogue A2228; Sigma-Aldrich) were used at the 1:500, 1:500, and1:1000 concentrations, respectively.

Confocal microscopy

HeLa cells were grown on 12-mm glass coverslips (#1, 0.17 mm thickness) at50% density in 24-well plate, 1 d before infection with 1 MOI of C. mur-idarum or C. trachomatis. Twenty-four hours p.i, cells were fixed with 4%paraformaldehyde (Electron Microscopy Sciences, Hatfield, PA) for immu-nostaining, as described previously (25). Rabbit anti-cGAS was used at1:200 dilution. Chlamydiae were stained with C. muridarum antiserumobtained from convalescent mice p.i diluted 1:300, whereas Alexa Fluor(AF) 488–conjugated anti-mouse and AF568 anti-rabbit (Invitrogen) wereused at 1:1000 dilution as secondary Abs for detection. AF488-conjugatedmouse mAb for STING (R&D Systems) was used at 1:100 dilution in con-junction with anti-cGAS polyclonal Ab for colocalization studies. AF647anti-GM130 (BD Biosciences) was used for Golgi staining at 1:100 dilutionin some experiments. Transfected cells were stained using mouse anti-FLAG(Origene). Cells were washed and mounted using Prolong antifade containingDAPI (Invitrogen). Confocal images were acquired with the 633 oil 0.8numerical aperture objective using Zeiss confocal microscope (LSM 510META), and images were analyzed using AxioVision software (Thorn-wood, NY).

Statistical analysis

At least three independent repeats were performed for each siRNA ex-periment, and a representative experiment was shown. Error bar indicates

The Journal of Immunology 2395

by guest on August 31, 2022

http://ww

w.jim

munol.org/

Dow

nloaded from

the SE for technical replicates for qRT-PCR. To determine statistical sig-nificance in siRNA experiments, percentage of decrease/increase in ex-pression levels of IFN-b relative to NT siRNA (100%) was averaged frommultiple experiments, and significance was determined by paired t testor one-way ANOVA with Holm-Sidak multiple comparison test usingGraphpad Prism. For transfection experiments in HEK293T cells, foldchanges from three independent experiments were averaged and repre-sented with SD, and significance was determined by one-way ANOVAwithHolm-Sidak multiple comparison test.

ResultsCytosolic nucleic acid is a potential ligand for IFN-bexpression during chlamydial infection

During intracellular bacterial or viral infection, nucleic acids releasedinto the cytosol can result in IFN-b expression. Host exonucleases,such as TREX-1, cleave intracellular DNA and have been shownto regulate IFN-b expression during viral (28) or bacterial infec-tion (29). To determine whether cytosolic DNA contributes to IFN-bexpression during chlamydial infection, TREX-1 KO MEFs andtheir corresponding WT controls (Fig. 1A) were infected withC. muridarum. TREX-1 KO cells showed a significant 2- to 3-foldincrease in IFN-b expression compared with wild-type (WT) MEFsduring C. muridarum infection (Fig. 1B), without altering chlamydialgrowth, as measured by comparable levels of chlamydial 16S rRNA(Fig. 1C). TREX-1 KO cells were also hyperresponsive to the DNAanalog poly(dA-dT), used as a positive control (Fig. 1B). Further-more, complementation of TREX-1 KO cells with a cDNA constructexpressing human TREX-1 (Fig. 1D) resulted in 50–70% reductionin IFN-b expression during chlamydial infection or poly(dA-dT)treatment (Fig. 1E) relative to their respective vector controls, withoutaltering chlamydial replication (Fig. 1F).To further establish the contribution of cytosolic DNA to

IFN-b expression in chlamydial infection, siRNA knockdown ofTREX-1 was carried out in mouse oviduct epithelial cells BM1.11(Fig. 2A). A significant increase in IFN-b expression was ob-served in cells transfected with TREX-1 siRNA compared with

NT siRNA, during poly(dA-dT) treatment (p = 0.04) or C. muri-darum infection (p = 0.01) (Fig. 2B, Supplemental Fig. 1), againwithout affecting chlamydial replication (Fig. 2C). The role ofTREX-1 was further confirmed in HeLa cells, in which siRNAtransfection can be performed with high efficiency. Knockdown ofTREX-1 (Fig. 2D) resulted in significant increase (p = 0.045) inChlamydia-induced IFN-b expression relative to NT siRNA (Fig.2E, Supplemental Fig. 1), without altering chlamydial replication(Fig. 2F). The effect of TREX-1 siRNA was specific for ISD- butnot RNA-induced IFN-b expression, as evidenced in cells trans-fected with the RNA analog poly I:C (Fig. 2E, Supplemental Fig.1). Overall, these data suggest that cytosolic DNA is a ligandrequired for IFN-b expression during chlamydial infection.

The DNA sensor, cGAS, is essential for IFN-b expressionduring C. muridarum and C. trachomatis infection

That DNA serves as a ligand for IFN-b expression during chla-mydial infection suggested that host DNA sensors detect cytosolicDNA during infection. We have previously shown that the adaptormolecule STING is required for IFN-b expression during chla-mydial infection in HeLa cells (25). Using these cells, we investi-gated potential DNA sensors involved in recognition of this ligandduring infection. siRNA knockdown of several DNA sensors, spe-cifically DNA-dependent activator of IFN regulatory factors, IFN-g–inducible protein 16, leucine-rich repeat protein FLII-interactingprotein 1, DDX41, Lsm14A, and TRIM56 in HeLa cells, had noeffect on IFN-b expression during infection (data not shown). Re-cently, the host enzyme cGAS was shown to catalyze the generationof the STING ligand, cGAMP, after binding cytosolic DNA (20,21). We tested the contribution of cGAS to IFN-b expression duringchlamydial infection via knockdown using two independent siRNAsdirected against cGAS in comparison with a NT siRNA from thesame provider. siRNA-mediated knockdown for cGAS and STINGresulted in .90% and .99% reduction in mRNA expression ofcGAS and STING, respectively (data not shown). cGAS knock-

FIGURE 1. IFN-b expression during chlamydial infection is elevated in TREX-1 KO MEFs and reduced by TREX-1 overexpression. (A–C) Control

MEFs and TREX-1 KO MEFs were infected with C. muridarum at 1 MOI or transfected with poly(dA:dT) for 6 h, before harvest. Infected cells were

harvested at 24 h p.i. TREX-1 mRNA (A), IFN-b mRNA (B), and chlamydial 16S rRNA (C) levels were measured by qRT-PCR. (D–F) TREX-1 KO cells

were transfected with human TREX-1 or vector control. Twenty-four hours posttransfection, cells were infected with C. muridarum or transfected with poly

(dA:dT). Relative expression levels of TREX-1 (D), IFN-b (E), and chlamydial 16S rRNA (F) normalized to GAPDH are shown. Panels are representative

of three independent experiments, and error bars represent the range in technical replicates. Statistical significance for quantitative PCR data between

multiple experiments was determined by using paired t tests on percent change in IFN-b levels between WT and TREX-1 KO-infected cells (p = 0.01). For

transfection of TREX KO cells, percent change in IFN-b between vector and TREX-1 cDNA from multiple experiments was used to calculate significance

[infection p = 0.001, poly(dA:dT) p = 0.009]. UT, untreated.

2396 cGAS IN CHLAMYDIA-INDUCED IFN-b

by guest on August 31, 2022

http://ww

w.jim

munol.org/

Dow

nloaded from

down with cGAS1 siRNA and cGAS2 siRNA decreased the ex-pression of IFN-b by .75% (p , 0.001) and .97% (p , 0.001)relative to NT controls during C. muridarum infection (Fig. 3A,3K). Similar results were observed with human C. trachomatisserovar D and C. trachomatis L2 (Fig. 3A). Cells transfected withISD, in which induction of IFN-b was significantly compromisedin STING and cGAS knockdown cells, served as positive controlsfor functional siRNA knockdown, showing .95% decrease (p ,0.001) in IFN-b expression relative to NT controls (Fig. 3B, 3K).To demonstrate the specificity of the siRNAs used, cells weretransfected with poly I:C in parallel. cGAS and STING siRNA didnot decrease IFN-b induction in poly I:C-transfected cells, sug-gesting that RNA-sensing pathways or downstream effectors forIFN-b expression were not targeted (Fig. 3C, 3K), although someincrease in IFN-b expression was observed in cells transfectedwith one of the two cGAS siRNAs tested (p = 0.024). cGAS andSTING siRNA significantly reduced the expression of another IFNregulatory factor 3 (IRF3)-dependent gene IFN-l during infectionand ISD transfection (Fig. 3D–F), but did not reduce the expres-sion of IL-8, an IRF3 independent gene (Fig. 3G), demonstratinga specific effect on the IRF3 pathway. Chlamydial replicationremained unaltered in response to cGAS or STING knockdown, asevidenced by 16S rRNA levels (Fig. 3H). IFN-b protein was notdetectable in supernatants because it is most likely rapidly endo-cytosed by its receptor IFNAR. Therefore, protein levels of an IFNresponse gene, CXCL10, which serves as a functional surrogate forbiologically active IFN-b, were measured in the culture super-natants by ELISA and were found to correspond to IFN-b mRNAlevels in the cells (Fig. 3I). siRNA knockdown of STING andcGAS resulted in undetectable or low levels of the respectiveproteins in cell lysates detected by Western blot (Fig. 3J). Fur-thermore, cGAS knockdown did not alter STING levels and viceversa (Fig. 3J), confirming the effect of cGAS on IFN-b expressionwas independent of STING protein levels.The role of cGAS and STING was further assessed in mouse and

human oviduct epithelial cells infected with C. muridarum orC. trachomatis, respectively (Fig. 4). About 50 and 90% reduction

in IFN-b expression was observed in mouse oviduct epithelialcells (BM1.11) with cGAS and STING knockdown, respectively,during C. muridarum infection (Fig. 4A, 4D). These results par-alleled a 50 and 80% decrease of cGAS and STING expression,respectively (Fig. 4B, 4C), whereas chlamydial growth remainedunaffected by cGAS and STING knockdown (data not shown).IFN-b expression during ISD transfection in BM1.11 cells wasalso significantly reduced in cells transfected with cGAS andSTING siRNA relative to NT siRNA (Fig. 4A, 4D). In OE-E6/E7cells (27), a transformed human oviduct cell line, a similar effectof cGAS and STING knockdown, was observed (Fig. 4E, 4H),although these cells only induced low levels of IFN-b duringC. trachomatis infection. The reduction in IFN-b expressionparalleled ∼50% knockdown achieved for both cGAS and STINGexpression in these cells (Fig. 4F, 4G). cGAS was also found to beessential for IFN-b expression during chlamydial infection in themouse macrophage line J774 (data not shown), which supports thegrowth of C. muridarum. These data indicate that DNA sensing bycGAS during chlamydial infection occurs in multiple cell typesand with both mouse and human C. trachomatis strains.

cGAS is recruited to the inclusion membrane duringchlamydial infection

We have previously shown that STING and the ER protein Sec16alocalized in close proximity to the chlamydial inclusion (25). cGAShas been shown to colocalize with transfected DNA in the hostcytosol (20). To determine the intracellular niche(s) in which cGASrecognizes DNA during chlamydial infection, immunostaining ofendogenous cGAS and STING was performed in infected cells.Ab specificity was confirmed by transfecting cGAS or STING intoHEK293T cells (data not shown), which do not express eitherprotein at high levels (20). Endogenous cGAS was found to localizein close proximity to chlamydial inclusion membrane during bothC. muridarum (Fig. 5A) and C. trachomatis (Fig. 5B) infection. Inuninfected cells, cGAS was found distributed in the cytosol (Fig. 5C),whereas in infected cells cGAS was found enriched at certain pointsaround the chlamydial inclusion membrane (Fig. 5D). Colocalization

FIGURE 2. siRNA knockdown of TREX-1 in epithelial cells increases IFN-b expression during chlamydial infection. (A–C) Mouse BM1.11 cells were

transfected with TREX-1 siRNA or NT siRNA (A) and infected with C. muridarum at 1 MOI, 72 h posttransfection. TREX-1 (A) and IFN-b (B) mRNA

were measured at 24 h p.i. In parallel, cells were transfected with dsDNA analog poly(dA:dT), 6 h before harvest. Chlamydial replication was monitored by

comparing 16S rRNA levels (C) in infected cells between treatments. (D–F) HeLa cells were transfected with siRNA for human TREX-1 or NT siRNA.

Cells were infected with C. muridarum at 1 MOI for 24 h or transfected 6 h before harvest with ISD or RNA analog poly I:C. TREX-1 (D), IFN-b (E), and

chlamydial 16S rRNA levels (F) were measured by qRT-PCR. Panels are representative of three independent experiments, and error bars represent the mean6error of technical replicates. Statistical significance for quantitative PCR data between multiple experiments was determined by using paired t tests on percent

change in IFN-b levels between TREX-1 siRNA relative to NT siRNA in each experiment. For BM1.11 cells, p = 0.014 for infection and p = 0.047 for poly

dA-dT. For HeLa, p = 0.04 for infection. p = 0.04 for ISD. p = NS for poly I:C. UT, untreated.

The Journal of Immunology 2397

by guest on August 31, 2022

http://ww

w.jim

munol.org/

Dow

nloaded from

of STING and cGAS was observed at several contact points aroundthe inclusion (Fig. 5D), but no colocalization of STING or cGAS wasobserved with the Golgi marker GM130 (data not shown). To confirmthe data obtained by endogenous staining of cGAS, we examinedtrafficking and localization of cGAS in HeLa cells transfected withFLAG-cGAS. A distinct enrichment of FLAG-cGAS was observed aspunctate staining around the inclusion in infected cells (Fig. 5E). Thesedata are reminiscent of what was previously observed with STING(25). cGAS is a cytosolic protein and not associated with the mem-brane component of the cells. Therefore, these data suggest that cGASis trafficked to the inclusion membrane during chlamydial infection.

Evidence for cGAMP generation during chlamydial infection

Recognition of DNA by cGAS results in the generation of theSTING ligand, cGAMP (20, 21). To determine whether cGAMP is

generated during chlamydial infection, we used an indirect approachbased on the recent finding that cGAMP can cross gap junctionsbetween epithelial cells (30). HeLa cells knocked down for STINGexpression should retain the capacity to generate cGAMP, becausethey express cGAS, but are incapable of inducing IFN-b expressionin the absence of STING. Similarly, cells knocked down for cGAScannot generate cGAMP, but retain their ability to induce IFN-bexpression because STING can respond to cGAMP provided in trans.Therefore, we hypothesized that IFN-b expression would be rescuedif the cells knocked down for cGAS or STING were mixed prior toinfection (Fig. 6A). siRNA knockdown of these proteins (cGAS orSTING) was carried out in HeLa cells as before, and cells weremixed at varying ratios with equal cell number in all wells. IFN-bexpression was rescued by.10- to 20-fold in infected cells that werecocultured relative to cells with cGAS or STING knockdowns, re-

FIGURE 3. cGAS is required for IFN-b expression

in HeLa cells infected with C. muridarum or C. tra-

chomatis (D and L2). HeLa cells were transfected

with NT siRNA, two different siRNA for cGAS

(cGAS1 and cGAS2), or a siRNA for STING, as

described in Materials and Methods. Significant

knockdown of STING (99%) and cGAS (90%) mRNA

was achieved. Cells were infected with 1 MOI of

C. muridarum (C.M) or 5 MOI of C. trachomatis-

serovar D (C.T-D) or C. trachomatis L2 (C.T-L2).

Cells were harvested at 24 h p.i and analyzed for

expression of IFN-b (A), IFN-l (D), IL-8 (G), and

chlamydial 16S rRNA (H). In parallel, cells were

transfected with ISD (positive control) or poly I:C-

LyoVec (negative control), harvested at 6 h post-

transfection, and analyzed for expression of IFN-b

(B, C) and IFN-l (E, F). Culture supernatants from

infected or transfected cells were collected at 24 h,

and CXCL10 protein levels were assayed by ELISA

(I). Mean 6 SD of samples from three independent

experiments are shown for ELISA. A representative

Western blot showing the levels of STING and

cGAS following siRNA knockdown in uninfected

HeLa cells (J). A representative of five independent

experiments is presented in (A)–(H) for qRT-PCR

data, and error bars represent range in technical rep-

licates. Statistical significance for quantitative PCR

data between multiple experiments was determined

by one-way ANOVA with multiple comparison tests

on the percent decrease in IFN-b levels for the siRNA

used relative to NT siRNA in each experiment (K).

UT, untreated.

2398 cGAS IN CHLAMYDIA-INDUCED IFN-b

by guest on August 31, 2022

http://ww

w.jim

munol.org/

Dow

nloaded from

spectively (Fig. 6B). Chlamydial growth was similar in coculturedcells relative to individually plated cells, as evidenced by 16S rRNAlevels (data not shown). Cells transfected with ISD served as a posi-tive control and showed a similar trend of rescue of IFN-b expressionresulting from cGAMP transfer (Fig. 6C). Gap junction inhibitorssuch as Carbenoxolone were able to abrogate cGAMP transfer incocultured cells transfected with ISD (data not shown), as previouslyreported (30), but could not be used in conjunction with infectedcells as they also function as Pannexin inhibitor and abrogated chla-mydial growth (31). Cell–cell contact was essential for IFN-b rescueto occur during coculture, because IFN-b rescue was compromisedwhen the STING KD and cGAS KD cells were cultured separately ona Transwell (Supplemental Fig. 2). Together, these results provideindirect evidence for generation of cGAMP during infection andits transfer from infected cell to adjacent cell through gap junctions.

DNA is the predominant ligand inducing IFN-b expressionduring chlamydial infection

Recently, it was reported that chlamydial elementary bodies (EBs)produce the bacterial second-messenger cyclic di-AMP, which can

directly bind to STING to induce IFN-b expression (24). This wasshown by infecting HEK293T cells overexpressing STING andtransfected with IFN-b promoter-driven luciferase reporter con-struct. In light of our finding that cGAS is required for Chlamydia-induced IFN-b, we investigated the contribution of STING inthe presence or absence of cGAS, to determine the relative con-tributions of DNA and chlamydial cyclic di-AMP to IFN-b ex-pression. HEK293 cells (without T Ag) express STING, but notcGAS, whereas HEK293T (with T Ag) have undetectable cGASand STING protein, as reported earlier (20), contrasting withHeLa cells, which express high levels of both STING and cGAS(Fig. 7A). The STING protein in HEK293 cells was fully func-tional, as evidenced by high levels of IFN-b expression in re-sponse to the commercially available STING ligand 2939cGAMP(Fig. 7B), but these cells expressed very low levels of endogenousIFN-b expression (2-fold increase) in response to chlamydial in-fection (Fig. 7C). These data suggested that STING expressionalone was insufficient to drive biologically relevant endogenousIFN-b expression. HEK293T cells were then used to examine theindependent and combined role of cGAS and STING in IFN-b

FIGURE 4. cGAS is required for IFN-b ex-

pression during Chlamydia spp. infection of

mouse oviduct epithelial cells (BM1.11 cells)

and human oviduct epithelial cells (OE-E6/E7).

siRNA knockdown in mouse BM1.11 cells was

carried out using Accell NT, mouse cGAS, or

STING siRNA pools. Seventy-two hours after

transfection, cells were infected with C. muri-

darum or transfected with ISD (DNA) 6 h before

harvest and analyzed simultaneously at 24 h p.i

for expression of mouse IFN-b (A), cGAS (B),

and STING (C). Human oviduct epithelial cells

(OE-E6/E7) were transfected with NT, human

cGAS (cGAS siRNA 1 and 2), or human STING

siRNA. Seventy-two hours after transfection,

cells were infected with C. trachomatis (serovar

D) at 5 MOI or transfected with ISD (DNA) 6 h

before harvest and analyzed concurrently at 24 h

p.i for expression of human IFN-b (E), cGAS

(F), and STING (G). A representative of three

experiments for BM1.11 cells and OE cells is

presented for qRT-PCR data. Statistical signifi-

cance for quantitative PCR data between multi-

ple experiments was determined by one-way

ANOVA with multiple comparison tests on the

percent decrease in IFN-b levels for the siRNA

used relative to NT siRNA in each experiment

(D and H).

The Journal of Immunology 2399

by guest on August 31, 2022

http://ww

w.jim

munol.org/

Dow

nloaded from

expression during infection. Cotransfection of cGAS and STINGcDNA into HEK293T cells resulted in significant backgroundIFN-b expression—a consequence of recognition of transfectedDNA by cGAS protein in the presence of STING (20). Therefore,the alternative approach of examining the contribution of cGASin trans was employed. HEK293T cells were transfected withSTING or cGAS, which resulted in equivalent levels of STING orcGAS mRNA (data not shown). HEK293T cells transfected withSTING were responsive to exogenous cGAMP but did not in-duce IFN-b expression to transfected DNA. Conversely, cGAS-transfected cells were unresponsive to either cGAMP or ISDtransfection in the absence of STING signaling (Fig. 7D). WhencGAS-expressing cells were cocultured with STING-expressingcells, 6 h before infection at a 1:1 ratio, IFN-b expression wasrescued 40-fold, relative to infected cells expressing STING orcGAS alone (Fig. 7E). Chlamydial growth was similar in cocul-tured cells to individually plated cells, as evidenced by 16S rRNAlevels (data not shown). Cells transfected with ISD served aspositive control (Fig. 7E). These data suggest that cGAMP pro-duced in cGAS-expressing cells migrated into STING-expressingcells to induce IFN-b expression during infection and DNAtransfection. Together these data demonstrate that STING alone isinsufficient to rescue significant IFN-b expression in the absence

of cGAS, and that both cGAS and STING function cooperativelyto induce maximal IFN-b expression during chlamydial infection(Fig. 8).

DiscussionA previous study from our laboratory showed that STING, the centraladaptor molecule involved in cytosolic DNA recognition, is essentialfor IFN-b induction during chlamydial infection (25). Recent iden-tification of the DNA sensor, cGAS, that catalyzes the formation ofcGAMP, a STING ligand (20), led us to investigate whether this wasthe major pathway leading to IFN-b expression during chlamydialinfection. Our results show that cGAS is required for IFN-b ex-pression during infection with multiple Chlamydia strains in HeLa,epithelial cell lines that were derived from human Fallopian tube ormouse oviducts, mouse macrophages, and HEK293T cells. cGASmeets the requirements for the DNA sensor during chlamydial in-fection, because it is specifically required for IRF3-mediated IFN-bexpression and it is localized adjacent to inclusion membrane. To ourknowledge, this is the first report on a role for cGAS during intra-cellular bacterial infection, although its role as a DNA sensor duringviral infection has been described (32).The evidence that DNA was a potential ligand for IFN-b ex-

pression during infection came from TREX-1 KO cells. TREX-1,

FIGURE 5. cGAS localizes in punctate regions

around the chlamydial inclusion membrane. HeLa cells

infected with C. muridarum (A) or C. trachomatis,

serovar D (B), were fixed and stained for endogenous

cGAS (red) and Chlamydia (green). Cells were fixed

with Prolong gold with DAPI (blue) and analyzed by

confocal microscopy. Uninfected HeLa cells (C) and

cells infected with C. muridarum (D) for 18 h were

fixed and stained for endogenous cGAS (red) and

STING (green). In an independent experiment, HeLa

cells were transfected with FLAG-cGAS (E) and

infected with C. muridarum 24 h later. Infected cells

were fixed at 18 h p.i and stained using mouse mAb for

FLAG. Chlamydial inclusions and cell nucleus are

marked with an I and N, respectively, on the DAPI

staining. Images were collected with a 633 objective

and 43 digital zoom.

2400 cGAS IN CHLAMYDIA-INDUCED IFN-b

by guest on August 31, 2022

http://ww

w.jim

munol.org/

Dow

nloaded from

a member of TREX family of proteins, is essential for degradationof cytosolic DNA (33) and could function to reduce autoimmuneresponse to host DNA. During HIV infection, host TREX-1 targetsHIV DNA to abrogate IFN-b expression, a classic example ofviral hijacking of host machinery (28). Using siRNA knockdownof TREX-1 in mouse epithelial cells and HeLa cells and TREX-1KO cells, we determined that cytosolic DNA is targeted by thesenucleases during chlamydial infection. Subsequent screening ofmultiple host DNA sensors did not identify a Chlamydia-specific

sensor until we tested the role for cGAS in IFN-b expression andcorrelated its function to its cellular localization during chla-mydial infection. In addition to demonstrating a role for cGAS inChlamydia-induced IFN-b expression, we also provide indirectevidence for generation of cGAMP during chlamydial infectionand its transfer from infected cell to adjacent cells. These ex-periments were built on the recent finding that cGAMP can crossgap junctions and activate STING in neighboring cells (30). Thesedata suggest that, during in vivo infection, IFN-b induced in un-infected cells adjacent to infected epithelial cells could eithermake them refractory to infection, as during viral infection, orinduce cell death pathways to affect tissue pathology.It has been shown that the expression level of cGAS in multiple

cell types is correlated to IFN-b expression in response to ISD(20). Therefore, the contribution of cGAS to IFN-b expressionduring chlamydial infection, relative to other receptors, is mostlikely dependent on their expression levels in multiple cell types.We have shown that cGAS is active in multiple cell types duringchlamydial infection, including those where other candidate re-ceptors, for example, TLR3 (34), have been associated with IFN-bexpression. Interestingly, cGAS is a bonafide IFN-inducible pro-tein (35), and we have observed a significant increase in cGASexpression following infection or ISD transfection. Therefore,during in vivo infection, both IFN-b and IFN-g can induce cGASexpression in epithelial cells. Analysis of cGAS and STING geneknockout mice should provide a more definitive answer on therelative contribution of cGAS over other receptors on IFN-b ex-pression in the genital tract during in vivo genital infection.Barker et al. (24) recently showed that C. trachomatis strain

L2 generates the bacterial second-messenger cyclic di-AMP,which induced a IFN-b promoter-driven luciferase reporter inHEK293T cells overexpressing STING. Furthermore, the authorsconcluded that cyclic di-AMP is the predominant inducer of IFN-b during chlamydial infection using a mouse STING mutant(mR231A) defective for cyclic di-AMP binding (22, 36) in STINGKO mouse fibroblasts. Our findings conflict with the conclusionsof Barker et al. (24) because our data support a model in whichDNA sensing via cGAS rather than cyclic di-AMP is the pre-dominant pathway of IFN-b induction during chlamydial infectionof human epithelial cells, but without disregarding the findingsthat chlamydial EBs generate cyclic di-AMP. Our model is sup-ported by the following observations. First, if recognition of cyclicdi-AMP by STING was sufficient to drive IFN-b expressionduring chlamydial infection, cGAS knockdown should not havealtered IFN-b expression significantly in multiple cell types thatexpress abundant and functional STING. Second, STING ex-pression was insufficient to rescue endogenous IFN-b expressionduring infection in cell lines expressing a functional STING, butnot cGAS, such as HEK293 cells. Finally, HEK293T transfectedwith cGAS or STING did not induce significant endogenous IFN-b expression during chlamydial infection. However, coculturingthe cells rescued expression, further confirming the requirement ofcGAS-mediated cGAMP generation and transfer during infection.cGAMP binds to the same pocket in STING as cyclic di-AMP/di-GMP, but at much lower concentration with higher affinity (21).Indeed, the cGAS product, 2939cGAMP, is a much more potentligand of STING than all other bacterial cyclic dinucleotides de-scribed (37). Furthermore, human STING is responsive only tocGAMP and unresponsive to the STING ligands CMA (38) andcyclic di-AMP/cyclic di-GMP (39), unlike mouse STING, whichis responsive to both cyclic dinucleotides and cGAMP (40). Thesestudies combined with our findings would significantly shift theimportance of cGAMP over bacterial cyclic dinucleotides duringC. trachomatis infection in human cells.

FIGURE 6. Evidence of cGAMP transfer from infected cells to adjacent

cells. Schematic representation of the experimental plan (A). STING2

cGAS+ can produce cGAMP but cannot induce IFN-b because they lack

STING, whereas STING+cGAS2 cells cannot produce cGAMP upon in-

fection. Coculture can rescue IFN-b expression during infection if cGAMP

from cGAS+ cells can migrate into STING+ cells. HeLa cells knocked down

for STING (.99% KD) or cGAS (.95% KD) were cultured individually or

cocultured at different ratios 18 h before infection in 24-well dishes (2 3105 cells/well). The mix numbers (1–3) represent cGAS KD:STING KD

cells in the indicated ratios. Cells were infected with C. muridarum at

3 MOI and IFN-b mRNA measured at 24 h p.i (B). A parallel set of cells

was transfected with ISD as a positive control, 6 h before harvesting all the

cells for RNA (C). Data are represented as mean of percent decrease relative

to NT control from three experiments with SD. Significance determined by

one-way ANOVAwith multiple comparison tests and indicated. Differences

between mix 1:3 and 1:1 coculture were not significant.

The Journal of Immunology 2401

by guest on August 31, 2022

http://ww

w.jim

munol.org/

Dow

nloaded from

An important question that remains unclear at present is what isthe source of the cytosolic DNA during infection? We speculatethat the source of the DNA is chlamydial. Mitochondrial damagehas not been observed at 24 h p.i (data not shown). Furthermore,addition of chloramphenicol to block chlamydial growth afterinclusion formation abrogates IFN-b expression, confirming theconsistent requirement of bacterial growth (41). The localizationof cGAS in punctate regions around the inclusion is also sug-

gestive of a chlamydial source for the DNA. Manzanillo et al. (29)have shown that, during Mycobacterium tuberculosis infection,phagosomal permeabilization mediated by the bacterial ESX-1secretion system allows cytosolic recognition pathways access toDNA. Numerous studies have linked IFN-b expression to bacterialsecretion systems (42–45), and we have shown that IFN-b ex-pression is abrogated in C. muridarum–infected cells exposed todrugs that inhibit type III secretion system (T3SS) (41), suggest-ing a similar role for chlamydial T3SS in permeabilization ofinclusion membrane. Previous studies (46) have shown thatchlamydial reticulate bodies (RB) make direct contact with theinclusion membrane, most likely through T3SS. These could bepotential permeabilization points in which nucleic acids couldleak into cytosol and are made available for host recognition. It isimportant to note that only viable chlamydiae induce IFN-b ex-pression. This would suggest that condensed DNA from UV-killedchlamydial EBs do not provide an acceptable form of DNA forrecognition by cGAS. Whereas EB are a known source of cyclicdi-AMP (24), when these developmental forms transition into RB,and RB replicate, this may initiate and accelerate DNA recogni-tion, significantly amplifying IFN-b expression levels. Lateralgene transfer has been shown to occur between chlamydial RBs(47), consistent with a model in which extrachlamydial DNAis available for sensing and supporting the possibility of DNAtransfer into cytosol, although DNA could also be released asa passive process resulting from nonviable RBs inside the inclu-sion. It has been shown that Chlamydia hijacks the host ER, andseveral ER proteins were found localized on inclusion membrane(48, 49). The localization of the ER protein STING (25) and cy-tosolic cGAS in close proximity to the inclusion membrane sug-gest that STING could serve as a membrane scaffold for the

FIGURE 7. Both cGAS and STING are required for maximal IFN-b expression during chlamydial infection. Protein levels of cGAS and STING in

HeLa, HEK293T, and HEK293 cell lysates (A). Actin blots were carried out using one-tenth of cell lysates used for cGAS and STING Western blots.

HEK293 cells were permeabilized with cGAMP and IFN-b expression measured 6 h posttreatment (B). HEK293 cells were infected with C. muridarum or

transfected with ISD and IFN-b expression measured 24 h p.i or 6 h posttreatment, respectively (C). HEK293T cells were transiently transfected with

pcDNA3.1 (vector), STING, or cGAS expression constructs. Twenty-four hours posttransfection, cells were trypsinized and plated individually or mixed at

indicated ratio. Six hours after plating, individual cells were transfected with ISD or permeabilized with cGAMP (D). Individual or mixed cells were

transfected with ISD or infected with C. muridarum (E). Expression of IFN-b was measured at 24 h p.i or 6 h after ISD/cGAMP transfection. Data represent

average of three experiments, and error bars indicate SD. Significance was determined by one-way ANOVA with multiple comparison test. For (D), p =

0.004 for STING transfected-untreated (UT) versus cGAMP-treated cells.

FIGURE 8. A schematic model of cGAS recognizing DNA during

Chlamydia infection, in human epithelial cells. Left panel, Shows an

electron micrograph of an inclusion, in which chlamydial RBs are in close

contact with the inclusion membrane and rough ER (arrows). Right panel,

Shows a hypothetical model of DNA recognition by cGAS leading to

cGAMP generation, STING activation, and IRF3 phosphorylation, result-

ing in IFN-b expression during infection. Contribution of TREX-1 in

diminishing this response is also shown.

2402 cGAS IN CHLAMYDIA-INDUCED IFN-b

by guest on August 31, 2022

http://ww

w.jim

munol.org/

Dow

nloaded from

interactions between DNA and cGAS to take place. Activation ofSTING would then result in phosphorylation and nuclear trans-location of the critical transcription factor IRF3 for IFN-b ex-pression (Fig. 8).An IFN-b response occurs during a vast number of intracellular

infections, arising from bacteria that can occupy diverse niches inthe cell (cytosolic, lysosomal, or vacuolar). In all cases, cytosolicDNA could be available for sensing; thus, we predict that cGASmay be a common recognition mechanism for sensing other intra-cellular pathogens. Furthermore, the intensity of IFN-b expres-sion could be directly correlated with the availability of foreignDNA in the cytosol, with cytosolic pathogens inducing a strongerresponse relative to those sequestered in membrane organelles.Collectively, our data demonstrate that cGAS is a novel pathogenrecognition receptor involved in recognition of chlamydial infec-tion and implicate cytosolic DNA recognition during infection asan inducer of this response. The requirement of cGAS for Chla-mydia-induced IFN-b expression also provides a novel therapeutictarget to block this response using a cGAMP antagonist to protectagainst oviduct disease during genital chlamydial infection.

AcknowledgmentsA significant portion of the work was carried out at the Department of Pediat-

rics, University of Pittsburgh. TREX-1 KO andWTmouse embryonic fibro-

blasts were provided by Cancer Research UK (London, U.K.) BM1.11 cells

were provided by Dr. Raymond Johnson (Indiana University, Indianapolis,

IN). We thank Dr. Toni Darville and Dr. Catherine O’ Connell (University

of North Carolina at Chapel Hill) for helpful suggestions and Theodor

Danciu (University of Pittsburgh) for laboratory assistance.

DisclosuresThe authors have no financial conflicts of interest.

References1. Xia, M., R. E. Bumgarner, M. F. Lampe, and W. E. Stamm. 2003. Chlamydia

trachomatis infection alters host cell transcription in diverse cellular pathways. J.Infect. Dis. 187: 424–434.

2. Rasmussen, S. J., L. Eckmann, A. J. Quayle, L. Shen, Y. X. Zhang,D. J. Anderson, J. Fierer, R. S. Stephens, and M. F. Kagnoff. 1997. Secretion ofproinflammatory cytokines by epithelial cells in response to Chlamydia infectionsuggests a central role for epithelial cells in chlamydial pathogenesis. J. Clin.Invest. 99: 77–87.

3. Lad, S. P., E. Y. Fukuda, J. Li, L. M. de la Maza, and E. Li. 2005. Up-regulationof the JAK/STAT1 signal pathway during Chlamydia trachomatis infection. J.Immunol. 174: 7186–7193.

4. Nagarajan, U. M., D. Prantner, J. D. Sikes, C. W. Andrews, Jr., A. M. Goodwin,S. Nagarajan, and T. Darville. 2008. Type I interferon signaling exacerbatesChlamydia muridarum genital infection in a murine model. Infect. Immun. 76:4642–4648.

5. Qiu, H., Y. Fan, A. G. Joyee, S. Wang, X. Han, H. Bai, L. Jiao, N. Van Rooijen,and X. Yang. 2008. Type I IFNs enhance susceptibility to Chlamydia muridarumlung infection by enhancing apoptosis of local macrophages. J. Immunol. 181:2092–2102.

6. Prantner, D., J. D. Sikes, L. Hennings, A. V. Savenka, A. G. Basnakian, andU. M. Nagarajan. 2011. Interferon regulatory transcription factor 3 protects micefrom uterine horn pathology during Chlamydia muridarum genital infection.Infect. Immun. 79: 3922–3933.

7. Nagarajan, U. 2011. Induction and function of IFNb during viral and bacterialinfection. Crit. Rev. Immunol. 31: 459–474.

8. Stetson, D. B., and R. Medzhitov. 2006. Type I interferons in host defense.Immunity 25: 373–381.

9. Ablasser, A., F. Bauernfeind, G. Hartmann, E. Latz, K. A. Fitzgerald, andV. Hornung. 2009. RIG-I-dependent sensing of poly(dA:dT) through the in-duction of an RNA polymerase III-transcribed RNA intermediate. Nat. Immunol.10: 1065–1072.

10. Chiu, Y. H., J. B. Macmillan, and Z. J. Chen. 2009. RNA polymerase III detectscytosolic DNA and induces type I interferons through the RIG-I pathway. Cell138: 576–591.

11. Seth, R. B., L. Sun, C. K. Ea, and Z. J. Chen. 2005. Identification and charac-terization of MAVS, a mitochondrial antiviral signaling protein that activatesNF-kappaB and IRF 3. Cell 122: 669–682.

12. Takaoka, A., Z. Wang, M. K. Choi, H. Yanai, H. Negishi, T. Ban, Y. Lu,M. Miyagishi, T. Kodama, K. Honda, et al. 2007. DAI (DLM-1/ZBP1) is a cy-

tosolic DNA sensor and an activator of innate immune response. Nature 448:501–505.

13. Unterholzner, L., S. E. Keating, M. Baran, K. A. Horan, S. B. Jensen, S. Sharma,C. M. Sirois, T. Jin, E. Latz, T. S. Xiao, et al. 2010. IFI16 is an innate immunesensor for intracellular DNA. Nat. Immunol. 11: 997–1004.

14. Yang, P., H. An, X. Liu, M. Wen, Y. Zheng, Y. Rui, and X. Cao. 2010. Thecytosolic nucleic acid sensor LRRFIP1 mediates the production of type I in-terferon via a beta-catenin-dependent pathway. Nat. Immunol. 11: 487–494.

15. Zhang, Z., B. Yuan, M. Bao, N. Lu, T. Kim, and Y. J. Liu. 2011. The helicaseDDX41 senses intracellular DNA mediated by the adaptor STING in dendriticcells. Nat. Immunol. 12: 959–965.

16. Kondo, T., J. Kobayashi, T. Saitoh, K. Maruyama, K. J. Ishii, G. N. Barber,K. Komatsu, S. Akira, and T. Kawai. 2013. DNA damage sensor MRE11 rec-ognizes cytosolic double-stranded DNA and induces type I interferon by regu-lating STING trafficking. Proc. Natl. Acad. Sci. USA 110: 2969–2974.

17. Li, Y., R. Chen, Q. Zhou, Z. Xu, C. Li, S. Wang, A. Mao, X. Zhang, W. He, andH. B. Shu. 2012. LSm14A is a processing body-associated sensor of viral nucleicacids that initiates cellular antiviral response in the early phase of viral infection.Proc. Natl. Acad. Sci. USA 109: 11770–11775.

18. Ferguson, B. J., D. S. Mansur, N. E. Peters, H. Ren, and G. L. Smith. 2012. DNA-PK is a DNA sensor for IRF-3-dependent innate immunity. eLife 1: e00047.

19. Ishikawa, H., and G. N. Barber. 2008. STING is an endoplasmic reticulumadaptor that facilitates innate immune signalling. Nature 455: 674–678.

20. 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.Science 339: 786–791.

21. Wu, J., L. Sun, X. Chen, F. Du, H. Shi, C. Chen, and Z. J. Chen. 2013. CyclicGMP-AMP is an endogenous second messenger in innate immune signaling bycytosolic DNA. Science 339: 826–830.

22. Burdette, D. L., K. M. Monroe, K. Sotelo-Troha, J. S. Iwig, B. Eckert,M. Hyodo, Y. Hayakawa, and R. E. Vance. 2011. STING is a direct innateimmune sensor of cyclic di-GMP. Nature 478: 515–518.

23. Woodward, J. J., A. T. Iavarone, and D. A. Portnoy. 2010. c-di-AMP secreted byintracellular Listeria monocytogenes activates a host type I interferon response.Science 328: 1703–1705.

24. Barker, J. R., B. J. Koestler, V. K. Carpenter, D. L. Burdette, C. M. Waters,R. E. Vance, and R. H. Valdivia. 2013. STING-dependent recognition of cyclicdi-AMP mediates type I interferon responses during Chlamydia trachomatisinfection. MBio. DOI: 10.1128/mBio.00018-13.

25. Prantner, D., T. Darville, and U. M. Nagarajan. 2010. Stimulator of IFN gene iscritical for induction of IFN-beta during Chlamydia muridarum infection. J.Immunol. 184: 2551–2560.

26. Johnson, R. M. 2004. Murine oviduct epithelial cell cytokine responses toChlamydia muridarum infection include interleukin-12-p70 secretion. Infect.Immun. 72: 3951–3960.

27. Lee, Y. L., K. F. Lee, J. S. Xu, Y. L. Wang, S. W. Tsao, and W. S. Yeung. 2001.Establishment and characterization of an immortalized human oviductal cellline. Mol. Reprod. Dev. 59: 400–409.

28. Yan, N., A. D. Regalado-Magdos, B. Stiggelbout, M. A. Lee-Kirsch, andJ. Lieberman. 2010. The cytosolic exonuclease TREX1 inhibits the innate im-mune response to human immunodeficiency virus type 1. Nat. Immunol. 11:1005–1013.

29. Manzanillo, P. S., M. U. Shiloh, D. A. Portnoy, and J. S. Cox. 2012. Myco-bacterium tuberculosis activates the DNA-dependent cytosolic surveillancepathway within macrophages. Cell Host Microbe 11: 469–480.

30. Ablasser, A., J. L. Schmid-Burgk, I. Hemmerling, G. L. Horvath, T. Schmidt,E. Latz, and V. Hornung. 2013. Cell intrinsic immunity spreads to bystander cellsvia the intercellular transfer of cGAMP. Nature 503: 530–534.

31. McKuen, M. J., G. Dahl, and K. A. Fields. 2013. Assessing a potential role ofhost Pannexin 1 during Chlamydia trachomatis infection. PLoS One 8: e63732.

32. Gao, D., J. Wu, Y. T. Wu, F. Du, C. Aroh, N. Yan, L. Sun, and Z. J. Chen. 2013.Cyclic GMP-AMP synthase is an innate immune sensor of HIV and other ret-roviruses. Science 341: 903–906.

33. Chowdhury, D., P. J. Beresford, P. Zhu, D. Zhang, J. S. Sung, B. Demple,F. W. Perrino, and J. Lieberman. 2006. The exonuclease TREX1 is in the SETcomplex and acts in concert with NM23-H1 to degrade DNA during granzymeA-mediated cell death. Mol. Cell 23: 133–142.

34. Derbigny, W. A., R. M. Johnson, K. S. Toomey, S. Ofner, and K. Jayarapu. 2010.The Chlamydia muridarum-induced IFN-b response is TLR3-dependent inmurine oviduct epithelial cells. J. Immunol. 185: 6689–6697.

35. Schoggins, J. W., S. J. Wilson, M. Panis, M. Y. Murphy, C. T. Jones, P. Bieniasz,and C. M. Rice. 2011. A diverse range of gene products are effectors of the type Iinterferon antiviral response. Nature 472: 481–485.

36. Diner, E. J., D. L. Burdette, S. C. Wilson, K. M. Monroe, C. A. Kellenberger,M. Hyodo, Y. Hayakawa, M. C. Hammond, and R. E. Vance. 2013. The innateimmune DNA sensor cGAS produces a noncanonical cyclic dinucleotide thatactivates human STING. Cell Reports 3: 1355–1361.

37. Zhang, X., H. Shi, J. Wu, X. Zhang, L. Sun, C. Chen, and Z. J. Chen. 2013.Cyclic GMP-AMP containing mixed phosphodiester linkages is an endogenoushigh-affinity ligand for STING. Mol. Cell 51: 226–235.

38. Cavlar, T., T. Deimling, A. Ablasser, K. P. Hopfner, and V. Hornung. 2013.Species-specific detection of the antiviral small-molecule compound CMA bySTING. EMBO J. 32: 1440–1450.

39. Conlon, J., D. L. Burdette, S. Sharma, N. Bhat, M. Thompson, Z. Jiang,V. A. Rathinam, B. Monks, T. Jin, T. S. Xiao, et al. 2013. Mouse, but not humanSTING, binds and signals in response to the vascular disrupting agent 5,6-dimethylxanthenone-4-acetic acid. J. Immunol. 190: 5216–5225.

The Journal of Immunology 2403

by guest on August 31, 2022

http://ww

w.jim

munol.org/

Dow

nloaded from

40. Gao, P., M. Ascano, T. Zillinger, W. Wang, P. Dai, A. A. Serganov,B. L. Gaffney, S. Shuman, R. A. Jones, L. Deng, et al. 2013. Structure-functionanalysis of STING activation by c[G(29,59)pA(39,59)p] and targeting by antiviralDMXAA. Cell 154: 748–762.

41. Prantner, D., and U. M. Nagarajan. 2009. Role for the chlamydial type III se-cretion apparatus in host cytokine expression. Infect. Immun. 77: 76–84.

42. Stetson, D. B., and R. Medzhitov. 2006. Recognition of cytosolic DNA activatesan IRF3-dependent innate immune response. Immunity 24: 93–103.

43. Roux, C. M., H. G. Rolan, R. L. Santos, P. D. Beremand, T. L. Thomas,L. G. Adams, and R. M. Tsolis. 2007. Brucella requires a functional type IVsecretion system to elicit innate immune responses in mice. Cell. Microbiol. 9:1851–1869.

44. Stanley, S. A., J. E. Johndrow, P. Manzanillo, and J. S. Cox. 2007. The type I IFNresponse to infection withMycobacterium tuberculosis requires ESX-1-mediatedsecretion and contributes to pathogenesis. J. Immunol. 178: 3143–3152.

45. Crimmins, G. T., A. A. Herskovits, K. Rehder, K. E. Sivick, P. Lauer,T. W. Dubensky, Jr., and D. A. Portnoy. 2008. Listeria monocytogenes multidrugresistance transporters activate a cytosolic surveillance pathway of innate im-munity. Proc. Natl. Acad. Sci. USA 105: 10191–10196.

46. Wilson, D. P., P. Timms, D. L. McElwain, and P. M. Bavoil. 2006. Type IIIsecretion, contact-dependent model for the intracellular development of chla-mydia. Bull. Math. Biol. 68: 161–178.

47. DeMars, R., and J. Weinfurter. 2008. Interstrain gene transfer in Chlamydiatrachomatis in vitro: mechanism and significance. J. Bacteriol. 190: 1605–1614.

48. Dumoux, M., D. K. Clare, H. R. Saibil, and R. D. Hayward. 2012. Chlamydiaeassemble a pathogen synapse to hijack the host endoplasmic reticulum. Traffic13: 1612–1627.

49. Giles, D. K., and P. B. Wyrick. 2008. Trafficking of chlamydial antigens to theendoplasmic reticulum of infected epithelial cells. Microbes Infect. 10: 1494–1503.

2404 cGAS IN CHLAMYDIA-INDUCED IFN-b

by guest on August 31, 2022

http://ww

w.jim

munol.org/

Dow

nloaded from