A Novel Mitochondrial MAVS/Caspase-8 Platform Links RNA Virus−Induced Innate Antiviral Signaling to Bax/Bak-Independent This information is current as of October 2, 2021. Souhayla El Maadidi, Laura Faletti, Birgit Berg, Christin Wenzl, Katrin Wieland, Zhijian J. Chen, Ulrich Maurer and Christoph Borner J Immunol 2014; 192:1171-1183; Prepublished online 3

January 2014; Downloaded from doi: 10.4049/jimmunol.1300842 http://www.jimmunol.org/content/192/3/1171

Supplementary http://www.jimmunol.org/content/suppl/2014/01/03/jimmunol.130084 http://www.jimmunol.org/ Material 2.DCSupplemental References This article cites 54 articles, 21 of which you can access for free at: http://www.jimmunol.org/content/192/3/1171.full#ref-list-1

Why The JI? Submit online.

by guest on October 2, 2021 • Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• 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 © 2014 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

A Novel Mitochondrial MAVS/Caspase-8 Platform Links RNA Virus–Induced Innate Antiviral Signaling to Bax/Bak-Independent Apoptosis

Souhayla El Maadidi,*,† Laura Faletti,*,† Birgit Berg,*,†,1 Christin Wenzl,*,†,2 Katrin Wieland,* Zhijian J. Chen,‡ Ulrich Maurer,*,x,{ and Christoph Borner*,x,{

Semliki Forest virus (SFV) requires RNA replication and Bax/Bak for efficient apoptosis induction. However, cells lacking Bax/Bak continue to die in a caspase-dependent manner. In this study, we show in both mouse and human cells that this Bax/Bak-independent pathway involves dsRNA-induced innate immune signaling via mitochondrial antiviral signaling (MAVS) and caspase-8. Bax/Bak- deficient or Bcl-2– or Bcl-xL–overexpressing cells lacking MAVS or caspase-8 expression are resistant to SFV-induced apoptosis. The signaling pathway triggered by SFV does neither involve death receptors nor the classical MAVS effectors TNFR-associated factor-2, IRF-3/7, or IFN-b but the physical interaction of MAVS with caspase-8 on mitochondria in a FADD-independent manner. Downloaded from Consistently, caspase-8 and -3 activation are reduced in MAVS-deficient cells. Thus, after RNA virus infection MAVS does not only elicit a type I antiviral response but also recruits caspase-8 to mitochondria to mediate caspase-3 activation and apoptosis in a Bax/Bak-independent manner. The Journal of Immunology, 2014, 192: 1171–1183.

emliki Forest virus (SFV) is an enveloped, single-stranded, anticancer therapy and vaccination (5). It is therefore crucial to and positive-sense RNA virus belonging to the Alphavirus elucidate the exact molecular mechanism by which SFV induces http://www.jimmunol.org/ S genus of the Togaviridae family. Its pathogenicity in humans apoptosis in host cells. and animals varies from asymptomatic to fatal, causing encephalitis We previously showed that apoptosis induced by SFV is RNA or epidemic polyarthritis, depending on the virulence of the strain replication dependent and involves the intrinsic mitochondrial (1, 2). The virus has a broad host range, replicates to high titers, and signaling pathway, in particular Bak (4). Bak and Bax are required induces apoptosis of numerous mammalian cells (3, 4). Importantly, for mitochondrial outer membrane permeabilization (MOMP) and its strong death- and type I IFN–inducing capacity has spurred the the subsequent release of cytochrome c, which then activates caspase- development of various SFV vectors that are currently used for 3 via the Apaf-1/caspase-9 apoptosome (6). Effective MOMP de- pends on so-called BH3-only that sense the apoptotic by guest on October 2, 2021 stimulus and either directly bind to and activate Bax/Bak or in- *Institute of Molecular Medicine and Cell Research, Albert Ludwigs University † Freiburg, D-79104 Freiburg, Germany; Faculty of Biology, Albert Ludwigs Univer- teract with the Bcl-2–like survival factors Bcl-2, Bcl-xL, or Mcl-1 sity Freiburg, D-79104 Freiburg, Germany; ‡Department of Molecular Biology, to release Bax/Bak from inhibitory constraints (6). Howard Hughes Medical Institute, University of Texas Southwestern Medical Center, Dallas, TX 75390; xSpemann Graduate School of Biology and Medicine, Albert Interestingly, cells infected with SFV or other alphaviruses such Ludwigs University Freiburg, D-79104 Freiburg, Germany; and {Centre for Biolog- as Sindbis continue to die in an apoptotic manner when Bax/Bak ical Signaling Studies, D-79104 Freiburg, Germany are depleted (4) or when Bcl-2 is overexpressed (3, 7, 8). This 1Current address: greenovation Biotech, Freiburg, Germany. behavior is typical for apoptosis induced by death receptor sig- 2Current address: U3Pharma, Martinsried, Germany. naling (9). However, we were unable to inhibit SFV-induced ap- Received for publication March 29, 2013. Accepted for publication November 25, optosis by blocking TNF-a, (FasL), TRAIL, and/or 2013. their receptors with neutralizing Abs (4). Nevertheless, caspase-8, This work was supported by Spemann Graduate School of Biology and Medicine a crucial mediator of death receptor signaling clearly was processed Grant GSC-4 and Centre for Biological Signalling Studies Grant EXC-294 (to C.B. and U.M., both funded by the Excellence Initiative of the German Federal and State and activated in SFV-infected cells (4, 10), and previous studies Governments, Germany). L.F. was supported by the Virtual Liver Network, spon- showed that the caspase-8 inhibitors cytokine response modifier sored by the German Federal Ministry of Education and Research, and Z.J.C. was A or FADD-like IL-1 b–converting enzyme-inhibitory (FLIP) supported by National Institutes of Health Grant AI-093967. were able to interfere with alphavirus-induced apoptosis (11, 12). Address correspondence and reprint requests to Dr. Christoph Borner, Institute of Molecular Medicine and Cell Research, Albert Ludwigs University Freiburg, Stefan This pointed to a caspase-8–dependent but Bax/Bak- and death Meier Strasse 17, D-79104 Freiburg, Germany. E-mail address: christoph.borner@ receptor–independent apoptosis signaling pathway triggered by SFV. uniklinik-freiburg.de RNA viruses induce an antiviral type I IFN response in infected The online version of this article contains supplemental material. host cells (13). After receptor-mediated endocytosis, the viruses Abbreviations used in this article: DISC, death-inducing signaling complex; eIF2a, release their genomic RNA into the cytoplasm where it replicates, eukaryotic initiation factor 2 a; FasL, Fas ligand; FLIP, FADD-like IL-1 b–convert- producing dsRNA as a byproduct (14). Cytoplasmic viral dsRNA ing enzyme-inhibitory protein; HCV, hepatitis C virus; IFNAR, IFN-a/b receptor; MAVS, mitochondrial antiviral signaling; MDA-5, melanoma differentiation Ag-5; is sensed by a class of ubiquitous cytoplasmic RNA helicases, reti- MEF, mouse embryonic fibroblast; MOI, multiplicity of infection; MOMP, mitochon- noic acid inducible -I (RIG-I) (15), and melanoma differentia- drial outer membrane permeabilization; OAS, oligoadenylate synthetase; PARP, poly (ADP-ribose) polymerase; PI, propidium iodide; PKR, protein kinase R; PLA, prox- tion Ag-5 (MDA-5) (16), which trigger an antiviral signaling imity ligation assay; RIG-I, retinoic acid inducible gene-I; SFV, Semliki Forest virus; cascade via their common adaptor called mitochondrial antiviral sh-Ctrl, scrambled control; shRNA, short hairpin RNA; TRAF, TNFR-associated signaling (MAVS) (also called IFN-b stimulator 1, virus-induced factor; TX, Triton X-100; WT, wild-type. signaling adapter, or caspase activation and recruitment domain Copyright Ó 2014 by The American Association of Immunologists, Inc. 0022-1767/14/$16.00 adaptor inducing IFN-b) (17–20). This leads to the production www.jimmunol.org/cgi/doi/10.4049/jimmunol.1300842 1172 SFV APOPTOSIS BY MAVS/CASPASE-8 INDEPENDENT OF Bax/Bak of type I IFNs (IFN-a/b) via NF-kB and IRF-3/-7 transcription were grown in high-glucose DMEM (4.5 g/l glucose) supplemented with factors (13). IFN-a/b then induce, via a common cell surface IFN- 10% FCS. MEFs were obtained from different sources (see Acknowl- a/b receptor (IFNAR) and the JAK-STAT signaling pathway, the edgments); the other cells were purchased from American Tissue Cell Collection. expression of IFN-inducible, dsRNA-activated protein kinase R (PKR) and 29-59-oligoadenylate synthetase (2-5-OAS) (21, 22). Both SFV production and titration enzymes are strongly activated by dsRNA (23, 24). Although PKR The virulent SFV prototype strain L10 was used in this study. Adherent blocks protein synthesis by phosphorylating eukaryotic initiation mosquito A. albopictus cells were infected with 20 multiplicity of infection factor 2 a (eIF2a), the 29-59-oligoadenylates generated by OAS ac- (MOI) SFV in L15 medium supplemented with 10% FCS and 4% Difco tivate the latent endonuclease RNase L to degrade viral RNAs (24). Bacto phosphate tryptose broth. After incubating at 28˚C for 24 h, the viral Although overexpression of PKR was shown to induce apoptosis supernatant was harvested, centrifuged at 3000 3 g for 10 min, and stored 2 (25, 26) and PKR2/2 mouse embryonic fibroblasts (MEFs) were at 80˚C. Virus titers from the culture medium of infected MEFs (10 MOI) were determined by the plaque assay. Briefly, monolayers of Vero resistant to cell death induced by dsRNA and LPS (27), the role of cells with .90% confluency were infected with 10-fold serial dilutions of PKR in the context of RNA virus–induced apoptosis has remained virus in high-glucose DMEM supplemented with 2% FCS and 20 mM enigmatic (23). In fact, in response to SFV, PKR suppressed viral HEPES at 37˚C for 90 min. The inoculum was removed and replaced with production and enhanced the type I IFN response rather than an overlay of 0.4% noble agar (Difco; BD Diagnostics). After 48-h in- 2/2 cubation at 37˚C, the agar was removed, the monolayer was stained with inducing apoptosis as PKR cultures died even faster than wild- 1% crystal violet, and the plaques were counted. type (WT) controls (28). Also, the role of 2-5-OAS in alphavirus- induced apoptosis has remained unclear (29). By contrast, recent Cell death assays studies have implicated MAVS in apoptosis induction by Sendai

SV40-transformed MEFs, human HeLa or HEK293 cells were infected with Downloaded from virus, dengue virus, and reovirus (30–33), by dsRNA/polyinosinic- 10 MOI SFV at 80% confluence while shaking in DMEM plus 0.5% FCS. polycytidylic acid transfection (34), and by cell detachment–induced After 1 h at 37˚C, viral infection was stopped, and the cells were incubated apoptosis (anoikis) (35). However, the mechanism by which MAVS in DMEM plus 10% FCS until processed for further experiments. Alter- performs this function and how it links to Bax/Bax activation has natively, the cells were treated with 50 ng/ml FasL from the supernatant of Neuro 2A cells (37) or 100 ng/ml Flag-FasL cross-linked with anti-Flag remained elusive. Ab M2 as previously reported (39) or irradiated with UV light (100 J/m2) In this study, we show that SFV requires MDA-5 and MAVS but in a UV Stratalinker. Apoptosis was quantified by His-GFP-annexin-V/PI not TNFR-associated factor (TRAF)-2, IRF-3/7, IFN-b, IFNAR, FACS analysis, and caspase-3/-7 activity was measured by the DEVDase http://www.jimmunol.org/ PKR, or RNase L for apoptosis induction. This proapoptotic path- assay as described previously (36). Fluorescence was detected in the Fluo- way runs independently of Bax/Bak-induced MOMP and involves roskan Ascent equipment (Thermo Labsystems), and the relative fluores- cence units were normalized to the protein concentration. a SFV-induced recruitment of caspase-8 to MAVS on mitochondria where caspase-8 is activated and required to cleave and activate Total protein extraction and subcellular fractionation caspase-3. For total extracts, cells were directly lysed in buffer A (25 mM HEPES KOH [pH 7.4], 2 mM MgCl2, and 2 mM EGTA) containing 1% TX and protease Materials and Methods inhibitors, and the lysate was left on ice for 20 min before spinning down at 4˚C and 13,000 rpm for 5 min. The supernatant contained the cytosolic Reagents and Abs by guest on October 2, 2021 and TX-soluble membrane proteins. For subcellular fractionation into Rabbit polyclonal anti–caspase-3 Abs recognizing the 32-kDa proform cytosol and crude mitochondria, the cells were pelleted, resupended in (9661) and the cleaved active 17-kDa form (9662), rabbit polyclonal anti– Mannitol-Sucrose-HEPES buffer (210 mM mannitol, 70 mM sucrose, 20 Bcl-xL (54H6), rabbit polyclonal anti-mouse, and human MAVS (4983 and mM HEPES [pH 7.5], 1 mM EDTA, and complete protease inhibitors), and 3993); and rabbit monoclonal anti–phospho-eIF2a (pS51, 3597) and rabbit incubated on ice for 30 min. Afterward, the cells were lysed using a sy- polyclonal anti–poly(ADP-ribose) polymerase (PARP) (9542) were pur- ringe with a 23- to 27-gauge needle until 50% of the cells were broken chased from Cell Signaling Technology. Rat monoclonal anti–caspase-8 (trypan blue positive). The nuclei were removed by centrifugation at 500 3 g, Ab (1G12) were from Alexis Biochemicals Enzo Life Sciences, mouse and a crude mitochondria (heavy membrane) fraction was obtained by an monoclonal anti–Bcl-2 (10C4) from Invitrogen, mouse monoclonal anti- additional centrifugation step at 13,000 3 g. The mitochondrial fraction actin (clone C4) from BD Biosciences, mouse monoclonal anti-ATPase was washed twice in Mannitol-Sucrose-HEPES buffer and then resus- (7H10) from Molecular Probes, mouse monoclonal anti-dsRNA J2 from pended in buffer A containing 1% SDS. The postmitochondrial supernatant English and Scientific Consulting (14), mouse monoclonal anti-PKR and was centrifuged at 100,000 3 g, 4˚C for 60 min. The resulting supernatant rabbit polyclonal anti-RNase L (H-300) from Santa Cruz Biotechnology, contains the cytosolic proteins. For the isolation of high-purity plasma rat monoclonal anti-Fas (7C10) and mouse monoclonal anti-Na+/K+ membrane protein fractions, we used the Qproteome Plasma Membrane ATPase a-1 (clone 464.6) from Millipore, and mouse monoclonal anti- Protein kit of Qiagen, according to the protocol of the manufacturer. cytochrome c (clone 6H2.B4), HRP-conjugated anti-rabbit, anti-rat, or Protein concentrations were determined by the Bradford assay. anti-mouse secondary Abs from Jackson ImmunoResearch Laboratories. Propidium iodide (PI), BSA, Triton X-100 (TX), mouse monoclonal anti- SDS-PAGE and Western blot analysis FLAG M2 Abs, and the 33 FLAG peptide were bought from Sigma- Aldrich. Z-VAD.fmk was obtained from MP Biomedicals, acrylamide Equal amounts of protein were separated on reducing or nonreducing 15% and DTT from AppliChem, and PageRuler Prestained Protein Ladder from SDS-PAGE and transferred to nitrocellulose membranes. The membranes 3 Thermo Scientific. His-GFP-Annexin-V was generated as described pre- were blocked in 1 PBS supplemented with 5% milk and 0.05% Tween 20 viously (36). FLAG-tagged FasL was purchased from Axxora. Alterna- and incubated with primary Abs (see supplemental information) at 4˚C tively, FasL was used as a multimerized, exosomal form (N2A FasL) secreted overnight. After three washings in PBS, secondary HRP-conjugated Abs from Neuro2A cells stably transfected with a mouse FasL expression vector were added and the membranes incubated at room temperature for 90 min. and provided by A. Fontana (University Clinic of Zurich, Zurich, Switzer- Proteins were visualized with the ECL SuperSignal West Pico Chemilu- land) (37). It was quantified and used for apoptosis assays as described minescent Substrate system (Pierce). previously (38). The mouse monoclonal anti-FADD Ab (clone 7A2) was provided by A. Strasser (Walter and Eliza Hall Institute, Parkville, VIC, Immunoprecipitations Australia). To study the MAVS death-inducing signaling complex (DISC), 500 ml(1 2/2 Cells mg) TX-solubilized mitochondrial membrane extract from WT, Bax/Bak , MAVS2/2,orBax/Bak2/2;sh-MAVS MEFs infected with SFV for 0–14 h or The insect cell line Aedes albopictus was maintained at 28˚C in L15 treated with Flag-FasL/anti-Flag M2 for 0–12 h was precleared with 100 ml medium (Life Technologies) supplemented with 10% FCS and 4% Difco 50% slurry protein G–Sepharose 4 Fast Flow recombinant protein G beads Bacto phosphate tryptose broth. SV40-transformed and primary MEFs, (GE Healthcare Bio-Sciences) on a turning wheel at 4˚C for 1 h. After the baby hamster kidney cells (BHK-21), and Vero, HeLa, and HEK293 cells supernatants were incubated with 5 ml polyclonal rabbit anti-MAVS on ice The Journal of Immunology 1173 for 1 h, 50 ml 50% protein G–Sepharose beads was added, and the mixture CCGGCGGGAAGGCATAAACATA-39; scrambled mouse SHC002, 59- was rotated at 4˚C for 2 h. All beads were centrifuged at 8200 3 g,4˚C CAACAAGATGAAGAGCACCAA-39; and scrambled human SHC007, 59- for 3 min and washed three times with lysis buffer, and immune complexes CCGGCGCTGAGTACTTCGAAATGTCCTCGAGGACATTTCGAAGTA- were eluted by boiling in Laemmli buffer. The eluted samples were run on CTCAGCGTTTTT-39. nonreducing SDS-PAGE and subjected to anti-MAVS, anti–caspase-8, or anti-FADD Western blot analysis. To study the Fas DISC, 500 ml (500 mg) RNA extraction and RT-PCR TX-solubilized plasma membrane extract from WT MEFs was directly Total RNA was isolated from mock- or SFV-infected WT or knockout MEFs incubated with 50 ml 50% protein G–Sepharose beads on a shaker at 4˚C by the Qiagen RNA easy Mini Kit, according to the manufacturer’s protocol for 2 h to pull down the Flag-FasL-Fas receptor complex performed by (Qiagen). Five micrograms of total RNA was reverse transcribed into cDNA anti-Flag cross-linking. After the beads were centrifuged and washed, the using the Invitrogen Superscript First Strand Synthesis Kit (Invitrogen) and complex was eluted by 50 ml33 Flag peptide. The eluted sample was run the therein provided random hexamer primers as described by the manu- on nonreducing SDS-PAGE and subjected to anti-Fas, anti–caspase-8, and facturer. To digest the leftover RNA, 1 ml RNase H was added and incubated anti-FADD Western blot analysis. at 37˚C for 20 min. Two microliters of the reverse transcriptase–generated cDNA samples was mixed with 18 ml PCR mix containing 1 mlofeach Immunofluorescence forward and reverse primer for the candidate gene (10 mM stock) and the MEFs were permeabilized with 0.2% TX for 4 min, washed in 13 PBS, and Hotstar meteor Taq polymerase. As a negative control, H2O was used, and blocked in 0.05% saponin and 1% BSA for 10 min. The cells were then for normalization, the housekeeping gene actin was coamplified. After 35 incubated for 45 min with mouse anti-cytochrome c (1:50) and rabbit anti- cycles of amplification in a Thermal Cycler PTC-225 (MJ Research Bio- 3 cleaved caspase-3 (1:200), followed by secondary anti-mouse Alexa 488 zym), the PCR products were supplemented with 6 Orange Loading Dye and anti-rabbit Alexa 546 Abs, diluted 1:200 in blocking buffer for 1 h. For and subjected to 2% agarose gel electrophoresis. PCR products were visu- some analysis, the anti-dsRNA Ab was used at a dilution of 1:200. The alized under the UV light. The primers for RT-PCR were as follows: b-actin coverslips were washed in PBS and mounted in Mowiol containing DAPI (forward), 59-TGGCGCTTTTGACTCAGGAT-39; b-actin (reverse), 59-AG-

(Molecular Probes) to stain the cell nuclei. Confocal microscopy was CCCTGGCTGCCTCAAC-39;IFN-b (forward), 59-CATCAACTATAAGCA- Downloaded from carried out with a laser scanning microscope (LSM 710; Carl Zeiss). GCTCCA-39;andIFN-b (reverse), 59-TTCAAGTGGAGAGCAGTTGAG-39. The primers for genotyping were as follows: RIG-I (forward), 59-GCAT- Proximity ligation assay CATCTCTCAGCTGATGAAGGAG A-39; RIG-I (reverse), 59-CCTTACA- CTTTAGGACCCATAGTGG-39; MDA-5 (forward), 59-CTCTTCTAAGCGTT- The in situ proximity ligation assay (PLA) is a method recently developed CCCTGGCTAGTGT-39; MDA-5 (reverse), 59-CTTGGGAAACAGCTCAGT- by OLink Bioscience (Uppsala, Sweden) to detect intracellular protein– AAAACTG-39; MAVS (forward), 59-TAGCTGTGAGGCAGGACAGGTAA- protein interactions (in-cell colocalization or coimmunoprecipitation). The GG-39;andMAVS(reverse),59-AGCCAAGATTCTAGAAGCTGAGAA-39. interaction signal is enhanced by rolling circle amplification of the an- http://www.jimmunol.org/ nealed probes and finally visualized by adding fluorescent oligonucleo- Statistics tides. In this study, the Duolink II Red starter kit was used. A total of 5 3 4 Statistical significance (p values) was analyzed by a two-tailed Student t 10 mock or SFV-infected WT MEFs were cytospun at 1200 rpm for 10 test. Data are the means of at least three experiments 6 SEM. min (Shandon Cytospin 4) onto microscopy slides in a defined 6-mm circular area. Cells were fixed in 100 ml 4% paraformaldehyde for 7 min at room temperature, washed twice in 13 PBS, and permeabilized with 0.1% Results TX for 10 min at 37˚C. After incubating in 40 ml blocking solution pro- SFV infection triggers dsRNA production, cytochrome c vided by the kit, the cells were exposed to rabbit anti-MAVS (or rabbit release, and caspase-8 and -3 activation anti–Bcl-xL as negative control) and rat anti–caspase-8 primary Abs

(diluted 1:100 each) or buffer only (negative control) at 37˚C for 2 h. The We first monitored the kinetics of SFV-induced apoptosis sig- by guest on October 2, 2021 samples were washed, incubated with the anti-rabbit and anti-rat secondary naling in MEFs. Between 4 and 8 h postinfection with 10 MOI SFV, Abs (PLA probes) from the kit, and amplified as described in the Duolink we detected dsRNA production (Fig. 1A) concomitant with the re- II protocol. After washing in 0.01% buffer B, the samples were air-dried, mounted, sealed, and subjected to Axiovert fluorescence microscopy. lease of cytochrome c from mitochondria into the cytosol (Fig. 1A, Images were captured with the help of the AxioVision Rel. 4.8 software. 1C). From 8 h on, caspase-8 and caspase-3 were cleaved into their fully processed, active forms in the cytosol (Fig. 1C). Moreover, RNA interference and Bcl-x or Bcl-2 overexpression L cells that had released cytochrome c costained with an Ab against Lentiviral short hairpin RNAs (shRNAs) to knock down mouse PKR, RNase active caspase-3 and showed nuclear condensation/fragmentation L, FADD, or mouse/human caspase-8 or MAVS were obtained from Sigma typical of apoptosis (Fig. 1A, white arrows). Interestingly, during Open Labs. As a control, scrambled shRNAs were used (mouse Ctrl SHC002 or human Ctrl SHC007 luciferase). To produce lentiviruses, 70–80% con- this time, low amounts of procaspase-8 and -3 seemed to translo- fluent HEK293T cells were transfected with a combination of 1.5 mgpMD2. cate to mitochondria (Fig. 1C, right panels, also see Fig. 6A, 6B). G encoding the viral envelope, 3 mg pSPAX2 packaging plasmid, and 5 mg Thus, SFV triggers rapid dsRNA production associated with MOMP lentiviral vector encoding the shRNA of interest, using the SuperFect and subsequent caspase-3 activation and apoptosis induction. transfection reagent (Qiagen). After 3 d of culturing in 5 mM sodium bu- tyrate, the lentivirus-containing medium was removed and passed through SFV infection launches both Bax/Bak-dependent and 2 2 a0.45-mm filter. A total of 2 3 105 WT or Bax/Bak / MEFs or HeLa cells -independent signaling pathways to activate caspase-3 and were infected with 400 ml shRNA lentirviral supernatants in the presence of apoptosis 5 mg/ml polybrene (Sigma-Aldrich) and centrifuged at 2000 rpm at room temperature for 10 min. One to 3 h postinfection, the cells were washed with MOMP is executed by the pore-forming Bcl-2 family members Bax PBS, cultured in full media for at least 18 h, and then selected with puro- and Bak (6). We therefore studied SFV-induced apoptosis in Bax/ mycin (Sigma-Aldrich) for stable expression of the shRNA construct of 6 2/2 Bak single- and double-knockout MEFs. As expected, SFV-infected interest. For Bcl-xL or Bcl-2 overexpression, 1 3 10 WT or caspase-8 2/2 c MEFs or HeLa or HEK293 cells were transfected with 5 mghumanBcl-xL/ Bax/Bak MEFs retained cytochrome in mitochondria in a pcDNA3 or mouse Bcl-2/pcDNA3 plasmids (Invitrogen) using the SuperFect punctate pattern. However, they still displayed active caspase-3 transfection reagent, respectively. After 24 h, the cells were put in selection immunostaining (Fig. 1B, white arrows), caspase-3 processing (Fig. media containing G418 (PAA) and cultured for 5 d to obtain a mixed pop- 1D) and activation (Fig. 1E, left graph), PARP cleavage (Fig. 1D), ulation of cells stably overexpressing Bcl-xL or Bcl-2. Lentiviral shRNA sequences were as follows: mouse MAVS, 59-CCAGTGCTGATCTATTAG- as well as apoptosis induction (Fig. 1E, right graph), although at GAA-39;humanMAVS,59-CCGGATGTGGATGTTGTAGAGATTCCTCG- reduced levels as compared with WT cells (Fig. 1E). Moreover, this AGGAATCTCTACAACATCCACATTTTTTG-39; mouse caspase-8, 59-ACG- Bax/Bak-independent apoptosis was effectively blocked by the ACTGCACTGCAAATGAAA-39; human caspase-8, 59-CCGGGCCTTGAT- general caspase inhibitor Z-VAD.fmk (Fig. 2C, right graph), in- GTTATTCCAGAGACTCGAGTCTCTGGAATAACATCAAGGCTTTT-39; FADD1, 59-CCACACTTGGAGCCCAATAAAC-39; FADD2, 59-CGAGC- dicating that SFV activates a Bax/Bak-independent pathway of GCGTGAGCAAACGAAA-39; FADD3, 59-CCCAGGAATCTGTGAGCAA- caspase-3 activation and apoptosis. As expected, although UV- GA-39;PKR,59-GGAGACTTCTGAACAAGAGCAG-39;RNaseL,59- induced apoptosis was entirely Bax/Bak-dependent (Fig. 1E, 1174 SFV APOPTOSIS BY MAVS/CASPASE-8 INDEPENDENT OF Bax/Bak

right graph), FasL triggered caspase-3 and PARP cleavage (Figs. 1D, 2A) and caspase-3 activation (Fig. 1E, left graph) to similar extents in WT and Bax/Bak2/2 cells because it did not require the intrinsic mitochondrial pathway in MEFs (type I signaling) (9). Caspase-8 is required for SFV-induced caspase-3 activation and apoptosis in the absence of Bax/Bak or when Bcl-xL or Bcl-2 are overexpressed We assumed that caspase-8 was responsible for SFV-induced caspase-3 activation and apoptosis in Bax/Bak2/2 cells because it was processed/activated simultaneously with caspase-3 (Fig. 1C). We therefore downregulated caspase-8 in Bax/Bak2/2 MEFs by lentiviral shRNA transfer. Although we could not fully deplete caspase-8 with this technique (Fig. 2A, Supplemental Fig. 1A), SFV-infected Bax/Bak2/2;sh-Casp-8 MEFs lacked caspase-8 and caspase-3 processing and PARP cleavage (Fig. 2A) and cytosolic caspase-3/-7 (DEVDase) activity (Fig. 2B, right graph, black col- umns). Moreover, these cells almost fully were protected against SFV-induced apoptosis as compared with Bax/Bak2/2;sh-Ctrl cells (Fig. 2C, right graph, black columns). By contrast, UV-induced Downloaded from apoptosis was not influenced by caspase-8 levels (Fig. 2C, right graph). In contrast, as expected, FasL was still able to use the remaining levels of caspase-8 in Bax/Bak2/2;sh-Casp-8 to directly activate caspase-3 via the type I signaling pathway, although to a lesser extent (Fig. 2A, 2B, right graph). Importantly, a partial

requirement of caspase-8 for caspase-3 activation and apoptosis http://www.jimmunol.org/ after SFV infection already was seen when caspase-8 was down- regulated in WT MEFs (WT;sh-Casp-8) (Supplemental Figs. 1A, 2B, 2C, left graphs), indicating that SFV-triggered, caspase-8–me- diated caspase-3 activation occurs in parallel to the classical Bax/ Bak-dependent pathway but the latter seems to be dominant. To exclude a clonal effect of the Bax/Bak2/2 MEFs and to confirm the caspase-8–mediated effect on caspase-3 activation by another strategy, we generated caspase-82/2 MEFs overexpressing 2/2 by guest on October 2, 2021 Bcl-xL (Fig. 3A). As shown above for Bax/Bak cells, WT MEFs overexpressing Bcl-xL still showed caspase-3 processing (Fig. 3B) and cytosolic activity (Fig. 3C, left graph) as well as apoptosis induction (Fig. 3D, left graph) in response to SFV, although to a lower extent. Caspase-8 depletion partially diminished these parameters as well, but SFV could still trigger caspase-3 activation and apoptosis via Bax/Bak-mediated MOMP. By contrast, over- 2/2 expression of Bcl-xL in caspase-8 MEFs abrogated SFV-induced caspase-3 processing and activation (Fig. 3B, 3C, right graph,black columns) and effectively saved the cells from apoptosis (Fig. 3D, right graph, black columns). As expected, UV-induced apoptosis

(hpi). Active caspase-3 colocalizes with diffuse (A) or punctate, mito- chondrial (B) cytochrome c (see arrows). (C) Anti-cytochrome c, caspase-8 (recognizing both the proform and the cleaved p18 form), caspase-3, and cleaved caspase-3 Western blots of cytosolic and mitochondrial fractions of MEFs at 0–14 hpi. (D) Anti–caspase-3, cleaved caspase-3, and PARP Western blots of total extracts of SV40-transformed WT, Bax2/2, Bak2/2, or Bax/Bak2/2 MEFs infected with SFV for 0–36 h (hpi). As controls for effective caspase-8 and -3 processing and PARP cleavage, MEFs were treated with 50 ng/ml FasL (Neuro 2A supernatant) for 6 h. Anti-ATPase and -actin Western blots served as controls for equal loading of mitochondrial FIGURE 1. SFV provokes dsRNA formation and induces apoptosis by or cytosolic proteins, respectively. Molecular weight markers are shown on Bax/Bak-dependent cytochrome c release and caspase-3 activation as well the right. (E) Cytosolic caspase-3/-7 (DEVDase) activity assay (left graph)or as independently of Bax/Bak and cytochrome c release but still involving annexin-V/PI FACS analysis (right graph) of SV40-transformed WT or caspase-3. (A and B) Anti-dsRNA (red), cytochrome c (green), and cleaved Bax/Bak2/2 MEFs infected with SFV for 0–36 h (hpi), treated with 50 ng/ caspase-3 (red) immunofluorescence analysis of SV40-transformed WT ml FasL for 6 h, or exposed to 100 J/m2 UV light (UV) for 24 h. Data in (E) (A) and WT and Bax/Bak2/2 (B) MEFs infected with SFV for 0–8 h. are the means of at least three independent experiments 6 SEM. The Nuclei were stained with DAPI (blue). Original magnification 3400. In p values are the following: Bax/Bak2/2 versus WT cells, caspase-3/-7: p = (A), the middle and lower panels show the same cells, whereas the cells in 0.01 for SFV 6 h; p , 0.001 for SFV 10, 14, and 24 h. Bax/Bak2/2 versus the upper panel were different but taken at the same hour postinfection WT cells, apoptosis: p , 0.001 for SFV 14, 24, and 36 h and UV; for all n =4. The Journal of Immunology 1175

FIGURE 2. SFV-induced caspase-3 processing/activation and apoptosis in Bax/Bak2/2 MEFs is dependent on cas- pase-8. (A) Anti–caspase-8/cleaved caspase-8, caspase-3, cleaved caspase-3, and PARP Western blots of total extracts of SV40-transformed Bax/Bak2/2 MEFs stably expressing sh-Casp-8 or sh-Ctrl infected with SFV for 0–14 h (hour postinfection [hpi]) or treated with 50 ng/ml FasL for 6 h. Downloaded from Anti-actin immunoblots served as control for equal protein loading. Cytosolic caspase-3/-7 (DEVDase) activity assay (B) or annexin-V/PI FACS analysis (C) of SV40-transformed WT (left graphs)orBax/Bak2/2 MEFs (right graphs)ex- pressing sh-Casp-8 or sh-Ctrl (scrambled control), infected with SFV for 0–36 h (hpi), treated with 50 ng/ml FasL for 6 h http://www.jimmunol.org/ or exposed to 100 J/m2 UV light (UV) for 24 h. In (C), Bax/ Bak2/2 MEFs also were treated with 100 mM Z-VAD.fmk (dotted bars). Data in (B, C) are the means of at least three independent experiments 6 SEM. The p values are the fol- lowing: WT; sh-Casp-8 versus WT; sh-Ctrl cells (B): p = 0.02forSFV6hand10h;p = 0.01 for SFV 14 and 24 h and FasL. Bax/Bak2/2; sh-Casp-8 versus Bax/Bak2/2; sh- Ctrl cells (B): p , 0.001 for SFV 6, 10, 14, and 24 h; p = 0.003 for FasL. WT; sh-Casp-8 versus WT; sh-Ctrl cells (C): 2/2 p = 0.01–0.005 for SFV 14, 24, and 36 h. Bax/Bak ; sh- by guest on October 2, 2021 Casp-8 and Bax/Bak2/2; Z-VAD.fmk versus Bax/Bak2/2; sh-Ctrl cells (C): p , 0.001 for SFV 24 and 36 h and UV; for all n =4.

2/2 still occurred in caspase-8 cells but not when Bcl-xL was over- show that SFV induces caspase-3 activation and apoptosis via both expressed (Fig. 3D), and FasL-induced caspase-3 activation en- Bax/Bak-dependent and Bax/Bak-independent, caspase-8–dependent tirely depended on caspase-8 but not on Bcl-xL overexpression processes in MEFs, HeLa, and HEK293 cells. (type I cells) (Fig. 3C). Similar results were obtained in two en- The antiviral signaling components PKR and RNase L are not tirely different cellular systems, human HeLa and HEK293 cells. 2/2 In this study, also stable overexpression of Bcl-2 (Supplemental required for SFV-induced apoptosis in WT and Bax/Bak cells Fig. 1B) could not prevent SFV-induced caspase-3 processing We wondered by which signaling pathway SFV induced Bax/Bak- (Supplemental Fig. 1C), caspase-3 activation (Supplemental Fig. independent caspase-3 activation and apoptosis. Positive-sense 1D, right graphs), and apoptosis (Supplemental Fig. 1E, right ssRNA viruses activate innate antiviral signaling leading to type I graphs), whereas the shRNA-mediated depletion of caspase-8 IFN-a/b induction and subsequent virus inhibition via PKR and (Supplemental Fig. 1B) in these cells drastically diminished all RNase L activation. Both PKR (23) and RNase L (29) have been these events (Supplemental Fig. 1C–E). Thus, our data clearly suggested to sensitize cells for virus-induced apoptosis. We therefore 1176 SFV APOPTOSIS BY MAVS/CASPASE-8 INDEPENDENT OF Bax/Bak

FIGURE 3. Bcl-xL overexpression effectively inhibits SFV-induced caspase-3 activation and apoptosis only when caspase-8 is deleted. (A) Anti–caspase-8 and Bcl-x Western blots showing the deletion of caspase-8 and the overexpression of Bcl-xL in SV40-transformed WT or caspase-82/2 MEFs. (B) Anti–caspase-3 and cleaved caspase-3

Western blots of total extracts of SV40-transformed WT or caspase- Downloaded from 2/2 8 MEFs overexpressing Bcl-xL or not infected with SFV for 0–36 h (hour postinfection [hpi]). Anti-actin immunoblots served as control for equal protein loading. Cytosolic caspase-3/-7 activity assay (C)or annexin-V/PI FACS analysis (D) of SV40-transformed WT (left graphs) or caspase-82/2 (right graphs) MEFs in the absence or presence of Bcl-xL overexpression, either infected with SFV for 0–24 h (hpi), treated with 50 ng/ml FasL for 6 h, or exposed to 100 J/m2 UV http://www.jimmunol.org/ light (UV) for 24 h. Data in (C, D) are the means of at least three independent experiments 6 SEM. The p values are the following: WT +

Bcl-xL versus WT cells (C): p =0.03forSFV6and10h;p =0.01 2/2 2/2 for SFV 14 and 24 h. Casp-8 + Bcl-xL versus Casp-8 cells (C): p , 0.001 for SFV 6, 10, 14, and 24 h. WT + Bcl-xL versus WT cells (D): p = 0.01 for SFV 14 and 24 h; p = 0.03 for SFV 36 h; p , 0.001 2/2 2/2 for UV. Casp-8 + Bcl-xL versus Casp-8 cells (D): p , 0.001 for SFV 14, 24, and 36; for all n =4. by guest on October 2, 2021

tested whether these enzymes also were involved in SFV-induced bled control (sh-Ctrl), sh-PKR, sh-RNase L, or both sh-PKR and sh- apoptosis. SV40-transformed PKR2/2, RNase L2/2, and PKR/ RNase L (Supplemental Fig. 2E). This finding shows that neither RNase L2/2 MEFs clearly lacked PKR and/or RNase L expres- PKR nor RNase L is required for Bax/Bak-dependent or -inde- sion (Supplemental Fig. 2A), and SFV-induced eIF2a phosphor- pendent apoptosis induced by SFV. ylation, critical for PKR-mediated protein synthesis inhibition, was inhibited effectively in the PKR/RNase L2/2 cells (Supple- MAVS is crucial for both SFV-induced type I IFN induction and mental Fig. 2B). However, neither the absence of PKR or RNase L Bax/Bak-independent caspase-8 and caspase-3 activation and nor both had any delaying effect on SFV-induced apoptosis (Sup- apoptosis plemental Fig. 2C). Because a proapoptotic effect of PKR and/or To investigate whether the RIG-I/MDA-5 dsRNA sensing pathway RNase L may only be revealed in the absence of the dominant Bax/ was involved in SFV-induced apoptosis, we infected RIG-I2/2 and Bak-mediated signaling pathway, we knocked down PKR and/or MDA-52/2 MEFs with SFV for up to 36 h (Supplemental Fig. 3A). RNase L expression in Bax/Bak2/2 MEFs by shRNA either com- Although the lack of RIG-I did not have any impact on the sensi- pletely (PKR) or by 90% (RNase L), respectively (Supplemental tivity of SFV-induced apoptosis, MDA-5 deficiency significantly Fig. 2D). Despite that, the kinetics of SFV-induced apoptosis were delayed the death response at all time points (Supplemental Fig. indistinguishable between Bax/Bak2/2 cells expressing the scram- 3B). As expected, neither helicase was involved in apoptosis The Journal of Immunology 1177 induced by an unrelated death stimulus such as UV (Supplemental Similar to MDA-52/2 cells, both primary and SV40-transformed Fig. 3B). Because both RIG-I and MDA-5 mediate antiviral sig- MAVS2/2 MEFs were significantly delayed in SFV-induced apo- naling through MAVS, we analyzed SFV-induced apoptosis sen- ptosis (Fig. 4A) and showed diminished caspase-3 activities as sitivity in primary and SV40-transformed MAVS2/2 MEFs. Both compared with WT cells at all time points postinfection (Fig. 4B). cell lines clearly lacked MAVS expression (Supplemental Fig. 3C) However, the lack of MAVS did not affect UV-induced apoptosis or and did not show any IFN-b induction in response to SFV (Sup- FasL-induced caspase-3 activation, respectively (Fig. 4A, 4B). plemental Fig. 3D). This confirms that MAVS is the major me- Moreover, the kinetic of cytochrome c release was similar between diator of type I IFN expression after SFV infection. MAVS2/2 and WT (MAVS+/+) cells (Fig. 4C, 4D), indicating that Downloaded from http://www.jimmunol.org/ FIGURE 4. MAVS deficiency delays SFV-induced cas- pase-8 and -3 processing/activation and apoptosis but has no effect on cytochrome c release. Annexin-V/PI FACS analysis (A) or cytosolic caspase-3/-7 activity assay (B)of SV40-transformed or primary MAVS+/+ and MAVS2/2 MEFs, infected with SFV for 0–36 h (hour postinfection [hpi]), treated with 50 ng/ml FasL for 6 h, or exposed to 100 J/m2 UV light (UV) for 24 h. (C) Anti-cytochrome c Western blots of mitochondrial and cytosolic extracts of primary MAVS+/+ and MAVS2/2 MEFs infected with SFV by guest on October 2, 2021 for 0–14 h. (D) Anti-cytochrome c (green) and cleaved caspase-3 (red, white arrows) immunofluorescence analy- sis of primary MAVS+/+ and MAVS2/2 MEFs, mock-, or SFV-infected for 8 h. Original magnification 3400. (E) Anti–caspase-8/cleaved caspase-8, caspase-3, and cleaved caspase-3 Western blots of total extracts of primary MAVS+/+ and MAVS2/2 MEFs infected with SFV for 0– 14 h or treated with 50 ng/ml FasL for 6 h. Anti-ATPase and actin immunoblots served as controls for equal protein loading. Data in (A, B) are the means of at least three in- dependent experiments 6 SEM. The p values are the fol- lowing: Values MAVS2/2 versus MAVS+/+ cells (A): p = 0.01–0.02 for SFV 14, 24, and 36 h; not significant for SFV 36 h transformed. MAVS2/2 versus MAVS+/+ cells (B): p = 0.01–0.03 for SFV 6, 12, 14, and 24 h; for all n =3. 1178 SFV APOPTOSIS BY MAVS/CASPASE-8 INDEPENDENT OF Bax/Bak

MAVS regulated an apoptotic pathway different from Bax/Bak- (Supplemental Fig. 4E), and apoptosis (Supplemental Fig. 4F). mediated MOMP. Indeed, as shown in Fig. 4E, MAVS seemed to These data demonstrate that MAVS mediates SFV-induced caspase- control caspase-8–mediated caspase-3 processing and activation in 8 and caspase-3 activation and apoptosis in the absence of Bax/Bak response to SFV because both the appearance of the active p18 in three independent cellular systems. caspase-8 and p17 caspase-3 fragments were delayed significantly in MAVS2/2 MEFs. Again, no effect on the generation of these SFV, but not FasL triggers Bax/Bak-independent, MAVS- fragments was noted in response to FasL. Because it is known that dependent mitochondrial translocation of caspase-8 and the MAVS triggers an antiviral IFN-b response, we examined whether formation of a novel MAVS/caspase-8 DISC that does not apoptosis signaling by MAVS was through the downstream contain FADD adaptor protein TRAF-2, the transcriptional IFN-b inducers IRF-3 We envisaged the possibility that MAVS may activate caspase-8 by and IRF-7, IFN-b itself, or its receptor IFNAR. However, MEFs recruiting it from the cytosol to mitochondria. Although mitochondria- deficient in these mediators did not display any reduction of SFV- associated caspase-8 has been reported (40, 41), its predominant induced apoptosis (Supplemental Fig. 3E). sequestration and complex formation occurs at activated death Because we suspected that MAVS-mediated caspase-8 activation receptors in the plasma membrane. Formation of this complex, occurred independently of the Bax/Bak signaling pathway, we called DISC, requires the adaptor FADD, which binds to the cy- downregulated MAVS by shRNA in Bax/Bak2/2 MEFs. The drastic toplasmic side of activated death receptors via a and reduction of MAVS protein levels (Fig. 5A, Supplemental Fig. 4A, to caspase-8 via a death effector domain (9). We therefore com- left blot) almost completely ablated SFV-induced caspase-3 acti- pared the formation of the classical caspase-8/FADD DISC at the vation and apoptosis in Bax/Bak2/2 cells (Fig. 5B, 5C). Concom- plasma membrane in response to FasL treatment to a possible for- itantly, caspase-8 and caspase-3 processing as well as PARP mation of a novel MAVS/caspase-8 DISC on mitochondria after Downloaded from cleavage, which still occurred in the Bax/Bak doubly deficient SFV infection. In addition, we wanted to know whether the mito- cells, were entirely ablated when MAVS was downregulated (Fig. chondrial MAVS/caspase-8 DISC also could be formed in response 5D). Similar results were obtained in HeLa and HEK293 cells to FasL. As previously reported, MAVS constitutively localizes to overexpressing Bcl-2. The effective downregulation of MAVS by mitochondria and shows a double-band pattern on SDS-PAGE (Fig. shRNA (Supplemental Fig. 4A, middle and right blots) prevented 6A). Fas, in contrast, is known to be a cell surface receptor on the SFV-induced caspase-8 (Supplemental Fig. 4C) and caspase-3 plasma membrane (9). Next, we determined the change of subcel- http://www.jimmunol.org/ (Supplemental Figs. 4B, 4D) processing, caspase-3 activation lular distribution of caspase-8 in response to either SFV infection or by guest on October 2, 2021

FIGURE 5. SFV-induced, Bax/Bak-independent caspase-3 activation and apoptosis are mediated by MAVS. (A) Anti- MAVS Western blot showing efficient knockdown of MAVS expression (sh-MAVS) in SFV-infected, SV40-transformed Bax/Bak2/2 MEFs as compared with the sh-Ctrl. A total ex- tract of MAVS2/2 MEFs is shown as a control. Annexin-V/PI FACS analysis (B) and cytosolic caspase-3/-7 activity assay (C) of SV40-transformed Bax/Bak2/2 MEFs stably expressing a scrambled sh-Ctrl or sh-MAVS, infected with SFV for 0–36 h (hour postinfection [hpi]), treated with 50 ng/ml FasL for 6 h, or exposed to 100 J/m2 UV light (UV) for 24 h. (D) Anti– caspase-8/cleaved caspase-8, caspase-3, cleaved caspase-3, and PARP Western blots of total extracts of sh-Ctrl and sh-MAVS Bax/Bak2/2 MEFs infected with SFV for 0–36 h. Anti-actin immunoblots served as control for equal protein loading. Data in (B, C) are the means of at least three independent experi- ments 6 SEM. The p values are the following: Bax/Bak2/2; sh-MAVS versus Bax/Bak2/2;sh-Ctrl(B): p = 0.02 for SFV 14 h; p , 0.001forSFV24and36h.Bax/Bak2/2; sh-MAVS versus Bax/Bak2/2; sh-Ctrl (C): p , 0.001 for SFV 6, 12, 14, and 24 h; for all n =4. The Journal of Immunology 1179 the cellular treatment with a Flag-tagged FasL cross-linked with performed a PLA on mock and SFV-infected MEFs. No interaction anti-Flag Abs (39). Although the bulk of cytosolic p51 procaspase-8 between MAVS and caspase-8 could be detected in mock-infected rapidly redistributed to the plasma membrane after Flag-FasL treat- cells or when only secondary Abs were used for the PLA (Fig. ment and was processed into its active p43 form (Fig. 6B, top blot), 7D, right pictures). In addition, caspase-8 did not interact with an a minor portion of procaspase-8 accumulated on mitochondria in irrelevant protein such as Bcl-xL (Fig. 7D, middle pictures). How- response to SFV infection where it was slightly processed as well ever, at 8 h postinfection, anti-MAVS and anti–caspase-8 incubation (Fig. 6A, 6B, lower blot). No caspase-8 was detected on mito- revealed amplified red fluorescence signals (Fig. 7D, left pictures) chondria after FasL treatment, and SFV did not recruit any caspase- indicative of a close proximity of endogenous MAVS and caspase-8 8 to the plasma membrane (Fig. 6B). Importantly, the mitochondrial after viral infection. translocation of caspase-8 in response to SFV was dependent on To confirm that FADD did not play any role in caspase-8 ac- MAVS but independent of Bax/Bak, because it did not occur in tivation in the mitochondrial MAVSDISC, we downregulated FADD MAVS2/2 or Bax/Bak2/2;sh-MAVS but still in Bax/Bak2/2;sh- by shRNA in Bax/Bak2/2 MEFs. As shown in Fig. 8A–C, despite Ctrl MEFs (Fig. 6A). Moreover, it perfectly coincided with the a successful knockdown of FADD, the kinetics of SFV-induced time of SFV-induced caspase-8 and caspase-3 processing (Fig. 1C). caspase-3 activation and apoptosis in Bax/Bak2/2;sh-FADD cells This indicates that MAVS may attract caspase-8 to mitochondria were indistinguishable from the scrambled Bax/Bak2/2;sh-Ctrl cells. once it gets activated by the dsRNA-induced innate antiviral sig- By contrast, as expected, FADD depletion completely inhibited naling of SFV. This process is Bax/Bak-independent and cannot be FasL-induced caspase-3 activation and apoptosis (Fig. 8B, 8C). mimicked by FasL. Taken together, our findings suggest that although Fas induced Indeed, when we immunoprecipitated MAVS from the mito- a caspase-8/FADD DISC on the plasma membrane, SFV triggered chondria of SFV-infected MEFs, we found that both full-length and a novel MAVS/caspase-8 DISC on mitochondria that does not Downloaded from partially cleaved caspase-8 gradually associated with MAVS with include FADD but leads to caspsae-8 activation and downstream time of infection (Fig. 7A, 7C, right blots). Again, this interaction processing of caspase-3 independent of Bax/Bak. was dependent on MAVS but independent of Bax/Bak (Fig. 7A). Surprisingly, no FADD was found in these MAVS immunopreci- Discussion pitations, despite the fact that caspase-8 was even partially pro- In this study, we show that SFV, a positive-sense ssRNA virus, cessed to the active p43 form (cleaved Casp-8) (Fig. 7A, 7C, right stimulates the formation of a novel mitochondrial platform con- http://www.jimmunol.org/ blots). Conversely, after cells were treated with Flag-FasL cross- sisting of the innate immune signaling component MAVS and the linking, MAVS did not pull down caspase-8 from a mitochondrial initiator caspase-8. The purpose of this platform is to trigger fraction (Fig. 7C, left blots), whereas, expectedly, the plasma caspase-3 activation and apoptosis in an entirely Bax/Bak- and membrane Fas DISC immunoprecipitated with anti-Flag Abs con- death receptor/FADD–independent manner. The signaling path- tained caspase-8 and FADD (Fig. 7B). way leading to this event is initiated by dsRNA, transiently formed Because immunoprecipitations of membrane proteins may pro- during the RNA replication cycle of the virus. The dsRNA is im- duce artifacts because of the detergent and salt conditions chosen, we munodetected in the cytoplasm of infected cells before caspase-3 by guest on October 2, 2021

FIGURE 6. SFV but not FasL triggers mitochondrial translocation of caspase-8 with MAVS in a Bax/Bak-independent manner. (A) Anti-MAVS and caspase-8 Western blots of mitochondria (containing Bax/Bak) from SV40-transformed MAVS+/+ and MAVS2/2 MEFs (left blot) as well as from Bax/Bak2/2 MEFs stably expressing sh-MAVS or sh-Ctrl (right blot), infected with SFV for 0–24 h (hour postinfection [hpi]). (B) Anti–caspase-8 Western blots of cytosolic (Cyt), plasma membrane (PM), or mitochondrial (Mito) fractions of WT MEFs, either treated with 100 ng/ml Flag-FasL and anti-Flag Ab or infected with SFV for 0–8 h. A total of 250 ng Flag–FasL was cross-linked with 250 ng anti-Flag Ab. Anti-actin (43 kDa), anti-Na+/K+ ATPase a-1 (112 kDa), and anti- ATPase (50 kDa) immunoblots served as controls for equal protein loading of Cyt, PM, and Mito fractions, respectively. 1180 SFV APOPTOSIS BY MAVS/CASPASE-8 INDEPENDENT OF Bax/Bak

FIGURE 7. SFV but not FasL induces caspase-8/ MAVS association (MAVS DISC) on mitochondria in the absence of FADD and independent of Bax/Bak. (A) Anti-MAVS and caspase-8 Western blots of anti-MAVS immunoprecipitations from mitochondrial extracts of Downloaded from MAVS+/+ and MAVS2/2 cells (containing Bax/Bak, left blots) or sh-Ctrl and sh-MAVS Bax/Bak2/2 cells (right blots) infected with SFV for 0–14 h. IgL and asterisk mark cross-reactive L and H chains, respectively. (B) Anti-Fas, caspase-8, and FADD Western blots of 33 FLAG peptide-eluted anti-Flag immunoprecipitations from plasma membrane extracts of WT MEFs treated http://www.jimmunol.org/ with Flag-FasL and anti-Flag Ab (250 ng each) for 0– 12 h. (C) Anti-MAVS, anti–caspase-8, and anti-FADD Western blots of anti-MAVS immunoprecipitations from mitochondrial extracts of WT MEFs either treated with Flag-FasL/anti-Flag for 0–12 h (left blots) or infected with SFV for 0–24 h (hour postinfection [hpi]) (right blots). (D) PLA of endogenous MAVS and caspase-8 in mock and SFV-infected (8 hpi) MEFs, using rabbit polyclonal anti-MAVS (or rabbit polyclonal anti–Bcl-x by guest on October 2, 2021 as negative control) and mouse monoclonal anti–caspase- 8 primary Abs and the respective anti-rabbit and anti- mouse PLA probes or the PLA probes alone as con- trol (only secondary Ab). Original magnification 3400. CL Casp-8, Cleaved p43 caspase-8; FL Casp-8, full-length p51 procaspase-8.

activation, and cells deficient for MAVS or MDA-5, the preferred over, it is known that caspase-8 inhibits rather than activates nec- cytoplasmic receptor for dsRNA produced by positive-strand RNA roptosis (9). viruses, are partially protected from SFV-induced apoptosis. Im- We propose that the MAVS- and caspase-8–dependent apoptosis portantly, Bax/Bak2/2 cells dying in a MAVS/caspase-8–dependent signaling pathway runs in parallel with the canonical Bax/Bak- manner do not switch to an alternative death mode such as nec- regulated MOMP pathway for the following reasons: 1) SFV- 2/2 roptosis because they still show apoptotic features such as nuclear infected Bax/Bak and Bcl-xL overexpressing MEFs and Bcl-2 fragmentation and caspase-3 activation, and they are not protected overexpressing HEK293 and HeLa cells still show significant by the necroptosis inhibitor necrostatin-1 (data not shown). More- caspase-3 activation and apoptosis induction suggesting the existence The Journal of Immunology 1181

FIGURE 8. SFV-induced, Bax/Bak-independent caspase-3 activation and apoptosis do not require FADD. (A) Anti-FADD Western blot, showing efficient knockdown of FADD expres- sion in SFV-infected SV40-transformed Bax/Bak2/2 MEFs as compared with the sh-Ctrl using three different FADD shRNAs (sh-FADD1, shFADD2, and sh-FADD3). A total extract of FADD2/2 MEFs is shown for comparison. Anti-actin immu- noblots served as control for equal protein loading. Annexin-V/ PI FACS analysis (B) and cytosolic caspase-3/-7 activity assay (C) of SV40-transformed Bax/Bak2/2 MEFs stably expressing a scrambled sh-Ctrl or sh-FADD1, infected with SFV for 0–36 h (hour postinfection [hpi]) or treated with 50 ng/ml FasL for 6 h. Note that FADD depletion did not inhibit caspase-3 acti- vation and apoptosis induced by SFV but by FasL. Data in (B, C) are the means of at least three independent experiments 6 SEM. The p values are the following: Bax/Bak2/2; sh-FADD1 versus Bax/Bak2/2; sh-Ctrl (C, D): not significant for SFV 14, 24, and 36 h; p , 0.001 for FasL, n =4. Downloaded from

of another caspase-dependent pathway unaffected by MOMP. 2) also did not involve components activated by the IFNAR signaling Although caspase-82/2 and MAVS2/2 MEFs are retarded in cell such as PKR or RNaseL, which are additional targets of dsRNA death, they are not saved from SFV-induced caspase-3 activation previously implicated in RNA virus-induced apoptosis (23, 29). http://www.jimmunol.org/ and apoptosis. 3) Effective death protection is only achieved when Rather, the MAVS-mediated apoptotic pathway induced by SFV 2/2 MAVS or caspase-8 are depleted in Bax/Bak or Bcl-xL over- was independent of Bax/Bak and any type I IFN signaling com- expressing MEFs or Bcl-2 overexpressing HEK293 or HeLa cells. ponents downstream of MAVS but required caspase-8. Our finding therefore reveals that RNA viruses such as SFV can Caspase-8 is an obligatory component of death receptor sig- still kill their target cells via a MAVS/caspase-8 pathway when the naling and its activation requires a protein platform called the DISC major MOMP pathway is perturbed. This may be relevant for (9). Could caspase-8 also be involved in other signaling pathways eliminating cells in which Bax or Bak are defective or Bcl-2 forming other DISCs? We confirm that FasL induces the trans- survival factors are overexpressed such as in or autoim- location of most of cytosolic procaspase-8 to the FasR on the by guest on October 2, 2021 mune cells (6). Indeed, we previously reported that Bcl-2 over- plasma membrane where it gets processed and activated. SFV expressing cells are effectively killed by SFV because Bcl-2 is infection, in contrast, triggers the association of a minor part of cleaved by caspase-3 at its unstructured loop region, generating procaspase-8 with mitochondria where it is also processed and a N-terminally truncated Bcl-2 that even exhibits proapoptotic activated but by forming a DISC with MAVS. The opposite activity (3). In this study, we now show that the MAVS/caspase-8 effects, FasL triggering a MAVS DISC or SFV promoting a Fas pathway activates caspase-3 under these conditions, bypassing the DISC cannot be seen, indicating that the two protein complexes Bcl-2 block. Hence, Bcl-2 overexpressing tumors may be selective form under different apoptotic circumstances. Caspase-8 has in- targets for anticancer therapy using SFV-related viral vectors or deed been found on mitochondria (40, 41) and also reported to dsRNA derivatives, an approach that has been initiated recently (5, form protein complexes with Bap31/Bap29 on the endoplasmic 8, 42, 43). These agents also may be combined with currently used reticulum (44) and with Atg5 and p62 on autophagosomes (45). BH3 mimetics or Bcl-2 inhibitors such as ABT737 (6) to increase Moreover, caspase-8 recently has been implicated in Bax/Bak- the efficacy of tumor killing via both Bax/Bak- and MAVS/cas- independent apoptosis induced by glucose deprivation (46). Fi- pase-8–dependent pathways. nally, caspase-8 formed an atypical death complex with TLR3, MAVS is required for effective antiviral responses to RNA which stimulates a type I IFN response via TRIF, TRAF-6, and viruses via the induction of type I IFNs (IFN-a/b) (17–20), and we receptor-interacting protein 1 that is independent of RIG-I/MDA-5 confirm this function in our study for SFV. Moreover, MAVS was and MAVS, mainly in response to extracellular dsRNA released suggested to mediate apoptosis from its mitochondrial site of during viral infections (13, 47). With regard to MAVS signaling, action, but it has remained elusive how it collaborates with the caspase-8 was found to be cleaved and activated in response to classical Bax/Bak-mediated MOMP pathway on this organelle and dsRNA and required for NF-kB activation and IFN-b production then activates effector caspases-3/-7 in the cytosol. Sendai virus, (48). Moreover, FLIP, an inhibitor of caspase-8, suppressed these dengue virus, and reovirus were reported to use the same down- events, and FLIP2/2 cells were more susceptible to dsRNA-induced stream signaling components of MAVS for apoptosis and IFN-b apoptosis (49). Most strikingly, caspase-8 seemed to interact with production, in particular the transcription factor IRF-3 (30–34). MAVS on mitochondria in response to DAP-3–mediated anoikis Although in one case IRF-3 triggered MOMP and caspase-3 (35). However, MAVS depletion diminished cell death only by through transcriptional induction of the proapoptotic BH3-only 10–20%, and it was unclear how MAVS could be activated by proteins Puma and Noxa (34), in the other case, it seemed to di- a stimulus unrelated to dsRNA and whether caspase-8 activation rectly interact and activate Bax (32). We cannot confirm these and its association with MAVS were required for anoikis. Thus, IRF-3–dependent mechanisms because SFV triggered apoptosis our study clearly shows that both MAVS and caspase-8 are required entirely independent of the MAVS downstream signaling com- for SFV-induced apoptosis, that their association on mitochondria is ponents TRAF-2, IRF-3, IRF-7, IFN-b, or the common IFNAR. It induced after SFV infection, and that this process does not need the 1182 SFV APOPTOSIS BY MAVS/CASPASE-8 INDEPENDENT OF Bax/Bak

Bax/Bak pathway. It also solves a long-term conundrum that 2. Fazakerley, J. K. 2002. Pathogenesis of Semliki Forest virus encephalitis. J. Neurovirol. 8(Suppl. 2): 66–74. Sindbis- or SFV-induced apoptosis could be delayed by the caspase- 3. Grandgirard, D., E. Studer, L. Monney, T. Belser, I. Fellay, C. Borner, and 8 inhibitors crmA and FLIP without involving death receptors (4, M. R. Michel. 1998. Alphaviruses induce apoptosis in Bcl-2-overexpressing 11, 12). cells: evidence for a caspase-mediated, proteolytic inactivation of Bcl-2. EMBO J. 17: 1268–1278. The pertinent question remains how caspase-8 is recruited to 4. Urban, C., C. Rheˆme, S. Maerz, B. Berg, R. Pick, R. Nitschke, and C. Borner. and activated by MAVS on mitochondria after SFV infection. A 2008. Apoptosis induced by Semliki Forest virus is RNA replication dependent plausible activator may be FADD because it is an adaptor and and mediated via Bak. Cell Death Differ. 15: 1396–1407. 5. Quetglas, J. I., M. Ruiz-Guillen, A. Aranda, E. Casales, J. Bezunartea, and activator for caspase-8 on the death receptor DISC (9). Previous C. Smerdou. 2010. Alphavirus vectors for cancer therapy. Virus Res. 153: 179–196. studies suggested an association of FADD with caspase-8 and MAVS 6. Strasser, A., S. Cory, and J. M. Adams. 2011. Deciphering the rules of pro- grammed cell death to improve therapy of cancer and other diseases. EMBO J. after DAP-3–induced anoikis (35) or upon co-overexpression (17). 30: 3667–3683. Moreover, FADD was proposed to mediate NF-kB activation and 7. Ubol, S., P. C. Tucker, D. E. Griffin, and J. M. Hardwick. 1994. Neurovirulent strains IFN-b induction by MAVS overexpression (17) and in response to of Alphavirus induce apoptosis in bcl-2‑expressing cells: role of a single amino acid change in the E2 glycoprotein. Proc. Natl. Acad. Sci. USA 91: 5202–5206. dsRNA (48, 50). However, a role of FADD in dsRNA- or RNA vi- 8. Murphy, A. M., B. J. Sheahan, and G. J. Atkins. 2001. Induction of apoptosis in rus–induced apoptosis has not yet been reported, and we did not find BCL-2‑expressing rat prostate cancer cells using the Semliki Forest virus vector. any apoptosis protection of Bax/Bak2/2 cells from SFV when Int. J. Cancer 94: 572–578. 9. Dickens, L. S., I. R. Powley, M. A. Hughes, and M. MacFarlane. 2012. The FADD was completely downregulated, although caspase-8 deple- ‘complexities’ of life and death: death receptor signalling platforms. Exp. Cell tion effectively saved them. We were also unable to coimmuno- Res. 318: 1269–1277. 10. Barry, G., R. Fragkoudis, M. C. Ferguson, A. Lulla, A. Merits, A. Kohl, and precipitate FADD with endogenous MAVS in SFV-infected cells, J. K. Fazakerley. 2010. Semliki forest virus-induced endoplasmic reticulum although it perfectly associated with the Fas DISC upon FasL stress accelerates apoptotic death of mammalian cells. J. Virol. 84: 7369–7377. stimulation. Finally, in our hands, we have not seen a defect in IFN- 11. Nava, V. E., A. Rosen, M. A. Veliuona, R. J. Clem, B. Levine, and Downloaded from J. M. Hardwick. 1998. Sindbis virus induces apoptosis through a caspase- b induction by RNA viruses in the absence of FADD as suggested dependent, CrmA-sensitive pathway. J. Virol. 72: 452–459. by Balachandran et al. (50) (Z.J. Chen, personal communication), 12.Sarid,R.,T.Ben-Moshe,G.Kazimirsky,S.Weisberg,E.Appel,D.Kobiler, indicating that FADD is unlikely to be a crucial MAVS adaptor, S. Lustig, and C. Brodie. 2001. vFLIP protects PC-12 cells from apoptosis induced by Sindbis virus: implications for the role of TNF-a. Cell Death Differ. 8: 1224‑1231. neither for IFN-b nor for apoptosis signaling. It is also difficult to 13. Takeuchi, O., and S. Akira. 2009. Innate immunity to virus infection. Immunol. explain how MAVS would interact with FADD because it contains Rev. 227: 75–86.

14. Weber, F., V. Wagner, S. B. Rasmussen, R. Hartmann, and S. R. Paludan. 2006. http://www.jimmunol.org/ a CARD, whereas FADD links proteins via death and death effector Double-stranded RNA is produced by positive-strand RNA viruses and DNA domains. We therefore suggest a yet unknown adaptor forming the viruses but not in detectable amounts by negative-strand RNA viruses. J. Virol. MAVS/caspase-8 DISC on mitochondria in response to SFV. 80: 5059–5064. 15. Yoneyama, M., M. Kikuchi, T. Natsukawa, N. Shinobu, T. Imaizumi, Interestingly, RNA viruses such as hepatitis C virus (HCV) or M. Miyagishi, K. Taira, S. Akira, and T. Fujita. 2004. The RNA helicase RIG-I influenza virus target the MAVS signaling pathway to disable the has an essential function in double-stranded RNA-induced innate antiviral host cell’s antiviral response. In the case of HCV, the nonstructural responses. Nat. Immunol. 5: 730–737. 16. Kato, H., O. Takeuchi, S. Sato, M. Yoneyama, M. Yamamoto, K. Matsui, NS3/4A protein is a protease, which cleaves and inactivates MAVS S. Uematsu, A. Jung, T. Kawai, K. J. Ishii, et al. 2006. Differential roles of MDA5 (51). For influenza, the NS1 protein blocks the TRIM25 ubiquitin and RIG-I helicases in the recognition of RNA viruses. Nature 441: 101–105. E3 ligase of RIG-I required for its optimal downstream signaling 17. Kawai, T., K. Takahashi, S. Sato, C. Coban, H. Kumar, H. Kato, K. J. Ishii, O. Takeuchi, and S. Akira. 2005. IPS-1, an adaptor triggering RIG-I‑ and Mda5- by guest on October 2, 2021 (52) and the PB1-F2 protein interferes with the RIG-I/MAVS mediated type I interferon induction. Nat. Immunol. 6: 981–988. complex (53). Because both HCV and influenza viruses induce 18. Meylan, E., J. Curran, K. Hofmann, D. Moradpour, M. Binder, R. Bartenschlager, and J. Tschopp. 2005. Cardif is an adaptor protein in the RIG-I antiviral pathway only little host cell apoptosis, we speculate that they use inhibition and is targeted by hepatitis C virus. Nature 437: 1167–1172. of the herein described MAVS/caspase-8 apoptosis pathway as 19. Seth, R. B., L. Sun, C. K. Ea, and Z. J. Chen. 2005. Identification and charac- another strategy of immune evasion. Such strategy is known for terization of MAVS, a mitochondrial antiviral signaling protein that activates NF-kB and IRF 3. Cell 122: 669–682. DNA viruses such as adeno- or herpesviruses, which encode Bcl- 20. Xu, L. G., Y. Y. Wang, K. J. Han, L. Y. Li, Z. Zhai, and H. B. Shu. 2005. VISA is 2–like homologs in their genomes to protect infected cells from an adapter protein required for virus-triggered IFN-b signaling. Mol. Cell 19: 727–740. apoptosis (54). A further investigation of the MAVS-mediated ap- 21. Der, S. D., A. Zhou, B. R. Williams, and R. H. Silverman. 1998. Identification of optosis pathway, in particular the identification of the components differentially regulated by interferon a, b,org using oligonucleotide of the MAVS/caspase-8 DISC on mitochondria will shed more light arrays. Proc. Natl. Acad. Sci. USA 95: 15623–15628. 22. Geiss, G., G. Jin, J. Guo, R. Bumgarner, M. G. Katze, and G. C. Sen. 2001. A on the mechanisms by which RNA viruses break antiviral immunity. comprehensive view of regulation of by double-stranded RNA- mediated cell signaling. J. Biol. Chem. 276: 30178–30182. Acknowledgments 23. Garcı´a, M. A., E. F. Meurs, and M. Esteban. 2007. The dsRNA protein kinase We thank Hiroki Kato (Osaka, Japan) for the RIG-I2/2 and MDA-52/2 PKR: virus and cell control. Biochimie 89: 799–811. 2/2 2/2 24. Hovanessian, A. G. 2007. On the discovery of interferon-inducible, double- MEFs; Robert H. Silverman (Cleveland, OH) for the PKR ,RNase-L , stranded RNA activated enzymes: the 29-59oligoadenylate synthetases and the 2/2 and PKR/RNase-L MEFs; Ian Gentle (Freiburg, Germany) for the cas- protein kinase PKR. Cytokine Growth Factor Rev. 18: 351–361. 2 2 2 2 pase-8 / MEFs; Georg Ha¨cker (Freiburg, Germany) for the TRAF2 / , 25. Lee, S. B., and M. Esteban. 1994. The interferon-induced double-stranded RNA- IRF-32/2,andIRF-3/72/2 MEFs; Friedemann Weber (Marburg, Germany) activated protein kinase induces apoptosis. Virology 199: 491–496. 2/2 2/2 26. Srivastava, S. P., K. U. Kumar, and R. J. Kaufman. 1998. Phosphorylation of for the anti-dsRNA Ab and the IFNb and IFNAR MEFs; Andreas eukaryotic initiation factor 2 mediates apoptosis in response to ac- Strasser, (Walter and Eliza Hall Institute, Parkville, VIC, Australia) for the tivation of the double-stranded RNA-dependent protein kinase. J. Biol. Chem. anti-FADD Ab, the Bax2/2,Bak2/2, and Bax/Bak2/2 as well as the FADD2/2 273: 2416–2423. MEFs; Cristina Mun˜oz-Pinedo (Barcelona, Spain) for the Bax/Bak2/2;sh-Ctrl 27. Der, S. D., Y. L. Yang, C. Weissmann, and B. R. Williams. 1997. A double- 2/2 stranded RNA-activated protein kinase-dependent pathway mediating stress- and Bax/Bak ;sh-Casp-8 MEFs; and Gabriel Nunez (Ann Arbor, MI) induced apoptosis. Proc. Natl. Acad. Sci. USA 94: 3279–3283. for the human Bcl-xL cDNA. 28. Barry, G., L. Breakwell, R. Fragkoudis, G. Attarzadeh-Yazdi, J. Rodriguez- Andres, A. Kohl, and J. K. Fazakerley. 2009. PKR acts early in infection to Disclosures suppress Semliki Forest virus production and strongly enhances the type I in- terferon response. J. Gen. Virol. 90: 1382–1391. The authors have no financial conflicts of interest. 29. Bisbal, C., and R. H. Silverman. 2007. Diverse functions of RNase L and implications in . Biochimie 89: 789–798. 30. Holm,G.H.,J.Zurney,V.Tumilasci,S.Leveille,P.Danthi,J.Hiscott,B.Sherry,and References T. S. Dermody. 2007. Retinoic acid-inducible gene-I and interferon-b 1. Strauss, J. H., and E. G. Strauss. 1994. The alphaviruses: gene expression, stimulator-1 augment proapoptotic responses following mammalian reovirus infec- replication, and evolution. Microbiol. Rev. 58: 491–562. tion via interferon regulatory factor-3. J. Biol. Chem. 282: 21953–21961. The Journal of Immunology 1183

31. Lei, Y., C. B. Moore, R. M. Liesman, B. P. O’Connor, D. T. Bergstralh, 43. Cheng, Y. S., and F. Xu. 2010. Anticancer function of polyinosinic-polycytidylic Z. J. Chen, R. J. Pickles, and J. P.-Y. Ting. 2009. MAVS-mediated apoptosis and acid. Cancer Biol. Ther. 10: 1219–1223. its inhibition by viral proteins. PLoS One 4: e5466. 44. Breckenridge, D. G., M. Nguyen, S. Kuppig, M. Reth, and G. C. Shore. 2002. 32. Chattopadhyay, S., J. T. Marques, M. Yamashita, K. L. Peters, K. Smith, The procaspase-8 isoform, procaspase-8L, recruited to the BAP31 complex at A. Desai, B. R. G. Williams, and G. C. Sen. 2010. Viral apoptosis is induced by the endoplasmic reticulum. Proc. Natl. Acad. Sci. USA 99: 4331–4336. IRF-3-mediated activation of Bax. EMBO J. 29: 1762–1773. 45. Young, M. M., Y. Takahashi, O. Khan, S. Park, T. Hori, J. Yun, A. K. Sharma, 33. Yu, C. Y., R. L. Chiang, T. H. Chang, C. L. Liao, and Y. L. Lin. 2010. The S. Amin, C. D. Hu, J. Zhang, et al. 2012. Autophagosomal membrane serves as interferon stimulator mitochondrial antiviral signaling protein facilitates cell platform for intracellular death-inducing signaling complex (iDISC)-mediated death by disrupting the mitochondrial membrane potential and by activating caspase-8 activation and apoptosis. J. Biol. Chem. 287: 12455–12468. caspases. J. Virol. 84: 2421–2431. 46. Caro-Maldonado, A., S. W. G. Tait, S. Ramı´rez-Peinado, J.-E. Ricci, I. Fabregat, 34. Besch, R., H. Poeck, T. Hohenauer, D. Senft, G. Ha¨cker, C. Berking, V. Hornung, D. R. Green, and C. Mun˜oz-Pinedo. 2010. Glucose deprivation induces an S. Endres, T. Ruzicka, S. Rothenfusser, and G. Hartmann. 2009. Proapoptotic atypical form of apoptosis mediated by caspase-8 in Bax-, Bak-deficient cells. signaling induced by RIG-I and MDA-5 results in type I interferon-independent Cell Death Differ. 17: 1335–1344. apoptosis in human melanoma cells. J. Clin. Invest. 119: 2399–2411. 47. Estornes, Y., F. Toscano, F. Virard, G. Jacquemin, A. Pierrot, B. Vanbervliet, 35. Li, H.-M., D. Fujikura, T. Harada, J. Uehara, T. Kawai, S. Akira, J. C. Reed, M. Bonnin, N. Lalaoui, P. Mercier-Gouy, Y. Pache´co, et al. 2012. dsRNA induces A. Iwai, and T. Miyazaki. 2009. IPS-1 is crucial for DAP3-mediated anoikis apoptosis through an atypical death complex associating TLR3 to caspase-8. Cell induction by caspase-8 activation. Cell Death Differ. 16: 1615–1621. Death Differ. 19: 1482–1494. 36. Egger, L., J. Schneider, C. Rheˆme, M. Tapernoux, J. Ha¨cki, and C. Borner. 2003. 48. Takahashi, K., T. Kawai, H. Kumar, S. Sato, S. Yonehara, and S. Akira. 2006. Serine proteases mediate apoptosis-like cell death and phagocytosis under Roles of caspase-8 and caspase-10 in innate immune responses to double- caspase-inhibiting conditions. Cell Death Differ. 10: 1188–1203. stranded RNA. J. Immunol. 176: 4520–4524. 37. Shimizu, M., A. Fontana, Y. Takeda, H. Yagita, T. Yoshimoto, and A. Matsuzawa. 49. Handa, P., J. C. Tupper, K. D. Jordan, and J. M. Harlan. 2011. FLIP (Flice-like 1999. Induction of antitumor immunity with Fas/APO-1 ligand (CD95L)- inhibitory protein) suppresses cytoplasmic double-stranded-RNA-induced apo- transfected neuroblastoma neuro-2a cells. J. Immunol. 162: 7350–7357. ptosis and NF-kB and IRF3-mediated signaling. Cell Commun. Signal. 9: 16 38. Walter, D., K. Schmich, S. Vogel, R. Pick, T. Kaufmann, F. C. Hochmuth, 10.1186/1478-811X-9-16. A. Haber, K. Neubert, S. McNelly, F. von Weizsa¨cker, et al. 2008. Switch from 50. Balachandran, S., E. Thomas, and G. N. Barber. 2004. A FADD-dependent in- type II to I Fas/CD95 death signaling on in vitro culturing of primary hep- nate immune mechanism in mammalian cells. Nature 432: 401–405. atocytes. Hepatology 48: 1942–1953. 51. Li, X. D., L. Sun, R. B. Seth, G. Pineda, and Z. J. Chen. 2005. Hepatitis C virus Downloaded from 39. Croker, B. A., J. A. O’Donnell, C. J. Nowell, D. Metcalf, G. Dewson, protease NS3/4A cleaves mitochondrial antiviral signaling protein off the mi- K. J. Campbell, K. L. Rogers, Y. Hu, G. K. Smyth, J.-G. Zhang, et al. 2011. Fas- tochondria to evade innate immunity. Proc. Natl. Acad. Sci. USA 102: 17717– mediated neutrophil apoptosis is accelerated by Bid, Bak, and Bax and inhibited 17722. by Bcl-2 and Mcl-1. Proc. Natl. Acad. Sci. USA 108: 13135–13140. 52. Rajsbaum, R., R. A. Albrecht, M. K. Wang, N. P. Maharaj, G. A. Versteeg, 40. Qin, Z.-H., Y. Wang, K. K. Kikly, E. Sapp, K. B. Kegel, N. Aronin, and M. DiFiglia. E. Nistal-Villa´n, A. Garcı´a-Sastre, and M. U. Gack. 2012. Species-specific in- 2001. Pro-caspase-8 is predominantly localized in mitochondria and released into hibition of RIG-I ubiquitination and IFN induction by the influenza A virus NS1 cytoplasm upon apoptotic stimulation. J. Biol. Chem. 276: 8079–8086. protein. PLoS Pathog. 8: e1003059.

41. Schug, Z. T., F. Gonzalvez, R. H. Houtkooper, F. M. Vaz, and E. Gottlieb. 2011. 53. Dudek, S. E., L. Wixler, C. Nordhoff, A. Nordmann, D. Anhlan, V. Wixler, and http://www.jimmunol.org/ BID is cleaved by caspase-8 within a native complex on the mitochondrial S. Ludwig. 2011. The influenza virus PB1-F2 protein has interferon antagonistic membrane. Cell Death Differ. 18: 538–548. activity. Biol. Chem. 392: 1135–1144. 42. Ambrus, J. L., Sr., K. C. Chadha, A. Islam, S. Akhter, and J. L. Ambrus, Jr. 2006. 54. Postigo, A., and P. E. Ferrer. 2009. Viral inhibitors reveal overlapping Treatment of viral and neoplastic diseases with double-stranded RNA derivatives themes in regulation of cell death and innate immunity. Microbes Infect. 11: and other new agents. Exp. Biol. Med. (Maywood) 231: 1283–1286. 1071–1078. by guest on October 2, 2021