Innate Immune Responses in Human Monocyte-Derived Dendritic Cells Are Highly Dependent on the Size and the 5′ Phosphorylation of RNA Molecules This information is current as of September 25, 2021. Miao Jiang, Pamela Österlund, L. Peter Sarin, Minna M. Poranen, Dennis H. Bamford, Deyin Guo and Ilkka Julkunen J Immunol 2011; 187:1713-1721; Prepublished online 8 July 2011; doi: 10.4049/jimmunol.1100361 Downloaded from http://www.jimmunol.org/content/187/4/1713

Supplementary http://www.jimmunol.org/content/suppl/2011/07/08/jimmunol.110036 http://www.jimmunol.org/ Material 1.DC1 References This article cites 50 articles, 24 of which you can access for free at: http://www.jimmunol.org/content/187/4/1713.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision by guest on September 25, 2021

• 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 © 2011 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Innate Immune Responses in Human Monocyte-Derived Dendritic Cells Are Highly Dependent on the Size and the 59 Phosphorylation of RNA Molecules

Miao Jiang,*,†,‡,x Pamela O¨ sterlund,x L. Peter Sarin,†,‡ Minna M. Poranen,†,‡ Dennis H. Bamford,†,‡ Deyin Guo,* and Ilkka Julkunenx

Recognition of viral genetic material takes place via several different receptor systems, such as retinoic acid-inducible gene I-like receptors and TLRs 3, 7, 8, and 9. At present, systematic comparison of the ability of different types of to induce innate immune responses in human immune cells has been limited. In this study, we generated bacteriophage f6 and influenza A -specific ssRNA and dsRNA molecules ranging from 58 to 2956 nt. In human monocyte-derived dendritic cells (moDCs),

short dsRNAs efficiently upregulated the expression of IFN (IFN-a, IFN-b, and IFN–l1) and proinflammatory (TNF-a, IL-6, IL- Downloaded from 12, and CXCL10) cytokine genes. These genes were also induced by ssRNA molecules, but size-specific differences were not as pronounced as with dsRNA molecules. Dephosphorylation of short ssRNA and dsRNA molecules led to a dramatic reduction in their ability to stimulate innate immune responses. Such a difference was not detected for long ssRNAs. RNA-induced cytokine responses correlated well with IFN regulatory factor 3 phosphorylation, suggesting that IFN regulatory factor 3 plays a major role in both ssRNA- and dsRNA-activated responses in human moDCs. We also found that IFN gene expression was efficiently stimulated following recognition of short dsRNAs by retinoic acid-inducible gene I and TLR3 in human embryonic kidney 293 http://www.jimmunol.org/ cells, whereas ssRNA-induced responses were less dependent on the size of the RNA molecule. Our data suggest that human moDCs are extremely sensitive in recognizing foreign RNA, and the responses depend on RNA size, form (ssRNA versus dsRNA), and the level of 59 phosphorylation. The Journal of Immunology, 2011, 187: 1713–1721.

attern recognition receptors (PRRs) are perhaps the best RIG-I and melanoma differentiation-associated gene 5 (MDA5), can characterized microbial sensors in the innate immune sys- recognize foreign RNA in the cytoplasm and activate intracellular P tem. They can detect microbe-specific material known as signaling cascades by interacting with the adaptor molecule IFN-b pathogen-associated molecular patterns (PAMPs), which include promoter stimulator-1 (4–6). In contrast, TLR3, TLR7, and TLR8 by guest on September 25, 2021 microbial structural and secretory components and genetic material are transmembrane receptors expressed on the plasma membrane (1). Recognition of viral RNA can trigger the induction of type I and (TLR3) or endosomes and the endoplasmic reticulum (TLR7 and III IFNs through the activation of multiple host signaling pathways TLR8) (7). TLR7 and TLR8 sense viral ssRNA and use MyD88 as (2, 3). Retinoic acid-inducible gene I (RIG-I)–like receptors (RLRs), the adaptor molecule to transmit the signal, whereas TLR3 detects dsRNA triggering the innate immunity response via Toll/IL-1R *National Key Laboratory of Virology and Modern Virology Research Center, Col- domain-containing adaptor-inducing IFN-b (2, 8). Activation of lege of Life Sciences, Wuhan University, Wuhan 430072, China; †Institute of Bio- RLR and TLR pathways leads to activation of NF-kB and phos- technology, University of Helsinki, FIN-00014 Helsinki, Finland; ‡Department of Biosciences, University of Helsinki, FIN-00014 Helsinki, Finland; and xDepartment phorylation of IFN regulatory factor (IRF) 3 and IRF7, leading to the of Vaccination and Immune Protection, National Institute for Health and Welfare, expression of IFN genes (2, 5, 6, 9). FIN-00271 Helsinki, Finland RIG-I is essential for eliciting IFN responses by many ssRNA Received for publication February 3, 2011. Accepted for publication June 5, 2011. , including both negative-stranded (2)RNA (e.g., paramyxo-, This work was supported by China “973” Basic Research Program Grant rhabdo-, and orthomyxoviruses) and positive-stranded (+)RNA 2010CB911800 (to D.G.) and China Natural Science Foundation Grants 30925003 viruses (e.g., flaviviruses). Nonetheless, MDA5 plays an important and 30921001 (to D.G.); Academy of Finland Center of Excellence Program 2006- 2011 Grant 1129684 (to D.H.B.); and Research Council for Health Grants 115279 role in recognizing other types of viruses, such as (+)RNA con- (to I.J.) and 128080 (to P.O¨ .). M.J. is supported by the China Scholarship Council. taining picornaviruses (10). In contrast, some viruses can be M.M.P., D.H.B., and I.J. are supported by the Sigrid Juselius Foundation. detected by both RIG-I and MDA5, such as West Nile and Dengue Address correspondence and reprint requests to Prof. Ilkka Julkunen or Prof. Deyin viruses that belong to the Flaviviridae family, or reovirus belonging Guo, Department of Vaccination and Immune Protection, National Institute for Health and Welfare, P.O. Box 30, FIN-00271 Helsinki, Finland (I.J.) or National to Reoviridae, a family of segmented dsRNA viruses (11, 12). Key Laboratory of Virology and Modern Virology Research Centre, College of Life The ligands for RIG-I have extensively been studied during the Sciences, Wuhan University, Wuhan 430072, China (D.G.). E-mail addresses: ilkka. 9 julkunen@thl.fi (I.J.) and [email protected] (D.G.) last few years, revealing three major groups, as follows: 1) 5 tri- The online version of this article contains supplemental material. phosphate containing RNAs, such as viral genomic RNAs, in vitro transcribed RNAs, Pol III-transcribed RNAs, and blunt end 59- Abbreviations used in this article: Ct, threshold cycle; DC, dendritic ; HEK, human embryonic kidney; HMW, high m.w.; IAV, influenza A virus; IRF, IFN reg- triphosphate double-stranded oligoribonucleotides; 2) RNAs lack- ulatory factor; LMW, low m.w.; MDA5, melanoma differentiation-associated gene ing 59 triphosphates, such as short polyinosinic:polycytidylic acid 5; moDC, monocyte-derived DC; PAMP, pathogen-associated molecular pattern; poly I:C, polyinosinic:polycytidylic acid; PRR, pattern recognition receptor; RdRP, RNA- (poly I:C) molecules, 59-monophosphorylated double-stranded dependent RNA polymerase; RIG-I, retinoic acid-inducible gene I; RLR, RIG-I–like oligoribonucleotides, and RNase L-cleaved cellular RNAs; and 3) receptor; S, small. RNAs with varying 59-end characteristics, for example, viral Copyright Ó 2011 by The American Association of Immunologists, Inc. 0022-1767/11/$16.00 or replication intermediates and total cellular RNA www.jimmunol.org/cgi/doi/10.4049/jimmunol.1100361 1714 RNA-INDUCED INNATE IMMUNE RESPONSES IN HUMAN moDCs extracted from virus-infected cells (13). However, the character- Phusion HF DNA polymerase (Finnzymes). Full-length and truncated istics of RNA agonists for MDA5 are less clear. Poly I:C is known cDNAs of bacteriophage f6 S segment and chimeric IAV genes, covering 9 to be a ligand for MDA5 and, in mouse embryonic fibroblasts, a range from 58 to 2956 nt from the 3 -end (Fig. 1A, Supplemental Fig. 1A), were obtained by PCR amplification from pLM659 (27) and RIG-I and MDA5 preferentially recognize short and long dsRNAs, pLD19, respectively. The oligonucleotides used in PCR reactions are listed respectively (14). TLR3, 7, and 8 are mainly responsible for in Supplemental Table II. dsRNAs were generated by one-step in vitro detecting viral RNAs. The ligands for TLR3 involve synthetic transcription and replication using the Replicator RNAi kit (Finnzymes). dsRNA, poly I:C, viral genomic dsRNA, or transcription or rep- Enzymatically synthesized ssRNA and dsRNA molecules were isolated with TRIzol (Invitrogen)/chloroform (5:1) extraction, followed by analytical lication intermediates of ssRNA or DNA viruses. The recognition agarose gel electrophoresis and 3 M sodium acetate precipitation (58–108 of RNA is most likely not sequence specific, as also suggested by nt RNAs) (28), or stepwise lithium chloride precipitation (208–2956 nt recent structural analysis on TLR3–RNA complexes (15). RNAs) (24). All RNAs were washed with 70% ethanol, dissolved in sterile Dendritic cells (DCs) express several classes of PRRs, and they water, and precipitated with 0.75 M NH4Ac-70% ethanol (24, 28). The are the key cell types regulating both innate and adaptive immunity pellets were suspended in sterile water and desalted with NAP5 columns (GE Healthcare) prior to HPLC purification (GenFax PAC, Waters). (16). PRRs recognize PAMPs by two ways, as follows: extracel- lularly via TLRs, and intracellularly via NOD-like receptors, RNA secondary structure calculations RLRs, and certain TLR subtypes (17). Microbial (PAMP) stimu- lation or virus of DCs leads to the maturation of DCs, The simulated two-dimensional RNA structure predictions (minimum free induction of costimulatory molecules, and production of cyto- energy and energy parameter 37˚C) were performed using the RNAfold WebServer program from the Vienna RNA Package. kines, which together with MHC-dependent Ag presentation leads to the activation of adaptive immune responses. Dephosphorylation of RNA Downloaded from The effects of small interfering RNAs, ranging in size from 17 To remove 59-triphosphates, enzymatically synthesized ssRNAs and to 31 nt on antiviral responses, have been extensively studied (18– dsRNAs were treated with FastAP thermosensitive alkaline phosphatase 20). However, at present systematic analysis on the role of RNA (Fermentas). The reactions were carried out at standard reaction conditions length, form, specificity, or the extent of phosphorylation in ac- (10 mM Tris-HCl [pH 8.0], 5 mM MgCl2, 0.1 M KCl, 0.02% Triton X-100, tivating antiviral responses, especially in the human system, is 1 mM 2-ME, and 0.1 mg/ml BSA) for 30 min at 37˚C, followed by in- activation of the enzyme at 75˚C for 5 min. Dephosphorylated RNAs were lacking. In the current study, we have applied DNA-dependent T7 subsequently purified, as described above. http://www.jimmunol.org/ polymerase and RNA-dependent RNA polymerase (RdRP) of bacteriophage f6 (21–24) to produce f6 (25) and influenza A Stimulation of cells with ssRNA and dsRNA molecules virus (IAV)-specific ssRNA and perfectly duplexed dsRNA mol- moDC stimulations. To minimize interindividual variation, all experiments ecules in vitro. These RNA molecules and their 59-dephosphory- were performed with moDCs obtained from four blood donors, and the cells lated derivatives were used to stimulate human primary monocyte- were stimulated separately. In initial experiments, optimal concentrations for synthetic ssRNAs, dsRNAs, and poly I:C were determined. Enzymat- derived DCs (moDCs) and human embryonic kidney (HEK) 293 ically synthesized RNAs (50 ng/ml) or poly I:C (10 mg/ml; Invivogen) were cells expressing various RLRs or TLRs and the expression of IFN transfected with Lipofectamine 2000 (Invitrogen), according to manu- and other proinflammatory cytokine genes were analyzed. The facturer’s instructions. After , cells from different donors were data reveal that in human moDCs the RNA length, form, and the collected and pooled. Cells were used for isolation of total cellular RNA or by guest on September 25, 2021 9 lysed for samples with Passive buffer (Promega, Dual Glo extent of 5 -triphosphate greatly affect cytokine gene expression. kit). Because there is considerable individual variation in the responses in However, the origin of viral sequence (f6 versus IAV) or differ- different blood donors (Supplemental Fig. 2), we used pooled cellular ential secondary structure of ssRNA (based on prediction) is un- RNA specimens to obtain a more global view of RNA-stimulated re- likely playing a role in RNA-induced stimulatory responses. sponses in moDCs. HEK293 cell stimulations. HEK293 cells were plated in 96-well culture Materials and Methods plates 1 d before transfection, after which the reporter and expression plasmids of RNA sensors (RIG-I, MDA5, or TLR3) or empty vectors were Cell lines and plasmids transfected into cells using TransIT-LT1 transfection reagent (Mirus Bio). The HEK293 cell line (American Type Culture Collection; ATCC CLR Renilla luciferase (Promega) was used to control transfection m 1573) was maintained in continuous growth in Eagle MEM (Sigma-Aldrich) efficiency. RNAs (20 ng/well) or poly I:C (1.5 g/well) were transfected supplemented with antibiotics, L-glutamine, and 10% FCS (Integro). with Lipofectamine 2000 (Invitrogen) 4 h after transfection of reporter and Human primary monocytes were purified from freshly collected, leuko- expression plasmids, according to manufacturer’s instruction. Luciferase cyte-rich buffy coats obtained from healthy blood donors (Finnish Red assays were performed 20 h after RNA stimulation with a Dual Glo kit Cross Blood Transfusion Service, Helsinki, Finland), as described pre- (Promega), according to manufacturer’s instructions. viously (26). Monocytes were differentiated into immature DCs for 6 d in the presence of 10 ng/ml human rGM-CSF (BioSource International), and Quantitative RT-PCR 20 ng/ml human rIL-4 (R&D Systems) in RPMI 1640 medium supple- Total cellular RNAwas isolated from moDCs derived from four pooled donors m m mented with 0.6 g/ml penicillin, 60 g/ml streptomycin, 2 mM L-glu- using the RNeasy Mini RNA Isolation kit (Qiagen). DNase-treated total tamine, 20 mM HEPES, and 10% FCS. cellular RNA (1 mg) was reverse transcribed into cDNA by using the TaqMan Plasmid pLM659 was the template for the production of full-length and + Reverse Transcriptase kit (Applied Biosystems). cDNA samples were then truncated mRNA sense ssRNA (s ) and dsRNA molecules of the small (S) amplified using a TaqMan Universal PCR Mastermix buffer (Applied Bio- f genome segment of bacteriophage 6 (27). Plasmid pLD18 was con- systems) and the commercial gene expression system assay (Applied Bio- structed by removing the egfp gene from multiple cloning site of plasmid systems) with primers and probes for IFN-a1 (Hs00256882 s1), IFN-b 9f _ pPS9 (24) and introducing a T7 terminator downstream of the 3 6 rep- (Hs00277188_s1), IFN-l1 (Hs00601677_g1), CXCL10 (Hs00171042_m1), lication signal. A chimera of IAV sequences from various conserved regions IL-1b (Hs00174097_m1), IL-6 (Hs00174131_ml), IL-10 (Hs00174986_m1), of the viral genome encoding viral internal (Supplemental Fig. 1A, IL-12 p40 (Hs00233688_m1), TNF-a (Hs00174128_ml), and b-actin Supplemental Table I) were reverse transcribed into cDNA using Moloney (Hs99999903_ml). Each cDNA sample was amplified in duplicate with murine leukemia virus reverse transcriptase (Fermentas) and amplified by a Mx3005P QPCR System (Stratagene). The relative amount of cytokine PCR (28). The individual segments were ligated and cloned into the mul- mRNA to b-actin was calculated with the DD comparative threshold cycle tiple cloning site of pLD18, thereby creating plasmid pLD19. All plasmids (C ) method. were maintained and propagated in Escherichia coli strain DH5a. t Synthesis of ssRNA and dsRNA molecules Western blot analysis ssRNAs were produced by in vitro transcription with T7 RNA polymerase. Whole-cell lysates were prepared in Passive Lysis buffer of Dual Luciferase Templates for T7 transcription were prepared by PCR amplification using Assay kit (Promega) containing 1 mM Na3VO4. Protein aliquots of whole- The Journal of Immunology 1715 cell lysates (30 mg) were separated on 10% SDS polyacrylamide gels using the Laemmli buffer system (29). Proteins were transferred onto Immobilon-P membranes, followed by blocking with 5% milk in PBS (or in TBS for cell-signaling Abs). Guinea pig (anti-RIG-I, anti-MDA5, and anti-IRF7) and rabbit (anti-IRF3) Abs were used, as previously described (30–33). Abs against b-actin (actin H-300; SC-10731) were obtained from Santa Cruz Biotechnology. Abs against the phosphorylated forms of IRF3 (Ser396) were from Cell Signaling Technology. HRP-conjugated goat anti- rabbit or rabbit anti-guinea pig Abs (DakoCytomation) were used in sec- ondary staining. Ab signals were visualized on HyperMax films using the ECL system (GE Healthcare). ELISA Cytokine levels of TNF-a and CXCL10 from moDC culture supernatants were analyzed using Ab pairs and standards from BD Pharmingen (San Diego, CA). IFN-l1 was measured with a Duoset kit ELISA (R&D Sys- tems). Secreted levels of IFN-a (multisubtype) and IFN-b were measured with VeriKine ELISA kit supplied by PBL Biomedical Laboratories (Piscataway, NJ).

Results Downloaded from Establishing a tool kit by enzymatic RNA synthesis and modification Cellular RNA recognizing receptors (TLRs and RLRs) have dif- ferent specificities for different types and sizes of ssRNA and dsRNA molecules (7, 15, 34). To systematically study the effects f of RNA size, type (ssRNA versus dsRNA), specificity ( 6 or IAV), http://www.jimmunol.org/ or level of 59 phosphorylation, we enzymatically synthesized different type and size RNA molecules using bacteriophage f6S segment (Fig. 1A) or chimeric IAV-specific (Supplemental Fig. 1) cDNAs as templates. Using well-established T7 polymerase-based FIGURE 1. Enzymatically synthesized dsRNA and ssRNA molecules derived from PCR-amplified fragments of f6 S segment. A, Schematic in vitro transcription system, we were able to generate a collection f 9 representation of enzymatic synthesis of bacteriphage 6 S segment-spe- of 5 -phosphorylated ssRNA molecules of different molecular cific ssRNA and dsRNA molecules. T7 RNA transcription was performed sizes (Fig. 1B). As a novelty, we used the ability of bacteriphage to yield ssRNA using various cDNA fragments of the f6 S segment as f6 RdRP to produce dsRNA molecules from the ssRNA tem- templates (top panel). An identical set of dsRNAs was generated from the plates. The dsRNA molecules were produced in a combined one- ssRNA template by primer-independent f6 RdRP-catalyzed replication by guest on September 25, 2021 step T7 polymerase RNA transcription and bacteriophage f6 (lower panel). The full-length f6 S segment cDNA is depicted as an or- RdRP dsRNA synthesis reaction (Fig. 1C). ange bar; the dark green block indicates the location of the T7 promoter; The dsRNA products generated by the f6 RdRP are blunt ended the light green and blue bars represent ssRNA and dsRNA products, re- and completely base-paired dsRNA molecules (21; see also Fig. spectively. B and C, Native agarose gel electrophoresis of T7-transcribed f 1). According to the secondary structure calculations, the pro- ssRNA (B) and 6 RdRP-catalyzed dsRNA (C). Note that RNA runs slower than DNA, thereby appearing slightly longer based on the length duced f6 S segment-specific ssRNA molecules contain various indicated by the DNA marker (lane marker). The sizes of the ssRNA and types of hairpin structures (Fig. 1D). Such structures may favor dsRNA molecules are indicated above each lane. D, In silico calculations the recognition of these molecules by different types of RLRs. of minimum free energy secondary structures of the f6 ssRNAs. The structures are colored according to base-pairing probability, ranging from Short dsRNA molecules are strong activators of IFN and 0 (least likely; violet) to 1 (most likely; red). The 59 end is indicated by proinflammatory cytokine genes a black arrow. For unpaired regions, the color denotes the probability of Although recognition of dsRNA by RNA sensors has been pre- being unpaired. viously studied (14, 35), systematic analysis of dsRNA length de- pendence in triggering antiviral responses in the human system, However, the difference in the responses induced by the short and such as in moDCs, has remained uncharacterized. Thus, we long dsRNAs was not as pronounced. The expression of the IL-10 studied the ability of dsRNA molecules of three different size gene was not induced with any of the dsRNAs studied. classes, short (58, 88, and 108 nt), medium (208, 308, and 508 nt), We also measured the cytokine levels in the supernatants col- and long (708, 1808, and 2956 nt), to induce the expression of IFN lected at three different time points (Fig. 2B). The data indi- (IFN-a1, IFN-b, IFN-l1), proinflammatory (TNF-a, IL-1b, IL-6), cate that short dsRNAs (58, 88, and 108 bp) efficiently induced Th1-type (IL-12, CXCL10), and anti-inflammatory Th2-type (IL- IFN-a, IFN-l1, TNF-a, and CXCL10 protein production. Al- 10) genes. A set of bacteriophage f6-specific dsRNAs (Fig. 1) though medium-size and large dsRNAs induced IFN and TNF-a was transfected into moDCs at equal concentrations. Cytokine relatively weakly, CXCL10 production was almost equally well mRNA expression was analyzed at three different time points. IFN induced by short and longer RNAs (Fig. 2B). As a whole, there family genes (IFN-a1, IFN-b, IFN-l1) and TNF-a were very appeared to be a very good correlation between cytokine mRNA efficiently induced by short dsRNAs (58, 88, and 108 bp) as and protein production levels. compared with induction by medium-size (308 bp) or long dsRNAs (708 and 1808 bp; Fig. 2A). Interestingly, IL-1b mRNA Cytokine gene expression is more sensitive to the size variation expression was induced only by long dsRNA molecules at early of dsRNA than ssRNA molecules time points (peaking at 3 h). The kinetics and expression patterns The ability of ssRNA and dsRNA molecules to induce innate anti- of IL-6, IL-12 p40, and CXCL10 genes resembled those of IFNs. viral responses in moDCs was compared. Commercial low m.w. 1716 RNA-INDUCED INNATE IMMUNE RESPONSES IN HUMAN moDCs

FIGURE 2. Relative cytokine mRNA expression and cytokine production in hu- man moDCs stimulated with dsRNA mol- ecules of varying size. A, MoDCs from four different blood donors were separately transfected with indicated dsRNAs (50 ng/ ml) for 3, 8, or 24 h. Relative expression of Downloaded from cytokine mRNAs was measured by quan- titative RT-PCR. The values were normal- ized against b-actin mRNA, and relative expression levels were calculated with the

DDCt method using untreated cells as a calibrator. B, Cell culture supernatants were

collected at 3, 8, and 24 h after transfec- http://www.jimmunol.org/ tion, and protein levels of IFN- l1, IFN-a, TNF-a, and CXCL10 were determined by ELISA. The mean (6SD) cytokine levels produced by the cells of four blood donors are shown. The data are representative of two individual experiments. Cytokines, time points, and the sizes of dsRNAs are indi- cated in the figure. by guest on September 25, 2021

(LMW; 0.2–1 kb) and high m.w. (HMW; 1.5–8 kb) poly I:C and RNA stimulation activates IRF3 phosporylation and enhances selected in vitro produced short (88 nt) and long (708 and 1808 nt) the expression of RLRs and IRF7 dsRNA and ssRNA molecules were used to stimulate moDCs for To investigate which signaling pathways would contribute to different periods of time (Fig. 3A,3B, Supplemental Fig. 2). triggering of innate immune responses after dsRNA or ssRNA Consistent with the data above (Fig. 2), short dsRNA molecules stimulation in moDCs, analysis of phospho-IRF3, total IRF3, IRF7, were able to elicit higher IFN and CXCL10 mRNA expression and RIG-I, and MDA5 expression was analyzed by Western blotting. protein production levels as compared with long dsRNAs. In the Short (88-bp) dsRNA elicited efficient IRF3 phosphorylation case of ssRNAs, the differences were not as pronounced, because in moDCs, whereas ssRNA induced IRF3 phosphorylation rela- also long ssRNA molecules induced IFN (especially IFN-b and C IFN-l1) and CXCL10 mRNA and protein expression (Fig. 3A, tively weakly (Fig. 3 ). The phosphorylation of IRF3 was clearly 3B). Stimulation of cells with LMW poly I:C led to somewhat detected at 3 h after dsRNA or ssRNA stimulation, reaching its higher cytokine mRNA and protein levels (except that of IFN-a) peak at 8 h poststimulation and thereafter declining to basal as compared with responses induced by HMW poly I:C. unphosphorylated levels within 24 h (Fig. 3C). The kinetics of There were clear differences in the kinetics of IFN and CXCL10 LMW and HMW poly I:C-induced phospho-IRF3 expression was gene expression. Depending on the IFN type, the peak mRNA faster and more efficient as compared with that induced by in vitro levels were observed at 3 h and/or 8 h after stimulation, followed by produced short and long dsRNA and ssRNA molecules (Fig. 3C). a clear decline within 24 h after stimulation. Poly I:C rapidly The basal expression of IRF7 was low, but it was strongly elevated induced CXCL10 gene expression, whereas its response to dsRNA at 8 and 24 h after RNA stimulation. The type or length of the and ssRNA stimulation was slightly slower. RNA did not significantly affect IRF7 expression levels. Initially, The Journal of Immunology 1717

low basal MDA5 and RIG-I levels were strongly induced by all forms of RNAs with certain differences. Poly I:C and short dsRNAs and ssRNAs were more efficient in triggering the in- duction of MDA5 and RIG-I protein expression at the 8-h time point, but at the 24-h time point the differences between the stimuli were less evident (Fig. 3C). The 59-triphosphate group is important in RNA recognition by RNA sensors The role of the 59-triphosphate group of the in vitro produced dsRNA and ssRNA molecules in stimulating cytokine mRNA expression in moDCs was investigated and compared with the responses induced by commercial poly I:C. Equal amounts of phosphorylated and dephosphorylated short (88 nt long) dsRNA and ssRNA were transfected into human moDCs, and the ex- pression of cytokine mRNAs (IFN-b,IFN-l1, IFN-a1, and CXCL10) was analyzed at three different time points after trans- fection. Interestingly, the removal of the 59-triphosphate group

from the short dsRNAs significantly suppressed the induction of Downloaded from cytokine gene expression (Fig. 4A,4B, left panel). Moreover, dephosphorylated short ssRNAs were practically unable to induce cytokine mRNA expression (Fig. 4A,4B, right panels). Phosphorylation of IRF3 was hardly detectable by Western blot analysis when moDCs were stimulated with dephosphorylated

short dsRNA or ssRNA molecules, indicating that 59-phosphate http://www.jimmunol.org/ groups in short RNAs contribute to the activation of cytokine gene expression (Fig. 4C). In the case of long RNA molecules, the level of RNA 59 phosphorylation did not significantly affect the effi- ciency of IRF3 phosphorylation (Fig. 4C). Multiple RNA sensors recognize dsRNA and ssRNA in HEK293 cells A previous study conducted in RIG-I2/2 and MDA52/2 mouse embryonic fibroblasts demonstrated that RIG-I selectively recog- by guest on September 25, 2021 nizes medium-size poly I:C molecules (∼300 bp long), whereas long poly I:Cs (∼800 bp or longer) are ligands for MDA5 (14). To further elucidate whether the recognition of dsRNA by multiple RNA sensors is length dependent also in the human system, var- ious size dsRNA molecules were transfected together with RNA sensor–expression constructs and IFN promoter–reporter con- structs into HEK293 cells, with LMW and HMW poly I:C func- tioning as positive controls (Fig. 5). Both in RIG-I–, MDA5-, and TLR3-overexpressing HEK293 cells, short (58–108 bp) dsRNAs triggered better activation of FIGURE 3. Stimulation of human moDCs with different types of RNA. IFN-b and IFN-l1 promoters as compared with medium-size A, Human moDCs were stimulated via transfection with LMW or HMW (208–508 bp) or long (708–2948 bp) dsRNAs (Fig. 5A). The poly I:C (10 mg/ml) or short (88 nt long) and long (708 and 1808 nt long) dsRNA and ssRNA molecules (50 ng/ml) for 3, 8, or 24 h, followed by IFN promoter activation in MDA5-expressing cells was somewhat quantitative RT-PCR. The values were normalized against b-actin, and stronger than in RIG-I–expressing cells, especially when cells relative mRNA levels were calculated with the DDCt method using un- were challenged with LMW or HMW poly I:C. In RIG-I gene- treated cells as a calibrator. Data are representative of three individual transfected cells, the dsRNA size dependency was clear, whereas experiments. B, Cell culture supernatants were collected at 3, 8, and 24 h in MDA5-expressing cells there appeared to be a trend that long after transfection, and protein levels of IFN-b, IFN-l1, IFN-a, TNF-a, and dsRNAs induced slightly better responses as compared with CXCL10 were determined by ELISA. The mean (6SD) cytokine levels medium-size dsRNA molecules (Fig. 5A). It is noteworthy that produced by the cells of four blood donors are shown. C, Western blot mere overexpression of MDA5 led to significant IFN promoter analysis for the expression of phospho-IRF3 (P-IRF3), total IRF3, MDA5, activation, rendering the dynamic range of the assay relatively RIG-I, IRF7, and b-actin proteins in RNA-stimulated human moDCs. Cells were collected at 3, 8, or 24 h after RNA transfection, and whole-cell insensitive. lysates were prepared. Cellular proteins (30 mg/lane) were separated on In TLR3-expressing cells, the control poly I:C molecules 10% SDS-PAGE, followed by electrophoretic transfer of the proteins to and the short 88-bp dsRNA were the best activators of both IFN-b polyvinylidene difluoride membranes and visualization of the transferred and IFN-l1 promoters (Fig. 5A, right panel). It is of interest that proteins by protein-specific Abs, as indicated. The data of one represen- all tested dsRNA molecules, although with slightly different ef- tative experiment of three independent experiments are shown. ficiency, were able to activate IFN promoters via TLR3 (Fig. 5A). To study the length dependency of ssRNA recognition by RNA sensors, we stimulated RIG-I– or MDA5-overexpressing HEK293 cells by transfecting them withequalmassamountoff6S 1718 RNA-INDUCED INNATE IMMUNE RESPONSES IN HUMAN moDCs

FIGURE 4. Cytokine gene expression in- duced by 59-phosphorylated and dephos- phorylated ssRNA and dsRNA molecules. A and B, moDCs from four blood donors were transfected with LMW or HMW forms of poly I:C (10 mg/ml), 59-triphosphorylated (593P), or dephosphorylated (deP) short (88- bp) dsRNA or ssRNA molecules (50 ng/ml) (A) or with the same molar amounts (0.5 pmol/ml) of short (88 nt long) or long (708 or 1808 nt long) 59-triphosphorylated and dephosphorylated dsRNA (B, left panel)or ssRNA (B, right panel) molecules for 3, 8, or 24 h. The values were normalized against b-actin mRNA, and relative mRNA levels Downloaded from were calculated with the DDCt method using untreated cells as a calibrator. Data are rep- resentative of three individual experiments. C, Western blot analysis of phospho-IRF3 (P-IRF3), total IRF3, and b-actin expression in 59-triphosphorylated and dephosphory- http://www.jimmunol.org/ lated dsRNA- and ssRNA-stimulated human moDCs. Cells were collected at 3, 8, or 24 h after RNA transfection, and lysed in passive lysate buffer, followed by separation of 30 mg samples of cellular proteins on 10% SDS- PAGE and Western blot analysis with the indicated Abs. One representative experi- ment of three independent experiments is shown. by guest on September 25, 2021

segment-specific ssRNAs of different lengths (Fig. 5B). In contrast (data not shown) is unlikely to contribute to RIG-I– or MDA5- to dsRNA, the length of the ssRNA did not have a major effect on dependent IFN promoter activation. the level of IFN-b or IFN-l1 promoter activation in HEK293 cells (Fig. 5B). Apparently, RIG-I– or MDA5-mediated ssRNA recog- Discussion nition was less length dependent as compared with that of dsRNA. It is well established that different types of RNAs can elicit antiviral To further investigate the recognition of the 59-triphosphate group responses via RIG-I (14, 36–39)–, MDA5 (14, 40)-, and/or TLR3 in dsRNA and ssRNA molecules by various RNA-sensing receptors, (41)-dependent signaling pathways. In the current study, we pro- we transfected similar amounts of phosphorylated and dephos- vide clear evidence that human primary moDCs are very sensitive phorylated short (88-nt) dsRNAs and ssRNAs into HEK293 cells sensors of various types of RNAs, and especially in vitro produced overexpressing RIG-I, MDA5, or TLR3 (Fig. 5C). Dephosphory- short dsRNA and ssRNA molecules strongly activated the ex- lation of dsRNA or ssRNA molecules led to reduced induction of pression of IFN and proinflammatory cytokine genes, but not that IFN-b and IFN-l1 promoters in RIG-I– or MDA5-overexpressed of anti-inflammatory IL-10 gene (Fig. 2). The 59 triphosphor- HEK293 cells; especially in MDA5-expressing cells, dephosphory- ylation of short dsRNAs and ssRNAs played an important role in lated short dsRNAs induced IFN-l1 promoter very poorly. Moreover, gene activation, because dephosphorylated molecules were almost dephosphorylation of short dsRNA or ssRNA molecules led to devoid of cytokine stimulatory activity (Figs. 4, 5). However, completely abolished IFN-b and IFN-l1 promoter activation via the human moDCs were also responsive to larger ssRNA molecules TLR3 signal pathway. even in the absence of 59-phosphate groups. Interestingly, we Similar IFN expression patterns were also observed when equal found no evidence that the origin of viral sequences (f6 versus mass (Fig. 5A) or molar amounts (Fig. 6A) of dsRNAs were used IAV) or potential ssRNA secondary structures would play a major to transfect HEK293 cells. Furthermore, the activation patterns role in DC activation. Rather we found evidence that the type of were not significantly changed when the f6-specific dsRNA RNA, its size, and the level of 59 phosphorylation are the key molecules were replaced with IAV-specific RNAs (Supplemental factors involved in RNA-induced responses. Human moDCs also Fig. 1, Fig. 6B). This suggests that the sequence appeared to show relatively wide capacity to respond to various specificity of the dsRNA molecules and also the ssRNA molecules types of RNAs, which is most likely due to the expression of The Journal of Immunology 1719

FIGURE 5. RNA-induced IFN gene expression in RIG-I–, MDA5-, and TLR3-expressing HEK293 cells. Downloaded from HEK293 cells (in 96-well plates, 3 3 105 cells/well) were transfected with empty vector (EV) or RIG-I, MDA5, or TLR3 (20 ng/well) expression plasmid to- gether with IFN-b or IFN-l1 promoter firefly lucifer- ase (20 ng/well) and RSV-Renilla luciferase plasmids (3 ng/well). After 4 h of incubation, cells were stim- ulated by transfecting them with different sized http://www.jimmunol.org/ dsRNAs or control LMW (A–C) or HMW (A) poly I:C (1.5 mg/well) for an additional 20 h, as indicated. Equal amounts (10 ng/well) (A) or equal molar amounts (0.5 pmol/ml) (B) of different sized RNA molecules were applied. C, Cells were stimulated with 59-triphosphorylated or dephosphorylated 88-bp dsRNA or 88-nt ssRNA molecules (10 ng/well). Cells were collected and luciferase assays were carried out to de-

termine the relative promoter activation. The means (6 by guest on September 25, 2021 SD) of three parallel determinations are shown. The data are representative of three individual experiments. RNA molecule sizes and expression plasmids are indi- cated.

a wide repertoire of PRRs and their flexibility in recognizing ecules triggering the expression of IFN and other proinflam- different types of RNAs. matory cytokine genes. In the case of long dsRNA molecules, It has previously been suggested that RIG-I preferentially rec- dimerization of RIG-I could be prevented (44), which would lead ognizes 59-triphosphorylated RNAs (14, 37, 38), including blunt- to inefficient RIG-I signaling by large dsRNA molecules. Con- ended dsRNAs with 59-triphosphate groups (42, 43). Because sistent with this hypothesis, medium or long dsRNA molecules short 59-triphosphorylated dsRNAs and ssRNAs induced strong were thus quite poor inducers of IFN responses in human moDCs immune responses in human moDCs, it is likely that RIG-I plays (Fig. 2) and in RIG-I–overexpressing HEK293 cells (Fig. 5A). a dominant role in recognizing foreign dsRNA and ssRNA mol- However, long dsRNAs were able to induce inflammatory 1720 RNA-INDUCED INNATE IMMUNE RESPONSES IN HUMAN moDCs

FIGURE 6. RNA-induced IFN gene expression in RIG-I– or MDA5-expressing HEK293 cells. HEK293 cells (in 96-well plates, 3 3 105 cells/well) were transfected with empty vector (EV) or RIG-I or MDA5 (20 ng/well) expression plasmids together with IFN-b or IFN-l1 promoter firefly luciferase (20 ng/ well) and RSV-Renilla luciferase plasmids (3 ng/well). After 4 h of incubation, cells were stimulated by transfecting them with LMW or HMW poly I:C (1.5 mg/well) or with indicated 59-triphosphorylated f6-specific dsRNAs with the same molar amounts (0.5 pmol/ml) (A) or with indicated IAV-specific dsRNAs (10 ng/well) (B) for an additional 20 h. Cells were collected and luciferase assays were carried out to determine the relative promoter activation.

The means (6SD) of three parallel determinations are shown. The data are representative of three individual experiments. Downloaded from responses, possibly through other signal pathways than RIG-I, upregulation of TLR and IRF7 genes is required (48, 49). It is because they efficiently enhanced IL-1b, IL-6, and CXCL10 likely that certain TLR/RLR ligands, such as short ssRNAs, may gene expression (Fig. 2). activate multiple transcription factor systems, whereas long

We also observed that short dsRNAs induced more efficient ssRNAs mainly stimulate the IRF3 signaling pathway to trigger http://www.jimmunol.org/ phosphorylation of IRF3 than long dsRNAs, especially at early antiviral responses. stages of stimulation. Although it has been reported that in mouse It has been observed that the 59-triphosphate group is essential fibroblasts long dsRNA molecules trigger antiviral responses in an for ssRNA recognition by RIG-I (36). In this study, we demon- IRF3-independent fashion (35), our result suggests that stimula- strated that short 59-triphosphorylated dsRNA and ssRNA mole- tion by short dsRNAs and to a lesser extent by long dsRNAs is cules could elicit high antiviral responses in human moDCs (Figs. mediated by IRF3 (Fig. 3C). This is also consistent with short 3, 4). Removing the 59-triphosphate group strongly reduced both dsRNAs being able to elicit strong IFN responses in transfected dsRNA- and ssRNA-induced antiviral responses (Fig. 4A). This HEK293 cells through RIG-I–, MDA5-, and TLR3-mediated indicates that RIG-I–dependent recognition of small RNAs, es- pathways (Fig. 5A). We also showed that stimulation of moDCs pecially that of ssRNA molecules, takes place mainly in a 59-tri- by guest on September 25, 2021 with short dsRNAs led to increased expression of RIG-I, MDA5, phosphate–dependent fashion. These data are consistent with pre- and IRF7 (Fig. 3C), which may provide a positive feedback to vious publications, which demonstrated that partially duplexed cellular responses. The increased expression of RLRs and IRF7 is RNAs (e.g., RNase L digestion products) and longer siRNAs (27– most likely due to early IFN production, because these genes are 31 bp), which both lack 59-triphosphate groups, can activate RIG-I under the regulation of type I and type III IFNs (30, 45). It was and IFN responses, respectively (19, 50, 51). Because IFN gene also of interest that none of the tested dsRNAs was able to induce expression and phosphorylation of IRF3 were inefficiently in- IL-10 gene expression, indicating that moDCs preferentially re- duced by dephosphorylated dsRNA and ssRNA molecules, it can spond to these danger signals by inducing IFN and proinflam- be expected that RNA sensing by RIG-I followed by IRF3 phos- matory cytokine-mediated responses. phorylation are the dominant signaling pathways in human MoDC responses to stimulation by ssRNAs were somewhat moDCs for the detection of foreign or short RNAs. different from those induced by dsRNAs. In moDCs (Figs. 3A,4B) Dephosphorylation of RNA molecules reduced the activation and in RLR- or TLR3-expressing HEK293 cells (Fig. 5B), IFN of the IFN-b and IFN-l1 promoters following dsRNA induction induction was less dependent on the length of ssRNAs as com- through the RIG-I and MDA5 signal pathway in HEK293 cells pared with dsRNAs. This may be due to a more complex sec- (Fig. 5). This is consistent with the data obtained in human ondary structure of ssRNAs, which, in contrast to linear dsRNA, moDCs. However, dephosphorylation completely abolished IFN can form various hairpin structures that are recognized by RNA activation through the TLR3 signal pathway as well (Fig. 5C). sensors, rendering the actual length of the ssRNA of secondary This strongly suggests that the 59-triphosphate group is essential importance. Surprisingly, at early time points of stimulation, long also for TLR3-dependent recognition of dsRNA. ssRNAs were able to elicit more efficient IRF3 phosphorylation Based on previous studies (2, 6, 7, 13, 14) and the data presented than short ssRNAs when equal molar amounts of ssRNAs were in this work, it appears that dsRNA recognition in human moDCs used (Fig. 4C). Interestingly, the induction of IFN mRNA ex- is clearly length dependent with short 59-triphosphorylated dsRNA pression remained at a similar level regardless of the ssRNA ap- molecules eliciting the strongest antiviral responses. Moreover, plied (Fig. 4B). It is therefore possible that in human moDCs, the 59-triphosphate group is essential for the recognition of short short ssRNAs induce antiviral responses simultaneously via mul- ssRNAs. Long ssRNA molecules, instead, are capable of activat- tiple signaling pathways, including TLR3, TLR7/8, RIG-I, and ing cytokine gene expression also in the absence of 59-phosphate MDA5, instead of using only the RIG-I/IRF3 signaling pathway groups. Furthermore, we found no evidence for viral sequence- (46, 47). Recent evidence (48, 49) suggests that in human moDCs, specific recognition of RNAs in human moDCs, because no major TLR3 and TLR7/8 ligands induce the expression of proin- differences were found between f6 versus IAV RNA-induced flammatory cytokines synergistically. However, for enhanced IFN responses. The findings presented in the present work expand gene expression, positive feedback via early IFN production and our knowledge on how human moDCs recognize RNA molecular The Journal of Immunology 1721 patterns and viral pathogens, thereby extending our understanding 25. Poranen, M. M., R. Tuma, and D. H. Bamford. 2005. Assembly of double- stranded RNA bacteriophages. Adv. Virus Res. 64: 15–43. of innate immune responses during viral infection and facilitating 26. Veckman, V., M. Miettinen, J. Pirhonen, J. Sire´n, S. Matikainen, and I. Julkunen. the design and development of RNA-based drugs. 2004. Streptococcus pyogenes and Lactobacillus rhamnosus differentially induce maturation and production of Th1-type cytokines and chemokines in human monocyte-derived dendritic cells. J. Leukoc. Biol. 75: 764–771. Acknowledgments 27. Gottlieb, P., J. Strassman, X. Qiao, M. Frilander, A. Frucht, and L. Mindich. 1992. In vitro packaging and replication of individual genomic segments of We thank Zhilin Li and Junwei Li for generating a chimera construct of the bacteriophage f 6RNA.J. Virol. 66: 2611–2616. influenza A virus cDNAs, as well as Linda Degerth for constructing plas- 28. Sambrook, J., and D. Russell. 2001. Molecular Cloning: A Laboratory Manual, mids pLD18 and pLD19. Riitta Tarkiainen is greatly acknowledged for ex- 3rd Ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. pert technical support in RNA production and purification. We are also 29. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. 227: 680–685. grateful to Sari Maljanen and Hanna Valtonen for excellent technical assis- 30. O¨ sterlund, P., V. Veckman, J. Sire´n, K. M. Klucher, J. Hiscott, S. Matikainen, and tance in cell experiments and protein and gene expression assays. I. Julkunen. 2005. Gene expression and antiviral activity of alpha/beta interfer- ons and interleukin-29 in virus-infected human myeloid dendritic cells. J. Virol. 79: 9608–9617. Disclosures 31. Matikainen, S., T. Ronni, M. Hurme, R. Pine, and I. Julkunen. 1996. Retinoic The authors have no financial conflicts of interest. acid activates interferon regulatory factor-1 gene expression in myeloid cells. Blood 88: 114–123. 32. Veckman, V., P. O¨ sterlund, R. Fagerlund, K. Mele´n, S. Matikainen, and I. Julkunen. 2006. TNF-a and IFN-a enhance influenza-A-virus-induced che- References mokine gene expression in human A549 lung epithelial cells. Virology 345: 96– 1. Janeway, C. A., Jr. 1989. Approaching the asymptote? Evolution and revolution 104. in immunology. Cold Spring Harb. Symp. Quant. Biol. 54: 1–13. 33. Pietila¨, T. E., V. Veckman, A. Lehtonen, R. Lin, J. Hiscott, and I. Julkunen. 2007. 2. Takeuchi, O., and S. Akira. 2009. Innate immunity to virus infection. Immunol. Multiple NF-kappaB and IFN regulatory factor family transcription factors Downloaded from Rev. 227: 75–86. regulate CCL19 gene expression in human monocyte-derived dendritic cells. J. 3. Ank, N., H. West, and S. R. Paludan. 2006. IFN-l: novel antiviral cytokines. J. Immunol. 178: 253–261. Interferon Cytokine Res. 26: 373–379. 34. Kumar, H., T. Kawai, and S. Akira. 2009. Pathogen recognition in the innate 4. Takeuchi, O., and S. Akira. 2008. MDA5/RIG-I and virus recognition. Curr. immune response. Biochem. J. 420: 1–16. Opin. Immunol. 20: 17–22. 35. DeWitte-Orr, S. J., D. R. Mehta, S. E. Collins, M. S. Suthar, M. Gale, Jr., and 5. Kawai, T., and S. Akira. 2009. The roles of TLRs, RLRs and NLRs in pathogen K. L. Mossman. 2009. Long double-stranded RNA induces an antiviral response recognition. Int. Immunol. 21: 317–337. independent of IFN regulatory factor 3, IFN-b promoter stimulator 1, and IFN. J.

6. McCartney, S. A., and M. Colonna. 2009. Viral sensors: diversity in pathogen Immunol. 183: 6545–6553. http://www.jimmunol.org/ recognition. Immunol. Rev. 227: 87–94. 36. Pichlmair, A., O. Schulz, C. P. Tan, T. I. Na¨slund, P. Liljestro¨m, F. Weber, and 7. Akira, S., S. Uematsu, and O. Takeuchi. 2006. Pathogen recognition and innate C. Reis e Sousa. 2006. RIG-I-mediated antiviral responses to single-stranded immunity. Cell 124: 783–801. RNA bearing 59-phosphates. Science 314: 997–1001. 8. Noppert, S. J., K. A. Fitzgerald, and P. J. Hertzog. 2007. The role of type I 37. Hornung, V., J. Ellegast, S. Kim, K. Brzo´zka, A. Jung, H. Kato, H. Poeck, interferons in TLR responses. Immunol. Cell Biol. 85: 446–457. S. Akira, K. K. Conzelmann, M. Schlee, et al. 2006. 59-Triphosphate RNA is the 9. Ghosh, S., M. J. May, and E. B. Kopp. 1998. NF-kB and Rel proteins: evolu- ligand for RIG-I. Science 314: 994–997. tionarily conserved mediators of immune responses. Annu. Rev. Immunol. 16: 38. Myong, S., S. Cui, P. V. Cornish, A. Kirchhofer, M. U. Gack, J. U. Jung, 225–260. K. P. Hopfner, and T. Ha. 2009. Cytosolic viral sensor RIG-I is a 59-triphosphate- 10. Yoneyama, M., and T. Fujita. 2009. RNA recognition and signal transduction by dependent translocase on double-stranded RNA. Science 323: 1070–1074. RIG-I-like receptors. Immunol. Rev. 227: 54–65. 39. Wang, Y., J. Ludwig, C. Schuberth, M. Goldeck, M. Schlee, H. Li, S. Juranek, 11. Fredericksen, B. L., and M. Gale Jr. 2006. West Nile virus evades activation of G. Sheng, R. Micura, T. Tuschl, et al. 2010. Structural and functional insights 9 interferon regulatory factor 3 through RIG-I-dependent and -independent path- into 5 -ppp RNA pattern recognition by the innate immune receptor RIG-I. Nat. by guest on September 25, 2021 ways without antagonizing host defense signaling. J. Virol. 80: 2913–2923. Struct. Mol. Biol. 17: 781–787. 12. Loo, Y. M., J. Fornek, N. Crochet, G. Bajwa, O. Perwitasari, L. Martinez- 40. Pichlmair, A., O. Schulz, C. P. Tan, J. Rehwinkel, H. Kato, O. Takeuchi, Sobrido, S. Akira, M. A. Gill, A. Garcı´a-Sastre, M. G. Katze, and M. Gale Jr. S. Akira, M. Way, G. Schiavo, and C. Reis e Sousa. 2009. Activation of MDA5 2008. Distinct RIG-I and MDA5 signaling by RNA viruses in innate immunity. J. requires higher-order RNA structures generated during virus infection. J. Virol. Virol. 82: 335–345. 83: 10761–10769. 13. Rehwinkel, J., and C. Reis e Sousa. 2010. RIGorous detection: exposing virus 41. Alexopoulou, L., A. C. Holt, R. Medzhitov, and R. A. Flavell. 2001. Recognition through RNA sensing. Science 327: 284–286. of double-stranded RNA and activation of NF-kappaB by Toll-like receptor 3. 14. Kato, H., O. Takeuchi, E. Mikamo-Satoh, R. Hirai, T. Kawai, K. Matsushita, Nature 413: 732–738. A. Hiiragi, T. S. Dermody, T. Fujita, and S. Akira. 2008. Length-dependent 42. Schlee, M., A. Roth, V. Hornung, C. A. Hagmann, V. Wimmenauer, W. Barchet, recognition of double-stranded ribonucleic acids by retinoic acid-inducible C. Coch, M. Janke, A. Mihailovic, G. Wardle, et al. 2009. Recognition of 59 gene-I and melanoma differentiation-associated gene 5. J. Exp. Med. 205: triphosphate by RIG-I requires short blunt double-stranded RNA as 1601–1610. contained in panhandle of negative-strand virus. Immunity 31: 25–34. 15. Yoneyama, M., and T. Fujita. 2010. Recognition of viral nucleic acids in innate 43. Schmidt, A., T. Schwerd, W. Hamm, J. C. Hellmuth, S. Cui, M. Wenzel, immunity. Rev. Med. Virol. 20: 4–22. F. S. Hoffmann, M. C. Michallet, R. Besch, K. P. Hopfner, et al. 2009. 59-Tri- 16. Steinman, R. M., and H. Hemmi. 2006. Dendritic cells: translating innate to phosphate RNA requires base-paired structures to activate antiviral signaling via adaptive immunity. Curr. Top. Microbiol. Immunol. 311: 17–58. RIG-I. Proc. Natl. Acad. Sci. USA 106: 12067–12072. 17. Iwasaki, A., and R. Medzhitov. 2010. Regulation of adaptive immunity by the 44. Yoneyama, M., and T. Fujita. 2008. Structural mechanism of RNA recognition . Science 327: 291–295. by the RIG-I-like receptors. Immunity 29: 178–181. 18. Sledz, C. A., M. Holko, M. J. de Veer, R. H. Silverman, and B. R. Williams. 45. Sire´n, J., J. Pirhonen, I. Julkunen, and S. Matikainen. 2005. IFN-a regulates 2003. Activation of the interferon system by short-interfering RNAs. Nat. Cell TLR-dependent gene expression of IFN-a, IFN-b, IL-28, and IL-29. J. Immunol. Biol. 5: 834–839. 174: 1932–1937. 19. Reynolds, A., E. M. Anderson, A. Vermeulen, Y. Fedorov, K. Robinson, 46. Diebold, S. S., T. Kaisho, H. Hemmi, S. Akira, and C. Reis e Sousa. 2004. Innate D. Leake, J. Karpilow, W. S. Marshall, and A. Khvorova. 2006. Induction of the antiviral responses by means of TLR7-mediated recognition of single-stranded interferon response by siRNA is cell type- and duplex length-dependent. RNA RNA. Science 303: 1529–1531. 12: 988–993. 47. Heil, F., H. Hemmi, H. Hochrein, F. Ampenberger, C. Kirschning, S. Akira, 20. Sioud, M. 2009. Deciphering the code of innate immunity recognition of siR- G. Lipford, H. Wagner, and S. Bauer. 2004. Species-specific recognition of NAs. Methods Mol. Biol. 487: 41–59. single-stranded RNA via Toll-like receptor 7 and 8. Science 303: 1526–1529. 21. Makeyev, E. V., and D. H. Bamford. 2000. Replicase activity of purified 48. Ma¨kela¨, S. M., M. Strengell, T. E. Pietila¨,P.O¨ sterlund, and I. Julkunen. 2009. recombinant protein P2 of double-stranded RNA bacteriophage f6. EMBO J. 19: Multiple signaling pathways contribute to synergistic TLR ligand-dependent 124–133. cytokine gene expression in human monocyte-derived macrophages and den- 22. Makeyev, E. V., and D. H. Bamford. 2000. The polymerase subunit of a dsRNA dritic cells. J. Leukoc. Biol. 85: 664–672. virus plays a central role in the regulation of viral RNA metabolism. EMBO J. 49. Ma¨kela¨, S. M., P. O¨ sterlund, and I. Julkunen. 2011. TLR ligands induce syner- 19: 6275–6284. gistic interferon-b and interferon-l 1 gene expression in human monocyte- 23. Laurila, M. R., E. V. Makeyev, and D. H. Bamford. 2002. Bacteriophage f 6 derived dendritic cells. Mol. Immunol. 48: 505–515. RNA-dependent RNA polymerase: molecular details of initiating nucleic acid 50. Malathi, K., B. Dong, M. Gale, Jr., and R. H. Silverman. 2007. Small self-RNA synthesis without primer. J. Biol. Chem. 277: 17117–17124. generated by RNase L amplifies antiviral innate immunity. Nature 448: 816–819. 24. Aalto, A. P., L. P. Sarin, A. A. van Dijk, M. Saarma, M. M. Poranen, U. Aruma¨e, 51. Marques, J. T., T. Devosse, D. Wang, M. Zamanian-Daryoush, P. Serbinowski, and D. H. Bamford. 2007. Large-scale production of dsRNA and siRNA pools R. Hartmann, T. Fujita, M. A. Behlke, and B. R. Williams. 2006. A structural for RNA interference utilizing bacteriophage f6 RNA-dependent RNA poly- basis for discriminating between self and nonself double-stranded RNAs in merase. RNA 13: 422–429. mammalian cells. Nat. Biotechnol. 24: 559–565.