Title: Nucleotide-binding Oligomerization Domain Proteins Are Innate Immune Receptors for Internalized Streptococcus pneumoniae
Abstract: Streptococcus pneumoniae, the major cause of community-acquired pneumonia and bacterial meningitis, has been shown to transiently invade epithelial and endothelial cells. Innate immune receptors including Toll-like receptors recognize various pathogens, such as S. pneumoniae, by identifying conserved pathogen-associated molecular patterns. Recently, two members of a novel class of pattern recognition receptors, the cytosolic proteins nucleotide-binding oligomerization domain 1 (Nod1)/CARD4 and Nod2/CARD15, have been found to detect cell wall peptidoglycans. Here we tested the hypothesis that Nod proteins are involved in the intracellular recognition of pneumococci. Data indicate that pneumococci invade HEK293 cells. Genetic complementation studies in these cells demonstrate that NF-κB activation induced by S. pneumoniae depends on Nod2. Moreover, intracellular transfection of inactivated pneumococci yielded similar effects, confirming the Nod2 dependence of NF-κB activation by pneumococci in HEK293 cells. By dominant negative overexpression and small interfering RNA experiments, we show for the first time that interleukin-1 receptor-associated kinase participates in Nod2-dependent NF-κB activation. Additionally, dominant negative interleukin-1 receptor-associated kinase 2, tumor necrosis factor receptor-associated factor 6, NF-κB-inducing kinase, transforming growth factor-β-activated kinase-binding protein 2, and transforming growth factor-β-activated kinase 1 also inhibited Nod2-dependent NF-κB activation. We finally demonstrate that in C57BL/6 mouse lung tissue in vivo as well as in the bronchial epithelial cell line BEAS-2B, Nod1 and Nod2 mRNA expressions were up-regulated after pneumococcal infection. Data presented suggest that Nod proteins contribute to innate immune recognition of S. pneumoniae. Furthermore, Rip-2 and members of the Toll-like receptor-signaling cascade are involved in the Nod2-dependent activation of NF-κB induced by pneumococci. Streptococcus pneumoniae, the major cause of community-acquired pneumonia and bacterial meningitis, has been shown to transiently invade epithelial and endothelial cells. Innate immune receptors including Toll-like receptors recognize various pathogens, such as S. pneumoniae, by identifying conserved pathogen-associated molecular patterns. Recently, two members of a novel class of pattern recognition receptors, the cytosolic proteins nucleotide-binding oligomerization domain 1 (Nod1)/CARD4 and Nod2/CARD15, have been found to detect cell wall peptidoglycans. Here we tested the hypothesis that Nod proteins are involved in the intracellular recognition of pneumococci. Data indicate that pneumococci invade HEK293 cells. Genetic complementation studies in these cells demonstrate that NF-κB activation induced by S. pneumoniae depends on Nod2. Moreover, intracellular transfection of inactivated pneumococci yielded similar effects, confirming the Nod2 dependence of NF-κB activation by pneumococci in HEK293 cells. By dominant negative overexpression and small interfering RNA experiments, we show for the first time that interleukin-1 receptor-associated kinase participates in Nod2-dependent NF-κB activation. Additionally, dominant negative interleukin-1 receptor-associated kinase 2, tumor necrosis factor receptor-associated factor 6, NF-κB-inducing kinase, transforming growth factor-β-activated kinase-binding protein 2, and transforming growth factor-β-activated kinase 1 also inhibited Nod2-dependent NF-κB activation. We finally demonstrate that in C57BL/6 mouse lung tissue in vivo as well as in the bronchial epithelial cell line BEAS-2B, Nod1 and Nod2 mRNA expressions were up-regulated after pneumococcal infection. Data presented suggest that Nod proteins contribute to innate immune recognition of S. pneumoniae. Furthermore, Rip-2 and members of the Toll-like receptor-signaling cascade are involved in the Nod2-dependent activation of NF-κB induced by pneumococci. Streptococcus pneumoniae, a human pathogen that frequently colonizes the upper respiratory tract, is the leading bacterial cause of community-acquired pneumonia and meningitis (1Bartlett J.G. Mundy L.M. N. Engl. J. Med. 1995; 333: 1618-1624Crossref PubMed Scopus (748) Google Scholar, 2Schuchat A. Robinson K. Wenger J.D. Harrison L.H. Farley M. Reingold A.L. Lefkowitz L. Perkins B.A. N. Engl. J. Med. 1997; 337: 970-976Crossref PubMed Scopus (1005) Google Scholar). In the United States alone, an estimated 40,000 deaths per year are caused by pneumococcal pneumoniae or meningitis (3Centers for Disease Control and Prevention MMWR Recomm. Rep. 1997; 46: 1-24Google Scholar). In order to infect different organs, S. pneumoniae crosses blood-tissue barriers by invading and transmigrating epithelial and endothelial cells (4Cundell D.R. Gerard N.P. Gerard C. Idanpaan-Heikkila I. Tuomanen E.I. Nature. 1995; 377: 435-438Crossref PubMed Scopus (615) Google Scholar, 5Ring A. Weiser J.N. Tuomanen E.I. J. Clin. Invest. 1998; 102: 347-360Crossref PubMed Scopus (315) Google Scholar). Hereby, the cell surface protein platelet-activating factor (PAF) 1The abbreviations used are: PAF, platelet-activating factor; CARD, caspase recruitment domain; dn, dominant negative; IRAK, interleukin-1 receptor-associated kinase; MyD88, myeloid differentiation protein 88; Mal, MyD88 adapter-like; MDP, muramyl dipeptide; NIK, NF-κB-inducing kinase; Nod, nucleotide-binding oligomerization domain; Rip2, receptor-interacting protein 2; TAB, transforming growth factor-β-activated kinase-binding protein; TAK, transforming growth factor-β-activated kinase; TLR, Toll-like receptors; TRAF, tumor necrosis factor receptor-associated factor; TRIF, TIR domain-containing adapter-inducing interferon-β; TNF, tumor necrosis factor; PBS, phosphate-buffered saline; CFU, colony-forming units; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; siRNA, small interfering RNA; RT, reverse transcription. receptor as well as phosphorylcholin components of the pneumococcal cell wall are involved (4Cundell D.R. Gerard N.P. Gerard C. Idanpaan-Heikkila I. Tuomanen E.I. Nature. 1995; 377: 435-438Crossref PubMed Scopus (615) Google Scholar). The innate immune system recognizes pathogens by sensing so-called pathogen-associated molecular patterns via its pattern recognition receptors. The best studied innate immune receptors are the Toll-like receptors (TLRs), which also have been found to be involved in defense against S. pneumoniae (6Koedel U. Angele B. Rupprecht T. Wagner H. Roggenkamp A. Pfister H.W. Kirschning C.J. J. Immunol. 2003; 170: 438-444Crossref PubMed Scopus (198) Google Scholar, 7Malley R. Henneke P. Morse S.C. Cieslewicz M.J. Lipsitch M. Thompson C.M. Kurt-Jones E. Paton J.C. Wessels M.R. Golenbock D.T. Proc. Natl. Acad. Sci. U. S. A. 2003; 100: 1966-1971Crossref PubMed Scopus (571) Google Scholar, 8Schroder N.W. Morath S. Alexander C. Hamann L. Hartung T. Zahringer U. Gobel U.B. Weber J.R. Schumann R.R. J. Biol. Chem. 2003; 278: 15587-15594Abstract Full Text Full Text PDF PubMed Scopus (500) Google Scholar, 9Yoshimura A. Lien E. Ingalls R.R. Tuomanen E. Dziarski R. Golenbock D. J. Immunol. 1999; 163: 1-5Crossref PubMed Google Scholar). Additionally, we could recently show that lipopolysaccharide binding protein has an important role in defense against pneumococci (10Weber J.R. Freyer D. Alexander C. Schroder N.W. Reiss A. Kuster C. Pfeil D. Tuomanen E.I. Schumann R.R. Immunity. 2003; 19: 269-279Abstract Full Text Full Text PDF PubMed Scopus (89) Google Scholar). Another subset of pattern recognition receptors are the recently identified nucleotide-binding oligomerization domain (Nod) proteins, molecules that have been implicated in intracellular recognition of pathogens (11Girardin S.E. Tournebize R. Mavris M. Page A.L. Li X. Stark G.R. Bertin J. DiStefano P.S. Yaniv M. Sansonetti P.J. Philpott D.J. EMBO Rep. 2001; 2: 736-742Crossref PubMed Scopus (530) Google Scholar). Whereas Nod1, also called caspase recruitment domain (CARD)-4 protein, has been found to recognize peptidoglycans containing meso-diaminopimelate acid found mainly in Gram-negative bacteria (12Chamaillard M. Hashimoto M. Horie Y. Masumoto J. Su Q. Saab L. Ogura Y. Kawasaki A. Fukase K. Kusumoto S. Valvano M.A. Foster S.J. Mak T.W. Nunez G. Inohara N. Nat. Immunol. 2003; 4: 702-707Crossref PubMed Scopus (1013) Google Scholar, 13Girardin S.E. Boneca I.G. Carneiro L.A. Antignac A. Jehanno M. Viala J. Tedin K. Taha M.K. Labigne A. Zahringer U. Coyle A.J. DiStefano P.S. Bertin J. Sansonetti P.J. Philpott D.J. Science. 2003; 300: 1584-1587Crossref PubMed Scopus (1257) Google Scholar), Nod2 (CARD15) mediates responsiveness to the muramyldipeptide MurNAc-l-Ala-d-iso-Gln (MDP) conserved in peptidoglycans of basically all bacteria (14Girardin S.E. Boneca I.G. Viala J. Chamaillard M. Labigne A. Thomas G. Philpott D.J. Sansonetti P.J. J. Biol. Chem. 2003; 278: 8869-8872Abstract Full Text Full Text PDF PubMed Scopus (1916) Google Scholar, 15Inohara N. Ogura Y. Fontalba A. Gutierrez O. Pons F. Crespo J. Fukase K. Inamura S. Kusumoto S. Hashimoto M. Foster S.J. Moran A.P. Fernandez-Luna J.L. Nunez G. J. Biol. Chem. 2003; 278: 5509-5512Abstract Full Text Full Text PDF PubMed Scopus (1406) Google Scholar). Nod1 and Nod2 are composed of one or two N-terminal CARDs, a centrally located Nod, and a C-terminal leucine-rich repeat domain (16Bertin J. Nir W.J. Fischer C.M. Tayber O.V. Errada P.R. Grant J.R. Keilty J.J. Gosselin M.L. Robison K.E. Wong G.H. Glucksmann M.A. DiStefano P.S. J. Biol. Chem. 1999; 274: 12955-12958Abstract Full Text Full Text PDF PubMed Scopus (311) Google Scholar, 17Inohara N. Koseki T. del Peso L. Hu Y. Yee C. Chen S. Carrio R. Merino J. Liu D. Ni J. Nunez G. J. Biol. Chem. 1999; 274: 14560-14567Abstract Full Text Full Text PDF PubMed Scopus (628) Google Scholar, 18Ogura Y. Inohara N. Benito A. Chen F.F. Yamaoka S. Nunez G. J. Biol. Chem. 2001; 276: 4812-4818Abstract Full Text Full Text PDF PubMed Scopus (1170) Google Scholar). Nod1 is expressed in multiple tissues, whereas Nod2 expression is mainly restricted to leukocytes, dendritic cells, and epithelial cells (17Inohara N. Koseki T. del Peso L. Hu Y. Yee C. Chen S. Carrio R. Merino J. Liu D. Ni J. Nunez G. J. Biol. Chem. 1999; 274: 14560-14567Abstract Full Text Full Text PDF PubMed Scopus (628) Google Scholar, 19Gutierrez O. Pipaon C. Inohara N. Fontalba A. Ogura Y. Prosper F. Nunez G. Fernandez-Luna J.L. J. Biol. Chem. 2002; 277: 41701-41705Abstract Full Text Full Text PDF PubMed Scopus (397) Google Scholar). Furthermore, Nod1 mRNA has been shown to be up-regulated by interferon-γ in intestinal epithelial cells, and Nod2 expression was increased in vitro by various inflammatory stimuli like TNF-α, interferon-γ, or lipopolysaccharide (19Gutierrez O. Pipaon C. Inohara N. Fontalba A. Ogura Y. Prosper F. Nunez G. Fernandez-Luna J.L. J. Biol. Chem. 2002; 277: 41701-41705Abstract Full Text Full Text PDF PubMed Scopus (397) Google Scholar, 20Hisamatsu T. Suzuki M. Podolsky D.K. J. Biol. Chem. 2003; 278: 32962-32968Abstract Full Text Full Text PDF PubMed Scopus (83) Google Scholar, 21Rosenstiel P. Fantini M. Brautigam K. Kuhbacher T. Waetzig G.H. Seegert D. Schreiber S. Gastroenterology. 2003; 124: 1001-1009Abstract Full Text Full Text PDF PubMed Scopus (365) Google Scholar). Currently, the Nod-dependent signaling cascade is poorly understood. After activation of Nod proteins, the Rip2 kinase (also called RICK or CARDIAK) interacts through homophilic CARD-CARD interactions and eventually leads to nuclear translocation of the transcription factor NF-κB via IκB release by the IκB kinase complex (16Bertin J. Nir W.J. Fischer C.M. Tayber O.V. Errada P.R. Grant J.R. Keilty J.J. Gosselin M.L. Robison K.E. Wong G.H. Glucksmann M.A. DiStefano P.S. J. Biol. Chem. 1999; 274: 12955-12958Abstract Full Text Full Text PDF PubMed Scopus (311) Google Scholar, 17Inohara N. Koseki T. del Peso L. Hu Y. Yee C. Chen S. Carrio R. Merino J. Liu D. Ni J. Nunez G. J. Biol. Chem. 1999; 274: 14560-14567Abstract Full Text Full Text PDF PubMed Scopus (628) Google Scholar, 18Ogura Y. Inohara N. Benito A. Chen F.F. Yamaoka S. Nunez G. J. Biol. Chem. 2001; 276: 4812-4818Abstract Full Text Full Text PDF PubMed Scopus (1170) Google Scholar). However, there might be intermediate signaling molecules leading to NF-κB activation by Nod proteins that are still unknown. Because pneumococci penetrate epithelial as well as endothelial cells and Nod proteins mediate intracellular detection of bacterial components, we tested the hypothesis that Nod proteins are involved in host recognition of S. pneumoniae. Overexpression experiments indicate that Nod2 recognizes S. pneumoniae. We furthermore show that Nod protein expression is up-regulated in vivo and in vitro by pneumococcal infection. Moreover, this study suggests that in addition to Rip2, the signal-transducing molecules IRAK, IRAK2, TRAF6, NIK, TAB2, and TAK1 are involved in NF-κB activation by Nod proteins in pneumococci-stimulated cells. Bacterial Strains—Capsulated S. pneumoniae serotype 3 (NCTC 7978) were used for the murine pneumonia model. Bacteria were incubated at 37 °C with 5% CO2 on Columbia agar with 5% sheep blood at 37 °C. Single colonies were expanded by resuspension in Todd-Hewitt broth supplemented with 0.5% yeast extract and incubation at 37 °C for 3–4 h to midlogarithmic phase (A600 = 0.2–0.4). Bacteria were harvested by centrifugation at 1500 × g for 15 min at 4 °C and were washed twice in sterile PBS. Bacteria were then resuspended in sterile PBS, at ∼2.5 × 108 CFU/ml, as determined by plating serial 10-fold dilutions onto Columbia-agar plates. The S. pneumoniae R6x is the unencapsulated derivate of serotype 2 strain D39 (22Tiraby J.G. Fox M.S. Proc. Natl. Acad. Sci. U. S. A. 1973; 70: 3541-3545Crossref PubMed Scopus (124) Google Scholar). A pneumolysin-negative mutant of R6x ply– was generated by insertion-duplication mutagenesis using the pJDC9:ply construct. After transformation of the plasmid into S. pneumoniae strain R6x, erythromycin-resistant transformants were selected with 1 μg/ml erythromycin of Luria-Bertani agar containing 5% sheep blood. Insertion of the plasmid into the pneumolysin-encoding gene ply was confirmed by PCR analysis and DNA sequencing. For cell culture stimulation studies, R6x or R6x ply– pneumococci were expanded as described above, harvested by centrifugation, and resuspended in suitable cell culture medium or inactivated by heating at 56 °C for 30 min and sonication. The ethanol inactivation was performed as described before (7Malley R. Henneke P. Morse S.C. Cieslewicz M.J. Lipsitch M. Thompson C.M. Kurt-Jones E. Paton J.C. Wessels M.R. Golenbock D.T. Proc. Natl. Acad. Sci. U. S. A. 2003; 100: 1966-1971Crossref PubMed Scopus (571) Google Scholar). HEK293 Cell Invasion Assay and PAF Receptor Overexpression—The cDNA encoding the human PAF receptor (kindly provided by T. Shimizu, Tokyo, Japan) was subcloned into the human expression vector cDNA3.1 (Invitrogen). Human embryonic kidney HEK293 cells (ATCC) were cultured in 12-well plates with Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% fetal calf serum. In some experiments, subconfluent cells were transfected with 0.3 μg of PAF receptor plasmid using Effectene transfection reagent (Qiagen, Hilden, Germany). Pneumococcal invasion of HEK293 was assessed by antibiotic protection assay as described previously for endothelial and epithelial cells (4Cundell D.R. Gerard N.P. Gerard C. Idanpaan-Heikkila I. Tuomanen E.I. Nature. 1995; 377: 435-438Crossref PubMed Scopus (615) Google Scholar, 5Ring A. Weiser J.N. Tuomanen E.I. J. Clin. Invest. 1998; 102: 347-360Crossref PubMed Scopus (315) Google Scholar). Briefly, HEK293 cells were infected with 106 CFU/ml S. pneumoniae strain R6x ply–. After the indicated time intervals, cells were washed three times with PBS and incubated for 1 h with culture medium supplemented with penicillin G (10 μg/ml) and gentamicin (200 μg/ml) to kill extracellular bacteria. Subsequently, HEK293 cells were washed three times again with PBS and lysed by using ice-cold 0.025% Triton X-100, and 100-μl aliquots were plated on Columbia agar with 5% sheep blood to assess numbers of intracellular pneumococci. In a control, the antibiotics used were sufficient to kill 5 × 107 CFU/ml S. pneumoniae in 1 h. Expression Plasmids, HEK293 Cell Transfection, and Luciferase Assays—Expression plasmids for Nod1 and Nod2 were kindly provided by G. Nuñez (Ann Arbor, MI). Expression vectors encoding wild-type Myc-Rip2 and dominant negative (dn) Rip-2 were a generous gift from P. Dempsey (Los Angeles, CA). The constructs encoding dnMyD88, dnIRAK, dnIRAK2, dnTRAF6, and dnNIK were from C. Kirschning (Munich, Germany) and Tularik Inc. (San Francisco, CA); dnTAK1 and dnTAB2 were a kind gift from Tohru Ishitani (Nagoya, Japan). dnMal and dnTRIF were from K. Fitzgerald (Golenbock Laboratory, University of Massachusetts, Worchester, MA). Subconfluent HEK293 cells were cotransfected the following day using the calcium phosphate method according to the manufacturer's instructions (Clontech) with 0.1 μg of NF-κB-dependent luciferase reporter (23Krull M. Klucken A.C. Wuppermann F.N. Fuhrmann O. Magerl C. Seybold J. Hippenstiel S. Hegemann J.H. Jantos C.A. Suttorp N. J. Immunol. 1999; 162: 4834-4841PubMed Google Scholar), 0.1 μg of RSV β-galactosidase plasmid, 0.1 μg of PAF receptor, and 0.3 ng of Nod or 0.5 μg of Rip2 expression vectors. In some experiments, 0.5 μg/ml MDP (Calbiochem) or 5 × 107 CFU/ml inactivated S. pneumoniae were added to the cells prior to transfection to allow their entry into the cells. When indicated, cells were additionally transfected with 0.1 μg of dominant negative mutants of several signaling molecules. The experiment shown in Fig. 4B was performed using 5, 25, and 125 ng of dnIRAK or dnIRAK2. In some experiments, cells were incubated with S. pneumoniae strain R6x ply– for 5 h on the following day. Luciferase activity was measured by using the Luciferase Reporter-Gene Assay (Promega, Mannheim, Germany), and results were normalized with values obtained by RSV-β-galactosidase or protein concentrations (Fig. 1A).Fig. 1Internalization of HEK293 cells by S. pneumoniae and effects of Nod1 and Nod2 overexpression on NF-κB activation. A, HEK293 cells were transiently transfected with a control vector, Nod1 or Nod2 expression plasmids, and plasmids encoding PAF receptor and an NF-κB-dependent luciferase reporter. The next day, cells were infected with the indicated doses of S. pneumoniae for 5 h. Relative luciferase activities are shown as means of -fold inductions ± S.E. for one representative experiment of three, each transfection performed in triplicate. B, HEK293 cells overexpressing PAF receptor were infected with S. pneumoniae for the indicated time intervals, and an antibiotic protection assay was performed. The number of internalized pneumococci was determined by assessing CFU surviving antibiotic treatment.View Large Image Figure ViewerDownload (PPT) RNA Interference—Control nonsilencing siRNA (sense, UUCUCCGAACGUGUCACGUtt; antisense, ACGUGACACGUUCGGAGGAGA-Att) and different siRNAs targeting IRAK were purchased from Ambion (Huntingdon, UK). Different siRNAs were checked for their ability to silence IRAK protein expression in HEK293 cells by Western blots using anti-human IRAK antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) 72 h after siRNA transfection (RNAiFect; Qiagen). Sequence 3 (sense, GGUUGUCCUUGAGUAAUAAtt; antisense, UUAUUACUCAAGGACAACCtg) was most effective in inhibition of protein expression and was therefore used for further studies. 72 h before cDNA transfections, siRNA duplexes (final concentration 100 nm) were transfected by using RNAiFect (Qiagen). Murine Model of Pneumococcal Pneumonia—Pathogen-free, female C57BL/6 mice were obtained from Charles River (Sulzfeld, Germany). The animals were kept with a 12-h light/dark cycle and were given free access to food and water. The study was approved by the Institutional Review Board for the care of animal subjects. Mice were challenged at 10 weeks of age and 19–20 g of weight. Mice were lightly anesthetized by intraperitoneal injection of ketamine and xylazine and were inoculated intranasally with 20 μl of bacterial suspension (5 × 106 CFU of S. pneumoniae serotype 3). Control mice were challenged with 20 μl of sterile PBS. Groups of three mice were sacrificed 6, 12, 24, or 48 h postinfection, and lungs were removed aseptically and immediately frozen in liquid nitrogen. Control mice were sacrificed 6 h postinfection (n = 3). Frozen lung tissue was crushed using a sterile, nitrogen-cooled high grade steel plate homogenizer. Total RNA was extracted using an RNeasy Mini Isolation kit (Quiagen) following the manufacturer's protocol. For cDNA synthesis, 2 μg of pooled RNA from each group was reversed transcribed using a Pro Stare First Strand RT-PCR kit (Stratagene Europe, Amsterdam, The Netherlands) following the manufacturer's protocol by using oligo(dT) primers. Murine glyceraldehyde 3-phosphate dehydrogenase (GAPDH) served as housekeeping gene. After 35 (Nod1) or 40 (Nod2) PCR amplification cycles, samples were resolved on a 1.2% agarose gel. The PCR primers for Nod1 (expected size 469 bp) were 5′-CTTGCATTCAATGGCATCTC-3′ and 5′-ACATCGGTGTGCACTGTGGA-3′; primers for Nod2 (expected size 515 bp) were 5′-AGCAGAACTTCTTGTCCCTGA-3′ and 5′-TCACAACAAGAGTCTGGCGT-3′; and primers for GAPDH (expected size 496 bp) were 5′-TGATGGGTGTGAACCACGAG-3′ and 5′-TCAGTGTAGCCCAAGATGCC-3′. Total RNA from lipopolysaccharide-stimulated murine macrophages (RAW 264.7; murine macrophage cell line) served as positive control. Stimulation of BEAS-2B Cells, mRNA Isolation, and Reverse Transcription-PCR—The human bronchial epithelial cell line BEAS-2B (a generous gift of Dr. Curtis Harris, Bethesda, MD) was cultured in keratinocyte medium supplemented according to the manufacturer's instructions (ATCC, Manassas, VA). Cells were grown to confluence in 3.5-cm well plates, subsequently infected with 106 CFU/ml S. pneumoniae strain R6x for different durations, stimulated with 5 × 107 CFU/ml heat-inactivated pneumococci or 20 ng/ml TNF-α (R&D Systems, Wiesbaden, Germany) for 24 h. Total RNA was isolated with the RNeasy minikit (Qiagen) and reverse transcribed using avian myeloblastosis virus reverse transcriptase (Promega, Heidelberg, Germany). The generated cDNA was amplified by PCR using specific primers (Nod1-sense, 5′-AAGCGAAGAGCTGACCAAAT-3′; Nod1-antisense, 5′-TTCATAGACTTTGGCCTCCTC-3′; Nod2-sense, 5′-AGCCATTGTCAGGAGGCTC-3′; Nod2-antisense, 5′-CGTCTCTGCTCCATCATAGG-3′). All primers were purchased from TIB MOLBIOL (Berlin, Germany). After 35 (Nod1) or 40 (Nod2) amplification cycles, the PCR products were analyzed on 1.5% agarose gels, stained with ethidium bromide, and subsequently visualized. To confirm equal amounts of RNA in each experiment, all samples were checked for GAPDH mRNA expression. Statistics—Nod2-dependent stimulatory effects of active or inactive S. pneumoniae as well as inhibitory effects of dominant negative signaling molecule mutants were statistically evaluated employing Student's t test. Throughout the figures, p values of <0.05 are indicated by one asterisk, and p values of <0.01 are shown by double asterisks. S. pneumoniae Invade HEK293 Cells and Activate NF-κB in a Nod2-dependent Manner—To explore a possible function of Nod proteins in host recognition of S. pneumoniae, we performed overexpression assays in HEK293 cells. This cell line has been described as not expressing TLR2, a well known receptor for Gram-positive patterns (24Kirschning C.J. Wesche H. Merrill Ayres T. Rothe M. J. Exp. Med. 1998; 188: 2091-2097Crossref PubMed Scopus (655) Google Scholar). Moreover, we could not detect any Nod2 mRNA and detected only very low levels of Nod1 mRNAs in these cells by RT-PCR analysis (data not shown). We transiently transfected the cells with an NF-κB-dependent luciferase reporter construct and Nod1 or Nod2, respectively. Additionally, since the PAF receptor has been shown to be involved in internalization of pneumococci (4Cundell D.R. Gerard N.P. Gerard C. Idanpaan-Heikkila I. Tuomanen E.I. Nature. 1995; 377: 435-438Crossref PubMed Scopus (615) Google Scholar, 5Ring A. Weiser J.N. Tuomanen E.I. J. Clin. Invest. 1998; 102: 347-360Crossref PubMed Scopus (315) Google Scholar), we cotransfected a plasmid encoding the PAF receptor. On the next day, cells were exposed to different concentrations of viable S. pneumoniae strain R6x ply–. We used this pneumolysin-negative mutant for HEK cell experiments in order to avoid pneumolysin-induced damage of HEK293 cells. The bacteria failed to cause induction of reporter gene activity if only a control vector was transfected (Fig. 1A). After overexpressing of Nod1 or Nod2, we observed a constitutive reporter activity, which is in line with previously published results (14Girardin S.E. Boneca I.G. Viala J. Chamaillard M. Labigne A. Thomas G. Philpott D.J. Sansonetti P.J. J. Biol. Chem. 2003; 278: 8869-8872Abstract Full Text Full Text PDF PubMed Scopus (1916) Google Scholar). Pneumococcal infection further enhanced the reporter gene activity in Nod2-transfected cells in a dose-dependent manner, thus indicating a role of Nod2 in NF-κB activation by pneumococci. Nod2-dependent NF-κB activation by pneumococci was less pronounced in cells that were transfected with an empty control vector instead of the PAF receptor plasmid, indicating the importance of cell invasion of pneumococci for Nod signaling (data not shown). Taken together, the data indicate that Nod2 is involved in recognition of internalized S. pneumoniae. To further confirm internalization of pneumococci into HEK293 cells, we performed an antibiotic protection assay as described previously (4Cundell D.R. Gerard N.P. Gerard C. Idanpaan-Heikkila I. Tuomanen E.I. Nature. 1995; 377: 435-438Crossref PubMed Scopus (615) Google Scholar, 5Ring A. Weiser J.N. Tuomanen E.I. J. Clin. Invest. 1998; 102: 347-360Crossref PubMed Scopus (315) Google Scholar). As seen in Fig. 1B, S. pneumoniae strain R6x ply– was internalized into HEK293 cells overexpressing PAF receptor in a time-dependent manner. Nod2 Mediates NF-κB Activation by Inactivated S. pneumoniae—To further characterize the role of Nod proteins in the recognition of pneumococci, we used heat-inactivated, sonicated, or ethanol-inactivated pneumococci; MDP served as a positive control. In contrast to viable bacteria, inactivated pneumococci as well as MDP failed to induce NF-κB activation in Nod-transfected cells when extracellular stimulation was performed (data not shown), an observation compatible with the notion that Nod2 detects only intracellularly located pathogen-derived material. We therefore assessed Nod-dependent NF-κB activity after allowing the internalization of bacterial components into the cells (as described under "Experimental Procedures"). Under these conditions, Nod2 conferred a dose-dependent responsiveness to inactivated pneumococci and MDP (Fig. 2, A and B). Hereby, the ethanol-inactivated pneumococci were more potent then the heat-inactivated pneumococci, displaying Nod2 activity already at a concentration of 5 × 105 CFU/ml. Again, we observed a certain constitutive activity caused by overexpression of Nod1 and Nod2. The different quantities in these background activities may be caused by slightly variable transfection conditions. Involvement of Rip2, but Not of the TLR Adapter Proteins, in the NF-κB Activation by Inactivated S. pneumoniae and MDP—Following our observation that Nod2, and not the TLRs, mediates NF-κB induction by S. pneumoniae in our assay, we tried to further exclude any TLR involvement by using dominant negative mutants of TLR adapter proteins (25O'Neill L.A. Curr. Top. Microbiol. Immunol. 2002; 270: 47-61Crossref PubMed Scopus (250) Google Scholar). Therefore, we cotransfected HEK293 cells with an NF-κB-dependent reporter gene, Nod2, as well as dominant negative mutants of different signaling molecules and stimulated these cells with heat-inactivated pneumococci or MDP. Vectors containing dominant negative mutants of the TLR adapter molecules MyD88, Mal, and TRIF failed to exhibit significant inhibitory effects on Nod2-dependent NF-κB activation (Fig. 3). As expected and shown earlier, dnMyD88, dnMal, and dnTRIF clearly inhibited TLR-dependent NF-κB activity brought about by respective TLR ligands (24Kirschning C.J. Wesche H. Merrill Ayres T. Rothe M. J. Exp. Med. 1998; 188: 2091-2097Crossref PubMed Scopus (655) Google Scholar, 26Yamamoto M. Sato S. Mori K. Hoshino K. Takeuchi O. Takeda K. Akira S. J. Immunol. 2002; 169: 6668-6672Crossref PubMed Scopus (1022) Google Scholar) (data not shown). In contrast, dominant negative mutants of Rip2, a kinase known to associate with Nod proteins, suppressed Nod2-dependent NF-κB activation induced by inactivated pneumococci and MDP. We thus conclude that Nod proteins and Rip2, but not the TLRs or their adapter molecules, mediate NF-κB activation in our assay. The Signaling Cascade Downstream of Nod2—Little is known about the further downstream signaling cascade of Nod2. In order to identify potential molecules involved, experiments using dominant negative mutants of different signaling molecules were carried out. As shown in Fig. 4A, dnIRAK, dnIRAK2, dnTRAF6, dnNIK, dnTAB2, and dnTAK1 suppressed the Nod2-dependent NF-κB activity induced by pneumococcal components and MDP. Moreover, dominant negative mutants of IRAK and IRAK2 inhibited NF-κB activation in a dose-dependent manner (Fig. 4B). To further confirm the unexpected role of IRAK in Nod2 signaling, gene silencing experiments using siRNA were performed. To assess gene silencing activities, HEK293 cells were transfected with different siRNAs targeting IRAK as well as nonsilencing siRNA for 24 h, and Western blot analyses using anti-IRAK antibodies were performed 72 h after transfection. Sequence 3 was most effective in gene silencing (Fig. 4C) and was therefore used in subsequent reporter gene assays. As can be seen in Fig. 4D, siRNA targeting IRAK inhibited the NF-κB activation brought about by overexpression of Nod2. In contrast, IRAK siRNA duplexes had no inhibitory effect on TNF-α response. IRAK, IRAK2, TRAF6, NIK, TAB2, and TAK1 Are Located Downstream of Nod2 and Rip2—To examine whether IRAK, IRAK2, NIK, TAB2 and TAK1 lie upstream or downstream of Rip2, plasmids encoding wild-type RIP2, an NF-κB reporter, and the dominant negative mutants of signaling molecules examined were cotransfected in HEK293 cells. Overexpression of Rip2 leads to a strong NF-κB activation (Fig. 5). In line with the above mentioned studies, the reporter gene activation could not be suppressed by dnMyD88, dnMal, and dnTRIF. Furthermore, dnIRAK, dnIRAK2, dnTRAF6, dnNIK, dnTAB2, and dnTAK1 inhibited Rip2-induced NF-κB activation. Thus, IRAK, IRAK2, TRAF6, NIK, TAB2, and TAK1 seem to be located downstream of Rip2 and play essential roles in transducing information from Nod2 to NF-κB activation. Nod1 and Nod2 mRNA Levels Were Up-regulated in Vivo and in Vitro after Pneumococcal Infection—Finally, we analyzed the expression of Nod1 and Nod2 in mouse lung tissue and examined whether their mRNA levels changed during pneumococcal infection in vivo. C57BL/6 mice were infected with capsulated S. pneumoniae serotype 3 for different durations, and subsequently total RNA of mouse lung tissue was isolated and RT-PCR was performed. Our analyses detected Nod1 transcripts in lung tissue, whereas the expression of Nod2 was lower (Fig. 6A). After pneumococcal infection, a marked increase of Nod1 and Nod2 mRNA levels could be observed. To further confirm up-regulation of Nod1 and Nod2 expression during infection with S. pneumoniae, we used the human bronchial epithelial cell line BEAS-2B. Cells were infected with S. pneumoniae strain R6x or stimulated with heat-inactivated pneumococci or TNF-α. The levels of Nod1 and Nod2 transcripts were low in untreated epithelial cells examined (Fig. 6B). Increased Nod1 and Nod2 mRNA levels could be detected after infection. Furthermore, Nod1 and Nod2 mRNA up-regulation was also observed after stimulation with inactivated pneumococci for 24 h and when TNF-α was used as a control (Fig. 6B). Thus, in a mouse pneumococcal pneumonia model and in a bronchial epithelial cell line, Nod1 and Nod2 expression was up-regulated, suggesting a physiological role of Nod proteins in host defense against pneumococci. In the presented study, we show for the first time a role of Nod proteins in host recognition of S. pneumoniae. Overexpression experiments indicate that Nod2 mediates NF-κB activation by viable S. pneumoniae or when inactivated pneumococci were transfected within the cells. Moreover, Rip2, IRAK, IRAK2, TRAF6, NIK, TAB2, and TAK1 seemed to be involved in Nod2-dependent NF-κB activation. We finally demonstrate that Nod1 and Nod2 expression was up-regulated after pneumococcal infection in vivo and in vitro, suggesting a physiological role of Nod proteins in immune response against pneumococci in pulmonary tissue. TLR2 and TLR4 have been associated with host recognition of pneumococci (6Koedel U. Angele B. Rupprecht T. Wagner H. Roggenkamp A. Pfister H.W. Kirschning C.J. J. Immunol. 2003; 170: 438-444Crossref PubMed Scopus (198) Google Scholar, 7Malley R. Henneke P. Morse S.C. Cieslewicz M.J. Lipsitch M. Thompson C.M. Kurt-Jones E. Paton J.C. Wessels M.R. Golenbock D.T. Proc. Natl. Acad. Sci. U. S. A. 2003; 100: 1966-1971Crossref PubMed Scopus (571) Google Scholar, 8Schroder N.W. Morath S. Alexander C. Hamann L. Hartung T. Zahringer U. Gobel U.B. Weber J.R. Schumann R.R. J. Biol. Chem. 2003; 278: 15587-15594Abstract Full Text Full Text PDF PubMed Scopus (500) Google Scholar, 9Yoshimura A. Lien E. Ingalls R.R. Tuomanen E. Dziarski R. Golenbock D. J. Immunol. 1999; 163: 1-5Crossref PubMed Google Scholar). Whereas these TLRs are likely to serve as the first line receptors for S. pneumoniae, the Nod proteins might play a major role in a subsequent phase of infection. Since TLRs mediate NF-κB activation and NF-κB binding sites have been identified in the Nod2 promoter (19Gutierrez O. Pipaon C. Inohara N. Fontalba A. Ogura Y. Prosper F. Nunez G. Fernandez-Luna J.L. J. Biol. Chem. 2002; 277: 41701-41705Abstract Full Text Full Text PDF PubMed Scopus (397) Google Scholar, 21Rosenstiel P. Fantini M. Brautigam K. Kuhbacher T. Waetzig G.H. Seegert D. Schreiber S. Gastroenterology. 2003; 124: 1001-1009Abstract Full Text Full Text PDF PubMed Scopus (365) Google Scholar), recognition of pneumococci by the TLRs might cause the up-regulation of Nod2 (and probably even Nod1) and thereby facilitate the immune response of the host against this pathogen. In line with this hypothesis, the penetration of epithelial and endothelial cells by S. pneumoniae is initiated during the first hours after infection, and it is most pronounced after 4–6 h (5Ring A. Weiser J.N. Tuomanen E.I. J. Clin. Invest. 1998; 102: 347-360Crossref PubMed Scopus (315) Google Scholar). mRNA levels of Nod1 and Nod2 increased within a similar time frame, suggesting that Nod-mediated NF-κB activation might play an important role in this subsequent phase of host responses against S. pneumoniae. The downstream signaling cascade induced by Nod activation is currently poorly understood. By two different approaches, such as dominant negative overexpression and siRNA experiments, we demonstrated that IRAK participates in Nod2 signaling. Moreover, data presented suggest that molecules like IRAK2, TRAF6, NIK, TAB2, and TAK1 are additionally involved in NF-κB activation and seemed to be located downstream of Nod2 and Rip2. Thus, Nod-mediated signaling might be partly homologous to TLR/MyD88 downstream signaling (25O'Neill L.A. Curr. Top. Microbiol. Immunol. 2002; 270: 47-61Crossref PubMed Scopus (250) Google Scholar). Importantly, the TLR adapter proteins MyD88, Mal, and TRIF did not participate in NF-κB activation by Nod2, indicating that Nod and TLR pathways merge down-stream of MyD88, Mal, and TRIF. In the Rip2 knockout mice, even TLR-dependent signal transduction was attenuated, showing that Rip2 is not exclusively involved in Nod-regulated cell activation (27Chin A.I. Dempsey P.W. Bruhn K. Miller J.F. Xu Y. Cheng G. Nature. 2002; 416: 190-194Crossref PubMed Scopus (331) Google Scholar, 28Kobayashi K. Inohara N. Hernandez L.D. Galan J.E. Nunez G. Janeway C.A. Medzhitov R. Flavell R.A. Nature. 2002; 416: 194-199Crossref PubMed Scopus (745) Google Scholar), and suggesting a model in which Rip2 could be a junction. Nevertheless, analyses using knockout mice and further siRNA experiments should be initiated to clarify differences and homologies of Nod and TLR signaling pathways. The Nod-dependent NF-κB activation by intact or inactive S. pneumoniae is most likely due to cell wall peptidoglycan. Nod2 has been found to mediate cell activation by a muramyl-dipeptide conserved in basically all kinds of peptidoglycans (14Girardin S.E. Boneca I.G. Viala J. Chamaillard M. Labigne A. Thomas G. Philpott D.J. Sansonetti P.J. J. Biol. Chem. 2003; 278: 8869-8872Abstract Full Text Full Text PDF PubMed Scopus (1916) Google Scholar, 15Inohara N. Ogura Y. Fontalba A. Gutierrez O. Pons F. Crespo J. Fukase K. Inamura S. Kusumoto S. Hashimoto M. Foster S.J. Moran A.P. Fernandez-Luna J.L. Nunez G. J. Biol. Chem. 2003; 278: 5509-5512Abstract Full Text Full Text PDF PubMed Scopus (1406) Google Scholar), whereas meso-diaminopimelate acid is responsible for Nod1 utilization (12Chamaillard M. Hashimoto M. Horie Y. Masumoto J. Su Q. Saab L. Ogura Y. Kawasaki A. Fukase K. Kusumoto S. Valvano M.A. Foster S.J. Mak T.W. Nunez G. Inohara N. Nat. Immunol. 2003; 4: 702-707Crossref PubMed Scopus (1013) Google Scholar, 13Girardin S.E. Boneca I.G. Carneiro L.A. Antignac A. Jehanno M. Viala J. Tedin K. Taha M.K. Labigne A. Zahringer U. Coyle A.J. DiStefano P.S. Bertin J. Sansonetti P.J. Philpott D.J. Science. 2003; 300: 1584-1587Crossref PubMed Scopus (1257) Google Scholar). Our finding that pneumococci showed a clear Nod2 but no Nod1 activity is thus in line with the fact that lysine, and not meso-diaminopimelate acid, has been found as the third amino acid in peptidoglycans of various pneumococcal strains (29Severin A. Tomasz A. J. Bacteriol. 1996; 178: 168-174Crossref PubMed Google Scholar). The high constitutive activity of overexpressed Nod1 and Nod2 might be responsible for the limited effect of viable pneumococci on Nod-dependent NF-κB activation (in contrast to the more distinct effects of intracellularly transfected pneumococci). A recently published paper describes macrophages from Nod2 knockout mice as responding normally to different pathogen-associated molecules, leading these authors to challenge the role of Nod2 in host defenses (30Pauleau A.L. Murray P.J. Mol. Cell. Biol. 2003; 23: 7531-7539Crossref PubMed Scopus (230) Google Scholar). In this study, cells were stimulated with known TLR ligands from the extracellular side, and no viable pathogens were used. It remains to be determined how these Nod2-deficient cells would respond to living invasive bacteria. Thus, the Nod2-deficient mice as well as the recently generated Nod1 knockout mice will be of invaluable help to further elucidate the precise role of these proteins in host defense (12Chamaillard M. Hashimoto M. Horie Y. Masumoto J. Su Q. Saab L. Ogura Y. Kawasaki A. Fukase K. Kusumoto S. Valvano M.A. Foster S.J. Mak T.W. Nunez G. Inohara N. Nat. Immunol. 2003; 4: 702-707Crossref PubMed Scopus (1013) Google Scholar, 13Girardin S.E. Boneca I.G. Carneiro L.A. Antignac A. Jehanno M. Viala J. Tedin K. Taha M.K. Labigne A. Zahringer U. Coyle A.J. DiStefano P.S. Bertin J. Sansonetti P.J. Philpott D.J. Science. 2003; 300: 1584-1587Crossref PubMed Scopus (1257) Google Scholar, 30Pauleau A.L. Murray P.J. Mol. Cell. Biol. 2003; 23: 7531-7539Crossref PubMed Scopus (230) Google Scholar). Overall, besides S. pneumoniae recognition by TLRs, Nod2 activation seems to play an important role in host cell activation by internalized pneumococci. Downstream of Nod2 and Rip2, signal-transducing molecules like IRAK, IRAK2, TRAF6, NIK, TAB2, and TAK1 might mediate NF-κB-dependent cell activation. Knowledge about the molecular interaction of pneumococci with target cells may pave the way to innovative therapeutic strategies. We are grateful to G. Nuñez, P. Dempsey, K. Fitzgerald, C. Kirschning, Tularik Inc., T. Ishitani, and T. Shimizu for providing expression vectors. We further thank K. Mohr, V. Johnston, and S. Moderer for excellent technical assistance.