XIAP mediates NOD signaling via interaction with RIP2 Andreas Kriega,b, Ricardo G. Correaa, Jason B. Garrisona, Gae¨ lle Le Negratea,c, Kate Welsha, Ziwei Huanga, Wolfram T. Knoefelb, and John C. Reeda,1 aBurnham Institute for Medical Research, La Jolla, CA 92037; bDepartment of General, Visceral and Pediatric Surgery, University Hospital Duesseldorf, D-40225 Duesseldorf, Germany; and cInstitute of Medical Microbiology and Hospital Hygiene, University Hospital Duessledorf, D40225 Duesseldorf, Germany Communicated by Erkki Ruoslahti, Burnham Institute for Medical Research at University of California, Santa Barbara, CA, July 6, 2009 (received for review June 11, 2009) NOD1 and NOD2 are members of the NOD-like receptor (NLR) protein signal transduction processes, which include activation of MAPKs family that are involved in sensing the presence of pathogens and are (23) and NF-␬B through interactions of TAB1/TAK1 with its BIR1 a component of the innate immune system. Upon activation by domain (25). specific bacterial peptides derived from peptidoglycans, NODs inter- Hints that IAP-family members might be involved in innate act via a CARD-CARD interaction with the receptor-interacting protein immunity have come from studies demonstrating that flies depleted kinase RIP2, an inducer of NF-␬B activation. In this report, we show by shRNA of Drosophila IAP-2 (DIAP2) fail to activate NF-␬Bin that NOD signaling is dependent on XIAP, a member of the inhibitor response to bacterial challenge with Escherichia coli and show of apoptosis protein (IAP) family. Cells deficient in XIAP exhibit a decreased survival rates when exposed to Enterobacter cloacae (26, marked reduction in NF-␬B activation induced by microbial NOD 27). Recently, studies of xiapϪ/Ϫ mice have provided evidence of ligands and by over-expression of NOD1 or NOD2. Moreover, we a contribution of XIAP to NF-␬B and JNK activation induced by show that XIAP interacts with RIP2 via its BIR2 domain, which could TLRs and NLRs during infection with Listeria monocytogenes, be disrupted by XIAP antagonists SMAC and SMAC-mimicking com- further supporting the hypothesis that IAPs may participate in pounds. Both NOD1 and NOD2 associated with XIAP in a RIP2- innate immune responses (28). Here we show that XIAP is required dependent manner, providing evidence that XIAP associates with the at least in certain types of epithelial cells for NF-␬B activation NOD signalosome. Taken together, our data suggest a role for XIAP induced by NOD1 and NOD2, and demonstrate that XIAP binds MEDICAL SCIENCES in regulating innate immune responses by interacting with NOD1 and RIP2 thereby associating with NOD1/NOD2 signaling complexes. NOD2 through interaction with RIP2. Results he NOD-like receptors (NLRs) constitute a family of innate XIAP Is Required for NOD Signaling. Epithelial cells of the intestinal immunity proteins involved in sensing the presence of intracel- track are a first line of defense against many microorganisms. We T took advantage of human tumor cell lines derived from colonic lular pathogens and stimulating host defense responses (2). Two of epithelium in which the XIAP gene had been ablated by homolo- these family members, NOD1 and NOD2, share structural and gous recombination to ask whether XIAP is required for cellular functional characteristics. Both, NOD1 and NOD2, contain C- responses to synthetic NOD1 or NOD2 ligands. Accordingly, terminal leucine-rich repeats (LRRs) thought to act as receptors for isogenic pairs of XIAPϩ/ϩ and XIAPϪ/Ϫ HCT116 and DLD-1 pathogen-derived molecules, a central nucleotide-binding oli- cells were stimulated for 24 h with NOD1 and NOD2 ligands, gomerization domain (NACHT) (3, 4), and N-terminal caspase L-Ala-␥-D-Glu-mDAP (DAP) and muramyl dipeptide (MDP), recruitment domains (CARDs) that associate with down-stream respectively, then Interleukin-8 (IL-8) production was measured signaling proteins (5, 6). NODs activation is stimulated by bacterial (Fig. 1A and B). Both DAP and MDP induced increases in IL-8 peptides derived from peptidoglycans, with diaminopimelic acid production in the wild-type HCT116 and DLD-1 cells, with MDP (DAP) stimulating NOD1 (7, 8), and muramyl dipeptide (MDP) more potent than DAP. In contrast, neither of these NOD ligands activating NOD2 (9, 10). Upon recognition of these ligands, oli- induced IL-8 production in cultures of XIAP-deficient HCT116 gomerization of the NACHT domains initiates the recruitment of and DLD1. Whereas XIAPϪ/Ϫ cells failed to respond to NOD interacting proteins, binding the serine/threonine protein kinase ligands, they remained responsive to TNF, which induced robust RIP2/CARDIAK/RICK via CARD-CARD-interactions (5, 6). ␬ IL-8 production. RIP2 is critical for NF- B activation induced by NOD1 and NOD2 The observation that XIAP gene knock-out impairs NOD- (11), although the molecular details of how the NOD/RIP2 com- signaling was further confirmed by quantitative RT-PCR analysis of ␬ plex stimulates NF- B activation are only partly understood. the NF-␬B target genes I␬B␣ and IL-8, detecting decreased levels RIP2 not only binds to NOD1 and NOD2 via CARD-CARD of I␬B␣ and IL-8 mRNAs in XIAP-deficient HCT116 cells com- interactions, but it also reportedly associates with other signaling pared with wild-type HCT116 cells following stimulation with MDP proteins independently of the CARD, including members of the or DAP (Fig. 1C). In contrast, TNF-␣ induced expression of these TNFR-associated factor (TRAF) family and members of the in- NF-␬B target genes comparably in XIAPϩ/ϩ and XIAPϪ/Ϫ cells. hibitor of apoptosis protein (IAP) family, cIAP-1 and cIAP-2 (12, Similar observations were made using a NF-␬B reporter gene to 13). IAP-family proteins play prominent roles in regulating pro- monitor responses to NOD ligands. In XIAPϩ/ϩ HCT116 cells, grammed cell death by virtue of their ability to bind caspases (14–17), intracellular cysteine proteases responsible for apoptosis. A common structural feature of the IAPs is the presence of one or Author contributions: A.K. and J.C.R. designed research; A.K., R.G.C., and J.B.G. performed more baculoviral IAP-repeat (BIR) domains, which serve as scaf- research; A.K., R.G.C., G.L.N., K.W., Z.H., W.T.K., and J.C.R. contributed new reagents/ folds for protein interactions (18). One of the most extensively analytic tools; A.K., R.G.C., J.B.G., and J.C.R. analyzed data; and A.K. and J.C.R. wrote the paper. investigated members of the IAP-family is X-linked IAP (XIAP). The authors declare no conflict of interest. XIAP contains three BIR domains (19), followed by a ubiquitin 1To whom correspondence should be addressed at: Burnham Institute for Medical Re- binding domain (UBA) (20) and a C-terminal RING that functions search, 10901 North Torrey Pines Road, La Jolla, CA 92037. E-mail: reedoffice@ as E3-Ligase promoting ubiquitination and subsequent proteaso- burnham.org. mal degradation of distinct target proteins (21). In additional to its This article contains supporting information online at www.pnas.org/cgi/content/full/ anti-apoptotic role as a caspase inhibitor, XIAP functions in certain 0907131106/DCSupplemental. www.pnas.org͞cgi͞doi͞10.1073͞pnas.0907131106 PNAS Early Edition ͉ 1of6 Downloaded by guest on September 29, 2021 Fig. 1. XIAP is required for induction of cytokine pro- duction by NOD ligands. (A and B) HCT 116 XIAP ϩ/ϩ (WT ϭ Wild-Type) and XIAP Ϫ/Ϫ cells (KO ϭ knock-out) (A) or DLD-1 XIAPϩ/ϩ or XIAP Ϫ/Ϫ cells (B) were stimu- lated with MDP (20 ␮g/mL), DAP (20 ␮g/mL), TNF-␣ (5 ng/mL), or left untreated for 24 h. Cell free supernatants were collected after centrifugation and analyzed for IL-8 secretion by ELISA. Data represent means Ϯ SD of three independent experiments (pg/mL). (C) Reduced expres- sion of NOD ligand-inducible genes in XIAP-deficient cells. HCT116 XIAPϪ/Ϫ (white bars) and XIAPϩ/ϩ (black bars) were stimulated for 1 h with various NF-␬B inducers: 20 ␮g/mL ␥Tri-DAP, 20 ␮g/mL MDP-LD, or 10 ng/mL TNF-␣. RNA was isolated and relative levels of I␬B␣ and IL-8 mRNAs were measured by Q-RT-PCR, normalized relative to 18S rRNA, expressed as relative levels compared with unstimulated cells (mean value ϭ 1), and presented as mean ϩ std dev of triplicate determinations performed in at least two independent experiments. (D) HCT116 XIAPϪ/Ϫ cells (KO) were transfected with FLAG-XIAP- encoding plasmid or empty FLAG-plasmid, then stimu- lated 24 h posttransfection with MDP (20 ␮g/mL), ␥Tri- DAP (20 ␮g/mL), TNF-␣ (5 ng/mL), or left untreated. As a control, HCT116 XIAPϩ/ϩ (WT) were similarly stimulated. NF-␬B reporter gene activity was measured after 24 h using the Dual Luciferase assay method. Normalized val- ues represented mean Ϯ SD (n ϭ 3). (Inset) Lysates from the cells were prepared, normalized for total protein content, and analyzed by immunoblotting using anti- XIAP antibody. Reprobing blot with anti-beta-Actin an- tibody confirmed equal loading. (E) XIAP deficiency se- lective impacts NOD-mediated NF-␬B activation. HEK293T cells containing a stably integrated NF-␬B- luciferase reporter gene were infected with XIAP shRNA (KD ϭ knock-down) (white bars) or scrambled control (CNTL) (black bar) lentiviruses. After 24 h, cells were stim- ulated with 10 ␮g/mL MDP-LD (MDP), 5 ␮g/mL ␥Tri-DAP (DAP), 0.2 ␮g/mL doxorubicin (DOX), 10 ng/mL PMA/ ionomycin (PMA), or 2 ng/mL TNF-␣. NF-␬B activity was measured 24 h later by luciferase activity, and data were expressed as fold-induction relative to control unstimu- lated values for each cell line (mean value ϭ 1) and represent mean Ϯ std dev of triplicates performed in at least two independent experiments.