Centaurin β1 Down-regulates Nucleotide-binding Oligomerization Domains 1- and 2-dependent NF-κB Activation*

Centaurin β1 (CENTB1), a GTPase-activating protein, is a member of the ADP-ribosylation factor family encoded by a gene located on the short arm of human chromosome 17. A yeast two-hybrid screen first suggested a direct interaction between CENTB1 and NOD2. Co-immunoprecipitation experiments confirmed direct interaction between CENTB1 and NOD2 and demonstrated similar interaction between CENTB1 and NOD1. We also demonstrate that endogenous CENTB1 interacts with endogenous NOD2 and NOD1 in SW480 and HT-29 intestinal epithelial cells. CENTB1 partially co-localized with NOD2 and NOD1 proteins in the cytoplasm of mammalian cells. CENTB1 expression in epithelial cells was highly induced by tumor necrosis factor α, interleukin 1β, and the NOD1 and NOD2 ligands (γ-d-glutamyl-meso-diaminopimelic acid and muramyl dipeptide, respectively). In addition, CENTB1 mRNA level is increased in the inflamed mucosa of patients with inflammatory bowel disease. Functionally, CENTB1 overexpression inhibited NOD1- and NOD2-dependent activation of NF-κB, whereas small inhibitory RNA against CENTB1 increased NF-κB activation following NOD1- or NOD2-mediated recognition of the bacterial components γ-d-glutamyl-meso-diaminopimelic acid and muramyl dipeptide, respectively. In contrast, CENTB1 had no effect on NF-κB activation induced by Toll-like receptors. In conclusion, CENTB1 selectively down-regulates NF-κB activation via NODs pathways, creating a “feedback” loop and suggesting a novel role of CENTB1 in innate immune responses to bacteria and inflammatory responses.

Nucleotide-binding oligomerization domain (NOD) 3 containing proteins, including NOD1 and NOD2, have been impli-cated in intracellular recognition of bacterial components (1). The CARD15 (NOD2) gene is located on the long arm of human chromosome 16 and encodes a protein that contains two C-terminal caspase recruitment domains (CARDs), a central NOD (also known as a nucleotide-binding domain) and an N-terminal leucine-rich repeat region (2). NOD2 recognizes intracellular peptidoglycans from Gram-negative and Grampositive bacteria through the detection of the minimal motif muramyl dipeptide (MDP) (3). Recognition of this specific dipeptide by NOD2 leads to nuclear factor B (NF-B) activation (4). This is thought to be relevant to disease pathogenesis given that NF-B is a key intracellular signaling molecule in a variety of inflammatory pathways (5) and the finding of NOD2 mutations in patients with inflammatory diseases such as Crohn disease (CD) (6,7) and Blau syndrome (BS) (8).
The NOD1 (CARD4) gene is located on the short arm of human chromosome 7. NOD1 is a cytosolic protein that is constitutively expressed by many cell types including intestinal epithelial cells (9) and has also been implicated in innate immune recognition. NOD1 contains CARD, NOD/nucleotide-binding domain, and leucine-rich repeat region domains. The CARD and NOD domains are required for NF-B activation (10), whereas the leucine-rich repeat region domain recognizes the ␥-D-glutamyl-meso-diaminopimelic acid (iE-DAP) derived from peptidoglycans of Gram-negative bacteria (11).
Because intestinal epithelial cells (IECs) are constantly exposed to bacteria and their components, IECs are considered to be not only a structural but also a functional barrier as the front line of host defense against microorganisms. Although IECs are hyporesponsive to normal flora in vivo, pathogenic bacteria or internalized bacterial components can initiate an inflammatory response (12)(13)(14). These data suggest that presentation of internalized bacterial components is a critical point in bacteria-epithelial interaction. We have previously shown that the NOD2 acts as a defensive factor against intracellular bacteria in IECs (20). In addition, membrane recruitment of NOD2 in IECs is essential for NF-B activation in MDP recognition, which could be responsible for NOD2 anti-bacterial activity (21).
Interestingly, intraepithelial bacteria have been found in patients with inflammatory bowel disease (IBD) (15) and an increased prevalence of adherent-invasive Escherichia coli has been observed in ileal mucosa of CD patients (16). This is consistent with the hypothesis that recognition mechanisms of intracellular bacteria or their components likely play a key role in the pathophysiology of IBD. However, understanding of the mechanisms through which NOD proteins lead to effective responses remains substantially incomplete. Although some studies have suggested that NOD2 activation modulates TLR2 responses, which may be dysregulated when the NOD2 3020insC mutant associated with CD is present, others have not found comparable results (17)(18)(19).
A number of groups have endeavored to identify and characterize proteins that interact with NODs to better understand their functional effects and mechanisms of action. These efforts have previously suggested roles for GRIM-19 (22), TAK1 (23), Ipaf (24), and Erbin (25,26) in the regulation of NOD2 activation of NF-B. However, it is evident that these proteins are not sufficient to fully explain the regulation of NOD2-mediated NF-B activation. To identify new molecules that could interact with these intracellular sensors, a yeast two-hybrid screen was undertaken using NOD2 protein as bait. Hybrid constructs comprising the NOD and leucine-rich repeat region domain led to the finding that Centaurin ␤1 protein (CENTB1) interacts with both the NOD1 and NOD2 proteins.

EXPERIMENTAL PROCEDURES
Yeast Two-hybrid Cloning of NOD2-interacting Proteins-Yeast two-hybrid screening was performed using MATCH-MAKER GAL4 two-hybrid system 3 according to the manufacturer's protocol (BD Biosciences Clontech, Palo Alto, CA). Briefly, pGBKT7-NOD2 was transfected into the AH109 yeast strain. Expression of Myc-tagged NOD2 protein in yeast extract was confirmed by Western blot analysis using anti-Myc monoclonal antibody (Covance, Richmond, CA) and affinity-purified anti-NOD2 antisera (19). Screening was performed using a bone marrow pretransformed library (BD Biosciences Clontech) as described previously (20). Co-transformants were selected in S.D. medium lacking histidine, leucine, and tryptophan. Yeast ␣-galactosidase activity, expressed from the MEL1 gene response to GAL4 activation, was determined in plates containing X-gal (BD Biosciences Clontech).
Colonic mucosal biopsy samples from patients with ulcerative colitis and CD were immediately snap frozen in liquid nitrogen. Frozen biopsies were disrupted mechanically, and total RNA was isolated using TRIzol reagent (Invitrogen) following the manufacturer's protocols. cDNA from colon surgical specimens of IBD patients and normal controls were provided by Dr. Emiko Mizoguchi. All of the tissue samples were obtained under protocols approved by Partners IRB.
Construction of Expression Plasmids-cDNA of CENTB1 was obtained by a PCR method using an Expand High Fidelity PCR system (Roche Applied Science) from the above mentioned cell lines. The PCR primers were: 5Ј-atgacggtcaagctggatttcgagg-3Ј (forward) and 5Ј-cacagcgtgtggaggtcatgactg-3Ј (reverse). The sequence of CENTB1 cDNA was confirmed using an ABI 3700 PRISM (PerkinElmer Life Sciences) automated sequencer. An expression plasmid-encoding Xpress-tagged CENTB1 mammalian expression vector (pCDNA4/HisMAX-Centaurin ␤1) was generated by RT-PCR from SW480 cells. An expression plasmid-encoding FLAG-tagged NOD2 (pCMV FLAG-NOD2) was previously generated (20). pCI CARD4/NOD1-HA expression vector was provided by Dr. John Bertin (Millennium Pharmaceuticals Inc). Two oligonucleotides, 19 residues in length (siRNA-1, gtacaaggaccctgtgact; and siRNA-2, aattcaccgtgagcctgaa), specific to the human CENTB1 were selected for synthesis of siRNA. pSUPER vector for siRNA was purchased from Oligoengine (Seattle, WA). The depletion of endogenous CENTB1 expression was confirmed by RT-PCR.
The cells were grown in six-well plates. The medium was removed, and 300 l of 1% Triton-X lysis buffer (1% Triton X-100, 0.1 M NaCl, 10 mM Hepes, pH 5.6, 2 mM EDTA, 4 mM Na 3 VO 4 , and 40 mM NaF) supplemented with protease-inhibitor mixture, complete Mini (Roche Applied Science) was added. Supernatants of cell lysates were obtained by centrifu-gation at 12,000 ϫ g for 10 min. Protein concentration was determined using DC protein assay kit (Bio-Rad). Proteins were extracted from COS-7 cells (transiently co-transfected with pCMV FLAG-NOD2 and pCDNA4-HisMAX-CENTB1 plasmids). Two milligrams of total cell lysate protein were immunoprecipitated with 2 g of anti-FLAG biotinylated M2 monoclonal or anti-Xpress monoclonal antibodies and 100 l of Hiptrap protein A/G-Sepharose beads. After overnight incubation at 4°C, the immunoprecipitated proteins were separated on 4 -12% Tris-glycine gels (Invitrogen). The proteins were blotted onto polyvinylidene difluoride membranes and detected using anti-FLAG biotinylated M2 monoclonal antibody (Sigma-Aldrich) and anti-Xpress monoclonal antibody (Invitrogen).
Co-localization of CENTB1 and NOD Proteins-COS-7 cells were seeded on sterile Permanox coverslips, grown for 24 h, and then co-transfected with pCDNA4/HisMAX-CENTB1 and pCMV FLAG-NOD2. After 48 h, the cells were washed twice with ice-cold phosphate-buffered saline, fixed 20 min with cold methanol at Ϫ20°C, and washed three times with ice-cold phosphate-buffered saline. The cells were saturated for 30 min with phosphate-buffered saline containing 5% horse serum and then incubated 2 h with primary antibody (mouse monoclonal anti-Xpress antibody and/or rabbit polyclonal anti-FLAG antibody from Sigma). Immunostaining was performed with Texas Red-conjugated anti-mouse IgG or with FITC-conjugated antirabbit IgG secondary antibodies (Vector Laboratories). The coverslips were mounted in Vectashield (Vector Laboratories) and examined at room temperature with a confocal laser scanning microscope (Radiance 2000 model; Bio-Rad) using multitracking (line switching) for two-color imaging (40ϫ). Image acquisition was performed with LaserSharpScanning software (Bio-Rad).
mRNA Expression of CENTB1-Total RNA of intestinal epithelial cell lines was extracted by TRIzol (Invitrogen) following the manufacturer's instructions. For RT-PCR, 1 g of total RNA was reverse transcribed using a SuperScript first strand synthesis system (Invitrogen). Real time RT-PCR was performed in an ABI Prism 7000 sequence detector using a SYBR Green JumpStart TM detection system. Briefly, 50 ng of the reversed transcribed cDNA was used for each PCR with 200 nM of forward and reverse primers. Two sets of primers for CENTB1 were designed (5Ј-cgccttgatgacagcccccgggg-3Ј (forward) and 5Ј-cccatgagtcaagggtcagagaccg-3Ј (reverse); and 5Ј-gtgtcagaattggagacccgtc-3Ј (forward) and 5Ј-gctgtccagcttgtggttcagg-3Ј (reverse)). The PCR program with specific primers was as follows: 50°C for 2 min and then 95°C for 10 min followed by 44 cycles of 95°C for 15 s, 60°C for 15 s, and 72°C for 15 s. PCR products were sequenced using an ABI 3700 PRISM (PerkinElmer Life Sciences) automated sequencer. The sequences were analyzed using NCBI BLAST software.
Assessment of NF-B Activation.-HEK293 cells were transfected overnight with NOD2, NOD1, CENTB1, and CENTB1 siRNA constructs, plus pIV luciferase reporter and Renilla plasmids. At the same time, 1 g of MDP-LD (Sigma) (where L and D indicate levo and dextro, respectively) and 100 ng/ml of iE-DAP were added. After 24 h, luciferase activity was measured using the dual luciferase reporter assay system (Promega, Mad-ison, WI) according to the manufacturer's instructions and normalized relative to Renilla activity.
Statistical Analysis-For analysis of the significance of differences in NF-B, invasion, and mRNA levels, Student's t test was used for comparison of two groups of data. All of the experiments were repeated at least five times. A p value Յ0.05 was considered to be statistically significant.

Identification of CENTB1 as a Novel NOD2-interacting
Protein-A yeast two-hybrid screen was performed to identify cellular proteins that interact with CARD15/NOD2. A NOD2 construct without the CARD15 domain was used as bait. We screened a human cDNA library expressing proteins fused to the active domain transcriptional activation domain. One of the positive clones was found to encode the human CENTB1. To confirm a specific interaction between NOD2 and CENTB1, we performed AH109 yeast survival assays in S.D./ϪAde/ϪHis/ ϪLeu/ϪTrp/X-gal selective medium. Co-expression of NOD2 and CENTB1 in AH109 yeast showed a strong interaction between these two proteins (data not shown).
To confirm the interaction of NOD2 and CENTB1 in mammalian cells, COS7 and HEK293 cells were transiently transfected with FLAG-tagged NOD2 and Xpress-tagged CENTB1. After 48 h of transfection, NOD2 was immunoprecipitated with HM2559 anti-NOD2 serum (20), and CENTB1 was immunoprecipitated with an anti-Xpress monoclonal antibody. As shown in Fig. 1A, CENTB1 was detected in anti-HM2559 (20) immunoprecipitates from NOD2 co-transfectants, but not from cells co-transfected with a control plasmid. A reciprocal immunoprecipitation/blotting experiment using anti-Xpress monoclonal antibody also showed NOD2 co-precipitation with CENTB1, confirming the interaction of these two proteins in mammalian cells.
Interaction between NOD1 and CENTB1 in Mammalian Cells-To examine the extent of the interaction of CENTB1 with NOD proteins, the ability of NOD1, an other NOD family member, to co-immunoprecipitate with CENTB1 was investigated by Western blot (Fig. 1B). After 48 h of co-transfection of HEK293 cells with DNA constructs encoding HA-tagged NOD1 and Xpress-tagged CENTB1, immunoprecipitation experiments were performed using anti-NOD1 antiserum HM3851 (9) and anti-Xpress monoclonal antibody to pull down NOD1 and CENTB1, respectively. CENTB1 was detected in anti-NOD1 immunoprecipitates from NOD1 co-transfectants. A reciprocal immunoprecipitation/Western blotting experiment using anti-Xpress monoclonal antibody also showed NOD1 co-precipitation with CENTB1, confirming the interaction of these two proteins.
Real time PCR showed that CENTB1 mRNA was present in surgical specimens from ulcerative colitis and CD patients, but not from normal controls (Fig. 2B).
Induction of CENTB1 Expression in Intestinal Epithelial Cells by Inflammatory Stimuli-The finding that endogenous CENTB1 expression is increased in association with intestinal inflammation prompted an assessment of the effects of proinflammatory cytokines on CENTB1 expression. Two representative colonic cell lines (SW480 and HT-29) were stimulated with TNF-␣ (50 ng/ml) or IL-1␤ (15 ng/ml). CENTB1 mRNA was markedly increased after 1h of TNF-␣ stimulation (Fig.  3A). Comparable effects were observed following IL-1␤ stimulation. Similar results were also found in Jurkat and THP-1 cells when they were stimulated with these two proinflammatory cytokines (data not shown).
CENTB1 mRNA was also markedly induced in SW480 and HT-29 cells stimulated with the ligands for NOD1 and NOD2 MDP and iE-DAP, respectively (Fig. 3, B and C). Comparable increases in endogenous protein expression were also observed (Fig. 3, D and E). Indeed, CENTB1 expression was greater in epithelial cell lines stimulated with NODs ligands than when cells were stimulated with proinflammatory cytokines. In contrast, TLR2 and TLR4 prototype ligands lipoteichoic acid and lipopolysaccharide (LPS) had no effect on CENTB1 expression (data not shown).
Interaction between Endogenous NODs and CENTB1 Proteins in Mammalian Cells-To assess the possible physiological significance of the NOD2/CENTB1 interaction, we investigated the interaction between endogenous NOD2 and CENTB1 using lysates from SW480 cells that express endogenous NOD2 (21) and stimulated them with MDP-LD (1 g) to induce endogenous CENTB1 expression. For immunoprecipitation experiments, we used a newly produced rabbit antiserum against CENTB1. Anti-CENTB1 antibody specifically recognized CENTB1 in lysate of COS-7 cells transfected with the Xpress-tagged CENTB1 construct (data not shown). Endogenous NOD2 expression was observed in SW480 cells by using the rabbit antiserum against NOD2 HM2559 (20). Immunoprecipitation using anti-CENTB1 antiserum in SW480 cells stimulated with MDP-LD (1 g) (Fig. 4A) confirmed the interaction between endogenous CENTB1 with NOD2 protein.
Similarly, interaction between endogenous NOD1 and CENTB1 proteins was assessed using HT-29 cells stimulated with the NOD1 ligand (iE-DAP) to induce CENTB1 expression. The HT-29 intestinal epithelial cell line has been previously  demonstrated to exhibit endogenous expression of NOD1 (9). Immunoprecipitation and Western blot experiments using rabbit antiserum previously generated against NOD1 (HM3851) (9) and the newly produced rabbit antiserum against CENTB1 described above confirmed interaction between endogenous NOD1 and CENTB1 proteins in HT-29 lysate (Fig. 4B).

CENTB1 and NODs Proteins Co-localize in the Cytosystem-
To determine whether NOD2 and CENTB1 co-localize as implied in the finding of a direct biochemical interaction, subcellular localization was evaluated by immunofluorescence confocal microscopy. COS-7 cells were transfected with Xpress-tagged CENTB1 and FLAG-tagged NOD2 expression plasmids. NOD2 protein was observed throughout the cytoplasm and also associated with the plasma membrane as previously reported (21). In COS-7 cells co-expressing FLAG-tagged NOD2 and Xpress-tagged CENTB1, CENTB1 co-localized with the cytoplasmic NOD2 but not with the NOD2 membrane-associated fraction (Fig. 5A). This partial co-localization of NOD2 with CENTB1 is consistent with the evidence that specific interactions occur between these two proteins in vitro. Using similar approaches CENTB1 was also found to co-localize with NOD1 in the intracytoplasmic compartment (Fig. 5B).
Cellular localization of endogenous CENTB1 and NOD2 was investigated in SW480 cells stimulated with MDP-LD. Co-localization between endogenous CENTB1 and NOD2 was found in the cytoplasm (Fig. 5C). Similar co-localization of endogenous CENTB1 and NOD1 was found in SW480 cells transfected with HA-NOD1 and stimulated with iE-DAP (Fig. 5D). These observations are consistent with interactions of endogenous NODs/CENTB1 proteins, suggesting physiological significance of these interactions.
Mapping of CENTB1 Domain Required for NODs Interaction-To characterize the domain of CENTB1 responsible for interaction with the NODs, deletion constructs of CENTB1 protein were prepared that allowed discrimination between encompassed three regions: the PH domain, the Arf-GAP domain, and the Ankyrin repeats (Fig. 6A). Co-immunoprecipitation experiments were performed in HEK293 cells co-transfected with wild type NOD1 or NOD2 and the variant deletion constructs of CENTB1. Using this approach, the PH domain of CENTB1 was found to be able to co-immunoprecipitate fulllength of NOD1 and NOD2 (Fig. 6, B and C), indicating that NODs/CENTB1 interaction occurs via the CENTB1 PH domain.  CENTB1 Down-regulates NOD-mediated NF-B Activation-To assess the functional effect of NODs/CENTB1 interactions, constructs encoding CENTB1, NOD2, and different amounts of MDP-LD were transfected into HEK293 cells. Subsequent activation of NF-B was determined using the NF-Bdriven luciferase reporter assay. CENTB1 was found to downregulate NF-B activation in a concentration-dependent manner (Fig. 7A). To confirm the role of CENTB1 in the down-regulation of NOD2, a siRNA approach was used to selectively inhibit the expression of CENTB1. HEK293 cells were transfected with NOD2 and specific siRNA for CENTB1 and then stimulated with MDP-LD. Transfection with CENTB1 siRNA significantly increased the NOD2 ligand-driven response compared with cells transfected with the same amount of CENTB1 and NOD2 constructs, suggesting that this effect is due to siRNA suppression effects of CENTB1. The lack of effect of a siRNA control (pSUPER vector) confirmed the specificity of these effects (Fig. 7B).
Similar studies were undertaken to evaluate the effect of CENTB1 on NOD1-mediated NF-B activation. HEK-293 cells were co-transfected with different concentrations of CENTB1 and NOD1 DNA constructs and stimulated with iE-DAP. CENTB1 down-regulated NOD1-mediated NF-B activation in a concentration-dependent manner resembling the effects of FIGURE 5. Co-localization of NOD proteins and CENTB1. A, cellular localization of NOD2 and CENTB1 was assessed in COS-7 cells co-transfected with FLAG-tagged NOD2 and Xpress-tagged CENTB1. CENTB1 was detected using monoclonal anti-Xpress antibody, followed by Texas Red-conjugated antimouse IgG and NOD2 protein, was detected using polyclonal rabbit anti-FLAG antibody followed by staining with FITC-conjugated anti-rabbit IgG. B, cellular localization of NOD1 and CENTB1 in COS-7 cells co-transfected with HA-tagged NOD1 and Xpress-tagged CENTB1. CENTB1 was detected using monoclonal anti-Xpress and Texas Red-conjugated anti-mouse IgG for CENTB1, and NOD1 was detected using polyclonal rabbit anti-HA antibody and FITC-conjugated anti-rabbit IgG. C, co-localization of endogenous CENTB1 and NOD2. SW480 cells were stimulated with 1 g of MDP-LD to induce CENTB1 expression. CENTB1 was detected using anti-CENTB1 antibody as primary and Texas Red-conjugated anti-rabbit IgG. NOD2 was detected using anti-NOD2 monoclonal antibody and stained with FITC-conjugated anti-mouse IgG. D, cellular localization of endogenous CENTB1 and transfected NOD1. SW480 cells were transfected with HA-tagged NOD1 construct and stimulated with iE-DAP (100 ng) to induce CENTB1 expression. Endogenous CENTB1 was detected as described previously. NOD1 was detected using anti-HA monoclonal antibody and stained with FITC-conjugated anti-mouse IgG.

FIGURE 6. Identification of domain of CENTB1 interacting with NODs.
A, schematic representation and amino acid composition of the CENTB1 deletion mutants used in this study. B, PH domain of CENTB1 is necessary for interaction with NOD2. HEK293 cells were transfected with different deletion constructs of CENTB1 and full-length NOD2. The cell lysates were immunoprecipitated (IP) and blotted with either anti-Xpress or anti-NOD2 antibodies. C, PH domain of CENTB1 interacts with NOD1. The cell lysates were immunoprecipitated and blotted with anti-Xpress and anti-NOD1 antibodies from HEK293 cells transfected with different deletion constructs of CENTB1 and the full length of NOD1. CENTB1 on NOD2 mediated activation (Fig. 8A). Transfection of siRNA against CENTB1 significantly enhanced NF-B activation in response to the NOD1 ligand (iE-DAP) in HEK293 expressing NOD1 and CENTB1 compared with those measured in HEK293 cells transfected with pSUPER control vector (Fig. 8B).
To determine whether the action of CENTB1 was specific to the NOD family or also encompassed other innate immune pathways, the effect of CENTB1 on the Toll-like receptor (TLR2 and TLR4) pathways was assessed. Activation of NF-B was evaluated in HEK293 cells transiently transfected with different concentrations of CENTB1 and TLR2 DNA constructs and stimulated with lipoteichoic acid. TLR2 pathway-mediated NF-B activation was not affected by CENTB1 expression (Fig.  9A). A similar approach was undertaken to evaluate the effect of CENTB1 on TLR4-mediated NF-B activation. HEK293 cells were transiently transfected with different concentrations of DNA constructs encoding CENTB1, TLR4, and MD-2 and then stimulated with LPS at 100 ng/ml for 4 h. Expressed CENTB1 exerted no effect on NF-B activation via TLR4 (Fig. 9B). These findings were confirmed in colonic epithelial cells transfected with two different siRNA against CENTB1 and stimulation of endogenous TLR4 and TLR2 by the specific ligands LPS and lipoteichoic acid in SW480 and Caco-2 cells, respectively (data not shown). These findings confirm the specificity of CENTB1 in modulating NOD-mediated, but not TLRmediated, NF-B cell activation.
Analysis of the Interaction of CENTB1 with NOD2 Disease-associated Mutants-Specific mutations in the NOD2 gene have been associated with inflammatory diseases. To determine whether CD and Blau mutants exert an effect on the interaction and functional effect of CENTB1, we performed co-immunoprecipitation assays in HEK293 cells transfected with constructs encoding CENTB1 and disease-associated NOD2 mutants. CD and Blau mutations retained the ability to interact with CENTB1 (Fig. 10A). To assess the functional role of these mutations on CENTB1, NF-B assays were performed in HEK293 cells transfected with wild type NOD2 or CD and Blau syndrome mutants. Interestingly, CENTB1 down-regulates NF-B activation in Blau mutants, but the suppressive effect of NF-B activation was lost when interacting with CD mutants (Fig. 10, B and C).
Up-regulation of NF-B Activation by Endogenous NODs Pathways Using siRNA against CENTB1-To assess the functional effect of endogenous CENTB1 after stimulation with NODs ligands, we first confirmed that siRNA could specifically knock down endogenous CENTB1 induced by NOD ligands in TNF. Two siRNA against CENTB1 (siRNA-1 and siRNA-2) significantly knocked down the level of CENTB1 mRNA (around 75 and 60%, respectively) compared with cells transfected with control siRNA (p-SUPER) (Fig. 11A). Following confirmation of effective knockdown, SW480 cells were stimulated with MDP and iE-DAP and transfected with siRNA against CENTB1 or siRNA control. Luciferase and Renilla reporters were transfected, and NF-B assays were performed. SW480 cells transfected with siRNA-1 and -2 against CENTB1 showed significantly increased NF-B activity compared with those measured in untransfected or transfected cells with siRNA control. (Fig. 11, B and C). These results confirm that endogenous CENTB1 down-regulates NF-B activation after stimulation with NOD1 and NOD2 ligands in colonic epithelial cells.

DISCUSSION
In the present study, we describe a novel NOD2 interactor, CENTB1, found by yeast two-hybrid screening. Interaction between CENTB1 and NOD2 was confirmed in mammalian cells by co-immunoprecipitation. Of interest, CENTB1 was also found to interact with NOD1. NOD1 and NOD2 have been involved in the activation of NF-B pathway through intracellular recognition of the bacterial components iE-DAP and MDP-LD (4,11).
To determine whether these associations have a functional role, we assessed NF-B activation assays in mammalian cells co-transfected with NOD2 or NOD1 and CENTB1. These studies showed that CENTB1 down-regulates NF-B activation via NOD1 and NOD2 pathways in a concentration-dependent manner. This functional interaction was confirmed by the finding that ligand-dependent NF-B activation was enhanced when the expression of CENTB1 was inhibited using specific siRNA against CENTB1. In contrast, CENTB1 does not modify NF-B activation via TLR-2 or -4, indicating a selective differential effect on NF-B activation by these two families (NODs and TLRs) of innate immune pattern recognition receptors.
In addition to the demonstration of direct interactions between CENTB1 and the NODs proteins biochemically, characterization of the subcellular localization of NODs proteins and CENTB1 revealed intracellular co-localization, consistent with functional interaction between these proteins in situ. We previously reported that membrane localization of NOD2 is required for MDP-dependent activation of NF-B (21). How-ever, CENTB1 is located in the cytoplasm and interacts with NOD1 and NOD2 in this cellular compartment to modulate the NF-B signaling. Recent studies demonstrated a direct interaction between NOD2 and Erbin that occurs near the basolateral membrane in differentiated Caco-2 cells, and Nod2-dependent activation of NF-B is inhibited by Erbin overexpression (25,26). In light of this observation, we hypothesize that CENTB1 suppresses the activation of cytosolic NOD2, whereas Erbin may have an inhibitory effect on membrane-associated NOD2 fraction. These findings suggest that CENTB1 could play an important role in overall modulation of innate immune responses to intracellular bacteria by downregulating NF-B activation in mammalian cells. The functional effects of CENTB1 contrast with those observed for two other NOD2 interactors, GRIM19. GRIM19 is required for NF-B activation following NOD2-mediated recognition of bacterial MDP and controls pathogen invasion of intestinal epithelial cells (22). TAK1 up-regulates NOD2-mediated NF-B activation (23). Collectively, it appears that NOD-NF-B activation is highly controlled, both positively and negatively by several NOD-protein interactors.
Mutations in NOD2 result in altered activity of NOD2 and are associated with two major forms of inflammatory diseases in humans: CD and BS. We explored whether NOD2 mutants lead to altered interaction with CENTB1. CD and BS mutants still bind CENTB1 protein. However, CENTB1 has a differential effect on the regulation of NF-B activation. CENTB1 was still able to exert down-regulatory effects on BS-associated NOD2 mutants following ligand stimulation comparable with those observed with wild type NOD2. In contrast, CENTB1 had no effect on ligand-induced NF-B activation by CD mutants. Conceptually, this could lead to a loss of feedback inhibition of NF-B activation mediated by NOD2 in patients with CD. Paradoxically, the CD-associated mutant forms of NOD2 have previously been found to exhibit base line-reduced activation of NF-B compared with wild type. However, the loss of the down-regulating effect mediated by CENTB1 might have pathophysiological significance by facilitating persistent basal activation.
CENTB1 was identified by sequencing clones obtained from a size-fractionated myeloid cell cDNA library (27,28). The deduced protein contains 740 amino acids and shares 20% sequence identity with Centaurin ␣1, a protein thought to be FIGURE 8. Effect of CENTB1 on NF-B activation via NOD1. A, CENTB1 down-regulates NOD1 induced-NF-B activation in a concentration-dependent manner. HEK293 cells were seeded into 24-well plates and transfected the following day with NOD1 (10 ng), CENTB1 at different concentrations (10 -100 ng), NF-B luciferase, and Renilla reporter plasmids. The cells were stimulated with iE-DAP (100 ng/ml) and were lysed 18h later. The luciferase activity was determined using an automated luminometer. The results are provided as the means ϩ S.D. of fold induction of NF-B activity. *, p Ͻ 0.05; **, p Ͻ 0.01. B, specific siRNA against CENTB1 increased ligand-dependent NOD1-mediated activation of NF-B. HEK293 cells were transfected with above mentioned plasmids and two different siRNA against CENTB1 (100 ng) and pSUPER vector as a control. All of the experiments were repeated at least four times. *, p Ͻ 0.05. most highly expressed in neurons that has been associated with the formation of neuritic plaques in Alzheimer disease (29). CENTB1 (also named ACAP1), previously identified as a GTPase-activating protein, is a member of the ADP-ribosylation factor family. This protein is composed by a coiled-coil region at its N terminus, followed by an X-box domain, a PH domain, an ARF-GAP domain, and C-terminal ankyrin repeats (27). CENTB1 is encoded by a gene located on the short arm of human chromosome 17, and there has previously been little known about its function (28). To our knowledge this is the first protein interaction identified for CENTB1. A recent study showed that Centaurin ␣1 contributes to extracellular signalregulated kinase activation induced by epidermal growth factor (30). Of note, a selected partner, Centaurin ␤4 protein appears to be involved in a variety of cellular signaling pathways including control of growth, survival, and movement and may act to directly assemble or disassemble molecular complexes (31). Its expression is significantly greater in high grade uveal melanomas with more significant risk of metastasis than in low grade tumors (32). Expression has also been found in breast cancer cell lines (33).
A previous report described high CENTB1 expression in spleen, thymus, and bone marrow; intermediate expression in lung and testis; and low expression in prostate and ovary. No expression was detected in heart, brain, placenta, liver, skeletal muscle, kidney, and pancreas (28). However, expression in the intestine and immune cells was apparently not assessed. In the present study, base-line endogenous expression of CENTB1 was not detected in several colonic epithelial cell lines but was present in several immune system cell lines such as THP-1,  . CENTB1 interaction with human disease-associated NOD2 mutants. A, immunoprecipitation (IP) shows direct binding of CENTB1 and the three most common mutants (L1007fs, R702W, and G908R) associated with Crohn's disease or the two most common mutants (R334W and L496F) associated with Blau syndrome. HEK293 cells were co-transfected with Xpress-CENTB1 and FLAG-NOD2 wild type and different mutations. Immunoprecipitation showed interaction between CENTB1 with CD and BS mutants in HEK293 cells. B, CENTB1 does not down-regulate NF-B activity of CD NOD2 mutants. HEK293 cells were transfected with the following DNA constructs of NOD2 CD mutants (L1007fs, R702W, and G908R) or NOD2 wild type (1 ng), CENTB1 (100 ng), NF-B, and Renilla reporters. Luciferase activity was measured as described above. C, CENTB1 downregulates NF-B activity by BS NOD2 mutants. HEK293 cells were seeded in 24-well plates and transfected with 1 ng of DNA encoding NOD2 BS mutants (R334W and L496F), NOD2 wild type (1 ng), NOD1 (10 ng), CENTB1 (100 ng), NF-B luciferase, and Renilla reporters and stimulated with 1 g of MDP for 18 h. NF-B activity was determined by measurement of luciferase and normalization with Renilla. *, p Ͻ 0.01. WB, Western blot.
Jurkat, and B cells in basal conditions. Interestingly, endogenous expression of CENTB1 was detected in surgical specimens from active ulcerative colitis and CD patients in contrast to surgical specimens of normal tissue where expression could not be detected, suggesting that CENTB1 is increased in association with inflammation. To examine the relationship between CENTB1 and the inflammatory condition, colonic epithelial and immune cells were stimulated with proinflammatory cytokines. CENTB1 was up-regulated in response to TNF-␣ and IL-1␤ and highly in colonic epithelial cell lines after stimulation with NOD2 and NOD1 ligands (MDP and iE-DAP), both fragments of bacterial cell walls. Induction of CENTB1 expression by NOD ligands appears to be specific in the same manner that CENTB1 appears to selectively affect NOD induced NF-B activation. Thus, LPS did not induce CENTB1 expression, consistent with findings that CENTB1 does not regulate TLR-mediated activation. These findings suggest an auto-regulatory loop of NOD pathways in mammalian cells involving CENTB1 but not TLR-2 and TLR-4 pathways. Given that CENTB1 can down-regulate NF-B activity, it may have a role in host defense ameliorating the NF-B response and ultimately modulating inflammation (34).
In summary, CENTB1 acts as a component of the host innate immune response system by down-regulating NF-B activation via the NOD2 and NOD1 pathways. CENTB1 expression is up-regulated by bacterial components (MDP-LD and iE-DAP) and proinflammatory molecules such as TNF-␣ and IL-1␤ and could play a moderating influence in the context of infectious and inflammatory challenges. These studies identify CENTB1 as a regulator of NOD2 signaling and demonstrate a novel role for CENTB1 in inflammatory responses.