Interferon-γ Augments CARD4/NOD1 Gene and Protein Expression through Interferon Regulatory Factor-1 in Intestinal Epithelial Cells*

Although intestinal epithelial cells appear to be functionally hyporesponsive to normal intestinal flora, human intestinal epithelial cells can respond to enteroinvasive bacteria and induce an inflammatory response. This initial inflammatory response leads to the recruitment of polymorphonuclear leukocytes to the affected site in vitro and in vivo. CARD4/NOD1 is a potential cytosolic receptor for peptidoglycan in mammalian cells that resembles pathogen-resistant proteins of plants. In this context, CARD4/NOD1 is a candidate for a recognition protein of intracellular bacteria or peptidoglycan in intestinal epithelial cells. In this study, we demonstrate that CARD4/NOD1 is constitutively expressed in intestinal epithelial cell lines and isolated primary intestinal epithelial cells. Interferon-γ (IFNγ), which is a potent pro-inflammatory cytokine in intestinal mucosal inflammation, activates CARD4/NOD1 mRNA transcription in a time- and dose-dependent manner and results in augmentation of CARD4/NOD1 protein in SW480 cells. Promoter analysis of CARD4/NOD1 indicates that interferon regulatory factor-1 (IRF-1) binding motif (–791 to –782) is essential for the effect of IFNγ. Nuclear extracts from SW480 cells treated with IFNγ show specific binding of oligonucleotides corresponding to this IRF-1-binding motif, which was supershifted by anti-IRF-1 antibody in electrophoretic mobility shift assay. Overexpression of IRF-1 protein activates the CARD4/NOD1 promoter but not the deletion mutant of the IRF-1-binding site in a co-transfection assay of IRF-1 expression plasmid with CARD4/NOD1 promoter. These studies suggest that the Th1 cytokine, IFNγ, activates CARD4/NOD1 transcription and regulate innate immune mechanisms in the condition of intestinal mucosal inflammation.


Although intestinal epithelial cells appear to be functionally hyporesponsive to normal intestinal flora, human intestinal epithelial cells can respond to enteroinvasive bacteria and induce an inflammatory response. This initial inflammatory response leads to the recruitment of polymorphonuclear leukocytes to the affected site in vitro and in vivo. CARD4/NOD1 is a potential cytosolic receptor for peptidoglycan in mammalian cells that resembles pathogen-resistant proteins of plants. In this context, CARD4/NOD1 is a candidate for a recognition protein of intracellular bacteria or peptidoglycan in intestinal epithelial cells. In this study, we demonstrate that CARD4/NOD1 is constitutively expressed in intestinal epithelial cell lines and isolated primary intestinal epithelial cells. Interferon-␥ (IFN␥), which is a potent pro-inflammatory cytokine in intestinal mucosal inflammation, activates CARD4/NOD1 mRNA transcription in a time-and dose-dependent manner and results in augmentation of CARD4/NOD1 protein in SW480 cells.
Promoter analysis of CARD4/NOD1 indicates that interferon regulatory factor-1 (IRF-1) binding motif (؊791 to ؊782) is essential for the effect of IFN␥. Nuclear extracts from SW480 cells treated with IFN␥ show specific binding of oligonucleotides corresponding to this IRF-1-binding motif, which was supershifted by anti-IRF-1 antibody in electrophoretic mobility shift assay. Overexpression of IRF-1 protein activates the CARD4/NOD1 promoter but not the deletion mutant of the IRF-1-binding site in a co-transfection assay of IRF-1 expression plasmid with CARD4/ NOD1 promoter. These studies suggest that the Th1 cytokine, IFN␥, activates CARD4/NOD1 transcription and regulate innate immune mechanisms in the condition of intestinal mucosal inflammation.
In the gastrointestinal tract, intestinal epithelial cells form a structural and functional barrier against numerous luminal bacteria, pathogenic and nonpathogenic. Previous studies indicate that IECs 1 are relatively hyporesponsive to extracellular bacterial components such as lipopolysaccharide. However, when either bacterial components or invasive bacteria gain access to intracellular compartments, they may readily induce cellular responses by IECs (1)(2)(3).
Plants were first noted to have both cytosolic and membrane pathogen-binding proteins that contribute to resistance to pathogenic microorganisms (4 -6). More recently, mammalian cells have also been shown to express binding pattern recognition receptors including membrane toll-like receptors (TLRs) and cytosolic pathogen recognition proteins designated NODs. CARD4/NOD1 and CARD15/NOD2 were first identified as homologues of Apaf-1. CARD4/NOD1 and CARD15/NOD2 have highly conserved domain structures composed of caspase recruit domains, nucleotide-binding domains, and leucine-rich repeats (7)(8)(9). Leucine-rich repeats domain are responsible for recognition of pathogen-associated molecular patterns by TLRs. Similarly, the leucine-rich repeat domain in NOD2 provides the recognition of bacterial peptidoglycan (10,11). NODs activate NF-B through a RICK signaling pathway following peptidoglycan stimulation (12)(13)(14).
Disorders of mucosal inflammation, such as inflammatory bowel disease, are associated with destruction of the mucosal barrier, which results in increasing susceptibility to luminal pathogen and penetration of normal flora. The interaction between luminal microorganisms and intestinal epithelial cells has been considered a critical point for initiation and perpetuation of mucosal inflammation in murine colitis models and human inflammatory bowel disease (15,16). Therefore, we investigate the expression of the cytosolic bacterial pathogen recognition molecule, CARD4/NOD1 in IECs, and the regulation by the pro-inflammatory cytokine IFN␥, which has been implicated as a key regulatory cytokine in inflammatory bowel disease.
Northern Blot Analysis-For Northern blot analysis, 15 g/lane of total RNA was loaded on a formaldehyde gel and transferred to a Nytran SuPerCharge membrane (Schleicher & Schuell). The cDNA probe for CARD4/NOD1 was obtained by EcoRI and SmaI (Promega, Madison, WI) digestion (1.5 kb) of the pCl CARD4-HA plasmid kindly provided by Dr. John Bertin (Millennium Pharmaceuticals Inc., Cambridge, MA). cDNA fragment for IRF-1 (439 bp) was obtained by RT-PCR in SW480 treated with IFN␥ using forward primer (5Ј-GAC CCT GGC TAG AGA TGC AG-3Ј) and reverse primer (5Ј-GAG CTG CTG AGT CCA TCA G-3Ј). PCR product was subcloned into pCR2.1 TOPO vector (Invitrogen) and sequenced. cDNA probe for IRF-1 was obtained by EcoRI digestion. Labeling of probes with [␣-32 P]dCTP was performed using Ready To Go DNA Labeling Beads (Amersham Biosciences). Prehybridization and hybridization were performed in QuikHyb Hybridization Solution (Stratagene, La Jolla, CA).
Primary Intestinal Epithelial Cell Isolation-Intestinal epithelial cells were isolated from freshly obtained colonic mucosal biopsies by modifications of a technique previously described (17). Briefly, colonic mucosal biopsies from patients were immediately placed into cold RPMI 1640 medium (Cellgro Mediatech Inc.) and then washed with PBS without Ca 2ϩ and Mg 2ϩ . The samples were then washed with 1 mM dithiothreitol in PBS and then washed with PBS. The samples were immediately placed into 1 mM EDTA in PBS and incubated at 37°C. When crypts were obtained from the biopsy samples, they were placed into 0.3 mM sodium tetraphenylborate (Sigma-Aldrich) in PBS at 37°C. The epithelial cells were collected and centrifuged at 400 ϫ g. RPMI 1640 was added, and the cells were isolated individually by micropipetting under the microscope. Medium containing epithelial cells was placed in 0.5 ml of Trizol (Invitrogen) with 50 g of Escherichia coli transfer RNA (Roche Applied Science) as carrier for RNA extraction. Total cellular RNA was extracted by Trizol according to the manufacturer's instructions. Total RNA obtained from 10 -20 primary epithelial cells was eluted in 10 l of nuclease free water. RNA was reversetranscribed using 1 l (50 units) of SuperScript II transcriptase (Invitrogen), and 2 l (10% of total volume) of reverse transcription products from RNA of primary intestinal epithelial cells were used for PCR.
Expression Plasmid and Transfection-pCl CARD4/NOD1-HA expression plasmid vector was kindly provided by Dr. John Bertin (Millennium Pharmaceuticals Inc.). Protein expression of HA-tagged CARD4 was confirmed by Western blot using anti-HA monoclonal antibody (Covance Inc., Princeton, NJ) in COS7 cells transiently transfected with pCl CARD4-HA plasmid. The ability of the pCl CARD4-HA plasmid to activate NF-B was confirmed by NF-B reporter assay in transiently transfected HEK293 cells using PathDetect in Vivo Signal Transduction Pathway cis-Reporting Systems (Stratagene). pCDNA IRF-1 expression plasmid vector and pCDNA3.1 empty vector was kindly provided by Dr. Mark Perrella (Brigham and Women's Hospital, Boston, MA). Transfection was performed using LipofectAMINE Plus reagent (Invitrogen) according to the manufacturer's instructions.
Immunoprecipitation and Immunoblotting-The cells were grown in 6-well plates and stimulated with 10 ng/ml IFN-␥ for 18 h and washed, and then 300 l of lysis buffer (1% Triton X-100, phenylmethylsulfonyl fluoride, protease inhibitor, EDTA, Tris/HCl, NaCl, pH 7.4, supplemented with protease-inhibitor mixture, Complete Mini (Roche Applied Science)) was added. After 10 min of incubation, cell-free lysates were obtained and stored at Ϫ80°C. Protein concentration was determined using a Bio-Rad DC protein assay kit. For immunoprecipitation, a total 2 mg of protein was mixed with 2 g of affinity-purified rabbit anti-CARD4/NOD1 antiserum (HM3851) and 10 l of Hitrap protein A-Sepharose beads (Amersham Biosciences). After overnight incubation at 4°C, the immunocomplexes were washed with lysis buffer and solubilized in 30 l of SDS sample buffer (1.25% SDS, 2.5% glycerol, 62.5 Total RNA for each sample was obtained from 10 -20 isolated primary intestinal epithelial cells. RT-PCR for CD45 and CD68 was performed to confirm the lack of contamination by nonepithelial cells (data not shown). The sensitivity of RT-PCR of CD45 and CD68 was validated by detection of an appropriate product using total RNA from a single Jurkat cell and THP-1 cell. As an internal control, RT-PCR of GAPDH (440 bp) was performed.
Construction of CARD4 Luciferase Reporter Constructs-The genomic sequence adjacent to the human CARD4/NOD1 exon 1 and promoter sequences were found in a GenBank TM working draft sequence of chromosome 7 (see Fig. 5A). The primers were designed corresponding to sequences overlapping the CARD4 exon1 at ϩ21 and also a sequence Ϫ2,128 bp upstream. This putative promoter region was amplified from human genomic DNA (Promega) by PCR using Expand high Fidelity PCR System (Roche Applied Science). The 2,128-bp PCR fragment was cloned into the promoterless luciferase reporter vector pGL3-basic (Promega). This construct was confirmed by sequencing and named pGL-2128. The CARD4 promoter deletions (see Fig. 5B) were generated by restriction enzyme digestion using a blunting kit (Takara Biochemical, Berkeley, CA) or by PCR.
Luciferase Reporter Assay-1 ϫ 10 5 /well of SW480 cells were seeded on 12-well plates (Costar, Cambridge, MA). After 24 h, the cells were transiently transfected with reporter plasmids, together with the indicated expression vectors. After 10 h of transfection, the cells were treated with 100 ng/ml of IFN␥ for 16 h. The luciferase activity of total cell lysates was measured using the dual luciferase reporter assay system (Promega). The Renilla luciferase reporter gene (10 ng/well) (Promega) was used as an internal control.
Electrophoretic Mobility Shift Assay-Nuclear proteins were prepared as previously described (18). Both double-stranded oligonucleotides corresponding to Ϫ807 to Ϫ768 bp and mutated oligonucleotides were used as probes or cold competitors to analyze the interaction between IRF-1 protein and DNA (see Fig. 6, A and B). The probes were labeled with T4 polynucleotide kinase (Promega) with [␥-32 P]ATP. The binding reaction was performed as previously described (19). For a competition assay, a 100-fold excess of unlabeled oligonucleotide was added to the reaction. To perform supershift assay, the binding mixture were incubated for 10 min at room temperature in the presence of 1 l of affinity-purified polyclonal anti-IRF-1 antibody (C-20X, Santa Cruz). The samples were fractionated on 4% nondenaturing polyacrylamide gel in 0.5 ϫ TBE buffer (0.045 M Tris-borate, 1 mM Na 2 EDTA). The resultant DNA-protein complexes were detected by autoradiography.

RESULTS AND DISCUSSION
Although intestinal epithelial cells are exposed to numerous bacteria and their components, they appear functionally to be hyporesponsive to normal flora or their components in vivo, an adaptation that seems appropriate to protect host from constant immune and inflammatory activation. However, previous studies have indicated that intestinal epithelial cells can respond to enteroinvasive bacteria to stimulate the recruitment of polymorphonuclear neutrophils to an affected site in vivo and in vitro. Enteroinvasive bacteria, but not noninvasive bacteria, activate NF-B in intestinal epithelial cell lines and induce expression of several genes encoding inflammatory me- Lysates from nontransfected COS7 cells were used as a negative control. B, protein obtained from SW480 cells cultured with 100 ng/ml of IFN␥ for 6, 12, 24, and 48 h were immunoprecipitated with affinitypurified anti-CARD4/NOD1 antiserum (HM3851) and then immunoblotted with biotinylated anti-CARD4/NOD1 antiserum (HM3847). The lysates from COS7 cells transiently transfected with pCl CARD4-HA were used as a positive control. diators including IL-8, ICAM-1, monocyte chemoattractant protein-1, or inducible nitric-oxide synthase in vitro (1,2,20). These results suggest that intracellular presentation of bacteria may lead to initial responses of intestinal epithelial cells that do not occur when bacteria are only encountered in the extracellular environment. Thus, intestinal epithelial cells have mechanisms for the recognition of intracellular bacterial pathogenic components that are presented by invasive bacteria.
Recently, Philpott et al. (3) reported that bacterial lipopolysaccharide can induce NF-B activation in microinjected HeLa or Caco2 cells. Girardin et al. (21) demonstrated that invasive Shigella flexneri, but not noninvasive S. flexneri, lead to activation of c-Jun in HeLa cells. Activation of NF-B and JNK by invasive S. flexneri is independent of the IL-1/TLR pathway in HEK293 cells. After infection with invasive S. flexneri, JNK activation is significantly increased in CARD4/ NOD1 overexpressing HEK293 cells. CARD4/NOD1 oligomerization and RICK-IKK␣ complex are found in CARD4/NOD1overexpressing HEK293 cells following invasive S. flexneri infection, and these effects are inhibited by a dominant nega-tive CARD4/NOD1 (21). Collectively, these data suggest that CARD4/NOD1 is a candidate to serve as a pivotal receptor for intracellular bacteria. However, CARD4/NOD1 expression and its regulation have not been previously demonstrated in IEC lines. Therefore, we investigated CARD4/NOD1 mRNA expression in IEC lines.
Following demonstration of mRNA expression of CARD4/ NOD1 by IEC lines, we investigated its expression in primary isolated intestinal epithelial cells to assess the relevance of the observations using the in vitro models. Isolated whole crypts were obtained from colonic biopsies after treatment with EDTA. Following treatment with sodium tetraphenylborate, single epithelial cells were isolated from whole crypts (Fig. 2A). To avoid contamination by nonepithelial origin cells, isolated or without (open bars) 100 ng/ml of IFN␥, and then luciferase activity was measured. The experiments were performed in triplicate. The data are presented as the relative fold (means Ϯ S.D.) compared with SW480 cells transfected with pGL3 basic empty vector. The results were confirmed in three independent experiments. D, relative induction of luciferase activity by IFN␥. The data are presented as relative fold (means Ϯ S.D.) of luciferase activity of SW480 with IFN␥ compared that with medium in three independent experiments. Each experiment was performed in triplicate. epithelial cells were picked by micropipetting. Markers for lymphocytes and monocytic cells (CD45 and CD68) were not detected in these isolated intestinal epithelial cells by highly sensitive RT-PCR (data not shown). Total RNA from a single Jurkat cell and THP-1 cell served as positive controls for CD45 and CD68, respectively. As shown in Fig. 2B, CARD4/NOD1 (374 bp) was expressed in all isolated primary intestinal epithelial cells examined.
Effect of IFN␥ on the Expression of CARD4/NOD1-Regulation of CARD4/NOD1 expression has not been described previously. As shown in Fig. 3A, IFN␥ augmented CARD4/NOD1 mRNA expression in SW480 cells. Other cytokines examined (tumor necrosis factor ␣, IL-1␤, IL-4, and transforming growth factor-␤) did not affect CARD4/NOD1 mRNA expression. The  -1 A, B, and C in bold type) in CARD4/NOD1 promoter. B, oligonucleotides sequences of wild type and mutated IRF-1 consensus motifs in CARD4/NOD1 promoter (wild type (WT) and mutated sequences (Mut)) for electrophoretic mobility shift assay. IRF-1 consensus sequences are underlined, and mutations are in bold type. C, specific binding of IRF-1 to the IRF-1 consensus sequences in CARD4/NOD1 promoter. Electrophoretic mobility shift assay was performed using 9 g of nuclear protein from SW480 cells treated with 100 ng/ml of IFN␥ for the indicated time periods. The competition assay was performed with a 100-fold excess of unlabeled oligonucleotides. The supershift assays were done by the addition of 1 g of anti-IRF-1 antibody. effect of IFN␥ on CARD4/NOD1 mRNA expression is time-and concentration-dependent as assessed by Northern blot analysis (Fig. 3, B and C).
Although CARD4/NOD1 is considered important for recognition for intracellular pathogen, previous studies have not demonstrated active CARD4/NOD1 protein in vitro or in vivo. To investigate the expression of CARD4/NOD1 protein, we generated anti-CARD4/NOD1 sera. The specificity and the sensitivity were confirmed using lysates from COS7 cells transiently transfected with the HA-tagged CARD4 expression plasmid, pCl CARD4-HA (Fig. 4A). Consistent with the regulation of mRNA expression, CARD4/NOD1 protein was also augmented in SW480 cells by IFN␥ stimulation (Fig. 4B).
IRF-1 Is Essential for Up-regulation of CARD4/NOD1 Transcription by IFN␥-To identify the transcriptional regulation of CARD4/NOD1, we constructed a series of luciferase reporter vectors containing up to 2,128 bp corresponding to the DNA sequence upstream of base 1 and extending 21 bp into the first exon of CARD4/NOD1 (Fig. 5, A and B). Luciferase activity in SW480 cells transfected with a vector containing the entire 2,128-bp upstream DNA (pGL-2128) was 125 Ϯ 16-fold higher than that obtained with the empty pGL3 basic vector. This promoter lacks a TATA box but has several GC-rich motifs including several SP-1-binding sites, near the transcription start site based on results of computer analysis using TRANS-FAC data base of the University of Pennsylvania. Promoter activity was significantly decreased in pGL-26 (4.2 Ϯ 1.4) compared with in pGL-367 (34.4 Ϯ 1.1), indicating that Ϫ26 to Ϫ367 upstream of exon1 is essential for basal CARD4/NOD1 expression. As shown in Fig. 5 (C and D), IFN␥ increased 80% luciferase activity in SW480 cells transfected with pGL-2128. Luciferase activity of cells transfected with deletion constructs pGL⌬-837-546, pGL-837, and pGL-546 demonstrated that sequences within Ϫ837 to Ϫ546 promoter are essential for activation by IFN␥. Three IRF-1-binding motifs (IRF-1A, Ϫ791 to Ϫ782; IRF-1B, Ϫ787 to Ϫ778; and IRF-1B, Ϫ694 to Ϫ689) (Figs. 5B and 6A) are clustered in this region. Luciferase activity of cells transfected with pGL-837, pGL-773, and pGL-729 suggested that the most distal IRF-1-binding motif (IRF-1A, Ϫ791 to Ϫ782) is essential for the IFN␥ effect (Fig. 5, C and D).
To determine whether IRF-1 binds to this IRF-1-binding motif in the CARD4/NOD1 promoter sequences, we performed electrophoretic mobility shift assay using nuclear extracts from SW480 cells with IFN␥ treatment. Consistent with the promoter analysis, oligonucleotides corresponding to IRF-1 binding sequences (IRF-1A) in CARD4/NOD1 promoter bound nuclear IRF-1 protein in SW480 cells treated with 100 ng/ml of IFN␥ time-dependently (Fig. 6C, lanes 1-4). The binding of IRF-1 was inhibited by unlabeled oligonucleotides competitor but not by oligonucleotides with mutations at IRF-1-binding motifs (lanes 5 and 6). The band reflecting the complex with IRF-1 oligonucleotides was supershifted by anti-IRF-1 antibody (lanes 4 and 7). Signal transducers and activators of transcription (STAT)-1-binding motifs, IFN␥ activation site, and interferon-stimulated response element were not present as far as 5 kb upstream of the promoter region of CARD4/ NOD1, as determined by computer analysis using TRANSFAC data base of University of Pennsylvania. We hypothesized that the effect of IFN␥ on CARD4/NOD1 expression depends on IRF-1, which has an interferon-␥ activation site motif in the promoter region. Transcription of IRF-1 is regulated by IFN␥ through Janus tyrosine kinase/STAT pathway (22). To test this hypothesis, we first demonstrated that IRF-1 mRNA and nuclear protein expression were rapidly up-regulated by IFN␥ treatment in SW480 cells (Fig. 7). To confirm that overex-pressed IRF-1 can activate CARD4/NOD1 promoter, we next performed promoter analysis using SW480 cells co-transfected with IRF-1 expression plasmid. Promoter activity of pGL-2128 was activated 2.0-fold (71.8 Ϯ 0.6 versus 35.9 Ϯ 0.9) in SW480 cells co-transfected with pCDNA IRF-1 expression vector. The effect of IRF-1 was completely abrogated in pGL⌬-837-546, which lacks the IRF-1 cluster region (Fig. 8). Collectively, these results suggest that rapid augmentation of nuclear IRF-1 protein by IFN␥ treatment results in activation of CARD4/NOD1 transcription.
IFN␥ is a potent pro-inflammatory cytokines in acute and chronic mucosal inflammation in the intestine (23). In Crohn's disease and ulcerative colitis as well as murine models of colitis, the Janus tyrosine kinase/STAT pathway is activated in inflamed mucosa (24,25). As shown in recent studies, the membrane pathogen recognition molecules TLR4 and MD2 are regulated by IFN␥ in IECs (26,27). Augmentation of TLR4 and MD2 by IFN␥ priming resulted in acquisition of responsiveness to lipopolysaccharide in HT29 cells, which are relatively hyporesponsive to lipopolysaccharide without IFN␥ priming. Secondarily, CARD15/NOD2, another cytosolic receptor for peptidoglycan, is augmented by tumor necrosis factor ␣ in IECs (28 -30). These findings provide evidence that pathogen recognition molecules are modulated by pro-inflammatory cytokines and suggest that the latter promote host defenses to pathogens in inflammatory conditions. In acute and chronic intestinal inflammation, luminal pathogens or their components can invade more easily, because intestinal mucosal barrier function is impaired. In patients with inflammatory bowel disease, it has been suggested that luminal bacteria contribute to the perpetuation of chronic inflammation (31,32). In this context, the pro-inflammatory cytokine, IFN␥, may readily lead to CARD4/ NOD1 augmentation in IECs.
Here, we first demonstrate that IECs express CARD4/NOD1 mRNA and protein and that IFN␥ activates transcription of CARD4/NOD1 mRNA through IRF-1 in IECs. These data suggest that intestinal epithelial cells provide a functional as well as structural barrier to the entry of luminal bacteria or their components into the underlying lamina propria via the cytosolic protein CARD4/NOD1.