Caspase-1 Activates Nuclear Factor of the (cid:1) -Enhancer in B Cells Independently of Its Enzymatic Activity*

The proteolytic activity of caspases is involved in apoptosis and inflammation. In this regard, caspase-1 is required for pro-interleukin (IL)-1 (cid:2) and pro-IL-18 maturation. We report here on a novel function of caspase-1 as an activator of nuclear factor of the (cid:1) -enhancer in B-cells (NF- (cid:1) B) and p38 mitogen-activated protein kinase (MAPK). This function is not shared by the murine caspase-1 homologues caspase-11 and -12. In contrast to pro-IL-1 (cid:2) maturation, caspase-1-induced NF- (cid:1) B activation is not inhibited by the virus-derived caspase-1 inhibitor cytokine response modifier A and is equally induced by the enzymatically inactive caspase-1 C285A mutant. Although the general NF- (cid:1) B-inhibiting protein A20 inhibits caspase-1-derived activation of NF- (cid:1) B, dom-inant-negative forms of TRAF2 and RIP1 have no effect. We demonstrate that caspase-1 interacts with RIP2 and that dominant-negative forms of RIP2 and I (cid:1) B kinase complex- (cid:2) inhibit caspase-1-mediated NF- (cid:1) of pro-IL-1 (cid:2) and RIP2-dependent of NF- (cid:1) B and by the caspase recruitment domain. g by the addition of control plasmid DNA. Microscopy fluorographs show that caspase-1 induces apoptotic cell death of transfected cells that can be blocked by CrmA. As a control, empty vector-transfected cells do not die. B , 293T cells were transiently transfected with a plasmid encoding 0.1 (cid:4) g of pro-IL-1 (cid:2) , either alone or in combination with 0.2 (cid:4) g of a plasmid encoding caspase-1 ( CASP1 ), and 1 (cid:4) g of a CrmA-encoding plasmid. Total DNA was maintained at 1.3 (cid:4) g by the addition of control plasmid DNA. Supernatant was analyzed for the presence of biologically active IL-1 (cid:2) 24 h after transfection. C , 293T cells were transiently cotransfected with a NF- (cid:1) B-dependent luciferase reporter and the indicated plasmids. Cells were lysed 24 h after transfection, and NF- (cid:1) B activity was measured as described under “Experimental Procedures.”

NF-B 1 refers to a class of transcription factors involved in immune regulation, apoptosis, differentiation, inflammation, and cancer (1). NF-B is sequestered in the cytoplasm as an inactive complex bound by an inhibitor known as IB (2). In response to a variety of signaling events, the IB kinase complex (IKK) phosphorylates IB proteins. This post-translational modification targets IB for poly-ubiquitination and subsequent degradation by the 26 S proteasome (3). The degradation of IB proteins liberates NF-B, allowing this transcription factor to translocate to the nucleus and activate its target genes. Besides regulation by IB, NF-B-dependent gene expression is also negatively regulated by the zinc finger protein A20, although the molecular mechanism remains unclear (4 -6).
NF-B activation is involved in the transcriptional activation of numerous inflammation-related and antiapoptotic genes in response to cytokines, bacterial products, and cellular stress conditions (1). Pathogen-derived products such as LPS, peptidoglycan, and double-stranded RNA, respectively, lead to NF-B activation through stimulation of TLR-2, -3, and -4 at the cell surface (7). Besides pathogen-derived products, the pro-inflammatory cytokines TNF and IL-1 also induce NF-B activation by binding to their respective cell surface receptors (2,8). These receptors activate NF-B through distinct signaling pathways (7)(8)(9).
Caspases are an evolutionarily conserved family of aspartate-specific, cysteine-dependent proteases that are synthesized as pro-enzymes with a p20 and a p10 domain. Upon activation, the latter domains undergo a conformational change to form the active enzyme (10). The enzymatic activity of caspases has been implicated in apoptosis and inflammation (11). Caspase-1 plays a key role in inflammatory responses by cleaving pro-IL-1␤ and pro-IL-18 into secreted pro-inflammatory cytokines (12)(13)(14). It has recently been discovered that the latter cytokines are matured in a large caspase-1 containing protein complex, named the "inflammasome" (15). Caspase-1 contains an N-terminal caspase recruitment domain (CARD). This protein module of ϳ100 amino acids is a homotypic oligomerization domain shown to be involved in the assembly of protein platforms that promote proteolytic activation of recruited caspases in the context of apoptosis and inflammation. However, because the family of identified CARD-containing proteins has recently expanded, it has become apparent that the majority of these CARDs are not involved in caspase activation. Instead, many participate in NF-B signaling pathways associated with innate and adaptive immune responses (reviewed in Ref. 16). NOD1 and NOD2 are two prototypical examples that have recently been identified as CARD-containing sensors of peptidoglycan derived from intracellular pathogens (17)(18)(19). After oligomerization, both proteins associate with the CARD-containing kinase RIP2 through CARD-CARD interactions (20,21). RIP2 then recruits the IKK complex through a direct interaction of its intermediate domain with IKK-␥, leading to the activation of NF-B (22)(23)(24).
ICEBERG and COP/Pseudo-ICE are human-specific CARDonly proteins that share 93% and 73% sequence homology, respectively, with the prodomain of caspase-1 (25)(26)(27). Both ICEBERG and COP/Pseudo-ICE are encoded by caspase-like genes that have acquired premature nonsense mutations that lead to the production of essentially CARD-only molecules. It is interesting that their genes are mapped to chromosome 11q22, adjacent to the caspase-1 gene, and have probably arisen by a recent gene duplication event. Both proteins bind to and prevent caspase-1 activation and the subsequent generation of IL-1␤ (25)(26)(27). However, in contrast to ICEBERG, COP/Pseudo-ICE also interacts with RIP2 in a heterotypic CARD-CARD interaction, and activates the transcription factor NF-B (25,27).
Herein, we show that caspase-1, but not its closest murine homologues caspase-11 and -12, induces p38 MAPK and NF-B activation independently of its enzymatic activity. This suggests that caspase-1 is involved in additional pro-inflammatory pathways besides the maturation of pro-IL-1␤ and pro-IL-18.

EXPERIMENTAL PROCEDURES
Plasmids-The cloning of cDNAs encoding murine caspases-1, -11, and -12 has been described previously (28). pCAGGS-caspase-1 C284A, coding for the enzymatically inactive mutant of murine caspase-1, was constructed by site-directed mutagenesis PCR. pCAGGS-caspase-12 C298A has been described elsewhere (29). pCAGGS-caspase-11 C254A, encoding an inactive caspase-11 mutant, and pCAGGS-caspase-1 2D/E, in which the cleavage sites Asp-103 and Asp-122 were mutated to Glu, were kind gifts from Dr. P. Schotte (Ghent University, Ghent, Belgium). Caspase-1 deletion mutants were generated by PCR using modified complementary PCR adaptor primers. E-epitope tagging was done by cloning the PCR-generated cDNAs of the respective open reading frames into the pCAGGS-E vector. The cDNA encoding full-length human caspase-1 was amplified from a THP-1 cDNA library and cloned in frame with the N-terminal E-tag epitope into a pCAGGS-E vector. The enzymatically inactive human caspase-1 C285A mutant was made by site-directed mutagenesis PCR and cloned in frame with the Eepitope tag of pCAGGS-E vector.
The plasmid pNF-conLuc, encoding the luciferase reporter gene driven by a minimal NF-B responsive promoter, was a generous gift from Dr. A. Israël (Institut Pasteur, Paris, France). Plasmid pUT651, encoding ␤-galactosidase, was obtained from Eurogentec (Seraing, Belgium). pCAGGS-pro-IL-1␤ has been described previously (28). pEGFP-C3 was purchased from BD Biosciences Clontech. pCAGGS-CrmA has been described previously (30). pCR3-RIP2 and pCR3-RIP2-CARD were kindly provided by Dr. J. Tschopp (University of Lausanne, Epalinges, Switzerland) and have been described elsewhere (31). Plasmids encoding dominant-negative forms of IKK-␤ and TRAF2 were generous gifts from Dr. J. Schmid (University of Vienna, Vienna, Austria) and Dr. D. V. Goeddel (Genentech, South San Fransisco, CA), respectively. The plasmid encoding murine A20 has been described elsewhere (32) and was kindly provided by Dr. K. Heyninck (Ghent University). Plasmids encoding T7-epitope tagged COP/Pseudo-ICE and ICEBERG have been described previously (25) and were kindly provided by Dr. E. S. Alnemri (Thomas Jefferson University, Philadelphia, PA). All the PCR products described above were sequenced to ensure that no errors had been introduced by PCR.
Transfection, Co-immunoprecipitation, and Immunoblotting Assay-293T is a human embryonal kidney carcinoma cell line. 293T cells were routinely transfected using the calcium phosphate precipitation method (33). Cells were seeded the day before transfection at 2 ϫ 10 5 cells/well. Cells were transfected for 4 h, washed, and incubated for another 24 h before lysates were prepared and/or supernatants were collected and tested in a biological assay for IL-1␤. Lysates were prepared by harvesting the cells and lysing them in ice-cold Nonidet P-40 lysis buffer (10 mM HEPES, pH 7.4, 142.5 mM KCl, 0.2% Nonidet P-40, and 5 mM EGTA), supplemented with 1 mM dithiothreitol, 12.5 mM ␤-glycerophosphate, 1 mM Na 3 VO 4 , 1 mM phenylmethylsulfonyl fluoride, and 1ϫ protease inhibitor mixture (Roche Applied Science). Cell lysates (0.5 ml) were clarified by centrifugation at 14,000 ϫ g for 5 min, and subjected to immunoprecipitation using specific antibodies, including anti-vesicular stomatitis virus (Sigma) and anti-hemagglutinin antibodies (Babco, Richmond, CA) in combination with 15 l of protein A-Sepharose. Immune-complexes were fractionated by SDS-PAGE and transferred to nitrocellulose membranes. The blots were subsequently incubated with various antibodies, including anti-E antibodies (Amersham Biosciences) and anti-vesicular stomatitis virus antibodies (Sigma), followed by horseradish peroxidase-conjugated secondary antibodies, and detection by an enhanced chemiluminescence method. Otherwise, lysates were analyzed directly by immunoblotting after normalization for total protein content. Rabbit polyclonal antibodies against recombinant murine caspase-1, -11, and -12 were prepared at the Centre d'Economie Rurale (Laboratoire d'Hormonologie Animale, Marloie, Belgium). Anti-MAPK and anti-phospho-MAPK antibodies (c-Jun NH 2terminal kinase, extracellular signal-regulated kinase-1/2, and p38) were from Cell Signaling (Beverly, MA). Anti-GFP antibody was from BD Biosciences Clontech.
Pro-IL-1␤ Processing Assay-Biologically active IL-1␤ was determined using growth factor-dependent D10(N4)M cells (34). Cells were maintained in RPMI 1640 medium supplemented with 10% fetal calf serum, 2 mM L-glutamine, 100 IU/ml penicillin G, 100 g/ml streptomycin, 1 mM sodium pyruvate, 5 mM ␤-mercaptoethanol, and 10% supernatant of phorbol ester-stimulated EL-4 cells as a source of IL-2, and 10% supernatant of phorbol ester-stimulated P388D1 cells as a source of IL-1. The day before the assay, D10(N4)M cells were washed and transferred to fresh media containing 10% EL-4 supernatant. The next day, cells were washed again and added to serial dilutions of IL-1␤-containing samples (10 4 cells/well), followed by incubation for 24 h at 37°C in a CO 2 incubator. Proliferation was quantified by A, 293T cells were transiently cotransfected with combinations of plasmids overexpressing nuclear localization signalcontaining green fluorescent protein, caspase-1 (CASP1), CrmA, and empty vector (CTRL). Total DNA was maintained at 1 g by the addition of control plasmid DNA. Microscopy fluorographs show that caspase-1 induces apoptotic cell death of transfected cells that can be blocked by CrmA. As a control, empty vector-transfected cells do not die. B, 293T cells were transiently transfected with a plasmid encoding 0.1 g of pro-IL-1␤, either alone or in combination with 0.2 g of a plasmid encoding caspase-1 (CASP1), and 1 g of a CrmA-encoding plasmid. Total DNA was maintained at 1.3 g by the addition of control plasmid DNA. Supernatant was analyzed for the presence of biologically active IL-1␤ 24 h after transfection. C, 293T cells were transiently cotransfected with a NF-B-dependent luciferase reporter and the indicated plasmids. Cells were lysed 24 h after transfection, and NF-B activity was measured as described under "Experimental Procedures." plate scintillation counter (PerkinElmer Life and Analytical Sciences). IL-1␤ was quantified according to a standard preparation of IL-1␤ with a specific biological activity of 10 9 IU/mg (obtained from the National Institute for Biological Standards and Control, Potters Bar, UK).
Quantification of NF-B Activity-293T cells were transfected with the indicated expression vectors in combination with 100 ng of NF-Bluciferase and pUT651-␤-galactosidase reporter plasmids. In some experiments, cells were treated for 6 h with 500 IU/ml of TNF-␣ before harvesting. Twenty-four hours after transfection the cells were collected, washed in phosphate-buffered saline, and lysed in Tris phosphate (25 mM, pH 7.8), 2 mM dithiothreitol, 2 mM CDTA, 10% glycerol, and 1% Triton X-100. After addition of 50 l of substrate buffer (658 M luciferin, 378 mM co-enzyme A, and 742 M ATP) to 20 l of cell lysates, NF-B activity was assayed in a TopCount NXT microplate scintillation reader (PerkinElmer Life and Analytical Sciences). To normalize transfection efficiency, cell lysates were also subjected to ␤-galactosidase colorimetric assay. In brief, 20 l of cell lysate were incubated for 5 min at room temperature with 200 l of a solution containing 0.9 mg/ml o-nitrophenyl-␤-D-galactopyranoside, 1 mM MgCl 2 , 45 mM ␤-mercaptoethanol, and 100 mM sodium phosphate, pH 7.5. The optical density was read at a wavelength of 595 nm. Results are expressed as relative luciferase units per second/optical density for ␤-galactosidase activity. The data represent the average Ϯ S.E. of at least three independent experiments.

Inhibition of Caspase-1 Enzymatic Activity Revealed a
Caspase-1-dependent NF-B Pathway-Because CARDs have been implicated in NF-B modulation and caspase-1 is an important regulator of inflammation, we investigated whether caspase-1 could be involved in NF-B activation. First, we studied the effect of murine caspase-1 and its enzymatic activity on cell death, pro-IL-1␤ maturation, and activation of NF-B. Caspase-1 induced apoptotic cell death upon overexpression in 293T cells (Fig. 1A). Moreover, when co-expressed with pro-IL-1␤ in these cells, the protease caused maturation of pro-IL-1␤ into its biologically active form (Fig. 1B). Because cell death can mask the induction of NF-B activation (35), we made use of CrmA, a cowpoxderived serpin that covalently binds to the catalytic site and irreversibly inhibits the enzymatic activities of caspase-1 and -8. CrmA overexpression completely abrogated both caspase-1-dependent cell death and pro-IL-1␤ maturation effects (Fig.  1, A and B). It is interesting that when the cells were protected from caspase-1-induced apoptosis by CrmA, a significant activation of NF-B was observed (Fig. 1C). This activation of NF-B occured in a caspase-1-dependent manner and was not caused by CrmA, because even high doses of CrmA did not significantly induce NF-B activation (Fig. 1C). To further confirm that the activation of NF-B was independent of the catalytic activity of caspase-1, we generated an enzymatically inactive C284A mutant of the caspase. Indeed, caspase-1 C284A clearly induced NF-B activation ( Fig.  2A), demonstrating that NF-B activation occurred in a man-

FIG. 2. Caspase-1-induced NF-B activation is independent of pro-IL-1␤ maturation.
A, 293T cells were transiently cotransfected with an NF-Bdependent luciferase reporter and the indicated amounts of plasmid encoding enzymatically inactive murine caspase-1 C284A. Total DNA was maintained at 0.5 g by the addition of control plasmid DNA. 24 h after transfection, lysates were analyzed for NF-B activation as described under "Experimental Procedures." B, 293T cells were transiently transfected with the indicated plasmids. Supernatant was analyzed for the presence of biologically active IL-1␤ 24 h after transfection. C, 293T cells were transiently cotransfected with a NF-Bdependent luciferase reporter and the indicated amounts of plasmid encoding enzymatically inactive human caspase-1 C285A. Total DNA was maintained at 0.5 g by the addition of control plasmid DNA. 24 h after transfection, lysates were analyzed for NF-B activation as described under "Experimental Procedures." D, 293T cells were transiently cotransfected with an NF-B-dependent luciferase reporter, a plasmid encoding pro-IL-1␤ and the indicated amounts of plasmid encoding murine caspase-1 2D/E. Total DNA was maintained at 0.5 g by the addition of control plasmid DNA. 24 h after transfection, supernatants were analyzed for the presence of mature IL-1␤ (top) and cell lysates for NF-B activation (center), as described under "Experimental Procedures." Aliquots of the lysates were analyzed by SDS-PAGE/immunoblotting to confirm expression of caspase-1 2D/E (bottom). ner independent of the enzymatic activity of caspase-1. The possibility that the observed activation of NF-B occurred as a result of an endogenous autocrine loop of IL-1␤, a potent inducer of NF-B, can be excluded, because transfection of 293T cells with wild-type caspase-1 or caspase-1 C284A in the absence of proIL-1␤ did not lead to detectable IL-1␤ in the supernatant of the transfected cells (Fig. 2B). Moreover, caspase-1 C284A was incapable of maturating pro-IL-1␤ when co-expressed (Fig. 2B). To analyze the NF-B-inducing capacity of human caspase-1, its cDNA was cloned from a THP-1 cDNA library, and we generated an enzymatically inactive C285A mutant to exclude interference of cell death with the assay. Dose-dependent induction of NF-B activation was observed upon transfection of human caspase-1 C285A in 293T cells (Fig. 2C), demonstrating that the NF-B-inducing ability is conserved in mouse and man. All together, these results showed that NF-B activation is a novel function of caspase-1 and that, in contrast to the induction of apoptosis and the maturation of pro-IL-1␤, the proteolytic activity of caspase-1 is not needed for the activation of NF-B.
Caspase-1-induced NF-B Activation and Pro-IL-1␤ Maturation Can Occur Simultaneously-Our results indicated that pro-IL-1␤ maturation was dependent on caspase-1 enzymatic activity, whereas NF-B activation was independent of it. However, stimulation of macrophages with LPS led to a strong induction of both NF-B activation and maturation of pro-IL-1␤ (data not shown). This led us to examine whether or not caspase-1-dependent pro-IL-1␤ processing and NF-B activation were mutually exclusive. Assuming that the strong proapoptotic activity of wild-type caspase-1 masked the read-out in our NF-B reporter system, we used caspase-1 2D/E, a less cytotoxic variant of wild-type caspase-1 in which the cleavage sites between the prodomain and the p20 domain are mutated. Co-expression of caspase-1 2D/E and pro-IL-1␤ demonstrated that pro-IL-1␤ maturation and NF-B activation could occur together (Fig. 2D). These results demonstrated that the induction of NF-B activity was independent of the enzymatic activity of the caspase and thus of the maturation of pro-IL-1␤. Moreover, these results suggest that caspase-1-mediated NF-B activation and pro-IL-1␤ maturation can occur simultaneously.

Specificity of Caspase-1-induced NF-B Activation-Murine
caspase-1 clusters phylogenetically with caspase-11 and caspase-12 (36). Therefore, we tested whether caspase-11 and -12 were also capable of inducing NF-B activation. Overexpression of wild-type caspase-1 quickly led to apoptotic cell death of the transfected cells (Fig. 1A). As a result, the protein was hard to detect in cytosolic lysates by Western blotting analysis (Fig. 3) but was present in the supernatant of dying cells (15,37). Apoptotic cell death was much less extensive when caspase-11 or -12 was overexpressed (data not shown). This is in accordance with Western blotting analysis showing clear expression of these proteins in cell lysates. In the case of caspase-11, a 27-kDa fragment characteristic of its activation (38) was detected. However, neither caspase-11 nor caspase-12 was capable of activating NF-B (Fig. 3). To exclude the possibility of cell death interfering with the NF-B assays, we also analyzed the enzymatically inactive C/A mutants of caspase-1, -11, and -12. In contrast to the wild-type caspases, none of the C/A mutants induced apoptosis (data not shown). However, caspase-11 C254A and caspase-12 C298A were still completely unable to activate NF-B, whereas caspase-1 C284A strongly induced the expression of the NF-B-dependent reporter (Fig. 3).
Caspase-1-induced NF-B Activation Required IKK-␤ and Was Independent of TRAF2 and RIP1-We next analyzed the signaling pathway involved in caspase-1-induced NF-B activation, using several inhibitors and dominant-negative (DN) molecules of key proteins involved in known NF-B signaling pathways. Because most NF-B signaling pathways converge at the IKK-complex, we used a kinase-dead mutant (K44A) of IKK-␤ that functions as a dominant-negative inhibitor (IKK-␤ DN), to analyze whether caspase-1 signals through this central NF-B activating complex. Low levels of IKK-␤ DN completely abolished both TNF-and caspase-1-induced NF-B activation, suggesting that the IKK-complex is a central downstream mediator of caspase-1-induced NF-B activation (Fig. 4A). A20 is an inhibitor of several NF-B-activating pathways, including those induced by LPS (39,40), TNF (6,40), and IL-1 (41,42). A20 blocked both TNF-and caspase-1-induced activation of NF-B, suggesting that it is a downstream inhibitor of both pathways (Fig. 4A). However, the exact molecular pathway used by A20 remains poorly characterized. RIP1 and TRAF2 are central mediators of TNF-and CD40-induced NF-B acti-

FIG. 3. Specificity of caspase-1-induced NF-B activation.
293T cells were transiently cotransfected with an NF-B-dependent luciferase reporter and 0.4 g of the indicated plasmids. 24 h after transfection, lysates were analyzed for NF-B activation as described under "Experimental Procedures." Because of the death of transfected cells, activation of NF-B by caspase-1 could not be measured. Wild-type caspase-11 and -12 are much less cytotoxic (data not shown). Caspase-1 C/A leads to a dose-dependent induction of NF-B activation, but caspase-11 C/A and caspase-12 C/A do not. *, wild-type caspase-1 leads to excessive cell death and leakage of caspase-1 into the supernatant.
Caspase-1-induced NF-B Activation Was RIP2-dependent-RIP2 is the central mediator of NF-B activation originating from TLR-2, -3, and -4, and from the intracellular receptors NOD1 and NOD2 (22,23). TLR-4 signaling is known to induce the inflammasome, a caspase-1 containing signaling complex (15). Furthermore, RIP2 is known to physically interact with caspase-1 through a CARD-CARD interaction in vitro at concentrations that correspond to the endogenous levels found in cells (27,31). It has been suggested that RIP2 is an upstream activator of caspase-1, leading to enhanced pro-IL-1␤ maturation (27,31). However, no defects in pro-IL-1␤ maturation can be observed in RIP2-deficient macrophages (23). Therefore, we hypothesized that RIP2 might instead be a downstream effector of caspase-1-induced NF-B activation. A deletion mutant of RIP2 lacking its kinase domain (RIP2 DN) functioned as a dominant-negative molecule on RIP2-induced activation of NF-B (Fig. 4B). This RIP2 DN inhibited caspase-1-mediated NF-B activation to a similar extent (Fig. 4B) but did not block TNF-induced NF-B activation (Fig. 4B). Co-immunoprecipitation experiments further confirmed that full-length caspase-1 physically interacted with RIP2 and with the isolated CARD of RIP2 (Fig. 4C). Taken together, these results suggest a downstream role for RIP2 in caspase-1-induced NF-B activation.
Caspase-1 CARD Was Necessary and Sufficient for Caspase-1-induced NF-B Activation-To determine the part of caspase-1 responsible for the activation of NF-B, we generated plasmids encoding different domains of caspase-1 and compared their NF-B-inducing abilities with that of full-length caspase-1 C284A (Fig. 5A). Among these, only the caspase-1 fragments containing the CARD were capable of activating NF-B (Fig. 5A). Moreover, caspase-1 CARD on its own induced NF-B activity to the same extent as caspase-1 C284A, suggesting that the CARD was sufficient for NF-B activation. Neither the p20 domain, the p10 domain, nor the collinear combination of both was able to induce NF-B activity (Fig.  5A). Although the fragment containing the CARD and the p20 domain induced NF-B-activation, the extent of activation was lower than that of the CARD alone (Fig. 5A). This suggests that in the absence of the p10 domain, the p20 domain interferes with the NF-B-inducing activity of the caspase-1 CARD. Western blotting analysis confirmed the appropriate expression of all domains and their combinations (Fig. 5B). All together, these data suggest that the caspase-1 CARD is sufficient and necessary for caspase-1-induced NF-B activation.
Because we observed that caspase-1 CARD was necessary and sufficient for NF-B activation, we were interested in comparing the NF-B-activating ability of human caspase-1C285A with that of COP/Pseudo-ICE and ICEBERG, two human-specific CARD-only proteins that share 93% and 73% sequence homology, respectively, with the prodomain of caspase-1 (25)(26)(27). It is noteworthy that human caspase-1 C285A and COP/Pseudo-ICE were comparable in their NF-Binducing capacities (Fig. 5C, top). However, ICEBERG had very little effect on the basal NF-B activity, again demonstrating that caspase-1 C285A-and COP/Pseudo-ICE-induced NF-B activity was not provoked by a nonspecific cellular FIG. 4. Mechanism of caspase-1-induced NF-B activation. A, 293T cells were transiently cotransfected with an NF-B-dependent luciferase reporter, caspase-1 C/A, and plasmids encoding A20 or dominant-negative molecules of either TRAF2, RIP1, or IKK-␤. As a control, plasmids were treated with 500 IU/ml human TNF for induction of NF-B activation. 24 h after transfection, lysates were analyzed for NF-B activation as described under "Experimental Procedures." B, 293T cells were transiently cotransfected with an NF-B-dependent luciferase reporter and plasmids encoding RIP2 or caspase-1 C/A in the presence or absence of RIP2 DN. As a control, cells were treated with 500 IU/ml human TNF for induction of NF-B activation. 24 h after transfection, lysates were analyzed for NF-B activation as described under "Experimental Procedures." C, co-immunoprecipitation assays were performed using lysates (normalized for total protein content) from 293T cells that had been transiently transfected with plasmids encoding E-epitope tagged caspase-1 C285A and vesicular stomatitis virus (VSV)-tagged RIP2 or RIP2 CARD. Immunoprecipitates were prepared using anti-vesicular stomatitis virus antibody adsorbed to protein G-Sepharose and analyzed by SDS-PAGE/immunoblotting using anti-E-HRP antibody with enhanced chemiluminescence-based detection. Aliquots of the same lysates were also analyzed directly by SDS-PAGE/immunoblotting, as indicated. IP, immunoprecipitation; WB, Western blotting.

Caspase-1 Activates p38 MAPK and NF-B
stress mediated by transient expression of these proteins. Western blot analysis confirmed the appropriate expression of all proteins (Fig. 5C, bottom). These results suggest that the NF-B-activating capacity of caspase-1 CARD is evolutionarily conserved in COP/Pseudo-ICE but is lost in the more distant ICEBERG.
Caspase-1 CARD-induced NF-B Activation Was Independent of p38 MAPK Activation-Beside its role in NF-B signaling, RIP2 leads to the phosphorylation of p38 MAPK (22,23). However, its kinase activity is dispensable for the phosphorylation of MAPKs (31), suggesting the involvement of at least one additional kinase. Because RIP2 is involved in caspase-1-mediated NF-B signaling (Fig. 4B), we tested whether caspase-1 CARD was also capable of inducing p38 MAPK phosphorylation. Although both caspase-1 CARD and green fluorescent protein-overexpressing cells contained similar basal levels of p38 MAPK, a strong induction of p38 MAPK phosphorylation was observed only in the caspase-1 CARD-expressing cells, and no active p38 MAPK was detected in control green fluorescent protein-transfected cells (Fig. 6A). This indicates that the phosphorylation of p38 MAPK by caspase-1 CARD was not caused by a general stress induced by the transfection procedure or by other nonspecific stress factors. Given that both p38 MAPK and NF-〉 were activated in cells overexpressing caspase-1 CARD (Fig. 6), and p38 MAPK activation has been implicated in NF-B activation (50), we next analyzed whether the activation of p38 MAPK was involved in the induction of NF-B activity. A recent report elegantly demonstrated that the widely used synthetic p38 MAPK inhibitor SB203508 also potently inhibits RIP2 kinase activity (51). However, because RIP2 kinase activity is dispensable for MAPK activation (31), we made use of this inhibitor. We show that the levels of phosphorylated p38 MAPK were significantly reduced in SB203508-treated cells (Fig. 6B). However, in the same cells, the levels of NF-B activation remained unchanged (Fig. 6B), suggesting that NF-B activation by caspase-1 was independent of p38 MAPK activation and RIP2 kinase activity. DISCUSSION Modular protein-protein interaction domains play an important role in many intracellular signal transduction pathways (52). In apoptosis and inflammation signaling pathways, four major families of highly related protein modules have been identified: the death domain, the death effector domain, the CARD, and the recently identified PYRIN domain (reviewed in Refs. 53,54). These protein modules of ϳ100 amino acids mediate homotypic protein-protein interactions between signaling components, which leads to the activation of downstream targets in response to stress and developmental stimuli. A large body of in vitro data suggests that some of the proteins containing these interaction domains, particularly CARD and PYRIN, are involved in NF-B activation (reviewed in Refs. 16, 55, and 56). The crucial role of CARDs is supported by recent in vivo data. Mice deficient in the CARD-containing Bcl10, the CARD of CARMA1, or the death domain-containing paracaspase (MALT1) are defective in lymphocyte activation and FIG. 5. Caspase-1 CARD is necessary and sufficient for NF-B activation. A, 293T cells were cotransfected with an NF-B-dependent luciferase reporter and either empty vector, fulllength caspase-1 C284A or the indicated deletion constructs. 24 h later, lysates were analyzed for NF-B activation as described under "Experimental Procedures." B, aliquots of the same whole cell lysates were analyzed by SDS-PAGE/immunoblotting using anti-E antibodies to confirm the appropriate expression of all constructs. C, 293T cells were cotransfected with an NF-B-dependent luciferase reporter and the indicated amounts of plasmids encoding full-length caspase-1 C285A, COP/Pseudo-ICE or ICEBERG. Total DNA content of 0.5 g was maintained by the addition of the control empty vector. 24 h later, lysates were analyzed for NF-B activation as described under "Experimental Procedures" (top). Aliquots of the same lysates were also analyzed by SDS-PAGE/immunoblotting using anti-E and anti-T7 epitope tag antibodies to confirm the appropriate expression of all constructs (bottom). proliferation because of abrogated NF-B activation after antigen-receptor stimulation (57)(58)(59). There is increasing evidence that the death effector domain-containing caspases (caspase-8 and -10) and death effector domain-containing adaptors also have important roles in processes beyond cell death, such as cell cycle regulation (60), proliferation (61)(62)(63)(64), and differentiation (65,66). It is interesting that the death effector domains of both caspase-8 and -10, as well as those of FADD and FLIP, have been shown to signal to NF-B and MAPK activation independently of caspase protease activity (67,68). In this study, we showed that caspase-1 as well was capable of inducing NF-B and p38 MAPK activation independently of its enzymatic role in cytokine maturation, and that the CARD of caspase-1 was necessary and sufficient for these functions (Fig. 7).
Although caspase-1 induces cell death when overexpressed in cell lines, the role of this enzyme in cell death has been restricted to ischemic death of neurons (69) and to macrophages infected with specific invasive pathogens such as Salmonella typhimurium and Shigella flexneri (70,71). It is widely accepted that caspase-1 is essential for the maturation of the pro-inflammatory cytokines pro-IL-1␤ and pro-IL-18 in LPSactivated monocytes, because their mature forms are completely absent from monocytes derived from caspase-1-deficient mice (13,72). After maturation, these cytokines are released or secreted using mechanisms that have not yet been identified. An autocrine or paracrine action of IL-1␤ may be one way leading to the activation of NF-B. However, our results demonstrate that caspase-1-dependent activation of NF-B occurs even in the absence of mature IL-1␤ (Fig. 2). In this respect, caspase-1 may add to the overall NF-B activation in stimulated macrophages, resulting in a pleiotropic activation of NF-B. The suggestion that caspase-1 plays a role in the expression of NF-B-dependent gene products is supported by the results of experiments using monocytes derived from caspase-1-deficient mice (72). Caspase-1-deficient monocytes produce about half the amount of TNF-␣ and IL-6 in response to LPS stimulation compared with wild type monocytes (72). Furthermore, the focal cerebral ischemia-induced early activation of NF-B is inhibited in caspase-1-deficient mice (73). However, because both the CARD and the catalytic domain can induce NF-B activation, either directly, as shown in this report, or via the production of pro-inflammatory cytokines, it is difficult to assess the contribution of each of these pathways in caspase-1deficient mice.
Although the cells in which caspase-1-induced NF-B activation is crucial remain to be identified, it is conceivable that in monocytes and macrophages, major cell types that express caspase-1 and generate pro-IL-1␤, the CARD-dependent activation of NF-B may be related to the induction of survival mechanisms protecting them against cytotoxic proteases such as caspases. In this respect, it is remarkable that NF-B reg-ulates the expression of such caspase inhibitors as c-FLIP (74,75), c-IAP1, cIAP-2 (76 ,77), and XIAP (78). The survival of macrophages is required because they form part of the innate immune system, killing and phagocytosing bacteria and presenting antigenic bacterial peptides that engage the antigenspecific T cell response (79). Exposure of macrophages to bacteria or LPS mediates activation of death-inducing and -preventing pathways in macrophages at the same time (80). FIG. 7. Model of caspase-1 signaling pathways. Stimulation of Toll-like receptor (TLR) signals via an unknown pathway to the formation of the inflammasome, a large caspase-1-containing protein complex. Caspase-1 enzymatic activity is responsible for the proteolytic maturation of pro-IL-1␤, because this activity is inhibited by the cowpox-derived serpin CrmA. After maturation, IL-1␤ is somehow released or secreted. Caspase-1 interacts with RIP2 by a heterotypic CARD-CARD interaction, leading to the association of RIP2 with IKK-␥/NEMO. This interaction activates the IKK complex by the phosphorylation of IKK-␣ and IKK-␤. The IKK-complex phosphorylates IB, targeting it for degradation. Released NF-B thus translocates to the nucleus, where it induces gene transcription. A second pathway downstream of RIP2 recruitment starts with the phosphorylation of p38 MAPK. Although RIP2 is a kinase, RIP2-mediated activation of p38 MAPK occurs in an indirect manner involving an as-yet-unidentified kinase. Activated p38 MAPK leads to enhanced transcription activation.
FIG. 6. NF-B activation is independent of p38 MAPK activation. A, 293T cells were transfected with plasmids encoding either EGFP or caspase-1 CARD. 24 h later, lysates were analyzed by SDS-PAGE/immunoblotting using antibodies against the indicated proteins. B, 293T cells, untreated or pretreated with the p38 MAPK-specific inhibitor SB203508, were cotransfected with caspase-1 CARD and an NF-B-dependent luciferase reporter. 24 h later, lysates were analyzed for NF-B activation as described under "Experimental Procedures." Aliquots of the same lysates were also analyzed by SDS-PAGE/immunoblotting using antibodies against the indicated proteins to confirm the inhibition of p38 MAPK phosphorylation.
Monocytes, as major producers of inflammatory cytokines, are resistant to the cytotoxic activities of LPS, IL-18, TNF, and interferon-␥ by virtue of the activation of survival pathways such as p38 MAPK and NF-B signaling cascades (81)(82)(83). When the anti-apoptotic NF-B pathway is blocked, the cytotoxic proteolytic pathways dominate, and the macrophage undergoes apoptosis (84). Indeed, several pathogenic species, including Yersinia enterocolitica, trigger apoptosis in macrophages by impairing the activation of p38 MAPK and NF-B, showing the relevance of these proteins for survival pathways in activated macrophages (84,85).
COP/Pseudo-ICE and ICEBERG are CARD-only proteins that share a high degree of sequence homology to the prodomain of caspase-1 (25)(26)(27). Because both caspase-1 CARD and COP/Pseudo-ICE, but not ICEBERG, can induce NF-B activation, we conclude that the NF-B-activating capacity of caspase-1 CARD is conserved in COP/Pseudo-ICE but is lost in the evolutionarily more distant ICEBERG. Therefore, it is likely that caspase-1 CARD and COP/Pseudo-ICE share on their surfaces a common NF-B-activating interface that has been lost in ICEBERG. It is interesting that the ability of these CARDs to activate NF-B correlates with their ability to interact with RIP2 (25-27), a protein essential for LPS-, doublestranded RNA-and peptidoglycan-induced activation of NF-B (22,23). Hence, identification on the surfaces of caspase-1 CARD and COP/Pseudo-ICE of critical amino acids involved in RIP2 binding and NF-B activation may represent a first step in the design of drugs that specifically interfere with RIP2-dependent signaling pathways, leading to NF-B activation.
In conclusion, our data indicate that caspase-1 could be involved in inflammation responses at two different levels. One pathway is dependent on caspase-1 enzymatic activity and leads to the processing and activation of the pro-inflammatory mediators IL-1␤ and IL-18. The novel pathway that we describe here is mediated by the CARD and leads to the activation of NF-B and p38 MAPK through a RIP2-dependent mechanism. The upstream signaling pathways and the cellular and physiological context in which this novel function of caspase-1 is operating are the subjects of ongoing research.