TAK1 mediates the ceramide signaling to stress-activated protein kinase/c-Jun N-terminal kinase.

Ceramide has been proposed as a second messenger molecule implicated in a variety of biological processes. It has recently been reported that ceramide activates stress-activated protein kinase (SAPK, also known as c-Jun NH2-terminal kinase JNK), a subfamily member of mitogen-activated protein kinase superfamily molecules and that the ceramide/SAPK/JNK signaling pathway is required for stress-induced apoptosis. However, the molecular mechanism by which ceramide induces SAPK/JNK activation is unknown. Here we show that TAK1, a member of the mitogen-activated protein kinase kinase kinase family, is activated by treatment of cells with agents and stresses that induce an increase in ceramide. Ceramide itself stimulated the kinase activity of TAK1. Expression of a constitutively active form of TAK1 resulted in activation of SAPK/JNK and SEK1/MKK4, a direct activator of SAPK/JNK. Furthermore, expression of a kinase-negative form of TAK1 interfered with the activation of SAPK/JNK induced by ceramide. These results indicate that TAK1 may function as a mediator of ceramide signaling to SAPK/JNK activation.

kinase (MAPKK), which is phosphorylated and activated by a serine/threonine kinase, generally called MAPKK kinase (MAPKKK). Recently, two novel members of MAPK-related enzymes have been identified (4). One subgroup is called stress-activated protein kinase (SAPK) or c-Jun NH 2 -terminal kinase (JNK) (5,6), and the other is p38/MPK2 (7)(8)(9). MAPK, SAPK/JNK, and p38/MPK2 constitute the MAPK superfamily. It is generally thought that each subfamily of the MAPK superfamily is independently regulated by particular sets of extracellular stimuli and therefore may have distinct functions in various biological processes.
We have recently identified a novel member of mammalian MAPKKK, TAK1. Several lines of evidence suggested that TAK1 functions in the signal transduction pathways triggered by members of the transforming growth factor-␤ (TGF-␤) superfamily (10). During the course of further biochemical characterization of TAK1, we have found that TAK1 can be activated in response to several stimuli that have been shown to activate SAPK/JNK. Moreover, we noticed that these stimuli are known to induce an increase in ceramide. Ceramide has recently emerged as a second messenger molecule that induces multiple cellular responses such as cell cycle arrest, differentiation, and apoptosis (11,12). Most recently, it has been reported that ceramide activates SAPK/JNK and the ceramideinduced activation of SAPK/JNK is required for stress-and ceramide-initiated apoptosis (13). This may highlight the role of SAPK/JNK as a mediator of cellular responses induced by ceramide. But the molecular mechanism by which ceramide induces the activation of SAPK/JNK is unknown.
In this report, we first show that TAK1 is activated in cells treated with agents and stresses that induce the generation of ceramide. We then show that ceramide itself can activate TAK1. Furthermore, expression of an active form of TAK1 induced activation of SAPK/JNK and its activator, SEK1/ MKK4. In addition, expression of a kinase-deficient form of TAK1 interfered with the activation of SAPK/JNK induced by ceramide. These results indicate that TAK1 acts as a mediator for the ceramide-induced activation of SAPK/JNK.

MATERIALS AND METHODS
Cell Culture-A673 human rhabdomyosarcoma cells, COS7 cells, and rat 3Y1 cells were cultured at 37°C in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum.
Preparation of Recombinant Proteins-Both kinase-negative MPK2 (KN-MPK2) and wild-type SAPK␣ were expressed as His-tagged proteins and purified using a Ni 2ϩ -affinity column (14). Both wild-type XMEK2/SEK1 and kinase-negative MAPK were expressed as glutathione S-transferase fusion proteins and purified as described (14). Recombinant c-Jun was expressed and purified as described previously (14).
Immunoprecipitation of Endogenous TAK1-The cells were lysed in an extraction buffer (20 mM HEPES, pH 7.4, 150 mM NaCl, 12.5 mM ␤-glycerophosphate, 1.5 mM MgCl 2 , 2 mM EGTA, 10 mM NaF, 2 mM dithiothreitol, 1 mM Na 3 VO 4 , 1 mM phenylmethylsulfonyl fluoride, and * This work was supported by Grants-in-Aid from the Ministry of Education, Science and Culture of Japan (to E. N. and K. M.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Immunoprecipitation of HA-TAK1, HA-MAPK, HA-SAPK/JNK, HA-MAPKK, and HA-SEK1/MKK4 -For immunoprecipitation of HA-TAK1, the cells were homogenized in an extraction buffer. For immunoprecipitation of other HA-tagged kinases, the cells were homogenized in the buffer containing 0.5% Triton X-100. Lysates were then clarified by centrifugation at 15,000 ϫ g for 10 min. The supernatants were incubated with anti-HA monoclonal antibody (12CA5) for 30 min at 4°C. After the addition of protein A-Sepharose beads, the lysates were incubated for an additional 1 h. The immunoprecipitates were washed as above and subjected to in vitro kinase assays.
Protein Kinase Assays-The activity of immunoprecipitated TAK1 was assayed by its ability to activate recombinant XMEK2/SEK1, whose activity was assayed by its ability to phosphorylate recombinant KN-MPK2. The immunoprecipitate was incubated with recombinant XMEK2/SEK1 (0.5 mg/ml) in a solution (22 l) containing 20 mM Tris-Cl, pH 7.5, 10 mM MgCl 2 , and 100 M ATP for 20 min at 30°C. Then, KN-MPK2 (5 mg/ml, 3 l) and [␥-32 P]ATP (5 Ci) was added, and the mixtures were incubated further for 10 min at 30°C. The reaction was stopped by the addition of Laemmli's sample buffer and boiling. After electrophoresis, incorporation of phosphate into KN-MPK2 was detected by autoradiography and quantified by using an image analyzer (Fujix BAS 2000). The activities of immunoprecipitated SAPK, SEK1, MAPK, and MAPKK were assayed by their ability to phosphorylate recombinant c-Jun, recombinant SAPK␣, myelin basic protein, and recombinant kinase-negative form of MAPK (KN-MAPK), respectively.

RESULTS AND DISCUSSION
A673 human rhabdomyosarcoma cells were stimulated with TNF-␣, IL-1, or anti-Fas antibody, and the immunoprecipitate with anti-TAK1 antibody was subjected to an in vitro coupled kinase assay. This antibody specifically immunoprecipitated endogenous TAK1 (Fig. 1A). TAK1 activity increased in cells exposed to these agents, peaked after 5-10 min, and declined gradually with time (Fig. 1B). We next examined whether TAK1 could be activated by treatment of cells with stresses. An expression construct encoding epitope-tagged TAK1 (HA-TAK1) was transiently transfected into COS7 cells. After treatment of the cells with UV irradiation or sorbitol, TAK1 activity was determined in a coupled kinase assay. TAK1 activity increased within 5 min of exposure of cells to UV irradiation and peaked at 10 min (Fig. 1C). Exposure of the cells to hyperosmolarity with 0.5 M sorbitol also caused an increase in TAK1 activity (Fig. 1C). TAK1 activity was maximal after 30 min, and the high level of activity was sustained for 60 min (Fig. 1C).
Because TNF-␣ (17, 18), IL-1 (19,20), anti-Fas antibody (21,22), and several stresses (13) were all known to induce an increase in ceramide, we tested whether TAK1 could be activated by treatment of cells with ceramide itself. COS7 cells were transfected with HA-TAK1 and treated with the mem-FIG. 1. Activation of TAK1 in cells exposed to cytokines and stresses. A, immunoprecipitation was performed with anti-TAK1 antiserum from A673 human rhabdomyosarcoma cell extracts. The immunoprecipitate was immunoblotted with the anti-TAK1 antibody. B, A673 cells were treated with TNF-␣ (100 ng/ml), IL-1 (12.5 ng/ml), or anti-Fas antibody (500 ng/ml). After the indicated times, endogenous TAK1 was immunoprecipitated with anti-TAK1 antibody, and the kinase activity was determined in a coupled kinase assay as described under "Materials and Methods." The radioactivity of the KN-MPK2 bands was detected by autoradiography (lower panels) and quantified using an image analyzer (Fujix BAS2000) (results shown in upper panels). C, COS7 cells were transiently transfected with an HA-TAK1 expression plasmid and exposed to UV irradiation (120 J/m 2 ) or sorbitol (0.5 M). HA-TAK1 was immunoprecipitated with anti-HA monoclonal antibody (12CA5) and assayed as described under "Materials and Methods." The radioactivity of the KN-MPK2 bands was detected by autoradiography (lower panels) and quantified using an image analyzer (Fujix BAS2000) (results shown in upper panels). Immunoblot analysis of each immunoprecipitate with 12CA5 demonstrated that approximately the same amount of HA-TAK1 was recovered at each time point (data not shown).

TAK1 Mediates the Ceramide Signaling to SAPK/JNK
brane-permeable ceramide analogue, C 2 -ceramide. HA-TAK1 was isolated by immunoprecipitation, and its activity was determined. TAK1 activity increased within 5 min of exposure of cells to C 2 -ceramide, was maximal after 10 -30 min, and declined with time (Fig. 2B). C 2 -ceramide stimulated TAK1 activity in a dose-dependent manner (Fig. 2C). This activation was not observed in immune complexes from cells transfected with an HA epitope-tagged, kinase-negative form of TAK1 ( Fig.  2A). The endogenous TAK1 was also activated by treatment with C 2 -ceramide (Fig. 2D). Taken together, these results have demonstrated that ceramide is capable of inducing TAK1 activation.
We tested the effect of other lipids on the TAK1 activity. C 2 -dihydroceramide, which lacks the trans double bond at C-4 -C-5 of the sphingoid base backbone, and 1,2-dioctanoyl-snglycerol, which is an activator of protein kinase C, failed to activate TAK1 activity efficiently (Fig. 2E). Arachidonic acid also failed to activate TAK1 (Fig. 2E). Thus, C 2 -ceramide specifically activates TAK1.
TAK1 was originally identified as a mediator of the signal transductions triggered by the TGF-␤ superfamily (10). Most recently, one of TAK1-binding proteins, TAB1, has been identified as an activator for TAK1 (16). A truncated form of TAB1 lacking the TAK1-binding domain (TAB1 (1-418)) was shown to act as a dominant-negative inhibitor of TGF-␤-induced gene expression (16). Overexpression of TAB1 (1-418) almost completely suppressed the activation of TAK1 induced by TGF-␤ stimulation but had little effect on the ceramide-induced activation of TAK1 (data not shown). These results may indicate that ceramide and TGF-␤ activate TAK1 through TAB1-independent and -dependent pathways, respectively.
A recent report of Verheij et al. (13) suggested the central role of SAPK/JNK in the ceramide-induced signal transduction. We therefore examined whether TAK1 participates in SAPK/ JNK activation induced by ceramide. We have previously shown that a truncated form of TAK1 lacking NH 2 -terminal 20 amino acids (TAK1⌬N) is constitutively active (10). The expression vector encoding TAK1⌬N was co-transfected into COS7 cells with an expression vector encoding HA-tagged SAPK/JNK or HA-tagged MAPK, and the activity of SAPK/JNK or MAPK was determined in the immunoprecipitate obtained with anti-HA antibody. Co-expression of TAK1⌬N induced activation of SAPK/JNK (Fig. 3A) but not that of MAPK (Fig. 3B). Coexpression of TAK1⌬N with SEK1/MKK4 or MAPKK, direct activators of SAPK/JNK and MAPK, respectively, resulted in activation of SEK1/MKK4 (Fig. 3C) but not that of MAPKK (Fig. 3D). These results demonstrated that TAK1 is able to activate the SEK1/MKK4 3 SAPK/JNK cascade but unable to activate the MAPKK 3 MAPK cascade. Furthermore, when a kinase-negative form of TAK1 was co-expressed, the C 2 -ceramide-induced activation of SAPK/JNK was suppressed markedly (Fig. 3E), indicating that TAK1 is required for SAPK/JNK activation induced by ceramide.
In this study we have shown that TAK1, a recently identified MAPKKK family molecule, is activated in cells treated with cytokines and stresses that are known to induce the generation of ceramide (13,(17)(18)(19)(20)(21)(22). We have here further shown that ceramide itself is able to activate TAK1 (Fig. 2). TAK1 activation was not observed in cells treated with other types of lipids, including diacylglycerol and arachidonic acid (Fig. 2). These results thus suggest that TAK1 may participate in the signal transduction pathway initiated by ceramide.
In the present study, expression of a constitutively active mutant of TAK1 resulted in the activation of SAPK/JNK and its direct activator SEK1/MKK4 (Fig. 3, A and C). Expression of a kinase-negative mutant of TAK1 suppressed the activation of SAPK/JNK induced by ceramide (Fig. 3E). Therefore, TAK1 is necessary and sufficient for the activation of SAPK/JNK induced by ceramide. Although Raf-1, another member of MAP-KKK family, has been shown to be a mediator of ceramide signaling to classical MAPK (23)(24)(25), a molecular link between ceramide and SAPK/JNK has been missing. Our present results have identified TAK1 as a crucial mediator of ceramide signaling to SAPK/JNK.
The mechanism of TAK1 activation in response to ceramide should be elucidated. Several molecules have previously been reported as direct targets for ceramide, including ceramideactivated protein kinase (26,27), ceramide-activated protein phosphatase (28), protein kinase C- (29), and Raf-1 (30). These molecules might lie upstream of TAK1.