TAJ, a novel member of the tumor necrosis factor receptor family, activates the c-Jun N-terminal kinase pathway and mediates caspase-independent cell death.

We have isolated a novel member of the TNFR family, designated TAJ, that is highly expressed during embryonic development. TAJ possesses a unique cytoplasmic domain with no sequence homology to the previously characterized members of the TNFR family. TAJ interacts with the TRAF family members and activates the JNK pathway when overexpressed in mammalian cells. Although it lacks a death domain, TAJ is capable of inducing apoptosis by a caspase-independent mechanism. Based on its unique expression profile and signaling properties, TAJ may play an essential role in embryonic development.

Members of the TNFR 1 superfamily play an important role in regulating diverse biological activities, such as cell proliferation, differentiation, and programmed cell death or apoptosis. The majority of TNFR family members are type I membrane proteins that share significant sequence homology in their extracellular domains (1). This homology is due to the presence of highly conserved cysteine residues in so-called cysteine-rich pseudo-repeats, a hallmark of this family. The cytoplasmic domains of the various TNFR family members, on the other hand, differ greatly not only in their sequence but also in their length. Some of the apoptosis-inducing members of this family, such as TNFR1, CD95/Fas/APO-1, DR3/TRAMP/APO-3, DR4/ TRAIL/APO-2, and DR5/TRAIL-R2, contain a conserved domain of approximately 80 amino acids, called a death domain, which is essential for mediation of cell death (2)(3)(4). Although death domains are absent in other members of the TNFR family, some of these non-death domain-containing receptors, such as TNFR2, CD30, and LT-␤R, are nevertheless capable of inducing cell death under certain circumstances (5)(6)(7).
In the present report, we describe the isolation and characterization of a novel member of the TNFR family, designated TAJ (for Toxicity And JNK inducer). We present evidence that, based on its unique expression profile and signaling activities, TAJ may be a key regulator of cell activation and death during embryonic development.

EXPERIMENTAL PROCEDURES
Cell Lines and Reagents-293T cells were obtained from Dr. David Han (University of Washington, Seattle). 293EBNA cells were obtained from Invitrogen. Rabbit polyclonal antibodies against FLAG and HA tags were obtained from Santa Cruz Laboratories. FLAG beads were obtained from Sigma. The pull-down kinase assay kit for JNK was obtained from New England Biolabs, and the constructs for the Pathdetect luciferase reporter assay were purchased from Stratagene. YO-PRO-1 was obtained from Molecular Probes.
Cloning of TAJ cDNA-Two murine EST clones (IMAGE consortium clones 650744 and 664665) encoding the extracellular domain of a new member of the TNFR superfamily were identified by searching the data base of expressed sequence tags (dbEST) for sequences sharing homology to the extracellular domain of human DR3. The sequence corresponding to the cytoplasmic domain of this receptor was obtained by using 3Ј-rapid amplification of cDNA ends (RACE) on murine spleen Marathon Ready cDNA (CLONTECH) using the Marathon cDNA amplification kit (CLONTECH) and following the manufacturer's instructions. A repeat search of the EST data base for sequences with homology to the cytoplasmic tail of murine TAJ revealed the presence of a human EST clone (IMAGE consortium clone 340844). The sequence corresponding to the 5Ј end of human TAJ (hTAJ) was obtained by performing 5Ј-RACE on human fetal spleen Marathon-ready cDNA (CLONTECH).
Northern Blot Analysis-Northern blot analysis was performed using mouse embryo and human multiple tissue Northern blots (CLON-TECH). Blots were hybridized under high stringency conditions with 32 P-labeled probes corresponding to the protein-coding regions of mTAJ and hTAJ and following the instructions of the manufacturer.
Expression Vectors-To construct FLAG or Myc epitope-tagged receptors, the amino acids 23-424 of hTAJ were amplified using Pfu polymerase (Stratagene) with a primer containing a BglII site at the 5Ј end and a SalI site at the 3Ј end and then were ligated to a modified pSectag A vector (Invitrogen) containing a FLAG or a Myc epitope downstream of a murine Ig chain signal peptide. Expression constructs for dominant-negative FADD, caspase 8 C360S, CrmA, p35, MRIT, Orf-K13, and TRAFs have been described previously (8 -10).
Reporter Assays-For the c-Jun transcriptional activation assay, 293 EBNA cells (1.2 ϫ 10 5 ) were transfected in duplicate with various expression constructs (500 ng) along with a fusion transactivator plasmid containing the yeast GAL4 DNA-binding domain fused to transcription factor c-Jun (pFA-cJun) (50 ng), a reporter plasmid encoding the luciferase gene downstream of the GAL4 upstream activating sequence (pFR-luc) (500 ng), as well as a RSV/LacZ (␤-galactosidase) reporter construct (75 ng). Transfection was performed using calcium phosphate coprecipitation method. Forty hours later cell extracts were prepared using the Luciferase Cell Culture Lysis Reagent (Promega, Madison, WI), and luciferase assays were performed using 20 l of cell extract. The cell lysate was diluted 1:20 with phosphate-buffered saline, pH 7.4, and used for the measurement of ␤-galactosidase activity. Luciferase activity was normalized relative to the ␤-galactosidase activity to control for the difference in the transfection efficiency. Western * 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.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AF167552-AF167555. blot analysis on cell extracts prepared from similarly transfected cells was used to confirm that the various expression constructs lead to the production of the desired proteins. The NF-B reporter assay was performed essentially as described above using a luciferase reporter plasmid containing four copies of a consensus NF-B binding site (11).
Caspase Activation Assay-293T cells (2 ϫ 10 5 ) were transfected with 2 g of an empty vector or expression vectors encoding TNFR1 or hTAJ along with a GFP-encoding plasmid. Cells were examined under a fluorescent microscope 32 hours later to ensure equal transfection efficiency. Cells were subsequently lysed in 100 l of lysis buffer (0.1% Triton X-100, 20 mM sodium phosphate, pH 7.4, 150 mM NaCl). For assaying caspase 3 activation, 10 l of cell extract was mixed with 80 l of assay buffer (50 mM Hepes, pH 7.4, 100 mM NaCl, 0.1% CHAPS, 10 mM dithiothreitol, 1 mM EDTA, 10% glycerol) in a 96-well microtiter plate in triplicate, and the was reaction started by the addition of 10 l of Ac-DEVD-pNA substrate. Caspase 3-mediated cleavage of Ac-DEVD-pNA into p-nitroanilide was measured using a plate reader at 405 nm.
DNA Content Analysis-293T cells (2 ϫ 10 6 ) were transfected with the various test plasmids along with a GFP-encoding plasmid. Approximately 32 h post-transfection, cells were examined under a fluorescent microscope to ensure equal transfection efficiency. Cells were subsequently trypsinized, washed once with PBS, and fixed in 50% ethanol at 4°C. After washing with PBS, cells were treated with 500 g of RNase A (Sigma) in 100 l of PBS for 30 min at 37°C and resuspended in 0.5 ml of PBS containing 50 g/ml propidium iodide. After further incubation at 4°C for 15 min, cells were analyzed by flow cytometry.
Coimmunoprecipitation Assays-For studying in vivo interaction, 2 ϫ 10 6 293T cells were plated in a 100-mm plate and cotransfected 18 -24 h later with 5 g/plate of each epitope-tagged construct along with 1 g of a hemagglutinin-tagged GFP-encoding plasmid (HA-GFP) by calcium phosphate coprecipitation. Eighteen to thirty-six hours post- transfection, cells were lysed in 1 ml of lysis buffer containing 0.1% Triton X-100, 20 mM sodium phosphate, pH 7.4, 150 mM NaCl, and 1 EDTA-free mini-protease inhibitor tablet per 10 ml (Roche Molecular Biochemicals). For immunoprecipitation, cell lysates (500 l) were incubated for 1 h at 4°C with 10 l of FLAG or control mouse Ig beads precoated with 2% bovine serum albumin. Beads were washed twice with lysis buffer, twice with a wash buffer containing 0.1% Triton X-100, 20 mM sodium phosphate, pH 7.4, 500 mM NaCl, and again with lysis buffer. Bound proteins were eluted by boiling, separated by SDSpolyacrylamide gel electrophoresis, transferred to a nitrocellulose membrane, and analyzed by Western blot.

Cloning of Murine and Human TAJ cDNAs-In order to
identify new members of the TNFR family, we searched the EST data base (dbEST) for sequences with homology to the extracellular domain of Death Receptor 3 (DR3) and identified two mouse cDNA clones. Both clones originated from cDNA libraries made from 13.5-to 14.5-day mouse embryos and, upon sequencing, were found to represent the alternatively spliced forms of the same gene. One of the clones, named mTAJ-␣S, was found to encode an open reading frame of 214 amino acids (Fig. 1A). Sequence analysis of this clone revealed the presence of an N-terminal signal peptide (amino acids 1-23), cysteinerich pseudo-repeats with significant sequence homology to the extracellular domain of other members of the TNFR family (20 -25% sequence identity and 35-46% sequence similarity), a hydrophobic stretch of amino acids representing the transmembrane region (amino acids 171-193), and a short cytoplasmic tail (Fig. 1A). Based on the presence of the short cytoplas-mic tail, this clone may encode for a decoy receptor. The sequence of the second murine clone, named mTAJ-␤, was identical to mTAJ-␣S in the N-terminal 149 amino acids. This region contains its signal peptide, and the cysteine-rich pseudorepeats representing the majority of the ligand-binding domain (Fig. 1A). However, mTAJ-␤ has a stop codon after amino acid 150 and, therefore, may represent a soluble receptor lacking a transmembrane domain (Fig. 1A). Based on the sequence of mTAJ-␣S, primers were designed and used in 3Ј-RACE to clone the sequence representing the cytoplasmic tail of TAJ. The full-length cDNA clone, named mTAJ-␣L, was found to contain an open reading frame of 416 amino acids with a unique cytoplasmic domain having no significant homology to any other member of the TNFR family (Fig. 1A).
A repeat search of the EST data base for clones homologous to the cytoplasmic tail of mTAJ led to the identification of a human EST clone derived from an embryonic heart library constructed from a 19-week-old embryo. Based on the sequence of this clone, primers were designed, and 5Ј-RACE was used to clone the full-length human TAJ (hTAJ) clone. The full-length hTAJ clone encoded a protein of 423 amino acids having 68.4% amino acid identity and 79.2% amino acid similarity with mTAJ-␣L (Fig. 1A). In addition, hTAJ was found to have significant sequence similarity to other members of the TNFR family in its extracellular ligand-binding domain (Fig. 1B).
Expression of TAJ-As discussed above, almost all the EST clones encoding TAJ are derived from cDNA libraries originating from embryonic tissues. To test further the expression of D, cell death induced by TAJ is not accompanied by oligonucleosomal DNA fragmentation. 293T cells (2 ϫ 10 6 ) were transfected with the indicated plasmids (5 g), and DNA fragmentation assay was performed after 38 h essentially as described (48). E, DNA content frequency distribution. Cell death induced by TNFR1 is accompanied by an increase in the hypodiploid cell population (Ͻ2n) that is absent in the TAJ-transfected cells. The percentage of cells in various stages of cell cycle is also shown. DNA content analysis was performed using flow cytometry on ethanol-permeabilized cells stained with propidium iodide. The transfection efficiency of the tested plasmids, as determined by the expression of cotransfected GFP, ranged from 40 to 50%. TAJ during embryonic development, Northern analysis was performed on a multiple tissue Northern blot containing poly(A) RNA from 7-, 11-, 15-, and 17-day mouse embryos. The protein-coding region of mTAJ-␣L cDNA was used as a probe. A strong signal of 4.4 kilobase pairs was detected in lanes containing RNA from day 11, 15, and 17 embryos indicating the expression of TAJ during embryonic development (Fig. 2).
Expression of TAJ in adult human tissues was studied by Northern analysis using the protein-coding region of hTAJ cDNA as a probe. Major expression was seen only in the prostate gland with only very low expression seen in other tissues such as spleen, thymus, testis, uterus, small intestine, colon, and peripheral blood leukocytes (Fig. 2). Expression of TAJ was also detected in 293 (human embryonic kidney) and LNCaP (prostate cancer) cell lines (data not shown).
TAJ Activates the JNK Pathway-Activation of the JNK and the NF-B pathways is a common feature of the TNFR family members. Previous studies have also shown that members of this family can be activated in a ligand-independent fashion upon transient transfection-based overexpression (8,(12)(13)(14). Therefore, the ability of TAJ to activate the JNK pathway upon transient overexpression in 293EBNA cells was tested using a luciferase-based c-Jun transcriptional assay. In this assay, luciferase expression is driven by JNK-mediated phosphorylation of the activation domain of the transcription factor c-Jun that is fused to the GAL4 DNA-binding domain. TAJ could strongly activate the JNK pathway in these cells which was comparable in magnitude to that observed with CD40 (Fig. 3A). The ability of TAJ to activate the JNK pathway was also confirmed by a "pull-down" kinase assay based on in vitro phosphorylation of GST-cJun (Fig. 3B).
TAJ-induced JNK activation was successfully blocked by the JNK-binding domain of JIP1 (15), a recently isolated specific inhibitor of the JNK pathway (Fig. 3C). TRAF2 and its homologs have been shown to play an essential role in the JNK activation by TNFR family members by activating the protein kinase ASK1 (16). However, as shown in Fig. 3D, dominantnegative mutants of TRAF2 and ASK1 failed to block JNK activation via TAJ, while successfully blocking TRAF2-induced JNK activation.
We also investigated the ability of TAJ to activate the NF-B pathway. However, transient transfection of TAJ led to only a weak activation of NF-B in 293EBNA cells and completely failed to activate this pathway in 293T or MCF7 cells (not shown).
TAJ Induces Caspase-independent Cell Death-During the course of investigating the above signaling functions of TAJ, we noticed that it could also induce cell death. This result was somewhat unexpected since TAJ does not possess a death domain. To characterize further the death inducing property of TAJ, 293T cells were transfected with a control plasmid, or plasmids containing TNFR1 or TAJ, along with a reporter plasmid encoding ␤-galactosidase or the green fluorescent protein (GFP). Approximately 32 h post-transfection, TAJ-transfected cells became rounded and started to detach from the plate (Fig. 4, A and B). In addition to 293T cells, TAJ also induced cell death in 293EBNA cells. However, we have so far failed to observe significant TAJ-induced apoptosis in COS cells, which might reflect the tissue or species specificity of this response.
A comparison with TNFR1-transfected cells revealed several unique features of TAJ-induced apoptosis. First, cell death induced by TAJ was slightly delayed as compared with that induced by TNFR1 (32 versus 24 h). Second, whereas TAJtransfected cells showed cytoplasmic swelling and nuclear condensation, they demonstrated relatively little membrane budding as compared with the TNFR1-transfected cells (Fig. 4A). Third, based on nuclear staining with YOPRO-1, a cell-impermeant DNA-intercalating dye, we have recognized differences in the nuclear morphology of cells undergoing apoptosis in response to TAJ or TNFR1. As shown in Fig. 4C, whereas the majority of the vector-transfected cells did not stain with YO-PRO-1, a large number of TAJ or TNFR1-transfected cells showed nuclear staining with this dye reflecting the loss of membrane integrity. However, unlike the TNFR1-transfected cells, those dying in response to TAJ did not demonstrate nuclear fragmentation, a key feature of caspase-induced cell death (Fig. 4C). Finally, as compared with TNFR1-transfected cells, TAJ-transfected cells demonstrated a complete absence of oligonucleosomal DNA fragmentation, one of the essential features of caspase-induced apoptosis (Fig. 4D).
Absence of nuclear fragmentation in the TAJ transfected cells was confirmed by DNA content analysis using flow cytometry. As shown in Fig. 4E, transfection of TNFR1 in 293T cells is accompanied by the appearance of a hypodiploid population of cells, which is absent in cells transfected with an empty vector or TAJ. Furthermore, there was no significant difference in the cell cycle distribution of cells transfected with TAJ or control vector, making it unlikely that TAJ induces cell death via a block in cell cycle transition.
To characterize further TAJ-induced cell death, we tested the ability of several known inhibitors of apoptosis to block cell death induced by it. TAJ-induced cell death was not blocked by two virally encoded caspase inhibitors, CrmA and p35, and a synthetic caspase inhibitor, benzyloxycarbonyl-VAD-fmk, all of which successfully blocked TNFR1-induced apoptosis ( Fig. 5A and data not shown). Similarly, MRIT/cFLIP, Orf-K13, and dominant-negative inhibitors of FADD and caspase 8 (caspase 8 C360S) had no effect on TAJ-induced cell death (Fig. 5B). Finally, JBD-JIP1 failed to effectively block TAJ-induced cell death, ruling out a major role of the JNK pathway in TAJinduced cell death (Fig. 5C).
Lack of activation of the caspase cascade during TAJ-induced cell death was further supported by a chromogenic assay based on caspase 3-mediated cleavage of the chromogenic substrate, pNA-DEVD. Caspase 3 is one of the executioner caspases of the caspase cascade and is activated during death receptors-induced apoptosis (17). As shown in Fig. 5D, cell lysates from cells transfected with TNFR1 demonstrated caspase 3 activation, whereas TAJ-transfected cells failed to do so.
TAJ Interacts with TRAF Family Members-TRAF family members have been previously implicated in the signal transduction by various TNFR family members (18). To test the involvement of TRAFs in the signaling via TAJ, we tested their ability to interact with each other using a coimmunoprecipitation assay in 293T cells. As shown in Fig. 6, TAJ successfully coimmunoprecipitated with TRAF1, TRAF2, TRAF3, and TRAF5 in the above assay. However, consistent with the absence of a death domain in its cytoplasmic tail, TAJ failed to interact with FADD (Fig. 6).

DISCUSSION
Programmed cell death or apoptosis plays an important role in the development and morphogenesis of multicellular organisms by controlling cell number and removing mutated or defective cells (19). Despite the recent progress in the identification of downstream effector molecules involved in developmental cell death (20), the cell surface receptors regulating this process remain to be identified. Members of the TNFR family are well known for their role in the mediation of cell death in adult tissues. Based on its unique expression profile during development, TAJ may play a similar role during embryogenesis.
In addition to its full-length isoform, we have isolated two alternative spliced forms of murine TAJ that are likely to act as decoy and soluble receptors, respectively. Such receptors have been previously identified for other members of the TNFR family and shown to block signaling via the full-length receptors (14,21,22). It remains to be seen whether the mTAJ-␣S and mTAJ-␤ isoforms play a similar role in regulating signaling via the full-length TAJ receptor.
Despite sharing significant sequence homology with TNFR family members in the extracellular ligand-binding domain, TAJ possesses a unique cytoplasmic domain, which suggests that it utilizes novel signal transduction pathways for the activation of JNK and cell death pathways. Consistent with this hypothesis, TAJ-induced JNK activation was not blocked by dominant-negative inhibitors of TRAF2, TRAF5, or ASK1, which have been previously implicated in JNK activation via TNFR1 and CD40. However, coimmunoprecipitation assays revealed that TAJ is capable of binding a number of different TRAF family members, and it is possible that TAJ-induced JNK activation is mediated by an as yet untested TRAF homolog, such as TRAF6. Similarly, in addition to ASK1, alternative pathways for JNK activation via TNFR family members have been described (23)(24)(25), and it remains to be seen whether TAJ utilizes one of these alternative pathways.
A surprising result of this study was the ability of TAJ to induce cell death since it does not possess a death domain. Our results indicate that TAJ induces cell death by a mechanism independent of the classical death domain receptor-FADDcaspase 8/10 pathway. Although, caspase-mediated cell death is the best characterized form of apoptosis, several caspaseindependent forms of death are also known in the literature. For example, FADD, in addition to its role in the activation of caspases, can also induce caspase-independent cell death (26). This death is resistant to caspase inhibitors and is accompanied by cellular swelling without apoptotic bodies or fragmentation of chromatin (26). Similar forms of caspase-independent death have been previously reported for both Fas and TNFR1 as well (27,28). Several non-death domain-containing members of the TNFR family have been shown to induce cell death under certain conditions (5-7, 29, 30), and it remains to be seen whether TAJ shares a common caspase-independent mechanism of cell death induction with these receptors. Several additional examples of caspase-independent cell death have been recently reported as well (31)(32)(33)(34)(35)(36)(37).
There are several potential mediators of caspase-independent cell death via TAJ. First, Bax and Bax-like proteins have been known to kill mammalian cells even in the presence of caspase inhibitors, provoking chromatin condensation and membrane alterations but without caspase activation or DNA degradation (38 -40). It has been proposed that Bax and Baxlike proteins might mediate caspase-independent death via their channel forming ability that could promote mitochondrial permeability transition or puncture the outer mitochondrial membrane (41). Second, nitric oxide, like Bax, also induces cell death accompanied by chromatin condensation, nuclear compaction, and mitochondrial swelling but without caspase activation or DNA fragmentation (42). The recent isolation of apoptosis-inducing factor may provide an additional mechanism for caspase-independent cell death induction by TAJ (43,44). Apoptosis-inducing factor is a flavoprotein of approximately 57 kDa that is normally confined to mitochondrial intermembrane space but translocates to the nucleus when apoptosis is induced. Apoptosis-inducing factor can induce chromatin condensation in isolated nuclei, dissipation of mitochondrial ⌬ m , and exposure of phosphatidylserine in the plasma membrane. None of these effects are blocked by treatment with broad-spectrum caspase inhibitor benzyloxycarbonyl-VAD-fluoromethyl ketone (43,44). We are currently testing the contribution of the above mediators to TAJ-induced cell death.
The ligand for TAJ has not been identified yet. Recently, two novel ligands belonging to the TNF family, designated APRIL (45) and THANK/BlyS (46,47), respectively, have been identified. Besides Blys/THANK, TAJ could also be a receptor for VEGI, another member of the tumor necrosis factor family for which receptor has not been identified. Studies are in progress to test whether one of these represents the ligand for TAJ.