Molecular Cloning and Functional Analysis of the Mouse Homologue of the KILLER/DR5 Tumor Necrosis Factor-related Apoptosis-inducing Ligand (TRAIL) Death Receptor¹

In mammalian cells, apoptosis induction can be triggered by the interaction of a set of ligands and their corresponding death-domain-containing receptors belonging to the TNF⁴ superfamily (1). Upon ligand binding, receptors become trimerized, and associated molecules are recruited to form the DISC (2, 3). FADD is directly recruited into the DISC in Fas-mediated apoptosis, while TNF-mediated apoptosis involves both FADD and TRADD (2, 3). Once the DISC is formed, it activates procaspases resulting in a caspase cascade leading to apoptosis.

TRAIL is a member of the TNF superfamily that is expressed in most human and mouse tissues (4, 5). TRAIL selectively kills transformed cells but not normal cells, which suggests that TRAIL has a potential application for cancer therapy (4, 5). Thus far, four human TRAIL receptors have been identified: Death receptor 4 (DR4) or TRAIL-R1 (6), KILLER/DR5 (TRAIL-R2, TRICK2; 7, 8, 9, 10, 11), TRID (DcR1, TRAIL-R3 or LIT; 9, 12, 13, 14) and TRUNDD (DcR2 or TRAIL-R4; 15, 16, 17). TRAIL receptors have high homology with members of the TNFR family (2). DR4 and KILLER/DR5 are two proapoptotic receptors that contain a cytoplasmic death domain that is required for transducing the apoptotic signal. TRID and TRUNDD are two decoy receptors that antagonize TRAIL-induced apoptosis by competing for ligand binding (2). Both DR4 and KILLER/DR5 rapidly induce apoptosis of human cancer cell lines upon overexpression (6, 8, 9, 10, 13). Moreover, KILLER/DR5 can eliminate colony formation of cancer cells (8).

In a search for p53-downstream targets involved in apoptosis, KILLER/DR5 was identified as a DNA damage-inducible p53-regulated death receptor gene from a chemosensitive ovarian cancer cell line (PA-1) by a subtractive hybridization technique (8). In this study, we report the identification and characterization of MK, and explore its regulation by the tumor suppressor p53.

Mouse fibroblast NIH3T3 and the mouse embryonal carcinoma cell line F9 were obtained from the American Type Culture Collection (Rockville, MD). The human colon cancer cell line SW480 was maintained as described previously (8). The M3 murine lymphoma cell line expressing temperature-sensitive p53 has been described (18, 19). Cells were transfected using pCEP4 vector (Invitrogen, CA), pCEP4MK, pEGFP-N vector, and pEGFP-N-MK. In experiments involving contransfection of pCMV-β-gal (Clonetech, CA) and the pCEP4-expressing vector, a 1:4 ratio of plasmid was used. For NIH3T3 cells, LipofectAMINE (Life Technologies, Inc., Gaithersburg) was used for the transfection. In the case of SW480 cells, Lipofectin was used for the transfection.

BLAST search was performed of the human and mouse EST databases using a sequence of death domain of human TRAIL death receptor KILLER/DR5 (8). An EST clone (AI154278) showed a high degree of homology with death domains of DR4 and KILLER/DR5. On the basis of the EST sequence, primers (MO-P3: 5′-TGT CTG TGA GGC CCA ACT GCC GC-3′; MO4: 5′CGC ACT TTC GGG ATC CCG TTC ACA GCC-3′; MO8: 5′-GCC TGT AGC ACA GCA AGC AGC AAC GC-3′; and MO9: 5′-CTC CTC CGG CCG CTG TAG GCC AGC-3′) were designed and 5′ RACE-PCR was performed using a mouse kidney cDNA library (Clonetech, CA) as a template according to the manufacturer’s instructions. To express MK, two vectors—pCEP4 (Invitrogen, CA) and pEGFP-N (Clonetech, CA)—were used. The cDNA of MK was obtained by PCR with primer pairs of 5′-CCC AAG CTT GCT GAG ACC ATG GAG CCT-3′ and 5′-CCG CTC GAG AAA CGC ACT GAG ATC CTC-3′ for cloning into the pCEP4 vector or 5′-AGC CGG AAT TCC CAC CAT GGA GCC TCC AGG A-3′ and 5′-GCC GGT ACC GGA ACG CAC TGA GAT CCT CCT-3′ for cloning into the pEGFP-N vector. PCR was performed by using Pfu (Stratagene, CA) as enzyme and a mouse kidney library (Clonetech, CA) as a template. The PCR conditions were: 35 cycles of 94°C for 30 s, 57°C for 1 min, and 72°C for 4 min, with one cycle of 72°C for 10 min for the final extension. The PCR product was purified after double digestion by HindIII/XhoI for pCEP4 or EcoRI/KpnI for pEGFP-N.

Apoptosis induction was determined by either β-galactosidase staining or by visible fluorescence under a UV microscope as described previously (8, 20). Briefly, for the β-galactosidase assay, at 48 h after transfection, cells were stained with 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside to assess the number of remaining blue cells as well as to observe morphological characteristics of apoptosis as described previously (20). In the case of apoptosis induction by using the pEGFP vector, 2 days after transfection, cells displaying a visible fluorescence with morphological feature of apoptosis were observed.

The extracellular domain of mouse TRAIL receptor (residues 53–182) was fused inframe to the Fc portion of mouse IgG and then transfected into 293T cells. The conditioned medium was used for TRAIL binding as described previously (7).

At 24 h after transfection of NIH3T3 cells, hygromycin B (200 μg/ml) was added to the medium and maintained for another 10 days. Colonies were stained with Coomassie Blue as described previously (8).

The expression pattern of mouse TRAIL death receptor KILLER/DR5 was determined through hybridization to a mouse multiple tissue blot (Clonetech, CA) using a full-length cDNA of MK as probe. The probe was labeled by random priming as described previously (8). The expression level of mouse actin was used to demonstrate equal loading by reprobing the same membrane after stripping. In the experiment involving induction of MK by the tumor suppressor gene p53 and the DNA-damaging agent Adriamycin, RNA was extracted from M3 cells either at 37°C or 32°C for 10 h with or without Adriamycin treatment (0.2 μg/ml). RNA was extracted from F9 cells at 0, 12, and 24 h after treatment with 40 μg/ml etoposide. The full-length mouse p21 cDNA was used as a probe as described previously (21).

On the basis of the information from EST clone (AI154278), which showed a high degree of homology with the death domains of DR4 and KILLER/DR5, we performed RACE-PCR and obtained the complete sequence for the mouse TRAIL death receptor KILLER/DR5 cDNA (Fig. 1). The sequence predicts an open reading frame of 1146 bp encoding a protein of 381 amino acids. The protein contains a cytoplasmic COOH-terminal death domain, a transmembrane domain, and a cysteine-rich extracellular NH₂-terminal domain characteristic of the TNFR family (2). At the protein level, the death domain of MK shows high homology with human TRAIL receptors DR4 (76%) and KILLER/DR5 (79%; Fig. 2,B and Table 1). Overall, the whole sequence of MK shares 43 and 49% of similarity with DR4 and KILLER/DR5, respectively (Fig. 2,A and Table 1) as well as high homology with two decoy receptors TRID, TRUNDD (Table 1). MK is more closely related to TRAIL receptors than either Fas or TNFR-1 (Table 1). Analysis of the DNA sequence indicated that MK has a Kozak sequence upstream of the initiating methionine 5′-GACACCATGG-3′ (data not shown).

Because MK is more closely related to human TRAIL receptors than other death domain-containing receptors such as Fas or TNFR-1, we suspected that MK may be a novel TRAIL receptor. To test this possibility, we fused the extracellular domain of MK to the Fc portion of mouse IgG and then performed a TRAIL-binding assay (Fig. 2,C). The results revealed that TRAIL can bind MK with similar affinity as DR4 and KILLER/DR5 (Fig, 2 C, Lanes 2-4), supporting the hypothesis that MK is a mouse TRAIL receptor.

Northern blot analysis using the MK probe revealed the expression of a single MK TRAIL death receptor transcript of approximately 3.4 kb in the mouse multiple tissue blot (Fig. 2,D). MK is predominantly expressed in heart, lung, and kidney (Fig. 2 D). In humans, two proapoptotic TRAIL receptors DR4 and KILLER/DR5 are expressed in almost all tissues (6, 9), as are the two decoy receptors TRID and TRUNDD (9, 13, 15, 16). This preferential expression of MK in heart, lung, and kidney suggests that MK may play a specialized role in these tissues as compared with other tissues in which MK expression is low or undetectable.

Overexpression of the two proapoptotic human TRAIL receptors DR4 and KILLER/DR5 can rapidly induce apoptosis in human cells (6, 7, 8, 9). On the basis of its death domain-containing structure, we predicted that MK may be a proapoptotic death receptor. To investigate whether MK could induce apoptosis, we transfected MK-expressing vectors into mouse or human cells. As shown in Fig. 3, A and B, overexpression of pCEP4MK or pEGFP-N-MK rapidly induces apoptosis of SW480 cells (a human colon cancer cell line), NIH3T3 cells (a mouse fibroblast line), and 293 T cells (data not shown), which suggests that MK is a proapoptotic receptor.

The evidence presented above suggested that MK is a mouse homologue of the human TRAIL receptors DR4 and KILLER/DR5. We previously showed (8) that overexpression of KILLER/DR5 can eliminate colony formation of human cancer cell lines. To test the possibility that MK can inhibit colony formation, we transfected an empty vector pCEP4 or the MK-expressing vector pCEP4MK into NIH3T3 cells. After 10 days of selection in the presence of hygromycin B, we found that MK is a potent suppressor of colony formation (Fig. 3,C). In pCEP4MK transfected cells, there were extremely rare surviving colonies as compared with empty vector pCEP4 transfected cells (Fig. 3,C). A similar result was obtained in the human colon cancer cell line SW480 (data not shown). Inhibition of colony formation by overexpression of the mouse TRAIL receptor MK is probably due to apoptosis induction in these cells (Fig. 3, A and B).

We previously showed that the human TRAIL death receptor KILLER/DR5 is a candidate p53 target gene (8). After DNA damage, KILLER/DR5 was induced in cells with a functional p53 but not in cells with a mutant p53 or where p53 is degraded. To investigate whether MK is also regulated by p53, we used M3 murine lymphoid cells harboring a temperature sensitive-p53 to examine MK expression in the presence or absence of DNA damage. When the temperature is shifted from 37°C to 32°C, p53 adopts a wild-type conformation and M3 cells undergo massive apoptosis (18, 19). As shown in Fig. 4,A, MK is induced upon p53 activation at 32°C but not at 37°C (Fig. 4,A, Lane 1 versus Lane 3). In the presence of wild-type p53 at 32°C but not mutant p53 (37°C), expression of MK is further elevated after exposure of M3 cells to the DNA-damaging agent Adriamycin (Fig. 4,A, Lane 2 versus Lane 4). These results indicate that MK is a putative p53 target gene and that DNA damage induces MK expression in a p53-dependent manner. Previous studies (22, 23) performed with wild-type p53-expressing F9 teratocarcinoma cells have demonstrated that these cells undergo apoptosis upon exposure to DNA-damaging agents. Because we have found that the human KILLER/DR5 is induced only in cells undergoing p53-dependent apoptosis but not in cells undergoing p53-mediated cell-cycle arrest,⁵ we examined the expression of MK after DNA damage in F9 cells. As shown in Fig. 4,B, MK was induced after etoposide treatment and correlated with p53 stabilization and apoptosis as demonstrated by PARP cleavage (Fig. 4 C). Together, these results implicate a possible role for MK in p53-mediated apoptosis.

We report the isolation and characterization of a cDNA encoding the mouse receptor (MK) for TRAIL. Like the human proapoptotic TRAIL receptors DR4 and KILLER/DR5, the mouse TRAIL receptor MK displays all of the features characteristic of the TNFR family. In humans, there are four TRAIL receptors identified, and they are classified into two categories based on their ability to induce or inhibit apoptosis (2). Several lines of evidence support the conclusion that MK is a mouse homologue of KILLER/DR5:

(a) MK rapidly induces apoptosis of human and mouse cell lines and inhibits colony formation (Fig. 3);

(b) MK has more similarity to the human TRAIL receptor KILLER/DR5 than to DR4 (79 versus 76%, respectively; Fig. 2);

(c) MK is able to bind TRAIL, which suggests that it may in fact be a TRAIL receptor (Fig. 2 C); and

(d) MK is induced upon p53 activation, and this induction is further elevated in the presence of DNA-damaging agents.

In humans, KILLER/DR5 but not DR4 is regulated by the tumor suppressor gene p53 (Wu et al.⁵ and data not shown). Induction of MK in cells upon p53 activation and by DNA damage is consistent with our previous observation in humans that KILLER/DR5 seems to be a p53 target gene (8) and suggests that MK may be the mouse homologue of KILLER/DR5. On the basis of this reasoning, we conclude that MK is a DNA damage-inducible p53-regulated mouse TRAIL death receptor.

Interestingly, the expression of MK is particularly abundant in heart, lung, and kidney, which is different from its human counterpart KILLER/DR5 as well as from DR4, TRUNDD, and TRID, which are expressed in most tissues (7, 9, 10, 13). It is notable that the expression of MK in immune tissues such as spleen is not appreciable (Fig. 2 D). In mice, TRAIL is highly expressed in spleen, prostate, and lung but not in brain, liver, and testes (4, 5). Expression of MK is also not observed in brain, liver, and testes. Therefore, the high expression of both TRAIL and its receptor in the lung suggests a possibly important role for TRAIL and its receptor in this particular tissue.

In addition to providing a tool for studying the role of KILLER/DR5 in p53-dependent apoptosis in mouse cell lines and tissues, the isolation of MK makes possible a number of other future experiments. Among these is the possibility to determine whether other TRAIL receptors exist in mice. If in the future such homologues are not found through low stringency hybridization or database searching, the inhibition of MK expression through knockout or antisense strategies should provide a way to determine whether MK is the sole determinant of TRAIL sensitivity. The isolation of MK also makes it possible for the first time to examine the role of TRAIL receptors in tumor development in different genetic backgrounds, e.g., in p53-null or other death receptor-null mice.

In summary, we have isolated the mouse TRAIL death receptor KILLER/DR5 (MK) whose expression appears to be p53-dependent and is induced in cells undergoing p53-dependent apoptosis. Overexpression of MK rapidly induces apoptosis of mouse and human cells and suppresses colony growth. Future experiments will focus on elucidating the role of MK in p53-dependent apoptosis and mouse embryonic development.