A Newly Identified Member of the Tumor Necrosis Factor Receptor Superfamily with a Wide Tissue Distribution and Involvement in Lymphocyte Activation*

The tumor necrosis factor receptor (TNFR) superfamily consists of approximately 10 characterized members of human proteins. We have identified a new member of the TNFR superfamily, TR2, from a search of an expressed sequence tag data base. cDNA cloning and Northern blot hybridization demonstrated multiple mRNA species, of which a 1.7-kilobase form was most abundant. However, TR2 is encoded by a single gene which, maps to chromosome 1p36.22–36.3, in the same region as several other members of the TNFR superfamily. The most abundant TR2 open reading frame encodes a 283-amino acid single transmembrane protein with a 36-residue signal sequence, two perfect and two imperfect TNFR-like cysteine-rich domains, and a short cytoplasmic tail with some similarity to 4–1BB and CD40. TR2 mRNA is expressed in multiple human tissues and cell lines and shows a constitutive and relatively high expression in peripheral blood T cells, B cells, and monocytes. A TR2-Fc fusion protein inhibited a mixed lymphocyte reaction-mediated proliferation suggesting that the receptor and/or its ligand play a role in T cell stimulation.

The members of the tumor necrosis factor receptor (TNFR) 1 / nerve growth factor receptor (NGFR) superfamily are characterized by the presence of three to six repeats of a cysteine-rich motif that consists of approximately 30 -40 amino acids in the extracellular part of the molecule (1). The crystal structure of TNFR-I complexed with its ligand showed that a cysteine-rich motif (TNFR domain) was composed of three elongated strands of residues held together by a twisted ladder of disulfide bonds (2). These receptors contain a hinge-like region immediately adjacent to the transmembrane domain, characterized by a lack of cysteine residues and a high proportion of serine, threonine, and proline, which are likely to be glycosylated with O-linked sugars. A cytoplasmic part of these molecules shows limited sequence similarities, a finding that may be the basis for diverse cellular signaling. At present, the members identified from human cells include CD40 (3,4), 4 -1BB (5), OX-40 (6), TNFR-I (7,8), TNFR-II (9), CD27 (10), Fas (11), NGFR (12), CD30 (13), and LTBR (14). Viral open reading frames encoding soluble TNFRs have also been identified, such as SFV-T2 (9), Va53 (15), G4RG (16), and crmB (17).
Because of their biological significance and the diverse membership of this superfamily, we predicted that there would be further members of the superfamily. By searching an EST data base, we identified a new member of the TNFR superfamily. We report here the initial characterization of the molecule called TR2.

MATERIALS AND METHODS
Identification and Cloning of New Members of the TNFR Superfamily-An EST cDNA data base, obtained from over 500 different cDNA libraries (29,30), was screened for sequence homology with cysteinerich motif of the TNFR superfamily, using the blastn and tblastn algorithms (31). One EST was identified in a human T cell line library, which showed significant sequence identity to TNFR-II at the amino acid level. This sequence was used to clone the missing 5Ј end by RACE (rapid amplification of cDNA ends) using a 5Ј-RACE ends-ready cDNA from human leukocytes (Clontech, Palo Alto, CA). This sequence matched three further ESTs (HTOBH42, HTOAU65, and HLHA49). Complete sequencing of these and other cDNAs indicated that they contained an identical open reading frame homologous to the TNFR superfamily, and it was named TR2. Analysis of several other ESTs and cDNAs indicated that some cDNAs had additional sequences inserted into the open reading frame identified above and might represent various partially spliced mRNAs.
Cells-The myeloid and B cell lines studied represent cell types at different stages of the differentiation pathway. KG1a and PLB 985 (32,33) were obtained from Phillip Koeffler (UCLA School of Medicine), BJA-B was from Z Jonak (SmithKline Beecham), and TF 274, a stromal cell line exhibiting osteoblastic features, was generated from the bone marrow of a healthy male donor. 2 All of the other cell lines were obtained from the American Type Culture Collection (Rockville, MD). Monocytes were prepared by differential centrifugation of peripheral blood mononuclear cells (PBMC) and adhesion to tissue culture dish. CD19 ϩ , CD4 ϩ , and CD8 ϩ were isolated from PBMC by immunomag-* This work was supported by National Institutes of Health Grants AI 28175 and AR 40248 (to B. S. K.) and by a postdoctoral fellowship from the American Heart Association, Indiana Affiliate (to S. W.). 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 first two authors contributed equally to this work. ¶ To whom correspondence should be addressed. netic beads (Dynal, Lake Success, NY). Endothelial cells from human coronary artery were purchased from Clonetics (San Diego, CA). RNA and DNA Blot Hybridization-Total RNA of adult tissues was purchased from CLONTECH or extracted from primary cells and cell lines with TriReagent (Molecular Research Center, Inc., Cincinnati, OH). 5-7.5 g of total RNA was fractionated in a 1% agarose gel containing formaldehyde, as described (34), and transferred quantitatively to Zeta-Probe nylon membrane (Bio-Rad) by vacuum blotting. The blots were prehybridized, hybridized with 32 P-labeled Xhol/EcoRI fragment of TR2 or OX-40 probe, washed under high stringency conditions, and exposed to x-ray films.
High molecular weight human DNA was digested with various restriction enzymes and fractionated in 0.8% agarose gel. The DNA was denatured, neutralized, and transferred to nylon membrane and hybridized to 32 P-labeled TR2 or its variant cDNA.
In Situ Hybridization and FISH Detection-The in situ hybridization and FISH detection of TR2 location in human chromosomes were performed as described previously (35,36). FISH signals and the DAPI banding pattern were recorded separately by taking photographs, and the assignment of the FISH mapping data with chromosomal bands was achieved by superimposing FISH signals with a DAPI-banded chromosome (37).
Production of Recombinant TR2-Fc Fusion Proteins-The 5Ј portion of the TR2 containing the entire putative open reading frame of extracellular domain was amplified by polymerase chain reaction (38). For correctly oriented cloning, a HindIII site on the 5Ј end of the forward primer and a BglII site on the 5Ј end of the reverse primer were created. The Fc portion of human IgG 1 was PCR-amplified from ARH-77 (ATCC) cell RNA and cloned in the SmaI site of the pGem7 vector (Promega, Madison, WI). The Fc fragment, including hinge, CH 2 , and CH 3 domain sequences, contained a BglII site at its 5Ј end and an XhoI site at its 3Ј end. The HindIII-BglII fragment of TR2 cDNA was inserted upstream of human IgG 1 Fc and an in-frame fusion was confirmed by sequencing. The TR2-Fc fragment was released by digesting the plasmid with Hin-dIII-XhoI and cloned into pcDNA3 expression plasmid.
The TR2-Fc plasmid, linearized with PvuI, was transfected into NIH 3T3 by the calcium phosphate co-precipitation method. After selection in 400 g/ml G418, neomycin-resistant colonies were picked and expanded. Enzyme-linked immunosorbent assay with anti-human IgG 1 and Northern analysis with 32 P-labeled TR2 probe were used to select for clones that produce high levels of TR2-Fc in the supernatant. In some experiments, a slightly differently engineered TR2-Fc produced in Chinese hamster ovary (CHO) cells was used. The TR2-Fc was purified by protein G chromatography, and the amino acid sequence of the N terminus was determined by automatic peptide sequencer (ABI).
In Vitro Transcription and Translation-The full-length TR2 cDNA was inserted into HindIII-XhoI sites of pcDNA 3 vector (Invitrogen, San Diego, CA). TNT-coupled reticulocyte lysate system (Promega) was used to in vitro transcribe and translate the TR2 cDNA in pcDNA 3. The 35 S-labeled reaction product was fractionated on a 5-15% gradient SDS-polyacrylamide gel, transferred onto an Immobilon membrane (Millipore, Bedford, MA), and exposed to x-ray film.
Blocking MLR-mediated PBMC Proliferation-PBMC were isolated from three healthy adult volunteers by Ficoll gradient centrifugation at 400 ϫ g for 30 min. PBMCs were recovered, washed in RPMI 1640 (Life Technologies, Inc.) supplemented with 10% fetal bovine serum, 300 g/ml L-glutamine, and 50 g/ml gentamycin, and adjusted to 1 ϫ 10 6 cells/ml for two donors and to 2 ϫ 10 5 cells/ml for the third donor.
Fifty l of each cell suspension was added to 96-well (round bottom) plates (Falcon, Franklin Lakes, NJ) together with 50 l of TR2-Fc, IL-5R-Fc, anti-CD4 mAb, or control mAb. Plates were incubated at 37°C in 5% CO 2 for 96 h. One Ci of [ 3 H]methylthymidine (ICN Biomedicals, Costa Mesa, CA) was then added for an additional 16 h. Cells were harvested, and radioactivity was counted.

RESULTS AND DISCUSSION
TR2 Is a New Member of the TNFR Superfamily- Fig. 1a shows the amino acid sequence of TR2 deduced from the longest open reading frame of one of the isolated cDNAs (HLHA49). Comparison with other sequenced cDNAs and with ESTs in the data base indicated potential allelic variants that resulted in amino acid changes at positions 17 (either Arg or Lys) and 41 (either Ser or Phe) of the protein sequence.
The open reading frame encodes 283 amino acids with a calculated molecular weight of 30,417. The TR2 protein was expected to be a receptor. Therefore, the potential signal se- quence and transmembrane domain were sought. A hydrophobic stretch of 23 amino acids toward the C terminus (amino acids 203-225) was assigned as a transmembrane domain, because it made a potentially single helical span (Fig. 1a), but the signal sequence was less obvious. The potential ectodomain of TR2 was expressed in NIH 3T3 and CHO cells as a Fc-fusion protein, and the N-terminal amino acid sequence of the recombinant TR2-Fc protein was determined in both cases. The Nterminal sequence of the processed mature TR2 started from amino acid 37, indicating that the first 36 amino acids constituted the signal sequence (Fig. 1a).
As shown in Fig. 1b, the in vitro translation product of TR2 cDNA was 32 kDa in molecular size. Since the first 36 amino acids constituted signal sequence, and its calculated molecular size was ϳ4 kDa, the molecular size of the protein backbone of processed TR2 would be approximately 28 kDa. Recently, Montgomery et al. (39) published a herpesvirus entry mediator (HVEM) whose cDNA sequence was identical to TR2. They found that the transfected HVEM cDNA produced a 32-36-kDa protein. Since it is larger than the in vitro product, this suggests that the protein is modified posttranslationally. Two potential asparagine-linked glycosylation sites are located at amino acid positions 110 and 173, as indicated in Fig. 1a.
Along with the other members of the TNFR family, TR2 contains the characteristic cysteine-rich motifs that have been shown by x-ray crystallography (2) to represent a repetitive structural unit. Fig. 1c shows the potential TNFR domain aligned among TR2, TNFR-I, TNFR-II, CD40, and 4 -1BB. TR2 contained two perfect TNFR motifs and two imperfect ones.
The TR2 cytoplasmic tail (TR-2 cy) does not contain the death domain seen in the Fas and TNFR-I intracellular domains, and appears to be more related to those of CD40cy and 4 -1bbcy. Signals through 4 -1BB and CD40 have been shown to be co-stimulatory to T cells and B cells, respectively (40,41).
TR2 RNA Expression-A human tissue RNA blot was used to determine tissue distribution of TR2 mRNA expression. TR2 mRNA was detected in several tissues with a relatively high level in the lung, spleen, and thymus, but was not found in the brain, liver, or skeletal muscle (Fig. 2a). TR2 was also expressed in monocytes, CD19 ϩ B cells, and resting or PMA plus PHA-treated CD4 ϩ or CD8 ϩ T cells. It was only weakly expressed in bone marrow and endothelial cells (Fig. 2b), although expression was observed in the hematopoietic cell line KG1a. For comparison, the tissue distribution of OX-40, another member of the TNFR superfamily, was examined. Unlike TR2, OX-40 was not detected in the tissues examined and was detected only in activated T cells and KG1a. Several cell lines were negative for TR2 expression, including TF274 (bone marrow stromal), MG63, TE85 (osteosarcomas), RL 95-2 (endometrial sarcoma), MCF-7, T-47D (breast cancer cells), BE, HT 29 (colon cancer cells), HTB-11, and IMR-32 (neuroblastomas), although TR2 was found in the rhabdosarcoma HTB-82 (data not shown).
Several cell lines were examined for inducible TR2 expression. HL60, U937, and THP1, which belong to the myelomonocytic lineage, all increased TR2 expression in response to the differentiating agents PMA or Me 2 SO (Fig. 2c). Increases in expression in response to these agents were also observed in KG1a and Jurkat cells. In contrast, PMA did not induce TR2 expression in MG63, but unexpectedly TNF-␣ did.
In almost all cases, the predominant mRNA was approximately 1.7 kilobases in size, although several higher molecular weight species could be detected in some tissues (Fig. 2a), and many cDNAs and ESTs that were sequenced contained insertions in the coding region indicative of partial splicing. The abundance of higher molecular weight mRNAs raises the pos-sibility that TR2 may in part be regulated at the level of mRNA maturation.
TR2 Maps at 1P36.2-P36.3-The FISH mapping procedure was applied to localize the TR2 gene to a specific human chromosomal region. The assignment of a hybridization signal to the short arm of chromosome 1 was obtained with the aid of DAP I banding. A total of 10 mitotic figures were photographed, one of which is shown in Fig. 3a. The double fluorescent signals are indicated on the schematic diagram of chromosome 1 as shown in Fig. 3b. This result indicated that the TR2 gene is located on the chromosome 1 region p36.2-p36.3. The TR2  (45), and TNFR-II (46), suggesting that it evolved through a localized gene duplication event. Interestingly, all of these receptors have stimulatory phenotypes in T cells in response to cognate ligand binding, in contrast to Fas and TNFR-I, which stimulate apoptosis. This prompted us to test if TR2 might be involved in lymphocyte stimulation.
TR2-Fc Interferes with MLR-mediated Proliferation of PBMC-To determine the potential involvement of cell surface TR2 with its ligand in lymphocyte proliferation, we examined allogeneic MLR proliferative responses. As shown in Fig. 4, a and b, when TR2-Fc was added to the culture, a significant reduction of maximal responses was observed (p Ͻ 0.05). The addition of TR2-Fc at 100 g/ml inhibited the proliferation up to 53%. No significant inhibition of proliferation was observed with the control IL-5R-Fc. Surprisingly, at high concentrations (10 -100 g/ml) IL-5R-Fc was shown to enhance proliferation. The concentrations of TR2-Fc required to inhibit MLR prolif-eration (1-100 g/ml) are comparable with those of CD40-Fc required for inhibition in other lymphocyte assays (47)(48)(49)(50). An anti-CD4 mAb assayed simultaneously inhibited MLR-mediated proliferation up to 60%, whereas a control anti-IL-5 mAb failed to inhibit the proliferation. It is well known that a major component of the MLR proliferative response is T cell-dependent; hence, it would appear that inhibiting the interaction of TR2 with its ligand prevents optimal T lymphocyte activation and proliferation.
Hence, we have identified an additional member of the TNFR superfamily that either plays a direct role in T cell stimulation or binds to a ligand which can stimulate T cell proliferation through one or more receptors, which may include TR2. We are currently trying to identify this ligand to which TR2 binds to clarify its role.