Cloning of a second dendritic cell-associated C-type lectin (dectin-2) and its alternatively spliced isoforms.

Using a subtractive cDNA cloning strategy, we isolated previously five novel genes that were expressed abundantly by the murine dendritic cell (DC) line XS52, but not by the J774 macrophage line. One of these genes encoded a unique, DC-associated C-type lectin, termed "dectin-1." Here we report the characterization of a second novel gene that was also expressed in a DC-specific manner. Clone 1B12 encoded a type II membrane-integrated polypeptide of 209 amino acids containing a single carbohydrate recognition domain motif in the COOH terminus. The expression pattern of this molecule, termed "dectin-2," was almost indistinguishable from that for dectin-1; that is, both were expressed abundantly at mRNA and protein levels by the XS52 DC line, but not by non-DC lines, and both were detected in spleen and thymus, as well as in skin resident DC (i.e. Langerhans cells). Interestingly, reverse transcriptase-polymerase chain reaction and immunoblotting revealed multiple bands of dectin-2 transcripts and proteins suggesting molecular heterogeneity. In fact, we isolated additional cDNA clones encoding two distinct, truncated dectin-2 isoforms. Genomic analyses indicated that a full-length dectin-2 (alpha isoform) is encoded by 6 exons, whereas truncated isoforms (beta and gamma) are produced by alternative splicing. We propose that dectin-2 and its isoforms, together with dectin-1, represent a unique subfamily of DC-associated C-type lectins.

Dendritic cells (DC) 1 are far more potent than other antigenpresenting cells (e.g. macrophages and B cells) in their capacity to activate immunologically naive T cells, and DC are, indeed, responsible for initiating T cell-mediated immune responses to a variety of antigens (1,2). Members of the DC family are distributed to virtually all the organs (except the brain), where they serve as tissue resident antigen-presenting cells, playing critical roles in presenting environmental, microbial, and tumor-associated antigens to the immune system.
Several years ago, we developed stable DC lines from the mouse epidermis (3). These lines, termed as the XS series, maintain many important features of skin resident DC, i.e. Langerhans cells, and they have provided a useful tool for the application of modern technologies for studying DC biology (3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13). Most recently, we employed the subtractive cDNA cloning strategy to identify genes that were expressed preferentially by DC. 2 Briefly, we constructed a DC-specific cDNA library by subtracting cDNAs prepared from the XS52 DC line with excess amounts of mRNAs isolated from the J774 macrophage line. Following three rounds of screening of this library, we identified five novel genes that were expressed selectively by XS52 DC, but not by J774 macrophages. One of these genes encoded a type II membrane-associated polypeptide of 244 amino acids (aa) containing a single carbohydrate recognition domain (CRD) motif at the COOH-terminal end. This polypeptide was, therefore, designated as DC-associated C-type (Ca 2ϩdependent) lectin-1 or dectin-1. Dectin-1 mRNA was expressed in skin resident DC (from which the XS52 line was derived) as well as in spleen and thymus, the tissues known to contain relatively large numbers of DC. 2 Although the physiological function of dectin-1 remains unknown at present, the identification of a novel, DC-associated lectin is of particular interest, because one of the characteristic features of DC is the expression of many C-type lectins (1,2). For example, DC express DEC-205, a type I membrane-integrated glycoprotein that contains 10 distinct CRD motifs in the extracellular region (14). DC also express a macrophage mannose receptor (MMR), which is a type I membrane-associated glycoprotein with 8 CRD motifs (15). With respect to function, currently available data suggest that both DEC-205 and MMR mediate the uptake of glycosylated antigens by DC (1,2,15,16). Unlike DEC-205 and MMR, which contain multiple CRD motifs in the NH 2 -terminal ends, the second group of C-type lectins consists of the polypeptides that contain a single CRD in their COOH termini. Members of this group include: (a) hepatic lectins (HL) (or asialoglycoprotein receptors), (b) macrophage galactose/N-acetylgalactosamine-specific lectin (MGL), (c) CD23 (low affinity Fc⑀ receptor), and (d) various receptors encoded in the natural killer gene complex (e.g. CD69, CD94, NKR-P1, Ly-49, and NKG2) (17,18). Of these type II surface lectins, DC are known to express CD23 (19) and CD69 (20). More recently, Bates et al. (21) identified a novel, DC-associated type II surface lectin, termed DC immunoreceptor (DCIR). In summary, we now know that DC express both type I surface lectins (DEC-205 and MMR) and type II surface lectins (CD23, CD69, DCIR, and dectin-1). In this study, we have identified new members of the DC-associated type II surface lectin family.

EXPERIMENTAL PROCEDURES
Animals-Female BALB/c mice (6 -10-week-old) and female Lou/c rats (6 -20-week-old) were housed in the pathogen-free facility of the Animal Resource Center at the University of Texas Southwestern Medical Center. All the experiments were conducted according to the guidelines of the National Institutes of Health. To study tissue distributions of dectin-2 mRNA and protein, mice were sacrificed by overdose methoxyflurane inhalation.
Epidermal Cell Isolation-Epidermal cells were isolated from abdominal skin of BALB/c mice using two sequential trypsin treatment and then enriched for Langerhans cells by centrifugation over Histopaque (1.083, Sigma) as described previously (11,25). Langerhans cells were depleted from this preparation by anti-Ia mAb plus complement treatment as before (11).
Isolation of Dectin-2 cDNA Clones-A DC-specific cDNA library was constructed by subtracting the cDNA library prepared from the XS52 DC line with excess amounts of mRNAs isolated from the J774 macrophage line, and this library was screened by colony hybridization, slot blotting, and Northern blotting to identify the clones that were expressed selectively by the XS52 DC line. 2 The original clone 1B12 was one of the 50 clones that were selected in the above manner. To isolate additional clones encoding dectin-2 isoforms, the DC-specific library was screened again by using clone 1B12 as a probe.
Preparation of His-Dectin-2 Fusion Protein-His-dectin-2 fusion protein consisting of 6x histidine, and the extracellular domain of dectin-2 (clone 1B12) was produced in Escherichia coli as follows; the DNA fragment encoding an extracellular domain (aa 46-209) of dectin-2 was PCR amplified. BamHI and SmaI sites were then attached to the resulting DNA fragment at the 5Ј-and 3Ј-end, respectively. Using the BamHI and SmaI site, the fragment was ligated to immediately downstream of the 6x histidine sequence in pQE-30 vector (Quiagen, Chatsworth, CA). This recombinant vector was introduced into E. coli; His-dectin-2 protein was extracted from isopropyl-1-thio-␤-D-galactopyranoside-treated E. coli in 8 M urea/100 mM sodium phosphate/10 mM Tris-HCl (pH 8.0) buffer and purified using nickelnitrilotriacetic acid resin (Quiagen). After extensive dialysis against phosphate-buffered saline, relatively small fractions of the His-dectin-2 proteins were recovered in a soluble form and this fraction was used to test its carbohydrate binding potential. Briefly, His-dectin-2 proteins were labeled with 125 I (ICN, Costa Mesa, CA) and incubated for 120 min on ice with agarose beads that were conjugated with mannose, fucose, lactose, GluNAc, or GalNAc (all purchased from Sigma). Specific binding was then examined by counting the radioactivities that were eluted by the addition of the corresponding carbohydrates.
Preparation of Anti-Dectin-2 mAb-Lou/c rats were immunized initially by subcutaneous injection of insoluble fractions of recombinant His-dectin-2 proteins in complete Freund's adjuvant followed by biweekly injections of the same proteins in incomplete Freund's adjuvant.
One week after the fifth immunization, spleens were harvested from these animals, and B cell hybridomas were prepared using the standard protocol (27). Culture supernatants were tested for the presence of antibodies reactive to His-dectin-2 proteins in immunoblotting. Positive clones were recloned by limiting dilution microculture. Immunoglobulin fractions were purified from the culture supernatants of clone R4C2 by ammonium sulfate precipitation followed by affinity chromatography using anti-rat IgG-conjugated agarose.
Immunoblotting-XS52 DC were homogenized in 10 mM HEPES (pH 7.3) with 5-10 strokes with a 27-gauge needle on 1-ml syringe, centrifuged for 10 min at 1000 ϫ g, and the resulting supernatant ("crude lysates") was fractionated into cytosolic and membrane fractions by centrifugation for 40 min at 100,000 ϫ g. In some experiments, whole cell extracts were prepared from several different cell lines in 0.3% Triton X-100 in phosphate-buffered saline followed by centrifugation for 10 min at 1000 ϫ g. COS-1 cells were transfected using FuGene 6 (Roche Molecular Biochemicals) with pZeoSV2(ϩ) vector (Invitrogen, Carlsbad, CA) containing each of the three different dectin-2 cDNA clones; membrane fractions were prepared 72 h after transfection. These samples were separated by 10 -20% or 4 -20% SDS-polyacrylamide gel electrophoresis, transferred onto polyvinylidene difluoride membrane (Millipore, Bedford, MA), and then blotted with anti-dectin-2 mAb R4C2 or control rat IgG. After an extensive wash, the membrane was blotted with horseradish peroxidase-conjugated anti-rat IgG (Zymed Laboratories Inc., San Francisco, CA) and then developed with ECL system (Amersham Pharmacia Biotech).
Phage DNA was purified from plate lysates of phage-infected E. coli and subcloned into a plasmid vector. The plate lysate was centrifuged at 8000 ϫ g for 10 min and treated with 1 mg/ml of DNase I and 5 mg/ml of RNase A and followed by chloroform extraction. Phage particles were precipitated for 1 h at 4°C with 10% polyethylene glycol (6000) and 1 M NaCl, and the pellet was resuspended in buffer (100 mM NaCl, 10 mM MgSO 4 , 35 mM Tris-HCl (pH 7.5), 2% gelatin) and extracted with phenol and chloroform. Finally, the phage DNA was precipitated with ethanol and dissolved in TE buffer (10 mM Tris-HCl (pH 7.5), 1 mM EDTA).
Genomic DNA fragments were excised from the phage DNA by digestion with BamHI and subcloned into a plasmid vector, pGEM-7zf(-) (Promega, Madison, WI). To determine restriction fragments containing exons, the subcloned DNAs were digested with different restriction enzymes and southern hybridized with a 1B12 cDNA probe. The hybridized DNA fragments were eluted from agarose gel and further subcloned into a pBluescript II SK(-) (Stratagene, La Jolla, CA), and their DNA sequences were determined.

RESULTS
Cloning of Dectin-2 cDNA and Its Deduced Amino Acid Sequence-As described in the Introduction, using a subtractive cloning strategy (XS52 DC line minus J774 macrophage line), we isolated five cDNA clones that were expressed selectively by XS52 DC and that encoded novel polypeptides. One of these clones (clone 1C11-5) encoded a 244-aa polypeptide, termed dectin-1, which contained a single CRD motif in the COOH terminus. Another clone 1B12 contained 630 nt in its open reading frame. As shown in Fig. 1, its deduced amino acid sequence revealed a 209-aa polypeptide of type II membraneintegrated configuration, consisting of a cytoplasmic domain (aa 1-14), a putative transmembrane domain (aa , and extracellular domains (aa . The overall amino acid sequence encoded by clone 1B12 showed 33.5% sequence identity (Clustral method analyzed with Lasergene Program, DNA Star, Madison, WI) to murine (m) DCIR, a DC-associated surface receptor (21). The 1B12 polypeptide sequence also showed significant homology with Dectin-2 and Its Isoforms other molecules, including mMGL (23.4%) (28), mHL2 (22.0%) (29), mHL1 (21.1%) (30), mCD69 (16.6%) (31), and mCD23 (16.3%) (32). Each of these molecules contained a single CRD motif at the COOH-terminal end and, thus, belongs structurally to the type II membrane-associated C-type lectin family. As indicated by asterisks in Fig. 1, the COOH-terminal region (aa 79 -209) of the 1B12 polypeptide contained all the 13 invariant amino acid residues known to be conserved in the CRD motifs of many C-type lectins (33). Thus, we designated this polypeptide as DC-associated C-type lectin-2 or dectin-2.
Tissue and Cell Distributions of Dectin-2 Transcripts-In Northern blotting, we identified a major dectin-2 mRNA of 1.5 kb expressed by the XS52 DC line (Fig. 3A). Dectin-2 mRNA was undetectable in any of the tested non-DC lines, including two macrophage lines (J774 and Raw), three T cell lines (7-17, HDK-1, and D10), a B cell hybridoma (5C5), a keratinocyte line (Pam 212), and a fibroblast line (NS01). The observed cell type specificity was not simply, because of the fact that only the XS52 DC line was cultured in the presence of granulocyte/ macrophage-colony stimulating factor and CSF-1 (3,4), because this line continued to express dectin-2 mRNA even after removal of these cytokines from culture medium and because its expression was not inducible in the J774 macrophage line by exposure to either cytokine (Fig. 3B). As shown in Fig. 3C, dectin-2 mRNA was expressed most abundantly in the spleen and thymus, tissues known to contain relatively large numbers of DC (1, 2).
Unexpectedly, dectin-2 mRNA was not detectable by Northern blotting in the skin, despite the facts that the XS52 DC line was established from this tissue (3) and that this DC line Dectin-2 and Its Isoforms resembles skin resident DC (Langerhans cells) in many features (3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13). We then tested dectin-2 mRNA expression in the epidermis using a more sensitive, RT-PCR analysis. As shown in Fig. 3D, two dectin-2 mRNA species were clearly detected in epidermal cells freshly isolated from BALB/c mice. The upper band represented the PCR product with the predicted molecular size of 631 base pairs, whereas the identity of the lower band remained unclear at that time (see below). Mouse epidermis contains keratinocytes and resident ␥␦ T cells, in addition to Langerhans cells expressing the Ia (MHC class II) molecules. Importantly, depletion of the Ia ϩ epidermal cell population abrogated almost completely both PCR signals for dectin-2. This corroborates our observations with cell lines; dectin-2 mRNA was detected by Northern blotting in the Langerhans cell-like XS52 line, but was totally absent from the Pam 212 keratinocyte line and the 7-17 epidermal ␥␦ T cell line (Fig.  3A). Thus, dectin-2 transcripts are expressed constitutively in the epidermis, and skin resident DC appear to be the major source of this expression. It should be emphasized that the observed expression patterns of dectin-2 mRNA were almost indistinguishable from those for dectin-1 mRNA. 2 Identification of Dectin-2 Proteins-To raise anti-dectin-2 mAb, we prepared a fusion protein consisting of a 6x His tag and the extracellular domains (aa 46 -209) of dectin-2. This fusion protein, His-dectin-2, was produced in E. coli, extracted in 8 M urea, and purified by nickel affinity chromatography. One of the rat mAb (R4C2) raised against His-dectin-2 recognized multiple bands (ranging from 23 to 31 kDa) in the whole cell extracts prepared from the XS52 DC line (Fig. 4A). Consistent with the predicted molecular structure as shown in Fig.  1, these bands were detected primarily in the membrane fraction isolated from the XS52 DC line (Fig. 4B). Most importantly, none of these bands were detected in the extracts prepared from any of the tested non-DC lines, including macrophage, B cell hybridoma, T cell, keratinocyte, and fibroblast lines (Fig. 4A). The two bands (of about 50 and 25 kDa) observed in the extracts from the 2G9 B cell hybridoma represented immunoglobulin heavy and light chains being detected by our secondary Ab against rat IgG, because the same bands were also detected with the secondary Ab alone. Thus, dectin-2 proteins are produced selectively by the XS52 DC line, corroborating our observations at mRNA levels (Fig. 3A).
It was of our particular interest to determine whether Hisdectin-2 proteins would recognize one or more conventional carbohydrate moieties. Soluble fractions of His-dectin-2 fusion proteins (containing the CRD motif) were labeled with 125 I and tested for the binding to agarose beads that had been conjugated with mannose, fucose, lactose, GluNAc, or GalNAc. Hisdectin-2 failed to exhibit specific binding to any of these carbohydrate probes, as determined by counting the radioactivities that were eluted by adding the corresponding carbohydrates (data not shown).
Identification of Two Truncated Isoforms of Dectin-2-As noted previously, we detected significant molecular heterogeneity in RT-PCR analysis (Fig. 3D) and immunoblotting (Fig. 4,  A and B), suggesting the presence of different dectin-2 isoforms. This possibility was then tested by RT-PCR analyses using different sets of primers designed to amplify different regions of the dectin-2 gene. As shown in Fig. 5A (left panels), the primer set 3 (designed to amplify the entire coding se-  ). B, XS52 cells and J774 cells were cultured for 48 h in the presence or absence of granulocyte/macrophagecolony stimulating factor (GM-CSF) (10 ng/ml) or CSF-1 (10 ng/ml) before RNA isolation. These RNA samples (10 g/lane) were then examined by Northern blotting for dectin-2 or GAPDH. D, epidermal cells isolated from adult BALB/c mice were examined for dectin-2 mRNA expression by RT-PCR using the primer set 0 ("Experimental Procedures") designed to amplify nt 475-1106. Some samples were treated with anti-Ia mAb plus complement to deplete Langerhans cells; the extent of depletion was assessed by measuring IL-1␤ mRNA, which is known to be expressed exclusively by Langerhans cells within murine epidermal cells.

FIG. 4. Dectin-2 protein expression.
A, whole cell extracts were prepared from the indicated cell lines in the cell lysis buffer containing 0.3% Triton X-100. These samples were examined by immunoblotting with anti-dectin-2 mAb (R4C2) (top) and by Coomassie Blue staining for protein profiles (bottom). The data shown are representative of two independent experiments. B, crude lysates prepared from the indicated numbers (ϫ10 Ϫ5 cells/sample) of XS52 DC were centrifuged at 100,000 ϫ g for 40 min, and the pellets (membrane fractions) were tested for the presence of dectin-2. Data shown are representative immunoblots with anti-dectin-2 mAb (R4C2) (top) and with control rat IgG (bottom) from three independent experiments.

Dectin-2 and Its Isoforms
quence) amplified at least two PCR products with different molecular sizes from the XS52 DC line as well as from the spleen. The primer set 1 (designed to amplify the 5Ј-region encoding the intracellular domain, the transmembrane domain, the neck domain, and a part of the CRD domain) amplified two PCR products with difference sizes. Likewise, the primer set 2 (designed to amplify the 3Ј-region encoding primarily the CRD domain) amplified two different PCR products. In all cases, the upper bands of the RT-PCR products had the anticipated molecular sizes (369 nt with set 1, 429 nt with set 2, and 798 nt with set 3). Taken all together, these results suggested the presence of a unique dectin-2 transcript with truncation within the 5Ј-region (nt 115-483) and a second transcript with truncation within the 3Ј-region (nt 474 -902).
To isolate cDNA clones encoding truncated dectin-2 transcripts, we rescreened the DC-specific cDNA library (XS52 DC minus J774 macrophages) with the original clone 1B12. In fact, several additional clones (27 in total) showed strong hybridization with this probe. By using the same three sets of primers, we analyzed the nucleotide structures of these clones. As shown in Fig. 5A (right panels), these cDNA clones showed different PCR profiles. The first group of cDNA clones (including the original clone 1B12) exhibited the PCR products of the expected molecular sizes with all the three primer sets. The second group of clones (including clone 1A7) showed significant reduction in molecular size (about 100 nt) with the primer set 1 but not with the primer set 2. The third group (including clone 1E4) exhibited significant reduction in size (about 100 nt) with the primer set 2 but not with the primer set 1. The PCR products amplified with the primer set 3 were also smaller in the second and the third groups than in the first group. Based on these results, we concluded that there must exist at least three different dectin-2 transcripts: (a) the first transcript encoding a full-length polypeptide (termed ␣ isoform), (b) the second transcript encoding a polypeptide with truncation in a region from the intracellular domain to the CRD domain (␤ isoform), and (c) the third transcript encoding a polypeptide with truncation within the CRD domain (␥ isoform).
To determine the primary structures of the truncated dectin-2 isoforms, we sequenced clones 1A7 and 1E4. We identified the deletion of nt 263-362 in clone 1A7 and nt 569 -691 in clone 1E4. Their deduced amino acid sequences showed that the truncation occurred primarily within the neck domain for dectin-2 ␤ isoform and exclusively within the CRD domain for the ␥ isoform (Fig. 5B). We next introduced these cDNA clones into COS cells. As shown in Fig. 6, anti-dectin-2 mAb (R4C2) revealed a single band of 31 kDa in the extracts after transfection with the full-length dectin-2 clone 1B12. By contrast, we detected two bands of 28 and 24 kDa after transfection with clone 1A7. Likewise, clone 1E4 produced two bands of 26 and 23 kDa. Based on these observations, we concluded that dectin-2 consists of at least three isoforms: (a) ␣ isoform of 209 aa with the molecular mass of 31 kDa, (b) ␤ isoform with a 34 aa deletion in the neck domain (migrating as 28 and 24 kDa), and (c) ␥ isoform with 41 aa deletion in the CRD (migrating as 26 and 23 kDa). Mechanisms by which a single cDNA clone (encoding the ␤ or ␥ isoform) produced two proteins with different migration profiles remain unclear at this moment.
Genomic Analysis of Dectin-2 DNA-To determine mechanisms for the generation of three transcripts encoding different isoforms, we cloned and sequenced genomic DNA fragments (14 kb in total) of dectin-2, covering from the cytoplasmic domain to the CRD domain. As shown in Fig. 7, the full-length dectin-2 isoform was found to be encoded by six exons, with exon 1 encoding 7 aa in the cytoplasmic domain, exon 2 encoding 31 aa mainly in the transmembrane domain, exon 3 encoding 34 aa  (37) were both conserved in each exon-intron interface. Remarkably, exon 3 was entirely skipped from the nucleotide sequence of clone 1A7. Interestingly, exons 5 and 6 were partially deleted in clone 1E4, and we identified the consensus motifs for splicing donor and acceptor sites within exon 5 and exon 6 at both ends (nt 569 and 691) where the 123-nt deletion was identified in clone 1E4. Based on these observations, we concluded that dectin-2 ␤ and ␥ isoforms are generated through alternative splicing. DISCUSSION In the present study, we have identified a novel, DC-associated C-type lectin, termed dectin-2, which shared several important features with dectin-1. First, both dectin-2 and dectin-1 exhibited a common domain structure, consisting of a relatively short cytoplasmic domain, a transmembrane domain, an extracellular neck domain, and a single CRD in the COOH termini. Second, dectin-2 mRNA expression profiles were indistinguishable from those for dectin-1 mRNA; that is, both were expressed: (a) at relatively high levels in the XS52 DC line but not in other tested non-DC lines, (b) most abundantly in spleen and thymus, and (c) constitutively in the Ia ϩ epidermal cell population (i.e. Langerhans cells). Third, expression of dectin-2 and dectin-1 proteins also occurred exclusively in the XS52 DC line among the tested cell lines. Despite these similarities, the degree of sequence homology between dectin-2 and dectin-1 was relatively low (19.6% identity in the overall sequence and 24.8% within the CRD motif). Thus, dectin-2 and dectin-1 represent two structurally independent, DC-associated C-type lectins.
As described in the Introduction, DC have been shown to express both type I surface lectins (DEC-205 and MMR) and type II surface lectins (CD23, CD69, DCIR, and dectin-1). The present study adds three new members (dectin-2 ␣, ␤, and ␥ isoforms) to this list of DC-associated C-type lectin family. The dectin-2 ␣ isoform showed the highest degree of homology (33.5% identity in the overall sequence and 44.8% within the CRD motif) to murine DCIR, which was identified recently by Bates et al. (21) by searching nucleotide data bases with a consensus sequence derived from the CRD motifs of HL-1, HL-2, and MGL. Dectin-2 differs from DCIR in that the immunoreceptor tyrosine-based inhibitory motif found in the intracellular domain of DICR was absent from the relatively short intracellular domain (14 aa) of dectin-2. Interestingly, dectin-2 also lacked an immunoreceptor tyrosin-based activation motif, which was identified in the intracellular domain of dectin-1. Therefore, an oversimplified scenario would be that dectin-1 and DCIR deliver counteracting signals into DC, whereas dectin-2 has no apparent signaling potential. Unfortunately, no information is available with respect to the natural ligands that are recognized by CD69, DCIR, dectin-1, and/or dectin-2. In this regard, none of the tested carbohydrate probes showed specific binding to His-dectin-2 (or to His-dectin-1). We interpreted these results to suggest that dectin-2 may recognize rather unique carbohydrate moieties or that dectin-2 may have to form heterodimers with a second C-type lectin (perhaps dectin-1) to exert high affinity carbohydrate binding, as has been shown for other type II lectins with single CRD motifs (38,39). An even more extreme possibility is that dectin-2 may recognize polypeptide or glycopolypeptide ligand(s). CD23 has been shown to recognize a polypeptide motif, instead of carbohydrate residues, of the IgE molecule (40), documenting that carbohydrates do not necessarily serve as natural ligands of all the molecules that "structurally" belong to the C-type lectin family. Further studies are required to identify the ligands recognized by the DC-associated type II surface lectins (including dectin-2) and to determine their physiological functions.
One of the striking findings in this study was the identification of two truncated isoforms of dectin-2, i.e. the ␤ isoform with 34 aa deletions in the neck domain and the ␥ isoform with 41 aa deletions within the CRD domain. In this regard, the neck domain of CD69 has been shown to be required for dimer formation (41), whereas the CRD domains most likely mediate the recognition of putative ligands. Therefore, it is tempting to speculate that the ␤ and ␥ isoforms of dectin-2 may differ from the full-length isoform ␣ in one or more functional respects. Once again, one must first identify the ligand(s) and the function of dectin-2 before testing this attractive hypothesis.
Production of different isoforms by alternative splicing has been reported for other members of the type II surface lectin family. For example, Nunez et al. (42) identified a truncated transcript of CD23 from which exon 3 (encoding the transmembrane domain and a portion of the cytoplasmic tail) was deleted by alternative splicing. It was postulated that the resulting CD23 isoform lacking the transmembrane domain might serve as a soluble CD23 receptor controlling CD23-mediated immunological function. Ying et al. (43) identified several alternatively spliced transcripts lacking exon 3 (encoding the cytoplasmic tail) and/or exon 4 (encoding the transmembrane domain) of CD72, a type II surface lectin expressed primarily by B cells. More recently, Furukawa et al. (44) identified an alternatively spliced transcript lacking exon 2 (encoding the transmembrane domain) of CD94, one of the natural killer-associated type II surface lectins. In this regard, the isoforms we identified in dectin-2 are unique in that they lack the neck domain or a portion of the CRD motif. Thus, our observation provides additional mechanisms for creating molecular heterogeneity of type II surface lectins.
In summary, we postulate that dectin-2 and its isoforms, together with other molecules (CD23, CD69, DCIR, and dectin-1), may form a unique subfamily of DC-associated, type II surface lectins. The present study provides an important piece of information for revealing the potential functions of those DC-associated molecules.
Acknowledgments-We thank Drs. Nancy Street and Mansour Mohamadzadeh for providing T cell clones and B cell hybridoma clones, respectively. We are also grateful to Ms. Pat Adcock for her secretarial assistance.