Identification of a human homologue of the dendritic cell-associated C-type lectin-1, dectin-1

Gene 272 (2001) 51±60

www.elsevier.com/locate/gene

Identi®cation of a human homologue of the dendritic cell-associated C-type lectin-1, dectin-1 q Koichi Yokota 1, Akira Takashima, Paul R. Bergstresser, Kiyoshi Ariizumi* Department of Dermatology, The University of Texas Southwestern Medical Center, Dallas, TX, USA Received 20 November 2000; received in revised form 10 February 2001; accepted 14 May 2001 Received by V. Larionov

Abstract Previously we identi®ed the novel type II lectin receptor, dectin-1, that is expressed preferentially by murine antigen presenting dendritic cells (DC) and is involved in co-stimulation of T cells by DC. To identify the human homologue (DECTIN-1), we employed degenerative PCR ampli®cation of mRNA isolated from DC and subsequent cDNA cloning. DECTIN-1 is a type II lectin receptor with high homology to type II lectin receptors expressed by natural killer (NK) cells. It contains an immunoreceptor tyrosine-based activation motif within the cytoplasmic domain. Human DECTIN-1 mRNA is expressed predominantly by peripheral blood leukocytes and preferentially by DC. The mRNA likely encodes a 33 kDa glycoprotein. In human epidermis, the protein is expressed selectively by Langerhans cells, which are an epidermal subset of DC. A truncated form of DECTIN-1 RNA (termed Tb) encodes for a polypeptide lacking almost the entire neck domain, which is required for accessibility of the carbohydrate recognition domain to ligands. Genome analysis showed the deleted amino acid sequence in Tb to be encoded by an exon, indicating that Tb RNA is produced by alternative splicing. DECTIN-1 gene maps to chromosome 12, between p13.2 and p12.3, close to the NK gene complex (12p13.1 to p13.2) which contains genes for NK lectin receptors. Our results indicate that human DECTIN-1 shares many features with mouse dectin-1, including the generation of neck domain-lacking isoforms, which may down-regulate the co-stimulatory function of dectin-1. q 2001 Elsevier Science B.V. All rights reserved. Keywords: C-type lectin; Dendritic cells; Alternative splicing; Natural killer receptor; Natural killer gene complex; Langerhans cells

1. Introduction Dendritic cells (DC) belong to the family of antigen presenting cells (APC) characterized morphologically by long, lamellar dendrites (Steinman, 1991). Compared to other APC (e.g. macrophages and B cells), DC have unsurpassed potency in activating immunologically naive T cells (Steinman, 1991). DC are capable of presenting T cells with Abbreviations: aa, amino acid; Ab, antibody; APC, antigen presenting cells; ATCC, American Type Culture Collection; CRD, carbohydrate recognition domain; DC, dendritic cells; DCIR, DC immunoreceptor; HL, hepatic lectin; ITAM, immunoreceptor tyrosine-based activation motif; ITIM, immunoreceptor tyrosine-based inhibitory motif; LC, Langerhans cells; mAb, monoclonal antibody; NK, natural killer; NKC, NK gene complex; nt, nucleotide; PBMC, peripheral blood mononuclear cells; PCR, polymerase chain reaction q Nucleotide sequences derived from this article have been deposited with GenBank Data Libraries under the Accession numbers AF313468 (human DECTIN-1) and AF313469 (DECTIN-1 Tb isoform). * Corresponding author. 5323 Harry Hines Boulevard, Dallas, TX 753909069, USA. Tel.: 11-214-648-7552; fax: 11-214-648-0280. E-mail address: [email protected] (K. Ariizumi). 1 Present address: Department of Dermatology, Hokkaido University School of Medicine, Sapporo, Japan.

a wide variety of antigens including chemical haptens and those associated with viral and bacterial organisms as well as tumors. DC have been reported to more ef®ciently endocytose glycosylated protein antigens compared to nonglycosylated forms (Lanzavecchia, 1993), suggesting that DC abundantly express receptors that internalize antigens via recognition of their carbohydrate moieties. In the process of searching for DC-speci®c receptors, recently many novel C-type lectin receptors have been discovered. The C-type lectin receptors are structurally divided into type I lectin receptors containing multiple carbohydrate recognition domains (CRDs) on their aminotermini, and type II lectin receptors containing a single CRD on their carboxyl-termini. DEC-205 (Jiang et al., 1995) is a type I lectin receptor that is thought to endocytose glycosylated antigens via recognition of carbohydrate moieties, although its carbohydrate-binding capacity has not yet been de®ned. On the other hand, DC-SIGN (Geijtenbeek et al., 2000), langerin (Valladeau et al., 2000), DC immunoreceptor (DCIR) (Bates et al., 1999), C-type lectin-like receptor-1 (CLEC-1) and CLEC-2 (Colonna et al., 2000) are type II lectin receptors; DC-SIGN and langerin recognize

0378-1119/01/$ - see front matter q 2001 Elsevier Science B.V. All rights reserved. PII: S 0378-111 9(01)00528-5

52

K. Yokota et al. / Gene 272 (2001) 51±60

mannose/mannan, whereas the carbohydrate binding capacities of the other receptors have not been de®ned. It is possible that DC lectin receptors do not recognize carbohydrates as speci®c ligands, but rather recognize peptide sequences of surface molecules (e.g. adhesion and co-stimulatory molecules) expressed on T cells or other cells. Alternatively, DC lectin receptors recognize complex carbohydrate moieties (e.g. sialyl Lewis x tetrasaccharides recognized by selectins; Vestweber and Blanks, 1999) functioning as receptors for speci®c glycosylated antigens that lead to distinctive intracellular pathways for the antigen processing by traf®cking signals in their cytoplasmic tails. Recently, by subtractive cDNA cloning using mouse XS52 DC and J774 macrophage lines, we have isolated two new members of the C-type lectin superfamily, termed DC-associated C-type lectin-1 and -2 or dectin-1 (Ariizumi et al., 2000a) and dectin-2 (Ariizumi et al., 2000b). XS52 DC were established from newborn skin of BALB/c and resemble an epidermal subset of DC termed Langerhans cells (LC) in dendritic morphology, surface phenotype, and APC capacity (Xu et al., 1995a,b). Dectin-1 and dectin-2 are type II transmembrane proteins with a single CRD that are expressed preferentially or selectively by DC including LC. Despite identi®cation of their CRDs, the carbohydrate binding capacity has yet to be formally proven. On the other hand, we have shown that dectin-1 recombinant protein binds speci®cally to the surface of T cells in a saccharide-independent manner (Ariizumi et al., 2000a), suggesting recognition of peptide sequences by dectin-1. Using PCR ampli®cation of mRNA isolated from human DC with human guess-mers and subsequent cloning of the full-length cDNA, we have identi®ed the human homologue of dectin-1, which encodes for a novel type II transmembrane protein with a single CRD. The genetic and biochemical characteristics of the human DECTIN-1 are reported herein. 2. Materials and methods 2.1. Cell lines Monocyte (THP-1, U937, and HL-60), B cell (Raji), and T cell (Jurkat) lines were purchased from American Type Culture Collection (ATCC, Rockville, MD) and maintained in complete RPMI 1640 supplemented with 10% fetal bovine serum. The NK cell line (NKL) (Robertson et al., 1996) was a gift from Dr Michael J. Robertson (Division of Hematologic Malignancies, Dana Farber Cancer Institute) and was maintained and expanded in complete RPMI 1640. 2.2. Isolation of human DC from peripheral blood Human DC were isolated from peripheral blood using a MACS blood dendritic cell isolation kit (Miltenyi Biotech, Auburn, CA); DC were identi®ed as CD4 1 /HLA-DR1/ CD32/CD11b2/CD162 cells. Brie¯y, mononuclear cells were isolated from 200 ml of peripheral blood by density

centrifugation over Ficoll Paque and resuspended in 0.5% bovine serum albumin (BSA)/phosphate-buffered saline (PBS) containing 10 mg/ml of anti-Fc receptor antibody (Ab) and hapten-conjugated Ab against CD3 (T cells), CD11b (NK cells), and CD16 (monocytic cells). Following 10 min incubation, cells were washed and incubated for 15 min with micro-magnetic beads coated with anti-hapten Ab. The magnetically labeled cells were depleted from peripheral blood mononuclear cells (PBMC) and non-magnetic fractions were collected. Enriched cells were mixed with anti-CD4 Ab-bound magnetic beads and incubated for 30 min. After washing, cells were resuspended in 0.5% BSA/ PBS and applied to a RS 1 column, in which CD41 cells were retained by a magnetic ®eld. After removing CD42 cells by extensive washing, CD41 cells were recovered from the column and tested for purity by determining the number of HLA-DR1 cells by ¯ow cytometry. Typically, 2 £ 10 6 of 85±95% pure DC are recovered from 200 ml of peripheral blood. 2.3. PCR ampli®cation of whole spectrum mRNA To amplify the whole spectrum of mRNA expressed in DC, we employed the PCR-based Smart method (Clontech, Palo Alto, CA). Total RNA (approximately 1 mg) isolated from 1 £ 10 6 of blood DC was reverse-transcribed to cDNA from which full-length cDNA was selectively PCR-ampli®ed for 20 cycles using primers speci®c for mRNA at a 5 0 end cap site and a 3 0 end polyA tail. This method yielded 20 mg of ampli®ed cDNA from 0.2 mg of total RNA. 2.4. Cloning of a PCR fragment for human DECTIN-1 To selectively amplify cDNA fragments for human DECTIN-1 in the DC-derived cDNA, we employed PCR with guess-mers, which are oligonucleotides designed from sequences encoding for mouse dectin-1 CRD (GenBank Accession number: AF262985; Ariizumi et al., 2000a): 5 0 primer, 5 0 -TCCTGCTACCTGTTCTCC-3 0 (corresponding to nt 473±490); 3 0 primer, 5 0 -ATTGCGGGAAAGGCCTATCCA-3 0 (nt 623±643). An aliquot (about 200 ng) of the ampli®ed total cDNA was further PCR-ampli®ed using the Expand High Fidelity PCR system (Roche Diagnostics, Indianapolis, IN) for 35 cycles of a low stringency protocol: denaturing at 948C for 15 s, annealing at 358C for 45 s, and extension at 688C for 1 min. The resulting PCR products were separated on 2% agarose gel, transferred to membranes, and hybridized with 32P-labeled cDNA probe for mouse CRD under the low stringency conditions. The membranes were incubated in hybridization buffer (1 M NaCl, 0.05 M NaPO4, 5 £ Denhardt's, 10% dextran sulfate, 45% formamide, and 500 mg/ml of sheared salmon sperm DNA) containing 0.8 £ 10 6 cpm/ml of 32P-labeled probe. Following a 16 h incubation at 378C, the membranes were washed sequentially with 2 £ SSC/0.1% SDS at room temperature for 30 min and with 2 £ SSC/0.1% SDS at 428C for 30 min. The hybridized PCR fragment was extracted from agarose gel and subcloned

K. Yokota et al. / Gene 272 (2001) 51±60

into a pCR2.1 vector (Invitrogen, Carlsbad, CA) using the TA cloning method (Invitrogen). The nucleotide sequence of the fragment was determined by Taq dye deoxy termination cycle sequencing on an Applied Biosystems model 373A DNA sequencer (Applied Biosystems, Foster City, CA). 2.5. Isolation of the full-length cDNA for human DECTIN-1 Approximately 10 6 phage in a cDNA library (purchased from Clontech) prepared from human peripheral blood leukocytes was plated and plaque-hybridized with 32Plabeled PCR fragment for DECTIN-1. Plaques that showed strong hybridization signals were selected and cloned by two more rounds of screening. The isolated phage clones were converted into plasmid vectors, pTripleEX (Clontech) via cre-recombinase-mediated site-speci®c recombination at the loxP sites. BM25.8 E. coli expressing cre-recombinase were infected with positive phage clones and spread on carbenicillin-agar plates. Visible colonies were selected and expanded for large-scale DNA isolation. The nucleotide sequence of the full-length cDNA was determined at both sense and anti-sense strands by the above-automated sequencing of a series of deletion mutants produced by the Erase-a-base system (Promega, Madison, WI). 2.6. Northern blotting Two nylon membranes (Tissue Northern Blots) blotted with 2 mg of mRNA isolated from various human lymphoid and non-lymphoid organs were purchased from Clontech and hybridized in one bag according to the manufacturer's protocol. Brie¯y, membranes were hybridized in buffer (5£ SSPE, 10£ Denhardt's, 2% SDS, and 100 mg/ml sheared salmon sperm DNA) containing 1 £ 10 6 cpm/ml of 32Plabeled full-length cDNA for DECTIN-1 or b-actin. After a 16 h incubation at 428C, membranes were washed and autoradiographed. 2.7. Reverse transcriptase PCR and Southern blotting Reverse transcriptase (RT)-PCR was performed as described previously (Ariizumi et al., 1995). Total RNA (1 mg) isolated from different cell lines and freshly isolated blood DC was reverse-transcribed into cDNA and an aliquot (5%) was used for PCR ampli®cation using primers for human DECTIN-1: 5 0 primer, 5 0 -TTTGGGAATCCTATGCTTGGTAAT-3 0 (nt 201±224); 3 0 primer, 5 0 -TCTCCACCCTTCCTCTTACATTGA-3 0 (nt 790±813); or for bactin: 5 0 primer, 5 0 -ATCTGGCACCACACCTTCTACAATGAGCTGCG-3 0 (nt 294±325); 3 0 primer, 5 0 -CGTCATACTCCTGCTTGCTGATCCACATCTGC-3 0 (nt 1100± 1131). Following 25 cycles of PCR ampli®cation, the products were separated on 1.5% agarose gel, alkalineblotted on membranes, and hybridized with 32P-labeled cDNA probe for DECTIN-1 or b-actin.

53

2.8. Generation of anti-human DECTIN-1 Ab Anti-human DECTIN-1 peptide Ab was generated by immunizing rabbits with synthetic 19 aa peptide of SRNKENHSQPTQSSLEDS-C (Research Genetics, Huntsville, AL) (the carboxyl-terminal cysteine was attached for thiol coupling) corresponding to aa 86±103 in the neck domain of human DECTIN-1. After three rounds of immunization, serum was collected and subjected to af®nity puri®cation of peptide-speci®c Ab using the same 19-mer peptide as described before (Ariizumi et al., 2000a). 2.9. DNA transfection and immunoblotting of human DECTIN-1 protein The full-length cDNA for human DECTIN-1 was enzymatically excised from the hDec1-8 clone and ligated to a mammalian expression vector, pcDNA3.1(1) (Invitrogen). COS-1 cells were transfected with the resulting vector (pcDNA-hDec1-8) or an empty vector using FuGene 6 (Roche Diagnostics, Indianapolis, IN). At 3 days after transfection, the whole cell extracts were prepared from the COS-1 transfectants by lysis with 0.3% Triton X-100 in Dulbecco's PBS (2) for 15 min, followed by centrifugation for 20 min at 10,000 £ g. These samples were separated by 4±15% SDS-PAGE, transferred onto polyvinylidene ¯uoride membrane (Milipore, Bedford, MA), and then blotted with 1.7 mg/ml of af®nity-puri®ed rabbit anti-human DECTIN-1 or control rabbit IgG. After washing, the membrane was further blotted with horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG (Jackson Immuno Research Lab, West Grove, PA) and developed with the ECL system (Amersham, Piscataway, NY). 2.10. Immunostaining of human epidermis Human skin biopsies excised from normal healthy individuals were soaked in O.C.T. compound (Tissue-Tek, Miles Inc., Elkhart, IN) and snap-frozen in liquid nitrogen. Frozen specimens were sliced into 5 mm sections, and ®xed in acetone for 10 min, air-dried, and rehydrated in PBS. To block non-speci®c staining, the sections were incubated with calf serum at 238C for 5 min. The specimens were then incubated with 1.3 mg/ml of af®nity-puri®ed rabbit anti-human DECTIN-1 Ab and with monoclonal mouse IgG2b raised against CD1a (Becton Dickinson Immunocytometry System, San Jose, CA) at 48C overnight. After extensive washing with PBS, samples were incubated with ¯uorescein-conjugated goat F(ab 0 )2 anti-mouse IgG or with rhodamine-conjugated goat F(ab 0 )2 anti-rabbit IgG at 238C for 30 min. Immunostaining was observed under a ¯uorescence microscope using ®lters for either of these two colors. 2.11. Fluorescence in situ hybridization Fluorescence in situ hybridization (FISH) was used to map the DECTIN-1 gene on human chromosomes as

54

K. Yokota et al. / Gene 272 (2001) 51±60

described before (Heng et al., 1992). Brie¯y, phytohemagglutinin (PHA)-activated lymphocytes were treated with 0.18 mg/ml of BrdU to synchronize the cell population, then washed and cultured in a-MEM with 2.5 mg/ml of thymidine at 378C for 6 h. The cells were ®xed on slides that were subsequently air-dried, baked, and treated with RNase A. The slides were denatured in 70% formamide/ 2 £ SSC for 2 min at 708C and hybridized with biotinylated DECTIN-1 cDNA probe (1.6 kb) in buffer (50% formamide, 10% dextran sulfate). After overnight hybridization, slides were washed and detected as well as ampli®ed. FISH signals and a DAPI (4 0 -6-diamidino-2-phenylindole) banding pattern were recovered separately by taking photographs. Assignment of FISH mapping data on chromosomal bands was achieved by superimposing FISH signals on DAPIbanded chromosomes (Heng et al., 1992).

3. Results 3.1. Cloning of cDNA encoding human DECTIN-1 Because suf®cient numbers of human DC freshly isolated

from peripheral blood are not readily available for molecular biology techniques, we ampli®ed the whole spectrum of mRNA isolated from human DC using primers that bind speci®cally to cap sites and polyA tails of mRNA. We then ampli®ed the cDNA fragment encoding human DECTIN-1 using several sets of PCR primers (guessmers) for various short stretches of amino acid sequences in mouse dectin-1 CRD (e.g. aa 118±123, 129±134, 152± 157, 179±185, and 189±195), all of which are well conserved among C-type lectins (Ariizumi et al., 2000a). Among a number of PCR primer sets, guess-mers for aa 118±123 and for aa 179±185 generated three bands, one of which was clearly hybridized with the mouse CRD probe under the low (but not high) stringency conditions (data not shown). We then subcloned this PCR fragment, determined its nucleotide sequence, and found that it contained 201 bp with one reading frame encoding an amino acid sequence possessing 64.7% identity to the corresponding sequence in mouse dectin-1, including invariant amino acid residues highly conserved among C-type lectins (Fig. 1A). These results suggest that the PCR fragment is a part of the full-length cDNA of the human DECTIN-1 gene. Using the PCR fragment as a probe, a cDNA phage

Fig. 1. Comparison of amino acid sequences of human and mouse dectin-1. (A) Both human and mouse dectin-1 (GenBank Accession numbers AF313468 and AF262985, respectively) are membrane-integrated polypeptides with a type II con®guration consisting of (from their amino-termini) cytoplasmic domains (Cy, a ®rst line), transmembrane (Tr), and extracellular regions which contain neck domains (Ne) and carbohydrate recognition domains (Cr). Both ITAMs (closed circles) and invariant amino acid residues (asterisks) are conserved in human and mouse dectin-1. The amino acid sequence encoded by the PCR fragment ampli®ed with guess-mers and the sequence deleted in the isoform b (Fig. 6) are shown by a solid line and a dashed line, respectively, with arrowheads. (B) The amino acid sequence of human DECTIN-1 is compared structurally to that of mouse dectin-1 in each domain. Numbers between two sequences indicate amino acid sequence identity (%) in each domain and overall between human and mouse dectin-1.

K. Yokota et al. / Gene 272 (2001) 51±60

library prepared from human blood leukocytes was screened for full-length cDNA clones by plaque hybridization. Four independent phage clones were isolated, including one termed hDec1-8 which contained the longest cDNA insert. The hDec1-8 clone has a longest open reading frame spanning nt 55±798, which encodes a polypeptide of 247 amino acids with three additional residues; by comparison, mouse dectin-1 has 244 amino acids (Fig. 1). 3.2. hDec1-8-encoded polypeptide conserves structural features of mouse dectin-1 Our hDec1-8-encoded polypeptide is a novel transmembrane protein with a type II con®guration, consisting of (from the amino-terminus) a cytoplasmic domain (aa 1± 43), a transmembrane domain (aa 44±67), and an extracellular region containing a neck domain (aa 68±119) and CRD (aa 120±247) based on conserved invariant amino acid residues (10 of 13) in the C-type lectin superfamily (Spiess, 1990) (Fig. 2A). This segmentation of domains is almost

55

identical to that of mouse dectin-1. An immunoreceptor tyrosine-based activation motif (ITAM), Tyr-x-x-Leu (YxxL), required for internalization of many receptors (Reth, 1989; Vivier and Daeron, 1997), was also conserved in the cytoplasmic tail, suggesting that both this molecule and mouse dectin-1 are surface receptors that may transduce tyrosine-based signaling. In contrast, the polypeptide but not mouse dectin-1 contains in its CRD a QPD sequence characteristic of the CRD of galactose- and N-acetyl-galactosamine-binding C-type lectins (Drickamer, 1993). hDce18-encoded polypeptide has the highest homology (59.2%) with mouse dectin-1 among currently identi®ed molecules (GenBank database) and its CRD showed signi®cant homology with many members of the C-type lectin superfamily, including the oxidized LDL receptor (LOX-1) (38.3%), Ctype lectin-like receptors CLEC-1 and CLEC-2 (31.2 and 28.9%, respectively), NK receptors NKG2-B and NKG2-D (25.2 and 26.3%, respectively), DCIR (20.3%), and asialoglycoprotein receptors-1 and -2 (HL-1 and HL-2) (17.2 and 21.9%, respectively) (Fig. 2). It is noteworthy that all of

Fig. 2. Amino acid sequence alignment of members of the type II lectin family. Using the Clustal method in the Lasergene program (DNA Star, Madison, WI), the amino acid sequence of human DECTIN-1 CRD was aligned with those of human members of the type II lectin family, which showed the highest homology with human DECTIN-1. (A) CRD sequences were extracted from human DECTIN-1 (amino acid (aa) 120±247), oxidized LDL receptor (LOX-1) (GenBank Accession number: NP_002534, aa 144±273), CLEC-1 (AF200949, aa 134±280), CLEC-2 (AF124841, aa 99±229), NKG2-B (X54868, aa 101±215), NKG2-D (X54870, aa 99±216), asialoglycoprotein receptor-1 (HL-1) (A22509, aa 153±290), asialoglycoprotein receptor-2 (HL-2) (P07307, aa 177±311), and DCIR (CAB54001, aa 105±237). The QPD motif and invariant amino acid residues are marked with open circles and asterisks, respectively. Amino acid residues conserved in this group are shown as shaded boxes. (B) Based on their amino acid sequence identity scores, each member of the type II lectin superfamily is plotted in the phylogenic tree.

56

K. Yokota et al. / Gene 272 (2001) 51±60

contrast, mRNA was not detectable in non-lymphoid organs (e.g. heart, brain, placenta, lung, liver, skeletal muscle, kidney, and pancreas), thereby documenting lymphoidrestricted expression. To determine which cells express mRNA within lymphoid organs, total RNA isolated from monocyte (THP-1, U937, and HL-60), B cell (Raji), T cell (Jurkat), and NK cell (NKL) lines was ampli®ed for DECTIN-1 mRNA by RT-PCR and subsequent Southern blotting with the CRD probe. As expected, the PCR signal was detected in mRNA prepared from blood DC (Fig. 3B). It was also detected at lower levels in HL-60 monocyte and Raji B cell lines, at minimal levels in the two other monocyte lines, and at undetectable levels in the Jurkat T cell and NKL cell lines (Fig. 3B, long exposure). Like mouse dectin1, human DECTIN-1 mRNA is expressed preferentially by DC. 3.4. Expression of human DECTIN-1 protein We generated af®nity-puri®ed rabbit Ab raised against a synthetic oligopeptide of a neck domain in the human DECTIN-1. By immunoblotting, this Ab detected a speci®c band of 33 kDa in COS-1 cells transfected with the fulllength cDNA for human DECTIN-1 but not with an empty

Fig. 3. Expression of human DECTIN-1 mRNA. (A) mRNAs (2 mg) isolated from the indicated lymphoid organs (left membrane) and from non-lymphoid organs (right membrane) were examined for expression of DECTIN-1 or b-actin mRNA by Northern blotting. (B) Total RNAs were extracted from freshly isolated blood DC and THP-1, U937, HL-60 (all monocytes), Raji (B cell), Jurkat (T cell), and NKL (NK cell) lines. The RNA was reverse-transcribed to cDNA, and subjected to PCR ampli®cation with primers for DECTIN-1 or b-actin. The resulting PCR products were Southern hybridized with the respective cDNA probe. The relative intensities of DECTIN-1-hybridized signals were estimated by short or long exposure of hybridized membranes to X-ray ®lms.

these surface receptors have type II con®gurations. In sum, hDec1-8-encoded polypeptide is a new member of the Ctype lectin superfamily that closely resembles mouse dectin1 in structure and in functional motifs. We conclude that the hDec1-8-encoded polypeptide is the human homologue for mouse dectin-1. 3.3. Human DECTIN-1 mRNA is expressed preferentially by DC Using a cDNA insert of the hDec1-8 clone, we examined the tissue distribution of DECTIN-1 mRNA by Northern blotting. Fig. 3A shows that the DECTIN-1 gene is expressed as two transcripts, 4.0 and 2.5 kb. Both transcripts were expressed abundantly in peripheral blood leukocytes and minimally in other lymphoid organs (e.g. spleen, lymph node, thymus, appendix, and bone marrow). In sharp

Fig. 4. Expression of human DECTIN-1 protein in COS-1 transfectants. Whole cell extracts were prepared from COS-1 cells transfected with cDNA encoding the full-length human DECTIN-1 or an empty vector. These samples were analyzed for human DECTIN-1 protein expression by immunoblotting with af®nity-puri®ed rabbit Ab raised against a synthetic peptide of human DECTIN-1. An arrow indicates the molecular size of human DECTIN-1 protein.

K. Yokota et al. / Gene 272 (2001) 51±60

vector (Fig. 4). This estimated molecular size is larger than that predicted from the deduced amino acid sequence (28 kDa). This discordance may be due to N-linked glycosylation of a site (aa 91) in the neck domain (Fig. 1A). Despite the prediction of a similar molecular weight, the size of mouse dectin-1 protein (43 kDa) is larger than human DECTIN-1, presumably due to a difference in the numbers of N-linked saccharide chain: one glycosylation site for the human and two sites for mouse dectin-1 (Ariizumi et al., 2000a). Nonetheless, human DECTIN-1 is most likely a glycoprotein of 33 kDa with a single N-linked saccharide chain. To examine expression of human DECTIN-1 protein in epidermal LC, the skin biopsies derived from normal healthy individuals and patients with psoriasis were doubly stained with Abs against the human DECTIN-1 (red ¯uorescence) and CD1a (green ¯uorescence), a surface marker for human LC. In both normal and psoriatic human epidermis, most red ¯uorescent cells were also stained with green ¯uorescence (doubly stained cells are shown in yellow) (Fig. 5). These results document that human DECTIN-1 protein is expressed selectively by LC, consistent with similar ®ndings for the mouse dectin-1 mRNA (Ariizumi et al., 2000a).

57

3.5. Identi®cation of a truncated form of DECTIN-1 mRNA In Fig. 3B, RT-PCR produced two bands: an upper band (612 bp) predicted from the cDNA sequence of DECTIN-1, and a second band migrating at approximately 500 bp. Since the primers were designed to amplify a nucleotide sequence encoding a transmembrane domain and a CRD, the second band may be an isoform with deletion of a 100 bp within these domains. To identify the deleted region, we cloned the PCR fragment corresponding to the second band, and subsequently the full-length cDNA containing the sequence of the second band was also isolated, and their nucleotide sequences were determined. These analyses demonstrated that the second PCR band represents a part of the transcript (termed Tb) encoding a polypeptide of 202 amino acids with deletion of aa 68±113, a region corresponding to the neck domain (aa 68±119) hypothesized to be required for dimerization of C-type lectins and responsible for spatial ¯exibility of the CRD (Spiess, 1990) (Fig. 6A). On the other hand, we have not found RNA heterogeneity for mouse dectin-1. To examine whether such heterogeneity exists in the mouse gene, mRNA isolated from spleen DC was ampli®ed by PCR with the corresponding mouse primers. An additional PCR fragment with a 100 bp smaller molecular weight was consistently detected. We also isolated a mouse dectin-1 cDNA clone (termed Tb) with almost the same deletion (aa 68±112) in the neck domain (aa 68±118). These results suggest that human and mouse dectin-1 genes employ similar mechanisms for generating the truncated form of mRNA and likely have identical segmentation of exons and introns. 3.6. Tb is produced by alternative splicing To delineate mechanisms for the generation, we searched for the DECTIN-1 genome in the human genome database provided by the National Center for Biotechnology Information (NCBI). Most of the DECTIN-1 cDNA sequence was found in a yeast arti®cial chromosome (YAC) clone (AC024224), divided into seven exons (Fig. 6B): except for the ®rst exon, all other exons encode amino acid sequences. Exon 2 encodes for part of the cytoplasmic domain, exon 3 for the other part of the cytoplasmic domain and the entire transmembrane domain, exon 4 primarily for a neck domain, and exons 5±7 for CRD. In fact, the DECTIN-1 Ta transcript for the full-length isoform contains all exon sequences, whereas the Tb transcript for the isoform b lacks the entire exon 4, thus documenting that Tb is generated by alternative splicing of exon 4.

Fig. 5. Human DECTIN-1 protein is expressed predominantly by LC in epidermis. Frozen sections prepared from human normal and psoriatic skin were stained with af®nity-puri®ed rabbit anti-human DECTIN-1 Ab (Dec1, red ¯uorescence) and with monoclonal mouse IgG2b against CD1a (CD1a, green ¯uorescence), a surface marker for LC. Staining was observed under a ¯uorescence microscope with a ®lter for red or green ¯uorescence, and their images were taken by a computer. LC-speci®c expression of DECTIN-1 protein was examined by superimposing these two images (Dec1/CD1a).

3.7. Human DECTIN-1 gene is located in the vicinity of NK gene complex Further, we mapped the DECTIN-1 gene on human chromosomes by FISH and DAPI banding (Fig. 7). The FISH dots clustered on chromosome 12 between p13.2 and p12.3, close to the NK gene complex (NKC) (12p13.1 to p13.2)

58

K. Yokota et al. / Gene 272 (2001) 51±60

which contains a number of genes encoding type II lectin receptors expressed by NK cells (Brown et al., 1997; Sobanov et al., 1999) (e.g. CD69, NKG2, and NKRP1 families).

4. Discussion 4.1. Comparison of human and mouse dectin-1 Using degenerative PCR ampli®cation coupled with cDNA cloning, we have identi®ed the human homologue of mouse dectin-1. Human DECTIN-1 displays several conserved features with mouse dectin-1. It is a new member of the C-type lectin superfamily and a type II receptor that contains an ITAM (YxxL motif). DECTIN-1 mRNA is preferentially expressed by DC and generates its alternatively spliced form with deletion of a neck domain. However, human DECTIN-1 differs from mouse dectin-1 in several respects: it produces two major mRNA transcripts, is expressed predominantly by peripheral blood leukocytes, and includes a QPD motif required for binding of the asialoglycoprotein receptors-1 and -2 (HL-1 and HL-2) to galac-

tose and N-acetyl-galactosamine (Drickamer, 1993). Finally, the DECTIN-1 is a less glycosylated protein than the mouse dectin-1. Previously, we showed that mouse dectin-1 mRNA is expressed predominantly in thymus and spleen tissues (Ariizumi et al., 2000a), but we did not test expression in peripheral blood because of the unavailability of suf®cient cell numbers for Northern blotting. Very recently, we detected stronger PCR signals for mouse dectin-1 in blood leukocytes than in thymus and spleen (data not shown). Thus, abundant mRNA expression by peripheral leukocytes is also likely for the mouse dectin-1 gene. The mouse dectin-1 gene is expressed as a single transcript of 3.2 kb, whereas the human gene produces two mRNA species of 4.0 and 2.0 kb. Precedent for a similar discordance has been shown for the c-Jun gene (Sten et al., 1992). The QPD motif was originally identi®ed to be required for recognizing speci®c carbohydrates by HL-1 and HL-2. The QPD in DECTIN-1 is located in the amino-terminal half of the CRD, whereas those in HL-1 and HL-2 are in the carboxyl-terminal half (Fig. 4). Since the relative position of the QPD motif in the CRD frame may be critical for recognizing a particular

Fig. 6. Identi®cation of human DECTIN-1 isoforms and mechanisms for the generation of different transcripts. (A) The cDNA clone for isoform b (GenBank Accession number: AF313469) was isolated, and the deduced amino acid sequence was aligned with isoform a (the full-length, encoded by an original clone, hDec1-8). The deleted regions are well conserved in both human and mouse dectin-1. (B) The structure of the human DECTIN-1 genome (AC024224) is shown and compared to that of the Ta transcript for isoform a. The coding sequence of human DECTIN-1 cDNA is divided into six exons (indicated with boxes), each encoding the different peptide segment of human DECTIN-1. The bottom diagrams illustrate a possible mechanism for generation of Tb lacking the exon 4 encoding primarily for a neck domain. The 3 0 end sequence of exon 7 has not been clearly identi®ed (shown by an open box with dashed lines).

K. Yokota et al. / Gene 272 (2001) 51±60

59

carbohydrate structure (Drickamer, 1993), it is unlikely that the QPD in human DECTIN-1 has a functional role in carbohydrate recognition.

exhibit none or a signi®cantly reduced ability to capture speci®c ligands, thereby serving to block co-stimulatory function of dectin-1 molecules on DC.

4.2. Alternative splicing of DECTIN-1 mRNA

4.3. Structural and functional relationships between DC and NK lectin receptor genes

We also found that both isoforms for human and mouse dectin-1 lack almost the entire neck domain. This alternative splicing was also true for mouse dectin-2, a type II C-type lectin receptor that we previously identi®ed using the same subtractive cDNA cloning strategy (Ariizumi et al., 2000b). Since such neck domain-defective isoforms have not been found for other C-type lectins, this alternative splicing may be unique for dectin-1 and dectin-2 genes. Previous studies with asialoglycoprotein receptors (HL-1 and HL-2) have demonstrated the neck domain to be required for disul®de-linked dimerization and spatial ¯exibility of the CRD to increase accessibility to ligands (e.g. carbohydrates) (Drickamer, 1993). However, unlike these lectins, both human and mouse dectin-1 do not contain cysteine residues in their neck domains, thereby suggesting that the neck domain in the dectin-1 functions only as a spacer and also that dectin-1 is expressed as a monomer on DC. We therefore speculate that neck domain-defective dectin-1 may

As noted earlier, human DC express two subfamilies of the C-type lectin receptors: type I integral membrane proteins, and type II integral membrane proteins, such as DC-SIGN, Langerin, DCIR, CLEC-1, CLEC-2, and DECTIN-1. DC-SIGN and Langerin are expressed selectively by DC, whereas the other four lectins are more widely expressed by cells other than DC. Moreover, CRDs of these latter four proteins show similar amino sequences. To delineate physical relationships between these genes on human chromosomes, we searched for these four genes in the human genome database (NCBI). All four genes are found in close proximity to the chromosomal location of NKC (12p13.1 to p13.2) (data not shown) where genes encoding NK type II lectin receptors have been mapped. Thus, these DC lectin receptor genes may form a multigene family located close to the NKC. Importantly, this putative family closely resembles the NK gene family not only in their high homology among their CRDs but also in their inclusion of tyrosine-based signaling motifs (i.e. ITAMs in DECTIN-1, CLEC-1, and CLEC-2 and an immunoreceptor tyrosine-based inhibitory motif (ITIM) in DCIR). By contrast, their expression pro®les are completely different: the DC receptor family is expressed by DC but not by NK cells (CLEC-2 is expressed in NK cells but at lower levels compared to DC; Colonna et al., 2000), whereas NK receptors are expressed by NK cells but not by DC. With respect to function, NK receptors recognize MHC class I molecules and carbohydrates, whereas speci®c ligands for the DC receptors have not yet identi®ed. In sum, DC receptor genes may arise from gene duplication of NK receptor genes, acquiring diversity in gene regulation and function during evolution. In conclusion, we have isolated cDNA clones that encode human DECTIN-1, a new member of the DC type II lectin receptor family. Further study of the structural and functional relationships among the gene products of this gene family may yield insights into the molecular mechanisms responsible for the potency of human DC in antigen processing and presentation. Acknowledgements

Fig. 7. DECTIN-1 gene mapped on chromosome 12p13.1 to p12.3. Mapping of the DECTIN-1 gene on human chromosome was performed by the FISH method. Chromosomes in human blood leukocytes were stained (A) with biotinylated hDec1-8 cDNA probe (1.6 kb) to determine the location of hybridized signals and (B) with DAPI to identify speci®c chromosomes. In (C), each dot representing the double FISH signals was mapped on human chromosome 12 between p13.1 and p12.3 by superimposing FISH images and DAPI-banding patterns.

We thank Dr Michael J. Robertson for kindly providing the NKL line, and Dr Pociano D. Cruz Jr. for critical reading of this manuscript. We are also grateful to Dale Edelbaum for excellent technical assistance and Susan Millberger for secretarial assistance in the preparation of this manuscript. This work was supported by a research grant from the National Institutes of Health (RO1-AR44189).

60

K. Yokota et al. / Gene 272 (2001) 51±60

References Ariizumi, K., Meng, Y., Bergstresser, P.R., Takashima, A., 1995. IFNgamma-dependent IL-7 gene regulation in keratinocytes. J. Immunol. 154, 6031±6039. Ariizumi, K., Shen, G.-L., Ritter III, R., Kumamoto, T., Edelbaum, D., Morita, A., Bergstresser, P.R., Takashima, A., 2000a. Identi®cation of a novel, dendritic cell-associated molecule, Dectin-1, by subtractive cDNA cloning. J. Biol. Chem. 275, 20157±20167. Ariizumi, K., Shen, G.-L., Shikano, S., Ritter III, R., Zukas, P., Edelbaum, D., Morita, A., Takashima, A., 2000b. Cloning of a second dendritic cell-associated C-type lectin (Dectin-2) and its alternatively spliced isoforms. J. Biol. Chem. 275, 11957±11963. Bates, E.E., Fournier, N., Garcia, E., Valladeau, J., Durand, I., Pin, J.-J., Zurawski, S.M., Patel, S., Abrams, J.S., Lebecque, S., Garrone, P., Saeland, S., 1999. APCs express DCIR, a novel C-type lectin surface receptor containing an immunoreceptor tyrosine-based inhibitory motif. J. Immunol. 163, 1973±1983. Brown, M.G., Scalzo, A.A., Matsumoto, K., Yokoyama, W.M., 1997. The natural killer gene complex: a genetic basis for understanding natural killer cell function and innate immunity. Immunol. Rev. 155, 53±65. Colonna, M., Samaridis, J., Angman, L., 2000. Molecular characterization of two novel C-type lectin-like receptors, one of which is selectively expressed in human dendritic cells. Eur. J. Immunol. 30, 697±704. Drickamer, K., 1993. Ca 21-dependent carbohydrate-recognition domains in animal proteins. Curr. Opin. Struct. Biol. 3, 393±400. Geijtenbeek, T.B., Torensma, R., van Vliet, S.J., van Duijnhoven, G.C., Adema, G.J., van Kooyk, Y., Figdor, C.G., 2000. Identi®cation of DCSIGN, a novel dendritic cell-speci®c ICAM-3 receptor that supports primary immune responses. Cell 100, 575±585. Heng, H.H.Q., Squire, J., Tsui, L.-C., 1992. High resolution mapping of mammalian genes by in situ hybridization to free chromatin. Proc. Natl. Acad. Sci. USA 89, 9509±9513. Jiang, W., Swiggard, W.J., Heu¯er, C., Peng, M., Mirza, A., Steinman, R.M., Nussenzweig, M.C., 1995. The receptor DEC-205 expressed by dendritic cells and thymic epithelial cells is involved in antigen processing. Nature 375, 151±155.

Lanzavecchia, A., 1993. Receptor-mediated antigen uptake and its effect on antigen presentation to class II-restricted T lymphocytes. Annu. Rev. Immunol. 8, 773±793. Reth, M., 1989. Antigen receptor tail clue. Nature 338, 383±384. Robertson, M.J., Cochran, K.J., Cameron, C., Le, J.-M., Tantravahi, R., Ritz, J., 1996. Characterization of a cell line, NKL, derived from an aggressive human natural killer cell leukemia. Exp. Hematol. 24, 406± 415. Sobanov, Y., Glienke, J., Brostjan, C., Lehrach, H., Francis, F., Hofer, E., 1999. Linkage of the NKG2 and CD94 receptor genes to D12S77 in the human natural killer gene complex. Immunogenetics 49, 99±105. Spiess, M., 1990. The asialoglycoprotein receptor: a model for endocytic transport receptors. Biochemistry 29, 10009±10018. Steinman, R.M., 1991. The dendritic cell system and its role in immunogenicity. Annu. Rev. Immunol. 9, 271±296. Sten, B., Angel, P., van Dam, H., Ponta, H., Herrlich, P., van der Eb, A., Rahmsdorf, H.J., 1992. Ultraviolet-radiation induced c-jun gene transcription: two AP-1 like binding sites mediate the response. Photochem. Photobiol. 55, 409±415. Valladeau, J., Ravel, O., Dezutter-Dambuyant, C., Moore, K., Kleijmeer, M., Liu, Y., Duvert-Frances, V., Vincent, C., Schmitt, D., Davoust, J., Caux, C., Lebecque, S., Saeland, S., 2000. Langerin, a novel C-type lectin speci®c to Langerhans cells, is an endocytic receptor that induces the formation of Birbeck granules. Immunity 12, 71±81. Vestweber, D., Blanks, J.E., 1999. Mechanisms that regulate the function of the selectins and their ligands. Physiol. Rev. 79, 181±213. Vivier, E., Daeron, M., 1997. Immunoreceptor tyrosine-based inhibition motifs. Immunol. Today 18, 286±291. Xu, S., Ariizumi, K., Caceres-Dittmar, G., Edelbaum, D., Hashimoto, K., Bergstresser, P.R., Takashima, A., 1995a. Successive generation of antigen-presenting, dendritic cell lines from murine epidermis. J. Immunol. 154, 2697±2705. Xu, S., Ariizumi, K., Edelbaum, D., Bergstresser, P.R., Takashima, A., 1995b. Cytokine-dependent regulation of growth and maturation in murine epidermal dendritic cell lines. Eur. J. Immunol. 25, 1018±1024.