Molecular Cloning of a Novel Human Collectin from Liver (CL-L1)*

Collectins are a C-lectin family with collagen-like sequences and carbohydrate recognition domains. These proteins can bind to carbohydrate antigens of microorganisms and inhibit their infection by direct neutralization and agglutination, the activation of complement through the lectin pathway, and opsonization by collectin receptors. Here we report the cloning of a cDNA encoding human collectin from liver (CL-L1 (collectin liver 1)) that has typical collectin structural characteristics, consisting of an N-terminal cysteine-rich domain, a collagen-like domain, a neck domain, and a carbohydrate recognition domain. The cDNA has an insert of 831 base pairs coding for a protein of 277 amino acid residues. The deduced amino acid sequence shows that this collectin has a unique repeat of four lysine residues in its C-terminal area. Northern blot, Western blot, and reverse transcription-polymerase chain reaction analyses showed that CL-L1 is present mainly in liver as a cytosolic protein and at low levels in placenta. More sensitive analyses by reverse transcription-polymerase chain reactions showed that most tissues (except skeletal muscle) have CL-L1 mRNA. Zoo-blot analysis indicated that CL-L1 is limited to mammals and birds. A chromosomal localization study indicated that the CL-L1 gene localizes to chromosome 8q23-q24.1, different from chromosome 10 of other human collectin genes. Expression studies of fusion proteins lacking the collagen and N-terminal domains produced in Escherichia coli affirmed that CL-L1 binds mannose weakly. CL-L1 and recombinant CL-L1 fusion proteins do not bind to mannan columns. Analysis of the phylogenetic tree of CL-L1 and other collectins indicated that CL-L1 belongs to a fourth subfamily of collectins following the mannan-binding protein, surfactant protein A, and surfactant protein D subfamilies including bovine conglutinin and collectin-43 (CL-43). These findings indicate that CL-L1 may be involved in different biological functions.

hydrate recognition domain (CRD) 1 (1). These lectins are found in vertebrates from birds to humans (2). There are three groups of collectins: the MBP group, which includes MBP-A and MBP-C (3); the SP-A group (4); and the SP-D group (5), which includes bovine conglutinin (6) and CL-43 (7). MBP can destroy bacteria through activation of the complement pathway (8) or by opsonization by collectin receptors (9). Conglutinin is a ␤-inhibitor of influenza A viruses that exhibits hemagglutination inhibition and neutralization activities (10,11). SP-A enhances the phagocytosis of bacteria by macrophages (12) and opsonizes herpes simplex virus (13). SP-D causes agglutination of bacteria (14) and exhibits hemagglutination inhibition against influenza A virus (15). These data indicate that collectins play an important role in immunoglobulin-independent host defense (16).
The isolation and functional characterization of novel collectins in addition to those described above might provide further insights into their functions. Here we report the cloning and preparation of recombinant fusion proteins and the characterization of human CL-L1 (collectin liver 1), a new member of the collectin family. CL-L1 is expressed mainly in liver, placenta, and adrenal gland and is expressed ubiquitously in most tissues, except skeletal muscle. Surprisingly, this new collectin is a cytosolic protein, although other all collectins are secreted proteins.
Generation of a Probe for Screening by Polymerase Chain Reaction-Screening an expressed sequence tag (EST) data base for potential new collectin genes revealed a novel gene in EST clone R29493. The partial clone (F1-1006D) from a fetal liver cDNA was kindly provided by Dr. Hee-Sup Shin (Pohang Institute of Science and Technology) and used to screen a human liver cDNA library for full-length cDNAs by plaque hybridization. To generate a digoxigenin-labeled cDNA probe, we used the polymerase chain reaction (PCR). Primers amplifying the cDNA probe were synthesized based on the 5Ј-and 3Ј-end nucleotide sequences of the insert in clone F1-1006D. The primers synthesized were 5Ј-GGCCAACACACTCATCGC-3Ј for the reverse primer and 5Ј-TTACTTTTTTCTTCTTG-3Ј for the forward primer. PCR was carried out using a PCR DIG labeling kit (Roche Molecular Biochemicals). The reaction mixture (50 l) consisted of 10 mM Tris-HCl, pH 8.3; 50 mM KCl; 1.5 mM MgCl 2 ; 200 mM each dATP, dCTP, and dGTP and 130 mM dTTP (Takara Shuzo Co., Ltd., Tokyo, Japan); 70 mM digoxigenin-11-dUTP; 1.25 units of Taq DNA polymerase; a 1 M concentration of each primer; and 20 ng of cDNA clone F1-1006D. PCR was performed for 30 cycles in TaKaRa PCR Model 480 thermal cycler (Takara Shuzo Co., Ltd.), with each cycle consisting of denaturation for 45 s at 95°C, annealing for 1 min at 60°C, and extension for 2 min at 72°C. The PCR product was electrophoresed on a 1% (w/v) agarose gel (Wako Pure Chemical Industries) and then extracted from the gel using a Sephaglas BandPrep kit (Amersham Pharmacia Biotech).
Determination of a cDNA Encoding CL-L1 by Screening a Human Liver cDNA Library and "Cap Site Hunting"-A phage library was screened essentially as described previously (17). In brief, ϳ1 ϫ 10 6 plaque units of a human liver gt11 cDNA library (CLONTECH) were plated with E. coli Y1090r Ϫ and incubated at 42°C for 4 h. Nylon filters (Nytran 13N, Schleicher & Schü ll) were prehybridized in hybridization buffer (5ϫ SSC, 1% blocking reagent (Roche Molecular Biochemicals), 0.1% N-lauroylsarcosine, and 0.02% SDS) for 1 h at 68°C and then hybridized for 16 h at 55°C with a digoxigenin-labeled probe in hybridization buffer. The filters were washed twice for 5 min in 2ϫ SSC and 0.1% SDS at room temperature and then twice for 15 min in 0.5ϫ SSC and 0.1% SDS at 55°C. The hybridized probe was detected by 30 min of incubation at room temperature with alkaline phosphatase-conjugated anti-digoxigenin antibody (Fab; Roche Molecular Biochemicals) diluted 1:5000. The enzyme-catalyzed color reaction was carried out using a nitro blue tetrazolium salt/5-bromo-4-chloro-3-indolyl phosphate system (Wako Pure Chemical Industries) in buffer consisting of 100 mM Tris-HCl, pH 9.5, 100 mM NaCl, and 50 mM MgCl 2 . The cDNA inserts in the positive clones were amplified using the primers described above and then directly subcloned in the pCR2.1 vector of a TA cloning kit (Invitrogen). The subclones were sequenced using an Autoread DNA sequencing kit and an A.L.F. Autosequencer (Amersham Pharmacia Biotech).
To identify the sequence including the transcriptional start site, we took the cDNA including the transcriptional start site from the Cap Site cDNA TM (Nippon Gene Co., Ltd., Tokyo) of human liver by nested PCR (18). This procedure is named cap site hunting (18). The primer set for the first PCR was 5Ј-CAAGGTACGCCACAGCGTATG-3Ј (primer 1RC2; Nippon Gene Inc., Ltd.) and 5Ј-TCTTCAGTTTCCCTAATCCC-3Ј (primer PR1), synthesized by the manufacturer indicated. The primer set for the second PCR was 5Ј-GTACGCCACAGCGTATGATGC-3Ј (primer 2RC2; Nippon Gene Co., Ltd.) and 5Ј-CATTCTTGACAAACTTCA-TAG-3Ј (primer PR2), synthesized by the manufacturer indicated. The reaction mixture (50 l) consisted of long and accurate (LA) PCR Buffer II (Mg 2ϩ -free); 2.5 mM MgCl 2 ; 200 M each dATP, dCTP, dGTP, and dTTP; 1 l of Cap Site cDNA TM from human liver; 1.25 units of TaKaRa LA Taq DNA polymerase (Takara Shuzo Co., Ltd.); and 0.5 M each primer 1RC2 and PR1 for the first PCR and primer 2RC2 and PR2 for the second PCR. The first PCR was performed for 35 cycles in the TaKaRa PCR Model 480 thermal cycler, with each cycle consisting of denaturation for 20 s at 95°C, annealing for 20 s at 60°C, and extension for 20 s at 72°C. The second PCR was performed for 25 cycles in same buffer using 1 l of the first PCR products as a probe. The final PCR products were extracted from the agarose gel after gel electrophoresis and directly subcloned into the pT7Blue vector (Novagen). The subclones were sequenced using the Autoread DNA sequencing kit and the A.L.F. Autosequencer.
Southern Blot Analysis-Genomic DNA was purified from human placenta by a standard method. Routinely, 4 g of genomic DNA was completely digested with restriction enzymes, separated on 0.7% agarose gel, and vacuum-transferred to the Nytran 13N nylon filters. Blots were prehybridized in ExpressHyb hybridization solution (CLON-TECH) at 68°C for 30 min and then hybridized for 1 h at 68°C with a 10 ng/ml concentration of the digoxigenin-labeled probe corresponding to a fragment of the CRD in the same buffer as used for prehybridization. The filters were washed for 5 min in 2ϫ SSC and 0.1% SDS at room temperature and then for 15 min in 0.2 ϫ SSC and 0.1% SDS at 68°C. The hybridized probe was detected by a chemiluminescent technique (Roche Molecular Biochemicals) as described by the manufacturer.
Zoo-blot analysis was performed with the modified method described above using completely EcoRI-digested genomic DNAs (5 g) from human, rhesus monkey, cow, dog, rabbit, rat, mouse, chicken, and yeast (Saccharomyces cerevisiae). Prehybridization, hybridization, and washing were performed at 58°C for the same time periods. The probe (which corresponded to the CRD) for the zoo-blot analysis was made with the PCR DIG labeling kit.
Northern Blot and RT-PCR Analyses-A human multiple-tissue Northern blot membrane was purchased from CLONTECH. It contained 2 g of poly(A) ϩ RNAs from human heart, brain, placenta, lung, liver, skeletal muscle, kidney, and pancreas. The membranes were prehybridized at 65°C for 3 h in a solution containing 5ϫ SSC, 10ϫ Denhardt's solution, 10 mM sodium phosphate, pH 6.5, 0.5% SDS, 50% formamide, and 0.1 mg/ml denatured salmon sperm DNA. Hybridization was then performed for 18 h at 65°C with RNA synthesized in vitro and labeled with digoxigenin using a DIG RNA labeling kit (Roche Molecular Biochemicals). The template for the RNA probe was a whole cDNA insert subcloned into pSPT18. The filters were washed twice for 5 min in 2ϫ SSC and 0.1% SDS at room temperature and then for 15 min in 0.1ϫ SSC and 0.1% SDS at 68°C. The hybridized probe was detected as described above.
Reverse transcription was carried out using total RNAs (1 g) from brain, heart, kidney, liver, lung, trachea, bone marrow, colon, small intestine, spleen, stomach, thymus, mammary gland, prostate, skeletal muscle, testis, uterus, placenta, adrenal gland, pancreas, salivary gland, and thyroid. Oligo(dT)-adaptor primers (RNA LA PCR kit (avian myeloblastosis virus), Version 1.1, Takara Shuzo Co., Ltd.) were used for the reverse transcription reaction. The reverse transcription products were amplified by 28 or 35 cycles of PCR using degenerate first primers (0.2 g), TaKaRa LA Taq polymerase (1.25 units), and reverse transcription reaction products in a TaKaRa PCR Model MP thermal cycler. Nested PCRs were performed using the PCR products as a probe after 35 cycles of RT-PCR. The primer set for the first PCR was 5Ј-TAGCAAATACGTAGGATGAG-3Ј for the reverse primer and 5Ј-CCA-CAGCAATGAATGGCTTT-3Ј for the forward primer. The inner primer set for nested PCR was 5Ј-TTACTTTTTCTTCTTGATGA-3Ј for the reverse primer and 5Ј-ATGAATGGCTTTGCATCCTT-3Ј for the forward primer. These primers were located in the neck and CRD regions, which spanned an intron in genomic DNA. Amplicons were separated on 1.0% agarose gels.
Chromosomal Localization of the CL-L1 Gene-The ϳ10-kilobase pair human CL-L1 genomic DNA fragment from the neck domain to the CRD, biotin-labeled with a nick translation kit (Roche Molecular Biochemicals), was used as a probe. The gene was localized by fluorescence in situ hybridization and PCR analysis using DNAs from human monochromosomal fusion cells kindly provided by Dr. Hashimoto (National Institute of Infectious Diseases). Map position was determined by inspection of fluorescent signals on 4,6-diamidino-2-phenylindole-stained chromosomes. 25 metaphase preparations were analyzed.
Expression of Neck and CRD Fragments of Cloned cDNA in E. coli-The recombinant fusion protein with the maltose-binding protein of E. coli was expressed using the expression vector pMAL-c2 system (New England Biolabs Inc.). E. coli XL1-Blue, which carries the plasmid containing the proper insert of neck and CRD fragments as described previously (19), was grown to A 600 nm ϳ 0.5 in 200 ml of LB medium supplemented with 0.5% glucose. After addition of isopropyl-␤-D-thiogalactopyranoside to a final concentration of 1 mM, the culture was incubated for an additional 3 h, and the cells were harvested. The cell pellets were suspended in 10 ml of lysis buffer and lysed by sonication (15 s, 70% output, 10 times). After centrifugation at 9000 ϫ g for 30 min, the supernatant was diluted with 5 volumes of column buffer/T and applied to an amylose resin column (New England Biolabs Inc.). The column was washed with column buffer/T and column buffer, and the recombinant fusion protein (CL-L1-CRDmal) was eluted with column buffer containing 10 mM maltose. The eluate was dialyzed against three changes of 1000 volumes of TBS/C and used for further characterization. To obtain another truncated CL-L1 protein (CL-L1-CRDhis), a new expression vector (pPLH3) was constructed and used. This vector consisted of a P L-Trp fusion promoter and histidine hexamer coding sequence just downstream from the ATG initiation codon. Recombinant proteins expressed by this vector system will have N-terminal MHHH-HHH sequences. E. coli GI724, containing the plasmid with the neck and CRD fragments, was grown at 30°C to A 600 nm ϳ 0.5 in 200 ml of induction base medium with glucose. After addition of tryptophan to a final concentration of 0.1 mg/ml, the culture was incubated for 3 h at 37°C, and the cells were harvested. The cells were suspended in 20 ml of lysis buffer A and lysed by sonication (15 s, 70% output, 10 times). After centrifugation at 9000 ϫ g for 30 min, the supernatant was incubated with nickel-nitrilotriacetic acid-agarose (QIAGEN Inc.) for 15 min, and the gel was loaded onto a column. The column was washed with column buffers B and C. The histidine-tagged recombinant protein was eluted with column buffers D and E. The eluate was dialyzed against 1000 volume of TBS, followed by two changes of 1000 volumes of TBS/C. This CL-L1 fusion protein (CL-L1-CRDhis) was used to produce antisera in New Zealand White rabbits. Purification and identification of the recombinant CL-L1 CRDs were confirmed by SDS-PAGE and Western blotting using the rabbit anti-CL-L1 CRD serum.
Another characterization was performed using a sugar-blot method (20). Recombinant CL-L1 CRDs, maltose-binding protein, and recombinant human MBP (1 g of each) were dissolved in SDS sample buffer, separated by SDS-PAGE (10 -20% gradient polyacrylamide gel), and transferred to BioBlot-NC membranes (Coster Co.) by standard procedures. After treating the membranes with TBS/TC (20 mM Tris-HCl, 140 mM NaCl, 0.1% Triton X-100, and 5 mM CaCl 2 ) with or without EDTA (10 mM), they were incubated with the ␣-D-mannose BP-probe alone or together with EDTA (10 mM) or mannan (100 g/ml) at 4°C for 60 min, washed with TBS/TC, and then incubated with the VEC-TASTAIN Elite ABC kit in TBS/TC for 60 min. After washing in TBS/ TC, the membranes were stained with 3,3Ј5,5Ј-tetramethylbenzidine substrate solution.
Western Blotting of Liver Tissue Extracts and Immunofluorescence Analyses in Primary Hepatocyte Cells-Cytosolic and microsomal fractions from human liver were prepared as described previously (21). Solutions of recombinant CL-L1-CRDhis, CL-L1-CRDmal, and maltosebinding proteins were dissolved with SDS sample buffer in the presence of 2-mercaptoethanol and applied to a 10 -20% gradient polyacrylamide gel. After electrophoresis, gels were transferred to BioBlot-NC membranes. The blots were blocked with Block Ace in TBS containing 0.1% Triton X-100 and then treated with a 1:1000 dilution of rabbit anti-CL-L1-CRDhis serum or rabbit anti-human MBP serum. The bound antibodies were visualized using alkaline phosphatase-conjugated goat anti-rabbit IgG (Chemicon International, Inc.) diluted 1:5000 and 5-bromo-4-chloro-3-indolyl phosphate and nitro blue tetrazolium (Kirkegaard and Perry Laboratories).
The primary hepatocyte cells (Cell Systems Corp.) were plated and cultured at a density of 3 ϫ 10 4 cells/0.2 ml in a 14-mm hole of 35-mm plastic culture dishes (Matsunami Glass Industries, Ltd., Kishiwada, Japan) in CS-C complete medium (Cell Systems Corp.). They were fixed in PBS containing 4% paraformaldehyde, pH 7.4, and treated with PBS containing 0.1% Triton X-100, followed by incubation with rabbit anti-MBP serum or rabbit anti-CL-L1-CRDhis serum with or without PBS containing 400 g/ml recombinant CL-L1-CRDmal and 25% Block Ace and fluorescein-conjugated anti-rabbit IgG (Chemicon International, Inc.) diluted 1:200. The cells were then treated with 1 drop of SlowFade antifade reagent (Molecular Probes, Inc.), mounted, and sealed. The fluorescence images were observed with a Nikon Optiphot-2 microscope, and photographs were prepared with Filing Imaging Software U6341 (Hamamatsu Photonics K.K.).
Sequence Analysis and Construction of a Phylogenetic Tree-The comparison of the amino acid sequence of the CRD of CL-L1 with the GenBank TM sequences below was done using the DNASIS Sequence Analysis Program (Hitachi). A phylogenetic tree was constructed by the neighbor-joining method (22) (29), rat SP-D (30), and human SP-A (4). The phylogenetic relationships were analyzed using the computer program PHYLIP Version 3.57c package (25).

Identification of a New Human Liver Collectin (CL-L1)-We
screened DNA data bases to identify novel members of the collectin family. This resulted in the identification of a cDNA fragment from human EST data bases that showed carboxylterminal sequence homology to the collectins. The EST clone F1-1006D from a fetal liver cDNA library was used to screen a human liver cDNA library, and two positive clones (HL11-3M and HL11-9) were isolated. Furthermore, cap site hunting was performed to determine the complete 5Ј-terminal sequence including the transcriptional start site of CL-L1 mRNA using Cap Site cDNA TM (18). Restriction map analysis and sequencing of the clones revealed that they contained an open reading frame encoding a sequence of 277 amino acid residues. The cDNA contained a 75-nucleotide 5Ј-nontranslated sequence, followed by 831 nucleotides corresponding to a whole protein FIG. 1. Nucleotide and deduced amino acid sequences of human CL-L1 cDNA. Nucleotide residues are numbered in the 5Ј to 3Ј direction. Amino acid residues are numbered in the N-to C-terminal direction, beginning with the first Met residue and ending with Lys. The underlined portions were identified by cap site sequencing. PR1 is the reverse primer for the first PCR, and PR2 was used for the second PCR for cap site sequencing. and a 759-nucleotide 3Ј-nontranslated sequence with an AATAA incomplete polyadenylation signal (Fig. 1). The deduced amino acid sequence from the cDNA revealed a collectin structure consisting of an N-terminal region with cysteine residues, a collagen-like region, a neck region, and a CRD. A hydrophobic signal peptide sequence was not evident at the N-terminal region, and the amino-terminal residue of the mature protein was unknown.
Northern Blot and RT-PCR Analyses-To examine the distribution of CL-L1 mRNA, Northern blot analyses were performed with mRNAs from various human tissues. The Northern blot was hybridized at high stringency with a probe derived from a CL-L1 cDNA clone. Two bands of ϳ1.2 and 3.8 kilobases were detected in liver and also at low levels in placenta ( Fig.  2A). The 1.2-kilobase mRNA was considerably more abundant than the 3.8-kilobase mRNA. RT-PCR analyses showed that most tissues, except skeletal muscle, expressed CL-L1 mRNAs. Slightly high expression levels of CL-L1 mRNAs were found in liver, placenta, adrenal gland, lung, small intestine, and prostate (Fig. 2B).
Southern Blot Analysis-Genomic DNA analyses were performed with several animal DNAs. Zoo blotting showed that all mammalian and avian species have the CL-L1 gene, but it is absent in yeast (Fig. 3). Genomic Southern blotting with several restriction enzyme fragments did not show whether or not CL-L1 is a single copy gene (data not shown).
Localization of the CL-L1 Gene-The CL-L1 gene was localized to chromosome 8q23-q24.1 by fluorescence in situ hybridization (Fig. 4). In 25 of 25 metaphase preparations, hybridiza- The blot contains EcoRI-digested genomic DNAs from human, rhesus monkey, cow, dog, rabbit, rat, mouse, chicken, and yeast (S. cerevisiae). kbp, kilobase pairs. tion signals were observed to the long arm of chromosome 8 in band q23-q24.1. In 22 samples, both copies of chromosome 8 were labeled, and in three samples, a signal was detected on one copy of chromosome 8. Its position was confirmed by PCR analysis using DNA from human monochromosomal hybrid cells (data not shown).
Characterization of Recombinant CL-L1 Sequences of the Neck and CRD Domains in E. coli-Although CL-L1 has collectin organizations that predict mannose or N-acetylglucosamine binding, we sought to verify this lectin activity using recombinant CL-L1 CRDs. Previously, we made collectins lacking the collagen domain in E. coli and characterized their biological activities (19,31). In this experiment, we made two recombinant CL-L1 CRD fusion proteins: CL-L1 CRD-histidine tag fusion protein (CL-L1-CRDhis) and CL-L1 CRD-maltose binding fusion protein (CL-L1-CRDmal). After solubilization, the proteins were applied to mannan or maltose columns, but they did not bind (data not shown). Both recombinant CL-L1 fusion proteins were purified by tag ligands to be used for sugar-blot assays; only CL-L1-CRDmal was used for ELISA because CL-L1-CRDhis is insoluble in TBS. SDS-PAGE showed that the fusion proteins have different molecular sizes of 22 kDa (CL-L1-CRDhis) and 60 kDa (CL-L1-CRDmal) (Fig. 5). Both proteins were immunostained by rabbit polyclonal anti-CL-L1-CRDhis serum (see Fig. 8A). Sugar-blot analyses showed that the two recombinant CL-L1 CRDs from E. coli and recombinant human MBP from Chinese hamster ovary cells were stained; only maltose-binding protein used as a negative control was not stained (Fig. 5). CL-L1-CRDhis was stained more strongly than CL-L1-CRDmal and human MBP. Mannan (100 g/ml) and 10 mM EDTA inhibited the collectins from binding to the ␣-D-mannose BP-probe.
On the other hand, ELISA analyses showed that the ␣-Dmannose BP-probe bound to CL-L1-CRDmal coated on 96-well microwells at its high concentration (Fig. 6). EDTA (10 mM) and mannan (10 mg/ml) inhibited MBP from binding to the ␣-Dmannose BP-probe completely, but inhibited CL-L1-CRDmal binding only slightly. EDTA inhibited more strongly than mannan. A comparison of saccharide specificities by ELISA showed that CL-L1 has affinity for mannose, fucose, and galactose; a lower affinity for N-acetylglucosamine; and the lowest affinity for N-acetylgalactosamine (Fig. 7). The maltose fusion core protein itself at a high concentration exhibited weak affinity for sugar BP-probes (Fig. 7). Therefore, we understand that the difference between the binding activity of recombinant CL-L1-CRDmal and that of maltose fusion core protein would reveal the real lectin activity of CL-L1.
Expression of CL-L1 in Human Tissues--To examine the expression of CL-L1 at the protein level, we performed immunoblot analyses of cytosolic and microsomal fractions from liver using rabbit anti-CL-L1-CRDhis serum. CL-L1 was detected only in the cytosol, but not in microsomes (Fig. 8A) or mitochondria or extracts of nuclei (data not shown). On the other hand, MBP existed in the cytosol and microsomes. This antiserum reacted with CL-L1-CRDhis and CL-L1-CRDmal, but not with MBP. The antiserum detected a band corresponding to ϳ40 kDa in liver. The immunofluorescence analyses in human primary hepatocyte cells showed that CL-L1 was expressed in the cytoplasm, as was MBP (Fig. 8B). Its staining was inhibited by addition of another recombinant CL-L1-CRDmal fusion pro- The plates were coated with 10 g each of CL-L1-CRDmal (q), maltose-binding protein (OE), and human MBP (f). After blocking, the ␣-D-mannose BP-probe was added at various concentrations (0.01, 0.1, 1, and 10 g/ml) with or without mannan (10 mg/ml) or EDTA (10 mM). The binding of the proteins to sugar was determined as described under "Experimental Procedures." tein. Immunoblotting of cell culture medium from hepatocyte cells showed only the band of MBP, but no specific bands of CL-L1.
Sequence Alignment with Collectins from Other Animal Species-To compare the amino acid sequences of CL-L1, MBPs, SP-D, conglutinin, and SP-A, the sequences were aligned with that of CL-L1 (Fig. 9A). This new collectin has the four major domains: an N-terminal cysteine-rich domain, a collagen-like domain, a neck domain, and a carbohydrate recognition domain (Fig. 1). It is composed of 277 amino acids, whereas human, rabbit, and bovine MBPs are composed of 248, 247, and 249 amino acids, respectively. Collectins usually have two or more cysteines in their N-terminal domains that are conserved in all species and are involved in oligomerization. However, this new collectin has only a single cysteine in its N-terminal domain. A collagen domain of 24 Gly-X-Y amino acid repeats is found in CL-L1, without interruption at the eighth repeat, whereas this interruption is conserved in most MBPs. The collagen domain also has many prolines (five residues) and lysines (12 residues) for hydroxylation, like other collectins. The neck region is a variable domain that has hydrophobic amino acids, causing the triple helical structure (32,33). The four cysteine residues and 14 amino acid residues that form the CRD frame in CL-L1 are conserved and found in all collectins (Fig. 9A). The four repeated lysine residues constitute the most characteristic motif that is not found in any other protein examined to date. Analyses using the DNASIS Sequence Analysis Program show that CL-L1 has 29% homology to human MBP (23) when 10 gaps are allowed in the alignment.
Phylogenetic Tree of CL-L1 and Other Collectins-The phylogenetic relationships between the amino acid sequence of the CRD of CL-L1 and those of other collectins were analyzed (Fig.  9B). The tree shows that the collectin family is made up of four classes: the MBP class, consisting of MBP, MBP-A, and MBP-C; the SP-D class, including SP-D, conglutinin, and CL-43; the SP-A class; and finally, CL-L1, which may be the first member of a new group. These data suggest that CL-L1 is a unique group in the collectin family. DISCUSSION We have been interested in studying the function and structure of collectins and their role in the immune system. Much recent data suggest that collectins play an important role in innate immunity (16). The isolation and functional characterization of novel collectins in addition to MBP, SP-D, and SP-A might provide further insights on the functions of these collectins. We screened the human EST data base for cDNA fragments that showed sequence homology to most of the collectins in their carboxyl-terminal amino acid residues and identified a cDNA fragment encoding CL-L1. Analyses of the cDNA encoding CL-L1 suggest that CL-L1 has the same domain organiza- tion as collectins, namely an N-terminal cysteine region, a collagen-like domain, a neck domain, and a CRD (34). Furthermore, a comparison of the amino acid sequence of the CL-L1 CRD with those of collectins suggests that CL-L1 has a basic frame of CRD (four cysteines and 14 amino acid residues). The four C-terminal repeated lysine residues constitute the most characteristic motif in this collectin and are not found in other collectins or any other proteins examined to date. The phylogenetic relationship between CL-L1 and other collectins suggests that CL-L1 may belong to a novel group in the collectin family. The CL-L1 gene was located on chromosome 8q23-q24.1. Other collectins are located on chromosome 10q (35), and we are very interested in the genomic localization of CL-L1.
The genomic organization of the collectin gene family, which includes the MBP, SP-A, and SP-D groups, shows differences. The CRD and neck domains are encoded by a single exon in all collectins (34), whereas the collagen domain is encoded by two (MBP and SP-A) or five (SP-D) exons. Preliminary data on the genomic organization of CL-L1 indicate that the CRD and neck domains are also encoded by a single exon, but the collagen-like domain is encoded by five exons (data not shown), and the number of amino acids in each exon is different from that in other collectins. These data and chromosome mapping results indicate that this new collectin developed differently than other collectins. More detailed genomic studies of its organization are needed.
Northern and Western blot analyses indicated that CL-L1 is expressed in ubiquitous organs (mainly expressed in liver, placenta, and adrenal gland), whereas SP-D and SP-A are expressed in lung and gastrointestinal tracts, and RT-PCR studies show that MBP is expressed in murine kidney and liver. RNA blotting showed that CL-L1 is expressed only faintly in placenta, and Western blotting of placenta showed bands similar to those in liver, but in trace amounts (data not shown). RT-PCRs and nested PCR after RT-PCR showed more sensitive results. CL-L1 gene expression is found in most tissues, except skeletal muscle. The expression level is varied in individual tissues. The high expression organs are liver, placenta, and adrenal gland. This ubiquitous expression pattern is different from that of other collectins and galectins.
Western blot and immunofluorescence analyses indicated that CL-L1 protein is localized in cytosolic fractions from liver. Usually, collectins are secreted into the extracellular space through the endoplasmic reticulum pathway. Using anti-MBP serum with the same sample blot shows that MBP is found mainly in microsomal fractions and less in cytosolic fractions. The secreted collectins are considered to play an important role in innate immunity against pathogens invading from outside the organism (16). CL-L1 may react with internal ligands in contrast to other collectins.
Western blotting showed that MBP has molecular mass of 32 kDa, whereas its calculated molecular mass is 24.5 kDa. Other collectins showed slightly larger molecular masses than those estimated. These results indicate that CL-L1 of ϳ40 kDa on SDS-PAGE is not inconsistent. The amino acid residues in the neck domain of CL-L1 can form ␣-helices like other collectins (34). The ␣-helical bundle is very stable against denaturation by heat (T m Ͼ 55°C) or pH (pH 3.0 -8.5), indicating that CL-L1 might maintain dimer and trimer structures under the reducing conditions of the cytoplasm, like other collectins.
The expression studies using CL-L1-CRDhis and CL-L1-CRDmal indicated that lectin activity is preserved in CL-L1, but it is very weak. Two analyses (ELISA and sugar-blot) of weak lectin activity suggested that CL-L1 can bind to mannose at high concentrations and that this can be inhibited by mannan and EDTA. The carbohydrate-binding specificities of most collectins are for mannose-type saccharides, and saccharide specificities of the CL-L1 CRD include galactose as well as mannose, fucose, and N-acetylglucosamine. Previously, the recombinant collectin fusion proteins produced in E. coli were used in the analysis of carbohydrate-binding specificities in recombinant collectins. All of these lectin fusion proteins attached to saccharide columns due to their high affinity. However, CL-L1-CRDmal cannot bind to saccharide columns under any buffer conditions. CL-L1-CRDmal at a high concentration (10 g/ml) has binding activity. At such a high concentration, the maltose fusion core protein itself exhibits weak affinity for sugar BP-probes. Therefore, we understand that the difference between the binding activity of recombinant CL-L1-CRDmal and that of the maltose fusion core protein would reveal the real lectin activity of CL-L1.
The two amino acid residues of the five in collectins responsible for complexing the calcium ion involved in carbohydrate binding, namely Glu-185 and Asn-187 (36), when changed to Gln-185 and Asp-187, resulted in an increased affinity for galactose. The two amino acids in CL-L1 are Glu-238 and Ser-240, indicating that CL-L1 is a hybrid type, like SP-A, between collectins and other lectins specific for galactose (34). These sugar specificities and weak lectin activities are involved in various functions of the collectins. Furthermore, the polycharge islands of repeated lysine residues in the carboxylterminal area might indicate an alternative role for this new collectin. A crystallographic study and computer graphic model suggest that conserved KGQKGEKGS sequences in the collagen domain of the macrophage scavenger receptor make a "charged collagen" molecule with a coiled groove surrounded by lysine residues (37). These lysine clusters are considered to be the receptor site of the macrophage scavenger receptor. Scavenger receptors have characteristic broad ligand specificities and are able to bind various substances including degenerated lipoproteins, lipopolysaccharide, and microorganisms (38). These findings suggest that CL-L1 may play a role as a scavenger or chaperonin in the cytoplasm.
The searches of the EST data base with the CL-L1 sequence can identify related molecules already described in the Unigene data base. From the first hit to the third hit are fragments of the CL-L1 gene. The related collectin genes are hit 18 times. The SP-A gene is hit 12 times, the SP-D gene is hit five times, and MBP is hit once. Other hit sequences are collagen genes. The searches showed that the related genes with high homology are hit from the first. These data support the relationship between the CL-L1 gene and other groups (SP-A, SP-D, and MBP genes) in the phylogenetic tree (Fig. 9B). The EST data base is made from mRNA. The covering how much mRNAs is different in individual tissues. The expression of mRNA varies in different tissues. The high expression gene is easily picked up by the EST database. The highly developed organ's EST data base might be able to pickup the CL-L1 gene. The isolation of cDNA encoding CL-L1, the preparation of CL-L1 CRD fusion proteins, and the cytoplasmic localization of CL-L1 will provide the basis for future studies on the function and structure of this new collectin. The findings that CL-L1 has a novel carboxyl-terminal lysine cluster and weak lectin activities for galactose as well as mannose, fucose, and N-acetylglucosamine and the ubiquitous expression in RT-PCR analysis suggest that this protein might play some role in regulating cell functions.