Human Mannose-binding Lectin and L -Ficolin Function as Specific Pattern Recognition Proteins in the Lectin Activation Pathway of Complement*

The innate immune response in vertebrates and inver-tebrates requires the presence of pattern recognition receptors or proteins that recognize microbial cell components including lipopolysaccharide, bacterial peptidoglycan (PGN), and fungal 1,3- (cid:1) - D -glucan. We reported previously that PGN and 1,3- (cid:1) - D -glucan recognition proteins from insect hemolymph were able to induce the activation of the prophenoloxidase-activating system, one of the major invertebrate innate immune reactions. The goal of this study was to characterize the biochemical properties and effects of the human counterparts of these molecules. Soluble pattern recognition proteins were purified from human serum and identified as human mannose-binding lectin (MBL) and L -ficolin. The use of specific microbial cell component-coupled columns demonstrated that MBL and L -ficolin bind to PGN and 1,3- (cid:1) - D -glucan, respectively. Purified MBL and L ficolin were associated with

Innate immune responses in animals are mediated by specific pattern recognition receptors or proteins that recognize microbe-specific pathogen-associated molecular pattern molecules such as lipopolysaccharide, peptidoglycan (PGN), 1 lipoteichoic acid, mannan, and 1,3-␤-D-glucan (1)(2)(3)(4). This process also occurs in mammals where secreted pattern recognition proteins such as mannose-binding lectin (MBL) and ficolins (L-ficolin and H-ficolin) bind to microbial cells. In human plasma, MBL and ficolins are associated with three types of serine protease termed MBL-associated serine protease (MASP), MASP-1, MASP-2, and MASP-3, and a truncated form of MASP-2 called small MBL-associated protein (sMAP). The complexes promote the elimination of microbes through activation of phagocytosis or the lectin pathway of the complement system (5,6). Further, pattern recognition receptors expressed on the cell surface, including the family of Toll-like receptors, can activate signaling pathways that induce antimicrobial effectors responses and inflammation upon recognition of different pathogen-associated molecular pattern molecules (1)(2)(3)(4). Also, Toll-like receptors expressed on specific antigen-presenting cells play a critical role in the initiation of adaptive immune responses (3,4).
PGN is a key component of the bacterial cell wall. It is composed of polymeric sugar chains consisting of alternating 1,4-␤-linked N-acetylglucosamine (GlcNAc) and N-acetylmuramic acid residues that are cross-linked with shorts peptides (7). Recent studies have demonstrated that PGN activates host cells through the membrane-associated pattern recognition receptors such as Toll-like receptor-2 (8 -9) and CD14 (10) and induces the expression of several genes in human monocytes (11). Other investigators have reported (12) that PGN-induced tumor necrosis factor-␣ expression is directly mediated via the several transcription factors, which suggests that PGN activates the innate immune system of the host and induces the release of inflammatory molecules.
The cloning of genes encoding vertebrate PGN recognition proteins (PGRPs) demonstrated that there is a high degree of sequence conservation between insects and mammalian species (13)(14)(15)(16), indicating that these genes play an important role in the innate immune response. In insects, the recognition of PGN (13,14) and 1,3-␤-D-glucan (17) by PGRPs results in the activation of the prophenoloxidase (pro-PO) activating system, a major humoral defense reaction in arthropods. Moreover, mammalian PGRPs have been shown to possess N-acetylmuramoyl-L-alanine amidase activity (18,19) and mediate inhibition of bacterial growth (20). Even though several biological functions of PGRPs have been determined, the specific ligands of soluble PGRPs have not been clarified in detail. Further, Girardin et al. (21) report that a novel intracellular PGN recognition protein, Nod-1, detected a unique tripeptide motif found in Gram-negative bacterial PGN and subsequently induced the activation of the NF-kB pathway.
Here, to further characterize human PGN and 1,3-␤-D-glucan recognition molecules, we purified two kinds of complement-related proteins from the human serum by using PGNand 1,3-␤-D-glucan-coupled columns and characterizing their biochemical properties. We found that human MBL and Lficolin function as secreted pattern recognition proteins for PGN and 1,3-␤-D-glucan, respectively.

EXPERIMENTAL PROCEDURES
Preparation of Soluble PGN-coupled Sepharose 4B and 1,3-␤-D-Glucan-coupled Toyopearl Resin-A PGN-coupled column was prepared as described previously (17). Soluble PGN fragments were prepared from Staphylococcus aureus insoluble PGN (Fluka Chemie, Steinheim, Germany) according to the published method (30). One milligram of the lyophilized soluble PGN was coupled to 0.4 g of CNBr-activated Sepharose 4B (Amersham Biosciences) according to the manufacturer's recommended protocol, and the PGN-coupled Sepharose binding ability was assessed using recombinant Holotrichia diomphalia PGRP-1 (17). Resins with a binding capability of Ͼ10 g of H. diomphalia PGRP-1 per 1-g resin were used for the purification of PGN recognition proteins in human serum. As a control, CNBr-activated Sepharose 4B that had been inactivated with 100 mM Tris-HCl (pH 8.0) was prepared according to the manufacturer's recommended protocol.
A 1,3-␤-D-glucan-coupled column was also prepared as described previously (31). 3 ml of 1,3-␤-D-glucan (curdlan, Wako Pure Chemicals, Osaka, Japan) stock solution (500 mg of curdlan was dissolved in 15 ml of 1 N NaOH) was added to 20 ml of 0. 2 M K 2 HPO 4 , and the pH was adjusted to 8.4 with 6 M HCl. 2.5 g of suction-dried AF-Amino Toyopearl 650 M (Tosho, Tokyo, Japan) was suspended next in 23 ml of the soluble curdlan solution with 1 g of NaCNBH 3 and the mixture was incubated at 60°C overnight. To block any remaining free amino groups, the gels was acetylated by incubation with 8 ml of 0.2 M sodium acetate and 4 ml of acetic anhydride for 30 min on ice followed by the addition of 4 ml of acetic anhydride and another 30 min of incubation at room temperature. Finally, the resin was washed sequentially with 150 ml of 0.1 M NaOH, 150 ml of 1 M Tris-HCl (pH 8.0), and distilled water. Resin without curdlan served as a control. The amount of 1,3-␤-D-glucan coupled to the resins was quantified by the sulfuric acid-phenol method (32). Resins that coupled Ͼ3 g of glucose per 1-mg resin were used for purification of human 1,3-␤-D-glucan recognition proteins.
Purification of Human PGN Recognition Proteins-Human plasma was obtained from volunteer donors who had provided written informed consent at the Busan Blood Center (Busan, Korea). The study was performed under the approval for research on human subjects obtained from the Pusan National University. A final concentration of 10 mM CaCl 2 was added to the plasma sample followed by incubation at 37°C for 1 h. Resulting fibrin clots were removed using sterilized gauze and discarded. The fibrin-excluded serum (100 ml) was precipitated with 8 g of polyethylene glycol 4000 (Sigma). After centrifugation at 9,000 rpm for 15 min at 4°C, the precipitates were dissolved in 15 ml of buffer A (50 mM Tris-HCl (pH 7.8), 200 mM NaCl, and 20 mM CaCl 2 ) and then divided into two batches. Each batch then was loaded on a PGN-coupled Sepharose 4B and control-Sepharose 4B columns that had been equilibrated with buffer A. Both columns were washed with buffer A until the absorbance at 280 nm was zero. Bound proteins were eluted from each column using buffer A containing 0.3 M GlcNAc followed by buffer A containing 0.3 M mannose at a flow rate of 0.3 ml/min. The protein profiles in aliquots of eluted fractions were analyzed by SDS-PAGE under reducing conditions, and fractions containing a 30-kDa protein band were pooled. The pooled fractions were passed through a protein A-attached Sepharose CL 4B (Amersham Biosciences) column equilibrated with buffer A, and the flow-through fractions were passed subsequently through an anti-IgM-coupled Sepharose 4B column. The purified 30-kDa protein was separated by SDS-PAGE under reducing conditions and electroblotted onto a polyvinylidene difluoride membrane for determination of the amino-terminal sequence. For internal amino acid sequencing, proteins were separated by 12% SDS-PAGE under reducing conditions. The gels were stained with 0.2% Coomassie Brilliant Blue in 50% methanol and destained with 30% methanol. The 30-kDa protein band was excised and treated with lysylendopeptidase (Wako Pure Chemicals) according to the method described by Wilm et al. (33). The resulting digest was subjected to reverse phase high performance liquid chromatography (Gilson, WI) using a C 18 high performance liquid chromatography column (Tosho). High performance liquid chromatography was performed with a linear gradient of 0 -80% acetonitrile in 0.1% trifluoroacetic acid for 75 min at a flow rate of 0.3 ml/min, and the prominent peaks were sequenced using an automatic amino acid sequencer (Procise, Applied Biosystems, Foster City, CA).
Purification of 1,3-␤-D-Glucan Recognition Proteins-Human serum was precipitated with polyethylene glycol as described above and dissolved in buffer A. Samples then were loaded onto a 1,3-␤-D-glucancoupled Toyopearl column equilibrated with buffer A. The column was washed with buffer A until the absorbance at 280 nm was zero, and bound proteins were sequentially eluted with buffer A containing 0.3 M glucose, 0.3 M mannose, and 0.3 M GlcNAc. Eluted proteins were analyzed by SDS-PAGE under reducing conditions. The fractions most enriched in the 35-kDa band were blotted onto a polyvinylidene difluoride membrane, and the amino-terminal sequence or partial amino acid sequences of this protein were determined as described above.
Immobilization of Soluble PGN Fragments-Carboxyl-enzymelinked immunosorbent assay 96-well microplates (Sumitomo Bakelite, Tokyo, Japan) were activated with 100 l of water-soluble carbodiimide (10 g/ml) in phosphate-buffered saline (pH 5.8) by incubation at 37°C for 2 h. The wells were washed three times with 300 l of phosphatebuffered saline, and 100 l of soluble PGN was added to each well (ϳ3 g of lyophilized-soluble PGN fragments) followed by incubation at 37°C for 2 h. The wells then were washed with 300 l of phosphatebuffered saline containing 0.05% Tween 20 (pH 7.6). Microplates without PGN served as controls.
Immobilization of 1,3-␤-D-Glucan-Amino-enzyme-linked immunosorbent assay 96-well microplates coupled with 1,3-␤-D-glucan were prepared by adding 200 l of soluble curdlan solution to each well and incubating the plates at 60°C for 30 min. 100 l of a NaBH 3 CN solution (20 mg of NaBH 3 CN dissolved in 0.2 M K 2 HPO 4 , (pH 9.4)) was added next to each well and incubated at 60°C for 14 h. Wells were washed three times with distilled water, twice with 100 mM of boric acid (pH 8.2), and another three times with distilled water. 160 l of 0.4 M sodium acetate and 160 l of acetic anhydride were added next to each well, and the plates were incubated at 4°C for 30 min. The supernatant was discarded, and 300 l of acetic anhydride was added followed by incubation at room temperature for 30 min. The wells were washed three times with distilled water, three times with 0.1 M NaOH, three times with 1 M of Tris-HCl (pH 8.0), and another six times with distilled water.
Antibodies and Immunoblotting-Mouse monoclonal antibody against MASP-1 (1E2) was prepared as described previously (34). Rabbit polyclonal antibody recognizing the amino terminus of MASP-2/ sMAP (35,36) was provided by Dr. I. Terai (Health Sciences University of Hokkaido, Sapporo, Japan). For immunoblotting, protein samples were subjected to SDS-PAGE under non-reducing and reducing conditions and transferred electrophoretically to polyvinylidene difluoride membranes. Membranes were blocked by immersion in TBST (50 mM Tris-HCl (pH 7.4), 150 mM NaCl, and 0.02% Tween 20) containing 5% skim milk and 1% horse serum for 12 h. The membranes were then transferred to a solution containing antibodies against MASP-1 (50 g/ml) or MASP-2 (5 g/ml) and then incubated at 4°C for 2 h and then later with horseradish peroxidase-labeled rabbit-goat IgG conjugate as secondary antibody (1:3,000, Amersham Biosciences). The bound antibodies were detected using an ECL Western blotting reagent kit (Amersham Biosciences).
Binding Assays-PGN-and 1,3-␤-D-glucan-coupled microplates were blocked with Block-Ace (0.3 ml, Dainippon Pharmaceutical, Osaka, Japan), and the purified human MBL/MASP or L-ficolin/MASP (4 g of protein/100 l in TBST) complexes were diluted with TBST containing 10 mM CaCl 2 and incubated at 4°C for 2 h. After washing with TBST, a biotinylated anti-MBL monoclonal antibody (3E7) (37) or a biotinylated anti-L-ficolin-monoclonal antibody (2F5) (38) was added to the wells and the plates were incubated at 37°C for 1 h. The wells were washed again and then incubated with an avidin-peroxidase complex (Vector Laboratories, Burlingame, CA) at 37°C for 1 h. The reaction was developed with ABTS (azino-di-(3-methylbenzthiazoline) sulfonic acid, Zymed Laboratories Inc., San Francisco, CA). The amount of the MBL/MASP complexes or L-ficolin/MASP complexes bound to plates was determined by measuring the absorbance at 405 nm (A 405 ).
C4 Activation Assay-The activation of the lectin-complement path-way was determined by the C4 cleavage assay as described by Petersen et al. (39). PGN-and 1, 3-D-␤-glucan-coupled microplates were blocked with Block-Ace, and the purified MBL/MASP or ficolin/MASP complexes (8 g of protein/well) were incubated at 4°C for 2 h. The wells were washed first with TBST and then with mannitol gelatin veronal buffer (5 mM veronal, 74 mM NaCl, 0.1% gelatin, 2.3% mannitol, 2 mM CaCl 2 , and 0.5 mM MgCl 2 (pH 7.4)). Human C4 (0.3 g) was added to each well, and the plates were incubated at 37°C for 1 h. The wells then were washed with TBST, and C4 deposition was detected by incubation with horseradish peroxidase-conjugated polyclonal anti-C4 antibody (horseradish peroxidase-anti-C4) at 37°C for 1 h followed by development with ABTS. The amount of C4 deposition was determined by measuring the absorbance at 405 nm.

Isolation of Human MBL Using PGN-coupled Sepharose 4B-PGN-coupled
Sepharose 4B columns and the control-Sepharose 4B columns were loaded with human serum precipitate. Upon elution with 0.3 M GlcNAc, no differences were observed in the elution profiles of proteins eluted from the two columns as determined by SDS-PAGE analysis (data not shown). However, the elution of the PGN-Sepharose column with 0.3 M GlcNAc followed by 0.3 M mannose yielded a 30-kDa band (Fig. 1A, Band-a). Furthermore, the amino-terminal sequence and two partial amino acid sequences (M-1 and M-2) of Band-a were identical to the sequences of human MBL (Fig.  1B). Human IgG and IgM from the PGN column were removed through the use of the protein A and the anti-IgM column, which yielded a homogenate of MBL (Fig. 1C, lane 5). In all, ϳ148 g of human MBL was obtained when 100 ml of human plasma was loaded onto the PGN-coupled column. These results suggest that human MBL recognizes and binds to soluble PGN.
Association of Purified MBL and L-Ficolin with MASPs-To determine whether MBL and L-ficolin associate with MASPs, eluents were subjected to Western blot analysis using anti-MASP-1 and anti-MASP-2/sMAP antibodies. MASP-1, MASP-2, and sMAP were detected along with MBL in the pass-through fractions from the protein A column (Fig. 3, A and B, lane 1) and from the anti-IgM column (Fig. 3, A and B, lane 2) under non-reducing conditions and reducing conditions, respectively. Further, purified L-ficolin from the 1,3-␤-D-glucan-column was associated with MASP-1, MASP-2, and sMAP (Fig. 3, C and D).
Activation of the Lectin Pathway of Complement-To examine the complement activation capacity of the two purified lectins, we next examined whether the purified MBL/MASP complexes and L-ficolin/MASP complexes can bind to PGN and 1,3-␤-D-glucan, respectively. For this experiment, PGN and 1,3-␤-Dglucan were chemically coupled with carboxyl-and amino-enzyme-linked immunosorbent assay microplates, respectively. Purified L-ficolin/ MASP complexes exhibited stronger binding to 1,3-␤-D-glucancoupled microplates (Fig. 4A, second column) than to PGNcoupled plate (Fig. 4A, first column). In contrast, MBL/MASP complexes showed strong binding to PGN plates (Fig. 4B, first column) but did not bind significantly to 1,3-␤-D-glucan-coupled plates (Fig. 4B, second column). These results clearly support that the binding specificities of the MBL and L-ficolin against PGN and 1,3-␤-D-glucan, respectively, are consistent with the results of the column chromatographies.
Previous studies demonstrate that MASP-2 cleaves C4 to C4a and C4b when its complexes with MBL and L-ficolin engage complementary carbohydrate ligands on the surface of microorganisms, leading to the covalent linking of C4b to such pathogens (C4 deposition) (39). Using the fixed PGN and 1,3-␤-D-glucan plates, we measured the ability of cleaving C4 by the purified MBL and L-ficolin. C4 cleavage and subsequent C4b deposition occurred in PGN plates incubated with MBL/MASP (Fig. 5A, first column) or L-ficolin/MASP complexes (Fig. 5B, first column) as well as with 1,3-␤-D-glucan plates incubated with L-ficolin/MASP complexes (Fig. 5C, first column). There was also a low level of C4 cleavage when PGN plates were incubated with L-ficolin/MASP complexes (Fig. 5B, first column), reflecting the weaker binding of L-ficolin/MASP complexes to PGN as demonstrated above. However, this did not occur in 1,3-␤-D-glucan plates incubated with MBL/MASP complexes (Fig. 5D, first column). These results suggest that PGN and 1,3-␤-D-glucan may stimulate the lectin pathway of the complement via MBL/MASP and L-ficolin/MASP complexes, respectively.

DISCUSSION
Recent studies have demonstrated that human MBL and L-ficolin constitute a group of soluble pattern recognition proteins that can stimulate the lectin pathway of the complement following binding to several ligands, such as mannan (41), lipoteichoic acid (42), GlcNAc (38), lipopolysaccharide (43), and several pathogenic microbes including bacteria and fungi (44). However, no biochemical data regarding specific interactions between soluble PGN or 1,3-␤-D-glucan and MBL/MASP or L-ficolin/MASP complexes have been reported. In this result, we have shown that L-ficolin/MASP complexes bound to 1,3-␤-D-glucan and MBL/MASP complexes bound to S. aureus PGN can induce the activation of complement protein C4.
Our results provide biochemical evidence of the recent observation (45) that MBL/MASP complexes binding to S. aureus cells enhanced complement activation and opsonophagocytosis. Two recent reports (42,43) demonstrate that MBL/MASP complexes can bind to lipopolysaccharide and that L-ficolin/MASP complexes can bind to lipoteichoic acid, a cell component found in all Gram-positive bacteria. These interactions lead to C4 deposition, suggesting that lipopolysaccharide and lipoteichoic acid can activate the lectin pathway of the complement via MBL and Lficolin, respectively. The diverse recognition abilities of human MBL and L-ficolin against different pathogen-associated molecular pattern molecules and the co-existence of MBL and L-ficolin in serum are expected to enlarge the spectrum of pathogenic microorganisms that can be recognized and eliminated through the lectin pathway of the complement activation.
Recent data support that the arthropod pro-PO system has similarity with the mammalian lectin pathway of the complement system (46 -48). Both the pro-PO activation system and the lectin pathway of the complement are proteolytic cascades comprising pattern recognition proteins, several serine proteases, and their inhibitors and terminate with the zymogen, pro-PO, and the membrane attack complex, respectively (49). In our previous studies, when we tried to purify 1,3-␤-D-glucan recognition molecules from insect hemolymph, we obtained insect 1,3-␤-D-glucan recognition protein and PGRPs-like proteins. These molecules triggered the activation of insect pro-PO system in association with an unidentified serine protease (17,31). In this study, MBL and L-ficolin from human serum were shown to act as pattern recognition molecules for PGN and 1,3-␤-D-glucan. These molecules were also found to be associated with MASP and sMAP. Among them, MASP-2 could activate the C4 component of the complement. These data provide strong evidence for a highly conserved nature of innate immunity from insects to mammals.
L-Ficolin is shown to have a carbohydrate binding specificity for GlcNAc (38,50). Because ␤-glucan has a backbone structure of ␤-1,3-linked glucoses, the reason why L-ficolin bound to 1.3-␤-D-glucan and eluted with GlcNAc remains to be elucidated. In addition, L-ficolin can bind to the PGN column and also to the control-Sepharose (data not shown), indicating that L-ficolin is not a specific PGN recognition protein. Previously, Le et al. (51) reported that L-ficolin could bind to CNBr-activated Sepharose that had been inactivated with Tris and could be eluted from the preparation using GlcNAc. They also postulated that Lficolin might recognize an amide group (-CO-NH-) or a similar group with an amide linkage of the control-Sepharose (52). However, in this study, a considerable amount of L-ficolin was bound to PGN. Because PGN is composed of sugar chains consisting of alternating GlcNAc and N-acetylmuramic acid, the possibility remains that L-ficolin bound specifically to PGN. This point still is to be clarified.
In conclusion, this study shows that human MBL and L-ficolin specifically recognize soluble PGN and 1,3-␤-D-glucan, respectively, and that subsequently these PGN-and 1,3-␤-D-glucanrecognizing pattern proteins can induce C4 activation of the complement. In addition, PGN-coupled Sepharose or 1,3-␤-D-glucan-coupled Toyopearl may provide a step in an improved purification procedure for human MBL/MASP complexes or ficolin/ MASP complexes to study their biological and functional studies.