Lectin from the Manila Clam Ruditapes philippinarum Is Induced upon Infection with the Protozoan Parasite Perkinsus olseni*

Glycan-binding proteins (lectins) are widely expressed in many invertebrates, although the biosynthesis and functions of the lectins are not well understood. Here we report that Manila clam (Ruditapes philippinarum) synthesizes a lectin termed Manila clam lectin (MCL) upon infection with the protozoan parasite Perkinsus olseni. MCL is synthesized in hemocytes as a ∼74-kDa precursor and secreted into hemolymph where it is converted to 30and 34-kDa polypeptides. The synthesis of MCL in hemocytes is stimulated by one or more factors in Perkinsus-infected hemolymph, but not directly by Perkinsus itself. MCL can bind to the surfaces of purified hypnospores and zoospores of the parasite, and this binding is inhibitable by either EDTA or GalNAc. Fluorescent beads coated with purified MCL were actively phagocytosed by hemocytes from the clam. Immunohistochemistry showed that secreted MCL is concentrated within cyst-like structures. To define the glycan binding specificity of MCL we examined its binding to an array of biotinylated glycans. MCL recognizes terminal non-reducing β-linked GalNAc as expressed within the LacdiNAc motif GalNAcβ1–4GlcNAcβ1-R and glycans with terminal, non-reducing β-linked Gal residues. Our results show that the synthesis of MCL is specifically up-regulated upon parasite infection of the clams and may serve as an opsonin through recognition of terminal GalNAc/Gal residues on the parasites.

These lectins may be classified into two general groups. One group is Ca 2ϩ -dependent and includes the C-type lectin family. Recently, C-type GalNAc-specific lectins from Cucumaria echinata were shown to interact with target membranes and exhibit strong hemolytic activity and cytotoxicity through pore formation (22). In the ascidian, Halocynthia roretzi, mannosebinding lectin has been reported (23). It is associated with mannose-binding lectin-associated serine proteases 1 and 2, which are the enzymes responsible for complement activation initiated by mannose-binding lectin in the lectin pathway (24).
The second group of lectins comprises the galactosyl-binding lectins, which have structural homologies with galectins. The sponges Geodia cydonium and Suberites domuncular express galectin-related lectins that are similar in structure to molecules that are involved mammalian immune responses (25). In marine bivalve host defense, a heterogeneous sialic acid-binding lectin with affinity for bacterial lipopolysaccharides has been isolated and partially characterized from the hemolymph of the horse mussel Modiolus modiolus (21). The isolated horse mussel lectin exhibited strong antibacterial activity against tested Vibrio strains, which suggests that the lectin plays a role in the elimination of bacteria and participates in an innate immunity.
We recently discovered a novel lectin (MCL) 2 from the Manila clam (26). The lectin has an apparent molecular mass of 138 kDa in non-reducing SDS-PAGE and consists of 74-, 34-, and 30-kDa species. Inhibition assays showed that mucins containing GalNAc inhibit MCL binding (26). To better understand the biological roles of MCL in clams, we have explored MCL biosynthesis, carbohydrate-binding specificity, and its potential function in host defense.
Here we report that the infection of clams by the protozoan parasite Perkinsus olseni induces the synthesis of MCL by clam hemocytes. Newly synthesized MCL is secreted into hemolymph as a 74-kDa precursor protein that is processed to the 30and 34-kDa species. We have also defined the glycan-binding specificity of MCL and propose a potential role for the lectin in innate immunity in bivalve defense against the parasite.

MATERIALS AND METHODS
Purification of Perkinsus Trophozoites from Manila Clam-Trophozoites of P. olseni were extracted from the hemolymph withdrawn from the blood sinus of the adductor muscle of Manila clams, which were collected at Wando Island off the southern coast of Korea, where a high incidence of the parasite infection has been reported (27). Perkinsus in the hemolymph were propagated for 1 week in Dulbecco's modified Eagle's medium/Ham's F-12 (1:2) with Hepes buffer and 5% fetal bovine serum according to Ordas and Figueras (49).
Perkinsus Purification and Culture-One milliliter of hemolymph containing trophozoites of Perkinsus was withdrawn from the blood sinus of the adductor muscle of Manila clams with an insulin syringe and seeded in 24-well polystyrene plates. One milliliter of 35% sterilized artificial sea water prepared with Marine Mix (Sigma-Aldrich) with 4000 IU/ml penicillin/streptomycin was added to each well, and after 12 h of incubation at 26°C, the contents of the wells were pooled and centrifuged for 10 min at 1500 ϫ g. The pellet containing parasites was resuspended in artificial sea water with 4000 IU/ml penicillin/streptomycin and incubated for another 12 h at 26°C. Parasites were centrifuged again, and the supernatant was carefully removed from each well. To each well was added 1 ml of Dulbecco's Modified Eagle Medium/Ham's F-12 (1:2) with 50 mM Hepes buffer, 3.5 mM sodium bicarbonate, and 5% fetal bovine serum in 35% artificial sea water. During the first week the media was changed twice, and when the number of Perkinsus cells was Ͼ10 7 cells/ml, another subculture was initiated.
Biosynthesis of MCL in the Hemocytes-The hemolymph (1 ml) was withdrawn from the Perkinsus-infected clam and centrifuged (1000 ϫ g). The supernatant was analyzed for a hemocyte-free hemolymph, and the pellet contained the hemocytes. The hemocytes were resuspended and incubated in media, as described above, for the indicated times at 37°C and subjected to Western blot analysis for MCL.
Hemolytic Assay-The potential hemolytic activity of MCL was assessed using human erythrocytes. Erythrocytes were isolated from heparinized blood by centrifugation (1000 ϫ g, 5 min) and washed three times with phosphate-buffered saline (PBS). The cell suspension was adjusted to 1 ϫ 10 9 ml Ϫ1 in PBS. The cell suspensions (20 l) were incubated at 37°C for 30 min with equal volumes of MCL. The incubation media was centrifuged (1000 ϫ g, 5 min). The absorbance of the solution was measured at 420 nm. The absorbance obtained after treating erythrocytes with only PBS and 2% SDS were taken as 0 and 100%, respectively.
Glycan Binding Specificity of MCL-MCL was purified as described previously (26) and was dialyzed into 150 mM NaCl containing 100 mM Na-HEPES, pH 7.8, and 10 mM CaCl 2 , and reacted with fluorescein isothiocyanate (FITC) (50 g of FITC/mg of protein). Procedures for probing the full glycan arrays are available at www.functionalglycomics.org/static/ consortium/. Biotinylated glycans were bound to streptavidincoated plates (Pierce) by incubation at 4°C overnight at a concentration of 100 M. Purified MCL was labeled with FITC in the presence of ligand (GalNAc) to reduce the likelihood of labeling the active binding site. After labeling only the active lectin is purified by affinity chromatography and used for the experiments. Fluoresceinated MCL was incubated in the triplicate wells for 2 h at 4°C in Tris-buffered saline (TBS, 25 mM Tris-Cl, pH 7.8, 150 mM NaCl, 25 mM CaCl 2 , 0.1 mg/ml bovine serum albumin) and washed three times in the same buffer, without added albumin, before measuring fluorescence intensity in a fluorescence plate reader.
Phagocytosis Assay-The opsonizing ability of MCL was tested by a flow-cytometric assay that measures the phagocytosis of MCL-coated fluorescent beads by hemocytes. Fluorescent latex beads ( 2.0 m, 20 l, Polyscience, Inc., Warrington, PA) were prepared by incubating them for 2 h with either 20 l of MCL (10,000 titer/50 l to human blood group O) in 1 ml of TBS-Ca 2ϩ ; 20 l of MCL with 2 mg/ml GalNAc in 1 ml of TBS-Ca 2ϩ ; or 1 ml of filtered seawater (control group). After incubation the beads were washed 3 times with filtered seawater, and then 30 l of a suspension of fluorescent beads from each preparation was added to a suspension of hemocytes prepared from Manila clams. Hemocytes were prepared from clam hemolymph (100 l). The cells were incubated with different bead preparations at 20°C in the dark on a gentle shaker for 1 h. An equal volume of 6% neutral formalin was added to the fix hemocytes, which were then analyzed by flow cytometry. Another control group was prepared where beads that had been incubated with 1 ml of filtered seawater were added to the hemocyte suspension and then immediately mixed with 3 l of cytochalasin B at 0.1% final concentration to block further phagocytosis. Phagocytosis was then analyzed by two parameters, log amplifications of the complexity and cell size, and the log fluorescence frequency distribution histogram of the hemocyte population was determined. The data were collected for 100 s. The percentage of cells phagocytosing beads was noted. Statistical analyses were performed using statistical software with possible differences among groups defined by one-way analysis of variance followed by Duncan's multiple range tests. Differences were considered significant at a probability level of 0.05.
SDS-PAGE-To determine the molecular sizes of the proteins that were eluted from the columns, 10% polyacrylamide gel electrophoresis was used under non-reducing and reducing conditions. The molecular masses of the proteins were compared using high and low molecular weight markers (Sigma-Aldrich).
Lectin Activity Assay-The direct hemagglutination assay was performed by a standard procedure using trypsin-treated human blood group O erythrocytes in TBS-Ca 2ϩ (26). Aliquots of 50 l of an erythrocyte suspension, which were diluted 10-fold with TBS-Ca 2ϩ , were added to microplate wells and mixed with 100 l of serially diluted MCL with or without inhibitors. The mixture was shaken for 10 min on a microplate well shaker, and hemagglutination was scored by microscopic observation of visible cell clumps.
Western Blotting-After electrophoresis on 12.5% polyacrylamide gels, proteins were transferred to a 0.45-m polyvinylidene difluoride membrane (Pierce). The transfer was carried out at 110 V for 2 h using 25 mM Tris (pH 8.3), 192 mM glycine, 20% MeOH, and 1% SDS as the transfer buffer. The gel was placed in Coomassie Blue to verify that the transfer was successful. The membrane was incubated in a blocking solution of TBS-containing Tween (TTBS, 100 mM Tris-HCl, 0.9% NaCl, 0.1% Tween-20) plus 1% bovine serum albumin for 1 h at room temperature. The membrane was washed three times for 10 min each in TTBS and incubated with anti-MCL antibody (1:500 dilution in TTBS plus 1% bovine serum albumin) for 1 h at room temperature. After three washes, the membrane was incubated with the secondary antibody (horseradish peroxidase-conjugated goat anti-rabbit IgG (BioDom Co.), which was diluted 1:500 in TTBS. The membrane was re-washed, and the antigen-antibody complexes were detected using enhanced chemiluminescence (ECL, Amersham Biosciences).
Immunofluorescence Staining-Hypnospores and zoospores of Perkinsus were washed with PBS plus 20 mM EDTA (PBS-T), incubated with 100 ml of MCL in TBS-Ca 2ϩ for 1 h, then incubated in 5% bovine serum albumin in PBS-T and a mono-specific antibody against MCL (1:100 dilution in PBS-T) for 1 h. The cells were then washed three times in PBS-T and incubated with a 1:200 dilution in PBS-T of FITC-conjugated goat antibody against rabbit immunoglobulins (Sigma-Aldrich) for 1 h. The cells were incubated with pre-immune rabbit IgG as a control. For immunohistochemical staining, Perkinsus-infected clams collected from Wando Island were embedded in Jung Tissue Freezing Medium (Jung, Germany). Cryostat sections were fixed in ice-cold acetone and air-dried for 30 min. Tissues were equilibrated to 0.1 M PBS, and then incubated in 1% goat serum in PBS-T for 30 min to block nonspecific antibody binding. Samples were then incubated with rabbit polyclonal anti-MCL (1/200) diluted in 0.1 M PBS-T for 1 h at room temperature in a humidified chamber. A negative control was performed by replacing antibody by pre-immune rabbit IgG. Specific binding of the primary antibodies was visualized using a FITC-labeled goat anti-rabbit IgG (1/1000). The immunofluorescence images were captured on a microscope (BX50, Olympus, Tokyo, Japan).
Diagnosis of Perkinsus Infection-Two populations of the Manila clam were collected from Gomso and Gimyong on the west and south coasts of Korea. According to Park and Choi (27), Gomso is known to be a Perkinsus-infected area, whereas Gimyong is Perkinsus-free. Clams were incubated in fluid thioglycollate medium, according to Ray (28). After incubation for 1 week, the tissues were stained with Lugol's iodine and examined under a light microscope. Incidence of Perkinsus in clams was also confirmed by digesting the fluid thioglycollate medium-incubated clam tissues in 2 M NaOH, as described by Choi et al. (29) and examining Perkinsus hypnospores using a light microscope.
In-gel Digestion and Peptide Sample Preparation-All solvents used in this procedure were high-performance liquid chromatography grade. The SDS-polyacrylamide gels were stained with Coomassie Brilliant Blue. Protein bands were excised from the stained gel. Then the excised bands were washed three times with 1:1 (v/v) solution of acetonitrile/deionized water for 10 min and dehydrated with 100% acetonitrile. The bands were finally washed with 1:1 (v/v) solution of 100% acetonitrile/100 mM ammonium bicarbonate and dried using a SpeedVac. Proteins contained in the gel pieces were reduced by using 10 mM tris(2-carboxyethyl)phosphine in 0.1 M ammonium bicarbonate at 56°C for 45 min and then alkylated with 55 mM iodoacetamide in 0.1 M ammonium bicarbonate at room temperature for 30 min. Next, the washing step above was repeated on the alkylated sample. After the washing step, the gel pieces were dried and soaked in sequencing-grade trypsin solution (500 ng) on ice for 45 min. Then the gel pieces were immersed in 100 l of 50 mM ammonium bicarbonate, pH 8.0, at 37°C for 14 -18 h. The resulting peptides were extracted sequentially for 20 min with 45% acetonitrile in 20 mM ammonium bicarbonate, 45% acetonitrile in 0.5% trifluoroacetic acid, and 75% acetonitrile in 0.25% trifluoroacetic acid with agitation. The extracts containing tryptic peptides were pooled together and evaporated under vacuum.
Micro Liquid Chromatography-Mass Spectrometry/Mass Spectrometry Analysis and Protein Data Base Search-In-gel digested proteins were loaded onto fused silica capillary columns (100-m inner diameter, 360-m outer diameter) containing 8 cm of 5-m particle size Aqua C 18 reverse-phase column material. The column was placed in-line with an Agilent Technologies (Palo Alto, CA) HP 1100 quaternary liquid chromatography pump, and a splitter system was used to achieve a flow rate of 250 nl/min. Buffer A (5% acetonitrile and 0.1% formic acid) and buffer B (80% acetonitrile and 0.1% formic acid) were used to make a 90-min gradient. The gradient profile started with 5 min of 100% buffer A, followed by a 60-min gradient from 0 to 55% buffer B, a 25-min gradient from 55 to 100% buffer B, and a 5-min gradient of 100% buffer B. Eluted peptides were directly electrosprayed into an LTQ linear ion trap mass spectrometer (ThermoFinnigan, Palo Alto, CA) by applying 2.3 kV of direct current voltage. A data-dependent scan consisting of one full mass spectrometry (MS) scan (400 -1400 mass/ charge) and five data-dependent MS/MS scans were used to generate MS/MS spectra of eluted peptides. A normalized collision energy of 35% was used throughout the data acquisition. MS/MS spectra were searched against the National Center for Biotechnology Information rat protein sequence data base using Bioworks version 3.1 (Beckman Coulter, Fullerton, CA) and Sequest Cluster System (14 nodes, Thermo Electron, San Jose, CA). DTASelect was used to filter the search results, and the following Xcorr values were applied to different charge states of peptides: 1.8 for singly charged peptides, 2.2 for doubly charged peptides, and 3.2 for triply charged peptides. Manual assignments of fragment ions in each MS/MS spectra were performed to confirm the protein data base search results.

RESULTS
Induction of MCL upon Perkinsus Infection-MCL was purified from Manila clams as described previously (26). The protein has an apparent molecular mass of 138 kDa in non-reducing SDS-PAGE and consists of 74-, 34-, and 30-kDa species. However, during the purification of MCL from mucus fluid, we found that different batches of clams gave different yields, even when the amounts of starting material were similar. The different MCL yields appeared to correlate with the degree of infection of the Manila clam with Perkinsus. As shown in Table 1, body lysate from a Manila clam that was heavily infected with Perkinsus showed strong hemagglutinating activity, whereas that from a non-infected clam showed very little hemagglutinating activity. The mucus fluids from Perkinsus-infected or non-infected clams were analyzed by Western blotting using anti-MCL antibody (Fig. 1). The mucus fluid from the infected clams showed strong expression of MCL, which migrated as three species (apparent molecular masses of 74, 34, and 30 kDa). Non-infected clams hardly express MCL if any. The presence of MCL in the infected clams suggested two obvious possibilities; either Perkinsus itself produced the MCL during infection, or clams produced it upon infection, possibly for host defense. It seemed unlikely that Perkinsus makes large amounts of MCL and secretes it into mucus fluid, because Perkinsus is a intracellular endoparasitic protozoan.
To test the hypothesis that MCL might be induced by Perkinsus as part of the host defense mechanism in the clam, we purified Perkinsus trophozoites from infected clams and analyzed MCL production by Western blot analysis. A photograph of Perkinsus is shown in Fig. 2A. The infected clams contained high levels of MCL (Fig. 2B, lanes 1 and 4), whereas purified Perkinsus appeared to contain only a small amount of the 74-kDa species of MCL (Fig. 2B, lane 2). When Perkinsus were washed with 1 M NaOH for 2 min before Western blotting, the 74-kDa MCL band was lost (Fig. 2B, lane 3). The wash with NaOH did not affect the integrity of Perkinsus (data not shown), indicating that no MCL was contained within the parasite. These results indicate that MCL is bound to the surface of Per-kinsus and suggest that MCL is acquired rather than synthesized by Perkinsus.
MCL Is Synthesized in Hemocytes and Secreted as a 74-kDa Protein-Because it has been reported that invertebrate hemocytes can produce lectins (30 -32), we tested whether MCL is synthesized by Manila clam hemocytes. Hemocytes were purified from Perkinsus-infected Manila clams and analyzed by Western blotting. Body lysates and mucus fluids from inflected clams contained high levels of MCL (74-, 34-, and 30-kDa species, Fig. 3, lanes 1 and 2, respectively). By contrast, hemocytes    a Fresh clam meat (100 ϫ g) was measured and added to 100 ml of distilled water. The mixture was blended three times for 30 s at maximum speed, incubated 1 h at 4°C, and then centrifuged at 5000 ϫ g. The supernatant (100 l) was used for hemagglutination. The titer was determined by serial 2-fold dilution and defined as the dilution factor. The blood type did not affect the hemagglutination titer.
from infected Manila clams also contained MCL (Fig. 3, lane 3), but only the 74-kDa band was predominant, and the 30-and 34-kDa bands were present at only very low levels. These results suggest that MCL is synthesized as a 74-kDa precursor that may be processed to the 30-and 34-kDa species.
To test this possibility hemocytes were purified from infected clams and incubated in culture media for up to 6 h, and the MCL was probed by Western blot analyses. At the initial time isolated hemocytes contained predominantly the 74-kDa form of MCL, but upon incubation both the 30-and 34-kDa species were generated (Fig. 4A). These results indicate that the 74-kDa form of MCL is a precursor to the 30-and 34-kDa forms and that their production of the latter forms occurs by proteolytic cleavage after secretion. We also performed proteomic analyses of 74-, 34-, and 30-kDa polypeptides. MCL was purified and analyzed by reducing SDS-PAGE. Each peptide was digested with trypsin and analyzed by liquid chromatography-MS/MS. As shown in Table 2, we found the identical tryptic peptides from 74-, 34-, and 30-kDa proteins, which confirmed 34-and 30-kDa proteins came from the 74-kDa precursor. MCL seems to be a new protein, because BLAST searches of proteomic peptides showed no significant homology with any known protein. Interestingly, identical tryptic peptides were found from the 34-and 30-kDa bands. This result may be interpreted either that the 74-kDa protein is degraded to 34 then 30 kDa or the 74-kDa peptide comprises a tandem-repeated gene structure.
To test whether the cleavage occurs after secretion, we examined the form of MCL in fresh hemolymph and hemocytes from Perkinsus-infected clams. Only the 74-kDa species was detected in both samples, confirming that the MCL is generated as a 74-kDa species that is secreted and circulates as a 74-kDa protein prior to proteolytic cleavage (Fig. 4B). The nature of the protease(s) responsible for cleavage and their sites of action within the infected clam will be the subject of future studies.
Factors in the Hemolymph but Not in Perkinsus Induce MCL Synthesis by Hemocytes-Because the Perkinsus infection of the clam correlated with the synthesis of MCL in hemocytes, we tested whether live Perkinsus could trigger the synthesis of MCL in isolated hemocytes from uninfected clams. The live Perkinsus were purified and incubated for 12 h with hemocytes purified from an uninfected clam, and whole reaction mixtures were subjected to Western blotting. No MCL was detected in hemocytes incubated in the presence of Perkinsus (Fig. 5), whereas hemocytes from the infected clams contained the 74-kDa MCL. These results demonstrate that Perkinsus parasites do not directly induce MCL synthesis in hemocytes.
To explore factors that might induce MCL synthesis in hemocytes from infected clams, we tested the cell-free hemolymph from Perkinsus-infected clam. For these studies, we purified hemocytes from the Perkinsus-free clam or the Perkinsus-infected clam and incubated them with hemolymph from Perkinsus-infected clams. Hemocytes from Perkinsus-infected clams contained MCL, whereas hemocytes from Perkinsus-free clams lacked MCL (Fig. 6, left panel, 0 time). However, when hemocytes from Perkinsus-free clams were incubated with hemolymph from Perkinsus-infected clam, large amounts of MCL were generated (Fig. 6, right panel, 12 h). These results indicate that one or more unknown factors present in hemolymph from Perkinsus-infected clams induces MCL synthesis by hemocytes.   MCL Binds to GalNAc-containing Glycans with High Affinity-Previous hapten inhibition experiments showed that the lectin activity of MCL was not inhibited by a large variety of mono-and oligosaccharides, but it was inhibited weakly by Gal-NAc and glycans from the bacterial species Vibrio fisheris and Pseudoalteramonas haloplanktis. Interestingly, highly branched mannans from the marine bacteria Alteromonas aurantica, Alteromonas atlanticus, Vibrio fluvialis, and Marinomonas communis also showed high inhibitory activity (26). Because of the confusing nature of these results, we explored direct bind-ing of fluorescein-labeled MCL to a defined glycan array containing biotinylated glycans captured on streptavidin-coated microtiter plates. The detailed structures of glycans used in this experiment are listed in the home page of the Protein-Carbohydrate Interaction Core H in the National Institutes of Healthsponsored Consortium for Functional Glycomics (www.functionalglycomics.org/static/consortium/). As shown in Fig. 7, glycans containing terminal, non-reducing GalNAc were well recognized by MCL. Table 3 shows the detailed structures of glycans well recognized by MCL. Interestingly, glycans containing the LacdiNAc sequence GalNAc␤1-4GlcNAc␤1-R, which is a determinant that is abundantly expressed by many invertebrates and parasites, populated the list of highest binders to MCL, followed by glycans containing non-reducing terminal ␤-linked Gal. These results show that MCL is a highly specific lectin recognizing terminal ␤-linked GalNAc and Gal residues in glycoconjugates.
MCL Bind to Surface of Perkinsus-The studies above suggest the possibility that infection of clams by Perkinsus triggers the synthesis of MCL, which may act as a defensive molecule and/or opsonin by directly binding to Perkinsus. To investigate this possibility we examined MCL binding to purified Perkinsus hypnospores and zoospores using immunochemistry with the anti-MCL antibody. MCL bound well to both Perkinsus hypnospores and zoospores (Fig. 8, A and C, respectively) and gave typical patterns of cell-surface staining. By contrast, in control studies Perkinsus hypnospores and zoospores treated with the pre-immune IgG showed only dim fluorescence (Fig. 8, B and  D, respectively). These results demonstrate that MCL can bind to the surfaces of both Perkinsus hypnospores and zoospores.
To investigate the property of the ligands on the hypnospores, we performed binding assay in the presence of various sugars (Fig.  9). Glucose did not affect binding and mannose showed some effect on binding (Fig. 9A). Binding of MCL to hypnospores was FIGURE 6. Induction of MCL from hemocytes by unknown factors in the hemolymph. The hemolymph was purified from each Perkinsus-infected and non-infected clam, and hemocytes and hemocyte-free hemolymph were prepared as described in Fig. 4. The hemocytes from Perkinsus free-clam were incubated with indicated amounts of hemocyte-free hemolymph from Perkinsus-infected clam. After 12-h incubation, the hemocytes were collected and analyzed by Western blot analysis. inhibited by EDTA, indicating the Ca 2ϩ dependence of the lectin, and by GalNAc and mannose, demonstrating that MCL is recognizing unusual glycan epitopes containing both GalNAc and mannose on the hypnospore surface. Based on glycan array (Fig. 7) and inhibition of MCL binding to hypnospores (Fig. 9), we suggest that the structure of MCL ligand in hypnospores contains LacdiNAc as a major binding epitope and mannose as a minor contribution. However, we could not completely block all binding of MCL by these reagents, suggesting that residual binding may result from nonspecific interactions or interactions that are carbohydrate-independent.
MCL Does Not Contain Hemolytic Activity-The potential hemolytic activity of MCL was determined using human erythrocytes. Erythrocytes were isolated from heparinized blood and incubated with MCL. The absorbance of incubation media was measured at 420 nm. MCL did not cause hemolysis of human erythrocytes (data not shown), indicating that MCL does not contain intrinsic hemolytic activity.
MCL May Be Involved in Innate Immunity to Parasites-We explored whether MCL might have an alternative activity and could induce phagocytosis of MCL-coated particles. To this end fluorescent beads were coated with MCL, and their phagocytosis by hemocytes was investigated. Phagocytosis of fluorescent beads reacted with MCL was approximately twice that of the control group containing beads treated only with saline ( p Ͻ 0.05, Fig. 10, A and B). In addition, fluorescent beads that were incubated with MCL and its inhibitor free GalNAc showed reduced phagocytosis that was similar to the control saline group, indicating that GalNAc blocked phagocytosis (Fig. 10, A and  B). Soybean lectin (commercially available GalNAc-binding lectin) also triggered opsonization and phagocytic uptake but not as efficiently as MCL (Fig. 10B). These data suggest that MCL may act as an opsonin by potentially binding to Perkinsus or other parasites expressing appropriate glycan targets and by inducing their adhesion and/or phagocytosis by hemocytes.
MCL Is Localized in Encapsulated Perkinsus-Because we observed that MCL binds to Perkinsus and that MCL-coated beads enhance the phagocytosis, we investigated the actual localization of MCL in Perkinsus-infected clam tissue using immunohistochemistry. Frozen sections of infected clams were incubated with monospecific anti-MCL antibody and stained with FITC-labeled secondary antibody. Negative controls with secondary antibody only (Fig. 11, A and B) did not show fluorescence staining. However, a strong signal was detected around Perkinsus (Fig 11, C and D). The infiltrated hemocytes that produce MCL were stained strongly and were found around Perkinsus trophozoite. MCL appears to be highly enriched inside encapsulated bodies. These results strongly suggest MCL binding to Perkinsus may be involved in the encapsulation.

DISCUSSION
Our studies have explored the immunological responses of the commercially important bivalve the Manila clam (R. philippinarum) to an invading pathogen. Our results show that Perkinsus is a major pathogen of the Manila clam (R. philippinarum) and that MCL expression is a marker of Perkinsus infection. Infection of clams by Perkinsus induces the synthesis of MCL in the hemocyte as a 74-kDa precursor, which can be secreted into hemolymph where it is converted to 30-and 34-kDa species. MCL binds well to glycans containing terminal, non-reducing LacdiNAc sequences (GalNAc␤1-4GlcNAc␤1-R). Mature MCL binds to Perkinsus surfaces and may be involved in recruitment of hemocytes to Perkinsus. These stud-  ies significantly extend our earlier discovery of MCL and the finding that it exhibits Ca 2ϩ -dependent carbohydrate binding activity to mucins and Perkinsus hypnospore (26).
Perkinsus is a protozoan parasite of several commercially important marine mollusks, including the American oyster, Crassostrea virginica, the Manila clam, R. philippinarum, and the Australian black rib abalone, Haliotis rubra (33). In clams, cellular defensive activities to Perkinsus often induce pathological symptoms such as hemocytic infiltration, digestive tubule atrophy, and nodule formation on mantle and foot due to inflammation (27,29,34,35). Perkinsus is also believed to be responsible for mass mortalities in clam populations (27,34,35).
It has been reported that bacterial infection causes changes in the components of plasma or hemolymph in mollusks, such as agglutinins, anti-bacterial peptides, and lysosomal enzymes (36 -41). However, little is known about the nature of these changes and the factors responsible for their synthesis. In the crustacean horseshoe crab, a group of lectins termed tachylectins were found in the hemocytes and shown to recognize pathogen-associated molecular patterns. These lectins have different carbohydrate-binding specificities, and they were induced in response to lipopolysaccharide (42). In the tunicate Styela plicata in vivo challenge with the inflammatory agent zymosan causes generation of a collectin-like protein that is secreted into hemolymph after 96 h (30).
Ordas et al. (43) assayed defensive parameters of non-infected clams (Ruditapes decussatus) and clams heavily infected with Perkinsus atlanticus and found that the serum anti-bacterial activity and the agglutination titer were significantly different between infected and non-infected clams. In our study, lysates from non-infected clams (Rerkinsus philippinarum) had minimal agglutination activity, whereas lysates from Perkinsus-infected clams had high titer agglutinating activity (Table 1). Thus, the hemolymph of the infected clam contains a hemagglutinin that is only lowly expressed in uninfected clams, and we have identified at least one of these hemagglutinins as being the MCL lectin.
We found that Perkinsus itself does not directly induce the synthesis of MCL from hemocyte when they were incubated together; rather, hemolymph from infected clams was active in stimulating hemocytes to synthesize MCL (Figs. 5 and 6). These results suggest that infection of clams by Perkinsus leads to secretion of immune related factors, possibly cytokines, which trigger the synthesis of MCL in hemocytes. In the tunicate, S. plicata, an inflammatory cytokine, interleukin 1-like protein, has been reported (44). Our results strongly suggest the MCL may be involved in defense mechanisms and lectin synthesis is triggered by the infection.
MCL is synthesized as a 74-kDa protein by hemocytes from infected clams and secreted into hemolymph, and it is eventually cleaved to 30-and 34-kDa species. We previously reported that MCL consist of three subunits when purified from whole lysate in the presence of proteases inhibitors (30,34, and 74 kDa) (26). Our present results show that MCL is synthesized as a 74-kDa precursor protein that is processed to 30-and 34-kDa species after secretion into hemolymph. In the non-reducing SDS-PAGE, MCL behaves as a dimer (138 kDa) before and after cleavage of the 74-kDa species to the 30-and 34-kDa species (data not shown). Thus, MCL is synthesized in hemocytes where it is dimerized by disulfide bonds and secreted. Apparently, cleavage of the 74-kDa subunit does not affect carbohydrate-binding activity, based on the observation that MCL in body lysates binds to the affinity absorbents equivalently to pro- We reported earlier that mucin types 1 and 2, which contain GalNAc residues, exhibited the best inhibition of MCL-mediated hemagglutination and that highly branched mannan structures from marine-originated bacteria also showed strong inhibition (26). However, because these glycan structures are not well characterized, and high branched mannan from Escherichia coli did not inhibit hemagglutination of MCL, the results were not conclusive whether the general high mannose structure inhibited agglutination. In glycan array analyses MCL showed the highest binding to glycans containing the Lacdi-NAc motif (GalNAc␤1-4GlcNAc␤1-R) and to glycans with terminal ␤-Gal and ␤-GalNAc structures. Many lectins from marine invertebrates have affinity toward to GalNAc residues (45). Interestingly, a mannose-binding lectin-like lectin from invertebrates has also been reported to have binding specificity toward to GalNAc-containing glycans instead of mannose (48). Thus, the recognition of LacdiNAc-type glycans, which are abundantly expressed in many invertebrates and parasites, is consistent with our predicted role of MCL in innate immune responses in the clam.
Broad specificity of innate immune-related lectins may be a general biological phenomenon. A previous hapten inhibition assay showed that lipopolysaccharides derived from bacteria-infecting marine organisms but not lipopolysaccharides from other sources are potent inhibitors of MCL (26). The tachylectin-2 from horseshoe crab binds to both D-GlcNAc and D-GalNAc and lipopolysaccharide from Gram-negative bacteria (42). The C-type lectin perlucin from the abalone has been reported to have broad binding specificity with the ability to bind glycoproteins containing galactose or mannose/glucose (46). In the absence of immunoglobulin molecules, broad binding specificity of lectin may be required for effective humoral non-self-recognition in the marine invertebrate.
MCL can act as an opsonin to stimulate in vitro phagocytosis (Fig.  10, A and B). We also made the novel observation that MCL synthesis is induced by Perkinsus infection of clams. It has been reported that many lectins from marine invertebrates show antibacterial activity. The lectin from the horse mussel M. modiolus (21) exhibited strong antibacterial activity against tested Vibrio strains. The gigalin H and gigalin E lectins from oyster (Crassostrea gigas) acted as opsonins to stimulate in vitro phagocytosis of the marine bacterium Vibrio anguillarum (47). However, the induction of lectin involvement in defense against parasitic invasion, as we have seen for Perkinsus, has not been reported previously. Our present study shows that only the hemocyte from Perkinsus-infected clams produces MCL and that MCL is able to bind to the Perkinsus (Figs. 2 and 8). However, Perkinsus trophozoite is probably too large to allow direct engulfment by hemocytes. MCL may function to recruit hemocytes around Perkinsus and/or function in phagocytosis of parasite-derived material. The infiltration of hemocytes (Fig. 11) around Perkinsus is a well known symptom of Perkinsus-infected clam in the gill (50). Clusters of Perkinsus trophozoites were observed encapsulated in amorphous eosinophilic material, forming cyst-like structures (Fig. 11) (50). Parasite encapsulation is well characterized in insects (51). For example, the C-type lectin immulectin has FIGURE 10. Effect of MCL on phagocytosis by hemocytes. Hemocytes prepared from clam hemolymph were incubated with fluorescent latex beads that had been coated with MCL (with or without 2 mg/ml GalNAc) or control incubated in filtered seawater. In addition, one set of control incubations included beads incubated in filtered seawater and then mixed with cytochalasin B to block phagocytosis, as described under "Materials and Methods." The extent of phagocytosis by hemocytes was measured by flow cytometry. Examples of flow cytometric data showing the extent of phagocytosis of beads coated with MCL and MCL in the presence of GalNAc (A). Compiled data of the percent phagocytosis by hemocytes of beads coated as indicated (B). The data are the average of 18 experiments. been reported to trigger encapsulation of parasites as a pattern recognition receptor (51). MCL may act as a pattern recognition receptor to activate lectin pathways of complement or to trigger encapsulation of parasites. In summary, our studies indicate that MCL is a parasite-inducible lectin that can bind parasites and promote their opsonization or encapsulation.