A new fungal lectin recognizing α(1-6)-linked fucose in the N-glycan

In this report, we describe a new lectin from the fungus Rhizopus stolonifer that agglutinates rabbit red blood cells. Agglutinating activity was detected in the extract of mycelium-forming spores cultured on agar plates but not in the mycelium-forming no spores from liquid medium. This lectin, which we designated R. stolonifer lectin (RSL), was isolated by affinity chromatography with porcine stomach mucin-Sepharose. SDS-polyacrylamide gel electrophoresis and mass spectral analysis showed that RSL is ∼4.5 kDa, whereas gel filtration indicated a mass of 28 kDa. This indicates that the lectin is a hexamer of noncovalently associated RSL monomers. RSL activity was very stable, since it was insensitive to heat treatment at 70 °C for 10 min. Analysis of RSL binding specificity by both microtiter plate and precipitation assays showed that N-glycans with l-fucose linked to the reducing terminal GlcNAc residues are the most potent inhibitors of RSL binding, whereas N-glycans without α(1–6)-linked fucose residues are ∼100-fold weaker inhibitors of binding. Oligosaccharides with α(1–2, –3, and –4) linkages showed no inhibition of binding in these assays. In a mirror resonance biosensor assay, high affinity binding was observed between RSL and the glycopeptide of bovine γ-globulin, which has N-glycans with α(1–6)-linked fucose residues. However, RSL showed only a weak interaction with the glycopeptide of quail ovomucoid, which lacks fucose residues. Finally, capillary affinity electrophoresis studies indicated that RSL binds strongly to N-glycans with α(1–6)-linked fucose residues. Together, these results show that RSL recognizes the core structure of N-glycans with α(1–6)-linked l-fucose residues. This specificity could make RSL a valuable tool for glycobiological studies.

Glycoconjugates possess a variety of structures, including glycoproteins and glycolipids, and are found on the surfaces of animal, plant, and microorganism cells. Determination of the structures and the distribution of glycoconjugates on cell surfaces is important for understanding their biological function. Lectins are useful in this regard because they have a specificity for defined carbohydrate structures. Lectins are also used for isolation and fractionation of glycoconjugates, cell separation, identification of microorganisms, and drug delivery (1).
Fucose-containing glycoconjugates on cell surfaces are known to be important in cell-to-cell recognition and signaling. For example, sialyl Lewis X on endothelial cells is a ligand for selectins, which are involved in leukocyte recruitment to sites of inflammation (2). In addition, some pathogens can adhere to fucosylated glycoconjugates on surfaces of their host cells (3). Similarly, during murine fertilization, a fucosyl residue on the zona pellucida of the egg is required for adhesion to the high affinity sperm-binding ligand (4). Fucose-containing glycoconjugates, such as H-blood antigens on tumor cells, also modulate cell growth, invasion, and metastasis (5).
Although many lectins have been identified, only a few are fucosyl-specific, including the Lotus tetragonolobus (6) and Ulex europaeus (7) lectins from plants as well as the Anguilla lectin from eel (8) and the Aleuria aurantia lectin from mushroom (9). Therefore, an additional lectin that recognizes Lfucose-containing oligosaccharides should be a valuable tool for glycobiological research.
Lectins are found in a variety of organisms including fungi. Although many fungal lectins have been reported recently (10 -12), most of them are from mushroom. Therefore, in the current study, we surveyed the lectins in fungi other than mushroom. Using a rabbit erythrocyte hemagglutination assay, we identified an agglutinating activity in an extract of the mycelium of the fungus Rhizopus stolonifer. Of several monosaccharides tested, only L-fucose inhibited the agglutinating activity of the crude extract, suggesting that the agglutinin was a L-fucosespecific lectin. We purified the lectin and examined its physical-chemical properties and carbohydrate binding specificity.
Preparation of Oligosaccharides-Glycopeptides were prepared from bovine ␥-globulin, quail ovomucoid, and apotransferrin as described previously (13). The oligosaccharides were released from their glycopeptides by hydrazinolysis (14) and separated by HPLC on a Cosmosil 5NH 2 column (4.6 ϫ 250 mm, Nakarai Tesque, Kyoto, Japan) (15). Man 3 GlcNAc 2 (9) and Gal 2 GlcNAc 2 Man 3 GlcNAc 2 Fuc (16) were obtained from glycopeptides of quail ovomucoid and bovine ␥-globulin, respectively. The oligosaccharides GlcNAc 2 Man 3 GlcNAc 2 (10) and Gal 2 GlcNAc 2 Man 3 GlcNAc 2 (11) were prepared from the glycopeptide of apotransferrin. Briefly, the oligosaccharide (10 mg) obtained from the glycopeptide of apotransferrin was dissolved in 1 ml of 50 mM sodium acetate buffer, pH 5.5, and was incubated with neuraminidase (1 unit) at 37°C for 24 h. The mixture was heated in a boiling water bath for 3 min and centrifuged, and the supernatant was subjected to HPLC on a Cosmosil 5NH 2 column to obtain the asialo-oligosaccharide (11). This asialo-oligosaccharide (5 mg) was dissolved in 200 l of 50 mM sodium acetate buffer, pH 3.5, and digested with ␤-galactosidase (1 unit) at 37°C for 24 h. Finally, the product (10) was purified by HPLC. The structures of the three oligosaccharides (9 -11) obtained were confirmed by mass spectrometry and NMR (16,17). Oligosaccharides GlcNAc 2 Man 3 GlcNAc 2 Fuc (13) and Man 3 GlcNAc 2 Fuc (12) were prepared by enzyme digestion of oligosaccharide 16. Briefly, the oligosaccharide 16 (5 mg) was dissolved in 200 l of 50 mM sodium acetate buffer, pH 3.5, and digested with ␤-galactosidase (1 unit) at 37°C for 24 h. The product (13; ϳ4 mg) was purified by HPLC. Next, the oligosaccharide was digested with ␤-N-acetylhexosaminidase (2 units) in 100 l of 50 mM sodium citrate buffer, pH 5.0, at 37°C for 24 h to produce oligosaccharide 12, which was purified by HPLC. Each portion (50 g) of the products obtained from 16 was digested with L-fucosidase (200 milliunits) in 60 l of 50 mM sodium citrate buffer, pH 5.0, at 37°C for 24 h, and the aliquots were subjected to HPLC to confirm that the products were the oligosaccharide 10 and 9, respectively.
Preparation of Fluorescently Labeled Oligosaccharides-Oligosaccharides were released from AGP and bovine ␥-globulin with peptide-N 4 -(acetyl-␤-D-glucosaminyl) asparagine amidase and labeled with APTS for capillary affinity electrophoresis as described previously (18). Briefly, the glycoprotein (1 mg) was dissolved in 20 mM phosphate buffer, pH 7.0 (40 l). Peptide-N 4 -(acetyl-␤-D-glucosaminyl) asparagine amidase (5 milliunits, 5 l) was added, and the solution was incubated for 24 h at 37°C and then heated in a boiling water bath for 5 min. Following centrifugation, the supernatant was evaporated to dryness using a centrifugal vacuum evaporator (SpeedVac, Farmingdale, NY). The residue was dissolved in 2 M aqueous acetic acid (50 l), and the mixture was incubated for 3 h at 80°C to remove sialic acids from the oligosaccharides. The residue was dissolved in 15% aqueous acetic acid (5 l) containing 100 mM APTS. A freshly prepared solution of 1 M NaBH 3 CN in tetrahydrofuran (5 l) was added, and the mixture was overlaid with mineral oil (100 l, n D ϭ 1.4670, d ϭ 0.838; Aldrich) to prevent evaporation of the solvent. The mixture was kept for 90 min at 55°C. Water (200 l) was added to the mixture, and the fluorescent yellowish aqueous phase was collected. The aqueous layer was applied to a column of Sephadex G-25 (1 ϫ 50 cm) equilibrated with water. The fluorescent fractions containing oligosaccharides were collected and evaporated to dryness. The residue was dissolved in water (100 l), and a portion (20 l) was put aside for capillary affinity electrophoresis.
Lectin Purification-R. stolonifer was cultured on an agar plate for 2 weeks at 30°C. A 10-ml aliquot of PBS was poured onto the cultured plate, and mycelium was suspended with a glass pipette. The suspension was centrifuged at 15,000 ϫ g for 10 min, and the supernatant was applied to a PSM-Sepharose column (1.0 ϫ 10 cm) previously equilibrated with PBS. Elution of protein was monitored by A 280 . The column was washed with PBS containing 1 M NaCl until the A 280 of the eluate TABLE I List of oligosaccharides was below 0.01. The lectin was eluted with PBS containing 0.1 M L-fucose at 50°C. The fractions containing lectin were dialyzed against distilled water using dialysis membrane (molecular weight cut-off 3500; Spectra/Por) and then lyophilized. The lyophilized material was dissolved in 100 mM Tris/HCl buffer, pH 7.5, containing 8 M urea and applied to a Sephacryl S-200 column (1.0 ϫ 100 cm) previously equilibrated with the same buffer. The main lectin-containing peak was pooled, dialyzed against distilled water, and lyophilized.
Protein Estimation-Protein concentrations were determined by the method of Lowry et al. using bovine serum albumin as a standard (19).
Amino Acid Composition of the Lectin-The amino acid composition of the purified lectin was determined using an Applied Biosystems model 420/amino acid analyzer. Samples were hydrolyzed for 24, 48, and 72 h at 110°C in 6 M HCl. The purified lectin was reduced and alkylated to obtain the S-pyridylethylated lectin. Cysteine was determined as S-(4-pyridylethyl)-cysteine (20). Tryptophan was determined spectrophotometrically by the procedure of Edelhoch (21). The sugar content was determined using the phenol/sulfuric acid method of Dubois et al. with glucose as the standard (22).
SDS-PAGE and Amino Acid Sequencing of the Lectin-Denatured proteins were separated in the presence and absence of 2-mercaptoethanol by Tricine-SDS-PAGE using a 12.5% slab gel and stained with Coomassie Brilliant Blue R250 (23). After SDS-PAGE of the S-pyridylethylated lectin, followed by electrophoretic transfer onto polyvinylidene difluoride (0.45 m; Millipore Corp., Bedford, MA), the N-terminal amino acid sequence of the subunit was determined by automatic Edman degradation using a model 477A protein and peptide sequencing system (Applied Biosystems).
Determination of Molecular Mass-The molecular mass of the purified lectin was determined by SDS-PAGE in the presence and absence of 2-mercaptoethanol, gel filtration, and mass spectral analysis.
The molecular mass from SDS-PAGE was estimated by comparison with standard proteins (Polypeptide standards; Bio-Rad). Gel filtration for estimation of the molecular mass was performed using a TSK SW2000 column (7.8 ϫ 300 mm, TOSO; Tokyo) equilibrated with 0.1 M Tris/HCl, pH 7.0. Elution was carried out with the same buffer at a flow rate of 1 ml/min. The calibration proteins were used as follows: ovalbumin (45 kDa), chymotrypsinogen A (25 kDa), cytochrome c (12.5 kDa), and aprotinin (6.5 kDa).
The molecular mass of the purified lectin and S-pyridylethylated lectin was also measured by matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometry using a Voyager DE PRO (PerkinElmer Life Sciences) equipped with a 337-nm UV laser. The instrument was operated in linear operation using positive polarity. Samples (0.5 l) were applied to a polished stainless steel target, and a solution (0.5 l) of 2,5-dihydroxybenzoic acid in 1:1 ethanol/water was added. The mixture was allowed to air-dry at ambient temperature.
Agglutination Activity Assay-Agglutination activity was measured in microtiter plates using serial 2-fold dilutions of lectin as described previously (24). Each well contained 25 l of rabbit red blood cells (10% suspension), 25 l of lectin solution diluted with PBS, and 50 l of saline. To study effect of carbohydrate and metal ions on agglutination, carbohydrate or metal ion solution in saline (50 l) was added, replacing the equivalent volume of saline. Agglutination was observed 1 h later.
Biotinylation of Lectin and Glycopeptides-RSL and glycopeptides were biotinylated using biotinyl N-hydroxysuccinimide ester (25). The succiminidyl ester (20 l; 50 mg/ml in dimethylformamide) was added dropwise to a solution of lectin or glycopeptide (1.0 ml; 1.0 mg/ml in PBS) while it was agitated with a vortex mixer. After incubation for 90 min at room temperature, the reaction mixture was dialyzed against PBS using dialysis membrane (molecular weight cut-off of 10,000 and 1000 (Spectra/Por) for lectin and glycopeptides, respectively) and stored at Ϫ20°C until use.
Microtiter-based Binding and Inhibition Assay-Microtiter wells (96 wells, flat bottom, Asahi Tekunogurasu, Tokyo, Japan) were incubated with bovine thyroglobulin (200 l; 25 g/ml in water) overnight at 4°C. Next, the wells were blocked with PBS containing 0.25% BSA (PBS/ BSA) overnight at 4°C. After washing four times with PBS/BSA, the wells were incubated for 30 min at room temperature with 50 l of biotinylated RSL and the same volume of PBS. The wells were next washed four times with PBS/BSA. The wells were incubated with avidin-horseradish peroxidase conjugate (200 l; 2 g/ml in PBS containing 3% BSA) for 15 min at room temperature and then washed four times with PBS/BSA. Finally, the wells were incubated with 100 l of horseradish peroxidase substrate solution at room temperature. Color development was terminated after 10 min by the addition of 2 M sulfuric acid (100 l). The A 405 of each well was measured with a microplate reader (model 550; Bio-Rad). For the inhibition assay, PBS was replaced with PBS containing carbohydrate in the reaction mixture. Biotinylated lectin was used at 2 g/ml. To examine the effect of pH on the binding reaction, sodium acetate, sodium phosphate, and Tris/HCl buffers were used instead of PBS as the solvents for the pH ranges 3-5, 6 -7, and 8 -10, respectively. All binding assays were performed in triplicate, and the averages are presented.
Quantitative Precipitation and Inhibition Assay-The quantitative precipitation assay was carried out in microcentrifuge tubes essentially as described by Mo et al. (26). Various amounts of glycoproteins (0 -100 g) were added to 5 g of the purified RSL in 160 l of PBS, and the mixture was incubated at 37°C for 1 h, followed by 4°C for 48 h. The precipitate were collected by centrifugation and washed three times with 150 l of ice-cold PBS. The protein content of the precipitate was determined by the method of Lowry (19). For the inhibition assay, various amounts of inhibitors were added to the solution of the purified RSL (5 g) and thyroglobulin (50 g) in a final volume of 160 l of PBS. The protein content of the precipitate was determined as described above, and the percentage of inhibition was calculated.
Assay of Lectin Binding Using a Resonant Mirror Biosensor-A lectin-binding assay was performed using an IAsys resonant mirror biosensor (Affinity Sensors Ltd., Saxon Hill, Cambridge, UK). Streptoavidin was immobilized on the carboxymethyl dextran surface of a cuvette. Next, biotinylated glycopeptides of bovine ␥-globulin or quail ovomucoid were immobilized according to the manufacturer's instructions (0.7 and 1.2 ng of glycopeptide bound/mm 2 of bovine ␥-globulin and quail ovomucoid, respectively). Binding reactions were carried out in a final volume of 100 l of phosphate-buffered saline-Tween (PBST) (140 mM NaCl, 10 mM NaH 2 PO 4 , 0.02% (v/v) Tween 20, pH 7.2) at 25°C. The association phase of the binding reaction was initiated by the addition of lectin (1-5 l) to PBST in the cuvette. After 5 min, the dissociation phase was initiated by quickly washing the cuvette three times with 100 l of PBST. Regeneration of the cuvette was achieved by repeated washing with 20 mM HCl. FASTfit software (Affinity Sensors) was used to calculate the dissociation constant (K D ).
Capillary Affinity Electrophoresis-Capillary affinity electrophoresis was performed using a P/ACE MDQ glycoprotein system (Beckman Coulter) equipped with an argon-laser-induced fluorescence detection system according to the method of Nakajima et al. (18). Briefly, separation was performed at 25°C using an eCAP N-CHO-coated capillary (20-cm effective length, 30-cm total length, 50-m inner diameter; Beckman Coulter) and detection was performed using a 520-nm filter for emission and a 488-nm argon-laser for excitation. Tris acetate buffer (100 mM, pH 7.4) was used as the electrolyte. Prior to capillary affinity electrophoresis, APTS-labeled carbohydrates were analyzed. Next, the capillary was filled with same electrolyte containing a lectin at the specified concentration, and electrophoresis was performed. Data were collected and analyzed using 32 Karat software (version 4.0; Beckman Coulter).

RESULTS
Lectin Purification-Rabbit erythrocyte agglutinating activity was detected in the extract of mycelium-forming spores cultured on agar-medium. However, when the fungus was cultured in liquid medium, it did not form spores, and agglutinating activity was not detected in either the culture filtrate or the mycelium extract. Therefore, fungus cultured on agar medium was used as a source of the lectin.
The inhibitory effect of various carbohydrates on the agglutinating activity of the crude extract was similar to that of the purified lectin (Table II). Specifically, L-fucose but none of the other monosaccharides tested inhibited the agglutinating ac-tivity. Among the glycoproteins examined, thyroglobulin was the most potent inhibitor of agglutinating activity. PSM also demonstrated potent inhibitory activity in the agglutination assay.
We attempted to purify the lectin using thyroglobulin-Sepharose, but little lectin activity was recovered. PSM-Sepharose was effective at binding the lectin, but only a small amount of lectin was eluted with L-fucose even at a concentration of 0.2 M when carried out at room temperature. However, ϳ80% of the lectin activity was recovered using a 0.1 M L-fucose solution at 50°C (Fig. 1).
Several peaks were revealed in gel filtration by A 280 read- ings, but only the fractions in the main peak showed agglutinating activity (Fig. 2).
Therefore, the fractions in the main peak were pooled and used as the purified lectin (designated R. stolonifer lectin, RSL). After dialysis and lyophilization, about 2 mg of purified lectin were obtained from 20 culture plates.
Agglutinating Activity-RSL agglutinated rabbit erythro-cytes at minimum concentration of 4 g/ml. Metal ions such as 100 mM Ca 2ϩ , Mg 2ϩ , or Mn 2ϩ or metal ion chelation with 10 mM EDTA did not affect the agglutination activity of RSL. This indicates that the lectin does not require metal ions for binding. The lectin maintained full activity following heating at 70°C for 10 min, but approximately half of the activity was lost following heating at 80°C for 10 min. The inhibitory effect of various carbohydrates on the agglutinating activity of the purified lectin is shown in Table II. Only L-fucose completely inhibited the agglutinating activity at 25 mM. The other monosaccharides, lactose and N,NЈ-diacetylchitobiose, showed no inhibition at up to 100 mM. Fucoidan, a polymer of L-fucose, was also an inhibitor of RSL agglutination activity. Among the glycoproteins examined, thyroglobulin was the most potent inhibitor, and PSM also showed inhibition. Collectively, these results suggest that RSL is a fucose-binding lectin. Molecular Mass and Molecular Structure-The purified lectin showed one band of ϳ4.5 kDa on reducing or nonreducing SDS-PAGE (data not shown). Similarly, MALDI-TOF mass spectrometry showed that the molecular masses of the native lectin and S-pyrimidylethylated lectin are 4570 and 4783 Da, respectively (Fig. 3).
The apparent molecular mass of the native lectin was estimated by TSK G2000SW gel filtration to be ϳ28 kDa (data not shown). This suggests that the native lectin is a hexamer of noncovalently associated subunits. Neutral carbohydrate was not detected using the phenol/sulfuric acid assay.
Amino Acid Composition and N-terminal Amino Acid Sequence-As shown in Table III, the purified lectin contained  high amounts of aspartic acid/asparagine and glutamic acid/ glutamine as well as glycine and valine.
The lectin also contains two residues of cysteine but no methionine. Automatic Edman degradation was performed on the S-pyridylethylated lectin, and the first 28 amino acid residues were identified as H 2 N ϩ -IDPVNVKKLQCDGD-TYKCTADLDFGDGR. This sequence was analyzed using the BLAST P algorithm, but high homology sequences were not found in the protein sequence data banks.
Microtiter Plate Binding and Inhibition Studies-Binding of biotinylated RSL to immobilized bovine thyroglobulin was saturable, reaching a maximum at 3-5 g/ml lectin. We also found that the optimum pH for binding was between 6 and 8 (data not shown). As mentioned above, agglutination inhibition studies suggested that RSL is a fucose-binding lectin. To elucidate the detailed carbohydrate binding specificity of RSL, we examined the inhibitory effects of various oligosaccharides on the binding of biotinylated lectin to immobilized thyroglobulin using a microtiter plate assay (Fig. 4).
The most potent inhibitor was 12, followed by 13 and 16, all of which are N-linked oligosaccharides with fucose attached to the reducing terminal GlcNAc via an ␣(1-6) linkage. In contrast, the oligosaccharides without fucose 9 -11 were 100-fold weaker inhibitors than the corresponding oligosaccharides 12, 13, and 16. Among the monosaccharides examined, only Lfucose showed inhibition, although inhibition was weak. The other monosaccharides examined, lactose and N,NЈ-diacetylchitobiose, did not show inhibition at up to 400 mM (data not shown). Oligosaccharide 2 was a more potent inhibitor than fucose, but oligosaccharides 3-8, which contain fucose residues with ␣(1-2), ␣(1-3), or ␣(1-4) linkages, showed no inhibition at a concentration of 2 mM. Quantitative Precipitation and Inhibition-RSL formed a precipitate with bovine thyroglobulin, but it did not form a pronounced precipitate with porcine stomach mucin or bovine ␥-globulin (data not shown). Therefore, inhibition experiments were performed in solution using the reaction between RSL and bovine thyroglobulin. The results were similar to those obtained with the microtiter plate assay (Table IV).
Assay of Lectin Binding Using a Resonant Mirror Biosensor-Crocus sativus lectin (CSL) was reported to recognize Man 3 GlcNAc 2 in the core structure. However, ␣(1-6)-linked L-Fuc is irrelevant to CSL binding (13). Therefore, to compare the binding specificity of RSL and CSL, we analyzed the interactions of RSL or CSL with the oligosaccharides of bovine ␥-globulin and quail ovomucoid using the resonant mirror biosensor assay. Biotinylated glycopeptides were coupled to a streptoavidin-immobilized cuvette surface, and the binding reactions were monitored in real time by varying the concentrations of lectins (Fig. 5).
RSL clearly interacted with the glycopeptide of bovine ␥-globulin; however, we did not observe an interaction between CSL and the glycopeptide of bovine ␥-globulin. The lack of interaction between CSL and the glycopeptide of bovine ␥-globulin is probably due to masking of the ␣(1-3)-linked Man residue by GlcNAc (13). On the other hand, CSL strongly interacted with the glycopeptide of quail ovomucoid. RSL, in   showed no inhibition at 2 mM in both assays. b Concentrations of carbohydrates required for 50% inhibition were obtained from the inhibition curve (Fig. 4).
c Maximum concentration examined (percentage of inhibition observed). contrast, showed only a weak interaction with the quail ovomucoid glycopeptide. The dissociation constants (K D ) calculated were 8.1 ϫ 10 Ϫ7 , 3.2 ϫ 10 Ϫ4 , and 1.0 ϫ 10 Ϫ7 M for the interactions between RSL and bovine ␥-globulin oligosaccharide, RSL and quail ovomucoid oligosaccharide, and CSL and quail ovomucoid oligosaccharide, respectively. The K D for the interaction between CSL and bovine ␥-globulin oligosaccharide was not determined.
Capillary Affinity Electrophoresis-Capillary affinity electrophoresis is based on high resolution separation of fluorescently labeled carbohydrates by capillary electrophoresis in conjunction with laser-induced fluorescent detection in the presence of a various concentrations of lectin. This method allows the assessment of the lectin binding specificity of carbohydrates derived from a glycoprotein as well as kinetic determinations such as the binding affinity of each carbohydrate.
Using this method, we attempted to analyze the interactions of RSL with fluorescently labeled oligosaccharides derived from bovine ␥-globulin, quail ovomucoid, and asialo-AGP. Oligosaccharides of bovine ␥-globulin were isolated in capillary electrophoresis as shown in Fig. 6.
The addition of RSL clearly altered the migration of the four bovine ␥-globulin-derived oligosaccharides, 13-16. These all have ␣(1-6)-linked fucose residues. The migration speed of the oligosaccharides decreased, depending on the concentration of RSL added, indicating that RSL interacted with these oligosaccharides. Using the method of Taga  AGP were also isolated on capillary electrophoresis as reported previously. However, RSL did not significantly affect the migration of those carbohydrates even at 15 M. This indicates that there is little interaction between RSL and the oligosaccharides from quail ovomucoid or asialo-AGP (data not shown). DISCUSSION Lectins have been found in plant pathogenic fungi, including R. solani (28), Rhizoctonia crocoroum, Athelia rolfsi (29), and Sclerotium rofsii (30), and some of them may participate in the infection of host cells by the fungi. R. stolonifer is a mucoraceous fungus and usually reproduces asexually by forming spores. This fungus can spoil stored processed foods, fresh fruits, and vegetables (31). Agglutinating activity was detected in the extract of the mycelium-forming spores of R. stolonifer when it was cultured on agar medium. However, agglutinating activity was not detected in the extract of the mycelium when the fungus was cultured in liquid medium, where spores are not observed. These findings suggest that RSL may be produced developmentally and involved in spore formation by the fungus. Indeed, similar observations have been made in cellular slime molds, such as Dictyostelium discoideum (32) and Polyphondylium pallidum (33). In these cases, extracts of the aggregating cells had potent hemagglutinating activity, whereas extracts of feeding amoebae lacked activity. Furthermore, Cooper and Barondes (34) demonstrated the production of two different lectins by Dictyostelium discoideum that were developmentally regulated.
To elucidate the detailed carbohydrate binding specificity of RSL, inhibition studies were performed in a microtiter plate assay (solid phase method) as well as a precipitation assay (solution phase method). Similar results were obtained in the two assays. Among the monosaccharides examined, only Lfucose inhibited the binding of RSL to thyroglobulin. L-Fucose ␣(1-6)GlcNAc was a 100-fold more potent inhibitor than Lfucose. In contrast, oligosaccharides having L-fucose residues with ␣(1Ϫ2, Ϫ3, or Ϫ4) linkages including Lewis a and Lewis x showed no inhibition, even at 2 mM, indicating that RSL specifically recognizes L-fucose residues with an ␣(1-6)linkage. The most potent inhibitor of RSL was 12, which is the core structure of N-glycan with an L-fucose residue attached to GlcNAc at the reducing terminus via an ␣(1-6) linkage. Furthermore, the presence of ␤(1-2)-linked GlcNAc residues (13) and Gal␤(1-4)GlcNAc residues (16) had little effect on lectin binding, indicating that these additional GlcNAc residues do not interfere with the binding. On the other hand, a lack of fucose markedly reduced the binding affinity, and oligosaccharides with fucose residues (12, 13, and 16) were 100-fold more potent inhibitors than the corresponding oligosaccharides lacking fucose residues (9, 10, and 11). Thus, an L-fucose residue with ␣(1-6) linkage is critical for binding. Collectively, the inhibition studies show that both an L-fucose ␣(1-6)GlcNAc residue and the core structure of N-glycan, Man 3 GlcNAc 2 , are important for the binding of RSL. In addition, GlcNAc or Gal␤(1-3)GlcNAc residues linked to Man residues in the trimannosyl core do not interfere with the binding.
Bovine ␥-globulin and PSM were not precipitated with RSL. Probably, the number of carbohydrate chains involved in precipitating the lectin may be not enough in bovine ␥-globulin and PSM. However, at present we do not clearly understand why these gycoproteins failed to precipitate RSL. Although PSM did not precipitate RSL, RSL was absorbed to PSM immobilized on an affinity gel. The binding of the lectin to the affinity gel may be due to an enhanced binding affinity as a result of clustering of the PSM oligosaccharides on the affinity gel.
Bovine thyroglobulin is known to possess N-glycans including complex type oligosaccharides with ␣(1-6)-linked fucose residues, although most of the oligosaccharide structures remain unclear. Bovine ␥-globulin mainly contains complex type oligosaccharides with ␣(1-6)-linked fucose residues such as 13-16. Asialo-AGP also has N-glycans, including ␣(1-2)-linked fucose residues in the outer chain of oligosaccharides (17 and 18). In contrast to these glycoproteins, the main oligosaccharide of quail ovomucoid N-glycan lacks fucose residues (9). Therefore, to examine the specificity for natural oligosaccharides, we examined the interaction of RSL with oligosaccharides of bovine ␥-globulin, asialo-AGP, and quail ovomucoid. These experiments were carried out using capillary affinity electrophoresis (liquid phase assay) or resonant mirror biosensor (solid phase assay). In both assays, RSL clearly interacted with oligosaccharides having ␣(1-6)-linked fucose residues derived from bovine ␥-globulin but not with those lacking ␣(1-6)-linked fucose residues that were derived from quail ovomucoid. Capillary affinity electrophoresis did not show a clear interaction between RSL and the asialooligosaccharides of AGP, indicating that ␣(1-2)-linked fucose residues in the outer chain of oligosaccharides (17,18) have little influence on binding. This agrees with the observation that Lewis a was not inhibitory in the microplate and quantitative precipitation assays.
Using the resonant mirror biosensor assay, we determined dissociation constants (K D ) for the interactions of RSL and CSL with glycopeptides. The K D value (1.0 ϫ 10 Ϫ7 M) for the inter- action between CSL and the oligosaccharide of quail ovomucoid obtained in this study is very close to a previously reported value of 1.5 ϫ 10 Ϫ7 M for the interaction between CSL and Taka amylase (which has a Man 5 GlcNAc 2 chain) that was determined by a surface plasmon resonance assay (13).
We also used capillary affinity electrophoresis to determine binding affinity constants, but values were lower than those determined in the resonant mirror biosensor assay. This may be due to destruction of the GlcNAc residue at the reducing termini of oligosaccharides by reductive amination during modification with APTS. Regardless of the differences between the methods, our kinetic analyses strongly support the concept that RSL recognizes the core structure of N-glycan with ␣(1-6)linked fucose residues.
Three lectins, U. europaeus agglutinin I, Lotus tetragonolobus agglutinin, and Aleuria aurantia lectin are well known to be specific for fucose, and their carbohydrate binding has been reported in many studies. U. europaeus agglutinin I preferentially binds to Fuc␣(1-2)Gal␤(1-4) GlcNAc, whereas the other lectins are much less specific for this oligosaccharide structure (35). L. tetragonolobus agglutinin is inhibited by oligosaccharides containing fucosyl residues with various linkages such as Fuc␣(1-6)GlcNAc, Fuc␣(1-3)GlcNAc, and Fuc␣(1-2)Gal␤(1-4)Glc (36). In addition, oligosaccharides containing the Lewis x antigen Gal␤ (1)(2)(3)(4)[Fuc␣(1-3)]GlcNAc-R are potent inhibitors of L. tetragonolobus agglutinin. Thus, the carbohydrate-binding specificity of these two lectins is different from that of RSL. A. aurantia lectin, which was isolated from the fruiting bodies of the orange peel fungus, binds to N-linked oligosaccharides with L-fucose linked to the proximal GlcNAc residue, such as in RSL (9). However, these appear to be distinct lectin because A. aurantia lectin has a different molecular mass and N-terminal amino acid sequence than RSL. Although Lens culinaris lectin and Pisum sativum lectin (37) require an L-fucosyl unit on the N,N-diacetylchitobiosyl residue of complex oligosaccharides for high affinity binding, RSL appears to be distinct from these two lectins, because they are primarily specific for ␣-mannopyranosyl residues. Recently, a bacterial lectin produced by plant pathogen, Ralstonia solanacearum, was reported to be L-fucose-specific (38). Again, RSL appears to be distinct from this lectin, because R. solanacearum lectin is inhibited by D-galactose and D-arabinose in addition to L-fucose, because the molecular mass of the R. solanacearum subunit is ϳ9.9 kDa and because the N-terminal sequences of the two lectins differ.
RSL appears to have the smallest molecular mass subunits (ϳ4.5 kDa) that have been reported for a lectin. Due to this small size, it may be possible to chemically synthesize the lectin. Furthermore, it is very stable; no loss of activity was observed after heating at 70°C for 10 min. These properties and the specificity for L-fucose residues with ␣(1-6) linkages in N-glycan should allow RSL to be a valuable tool for glycobio-logical studies. Cloning studies are in progress to determine its complete amino acid sequence and for the production of large quantities of recombinant lectin.