Glycosidases of Aspergillus niger

A highly purified preparation of 1,2-oc-L-fucosidase, free of /3-galactosidase and P-N-acetylglucosaminidase activities, has been obtained from a commercial preparation of Aspergillus niger by a simple isolation procedure involving ammonium sulfate precipitation, pressure dialysis, repeated gel Cltration on Sephadex G-150, and chromatography on diethylaminoethyl Sephadex A-50. The enzyme has a pH optimum of 3.8 f 0.2 and K, and V,, values of 8.3 X low5 M and 16.0 Mmoles per mg per hour, respectively, at 37” for methyl Z-0-cu-L-fucopyranosyl-fi-D-galactoside as substrate. The detailed specificity studies indicate that it is highly specific for nonreducing terminal L-fucose residues linked to D-galactose residues by 1 + 2-a linkage. It hydrolyzes fucose residues quantitatively from Z-0-a-L-fucopyranosylD-galactose, 2-O-ac-L-fucosyllactose, and Iacto-N-fucopentaose I. It does not split p-nitrophenyl-cr-L-fucoside, Z-O-, 3-o-, and 4-0-cr-L-fucopyranosylfucoses, and 3-Oand 4-O-a-L-fucopyranosyl-o-galactoses. It also does not release any focuse from oligosaccharides of human milk such as lacto-N-fucopentaose II, lacto-N-fucopentaose HI, and 3’-0ar-L-fucosyllactose. The enzyme liberates 80 to 90% of fucose residues from porcine and canine submaxillary mucins. It has no action on orosomucoid, human chorionic gonadotropin, and fucan sulfate. The enzyme is extremely active against human blood group substance H, destroying virtually all of the detectable activity.

It also does not release any focuse from oligosaccharides of human milk such as lacto-N-fucopentaose II, lacto-N-fucopentaose HI, and 3'-0ar-L-fucosyllactose.
The enzyme liberates 80 to 90% of fucose residues from porcine and canine submaxillary mucins. It has no action on orosomucoid, human chorionic gonadotropin, and fucan sulfate.
The enzyme is extremely active against human blood group substance H, destroying virtually all of the detectable activity.
ZL-Fucose is frequently located on the nonreducing termini of the carbohydrate moiety of several biologically active molecules. These include serum glycoproteins, immunoglobulins, gastric and submaxillary mucins, gonadotropic hormones, and blood group substances (1). The serological activity of certain blood group substances is associated with L-fucose residues (2). It is also present in oligosaccharides of human milk and of bovine colostrum (3), marine algae polysaccharides such as fucan sulfate, and plant gums (4).
L-Fucosidases from various sources have been found to release * Supported by United States Public Health Research Grant AM-10273.
The preceding paper in this series is Reference 14.
fucose* from these molecules by splitting terminal oc-L-fucosidic bonds. Levvy and M&Ian (5) have studied the distribution of ar-L-fucosidase in rat and mouse tissues. The epididymis and kidney showed the highest activity of the enzyme toward pnitrophenyl-ar-L-fucopyranoside.
A cell-free extract of T&homonas fetus was found to release fucose from blood group substances but failed to hydrolyze p-nitrophenyl and methyl ar-n-fucosides (6). Obviously, the mammalian and bacterial enzymes possess different substrate specificity.
No extensive purification and characterization of the enzymes from any of the above sources has been carried out. Recently, Aminoff (7) has reported the purification of a-n-fucosidase from Clostridium perfringens which exhibited substrate specificity similar to that of T. fetus enzyme. More recently, Tanaka. et al. (8) have partially purified two oc-n-fucosidases from abalone livers which hydrolyzed p-nitrophenyl cu-L-fucoside, but only one of them liberated fucose from porcine submaxillary mucin.
We have previously reported (14) the purification and characterization of P-acetylglucosaminidase and CX-and &galactosidases from a commercial enzyme product of Aspergillus niger. This communication describes the procedure for the purification of a highly specific enzyme, 1,2-a-L-fucosidase, and its kinetic and detailed substrate specificity properties, including its action on p-nitrophenyl o(-and /3-L-fucopyranosides, various fucose-containing oligosaccharides, porcine and canine submaxillary mucins, orosomucoid, human chorionic gonadotropin, and finally on fucan sulfate.

MATERIALS AND METHODS
A commercial enzyme product from A. niger, Rhozyme HP-150 (without diluent), was obtained from Rohm and Haas Company, Philadelphia, Pennsylvania. p-Nitrophenyl OL-and p-n-fucopyranosides, other p-nitrophenyl glycosides, and fucan sulfate were purchased from Pierce Chemical Co., Rockford, Illinois. The disaccharides, 2-O-, 3-O-, and 4-O-a-n-fucopyranosyl-nfucopyranoses were prepared essentially by CBte's method (9), except that the final purification of the disaccharides was carried out on a charcoal column using stepwise elution with ethanolwater mixtures ranging in concentration from 0 to 10% alcohol. Porcine submaxillary mucin (lo), orosomucoid (ll), and human chorionic gonadotropin (12) were prepared according to the published procedures.
The details of the syntheses will be described elsewhere.2 Gas Chromatographic Determination of Fucose L-Fucose liberated during the enzymatic hydrolysis of various substrates was quantitatively determined by gas-liquid chromatography.
The enzymatic digest was deionized and freed of the unreacted substrate and protein by passing through a column of charcoal (1 x 0.25 cm) and a mixed bed resin, MB-3 (1 x 0.5 cm). The column was eluted with 8 to 10 ml of 10% ethanol (14). The eluate was dried in a Virtis Biodryer (The Virtis Company, Inc., Gardiner, New York), and fucose in the residue was determined as its trimethylsilyl ether derivative by gasliquid chromatcgraphy (15). A 25+1 portion of the silylating agent (prepared by mixing 1 ml of pyridine, 0.2 ml of hexamethyldisilazane, and 0.1 ml of trimethylchlorosilane) was added to the residue, and the reaction mixture was stirred by a Vortex mixer for 10 min. The solution was evaporated to dryness at room temperature under reduced pressure, the residue was dissolved in 25 ~1 of cyclohexane, and a l-p1 aliquot was injected into a Varian Aerograph model 1500 gas chromatograph, equipped with flame ionization detector. A 5% coating of SE-52 on hexamet,hyldisilazane-treated Chromosorb (60 to 80 mesh) was employed for column packing (+ inch X 10 feet). The column was programmed isothermally at 175". The retention times for (Y-and P-peaks of fucose were 3.2 and 3.7 min. The use of cyclohexane eliminates the problem of tailing, as encountered with pyridine, and consequently reduces drastically the time of each run. The coconut charcoal (Fisher Product 50-250 mesh), used above in the procedure for the gas chromatographic analysis of fucose, was treated as follows.
After the removal of the fine particles by screening through a loo-mesh sieve, the charcoal was boiled with 6 N HCl for 1 hour and washed with water.
It was then boiled with 1 N NaOH for 1 hour, neutralized with dilute HCl, and finally washed with water, ethanol, and acetone.
To a 100~~1 sample of a 5.5 mM solution of the substrate in 0.01 M sodium acetate buffer, pH 4.0, 10 ~1 of the enzyme solution were added. After incubation of the reaction mixture for 1 hour at 37", the fucose released was determined as described above. During the purification procedure, the assay of the enzyme in the fractions from the columns was slightly modified by using 200 ~1 of 1 mM solution in 0.25 N 2 K. L. Matta and 0. P. Bahl, manuscript in preparation. sodium acetate buffer, pH 4.0, and 10 to 25 ~1 of the enzyme solution.
One unit of the enzyme was defined as the amount which would liberate 1 pmole of fucose per hour at 37". Specific activity was expressed as units per mg of protein.

Assay of Other Glycosidases and Proteases
The enzymes /3-galactosidase and /?-acetylglucosaminidase were assayed as previously described, using appropriate p-nitrophenyl glycosides as substrates (16).
Protease activity was ascertained by using Azocoll as a substrate.3 A suspension of 6.25 mg of Azocoll in 1.2 ml of 0.1 M sodium phosphate buffer, pH 7.5, was incubated with 50 ~1 of the enzyme (75 pg) at 37" for 2 to 12 hours. h blank of the substrate was run concurrently under identical conditions without the enzyme.

PuriJication of 1 ,2-ol-L-Fucosidase
All steps of purification were carried out at 4" unless otherwise specified.
Fractions from the column were monitored for protein by measuring absorbance at 280 rnp (17). For the specific activity measurements, the protein was determined by the method of Lowry et al. (18). The enzyme solution at each stage of purification was concentrated by pressure dialysis using XM-50 membrane (Amicon Corporation, Cambridge, Massachusetts) as previously described (14). Since the commercial enzyme product contained large amounts of ,&glucosidase or cellulase, prolonged dialysis of the enzyme solutions in cellophane tubing was avoided because of the hazard of rupturing the dialysis tubing.
Step 1: Extraction and Ammonaum Sulfate Precipitation-A sample of 250 g of the commercial enzyme product (Rhozyme HP-150) was extracted with 2 liters of cold 0.1 M NaCl solution over a period of 2 hours. The extract was centrifuged for 1 hour at 16,500 rpm in a Servall refrigerated centrifuge, and the residue was further extracted twice, with 1 liter of 0.1 M NaCl solution each time.
The combined supernatant solution (4 liters, 72 g of protein) was treated slowly by stirring with 2,808 g of ammonium sulfate to bring about 100% saturation. After allowing the precipitate to settle overnight, the solution was centrifuged again at 16,500 rpm. The precipitate was collected in 1 liter of water, and the resulting solution was further subjected to another ammonium sulfate treatment (702 g). Finally, the precipitate was dissolved in 1,000 ml of 0.04 M sodium phosphate buffer, pH 6.8, and concentrated to 400 ml by pressure dialysis.
Subsequently, two additions of the buffer were made, and the solution was concentrated to a final volume of about 400 ml.
Step 2: Column Chromatography on DEAE-Sephadex-A column (10 x 110 cm) was packed with DEAE-Sephadex A-50 in 0.04 M phosphate buffer, pH 6.8. After equilibration of the column with the same buffer, 400 ml of the solution (17.0 g of protein) from the preceding step were applied to the column The column was eluted with a stepwise salt gradient from 0 to 1 M NaCl in 0.04 M sodium phosphate buffer, pH 6.8. The buffer changes are indicated in Fig. 1. Fractions of 22 ml were collected. Fractions 350 to 450, containing the enzyme, were pooled and concentrated by pressure dialysis to 40 ml. To the concentrated solution, 800 ml of 0.25 M sodium acetate buffer were added, and the solution was concentrated to 40 ml again.
Step 3: Gel Filtration on Saphadex G-150-The above solution (3.4 g of protein) was applied to a column of Sephadex G-150 (5 X 110 cm) packed in 0.25 M sodium acetate buffer, pH 4.6, as described previously (14). The column was eluted with the same buffer. Fractions of 4.8 ml were collected.
Step 4: DEAE-Sephadex A-50-A column was packed with DEAE-Sephadex in 0.04 M sodium phosphate buffer, pH 6.55. A 2-ml sample (100 mg of protein) was applied to a column (2.5 X 40 cm), and the column was eluted with a continuous salt gradient between 200 ml of 0.04 M sodium phosphate buffer, pH 7.2, and 200 ml of 1 M NaCl in the same buffer. Fractions of 4.5 ml were collected.
Fractions 67 to 75, containing the enzyme 1,2-mu-~fucosidase, were pooled and concentrated by pressure dialysis to 4 ml. The enzyme solution was diluted with 25 ml of 0.25 M sodium acetate, pH 4.6, and concentrated to 4 ml again.
Gel Filtration on Xephadex G-150-A 4.ml sample (70 mg of protein) from the preceding step was applied to a Sephadex G-150 column (2 x 150 cm). A head pressure of 40 cm was used, and the column was developed as described above.

Substrate XpeciJicity Studies
Action of 1 ,2-oc+Fucosidase on Oligosaccharides-A lOO+l solution of the oligosaccharide, containing 50 to 100 pg of the material and 10 to 20 ~1 of the enzyme solution (25 to 50 pg of protein), was added, and the reaction mixture was inculcated at 37" for 2 to 6 hours.
The digest was deionized and freed of the oligosaccharide and protein by passing through a bed of charcoal and mixed bed resin (MB-3) as described above. The eluate was evaporated to dryness on a rotary evaporator, and the resulting residue was estimated for fucose by gas-liquid chromatography.
Action of 1 ,2-oc-L-Fucosidase on Porcine and Ca.nine Submaxillary Mu&s, Orosomucoid, and Human Chorionic Gonadotropin-A sample of 250 to 500 ~1 of an aqueous solution containing 0.5 to 3.5 mg of the material was treated with 50 to 250 ~1 of the enzyme solution (0.1 to 0.5 mg) in 0.05 M sodium acetate buffer, pH 4.0. After the addition of 20 ~1 of toluene, the digest was incubated from & to 48 hours. Aliquots of 50 to 100 ~1 were withdrawn at regular time intervals and applied to a column (1 x 0.5 cm) of charcoal and mixed bed resin (MB-3).
The column was washed with 8 to 10 ml of 5y0 ethanol.
The washings were evaporated to dryness, and the fucose in the residue was estimated by gas-liquid chromatography.

Puri&ation
of 1 ,2-c~-L-Fucosidase-The purification scheme described above yielded a preparation which was almost free of /3-acetylglucossminidase, ,&galactosidase, and protease activities. The scheme was based on precipitation with ammonium sulfate, pressure dialysis, and repeated chromatography on DEAE-Sephadex A-50 and Sephadex G-150. A 250-g sample of the commercial enzyme product was employed in the preparation.
After three extractions of the sample with a total of 4.0 liters of 0.1 N NaCl solution, the combined extracts were subjected to precipitation twice with ammonium sulfate at 100% saturation.
During the ammonium sulfate precipitation step, not only did enrichment of the enzyme occur, but also the polysaccharide contaminants were eliminated (19). The recovery of the enzyme units after thr precipitation step was higher than that in the extract. This discrepancy was probably due to inaccuracy in the fuc3se determination by gas-liquid chromatography, which may be due to the presence of carbohydrate contaminants in the crude extract which interfered miith the fucose peaks in the gas chromatogram.
However, the possibility of the presence of an enzyme inhibitor in the crude rraterial can not be ruled out. The partially purified enzyme was applied to a DEAE-Sephadex A-50 column which was developed by a linear stepwise salt gradient.
The enzyme appeared in the first peak shown in Fig. 1, along with ,L-galactosidase and P-acetylglucosaminidase.
The fractions containing the enzyme were pooled, concentrated by pressure dialysis, and further purified by gel filtration on Sephadex G-150, which effected a partial separat'ion of 1,2-a-L-fucosidase fro=n ,&acetylglucosaminidase and fl-galactosidase (Fig. 2). The resulting enzyme fraction was subjected to chromatography on DEAELSephadex A-50, using continuous linear salt gradient (Fig. 3) instead of the stepwise gradient used above, and finally to the gel filtration on Sephadex G-150 (Fig. 4). Fractions pooled in each case are represented in Figs. 3 and 4. A summary of the purification is given in Table I.  shown in Fig. 5 indicates a pH optimum of 3.8 f 0.2 for the enzyme.
E$ect of Substrate Concentration on Enzyme Catalysis--Using the above substrate at concentrations varying from 0.6 to 5 mM in 0.01 M sodium acetate buffer, pH 4, a Lineweaver-Burk plot was obtained.
The plot of l/v versus l/s showed a straight line relationship.
The value for the apparent K, as computed from the plot was 8.3 X lo+ M, and that of Vm,,, 16.0 pmoles per mg per hour (Fig. 6).
Substrate SpeciJicity of 1 ,b-a+Fucosidase-The enzyme was found to be highly specific for 1,2-oc-L-fucosidic linkage to Dgalactose.
Finally, the enzyme did not liberate fucose from lacto-N-fucopentaose II, in which fucose is linked by 1 + 4-or-D-linkage to N-acetylglucosamine.
The enzyme A. niger 1,2-or-D-fucosidase hydrolyzed 80 to 90% of fucose residues from the intact as well as desialyzed canine and porcine submaxillary mucins (Fig. 7), indicating that D-fucose is  Recent chemical studies based on methylation and periodate oxidation of the oligosaccharides derived from porcine submaxillary mucin have shown (20,21) that fucose is indeed terminal and is linked to the D-g&CtOSe residue in the oligosaccharide chains by 1,2-a-~ type linkage.
Little is known about the structure of the carbohydrate chains of canine submaxillary mucin. However, the Glycosidases of A. niger. I1 Vol. 245, No. 2 release of fucose from this mucin by 1,2-Lu-L-fucosidase indicates clearly the position and the linkage of the fucose residues in the mucin. Since the enzyme did not cause the hydrolysis of fucose residues in human chorionic gonadotropin and orosomucoid, this may indicate that fucose residues in these glycoproteins have a different linkage than that in porcine or canine submaxillary mucin.
Structure of a glycopeptide from orosomucoid, proposed by Yosizawa and his co-workers recently (22), indicates the presence of 1, S-au-L-fucosidic linkage.
Another fraction obtained in the course of the present studies contained a fucosidase which was found to release fucose from human chorionic gonadotropin and orosomucoid; this fraction showed no activity toward porcine or canine submaxillary mucin and, therefore, appears to be a different enzyme, presumably specific for a different linkage. This fraction also did not hydrolyze p-nitrophenyl-oc-r-fucopyranoside.
The detailed characterization of the enzyme is currently under investigation in our laboratory.
Finally, the A. niger 1 ,2-ac-L-fucosidase did not hydrolyze fucan sulfate. A summary of the substrate specificity studies is given in Table II.

DISCUSSION
Recently, a great deal of interest has been aroused in glycosidases which are capable of hydrolyzing specific sugar residues from the nonreducing termini of the carbohydrate chains. Several such enzymes as P-N-acetylglucosaminidase (14,16,(23)(24)(25)(26), ,&galactosidase (19,27,28), oc-mannosidase (19,28,29), and ac-L-fucosidase (5,s) have been highly purified and characterized. These enzymes require for their action a specific nonreducing sugar residue and a specific anomeric configuration of the glycosidic bond, regardless of the nature of the aglycon.
The enzyme 1,2-oc+fucosidase, on the other hand, shows a much higher degree of specificity than the enzymes mentioned above. It requires a specific linkage of l-+2 type between the nonreducing terminal L-fucose residue and the adjacent n-galactose residue in addition to the ok anomeric configuration of the glycosidic linkage.
Consequently, a glycopeptide, a glycoprotein, or a polysaccharide which is susceptible to the action of this enzyme most likely contains a nonreducing terminal fucose residue, linked by l--+2-0( type linkage to the o-galactose residue in the carbohydrate chain. Such an inference would be strongly supported by the specificity properties of the enzyme summarized in Table II. Obviously, the enzyme offers a great potential in the study of fucose-containing macromolecules of the type described above.
It is quite interesting to note that the enzyme does not hydrolyze at all, even on prolonged incubation, 3-O-& 4-O-0(-, and 6-0-fi-n-fucopyranosyl galactoses, as well as any of the fucosyl fucoses, suggesting that 1,2-a linkage as well as both monosaccharide units are essential for enzyme activity.
Furthermore, since the conformation of the 1,2-a! linked disaccharide is quite different from that of the other fucosyl galactoses having l-t3-/?-, 1 -&(Y-, and 1+6-@-n-linkages, it is conceivable that probably a particular conformation is necessary for the binding of the substrate to the enzyme. The studies to delineate the hydroxyl groups which are involved in the binding and those which might be involved in the catalytic activity of enzyme are currently under investigation. The L-fucosidases (exo), on the basis of the available information on spec;fi.city, can be broadly classified into two groups: those which are specific for the L-fucopyranosyl group and the anomeric configuration of the fucosidic linkage, and those which require, in addition, the presence of a specific linkage of fucose to the next sugar residue in the carbohydrate chain. The former class of fucosidases hydrolyze alkyl or aryl fucosides and may also liberate fucose from the nonreducing ends of the oligo-or polysaccharide chains, regardless of the intersugar linkage. The mammalian fucosidases reported by Levvy and McAllan (5) and those reported by Tanaka et al. (8) belong to this class. The latter class of fucosidases will not hydrolyze alkyl or aryl fucosides, but will hydrolyze only specific fucosidic linkages in oligoand polysaccharides.
Bacterial fucosidases reported by Watkins (6) and that reported by Aminoff (7) show this type of specificity, although, hitherto, complete characterization of the enzymes has not been reported.
The A. niger enzyme, being highly specific for 1,2-a-~ type linkages, falls in the second category of L-fucosidases.