The isolation by ligand affinity chromatography of a novel form of alpha-L-fucosidase from almond

THE JOURNAL OF BIOLOGICAL CHEMISTRY Q 1990 by The American Society for Biochemistry Vol. 265, No. 27, Issue of September 25, and Molecular Biology, Inc. The Isolation by Ligand Affinity cu-L-Fucosidase from Almond* pp. 16472-16477,199O Printed in U S. A. Chromatography of a Novel Form of (Received for publication, Peter Scudderj$jY, David C. A. NevilleS, Terry Raymond A. Dwek**, Thomas W. Rademacher**, D. Butters+, and Gary February 5, 1990) George W. J. Fleet]], S. Jacob*4 From the $Department of Biochemistry, G. D. Searle & Company, Oxford Research Group, University of Oxford, Oxford OXI 3QU, United Kingdom, the l/Dyson Perrins Laboratory and Oxford Centre for Molecular Sciences, University Oxford, South Parks Road, Oxford OX1 3QU, United Kingdom, and the **Glycobiology Unit, Department of Biochemistry, University of Oxford, South Parks Road, Oxford OX1 3QU, United Kingdom Almond emulsin (a commercial preparation of partially purified /3-glucosidase) is a convenient source of two linkagespecific exoglycosidases, cu-fucosidase I and cu-fucosidase II (l), which show activity toward natural oligosaccharides but do not hydrolyze synthetic substrates such as p-nitrophenyl cY-L-fucopyranoside. The cr-fucosidase I hydrolyzes Fuc(c~1 + 3)GlcNAc and Fuc(ot1 + 4)GlcNAc linkages whereas cy-fucosidase II only demonstrates activity toward Fuc(crl ---* 2)Gal. The narrow specificity of fucosidase I in particular has made it an important reagent for the sequencing and characterization of oligosaccharides that contain the Le” antigen Gal@1 -+ 4)(Fuc(al --f 3))GlcNAc (2, 3). Thus far, however, despite the use of affinity chromatographic methods (4, 5), neither enzyme has been purified to homogeneity, completely free of contaminating glycosidase activities. In this paper we report the first use of the affinity ligand N-(5-carboxy-1-pentyl)-1,5dideoxy-1,5-imino-L-fucitol to purify to homogeneity an 01fucosidase from almond. This enzyme, termed ol-fucosidase III, is shown to have similar substrate specificity to, but * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 5 Present address: Monsanto Company, 800 N. Lindbergh Blvd., St. Louis, MO 63167. Yl To whom correspondence should be sent. different physicochemical properties described earlier (1, 4, 5). MATERIALS AND from, the a-fucosidase I METHODS Oligosaccharides LNT,’ human LNNT, milk LNFPI, essentially LNFPII, as described and Z’FL were isolated from by Donald and Feeney (6). The trisaccharide GlcNAc(P1 -+ 3)Gal(Pl + 4)Glc was isolated by BioGel P-4 chromatography following digestion of LNNT with jack bean /3-galactosidase. Man(cY1 - J)Man(@l - 4)GlcNAc was isolated from the urine of a patient having mannosidosis by Bio-Gel P-4 chromatography and I?PAEC (Dionex BioLC systed, CarboPac PA-l, 9 x 250-mm column eluted isocraticallv with 150 mM NaOH, 30 mM NaOAc at a flow rate of 4 ml/min, detection by A&. A biantennary complex oligosaccharide Gal@1 - 4)GlcNAc@l + Z)Man(Lul + 4))GlcNAc@l 2)Man(oll -+ B)Man(pl + G)(Gal(Pl 4)GlcNAc(Bl + 4)GlcNAc was purified using Bio-Gel P-4 chromatography and HPAEC (conditions as described above) from human asialotransferrin (obtained from Sigma) by treatment with anhydrous hydrazine (7, 8). The heptaan&pen&saccharides GlcNAc@l + B)Man(cul + G)(GlcNAc(fll + P)Man(al + 3))Man(Bl + 4)GlcN’AQl -+ &GlcNAc and Man(al -+ i)(Mancul ‘i 3))Mar@l + 4)GlcNAc@l+ 4)GlcNAc were also isolated by HPAEC following digestion of the nonasaccharide with jack bean P-galactosidase or a mixture of jack bean P-galactosidase and &hexosaminidase, respec--f 4)GlcNAc was isolated by Biotively. The disaccharide Man@1 Gel P-4 chromatography following digestion of Man(a1 - S)Man@l - 4)GlcNAc with jack bean a-mannosidase. The structure and purity of all oligosaccharides were confirmed using 500-MHz proton NMR spectroscopy. LNFPIII, 3’FL, GalNAc(ot1 + 3)(Fuc(al+ 2))Gal(Pl + 4)Glc and Fuc(a1 + 6)GlcNAc were obtained from BioCarb, Lund, Sweden. Synthesis of N-(5-Carbomethoxy-l-pentyl)-l,5-dideoxy-1,5-imino-~fucitol One hundred fifty mg of 1,5-dideoxy-1,5-imino-l-fucitol (deoxyfuconojirimycin, DFJ) prepared from D-gh!OSe as described previously (9) was dissolved in 2.3 ml of methanol:water:methyl-6-oxohexanoate:acetic acid (16:4:2:1, v/v) and the mixture stirred overnight under an atmosphere of hydrogen in the presence of 100 mg of palladium black. TLC (ethyl acetate, methanol, 1 M aqueous ammonium hy’ The abbreviations used are: LNT, lacto-iV-tetraose (Gal@1 -+ 3)GlcNAc(/31 -+ 3)Gal(@l + 4)Glc); LNNT, lacto-N-neotetraose (Gal(p1 + 4)GlcNAc(Bl -f 3)Gal(Pl + 4)Glc); LNFPI, lacto-Nfucopentaose I (Fuc(oll + 2)Gal@l -+ 3)GlcNAc@l - 3)Gal@l + 4)Glc); LNFPII, lacto-N-fucopentaose II (Gal@1 + 3)(Fuc(otl + 4))GlcNAc(Bl - 3)Gal@l - 4)Glc; LNFPIII, lacto-N-fucopentaose Iii (GalB(l + 4)(Fuc(al3))GlcNAc(/313)Gal(Pl+ 4Glc); 2’FL, 2’-fucosvllactose (Fuc(crl + 2)GalW -f 4)Glc); 3’FL, 3’-fucosyllactose (G&31 -+ 4)Fuc(‘al+ 3)Glc; HPAEC, high performance &ion exchange chromatography; CPDFJ, carboxypentyldeoxyfuconojirimycin (N-(5-carboxy-l-pentyl)1,5-dideoxy-l,5-imino-L-fucitol);DFJ, deoxyfuconojirimycin (1,5-dideoxy-1,5-imino-L-fucitol); BSA, bovine serum albumin; SDS, sodium dodecyl sulfate. 16472 Downloaded from www.jbc.org by guest, on July 10, 2011 An cr-fucosidase has been extracted from almond meal and purified 163,000-fold to apparent homogeneity using a novel affinity ligand, N-(5-carboxy-lpentyl)-1,5-dideoxy-1,5-imino-L-fucitol, coupled to AffLGellO2. Substrate specificity studies demonstrate that the enzyme hydrolyzes the a-fucosidic linkages in Gal(j31 + 3)(Fuc(al + 4))GlcNAc(@l + 3)Gal(Bl + 4)Glc and Gal(B1 + 4)(Fuc((~l + 3))GlcNAc(Bl + S)Gal(Bl -+ 4)Glc at similar rates. but is unable to hydrolyze Fuc(cul + 2)Gal, Fuc(c~1 + G)GlcNAc, or the synthetic substrate,p-nitrophenyl a-L-fucopyranoside. Hence, the enzyme closely resembles an cw-fucosidase I isolated previously from a commercial preparation of partially purified almond &glucosidase (Ogata-Arakawa, M., Muramatsu, T., and Kobata, A. (1977) Arch. Biochem. Biophys. 181, 353-358). However, native and subunit relative molecular masses of 106,000 and 54,000, respectively, different charge and hydrophobicity properties, and the absence of stimulation by NaCl clearly distinguish this enzyme, designated (Yfucosidase III, from other almond cr-fucosidases reported previously. of Novel a-Fucosidase from Almond Meal droxide, 2:2:1, v/v) of the resulting mixture showed that the starting material (RF = 0.23) had been converted to a single product (RF = 0.72) that was purified by ion exchange chromatography (250 mg, yield 95%) and shown by 500-MHz proton NMR spectroscopy to be N-(5-carbomethoxy-l-pentyl)-1,5-dideoxy-l,5-imino-~-fucitol. A similar synthesis of this compound has been described recently by Paulsen and Matske (10). Preparation of Affinity Gel Enzymes P-N-Acetylglucosaminide ol-3/4-L-fucosyltransferase was purified 500.fold from 1 liter of human milk by a two-step batch absorption procedure involving SP-Sephadex C-50 (11) and phenyl-Sepharose CL-4B. Extensive washing of the phenyl-Sepharose with 50 mM cacodylate buffer, pH 7.2, and subsequent elution with the same buffer containing 25% glycerol and 1% Triton X-100 gave 8 milliunits of enzyme that was shown, by assaying against phenyl-fl-n-galactoside (12), to be free from @-galactoside or-2-r,-fucosyltransferase activity. Jack bean p-galactosidase, P-N-acetylhexosaminidase, and cymannosidase were purified by adaptations of methods described previously (13, 14). Generation of Asialo-oil-acid Glycoprotein Gal(~l-4(f’4C]Fuc(ul-3~,JGlcNAc Having Sequences Terminal Asialo-oi-acid glycoprotein (3.5 mg) and 37 nmol of GDP[“C] fucose (268 Ci/mol) were incubated in 0.5 ml of 50 mM cacodvlate buffer, pH 7.2, containing 25% glycerol, 1% Triton X-100, 1dmM MnCb, and 0.5 milliunits of cu-3/4-L-fucosvltransferase at 37 “C for 16 h, at which time a further 0.5 milliunits of enzyme was added and incubation continued for 6 h. The i4C-labeled glycoprotein (13.8 X lo6 cpm) was recovered by fast protein liquid chromatography using a fast desalting HRlO/lO column (Pharmacia LKB Biotechnology Inc.) with 20 mM NaCl as eluant. Enzyme Assays Human Milk cu-3/4-L-Fucosyltransferase-Enzyme was incubated for 15 min at 37 “C in a final volume of 80 ~1 containing 50 mM cacodylate buffer, pH 7.2, 10 mM MnCIZ, 160 mg of BSA, 160 nmol of LNFPI, and 0.5 nmol of GDP[‘4C]fucose. The reaction was stopped by the addition of 1 ml of distilled water, and the [“Clfucose incorporated into LNFPI was measured as described previously (15). Almond cu-3/4-Fucosidase-Enzyme activity ai each stage of purification was monitored by incubating aliquots of 5-90 ~1 at 37 “C for between 15 and 240 min in a final volume of 100 ~1 containing 50 mM NaOAc buffer, pH 5.0, and 10.000 cum of 1°C-Fuclasialo-al-acid glycoprotein (17 pmol of terminal [iiC]fucose). The’ reaction was stopped by the addition of 200 ~1 of 10% trichloroacetic acid, 5% phosphotungstic acid, and 100 ~1 of 25 mg of BSA/ml of distilled water then added as a carrier before centrifuging at 15,000 X g for 10 min. The supernatant was recovered, neutralized by the addition of 0.25 ml of 10% NaHCO,, and the radioactivity (shown by paper chromatography (15) to correspond to free 1”Clfucose) measured bv . scintillation counting. Under the conditions used, the rate of release of [‘Clfucose was linear up to 10% hydrolysis of the glycoprotein substrate. The purified enzyme was assayed at 37 “C against 1.6 mM LNFPII in 50 ~1 of the above buffer. An aliquot equivalent to 10 nmol of oligosaccharide was desalted using Dowex AG 5OW-X12 (H’ form, 100-200 mesh) and AG 3-X4 (OHform, 100-200 mesh), subjected to HPAEC, and the reaction products monitored using pulsed amperometric detection (conditions as described above). The degree of hydrolysis was calculated from the response factors for LNFPII and fucose, which were determined to be 1.72:l.O. One unit of activity is defined as that amount of enzyme required to release 1 pmol of [“Cl fucose per min under the assay conditions described above. Other Glycosidases-The purified a-fucosidase (9.5 milliunits/ml) was screened for contaminating glycosidase activities by incubation for 24 h at pH 5.0 with an appropriate p-nitrophenyl glycoside (3 mM) or oligosaccharide substrate (1 mM) as follows: @-mannosidase, Man@1 ---f 4)GlcNAc; @-N-acetylglucosaminidase, GlcNAc@l + 3)Gal(Bl-+ 4)Glc, and GlcNAc(@l + 2)Man(oll+ G)(GlcNAc(@l + Z)Man(a3))Man@l + 4)GlcNAc(fll - 4)GlcNAc; @-galactosidase, Gal@1 - 4)GlcNAc@l+ P)Man(al --t G(Gal(P1 + 4)GlcNAc@l+ 2)Man(cul + 3))Man@l + 4)GlcNAc@l + 4)GlcNAc, LNT and LNNT; a-mannosidase, Mama1 -t 3)Man@l -+ 4)GlcNAc and Man(n1 + G)(Man(oll + B))Man@l --f 4)GlcNAc(/31 + 4)GlcNAc; and a-fucosidase II, 2’FL. For natural oligosaccharide substrates, hydrolysis was monitored by HPAEC using an eluant of 150 mM NaOH, 30 mM NaOAc and pulsed amperometric detection (conditions as described above). The hydrolysis of p-nitrophenyl glycosides was determined as described previously (14). Determination of Kinetic Con&on&-Enzyme, 0.6 milliunit/ml, was incubated in 50 mM NaOAc buffer, pH 5.0, containing 1 mg of BSA per ml and either LNFPII or LNFPIII at concentrations that ranged from 0.2 to 1.2 mM. After 30 min, the reaction was stopped by the addition of an equal volume of 1 M citric acid, and the mixture was desalted by treatment with Dowex AG 5OW-X12 and AG 3-X4. An aliquot equivalent to 5 nmol of oligosaccharide was subjected to HPAEC and the reaction products monitored using pulsed amperometric detection (conditions as described above). The degree of hydrolysis was calculated from the response factors for the substrates and enzyme products which were determined to be as follows: LNFPII:Fuc, 1.72:l.O and LNFPIII:Fuc, 1.6:l.O. Values for Km2 and V were calculated from Hanes-Woolf plots. Purification of Almond Meal a-Fucosidase III Almond meal (Sigma) 500 g, was extracted by stirring for 80 min at room temperature in 3 liters of 0.1 M NaOAc buffer, pH 5.0, containing 0.1 M NaCl, 0.3 g of phenylmethylsulfonyl fluoride, 60 mg of leupeptin, and 30 mg of pepstatin A. The extract was filtered through muslin and the filtrate centrifuged at 13,000 X g for 45 min at 4 “C. Unless otherwise stated all further purification steps were carried out at 4 “C. Ammonium Sulfate Precipitation-The extract was brought to 80% saturation with (NH&SOs and stirred for 30 min. The precipitate was recovered by centrifugation, dissolved in 200 ml of 50 mM NaOAc buffer, pH 5.0 (buffer A), and dialyzed for 16 h against the same buffer. S-Sepharose Chromatography-The dialyzed material was chromatographed as two equal portions on a Fast-Flow S-Sepharose column (5.0 X 30 cm) equilibrated with buffer A. After applying the sample, the column was eluted for 3 h with buffer A, 12 h with a linear gradient of 0.0-0.5 M NaCl in buffer A, and 3 h with buffer A containing 0.5 M NaCl (flow rate, 240 ml/h; fraction size, 50 ml). Fractions containing cu-3-fucosidase which eluted first and did not bind to the column were pooled (see Fig. 1, peak A) and concentrated to 90 ml using an Amicon ultrafiltration cell with a YM-30 membrane. Phenyl-Sepharose CL-4B Chromatography-The concentrated enzyme solution was adjusted to 0.5 M (NH&SO, and applied at a flow rate of 75 ml/h to a column (2.6 X 31 cm) of phenyl-Sepharose CL4B equilibrated with 50 mM NaOAc, pH 5.0, containing 0.5 M (NH,),SO, (buffer B). The column was washed for 1 h with buffer B followed by a 4-h linear gradient to 100% buffer A. After eluting the column isocratically for 3 h with buffer A, the a-fucosidase was batch eluted using distilled water as eluant. Enzyme-active fractions were pooled, adjusted to 50 mM NaOAc, pH 5.0, and concentrated to 10 ml. Fast Protein Liquid Chromatography-Chromatofocusing-The enzyme was dialyzed against 25 mM methylpiperazine/HCl buffer, pH 5.7, and four 2.5-ml aliquots were chromatographed separatelyat room temperature on a Mono P HR5/20 column (Pharmacia) bv eluting with Polybuffer 74, pH 4.0, at a flow rate of 1 ml/mm. Fractions (1 ml) containing ru-fucosidase were pooled and concentrated to 1.4 ml. Downloaded from www.jbc.org by guest, on July 10, 2011 N-(5-Carbomethoxy-l-pentyl)-1,5-dideoxy-1,5-imino-L-fucitol (160 rmol) was converted to the free carboxylic acid form N-(5carboxy-l-pentyl)-1,5-dideoxy-1,5-imino-L-fucitol (CPDFJ) by incubating in 2.2 ml of 0.1 M NaOH for 20 min at room temperature. After adjusting the pH to 4.8, the solution was added to 2.2 ml of Affi-Gel-102 plus 2.2 ml of 20 mM l-ethyl-3-(3-dimethylaminopro60 mM lactose (internal standard). _uH 4.8. The DH _pvl)carbodiimide. _ of the reaction mixture was maintained at 4.8 for 1 h by the addition of NaOH and coupling allowed to continue for 16 h. The gel was then washed with 100 ml of 0.1 M Tris/HCl buffer, pH 8.0, containing 0.5 M NaCl, followed by 0.1 M NaOAc buffer, pH 4.0, containing 0.5 M NaCl, and finally, 0.02% aqueous NaNs. From the decrease in the ratio of CPDFJ to lactose in the supernatant (determined by HPAEC, using a CarboPac PA-1 4 x 250-mm column eluted at 1 ml/min with 150 mM NaOH, 100 mM NaOAc and using triple-pulsed amperometric detection with the following pulse potentials and durations: E, = 0.01 V (ti = 120 ms), EP = 0.6V (t2 = 120 ms), and E3 = -0.93 V (tS = 180 ms)). it was calculated that 36% of the CPDFJ was coualed. giving a ligand concentration of 20 rmol/ml Affi-Gel. 16473 Novel a-Fucosidase from Almond Meal 16474 Fast Protein Liquid Chromatography-Gel Permeation Chromatography-The enzyme was chromatographed at room temperature at a flow rate of 0.3 ml/min as four separate 0.35-ml aliquots on a Superose 12 HR10/30 column (Pharmacia) equilibrated with 50 XnM NaOAc buffer, pH 5.0, containing 0.1 M NaCl (buffer C). Fractions (0.45 ml) containing a-fucosidase were pooled and concentrated to 1.8 ml. Affinity Chromatography-The enzyme was applied to a column (0.5 x 3.8 cm) of Affi-Gel 102 coupled with CPDFJ and equilibrated with buffer C and the column washed with the same buffer until the absorbance of the eluate (measured at 254 nm) had returned to base line. The cr-fucosidase was then eluted with 20 ml of buffer C containing 10 mM CPDFJ. The absorbance due to the eluted protein could not be distinguished from that due to ligand, therefore the enzyme was recovered by pooling the fractions that contained CPDFJ (see Fig. 4). The final preparation of ol-fucosidase was concentrated by ultrafiltration and dialyzed against buffer A. Analytical Procedures RESULTS Isolation of Almond a-Fucosidase-Table I summarizes the purification of a-fucosidase from almond meal. The first chromatography step, using S-Sepharose, gave two peaks of a-fucosidase activity (Fig. 1); a major peak (A) that did not bind to the column, accounting for 65% of the total activity, and a minor peak that corresponds to the a-fucosidase I isolated previously by Imber et al. (5), which was eluted with a linear gradient of 0.0-0.5 M NaCl. Further purification was restricted to peak A, which also contained a-fucosidase II activity (detected by assaying against 2’FL). Although the majority of the fucosidase II activity was subsequently removed by hydrophobic interaction chromatography on phenyl-Sepharose (see Fig. 2), two major contaminating glycosidase activities, @-N-acetylglucosaminidase and /3-galactoTABLE Purification procedures and assay step Purification conditions Total I a-fucosidase are as described under activity” pm01 [‘4Clfucose relea.sed/min Buffer extract 80% saturation (NH&SO, precipitate S-Sepharose chromatography Phenyl-Sepharose chromatography Mono P chromatography Superose 12 chromatography Affinity chromatography 283 293 122 82 40 23 17 of almond Total protein III “Materials Specific and Methods.” activity Recovery Purification mg pm01 /‘~C]jucose relea.sed/min/mg % -fold 32,480 11,250 1,359 84 23 0.9 0.012 0.0087 0.026 0.089 0.98 1.74 25.5 1,417 100 100 43 30 14 8 6 1 3 10.7 113 203 2,966 163,000 a It should be emphasized that during purification, assay of the oc-fucosidase was performed at an estimated glycoprotein substrate concentration of 0.17 pM with respect to [14C]Fuc( cy 1 + 3)GlcNAc. This value (by analogy with data obtained using the pentasaccharide LNFPII) is likely to be significantly below the substrate’s K,,,. When assayed against a saturating concentration (1.6 mM) of LNFPII the total amount of enzyme recovered was determined to be 20.9 milliunits (see “Results”). Downloaded from www.jbc.org by guest, on July 10, 2011 Determination of Protein-Protein concentrations were determined using the BCA or micro-BCA assay system (Pierce Chemical Co.) with bovine serum albumin as a standard. Native Molecular Weight Estimation-Gel filtration of the oc-fucosidase was performed using a Superose 12 column that had been calibrated with dextran T2000 and glucose (to determine VO and V,, respectively), and the following molecular weight markers (obtained from Sigma): myoglobin (iUr = 17,700), ovalbumin (Mr = 45,000), conalbumin (Mr = 77,000), and lactate dehydrogenase (A4, = 140,000). The calibration curve is shown as an inset in Fig. 3. Polyacrylamide Gel Electrophoresis-Purified ol-fucosidase (0.125 pg of protein) in 2% SDS, 5% mercaptoethanol was denatured by heating at 100 “C for 5 min and electrophoresed in a 7.5% polyacrylamide slab gel (16) together with the following molecular weight markers (obtained from Sigma) (subunit M, shown in parentheses): myosin (205,000), P-galactosidase (116,000), phosphorylase a (94,000), BSA (66,000), egg albumin (43,000), and carbonic anhydrase (29,000). Protein bands were detected by silver staining (17). sidase, were still present. The P-galactosidase activity was removed by preparative chromatofocusing between pH 5.0 and 4.0 (p1 of @-galactosidase and a-fucosidase measured as 4.9 and 4.4, respectively) but, despite a large difference in the M, values of @-N-acetylglucosaminidase and oc-fucosidase (214,000 and 106,000 respectively), fast protein liquid chromatography-gel filtration using a Superose 12 column failed to remove completely the contaminating glycosidase activity (see Fig. 3). Finally, chromatography on a CPDFJ affinity column (Fig. 4) enabled the complete separation of the afucosidase, which remained bound to the column when it was developed with buffer C from the P-N-acetylglucosaminidase activity that eluted under these conditions. The a-fucosidase was subsequently eluted from the column using CPDFJ to give a total of 20.9 milliunits of activity as assayed against LNFPII. Purity-The final preparation of almond meal a-3-fucosidase, purified 163,000-fold by the above procedure, gave a single band when analyzed by SDS-polyacrylamide gel electrophoresis under reducing conditions (Fig. 5). The subunit M, was determined to be 54,000, suggesting that the native enzyme (Mr = 106,000) is a dimer composed of essentially identical subunits. When assayed against appropriatep-nitrophenyl glycosides and oligosaccharide substrates of 2’FL, LNT, LNNT, Man(a1 + B)Man@l+ 4)GlcNAc, Man(otl+ 3)Man(al+ B)Man(@ + 4)GlcNAc, Man@1 ---, 4)GlcNAc, GlcNAc(@l -+ 3)Gal(P4)Glc, Gal@1 + 4)GlcNAc(@l + S)Man(al ---f G)(Gal(Pl + 4)GlcNAc(@l + 2)Man(oll + 3))Man@l + 4)GlcNAc(@l + 4)GlcNAc, GlcNAc(P1 --, 2)Man(al + G)(GlcNAc@l + 2)Man(otl+ 3))Man(@l+ 4)GlcNAc(Pl+ 4)GlcNAc, and Man(cY1 -+ G)(Man(cul + 3))Man(pl -+ 4)GlcNAc@l + 4)GlcNAc, the purified a-fucosidase was found to be free of detectable a-fucosidase II, ,&galactosidase, o(- and /3-mannosidase and P-N-acetylglucosaminidase activities. In addition, no N-glycanase or endoglycosidase activities could be demonstrated (as monitored by Bio-Gel P-4 chromatography of the reaction products) following incubation of [‘4C-Fuc]asialo-ocl-acid glycoprotein with a-fucosidase for 24 h. Stability-The purified enzyme (protein concentration of 2.6 pg/ml 50 mM NaOAc buffer, pH 5.0, containing 0.02% NaN3) was relatively stable at 4 “C and exhibited a tlh of 5 months. In contrast, it was rapidly inactivated at 37 “C ( tlh of 10 min) but could be stabilized effectively by the addition of 3 mg/ml BSA, resulting in a preparation that showed no loss of activity after incubation for 24 h at 37 “C or 3 months at Novel a-Fucosidase from Almond Meal 16475 r vo 20 40 60 80 0 1. S-Sepharose cation exchange chromatography of almond a-fucosidase. The column eluate was monitored by absorbance at 254 nm (-), and alternate fractions were assayed (90-rl aliquots) for enzyme activity (U). Subsequent purification was restricted to peak A. Full details appear under “Materials and Meth- IO ods.” 50 30 40 50 Fraction number FIG. Or-----lb0 Fraction FIG. 2. Phenyl-Sepharose chromatography of almond 260 number 360 CL-4B hydrophobic cy-fucosidase. The interaction was monitored by absorbance at 254 nm (--), and fractions were assayed (50-~1 aliquots) for enzyme activity (U). An arrow indicates when elution with water was started. Full details eluate appear under “Ma- 0 20 terials and Methods.” 4 “C. In 50 mM NaOAc, pH 5.0, the activity of the cr-fucosidase was unaffected by rapid freezing in liquid nitrogen and, in the presence of 3 mg/ml BSA, was completely stable to lyophilization. The Effect of pH and Ammonium Sulfate on a-Fucosidase Activity-The activity of a-fucosidase toward [i4C-Fuc]asialool,-acid glycoprotein was studied from pH 3.0 to 7.0 in 0.1 M citric acid/Na,HPO, buffer. No activity was detected at or below pH 3.0; and at pH 7.0 only 25% of the maximal activity detected at pH 5.3 was observed. In marked contrast to the almond a-fucosidase I described previously (5), the activity of the enzyme was not stimulated by the presence of NaCl. There was, however, a strong concentration-dependent stimulation of enzyme activity at (NH&SO4 levels of 0.25 M and above, with a 500% increase in enzyme activity seen at an (NH&SO4 concentration of 2 M (Fig. 6). This effect may be restricted to glycoprotein substrates since it was not seen when the cy-fucosidase was assayed against an oligosaccharide substrate, LNFPII; in fact, enzyme assayed in the presence of 2 M (NH&SO4 showed a 20% reduction of activity. Substrate Specificity of a-Fucosidase-An HPAEC approach was used to separate and monitor simultaneously the products obtained following treatment of a variety of (Yfucosides with the purified enzyme. As can be seen from Table II, the cu-fucosidase readily hydrolyzed the Fuc(a1 -+ 3)Glc and Fuc(~yl + 3)GlcNAc linkages in 3’FL and LNFPIII as 60 40 Fraction number 4. Affinity chromatography of cY-fucosidase on Affi102 coupled with carboxypentyldeoxyfuconojirimycin. column eluate was monitored by absorbance at 254 nm (-). enzyme was pulsed off the column with buffer C containing 10 InM CPDFJ at the position indicated by an arrow. The resulting increase in absorbance of the eluate is almost completely due to the FIG. Gel The The presence of ligand and not indicative column. Full details appear under of protein “Materials eluted from the and Methods.” well as the Fuc(cy1 + 4)GlcNAc linkage in LNFPII. However, several al + 2-linked fucosides and Fuc(a1 + 6)GlcNAc were completely resistant to hydrolysis by the enzyme. The apparent Michaelis constants and maximum velocities for the reaction using LNFPII and LNFPIII as substrates were 0.23 mM and 4.5 pmol min-’ mg of protein-’ and 0.1 mM and 3.5 pm01 min-l mg of protein-‘, respectively. DISCUSSION The present paper describes (a) the use of an N-alkylated imino sugar analogue of a-fucose, carboxypentyldeoxyfuconojirimycin (for structure see Fig. 7) to purify to apparent homogeneity an a-fucosidase from almond meal; and (b) characterization studies performed on this enzyme which demonstrate that it is a form of a-fucosidase I different from that described previously (1, 4, 5). Two affinity chromatographic methods for the isolation of almond oc-fucosidase I have been reported previously. Yosh- Downloaded from www.jbc.org by guest, on July 10, 2011 FIG. 3. Superose 12 gel permeation chromatography of almond a-fucosidase. The column eluate was monitored by absorbance at 254 nm (-), and fractions were assayed (lo-~1 aliquots) for enzyme activity (U). The elution position of @N-acetylglucosaminidase (assayed using p-nitrophenyl P-N-acetylglucosaminide as substrate) is indicated by hexase. The inset contains the calibration curve used to determine A4, values of a-fucosidase and P-N-acetylglucosaminidase; the molecular weight markers are indicated by the numbers 1-4 (myoglobin, ovalbumin, conalbumin, and lactate dehydrogenase, respectively). Full details appear under “Materials and Methods.” 16476 Novel Lu-Fucosidase from Almond Meal TABLE Hydrolysis w, 94000----t a-fucosidase III Enzyme, 0.2 milliunit, was incubated at 37 “C in 50 ~1 of 50 mM NaOAc buffer, pH 5.3, containing 0.15 mg of BSA and 50 nmol of oligosaccharide. Aliquots (10 ~1) were removed at the times indicated and the reaction products separated by HPAEC and monitored using pulsed amperometry. The amount of enzyme-released fucose was calculated from experimentally determined response factors as described under “Materials and Methods.” 205000----t 116000~ II by almond of oligosacchurides Hydrolysis Substrate 4h Gal@1 + 3)GlcNAc(@l + 3)Gal(fll + 4)Glc 12h 81 100 I 29OOON y9 P I Fuc(cul.4) Gal@1 + 4)GlcNAc@i + 3)Gal@l + 4)Glc I Fuc(al,B) Gal@1 --, 4)Glc 4 83 100 I FIG. Fuc(c~1,3) Fuc(c~1 + Z)Gal@l + 4)Glc ND” ND Fuc(~ul + 2)Gal@l+ 3)GlcNAc@l+ 3) ND ND Gal@1 + 4)Glc GalNAc(Lu1 + 3)Gal(@l + 4)Glc ND ND I Fuc(cul,Z) Fuc(cu1 + 6)GlcNAc ND ND ’ ND, no hydrolysis detected. The limit of detection for fucose is approximately 5 pmol, which is equivalent to 0.05% hydrolysis. HO o--OH CH, OH k 0.0 0.5 1.0 Ammonium Sulphate 1.5 (M) 2.0 FIG. 6. The effect of ammonium sulfate on the activity of almond cY-fucosidase. Fucosidase, 0.36 milliunit, was incubated at 37 “C in 0.1 ml of 50 mM NaOAc buffer, pH 5.3, containing 10,000 cpm [“C-Fuclasialo-al-acid glycoprotein, 0.3 mg of BSA, and varying concentrations (O-2.0 M) of ammonium sulfate. After 15 min, the reaction was stopped by the addition of trichloroacetic acid/phosphotungstic acid and the [W]fucose released from the glycoprotein determined as described under “Materials and Methods.” Enzyme activity is normalized relative to “control” activity measured in the absence of ammonium sulfate. ima et al. (4), using the l-amino derivative of LNFPII as a ligand, were able to achieve about a X0-fold purification of the enzyme whereas Imber et al. (5), using Cibacron blue FG3A achieved a purification factor of 1250. Unfortunately, both preparations showed considerable heterogeneity and still contained trace amounts of contaminating glycosidase activities. Prompted by the successful use of carboxypentyldeoxynojirimycin (an N-alkylated imino sugar analogue of glucose) as an affinity ligand for the purification of a-glucosidase I (18, 19) and also the knowledge that DFJ is an inhibitor of almond cr3/4-fucosidase: we explored the utility of CPDFJ for the isolation of a-fucosidases from almond and C. lampas (20). The efficiency of CPDFJ as a specific ligand for this class of enzyme is exceptionally good, as indicated by the degree of purification (163,000-fold) of the fucosidase obtained and by the purification to apparent homogeneity of an a’fucosidase derived from C. lumpas (20). 2 P. &udder, unpublished observations. FIG. 0 OH 7. Carboxypentyldeoxyfuconojirimycin. To be of value as a reagent for the sequencing of oligosaccharides, it is of particular importance that the enzyme isolated using the above protocol be free from additional exoglycosidase activities. To assay for these contaminants, we monitored the hydrolysis of appropriate oligosaccharide substrates using pulsed amperometric detection following separation of the reaction products by HPAEC. The sensitivity of this detection system is such that using only 30 nmol of an appropriate substrate, we can readily detect a contaminant activity at a level of 0.001% relative to that of the a-fucosidase. However, none of the glycosidases assayed for was detected, establishing that the cY-fucosidaseis operationally pure, ie. in a form suitable for the structural analysis of glycans. The purified enzyme exhibits a linkage specificity similar to the almond cr-fucosidase I isolated previously (1,5) in that it readily hydrolyzes Fuc(a1 + 4)GlcNAc and Fuc(cr1 + 3) GlcNAc linkages but demonstrates no activity toward Fuc(a1 + 2)Gal or Fuc(cu1 + 6)GlcNAc. In addition, K,,, values for LNFPII and LNFPIII are similar to those determined in an earlier study (1) for the corresponding alditols of these oligosaccharides. However, there are significant differences between the present enzyme, designated fucosidase III, and the cu-fucosidaseI described earlier (5) which can be summarized as follows. The enzyme described here displays a significantly different native molecular weight from that of the a-fucosidase I reported previously (106,000 compared with 73,000); its activity is not dependent on, nor is it stimulated by, the presence of NaCl; and it has different physicochemical prop- Downloaded from www.jbc.org by guest, on July 10, 2011 5. Reductive SDS-polyacrylamide gel (7.5%) electrophoresis of purified almond a-fucosidase. Electrophoresis of the affinity-purified cz-fucosidase was performed as described under “Materials and Methods.” Protein was detected using silver stain. The arrows indicate the positions of molecular weight markers. 60 100 Novel a-Fucosidase from Almond Meal Acknowledgments-We are grateful to Dr. Wrenn Wooten for performing NMR spectroscopicanalyseson purified oligosaccharides, to Brian Mathews for performing View publication stats hydrazinolysis of transferrin sam- ples, and to Jean Rotsaert manuscript. 1. 2. 3. 4. 5. 6. I. 8. 9. for help with the preparation of the REFERENCES Ogata-Arakawa, M., Maramatsu, T. & Kohata, A. (1977) Arch. Biochem.Biophys.181,353-358 DiCiocci, R. A., Piskorz, C., Salamadi, G., Barlow, J. J. & Matta, K. J. (1981) Anal. Biochem. 111, 176-183 Yamashita, K., Koide, N., Endo, T., Iwaki, Y. & Kobata, A. (1989) J. Biol. Ckem. 264,2415-2423 Yoshima, H., Takasaki, S., Ito-Mega, S. & Kobata, A. (1979) Arch. Biochem. Biophys. 194, 394-398 Imber, M. J., Glasgow, L. R. & Pizzo, S. V. (1982) J. Biol. Chem. 257,8205-8210 Donald, A. S. R. & Fenney, J. (1988) Carbohydr. Res. 178, 7991 Ashford, D., Dwek, R. A., Welply, J. K., Amatayakul, S., Homans, S. W., Lis, H., Taylor, G. N., Sharon, N. & Rademacher, T. W. (1987) Eur. J. Biochem. 166.311-320 Hdmans, S. W., Dwek, R. A., Fernandes, D. L. & Rademacher, T. W. (1984) Proc. Natl. Acad. Sci. U. S. A. 81, 6286-6289 Fleet, G. W. J., Petursson, S., Campbell, A. L., Mueller, R. A., Behline. J. R.. Babiak. K. A.. Ne. J. S. & Scares. M. G. (1989) J. Ckem. Sot. ‘Perkin &ans. i, g65-667 10. Paulsen, H. & Matske, M. (1988) Ann&n 12, 1121-1126 11. Prieels, J.-P., Monnom, D., Dolmans, M., Beyer, T. A. & Hill, R. L. (1981) J. Biol. Ckem. 256, 10456-10463 12. Chester, M. A., Yates, A. D. & Watkins, W. M. (1976) Eur. J. Biochem. 69, 583-592 13. Li, S.-C., Mazzotta, M. Y., Chien, S.-F. & Li, Y.-T. (1975) J. Biol. Chem. 250,6786-6791 14. Li, Y.-T. & Li, S.-C. (1972) Methods Enzymol. 28, 702-713 15. Scudder. P. R. & Chantler. E. (1981) Biochim. Bionhvs. Acta 660. 128-135 16. Laemmli. U. K. (1970) Nature 227.680-685 17. Morrissey, J. H. (1981) Anal. Biochem. 117,307-310 18. Hettkamn. H.. Leeler. G. & Bause. E. (1984) Eur. J. Biochem. 142,8&90' 19. Shailubhai. K.. Pratta. M. A. & Viiav. _ “I I. K. (1987) Biochem. J. 247, 555-562 20. Butters. T. D.. Scudder. P.. Willenbrock. F. W.. Rotsaert. J. M. V., Rademacher, T., ‘Dwek, R. A. & Jacob,‘G. S. (1989) in I Proceedings of the 10th International jug&es (Sharon, N., Lis, H., Duksin, . Symposium I on Glycocon- D. & Kahane, I., eds) pp. 312-313, Sept. 10-15, Jerusalem, Israel Downloaded from www.jbc.org by guest, on July 10, 2011 e&es, as demonstrated by its inability to bind to S-Sepharose at pH 5.0 and its strong affinity for phenyl-Sepharose CL-4B (the a-fucosidase I is considerably less hydrophobic and is only weakly bound (results not shown)). The stimulation of cu-fucosidase activity by (NH&S04 when measured against [W-Fuclasialo-al-acid glycoprotein but not the free sugar LNFPII illustrates a key difference between the use of exo- or endoglycosidases to digest sugars on glycoproteins rather than in their free state, namely the accessibility of the sugar substrate to the enzyme. In the present example, (NH&SO4 appears to exert its effect directly on the glycoprotein substrate, asialo-al-acid glycoprotein. This conclusion is based on the fact that (NH&SO4 concentrations up to 2 M had no measurable effect on the catalytic activity of the cu-fucosidase (as monitored by activity toward the oligosaccharide substrate LNFPII) and thus was not exerting an effect via an alteration of the enzyme’s conformation. Stimulation of activity against asialo-al-acid glycoprotein is likely to be the result of an increased accessibility of the N-linked sugars to the cu-fucosidase. One possible explanation for this is that the salt is “freeing up” proteincarbohydrate interactions of the substrate which interfere with the enzyme’s ability to bind and hydrolyze the fucosylated sugars attached to the protein backbone. Of course, the well known “salting-out” effect of (NH&SO4 might also be playing a role by altering the nature of the solvent interactions with protein and carbohydrate. It will be of interest to see if this phenomenon observed with (NH&SO, can be extended to other glycosidases and whether a similar effect is seen with other glycoproteins. Further substrate specificity studies are currently being performed to optimize the value of cr-fucosidase III as a reagent for oligosaccharide sequencing.