THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1989 hy The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 264, No. 21, Issue of July 25, pp. 12272-12277,1989 Printed in U.S.A.

Isolation and Characterization of Ceramide Glycanase from the Leech, Macrobdella decoru*

(Received for publication, January 17, 1989)

Bing Zhoul, Su-Chen Lis, Roger A. Laineg, Richard T. C. Huangq, and Yu-Teh Lis(( From the $Department of Biochemistry, Tulane University School of Medicine, New Orleans, Louisiana 70112, the §Department of Biochemistry, Louisiana State University, Baton Rouge, Louisiana 70803, and the nlnstitut fur Molekularbiologie und Biochemie, Freie Universitat, Berlin 33, Federal Republic of Germany

We have devised a simple method for achieving 890- fold purification of ceramide glycanase with 17% re- covery from a North American leech, Macrobdella dec- ora. The method includes water extraction, ammonium sulfate fractionation, and chromatography on octyl- Sepharose, Matrex gel blue A, and Bio-Gel A-0.5m columns. The final preparation showed one major pro- tein band at 54 kDa by sodium dodecyl sulfate-poly- acrylamide gel electrophoresis. By using Bio-Gel A- 0.5m filtration, the native enzyme was found to have a molecular mass of 330 kDa. With GM1 as substrate, the optimum pH of this enzyme was determined to be 5.0; the enzyme was stable between pH 4.5 and 8.5. Zn2+ at 5 mM and Cu2+, Ag+, and Hg2+ at 1 mM strongly inhibited the hydrolysis of GMl by ceramide glycanase.

The ceramide glycanase released the intact glycan chain from various glycosphingolipids in which the glycan chain is linked to the ceramide through a /3- glucosyl linkage. This enzyme also cleaved lyso-gly- cosphingolipids such as ~YSO-GM~ and lyso-LacCer and synthetic alkyl P-lactosides. Among seven alkyl &lac- tosides tested, the enzyme only hydrolyzed the ones with an alkyl chain length of four or more carbons. The enzyme also hydrolyzed 2-(octadecy1thio)ethylO- B-lactoside and 2-(2-carbomethoxyethylthio)ethyl 0- B-lactoside. p-Nitrophenyl, benzyl, and phytyl &lac- tosides, on the other hand, were not hydrolyzed. These results suggest that the enzyme can recognize the hy- drophobic portion of glycolipid substrates.

The fact that 2-(2-carbomethoxyethylthio)ethyl 0- P-N-acetyllactosaminide and DiGalCer were refrac- tory to the enzyme indicated that in the substrate the first sugar attached to the hydrophobic chain cannot be N-acetylglucosamine and galactose. Furthermore, dodecyl maltoside, Galal+6Glcj3Cer, and the LacCer in which the -CH20H of the galactose was converted into “CHO were also resistant to the enzyme, and Man@1+4GlcBCer was hydrolyzed at a much slower rate than LacCer. These results indicate that the nature and the linkage of the sugar attached to the glucose have a profound effect on the action of this enzyme.

The hydrolysis of glycosphingolipids by ceramide glycanase is stimulated by bile salts. Among various bile salts tested, sodium cholate at a concentration of 1 pg/pl was found to be most effective in stimulating the

* This investigation was supported by Grants DMB8617033 from the National Science Foundation and NS 09626 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

11 To whom correspondence should be addressed Dept. of Biochem- istry, Tulane University School of Medicine, 1430 Tulane Ave., New Orleans, LA 70112. Tel: 504-584-2459.

hydrolysis of various glycosphingolipids with the ex- ception of LacCer. For LacCer, sodium taurodeoxycho- late at a concentration of 2-3 pglpl was most effective. Tween 20, Nonidet P-40, and Triton X-100 did not stimulate the hydrolysis of GMl.

Hydrolysis of GMl in Hi80 and analysis of the re- leased glycan chain and ceramide by fast atom bom- bardment mass spectrometry showed that “0 was in- corporated into the released oligosaccharide, indicat- ing that the cleavage is between the sugar chain and the oxygen atom and characterizing the enzyme as an endoglycosidase.

Ceramide glycanase is an enzyme that releases the intact glycan chain from various glycosphingolipids by cleaving the linkage between the ceramide and glycan chain. We have reported the occurrence of ceramide glycanase in the Euro- pean leech Hirudo medicinalis (1) and the earthworm Lum- bricus terrestris (2). A similar enzyme (endoglycoceramidase) has also been induced in Rhodococcus sp. by using bovine brain gangliosides as the inducer (3). However, this novel enzyme has not been well characterized. In this paper, we report the purification and characterization of ceramide gly- canase isolated from a North American leech, Macrobdella decora.

EXPERIMENTAL PROCEDURES

Materials-Leeches (M. decora) were collected locally and stored at -60 “C until use. The following glycosphingolipids were isolated in our laboratory: GM2’ from Tay-Sachs brain (4); GbOsesCer from dog intestine ( 5 ) ; DiGalCer from the kidney of a patient with Fabry’s disease (6); LacCer, GbOse3Cer, and GbOse,Cer from human eryth- rocytes (7); L a ~ c e r - ~ I I - S o ~ from salmon milt.’ Tritium-labeled GMl was prepared according to the method described by Radin (8). Gg- Ose4Cer and GgOse3Cer were prepared from GMl and GM2, respec- tively, by formic acid hydrolysis (9); n-butyl @-lactoside, n-octyl @- lactoside, myristyl @-lactoside, stearyl @-lactoside, and phytyl @-lac- toside were prepared as described by Koenigs and Knorr (10). Briefly, lactose was converted into its acetobromo derivative as described by Conchie and Levvy (11). This acetobromo derivative was then gly-

The abbreviations used are: GgOse3Cer, gangliotriaosylceramide; GgOse4Cer, gangliotetraosylceramide; G M ~ , I13aNeuAc-GgOse3Cer; GMl, I13aNeuAc-GgOse4Cer; GM3, I13nNeuAc-LacCer; GT~I,, I13a(Neu- A~)~,IV~uNeuAc-GgOse~Cer; GDI., I13aNeuAc,IV3aNeuAc-GgOse4Cer; LacCer, lactosylceramide; DiGalCer, galactobiosylceramide; GbOse5- Cer, globopentaosylceramide; GbOse3Cer, globotriaosylceramide; GbOseqCer, globotetraosylceramide; LacCer-’I1-SO4, lactosylceram- ide 3’-sulfate; nLcOserCer, neolactotetraosylceramide; nLcOse&er, neolactotriaosylceramide; IV3(3’-S03-GlcA)-nLcOse4Cer, neolacto- tetraosylceramide 3’-sulfateglucuronic acid GlcCer, glucosylceramide. ’ S.-C. Li, R. DeGasperi, Y. Ishikawa, T. Toida, Y. Kushi, S. Handa,

I. Ishizuka, and Y.-T. Li, manuscript in preparation.

12272

Leech Ceramide Glycanase 12273

cosidically linked to an alkyl alcohol using silver oxide as catalyst and a mixture of anhydrous diethyl ether/chloroform (l:l, v/v) as the solvent. After deacetylation in methanolic sodium hydroxide, the alkyl P-lactosides were purified by silicic acid column chromatography using increasing amounts of methanol in chloroform (from 0 to 5%) as the eluting solvent. The fractions were monitored by thin-layer chromatography using chloroform/methanol/water (65:35:8, v/v/v) as the developing solvent. Fractions containing chromatographically pure alkyl 8-lactosides were pooled and evaporated to dryness. Lyso- G,, and lyso-LacCer were prepared from G M ~ and LacCer, respec- tively, according to the procedure described by Neuenhofer et al. (12). Manpl4GlcCer was generated from GlcNAc-/31+3Man@l+ 4GlcCer by using jack bean @-hexosaminidase (13). Al-LacCer (the LacCer in which the primary alcohol of the terminal galactose is converted into an aldehyde group) was prepared by treating LacCer with galactose oxidase as described by Leskawa et al. (14). The following were generous gifts: polyglycosylceramides (15), Dr. A. Gardas, Medical Center of Postgraduate Education, Warsaw, Poland; nLcOse,Cer and nLcOsesCer, Dr. S. Basu, University of Notre Dame, Notre Dame, IN; 3-OMeGal~l+3GalNAcd+3[6'-0-(2-aminoethyl phosphonyl)Galal+2](2-aminoethylphosphonyl l+6)Gal~l~ 4Glcpl+lCer (16), Dr. S. Araki, Brain Research Institute, Niigata University, Asahimachi, Niigata, Japan; IV3(3'-S03-GlcA)-nLc- Ose&er, Drs. D. K. H. Chou and F. B. Jungalwala, Shriver Center for Mental Retardation, Waltham, M A GalNAcal+3Gal(2+ 1~Fuc)pl+3GlcNAc/3l+3Gal~l4Ga1/314Glc+Cer (17), Dr. A. Slomiany, Dental Research Center, New Jersey Dental School, Uni- versity of Medicine and Dentistry of New Jersey, Newark, NJ; GlcNAc~l+3Man~l+4GlcpCer (18), Dr. T. Hori, Shiga University, Otsu, Japan; a separate sample of lyso-LacCer, Dr. T. Taketomi, Shinshu University, Matsumoto, Japan; Galal+GGlcpCer, Dr. M. Hoshi, Tokyo Institute of Technology, Meguro-ku Tokyo. The follow- ing were purchased from commercial sources: G M ~ , G M ~ , GTlb, Supelco; GD,., Biocarb Chemicals, Lund, Sweden; OTE P-lactoside, CETE 0- lactoside, and CETE P-N-acetyllactosaminide, Pierce Chemical Co.; octyl-Sepharose, p-nitrophenyl glycosides, 4-methylumbelliferyl-N- acetylneuraminic acid, methyl P-lactoside, ethyl P-lactoside, n-propyl P-lactoside, benzyl P-lactoside, ceramide type I11 (primarily nonhy- droxyl fatty acids), ceramide type IV (mainly a-hydroxyl fatty acids), sodium cholate, sodium deoxycholate, Tween 20, Nonidet P-40, Tri- ton X-100, and octyl p-glucoside, Sigma; sodium taurodeoxycholate, sodium taurodehydrocholate, sodium taurolithocholate, sodium tau- rocholate, sodium taurochenodeoxycholate, and dodecyl 8-maltoside, Calbiochem; Matrex gel blue A, Amicon Corp; Rotofor Cell isoelec- trofocusing unit and Bio-Gel A-0.5m, Bio-Rad. Hi80 (purity, >95%), ICN Biochemicals; Silica Gel 60 precoated plates, Merck, Darmstadt, Federal Republic of Germany; Iatrobeads, Iatron Laboratories, To- kyo, Japan; and Sep-Pak (C18 cartridge), Waters Associates.

Enzyme Assay-The incubation mixture contained [3H]G~1, 30 nmol (10,000 cpm); sodium cholate, 200 pg, and an appropriate amount of enzyme in 200 pl of 50 mM sodium acetate buffer, pH 5.0. After incubation at 37 "C for a preset time, the liberated radioactive oligosaccharide was analyzed as described (1). One unit is defined as the amount of enzyme which hydrolyzes 1 nmol of GM,/min at 37 "C. The specific activity of the enzyme is expressed as units/mg of protein. Protein concentrations were determined by the method of Lowry et al. (19) with bovine serum albumin as the standard.

When a nonradioactive glycosphingolipid was used as the sub- strate, the reaction was terminated by adding 5 volumes of chloro- form/methanol (2:1, v/v). The mixture was vortexed and centrifuged at 7000 X g for 5 min to separate the organic phase (lower) and the aqueous phase (upper). The two phases were separately evaporated to dryness. For the detection of the released oligosaccharide, the aqueous phase was analyzed by thin-layer chromatography (TLC) using n-butyl alcohol/acetic acid/H20 (2:1:1, v/v/v) as the developing solvent (20). Oligosaccharides containing sialic acid were revealed by resorcinol (21), whereas those containing neutral sugars were visual- ized by diphenylamine (22). For the analysis of the ceramides released by the enzyme, the organic phase was analyzed by TLC using chlo- roform/methanol (9:1, v/v) as the developing solvent (1). The cer- amides were revealed by staining the plate with Coomassie Brilliant Blue (23). Since the sphingosine and alkyl alcohols are not stained by Coomassie Brilliant Blue, the release of oligosaccharides (1) was used to determine the hydrolysis of lyso-GMl, lyso-LacCer, and alkyl P-lactosides. The quantitative estimation of the stained bands on a TLC plate was accomplished by scanning the plate using a Shimadzu CS-930 TLC scanner as described by Ando et al. (24). The lactose

released from alkyl 0-lactosides was also analyzed by gas chromatog- raphy (25).

Other Methods-Exoglycosidase activities were assayed using p - nitrophenyl glycosides as substrates (13). Neuraminidase was assayed using 4-methylumhelliferyl-N-acetylneuraminic acid as substrate (26). The isoelectric point of ceramide glycanase was determined by using Rotofor Cell isoelectrofocusing (27).

Purification of Ceramide Glycanase from M. decora-All operations were performed at 0-5 "C except that chromatography on octyl- Sepharose and Matrex gel blue A was run at room temperature. Centrifugations were carried out at 20,000-30,000 X g for 30 min in a Sorvall RC5C refrigerated centrifuge. Ultrafiltrations were carried out using Amicon stirred cells and PM-10 membranes.

Leeches (1 kg) were homogenized with 2.5 liters of distilled water in a Waring blender at 30-s intervals for a total of 2 min. The homogenate was centrifuged to obtain a water extract. After adjusting the pH of the extract to 4.8 with saturated citric acid, the precipitate was removed by centrifugation, and the supernatant was adjusted to pH 6.5 with a saturated Na2HP0,. To this solution, protamine sulfate (2 g/100 ml of H20) was added dropwise until the precipitate ceased to appear. At this point, the final concentration of protamine sulfate was about 0.5 mg/ml. The precipitate was removed by centrifugation, and the supernatant was brought to 30% saturation with solid am- monium sulfate (176 g/liter). After standing for 2 h, the precipitate was removed by centrifugation, and the supernatant was further brought to 55% saturation with solid ammonium sulfate (162 g/liter). After standing overnight, the precipitate was collected by centrifu- gation and dissolved in 100 ml of 50 mM sodium phosphate buffer, pH 7.0, to obtain a crude ceramide glycanase preparation. An equal volume of 2% sodium cholate solution was mixed with this prepara- tion and applied to an octyl-Sepharose column (2.5 X 20 cm) equili- brated with 1% sodium cholate solution. The column was subse- quently washed extensively with the same solution to remove the unadsorbed proteins. Under this condition, all exoglycosidases were washed off from the column. After the absorbance a t 280 nm of the effluent reached below 0.01, the column was washed with approxi- mately 4 liters of water. Ceramide glycanase retained by the column was then eluted with 500 ml of 1% octyl 0-glucoside. Fractions of 10 ml were collected. The fractions containing ceramide glycanase were pooled and concentrated to approximately 20 ml by ultrafiltration. This preparation was dialyzed against 50 mM sodium acetate buffer, pH 6.0, and applied to a Matrex gel blue A column (2.5 X 20 cm) equilibrated with 50 mM sodium acetate buffer, pH 6.0. The column was washed with the same buffer and then with water to remove the unadsorbed proteins, and the ceramide glycanase was eluted with 1 % sodium cholate. The active fractions were pooled, concentrated to 4.8 ml by ultrafiltration, and 2.4 ml of this solution was applied to a Bio- Gel A-0.5m column (1.5 X 80 cm) equilibrated with 50 mM sodium phosphate buffer, pH 7.0. The column was eluted with the same buffer, and the first protein peak that contained ceramide glycanase activity was pooled and concentrated to 1.6 ml by ultrafiltration. Table I summarizes a typical purification scheme and the recovery of ceramide glycanase from 1 kg of leeches.

Hydrolysis of G,, by Ceramide Glycanase in the Presence of Hi'0- To determine whether ceramide glycanase catalyzes the insertion of water between the C-1 of the glucose and the oxygen or between the oxygen and the C-1 of the sphingosine, we carried out the enzymatic hydrolysis of G M ~ in Hi80 and analyzed the released ceramide and glycan chain by mass spectrometry according to the following scheme. One milligram of G M ~ was dissolved in 0.5 ml of 50 mM sodium acetate buffer, pH 5.0, containing 500 p g of sodium cholate and freeze-dried. The solid was then dissolved in 0.5 ml of Hi'0. After brief sonication in an ultrasonic bath, the mixture was subsequently transferred to a

TABLE I Purification of ceramide glycanase from 1 kg of leeches (M. decora)

n Recovery

mg nmollmin unitslmg -fold % Crude extract 18,860 4,160 0.22 1.00 100 Protamine sulfate 11,450 3,320 0.29 1.32 80

treatment (NH4)zSOa (30-55%) 4,750 2,580 0.54 2.45 62 Octyl-Sepharose 84 1,760 20.91 95.23 42 Matrex gel blue A 28.4 1,370 48.27 219.27 33 Bio-Gel A-0.5 m 3.6 710 197.22 896.45 17

12274 Leech Ceramide Glycanase tube that contained 5 units of freeze-dried ceramide glycanase and incubated a t 37 "C for 48 h. For checking the extent of hydrolysis, a 5-pl aliquot of the reaction mixture was directly analyzed by TLC to monitor the release of the oligosaccharide and the remaining G M ~ . Under this condition, more than 95% of GMI was found to he hydro- lyzed. The reaction mixture was then partitioned by the addition of 2.5 ml of chlordform/methanol (2:1, v/v). The organic phase and the aqueous phase were separately evaporated to dryness. The aqueous phase, which contained the released oligosaccharide and a trace of GvI, was dissolved in 1 ml of 0.1 M KC1 solution and applied to a Sep-Pak that had been equilibrated with 0.1 M KCI. The Sep-Pak was subsequently washed with 5 ml of water to remove the unadsorhed oligosaccharide. The water wash was lyophilized, dissolved in 0.2 ml of water, and applied to a Sephadex G-15 column (0.4 X 28 cm) to separate the oligosaccharide from inorganic salts. The column was eluted with water a t 2 ml/h, and a 0.2-ml fraction was collected. T o locate the oligosaccharide, a 3-pl aliquot of each fraction was analyzed by TLC using n-butyl alcohol/acetic acid/water (2:l:l. v/v/v) as the developing solvent (20). The sialo-oligosaccharide on the plate was revealed by resorcinol (21). The fractions that contained the oligo- saccharides were pooled, lyophilized, and analyzed by fast atom bombardment mass spectrometry (FAB-MS) using a Finnigan TSQ- 70 instrument. Ionization of the sample was performed with bom- barding xenon atoms a t 8.5 keV and positive ion detection mode. The Folch lower phase (organic phase) which contained the released ceramide and a trace of GMl was evaporated to dryness, dissolved in 1 ml of chloroform, and applied onto an Iatroheads column (0.5 X 7 cm) which had been equilibrated with chloroform. The column was washed with 3 ml of chloroform followed by 3 ml of chloroform/ methanol (95:5, v/v) to elute the ceramide. The two eluates were combined, evaporated to dryness, and analyzed by FAB-MS.

RESULTS AND DISCUSSION

Purification of Ceramide Glycanase-Using the above puri- fication scheme, we were able to achieve 890-fold purification with 17% recovery of the enzyme. This purification scheme was based on the use of the hydrophobic interaction chro- matography (octyl-Sepharose and Matrex gel blue A) and gel filtration (Bio-Gel A-0.5m). Our efforts to use ion-exchange chromatography and preparatory isoelectrofocusing have not proved successful. By Rotofor Cell isoelectrofocusing, the enzyme was focused between pH 4.7 and 4.9. However, this method caused the enzyme to precipitate at this pH range and caused thereby the loss of activity due to electrolysis. We found, therefore, that during the purification, after removing the precipitable material from the crude extract at pH 4.8, the pH of the supernatant should be immediately brought back to 6.5 to minimize the loss of the activity due to the coprecipitation of the enzyme with other proteins that precip- itate at this pH. As shown in Table I, octyl-Sepharose chro- matography is the key step in our purification scheme. The inclusion of 1% sodium cholate in the column equilibration solution and also in the enzyme solution was to reduce the adsorption of the contaminating proteins by the column. This was based on the fact that the enzyme is more hydrophobic than most of the contaminating proteins. Washing the column with 1% sodium cholate followed by H 2 0 resulted in removing the bulk of the contaminating proteins and exoglycosidases. This step alone resulted in a 38-fold purification with 68% recovery of the enzyme. For practical purposes, the enzyme obtained from a single step chromatography using an octyl- Sepharose column can be used to release glycan chains from various glycosphingolipids. In fact, this preparation contained only two major protein bands (Fig. 1). Before the octyl- Sepharose chromatography step, the units of ceramide glycan- ase can only be determined by analyzing the released ceramide using quantitative TLC because of the interference of exogly- cosidases. After this step, both the oligosaccharide and the ceramide released can be used to assay the enzyme activity.

General Properties of Leech Ceramide Glycanase-The final enzyme preparation showed one major band stained by Coo-

FIG. 1. Sodium dodecyl sulfate-polyacrylamide gel electro- phoresis of ceramide glycanase preparations. Electrophoresis was carried out in 5% gel. Lane 2,60 pg of the crude enzyme fraction precipitated by (NH,),SO, (30-55%); lane 3,20 pg of the preparation obtained after octyl-Sepharose chromatography; lane 4, 25 pg of the preparation ohtained after Matrex gel blue A chromatography; lane 5 1 2 pg of the preparation ohtained after Bio-Gel A-0.5m chromatog- raphy. Lanes 1 and 6 are the marker proteins. From bottom to top (in kDa): a-lactalubumin (14.4), soybean trypsin inhibitor (20.1), car- bonic anhydrase (30), ovalbumin (43), bovine serum albumin (671, and phosphorylase b (94).

massie Blue on sodium dodecyl sulfate-polyacrylamide gel electrophoresis with a molecular mass of 54 kDa (Fig. 1). This enzyme, however, showed a molecular mass of 330 kDa on Bio-Gel A-0.5m gel filtration. For some unknown reason, the enzyme did not enter the gel on native polyacrylamide gel electrophoresis even with 5% gel. By incubating 0.2 units of the enzyme with various p-nitrophenyl glycosides, 4-meth- ylumbelliferyl-N-acetylneuraminic acids, and glycosphingo- lipids at 37 "C for 17 h, the purified ceramide glycanase was found to be free from a- and @-glucosidases, a- and 8-galac- tosidases, a- and 8-mannosidases, a-N-acetylglucosamini- dase, a-N-acetylgalactosaminidase, &hexosaminidase, a-L- fucosidase, and neuraminidase. The enzyme was also devoid of protease activity using Azocasine (28) and Azocoll (29) as substrates.

Ceramide glycanase was found to be stable between pH 4.5 and 8.5 with the highest activity at pH 5.0 using Gnnl as substrate. The PI of this enzyme was found to be between pH 4.7 and 4.9. At a concentration of 5 mg/ml, the enzyme did not lose any appreciable activity at -20 "C for 3 months or at 37 "C for 1 week.

The effect of substrate concentration on the rate of hydrol- ysis was examined by using various concentrations of G M ~ under the standard assay conditions (see "Enzyme Assay"). As shown in Fig. 2, the rate of hydrolysis increased with the increment of GMl concentration and reached the maximum a t the concentration of 150 p ~ . The K,,, for GM1 was estimated to be 1.54 X M from the double-reciprocal plot (30). We chose the substrate concentration of 30 nmo1/200 pl (150 pM) to ensure that the reaction was carried out at an excess amount of substrate. This concentration was above the critical micellar concentration of monosialogangliosides reported by Ulrich-Bott and Wiegandt (31). However, in our incubation mixture, the presence of a high concentration (about 2.3 mM)

Leech Ceramide Glycanase 12275

hydrolysis of LacCer. Under the same conditions, the hydrol- ysis of G M ~ required a much lower concentration (1.5 gg/pl) of the detergent to reach the maximal activity. These results may reflect the difference in the solubility of these two gly- cosphingolipids.

Nonionic detergents such as Triton X-100, Tween 20, and Nonidet P-40 were not effective in stimulating the hydrolysis of GMl by ceramide glycanase.

Hydrolysis of Various Glycosphingolipids by Ceramide Gly- canase-Table 111 summarizes the substrate specificity of this enzyme. Among various glycosphingolipids studied, Gg- Ose3Cer and GgOselCer were the two best substrates followed by GMr and GM~. As evident from Table 111, the rate of hydrolysis of G M ~ is considerably slower than that of Gg- Ose,Cer. Gol,, in turn, is hydrolyzed much slower than G M ~ . Also, the rate of hydrolysis of GT1 is slower than that of Gol,, suggesting that there is an inverse relationship between the rate of hydrolysis and the number of sialic acids attached to the glycolipid substrates. These results may indicate that the negatively charged sialic acids interfere with the interaction between the substrate and the enzyme. With the exception of GlcCer, the rate of hydrolysis of LacCer, which was at 20% the rate of GM1, was the slowest among various glycosphin- golipids. For practical purposes, 0.5 units of the enzyme would be sufficient to cleave completely those glycosphingolipids isolated from higher animals in 13 h under our assay condi- tions.

The purified ceramide glycanase was found to cleave GlcCer at a very slow rate, about 5% that of GMI, even though it was free from /3-glucosidase activity as assayed by using p-nitro- phenyl P-glucose. Whether or not the hydrolysis of GlcCer is intrinsic activity of leech ceramide glycanase remains to be determined.

In addition to the glycosphingolipids listed in Table 111, the leech ceramide glycanase also hydrolyzed unusual glycosphin- golipids such as 3-OMeGal~1+3GalNAca1~3[6”0-(2-ami- noethylphosphonyl)Gal~l~2~-(2-aminoethylphosphonyll -+6)Gal~1-+4Glc/31-+1Cer, IV3-(3’-S03-GlcA)-nLcOse&er, GalNAca1~3Gal(2tlcFuc)~1+3GlcNAc~1~3Gal~~+ 4Gal/31+4GlcCer, and polyglycosylceramides, although at a much slower rate. This indicates that the leech ceramide glycanase has a broad specificity that makes this enzyme useful for structural analysis of glycosphingolipids.

To investigate the effect of the fatty acyl residue in glyco- sphingolipids on the rate of hydrolysis, we studied the hy- drolysis of lyso-GMl and lyso-LacCer by ceramide glycanase.

- 0.12 - A

I 0.1 0 . 2 0 . 3 0 . 4

l/ [ G M l i uM-’ U

0 I I I 1 1 1 0 5 0 100 150 2 00 2 5 0

IGM11 v M FIG. 2. Effect of GM, concentration on the rate of hydrolysis.

Ceramide glycanase (0.2 units) was incubated with various concen- trations of GM1 from 2.5 to 225 WM for 10 min. The detailed assay conditions are described under “Enzyme Assay.” The inset shows the double-reciprocal plot.

of sodium cholate might have changed the micellar property of the glycolipid substrates.

The effects of metal ions on ceramide glycanase showed that ZnZ+ at 5 mM inhibited 60% of the activity, whereas Cu2+, Ag’, and Hg2+ at 1 mM inhibited 64, 80, and 95% of the activity, respectively. Ba2+, Ca2+, Co2+, Mg2+, and Mn2+ at 5 mM did not have an appreciable effect.

Effect of Detergents on the Hydrolysis of Glycosphingolipids by Ceramide Glycanase-The hydrolysis of glycosphingolipids by ceramide glycanase in the absence of a detergent is very slow. This reaction is stimulated by bile salts. Among various bile salts tested, sodium cholate was most effective except for the hydrolysis of LacCer (Table 11). It should be noted that the maximal solubility of sodium cholate in 50 mM sodium acetate buffer, pH 5.0, is close to 1 wg/wl. Above this concen- tration, ceramide glycanase activity was severely inhibited. For the hydrolysis of LacCer, sodium taurodeoxycholate was better than sodium cholate. Fig. 3 shows the effect of sodium taurodeoxycholate concentration on the hydrolysis of LacCer. For comparison, a parallel study on the hydrolysis of GM1 was also performed. The results indicated that sodium taurodeox- ycholate at 2.5 pg/pl was most effective in stimulating the

TABLE I1 Effects of detergents on the hydrolysis of glycosphingolipids by cerarnide glycanase

In all cases, the reaction mixture contained 30 nmol of the glycosphingolipid substrate. For the hydrolysis of GM, G M ~ , G m GD1., GbOse,Cer, and GbOserCer, each glycosphingolipid was incubated with 0.2 units of the enzyme a t 37 “C for 1 h, whereas LacCer was incubated with 0.4 units of the enzyme a t 37 “C for 2 h. Each incubation mixture contained one of the detergents shown at a final concentration of 1 pg/pl. Ceramide glycanase activity was assayed by quantitative TLC according to the conditions described under “Experimental Procedures.”

Detergents GML G M ~ GMM~ GD,. GbOse&er GbOse,Cer LacCer

% hydrolysis Sodium taurodeoxycholate 100 100 100 100 Sodium cholate 232 204 120

100 100

100 209

Sodium deoxycholate 16 41 0 0 65

24 4

Sodium taurocholate 2 10 0 0 12 3

64 44

49 Sodium taurochenodeoxycholate 61 37 Sodium taurodehydrocholate

112 8 15

286 36 21 0 0

Sodium taurolithocholate 0 0 0 2

0 0 0 0 Tween 20

0 0

Triton X-100 0 Nonidet P-40 0 - -

100 423

0

- - - -

- -

- - -

- - -

- - -

-

a -. not determined.

12276 Leech Ceramide Glycanase

‘0 1 . b 2:O 310 4:O ’ TDC IrW rll

FIG. 3. Effect of sodium taurodeoxycholate concentration on the hydrolysis of LacCer and GMl by ceramide glycanase. The hydrolysis of LacCer (0) was carried out by incubating 0.4 units of ceramide glycanase with 30 nmol of the substrate and varying amounts of sodium taurodeoxycholate in 200 pl of 50 mM sodium acetate buffer, pH 5.0, at 37 “C for 2 h. The hydrolysis of GMl (0) was performed under the same conditions except 0.2 units of the enzyme and 30 min were used for incubation. The detailed assay conditions are described under “Experimental Procedures.” TDC, sodium taurodeoxycholate.

TABLE 111 Hydrolysis of various glycosphingolipids by ceramide glycanase

Ceramide glycanase (0.2 units) was incubated with 30 nmol each of different glycosphingolipids in 200 p1 of 50 mM sodium acetate buffer, pH 5.0, containing 200 pg of sodium cholate at 37 “C. The incubation time varied depending on the substrate: for GM1, Gg- OsesCer, GgOse4Cer, GMnp, and GbOseaCer, 30 min; for GbOse4Cer, GbOsesCer, nLcOse4Cer, G M ~ , LacCer311-SO4, nLcOseaCer, and GTlb, 1 h; for LacCer and GlcCer, 2 h. The ceramide glycanase activity was determined by quantitative TLC as described under “Experimen- tal Procedures.”

Glycosphingolipids Relative activity

%

G M ~ 100 GgOsesCer 197 GgOse4Cer 185 GMV~~ 115 GbOsesCer 67 GbOse4Cer 50 GbOse&er 46 GDI. 41 nLcOse4Cer 38 G M ~ 35 LacCer311-S04 34 nLcOse3Cer 31

LacCer 19 GlcCer 5

GTlb 22

In both cases, the rate of hydrolysis of the deacyl derivative was only 10% of the rate of intact glycosphingolipids. Hy- drolysis of sphingosine p-lactoside (lyso-LacCer) by ceramide glycanase prompted us to examine the hydrolysis of stearyl p-lactoside and other synthetic P-lactosides by the enzyme.

Hydrolysis of Synthetic p-Lactosides by Ceramide Glycan- use-To understand the effect of the structure and hydropho- bicity of the aglycon portion of glycolipids on the rate of hydrolysis, we examined the hydrolysis of several synthetic P-lactosides by ceramide glycanase. As shown in Table IV, short chain alkyl P-lactosides such as methyl, ethyl, and n- propyl P-lactosides were found to be refractory to ceramide glycanase. The increase in chain length of the alkyl group starting from n-butyl P-lactoside resulted in the concomitant increase in the rate of hydrolysis which reached a maximum a t 14 carbons. Based on this result, if dodecyl P-lactoside was available, it should also be hydrolyzed by this enzyme. The rate of hydrolysis of stearyl p-lactoside, on the other hand, was much slower than that of myristyl p-lactoside. Other

TABLE IV Hydrolysis of synthetic lactosides by ceramide glycanase

Synthetic p-lactosides (30 nmol) were incubated with 0.9 units of ceramide glycanase in 200 pl of 50 mM sodium acetate buffer, pH 5.0, at 37 “C for 3 h. The TDC concentration was 1 pg/+l. The lactose released was analyzed by gas chromatography as described under “Experimental Procedures.” TDC, sodium taurodeoxycholate; PNP, D-nitrouhenvl.

Svnthetic S-lactoside -TDC +TDC

PNP lactoside Benzyl lactoside Phytyl lactoside Methyl lactoside Ethyl lactoside n-Propyl lactoside n-Butyl lactoside n-Octyl lactoside Myristyl lactoside Stearyl lactoside OTE lactoside CETE lactoside

% hydrolysis 0 0 0 0 0 0 0 0 0 0 0 0 0.3 0.1 2.5 6.7

13.3 34.5 0.5 4.6 1.7 3.5 2.8 13.8

TABLE V Hydrolysis of glycolipids with different disaccharides

by ceramide glycanase Each substrate (30 nmol) was incubated with the indicated amount

of ceramide glycanase at 37 “C. The assay conditions are described under “Experimental Procedures.”

Substrate enzvme used time Amount of Incubation % Hydrolysis

units h

Galpl4Glcp-Cer 0.5 2 100 Manpl“AGlcp+Cer 0.5 2 40 Al-Gal@l+4Glcp+Cer 0.5 2 5 Glcal+4Glcp+dodecanol 0.9 4 0 Galal+6Glc@+Cer 0.5 2 0 Galal+4Gal@+Cer 0.5 2 0 Galpl+4GlcNAcp+CETE 0.9 4 0

synthetic lactosides such as OTE and CETE P-lactosides were also hydrolyzed. p-Nitrophenyl, benzyl, and phytyl p-lacto- sides were resistant to hydrolysis. The above results indicate that the chain length, structure, and the hydrophobicity of the aglycon portion of the glycolipids have a profound effect on the rate of hydrolysis.

Effect of the Carbohydrate Moiety of Glycolipids on the Rate of Hydrolysis-As shown in Table 111, the shortest carbohy- drate chain on a glycosphingolipid which could be efficiently hydrolyzed by ceramide glycanase was found to be LacCer. In order to understand the structural requirement of the disac- charide unit for the enzyme, we studied the hydrolysis of five disaccharides, Galpl+4GlcNAcp+, Galal+4Galp+, Manpl-+IGlcp+, Glcal+4Glcp+, and Galal-&Glcp+ at- tached to ceramide or to an alkyl chain. As shown in Table V, GalaldGalpCer and Galpl4GlcNAcpCETE were not hydrolyzed by the enzyme. In these two glycolipids, the first sugar attached to the ceramide or CETE is either galactose or GlcNAc instead of glucose. These results indicate that the identity of the first sugar attached to the hydrophobic portion of the glycolipid substrate is recognized by the enzyme. In addition, the second sugar and its linkage to the glucose moiety also affect the rate of hydrolysis. Manpl4GlcpCer was hydrolyzed a t a rate 40% of that for LacCer, and the rate of hydrolysis of Al-LacCer was only 5% in comparison with LacCer. GlcaldGlc/3dodecanol and Galal+6GlcpCer, on the other hand, were completely refractory to ceramide gly- canase. These results clearly indicate that the nature and linkage of the sugar unit next to the glucose moiety can also

Leech Ceramide Glycanase 12277

% 100"

1 80 - -

I Al

I I I 1020 1021 1022 1023 1024 1025

FIG. 4. FAB-MS analysis of the oligosaccharide released from the hydrolysis of Gul by ceramide glycanase in H20 and Hi80. The experimental conditions were described under "Experi- mental Procedures." A , mass spectrum of the oligosaccharide released when the experiment was carried out in H20. B, mass spectrum of the oligosaccharide released when the experiment was performed in Hi80.

affect the rate of hydrolysis. We also studied the effect of the conversion of the 6th carbon of galactose in LacCer into an aldehyde group (Al-LacCer) on the rate of hydrolysis. The fact that the conversion of LacCer to Al-LacCer rendered the substrate resistant to ceramide glycanase may suggest the possible interaction of the enzyme with the C-6 region of the galactose residue in LacCer.

The Mode of Cleavage of the Linkage between the Ceramide and the Glycan Chain-The above results indicate that cer- amide glycanase recognizes the glycan chain as well as the hydrophobic portion of the glycolipid substrates. Is the cleav- age of the glycosidic linkage between the glucose and the oxygen or between the oxygen and ceramide? To answer this question, we carried out the hydrolysis of GM, in the presence of Hi80 and analyzed the released oligosaccharide and cer- amide by FAB-MS. Fig. 4 shows the FAB-MS analysis of the oligosaccharide released by ceramide glycanase in H20 and in Hi80. When the enzymatic hydrolysis was carried out in H20, the ion at m/z 1021 corresponding to the sodium adduct of the molecular ion [M+Na]+ of the oligosaccharide was de- tected (Fig. 4A), whereas an [M+Na]+ at m/z 1023 was detected when the enzymatic hydrolysis was carried out in Hf80 (Fig. 4B). In the case of ceramide, no such shift was observed, and the mass spectra of the ceramides released in both cases were found to be identical (results not shown). These results support the enzymatic cleavage at the glucose C-0 bond.

In conclusion, our studies revealed that in addition to ceramide glycans (glycosphingolipids), the leech ceramide gly- canase can also hydrolyze sphingosine glycans (lyso-glyco- sphingolipids) and alkyl glycans. Our results also showed that ceramide glycanase has binding domains that recognize both the glycan chain and the hydrophobic portion of the glycolipid substrates. Based on the fact that the enzyme exerted the fission at the glucose C-0 bond of the glycosidic linkage, this enzyme is an endoglycosidase.

We have chosen to isolate ceramide glycanase from M. decora because of its abundance in North America. The spec- ificity of the leech ceramide glycanase toward glycosphingo- lipids is similar to that of Actinomycetes endoglycoceramidase (3). The Actinomycetes enzyme hydrolyzed GgOse4Cer (asialo- GM1) at a rate slightly slower than that of GM1 (3), whereas the leech enzyme hydrolyzed GgOse4Cer at a rate twice that

The Actinomycetes enzyme was assayed by measuring the production of reducing power using bovine brain ganglioside mixture as the substrate. The assay of the leech ceramide glycanase was based on the liberation of the radioactive oligosaccharide from the tritium-labeled G;cli. Therefore, it would be not possible to compare directly the activity of these two enzymes based on those derived units. However, we found that 1 unit of the leech ceramide glycanase is roughly equiv- alent to 1 milliunit of the Actinomycetes endoglycoceramidase.

of GM1.

REFERENCES 1. Li, S.-C., DeGasperi, R., Muldrey, J. E., and Li, Y.-T. (1986) Biochem.

2. Li, Y.-T., Ishikawa, Y., and Li, S.-C. (1987) Biochem. Biophys. Res. Com-

3. Ito, M., and Yamagata, T. (1986) J. Biol. Chem. 261,14278-14282 4. Svennerholm, L. (1972) Methods Carbohydr. Chem. 6,464-474 5. McKibbin, J. M. (1969) Biochemistry 8,679-685 6. Li, Y.-T., Li, S.-C., and Dawson, G. (1972) Biochim. Biophys. Acta 260,

Biophys. Res. Commun. 141,346-352

mun. 1 4 9 , 167-172

88-99 7. Esielm&, W. J., Laine, R. A., and Sweeley, C. C. (1972) Methods Enzymol.

8. Radin, N. S. (1972) Methods Enzymol. 28,300-306 9. Svennerholm, L., Minsson, J.-E., and Li, Y.-T. (1973) J. Biol. Chem. 248 ,

28,140-156

740-749 . .. . " 10. Koenigs, W., and Knorr, E. (1901) Ber. 34, 957-981 11. Conchie, J., and Lewy, G. A. (1963) Methods Carbohydr. Chem. 2, 335-

12. Neuenhofer, S., Schwarzmann, G., Egge, H., and Sandhoff, K. (1985)

13. Li, Y.-T., and Li, S.-C. (1972) Methods Enzymol. 28, 702-713 14. Leskawa, K. C., Dasgupta, S., Chien, J.-L., and Hogan, E. L. (1984) Anal.

15. Gardas, A., and Koscielak, J. (1973) Eur. J. Biochem. 3 2 , 178-187 16. Araki, S., Satake, M., Ando, S., Hayashi, A., and Fujii. N. (1986) J. Biol.

337

Biochemistry 24, 525-532

Biochem. 140 , 172-177

C h m 281. 5138-5144 17. Slomiany, A., Slomiany, B. L., and Horowitz, M. I. (1974) J. Biol. Chem.

18. Sugita, M., Nishida, M., and Hori, T. (1982) J. Biochem. (Tokyo) 92,327-

. . . . . . .. - - - , - - - - - - - -

249 , 1225-1230

RRA 19. Lo;;, 0. H., Rosehrough, N. J., Farr, A. L., and Randall, R. J. (1951) J.

20. Kitamikado, M., Ito, M., and Li, Y.-T. (1981) J. Bid. Chem. 266, 3906- Bid. Chem. 193,265-275

74n4

21. Svennerholm, L. (1957) Biochim. Biophys. Acta 24,604-611 22. Harris, G., and MacWilliams, 1. C. (1954) Chem. Ind. ( L o r d . ) 249 23. Nakamura, K., and Handa, S. (1984) Anal. Biochen. 142,406-410 24. Ando, S., Chang, N.-C., and Yu, R. K. (1978) Anal. Biochem. 89,437-450 25. Laine, R. A., Esselman, W. J., and Sweeley, C. C. (1972) Methods EnzymoL

26. Potier, M., Mameli, L., Belisle, M., Dallaire, L., and Melancon, S. B. (1979)

""-"

28,159-167

Anal. Biochem. 94. 287-296 27. Egen, N.~B., Thormann, W., Twitty, G. E., and Bier, M. (1984) in Electro-

28. Tomarelli, R. M., Charney, J., and Harding, M. L. (1949) J. Lab. Clin. Med.

~ ~, ~~ ~~~

phoresis (Hirai, H., ed) pp. 547-550, Walter de Gruyter, Berlin, FRG

34.428-433 29. Moore, G . L. (1969) Anal. Biochem. 32, 122-127 30. Dixon, M., and Webb, E. C. (1979) Enzymes 3rd Ed., pp. 55-68, Academic

31. Ulrich-Bott, B., and Wiegandt, H. (1984) J. Llpid Res. 25, 1233-1245

. _ I ~" ."

Press, New York