Eur. J. Biochem. 146,315-321 (1985) 0 FEBS 1985

Characterization of human glucosylsphingosine glucosyl hydrolase and comparison with glucosylceramidase

Anna Maria VACCARO, Michele MUSCILLO, and Kunihiko SUZUKI Laboratory of Metabolism and Pathological Biochemistry, Istituto Superiore di Sanita, Roma; and The Saul R. Korey Department of Neurology, Department of Neuroscience, The Rose F. Kennedy Center, Albert Einstein College of Medicine, Bronx, New York

(Received May 15/September 17, 1984) - EJB 84 0505

Properties of glucosylsphingosine (gluco-psychosine) glucosyl hydrolase were studied in detail in cultured human fibroblasts and placenta and were compared with those of glucosylceramidase. The two activities, that are deficient in tissues of Gaucher patients, showed minor but consistent differences. The pH optima were 4.8 for psychosine hydrolysis and 5.3 for glucosylceramide hydrolysis. In the presence of oleic acid, taurocholate activated glucosylceramidase more than 10-fold, while it activated psychosine hydrolysis only by about 30%. Triton X-100 was stimulatory for glucosylceramidase but was strongly inhibitory for psychosine hydrolysis. Phospholipids, that increase many times glucosylceramidase activity, were moderately inhibitory to enzymatic hydrolysis of psychosine. The psychosine hydrolase activity was slightly more heat-stable than the glucosylceramidase activity. The K, values for the two substrates were similar; 1 . 7 ~ M for glucosylceramide. The V for glucosylceramide was, however, 100-times that for psychosine hydrolysis. Psychosine acted as a potent non-competitive inhibitor (Ki = 1.8 x M), while glucosylceramide was a weak inhibitor against psychosine hydrolysis. Within the limit of glucosylceramide solubility, psychosine hydrolysis could not be inhibited by more than 50%. Furthermore, the Dixon plot of glucosylceramide inhibition showed an anomalous slope. The ratio of the two activities was similar in fibroblasts, in the placenta mitochondria-lysosomal fraction and in a partially purified placental preparation. These findings are best explained by the hypothesis that, although the two substrates are hydrolyzed by a single enzyme, they share an overlapping but not identical catalytic site while binding to hydrophobic sites unique for the respective substrates.

M for psychosine and 2 . 7 ~

Gaucher disease is an autosomal recessive disorder due to a deficiency of glucosylceramidase activity [l, 21. Abnormal accumulation of glucosylceramide occurs in tissues of patients. Glucosylsphingosine (gluco-psychosine) is chemical- ly and metabolically closely related to glucosylceramide [3]. It is not a constituent of normal tissues, but its presence in the spleen and brain of patients with Gaucher disease has been reported [4, 51. Glucosylsphingosine-hydrolyzing activity was deficient in the spleen and fibroblasts of Gaucher patients [6]. This provided prima facie evidence that a single enzyme is responsible for hydrolysis of both glucosylceramide and glucosylsphingosine. However, the nature of the enzymatic hydrolysis of glucosylsphingosine has not been studied in detail.

In thls report we describe the enzymological characteristics of gluco-psychosine hydrolysis by human fibroblasts and placenta and compare them with those of glucosylceramide hydrolysis. The results indicate that, although a single enzyme is probably responsible for hydrolysis of both of the substrates, they may not share identical binding sites.

MATERIALS AND METHODS

Sodium taurocholate (crude and pure), taurodeoxy- cholate, cholate, deoxycholate and oleic acid were purchased from Sigma Chemical Co. (St Louis, MO, USA), Triton X-

Enzymes. Glucosylceramidase (EC 3.2.1.45).

100 was from BDH Chemicals Ltd (Poole, UK), phospha- tidylserine, phosphatidylcholine, phosphatidylethanolamine, and phosphatidylinositol (from plants) were from Supelco Inc. (Bellefonte, PA, USA), phosphatidylinositol from yeast was from Koch-Light Labs (Colnbrook, UK).

Enzyme source

Skin fibroblasts were cultured from skin biopsies obtained from the forearm of normal individuals and from patients with Gaucher disease. They were grown in Ham’s F-10 medi- um supplemented with 15% fetal calf serum, penicillin (100 U/ ml) and streptomycin (100 yglml). The cells were harvested by trypsinization, homogenized in distilled water with an Ultra- Turrax homogenizer (Ika-Werk, Staufen, FRG) and sonicated in a Bransonic 52 ultrasonic bath. The concentration of pro- tein was adjusted to 2 mg/ml with water.

Human placenta was obtained at delivery. After rinsing with cold distilled water it was cut in small pieces and homogenized in 9 vol. of cold water in a Waring blender (‘total homogenate’). When the crude mitochondria-lysosome fraction was prepared, placenta was homogenized in ice-cold 0.32 M sucrose and centrifuged at 900 x g for 10 min. The supernatant was centrifuged at 11 000 x g for 20 min. The crude mitochondria-lysosomal pellet was suspended in water (3 ml/g placenta) and frozen-thawed three times. After 2-min sonication in the ultrasonic water bath, the suspension was centrifuged at I00000 x g for 60 min. The supernatant was

316

concentrated by ultrafiltration on the Amicon cell system with YM-10 membrane up to a protein concentration of 2 mg/ml. A partially purified placental glucosylceramidase was pre- pared essentially according to Furbish et al. [7] up to the step of butanol extraction.

The protein contents of the enzyme sources were determined by the dye-binding method according to Bradford [8] with bovine serum albumin as standard.

Preparation of substrates

Glucosylceramide, purified from spleen of Gaucher patients, was labelled in our laboratory with tritium at the 6-position of glucose following the method of McMaster and Radin [9]. It was diluted with the starting unlabelled gluco- sylceramide to a specific activity of 463 dpm/nmol.

[gluc0se-6-~H]Glucosylsphingosine was prepared from the labelled glucosylceramide (1 00 mg) by alkaline hydrolysis in an analogous procedure used for galactosylsphingosine [lo, 113. After hydrolysis, glucosylsphingosine was isolated by activated silicic acid (200 - 325 mesh) column chromatogra- phy (1.5 x 20 cm). After application of the sample in chloroform, the column was washed sequentially with mixtures of chloroform and methanol (150 ml 98:2, then 150 ml 90:lO). Psychosine was eluted with 200 ml of chloroform/methanol(80 : 20) and purified through a column of Dowex AG 50W-X8 (200 - 400 mesh, hydrogen form) (0.8 x 4 cm). The column was equilibrated with methanol. The labelled psychosine from the silicic acid column was applied in methanol/chloroform (90: 10). The column was washed with 36ml of the same solvent and the labelled glucosyl- sphingosine was eluted with 30 ml of methanol/l4 M ammo- nia (3 : 1). The final product gave a single spot on high-perfor- mance thin-layer plates, coated with 0.2-mm layer of silica gel 60 (E. Merck, Darmstadt, FRG), in the solvent systems chloroform/methanol/2.5 M ammonia (60:40 : 9, v/v/v) and chloroform/methanol/water (65 : 25 : 4, v/v/v). The structure of the product was confirmed to be that of glucosyl- sphingosine by mass spectrometric analysis of its trimethyl- sylyl derivative performed by a low resolution mass spectrom- eter and mass spectral data system (LKB model 2091/2130). Ionization was obtained by electron impact. The mass spectra was recorded at 70 eV and at a source temperature of 250°C by direct inlet probe. The probe temperature was raised from 25°C to 250°C in 30 min. A repetitive scan, from mass 1 to 850, with a scan time of 2 s was used.

The purified psychosine was adjusted to a specific activity of 2700 dpm/nmol with the unlabelled compound prepared in the same way.

Enzyme assays The following standard assay procedures were used to

measure the enzymatic activities. Glucosylceramidase. Each test tube contained, in a final

volume of 0.2 ml, 40 pg of [3H]glucosylceramide, 0.5 mg of pure sodium taurocholate, 50 pg of oleic acid, 0.1 ml of sodium citrate/phosphate buffer (0.2/0.4 M) pH 5.3 and 20 pg of protein (fibroblast homogenate or placenta mitochondria- lysosome fraction). The reaction time was 60 min at 37°C. When the partially purified placenta glucosylceramidase was the enzyme source, protein was 1 pg and the incubation time 30 min. The liberated glucose was isolated and its radioactiv- ity determined as described [12].

Gluco-psychosine hydrolase. Each test tube contained, in a final volume of 0.2 ml, 10 pg of [3H]psychosine, 0.5 mg of

pure sodium taurocholate, 50 pg of oleic acid, 0.1 ml of citrate/phosphate buffer (0.2/0.4 M) pH 4.8 and 100 pg pro- tein (fibroblast homogenate or placenta mitochondria-lyso- some fraction). The reaction time was 3 h at 37°C. Under these conditions the minimum amount of protein that allowed a reliable quantification of the enzymatic activity was 50 pg. When the partially purified placenta glucosylceramidase was tested for psychosine-hydrolyzing activity, protein was 5 pg and the reaction time 30min. The enzymatic reaction was stopped by the addition of 2.5 ml of chloroform/methanol (2: l), 0.3 ml of a 0.1% aqueous glucose solution and 0.05 ml of concentrated ammonia. After thorough mixing, the two phases were separated. The upper phase was washed twice with 2 ml of chloroform, transferred to a counting vial and dried. The residue was dissolved in 0.5 ml of water and, after addition of 12 ml of Insta-gel (Packard Becker, Groningen, Netherlands), the radioactivity determined. Blank counts were generally less than 15% of the sample counts.

Heat stability

Fibroblast homogenate was diluted 1 : 1 with the citrate- phosphate buffer pH4.8 and heated at 50°C or 55°C for varying periods of time. In another series of experiments, the enzyme source was heated for 5 min at different temperatures. After heating, activities to hydrolyze glucosylceramide and glucosylsphingosine were tested in the standard procedures described above. For psychosine hydrolysis, 100 pl (= 100 pg protein) of the heated sample was used directly as the enzyme source. For glucosylceramidase activity, the heated sample was diluted fivefold with citrate-phosphate buffer (0.2/0.4 M) pH 5.35 (the final pH 5.30) and 20 pg of protein was used for each assay tube.

RESULTS

Psychosine and glucosylcerarnide hydrolysis by fibroblasts

The pH optima for hydrolysis of psychosine and glucosylceramide by fibroblast homogenate were 4.8 and 5.3 respectively (Fig. 1). Neither the optima of pH nor the reac- tion velocities changed when the citrate/phosphate buffer con- centration varied in the range 0.05/0.1 M-0.1/0.2 M (final concentration in the assay mixtures). Sodium citrate buffer gave 10% lower values than the citrate-phosphate buffer, and sodium acetate buffer gave 50% lower values. The dependence of psychosine hydrolase activity on protein concentration was linear up to 200 pg protein per tube (Fig. 2) while that of glucosylceramidase was linear up to 50 pg/tube in confirma- tion of our previous results [13]. The reaction was linear for at least 3 h at 37°C for both reactions (Fig. 3, data for glucosylceramidase not shown). A single fibroblast prepara- tion gave specific activities of 175 nmol x h ~ x mg protein- and 1.8 nmol x h-' x mg protein-' toward glucosylceramide and glucosylsphingosine, respectively. Fibroblasts from Gaucher disease patients showed no activity to hydrolyze gluco-psychosine. When the fibroblast homogenate was centrifuged at I00000 x g for 60 min and the two enzymatic activities were determined in the supernatant and in the pellet, approximately 90% of the total activity was found still associ- ated with the pellet for both activities.

Effect of detergents

Pure taurocholate was an effective activator for both enzymatic activities (Fig. 4) : the psychosine hydrolase activity

317

3 4 5 6 7 PH

Fig. 1. Effect o fpH on the activities offibroblast psychosine hydrolase and glucosylccrramidase. The activities were assayed according to the standard procedures except that the pH was varied as indicated. (0-0) Psychosine hydrolase; ( A-A) glucosylceramidase

// J

O V L $ 0 10 0.20

Protein content, mg/tube

Fig. 2. Ejf'ct of enzyme concentrations on fibroblast psychosine h,ydrolase activity. The assays were carried out according to the stan- dard procedure except that the amount of the enzyme source was varied as indicated

Incubation Time, hours

Fig. 3. Linearity of fibroblast psychosine hydrolase activity according to the incubation period. The assays were carried out according to the standard procedure except that the period of incubation was varied as indicated

0 L 0 05 1

Sodium Taorochoiate. mg/fube

Fig. 4. Effect of sodium taurocholate on fibroblast psychosine hydrolase nndglucosylceramidase activities. The assays were carried out accord- ing to the standard procedures except that the amount of sodium taurocholate was varied as indicated. (0-0) Psychosine hydrolase; ( A-A) glucosylceramidase

0 50 100 150 200 Oleic acid, y g / t u b e

Fig. 5. Effect of oleic acid on fibroblast psychosine hydrolase and glucosylceramidase activities. The assays were carried out according to the standard procedures except that the amount of oleic acid was varied as indicated. (0-0) Psychosine hydrolase; ( A-A) glucosylceramidase

of the fibroblast homogenate increased by 50% at the optimum taurocholate concentration (0.25%) while the glucosylceramidase increased 1 0-fold at the same taurocholate concentration. Oleic acid increased the psychosine hydrolysis by 20% but glucosylceramide hydrolysis by fourfold (Fig. 5). Only 1 % of the glucosylceramidase activity under the optimal conditions could be detected in the absence of both taurocholate and oleic acid, while 40% of the optimal psychosine hydrolysis was measured without these two con- stituents. Effects of several other bile salts were tested for psychosine hydrolysis by the fibroblast homogenate (Table 1).

318

Table 1. Effect of bile salts on the psychosine hydrolase activity The assays were carried out according to the standard procedure except that the nature and the amount of bile salts were varied as indicated. The activities are expressed as a percentage of that mea- sured according to the standard system without taurocholate

Bile salt Amount Psychosine of bile salt hydrolase

activity

Table 2. Effect of phospholipids on the fibroblast psychosine hydrolase activity The assays were carried out according to the standard procedure except that taurocholate was omitted and the designated amount of phospholipids were added. The phospholipids were dissolved in chloroform/methanol(2: 1) and dried together with the oleic acid and the psychosine in the test tubes. They were then dispersed in the buffer by sonication. The activities are expressed as a percentage of that measured according to the standard system without taurocholate

None mg Yo

100

Pure taurocholate 0.5 1 .o

Crude taurocholate 0.5 1 .o

Taurodeoxycholate 0.5 1 .o

Cholate 0.5 1 .o

Deoxycholate 0.5

150 101 108 51 38 10 76 32 90

Phospholipids Amount Psychosine of phospholipid hydrolase

activity

Pg None P hosphatidylserine 25

from bovine brain 100 150

Pure phosphatidylinositol 25

150 from yeast 100

Pure phosphatidylinositol 25 from plant 100

1 .o 44

3001

‘ I -A

I -. I

0 5 10 15 20 Tr i ton X-100, mg/tube

Fig. 6. Effect of Triton X-100 on fibroblast psychosine hydrolase and glucosylceramidase activities. The assays were carried out according to the standard procedures except that sodium taurocholate was omitted and different amounts of Triton X-100 were added as indicat- ed. The relative activities are expressed as a percentage of that mea- sured according to the standard systems without taurocholate. (0-0) Psychosine hydrolase; (A- A) glucosylceramidase

Only the pure taurocholate had the capacity to activate psychosine hydrolysis.

The effect of a non-ionic detergent, Triton X-100, was of interest. When Triton X-100 was added to the reaction mixture at 0.25%, glucosylceramidase activity was stimulated by 20% while glucosylsphingosine hydrolysis was inhibited by 60%. If taurocholate was omitted from the reaction mixture, the differential effect of Triton X-100 on the two activities was even more dramatic (Fig. 6). I t increased the glucosylceramidase activity 2 - 3-fold while completely in- hibiting psychosine hydrolysis.

Eflect of phospholipids

Effects of several phospholipids on psychosine hydrolysis by the fibroblast homogenate were tested (Table 2). The purity

Yo

100 87 99 95 61 70 58 91 95

150 85

100 60 150 50

Phosphatidylethanolamine 25 61 100 57 150 39

Phosphatidylcholine 25 49

of the phospholipids was examined by thin-layer chromatog- raphy in the solvent systems of chloroform/methanol/14 M ammonia (70: 30: 5, v/v/v) [14] or chloroform/methanol/ water/acetic acid (50:40: 5: 3, v/v/v/v) [15]. All of them appeared pure when 20 - 30 pg was applied to the plate. All of the phospholipids tested showed relatively minor effects on psychosine hydrolysis. No effects were detected when 50 pg, 100 pg and 150 pg of the phospholipids were added to the standard assay system. In the absence of taurocholate, they were generally slightly to moderately inhibitory (Table 2).

Heat .stability

There was a small but consistent difference in heat stability of glucosylceramide-hydrolyzing and glucosylsphingosine- hydrolyzing activities of the fibroblast homogenate. When heated at either 50°C or 55°C for varying periods of time, the psychosine hydrolase activity was always slightly more heat- stable than the glucosylceramidase activity (Fig. 7). When the fibroblast homogenate was heated for 5 min at different temperatures, similar differences in heat-stability were observed; 50% inactivation was achieved at 50°C for gluco- sylceramidase and at 60°C for psychosine hydrolase.

Kinetic studies

The relationship between fibroblast psychosine hydrolase activity and the substrate concentration was of the Michaelis- Menten type, the K , value being 1.7 x M (Fig. 8). This is slightly lower than the K, of fibroblast glucosylceramidase (2.7 x M) [16]. The two K, have been determined under

319

$/ 1. ax 10 - 5 ~

L i P

30 60 Incubation t ime (rnin)

Fig. 7. Heat stability of fibroblast psychosine hydrolase and glucosylceramidase. The same fibroblast homogenate was heated at 50°C or 55°C for different times and than assayed for the two enzymatic activities according to the standard procedures. (0----0 , 0-0) Psychosine hydrolase; (A----A, A-A) glucosylceramidase. (----) 55°C; (-) 50°C

0 100 200 300 LOO l/lPsychosinel (rnM-’)

Fig. 8. Lineweaver-Burk plot for fibroblast psychosine hydrolase. The assay was performed according to the standard procedure except that the amount of the substrate was varied as indicated

the same conditions except the pH that was 4.8 and 5.3 re- spectively (optima of pH for psychosine and glucosylceramide hydrolysis). Psychosine was a non-competitive inhibitor against glucosylceramide hydrolysis with a Ki of 1.8 x lo-’ M (Fig. 9). On the other hand, glucosylceramide was a weak inhibitor against psychosine hydrolysis. Within the limit of glucosylceramide solubility in the reaction mixture, psychosine hydrolysis could not be inhibited by more than 50%. Furthermore, the Dixon plot of glucosylceramide in- hibition showed an anomalous slope (Fig. 10). This finding was consistent in many repeated experiments. The maximum concentration of glucosylceramide reported in Fig. 10 (30 x M) is almost 10-times the K,,, for glucosyl- ceramidase and is similar to the concentration we used in the standard assay which is within the linearity of the Lineweaver- Burk plot [I 3, 161. When the glucosylceramide concentration was doubled, the same 50% inhibition was obtained (data not shown). If only the portion of the curves below the inhibitor concentration of 10 x lo-’ M is taken, the inhibition is of a non-competitive type with an apparent Ki of 22 x l op5 M (Fig. 10).

Placental psychosine hydrolase and glucosylcerarnidase

The activities of psychosine hydrolase and glucosyl- ceramidase have been tested also in human placenta. When the enzymatic activities were determined in the total homogenate, the linearity of the reaction with respect to the

0 50 100 [Psychosinel (pM1

Fig. 9. Dixon plot of the inhibition by psychosine of fibroblast glucosylceramidase activity. The reaction rate v is expressed in nrnol/ tube. (0-0) 250 pM Glucosylceramide; (m-m) 62 pM glucosylceramide; ( A-A) 6.2 pM glucosylceramide

0 100 200 300 lGlucosylceromide1 (pM)

Fig. 10. Dixon plot of the inhibition by glucosylceramide of fibroblast psychosine hydrolase. The reaction rate v is expressed in nmol/tube. (0-0) 217 pM Psychosine; (ap=) 38 pM psychosine; (A-A) 16 pM psychosine

enzyme source was poor. The glucosylceramidase activity was linear only with less than 50 pg protein. The much lower specific activity of psychosine hydrolase together with the lack of linearity of the reaction rate as a function of the concentration of the enzyme source made reliable quantifica- tion of the enzymatic activity in the total homogenate impos- sible. The crude mitochondria-lysosomal fraction was pre- pared in order to overcome this difficulty. With this prepara- tion and with a partially purified placental glucosylcer- amidase, the pH optimum of both reactions, their time course and their dependence on the protein concentration were ident- ical to those of the fibroblast homogenate. Thus, the same standard assay systems could be used. The glucosylceramide- hydrolyzing and psychosine-hydrolyzing activities of human placental preparations together with that of fibroblast homogenate are given in Table 3. The ratio of the two activi- ties in all the preparations was similar.

DISCUSSION The psychosine hydrolase activity that we found in

cultured human fibroblasts (1.8 nmol x h-’ xmg protein-l) is similar to that reported earlier (1.35 nmol x h-’ xmg

320

Table 3. Psychosine hydrolase and glucosylceramidase activities in human placenta preparations and fibroblast homogenate The assays were carried out according to the standard procedures reported under Materials and Methods. Partially purified glucosylceramidase was prepared according to Furbish et al. [7] up to the step of butanol extraction

Enzyme source Glucosyl- Psychosine Glucosyl- ceramidase hydrolase ceramidase/ activity activity psychosine

hydrolase

nmol/h/mg

Placenta mitochondria- 158 1.51 104 lysosome fraction

glucosylceramidase Placenta partially purified 4390 44.2 100

Fibroblast homogenate 175 1.8 97

protein-') [6]. Its relative activity is only 1% of the glucosylceramidase activity. There are many consistent and perhaps significant differences in the properties of the psychosine hydrolase activity and glucosylceramidase activity in fibroblasts. The gluco-psychosine hydrolysis shows a sharp optimum at pH 4.8, while glucosylceramidase activity has a broad peak at pH 5.3. As reported earlier, glucosylceramidase is strongly activated by sodium taurocholate and oleic acid [17], without which it is almost inactive. In contrast, glucosylsphingosine hydrolysis is only moderately stimulated by this combination with 40% of the optimal activity still measurable without these constituents. These findings may be primarily due to the solubility and/or charge difference between the two substrates. However, other bile salts that have been reported to be effective activators of glucosylceramidase, such as cholate [18, 191 and taurodeoxycholate [20], inhibit psychosine hydrolase. Triton X-100, another well known stimulator of glucosylceramide hydrolysis [19], strongly in- hibits enzymatic hydrolysis of glucosylsphingosine either in the presence or absence of taurocholate. Phospholipids, such as phosphatidylserine and phosphatidylinositol that have been reported to increase many times the glucosylceramidase activity [21- 231 are also inhibitory to enzymatic hydrolysis of psychosine.

Despite the many points of differences, the possibility that two different enzymes are responsible for hydrolysis of glucosylceramide and glucosylsphingosine is unlikely because of the absence of psychosine hydrolase activity in fibroblasts of Gaucher patients and the constant ratio of the two activities in different preparation of the enzyme (Table 3).

Miyatake and Suzuki [I 1,241 studied enzymatic hydrolysis of analogous substrates, galactosylsphingosine and galacto- sylceramide, and concluded that they are hydrolyzed by a single enzyme. Among many pieces of evidence in support of their conclusion was the finding that galactosylcerainide was a powerful competitive inhibitor of galactosylsphingosine hy- drolysis. The behaviour of glucosylceramide as an inhibitor of glucosylsphingosine hydrolysis is completely different. It can inhibit psychosine hydrolase activity by no more than 50%, and its high apparent Ki value of 22 x lo-' M indicates that glucosylceramide is a very poor inhibitor of psychosine hydrolysis. On the other hand, psychosine can strongly inhibit glucosylceramide hydrolysis but the mode of inhibition is non-competitive. Other authors also found strong but non- competitive inhibition of hydrolysis of a glucosylceramide

derivative { 12- [N-methyl-N-(7-nitrobenz-2-oxa- 1,3 -diazol- 4-yl)]-am~nododecanoyl-sphingosyl-~-~-glucopyranoside} by psychosine [25, 261. The Ki value reported by these authors was 0.28 mM, about 15-times greater than ours (Ki 1.8 x M). The difference may well be just due to the different substrates.

To explain the non-competitive inhibition, these authors hypothesized that glucosylsphingosine binds not only to the catalytic site where it is hydrolyzed, but also to a hydrophobic site in the enzyme molecule blocking the interaction of the catalytic site with the substrate. This hypothesis cannot easily explain why glucosylceramide, which presumably binds only to the catalytic site, is such a poor inhibitor of psychosine hydrolysis. The fact that psychosine is a non-competitive in- hibitor of glucosylceramidase activity and that glucosylcer- amide has only very weak effect on psychosine hydrolysis suggests that the affinity of one substrate to the binding site of the other is very low. A more likely hypothesis might be that the two substrates share an overlapping but not identical catalytic site and have hydrophobic binding sites unique to the respective substrates. The aglycon portions of the two substrates are quite different in shape, and in particular, in the charge. The free amino group of psychosine carries a strong positive charge and is highly reactive, while the ceramide portion of glucosylceramide is very hydrophobic. The non-catalytic binding sites may be the sites for different effects of bile salts, phospholipids and Triton X-100 on hy- drolysis of these compounds. The slight difference in heat stability in hydrolysis of these substrates may indicate differ- ence in heat stability between the two unique binding sites. The Gaucher mutation could have occurred within the over- lapping portion of the active sites so that the catalytic activity for both of the substrates was destroyed.

The authors wish to thank Prof. L. Boniforti for the mass spectrometric analysis of psychosine, Mr E. Raia for fibroblast culti- vation and Mr E. Mazzeo for technical assistance. This investigation was supported in part by the North Atlantic Treaty Organisation research grant 208.81, by grant 83.01021 .51 from Consiglio Nuzionale Ricerche (Progetto Finalizzato Ingegneria Genetica e Basi Molecolari delle Malattie Ereditarie), and NS-10885, NS-19321, NS-03356, HD- 01799 from the United States Public Health Service.

REFERENCES 1. Brady, R. O., Kanfer, J . N. & Shapiro, D. (1965) Biochem.

2. Patrick, A. D. (1965) Biochem. J . 97,17C-l8C. 3. Suzuki, K. (1982) in Gaucher Disease: a Century ofDelineation

and Research (Desnick, R. J., Gatt, S. & Grabowski, G. A., eds) pp. 219-230, Alan R. Liss Inc., New York.

4. Raghavan, S. S., Mumford, R. A. & Kanfer, J. N. (1974) J . Lipid Res. 15,484-490.

5 . Nilsson, 0. & Svennerholm, L. (1982) J . Neurochem. 39, 709- 718.

6. Raghavan, S. S., Mumford, R. A. & Kanfer, J. N. (1973) Biochem. Biophys. Res. Commun. 54,256-263.

7. Furbish, F. S., Blair, H. E., Shiloach, J., Pentchev, P. G. & Brady, R. 0. (1977) Proc. Natl Acad. Sci. USA 74, 3560-3563.

8. Bradford, M. M. (1976) Anal. Biochem. 72, 248-2254, 9. McMaster, M. C. Jr. & Radin, N. S. (1977) J. Labelled Comp.

10. Taketomi, T. & Yamakawa, T. (1963) J . Biochem. (Tokyo) 54,

11. Miyatake, T. & Suzuki, K. (1972) J. Biol. Chem. 247, 5398-

12. Suzuki, K. (1978) Methods Enzymol. 5OC, 456-488.

Biophys. Res. Commun. 18, 221 -225.

Radiopharm. 13, 353 - 357.

444 -451.

5403.

321

13.

14. 15.

16.

17.

18.

19.

20.

Vaccaro, A. M., Kobayashi, T. & Suzuki, K. (1982) Clin. Chim.

Svennerholm, L. (1968) J . Lipid Res. 9, 570- 579. Harzer, K., Wassle, W., Sandhoff, K. & Jatzkewitz, H. (1968) Z .

Vaccaro, A. M., Muscillo, M. & Suzuki, K. (1983) Clin. Chim.

Wenger, D. A., Sattler, M., Clark, C., Wharton, C. (1976) Life

Svennerholm, L., Hakansson, G., Mansson, J. E. & Vanier, M.

Blonder, E., Klibansky, C. & De Vries, A. (1976) Biochim.

Raghavan, S. S., Topol, J. & Kolodny, E. H. (1980) Am. J . Hum.

21. Acta 118, 1-7. 22.

Anal. Chenz. 243, 527- 536.

Acta 131, 1 - 13.

Sci. 19,413-420.

T. (1979) Clin. Chim. Acta 92, 53-64.

Biophys. Acta 431, 45 - 53.

Genet. 32, 158-173.

23.

24. 25.

26.

Ho, M. W. & Light, N. D. (1973) Biochem. J . 136, 821 -823. Peters, S. P., Coyle, P., Cofee, C. J., Clew, R. H., Kuhlenschmidt,

M. S., Rosenfeld, L. & Lee, Y. C. (1977) J . Biol. Chem. 252,

Shafit-Zagardo, B. &Turner, B. M. (1981) Biochim. Biophys. Acta

Miyatake, T. & Suzuki, K. (1974) J . Neurochem. 22,231 -237. Gatt, S., Dinur, T. & Desnick, R. J. (1982) in Gaucher Disease:

a Century qf Delineation and Research (Desnick, R. J., Gatt, S. & Grabowski, G. A.) pp. 315-331, Alan R. Liss Inc., New York.

Desnick, R. J., Grabowski, G. A., Dinur, T., Fabbro, D., Goldblatt, J. & Gatt, S. (1982) in Membranes and Genetic Disease (Sheppard, J. R., Anderson, V. E. & Eaton, J. W., eds) pp. 193-215, Alan R. Liss Inc., New York.

563 - 573.

659,7- 34.

A. M. Vaccaro and M. Muscillo, Laboratorio Metabolism0 e Biochimica Patologica, Istituto Superiore di Sanit;, Viale Regina Elena 299,I-00161 Roma, Italy K. Suzuki, The Saul R. Korey Department of Neurology, Department of Neuroscience, The Rose F. Kennedy Center, Albert-Einstein College of Medicine, Yeshiva University, 1300 Morris Park Avenue, Bronx, New York, USA 10461