Bioch#nica et Biophysica Acta, 1i 19 (1992) 239-246 © 1992 Elsevier Science Publishers B.V. All rights reserved 0167-4838/92/$(15.00

239

BBAPRO 34121

Reconstitution of glucosylceramidase on binding to acidic phospholipid-containing vesicles Anna Maria Vaccaro, Massimo Tatti, Fiorella Ciaffoni, Rosa Salvioli and Paola Roncaioli Department of Metabolism and Pathological Biochemistry, Istituto Superiore Sanitd, Roma (Italy) (Received 18 July 1991) (Revised manuscript received 28 Septembe~ 1991)

Key words: Lysosomal enzyme; Glucosylceramidase; Reconstitution; Lipid binding; Acidic phospholipid

Studies were conducted to investigate the mechanism by which acidic phospholipid-containing vesicles stimulate purified placental glucosylceramidase activity towards the water-soluble substrate 4-methylumbelliferyI-B-D-glucopyranoside (MUGlc). Vesicles composed of pure phosphatidic acid (PA) or pure phosphatidylserine (PS) stimulated the activity of the enzyme about 20ofold. The inclusion of cholesterol and phosphatidylcholine (PC), beside PA, into the vesicles slightly improved their stimulatory effect. Further addition of oleic acid (OA) markedly increased the stimulation (50-fold). By ultracentrifugation and gel permeation procedures it was shown that, under optimal conditions for stimulation of the MUGIc hydrolysis by acidic phospholipid-containing vesicles, purified glucosylceramidase spontaneously binds to their surface. Interestingly, the molar fraction of the acidic phospholipid into the mixed vesicles, rather than its concentration in the assay, is the crucial parameter for activation and binding of the enzyme. The importance of glucosylceramidase association with appropriate vesicles for enzyme activation was indicated by observing that the presence of 0.2 M citrate-phosphate buffer (pH 5.5), that prevented the binding to PA.containing surfaces, also inhibited the enzyme activity. Our results indicate that the reconstitution of glucosylceramidase activity occurs through the spontaneous tight association of the enzymatic protein ~ith preformed acidic phospholipid-containing vesicles.

Introduction Glucosylceramidase (EC 3.2.1.45) is the enzyme which hydrolyzes glucosylceramide to ceramide and glucose [1,2]. The enzyme is tightly bound to the lysosomal membrane [3]. A profound deficiency of glucosylceramidase activity is the cause of Gaucher's disease, an autosomal recessive disorder which may present different clinical manifestations. At least three clinical subtypes have been defined: type 1 or 'adult', type 2 or "infantile' and type 3 or 'juvenile' forms. Type

Abbreviations: MUGIc, 4-methylumbelliferyl-/3-D-glucopyranoside; PC, phosphatidylcholine; PA, phosphatidic acid; PS, phosphatidylserine; OA, oleic acid; TC. sodium taurocholate. Correspondence: A.M. Vaccaro, Department of Metabolism and Pathological Biochemistry, Istituto Superiore di Sanit~, Viale Regina Elena 299, 00161 Roma, Italy.

1 is distinguished from type 2 and 3 by the absence of a neurological involvement [1,2,4]. Glucosylceramidase is able to hydrolyze not only its natural lipid substrate, glucosylceramide, but also artificial water-soluble compounds such as 4-methylumbelliferyl-fl-D-glucopyranoside (MUGIc) [5,6]. Delipidation and purification, that make the enzyme watersoluble, markedly impair the enzyme activity towards both lipid and water-soluble substrates; the activity can be restored by the addition of detergents or acidic iipids to the assay mixture [7-10]. When bile salts are used as activators no correlation has been found between the clinical manifestations of Gaucher's disease and the residual enzyme activity [11]. Instead the use of appropriate acidic lipids may help to distinguish neurological from non-neurological cases, as shown by the much better restoration of the residual glucosylceramidase activity from type I than from type 2 Gauchcr's patients on addition of phosphatidylserine (PS) or galactocerebroside 3-sulfate [9-11].

240 Acidic phospholipids are the most effective lipids in reconstituting the enzyme activity [10,12]. Since PS and phosphatidylinositol are the most abundant acidic phospholipids of the lysosomal membrane [13], they probably represent the physiological activators of glucosylceramidase [10,13]. Although t h e importance of acidic phospholipids in prom0ting the glucosylceramidase reconstitution is well established; a question still remains as to the exact molecular mechanism underlying the stimulation of the enzyme activity. The activation of the purified watersoluble enzyme by sonic dispersions of acidic phospholipids is most likely the result of hydrophobic interactions between the enzyme and the acidic phospholipid vesicles. These interactions might occur either by transfer of phospholipid molecules from the vesicles to the glucosylceramidase solution or by displacement of glucosylceramidase from the aqueous solution to the vesicle surface. In a previous report from this laboratory direct evidence was obtained that the latter is the case for the enzymatic hydrolysis of glucosylceramide, the lipid substrate, stimulated by acidic phospholipids [14]. Actually, purified glucosylceramidase spontaneously incorporates into preformed liposomes containing phosphatidic acid (PA) besides glucosyiceramide [14]. On the other hand it has been suggested that the PS activation of delipidated glucosylceramidase acting on the MUGIc substrate, was due to the formation of soluble, high molecular weight, enzyme-PS complexes [15-17], with the possible participation of a heat-stable activating factor [18], also called 'Gaucher's factor' or SAP-2, capable of promoting the transfer of PS to glucosylceramidase [15]. When either MUGlc or glucosylceramide are used as glucosylceramidase substrates, the two compounds are localized in different phases, the first being dissolved in the aqueous phase, while the second is present at the lipid surface together with the activating phospholipid. We thus considered the possibility that in the case of the water-soluble substrate MUGIc the stimulation of hydolysis occurs by a mechanism different from the one found for the lipid substrate [14]. The objective of the present paper was to define how acidic phospholipids reconstitute glucosylceramidase activity towards the soluble MUGIc substrate. In particular the studies reported in this paper were aimed at elucidating which type of physical interaction occurred between the enzyme and the acidic pnospholipidcontaining vesicles used as activators. Materials and Methods

Materials Oleic acid (OA), sodium taurocholate (TC) (synthetic, > 98% pure) cholesterol, 3-sn-phosphatidylcholine (PC) from egg yolk and 1,2-dipalmitoyl-3-sn-PA

were from Sigma (St. Louis, MO, U.S.A.). PS from bovine spinal cord and Triton X-100 were from Calbiochem (San Diego, CA, U.S.A.). MUGIc was obtained from Koch-Light Labs. (Colnbrook, U.K.). [414C]Cholesterol (50 mCi/mmoi), 1,2-dipalmitoyl[2palmitoyl-l-14C]-PC (90 mCi/mmol) and 1,2-dipalmitoyl[glycerol-'4C(U)]-PA (120 mCi/mmol) were from DuPont de Nemours (Germany), Biotechnology Systems Division, New England Nuclear Research Products. 1,2-Dioleoyl-3-sn-phosphatidyl-L[3-14C]serine was from Amersham International (U.K.). The radiopurity of labelled compounds, determined by thin-layer chromatography followed by counting of radioactivity, exceeded 98%. The other chemicals were of the purest available grade.

Enzyme preparation Glucosylceramidase was purified from human placenta as in our previous paper for preparation II [19]. The specific activity of the en~'me, measured according to the standard assay, was about 1 • 10~ n m o i / h per mg.

Vesicle preparation Vesicles were essentially prepared as previously described [14]. Appropriate aliquots of lipids and OL-Ottocopherol acetate (40 nmol) were mixed and the organic solvent evaporated under a stream of nitrogen. Vesicles were composed either of a single phospholipid or of a lipid mixture. In the first case, 1000 nmol of phospholipid supplemented with about 70000 dpra of the same radiolabelled compound as tracer, were suspended in 300/.tl of water. In the second case, the lipid mixture dispersed in the same volume of water consisted of cholesterol (600 nmol supplemented with about 70000 dpm of the radiolabelled analog), phospholipids (PC + PA, total phospholipids 1940 nmol) and OA (1640 nmol). This preparation was called vesicles A. When OA was not included in the mixture, the preparation was called vesicles B. The molar percentage of PA into vesicles A and B was varied as indicated in the experiments, keeping the molar percentage of total phospholipids constant (PA + PC, 46 mol% into vesicles A and 76 mol% into vesicles B) by appropriate variations of PC. The lipid suspensions were submitted to sonication under nitrogen in a Branson B 15 Sonifier (3 min with a Cup Horn at a power setting of 100 W, followed by 6 min with a Microtip at a power setting of 30 W). The sonication temperature was about 20°C. The lipid dispersion was centrifuged at 100 000 × g for 30 min and the resulting supernatant was left at room temperature for about 20 h before it was used for the experiments. More than 90% of the radioactivity present in the dried lipids was found in the final preparation. The concentration of vesicles was determined by measuring their radioactivity.

241

Glucosylceramidase assay The standard assay mixture contained, in a final vol. of 0.2 ml: 0.2 M sodium acetate buffer (pH 5.5), 2.5 mM MUGlc, 5-10 U of purified enzyme, 0.1% (v/v) Triton X-100 and 0.25% ( w / v ) tauroch~late. When the activating effect of phospholipids was checked the detergents were not added to the assays. The mixtures were incubated for 30 min at 37°C. The extent of reaction was estimated fluorometrically as previously described [19]. All assays were carried out in duplicate and generally agreed within 5%. 1 U of glueosylceramidase is defined as the amount of enzyme which hydrolyzes 1 nmol of M U G l c / h under the standard assay conditions.

Physical interaction between t'esicles and glucosylceramidase analyzed by a flotation procedare Ultracentrifugation was performed as previously described [14] with a few modifications. An aliquot of vesicles (150 #1 of the preparations described above) and of glucosylceramidase (800-1500 U) preparations were diluted to 0.5 ml, either separately or after mixing with each other. The final samples contained, beside vesicles a n d / o r glucosyiceramidase, 0.2 M acetate buffer (pH 5.5) supplemented with 5 mM EDTA and 5 mM dithioerythritol (buffer A), 35% w / v sucrose and 500 # g of trypsin inhibitor, added as stabilizer of the enzyme. After 15 rain incubation at 37°C, the samples (0.5 ml) were placed into an Ultra-clearTM centrifuge tube (13 x 51 ram) over a cushion of 1 ml of 60% w / v sucrose. 2.3 ml of 20% w / v sucrose were carefully layered over, followed by 1 ml of buffer A. All sucrose solutions were made up in buffer A. Centrifugation was performed in a Beckman SW 55 Ti rotor for 1 h at 300 000 × g at 25°C. Fractions of 0.2 ml were pumped out from the bottom of the centrifuge tubes. To determine the enzyme content of the fractions, aliquots were added to the standard assay mixture and tested as described above; the presence of vesicles was determined by measuring the radioactivity of the fractions.

Physical interaction between resicles and glucosylceramidase analyzed by chromatography on a Bio-Gel A-5m column The samples containing vesicles a n d / o r glucosylceramidase were prepared exactly as that for ultracentrifugation experiments except that sucrose was 20%. The samples were loaded on a Bio-Gel A-5m column (10 x 210 ram; Bio-Rad, Richmond, U.S.A.), equilibrated with buffer A containing 20% sucrose. After elution with 24 ml of this buffer, 2% (w/v) TC was added to the eluent. The flow rate was 0.25 ml/min. Fractions of 0.8 mi were collected. In some specified experiments (see figures) the acetate buffer was substituted either with citrate/phosphate 0.2 M in citrate

(pH 5.5) or with 0.1 M Tris-HCl (pH 7.0). Immediately after elution the fractions were tested for enzyme activity according to the standard assay. A new column was used for every experiment. The void volume and the total volume of the column (7-8 ml and 20-21 ml, respectively) were determined using very low density iipoprotein and cytochrome c as markers. Results

Stimulation by acidic phospholipids of glucosylceramidase actil'ity towards MUGlc In order to evaluate the efficiency of different types of PA-containing vesicles in the activation of glucosylceramidase, we have measured the enzymatic hydrolysis of MUGIc in the presence of vesicles prepared from pure PA and of vesicles containing PA as one of their lipid components. Since we had previously found that OA synergistically increases the efficiency of PA as activator of the glucosylceramide hydrolysis [14], vesicles containing a combination of PA and OA, besides cholesterol and PC, were also tested (vesicles A, see Materials and Methods). Incubations were performed in acetate buffer (pH 5.5), known to bring about the highest stimulation of glucosylceramidase activity towards MUGIc in the presence of acidic phospholipids [20]. Fig. 1 shows that vesicles A are the more effective activators; vesicles devoid of OA (vesicles B) or composed of pure PA gave 50 and 35%, respectively, of the

~> - - ~ , o - - - - ' - - ' - - - ~

"

/

15-

10" E

.

.

.

.

.

.

.

.

.

W' nmolPA/ml Fig. 1. Stimulation of glucosylceramidase activity by different PAcontaining vesicles. The enzyme activity was measured in the absence of detergents (see Materials and Methods). Each assay contained, in a final vol. of 0.2 ml, 8 U of glucosylceramidase and different activating vesicles. The lipid additions were the following: a constant amount of vesicles A (total lipids 333 nmol/assay) containing different percentages of PA (0-24%) (o o); different amounts of vesicles A containing 12 mot% PA ( o - - - - - - o); a constant amount of vesicles B (total lipids 200 nmol/assay) containing different percentages of PA (0-40%) (11 li ); different amounts of vesicles B containing 20 tool% PA (D - - - - - - t3): different amounts of vesicles composed of pure PA ( • • ). When the percentage of PA was varied, the percentage of total phospholipids into vesicles A or B was kept constant by appropriate variations of PC (see Materials and Methods).

242 TABLE !

Stimulation of glucosylceramidaseactirio' by t'arious amounts of i'esich's cmltainb~gdifferentpercentagesof PA The enzymeactivitywas measured in the absence of detergents as reported in the Materialsand Methods.Each assaycontained8 U of glucosylceramidase and varying amounts of vesicles A or B with different mole fractionsof PA as indicated.The percentage of total phosphollpidsinthe vesicleswas kept constant by appropriate varialions of PC (see Materialsand Methods) % Pa into the vesicles VesiclesA 12% 2% 0.5% Vesicles B 20% 5% 2%

nmol PA/ ml of assay

Enzymeactivity (nmol/h)

35 15 35 70 15 35

16 10.5 12.2 12.4 4.5 4.2

50 25 50 I00 25 50

8.2 4.(1 4.6 4.6 2.1 2.0

activity observed with vesicles A. The same stimulation was obtained with either pure PS or pure PA vesicles (data not shown). Using vesicles A or B as activators, the dependence of enzyme activity on PA concentration in the assay mixture was different according to whether increasing amounts of vesicles with a constant and optimal percentage of PA (12 mol% into vesicles A and 20 mol% into vesicles B), or a constant amount of vesicles with increasing percentages of PA (from 0 to 24 mol% into vesicles A and from 0 to 40 mol% into vesicles B) were added (Fig. 1). A similar result was obtained oy Sarmientos et al. [21] who found a different stimulation pattern when the MUGIc enzymatic hydrolysis was measured as a function of the molar percentage of PS present in unilamellar liposomes or of the amount of liposomes in the assay. To verify these observations different amounts of vesicles A or B containing optimal and suboptimal percentages of PA were added to the enzyme assay. As shown in Table I, the glucosylceramidase activity markedly changed when the percentage of PA into the vesicles was modified, while it was not significantly influenced by variations of the amount of vesicles in the assay mixture, at least in the concentration ranges examined. Consequently, the same amount of PA in the assay mixture can have quite a different effect on the enzymatic reaction, depending on the molar fraction of PA into the vesicles and on the type of vesicles employed as activators (Fig. 1 and Table I).

Physical interaction of glucosylceramidase with preformed phospholipid-containing vesicles: analysis by chromatography on a Bio.Gel A-5m column The physical interaction of purified placental glucosylceramidase with acidic phospholipid-containing vesicles was studied under the conditions for maximal stimulation of MUGIc hydrolysis. Separation of the vesicle-bound from the free enzyme was achieved by chromatography on a Bio-Gel A-5m column equilibrated and eluted with the same buffer used in the glucosylceramidase assay (0.2 M acetate, pH 5.5) (Fig. 2). We have previously reported that placental glucosylceramidase shows an anomalous behaviour on gelpermeation chromatography and that a strong hydrophobic interaction, which can be reversed by detergent addition, occurs between the purified enzyme and the matrices of two HPLC columns, namely Superose 6 and TSK-40-XL [22]. As shown in Fig. 2A, glucosylceramidase also binds to a Bio-Gel A-5m column and can be recovered (80-90% yeld) only on addition of taurocholate to the eluting buffer. Thus the Bio-Gel A-Sm column separates glucosylceramidase on the basis of hydrophobic interactions rather than molecular size. When glucosylceramidase was preincubated with pure PA vesicles, the enzyme was not retained by the column and instead coeluted with the vesicles (Fig. 2B). The same result was obtained on incubation of the enzyme with PS (Fig. 2C). On the contrary, when the experiment was performed with PC, the enzyme was retained by the column and could be recovered only by adding taurocholate to the buffer (Fig. 2D) as when it was chromatographed in the absence of phospholipids. These results indicate that PS and PA vesicles absorb glucosylceramidase on their surface thus preventing the hydrophobic interaction of the enzyme with the gel matrix, while PC vesicles are unable to interact with the enzyme. The association of glucosylceramidase with vesicles A and B was investigated next. Glucosylceramidase coeluted with vesicles A, when they contained 12 mol% PA (Fig. 2F), while the ability of this lipid mixture to interact with the enzyme was drastically decreased when the PA content into vesicles A was lowered to 0.5 mol% (Fig. 2E). A similar phenomenon occurred with vesicles B: more than 60% of glucosylceramidase coeluted with the peak of the vesicles when they contained 40 mol% PA, (Fig. 2H), while the interaction of the enzyme with the vesicles was not observed when the PA molar ratio was lowered to 2% (Fig. 2G).

Physical interaction of glucosylceramidase with preformed PA-comaining vesicles: analysis by ultracentrifugation To confirm the binding of glucosylceramidase to PA-containing vesicles, a different non perturbing method was sought to separate free from bound en-

243

L

150-

- 20 ii it II

100.lO

I'

ft i I

50-

-20

300200-

-10

100-

G

150-

"~ o ¢j

100-

1

Q.

-10

i=

-5

50.

f~

starting layer toward the top of the gradient (Fig. 3). Glucosylceramidase sedimented toward the bottom of the tube when centrifuged either alone or together with vesicles A or B containing low percentages of PA (Fig. 3A and D). When the PA content of the vesicles was increased, part (Fig. 3B, C and E) or all of the enzyme (Fig. 3F) floated to the upper part of the gradient. Differences in the distribution of floating glucosylceramidase were observed, after incubation with either vesicles A or B containing an optimal amount of PA. Actually, the enzyme distribution coincided with that of vesicles A (12 mol% PA), but not with that of vesicles B (20 or 40 mol% PA). This difference may be caused either by a lower affinity of the enzyme towards vesicles B than vesicles A or by a preferential binding of glucosylceramidase to vesicles B with higher density. Vesicles composed of either p u r e PA or PS could not be submitted to the ultraccntrifugation experi-

150" -10

oooo ooi :

10050-

lO

30

50

70

10

3o

5o

.....

70

Fraction number Fig. 2. Physical interaction of glucosylccramidase with vesicles of different composition as analyzed by chromatography on a Bio-Oel A-Sin column. The purified glucosyiceramidase (1000-1300 U) was loaded onto a Bio-Gel A-5m column after incubation in the absence (A) or in the presence of vesicles. (B) Vesicles of pure PA; (C) vesicles of pure PS; (D) vesicles of pure PC; (E) vesicles A containing 0.5 tool% PA; (F) vesicles A containing 12 tool% PA; (O) vesicles B containing 2 tool% PA; (H) vesicles B containing 40 tool% PA. The vesicles were supplemented with the analogous labelled phospholipid ( B - D ) or with labelled PA ( E - H ) as tracer. When labelled cholesterol instead of labelled PA was used the same distribution of radioactivity was observed (E-H). The elution was performed as reported in Materials and Methods. The arrows indicate when taurocholate was added to the elution buffer. The vesicles ( o - - o) and the glucosyiceramidase (e e) distributions were determined by measuring the radioactivity and the enzyme activity in the fractions.

",:;:) " 150-

'B

~

-10

E 9,

q4,

100.

?

f -5

E ~

E

5o.

150-

,

',

o. "o

°

c

1

5'°

100-

50-

2/-

,

,

20

16

12

,

.--..

8

&

24

20

16

12

8

4

Fracfion number

zyme. Flotation experiments in discontinuous density gradients, previously utilized by us to investigate the interaction of glucosylceramidase with lipid surfaces [14] were performed. The general protocol involved the incubation of vesicles A or B with purified glucosylceramidase in the acetate buffer used for the enzymatic assay. The vesicle-enzyme mixture was suspended in a dense-buffered medium and placed in a centrifuge tube below a layer of less dense medium. After a short ultracentrifugation, vesicles A and B, floated out of the

Fig. 3. Physical interaction of glucosylccramidase with vesicles of different composition as analyzed by ultracentrifugation. The purified glucosylccramidasc (600-800 U) was ultracentrifuged after incubation with buffer containing vesicles of different composition (see Materials and Methods). Vesicles B containing either 2 tool% PA (A) or 20 tool% PA (B) or 40 tool% PA (C). Vesicles A containing either 0.5% PA (D) or 4 tool% PA (E) or 12 tool% PA (F). The vesicles were supplemented with labelled PA as tracer. The arrows indicate where the samples were originally layered. The glucosylccramidase (e e) and the vesicles ( o - - - - - - o ) distributions were determined as in Fig. 2. When glucosylceramidase was incubated and centrifuged in the absence of vesicles the enzyme distribution was identical to that reported in A.

244 ments since they do not float under the reported experimental conditions.

200-

Effect of the buffer composition on the abilio' of PA-contahzing cesicles to stbnulate and bhtd ghwosylceramidase With both vesicles A or B as activators the optimal glucosylceramidase activity was observed in the 5.5-6.0 p H range. The substitution of the 0.2 M acetate buffer with citrate phosphate buffer 0.03 M in citrate did not affect the stimulating efficiency of the two types of iiposomes. Table I! shows the effect of the buffer molarity on the activation of glucosylceramidase by PA-containing liposomes at pH 5.5. When the molarity of the acetate buffer-was doubled the enzyme activity was unchanged, while both vesicles A and B lost their activating capacity when the citrate-phosphate buffer molarity was increased. A negative effect of high concentrations of citrate and phosphate ions on the glucosylceramidase activation by a PS sonic dispersion has been previously reported [20]. in order to investigate the cause of the enzyme inhibition by these ions, the interaction between glucosylceramidase and either vesicles A (PA 12 mol%) or B (PA 40 mol%) was analyzed by chromatography on Bio-Gel A-5m in 0.2 M citrate-phosphate buffer (pH 5.5) (Fig. 4A, B and C). As opposed to what occurred in acetate buffer, less than 5% of the enzyme coeluted with the peak of the vesicles. Both in the presence and in the absence of vesicles, most of the enzyme activity was eluted from the Bio-Gel A-5m column after the addition of taurocholate to the buffer (Fig. 4A, B and C). Thus the 0.2 M citrate-phosphate buffer prevents the binding of the enzyme to both vesicles A and B. The lack of interac-

A

1

B

(

o

1

E

1

100-

::b

-20 ,

200-

-lO ,I

1oo-

5

?

~J ::3

20

C 9

F

200-

" 100-

1 t0

1

-5

,i

10

E

30

~

70

10

3o

50

70

Fraction number

Fig, 4. Effect of the buffer composition on the physical interaction between glucosylceramidase and vesicles A and B. The purified glucosylceramidase (10(}0-1300 U ) w a s loaded on a Bio-Gel A-5m column equilibrated and eluted with either citrate-phosphate buffer 0.2 M in citrate (pH 5,5) (A, B, C) or with 0.1 M Tris-HCI buffer (pH 7) (D. E, F). The enzyme was previously incubated with the same buffer of the column either in the absence (A, D) or in the presence of vesicles. (B. E) Vesicles A containing 12 m o l ~ PA; (C. F)vesicles B containing 40 mol'~[ PA. The vesicles were supplemented with labelled PA. The elulkm was performed as in Fig. 2. The arrows indicate when taurocholate was added to the elution buffer. The glucosylceramidase (e e) and the vesicles ( o - - - - - - o ) distributions were determined as in Fig. 2.

TABLE II

lnfluem'e of buffer compositmn on the stimulation of gltwosylceramidase aeticity b.v cesicles A and B The enzyme activity was measured in the absence of detergents using the indicated buffers at pH 5.5. Each assay contained 8 U of glucosylceramidase and either vesicles A (12 mol% PA, total lipids 333 nmol) or vesicles B (20 mol% PA, total lipids 200 nmol). The percentage of enzyme activity is relative to that measured in 0.2 M acetate buffer, using vesicles B as activator. Buffer

Enzyme a c t i v i t y ( Jt ) vesicle A

vesicle B

Acetate 0.2 M 0.4 M

20O 200

100 100

Citrate/Phosphate 0.03 M * 0.05 M * 0.10 M * 0.20 M *

200 140 45 15

100 65 26 10

* Final molarity of the citrate ions in the buffer.

tion between enzyme and vesicles was observed also when the 0.2 M citrate-phosphate buffer (pH 5.5) was used in the flotation-ultracentrifugation experiments (data not shown). At pH 7 (Tris-HC! buffer) the enzyme is inactive towards the MUGIc substrate both in the presence and in the absence of PA-containing vesicles. In order to investigate the interaction between enzyme and vesicles A and B at this pH, their mixture was incubated and analyzed by gel chromatography on the Bio-Gel A-5m column in Tris-HCl buffer (Fig. 4D, E and F). In the absence of vesicles the enzyme was mostly retained by the column, as occurred in acetate buffer at pH 5.5. After incubation with either vesicles A or B, the enzyme eluted with the vesicles, indicating that also at pH 7 the enzyme is bound to the PA-containing lipid surface.

245

Discussion It is well known that the activity of purified glucosylceramidase towards either the lipid substrate glucosylceramide or the water-soluble substrate MUGIc depends on the presence of activators such as detergents or acidic lipids [5-10]. In a previous paper we have shown that the stimulation of glucosylceramide hydrolysis by acidic phospholipids was due to their ability to promote the binding of glucosylceramidase to glucosylceramide and PA containing liposomes [14]. The present study reports results obtained by two independent procedures (ultracentrifugation in discontinuous density gradient and chromatography on a Bio-Gel A-5m column) which indicate that the acidic phospholipidactivated hydrolysis of the water-soluble substrate MUGIc is related to the association of glucosyleeramidase with the acidic phospholipid-containing vesicles. The enzyme activity was more efficiently stimulated when an acidic phospholipid such as PA was inserted into a mixed lipid system (vesicles A and B, see Materials and Methods) instead of being present in isolation. OA, previously shown to synergistically potentiate the PA stimulation of glucosylceramide hydrolysis [14], exhibits a similar effect on the MUGlc hydrolysis, Interestingly, the glucosylceramidase binding to the lipid surface and the stimulation of the enzyme activity depend on the relative amount of PA into the vesicles, rather than on the phospholipid concentration in the assay mixture. In other words, the enzyme requires a minimal acidic phospholipid environment, in quantitative terms, in order to express its activity. it has been often reported that glucosylceramidase activity can vary tremendously with small changes in the reaction conditions [10,20,23]. Gonzales et al. [20] reported that glucosylceramidase was markedly activated by PS in sodium acetate buffer (pH 5.5), but completely inhibited in the presence of high concentrations of citrate a n d / o r phosphate ions. We too have observed that the enzyme activation by vesicles A and B is dramatically impaired in 0.2 M citrate-phosphate buffer and in particular we have shown that the binding of the enzyme to acidic phospholipid-containing vesicles is prevented by this buffer. Most likely every condition preventing this binding, inhibits the activating effect of acidic phospholipids. The activation of glucosylceramidase on binding to acidic-phospholipid containing vesicles most probably results from the sum of several factors. In a previous study concerned with the stimulation of the glucosylceramide enzymatic hydrolysis we explained the positive effect of PA with a promotion of the physical contact between the purified, water soluble enzyme and its lipid, water insoluble substrate incorporated into PA-containing liposomes [14]. The fact that the hydrolysis ef the water soluble substrate MUGIc re-

quires also the binding of the enzyme to the PA-containing vesicles indicates that the activation is not exclusively related to the promotion of physical contact with the substrate. Most likely, on binding to acidic phospholipid-containing vesicles, glucosylccramidase undergoes a conformationai change essential for its catalytic efficiency. Our results lead to an interpretation of the mechanism of glucosylceramidase reconstitution by acidic phospholipids different from that formulated by other authors. In fact, it had been reported that the activation of the enzyme by PS is accompanied by an increase in the size of the enzyme, from a 56 kDa form to an activated, 200 kDa form, a transformation greatly favoured by the presence of a heat-stable protein factor [15,16]. Although the composition and stoichiometry of the 200 kDa form had not been well defined, the authors assumed the formation of a soluble complex composed of one or more molecules of glucosylceramidase and several molecules of phosphatidylserine, possibly also containing the heat-stable factor [15]. Recently an other group has claimed that PS is bound to glucosylceramidase in the soluble complex at a molar ratio of 4.2 to 1 [17]. The formation of a high M r activated complex was supported by an observed increase of the enzyme sedimentation rate in sucrose gradient centrifugation after incubation of glucosylceramidase with a sonic dispersion of PS [15,16]. Moreover, it was found that PS coated on Celite was partly transferred into an aqueous solution after incubation with the enzyme [17]. There are several alternatives in the interpretation of the above data. In our opinion, they could as well result from the binding of glucosylceramidase to PS vesicles. A review of the literature provides many examples of the spontaneous assembly of proteins into preformed lipid bilayers [24], e.g phospholipase A 2 is activated several thousand-fold when the enzyme binds to a lipid-water interface [25], cytochrome oxidase is selectively incorporated into PS-containing iiposomes [26]. According to our results glucosylceramidase might also be listed among the proteins reconstituted on binding to appropriate vesicles.

Acknowledgments This work was partly supported by CNR (Progetto Finalizzato lngegneria Genetica). The authors would like to thank Mr. E. Raia for technical assistance.

References 1 Brady, R.O., Kanfer, J.N. and Shapiro. D.(1965) J. Biol. Chem. 240, 39-43. 2 Patrick, A.D. (1965) Biochem. J. 97, 17c-18c.

246 3 Van Dongen, J.M., Willemsen, R., Ginns, E.I., Sips, H.J., Tager, J.M, Barranger, J.A. and Reuser. A.J.J. (1985) Eur. J. Cell Biol. 39, 179-189. 4 Barranger, J.A. and Ginns, E.I. (1989) in The Metabolic Basis of Inherited Disease (Scriver, C.R., Beaudet, A.L, Sly, W.S. and Valle D., eds.), pp. 1677-1698, McGraw-Hill, New York. 5 Daniels, L.B., Glew, R.H., Diven, W.Y., Lee, R.E. and Radin, N:S: (1981) Clin. Chim. Acta 115, 369-375. 6 Raghavan, S.S., Topoi, J. and Kolodny, E.H. (1980) Am. J. Hum. Genet. 32, 158-173. 7 Strasberg, P.M. and Lowden, J.A. (1982) Clin. Chim. Acta 118, 9-20. 8 Wenger, D.A. and Roth, S. (1982) in Gaucher Disease: a Century of Delineation and Research (Desnick, R.J., Gatt, S. and Grabowski, G.A., eds.), pp. 551-572, Alan R. Liss, New York. 9 Basu, A., Glew, R.H., Daniels, L.B., Clark, L.S. (1984) J. Biol. Chem. 259, 1714-1719. 10 Glew, R.H., Basu, A., La Marco, K. and Prence, E. (1988) Lab. Invest. 58, 5-25. 11 Glew, R.H., Daniels, L.B., Clark, L.S. and Hoyer, S.W. (1982) J. Neuropathol. Exp. Neurol. 41,630-641. 12 Dale, G.L., Villacorte, D.G. and Beutler, G. (1976) Biochem. Biophys. Res. Commun. 71. 1048-1053. 13 Basu, A., Glew, R.H., Wherrett, J.R. and Hulerer. S. (1986) Arch. Biochem. Biophys. 245. 464-469. 14 Vaccaro, A.M., Tatti, M., Salvioli, R., Ciaffoni, F. and Gallozzi, E. (1990) Biochim. Biophys. Acta 1033, 73-79. 15 Prence, E.M., Chakravorti, S., Basu, A., Clark, L.S., Glew, R.H.

and Chambers. J.A. (1985) Archiv. Biochem. Biophys. 236, 98109. 16 Garrett, K.O., Prence, E.M. and Glew, R.H. (1985) Archly. Biochem. Biophys. 238, 344-352. 17 Morimoto, S., Kishimoto, Y., Tomich, J., Weiler, S., Ohashi, T., Barranger, J.A., Kretz, K.A. and O'Brien, J.S. (199{I) J. BioI.Chem., 265, 1933-1937. 18 Ho, M.W. and O'Brien, J.S. (1971) Proc. Natl. Acad. Sci. USA, 68, 2810-2813. 19 Vaccaro, A.M., Muscillo, M., Galiozzi, E., Salvioli, R., Tatti, M. and Suzuki K. (1985) Biochim. Biophys. Acta 836, 157-166. 20 Gonzales, M.L., Basu, A., de Haas, G.H., Dijkman, R., van Oort, M.G., Okolo A.A. and Glew, R.H. (1988) Archly. Biochem. Biophys. 262, 345-353. 21 Sarmientos, F., Schwarzmann, G. and Sandhoff, K. (1986) Eur. J. Biochem. 160, 527-535. 22 Vaccaro, A.M., Salvioli, R., Gallozzi, E., Ciaffoni, F. and Tatti, M. (1990) Biochim. Biophys. Acta 1047, 57-62. 23 Aerts, J.M.F.G, Sa Miranda, M.C., Brouwer-Kelder, E.M., Van Weely, S., Barranger, J.A. and Tagcr, J.M. (1990) Biochim. Biophys. Acta 10,,i, 55-63. 24 Jain, M.K. and Zakim, D. 0987) Biochim. Biophys. Acta 906, 33-68. 25 Slotboom, A.J., Verheij, H.M. and DeHaas, G.H. (1982) in Phospholipids, Vol 4 (Hawthorne, J.N. and Anseil, G.B., eds.), pp. 359-434, Elsevier Biomedical Press, Amsterdam. 26 Eytan, G.D. and Racker, E. 0977) J. Biol. Chem. 252, 3208-3213.

Reconstitution of glucosylceramidase on binding to acidic phospholipid-containing vesicles.

Studies were conducted to investigate the mechanism by which acidic phospholipid-containing vesicles stimulate purified placental glucosylceramidase a...
726KB Sizes 0 Downloads 0 Views