Biochimica et Biophysics Acta, 1047 (1990) 57-62 Elsevier

57

BBALIP 53507

Effect of experimental conditions on the appearance of distinct forms of placental glucosylceramidase: use of gel filtration

analysis as a means of ascertaining their occurrence *

Anna Maria Vaccaro, Rosa Salvioli, Elisabetta Gallozzi, Fiorella Ciaffoni and Massimo Tatti

Department of Metabolism and Pathological Biochemistq, Isiituto Superiore Saniia’, Roma (Italy)

(Received 29 June 1990)

Key words: Glucosylceramidase; Gaucher disease; Placental activating factor

We have found that, under some experimental conditions, the placental glucosylceramidase shows an anomalous behaviour on gel filtration chromatography. At pH 5.6, the optimal pH of the enzymatic assay, the purified enzyme remains bound to either Superose 6 or TSK40-XL HPLC columns, while the interaction of the crude glucosylcerami- dase contained in the water extract of the lysosome-mitochondrial fraction of placenta with the two HPLC gel matrices is much weaker. The quite different behaviour of the crude compared to the purified enzyme may be explained by the formation in the crude preparation of associated form(s) of glucosylceramidase with suitable endogenous compound (s), which compete with the gel matrices for the binding to the enzyme. The most likely one component of the enzyme complex is the placental activating factor, previously reported by us (Vaccaro et al. (1985) B&him. Biophys. Acta 836, 157-166), as indicated by the negligible stimulation of the crude enzyme activity on addition of the factor, either before or after passage through the HPLC columns. On the assumption that the behaviour of crude glucosylceramidase on gel filtration becomes similar to that of the purified enzyme when its interaction with endogenous substance(s) is impaired, we have identified some conditions which prevent the formation of the enzyme associated form(s): (a) the addition of guanidine chloride (0.2 M), a cahotropic agent, to the crude preparation; and (b) the increase of pH up to 8. In conclusion, taking advantage of the anomalous behaviour of glucosylceramidase on gel filtration chromatography, evidence has been obtained that placental glucosylceramidase may occur under several forms which had not been previously reported; a difference in experimental conditions can promote the formation of one or another form, by possibly affecting the composition and/or the stoichiometry and/or the stability of the enzyme complex.

Introduction

Glucosylceramidase (EC, 3.2.1.45) is the lysosomal enzyme deputed to the hydrolysis of glucosylceramide. A deficiency of this enzymatic activity is responsible for

Gaucher’s disease [1,2]. Glucosylceramidase is a membrane-bound enzyme

and direct evidence of its association with the lysosomal

membrane has been obtained [3]. A considerable amount of enzyme may be solubilized in detergent-free buffers [4,5], but its complete solubilization requires the pres-

ence of detergents.

* This work is dedicated to Professor G.B. Marini Bettolo on the

occasion of his 75th birthday.

Correspondence: A.M. Vaccaro, Department of Metabolism and

Pathological Biochemistry, Istituto Superiore di Sanita’, Viale Regina

Elena 299, 00161 Roma, Italy.

The subunit size of glucosylceramidase is approx. 60 kDa [6,7], but, according to several authors [g-11], the enzyme must be in the form of a larger active complex in order to express optimal activity. The components of

the active complexes reported up to now are, beside glucosylceramidase, acid lipids and/or the Gaucher fac- tor, a P-glucosidase activator protein, particularly abun- dant in Gaucher spleen [12]. An active enzyme complex consisting of two glucosylceramidase subunits plus enough lipid activator (ganglioside G,,) to yield a molecular weight of 178 kDa has been described [lo]. Furthermore, large active complexes of normal human spleen [8] or rat liver [ll] glucosylceramidase with phos- phatidylserine and the Gaucher factor have also been obtained. In addition, two immunologically dis- tinguishable forms of glucosylceramidase (I and II) have been recently identified in the water and detergent extracts of several human tissues [13,14].

0005-2760/90/%03.50 0 1990 Elsevier Science Publishers B.V. (Biomedical Division)

58

We have previously shown that glucosylcera~dase is present in the water extract of the lysosome-~tochon- drial fraction of placenta, together with an activator protein, the placental factor, which stimulates several- fold the enzymatic hydrolysis of the glucosylceramide [5]. When glucosylceramidase was separated from the

factor during pu~fication, the enzymatic ‘activity de-

creased dramatically, but could be completely restored by subsequent addition of the placental factor to the

purified enzyme [5]. Several features differentiate the

placental factor from the Gaucher factor, one of the most important differences being that the first is mainly active on the e~ymatic hydrolysis of the artificial sub-

strate, 4-methylumbelliferyl-~-glucoside, while the sec- ond stimulates only the hydrolysis of glucosylceramide

v51. To obtain informations on the interactions between

glucosylcera~dase, the placental factor and other com- pounds present in the water extract of the lysosome-

mitochondrial fraction of placenta, we have investigated in the present paper: (a) which forms of gluco-

sylceramidase occur in placenta extracts; (b) the in- fluence of the experimental conditions on the ap- pearance of one or another form of the enzyme and (c)

the possible presence of the placental activator in some of the glucosylceramidase forms.

In addition, we have shown that glucosylceramidase exhibits an anomalous behaviour on gel filtration chro-

matography; the possibility of taking advantage of this phenomenon for ascertai~ng the occurrence of distinct enzyme forms has been explored.

Materials and Methods

~ateriais Glucosylcera~de purified from spleen of patients

with Gaucher disease was labelled with tritium in the

glucose moiety according to McMaster and Radin [16]. The 4-MU-glycosides were obtained from Koch-Light Labs. (Colnbrook, U.K). Sodium taurocholate (syn- thetic, > 98% pure), oleic acid and trypsin inhibitor from soybean were from Sigma (St. Louis, MO). The other chemicals were of the purest available grade.

Glucosylceramidase preparations Purified preparation. Glucosylceramidase was puri-

fied from human placenta as in our previous paper for preparation II 151.

Crude preparation. The water extract of the lyso- some-mitochondrial fraction of human placenta was prepared as previously described [5].

Glucosylceramidase. The standard assay mixture contained, in a final volume of 0.2 ml, 40 pg [3H]gluco- sylceramide (spec. act. 3000 dpm/nmol), 0.25 mg

sodium taurocholate, 0.2 mg oleic acid, sodium citrate/ phosphate buffer (pH 5.6) 50 mM in citrate and the enzyme source. In some assays 1 pg of the placental activating factor [5] was also added. The mixture was

incubated 1 h at 37” C. The released glucose was esti- mated as described [17].

/3-Galactosidase, cY-galactosidase and ~-hexosamini- dase were assayed with 4-MU-glycosides as fluorogenic

substrates. The assays for the first two enzyme activities were carried out according to the standard procedures

of this laboratory [18]. A published procedure was fol- lowed for the assay of /3-hexosaminidase [17].

One enzyme unit is defined as the amount of enzyme which hydrolyzes 1 nmol of substrate per hour, under

the standard assay conditions.

High performance liquid chromatography (HPLC) gel titration

HPLC gel filtration of glu~osyl~era~dase was per- formed either with a Superose 6 column (10 x 300 mm; Pharmacia, Sweden) or with a Bio-Gel TSK-40-XL col- umn (7.5 X 300 mm; Bio-Rad, Richmond, U.S.A.). The enzyme preparations, previously dialyzed against the equilibrating buffers, were applied to the columns through a 200 ~1 sample injection loop. The flow rate

was 0.3 ml/mm. Fractions of 0.4 ml were collected. Both columns were equilibrated and first eluted with pH 5.6 sodium citrate/phosphate buffer, 50 mM in citrate, 5 mM EDTA, 1 mM DTE, 10% ethylene glycol

(buffer A), followed by the same eluent containing 1% taurocholate. In some specified experiments {see fig- ures) the buffer had a different pH or contained 0.2 M

guanidine chloride. In the experiments performed at pH 7, buffer A was adjusted to the desired pH with dis- odium hydrogen phosphate 200 mM. In the experiments

at pH 8.0, the buffer was Tris-HCI 50 mM, containing 5 mM EDTA, 1 mM DTE and 10% ethylene glycol.

When the purified enzyme preparation was analyzed, the collection tubes contained 1 pg of the placental factor in order to minimize the loss of activity observed

in preliminary experiments performed without this ad- dition. The fractions were tested for enzyme activity

immediately after elution. The columns were calibrated with thyroglobulin,

bovine serum albumin, trypsin inhibitor from soybean and glycyl-tyrosine as molecular weight markers.

Protein determination Protein was measured according to the method of

Bradford [19], using bovine serum albumin as standard.

Results

Gel titration analysis of purified glucosylceramidase In order to detect differences between the molecular

weight of purified glucosylceramidase and that of possi-

59

C TC V

TC V

0.6

0.2

a E

2

.s

0.6 g

Fraction number

Fig. 1. HPLC gel filtration of purified and crude glucosylceramidase on Superose 6 and Bio-gel TSK-40-XL columns. Either the purified (A,B)

(1500-2CKMl U as measured in the presence of the placental activating factor) or the crude (C,D) (1.5-2.0 mg, 1000-1500 U) glucosylceramidase

preparation was loaded onto either a Superose 6 (A,C) or a TSK-40-XL (B,D) column. Each sample of the purified enzyme was supplemented with

1 mg of trypsin inhibitor. The eluent was buffer A. The void volume and the total volume of the Superose 6 column were 11 ml and 24 ml,

respectively, while those of the TSK-40-XL column were 7 and 15 ml. When indicated by the arrow, TC (l%, w/v) was added to the eluent. The

enzyme activity was tested according to the standard assay either in the presence (0) or in the absence (0) of the placental factor (see Materials and

Methods). The dotted line (- - -) indicates the protein distribution; it represents the elution of the trypsin inhibitor when purified glucosylcerami-

dase is analyzed.

ble associated forms of the enzyme present in the water extract of the lysosome-mitochondrial fraction of hu- man placenta, we have tried to determine the elution volume of purified placental glucosylceramidase from a gel filtration HPLC Superose 6 column. Unexpectedly, at pH 5.6, the pH of the standard enzymatic assay, no recovery of enzyme activity was obtained with a volume of eluent corresponding to the total volume of the column (Fig. 1A). The trypsin inhibitor, also added to

the sample as a chromatographic marker and as a stabilizer, was eluted as expected for a protein of 20 kDa. In the hypothesis that the enzyme might have been

adsorbed to the gel matrix, the column was subse- quently washed with the eluent supplemented with 1% taurocholate. This treatment eluted the glucosylcerami- dase activity in a sharp peak with about 90% recovery (Fig. 1A). It is thus evident that a strong interaction, which can be reversed by detergent addition, occurs between purified glucosylceramidase and the gel matrix.

As expected from our previous findings [5,15], the glucosylceramidase activity of the purified preparation, as well as that present in the eluted fractions could be

fully appreciated only when exogenous placental activating factor [5] was added to our standard assay (see Materials and Methods). The enzyme activity of the fractions was markedly reduced after 24 h storage at

4°C. Also when the experiment was repeated with an

other HPLC gel filtration column, namely a TSK-40-XL,

purified glucosylceramidase was retained by the matrix; again, the enzyme could be recovered on addition of

taurocholate to the eluent (Fig. 1B).

Gel filtration analysis of crude glucosylceramidase

The observed tendency of the enzyme to hydrophobi- tally interact with the gel filtration matrices offered the possibility to compare, under this new point of view, the properties displayed by different preparations of

60

placental glucosylceramidase. An aliquot of the crude preparation containing about the same enzyme units as the previously analyzed purified preparation was loaded either on the Superose 6 or on the TSK-40-XL column. At difference with the purified enzyme which, under the same conditions, was adsorbed to the gel matrices, the

crude enzyme was only retarded, eluting from both columns as a broad peak after the bulk of the proteins (Fig. 1C and 1D). Only a very low amount of gluco- sylceramidase activity was further eluted when the two

columns were washed with the taurocholate containing

buffer. Three other lysosomal enzymatic activities, namely P-hexosaminidase, ol-galactosidase and p-

galactosidase, assayed in the fractions obtained from

the two HPLC columns, were eluted together in a sharp

peak corresponding to an elution volume of 17 ml for Superose 6 and of 11 ml for TSK-40-XL (unpublished).

The markedly decreased interaction with the HPLC gel matrices of the crude compared to the purified enzyme indicates that distinct glucosylceramidase forms occur in the two preparations. Most likely the cause of

the different behaxiour is the presence in the crude preparation of some compound(s) associated with glu- cosylceramidase.

The activity of the crude glucosylceramidase was independent of exogenous placental activator, in accor-

dance with our previous data [5]. Also most of the

activity present in the main peak eluted from either the Superose 6 or the TSK40-XL column was poorly

stimulated by the addition of the activator (Fig. 1C and 1D). This observation indicates an association of the enzyme with the activator protein unimpaired by the

passage through the columns.

Effect of guanidine chloride on the gel filtration behaviour of glucosylceramidase

We have then investigated under which conditions the enzyme complex eluted from the two HPLC col-

umns at pH 5.6 could be dissociated. Guanidine chlo- ride (0.2 M) was added to both the crude and the purified glucosylceramidase preparation and the sam- ples were analyzed on the TSK-40-XL column, run with

buffer A containing also guanidine chloride. The pres- ence of the chaotropic agent decreased the interaction between the gel matrix and the purified enzyme, which,

although still markedly retarded, could be eluted without the help of taurocholate (Fig. 2A). Under the same conditions the crude enzyme exhibited an elution pro- file almost identical to that of the pure enzyme (Fig. 2B); after passage through the column most of the activity became dependent upon addition of the placen- tal activator. It thus appears that guanidine chloride can impair the interactions involved in the formation of the enzyme complex.

In order to investigate whether the chaotropic agent affected specifically the behaviour of glucosylcerarni-

A IOO-

50 -

di i i

Fraction number

Fig. 2. Effect of guanidine chloride on the behaviour of gluco-

sylceramidase on TSK-40-XL. Either the purified (A) (about 2000 U

supplemented with 1 mg of trypsin inhibitor) or the crude (B) (about

1000 U) enzyme preparation was added with 0.2 M guanidine chloride

and loaded onto a TSK-40-XL column. The eluent was buffer A,

containing also 0.2 M guanidine chloride. The glucosylceramidase

activity was tested according to the standard assay either in the

presence (0) or in the absence (0) of the placental factor. /3-Hexosa-

minidase (A), p-galactosidase and cY-galactosidase were eluted in the

same fractions; the elution profile of the last two enzymes was

omitted for clarity.

dase on the TSK40-XL column, the fractions were also analyzed for the occurrence of other lysosomal enzyme activities present in the crude placental preparation. The tested enzymes, namely /%hexosaminidase, (Y- galactosidase and /3-galactosidase behaved identically both in the presence (Fig. 2B) and in the absence of 0.2

M guanidine chloride.

Effect of pH increase on the gel filtration behaviour of glucosylceramidase

The possible influence of the pH increase on the interaction of glucosylcerarnidase with the gel matrices and with endogenous compounds was investigated by carrying out the gel filtration analysis of the purified and crude enzyme preparation at pH 7 and 8 (Fig. 3). Although at pH 7 and, more substantially, at pH 8 a small peak of activity was eluted from the TSK-40-XL column before the trypsin inhibitor, most of the puri- fied enzyme remained adsorbed to the gel (Fig. 3C and 3D) as at pH 5.6. Thus, the strong interaction between

61

0.6

Fig. 3. Effect of the pH increase on the behaviour of glucosylceramidase on TSK-40-XL. Either the crude (A,B) (1.5-2.0 mg, 1000-1500 U) or the purified (CD) (1500-2000 U supplemented with 1 tug of trypsin inhibitor) enzyme preparation was loaded onto a TSK-40-XL column. The ehrtion was performed exactly as in Fig. 1 except that the pH of the buffer was either 7.0 (A and C) or S.O.(B and D) as described in Materials and Methods. The enzyme activity was tested according to the standard assay either in the presence (O), or in the absence (0), of the placental factor.

The elution volume of /3-hexosaminidase at pH 7 is indicated by the arrow (HEX). The dotted line (- - -) indicates the protein distribution.

the matrix of the TSK40-XL column and the purified enzyme occurs also when the pH of the buffer is raised.

Discussion

Instead, the elution pattern of the crude enzyme We have found that, under several experimental con- preparation was dramatically affected by the pH in- ditions, the behaviour of glucosylceramidase on HPLC crease (Fig. 3A and 3B). Both at pH 7 and 8 only a gel filtration chromatography is more influenced by its small percentage of the enzyme was eluted together with hydrophobic and/or conformational features rather the bulk of proteins; this activity, as well as that of the than by its size. This phenomenon offered the possibil- loaded sample, was insensitive to the addition of the ity of using the gel filtration technique for ascertaining placental activator. The rern~~g enzyme eluted after the occurrence of distinct ~ucosyl~era~dase forms in the inclusion of taurocholate in the buffer and was different placental preparations. Actually, the purified stimulated several-fold by the activator. Thus, the crude enzyme was adsorbed on the gel matrices of the two enzyme preparation analyzed by the TSK-40-XL col- HPLC colums we have used, while the crude enzyme umn at pH 7 and 8 exhibits an elution pattern quite was only retarded. The comparison of the elution pat- different from that obtained at pH 5.6 (Fig. 1D) but terns of pure and crude glucosylcer~da~ indicates very similar to that of the purified preparation (Fig. 3C that at pH 5.6, the pH of the enzymatic assay, the crude and 3D). These findings indicate that the interaction enzyme is associated with some endogenous substances, between glucosylceramidase and the endogenous sub- which change the enzyme conformation and/or hydro- stances forming the complex are weakened by the in- phobicity. Since the activity of the enzyme associated crease of pH. The elution volume of /?-hexosaminidase, form is independent on addition of the activating factor assayed as a marker enzyme, is the same either at pH recently found in placenta [5,15], it is most likely that 5.6 or at pH 7 (Fig. 3A). the activator participates in the complex.

62

At least two conditions are able to convert the gluco- sylceramidase complex to a form which behaves on gel filtration columns as the purified enzyme and which,

after elution, is stimulated several-fold by the placental activator: (a) the addition of guanidine chloride, a well known chaotropic agent and (b) the pH increase up to 8

of the enzyme solution. We conclude that, under these conditions, the interaction between glucosylceramidase and endogenous compounds are markedly impaired.

Our findings show that placental glucosylceramidase can occur under different forms and that a change of

the experimental conditions may determine the ap- pearance of one form or another form by possibly affecting the composition and/or the stoichiometry

and/or the stability of the glucosylceramidase complex with suitable substance/s present in the lysosome-mito- chondrial fraction of placenta.

In crude preparations glucosylceramidase has been

previously reported to exist under multiple forms which result from aggregation of the enzyme with proteins

and/or lipids [7-11,141. The occurrence of two enzyme forms (I and II) in the extracts of several human tissues has been recently described by Aerts et al. [13,14]. Form I, a 60 kDa protein, is considered a dissociated mono-

meric form, whereas form II is a glucosylceramidase complex in association with the Gaucher factor [14,20]; in the placenta extract form I is the predominant one,

the contribution of form II to the total glucosylcerami- dase activity being less than 5% [14,20]. On the con- trary, we observe that in the water extract of the lyso-

some-mitochondrial fraction of placenta at pH 5.6 most of the glucosylceramidase is associated with endogenous

compounds, the placental activator included. It is possi- ble that our identification of new glucosylceramidase associated forms derive from a different experimental

approach. We have previously described the quite different

ability of crude and purified placental glucosylcerami- dase to bind to a glucosylceramide-oleic acid dispersion

at the pH of the enzymatic assay: the enzyme present in the water extract of the lysosome-mitochondrial frac-

tion of placenta was completely adsorbed on the disper- sion, while the pure enzyme interacted weakly with the lipid-water interface [21,22]. These results, together with

the evidence provided by the present paper that the crude and the pure placental glucosylceramidase occur at pH 5.6 under distinct forms, suggest that these forms have a different affinity for suitable lipid surfaces, the aggregated form present in the crude preparation being preferentially bound. It seems premature to envisage a physiological role for the different glucosylceramidase forms. Nevertheless, our previous [21,22] and present results favour the hypothesis that the formation of associated enzyme forms may promote the binding of glucosylceramidase to appropriate lipid surfaces and

consequently influences its localization on the mem- branes.

Acknowledgments

This work was partly supported by grants from Con- siglio Nazionale delle Ricerche (Progetto Finalizzato

Ingegneria Genetica, Sottoprogetto Malattie Ereditarie). The authors would like to thank Mr. E. Raia for techni- cal assistance.

References

1 Brady, R.O., Kanfer, J.N. and Shapiro, D. (1965) J. Biol. Chem.

240, 39-49.

2 Patrick, A.D. (1965) B&hem. 97, 17C-l8C.

3 Van Dongen, J.M., Willemsen, R., Ginns, EL, Sips, H.J., Tager,

J.M., Barranger, J.A. and Reuser, A.J.J. (1985) Eur. J. Cell Biol.

39, 179-189.

4 Aerts, J.M.F.G, Donker-Koopman, W.E., Koot, M., Murray, G.J.,

Barranger, J.A., Tager, J.M. and Schram, A.W. (1986) B&him.

Biophys. Acta 863, 63-70.

5 Vaccaro, A.M., Muscillo, M., Gallozzi, E., Salvioli, R., Tatti, M.

and Suzuki, K. (1985) B&him. Biophys. Acta 836, 157-166.

6 Ginns, E.I., Brady, R.O., Pirrucello, S., Moore, C., Sorrell, S.,

Furbish, F.S., Murray, G.J., Tager, J. and Barranger, J.A. (1982)

Proc. Natl. Acad. Sci. USA 79, 5607-5610.

7 Pentchev, P.G., Brady, R.O., Hibbert, S.R., Gal, A.E. and Shapiro,

D. (1973) J. Biol. Chem. 248, 5256-5261.

8 Garrett, K.O., Prence, E.M. and Glew, R.H. (1985) Arch. Bio-

them. Biophys. 238, 344352.

9 Glew, R.H., Basu, A., La Marco, K. and Prence, E. (1988) Lab.

10

11

12

13

14

15

16

17

18

19

20

21

22

Invest. 58, 5-25.

Basu, A. and Glew, R.H. (1985) J. Biol. Chem. 260, 13067-13073.

Prence, E.M., Chakravorti, S., Basu, A., Clark, L.S., Glew, R.H.

and Chambers, J.A. (1985) Arch. Biochem. Biophys. 236, 98-109.

Ho, M.W. and O’Brien, J.S. (1971) Proc. Natl. Acad. Sci. USA 68,

2810-2813.

Aerts, J.M.F.G., Donker-Koopman, W.E., Van Der Vliet, M.K.,

Jonsson, L.M.V., Ginns, E.I., Murray, G.J., Barranger, J.A., Tager,

J.M., Schram, A.W. (1985) Eur. J. B&hem. 150, 565-574.

Aerts, J.M.F.G., Donker-Koopman, W.E., van Laar, C., Brul, S.,

Murray, G.J., Wenger, D.A., Barranger, J.A., Tager, J.M., Schram,

A.W. (1987) Eur. J. B&hem. 163, 583-589.

Vaccaro, A.M., Muscillo, M., Gallozzi, E., Salvioli, R., Tatti, M.

and Suzuki, K. (1986) in Enzyme of Lipid Metabolism II (Freysz,

L., Dreyfus, H., Massarelh, R. and Gatt, S., eds.), pp. 371-381,

Plenum, New York.

McMaster, M.C., Jr. and Radin, N.S. (1977) J. Label. Comp.

Radiopharmaceut. 13, 353-357.

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

Vaccaro, A.M., Salvioli, R. and Muscillo, M. (1981) LAB, J. Res.

Lab. Med. 8, 465-469.

Bradford, M.M. (1976) Anal. B&hem. 72, 248-254. Aerts, J.M.F.G., Donker-Koopman, W.E., Koot, M., Brouwer-

Kelder, E.M., Murray, G.J., Barranger, J.A., Tager, J.M., Schram,

A.W. (1986) in Enzyme of Lipid Metabolism, II (Frey=, L.,

Dreyfus, H., Massarelli, R. and Gatt, S., eds.), pp. 747-752,

Plenum, New York.

Vaccaro, A.M., Muscillo, M., Salvioti, R., Tatti, M., Gallozzi, E.

and Suzuki, K. (1987) FEBS Lett. 216, 190-194.

Vaccaro, A.M., Muscillo, M., SalvioIi, R., Tatti, M. and Gallozzi,

E. (1988) in Lipid Storage Disorders (Salvayre, R., Douste-Blazy,

L. and Gatt, S., eds.), pp. 347-352, Plenum, New York.