Characterization of the phospholipid requirement of a rat liver beta-glucosidase

Further studies on the reconstitution of glucosylceramidase activity by Sap C and anionic phospholipids

The reconstitution of the activity of the lysosomal enzyme glucosylceramidase requires anionic phospholipids and, at least, a protein factor, saposin C (Sap C). We have previously proposed a mechanism for the glucosylceramidase activation [Vaccaro et al. (1993) FEBS Lett. 336, 159-162] which implies that Sap C promotes the association of the enzyme with anionic phospholipid-containing membranes, thus favoring the contact between the enzyme and its lipid substrate, glucosylceramide. We have further investigated the properties of Sap C using a fluorescent hydrophobic probe such as 4, 4'-dianilino-1,1'-binaphthyl-5,5'-disulfonic acid (bis-ANS). The binding between bis-ANS and Sap C was pH-dependent, indicating that protonation leads to increased exposure of hydrophobic surfaces of Sap C. The interaction of Sap C with membranes, triggered by the development of hydrophobic properties at low pH values, was affected by the content of anionic phospholipids, such as phosphatidylserine or phosphatidylinositol, suggesting that anionic phospholipids have the potential to modulate the insertion of Sap C in the hydrophobic environment of lysosomal membranes. We previously showed that Sap C and anionic phospholipids are both required for the binding of glucosylceramidase to large vesicles. We have presently observed that Sap C is able to promote the association of glucosylceramidase with the lipid surface only when anionic phospholipids exceed a concentration of 5-10%. This level can be reached by summing lower amounts of individual anionic phospholipids, since they have additive effects. The present data extend and refine our model of the mechanism of glucosylceramidase activation and stress the key role of pH, Sap C and anionic phospholipids in promoting the interaction of the enzyme with membranes.

Acid beta-glucosidase: intrinsic fluorescence and conformational changes induced by phospholipids and saposin C

Acid beta-glucosidase is a lysosomal membrane protein that cleaves the O-beta-D-glucosidic linkage of glucosylceramide and aryl-beta-glucosides. Full activity reconstitution of the pure enzyme requires phospholipids and saposin C, an 80 aa activator protein. The deficiency of the enzyme or activator leads to Gaucher disease. A conformational change of acid beta-glucosidase is shown to accompany activity reconstitution by selected phospholipids or, particularly, phospholipid/saposin C complexes by intrinsic fluorescence spectral shifts, fluorescence quenching, and circular dichroism (CD). Negatively charged phospholipid (NCP) interfaces with unsaturated fatty acid acyl chains (UFAC) induced concordant blue-shifts in tryptophanyl fluorescence spectra and a loss of beta-strand structure by CD. The enzyme required an unsaturated fatty acid acyl chain in proximity (10-11 A) within liposomal membranes for activation, fluorescence blue-shifts, and changes in CD spectra. Activity enhancements were greatest when UFAC and the negatively charged headgroup were present on the same phospholipid. NCPs with UFAC protected the enzyme from fluorescence quenching by aqueous agents (I-, Cs+, acrylamide, TEMPO). Phosphatidylcholine with doxyl spin-labeled fatty acid acyl chains at carbons 7, 10, or 16 quenched enzyme fluorescence only when in NCP/PC liposomes. Saposin C (Trp-free) induced additional activity and fluorescence spectral changes in the enzyme only in the presence of NCP liposomes containing UFA. CD spectral changes indicated saposin C and acid beta-glucosidase interaction only in the presence of NCPs with UFA. These studies show that acid beta-glucosidase requires interfaces composed of NCPs, containing UFAC, for penetration into the outer leaflet of membranes. Furthermore, this interaction induces essential conformational changes for saposin C binding and further enhancement of acid beta-glucosidase catalytic activity.

Identification of the binding and activating sites of the sphingolipid activator protein, saposin C, with glucocerebrosidase

Saposin C is a sphingolipid activator protein of 8.5 kDa that activates lysosomal glucocerebrosidase. Previously, we synthesized and characterized a synthetic full-length human saposin C protein that displays 85% of the activity of the native saposin C. In this study we use shorter synthetic peptides derived from the saposin C sequence to map binding and activation sites. By determining the activity and kinetic constant (Kact) values of these peptides, we have identified two functional domains, each comprising a binding site adjacent to or partially overlapping with an activation site. Domains 1 and 2 are located within amino acid positions 6-34 and 41-60, respectively. The activation sites span residues 27-34 and 41-49, whereas binding sites encompass residues 6-27 and 45-60. Peptides containing the sequences of either domain displayed 90% of the activity of the full-length synthetic saposin C. Domain 2, however, bound to glucocerebrosidase by at least an order of magnitude more strongly than domain 1. Binding sites within these domains contain sequences that are excellent candidates for forming amphipathic helical structures. Competition assays demonstrated that the binding of one domain to glucocerebrosidase prevents binding of the other domain, and that saposin A and saposin C bind to the same sites on glucocerebrosidase. A model predicting a saposin C:glucocerebrosidase complex with a stoichiometry of 4:2, respectively, is presented.

Human glucocerebrosidase catalyses transglucosylation between glucocerebroside and retinol

The basal activity of human placental glucocerebrosidase is elevated 16-fold by n-pentanol when assayed using p-nitrophenyl beta-D-glucopyranoside (pNPGlc) as the beta-glucosidase substrate. This enhancement of activity is the result of the formation of a transglucosylation product, n-pentyl beta-D-glucoside, in rate-determining competition with the hydrolytic reaction. The transglucosylation product accounts for approximately 80% of the reaction product generated in the presence of n-pentanol (0.18 M) when either glucocerebroside or pNPGlc was used as the substrate. This stimulatory effect can be increased an additional 3-fold by the inclusion of phosphatidylserine (20 micrograms/ml) or sodium taurodeoxycholate (0.3%, w/v) in the incubation medium. In the presence of retinol, glucocerebrosidase also catalyses the synthesis of a novel lipid glucoside, retinyl glucoside, when either glucocerebroside or pNPGlc serves as the substrate. The reaction product was identified as retinyl beta-D-glucoside, based on its susceptibility to hydrolysis by almond beta-D-glucosidase and the subsequent release of equimolar amounts of retinol and glucose. The rate of retinyl-beta-glucoside formation is dependent on the concentration of retinol in the incubation medium, reaching saturation at approximately 0.3 mM retinol. Retinyl beta-D-glucoside is a substrate for two broad-specificity mammalian beta-glucosidases, namely the cytosolic and membrane-associated beta-glucosidases of guinea pig liver. However, retinyl beta-D-glucoside is not hydrolysed by placental glucocerebrosidase. These data indicate that the glucocerebrosidase-catalysed transfer of glucose from glucocerebroside to natural endogenous lipid alcohols, followed by the action of a broad-specificity beta-glucosidase on the transglucosylation product, could provide mammals with an alternative pathway for the breakdown of glucocerebroside to glucose and ceramide.

Studies on glucosylceramidase binding to phosphatidylserine liposomes: the role of bilayer curvature

The influence of phosphatidylserine (PS) liposome size on their capacity to activate and bind purified glucosylceramidase was investigated. Gel filtration and flotation experiments showed that large unilamellar vesicles (LUV) of either pure PS or PS in admixture with phosphatidylcholine (PC) are unable to tightly bind purified glucosylceramidase, and thus, to fully stimulate its activity. By contrast, small unilamellar vesicles (SUV) of PS adsorb glucosylceramidase can either be favoured or inhibited by factors affecting the bilayer curvature of PS liposomes. An increase of PS vesicle size induced by a fusogenic agent such as poly(ethylene glycol) (PEG), decreased enzyme binding and activity. On the contrary, the reduction of PS LUV size by sonication increased their stimulating ability. Enzyme association with PS SUV is reversible. In fact, glucosylceramidase bound to PS SUV was released from the lipid surface when the SUV were transformed into larger vesicles by PEG; dissociation from the vesicles resulted in a dramatic decrease of enzyme activity. Although PS LUV are unable to reconstitute glucosylceramidase, their association with oleic acid (OA) promotes the interaction with glucosylceramidase. This phenomenon is best explained in terms of OA-induced surface defects of PS LUV, with consequent exposure of the more hydrophobic part of the membrane and hence the improved binding of hydrophobic region/s of glucosylceramidase. Our data indicate that the physical organization of the PS-containing liposomes is of critical importance of glucosylceramidase reconstitution. The observation that physical changes of the lipid surface can markedly affect the enzyme activity offers a new approach to the study of glucosylceramidase regulation.

Correlation between the activity of glucosylceramidase and its binding to glucosylceramide-containing liposomes. Role of acidic phospholipids and fatty acids

Optimal enzymatic hydrolysis of glucosylceramide inserted into liposomes has been obtained when both acidic phospholipids and the appropriate fatty acids were added to glucosylceramide-containing liposomes. In fact, the stimulation of glucosylceramidase by acidic phospholipids was synergistically enhanced by fatty acids, whose effect was dependent upon chain length and increased on unsaturation. By following the partition of glucosylceramidase between the aqueous phase and the liposome-associated state with a flotation procedure, it has been found that phosphatidic acid (PA) and oleic acid (OA), as representatives of acidic phospholipids and activating fatty acids, respectively, were both required not only for optimal glucosylceramidase activity, but also for a tight binding of the enzyme to the liposomes. The binding was significantly less effective in the absence of either PA or OA. In the absence of both PA and OA no physical interaction between the enzyme and the liposomes was observed. Under all conditions, the glucosylceramidase activity directly correlated with the enzyme binding to the substrate-containing liposomes. Additionally, we have obtained evidence that the site(s) of the enzyme involved in the binding to the liposomes is distinct from the catalytic site; in fact, the enzyme could still associate with liposomes containing PA and OA but devoid of glucosylceramide, while it was incapable of binding to glucosylceramide-containing liposomes in the absence of PA and OA. In conclusion, the presence in liposomes of acidic phospholipids together with the appropriate fatty acids plays a key role in promoting the binding of glucosylceramidase. Consequently, when glucosylceramide is also included in the liposomes, its hydrolysis is markedly enhanced by these acidic lipids.

Acid beta-glucosidase: enzymology and molecular biology of Gaucher disease

Human lysosomal beta-glucosidase (D-glucosyl-acylsphingosine glucohydrolase, EC 3.2.1.45) is a membrane-associated enzyme that cleaves the beta-glucosidic linkage of glucosylceramide (glucocerebroside), its natural substrate, as well as synthetic beta-glucosides. Experiments with cultured cells suggest that in vivo this glycoprotein requires interaction with negatively charged lipids and a small acidic protein, SAP-2, for optimal glucosylceramide hydrolytic rates. In vitro, detergents (Triton X-100 or bile acids) or negatively charged ganglioside or phospholipids and one of several "activator proteins" increase hydrolytic rate of lipid and water-soluble substrates. Using such in vitro assay systems and active site-directed covalent inhibitors, kinetic and structural properties of the active site have been elucidated. The defective activity of this enzyme leads to the variants of Gaucher disease, the most prevalent lysosomal storage disease. The nonneuronopathic (type 1) and neuronopathic (types 2 and 3) variants of this inherited (autosomal recessive) disease but panethnic, but type 1 is most prevalent in the Ashkenazi Jewish population. Several missense mutations, identified in the structural gene for lysosomal beta-glucosidase from Gaucher disease patients, are presumably casual to the specifically altered posttranslational oligosaccharide processing or stability of the enzyme as well as the altered in vitro kinetic properties of the residual enzyme from patient tissues.

The inhibition of glucosylceramide beta-glucosidase and other acid hydrolases by nucleic acids

During the isolation of the activator protein for glucosylceramide beta-glucosidase, we found that certain column fractions contained an inhibitor of the enzyme. After separation from the activator protein by a DEAE-Sephacel column, the inhibitor was purified further with a Spehadex G-75 column. The u.v. absorption spectrum of the purified material was similar to that of nucleic acids and the protein content of the purified material was negligible. Furthermore the purified inhibitor reacted with orcinol but not with diphenylamine, and its inhibitory activity was completely destroyed by treatment with RNAases. It seems likely that the purified inhibitor was tRNA. Authentic RNA, tRNA and DNA had similar inhibitory effects on beta-glucosidase (Ki 17 micrograms/ml for tRNA, noncompetitive toward the substrate). The inhibitory effect of nucleic acids was not fully overcome by an excess amount of the activator protein, but phosphatidylserine could restore the activity to normal. Tests with several other hydrolases revealed that the inhibitory effect of nucleic acids was fairly general.

Activation of human spleen glucocerebrosidases by monoacylglycol sulfates and diacylglycerol sulfates

The study of the acidic lipid requirement of human spleen glucocerebrosidase was extended to include two new series of acidic lipids, namely, monoacylglycol sulfates and diacylglycerol sulfates. Lysosomal glucocerebrosidase was extracted with sodium cholate and 1-butanol to render its beta-glucosidase activity dependent upon exogenous lipids. Maximum reactivation of control glucocerebrosidase was obtained with nonanoylglycol sulfate (NGS) and diheptanoylglycerol sulfate (DHGS). However, the effects of these lipids were markedly dependent on the nature of buffer used in the assay medium; specifically, 0.2 M sodium citrate-phosphate (pH 5.5) was much more effective than 0.2 M sodium acetate (pH 5.5) in permitting these lipids to reactivate glucocerebrosidase. In contrast, the marked activation of glucocerebrosidase by phosphatidylserine and galactocerebroside 3-sulfate (sulfatide) that was achievable in the sodium acetate buffer was totally inhibited by citrate or phosphate ions. The effects of NGS and DHGS on the kinetic parameters of control glucocerebrosidase were to lower the Km for the substrate, 4-methylumbelliferyl-beta-D-glucoside from 5.5 mM to approximately 2 mM (in sodium citrate-phosphate buffer) and markedly increase the Vmax. Furthermore, with DHGS, significant activation was achieved at concentrations below the lipid's critical micellar concentration. None of the monoacylglycol- or diacylglycerol sulfates were capable of stimulating mutant glucocerebrosidases from either type 1 (Ashkenazi-Jewish) or type 2 Gaucher's disease patients. Like control glucocerebrosidase, the type 1 glucocerebrosidase was unresponsive to phosphatidylserine and sulfatide when the beta-glucosidase assay was conducted in 0.2 M sodium citrate-phosphate buffer. Based on the differential action of these lipid activators in the two buffers and their effects on the mutant enzymes, we propose that, with regard to the lipid requirement of glucocerebrosidase, there are two classes of acidic lipids--one comprised of phosphatidylserine and sulfatide and the other comprised of the likes of NGS, DHGS, or sodium taurodeoxycholate. It appears that control glucocerebrosidase and the mutant enzyme of the patient with type 1 Gaucher's disease is reconstitutable with the first class of lipids whereas the glucocerebrosidase of the type 2 patient is not. The observations in this report are interpreted in terms of a model which postulates that normal glucocerebrosidase possesses at least two distinct lipid binding domains.

Relationship between the two immunologically distinguishable forms of glucocerebrosidase in tissue extracts

Extracts of human spleen contain two immunologically distinguishable forms of glucocerebrosidase: form I is precipitable by polyclonal or monoclonal anti-(placental glucocerebrosidase) antibodies, whereas form II is not [Aerts, J. M. F. G., Donker-Koopman, W. E., Van der Vliet, M. F. K., Jonsson, L. M. V., Ginns, E. I., Murray, G. J., Barranger, J. A., Tager, J. M. & Schram, A. W. (1985) Eur. J. Biochem. 150, 565-574]. The proportion of form II glucocerebrosidase was high in extracts of spleen, liver and kidney and low in extracts of brain, placenta and fibroblasts. Furthermore, the proportion of form II enzyme was higher in a detergent-free aqueous extract of spleen than in a Triton X-100 extract of total spleen or splenic membranes. When form II glucocerebrosidase in a splenic extract was separated from form I enzyme by immunoaffinity chromatography and stored at 4 degrees C, a gradual conversion to form I enzyme occurred. The conversion was almost immediate if 30% (v/v) ethylene glycol was present. In the denatured state both forms of glucocerebrosidase reacted with anti-(placental glucocerebrosidase) antibodies. Form I glucocerebrosidase was stimulated by sodium taurocholate or sphingolipid-activator protein 2 (SAP-2), whereas form II enzyme exhibited maximal activity in the absence of the effectors. The pH activity profile of form II glucocerebrosidase was almost identical to that of form I enzyme in the presence of SAP-2. In the native state, form I glucocerebrosidase had a molecular mass of 60 kDa whereas that of form II glucocerebrosidase was about 200 kDa. After gel-permeation high-performance liquid chromatography of splenic extracts, the fractions with form II glucocerebrosidase contained material cross-reacting with both anti-(placental glucocerebrosidase) and anti-(SAP-2) antibodies. Preincubation of form I glucocerebrosidase with SAP-2 at pH 4.5 led to masking of the epitope on glucocerebrosidase reacting with monoclonal anti-(placental glucocerebrosidase) antibody 2C7. Furthermore, preincubation of form I glucocerebrosidase with monoclonal antibody 2C7 prevented activation of the enzyme by SAP-2. We propose that form I glucocerebrosidase is a monomeric form of the enzyme, whereas form II glucocerebrosidase is a high-Mr complex of the enzyme in association with sphingolipid-activator protein 2.

A kinetic study of the effects of galactocerebroside 3-sulphate on human spleen glucocerebrosidase. Evidence for two activator-binding sites

Extraction of control human spleen glucocerebrosidase with sodium cholate and butan-l-ol reversibly inactivates the enzyme in terms of its ability to hydrolyse the water-soluble substrate 4-methylumbelliferyl beta-D-glucopyranoside (MUGlc). The acidic brain lipid galactocerebroside 3-sulphate (sulphatide) reconstitutes beta-glucosidase activity in a strongly concentration-dependent manner. In this study we show that sulphatide exhibits three critical micellar concentrations (CMCs): CMC1, 3.72 microM; CMC2, 22.6 microM; CMC3, 60.7 microM. We designate the aggregates formed at these CMCs as primary, secondary and tertiary micelles respectively. From the results of kinetic studies performed at various sulphatide concentrations (0.012-248 microM), we found that sulphatide monomers (less than 3 microM) decreased the Km (for MUGlc) of control glucocerebrosidase from 11 to 4.6 mM, and lowered the Vmax. 2-fold. However, secondary and tertiary micelles were required for expression of high control glucocerebrosidase activities. Glucocerebrosidase prepared from the spleen of a patient with non-neuronopathic type 1 Gaucher's disease exhibited a very low Km (2.8 mM) even in the absence of exogenous lipid, and sulphatide monomers had no effect on the mutant enzyme's Km or Vmax. However, secondary or tertiary micelles markedly increased the Vmax. of the type 1 glucocerebrosidase to 60% of the corresponding control enzyme value. In contrast, for the glucocerebrosidase of the neuronopathic type 2 case, although sulphatide decreased the Km from 9.2 to 1.7 mM, the Vmax. never reached more than 5% that of the control enzyme, even at high concentrations of sulphatide. In addition, we found that secondary and tertiary sulphatide micelles enhanced the rate of inactivation of all three glucocerebrosidase preparations by chymotrypsin. Collectively, these results indicate the presence of two sulphatide-binding sites on glucocerebrosidase: one that enhances substrate binding, and another that enhances catalysis.

Sulfogalactocerebroside and bis-(monoacylglyceryl)-phosphate as activators of spleen glucocerebrosidase

Sequential extraction of human spleen membranes with sodium cholate and n-butanol removes endogenous lipids and renders glucocerebrosidase activity dependent upon exogenous acidic lipids (e.g., phosphatidylserine, gangliosides) and a heat-stable activator protein (HSF). In the present report, we show that two previously untested lysosomal acidic lipids, namely sulfogalactocerebroside and bis-(monoacylglyceryl)-phosphate (BMP), also activate normal human glucocerebrosidase. In addition, sulfogalactocerebroside also markedly enhanced the activity of glucocerebrosidase isolated from a patient with type 1 (non-neuronopathic) Gaucher's disease, resulting in a specific activity which was 60-80% that of control glucocerebrosidase. Furthermore, when the sulfolipid was used as the activator, glucocerebrosidase from the type 1 patient was 30 times more active than the corresponding glucocerebrosidase from a person with type 2 (neuronopathic) Gaucher's disease. In contrast, the two BMPs, one rich in C26 saturated fatty acid and another rich in C18 unsaturated fatty acids, were relatively poor activators of both mutant glucocerebrosidases while providing excellent reconstitution of control activity.

Lipid-protein interactions in biomembranes studied through the phospholipid specificity of D-beta-hydroxybutyrate dehydrogenase

Since the biological membranes are fundamental units in the living cells, the studies of lipid-protein interactions are crucial for the understanding of their structure, functions and properties. Beside hydrophobic interactions between fatty acids chain of phospholipids and intrinsic membrane proteins, the interactions between charged groups of the protein with the polar heads of phospholipids generally confer the specificity which may be absolute or preferential. This paper reports essential results obtained these last few years with D-beta-hydroxybutyrate dehydrogenase (BDH) from inner mitochondrial membrane, one of the most interesting and best documented examples of a lipid-requiring enzyme. This is a review of the molecular basis--knowledge and strategy of study--of the lipid specificity for membrane protein functions.

Comparison of the ability of phospholipids from rat liver lysosomes to reconstitute glucocerebrosidase

The in situ lipid activator of rat liver glucocerebrosidase was investigated. Rat liver lysosomes were purified (42.9-fold relative to the crude homogenate) by sequential isopycnic centrifugation in sucrose and metrizamide gradients. Lipids were extracted with chloroform:methanol (2:1) and phospholipids were separated by one-dimensional thin-layer chromatography. The phospholipid content of the lysosome preparation was 0.28 mumol lipid phosphorus/mg protein. Phosphatidylcholine was present as the major nonacidic phospholipid (39.3%). Of the acidic phospholipids, phosphatidylinositol and phosphatidylserine were present in the greatest amounts (12.0 and 19.7%, respectively). The resolved phospholipids were tested separately and in the presence of a heat-stable factor from Gaucher spleen for their ability to reconstitute butanol-delipidated rat liver glucocerebrosidase activity. Alone or in the presence of the heat-stable factor, phosphatidylserine and phosphatidylinositol were the most effective activators of glucocerebrosidase. Bis(monoacylglyceryl) phosphate derived from rat liver tritosomes or rabbit lung macrophages was also effective in reconstituting beta-glucosidase activity.

Comparison of N-acyl phosphatidylethanolamines with different N-acyl groups as activators of glucocerebrosidase in various forms of Gaucher's disease

The acidic phospholipid requirement of the predominant particulate beta-glucosidase of mammalian spleen and liver was investigated using a series of N-acyl derivatives of dioleoyl phosphatidylethanolamine (PE). The PE, a neutral phospholipid, was converted to an acidic lipid, (N-acyl)-phosphatidylethanolamine (NAPE) by acylation of the amino group with different fatty acyl chains. Lysosomal beta-glucosidases from rat liver and spleens of controls and patients with various types of Gaucher's disease were solubilized and delipidated by extraction with sodium cholate and 1-butanol. All members of the NAPE series tested were effective activators of the delipidated rat liver beta-glucosidase, and the stimulatory power of the NAPE family increased with increasing chain length of the fatty acid substitution. In contrast, dioleoyl-PE had no effect on beta-glucosidase activity. A heat-stable factor (HSF) purified from the spleen of a patient with Gaucher's disease significantly increased the sensitivity of the rat liver beta-glucosidase to all of the NAPE derivatives. The maximum stimulation achieved in the presence of HSF was independent of N-acyl chain length. Compared to the potent activator, phosphatidylserine (PS), (N-acetyl)-PE and (N-linoleoyl)-PE were half as effective as activators of beta-glucosidase from control human spleen. PS stimulated the beta-glucosidase of type 1 nonneurologic Gaucher's disease, but none of the NAPE compounds activated it. Neither PS nor any of the (N-acyl)-PE compounds could activate a delipidated preparation of beta-glucosidase obtained from the spleen of a neurologic case. These results indicate that although the presence of a net negative charge on a phospholipid confers upon it an ability to reconstitute beta-glucosidase activity to the normal, nonmutant enzyme, it is insufficient to permit differentiation of the various types of Gaucher's disease.