The role of the ganglioside lipid moiety in the process of ganglioside-cell interactions

Chemistry and Physics of Lipids, 62 (1992) 1-9 Elsevier Scientific Publishers Ireland Ltd.

1

The role of the ganglioside lipid moiety in the process of ganglioside-cell interactions Rosaria

Bassi and Sandro

Sonnino

Study Center for the Functional Biochemistry of Brain Lipids, Department of Medical Chemistry and Biochemistry, Medical School, University of Milan, Via Saldini 50, 20133 Milan (Italy)

(Received September 2nd, 1991; revision received December 13th, 1991; accepted April 6th, 1992) The role of the ceramide moiety of gangliosides, together with the deriving aggregative properties of ganglioside in solution, in the process of ganglioside-cell interactions was studied. The natural GMl(stearoyl) and the synthetic GMl(acetyl), containing the stearoyl and aeetyl groups as the acyl moiety, respectively, were used in binding experiments to rat cerebellar granule cells. Regardless oftbe cell culture conditions, such as the presence or absence of fetal calf serurn, the association of GM l(acetyl) to the cells was much greater than that of GMl(stearoyl). GMl(acetyl) was present in the incubation medium as monomers. After incubation, a large part of the total GM l(acetyl) associated to cells, 76-93% depending on the experimental conditions, was removed by washing with protein solutions. The remaining associated ganglioside was not removed by repeating washing with protein solutions or trypsin treatments and was considered as a component of the membrane. The cell association of GMl(stearoyl), present in solution as monomers as well as micelles, could be classified as serum-labile, trypsin-labile and trypsin-stable. The trypsin-stable form of association, corresponding to the molecules stably inserted into the membrane, was proportionally higher, the proportions varying with increasing incubation time and decreasing ganglioside concentration. This form of association was particularly high when incubation was performed in the presence of fetal calf serum. Incubation experiments performed with a mixture of GMl(stearoyl) and GMl(acetyl) in a molar ratio which allowed their presence in the medium as monomers as well as mixed micelles, led to a ganglioside association suggesting that besides the aggregative properties of the molecule other gangli0side properties are involved in the ganglioside-cell interaction process. Key words: gangliosides; ceramide; lipid moiety; aggregation properties; critical micellar concentration; granule cells

Introduction Gangliosides, sialic acid containing glycosphingolipids, are c o m p o n e n t s o f the cell membranes o f vertebrates and are particularly a b u n d a n t in neuronal cells. They contain a h y d r o p h o b i c portion, the ceramide, inserted into the m e m b r a n e outer lipid layer and a hydrophilic portion, the sialosyloligosaccharide, which protrudes from the m e m b r a n e surface [ I]. Gangliosides appear to be involved in the biotransduction o f membrane-mediated information, presumably acting as modulators o f some Correspondence to: Sandro Sonnino, Dipartimento di Chimica e Biochimica Medica, Via Saldini 50, 20133 Milano, Italy.

proteins instrumental to the process [2-4]. A widely used a p p r o a c h in studying the functional implications o f gangliosides is their addition to in vitro cellular systems and the determination o f the biological effects, which, till now, have been assumed to arise f r o m the preliminary insertion o f gangliosides into plasma membranes, thus modifying membrane properties [5-8]. Uptake o f gangliosides by cells also occurs in in vivo systems. Gangliosides administered to animals reach all tissues t h r o u g h the bloodstream [9]. Moreover it has recently been proved in the rat that gangliosides are released f r o m liver, thus modifying the serum content [10]. The molecular aspects o f ganglioside-cell interactions and the modalities o f the insertion of

0009-3084/92/$05.00 © 1992 Elsevier Scientific Publishers Ireland Ltd. Printed and Published in Ireland

gangliosides into the cells have been studied using the GM1 ganglioside [6-8] or GM1 ganglioside derivatives containing different probes [11-13]. The main findings are the following. GM1 ganglioside present in solution as micelles [14], become loosely bound to the cell surface and can be removed by treatment with solutions of proteins, with which it is known to form very tight and stable lipoproteic complexes [15]; a smaller ganglioside portion interacts strongly with proteins protruding from the membrane surface and can be released by trypsin treatment. Finally, a small portion of associated gangliosides is made up of molecules inserted into the membrane layer, where they are interacting with specific proteins [13]. It seems that the insertion can be mediated by the surface-located specific membrane-bound proteins [13,16]. The above scheme has often been correlated to the aggregative properties in solution of gangliosides. Gangliosides are present in solution as micelles of large molecular weight, in equilibrium with a small number of molecules in monomeric form [8]. The aim of this work was to verify the role of the ganglioside ceramide moiety in the ganglioside-cell membrane interaction process. For this purpose we used rat cerebellar granule cells in culture and GM1 ganglioside molecular species containing stearic and acetic acid, as the acyl moiety: GM l(stearoyl) and GM l(acetyl), respectively. The aggregative properties of the amphiphilic compounds are mainly determined by the geometrical constraints of the monomers inside the aggregate which are correlated to the geometric or packing properties of the amphiphilic molecule. In the case of the two above ganglioside molecular species there is a quite different contribution made by the lipid moieties. The aggregative properties of GMl(stearoyl) and of GMl(acetyl) are known. GMl(stearoyl) is present in solution, over 2 × 10-8 M [14], as non-spherical micelles showing a hydrodynamic radius and molecular weight of 58.7 /k and 470 000 Da, respectively [17]; GMl(acetyl) is present in solution, over 2 × 10-5 M, as nearly spherical micelles showing a hyrodynamic radius and molecular weight of 34 A, and 102 000 Da, respectively [18].

Materials and Methods Reagents

Commercial chemicals were the purest available, common solvents were distilled and water was doubly distilled in a glass apparatus. Poly-Llysine, trypsin, 1-ot-D-arabinofuranosylcytosine, bovine serum albumin, N-acetylneuraminic acid and galactose oxidase (EC 1.1.3.9) were from Sigma, USA; sodium boro[3H]hydride, 8 Ci/mmol, from Amersham, UK; silica gel pre-coated thin layer plates (HPTLC, Kieselgel 60, 10 × 10 cm) from Merck, Germany; basal modified Eagle's medium and fetal calf serum from Flow Laboratories, UK. GMl(stearoyl) ganglioside, /~-Gal-(1 ---3)-/~-GalNAc-(1-- 4)-[~-Neu5Ac-(2--, 3)l/~-Gal-(1 -- 4)-/3-G1c-(1 -- 1)-[2-(stearoyl)-amino-3hydroxy-octadec/eicos-4t-ene],was extracted from calf brain [19], purified to over 99% and characterized [20]; its content of stearic acid determined by gas chromatographic analysis [20] was 91% of the total fatty acid content, the remainder being covered by palmitic and arachic acids. GMl(acetyl) was prepared by alkaline hydrolysis of GM 1(stearoyl) followed by N-acetylation of the lysoderivative formed [18,21]. Both GM1 (stearoyl) and GMl(acetyl) were tritium labeled at position 6 of external galactose according to the galactose oxidase-sodium boro[3H]hydride procedure [22]. The specific radioactivity of GMl(stearoyl) and GMl(acetyl) was 1.2 and 1.4 Ci/mmol, respectively. Granule cells were prepared from the cerebellum of 8-day-old Sprague-Dawley rats and cultured as described [23]. The cells were plated into 60-ram-diameter Falcon dishes (about 5 × 106 cells/dish; about 300 t~g of protein/dish) previously coated with 0.001% poly-L-lysine and maintained at 37°C in a 5% CO2 humidified atmosphere. 1-~-Darabinofuranosylcytosine (10 t,M) was added 18 h later to inhibit nonneuronal cell growth. Cell culture treatments

Incubation, up to 2 h, of the ganglioside molecular species with cerebellar granule cells was carried out on day 7 in culture. (a) Cell culture

dishes were washed twice with pre-warmed supplemented basal modified Eagle's medium (EMEM) without fetal calf serum (FCS) and then incubated in the same medium (2 ml/dish) containing: (i) l 0 -6 to 10 -8 M o f GMl(stearoyl) or GMl(acetyl), (ii) 10-5 to 10 -7 M mixture of GMl(stearoyl) and GMl(acetyl) at a molar ratio of one; (b) cells were incubated with 10% FCSEMEM containing 10-6 to 10-7 M of GM 1(stearoyl) or GM 1(acetyl). Ganglioside solutions at 10-5 M were obtained by previous dilution, 1:10, of tritiated gangliosides with cold gangliosides; at lower ganglioside concentrations only the tritiated compounds were used. Granule cell viability in the presence of gangliosides was verified by the Trypan Blue exclusion assay [24]. The morphology of cultured granule cells was examined by phase-contrast microscopy. At the end of the incubation the medium containing the ganglioside molecular species was discarded and the cells were washed for 30 rain with supplemented basal modified Eagle's medium containing 10% FCS. Cells were then treated for 5 rain with 2 ml of EMEM containing 0.1% trypsin; after careful removal of the medium the cells were scraped off the plates with cold saline solution and centrifuged. Cells subjected to the above treatment or to part of it were analyzed to determine cellassociated radioactivity for radioactive ganglioside content.

Aggregative parameters of GMl(stearoyl) and GM1 (acetyl) ganglioside molecular species in solution The critical micellar concentration (cmc) of gangliosides has been studied by several authors, but there is some disparity as the values found vary greatly [25]. From the first studies [26] the cmc of GMl(stearoyl) was determined to be very low, 10-9 M and then [14] modified by convincing procedures to 2 x 10-8 M. This is the value we adopted for our purposes. The cmc of GMl(acetyl) is 2 x 10-5 M [18]. The theory developed for the micellization in systems of mixed surfactants [27] and then also applied to gangliosides [28] has been used for the determination of the aggregative parameters of the GMl(stearoyl)/GMl(acetyl) mixed system. The cmc of the mixed system CmCmix can be calculated using the expression: 1

a

(I - oO

CmCmi x

cmc I

cmc2

where a is the mole fraction of GM 1(stearoyl) and cmc~ and cmc2 are the cmc values of the two pure ganglioside molecular species GMl(stearoyl) and GMl(acetyl), respectively. For a = 0.5 the CmCmix is 3.4 × 10-8 M. At a ganglioside total concentration c lower than the CmCmix,when a = 0.5 the concentration

TABLE I Distribution of the molecules of gangliosides between miccllar aggregates and monomers in the mixed system GMl(stearoyl)GMl(acetyl), 1:1 by mol. System cone.

Micclles

Monomers

6.15 x 10 -6 3.01 x 10 -8 61.5 30.1

3.85 x 10 -6 6.99 x 10 -8 38.5 69.9

[c] (M) Moles Percentage (%)

Moles Percentage (%)

10 -5 10 -7 10 -5 10 -7

10 -5 10 -7 10 -5 10 -7

G M 1(stearoyl)

GM 1(acetyl)

GM 1(stearoyl)

GM i (acetyl)

4.98 × 10 -6 3.0 x 10 -8 81.0 99.7

1.17 × 10 -6 0.01 x 10 -8 19.0 0.3

0.02 x 10 -6 2.0 x 10 -8 0.5 28.6

3.83 x 10 -6 4.99 x 10 -8 99.5 71.4

Analytical procedures

of monomers of GMl(stearoyl), c~, or GMl(acetyl), c~, is = 0.5 x c. When c is above the CmCmixc~ and c~ can be calculated with the expressions

Cell-associated radioactivity was determined by liquid scintillation counting of the cells dispersed in 1 ml sodium hydroxide (3 mg protein/ml). Gangliosides were extracted and purified from pelletted cells according to the phosphate buffertetrahydrofuran procedure [19]. Gangliosides were separated by TLC using 10-cm-long HPTLC plates developed at room temperature, the solvent system chloroform-methanol-0.2% aqueous CaCI2 (50:42:11, v/v/v/) reaching the top edge of the plate and radioactive gangliosides were quantified by radiochromatographic scanning of the plate using a Berthold Digital Autoradiographic Radiochromatoscanner (DAR). GM1 ganglioside molecular species were assayed as bound Neu5Ac by the resorcinol-HC1 method [29-30], pure Neu5Ac being used as the reference standard. Protein content was determined in a sodium hydroxide solubilized cell pellet, bovine serum albumin being used as the reference standard [31].

- (c - A) + [(C - A) 2 + 4otcA] 1/2

1)

2 ( .cmc2 cmc I

c~=(1

cmc ci l ) cmc2

where A is the difference cmc2 cmcl. Following the above expression and with tx = 0.5 we determined the amount of GMl(stearoyl) and GM l(acetyl) in monomeric form, c~ and c~ and in micellar form, c~ and c M, at the concentrations used for binding experiments (Table I). -

T A B L E II Binding of G M l(stearoyl) and G M 1(acetyl) to rat cerebellar granule cells in culture in the absence o f fetal calf serum and any proteins. Distribution within the three forms of association: as pmoles/mg protein and as % on the total associated ganglioside. $L, serum labile; TL, trypsin labile; TS, trypsin stable. G M 1(stearoyl)

G M I (acetyl)

System concentration

[c] (M)

Incubation time (h) 0.5

1

2

pmol/ m g protein

0.5

1

2

%

0.5

1

2

0.5

pmoi/ mg protein

SL TL TS

111 50 31

143 86 35

140 113 53

58 26 16

54 33 13

46 37 17

780 . 81

.

SL TL TS

2.2 0.9 1.2

2.5 i.1 2.0

2.9 0.8 3.2

51 22 27

45 19 36

42 12 46

109 . 8

.

SL TL TS

0.2 0.5 0.7

0.5 0.3 0.9

0.5 0.2 1.0

14 36 50

29 18 53

29 12 59

12 . 1.3

937 . 147

.

114 . 11

131 . 15

.

12

14 .

1.5

2

91

90

86

9

l0

14

93

91

90

7

9

10

90

89

88

10

11

12

%

790 . 91

.

l

. 2

10 -6 10 -6 l0 -6 l0 -7 l0 -7 10-7 10 -8 10 -8 10 -s

TABLE III Binding of GMl(stearoyl) and GMl(acetyl) to rat cerebellar granule ceils in culture in the presence of fetal calf serum. Distribution within the three forms of association: as pmoles/mg protein and as % on the total associated ganglioside. SL, serum labile; TL, trypsin labile; TS, trypsin stable. GM 1(stearoyl)

GM 1(acetyl)

System concentration [c] (M)

Incubation time (h) 0.5

1

2

0.5

pmoi/ mg protein SL TL TS

0.2

SL TL TS

0.0 . 0.4

.

1.0

0.2 .

1.7

0.1 .

0.4

17 .

3.3

0.1 .

2

0.5

%

0.4 .

1

0.6

19

6

0.5

81

94

2

20 . 80

14

1

2

76

77

21

24

23

10-6

75 . 25

81

76

19

24

10-7 10-7 10-7

%

93

94

111

79

.

.

.

.

25

30

33

9 . 3

13 . 3

13

.

83

98

2

pmol/ nag protein

.

.

1

. 86

Results

Viability of cerebellar granule cells cultured in the presence of GMl(stearoyl) and GMl(acetyl) gangliosides Cerebellar granule cells in culture, incubated for up to 2 h with pure GM I (stearoyl) or GM 1(acetyl) over 10-6 to 10-s M concentrations or with a mixture of GMl(stearoyl)-GMl(acetyl), 1:1 by mol, over 10-5 to 10 -7 M concentrations, showed 100% viability. After incubation with GMI(acetyl) at a concentration of 10-5 M, neuronal cells deteriorated progressively as a function of time. When used at concentrations higher than its cmc, 2 x 10-5 M [18], all the cells died. At a GMl(acetyl) concentration of 5 x 10 -6 M cells did not show any suffering, this value was then considered as being safe for cell viability and binding experiments.

Binding of GMl(stearoyl) and GMl(acetyl) to cerebellar granule cells in culture (Tables H and Ill) Under the different experimental conditions, the total amount of ganglioside associated to cells never exceeded 2-3% of the total ganglioside present in the medium; this means that the medium,

. 4

10-6 10 -6

also after 2 h of incubation, was not exhausted of ganglioside. Incubations performed in the absence of proteins led to a very rapid binding of GM 1(stearoyl) to cells and after 30 min incubation the total amount of associated ganglioside was over 65% of that found after 2 h. Associated gangliosides could be fractionated by means of cell washings with 10% FCS-EMEM and with 0.1% trypsin solution into three different portions, defined as 'serum-labile', 'trypsin-labile' and 'trypsin-stable' forms of associated gangliosides. The serum-labile form of association predominated at high ganglioside concentration while a decrease in the ganglioside concentration and prolonged times of incubation resulted in an enhancement of the trypsin-stable form of association. These results largely overlap those found in binding experiments of GMl(stearoyl) to human fibroblast cells in culture [6,8]. The trypsin-labile form of association was very noticeably present at all ganglioside concentrations and incubation times. At high ganglioside concentration the trypsin-labile form of association increased thereby prolonging the incubation time, while at low ganglioside concentration the reverse behavior was found. When incubations were performed in the

6 TABLE IV

Binding of the mixed system G M l(stearoyl)-GMl(aeetyl), 1:1 by mol, to rat cerebellar granule cells in culture in the absence of fetal calf serum and any other protein. Distribution in the three forms of association. SL, serum labile; TL, trypsin labile; TS, trypsin stable. Incubation time (h) 0.5

i

2

16 260 430 1230

SL TL TS

46 0.7 6

2

91 2 7

87 3 10

83 4 13

10-5 10-5 10-5

87 i 12

86 3 11

86 2 12

10-7 10 -7 10 -7

0.5

pmol/mg protein SL TL TS

I

System concentration [cl (M)

% 16 890 590 1970 69 2.1 9

18 750 920 2900 81 1.5 11

presence of fetal calf serum the amount of associated gangliosides was strongly reduced. Due to the very low amount of associated radioactivity the experiments of binding in the presence of proteins could not be performed at a ganglioside concentration of 10 -s M. The trypsin-labile form of association was completely lacking and the trypsin-stable form always predominated covering 80-90% or more of the total association. The total amount of associated GMl(acetyl)

was always higher than that found for GMl(stearoyl): 5-50 times higher. The trypsinlabile form of associated gangliosides was lacking and the trypsin-stable form found at all concentrations was about 10% of the total ganglioside amount associated to the cells. Incubation performed in the presence of proteins reduced to a tenth the total association but did not greatly change the distribution of the three forms of association.

TABLE V

Binding of the mixed system G M 1(stearoyl)-GM 1(acetyl), 1:1 by mol, to rat cerebellar granule cells in culture. Percent distribution of GMl(stearoyl) and GMl(acetyi) in the three forms of ganglioside association to the cells. SL, serum labile; TL, trypsin labile; TS, trypsin stable. Incubation time (h) 0.5

1

2

0.5

10 -5 M

1

2

10-7 M

SL GMl(stearoyl) SL G M l(aeetyl)

48 52

46 54

46 54

9 91

11 89

5 95

TL GM l(stearoyl) TL GM l(acetyl)

100 0

0

0

0

0

0

TS GM1 (stearoyl) TS GM l(aeetyl)

31 69

30 70

29 71

29 71

17 83

17 83

Binding of the GMl (stearoyl)-GMl (acetyl) system to cerebellar granule cells in culture (Tables IV and V) GMl(stearoyl and GMl(acetyl), mixed in a ratio of 1"1 by mol, were incubated as described above, with cells in the absence of fetal calf serum. The three forms of association were all present. At 10-5 M the total ganglioside concentration and at 10-7 M the serum-labile form of association covered about 90% of the total associated gangliosides and the trypsin labile form was present in low amount. At both 10-5 and 10-7 M concentrations the trypsin-labile form contained only GM 1(stearoyl), in the trypsin-stable form the GM 1(acetyl) prevailed; 70-80% of the total. In the serum-labile form at the two concentrations the distribution of the two gangliosides was quite different; at the higher concentration the two gangliosides were present in the same amount but, at the lower concentration GMl(acetyl) greatly predominated. Discussion

In the present work we studied the binding process of two GM1 ganglioside molecular species, containing a short and a long acyl chain, to rat cerebellar granule cells in culture and we tried to correlate the process with the chemical and physico-chemical features of the molecules. The incubation of the gangliosides with the cells lasted for up to 2 h. Within this time only a very small part of the gangliosides enters the cells and is thus metabolized [32]. The reduced cell ganglioside metabolism, together with the fact that gangliosides were tritium-labeled at the galactose residue, is the reason why the radioactivity associated to the cell pellets is considered as the true bound GM1 ganglioside radioactivity. When GMl(stearoyl) was added to the proteindeprived cell medium, it bound cells very rapidly and the association could be divided, according to the experimental protocol reported in the Methods section, into the three forms of association coded as 'serum-labile', 'trypsin-labile' and 'trypsinstable'. Of these three forms, the proportion of the trypsin-stable form, considered as that corresponding to the molecules stably inserted into the cell

membrane, increased with the binding, prolonging the incubation time and decreasing the concentration. At the same time the serum-labile form of association followed an opposite trend. This confirms previous results which suggest that, while the monomers present in solution can become components of the membrane, the micelles can only loosely bind to the cell surface. Moreover the trypsin-labile form of association, which was always present, seemed to be an intermediate form. It has been postulated [13,16] that the proteins involved in the trypsin-labile association can modify the natural equilibrium between micelles and monomers, thus causing the increase of the amount of monomers in solution. Above 2 × l0 -s M, which we considered to be the cmc of GMl(stearoyl) [14], the serum-labile form of association was still present. A previous paper [26] reported the cmc of GMl(stearoyl) as being lower the 2 × l0 -s M. This could explain the presence of the serum-labile form of association at l0 -s M. However we must also consider that when the ganglioside monomers interact with the cell surface their concentration near the surface increases, thus allowing the possibility of overcoming the cmc and the formation of micelles. At 10-6 M concentration the majority of molecules are aggregated as micelles and only 2% of the total ganglioside concentration is present in solution as monomers, that is, in our experiments, 40 pmoles in the 2 ml of medium. After 2 h of incubation 53 pmol of ganglioside, more than that available in solution, were stably inserted into the membrane; mindful of the fact that the natural formation of monomers from micelles seems a very slow process [33], we need to postulate that other forms of interaction besides that involving the monomers could occur for the insertion of gangliosides into the cell membrane. Also, the results of the binding experiments performed with GMl(acetyl) disagree with the hypothesis that the interactions between gangliosides and cells depend only on the availability in solution of micelles and monomers and on the micelle-monomer equilibrium, as determined for the pure compound in a simple model of solution. GMl(acetyl) was found to bind to the cells in a very high amount and to be associated in the serum-labile form for 90% of the total associa-

tion; the same occurred at 10-8 M concentration. Micelles are not present in solution and therefore interaction would involve the ganglioside monomers. On the other hand it seems quite difficult at 10-8 M concentration to have a ganglioside density gradient with an increase of concentration in the proximity of the cell sufficient to overcome the cmc which is three orders of magnitude higher. The trypsin-labile form of association is lacking, thus indicating that this form of association does not depend on the monomeric state as was the hypothesis for GMl(stearoyl). On the other hand, if we suppose a process of micellization mediated by surface proteins, it would also mean that the micellar state is not necessary for the availability of the trypsinlabile form of association. To overcome the impossibility of having micelles of GMl(acetyl) without having cytotoxic effects we introduced the GMl(stearoyl)GMl(acetyl) mixed system. In a molar ratio of 1 at a concentration of 10-5 M, about 60% of the total molecules are aggregates of micelles containing 20% GM 1(acetyl), the remainder being covered by GMl(stearoyl) (Table I). GMl(stearoyl) and GMl(acetyl) were equaly present in the serumlabile form of association. The 99.6% of GMl(stearoyl) is present in solution as miceUar aggregates, thus the formation of the serum-labile form of ganglioside association should involve ganglioside micelles. The distribution of the ganglioside association within the three forms was similar to that found using pure GMl(acetyl), the serum-labile form of association covering 80-90% of the total ganglioside association. The predominance of the above form of association could depend on the presence of GMl(acetyl) in the mixed micelles or on the fact that the parameters, such as size and molecular weight, of the mixed micelles are different from those of pure GMl(stearoyl) and GMl(acetyl). According to an ideal mixing we calculated that the molecular mass of the mixed micelles would be about 286 000 kDa, a value 39% lower than that of the micelles of GMl(stearoyl) and 2.8 times higher than that of the micelles of GMl(acetyl). Moreover, due to the very different packing parameter of the two gangliosides [18] and according to recent sugges-

tions [34] we can suppose that the GMl(acetyl) is segregated at the edges of the mixed mieelle where a strong curvature is available, thus slightly increasing the ideal molecular mass and the axial ratio of the micelle. We found small proportions of the trypsin-labile form of association but this contained only GMl(stearoyl); this supports the idea that a double long chain moiety is necessary for the availability of strong interactions between gangliosides and surface membrane proteins. It was determined (see Methods) that GMl(acetyl) covers 99.6% of the total ganglioside monomers, but only 70% of the trypsin-stable form of association was covered by this ganglioside, GMl(stearoyl) being present in large amount. All this suggests that the total chemical amphiphilic structure of the ganglioside must be considered in the interaction process with the membrane. At very low concentrations, where the micelles contain practically only GMl(stearoyl) (Table I), the two gangliosides behave as separate entities. It should be noted that, for the validity of the above discussion, the use of a GMl(stearoyl) cmc value lower than 2 x 10-SM modifies by a few percent the composition of the mixed micelles reported in Table I. In the presence of fetal calf serum different forms of lipoproteic aggregates can be present in solutions [15,35], thus reducing the amount of ganglioside available for the interaction with the cell. In the case of GMl(stearoyl) the trypsinstable form of association becomes predominant; also the small variations, time- and concentrationdependent, are not easily explained. In the case of GMl(acetyl) the serum-labile form of association still remained the predominant form of ganglioside association. This would suggest that the monomers of GM 1(acetyl) are directly involved in the serum-labile form of association and that the affinity to GMl(stearoyl) and GMl(acetyl) shown by the cell and FCS proteins are quite different.

Acknowledgments The authors wish to thank Prof. Guido Tettamanti for his help and friendly discussion.

References 1 H. Wiegandt, (1985) New Comp. Biochem. 10, 199-260. 2 R.O. Brady and P.H. Fishman (1979) Adv. Enzymol. Relat. Areas Mol. Biol. 50, 303-323. 3 P.H. Fishman (1982) J. Membr. Biol. 69, 85-98. 4 S.-I. Hakomori (1986) Fidia Res. Ser. 6, 201-214. 5 J. Moss, P.H. Fishman, V.C. Manganello, N. Vaughan and R.O. Brady, (1976) Proc. Natl. Acad. Sci. U.S.A. 73, 1034-1037. 6 K. Radsak, G. Schwarzmann and H. Wiegandt (1982) Hoppe Seyler's Z. Physiol. Chem. 363, 263-272. 7 L. Facci, A. Leon, G. Toffano, S. Sonnino, R. Ghidoni and G. Tettamanti (1984) J. Neurochem. 42, 299-305, 8 V. Chigorno, M. Pitto, G. Cardace, G. Kirschner, S. Sonnino, R. Ghidoni and G. Tettamanti (1985) Glycoconjugate J. 2, 279-291. 9 G. Tettamanti, B. Venerando, S. Roberti, V. Chigorno, S. Sonnino, R. Ghidoni, P. Orlando and P. Massari (1981) in: M.M. Rapport and A. Gorio (Eds.), Gangliosides in Neurological and Neuromuscular Function, Development and Repair, Raven Press, New York, pp. 225-240. 10 S.C. Kivatinitz, A. Miglio and R. Ghidoni (1991) Biochem. J. 274, 581-585. 11 S. Kanda, K. Inoue, S. Nojima, H. Utsumi and H. Wiegandt (1982) J. Biochem. (Tokyo) 91, 2095-2098. 12 G. Schwarzmann, P. Hoffmann-Bleihauer, J. Shubert, K. Sandhoff and D. Marsh (1983) Biochemistry 22, 5041-5048. 13 S. Sonnino, V. Chigorno, D. Acquotti, M. Pitto, G. Kirshner and G. Tettamanti (1989) Biochemistry 28, 77-84. 14 B. Ulrich-Bott and H. Wiegandt (1984) J. Lipid Res. 25, 1233-1245. 15 M. Tomasi, G. Roda, C. Ausiello, G. D'Agnolo, B. Venerando, R. Ghidoni, S. Sonnino and G. Tettamanti (1980) Eur. J. Biochem. 111, 315-324. 16 K.C. Leskawa, R.E. Erwin, A. Leon, G. Toffano and E.L. Hogan (1989) Neurochem. Res. 14, 547-554. 17 L. Cantu', M. Corti, S. Sonnino and G. Tettamanti (1986) Chem. Phys. Lipids 41, 315-328.

18

19 20

21 22 23 24

25 26

27 28 29 30 31 32 33 34 35

S. Sonnino, L. Cantu', M. Corti, D. Acquotti, G. Kirschner and G. Tettamanti (1990) Chem. Phys. Lipids 56, 49-57. G. Tettamanti, F. Bonali, S. Marchesini and V. Zambotti, (1973) Biochim. Biophys. Acta 296, 160-170. R. Ghidoni, S. Sonnino, G. Tettamanti, N. Baumann, G. Reuter and R. Schauer (1980) J. Biol. Chem. 255, 6990-6995. S. Sonnino, G. Kirschner, R. Ghidoni, D. Acquotti and G. Tettamanti (1985) J. Lipid Res. 26, 248-257. G. Gazzotti, S. Sonnino, R. Ghidoni, P. Orlando and G. Tettamanti (1984) Glycoconjugate J. 1, 111-121. V. Gallo, M.T. Ciotti, A. Coletti, F. Aloisi and G. Levi (1992) Proc. Natl. Acad. Sci. U.S.A. 79, 7919-7923. H.J. Phillips (1973) in: P.F. Kruse, Jr. and M.K. Patterson, Jr. (Eds.), Tissue Culture Methods and Applications, Academic Press, New York. M. Corti, L. Cantu', S. Sonnino and G. Tettamanti (1988) Fidia Res. Ser. 14, 79-91. W. Mraz, G. Schwarzmann, J. Sattler, T. Momoi, B. Seemann and H. Wiegandt (1980) Hoppe Sayler's Z. Physiol. Chem. 361, 177-185. J.H. Clint (1975) J. Chem. Soc. 71, 1327-1334. M. Corti, V. Degiorgio, R. Ghidoni and S. Sonnino (1982) J. Phys. Chem. 86, 2533-2537. L. Svennerholm (1957) Biochim. Biophys. Acta 13, 856-858. J. Miettinen and J.T. Takki Lukkainen (1959) Acta Chem. Scand. 13, 856-858. O.H. Lowry, N.J. Rosebrough, A.L Farr and R.J. Randall (1951) J. Biol. Chem. 193, 265-275. R. Ghidoni, L. Riboni and G. Tettamanti (1989) J. Neurochem. 53, 1567-1574. S. Formisano, M.L. Johnson, G.L. Salvatore, M. Aloj and H. Edelhoch (1979) Biochemistry 18, 1119-1124. L. Cantu', M. Corti, S. Sonnino and G. Tettamanti (1990) Chem. Phys. Lipids 55, 223-229. B. Venerando, S. Roberti, S. Sonnino, A. Fiorilli and G. Tettamanti (1982) Biochim. Biophys. Acta 692, 18-26.