136

Bmchimica

et Biophysics

Acto, 1046 (1990) 136- 143 Elsevier

BBALIP 53473

Effect of lipid composition on the transfer of sterols mediated by non-specific lipid transfer protein ( sterol carrier protein,) Jeffrey T. Billheimer and James L. Gaylor * Cardiovascular

Group, E. I. Du Pont de Nemours

and Company,

Experimental

Statron,

Wilmington,

DE (U.S.A.)

(Received 26 March 1990)

Key words: Sterol; Lipid transfer protein; Lipid composition

The rate of non-specific lipid transfer protein (nsLTP)-mediated exchange is independent of structure for dissimilar sterols: cholesterol, lanosterol, sitosterol and vitamin D-3. Conversely, the nsLTP-mediated exchange of cholesterol is markedly affected by the phospholipid composition of the donor liposome. Negatively charged phosphatidylglycerols strikingly increase cholesterol exchange in the presence of nsLTP while not altering the exchange in the absence of nsLTP. The presence of unsaturated acyl chains in the phospholipid enhances exchange. Sphingomyelin drastically decreases cholesterol exchange, as does di-O-alkylphosphatidylcholine. Decreased exchange produced by these substitutions can be reversed by addition of phosphatidylcholine. The presence of an acyl group and a negative charge in the phospholipid are critical for the nsLTP-mediated transfer of cholesterol. In addition to these studies on composition of the donor membrane, the charge on the membrane also appears critical. Maximal exchange rates accompany optimization of potential interaction of negatively charged surface and the basic nsLTP. The nsLTP also mediates an approximately equal rate of exchange of cholesterol and phosphatidylcholine. However, approaching equilibrium, only half of the phospholipid can be exchanged while there is exchange of about 90% of cholesterol. Thus, it appears that only the phospholipid in an outer membrane layer may be available whereas cholesterol is fully available. Therefore, in contrast to a ‘carrier’ model we suggest that nsLTP facilitates exchange by binding to the membranes, and binding is highly dependent upon lipid composition. Once bound the protein functions as a bridge between membranes, thus, facilitating exchange.

Introduction Sterols and other lipids are transported in biological systems composed primarily of water. Understanding extracellular transport of cholesterol came with the isolation and characterization of the various classes of circulating lipoproteins. However, much less is known about intracellular transport of cholesterol from either its site of synthesis in membranes of the endoplasmic reticulum or from lysosomes (exogenously derived cholesterol) to the plasma membrane where cellular cholesterol concentration is greatest. To explain facilitated intracellular transfer of cholesterol between membranes, cytosolic transfer proteins have been sought. About 15 years ago, two putative transfer proteins were

* Present address: Corporate Office of Science and Technology, Johnson & Johnson, New Brunswick, NJ, U.S.A. Correspondence: J.T. Billheimer, Medical Products Department, Experimental Station, P.O. Box 80400, Wilmington, DE 19880-0400, U.S.A. 0005-2760/90/$03.50

identified in rat liver cytosol. Both have about the same molecular weight and, thus, have been difficult to separate; unfortunately, both were given essentially the same name (sterol carrier protein (SCP) and sterol carrier protein 2 (SCP,)) [1,2]. This situation unnecessarily led to considerable confusion, not only because initial experiments were carried out with mixtures containing both proteins, but the name ‘sterol carrier’ protein(s) conceptually suggests a mechanism of action of binding or solubilizing cholesterol and ‘carrying’ bound cholesterol across aqueous barriers between membranes. We have been interested in the possible function of these proteins in the biosynthesis of cholesterol from lanosterol that had been suggested for both SCP and SCP, [1,2]. Accordingly, we purified to homogeneity the cytosolic protein from rat liver that affected the activity of 4-methyl sterol oxidase in vitro [3] and showed that the purified protein is identical to both SCP and the well-characterized cytosolic fatty-acid binding protein (FABP) _ The latter is also known as Z-protein, and various functions in lipid metabolism have been ascribed to its strong affinity for fatty-acids and fatty acyl-CoA’s

0 1990 Elsevier Science Publishers B.V. (Biomedical Division)

137 [4,5]. In the last step of SCP (FABP) purification reported by us in 1980 [3], a minor basic protein of the same size was removed from SCP. The basic protein was capable of facilitating cholesterol transfer, and in 1983 we reported the purification to homogeneity of that basic cytosolic protein [6] which was identical to SCP,. A study of SCP, its identity with FABP, and its non-identity with SCP, was more recently reported in detail [7]. Furthermore, SCP, has been shown to be identical to a phospholipid exchange protein first isolated by Zilversmit and co-workers, and because of the protein-facilitated exchange of a wide variety of membrane lipids, it was named non-specific lipid transfer protein (nsLTP) [8]. With these publications and other studies reported by the mid-1980’s, there is no longer confusion resulting from either identification or separation of the proteins. However, even with pure SCP, protein available for about 10 years, neither the mechanism of facilitated transfer nor the physiological significance of SCP, has been revealed. SCP, (nsLTP) has been shown in vitro to increase the transfer of cholesterol from lipid droplets and liposomes to mitochondria and to affect several enzymes involved in the metabolism of sterols including: microsomal enzymes involved in the conversion of lanosterol to cholesterol [6,9], the esterification of cholesterol by acylCoA:cholesterol acyltransferase [6,10], and the oxidation of cholesterol by side-chain cleavage enzyme or 7-hydroxycholesterol oxidase [11,12]. In addition SCP, has been shown to facilitate the transfer of a number of different phospholipids [13,14]. These transfers of key metabolites under physiologically significant conditions at least suggest critical functions for nsLTP in intracellular lipid metabolism. On the other hand, the mechanism of facilitated transfer of lipids between membranes remains to be determined. We have continued to study the mechanism with pure rat hepatic nsLTP (SCP,) that is now abundantly available. Herein we report the very striking effect of altered phospholipid composition of liposomes on the transfer of cholesterol from these liposomal membranes to an acceptor membrane. Furthermore, careful quantitative analysis of exchanged lipids support the mechanistic suggestion that nsLTP facilitates direct membrane-to-membrane transfer rather than binding and travel across an aqueous barrier. Materials and Methods General

Egg lecithin, dioleoylphosphatidylcholine, l-oleoyl2-palmitoylphosphatidylcholine, l-palmitoyl-2oleoylphosphatidylcholine, dipalmitoylphosphatidylcholine, dimyristylphosphatidylcholine, phosphatidylserine (bovine brain), phospatidylinositol (bovine liver), phosphatidylglycerol (prepared from egg lecithin),

sphingomyelin (bovine brain) and 1,2 dioleoylglycerol were obtained from Avanti Polar Lipids. Sitosterol was supplied by Applied Science. Cholesterol, 25-hydroxycholesterol , dihydrolanosterol and vitamin D-3 were obtained from Steraloids. The dihydrolanosterol was further purified by HPLC. Triolein and pregnenolone were supplied by Sigma Chemical. 55 mCi/mmol [14C]and 50 Ci/mmol [ 3H]cholesterol, 87 Ci/mmol 25-hydroxy[3H]cholesterol, 35 Ci/mmol [ 3H]vitamin D-3, (50 mCi/ mmol) l-palmitoyl-2-oleoyl[‘4C]phosphatidylcholine, (137 Ci/mmol) 1-stearoyl-2-arachidonoyl [‘HISA-glycerol, (15 Ci/ mmol) [ 3H]- and 100 mCi/ mm01 [‘4C]triolein, 500 mCi/nmol [ 3H]dihydrolanosterol and 15 Ci/ mmol [ 3H]pregnenolone were obtained from New England Nuclear. [‘4C]Sitosterol (56 mCi/mmol) was supplied by Amersham. Melittin was the kind gift of Dr. J. Lear, pure fatty acid binding protein (SCP) was from Dr. G. Vahouny, apo A-I was supplied by Dr. M. Phillips and apo E was the gift of Dr. D. Usher. Liposome preparation

Radiolabeled sterol-containing liposomes were prepared by mixing 2.8 pMo1 of sterol(l.0 Ci/mol), a trace of labeled triolein (2.8 PCi) as a recovery standard, 3.2 ~Mol of phospholipid, and 25 pg of butylated hydroxytoluene in 4.0 ml of Tris-buffered sucrose as described by Bloj and Zilversmit [8,13]. Liposomes containing radiolabeled phospholipid (1.0 Ci/mol) as well as sterol-free liposomes were prepared in a similar manner. Liposomes were prepared by sonication using a bath sonicator (Laboratory Supplies Company). All liposome preparations were optically clear. Cholesterol liposomes containing different phospholipids were examined by transmission electron microscopy; a mean diameter between 30 and 45 nm was found. Lipid exchange assay

Radiolabeled lipid exchange was measured in a liposomal/mitochondria assay system similar to the method of Bloj and Zilversmit [13]. Rat liver mitochondria were prepared from the 27000 x g pellet obtained from rat liver homogenate. The pellet was resuspended in 20 mM Tris (pH 7.4) containing 0.25 M sucrose and 1 mM EDTA and centrifuged at 750 X g for 10 min. The resulting supernatant was centrifuged at 8500 x g for 10 min and the resulting pellet was washed twice and stored at - 70” C in Tris-sucrose EDTA buffer (1 : 1, v/v). Prior to use, samples of the frozen mitochondria (lo-15 ml) were thawed and heat-treated for 30 min at 60°C. The samples were cooled, homogenized in 30 ml of Tris-sucrose EDTA buffer and centrifuged for 10 min at 10000 X g. The corresponding pellet was resuspended in 4 ml of buffer: final protein concentration about 10 mg/ml. The lipid exchange assay contained in a final volume of 500 ~1 20 mM Tris (pH 7.4) 0.25 M sucrose, 1 mM EDTA, 3 mg of heat-treated mitochon-

138 drial protein, 1 mg bovine serum albu~n, liposomes (30 nMo1 total Lipid) and when added 5 pg of pure nsLTP. The contents were incubated for 20 min at 37°C. The samples were cooled and centrifuged at 10000 x g for 10 min. The radioacti~ty present in samples (150 ~1) of the supernatant was then determined. Radioactivity in control samples (150 ~1) was determined prior to centrifugation. Net exchange activity was defined as the difference in % exchange of radiolabeled lipid between incubations to which nsLTP had been added and incubations without nsLTP. Pur~ication of nsLTP

The nsLTP was purified from rat liver cytosol according to the procedure of Trzaskos and Gaylor [6]. The final nsLTP preparation gave one band on SDS gel electrophoresis, one band on isoelectric focusing gels, and only one N-terminal amino acyl sequence. Protein dete~~inat~o~

Protein was determined by the method of Bradford [15] using bovine serum albumin as a standard.

Fig. 1. Effect of phospholipid composition of donor liposome on the facilitated transfer of cholesterol by nsLTP. Transfer assays were carried out as described in Material and Methods. Donor liposomes contained 16 nmol of the indicated phospholipid and 14 nmol of cholesterol.

Rt?SUltS

Ability

of nsLTP

to increase the exchange of various lipids

The effect of sterol structure on rate of exchange by nsLTP is shown in Table 1. Variation from the structure of cholesterol in either the side-chain (sitosterol), nucleus (lanosterol), or scission of the AB ring (vitamin D-3) had little effect on the rate of percent net sterol exchange facilitated by nsLTP, whereas the unfacilitated rate differed by more than 4-fold. The transfer of diacylglycerol whose structural similarity with cholesterol is limited to a free hydroxyl group and a hydrophobic tail was also increased by nsLTP. The transfer of more hydrophilic steroids, 25hydroxycholesterol and

TABLE I Substrate specificity of nsLTP

Lipid exchange assay was carried out as described in Materials and Methods. Donor liposomes contained a total of 16 nmol of phosphatidylcholine and 14 nmol of cholesterol or other steroid, with the exception of those assays in which diacylglycerol was used, in which the final amount was 3.5 nmol. Compound

% Net exchange /lO mm

Cholesterol Sitosterol Dihydrolanosterol Vitamin D, Diacyglycerol Pregnenolone 25-Hydroxycholesterol Phosphatidylchohne

14.4+ 1.4 11.5jro.9 13.9+ 1.1 14.9kO.8 4.7 f 0.9 (2 25 < 0.5

pregnenolone, was not enhanced by nsLTP, but these substances exhibited high rates of exchange without nsLTP [16,17]. Recently, Van Amerongen et al 1181 similarly showed that nsLTP increased the transfer of 25-hydroxycholesterol from monolayer to vesicles only by 1.7-fold under conditions of transfer of 25-hydroxycholesterol that were 400-times the rate of cholesterol in the absence of nsLTP. Cholesterol ester and triacylglycerol, which are more hydrophobic than cholesterol, were not transferred by nsLTP f14]. Effect of phospholipid on cholesterol exchange

The cholesterol-phospholipid mole ratio of the donor liposomes did not affect the net exchange rate of cholesterol. The initial rate of exchange was the same (12.2 + 0.3% (10 mm) from liposomes having a 4-fold difference in ratio of cholesterol-phosphatidylcholine between 0.22 and 0.88. On the other hand, the exchange rate of cholesterol was markedly affected by the composition of the phospholipid. Exchange was increased about 2-fold by the substitution of negatively charged glycerol phospholipids for phosphatidylcholine (Fig. 1). Strikingly, only negligible exchange of cholesterol was observed when sp~ngomyelin was used as the source of phospholipid. To further test the marked effect of sphingomyelin in the donor membrane, liposomes were prepared in which the phospholipid composition was composed of a mixture of sp~ngomyelin and phosphatidylcholine (Fig. 2). Once again, little exchange was observed until phosphatidylcholine was greater than 50% of the phospholipid

139

100

% Phosphatidyl Choline 50 % Sphingomyelin

0

Fig. 2. Net exchange of cholesterol from liposomes containing a mixture of phosphotidylcholine and sphingomyelin. Liposomes contain the indicated amount of phosphatidylcholine and spbingomyelin. As a control (0) one sample was run in which one half of the Iiposomes contained 100% phosphatidylcho~ne and the other half of the liposomes contain 100% sphingomyelin. (a) % net exchange of cholesterol, (W) % exchange of cholesterol in the absence of nsLTP. In all cases donor liposomes contained a total of 16 nmol phospholipid and 14 nmol of cholesterol. Transfer assays were carried out for 30 min as described in Material and Methods. Data are the average of two experiments performed in duplicate.

present. Sphingomyelin did not appear to have a direct inhibitory effect on nsLTP because addition of liposomal vesicles prepared with sphingomyelin did not effect the nsLTP-mediated exchange of cholesterol from liposomal vesicles containing only phosphatidylcholine (Fig. 2). Sphingomyelin repeatedly has been shown to have an apparently greater affinity for cholesterol versus glycerol phospholipids. Because a part of the effect may have been due to the more saturated acyl chains of sphingomyelin 119,201, the effect of altered fatty acyl composition of the phospholipid on cholesterol exchange was studied. The rate of exchange of cholesterol from phosphatidylcholine liposomes was increased with increasing unsaturation of the acyl chains (Table II). This increase

Time (mm)

Fig. 3. Comparison of nsLTP-facilitated cholesterol exchange from liposomes containing dipahnitoyl phophatidylchol~e versus those containing sp~ngomyetin. Donor liposomes contained 16 nmol of the indicated phospholipid and 14 nmo1 of cholesterol. Data are the average of two experiments performed in duplicate. Over all time periods, the % exchange of cholesterol in the absence of nsLTP was 1.64k0.46 for dipahnitoyl phosphatidylcholine liposomes and 0.465 0.15 for sp~ngomyelin Iiposomes.

was observed both in the presence and absence of nsLTP; the latter is similar to the results reported by Bloj and Zilversmit [21]. The marked effect of sphingomyelin on exchange in the presence of nsLTP cannot be explained by the somewhat greater amount of saturated acyl chains in sp~ngo~pids because even from liposomes containing the fully saturated dipal~toylphosphatidylcholine, cholesterol was exchanged at lo-times the rate of cholesterol from liposomes prepared with sphingomyelin liposomes (Fig. 3). Because major effects of liposome composition were observed other structural differences, e.g., the presence of esters were studied by examining the exchange of cholesterol from liposomes containing ether lipids (Table II). Replacement of the palmitoyl ester with a

TABLE II Effect of forty ucyl (alkyl) composifion on cholesterol exchange acrivity

Lipid exchange assay was carried out as described in Materials and Methods. Donor liposomes contain a total of 16 nmol of designated phospholipid and 14 nmol of cholesteroi. Compound

% Net exchange with nsLTP/lO tin

Egp lecithin D&oylphosphatidylcholine Oleoylpalmitoylphosphatidylcholine Palmitoyloleoylphosphatidylcholine D~p~~toylph~phatidylcholine Dimy~st~hosphatidylchotine 0-Hexadecyloleoyl phosphatidylcholine Di-o_Hexadecylphosphatidylcholine

16.4&0.5 24.2 f 1.0 13.6k2.0 11.3*1.5 4.3kO.6 5.7i1.3 6.9zk2.1 85% of the cholesterol was exchanged. Thus, even approaching equilibrium, phosphatidylcholine appeared to be transferred only from the outer leaflet of the liposome. Discussion

Comparison of analytical data earlier revealed that SCP, is the same as nsLTP originally isolated by Zilversmit as a non-specific transfer protein of both phospholipid and cholesterol [6]. Although Scallen et al [7] suggested that nsLTP-mediated transfer of phospholipid is secondary to the transfer of cholesterol and may not be physiologically important, both rates and amounts of total transfer of cholesterol and phospholipids had not been measured under identical conditions. Accordingly, in this report we have studied the exchange of both sterols and phosphatidylcholine under identical conditions utilizing pure nsLTP to get a better understanding of the role of nsLTP in the exchange process. Under identical conditions nsLTP exchanged both cholesterol and phosphatidycholine at about the same rate from liposomes containing [ ‘HIcholesterol and [‘4C]phosphatidylcholine. Any proposed mechanism must now include results reported above. Furthermore, all data also were consistent with the suggestion that nsLTP must bind to the liposome, and ionic interactions appeared to be highly critical (Figs. 2-4). The exchange rate was markedly affected by structure of the liposomal phospholipid, but the sterol structure does not appear to be critical. The suppressive effect of addition of sphingomyelin and the stimulating effect of addition of phosphatidylserine were most striking. In the latter case, unlike essentially ail observations reported here and to date, phosphatidylserine enhanced the nsLTP-facilitated exchange while not concomitantly enhancing unfacilitated exchange, thus emphasizing the potential interaction of the liposomal phospholipid and nsLTP. In other words, charge was only important under conditions of exchange facilitated by the highly basic nsLTP [6]. Four mechanisms have been suggested for this type of protein-mediated transfer of lipids between membranes: (a) protein promoting membrane fusion, (b) protein increasing the off-rate of lipids from membranes, (c) protein carrying bound lipid across an aqueous barrier and (d) protein facilitating close association of lipids in two membranes without either fusion or an aqueous space. Triacylglycerol, an internal negative control, is not transferred in the presence of nsLTP. Fusion of membranes would allow equilibration of all

components. These observations appear to rule out fusion as the mechanism of transfer. McLean and Phillips [26] have shown that the exchange of cholesterol and phosphatidylcholine between vesicles in the absence of transfer proteins is consistent with aqueous diffusion of the lipids with the limiting step being the desorption of the lipid from the bilayer. The half time of transfer was calculated as 2.3 h for cholesterol and 48 h for phosphatidylcholine. Sequence data [27] and monolayer studies [28] demonstrate that nsLTP is an amphipathic protein. Thus, it is possible that nsLTP may interact with the membrane and disrupt the bilayer with an increase in the off-rate of lipids. In agreement with this mechanism is the observation that an increased transfer of cholesterol from liposomes containing more unsaturated acyl chains is observed both in the presence and absence of nsLTP (Table II). On the other hand to have approximately equal rates of exchange nsLTP would have to increase the desorption rate of phospholipid by 56-fold while increasing the off-rate of cholesterol only 3.5-fold (data from Table I and Ref. 25). This still leaves the possibility that nsLTP increases the off-rate of cholesterol but facilitates phospholipid transfer by a different mechanism. Also, several proteins which are known to perturb organized membrane structure such as apo E, apo A-I, fatty acid binding protein, and melittin in our hands do not increase the transfer of cholesterol in this assay system (data not shown). Purified nsLTP does not contain any bound lipid [ll]. Phospholipid binding to nsLTP has not been observed previously using radiolabeled phopholipids [14,24]. Similarly, when pure nsLTP is incubated with negatively charged liposomes containing radiolabeled cholesterol and the liposomes removed by ion-exchange chromatography, no cholesterol is found in the eluent which contains the nsLTP (data not shown). Recently, Nichols suggested that the fluorescent-labeled phospholipid, P-C,,-NBD-PC, binds to nsLTP [25]. In these experiments, however, a nsLTP-NBD-PC complex was not isolated but binding was calculated indirectly from decreased self-quenching when nsLTP was added to liposomes containing only NBD-PC. To do those experiments, Nichols used an extremely high molar ratio of nsLTP to phospholipid, between 1 : 1 and 10: 1. Because nsLTP self-associates and at this high ratio one might expect the amount of the amphipathic nsLTP alone could be sufficient to decrease the self-quenching of the NBD-PC, the results may be beyond comparison to the more catalytic functions of nsLTP. Binding of radiolabeled cholesterol to nsLTP has been proposed by Chanderbahn, et al [ll]. Again, the nsLTP-cholesterol complex was not isolated but binding was concluded from an increase in radiolabeled cholesterol found in the subnatant when nsLTP was added to a solution containing adrenal lipid droplets. Under similar condi-

142 tions, phospholipid binding to nsLTP was not detected [29]. Non-lipid components of the lipid droplet which may affect the movement of cholesterol were not studied and even in the absence of nsLTP, cholesterol was detected in the subnatant (about 30% of that observed in the presence of 80 pg nsLTP) (Fig. 4, Ref. 11). These studies suggest that if bound cholesterol and phospholipid rapidly dissociate from the nsLTP, nsLTP does not meet the criteria of a binding protein. We suggest therefore, that nsLTP has two binding domains. Upon interaction of one domain with a lipid bilayer (donor membrane) the second domain can interact with a second membrane to promote the intermembrane exchange of lipids. A similar mechanism has been proposed for phospholipid exchange [26] by suggesting that one molecule of nsLTP associates with each of the donor and acceptor membranes and then dimerization of the nsLTP causes a close association of the two bilayers. Because nsLTP self-associates and removal of dimer is the final step in pu~fication [6] our observations would be consistent with this suggestion and we have not carried out studies to discriminate between active monomer and dimer. Membrane-binding of nsLTP is strongly suggested in this current work. The transfer of cholesterol was facilitated by a negatively charged phospholipid and under conditions where the non-facilitated transfer was not affected (Fig. 4). Ionic attraction appears essential. Furthermore, DiCorleto and Zilversmit [30] have shown that in their assay system, little phospholipid was transferred in absence of negatively charged phospholipid in the liposome. This model predicts that the nsLTPfacilitated transfer of cholesterol and phospholipid would be identical. If nsLTP is important physiologically in the transfer of intracellular lipids, the effect of other lipids, especially phosphatidylserine and sphingomyelin, on the transfer process would be consistent with a widely varying distribution of cholesterol among membranes and organelles. Indeed, this mechanistic study may actually reveal some elements of physiological significance based on effects of variable lipid compositions of functionally different membranes. For example, in plasma membrane which contains the largest concentration of cholesterol, phosphatidylserine is located primarily in the inner leaflet and sphingomyelin in the outer leaflet 131J, The internal phosphatidylse~ne might facilitate rapid intracellular ~ui~bration of cholesterol, whereas sphingomyelin in the outer leaflet might conserve cellular cholesterol by retention. However, physiological significance of nsLTP still remains an enigma. We are reminded of another cytosolic lipid-binding protein, ligandin, which binds fattyacids and other organic anions [32]. A number of suggestions of its involvement in fatty-acid metabolism

were put forth until ligandin was shown to be glutathione S-transferase 1337. Acknowledgments

We thank Joan Binkley, Debra Cromley and Susan Fager for excellent technical assistance and Claire Stecher for manuscript preparation. References 1 Ritter, M.C. and Dempsey, M.E. (1971) J. Biol. Chem. 246, 1536-1547. 2 Scallen, T.J., Seetharam, B., Srikantaiah, M.V. Hansbury, E. and Lewis, M.K. (1975) Life Science 16, 853-874. 3 Billheimer, J.T. and Gaylor, J.L. (1980) J. Biol. Chem. 255. 81X8135. 4 Warner, M. and Neims, A.H. (1975) Can. J. Phys. Pharm. 53, 493-500. 5 Ockner, R.K., Manning, J.A., Poppenhauser, R.B. and Ha, W.K.Z. (1972) Science 177, 56-68. 6 Trzaskos, J. and Gaylor, J.L. (1983) Biochim. Biophys. Acta 751, 52-65. 7 Scallen, T.J., Noland, B.J., Gavey, K.L., Bass, N.M., Ockner, R.K., Chanderbhan, R. and Vahouny, G.V. (1985) J. Biol. Chem. 260, 4733-4739. 8 Bloj, B., Hughes, M.E., Wilson, D.B. and Zilversmit, P.B. (1978) FEBS Lett. 96, 87-89. 9 Noland, B.J., Arebalo, R.E., Hansbury, E. and Scallen, T.J. (1980) J. Biol. Chem. 255,4282-4289. 10 Gavey, K.L., Noland, B.J. and Scallen, T.J. (1981) J. Biol. Chem. 256. 2993-2999. 11 Chanderbhan, R., Noland, B.J., Scallen, T.J. and Vahouny, G.V. (1982) J. Biol. Chem. 257, 8928-8934. R.. 12 Seltman, H., Divan, W., Risk, M., Noland, B.J., Chanderbhan, Scallen, T.J., Vahouny, G. and Sanghvi, A. (1985) Biochem J. 230, 19-24. 13 Bloj, B. and Zilversmit, D.B. (1977) J. Biol. Chem. 252, 1613-1619. D.B. (1980) Biochemistry 19, 143314 Crain, R.C. and Zilversmit, 1439. 72.248-254. 15 Bradford, M.M. (1976) Anal. B&hem. Biophys. 209, 321-324. 16 Sinensky, M. (1981) Arch. B&hem. J.J.H., Jackson, R.L., Kempen, H.J.M. and Demel, 17 Theunissen, R.A. (1986) Biochim. Biophys. Acta 860, 66-74. A., Demel, R.A., Westerman, J. and Wirtz, 18 Van Amerongen, K.W.A. (1989) Biochim. Biophys. Acta 1004, 36-43. 19 Fagler, L., Chejan, S. and Bittmen, R. (1985) J. Biol. Chem. 260, 4098-4102. 20 Lange, Y., D’Allessandro, J.S. and Small, D.M. (1979) Biochim. Biophys. Acta. 556, 338-398. 21 Bloj, B. and Zilversmit, D.B. (1977) Biochemistry 16, 3943-3947. 22 Cram, R.C. (1982) Lipids 17, 935-943. 23 Bloj. B. and Zilversmit, D.B. (1981) J. Biol. Chem. 256,598&599X. 24 Van Amerongen, A., Teerlink, T., Van Heusden, G.P.H. and Wirtz, K.W.A. (1985) Chem. Phys. Lipids 38, 195-204. 25 Nichols, J.W. (1987) J. Biol. Chem. 262,14172-14177. 26 McLean, L.R. and Phillips, M.C. (1981) Biochemistry 20, 28932900. 27 Pastuszyn, A., Noland, B.J., Bazon, F., Fletterick, R.J. and Scallen, T.J. (1987) J. Biot. Chem. 262, 13219-13227. 28 Demel, R.A., Lowers, H., Jackson, R.L. and Wirtz, K.W.A. (1984) Colloids and Surfaces 10, 301-311.

143 29 Scallen, T.J. and Vahouny, G.B. (1985) in Sterols and Bile acids (Danielson, H. and Sjovall, J., eds), pp. 73-93, Elsevier, Amsterdam. 30 DiCorleto, P.E. and Zilversmit, D.B. (1977) Biochemistry 16, 2145-2150. 31 Verkleij, A.J., Zwaal, R.F.A., Roelofsen, B., Comfurius, P., Kastelijn, D. and Van Deenen, L.L.M. (1973) Biochim. Biophys. Acta 323, 178-193.

32 Litwack, G., Ketterer, B. and Arias, I.M. (1971) Nature 234, 466-467. 33 Habig, W.H., Pabst, M.J., Fleischner, G., Gatmaitan, Z., Arias, I.M. and Jakoby, W.B. (1974) Proc. Natl. Acad. Sci. USA 71, 3879-3882.