Plasma protein-facilitated coupled exchange of phosphatidylcholine and cholesteryl ester in the absence of cholesterol esterification.

A protein(s) which catalyzes the exchange of phosphatidylcholine and cholesteryl ester between plasma lipoproteins has been purified 10,000-fold from lipoprotein-free human plasma. The apparent molecular weight of the protein of the active fraction, designated lipid transfer complex (LTC), is approximately 61,000; when electrophoresed in 6 M urea, 0.1% sodium dodecyl sulfate on a 3-20% polyacrylamide gradient, the protein appears as a doublet of molecular weights 58,000 and 63,000. The active material is a glycoprotein which binds to concanavalin A. Human LTC is a lipid-protein complex with phospholipid, cholesterol, cholesteryl ester, and glyceride comprising 7% of the total mass. A similar glycoprotein (or glycoproteins) exists in rat plasma, although the fold-purification thus far achieved is low: about 500-fold. Moreover, the rat preparation enhances exchange of phosphatidylcholine, but does not appreciably enhance exchange of cholesteryl ester. Partially purified LTC (less than or equal to 3500-fold) exists in a complex with lecithin: cholesterol acyltransferase. Active lecithin: cholesterol acyltransferase is not, however, required for exchange of phosphatidylcholine or cholesteryl ester facilitated by human LTC. The rates of exchange of phosphatidylcholine and cholesteryl ester facilitated by human LTC are equal. Coupled lipid exchange occurs at all stages of LTC purification, at values of pH between 5 and 10, and at ionic strengths as great as 0.9. Moreover, phosphatidylcholine and cholesteryl ester are exchanged with 1:1 stoichiometry in the presence of thiol group reagents such as 5,5'-dithiobis-(2-nitrobenzoic acid). Both lipid exchange activities are relatively resistant to elevated temperatures. Coupled exchange of phospholipid and neutral lipid is not dictated by the nature of the lipoprotein donor and acceptor substrates: bovine liver phospholipid exchange protein catalyzes exchange of phosphatidylcholine but not cholesteryl ester between low and high density lipoproteins under conditions identical with those in which human LTC facilitates exchange of both lipids.

1 Established Investigator of the American Heart Association. capsulating agents for site-specific delivery of therapeutic molecules via the blood stream, have stimulated investigation of plasma proteins which bind lipids, which catalyze lipid synthesis and degradation, and which enhance the exchange or transfer of lipids between biosurfaces. The primary focus of these investigations has been the plasma lipoproteins, well recognized classes of particles which transport lipids and which regulate lipid metabolism extracellularly and intracellularly. Our understanding of the structure and function of the lipoproteins has been simultaneously facilitated and hampered by the fact that many of their constituents are in dynamic equilibrium. Proteins (1) and certain lipids such as cholesterol (2), monoacyl phospholipids (3), and nonesterified fatty acids (4) exchange rapidly between lipoprotein classes and between lipoproteins and cells. On the other hand, cholesteryl esters, triglycerides, and diacyl phospholipids redistribute very slowly when lipoproteins are incubated in vitro (5-7) but rapidly in plasma (8-lo), prompting the search for water-soluble plasma proteins which catalyze the migration of these lipid constituents.
Indeed, plasma proteins which facilitate the transfer' of cholesteryl esters, phospholipids, and triglycerides have been identified, isolated, and partially characterized. Zilversmit and colleagues (11,12) isolated a glycoprotein, with a molecular weight of 80,000 and an isoelectric point of 5, from human plasma which catalyzes the exchange of cholesteryl esters between all classes of lipoproteins. Ihm et al. (13) reported that a glycoprotein complex, with an apparent molecular weight of 150,000 and isoelectric point of 5.2, partially purified from human plasma d > 1.21 g/ml infranatant mediates the exchange of both cholesteryl ester and phosphatidylcholine between lipoproteins. The exchange occurs with a 1:l stoichiometry, cholesteryl ester:phosphatidylcholine. A glycoprotein with an apparent molecular weight of 35,000, identified as apolipoprotein D, has been purified by Chajek and Fielding (14); this factor enhances the net transfer of cholesteryl ester from high density to very low or low density lipoproteins with concomitant and equimolar reverse-transfer of triglyceride to HDL.2 However, evidence presented by Morton and Zilversmit (15) indicates that the human transfer protein is not apolipoprotein D. Finally, Rajaram et al. (16) achieved a 500-The terms "exchange" and "transfer" are used interchangeably throughout to mean bidirectional lipid movement. "Net transfer" refers to the unidirectional movement of lipid. 4819 fold purification of a triglyceride transfer protein from rabbit serum. The active fraction contained a glycoprotein with a molecular weight of 100,OOO-115,000 and an isoelectric point of 9. The experiments reported herein were designed to purify further the human lipid exchange protein described initially by Ihm et al. (13) which simultaneously facilitates exchange of cholesteryl ester and phosphatidylcholine, to assess the tightness of coupling between transfer of neutral lipid and phospholipid, and to probe the relationship between the lipid exchange protein(s) and 1ecithin:cholesterol acyltransferase. For comparison, a lipid exchange protein was isolated by similar procedures from rat plasma. The data suggest that both human and rat lipid exchange proteins have molecular weights of about 61,000 as determined by electrophoresis in polyacrylamide gels containing sodium dodecyl sulfate. The human protein has been analyzed for lipid and is found to have associated phospholipid, cholesterol, cholesteryl ester, and glyceride. Under all conditions of assay, the human lipid exchange protein(s) catalyzes exchange of cholesteryl ester and phosphatidylcholine with 1:l stoichiometry between human plasma LDL and HDL. The stoichiometry of lipid exchange catalyzed by the rat plasma protein is about 1:15, cholesteryl ester:phosphatidylcholine. Although 1ecithin:cholesterol acyltransferase activity is associated with the lipid exchange protein(s), lipid exchange and 1ecithin:cholesterol acyltransferase activity are uncoupled, indicating that active acyltransferase is not required for facilitated lipid exchange.

EXPERIMENTAL PROCEDURES
Materials-Cholesterol and Triglyceride Test Combination Kits were obtained from Boehringer Mannheim. [1,2-'H]Cholesterol (40-60 Ci/mmol), [dip~lmitoyZ-l-'~C]phosphatidylcholine (60-100 mCi/ m o l ) , and Aquasol-2 were purchased from New England Nuclear. L-a-Dipalmitoyl phosphatidylcholine and L-a-dilinolenoyl phosphatidylcholine were obtained from Avanti Biochemicals; bovine serum albumin (Fraction V and essentially fatty acid-free) and human serum albumin from Sigma. Liquid chromatography supports were purchased as follows: phenyl-Sepharose and concanavalin A-Sepharose from Pharmacia Fine Chemicals; Bio-Gel P-4 and A-I5 m from Bio-Rad; and CM52 cellulose from Whatman. Ultradex was obtained from LKB. Silica Gel IB2 strips and Aquacide were obtained from J. T. Baker Chemical Co. and Calbiochem, respectively. Rat plasma was purchased from Pel-Freeze Biologicals. The phosphatidylcholine exchange protein was purified from beef liver by the method of Kamp and Wirtz (17). Purified apo-AI was obtained from Dr. Richard L. Jackson (University of Cincinnati).
Isolation of Plasma Lipoproteins and d > 1.21 g/ml Plasma Infranatant-Lipoproteins were isolated from the freshly collected plasma of normolipemic fasted human volunteers by sequential ultracentrifugal flotation in KBr containing 4 X M EDTA, pH 7.0 (18). A Beckman type 50.2 Ti rotor was used for all centrifugation steps. The purity of each lipoprotein class was assessed by electrophoresis on agarose (176, pH 8.6) and by immunodiffusion against antibodies raised against LDL, HDL, and apo-B, apo-E, apo-AI, apo-AII, and serum albumin. LDL were isolated between d 1.019 and d 1.063 g/ml by centrifugation for 18 h at 48,000 rpm. KBr was added to the d 1.063 infranatant to increase the density to 1.21 g/ml and the solution was centrifuged for 24 h at 48,000 rpm to float lipoproteins of high density class. The infranatant, excluding the clear zone immediately below the lipoprotein layer, termed the d > 1.21 g/ml plasma, was employed as the source of the lipid transfer complex for most of the experiments outlined. Lipoproteins and d > 1.21 g/ml plasma were stored at 4 "C in the KBr solution in which they were isolated; all lipoproteins were dialyzed immediately prior to use.
Preparation of LDL and HDL Labeled with [14CJDipalmitoyl Phosphatidylcholine and PHICholesteryl Ester-To prepare double-labeled ['4C]DPPC/[3H]CE-LDL or -HDL, the cholesteryl esters and the phospholipids were labeled sequentially. Freshly isolated plasma (35 m l ) was incubated at 37 "C for 72 h with 1 mCi of [1,2-'H] cholesterol impregnated on a Whatman No. 1 filter paper disc (19). Sodium azide (0.02%) was included in the incubation medium. At the conclusion of the incubation, the density of the solution was adjusted with KBr, and LDL and HDL were obtained by sequential ultracentrifugation in KBr at solvent densities 1.019. 1.063, and 1.21 g / d , respectively, employing a Beckman 50 Ti rotor operating at 48,000 rpm for 18-24 h at 15 "C. About 65-75% of the radioactivity in the The radiolabeled unesterified cholesterol in LDL was removed by lipoproteins was cholesteryl ester, the remainder being cholesterol. exchange with unlabeled cholesterol of HDL. LDL were incubated with a 20-fold excess (cholesterol) of human HDL for 6 h at 37 "c. Similarly, the radiolabeled cholesterol in HDL was removed by incubation with a 5-fold excess of LDL (protein). The mixtures were then centrifuged at density 1.063 g / d to reisolate the ["HICE-lipoproteins. The LDL typically had a radioactivity of 7,000 cpm/pg of total cholesterol; the HDL, about 6,000 dpm/pg of cholesteryl ester. In general, less than 4% of the radioactivity occurred as unesterified cholesterol.
[ Assay of the Lipid Transfer Complex-The routine assay of column chromatographic fractions for transfer of phosphatidylcholine and cholesteryl ester was conducted in Tris-buffered saline at 44 "C with [I4C]DPPC-and ["ICE-labeled human LDL (usually 50-120 nmol of phospholipid per assay) and human HDL (500-1000 nmol of phospholipid per assay) as lipid donors and acceptors, respectively. Experiments designed to probe stoichiometry of exchange of phosphatidylcholine and cholesteryl ester and to assess fold-purification utilized double-labeled [14C]DPPC/[JH]CE-LDL. In some experiments, lipid transfer was also monitored in the reverse reaction, i.e. from ['4C]DPPC/[JH]CE-HDL (120 nmol of phospholipid per assay) to LDL (600 nmol of phospholipid per assay). After a 3-h incubation in the absence and presence of LTC, LDL and HDL were reisolated by ultracentrifugation. The density of the incubation mixture was raised to 1.21 g/ml by the addition of 0.32 g of KBr per ml of solution. The d 1.21 g/ml mixture was layered under d 1.063 g/ml KBr and the mixture was centrifuged for 18 h at 48,000 rpm in a 50 Ti rotor. LDL floated as a thin yellow surface fdm and HDL sedimented to the bottom of the tube. The LDL and HDL were collected and 0.5 ml of each was added to 10 ml of Aquasol-2. The amount of radioactivity (3H and "C) in LDL and HDL was determined in a Beckman LS-230 liquid scintillation counter equipped with an adjustable discriminator. The discriminator was set so that 'H counts in the I4C channel were negligible. I4C counts detected in the 'H channel were 10-12% of I4C counts in the I4C channel. The percentage of overlap was determined in each experiment by counting standard I4C-containing solutions under the same conditions. The I4C counts in the 'H channel were subtracted to obtain ' H counts. T o reduce the counting error, the ratio of 3H to I4C was usually maintained relatively high (between 7 lated LDL and HDL was usually 92-9696 of the total amount added. and 10 to 1). The percentage of radioactivity recovered in the reiso-Lipids extracted from both donor and acceptor lipoproteins at all assay conditions employed were analyzed by thin layer chromatography and it was determined that >95% of the transferred phospho- Assay of Lecithin:Cholesterol Acyltransferase-The substrate for 1ecithin:cholesterol acyltransferase was that described by PownalL3 A complex of apo-AI, dilinolenoyl phosphatidylcholine, and [3H]cholesterol was prepared in a molar ratio of 1:100:2. The desired quantities of DLLPC, [1,2-3H]cholesterol (2 mol %), and organic solvent were dried under Nz. Sufficient Tris-buffered saline was added to give a final DLLPC concentration of 40 m. The DLLPC/cholesterol mixture (0.25 ml) was mixed with 0.25 ml of apo-AI (440 PM) at room temperature and sufficient sodium cholate (-12.5 mg in 60 p1) was added to clarify the solution. The mixture was then eluted from a column of Bio-Gel P-4 (0.9 X 19 cm) to remove the cholate.
The standard assay was conducted using 100-pl aliquots of Trisbuffered saline containing 1 m~ dithiothreitol, 500 pg of human serum albumin, 15 p1 of the stock substrate solution, and a suitable amount of assay sample. The final volume was 0.1-0.4 ml . The assay mixtures were incubated at 37 "C for 6-20 h. These conditions are optimal for assessing lecithin:cholesterol acyltransferase activity. At the end of the incubation, the reactions were terminated by addition of 0.5 ml of methanol. Unesterified and esterified cholesterol were extracted into 4 ml of hexane. The top layer of hexane was withdrawn and evaporated to dryness under Nz. Lipids were redissolved in 0.5 ml of hexane, a 50-pl aliquot was transferred to a scintillation vial to determine total radioactivity, and the remainder was loaded onto a silica gel minicolumn. Eluant (2.0 ml of hexane:diethyl ether, 6:1, v/v) was added to the column and collected directly in a scintillation vial. Minicolumns of silica gel (0.5 X 3.0 cm) were constructed from 14.5-cm Pasteur pipettes; the silica gel was preheated to 100 "C for 2 h.
Analytical Procedures-Protein concentrations were determined by the method of Lowry et al. (20), modified by the addition of 1% SDS to clarify the samples; bovine serum albumin was the standard. Total glyceride was determined enzymatically (21) using the Triglyceride Test Combination Kit. Cholesterol and cholesteryl esters were determined by the method of Roeschlau et al. (22) using the Cholesterol Test Combination Kit. Total phospholipid was measured by phosphorus by the method of Bartlett (23).
Cholesterol and cholesteryl esters were extracted with etha-no1:hexane (40:60, v/v) and were separated by thin layer chromatography on Silica Gel IB2 in a solvent system of hexane:diethyl etheracetic acid (5050:1, v/v). Phospholipids were extracted with 10fold excess (by volume) chloroform:methanol (2:1, v/v). Individual phospholipids were separated by thin layer chromatography on Silica Gel IB2 in a solvent system of ch1oroform:methanol:acetic acidwater (25:154:2, v/v). Each phospholipid was identified by comparison with a commercially available standard. A l l lipids were detected on chromatography plates with iodine vapor. Areas containing lipids were marked and scraped off into scintillation vials after the iodine stain had disappeared, and the radioactivity was determined by liquid scintillation in 10 ml of Aquasol-2. Phospholipids on silica gel were solubilized with 0.1 ml of chloroform and the amount of phospholipid in each band was determined as phosphorus.
Flat-bed electrofocusing was performed in Ultradex as described by LKB (Application Note 198) with ampholines pH 3.5-8.0; the temperature was maintained at 4 "C. Protein bands were visualized with Coomassie brilliant blue R-250 following transfer of the protein pattern to a fiiter paper strip. Zones of the gel containing protein were scraped into 5-ml columns and the protein was eluted from the gel with Tris-buffered saline. Gel electrophoresis was performed in tubes of 7.5% acrylamide containing 0.25% N,N"methylenebisacrylamide and 0.1% SDS as described by Weber and Osborn (24) and in slabs of a gradient of 3-20% acrylamide in 6 M urea and 0.1% SDS essentially as described by Camejo and Socorro (25). Electrophoresis was performed for 6 h at 12 "C at a constant current of 10 mA/slab. Following electrophoresis, the gel was fixed in isopropyl alcoho1:acetic acid:water (251065, v/v/v), then stained in methanokacetic acidwater (5010:40, v/v/v) containing 0.05% Coomassie brilliant blue G. The gel was destained in isopropyl alcoho1:acetic acidwater (101080, v/v/v) and stored in 7% acetic acid. Gel scans were performed with a Zeineh soft laser scanning densitometer (Biomed Instruments). Protein samples were prepared in the presence of /3-mercaptoethanol. Molecular weights of proteins were estimated from a plot of log (molecular weight) versus mobility (distance of protein migration-distance of dye migration) using standards of known molecular weight.

RESULTS
The lipid transfer complex was isolated from the d > 1.21 g/ml infranatant of human and rat plasma by sequential chromatography on phenyl-Sepharose, CM-cellulose, and concanavalin A-Sepharose essentially as outlined by Ihm et al. (13) except that /3-mercaptoethanol was added to the d > 1.21 g/ml infranatant at a final concentration of 50 mM prior to the first purification step. This modification of the procedure alters the behavior of LTC on ConA-Sepharose so that the major active fraction is well separated from the peak containing most of the ConA-bound proteins, resulting in an LTC preparation with high specific activity. The elution of human and rat LTC from ConA-Sepharose is illustrated in Fig. 1. In the case of human lipid transfer activities, cholesteryl ester and phosphatidylcholine exchange activities co-elute from ConA-Sepharose. Stoichiometry of exchange of neutral and polar lipid cannot be determined from the data of Fig. 1 since different assays were employed to assess exchange of the two lipids. Rat LTC has very little cholesteryl ester exchange activity.
Chromatographic fractions from ConA-Sepharose were combined to give two LTC preparations, one of low specific activity designated LTC-A (human, fractions 52-54; rat, fraction 47) and one of high specific activity designated LTC-B (human, fractions 58-67; rat, fractions 55-60). As is indicated in Fig. 1, lecithimcholesterol acyltransferase activity is pronounced in human LTC-A; although the acyltransferase activity is measurable in LTC-B, its amount is considerably diminished. The fractions containing rat 1ecithin:cholesterol acyltransferase activity are concentrated between fractions designated LTC-A and -B. With respect to phosphatidylcholine exchange as a basis for comparison, the fold-purification from the d > 1.21 g/ml infranatant of human plasma LTC-B is 10,000; the amount of lipid transfer activity recovered is 30%. Rat LTC-B is purified only about 500-fold with recovery of 11% of the total activity. In an LDL to HDL assay, purified human and rat LTC transfer typically 1-2 pmol and 300-800 m o l of phospholipid, respectively, per mg of protein in 3 h at 37 "C.
Electrophoresis of human and rat LTC-B in 7.5% polyacrylamide gels containing SDS produces a single broad stained band of an apparent molecular weight of 61,000; chromatography of human LTC-B on a calibrated Bio-Gel A-15m column results in one protein peak with an apparent molecular weight of 70,000 (data not shown). It was possible to obtain sufficient quantities of the human LTC-B to characterize it in more detail. The protein was subjected to gradient gel electropho-  Fig. 2. Two major staining bands account for ferase and less than 4% of the total protein is apo-D. The lipid r95% of the staining intensity; these bands, present in almost composition of LTC-B was determined and is presented in equimolar amounts, have apparent molecular weights of 58,- Table I   total. Although these calculations are certainly crude, three different LTC-B preparations had very similar lipid and protein compositions (indicated by the values of per cent error in Table I), suggesting that LTC-B is a lipid-protein complex with a relatively stable composition. It is intriguing that cholesterol and cholesteryl ester are present in this complex in nearly equimolar amounts. Amino acid analysis of two preparations of LTC-B indicates the following composition (mole %, subject to 5% uncertainty): Asp, 12%; Glu, 13%; Lys, 5%; Arg, 4%; Thr, 7%; Ser, 11%; Pro, 6%; Gly, 9%; Ala, 8%; Val, 5%; Met, 2%; Ileu, 3%; Leu, 9%; Tyr, 2%; Phe, 3%; and His, 1%. The content of cysteine and/or cystine has not been determined. Since LTC-B may be a mixture of at least two proteins, the importance of the amino acid composition is doubtful.
T o determine whether LTC facilitates net lipid transfer or lipid exchange in the assay system employed, LDL and HDL were incubated a t 37 "C for 3 h, either in the presence or the absence of human LTC. After incubation, LDL and HDL were separated at density 1.063 g/ml by ultracentrifugation. The amount of protein, phospholipid, and cholesterol (free plus esterified) recovered in the reisolated LDL and HDL was then determined. There is no significant alteration of composition of the reisolated lipoproteins as is indicated by the data of Table I1 obtained      as is the case for the more purified rat LTC-B. The flux rate ratios vary from 5.8 to 26.7, phosphatidy1choline:cholesteryl ester. The possibility exists that rat plasma contains an inhibitor of cholesteryl ester exchange. Rat cl > 1.21 g/ml infranatant, and the unbound-inactive and partially purified boundactive preparations obtained from the phenyl-Sepharose chromatography step were added to an assay containing ["ICE-HDL, LDL, and human cl > 1.21 g/ml infranatant (Table IV).
None of the fractions isolated from rat plasma inhibits exchange of cholesteryl ester catalyzed by human lipid exchange protein(s).
In the next series of experiments, the reaction conditions were varied in an effort to uncouple transfer of phosphatidylcholine and cholesteryl ester and to assess the importance of 1ecithin:cholesterol acyltransferase in lipid transfer. It was first established that the course of the reaction was lipid exchange and not net lipid transfer at all reaction conditions. The partially purified preparation of human LTC, LTC-A, was utilized in these studies since it contains sufficient lecithin:cholesterol actyltransferase activity to ensure accurate measurement (Fig. 1).
To determine the effect of ionic strength on the exchange of phosphatidylcholine and cholesteryl ester, the ionic strength of the reaction solutions was varied by addition of appropriate amounts of NaCl or CaC12. As is shown in Fig. 3, phosphatidylcholine and cholesteryl ester exchange and lecithin:cholesterol acyltransferase activities are similarly influenced by ionic strength. In NaCl buffer, both phosphatidylcholine and cholesteryl ester exchange activities decrease by about 20% from p = 0.3-0.9. The acyltransferase activity decreases steadily as the concentration of NaCl increases, and at an ionic strength of 0.9, 28% of the original lecithin: cholesterol acyltransferase activity is inhibited. When CaCl is used to adjust the solution ionic strength, the results are qualitatively similar although both lipid exchange activities and particularly 1ecithin:cholesterol acyltransferase activity are more sensitive to CaC12 relative to NaCl. Both phosphatidylcholine and cholesteryl ester exchange activities decrease by approximately 40% between ionic strengths of slightly greater than 0 to 0.3 (which corresponds to 0.1 M CaCl,); no further decrease in lipid exchange activity occurs at higher ionic strengths. Lecithin:cholesterol acyltransferase activity decreases rather drastically as the CaC4 ionic strength is increased to 0.9; only 50% of the original activity remains at  an ionic strength 0.3, and only 24% remains at p = 0.9. In these experiments, the mole ratio of phosphatidylcho1ine:cholesteryl ester exchanged at the various ionic strengths ranges from 0.9 to 1.2.
To test the effect of pH on facilitated phosphatidylcholine or cholesteryl ester exchange, assays were performed at pH 5 through pH 10. As is shown in Fig. 4, rates of facilitated phosphatidylcholine and cholesteryl ester exchange do not vary markedly over the pH range 7 through 9. Decreases in both activities occur at pH 5-7 and 9-10, with facilitated exchange of both lipids considerably diminished at pH 5. Lecithin:cholesterol acyltransferase activity, on the other hand, has a fairly sharp pH optimum, pH 7 . Relative to the optimum, 1ecithin:cholesterol acyltransferase is only 40% active at pH 6 and pH 9.
To investigate the heat stability of the lipid exchange activ-of Phosphatidylcholine and Cholesteryl Ester ities and of 1ecithin:cholesterol acyltransferase, partially purified human LTC-A was heated at 37-72 "C for 15 min prior to the routine assay. Results are presented in Fig. 5. The phosphatidylcholine exchange activity is not influenced by temperatures as great as 55 "C for 15 min. Activity decreases only slightly at higher temperatures such that the complex heated at 72 "C retains 89% of the activity of the nonheated material. Cholesteryl ester exchange activity is similarly heatstable. The complex heated at 62 "C retains 95% of its original activity; heating at 72 "C reduces the cholesteryl ester exchange to 84% of the original. In contrast to the lipid exchange activities, 1ecithin:cholesterol acyltransferase activity is very sensitive to elevated temperature. After 15 min at 55 "C, the acyltransferase loses 94% of the original activity. Incubation for 15 min at 62 "C or greater completely inactivates lecithin:cholesterol acyltransferase. The acyltransferase is sensitive to thiol group blockers (26). In addition, Hopkins and Barter (10) report that p-chloromercuriphenyl sulfonate added to rabbit serum inhibited transfer of triglyceride; the transfer of cholesteryl ester was much less sensitive. It was therefore of interest to assess the effect of a thiol group reagent on phosphatidylcholine and cholesteryl ester exchange and on 1ecithin:cholesterol acyltransferase. Thus, the standard assays were performed in the absence and presence of DTNB. The results are presented in  Finally, LTC-A obtained from the Cod-Sepharose step was subjected to preparative isoelectric focusing. The gel was sliced in pieces (0.8 cm X 10.5 cm) and the protein was eluted with 2.0 ml of Tris-buffered saline. The protein mixtures were incubated with ['4C]DPPC/[3H]CE-HDL and LDL to identify the isoelectric points of the phosphatidylcholine and cholesteryl ester exchange activities. The data presented in Fig. 7 indicate that both lipid exchange activities occur over the pH range 4.2-5.0. Over this range, the activities exist in equimolar stoichiometries.
To evaluate the possibility that coupled exchange of phosphatidylcholine and cholesteryl ester is dictated by the nature of lipoprotein substrates, phosphatidylcholine and cholesteryl ester exchange facilitated by LTC was compared to that facilitated by bovine liver phosphatidylcholine-specific exchange protein. The routine assay employing ['4C]DPPC/[3H] CE-LDL and HDL was conducted at 37 "C and at 44 "C. As is shown in Fig. 8, the liver PLEP greatly facilitates the transfer of phosphatidylcholine both at 37 "C and 44 "C, while it does not enhance the transfer of cholesteryl ester even at the higher incubation temperature. Under the same conditions in the presence of human plasma lipid exchange complex, exchange of both lipids occurs with a mole ratio of phosphatidylcholine to cholesteryl ester about 1:l. DISCUSSION In the lipoprotein assay system employed, LTC clearly catalyzes lipid exchange rather than net lipid transfer. In three consecutive steps of chromatography of LTC through a hydrophobic column, a cation exchange column, and a carbohydrate affinity column, phosphatidylcholine and cholesteryl  acetylglucosamine. In addition, LTC has associated lipid and must therefore be considered a lipid-protein complex. The lipid composition is surprisingly reproducible which suggests that it may be relevant to the biological function of this entity. LTC has an apparent molecular weight of 61,000 as judged by electrophoresis in 7.5% polyacrylamide-SDS. This value is consistent with that determined by chromatography of human LTC-B on Bio-Gel A-5m; the gel filtration molecular weight is approximately 70,000. In polyacrylamide gradient gels in urea/SDS, the 61,000 molecular weight band exists as two closely spaced bands of apparent molecular weights 58,000 and 63,000. LTC purified from rat d > 1.21 g/ml infranatant has an almost identical molecular weight based on electrophoresis in 7.5% acrylamide-SDS. It is not yet known whether the broad protein band of apparent molecular weight 61,000 resolves into a doublet when electrophoresis is performed in an acrylamide gradient. The question is an interesting one since the rat LTC does not appreciably facilitate exchange of cholesteryl ester. Lack of catalyzed exchange of the neutral lipid is probably not due to the presence of an inhibitor in rat plasma, although it is possible that an inhibitor is specific for the rat LTC.
Does LTC require 1ecithin:cholesterol acyltransferase to catalyze the transfer of cholesteryl esters? Since lecithin: cholesterol acyltransferase plays a role in both cholesteryl ester and phospholipid metabolism and is present in d > 1.21 g/ml plasma, the role of the acyltransferase in LTC-facilitated lipid exchange was investigated. Lecithin:cholesterol acyltransferase activity co-elutes with that of phosphatidylcholine and cholesteryl ester exchange activity through the fwst two steps of purification consistent with the report of Fielding and Fielding (27) that cholesteryl ester transfer activity is associated with 1ecithin:cholesterol acyltransferase. In the ConA-Sepharose step, however, lipid exchange activities separate from 1ecithin:cholesterol acyltransferase activity, although the fraction containing high specific activity LTC still has a small amount of acyltransferase activity. In any event, the data demonstrate conclusively that 1ecithin:cholesterol acyltransferase activity is not required for lipid exchange facilitated by LTC. First, 1ecithin:cholesterol acyltransferase is very sensitive to heat inactivation, whereas lipid exchange activities are stable to exposure to temperatures as great as 72 "C. Second, while lipid exchange activities do not vary markedly over the pH range 7 through 9, 1ecithin:cholesterol acyltransferase activity has a sharp pH optimum at pH 7. Finally, the sulfhydryl group reagent, DTNB, inhibits 1ecithin:cholesterol acyltransferase activity almost completely, while both lipid exchange activities are little influenced. Taken together, the data strongly indicate that active LCAT is not a prerequisite for phosphatidylcholine and cholesteryl ester exchange.
The most significant finding is that phosphatidylcholine and cholesteryl ester are exchanged with a 1:l stoichiometry by human LTC. For example, facilitated exchange occurs with about a 1:1 stoichiometry in the presence of increasing amounts of d > 1.21 g/ml infranatant plasma protein and of LTC purified to high specific activity. Phosphatidylcholine and cholesteryl ester exchange is also stoichiometric at all values of ionic strength and pH tested. The activities are present in equimolar stoichiometries in protein-containing bands obtained by preparative isoelectric focusing. Furthermore, the activities are not uncoupled by elevated temperatures or by thiol group reagents. Under conditions in which human LTC facilitates exchange of phosphatidylcholine and cholesteryl ester equally between LDL and HDL, bovine liver phospholipid exchange protein catalyzes exchange of phosphatidylcholine only, suggesting that stoichiometric exchange is not dictated by the structure of the lipoprotein substrates. Available data may be interpreted to indicate that exchange/ transfer of cholesteryl ester by LTC depends on the presence of phosphatidylcholine, but that facilitated exchange/transfer of phosphatidylcholine occurs independently of cholesteryl ester. In a liposome -+ mitochondria assay system, for example, LTC-facilitated phosphatidylcholine transfer occurs from liposomes to mitochondria; both assay substrates are deficient in cholesteryl ester (28). Furthermore, the rat LTC facilitates exchange of phosphatidylcholine under conditions in which little cholesteryl ester exchange occurs. Although cholesteryl ester transfer activity emerges as the fold-purification of LTC from rat d > 1.21 g/ml infranatant increases, it remains very low such that instead of the 1:1 stoichiometric exchange of phosphatidylcholine and cholesteryl ester facilitated by human LTC, the rat LTC catalyzes transfer in a mole ratio of about 17:l. Finally, human LTC removes cholesteryl oleate from a dioleoylphosphatidylcholine monolayer containing 1-6 mol '% cholesteryl 01eate.~ When a sphingomyelin monolayer contains the cholesteryl ester, the sterol is not removed by LTC. However, as increasing amounts of phosphatidylcholine are added to the sphingomyelin-cholesteryl ester monolayer, the rate of cholesteryl ester removal by LTC is markedly increased. Furthermore, phosphatidylcholine, as well as cholesteryl oleate, is removed from the monolayer by LTC, but removal of phosphatidylcholine does not require the presence of cholesteryl ester.
An important question which arises is: What is the biological role of the plasma LTC? Any hypothesis must accommodate the fact that human LTC facilitates equimolar exchange of phosphatidylcholine and cholesteryl ester while rat LTC does not. Resolution of this question must await the availability of additional information.