Lipoprotein products of lecithin: Cholesterol acyltransferase and cholesteryl ester transfer

Biochimica Elsevier

et Biophysics

Biomedical

541

Acta, 112 (1982) 541-556

Press

BBA 51193

LIPOPROTEIN PRODUCTS OF LECITHIN: CHOLESTEROL CHOLESTERYL ESTER TRANSFER HERBERT

G. ROSE and POLLY

Lipid Research Laboratory (Received December (Revised manuscript

ACYLTRANSFERASE

AND

ELLERBE

Veterans Administration

Medical Center, Bronx, NY 10468 (U.S.A.)

8th, 1981) received May 26th, 1982)

Key words: Lecithin: cholesterol acyltransferase;

Cholesterol ester transferase;

Lipoprotein;

HDL;

(Phosphatidylcholine

vesicle)

High-density lipoprotein substrates and products of human plasma lecithin: cholesterol acyltransferase have been labelled with radioisotopic cholesteryl esters in order to facilitate identification. [ 3H]Cholesteryl esters were formed by endogenous HDL,/VHDL enzyme (d > 1.125 g/ml) following incubation with mixed vesicles of phosphatidylcholine, unesterified cholesterol and 3H-labelled unesterified cholesterol. Transfer of labelled esters to acceptor lipoproteins (VLDL+LDL, d -=c1.063 g/ml) was employed to distinguish a hypothetical transfer complex. Separation of labelled HDL3 /VHDL was by gel-permeation chromatography. The results indicate that a subpopulation of labelled HDL, /VHDL cholesteryl esters (43-61s of total) were removed by VLDL/LDL during a 3 h transfer period and these derive from the smaller lipoproteins of the spectrum. HDL carrying non-transferable [ 3H]cholesteryl esters localize to the larger HDL,. Transfer rates were proportional to ratios of acceptor to donor lipoproteins. Net transfer of cholesteryl esters from the smaller HDL, also occurred, but was smaller in magnitude (about 10.5% of total). Acyltransferase assays indicated that enzyme distribution is skewed to larger-sized HDL,, suggesting that the non-transferable components might be lecithin: cholesterol acyltransferase-containing parent complexes, while the smaller transfer products contain little acyltransferase. The results fit the hypothesis that a parent HDL,-lecithin: cholesterol acyltransferase complex generates a smaller-sized lipoprotein product which is active in cholesteryl ester transport.

Introduction

Evidence for these processes has been derived largely from in vitro experiments, but in vivo studies in human subjects are in good agreement [3,4]. Transfer rates are dependent upon specific exchange proteins and the ratios of acceptor to donor lipoproteins [4,5]. Even though transfer occurs when lecithin : cholesterol acyltransferase is inhibited [6], the enzyme may play a role in the regulation of transport [6,7]. Lipoprotein comcontaining lecithin : cholesterol plexes acyltransferase or produced by lecithin : cholesterol acyltransferase could serve as the principal transport vehicles. Indeed, Fielding and Fielding [8] have reported data in keeping with a cholesteryl

Plasma lecithin : cholesterol acyltransferase (EC 2.3.1.43) exists in a lipoprotein complex in the HDL, fraction of human plasma, with HDL, lipids serving as substrates. Cholesteryl esters generated by the enzyme distribute among all lipoprotein classes by transfer and exchange mechanisms [ 1,2].

Abbreviations:VLDL, very-low-density lipoprotein; LDL, low-density lipoprotein; HDL, high-density lipoprotein; VHDL, very-high-density lipoproteins (d 1.121- 1.250 g/ml); PC, phosphatidylcholine; DTNB, 5,5-dithiobis nitrobenzoic acid; PPO, 2,5-diphenyloxazole. OOOS-2760/82/0000-0000/$02.75

Q 1982 Elsevier Biomedical

Press

548

ester transfer complex consisting of lecithin : cholesterol acyltransferase, apolipoprotein A-I, apolipoprotein D and HDL lipids in stoichiometric proportions. In a recent review of the relationship between HDL and plasma lecithin : cholesterol acyltransferase [9], we postulated the existence of a parent H D L ,-lecithin : cholesterol substrate acyltransferase complex and two lipoprotein products, viz, a small primary product which is active in exchange reactions and a larger, more inert particle, comprising part of the HDL,. Lecithin: cholesterol acyltransferase would play a necessary role in exchange reactions by controlling the pool size of the reactive product lipoproteins. The experiments described in the present report constitute an attempt at localizing a reactive lecithin : cholesterol acyltransferase product lipoprotein within the size distribution of the human plasma HDL, and VHDL (d> 1.125 g/ml fraction) by gel-permeation chromatography. These experiments are based on earlier work from this laboratory indicating that radioactive cholesteryl esters formed by endogenous lecithin : cholesterol acyltransferase when plasma is incubated with unilamellar PC vesicles carrying radioisotopic free cholesterol can exchange to equilibrium among lipoproteins [l], and that the probable HDL site for ester transport was the HDL, and VHDL subfractions [4]. An in vitro system to assess transfer has been devised consisting of a radioisotopic cholesteryl ester-generating system (lecithin : cholesterol transferase-HDL,-radioactively labelled mixed vesicles of PC and free cholesterol) and acceptor lipoproteins (VLDL + LDL) for the cholesteryl esters formed. The results of these studies are compatible with the presence within human of a parent lecithin : cholesterol HDL, acyltransferase substrate complex of larger size than a product lipoprotein which is highly reactive in cholesteryl ester transport. Materials and Methods Experimental design. The basic incubation experiments for evaluation of cholesteryl ester transfer consisted of two parts: firstly, a 1 h labelling incubation at 37°C in which radioisotopic cholesteryl esters were incorporated into the plasma

HDL, and VHDL (d > 1.125 g/ml fraction) and, secondly, after labelling a VLDL + LDL fraction (d < 1.063 g/ml) is added as acceptor for HDL esters as measured after a 3 h transfer period. During this time interval lecithin : cholesterol acyltransferase is inhibited by Paraoxon (diethyl p-nitro phenyl phosphate, K&K Laboratories, Plainview, NY) at a concentration of 1.4 mM. Preparation of vesicles. Single bilayer PC/unesterified cholesterol (10 : 1, mol/mol) were prepared without sonication in standard Tris buffer (50 mM Tris, 0.15 M NaCl, 2.0 mM EDTA, 200 mg/l sodium azide, pH 7.40) as described by Soutar et al. [lo], based on the original procedure of Batzri and Korn [ll]. Egg PC was obtained from Applied Sciences Laboratories, State College, PA. The final clear vehicle solution contained 1.2 mM PC and 0.12 mM unesterified cholesterol. Radioactive [7(n)-3H]cholesterol, lo-25 Ci/mmol (New England Nuclear, Boston, MA) was incorporated into the ethanolic solution used to prepare the vesicles to provide approximately 7.5 . 10” dpm per ml of vesicle solution. Thin-layer chromatography indicated that 99.0% of 3H label was in the unesterified cholesterol region and under 0.2% in the ester cholesterol region. Less than 4% vesicle 3H-labelled unesterified cholesterol was recovered in the void fraction of a Sepharose 4B column, and essentially all eluted in the void volume of a Biogel A 1.5 M column (Bio-Rad Laboratories, Richmond, CA). Vesicles prepared by this procedure have a mean diameter of about 26.5 nm [ 111. Lipoprotein preparation. The lipoprotein fraction containing the endogenous HDL,-lecithin : cholesterol acyltransferase complex was prepared by ultracentrifugation at d 1.125 g/ml of freshly obtained, post-absorptive human plasma (EDTA, 1.0 mg/ml) from nomolipidemic subjects after obtaining informed consent. 7 ml of plasma was raised to d 1.125 g/ml with solid KBr/NaCl and overlayered with 4.0 ml Tris buffer made to the same density. Tubes were centrifuged in a 50 Ti rotor in a Beckman L3-50 ultracentrifuge for 24 h at 40000 rpm (106500 X g). Another sample of autologous plasma at d 1.063 g/ml was centrifuged to isolate the combined VLDL/LDL fraction. After tube slicing, the d 1.125 g/ml infranate and d 1.063 g/ml supernate were dialyzed against standard Tris buffer and concentrated over a PM-

549

10 filter in an Amicon cell to 1.2 and 3.5 times original plasma level, respectively. Essentially all plasma lecithin : cholesterol acyltransferase was recovered in the infranate and no activity could be detected in the supernate. Loss corrections and adjustments relative to original plasma were made using triacylglycerol levels as a marker for the VLDL/LDL fraction and plasma protein levels as a marker for the d 1.125 g/ml infranate. Adequacy of separation and purity of the final products was monitored by lipoprotein electrophoresis on agarose films (Pfizer Diagnostics, New York). Incubation conditions. Labelled vesicles (0.2 vol.) were added to the d > 1.125 g/ml fraction for the labelling incubation and 0.43 vol. of VLDL/LDL lipoprotein solution, made to various dilutions as appropriate, for the transfer period. As a control experiment, radioactive vesicles were incubated with partially purified (lOOO- 1600 X ) human lecithin : cholesterol acyltransferase prepared according to the method of Soutar et al. [lo]. The enzyme preparation contained 0.2- 1.4 mg protein per ml and was free of lipoproteins. It actively esterified vesicle unesterified cholesterol without supplemental apolipoprotein cofactor. Incubation times were varied in order to obtain levels of esterification varying from 25 to 38%. Vesicles (0.2 vol.) so treated were equilibrated with the d > 1.125 g/ml fraction in the presence of Paraoxon for 1 h, followed by addition of VLDL/LDL for the 3 h transfer period. As a further control experiment [ l-‘4C]cholesteryl oleate (New England Nuclear, Boston, MA) was incorporated into vesicles at 1 mol% of the PC concentration, followed by 1 h equilibration with d 1.125 g/ml infranate in the presence of 1.4 mM 5,5’-dithiobis nitrobenzoic acid (K&K Laboratories, Plainview, NY). Analytical procedures. Assay of lecithin : cholesterol acyltransferase activity in HDLJVHDL subfractions separated by agarose column chromatography on Bio-Gel A 1.5 M was accomplished by incubation of eluates, vol. for vol., with vesicle solution and esterification of 3H-labelled unesterified cholesterol determined at 1 and 6 h. Activity was linear for 1 h. Assays of activity with endogenous substrats were also performed by incubation of pooled eluates and determination of unesterified cholesterol consumption as previously described [ 121. Rates were linear for 3 h. Samples of

incubation mixtures, before and after the transfer period, were filtered through agarose columns (Bio-Gel A 1.5 M, 2.0 X 90 cm) in standard Tris buffer in order to separate HDL, from VLDL/LDL or unbound vesicles. In all experiments, 3.5 ml of labelled incubation mixture diluted with 1.5 ml Tris buffer, was loaded onto the column. Fractions (2.5-4.5 ml) were collected with an automatic fraction collector and extracted with chloroform/methanol (2 : 1, v/v). In some studies separation was also achieved using heparin-Mn2+ precipitation as described by Warnick and Albers [ 131. cholesterol was determined by ophthaldialdehyde method of Rude1 and Morris [14]. Ester and free cholesterol were separated on thin-layer plates of silica gel H using hexane/diethyl ether/acetic acid (90: 10: 1, v/v). The position of each sterol was identified with marker compounds in separate channels by staining with iodine vapor. Spots were scraped and eluted with chloroform. Recovery of radioactivity from thin-layer plates was 88.2 2 3.0% (S.D.). Recovery of radioactive cholesterol from agarose columns was 92.1 2 7.4% (S.D.), without consistent differences for esterified and unesterified isotope. Aliquots taken for radioactivity were counted in 10 ml PPO/toluene (0.4% PPO) using a Beckman LS-100 scintillation counter. Triacylglycerols were determined by an enzymatic method (Dow Chemical Co.) and protein by the method of Lowry et al. [15] with bovine serum albumin as standard. Results Molecular size distribution of HDL,/ VHDL [ ‘H ] cholesterol after labelling with endogenous lecithin: cholesterol acyltransferase and localization of transferable components Four experiments were performed in order to evaluate the pattern of HDL, labelling by endogenous HDL,-bound lecithin : cholesterol acyltransferase, using radioisotopic vesicles for labelling. Following the 1 h labelling incubation, VLDL/LDL was added at a ratio to HDL, of 1.5 : 1.0 to assess the size distribution of labelled lipoproteins which are transferred. Esterification of vesicle 3H-labelled unesterified cholesterol during the 1 h labelling period varied from 7.3 to 15.1%. A representative study is depicted in Fig. 1.

25

30

35

40

45

35 FRACTION

40

45

B

61

0

25

30

50

Fig I. Gel separation of vesicle radioactive cholesterol after labelling of d > 1.125 g/ml lipoproteins by endogenous lecithin : cholesterol acyltransferase, before and after a transfer period. Panel A shows distribution of 3H-labelled unesterified cholesterol before (Cl) and after transfer ( n ). Panel B shows 3H-labelled ester cholesterol distribution before (0) and after Results of subtraction of after-transfer transfer (0 -0). curve from before-transfer is drawn, (O- - -0). Panel A also depicts column position of HDL, by total cholesterol content (X).

The distribution of 3H-labelled unesterified cholesterol after the labelling period indicated that about 2/3 of the isotope is associated with HDL, (Fig. 1, panel A, open symbols), while l/3 remained in the column void. Separate experiments with freshly prepared vesicles alone showed that virtually all vesicle [3H]cholesterol is eluted in the void peak. After the 3 h transfer period, a marked reduction in HDL, 3H-labelled unesterified cholesterol had resulted, amounting to 61.0% of the pre-transfer HDL, level (closed symbols). In four studies removal varied from 52 to 78%, with corresponding increases occurring in the void peak along with VLDL/LDL. Removal was more extensive from hind-limb components. After the labelling period, the void peak was devoid of ‘H-labelled ester cholesterol in every

experiment (Fig. 1, panel B, open symbols) and all radioactivity was uniformly associated with the HDL, peak. 61% of 3H-labelled ester cholesterol was removed from HDL, during the transfer period and appeared in the void peak (panel B, closed symbols). In four experiments, fractional removal varied from 43 to 61%. Of significance is that this removal was distinctly in favor of the smallest HDL,/VHDL components. At the exall of the hind-limb labelled esters tremes, transferred and none from the fore limb. When the post-transfer curve was subtracted from pretransfer, a curve representing the transferable components could be drawn (Fig. 1, panel B). Selective transfer was also observed when HDL was labelled in whole plasma, followed by transfer to endogenous VLDL/LDL after addition of Paraoxon. Thus, ester cholesterol-labelled HDL,/VHDL can be resolved into two components differing in molecular size after labelling by endogenous lecithin : cholesterol acyltransferase. Transfer of ‘H-labelled unesterified cholesterol and ‘H-labelled ester cholesterol from HDL,/VHDL to VLDL/LDL after labelling by equilibration with cholesteryl ester-carrying radioactive vesicles In order to determine whether selective transfer of [ 3H]cholesterol esters from labelled hind-limb lipoproteins was dependent on endogenous lecithin: cholesterol acyltransferase, two control experiments were performed. In the first, vesicles were incubated with purified lecithin : cholesterol acyltransferase for 3 h to yield 25.0% esterification of 3H-labelled unesterified cholesterol and then incubated for 1 h with d > 1.125 g/ml lipoproteins and Paraoxon. HDL, becomes labelled as depicted in Fig. 2. Of the 3H-labelled unesterified cholesterol, 31.4% was in the column void and 68.6% in the gel (Fig. 2, panel A, open symbols). For 3Hlabelled ester cholesterol, 89.1% associated with the HDL, in a uniform manner and 10.9% remained in void particles (Fig. 2, panel B, open symbols). After the labelling period, transfer to VLDL/LDL was evaluated by addition of the d c 1.063 g/ml fraction. The results are shown in Fig. 2, panels A and B. For 3H-labelled unesterified cholesterol, (Fig. 2, panel A, closed symbols) there occurred a large reduction in HDL, ‘Hlabelled unesterified cholesterol, amounting to

551

25

0

30

35

40

45

50

55

6

iz.o[, AK !t

0

25

30

HDL3

35

40

45

50

55

In a second group of experiments, the labelling and transfer pattern of HDL, was determined after incorporation of [ “C]cholesteryl oleate directly into the vesicle preparation. In these experiments equilibration with the d > 1.125 g/ml fraction was accomplished over 1 h with (1.4 mM) DTNB as lecithin : cholesterol acyltransferase inhibitor. Transfer from HDL, to VLDL/LDL over 3 h was also tested in the presence of DTNB. The extent of “C-labelled ester cholesterol transfer was comparable to that observed with lecithin : cholesterol acyltransferase-treated vesicles and Paraoxon as inhibitor. Transfer was extensive from both ascending and descending limb fractions, without evidence of selective removal. Thus, selective transport could not be demonstrated unless radioactive cholesteryl esters were formed by endogenous lipoprotein-bound lecithin : cholesterol acyltransferase. Cholesteryl ester transfer from HDL,/ VHDL Two transfer experiments were performed to determine if 3H-labelled ester cholesterol transfer

FRACTION

Fig. 2. Gel separation of lecithin : cholesterol acyltransferasetreated vesicles after 1 h equilibration with d > 1.125 g/ml fraction and after 3 h transfer period with VLDL/LDL present. The column distribution of ‘H-labelled unesterified cholesterol is shown in panel A, after both labelling of HDL,/VHDL (O), and transfer to VLDL/LDL (W). The distribution of 3H-labelled ester cholesterol is shown in panel B after labelling (0) and after transfer (0). The position of HDL, is indicated in panel A by total cholesterol content (.---

0).

72.1% with a corresponding increase in void volume isotope. Removal of isotope from HDLJVHDL was uniform across the spectrum of lipoprotein molecular diameters. Similarly, there resulted extensive transfer of 3H-labelled ester cholesterol from HDL, to the void peak, amounting to 56.3% of HDL-bound label (Fig. 2, panel B, closed symbols). This removal, like that for 3H-labelled unesterified cholesterol, was generally uniform, but was more extensive for the fore-limb components. Selective removal of smaller HDL, components was not seen, indicating that lipoprotein-bound lecithin : cholesterol acetyltransferase is a requirement for selective transport.

0

I5

20

25

30

36

40

45

50

FRACTION

Fig. 3. Agarose column analysis of HDL,/VHDL cholesteryl ester changes resulting from incubation with VLDL/LDL. Distribution of HDL, cholesteryl esters on Bio-Gel A 1.5 M after l-h incubation with endogenous lecithin : cholesterol acyltransferase (0) and after 3 h in presence of VLDL/LDL (1.5 X original plasma level) plus Paraoxon (O0). Contamination of HDL, fractions by void peak VLDL/LDL is shown (O- - -0) as determined by fractionation of the identical quantity of VLDL/LDL in the absence of HDL,.

552

from descending limb HDL, particles was accompanied by net transfer of cholesteryl ester mass. One of these experiments is illustrated in Fig. 3. Prior to the 3 h transfer period, only a small cholesteryl ester peak was present in the void, followed by the large HDL, ester peak in the gel (open symbols). After transfer, the large VLDL/LDL ester peak appeared in the void, with addition of esters to fore-limb fractions of the HDL, and subtractions from the hind-limb fractions (closed symbols). When an identical quantity of VLDL/LDL was filtered alone, a trail of cholesteryl esters extended into the gel, which was sufficient to account for the increases in HDL, fore-limb cholesteryl esters. That these fore-limb increments, in whole or large part, represent contamination was also supported by independent measures of transfer from HDL, by the heparinMn2+ procedure which indicated net loss of HDL, esters (vide infra). Removal from individual hindlimb fractions ranged from 43 to 69% of the pre-transfer levels. The total quantity removed, after correction for contamination, as a pecentage of the HDL, was 16.1% in this study and 12.9% in the repeat study (mean, 14.6%). Since contamination could not be entirely avoided in view of the large quantity of cholesteryl esters present in the VLDL/LDL, the possibility that non-transfer of forelimb 3H-labelled ester cholesterol (Fig. 2, Panel B) was a consequence of contamination by labelled VLDL/LDL esters was considered. This explanation is unlikely for the following reasons: firstly, contamination by VLDL/LDL ester and free cholesterol would not explain selective non-transfer of forelimb 3H-labelled ester cholesterol without concomitant non-transfer of 3H-labelled unesterified cholesterol. Secondly, calculating possible 3Hlabelled ester cholesterol contamination from forelimb ester mass increments and void peak specific radioactivities indicated that insufficient 3Hlabelled ester cholesterol would contaminate forelimb fractions to explain apparent failure of transfer and, thirdly, contaminating 3H-labelled ester cholesterol would be expected to vary from experiment to experiment and net addition of 3H-labelled ester cholesterol might have resulted in at least some of the four transfer experiments. Yet, this was not observed. Lastly, the control experiments in which HDL, was similarly labelled but, by

1

30

60

SO

120

150

180

30

60

so

120

150

I 180

e y ;, $

60 0 TIME IN MINUTES

Fig. 4. Rate measurements for transfer of cholesteryl esters from HDL,/VHDL to VLDL/LDL. Time curves for transfer of 3H-labelled ester cholesterol expressed as percentage of initial ‘H-labelled ester cholesterol associated with HDL, at each time point, are shown for a representative experiment in panel A. Varying ratios of VLDL/LDL:HDL,/VHDL are indicated: X, 0.5; W, 1.0; 0, 1.5. Curves are fitted by a computer to a single exponential function. The same data are shown in panel B for ester cholesterol mass. Results of a single representative experiment are presented. EC, ester cholesterol.

equilibration with cholesteryl-ester-carrying vesicles, failed to show non-transfer from fore-limb particles. Therefore, the lack of 3H-labelled ester cholesterol transfer from fore-limb HDL, fractions cannot be explained as the result of incomplete separation of VLDL/LDL from HDL, cholesteryl esters. It appears that 3H-labelled ester cholesterol transfer from small HDL, is associated with mas transfer from these same fractions. Transfer of isotope exceeded that of ester mass, since specific radioactivities of the hind-limb components fell after the transfer period in these experiments.

553

Measurement of HDL, radioisotopic cholesteryl ester transfer rates using heparin-Mn2’ precipitation In order to determine rates of transfer of HDL, lipids to VLDL/LDL, HDL, must be re-isolated at multiple time points. This can be accomplished by agarose gel filtration or by heparin-Mn2+ precipitation. The latter procedure would be expected to precipitate VLDL/LDL, allowing sampling of the supernatant HDL,/VHDL. In order to assess the utility of the heparin-Mn*+ procedure, four experiments were performed comparing transfer at the end of 3 h, using the agarose filtration method as the reference procedure. In these experiments, individual comparisons of 3H-labelled ester cholesterol 3-h fractional transfer rates showed slightly higher values by the precipitation method (+ 6.7%, but when total 3H-labelled ester cholesterol transfer rates were compared, agreement was excellent (-0.9%). Three experiments were then performed to determine the time course of HDL, 3H-labelled ester cholesterol and cholesteryl ester mass transfer to VLDL/LDL. In each experiment, measurements were made of transfer rates at three different ratios of VLDL/LDL to HDL,, viz, 1.5, 1.0 and 0.5 times original plasma levels. Fig. 4, panel A, shows the time curve for transfer of ‘H-labelled ester cholesterol in one representative experiment. The slopes of these curves could be fitted to exponential functions with good agreement (R’, 0.82-0.99). As illustrated in the figure and in Table I, the transfer rates were proportional to the ratios of VLDL/LDL to HDL,. Transfer of 3Hlabelled ester cholesterol was rapid (16.6 ? 4.6% (S.E.) per h) at the ratio of unfractionated plasma

TABLE

Localization of lecithin: cholesterol acyltransferase in the HDL,/ VHDL fraction Four experiments were performed to localize the position of plasma lecithin : cholesterol acyltransferase within the spectrum of HDL, as fractionated by Bio-Gel A 1.5 M. In two experiments the d > 1.125 g/ml fraction was employed and in two whole plasma. Whole plasma was studied because of possible translocation of lecithin : cholesterol acyltransferase resulting from ultracentrifugation. Lecithin : cholesterol acyltransferase was determined in column fractions by assay with unequilibrated vesicles as substrate as well as by fall in endogenous unesterified cholesterol level during a 3 h incubation. Recovery of activity from columns ranged from 74 to 86%.

I

RATES OF HDL,/VLDL

TRANSFER OF TO VLDL/LDL

Rates were determined for three experiments. Ratio:

(i.e., 1: 0) and increased to 21.8 * 4.6% at a ratio of 1.5 : 1.0. Transfer at a ratio of 0.5 slowed to 9.2 * 2.7%/h. In contrast, transfer of cholesteryl ester mass was much slower (Fig. 4, panel B and TableI) and fit with an exponential function with variable precision, particularly at the lower ratio of acceptor lipoproteins ( R2, 0.06-0.78). For transfer of mass, an influence of VLDL/LDL : HDL, ratios was suggested but not certain. Transfer of 3Hlabelled ester cholesterol was 4.5-6 times more rapid than that of mass. This result for the total HDL, agrees with that from column experiments, and suggests that endogenous lecithin : cholesterol acyltransferase during a short labelling period labels cholesteryl esters in HDL, particles that are preferentially predisposed to transfer their constituent esters.

VLDL/LDL:

3H-LABELLED

using heparin-Mn’+

HDL,/VHDL

ESTER

precipitation

CHOLESTEROL

procedure

3 H-labelled ester cholesterol

AND

for HDL separation

CHOLESTERYL

as described

ESTER

21.8k4.6 16.6ir4.6 9.2t2.1

FROM

in text. Values are mean k S.E.

Ester cholesterol

Ester cholesterol

(g/h)

(PM/I

3.5 kO.5 2.6kO.9 2.1kO.7

28.4” 3.9 20.9k6.9 17.Ok5.4

per h)

(g/h) 1.5 1.0 0.5

MASS

554

r25D HDL

I

FRACWN

Fig. 5. Distribution of human plasma lecithin: cholesterol acyltransferase on a Bio-Gel A 1.5 M column loaded with whole plasma. The position of HDL is indicated by total cholesterol content of isolated fractions and plotted with the vertical scale on the right (X). Lecithin : cholesterol acyltransferase activity in each fraction (0) is plotted with the vertical scale on the left. Enzyme activity was determined with radioactive PC/3H-labelled unesterified cholesterol vesicles in each fraction and expressed in units. A unit is defined as 1% esterification during the 6 h incubation period.

Both assay procedures gave similar results. Fig. 5 shows these results in an experiment with whole plasma and radioisotopic vesicle assay. Although some lecithin: cholesterol acyltransferase activity was demonstrable throughout the spectrum of HDL/VHDL, the peak activity occurred slightly before the HDL peak and the largest part of the activity was localized to the ascending limb. Similar results were obtained with the d > 1.125 g/ml lipoproteins. The smallest HDL, which were most active in cholesteryl ester transfer to acceptor lipoproteins, contained little enzymatic activity. Discussion Incubation of unilamellar lecithin vesicles with plasma or HDL results in transfer of PC to the HDL [16-211. This interaction is dependent upon temperature, incubation time and ratio of reactants [ 17,201. When the phospholipid vesicles also carry free cholesterol or cholesteryl esters, these lipids are transferred to the HDL, the rate of cholesterol ester transport exceeding that of lecithin [ 17,211. In previously reported experiments, we utilize radioactively labelled unesterified cholesterol incorporeated into unilamellar lecithin vesicles as substrates for endogenous HDL-bound

plasma lecithin : cholesterol acyltransferase [ 11. After a 3 h labelling incubation, plasma cholesteryl ester specific radioactivities were highest in the HDL, and subfractionated by agarose filtration showed peak values in HDL subfractions containing the smallest lipoproteins, suggesting localization of enzyme to these particles. It was further demonstrated that the labelled cholesteryl esters reached equilibrium among esters in each lipoprotein class after 19-h incubation in the presence of Paraoxon. Studies of lipoprotein composition changes during alimentary lipemia in normal volunteers suggested that transfer of cholesteryl esters to VLDL was greatest from the VHDL subfraction and, to a lesser extent, from HDL, [4]. Companion in vitro experiments substantiated this interpretation and also revealed that HDL, esters do not participate in transfer [4]. Since transfer was most active from HDL subfractions which were thought to contain the most active lecithin : cholesterol acyltransferase substrates, the possibility that lecithin: cholesterol acyltransferase product lipoproteins were involved in transfer was suggested [4,9]. Furthermore, it seemed probable that lecithin : cholesterol acyltransferase-generated labelled cholesteryl esters might serve as a useful marker for parent and product lipoproteins. The value of the ester label as a marker is enhanced by the finding in the present report that labelled esters are localized to HDL,/VHDL rather than remain in the vesicle bilayer. This finding simplified interpretation of transfer experiments and supports the validity of the label. In the current study we have sought to determine if the early-time lecithin : cholesterol acyltransferase product lipoproteins include particles smaller than the parent lecithin : cholesterol acyltransferase-HDL, complex and if such products are preferentially disposed to transfer labelled chlolesteryl ester to acceptor lipoproteins. While the cholesteryl ester-labelled lipoprotein products of lecithin : cholesterol acyltransferase appeared to distribute uniformly across the size spectrum of HDLJVHDL, as fractionated by molecular sieve chromatography, transfer to VLDL/LDL during a 3 h period was selective for the smaller components, with no appreciable transfer from the larger. Thus, the labelled HDLJVHDL were subdivided into two components, based on transfer to accep-

55.5

tor lipoproteins. Control experiments with vesicles carrying labelled cholesteryl ester, either resulting from treatment with purified lecithin: cholesterol acyltransferase or from incorporation of radioactive cholesteryl oleate, which are equilibrated with HDL,/VHDL, could not replicate this result. This was true both with Paraoxon and DTNB as inhibitors during the transfer period. The same selectivity of transport was also demonstrable from whole plasma, eliminating the possibility of artifacts produced by ultracentrifugation. At the elevated ratio of VLDL/LDL to HDL, in these experiments (i.e. 1.5 : 1.O), it was possible to show that isotope transfer from the smallest HDLJVHDL is associated with ester mass transfer (mean, 14.6% of total HDL) from among these same components. This figure agrees reasonably well with net transfer from HDL,/VHDL, determined after heparinMn*+ precipitation of VLDL/LDL, during 3 h, which averaged 10.5% in three experiments (TableI). In in vivo studies of cholesteryl ester turnover in humans, Barter and Connor 1221concluded that an HDL-ester cholesterol subpool comprising 5-20% of the total fraction was a rapidly-turning-over subfraction. Since the fraction of isotope transferred in our experiments is 4.5-6.0 times greater than mass, it is clear that the specific radioactivity of the transferred particles is higher than the mean and that the remaining non-transferred small components would be at reduced specific activity, which was, in fact, observed. The isotope, therefore, is concentrated in small-sized particles which demonstrate selective net transport of constituent esters to acceptor lipoproteins. This conclusion is strengthened by experiments showing selective removal of cholesteryl ester mass from among these same small components. When the size distribution of lecit~n:cholesterol acyltransferase bound to HDL,/VHDL or to HDL in whole plasma was determined by isotopic vesicle assay and unesterified cholesterol consumption assay, there was little enzymic activity in the small particle region, with a skewed distribution favoring the larger sized HDL,. These findings are consistent with the original hypothesis that an lecithin : cholesterol acyltransferase product lipoprotein of smaller size than the parent lecithin : cholesterol acyltransferase complex is

formed through activity of the enzyme and that this product is characterized by predisposition for transfer to acceptor lipoproteins. Implicit in this interpretation is the assumption that the labelled cholesteryl esters, which are lecithin : cholesterol acyltransferase product lipids, are localized in lecithin: cholesterol acyltransferase product lipoproteins. Since transfer of esters between HDL particles is likely [S], pre-existing particles, whether generated by lecithin: cholesterol acyltransferase or not, might become labelled. Furthermore, extensive transfer of esters incorporated by equilibration with cholesteryl ester-containing vesicles, although not specific for the smaller HDL, suggests either that the transferable lecithin : cholesterol acyltransferase product lipoproteins are preferentially labelled by this non-enzymatic means or that lecithin: cholesterol acyltransferase product lipoproteins do not play a unique role in transport of cholesteryl esters. Since PC dispersions have an affinity for lecithin : cholesterol acyltransferase in the lipoprotein-free fraction of plasma [23], it may be that PC vesicles selectively bind to lecithin : cholesterol acyltransferase-containing lipoproteins and thereby label pre-existing transfer lipoproteins. Alternatively, the HDL-bound cholesteryl ester-carrying vesicles may be loosely held and as readily removed as esters incorporated by endogenous lecithin: cholesterol acyltransferase. In any case, the existence of small-sized lecithin : cholesterol acyltransferase product lipoproteins remains to be demonstrated and experiments are in progress towards this end. Rates of transfer of ester cholesterol from HDL, to VLDL/LDL have been determined for radioactive and non-radioactive esters (Table I). Labelled ester transfer is rapid, and increased with increasing ratios of acceptor lipoproteins to HDL,. Transfer of ester cholesterol mass is much slower and is perhaps also influenced by acceptor lipoprotein levels. In Table I, mass rates are also expressed in terms of pM/l per h for comparison with lecithin: cholesterol acyltransferase rates and transfer rates reported in the literature. Plasma lecithin: cholesterol acyltransferase activity is from 45-98 PM/I per h in normal male subjects [ 121. In several reports [3,24,25], rates of ester transfer and ester exchange are close to that of lecithin : cholesterol acyltransferase, suggesting that these

556

processes are coupled. In the present report, ester transfer, at a ratio of VLDL/LDL: HDL, of 1.0, was 20.9 * 11.9 pM/l per h, which is appreciably lower than plasma level. However, lecithin : cholesterol acyltransferase activity in the isolated 1.125 g/ml infranate is 30-50% that of plasma (Ellerbe and Rose, unpublished data) and agrees well with transfer from this fraction. This finding suggests that removal of lipoproteins less dense than d 1.125 g/ml also reduced both lecithin : cholesterol activity and transfer rates pari passu. This observation is compatible with the view that lecithin : cholesterol acyltransferase product lipoproteins generated during the labelling incubation influence transfer rates. Acknowledgements This work was supported by the Service of the Veterans Administration.

Research

References 1 Rose, H.G. (1978) Stand. J. Clin. Lab. Invest. 38, Suppl. 150,91-97 2 Barter, P.J., Ha, Y.C. and Calvert, G.D. (1981) Atherosclerosis 38, 165-175 3 Nestel, P.J., Reardon, M. and Billington. T. (1979) Biochim. Biophys. Acta 573, 403-407 4 Rose, H.G. and Juliano, J. (1979) J. Lipid Res. 20, 399-407 5 Barter, P.J. and Jones, M.E. (1980) J. Lipid Res. 21, 238-249 6 Nichols, A.V. and Smith, L. (1965) J. Lipid Res. 6, 206-210 7 Glomset, J.A., Norum, K.R. and King, W. (1970) J. Clin. Invest. 49, 1827-1837

8 Fielding, P.E. and Fielding, C.J. (1980) Proc. Natl. Acad. Sci. U.S.A. 77, 3327-3330 9 Rose, H.G. (1981) in High Density Lipoproteins (C.E. Day, ed.), pp. 214-263, Marcel Dekker, New York 10 Soutar, A.K., Pownall, H.J., Hu, A.S. and Smith. L.C. (1974) Biochemistry 13, 2828-2836 11 Batzri, S. and Korn, E.D. (1973) Biochim. Biophys. Acta 298, 1015-1019 12 Rose. H.G. and Juliano, J. (1976) J. Lab. Clin. Med. 88, 29-43 13 Warnick, G.R. and Albers, J.J. (1978) J. Lipid Res. 19, 65-76 14 Rudel, L.L. and Morris, M.D. (1973) J. Lipid Res. 14, 364-366 15 Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J. (1951) J. Biol. Chem. 193, 265-275 16 Nichols, A.V., Gong, E.L., Forte, T.M. and Blanche, P.J. (1978) Lipids 13, 943-950 17 Chobanian, J.V., Tall, A.R. and Brecher, P.I. (1979) Biochemistry 18, 180-187 18 Jonas, A. (1979) J. Lipid Res. 20, 817-823 19 Scherphof, G., Roerdink, F., Waite, M. and Parks, J. (1978) Biochim. Biophys. Acta 542, 296-307 20 Tall, A.R. and Green, P.H.R. (1981) J. Biol. Chem. 256, 2035-2044 21 Young, P.M. and Brecher, P. (1981) J. Lipid Res. 22, 944-954 22 Barter, P.J. and Connor, W.E. (1976) J. Lab. Clin. Med. 88, 627-639 Biophys. 23 Ho, W.K.K. and Nichols, A.V. (1971) B&him. Acta 231, 185-193 A. 24 Marcel, Y.L., Vezina, C., Teng, B. and Sniderman, (1980) Atherosclerosis 35, 127- 133 25 Chajek, T. and Fielding, C.J. (1978) Proc. Natl. Acad. Sci. U.S.A. 75, 3445-3449