47

Biochimica et Biophysics Acta, 528 (1978) 47-57 @ Elsevier/North-Holland Biomedical Press

BBA 57114

INCORPORATION OF EXCESS CHOLESTEROL SERUM LIPOPROTEINS

BY HIGH DENSITY

ANA JONAS, LYNDAL K. HESTERBERG and SUSAN M. DRENGLER Department of Biochemistry, Urbana, Ill. 61801 (U.S.A.)

School of Basic Medical Sciences,

University of Illinois,

(Received June 14th, 1977)

Summary Excess cholesterol was added to human HDL3 and to bovine mammalian high density serum lipoprotein (HDL) by incubating aqueous lipoprotein solutions with solid dispersions of [4-14C]cholesterol on Celite. Lipoprotein * cholesterol complexes were isolated by centrifugation and filtration through a Sepharose 4B column. The pure complexes were analyzed for protein and lipid content and composition and were subsequently investigated by physical methods (analytical ultracentrifugation, circular dichroism, and fluorescence spectroscopy), in order to detect any structural changes induced by added cholesterol. The rates of cholesterol uptake varied as an inverse function of the intrinsic cholesterol present in the native lipoproteins. The maximum cholesterol taken up by human HDL3 increased the free cholesterol content from 3-4s (initial) up to 22% of the total lipoprotein weight. Bovine HDL was observed to increase its free cholesterol content from 2-4% (initial) up to ll-17% of the total lipoprotein weight, before denaturation. At maximum levels of added cholesterol, both lipoproteins had increased molecular weights and sedimentation velocity coefficients corresponding to the increased mass of the particles. No major changes in the hydrodynamic properties were observed. At the molecular level, the protein components only showed a 15-20s decrease in fluorescence intensity, possibly a consequence of a modified environment of the aromatic amino acid residues. In the human HDL3, added cholesterol increased the microviscosity of the lipid domains by 1.2 P at 25°C (from 3.4 to 4.6 P), but did not affect the fluidity of bovine HDL lipids (5.9 P).

Abbreviations: HDL, mammalian high density serum lipoprotein; isolated between densities 1.125 and 1.210 g/cm3.

HDL3, subclass of human HDL

48

Introduction The generally accepted structural model of high density serum lipoproteins (HDL) consists of a spherical, polar shell of proteins and phospholipid head groups surrounding a core of non-polar lipids, primarily cholesteryl esters and triglycerides [ 1,2]. Free cholesterol, which is present in low levels in circulating HDL (4.5--2X% of total weight) [2,3], is thought to be located near the lipoprotein surface, with the hydroxyl group directed toward the exterior of the particle [ 1,2]. The average physical and chemical properties of HDL have been well characterized in several animal species besides man (e.g. refs. 4-7). From such comparative studies, it is clear that each species has a distinct HDL class which exhibits considerable heterogeneity among different donors, particularly in the lipid components. Since the primary function of HDL, as of the other lipoprotein classes, is lipid solubilization and transport in blood, it is not surprising that their lipid composition should be variable. Furthermore, HDL is thought to be responsible for cholesterol removal from cell membranes and for its transport to the liver in the form of cholesteryl esters [8,9]. In view of the variability of lipid composition in HDLs, and their supposed ability to change it while in circulation, it is of considerable importance to know the capacity of these lipoproteins for binding lipids, in particular cholesterol. In this paper, we describe the binding of free cholesterol by human and bovine HDL and the structural consequences of adding this lipid component to native lipoproteins. Experimental

Procedure

Crystalline cholesterol, 99% pure, was purchased from Sigma Chemical Co.; [4-14C]cholesterol was a product of New England Nuclear with a specific activity of 55 Ci/mol. Both lipids were tested by thin-layer chromatography on Eastman silica gel G plates in light petroleum/diethyl ether/acetic acid (90 : 10 : 1, v/v) and were found to be pure by dichromate/H,SO, charring and by autoradiographic techniques. Lipid extraction with chloroform/methanol (2 : 1, v/v) and thin-layer analyses were performed at various stages of this work in order to monitor any changes in the radioactively labeled cholesterol. No chemical changes were observed for at least 2 months. Bovine HDL and human HDL3 were prepared by ultracentrifugal flotation in the density ranges: 1.063-1.170 and 1.125-1.210 g/ml, respectively. The procedures used in the preparations were described previously in our work [5,10] and the work of other investigators [11,12]. Purity of the lipoprotein preparations was judged from gel filtration experiments on a column of Sepharose 6B and from polyacrylamide gel electrophoresis of the isolated protein components. The Celite (grade 545) used in dispersing cholesterol was a product of John Manville Co. Prior to use, the Celite was washed repeatedly with dilute HCl, deionized water, and methanol, and dried under vacuum. The dispersion of lipid on Celite and incubation of protein and lipoprotein solutions with this solid dispersion of lipids was described in detail by Avigan [ 131, Ashworth and

49

Green [ 141, and more recently by our laboratory [ 15,161. In attempts to determine the maximum capacity of the HDLs for cholesterol, we used variable amounts of cholesterol on Celite (from 1 to 8%), variable weights of lipid dispersion per ml of lipoprotein solution (from 50 to 200 mg/ml), different temperatures (from 20 to 4O”C), and concentrations of HDL containing from 0.5 to 3.0 mg protein/ml. The “optimum” conditions (in terms of rates, recovery of products, and economy of reagents) were: 4% cholesterol (specific activity, 6.4 . lo7 dpm/mmol) on Celite, 50 mg of Celite dispersion per ml of HDL solution in 0.1 M Tris - HCl buffer (pH 8.0), 0.005% EDTA, low3 M NaN,, and 1 mg/ml protein. The mixtures were incubated under Nz, at 3O”C, in a shaking water bath for up to 48 h. After incubation, each sample was centrifuged at 15 000 rev./min for 30 min. The supernatant was filtered through Millipore filters (0.45 pm, pore size) and was then applied to a Sepharose 4B column (1.8 X 32 cm), equilibrated with the same buffer and operated at 23°C. The function of the column was 2-fold: (1) To ensure the separation of cholesterol bound to the lipoprotein from aggregated free cholesterol that apparently passes through Millipore filters in the presence of HDL, in particular the bovine HDL. The aggregated cholesterol appeared at the excluded volume of the Sepharose 4B column. In controls where lipoproteins were absent, only insignificant amounts of cholesterol remained in the supernatant and passed through Millipore filters. (2) The second function of the calibrated Sepharose 4B column was to provide the first indication of major structural changes in the lipoproteins upon cholesterol addition. The column fractions were analyzed by measuring absorbance at 280 nm and scintillation counting of [‘4C]cholesterol. Scintillation counting was performed in a Beckman LS-1OOC counter using a dioxane/naphthalene-based scintillation fluid. Chemical analyses were performed by the calorimetric procedures of Lowry et al. [ 171 for protein, Chen et al. [ 181 for organic phosphate, and Sperry and Webb [19] for cholesterol and cholesteryl esters. The effects of cholesterol additions on the protein structure of the lipoproteins were observed by determining intrinsic fluorescence spectra with a Hitachi-Perkin-Elmer MPF3 recording spectrofluorometer, exciting at 280 nm, and using 5-nm slits. CD spectra in the range from 260 to 190 nm were recorded with a Jasco J-40A automatic spectropolarimeter at 23°C. Fluorescence polarization of the lipophilic probe 1,6-diphenyl-1,3,5-hexatriene (Aldrich Chemical Co.) as a function of temperature, was used as a measure of the fluidity of the lipid domains of the lipoproteins with and without added cholesterol. The theory and applications of this technique were previously described by Shinitzky and co-workers [20-221 and were applied to lipoproteins in our own laboratory [lo]. The fluorescence polarization measurements were carried out with the SLM, Series 400 fluorescence polarization instrument, regulating temperature to +O.l”C with a circulating water bath. Fluorescence lifetimes were measured in the laboratory of Professor G. Weber, University of Illinois, using the cross-correlation phase fluorimeter described by Spencer and Weber [23]. Weight average molecular weights and sedimentation velocity coefficients of lipoproteins were determined using a Beckman Model-E 969 ultracentrifuge,

50

equipped with a photoelectric scanner. Analysis of sedimentation velocity and “conventional” sedimentation equilibrium results were performed using the methods summarized by Van Holde [ 241. Partial specific volumes for the lipoproteins were obtained from the literature: 0.870 ml/g for the human HDL, [ 21 and 0.910 ml/g for bovine HDL [ 51. Partial specific volumes for cholesterolcontaining samples were calculated by adding the fractional cholesterol contribution (cholesterol, u = 0.937 ml/g [25]) to the partial specific volume of the intact lipoprotein. The assumption that the volumes of the native lipoprotein components are not altered upon addition of cholesterol, appears to be justified, since there are no major changes in the structure of both lipoproteins even after addition of 120-170 cholesterol molecules (see Results). Results Before cholesterol binding experiments could be initiated, we had to test the effects of Celite exposure on the native structure and chemical composition of the human and bovine HDLs. Under the “optimum” cholesterol uptake conditions described in Experimental Procedure, both the human and bovine HDLs remained stable in terms of chemical composition and all physical parameters employed in this study, for 30 h. After 40 h, signs of denaturation appeared: turbidity and loss of protein and lipid from the supernatant fraction. Similar tests carried out on human low density and very low density lipoprotein fractions indicated that these lipoproteins bound very readily to Celite, suggesting that they can adsorb to non-specific surfaces. No further experiments were performed with the latter lipoprotein classes, although there were indications that low density lipoproteins take up large amounts of chdesterol. The kinetics of cholesterol uptake by human and bovine HDL preparations (containing 2.8 and 1.7% initial free cholesterol, respectively), under “optimum” conditions (as defined in Experimental Procedure) are shown in Fig. 1. The rates appear linear over a period of 20 h; however, the slight decrease observed at about 5 h is real. A closer examination of the kinetics in this region for both lipoproteins gave a reproducible decline in rate from 30 min to 5 h of incubation and then an increase between 5 and 7 h to the final linear rate. Other preparations of similar initial free cholesterol contents gave comparable results. Human HDLs reaches saturation after addition of 130 cholesterol molecules per lipoprotein, i.e. a total free cholesterol content of 22% of the lipoprotein weight. With several different bovine HDL preparations, complete saturation was never reached before denaturation set in; however, the maximum free cholesterol added varied between 120 and 180 cholesterol molecules/HDL, increasing the cholesterol content to 11-17s of the total lipoprotein weight. The rate of cholesterol incorporation by the lipoproteins was a function of several experimental variables: temperature, amount of cholesterol present in the incubation mixture, and concentration of the lipoprotein. We also observed that the initial rates of cholesterol uptake (up to 5 h) varied considerably for lipoprotein preparations from different donors. These rate differences were attributed to differences in the free cholesterol levels in the native lipoproteins. Lipoprotein preparations with high free cholesterol contents had low initial rates of cholesterol uptake, whereas lipoprotein preparations low in cholesterol had

51

Fig. 1. Rates of cholesterol uptake by human HDL3 (o--O) and bovine HDL (M) (mol cholesterol/mol HDL X h-1). The conditions for this experiment were: 4% cholesterol on Celite; 50 mg Celite . lipid dispersion per ml of lipoprotein solution; 1 mg/ml protein; 0.1 M Tris . HCl. pH 8.0,0.005% EDTA, 10m3 M NaN3; incubation at 3O’C. under N2 and in the dark. After 40 h of incubation. SOme denaturation was evident.

high initial rates of uptake. Table I shows cholesterol analysis data, together with the corresponding initial uptake rates for several lipoprotein preparations. Similar correlations were not observed between phospholipid or cholesteryl ester contents of the lipoprotein preparations and the rates of cholesterol incorporation. To determine the physical effects of cholesterol addition to the HDL structures, we selected human and bovine HDL, each having relatively low levels of intrinsic free cholesterol: 2.8 and 1.7%, respectively. Then we added cholesterol, up to 123 molecules to the human lipoprotein, and up to 165 molecules to the bovine lipoprotein. These preparations, plus several interTABLE I INITIAL CONTENTS OF CHOLESTEROL VARIOUS HDL PREPARATIONS

AND INITIAL

RATES

OF CHOLESTEROL

UPTAKE FOR

The initial rates were obtained under “optimal” experimental conditions. i.e. 4% cholesterol on Celite. 50 mg of solid per 1 ml of lipoprotein solution at 1 mg/ml protein content, and 30°C incubation temperature, after 2-5 h of incubation. The results are the mean of three determinations on a single preparation. The errors represent the variability of the results. HDL preparation

Initial cholesterol (percent of total weight)

Initial rate of uptake (mol cholesterollmol HDL X h-l)

Human HDL3 1 2

2.8 (tlO%) 4.3

4.6 (rlO%) 2.5

Bovine HDL 1 2 3 4

1.7 (+10x?) 2.3 3.1 4.2

5.0 (510%) 3.3 2.6 1.9

52

mediate ones, and the corresponding controls were used in all the experiments described below. Gel filtration on the Sepharose 4B column (Fig. 2) gave identical elution profiles and positions for the lipoproteins with and without added cholesterol; however, as expected, the larger bovine lipoprotein eluted a few fractions ahead of the human lipoprotein. Chemical analysis of the peak column fractions agreed very well with the cholesterol/HDL values calculated from radioactivity and protein determinations. In order to test the stability of the lipoproteins with added cholesterol, we repeated the gel filtration experiments on the same preparations which had been stored at 4”C, under N2 for 1 month. The results were identical in terms of peak positions and contents of cholesterol/HDL, indicating that these preparations are stable for at least that length of time. No lecithin : cholesterol acyltransferase activity was detected. Molecular weights were determined by the “conventional” sedimentation equilibrium method at 10 000-12 000 rev./min. Weight average molecular weights calculated from log c vs. r2 plots, agree well with the published molecular weights for human HDL, (180 000) [Z] and bovine HDL (380 000) [ 51. The molecular weights determined in this work are given in Table II together with sedimentation velocity coefficients corrected to water at 20°C (~se,~). These results indicate that there are no major differences between the hydrodynamic properties of the native lipoproteins and those with added cholesterol, except for the increases in molecular weights and sedimentation velocity coefficients due to the increased mass of cholesterol. At the level of the protein and lipid constituents, we employed spectroscopic

Fraction Fig.

2.

Column

indicate

the

Sepharose NaN3, ------,cpm.

and

elution

position 4B was

(1.8

profiles: of

elution

X 32 cm)

operated

at

(A) of was

23’C.

Number human the

native

equilibrated Column

HDL3

and,

(B)

lipoproteins; with fractions

0.1

bovine

Vo M

were

Tris 1.5

HDL,

denotes

added

cholesterol.

volume.

PH

0.005%

. HCI, ml

with

excluded

each.

F,

8.0,

The EDTA,

absorbance

Arrows column

of

10m3

at 280

M

nm;

53 TABLE II PHYSICAL

PROPERTIES

OF HDLs WITH AND WITHOUT

ADDED

CHOLESTEROL

The indicated errors are the approximate uncertainty of each instrumental method. 1,3,5-hexatriene. HDL sample

Molecular weight

DPH. 1,6-diphenyl-

Sedimentation coefficient

Lifetime (DPH fluorescence)

‘2O.w (S)

(ns)

Microviscosity at 25% (P)

Human HDL3

192 000 (i15%)

4.37 (&o)

8.7 (i3%)

3.4 (?5%)

Human HDL3 +123 cholesterol molecules

236 000

4.42

8.8

4.6

Bovine HDL

388 000

4.85

8.5

5.9

Bovine HDL +165 cholesterol molecules

439 000

5.69

8.5

5.8

techniques to monitor any structural changes induced by added cholesterol. Fig. 3 shows the intrinsic fluorescence spectra for the human and bovine HDLs and for various complexes with cholesterol. At low levels of added cholesterol, up to 50 cholesterol molecules/HDL, there is a slight increase in fluorescence intensity followed by a 25% decrease at the maximum cholesterol levels. No shifts in the wavelength of maximum fluorescence could be detected. Circular dichroism spectra recorded between 260 and 190 nm, for the same series of samples as those used in the determination of fluorescence spectra, did not

OEO I

300 1

350 I

4co250 / I Wavelength

x)0 I (nm)

350 /

.1

4(D

Fig. 3. Intrinsic fluorescence spectra of human HDLJ (A) and bovine HDL (B) with and without added cholesterol. Fluorescence spectra are uncorrected and are given in arbitrary fluorescence intensity units. Panel A: (1) HDL3 plus 35 molecules of cholesterol; (2) native HDL3; (3) HDL3 plus 123 molecules of cholesterol. Panel B: (1) HDL plus 50 molecules of cholesterol: (2) native HDL; (3) HDL Plus 165 molecules of cholesterol.

54

oil0

10

20 Temperature

30

40

50

(“Cl

Fig. 4. Fluorescence polarization of the 1,6-diphenyl-1,3,5-hexatriene lipoprotein probe as a function of temperature. 0, native human HDL3; A, HDL3 plus 123 molecules of cholesterol; 0, native bovine HDL; +, bovine HDL plus 165 molecules of cholesterol; at 25’C the corresponding microviscosities are: 3.4, 4.6, 5.8, and 5.9 P, respectively. These results were not corrected for depolarization due to lipoprotein rotations.

reveal significant differences between each lipoprotein and the corresponding samples with added cholesterol. To probe the lipid domains of the lipoprotein samples, we dissolved 1,6-diphenyl-1,3,5-hexatriene into the lipoproteins up to a final concentration of one or two molecules of probe per HDL. The fluorescence polarization of the probe, as a function of temperature, is shown in Fig. 4. For the bovine HDL, addition of cholesterol up to 165 molecules per lipoprotein has no effect on the fluidity of the core lipids; on the other hand, for the human HDL3, there is a significant decrease in the mobility of the lipids as indicated by the increase in polarization at all temperatures. Microviscosity values at 25°C were calculated using the procedure of Shinitzky and co-workers [21,22] and measured fluorescence lifetimes. The microviscosity values together with lifetime data are included in Table II. Discussion In this work, we demonstrated that human HDL3 can incorporate up to 22% of its weight in cholesterol and that the capacity of bovine HDL for cholesterol is at least ll-17% of the lipoprotein weight. Furthermore, the initial rate of cholesterol incorporation depends on the initial cholesterol concentration of the lipoproteins. Since, in circulation, the average content of cholesterol is only 2-4% of the HDL weight, the amounts of free cholesterol present are not limited by the binding capacity of HDL; rather, they are determined by the relative rates of cholesterol exchange between HDL and cell membranes and the other classes of lipoproteins (see ref. 8), as well as by the rate of cholesterol transformation into cholesteryl esters by the action of lecithin : cholesterol acyltransferase [8]. It also appears, from our in vitro results, that intrinsic HDL cholesterol levels

55

may be able to regulate the uptake of cholesterol by affecting the rate of incorporation. Regarding the kinetics of cholesterol uptake shown in Fig. 1, the discontinuity in rates observed around 5 h may indicate the existence of more than one type of binding site for cholesterol on each HDL. Since the inverse correlation of initial rates (up to 5 h) and intrinsic free cholesterol contents requires the presence of saturable binding sites, it is logical to assume that the first part of the curve in Fig. 1 represents the filling of these sites. The origin and nature of the sites that fill up later is unknown, but we are presently conducting experiments on native HDL and on chemically defined complexes of the A-I protein from HDL with lipids, that should resolve this problem. Changes in protein fluorescence intensity from increase to decrease, relative to the native lipoprotein, take place in the same region of cholesterol incorporation where the discontinuity in the kinetics is observed, supporting the possible existence of different binding sites for cholesterol. The structure of lipoproteins is determined by the physical properties of its components, by their interactions, and by their volume and packing characteristics [ 1,2,26]. The location and structural role of free cholesterol in HDL is not known with certainty. By analogy with phospholipid : cholesterol bilayer structures [27], cholesterol in HDL may be localized near the lipid-water interface, with the hydroxyl groups pointing toward the surface and in close proximity to the phosphate groups. Recently, on the basis of volume and composition considerations, cholesterol in human HDL was localized in contact with the inner surfaces of protein components [2]. Whatever the location of cholesterol in HDL, its structural role is still uncertain. These experiments are, in fact, the first to probe the effect of excess cholesterol on the structure of HDL. From ultracentrifugal analysis and gel filtration experiments, we demonstrated that the size and shape of the lipoproteins is affected very little by the addition of 120-170 cholesterol molecules/HDL. Changes in molecular weights and sedimentation velocity coefficients can essentially be accounted for by the increase in mass due to the added cholesterol. Sedimentation velocity coefficients calculated for the cholesterol-containing samples from the szo.., values of native lipoproteins and the molecular weights given in Table II, using the relationship sio,w = (12.0 * 10m3) M,2’3(1 - u~)/u”~, for a sphere [24] were: 4.48 S for the human HDL, + 123 cholesterol molecules sample, and 5.09 S for the bovine HDL + 165 cholesterol molecules sample. The calculated (4.48 S) and experimental (4.42 S) values of sZo,w for the human sample agree very well and indicate no changes of shape or hydration. The discrepancy for the bovine HDL sample between the measured (5.69 S) and calculated (5.09 S) values is only 10% and could be due to a slight difference in shape or hydration, or both, between the native HDL and the cholesterol-containing sample. In terms of the structure at the level of the protein and lipid components, the only effects of added cholesterol are on the fluorescence spectra of both lipoproteins and on the lipid fluidity of human HDL3. At maximum levels of cholesterol addition, there is a 15-20% decrease in the fluorescence intensity of the protein components, whereas at low levels of added cholesterol (up to 50 cholesterol molecules/HDL) there is in fact a 5-10% increase in the

56

fluorescence yield. Because the mechanisms for fluorescence quenching in proteins are not well understood [28], it is not possible to interpret these results in terms of molecular events; however, the most reasonable explanation is that added cholesterol has either a direct or an indirect effect on the aromatic amino acid residues by changing their local environment. Since we do not know what is the precise location of these amino acid residues, we cannot say more about the localization of cholesterol. Secondary structure changes in the protein components, that may be revealed by circular dichroism spectra, were not observed upon addition of cholesterol to the lipoproteins. Microviscosity results indicated that addition of cholesterol to the highly viscous lipids of bovine HDL (r) (at 25°C) = 5.9 P) had no effect on the average mobility of the fluorescent probe. Apparently, this lipoprotein with a relatively high content of lipids, in particular cholesteryl esters, can accommodate the added cholesterol without affecting significantly their fluidity. The human HDL3, on the other hand, with a n (at 25°C) = 3.4 P, increases in microviscosity up to 4.6 P as 123 excess cholesterol molecules are added per lipoprotein. As in bilayer structures, the probable effect of cholesterol, is to restrict the mobility of lipids (particular, acyl chains of phospholipids) above their phase transition temperature [ 21,221. The microviscosities reported previously for total human HDL at 25°C were 5.1 + 0.7 P [lo]; the present results, however, were not corrected for particle rotations. In their early work Ashworth and Green [14] reported that human a-lipoproteins take up 22 mol of cholesterol/mol in 18 h. This rate of uptake is compatible with our results (see Table I). Stein et al. [29], in recent studies of human skin fibroblasts and rat aortic smooth muscle cells in tissue culture, showed that replacement of the fetal calf serum with human HDL resulted in a loss of 15-27s of the cellular cholesterol. The loss was enhanced further by addition to the medium of HDL-apolipoprotein . phospholipid mixtures. It appears, therefore, that HDL and apolipoprotein . phospholipid complexes are important in cholesterol uptake and that rates may be determined by the cholesterol content of the acceptor. Our in vitro experiments support this observation by showing that initial HDL content of free cholesterol is inversely related to the rates of exogenous cholesterol uptake. We also demonstrated that human and bovine HDL have a high capacity for binding excess cholesterol without causing major changes in the lipoprotein structure. Acknowledgements This work was supported by Grant HL-16059 from the National Institutes of Health, U.S.A., and Grant No. C-l from the Illinois Heart Association. A.J. is an Established Investigator of the American Heart Association.

References 1

Morris&t,

2

Shen,

3

J.D.,

B.W.,

Skipski, Interscience,

V.P.

Jackson,

Scanu. (1972) New

R.L.

A.M. York

in

and

and

Gotto,

Kezdy,

Blood

F.J.

Lipids

Jr.,

A.M.

(1977) and

(1975) Proc.

Natl.

Lipoproteins

Anna

Rev.

Acad. (Nelson,

Biochem.

Sci.

U.S.

G.J.,

74. ed.),

44,

183-207

837-841 PP.

471-583,

Wiley-

57

4 Koga,

S..

Horwitz,

5

Jonas,

A.

6

Fidge.

N. (1973)

7

Blaton,

D.L.

(1972)

V..

and

J. Biol.

Scanu.

Chem.

Biochim.

Vercaemst,

A.M.

247,

Biophys. R.,

(1969)

J. Lipid

Res.

10,

577-588

7767-7772 Acta

295,258-273

Vandecasteele,

Caster,

N..

H.

and

Peeters,

H.

(1974)

Biochemistry

13,

1127-1135 8

Glomset,

J.A.

(1972)

Interscience. 9

New

Hamilton. 572.

R.L.

M.

10

Jonas,

11

DeLaIla,

Blood

Kayden,

New

(1977) 0.

Lipids

and

H.J.

and

J.

F.T.

181-274,

Wiley-Interscience,

Jensen.

L.C.

(1972)

New

J. Biol.

15

Jonas,

A.

(1975)

Biochem.

Biophys.

Res.

16

Jonas,

A.

(1976)

Biochim.

Biophys.

Acta

17

Lowry,

18

Chen,

19

Sperry, W.M.

20

Shinitzky,

M.,

21

Cogan,

Shinitzky,

and

O.H..

Rosebrough,

Jr.. P.S.,

Toribara,

Shinitzky,

24

Van

and

Holde,

Academic

of Chemistry R.,

27

Papahadjopoulos,

Morrod,

28

Eisinger,

J., Lamola,

29

Stein,

Vanderhoek,

Disease

(Becker,

F.F.,

ed.),

PP. 531-

Biochemical

Analysis

(Glick,

D.,

ed.),

Vol.

I.

Lipids

a’nd

Lipoproteins

Acta

70,

(Nelson,

G.J..

ed.),

PP.

The

Randall,

H. (1956) Chem.

N.Y.

Proteins

R.J.

Anal.

187,

Weber,

(1951) Chem.

Chem. Acad.

(Neurath.

J. Biol. 28,

Chem.

193,

265-275

1756-1758

97-106

G. (1971)

T. (1973)

J. Biol.

Ann.

68-74

1003-1008

325-336

Nishida,

(1974)

64.

Biochemistry Biochemistry

249, Sci. H.

10. 12,

2106-2113

521-528

2652-2657 158.361-376

and

Hill,

R.L.,

eds.),

Vol.

I,

pp.

225-291,

York

Handbook

D.

of

Biophys.

and

C. and

G. and

in

Blood

427.

J. Biol.

(1975)

Schneider,

PP. 180-190,

Y.

of

Commun.

A.L.

Gitler,

G. (1969)

26

O.,

WikY-

York

Biochim.

Warner,

Weher,

New

IV,

and

Weber,

Barenholz,

K.E. Press,

PP. 745-787.

787-790

Far,

M. (1950) A.C.,

M..

25

Vol.

T.Y.

Webb,

and

234,

C. (1963)

N.J.,

Dianoux,

M. and R.D.

Chem.

Green,

New in

Avigan,

Spencer,

ed.).

York

Ashworth,

22

G.J..

lo-22

Methods

Inc.,

13

23

(Nelson,

Biochemistry

486.

in

14

L.A.E.

in

Acta

(1954)

Publishers

Lindgren,

J. (1959)

(1974)

Biophys.

Gofman,

Interscience

U.,

Lipoproteins

York

Biochim.

and

PP. 459-478, 12

and

Dekker,

A.

in

York

and R.S., and

Physics

(1975/1976).

Calvin,

J.R.

Kimelberg,

Pergamon A.A.,

H.K.

Press.

Longworth.

J. and

Stein,

and

(1973)

CRC

N.H.

Press.

(1973)

in Progress

Cleveland,

Chem.

Phys.

in Surface

Ohio Lipids

Science

10,

(Davison,

Oxford J.W.

Y.

C-240. Tattrie,

(1976)

and

Gratzer. Biochim.

W.B.

(1970)

Biophys.

Acta

Nature

226,

431,

347-358

113-118

328-353 S.G.,

ed.).