Distribution of glycosphingolipids in the serum lipoproteins of normal human subjects and patients with hypo- and hyperlipidemias.

Five glycosphingolipids (GSL), glucosylceramide, lactosylceramide, trihexosylceramide, globoside, and hematoside (GM3) were studied in serum from normal human subjects and patients with dyslipoproteinemia and found to be exclusively associated with the various classes of serum lipoproteins. Based on a unit weight of lipoprotein protein, the total amount of GSL in serum normal subjects was twice as high in very low density lipoprotein (VLDL) (d less than 1.006 g/ml) and low density lipoprotein (LDL) (d 1.019-1.063 g/ml) as in high density lipoproteins HDL2 (d 1.063-1.125 g/ml) or HDL3 (d 1.125-1.21 g/ml). In abetalipoproteinemia the levels of serum GSL were slightly reduced when compared to normal serum and were all found in the only existing lipoprotein, HDL; this contained 2-3 moles of GSL/ mole of lipoprotein as compared to 0.5 GSL/mole in normal HDL. In hypobetalipoproteinemia and Tangier disease, the serum glycosphingolipids were 10 to 30% reduced in concentration compared to the 75% reduction in other lipids, and were again found to be associated only with the serum lipoproteins. The relative proportions of GSL did not vary substantially in the normo- and hypolipidemic subjects studied. Only in patients with homozygous familial hypercholesterolemia was there a significant (3-4-fold) elevation of all of the five GSL species and this elevation of all of the five GSL species and this elevation correlated well with that of the circulating cholesterol and LDL. On a molar basis the LDL of these patients contained the same amount of GSL as normal subjects (5 moles GSL/mole protein). It is concluded that: (1) glycosphingolipids are associated only with the major lipoprotein classes in both normal and dyslipoproteinemic serum; (2) the relative proportions of the five glycosphingolipids are not significantly affected by dyslipoproteinemia; (3) only in severe hypolipoproteinemia do the remaining serum lipoproteins carry a complement of glycosphingolipids greater than normal. Although our results establish that glycosphingolipids are intimately associated with serum lipoproteins, the mode of association or the structural and functional significance of such an association remains undetermined.     

The sites of degradation of purified rat low density lipoprotein and high density lipoprotein in the rat

Low density lipoprotein and high density lipoprotein were isolated from rat serum by sequential ultracentrifugation in the density intervals 1.025-1.050 g/ml and 1.125-1.21 g/ml, respectively. The isolated lipoproteins were radioiodinated using ICl. Low density lipoprotein was further purified by concanavalin A affinity chromatography and concentrated by ultracentrifugation. 95% of the purified low density lipoprotein radioactivity was precipitable by tetramethylurea, while only 4% was associated with lipids. The radioiodinated high density lipoprotein was incubated for 1 h at 4 degrees C with unlabelled very low density lipoprotein, followed by reisolation by sequential ultracentrifugation. Only 3% of the radioactivity was associated with lipids and 90% was present on apolipoprotein A-I. The serum decay curves of labelled and subsequently purified rat low and high density lipoprotein, measured over a period of 28 h, clearly exhibited more than one component, in contrast to the monoexponential decay curves of iodinated human low density lipoprotein. The decay curves were not affected by the methods used to purify the LDL and HDL preparations. The catabolic sites of the labelled rat lipoproteins were analyzed in vivo using leupeptin-treated rats. In vivo treatment of rats with leupeptin did not affect the rate of disappearance from serum of intravenously injected labelled rat low density lipoprotein and high density lipoprotein. Leupeptin-dependent accumulation of radioiodine occurred almost exclusively in the liver after intravenous injection of iodinated low density lipoprotein, while both the liver and the kidneys showed leupeptin-dependent accumulation of radioactivity after injection of iodinated high density lipoprotein.     

Isoelectric focusing of human serum high density lipoprotein in the presence of radioactive ampholines.

Preparative isoelectric focusing of a serum high density lipoprotein fraction (rho 1.063-1.125 g/ml) in the presence of radioactive Ampholines showed that there were no significant 14-C-activity peaks in the lipoprotein region of the pH gradient. This suggests that the discrete lipoprotein subspecies obtained by electrofocusing are not artefacts caused by the binding of Ampholines to serum high density lipoprotein.     

Purification and characterization of proinsulin mRNA from rat B-cell tumor.

1. Photoaffinity labelling of human serum albumin with the sodium salts of (3 beta-azido-7 alpha,12 alpha-dihydroxy-5 beta-cholan-24-oyl)-2-amino[2(-3)H (N)]ethanesulfonic acid, (7,7-azo-3 alpha,12 alpha-dihydroxy-5 beta-cholan-24-oyl)-2-amino[2(-3)H (N)]ethanesulfonic acid and (11 zeta-azido-12-oxo-3 alpha,7 alpha-dihydroxy-5 beta-cholan-24-oyl)-2-amino[2(-3)H (N)]ethanesulfonic acid resulted, in each case, in a considerable covalent incorporation of radioactivity into the protein. 2. Photoaffinity labelling of whole serum, obtained from fasting test persons, revealed with all three photolabile derivatives of taurocholate at the physiological concentration of 2.1 microM the incorporation of radioactivity not only into albumin but also into high-density lipoprotein, as demonstrated by density gradient centrifugation and by immunological characterization. 3. The bulk of radioactivity incorporated into high-density lipoprotein by photoaffinity labelling of whole serum was found to have been associated with the lipids. Only 10-20% of the label was covalently bound to apolipoproteins, predominantly to the apolipoproteins A-I and A-II. 4. The interaction of taurocholate with high-density lipoprotein has been confirmed by density gradient centrifugation using 14C-labelled taurcholate. It is assumed that the interaction of taurocholate with high-density lipoprotein is physiologically of significance.     

Mycoplasma growth factors in bovine serum fraction.

Mycoplasma growth factors in bovine serum fraction were separated by Sephadex G150 column chromatography and density ultracentrifugation. The major growth factor of bovine serum fraction eluted from the Sephadex column in the void volume. Its growth-supporting activity was greatly enhanced by the presence of bovine serum albumin in the mycoplasma culture media. Other investigators had previously identified the major growth factor in serum as an alpha-lipoprotein. Although density ultracentrifugation revealed the presence of traces of a high-density lipoprotein in bovine serum fraction, another, less dense component, isolated by ultracentrifugation (component 3) and containing cholesterol, cholesteryl esters, free fatty acids, triglycerides, and protein, but no lipoprotein, exhibited considerably more growth-supporting activity than did the high-density lipoprotein, thus indicating that at least two mycoplasma species do not require intact serum lipoprotein for growth. Both the high-density lipoprotein and component 3 exhibited maximum activity only in the presence of bovine serum albumin. A chloroform extract containing component 3 lipids combined with bovine serum albumin to form an effective, partially defined, less complex substitute for serum in mycoplasma culture media.     

Heat denaturation of human high density lipoproteins and chylomicrons

1. The objective of this investigation was to determine whether structural differences between apolipoproteins could be detected by heat denaturation. 2. The apoproteins of human serum high density lipoprotein (HDL2, d = 1.070-1.125 and HDL3, d = 1.125-1.21 g/ml), their major polypeptide constituents (R-Thr and R-Gln), and apochylomicrons were investigated. 3. Heat denaturation was found to be reversible in the temperature range from 20 to 80 degrees. 4. The thermodynamic parameters of heat denaturation delta F, delta H, delta S and delta Cp were calculated on the basis of a single transition from the "native" to "denatured" state for apo-HDL2, apochylomicrons, R-Thr and R-Gln; for apo-HDL3 these parameters were calculated on the basis of two transitions. 5. The thermodynamic parameters, with the exception of delta F, which describe heat denaturation of high density apolipoprotein, of high density apolipoprotein polypeptides and of apochylomicrons were found to be similar on a molar basis and to have approximately the same values as the thermodynamic parameters which describe heat denaturation of non-lipid binding proteins; on a weight basis differences were apparent between the apolipoproteins and the polypeptides or non-lipid binding proteins.     

Studies on the composition of pig serum lipoproteins. Isolation and characterization of different apoproteins.

1. Different lipoprotein density fractions from pig serum were isolated by phosphotungstate precipitation followed by purification in the preparative ultra-centrifuge. 2. The protein part of very low density lipoproteins was composed of approximately 52 percent lipoprotein B apoprotein and the rest of lipoprotein C II apoprotein and other as yet unidentified peptides. 3. The protein moiety of low density lipoproteins consisted primarily of lipoprotein B apoprotein (over 95 percent); the amino acid compositions of lipoprotein B apoprotein of very low and low density lipoproteins were practically identical. 4. The predominant polypeptide of pig serum high density lipoproteins exhibited an amino acid composition and a molecular weight very similar to human liprotein A I apoprotein. In contrast to human lipoprotein A I apoprotein, the apoprotein from pigs was found to release leucine first followed by alanine, threonine, and lysine upon incubation with carboxypeptidase A. 5. In pig serum the major lipoprotein C apoprotein was found to be a polypeptide similar in amino acid composition to lipoprotein C II apoprotein from human serum. The molecular weight of this polypeptide is approximately 8000. Incubation experiments with carboxypeptidase A indicate serine to be the most likely C-terminal amino acid.     

Substitution in vitro of lecithin-cholesterol acyltransferase. Analysis of changes in plasma lipoproteins

Lecithin-cholesterol acyltransferase (EC 2.3.1.43) was purified 15 000-fold from human plasma. The active material was homogeneous in different gel electrophoretic systems but separated into three major bands with apparent pI values of 4.28, 4.33 and 4.37 in isoelectrofocusing. The apparent Mr of the enzyme is 67 000 +/- 2000. An antiserum prepared against the purified enzyme specifically inhibited the activity of lecithin-cholesterol acyltransferase in whole serum. Serum from a patient with familial deficiency of lecithin-cholesterol acyltransferase was substituted in vitro with the highly purified enzyme. The serum from this patient did not contain immunochemically detectable enzyme protein. Substitution of enzyme resulted in the following major changes. 1. Cholesteryl ester content in serum increased by 36-89 mg/100 ml depending on the experimental conditions. The enzyme-mediated formation of cholesteryl ester led to an increase of cholesteryl ester content in high-density and very-low-density lipoproteins and in low-density lipoproteins containing apoprotein-B. No increase occurred in fractions containing very large flattened structures and the abnormal lipoprotein-X and in lipoprotein-E. Incubation of isolated fractions with lecithin-cholesterol acyltransferase led to significant cholesterol esterification only in high-density lipoproteins. 2. The characteristic disc-shaped rouleaux-forming high-density lipoproteins of enzyme-deficient serum disappeared. Instead a single homogeneous population of high-density lipoproteins formed. The particles generated were spherical and had the electrophoretic properties, density (1.080 g/ml), diameter (12.5 nm) and apoprotein composition of normal high-density lipoproteins-2. 3. The concentration of spherical particles containing apolipoprotein E (density 1.040-1.080 g/ml) and the lamellar lipoprotein-X-like structures in the low-density lipoprotein fraction were not affected by the enzyme substitution. 4. A single homogeneous population of spherical lipoprotein-B particles of 26.5-nm diameter occurred at density 1.029 g/ml. The data suggest that the discoidal high-density lipoproteins are the major site of cholesteryl ester formation that apolipoprotein-E is not involved in an undirectional transport of newly formed cholesteryl ester from high-density lipoproteins to other lipoproteins and that lipoprotein-X and lipoprotein-E are not preferential substrates for the acyltransferase.     

Reverse cholesterol transport in the isolated perfused rat spleen.

1. A method has been developed which enables the rat spleen to be loaded in vivo with [3H]cholesterol to a high specific radioactivity using cholesterol-labelled erythrocytes. The erythrocytes were shown to be rapidly degraded by the spleen and not released intact during subsequent perfusion. 2. When labelled spleens were perfused with whole blood or serum, lipoproteins in the high-density lipoprotein (HDL) range were shown to be the principal lipoprotein vehicles for the removal of cholesterol, the specific radioactivity of cholesterol being much greater in the HDL fractions than in other lipoproteins, particularly in the d 1.175-1.210 fraction. 3. The formation of [3H]cholesteryl ester was restricted to the major HDL fractions. 4. Experiments utilizing individual HDL fractions added to a basal perfusate indicated that HDL1 (d 1.050-1.085) was of less importance in the removal of cholesterol from the spleen than HDL subfractions of higher density. Also, a decrease in density of the lipoproteins was observed during perfusion, concurrent with uptake of cholesterol, especially in the d 1.085-1.125 subfraction. 5. When [3H]cholesterol-labelled spleens were perfused with whole blood, about half of the radioactivity released was detected in erythrocytes, indicating a rapid exchange or transport of cholesterol. Thus erythrocytes could play an important role in the transfer of unesterified cholesterol when the chemical potential gradient is favourable.     

Esterification of cholesterol in high density lipoprotein decreases its ability to support ACTH-stimulated steroidogenesis by rat adrenocortical cells

Addition of high density lipoprotein 3 (HDL3) isolated from human plasma of d greater than 1.125 g/ml which had been preincubated for 24 h at 37 degrees C enhanced steroidogenesis by cultured rat adrenal cells only 38% as well as HDL3 isolated from unincubated plasma. Loss of steroidogenic activity due to preincubation was associated with a decrease in the percent HDL3 cholesterol remaining unesterified. Inhibition of lecithin-cholesterol acyltransferase activity by heating (60 degrees C, 1 h) or addition of dithionitrobenzoic acid (1.4 mM) prevented esterification of cholesterol in HDL and also prevented loss of steroidogenic activity. Although incubation of plasma of d greater than 1.125 g/ml prior to isolation caused cholesterol esterification, there was no change in the ratio of total cholesterol to protein in HDL, size and shape of the HDL particle as assayed by measurement of sedimentation velocity, nor affinity for the putative HDL receptor. Addition of unesterified cholesterol to preincubated HDL restored steroidogenic activity. These results indicate that unesterified cholesterol in HDL is preferentially used as substrate for rat adrenal steroidogenesis. The effects of nonlipoprotein serum proteins on HDL action in the adrenal were also examined. The ability of HDL3 to enhance rat adrenal steroidogenesis was not significantly less in serum-free media than in media supplemented with lipoprotein-poor fetal calf serum or human plasma of d greater than 1.21 g/ml, suggesting that rat adrenal uptake of HDL cholesterol does not depend on participation of plasma enzymes or transport proteins.     

Rapid transformation of [3H]cholesteryl ester in rat high-density lipoprotein: in vivo and in vitro studies

[24,25-3H]Cholesteryl ester-labeled rat high-density and low-density lipoproteins were administered to recipient rats. Following death of the rats, a major portion of the radioactivity in administered [3H]cholesteryl ester-high-density lipoprotein rapidly appeared in less dense [3H]cholesteryl ester-lipoproteins and was isolated with the low-density lipoprotein fraction. The specific activity of the esterified cholesterol in the product lipoproteins found with the low-density lipoproteins exceeded that of the precursor high-density lipoproteins. In vitro, the addition of [3H]cholesteryl ester-high-density lipoprotein to plasma resulted in a five- to six-fold increase in radioactivity recovered in the low-density lipoprotein. These results demonstrate that, under a variety of experimental conditions, isolated high-density lipoprotein particles (both in vitro and in vivo) tend to become larger and less dense. Rapid changes in the density of lipoproteins labeled with [3H]cholesteryl ester must be considered when interpreting physiologic studies using this label.     

Relation of angiographically defined coronary artery disease to plasma lipoprotein subfractions and apolipoproteins.

The relation of coronary artery disease to plasma lipoproteins was examined in 104 men aged 35-65 years undergoing coronary angiography for suspected myocardial ischaemia. A score reflecting the number, degree, and length of stenoses in seven major coronary arteries was assigned to each angiogram. Lipid concentrations in lipoprotein subfractions were measured after preparative ultracentrifugation; plasma apolipoprotein concentrations were measured by electroimmunoassay. Men with high coronary scores tended to have lower plasma high-density lipoprotein (HDL) cholesterol concentrations and higher low-density lipoprotein (density 1.019-1.063 g/ml) cholesterol concentrations than subjects of similar age with low coronary scores (p approximately equal to 0.1). The strongest relation, however, was with the cholesterol concentration in the HDL2 subfraction (density 1.063-1.125 g/ml) of HDL, which averaged 44% lower in the severely affected patients (p less than 0.005). No associations were found between the coronary score and HDL3 cholesterol, the cholesterol content of lipoproteins of density less than 1.019 g/ml, plasma triglyceride, or the concentrations of apolipoproteins AI, AII, and E. The high coronary scores associated with low HDL2 concentrations reflected an increase in the number of both partial and complete stenoses distributed throughout the coronary tree. In contrast the sizes of the lesions and the proportion producing complete occlusion were unrelated to HDL2.     

Isolation, characterization and quantification of apolipoproteins A-1 and B of the Golden Syrian hamster (Mesocricetus auratus) and modification of their levels by dietary cholesterol

1. Apolipoprotein A-1, isolated from hamster high density lipoprotein, possessed a molecular weight of approximately 27,000. 2. Its amino acid composition differed from human apo A-1 and it contained a higher threonine to serine ratio and a higher methionine and leucine content. 3. The concentration in normal serum was 126.0 +/- 1.9 mg/dl. 4. Apolipoprotein B, isolated from hamster low density lipoprotein consisted of three major components when analyzed by SDS-polyacrylamide gel electrophoresis with Mrs of 635 Kd, 460 Kd and 305 Kd respectively. 5. Hamster apo B possessed a higher aspartic acid to glutamic acid ratio and a higher methionine and valine content than human apo B. 6. The concentration in normal serum was 20.9 +/- 1.0 mg/dl. 7. The apolipoprotein and lipoprotein profile of hamsters fed a high cholesterol diet for 30 days changed considerably. 8. Total serum cholesterol levels increased 7 fold; LDL levels increased 14 fold; HDL levels doubled and total serum triglyceride increased 3 fold. 9. Apo A-1 levels increased by 45% and apo B levels increased 5 fold.     

The isolation and partial characterization of the serum lipoproteins and apolipoproteins of the rainbow trout.

1. VLD (very-low-density), LD (low-density) and HD (high-density) lipoproteins were isolated from the serum of trout (Salmo gairdneri Richardson). 2. Each lipoprotein class resembled that of the human in immunological reactivity, electrophoretic behaviour and appearance in the electron microscope. Trout LD lipoprotein, however, was of greater density than human LD lipoprotein. 3. The trout lipoproteins have lipid compositions which are similar to those of the corresponding human components, except for their high contents of long-chain unsaturated fatty acids. 4. HD and LD lipoproteins were immunologically non-identical, whereas LD lipoproteins possessed antigenic determinants in common with VLD lipoproteins. 5. VLD and HD lipoproteins each contained at least seven different apoproteins, whereas LD liprotein was composed largely of a single apoprotein which resembled human apolipoprotein B. 6. At least one, and possibly three, apoprotein of trout HD lipoprotein showed features which resemble human apoprotein A-1.7. The broad similarity between the trout and human lipoprotein systems suggests that both arose from common ancestral genes early in evolutionary history.     

A 70-kDa apolipoprotein designated ApoJ is a marker for subclasses of human plasma high density lipoproteins

A new apolipoprotein, termed apolipoprotein J (apoJ), was purified from human plasma by immunoaffinity chromatography. ApoJ is a glycoprotein consisting of disulfide-linked subunits of 34-36 and 36-39 kDa. Each subunit is glycosylated and has a pI range of 4.9-5.4. ApoJ exists in the plasma associated with high density lipoproteins (HDL) and specifically with subclasses of HDL which also contain apoAI and cholesteryl ester transfer protein activity. Immunoaffinity purified apoJ-HDL subclasses have apparent molecular masses of 80, 160, 240, 340, and 520 kDa, as determined by gradient gel electrophoresis. By negative staining electron microscopy, apoJ-HDL range in diameter from 5 to 16 nm. Fractionation of plasma by vertical gradient density centrifugation revealed apoJ-HDL in HDL2 (d 1.063-1.125 g/ml) with the majority overlapping HDL3 (d 1.125-1.21 g/ml) and very high density lipoprotein (d 1.21-1.25 g/ml). The bimodal density distribution of apoJ-HDL suggests that these subclasses have a unique metabolic relationship and may play a role in the transport of cholesterol from peripheral tissues to the liver.     

The levels of apolipoprotein-E in hypercholesterolemic rat serum.

The levels of apolipoprotein-C (apo-E) in serum and isolated liproproteins from diet-induced hypercholesterolemic, and to some extent hypertriglycerdemic rats were measured by electroimmunoassay. The hypocholesterolemia was accompanied by a mild hypertriglyceridemia. The apo-E was increased by 60% in the hypercholesterolemic serum with a 5- and 50-fold increase in very low density lipoproteins (VLDL) and low density lipoproteins (LDL) respectively. However, the proportion of apo-E in nascent VLDL isolated from the hepatic Golgi apparatus of hypercholesterolemic rats was significantly decreased. In control serum, 40--50% of the apo-E is found in the density greater than 1.21 g/ml fraction, although this is at least partially due to ultracentrifugation. The aproprotein is absent from the density greater than 1.21 g/ml fraction from hypercholesterolemic serum, suggesting that it is bound more firmly to the lipoprotein complex. It is concluded that the large increases in apo-E in the VLDL and LDL density ranges of serum from hypercholesterolemic rats may in part be accounted for by the utilization of apo-E normally found at higher densities.     

Characterization of lipoprotein secreted by cultured eel hepatocytes and its comparison with serum lipoproteins.

The lipoprotein secreted by cultured eel hepatocytes was fractionated by density gradient ultracentrifugation and compared with eel serum lipoproteins. Eel hepatocytes were cultured for 7 to 10 days as a monolayer in Williams' medium E containing 5% fetal bovine serum and 0.16 microM insulin on a dish precoated with fibronectin of horse serum. The only lipoprotein secreted by eel hepatocytes was a very-low-density lipoprotein like one which consisted of 69% triglyceride, 15% phospholipid, 4% cholesterol, and 12% protein. On the other hand, very-low-density lipoprotein and high density lipoprotein were found in eel serum, in which high density lipoprotein was a main lipoprotein. The secreted lipoprotein contained apo B and apo A as the main protein components. Furthermore, the lipoprotein contained proapo A-I in addition to apo A-I, which was proved by comparing the amino acid composition of both proteins. In our discussion, we noted that the lipoprotein secreted by eel hepatocytes was a good material for the study of high-density lipoprotein formation.     

Characterization of an unusual lipoprotein similar to human lipoprotein a isolated from the baboon, Papio sp

An unusual lipoprotein was detected and purified from the blood of some members of a large colony of baboons, Papio sp. This lipoprotein was found to be similar to human lipoprotein a in all respects and is therefore termed lipoprotein a. Baboon lipoprotein a had a density of 1.052 g/ml and was located between low- and high-density lipoproteins in a density gradient ultracentrifugation. However, despite its greater density, baboon lipoprotein a was larger than low-density lipoprotein, based on gradient gel electrophoresis and gel filtration. The lipoprotein contained a very large apolipoprotein (apolipoprotein-lipoprotein a) which was found to consist of an apolipoprotein B linked to another protein called apolipoprotein a by a disulfide bridge(s). In all these characteristics, baboon lipoprotein a was similar to human lipoprotein a.     

A rapid electrophoretic technique for identification of subunit species of apoproteins in serum lipoproteins

The denaturing solvent tetramethylurea (TMU) delipidates and quantitatively liberates the apoproteins of human serum high-density lipoprotein (HDL) in soluble form while virtually the whole apoprotein of human lowdensity lipoprotein (LDL) is precipitated. A fraction of the apoprotein of very low density lipoprotein (VLDL) which appears to represent its content of LDL-like protein (apo B) is precipitated by this reagent, while the remaining apoprotein species are liberated in soluble form.The dissociation of the soluble apoproteins from lipid by TMU obviates the need for time-consuming delipidation by organic solvents, permitting immediate electrophoretic analysis in polyacrylamide gels. Bands are observed with mobilities corresponding to those of all the major soluble polypeptide species isolated from serum lipoproteins by ion-exchange chromatography. The apparent distribution of these elements in the different classes of lipoproteins is in agreement with findings of studies employing chromatographic methods. The predominant apoprotein of HDL, which has been identified immunochemically in VLDL, appears to comprise less than 1% of the apoprotein of VLDL from normal serum.     

Thermal behavior of human plasma high density lipoprotein.

Human plasma low density lipoprotein displays a reversible thermal transition between 20 and 40 degrees C, due to a phase transition of its core cholesterol ester from a smectic to a more liquid-like state. To determine if the cholesterol of high density lipoprotein (HDL) displays similar thermal behavior, the human lipoprotein and its extracted lipid have been examined by differential scanning calorimetry, low angle X-ray scattering and polarizing microscopy. Neither HDL2**(d 1.063--1.125--1.21 g/ml) nor HDL3(d1.125--1.21g/ml) show thermal transitions between O and 60 degrees C. By contrast cholesterol ester isolated from HDL and mixtures of cholesterol oleate and linoleate show reversible liquid crystalline transitions between 20 and 40 degreesC. X-ray scattering studies of HDL2 and HDL3 performed at 10 degreesC show no scattering fringes attributable to a smectic phase of cholesterol ester. When HDL is heated to temperatures above 60 degreesC a broad, double-peaked endotherm is observed. The first component (peak temperature=71 degreesC) corresponds to a selective release of apoprotein A-1 from the lipoprotein, and the second component (peak temperature=90 degreesC) to a more generalized disruption of lipoprotein structure with release of cholesterol ester and apoprotein A-2. Following the thermal disruption of HDL, reversible liquid crystalline transitions of cholesterol ester can be seen by differential scanning calorimetry and polarizing microscopy, showing the presence of large domains of cholesterol ester. The absence of cholesterol ester transitions in intact HDL may indicate an interaction of cholesterol ester molecules with the protein-phospholipid surface of HDL that prevents the formation of an organized lipid phase. The high temperature behavior of HDL indicates that apoprotein A-1 is less important than apoprotein A-2 in maintaining the HDL apolar lipids in the form of a stable miroemulsion.