Presence and Formation of ‘Free Apolipoprotein A-I–Like’ Particles in Human Plasma
Abstract The influence of dilution on apolipoprotein (apo) A-I–containing subpopulations was studied in human plasma. Agarose electrophoresis and two-dimensional agarose nondenaturing gradient polyacrylamide gel electrophoresis were used. Both in one- and two-dimensional electrophoresis, an increase of charge was observed that resulted in an increase of subpopulations with pre-α mobility. Dilution of plasma also resulted in a decrease in the size of apo A-I–containing pre-β1 subpopulations. The existence of smaller pre-β1 particles was confirmed by subjecting undiluted and 8× diluted plasma to 3% to 16% nondenaturing gradient gel electrophoresis for 4 hours. In addition to the generally observed pre-β1 subpopulations, smaller particles similar in size to the free apo A-I were detected even in the undiluted plasma. During dilution, the proportion of larger pre-β1 particles decreased while the smaller ones increased, and in 8× diluted plasma, almost all the pre-β1 was present in smaller sizes. Using 3% to 35% nondenaturing polyacrylamide gels run for 24 hours, no pre-β1 particles could be detected in 8× diluted plasma because the small pre-β1 electrophoresed out. These studies show that pre-β1 particles can be converted to smaller ones during dilution. It also was demonstrated that “free apo A-I–like” pre-β1 particles are present in undiluted plasma. The presence of these particles may have important physiological and pathophysiological functions.
- Received April 11, 1995.
- Accepted June 6, 1995.
Initial stages of reverse cholesterol transport take place between the peripheral cells and interstitial fluid (IF) HDL.1 The smallest HDL particles, the pre-β1, are believed to participate in the first steps of reverse cholesterol transport,2 3 4 5 that is, transfer of cholesterol and possibly phospholipids from the peripheral cells to the HDL present in the IF.
We developed a quantitative, two-dimensional electrophoretic system and identified 12 HDL subpopulations in human plasma.6 The subpopulations were divided into three major groups according to their charge—pre-β, α, and pre-α. We used this system for our studies of IF lipoprotein metabolism in dogs by using prenodal peripheral lymph as the representative of IF.7 We observed major differences between the composition of plasma and peripheral lymph IF manifesting in a relative increase of both pre-β1 and pre-α subpopulations.
Because peripheral lymph HDL is 10-fold diluted when compared with plasma HDL concentrations,1 we undertook studies to evaluate the influence of dilution on the distribution of HDL subpopulations. These studies show that dilution does affect the distribution and characteristics of HDL subpopulations. We also obtained evidence that as a result of dilution, the size of pre-β1 subpopulations will change and become similar to the size of free apolipoprotein (apo) A-I.
Processing of Blood
Fasting blood samples were collected from normolipidemic human subjects and kept on ice with the following additives in final concentrations: ethylenediaminetetraacetate, 1.2 g/L; sodium azide, 0.1 g/L; gentamicin sulfate, 80 mg/L; and kallikrein inactivator (aprotinin), 10 KU/L. Plasma was separated immediately by centrifugation at 5°C. Additional preservatives were added to the plasma in 1 mmol/L final concentration: phenylmethylsulfonylfluoride in dimethyl sulfoxide (DMSO), benzamidine in DMSO, and N′ ethylmaleimide (NEMI) in water. NEMI must be dissolved just before use because it is only stable in the water phase for about 1 hour. Plasma was kept on ice and used immediately. When necessary, plasma was diluted with Tris-tricine buffer and electrophoresed promptly after dilution was obtained.
Agarose and nondenaturing 3% to 35% concave polyacrylamide gradient gel electrophoresis was carried out as described previously6 except that the samples were applied in 4% sucrose. We also used 3% to 16% linear gradient gels run in a SE600 Hoefer unit (Hoefer Scientific Instruments) without recirculation.
Nondenaturing 3% to 16% gels were run at a constant 250 V for 4 hours and at a constant temperature of 10°C; under these conditions, particles like free apo A-I did not electrophorese out of the gel. In some experiments, only the pre-β segment of the agarose gel was cut out and placed onto the 3% to 16% gradient gel and sealed with the same agarose. Three μL of purified apo A-I (10 μg/mL) was applied to each gel as an internal standard and electrophoresed similarly to the samples (both in the first and second dimensions).
In some experiments, linear 3% to 20% gradient sodium dodecyl sulfate (SDS)–polyacrylamide gels8 were used in the second dimension instead of the nondenaturing gels. Electrophoresis was carried out at a constant voltage (200 V) until the free dye front was 1 cm from the end of the gel.
Electrophoretic transfer, fixing, blocking, and immunolocalization were performed as described previously except that the 3% to 16% nondenaturing and SDS gels were transferred at 20 V for 16 hours. The measurement of the distribution of apo A-I–containing subpopulations also has been described in detail.6
Plasma was diluted 0, 2, 4, and 8 times with phosphate-buffered saline and subjected to one-dimensional agarose electrophoresis, and the distribution of HDL subpopulations was examined (Fig 1⇓ and Table 1⇓). A gradual shift of apo A-I from α mobility to pre-α occurred. The samples also were subjected in different dilutions to two-dimensional electrophoresis (Fig 2⇓). The two-dimensional electrophoresis also demonstrated an increase of the subpopulations with pre-α mobilities, with concomitant decrease in the α migrating subpopulations (Table 2⇓).
The unexpected finding was that the pre-β1 particles could not be detected in the 8× diluted plasma. We hypothesized that the reason for our inability to recognize the apo A-I–containing pre-β1 particles is that the dilution causes a modification of the pre-β1 particle, resulting in the appearance of smaller particles of pre-β1 mobility. These small apo A-I–containing particles most likely electrophoresed out of the 3% to 35% polyacrylamide gels because the gels were run to completion. To test this hypothesis, we performed two-dimensional electrophoresis in which the nondenaturing 3% to 35% polyacrylamide gel electrophoresis (PAGE) was replaced in the second dimension by 3% to 20% SDS-PAGE. Apo A-I was detected even in eightfold dilution in the SDS-PAGE gels with pre-β mobility (Fig 3⇓). This suggests that the pre-β1 particles decreased in size in high dilution, and the 3% to 35% gel could not retain the smaller pre-β1 particles.
To follow the movement of the pre-β1 particles, we electrophoresed samples in the 3% to 35% gel and terminated the electrophoresis at 12 and 24 hours. This study provided us with information on how the small pre-β1 particles separate as a function of time during electrophoresis. After 12 hours of electrophoresis, pre-β1 particles were detected with widely different migration spanning from albumin to below ovalbumin. After 24 hours, the faster-moving small pre-β1 particles could not be detected, suggesting that these particles had run out of the gel (Fig 4⇓).
To further clarify that a size change occurred in the pre-β1 subpopulations, we subjected undiluted and 8× diluted plasma to electrophoresis in a 3% to 16% nondenaturing polyacrylamide gel for 4 hours. Fig 5⇓ demonstrates that in the undiluted samples there were two large (pre-β1a and pre-β1b) particles with pre-β1 mobilities,6 whereas in the 8× diluted sample only the smaller-sized pre-β1 particle was present. It should be noted that the small particles with similar migration to free apo A-I also were detected in the undiluted sample. To better compare the size changes of pre-β1 particles, the pre-β bands from the one-dimensional electrophoresis were cut out and the agarose segments were placed side by side and electrophoresed in the 3% to 16% gel. In the undiluted samples, the majority of apo A-I was in the large pre-β1 particle, but a small amount was always present in the smaller-sized pre-β1 particles (Fig 6⇓). When the samples were diluted, the larger pre-β1 particle decreased gradually and the smaller-sized particle increased (Table 3⇓). We also compared the mobility of the smaller pre-β1 particles with free apo A-I and observed that they had a similar rate of migration (Fig 6⇓).
Fielding9 described several apo A-I–containing subpopulations with specific physiological functions by using two-dimensional electrophoresis. The most intriguing finding was the demonstration of an apo A-I–containing particle (pre-β1) capable of accepting free cholesterol from the cell membrane when plasma was incubated with cells in tissue culture for a short time.2 They also postulated that movement of cholesterol from the cell membrane to the pre-β1 particles represents the initial stages of reverse cholesterol transport. To follow this observation in vivo, we measured the apo A-I–containing subpopulations in the dog prenodal peripheral lymph and compared it with the plasma apo A-I–containing subpopulations. The prenodal peripheral lymph is an accepted model of IF in which the initial stages of reverse cholesterol transport take place in vivo.1 The relative concentration of pre-β1 particles more than doubled in the dog prenodal peripheral lymph (IF), which suggests that the interaction of peripheral cells with the plasma filtrate (IF) results in an increase of the apo A-I–containing pre-β1 subpopulations.7 We also reported that in dog prenodal peripheral lymph, the lymph/plasma concentration ratio is 0.1 both in HDL and apo A-I.10 To properly interpret changes observed in HDL subpopulations in lymph, we must consider that the prenodal peripheral lymph apo A-I concentration is 10× diluted when compared with plasma apo A-I concentrations.1 It is possible that as a consequence of dilution, certain changes take place in the structure of the HDL particles, resulting in an increase of the pre-β1 subpopulations.
The structure of HDL can be influenced by several factors both in vitro and in vivo. Continuous in vivo remodeling of HDL particles has been demonstrated, resulting in different physiological and physical characteristics.11 12 13 14 15 16 17 18 Because of the unstable nature of the pre-β1 particles, samples had to be preserved to minimize any in vitro changes that could occur; therefore, immediately at the blood sampling, specific cocktails were added as described in “Methods.”
To test the influence of the dilution on the distribution of HDL subpopulations, human plasma was diluted to 2, 4, and 8 times, and the distribution of apo A-I in the different subpopulations was determined. On agarose electrophoresis, practically no change was noted in the distribution of apo A-I–containing pre-β1 subpopulations; however, in the second dimension (3% to 35% nondenaturing PAGE), pre-β1 particles could not be detected. To confirm that pre-β1 subpopulations are present in the second dimension, we used SDS-PAGE, and the results showed that pre-β1 particles were still present in the second dimension. Because the pre-β1 particles did not disappear in the first dimension (agarose electrophoresis) or on SDS electrophoresis (second dimension) (Fig 3⇑), we concluded that as a result of dilution, the size of pre-β1 subpopulations decreased and particles were electrophoresed out of the 3% to 35% gel under the condition we used.
To follow the movement of small and large pre-β1 particles, we could obtain much better resolution using 3% to 16% linear nondenaturing PAGE run only for 4 hours. We also used free apo A-I as a reference point in this system because the molecular weight internal standards could not provide adequate information on particle size because electrophoresis was not run to completion. In the 8× diluted plasma, practically all of the pre-β1 particles were present in the small form with similar migration to the free apo-A-I (Table 3⇑). These observations also demonstrate the unstable nature of the pre-β1 subpopulations.
It is important that in the undiluted plasma sample, the small-sized pre-β1 particle could be detected. This small-sized pre-β1 particle was similar in size and migration to the free apo A-I. We have designated this particle as “free apo A-I–like” pre-β1 because in this system we cannot be certain that these particles do not contain some lipid, especially phospholipids. The chemical analysis of this free apo A-I–like pre-β1 particle is not feasible because of the small amounts of material available.
From these data, it is clear that apo A-I–containing pre-β1 particles respond differently to dilution than any other apo A-I–containing subpopulation. As opposed to other particles, the size of the pre-β1 subpopulation decreases and approaches the size of free apo A-I without change in charge. The dynamic nature, coupled with the small quantities of pre-β1 particles available, makes it difficult to obtain accurate information on its structure.12 13 14 It is unlikely that the particle is a dimer because of its size.6 As a result of dilution, a decrease in size takes place and the size of the particle becomes similar to the size of purified apo A-I (see Fig 5⇑). We hypothesized that conformational changes occur in the pre-β1 during dilution, resulting in the loss of lipids from these particles.
It is important to know whether the influence of dilution is a generalized phenomenon and applicable to other apolipoproteins. Therefore, we determined the distribution of apo A-IV in undiluted and diluted samples and performed one- and two-dimensional gel electrophoresis, as we did with apo A-I. We did not find any changes in charge, size, or distribution as a result of dilution. During dilution, we also observed an increase in charge of α particles; the distribution shifted from α to pre-α. It is possible that the lipid associated with pre-β1 could be incorporated into the α particles, altering the conformation of apo A-I and increasing its charge. Increased electrophoretic mobilities of isolated HDL3 particles have been observed when certain lipids were incorporated into the particle.19 The actual mechanism of the increase of pre-α in the prenodal peripheral lymph could be the result of dilution, but other mechanisms could not be excluded.
These data demonstrate that the increase in pre-β1 particles in the prenodal peripheral lymph is not the result of dilution, because during dilution a small decrease in pre-β1 particles was found along with a decrease in the size of the pre-β1 particles approaching the size of free apo A-I (Table 1⇑ and Figs 4⇑ and 5⇑). We hypothesize that in the original plasma filtrate at the albuminal site of the capillaries, most of the pre-β1 particles are in the free apo A-I–like form because of the dilution and as a result of the interaction with the capillary wall. The free apo A-I–like particles have the ability to accept cholesterol from peripheral cells; during this process, the particle is transformed to pre-β1. It should be noted that the lipid affinity of the monomeric free apo A-I particles is greater than the larger aggregates, and at higher concentrations apo A-I undergoes self-association.20 It has been shown that a certain fraction of apo A-I readily dissociates from the intact particle over a wide range of HDL concentrations.21
When isolated HDL was incubated with cholesteryl ester transfer protein in the presence of VLDL and LDL, a reduction in size and a progressive dissociation from HDL to essentially lipid-free apo A-I was observed. However, lipid-free apo A-I has not been identified in human plasma.16 In this study, we were able to detect a small amount of free apo A-I–like particle in human plasma. The low concentration of free apo A-I–like particle probably is due to its fast rate of turnover and possibly its incorporation to other HDL fractions. This small particle also would be filtered through the capillaries far more efficiently than the pre-β1 particles, which are at least twice as large. We postulate that this free apo A-I–like particle also contributes to the formation of pre-β1 by the interaction with plasma membranes of the peripheral cells and accepting cholesterol and phospholipid; thus, they may have an important role in reverse cholesterol transport.22 It is our belief that the free apo A-I–like particles may have important physiological and pathophysiological roles.
This work was supported by a grant from the National Institutes of Health, National Heart, Lung, and Blood Institute (HL-25596). The authors wish to thank Dr Howard Eder for his suggestions and Gae Garrard and Debbie Barra for their editorial assistance and manuscript preparation. The expert technical assistance of Colleen Tierney and Katalin Horvath is gratefully acknowledged.
Sloop CH, Dory L, Roheim PS. Interstitial fluid lipoproteins. J Lipid Res. 1987;28:225-237.
Castro GR, Fielding CJ. Early incorporation of cell-derived cholesterol into pre-beta-migrating high-density lipoprotein. Biochem Mol Biol Int. 1988;27:25-29.
Francone OL, Fielding CJ. Initial steps in reverse cholesterol transport: the role of short-lived cholesterol acceptors. Eur Heart J. 1990;2:218-224.
Kawano M, Miida T, Fielding CJ, Fielding PE. Quantitation of pre beta-HDL-dependent and nonspecific components of the total efflux of cellular cholesterol and phospholipid. Biochem Mol Biol Int. 1993;32:5025-5028.
Fielding CJ, Fielding PE. Molecular physiology of reverse cholesterol transport. J Lipid Res. 1995;36:211-228.
Sloop CH, Dory L, Hamilton R, Krause BR, Roheim PS. Characterization of dog peripheral lymph lipoproteins: the presence of a disc-shaped ‘nascent’ high density lipoprotein. J Lipid Res. 1983;24:1429-1440.
Lefevre M, Sloop CH, Roheim PS. Characterization of dog prenodal peripheral lymph lipoproteins: evidence for the peripheral formation of lipoprotein-unassociated apoA-I with slow pre-beta electrophoretic mobility. J Lipid Res. 1988;29:1139-1148.
Ishida BY, Albee D, Paigen B. Interconversion of prebeta-migrating lipoproteins containing apolipoprotein A-I and HDL. J Lipid Res. 1990;31:227-236.
Kunitake ST, Mendel CM, Hennessy LK. Interconversion between apolipoprotein A-I-containing lipoproteins of pre-beta and alpha electrophoretic. J Lipid Res. 1992;33:1807-1816.
Miida T, Kawano M, Fielding CJ, Fielding PE. Regulation of the concentration of pre beta high-density lipoprotein in normal plasma by cell membranes and lecithin-cholesterol acyltransferase activity. Biochem Mol Biol Int. 1992;31:11112-11117.
Liang HQ, Rye KA, Barter PJ. Dissociation of lipid-free apolipoprotein A-I from high density lipoproteins. J Lipid Res. 1994;35:1187-1199.
Roheim PS, Asztalos BF. Clinical significance of lipoprotein size and risk for coronary atherosclerosis. Clin Chem. 1995;41:147-152.
Davidson WS, Sparks DL, Lund-Katz S. The molecular basis for the difference in charge between pre-beta- and alpha-migrating high density lipoproteins. J Biol Chem. 1994;269:8959-8965.
Ritter MC, Scanu AM. Role of apolipoprotein A-I in the structure of human serum high density lipoproteins. J Biol Chem. 1977;252:1208-1216.
Hara H, Yokoyama S. Role of apolipoproteins in cholesterol efflux from macrophages to lipid microemulsion: proposal of a putative model for the pre-beta high-density lipoprotein pathway. Biochem Mol Biol Int. 1992;31:2040-2046.