the interaction of cholesteryl ester transfer protein and unesterified fatty acids promotes a...
TRANSCRIPT
Biochimicu et Biophysics AC&, 1045 (1990) 81-89
Elsevier
81
BBALIP 53418
The interaction of cholesteryl ester transfer protein and unesterified fatty acids promotes a reduction
in the particle size of high-density lipoproteins
P.J. Barter, L.B.F. Chang, H.H. Newnham, K.-A. Rye and O.V. Rajaram Baker Medical Research Institute, Melbourne (Australia)
(Received 10 August 1989)
(Revised manuscript received 27 February 1990)
Key words: HDL conversion; Cholesteryl ester transfer protein; Unesterified fatty acid
Purified human cholesteryl ester transfer protein (CETP) has been found, under certain conditions, to promote changes to the particle size distribution of high-density lipoproteins (HDL) which are comparable to those attributed to a putative HDL conversion factor. When preparations of either the conversion factor or CETP are incubated with HDL, in the presence of very-low-density lipoproteins (VLDL) or low-density lipoproteins (LDL), the I-IDL, are converted to very small particles. The possibility that the conversion factor may be identical to CETP was supported by two observations: (1) CETP was found to be the main protein constituent of preparations of the conversion factor and (2) an antibody to CETP not only abolished the cholesteryl ester transfer activity of the conversion factor preparations but also inhibited changes to HDL particle size. In additional studies, the changes to HDL particle size promoted by purified CETF’ were inhibited by the presence of fatty-acid-free bovine serum albumin; by contrast, albumin had no effect on the cholesteryl ester transfer activity of the CETP. The possibility that albumin may inhibit changes to HDL particle size by removing unesterified fatty acids from either the lipoproteins or CETP was tested by adding exogenous unesterified fatty acids to the incubations. In incubations of HDL with either VLDL or LDL, sodium oleate had no effect on HDL particle size. However, when CETP was also present in the incubation mixtures the capacity of CETP to reduce the particle size of HDL was greatly enhanced by the addition of sodium oleate. It is concluded that the changes in HDL particle size which were previously attributed to an HDL conversion factor can be explained in terms of the interacting effects of CETP and unesterified fatty acids.
Introduction
It is generally accepted that high density lipoproteins (HDL) play on obligatory role in the transport of cholesterol from peripheral tissues into the plasma in the first stage of the pathway known as reverse cholesterol transport. Human HDL, however, are heterogeneous, and consist of several discrete subpopu- lations of particles which differ in size, density and composition [1,2]. There is growing evidence that differ- ent HDL subpopulations have different metabolic func-
CETP, cholesteryl ester transfer protein; VLDL, very-low-density
lipoproteins; LDL, low-density lipoproteins; HDL, high-density lipo-
proteins; HDL,, high-density lipoprotein - subfraction 2; HDL,,
high-density lipoprotein - subfraction 3; SDS, sodium dodecyl
sulphate; PAGE, polyacrylamide gel electrophoresis.
Correspondence: P.J. Barter, Baker Medical Research Institute, Com-
mercial Road, Prahran, Victoria 3181, Australia.
tions [3]. This heterogeneity also has pathological impli- cations, with a low concentration of HDL cholesterol, a well documented risk factor for coronary heart disease [4], tending to reflect a selective reduction in subpopula- tions of larger, less dense HDL [5]. Thus, factors which modulate HDL subpopulation distribution are poten- tially of major importance in terms of both normal metabolism and predisposition to atherosclerosis.
Several plasma factors have been implicated in changing the particle size of HDL. For example, activi- ties of either lipoprotein lipase [6] or lecithin: cholesterol acyltransferase [7] have been reported to promote HDL particle enlargement, while the combined activities of CETP and hepatic lipase in the presence of VLDL have been shown to reduce the particle size of HDL [8,9]. There have also been reports of changes to the particle size of HDL during incubation in vitro, under condi- tions in which there is no activity of lecithin: cholesterol acyltransferase, no transfers of lipid between HDL and other lipoprotein fractions and in which there is no
00052760/90/$03.50 0 1990 Elsevier Science Publishers B.V. (Biomedical Division)
82
lipolytic activity [lO,ll]. Such changes have been attri- buted to the activity of a putative HDL conversion factor [lo-121. When added to incubations of isolated
HDL,, the HDL conversion factor promotes the forma- tion of new populations of particles, some of which are larger and others smaller than those originally present
[ll]. In the present studies, we show that the addition of
VLDL or LDL to incubations of HDL, and the conver- sion factor, results in an inhibition of the formation of
larger HDL particles while markedly increasing the appearance of smaller particles. We have also found
that preparations of the conversion factor contain CETP as their major protein component. We now present evidence that the changes to HDL particle size which have been attributed to the activity of a putative HDL
conversion factor can be explained in terms of an interaction of CETP and the unesterified fatty acids
which reside on the surface of plasma lipoproteins.
Materials and Methods
Lipoproteins
Blood from healthy male subjects aged 28-47 years was collected into tubes containing EDTA-Na, (final
concentration 1 mg/ml) and placed immediately on ice. Blood for isolation of VLDL was collected from sub- jects who had fasted for at least 12 h, while that for isolating other lipoprotein fractions was collected after a light breakfast. Plasma was separated by centrifuga-
tion at 4OC, then adjusted to appropriate densities by adding solid KBr [13]. Lipoproteins were isolated by sequential ultracentrifugation [14] using a Beckman Ti
50 rotor at a speed of 48000 rpm. VLDL (d < 1.006 g/ml) were isolated by layering plasma under 0.15 M
NaCl and spinning for 18 h. LDL (d 1.019-1.055 g/ml) were isolated with a single spin of 24 h at the lower density and two successive spins for the same period at
the higher density. HDL, (d 1.13-1.21 g/ml) and the
combined LDL and HDL fraction (d 1.019-1.21 g/ml) were isolated with a single 24 h spin at the lower
densities and two 40 h spins at the higher density. The isolated lipoproteins were dialysed against 0.01 M Tris- HCl (pH 7.4) containing 0.15 M NaC1/0.005 M EDTA- Na ,/0.003 M NaN, (Tris-buffered saline).
Isolation of preparations containing activity of the puta- tive HDL conversion factor
The presence or absence of HDL conversion activity was determined as described previously [ll]. This assay depends on the capacity of the putative conversion factor to convert a relatively homogeneous population of HDL, into new populations of particles, some of which are smaller and others larger than the parent particles. The conversion factor was purified from pooled, titrated human plasma as described [ll]. In brief, 1800 ml human plasma was treated with
(NH4)2S04 to precipitate proteins between 35% and 55% saturation; these proteins were recovered and sub-
jected to ultracentrifugation to obtain the fraction of density 1.21-1.25 g/ml. This fraction was subjected to cation-exchange, anion-exchange and hydroxyapatite
chromatography to yield 150-350 pg of protein in a final volume of 3 ml. As reported previously [ll], when
such preparations were subjected to SDS-PAGE fol-
lowed by silver staining, there was one major band equating with a protein of apparent M, of 74000 and two or three minor bands of lower molecular weight. To gain insight into the identity of these bands, samples were subjected to size exclusion liquid chromatography on two zorbax GF-250 columns (DuPont) connected in
series to a standard Beckman HPLC system. The major
protein (but not the minor proteins) was recovered in amounts sufficient to determine the N-terminal amino
acid sequence (Applied Biosystems Model 470A protein sequencer). The sequence of the first 25 amino acids of this protein was identical to that reported for CETP
[15]. The conclusion that the major protein constituent was in fact CETP was supported by the observation
that preparations containing activity of the putative
HDL conversion factor invariably contained a high level of cholesteryl ester transfer activity. Using an assay system described previously [16], the specific ac- tivity of CETP in the preparations recovered after chro- matography on hydroxyapatite (see above) was in the range of 50-100 transfer units per mg protein compared with 0.01 unit per mg in human lipoprotein-free plasma.
Assay of cholesteryl ester transfer activity
Transfer activity in a given sample was measured as the capacity of the sample to facilitate transfer of [3H]cholesteryl ester from LDL to HDL during an
incubation at 37” C [16]. Transfer activity was ex- pressed as units per ml or per mg protein, the number of units being the rate constant li for the transfer of
LDL tracer to HDL per 3 h using the formula of
Pattnaik et al. [16].
Isolation of the cholestetyl ester transfer protein Cholesteryl ester transfer protein was purified from
titrated human plasma. Ammonium sulphate precipita- tion of proteins, ultracentrifugation to remove the lipo- proteins, hydrophobic interaction chromatography on phenyl-Sepharose CL-4B and cation exchange chro- matography on CM-52 cellulose were performed as described by Pattnaik et al. [16] except that the plasma proteins were precipitated with ammonium sulphate between 35 and 55% saturation. The subsequent purifi- cation was performed by anion-exchange chromatogra- phy on a Mono Q HR 5/5 column (Pharmacia, Up- psala, Sweden) and by hydroxyapatite chromatography as described in the purification of the putative HDL conversion factor [ll]. This procedure afforded a sam-
83
ple which appeared after SDS-PAGE and silver staining as a single band of apparent M, 74000. In terms of
cholesteryl ester transfer activity, this preparation had a
specific activity of about 200 transfer units per mg protein; i.e. it was purified about 20000-fold compared with human lipoprotein-free plasma. In terms of effects
on HDL particle size, the changes promoted by this highly purified CETP were identical to those of less
pure samples isolated after chromatography on Mono
Q. In view of this and because the stability of CETP activity was much greater in the post-Mono Q than in the post-hydroxyapatite samples, the experiments shown in Figs. 3-6 were performed with samples of CETP obtained after chromatography on Mono Q. The specific
activity of those preparations was about 50 transfer units per mg protein.
Antibody to CETP
Immunoglobulin G prepared from goat antisera raised against rabbit CETP was donated by Dr. M. Abbey, CSIRO Division of Human Nutrition, Adelaide,
South Australia [17]. This antibody had been shown previously to inhibit completely the activity of CETP in rabbit plasma [17]. It was established, in the present studies, that it also completely inhibited the cholesteryi
ester transfer activity in preparations of both the HDL conversion factor and the CETP isolated from human plasma.
Incubations
All incubations were carried out in stoppered plastic tubes in a shaking water bath at 37” C. Nonincubated
control samples were stored at 4°C. Some experiments were performed in the presence of fatty acid-free bovine
serum albumin (Sigma, St. Louis, MO, U.S.A.). In other experiments, the incubations were supplemented with sodium oleate (Sigma) [18].
Processing of samples
Following incubation, samples were subjected to 16 h of ultracentrifugation at a density of 1.25 g/ml in a Beckman TL-100 table top ultracentrifuge using a Beck- man TLA-100.2 rotor at a speed of 100000 rpm. The
supernatant fractions were then subjected to gradient gel electrophoresis on nondenaturing 4-308 poly- acrylamide gels (Pharmacia-LKB, Uppsala, Sweden) [193 to define the particle size distribution of HDL [20].
Composition of HDL At the termination of incubation in some experi-
ments, samples were subjected to size exclusion chro- matography on a column of Superose 6 HR lo/30 (Pharmacia-LKB, Uppsala, Sweden) [20]. The fractions containing HDL were recovered as described [20] and assayed for chemical composition. All assays were per-
formed on a Cobas-Bio Centrifugal Analyser (Roche Diagnostics, Zurich, Switzerland). Concentrations of
total and unesterified cholesterol, triacylglycerol and phospholipids were measured using enzymatic kits
(Boehringer-Mannheim, F.R.G.). The concentration of esterified cholesterol was calculated as the difference between the concentration of total and unesterified
cholesterol. Concentrations of apolipoprotein (apo) A-I and apoA-II were determined by immunoturbidometric
assay [21]. ApoA-I and apoA-II standards were purchased from Boehringer Werke, Marburg, France and antibodies to these apolipoproteins were from Boehringer-Mannheim, F.R.G. Protein concentration was assayed by the method of Lowry et al. [22] adapted for use in the Cobas-Bio Centrifugal analyser [23].
Results
Incubation of HDL, and preparations containing activity
of the putative HDL conversion factor
In the presence of VLDL (Fig. IA). The non-in- cubated HDL, consisted of a major population of par-
ticles of radius 4.3 nm as assessed by gradient gel electrophoresis (profile I). When the HDL, were in-
cubated for 24 h at 37 o C in the presence of VLDL at a triacylglycerol concentration of 1.08 mM but in the absence of the conversion factor, the particle size of HDL did not change (profile II). Incubation of the HDL, in the presence of the conversion factor but in the absence of VLDL resulted in the formation of new
populations of both smaller (radii 3.7 and 3.9 nm) and
larger (radius 4.8 nm) HDL. Addition of increasing
amounts of VLDL to incubation mixtures which con- tained the conversion factor promoted a progressive decrease in the formation of larger HDL particles and
an increase in the formation of smaller particles. When the concentration of VLDL triacylglycerol in the in- cubation mixture was 1.08 mM, conversion to large
particles was completely inhibited. At this concentra-
tion of VLDL most of the HDL particles of radii 4.3 nm and 3.9 nm had also disappeared; the very small particles of radius 3.7 nm now predominated.
In the presence of LDL (Fig. 1B). In the absence of LDL, the conversion factor converted the HDL, into populations of both larger and smaller particles as described above. The addition of increasing amounts of LDL had an effect comparable to that of VLDL. At an
LDL cholesterol concentration of 0.58 mM the forma- tion of larger HDL particles of radius 4.8 nm was inhibited. At higher concentrations of LDL there was also a reduction in the proportion of particles of radius 4.3 nm, such that at the highest concentration of LDL cholesterol (1.62 mM) the HDL particles of radius 4.3 nm had disappeared, leaving only small particles of radii 3.7 and 3.9 nm.
4.8 4.3 3.9 3.7 4:8&3 5.93.7
STOKES’ RADIUS (nm)
Fig. 1. The particle size dist~bution of HDL after incubating HDL,
and preparations containing activity of the putative HDL conversion
factor. (A) Effect of increasing concentrations of VLDL; (B) effect of
increasing concentrations of LDL. The profiles show the particle size
distribution of HDL as assessed by gradien: gel electrophoresis.
Aliquots of isolated HDL, (final concentration of HDL cholesterol
0.43 mM in (A) and 0.40 mM in (B)) were mixed with the HDL
conversion factor (1.9 units of cholesteryl ester transfer activity per ml
of incubation mixture ml in (A) and 3.1 units per ml in (B)) and
varying concentrations of VLDL (to provide final VLDL tri-
acylglycerot concentrations ranging from 0 to 1.08 mM as indicated in
(A) or varying concentrations of LDL (to provide final LDL
cholesterol concen~ations ranging from 0 to 1.62 mM as indicated in
(B)) and incubated at 37” C for 24 h. Profiles I show samples of
nonincubated HDL, which had been kept at 4OC; profiles II show
other control samples which were incubated at the highest concentra-
tion of VLDL (A) or LDL (B) but in the absence of the conversion
factor. The remaining profiles show the size distribution of HDL after
incubating HDL, and the conversion factor in the presence of in-
creasing concentrations of VLDL (A) or LDL (B) as indicated.
Relationship between activity of the putative HDL conver- sion factor and that of CETP (Figs. 2, 3 and 4)
Since the preparations of the HDL conversion factor
used in the previous experiment contained CETP as their major protein constituent (see Materials and
Methods), it was logical to investigate whether CETP was involved in HDL conversion. As one approach to this, we used a polyclonal antibody raised against rabbit CETP (171. This antibody was shown previously to inhibit completely the cholesteryl ester transfer activity in rabbit plasma [17] and was found in the present studies also to inhibit completely the cholesteryl ester transfer activity in preparations of both human lipopro- tein-deficient plasma and the isolated HDL conversion factor (result not shown). It also inhibited the capacity of the conversion factor to promote changes in HDL
particle size (Fig. 2). In an incubation of HDL, and the conversion factor, a proportion of the original HDL of Stokes’ radius 4.3 nm (Fig. 2, profile I) were converted into new populations of larger and smaller particles
(profile II) as described above. When the mixture of
HDL, and conversion factor was incubated in the pres- ence of the anti-CETP antibody, the change in HDL particle size was completely abolished (profile III).
As a second approach to investigating the possible involvement of CETP in HDL conversion, a prepara- tion of CETP was purified from the 1.25 g/ml infra- natant of human plasma. This preparation was found to promote changes to HDL particle size (Fig. 3) which
were virtually identical to those promoted by the HDL conversion factor (Fig. 1); these effects of CETP were
apparent both in the absence (Fig. 3A) and in the presence (Fig. 3B) of LDL.
It has been reported previously that the crude 1.25
g/ml infranatant of human plasma is a rich source of
4.8 4.3 3.9 3.7
STOKES’ RADIUS (nm)
Fig. 2. Effects of an anti-CETP antibody on the capacity of the putative HDL conversion factor to change the particle size of HDL,.
Aliquots of a preparation of HDL, (to provide a final concentration
of HDL cholesterol of 0.40 mM) were mixed with the HDL conver-
sion factor (3.4 units of cholesteryl ester transfer activity per ml of
incubation mixture) and either kept at 4’C (profile I) or incubated at
37’C for 24 h in the absence (profile II) or in the presence (profile
III) of IgG (4.5 mgfml) prepared from goat antisera raised against rabbit CETP 1171. The final incubation volume was 200 ~1.
/ / \ \ 4.8 4.3 3.9 3.7
I I i i
I I I I I I I
/ I \ 1
4.8 4.3 3.9 3.7
STOKES’ RADIUS (nm)
Fig. 3. The particle size distribution of HDL after incubating HDL,
and CETP. Aliquots of isolated HDL, (final concentration of HDL
cholesterol 0.29 mM) were incubated in the absence (A) or the
presence (B) of LDL added to provide a final LDL cholesterol
concentration of 1.53 mM. Samples were either kept at 4” C (profiles
I) or incubated for 24 h at 37 o C in the absence of CETP (profiles II),
in the presence of CETP at a final concentration of 2.1 transfer
units/ml (profile III) or in the presence of CETP at a concentration
of 4.2 units/ml (profile IV). The final incubation volume was 200 ~1.
cholesteryl ester transfer activity [16] but has minimal effect on HDL particle size [lo]. In studies designed to investigate this anomaly, we examined the effect of
albumin on HDL particle size in incubations containing CETP and VLDL. In the experiment shown in Fig. 4,
incubation of HDL, in the presence of VLDL and CETP resulted in the conversion of a large proportion
of the original HDL, of Stokes’ radius 4.3 nm (profile I) into particles of radius 3.7 nm (profile II). When, how- ever, the incubation mixture also contained bovine
serum albumin, the formation of very small HDL par- ticles was substantially inhibited and a population of larger particles of Stokes’ radius 4.7 nm appeared (pro- file III). The presence of bovine serum albumin at the same concentration had no effect on the cholesteryl ester transfer activity of CETP (result not shown).
Interaction of CETP and unesterijied fatty acids in pro-
moting changes to HDL particle size (Figs. 5 and 6)
To account for the observation that bovine serum albumin had no effect on the cholesteryl ester transfer activity of CETP while markedly inhibiting its capacity to change HDL particle size, we postulated that the HDL conversion process represents an interaction be-
85
tween CETP and the unesterified fatty acids which reside on the surface of plasma lipoproteins [18]. This
issue was addressed by pre-incubating, preparations of HDL, and VLDL at 37’C for 4 h in the presence of fatty acid-free bovine serum albumin in an attempt to remove a proportion of the unesterified fatty acids
which may have been present on the surface of the
lipoproteins. The VLDL and HDL, were subsequently separated from the albumin by ultracentrifugation. When the mixture of pretreated HDL, and VLDL was
incubated in the presence of CETP, a minority of the original HDL, particles of radius 4.3 nm (Fig. 5, profile I) was converted into populations of smaller particles of radii 3.9 and 3.7 nm (profile II). When this incubation mixture was supplemented with sodium oleate at a final
concentration of 0.06 mM, the HDL, were completely converted to very small particles of radii 3.9 nm and 3.7
nm (profile III). By contrast, when the HDL, and
VLDL were incubated in the presence of 0.06 mM
4.7 4.3 3.9 3.7
STOKES’ RADIUS (nm)
Fig. 4. Effects of albumin on the changes to HDL particle size when incubated with VLDL and CETP. Aliquots of isolated HDL, (final
concentration of HDL cholesterol 0.63 mM) were mixed with VLDL
(final VLDL triacylglycerol concentration 2.4 mM) and CETP (final
concentration 1.1 units per ml) and either kept at 4OC (profile I) or
incubated at 37°C for 24 h in the absence (profile II) or in the
presence (profile III) of fatty acid-free bovine serum albumin at a
final concentration of 39 mg/ml. The incubation volume was 200 ~1.
86
sodium oleate without CETP, the particle size of the HDL, did not change (profile IV).
Synergistic effects of CETP and unesterified fatty
acids were also observed in incubations of a mixture of HDL and LDL. In the experiment shown in Fig. 6,
LDL and HDL were co-isolated as the plasma fraction
of density 1.019-1.21 g/ml. This mixture was then supplemented with CETP and incubated either in the
absence of exogenous unesterified fatty acid (panel A) or in the presence of added sodium oleate at a final concentration of 0.06 mM (panel B) or 0.24 mM (Panel C). The HDL in nonincubated samples (profiles I) contained two populations of particles of radii 5.2 nm
(HDL,) and 4.3 nm (HDL,). When incubated at 37’ C in the absence of CETP (profiles II) the size distribution of the HDL did not change whether or not the mixture
IV
4.3 3.9 3.7
STOKES s RADIUS (nm)
Fig. 5. Effects of sodium oleate. CETP and VLDL on the particle size distribution of HDL. Preparations of HDL, and VLDL were pre-in-
cubated at 37O C for 4 h in the presence of fatty acid-free bovine
serum albumin. The lipoproteins were subsequently reseparated from
the albumin by ultracentrifugation. Aliquots of the pretreated HDL,
(final concentration of HDL cholesterol 0.44 mM) and VLDL (final concentration of VLDL triacylglycerol 2.5 mM) were mixed and
either kept at 4’C (profile I) or incubated at 37O C for 24 h in the
presence of CETP alone (3.5 units/ml) (profile II), CETP and sodium
oleate (0.06 mM) (profile III) or sodium oleate alone (profile IV). The incubation volume was 200 pl.
contained exogenous unesterified fatty acids. When in- cubated in the presence of CETP at a concentration of 1.0 units/ml (profiles III) or 2.4 units/ml (profiles IV)
there was a concentration-dependent appearance of populations of smaller HDL particles of radii 3.9 and
3.7 nm. In addition, at each concentration of CETP, the extent of the conversion to smaller particles was also
dependent on the concentration of sodium oleate. At the highest concentrations of sodium oleate (0.24 mM) and CETP (2.4 units/ml), virtually all of the original HDL, and HDL, particles had disappeared, with very
small particles of radius 3.7 nm now predominating (panel C, profile IV).
Effects of incubation on the chemical composition of HDL,
(Table I)
Three experiments are shown in Table I. In each,
incubation resulted in quantitative conversion of the original HDL, of Stokes’ radius 4.3 nm into particles of
radii 3.7 and 3.9 nm, with a distribution comparable to that observed in Profile III of Fig. 5. The HDL, were mixed either with VLDL (experiment A) or LDL (ex-
periments B and C) and incubated in the presence either of samples prepared according to the protocol
designed to isolate the putative HDL conversion factor (experiments A and B) or of a mixture of CETP and sodium oleate (experiment C). In all three experiments, incubation resulted in the HDL becoming markedly depleted of cholesteryl esters (and of total core lipids) and relatively enriched in apolipoproteins and phos-
pholipids.
Discussion
It has been reported previously [ll] and confirmed in the present studies (Fig. 1) that when preparations containing activity of a putative HDL conversion factor
are incubated with isolated HDL,, there is an ap- pearance of new populations of HDL, some of which
are larger and others smaller than the starting material. The present study shows that when either VLDL or
LDL are added in physiological concentrations to a mixture of HDL, and the putative conversion factor, the formation of larger HDL particles is inhibited and conversion to populations of small HDL is enhanced. It has also been found in preparations containing activity of the HDL conversion factor that CETP is the main protein constituent and that an antibody which inhibits cholesteryl ester transfer activity also inhibits the changes to HDL size. Furthermore, a preparation of highly purified CETP was found to promote changes to HDL particle size which were identical to those attri- buted to activity of the conversion factor. Thus, there are several lines of evidence implicating CETP in the process previously ascribed to the HDL conversion factor. There are, however, circumstances in which the
87
- 5.2 4.3 3.7
Fig. 6. Incubation of HDL, LDL and CETP in the presence of varying concentrations of sodium oleate. A mixture of LDL and HDL was recovered
as the plasma fraction of d 1.019-1.21 g/ml. Aliquots of the LDL-HDL mixture (final cholesterol concentration 1.96 mM) were either kept at 4O C
(profiles I) or incubated at 37 o C for 24 h in the absence of CETP (profiles II), in the presence of CETP at a concentration of 1.0 units/ml (profiles
III) and in the presence of CETP at a concentration of 2.4 units/ml (profiles IV). The incubations were performed either in the absence of added
sodium oleate (A), in the presence of 0.06 mM sodium oleate (B) or in the presence of 0.24 mM sodium oleate (C). Incubation volume was 200 ~1.
+ 5.2 4.3 3.7
STOKES’ RADIUS (nm)
I \ \ 5.2 4.3 3.7
cholesteryl ester transfer activity does not correlate with HDL conversion activity.
The lipoprotein-free fraction of plasma is a rich source of CETP [16] but, so long as lecithin : cholesterol
acyltransferase is inhibited and other lipoproteins are excluded, it promotes relatively little change to the particle size of HDL [lo]. When VLDL are added to
incubations of HDL, and lipoprotein-free plasma, lipid transfers leave the HDL fraction depleted of cholesteryl esters and enriched in triacylglycerol, although the total
mass of HDL core constituents (cholesteryl esters plus triacylglycerol) remains essentially unchanged [24]. Since the molecular volume of triacylglycerol is greater than that of cholesteryl esters [25], the heteroexchange of the two moieties results in an increase in the core volume
and the particle size of HDL [24], although an ap- pearance of a minor population of very small, core lipid-depleted HDL particles has also been observed under such conditions [24]. When, by contrast, VLDL
were incubated with HDL, in the presence of either the putative HDL conversion factor or purified CETP, the larger HDL particles did not appear; rather, the change was now limited to an appearance of populations of small HDL particles (Figs. 1 and 4), in which the total core lipid content was reduced (Table I). This may suggest that purified CETP promotes a net mass trans- fer of cholesteryl esters from HDL to VLDL which is not accompanied by the mole for mole reciprocal trans-
fers of triacylglycerol which are observed during incuba- tion with crude preparations of CETP. However, the observation is also consistent with a proposition that HDL are modified by unesterified fatty acids in such a way that HDL particles have a reduced capacity to
accomodate core lipids. In these terms, when HDL and VLDL are incubated with crude preparations of CETP,
it would be predicted that contaminating albumin would remove a proportion of the unesterified fatty acids and
thus enhance the capacity of HDL to receive tri- acylglycerol. In the case of HDL and LDL, incubation in vitro in the presence of crude lipoprotein-free plasma
as a source of CETP results in bidirectional transfers of isotopically labelled esters between the two fractions [26,27], but minimal net mass transfer of cholesteryl esters in either direction [26,28] and no demonstrable change in the core lipid content of HDL particles. When, however, HDL and LDL were incubated with preparations of either the conversion factor or purified CETP, there was again a marked reduction in the core lipid content of HDL (Table I) which coincided with a decrease in the HDL particle size (Figs. 1, 3 and 6).
The observation (Fig. 4) that albumin inhibited the CETP-mediated reduction in HDL size, while having no demonstrable effect on the cholesteryl ester transfer activity of the CETP, led to the speculation that un- esterified fatty acids and CETP interact in the process of HDL conversion. This proposition was supported by
88
TABLE I
EJJkts oJ rncuhution on the chemicul composition of HDL
Experiment A: Incubation mixtures (600 ,uI) containing HDL, (total
cholesterol concentration 0.70 mM). VLDL (triacylglycerol concentra-
tion 3.5 mM) and a sample prepared according to the protocol
designed to isolate the putative HDL conversion factor (3.2 units per
ml of cholesteryl ester transfer activity) were either kept at 4°C or
incubated at 37OC for 24 h.
Experiment B: Incubation mixtures (600 ~1) containing HDL, (total
cholesterol concentration 0.37 mM). LDL (total cholesterol concentra-
tion 1.30 mM) and the putative HDL conversion factor (3.1 units/ml
of cholesteryl ester transfer activity) were either kept at 4O C or
incubated at 37OC for 24 h.
Experiment C: Incubation mixtures (800 ~1) containing HDL, (total
cholesterol concentration 0.50 mM), LDL (total cholesterol 2.24 mM).
CETP (3.2 units/ml of transfer activity) and sodium oleate (0.24 mM)
here either kept at 4OC or incubated at 37OC for 12 h.
Experiment Hours of HDL ‘: percentage by weight
incubation Proteinh PL’ UC ’ CE ’ TG L
A 0 45 23 3 22 6
24 54 31 2 2 11
B 0 51 22 4 23 1
24 60 31 4 3 1
C 0 49 23 3 20 5
12 63 28 3 6 1
” HDL were isolated from the incubation mixture by size exclusion
chromatography as described [20].
h Protein content was taken as the sum of apoA-I plus apoA-II.
’ PL denotes phospholipids: UC. unesterified cholesterol: CE.
cholesteryl esters; TG. triacylglycerol.
the finding that sodium oleate markedly accentuated the CETP-mediated reduction in HDL particle size but
was without effect in the absence of CETP (Figs. 5, 6). However, such an observation does not exclude a possi-
ble involvement of other products such as lysophosphatidylcholine which would also have been
removed by albumin. Unesterified fatty acids are known to be a compo-
nent of the surface of plasma lipoprotein fractions [18],
presumably as a consequence of the hydrolysis of lipo- protein triacylglycerol by lipoprotein lipase and hepatic lipase. Furthermore, there is evidence that such un- esterified fatty acids may interact with and modify the activity of CETP. For example, it has been reported that when a mixture of HDL and CETP is supple- mented by the addition of VLDL which had been pretreated with lipoprotein lipase, there is an enhanced formation of small HDL particles [29]. It has also been found that lipolytic products, specifically unesterified fatty acids, result in both an increase in the binding of CETP to lipoproteins and an increase in the rate of cholesteryl ester transfer from HDL to VLDL [18]. The synergistic effect of hepatic lipase and CETP in the reduction in size of triacylglycerol-enriched HDL [8] may also reflect an involvement of unesterified fatty
acids in the process.
The mechanism by which CETP and unesterified fatty acids lead to changes in HDL particle size is beyond the scope of the present studies. On the one hand, it is possible that unesterified fatty acids and
CETP act separately, but synergistically. For example, unesterified fatty acids may act as a primer of the
process by altering the structure of HDL particles in
such a way that their capacity to accommodate core lipids is reduced. An actual depletion of the core lipid content (and a consequent reduction in size) of the
modified particles may then be dependent on subse- quent interactions with CETP. On the other hand, how- ever, it is conceivable that the actions of unesterified
fatty acids and CETP are much more closely linked. It is possible that CETP acquires unesterified fatty acids from the surface of plasma lipoproteins and circulates in the plasma as a CETP-unesterified fatty acid com-
plex which subsequently interacts with and modifies HDL. If this is the case, the HDL conversion factor,
which has been shown to modify the size of discoidal complexes of apolipoprotein A-I and egg phosphati- dylcholine in the absence of exogenous unesterified
fatty acids [30], may be a preformed complex of CETP and unesterified fatty acid.
The physiological significance of the HDL conver- sion process is not addressed by these studies. Neverthe- less, it is tempting to speculate that during lipolysis of lipoproteins, the unesterified fatty acids released onto the lipoprotein surface may exist transiently at a high local concentration and at a site which favours interac- tion with CETP in preference to binding to albumin. If so, the process may prove to be of considerable physio-
logical importance.
Acknowledgements
This work was supported by grants from the Na- tional Health and Medical Research Council of Australia and the National Heart Foundation of Australia.
References
Kostner, G.M. (1981) in High-Density Lipoproteins (Day, E.C.,
ed.), pp. l-42, Marcel-Dekker, New York.
Blanche, P.J.. Gong, E.L., Forte, T.M. and Nichols, A.V. (1981)
Biochim. Biophys. Acta 665, 408-419.
Castro, G.R. and Fielding, C.J. (1988) Biochemistry 27. 25-29.
Miller, G.J. and Miller, N.E. (1975) Lancet i, 16-19. Anderson, D.A., Nichols, A.V., Pan, S.S. and Lindgren. F.T.
(1978) Atherosclerosis 29, 161-179.
Patsch, J.R., Gotto. A.M., Jr., Olivecrona, T. and Eisenberg, S.
(1978) Proc. Natl. Acad. Sci. USA 75, 4519-4523.
Rajaram, O.V. and Barter, P.J. (1986) Biochim. Biophys. Acta 877,
406-414.
Hopkins, G.J. and Barter, P.J. (1986) J. Lipid Res. 27. 1265-1277.
Patsch. J.R.. Prasad, S.. Gotto, A.M., Jr. and Bengtsson-Olivecrona
G. (1984) J. Clin. Invest. 74, 2017-2023.
89
10 Rye, K.A. and Barter, P.J. (1986) B&him. Biophys. Acta 875,
429-438.
11 Barter, P.J., Rajaram, O.V., Chang, L.B.F., Rye, K.A., Gambert,
P., Lagrost, L., Ehnholm, C. and Fidge, N.H. (1988) B&hem. J.
254.179-184.
12 Gambert, P., Lallemant, C., Athias, A. and Padieu, P. (1982)
B&him. Biophys. Acta 713, 1-9.
13 Hatch, F.T. and Lees, R.S. (1968) Adv. Lipid Res. 6, l-68.
14 Havel, R.J., Eder, H.A., Bragdon, J.H. (1955) J. Clin. Invest. 34,
1345-1353.
15 Drayna, D., Jamagin, A.S., McLean, J., Henzel, W., Kohr, W.,
Fielding, C. and Lawson, R. (1987) Nature 327, 632-634.
16 Pattnaik, N.M., Montes, A., Hughes, L.B. and Zilversmit, D.B.
(1978) B&him. Biophys. Acta 530, 428-438.
17 Abbey, M. and Calvert, G.D. (1989) B&him. Biophys. Acta 1003,
20-29.
18 Sammett, D. and Tall, A.R. (1985) J. Biol. Chem. 260, 6687-6697.
19 Blanche, P.J., Gong, E.L., Forte, T.M. and Nichols, A.V. (1981)
B&him. Biophys. Acta 665, 408-419.
20 Clay, M.A., Hopkins, G.J., Ehnholm, C.P. and Barter, P.J. (1989)
B&him. Biophys. Acta 1002, 173-181.
21 Austin, G.E. and Maznicki, E. (1983) Clin. Biochem. 16, 338-340.
22 Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J.
(1951) J. Biol. Chem. 193, 265-275.
23 Clifton, P.M., Chang, L. and Mackinnon, A.M. (1988) Anal.
B&hem. 172. 165-168.
24 Hopkins, G.J., Chang, L.B.F. and Barter, P.J. (1985) J. Lipid Res.
26, 218-229.
25 Shen, B.W., Scanu, A.M. and Kezdy. F.J. (1977) Proc. Natl. Acad.
Sci. USA 74, 837-841.
26 Sniderman, A., Teng, B., Vezina, C. and Marcel, Y.L. (1978)
Atherosclerosis 31, 327-333.
27 Barter, P.J. and Lally. J.I. (1979) Metabolism 28, 230-236.
28 Barter, P.J. and Jones, M.E. (1979) Atherosclerosis 34, 67-74.
29 Ellsworth, J.L., Kashyap. M.L., Jackson, R.L. and Harmony.
J.A.K. (1987) Biochim. Biophys. Acta 918, 260-266.
30 Rye, K.A. (1988) J. Lipid Res. 30, 335-346.