Replacement of neutral lipids of low density lipoprotein with esters of long chain unsaturated fatty acids.

A method has recently been described by which the neutral lipids of plasma low density lipoprotein (LDL) can be extracted with heptane and replaced with exogenous cholesteryl esters. In the current studies we show that, in addition to cholesteryl esters, other esters of long chain fatty acids, including triacylglycerols and methyl esters, can be used to reconstitute the core of heptane-extracted LDL. For each of these classes of esters, the common structural requirement for substantial incorporation into LDL was the presence of at least one &-double bond in the fatty acyl chain. For example, cholesteryl oleate, triolein, and methyl oleate could each be incorporated into LDL to yield a final lipid to protein mass ratio that was greater than one. In contrast, only trace amounts of esters of saturated fatty acids, such as cholesteryl stearate, tristearin, or methyl stearate, could be incorporated into LDL despite the use of a variety of solvents and different temperatures of incubation. Incorporation of these saturated compounds was not enhanced by the inclusion of unsaturated cholesteryl esters or unsaturated triacylglycerols in the reconstitution reaction. Another class of compounds that can be incorporated into heptane-extracted LDL consists of lipids that contain a polyisoprenyl side chain, such as retinyl palmitate and ubiquinone-lo. Each of the reconstituted LDL preparations retained the ability to bind to the LDL receptor of human fibroblasts and thus to deliver its respective core lipid to cells. The current data establish that plasma LDL can be made to function as a carrier for a variety of hydrophobic compounds that contain either long chain unsaturated fatty acyl or polyisoprenyl groups. The preferential incorporation of such compounds into LDL as compared with compounds containing long chain saturated fatty acids suggests that the protein or phospholipid component of LDL may have a specific ability to interact with long chain hydrocarbons that contain one or more double bonds.

ma1 hydrolysis of the LDL'-bound cholesteryl esters satisfies the cholesterol requirements of the cell and elicits a series of regulatory responses that control the cellular cholesterol content (2).
The cholesteryl esters of LDL, which account for approximately 50% of the mass of the lipoprotein, consist chiefly of two &-unsaturated fatty acid esters, cholesteryl linoleate [C 18:2(9c,12c)] and cholesteryl oleate [C 18:1(9c)] (3). These cholesteryl esters are believed to form an apolar core that is surrounded by the more polar constituents of the lipoprotein, including various phospholipids, small amounts of free cholesterol, and a protein called apoprotein B (3, 4). The LDL receptor on cell surfaces recognizes the apoprotein B component of the lipoprotein (1,5,6). We recently demonstrated that the endogenous cholesteryl esters of LDL could be extracted with heptane and replaced with exogenous cholesteryl linoleate (7). Inasmuch as the resulting reconstituted LDL particle contained its original complement of phospholipid and protein, it retained the ability to bind to the LDL receptor in cultured human fibroblasts. As a consequence of this binding, the reconstituted LDL was taken up by the cells and the cholesterol released from the lysosomal hydrolysis of the cholesteryl linoleate elicited the same regulatory actions that occur when native LDL is incubated with fibroblasts (7).
The ability to replace the cholesteryl esters of LDL while retaining the functional activity of the LDL particle raised the possibility that molecules other than cholesteryl esters might be introduced into the LDL particle and hence delivered to cells via the LDL receptor. In an initial series of studies along these lines, we incorporated 25-hydroxycholesteryl oleate into LDL and showed that uptake of this reconstituted lipoprotein by fibroblasts was receptor-dependent (8

Reconstitution
of Low Density Lipoprotein these fatty acid esters into LDL is the presence of at least one double bond in the fatty acyl chain. Chen et al. (16).

RESULTS
In the previously described reconstitution procedure, LDL is adsorbed onto starch and its neutral lipids are extracted with heptane (11). Exogenous cholesteryl esters in heptane solution are then added to the LDL-starch residue. The heptane carrier is evaporated and the exogenous cholesteryl esters are deposited in association with LDL residue on the starch (7). To determine whether the evaporation step is required for reconstitution, we performed an experiment in which the heptane-extracted LDL was incubated for 1 h with increasing amounts of ["Hlcholesteryl linoleate dissolved in heptane ( Fig.  1). The samples were then treated in one of two ways: in one group of samples the heptane was evaporated by the standard procedure and in the second group of samples the heptanecholesteryl linoleate solution was removed by aspiration without any evaporation.
Both groups of lipoproteins were then solubilized in an aqueous buffer. Fig. 1A  Each of the reconstituted lipoprotein particles described in Table I was also tested for its ability to bind to the LDL receptor and to be taken up and hydrolyzed by fibroblasts. These experiments were performed by incubating the reconstituted lipoproteins with fibroblast monolayers at a concentration of 10 pg of protein/ml and measuring the amount of ["Hlcholesteryl linoleate hydrolyzed after 6 h. The data show that all of the recovered lipoproteins were biologically active. Even in the case of the chloroform-or acetone-treated materials, in which the amounts of lipoprotein recovered were low, the lipoprotein that was recovered in the aqueous supernatant retained biologic activity.
To determine the effect of the fatty acid moiety of the cholesteryl ester on the reconstitution reaction, we incubated heptane-extracted LDL with cholesteryl esters of saturated and cis-monounsaturated fatty acids of varying chain length. Carbon tetrachloride was employed as a solvent for these reconstitutions because the cholesteryl esters of long chain saturated fatty acids were more soluble in this solvent than in heptane. As shown in Fig. 2A, cholesteryl esters of saturated fatty acids with chain length less than 14 carbons partially protected the lipoprotein against denaturation during the solvent evaporation step and hence allowed the recovery of protein in the aqueous supematant.
" The value in parentheses represents the dielectric constant for the indicated solvent (20) except for the value for heptane. Its dielectric constant is assumed to be the same as the published values for hexane and octane, which are both 1.9 (20). lesteryl esters of saturated fatty acids with chain length greater than 14 carbons failed to protect the lipoprotein, and the protein recovery declined precipitously. Fig. 2B shows that despite their ability to protect the lipoprotein against denaturation, cholesteryl esters of short and medium chain saturated fatty acids were incorporated only to a small extent into the lipoprotein.
At chain lengths above 14 carbons, no detectable cholesteryl esters were incorporated. The low incorporation of all of the saturated cholesteryl esters into LDL was reflected in relatively low mass ratios of cholesteryl ester to protein in the reconstituted LDL particles (Fig. 2C). For the fatty acyl chains below 16 carbons, the mass ratio of cholesteryl ester to protein (mg/mg) varied in the range of 0.5 to 1. Above a chain length of 14 carbons, this ratio was 0.
A striking difference was seen when a series of cis-monounsaturated cholesteryl esters ranging in chain length from 14 to 24 carbons was utilized. All of these cholesteryl esters not only protected the lipoprotein during the solvent evaporation step ( Fig. 2A) but also were incorporated into the lipoprotein in relatively large amounts (Fig. 2B). The mass ratio of cholesteryl ester to protein for these cis-monounsaturated cholesteryl esters was approximately 2.0 (Fig. 2C).
The difference in behavior of the saturated and unsaturated fatty acids was also observed in experiments comparing cholesteryl esters of 18 carbon fatty acids that possess varying numbers and configurations of double bonds (Fig. 3) The lipoproteins recovered in various reconstitution experiments as described above were tested for biologic activity by measuring their ability to suppress HMG-CoA reductase activity in monolayers of growing human fibroblasts (Fig. 4). Because of the varying ratios of cholesteryl ester to protein in the different preparations, the concentration of cholesterol in the incubations varied from 0.2 kg/ml to 15 pg/ml. The data are plotted to show the percentage suppression of HMG-CoA reductase in each experiment as a function of the sterol concentration in the medium. The results were compared to the suppression achieved by native LDL over the same range of sterol concentrations.
In general, all of the reconstituted LDL preparations suppressed HMG-CoA reductase in proportion to their sterol content, irrespective of the chain length of the fatty acid and irrespective of whether that fatty acid was saturated or unsaturated.
These data indicate that even when low amounts of reconstituted lipoprotein were recovered and even when the mass ratio of cholesterol to protein in the recovered material was low (as in the case of the saturated cholesteryl esters) the material that was recovered remained biologically active. Fig. 5 shows that triacylglycerols containing fatty acyl groups with k-double bonds, such as triolein and trilinolein, were incorporated into LDL in amounts that were similar to those for cholesteryl linoleate. On the other hand, tripalmitin and tristearin, two triacylglycerols containing saturated fatty acyl groups, were not incorporated to any measurable extent. The requirement for unsaturated fatty acids in the reconstitution of LDL also applied when methyl esters of fatty acids were used to reconstitute LDL. Thus, methyl palmitate and methyl stearate, two saturated fatty acid esters, were not incorporated into LDL, whereas the unsaturated fatty acid esters methyl palmitoleate and methyl linoleate were incorporated in amounts that were similar to those for cholesteryl linoleate (data not shown). Fig. 6 shows the incorporation of methyl linoleate into LDL as a function of the amount of lipid added to the heptane-extracted LDL. The protein recovery (Fig. 6A) and the mass ratio of lipid to protein (Fig. 6C)  similar to values obtained in other experiments with cholesteryl linoleate (cf Fig. 5). The data in Fig. 6 also show that appreciable amounts of linoleyl alcohol could be incorporated into LDL. Two experiments were performed to determine whether the presence of an unsaturated fatty acid ester would enhance the incorporation of a saturated fatty acid ester into LDL. In the first experiment shown in Fig. 7, samples of heptane-extracted LDL were incubated with a total of 4 mg of cholesteryl ester that was made up of varying proportions of [3H]cholesteryl linoleate and unlabeled cholesteryl stearate. When the tubes contained only cholesteryl stearate, the recovery of protein was low (Fig. 7A) and no cholesteryl ester was recovered in the aqueous supernatant (Fig. 7B). As the proportion of cholesteryl linoleate was increased, the yield of protein increased. All of the cholesteryl esters in the reconstituted LDL could be accounted for by the measured amount of ["Hlcholesteryl linoleate.
No cholesteryl stearate was recovered in any of the aqueous supernatants (Fig. 7B did not suppress HMG-CoA reductase activity (data not shown) even though these reconstituted lipoproteins were taken up and hydrolyzed by the cells at rates that were similar to those of native LDL (see below). These data are in keeping with earlier studies indicating that the cholesterol of LDL is the component responsible for lipoprotein-mediated suppression of HMG-CoA reductase activity (2,6,21).
In addition to compounds that contain unsaturated fatty acids, another class of compounds that could be incorporated into LDL were those that contained polyisoprenyl chains. Table II shows that retinyl palmitate, a polyisoprenyl alcohol ester of a saturated fatty acid was incorporated into LDL in an amount similar to that obtained with cholesteryl linoleate. Ubiquinone-10, another molecule with a polyisoprenyl side chain, could also be incorporated into LDL in appreciable amounts (Table II). We have also been able to reconstitute heptane-extracted LDL with retinol and chlorophyll, both of which contain a polyisoprenyl side chain (data not shown). To test the biologic activity of the LDL preparations that were reconstituted with molecules other than cholesteryl esters, we carried out the reconstitution procedure with IZ51-LDL. The resulting lz51-labeled reconstituted LDL preparations were then incubated with normal human fibroblasts and the amount of '"'I-labeled protein that was degraded was measured (Table III) To demonstrate that the degradation was mediated by the LDL receptor, each of the ""I-labeled reconstituted LDL preparations was also incubated with cells from a patient with homozygous familial hypercholesterolemia. These mutant cells lack LDL receptors and hence are unable to take up and degrade '""I-LDL through the LDL receptor pathway (1, 2). The data in Table III show that  preparations of '251-LDL reconstituted with cholesteryl linoleate, trilinolein, methyl linoleate, and retinyl palmitate were each able to bind to the LDL receptor and be delivered to  lysosomes, as indicated by their normal rates of degradation in normal fibroblasts. In each case, the degradation was markedly reduced in the homozygous familial hypercholest,erolemia cells.

DISCUSSION
Several conclusions can be drawn from the data in the current paper. First, the neutral lipids in the core of LDL can be replaced with a wide variety of lipid-soluble molecules whose structures differ markedly from those of the cholesteryl esters that normally occupy this core. Second, the exogenous neutral lipids can be delivered in a variety of nonpolar solvents. Third, in order for the exogenous neutral lipid to be incorporated into LDL, it is necessary that the solvent be removed by evaporation so that the added lipid deposits in association with the heptane-extracted LDL residue. Fourth, the incorporation of esters of long chain fatty acids into LDL, including cholesteryl esters, triacylglycerols, and methyl esters, is markedly enhanced when the fatty acyl chain contains at least one double bond. Fifth, compounds containing polyisoprenyl chains can also be incorporated into LDL with high efficiency.
The inability to incorporate cholesteryl, glyceryl, and methyl esters of saturated fatty acids might theoretically have been due to two physical properties of the saturated esters that differ from those of the unsaturated esters. First, the esters of saturated fatty acids are generally somewhat less soluble in heptane than the corresponding esters of the unsaturated fatty acids. This limited heptane solubility might have limited the incorporation of these compounds into LDL by preventing them from reaching some critical concentration in the solvent during the evaporation step. This possibility seems unlikely in view of the finding that cholesteryl esters of long chain saturated fatty acids could not be incorporated into LDL even when they were dissolved in solvents such as carbon tetrachloride or benzene, two solvents in which these cholesteryl esters were much more soluble than they were in heptane. A second reason for the failure to incorporate the saturated fatty acid esters could have been due to the fact that esters of unsaturated fatty acids, but not saturated fatty acids, will form thermotropic liquid crystals or mesophases at temperatures below 60°C (22). This possibility seems unlikely in view of the fact that the evaporation step was routinely conducted at O"C, a temperature at which both the saturated and unsaturated cholesteryl esters exist in a solid phase (22). Moreover, cholesteryl linoleate was incorporated into LDL at both 0°C and at 37"C, whereas cholesteryl stearate was not incorporated at either temperature. Considered together, the above data suggest that neither the differences in solubility of saturated and unsaturated cholesteryl esters nor the differences in phase behavior are likely to account for the all-or-none difference in the incorporation of the saturated and unsaturated fatty acid esters into LDL. Rather, the data are more compatible with the notion that the physical characteristics of the LDL particle itself favor the incorporation of unsaturated rather than saturated fatty acid esters, thus implying that differences in the three-dimensional structures of these two classes of fatty acids are important.
Consistent with this hypothesis is the observation that the cholesteryl ester of the trans-isomer of the 18carbon monounsaturated fatty acid (elaidate) was incorporated less well than the &-isomer (oleate), suggesting that the link in the fatty acyl chain induced by the c&--configuration about the double bond plays some role in allowing the molecule to be incorporated into LDL. The observation that heptane-extracted LDL could be reconstituted with ubiquinone-109 and retinol, each of which contains a long polyisoprenyl chain, suggests that the polyisoprenyl group acts like the unsaturated fatty acyl group in facilitating the incorporation of lipophilic compounds into LDL. When a long chain polyisoprenyl alcohol (retinal) was attached to a long chain saturated fatty acid (palmitate), the polyisoprenyl chain appeared to compensate for the lack of a double bond in the saturated fatty acid, and the resultant ester (retinyl palmitate) was incorporated into LDL in large amounts. These data suggest that the lack of incorporation of cholesteryl, glyceryl, or methyl esters of saturated fatty acids