Opposing Effects of Apolipoproteins E and C on Lipoprotein Binding to Low Density Lipoprotein Receptor-related Protein*

The low density lipoprotein receptor-related protein (LRP) from rat liver membranes binds apoprotein E (apoE)-enriched rabbit o-migrating very low density lipoproteins (j3-VLDL) in a ligand blotting assay on nitrocellulose membranes. Binding was markedly ac- tivated when the &VLDL was preincubated with recombinant human native

The effects of apoE and apoC on the ligand blotting assay were paralleled by similar effects in the ability of @-VLDL to stimulate cholesteryl ester synthesis in mutant human fibroblasts that lack low density lipoprotein receptors. These properties of LRP are consistent with the known effects of apoE and apoC on uptake of chylomicron and very low density lipoprotein remnants in the liver and raise the possibility that LRP functions as a receptor for apoE-enriched forms of these lipoproteins in intact animals.
The low density lipoprotein receptor-related protein (LRP)' is a 4526-amino acid integral membrane glycoprotein that occupies the cell surface and intracellular vesicles of many animal cells (l-3). The protein was first identified when its cDNA was observed to cross-hybridize with an oligonucleotide that was complementary to a sequence encoding a cysteinerich element that occurs multiple times in the ligand-binding domain of the low density lipoprotein (LDL) receptor and once in several proteins of the terminal complement cascade (1). The external domain of LRP contains 4400 amino acids that consist of four imperfect copies (1) of the 767-amino acid external domain of the LDL receptor (4). Each copy contains multiple cysteine-rich ligand-binding repeats followed by an epidermal growth factor precursor homology region that contains a second class of cysteine-rich sequences designated growth factor repeats. The LRP contains a total of 31 ligand binding repeats and 22 growth factor repeats as opposed to seven and three, respectively, in the LDL receptor. Through use of different combinations of cysteine-rich ligand binding repeats, the LDL receptor is able to bind two different apolipoproteins, apoB-100 and apoE (5,6). ApoB-100, a glycoprotein of 4536 amino acids, is the sole apoprotein of LDL (7). ApoE, a 299-amino acid protein, is found in several lipoproteins, including /?-migrating very low density lipoproteins (/3-VLDL), which are cholesteryl ester-rich lipoproteins that accumulate in the plasma of cholesterol-fed animals (8). Each particle of @-VLDL contains multiple molecules of apoE plus one molecule of either apoB-100 or apoB-48. Most apoB-4%containing /3-VLDL are derived from the metabolism of dietary chylomicrons, whereas most apoB-lOOcontaining P-VLDL are derived from the metabolism of endogenous VLDL (9).
The function of LRP has been studied in mutant fibroblasts from an individual with homozygous familial hypercholesterolemia (FH) who has no LDL receptors, owing to the inheritance of two null alleles at the LDL receptor locus (2). Lipoprotein uptake was estimated by measuring the rate at which the mutant cells incorporated [%]oleate into cholesteryl ['4C]oleate. This reesterification reaction is stimulated many-fold when cells have taken up a lipoprotein through receptor-mediated endocytosis and have hydrolyzed its cholesteryl esters in lysosomes (10). When the LDL receptornegative cells were incubated with rabbit P-VLDL, there was no stimulation of cholesterol esterification (2). However, when the /3-VLDL was enriched by prior incubation with recombinant human apoE, the P-VLDL acquired the ability to stimulate the synthesis of cholesteryl esters in the mutant fibroblasts. This stimulation was blocked by an antibody against LRP, indicating that LRP mediated this uptake. The stimulation was also blocked by chloroquine, suggesting that lysosomes were involved (2). These studies were consistent with the concept that LRP may serve as a receptor for the removal of apoE-enriched remnant particles derived from the metabolism of dietary chylomicrons (2). The ability of LRP to bind apoE was recently demonstrated directly by Beisiegel et al. (ll), who showed that apoE contained in phospholipid vesicles could be chemically cross-linked to LRP in isolated membranes and in intact cells.
In the current studies we have sought to further characterize 10771 This is an Open Access article under the CC BY license.
the apoE-dependent binding of P-VLDL to LRP using a solidphase in uitro ligand blotting assay. We have also used the solid-phase assay and the intact fibroblast assay to explore the ability of various polymorphic forms of apoE to interact with LRP. Human apoE exists in three common polymorphic forms (8). The most common form, designated apoE-3, has an arginine at position 158 and binds with very high affinity to LDL receptors (8,12,13). Another form, found in patients with familial type III hyperlipoproteinemia, contains a cysteine substituted for the arginine at position 158. This form, designated apoE-2, binds poorly to LDL receptors. The third form, designated apoE-4, has an arginine at position 158 and an arginine substituted for the sole cysteine at position 112. It binds to LDL receptors with an affinity similar to that of apoE-(13). Here, we show that LRP binds all three apoEs, but that its binding capacity for apoE-is lower than that for the other two proteins.
We have also studied the effect of the C apoproteins on the apo E-dependent binding of P-VLDL to LRP. ApoCs are a mixture of three small polypeptides (designated C-I, C-II, and C-III) that are found in VLDL and high density lipoproteins (9). Previous studies by Windler and Have1 (14) and Shelburne et al. (15) showed that addition of apoE to lipoprotein particles or lipid emulsions enhances their uptake by the liver and that this effect can be overcome by adding excess C apoproteins.
The parallel effects of apoE and apoC on p-VLDL binding to LRP observed in the current study further supports the notion that LRP may serve as a receptor in the liver for apoE-enriched lipoproteins, such as fl-VLDL and chylomicron remnants.
Lipoproteins-@-VLDL (d < 1.006 g/ml) was prepared from the plasma of male New Zealand White rabbits (2-3 kg Fig. 2 shows a ligand blot performed with a partially purified fraction of rat liver membrane proteins. This fraction, which was obtained after DEAE-cellulose chromatography, contains the LDL receptor and LRP. The proteins were subjected to SDS-electrophoresis, blotted onto nitrocellulose, and incubated with biotinylated rabbit P-VLDL. After washing, the bound P-VLDL was visualized by incubation with ""I-streptavidin followed by autoradiography. In the absence of apoE  (25). When the /3-VLDL was preincubated with recombinant human apoE-3, it acquired the ability to bind to LRP, which migrated at approximately 515-600 kDa on these gels, and there was no change in its ability to bind to the LDL receptor. The binding of fl-VLDL to LRP rose progressively as the apoE-concentration was increased up to 20 pg/ml (Fig. 2).
Binding of apoE-enriched P-VLDL to LRP on ligand blots required Ca'+. Fig. 3 shows that the apoE-enriched P-VLDL did not bind either to the LDL receptor or to LRP when the incubation was performed in the absence of Ca'+ and the presence of 1 mM EDTA.
Binding to both proteins was maximal when 1.3 mM Ca2+ was present, which was only a 0.3 mM excess over the EDTA concentration.
The ability of recombinant apoE-3-enriched P-VLDL to stimulate cholesteryl [Yloleate formation by FH 808 fibroblasts was increased when the incubation was performed in the presence of /3-mercaptoethanol (Fig. 4A). The presence of reducing agent presumably disrupted intermolecular disulfide bonds that form during purification of recombinant apoE or during its incubation with /3-VLDL (19). Native apoE isolated from rabbit fl-VLDL under reducing conditions was as potent as recombinant human apoE in stimulating cholesteryl ['"Cl oleate formation mediated by rabbit p-VLDL, and there was no requirement for P-mercaptoethanol in the incubation medium (Fig. 4B). Fig. 5  was detected by incubation with "'I-streptavidin (10" cpm/ ml: 5700 cpm/ng) after which the strips were exposed to XAR-1 film for 9 h at -70 "C with an intensifying screen. The gels were calibrated vvith molecular weight markers as described in the legend to Fig. 2 three native isoforms of human apoE in enhancing the ability of @VLDL to stimulate cholesteryl ester synthesis in FH 808 fibroblasts. At low concentrations apoE-was less effective than native or recombinant apoE-3. This finding was reproduced in another experiment using a different preparation of apoE-(not shown). These preparations of apoE-were only 1% as active as apoE-and E-4 in competing for the cell surface binding of '""I-LDL to the LDL receptor of normal human fibroblasts at 4 "C (data not shown). In the experiment of Fig. 5 Bound /SVLDL was detected by incubation with ""I-streptavidin (10" cpm/ml; 20,000 cpm/ng) after which the strips were exposed to XRP-1 film for 13 h at -70 "C with an intensifying screen. The strips were calibrated with molecular weight markers as described in the legend to Fig. 2. the two apoE-preparations.
However, in other experiments, apoE-was equally as effective as E-3. We also noted that high concentrations of apoE-and apoE-inhibited the @-VLDL-stimulated enhancement of cholesteryl ester synthesis, which was a consistent finding.
In agreement with the fibroblast assays, apoE-was approximately one-third to one-half less effective than apoEand E-4 in stimulating /'SVLDL binding to LRP on nitrocellulose blots as quantified by densitometric scanning of the gels (Fig. 6). ApoE-and E-4 were approximately equivalent in this regard.
A mixture of C apoproteins isolated from human plasma abolished the ability of apoE-enriched P-VLDL to stimulate cholesteryl ester formation in FH 808 fibroblasts (Fig. 7A). Inhibition was complete at 17 pg/ml. The effect of apoC could not be overcome by increasing amounts of apoE, indicating that apoC acts in a noncompetitive fashion with respect to apoE (Fig. 7B). The inhibitory effect of the apoC apoproteins was specific; similar concentrations of human apoA-I or apoA-II had markedly less inhibitory effect (Fig. 8). The apoC apoproteins did not block the ability of 25-hydroxycholesterol to stimulate cholesteryl ester synthesis, indicating that the proteins were not affecting the esterification process itself ( Table I). The apoCs also did not inhibit the synthesis of triglyceride (Table I). A, on day 7, each monolayer received 2 ml of medium A containing 20 pg of protein/ml of fi-VLDL that had been preincubated with 20 pg/ml of r-apoE-and the indicated concentration of apoC. R. on day 7 each monolayer received 2 ml of medium A containing 20 pg of protein/ml of &VLDL that had been preincubated with the indicated concentration of r-apoE-and apoC as described under "Experimental Procedures." After 5 h at 37 "C, the cells were pulselabeled for 2 h with 0.  was detected by incubation with 'Ystreptavidin (10" cpm/ ml; 20,000 cpm/ng) after which the strips were exposed to XRP-1 film for 13 h at room temperature with an intensifying screen. The strips were calibrated with molecular weight markers as described in the legend to Fig. 2. The inhibitory effect of apoC was due to its inhibition of the binding of apoE-enriched /3-VLDL to LRP. Fig. 9 shows that increasing concentrations of apoC abolished this binding on nitrocellulose blots, whereas apoA-I and apoA-II had no effect. At a concentration of apoC (5 pg/ml) that markedly inhibited /3-VLDL binding to LRP (70% inhibition by scanning densitometry), apoC did not inhibit @-VLDL binding to the LDL receptor (Fig. 9), nor did it inhibit the fi-VLDLmediated enhancement in cholesteryl ester synthesis in normal human fibroblasts (Table II), which take up fl-VLDL predominantly through the LDL receptor (2). In two ligand blotting experiments with different apoprotein preparations, apoC was at least S-fold more potent in inhibiting /3-VLDL binding to LRP than to the LDL receptor (data not shown). The apoC concentration that gave 50% inhibition averaged <5 pg/ml for LRP compared with >40 pg/ml for the LDL receptor. and lanes 3 and 6, void fraction containing 13 A.'x,, units of gel-filtered &VLDL that had been incubated with r-apoE as described above. The filters were incubated for 1 h at room temperature with 1.5 pg/ml of the indicated monoclonal IgG antibody directed against apoE, followed by incubation for 30 min with Ylabeled rabbit anti-mouse IgG (10" cpm/ml). After washing and drying, the filters were exposed to Kodak XAR film for 4 h at -70 "C with an intensifying screen. The position of migration of molecular weight markers is shown.
The experiment of Fig. 10 was performed to determine whether rabbit /?-VLDL and human apoE form a stable complex after preincubation. For this purpose we employed two monoclonal antibodies prepared against human apoE (17,18 in the legend to Fig. 10 except that biotinylated @VLDL rather than untreated fl-VLDL was preincubated with r-apoE prior to gel filtration. Bound @-VLDL was detected by incubation with "'Istreptavidin (10' cpm/ml; 30,000 cpm/ng) after which the strips were exposed to XRP film for 12 h at -70 "C with an intensifying screen. The position of migration of LRP and LDL receptors (LDLR) was verified as described in the legend to Fig. 2.
whereas antibody 3B7 reacts only with human apoE. As shown in Fig. 10, recombinant human apoE reacted with both antibodies (lanes 1 and 4). Rabbit P-VLDL reacted with 3Hl (lane 2), but not with 3B7 (lane 5). When /3-VLDL had been incubated with recombinant human apoE and was then reisolated by gel filtration, the @-VLDL in the void fraction contained human apoE as indicated by its equal reactivity with the 3Hl antibody (lane 3) and the 3B7 antibody (lane 6). The apoE-enriched P-VLDL that was isolated in the void fraction of the gel filtration column bound to LRP on nitrocellulose blots (Fig. 11). We previously described a polyclonal rabbit antibody that was prepared against LRP purified from rat liver (2). When incubated with FH 808 fibroblasts, this antibody prevented the stimulation of cholesteryl esterification by apoE-enriched /3-VLDL. Fig. 12 shows that this antibody did not directly block the binding of apoE-enriched @-VLDL to rat LRP on nitrocellulose blots. The antibody did, however, cause the rapid degradation of LRP in intact cells. Fig. 13 shows immunoprecipitates of NRK cells that were pulse-labeled with ["'Slcysteine and then incubated for varying times with anti-LRP. In the absence of antibody treatment (lane I), three immunoprecipitable forms of LRP were observed: a 600-kDa band that represents the intact protein and a 515-kDa band and 85-kDa band that result from a specific proteolytic cleavage.' After incubation of the intact cells for 3 h with anti- with ""I-streptavidin (IO6 cpm; 20,000 cpm/ng), after which the strips were exposed to XRP-1 film for 14 h at -70 "C with an intensifying screen. The strips were calibrated with molecular weight markers as described in the legend to Fig. 2.
LRP (lane 4), the amount of immunoprecipitable LRP was markedly reduced when compared with cells that had been incubated for 5 h with nonimmune serum (designated zero time in Fig. 13, lane I), and a new degradative product migrating at -110 kDa appeared (lane 4). This rapid degradation of LRP presumably accounted for the previously observed antibody-induced inhibition of /3-VLDL/apoE-mediated stimulation of cholesteryl esterification in fibroblasts (2).

DISCUSSION
The current paper describes several properties of LRP that are consistent with its postulated role as a remnant receptor. These include 1) its ability to bind apoE-enriched &VLDL on nitrocellulose blots, which correlates with the ability of apoE-enriched fl-VLDL to stimulate cholesteryl ester formation in fibroblasts, 2) its ability to distinguish between apoE-2 and the other apoE isoforms, and 3) its sensitivity to inhibition by the C apoproteins. All of these properties correlate with known parameters for the uptake of chylomicrons and VLDL remnants in intact livers (8,14,15,33,34).
The opposing effects of apoC and E on remnant uptake by the liver have been studied most extensively by Have1 and coworkers (14,33,34). These workers showed that freshly secreted VLDL and chylomicrons are taken up at measurable but relatively slow rates by perfused rat livers. Incubation of (515 and 85 kDa these lipoproteins with VLDL-free plasma produced an increase in the apoC content and a concomitant decrease in the rate of hepatic uptake, even though apoE remained on the particles. The inability of apoE to overcome the inhibitory effect of apoC was particularly evident in the case of small chylomicrons, which showed a 20-fold increase in apoE content after incubation with VLDL-free plasma. Nevertheless, this incubation decreased hepatic uptake, apparently owing to a simultaneous increase in apoC. When VLDL or chylomicrons were injected into eviscerated rats, the apoCs came off of the particle, and this was followed by rapid hepatic uptake (34). These studies and others (14,15,33) have led to the concept that hepatic uptake of remnants is governed by the balance of apoE and C on the particle. Although the characteristics of the remnants that favor hepatic uptake have been well defined by the studies discussed above, the putative hepatic receptor that mediates this uptake has remained obscure. In particular, it has not been clear as to whether this uptake is mediated by the LDL receptor or by another receptor with partially overlapping properties. The notion of a separate receptor emerged from observations on individuals with homozygous FH, who lack LDL receptors and have a marked delay in clearance of intermediate density lipoproteins and LDL from plasma. Despite this deficiency, there is no evidence for delayed clearance of chylomicron remnants (35). The same is true in Watanabe-heritable hyperlipidemic rabbits, whose mutation in the LDL receptor gene severely impairs its function (36). Intravenous infusion of apoE into Watanabe-heritable hyperlipidemic rabbits (37, 38) or into cholesterol-fed normal rabbits that have downregulated their LDL receptors (37) lowers plasma cholesterol levels, further suggesting that apoE may mediate the uptake of lipoproteins through pathways independent of the LDL receptor. The initial phase of this uptake appears to involve the sequestration of lipoproteins in the space of Disse (8,39). This sequestration is followed by uptake of the particle into hepatocytes, an event that is likely to be mediated by a receptor with the properties of LRP.
Although the current studies are consistent with the notion that LRP may function as a chylomicron remnant receptor, they leave many questions unanswered.
Why is it necessary to add exogenous apoE to @-VLDL, which already contains abundant apoE? Coomassie Blue-stained electrophoresis gels of the unincubated rabbit @-VLDL used in these studies showed that apoE was the most abundant protein, staining much more darkly than either apoB-100 or apoB-48. It is possible that the endogenous apoE is inactive by virtue of an association with C apoproteins and that additional apoE is required in order to exceed the inhibitory capacity of the apoCs. Arguing against this notion is the finding that the inhibitory effect of additional apoC cannot be overcome by raising the amount of apoE (Fig. 7B).
An alternative possibility is that the endogenous apoE on circulating fl-VLDL is in a conformation that is inactive for binding to LRP. When we extracted apoE from rabbit @-VLDL and then incubated it with fresh rabbit /3-VLDL, the native rabbit apoE was just as effective as human recombinant apoE in stimulating cholesterol esterification in FH 808 fibroblasts (Fig. 4B). Thus, the ineffectiveness of the apoE on p-VLDL is not due to a permanent modification but rather to some modification that is reversed when the apoE is removed from the lipid/apoprotein environment of P-VLDL and then reconstituted.
It seems likely that the apoE on the /3-VLDL can also be activated in vivo and that this may trigger the final endocytic uptake of a partially metabolized chylomicron or VLDL remnant. The activation might be produced by the acquisition of additional apoE, as might occur if the lipoprotein was trapped in the space of Disse where it would encounter newly secreted apoE. Such activation might involve a loss of C proteins that unmasks the apoE already resident on /3-VLDL. Or, it might be triggered by a change in the phospholipid or core lipid components of P-VLDL resulting from the action of hepatic lipase (40). In order to solve this problem, it will be important in the future to determine the stoichiometry of the interaction between apoC and apoE on P-VLDL. Can one apoC molecule inactivate multiple apoEs? Are all of the apoCs effective, or is there selectivity for one of the different apoCs?
Another question relates to the tissue distribution of LRP.
LRP is found nearly ubiquitously in animal cells (Ref. 1).3 The bulk of chylomicron remnants, however, are taken up by the liver and by the bone marrow in some species (41,42). These relative uptake rates do not correlate with the relative abundance of LRP in the different organs. This discrepancy may arise because a rate-limiting step in remnant removal is penetration across capillary endothelia. The fenestrated endothelium of the hepatic sinusoids would allow relatively rapid access to the LRP on the hepatocyte surface. Such access may be much slower in other tissues. It is also possible that in most tissues LRP is not displayed predominantly on the cell surface but functions more restrictively in intracellular lipoprotein transport.
ApoE-is clearly defective in its ability to bind to LRP, but the degree of reduction is less than the loss of binding to the LDL receptor (12). It has recently been shown that the arginine to cysteine substitution at position 158 in apoEdoes not directly disrupt binding to the LDL receptor. Rather, this substitution appears to alter the conformation of the receptor-binding domain (residues -140-150) of apoE and prevents normal receptor interaction (43,44). Furthermore, it has been shown that the receptor-binding activity of the 158 variant of apoE-can be modulated under a variety of conditions from very inactive to active (44). Most subjects who are homozygous for the apoE-isoform do not accumulate large amounts of chylomicron and VLDL remnants in their plasma unless they have some other defect such as hypothyroidism or heterozygous FH (45). These latter events may somehow alter the activity of apoE-2, perhaps by changing the size or composition of the remnant lipoproteins (44). It is possible that the intermediate result obtained with apoE-2 binding to LRP relates to the fact that some of the apoprotein is present in an active form and some in an inactive form or that the multiple ligand-binding domains in the LRP provide additional opportunities for the binding of the defective apoE-2. Regardless of m, the defective apoE could lead to increased pla remnant lipoproteins seen in patients with type III hyperlipoproteinemia if, in fact, the LRP is involved in remnant metabolism.