Only Multimers of a Synthetic Peptide of Human Apolipoprotein E Are Biologically Active*

Plasma apolipoprotein E (apoE) is a ligand for the cellular uptake of cholesterol-rich plasma lipoproteins. ApoE also inhibits mitogen-stimulated lymphocyte proliferation and gonadotropin-stimulated ovarian thecalinterstitial cell androgen production. To address the mechanism(s) by which apoE is active and to understand its interaction with the target cells, we pre- pared and examined the inhibitory activity of a series of apoE synthetic peptides. ApoE peptides representing aminoacidresidues93-112,141-155,161-171,172- 182, and 174-193 were not active in either bioassay. However, specific inhibition of both lymphocyte pro- liferation and ovarian androgen production was observed with a self-conjugate of peptide-(141-155). Furthermore, a synthesized dimeric peptide represent- ing two repeats of sequence-(141-155) (Le. (141- 155)-(141-155)) was active as well. In both bioassays, the inhibition observed was not a result of direct cell killing. Furthermore, these apoE peptides exhibited activities with characteristics that were shared with those of native apoE. The results indicate that amino acid residues 141-155 of apoE are responsible for the biological activity of apoE. Furthermore, the results suggest that dimers or multimers of native apoE may be a biologically important species.

lung, spleen, testes, ovary, and brain. Although in some circumstances locally produced apoE probably participates in the regional tissue or cellular transport of cholesterol, it is clear that this is not the only function for locally produced apoE. In fact, we have described two in vitro functional systems, which are regulated by apoE, that do not appear to involve cholesterol transport (8-10). ApoE-containing lipoproteins as well as delipidated apoE inhibit mitogenor antigen-stimulated lymphocyte proliferation (8,9). ApoE also inhibits luteinizing hormone (LH)-stimulated ovarian thecal interstitial cell androgen production (10). In both systems, the regulatory activity of apoE is independent of its role in the transport and cellular uptake of cholesterol-rich lipoproteins. Furthermore, in both systems, there is local and regulated production of apoE by nearby accessory cells, specifically the monocyte-derived macrophage (3) and the ovarian granulosa cell (11). Both of these accessory cells work in collaboration with their respective apoE-sensitive effector cells (lymphocytes and ovarian theca/interstitial cells) to help regulate expression of differentiated functions of the immune and reproductive systems.
Our ultimate goal is to understand the mechanism of inhibition by apoE. To this end, we must identify the essential structural requirements of the ligand and characterize its mode of interaction with the target cell. The cellular receptor(s) that mediate apoE inhibition of lymphocyte proliferation and ovarian androgen production have not been identified. There are several cell-surface proteins that specifically bind apoE with high affinity. The best characterized is the LDL receptor (2). A number of studies that have made use of natural mutants (I), site-directed mutants ( E ) , monoclonal antibodies (13), amino acid modifications (l), and proteolytic fragments (14) have localized the LDL receptor-binding region of apoE to amino acid residues 140-160. Furthermore, it has been demonstrated that multiple copies of apoE (up to four) can bind one LDL receptor (15, 16). However, it is unlikely that the inhibitory activity of apoE is mediated by LDL receptors because lymphocytes, which lack LDL receptors, are inhibited by apoE (17) and ovarian androgen production is stimulated rather than inhibited in the presence of excess LDL (18). Recent reports describe additional membrane-binding proteins that bind apoE including nucleolin (19) and the chicken oocyte receptor specific for very low density lipoproteins and vitellogenin (20). The participation of these receptors in mediating the inhibitory activity of apoE has not been assessed. However, a promising candidate for a receptor that could mediate the inhibitory activity is the LDL receptor-related protein (LRP) (21-23). LRP is widely distributed and appears to specifically bind apoE-enriched lipoproteins. But, unlike the LDL receptor, it does not bind apoEfree LDL. Nevertheless, LRP has >40% homology to the LDL receptor and in a simplified comparison can be thought of as 15009 Biologically Active Synthetic Peptides of ApoE containing the equivalent of at least four LDL receptor ligandbinding domains (21). Also, it appears that multiple apoE molecules are required for binding to one LRP because only apoE-enriched @-very low density lipoproteins will interact with LRP (22, 24). The ligand-binding domain of apoE that binds LRP has not been determined. But given the extensive homology between LRP and the LDL receptor, it is possible that the same region of apoE (residues 140-160) is involved in binding to LRP. In support of this prediction, it has been shown that apoE2, a naturally occurring isoform that has a cysteine substituted for an arginine at position 158 and has 1% of normal LDL receptor binding, also has diminished interaction with LRP (26). Therefore, to understand the mechanism of action of apoE inhibition of lymphocyte proliferation and/or ovarian androgen production and to identify which receptors mediate this inhibition, we have worked to define a more restricted region of apoE that has inhibitory activity. To this end, we prepared several synthetic peptides that represented specific regions of apoE. Furthermore, to reproduce the multimeric property of apoE that may enable it to bind to cellular receptors such as LRP, self-conjugates were made of these same synthetic peptides. To document biological activity, we screened these monomeric synthetic peptides and chemical self-conjugates for their ability to inhibit mitogen-induced lymphocyte proliferation and hormone-induced theca/interstitial cell androgen production.

MATERIALS AND METHODS
Peptide Synthesis-Each peptide was synthesized from the primary sequences of the native proteins as published by Rall et al. (27), Sodroski et el. (28),and Brewer et al. (29) as obtained from the GenBank Data Bank of the General Clinical Research Center of Scripps Clinic and Research Foundation (La Jolla, CA). The peptides were synthesized according to the solid-phase method of Merrifield (30) on an Applied Biosystems Model 430A automated peptide synthesizer using hydroxybenzotrizole hydrate/dicyclohexylcarbodiimide activation as described previously (31). The resultant peptide-resins were treated with 10% anisole/hydrogen fluoride at -4 "C for 1 h (32). The peptide preparations (10 pg in 0.01 ml) were analyzed by HPLC using a reverse-phase Vydac CIR column. The starting buffer contained 80% of 0.1% trifluoroacetic acid in HI0 (solvent system A) and 20% of 0.1% trifluoroacetic acid in acetonitrile (solvent system B). The run consisted of a 20-70% gradient increase in solvent system B over 20 min at 40 "C. The separations were monitored at an absorbance of 214 nm. Preparative purification of the peptides was obtained by chromatography on a Waters Auto 500 preparative HPLC column (50 X 250-mm reverse-phase Vydac CIS column, 15-20 pm) using the same conditions described above for analytical chromatography. The amino acid compositions were determined on peptide preparations hydrolyzed according to Hirs et al. (33) on a Beckman Model 6300 high performance analyzer operated with internal standards, and the absorbance was recorded at 214 nm. The purity as obtained from the HPLC chromatograms and composition of all peptides used in this study are listed in Table I. Peptides were lyophilized and stored in a dark environment under vacuum. To prepare them for addition to the cultures, all peptides were dissolved in phosphate-buffered saline (PBS) and dialyzed overnight in M , 1000 cutoff dialysis tubing (Spectrum Medical, Los Angeles, CAI. Peptide concentrations were initially obtained from dry weight measurements. However, each peptide solution was subsequently assayed for protein using a modified Lowry assay (see Ref. 34) so that recoveries could be monitored following dialysis and self-conjugation as discussion below. Peptide Conjugation-Peptides were self-conjugated following the procedure of Hoare and Koshland (35). In a typical reaction, a 1-2 mM solution of peptide in water was mixed with l-ethyl-343-di-methylaminopropy1)carbodiimide hydrochloride (EDC) to a final concentration of 0.5 M EDC. After 5 min, 0.1 N HCI was added to the reaction mixture to maintain the pH at <4.0. The reaction mixture was rotated continuously for 1-2 h at room temperature and quenched with acetate buffer (NaC2H302-CH,COOH) at pH 4.75 a t a final concentration of 1 M. The self-conjugated peptide was finally dialyzed in an M , 2000 cutoff membrane against 0.5 M acetic acid to remove unreacted monomer and lyophilized. Yields from the self-conjugation were consistently between 30 and 70%. A modified SDS-polyacrylamide gel electrophoresis method employing a 20% uniform polyacrylamide gel with 0.5% cross-linking and 10% glycerol (36) was used to evaluate the size of the molecular species formed during self-conjugation. All conjugates were dissolved in PBS and added to cell culture on the basis of protein measured in a modified Lowry assay (see Ref. 34).
Lymphocyte Cultures-Human peripheral blood mononuclear (PBM) cells were isolated and assayed for phytohemagglutinin (PHA)-stimulated proliferation according to the method of Pepe and Curtiss (9) with two changes. The fetal bovine serum was reduced to 5%, and the antibiotics were changed to penicillin (200 units/ml) and streptomycin (0.2 mg/ml). Briefly, PBM cells (200,000/0.2 ml) were incubated at 37 "C with 0.05 ml of peptide for 18-24 h before the addition of PHA. After 72 h of incubation, the cultures were pulsed for 18 h with 0.5 pCi of ["Hlmethylthymidine and harvested onto glass fiber filters with a MASH unit. The radioactivity of the filters was counted in the presence of 1.0 ml of 3a20 complete counting mixture (Research Products International Corp., Mt. Prospect, IL) in a Beckman liquid scintillation counter. The results were expressed as the mean cpm f S.D. or as percent inhibition that was derived from 100 X (1 -(cpm test culture/cpm control culture)). All data shown are representative of a single experiment, and all experiments were repeated between two and four times to verify that the results reported were reproducible.
PBM cell viability following exposure to the peptides was assessed by measuring lactate dehydrogenase (LDH) release using a Sigma diagnostic LDH assay kit. Following culture, the cell supernatants were diluted 10-fold in 0.1 M sodium phosphate buffer, pH 7.4 (0.4 ml), and 6 mM NADH (0.4 ml) was added. Immediately after adding 0.2 ml of 6 mM sodium pyruvate, the absorbance was measured kinetically at 340 nm in a Pharmacia LKB Biotechnology UV spectrophotometer. The relative concentration of LDH in the culture supernatants was compared to total LDH release obtained by freezethaw lysing of the cells in water. The results were expressed as percent cytolysis.
Ouarian Cell Cultures-Rat ovarian theca/interstitial cells were isolated and cultured as described by Dyer and Curtiss (10). Briefly, the cells from collagenase/DNase-dispersed ovaries from female hypophysectomized immature rats were cultured (20,000/0.25 ml) in serum-free McCoy's 5a medium supplemented with penicillin (200 units/ml) and streptomycin (0.2 mg/ml), 300 Fg/ml human high density lipoproteins (IO), and 4 ng/ml ovine LH (National Institute of Diabetes and Digestive and Kidney Diseases ovine LH 13V25). The peptides (0.05 ml) were added at the initiation of the cell culture period. Supernatants were collected from the cells after 48 h and stored frozen until the steroids (progesterone and androstenedione) were assayed by radioimmunoassay (RIA) as described (10). The results were expressed as the mean nanograms of steroid/milliliter f S.D. or as percent inhibition that was derived from 100 X (1 -(nanogram/milliliter test culture/nanogram/milliliter control culture)). All data shown are representative of a single experiment. However, all experiments were repeated multiple times to ensure reproducibility.
HepG2 and Rabbit Arterial Smooth Muscle Cell Cultures-HepG2 cells were acquired from the American Type Culture Collection (HB8065) and grown in Dulbecco's modified Eagle's medium (GIBCO) with sodium pyruvate (1 mM), minimal essential medium nonessential amino acids (0.1 mM), and 10% fetal bovine serum. Rabbit arterial smooth muscle cells were kindly provided by the Specialized Center for Research in Atherosclerosis tissue culture facility at The University of California (San Diego). The aortic arch was removed from New Zealand White rabbits and washed in PBS. The upper cell layer was removed with a scalpel, and the tissue pieces were grown under glass coverslips in the same Dulbecco's modified Eagle's medium.
The peptides were tested by adding them to cultures of growing HepG2 and rabbit arterial smooth muscle cells (0.5 X lo6 cells/ml) maintained in 12-well Costar tissue culture plates. After 96 h of incubation, the cells were harvested. Viability was assessed by trypan blue dye exclusion, and growth was assessed by measuring increases in cellular DNA (37). 93-112, 141-155, 161-171, 172-182, and 174-193 were synthesized and characterized (Table I). These different regions of apoE were selected because they represented the putative LDL receptorbinding region and putative heparin-binding sites (1) or they were predicted to be antigenic determinants by the program of Jameson and Wolf (38) and therefore had a better probability of being expressed on the surface of native apoE. When tested for biological activity, none of these synthetic apoE peptides inhibited either lymphocyte proliferation or ovarian androgen production. Because the monomeric apoE synthetic peptides were consistently inactive, we prepared a series of self-conjugates of each of the peptides using EDC-catalyzed coupling. In addition, we prepared self-conjugates of synthetic peptides that represented three regions of another apolipoprotein, apoA-I, which does not have immunoregulatory activity (8). These apoA-I peptides contained multiple lysine, arginine, and aspartic acid residues, making them good candidates for EDC-catalyzed self-conjugation. As illustrated in Fig. 1, a self-conjugate of the apoE peptide containing amino acid residues 141-155 was inhibitory. Only increasing amounts of this self-conjugate inhibited both the thymidine uptake of PHA-stimulated lymphocytes (Fig. 1A) and the androgen production of LH-stimulated ovarian theca/interstitial cells (Fig. 1B). Self-conjugates of all other peptides tested had no inhibitory activity including: apoE-(93-112) and apoE-(172-182) (Fig. 1) and apoE-(161-171), apoE-(174-193), apoA-I-(79-95), apoA-I-(95-105), and apoA-1-(100-105) (data not shown). The absence of inhibition with all but one of the peptides indicated that the self-conjugation procedure did not generate active peptides. More important, it suggested that the inhibitory activity was restricted to a discrete amino acid sequence within the apoE molecule.

Peptides representing human apoE residues
Two assessments of cell viability were used to verify that the inhibitory activity of self-conjugated apoE-( 141-155) was attributable to a specific biological activity. In the lymphocyte system, cell viability was assessed by monitoring trypan blue exclusion and the loss of intracellular LDH activity. Doses of self-conjugated peptide (3-4 pg/ml), which resulted in >90% inhibition of PHA-stimulated [3H]thymidine uptake, had no effect on cell viability as assessed by trypan blue exclusion or on the retention of cell-associated LDH activity (Fig. 2). Therefore, the inhibition of lymphocyte proliferation by selfconjugated apoE-(141-155) appeared to be peptide-specific.
An equivalent assessment of cell viability in the ovarian system was accomplished by the simultaneous measurement of progesterone production. Theca/interstitial cells synthesize androstenedione from progesterone precursor, and we have reported that apoE-containing lipoproteins selectively inhibit this conversion (10). Therefore, both progesterone and androstenedione were assayed in culture supernatants from theca/ interstitial cells that had been treated with increasing concentrations of self-conjugated apoE-( 141-155). Whereas androstenedione production was inhibited at 20 pg/ml self-conjugated apoE-(141-155), this same concentration of peptide had no effect on progesterone production (Fig. 3). Because these cells were able to produce progesterone, the inhibition observed with the self-conjugated apoE-(141-155)-peptide could not have been a result of decreased cell viability.
We next attempted to demonstrate that the self-conjugate was a true apoE functional mimic, i.e. it could mimic the inhibition observed with intact apoE. It has been repeatedly demonstrated that the inhibition by apoE of mitogen-stimulated lymphocyte proliferation is not readily reversible (9). Reversibility of the inhibition by the self-conjugated peptide was tested by incubating lymphocytes overnight with peptide and then washing the cultures before adding PHA. As shown in Table 11, a self-conjugate of an apoA-I peptide (apoA-1-(95-105)) had no activity, whereas the apoE peptide was maximally inhibitory. More important, the inhibition of [3H] thymidine uptake by PBM cells exposed to self-conjugated apoE-(141-155) was not reversed by washing away noncellassociated peptide before PHA stimulation. Therefore, the inhibitory activity of self-conjugated apoE-( 141-155) mimicked the irreversible inhibitory activity that we had observed previously (9) with native apoE. Another property of apoE inhibition of PBM cell proliferation is the attenuation of its activity if it is added within 12 h of PHA stimulation (9). However, this was not the case with the self-conjugate of         "At 24 h, the cells were washed three times in medium and resuspended to their original volume in RPMI 1640 medium containing 5% fetal bovine serum and PHA was added.
Peptides were used at a final concentration of 40 Fg/ml. The inhibitory activity of apoE in the ovarian steroidogenesis system has a number of distinctive features. The inhibition is selective for only androstenedione production because progesterone production by these cells is unaffected (10). As shown above (Fig. 3), the self-conjugated apoE-(141-155)peptide had no effect on progesterone production. Another hallmark of apoE inhibition of theca/interstitial cell androstenedione production is that the inhibition is reversible (10). To test if the activity of self-conjugated apoE-(141-155) was reversible, theca/interstitial cells were cultured with the peptide for 48 h. At the end of this culture period, the medium was removed and replaced with fresh medium without the self-conjugated peptide. As shown in Table 111, theca/interstitial cell androstenedione production was inhibited 50% by 0.25 pg/ml self-conjugated apoE-( 141-155) during the first 48 h of culture. However, after the peptide was removed and the theca/interstitial cells were refed with fresh medium, their androstenedione production during the subsequent 48 h was equivalent to control cells that had not been exposed to peptide. This experiment supported our earlier conclusion that the inhibition caused by self-conjugated apoE4141-155) was not a result of decreased cell viability because the cells restored their androstenedione production. More important, it supported our hypothesis that the self-conjugated peptide may be a true apoE mimic.
EDC catalyzes the formation of a peptide bond between activated carboxyl groups and amino groups. The presence of

FIG. 4. Size analysisof EDC-cross-linked apoE-(141-155)peptides.
Fifty pg of the monomer or EDC-conjugated monomer was electrophoresed on a 20% polyacrylamide gel with 0.5% cross-linking in the presence of 10% glycerol, 0.1% SDS, and a separating gel buffer of pH 9.3. Following electrophoresis, the peptides and molecular weight markers were fixed with glutaraldehyde for 12 h and visualized by staining with Coomassie Blue R-250. The three species observed in the EDC conjugation reaction mixtures were estimated to have molecular weights (1800,3800, and 5800) that were consistent with the existence of monomeric, dimeric, and trimeric peptides, respectively. multiple lysine, arginine, and aspartic acid residues in the apoE-( 141-155)-peptide (Table I) makes it difficult to predict all of the molecular species that were generated during the self-conjugation reaction. All primary amine and terminal carboxyl groups as well as the amine and carboxyl side groups could potentially react to yield multiple cross-linked species. Analysis of self-conjugated apoE-(141-155) on SDS-polyacrylamide gel electrophoresis indicated that the majority of the cross-linked peptide was in a dimer-size species (Fig. 4). One dimer species that could be easily synthesized was a linear tandem repeat, i.e. (141-155)-(141-155). Therefore, to identify at least one of the active species within the selfconjugated apoE-(141-155) reaction mixture, we synthesized tandem peptide-(141-155). If a linear tandem peptide had biological activity, it would enable us to by-pass the costly and time-consuming task of isolating and characterizing all of the active congeners in the self-conjugated peptide reaction mixture.
As shown in Table I, this synthesis yielded a peptide following preparative HPLC that was >99% pure. More important, this peptide inhibited the thymidine uptake of PHA-stimulated lymphocytes in a concentration-dependent manner (Fig.  5A). Thymidine uptake was inhibited by 50% at a concentration of -10 pg/ml tandem peptide and was thus active over a concentration range of 2-25 p~. This peptide also inhibited theca/interstitial cell androstenedione production in a concentration-dependent manner (Fig. 5B). In this system, androgen production was inhibited 50% at a concentration of -1 pg/ml tandem peptide and was therefore active over a range of 0.2-2.5 P.M. To verify the specificity of the tandem apoE-(141-155)-peptide, we also synthesized both a monomeric and a dimeric tandem repeat of another immunosuppressive peptide, LQNRRGLD, a sequence found in both HTLV-I (amino acids 376-383) and HTLV-I1 (amino acids 372-379) that is active when it is coupled to bovine serum albumin (39-41). The purity and composition of these peptides are shown in Table I. Neither the monomeric nor the dimeric HTLV peptides inhibited the thymidine uptake of mitogen-stimulated lymphocytes (data not shown). Therefore, the inhibitory activity of apoE-(141-155) was not simply an outcome of the synthesis of a tandem repeat of an amino acid sequence that contained multiple basic amino acid residues. Finally, we were able to demonstrate that the observed inhibitory activity of the tandem peptide was not a result of a loss of cellular viability because there was no significant release of LDH activity by the tandem peptide-treated PBM cells, and progesterone production by the tandem peptide-treated ovarian cells was unaffected (data not shown).
Interestingly, at low concentrations, the dimeric tandem peptide caused an enhanced response in both cell systems (Fig. 5, A and B). There was a significant enhancement of thymidine uptake between 2 and 6 pg/ml tandem apoE peptide. Enhancement of mitogen responsiveness has been observed with low concentrations of apoE.* There was also a significant enhancement of androstenedione production at concentrations of <0.2 pg/ml tandem apoE peptide. Enhance-* M. G. Pepe

Biologically Active
Synthetic Peptides of ApoE ment of androstenedione production in the presence of native apoE has not been observed (10). In a separate cell system to corroborate the absence of direct cytotoxicity, the tandem apoE peptide (20 pg/ml) was added to proliferating cultures of rabbit arterial smooth muscle cells and human liver tumor-derived HepG2 cells. After 96 h of culture, there was no effect on either the cell viability (determined by trypan blue dye exclusion) or proliferation (determined by cellular DNA) in either cell type. Therefore, the inhibition observed appeared to be specific for stimulusinduced cellular differentiation as exhibited by the proliferation of mitogen-treated lymphocytes or the androgen production of LH-stimulated theca/interstitial cells.

DISCUSSION
A step toward understanding the mechanism(s) by which apoE is able to inhibit either lymphocyte proliferation or ovarian theca/interstitial cell androgen production is the identification of the functionally important domains of this apoprotein. To this end, we have sought to identify a fragment or synthetic peptide of apoE that would mimic the activity of the native apoprotein. Furthermore, if a functional fragment could be defined, it could be used to identify the nature of the cellular interactions that lead to changes in intracellular signaling.
Using thrombin digestion, we were able to verify that a biologically active portion of the apoE molecule resided in a 22-kDa thrombin fragment that represented amino acid residues 1-192.' To further define this biologically active region of apoE, six synthetic peptides were prepared (Table I) and tested for their capacity to inhibit stimulus-induced lymphocyte proliferation and ovarian androgen production. None of the monomeric peptides had activity up to concentrations of 500 PM. In fact, only a self-conjugate or a dimeric synthetic repeat of sequence-(141-155) was active. Both multimeric preparations inhibited the [3H]thymidine uptake of PHAstimulated lymphocytes in a concentration-dependent manner ( Figs. 1 and 5). This inhibition was not a result of cytotoxicity (Fig. 2) and was irreversible (Table 11), properties typical of native apoE (8,9). When added to the LH-exposed ovarian androgen-producing cells, again both the self-conjugate and the dimeric tandem peptide inhibited androgen production in a concentration-dependent manner (Figs. 1 and 5). More important, androstenedione production was blocked without decreasing progesterone production (Fig. 3), a property that is shared with native apoE (10). When this ongoing progesterone synthesis was considered with the fact that the inhibition was reversible (Table 111), there was little doubt that the cells were healthy. Furthermore, in both systems, the apoE-( 141-155) self-conjugate and the dimeric tandem peptide were active in the absence of either cholesterol or phospholipid, another property that is characteristic of the inhibition of lymphocyte proliferation (9) and ovarian androgen production (10) by intact apoE. Taken together, the data support the idea that either a chemical self-conjugate or a dimeric synthesized peptide can mimic the activity of native apoE.
Although it appeared that the lymphocytes were less sensitive to the inhibitory activity than the ovarian cells, differences in the effective inhibitory concentration may be explained by differences between the two culture systems. The PBM cells were cultured at a density of 200,000 cells/0.25 ml in the presence of 5% fetal bovine serum (total protein = 1.95 mg/ml), whereas the theca/interstitial cells were cultured at 20,000 cells/0.25 ml in serum-free medium (total protein = 0.54 mg/ml). The requirement for 10 times more tandem peptide in the PBM cell cultures compared with the theca/ interstitial cell cultures may reflect differences in cell number, differences in total protein, or differences in peptide stability (e.g. the serum probably contained peptidases).
A number of approaches including the use of monoclonal antibodies (13), natural mutants (I), site-specific mutants generated in vitro (12), and synthetic peptides (42) have indicated that residues 140-160 of apoE are important. This region of apoE has a high affinity heparin-binding site (43). A peptide representing amino acid residues 139-153 can inhibit lipoprotein lipase (44), and a synthetic peptide of apoE residues 141-155 was found to be immunosuppressive when it was conjugated to bovine serum albumin (42). Furthermore, this region of the molecule is involved in the binding of apoE to the LDL receptor (1). It is intriguing that inhibitory sequence-(141-155) is included within this LDL receptorbinding region. However, it is also unlikely that the LDL receptor mediates or is required for the expression of the inhibitory activity. First, it has been shown that LDL receptor-deficient lymphocytes are sensitive to inhibition by apoE (17). This indicates that binding to the LDL receptor is not required for immunosuppressive activity. Second, rat ovarian theca/interstitial cells respond to culture in apoE-depleted LDL with increased rather than decreased androstenedione production (18). Therefore, the LDL receptor is probably not involved in the inhibitory activity of apoE. However, the recent description of an LRP that binds apoE (21) suggests the possibility that apoE could exert its biological activity via this receptor. The LRP receptor binds apoE-enriched very low density lipoproteins and has been proposed to be the putative hepatic chylomicron remnant receptor (21)(22)(23)(24)26). In contradiction to its liver-specific and specialized function, the LRP receptor is distributed widely throughout the body in sites that are not involved in chylomicron remnant removal (23) and may be identical to the a,-macroglobulin receptor (45,46). The domain of apoE that binds LRP has not been mapped. However, because LRP contains redundant reiterations of ligand-binding regions that are similar to those of the LDL receptor (21) and because the LDL receptor-defective apoE2 isoform with cysteine substituted for arginine at position 158 also binds poorly to LRP (26), it seems likely that amino acid residues 141-155 of apoE are involved in binding to the LRP receptor.
An apparent essential property for observing inhibitory activity was a reiteration of amino acid sequence-(141-155) of native apoE. Large amounts of monomer-(141-155) (500 p~) had no effect, whereas multimeric forms of the same sequence were inhibitory at 2-20 p~. In all cases, the inhibitory activity observed with homogeneous tandem peptide-(141-155) was similar to that observed with the heterogeneous self-conjugation reaction mixture. However, the caveat should be stated that the synthetic tandem peptide probably represents only one of several active multimeric peptide species found in the self-conjugated preparation. Further study of additional multimeric peptides, such as linear trimers or other branched multimers, will make possible a more complete modeling of the structural requirements for activity. Nevertheless, because the tandem peptide is active, it can provide a uniform, defined, and unlimited supply of a biologically active apoE mimic.
Perhaps the most perplexing question to come from this study is why sequence-(141-155) must be repeated as a selfconjugate or as a tandem peptide for it to exhibit biological activity. Although specific reasons for observing biological activity with the tandem or self-conjugated peptide and not with the monomeric peptide are unknown, it seems appropri-ate to speculate. It is possible that the inhibition observed with the self-conjugated and synthetic tandem repeat peptides was a result of only stabilization of a requisite amino acid sequence in an active conformation, suggesting that it may not necessarily be the number of repeats that are critical for activity, but the structure of the multimer. However, a number of observations have been made to suggest that the biologically active form of apoE may be multimeric. For example, apoE exists in aqueous solutions as a tetramer (25). Furthermore, it has been reported that isolated apoE can function as a receptor-binding ligand only after its association with phospholipids (1). Although the requirement for phospholipids may be important to the solubility and conformation of a single molecule of apoE, phospholipids also could be required to properly assemble multiple copies of apoE within a confined space on the liposome surface. These observations suggest that the biologically important conformation of apoE may not be monomeric. Furthermore, this requirement could reflect the fact that the interaction of apoE with cells is mediated by a cellular binding site (e.g. LRP) that recognizes or comes in contact with multiple copies of domain-(141-155).
In conclusion, we have identified an active molecular species using synthetic peptides of apoE that mimic the activity of native apoE. The active species is apoE-(141-155), either synthesized in tandem as peptide-(141-155)-(141-155) or as a chemical self-conjugate. The availability of unlimited quantities of a homogenous and biologically active mimic of apoE will enable future investigations of receptor interactions and the mechanisms of apoE inhibition of cell differentiation.