Binding of Apolipoprotein A-I and A-I1 after Recombination with Phospholipid Vesicles to the High Density Lipoprotein Receptor of Luteinized Rat Ovary*

To determine the apolipoprotein specificity of high density lipoprotein (HDL) receptor, apolipoprotein A-I (apo-AI) and apolipoprotein A-I1 (apo-AII) purified from high density lipoprotein3 (HDLs) were reconstituted into dimyristoyl phosphatidylcholine vesicles (DMPC) and their ability to bind to luteinized rat ovarian membranes was examined. Both '261-apo-A- 1-DMPC and '261-apo-A-II*DMPC were shown to bind to ovarian membranes with Kd = 2.87 and 5.70 pg of protein/ml, respectively. The binding of both 1261-ap0-A-I-DMPC and '2SI-apo-A-II*DMPC was inhibited by unlabeled HDL3, apo-A-1-DMPC, apo-A-11-DPMC, apo-C-1-DMPC, apo-C-IIeDMPC, apo-C-III1-DMPC, and apo-C-II12*DMPC, but not by DMPC vesicles, bovine serum albumin-DMPC or low density lipoprotein. Since the binding labeled apo-A-1-DMPC and apo-A-11-DMPC was inhibited by the DMPC complexes of apo-C groups, the direct binding of '261-apo-C- IIIl*DMPC was also demonstrated with Kd = 9.6 pg of protein/ml. In addition, unlabeled apo-A-1-DMPC, and apo-A-11-DMPC, as well as apo-C-DMPC, inhibited "'I-HDL3 binding. '261-apo-A-I, 1261-apo-A-II, and "'1- apo-C-1111 in the absence of DMPC also bind to the membranes. These results suggest that HDL receptor recognizes apolipoprotein AI, AII, and the C group and that the binding specificity of the reconstituted lipo- proteins is conferred by their apolipoprotein moiety rather than the lipid environment. In vivo pretreatment of rats with human chorionic gonadotropin re- sulted in an increase of '261-apo-A-I*DMPC, '251-apo-A-11-DMPC, and '261-apo-C-II11*DMPC binding activ- ities. However, no induction of binding activity was observed when the apolipoprotein was not included in DMPC vesicles. An examination of the equilibrium dissociation constant and binding capacity for '"I-apo- A-1-DMPC and '261-apo-A-II*DMPC after human chorionic gonadotropin treatment revealed that the in- crease in binding activity was due to an increase in the number of binding sites rather than a change in the binding affinity. These results further support our contention that apo-A-I, apo-A-11, and the apo-C group bind to HDL receptor. In conclusion, the HDL receptor of luteinized rat ovary recognizes apolipoproteins A-I, A-11, and the C group but not low density lipoprotein, and the binding is induced by human chorionic gonad- otropin in vivo.

To determine the apolipoprotein specificity of high density lipoprotein (HDL) receptor, apolipoprotein A-I (apo-AI) and apolipoprotein A-I1 (apo-AII) purified from high density lipoprotein3 (HDLs) were reconstituted into dimyristoyl phosphatidylcholine vesicles (DMPC) and their ability to bind to luteinized rat ovarian membranes was examined. Both '261-apo-A-1-DMPC and '261-apo-A-II*DMPC were shown to bind to ovarian membranes with Kd = 2.87 and 5.70 pg of protein/ml, respectively. The binding of both 1261-ap0-A-I-DMPC and '2SI-apo-A-II*DMPC was inhibited by unlabeled HDL3, apo-A-1-DMPC, apo-A-11-DPMC, apo-C-1-DMPC, apo-C-IIeDMPC, apo-C-III1-DMPC, and apo-C-II12*DMPC, but not by DMPC vesicles, bovine serum albumin-DMPC or low density lipoprotein. Since the binding labeled apo-A-1-DMPC and apo-A-11-DMPC was inhibited by the DMPC complexes of apo-C groups, the direct binding of '261-apo-C-IIIl*DMPC was also demonstrated with Kd = 9.6 pg of protein/ml. In addition, unlabeled apo-A-1-DMPC, and apo-A-11-DMPC, as well as apo-C-DMPC, inhibited "'I-HDL3 binding. '261-apo-A-I, 1261-apo-A-II, and "' 1apo-C-1111 in the absence of DMPC also bind to the membranes. These results suggest that HDL receptor recognizes apolipoprotein AI, AII, and the C group and that the binding specificity of the reconstituted lipoproteins is conferred by their apolipoprotein moiety rather than the lipid environment. In vivo pretreatment of rats with human chorionic gonadotropin resulted in an increase of '261-apo-A-I*DMPC, '251-apo-A-11-DMPC, and '261-apo-C-II11*DMPC binding activities. However, no induction of binding activity was observed when the apolipoprotein was not included in DMPC vesicles. An examination of the equilibrium dissociation constant and binding capacity for '"I-apo-A-1-DMPC and '261-apo-A-II*DMPC after human chorionic gonadotropin treatment revealed that the increase in binding activity was due to an increase in the number of binding sites rather than a change in the binding affinity. These results further support our contention that apo-A-I, apo-A-11, and the apo-C group bind to HDL receptor. In conclusion, the HDL receptor of luteinized rat ovary recognizes apolipoproteins A-I, A-11, and the C group but not low density lipoprotein, and the binding is induced by human chorionic gonadotropin in vivo.
Plasma lipoproteins are known to bind to receptors with high affinity on the plasma membranes in several different cell types (for review see Refs. 1 and 2). The two classes of lipoproteins that have been demonstrated to bind to rat luteal cell surface receptors are the low density lipoprotein (LDL') and high density lipoprotein (HDLJ (3,4). These lipoproteins have been shown to enhance steroidogenesis in the enzymedispersed luteal cells (5)(6)(7)(8). The stimulation of steroidogenesis by lipoproteins involves the binding of lipoproteins to specific cell surface receptors followed by endocytosis and lysosomal hydrolysis to release free cholesterol which is then utilized as a substrate for progesterone synthesis (4,9).
Several studies have indicated that the binding specificity of lipoprotein is conferred by its protein component (10, 11). Recent investigations have shown that LDL receptor can recognize both apo-B and apo-E (12-14). However, which apolipoprotein is recognized by HDL receptor is not well defined although it has been suggested that apo-A-I may be the HDL apolipoprotein recognized by the HDL receptor on rat testicular membranes (15). This study was, therefore, undertaken in an effort to examine the apolipoprotein binding specificity of HDL receptor. Since HDL3 contains two majqr apolipoproteins, apo-A-I and apo-A-11, experiments were carried out to purify apo-A-I and apo-A-11 from human HDL3, followed by an examination of their ability to bind to rat ovarian membranes. It has also been reported that both apo-A-I and apo-A-I1 can insert into dimyristoyl phosphatidylcholine vesicles (16)(17)(18)(19). We have followed these procedures in reconstituting apo-A-I and apo-A-I1 into phospholipid vesicles to examine their ability to bind to ovarian membranes and to investigate the binding specificity of the resulting complexes. Our results suggest that HDL receptor recognizes apolipoprotein A-I, A-11, and the C group but not LDL, and that the lipid environment may not be required for binding.

Methods
Preparation of Serum Lipoprotein and Purification of Apo-A-I and Apo-A-ZI from Human HDL,-Human LDL (d, 1.019-1.063 g/ml) and HDL, (cf, 1.125-1.210 g/ml) were fractionated by sequential flotation in a Beckman preparative ultracentrifuge, fitted with a 60 Ti rotor, using KBr and NaCl for density adjustment as described by Have1 et al. (20). Human HDL3 was delipidated by solvent extraction (21) and the apolipoproteins were purified by Sephadex G-200 column chromatography (22,23). Prior to use, apo-A-I and apo-A-I1 were denatured in 8 M urea and refolded by slow removal of the denaturing agent (18).
Iodination of Apolipoproteim-Apo-A-I, apo-A-11, and apo-C-III,, were iodinated with "' I using a modification of the iodine monochloride procedure (24). Free I w I ion was removed from the "'I-apo- Briefly, 1 ml of buffer, containing 0.15 M NaC1, 10 mM Tris-HC1, 1 mM EDTA, with 1 mM NaN,, pH 7.6, was added to 10 mg of DMPC that had been dissolved in 1 ml of benzene and dried by lyophilization. The solution was sonicated for 20 min at 30 "C using a Sonifier Cell Disruptor, Model W140 (Heat Systems-Ultrasonics, Inc.) equipped with a microtip at a setting of 8 (30 watts). The sonicated preparation of DMPC was then added to the 1251-apo-A-I, lwI-apo-A-II, or lwI-apo-C-III1 (specific activity for 12SI-apo-A-I, 'Z51-apo-A-II, and 1251-apo-C-II11 was adjusted to 50, 40, and 20 cpm/ng, respectively, by adding an appropriate amount of unlabeled apo-A-I, apo-A-11, or apo-C-111,) in the original buffer at a lipid-to-protein ratio of 4 1 (w/w). The mixture was vortexed for 10 s and left overnight at room temperature. The proteinlipid complexes were isolated by density gradient ultracentrifugation performed in a SW 50.1 swinging bucket rotor at 45,000 rpm and 25 "C for 60 h as described by Roth (26). Potassium bromide was removed by dialysis prior to the use of the protein-lipid complexes for binding studies.
Preparation of Unlabeled Apo-A-I.DMPC, Apo-A-IZ. DMPC, Apo-DMPC Vesicles-Unlabeled DMPC vesicles containing the above apoproteins were prepared by the procedure of Roth et at. (26) followed by the isolation of individual lipid-protein complex by sucrose density gradient centrifugation as described in the previous section dealing with the preparation of 1251-apoprotein complexes. Electrophoresis and Autoradiography-The sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE) described by Laemmli (27) was used to examine the purity of apo-A-I and apo-A-11. Electrophoresis was typically carried out at 3 mA/sample for approximately 5 to 6 h on 1.5 mm-thick gels containing 15% acrylamide with 4% stacking gel. The gels were stained overnight in a staining solution containing 0.05% Coomassie Blue R-250, 25% isopropanol, and 10% acetic acid and were destained in 10% acetic acid. The gel was dried prior to autoradiography. Autoradiographic exposure of the gels was performed at -70 "C for 20 h with Kodak XAR-5 film using the Dupont Cronex Lightning Plus enhancing screen which has been shown to improve the sensitivity and speed of autoradiography (28).
Animals and Hormonal Treatment-Twenty-one-day-old female Sprague-Dawley rats (Spartan Research Inc., Hazlett, MI) were employed in the present studies. Highly luteinized ovaries were obtained following a regimen described by Parlow (29). Rats were injected subcutaneously with 50 IU of PMSG followed 56 h later with 25 IU of hCG. Day 0 was taken as the day of hCG injection and animals were usually killed on day 7.
Isolation of Partially Purified Plasma Membranes-Fresh, highly luteinized rat ovaries were weighed and homogenized at 4 "C in an all-glass homogenizer fitted with a Teflon pestle, and plasma mem-C -I . DMPC, A~o-C-ZI. DMPC, Apo-C-IIII. DMPC, and A~o-C-III~. branes were prepared according to the procedure described by Gospodarowicz (30). The relative purification of membrane preparations was ascertained by assaying 5'-nucleotidase activity. The extent of mitochondrial, lysosomal, and microsomal contamination was checked by assaying specific marker enzymes (31). Relative to 5'nucleotidase activity present in the original rat ovarian homogenate, enrichment of this marker enzyme in isolated plasma membranes averaged 10-fold with negligible mitochondrial, lysosomal, and microsomal contamination. Membranes were prepared on the same day as the binding assay.
Binding Assays-Aliquots of membrane preparations were incubated in disposable glass tubes (12 X 75 mm) precoated overnight with 10% BSA. The incubations usually were carried out in a total volume of 0.25 ml of 25 mM Tris-HC1 buffer, pH 7.4, containing "' Iapo-A-I.DMPC, 1251-apo-A-II.DMPC, or 12SI-apo-C-II11.DMPC (10 pg of protein/ml; specific activity for '251-apo-A-I, 1251-apo-A-II, and '251-apo-C-II11, was 50, 40, and 20 cpm/ng, respectively, either alone (total binding) or in the presence of 3 mg of unlabeled HDLa/ml (nonspecific binding). All the incubations were performed at 37 "C, for 60 min unless otherwise specified. Following incubation, the receptor-bound '*'I-apo-A-I .DMPC, 1251-apo-A-II. DMPC, or "' Iapo-C-III1. DMPC was separated from free '251-apo-A-I. DMPC, 9apo-A-11. DMPC, or 1z51-apo-C-II11 .DMPC by the addition of 1 ml of precooled buffer (25 mM Tris, pH 7.4) to each tube followed by centrifugation at 3000 X g for 20 min at 4 "C in an IEC (DPR-6000) centrifuge, fitted with a 949 rotor. The supernatants were removed by suction. The pellets were washed once and then counted in an automatic y counter (Searle Analytic, Model 1195). The specific binding was determined by subtracting the nonspecific binding from the total binding. All assays were carried out in triplicate. The slope of the line obtained from Scatchard plot is equal to -1/& (32).
Protein Assay-Protein was determined by the colorimetric procedure of Lowry et al. (33).

RESULTS
The identity and purity of apo-A-I and apo-A-11, isolated from human HDL3 by chromatography apo-HDL, on a Sephadex G-200 column (100 X 2.5 cm) equilibrated with buffer, containing 10 m M Tris-HC1, 1 mM EDTA, and 8 M urea, pH 8.6, was examined by SDS-PAGE (27). Thirty micrograms of apo-A-I and apo-A-I1 gave single bands with molecular weights of 28,000 and 17,000, respectively, when analyzed by electrophoresis on a 0.1% SDS, 15% polyacrylamide gel under nonreducing conditions (Fig. L4). The molecular weights of apo-A-I and apo-A-I1 obtained in this study are consistent with the values reported by other investigators (34-36). Bovine serum albumin (BSA), purified apo-A-I, and apo-A-I1 were then iodinated with radioactive lz5I. The iodinated products were then examined by electrophoresis on a 0.1% SDS, 15% polyacrylamide gel under nonreducing conditions followed by autoradiography. The 1251-apo-A-I and 1251-apo-A-II appeared as single bands in the autoradiogram, and the lZ5I-BSA preparation had slight impurity as shown in Fig. 1B. These results further show that apo-A-I and apo-A-I1 are homogeneous preparations.
Since HDL3 contains two major apolipoproteins, apo-A-I and apo-A-I1 and apo-C group (apo-C-I, -C-11, -C-1111, and -C-IIIz) as minor constituents, it was of interest to examine which apolipoprotein in HDL3 recognizes the HDL receptor.
The result of an examination of the ability of BSA, apo-A-I, apo-A-11, and apo-C-111, to bind to rat ovarian membranes is shown in Table I. The 'T-BSA and lZ5I-BSA. DMPC were used as controls, since BSA is not expected to specifically bind to the HDL receptor. As expected, 'T-BSA and lz5I-BSA . DMPC failed to bind to plasma membranes. However, '2'I-apo-A-I, '251-apo-A-II, and 1251-apo-C-II11, irrespective of whether the radiolabeled ligand was present in a reconstituted form with DMPC vesicle or solely as a protein in solution, bound significantly to membrane preparations. In addition, the binding of '251-apolipoprotein and '251-apolipoprotein. To examine whether the binding sites of apo-A-I-DMPC, apo-A-II . DMPC, and apo-C-III, . DMPC are the putative HDL receptors, the binding of these labeled apolipoprotein. DMPC complexes to plasma membranes as a function of ligand concentrations, membrane protein concentration, incubation time, and binding specificity was examined. If apo-A-I, apo-A-I. DMPC, apo-A-11, apo-A-I1 .DMPC, apo-C-1111, and apo-C-1111 .DMPC bind to the putative receptor, the binding of these ligands to a fixed amount of plasma membranes should reach saturation with increasing concentrations of ligand. For this study, the lipid-protein complexes, 1251-ap0-A-I .DMPC, 1251-apo-A-II. DMPC, and 1251-apo-C-II11. DMPC, were used instead of lZ5I-apo-A-I, '251-apo-A-II, and '251-apo-C-II11, to compare the binding kinetics obtained with HDL3 since native HDL3 is also a protein-lipid complex. The specific activity of 1251-apo-A-I. DMPC, '251-apo-A-II. DMPC, and '251-apo-C-IIIl.DMPC was adjusted to 50, 40, and 20 cpm/ng of protein, respectively, by the addition of unlabeled apo-A-I. DMPC, apo-A-11. DMPC, or apo-C-IIIl. DMPC complex. When aliquots of plasma membranes (200 pg of membrane proteins) were incubated with various concentrations of 1251-apo-A-I .DMPC, the specific binding of 1251-apo-A-I. DMPC increased with increasing concentrations of the ligand up to 5.0 pg of protein/ml and the binding process was saturated at a concentration of 7.5-12.5 pg of protein/ml ( Fig.  2A). The nonspecific binding was also increased linearly as a function of labeled ligand concentration. Scatchard analysis of '251-apo-A-I. DMPC binding showed a linear plot indicating that the membranes contain one class of binding sites for 1251apo-A-I.DMPC ( Fig. 2A, inset). The Kd for lZ5I-apo-A-I. DMPC binding to rat ovarian membranes, as calculated from the Scatchard plot, was 2.87 pg of protein/ml. Similar results were also obtained for 1251-apo-A-II.DMPC binding. When aliquots of plasma membranes (135 pg of membrane protein) were incubated with different concentrations of '251-apo-A-II. DMPC, the amount of membrane bound ligand was increased with increasing concentrations of the labeled ligand up to 8.0 pg of protein/ml (Fig. 2B). The binding reaction was saturated at a concentration of 12.5-15 pg of protein/ml and the nonspecific binding increased linearly as a function of the labeled ligand concentration. Scatchard analysis of 1251-apo-A-II. DMPC binding showed a linear plot indicating that the membranes contain one class of binding sites for 1251-apo-A-II. DMPC with Kd = 5.70 pg of protein/ml (Fig. 2B, inset). The binding of 1251-apo-C-II11 DMPC to the plasma membranes (150 pg of membrane protein) with increasing concentrations of the labeled ligand concentration is shown in Fig. 2C. The binding was increased with increasing concentrations of the labeled ligand up to 25 pg of protein/ml and reached saturation at 40-90 pg/ml. Nonspecific binding increased as a function of ligand concentration. The Scatchard analysis indicated one class of binding sites with K d = 9.6 pg/ml (Fig. 2C, inset).
To characterize the binding properties of the major apolipoprotein of HDL3, the binding of 1251-apo-A-I.DMPC and '251-apo-A-II-DMPC, as a function of membrane protein concentration and incubation time, was examined. Fig. 3, A  The binding of both labeled apolipoprotein. DMPC complexes was dependent on incubation time (Figs. 4, A and B). The binding reached equilibrium after 45 min and remained constant up to 120 min in both binding assays. The small amount of binding observed at time 0 is probably due to residual binding occurring during the 20 min centrifugation of the reaction mixture employed to separate the bound ligand from the free fraction.

Time (Minutes)
lZ5I-HDL3 binding. Since HDL3, apo-A-I. DMPC, apo-A-11. DMPC, and apo-C-III1.DMPC can mutually inhibit each other's binding, it appears that these ligands may bind to the HDL receptor. In addition, the binding of lZ5I-apo-A-I.

DMPC,'251-apo-A-II.DMPC,and'251-apo-C-II11~DMPC was
inhibited by unlabeled reconstituted apolipoprotein. DMPC complexes but not by DMPC vesicles, suggesting that the binding ability of the reconstituted lipoproteins is conferred by their apolipoprotein moiety rather than the lipid environment. Taken together, these results suggest that HDL receptor recognizes apolipoprotein A-I, -A-11, and the C group, but not LDL.
Since we have previously reported that in vivo pretreatment of rats with hCG results in induction of HDL receptors in the ovary (3), this characteristic was used as an additional criterion to examine whether apo-A-1. DMPC, apo-A-11. DMPC, and apo-C-III,.DMPC bound to the HDL receptor. If the binding site for apo-A-1-DMPC, apo-A-I1 .DMPC, and apo-C-1111 .DMPC is the HDL receptor, then the number of binding sites for these ligands should also be increased following hCG treatment. The binding of 1251-apo-A-I. DMPC and 1251apo-A-II.DMPC to the control and the hCG-treated groups as a function of the concentrations of ligand is shown in Fig.  5, A and B. As expected, the binding of '251-apo-A-I.DMPC

Unlabeled
Binding of Binding of substance"

Binding of '%I-apo-A-I.DMPC '%I-apo-A-11. DMPC 'mI-apo-C-IIIl. DMPC
and 12sI-apo-A-II .DMPC to the ovarian membranes was significantly higher in the hCG-treated group than in the control group.
T o examine whether the increase in '2sI-apo-A-I.DMPC and '2sI-apo-A-II. DMPC binding activity to hCG-treated ovarian membranes was caused by an increase in the number of binding sites or by a change in the binding affinity, the data obtained from Fig. 5, A and B, were transformed to a Scatchard plot (Fig. 6). Two linear parallel lines were obtained as shown in Fig. 6A with K d of '251-apo-A-I.DMPC = 2.87 pg of protein/ml for the control group and 2.83 pg of protein/ml for the hCG-treated group. The binding capacity for '2sI-apo-A-I. DMPC was 4.36 and 7.36 pg of protein/mg of membrane protein in control and hCG-treated group, respectively. Similar results were also obtained for '251-apo-A-II. DMPC binding. Two linear parallel lines were obtained in Fig. 6B with & = 5.70 and 5.86 pg of protein/ml for control and hCGtreated group, respectively. The binding capacity for '2sI-apo-A-I1 . DMPC was 16.4 pg of proteinlmg of membrane protein in control group and 23.6 pg of protein/mg of membrane protein in hCG-treated group. The data derived from Fig. 6,  A and B , are summarized in the inset. These results further support the finding that the binding site for apo-A-1. DMPC and apo-A-11. DMPC is the HDL receptor.
Experiments were then carried out to examine the specificity of the hCG effect in inducing apo-A-I, apo-A-11, and apo-C-1111 binding activities. The binding of '2sI-apo-A-I. DMPC, 12sI-apo-A-II .DMPC. '251-apo-C-II11 .DMPC, lZ5I-BSA. DMPC, and [I4C]DMPC vesicles to hCG-treated and control rat ovarian membranes is shown in Fig. 7. Administration of the maximum stimulatory dose of 25 IU of hCG did not result in an increase in the binding activities of lZ5I-BSA.DMPC and [14C]DMPC vesicles. However, the binding of "'I-apo-A-I . DMPC, '2sI-apo-A-II .DMPC, and 1251-apo-C-IIIl .DMPC in the hCG-treated group did result in an increase, suggesting that the hCG effect to induce lipoprotein binding activity is apolipoprotein-specific. DISCUSSION In the present study, we have used apo-A-I and apo-A-I1 isolated from human HDL3, due to the fact that 1) the stimulation of steroidogenic response was previously performed using human HDL (5-7), 2) the characterization of HDL receptor was carried out by using human HDL, (3), 3) rat and human HDL are similar in apolipoprotein content with apo-A-I being the major apolipoprotein component (37, 38) while rat HDL possesses more of apo-A-IV and apo-E (39, 40) than that seen in human HDL, and 4) unlabeled human HDL inhibits 1251-rat HDL binding as effectively as unlabeled rat HDL (4).
Previous studies have shown that rat corpus luteum possesses discrete receptors for LDL and HDLB (3). This conclusion has been further supported by the observation in this study that unlabeled LDL failed to inhibit the binding of both '2sI-apo-A-I .DMPC and 1251-apo-A-II. DMPC. Furthermore, apo-A-I and apo-A-I1 bound significantly to ovarian membranes as proteins in solution, unlike apo-E which possesses binding activity only in a form of protein-lipid complex (41). This would suggest that the lipid environment may not be required for apo-A-I and apo-A-I1 binding. Since the binding assays were performed at 37 "C, it could be argued that at this temperature binding assays of proteoliposomes might be artificially affected by phospholipid or protein exchange with membranes or lipoproteins. However, both apo-A-I and apo-A-11, when incubated at 4 "C to minimize such exchange reaction, also bound to ovarian membranes (data not shown). This observation may rule out the possibility that the binding activity of apo-A-I and apo-A-I1 was due to lipid-protein exchange.
From the binding specificity study it is apparent that the HDL receptor has broad specificity since apolipoprotein A-I, A-11, and the C group inhibited "'I-HDL3 binding. It has been shown that apolipoprotein A-I, A-11, and the C group share no primary sequences in common (42). However, the secondary structure of these apolipoproteins shows typical amphipathic helical structure (42). Thus, this amphipathic helical structure may represent the common binding domain for the HDL receptor which in turn confers broad binding specificity of Rat Ovarian HDL Receptor binding assay solution (total binding). Incubation was carried out at 37 "C for 60 min. For estimation of nonspecific binding, unlabeled HDL, was added at a final concentration of 3 mg/ml. The specific binding was calculated by substracting the nonspecific binding from the total binding. The data points of the control are the same as those in Fig. 2 . 0 , control group; 0, hCG-treated group.
(43). Since apolipoprotein alone bound to ovarian membranes as effectively as apolipoprotein. DMPC complexes, this may suggest that the binding domain of these apolipoproteins resides in the hydrophilic portion rather than in the lipidbinding determinants. Since apolipoproteins have been shown to increase amphipathic helical formation upon recombination with DMPC (42), a question may be then raised as to why apo-A-I, apo-A-11, and apo-C-111, effectively bind to the HDL receptor when supplied solely as a protein in solution. This may be due to the fact that apo-A-I, apo-A-11, and apo-C-111, were already in an amphipathic helical structure since they were denatured in 8 M urea and refolded by slow removal of the urea before being used for the binding assay.
lZ5I-apo-A-I. DMPC, '251-apo-A-II.DMPC, and '251-apo-C-  T-apo-A-I . DMPC, '251-apo-A-II. DMPC, and apo-C-III, . DMPC, respectively) showed differences. If labeled apolipoprotein .DMPC complexes bind to the same receptor, then the binding capacity for these ligands should be comparable on a stoichiometric basis. Thus, the difference in binding capacity for lZ5I-apo-A-I. DMPC, 1251-apo-A-II. DMPC, and '251-apo-C-III, .DMPC observed in this study may be attributed to the following reasons: 1) the average number of apolipoprotein molecules in each DMPC vesicle may be different or 2) the stoichiometry of the ligand-receptor interaction may be different for different apolipoprotein. DMPC vesicles. The former possibility is supported by the finding that the average number of apo-A-I molecules in the lZ5I-ap0-A-1. DMPC complex is less than that of apo-A-I1 in lZ5I-ap0-A-11-DMPC complexes (data not shown). In view of the interaction of apo-A-I and apo-A-I1 with lipid vesicles, it has been shown that lipid vesicles have higher affinity for apo-A-I1 than apo-A-I (44-46). This may result in an increase in the average number of apo-A-I1 molecules in one apo-A-II . DMPC complex. However, in spite of the evidence in support of the first possibility, we cannot completely rule out the latter possibility at the present time.
In addition to the finding that '251-apo-A-I-DMPC, lZ5Iapo-A-11. DMPC, and 1251-apo-C-IIIl .DMPC bind to the same receptor, it was also shown that the binding site for these ligands probably is the HDL receptor, since their binding was inhibited by unlabeled HDL3. This conclusion was further substantiated by the fact that in vivo pretreatment of rats with hCG resulted in an induction of binding sites of 'z51-ap0-

A-I~DMPC,'25I-apo-A-II~DMPC,and'25I-apo-C-III~~DMPC
complexes. Since HDL3 contains two major (apo-A-I and apo-A-11) and minor (apo-C group) apolipoprotein constituents, and assuming that apo-A-I, apo-A-11, and apo-C group bind to the HDL receptor, then the apparent Kd of HDL3 to its receptor should be the combination of the Kd values of apo-A-I. DMPC, apo-A-II . DMPC, apo-C . DMPC. However, in our previous study it was shown that HDL, binds to ovarian membranes with a single Kd. This may be due to the fact that the apparent Kd of HDL, cannot be resolved into separate components by Scatchard analysis, since the difference of the Kd values for apo-A-I. DMPC, apo-A-11. DMPC, and apo-C-IIIl. DMPC is small.
Although the Kd values for '251-apo-A-I. DMPC, lZ5I-apo-A-11. DMPC, and 1251-apo-C-II11. DMPC obtained in this study were slightly lower than the apparent Kd of native HDL3, these values were comparable (Kd for apo-A-I.DMPC, apo-A-II.DMPC, apo-C-1111 .DMPC, and native HDL3 are 2.87, 5.70, 9.6, and 17.8 pg of proteinlml, respectively). A likely explanation for this difference, however, is that apo-A-I in the reconstituted apo-A-I.DMPC complex may be more exposed to the medium than in the native HDL3. Thus, apo-A-I .DMPC may strongly interact with HDL receptor with higher affinity than native HDL3.
Taken together, the present data support the contention that apo-A-I, apo-A-11, and apo-C group bind to the HDL receptor, based on their binding ability, binding specificity, and the fact that in uiuo hCG treatment resulted in an increase  cles binding to rat ovarian membranes. Rats were killed 12 h after subcutaneous injection with saline or 25 IU of hCG, and plasma membranes fractions were prepared from the isolated ovaries. Aliquots (100 pg of membrane protein) of membranes from both control and hCG-treated groups were incubated with 9-apo-A-1. DMPC, (10.0 pg of protein/ml, 50 cpm/ng), 1251-apo-A-II.DMPC, (10.4 pg, of protein/ml, 40 cpm/ng), 1251-apo-C-II11.DMPC, (10.0 pg of protem/ ml, 20 cpm/ng), "'1-BSA.DMPC (10.2 pg of protein/ml, 50 cpm/ng), or ["CIDMPC vesicles (9.6 pg of DMPC/ml, 20 cpm/ng) in a total volume of 0.3 ml of binding assay solution at 37 'C for 60 min either alone (total binding) or together with unlabeled HDL, a t a concentration of 3 mg/ml (nonspecific binding). The specific binding was calculated by subtracting the nonspecific binding from the total binding. The binding activity ratio for 1Z61-apo-A-I.DMPC, 1251-ap0-A-I1 .DMPC, and '251-apo-C-II11.DMPC was obtained by dividing the specific binding in the hCG-treated group by that in the control group. Since the binding of '*'I-BSA.DMPC and ['TIDMPC to the membranes showed no significant difference between total binding and nonspecific binding, the binding activity ratio of these two ligands was obtained by dividing the total binding in the hCG-treated group by that in the control group.
of their binding sites in the ovary. We conclude from this study that HDL receptor recognizes apolipoprotein A-I, A-11, and the C group, that the apolipoprotein binding activity can be induced by hCG treatment, and that lipid environment probably is not required for apo-A-I and apo-A-I1 binding to the ovarian membranes.