Mechanism of Scavenger Receptor Class B Type I-mediated Selective Uptake of Cholesteryl Esters from High Density Lipoprotein to Adrenal Cells*

Despite extensive studies and characterizations of the high density lipoprotein-cholesteryl ester (HDL-CE)-selective uptake pathway, the mechanisms by which the hydrophobic CE molecules are transferred from the HDL particle to the plasma membrane have remained elusive, until the discovery that scavenger receptor BI (SR-BI) plays an important role. To elucidate the molecular mechanism, we examined the quantitative relationships between the binding of HDL and the selective uptake of its CE in the murine adrenal Y1-BS1 cell line. A comparison of concentration dependences shows that half-maximal high affinity cell association of HDL occurs at 8.7 ± 4.7 μg/ml and the K m of HDL-CE-selective uptake is 4.5 ± 1.5 μg/ml. These values are similar, and there is a very high correlation between these two processes (r 2 = 0.98), suggesting that they are linked. An examination of lipid uptake from reconstituted HDL particles of defined composition and size shows that there is a non-stoichiometric uptake of HDL lipid components, with CE being preferred over the major HDL phospholipids, phosphatidylcholine and sphingomyelin. Comparison of the rates of selective uptake of different classes of phospholipid in this system gives the ranking: phosphatidylserine > phosphatidylcholine ≈ phosphatidylinositol > sphingomyelin. The rate of CE-selective uptake from donor particles is proportional to the amount of CE initially present in the particles, suggesting a mechanism in which CE moves down its concentration gradient from HDL particles docked on SR-BI into the cell plasma membrane. The activation energy for CE uptake from either HDL3 or reconstituted HDL is about 9 kcal/mol, indicating that HDL-CE uptake occurs via a non-aqueous pathway. HDL binding to SR-BI allows access of CE molecules to a “channel” formed by the receptor from which water is excluded and along which HDL-CE molecules move down their concentration gradient into the cell plasma membrane.

In humans, peripheral cells receive exogenous cholesterol from low density lipoprotein (LDL) 1 particles, which are incorporated into the cells by receptor-mediated endocytosis (1). The cholesteryl ester (CE) in the LDL particles is hydrolyzed to free cholesterol in lysosomes. Since non-steroidogenic cells cannot catabolize cholesterol, homeostasis is maintained by efflux of free cholesterol to extracellular high density lipoprotein (HDL) particles (2). The concentration gradient required for net efflux of cell cholesterol mass is maintained by the action of lecithincholesterol acyltransferase, which converts free cholesterol in HDL to CE. This CE is then returned to the liver by the reverse cholesterol transport pathway (3). In addition to this flux of HDL-CE to the liver, where it can be converted into bile acids, HDL-CE is used by steroidogenic cells such as adrenal, ovary, and testis for production of steroid hormones (4). In rodents, HDL cholesterol is taken up by adrenal cells in vitro (5) and in vivo (6,7) in preference to LDL cholesterol. The HDL-dependent increase in corticosterone production in cultured rat adrenocortical cells greatly exceeds what can be accounted for by HDL particle uptake and degradation by the cell, suggesting that HDL-CE is taken up by the cell without concomitant uptake of the HDL particle (8). Subsequent studies in vivo and in cell culture have shown directly that HDL-CE is taken up by adrenal cells without HDL particle uptake and degradation (9 -12).
This HDL-CE-selective uptake pathway is distinct from the LDL receptor-mediated pathway wherein LDL binds to the receptor, is internalized in coated pits, and is directed to lysosomes for whole particle degradation (13). In contrast, HDL-CE is preferentially taken up by a variety of tissues and cell types by a non-endocytic mechanism without either degradation of apolipoproteins or whole particle uptake (10). CE from HDL is transferred to cell plasma membranes by a passive process that is dependent on the cholesterol content of the membrane (14,15). Thereafter, CE molecules are irreversibly transferred into the cell interior (16) and are hydrolyzed by a non-lysosomal mechanism (17,18). When reconstituted HDL particles containing different apolipoproteins were studied in cultured adrenal cells, the HDL-CE-selective uptake process showed only a modest preference for apolipoprotein (apo) A-I (19) although apoA-I appears to be essential in vivo (20). Although the HDL-* This work was supported in part by Grants HL22633, HL07443, HL32868, and HL58012 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ Supported during portions of this work by a research fellowship from the Heart and Stroke Foundation of British Columbia and Yukon (Canada) and a postdoctoral fellowship from the American Heart Association-Southeastern Pennsylvania Affiliate. Present address: Dept. of Pharmacology, Esperion Therapeutics, Inc., Ann Arbor, MI 48108.
ʈ To whom correspondence should be addressed: Dept. of Biochemistry, MCP Hahnemann University, 2900 Queen Ln., Philadelphia, PA 19129. Tel.: 215-991-8280; Fax: 215-843-8849; E-mail: michael.phillips @drexel.edu. CE-selective uptake pathway has been extensively studied and characterized, understanding of the mechanisms by which the hydrophobic CE molecules are transferred from the HDL core to the plasma membrane has remained elusive.
The recent demonstration that scavenger receptor BI (SR-BI) mediates the selective uptake of HDL-CE in transfected cells provides an important link between the selective uptake pathway and a specific cell surface receptor (21). These studies showed that expression of SR-BI in Chinese hamster ovary (CHO) cells resulted in HDL binding and the selective uptake of a fluorescent lipid probe and CE (21). This report provided strong evidence that a cell surface receptor may mediate the uptake of HDL-CE into cells; similar results have been reported with the human homologue, CLA-1 (22). In both mice (21) and rats (23), SR-BI expression occurs predominantly in tissues exhibiting high levels of HDL-CE-selective uptake, namely steroidogenic cells and the liver (9,24). SR-BI expression is induced by adrenocorticotropic hormone (ACTH) in mouse adrenal gland in vivo (25) and in Y1-BS1 (25) and Y1 (26) adrenocortical cells in culture. Direct evidence for SR-BI function in steroidogenic cells is provided by recent results showing that antibody to the extracellular domain of SR-BI blocks HDL-CE-selective uptake and the delivery of HDL cholesterol to the steroidogenic pathway in cultured adrenocortical cells (27). In addition, inactivation of the SR-BI gene by a targeted mutation in mice results in a dramatic reduction in neutral lipid stores in adrenal glands (28). These mice also show increased plasma levels of large CE-rich HDL particles. Furthermore, adenovirus-mediated overexpression of SR-BI in the liver results in a marked reduction of HDL cholesterol in plasma and enhanced biliary secretion of cholesterol (29). Taken together, these observations provide strong evidence that SR-BI plays an important role in mediating HDL-CEselective uptake in steroidogenic cells and the liver.
Our goal was to define the molecular mechanisms of SR-BImediated HDL-CE-selective uptake. In order to achieve this, we examined the relationship between the binding of HDL, and the selective uptake of its CE in the Y1-BS1 cell; the quantitative data show a tight correlation among these parameters. The stoichiometry and temperature dependence of CE uptake from HDL and recombinant HDL particles of defined composition support a model in which HDL binding to SR-BI results in the formation of a non-aqueous pathway through which HDL-CE molecules move down their concentration gradient into the adrenocortical cell membrane. By obtaining novel quantitative data with the Y1-BS1 cell line with which Pittman and colleagues (9 -11, 14 -16, 18, 19, 24) made extensive kinetic measurements to establish the phenomenon of HDL-CE-selective uptake, it has been possible to build on their work and develop the first model of the mechanism by which SR-BI facilitates HDL-CE-selective uptake in this physiologically relevant system.

Preparation of 125 I-Apolipoprotein-and [ 3 H]Cholesteryl Hexadecyl
Ether-labeled HDL-HDL 3 and lipoprotein-deficient plasma (d Ͼ 1.21 g/ml fraction) were isolated from normolipidemic subjects by sequential ultracentrifugation as described previously (30). These samples were dialyzed exhaustively against 150 mM NaCl, and the lipoprotein-deficient plasma containing cholesteryl ester transfer protein activity was heat-inactivated for 10 min at 60°C to eliminate lecithin-cholesterol acyltransferase activity (31). Typically, 5 mg of HDL 3 (31). Following the equilibration of [ 3 H]CHE label into the neutral lipid core of HDL, HDL was re-isolated by adjusting the density of the mixture and col-lecting the d Ͻ 1.21 g/ml fraction following two sequential isolations after 24-h centrifugation cycles. Using this procedure, essentially all of the albumin present in the lipoprotein-deficient plasma fraction was eliminated from the re-isolated HDL. Radioiodination of the protein moiety of [ 3 H]CHE-labeled HDL was performed by the iodine monochloride method as described by Goldstein et al. (13). Briefly, 3 mg of HDL protein was placed in a glass tube, the pH adjusted with glycine buffer (pH 10), then approximately 1 mCi of carrier-free Na 125 I (15)(16)(17)(18) Ci/mg, NEN Life Science Products) and 10 l of a 2.64 mM iodine monochloride solution was added to initiate the labeling process. After approximately 30 s, the mixture was passed through a 10-ml Sephadex G-25 (Amersham Pharmacia Biotech PD-10) column equilibrated with 150 mM NaCl, 0.25 mM EDTA, pH 7.4 (saline buffer), to separate radiolabeled material from free 125 I. Thereafter, fractions corresponding to the radioiodinated HDL were pooled and exhaustively dialyzed overnight against saline buffer containing 97 mM potassium iodide, then with saline buffer without potassium iodide. Specific activities were 300 -1000 dpm/ng of protein for 125 I-labeled HDL and 5-25 dpm/ng of CE for [ 3 H]CHE-HDL.
Preparation of 125 I-Apolipoprotein-and [ 3 H]Cholesteryl Hexadecyl Ether-labeled Reconstituted HDL Particles-Briefly, human apoA-I/ phospholipid (PL) discoidal complexes were prepared by adding 5.5 mg of total lipid composed of 1-palmitoyl,2-oleoyl-phosphatidylcholine (POPC, ϩ99% grade, Avanti Polar Lipids) and cholesteryl oleate taken from stock solutions in chloroform to a glass tube and mixing them so that an initial mole ratio of 99.5:0.5, 99.0:1.0, or 97:3 PL-to-CE was achieved. Thereafter, 40 Ci of [ 3 H]CHE was added as a trace label to monitor CE transfer. This mixture of lipid and radiolabeled markers was dried into a film under a stream of nitrogen, then dried under vacuum. The lipid was subsequently hydrated in saline buffer, vortexed to generate multilamellar vesicles, and sodium cholate (4.2 mg from a 30 mg/ml solution in saline buffer) was added to generate mixed micelles; the mixture was then incubated at 37°C with immersion for at least 30 min in a bath sonicator to clear the initially turbid solution. The PL vesicles and sodium cholate were then allowed to further react for the remainder of 1.5 h with vortex mixing every 15 min. Next, apoA-I in saline buffer was added to the detergent/lipid mixture at a mole ratio of 100:1 lipid-to-protein, which typically required 2 mg of apoA-I. The lipid/detergent/protein complexes were then aspirated into a syringe containing Bio-beads (Bio-Rad), and the solution was rotated for 3 h at 4°C to facilitate the removal of cholate; the apoA-I/PL/CE discoidal complexes were then isolated by filtration of the contents of the syringe through a 0.45-m sterile filter. The sizes of the reconstituted HDL particles (rHDL) generated were assessed by native gel polyacrylamide gradient gel electrophoresis (PAGGE) using a Phast System as described previously (32,33). The [ 3 H]CHE reconstituted HDL particles were radioiodinated as described above, and a homogenous population of particles was isolated following elution from a Superdex HR-200 (30 ϫ 1 cm) column to remove any remaining vesicular structures and lipid-free apoA-I. Specific activities were typically 400 -1000 dpm/ng of apoA-I and 3-4 dpm/ng of PL for recombinant HDL particles. Protein content was determined by a modified Lowry assay (34), PL content was determined by the Bartlett assay, and unesterified cholesterol and CE were measured by gas-liquid chromatography, without and with prior saponification, as described (33).
In some instances, particles were dual-labeled with [  Determination of HDL Cell Association and HDL-CE-selective Uptake-Mouse Y1 and Y1-BS1 adrenal cells (27) were grown in Ham's F-10 medium supplemented with 12.5% horse serum and 5% heatinactivated fetal bovine serum (FBS) and 50 g of Gentamycin/ml of medium (complete medium) in a humidified incubator equilibrated to 5% CO 2 , 95% air at 37°C. For experiments, cells were seeded at 1.5 ϫ 10 6 cells/well in six-well plates or at 0.5 ϫ 10 6 cells/well in 12-well plates. After 48 h, medium was replaced with 1.5 ml of serum-free medium containing either 100 nM ACTH (Sigma) or Cortrosyn, a synthetic 1-24 ACTH analogue (Organon), to up-regulate SR-BI expression (25). After 24 h, wells containing cells and empty wells that served as "no cell" controls were washed with serum-free medium, and serum-free medium containing 100 nM ACTH (or Cortrosyn) plus either duallabeled HDL or dual-labeled discoidal rHDL particles was added, and wells were incubated for 2 h at 37°C. Time-course experiments were carried out over a 4-h period. To eliminate loss of HDL by binding to plastic, the polypropylene test tubes used to prepare serum-free test media were initially rinsed with a 10% bovine serum albumin solution, exhaustively washed with phosphate-buffered saline (PBS), then allowed to air-dry prior to placing 125 I-HDL, dual-labeled HDL, or duallabeled discoidal particles into them. All incubations were performed in a humidified incubator equilibrated to 5% CO 2 , 95% air at 37°C unless otherwise stated.
At the end of the incubation period, medium was removed and placed in glass tubes; 250 l of a 50% trichloroacetic acid solution was added, and 125 I-degradation products released from the cells were measured as monoiodotyrosine by the method described by Goldstein et al. (13). Cells were washed four times with 2 ml of ice-cold PBS containing 0.1% bovine serum albumin and then with a final wash of 4 ml of PBS. Cells were lysed with 1.25 ml of 0.1 N NaOH, and aliquots of the lysates were taken for: 1) heptane extraction to determine cell-associated HDL-CE (27), 2) trichloroacetic acid precipitation to distinguish between HDLcell association (trichloroacetic acid-insoluble) and internalized and degraded apolipoprotein (trichloroacetic acid supernatant), and 3) protein determination by a modified Lowry (34) to measure cellular protein mass. This analysis assumes that cell-associated HDL is intact and trichloroacetic acid-precipitable, whereas internalized apolipoprotein is degraded and either remains in the cells or is released from the cells as 125 I-degradation products.
The extent of HDL whole particle association with cells was traced by the 125 I-labeled apolipoprotein tracer, and the total uptake of HDL-CE was traced by [ 3 H]CHE. The calculation of cell-associated HDL and HDL-CE-selective uptake was carried out as follows. 125 I and 3 H radioactivity obtained from ␥ and scintillation counting were adjusted for the volume of cell lysate used and for milligrams of cell protein per well. Cell-associated HDL apolipoprotein was monitored by the trichloroacetic acid-precipitable 125 I cpm in the cell lysate, and HDL apolipoprotein endocytosis and degradation was determined as the sum of the cell lysate trichloroacetic acid-soluble material and the 125 I degradation products in the overlaying medium after subtracting "no cell" controls; these values are expressed as nanograms of HDL protein per mg of cell protein. These values were converted to HDL-CE/mg of cell protein on the basis of the CE-to-protein ratio of the particles. Total HDL-CE uptake was determined by heptane extraction of cell lysates and expressed as nanograms of CE/mg of cell protein. HDL-CE-selective uptake was calculated by subtracting the sum of cell-associated HDL-CE and HDL-CE taken up by endocytosis from the total cellular HDL-CE as determined from the 3 H radioactivity. Data were obtained from triplicate wells, and each experiment was repeated at least twice. The cellular uptake of DPPC relative to CE was measured using duallabeled rHDL particles containing [ 14 C]DPPC and [ 3 H]CHE. The processing of these particles by the cells was monitored in parallel experiments using a subset of the HDL particles that were dual-labeled with HDL binding isotherms were analyzed as follows. The total cellassociated 125 I-HDL as a function of HDL concentration was analyzed by nonlinear regression (Prism) and tested for best fit to equations for three models, which describe: 1) high affinity and saturable binding to a single site (SR-BI), 2) high affinity and saturable binding to 2 sites (SR-BI being one of them), or 3) high affinity and saturable binding to one site (SR-BI) plus low affinity and nonsaturable binding to other sites. The best fit was found for the model with high affinity binding to a specific site (SR-BI) plus low affinity binding to other sites. The equation is given below, where B total is the measured amount of HDL bound, [B max ] is the amount of HDL bound at saturating concentrations of HDL, K HA is the apparent high affinity K d , and C is the slope of the low affinity nonsaturable process. B total was resolved into high and low affinity components by determining C and subtracting C [HDL] from B total to generate the binding isotherm for the high affinity HDL interaction. The same equation (in which the high affinity term is analogous to the Michaelis-Menten equation) was also used to estimate K HA , the apparent K m for the rate of CE-selective uptake. In some experiments, specific cellassociated 125 I-HDL was calculated after subtracting values for bound HDL obtained in the presence of a 40-fold excess of unlabeled HDL from total cell-associated HDL (cf. Ref. 21). In this case, the binding isotherm was subjected to Scatchard analysis to estimate B max and K HA . Temperature Dependence of HDL-CE-selective Uptake-For temperature dependence studies, cells were grown in six-well plates and treated with either ACTH or Cortrosyn as described above. On the day of the experiment, cells were washed with 2 ml of Ham's F-10 medium supplemented with 100 mM HEPES (pH 7.4). Plates were equilibrated in water baths of the appropriate temperature for at least 5 min, and the test medium, equilibrated at the same temperature, was added. Visual examinations of the cells prior to, during, and after the 2-h incubation period showed no obvious cell toxicity. Uptake studies were done at the following temperatures (°C): 4, 15, 23, and 37. The activation energy for HDL-CE-selective uptake was calculated from the Arrhenius equation.

Relationship between HDL Binding and HDL-CE-selective
Uptake-In order to test the relationship between HDL particle binding and HDL-CE-selective uptake, we examined the concentration dependences for HDL cell association and HDL-CEselective uptake. The data in Fig. 1 for HDL cell association represent the average from six separate experiments in which triplicate wells were incubated for 1 h. These experiments include data using single-labeled 125 I-HDL and dual-labeled HDL as these particles exhibited similar extents of cell association. For selective uptake measurements, data from 2-h incubations were used to enhance the sensitivity. As shown in Fig.  1, both processes showed a concentration dependence indicative of high and low affinity components. The high affinity component for each process was resolved as described under "Experimental Procedures." These results are consistent with those reported by Temel et al. (27). From non-linear regression analysis, half-maximal high affinity cell association of HDL occurred at 8.7 Ϯ 4.7 g/ml and the K m of HDL-CE-selective uptake was 4.5 Ϯ 1.5 g/ml, as traced by [ 3 H]CHE. These values are similar, suggesting that the two processes are linked and, indeed, there is a very high correlation between high affinity HDL-cell association and high affinity HDL-CE-selective uptake (Fig. 2). A similar correlation is also observed if total HDL-cell association and total HDL-CE-selective uptake are plotted (data not shown). When high affinity cell-associated 125 I-HDL was determined by subtracting values obtained in the presence of a 40-fold excess of unlabeled HDL, half-maximal cell association of HDL occurred at 38 Ϯ 14 g/ml and the K m of HDL-CE-selective uptake was 30 Ϯ 10 g/ml. These K d and K m values are similar to those reported by Pittman et al. (10). The use of excess unlabeled ligand often causes problems in estimating the slope of the low affinity component so that K d is overestimated (35). Therefore, we believe the K d and K m values derived by fitting the data to a two-site model by nonlinear regression are probably more accurate. In selected experiments, the uptake of [  Mechanism of CE Transfer into Cells-It is well established that, during CE-selective uptake from donor particles, CE is taken up into cells without the uptake and lysosomal degradation of the apolipoproteins (see Ref. 36 for review). Much less is known about the uptake of the other lipid components of the donor particles. To examine the selectivity of SR-BI-mediated uptake for CE, the rates of PL and CE uptake from rHDL particles into cells were compared. Discoidal rHDL was radiolabeled with [ 3 H]CHE to track CE uptake and [ 14 C]DPPC to monitor PC uptake. Table I summarizes the composition of the particles used, the rates of CE and PC total cell association, the percentage of total and selective uptake of CE and PC and the relative fractional selective uptake of CE and PC. Note that the absolute rates of PC uptake are 25-fold higher than those for CE; however, the percentage of uptake of PC from rHDL as tracked by the DPPC label is about 4-fold lower than the percentage of uptake of CE. These results indicate that there is a non-stoichiometric uptake of HDL lipid components with CE being preferred over PC. In parallel experiments using [ 3 H]CHE/ 125 I-rHDL, we determined that 12% of the total CE uptake was due to rHDL cell association plus endocytosis and degradation and 88% was due to selective uptake. These values were used to calculate the rate constant (percentage of selective uptake per 2 h) for selective uptake of CE and PC. The ratio of the CE and PC rate constants is about 6/1, indicating that the selective uptake mechanism mediated by SR-BI discriminates for CE over PC (about 55% of total PC uptake was due to selective uptake). The kinetic data summarized in Table II further demonstrate that the structure of the transferring lipid molecule affects the rate of selective uptake into the Y1-BS1 cells. In particular, the rate of uptake of PL molecules depends upon the structure of their polar group. Thus, the rate constant for PS-selective uptake is about 8 times that of DPPC (Table  II), which is similar to the ratio observed for CE (Table I). In contrast, the rate constant for PI-selective uptake is similar to that for DPPC whereas the rate constant for SM-selective uptake is much smaller (Table II); in fact, only about 10% of the total SM uptake is due to the selective pathway. It is important to note that the rates of radiolabeled PL-selective uptake were all measured using a common rHDL particle (containing Յ1.5 mol% of PL radiolabel) so that there are no confounding issues of either variations of particle size and charge, or differences in rHDL-cell interactions.
To test the influence of different CE contents on HDL-CEselective uptake, reconstituted discoidal particles were prepared with different ratios of CE to apoA-I. Table III summarizes the rates of CE-selective uptake to Y1-BS1 adrenal cells from rHDL particles containing 0.6 mol% CE and 1.3 mol% CE. The binding of both rHDL particles to the cells was similar (data not shown) but increasing the CE content of rHDL particles resulted in an increase in the rate of CE-selective uptake. This increase in uptake was proportional to the amount of CE initially present in the donor particles; the ratio of the CE contents of the two types of particles was 2.3 to 1, while the ratio of CE-selective uptake was 2.6 to 1 (Table III). It is interesting to note that the uptake of CE is much greater from spherical human HDL 3 particles, which are relatively enriched in CE relative to discoidal rHDL. Apparently, this is in contrast to the results of Pittman and colleagues (19), who found that larger HDL (with more CE) gave less CE-selective uptake. The reason for this discrepancy is not entirely clear, but it may be due to effects of HDL particle size on binding to SR-BI.
The temperature dependence for the selective uptake process was measured using particles containing tracers for apolipoprotein and CE. For these experiments, cells were equilibrated at four temperatures between 4 and 37°C, and selective CE uptake was determined in a 2-h assay. Fig. 3 illustrates a typical Arrhenius plot generated when the temperature dependence of HDL-CE-selective uptake was measured; it is apparent that the rate of HDL-CE-selective uptake is reduced by cooling the system. As shown in Table IV, the activation energies for the uptake of CE from HDL 3 and discoidal rHDL into Y1-BS1 cells were both approximately 9 kcal/mol. The degree of binding of the HDL particles to the cells was also decreased progressively by reducing the temperature. However, the temperature dependence of HDL binding was less than that of HDL-CE-selective uptake, so that the ratio of selective uptake to binding decreased by a factor of about 2 on reducing the temperature from 37°C to 4°C. This result is consistent with the transfer of CE molecules out of the bound HDL particles being the rate-limiting step.

Role of HDL Binding to SR-BI in CE-selective Uptake-
Despite extensive investigations of the phenomenon of CEselective uptake from HDL, the molecular mechanism involved has not been defined in detail. In this study, we obtained quantitative values for the parameters characterizing the phenomenon in order to elucidate both the relationship between HDL binding to SR-BI and the selective uptake of CE, as well as the molecular basis for the facilitation of this CE uptake from HDL. The recent demonstration that SR-BI mediates both HDL binding and HDL-CE-selective uptake has provided an important link between the selective uptake pathway and a specific cell surface receptor (21). Prior to this finding, there was no convincing evidence that a specific receptor is involved because there was no apparent specificity for an HDL apolipoprotein (19), and model membranes devoid of protein also exhibited HDL-CE-selective uptake, albeit to a much lesser extent (37). We sought to test whether the interaction of HDL with SR-BI is crucial for HDL-CE-selective uptake and hence examined the concentration dependence of these two processes.
These studies were carried out with the Y1-BS1 adrenocortical cell line to provide a physiological model for the selective uptake process. In addition, previous studies with Y1-BS1 cells have shown that SR-BI is the major component responsible for the cell association of HDL particles and for the delivery of HDL-CE to the steroidogenic pathway (27). Thus, the observation that at least 78% of the HDL-CE-selective uptake in these cells could be blocked by an antibody directed against SR-BI is consistent with this protein being the only receptor involved.
Using non-linear regression analyses of the concentration dependences of cell associated HDL and HDL-CE-selective uptake, we found that the K d for cell association was 8.7 Ϯ 4.7 g/ml (equivalent to 311 Ϯ 168 nM apoA-I) and the K m of HDL-CE-selective uptake was 4.5 Ϯ 1.5 g/ml (161 Ϯ 53 nM apoA-I). The dissociation constant (K d ) for the HDL (apoA-I) SR-BI interaction is ϳ150-fold higher than that of LDL with its receptor (K d ϳ2 nM apo B-100). The dissociation constant (K d ) is a function of both the on and off rates (K d ϭ k off /k on ) of a ligand and its receptor. For an approximately spherical ligand, the Stokes-Einstein equation predicts that the diffusion coefficient is inversely proportional to particle radius, so that the diffusion-limited on-rate for HDL (ϳ10 nm diameter) should be some 2 times faster than that of LDL (ϳ25 nm). Knowing the ratio of the HDL and LDL K d values, it can be estimated that the off-rate of HDL from SR-BI is about 300 times higher than the off-rate of LDL from its receptor. These findings are consistent with the idea that HDL comes on and off SR-BI relatively readily and may explain, in part, why HDL particles are not endocytosed by SR-BI. These rapid on and off interactions would facilitate lipid delivery by releasing the HDL particle once the HDL core CE has been transferred to the membrane, thereby freeing SR-BI for another round of interaction.
The similarity of the K d and K m values for HDL binding and HDL-CE-selective uptake argues that the two processes are linked. A high degree of correlation between HDL binding and CE-selective uptake is seen over the entire range of HDL binding (Fig. 2), indicating a tight correspondence between HDL binding to SR-BI and the subsequent transfer of HDL-CE into the cell.
Mechanism of SR-BI-mediated Uptake of Lipids-To determine the mechanism of SR-BI-mediated CE uptake, the fractional uptake of different components of recombinant HDL particles was examined. There are three possible pathways by which lipid molecules in the rHDL particles can become cellassociated: 1) binding and uptake of intact rHDL, 2) diffusion of individual lipid molecules from rHDL through the aqueous phase to the cell plasma membrane, and 3) the selective uptake pathway mediated by SR-BI. The first pathway leads to stoichiometric uptake of the constituents of the rHDL particle. Thus, any non-stoichiometric lipid uptake is due to the second and third pathways. The situation is relatively simple for CE a The rHDL particles were the same as those described in Table I   because this very hydrophobic molecule does not transfer significantly by diffusion through the aqueous phase (38) so that the non-stoichiometric uptake of CE is entirely due to the SR-BI-mediated selective pathway. Uptake of PL molecules can occur by both aqueous diffusion and selective uptake pathways. Consequently, the assumption that non-stoichiometric uptake of PL is entirely due to SR-BI-mediated selective uptake leads to the PL rate constants in Tables I and II being upper limits for PL-selective uptake. However, the differences between the rate constants for PL-selective uptake (Table II) must arise from SR-BI-mediated uptake. This follows because the rate constants for the spontaneous transfer of PL molecules such as PC, PS, and SM from rHDL particles via aqueous diffusion are affected primarily by the hydrophobic content and not the nature of the polar headgroup (39,40). With respect to the mechanism of HDL lipid-selective uptake, the marked differences in selective uptake between PL classes indicate that transient fusion of the HDL PL with the outer leaflet of the plasma membrane does not occur. If this were the case, it is difficult to envision how this process could give rise to the observed specificity of PL transfer.
The results in Table I show that the rate constant for selective uptake of CE is approximately 6-fold higher than that for PC. Thus, in Y1-BS1 adrenocortical cells, the SR-BI-mediated selective uptake process discriminates in favor of CE as compared with this class of PL. This result indicates that the selective uptake process does not arise simply from bringing the SR-BI-docked HDL particle close to the plasma membrane, but must involve subsequent steps that permit the selective removal of CE from the particle while leaving the PC behind. It is striking that the SR-BI-facilitated selective uptake seems not to discriminate between CE and PS molecules but does discriminate against PC, PI, and SM molecules relative to CE. The fact that the rate constants for selective uptake vary as shown in Tables I and II suggests that the putative non-aqueous channel created by SR-BI (see below and Fig. 4) can distinguish aspects of the structure of the transferring lipid molecules. The nonpolar CE molecule can transfer readily from the bound rHDL particle into the channel, whereas the more polar PC, PI, and SM molecules transfer less well. The charge on the PL molecule seems not to be critical, because PI and PS are both negatively charged at neutral pH but PS transfers as well as CE. Understanding of the molecular basis for this specificity requires more knowledge of the structure of SR-BI, but it is possible that the receptor contains a recognition site for PS that facilitates transfer of PS molecules out of the bound rHDL particles.
Comparison of reconstituted HDL particles containing different amounts of CE showed that CE-selective uptake increases proportionally with particle CE content (Table III). This was demonstrated by exposing cells to equivalent protein concentrations of donor rHDL which were virtually identical in composition except that one was 2.3-fold-enriched in CE versus the other. In this way, cells were exposed to identical numbers of rHDL particles. CE uptake was even greater from spherical HDL 3 particles, which are even more enriched in CE, suggesting a mechanism in which CE moves down its concentration FIG. 4. Model of SR-BI-mediated selective uptake of CE from HDL. This model proposes that SR-BI contains a non-aqueous channel, which excludes water, and serves as a conduit for hydrophobic CE molecules diffusing from bound HDL down their concentration gradient to the cell plasma membrane. The scheme depicts a channel formed by a single SR-BI molecule, but it is possible that self-association of SR-BI is required to create the channel.   gradient from the SR-BI-docked HDL particle to the plasma membrane. Since the capacity of the plasma membrane to accommodate CE is limited to 2-3 mol% with respect to the membrane PL, continued selective uptake must involve the removal of CE from the cytoplasmic side of the membrane either by a CE transfer mechanism or via CE hydrolysis. In the case of the adrenocortical cell, cholesterol utilization by the steroidogenic pathway would serve to drive this process. The temperature dependence studies provided important quantitative information relative to the mechanism of the SR-BI-mediated uptake process, for which the rate-limiting step is presumably the selective transfer of CE molecules out of the HDL particle. The activation energy for the rate-limiting step in CE-selective uptake from either HDL 3 or from reconstituted HDL (ϳ8 -9 kcal/mol) is lower than the activation energy for the desorption of free cholesterol from HDL particles into the aqueous phase (Table IV). The aqueous solubility of CE is very low, and thus its movement into the cells via the aqueous phase would be extremely energetically unfavorable. It has been estimated that if CE transfers via the aqueous phase, then the activation energy associated with this process would be enormous and the process would require years (38). The finding that the activation energy for HDL-CE uptake is relatively low strongly suggests that HDL-CE uptake occurs via a non-aqueous pathway. We postulate that SR-BI binds HDL at the cell surface in such a way as to provide a "channel" from which water is excluded and along which CE molecules diffuse down the concentration gradient from HDL particles to the cell plasma membrane (Fig. 4). At this time, it is not clear whether one or several SR-BI molecules are involved in formation of the channel. This model is consistent with the results of a recent mutagenesis studies of SR-BI showing that its extracellular domain is responsible for facilitating transfer of lipid from bound HDL particles to the cell (41,42).