Separation of Lipid Transport Functions by Mutations in the Extracellular Domain of Scavenger Receptor Class B, Type I*

Scavenger receptor class B, type I (SR-BI) shows a variety of effects on cellular cholesterol metabolism, including increased selective uptake of high density lipoprotein (HDL) cholesteryl ester, stimulation of free cholesterol (FC) efflux from cells to HDL and phospholipid vesicles, and changes in the distribution of plasma membrane FC as evidenced by increased susceptibility to exogenous cholesterol oxidase. Previous studies showed that these multiple effects require the extracellular domain of SR-BI, but not the transmembrane and cytoplasmic domains. To test whether 1) the extracellular domain of SR-BI mediates multiple activities by virtue of discrete functional subdomains, or 2) the multiple activities are, in fact, secondary to and driven by changes in cholesterol flux, the extracellular domain of SR-BI was subjected to insertional mutagenesis by strategically placing an epitope tag into nine sites. These experiments identified four classes of mutants with disruptions at different levels of function. Class 4 mutants showed a clear separation of function between HDL binding, HDL cholesteryl ester uptake, and HDL-dependent FC efflux on one hand and FC efflux to small unilamellar

Scavenger receptor class B, type I (SR-BI) 1 is an ϳ82-kDa cell-surface glycoprotein that was first identified by its sequence homology to CD36 (1,2) and later characterized as one of the first physiologically relevant receptors for high density lipoprotein (HDL) particles (Ref. 3; for review, see Refs. 4 -6). Early analyses of SR-BI knockout mice revealed altered plasma HDL metabolism and reduced adrenal gland cholesteryl ester (CE) accumulation (7,8). Several recent reports have shown that alterations in murine SR-BI expression can have profound effects on biliary cholesterol excretion, red blood cell development, female infertility, atherosclerosis, and the development of coronary heart disease (9 -17), all of which seem to be related to changes in HDL and/or cholesterol. Taken together, these studies are consistent with SR-BI playing a major role in HDL cholesterol metabolism in vivo.
SR-BI mediates its effects on HDL CE metabolism by facilitating the transport of lipids to cells in a process termed selective uptake (3-5, 18 -23). Contrary to the classic low density lipoprotein receptor endocytic pathway, in which the entire lipoprotein is internalized in clathrin-coated pits and degraded (24), HDL binds SR-BI, and the core CE is delivered to the plasma membrane without the concomitant uptake and degradation of the entire HDL particle (18 -23). Furthermore, low density lipoprotein CE delivered by the low density lipoprotein receptor is hydrolyzed in the lysosomal pathway by an acidic CE hydrolase (24,25), whereas HDL CE delivered by SR-BI is hydrolyzed extralysosomally (26) by a neutral CE hydrolase (27,28). In fact, SR-BI has been shown to deliver HDL CE into a metabolically active membrane pool, where it is efficiently hydrolyzed by cell type-specific neutral CE hydrolases (29).
In addition to the uptake and metabolism of HDL CE, SR-BI stimulates the bidirectional flux of free cholesterol (FC) between cultured cells and lipoproteins (30 -34), an activity that may be responsible for net cholesterol efflux from peripheral cells as well as the rapid hepatic clearance of FC from plasma HDL and its resultant secretion into bile (9,35). SR-BI also increases cellular cholesterol mass and alters cholesterol distribution in plasma membrane domains as judged by the enhanced sensitivity of membrane cholesterol to extracellular cholesterol oxidase (31,36). Moreover, the delivery of HDL FC by SR-BI results in efficient delivery of FC for esterification (36,37). Together, these data support the idea that, similar to HDL CE, SR-BI delivers HDL FC into a metabolically active membrane pool. Whether this membrane pool reflects the localization of SR-BI in a physically distinct membrane domain or its interaction with other membrane proteins is not known.
Previous studies have shown that several key lipid transport functions of SR-BI, including the ability to alter membrane cholesterol domains, are dependent on its extracellular region (37)(38)(39). However, to date, it is unclear whether 1) the extracellular domain of SR-BI mediates multiple distinct activities by virtue of discrete functional domains or 2) the distinct func-tions of SR-BI are, in fact, secondary to, and driven by, changes in cellular cholesterol content and distribution. To address this question, we mutagenized the extracellular domain of murine SR-BI by strategically placing an epitope tag from the adenovirus E4/5 protein into nine sites in the SR-BI extracellular domain (designated A-II through A-X in Fig. 1 and "Experimental Procedures"). We expressed the mutant receptors in COS-7 cells, checked for cell-surface expression of the mutant receptors, and assessed the ability of these epitope-tagged mutants to mediate multiple aspects of cellular lipid metabolism. These experiments identified four classes of SR-BI mutants that revealed that 1) all SR-BI activities do not derive from the ability of SR-BI to load cholesterol into the plasma membrane, and 2) there is a separation of function between SR-BI-mediated HDL CE uptake and HDL FC efflux on one hand and FC efflux to small unilamellar vesicle (SUV) acceptors and an increased cholesterol oxidase-sensitive pool of membrane FC on the other. These data provide clear evidence for a difference in the pathways for SR-BI-mediated FC efflux to HDL and to phospholipid vesicles.
All plasmids were prepared using endotoxin-free QIAGEN maxi-prep kits and sequenced throughout the SR-BI coding region to confirm the correct epitope insertion and to ensure that no point mutations had been generated during the amplification process. DNA sequencing was

FIG. 2. Cell-associated HDL and selective HDL COE uptake mediated by wild-type SR-BI and epitope tag insertion receptors.
Immunoblot analysis of protein lysates from COS-7 cells transiently expressing SR-BI or the tag insertion mutants detected with antibody directed against the SR-BI C-terminal domain shows the expression level of each receptor compared with that of wild-type SR-BI (A). Three immunoblots from separate selective uptake experiments were scanned; densitometric values in arbitrary units (AU) were obtained; and means Ϯ S.E. were calculated to show relative total receptor expression levels (B). Flow cytometry was also performed on parallel wells of cells, and the mean of the fluorescence-positive cells Ϯ S.E. from two separate experiments is shown to represent relative receptor expression levels on the cell surface (C). Parallel wells of cells were incubated at 37°C for 1.5 h with 125 I-DLT-and [ 3 H]COE-labeled HDL (10 g/ml HDL protein), after which cells were processed to determine cell-associated HDL COE (D) and selective HDL COE uptake (E). The efficiency of selective HDL COE uptake was determined by normalizing the amount of selective HDL COE uptake to the amount of cell-associated HDL COE (F). Values represent means Ϯ S.E. of 12 replicates from four separate experiments after subtraction of values obtained with vector-transfected cells. Values in all graphs are expressed relative to wild-type SR-BI values, which were set at 100%. The mean experimental values for a representative data set for D (before subtraction of the value for vector-transfected cells) for vector, SR-BI, and mutant A-VI were 4.3, 70.6, and 16.6 ng of HDL COE/mg of cell protein, respectively. The mean experimental values for a representative data set for E (before subtraction of the value for vector-transfected cells) for vector, SR-BI, and mutant A-VI were 32, 1470, and 296 ng of HDL COE/mg of cell protein, respectively. performed by the automated sequencing facility at the State University of New York (Stony Brook). Reactions were prepared using a dye termination cycle sequencing kit and analyzed on an Applied Biosystems Model 373 DNA Sequencer with an Excel upgrade as recommended by the manufacturer (PE Applied Biosystems).
Transient Transfection of COS-7 Cells-COS-7 cells were maintained in Dulbecco's modified Eagle's medium (Invitrogen), 10% calf serum (Atlanta Biologicals, Inc.), 2 mM L-glutamine, 50 units/ml penicillin, 50 g/ml streptomycin, and 1 mM sodium pyruvate and transfected as described previously (39). The following day, two 10-cm dishes of transfected cells were trypsinized and resuspended in a total volume of 12 ml with fresh medium, and 0.5 or 1 ml was dispensed to each 22-mm (12-well plate) or 35-mm (6-well plate) well, respectively (one 10-cm dish is equivalent to one 12-well or one 6-well plate). The cells were assayed 48 h post-transfection unless otherwise indicated.
Immunoblot Analysis-Transiently transfected cells expressing SR-BI (in 35-mm wells) were washed twice with phosphate-buffered saline (PBS; Invitrogen), pH 7.4, and lysed with 300 l of Nonidet P-40 cell lysis buffer (41, 42) containing 1 g/ml pepstatin, 0.2 mM phenylmethylsulfonyl fluoride, 1 g/ml leupeptin, and 10 g/ml aprotinin. Protein concentrations were determined by the method of Lowry et al. (43). Immunoblots with antibodies directed to SR-BI confirmed mutant and wild-type receptor expression. Equal amounts of total cell proteins were separated on precast 10% SDS-polyacrylamide gels (Bio-Rad), blotted onto nitrocellulose membranes (Bio-Rad), and detected using antibody directed against the C-terminal domain of SR-BI or the respective monoclonal antibody epitope, horseradish peroxidase-conjugated anti-rabbit secondary antibody, and SuperSignal West Pico reagent (Pierce). Blots were quantified using a Bio-Rad Model GS-700 imaging densitometer and MultiAnalyst software.
Cell-surface Receptor Expression Levels by Flow Cytometry-Transiently transfected COS-7 cells (in 35-mm wells) were washed with 2 ml of cold PBS. Cells were removed from plates by the addition of 1 ml of PBS and 0.5 mM EDTA and incubation for 5-7 min at room temperature. Cells were placed in a microcentrifuge tube, centrifuged at 200 ϫ g for 2-3 min, and resuspended in 100 l of PBS and 1% bovine serum albumin (BSA). Anti-SR-BI primary antibody 356 at a concentration of 0.48 mg/ml IgG or anti-M45 hybridoma medium at a 1:250 dilution was added to the cells and incubated for 1 h at 4°C. The cells were centrifuged at 200 ϫ g for 2-3 min, and the supernatant was aspirated. Cells were washed twice with 0.5 ml of PBS and 1% BSA before incubation with secondary antibody (3 l of fluorescein-or phycoerythrin-conjugated antirabbit antibody) in 300 l of PBS and 1% BSA for 30 min at 4°C. Cells were washed three times with 0.5 ml of PBS and 1% BSA and fixed in 0.5 ml of 1% formaldehyde in PBS and 1% BSA for 15 min at 4°C with gently shaking. Following incubation with fixative, the cells were centrifuged at 200 ϫ g for 2-3 min and resuspended in 0.5 ml of PBS and 1% BSA.
Fluorescence intensities were measured using a BD Biosciences FACS-Advantage cell sorter or a FACScan flow cytometer.
Immunofluorescence-Transiently transfected COS-7 cells were replated 24 h post-transfection onto a 12-well plate containing glass microscope slide coverslips. After 24 h, the medium was removed, and the cells were washed at room temperature with PBS. The cells were fixed for 1 h in 4% (w/v) paraformaldehyde in 77 mM PIPES, pH 7.5 (44); washed with PBS; and blocked for 1 h with 3% BSA and 10 mM glycine in PBS. Antibody against the extracellular domain of SR-BI was diluted in 3% BSA and applied to cells for 1 h at room temperature. Cells were washed as described above and incubated with Alexa 488-conjugated secondary antibodies for 30 min at room temperature. Cells were then washed with PBS and mounted using ProLong antifade mounting medium (Molecular Probes, Inc.). Cells were examined using a Leica DMIRE2 confocal microscope, and images were collected using Leica confocal software.

Preparation of 125 I-Dilactitol Tyramine (DLT)-and [ 3 H]Cholesteryl
Oleyl Ether (COE)-labeled HDL and 125 I-HDL-Human HDL 3 (1.125 g/ml Ͻ Ͻ 1.210 g/ml), herein referred to as HDL, was isolated by sequential ultracentrifugation (45). The HDL was labeled with [ 3 H]COE (Amersham Biosciences) using recombinant CE transfer protein (Cardiovascular Targets, Inc.) as described (46) with the following modifications. HDL and CE transfer protein were incubated with [ 3 H]COE (dried down on the glass vial) for 5 h at 37°C. Labeled particles were reisolated by gel exclusion chromatography on a 25-ml Superose 6 column (Amersham Biosciences). The HDL was then labeled with 125 I-DLT as described previously (39). Particles were dialyzed against four changes of 150 mM NaCl, 10 mM potassium phosphate buffer, pH 7.4, and 1 mM EDTA and stored at 4°C under argon. The average specific activity of the 125 I-DLT-and [ 3 H]COE-labeled HDL was 650 dpm/ng of protein for 125 I and 37 dpm/ng of protein for 3 H. For some experiments, HDL was labeled using the iodine monochloride method (47), and the average specific activity of the 125 I-HDL was ϳ500 dpm/ng of protein.
HDL Cell Association, Selective COE Uptake, and Apolipoprotein Degradation-Transiently transfected COS-7 cells (in 35-mm wells) were washed once with serum-free Dulbecco's modified Eagle's medium and 0.5% BSA. 125 I-DLT-and [ 3 H]COE-labeled HDL particles were added at a concentration of 10 g/ml protein (unless otherwise indicated) in serum-free Dulbecco's modified Eagle's medium and 0.5% BSA. After incubation for 1.5 h at 37°C, the medium was removed, and the cells were washed three times with PBS and 0.1% BSA, pH 7.4, and one time with PBS, pH 7.4. The cells were lysed with 1.1 ml of 0.1 N NaOH, and the lysate was processed to determine trichloroacetic acidsoluble and -insoluble 125 I radioactivity and organic solvent-extractable 3 H radioactivity. The values for cell-associated HDL apolipoprotein, total cell-associated HDL COE, and the selective uptake of HDL COE were obtained as described previously (39).
Cholesterol Efflux Assay-Transiently transfected COS-7 cells were replated onto 11-mm wells in growth medium. Cells were labeled for 24 h with 5 Ci/ml [ 3 H]cholesterol (PerkinElmer Life Sciences) in Dulbecco's modified Eagle's medium containing 10% calf serum immediately after reseeding. Cells were washed, and [ 3 H]cholesterol efflux was measured at 2 h in triplicate using different concentrations of HDL acceptor or palmitoyloleoylphosphatidylcholine-containing SUV as described previously (31). The release of radioactive cholesterol was measured by scintillation counting of filtered aliquots of acceptor-containing medium and expressed as the fraction of the total 2-propanol-soluble label in the cells plus the label that was released into the medium. Fractional efflux values were corrected for the small amount of radioactivity released in the absence of acceptor.
To normalize the data for FC efflux to the amount of cell-surface receptor expressed in the transient transfections, modified HDL cell association assays were performed in parallel with the FC efflux studies. COS-7 cells (in 22-mm wells) were washed once with serum-free minimal essential medium/HEPES and 1% BSA, and monochloridelabeled 125 I-HDL was added at 10 or 25 g of protein/ml of minimal This graph was plotted from 0 to 20 g/ml to emphasize the high efficiency portion of the selective HDL COE uptake curve. The K m value for wild-type SR-BI is 6.6 Ϯ 1 g/ml HDL protein, and that for mutant A-VI is 0.8 Ϯ 1.5 g/ml HDL protein using a one-site model. essential medium/HEPES. After incubation for 1.5-2.0 h at 4 or 37°C, the medium was removed, and cells were washed three times with PBS and 0.1% BSA, pH 7.4, and twice with PBS, pH 7.4. Cells were lysed with 1 ml of 0.1 N NaOH; the wells were washed with an additional 0.5 ml of 0.1 N NaOH; and the wash and lysate were counted for ␥-radiation. After counting, an aliquot was removed for protein determination (43).
Cholesterol Oxidase Assays and Cholesterol Mass-Transiently transfected COS-7 cells were replated onto 22-mm wells and labeled with 5 Ci/ml [ 3 H]cholesterol as described above. Cholesterol oxidase assays were performed 24 h post-labeling with live cells as described by Smart et al. (48) and as modified by Kellner-Weibel et al. (36). Following cholesterol oxidase treatment for 4 h at 37°C, the cell monolayers were extracted with 2-propanol, and the [ 3 H]cholesterol and [ 3 H]cholestenone were quantitated after separation by TLC (36). Cholesterol oxidase sensitivity was normalized to HDL binding as described above.

Mutagenesis of the SR-BI Extracellular Domain by Epitope
Tag Insertion-Previous studies showed that several key lipid transport functions of SR-BI, as well as its ability to alter membrane cholesterol domains, are dependent on its extracellular region (37)(38)(39). However, to date, it is unclear whether 1) the extracellular domain of SR-BI mediates multiple distinct activities by virtue of discrete functional domains or 2) the distinct functions of SR-BI are, in fact, secondary to, and driven by, changes in membrane cholesterol content and distribution.
To address this question, we mutagenized the murine SR-BI extracellular domain by strategically placing monoclonal antibody epitope sequences or "tags" into the SR-BI coding sequence, expressed the mutant receptors in COS-7 cells, confirmed cell-surface expression of the mutant receptors, and assessed the ability of these tagged mutants to mediate lipid transport. To minimize the possibility of disrupting the overall tertiary structure of SR-BI, we chose sites for epitope tag insertion by locating regions with 1) as little homology as possible between the amino acid sequences of SR-BI, CD36, and lysosomal integral membrane protein II from several different species and 2) a low probability of having ␣-helical or ␤-pleated sheet structure when comparing the predicted secondary conformations of SR-BI and CD36 (Fig. 1). We inserted a 14-amino acid sequence from the adenovirus E4/5 protein, an epitope for monoclonal antibody M45, into nine sites in the extracellular domain (designated A-II through A-X in Fig. 1 and under "Experimental Procedures"). Immunoblot analysis of COS-7 cells transiently transfected with the epitope-tagged receptors revealed that all but two of the tagged receptors, A-IX and A-X, were expressed (Fig. 2, A and B). Furthermore, all of the tagged receptors that were expressed migrated approximately the same distance upon SDS-PAGE, suggesting that none of the tag insertions had a major effect on N-linked glycosylation of Experimental values were normalized to cell-associated HDL and are expressed relative to SR-BI as described in Fig. 5. POPC, palmitoyloleoylphosphatidylcholine.
the mature protein (49). Flow cytometry using antibodies directed against the extracellular domain of SR-BI showed that all of the tagged receptors, except A-IX and A-X, were expressed to varying degrees on the cell surface (Fig. 2C). This was confirmed with the tagged receptors by flow cytometry using monoclonal antibody M45 (data not shown).
Comparison of cells expressing mutant receptors with cells expressing wild-type SR-BI revealed that insertions of the tag sequence impaired HDL binding to varying degrees, although most showed high affinity HDL binding (Fig. 2D). All of the mutant receptors were inactive in mediating selective HDL CE uptake, except A-VI, which showed activity in the range of 10 -20% compared with SR-BI in multiple assays (Fig. 2E).
When the selective uptake data were normalized to the amount of cell-associated HDL to estimate selective uptake efficiency (39), mutant A-VI showed ϳ60% of the efficiency of the wildtype receptor (Fig. 2F). The other mutants had selective uptake efficiencies less than or similar to that of CD36, indicating that, like CD36, they were unable to mediate high efficiency transfer of HDL CE to the cells.
We chose to study the A-III and A-VI mutants further for two reasons. First, they expressed well on the cell surface compared with some of the other tagged mutant receptors; and second, they displayed phenotypes that might speak to the different functions of SR-BI. For instance, A-III bound HDL particles, but did not mediate high efficiency selective HDL COE uptake. Like A-III, CD36 was previously shown to bind HDL with high affinity, but it had a greatly reduced ability to mediate selective HDL uptake compared with SR-BI (38,39). It was concluded that low efficiency selective HDL uptake by CD36 was primarily due to tethering HDL particles on the cell surface. The greater efficiency of SR-BI-mediated selective HDL uptake suggests that the extracellular domain of SR-BI participates in a unique second step in which CE is rapidly transferred from the HDL particle to the cell membrane (38). Mutant A-III may be an SR-BI in which the second step of lipid transfer is selectively disrupted. On the other hand, mutant A-VI bound HDL and mediated selective HDL COE uptake with an efficiency averaging 60% of that of the wild-type receptor.

FIG. 7. Effects of expression of SR-BI and mutants A-VI and H-VI on FC efflux to HDL and SUV and cholesterol oxidase sensitivity of membrane cholesterol. COS-7 cells transfected with vector or transiently expressing SR-BI, A-VI, or H-VI were prelabeled with [ 3 H]cholesterol and incubated with increasing concentrations of HDL (A and
Plasma Membrane Expression of A-III and A-VI-Immunofluorescence experiments were performed to support the flow cytometric data showing that the A-III and A-VI tag insertion mutants were expressed at the cell surface. Non-permeabilized COS-7 cells transiently expressing SR-BI, A-III, or A-VI were stained with antibody directed against the extracellular domain of SR-BI and viewed by confocal microscopy. As shown in Fig. 3, cells expressing each of the three receptors showed significant cell-surface staining, indicating that, like wild-type SR-BI, the epitope insertion mutants were expressed at the plasma membrane in COS-7 cells. No staining was seen with non-transfected COS-7 cells (data not shown). In addition, nonexpressing cells within the same field as receptor-expressing cells showed no antibody staining (Fig. 3). Similar results were observed with mutants A-III and A-VI using the anti-M45 epitope antibody, confirming the results obtained with the anti-SR-BI extracellular domain antibody (data not shown). At the light microscopic level, the overall cell-surface pattern of expression for the tagged mutants appeared similar to that for wild-type SR-BI.
Cell-associated HDL and Selective HDL COE Uptake Mediated by A-III and A-VI-To determine whether the reduction in HDL binding to the A-III and A-VI tagged receptors was due to a decrease in the number of cell-surface receptors or to a reduction in binding affinity for HDL, assays were performed at varying HDL concentrations. COS-7 cells transiently expressing SR-BI, A-III, or A-VI were incubated with increasing concentrations of HDL, after which cells were processed to determine cell-associated HDL protein (Fig. 4A) and selective HDL COE uptake (Fig. 4B). Fig. 4A shows that, similar to wild-type SR-BI, the A-III and A-VI receptors bound HDL with high affinity. In fact, the K d values for A-III and A-VI were comparable to that for wild-type SR-BI (micrograms/ml HDL protein (mean Ϯ S.E.): A-III, 14.6 Ϯ 1.1; A-VI, 9.9 Ϯ 3.9; and SR-BI, 12.5 Ϯ 2.8). However, the B max values were quite different (nanograms of HDL protein/mg of cell protein: A-III, 17.8 Ϯ 4.7; A-VI, 32.7 Ϯ 4.0; and SR-BI, 165.9 Ϯ 12.4), suggesting that not as many A-III and A-VI receptors capable of high affinity HDL binding were on the cell surface compared with wild-type SR-BI-expressing cells. This point was confirmed by the flow cytometric measurements (Fig. 2C) compared with the HDL cell association data (Fig. 2D) showing reduced HDL binding compared with wild-type SR-BI even when A-III and A-VI were on the cell surface at similar levels compared with SR-BI.
Measurements of selective HDL COE uptake showed that A-III-expressing cells were devoid of high efficiency selective uptake (Fig. 4B). In the HDL concentration range of 2.5-20 g/ml, A-VI, like wild-type SR-BI, showed saturable high affinity selective HDL COE uptake (Fig. 4B). Interestingly, however, A-VI did not show the additional low affinity selective uptake component seen with SR-BI that is characterized by a gradual linear increase in selective uptake at higher HDL concentrations (41,50). With HDL concentrations up to 250 g/ml, SR-BI-mediated selective uptake continued to increase, but A-VI-mediated uptake did not (data not shown). A-VIexpressing cells seemed to mediate high efficiency (but not low efficiency) selective HDL COE uptake (Fig. 4B). This led us to speculate that other lipid transport functions of A-VI may be affected by the epitope tag insertion at this site.
Effect of A-VI Expression on Cholesterol Efflux to HDL and SUV as Well as Plasma Membrane Cholesterol Distribution-In addition to mediating selective HDL COE uptake, SR-BI accelerates the efflux of FC from cells (30 -32). To determine whether the epitope insertions at amino acids 283 (A-III) and 424 (A-VI) had an effect on SR-BI-mediated cholesterol efflux, COS-7 cells expressing SR-BI, A-III, or A-VI were assayed for efflux to increasing concentrations of HDL (Fig. 5A) or SUV (Fig. 5B). Parallel dishes of the same cells were assayed for binding of 125 I-HDL to normalize cholesterol efflux to the levels of SR-BI-bound HDL. As with selective HDL COE uptake, there was no reproducible cholesterol efflux to HDL or SUV with A-III-expressing cells (data not shown). However, A-VI-expressing cells showed a somewhat reduced efflux of cholesterol to HDL and, interestingly, no efflux of cholesterol to SUV (Fig. 5, A and B). Moreover, in contrast to wild-type SR-BI, expression of A-VI did not increase the pool of plasma membrane FC available for oxidation by cholesterol oxidase (Fig. 5C). Mutant A-III also showed no FC efflux to SUV and no increase in the cholesterol oxidase-sensitive pool of membrane FC (data not shown).

Cell-associated HDL, Selective HDL COE Uptake, Cholesterol Efflux, and Changes in Plasma Membrane Cholesterol Mediated by Epitope Tag Insertion Mutant H-VI-To test
whether the altered function of A-VI was due to an amino acid insertion per se or to the particular amino acids in the adenovirus E4/5 epitope, we engineered another epitope at this site by replacing amino acids 423-428 with six histidines to form an epitope for an anti-His 6 monoclonal antibody (H-VI) (Fig. 1). H-VI showed strong expression upon Western blotting and a cell-surface expression pattern similar to that of wild-type SR-BI upon immunofluorescence analysis (data not shown). COS-7 cells expressing SR-BI or H-VI were assayed for their ability to bind HDL (Fig. 6A) and to mediate selective uptake of HDL COE (Fig. 6B). The results reveal that H-VI-expressing cells exhibited wild-type levels of HDL binding and selective HDL COE uptake as well as selective uptake efficiency (Fig.  6C).
To determine whether H-VI has altered ability to mediate cholesterol efflux, COS-7 cells expressing SR-BI, A-VI, or H-VI were assayed for FC efflux to increasing concentrations of HDL (Fig. 7, A and B) or SUV (Fig. 7, C and D). In addition, these mutants were tested for their ability to increase the cholesterol oxidase sensitivity of membrane FC (Fig. 7, E and F). Panels A, C, and E show these data in comparison with vector-transfected cells, and panels B, D, and F show the data after subtraction of control values from vector-transfected cells. As shown in Fig. 7B, mutants H-VI and A-VI showed enhanced FC efflux to increasing concentrations of HDL. Relative to SR-BI or relative to the respective values for each mutant for FC efflux to HDL, mutants A-VI and H-VI showed reduced FC efflux to SUV (Fig. 7D), with A-VI being more disabled than H-VI.
Similarly, Fig. 7F shows that the cholesterol oxidase sensitivities with mutants A-VI and H-VI were reduced much more compared with those with SR-BI relative to the respective reductions in FC efflux to HDL (Fig. 7B).
To normalize the data of Fig. 7 to the levels of SR-BI-bound HDL, parallel dishes of the same cells were assayed for binding of 125 I-HDL. The normalized data for FC efflux to 10 g/ml HDL protein or 1000 g of phospholipid/ml of SUV were pooled with data from other experiments and are shown in Fig. 6 (D  and E), respectively. Normalized and pooled data for cholesterol oxidase sensitivity are shown in Fig. 6F. These data show that, like A-VI-expressing cells, H-VI-expressing cells showed wild-type activity for FC efflux to low concentrations of HDL (Fig. 6D) and greatly reduced efflux to SUV (Fig. 6E). Moreover, expression of H-VI only slightly increased the plasma membrane pool of FC available for oxidation by cholesterol oxidase (Fig. 6F). These data confirm that the region of the SR-BI extracellular loop just before the C-terminal transmembrane domain is essential for SR-BI to mediate FC efflux to phospholipid vesicles and to enhance the cholesterol oxidasesensitive pool of membrane FC, but is not essential for HDL binding, selective HDL COE uptake, or FC efflux to HDL. DISCUSSION Insertion of epitope tags into the extracellular domain of SR-BI revealed four classes of SR-BI mutants. Class 1 mutants (A-IX and A-X) (Fig. 1), with epitope insertions near the Nterminal transmembrane domain, were not expressed at detectable levels ( Fig. 2A). We have not determined whether the lack of expression reflects low mRNA levels or degradation of the protein by quality control pathways in the endoplasmic reticulum. Class 2 mutants (A-IV and A-VIII) were expressed on the cell surface at ϳ50% of the wild-type levels, but showed no (A-VIII) or very little (A-IV) HDL binding (Fig. 2, C and D). Class 3 mutants (A-II, A-III, A-V, and A-VII), with insertions in the central region of the extracellular domain, were expressed on the cell surface at 25-100% of the wild-type levels and bound HDL to varying degrees, but showed no selective HDL COE uptake (Fig. 2E). When selective HDL COE uptake measurements with Class 3 mutants were normalized to receptorbound HDL, these receptors all showed selective uptake efficiencies less than that seen with CD36, in which case the low efficiency uptake may be due to simply tethering HDL on the cell surface (39). Class 4 mutants (A-VI and H-VI), with epitope insertions in the extracellular domain near the C-terminal transmembrane domain, bound HDL and showed wild-type or modestly reduced selective HDL COE uptake and FC efflux to HDL, but showed no or greatly reduced FC efflux to SUV and failed to increase the plasma membrane pool of FC sensitive to cholesterol oxidase. Class 4 mutants thus separate some cholesterol transport and membrane activities of SR-BI from HDLdependent activities for ligand binding, selective HDL uptake, and FC efflux to HDL.
Mutant A-VI, which has an epitope insertion C-terminal to residue 424, exhibited high affinity HDL binding and mediated selective HDL COE uptake with an efficiency of ϳ60% compared with SR-BI. A-VI also showed ϳ60% of the wild-type activity in mediating cholesterol efflux to HDL, but this deficit was seen mostly at higher concentrations of HDL (Fig. 5A). With HDL at 10 g/ml, FC efflux to HDL was not different between A-VI and SR-BI (Fig. 6D). This is similar to the situation with selective HDL COE uptake, in which A-VI did not show the lower efficiency component at higher HDL concentrations (Fig. 4B). One interpretation of these data is that the lower efficiency components for selective HDL uptake and for FC efflux to HDL may reflect contributions to these processes due to SR-BI-mediated changes in the plasma membrane. In the case of FC efflux, these changes may facilitate cholesterol desorption from the membrane, an effect that would be expected to enhance FC efflux to phospholipid vesicles or to HDL enriched with phospholipid. Previous studies showed that phospholipid enrichment of HDL 3 stimulates SR-BI-mediated FC efflux up to 5-fold without changing the K d for HDL binding to SR-BI (51). The loss of or disproportionate reduction in FC efflux to phospholipid vesicle acceptors versus HDL with mutants A-VI and H-VI may reflect a FC efflux activity of SR-BI that is independent of SR-BI/HDL binding. We hypothesized previously that SR-BI-mediated FC efflux includes two components: one dependent on HDL binding and another independent of HDL binding (5,31). Mutants A-VI and H-VI should prove useful in further testing this hypothesis.
Mutants A-VI and H-VI identify a region of the extracellular loop of SR-BI near the C-terminal transmembrane domain that is important for enhancement of the cholesterol oxidase-sensitive pool of membrane FC and for FC efflux to SUV acceptors. However, these mutations did not affect the ability of the receptor to form productive complexes with HDL and to mediate high efficiency selective uptake of HDL COE and efficient cholesterol efflux to HDL. Thus, A-VI shows a clear separation of function between HDL CE uptake and HDL-dependent FC efflux on one hand and FC efflux to SUV and the oxidasesensitive pool of membrane FC on the other. This novel result has several implications. First, it argues against the idea that all SR-BI activities derive from the ability of SR-BI to load cholesterol into the plasma membrane. Second, it provides clear evidence for a difference in the pathways for SR-BImediated FC efflux to HDL and to phospholipid vesicle acceptors. Whether this difference reflects two very different FC efflux pathways or a divergent step in an otherwise common pathway remains to be tested. Third, the SR-BI-mediated increase in the cholesterol oxidase-sensitive pool of membrane FC suggests that SR-BI alters, in some manner, the distribution of FC within plasma membrane domains. The loss of this activity in A-VI suggests that this is an inherent property of SR-BI that is independent of its ability to facilitate lipid movement between cells and HDL. Additionally, the loss of the cholesterol oxidase-sensitive FC pool and FC efflux to SUV in A-VI and the similar reductions of these two activities in H-VI suggest that these activities may be mechanistically related. Further analysis of these mutants in cell culture should prove valuable in understanding the significance of changes in membrane lipid organization due to SR-BI expression. Additionally, expression of these mutants in vivo may provide insight into the disruption of plasma membrane organization in adrenal glands of SR-BI-deficient mice (52).
Mutant A-III, which has an epitope insertion C-terminal to residue 283, exhibited high affinity HDL binding, but appeared devoid of other SR-BI-mediated activities. This mutant is phenotypically similar, in loss of selective HDL uptake activity, to the N173Q mutant that blocks glycosylation at residue 173 as recently described by Viñ als et al. (49). These mutants, as well as previous studies with apoA-I mutants (53) and apoA-I Ϫ/Ϫ HDL particles (54,55), show that high affinity HDL binding to SR-BI is necessary but not sufficient for efficient lipid transfer. Presumably, apoA-I on the HDL particle surface must be properly positioned or registered with SR-BI to facilitate efficient lipid transfer.
In this study, we attempted to minimize the possibility of major structural disruptions in SR-BI by inserting epitope tags at positions with as little homology as possible between the amino acid sequences of SR-BI, CD36, and lysosomal integral membrane protein II from several different species. Additionally, these sites had a low probability of having ␣-helical or ␤-pleated sheet structure in predicted secondary conformations of SR-BI. Nevertheless, most of the insertions produced major reductions in HDL binding even when the mutants were expressed efficiently on the cell surface, the latter point indicating that protein folding occurred well enough to pass the quality control mechanisms of the endoplasmic reticulum. Furthermore, all but one insertion essentially inactivated the lipid transport functions of SR-BI. These results, as well as the N173Q glycosylation mutant described by Viñ als et al. (49), suggest that SR-BI is very intolerant of changes that may alter protein conformation in the extracellular domain, irrespective of where the change occurs. Additionally, when mutants A-III and A-VI were examined in detail, the reduction in HDL binding activity was due to reduced B max values and not to binding affinity, implying that only a fraction of the cell-surface receptors were competent to bind HDL. The basis for this behavior is currently not known, but might reflect the inability of mutant receptors to cluster properly in plasma membrane domains or the inability to form oligomeric receptor complexes. The mutants identified here will be useful in testing these ideas.