Cholesteryl Ester Is Transported from Caveolae to Internal Membranes as Part of a Caveolin-Annexin II Lipid-Protein Complex*

We previously demonstrated that in Chinese hamster ovary cells scavenger receptor, class B, type I-dependent selective cholesteryl ester uptake occurs in caveolae. In the present study we hypothesized that cholesteryl ester is transported from caveolae through the cytosol to an internal membrane by a caveolin chaperone complex similar to the one we originally described for the transport of newly synthesized cholesterol. To test this hypothesis we incubated Chinese hamster ovary cells expressing scavenger receptor, class B, type I with [3H]cholesteryl ester-labeled high density lipoprotein, subfractionated the cells and looked for a cytosolic pool of [3H]cholesteryl ester. The radiolabeled sterol initially appeared in the caveolae fraction, then in the cytosol, and finally in the internal membrane fraction. Caveolin IgG precipitated all of the [3H]cholesteryl ester associated with the cytosol. Co-immunoprecipitation studies demonstrated that in the presence of high density lipoprotein, but not low density lipoprotein or lipoprotein-deficient serum, caveolin IgG precipitated four proteins: annexin II, cyclophilin 40, caveolin, and cyclophilin A. Caveolin acylation-deficient mutants were used to demonstrate that acylation of cysteine 133 but not cysteine 143 or 156 is required for annexin II association with caveolin and the rapid transport of cholesteryl esters out of caveolae. We conclude that a caveolin-annexin II lipid-protein complex facilitates the rapid internalization of cholesteryl esters from caveolae.

We previously demonstrated that in Chinese hamster ovary cells scavenger receptor, class B, type I-dependent selective cholesteryl ester uptake occurs in caveolae. In the present study we hypothesized that cholesteryl ester is transported from caveolae through the cytosol to an internal membrane by a caveolin chaperone complex similar to the one we originally described for the transport of newly synthesized cholesterol.

To test this hypothesis we incubated Chinese hamster ovary cells expressing scavenger receptor, class B, type I with [ 3 H]cholesteryl ester-labeled high density lipoprotein, subfractionated the cells and looked for a cytosolic pool of [ 3 H]cholesteryl ester. The radiolabeled sterol initially appeared in the caveolae fraction, then in the cytosol, and finally in the internal membrane fraction. Caveolin IgG precipitated all of the [ 3 H]cholesteryl ester associated with the cytosol. Co-immunoprecipita-
tion studies demonstrated that in the presence of high density lipoprotein, but not low density lipoprotein or lipoprotein-deficient serum, caveolin IgG precipitated four proteins: annexin II, cyclophilin 40, caveolin, and cyclophilin A. Caveolin acylation-deficient mutants were used to demonstrate that acylation of cysteine 133 but not cysteine 143 or 156 is required for annexin II association with caveolin and the rapid transport of cholesteryl esters out of caveolae. We conclude that a caveolin-annexin II lipid-protein complex facilitates the rapid internalization of cholesteryl esters from caveolae.
Caveolae are plasma membrane domains found in most types of cells and are identified biochemically by the presence of a 22-kDa protein called caveolin (1). Caveolin plays a pivotal role in the formation, structural integrity, and function of caveolae (1). Caveolin has multiple functions, but the function of caveolin relevant to the present studies is its role in the trafficking of intracellular sterol. Caveolin can directly bind to cholesterol (2,3), and in an earlier study, we demonstrated that acylation of caveolin was required for the binding of cholesterol to caveolin (4). We speculated that the acylation of caveolin, which occurs adjacent to the hydrophobic membrane domain of caveolin, forms a binding pocket that sequesters cholesterol from the aqueous environment (4). These studies were ex-tended to demonstrate that caveolin is part of a lipid-protein chaperone complex that transports newly synthesized cholesterol from the endoplasmic reticulum directly to caveolae (2). The lipid-protein complex consists of cholesterol, caveolin, heat shock protein 56 (HSP56), 1 cyclophilin 40, and cyclophilin A. Acylation of caveolin at cysteine residues 143 and 156, but not at 133, was required for cholesterol to associate with caveolin (4) and for the assembly of the lipid-protein complex. The lipid-protein complex rapidly (ϳ10 min) transported newly synthesized cholesterol to caveolae where it remained for some time before diffusing throughout the plasma membrane (2) or was effluxed to extracellular acceptors (5,6). Pharmacological disruption of the lipid-protein complex with cyclosporin A or rapamycin prevented the rapid translocation of newly synthesized cholesterol to caveolae and resulted in a net decrease in the mass of cholesterol associated with caveolae over time. The decrease in caveolae cholesterol mass was presumably caused by the diffusion of cholesterol from caveolae to the rest of the plasma membrane without replenishing the cholesterol from intracellular stores.
We recently demonstrated that caveolae in CHO cells are also involved in the selective uptake of exogenous cholesteryl esters from HDL (7,8). Selective uptake refers to the uptake of cholesteryl esters without the internalization and degradation of the entire lipoprotein particle (9). The scavenger receptor, class B, type I (SR-BI) is a physiological receptor for HDL that facilitates the selective uptake of HDL cholesteryl esters (10). Babitt et al. (11) demonstrated that SR-BI is preferentially localized to caveolae in CHO cells, and we recently demonstrated that HDL-derived cholesteryl ester is initially transferred to caveolae (7). In the presence of HDL, caveolae rapidly saturates (ϳ7 min) with cholesteryl ester, and the sterol is not found in other compartments (7). The cholesteryl ester associated with caveolae is reversible, that is, it can be effluxed to extracellular acceptors (7,12). Once the caveolae are saturated with cholesteryl ester, the sterol begins to appear in an intracellular membrane compartment where it is irreversible, that is, it does not efflux to extracellular acceptors. The mechanism of how cholesteryl ester translocates from caveolae to internal membranes is not known.
In the present study we hypothesized that cholesteryl ester is transported from caveolae to internal membranes by a lipidprotein complex similar to the complex that transports newly synthesized cholesterol to caveolae. We demonstrate that caveolin directly binds to cholesteryl ester and that a HDL-de-pendent and SR-BI-dependent intracellular lipid-protein complex is involved in the translocation of cholesteryl esters from caveolae to an intracellular membrane compartment. These data demonstrate 1) a mechanism for the intracellular trafficking of cholesteryl esters, 2) the regulation of caveolin-dependent sterol trafficking, and 3) the requirement of caveolin acylation in the directional transport of sterol.
Isolation and Labeling of Lipoproteins-LDL (d ϭ 1.019 -1.05 g/ml) and HDL (d ϭ 1.063-1.21 g/ml) were isolated from fresh human plasma by density gradient ultracentrifugation as described previously (14). The HDL 3 subfraction (d ϭ 1.13-1.18 g/ml) was isolated from other HDL subfractions by centrifugation. The HDL 3 subfraction was used for all of the experimental treatments described herein. SDS-PAGE and Coomassie staining of the gels was used to check the purity of each lipoprotein fraction. HDL 3 apolipoproteins were then iodinated with iodine monochloride (15) to a specific radioactivity of 400 -600 cpm/ng protein. [ (16). The specific radioactivity of [ 3 H]cholesteryl label in the dually labeled particles ranged between 32 and 35 dpm/ng cholesterol.
Selective Uptake Assays-The selective uptake assays were done as described previously (7,17). The cells were rinsed twice with phosphate-buffered saline (37°C) and then placed in medium (Ham's F-12 or RPMI 1640) containing 5% human lipoprotein-deficient serum (LPDS) and 10 g/ml 125 I-[ 3 H]cholesteryl ester HDL for the indicated times. The assay was terminated by removing the medium and washing the cells four times with Tris saline at 4°C. The total cholesteryl ester associated with the cells was determined by liquid scintillation counting of the organic extracts. The amount of HDL 3 degraded was determined by measuring the amount of non-trichloroacetic acid-precipitable 125 I in the cell medium. All of the experiments using cholesteryl ester were also done with cholesteryl ether, and similar results were obtained (data not shown).
Isolation of Caveolae-Caveolae, cytosol, and internal membranes were isolated as described previously (2,18). This method has been used extensively to characterize caveolae. The procedure does not separate the different internal membranes; thus "internal membranes" refers to endoplasmic reticulum, Golgi, lysosomes, etc. To ensure complete removal of possible contaminating membranes, the cytosol fraction was further purified by centrifugation at 400,000 ϫ g for 1 h.
Immunoisolation-Protein A-Sepharose beads were first blocked by incubating them for 4 h at 4°C with the appropriate cell lysate (200 g/ml) plus 30 mg/ml of bovine serum albumin in 25 mM MES, pH 6.5, 0.15 M NaCl, 1% (v/v) Triton X-100, 60 mM octylglucoside, and 0.1% (w/v) SDS or RIPA buffer (150 mM NaCl, 1.0% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS, 50 mM Tris, pH 8.0). Blocked beads were then used to preclear the experimental fractions that had been adjusted to 1% (v/v) Triton X-100, 60 mM octylglucoside, and 0.1% (w/v) SDS. Precleared fractions were then incubated for 18 h at 4°C with the appropriate antibody (2 g of IgG/sample) before adding blocked protein A-Sepharose beads and incubating an additional 2 h at 4°C. The beads were collected by centrifugation and washed five times in high salt RIPA buffer (500 mM NaCl). Precipitated proteins were detected by immunoblots or silver stains. For the silver staining experiment 30 mg/ml of L-lysine was substituted for the bovine serum albumin.
Electrophoresis and Immunoblots-The samples were concentrated by trichloroacetic acid precipitation and washed in acetone. Pellets were suspended in Buffer C that contained 1.2% (v/v) ␤-mercaptoethanol and heated at 95°C for 3 min before being loaded onto gels. The proteins were separated in a 12.5% SDS-polyacrylamide gel using the method of Laemmli (19). The separated proteins were then transferred to PVDF. The PVDF was blocked in Buffer A that contained 5% dry milk for 1 h at room temperature. The primary antibodies were diluted in Buffer A that contained 1% dry milk and incubated with the PVDF for 1 h at room temperature. The PVDF was washed four times for 10 min each time in Buffer A with 1% dry milk. The secondary antibodies (all conjugated to horseradish peroxidase) were diluted 1:20,000 in Buffer A with 1% dry milk and incubated with the PVDF for 1 h at room temperature. The PVDF was then washed, and the bands were visualized by chemiluminescence.
Gas Chromatography and Mass Spectrometry-Cholesteryl heptadecanoate was added to each caveolin IgG immunoprecipitate to serve as an internal standard, and the samples were then extracted with isopropanol-hexane (20). The extracted lipid was derivatized by suspending the dried lipid in N,O-bis(trimethylsilyl) trifluoroacetamide/trimethylchlorosilane/acetonitrile (89:1:10). The material was heated at 80°C for 5 min, dried, suspended in iso-octane, and used for gas chromatography (Protocol T496125B; Supelco, Sigma-Aldrich). Authentic cholesteryl oleate (Sigma) was dissolved in iso-octane and used as a standard for the retention time of cholesteryl oleate. The samples were injected (splitless) onto an Agilent 6890 GC G2579A system (Agilent, Palo Alto, CA) equipped with a SGE HT5 aluminum clad fused silica capillary column (12 m ϫ 0.32 mm ϫ 0.1 m; Supelco, Bellefonte, PA). The GC temperature program was as follows: the initial temperature was 220°C for 3 min and then increased to 310°C (20°C/min) and then to 400°C (10°C/min) and held for 3.5 min. A model 5973 mass-selective detector (Agilent Technologies, Palo Alto, CA) was used in both scan and selected ion monitoring modes to identify the samples.
Statistical Analysis-Least squares analysis of variance was used to evaluate the data with respect to cell fraction, time, and their interaction using the analysis of variance procedure of SigmaPlot. When appropriate, the fractions were compared within a given time using the Tukey's HSD (honestly significant difference) test. The means were considered different at p Ͻ 0.01.

RESULTS
Cytosolic Pool of Cholesteryl Ester-We previously used CHO cells to demonstrate that SR-BI-dependent selective cholesteryl ester uptake occurs in caveolae and that the sterol is subsequently translocated to an internal membrane compartment; however, we did not elucidate the transport mechanism (7,8,11). In addition, we previously described a novel lipidprotein complex that transports cholesterol from the endoplasmic reticulum through the cytosol to caveolae independent of vesicles (2,4). We hypothesized that a similar lipid-protein complex may be involved in the translocation of cholesteryl esters from caveolae to an intracellular membrane compartment. If this hypothesis is correct than HDL-derived cholesteryl ester should first appear in a caveolae subcellular fraction, followed by the cytosol, and finally internal membranes. To test this, CHO cells expressing SR-BI (7, 11) were incubated with 10 g/ml of [ 3 H]cholesteryl ester or [ 3 H]cholesterol ether-labeled HDL for the indicated times. The cells were then washed and fractionated as described (2, 18). Fig. 1 demonstrates that radiolabeled cholesteryl ester (panel A) and ra-diolabeled cholesterol ether (panel B) initially appeared in the caveolae fraction and that this fraction saturated in ϳ7 min (7). The sterol next appeared in the cytosol fraction, and the amount of radiolabel in this fraction saturated in ϳ15 min. Finally, the sterol appeared in the internal membrane fraction, and this fraction continued to increase linearly for up to 5 h (data not shown).
We next determined whether the cytosolic pool of radiolabeled sterol was associated with caveolin. To do this, CHO cells expressing SR-BI were incubated with [ 3 H]cholesteryl ester or [ 3 H]cholesterol ether-labeled HDL for 30 min, washed, and then processed to isolate cytosol. Equal amounts of cytosol (200 g of protein) were then incubated with IgG for caveolin or an isotype-matched nonspecific IgG. Aliquots of the immunoprecipitate and the resulting supernatant fraction were then used to determine the relative efficiency of the immunoisolation (immunoblots) and for quantifying the amounts of [ 3 H]sterol associated with each fraction. Fig. 2 demonstrates that caveolin IgG quantitatively precipitated caveolin and essentially all of the [ 3 H]cholesteryl ester ( Fig. 2A) and [ 3 H]cholesterol ether (Fig. 2B) found in the starting cytosol fraction (compare Pellet to Cytosol). In contrast, the nonspecific IgG did not precipitate a significant amount of the caveolin or of the [ 3 H]sterol (compare Sup to Cytosol). These data suggest that the cytosolic pool of [ 3 H]sterol is associated with caveolin.
To confirm that the sterol that immunoprecipitated with caveolin was bona fide cholesteryl ester we performed gas chromatography-mass spectrometry analysis on the caveolin immunoprecipitate. CHO cells expressing SR-BI were incubated with HDL for 30 min, washed, and then processed to isolate cytosol. The samples were then prepared as described for Fig. 2, and the immunoprecipitate was extracted with isopropanol-hexane and derivatized with N,O-bis(trimethylsilyl) trifluoroacetamide. Fig. 3 demonstrates that the retention time on the GC column was identical for an authentic cholesteryl oleate standard (Panel A) and the lipid associated with caveolin (Panel C). Four distinctive peaks were generated in the mass spectra of the cholesteryl oleate standard (Panel B). Importantly, the same four peaks were generated from the lipid associated with immunoprecipitated caveolin, demonstrating that cholesteryl oleate is bound to caveolin (Panel D).
Lipid-Protein Complex-We next determined whether the [ 3 H]cholesteryl ester-caveolin complex associated with any other proteins (2,4). The cells were treated with 5% LPDS, 10 g/ml HDL, or 10 g/ml LDL for 30 min and then processed to isolate cytosol. Caveolin along with caveolin-associated proteins were immunoisolated from the cytosol with caveolin IgG, and the precipitated material was resolved by SDS-PAGE and silver-stained to visualize proteins. Fig. 4 demonstrates that four bands were detected in LPDS-and LDL-treated cells at 18, 22, 40, and 56 kDa. Caveolin IgG also precipitated four bands from cells treated with HDL; however, the 56-kDa band was replaced with a 36-kDa band.
To identify the protein bands, the cells were treated as described in the legend to Fig. 4, caveolin and caveolin-associated proteins were immunoisolated, and the precipitates were used for immunoblot analysis. Fig. 5 demonstrates that the 18-, 22-, 40-, and 56-kDa bands that immunoisolated with caveolin IgG from LPDS-or LDL-treated cells were cyclophilin A, caveolin,

FIG. 2. Cholesteryl ester co-immunoprecipitates with caveolin from the cytosol. CHO cells transfected with human SR-BI (7) were incubated with 10 g/ml of [ 3 H]cholesteryl ester or [ 3 H]cholesterol
ether-labeled HDL for 30 min at 37°C. The cytosol was then isolated by centrifugation on a Percoll gradient followed by a 400,000 ϫ g centrifugation to remove any residue-contaminating membranes. Caveolin IgG (2 g/ml) and isotype-matched nonspecific IgG (2 g/ml) were added to equal aliquots of cytosol (200 g). The IgGs were precipitated with protein A-Sepharose, and the pelleted material and the remaining supernatant (Sup) were resolved by SDS-PAGE and immunoblotted for the presence of caveolin. In addition, an aliquot of each fraction was quantified by liquid scintillation spectroscopy to determine the amount of cyclophilin 40, and HSP56, respectively. Cyclophilin A, caveolin, and cyclophilin 40 were also found in the cytosol of HDLtreated cells, but HSP56 no longer co-precipitated with caveolin. The 36-kDa band found in the cytosol of HDL-treated cells was identified as annexin II. Identical results were obtained when IgGs for the different proteins were used for the immunoprecipitations (data not shown).
To determine whether the HDL-dependent association of annexin II with caveolin was SR-BI-dependent, we added 10 g/ml HDL to control CHO cells (not transfected with SR-BI) and cells expressing SR-BI for 30 min. The cytosol was isolated, and caveolin and caveolin-associated proteins were then immunoprecipitated with caveolin IgG or annexin II IgG as described above. The precipitated material was resolved by SDS-PAGE, and the proteins were identified by immunoblot analysis. Fig. 6 demonstrates that HDL did not induce the association of caveo-lin and annexin II in CHO cells not transfected with SR-BI. In contrast, when HDL was added to cells transfected with SR-BI (CHO-SR-BI) caveolin IgG precipitated both caveolin and annexin II, and annexin II IgG precipitated both annexin II and caveolin.
Caveolin Palmitoylation-Caveolin is acylated at three positions (Cys 133 , Cys 143 , and Cys 156 ). We previously demonstrated that acylation of positions 143 and 156, but not position 133, was necessary for the formation of the caveolin-cholesterol chaperone complex (4). To determine whether acylation of caveolin is necessary for the association of annexin II with caveolin, we used our established cell lines (L1210) that express wild-type caveolin, caveolin lacking acylation at either position 133, 143, or 156, or caveolin lacking all three sites of acylation (4). L1210 cells do not endogenously express caveolin but do express SR-BI (4). These cells were treated with 10 FIG. 3. Gas chromatography-mass spectrometry analysis of caveolin-associated lipid. CHO cells transfected with human SR-BI (7) were incubated with 10 g/ml of HDL for 30 min at 37°C. The cytosol was then isolated by centrifugation on a Percoll gradient followed by a 400,000 ϫ g centrifugation to remove any residue-contaminating membranes. Caveolin IgG (2 g/ml) was added to aliquots of cytosol (200 g). The IgGs were precipitated with protein A-Sepharose, and the pelleted material was extracted with isopropanol-hexane and derivatized with N,O-bis(trimethylsilyl) trifluoroacetamide. The sample was then resolved by gas chromatography, and the lipid was identified by mass spectrometry. A, GC retention time with an authentic cholesteryl oleate standard. B, mass spectra of an authentic cholesteryl oleate standard. C, GC retention time of lipids associated with caveolin IgG immunoprecipitates. D, mass spectra of lipids associated with caveolin IgG immunoprecipitates. The data are representative of three independent experiments. g/ml of HDL for 30 min and then processed to isolate cytosol. Caveolin IgG was used to immunoprecipitate putative caveolincontaining lipid-protein complexes. Fig. 7 demonstrates that wild-type caveolin, and acylation-deficient mutants 143 and 156 associated with annexin II in the presence of HDL. In contrast, the triple acylation-deficient caveolin and mutant 133 did not associate with annexin II in the presence of HDL. As previously demonstrated, mutant 133 associated with cyclophilin 40 and cyclophilin A; mutant 143 and the triple acylation mutant did not associate with cyclophilin A; and mutant 156 did not associate with cyclophilin 40 (4).
We next determined whether caveolin acylation was required for the rapid internalization of cholesteryl esters from caveolae. To do this the various caveolin-acylation deficient cell lines were incubated with 10 g/ml [ 3 H]cholesteryl ester-labeled HDL for the indicated times and then fractionated to isolate caveolae, cytosol, and internal membranes (identical to Fig. 1). In cells expressing wild-type caveolin the [ 3 H]cholesteryl ester first appeared in caveolae, followed by cytosol and finally accumulated in internal membranes (Fig. 8).

DISCUSSION
The selective uptake of HDL-derived cholesteryl ester is mediated by SR-BI; however, the mechanism(s) whereby cholesteryl ester is transferred from HDL to the plasma membrane and then subsequently transported to intracellular locations is not understood. In the present study we focused on determining the mechanism for the internalization of cholesteryl ester from the plasma membrane to internal membranes. Because of the limited solubility of cholesteryl esters in aqueous environments, it seems likely that trafficking would occur in a vesicle. However, selective uptake of cholesteryl esters occurs independently of HDL particle uptake and SR-BI internalization (10), suggesting that a classical endocytoic mechanism is not involved. In CHO cells caveolae are the initial sites of cholesteryl ester accumulation in the plasma membrane (7,8,11), which suggests a possible role of caveolae in cholesteryl ester internalization. We previously demonstrated that caveolin can form a lipid-protein chaperone complex and facilitate the direct  (7) were incubated with 5% LPDS, 10 g/ml HDL, or 10 g/ml LDL for 30 min at 37°C. The cytosol was then isolated by centrifugation on a Percoll gradient followed by a 400,000 ϫ g centrifugation to remove any residue-contaminating membranes. Caveolin and caveolin-associated proteins were immunoprecipitated with caveolin IgG (2 g/ml). The precipitate was resolved by SDS-PAGE and silver-stained to visualize proteins. The data are representative of five independent experiments.
FIG. 5. Immunoblot analysis of proteins associated with cytosolic caveolin. CHO cells transfected with human SR-BI (7) were incubated with 5% LPDS, 10 g/ml HDL, or 10 g/ml LDL for 30 min at 37°C. The cytosol was then isolated by centrifugation on a Percoll gradient followed by a 400,000 ϫ g centrifugation to remove any residue-contaminating membranes. Caveolin and caveolin-associated proteins were immunoprecipitated with caveolin IgG (2 g/ml) and resolved by SDS-PAGE. The proteins were transferred to PVDF and immunoblotted with IgGs for HSP56, cyclophilin 40 (Cyp40), annexin II, caveolin-1, and cyclophilin A (CypA). The immunoblots were developed by the method of chemiluminescence (2-min exposures). The data are representative of five independent experiments.
FIG. 6. SR-BI is required for the association of annexin II with caveolin. CHO cells transfected with human SR-BI (7) or control nontransfected CHO cells were incubated with 10 g/ml HDL for 30 min at 37°C. The cytosol was then isolated by centrifugation on a Percoll gradient followed by a 400,000 ϫ g centrifugation to remove any residue-contaminating membranes. Caveolin (Cav) and annexin II IgGs (Anx, 2 g/ml) were used to immunoprecipitate (IP) proteins from the cytosol. The precipitated proteins were resolved by SDS-PAGE, transferred to PVDF, and immunoblotted with IgGs for annexin II and caveolin. The immunoblots were developed by the method of chemiluminescence (2-min exposures). The data are from four independent experiments. FIG. 7. Acylation of caveolin residue 133 is required for the association of annexin II with caveolin. L1210 cells transfected with wild-type caveolin (WT), acylation mutant 133, 143, or 156, or a triple acylation mutant (Triple) were incubated with 10 g/ml HDL for 30 min at 37°C. The cytosol was then isolated by centrifugation on a Percoll gradient followed by a 400,000 ϫ g centrifugation to remove any residue-contaminating membranes. Caveolin and caveolin-associated proteins were immunoprecipitated with caveolin IgG (2 g/ml) and resolved by SDS-PAGE. The proteins were transferred to PVDF and immunoblotted with IgGs for HSP56, cyclophilin 40 (Cyp40), annexin II, caveolin, and cyclophilin A (CypA). The immunoblots were developed by the method of chemiluminescence (2-min exposures). The data are representative of six independent experiments. transport of newly synthesized cholesterol from the endoplasmic reticulum to caveolae (2,4). We hypothesized that the same or a similar lipid-protein complex may transport cholesteryl esters from caveolae to internal membranes. In the present study we have described a novel mechanism whereby caveolin binds to and translocates cholesteryl ester from caveolae to intracellular membranes.
We (2,4,13) and others (21) have demonstrated that caveolin can exist in the cytosol as part of a nonvesicle lipid-protein complex. The present study demonstrates that cholesteryl ester associates with caveolin in a protein complex consisting of cyclophilin 40, cyclophilin A, and annexin II. Annexin II has been associated with lipid rafts and sterol-dependent membrane cycling previously (22)(23)(24). These previous studies and the presence of a 36-kDa band in the silver stain gel lead us to investigate the possibility of the involvement of annexin II in caveolin-dependent cholesteryl ester trafficking. The exact site(s) of interaction between annexin II and caveolin is not known. Annexin II does not contain a classical caveolin-binding motif that mediates the interaction of numerous proteins with caveolin (1). It is possible that annexin II does not interact with caveolin directly but associates with caveolin by binding cholesteryl ester. The mechanism of interaction between these proteins and cholesteryl ester requires further study.
The ability of HDL (outside the cell) to alter the composition of the caveolin-containing cytosolic lipid-protein complex (inside the cell) suggests that HDL somehow generates a signal that modulates the composition of the complex. In addition, this signal is specific for HDL because lipoprotein-deficient serum and LDL did not alter the lipid-protein complex. Although we do not know what this putative signal consists of, we did demonstrate that the ability of HDL to alter the lipidprotein complex is SR-BI-dependent (Fig. 6). At least two mechanisms are possible. The binding of HDL to SR-BI may stimulate a signaling cascade that indirectly alters the lipid-protein complex. Alternatively, the uptake and association of cholesteryl ester with caveolin may somehow directly alter the composition of the lipid-protein complex. Although to date it has not been demonstrated that SR-BI is directly involved in any signaling mechanism, CD36, a similar class B scavenger receptor, has been shown to activate nonreceptor tyrosine kinases (25)(26)(27). The potential role of SR-BI in signaling will require additional study.
The acylation of caveolin is critical for the binding of cholesterol to caveolin and for the formation of a functional cholesterol transport complex (4). We used our established cell systems to determine whether caveolin acylation is involved in the formation of a cholesteryl ester lipid complex and/or the trafficking of cholesteryl ester (Figs. 7 and 8). We previously demonstrated that acylation at positions 143 and 156 in caveolin is essential for the formation of a functional lipid-protein complex that can traffic cholesterol from the endoplasmic reticulum to caveolae (4). The present study demonstrates that acylation of residues 143 and 156 are not required for the association of caveolin and annexin II (Fig. 7) or in the rapid transport of HDL-derived cholesteryl esters to internal membranes (Fig. 8). However, acylation of residue 133 is required for the association of annexin II with caveolin and for the internalization of HDL-derived cholesteryl esters. Interestingly, acylation of residue 133 is not involved in the formation of the cholesterol lipid-protein complex (4). As we previously published, the 143 mutant does not associate with cyclophilin A, and the 156 mutant does not associate with cyclophilin 40 (4); however, both of these mutant caveolins were able to bind annexin II and facilitate the uptake of cholesteryl esters. These data demonstrate that the association of cyclophilin 40 and cyclophilin A with caveolin and annexin II is not required for the uptake of cholesteryl esters.
The current data demonstrate that caveolin associates with cholesteryl ester and annexin II in a lipid-protein complex that appears to be involved in the internalization of cholesteryl esters from caveolae to internal membranes. The present data and previously published data (2,4) suggest that two types of caveolin-containing lipid-protein complexes can exist: one for transporting newly synthesized cholesterol from the endoplasmic reticulum to caveolae (efflux) and one for transporting cholesteryl esters from caveolae to internal membranes (uptake). In addition, the binding of HDL to SR-BI can promote the switch from a cholesterol-transporting complex to a cholesteryl ester-transporting complex, although the mechanism is not known. The cholesterol-transporting complex would most likely decrease cellular sterol concentrations, whereas the cholesteryl ester-transporting complex would most likely increase cellular sterol levels. It is important to emphasize that a caveolin-containing lipid-protein complex is not the only mechanism for the internalization of HDL-derived cholesteryl esters. Tall and co-workers (28) have provided data in hepatocytes, which suggest that HDL-derived cholesteryl es-FIG. 8. Acylation of caveolin residue 133 is required for the rapid internalization of cholesteryl ester to internal membranes. L1210 cells transfected with wild-type caveolin, acylation mutant 133, 143, or 156, or a triple acylation mutant (triple) were incubated with 10 g/ml HDL for the indicated times. The cells were washed and subfractionated to isolate caveolae, cytosol, and internal membranes (see "Experimental Procedures"). The amount of [ 3 H]cholesteryl ester in each fraction was quantified by liquid scintillation spectroscopy. Similar data were obtained with [ 3 H]cholesteryl ether labeled HDL (data not shown). The data are from four independent experiments (means Ϯ S.D.). f, caveolae; OE, cytosol; q, internal membranes. ters are internalized by the endocytosis and retro-endocytosis of HDL particles. In addition, our own data demonstrate that cells that do not contain caveolin can still internalize cholesteryl esters, albeit at a greatly reduced rate (data not shown). Additional studies are required to determine the extent to which caveolin-containing lipid-protein complexes influence cellular sterol metabolism.