Transfer of Cholesterol from Its Site of Synthesis to the Plasma Membrane*

We have followed the transfer of newly synthesized cholesterol to the plasma membrane in cultured fibro- blasts using cholesterol oxidase as a probe. Since the enzyme has access only to the plasma membrane in intact cells, it permits the discrimination of cell surface and endogenous cholesterol. Cholesterol synthesized from radiolabeled acetate was transferred to the plasma membrane in a strictly first order fashion with a half-time of 1-2 h at 37 OC. The rate of transfer was similar in rapidly growing and confluent cells and was not affected by preincubating the cells in lipoprotein- deficient serum which greatly stimulated cholesterol synthesis. We used equilibrium density gradient cen- trifugation of homogenates from cholesterol oxidase-treated cells to examine further the distribution of newly synthesized cholesterol between cellular pools. We identified membrane fractions enriched in newly synthesized cholesterol yet inaccessible to cholesterol oxidase. The cholesterol in these membranes eventu-ally moved to the plasma membrane. The movement of exogenous radiocholesterol from the plasma membrane to the cell interior also was examined by this method. No detectable transfer was observed over several hours, during which time endogenous cholesterol moved to the plasma membrane. We conclude that the transfer of newly synthesized cholesterol to the plasma membrane is a vectorial process and is not mediated by a simple diffusional equilibrium.


Transfer of Cholesterol from Its Site of Synthesis to the Plasma Membrane*
(Received for publication, May 2, 1984)

Yvonne Lange and Heinrich J. G. Matthies
From the Departments of Pathology and Biochemistry, Rush Medical College, Chicago, Illinois 60612 We have followed the transfer of newly synthesized cholesterol to the plasma membrane in cultured fibroblasts using cholesterol oxidase as a probe. Since the enzyme has access only to the plasma membrane in intact cells, it permits the discrimination of cell surface and endogenous cholesterol. Cholesterol synthesized from radiolabeled acetate was transferred to the plasma membrane in a strictly first order fashion with a half-time of 1-2 h at 37 OC. The rate of transfer was similar in rapidly growing and confluent cells and was not affected by preincubating the cells in lipoproteindeficient serum which greatly stimulated cholesterol synthesis. We used equilibrium density gradient centrifugation of homogenates from cholesterol oxidasetreated cells to examine further the distribution of newly synthesized cholesterol between cellular pools. We identified membrane fractions enriched in newly synthesized cholesterol yet inaccessible to cholesterol oxidase. The cholesterol in these membranes eventually moved to the plasma membrane. The movement of exogenous radiocholesterol from the plasma membrane to the cell interior also was examined by this method. No detectable transfer was observed over several hours, during which time endogenous cholesterol moved to the plasma membrane. We conclude that the transfer of newly synthesized cholesterol to the plasma membrane is a vectorial process and is not mediated by a simple diffusional equilibrium.
While there is considerable information concerning the movement of newly synthesized proteins to the plasma membrane (1,2), the pathway by which lipid is incorporated into the plasm? membrane is not well understood. The transfer of newly synthesized phospholipid to the plasma membrane has been examined recently in Dictyostelium discoideum (3) and cultured Chinese hamster fibroblasts (4); however, different mechanisms appear to be involved in these two systems.
More than 90% of the cholesterol in human fibroblasts and Chinese hamster ovary cells is localized in the plasma membrane ( 5 ) , but the mechanism of this segregation is unknown. Studies of the rate of transfer of cholesterol from its site of synthesis to the plasma membrane have given conflicting results. In a study of radiocholesterol movement between human skin fibroblasts and sonicated liposomes, Poznansky and Czekanski (6) found no movement of endogenously synthesized cholesterol to the cell surface in the course of several hours. However, DeGrella and Simoni (7) showed that the * This work was supported b~ National Institutes of Health Grants HL-28448 and HL-32466. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. movement of newly synthesized cholesterol to the plasma membrane in Chinese hamster ovary cells occurred with a half-time of approximately 10 min.
We recently have developed a technique which permits plasma membrane and endogenous cholesterol pools to be distinguished in intact cells (5). The method is based on the use of the enzyme cholesterol oxidase, which attacks surface cholesterol under certain conditions, leaving internal cholesterol pools untouched. Because of the rapid equilibration of cholesterol across the plasma membrane (5,8), the entire pool of cell surface cholesterol can not only be quantified, but can be physically separated, as cholestenone, from the internal cholesterol. This finding provides an approach to the measurement of the rate at which newly synthesized cholesterol moves to the plasma membrane. Our strategy was to incubate cells in culture with radiolabeled acetate and use cholesterol oxidase to follow the distribution of labeled cholesterol between plasma membrane and internal pools.
Cholesterol oxidase (EC 1.1.3.6; Breuibacterium sp.) was used as obtained from Beckman Instruments. Analysis of the enzyme by gel electrophoresis (9) showed a single band. Thin layer chromatography plates were obtained from Whatman.
Cell Culture-Cultured human fibroblasts were derived from the foreskin of healthy newborns. Cells were grown in monolayer and used between the 5th and 11th passages. Cultures were maintained at 37 "C in a humidified incubator (5% C02) in 75-cm2 or 175-cm2 flasks containing Dulbecco's modified Eagle's minimal essential medium plus 10% fetal calf serum. Twenty-four to forty-eight h prior to the initiation of most experiments, the medium was removed from the flasks, the monolayers of cells were rinsed twice with serum-free medium, and fresh medium containing 5 % (v/v) fetal calf lipoproteindeficient serum prepared as described (10) was added.
Chinese hamster ovary cells were obtained from the American Type Culture Collection and grown in 75-cm2 flasks using the conditions described above for fibroblasts.
Lubeling Cellular Lipids-Radiolabeled acetate (0.075-0.25 mCi) was added in ethanol to the 7-ml or 15-ml medium of each flask and incubated at 37 "C for various times. The medium then was removed and in some experiments the label was chased by the addition of fresh medium containing 10 mM Na acetate for varied intervals. The cells then were washed twice in the flask with phosphate-buffered saline:0.15 M NaCI, 5 mM NaPi @H 7.5). 0.25% trypsin, 0.03% EDTA in phosphate-buffered saline prewarmed to 37 "C were added and the flask was incubated for 5 min at 37 "C to release the cells. After removal from the flask, the cells were washed twice more in phosphate-buffered saline at 0 "C prior to cholesterol oxidase treatment.
Treatment of Cells with Cholesterol Oxidase-Washed cells were suspended in approximately 10 volumes of phosphate-buffered saline and duplicate aliquots were taken for the determination of protein.
The remainder was made 1% in glutaraldehyde and incubated for 10 min on ice. The cells were washed twice in 0.5 mM NaPi, 310 mM sucrose (pH 7.5) and resuspended in this buffer at 10 "c. This pretreatment greatly increases the sensitivity of membrane cholesterol to this enzyme (11). Cholesterol oxidase was added to 8-10 IU/ ml and the cells were incubated for 45 min at 10 "C. The mixture was placed on ice, brought to isotonicity by the addition of 1.5 M NaC1, 5 mM Napi, and extracted with 5 volumes of chloroform:methanol(kl, v/v). "C-labeled cholesterol and cholestenone were added with the organic solvent to permit determination of recovery which averaged 85%.
Density Gradient Fractionation-Radiolabeled cells (approximately 4-6 mg of protein) were washed twice in 0.5 mM NaPi, 310 mM sucrose (pH 7.5) and resuspended to 0.4 ml in this buffer.
Cholesterol oxidase was added to a final concentration of 8-10 IU/ ml and the cells were incubated for 45 min at 10 "C. The suspension was brought to 1 ml in 0.25 M sucrose, 5 mM Napi and the cells were homogenized in a glass-glass coaxial pestle homogenizer using 25-35 strokes. Homogenates were spun for 5 min at 800 X g to remove unbroken cells and nuclei and the resulting supernatants were layered on linear gradients of 19.5%-62% sucrose (w/v) in 5 mM NaPi (pH 8). The gradients were spun in a Beckman SW40 rotor for a minimum of IO8 x g.,.min at 2 "C. Gradients were collected in equal volume fractions from the bottom of the tube. Aliquots of each fraction were taken for refractive index determination and the remainder was extracted for the analysis of radioactivity.
Determination of Radioactivity in Cholesterol and Cholestemm-Lipid extracts were spotted together with standards on Silica Gel G plates. The plates were developed in ch1oroform:methanol (1002, v/ v) and visualized with I2 vapor. The appropriate region of the plate was scraped and the material was eluted twice with 2 ml of chloroform:methanol(2:1). The pooled extracts were dried under Nz, spotted on a reversed-phase plate (12), and developed in acetonitrile/chloroform (40:35, v/v). The radioactivity in the spots was determined by scraping the gel directly into vials to which Aquasol (New England Nuclear) was added for scintillation counting.
Other Methods-Protein was determined by the method of Lowry et al. (13) using bovine serum albumin as standard. Cholesteryl esters were eluted from Silica Gel G plates as described above and saponified by heating for 1 h at 65 "C with ethanolic KOH. Two volumes of hexane and one volume of water were added and the upper phase (sterol fraction) was dried for the determination of radioactivity.

RESULTS AND DISCUSSION
Time Course of Incorporation of PHIAcetate into Cellular Cholesterol-Human foreskin fibroblasts synthesize radiocholesterol from labeled acetate. In agreement with previous reports (14), we found that the incorporation of radiolabeled acetate into cholesterol was dramatically enhanced if cells were preincubated in a medium containing lipoprotein-deficient serum. After 48 h of culture under these conditions, the synthesis of radiocholesterol was essentially linear over a period of several hours (Fig. 1).
Cholesterol oxidase was used to measure the time course of appearance of newly synthesized cholesterol at the plasma membrane. Incorporation into the plasma membrane initially lagged behind total synthesis ( Fig. 1) but after about 2 h the rate of appearance of radiocholesterol at the cell surface became parallel to the rate of synthesis (steady state). These kinetics suggest that essentially all of the newly synthesized cholesterol migrates to the cell surface by a single pathway.
Our data can be analyzed in terms of a simple kinetic model to determine the rate of transfer of cholesterol from its site of synthesis to the plasma membrane. The model is described by were added to each flask and the cells were incubated at 37 "C. At the times indicated, the cells were dissociated by trypsin treatment, washed, and suspended in phosphate-buffered saline at 0 "C. Aliquots were taken for protein determination and the remainder was fixed for 10 min in 1% glutaraldehyde. The cells were washed twice in 310 mM sucrose, 0.5 mM Napi (pH 7.5) and resuspended in this buffer. Cholesterol oxidase was added to 10 XU/ml and the suspension was incubated for 45 min at 10 "C. Lipid extraction and the determination of radioactivity in cholesterol, cholestenone, and cholesteryl esters (negligible in this experiment) were as described under "Experimental Procedures." Incorporation of radiolabel into cellular cholesterol We used also a simpler method for determining k2 which did not depend on the lag and hence did not require determination of complete time courses like that shown in Fig. 1.
Using the formalism given in Ref. 15, the fraction F of cholesterol in the plasma membrane at time t is given by: The per cent of cholesterol oxidized was plotted as a function of time and the value of kz which best fit the data was determined for several experiments (Fig. 2). In four experiments, we found a half-time of (1.0 & 0.1) h. This method had the advantage that variations between flasks in the incorporation of [3H]acetate into cholesterol did not affect the data since the parameter fit was the fraction susceptible to oxidation.  ). In a separate experiment, 6 confluent monolayers of fibroblasts were used. Three of them (0) were preincubated for 48 h in medium containing 5% lipoprotein-deficient serum while the other three were maintained in medium containing 10% fetal calf serum (0). On the day of the experiment, the medium was removed from all the flasks and replaced with fresh medium containing 5% lipoprotein-deficient serum to which the 13H]acetat,e was added. One experiment (0) used confluent monolayers of Chinese hamster ovary cells which were preincubated for 48 h in medium containing 5% lipoprotein-deficient serum prior to the addition of [3H]acetate. The curves were derived from Equation 2 and give the theoretical time courses of oxidation for half-times of 0.9 h (upper curve) and 1.3 h (lower curue).
DeGrella and Simoni (7) studied the transfer of newly synthesized cholesterol to the plasma membrane in Chinese hamster ovary cells using a plasma membrane isolation method. They observed kinetics similar to those we report here; however, the half-time of transfer was only 10 min. This difference does not appear to be species-related since we observed similar time courses in both cell types (Fig. 2).

Rate of Movement of Cholesterol to the Plasma Membrane;
Pulse-Chase Study-Cells in culture flasks were pulsed with [3H]acetate and chased by the addition of 10 mM unlabeled sodium acetate to the medium. At intervals thereafter, the cells were removed from the culture flasks and treated with cholesterol oxidase and the amounts of radioactivity associated with cholesterol and cholestenone were measured. The fraction of radiolabel susceptible to oxidation was found to increase with chase time at 37 "C but not at 4 "C, reflecting the transfer of newly synthesized cholesterol to the plasma membrane (Fig. 3B). The effectiveness of the chase was variable among experiments and was not always complete as seen by the continued incorporation of radiolabel into sterol after the initiation of the chase (Fig. 3A). This phenomenon, which also was observed by DeGrella and Simoni (7), may relate to the prolonged run-off of the numerous intermediates in the complex pathway of sterol biosynthesis. Nevertheless, we observed time courses similar to that shown in Fig. 3B in four experiments.
Careful kinetic analysis (Fig. 4) showed that the transfer in pulse-chase experiments was a strictly first order process with a final value of 95% of radiocholesterol in the plasma mem- chase. It is noteworthy that there was no detectable lag in the transfer process following a 20-min pulse (Fig. 4).
The two methods used to estimate the rate of transfer of newly synthesized cholesterol t,o the plasma membrane yielded results which were in reasonable agreement given the assumptions involved in each of these experiments. Derivation of a half-time from the lag in the appearance of newly synthesized cholesterol at the plasma membrane (Fig. I) depended on the assumption that the amount of [3H]acetate in the cells remained constant throughout the time course. In pulse-chase experiments (Figs. 3 and 4), it was assumed that the chase was totally effective. In both cases it was assumed that there was only one rate-limiting step: the time of transfer of radiolabeled cholesterol to the plasma membrane. The qualitative agreement of the two methods is reassuring although the shorter half-time, l h, is probably the more accurate. We conclude that the half-time of transfer is 1-2 h at 37 "C.
Effect of Cell Metabolism on Transfer of Cholesterol to the Plasma Membrane-The studies of transfer of newly synthesized cholesterol to the plasma membrane described above used confluent cells which had been preincubated for 48 h in lipoprotein-deficient serum to stimulate cholesterol synthesis (14), It also has been reported that cholesterol synthesis in rapidly dividing cells is enhanced relative to that in confluent cultures (16). To determine whether cellular metabolic state influences cholesterol movement, we measured the rate of transfer of newly synthesized cholesterol to the plasma membrane in cells which had not been pre-exposed to lipoprotein deficient serum, as well as in preconfluent monolayers.
Although incorporation of 13H]acetate into cholesterol was stimulated up to 300-fold by preincubating confluent monolayers of cells for 48 h in lipoprotein-deficient serum, the rate of transfer of newly synthesized cholesterol to the plasma membrane was not markedly different in starved cells and in cells maintained in the presence of serum. In a typical experiment (Fig. 2), the half-time of transfer in starved cells was 0.9 h compared to a value of 1.3 h for cells in serum. This difference, while significant, was not greater than the dispersion in half-time values measured in different experiments.
The rate of transfer of newly synthesized cholesterol to the plasma membrane in subconfluent cells was found to be similar to that measured in confluent monolayers ( Table I).
[3H]Acetate was added to the medium of cells which were either confluent (Table I, flasks 1-3) or less than 50% confluent ( Table I, flasks 4-6). The fraction of newly-synthesized cholesterol susceptible to oxidation was determined after incubation of from 1 to 4 h at 37 "C. As can be seen from the

Transfer of newly synthesized cholesterol to the plasma membrane in subcanfluent cultures
On day 1, cells were seeded a t a concentration of 2.5 X lo6 cells (flasks 1-3) or 1.0 X lo6 cells (flasks 4-6) per 75-cm2 flask and grown in medium containing 10% fetal calf serum. On day 2, the medium was replaced with fresh medium containing 5% lipoprotein-deficient serum. Forty-eight h later, when the cells seeded a t a higher concentration were confluent, 75 pCi of [3H]acetate was added to each flask. After incubation of 1-4 h a t 37 "C, the medium was removed from the flasks and the cells were dissociated with trypsin. An aliquot of the washed cells was taken for the determination of protein, and the remainder was glutaraldehyde-fixed, treated with cholesterol oxidase, and assayed as described in the legend to Fig. 1.  , the extent to which newly synthesized cholesterol was oxidized was similar for confluent and preconfluent cells at each time point. We conclude that the transfer of cholesterol to the cell surface is not strongly affected by the metabolic state of the cell but is inhibited at low temperature. Equilibrium Density Gradient Analysis of the Distribution of Newly Synthesized Cholesterol-We sought to examine the subcellular distribution of cholesterol by sucrose density gradient fractionation. This analysis was made difficult by the fact that 95% of cellular cholesterol is in the plasma membrane ( 5 ) and that plasma membrane fragments are broadly distributed in such gradients, thus imposing considerable contamination on the cholesterol-poor internal membrane fractions. However, we found that this contamination can be eliminated by selectively converting plasma membrane cholesterol to cholestenone by cholesterol oxidase treatment of the intact cells prior to homogenization. This system provides a sensitive approach to the study of the distribution of newly synthesized cholesterol in both intracellular and plasma membrane pools.

Flask
We traced the pathway of newly synthesized cholesterol through the cell by labeling cholesterol synthesized at two different times with two different radioisotopes. In one such experiment, cells were incubated with [ 14C]acetate for 2 h, the radiolabel was removed, and [3H]acetate was added for an additional 1-h incubation. The cells then were treated with cholesterol oxidase, homogenized, and spun on a sucrose density gradient. The amounts of each radiolabel in cholesterol and cholestenone were determined throughout the gradient (Fig. 5). At the moment of cholesterol oxidase treatment, most of the cholesterol synthesized from the first added radiolabel (14C) had reached the plasma membrane, as evidenced by its extensive oxidation (panel A). However, only a small fraction of the cholesterol synthesized from the second added radiolabel (3H) had reached the plasma membrane, since most of it was unoxidized (panel B). Furthermore, we observe that the [3H]cholesterol profile in panel B showed a peak in fraction 4 not present in the ['4C]cholesterol profile. This was a reproducible finding which may reflect the presence of intracellular membranes enriched in newly synthesized cholesterol.
Is Exogenous Cholesterol Transferred from the Plasma Membrane to the Cell Interior?-The plasma membrane can be labeled rapidly and selectively by introducing radiocholesterol into the medium (5). This exogenous probe initially was 95-97% susceptible to cholesterol oxidase, documenting its plasma membrane localization (5). The cholesterol oxidase method described above can potentially be used to examine the transfer of this cholesterol into the cell interior. However, since only about 5% of fibroblast cholesterol is intracellular, the disappearance of radiocholesterol from the plasma membrane into internal pools may not be reliably measured by the cholesterol oxidase method against a comparable background of unoxidizable plasma membrane radiolabel.
We approached this problem in two ways. In the first, we employed two different exogenous radiocholesterol probes to correct for incomplete oxidation of the plasma membrane. The strategy was to label the plasma membrane with exogenous [3H]cholesterol and allow several hours to elapse during which time the radiolabel potentially could equilibrate with internal pools. ['4C]Cholesterol then was introduced in a like manner and the cells were immediately treated with cholesterol oxidase to determine the distribution of the two radiolabels between oxidizable and unoxidizable pools. The [l4CC] cholesterol provided an indicator of the plasma membrane cholesterol which was not oxidized, so that the excess of Transfer of exogenous cholesterol from plasma membrane to intracellular pools The medium was removed from three 25-cm2 flasks of confluent cells growing in medium containing 5% lipoprotein-deficient serum and replaced with 3 ml of unsupplemented medium. 0.2 pCi of [3H] cholesterol in 10 pi of ethanol was added to the flask. After a 10-min incubation at room temperature, the medium was removed from the flask and the cell monolayer was rinsed twice with sterile phosphatebuffered saline. Fresh medium was added and the cells were incubated for 0-16 h at 37 "C. Finally, 0.5 pCi of ["C]cholesterol was added to the medium and after a 10-min incubation at room temperature, the cells were dissociated from the flask, treated with cholesterol oxidase, and assayed as described in the legend to Fip, 1. Medium containing 5% lipoprotein-deficient serum was added to a 175-cm2 flask of confluent cells and after a 46-h incubation at 37 "C, 0.1 mCi of ["Clacetate was added to the medium. The cells were incubated for 2 h at 37 "C, the medium was removed, and the monolayer was rinsed. Fresh medium containing 250 NCi of [' HI acetate was added to the flask and the cells were incubated for a further 1 h at 37 "C. The medium was removed, and the monolayer was rinsed and dissociated by trypsin treatment. The cells were treated with cholesterol oxidase, homogenized, subjected to equilibrium density gradient centrifugation, and assayed as described under "Experimental Procedures." Fraction 1 is the bottom of the gradient. The densities of fractions 4 and 8 were 1.22 g/ml and 1.14 g/ml, respectively. Panel A shows the distribution of ["C]cholesterol (0) and [14C]cholestenone (0); panel B shows the distribution of 13H] cholesterol (0) and [3H]cholestenone (0) in the same gradient. unoxidized 3H over 14C would reflect transfer of [3H]cholesterol to internal pools. The results of a representative experiment are given in Table 11. The first added radiocholesterol (3H) was 1-296 more susceptible to oxidation than the second added at all times. We do not take this tiny differential as significant, since it shows no time-dependent decrease as would be expected from a transfer process. Therefore, these data demonstrate that no significant transfer of [3H]cholesterol from the plasma membrane to the cell interior occurred during a 16-h incubation at 37 "C.
The data in Table I1 notwithstanding, it could be that plasma membrane cholesterol does equilibrate with a subfraction of the internaI cholesterol pool. To examine this possibility, the cell plasma membrane was labeled selectively by introducing ['4C]cholesterol into the medium. The label was given the opportunity to equilibrate for several hours. Intracellular cholesterol then was labeled by the brief addition of [3H]acetate followed by a chase with unlabeled acetate. [3H]cholesterol between plasma membrane and intracellular pools Confluent fibroblasts were preincubated for 48 h at 37 "C in medium containing 5% lipoprotein-deficient serum. The cells were labeled with ["C]cholesterol by the addition of the radiolabel in ethanol to the flask for 10 min as described (7). The medium was removed and the cells were incubated for 3 h to let the label equilibrate. Fresh medium containing 200 qCi of [3H]acetate then was added. After a 45-min incubation at 37 "C, the labeled medium was removed and replaced with medium containing 10 mM Na acetate. Following a 1h incubation at 37 "C, the cells were dissociated by trypsin treatment, treated with cholesterol oxidase, homogenized, and spun to equilibrium on a sucrose density gradient as described in the legend to Fig.  5. The radioactivity in cholesterol and cholestenone was determined for each fraction as described under "Experimental Procedures." In these double label experiments, externa1 standards could not be used to correct for recovery after TLC; however, the ratio of the two radiolabels in each spot is precisely determined. Fraction 1 is the most dense fraction. The major peaks in the cholestenone and cholesterol profiles occurred in fractions 6 (1.16 g/ml) and 7 (1.14 g/ml), respectively. There was a small cholesterol peak in fraction 3 (1.22 g/ml) as in Fig. 5 and fractionated by sucrose gradient centrifugation. The use of two radiolabels in these experiments made it possible to determine the ratio of their specific activities in the plasma membrane and cell interior without knowledge of the cholesterol mass in either pool. As shown in Table 111, the ratio of exogenous to endogenous radioactivity ( i e . 14C/3H) in choles-tenone was much higher than in cholesterol in every fraction of the gradient. Thus, the two radiolabels did not equilibrate between the cell surface and any detectable intracellular pool. These data therefore are inconsistent with a cholesterol transfer process involving one-for-one exchange of plasma membrane cholesterol with a substantial pool of intracellular cholesterol.
Esterification of Exogenously Incorporated P4C]Cholesterol-The mingling of plasma membrane cholesterol with intracellular pools also can be examined by measuring its availability for esterification. The intracellular esterification of cholesterol is mediated by acyl-CoAcholesteryl acyltransferase, an enzyme which is associated with the endoplasmic reticulum (17). Poznansky and Czekanski (6) reported that in human skin fibroblasts, exogenously introduced cholesterol was not esterified during a 24-h incubation. In contrast, Slotte and Lundberg (18)  Brown and Goldstein (10) showed that the addition to the medium of 2-15 pglml cholesterol in ethanol suppressed 3hydroxy-3-methylglutaryl coenzyme A reductase activity in cultured fibroblasts during a 24-h incubation, suggesting that exogenous cholesterol entered the cell. This finding must be reconciled with our observation that exogenously added radiocholesterol is transferred slowly, if at all, from the plasma membrane to the cell interior. The explanation may lie in the relatively high concentration of cholesterol added in the study cited (10). Because cholesterol forms aggregates in aqueous solution above about lo-' M (19), it is possible that in these experiments the sterol entered the cell by endocytosis and not by a molecular transfer mechanism. We tested this hypothesis by examining the esterification of exogenous ['4C]cholesterol in the presence of added cholesterol carrier. Unlabeled cholesterol was added to the medium in the culture flask together with ['4C]cholesterol and [3H]acetate and the cells were incubated for 18 h at 37 "C. The addition of 10 pg/ml unlabeled cholesterol to the medium led to a 6-fold increase in the incorporation of the exogenous label into cholesteryl esters (Table IV, experiment 2). A concomitant reduction ofapproximately 40% in the esterification of the synthesized cholesterol was observed as expected (10). These data suggest that above a critical threshold concentration, cholesterol can be internalized by the cells even though trace amounts of radiocholesterol are not transferred from the plasma membrane to intracellular pools.

TABLE IV
Esterijication of exogenous cholesterol For experiment 1-The medium in a 75-cm2 flask of confluent fibroblasts growing in medium containing 10% fetal calf serum was removed and replaced with 7.5 ml of unsupplemented medium. Approximately 0.5 pCi of ['4C]cholesterol was added in 30 pI of ethanol and the flask was incubated for 10 min a t 25 "C. The medium was removed, the cell monolayer was rinsed twice, and fresh medium containing 75 pCi of [3H]acetate was added. After a 1-h incubation at 37 "C, the medium was removed, and the cells were rinsed and then incubated for 1 h at 37 "C in medium containing 10 mM Na acetate. The medium was removed, the cells were dissociated from the flask by trypsin treatment, and the cell lipids were extracted. The radioactivity in cholesterol and cholesteryl esters (sterol moiety) were determined as described under "Experimental Procedures." For experiment 2-Three 75-cm2 flasks of confluent fibroblasts were preincubated for 24 h in medium containing 5% lipoprotein-deficient serum. The cell monolayers were rinsed and 7 ml of medium containing 5% lipoprotein-deficient serum was added. Thirty pl of ethanol was added to flask A and 16 p1 or 30 p1 of a solution of cholesterol in ethanol (2.5 p g / p l ) was added to the other flasks to give cholesterol concentrations of 5 pg/ml (B) and 10 pg/ml (C). Approximately 0.5 pCi of (J4C]cholesterol in 20 pl of ethanol, followed by 10 pCi of [3H] acetate in 2 pl of ethanol were added to each flask. After an 18-h incubation at 37 "C, the medium was removed and the cells were dissociated from the flasks by trypsin treatment, Aliquots were taken for the determination of protein and the remaining cells were extracted for the determination of radioactivity as in Experiment 1.

CONCLUDING COMMENTS
The mechanism by which the striking nonuniformity of cholesterol in mammaIian celI membranes is maintained can now be understood in terms of a simple working hypothesis. Plasma membrane cholesterol is in equilibrium with plasma lipoproteins; indeed, in erythrocytes it appears that membrane cholesterol levels are determined primarily by plasma-unesterified cholesterol (20, 21). Despite the internal synthesis of sterol in tissue cells, it is conceivable that their plasma membrane cholesterol is similarly a function of rapid equilibration with plasma lipoproteins. It has been reported that cholesterol transfer between red cells and plasma lipoproteins involves diffusion of the sterol through the aqueous phase (22, 23). However, the movement of radiolabeled cholesterol between red cells is negligibly slow compared to its transfer to lipoproteins (241, suggesting that it cannot simply be released into the aqueous medium. Furthermore, Wattenberg and Silbert (25) have shown that the cholesterol in various subcellular fractions of LM cells is not in diffusional equilibrium. The data presented here suggest that cholesterol does not diffuse between plasma membranes and internal organelles, in that exogenously introduced radiolabel does not reach internal pools during the time its takes newly synthesized cholesterol to reach the plasma membrane. Rather, cholesterol appears to be transported unidirectionally from its site of synthesis to the plasma membrane with a half-time of 1-2 h at 37 "C.
The mechanism by which newly synthesized cholesterol is transferred to the plasma membrane remains to be elucidated.
Cytoplasmic proteins such as those which promote intermembrane transfer of certain intermediates in cholesterol and steroid biosynthesis (26,27) could be involved. However, we have been unable to detect newly synthesized cholesterol in the soluble protein fraction of cell homogenates. We feel that it is more likely that newly synthesized cholesterol is transported to the plasma membrane by specific membrane vesicles as has recently been suggested for phospholipids (3). If this is the case, our demonstration of first order kinetics and no detectable lag in the appearance of newly synthesized cholesterol at the plasma membrane may signify that the insertion of this cholesterol into transport vesicles is not rate-limiting in the overall process; rather, delivery of the cholesterol to the cell surface by such vesicles could be the slow step. That is, the first order time course suggests that the bulk of intracellular cholesterol constitutes a single pool in rapid equilibrium compared with the unidirectional transfer to the cell surface.
Our experiments using serial additions of ["Clacetate and [3H]acetate to the medium of cells in culture showed that newly synthesized cholesterol was chased from a high density fraction into a lower density fraction (Fig. 5). The high density membranes (-1.22 g/ml) could derive from endoplasmic reticulum since this organelle is the site of cholesterol synthesis (28) but also could be a transport vesicle. It is noteworthy that the low density fraction to which newly synthesized cholesterol moved prior to becoming available for oxidation generally co-migrated with one of the major plasma membrane peaks (Fig. 5). While this juxtaposition could be fortuitous, it also could indicate the association of vesicles containing newly synthesized cholesterol with the plasma membrane prior to actual delivery, perhaps by fusion. Accessibility to cholesterol oxidase in intact cells would signify the incorporation of this new sterol into the plasma membrane.