Differential Mobilization of Newly Synthesized Cholesterol and Biosynthetic Sterol Precursors from Cells*

Previous work demonstrates that the biosynthetic precursor of cholesterol, desmosterol, is released from cells and that its efflux to high density lipoprotein or phosphatidylcholine vesicles is greater than that of newly synthesized cholesterol (Johnson, W. J., Fischer, R. T., Phillips, M. C., and Rothblat, G. H. (1995) J. Biol. Chem. 270, 25037–25046). Here we report that the release of individual precursor sterols varies with the efflux of newly synthesized zymosterol being greater than that of lathosterol and both exceeding that of newly synthesized cholesterol when using either methyl- (cid:1) -cyclodex-trin or complete serum as acceptors. The transfer of newly synthesized lathosterol to methyl- (cid:1) -cyclodextrin was inhibited by actin polymerization but not by Golgi disassembly whereas that of newly synthesized cholesterol was inhibited by both conditions. Newly synthesized lathosterol associated with cellular detergent-re-sistant membranes more rapidly than newly synthesized cholesterol. Upon efflux to serum, newly synthesized cholesterol precursors associated with both high and low density lipoproteins. Stimulation of the formation of direct endoplasmic reticulum-plasma membrane contacts was accompanied by enhanced efflux of newly synthesized lathosterol but not of newly synthesized cholesterol to serum acceptors. The data indicate that the efflux of cholesterol precursors differs not only from that of cholesterol but also from each other,

Virtually all the organs of the body synthesize cholesterol, with extrahepatic tissues accounting for a significant fraction of whole body sterol production (1). Cholesterol is synthesized from acetyl-CoA via the mevalonate pathway that initially produces farnesyl diphosphate, a precursor for squalene, dolichol, heme a, ubiquinone, and isoprenylated proteins. The committed step in cholesterol synthesis is the cyclization of squalene to lanosterol. From this compound, cholesterol is synthesized in a 19-step process involving the activity of nine different enzymes (2). Recent data indicate that sterols regulate the pathway both at the early (i.e. via hydroxymethylglutaryl-coenzyme A reductase) and postlanosterol steps (3). The late steps of cholesterol synthesis can proceed via lathosterol and 7-dehydrocholesterol or via desmosterol to cholesterol. Interestingly, the relative importance of the two pathways may shift in vivo, e.g. during aging (4).
Cholesterol biosynthesis is critically important for human development and cannot be compensated for by increasing the uptake of cholesterol from exogenous sources. This is exemplified by an increasing number of inborn errors of metabolism that are attributed to mutations in cholesterol biosynthetic enzymes (5). The prototype of these disorders, Smith-Lemli-Opitz syndrome (SLOS) is caused by deficiency of 7-dehydrocholesterol reductase, the last step in cholesterol synthesis via the lathosterol pathway. More recently, other multiple malformation/mental retardation syndromes, including lathosterolosis, have been characterized (6). In these patients, cholesterol precursors may constitute up to ϳ10% of total cellular and plasma sterols. Importantly, cholesterol precursors are also found in normal human plasma, at concentrations roughly 1:1000 of that of cholesterol (7,8). The plasma levels of lathosterol and desmosterol are commonly used as measures of the cholesterol biosynthetic activity of the individual. Recent data suggest that these values are highly heritable (9,10) and could potentially be used to predict individual responsiveness to the cholesterol-lowering regimen (11,12).
Cholesterol biosynthetic enzymes are localized in the cytosol as well as rough and smooth endoplasmic reticulum (ER), 1 both the rate-limiting and the last enzyme of the pathway (hydroxymethylglutaryl-CoA reductase and 7-dehydrocholesterol reductase, respectively) being integral membrane proteins of the ER (13)(14)(15). Several steps of the pathway also occur in peroxisomes. However, the absence of functional peroxisomes does not lead to deficiency of cholesterol biosynthetic enzymes (16). The transfer of cholesterol from its site of synthesis in the ER to the plasma membrane and extracellular acceptors has been investigated in a number of studies (for reviews see Refs. 17 and 18). Instead, the transfer of sterol precursors has so far received little attention despite the pioneering observations by Lange et al. (19) and Johnson and co-workers (20,21) that indicate clear differences in the behavior of cholesterol and its biosynthetic precursors.
Lange and co-workers (22,23) reported that in fibroblasts at least three cholesterol precursors, lanosterol, zymosterol, and 7-dehydrocholesterol were highly concentrated in the plasma membrane. Moreover, newly synthesized zymosterol was found to move to the plasma membrane faster than cholesterol, with a half-time of 9 min (that of cholesterol being 18 min). In contrast, in McA-RH7777 cells the rate of transport of newly synthesized desmosterol was found to be roughly equal to that of cholesterol, with a half-time of ϳ30 min for cholesterol and ϳ40 min for desmosterol (24). Lange et al. (19) further reported that plasma membrane zymosterol turned over rapidly by internalization and became converted to cholesterol. On the other hand, Johnson et al. (20) showed that sterol precursors were not only enriched in the plasma membrane but were rapidly effluxed from cells to high density lipoprotein and phosphatidylcholine vesicles. The major biosynthetic sterol released from Chinese hamster ovary cells was reported to be desmosterol or a closely related sterol. The rapid efflux of biosynthetic desmosterol was attributed to its more efficient desorption from the plasma membrane rather than to its more efficient delivery to the plasma membrane (21).
In the present work, the synthesis, intracellular partitioning, and cellular release of cholesterol and its select precursors were further studied. The efflux of sterols to both methyl-␤cyclodextrin and to serum was analyzed. We report that the faster efflux compared with newly synthesized cholesterol is observed for lathosterol, the major sterol precursor in the circulation, but in particular for zymosterol. Moreover, we provide evidence suggesting that the faster efflux of lathosterol is coupled to a cellular circuit different from that of newly synthesized cholesterol and that its release to physiological acceptors can be modulated differently from that of newly synthesized cholesterol. glutamine, 100 units/ml penicillin, 100 g/ml streptomycin, (experiment medium) for 5 or 15 min at 37°C, followed by chasing in experiment medium containing 10 M lovastatin and 25 mM mevalonate at 37°C. For cyclodextrin extractions, the cells were incubated with 5 mM methyl-␤ -cyclodextrin during the last 5 min of chase in MEM supplemented with 10 mM Hepes pH 7.4, 0.35 g/liter NaHCO 3 , 2 mM glutamine, 100 units/ml penicillin, 100 g/ml streptomycin (air medium) on a shaking water bath at 37°C. The medium was collected, and the cells were scraped into ice-cold PBS, harvested by centrifugation, and resuspended in 2% NaCl. Aliquots of the medium and the cell suspension were analyzed by liquid scintillation counting to determine the [ 14 C]cholesterol content. The lipids from the cells and the medium were extracted and analyzed by thin layer chromatography (TLC), silver ion HPLC, and liquid scintillation counting as previously described (27). Where indicated, jasplakinolide was added to the experiment medium to a final concentration of 3 M (from a 3 mM stock in methanol) for 1 h and BFA to a final concentration of 5 g/ml (from a 5 mg/ml stock in ethanol) for 15 min before [ 3 H]acetate labeling. The labeling, chase, and cyclodextrin extraction were performed in the continued presence of the drugs.

Materials-Media
Efflux of Cholesterol and Biosynthetic Sterol Intermediates to Serum-Cells were grown on 55-mm dishes in LPDS-containing medium supplemented with [ 14 C]cholesterol (20 nCi/ml) for 48 h. The cells were then washed with PBS, labeled with [ 3 H]acetate in experiment medium (250 Ci/ml) for 5 or 15 min at 37°C and chased for 30 min to 4 h in experiment medium containing 20% human serum, 10 M lovastatin, and 25 mM mevalonate at 37°C. The medium and the cells were collected as above, and aliquots of both were analyzed by liquid scintillation counting to determine the [ 14 C]cholesterol content. Lipids from the medium and cells were extracted and separated by TLC and HPLC. Notably, as various [ 3 H]acetate-derived cellular products were released to serum, the medium was analyzed by TLC prior to HPLC. The procedural losses were corrected for based on the recovery of the [ 14 C]cholesterol label as in Ref. 27. Where indicated, the serum-containing chase medium was supplemented with 0.8-m diameter latex beads (1:10 dilution of a 10% bead suspension).
Analysis of Albumin Secretion-Confluent HuH7 cultures in 25-mm diameter wells were preincubated with Met-and Cys-free MEM supplemented with 10 mM Hepes pH 7.4, 0.35 g/liter NaHCO 3 , 2 mM glutamine, 100 units/ml penicillin, 100 g/ml streptomycin for 1 h at 37°C. The cells were then pulse-labeled for 10 min with [ 35 S]Met/Cys labeling mixture (100 Ci/ml). Chase was performed in 1 ml of serumfree culture medium containing 10-fold excess of unlabeled Met and Cys and 20 g/ml cycloheximide. Where indicated jasplakinolide (3 M) was present during the preincubation, pulse and chase. The dishes were then placed on ice, and the medium was collected. The cells were washed with PBS, lysed in 50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Triton X-100, 5 mM EDTA, and 25 g/ml each of chymostatin, leupeptin, antipain, and pepstatin, and insoluble material was removed by centrifugation. The medium and cell lysates were then incubated with anti-albumin antibodies for 16 h at 4°C. The immunocomplexes were captured by protein A-Sepharose (2 h at 4°C), and the bound material was washed five times with 10 mM Tris-HCl, pH 7.4, 0.1% SDS, 0.1% Triton X-100, and 2 mM EDTA. The proteins were boiled in reducing Laemmli sample buffer, resolved by SDS-PAGE (8% gels), and the albumin bands quantitated by Fujifilm BAS-1500 Imaging system. Triton X-100 Extraction-HuH7 cells were labeled with [ 14 C]cholesterol for 48 h and with [ 3 H]acetate for 5 min, and chased for 5-60 min as described above. The cells were washed and scraped in ice-cold PBS, harvested by centrifugation, resuspended in 1% Triton X-100-containing buffer on ice, and fractionated in 0 -40% Optiprep flotation gradient in the presence of 1% Triton X-100 as described previously (27). Six fractions were collected from the top, and the lipids extracted and analyzed as above.
Size-exclusion Chromatography-Serum lipoproteins were fractionated using Superose 6HR column (Amersham Biosciences) with PBS as elution buffer. The flow rate was 0.5 ml/min and 1-min fractions were collected. One-tenth of each fraction was used to determine the [ 3 H]acetate-derived and [ 14 C]cholesterol radioactivity. The major peaks of [ 3 H]acetate-derived radioctivity were pooled, and lipids extracted and analyzed by TLC followed by HPLC as above to resolve the biosynthetic sterols. Procedural losses were corrected based on the recovery of Other Methods-Protein concentrations were measured according to Lowry (28). Statistical significance of differences was determined using the Student's t test.

Analysis of Cholesterol and Its Biosynthetic Precursors in
Cell Lines-Newly synthesized cholesterol is often resolved only by TLC although this method is inadequate to separate cholesterol from its biosynthetic precursors as reported in several studies (20,23,27,29,30). To evaluate the resolving power of TLC, we pulse-labeled sterol-starved BHK cells with Prolonged incubation at 14°C during [ 3 H]acetate labeling has been used to accumulate newly synthesized cholesterol intracellularly prior to chasing at 37°C (26,31). Therefore, the fraction of cholesterol present in the TLC spot was also monitored under these conditions. At chase times under 30 min when cholesterol was postulated to undergo rapid movement, less than 50% of the TLC spot represented [ 3 H]cholesterol (Fig.  1b). According to HPLC analysis, [ 3 H]lathosterol was one of the major [ 3 H]acetate-derived products co-migrating in the TLC spot, representing 35% of the dpms analyzed at 0 -10 min and 25% at 30 min of chase (data not shown). The result suggests that the use of TLC alone could yield misleading results and reinforces the necessity of using methods with high resolving ability for accurate separation of cellular sterols.
Next, the levels of [ 3 H]acetate-derived newly synthesized cholesterol and its biosynthetic precursor sterols were measured by HPLC from fibroblastic (BHK and NIH3T3) and hepatic (HuH7) cell lines. Prior to [ 3 H]acetate labeling, the cells were sterol-starved by culturing for 2 days in lipoprotein-deficient medium. The cells were pulse-labeled with [ 3 H]acetate for 15 min and chased for 30 or 60 min. Both the rate and efficiency of [ 3 H]acetate incorporation into cholesterol varied considerably between similarly cultured cells (Fig. 2). HuH7 and NIH3T3 cells produced [ 3 H]cholesterol more efficiently than BHK cells that were slower in synthesizing cholesterol and contained larger fractions of the precursor sterols. In BHK cells, [ 3 H]zymosterol represented the major sterol peak by HPLC analysis both at 30 and 60 min of chase (Fig. 2, a and b). BHK cells also contained significant levels of both [ 3 H]lathosterol and [ 3 H]desmosterol, whereas in HuH7 and NIH3T3 cells, lathosterol represented the major precursor sterol, and only minor amounts of desmosterol were detected (Fig. 2b) hepatic cell line, HepG2, with cholesterol as the main biosynthetic sterol product (20). On the other hand, some fibroblastic cells synthesize cholesterol efficiently while others do not, as exemplified by NIH3T3 and BHK cells, respectively.
Efflux of Cholesterol and Its Biosynthetic Precursors to Methyl-␤-cyclodextrin-To examine the efflux of newly synthesized sterols, the release of [ 3 H]acetate-derived sterols to methyl-␤cyclodextrin was analyzed from BHK and HuH7 cells after increasing chase times. For HuH7 cells, the [ 3 H]acetate labeling was shortened to 5 min to increase the proportion of radiolabeled precursor sterols. Methyl-␤-cyclodextrin was added to the cells for 5 min at 37°C either immediately after the labeling or after increasing chase times in serum-free medium (in the presence of statin and unlabeled mevalonate). The cyclodextrin concentration was titrated such that the efflux of prelabeled cellular [ 14 C]cholesterol was ϳ25%. The efflux of [ 3 H]cholesterol increased gradually with increasing chase times in both cell lines, being somewhat faster in HuH7 than in BHK cells as shown previously (27) (Fig. 3a). In both cell lines, the efflux of newly synthesized [ 3 H]lathosterol was greater than that of newly synthesized [ 3 H]cholesterol at all time points analyzed (Fig. 3a). In BHK cells, the major precursor [ 3 H]zymosterol was very efficiently recovered by cyclodextrin and was virtually absent from the cells, yielding a very high efflux percentage (82.4 Ϯ 0.3% at 5 min, 80.5 Ϯ 1.0% at 60 min, and 52.8 Ϯ 2.39% at 120 min of chase). Accordingly, when the dpms of 3 H-labeled sterols in the medium were plotted in BHK cells the overwhelming majority represented zymosterol at short chase times (5-60 min) (Fig. 3b) (Fig. 3a). The more efficient completion of cholesterol biosynthesis in HuH7 cells was paralleled by the release of fewer precursors and correspondingly more of newly synthesized cholesterol to the acceptor (Fig. 3b).
We have previously shown that the efflux of newly synthesized cholesterol to cyclodextrin is moderately inhibited in both BHK and HuH7 cells by BFA (27). However, we now observed that BFA had no effect on the cyclodextrin availability of newly synthesized lathosterol under the same conditions (Fig. 4, a  and b). In search of additional modulators of newly synthesized sterol efflux we tested the effect of a membrane-permeant promoter of actin polymerization, the marine sponge toxin, jasplakinolide. We found that this compound inhibited slightly but reproducibly the efflux of both newly synthesized cholesterol and lathosterol (Fig. 4, a and b). Interestingly, for newly synthesized cholesterol the effect was apparently additive with that of BFA, suggesting that jasplakinolide affected a Golgibypass route of cholesterol transport (Fig. 4a). This was also in line with the observation that the jasplakinolide treatment had no effect on albumin secretion from the cells (Fig. 4c). The combination of BFA and jasplakinolide was not significantly more effective than jasplakinolide alone in inhibiting the efflux of lathosterol (Fig. 4b). Similar inhibition by jasplakinolide on the efflux of newly synthesized lathosterol to methyl-␤-cyclodextrin was observed in BHK cells (data not shown).
Association of Newly Synthesized Lathosterol with Detergentresistant Membranes-Next, the association of newly synthesized lathosterol and cholesterol with detergent-resistant membrane fractions (DRMs) was compared. We have earlier shown that newly synthesized cholesterol was initially found in Triton X-100 soluble membranes but upon chasing, gradually associated with DRMs, kinetically closely paralleling its availability for efflux to cyclodextrin (27). We now observed that newly synthesized lathosterol acquired detergent resistance more rapidly than newly synthesized cholesterol in the same cells, with 35-40% found in DRMs already at 5 min of chase while at that time point, only ϳ20% of newly synthesized cholesterol was detergent resistant (Fig. 5). At 15 min of chase, the difference between the detergent resistance of newly synthesized lathosterol and cholesterol was still considerable (50 -55% and 30 -35% in DRMs, respectively; Fig. 5c) but started to level off at longer chase times (Fig. 5b). At 1-2 h of chase, the Triton X-100 solubility of newly synthesized lathosterol and cholesterol were closely similar ( Fig. 5b and data not shown) suggesting that the differences in the phase behavior of the two lipids were not due solely to their structural differences. Rather, the differential, time-dependent partitioning between the detergent soluble and resistant phases could reflect differential cellular itineraries of the biosynthetic sterols.
Efflux of Cholesterol and Its Biosynthetic Precursors to Serum-Next, we analyzed the release of newly synthesized cholesterol and sterol precursors to physiological acceptors. Complete serum was used as the acceptor because efflux to serum is more efficient than to isolated particles, such as HDL or apolipoprotein A-I. In this system, HuH7 or BHK cells prelabeled for 48 h with [ 14 C]cholesterol and thereafter pulse-labeled with [ 3 H]acetate for 5 or 15 min, were incubated with increasing chase times in the presence of 20% serum containing lovastatin and mevalonate. Lipids from the cells and medium were extracted and analyzed by TLC followed by HPLC.
As expected, an increasing fraction of cellular [ 14 C]cholesterol was released to serum with increasing chase time (Fig.  6a). The shortest efflux time at which newly synthesized sterols could reliably be detected from the efflux medium was 30 min (Fig. 6a) (Fig. 6a). In HuH7 cells the amount of [ 3 H]zymosterol was negligible, and its efflux could not be reliably measured. The same was true for [ 3 H]desmosterol in both cell types (data not shown). Evidently, the high efflux percentages for [ 3 H]lathosterol and in particular [ 3 H]zymosterol, reflect efficient removal of the sterol precursors from the cells. However, the proportion of newly synthesized cholesterol in the medium [ 3 H]-labeled sterols increased steadily so that at 2-4 of chase, cholesterol was the major newly synthesized sterol effluxed from the cells (Fig. 6b).
To identify which lipoprotein fractions newly synthesized cholesterol, zymosterol, and lathosterol associated with, cells labeled with [ 14 C]cholesterol and [ 3 H]acetate as above were incubated with 20% serum-containing medium for 2 h. The efflux medium was then analyzed by gel filtration on a Superose 6HR column (32). The experiment was carried out in BHK cells because of the small amount of radiolabeled precursors effluxed from HuH7 cells. Moreover, lipoprotein assembly and secretion in hepatic cells might complicate the interpretation of the data. The elution profiles of the prelabeled [ 14 C]cholesterol, 3 H-radiolabeled products, and elution positions of the major serum lipoproteins are shown in Fig. 7a. The major peaks of [ 14 C]cholesterol radioactivity appeared in fractions that corresponded to the positions of LDL and HDL. To resolve the [ 3 H]acetate-derived sterols, the fractions corresponding to LDL and HDL and the major 3 H radioactivity peak (that eluted after the smallest standard of 1,300 Da) were analyzed by HPLC. This revealed that cholesterol was the major newly synthesized sterol associated with both LDL and HDL but that substantial proportions of newly synthesized lathosterol and zymosterol were also detected in both lipoprotein fractions (Fig. 7, b and c). Although the overwhelming majority of [ 3 H] radioactivity was recovered in the non-resolving end volume of the column no newly synthesized sterols were found in these fractions (Fig.  7d). All of the three sterols were more enriched in LDL, suggesting that similarly to cholesterol they may be initially ac- quired by HDL and then transferred to LDL (Fig. 7e). The proportions of esterified newly released cellular sterols at 2 h of incubation were negligible (data not shown).

Promotion of ER-Plasma Membrane Contacts Is Associated with Enhanced Transfer of Newly Synthesized Lathosterol to
Serum Acceptors-Considering the rapid mobilization of cholesterol precursors from cells, we speculated whether direct ER-plasma membrane contacts could facilitate this movement. It has recently been shown that ER can fuse with the plasma membrane to provide a source of membrane for the uptake of foreign material by phagocytosis (33). Within 15 min of feeding cells with inert particles (latex beads), ER chaperones such as GRP94, BiP, PDI, and calreticulin are redistributed to the phagocytic cup that forms as a specialization of the plasma membrane. Upon phagosome maturation, successive waves of ER become associated with its membrane. Although professional phagocytes, such as monocytemacrophages and polymorphonuclear granulocytes, are most efficient in engulfing foreign particles the process occurs in a wide variety of eukaryotic cells. Moreover, phagosomes from BHK cells have also been shown to contain ER proteins (34).
We therefore tested whether incubation of BHK cells with latex beads would affect the release of newly synthesized sterols from the cells to serum acceptors. Cells prelabeled with [ 14 C]cholesterol and pulse-labeled with [ 3 H]acetate were incubated for 60 min in 20% serum-containing medium supplemented with 0.8-m diameter latex beads in near saturating conditions (1:10 dilution of a 10% bead suspension). The efflux medium was collected and cholesterol and its precursors analyzed from the cells and medium as above. Surprisingly, we found that while the efflux of newly synthesized [ 3 H]cholesterol was not affected upon incubation with the beads, the efflux of [ 3 H]lathosterol was enhanced by ϳ25% (Fig. 8). Furthermore, although newly synthesized zymosterol was very efficiently effluxed already in the absence of beads (see Fig. 6) the amount of [ 3 H]zymosterol in the serum further increased by 26.9 Ϯ 7.7% upon the phagocytic stimulus. Instead, the efflux of prelabeled [ 14 C]cholesterol was not affected (data not shown).  Fig. 6 were chased for 2 h in the presence of 20% serum. The medium was collected, concentrated in a Centricon-10 concentrator, and analyzed by Superose 6HR gel-filtration. One-tenth of each fraction was used to determine the [ 3 H]acetate-derived and [ 14 C]cholesterol radioactivity (a). The peaks corresponding to LDL, HDL, and V t were pooled, lipids extracted, and analyzed by TLC followed by HPLC to resolve the biosynthetic sterols (b-d).

DISCUSSION
In this study, the separation of individual cholesterol precursors from each other and from cholesterol enabled us to compare their ratio and characteristics within cells as well as their release to extracellular acceptors. These analyses revealed firstly, that the efficiency of cholesterol synthesis from acetate varied significantly between cell lines, with BHK cells containing severalfold more of sterol precursors than cholesterol even after 1 h of acetate labeling. Secondly, the cellular partitioning of sterol precursors may differ from that of cholesterol immediately postsynthesis. This was suggested by the more rapid recovery of newly synthesized lathosterol from membranes resisting detergent solubilization. The finding implies that lathosterol moves more rapidly than cholesterol within the cell to reach sites enriched in cholesterol-sphingolipid-rich membrane domains. Thirdly, the release of newly synthesized sterol precursors to both methyl-␤-cyclodextrin and serum was greater than that of newly synthesized cholesterol. Moreover, there were significant differences between individual precursors, with zymosterol efflux being much greater than that of lathosterol. This correlates with sterol hydrophilicity, with the more polar zymosterol (due to its additional double bond) being more readily effluxed. Finally, the efflux of newly synthesized sterol precursors could be modulated differently from that of newly synthesized cholesterol.
Our findings emphasize the necessity to accurately separate structurally different sterols when studying the fate of cholesterol immediately following its synthesis. Although the actual amounts of cholesterol precursors present in the cells and in the circulation are small their proportion of the newly synthesized sterol pool may be substantial. As cholesterol precursors apparently move about more freely than cholesterol, pooling of data could result in the overestimation of the transport rate of cholesterol and in misinterpretations regarding the insensitivity of cholesterol to perturbations that do affect the movement of cholesterol but not that of its abundant precursor sterols.
The extractability of newly synthesized cholesterol and sterol precursors varied considerably when cells were rapidly incubated with methyl-␤-cyclodextrin after different periods of chasing. We have previously used cyclodextrin extractability of newly synthesized cholesterol as a measure of its plasma membrane arrival (27). Evidently, this cannot be directly used to compare the plasma membrane arrival of cholesterol and its precursors because the differential desorption of sterols from the membrane (21) as well as differential affinity toward cyclo-dextrin may contribute to the process. Yet, the time-dependent changes in the cyclodextrin extractability of each sterol reveal interesting differences, with newly synthesized lathosterol and desmosterol behaving more similarly to cholesterol than zymosterol (Fig. 3). The increasing release of newly synthesized lathosterol, desmosterol, and cholesterol to cyclodextrin presumably reflects their increasing appearance in efflux accessible membrane domains. On the other hand, zymosterol efflux to cyclodextrin was initially very high but gradually decreased. This could derive from the combination of inefficient cholesterol synthesis, rapid zymosterol desorption from the membrane and its eventual, relatively slow conversion to other sterols. The data would be in accordance with the rapid delivery of zymosterol to the plasma membrane and slow metabolism to cholesterol as reported by Lange et al. (19).
The use of serum as sterol acceptor has several important distinctions compared with methyl-␤-cyclodextrin. The efflux is slower and consists of both diffusion and apolipoprotein-mediated components (35). In addition, serum incubation is known to activate physiological signaling and trafficking pathways implicated in the efflux process. The preferential efflux of sterol precursors compared with cholesterol also applied when using serum as an acceptor. However, serum incubation enhanced the efflux of newly synthesized cholesterol relative to the precursors especially in BHK cells. This is observed when comparing the ratios of radiolabeled sterols in the medium at the same chase time following either cyclodextrin extraction or serum incubation (compare Fig. 3b and Fig. 6b). The finding supports the idea that the efflux of newly synthesized cholesterol could be regulated by HDL or other biological acceptors (20).
Consistently with the enhancement of cholesterol efflux by serum, newly synthesized cholesterol was also more abundant than the precursors lathosterol and zymosterol in the major lipoprotein acceptors. According to the current view, small pre-␤ migrating HDL particles are likely to serve as the initial acceptors followed by further lipidation of HDL and gradual transfer of the sterols to LDL (36). As the majority of sterols recovered from LDL were free, i.e. unesterified sterols the initial transfer did not seem to require the actions of the plasma lipoprotein modifying enzymes lecithin:cholesteryl acyl transferase and cholesteryl ester transfer protein. The low levels of cholesterol precursors in normal human serum compared with the levels that appear to be produced by tissues probably reflect efficient secretion of precursors into the bile as well as their conversion to cholesterol (7,37).
Earlier studies have shown that the transport of newly synthesized cholesterol is relatively resistant to Golgi disassembly (27,38,39). However, the molecular machineries operating in this Golgi-bypass route have remained elusive. In the present work, we found that pharmacologically induced actin polymerization inhibited the release of newly synthesized cholesterol to methyl-␤-cyclodextrin, and that this effect was additive with the inhibitory effect of BFA. Although quantitatively minor, this efflux inhibition provides the first indication for a potential involvement of the actin cytoskeleton in the Golgi-bypass route of cholesterol transfer. Moreover, actin may also play a role in the mobilization of sterol precursors as similar inhibition was observed for lathosterol efflux upon enhanced actin polymerization. Whether newly synthesized cholesterol and lathosterol use the same Golgi-bypass mechanism remains open. The more rapid association of lathosterol with DRMs and the enhancement of lathosterol but not cholesterol efflux upon stimulation of phagocytic uptake would nonetheless argue for a more restricted mobility of cholesterol.
In using latex bead internalization as a means to promote ER-plasma membrane contacts we cannot rule out the possi- bility that the enhanced lathosterol efflux is due to some other reorganization in the cell upon phagocytic uptake. However, direct fusion of ER subdomains with the plasma membrane could potentially provide a rapid means to transfer precursor sterols between membranes, allowing them to bypass the cytoplasmic environment. Newly synthesized cholesterol is thought to be delivered to the plasma membrane by vesicular transport (38,39) and may not be able to employ such a pathway. Interestingly, Patterson et al. (40) found that jasplakinolide-induced redistribution of F-actin into a tight cortical layer subjacent to the plasma membrane prevented coupling between ER and plasma membrane Ca 2ϩ entry channels. It could be speculated that actin-dependent ER-plasma membrane transfer may also operate in the case of some of the sterol precursors.
An intriguing question raised by earlier studies and reinforced by our work is the apparent "leakiness" of the cholesterol biosynthetic system. The rapid cellular transfer and efficient desorption of sterol intermediates facilitate their removal to lipoprotein acceptors. However, this seems poorly compatible with the complexity and high energy expenditure of cholesterol biosynthesis. It is possible that delivery of biosynthetic sterol intermediates to the liver for catabolism or further conversion to cholesterol constitutes a mechanism to prevent their pathological accumulation in peripheral tissues (21). Indeed, the variable manifestations in the inborn errors of cholesterol biosynthesis are probably partly due to the harmful effects of the accumulating precursors and not solely to the deficiency of the end product (5). There is also evidence that some of the sterols function as precursors for other metabolites in specific tissues, an example being the conversion of 7-dehydrocholesterol via cholecalciferol to vitamin D in the skin. Finally, considering the short half-life of precursor sterols in serum it could be envisaged that specific precursors could have local functions, e.g. in paracrine signaling. The action of dimethyl-zymosterol as a meiosis-activating sterol in oocytes and spermatozoa apparently represents such a function (41). Considering the increasing appreciation of the sterol fine structure in determining its biological effects and the potential to quantitatively analyze specific sterols by HPLC, gas chromatography, and mass spectrometry, greater understanding of the actions of cholesterol precursors should be anticipated.