INTRODUCTION

The endoplasmic reticulum (ER)¹ is the site of synthesis, folding, and modification for proteins which transit the secretory pathway or are residents of the secretory organelles themselves (reviewed in Ref. 1). Newly synthesized proteins are evaluated in the ER through a cellular process known as quality control to determine their suitability for transport from the ER. Proteins being judged as unsuitable for export are selectively degraded by a pre-Golgi protein degradation pathway, probably within the ER itself. Misfolded or abnormal secretory proteins are not the only substrates of this pathway, however, as evidenced by the selective degradation of resident ER proteins as well.

One such resident which is subject to ER proteolysis is the enzyme HMG-CoA reductase (EC1.1.1.34; GenBank accession number M12705[GenBank]). HMG-CoA reductase is responsible for the synthesis of mevalonate, a key intermediate in the cellular synthesis of cholesterol and other isoprenoid compounds, such as dolichol and prenyl groups used to modify many proteins (2). Isoprenoid biosynthesis is tightly regulated in animal tissues and the key focal point for this regulation is HMG-CoA reductase. HMG-CoA reductase is subject to regulation at virtually all levels available to the cell, including transcriptional control of gene expression, translational regulation, protein phosphorylation, and regulated degradation of the enzyme (2). HMG-CoA reductase is normally a relatively stable enzyme which is degraded slowly. However, in the presence of excess mevalonate or sterols HMG-CoA reductase is degraded rapidly and selectively (2). This degradation does not result in the production of detectable proteolytic intermediates, despite the complicated topology of the enzyme. HMG-CoA reductase consists of an amino-terminal transmembrane domain which spans the membrane eight times and which anchors the catalytically active carboxyl-terminal cytosolic domain to the ER membrane (3). The amino-terminal membrane domain is necessary for the regulated degradation of the protein as evidenced by the stability of the catalytic domain when expressed free of the transmembrane anchor sequences. Expression of this free cytosolic domain results in the production of a catalytically active protein which is not subject to the normal regulated degradation (4), highlighting the importance of the membrane domain in the physiologically relevant degradative process (5).

Previous investigation into the mechanisms which are responsible for the regulated accelerated degradation of HMG-CoA reductase have revealed that the proteolysis is not inhibited by brefeldin A, an inhibitor of protein transport through the Golgi (6). These results support the hypothesis of a pre-Golgi site of proteolysis. Other pharmacological studies of HMG-CoA reductase degradation demonstrated that the degradative process is inhibited by cysteine protease inhibitors and calcium perturbants (7, 8). Reductase degradation was also shown to be unaffected by inhibitors of lysosomal proteolysis and to require on-going protein synthesis (7, 9). Such studies, while informative, have nevertheless failed to identify the protease(s) which catalyze the regulated degradation of HMG-CoA reductase or the mechanism by which the proteolysis is accelerated.

Early attempts to dissect the process of regulated ER proteolysis succeeded in the biochemical analysis of HMG-CoA reductase degradation in detergent permeabilized cells (10, 11). Meigs and Simoni (10) demonstrated that Chinese hamster ovary (CHO) cells permeabilized with digitonin and substantially free of cytosol, carry out regulated degradation of HMG-CoA reductase in a fashion which mimics the in vivo process (10). The degradation of HMG-CoA reductase in these permeabilized cells was more rapid if the cells had been supplemented with mevalonate or sterols prior to detergent permeabilization. The degradation was also partially inhibited by the cysteine protease inhibitor ALLN, an efficient inhibitor of the in vivo proteolysis of HMG-CoA reductase (7, 10). These results demonstrated that cytosolic proteins are not required to accomplish the proteolysis of HMG-CoA reductase once the process has been initiated by mevalonate or mevalonate-derived products.

Work by Correll and Edwards (11) extended the examination of HMG-CoA reductase degradation in vitro, examining the process in hepatic microsomes prepared from rats which had been injected with mevalonate prior to the harvesting of the liver. These workers demonstrated that hepatic microsomes prepared from rats which had received mevalonate degraded HMG-CoA reductase more rapidly than microsomes isolated from rats which had not received mevalonate. This degradation, however, was not sensitive to protease inhibitors known to block HMG-CoA reductase degradation in cultured cells, nor was it sensitive to other protease inhibitors, suggesting either a novel class of proteolytic enzyme or the presence of multiple proteases in the microsomal preparations (11).

In this report, we demonstrate an improved system for the study of the degradation of HMG-CoA reductase in purified ER membranes in vitro. Pretreatment of the cells from which the ER membranes are prepared with either mevalonate or 25-hydroxycholesterol results in the acceleration of the in vitro proteolysis. Degradation of HMG-CoA reductase in vitro is inhibited by the cysteine protease inhibitors, ALLN and E64, as well as the inhibitor of the multicatalytic proteasome, lactacystin (12). In vitro degradation in this system does not require the presence of cytosol, ATP, or the soluble lumenal contents of the ER and proceeds in the presence of solubilizing detergents. Finally, we confirm that purified ER membranes contain subunits of the multicatalytic protease, the proteasome. This biochemical system for the study of HMG-CoA reductase degradation has allowed us to make clear and testable predictions as to the mechanism of ER proteolysis and suggest that the acceleration of degradation is due to the increased susceptibility of reductase to proteolysis rather than the induction of a novel proteolysis system.

EXPERIMENTAL PROCEDURES

Minimal essential medium (MEM) without methionine and cysteine was obtained from ICN Biomedicals, Inc. 25-Hydroxycholesterol, DL-mevalonolactone, and phenylmethylsulfonyl fluoride (PMSF) were purchased from Sigma. DL-Mevalonolactone was converted to sodium mevalonate as described by Brown and Goldstein (9). N-Acetyl-leucyl-leucyl-norleucinal (ALLN) was purchased from Calbiochem and N-[N-(L-3-trans-carboxyoxiran-2-carbonyl)-L-leucyl]-agmatine (E64) was obtained from Boehringer Mannheim. Compactin was the generous gift of Akira Endo, Department of Agricultural and Biological Chemistry, Tokyo Noko University, Tokyo, Japan. Antisera specific for connexin43 was the kind gift from Dale Laird, McGill University. ER72-specific sera was the gift of Michael Green, Washington University, and proteasome antibodies were provided by Martin Rechsteiner and Katherine Ferrell, University of Utah. E36 ts20 cells were the generous gift of Ron Kopito, Stanford University. Polyvinylidene difluoride blotting membrane was purchased from Millipore. Donkey anti-rabbit Ig conjugated to horseradish peroxidase, ECL Western blotting detection reagents, and Hyperfilm-MP were purchased from Amersham.

C100 cells, a compactin-resistant SV40- transformed baby hamster kidney cell line that over-expresses HMG-CoA reductase (13), were maintained in minimal essential medium supplemented with nonessential amino acids and 5% fetal calf serum (MEM, 5% FCS) or 5% lipid- poor serum (MEM, 5% LPS). Lipid-poor serum (LPS) was prepared by solvent extraction as described previously by Rothblat et al. (14). Compactin was added to cells grown in MEM, 5% LPS to increase expression of HMG-CoA reductase. Roller bottle cultures grown in a 5% CO₂ atmosphere and incubated on a rolling apparatus at 37 °C. CHO cells were grown in MEM, 5% FCS at 37 °C. Chinese hamster E36 ts20 cells were grown in minimal essential medium supplemented with 10% fetal calf serum (MEM, 10% FCS) at 30 °C in a 5% CO₂ humidified incubator.

Cultures for ER membrane fractionation were grown in MEM, 5% FCS to 80% confluency in 850-cm² roller bottles. Cell monolayers were washed once with phosphate-buffered saline (PBS) followed by an overnight incubation in MEM, 5% LPS supplemented with 25 µg/ml compactin. For degradation studies, sodium mevalonate (20 mM) or 25-hydroxycholesterol (2.5 µM) were added 3 h prior to cell harvest. Fractionation of ER membranes was carried out as described by Urbani and Simoni (15) with all subsequent steps carried out on ice in a 4 °C cold room. Briefly, cells were harvested at 4 °C, concentrated by centrifugation at 500 × g for 4 min and washed with ice-cold MEM, 5% FCS, PBS, and 10 mM Tris-HCl (pH 7.5), with re-centrifugation between each wash. Cell pellets were resuspended in 2.0 ml of 10 mM Tris-HCl (pH 7.5) and incubated at 0 °C for 20 min. The cell suspension was then mixed with 2.0 ml of 10 mM Tris-HCl (pH 7.5), 300 mM sucrose and the cells disrupted by Dounce homogenization. 2.5 ml of the lysate was then mixed with 6.5 ml of 1 mM Tris-HCl (pH 7.5), 1 mM EDTA, 63.5% sucrose and layered under a sucrose density gradient (15). Gradients were centrifuged at 100,000 × g for 3 h and the microsome fractions collected. Pooled microsome fractions were diluted in 2 volumes of Tris-HCl (pH 7.5) and centrifuged at 100,000 × g for 1 h. After concentrating the microsomes and resuspension in 10 mM Tris-HCl (pH 7.5), 150 mM sucrose, the membrane fractions were quick frozen and stored in a liquid nitrogen freezer.

Protease assays were performed essentially as described by Twining (16). Cell lysates or purified microsome fractions were diluted in 50 mM Tris-HCl (pH 7.5), 5 mM CaCl₂ and incubated with resorufin-labeled casein at 37 °C for 4-16 h. Trichloroacetic acid was added to a final concentration of 5% and the samples incubated for 20 min at 37 °C. Samples were then centrifuged at 12,000 × g for 10 min and the supernatants collected and neutralized by the addition of 0.5 M Tris-HCl (pH 8.8). Fluorescence at 584 nm with an excitation wavelength of 574 nm was then detected using a Perkin-Elmer fluorescent spectrophotometer. Proteolytic release of trichloroacetic acid non-precipitable peptides was expressed as fluorescent units produced per min per mg of total protein added to each assay. Protein assays were carried out using the BCA Protein Assay (Pierce).

The degradation of HMG-CoA reductase was examined in purified ER fractions. ER fractions were thawed on ice and diluted 1:1 with ice-cold 10 mM Tris-HCl (pH 7.5). Aliquots of the diluted microsome fractions were dispensed into ice-cold Eppendorf tubes and incubated at 37 °C for various time periods. At the end of each incubation, an equal volume of sample buffer (62.5 mM Tris-HCl (pH 6.8), 8 M urea, 15% sodium dodecyl sulfate, 20% glycerol, 0.25% bromphenol blue, 25 mg/ml dithiothreitol, 100 µM PMSF, 25 µM ALLN, 100 µM leupeptin) was added to each tube and the sample quick frozen in a dry ice-ethanol bath. Samples were stored frozen until all samples from all time points were collected. Samples were then incubated at 37 °C for 30 min and the proteins separated by SDS-PAGE on a 5-15% polyacrylamide gradient gel (17). Prestained molecular weight standards (Amersham) were run on each gel to provide standards for molecular weight determination. Proteins were then transferred to a polyvinylidene difluoride blotting membrane and HMG-CoA reductase was detected by Western blotting with mixed antisera specific for the membrane anchor domain of HMG-CoA reductase (18) as well as antisera generated against the intact catalytic domain of reductase (19). Bound antibodies were detected using donkey anti-rabbit IgG conjugated to horseradish peroxidase and ECL Western blotting and were visualized using Hyperfilm-MP. The developed films were manually aligned with the blotted membrane and the position of each prestained standard marked. These standards are not substrates for the chemiluminescent detection system and so do not appear in the figures.

CHO cells were analyzed in a pulse-chase regimen as described previously (17). Briefly, cells were grown to near-confluency in MEM, 5% FCS medium. The media was then changed and the cells incubated overnight in MEM, 5% LPS supplemented with compactin (10 µM) and sodium mevalonate (100 µM). Cells were then starved in methionine-, cysteine-, and glutamine-free media for 1 h and pulse-radiolabeled with Tran³⁵S-label (ICN). Labeling media was removed and the cells chased in MEM, 5% LPS supplemented with 10 µM compactin, 2 mM methionine, and 2 mM cysteine. Accelerated degradation was observed by supplementing the chase media with mevalonate (20 mM) or 25-hydroxycholesterol (2.5 µM). At the indicated time points, samples were washed with ice-cold PBS and the labeled cells lysed and collected in solubilization buffer (17). Lysates were clarified by centrifugation at 16,000 × g and immunoreactive proteins precipitated with specific antisera and protein A-Sepharose. Immunoprecipitates were resolved by SDS-PAGE and visualized by autoradiography.

Pulse-chase of C100 cells was accomplished as above with minor alterations in protocol. C100 cells were preincubated with 20 mM sodium mevalonate or 2.5 µM 25-hydroxycholesterol for 2 h and pulse-radiolabeled for 30 min with Trans³⁵S-label. Labeling medium was then removed and cells incubated in chase medium supplemented with 2 mM methionine, 2 mM cysteine plus 20 mM sodium mevalonate or 2.5 µM 25-hydroxycholesterol where indicated. At the indicated time points, cell samples were processed for immunoprecipitation as described above and the recovered proteins resolved by SDS-PAGE.

E36 ts20 cells were grown in 60-mm dishes to 70-80% confluency in MEM, 10% FCS. 20 h prior to heat treatment and labeling, cells were washed once with PBS, and minimal essential medium supplemented with 10% LPS, 10 µM compactin, and 100 µM sodium mevalonate was added. Heat-treated cells were shifted to 44 °C for 1 h in MEM, 10% LPS supplemented with 10 µM compactin, 100 µM sodium mevalonate, and 25 mM HEPES (pH 7.5). The heat-treated cells were then immediately shifted to 40 °C for starvation, labeling, and chase. All cells were starved 1 h in methionine/cysteine-free minimal essential medium supplemented with 10 µM compactin and 100 µM sodium mevalonate, and labeled for 1 h in methionine/cysteine-free minimal essential medium containing 10 µM compactin, 100 µM sodium mevalonate, and 100 µCi/ml Tran³⁵S-label. Cells were chased in pre-warmed MEM, 10% LPS supplemented with 10 µM compactin, 100 µM sodium mevalonate, 2 mM methionine, and 2 mM cysteine for 0, 4, 8, and 16 h. At each chase time point, cells were collected by washing three times in ice-cold PBS followed by lysis in ice-cold solubilization buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 2 mM phenylmethylsulfonyl fluoride, 0.1 mM leupeptin, 2 µg/ml calpain inhibitor I, 10 mM sodium fluoride, 10 mM sodium orthovanadate, 100 mM dithiothreitol). Lysates were centrifuged at 16,000 × g for 15 min at 4 °C to remove insoluble material and supernatants were collected and processed for immunoprecipitation with antisera specific for HMG-CoA reductase or connexin43 antisera. Connexin43 samples were first precleared with preimmune sera and protein A-Sepharose, then incubated overnight with anti-connexin43 antisera at 4 °C. Immunoreactive species were immunoprecipitated as described previously (17). Immunoprecipitates were resolved by SDS-PAGE and visualized by autoradiography. Radioactive species were quantitated using the Bio-Rad PhosphorImager.

RESULTS

To prepare ER membranes for the analysis of HMG-CoA reductase degradation, we have used the procedure of Urbani and Simoni (15) to purify ER membranes from C100 cells. This fractionation procedure has been shown to reliably yield fractions substantially enriched in ER membranes. To verify that this procedure would yield substantial amounts of ER membrane containing intact HMG-CoA reductase, we grew C100 cells in 850-cm² roller bottles to near confluency. Cells were then lysed by Dounce homogenization, the nuclei removed by centrifugation at 1000 × g and the clarified lysate applied to the bottom of a discontinuous sucrose step gradient. After ultracentrifugation for 3 h at 100,000 × g, membrane bands were observed at the interfaces between the gradient layers. The bands were collected manually and the membranes concentrated for further analysis. Shown in Fig. 1A are the relative amounts of HMG-CoA reductase-immunoreactive material recovered in the total cell lysate and membrane fractions recovered from the gradient. The most dense membrane band resolved on this gradient, recovered at the interface between 55 and 38% sucrose, is highly enriched in HMG-CoA reductase and contains the majority of the HMG-CoA reductase-immunoreactive material resolved by this technique.

The ER membranes prepared by this procedure were also examined for proteolytic activity toward a model substrate, casein-resorufin (16). As shown in Fig. 1B, the proteases present in the final ER preparation retain approximately 10% of the total proteolytic activity present in the original cell lysate, as measured by this method. Furthermore, the proportion of the proteolytic activity which was inhibited by ALLN and PMSF is markedly lower in the ER preparation than in the total lysate.

The C100 line of baby hamster kidney cells was isolated in a selection for compactin resistance and were found to overexpress HMG-CoA reductase (13). Previous reports in the literature have differed over the regulation of HMG-CoA reductase degradation in these cells. While Hardemann et al. (22) concluded that reductase degradation in C100 cells was regulated by mevalonate and sterols, Peffley (21) reported that the addition of regulatory molecules to these cells did not accelerate the degradation of reductase. To resolve this apparent difference and to determine if the reductase protein expressed in these cells was subject to the normal process of regulated degradation, C100 cells were analyzed in a pulse-chase radiolabeling protocol. As shown in Fig. 2A, C100 cells express HMG-CoA reductase that is degraded relatively rapidly, with a half-life of approximately 3 h. This degradation rate is considerably faster than that reported for CHO cells (6, 7, 8). Supplementation of C100 cells with sodium mevalonate or 25-hydroxycholesterol, however, results in a significant acceleration of the degradation rate (Fig. 2A). C100 cells pretreated with either of these regulatory molecules degrade HMG-CoA reductase with a half-life of approximately 1 h. Both the basal level and accelerated degradation rates are considerably faster than those exhibited by CHO cells (6, 7, 8), although the sensitivity of the degradation rate to the presence of regulatory molecules suggests that C100 cells employ a similar mechanism for the degradation of reductase. The degradation of HMG-CoA reductase in C100 cells is also sensitive to the protease inhibitor ALLN (data not shown), similar to the degradative process in other cell types (7). Collectively, these data suggest that C100 cells are a suitable system for the study of HMG-CoA reductase degradation.

Previous results have shown that cytosolic components are not required for the regulated degradation of HMG-CoA reductase (10, 11) in detergent-permeabilized cells or crude hepatic rat microsomes (11). These results have strongly suggested that the cellular machinery necessary to accomplish the mevalonate-accelerated degradation of HMG-CoA reductase are localized to the endoplasmic reticulum. To further investigate the degradation of HMG-CoA reductase in vitro, we purified ER membranes (Fig. 1) from cells grown under varying conditions in which HMG-CoA reductase would be first maximally stable and then rapidly degraded. Cells were incubated in MEM, 5% LPS supplemented with compactin to increase expression of HMG-CoA reductase. Compactin is added to the growth media to increase the expression of HMG-CoA reductase. Three hours prior to cell harvest, mevalonate or 25-hydroxycholesterol were added to cultures for accelerated degradation studies. Cells were harvested, ER microsomes purified, and the degradation of HMG-CoA reductase examined in vitro. As shown in Fig. 2C, the predominant immunoreactive species present in these membrane preparations is a 97-kDa peptide which corresponds to intact HMG-CoA reductase (2). However, minor bands are also observed at 62 and 66 kDa, consistent with earlier reports of these proteolytic products in cells lysed in the absence of protease inhibitors (4).

Incubation of purified ER membranes at 37 °C results in the in vitro degradation of the intact 97-kDa intact HMG-CoA reductase protein but not the lower molecular weight proteolytic products. As shown in Fig. 2B, ER membrane proteins prepared from C100 cells grown only in the presence of compactin degrade HMG-CoA reductase slowly. At the end of a 4-h incubation at 37 °C, there is little change in the amount of HMG-CoA reductase in the ER sample, as compared to the zero time starting material. The HMG-CoA reductase in these membranes is degraded, however, if the 37 °C incubation was extended to 7-8 h (not shown). Supplementation of the cells with mevalonate or 25-hydroxycholesterol prior to cell lysis, however, results in a significant acceleration in the rate of HMG-CoA reductase degradation. ER membrane preparations from cells pretreated with mevalonate or 25-hydroxycholesterol degrade virtually all of the intact HMG-CoA reductase present after a 1-h incubation at 37 °C (Fig. 2C). The 62- and 66-kDa fragments of HMG-CoA reductase in these same membranes are stable throughout the incubation (Fig. 2C). These data show that ER membrane fractions retain proteases capable of degrading intact HMG-CoA reductase and that this degradation proceeds much more quickly if the microsomes are prepared from cells which have been pretreated with mevalonate or sterols. It is interesting to note that the degradation of HMG-CoA reductase in vitro proceeds more quickly than the in vivo process. We have no clear explanation for this phenomenon and have applied further criteria to establish the physiological relevance of the in vitro degradation.

The acceleration of HMG-CoA reductase degradation by mevalonate or sterols in vitro suggests that the mechanism of in vitro degradation is similar to the process of in vivo regulated proteolysis. To further investigate this, we performed degradation assays in the presence of various protease inhibitors. As seen in Fig. 3, ER membranes prepared from cells which received neither mevalonate nor sterols display a slow in vitro degradation rate in this assay and the amount of reductase remaining at the end of a 2-h incubation at 37 °C is comparable to that in the starting, zero time, sample. A preincubation with the protease inhibitors ALLN (50 µg/ml), E64 (60 µg/ml), APMSF (100 µM), or lactacystin (50 µg/ml) does not affect the stability of HMG-CoA reductase under these conditions (Fig. 3). In contrast, ER samples obtained from cells which had been supplemented with mevalonate or sterols display rapid reductase degradation, as shown by the nearly complete loss of HMG-CoA reductase in the ER samples after an incubation for 2 h at 37 °C (Fig. 3, lanes 7 versus 8 and 13 versus 14). The accelerated in vitro degradation of HMG-CoA reductase is blocked by a preincubation with the cysteine protease inhibitors ALLN, E64, or the serine protease inhibitor APMSF (Fig. 3, lanes 9-11 and 15-17). Although the inhibition of degradation by ALLN appears to be less than that observed for E64 or APMSF, in subsequent experiments it was determined that these inhibitors are equally effective if incubated with the samples at 4 °C for 15 min prior to warming to 37 °C. It is notable that the in vivo degradation of HMG-CoA reductase is also inhibited by ALLN and E64-derivatives (7). Addition of chymostatin, pepstatin, phosphoramidon, or aprotinin does not effect the in vitro degradation of HMG-CoA reductase (not shown). These results demonstrate that the process of in vitro degradation of HMG-CoA reductase resembles the in vivo process both in the acceleration by mevalonate or sterol supplementation as well as in protease inhibitor sensitivity. From these data, we conclude that the degradation of HMG-CoA reductase observed in vitro is physiologically relevant.

The accelerated in vitro degradation of HMG-CoA reductase is also inhibited by lactacystin (Fig. 3, lanes 12 and 18), a newly defined inhibitor of the multicatalytic proteasome (12). These results suggest a role for the multicatalytic protease, the proteasome, in HMG-CoA reductase proteolysis. It is significant that several other groups have recently reported a similar role for the proteasome in the degradation of ER proteins, including the cystic fibrosis transporter gene product, CFTR (23, 24), and major histocompatibility complex class I heavy chains (25).

The previous results have established that all components necessary to degrade HMG-CoA reductase are contained in the purified ER, either as membrane components or as soluble components contained within the ER lumen. To further define the location of the reductase protease(s), we undertook to separate the soluble proteins of the ER lumen from the integral and peripheral proteins of the ER membrane. To effect this separation, ER samples were sonicated and the resulting disrupted ER membranes isolated by ultracentrifugation. After sedimenting the membrane fragments, the distribution of known ER proteins was assessed by Western blotting with specific antisera. As shown in Fig. 4A, after vigorous sonication all detectable ER72, a known lumenal protein, is recovered in the non-sedimentable phase. These results are consistent with the disruption of the ER and release of soluble lumenal material. While the majority of HMG-CoA reductase remains sedimentable under these conditions, vigorous sonication is sufficient to render a fraction of HMG-CoA reductase into the non-sedimentable phase (Fig. 4B). These results suggest that some of the membrane fragments produced by this sonication are incapable of sedimentation under these conditions or are easily disturbed when the soluble phase is removed.

To determine if the ``lumen-free'' membrane fragments retain the proteases necessary for the degradation of HMG-CoA reductase, the pelletable material in Fig. 4B was analyzed in an in vitro degradation assay as described above. Membrane fragments obtained from cells which had not received either mevalonate or sterols continue to display a slow, basal rate of proteolysis of HMG-CoA reductase (Fig. 4C). The amount of HMG-CoA reductase in control membrane fragments incubated for 2 h at 37 °C is essentially the same as that present in the starting material and this basal rate of turnover is unaffected by the presence of ALLN (Fig. 4C, lane 2 versus 3). However, lumen-free ER membrane fragments prepared from cells which had been pretreated with either mevalonate or sterols retain the ability to degrade HMG-CoA reductase in a fashion which is both rapid and sensitive to ALLN (Fig. 4C). Lumen-free membrane fragments from mevalonate-pretreated cells degrade nearly all HMG-CoA reductase present after a 2-h incubation at 37 °C (Fig. 4C, lane 4 versus 5). This degradation is completely inhibited in the presence of ALLN (Fig. 4C, lane 6). The same pattern of accelerated degradation and inhibition is obtained for the microsomes prepared from sterol-pretreated cells (Fig. 4C, lanes 7-9). These data suggest that the cellular components necessary to degrade HMG-CoA reductase are localized to the membranes of the ER and that neither cytosol nor soluble lumenal contents are required for physiologically relevant accelerated proteolysis.

Two possibilities have been presented for the mechanism by which HMG-CoA reductase proteolysis is accelerated. First, it has been considered that addition of exogenous mevalonate or sterols might induce the synthesis or activation of a novel protease dedicated to the degradation of HMG-CoA reductase within the ER. Alternatively, supplementation with these regulatory molecules might elicit a change in the susceptibility of reductase to pre-existing ER proteases, rendering it a better substrate for ER proteolysis. The existence of a highly refined in vitro degradation system has allowed us to probe this essential question as to how HMG-CoA reductase is rendered labile. If supplementation with exogenous mevalonate increases the activity of a dedicated protease, proteases derived from mevalonate-accelerated ER would be predicted to degrade all HMG-CoA reductase molecules equally well, regardless of the source of the reductase protein. If, however, reductase degradation is determined by susceptibility and the proteases responsible for the degradation of HMG-CoA reductase are normal constituents of the ER, proteases from a mevalonate-accelerated ER sample would be predicted to have little effect when supplied in trans to a reductase population undergoing degradation at a basal rate. To investigate this, ER membrane samples displaying basal and accelerated degradation rates were solubilized in 1.0% Triton X-100, conditions known to solubilize HMG-CoA reductase (data not shown), and the detergent-solubilized samples used in an in vitro degradation assay. As shown in Fig. 5, ER membranes prepared from cells which had received no pretreatment degrade HMG-CoA reductase slowly in the presence of Triton X-100, as was the case in the absence of solubilizing detergent (Fig. 3). Basal rate samples which had been incubated at 37 °C for 3 h in the presence of 1.0% Triton X-100 contained similar amounts of HMG-CoA reductase as the starting material. Solubilized membranes prepared from mevalonate-pretreated cells, however, continue to degrade HMG-CoA reductase rapidly in the presence of Triton X-100 (Fig. 5). This rapid degradation by detergent-solubilized ER membrane fractions is blocked by ALLN, as in non-solubilized ER (not shown). Mixing the two membrane samples during detergent solubilization results in an intermediate degradation result. Mixed ER microsomes, solubilized with Triton X-100 and incubated at 37 °C for 3 h, retained approximately 40% of the total amount of reductase present in the starting material (Fig. 5), as opposed to the nearly complete degradation of reductase when the mevalonate-accelerated membranes were analyzed alone. Additionally, the degradation of reductase in these mixed-membrane samples displays a noticeably biphasic rate, suggesting that one population of reductase is degraded rapidly while a second population is more resistant to proteolysis. Centrifugation of the samples after detergent-addition at 200,000 × g, followed by the removal of the sedimentable material and mixing of only the solubilized proteins resulted in a similar pattern of reductase degradation (data not shown). ER membranes mixed in the absence of Triton X-100 also show a biphasic degradation pattern, indicating that the detergent does not cause this pattern and that the proteases of one membrane sample do not have access to HMG-CoA reductase from another sample when not solubilized. The interpretation of these experiments assumes that the detergent-solubilized proteins from one membrane sample mix completely with the solubilized proteins from the other sample. That the proteases supplied by the accelerated membranes were unable to degrade all reductase present, despite the longer incubation time of this experiment, suggests that the reductase supplied by the non-pretreated membranes is not an efficient substrate for the proteases present in the mevalonate-accelerated membranes. These results suggest that the acceleration of reductase degradation is mediated by changes in the susceptibility of the reductase to ER proteases, not the production of a new protease capable of degrading all HMG-CoA reductase, regardless of the source.

Lactacystin has recently been identified as a specific inhibitor of the multicatalytic protease, the proteasome (12), which has been implicated in the degradation of many cellular proteins, most notably substrates of the ubiquitin pathway for protein turnover, such as cyclin B and connexin43 (26, 27). The identification of lactacystin as an inhibitor of in vitro (Fig. 3) degradation of HMG-CoA reductase suggests that lactacystin might also block the in vivo process of regulated degradation. The degradation of HMG-CoA reductase in vivo was examined by pulse-chase analysis in CHO cells. CHO cells degrade HMG-CoA reductase considerably more slowly than C100 cells but retain the general pattern of acceleration by mevalonate or sterols (6, 7, 8). As shown in Fig. 6, CHO cells pulse-radiolabeled for 30 min followed by an 8-h chase retain approximately 50% of the total labeled HMG-CoA reductase, as compared to the zero time point. If, however, the chase medium was supplemented with either mevalonate or 25-hydroxycholesterol, the cells rapidly degrade the radiolabeled HMG-CoA reductase (Fig. 6, lane 3 versus 4) in this time period. The addition of lactacystin (50 µM) in the chase medium results in a significant inhibition of the accelerated degradation of HMG-CoA reductase (Fig. 6, lanes 5-7). These results suggest a role either for the multicatalytic proteasome or some other lactacystin-sensitive ER protease in the degradation of HMG-CoA reductase.

The involvement of proteasomes in the degradation of HMG-CoA reductase has been investigated using the in vitro degradation system described earlier. If the proteasome is required for the proteolysis of HMG-CoA reductase, proteasome subunits should be associated with both purified ER fractions and ER membranes which have been sonicated to release soluble lumenal contents. Although Rivett and co-workers (28) have previously shown the association of proteasomes with the ER membrane by immunoelectron microscopy, it was not known if these proteasomes were active in the degradation of ER proteins. ER membranes were sonicated as described above and the disrupted membranes separated into soluble and sedimentable fractions. The starting material and resulting subfractions of the ER were then evaluated for the presence of proteasome subunits by Western blotting with antisera specific for subunits of the 20S and 26S proteasome. As shown in Fig. 7B, the S4 subunit of the 26S proteasome (29) was detectable in the purified ER and the lumen-free ER membrane fragments but was not detected in the purified soluble material released from the membranes by sonication. Subunits of the 20S proteasome were also detected in the total ER and lumen-free ER samples (Fig. 7A) with a lesser amount of 20S proteasome subunits also being released from the membrane by sonication. The immunoreactive species observed in these experiments were not recognized by other rabbit antisera or by the secondary antibody-conjugate used in the ECL detection. It is important to note that the presence of these proteasome subunits in these samples does not demonstrate that the subunits are assembled into active multicatalytic proteasomes but, given the presence of both 20S and 26S subunits, we feel that the possibility is compelling. The presence of proteasome subunits in these samples, in combination with the inhibition of reductase degradation by lactacystin, demonstrate that the proteasome is stably associated with the ER membranes under the conditions used in the in vitro degradation experiments and suggest that the proteasome may be involved in the degradation of HMG-CoA reductase. However, the amount of proteasome present in the ER membranes did not appear to vary with pretreatment of cells with mevalonate or sterols prior to ER preparation (not shown), suggesting that recruitment of the proteasome to the ER was not a mechanism for the acceleration of HMG-CoA reductase degradation. It is possible, given earlier results, that the acceleration of HMG-CoA reductase degradation is mediated by increased recognition of reductase by the proteasome.

The degradation of the cystic fibrosis transmembrane conductance regulator (CFTR) within the ER has recently been shown to be dependent upon the ubiquitin conjugation system (23, 24). In contrast, HMG-CoA reductase degradation in vitro is carried out under conditions under which the system of ubiquitin activation and conjugation would not operate efficiently. Specifically, the degradation of HMG-CoA reductase occurs in the absence of exogenous ATP, which is required for the cycling of the ubiquitin system. The in vitro degradation is also efficiently achieved in the absence of cytosolic proteins, such as free ubiquitin and other components of the ubiquitin system. However, we have now directly evaluated the involvement of the ubiquitin system in the degradation of HMG-CoA reductase by two independent means. The presence of methyl-ubiquitin has been shown to inhibit the polyubiquitination and degradation of cyclins A and B in clam embryo extracts (30). We have exploited this phenomenon to test whether HMG-CoA reductase is ubiquitinated in the in vitro degradation assay. We have determined that the addition of methylated ubiquitin to ER membranes has no effect on the degradation of HMG-CoA reductase, regardless of the source of the ER (data not shown). The interpretation of these results is complicated, however, by the lack of a positive control whose degradation in the ER is blocked by methyl-ubiquitin.

We have independently tested the involvement of the ubiquitin system in the degradation of HMG-CoA reductase in vivo. The temperature-sensitive cell line E36 ts20 (32) expresses a mutant form of the hamster E1 enzyme, the protein required for ATP-dependent activation of the C-terminal glycine residue of ubiquitin, an essential reaction in the ubiquitin cycle (40). At the nonpermissive temperature (40 °C), these cells accumulate proteins which would normally be targeted for degradation by the ubiquitin pathway. One of the proteins which is stabilized by this treatment is connexin43, a protein which is normally degraded with a half-life of approximately 3 h (27). Incubation of E36 ts20 cells at 40 °C results in the stabilization of connexin43, even after 8 h of chase (27).

HMG-CoA reductase degradation was investigated in E36 ts20 cells by pulse-chase analysis at both the permissive (30 °C) and non-permissive (40 °C) temperatures. Cells were pulse-radiolabeled and chased at the respective temperatures. Samples were collected at the indicated time points and evaluated by immunoprecipitation. As shown in Fig. 8A, E36 ts20 cells degrade HMG-CoA reductase slowly, with a t_1/2 of approximately 8 h when grown in the absence of regulatory sterols. Supplementation of these cells with 25-hydroxycholesterol, however, results in the rapid acceleration of HMG-CoA reductase degradation, whether the chase is performed at 30 °C or 40 °C (Fig. 8B). In contrast, however, connexin43 recovered from these lysates is significantly stabilized by shift to 40 °C during the chase period relative to the sibling cells chased at 30 °C (Fig. 8C). These results demonstrate that HMG-CoA reductase degradation is unaffected by conditions which block the ubiquitin pathway. These data are in good agreement with previous reports that mouse mammary cells expressing a defective E1 enzyme were also capable of the efficient degradation of HMG-CoA reductase (42). On the basis of these data we conclude that the degradation of HMG-CoA reductase is not mediated by the ubiquitin system.

DISCUSSION

We have employed three different criteria to distinguish physiologically relevant degradation of HMG-CoA reductase from artifactual proteolytic cleavage events. These criteria are: induction by regulatory molecules known to accelerate reductase degradation in vivo, sensitivity to protease inhibitors which block in vivo degradation, and specificity for intact HMG-CoA reductase. The degradation of HMG-CoA reductase in vivo is substantially accelerated if the cells are treated with either mevalonate or sterols (Fig. 2A). As shown in Fig. 2C, a pretreatment of cells with regulatory molecules also accelerates HMG-CoA reductase degradation in vitro. The physiological relevance of the in vitro HMG-CoA reductase degradation is also suggested by the sensitivity of the degradation process toward various protease inhibitors. The proteolysis of HMG-CoA reductase in vivo is blocked by cysteine protease inhibitors as well as lactacystin (Fig. 6) (7). Similarly, addition of ALLN, E64, or lactacystin to the in vitro degradation assay efficiently blocks the degradation of HMG-CoA reductase in purified ER. This sensitivity to multiple proteolytic inhibitors with differing specificities suggests that multiple proteases may be involved in the degradative process or that a single protease with multiple sensitivities, such as the multicatalytic proteasome, may be required for the process.

The third criteria we have employed to distinguish physiolgically relevant degradation of HMG-CoA reductase from artifactual proteolysis is the specificity of the accelerated degradation. Incubation of purified ER membranes does not result in the rapid degradation of total ER protein (not shown) or of the 62-kDa fragment of HMG-CoA reductase (Fig. 2B). The 62-kDa fragment present in the ER membrane preparations contains portions of all three domains of reductase as demonstrated by recognition of this fragment by antisera raised against either the membrane-spanning or catalytic domains (data not shown). Despite this, the reductase fragment is stable in the in vitro degradation assay (Fig. 2C). These data suggest that the fragment lacks sequences necessary for rapid degradation and that rapid degradation is not a general phenomenon but is highly sequence specific. Collectively, these three criteria establish that the in vitro degradation of HMG-CoA reductase in purified ER membranes is reflective of the physiological process of regulated degradation in vivo.

Lumenal proteases, such as ER72 and ER60, have been implicated in the process of quality control within the ER (1, 20, 34). Likewise, membrane proteases, such as signal peptidase and the KEX2p of yeast, have been implicated in the processing of secretory and membrane proteins (33, 35, 36). In this work we show that ER membranes, free of soluble lumenal contents, retain the proteases necessary degrade HMG-CoA reductase in a mevalonate- or sterol-stimulated, ALLN-sensitive fashion (Fig. 4C). These results demonstrate that the proteases which degrade HMG-CoA reductase are restricted to the membrane of the ER. Such a localization raises several questions as to how HMG-CoA reductase, with both lumenal and cytoplasmically disposed domains, is degraded without detectable proteolytic intermediates. The mechanism(s) by which these topological challenges are met, however, has yet to be elucidated.

The mechanism by which the degradation of reductase is accelerated has been of primary interest for several years. Most simply, degradation can be accelerated by either enhanced protease-susceptibility or by activation of a reductase-specific protease. The data presented in Fig. 5 strongly suggest that the accelerated degradation of HMG-CoA reductase is mediated by increased susceptibility of the reductase to resident proteases of the ER. HMG-CoA reductase proteolytic susceptibility might be increased through any of a variety of mechanisms. Among these are binding to chaperone proteins which participate in the quality control process (38), a protein similar in function to the ornithine decarobxylase antizyme (39), or structural changes in the reductase protein due to changes in membrane composition. Chaperone binding is known to be an important part of the quality control process and a simple model for the regulation of proteolysis would predict the binding of an ER chaperone to the reductase in response to excess cholesterol or mevalonate-derived products. Such a binding would be predicted to result in an unfolding of the reductase protein and recognition by constitutive ER proteases. It is also possible, however, to envision structural changes in the reductase in direct response to the cholesterol content of the ER membrane, obviating the need for a proteinaceous chaperone to mediate the unfolding of reductase. HMG-CoA reductase spans the ER membrane eight times and these membrane spans represent an excellent target for changes in the reductase structure. Excess cholesterol in the ER membrane, due to endogenous biosynthesis or exogenous sources (45), could bind to the membrane spanning portions of reductase, resulting in a structural change in the membrane domain and increased susceptibility to membrane-bound proteases. This model is especially appealing since it potentially involves the fewest number of necessary components, requiring only HMG-CoA reductase, excess sterols, and the protease(s) which degrade reductase. With an efficient in vitro system for studying reductase degradation, it should be possible to dissect such a system, possibly even to the extent of biochemical reconstitution using purified components. It is also possible that the susceptibility of reductase to ER proteases is mediated by covalent modification of the protein.

The 26S proteasome and ubiquitin system have been implicated in the degradation of many cellular proteins, including the proteolysis of the CFTR within the ER (23, 24). Our observation that the degradation of HMG-CoA reductase is inhibited by lactacystin in vivo (Fig. 6) and in vitro (Fig. 3) suggests that the proteasome, or some other as yet unidentified lactacystin-sensitive protease, is required for the proteolysis of reductase. This suggestion is supported by the presence of proteasome subunits in the ER (Fig. 7). In contrast to CFTR, however, the ubiquitin system is not required for the degradation of reductase (Fig. 8). These results may indicate a role for the 20S proteasome in the degradation of HMG-CoA reductase or indicate non-ubiquitin-mediated proteolysis by the 26S proteasome. It is noteworthy that the ornithine decarboxylase, the best characterized substrate of regulated proteolysis, is degraded by the 26S proteasome in a ubiquitin independent fashion (43, 44). Further experimentation will be required to address the involvement of proteasome in reductase degradation.

The development of a refined in vitro system for the investigation of regulated protein degradation has allowed us to make testable predictions as to the mechanism of the process by which HMG-CoA reductase is degraded in response to regulatory molecules. Exploitation of this system has revealed important clues as to the mechanism by which proteolysis is accelerated and as to the subcellular localization of the proteases which degrade HMG-CoA reductase. Despite these advances, however, the identity of the protease(s) which degrade HMG-CoA reductase remains elusive. Although we have indications that the activity of the proteasome may be required for the degradative process this evidence is correlative and will require further confirmation. Continued exploitation of the combined in vitro and in vivo degradative systems is expected to reveal the mechanism by which HMG-CoA reductase is rendered susceptible to ER proteolysis and further define the proteolytic machinery of the ER. It will be of particular interest to determine the relationship between the proteases which degrade HMG-CoA reductase and the proteases of the general quality control process.