Insig-dependent Ubiquitination and Degradation of Mammalian 3-Hydroxy-3-methylglutaryl-CoA Reductase Stimulated by Sterols and Geranylgeraniol*

The endoplasmic reticulum enzyme 3-hydroxy-3-methylglutaryl-CoA reductase produces mevalonate, which is converted to sterols and to other products, including geranylgeraniol groups attached to proteins. The enzyme is known to be ubiquitinated and rapidly degraded when sterols and nonsterol end products of mevalonate metabolism accumulate in cells. Here, we use RNA interference to show that sterol-accelerated ubiquitination of reductase requires Insig-1 and Insig-2, membrane-bound proteins of the endoplasmic reticulum that were shown previously to accelerate degradation of reductase when overexpressed by transfection. Alanine substitution experiments reveal that binding of reductase to Insigs and subsequent ubiquitination require the tetrapeptide sequence YIYF in the second membrane-spanning helix of reductase. The YIYF peptide is also found in the sterol-sensing domain of SCAP, another protein that binds to Insigs in a sterol-stimulated fashion. When lysine 248 of reductase is substituted with arginine, Insig binding persists, but the reductase is no longer ubiquitinated and degradation is markedly slowed. Lysine 248 is predicted to lie immediately adjacent to a membrane-spanning helix, suggesting that a membrane-bound ubiquitin transferase is responsible. Finally, we show that Insig-dependent, sterol-stimulated degradation of reductase is further accelerated when cells are also supplied with the 20-carbon isoprenoid geranylgeraniol, but not the 15-carbon farnesol, raising the possibility that the nonsterol potentiator of reductase regulation is a geranylgeranylated protein.

The enzyme 3-hydroxy-3-methylglutaryl-CoA reductase (HMG-CoA reductase) 1 catalyzes the two-step reduction of HMG-CoA to mevalonate, a crucial intermediate in the synthesis of cholesterol and nonsterol isoprenoids (1). Mevalonatederived products include ubiquinone, heme, and the farnesyl and geranylgeranyl groups that become attached to many proteins, directing them to membranes. HMG-CoA reductase is subject to tight regulation by a multivalent feedback mechanism mediated by nonsterol and sterol end products of mevalonate metabolism (2). These end products decrease reductase activity by inhibiting transcription of the reductase gene, blocking translation of the reductase mRNA, and accelerating degradation of the reductase protein. The transcriptional effects are mediated through the action of sterol regulatory elementbinding proteins (SREBPs), membrane-bound transcription factors that enhance transcription of genes encoding cholesterol biosynthetic enzymes and the low density lipoprotein receptor (3). The translational effects are mediated by nonsterol mevalonate-derived isoprenoids, which act by an undefined mechanism (4). The accelerated degradation of reductase is mediated both by sterols and nonsterol end products of mevalonate metabolism (4,5) and appears to be carried out by the 26 S proteasome (6,7).
Hamster HMG-CoA reductase is an endoplasmic reticulum (ER) resident glycoprotein of 887 amino acids and consists of two distinct domains (8). The NH 2 -terminal domain of 339 amino acids contains eight membrane spanning regions separated by short loops (9); it anchors the protein to ER membranes. The COOH-terminal domain of 548 amino acids projects into the cytoplasm and exerts all of the catalytic activity (8). Two lines of evidence indicate that the membrane domain is crucial for sterol-accelerated degradation of the enzyme: 1) the truncated cytoplasmic COOH-terminal domain, expressed by transfection, retains full catalytic activity and exhibits a long half-life that is not shortened by sterols (10); and 2) the transmembrane domain of reductase fused to soluble ␤-galactosidase exhibits sterol-regulated degradation in a manner similar to wild-type reductase (11). Sterol-accelerated degradation of reductase occurs in the ER and is blocked by inhibitors of the 26 S proteasome, which lead to the accumulation of ubiquitinated forms of the enzyme (6,7).
A series of studies in the yeast Saccharomyces cerevisiae support a role for the ubiquitin-proteasome pathway in the control of HMG-CoA reductase degradation (12). HMG2p, one of two reductase isozymes expressed in yeast, undergoes accelerated degradation when flux through the mevalonate pathway is high (12,13). Degradation of HMG2p is dependent upon a membrane-bound ubiquitin ligase complex that interacts with HMG2p and mediates its ubiquitination in response to degradative signals (14).
Although accelerated degradation of mammalian reductase and that of yeast HMG2p seem similar, major differences exist. Although the membrane domain of yeast HMG2p includes multiple membrane spanning segments (13), it bears no primary sequence resemblance to that of mammalian reductase (8,15). Moreover, accelerated degradation of HMG2p in yeast is triggered maximally by a nonsterol isoprenoid (16), although maximal degradation of reductase in mammalian cells requires sterols plus nonsterol isoprenoids (4,5).
Recent clues to the mechanism of degradation of mammalian HMG-CoA reductase have emerged from the study of the transcriptional axis of sterol regulation as mediated by SREBPs.
The key discovery was SREBP cleavage-activating protein (SCAP), a polytopic membrane protein that resembles the reductase by virtue of eight membrane spanning segments. SCAP is a sterol-responsive escort protein that forms a complex with SREBPs and escorts them from the ER to the Golgi where SREBPs are proteolytically released from membranes, thereby allowing transcriptionally active fragments to enter the nucleus where they activate the reductase gene (3,17,18). Sterols block SREBP activation by preventing the exit of SCAP-SREBP complexes from the ER, leading to decreased synthesis of cholesterol and other lipids. The block in SREBP export is achieved through sterol-induced binding of SCAP to one of a pair of ER retention proteins called Insig-1 and Insig-2 (19,20). The binding site for Insig in SCAP has been localized to a ϳ170-amino acid segment that comprises transmembrane segments 2-6. These segments show significant sequence similarities to helices 2-6 of the membrane domain of HMG-CoA reductase, and this region has been termed the sterol-sensing domain (21)(22)(23). Point mutations within the sterol-sensing domain of SCAP prevent it from binding to Insigs (19,20), thus preventing the sterol-mediated ER retention of SCAP/SREBP (18,24,25).
The resemblance between the sterol-sensing domains of SCAP and HMG-CoA reductase led us to test the possible role of Insig-1 in the sterol-regulated degradation of the reductase (26). When reductase is overexpressed in Chinese hamster ovary (CHO) cells, the enzyme is no longer degraded rapidly in response to sterol treatment. Overexpression of Insig-1 restores regulated degradation of overexpressed reductase, and this degradation can be inhibited by overexpressing the sterolsensing domain of SCAP, suggesting that reductase and SCAP compete for limiting amounts of Insig when sterol levels are high. Although both proteins bind to Insig-1, the binding leads to markedly different consequences: binding leads to ER retention of SCAP and to accelerated degradation of reductase. How Insigs orchestrate seemingly discordant modes of cellular regulation requires further knowledge of the mechanisms underlying the sterol-mediated effects on reductase and SCAP. In this paper, we explore the role of Insigs, sterols, and nonsterol isoprenoids in regulating degradation of reductase.
Expression Plasmids-The following plasmids were described in the indicated reference: pCMV-Insig-1-Myc, which encodes amino acids 1-277 of human Insig-1 followed by six tandem copies of a c-Myc epitope tag under control of the cytomegalovirus promoter (19); pRc/CMV7S-Mev, which encodes amino acids 1-494 of the hamster Mev protein under control of the CMV promoter (28,29); and pCMV-HMG-Red-T7(TM1-8), which encodes amino acids 1-346 of hamster HMG-CoA reductase followed by three tandem copies of the T7 epitope tag under control of the CMV promoter (26). pEF1a-HA-ubiquitin (provided by Dr. Zhijian Chen, University of Texas Southwestern Medical Center) encodes amino acids 1-76 of human ubiquitin preceded by an epitope tag derived from the influenza hemagglutinin (HA) protein (YPYDVPDY) under the control of the EF1a promoter.
Cell Culture-Monolayers of Chinese hamster ovary cells (CHO-K1) were maintained in tissue culture at 37°C in 8-9% CO 2 . Stock cultures of CHO-K1 cells were maintained in medium A (1:1 mixture of Ham's F-12 medium and Dulbecco's modified Eagle's medium containing 100 units/ml penicillin and 100 g/ml streptomycin sulfate) supplemented with 5% (v/v) fetal calf serum (FCS).
Monolayers of SV-589 cells, an immortalized line of human fibroblasts expressing the SV40 large T antigen (30), were grown at 37°C in 5% CO 2 . Stock cultures of SV-589 cells were maintained in medium B (Dulbecco's modified Eagle's medium containing 100 units/ml penicillin and 100 g/ml streptomycin sulfate) supplemented with 10% FCS. SV589/pMev cells are derivatives of SV-589 cells stably expressing pMev, a cDNA encoding a mutated monocarboxylate transporter that confers efficient uptake of mevalonate from tissue culture medium (28). SV-589/pMev cells were generated as follows. On day 0, SV-589 cells were set up at 2 ϫ 10 5 cells per 60-mm dish in medium B supplemented with 10% FCS. On day 1, the cells were transfected with 2 g of pRc/CMV7S-Mev using the FuGENE 6 transfection reagent as described below. Twenty-four hours after transfection, cells were switched to selection medium that consisted of medium B supplemented with 10% LPDS, 50 M sodium compactin, 200 M sodium mevalonate, and 700 g/ml G418. Fresh medium was added every 2-3 days until colonies formed after about 10 days. Individual colonies were isolated with cloning cylinders, and pMev expression was assessed by the ability of exogenous mevalonate to induce reductase degradation. Once stably transfected cell lines were established, they were each maintained at 37°C, 5% CO 2 in medium B containing 10% LPDS, 50 M sodium compactin, 200 M sodium mevalonate, and 500 g/ml G418. Sterols were added to the culture medium at a final concentration of 0.1 or 0.2% (v/v) ethanol.
Transient Transfection, Cell Fractionation, and Immunoblot Analysis-Transfections were performed as described (31) with minor modifications. CHO-K1 cells were transfected with 3 g of DNA per 60-mm dish. For each transfection, FuGENE 6 DNA transfection reagent (Roche Diagnostics) was added to 0.2 ml of medium A at a ratio of 3 l of FuGENE-6 per 1 g of DNA. Conditions of incubation are described in the figure legends. At the end of the incubation, dishes of cells were harvested and pooled for analysis.
Blue Native-PAGE for Detection of HMG-CoA Reductase-Insig-1 Complex-Monolayers of SRD-13A cells, a line of mutant CHO cells that lacks SCAP (31), were set up on day 0 (8 ϫ 10 5 cells/100-mm dish) and cultured in 8 -9% CO 2 at 37°C in medium A supplemented with 5% FCS, 5 g/ml cholesterol, 1 mM sodium mevalonate, and 20 M sodium oleate. On day 2, the cells were transfected as described in the legend to Fig. 5, after which they were harvested for preparation of a 1.6 ϫ 10 4 -g membrane pellet, which was solubilized with 1% (w/v) digitonin and centrifuged as described (19). One aliquot of the resulting 20,000 ϫ g supernatant was mixed with a 6-amino-n-hexanoic acid containing buffer and subjected to electrophoresis on a 4 -16% blue native gel as described (19). A second aliquot of the 20,000 ϫ g supernatant was mixed with SDS-containing buffer and subjected to 10% SDS-PAGE.
On day 0, SV-589 cells were set up at a density of 5 ϫ 10 4 cells per 60-mm dish in medium B supplemented with 10% FCS. On day 1, the cells were transfected with siRNA using a ratio of 6 l of OligofectAMI-NE TM reagent (Invitrogen) to 400 pmol of siRNA duplexes per dish as described by the manufacturer. Cells were washed once with 2 ml of medium B (without antibiotics) and refed with 1.6 ml of medium B (without antibiotics). The siRNA/OligofectAMINE mixture (in a total volume of 0.4 ml) was added to each dish and incubated at 37°C. After 6 h, the cells received a direct addition of 1 ml of medium B containing (final concentrations) 10% FCS, 5 g/ml cholesterol, 1 mM sodium mevalonate, and 20 M sodium oleate, and were then incubated at 37°C. On day 3, the cells were transfected with siRNA duplexes as described above. Six hours after the final transfection, the cells received a direct addition of 1 ml of medium B containing (final concentration) 10% LPDS, 50 M sodium compactin, and 50 M sodium mevalonate. On day 4, the cells were treated as described in the figure legends, and harvested for analysis.
Real-time PCR-The protocol was identical to that described by Liang et al (35). Triplicate samples of first-strand cDNA were subjected to real-time PCR quantification using forward and reverse primers for human Insig-1, human Insig-2, and human 36B4 (invariant control) ( Table I). Relative amounts of mRNAs were calculated using the comparative C T method.
Pulse-chase Analysis of HMG-CoA Reductase-Cells were pulse-labeled in medium C (methionine/cysteine-free Dulbecco's modified Eagle's medium containing 100 units/ml penicillin, 100 g/ml streptomycin sulfate, 10% LPDS, and 50 M sodium compactin), and pulse-chase analysis was carried out as described (26). Immunoprecipitation of labeled reductase was carried out with 30 g of a polyclonal antibody directed against the 60-kDa COOH-terminal domain of human reductase (36). Immunoprecipitates were subjected to SDS-PAGE and transferred to Hybond C-extra nitrocellulose filters. Dried filters were exposed to an imaging plate at room temperature and scanned in a Fuji X Bas 1000 Phosphorimager.
Ubiquitination of HMG-CoA Reductase-Conditions of incubations are described in the figure legends. At the end of the incubation, cells were harvested, and immunoprecipitations with either polyclonal antibody against endogenous human reductase (see above) or monoclonal anti-T7 IgG-coupled agarose beads (Novagen) against transfected reductase were carried out as described (26) with two modifications: the lysis buffer used to harvest the cells was supplemented with 10 mM N-ethylmaleimide and the pelleted beads were washed three times (30 min each). Immunoprecipitates were subjected to SDS-PAGE on 5-12% gradient gels and transferred to nitrocellulose filters.

RESULTS
The experiment in Fig. 1 was designed to examine the relationship between ubiquitination and sterol-accelerated degradation of HMG-CoA reductase. SV589 cells, a line of SV-40transformed human fibroblasts, were depleted of sterols by incubation for 16 h in LPDS containing the reductase inhibitor, compactin, and a low level of mevalonate (50 M), which is the lowest level that assures viability. Cells were then treated for either 2 or 5 h with a mixture of sterols (1 g/ml 25-hydroxycholesterol and 10 g/ml cholesterol) and a high concentration of mevalonate (10 mM) in various combinations. Additionally, some of the cells received the proteasome inhibitor MG-132 to block the degradation of any ubiquitinated reductase molecules. Detergent-solubilized extracts were subjected to immunoprecipitation with polyclonal antibodies against reductase, and the resulting immunoprecipitates were subjected to SDS-PAGE and blotted with anti-ubiquitin ( or presence (ϩ) of sterols (1 g/ml 25-hydroxycholesterol and 10 g/ml cholesterol), 10 mM sodium mevalonate, and 10 M MG-132 as indicated. After 2 (lanes 1-8) or 5 h (lanes 9 -16) at 37°C, cells were harvested, lysed, and subjected to immunoprecipitation with polyclonal anti-reductase as described under "Experimental Procedures." Aliquots of immunoprecipitates were subjected to SDS-PAGE, transferred to nylon membranes, and immunoblotted with 5 g/ml monoclonal IgG-A9 (against reductase) or 0.2 g/ml monoclonal IgG-P4D1 (against ubiquitin). Filters were exposed to film for 2 (top gel) and 30 s (bottom gel). B, cells were incubated at 37°C for the indicated times in the absence (Ϫ) or presence (ϩ) of sterols and 10 M MG132 as indicated, harvested, and subjected to immunoprecipitation and immunoblot analysis as in A. Filters were exposed to film for 30 (top gel) and 10 s (bottom gel).  [5][6][7][8], indicating that this decrease reflected accelerated degradation mediated by proteasomes. After 2 h of sterol treatment in the presence of MG-132, we observed the appearance of ubiquitinated reductase as indicated by the high molecular weight smear in the anti-ubiquitin immunoblots of the reductase immunoprecipitates (lanes 6 and 8, upper panel), but not in cells treated with mevalonate alone (lane 7). The addition of mevalonate did not enhance the sterol effect (compare lanes 6 and 8). After 5 h in the presence of sterols alone, reductase was still detectable, although diminished (lower panel, lane 10). In the presence of sterols plus mevalonate, the reductase was further reduced (lower panel, lane 12). This result is consistent with other studies demonstrating that sterols alone accelerate the degradation of reductase and that nonsterol mevalonate-derived products further accelerate degradation (4,5). Accelerated degradation of reductase was completely blocked by MG-132 (lower panel, lanes [13][14][15][16]. In the presence of MG-132, sterols caused a marked increase in the amount of ubiquitinated reductase, and there was no further effect with mevalonate addition (compare lanes 14  The sterol stimulation of reductase ubiquitination was remarkably rapid (Fig. 1B). In the absence or presence of the proteasome inhibitor, sterol-dependent ubiquitination of reductase was observed after as little as 5 min of sterol treatment (lanes 2 and 3) and reached a maximum after 10 min (lanes 5 and 6). Thereafter, in the absence of MG-132 ubiquitinated reductase declined, presumably because of degradation (lanes 8 and 11). This decline was prevented by MG-132 (lanes 9 and 12).
Although the data of Fig. 1 indicate that reductase is ubiquitinated and degraded in a sterol-dependent manner, the proportion of reductase that was detectably polyubiquitinated at any one time was very small, even when degradation was blocked by MG-132. Thus, in the presence of MG-132 most of the reductase that accumulated was not polyubiquitinated, as indicated by its continued migration on SDS-PAGE as a 97-kDa band (lower panel in Fig. 1). We interpret these findings to indicate a balance between ubiquitination and de-ubiquitination reactions. According to this hypothesis, sterols stimulate the ubiquitination of reductase. When proteasomal function is intact, most of the ubiquitinated reductase is degraded. When proteasomal degradation is blocked, the amount of ubiquitinated reductase at any instant reflects an equilibrium between ubiquitination and de-ubiquitination reactions. Similar conclusions have been reached by others in studies of the ubiquitination of other proteins (37). The addition of mevalonate enhances reductase degradation, but it is not clear whether mevalonate increases reductase ubiquitination or whether it acts to enhance degradation of reductase that has become ubiquitinated under the influence of sterols. It is clear, however, that nonsterol mevalonate products cannot act by themselves. They only act in the presence of sterols.
To demonstrate directly that the RNAi knockdown of Insig-1 and Insig-2 abolishes sterol-accelerated degradation of reductase, we performed a pulse-chase experiment. Cells transfected with siRNA duplexes were pulse-labeled for 1 h with 35 Slabeled methionine plus cysteine, after which they were washed and switched to medium containing an excess of unlabeled methionine and cysteine in the absence or presence of sterols plus mevalonate. Radiolabeled reductase was monitored by immunoprecipitation with anti-reductase antibodies followed by SDS-PAGE and visualization in a phosphorimager. Fig. 2C shows the time course of disappearance of 35 S-labeled reductase in cells treated with siRNA targeting VSV-G (upper panel) or Insig-1 plus Insig-2 (lower panel). In control cells treated with VSV-G siRNA, sterols plus mevalonate caused a rapid disappearance of 35 S-labeled reductase that was nearly complete after 5 h (upper panel, lanes 5-7). However, in cells transfected with Insig-1 plus Insig-2 siRNA, labeled reductase continued to be present throughout the chase (lower panel, lanes 5-7). These results were quantified by scanning of images in a phosphorimager, and the data are plotted in the lower panel of Fig. 2C. The half-life of labeled reductase in control cells declined from 11.5 to 0.9 h upon treatment with sterols plus mevalonate. In cells treated with siRNAs targeting the two Insigs, reductase half-life was 10.4 h in the absence or presence of sterols plus mevalonate.
We next addressed the question of whether Insigs are necessary for the ubiquitination of reductase or whether they are required for a post-ubiquitination step. To this end, we transfected cells with siRNA duplexes targeting VSV-G (Fig. 2D,  lanes 1-4) or Insig-1 plus Insig-2 (lanes 5-8). SV-589 cells were set up on day 0 at 5 ϫ 10 4 cells per 60-mm dish in medium B with 10% FCS. On days 1 and 3, cells were transfected with the indicated siRNA as described under "Experimental Procedures." The total amount of siRNA was adjusted to 400 pmol/dish by addition of VSV-G siRNA. After the second transfection on day 3, cells were incubated for 16 h at 37°C in medium B containing 10% LPDS, 50 M compactin, and 50 M mevalonate. A and B, on day 4, cells were switched to medium B containing 10% LPDS and 50 M compactin in the absence (Ϫ) or presence (ϩ) of sterols (1 g/ml 25-hydroxycholesterol and 10 g/ml cholesterol) and/or 10 mM mevalonate as indicated. After 5 h, cells were harvested for preparation of whole cell extracts (A) and total RNA (B). Aliquots of extracts (10 g of protein) were subjected to SDS-PAGE and immunoblot analysis with 5 g/ml monoclonal IgG-A9 (against reductase) and 1 g/ml monoclonal IgG-CD71 (against transferrin receptor). Filters were exposed to film for 1 min (top) and 1 s (bottom). Total RNA was prepared from cells treated with sterols and mevalonate (lanes 4, 8, 12, and 16). Aliquots of reverse-transcribed total RNA (20 ng) were subjected to real-time PCR. Each value for cells transfected with the indicated siRNA represents the amount of Insig-1 mRNA or Insig-2 mRNA relative to that in control cells transfected with siRNA targeting VSV-G. C, on day 4, cells were preincubated for 1 h in medium C, after which they were pulse-labeled for 1 h with 100 Ci/ml [ 35 S]methionine in medium C. Cells were then chased in medium C supplemented with 0.5 mM unlabeled methionine and 1 mM unlabeled cysteine in the absence (Ϫ) or presence (ϩ) of sterols (1 g/ml 25-hydroxycholesterol and 10 g/ml cholesterol) plus 10 mM mevalonate as indicated. After the indicated chase period, cells were harvested, lysed, and subjected to immunoprecipitation with polyclonal anti-reductase. Aliquots of immunoprecipitates (normalized for equal 35 S radioactivity incorporated into total cell protein) were subjected to SDS-PAGE and transferred to nylon membranes. Filters were exposed for 48 h to an imaging plate at room temperature, scanned, and quantified in a Fuji X Bas 1000 Phosphorimager, and the image was photographed. The amount of 35 S-labeled reductase during the pulse (lane 1 of both gels) was arbitrarily set at 100%. D, on day 4, cells were switched to medium B containing 10% LPDS, 50 M compactin, and 10 M MG-132 in the absence (Ϫ) or presence (ϩ) of sterols (1 g/ml 25-hydroxycholesterol and 10 g/ml cholesterol) and/or 10 mM mevalonate as indicated. After 1 h cells were harvested, lysed, and subjected sequentially to immunoprecipitation, SDS-PAGE, and immunoblot analysis as described in the legend to Fig. 1. Filters were exposed to film for 3 (top gel) and 30 s (bottom gel).
FIG . 3. Lysines 89 and 248 are required for regulated ubiquitination and degradation of HMG-CoA reductase. A, amino acid sequence and topology of the membrane domain of hamster reductase. Lysine residues are shown in black with those located in the cytosol denoted by arrows. B and C, CHO-K1 cells were set up on day 0 at 5 ϫ 10 5 cells/60-mm dish in medium A with 5% FCS. On day 1, cells were transfected in 2 ml of medium A containing plasmid encoding full-length reductase with three tandem copies of the T7 epitope tag at the COOH terminus of the protein.
To enhance sensitivity of ubiquitin detection, cells were also transfected with pEF1a-HA ubiquitin, an expression plasmid encoding human ubiquitin with an NH 2 -terminal tag consisting of two copies of the HA epitope. In the experiment shown in Fig. 3B, CHO cells were transfected with pEF1a-HA-ubiquitin and wild-type (lanes 1-4) or mutant versions (lanes 5-16) of pCMV-HMG-Red-T7 in the absence or presence of pCMV-Insig-1-Myc. After treatment for 2 h with MG-132, cells were harvested, transfected reductase was immunoisolated from detergent lysates with anti-T7 agarose beads, and samples were subjected to SDS-PAGE. For simplicity, the cells were incubated with a mixture of sterols plus mevalonate instead of adding each one separately. Immunoblot analysis with anti-HA revealed that ubiquitination of wild-type reductase required Insig-1 plus sterols and mevalonate (upper panel, lanes 1-4).
The overexpressed sterol-sensing domain of SCAP blocks sterol-mediated degradation of reductase, suggesting that the two proteins compete for the same site on Insig (26). Binding of SCAP to Insig is abolished upon the substitution of a cysteine for tyrosine 298 (Y298C mutant), which lies within one of the transmembrane helices in the sterol-sensing domain of SCAP (18,20,25). The Y298C mutant of SCAP fails to competitively inhibit sterol-mediated degradation of reductase presumably because it fails to bind to Insigs (19). Fig. 4A shows an amino acid alignment between the second transmembrane helices of reductase and SCAP. Reductase and SCAP share a tetrapeptide (YIYF) beginning at residue 75 of reductase, which corresponds to the crucial Tyr-298 residue in SCAP. To determine whether this tetrapeptide sequence is essential for reductase interaction with Insigs, we generated a series of mutations in pCMV-HMG-Red-T7 and tested their effects on sterol-acceler-ated degradation (Fig. 4B). CHO cells were transfected with pCMV-Insig-1-Myc and wild-type or mutant versions of pCMV-HMG-Red-T7 containing various combinations of alanine substitutions in the YIYF motif. The level of Insig expression is sufficient to permit nearly complete degradation of reductase in 5 h in the presence of 25-hydroxycholesterol without added mevalonate. After incubation with various concentrations of 25-hydroxycholesterol for 5 h, cells were harvested, fractionated, and subjected to SDS-PAGE followed by immunoblot analysis with anti-T7 antibody. Under the conditions of this experiment, 0.1 g/ml 25-hydroxycholesterol was sufficient to accelerate the degradation of wild-type reductase, and complete disappearance was observed at 0.3 g/ml of the oxysterol (Fig. 4B, lanes 1-4 and 17-20). The Y75A mutant (corresponding to the Y298C mutant in SCAP) reduced only slightly the effect of sterols on degradation of reductase (Fig. 4B, lanes  5-8). However, substitution of the other tyrosine in this tetrapeptide (Y77A) rendered reductase significantly resistant to the sterol effect (lanes 9 -12). The combined effect of the two tyrosine mutations (Y75A/Y77A) was more potent than either of the two substitutions alone. The isoleucine at the second position of the tetrapeptide was also important for sterol-accelerated degradation (lanes 21-24), but the phenylalanine at the fourth position was not (F78A, lanes [25][26][27][28]. Substitution of all four residues of the tetrapeptide (YIYF to AAAA) prevented sterols from exerting any effect on the degradation of reductase (lanes 33-36).
We next tested whether ubiquitination of the various mutants of reductase correlates with their resistance to sterolaccelerated degradation (Fig. 4C). Cells were transfected with pCMV-Insig-1-Myc, pEF1a-HA-ubiquitin, and wild-type (lanes 1 and 2) or mutant versions of pCMV-HMG-Red-T7 (lanes [3][4][5][6][7][8][9][10][11][12]. The cells were treated for 2 h with MG-132 in the absence or presence of sterols plus mevalonate and harvested, and detergent extracts were prepared. Transfected reductase was immunoprecipitated with anti-T7-coupled agarose beads. Immunoblotting the precipitates with anti-HA monoclonal antibody revealed that the Y75A mutant of reductase was ubiquitinated in a regulated manner, similar to that of the wildtype protein (lanes 1-4). In contrast, ubiquitination of all of the degradation-resistant mutants was markedly reduced (lanes [5][6][7][8][9][10][11][12]. These results strengthen the view that the conserved YIYF sequence in reductase is essential for the ability of Insig-1 to mediate the regulated ubiquitination and the subsequent proteasomal degradation of reductase. The results in Figs. 3 and 4 show that lysine 248 and the YIYF sequence are necessary for sterol-mediated reductase ubiquitination. We utilized blue native-polyacrylamide gel electrophoresis to test the hypothesis that mutations in the YIYF sequence of reductase decrease the binding of protein to Insig-1, whereas substitution of lysine 248 preserves binding but prevents ubiquitination. This technique was recently employed to identify sterol-induced complexes between Insig-1 and SCAP (19). To eliminate the potential competition between SCAP and reductase for Insig, we utilized SRD-13A cells, a 5% LPDS and 1 g of wild-type (lanes 1-4), K89R (lanes 5-8), K248R (lanes 9 -12), or K89R/K248R (lanes 13-16) versions of pCMV-HMG-Red-T7 in the absence or presence of 30 ng of pCMV-Insig-1-Myc as indicated. In B, cells in each lane were also transfected with 0.1 g of pEF1a-HAubiquitin. The total DNA in each lane was adjusted to 3 g/dish by the addition of pcDNA3 mock vector. Six hours after transfection, cells received a direct addition of 2 ml of medium A containing 5% LPDS, 10 M compactin, and 50 M mevalonate (final concentrations). After 16 h at 37°C, cells were switched to medium A containing 5% LPDS and 50 M compactin in the absence (Ϫ) or presence (ϩ) of sterols (1 g/ml 25hydroxycholesterol and 10 g/ml cholesterol) plus 10 mM mevalonate as indicated. In B, the medium also contained 10 M MG-132. B, following incubation at 37°C for 2 h, cells were harvested, lysed, and subjected to immunoprecipitation with monoclonal anti-T7-coupled agarose beads. Aliquots of immunoprecipitates were subjected to SDS-PAGE and transferred to nylon membranes. Immunoblot analysis was carried out with 1:1000 dilution of monoclonal anti-HA-IgG (against ubiquitin) or 1 g/ml monoclonal anti-T7 IgG (against reductase). Filters were exposed to film for 5 s. C, after incubation for 5 h, cells were harvested, membrane fractions were prepared, and aliquots (10 g) were subjected to SDS-PAGE. Immunoblot analysis was carried out with 1 g/ml monoclonal anti-T7 IgG (against reductase) and 1 g/ml monoclonal anti-Myc IgG (against Insig-1). Filters were exposed to film for 7-10 s.
CHO-derived mutant cell line that lacks SCAP (31). In the experiment shown in Fig. 5, we transfected SRD-13A cells with pCMV-Insig-1-Myc in the absence (lanes 1 and 2) or presence of wild-type or mutant versions of pCMV-HMG-Red-T7(TM1-8) (lanes 3-8). pCMV-HMG-Red-T7(TM1-8) is an expression plasmid encoding the membrane domain of hamster reductase (amino acids 1-346) with a COOH-terminal T7 epitope tag (26). Following sterol depletion with hydroxypropyl-␤-cyclodextrin, the cells were incubated in medium with or without sterols. The medium also contained MG-132 to prevent proteasomal degradation of the ubiquitinated reductase membrane domain. After treatments, the cells were harvested, membranes were isolated and solubilized with digitonin. Soluble material was mixed with Coomassie Blue and 6-amino-n-hexanoic acid (to impart a negative charge and prevent precipitation of proteins) and subjected to PAGE. The fractionated proteins were transferred to membranes and blotted with various antibodies. In the absence of the reductase membrane domain, Insig-1 appeared as a single band as revealed by immunoblotting with anti-Myc. Its migration was not altered upon sterol treatment (compare lanes 1 and 2 in the upper panel of Fig. 5). This band was designated as unbound Insig-1. When the cells expressed the wild-type reductase membrane domain, some of the Insig-1 migrated more slowly (lane 3). This slower migrating band contained Insig-1 bound to the membrane domain of reductase, as demonstrated by the fact that it was supershifted when the digitonin-solubilized material was incubated prior to PAGE with anti-T7 monoclonal antibody, but not with a control antibody (data not shown). Thus, we designated the slower migrating band as the reductase-Insig complex. In the presence of sterols, unbound Insig-1 disappeared almost completely (lane 4). The reductase-Insig-1 complex was also diminished, but to a lesser extent (lane 4). The decrease in the reductase-Insig-1 complex is likely caused by incomplete inhibition of proteasomal activity, as revealed by immunoblotting of another aliquot of the same extract that was subjected to SDS-PAGE. This blot revealed a decrease in the total amount of the reductase membrane fragment in the presence of sterols (SDS-PAGE, lower  panel, lanes 3 and 4).
In cells that expressed the K248R mutant of reductase, about FIG. 4. Mutations in the sterol-sensing domain of HMG-CoA reductase confer resistance to sterol-mediated degradation. A, comparison of the amino acid sequence of the second transmembrane domain of reductase with that of the comparable sterol-sensing region in SCAP. Sequences are from Chinese hamster reductase (amino acids 57-84) and Chinese hamster SCAP (amino acids 280 -307). GenBank TM accession numbers are L00165 and U67060, respectively. Identical amino acids are highlighted in black. Asterisk denotes the site of sterol-sensing mutation (Y298C) in SCAP. B and C, on day 0, CHO-K1 cells were set up at 5 ϫ 10 5 cells/60-mm dish in medium A with 5% FCS. On day 1, cells were transfected with 1 g of the indicated wild-type or mutant version of pCMV-HMG-Red-T7 together with 30 ng of pCMV-Insig-1-Myc. Cells in C were also transfected with 0.1 g of pEF1a-HA-ubiquitin. Six hours after transfection, cells received a direct addition of 2 ml of medium A to produce final concentrations of 5% LPDS, 10 M compactin, and 50 M mevalonate. B, after 16 h, cells were switched to 3 ml of medium A containing 5% LPDS, 50 M compactin, and the indicated concentration of 25-hydroxycholesterol. After incubation for 5 h, cells were harvested, membrane fractions were prepared, and aliquots (15-20 g) were subjected sequentially to SDS-PAGE and immunoblot analysis as described in the legend to Fig. 3. Filters were exposed to film for 5-20 s. C, following incubation for 16 h, cells were switched to 3 ml of medium A containing 5% LPDS, 50 M compactin, 10 M MG-132 in the absence (Ϫ) or presence (ϩ) of sterols (1 g/ml 25-hydroxycholesterol and 10 g/ml cholesterol) plus 10 mM mevalonate. After 2 h at 37°C, cells were harvested, lysed, and subjected sequentially to immunoprecipitation, SDS-PAGE, and immunoblot analysis as described in the legend to Fig. 3. Filters were exposed to film for 1-5 s. half of the Insig-1 was found in the reductase-Insig-1 complex in the absence of sterols (lane 5). Addition of sterols caused all of the Insig-1 to migrate as a complex with reductase (lane 6), indicating that this mutation does not block the ability of sterols to enhance binding of the reductase membrane fragment to Insig-1. Sterol addition caused only a slight decline in the total amount of the K248R reductase membrane fragment (SDS-PAGE, lane 6), presumably because this mutant is resistant to ubiquitination and degradation. A very different result was obtained with the tetrapeptide mutant (YIYF to AAAA). In cells that expressed this mutant protein, the majority of Insig-1 was in the unbound state, and there was no increase in the complex when sterols were added (blue native-PAGE, lanes 7 and 8), nor was there a detectable decrease in the total amount of the reductase membrane domain as revealed by SDS-PAGE (lanes 7 and 8).
We next conducted a series of studies designed to determine which product of mevalonate metabolism, in addition to sterols, is required to accelerate the degradation of reductase. To facilitate cellular uptake of mevalonate, we used a permanently transfected line of human SV-589 cells that expresses a cell surface mevalonate transporter (SV589-pMev cells). This transporter is a mutant version of a monocarboxylate transporter (MCT-1) with a substitution of phenylalanine for cysteine at position 360, which lies in a putative transmembrane helix (28). The mutant protein, but not the native protein, transports mevalonate into cells with high efficiency and thus facilitates studies of the effects of mevalonate and its products on reductase. Fig. 6A show three of the branches of the mevalonate pathway in animal cells. Mevalonate, produced by reductase, is converted to the five-carbon intermediate isopentenyl pyrophosphate (isopentyl-PP), which is polymerized to form geranyl-PP (10 carbons) and farnesyl-PP (15 carbons). Two farnesyl-PP molecules can undergo head-to-head condensation to produce squalene (30 carbons), which is the first committed intermediate in the cholesterol synthetic pathway. Conversion of squalene to cholesterol is blocked by NB-598, which inhibits squalene epoxidase (38). Farnesyl-PP can donate a farnesyl group to certain proteins, and it can condense with another isopentenyl-PP to form geranylgeranyl-PP (20 carbons), which is added to other proteins by two distinct geranylgeranyl transferases (39).
When SV589-pMev cells were incubated with a higher concentration of mevalonate (10 mM) in the presence of the proteasome inhibitor MG-132, reductase was ubiquitinated as determined by immunoprecipitation and blotting with the antiubiquitin antibody (Fig. 6B, lane 2) The total amount of DNA in each experiment was adjusted to 5 g/dish by addition of pcDNA3 mock vector. On day 3, cells were switched to medium A containing 5% LPDS, 1% hydroxypropyl-␤-cyclodextrin, and 50 M compactin. After incubation for 1 h at 37°C, the cells were washed and switched to the same medium (without hydroxypropyl-␤-cyclodextrin) containing 10 M MG-132 in the absence or presence of 1 g/ml 25-hydroxycholesterol plus 10 g/ml cholesterol (sterols) as indicated. After incubation for 2 h at 37°C, cells from 4 dishes were pooled, and the resulting digitonin-solubilized extracts were subjected to blue native-PAGE and immunoblot analysis with 5 g/ml monoclonal anti-Myc IgG for measurement of reductase-Insig-1 complex. Filter was exposed to film for 30 s at room temperature. Digitonin-solubilized extracts were also subjected to SDS-PAGE and immunoblot analysis with either 5 g/ml monoclonal anti-Myc IgG (against Insig-1) or 1:1000 dilution of monoclonal anti-T7 IgG (against reductase). Filters were exposed to films for 30 s. also provoked ubiquitination of reductase (lane 5), but this effect was not blocked by NB-598 (lane 7). In the presence of maximal levels of sterols, mevalonate had no additional effect on ubiquitination (lanes 6 and 8).
To determine the relation between ubiquitination and degradation of reductase, we made the same additions to SV589-pMev cells in the absence of MG-132 and measured the total amount of reductase by SDS-PAGE (Fig. 7C). Mevalonate at 10 mM caused the complete disappearance of reductase (Fig. 7C, lane 2) and this was blocked by NB-598 (lane 4). Exogenous sterols reduced the amount of reductase only partially (lane 5). Complete disappearance required the addition of mevalonate (lane 6), and this mevalonate effect was not blocked by NB-598 (lane 8), indicating that it did not require conversion of mevalonate to sterols.
To determine which of the nonsterol mevalonate-derived products is required for accelerated degradation of reductase, we tested the ability of externally added farnesol or geranylgeraniol (GG-OH) to reproduce the effects of mevalonate (Fig. 7). For these experiments, we used native SV-589 cells that lack the mevalonate transporter. Sterols added to these SV-589 cells led to a partial decrease in reductase protein after 5 h (Fig.  7A, lane 2). Addition of 10 mM mevalonate alone had no effect (lane 3), presumably because of the relatively slow uptake of mevalonate in these cells. However, mevalonate synergized with sterols to give a complete disappearance of reductase (lane 4). Farnesol at 10 or 100 M had no effect on reductase, either in the absence or presence of sterols (lanes [5][6][7][8]. In contrast, GG-OH at 10 M synergized with sterols (lane 10), and the effect was complete at 100 M (lane 12).
As discussed above, Insig proteins are required for the regulated degradation of reductase and for the ability of sterols to block SCAP-dependent processing of SREBPs. To determine whether mevalonate and GG-OH contribute to SCAP regulation as they do to reductase regulation, we added these compounds to cells in the absence or presence of sterols and then measured the amount of membrane-bound and nuclear SREBP-2 as well as membrane-bound reductase (Fig. 7B). As described previously, both mevalonate and GG-OH synergized with sterols in eliminating reductase protein from membranes (Fig. 7B, lanes 4 and 6). On the other hand, sterols alone blocked SREBP processing completely (lane 8). Mevalonate and GG-OH had no additional effect (lanes 10 and 12). To confirm that GG-OH was acting by enhancing reductase degradation, we performed a pulse-chase experiment (Fig. 7C). Sterols alone accelerated the degradation of reductase (compare lanes e-g with lanes b-d in Fig. 7C), but some 35 S-labeled reductase remained detectable at 5 h (lane g). GG-OH had little effect on degradation by itself (lanes i-k), but it synergized with sterols so that the reductase was markedly reduced at 1.5 h (lane l) and barely detectable at 3 h (lane m).

DISCUSSION
The data in this paper confirm and extend previous observations by others that eukaryotic HMG-CoA reductase is degraded by a ubiquitin-proteasome mechanism that is enhanced by sterols and nonsterol mevalonate-derived products (5,7,13). The new findings in the current paper are as follows: 1) ubiquitination and degradation of mammalian reductase is dependent upon Insigs, as revealed by the elimination of both processes when the two Insig mRNAs are removed through RNA interference (Fig. 2). 2) Ubiquitination occurs primarily on lysine 248 of reductase (Fig. 3). 3) Replacement of lysine 248 with arginine does not interfere with sterol-stimulated binding of reductase to Insigs (Fig. 5), but it does block the subsequent ubiquitination and degradation (Fig. 3). 4) Sterol-stimulated In B, the medium also contained 10 M MG-132. B, after incubation for 1 h at 37°C, cells were harvested, lysed, and sequentially subjected to immunoprecipitation, SDS-PAGE, and immunoblot analysis as described in the legend to Fig. 1. Filters were exposed to film for 2 min (top gel) and 30 s (bottom gel). C, following incubation at 37°C for 5 h, the cells were harvested and subjected to cell fractionation as described under "Experimental Procedures." Aliquots of membranes (3 g of protein) were subjected to SDS-PAGE, transferred to nylon membranes, and immunoblotted with 5 g/ml monoclonal IgG-A9 (against reductase). The filter was exposed to film for 1 min.
binding of reductase to Insigs requires the tetrapeptide YIYF that is also present in the sterol-sensing domain of SCAP (Fig.  5). 5) Mutations within the YIYF tetrapeptide prevent sterolmediated ubiquitination and degradation (Fig. 4). 6) GG-OH, but not farnesol, can synergize with sterols in accelerating degradation of reductase, suggesting that a geranylgeranylated protein may be involved (Fig. 7).
The data of Fig. 1B show that the kinetics of sterol-stimulated reductase ubiquitination in human fibroblasts are complex. The initial response is unexpectedly rapid, with ubiquitination detectable within 5 min of sterol addition, suggesting that the added sterols quickly reach the ER membrane. In the absence of a proteasome inhibitor, the amount of ubiquitinated reductase reaches a maximum at 10 -30 min, then declines dramatically by 60 min even though substantial reductase remains within the cell. The fall in ubiquitinated reductase is blocked by MG-132, implicating proteasomal degradation. But why is the remaining reductase not ubiquitinated? One explanation is that the rate of ubiquitination slows after 30 min; another is that the rate of de-ubiquitination increases. It is also possible that ubiquitination continues, but the partial decline in total reductase lowers the absolute amount of ubiquitinated reductase below the threshold of detection by the immunoblotting assay. Part of the difficulty in interpretation arises because only a small amount of reductase appears to be ubiquitinated at any instant in time, perhaps because of deubiquitinating enzymes operating in the cell.
Nonsterol, mevalonate-derived products are not capable of stimulating reductase degradation when cells are sterol-deficient, but they accelerate degradation when sterols are abundant (4,5,40). Two lines of evidence in the current study indicate that the major nonsterol product is geranylgeranyl pyrophosphate. First, the nonsterol acceleration is not blocked by the squalene epoxidase inhibitor NB-598 (Fig. 6), suggesting that it occurs at a branch prior to sterol synthesis. Second, the effect can be reproduced by addition to the culture medium of GG-OH, but not farnesol. Previous studies have shown that cells have a salvage pathway by which they can activate GG-OH to the metabolically active pyrophosphate (41). Although there is some suggestion that GG-OH can be converted to cholesterol, there is no known enzymatic pathway for this conversion. Moreover, the failure of NB-598 to inhibit the mevalonate effect suggests that sterol synthesis is not required. It seems more likely that GG-OH, after conversion to the pyrophosphate, is being incorporated into a geranylgeranylated protein that is required for maximally rapid reductase degradation. Possible candidates include geranylgeranylated Rab proteins, which play roles in vesicular transport (42). We cannot exclude the possibility that GG-OH itself, or its pyrophosphate derivative, is the actual regulator.
Our finding that GG-OH, but not farnesol, synergizes with sterols in accelerating degradation of mammalian reductase differs from previous studies. In one study farnesol, but not GG-OH, promoted degradation of reductase in permeabilized FIG. 7. Geranylgeraniol accelerates sterol-induced degradation of endogenous HMG-CoA reductase in SV-589 cells. A and B, SV-589 cells were set up on day 0 at 2 ϫ 10 5 cells/100-mm dish in medium B with 10% FCS. On day 2, the cells were refed medium B with 10% LPDS, 50 M compactin, and 50 M mevalonate. Following incubation for 16 h at 37°C, the cells were switched to medium B containing 10% LPDS and 50 M compactin in the absence (Ϫ) or presence (ϩ) of sterols (1 g/ml 25-hydroxycholesterol and 10 g/ml cholesterol), 10 mM mevalonate, and the indicated concentration of farnesol or GG-OH. After incubation for 5 h at 37°C, the cells were harvested and subjected to cell fractionation. A, aliquots of membrane fractions (5 g) were subjected to SDS-PAGE and immunoblot analysis as described in the legend to Fig.  6. Filters were exposed to film for 30 s. B, aliquots of membrane and nuclear extract fractions (3-12 and 20 g, respectively) were subjected to SDS-PAGE and immunoblot analysis as above. Filters were exposed to film for 30 s (top gel) and 1 min (bottom gel). P and N denote precursor and nuclear cleaved forms of SREBP-2, respectively. C, SV-589 cells were set up on day 0 at 2 ϫ 10 5 cells/60-mm dish in medium B with 10% FCS. On day 1, the cells were refed medium B with 10% LPDS, 50 M compactin, and 50 M mevalonate. Following incubation for 16 h at 37°C, cells were preincubated and pulselabeled with [ 35 S]methionine as described in the legend to Fig. 2. Cells were then chased in medium C with 0.5 mM methionine and 1 mM unlabeled cysteine in the absence (Ϫ) or presence (ϩ) of sterols (1 g/ml 25-hydroxycholesterol and 10 g/ml cholesterol) and 30 M GG-OH as indicated. Following the indicated chase period, cells were harvested, lysed, and subjected to immunoprecipitation with polyclonal anti-reductase and analyzed as described in the legend to Fig. 2. cells (43). However, subsequent studies showed that farnesol caused the reductase to become detergent-insoluble and thus resistant to immunoprecipitation, which masqueraded as an apparent increase in degradation (44). In the same study, farnesol synergized with sterols to accelerate reductase degradation in intact cells, but GG-OH was not tested and kinetic data were not reported.
The data of Figs. 4 and 5 implicate the YIYF sequence in reductase as a binding site for Insigs. We were attracted to the YIYF sequence because this sequence is conserved in the sterol-sensing domain of SCAP (25) and because the first tyrosine of this sequence is required for SCAP interaction with Insigs (19,20). Surprisingly, the first tyrosine in the reductase YIYF sequence is not required for sterol-stimulated ubiquitination and degradation, but the isoleucine and the second tyrosine are required (IY sequence, Fig. 4). Replacement of the entire YIYF sequence with AAAA blocks reductase binding to Insigs (Fig.  5), which explains the lack of ubiquitination and degradation. Whether the IY sequence in SCAP is required for its interaction with Insigs is currently being explored.
The conservative substitution of arginine for lysine 248 in reductase does not affect Insig binding (Fig. 5), but it abolishes detectable reductase ubiquitination and delays degradation (Fig. 3). A similar replacement of lysine 89, the only other lysine in the cytoplasmic loops of reductase, has little effect on its own, but it further delays degradation when lysine 248 is also replaced (Figs. 3 and 5). These findings indicate that lysine 248 is the primary site of reductase ubiquitination and that lysine 89 can become ubiquitinated to a small degree when lysine 248 is absent. It is striking that both of these lysines are predicted to lie at the COOH-terminal ends of cytoplasmic loops where they join membrane-spanning helices (see diagram of Fig. 3). This juxtamembrane location raises the possibility that ubiquitin is transferred to reductase by a membranebound ubiquitin transferase. This would resemble the situation in S. cerevisiae where, Hrd1p, a membrane-bound ubiquitin transferase, is believed to attach ubiquitins to reductase, earmarking it for degradation (12). In contrast to the mammalian system, the yeast ubiquitination reaction appears to be stimulated primarily by a nonsterol mevalonate metabolite, and not a sterol (12). Moreover, the yeast genome does not contain an easily recognizable counterpart of Insigs.
A major remaining puzzle is how Insigs can interact with similar YIYF sequences in the membrane domains of both SCAP and HMG-CoA reductase while producing dramatically different consequences. When SCAP binds to Insigs, it is not degraded, but rather it is retained in the ER (19). Furthermore, we have been unable to find evidence that SCAP is ubiquitinated in response to sterols. 3 In contrast, the Insig-reductase interaction leads to ubiquitination and degradation. Further progress in understanding these important differences will come through the identification of other proteins that are recruited to the Insig-reductase and Insig-SCAP complexes that form in response to sterols.