2342 Research Article

IntroductionThe endoplasmic reticulum (ER) is the center of biosynthesis,folding and quality control of a variety of secretory proteinsand secretory pathway-localized proteins, including membraneproteins (Sitia and Braakman, 2003). Perturbations that disruptprotein folding lead to an accumulation of unfolded proteins orprotein aggregates in the ER, and eukaryotic cells have severaldistinct mechanisms to cope with these situations (Mori, 2000;Kaufman et al., 2002): (1) retention of unfolded proteins in theER by the ER-Golgi retrieval pathway, (2) upregulation of ERchaperones, the so-called unfolded protein response, (3)transient repression of de novo protein biosynthesis to maintainthe fidelity of protein folding (attenuation), and (4) removal ofmisfolded proteins by the proteasome-dependent ER-associated degradation (ERAD) pathway. In the process ofprotein elimination by proteasome-dependent ERAD, proteinsubstrates are polyubiquitylated by the E2 ubiquitin-conjugating enzymes and E3 ubiquitin ligases, then retro-translocated to the cytosol by the concerted action of AAA-ATPase Cdc48p/p97 and the cofactor Npl4p-Ufd1p complexthrough the retrograde translocation channel, presumablyincluding Sec61p, and finally degraded by proteasomes(Ellgaard and Helenius, 2003; Hampton, 2002). Recentexperiments in yeast revealed that the ER membrane proteinUbx2p recruits the Cdc48p complex to ubiquitin ligases andERAD substrates (Schuberth and Bachberger, 2005; Neuber etal., 2005). In mammals, the ER membrane protein VIMPrecruits the p97-Ufd1-Npl4 complex to the putative dislocationchannel protein derlin-1 (Ye et al., 2004; Lilley and Ploegh,2004). The proteasome-dependent ERAD system is used not

only for the removal of unfolded or unassembled membraneproteins, such as the �F508 mutant of the cystic fibrosistransmembrane conductance regulator or T-cell receptor �-subunit (TCR�), but also for degradation of the ER-residentnormal proteins such as hydroxymethylglutaryl-CoA reductase(HMGR), Ubc6p, and yeast �9-fatty acid desaturase (Walter etal., 2001; Hampton, 2002; Braun et al., 2002).

Mammalian �9 stearoyl-CoA desaturase 1 (SCD1) is ashort-lived integral membrane protein of the ER (Oshino andSato, 1972). It is a key enzyme in the biosynthesis of mono-unsaturated fatty acids and catalyzes the introduction of the cisdouble bond at the �9 position of palmitoyl- or stearoyl-CoAto form palmitoleyl- or oleyl-CoA, respectively, by theoxygenase reaction, in which the reducing equivalents areprovided by NADH via NADH-cytochrome b5 reductase andcytochrome b5 (Shimakata et al., 1972; Strittmatter et al.,1974). A key regulator of SCD1 gene expression is sterol-regulatory element-binding protein-1c (SREBP-1c), whichmediates insulin-induced transcriptional activation of theSCD1 gene (Ntambi, 1999; Bene et al., 2001; Horton et al.,2002). SCD1 protein is induced 50- to 100-fold when dietaryfat is restricted, whereas it rapidly decreases to undetectablelevels with a half-life of 3-4 hours after the restricted-fatdietary regimen is stopped (Oshino and Sato, 1972).

In yeast, the abundance of yeast �9 fatty acid desaturase,Ole1p, in the ER is transiently regulated at the genetranscription level and by degradation of Ole1p. OLE1 isactivated by the OLE pathway, in which transcription factorsSpt23p and Mga2p are synthesized as ER-bound precursorproteins. Upon unsaturated fatty acid starvation, they are

Mammalian ��9 stearoyl-CoA desaturase 1 (SCD1) is a keyenzyme in the biosynthesis of mono-unsaturated fatty acidsin the endoplasmic reticulum (ER). It is a short-livedmultispanning ER membrane protein, reported to bedegraded by the ubiquitin-proteasome-independentpathway. We have examined SCD1 protein degradationusing cultured mammalian cells. Exogenously expressedSCD1 in CHO-K1 cells was localized to the ER and turnedover with a half-life of ~3 hours. Unexpectedly, proteasomeinhibitors increased the half-life of SCD1 to ~6 hours.Endogenously expressed SCD1 in adipocyte-differentiatedNIH 3T3-L1 cells was also rapidly degraded in aproteasome inhibitor-sensitive manner. In the presence ofproteasome inhibitors, polyubiquitylated SCD1

accumulated in the ER and interacted with AAA-ATPasep97, which is involved in ER-associated degradation(ERAD). The 66-residue N-terminal segment carrying thePEST sequence is mainly responsible for SCD1degradation and this segment induced instability in anotherwise stable ER membrane protein. Furthermore,SCD1 was degraded constitutively irrespective of thecellular levels of unsaturated fatty acids, which strictlyregulate SCD1 gene expression. These findings indicatethat the ubiquitin-proteasome-dependent ERAD system isalso involved in constitutive SCD1 degradation.

Key words: Stearoyl-CoA desaturase, Endoplasmic reticulum,Proteasome, Ubiquitylation, ERAD, Fatty acid desaturation

Summary

Ubiquitin-proteasome-dependent degradation ofmammalian ER stearoyl-CoA desaturaseHiroki Kato, Kenjiro Sakaki and Katsuyoshi Mihara*Department of Molecular Biology, Graduate School of Medical Science, Kyushu University, Fukuoka 812-8582, Japan*Author for correspondence (e-mail: mihara@cell.med.kyushu-u.ac.jp)

Accepted 21 February 2006Journal of Cell Science 119, 2342-2353 Published by The Company of Biologists 2006doi:10.1242/jcs.02951

Jour

nal o

f Cel

l Sci

ence

2343Ubiquitin-proteasome-dependent degradation of stearoyl-CoA desaturase

activated by a regulated ubiquitin-proteasome-dependentproteolysis, and translocated to the nucleus to activate OLE1transcription (Rape et al., 2001). Furthermore, Ole1p isnaturally short-lived and is degraded by ERAD (Braun et al.,2002).

Mammalian SCD1 is degraded by a proteasome-independent pathway (Ozols, 1997). Mziaut et al. (Mziaut etal., 2000) demonstrated that SCD1 fused to the N-terminus ofgreen fluorescent protein (GFP), when expressed in CHO-K1cells, was correctly targeted to the ER and, in contrast to GFP,became unstable in the ER and degraded with a half-life of afew hours. Mziaut et al. also demonstrated that deletion of the26- or 44-residue N-terminal segment from the SCD1-GFPfusion protein led to stabilization of the truncated proteins. Inturn, the 1 to 33-residue N-terminal segment transplanted tothe N-terminus of the GFP rendered the reporter extremelyshort-lived. They thus concluded that the ~30-residue N-terminal segment of SCD1 constitutes a motif responsible forthe rapid degradation of SCD1. In a recent report, Ozols andcollaborators characterized the protease responsible for SCD1degradation using an in vitro assay system with SCD1-inducedrat liver microsomes (Ozols, 1997; Heinemann and Ozols,1998), and purified a plasminogen (PLG)-like protease as theSCD1-specific protease (Heinemann et al., 2003b). SCD1protease is presumably in the 90 kDa proform, and itsconversion to a series of smaller proteins markedly increasesits proteolytic activity (Heinemann et al., 2003b). In vitrodegradation of SCD1 is not inhibited by a wide variety ofprotease inhibitors; lysosomotrophic agents (chloroquine,leupeptin and pepstatin A), the calpain inhibitor N-acetyl-leucyl-methionyl, and the proteasome inhibitor lactacystin donot inhibit the in vitro degradation of SCD1 in isolatedmicrosomal membranes (Ozols, 1997). The responsibleprotease appears to be a serine or thiol protease, because SCD1is sensitive to dithiothreitol, diisopropyl phosphofluoridate, andleupeptin (Heinemann et al., 2003a; Heinemann et al., 2003b).In spite of these findings, however, there is limited informationon SCD1 degradation in vivo.

SCD1 is a typical PEST sequence-rich membrane protein(Rogers et al., 1986). Generally, PEST sequence-containingproteins are targets of rapid protein degradation (Rechsteinerand Rogers, 1996). Although the pathways (or enzymes)responsible for degrading PEST proteins remain controversial,a considerable body of evidence supports the idea that PESTsequences target proteins for degradation by the ubiquitin-proteasome system (Rechsteiner and Rogers, 1996). In light ofthe apparently distinct mechanisms for degradation of SCDproteins between yeast and mammalian cells, we used culturedmammalian cells to examine whether SCD1 is degradedexclusively by the ubiquitin-independent pathway, or whetherthe ubiquitin-dependent pathway is also involved in parallel.

Here, we demonstrate that exogenously expressed SCD1 inCHO-K1 cells, as well as the endogenously expressed form inadipocyte-differentiated 3T3-L1 cells, was degraded rapidlywith a half life of less than 3.5 hours, and this degradation wassignificantly inhibited by the proteasome inhibitors, MG132,and epoxomicin (P<0.01). Consistent with these findings, thepolyubiquitylated SCD1 intermediates accumulated in the ERin the presence of MG132, and interacted with AAA-ATPasep97, which is required for the retro-translocation of the ERAD(Ellgaard and Helenius, 2003), thus indicating involvement of

the ERAD pathway in controlling cellular levels of mammalianSCD1, as is the case for yeast Ole1p (Braun et al., 2002).Furthermore, we demonstrated that the 66-residue N-terminalsegment of SCD1 containing the PEST sequences has animportant role in proteasomal degradation of ER-localizingproteins.

ResultsSCD1 is degraded by a proteasome-dependent pathwayUsing a cell-free assay system, Ozols and collaboratorsdemonstrated that SCD1 protein is degraded through aproteasome-independent pathway, although in vivo evidencehas been lacking (Ozols, 1997; Heimemann and Ozols, 1998).We, therefore, addressed this point in cultured mammaliancells. CHO-K1 cells transiently expressing HA-tagged SCD1were subjected to pulse-chase analysis with [35S]methionine,followed by immunoprecipitation using an anti-HA antiserum.In the absence of the proteasome inhibitor MG132, SCD1protein was degraded with a half life of ~3 hours (Fig. 1A). Inthe presence of MG132, the half-life of SCD1 increased to ~6hours. This effect was also observed with another proteasomeinhibitor, epoxomicin (Fig. 1B). Similarly, exogenouslyexpressed SCD1 was degraded rapidly in HeLa cells, andMG132 inhibited the degradation (Fig. 1A). The same resultswere obtained in HEK-293 cells (data not shown). As a control,the degradation of TCR�, a typical ERAD substrate (Huppaand Ploegh, 1997), was also efficiently inhibited by MG132(t1/2<2 hours in the absence of the inhibitor, and t1/2>8 hoursin the presence of the inhibitor; Fig. 1A). Heinemann et al.(Heinemann et al., 2003a) demonstrated that SCD1degradation is inhibited by leupeptin in vitro. By contrast,however, in our study the stability of SCD1 in vivo was notaffected by an even higher concentration (100 �M) ofleupeptin (Fig. 1B). These results suggest that proteasomes areinvolved in the degradation of SCD1 protein in cultured cells.

Endogenously expressed SCD1 protein is alsodegraded by a proteasome-dependent pathwayTo exclude the possibility that the rapid degradation of SCD1and its inhibition by proteasome inhibitors were due toexogenous overexpression of SCD1 in CHO-K1 cells, weanalyzed the metabolic stability of endogenously expressedSCD1 protein in adipocyte-differentiated NIH 3T3-L1 cells(Green and Meuth, 1974). NIH 3T3-L1 cells expressed SCD1during differentiation into adipocytes (Fig. 2A). The adipocyte-differentiated cells were subjected to pulse labeling with[35S]methionine and chase reactions, followed byimmunoprecipitation with anti-SCD1 antiserum (Fig. 2B). Theendogenously expressed SCD1 was degraded with a half-lifeof ~3.5 hours, and MG132 effectively inhibited thisdegradation (t1/2~7.5 hours; Fig. 2B). Together, these resultsindicate that proteasomes are involved in the degradation ofendogenous SCD1 protein.

Accumulation of ubiquitylated SCD1 in the presence ofMG132We then examined whether SCD1 is modified by polyubiquitinchains, using immunoprecipitation. CHO-K1 cells transientlytransfected with empty plasmid or plasmids harboring SCD1-HA cDNA were cultured in the presence or absence of MG132.They were lysed in buffer containing SDS and the cell lysates

Jour

nal o

f Cel

l Sci

ence

2344

were subjected to immunoprecipitation using anti-HA serumand subsequent immunoblot analysis with anti-polyubiquitinmonoclonal antibody (Fig. 3A, lanes 1-5). PolyubiquitylatedSCD1-HA was detected in the presence of MG132 (Fig. 3A,lane 5). Immunoblot analysis with anti-HA antiserum revealedthat the steady state level of SCD1-HA increased in thepresence of MG132 (Fig. 3A, compare lanes 9 and 10).However, only a faint band of polyubiquitylated SCD1-HAwas detected with anti-HA antiserum (Fig. 3A, lower panel);detection of ubiquitin-modified proteins at steady state seemedto be difficult, as reported previously (Liao et al., 1998). Thesame results were obtained for endogenously expressed SCD1in differentiated 3T3-L1 cells (Fig. 3B); a polyubiquitylatedband for the immunoprecipitate with anti-SCD1 peptideantibodies was detected only for cells grown in the presence ofMG132 (lane 6), but not in its absence (lane 5; DMSO). Theimmunoprecipitates with control IgGs did not have suchsignals, irrespective of the presence or absence of MG132 (Fig.3B, lanes 1-3), indicating that this ubiquitylation was not dueto exogenous overproduction of SCD1 protein. To furtherconfirm the polyubiquitylation of SCD1, SCD1-HA and Myc-tagged ubiquitin were coexpressed in the presence of MG132in CHO-K1 cells, and the cell lysates were subjected to

Journal of Cell Science 119 (11)

immunoprecipitation using anti-Myc IgGs, which were thensubjected to immunoblotting with anti-HA antibodies (Fig.3C). Again, the polyubiquitylated bands were detected for theimmunoprecipitates processed with anti-HA antibodies (Fig.3C, lane 4). The lysates prepared from the cells that weretransfected with either empty plasmid or plasmids harboringSCD1-HA cDNA or Myc-ubiquitin cDNA did not show anyubiquitylation, even in the presence of MG132 (Fig. 3C, lanes1-3). Taken together, these results clearly indicate that SCD1protein is degraded via a ubiquitin-proteasome pathway.

Ubiquitinated SCD1 accumulates in the ER andinteracts with p97Typical proteasome-dependent ERAD substrates areubiquitylated on the ER and retro-translocated from the ER tothe cytoplasm by the p97-Ufd1-Npl4 complex (Meyer et al.,2002; Ye et al., 2003). We therefore examined whether SCD1degradation depends on the ERAD system. CHO-K1 cellstransfected with either pcDNA3.1 or plasmids harboringSCD1-HA cDNA were cultured in the presence or absence ofMG132 and fractionated into supernatant and microsomalfractions, which were subjected to immunoprecipitation usinganti-HA antiserum. Immunoblot analysis of the

Fig. 1. Degradation of exogenously expressed SCD1 is inhibited by proteasome inhibitors. CHO-K1 cells or HeLa cells were transfected withplasmids harboring SCD1-HA or TCR�-flag cDNAs. After 48 hours culture, the cells were pulsed with [35S]methionine in the presence ofDMSO or 50 �M MG132 (A), and 20 or 100 �M leupeptin or 2 �M epoxomicin (epox) (B) for 30 minutes, then subjected to the chasereaction under the indicated conditions for the indicated times. Proteins (150 �g) were immunoprecipitated using anti-HA (for SCD1-HA) oranti-FLAG (for TCR�-FLAG) antibodies and analyzed with an SDS-PAGE and BAS-2500 imager. Band intensities corresponding to SCD1-HA and TCR�-flag were quantified by Image Gauge and shown by percentages of the signals setting those at time zero as 100%. The resultsshown are the average of three independent experiments. Asterisks: nonspecific immunoprecipitation bands.

Jour

nal o

f Cel

l Sci

ence

2345Ubiquitin-proteasome-dependent degradation of stearoyl-CoA desaturase

immunoprecipitates with anti-polyubiquitin monoclonalantibody revealed that the polyubiquitylated SCD1-HA wasenriched in the microsomal fraction, but not in the supernatantfraction (Fig. 4A, compare lanes 11 and 12). By contrast, thepolyubiquitylated smear was undetectable in cells transfectedwith empty plasmid, irrespective of the presence or absence ofMG132 (Fig. 4A, lanes 5-8), or detected only weakly in themicrosomes of cells expressing SCD1-HA in the absence ofMG132 (lane 10). Furthermore, recovery of SCD1-HA in theER fraction clearly increased in the presence of MG132(compare lanes 23 and 25), confirming the results in Fig. 3A.

Of note, p97 co-immunoprecipitated with SCD1-HA fromthe microsomal fraction of the SCD1-HA-expressing cells thatwere grown in the presence of MG132 (Fig. 4A, lane 38),whereas the signal was not detectable in the microsomes ofcells grown under all the other conditions (Fig. 4A, lanes 32,34, and 36). Approximately 1.3% of the total amount of p97

was recovered to the ER of the cells expressing SCD1-HA (Fig.4B). We cannot, however, rigorously rule out the trivialpossibility that p97 became associated with SCD1-HA aftermembrane solubilization. Taken together, these results stronglysuggest that SCD1 is degraded by proteasome-dependentERAD.

The N-terminal segment is important for SCD1degradation by the ubiquitin-proteasome systemThe rapid turnover of the SCD1-GFP fusion protein in CHO-K1 cells is compromised by the deletion of either the 27- or45-residue N-terminal segment of SCD1 (Mziaut et al., 2000).Furthermore, appending the 1 to 33-residue N-terminalsegment of SCD1 onto the N-terminus of GFP destabilizes thefusion protein, although it is not targeted to the ER but remainsin the cytoplasm, indicating that the ~30-residue N-terminalsegment fused to a reporter protein functions as a rapiddegradation signal, even if the fusion protein remains in thecytoplasm (Mziaut et al., 2000). A web-based algorithm PEST-Find (http://www.at.embnet.org/embnet/tools/bio/PESTfind)revealed PEST sequences in the N-terminus of SCD1 (Fig.5A), suggesting that the N-terminal region has an importantrole in proteasome-dependent SCD1 degradation. Weexamined this hypothesis using a 66-amino acid N-terminal-truncated SCD1 mutant (�66 SCD1). When expressed inCHO-K1 cells, �66 SCD1 (Fig. 5B) localized to the ER, justlike the wild-type protein (Fig. 5C). In pulse-chaseexperiments, �66 SCD1 was stable compared with wild-typeSCD1 (Fig. 5D, compare left and right panels), confirming theprevious report by Mziaut et al. (Mziaut et al., 2000). Inaddition, the residual degradation reaction of �66 SCD1 wasinsensitive to epoxomicin (Fig. 5D, right), indicating that theN-terminal segment of SCD1 is important for proteasome-dependent degradation of SCD1.

The 66-residue N-terminal segment of SCD1destabilizes other ER-localized membrane proteinsTo further examine the function of the N-terminal region ofSCD1, we constructed fusion proteins in which the 66-residueN-terminal segment was fused to the N-terminus of EGFP,SCD66-EGFP. When expressed in CHO-K1 cells, EGFP andSCD66-EGFP both localized to the cytoplasm (Fig. 6A) anddegraded slowly, and both degradations were epoxomicin-insensitive (Fig. 6B). Analysis of the stability of the EGFPconstructs including SCD33-EGFP by FACScan confirmedthese results (Fig. 6C). These observations were in markedcontrast to the findings of Mziaut et al. (Mziaut et al., 2000),who reported that the fusion construct in which the 1 to 33-residue N-terminal segment of SCD1 was fused to the N-terminus of EGFP (SCD33-EGFP), and was extremely short-lived (t1/2 ~1.5 hour) in CHO-K1 cells. We have no adequateexplanation for this discrepancy. It should be noted, however,that the expression efficiency as assessed by flow cytometrywas 40% for EGFP and 15% for SCD1wt-EGFP: similar levelsto those reported by Mziaut et al. (Mziaut et al., 2000); 45.8%for the former and 14.4% for the latter). By contrast, theexpression efficiency was distinctly different for SDC33-EGFP: 35% in the present study, but 6.2% in the previousreport.

We next examined whether the ER localization signaldestabilizes the SCD66-appended substrates. For this purpose,

Fig. 2. Endogenously expressed SCD1 is also degraded byproteasomes. NIH 3T3-L1 cells were grown in differentiationcocktail containing methylisobutylxanthine, dexamethasone, andinsulin. After 10 d, the cells were collected and subjected to (A)SDS-PAGE and immunoblot analysis using anti-SCD1 antibody, and(B) pulse-chase analysis using [35S]methionine in the absence(DMSO) or presence of 50 �M MG132, as described in Fig. 1 andMaterials and Methods. Proteins were immunoprecipitated by anti-SCD1 antibody and analyzed by SDS-PAGE and subsequent digitalautoradiography using BAS2500. Other conditions were the same asin Fig. 1.

Jour

nal o

f Cel

l Sci

ence

2346 Journal of Cell Science 119 (11)

Fig. 3. Polyubiquitylated SCD1 accumulates in the presence of MG132. (A) CHO-K1 cells were transfected with pcDNA3.1 or the plasmidharboring SCD1-HA and grown for 48 hours. DMSO or 50 �M MG132 was then added to the cells and culture was continued for 5 hours. Thecells were solubilized with buffer containing SDS (see Materials and Methods) and subjected to immunoprecipitation using anti-HA antibodies.The immunoprecipitates were analyzed by SDS-PAGE and subsequent immunoblotting with anti-polyubiquitin antibodies (upper panel) or anti-HA antibodies (lower panel). Lane 1 (IP control): immunoprecipitation without cell lysate. (B) NIH 3T3-L1 cells were grown in thedifferentiation cocktail for 10 days, DMSO or 50 �M MG132 was added, and culture continued for 5 hours. The cells were subjected toimmunoprecipitation using anti-SCD1 C-terminal peptide antibodies or preimmune IgG and the immunoprecipitates were analyzed by SDS-PAGE and subsequent immunoblotting with anti-polyubiquitin (upper panels) or SCD1-C-peptide antibodies (lower panels). Lane 1 (IPcontrol): immunoprecipitation without cell lysate. (C) CHO-K1 cells were transfected with the indicated plasmids and grown for 48 hours.MG132 (50 �M) was added to the cells and culture was continued for 5 hours. The cells were solubilized using SDS-buffer and subjected toimmunoprecipitation using anti-Myc antibody. The immunoprecipitates were analyzed by SDS-PAGE and subsequent immunoblotting withanti-HA antibody. Magic Mark XP (Invitrogen) containing IgG-binding sequence-tagged proteins was used as molecular mass markers;detected by immunoblotting.

Jour

nal o

f Cel

l Sci

ence

2347Ubiquitin-proteasome-dependent degradation of stearoyl-CoA desaturase

EGFP was fused to the C terminus of cytochrome b5, the C-terminal tail anchored ER membrane protein, to construct b5-EGFP. Furthermore, SCD66 was fused to the N-terminus ofb5-EGFP to create SCD66-b5-EGFP. When expressed inCHO-K1 cells, SCD66-b5-GFP and b5-GFP both co-localized

with the ER marker protein calnexin (Fig. 7A). Analysis of themetabolic stability of these fusion proteins by [35S]methioninepulse-chase experiments, indicated that the turnover rate of b5-EGFP was relatively slow (Fig. 7B, left panel). By contrast,SCD66-b5-GFP was rapidly degraded with a half life of 4hours, and the degradation was significantly inhibited byepoxomicin (Fig. 7B, right panel) (P<0.05). Taken together,these results indicated that the 66-residue N-terminal segmentof SCD1 functions as a degradation signal on the ER.

SCD1 degradation is not affected by cellular levels ofunsaturated fatty acids or sterolsSCD1 gene transcription is regulated by unsaturated fattyacids, and exogenous polyunsaturated fatty acids repressSCD1 gene transcription (Ntambi, 1995; Waters and Ntambi,1996). Furthermore, the SCD1 gene is SREBP-responsivegene that is involved in cholesterol homeostasis (e.g. HMG-CoA synthase, HMGR, squalen synthase or low densitylipoprotein receptor), fatty acid synthesis (e.g. fatty acidsynthase, acetyl-CoA carboxylase) and fatty acid desaturation(Shimano, 2001), and transcription of these genes is regulatedby cellular sterols; transcription is repressed by excess sterols,and induced by sterol depletion (Tabor et al., 1999). Inparticular, cellular sterol levels exert feed-back regulation onHMGR in the ER through multiple mechanisms; high sterollevels shut-down transcription of the HMGR geneconcomitant with accelerated protein degradation (Goldsteinand Brown, 1990). The HMGR gene should also be regulatedby unsaturated fatty acids, given that SREBP isoforms aredown-regulated by unsaturated fatty acids (Hannah et al.,2001). The degradation of HMGR is accelerated by the sterol-induced binding of its sterol-sensing domain to Insig-1 (Severet al., 2003). The apparent similarity of the regulation of geneexpression between the HMGR and SCD1 genes led us toexamine whether SCD1 degradation is regulated by cellularlevels of sterols or polyunsaturated fatty acids. For thispurpose, the SCD1-HA-transfected HEK-293 cells (CHO-K1cells could not be used in this experiment because mediumcontaining linoleic acid must be used for the culture) werecultured in the presence of either lipid-depleted fetal bovineserum (FBS) or untreated FBS, and the turnover rate of SCD1-HA was examined after pulse-chase analysis using[35S]methionine (Fig. 8). Real-time PCR revealed that genetranscription for HMGR, low density lipoprotein receptor andSCD1 was induced in response to the presence of delipidatedFBS, whereas transcription of the glyceraldehyde 3-phosphatedehydrogenase (GAPDH) gene, which is unrelated to lipidbiosynthesis, was not affected (Fig. 8A). The turnover ofSCD1, however, was not altered in the presence of delipidatedFBS (Fig. 8B,C). Similarly, overexpression of Insig-1 orreplenishment of polyunsaturated fatty acids in cultured CHO-K1 cells did not significantly affect the SCD1 degradation rate(data not shown).

These results suggested that rapid degradation of SCD1 isnot affected by either polyunsaturated fatty acids or sterols, andproceeds constitutively. Taken together, the data demonstratedthat SCD1 is a naturally short-lived protein that is degradedconstitutively by the ubiquitin-proteasome pathway. Thus,cellular levels of SCD1 seem to be determined by the SCD1gene transcription rate and SCD1 mRNA stability (Ntambi,1995; Sessler et al., 1996).

Fig. 4. Polyubiquitylated SCD1 accumulates in the ER in thepresence of MG132. (A) CHO-K1 cells were transfected withpcDNA3.1 or the plasmid harboring SCD1-HA cDNA, and grownfor 48 hours. DMSO or 50 �M MG132 was then added to the cellsand the culture continued for 5 hours. Post-nuclear supernatant(PNS) was prepared from the cells, and then fractionated into thesupernatant (S) and membrane (P) fractions. Both fractions weresubjected to immunoprecipitation using anti-HA antibody and theimmunoprecipitates were then analyzed by SDS-PAGE andsubsequent immunoblotting using anti-polyubiquitin antibody (upperpanel). The filter was re-blotted with anti-HA antibody (middlepanel), or anti-p97 antibody (lower panel). Lanes 1-4: 5% of PNSprepared from the cells grown under the indicated conditions weredirectly applied. Lane 13 (IP control): immunoprecipitation withoutcell lysate. (B) Quantification of p97 recovered to the ER of theSCD1-HA-expressing cells. Immunoprecipitation was performed asdescribed in A, the gel was analyzed by digital autoradiography, andband intensities were calculated setting that of the total amount ofp97 as 100%. The average of three independent experiments isshown.

Jour

nal o

f Cel

l Sci

ence

2348

Exogenous expression of plasminogendoes not influence SCD1 proteindegradation in cultured cellsSCD1 is degraded by plasminogen (PLG)-likeprotein in rat liver microsomes (Heinemann etal., 2003b). Immunoblot analysis using anti-human PLG antibodies, which gave strongsignals for Chinese hamster liver extracts or itsmicrosomes, or recombinant human PLG-flag,revealed that PLG was scarcely expressed inHeLa cells, CHO-K1 cells or HEK-293 cells(Fig. 9A). We therefore examined whetherexogenously expressed PLG influences SCD1degradation in cultured cells. SCD1-HA andflag-tagged human PLG (PLG-flag) were co-expressed in CHO-K1 cells, and the steady-statelevel and turnover rate of SCD1-HA were bothmeasured. The expressed PLG-flag localizedmainly in P1, composed of mitochondria andheavy microsomes, and P2, composed ofmicrosomes, indicating that significant amountsof the expressed PLG-flag localized within thesecretory organelles (Fig. 9B). We thenexamined if the steady-state level of SCD1-HAis influenced by co-expression with increasingamounts of PLG-flag. There were no clear dose-dependent decreases, however, in SCD1-HA(Fig. 9C). Similarly, [35S]methionine pulse-chase experiments revealed that SCD1-HAdegradation was not influenced by co-expressionwith PLG-flag (Fig. 9D).

Knockdown of PLG does not affect SCD1protein degradation in HepG2 cellsIn contrast to the cells examined above, wefound that PLG was expressed to a detectablelevel in HepG2 cells (Fig. 9E). Therefore, weexamined the effect of PLG-knockdown onSCD1 turnover in HepG2 cells. Exogenouslyexpressed SCD1-HA was degraded in controlHepG2 cells with a half-life similar to those inHeLa cells or CHO-K1 cells. Unexpectedly,degradation rate of SCD1-HA was not affectedeven after ~93% depletion of PLG (Fig. 9F,G).Together, these results indicate that PLG is notinvolved in SCD1 degradation, at least incultured cells.

DiscussionOzols and collaborators developed an in vitro assay for SCD1degradation using microsomes isolated from SCD1-induced ratliver, and demonstrated that the degradation is rapid (half-lifeof ~5 hours), selective for SCD1 and proceeds by a non-proteasome pathway; the reaction is not inhibited by eitherlysosomal protease inhibitors, calpain inhibitors or proteasomeinhibitors (Ozols, 1997; Heinemann and Ozols, 1998). Usingthis system, they purified a ~90 kDa microsomal endopeptidaseas the SCD1 protease, which exhibited serine or thiol proteasecharacteristics, as it was inhibited by diisopropylphosphofluoridate, dithiothreitol or leupeptin (Heinemann etal., 2003a). The ~90 kDa form seems to be a proform, because

Journal of Cell Science 119 (11)

it is degraded to smaller-size fragments during purification,concomitant with an increase in its activity (Heinemann et al.,2003b). The protease was identified as a microsomal form ofPLG, because the reaction is inhibited by a plasmin-specificinhibitor, Bdellin (Heinemann et al., 2003b). In this regard, invitro degradation of SCD1 in liver microsomes isolated fromPLG-deficient mice was markedly compromised (Heinemannet al., 2003b).

To our knowledge, however, there are no in vivo data on thecellular stability of SCD1; even data on the effects ofproteasome inhibitors have been lacking. Therefore, thequestion has remained as to whether the microsomal form of

Fig. 5. The N-terminal PEST sequence-containing region of SCD1 is responsiblefor the proteasomal degradation. (A) The PEST sequence regions in the N-terminalsegment of SCD1 predicted by the PEST-Find algorithm (shown in red characters).(B) CHO-K1 cells were transfected with plasmids carrying SCD1-HA cDNA or�66 SCD1-HA cDNA and grown for 48 hours. The cells were lysed with SDSbuffer and analyzed by SDS-PAGE and subsequent immunoblotting with anti-HAantibody. (C) Indirect-immunofluorescence microscopy. Fluorescence images ofSCD1-HA and �66 SCD1-HA (green) and calnexin (red) were taken using aconfocal microscope. Merged images are also shown. Bar, 10 �m. (D) The cellsobtained above were subjected to pulse-chase analysis with [35S]methionine in thepresence of DMSO or 2 �M epoxomicin. At the indicated time points, proteinswere immunoprecipitated using anti-HA antibody, resolved by SDS-PAGE, andanalyzed using a BAS2500 Image Analyzer. Band intensities were quantified byImage Gauge and are shown as the percentage of the signal at time zero. The resultsshown are the average of three independent experiments.

Jour

nal o

f Cel

l Sci

ence

2349Ubiquitin-proteasome-dependent degradation of stearoyl-CoA desaturase

plasminogen is the enzyme solely responsible for the rapiddegradation in vivo. In this study, we examined the cellularstability of SCD1 using cultured mammalian cells anddemonstrated that there is a ubiquitin-proteasome-dependentpathway for SCD1 degradation. The degradation of SCD1 wasinhibited by proteasome inhibitors, and as such, the level ofmicrosomal SCD1 was increased. Furthermore,polyubiquitylated SCD1 accumulated in the ER in the presenceof proteasome inhibitors and interacted with AAA-ATPasep97, which is essential for proteasome-dependent ERAD.These results suggest that SCD1 is degraded by proteasome-dependent ERAD.

We addressed the involvement of a PLG-like protease inthe SCD1 degradation in cultured cells. The reportedsequence of SCD1 protease spans the entire PLG sequenceexcept for the 68-residue N-terminal segment of the secretedform (Heinemann et al., 2003b), although neither thetopogenic sequence nor the membrane topology have beenreported. Consistent with this notion, knockout of the PLGgene influenced SCD1 degradation as examined by an in vitrodegradation assay (Heinemann et al., 2003b). We, therefore,

Fig. 6. The 66-residue N-terminal segment of SCD1 appended to asoluble reporter protein fails to stimulate degradation. (A) CHO-K1cells were transfected with the plasmids harboring the indicatedcDNAs, grown for 48 hours, and subjected to immunofluorescencemicroscopy. Fluorescence images of EGFP (green) and calnexin(red) were taken using a confocal microscope. Merged images arealso shown. Bar, 10 �m. (B) The cells grown as above weresubjected to pulse-chase analysis. Proteins were immunoprecipitatedby anti-EGFP antibody. Other conditions were as described in Fig.5D. (C) CHO-K1 cell were transfected with EGFP, SCD66-EGFP,SCD33-EGFP, or SCD1wt-EGFP constructs. After 48 hours, thetransfected cells were treated with cycloheximide (10 �g/ml) for 0,2, and 4 hours and their fluorescence was determined using flowcytometric analysis. The results shown are the average of threeindependent experiments.

Fig. 7. The 66-residue N-terminal region of SCD1 makes anotherwise stable ER membrane protein unstable. (A) CHO-K1 cellswere transfected with plasmids carrying b5-EGFP cDNA or SCD66-b5-EGFP cDNA, grown for 48 hours, and analyzed byimmunofluorescence microscopy. Fluorescence images of b5-EGFPand SCD66-b5-EGFP (green) and calnexin (red) were taken using aconfocal microscope. Merged images are also shown. Bar, 10 �m.(B) Cells, prepared as above, were subjected to pulse-chase analysiswith [35S]methionine in the absence (DMSO) or presence of 2 �Mepoxomicin. Other conditions were as described in Fig. 5.

Jour

nal o

f Cel

l Sci

ence

2350 Journal of Cell Science 119 (11)

exogenously expressed human PLG in CHO-K1 cells andexamined whether the expression affected the cellularstability of SCD1. PLG was expressed in the cells and wasfound to be localized in the secretory organelles. There wasno significant stimulation, however, of SCD1 degradation,although it remains unknown whether the expressed PLGassumed the correct topology in the ER. In CHO-K1 cells andother cultured mammalian cells including HeLa and HEK-293, immunoblotting revealed no significant expression ofPLG; nevertheless SCD1 was rapidly degraded. Conversely,we found that PLG was expressed to a detectable level inHepG2 cells and exogenously expressed SCD1-HA wasdegraded with a half-life similar to that in CHO-K1 or HeLacells, and SCD1-HA degradation was not affected by PLGknockdown. Taken together, we considered that the ubiquitin-proteasome system is mainly responsible for the rapid SCD1

degradation in cultured mammalian cells, although theseresults do not necessarily exclude a PLG-dependentdegradation pathway.

We demonstrated that the 66-residue N-terminal segment ofSCD1 is rich in PEST sequences and is important forproteasome-dependent degradation. Deletion of the 66-residueN-terminal segment made SCD1 more stable than the wild-type enzyme, but the degradation was not completely arrestedby this manipulation and the mutant (�66SCD1) was degradedat a slower rate in a proteasome inhibitor-resistant manner,suggesting that �66 SCD1 was degraded by some non-proteasomal pathway.

Similarly, several membrane proteins are degraded by a non-proteasomal pathway, e.g. a type I transmembrane proteinthyroperoxidase and polytopic membrane protein voltagesensor mutant of the Shaker K+-channel subunit (Fayadat et al.,2000; Myers et al., 2004). The functional division, however, ofthe proteasomal and non-proteasomal degradation pathways inthe ER remains unknown.

We demonstrated here that the 66-residue N-terminalsegment of SCD1 containing PEST sequences (SCD66) madea stable ER membrane protein, b5-EGFP, susceptible toproteasome-dependent degradation. By contrast, it failed tostimulate the degradation of cytoplasm-localizing EGFP. Theseresults indicated that SCD66 is not a general degradationsignal, but functions as a specific degradation signal in the ERmembrane. In the ER, there might be a mechanism thatrecognizes the SCD66 segment and leads to proteasome-mediated degradation.

HMGR is a polytopic ER membrane protein and a rate-limiting enzyme of the cholesterol biosynthesis pathway. It iswell established that its cellular level is regulated by at leasttwo independent control systems that respond to cellular levelsof sterols or unsaturated fatty acids. One involves sterolrepression of the HMGR gene through regulated cleavage ofthe ER membrane-bound transcription factor SREBP, thesecond involves sterol-induced rapid degradation of HMGR,the regulated degradation via the ERAD system.

In spite of the similar gene regulation between HMGR andSCD1, the SCD1 half-life in cultured cells remained unaffectedby feeding unsaturated fatty acids or culture in the presence oflipid-depleted FBS, albeit expression of the SCD1 generesponded properly. We thus concluded that SCD1 is naturallyshort-lived and is constitutively degraded within the ER by theubiquitin-proteasome system, as is also the case for the yeastSCD, Ole1p (Braun et al., 2002).

Materials and MethodsMaterialsThe following reagents were obtained from the companies shown in parentheses.MG132 (Sigma Chemical Co., St Louis, MO, USA), epoxomicin and leupeptin(Peptide Institute, Osaka, Japan), an expression vector in mammalian cells,pcDNA3.1 (Invitrogen, Carlsbad, CA, USA), rabbit anti-hemagglutinin (HA)antiserum (Covance, Rapids, MI, USA), rabbit anti-calnexin polyclonal antibodies(Stress Gen, Victoria, Canada), mouse anti-Myc antibody (Upstate, Waltham, MA,USA), mouse anti-polyubiquitin monoclonal antibody (Medical & BiologicalLaboratories, Aichi, Japan), mouse anti-p97 monoclonal antibody (ProgenBiotechnik, Heidelberg, Germany), goat anti-human PLG antibodies (CedarlaneLaboratories, Ontario, Canada), protein A-Sepharose (Amersham Biosciences,Piscataway, NJ, USA), and Pfu turbo DNA polymerase (Stratagene, La Jolla, CA,USA). Antibodies against SCD1 were raised in rabbits against a synthetic 15-merpeptide corresponding to the C-terminal 15 amino acid residues of rat SCD1.Antibodies against rat Tim23 and EGFP were as described by Ishihara and Mihara(Ishihara and Mihara, 1998) and Miyazaki et al. (Miyazaki et al., 2001),respectively.

Fig. 8. SCD1 is naturally short-lived and is constitutively degradedirrespective of cellular lipid levels that regulate SCD1 geneexpression. (A) HEK-293 cells (untransfected) were grown for 36hours in standard culture medium containing untreated FBS(delipidated, –) or delipidated FBS (delipidated, +). Total RNAs wereisolated from the cells and the mRNA amount for the indicatedproteins was determined by real-time quantitative PCR. Levels ofmRNAs are shown relative to the amounts of mRNA from the cellscultured in untreated FBS-containing medium. The GAPDH genewas used as a negative control. (B) HEK-293 cells were transfectedwith plasmid carrying SCD1-HA cDNA, grown for 48 hours, andsubjected to pulse-chase analysis with [35S]methionine under theindicated conditions. At the indicated chase times, cells weresolubilized and subjected to immunoprecipitation with anti-HAantibodies. The immunoprecipitates were analyzed by SDS-PAGEand subsequent digital autoradiography. (C) Band intensities in Bwere quantified. Relative radioactivity is shown with the levels attime zero defined as 100%. The results shown are the average ofthree independent experiments.

Jour

nal o

f Cel

l Sci

ence

2351Ubiquitin-proteasome-dependent degradation of stearoyl-CoA desaturase

Vector and cDNATo construct pCMV SCD1-HA and pCMV �66SCD1-HA, the cDNA fragments ofrat SCD1 were amplified by polymerase chain reaction (PCR) using SCD1 cDNA(Mihara, 1990) as the template and the following oligonucleotides as the primers,and cloned between the HindIII and XbaI sites in pRcCMV-HA (Ukaji et al., 2002).SCD1-wt-F: CACAAGCTTCCCACCATGCCGGCCCACATGCTC; SCD1-wt-R:GTGTCTAGAGCTACTCTTGTGGCTCCC; SCD1-�66-F: CACAAGCTTCCC-ACCATGAAGCTGGAGTACGTCTGGAGG.

To construct pCMV b5-EGFP, the cDNA fragment was amplified by PCR usingrat cytochrome b5 cDNA (Mitoma and Ito, 1992) as the template and the followingoligonucleotides as the primers, and cloned between the HindIII and BamHI sites inEGFP-N1 (Clontech, Palo Alto, CA, USA). b5-F: CACAAGCTTCCCACCATG-GCCGAGCAGTCA; b5-R: GCTTGGATCCCGATCTTCTGCCATGTAGAGGC.

To construct pCMV SCD66-b5-EGFP, cDNA fragments encoding the 66-residueN-terminal segment (SCD66) and cytochrome b5 were generated by PCR, and thenconnected by overlapping PCR. The obtained PCR fragment was cloned betweenthe HindIII and BamHI sites in EGFP-N1 (Clontech). To construct pCMV TCR�-3xflag, mouse TCR� cDNA was amplified by PCR using pHDS58 (Saito et al.,1984) as the template and the following oligonucleotides as the primers, and clonedbetween the EcoRI and BamHI sites in p3xFLAG-CMV-14 (Sigma). TCR-F:GAGCGAATTCCCACCATGCTCCTGGCACTCCTCCCA; TCR-R: CCGGGA-TCCACTGGACCACAGCCTCAGCGT.

To construct SCD1-wt-enhanced GFP (EGFP), SCD66-EGFP and SCD33-EGFPwere amplified by PCR using SCD1 cDNA as a template, and SCD1-wt-F, SCD1-wt-GR; SCD1-wt-F, SCD1-66-R; and SCD1-wt-F, SCD1-33-R as the primers, andthe obtained cDNA fragment was cloned between HindIII and AgeI sites in EGFP-

Fig. 9. Plasminogen does not influencedegradation of SCD1 in cultured cells.(A) Hamster liver, hamster livermicrosomes, CHO-K1 cells expressingPLG-flag or the indicated cells weresubjected to SDS-PAGE and subsequentimmunoblot analyses with anti-PLGantibodies. (B) CHO-K1 cells weretransfected with a PLG-flag construct andsubjected to subcellular fractionation bydifferential centrifugation after 48 hoursculture. The fraction containing unbrokencells and nuclei was removed bycentrifugation at 1,000 g for 10 minutes.The supernatant was fractionatedsequentially at 6,000 g for 10 minutes andthen 100,000 g for 30 minutes to obtain P1and P2, respectively. P1 is composed ofmitochondria and heavy microsomes,whereas P2 is composed of microsomes.Fractionated samples were analyzed bySDS-PAGE, and subsequentimmunoblotting using anti-FLAG, anti-calnexin antibody (ER marker), or anti-Tim23 antibody (mitochondrial marker).(C) CHO-K1 cells were transfected withthe indicated amounts of the empty vector(pcDNA3.1), or those harboring SCD1-HAand PLG-flag cDNAs. After 48 hours, thetransfected cells were analyzed by SDS-PAGE and subsequent immunoblottingusing anti-FLAG antibody or anti-HAantiserum. (D) CHO-K1 cells weretransfected with pcDNA3.1 harboringSCD1-HA cDNA and different amounts ofPLG-flag cDNA (total amounts wereadjusted to 5 �g with pcDNA3.1). After 48hours, the cells were subjected to[35S]methionine pulse-chase analysis forthe indicated time intervals. The resultsshown are the average of three independentexperiments. Other conditions were asdescribed in Fig. 5D. (E) The extractsobtained from CHO-K1 cells expressingPLG-flag or from untransfected HepG2cells were resolved by SDS-PAGE andanalyzed by immunoblotting using anti-PLG, or anti-calnexin antibodies. (F) Theextracts obtained from control (RNAi forEGFP) or PLG-knockdown HepG2 cellswere analyzed by SDS-PAGE andsubsequent immunoblotting using anti-PLG- or anti-GAPDH antibodies. PLG was depleted by ~93%. (G) Hep G2 cells were subjected to RNAi for PLG and pCMV SCD1-HAplasmid transfection as described in Materials and Methods. The cells were then subjected to [35S]methionine pulse-chase experiments as

Jour

nal o

f Cel

l Sci

ence

2352 Journal of Cell Science 119 (11)

N1 (Clontech). SCD1-wt-GR: GGCGACCGGTGGCCGGCTACTCTTGTGGCTC-CCATC; SCD1-66-R: GGCGACCGGTGGCGCGGGCGGGGGCCCCTCCT-CATC; SCD1-33-R: GGCGACCGGTGGCTTCTCTCGTCCATTCTGCAG.

To construct flag-tagged plasminogen (PLG-flag), human plasminogen wasamplified by PCR using a human kidney cDNA library (Clontech) as a template,with PLG-F and PLG-R as the primers, and the obtained cDNA fragment was clonedbetween the HindIII and XbaI sites in p3xFLAG-CMV-14 (Sigma). PLG-F: GAG-CAAGCTTCCACCATGGAACATAAGGAAGTGGTTCTT; PLG-R: CTGTCTA-GAATTATTTCTCATCACTCCCTCAATCCAAGT.

To construct Myc-tagged ubiquitin, human ubiquitin was amplified by PCR ausing a human kidney cDNA library (Clontech) as a template, and ubi-F and ubi-Ras the primers, and the obtained cDNA fragment was cloned between the EcoRI andKpnI sites in pCMV-Myc (Clontech). ubi-F: GGACGAATTCGGATGCA-GATCTTCGTGAAGACTCTGACT; ubi-R: CCTTGGTACCTTACCCACCTCT-GAGACGGAGCACCAGGT

Cell culture and transfectionCHO-K1 cells were cultured in F-12 medium supplemented with 10% fetal bovineserum (FBS) and 10 �g/ml gentamycin (Invitrogen). HEK-293 and HeLa cells werecultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10%FBS and 10 �g/ml gentamycin. HepG2 cells were cultured in minimum essentialmedium (MEM) supplemented with 1% non essential amino acids, 1 mM sodiumpyruvate, 2 mM glutamine and 10% FBS. Liposome-mediated transient transfectionwas performed using lipofectamine (Invitrogen) according to the manufacturer’sinstructions. The cells were cultured at 37°C under 5% CO2 after transfection. NIH3T3-L1 cells were cultured in DMEM supplemented with 10% FBS and 10 �g/mlgentamycin. For differentiation of cells into adipocytes, confluent preadipocytemonolayers were incubated for 48 hours in DMEM containing 10% FBS anddifferentiation cocktail consisting of 115 �g/ml methylisobutylxanthine (Sigma),390 ng/ml dexamethasone (Sigma), and 10 �M insulin (Sigma). After 48 hoursculture, the cells were maintained in DMEM containing 10% FBS and 10 �Minsulin, changing the medium every 2 days. The cell morphology was monitoreddaily for the appearance of cytoplasmic lipid droplets using a phase-contrastmicroscope and oil Red O staining.

RNA interferenceSmall interference RNAs (siRNAs) targeting EGFP (control) and PLG werechemically synthesized: EGFP sense: 5�-CUACAACAGCCACAACGUCTT-3�;EGFP antisense: 5�-GACGUUGUGGCUGUUGUAGTT-3�; PLG sense: 5�-CCGCAAUCCUGACGGAAAATT-3�; PLG antisense: 5�-UUUUCCGUCAGG-AUUGCGGTT-3�.

HepG2 cells were transfected with siRNA using Lipofectamine 2000 reagent(Invitrogen). After 12 hours, the cells were transfected with pCMV SCD1-HA usingFugene 6 (Roche Diagnostics, Penzberg, Germany). After 12 hours, the cells weretransfected with siRNA as described above, cultured for 24 hours, and thensubjected to pulse-chase experiments.

ImmunocytochemistryCells on coverslips were fixed with 4% paraformaldehyde at room temperature for 20minutes and permeabilized with 0.1% Triton X-100. They were incubated with 1%bovine serum albumin in phosphate-buffered saline (PBS) for 20 minutes, and thenincubated with primary antibodies for 1 hour at room temperature. After several washesin PBS, the coverslips were incubated with fluorescein isothiocyanate-conjugated goatanti-mouse antibodies (BioSource, Camarillo, CA, USA) or Texas-Red-conjugatedgoat anti-rabbit antibodies (Cappel, Durham, NC, USA) for 1 hour at roomtemperature, followed by washing in PBS. Fluorescent images were taken using aconfocal laser microscope Radiance 2000 (BioRad, Hercules, CA) and analyzed.

Pulse-chase assayCells (in a 3.5 cm dish) were incubated for 1 hour in cysteine- and methionine-freemedium. They were then pulse-labeled for 30 minutes with 1.85 MBq 35S-labelingmix (Perkin Elmer, Wellesley, MA, USA) in 1 ml culture medium in the absenceor presence of proteasome inhibitors. The labeling medium was removed, and thecells were subjected to a chase reaction in normal medium. At the indicated timepoints, cells were washed with ice-cold PBS, lysed in 50 �l solubilization buffer[50 mM Tris-HCl buffer (pH 7.5) containing 2% sodium dodecyl sulfate (SDS)] anddenatured at 95°C for 5 minutes. The mixture was then diluted 20-fold withimmunoprecipitation buffer [50 mM Tris-HCl buffer (pH 7.5) containing 150 mMNaCl, and 1% Triton X-100]. The lysates were clarified by brief centrifugation. Thecell lysates (150 �g) were incubated with specific antibodies and protein A-Sepharose at 4°C for 5 hours. Protein A-Sepharose was collected by centrifugation,washed with the washing buffer [50 mM Tris-HCl buffer (pH 7.5) containing 150mM NaCl, 1% Triton X-100, and 0.2% SDS], and the antigen was extracted withSDS-polyacrylamide gel electrophoresis (PAGE) loading buffer [62.5 mM Tris-HClbuffer (pH 6.8) containing 2% SDS, 5% �-mercaptoethanol and 10% glycerol]. Theimmunoprecipitates were resolved by SDS-PAGE and analyzed using a BAS 2500image analyzer (Fuji Photo Film, Tokyo, Japan). Band intensities were quantifiedusing Image Gauge software (Fuji Photo Film).

Immunoprecipitiation for detection polyubiquitylated SCD1Cells were transfected with pcDNA3.1 or the plasmid harboring SCD1-HA cDNAand grown for 48 hours. They were incubated for 5 hours in the presence or absenceof 50 �M MG132, washed with ice-cold PBS, lysed in 100 �l solubilization buffer[50 mM Tris-HCl buffer (pH 7.5) containing 2% SDS and 5 mM N-ethylmaleimide(NEM)], and incubated at 95°C for 10 minutes. The mixture was then diluted 10-fold with immunoprecipitation buffer [50 mM Tris-HCl buffer (pH 7.5) containing150 mM NaCl, 1% Triton X-100 and 5 mM NEM]. The lysates were clarified bybrief centrifugation. The supernatants (800 �g) were incubated with specificantibodies and Protein A-Sepharose at 4°C for 5 hours. Protein A-Sepharose wascollected by centrifugation, washed with washing buffer [50 mM Tris-HCl buffer(pH 7.5) containing 150 mM NaCl, 1% Triton X-100, 0.2% SDS and 5 mM NEM],and the antigen was extracted with SDS-PAGE loading buffer. Theimmunoprecipitates were analyzed by SDS-PAGE and subsequent immunoblottingusing specific antibodies.

Cell fractionation and immunoprecipitationCells transfected with pcDNA3.1 or the plasmid harboring SCD1-HA cDNA andgrown for 48 hours were incubated for 5 hours in the presence or absence of 50 �MMG132. They were washed with PBS and collected by centrifugation at 700 g for5 minutes. The cells were washed once with homogenization buffer [10 mM Hepes-KOH buffer (pH 7.5) containing 0.25 M sucrose and 5 mM NEM], homogenizedin 1 ml homogenization buffer by passing through a 27-gauge needle 20 times, andthen centrifuged at 1,000 g for 10 minutes to obtain a post-nuclear supernatant. Thepost-nuclear supernatant (400 �g) was centrifuged at 100,000 g for 30 minutes toseparate the cytosolic and membrane fractions, which were lysed with the lysisbuffer [20 mM Hepes-KOH buffer (pH 7.5) containing 150 mM NaCl, 2 mM MgCl2,1% Triton X-100 and 5 mM NEM] at 4°C for 1 hour. The lysates were clarified bycentrifugation and incubated with specific antibodies and protein A-Sepharose at4°C for 5 hours. The immunoprecipitates were analyzed by SDS-PAGE andsubsequent immunoblotting using specific antibodies.

Real-time quantitative PCRTotal RNA was extracted from HEK-293 cells using an RNeasy Mini Kit (Qiagen,Hilden, Germany). First-strand cDNA was synthesized by a SuperScript III First-Strand Synthesis System for RT-PCR (Invitrogen) according to the manufacturer’sinstructions. Primers for real-time quantitative PCR were designed using PrimerExpress software (Applied Biosystems, Foster City, CA. USA). The sequence ofprimer sets for human �-actin were: 5� TGT CCC CCA ACT TGA GAT GTA TG3� (actin forward) and 5� CCT CAT TTT TAA GGT GTG CAC TTT T 3� (actinreverse); for human SCD: 5� GAG GTA CTA CAA ACC TGG CTT GCT G 3�(SCD forward) and 5� CCA CTC TTG TAG TTT CCA TCT CCG G 3� (SCDreverse); for human HMGR: 5� TAC CAT GTC AGG GGT ACG TC 3� (HMGRforward) and 5� CAA GCC TAG AGA CAT AAT CAT C 3� (HMGR reverse); forhuman low density lipoprotein receptor (LDLR): 5� CAA TGT CTC ACC AAGCTC TG 3� (LDLR forward) and 5� TCT GTC TCG AGG GGT AGC TG 3� (LDLRreverse); for human GAPDH: 5� TGG AGT CCA CTG GCG TCT TC 3� (GAPDHforward) and 5� TTC ACA CCC ATG ACC AAC ATG 3� (GADPH reverse). Real-time quantitative PCR was performed using QuantiTect SYBR Green PCR (Qiagen)and analyzed with an ABI PRISM 7000 Sequence Detection System (AppliedBiosystems).

Preparation of delipidated FBSFetal bovine serum was delipidated using the method of Hannah et al. (Hannah etal., 2001). Serum (50 ml) was mixed with 40 ml of n-butanol and 60 ml of isopropylether at room temperature for 60 minutes, followed by 20 minutes incubation onice. The mixture was centrifuged at 2,000 g for 5 minutes at room temperature andthe aqueous phase was re-extracted with 20 ml of isopropyl ether. The aqueousphase was lyophilized and reconstituted in 20 ml of distilled water, which was thendialyzed against PBS and stored in aliquots at –20°C. Delipidated FBS containedno detectable free fatty acids compared with untreated FBS (787 �M) as determinedby ‘Free fatty acids, Half-microtest’ (Roche Diagnostics).

Flow cytometryThe cells were transfected with N-terminal segments of SCD1 and EGFP fusionsand grown for 48 hours. The cells were treated with 10 �g/ml cycloheximide for 0,2 and 4 hours. They were then washed with PBS and collected by centrifugation at700 g for 5 minutes. A population of 10,000 cells was then analyzed for EGFPfluorescence intensity using a FACScan flow cytometer (Becton Dickinson, FranklinLakes, NJ, USA).

We thank Takeshi Watanabe (RIKEN, RCAI, Research Unit forImmune Surveillance) for providing us with mouse TCR� subunitcDNA. We also thank J. L. Degen (Children’s Hospital ResearchFoundation, Cincinnati, OH, USA) for the information on PLGknockout mice. This work was supported by grants from the ministryof Education, Science and Evolutional Science and Technology, and

Jour

nal o

f Cel

l Sci

ence

2353Ubiquitin-proteasome-dependent degradation of stearoyl-CoA desaturase

Specially Promoted Research from the Ministry of Education,Science, and Culture of Japan, and Takeda Science Foundation toK.M.

ReferencesBene, H., Lasky, D. and Ntambi, J. M. (2001). Cloning and characterization of the

human stearoyl-CoA desaturase gene promoter: transcriptional activation by sterolregulatory element binding protein and repression by polyunsaturated fatty acids andcholesterol. Biochem. Biophys. Res. Commun. 284, 1194-1198.

Braun, S., Matuschewski, K., Rape, M., Thoms, S. and Jentsch, S. (2002). Role of theubiquitin-selective CDC48 (UFD1/NPL4) chaperone (segregase) in ERAD of OLE1and other substrates. EMBO J. 21, 615-621.

Ellgaard, L. and Helenius, A. (2003). Quality control in the endoplasmic reticulum. Nat.Rev. Mol. Cell Biol. 4, 181-191.

Fayadat, L., Siffroi-Fernandez, S., Lanet, J. and Franc, J. L. (2000). Degradation ofhuman thyroperoxidase in the endoplasmic reticulum involves two different pathwaysdepending on the folding state of the protein. J. Biol. Chem. 275, 15948-15954.

Goldstein, J. L. and Brown, M. S. (1990). Regulation of the mevalonate pathway. Nature343, 425-430.

Green, H. and Meuth, M. (1974). An established pre-adipose cell line and itsdifferentiation in culture. Cell 3, 127-133.

Hampton, R. Y. (2002). ER-associated degradation in protein quality control and cellularregulation. Curr. Opin. Cell Biol. 14, 476-482.

Hannah, V. C., Ou, J., Luong, A., Goldstein, J. L. and Brown, M. S. (2001).Unsaturated fatty acids down-regulate srebp isoforms 1a and 1c by two mechanismsin HEK-293 cells. J. Biol. Chem. 276, 4365-4372.

Heinemann, F. S. and Ozols, J. (1998). Degradation of stearoyl-coenzyme A desaturase:endoproteolytic cleavage by an integral membrane protease. Mol. Biol. Cell 9, 3445-3453.

Heinemann, F. S., Mziaut, H., Korza, G. and Ozols, J. (2003a). A microsomalendopeptidase from liver that preferentially degrades stearoyl-CoA desaturase.Biochemistry 42, 6929-6937.

Heinemann, F. S., Korza, G. and Ozols, J. (2003b). A plasminogen-like proteinselectively degrades stearoyl-CoA desaturase in liver microsomes. J. Biol. Chem. 278,42966-42975.

Horton, J. D., Goldstein, J. L. and Brown, M. S. (2002). SREBPs: activators of thecomplete program of cholesterol and fatty acid synthesis in the liver. J. Clin. Invest.109, 1125-1131.

Huppa, J. B. and Ploegh, H. L. (1997). The alpha chain of the T cell antigen receptoris degraded in the cytosol. Immunity 7, 113-122.

Ishihara, N. and Mihara, K. (1998). Identification of the protein import components ofthe rat mitochondrial inner membrane, rTIM17, rTIM23, and rTIM44. J. Biochem.Tokyo 123, 722-732.

Kaufman, R. J., Scheuner, D., Schroder, M., Shen, X., Lee, K., Liu, C. Y. and Arnold,S. M. (2002). The unfolded protein response in nutrient sensing and differentiation.Nat. Rev. Mol. Cell Biol. 3, 411-421.

Liao, W., Yeung, S.-C. J. and Chan, L. (1998). Proteasome-mediated degradation ofapolipoprotein B targets both nascent peptides cotranslationally before translocationand full-length apolipoprotein B after translocation into the endoplasmic reticulum. J.Biol. Chem. 273, 27225-27230.

Lilley, B. N. and Ploegh, H. L. (2004). A membrane protein required for dislocation ofmisfolded proteins from the ER. Nature 429, 834-840.

Meyer, H. H., Wang, Y. and Warren, G. (2002). Direct binding of ubiquitin conjugatesby the mammalian p97 adaptor complexes, p47 and Ufd1-Npl4. EMBO J. 21, 5645-5652.

Mihara, K. (1990). Structure and regulation of rat liver microsomal stearoyl-CoAdesaturase gene. J. Biochem. Tokyo 108, 1022-1029.

Mitoma, J. and Ito, A. (1992). The carboxy-terminal 10 amino acid residues ofcytochrome b5 are necessary for its targeting to the endoplasmic reticulum. EMBO J.11, 4197-4203.

Miyazaki, E., Sakaguchi, M., Wakabayashi, S., Shigekawa, M. and Mihara, K.(2001). NHE6 protein possesses a signal peptide destined for endoplasmic reticulummembrane and localizes in secretory organelles of the cell. J. Biol. Chem. 276, 49221-49227.

Mori, K. (2000). Tripartite management of unfolded proteins in the endoplasmicreticulum. Cell 101, 451-454.

Myers, M. P., Khanna, R., Lee, E. J. and Papazian, D. M. (2004). Voltage sensormutations differentially target misfolded K+ channel subunits to proteasomal and non-proteasomal disposal pathways. FEBS Lett. 568, 110-116.

Mziaut, H., Korza, G. and Ozols, J. (2000). The N-terminus of microsomal delta 9stearoyl-CoA desaturase contains the sequence determinant for its rapid degradation.Proc. Natl. Acad. Sci. USA 97, 8883-8888.

Neuber, O., Jarosch, E., Volkwein, C., Walter, J. and Sommer, T. (2005). Ubx2 linksthe Cdc48 complex to ER-associated protein degradation. Nat. Cell Biol. 7, 993-998.

Ntambi, J. M. (1995). The regulation of stearoyl-CoA desaturase (SCD). Prog. Lipid Res.34, 139-150.

Ntambi, J. M. (1999). Regulation of stearoyl-CoA desaturase by polyunsaturated fattyacids and cholesterol. J. Lipid Res. 40, 1549-1558.

Oshino, N. and Sato, R. (1972). The dietary control of the microsomal stearyl CoAdesaturation enzyme system in rat liver. Arch. Biochem. Biophys. 149, 369-377.

Ozols, J. (1997). Degradation of hepatic stearyl CoA delta 9-desaturase. Mol. Biol. Cell8, 2281-2290.

Rape, M., Hoppe, T., Gorr, I., Kalocay, M., Richly, H. and Jentsch, S. (2001).Mobilization of processed, membrane-tethered SPT23 transcription factor byCDC48(UFD1/NPL4), a ubiquitin-selective chaperone. Cell 107, 667-677.

Rechsteiner, M. and Rogers, S. W. (1996). PEST sequences and regulation byproteolysis. Trends Biochem. Sci. 21, 267-271.

Rogers, S., Wells, R. and Rechsteiner, M. (1986). Amino acid sequences common torapidly degraded proteins: the PEST hypothesis. Science 234, 364-368.

Saito, H., Kranz, D. M., Takagaki, Y., Hayday, A. C., Eisen, H. N. and Tonegawa, S.(1984). A third rearranged and expressed gene in a clone of cytotoxic T lymphocytes.Nature 312, 36-40.

Schuberth, C. and Buchberger, A. (2005). Membrane-bound Ubx2 recruits Cdc48 toubiquitin ligases and their substrates to ensure efficient ER-associated proteindegradation. Nat. Cell Biol. 7, 999-1006.

Sessler, A. M., Kaur, N., Palta, J. P. and Ntambi, J. M. (1996). Regulation of stearoyl-CoA desaturase 1 mRNA stability by polyunsaturated fatty acids in 3T3-L1 adipocytes.J. Biol. Chem. 271, 29854-29858.

Sever, N., Yang, T., Brown, M. S., Goldstein, J. L. and DeBose-Boyd, R. A. (2003).Accelerated degradation of HMG CoA reductase mediated by binding of Insig-1 to itssterol-sensing domain. Mol. Cell 11, 25-33.

Shimakata, T., Mihara, K. and Sato, R. (1972). Reconstitution of hepatic microsomalstearoyl-Coenzyme A desaturase from solubilized components. J. Biochem. 72, 1163-1174.

Shimano, H. (2001). Sterol regulatory element-binding proteins (SREBPs):transcriptional regulators of lipid synthetic genes. Prog. Lipid Res. 40, 439-452.

Sitia, R. and Braakmanm, I. (2003). Quality control in the endoplasmic reticulumprotein factory. Nature 426, 891-894.

Strittmatter, P., Spatz, L., Corcoran, D., Rogers, M. J., Setlow, B. and Redline, R.(1974). Purification and properties of rat liver microsomal stearyl coenzyme Adesaturase. Proc. Natl. Acad. Sci. USA 71, 4565-4569.

Tabor, D. E., Kim, J. B., Spiegelman, B. M. and Edwards, P. A. (1999). Identificationof conserved cis-elements and transcription factors required for sterol-regulatedtranscription of stearoyl-CoA desaturase 1 and 2. J. Biol. Chem. 274, 20603-20610.

Ukaji, K., Ariyoshi, N., Sakaguchi, M., Hamasaki, N. and Mihara, K. (2002).Membrane topogenesis of the three amino-terminal transmembrane segments ofGlucose-6-phosphatase on endoplasmic reticulum. Biochem. Biophys. Res. Commun.292, 153-160.

Walter, J., Urban, J., Volkwein, C. and Sommer, T. (2001). Sec61p-independentdegradation of the tail-anchored ER membrane protein Ubc6p. EMBO J. 20, 3124-3131.

Waters, K. M. and Ntambi, J. M. (1996). Polyunsaturated fatty acids inhibit hepaticstearoyl-CoA desaturase-1 gene in diabetic mice. Lipids 31, S33-S36.

Ye, Y., Meyer, H. H. and Rapoport, T. A. (2003). Function of the p97–Ufd1–Npl4 complexin retrotranslocation from the ER to the cytosol dual recognition of nonubiquitylatedpolypeptide segments and polyubiquitin chains. J. Cell Biol. 162, 71-84.

Ye, Y., Shibata, Y., Yun, C., Ron, D. and Rapoport, T. A. (2004). A membrane proteincomplex mediates retro-translocation from the ER lumen into the cytosol. Nature 429,841-847.

Jour

nal o

f Cel

l Sci

ence