Distinguishing between retention signals and ... - Semantic Scholar

Research Article

4377

Distinguishing between retention signals and degrons acting in ERAD Ilana Shapira1, Dana Charuvi1,*, Yechiel Elkabetz1,‡, Koret Hirschberg2 and Shoshana Bar-Nun1,§ 1

Department of Biochemistry, George S. Wise Faculty of Life Sciences and 2Department of Pathology, Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel

*Present address: The Robert H Smith Institute of Plant Sciences and Genetics in Agriculture, The Hebrew University of Jerusalem, Rehovot, Israel ‡ Present address: Laboratory of Stem Cell and Tumor Biology, Division of Neurosurgery and Developmental Biology Program, Sloan-Kettering Institute, New York, NY, USA § Author for correspondence (e-mail: [email protected])

Journal of Cell Science

Accepted 3 October 2007 Journal of Cell Science 120, 4377-4387 Published by The Company of Biologists 2007 doi:10.1242/jcs.011247

Summary Endoplasmic reticulum-associated degradation (ERAD) eliminates aberrant proteins from the secretory pathway. Such proteins are retained in the endoplasmic reticulum and targeted for degradation by the ubiquitin-proteasome system. Cis-acting motifs can function in ERAD as retention signals, preventing vesicular export from the endoplasmic reticulum, or as degrons, targeting proteins for degradation. Here, we show that ␮stp, the C-terminal 20-residue tailpiece of the secretory IgM ␮s heavy chain, functions both as a portable retention signal and as an ERAD degron. Retention of ␮stp fusions of secreted versions of thyroid peroxidase and yellow fluorescent protein in the endoplasmic reticulum requires the presence

Introduction Quality control mechanisms that operate in the secretory pathway prevent deployment of aberrant proteins to distal compartments. In the endoplasmic reticulum (ER), folding, Nglycosylation and assembly of nascent chains are tightly monitored. Misfolded proteins and orphan subunits of oligomeric proteins, as well as metabolically or developmentally unwanted proteins, are eliminated by the ubiquitin-proteasome system by means of the ER-associated protein degradation (ERAD) pathway (for reviews, see Bonifacino and Weissman, 1998; Ellgaard and Helenius, 2003; Kostova and Wolf, 2003; McCracken and Brodsky, 2003; Sitia and Braakman, 2003; Trombetta and Parodi, 2003; Sayeed and Ng, 2005; Bar-Nun, 2005). Retention of proteins destined for vesicular export is a prerequisite for ERAD; however, retention does not necessarily culminate in degradation. Therefore, it is of interest to determine whether cis-acting motifs that are recognized as ‘retention signals’, preventing protein secretion, overlap with cis-acting motifs that act as ‘degrons’, targeting proteins for degradation. Degrons are defined as sequences or domains that are necessary and sufficient for directing otherwise stable proteins for degradation (Varshavsky, 1991). A degron that targets soluble lumenal proteins for ERAD might play a role in any step along this complex pathway, steps that are not necessarily coupled and schematically entail substrate selection, dislocation to the cytosol, ubiquitylation and degradation by the proteasome. Thus, a degron might constitute the recognition signal for the substrate, it might lead the

of the penultimate cysteine of ␮stp. In its role as a portable degron, the ␮stp targets the retained proteins for ERAD but does not serve as an obligatory ubiquitin-conjugation site. Abolishing ␮stp glycosylation accelerates the degradation of both ␮stpCys-fused substrates, yet absence of the N-glycan eliminates the requirement for the penultimate cysteine in the retention and degradation of the unglycosylated yellow fluorescent protein. Hence, the dual role played by the ␮stpCys motif as a retention signal and as a degron can be attributed to distinct elements within this sequence. Key words: ERAD, Retention signals, Degrons, Proteasome

dislocation process, it might recruit the ubiquitylation machinery and/or provide the platform and Lys residue(s) for ubiquitin conjugation and/or it might target the substrate to the proteasome. The heavy chain of the secretory immunoglobulin M (sIgM), ␮s, is one of the few soluble lumenal ERAD substrates studied in mammalian cells (Amitay et al., 1992; Elkabetz et al., 2003). This protein provides an attractive model to study the interrelations between retention signals and degrons that operate in ERAD. During differentiation of B lymphocytes, the fate of ␮s switches from retention and degradation in B cells to stability and efficient secretion in plasma cells (Amitay et al., 1991; Amitay et al., 1992; Shachar et al., 1992). Moreover, in the same B cell, where ␮s is retained and degraded, the membrane isoform, ␮m, is a stable protein that reaches the cell surface, where it constitutes the B cell receptor. Thus, the intracellular fate of ␮s appears to be mediated by B-cellspecific components and to rely on ␮stp, the C-terminal tailpiece of ␮s that distinguishes it from ␮m. This unique sequence comprises 20 residues that include a penultimate Cys575 and is designated ‘␮stpCys’. The ␮stpCys motif is a necessary cis-acting retention signal; either its truncation or Cys575 mutation to Ser or Ala hampers retention of ␮s in B cells (Sitia et al., 1990). Moreover, ␮stpCys, unlike ␮stpSer or ␮stpAla, is also sufficient to prevent secretion of IgG2b in B cells or in the non-lymphoid COS-7 cells and to retain the lysosomal protein cathepsin D to a pre-Golgi compartment in COS-7 cells (Sitia et al., 1990; Fra et al., 1993; Isidoro et al., 1996). In addition, the ␮stpCys motif targets the retained


4378

Journal of Cell Science 120 (24)

cathepsin D, but not the retained IgG2b, for rapid degradation in a pre-Golgi compartment (Fra et al., 1993). In this work, we fuse the ␮stpCys to two bona fide secretory proteins and study, in various non-lymphoid cells, its role as a retention signal, we localize the retained protein to the ER and we show that the ␮stpCys sequence acts as an ERAD degron. That ␮stpCys acts as a cis-acting motif also allows investigation of the relevance of degron-linked N-glycans to ERAD. It is well established that N-glycans are key players in ER quality control, both as monitors of the folding of glycoproteins in the ER and as signals that target glycoproteins to ERAD (Helenius and Aebi, 2001; Ellgaard and Helenius, 2003; Helenius and Aebi, 2004; Moremen and Molinari, 2006; Olivari et al., 2006). Furthermore, correct positioning of Nglycans appears crucial for retention and degradation, as shown for the yeast lumenal ERAD substrate carboxypeptidase Y* (CPY*), whose most C-terminal of four N-glycans is necessary and sufficient to direct this protein for degradation (Kostova and Wolf, 2005; Spear and Ng, 2005). Nonetheless, it is currently unknown whether actual linking of N-glycans to degrons is of any significance. The ␮stp element harbors a single, most C-terminal N-glycan linked to Asn563 in the context of ␮s, in addition to four N-glycans linked to ␮s upstream of the ␮stp element (Brenckle and Kornfeld, 1980; Sidman et al., 1981). Both the ␮stp N-glycan as well as the penultimate Cys residue are highly conserved throughout evolution and are found in all species (de Lalla et al., 1998), although this N-glycan plays no role in maintaining the function of Cys575 in retention (Fra et al., 1993). Nevertheless, the ␮stp N-glycan was shown to be important for sIgM J-chain association to form pentamers, yet it was dispensable for secretion of sIgM hexamers or a ␮s double mutant (S565A C575A) (Cals et al., 1996; de Lalla et al., 1998). Taking into account the contribution of N-glycans linked to ␮s to the secretion and stability of this heavy chain (Sidman et al., 1981; Sidman, 1981), here we examine whether the N-glycan linked to the ␮stp motif, being the most C-terminal or the sole Nglycan, plays any role in sorting two bona fide secretory proteins to either a secretory or ERAD fate. Our results establish the role of the ␮stpCys element as a portable retention signal. We also demonstrate that the ␮stpCys sequence acts as a portable degron that targets secretory proteins to ERAD through proteasomal degradation. Two interesting features of ␮stpCys are studied with regard to its role as a degron. First, we exclude the ␮stpCys degron as an obligatory ubiquitylation site. Second, we show that the Nglycan linked to the ␮stp element plays an important role in the sorting of ␮stp-fusion proteins for either secretion or ERAD. Results The ␮stpCys sequence acts as a portable retention signal and a portable ERAD degron To investigate the various roles of the ␮stpCys element in ERAD, we selected two bona fide secretory proteins as reporters: a secreted version of the relatively unstable thyroid peroxidase (Fayadat et al., 2000) from which the transmembrane segment was truncated (designated TPO) and a secreted version of the relatively stable yellow fluorescent protein that is led to the ER lumen by the hen egg lysozyme signal sequence (designated ssYFP). Two versions of ␮stp,

wild-type ␮stpCys (penultimate Cys) and mutant ␮stpSer (penultimate Ser), were appended to the C-termini of TPO and ssYFP, downstream of a Myc tag, generating TPO-␮stpCys and TPO-␮stpSer (Fig. 1A) and ssYFP-␮stpCys and ssYFP␮stpSer (Fig. 2A). Importantly, the fluorescence of the ssYFP fusion proteins reflects their proper folding and allows direct visualization of their localization within living cells by laser scanning confocal microscopy. First we demonstrated that the ␮stpCys sequence conferred intracellular retention upon secretory TPO in several nonlymphoid cell lines. Radiolabeled cells were chased and the TPO-␮stp fusion proteins were immunoprecipitated from cells and medium with an antibody against Myc. Throughout the chase, the intracellular levels of both TPO-␮stpCys and TPO␮stpSer declined. However, after 6 hours of chase, ~70% of TPO-␮stpSer was recovered from the medium, fully accounting for its decrease within the cells, whereas less than 10% of TPO-␮stpCys was secreted from HEK293T or COS-7 cells (Fig. 1B,C), suggesting that this protein was degraded. Grafting the ␮stpCys sequence onto TPO allows us to test directly whether this cis-acting retention signal acts also as a portable ERAD degron. We have previously shown in 38C B cells that ␮s is a bona fide lumenal ERAD substrate because this retained protein is eliminated rapidly by the ubiquitinproteasome system (Amitay et al., 1991; Amitay et al., 1992; Shachar et al., 1992; Elkabetz et al., 2003). To determine whether TPO-␮stpCys was degraded in a fashion similar to that of ␮s, we performed pulse-chase experiments in the presence of proteasome inhibitors. The rapidly degraded TPO␮stpCys was stabilized in the presence of proteasome inhibitors, extending its half-life from ~4 to ~10 hours, and its secretion was improved slightly (Fig. 1D,E). Similar experiments were performed with Myc-tagged ssYFP-␮stpCys and ssYFP-␮stpSer, and, for comparison, with ssYFP-Myc that lacked ␮stp altogether. Radiolabeled cells were chased and ssYFP-␮stpCys, ssYFP-␮stpSer and ssYFPMyc were immunoprecipitated from cells and medium. After 7.5 hours of chase, more than 50% of ssYFP-␮stpCys was retained within cells, whereas 70% of ssYFP-␮stpSer and 63% of ssYFP-Myc were secreted to the medium (Fig. 2B,C). Notably, in agreement with the previously reported inefficient processing of the ␮stp N-glycan (Brenckle and Kornfeld, 1980; Cals et al., 1996), the N-glycans of the secreted ssYFP-␮stpSer were a mixture of high-mannose (Fig. 2B, lower panel, arrow) and complex (Fig. 2B, lower panel, arrowhead) species. Again, ssYFP-␮stpCys was not secreted, yet its intracellular levels decreased during the chase (Fig. 2B,C), suggestive of degradation. Indeed, in the presence of proteasome inhibitors, the ~9 hour half-life of ssYFP-␮stpCys was prolonged tenfold (Fig. 2B,C). Thus, ␮stpCys acts as a portable ERAD degron that targets both TPO and ssYFP for proteasomal degradation. The function of ␮stpCys as an ER retention signal was demonstrated by real-time visualization of ssYFP in COS-7 cells (Fig. 3). Fluorescence microscopy revealed that, while the wild-type ssYFP-␮stpCys was restricted to the ER, the mutant ssYFP-␮stpSer was distributed both within the ER and Golgi, as was ssYFP-Myc (Fig. 3). These latter two fusion proteins colocalized with the trans-Golgi resident galactosyl transferase fused to cyan fluorescent protein (GalT-CFP), whereas the ssYFP-␮stpCys was excluded from the GalT-CFP sites (Fig. 3). Importantly, the comparable fluorescence of the ssYFP


Retention signals and degrons in ERAD

4379

Fig. 1. The ␮stpCys motif prevents secretion of TPO and targets it to ERAD. (A) Schematic presentation of the Myc-tagged TPO vectors. (B) COS-7 or HEK293T cells transfected with vectors encoding Myc-tagged TPO-␮stpCys (Cys) or TPO-␮stpSer (Ser) were pulselabeled with [35S]methionine-[35S]cysteine and chased for the indicated time. Cells and medium were separated, cells were lysed and substrates were immunoprecipitated (IP) from lysed cells or medium by an antibody against Myc. Immunoprecipitates were resolved by SDS-PAGE, electroblotted and blots were exposed to autoradiography ([35S]). (C) The graph, representative of five independent experiments, illustrates the amounts of TPO␮stpCys (Cys; open symbols) or TPO-␮stpSer (Ser; filled symbols) remaining in HEK293T cells (circles), recovered from medium (squares) or their sum (triangles). Amounts estimated by densitometry of autoradiograms in B were calculated as a percentage of their levels at the end of the pulse (100%). (D) HEK293T cells transfected with a vector encoding Myc-tagged TPO-␮stpCys were pulse-labeled with [35S]methionine[35S]cysteine, chased for the indicated time with (+) or without (–) ALLN and MG-132 and analyzed as described in B. (E) The graph, representative of five independent experiments, illustrates the amounts of TPO-␮stpCys remaining in untreated (open symbols) or ALLN- and MG-132-treated (+ Inhib; filled symbols) HEK293T cells. The amounts detected in cells (circles), recovered from medium (squares) or their sum (triangles) were estimated by densitometry of autoradiogram D, calculated as a percentage of their levels at the end of the pulse (100%), and the half-life values (see text) were calculated.

fusion proteins indicates that ␮stpCys does not interfere with the proper folding of this reporter protein. Therefore, the retention of the fusion proteins at the ER is brought about by ␮stpCys per se. Hence, by analogy to the fate of ␮s in B cells (Sitia et al., 1990), wild-type ␮stpCys, unlike mutant ␮stpSer, is a sufficient cis-acting retention signal when fused to several reporter secretory proteins. The ␮stpCys sequence retains the reporter secretory proteins at the ER and appears to act as a retention signal and an ERAD degron recognized in various non-lymphoid cell lines. The single lysine in ␮stp is irrelevant for degradation Most ERAD substrates, including ␮s (Amitay et al., 1991; Amitay et al., 1992; Shachar et al., 1992; Elkabetz et al., 2003), are ubiquitylated on lysine residues before their proteasomal degradation. Following our finding that the ␮stpCys fusion proteins studied here are degraded by the proteasome (Figs 1, 2), we asked whether ␮stpCys, acting as an ERAD degron, also serves as the ubiquitylation site of its fusion proteins. Conceivably, sequences that function as degrons could recruit the ubiquitylation machinery to the ERAD substrates and/or provide the site for ubiquitin conjugation. As ␮stpCys contains a single lysine, we examined whether this residue serves as the ubiquitylation site by substituting it with arginine, generating the TPO-␮stpCysKR mutant. Pulse-chase experiments showed indistinguishable proteasomal degradation of wild-type TPO␮stpCys (Fig. 4A) and the TPO-␮stpCysKR mutant (Fig. 4B). The similar short half-life of both substrates was prolonged equally by 2.5-fold in the presence of proteasome inhibitors (Fig. 4C). Hence, the ␮stp lysine residue appears to play no role in degradation of TPO-␮stpCys, suggesting that degrons,

which by definition confer degradation, are not necessarily the sequences that undergo ubiquitylation. The N-glycan of ␮stp plays a key role in secretion and degradation of ␮stp fusion proteins N-glycans play an important role in ER quality control and ERAD (Helenius and Aebi, 2001; Ellgaard and Helenius, 2003; Helenius and Aebi, 2004), and their positioning within protein substrates might be crucial for degradation (Kostova and Wolf, 2005; Spear and Ng, 2005). Here, we investigated whether the N-glycan linked to the ␮stp motif plays any role in the retention and degradation of ␮stp-fusion proteins. This is especially relevant because the single N-glycan of ␮stp is the most Cterminal also in the context of the ␮s heavy chain and is highly conserved throughout evolution (Sidman et al., 1981; de Lalla et al., 1998). Substituting the asparagine for glutamine in the tripeptide sequon Asn-Val-Ser in the ␮stp of the Cys and Ser versions of the two reporter fusion proteins generated four ‘NQ’ mutants. In TPO-␮stp, this asparagine residue is most Cterminal of four additional potential N-glycosylation sites (Rawitch et al., 1992). For ssYFP, which has no Nglycosylation sites of it own, ␮stp provides the only site for Nglycan attachment in ssYFP-␮stp. To confirm that the asparagine residue within the ␮stp motif was glycosylated also in the context of the fusion proteins, the electrophoretic mobility of the NQ mutants was compared with that of their wild-type counterparts. This analysis is based on mobility retardation, which is contributed by each of the N-glycans. The slightly faster migration of the TPO-␮stpCysNQ mutant, compared with the mobility of wild-type TPO-␮stpCys and the TPO-␮stpCysKR mutant (Fig. 5A), suggested that the


4380


Fig. 2. The ␮stpCys element prevents secretion of ssYFP and targets it to ERAD. (A) Schematic presentation of the Myc-tagged ssYFP vectors. (B) Hela cells expressing a vector encoding Myc-tagged ssYFP (myc), or HEK293 cell vectors encoding ssYFP-␮stpCys (Cys) or ssYFP-␮stpSer (Ser), were incubated with (+) or without (–) ALLN and MG-132, pulse-labeled with [35S]methionine-[35S]cysteine and chased for the indicated time with (+) or without (–) ALLN and MG-132. Cells and medium were separated, cells were lysed and substrates were immunoprecipitated (IP) from lysed cells or medium by antibodies against GFP or Myc. Immunoprecipitates were resolved by SDS-PAGE, electroblotted and blots were exposed to autoradiography ([35S]). ssYFP-␮stpCys without N-glycan (*), resulting from de-glycosylation or inefficient glycosylation. ssYFP␮stpSer with high-mannose (arrow) or complex (arrowhead) Nglycans are indicated. (C) The graphs, representative of three independent experiments, illustrate the amounts of the indicated ssYFP fusion proteins remaining in untreated (open symbols) or ALLN- and MG-132-treated (+ Inhib; filled symbols) cells. The amounts detected in cells (circles) or recovered from medium (squares) were estimated by densitometry of autoradiograms in B, calculated as a percentage of their levels at the end of the pulse (100%), and the half-life values (see text) were calculated.

asparagine residue within ␮stp was indeed glycosylated. Mutant TPO-␮stpCysNQ was still sensitive to endoglycosidase H (endo H) (Fig. 5A), indicating that N-glycans were linked to at least some of the additional four N-glycosylation sites in the TPO protein. Likewise, the ssYFP-␮stpCysNQ mutant migrated faster than wild-type ssYFP-␮stpCys (Fig. 6A), and its electrophoretic mobility, which was similar to that of deglycosylated ssYFP-␮stpCys, was not altered upon digestion with endo H (Fig. 6A). This demonstrated that the asparagine residue in ␮stp underwent glycosylation also in the context of ssYFP-␮stpCys, establishing it as the sole N-glycosylation site in this fusion protein.

Fig. 3. The ␮stpCys element retains ssYFP in the ER. COS-7 cells were transfected with vectors encoding Myc-tagged ssYFP (myc), ssYFP-␮stpCys (Cys) or ssYFP-␮stpSer (Ser), together with galactosyl transferase-CFP (GalT-CFP). YFP (left panels) and CFP (middle panels) were visualized by confocal fluorescence microscopy individually or as merged images (right panels). Bar, 5 ␮m.

Ablation of the N-glycosylation site in ␮stp exerted distinct effects on the intracellular fate of the ␮stp-fusion proteins. It significantly accelerated the turnover of both TPO-␮stpCysNQ (Fig. 5B,C) and ssYFP-␮stpCysNQ (Fig. 6B,C), relative to their corresponding wild-type proteins. The half-life of TPO␮stpCysNQ was shortened to ~2.5 hours (Fig. 5C) and that of ssYFP-␮stpCysNQ was ~2 hours (Fig. 6C). Moreover, this mutation also considerably altered the intracellular fate of the SerNQ chimeras. While TPO-␮stpSerNQ was still secreted, albeit poorly (data not shown), ssYFP-␮stpSerNQ was tightly retained within cells (Fig. 6D). Confocal microscopy revealed that, unlike the ssYFP-␮stpSer that was localized both to the ER and the trans-Golgi (Fig. 3), ssYFP-␮stpSerNQ was restricted to the ER (Fig. 6E), as was the wild-type ssYFP␮stpCys (Fig. 3). Moreover, ssYFP-␮stpSerNQ became a rapidly degraded protein (Fig. 6B,C), exhibiting turnover rates (t1/2=1.7 hours) similar to those measured for the CysNQ construct (t1/2=2.1 hours), and both substrates were stabilized in the presence of proteasome inhibitors (Fig. 6B,C). Thus, in the case of the ␮stp-fusion reporter proteins, the absence of the most C-terminal N-glycan did not hamper their elimination by ERAD, but in fact, it accelerated the turnover of these NQ chimeras and even converted secreted ssYFP-␮stpSer to an ERAD substrate. Substituting the asparagine for glutamine in the ␮stp motif abolished the N-glycosylation site, but also introduced a mutation that could have generated a retention signal that is



4381

secreted ssYFP-␮stpSer in untreated cells (Fig. 7D, compare arrow and arrowhead; see also Fig. 2B, lower panel). Moreover, unlike the residual glycosylated fusion protein that was relatively stable, the unglycosylated ssYFP-␮stpSer was degraded rapidly (t1/2=3.5 hours), and this degradation was inhibited by proteasome inhibitors (Fig. 7D,E). To confirm further that the cysteineindependent retention and accelerated degradation were the consequence of the absence of the ␮stp N-glycan, we employed an additional mutagenesis to abolish specifically the glycosylation of the ␮stp motif, by substituting the serine in the tripeptide sequon Asn-Val-Ser for alanine, generating ssYFP-␮stpCysSA and Fig. 4. The ␮stpCys element, acting as a portable ERAD degron, does not serve as ssYFP-␮stpSerSA. This approach was used the ubiquitin conjugation site. HEK293T cells were transfected with vectors previously to generate ␮s that lacked the ␮stp Nencoding (A) the wild-type (WT) version of the Myc-tagged TPO-␮stpCys or (B) glycan and to show that this N-glycan was the TPO-␮stpCysKR mutant (KR), in which the only K residue in ␮stp was replaced by arginine. Cells were pulse-labeled with [35S]methionine-[35S]cysteine dispensable for secretion of sIgM hexamers or the and chased for the indicated time with (+) or without (–) ALLN and MG-132. ␮s double mutant S565A C575A (Cals et al., Substrates were immunoprecipitated (IP) from lysed cells by an antibody against 1996; de Lalla et al., 1998). The electrophoretic Myc, immunoprecipitates were resolved by SDS-PAGE, electroblotted and blots mobility and sensitivity to endo H confirmed that 35 were exposed to autoradiography ([ S]). (C) The graph, representative of three the ␮stp N-glycosylation was ablated by this independent experiments, illustrates the amounts of TPO-␮stpCys (WT; circles) or substitution (Fig. 8A). This ablation exerted TPO-␮stpCysKR (KR; triangles) remaining in untreated (open symbols) or ALLNeffects similar to those observed with the NQ and MG-132-treated (+ Inhib; filled symbols) HEK293T cells. The amounts mutants or the tunicamycin treatment. The estimated by densitometry of autoradiograms A and B were calculated as a turnover of ssYFP-␮stpCysSA was significantly percentage of their levels at the end of the pulse (100%), and the half-life values accelerated (t1/2 ~2.5 hours) and ssYFP(see text) were calculated. ␮stpSerSA was retained within the cells (Fig. 8D) and rapidly degraded (t1/2=2.6 hours) (Fig. 8B,C). unrelated to the glycosylation status of ␮stp. To demonstrate Degradation of both SA mutants was proteasomal, as indicated that the improved retention and accelerated degradation were by its inhibition by proteasome inhibitors (Fig. 8B,C). Finally, indeed related to the absence of the N-glycan, we generated the ER retention of ssYFP-␮stpSerSA (Fig. 8E) observed by unglycosylated ssYFP-␮stpCys and ssYFP-␮stpSer by two confocal microscopy resembled the ER localization of ssYFPadditional approaches. First, N-glycosylation was abolished ␮stpSerNQ (Fig. 6E) and wild-type ssYFP-␮stpCys (Fig. 3), altogether by treating cells with tunicamycin. As expected, this contrary to the ER and trans-Golgi localization of the wild-type treatment generated mostly unglycosylated ssYFP-␮stp ssYFP-␮stpSer (Fig. 3). To conclude, in the absence of the sole nascent proteins, but a small fraction (~20%) was still N-glycan of the ␮stp motif, ssYFP-␮stpSer was no longer glycosylated, probably reflecting the pool of the dolichol secreted and both ssYFP-␮stpSer and ssYFP-␮stpCys were pyrophosphate oligosaccharide donor (Fig. 7A,D). The rapidly degraded proteins (Figs 6-8). Thus, the ␮stp N-glycan residual glycosylation of ssYFP-␮stp chimeras provided an facilitates secretion and impedes elimination by ERAD of interesting reference to the retention and degradation of the secretory reporter proteins fused to the ␮stp sequence. unglycosylated fusion proteins. Accelerated proteasomal degradation (t1/2=2.2 hours) was observed for the Discussion unglycosylated ssYFP-␮stpCys (Fig. 7A,B, triangles), whereas Quality control mechanisms that eliminate proteins from the the 20% glycosylated ssYFP-␮stpCys was more stable (Fig. secretory pathway perform two main tasks. They prevent 7A). Importantly, tunicamycin exerted no effect on the secretion and, subsequently, target the retained proteins for turnover of the already unglycosylated ssYFP-␮stpCysNQ degradation by ERAD. It is well established that retention and mutant (Fig. 7C). This indicates that the effect of tunicamycin degradation are conferred by two types of cis-acting motifs is specific to the glycosylation status of the ssYFP-␮stp known, respectively, as ‘retention signals’ and ‘degrons’. chimeras, rather than being a consequence of abolished NHowever, it is less clear whether retention signals and degrons glycosylation of other glycoproteins. The effect of tunicamycin necessarily overlap or can be represented by separate on ssYFP-␮stpSer revealed that the secretion of the sequences, and, if so, what the interrelations between these unglycosylated form was severely hampered (Fig. 7D,E, functional motifs are. Here we show that ␮stpCys is a ciscompare triangles and circles), resembling the tight retention acting element that performs both tasks, acting as a portable of ssYFP-␮stpSerNQ (Fig. 6D). Conversely, a small fraction retention signal and also as a portable degron and therefore can of the glycosylated ssYFP-␮stpSer was secreted (Fig. 7D,E, be considered as an ERAD signal. Importantly, these functions squares) and was even processed from a high-mannose to a can be attributed to the ␮stpCys motif itself and not to its complex (Fig. 7D, arrowhead) N-glycan. However, this potential interference with the correct folding of the reporter processing was inefficient, resembling the processing of the proteins, as indicated by the fluorescence of YFP. In its role as


4382


A particularly intriguing finding in this study is the role played by the ␮stp N-glycan in sorting for retention or secretion (Table 1). Our results indicate that this N-glycan is required for secretion of the fusion protein, as its absence results in intracellular retention. However, the latter is observed only when this N-glycan is the sole one that is linked to the protein, as in the case of ssYFP. This observation is independent of the mode in which the N-glycan is abolished. Substitutions of the asparagine or serine in the tripeptide sequon Asn-Val-Ser as well as Nglycosylation inhibition by tunicamycin resulted in retention of ssYFP-␮stpSer. Conversely, proteins bearing additional glycosylation sites, albeit lacking the ␮stp N-glycan, such as TPO␮stpSerNQ or the ␮s double mutant (S565A C575A) (Fra et al., 1993; Cals et al., 1996; de Lalla et al., 1998), were still secreted. Notably, the contribution of the N-glycan to secretion is revealed only in the context of the ␮stp, as Fig. 5. The ␮stpCys N-glycan decelerates degradation of TPO. HEK293T were ssYFP-Myc, lacking ␮stp altogether, is secreted. transfected with an empty vector (mock), or with vectors encoding Myc-tagged Nonetheless, the involvement of the N-glycan in TPO-␮stpCys (WT), TPO-␮stpCysKR (KR) or TPO-␮stpCysNQ (NQ). In the NQ secretion is counteracted by the penultimate Cys mutant, the only N residue in the ␮stp motif was replaced by Q, abolishing the residue, and the net result is retention of ssYFP␮stp glycosylation site. (A) Cells were lysed, samples of cell lysate (10%) were ␮stpCys. collected and substrates were immunoprecipitated (IP) from the remaining 90% of the lysate by an antibody against Myc. Immunoprecipitates were treated with (+) or The contribution of the ␮stp N-glycan to without (–) endo H, and immunoprecipitates and cell lysates were resolved by secretion is revealed only when the Cys residue SDS-PAGE and immunoblotted (IB) with an antibody against ␮stp. Fully is replaced by Ser, as ssYFP-␮stpSer is secreted, glycosylated WT and KR, under-glycosylated NQ and endo-H-treated dewhereas ssYFP-␮stpSer lacking N-glycan is glycosylated substrates are indicated. (B) HEK293T transfected with an empty retained (Table 1). This indicates that abolishing vector (mock) or with vectors encoding Myc-tagged TPO-␮stpCys (WT) or TPON-glycosylation renders the ␮stp a retention 35 35 ␮stpCysNQ (NQ) were pulse-labeled with [ S]methionine-[ S]cysteine, chased signal, regardless of the penultimate Cys. This for the indicated time, lysed and substrates were immunoprecipitated (IP) by an finding is especially interesting as both the ␮stp antibody against Myc. Immunoprecipitates were resolved by SDS-PAGE, N-glycan and penultimate Cys are highly electroblotted and blots were exposed to autoradiography ([35S]; upper panels) to conserved throughout evolution and the ␮stp Nmonitor degradation and then immunoblotted (IB) with an antibody against ␮stp (lower panels) to estimate steady-state levels. (C) The graph, representative of glycan is important for sIgM association with the three independent experiments, illustrates the amounts of TPO-␮stpCys (WT; J-chain for pentamerization (de Lalla et al., circles) or TPO-␮stpCysNQ (NQ; triangles) remaining in HEK293T cells. The 1998). However, in the context of the ␮s heavy amounts estimated by densitometry of autoradiograms in B were calculated as a chain, this N-glycan is relevant neither for percentage of their levels at the end of the pulse (100%), and half-life values (see maintaining the Cys575 function in retention nor text) were calculated. for secretion of sIgM hexamers or the ␮s double mutant (S565A C575A) (Fra et al., 1993; Cals et a degron, the ␮stpCys element targets proteins from the al., 1996; de Lalla et al., 1998), indicating that the secretory pathway to the proteasome for degradation. unglycosylated ␮stp adopts a secretion-competent However, the functional domains within ␮stpCys can be conformation. Therefore, retention of the unglycosylated discerned, as the penultimate Cys residue plays a key role in ssYFP-␮stp suggests that, in the context of ␮stp, the N-glycan retention, whereas degradation is maintained even when this plays a positive role in secretion, rather than a negative role in residue is replaced by Ser (Table 1). generating a misfolded ␮stp that can serve as a retention signal. The function of ␮stpCys as an element necessary for Our study demonstrates that ␮stpCys is a portable ERAD retention of ␮s in B cells (Sitia et al., 1990) and sufficient for degron, acting in cis and conferring proteasomal degradation retention of lysosomal cathepsin D in COS-7 cells (Isidoro et upon two secretory reporter proteins in non-lymphoid cells. al., 1996) is extended in the current study. We show that Moreover, the turnover rates of these ␮stp fusion proteins are ␮stpCys confers ER retention upon two bona fide secretory similar to those measured for the endogenous ␮s heavy chain proteins, TPO and ssYFP, when fused to their C-termini. in B cells (Amitay et al., 1991; Amitay et al., 1992; Elkabetz Secretion of the chimeras is prevented with similar efficiency et al., 2003). Another portable C-terminal ERAD degron was in several non-lymphoid cell lines (COS-7, HEK293T or discovered recently in cyclooxygenase 2 (COX2) (Mbonye et HeLa). Consistent with previous reports, this ␮stpCysal., 2006). A segment of 19 amino acids (19-AA), located just mediated retention depends on the penultimate Cys residue, as inside the C-terminus, mediated the entry of COX2 to ERAD. its replacement by a Ser residue allows secretion of TPOThis degron was necessary for rapid degradation of COX2 ␮stpSer and ssYFP-␮stpSer. and conferred rapid turnover upon the otherwise stable



cyclooxygenase 1 (COX1) when inserted near its C-terminus (ins594-612 COX1) (Mbonye et al., 2006). Similar to the COX2 degron, ␮stp shares no apparent properties with other degrons. Unlike the artificial C-terminal degron SL17 (Gilon et al., 1998) or the N-terminal Sgk1 degron (Arteaga et al., 2006), neither the COX2 degron nor ␮stp are hydrophobic. Also, neither the COX2 degron nor ␮stp forms amphipathic helices, as opposed to the N-terminal Deg1 (Johnson et al., 1998) or Sgk1 degrons. Nevertheless, the ␮stp element is extremely conserved in mammals, indicating the significance of its entire sequence. ERAD degrons might recruit the ubiquitylation machinery and/or provide the platform for ubiquitin conjugation. It has Table 1. Roles of elements within the ␮stp motif in secretion and degradation ␮stp N-glycan

Penultimate residue

Secretion

Degradation

–* + + – –‡

–* Cys Ser Cys Ser

+ – + – –

– + – ++ ++

*Myc-tagged substrate lacking the ␮stp motif. ‡ Valid only for the ssYFP constructs that have no additional N-glycans and is irrespective of the method of N-glycan abolishment (NQ, SA or tunicamycin; see main text for further details).

4383

Fig. 6. Hampered secretion and accelerated degradation of the ssYFP-␮stpNQ mutants. (A) HEK293T cells were transfected with an empty vector (mock), with vectors encoding Myc-tagged wildtype versions of ssYFP-␮stpCys (Cys) or ssYFP-␮stpSer (Ser), or the NQ mutants of ssYFP-␮stpCys (CysNQ) or ssYFP-␮stpSer (SerNQ). In the NQ mutants, the N residue in the ␮stp was replaced by a Q residue, abolishing the only glycosylation site in the chimera. Cells were lysed, substrates were immunoprecipitated (IP) by an antibody against Myc, immunoprecipitates were treated with (+) or without (–) endo H, resolved by SDS-PAGE and immunoblotted (IB) with an antibody against Myc. Fully glycosylated WT, unglycosylated NQ and endo-H-treated de-glycosylated substrates are indicated. (B) HEK293T cells expressing ssYFP-␮stpCysNQ (CysNQ) or ssYFP-␮stpSerNQ (SerNQ) were preincubated for 1 hour and pulse-labeled with [35S]methionine-[35S]cysteine and chased for the indicated time with (+) or without (–) ALLN and MG132. Cells were lysed, substrates were immunoprecipitated (IP) by an antibody against Myc, resolved by SDS-PAGE, electroblotted and blots were exposed to autoradiography ([35S]). (C) The graph, representative of three independent experiments, illustrates the amounts of ssYFP-␮stpCysNQ (CysNQ; circles) or ssYFP␮stpSerNQ (SerNQ; squares) remaining in untreated (open symbols) or ALLN- and MG-132-treated (+ Inhib; filled symbols) cells. The amounts estimated by densitometry of autoradiograms in B were calculated as a percentage of the levels of the NQ mutants at the end of the pulse (100%), and half-life values (see text) were calculated. (D) Hela cells were transfected with an empty vector (mock), with vectors encoding Myc-tagged wild-type versions of ssYFP-␮stpCys (Cys) or ssYFP-␮stpSer (Ser), or the NQ mutants of ssYFP-␮stpCys (CysNQ) or ssYFP-␮stpSer (SerNQ). Medium was collected 40 hours post-transfection, and the cells were lysed. Substrates were immunoprecipitated (IP) from the medium by an antibody against Myc, and cell lysates (10%) and immunoprecipitates from the medium were resolved by SDS-PAGE and immunoblotted (IB) with an antibody against Myc. Substrates with complex (arrowhead) or high-mannose (arrow) N-glycans and unglycosylated NQ are indicated. (E) COS-7 cells were transfected with a combination of vectors encoding Myc-tagged ssYFP-␮stpSerNQ mutant (SerNQ) and galactosyl transferase-CFP (GalT-CFP). YFP (left panel) and CFP (middle panel) were visualized by confocal fluorescence microscopy individually or as merged images (right panels). Bar, 5 ␮m.

been shown that the N-terminal transmembrane segment of 3hydroxy-3-methylglutaryl-coenzyme A reductase serves as a necessary and sufficient ERAD degron that also provides the lysine residues required for the sterol-dependent ubiquitylation and degradation of this protein (Doolman et al., 2004; Sever et al., 2003). However, in the case of ␮stpCys, it appears that the polyubiquitin chain(s) are not conjugated to ␮stpCys itself as degradation of TPO-␮stpCys is not affected by substituting the single lysine residue within the ␮stp sequence. Hence, ubiquitin is either always conjugated to other lysine residues located within the reporter protein or ubiquitin conjugation is promiscuous enough to select alternative lysine(s) as acceptor site(s) if the preferred lysine within the ␮stp is mutated. The notion that ubiquitin conjugation to site(s) proximal to the degron is not obligatory for degradation was also reported for the yeast ERAD substrate CPY* (Spear and Ng, 2005). The dual role played by ␮stp-Cys as a retention signal and a degron can be attributed to distinct elements within the ␮stpCys motif (Table 1). As described above, the penultimate Cys is crucial for retention, and the N-glycan of ␮stp is required


4384


Fig. 7. Tunicamycin hampers secretion and accelerates degradation of the ssYFP-␮stp fusion proteins. (A) HEK293 cells expressing a vector encoding ssYFP-␮stpCys (Cys) were preincubated for 1 hour with tunicamycin. Cells were pulse-labeled with [35S]methionine-[35S]cysteine and chased for the indicated time with tunicamycin together (+) or without (–) ALLN and MG-132. Cells and medium were separated, cells were lysed and ssYFP-␮stpCys was immunoprecipitated (IP) from lysed cells or medium by an antibody against Myc. (B) The graph, representative of three independent experiments, illustrates the amounts of ssYFP-␮stpCys remaining in untreated (open symbols) or ALLNand MG-132-treated (+ Inhib; filled symbols) cells. The amounts recovered from medium (M; squares) or detected in cells as combined unglycosylated and glycosylated forms (C; circles) or as unglycosylated forms (C ung; triangles) were estimated by densitometry of autoradiogram A, calculated as a percentage of the ssYFP-␮stpCys level at the end of the pulse (100%), and half-life values (see text) were calculated. (C) HEK293T cells expressing a vector encoding ssYFP-␮stpCysNQ (CysNQ) were preincubated for 1 hour and chased with (+) or without (–) tunicamycin (Tm). Cells were pulse-labeled with [35S]methionine-[35S]cysteine and chased for the indicated time with or without tunicamycin together (+) or without (–) ALLN and MG-132. ssYFP-␮stpCysNQ was immunoprecipitated (IP) from lysed cells by an antibody against Myc. (D) HEK293 cells expressing ssYFP-␮stpSer (Ser) were preincubated for 1 hour and chased with (+) or without (–) tunicamycin (Tm). Cells were pulse-labeled with [35S]methionine-[35S]cysteine and chased for the indicated time with or without tunicamycin, together (+) or without (–) ALLN and MG-132. Cells and medium were separated, cells were lysed and substrates were immunoprecipitated (IP) from lysed cells (upper panel) or medium (lower panel) by an antibody against Myc. Immunoprecipitates were resolved by SDS-PAGE, electroblotted and blots were exposed to autoradiography ([35S]). The unglycosylated and the residual glycosylated substrates are indicated. ssYFP-␮stpSer with high-mannose (arrow) or complex (arrowhead) N-glycans are indicated. (E) The graphs, representative of three independent experiments, illustrate the amounts of ssYFP-␮stpSer remaining in untreated (open symbols) or ALLN- and MG-132-treated (+ Inhib; filled symbols) cells (upper graph) or recovered from medium (lower graph). Glycosylated forms from untreated cells (gly; circles) or unglycosylated (Tm ung; triangles) and glycosylated (Tm gly; squares) forms from tunicamycin-treated cells were estimated by densitometry of autoradiogram D, calculated as a percentage of the level of ssYFP-␮stpSer at the end of the pulse (100%), and half-life values (see text) were calculated.

for secretion. However, reporter proteins that are tightly retained, either owing to the presence of the Cys residue or absence of the N-glycan, are degraded by the proteasome, suggesting that elements other than the penultimate Cys or the N-glycan within ␮stp also contribute to the targeting to ERAD. The penultimate Cys appears to function mostly in retention, possibly by means of its association with ER thiol oxidoreductases (Anelli et al., 2002; Anelli et al., 2003), but does not seem to contribute to degradation. Even when the penultimate Cys is replaced, ␮stpSer can still target ssYFP for degradation provided retention is conferred by the absence of the N-glycan. On the other hand, the N-glycan appears to play

two roles, contributing both to secretion and stability of the reporter proteins. Clearly, in the absence of the N-glycan, degradation of the already tightly retained ␮stpCys fusion proteins is accelerated, and ssYFP-␮stpSer is converted from a secreted protein into one that is retained and rapidly degraded. Our finding that the ␮stp N-glycan plays a stabilizing role adds another layer to the role played by N-glycans in ER quality control and in ERAD (Helenius and Aebi, 2001; Ellgaard and Helenius, 2003; Helenius and Aebi, 2004; Molinari et al., 2003; Moremen and Molinari, 2006; Olivari et al., 2006). Results similar to ours were reported for RI332, a



lumenal truncated variant of ribophorin I, which was an ERAD substrate in HeLa cells (de Virgilio et al., 1999). Removal of its single N-glycosylation site resulted in degradation that was no longer biphasic, but turned into a rapid monophasic turnover of the mutant RI332-Thr. Interaction with calnexin was implicated in regulating the slower proteolytic phase (de Virgilio et al., 1999). The contribution of the most C-terminal N-glycan to ERAD was shown in yeast as well as in mammalian cells. Truncation of the C-terminal domain of CPY* stabilized this ERAD substrate. As it turned out, the lysine-devoid C-terminal domain still allowed degradation of CPY*, whereas the most C-terminal N-glycan was necessary and sufficient for ERAD (Kostova and Wolf, 2005; Spear and Ng, 2005). Interesting interrelationships between an ERAD Table 2. Sequences of oligonucleotides used for PCR or direct cloning Primer no. 1 2 3 4 5 6 7 8 9 10 11

Sequence 5⬘-GGAATTCCATATGGTAAACCCACACTGTAC-3⬘ 5⬘-ACATACATGCATGCTCAATAGCAGGTGCCGCC-3⬘ 5⬘-GGAATTCCATATGGTAAACCCACACTGTACCAGGTCTCC3⬘ 5⬘-ACATACATGCATGCTCAATAGCAGGTGCCGCC-3⬘ 5⬘-GCGAAGCTTGAGCAAAAGCTC-3⬘ 5⬘-CGCGGATCCTCAATAGCAGGTGCC-3⬘ 5⬘-CGCGGATCCTCAATAGCTGGTGCC-3⬘ 5⬘-AGCTTGAGCAAAAGCTCATTTCTGAAGAGGACTTGTGAG-3⬘ 5⬘-GATCCTCACAAGTCCTCTTCAGAAATGTGCTTTTGCTCA3⬘ 5⬘-CCACACTGTACAATGTCGCCCTGATCATGTCTGAC-3⬘ 5⬘-GTCAGACATGATCAGGGCGACATTGTACAGTGTGG-3⬘

4385

Fig. 8. Hampered secretion and accelerated degradation of the ssYFP-␮stpSA mutants. (A) HEK293T were transfected with an empty vector (mock), with vectors encoding Myc-tagged wild-type versions of ssYFP-␮stpCys (Cys) or ssYFP-␮stpSer (Ser), the NQ mutants of ssYFP-␮stpCys (CysNQ) or ssYFP-␮stpSer (SerNQ) or the SA mutants of ssYFP-␮stpCys (CysSA) or ssYFP-␮stpSer (SerSA). In the SA mutants, the S residue in ␮stp was replaced by an alanine residue, abolishing the only glycosylation site in the chimera. The cells were lysed, substrates were immunoprecipitated (IP) by an antibody against Myc, immunoprecipitates were treated with (+) or without (–) endo H, resolved by SDS-PAGE and immunoblotted (IB) with an antibody against Myc. Fully glycosylated WT, unglycosylated SA or NQ and endo-H-treated de-glycosylated substrates are indicated. Asterisk, anti-Myc antibody heavy chain. (B) HEK293T cells expressing ssYFP-␮stpCysSA (CysSA) or ssYFP-␮stpSerSA (SerSA) were preincubated for 1 hour, pulselabeled with [35S]methionine-[35S]cysteine and chased for the indicated time with (+) or without (–) ALLN and MG-132. The cells were lysed, substrates were immunoprecipitated (IP) by an antibody against Myc, resolved by SDS-PAGE, electroblotted and blots were exposed to autoradiography ([35S]). (C) The graph, representative of three independent experiments, illustrates the amounts of ssYFP␮stpCysSA (CysSA; circles) or ssYFP-␮stpSerSA (SerSA; triangles) remaining in untreated (open symbols) or ALLN- and MG-132treated (+Inhib; filled symbols) cells. The amounts estimated by densitometry of autoradiograms in B were calculated as a percentage of the levels of the SA mutants at the end of the pulse (100%), and half-life values (see text) were calculated. (D) Hela cells were transfected with an empty vector (mock), with vectors encoding Myc-tagged wild-type versions of ssYFP-␮stpCys (Cys) or ssYFP␮stpSer (Ser), or the SA mutants of ssYFP-␮stpCys (CysSA) or ssYFP-␮stpSer (SerSA). Medium was collected 40 hours posttransfection, cells were lysed, and substrates were immunoprecipitated (IP) from lysed cells and medium by an antibody against Myc. Immunoprecipitates were resolved by SDSPAGE and immunoblotted (IB) with an antibody against Myc. Substrates with complex (arrowhead) or high-mannose (arrow) Nglycans and unglycosylated SA are indicated. (E) COS-7 cells were transfected with a combination of vectors encoding Myc-tagged ssYFP-␮stpSerSA mutant (SerSA) and galactosyl transferase-CFP (GalT-CFP). YFP (left panel) and CFP (middle panel) were visualized by confocal fluorescence microscopy individually or as merged images (right panels). Bar, 5 ␮m.

degron and its N-glycan are shown here and were recently reported for COX1 and COX2 (Mbonye et al., 2006). Mutation of Asn594, an N-glycosylation site at the beginning of the 19AA degron, stabilized both COX2 and ins594-612 COX1. Nonetheless, COX mutants that were glycosylated at Asn594 but lacked the remainder of the 19-AA cassette were also stable, suggesting that glycosylation of Asn594 was necessary for COX2 degradation, but that at least some other part of the 19-AA segment was also required (Mbonye et al., 2006). Conversely, in our study, abolishing the single N-glycan in ␮stp, which is the only site for N-glycosylation in ssYFP␮stpCys and the most C-terminal (out of five potential sites) in TPO-␮stpCys, did not hamper degradation but, in fact, accelerated the turnover of both chimeras. Hence, the role of N-glycans in ERAD cannot be generalized and appears to be affected by other elements in the degron and/or reporter protein. Comparison between the turnover rates of the ssYFP␮stpCys with or without the N-glycan points to N-glycanrelated processes as potential rate-limiting steps and might

4386



provide mechanistic insights into their degradation. Within the ER lumen, ERAD substrates are subjected to several Nglycan-related quality control processes, including the calnexin-calreticulin cycle, trimming by mannosidase and recognition by EDEM (Helenius and Aebi, 2001; Ellgaard and Helenius, 2003; Helenius and Aebi, 2004; Moremen and Molinari, 2006; Olivari et al., 2006). Passage across the ER membrane, poly-ubiquitylation and de-glycosylation by peptide N-glycanase (Suzuki et al., 2002) appear to occur in a coupled fashion coordinated by the p97/Cdc48 complex (Bar-Nun, 2005). This AAA-ATPase provides the driving force for dislocation (Ye et al., 2001; Elkabetz et al., 2004), associates with ERAD-dedicated E3 ligases (Zhong et al., 2004; Ye et al., 2005; Lilley and Ploegh, 2005; Schuberth and Buchberger, 2005; Neuber et al., 2005), mediates the binding of peptide N-glycanase to proteasomes (Li et al., 2005; Allen et al., 2006) and interacts with ubiquitin (Dai and Li, 2001) and deubiquitylating enzymes such as Otu1 and ataxin-3 (Doss-Pepe et al., 2003; Rumpf and Jentsch, 2006; Zhong and Pittman, 2006; Wang et al., 2006). It remains to be established which of the N-glycan-related processes is the rate-limiting step that is bypassed by the proteins fused to the unglycosylated ␮stp. Materials and Methods DNA constructs Plasmids encoding the fusion proteins were constructed in the pcDNA3 vector (Invitrogen). From a plasmid containing cDNA of the full-length human thyroid peroxidase [accession # J02970 (Magnusson et al., 1987); generously provided by R. Magnusson], the transmembrane domain (residues 849-933) was truncated and residues 1-848 were fused to a Myc tag sequence (EQKLISEEDLN) followed by the ␮stpCys/Ser sequence. The latter were generated by PCR from plasmid pSV-V␮1 (Neuberger, 1983), generously provided by R. Sitia. To introduce mutations into ␮stpCys, a fragment of about 1 kb encoding the last third of TPOMyc-␮stpCys was excised from the TPO-Myc-␮stpCys pcDNA3 vector, using ClaI and XbaI (ClaI-XbaI fragment), and cloned into pBCSK(+) (Stratagene) at these sites. PCR with TPO-Myc-␮stpCys as a template was used to generate the KR mutation, using forward primer no. 1 and reverse primer no. 2, and the NQ mutation, using forward primer no. 3 and reverse primer no. 4 (for sequences of PCR primers and oligonucleotides, see Table 2). The PCR products were cut with SphI and NdeI and cloned into these sites in the pBCSK(+) vector containing the ClaI-XbaI fragment. The ClaI-XbaI fragment was then cut and cloned back into the TPO-Myc-␮stpCys pcDNA3 vector and the resulting plasmids were sequenced. Plasmid pEYFP-C1 (Clontech) encodes yellow fluorescent protein (YFP), and ssYFP was generated by inserting the hen egg lysozyme signal sequence upstream of the YFP sequence. The ssYFP-encoding sequence was excised with NheI and HindIII and cloned into NheI- and HindIII-digested pcDNA3, generating pcDNA3-ssYFP. The Myc-␮stp was amplified by PCR using TPO-Myc-␮stpCys pcDNA3 as a template, introducing HindIII and BamHI sites at the 5⬘ and 3⬘ ends, respectively. Both Myc-␮stpCys and Myc-␮stpSer were generated using the same forward primer no. 5 and two different reverse primers, no. 6 for ␮stpCys and no. 7 for ␮stpSer. The ssYFP-Myc-␮stp expression plasmids (pcDNA3-ssYFP-Myc␮stpCys/Ser) were constructed by cloning the HindIII- and BamHI-digested PCR products into the HindIII- and BamHI-digested pcDNA3-ssYFP. For comparison, pcDNA3-ssYFP-Myc (with no ␮stp) was constructed by ligating the HindIII- and BamHI-digested Myc sequence (sense strand no. 8 and antisense strand no. 9) into the HindIII- and BamHI-digested pcDNA3-ssYFP. For constructing the ssYFP-Myc-␮stp NQ mutants, an intermediate plasmid, pBC-Myc-␮stp, was constructed by cloning a PCR product into pBCSK(+). The PCR reaction used primers no. 5 and no. 6 and pcDNA3-ssYFP-Myc-␮stpCys as a template, and the HindIII- and BamHI-digested PCR products were inserted into pBCSK(+) digested with the same enzymes. To introduce the NQ mutation into the ␮stp motif, both ␮stpCysNQ and ␮stpSerNQ were amplified using pBC-Myc␮stp as a template, the same forward primer no. 3 and two different reverse primers, no. 6 for ␮stpCysNQ and no. 7 for ␮stpSerNQ. The NdeI- and BamHIdigested PCR products were cloned into the same sites in the pBC-Myc-␮stpCys. The Myc-␮stpCysNQ and Myc-␮stpSerNQ fragments were excised from the respective pBC-Myc-␮stpNQ plasmids by HindIII and BamHI and cloned back into the pcDNA3-ssYFP. The ssYFP-Myc-␮stp SA mutants were constructed by PCR reaction using primers no. 11 and no. 12 and pcDNA3-ssYFP-Myc-␮stpCys

and ssYFP-Myc-␮stpSer as templates, employing a QuikChange site-directed mutagenesis kit (Stratagene).

Cell lines and transfection COS-7, HeLa or HEK293(T) cells were grown at 37°C in a humidified 5% CO2 incubator in Dulbecco’s modified Eagle’s medium (DMEM; Invitrogen Life Technologies), supplemented with 10% (v/v) fetal calf serum, 1 mM L-glutamine, 2 ␮g/ml penicillin, 20 ␮g/ml streptomycin and 2.5 ␮g/ml nystatin (Biological Industries, Beit Ha’Emek). When they reached ~40% confluence, the cells were transiently transfected with 4-10 ␮g of the appropriate DNA construct by the calcium phosphate method (Chen and Okayama, 1988) or using jetPEI (PolyplusTransfection). Analysis of the cells was performed 24-48 hours post-transfection. Stable transfectants were selected with G418.

Protein secretion and degradation, immunoprecipitation, immunoblotting and endoglycosidase H treatment Secretion was followed by estimation of radiolabeled or unlabeled proteins accumulated in the medium. Degradation was followed by pulse-chase experiments of radiolabeled cells. Cells were starved for 1 hour in methionine/cysteine-deficient medium, pulse-labeled for 1 hour with 50 ␮Ci/ml of [35S]methionine-[35S]cysteine (Promix), and chased in the presence of excess unlabeled methionine-cysteine. Where indicated, tunicamycin (10 ␮g/ml) or proteasome inhibitors N-acetyl-leucylleucyl-norleucinal (ALLN; 50 ␮M) and carboxybenzyl-leucil-leucil-leucinal (MG132; 5 ␮M) were present throughout the starvation, pulse-labeling and chase. At various time points, medium was collected, cells were washed twice with PBS and lysed in fresh ice-cold lysis buffer, as described previously (Amitay et al., 1991). The ␮stp fusion proteins were immunoprecipitated from cell lysates and media using an excess of mouse anti-Myc (clone 9E10) or rabbit anti-GFP (Abcam), followed by protein A-sepharose (Repligen). Immunoprecipitated material was collected by centrifugation (500 g, 2 minutes, 4°C) and washed three times with PBS containing 1% (v/v) Nonidet P-40. Immunoprecipitated proteins or cell extracts were resolved by SDS-PAGE and electroblotted onto nitrocellulose. Radiolabeled blots were first exposed to autoradiography and then probed with either horseradish peroxidase (HRP)-conjugated specific antibodies or with specific primary antibodies followed by the respective HRP-conjugated secondary antibodies. The HRP was visualized by the enhanced chemiluminescence (ECL) reaction. The primary antibodies used were: mouse anti-Myc (clone 9E10) and rabbit anti-␮stp (Rabinovich et al., 2002). The HRP-conjugated secondary antibodies used were: goat anti-mouse IgG (Jackson) and goat-anti-rabbit IgG (Sigma). The HRPconjugated specific antibody used was mouse anti-Myc-HRP (Upstate Millipore). For endo H treatment, immunoprecipitated material was incubated for 1 hour at 37°C with 0.25 IU/ml endo H (New England Biolabs) according to the manufacturer’s protocol.

Confocal laser-scanning microscopy and live-cell imaging Transfected cells, grown in Labtek chambers in DMEM (without phenol red) supplemented with 20 mM HEPES pH 7.4, were imaged with a Zeiss LSM PASCAL equipped with an Axiovert 200 inverted microscope. Argon 458 nm and 514 nm laser lines were used for ECFP and EYFP, respectively. Confocal images were captured using a 63⫻ 1.4 NA objective with a pinhole diameter between 1 and 2 Airy units. Images were generated and analyzed using the Zeiss LSM software and modified using Adobe Photoshop.

We thank the members of our groups and Joseph Roitelman for critical reading of the manuscript, Eran Bosis for helping with bioinformatic analyses, Yair Argon for suggesting the TPO and R. Magnusson for providing the TPO plasmid. This work was supported in part by grants from ISF, BSF and the Public Committee for the Allocation of Estate Fund, The Israeli Ministry of Justice (to S.B.-N.) and a grant from ISF (to K.H.).

References Allen, M. D., Buchberger, A. and Bycroft, M. (2006). The PUB domain functions as a p97 binding module in human peptide N-glycanase. J. Biol. Chem. 281, 25502-25508. Amitay, R., Bar-Nun, S., Haimovich, J., Rabinovich, E. and Shachar, I. (1991). Posttranslational regulation of IgM expression in B lymphocytes. J. Biol. Chem. 266, 12568-12573. Amitay, R., Shachar, I., Rabinovich, E., Haimovich, J. and Bar-Nun, S. (1992). Degradation of secretory IgM in B lymphocytes occurs in a post-endoplasmic reticulum compartment and is mediated by a cysteine protease. J. Biol. Chem. 267, 20694-20700. Anelli, T., Alessio, M., Mezghrani, A., Simmen, T., Talamo, F., Bachi, A. and Sitia, R. (2002). ERp44, a novel endoplasmic reticulum folding assistant of the thioredoxin family. EMBO J. 21, 835-844. Anelli, T., Alessio, M., Bachi, A., Bergamelli, L., Bertoli, G., Camerini, S., Mezghrani, A., Ruffato, E., Simmen, T. and Sitia, R. (2003). Thiol-mediated protein retention in the endoplasmic reticulum: the role of ERp44. EMBO J. 22, 5015-5022.


Retention signals and degrons in ERAD Arteaga, M. F., Wang, L., Ravid, T., Hochstrasser, M. and Canessa, C. M. (2006). An amphipathic helix targets serum and glucocorticoid-induced kinase 1 to the endoplasmic reticulum-associated ubiquitin-conjugation machinery. Proc. Natl. Acad. Sci. USA 103, 11178-11183. Bar-Nun, S. (2005). The role of p97/Cdc48p in endoplasmic reticulum-associated degradation: from the immune system to yeast. Curr. Top. Microbiol. Immunol. 300, 95-125. Bonifacino, J. S. and Weissman, A. M. (1998). Ubiquitin and the control of protein fate in the secretory and endocytic pathways. Annu. Rev. Cell Dev. Biol. 14, 19-57. Brenckle, R. and Kornfeld, R. (1980). Structure of the oligosaccharides of mouse immunoglobulin M secreted by the MOPC 104E Plasmacytoma. Arch. Biochem. Biophys. 201, 160-173. Cals, M.-M., Guenzi, S., Carelli, S., Simmen, T., Sparvoli, A. and Sitia, R. (1996). IgM polymerization inhibits the Golgi-mediated processing of the mu-chain carboxyterminal glycans. Mol. Immunol. 33, 15-24. Chen, C. A. and Okayama, H. (1988). Calcium phosphate-mediated gene transfer: a highly efficient transfection system for stably transforming cells with plasmid DNA. Biotechniques 6, 632-638. Dai, R. M. and Li, C. C. (2001). Valosin-containing protein is a multi-ubiquitin chaintargeting factor required in ubiquitin-proteasome degradation. Nat. Cell Biol. 3, 740744. de Lalla, C., Fagioli, C., Cessi, F. S., Smilovich, D. and Sitia, R. (1998). Biogenesis and function of IgM: the role of the conserved mu-chain tailpiece glycans. Mol. Immunol. 35, 837-845. de Virgilio, M., Kitzmuller, C., Schwaiger, E., Klein, M., Kreibich, G. and Ivessa, N. E. (1999). Degradation of a short-lived glycoprotein from the lumen of the endoplasmic reticulum: the role of N-linked glycans and the unfolded protein response. Mol. Biol. Cell 10, 4059-4073. Doolman, R., Leichner, G. S., Avner, R. and Roitelman, J. (2004). Ubiquitin is conjugated by membrane ubiquitin ligase to three sites, including the N terminus, in transmembrane region of mammalian 3-hydroxy-3-methylglutaryl coenzyme A reductase: implications for sterol-regulated enzyme degradation. J. Biol. Chem. 279, 38184-38193. Doss-Pepe, E. W., Stenroos, E. S., Johnson, W. G. and Madura, K. (2003). Ataxin-3 interactions with rad23 and valosin-containing protein and its associations with ubiquitin chains and the proteasome are consistent with a role in ubiquitin-mediated proteolysis. Mol. Cell. Biol. 23, 6469-6483. Elkabetz, Y., Kerem, A., Tencer, L., Winitz, D., Kopito, R. R. and Bar-Nun, S. (2003). Immunoglobulin light chains dictate vesicular transport-dependent and -independent routes for IgM degradation by the ubiquitin-proteasome pathway. J. Biol. Chem. 278, 18922-18929. Elkabetz, Y., Shapira, I., Rabinovich, E. and Bar-Nun, S. (2004). Distinct steps in dislocation of luminal endoplasmic reticulum-associated degradation substrates: roles of endoplamic reticulum-bound p97/Cdc48p and proteasome. J. Biol. Chem. 279, 3980-3989. 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 of human thyroperoxidase in the endoplasmic reticulum involves two different pathways depending on the folding state of the protein. J. Biol. Chem. 275, 15948-15954. Fra, A. M., Fagioli, C., Finazzi, D., Sitia, R. and Alberini, C. M. (1993). Quality control of ER synthesized proteins: an exposed thiol group as a three-way switch mediating assembly, retention and degradation. EMBO J. 12, 4755-4761. Gilon, T., Chomsky, O. and Kulka, R. G. (1998). Degradation signals for ubiquitin system proteolysis in Saccharomyces cerevisiae. EMBO J. 17, 2759-2766. Helenius, A. and Aebi, M. (2001). Intracellular functions of N-linked glycans. Science 291, 2364-2369. Helenius, A. and Aebi, M. (2004). Roles of N-linked glycans in the endoplasmic reticulum. Annu. Rev. Biochem. 73, 1019-1049. Isidoro, C., Maggioni, C., Demoz, M., Pizzagalli, A., Fra, A. M. and Sitia, R. (1996). Exposed thiols confer localization in the endoplasmic reticulum by retention rather than retrieval. J. Biol. Chem. 271, 26138-26142. Johnson, P. R., Swanson, R., Rakhilina, L. and Hochstrasser, M. (1998). Degradation signal masking by heterodimerization of MATalpha2 and MATa1 blocks their mutual destruction by the ubiquitin-proteasome pathway. Cell 94, 217-227. Kostova, Z. and Wolf, D. H. (2003). For whom the bell tolls: protein quality control of the endoplasmic reticulum and the ubiquitin-proteasome connection. EMBO J. 22, 2309-2317. Kostova, Z. and Wolf, D. H. (2005). Importance of carbohydrate positioning in the recognition of mutated CPY for ER-associated degradation. J. Cell Sci. 118, 1485-1492. Li, G., Zhou, X., Zhao, G., Schindelin, H. and Lennarz, W. J. (2005). Multiple modes of interaction of the deglycosylation enzyme, mouse peptide N-glycanase, with the proteasome. Proc. Natl. Acad. Sci. USA 102, 15809-15814. Lilley, B. N. and Ploegh, H. L. (2005). Multiprotein complexes that link dislocation, ubiquitination, and extraction of misfolded proteins from the endoplasmic reticulum membrane. Proc. Natl. Acad. Sci. USA 102, 14296-14301.

4387

Magnusson, R. P., Chazenbalk, G. D., Gestautas, J., Seto, P., Filetti, S., DeGroot, L. J. and Rapoport, B. (1987). Molecular cloning of the complementary deoxyribonucleic acid for human thyroid peroxidase. Mol. Endocrinol. 1, 856-861. Mbonye, U. R., Wada, M., Rieke, C. J., Tang, H. Y., DeWitt, D. L. and Smith, W. L. (2006). The 19-amino acid cassette of cyclooxygenase-2 mediates entry of the protein into the endoplasmic reticulum-associated degradation system. J. Biol. Chem. 281, 35770-35778. McCracken, A. A. and Brodsky, J. L. (2003). Evolving questions and paradigm shifts in endoplasmic-reticulum-associated degradation (ERAD). BioEssays 25, 868-877. Molinari, M., Calanca, V., Galli, C., Lucca, P. and Paganetti, P. (2003). Role of EDEM in the release of misfolded glycoproteins from the calnexin cycle. Science 299, 13971400. Moremen, K. W. and Molinari, M. (2006). N-linked glycan recognition and processing: the molecular basis of endoplasmic reticulum quality control. Curr. Opin. Struct. Biol. 16, 592-599. Neuber, O., Jarosch, E., Volkwein, C., Walter, J. and Sommer, T. (2005). Ubx2 links the Cdc48 complex to ER-associated protein degradation. Nat. Cell Biol. 7, 993-998. Neuberger, M. S. (1983). Expression and regulation of immunoglobulin heavy chain gene transfected into lymphoid cells. EMBO J. 2, 1373-1378. Olivari, S., Cali, T., Salo, K. E., Paganetti, P., Ruddock, L. W. and Molinari, M. (2006). EDEM1 regulates ER-associated degradation by accelerating demannosylation of folding-defective polypeptides and by inhibiting their covalent aggregation. Biochem. Biophys. Res. Commun. 349, 1278-1284. Rabinovich, E., Kerem, A., Frohlich, K. U., Diamant, N. and Bar-Nun, S. (2002). AAA-ATPase p97/Cdc48p, a cytosolic chaperone required for endoplasmic reticulumassociated protein degradation. Mol. Cell. Biol. 22, 626-634. Rawitch, A. B., Pollock, G., Yang, S. X. and Taurog, A. (1992). Thyroid peroxidase glycosylation: the location and nature of the N-linked oligosaccharide units in porcine thyroid peroxidase. Arch. Biochem. Biophys. 297, 321-327. Rumpf, S. and Jentsch, S. (2006). Functional division of substrate processing cofactors of the ubiquitin-selective Cdc48 chaperone. Mol. Cell 21, 261-269. Sayeed, A. and Ng, D. T. (2005). Search and destroy: ER quality control and ERassociated protein degradation. Crit. Rev. Biochem. Mol. Biol. 40, 75-91. Schuberth, C. and Buchberger, A. (2005). Membrane-bound Ubx2 recruits Cdc48 to ubiquitin ligases and their substrates to ensure efficient ER-associated protein degradation. Nat. Cell Biol. 7, 999-1006. Sever, N., Song, B. L., Yabe, D., Goldstein, J. L., Brown, M. S. and DeBose-Boyd, R. A. (2003). Insig-dependent ubiquitination and degradation of mammalian 3-hydroxy3-methylglutaryl-CoA reductase stimulated by sterols and geranylgeraniol. J. Biol. Chem. 278, 52479-52490. Shachar, I., Amitay, R., Rabinovich, E., Haimovich, J. and Bar-Nun, S. (1992). Polymerization of secretory IgM in B lymphocytes is prevented by a preceding targeting to a degradation pathway. J. Biol. Chem. 267, 24241-24247. Sidman, C. (1981). Differing requirements for glycosylation in the secretion of related glycoproteins is determined neither by the producing cell nor by the relative number of oligosaccharide units. J. Biol. Chem. 256, 9374-9376. Sidman, C., Potash, M. J. and Kohler, G. (1981). Roles of protein and carbohydrate in glycoprotein processing and secretion. Studies using mutants expressing altered IgM mu chains. J. Biol. Chem. 256, 13180-13187. Sitia, R. and Braakman, I. (2003). Quality control in the endoplasmic reticulum protein factory. Nature 426, 891-894. Sitia, R., Neuberger, M., Alberini, C., Bet, P., Fra, A., Valetti, C., Williams, G. and Milstein, C. (1990). Developmental regulation of IgM secretion: the role of the carboxy-terminal cysteine. Cell 60, 781-790. Spear, E. D. and Ng, D. T. (2005). Single, context-specific glycans can target misfolded glycoproteins for ER-associated degradation. J. Cell Biol. 169, 73-82. Suzuki, T., Park, H., and Lennarz,W. J. (2002). Cytoplasmic peptide:N-glycanase (PNGase) in eukaryotic cells: occurrence, primary structure, and potential functions. FASEB J. 16, 635-641. Trombetta, E. S. and Parodi, A. J. (2003). Quality control and protein folding in the secretory pathway. Annu. Rev. Cell Dev. Biol. 19, 649-676. Varshavsky, A. (1991). Naming a targeting signal. Cell 64, 13-15. Wang, Q., Li, L. and Ye, Y. (2006). Regulation of retrotranslocation by p97-associated deubiquitinating enzyme ataxin-3. J. Cell Biol. 174, 963-971. Ye, Y., Meyer, H. H. and Rapoport, T. A. (2001). The AAA ATPase Cdc48/p97 and its partners transport proteins from the ER into the cytosol. Nature 414, 652-656. Ye, Y., Shibata, Y., Kikkert, M., van Voorden, S., Wiertz, E. and Rapoport, T. A. (2005). Inaugural Article: Recruitment of the p97 ATPase and ubiquitin ligases to the site of retrotranslocation at the endoplasmic reticulum membrane. Proc. Natl. Acad. Sci. USA 102, 14132-14138. Zhong, X. and Pittman, R. N. (2006). Ataxin-3 binds VCP/p97 and regulates retrotranslocation of ERAD substrates. Hum. Mol. Genet. 15, 2409-2420. Zhong, X., Shen, Y., Ballar, P., Apostolou, A., Agami, R. and Fang, S. (2004). AAA ATPase p97/valosin-containing protein interacts with gp78, a ubiquitin ligase for endoplasmic reticulum-associated degradation. J. Biol. Chem. 279, 45676-45684.