STT3B-Dependent Posttranslational N-Glycosylation as a Surveillance System for Secretory Protein

Molecular Cell

Article STT3B-Dependent Posttranslational N-Glycosylation as a Surveillance System for Secretory Protein Takashi Sato,1,2,7,8,* Yasuhiro Sako,1,8 Misato Sho,1,8 Mamiko Momohara,1 Mary Ann Suico,1 Tsuyoshi Shuto,1 Hideki Nishitoh,3 Tsukasa Okiyoneda,4 Koichi Kokame,5 Masayuki Kaneko,6 Manabu Taura,1 Masanori Miyata,1 Keisuke Chosa,1 Tomoaki Koga,1 Saori Morino-Koga,1 Ikuo Wada,2 and Hirofumi Kai1,* 1Department of Molecular Medicine, Global COE Cell Fate Regulation Research and Education Unit, Graduate School of Pharmaceutical Sciences, Kumamoto University, 5-1 Oe-Honmachi, Kumamoto 862-0973, Japan 2Department of Cell Science, Institute of Biomedical Sciences, Fukushima Medical University School of Medicine, Fukushima 960-1295, Japan 3Division of Biochemistry and Molecular Biology, Department of Medical Sciences, University of Miyazaki, 5200 Kihara, Kiyotake, Miyazaki 889-1692, Japan 4Department of Physiology, McGill University, Montreal, QC H3G 1Y6, Canada 5Department of Molecular Pathogenesis, National Cerebral and Cardiovascular Center, 5-7-1 Fujishirodai, Suita, Osaka 565-8565, Japan 6Department of Pharmacology, Faculty of Pharmaceutical Sciences, Chiba Institute of Science, Choshi, Chiba 288-0025, Japan 7Present Address: Department of Cell Fate Control, Institute of Molecular Embryology and Genetics, Kumamoto University, 2-2-1 Honjo, Kumamoto 860-0811, Japan 8These authors contributed equally to this work *Correspondence: [email protected] (T.S.), [email protected] (H.K.) DOI 10.1016/j.molcel.2012.04.015

SUMMARY

Nascent secretory proteins are extensively scrutinized at the endoplasmic reticulum (ER). Various signatures of client proteins, including exposure of hydrophobic patches or unpaired sulfhydryls, are coordinately utilized to reduce nonnative proteins in the ER. We report here the cryptic N-glycosylation site as a recognition signal for unfolding of a natively nonglycosylated protein, transthyretin (TTR), involved in familial amyloidosis. Folding and ERassociated degradation (ERAD) perturbation analyses revealed that prolonged TTR unfolding induces externalization of cryptic N-glycosylation site and triggers STT3B-dependent posttranslational N-glycosylation. Inhibition of posttranslational N-glycosylation increases detergent-insoluble TTR aggregates and decreases cell proliferation of mutant TTR-expressing cells. Moreover, this modification provides an alternative pathway for degradation, which is EDEM3-mediated N-glycan-dependent ERAD, distinct from the major pathway of Herpmediated N-glycan-independent ERAD. Hence we postulate that STT3B-dependent posttranslational N-glycosylation is part of a triage-salvage system recognizing cryptic N-glycosylation sites of secretory proteins to preserve protein homeostasis. INTRODUCTION In eukaryotic cells, secretory proteins are translocated into the endoplasmic reticulum (ER), where they undergo several cova-

lent modifications including cotranslational N-glycosylation catalyzed by the oligosaccharyltransferase (OST) complex. This occurs in synchronization with protein folding and oligomerization assisted by molecular chaperones and folding enzymes. While correctly folded and assembled proteins are prompted to their final destinations, partially folded or misfolded proteins are retained in the ER. To distinguish mature proteins from immature proteins, exposure of structural elements such as hydrophobic patches that are usually buried within proteins and/or unpaired sulfhydryls on normally disulfide-bonded cysteines are utilized as recognition signatures for unfolding (Hegde and Ploegh, 2010). This mechanism, collectively called ER quality control (QC) system (Ellgaard and Helenius, 2003), facilitates the retrotranslocation of terminally misfolded proteins across the ER membrane to be degraded by the cytosolic proteasome in a process referred to as ER-associated degradation (ERAD) (Ellgaard and Helenius, 2003; Vembar and Brodsky, 2008). Hence, strict selection of potential ERAD substrates from properly folded proteins is required to preserve protein homeostasis (proteostasis) (Balch et al., 2008). Extensive studies demonstrate that molecular mechanism of ERQC depends on the cargo’s characteristics (i.e., topology and N-glycosylation) (Bernasconi et al., 2010; Christianson et al., 2008; Nakatsukasa and Brodsky, 2008). For glycoproteins, a dynamic ‘‘glycan code’’ displays the folding status of a multitude of glycoproteins, and the progressive trimming of terminal mannose residues on mannose9-N-acetylglucosamine2 (Man9 GlcNAc2) asparagine-linked glycans by mannosidases serves as a molecular timer that marks a period during which the polypeptide is provided with multiple attempts to fold properly (Helenius and Aebi, 2004; Liu et al., 1999). In contrast to the well-characterized ERQC of glycoproteins, the precise mechanism of ERQC for nonglycosylated ERAD substrates remains largely unknown. Moreover, while N-glycan-dependent and -independent ERAD pathways may couple to overcome Molecular Cell 47, 99–110, July 13, 2012 ª2012 Elsevier Inc. 99

Molecular Cell Posttranslational N-Glycosylation of TTR Variants

A

B

UBE2J1

C

UBE2G2

HRD1

Mock WT C89S

Mock WT C91S

Mock WT C329S

(kD) 150

(kD) 150

150

50

50

50

(kD)

TTR

TTR 20

TTR

20

D18G TTR

15

Fold

20

D18G TTR

15

1.0 1.9 2.2

HMW D18G TTR D18G TTR

15

Fold 1.0 1.6 2.0

Fold 1.0 1.8 2.5

Myc

Myc-UBE2J1

Myc

Myc-UBE2G2

Myc

HRD1-Myc/His

Actin

Actin

Actin

Actin

Actin

Actin

gp78

D

RMA1

E

Mock WT R2M (kD) 150

(kD) 150

50

50

50

TTR 20

20

20

D18G TTR

15

Fold

HMW D18G TTR D18G TTR

15

D18G TTR

15

Fold

1.0 0.9 1.0

Mock WT QQ

(kD) 150

TTR

TTR

p97

F

Mock WT C42S

Fold

1.0 1.0 1.1

1.0 1.8 2.6

Myc

gp78-Myc/His

Flag

RMA1-Flag

His

His-p97

Actin

Actin

Actin

Actin

Actin

Actin

H

20

HMW D18G TTR D18G TTR 1 0.8 1.1 1.6

Derlin-1

Derlin-1

Derlin-2

Derlin-2

J

Actin

Actin

Remaining D18G TTR (%)

Fold

Remaining D18G TTR (%)

TTR

*

100 80

**

60 40 Mock C91S UBE2J1 C89S UBE2G2

20 0 0

1.5 3 Chase (h)

K **

100

***

80 60 40 Mock C329S HRD1 p97 QQ

20 0 0

1.5 3 Chase (h)

Remaining D18G TTR (%)

50

1.2 1.0 0.8 0.6 0.4 0.2 0

siDerlin-3

(kD) 150

siCON

siRNA

15

I

Relative quantity of Derlin-3 mRNA

G

100 **

80

***

60 40

siCON siDerlin-1 siDerlin-2 siDerlin-3

20 0 0

1.5 3 Chase (h)

Figure 1. Identification of a Set of Molecules Involved in ERAD of D18G TTR (A–F) Steady-state expression of D18G TTR and high molecular weight (HMW) D18G TTR in HEK293 cells cotransfected with D18G TTR and epitope-tagged wildtype (WT) or dominant negative form of E2 enzymes (A and B), E3 ligases (C–E), and p97 (F) were determined by immunoblotting. (G) Steady-state expressions of D18G TTR in 50 nM Derlin siRNA-transfected HEK293 cells stably expressing D18G TTR were determined by immunoblotting. (H) Knockdown efficiency of Derlin-3 gene by siRNA was assessed by real-time quantitative PCR. Derlin-3 mRNA level was normalized to GAPDH (internal control). Data are means ± SEM from three independent experiments.

100 Molecular Cell 47, 99–110, July 13, 2012 ª2012 Elsevier Inc.

Molecular Cell Posttranslational N-Glycosylation of TTR Variants

accumulation of numerous misfolded proteins, the relationship between both pathways is unclear. Transthyretin (TTR) is a soluble nonglycosylated secretory protein, which forms a homotetramer and functions as a transporter of thyroxine and holo-retinol binding protein in the blood and cerebral spinal fluid (Hamilton and Benson, 2001). Mutation in the TTR gene is responsible for familial amyloid polyneuropathy (FAP) and familial amyloid cardiomyopathy (Sousa and Saraiva, 2003). The vast majority of TTR variants, all kinetically and thermodynamically less stable than wild-type (WT) TTR, escape ERQC and form aggregates or amyloid fibrils in the extracellular space. In contrast, the most destabilized TTR variant, D18G TTR, is subject to ERAD, resulting in very low level secretion and mild late-onset phenotype (Sato et al., 2007; Sekijima et al., 2005; Susuki et al., 2009). Therefore, ERQC protects against severe early-onset systemic amyloidosis caused by highly destabilized TTR variants (Hammarstro¨m et al., 2003; Sekijima et al., 2005). However, the molecular mechanism of how cells manipulate and efficiently remove the highly amyloidogenic TTR in the ER is obscure. Here we show that excessive unfolding of a nonglycosylated protein TTR exposes the cryptic N-glycosylation site and triggers STT3B-dependent posttranslational N-glycosylation in the ER. This increases the solubility of the aggregation-prone TTR and shifts the ERAD of TTR variants from an N-glycan-independent to an N-glycan-dependent pathway. We suggest that a cryptic glycosylation site functions as a recognition signal for unfolding, and posttranslational N-glycosylation plays a crucial role in ERQC to strictly preserve proteostasis. RESULTS D18G TTR Is Degraded through UBE2J1, UBE2G2, HRD1, Derlin-3 and p97 To identify the molecules involved in ERAD of TTR variants, we investigated the effect of perturbation of ERAD components (e.g., ubiquitin E2, E3 enzymes, and retrotranslocon complex [Chen et al., 2006; Kaneko et al., 2002; Lenk et al., 2002; Lilley and Ploegh, 2004; Oda et al., 2006; Tiwari and Weissman, 2001; Ye et al., 2001; Ye et al., 2004; Younger et al., 2006]) on the steady-state expression of D18G TTR in HEK293 cells using dominant negative mutants and small interfering RNA (siRNA). For dominant negative mutants of ubiquitin E2 and E3 enzymes, a mutation was introduced, respectively, at the catalytic cysteine residue that accepts a charged ubiquitin from E1 enzyme and at the catalytic cysteine residue in the RING finger domain that is required for substrate ubiquitination. E2 mutants C91S UBE2J1 and C89S UBE2G2, and E3 mutant C329S HRD1, but not R2M gp78 or C42S RMA1, induced the accumulation of D18G TTR compared with mock-transfected cells (Figures 1A– 1E). Among the retrotranslocon complex components, dominant negative p97 mutant (p97QQ), which inhibits ATP hydrolysis required for retrotranslocation of ERAD substrate, and siDerlin-3

increased the level of D18G TTR (Figures 1F and 1G). Knockdown efficiency of siDerlin-3 was assessed (Figures 1H and S1A available online). The ability of WT UBE2J1, WT UBE2G2, and WT HRD1 to induce accumulation of D18G TTR is consistent with previous studies where the overexpression of ubiquitination enzymes or ligase actually inhibited ERAD instead of increasing protein turnover (Kikkert et al., 2004; Lenk et al., 2002; OkudaShimizu and Hendershot, 2007; Shen et al., 2007; Tiwari and Weissman, 2001). The degradation kinetics of D18G TTR determined by cycloheximide (CHX) chase was significantly slowed by dominant negative mutants of UBE2J1, UBE2G2, HRD1, and p97 or siDerlin-3 (Figures 1I–1K and S1B–S1D). Moreover, D18G TTR was coimmunoprecipitated with these ERAD components except UBE2G2 (Figures S1E–S1H). While Myc-UBE2J1 was coimmunoprecipitated with D18G TTR through endogenous HRD1 (Figure S1I, lanes 1 and 2), association of D18G TTR with Myc-UBE2G2 was hardly detected in normal condition (Figure S1J, lane 1). When retrotranslocation of D18G TTR was inhibited by p97QQ, interaction between D18G TTR and MycUBE2G2 was slightly observed (Figure S1J, lane 2). These results suggest that UBE2J1, UBE2G2, HRD1, p97, and Derlin-3 are the elements controlling the ERAD of D18G TTR. ERAD Inhibition Accumulates N-Glycosylated Form of D18G TTR In the screening process, we noticed the appearance of a high molecular weight (HMW) form of D18G TTR when cotransfected with C329S HRD1, p97QQ, or siDerlin-3 (Figures 1C, 1F, and 1G). Because the increase in molecular weight was about 2–3 kDa, we hypothesized that this modification may be N-glycosylation rather than ubiquitination. As expected, the HMW band was Endo H and PNGase F sensitive (Figures 2A and 2B) and abolished by an N-glycosylation inhibitor tunicamycin (Figures 2C and 2D), indicating that the HMW band was indeed the N-glycosylated TTR (designated as TTR+CHO in the figures). TTR contains a consensus glycosylation site (sequon), NDS, near the C terminus, which is conserved among mammals (Figures 2E and 2F). To confirm the glycosylation site of TTR, we mutated an asparagine residue (N98) of the putative glycosylation site to aspartic acid or glutamine. As expected, the HMW band was not observed in the mutants in p97QQ- or C329S HRD1-transfected cells (Figures 2G and 2H), indicating that N-glycosylation occurs at N98 residue if D18G TTR is accumulated in the ER. D18G TTR Is Subject to STT3B-Dependent Posttranslational N-Glycosylation While most of N-glycosylation occurs cotranslationally and is mediated by STT3A (Ruiz-Canada et al., 2009), a catalytic subunit of OST complex in mammalian cells (Helenius and Aebi, 2004), metazoa is equipped with the isoform STT3B in OST complex, which catalyzes posttranslational N-glycosylation. We examined which OST catalytic subunit was involved in the

(I–K) Stability of D18G TTR in HEK293 cells cotransfected with D18G TTR and the epitope-tagged dominant negative form of UBE2J1, UBE2G2, HRD1, and p97 or 50 nM Derlin siRNA was determined by CHX chase. D18G TTR was quantified and presented as the percentage of the amount detected at 0 hr. Data are means ± SEM from three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, versus mock-transfected cells; Student’s t test. See also Figures S1 and S7.

Molecular Cell 47, 99–110, July 13, 2012 ª2012 Elsevier Inc. 101

Molecular Cell Posttranslational N-Glycosylation of TTR Variants

A

B

C

D

E

F

G

H

N-glycosylation of the TTR using STT3 isoform-specific siRNA. Immunoblotting showed that N-glycosylated form of D18G TTR induced by p97QQ or C329S HRD1 was abolished by STT3B knockdown but not by STT3A siRNA (Figures 3A and 3B and S2A–S2D), suggesting that N-glycosylation of the TTR is posttranslational. To confirm that N-glycosylation indeed occurs after completion of synthesis, we performed pulse-chase experiments. Glycosylated D18G TTR was observed at chase 0 hr when pulse-labeling time was carried out for 30 min (Figure 3C). But it was not observed in a 5 min pulse-labeling period, and only faintly detectable by longer exposure (Figure S2E). Moreover, the ratio of glycosylated to nonglycosylated form increased upon longer pulse-labeling time (Figure S2F). In mock-transfected cells, the glycosylated form of D18G TTR accumulated during the chase period until 1 hr chase, and decreased quickly thereafter (Figure 3C, upper panel). In contrast, the glycosylated form continuously accumulated during the chase period in the p97QQ-transfected cells (Figure 3C, lower panel). We confirmed that neither nonglycosylated D18G TTR nor glycosylated D18G TTR was secreted from the cells regardless of the presence or absence of p97QQ (Figure S2G). These results suggest that the N-glycan is transferred to TTR posttranslationally and the glycosylated form undergoes ERAD. Glycosylated D18G TTR Is Degraded through EDEM3Mediated N-Glycan-Dependent Pathway Since oligosaccharides play an important role in quality control of secretory proteins, we speculate that the posttranslationally glycosylated D18G TTR is transferred to the machinery of N-glycan-dependent ERAD (GERAD) pathway for disposal. 102 Molecular Cell 47, 99–110, July 13, 2012 ª2012 Elsevier Inc.

Figure 2. HMW D18G TTR Is N-Glycosylated Form at Asn98 Residue (A and B) Endo H (‘‘H’’) and PNGase F (‘‘F’’) sensitivity of high MW D18G TTR induced by His-p97QQ (A) or C329S HRD1-Myc/His (B). N-glycosylated TTR is represented as TTR+CHO. (C and D) D18G TTR and His-p97QQ- (C) or C329S HRD1Myc/His- (D) cotransfected HEK293 cells were treated with 0.5 mg/ml tunicamycin (TM) for 12 hr. (E) The amino acid sequence of human TTR. N-glycosylation sequon in TTR sequence is indicated in italic and boldface type. (F) Comparison of the amino acid sequences from 90 to 110 residues of transthyretins from nine vertebrates. (G and H) HEK293 cells were cotransfected with D18G TTR or N98 mutants (D18G/N98D or D18G/N98Q TTR) and His-p97QQ (G) or C329S HRD1-Myc/His (H). All assays were determined by immunoblotting with the indicated antibodies. See also Figure S7.

Because GERAD requires mannose trimming (Helenius and Aebi, 2004; Liu et al., 1999), we examined the effect of kifunensine (KIF) and deoxymannojirimycin (DMJ), which are ER a-mannosidase inhibitor (Marcus and Perlmutter, 2000), on the steady-state expression of glycosylated D18G TTR in mock- or p97QQtransfected cells. In mock-transfected cells, KIF or DMJ treatment induced the accumulation of glycosylated D18G TTR (Figures 3D and S2H), suggesting that glycosylated D18G TTR is degraded through GERAD pathway. In p97QQ-transfected cells, low mobility glycosylated D18G TTR band was observed in KIF or DMJ treatments (Figures 3D and S2H), suggesting that the glycosylated form of D18G TTR in the presence of p97QQ may be subject to extensive mannose trimming. We confirmed that STT3B knockdown decreased the level of glycosylated D18G TTR accumulated by KIF or DMJ treatments (Figures S2I and S2J). Consistently, the pulse-chase experiments clearly showed that KIF treatment inhibited the degradation of glycosylated D18G TTR, but not that of the nonglycosylated form (Figures 3E–3G). To further confirm the involvement of GERAD components, we examined the effects of EDEM family (Hirao et al., 2006; Molinari et al., 2003; Oda et al., 2003) using EDEM isoform-specific siRNA. We found that knockdown of EDEM3, but not of EDEM1, increased the steady-state expression of glycosylated D18G TTR (Figures 3H, 3I, S3A, and S3B) and attenuated its degradation (Figures S3C–S3E). Moreover, endogenous EDEM3 was coimmunoprecipitated with D18G TTR but not with WT TTR (Figure S3F). Endogenous EDEM3 also interacted with D18G/N98D and D18G/N98Q TTRs (Figure S3G), implying that EDEM3 binds to D18G TTR in N-glycan-independent manner. These results suggest that posttranslationally glycosylated D18G TTR is eliminated by EDEM3-mediated GERAD pathway. Role of Posttranslational N-Glycosylation on ERQC of D18G TTR Posttranslational N-glycosylation could provide alternative ERAD (GERAD) pathway in addition to N-glycan-independent

Molecular Cell Posttranslational N-Glycosylation of TTR Variants

siSTT3B

siSTT3A

siCON

siSTT3A

siSTT3B

siCON

Figure 3. D18G TTR Is Subject to Posttranslational N-Glycosylation and EDEM3-Mediated GERAD

C329S HRD1

B

p97QQ

A

TTR

D18G TTR+CHO D18G TTR

TTR

D18G TTR+CHO TTR D18G TTR

His

His-p97QQ

His

C329S HRD1-Myc/His HR

Tubulin

Tubulin

Tubulin

Tubulin

D

Chase (h) h)

0 0.5 1

2

Mock

KIF (µg/mL) µ

Mock ck

D18GTTR+CHO * D18G TTR

His-p97QQ QQ

D18GTTR+CHO * D18G TTR

1

His-p97QQ 5

0

1

D18G TTR+CHO D18G TTR

His

His-p97QQ

Tubulin

Tubulin

D18G TTR+CHO

CON

D18G TTR D18G TTR+CHO

80 60 40 CON KIF

20 0

0

I

siEDEM3

siCON

H

siEDEM1

D18G TTR

100

TTR

D18G TTR+CHO D18G TTR

EDEM1 1

EDEM1

EDEM3 3

EDEM3

Tubulin

Tubulin

Relative expression of D18G TTR+CHO

KIF

G TTR

2.0

1 2 Chase (h)

3

TTR+CHO ***

1.5 1.0

TTR+CHO 17 175 15 150 12 125 10 100 75 7 5 50 2 25 0

*

* *

CON KIF 0

1 2 Chase (h)

3

0.5 0

siEDEM3 EDEM3

3

siCON

2

siEDEM1 EDEM1

0 0.5 1

5

TTR

F

Chase (h))

Relative radioactivity (%)

E

0

Relative radioactivity (%)

C

(A and B) STT3 siRNA (100 nM)-transfected HEK293 cells were cotransfected with D18G TTR and His-p97QQ (A) or C329S HRD1-Myc/His (B). Cell lysates were analyzed by immunoblotting. (C) D18G TTR- and His-p97QQ-cotransfected HEK293 cells were pulse labeled for 30 min and chased at the indicated time. Cell lysates were immunoprecipitated with anti-TTR antibody and analyzed by autoradiography. The asterisk indicates immature TTR that has an uncleaved bipartite signal sequence. (D) D18G TTR- and His-p97QQ-cotransfected HEK293 cells were treated with the indicated concentration of kifunensine (KIF) for 16 hr. Cell lysates were analyzed by immunoblotting. (E) Pulse chase of HEK293 cells stably expressing D18G TTR (D18G TTR-HEK293 cells) untreated or treated with 5 mg/ml kifunensine (KIF) for 1 hr before and during pulse chase. (F and G) Relative amount of D18G TTR (F) and D18G TTR+CHO (G) during the chase periods. Data are means ± SEM from three independent experiments. *p < 0.05; Student’s t test. (H) D18G TTR-HEK293 cells were transfected with 50 nM EDEM siRNA. Cell lysates were analyzed by immunoblotting. (I) Quantification of D18G TTR+CHO in immunoblots (H). Data are means ± SEM from three independent experiments. ***p < 0.001; ANOVA with Tukey-Kramer. See also Figures S2 and S3.

ERAD pathway to ensure the D18G TTR disposal. However, it is unlikely that the fraction of TTR undergoing STT3B-mediated N-glycosylation is degraded by the N-glycan-independent ERAD pathway because STT3B knockdown neither affected the degradation kinetics (Figures S4A and S4B) nor increased the steady-state level of nonglycosylated D18G TTR (Figures S4C and S4D). N-glycosylation is known to have several functions such as accelerating protein folding (Jitsuhara et al., 2002), stabilizing protein structure (Culyba et al., 2011), and reducing aggregation by increasing solubility (Petrescu et al., 2004). Because D18G TTR is highly prone to aggregate (Hammarstro¨m et al., 2003), N-glycosylation may increase the solubility to prevent aggregation. As expected, most N-glycosylated D18G TTR was recovered in detergent-soluble fraction (S) (Figure 4A). STT3B knockdown, which inhibited N-glycosylation, increased the nonglycosylated form in the insoluble fraction (P) (Figures 4A and 4B). Furthermore, the ratio of nonglycosylated TTR in insoluble fraction to that in soluble fraction was increased in siSTT3B dose-dependent manner (Figures S4E and S4F). In contrast, the solubility of D18G/N98Q TTR, which cannot be N-glycosylated, was not affected by siSTT3B (Figures S4G and S4H), implying that STT3B knockdown reduced the D18G

TTR solubility by inhibiting N-glycosylation. These results suggest that STT3Bmediated posttranslational N-glycosylation increases the solubility of aggregation-prone D18G TTR to prevent aggregation. To gain more insight into the physiological role of posttranslational N-glycosylation of D18G TTR by STT3B, we examined the cell proliferation of HEK293 cells stably expressing D18G TTR (D18G TTR-HEK293 cells) using STT3B siRNA. The cell proliferation of D18G TTR-HEK293 cells decreased relative to that of parental HEK293 cells likely due to the expression of aggregation-prone D18G TTR (Figure 4C; black diamonds versus green triangles). This decrease was explained in part by the downregulation of cyclin D1 (Figure 4D). Intriguingly, STT3B knockdown further decreased the cell proliferation of D18G TTR-HEK293 cells compared with control siRNA (Figure 4C; green triangles versus blue circles). To characterize STT3B knockdown-induced cell proliferation delay, we biochemically assessed the expression of cell-cycle regulators. In D18G TTR-HEK293 cells, STT3B knockdown upregulated the expression of p27 (Figure 4D), which is a Cip/Kip type CDK inhibitor known to suppress both cyclin D-CDK4/6 and cyclin E-CDK2 and is a critical negative regulator of G1/S phase progression (Polyak et al., 1994). Coinciding with the increase of p27, the expression level of cyclin E, but not of cyclin D1, was decreased by siSTT3B compared Molecular Cell 47, 99–110, July 13, 2012 ª2012 Elsevier Inc. 103

Molecular Cell

A

B siRNA

CON STT3A STT3B S P S P S P D18G TTR+CHO D18G TTR

TTR

C

2.0

Parental HEK293: siCON Parental HEK293: siSTT3B D18G TTR-HEK293: siCON D18G TTR-HEK293: siSTT3B

1.8 1.7

1.0 0.5 0

Parental D18G TTR HEK293 -HEK293

siSTT3B

1.6 1.5

-

+

-

+

p27

p27

Cyclin E

Cyclin E

Cyclin D1

Cyclin D1

Tubulin

Tubulin

1.4 1.3 ** ** *

1.2

* ** * * * *

** *

1.1 * * * * * ** *

1 0.9 0

4

8

**

**

*

12 16 20 24 28 32 36 40

Relative quantity of Skp2 mRNA

Time (h)

E

Figure 4. Inhibition of Posttranslational N-Glycosylation Decreases Solubility of Aggregation-Prone TTR and Delays Cell Proliferation

**

1.5

D 1.9

Relative cell number

Relative ratio of insoluble TTR to soluble TTR (%)

Posttranslational N-Glycosylation of TTR Variants

1.6

*

1.4 1.2 1.0 0.8 0.6 0.4 0.2

0 siSTT3B

-

+

-

(A) D18G TTR in 1% Triton X-100-soluble (S) and -insoluble (P) fractions was recovered from D18G TTR-HEK293 cells transfected with 100 nM STT3 siRNA and determined by immunoblotting. (B) Quantification of D18G TTR (nonglycosylated form) of each fraction in immunoblots (A). Bars represent the relative ratio of insoluble TTR to soluble TTR. Data are means ± SEM from three independent experiments. **p < 0.01; ANOVA with Tukey-Kramer. (C) Cell proliferation of D18G TTR-HEK293 or parental HEK293 cells that were transfected with siCON or siSTT3B. Relative cell number, derived from quantitative analyses by IncuCyte live-cell imaging system, is plotted against time. Data are means ± SEM from three independent experiments. *p < 0.05, yp < 0.001, siCON-transfected parental HEK293 cells versus siCONtransfected D18G TTR-HEK293 cells; siCONtransfected D18G TTR-HEK293 cells versus siSTT3B-transfected D18G TTR-HEK293 cells; Student’s t test. (D) Expression of cell-cycle regulators in D18G TTR-HEK293 or parental HEK293 cells that were transfected with siCON or siSTT3B was determined by immunoblotting. (E) Expression of Skp2 gene in D18G TTR-HEK293 or parental HEK293 cells that were transfected with siCON or siSTT3B was assessed by real-time quantitative PCR. Skp2 mRNA level was normalized to 18S ribosomal RNA (internal control). Data are means ± SEM from three independent experiments. *p < 0.05; Student’s t test. See also Figure S4.

+

Parental D18G TTR HEK293 -HEK293

with control siRNA (Figure 4D). Moreover, STT3B knockdown in D18G TTR-HEK293 cells decreased the messenger RNA (mRNA) level of Skp2, a SCF-type E3 ligase for p27 (Kitagawa et al., 2009) (Figure 4E), suggesting that p27 is stabilized by STT3B knockdown through downregulation of Skp2. These results support the physiological relevance of posttranslational N-glycosylation in maintaining cellular homeostasis. Posttranslational N-Glycosylation-Mediated QC Mechanism Probably Functions as an Auxiliary System Next, we focused on the link between N-glycan-dependent and -independent ERAD pathways to understand the significance of their roles in the disposal of D18G TTR. Considering that inhibition of downstream ERAD machinery increases the N-glycosylated form, D18G TTR that is accumulated in the ER may be posttranslationally glycosylated and be transferred to GERAD pathway to rescue the disposal process. To prove this hypothesis, we tested whether selective inhibition of N-glycanindependent ERAD per se could stimulate posttranslational N-glycosylation. However, since perturbation of ERAD components (HRD1, p97, Derlin-3) stabilized both glycosylated and non104 Molecular Cell 47, 99–110, July 13, 2012 ª2012 Elsevier Inc.

glycosylated D18G TTR (Figures S1C and S1D), the rescue effect by STT3B-mediated N-glycosylation could not be ascertained with these components. Therefore, we focused on Herp, which is involved in ERAD of nonglycosylated BiP substrates because D18G TTR is a BiP substrate (Okuda-Shimizu and Hendershot, 2007; Susuki et al., 2009). When degradation of both nonglycosylated and glycosylated D18G TTR was inhibited by p97QQ, nonglycosylated D18G TTR was found in Herp complex (Figure 5A). As expected, Herp knockdown attenuated the degradation of nonglycosylated form (Figures 5B and S5A) and increased its steady-state level (Figures 5D, 5E, and S5B). Importantly, it also increased the steady-state level of the glycosylated form (Figures 5D, 5F, and S5A). Although the apparent rate of glycosylated D18G TTR GERAD during CHX chase was also attenuated at a later time point by Herp knockdown (Figures 5C and S5A), it is unlikely that siHerp inhibited the GERAD because KIF treatment accumulated the glycosylated D18G TTR at relatively equal level in both siHerp- and siCON-transfected cells (Figure 5G). Thus, Herp knockdown likely induced the accumulation of the nonglycosylated D18G TTR in the ER, resulting in increased production of the N-glycosylated form and degradation via

Molecular Cell Posttranslational N-Glycosylation of TTR Variants

+ + +

+ +

+ + -

+ + +

+ +

TTR

TTR+CHO TTR

Flag

Herp-Flag His-p97QQ

Tubulin

Tubulin

-

+

-

+

TTR

D18G TTR+CHO D18G TTR

Herp

Herp

Tubulin

Tubulin TTR+CHO

G

4

H

3

siSTT3B siHerp

2

0 KIF: - + - + siCON siHerp

Relative expression of TTR

Soluble fraction TTR ** 1.6 * 1.2 0.8 0.4

0 siSTT3B siHerp

*

60

***

***

40 20 0

0

1.5 3 4.5 Chase (h)

F

TTR * *

2.0 1.5 1.0 0.5

0 - + KIF: - + siCON siHerp

TTR+CHO 100 80

**

60 40 siCON siHerp

20 0 0

1.5 3 Chase (h)

4.5

TTR+CHO

-

+ -

+

+ +

I

Soluble fraction -

+ -

+

+ +

*

8 ***

6 4 2

0 - + KIF: - + siCON siHerp

TTR

D18G TTR+CHO D18G TTR

Herp

Herp

Tubulin

Tubulin

K 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0 siSTT3B siHerp

Insoluble fraction

siSTT3B siHerp

1

J

80

siCON siHerp

n.s.

Relative expression of TTR

Relative expression of TTR

Normalize with KIF (-) of each siRNA

5

100

Soluble fraction TTR+CHO *

*

-

+ -

+

+ +

-

+ -

+

+ + D18G TTR

TTR

L

Insoluble fraction TTR ***

Relative expression of TTR

KIF

TTR

E

siCON siHerp

Relative expression of TTR

D

His

C Remaining D18G TTR (%)

Herp-Flag + p97QQ + D18G TTR -

B

Input (1/20)

Relative expression of TTR

IP: Flag

Remaining D18G TTR (%)

A

3.5 3.0 2.5 2.0 1.5 1.0 0.5 0 siSTT3B siHerp

**

***

-

+ -

+

+ +

Figure 5. Posttranslational N-Glycosylation Regulates the Efficient Removal of D18G TTR (A) Plasmids-transfected HEK293 cells were immunoprecipitated with anti-Flag antibody and were analyzed by immunoblotting. (B and C) Stabilities of D18G TTR and D18G TTR+CHO in the presence of siHerp were determined by CHX chase analysis. D18G TTR (B) and D18G TTR+CHO (C) were quantified and presented as the percentage of the amount detected at 0 hr. (D) D18G TTR-HEK293 cells transfected with 50 nM Herp siRNA were treated with 5 mg/ml KIF for 16 hr. Cell lysates were analyzed by immunoblotting. (E–G) Quantification of D18G TTR (E) and D18G TTR+CHO (F and G) in immunoblots shown in (D). Bars represent the relative expression of TTR or TTR+CHO normalized to siCON/KIF(–) (E and F) or to KIF(–) in siCON or siHerp (G). (H and I) D18G TTR in 1% Triton X-100-soluble and -insoluble fractions was recovered from D18G TTR-HEK293 cells transfected with 10 nM siHerp and 100 nM siSTT3B or each siRNA alone and determined by immunoblotting. (J–L) Quantification of soluble D18G TTR (J) and soluble D18G TTR+CHO (K) in immunoblots shown in (H) and insoluble D18G TTR (L) in immunoblots shown in (I). Bars represent the relative expression of TTR or TTR+CHO normalized to siHerp(–)/siSTT3B(–). All data are means ± SEM from three independent experiments. n.s., not significant; *p < 0.05, **p < 0.01, ***p < 0.001; Student’s t test. See also Figure S5.

GERAD pathway to provide an auxiliary pathway for D18G TTR disposal. To investigate this issue, we examined the effect of Herp and STT3B double knockdown on the steady-state expression of nonglycosylated and glycosylated D18G TTR in deter-

gent-soluble and -insoluble fraction. As expected, Herp and STT3B double knockdown decreased glycosylated D18G TTR in detergent-soluble fraction and slightly increased nonglycosylated D18G TTR compared with Herp single knockdown (Figures Molecular Cell 47, 99–110, July 13, 2012 ª2012 Elsevier Inc. 105

Molecular Cell Posttranslational N-Glycosylation of TTR Variants

A

p97QQ

B

Cell lysate WT V30M V30M M-TTR M-TTR TTR + + +

rotate TTR+CHO TTR

TTR His

His-p97QQ

Actin

Actin

p97QQ

Medium WT V30M V30M M-TTR M-TTR TTR + + +

TTR

rotate

TTR

C

Cell lysate

D

WT M-TTR Mock p97QQ Azc (mM) 0 5 10 0 5 10 TTR

Tyr105 FG loop 3.8 Å

Asn98

TTR+CHO TTR

His

His-p97QQ

Tubulin

Tubulin

Medium WT M-TTR Mock p97QQ Azc (mM) 0 5 10 0 5 10 TTR

TTR

Figure 6. Unfolding that Exposes the Cryptic N-Glycosylation Site Triggers Posttranslational N-Glycosylation of TTR (A) HEK293 cells were cotransfected with WT monomeric mutant (F87M/L110M) TTR (M-TTR), V30M M-TTR (V30M/F87M/L110M), or V30M TTR and His-p97QQ. Samples were analyzed by immunoblotting. (B) The structures of WT TTR monomer are shown in cartoon style (upper panel) or surface representation (lower panel). The figure was rendered in PyMOL with the WT TTR structure (PDB 1BMZ). The blue circle indicates the side chain nitrogen of Asn98 residue. Nitrogen, oxygen, and carbon are colored blue, red and green, respectively. (C) Close-up view of the region around b strand F and G in monomer A. (D) WT M-TTR and His-p97QQ-cotransfected HEK293 cells were treated with the indicated concentration of L-Azetidine-2-carboxylic acid (AZC) for 16 hr. Samples were analyzed by immunoblotting. See also Figure S6.

5H, 5J, and 5K). Moreover, the nonglycosylated form in detergent-insoluble fraction was significantly increased by Herp and STT3B double knockdown relative to Herp knockdown or control (Figures 5I and 5L). These results suggest that nonglycosylated D18G TTR is disposed by Herp, and its failure increases the chance of D18G TTR glycosylation by STT3B so that its disposal is handled by the GERAD pathway. Excessive Unfolding Externalizes a Cryptic N-Glycosylation Site for the Posttranslational N-Glycosylation Given that posttranslational N-glycosylation is the molecular switch to ensure the disposal of misfolded TTR, other TTR mutants might undergo N-glycosylation. Thus, we examined several TTR variants that show distinct fates in the cells (Sato et al., 2007; Sekijima et al., 2005; Susuki et al., 2009). WT monomeric TTR (M-TTR) that has artificial monomeric mutations (F87M/L110M) is secreted from the cells. V30M TTR, a common variant observed in FAP patients, forms tetramer, and is secreted 106 Molecular Cell 47, 99–110, July 13, 2012 ª2012 Elsevier Inc.

from the cells. In contrast, V30M M-TTR, having monomeric mutations in V30M TTR, is retained in the ER and subjected to ERAD, which is similar to D18G TTR. Immunoblotting showed that the N-glycosylated form in p97QQ- or C329S HRD1-transfected cells was only observed in V30M M-TTR (Figures 6A and S6A), suggesting that ERAD substrate is a target of posttranslational N-glycosylation. Next, we examined whether substrate accumulation in the ER per se stimulated posttranslational N-glycosylation. When ER retention of WT TTR was forcibly induced by brefeldin A (BFA) in combination with p97QQ, glycosylated WT TTR was not observed (Figure S6B), suggesting that ER retention itself is not the cause of posttranslational N-glycosylation. Structural data of WT TTR monomer shows that N98 residue is faced toward the interior of TTR and that the side-chain nitrogen of N98 residue forms Van der Waals interaction with the side chain oxygen of Y105 residue (Figures 6B and 6C). Thus, TTR unfolding might expose N98 residue on the molecular surface for posttranslational N-glycosylation. To examine this possibility,

Molecular Cell Posttranslational N-Glycosylation of TTR Variants

A

B

C

D

we used L-Azetidine-2-carboxy acid (Azc), a proline analog, to forcibly induce TTR to misfold (Cameron et al., 2009). Azc induced the N-glycosylation of WT M-TTR, V30M, and L55P TTRs but not WT TTR (Figures 6D and S6C–S6F), strongly indicating that unfolding indeed induces posttranslational N-glycosylation. Because WT TTR is more energetically stable than the TTR variants used in this study (Sekijima et al., 2005), the effect of Azc was probably insufficient to induce WT TTR unfolding. Although D18G TTR is a highly destabilized variant, the ratio of glycosylated form to nonglycosylated D18G TTR was relatively low in the presence of p97QQ, suggesting that most D18G TTR exists as a modestly unfolded form, and only a small fraction is severely unfolded and exposing a cryptic N-glycosylation site. To test this hypothesis, HEK293 cells cotransfected with D18G

Figure 7. Occurrence of Posttranslational N-Glycosylation Correlates with Improper Maturation of TTR (A and B) HEK293 cells were transiently transfected with the indicated TTR variants. Twentyfour hours posttransfection, cells were treated with 5 mg/ml KIF for 24 hr. Cell lysates and media were analyzed by immunoblotting. TTR and TTR+CHO were quantified by densitometry of immunoblots. Relative ratio of TTR+CHO and relative secretion efficiency were calculated as follows: Relative ratio of TTR+CHO = the indicated TTR+CHO/D18G TTR+CHO; D18G TTR+CHO = 1.0. Relative secretion efficiency (%) = (the indicated secreted TTR/ [the indicated intracellular TTR + the indicated secreted TTR])/(secreted WT TTR/[intracellular WT TTR + secreted WT TTR]); WT TTR = 100%. Data are means ± SEM from three independent experiments. (C) Relationship between N-glycosylation ratio and secretion efficiency. Relative ratio of TTR+CHO and relative secretion efficiency of 18 TTR variants were calculated and plotted. The spearman rank correlation between N-glycosylation ratio and secretion efficiency was rs = 0.886, p < 0.01. (D) Model for ERQC system of mutant TTR by posttranslational N-glycosylation. TTRs are synthesized through translocon and properly folded in the ER. Folded monomer forms tetramer and tetramer is exported from the ER to the Golgi (1). Mutant TTR that cannot form tetramer due to low energetic stability are unfolded and retained in the ER (2). Modest unfolded TTR (nonglycosylated TTR) is degraded by the proteasome through ERAD complex consisting of UBE2J1 or UBE2G2, HRD1, Herp, Derlin-3, and p97 (3). Prolonged ER residence induces severe unfolding of TTR that exposes cryptic N-glycosylation site (sequon: NDS) (4), and severely unfolded form is subject to posttranslational N-glycosylation by the OST complex with STT3B (5). Owing to N-glycosylation, solubility of severely unfolded TTR can be temporarily increased (6). N-glycosylated TTR is subject to demannosylation, followed by degradation through ERAD complex, that is distinct from the ERAD complex for nonglycosylated TTR, consisting of UBE2J1 or UBE2G2, HRD1, Derlin-3, and p97. EDEM3 is likely required for the ERAD of N-glycosylated TTR (7).

TTR and p97QQ were treated with Azc to severely misfold D18G TTR. Azc treatment increased the ratio of glycosylated form to total D18G TTR (Figures S6G and S6H), suggesting that severely unfolded form is targeted for posttranslational N-glycosylation. We further examined whether N-glycosylation susceptibility of TTR correlated with the secretion efficiency in 18 different TTR variants because the relative secretion efficiency of TTR correlates with its combined stability score that is defined by the combination of thermodynamic and kinetic stability of TTR in vitro (Sekijima et al., 2005). While the TTR variants except for D18G TTR were mostly secreted and less susceptible to N-glycosylation, M-TTR variants as well as D18G TTR were barely secreted and highly susceptible to N-glycosylation (Figures 7A and 7B). When correlation analysis Molecular Cell 47, 99–110, July 13, 2012 ª2012 Elsevier Inc. 107

Molecular Cell Posttranslational N-Glycosylation of TTR Variants

was performed by Spearman rank correlation, N-glycosylation susceptibility inversely correlated with the secretion efficiency (rs = 0.886, p < 0.01; Figure 7C), supporting the notion that conformational instability triggers posttranslational N-glycosylation. Taken together, these data indicate that unfolding-induced externalization of a cryptic N-glycosylation site of TTR stimulates posttranslational N-glycosylation. DISCUSSION Our study shows that a cryptic N-glycosylation sequon functions as a recognition signal for unfolding and that STT3B-mediated posttranslational N-glycosylation plays a crucial role in the ERQC of a nonglycosylated protein. Based on our results, the following model is proposed for the ERQC of nonglycosylated protein TTR (Figure 7D). Modestly unfolded TTR is sensed by Hsp70 family chaperone BiP that recognizes the exposed hydrophobic patch and is degraded by Herp-mediated N-glycanindependent ERAD pathway. In contrast, excessively unfolded TTR, which fails to be eliminated by N-glycan-independent ERAD pathway, undergoes STT3B-mediated posttranslational N-glycosylation at the cryptic glycosylation site, which is normally masked in more native conformation. In addition to the severity of the unfolding state, STT3B-mediated posttranslational N-glycosylation is likely enhanced by the accumulation of nonglycosylated ERAD substrates in the ER. N-glycosylation increases the solubility of aggregation-prone protein and directs it to the EDEM3-mediated GERAD pathway. The posttranslational N-glycosylation may also facilitate the engagement of lectin-like chaperones (e.g., calnexin and calreticulin) that further improves the ERQC surveillance system (Helenius and Aebi, 2004). We speculate that the ERAD and the GERAD pathways have distinct contribution to the degradation of misfolded TTR, with the former functioning as major disposal system and the latter functioning as minor or auxiliary system. Given that the steady-state expression of D18G/N98Q TTR in whole-cell lysate was higher than that of D18G TTR (Figure S7A) while the degradation kinetics of nonglycosylated form was almost identical (Figure S7B), the GERAD pathway may have a crucial role in misfolded TTR disposal under normal physiological condition. We think that the contribution of GERAD pathway to D18G TTR degradation may vary depending on the condition in the ER such as a crowded environment where exposure of cryptic N-glycosylation site could be enhanced by severe unfolding. Together, the coordinated ERAD and GERAD ensure efficient clearance of misfolded TTR to maintain cellular proteostasis. It is generally thought that hydrophobic patches or unpaired sulfhydryls serve as signature of unfolding (Hegde and Ploegh, 2010). Based on the current results, we propose that the externally exposed cryptic N-glycosylation site(s) of nonglycosylated protein is an alternate recognition motif of unfolding for ERQC. Considering that the N-glycosylation of TTR was enhanced by inhibition of the retrotranslocation machineries, but not of proteasome (Figure S7C), residence time of unfolded TTR in the ER lumen must be a critical factor for posttranslational N-glycosylation. Although we did not observe accumulation of glycosylated D18G TTR (HMW band) by overexpression of dominant negative UBE2J1 or UBE2G2 (Figures 1A and 1B), which 108 Molecular Cell 47, 99–110, July 13, 2012 ª2012 Elsevier Inc.

would inhibit ubiquitination and retrotranslocation, coexpression of both E2 mutants induced the accumulation of glycosylated TTR (Figure S7D), suggesting that UBE2J1 and UBE2G2 may have redundant activities in the removal of TTR from the ER. Prolonged ER retention of D18G TTR should increase the chance of open conformation in which a side chain of N98 is externalized, which is trapped by OST complex with STT3B. We speculate that the sequon externalization is a stochastic process, and the on-time reaction rate of STT3B is slow. As the frequency of N98 externalization is increased, the molecule should have more chance to be recognized by molecular chaperones or cochaperones and allow enough time for STT3B to catalyze the reaction. Posttranslational N-glycosylation by STT3B may act as a kind of biological timer, similar to ER mannosidase I (Helenius and Aebi, 2004; Liu et al., 1999), for determining the fate of cargo proteins in the ER. These coordinated systems for recognition of unfolded state would be important to distinguish the nonproductive molecules from properly folding nascent proteins. STT3B-mediated posttranslational N-glycosylation pathway is probably stimulated to increase the solubility of aggregationprone TTR. This cellular strategy is somewhat analogous to the O-mannosylation system for misfolded proteins in yeast (Nakatsukasa et al., 2004). In S. cerevisiae, O-mannosylation is expected to function as a fail-safe mechanism for the ERAD by solubilizing misfolded proteins that overflowed from the ERAD pathway. In mammals, although O-mannosylation occurs in a limited number of glycoproteins in brain, nerve, and skeletal muscle, the fail-safe function of O-mannosylation in the ERAD pathway has not been reported. Perhaps posttranslational Nglycosylation instead of O-mannosylation may have a role in protecting the ER against aggregate of misfolded proteins in mammals. Consistent with this hypothesis, although a relatively small fraction of D18G TTR is targeted for posttranslational N-glycosylation in normal condition, inhibition of posttranslational N-glycosylation by STT3B knockdown induces cell proliferation delay. Considering that N-glycosylation substrate is a severely misfolded form of D18G TTR, this TTR species may form aggregates in the STT3B knockdown condition. Although the mechanism underlying cell proliferation delay induced by STT3B knockdown still remains to be elucidated, Chen et al. recently reported that prolonged ER stress triggers cell-cycle delay during the G1 phase through the regulation of Skp2-p27 axis, that is, decrease in Skp2 leading to stabilization of p27 (Chen et al., 2011). This system facilitates efficient clearance of misfolded proteins through ERAD pathway to restore homeostasis. Because STT3B knockdown did not show apparent cell death of D18G TTR-HEK293 cells in this experimental time course (data not shown), the cell proliferation delay observed in this study may contribute to restore homeostasis as Chen et al. showed. As far as we know, the only known substrate for posttranslational N-glycosylation is human blood coagulation factor VII (FVII). FVII is originally a glycoprotein and has two sequons that are modified cotranslationally at N183 or posttranslationally at N360 (Bolt et al., 2005). Since cotranslationally glycosylated FVII remains unfolded for more than 30 min after the protein enters the ER lumen, posttranslational glycosylation of N360 in

Molecular Cell Posttranslational N-Glycosylation of TTR Variants

FVII is permitted, followed by folding into a secretion-competent conformation (Ruiz-Canada et al., 2009). Thus, posttranslational N-glycosylation of FVII is required for acquisition of native protein structure. At present, it remains unclear whether other ERAD substrates utilize cryptic N-glycosylation site(s) for surveillance of terminal misfolding. We think that posttranslational N-glycosylation-mediated QC system may be applied to not only nonglycosylated proteins but also to some glycoproteins, which have cryptic N-glycosylation sites because substrate specificity of STT3B is less selective (Ruiz-Canada et al., 2009). Since the rate of sequon occupancy is only 60% (Petrescu et al., 2004), it is likely that this posttranslational N-glycosylationmediated QC system may function to rescue the secretory system. Genome-wide, systematic studies would be needed to clarify the detail of this pathway and to identify other substrates. EXPERIMENTAL PROCEDURES Plasmid DNA and siRNA Transfections Transient transfections of plasmid DNA were performed with TransIT-LT-1 (Mirus), as described previously (Sato et al., 2007). Most of the experiments were performed after 48 hr of transfection. Fifty nanomolar (except for siSTT3A and siSTT3B) and 100 nM siRNA (siSTT3A and siSTT3B) Stealth siRNA (Invitrogen) listed in Table S1 were transiently transfected into HEK293 cells or HEK293 stably expressing D18G TTR using Lipofectamine RNAiMAX (Invitrogen) according to the manufacturer’s instructions. Stealth RNAi Low GC Duplex was used as negative control siRNA. Cells were typically assayed 48 or 72 hr after transfection. In the case of cotransfection of siRNA and complementary DNA plasmids, plasmid transfection was performed after 24 hr of siRNA transfection, and then cells were cultured for further 48 hr. The knockdown efficiency was confirmed by immunoblotting or quantitative real-time RT-PCR. Cell Lysis, Cycloheximide Chase, and Densitometric Quantification Culture media were centrifuged to remove the debris and prepared as medium samples. Cells were washed twice with ice-cold PBS and lysed in 1% Triton Buffer (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, and 1% Triton X-100, supplemented with 1% protease inhibitor cocktail [Sigma]). The lysates were cleared by centrifugation at 15,000 g for 20 min. For the recovery of 1% Triton bufferinsoluble fraction, cell pellets were isolated and dissolved in 1% SDS buffer (10 mM Tris-HCl [pH 6.8] and 1% SDS, supplemented with 1% protease inhibitor cocktail), followed by sonication at 4 C. The supernatants were prepared as detergent-insoluble fraction. For the recovery of whole-cell lysate, cells were lysed on ice for 30 min with immunoprecipitation lysis buffer containing 1% protease inhibitor cocktail, as described previously (Suico et al., 2004). For the CHX chase analysis, cells were treated with 200 mM CHX for the time periods indicated and lysed and analyzed by western blotting. The density of the bands was quantified with Image Gauge software (version 4.23; Fujifilm). SUPPLEMENTAL INFORMATION Supplemental Information includes Supplemental Experimental Procedures, seven figures, and one table and can be found with this article online at doi:10.1016/j.molcel.2012.04.015. ACKNOWLEDGMENTS We thank T. Sommer for UBE2J1 plasmids, A.M. Weissman for gp78 plasmids, and Y. Ye for p97 plasmids. We also thank M. Mizuguchi, S. Ikemizu, and the members of Dr. Wada’s laboratory at Fukushima Medical University for helpful discussions. This work was supported by Grants-in-Aid for Scientific Research from the Ministry of Education, Sciences, Sports, and Culture (MEXT) of Japan, and grants from the Global COE Program (Cell Fate Regulation Research and Education Unit).

Received: August 24, 2011 Revised: January 3, 2012 Accepted: April 12, 2012 Published online: May 17, 2012 REFERENCES Balch, W.E., Morimoto, R.I., Dillin, A., and Kelly, J.W. (2008). Adapting proteostasis for disease intervention. Science 319, 916–919. Bernasconi, R., Galli, C., Calanca, V., Nakajima, T., and Molinari, M. (2010). Stringent requirement for HRD1, SEL1L, and OS-9/XTP3-B for disposal of ERAD-LS substrates. J. Cell Biol. 188, 223–235. Bolt, G., Kristensen, C., and Steenstrup, T.D. (2005). Posttranslational N-glycosylation takes place during the normal processing of human coagulation factor VII. Glycobiology 15, 541–547. Cameron, P.H., Chevet, E., Pluquet, O., Thomas, D.Y., and Bergeron, J.J. (2009). Calnexin phosphorylation attenuates the release of partially misfolded alpha1-antitrypsin to the secretory pathway. J. Biol. Chem. 284, 34570–34579. Chen, B., Mariano, J., Tsai, Y.C., Chan, A.H., Cohen, M., and Weissman, A.M. (2006). The activity of a human endoplasmic reticulum-associated degradation E3, gp78, requires its Cue domain, RING finger, and an E2-binding site. Proc. Natl. Acad. Sci. USA 103, 341–346. Chen, M., Gutierrez, G.J., and Ronai, Z.A. (2011). Ubiquitin-recognition protein Ufd1 couples the endoplasmic reticulum (ER) stress response to cell cycle control. Proc. Natl. Acad. Sci. USA 108, 9119–9124. Christianson, J.C., Shaler, T.A., Tyler, R.E., and Kopito, R.R. (2008). OS-9 and GRP94 deliver mutant alpha1-antitrypsin to the Hrd1-SEL1L ubiquitin ligase complex for ERAD. Nat. Cell Biol. 10, 272–282. Culyba, E.K., Price, J.L., Hanson, S.R., Dhar, A., Wong, C.H., Gruebele, M., Powers, E.T., and Kelly, J.W. (2011). Protein native-state stabilization by placing aromatic side chains in N-glycosylated reverse turns. Science 331, 571–575. Ellgaard, L., and Helenius, A. (2003). Quality control in the endoplasmic reticulum. Nat. Rev. Mol. Cell Biol. 4, 181–191. Hamilton, J.A., and Benson, M.D. (2001). Transthyretin: a review from a structural perspective. Cell. Mol. Life Sci. 58, 1491–1521. Hammarstro¨m, P., Sekijima, Y., White, J.T., Wiseman, R.L., Lim, A., Costello, C.E., Altland, K., Garzuly, F., Budka, H., and Kelly, J.W. (2003). D18G transthyretin is monomeric, aggregation prone, and not detectable in plasma and cerebrospinal fluid: a prescription for central nervous system amyloidosis? Biochemistry 42, 6656–6663. Hegde, R.S., and Ploegh, H.L. (2010). Quality and quantity control at the endoplasmic reticulum. Curr. Opin. Cell Biol. 22, 437–446. Helenius, A., and Aebi, M. (2004). Roles of N-linked glycans in the endoplasmic reticulum. Annu. Rev. Biochem. 73, 1019–1049. Hirao, K., Natsuka, Y., Tamura, T., Wada, I., Morito, D., Natsuka, S., Romero, P., Sleno, B., Tremblay, L.O., Herscovics, A., et al. (2006). EDEM3, a soluble EDEM homolog, enhances glycoprotein endoplasmic reticulum-associated degradation and mannose trimming. J. Biol. Chem. 281, 9650–9658. Jitsuhara, Y., Toyoda, T., Itai, T., and Yamaguchi, H. (2002). Chaperone-like functions of high-mannose type and complex-type N-glycans and their molecular basis. J. Biochem. 132, 803–811. Kaneko, M., Ishiguro, M., Niinuma, Y., Uesugi, M., and Nomura, Y. (2002). Human HRD1 protects against ER stress-induced apoptosis through ERassociated degradation. FEBS Lett. 532, 147–152. Kikkert, M., Doolman, R., Dai, M., Avner, R., Hassink, G., van Voorden, S., Thanedar, S., Roitelman, J., Chau, V., and Wiertz, E. (2004). Human HRD1 is an E3 ubiquitin ligase involved in degradation of proteins from the endoplasmic reticulum. J. Biol. Chem. 279, 3525–3534. Kitagawa, K., Kotake, Y., and Kitagawa, M. (2009). Ubiquitin-mediated control of oncogene and tumor suppressor gene products. Cancer Sci. 100, 1374– 1381.

Molecular Cell 47, 99–110, July 13, 2012 ª2012 Elsevier Inc. 109

Molecular Cell Posttranslational N-Glycosylation of TTR Variants

Lenk, U., Yu, H., Walter, J., Gelman, M.S., Hartmann, E., Kopito, R.R., and Sommer, T. (2002). A role for mammalian Ubc6 homologues in ER-associated protein degradation. J. Cell Sci. 115, 3007–3014. Lilley, B.N., and Ploegh, H.L. (2004). A membrane protein required for dislocation of misfolded proteins from the ER. Nature 429, 834–840. Liu, Y., Choudhury, P., Cabral, C.M., and Sifers, R.N. (1999). Oligosaccharide modification in the early secretory pathway directs the selection of a misfolded glycoprotein for degradation by the proteasome. J. Biol. Chem. 274, 5861– 5867. Marcus, N.Y., and Perlmutter, D.H. (2000). Glucosidase and mannosidase inhibitors mediate increased secretion of mutant alpha1 antitrypsin Z. J. Biol. Chem. 275, 1987–1992. 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, 1397–1400. Nakatsukasa, K., and Brodsky, J.L. (2008). The recognition and retrotranslocation of misfolded proteins from the endoplasmic reticulum. Traffic 9, 861–870. Nakatsukasa, K., Okada, S., Umebayashi, K., Fukuda, R., Nishikawa, S., and Endo, T. (2004). Roles of O-mannosylation of aberrant proteins in reduction of the load for endoplasmic reticulum chaperones in yeast. J. Biol. Chem. 279, 49762–49772. Oda, Y., Hosokawa, N., Wada, I., and Nagata, K. (2003). EDEM as an acceptor of terminally misfolded glycoproteins released from calnexin. Science 299, 1394–1397. Oda, Y., Okada, T., Yoshida, H., Kaufman, R.J., Nagata, K., and Mori, K. (2006). Derlin-2 and Derlin-3 are regulated by the mammalian unfolded protein response and are required for ER-associated degradation. J. Cell Biol. 172, 383–393. Okuda-Shimizu, Y., and Hendershot, L.M. (2007). Characterization of an ERAD pathway for nonglycosylated BiP substrates, which require Herp. Mol. Cell 28, 544–554. Petrescu, A.J., Milac, A.L., Petrescu, S.M., Dwek, R.A., and Wormald, M.R. (2004). Statistical analysis of the protein environment of N-glycosylation sites: implications for occupancy, structure, and folding. Glycobiology 14, 103–114. Polyak, K., Lee, M.H., Erdjument-Bromage, H., Koff, A., Roberts, J.M., Tempst, P., and Massague´, J. (1994). Cloning of p27Kip1, a cyclin-dependent kinase inhibitor and a potential mediator of extracellular antimitogenic signals. Cell 78, 59–66. Ruiz-Canada, C., Kelleher, D.J., and Gilmore, R. (2009). Cotranslational and posttranslational N-glycosylation of polypeptides by distinct mammalian OST isoforms. Cell 136, 272–283.

110 Molecular Cell 47, 99–110, July 13, 2012 ª2012 Elsevier Inc.

Sato, T., Susuki, S., Suico, M.A., Miyata, M., Ando, Y., Mizuguchi, M., Takeuchi, M., Dobashi, M., Shuto, T., and Kai, H. (2007). Endoplasmic reticulum quality control regulates the fate of transthyretin variants in the cell. EMBO J. 26, 2501–2512. Sekijima, Y., Wiseman, R.L., Matteson, J., Hammarstro¨m, P., Miller, S.R., Sawkar, A.R., Balch, W.E., and Kelly, J.W. (2005). The biological and chemical basis for tissue-selective amyloid disease. Cell 121, 73–85. Shen, Y., Ballar, P., Apostolou, A., Doong, H., and Fang, S. (2007). ER stress differentially regulates the stabilities of ERAD ubiquitin ligases and their substrates. Biochem. Biophys. Res. Commun. 352, 919–924. Sousa, M.M., and Saraiva, M.J. (2003). Neurodegeneration in familial amyloid polyneuropathy: from pathology to molecular signaling. Prog. Neurobiol. 71, 385–400. Suico, M.A., Yoshida, H., Seki, Y., Uchikawa, T., Lu, Z., Shuto, T., Matsuzaki, K., Nakao, M., Li, J.D., and Kai, H. (2004). Myeloid Elf-1-like factor, an ETS transcription factor, up-regulates lysozyme transcription in epithelial cells through interaction with promyelocytic leukemia protein. J. Biol. Chem. 279, 19091–19098. Susuki, S., Sato, T., Miyata, M., Momohara, M., Suico, M.A., Shuto, T., Ando, Y., and Kai, H. (2009). The Endoplasmic Reticulum-associated Degradation of Transthyretin Variants Is Negatively Regulated by BiP in Mammalian Cells. J. Biol. Chem. 284, 8312–8321. Tiwari, S., and Weissman, A.M. (2001). Endoplasmic reticulum (ER)associated degradation of T cell receptor subunits. Involvement of ERassociated ubiquitin-conjugating enzymes (E2s). J. Biol. Chem. 276, 16193– 16200. Vembar, S.S., and Brodsky, J.L. (2008). One step at a time: endoplasmic reticulum-associated degradation. Nat. Rev. Mol. Cell Biol. 9, 944–957. 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., Yun, C., Ron, D., and Rapoport, T.A. (2004). A membrane protein complex mediates retro-translocation from the ER lumen into the cytosol. Nature 429, 841–847. Younger, J.M., Chen, L., Ren, H.Y., Rosser, M.F., Turnbull, E.L., Fan, C.Y., Patterson, C., and Cyr, D.M. (2006). Sequential quality-control checkpoints triage misfolded cystic fibrosis transmembrane conductance regulator. Cell 126, 571–582.