Mnl1p, an α-Mannosidase-like Protein in YeastSaccharomyces cerevisiae, Is Required for Endoplasmic Reticulum-associated Degradation of Glycoproteins*

The endoplasmic reticulum (ER) has a mechanism to block the exit of misfolded or unassembled proteins from the ER for the downstream organelles in the secretory pathway. Misfolded proteins retained in the ER are subjected to proteasome-dependent degradation in the cytosol when they cannot achieve correct folding and/or assembly within an appropriate time window. Although specific mannose trimming of the protein-bound oligosaccharide is essential for the degradation of misfolded glycoproteins, the precise mechanism for this recognition remains obscure. Here we report a new α-mannosidase-like protein, Mnl1p (mannosidase-like protein), in the yeast ER. Mnl1p is unlikely to exhibit α1,2-mannosidase activity, because it lacks cysteine residues that are essential for α1,2-mannosidase. However deletion of the MNL1 gene causes retardation of the degradation of misfolded carboxypeptidase Y, but not of the unglycosylated mutant form of the yeast α-mating pheromone. Possible roles of Mnl1p in the degradation and in the ER-retention of misfolded glycoproteins are discussed.

The endoplasmic reticulum (ER) 1 is the site of entry for proteins destined for the secretory pathway. The ER provides an optimized environment for correct maturation, including correct folding, oligomerization, N-and O-linked glycosylation, and disulfide bond formation of proteins imported into the ER (1,2). Several components, including molecular chaperones and folding enzymes that mediate these processes, have been identified in the ER. Shortage or defects of these components as well as mutations in secretory proteins and environmental stress tend to result in failure and therefore misfolding of secretory proteins. The misfolded proteins are retained in the ER and are subjected to retrial for correct maturation with the aid of molecular chaperones and folding enzymes. However if this process is not successful, prolonged retention of the misfolded proteins in the ER eventually leads to their degradation (3,4). The ER-associated degradation (ERAD) requires the retrotranslocation of misfolded proteins through the Sec61 channel from the ER lumen to the cytosol and subsequent degradation by the 26 S proteasome located in the cytosol (5)(6)(7)(8)(9).
For ERAD of aberrant proteins, the question of how proteins that are misfolded and to be degraded are identified and targeted for the retrotranslocation system has not been resolved. In yeast Saccharomyces cerevisiae, slow removal of ␣1,2-mannose from the middle branch of the protein-bound oligosaccharide, or formation of Man 8 GlcNAc 2 , has been shown to be critical for the degradation of misfolded carboxypeptidase Y (CPY), a vacuolar glycoprotein (10). This is consistent with the reports that inhibition of ␣-mannosidase trimming stabilizes specific misfolded glycoproteins in the mammalian ER (11,12). On the basis of these observations, it has been proposed that the trimming of the glycoprotein-bound oligosaccharide may well function as the biological timer for the onset of the glycoprotein degradation that prevents permanent residence of misfolded glycoproteins in the ER (2). This naturally suggests the presence of a Man 8 GlcNAc 2 -binding lectin involved in ERAD of misfolded glycoproteins, which remains to be identified (10).
In the present study, we identified a new ␣-mannosidase-like protein, Mnl1p (mannosidase-like protein), in the yeast ER. Although it shows some homology with Mns1p, yeast ␣1,2mannosidase, Mnl1p is unlikely to exhibit the ␣1,2-mannosidase activity, because it lacks cysteine residues that are essential for the ␣1,2-mannosidase activity (13). However deletion of the MNL1 gene resulted in retardation of the degradation of misfolded CPY, but not of the unglycosylated mutant form of the yeast ␣-mating pheromone (⌬Gp␣F). Possible roles of Mnl1p in ERAD and the ER retention of misfolded glycoproteins will be discussed.

EXPERIMENTAL PROCEDURES
Plasmids, Strains, and Culturing Conditions-Standard recombinant techniques were employed using an Escherichia coli strain TG1 (supE hsd⌬5 thi⌬ (lac-proAB) FЈ [tra⌬36 proAB ϩ lacI q lacZ⌬M15]). Yeast strains used in this study were SEY6210 (MAT␣ ura3 leu2 trp1 his3 lys2 suc2) (14) and KYSC1 (MAT␣ prc1-1 ura3 leu2 trp1 his3 lys2 suc2). 2 Yeast strain BJ3505 (MATa pep4::HIS3 prb1 lys2 trp1 ura3 gal2 can1) was used for preparation of the cytosol (9). Cells were grown in YPD medium containing 1% yeast extract, 2% polypeptone, and 2% glucose. A sulfate-free synthetic minimal medium (15) was used for metabolic labeling of yeast cells. The gene for the C-terminally HA (influenza hemagglutinin) epitope-tagged version of Mnl1p was cloned by PCR from the yeast genomic DNA. The amplified DNA was sub-* This work was supported by grants-in-aid for Scientific Research from the Ministry of Education, Science and Culture of Japan. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
ERAD Assays-Metabolic labeling of yeast cells with Tran 35 S-label (ICN) and preparation of cell extracts were performed as described previously (19). Immunoprecipitation was performed as described by Nishikawa et al. (20). Yeast microsome fractions were prepared from wild-type cells and ⌬mnl1 cells (9) and the yeast cytosol S100 fraction from yeast BJ3505 cells. 35 S-Labeled mutant prepro-␣ factor lacking the three consensus glycosylation sites (⌬Gpp␣F) (9) was synthesized in vitro using a yeast cell-free translation. The in vitro ERAD assay was performed as described previously (9).
Fluorescence Microscopy-Immunofluorescent staining of yeast cells was performed as described previously (20) with minor modifications. Incubation with the primary and the secondary antibodies was performed as follows: 1) 1:250 dilution of the rabbit anti-BiP polyclonal antibodies (21) and 1:250 dilution of the 16B12 mouse monoclonal antibody and 2) 1:100 dilution of the rhodamine-conjugated goat antirabbit IgG antibody, 1:100 dilution of the fluorescein isothiocyanateconjugated goat anti-mouse IgG antibody.

Mnl1p
Is an ER Protein-MNL1 (YHR204W) on yeast chromosome VIII encodes an ␣-mannosidase-like protein, which is 796 amino acids long, with a calculated molecular weight of 92,206. The predicted amino acid sequence of Mnl1p shows 25% identity with Mns1p, yeast ␣1,2-mannosidase that converts glycosyl residues from Man 9 GlcNAc 2 to Man 8 GlcNAc 2 in the ER (Fig. 1A). Mns1p is a type II ER transmembrane glycoprotein with its N-terminal hydrophobic segment functioning as a signal-anchor sequence. The question of whether the segment of hydrophobic residues 1-21 ( Fig. 1A) functions as a signalanchor sequence to make Mnl1p a type-II membrane protein should await future studies. The amino acid residues involved in the substrate binding, Ca 2ϩ binding, and catalytic activity in Mns1p (22) are conserved in Mnl1p except for the two cysteine residues (Cys 340 and Cys 385 in Mns1p) that form a disulfide bridge and are essential for the ␣-mannosidase activity (13,22). The lack of the essential cysteine residues in Mnl1p suggests that Mnl1p is unlikely to have the ␣-mannosidase activity. The C-terminal part of Mnl1p shows no homology with any proteins deposited in the data base.
To assess the functions of Mnl1p, we first analyzed its subcellular location by indirect immunofluorescent microscopy. Mnl1p was tagged at the C terminus with the HA epitope for recognition by the monoclonal antibody 16B12. Cells that express the HA-tagged version of Mnl1p from a multicopy plasmid were fixed, permeabilized, and subjected to staining with the anti-HA antibody. Staining with the anti-HA antibody showed perinuclear staining with tubular extensions in the cytoplasm (Fig. 1B, panel b). This staining is typical for yeast ER proteins, and nearly identical staining was observed with the anti-BiP antibodies (Fig. 1B, panel a). Cells that contain a multicopy plasmid without expression of the HA-tagged Mnl1p did not show such staining (data not shown). These results indicate that Mnl1p resides exclusively in the ER.
Deletion of MNL1 Causes a Defect in ERAD of CPY*-The trimming of N-linked oligosaccharides is critical for proteasomedependent ERAD of misfolded glycoproteins. For example in yeast, removal of ␣1,2-glucose by glucosidase I and glucosidase II and that of ␣1,2-mannnose by ER-␣1,2-mannosidase, Mns1p, to yield Man 8 GlcNAc 2 are essential for the efficient degradation of CPY*, mutated and therefore misfolded carboxypeptidase Y (10,23). Although Mnl1p is not expected to have the ␣-mannosidase activity, its homology with Mns1p raises the possibility that Mnl1p is involved in the recognition of specific oligosaccharide structures and/or ERAD of misfolded glycopro-teins. We thus analyzed the effects of deletion of the MNL1 gene on ERAD of a model misfolded protein, CPY*, in vivo. The ⌬mnl1 null mutant strain did not show any detectable growth phenotypes as reported previously (24).
CPY is known to receive distinct posttranslational modifications in different cellular compartments on its transport pathway to the vacuole (25,26). On translocation across the ER membrane, a prepro form of CPY (prepro-CPY) receives proteolytic cleavage of the signal sequence and addition of four N-linked oligosaccharide chains to generate a 67-kDa ER form, p1CPY. In the Golgi complex, p1CPY is converted to a 69-kDa form, p2CPY, by addition of mannose to the N-linked oligosac- Two conserved cysteine residues that are essential for the ␣-mannosidase activity are indicated with asterisks. B, localization of Mnl1p by immunofluorescence microscopy. Cells of SEY6210/pKNM1 were grown in SCD medium (minimal medium containing 2% (w/v) glucose and 0.5% (w/v) casamino acids) lacking uracil at 30°C and analyzed by double label immunofluorescence microscopy using anti-BiP polyclonal antibodies and the 16B12 monoclonal antibody. Panels a, b, and c show the same field of the fluorescent images stained with the anti-BiP antibodies, and the 16B12 antibody, and 4Ј,6-diamidino-2phenylindole, respectively. Bar, 2 m. ␣-Mannosidase-like Protein of the Yeast ER Mediates ERAD charide chains (26). After reaching the vacuole, p2CPY becomes a 61-kDa mature form, mCPY (26,27). When wild-type cells or ⌬mnl1 cells were metabolically labeled with 35 S-containing amino acids for 5 min at 30°C, the p1 form of labeled CPY* was recovered as pellets by immunoprecipitation ( Fig.  2A, upper panel, chase 0 min). When the fate of labeled p1CPY* in wild-type cells was followed, the amount of p1CPY* decreased rapidly without changing its molecular mass during the chase period. This indicates that misfolded p1CPY* was prevented from its exit from the ER for the Golgi complex and was degraded efficiently with a half-life of 18 min at 30°C (Fig.  2, A and B). In contrast, the p1 form of labeled CPY* in ⌬mnl1 cells was degraded 3-fold more slowly than in wild-type cells (Fig. 2B). This suggests that Mnl1p is, whether directly or indirectly, involved in ERAD of CPY*.
We also noticed that p1CPY* in ⌬mnl1 null mutant cells was converted to a higher molecular weight form, which was designated as p1ЈCPY* and was observed as a smear on the SDS-PAGE gel, during the chase period ( Fig. 2A). When p1ЈCPY* was treated with endoglycosidase H, it was converted to a deglycosylated form with the same electrophoretic mobility as that for p1CPY* (Fig. 2A). This means that p1ЈCPY* arose from glycosyl modifications of p1CPY*. As a control, we confirmed that the deletion of MNL1 does not affect the exit of p1CPY from the ER for the Golgi complex and for the vacuole (not shown).
A possible explanation for the glycosyl modifications of p1CPY* in ⌬mnl1 cells is that a part of misfolded p1CPY* escaped the ER and received addition of ␣136and/or ␣133linked mannose to core oligosaccharides in the Golgi complex (25,28). To test if p1ЈCPY* had been modified by the Golgi enzymes, reactivity of p1ЈCPY* to the anti-␣136 mannose and the anti-␣133 mannose antibodies was examined. Wild-type cells expressing CPY or CPY* and ⌬mnl1 cells expressing CPY* were metabolically labeled for 5 min and chased, and the CPY or CPY* species were immunoprecipitated with the anti-CPY antibodies from the cell extracts. The immunoprecipitated proteins were subjected to the second-round immunoprecipitation with the antibodies against CPY, ␣136 mannose, or ␣133 mannose. As shown in Fig. 4, p1ЈCPY* in ⌬mnl1 cells was immunoprecipitated with the anti-␣136 mannose antibodies but not with the anti-␣133 mannose antibodies (Fig. 3, lanes  14 and 15), whereas p1CPY* was not recognized by these antibodies (Fig. 3, lane 13). p1CPY* in wild-type cells, which was retained in the ER and therefore not glycosylated by the Golgi enzymes, was not precipitated with the anti-␣136 mannose or anti-␣133 mannose antibodies (Fig. 3, lanes 9 and 10). As a control, it was confirmed that mCPY in wild-type cells, which had received carbohydrate modification in the Golgi complex, were precipitated with these antibodies (Fig. 3, lanes  4 and 5). These results indicate that p1ЈCPY* was modified by the early Golgi enzymes in ⌬mnl1 cells and that it is the intermediate form of the mannose addition in the Golgi complex. In other words, deletion of the MNL1 gene allowed exit of a fraction of p1CPY*, which is strictly retained in the ER in wild-type cells, from the ER for at least the early Golgi cisternae (29).
Mnl1p Is Not Involved in ERAD of ⌬Gp␣F-We next asked if Mnl1p is involved in ERAD of misfolded but unglycosylated proteins. For this purpose, we performed in vitro export/degradation assays with an unglycosylated mutant form (⌬Gp␣F) of the yeast ␣-mating pheromone, pro-␣-factor (p␣F), as a substrate (8,9,30,31). A radiolabeled precursor form of ⌬Gp␣F (⌬Gpp␣F) was translocated into microsomes prepared from the ⌬mnl1 strain or from the isogenic wild-type strain. Upon signal sequence cleavage, the resulting product, ⌬Gp␣F, becomes an ERAD substrate when the washed vesicles are incubated in the presence of the cytosol and ATP (9). When this assay was performed using microsomes prepared from ⌬mnl1 cells, degradation of ⌬Gp␣F was as efficient as that with wild-type microsomes at 30°C (Fig. 4A).
In parallel with these in vitro degradation assays, we followed the degradation of p␣F in vivo by pulse-chase experiments. The previous study showed that tunicamycin, an inhibitor of the N-linked glycosylation in the ER, causes formation of the unglycosylated and therefore misfolded ␣F precursor (32). 2 The unglycosylated p␣F is rapidly degraded in a Sec61p-dependent manner, suggesting that it is a substrate for ERAD in vivo ((32) ; Fig. 4C). Wild-type cells and ⌬mnl1 cells were metabolically labeled for 5 min in the presence of tunicamycin, and p␣F was recovered as pellets by immunoprecipitation by the anti-␣-factor antibodies. As shown in Fig. 4B, unglycosylated p␣F was rapidly degraded in ⌬mnl1 cells with nearly the same kinetics as that for wild-type cells. Taken together, these re- cells (SNY1080, ⌬mnl1) expressing CPY* (CPY*/⌬mnl1) were pulselabeled with 35 S-containing amino acids for 5 min at 30°C and chased for the indicated times. Cell extracts were prepared from the labeled cells and subjected to immunoprecipitation with the anti-CPY antibodies (1st antibody). The immunocomplexes were solubilized, and aliquots were subjected to the second round of immunoprecipitation with the indicated antibodies (2nd antibody; the 1-6 and 1-3 antibodies are specific for the ␣1, 6-, and ␣1,3-mannose linkages). The amounts of cell extracts for lanes 2-5, 7-10, and 12-15 were twice as much as those for lanes 1, 6, and 11, respectively. The immunoprecipitated proteins were analyzed by SDS-PAGE and radioimaging. p1, p1CPY; p1Ј, p1ЈCPY*; p2, p2CPY; m, mCPY.

␣-Mannosidase-like Protein of the Yeast ER Mediates ERAD
sults of the in vivo as well as the in vitro degradation assays strongly suggest that Mnl1p is not required for ERAD of unglycosylated proteins. DISCUSSION In the present study, we have characterized the functions of Mnl1p, an ␣-mannosidase-like protein in the ER. Deletion of the MNL1 gene led to retardation of ERAD of CPY*, a misfolded glycoprotein, but not of ⌬Gp␣F, a misfolded unglycosylated protein, suggesting that Mnl1p plays important roles in ERAD of misfolded proteins with N-linked oligosaccharides. Then what is the role of Mnl1p in ERAD? The lack of two cysteine residues strongly suggests that Mnl1p does not have the ␣-mannosidase activity. Mnl1p may be involved in the recognition of oligosaccharide structures, including Man 8 GlcNAc 2 , which is essential for ERAD substrates, of misfolded proteins. Indeed, a recently identified mammalian counterpart of Mnl1p does not have the mannosidase activity and probably functions as a lectin that specifically recognizes the oligosaccharide structure of Man 8 GlcNAc 2 (34).
The fate of CPY* was followed in detail when its degradation was impaired by the deletion of the MNL1 gene. Whereas CPY* was partly retained as p1CPY*, the ER form to be degraded, a part of p1CPY* molecules were converted to the early-Golgi form, p1ЈCPY*, containing ␣136 mannose. This indicates that at least some p1CPY* was released from the ER to the Golgi complex. However, since ␣133 mannose was not attached to p1ЈCPY* even after the prolonged chase, CPY* most likely undergoes recycling through the early-Golgi cisternae into the ER. Deletion of DER1 encoding an ER protein involved in ERAD also leads to escape of some CPY* from the ER for the early-Golgi compartment (33). Therefore ER-resident proteins that are involved in ERAD are also responsible for retaining their substrates in the ER probably by their affinity for misfolded proteins. FIG. 4. Deletion of MNL1 does not affect the degradation of ⌬Gp␣F either in vitro or in vivo. A, in vitro ERAD assays for radiolabeled ⌬Gpp␣F. 35 S-Labeled ⌬Gpp␣F was translocated into microsomes prepared from wild-type (SEY6210, wt) and ⌬mnl1 (SNY1079, ⌬mnl1) cells at 20°C for 50 min. Microsomes containing ⌬Gp␣F were collected by centrifugation, washed once with buffer 88 (20 mM HEPES-KOH, pH 6.8, 250 mM sorbitol, 150 mM KOAc, 5 mM Mg(OAc) 2 ) (9), and resuspended in buffer 88. The microsomes were further incubated in buffer 88 containing an ATP-regenerating system (1 mM ATP, 50 M GDP-mannose, 40 mM creatine phosphate, 0.2 mg/ml creatine phosphokinase) and the yeast cytosol (5 mg of protein/ml) at 30°C. At the indicated time points, proteins were precipitated with trichloroacetic acid and subjected to SDS-PAGE in the presence of 4 M urea. The upper band is ⌬Gpp␣F, which is bound to the cytosolic surface of the microsomes and escaped ERAD. B, quantification of the results shown in A. The amount of ⌬Gp␣F at 0-min incubation was set to 100%. Data represent the average of the results of two independent assays. C, in vivo ERAD assays for radiolabeled p␣F. Wild-type (SEY6210, wt) cells and ⌬mnl1 (SNY1079, ⌬mnl1) cells were pulse-labeled with 35 S-containing amino acids for 5 min at 30°C in the presence of 10 g/ml of tunicamycin, chased for the indicated times, and subjected to immunoprecipitation with the anti-␣F antibodies. D, quantification of the results shown in C. The amount of ⌬Gp␣F at 0-min chase was set to 100%. Data represent the average of the results of two independent assays. ␣-Mannosidase-like Protein of the Yeast ER Mediates ERAD 8638