Biochimica et Biophysica Acta 1820 (2012) 1457–1462

Contents lists available at SciVerse ScienceDirect

Biochimica et Biophysica Acta

j ourna l homepage: www.e lsev ie r .com/ locate /bbagen

A plant peptide: N-glycanase orthologue facilitates glycoprotein ER-associateddegradation in yeast

Yuki Masahara-Negishi a,1, Akira Hosomi a,1, Massimiliano Della Mea b,Donatella Serafini-Fracassini b, Tadashi Suzuki a,⁎a Glycometabolome Team, RIKEN Advanced Science Institute, 2–1 Hirosawa, Wako, Saitama 351–0198, Japanb Dipartimento di Biologia Evoluzionistica Sperimentale, Università di Bologna, 40126 Bologna, Italy

Abbreviations: PNGase, peptide:N-glycanase; ERADTG and TGase, transglutaminase; RTAΔ, ricin A chain ntransmembrane-Leu2; SC, synthetic complete; endo H⁎ Corresponding author. Tel.: +81 48 467 9628; fax:

E-mail address: tsuzuki_gm@riken.jp (T. Suzuki).1 Both authors contributed equally to this work.

0304-4165/$ – see front matter. Crown Copyright © 20doi:10.1016/j.bbagen.2012.05.009

a b s t r a c t

a r t i c l e i n f o

Article history:

Received 24 February 2012Received in revised form 14 May 2012Accepted 21 May 2012Available online 31 May 2012

Keywords:Peptide:N-glycanaseER-associated degradationYeast

Background: The cytoplasmic peptide:N-glycanase (PNGase) is a deglycosylating enzyme involved in the ER-associated degradation (ERAD) process, while ERAD-independent activities are also reported. Previous bio-chemical analyses indicated that the cytoplasmic PNGase orthologue in Arabidopsis thaliana (AtPNG1) canfunction as not only PNGase but also transglutaminase, while its in vivo function remained unclarified.Methods: AtPNG1 was expressed in Saccharomyces cerevisiae and its in vivo role on PNGase-dependent ERADpathway was examined.Results: AtPNG1 could facilitate the ERAD through its deglycosylation activity. Moreover, a catalytic mutantof AtPNG1 (AtPNG1(C251A)) was found to significantly impair the ERAD process. This result was found tobe N-glycan-dependent, as the AtPNG(C251A) did not affect the stability of the non-glycosylated RTAΔ

(ricin A chain non-toxic mutant). Tight interaction between AtPNG1(C251A) and the RTAΔ was confirmedby co-immunoprecipitation analysis.Conclusion: The plant PNGase facilitates ERAD through its deglycosylation activity, while the catalytic mutantof AtPNG1 impair glycoprotein ERAD by binding to N-glycans on the ERAD substrates.General significance: Our studies underscore the functional importance of a plant PNGase orthologue as adeglycosylating enzyme involved in the ERAD.

Crown Copyright © 2012 Published by Elsevier B.V. All rights reserved.

1. Introduction

Peptide:N-glycanase (PNGase) is a deglycosylating enzyme thatcleaves the β-aspartyl glycosylamine bond of N-linked glycoproteins,releasing intact N-glycans from proteins. PNGase activity was originallydiscovered in almond [1] and subsequently confirmed in bacteria [2].Since then, this enzyme has been widely used as a tool to analyse thestructure and functions of N-linked glycans on glycoproteins. The pres-ence of PNGase activity in animals was first reported in Medaka fish(Oryzias latipes) [3], and later in various mammalian-derived culturedcells [4,5]. While the fish enzyme is believed to be of lysosomal origin[6], mammalian PNGase is localised in the cytosol (hence, cytoplasmicPNGase) and its optimal activity is at neutral pH [5]. A gene encoding cy-toplasmic PNGase (PNG1) was identified in Saccharomyces cerevisiae[7]. The orthologous genes of PNG1 are widely distributed throughouteukaryotes.

, ER-associated degradation;on-toxic mutant; RTL, RTAΔ-, endoglycosidase H+81 48 467 9626.

12 Published by Elsevier B.V. All rig

Endoplasmic reticulum (ER)-associated degradation (ERAD) is acomponent of the quality control system for newly synthesised pro-teins. In this system, proteins that fail to fold correctly are degraded,while functional proteins are delivered to their intended destina-tions through the secretory pathway [8,9]. ERAD involves the extrac-tion of proteins from the ER to the cytosol, followed by proteasomaldegradation. Cytoplasmic PNGase is involved in the efficient degra-dation of some ERAD substrates [10–12]. The PNGase-mediateddeglycosylation during ERAD was also suggested to play an impor-tant role in antigen presentation by class I major histocompatibilitycomplex in mammalian cells [13–16]. However, some of the PNGaseorthologues were reported to be catalytically inactive, while its mu-tant exhibited severe phenotypic consequences [17,18], raisingthe possibility that, in addition to its enzyme activity, PNGaseorthologues might have significant enzyme-independent roles.

From an evolutionary view, cytoplasmic PNGase is an interestingprotein with a diverse structural arrangement [19,20]. The core cat-alytic domain of cytoplasmic PNGase is highly conserved throughouteukaryotes and, because of its homology with transglutaminase(TGase), cytoplasmic PNGase was categorised as a member of theTGase superfamily [19,21]. The plant orthologue of cytoplasmicPNGase was first identified in Arabidopsis thaliana (AtPNG1),based on the homology of a TGase domain [22], while the regions

hts reserved.

1458 Y. Masahara-Negishi et al. / Biochimica et Biophysica Acta 1820 (2012) 1457–1462

outside the TGase domain are quite unique and show no apparenthomology with the primary structures of animal/fungalorthologues (Fig. 1A). This indicates that the plant cytoplasmicPNGase may have followed a distinct evolutional route from othersto acquire its unique function. Interestingly, AtPNG1 was reportedto possess TGase activity [23], besides PNGase activity [24], in vitro.However, the functional role of this plant protein in vivo remainspoorly understood.

In this study, we utilized the assay system for PNGase-dependentERAD in yeast, using RTAΔ (ricin A chain non-toxic mutant; [10]) orRTL (RTAΔ-transmembrane-Leu2; [12]) as substrates to examine ifAtPNG1 can function as PNGase in vivo. Our results clearly indicatedthat AtPNG1, when expressed in yeast, can act as a deglycosylatingenzyme and facilitate the degradation of RTAΔ/RTL. Interestingly,its catalytic mutant, AtPNG1(C251A), was found to significantlystabilise the RTAΔ. Most importantly, these results were found tobe N-glycan specific, since non-glycosylated RTAΔ did not requireAtPNG1 for efficient degradation. Consistent with this finding,stable binding of AtPNG1(C251A) to RTAΔ was confirmed by co-immunoprecipitation experiments. Taken all results together, wedemonstrated that AtPNG1 can function as a PNGase in vivo and facil-itate the degradation of ERAD substrate in an N-glycan-dependentfashion.

A

B

C

Fig. 1. Heterologous expression of AtPNG1 and deglycosylation of RTAΔ. (A) Schematic repmain (catalytic PNGase domain) is indicated. (B) Western blotting analysis of RTAΔ. RTAΔw1 png1Δ cells. Cell extracts were resolved by SDS-PAGE and RTAΔwas visualised by immunoor no (g0) glycan. The immunoblot was also probed with anti-PGK antibody as a loading conexpressing AtPNG1 or its catalytic mutant, AtPNG1(C251A). AtPNG1 or AtPNG1(C251A) weror without leucine. The plates were incubated for 3 days at 30 °C.

2. Material and methods

2.1. Yeast strains and media

We used the following yeast strains: cim5-1 png1Δ cells (MATacim5-1 png1:: URA3 ura3-52 leu2Δ1 his3Δ200 FOAR; [25]) and png1Δcells (MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 png1Δ::kanMX4; [7]).Standard yeast media and genetic techniques were used [26,27].

2.2. Plasmid construction

Mutation of the catalytic Cys residue 251 to Ala was introduced to thepET28b-AtPNG1 plasmid [23] using QuikChange (Stratagene) to establishpET28b-AtPNG1(C251A). To construct pRS423GPD-AtPNG1-FLAG andpRS423GPD-AtPNG1(CA mutant)-FLAG, cDNA of A. thaliana AtPNG1 andAtPNG1(C251A) were amplified from pET28b-AtPNG1 and pET28b-AtPNG1(C251A), respectively, using the following primers, 5′-CACCA-TGGGAGAGGTATACGAA-3′ and 5′- ATGCGGCCGCCTACTTATCGTCGTC-ATCCTTGTAATCCTGGTGACTTCTGTACAGAT-3′, in which the secondprimer was designed to add a C-terminal FLAG tag, and were clonedinto pENTR™/D-TOPO (Invitrogen). The DNA sequences of the constructswere confirmed using BigDye ver. 3.1 and an ABI DNA sequencer(3730xl). The AtPNG1 and AtPNG1(C251A) genes in the pENTR vectors

resentation of AtPNG1 and ScPng1 (S. cerevisiae Png1). The transglutaminase (TG) do-as co-expressed with an empty vector (control; V), ScPng1 (Sc) or AtPNG1 (At) in cim5-blotting using an anti-ricin antibody. The arrows indicate RTAΔmodified with one (g1)trol. At, Arabidopsis thaliana; Sc, Saccharomyces cerevisiae. (C) RTL assay for png1Δ cellse co-expressed with RTL in png1Δ cells. Cells were plated on SC-galactose medium with

1459Y. Masahara-Negishi et al. / Biochimica et Biophysica Acta 1820 (2012) 1457–1462

were then transferred to the destination vector (pRS423GPD) [17] by LRClonase II reactions (Invitrogen) to generate pRS423GPD-AtPNG1-FLAGand pRS423GPD-AtPNG1(C251A)-FLAG. To generate pRS316-GPDRTAΔ,pRS315GPD-RTAΔ [11] was cut by SalI/NotI digestion, and the insert wascloned into the equivalent site of pRS316. pRS316-GPDHA-RTAΔwas gen-erated by site-directed mutagenesis (Quikchange).

2.3. Cycloheximide (CHX) decay assay

Strains harbouring the RTAΔ expression plasmid were grown andCHX was added to the cultures (final concentration, 250 μg/ml). The cul-tures were collected at the indicated times, and the cells were subjectedto western blotting analysis.

2.4. Preparation of yeast cell extracts and western blotting

Preparation of yeast cell extracts and western blotting were carriedout as previously described [12]. Antibodies were used at the following

Fig. 2. AtPNG1 acts as a PNGase and facilitates degradation of RTAΔ, while AtPNG1(C251A) sScPng1, AtPNG1 or AtPNG1(C251A). ScPng1, AtPNG1 and AtPNG1(C251A) were co-expressedcell line. CHXwas added at t=0min. Samples were collected at the indicated times and subjecblot was also probed with anti-PGK antibody as a loading control. (B) Quantitation of the CHX dLinear regression lines for each experiment are shown.

dilutions: 1:10,000 for anti-PGK (22C5) (Invitrogen) and anti-DYKDDDDK (Wako Chemicals Co.), and 1:500 for anti-HA (sc-805)(Santa Cruz).

2.5. Spotting assay

RTL spotting assay was carried out as previously described [12]. Inbrief, strains harbouring the RTL expression plasmid were spottedonto SC-uracil-histidinemedium or SC-uracil-histidine-leucinemediumcontaining galactose, and cells were incubated at 30 °C for 3 days. Pho-tographs of the gel were taken using FUJIFILM LAS-3000 mini (FujifilmCo., Tokyo, Japan).

2.6. Immunoprecipitation

Immunoprecipitation was carried out as previously described [12].

tabilises RTAΔ. (A) Cycloheximide (CHX) decay assay for RTAΔ in png1Δ cells expressingwith RTAΔ in png1Δ cells. The CHX decay assays were performed in each transformantted to SDS-PAGE, followed by immunoblotting using an anti-ricin antibody. The immuno-ecay assay shown in A. The log scale of the remaining protein (%) was plotted versus time.

Fig. 3. Effects of AtPNG1 or AtPNG1(C251A) on degradation of non-glycosylated RTAΔ, RTAΔ(N10QN236Q). AtPNG1 or AtPNG1(C251A) were co-expressed with the non-glycosylated RTAΔ, RTAΔ(N10QN236Q), in png1Δ cells. The cycloheximide decay assays were performed as shown in Fig. 2A.

1460 Y. Masahara-Negishi et al. / Biochimica et Biophysica Acta 1820 (2012) 1457–1462

3. Results

3.1. AtPNG1 acts as a PNGase in yeast and facilitates ERAD processes

It has been reported that cytoplasmic PNGase possesses a core TGasedomain, which contains a catalytic triad of Cys, His and Asp, and is es-sential for enzyme activity [21]. Previous studies showed that AtPNG1has both TGase [23] and PNGase activity [24]. However, the role ofAtPNG1 in living cells is still unclear, as its enzymatic activities wereonly detected by in vitro assays. In fact, PNGase activity was not detectedin a crude extract of Arabidopsis leaves [24] and we could not detectPNGase activity in AtPNG1 expressed in E. coli either (data not shown).These features make the functional assessment of AtPNG1 as a PNGasedifficult. Therefore, we aimed to utilize an in vivo assay using a yeast sys-tem to examine whether AtPNG1 can function as a deglycosylating en-zyme in the ERAD processes. To achieve this aim, we used a modelERAD substrate, RTAΔ, a non-toxic mutant of ricin A-chain [10]. This sub-strate is known to be degraded by ERAD in a PNGase (ScPng1)-dependentfashion [10–12], and provides an excellent system to measure in vivoPNGase activity. To determine whether AtPNG1 can facilitate the degra-dation of RTAΔ, C-terminal FLAG-tagged ScPng1 or AtPNG1 were co-expressed with RTAΔ in the cim5-1 png1Δ cells, which bear a mutationin a proteasome subunit and deletion of the PNG1 gene [25]. Three inde-pendent clones were picked up from ScPng1-expressing, AtPNG1-expressing and vector-bearing (control) cells, and the accumulation ofRTAΔ in each clone was examined by western blotting. As shown inFig. 1B, the g0 form of RTAΔ (i.e., un- or deglycosylated RTAΔ) wasdetected in all cells expressing ScPng1 or AtPNG1 (lanes 2, 3, 5, 6, 8 and9), whereas the g1 form of RTAΔ (i.e., glycosylated RTAΔ) was almost ex-clusively detected in control cells (lanes 1, 4 and 7). These resultssuggested that AtPNG1 can act as a PNGase and deglycosylates RTAΔ invivo during ERAD. It should also be noted that the amount of RTAΔ waslower in AtPNG1-expressing cells compared with control cells (comparelanes 1, 4, 7 and 3, 6, 9), indicating that AtPNG1 can, to some extent, sub-stitute ScPng1 in facilitating degradation of RTAΔ in yeast.

RTL (RTAΔ-transmembrane-Leu2 protein) is another Png1-dependent ERAD substrate [12]. RTL is a membrane protein that con-sists of a luminal RTA, transmembrane domain of Pdr5, and a cytoplas-mic Leu2 protein involved in biosynthesis of leucine. RTL assay is amethod for detection of defective ERAD by yeast cell growth. Thelogic behind this assay is that strains with leu2 mutation (bearingRTL-expressing plasmid) are usually unable to grow in media lackingleucine, whereas yeast mutants with defective ERAD pathway cansupport cell growth under such conditions because they are defectivein degrading Leu2 proteins on RTL. It has been shown that RTL, as wasthe casewith RTAΔ, was degraded in yeast in ScPng1-dependentman-ner [12]. RTL spotting assay therefore is a genetic device to character-ize the Png1-dependent ERAD pathway in yeast. To confirm whetherAtPNG1 is also required for RTL degradation, RTL spotting assay wasperformed. As shown in Fig. 1C, AtPNG1-expressing cells resulted inimpaired growth onmedia lacking leucine (Fig. 1C, middle row), indi-cating the efficient degradation of RTL. On the other hand, cells ex-pressing AtPNG1(C251A), where a catalytic Cys [28] is converted toAla, exhibited defect in degrading RTL, and therefore could grow onmedia lacking leucine to the level of vector control (Fig. 1C, bottomrow). This result clearly indicated that PNGase activity of AtPNG1 isalso required for efficient RTL degradation.

3.2. A catalytic mutant of AtPNG1 stabilises glycosylated ERAD substrates

To further examine whether the degradation of RTAΔ is indeed facili-tatedbyAtPNG1-mediateddeglycosylation,weperformed cycloheximide(CHX)-decay experiments. As shown in Fig. 2A, the g0 formwas detected,as was efficient degradation, in ScPng1- (top-right panel) and AtPNG1-expressing cells (bottom-left panel). By comparison, RTAΔ remained inthe g1 form during the chase period in control cells (top-left panel) andin cells expressing a catalytic mutant AtPNG1(C251A) (bottom-rightpanel) (Fig. 2A). By quantifying the amount of the g1 form of RTAΔ, itseemed that facilitated degradation was occurring in ScPng1- andAtPNG1-expressing cells (Fig. 2B). Surprisingly, expression of the catalyticmutant AtPNG1(C251A) resulted in significant stabilization of RTAΔ, asdemonstrated by the accumulation of RTAΔ in the steady state (Fig. 2A)and in the CHX-chase experiment (Fig. 2B; compare AtPNG1(C251A)with vector control). Taken together, these results indicate that AtPNG1facilitates the degradation of RTAΔ, while its catalytic mutant stabilisesRTAΔ.

3.3. Non-glycosylated RTAΔ was degraded in an AtPNG1-independentfashion

Previously it was shown that the cytoplasmic PNGase is only requiredfor degradation of RTAΔ, and its non-glycosylated mutant (RTAΔ-(N10QN236Q)), in which Asn in potential N-glycosylation sites are mu-tated into Gln, was not degraded in an ScPng1-dependent manner [12].To find out that it is also the case with AtPNG1, the effect of AtPNG1 ondegradation of non-glycosylated version of RTAΔ, RTAΔ(N10QN236Q)[12], was examined. As shown in Fig. 3, the nonglycosyled RTAΔ wasfound to be degraded in the presence of AtPNG1 as efficient as the vectorcontrol (compare the left panel and middle panel). As expected, expres-sion of the catalytic mutant, AtPNG1(C251A), also did not alter the effi-ciency of the degradation (right panel). This result contrasted with thecase with glycosylated RTAΔ (Fig. 2A and B), strongly indicating thatthe stabilization effect of AtPNG1(C251A) was specific to N-glycosylatedERAD substrates.

3.4. The catalytic mutant of AtPNG1 stabilises RTAΔ by binding to N-glycans on the substrate

Since AtPNG1(C251A) stabilised RTAΔ in an N-glycan dependentmanner, we assumed that the catalytic mutant may form a stable com-plexwith RTAΔ bybinding toN-glycans on RTAΔ. To investigate this hy-pothesis, we performed co-immunoprecipitation experiments usingAtPNG1 or AtPNG1(C251A). For this purpose, HA-tagged RTAΔ wasconstructed and AtPNG1(C251A) was immunoprecipitated with theanti-FLAG antibody. As shown in Fig. 4A, co-precipitation was observedfor HA-tagged RTAΔ and AtPNG1(C251A) (lane 2, top panel). The inter-action was specific because no band was observed in the absence ofAtPNG1(C251A) (lane 1, top panel). Moreover, wild-type AtPNG1could not be immunoprecipitated with RTAΔ under the same experi-mental conditions (data not shown), suggesting that mutation in thecatalytic residue render the AtPNG1 to form the stable complex withRTAΔ.

It should be noted that larger species of RTAΔ tended to be con-centrated in the precipitated fraction (compare top panel and middlepanel of lane 2). The upper band on precipitates was confirmed to be

1461Y. Masahara-Negishi et al. / Biochimica et Biophysica Acta 1820 (2012) 1457–1462

di-N-glycosylated RTAΔ, because endo H treatment resulted in com-plete deglycosylation of RTAΔ around the same position of the g0form (compare lanes 3 and 4 in the top panel). It has been shownthat, of two potential N-glycosylation sites, only Asn10 was efficientlyN-glycosylated, while the other site (Asn236) was scarcely used forN-glycosylation [12]. The fact that di-N-glycosylated RTAΔwas signif-icantly concentrated in the immunoprecipitates further suggests thatAtPNG1 preferentially binds to N-glycans. Considering all of these re-sults together, we concluded that the catalytic mutant of AtPNG1 sta-bilises RTAΔ by binding to the substrate in an N-glycan-dependentfashion. Consistent with this observation, the amino acids critical forcarbohydrate binding are all conserved in AtPNG1 [29] (Fig. 4B).

4. Discussion

While recent studies have unequivocally demonstrated that cyto-plasmic PNGase is involved in ERAD, the biological significance of thisenzyme has not been fully established in any system. For example,png1Δ cells derived from S. cerevisiae show no apparent growth defectsunder various experimental conditions [7], but they do exhibit delayeddegradation of some ERAD substrates [7,10,12]. More recently, using C.elegans, it was shown that a defect in a gene orthologue (png-1) causedexcessive axon branching in specific neurons [30], and this phenotypeappeared to be associated with the deglycosylation activity [31,32].On the other hand, a mutation of the PNGase orthologue in N. crassacaused temperature-sensitive growth/hyphae malformation pheno-types [33], despite the fact that PNGase is inactive because of theamino acid substitutions at the catalytic amino acids [17]. It was alsoreported that, while the catalytic triad was conserved, an orthologuein fruit fly (Pngl) exhibited no detectable PNGase activity possibly be-cause of the absence of a CXXC motif critical for enzyme activity, andthe mutants exhibited severe developmental delay [18]. Considering

Fig. 4. AtPNG1(C251A) can bind to RTAΔ. (A) Immunoprecipitation of the AtPNG1 catalyticcells. (B) Sequence alignment of the catalytic domains important for carbohydrate binding.indicated.

the remarkable structural diversity among species, researchers musttake carewhen evaluating the functional importance of this protein, be-cause we need to establish whether the observed properties are com-mon for all orthologues or unique for a certain subset of organisms.

In previous studies, it was reported that AtPNG1 retains PNGase ac-tivity [24] as well as TGase activity [23] in vitro. While multiple roles ofplant transglutaminases have been suggested, it remained to be demon-strated which activity can be attributed to AtPNG1 in vivo [34]. In thisstudy, we examined whether AtPNG1 has PNGase activity in vivo,using an ERAD assay for cytoplasmic PNGase-dependent substrates[10,12]. Our results indicate that AtPNG1 can function as a cytoplasmicPNGase and facilitates the degradation of RTAΔ/RTL in similar mannerto the yeast orthologue ScPng1. These results are consistentwith an ear-lier observation that ricin protein undergoes deglycosylation duringERAD in tobacco cells [35,36].

What surprised us, however, was the inhibitory effect ofAtPNG1(C251A), a catalytic mutant, on RTAΔ degradation. This resultclearly suggested that the catalytic mutant was not only functional,but can also impair substrate degradation. This activity was specificto N-glycosylated substrates, because the degradation of a non-glycosylated RTAΔ mutant was not affected by the expression ofAtPNG1(C251A). Furthermore, co-immunoprecipitation analysis re-vealed that the catalytic mutant can form a stable complexwith glyco-sylated RTAΔ. Consistent with this result, cytoplasmic PNGase isknown to bind carbohydrates with a preference for N-glycans [29,37].

Although AtPNG1(C251A) did not result in enhanced stability forRTL, it should be noted that RTL assay depends on stability of the cy-toplasmic Leu2 proteins, and while most of the components requiredfor RTL and RTAΔ are shared, there may be a difference, and such dis-crepancy has been observed for degradation of CPY* (a mutant formof carboxypeptidase Y), a well-studied ERAD substrate, and CTL*, aLeu2-bearing, transmembrane version of CPY* similar to RTL [38,39].

mutant with RTAΔ. AtPNG1(C251A) was coexpressed with HA-tagged RTAΔ in png1ΔThe catalytic residues (*) and residues important for carbohydrate binding (#) [29] are

1462 Y. Masahara-Negishi et al. / Biochimica et Biophysica Acta 1820 (2012) 1457–1462

It should be noted that, although our data clearly indicate thatAtPNG1 functions as a PNGase in vivo, it does not necessarily excludethe possibility that AtPNG1 may also function as a TGase in certaincellular processes. On-going research should be directed at dissectingthe physiological function of AtPNG1 and whether defects in this genehave any phenotypic consequences. If so, one must clarify whetherthe effects are due to enzyme (PNGase or other activities)-dependentor enzyme-independent activities, as is the case with N. crassa [17].

Acknowledgements

We thank the members of the Glycometabolome Team (RIKEN),particularly Dr. Hiroto Hirayama, for helpful discussion. We alsothank Dr. Yoshiki Yamaguchi (RIKEN) for valuable comments onthis manuscript. The work was supported in part by a Grant-in-Aidfor Scientific Research from the Ministry of Education, Culture, Sports,Science and Technology of Japan (to T.S. and A.H.) and RFO 2009 Uni-versity of Bologna (Italy) to D.S.F. We also thank COE (Center of Excel-lence) Program, Osaka University for their support to M.D. to carryout experiments in Japan.

References

[1] N. Takahashi, Demonstration of a new amidase acting on glycopeptides, Biochem.Biophys. Res. Commun. 76 (1977) 1194–1201.

[2] T.H. Plummer Jr., J.H. Elder, S. Alexander, A.W. Phelan, A.L. Tarentino, Demonstra-tion of peptide:N-glycosidase F activity in endo-beta-N-acetylglucosaminidase Fpreparations, J. Biol. Chem. 259 (1984) 10700–10704.

[3] A. Seko, K. Kitajima, Y. Inoue, S. Inoue, Peptide:N-glycosidase activity found in theearly embryos of Oryzias latipes (Medaka fish). The first demonstration of the oc-currence of peptide:N-glycosidase in animal cells and its implication for the pres-ence of a de-N-glycosylation system in living organisms, J. Biol. Chem. 266 (1991)22110–22114.

[4] T. Suzuki, A. Seko, K. Kitajima, Y. Inoue, S. Inoue, Identification of peptide:N-glycanase activity in mammalian-derived cultured cells, Biochem. Biophys.Res. Commun. 194 (1993) 1124–1130.

[5] T. Suzuki, A. Seko, K. Kitajima, Y. Inoue, S. Inoue, Purification and enzymatic prop-erties of peptide:N-glycanase from C3H mouse-derived L-929 fibroblast cells.Possible widespread occurrence of post-translational remodification of proteinsby N-deglycosylation, J. Biol. Chem. 269 (1994) 17611–17618.

[6] A. Seko, K. Kitajima, T. Iwamatsu, Y. Inoue, S. Inoue, Identification of two discretepeptide: N-glycanases in Oryzias latipes during embryogenesis, Glycobiology 9(1999) 887–895.

[7] T. Suzuki, H. Park, N.M. Hollingsworth, R. Sternglanz, W.J. Lennarz, PNG1, a yeast geneencoding a highly conserved peptide:N-glycanase, J. Cell Biol. 149 (2000) 1039–1052.

[8] W. Xie, D.T. Ng, ERAD substrate recognition in budding yeast, Semin. Cell Dev.Biol. 21 (2010) 533–539.

[9] R. Bernasconi, M. Molinari, ERAD and ERAD tuning: disposal of cargo and of ERADregulators from the mammalian ER, Curr. Opin. Cell Biol. 23 (2011) 176–183.

[10] I. Kim, J. Ahn, C. Liu, K. Tanabe, J. Apodaca, T. Suzuki, H. Rao, The Png1-Rad23 com-plex regulates glycoprotein turnover, J. Cell Biol. 172 (2006) 211–219.

[11] K. Tanabe, W.J. Lennarz, T. Suzuki, A cytoplasmic peptide:N-glycanase, MethodsEnzymol. 415 (2006) 46–55.

[12] A. Hosomi, K. Tanabe, H. Hirayama, I. Kim, H. Rao, T. Suzuki, Identification of anHtm1(EDEM)-dependent, Mns1-independent Endoplasmic Reticulum-associated Degra-dation (ERAD) pathway in Saccharomyces cerevisiae: application of a novel assayfor glycoprotein ERAD, J. Biol. Chem. 285 (2010) 24324–24334.

[13] M.L. Altrich-VanLith,M. Ostankovitch, J.M. Polefrone, C.A.Mosse, J. Shabanowitz, D.F.Hunt, V.H. Engelhard, Processing of a class I-restricted epitope from tyrosinase re-quires peptide N-glycanase and the cooperative action of endoplasmic reticulumaminopeptidase 1 and cytosolic proteases, J. Immunol. 177 (2006) 5440–5450.

[14] E. Kario, B. Tirosh, H.L. Ploegh, A. Navon, N-linked glycosylation does not impairproteasomal degradation but affects class I major histocompatibility complexpresentation, J. Biol. Chem. 283 (2008) 244–254.

[15] M. Ostankovitch, M. Altrich-Vanlith, V. Robila, V.H. Engelhard, N-glycosylationenhances presentation of a MHC class I-restricted epitope from tyrosinase,J. Immunol. 182 (2009) 4830–4835.

[16] A. Dalet, P.F. Robbins, V. Stroobant, N. Vigneron, Y.F. Li, M. El-Gamil, K. Hanada, J.C.Yang, S.A. Rosenberg, B.J. Van den Eynde, An antigenic peptide produced by re-verse splicing and double asparagine deamidation, Proc. Natl. Acad. Sci. U. S. A.108 (2011) E323–E331.

[17] S. Maerz, Y. Funakoshi, Y. Negishi, T. Suzuki, S. Seiler, The Neurospora peptide:N-glycanase ortholog PNG1 is essential for cell polarity despite its lack of enzy-matic activity, J. Biol. Chem. 285 (2010) 2326–2332.

[18] Y. Funakoshi, Y. Negishi, J.P. Gergen, J. Seino, K. Ishii, W.J. Lennarz, I. Matsuo, Y. Ito,N. Taniguchi, T. Suzuki, Evidence for an essential deglycosylation-independentactivity of PNGase in Drosophila melanogaster, PLoS One 5 (2010) e10545.

[19] T. Suzuki, H. Park, W.J. Lennarz, Cytoplasmic peptide:N-glycanase (PNGase) in eu-karyotic cells: occurrence, primary structure, and potential functions, FASEB J. 16(2002) 635–641.

[20] T. Suzuki, Cytoplasmic peptide:N-glycanase and catabolic pathway for freeN-glycans in the cytosol, Semin. Cell Dev. Biol. 18 (2007) 762–769.

[21] K.S. Makarova, L. Aravind, E.V. Koonin, A superfamily of archaeal, bacterial, andeukaryotic proteins homologous to animal transglutaminases, Protein Sci.8 (1999) 1714–1719.

[22] T. Suzuki, H. Park, E.A. Till, W.J. Lennarz, The PUB domain: a putativeprotein-protein interaction domain implicated in the ubiquitin-proteasome path-way, Biochem. Biophys. Res. Commun. 287 (2001) 1083–1087.

[23] M. Della Mea, D. Caparros-Ruiz, I. Claparols, D. Serafini-Fracassini, J. Rigau,AtPng1p. The first plant transglutaminase, Plant Physiol. 135 (2004) 2046–2054.

[24] A. Diepold, G. Li, W.J. Lennarz, T. Nurnberger, F. Brunner, The Arabidopsis AtPNG1gene encodes a peptide:N-glycanase, Plant J. 52 (2007) 94–104.

[25] T. Suzuki, W.J. Lennarz, Glycopeptide export from the endoplasmic reticulum intocytosol is mediated by a mechanism distinct from that for export of misfoldedglycoprotein, Glycobiology 12 (2002) 803–811.

[26] M.D. Rose, F. Winston, P. Hieter, Methods in Yeast Genetics, Cold Spring HarborLaboratory, Cold Spring Harbor, NY, 1990.

[27] F. Sherman, Getting started with yeast, Methods Enzymol. 194 (1991) 3–21.[28] S. Katiyar, T. Suzuki, B.J. Balgobin, W.J. Lennarz, Site-directed mutagenesis study

of yeast peptide:N-glycanase. Insight into the reaction mechanism ofdeglycosylation, J. Biol. Chem. 277 (2002) 12953–12959.

[29] G. Zhao, G. Li, X. Zhou, I. Matsuo, Y. Ito, T. Suzuki, W.J. Lennarz, H. Schindelin,Structural and mutational studies on the importance of oligosaccharide bindingfor the activity of yeast PNGase, Glycobiology 19 (2009) 118–125.

[30] N. Habibi-Babadi, A. Su, C.E. de Carvalho, A. Colavita, The N-glycanase png-1 actsto limit axon branching during organ formation in Caenorhabditis elegans, J. Neu-rosci. 30 (2010) 1766–1776.

[31] T. Suzuki, K. Tanabe, I. Hara, N. Taniguchi, A. Colavita, Dual enzymatic propertiesof the cytoplasmic peptide: N-glycanase in C. elegans, Biochem. Biophys. Res.Commun. 358 (2007) 837–841.

[32] T. Kato, A. Kawahara, H. Ashida, K. Yamamoto, Unique peptide:N-glycanase ofCaenorhabditis elegans has activity of protein disulphide reductase as well as ofdeglycosylation, J. Biochem. 142 (2007) 175–181.

[33] S. Seiler, M. Plamann, The genetic basis of cellular morphogenesis in the filamen-tous fungus Neurospora crassa, Mol. Biol. Cell 14 (2003) 4352–4364.

[34] D. Serafini-Fracassini, M. Della Mea, G. Tasco, R. Casadio, S. Del Duca, Plant and an-imal transglutaminases: do similar functions imply similar structures? AminoAcids 36 (2009) 643–657.

[35] A. Di Cola, L. Frigerio, J.M. Lord, A. Ceriotti, L.M. Roberts, Ricin A chain without itspartner B chain is degraded after retrotranslocation from the endoplasmic reticu-lum to the cytosol in plant cells, Proc. Natl. Acad. Sci. U. S. A. 98 (2001)14726–14731.

[36] A. Di Cola, L. Frigerio, J.M. Lord, L.M. Roberts, A. Ceriotti, Endoplasmicreticulum-associated degradation of ricin A chain has unique and plant-specificfeatures, Plant Physiol. 137 (2005) 287–296.

[37] T. Suzuki, I. Hara, M. Nakano, G. Zhao, W.J. Lennarz, H. Schindelin, N. Taniguchi, K.Totani, I. Matsuo, Y. Ito, Site-specific labeling of cytoplasmic peptide:N-glycanaseby N, N'-diacetylchitobiose-related compounds, J. Biol. Chem. 281 (2006)22152–22160.

[38] B. Medicherla, Z. Kostova, A. Schaefer, D.H. Wolf, A genomic screen identifiesDsk2p and Rad23p as essential components of ER-associated degradation,EMBO Rep. 5 (2004) 692–697.

[39] S. Kohlmann, A. Schafer, D.H. Wolf, Ubiquitin ligase Hul5 is required forfragment-specific substrate degradation in endoplasmic reticulum-associateddegradation, J. Biol. Chem. 283 (2008) 16374–16383.