Analysis of the vacuolar luminal proteome of Saccharomyces cerevisiae Jean-Emmanuel Sarry1*, Sixue Chen2*, Richard P. Collum1, Shun Liang1, Mingsheng Peng1, Albert Lang1, Bianca Naumann1, Florence Dzierszinski1, Chao-Xing Yuan3, Michael Hippler1 and Philip A. Rea1

1 Department of Biology, University of Pennsylvania, Philadelphia, PA, USA 2 Department of Botany, Genetics Institute, University of Florida, Gainesville, FL, USA 3 Proteomics Core Facility, University of Pennsylvania, Philadelphia, PA, USA

Keywords Despite its large size and the numerous processes in which it is implicated, 2D gel electrophoresis; luminal ; neither the identity nor the functions of the proteins targeted to the yeast mass spectrometry; proteome; vacuole vacuole have been defined comprehensively. In order to establish a method- purification ological platform and inventory to address this shortfall, we refined Correspondence techniques for the purification of ‘proteomics-grade’ intact vacuoles. As P. A. Rea, Plant Science Institute, confirmed by retention of the preloaded fluorescent conjugate glutathione– Department of Biology, Carolyn Hoff Lynch bimane throughout the fractionation procedure, the resistance of soluble Biology Laboratory, 433 South University proteins that copurify with this fraction to digestion by exogenous extra- Avenue, University of Pennsylvania, vacuolar proteinase K, and the results of flow cytometric, western and mar- Philadelphia, PA 19104, USA ker activity analyses, vacuoles prepared in this way retain most of Fax: +1 215 898 8780 their protein content and are of high purity and integrity. Using this mate- Tel. +1 215 898 0807 E-mail: [email protected] rial, 360 polypeptides species associated with the soluble fraction of the vacuolar isolates were resolved reproducibly by 2D gel electrophoresis. Of *These authors contributed equally to this these, 260 were identified by peptide mass fingerprinting and peptide work sequencing by MALDI-MS and liquid chromatography coupled to ion trap or quadrupole TOF tandem MS, respectively. The polypeptides identified (Received 28 March 2007, revised 30 May in this way, many of which correspond to alternate size and charge states 2007, accepted 27 June 2007) of the same parent translation product, can be assigned to 117 unique doi:10.1111/j.1742-4658.2007.05959.x ORFs. Most of the proteins identified are canonical vacuolar , glycosidases, phosphohydrolases, lipid-binding proteins or established vacu- olar proteins of unknown function, or other proteases, glycosidases, lipid- binding proteins, regulatory proteins or proteins involved in intermediary metabolism, protein synthesis, folding or targeting, or the alleviation of oxidative stress. On the basis of the high purity of the vacuolar prepara- tions, the electrophoretic properties of the proteins identified and the results of quantitative proteinase K protection measurements, many of the noncanonical vacuolar proteins identified are concluded to have entered this compartment for breakdown, processing and ⁄ or salvage purposes.

The vacuole of the budding yeast Saccharomyces cere- processes ranging from macromolecule degradation visiae, which can occupy as much as 25% of the total and salvage, pH and general ion homeostasis, osmo- intracellular volume, participates in numerous cellular regulation and volume regulation, to the storage of

Abbreviations APE1, aminopeptidase I; APE3, aminopeptidase Y; BLH1, bleomycin ; 2-DE, 2D gel electrophoresis; ER, endoplasmic reticulum; GSH, glutathione; HSP, heat shock protein; LC, liquid chromatography; PRB1, vacuolar b; PRC1, carboxypeptidase Y; QTOF, quadrupole time-of-flight; SGD, Saccharomyces Genome Database; SOB, sorbitol buffer; SUC, sucrose buffer.

FEBS Journal 274 (2007) 4287–4305 ª 2007 The Authors Journal compilation ª 2007 FEBS 4287 Proteomics of yeast vacuolar lumen J.-E. Sarry et al. amino acids, carboxylic acids, carbohydrates and some of the vacuoles of other fungal systems and plants, vitamins, and the sequestration of endogenous and and also the lysosomal compartments of animal cells. exogenous toxins. Given this degree of multifunctional- Here we describe refinement of a procedure to purify ity, what is perhaps surprising is how little is known of intact vacuoles from yeast and elucidate the identity of the range of proteins found in this compartment and the protein species found in the lumen by the com- the types of modifications to which they are subject. bined application of 2D gel electrophoresis (2-DE), The vacuole is known to be a major site for protein MALDI-TOF-MS and liquid chromatography (LC) turnover in the yeast cell. As such, several alternate coupled to electrospray ion trap or quadrupole time- but overlapping protein-transport pathways converge of-flight (QTOF) tandem MS. In so doing, we establish on this organelle [1–3]. Several canonical vacuolar pro- a methodological platform for defining the proteome teases, as newly synthesized proteins, are transported of the vacuolar lumen of S. cerevisiae and a compre- via the secretory pathway to the endoplasmic reticulum hensive data set containing polypeptide species that (ER) lumen or membrane, transit from the ER to the may be subjected to further functional characteriza- Golgi apparatus, and in the late Golgi are diverted to tion. the vacuole. Other proteins, as exemplified by alkaline To date, there have been three published analyses of phosphatase, enter the vacuole via the so-called ‘alka- the vacuolar proteome in the model plant Arabidopsis line phosphatase’ pathway by the direct fusion of thaliana. Two of these focus exclusively on the vacuo- Golgi-derived vesicles, or in the case of those coming lar membrane [5,6], and the third, which examines from the cell surface, are delivered to the vacuole luminal as well membrane proteins [7], though seminal, endocytotically by the formation of multilamellar is difficult to assess and put into context for two rea- bodies that are released into the vacuole lumen. Oth- sons. First, it does not systematically address the ers, specifically some cytoplasmic proteins, enter the purity of the luminal protein fraction or take special vacuole by cytoplasm-to-vacuole targeting which over- precautions to guard against contamination by non- laps with the autophagic pathway, the mechanism vacuolar luminal proteins, so compromising assessment responsible for the nonselective delivery of cytosolic of the tightness of association of the proteins in ques- proteins and organelles in their entirety to the vacuole tion with the vacuolar compartment. Second, it is for degradation under stress conditions. based exclusively on the results of nonquantitative Despite this wealth of information on the mecha- shotgun approaches: multidimensional LC-tandem MS nisms of protein trafficking into the vacuole there is a or the combination of 1D gel electrophoresis-coupled lack of fundamental systems-level knowledge of the LC-MS. As a result, there is little or no information range of proteins found in this compartment. This on the relative levels of the proteins identified or a may pose an impasse for the rational analysis of many 2-DE-based map to which other researchers, who do cellular processes and ultimately cellular engineering. If not have immediate access to MS resources, might vacuolate cells, for instance those of yeast, are to be refer. Here, we present investigations directed at over- manipulated for enhanced nutritional quality, the pro- coming limitations of this type and assembly of the vision of pharmaceuticals or their precursors, the first protein map of the yeast vacuolar lumen. provision of precursors for manufacturing purposes, or environmental remediation applications, ready access Results and Discussion to a luminal proteomics toolbox detailing the protein profile of the vacuole lumen and how the latter is Assessment of vacuolar purity and integrity established and maintained by intravacuolar reactions and vacuolar protein-trafficking pathways is critical. Intact vacuoles were routinely purified from S. cerevi- S. cerevisiae is a model system for the identification siae strain SEY6210 using a procedure based on those and definition of eukaryotic protein functions because described by Wiemken et al. [8], Roberts et al. [9] and it is especially easy to manipulate molecularly and pos- Kim et al. [10], in which lysed spheroplasts were sub- sesses only a moderate number of ORFs, most of jected to multiple cycles of density-gradient and flota- which are devoid of the interpretatively complicating tion centrifugation on sorbitol–sucrose and Ficoll step introns found in the genomes of most other eukaryotes gradients. As determined by flow cytometry, fluores- [4]. Moreover, because it is vacuolate and the core cence microscopy, western blot analysis and marker machinery for protein delivery into and processing enzyme analyses, this procedure yields intact vacuoles within the vacuole is likely conserved in other vacuo- of high purity. lysosomal structures [3], investigations of S. cerevisiae The integrity of the final vacuole preparation was have the potential to contribute to our understanding assessed by comparison of its forward-scatter plots

4288 FEBS Journal 274 (2007) 4287–4305 ª 2007 The Authors Journal compilation ª 2007 FEBS J.-E. Sarry et al. Proteomics of yeast vacuolar lumen

A total signal. Whereas the unfractionated spheroplast preparation consists of two components, the smaller of which likely corresponds to vacuoles that have under- gone premature liberation from some of the sphero- plasts, the final vacuolar isolate has the properties of a monodisperse suspension of relatively uniformly sized particles characteristic of a reasonably homogeneous preparation. Maintenance of vacuolar integrity throughout the purification procedure, which is critical for definition of the luminal protein profile, was monitored using monochlorobimane as an indicator. Free, unsubstituted monochlorobimane is lipophilic, membrane permeant and nonfluorescent, but when S-conjugated with cyto- solic glutathione (GSH) to yield bimane–GS it assumes a hydrophilic, membrane-impermeant, fluorescent state. In this state, it is amenable to ATP-dependent vacuo- B lar accumulation by the GS-conjugate pump yeast cadmium factor 1, a vacuolar membrane ATP-binding cassette transporter [11]. Incubation and equilibration of viable yeast cells in media containing monochloro- bimane thereby enables visualization of the vacuolar compartment as the site of accumulation of bimane–GS by fluorescence microscopy [11]. Providing the medium is washed free of the probe before cell fractionation, maintenance of vacuolar integrity can be monitored as retention of the fluorescence associated with bimane– GS within this compartment throughout the isolation procedure. Using this approach it was determined that the vac- uoles in the final isolates had retained the bulk of their contents throughout the isolation protocol. Fluores- cence microscopy of cells to be fractionated after Fig. 1. Assessment of integrity of intact vacuoles purified from incubation in medium containing monochlorobimane S. cerevisiae SEY6210. (A) Flow cytometric forward scatter analysis revealed an intense fluorescence that localizes to the of samples of spheroplasts (inset) and intact vacuoles (main figure) vacuole of these and spheroplasts (Fig. 1B), which is from the same batch of cells as the spheroplasts and vacuoles retained in the intact vacuoles isolated from this shown in (B). Each of the forward scatter plots was derived from 5 · 105 events and is representative of two independent vacuole source (Fig. 1B). Because bimane–GS, a relatively low purifications and cytometric analyses. Numbers beside and above molecular mass-substituted tripeptide, is retained by ‘cell number’ peaks denote percentage of the total signal that is the vacuole fraction, it is reasonable to conclude that attributable to the peak indicated. (B) Fluorescence microscopy of there is also substantial retention of the luminal spheroplasts and isolated vacuoles from cells stained with mono- protein complement when vacuoles are isolated in this chlorobimane. The spheroplasts and vacuoles shown are from way. yeast cultures grown in YPD medium containing 150 lM monochlo- As expected of vacuoles that are relatively free of robimane for 18 h as described in Experimental procedures. contaminants, western analyses demonstrate marked enrichment of an intensely immunoreactive molecular with those of the spheroplasts from which it was mass 61 000 Da polypeptide species in the vacuolar derived by flow cytometry (Fig. 1A). In this way, it lysate versus whole-cell lysates after probing with mAb was determined that the spheroplast suspensions con- raised against carboxypeptidase Y (PRC1), a vacuolar sist of two refractive components corresponding to 18 lumen marker. There was no immunoreaction with any and 82% of the total forward scatter signal. By con- of the other antibodies tested with the exception of the trast, the final vacuole preparation consists of only one one raised against the cytosolic enzyme 3-phosphoglyc- major refractive component representing 89% of the erate kinase in some of the preparations (Fig. 2).

FEBS Journal 274 (2007) 4287–4305 ª 2007 The Authors Journal compilation ª 2007 FEBS 4289 Proteomics of yeast vacuolar lumen J.-E. Sarry et al.

vacuolar membrane and lumen markers are equivalent (Table 1) it is apparent that there is strict copurifica- tion of the membrane bounding this organelle and its contents, implying that those vacuoles which undergo disruption and release of their contents are excluded from the final isolate.

Resistance of vacuolar luminal proteins to proteinase K action Fig. 2. Western blot analyses of total lysates from intact cells (IC) and isolated vacuoles (vacuolar lumen, VL) from S. cerevisiae When attempting to define the vacuolar luminal pro- SEY6210. The blots were probed with monoclonal antibodies raised tein complement, it is important to guard against against PRC1 (61 000 Da), 3-phosphoglycerate kinase (45 000 Da), potential contamination by extravacuolar and surface- dolichol-P mannose synthase (30 000 Da), mitochondrial porin adherent proteins. This is an inevitable concern, which (30 000 Da) and the 100 000 Da intrinsic subunit of the vacuolar must be addressed by direct experimentation to maxi- + H -ATPase. All lanes were loaded with 10 lg protein and subjected mize the reliability of the protein database because of to SDS ⁄ PAGE and western blot analysis as described in Experi- the broad range of cytosolic and other proteins that mental procedures. The immunoreactive bands shown were the principal bands detected. undergo transport into this compartment. This issue was addressed by adopting a protease- protection procedure similar to that first used for Although mAbs raised against dolichol-P mannose in vitro yeast vacuolar protein import assays [16]. In synthase, mitochondrial porin and the 100 kDa sub- this procedure, proteins associated with the surface of unit of the vacuolar H+-ATPase (ER, outer mitochon- isolated vacuoles are digested by incubating the iso- drial membrane and vacuolar membrane markers, late in media containing an excess of the broad-range respectively) immunoreact with polypeptides of the serine protease, proteinase K, on ice for 20 min appropriate molecular mass from whole-cell lysates, no before terminating the reaction by the addition of the immunoreactive species are discernible in the vacuolar serine protease inhibitor phenylmethanesulfonyl fluo- lysates (Fig. 2). The presence of 3-phosphoglycerate ride. Thereafter, vacuoles are lysed and their contents kinase in the vacuolar lumen, albeit intermittently and subjected to 2-DE profiling to distinguish those pro- not in the preparation depicted in Fig. 2, is attribut- teins that are resistant to proteinase K action and able to the entry of this cytosolic protein into this likely have an intravacuolar localization from those compartment for breakdown, processing and ⁄ or sal- that undergo digestion and likely have an extravacuo- vage purposes. lar localization. Marker enzyme assays of the intact vacuole isolates Most, if not all, of the major polypeptide species for PRC1 and aminopeptidase I (APE1), two canoni- identified by 2-DE in the vacuolar luminal fraction are cal vacuolar luminal proteases [12,13] and a-mannosi- refractory to proteinase K action providing the prote- dase, a vacuolar membrane-localized oligosaccharide ase is added before disruption of the bounding vacuo- hydrolase [14,15], yield enrichment factors of 53-, 59- lar membrane (Fig. 3 and Table 2). The 2D gels of and 43-fold, respectively, compared with spheroplasts lysates from control vacuoles and proteinase K-pre- (Table 1). These values are commensurate with those treated vacuoles are virtually indistinguishable (Fig. 3 of the purest intact vacuole preparations reported pre- and Table 2). Regardless of their origin, charge or size, viously [16]. Because the enrichment factors for the the proteins identified are similarly represented in both

Table 1. Enrichment of vacuolar marker enzyme activities in intact vacuoles isolated from S. cerevisiae SEY6210. Intact vacuoles were puri- fied from spheroplasts as described in Experimental procedures and the total spheroplast and vacuolar lysates were assayed for a-mannosi- dase activity, a vacuolar membrane enzyme marker, and PRC1 and APE1 activity, two vacuolar lumen markers, as described by Jelinek-Kelly et al. [14], Stevens et al. [12] and Jones [13], respectively. Values shown are means ± SE for n ¼ 12 (a-mannosidase), n ¼ 11 (PRC1) or n ¼ 7 (APE1).

a-Mannosidase Specific Purification PRC1 Specific activity Purification APE1 Specific activity Purification Lysate activity (lmolÆmin)1Æmg)1) (-fold) (lmolÆmin)1Æmg)1) (-fold) (lmolÆmin)1Æmg)1) (-fold)

Spheroplast 0.015 ± 0.002 1.0 1.3 ± 0.1 1.0 0.010 ± 0.001 1.0 Vacuolar 0.649 ± 0.130 43.3 69.1 ± 6.6 53.2 0.593 ± 0.068 59.3

4290 FEBS Journal 274 (2007) 4287–4305 ª 2007 The Authors Journal compilation ª 2007 FEBS J.-E. Sarry et al. Proteomics of yeast vacuolar lumen

A ation and are not simply contaminants that have copu- rified with the preparation. By contrast, if proteinase K is added after vacuolar lysis, no protein spots except those corresponding to proteinase K and its autolysis products are discernible (data not shown), confirming that it is the intervening membrane that protects the bulk of the vacuolar luminal protein complement from digestion by this enzyme.

Identification of vacuolar luminal proteins Having established that intact vacuoles prepared in this way are of high purity and integrity, their luminal B protein profiles were determined. For this, prepara- tions were lysed and the soluble proteins released were separated by 2-DE. After visualization of the separated proteins with Coomassie Brilliant Blue, protein spots were excised for in-gel tryptic digestion and MS analy- sis. By this means, 430 spots were resolved of which 360 reproduced in all three independent replicates (Fig. 4) and 260 were identified unambiguously by MALDI-TOF-MS, LC-IT-MS ⁄ MS and QTOF- MS ⁄ MS through the deployment of MASCOT- and SEQUEST-based protein sequence database searches in combination with molecular mass and isoelectric point (pI) considerations (Fig. 4 and Table S1). The polypeptides identified fall into five broad cate- Fig. 3. 2-DE of soluble fraction from lysates of isolated intact gories: (a) canonical vacuolar and other proteases, S. cerevisiae SEY6210 vacuoles before (A) and after (B) pretreat- (b) other vacuolar proteins, (c) glycolytic ⁄ gluconeo- ment with proteinase K. For the pretreatments with proteinase K, genic , (d) other metabolic enzymes and intact vacuoles were incubated with the protease for 20 min on ice (e) other proteins (Table 3). With the exception of the before terminating the reaction with phenylmethanesulfonyl fluo- first two, these categories are largely populated by ride. For the pretreatment controls, intact vacuoles were treated in the same way but without the addition of proteinase K. After vacu- proteins and their breakdown products that have olar lysis, membrane sedimentation and precipitation of the soluble likely entered the vacuole for breakdown, processing luminal proteins, the pellets were resuspended in isoelectric focus- and ⁄ or salvage. ing loading buffer and aliquots (200 lg protein) were subjected to 2-DE as described in Experimental procedures. The gels were stained with colloidal Coomassie Brilliant Blue G-250 and scanned Canonical vacuolar and other proteases with an Epson flatbed scanner. Eleven proteases were identified in the vacuolar lumen, of which five – PRC1 [17], APE1 [18], aminopepti- dase Y (APE3) [19], proteinase A alias saccharopepsin control and proteinase K-pretreated vacuolar prepara- (PEP4, aspartate protease) [20], and vacuolar prote- tions. Both the canonical vacuolar proteases and repre- ase b (PRB1) [19] – are canonical luminal proteins sentative polypeptides belonging to all three of the (Table 3 and Table S1). The remaining six proteases other main protein categories identified below, with (‘other proteases’) identified, for which localization to the exception of APE1 and heat shock protein 60 the vacuole other than for salvage purposes is unlikely (HSP60), have percentage abundance ratios of between are: (a) bleomycin hydrolase (BLH1, alias LAP3, cyste- 0.82 and 1.12 when the luminal lysates from control ine protease), a nonessential aminopeptidase responsi- and proteinase K-pretreated vacuoles are compared ble for inactivation of the glycopeptide antibiotic directly (Table 2). This is as predicted if the bulk of bleomycin [21,22]; (b) a putative aspartyl aminopepti- the polypeptides subsequently identified have indeed dase (DNPE), which is competent in the release of entered the vacuolar lumen before subcellular fraction- N-terminal Asp or Glu from peptides with preference

FEBS Journal 274 (2007) 4287–4305 ª 2007 The Authors Journal compilation ª 2007 FEBS 4291 Proteomics of yeast vacuolar lumen J.-E. Sarry et al.

Table 2. Percentage abundances of representative spots on 2D gels of lysates from isolated intact vacuoles purified from S. cerevisiae SEY6120 before (– PK) and after (+ PK) pretreatment with proteinase K. The protease and control pretreatments were performed as described in the legend to Fig. 3 and Experimental procedures. The abundances were estimated from three independent vacuolar lysate preparations before and after treatment with proteinase K as described in the legend to Table S1 and Experimental procedures. The spots that were subjected to this analysis were selected to ensure that representatives from each of four of the main protein classes identified in Fig. 4, Table 3 and Table S1 were included.

Abundance – PK Abundance + PK (%) (%) Functional category PDQUEST

Gene Protein Access no. Mr pI Average SEM Average SEM Ratio

Vacuolar proteases APE1 Aminopeptidase I P14904 44.8 5.7 0.73 0.31 0.73 0.45 1.00 APE1 Aminopeptidase I P14904 45.3 5.8 0.63 0.15 0.56 0.10 0.88 APE1 Aminopeptidase I P14904 44.6 5.9 0.36 0.09 0.33 0.10 0.91 APE3 Aminopeptidase Y P37302 71.0 5.5 0.36 0.21 0.24 0.09 0.68 BHL1 Cysteine proteinase 1 Q01532 44.8 7.7 0.36 0.01 0.36 0.02 0.99 PEP4 Proteinase A P07267 37.9 4.3 3.14 0.26 2.98 0.02 0.95 PRB1 Vacuolar protease B P09232 27.1 5.9 1.24 0.19 1.21 0.26 0.98 PRB1 Vacuolar protease B P09232 27.6 6.2 0.48 0.20 0.40 0.18 0.82 PRB1 Vacuolar protease B P09232 26.7 6.4 0.71 0.24 0.74 0.33 1.05 PRC1 Carboxypeptidase Y P00729 49.2 4.5 3.44 0.49 3.16 0.09 0.92 Glycolytic ⁄ gluconeogenic enzymes ADH1 Alcohol dehydrogenase I P00330 37.2 6.6 0.98 0.18 0.88 0.03 0.90 ENO1 1 P00924 42.0 6.6 0.91 0.11 0.85 0.08 0.94 ENO2 P00925 40.3 5.8 0.63 0.04 0.64 0.05 1.03 FBA1 Fructose biphosphate aldolase P14540 35.3 5.8 0.87 0.15 0.79 0.09 0.91 PCK1 Phosphoenolpyruvate carboxykinase P10963 50.3 6.1 0.46 0.17 0.49 0.23 1.08 PGI1 Glucose 6-phosphate P12709 47.8 6.1 0.41 0.08 0.44 0.07 1.09 PGK1 Phosphoglycerate kinase P00560 40.2 8.2 0.63 0.09 0.69 0.03 1.08 PGM1 Phosphoglycerate mutase 1 P33401 25.3 9.3 0.37 0.08 0.41 0.02 1.12 PGM1 Phosphoglycerate mutase 1 P33401 25.8 9.2 0.48 0.11 0.52 0.11 1.09 TPI1 Triose-phosphate isomerase P00942 23.7 5.8 0.63 0.03 0.63 0.04 1.00 Other metabolic enzymes CPR1 Peptidylprolyl isomerase P14832 11.6 6.8 0.56 0.10 0.52 0.11 0.93 EFT1 ⁄ 2 Elongation factor P32324 32.5 5.5 0.27 0.05 0.30 0.04 1.10 ICL1 Isocitrate 1 P28240 51.1 6.2 0.87 0.09 0.83 0.07 0.95 ILV2 Acetolactate dehydrogenase P07342 73.2 6.8 0.28 0.05 0.27 0.07 0.98 ILV2 Acetolactate dehydrogenase P07342 73.9 6.9 0.28 0.07 0.26 0.08 0.94 ILV3 Dihydroxy-acid P39522 51.5 6.9 0.30 0.08 0.28 0.10 0.94 MLS1 Malate synthase P30952 51.2 7.4 0.76 0.13 0.75 0.18 0.99 Other proteins HSP60 Heat shock protein 60 P19882 51.1 5.0 0.47 0.33 0.31 0.22 0.65 SOD1 Superoxide dismutase P00445 14.1 5.7 0.52 0.10 0.46 0.03 0.89 SSA1 Heat shock protein SSA1 P10591 38.3 5.2 0.29 0.05 0.26 0.02 0.90 SSA1 Heat shock protein SSA1 P10591 39.5 5.1 0.29 0.13 0.32 0.17 1.11 SSA2 Heat shock protein SSA2 P10592 51.6 5.2 0.42 0.24 0.40 0.34 0.97 SSC1 Heat shock protein SSC1 P12398 77.4 5.1 0.45 0.08 0.44 0.11 0.97

for Asp [23]; (c) a putative Xaa-Pro aminopeptidase YDR415c), an M28 peptidase family member of (XPPE), thought to cleave N-terminal amino acids, unknown function [27] (Fig. 4, Table 3 and Table S1). including Pro, from peptides when the penultimate res- Based on their staining reaction with Coomassie idue is Pro [24]; (d) subtilisin-like protease 3 (YSP3), a Brilliant Blue and the quantitative pdquest analyses, serine protease of unknown function [25]; (e) a proba- the canonical proteases represent some of the most ble peptidase (encoded by YFR006w), a member of abundant vacuolar proteins identified. PRC1, PEP4 the M24B peptidase family of unknown function [26]; and PRB1, in that order, are among the most abun- and (f) a probable aminopeptidase (encoded by dant luminal polypeptide species, followed by APE1

4292 FEBS Journal 274 (2007) 4287–4305 ª 2007 The Authors Journal compilation ª 2007 FEBS J.-E. Sarry et al. Proteomics of yeast vacuolar lumen

Fig. 4. Annotated 2-DE map of soluble vacuolar luminal proteins from S. cerevisiae SEY6210. The vacuolar lysates were prepared and aliquots (200 lg protein) of the soluble fraction from the lysate were subjected to 2-DE as described in the legend to Fig. 3 and Experimental procedures. The image shown is the PDQUEST-generated matchset image based on three replicate gels. The Coomassie Brilliant Blue-stained protein spots were excised and subjected to tryptic digestion followed by protein identification by MALDI-TOF MS, LC-IT-MS ⁄ MS and QTOF-MS ⁄ MS as described in Experimental procedures. Refer to Table 3 and Table S1 for definition of protein acronyms and criteria of identity. and APE3 (Fig. 4 and Table S1). By comparison, all Other canonical vacuolar proteins of the other proteases identified are rarer (Fig. 4 and Table S1). All five of the canonical proteases, PRC1, Canonical vacuolar proteins other than proteases PEP4, PRB1, APE1 and APE3, are annotated as vacu- include the vacuolar marker a-mannosidase (AMS1), olar proteins in SwissProt and the Saccharomyces acid trehalase (ATH1), a nonspecific repressible alkaline Genome Database (SGD) but only two, PRC1 and phosphatase (PHO8), a GTP-binding protein (YPT7), a PEP4, are listed as having a vacuolar localization in phosphatidyglycerol ⁄ phosphatidylinositol transfer pro- the Yeast GFP Fusion Localization Database (yeast- tein (NPC2), a vacuolar transporter chaperone (VTC4), gfp); the other four are listed as having a cytoplas- and a hypothetical 69.0 kDa protein (encoded by mic ⁄ punctate composite localization. YHR202w) (Fig. 4, Table 3 and Table S1). In terms of Some investigators contend that BLH1 is a cytosolic their novelty and potential for contributing to a more enzyme [22] and others propose that it has a dual thorough understanding of both yeast vacuolar func- localization in the cytosol and mitochondria [28], tion and animal lysosomal function, YPT7, NPC2 and whereas SwissProt, SGD and yeastgfp list it as mito- VTC4 are of special interest. chondrial and both cytosolic and mitochondrial, YPT7 is implicated in the regulation of membrane respectively (Table S1). All of the other noncanonical fusion and plays a role in determining the size of this proteases, with the exception of YSP3 and YDR415c compartment and its capacity to contribute to cellular for which there are no localization data, are tentatively pH homeostasis. It is speculated that the proteasomal listed as cytosolic enzymes in the SwissProt, SGD degradation of ubiquitinated YPT7 is necessary and ⁄ or yeastgfp databases (Table 3 and Table S1). for membrane fusion [29]. NPC2 is a functionally

FEBS Journal 274 (2007) 4287–4305 ª 2007 The Authors Journal compilation ª 2007 FEBS 4293 Proteomics of yeast vacuolar lumen J.-E. Sarry et al.

Table 3. Unique ORFs identified by 2-DE separation and MS analysis of proteins extracted from the lumen of intact vacuoles purified from S. cerevisiae SEY6210 and grouped according to functional category. The protein acronyms and locus numbers listed are based on annota- tions in SGD (http://www.yeastgenome.org) and grouped according to functional category together with their corresponding SwissProt accession numbers. The ‘localizations’ shown are those listed in the GFP (http://yeastgfp.ucsf.edu), SwissProt (http://us.expasy.org/sprot/) and SGD, respectively. cw, cell wall; cyt, cytoplasm; gly, glyoxysome; mit, mitochondrion; nuc, nucleus; per, peroxisome; pun, punctuate composite; vac, vacuole. The GFP ascription ‘punctate composite’ is intended to indicate that the distribution of the GFP-tagged protein is punctate but, without the results from colocalization experiments with a second marker, it is not possible to determine the structure or structures with which the protein is most closely associated. Refer to Table S1 for complete compilation inclusive of multiple spots for the same parent species, experimental MW and pI, percentage abundances, and criteria of identity.

Localization Theoretical Functional category Access No.

Gene Protein Locus no. GFP SwissProt SGD Mr pI spots

Canonical vacuolar and other proteases APE1 Aminopeptidase I KL103c P14904 cyt, pun vac vac 57.1 5.6 7 APE3 Aminopeptidase Y YBR286w P37302 – vac vac 60.1 5.1 19 PEP4 Proteinase A YPL154c P07267 vac vac vac, mit 44.5 4.7 5 PRB1 Vacuolar protease B YEL060c P09232 cyt, pun vac vac 69.6 6.3 14 PRC1 Carboxypeptidase Y YMR297w P00729 vac vac vac 59.8 4.6 4 BHL1 Cysteine proteinase 1 YNL239w Q01532 cyt, mit mit cyt, mit 52.1 7.6 5 DNPE Putative aspartyl aminopeptidase YHR113w P38821 cyt cyt cyt 54.2 6.4 2 XPPE Putative Xaa-Pro aminopeptidase YLL029w Q07825 cyt cyt cyt 84.9 6.4 2 YSP3 Subtilisin-like protease 3 YOR003w P25036 – – – 52.1 5.3 1 YFR006w Probable peptidase YFR006w P43590 cyt – cyt 61.8 5.8 1 YDR415c Probable aminopeptidase YDR415c Q04033 – – – 42.5 5.5 1 Other canonical vacuolar proteins AMS1 Alpha-mannosidase YGL156w P22855 pun vac vac 124.5 6.8 1 ATH1 Acid trehalase YPR026w P48016 – vac cw, vac 136.9 5.2 1 PHO8 Alkaline phosphatase YDR481c P11491 vac vac vac 63.0 5.3 3 VTC4 Vacuolar transporter chaperone 4 YJL012c P47075 – vac vac 83.2 6.4 2 YPT7 GTP-binding protein YPT7 YML001w P32939 vac, cyt vac vac 23.0 4.8 2 NPC2 Phosphatidylglycerol ⁄ phosphatidylinositol YDL046w Q12408 vac vac vac 19.1 4.4 1 transfer protein YHR202w Hypothetical 69.0 kDa protein YHR202w P38887 vac – vac 69.0 5.8 2 Glycolytic ⁄ gluconeogenic enzymes ADH1 Alcohol dehydrogenase I YOL086c P00330 – cyt cyt 36.7 6.3 8 ADH2 Alcohol dehydrogenase II YMR303c P00331 cyt cyt cyt 36.7 6.3 2 ALD6 Aldehyde dehydrogenase 6 YPL061w U56604 cyt cyt cyt 54.3 5.3 1 ENO1 Enolase 1 YGR254w P00924 cyt cyt cyt 46.6 6.2 2 ENO2 Enolase 2 YHR174w P00925 cyt cyt cyt 46.8 5.7 4 FBA1 Fructose biphosphate aldolase YKL060c P14540 cyt cyt cyt 39.6 5.7 5 FBP1 Fructose-1,6-bisphosphatase YLR377c P09201 – – cyt 38.3 5.6 1 GLK1 Glucokinase YCL040w P17709 cyt, pun – cyt 55.3 5.8 2 PGI1 Glucose 6-phosphate isomerase YBR196c P12709 cyt cyt cyt 61.1 6.0 2 TDH1 Glyceraldehyde 3-phosphate YJL052w P00360 cyt, nuc cyt cyt, cw 35.7 8.3 3 dehydrogenase 1 TDH2 Glyceraldehyde 3-phosphate YJR009c P00358 cyt, nuc cyt cw, cyt 35.8 6.5 6 dehydrogenase 2 TDH3 Glyceraldehyde 3-phosphate YGR192c P00359 cyt, nuc cyt cw, cyt 35.7 6.5 7 dehydrogenase 3 HXK1 Hexokinase 1 YFR053c P04806 cyt cyt cyt 53.7 5.3 3 HXK2 Hexokinase 2 YGL253w P04807 cyt cyt, nuc cyt, nuc 53.8 5.2 2 CYB2 L-Lactate dehydrogenase YML054c P00175 – mit mit 65.5 8.6 1 PCK1 Phosphoenolpyruvate carboxykinase YKR097w P10963 – – cyt 61.0 5.9 3 PGK1 Phosphoglycerate kinase YCR012w P00560 cyt, nuc cyt cyt 44.8 7.7 5 PGM1 Phosphoglycerate mutase 1 YKL127w P33401 cyt cyt cyt 26.4 8.3 2 PGM2 Phosphoglycerate mutase 2 YMR105c P37012 cyt, nuc cyt cyt 63.0 6.2 1 PDC1 Pyruvate decarboxylase 1 YLR044c P06169 cyt, nuc cyt, nuc cyt, nuc 61.4 5.8 5 PYK1 Pyruvate kinase 1 YAL038w P00549 cyt – cyt 54.6 8.0 4 TKL1 Transketolase 1 YPR074c P23254 cyt, nuc cyt cyt 73.8 6.5 1

4294 FEBS Journal 274 (2007) 4287–4305 ª 2007 The Authors Journal compilation ª 2007 FEBS J.-E. Sarry et al. Proteomics of yeast vacuolar lumen

Table 3. (Continued).

Localization Theoretical Functional category Access No.

Gene Protein Locus no. GFP SwissProt SGD Mr pI spots

TPI1 Triosephosphate isomerase YDR050c P00942 cyt, nuc – cyt 26.6 5.8 4 UGP1 UDP-glucose pyrophosphorylase YKL035w Q9P966 cyt cyt cyt 56.0 7.0 1 Other metabolic enzymes ACO1 1 YLR304c P19414 cyt, mit cyt, mit cyt, mit 85.3 8.2 3 ACS2 Acetyl-CoA synthetase 1 YLR153c P52910 nuc cyt cyt 75.5 6.2 1 KGD1 Lipoamide oxoglutarate dehydrogenase YIL125w P20967 mit mit mit 114.3 6.8 2 PDB1 Pyruvate dehydrogenase E1 component YBR221c P32473 mit mit mit 40.1 5.2 1 subunit beta SDH1 Succinate dehydrogenase ubiquinone YKL148c Q00711 – mit – 70.2 7.1 1 GDH1 Glutamate dehydrogenase 1 YOR375c P07262 cyt, nuc cyt, nuc cyt, nuc 49.6 5.6 1 ILV2 Acetolactate dehydrogenase YMR108w P07342 mit mit mit 74.9 8.6 2 ILV3 Dihydroxy-acid dehydratase YJR016c P39522 mit mit mit 62.8 7.9 1 ILV5 Ketol-acid reductoisomerase YLR355c P06168 cyt, nuc mit mit 44.3 9.1 1 MET6 Methionine synthase YER091c P05694 cyt – cyt 85.9 6.2 2 MET17 O-acetylhomoserine sulfhydrylase YLR303w P06106 – cyt cyt 48.6 6.0 2 SAM1 S-adenosylmethionine synthetase 1 YLR180w P10659 cyt – cyt 41.8 5.0 1 SER1 Phosphoserine aminotransferase YOR184w P33330 cyt cyt cyt 43.4 6.1 1 SHM2 Serine hydroxymethyltransferase 2 YLR058c P37291 cyt, nuc cyt cyt 52.2 7.0 2 ACH1 Acetyl-coA hydrolase YBL015w P32316 mit cyt mit 58.7 6.3 1 GND1 6-Phosphogluconate dehydrogenase 1 YHR183w P38720 cyt mit cyt, mit 49.6 5.6 1 ICL1 Isocitrate lyase 1 YER065c P28240 – cyt – 62.4 6.0 1 MDH2 Malate dehydrogenase 2 YOL126c P22133 per, pun cyt cyt 40.7 6.4 1 MLS1 Malate synthase YNL117w P30952 – gly cyt, per 62.8 6.7 1 ADE13 Adenylosuccinate lyase YLR359w Q05911 mit – – 54.5 6.0 1 ADO1 Adenosine kinase putative YJR105w P47143 cyt, nuc cyt, nuc cyt, nuc 36.4 5.0 2 PRO3 Pyrroline-5-carboxylate reductase YER023w P32263 cyt – cyt 30.1 5.4 1 AAT2 Aspartate aminotransferase 2 YLR027c P23542 cyt cyt cyt, per 46.1 8.5 1 CAR1 Arginase YPL111w P00812 cyt, nuc – – 35.7 5.4 1 CAR2 Ornithine aminotransferase YLR438w P07991 cyt, nuc cyt cyt, nuc 46.1 6.5 1 URA1 Dihydroorotate dehydrogenase YKL216w P28272 cyt, nuc cyt cyt 34.8 5.8 1 AEP3 ATPase expression protein 3 YPL005w Q12089 mit mit mit 70.3 9.8 2 DPS1 Aspartyl-tRNA synthetase YLL018c P04802 cyt cyt cyt 63.4 6.2 1 EFT1 ⁄ 2 Elongation factor YOR133w P32324 cyt cyt cyt 93.2 5.9 2 YEF3 Elongation factor 3A YLR249w P16521 cyt cyt cyt 116.0 5.7 1 ANB1 Eukaryotic translation initiation factor 5A-1 YJR047c P19211 cyt cyt cyt 17.1 4.8 1 FPR1 Peptidylprolyl cis–trans isomerase YNL135c P20081 cyt, nuc cyt cyt, nuc 12.2 5.7 1 CPR1 Peptidylprolyl isomerase YDR155c P14832 cyt, nuc cyt nuc 17.3 6.9 2 RNR4 Ribonucleoside-diphosphate reductase YGR180c P49723 nuc, cyt nuc nuc, cyt 40.1 5.1 1 small chain 2 IPP1 Inorganic pyrophosphatase YBR011c P00817 cyt, nuc cyt cyt 32.3 5.2 1 PLB1 Lysophospholipase 1 YMR008c P39105 – – cw 71.7 4.7 1 Other proteins ACT1 Actin YFL039c P60010 – cyt – 41.7 5.4 1 ASC1 Guanine nucleotide-binding protein YMR116c P38011 cyt cyt cyt 34.8 5.8 1 subunit beta-like protein RAS2 Ras-like protein 2 YNL098c P01120 nuc, cyt PM PM 34.7 6.8 1 SEC4 Ras-related protein SEC4 YFL005w P07560 – cyt, PM PM 23.5 6.6 1 YPT1 GTP-binding protein YPT1 YFL038c P01123 – ER, cyt ER, mito 23.2 5.2 2 BGL2 Glucan 1,3- b-glucosidase YGR282c P15703 vac cw cw 34.1 4.3 3 KRE6 Beta-glucan synthesis-associated protein YPR159w P32486 vac golgi golgi 80.1 4.5 2 CRH1 Probable glycosidase CRH1 YGR189c P53301 cyt cw cw 52.8 4.5 1 EXG2 Glucan 1,3-b-glucosidase 2 YDR261c P52911 – PM cw 63.5 5.2 1 YGP1 Protein YGP1 YNL160w P38616 – cw cw 37.3 5.3 2 GAS5 Glycolipid-anchored surface protein 5 YOL030w Q08193 nuc cw cw 51.8 4.5 2

FEBS Journal 274 (2007) 4287–4305 ª 2007 The Authors Journal compilation ª 2007 FEBS 4295 Proteomics of yeast vacuolar lumen J.-E. Sarry et al.

Table 3. (Continued).

Localization Theoretical Functional category Access No.

Gene Protein Locus no. GFP SwissProt SGD Mr pI spots

CDC3 Cell division control protein 3 YLR314c P32457 – PM PM 60.0 5.3 2 ECM33 Protein ECM33 YBR078w P38248 – cw mito, cw 48.4 4.8 3 PST2 Protoplast secreted protein 2 YDR032c Q12335 pun cw mit 21.0 5.5 1 RHO1 GTP-binding protein RHO1 YPR165w P06780 – per, mito per, mito 23.2 6.0 5 YHB1 Flavohemoprotein YGR234w P39676 cyt cyt, mit cyt, mit 44.6 5.9 1 HSP31 Chaperone HSP31 YDR533c Q04432 cyt, nuc – – 25.7 5.3 1 HSP60 Heat shock protein 60 YLR259c P19882 mit mit mit 60.7 5.2 1 HSP78 Heat shock protein 78 YDR258c P33416 mit mit mit 91.3 8.2 1 HSP150 150 kDa heat shock YJL159w P32478 – cw cw 41.1 5.o 1 glycoprotein SSA1 Heat shock protein SSA1 YAL005c P10591 cyt cyt cyt, cw, 37.4 5.2 5 nuc, vac SSA2 Heat shock protein SSA2 YLL024c P10592 cyt cyt cyt, cw, 69.3 5.0 4 vac SSA3 Heat shock protein SSA3 YBL075c P09435 cyt cyt cyt 70.5 5.1 1 SSA4 Heat shock protein SSA4 YER103w P22202 cyt cyt cyt, nuc 69.7 5.0 1 SSB1 Heat shock protein SSB1 YDL229w P11484 cyt cyt cyt 66.4 5.3 1 SSB2 Heat shock protein SSB2 YNL209w P40150 cyt cyt cyt 66.5 5.4 1 SSC1 Heat shock protein SSC1 YJR045c P12398 – mit mit 70.6 5.5 1 SSE1 Heat shock protein homolog YPL106c P32589 cyt cyt cyt 77.4 5.1 1 SOD1 Superoxide dismutase YJR104c P00445 cyt, nuc cyt cyt, mit 15.8 5.6 2 TSA1 Thioredoxin peroxidase 1 YML028w P34760 cyt cyt cyt 21.4 5.0 2 CTT1 Catalase T YGR088w P06115 cyt cyt cyt 65.7 6.2 1 GTT1 Glutathione S- I YIR038c P40582 – ER ER, mit 26.8 6.2 1 MCR1 Cytochrome b5 reductase YKL150w P36060 mit mit mit 34.1 8.7 1 NCP1 NADPH-cytochrome P450 YHR042w P16603 ER ER mit 76.8 5.1 1 reductase YDL124w a-Keto amide reductase, putative YDL124w Q07551 cyt, nuc cyt, nuc cyt, nuc 35.5 5.8 1 YNL134c Alcohol dehydrogenase, putative YNL134c P53912 cyt, nuc cyt, nuc cyt, nuc 41.2 5.8 2 YLR179c Uncharacterized protein YLR179c Q06252 – – cyt, nuc 22.2 4.8 1 YHR029c Uncharacterized isomerase YHR029c P38765 – – – 32.6 5.5 1 YER004w Uncharacterized protein YER004w P40008 ER ER, mit ER, mit 25.1 9.3 1 conserved homolog of the human lipid-binding lyso- tion of the nonreducing dissacharide, trehalose, as sole somal glycoprotein NPC2, mutation of which causes carbon source [32]; PHO8, an enzyme involved in the fatal congenital neurovisceral disease Niemann– phosphate-regulated organic phosphorus mobilization Pick disease type C, characterized by the lysosomal [33]; and the hypothetical 69.0 kDa protein about accumulation of cholesterol and its derivatives [30]. which nothing is known at the time of writing. VTC4 which predominantly localizes to the vacuole but is also found on the ER and at the cell periphery Glycolytic/gluconeogenic enzymes is a constituent of the vacuolar transporter chaperone complex which is recruited to and concentrated by the Among the glycolytic ⁄ gluconeogenic enzymes identified vacuole during nutrient limitation. Deletion of the in the vacuolar lumen, most of which have been shown encoding components of the vacuolar transporter to enter the vacuole for salvage purposes [16,34–38], chaperone abolishes vacuolar microautophagy [31]. are: aldehyde dehydrogenase (ALD6), alcohol dehy- Without exception, these polypeptides are listed as drogenase (ADH1, ADH2), enolase (ENO1, ENO2), having a vacuolar localization in at least two of the fructose bisphosphate aldolase (FBA1), fructose- SwissProt, SGD and yeastgfp databases (Table 3 and 1,6-bisphosphatase (FBP1), glucokinase (GLK1), Table S1). The same applies to the longer known glucose-6-phosphate isomerase (GPI1), glyceraldehyde- members of this category: AMS1, a key enzyme for 3-phosphate dehydrogenase (TDH1, TDH2, TDH3), oligosaccharide processing [14]; ATH1, a glycosidase hexokinase (HXK1, HXK2), lactate dehydrogenase with an acid pH optimum which is essential for utiliza- (CYB2), phosphoglycerate mutase (PGM1, PGM2),

4296 FEBS Journal 274 (2007) 4287–4305 ª 2007 The Authors Journal compilation ª 2007 FEBS J.-E. Sarry et al. Proteomics of yeast vacuolar lumen phosphoglycerate kinase (PGK1), phosphoenolpyruvate mitochondria, glyoxysomes, peroxisomes and possibly carboxykinase (PCK1), pyruvate decarboxylase the nucleus, as judged by their known sites of action (PDC1), pyruvate kinase (PYK1), transketolase and localizations as listed in the SwissProt, SGD and (TKL1), triose phosphate isomerase (TPI1) and UDP- yeastgfp databases (Table 3 and Table S1), also make glucose pyrophosphorylase (UGP1) (Fig. 4, Table 3 a large contribution to the luminal proteome. and Table S1). The most abundant of these polypep- Although the possibility of coalescence of these organ- tide species in the luminal fraction are TDH2 and ⁄ or elles with the vacuole during spheroplast lysis or sub- TDH3 and ADH1 and ⁄ or ADH2, closely followed cellular fractionation cannot be discounted completely, by ENO2, FBA1, PGK1 and PGM1 (Fig. 4 and the most likely explanation for the presence of their Table S1). All of these polypeptides are listed as cyto- polypeptide constituents in the vacuolar lumen, is solic or cytosolic and nuclear species in the SwissProt, uptake by autophagy in the intact cell before cell wall SGD and yeastgfp databases (Table 3 and Table S1). digestion and fractionation.

Other metabolic enzymes Other proteins The other metabolic enzymes identified in the vacuolar The ‘other proteins’ identified in the vacuolar lumen, lumen (Fig. 4, Table 3 and Table S1), whose delivery a heterogeneous category, encompasses a cytoskeletal to this compartment has largely been neglected by protein, several proteins involved in intracellular tar- comparison with glycolytic ⁄ gluconeogenic enzymes, are geting, membrane assembly, cell division and cell wall those associated with: (a) the (ACO1, biogenesis, a flavohemoprotein, several HSPs and sev- aconitase; ACS2, acetyl-CoA synthetase; KGD1, lipo- eral enzymes implicated in the alleviation of oxidative amide oxoglutarate dehydrogenase; PDB1, lipoamide stress and related detoxification processes (Fig. 4, pyruvate dehydrogenase; SDH1, succinate dehydroge- Table 3 and Table S1). nase flavoprotein subunit); (b) amino acid metabolism The sole cytoskeletal protein identified is actin (AAT2, cytoplasmic aspartate aminotransferase; (ACT1), a protein that is abundant at the whole-cell GDH1, glutamate dehydrogenase; ILV2, acetolactate level and also associated with the surface of purified synthase; ILV3, dihydroxy-acid dehydratase; ILV5, vacuoles where it is considered to participate in mem- ketol-acid reductoisomerase; MET6, methionine syn- brane fusion [39]. Among the proteins implicated in thase; MET17, O-acetyl-homoserine sulfhydrylase; membrane trafficking, cell division and membrane and PRO3, D-1-pyrroline-5-carboxylate reductase; SAM1, wall fabrication are: (a) four GTPases (ASC1, RAS2, S-adenosylmethionine synthetase; SER1, phospho- SEC4 and YPT1) that likely participate in the secretory serine aminotransferase; SHM2, serine hydroxy- and endocytotic pathways [40]; (b) a glycosidase methyltransferase); (c) fatty acid oxidation (ACH1, (CRH1), a glucan 1,3-b-glucosidase (EXG2), a glucan- acetyl-CoA hydrolase); (d) the pentose phosphate shunt ase (BGL2), and a b-glucan synthesis associated protein (GND1, 6-phosphogluconate dehydrogenase); (e) the (KRE6) that have been established to contribute to glyoxylate cycle (ICL1, isocitrate lyase; MLS1, malate glucan biosynthesis and breakdown; (c) two glycolipid- synthase; MDH2, malate dehydrogenase, cytoplasmic anchored proteins, a glycolipid-anchored surface isoform); (f) nucleotide metabolism (ADE13, adenylo- protein (GSA5) and a glycosylphosphatidylinositol- succinate lyase; ADO1, adenosine kinase; CAR1, argi- anchored protein (ECM3); (d) a protoplast secreted nase; CAR2, ornithine aminotransferase; URA1, protein (PST2), a protein of unknown function induced dihydroorotate dehydrogenase); (g) RNA and protein by oxidative stress; (e) a cell-division control protein synthesis (AEP3, ATPase expression protein; ANB1 (CDC3), a component of the septin ring required for (alias HYP1), hypusine-containing protein, eukaryotic cell division [41]; and (f) a GTP-binding protein translation initiation factor 5 A-1; DPS1, aspartyl- (RHO1) that serves to coordinate b-glucan synthesis tRNA synthetase; YEF3 and EFT1 ⁄ 2, translation and protein and secretory vesicle-targeting to the elongation factors; CPR1 and FPR1, peptidyl-prolyl growth site during cell wall remodeling in parallel with cis-trans ; RNR4, ribonucleotide-diphos- organization of the actin cytoskeleton [42]. The major- phate reductase); and (h) phosphate and phospholipid ity of these proteins localize to structures other than metabolism (IPP1, inorganic pyrophosphatase; PLB1, the vacuole as GFP fusions, with the exception of lysophospholipase). It is apparent that although a BGL2 and KRE6 which are assigned a vacuolar locali- large fraction of the proteins destined for delivery to zation in the yeastgfp database (Table 3 and Table S1). and degradation in the vacuole are of cytosolic origin, YHB1, the one flavohemoprotein identified, is an the polypeptides from other organelles, for instance intermediate abundance vacuolar lumen component

FEBS Journal 274 (2007) 4287–4305 ª 2007 The Authors Journal compilation ª 2007 FEBS 4297 Proteomics of yeast vacuolar lumen J.-E. Sarry et al.

(Fig. 4, Table 3 and Table S1) which likely initially within the vacuole. Of the 260 polypeptide species localizes to the cytosol and ⁄ or mitochondria based on identified, which correspond to a total of 117 unique its annotation in the various yeast protein databases ORFs, only 64 migrate, or are at least identified, as sin- (Table 3 and Table S1). Very little is known of the gle spots. The remaining 196 either appear as horizontal function of YHB1, despite its relatively high abun- trains of spots differing only in pI, consistent with dance at the whole cell level and identity as a flavopro- protein-charge modification through acetylation, phos- tein possessing both heme and flavin binding domains phorylation and ⁄ or minor proteolysis, or as spots associated with the capacity to transfer electrons from staggered with respect to both pI and molecular mass, NADPH to heme iron, other than it might contribute consistent with the initiation of intravacuolar degrada- to the alleviation of nitrosative stress by catalyzing the tion or a combination of degradation and other forms oxidative decomposition of nitric oxide [43]. of covalent modification (Fig. 4, Table 3 and Table S1). Heat shock proteins, as exemplified by HSP31, A majority of the spots identified, 86 of a total of 260 HSP60, HSP78, HSP150, SSA1, SSA2, SSA3, SSA4, migrate at molecular masses lower than expected from SSB1, SSB2, SSC1 and SSE1, constitute a significant the computed masses of the translation products of fraction of the vacuolar luminal proteome (Fig. 4, their corresponding ORFs, which is what would be Table 3 and Table S1). Cytosolic chaperones of the expected of the bulk of the polypeptide species in a lytic HSP70 family participate in the vacuolar uptake of compartment. Examples of exceptions to this pattern cytosolic enzymes, for instance, the major luminal are ACO1, ADH1, ADH2, TDH1, TDH2, TDH3, component glyceraldehyde 3-phosphate dehydrogenase PGM1, SDH1, TKL1, ILV2, ACH1, HYPO1, MET6, (TDH isoforms) [16], whereas others, for example HSP31, SSA1, SSA2, SSC1 and EFT1 ⁄ 2 whose theoret- SSA1, are required for targeting of certain vacuolar ical and experimental molecular masses coincide despite proteases, APE1 in this case [44]. speciation in the IEF dimension (Fig. 4, Table 3 and The four intermediate abundance luminal polypep- Table S1), and the canonical vacuolar proteases which tides identified that participate in the alleviation of oxi- have precisely the migration properties that would be dative stress and related processes are superoxide predicted from the results of previous, targeted molecu- dismutase (SOD1), thioredoxin peroxidase (TSA1), lar investigations (Fig. 4 and Table S1). APE1 has an catalase T (CTT1) and glutathione S-transferase apparent molecular mass of 45 kDa consistent with its (GTT1), enzymes that catalyze the dismutation of cleavage from a 57 kDa precursor to the mature form superoxide radicals, the reduction of organic hydroper- by PRB1 [47]. APE3 migrates at an apparent mass of oxides, the decomposition of H2O2, and the conjuga- 71, 94, 114, 146 or 176 kDa, versus a mass of 60 kDa tion of cytotoxic electrophiles, respectively. predicted from its full-length ORF, consistent with its Three of the seven relatively minor luminal compo- susceptibility to Asn-glycosylation at several positions nents of this category that were identified, are the during its migration through the ER [48,49]. PRB1 translation products of ORFs YDL124w, YLR179c migrates at an apparent mass of 27–28 kDa consistent and YNL134c which have been tentatively identified as with its derivation from a 70 kDa precursor through hypothetical proteins with a-keto amide reductase the removal of a 20 amino acid signal peptide in the ER activity [45] and as proteins bearing a resemblance to and a large N-terminal propeptide of about 260 amino phosphatidylethanolamine binding and alcohol dehy- acid residues by proteinases A and B in the vacuolar drogenase proteins from other model organisms lumen [50,51]. APE3 migrates at an apparent mass of (SGD), respectively. The remaining four proteins are 61 kDa consistent with the proteinase A and protein- two about which nothing is known, YER004w and ase B-mediated removal of an N-terminal stretch of 91 YHR029c, and two others, NCP1, an ER-localized amino acid residues from the multiply Asn-glycosylated NADPH-cytochrome P450 reductase and MCR1, cyto- 69 kDa species [52–55]. Proteinase A migrates at an chrome b5 reductase, a nuclearly encoded mitochon- apparent mass of 38 kDa, the size predicted for the drial protein implicated in antimycin-insensitive active mature species which is subject to only limited NADH oxidation [46]. glycosylation and autocatalytic and proteinase B-medi- ated truncation [56,57]. Multiple spots, protein maturation, modification and degradation Concluding remarks Many of the spots identified correspond to alternate The primary objectives of these investigations were to isoelectric charge or size states of the translation prod- establish a methodological platform for the fraction- ucts, presumably because of processing and degradation ation of intact vacuoles from yeast, to apply a broad

4298 FEBS Journal 274 (2007) 4287–4305 ª 2007 The Authors Journal compilation ª 2007 FEBS J.-E. Sarry et al. Proteomics of yeast vacuolar lumen range protein separation technique to the soluble com- ment, the vegetative vacuoles isolated from protoplasts ponents, and to identify a sizeable fraction of protein derived from mature plant tissues such as Arabidopsis species in the vacuole. The overall goal was to not only rosette leaves serve both a lytic and protein storage provide information on subcellular localization, but function. Unlike yeast, plant cells contain at least two also a means for the enhanced detection and identifica- types of vacuoles: so called lytic vacuoles and protein tion of lower abundant proteins in yeast cells through storage vacuoles [58]. Lytic vacuoles are considered to organelle purification. If the function and ⁄ or functional be yeast vacuole and mammalian lysosome analogs; context of proteins, especially those that have not been protein storage vacuoles are considered to be the main identified experimentally before, are to be elucidated, sites of protein storage. The large central vacuoles of knowledge of the intracellular organelle or organelles mature tissues such as those subjected to the proteomic with which they are associated is necessary. analyses presented by Carter et al. [7] arise from the All three of these objectives have largely been ful- fusion of lytic vacuoles and protein storage vacuoles filled. Vacuoles of high purity and high integrity have [59]. In short, although the proteome of the yeast vac- been obtained, as judged by the enrichment factors uole lumen is very much biased in favor of the process- achieved, their light-scattering properties and retention ing and degradation of extravacuolar proteins, the of the relatively low molecular mass membrane-imper- majority of which participate in intermediary metabo- meant fluorescent indicator, bimane–GS throughout the lism, the proteome of the plant vegetative vacuole fractionation procedure. Moreover, most of the 360 lumen, though active in these lytic functions, is also spots separated reproducibly on the 2D gels, of which richly populated by proteins that participate in a wide > 70% have been identified with high confidence range of other processes including protein storage. (Table S1), correspond to canonical vacuolar luminal With the benefit of the insights gained from our analy- proteins or proteins that have entered this compartment ses of a predominantly lytic vacuolar compartment, in for breakdown, processing and ⁄ or salvage purposes. which organellar purity and integrity were overriding The principal conclusion to come from this system- concerns, it is probable that although many of the atic global approach is that the yeast vacuole is pre- extravacuolar proteins identified in the Arabidopsis dominantly a lysosomal compartment populated by vacuole proteome were listed as potential contaminants high-abundance canonical vacuolar proteases plus sev- [7] they are in fact targets destined for vacuolar pro- eral other luminal and factors that contrib- cessing or degradation. ute to cellular turnover, and a broad range of proteins In conducting these investigations, we have gener- from most of the other intracellular compartments. ated the first comprehensive 2-DE gel map of the lumi- Because many of the proteins identified were present in nal proteome of the yeast vacuole and established a multiple charge states and were often of lower molecu- baseline against which the protein profiles of vacuoles lar mass than their full-length translation products, from wild-type and mutant yeast under a wide range they were inferred to be products of intravacuolar pro- of conditions may be examined. The annotated gels cessing or degradation. This of course does not exclude inclusive of displacements associated with post-trans- the possibility, pending the results of other investiga- lational modifications will be a resource that many tions, that at least in some cases the proteins investigators will be able to draw upon for tracking concerned contribute to other processes within or inti- the modulation of protein levels and ⁄ or their modifica- mately associated with the vacuole. A case in point is tions under different conditions and in defining the some of the many HSPs identified which might serve true in vivo functions of the proteins in question. to chaperone some proteins into this compartment. Notable is the prevalence of proteins involved in Experimental procedures intermediary metabolism in the vacuolar lumen of yeast by comparison with the corresponding fraction Chemicals and yeast strain from Arabidopsis [7]. Although several enzymes belonging to this category, for example, aldehyde All of the proteomics and general reagents were obtained dehydrogenase, hexokinase and triosephosphate isom- from Fisher Scientific (Pittsburgh, PA), Becton-Dickinson erase, were identified in the Arabidopsis vacuole prote- (Franklin Lakes, NJ), Sigma-Aldrich (St. Louis, MO), Bio- ome, as a group they and others were only a minor Rad Laboratories (Hercules, CA) or GE Healthcare (Pis- fraction of the total number of protein species identi- cataway, NJ). S. cerevisiae strain SEY6210 (MATa, ura3– fied [7]. The full significance of this difference remains 52, leu2–3, 112, his3-D200, trp1-D901, lys2–801, suc2-D9, to be defined but the most likely explanation is that mel, GAL) which is isogenic to strain DTY165 [60,61] was whereas the yeast vacuole is largely a lytic compart- employed for all of the manipulations described.

FEBS Journal 274 (2007) 4287–4305 ª 2007 The Authors Journal compilation ª 2007 FEBS 4299 Proteomics of yeast vacuolar lumen J.-E. Sarry et al.

morphological homogeneity, parallel samples of the same Isolation of intact vacuoles preparations of cells, spheroplasts and vacuoles used for Intact vacuoles were isolated using a procedure optimized fluorescence microscopy (below) were fixed in 3.7% (v ⁄ v) to maximize both purity and integrity. Of the several paraformaldehyde, washed in NaCl ⁄ Pi and introduced into approaches examined experimentally, one based on a com- a FACSCalibur Flow Cytometer (BD Biosciences, San posite of the procedures described by Wiemken et al. [8], Jose, CA). Forward scatter plots were generated after ana- Roberts et al. [9] and Kim et al. [10] was determined to lyzing the signal using flowjo software (Tree Star, Ash- most satisfactorily meet these requirements. In this proce- land, OR). For the assessment of vacuolar integrity, yeast dure, liquid cultures were grown at 30 CtoanD600 of 1.8 liquid cultures were grown in YPD medium containing in YPD medium, harvested by centrifugation at 2800 g for 150 lm monochlorobimane (syn-(ClCH2.CH3)-1,5-diazabu- 5 min, and washed and resuspended in cell-wall digestion cyclo-[3.3.0]-octa-3,6-dione-2,8-dione; Molecular Probes, medium (0.7 m sorbitol, 100 mm Tris ⁄ HCl pH 7.5, 5 mm Eugene, OR) [62] to a D600 of 1.8 after which time the cells dithiothreitol) containing Zymolyase (5 mgÆg)1 cells; Zymo- were harvested and converted to spheroplasts for the purifi- lyase 20T, Seikagaku-Kogyo, Japan). For the liberation of cation of intact vacuoles as described above. Cells, sphero- vacuoles after 2 h of digestion at 30 C, the spheroplasts plasts and vacuoles from this source were observed without were pelleted by centrifugation at 1600 g for 5 min, resus- fixation under a Leica DM IRBE inverted fluorescence pended in sorbitol buffer (SOB: 0.6 m sorbitol, 1 mm di- microscope equipped with a 100 W Hg-vapor lamp and thiothreitol, 10 mm Tris-Mes, pH 6.8) containing protease appropriate barrier ⁄ emission filters. Images were collected inhibitor cocktail for fungal and yeast extracts [fungal pro- on an inline transfer chip CCD camera (Hamamatsu) and tease inhibitor cocktail; 0.1 mm 4-(2-aminoethyl)benze- color-adjusted using openlab software (Inprovision) and nesulfonyl fluoride, 2.2 lm pepstatin A, 1.4 lm E-64, and adobe photoshop 7.0 (Adobe Systems, San Jose, CA). 0.5 mm 1,10-phenanthroline; Sigma-Aldrich] and disrupted by Dounce homogenization. Intact vacuoles were fraction- Vacuolar marker enzyme assays ated from the spheroplast lysate by layering it onto a step gradient consisting of one layer of 3 vol SOB + 1 vol a-mannosidase a vacuolar membrane marker, carboxypepti- sucrose buffer (SUC: 0.6 m sucrose, 1 mm dithiothreitol, dase Y and aminopeptidase I, two vacuolar lumen markers, 10 mm Tris-Mes, pH 6.8, containing fungal protease inhibi- were employed to enumerate enrichment of the vacuolar tor cocktail), one layer of 2 vol SOB + 1 vol SUC, and fractions. a-mannosidase was assayed as the rate of one layer of 1 vol SUC +1 vol 2.5% (w ⁄ v) Ficoll-400 dis- liberation of p-nitrophenol from p-nitrophenyl-a-d-manno- solved in SUC. After centrifugation at 80 000 g for 1 h in a pyranoside [14]. p-Nitrophenol was monitored spectrophoto- swing-out rotor, the crude vacuole fraction which parti- metrically at 420 nm. Carboxypeptidase Y activity was tioned at the 2 : 1 SOB ⁄ 2.5% Ficoll interface was collected assayed as the rate of liberation of Leu from N-CBZ-l-Phe- and centrifuged at 3500 g for 20 min through a (2 vol l-Leu [12]. Leu was determined spectrophotometrically with SOB + 1 vol SUC) ⁄ (1.5 vol SOB + 1 vol SUC) step gra- l-amino acid oxidase in an o-dianisidine-coupled reaction. dient. Intact vacuoles pelleted by this procedure were resus- Aminopeptidase I activity was assayed as the rate of libera- pended in SOB and washed once in the same buffer. tion of p-nitroaniline from l-Leu p-nitroanilide [13]. P-Nitro- aniline was monitored spectrophotometrically at 405 nm. Preparation of vacuolar lysates Proteinase K protection assays Purified vacuoles were lysed by resuspension in lysis buffer (5 mm Tris ⁄ HCl, pH 7.0 containing 1 mm dithiothreitol), For the proteinase K protection assays, intact vacuoles vigorous agitation on a vortex mixer and flash-freezing in were isolated as described above except that generic serine liquid nitrogen. After thawing on ice, the lysates were cen- protease inhibitors were excluded from the isolation media trifuged at 100 000 g for 30 min to yield a supernatant by substituting the Sigma-Aldrich fungal protease inhibitor (vacuolar luminal protein fraction) and pellet (vacuolar cocktail with a mixture of 1 mm leupeptin, 1 mm aprotinin membrane fraction), the former of which was used for the and 1 mm pepstatin. The intact vacuole preparations were investigations described here. pretreated with proteinase K by incubating 500 lg aliquots of the suspensions with 50 lg proteinase K (15.6 mgÆmL)1; Roche, Mannheim, Germany) for 20 min on ice with gentle Fluorescence microscopy and flow cytometry mixing by inversion every 5 min. The digestions were The morphological homogeneity of the final vacuole isolate arrested by the addition of 1 mm phenylmethanesulfonyl and the maintenance of vacuolar integrity during the course fluoride, and the vacuoles were pelleted by centrifugation at of its purification were assessed by flow cytometry and fluo- 3500 g for 20 min in SOB containing Sigma-Aldrich fungal rescence microscopy, respectively. For the assessment of protease inhibitor cocktail. Negative control samples were

4300 FEBS Journal 274 (2007) 4287–4305 ª 2007 The Authors Journal compilation ª 2007 FEBS J.-E. Sarry et al. Proteomics of yeast vacuolar lumen treated in the same way except that proteinase K was omit- Estimation of spot molecular mass and pI ted from the digestion medium. Positive control samples The apparent molecular sizes of the gel-separated proteins were treated with proteinase K after lysis instead of before. were estimated by calibration of the SDS ⁄ PAGE dimension with a prestained SDS ⁄ PAGE standard mixture (Bio-Rad) 2D gel electrophoresis consisting of myosin (calibrated molecular mass on Tris ⁄ HCl gel, 200 768), b-galactosidase (115 281), BSA The vacuolar lysates were prepared for 2-DE as described by (96 190), ovalbumin (51 783), (37 659), Sarry et al. [63]. The soluble luminal proteins in the lysates soybean trypsin inhibitor (29 054), lysozyme (20 461) and were precipitated by the addition of 10% (w ⁄ v) trichloroace- aprotinin (7100). The apparent pI values of the separated tic acid dissolved in ice-cold acetone and centrifuged at proteins were estimated by calibrating the isoelectric focus- 16 000 g for 20 min. After two washes with ice-cold acetone ing dimension against nine of the proteins identified in the containing 0.05% (w ⁄ v) dithiothreitol, the protein pellets luminal samples whose pI values fell on a common calibra- were resuspended in isoelectric focusing loading buffer con- tion curve (ACO1, ANB1, CPR1, KGD1, PEP4, PGM1, taining 2 m thiourea, 6 m urea, 4% (w ⁄ v) Chaps and 50 mm SOD1, TDH3, TPI1). In both cases pdquest was employed dithiothreitol, incubated with stirring at room temperature for the calibrations and in assigning molecular mass and pI for 1 h, and centrifuged at 16 000 g for 15 min. For most of values to the spots on the gels. the experiments described, 300 lL aliquots of the isoelectric focusing buffer containing 200 lg protein and supplemented with 0.2% (v ⁄ v) pH 3–10 ampholines were loaded onto the In-gel tryptic digestions IPG strips (pH 3–10 NL; 11 cm; Bio-Rad) by passive For identification purposes, the Coomassie Brilliant Blue- rehydration for 12 h and active rehydration at 50 V for 8 h. stained protein spots on the gels were excised manually and Isoelectric focusing was performed for 120 kVh in a Protean transferred to 96-well microtiter plates for fully automated IEF cell (Bio-Rad) at 20 C. Voltage was increased linearly washing, destaining, tryptic digestion, elution, and pre- from 50 to 200 V for the first 3 h, held at 200 V for another paration for MALD-TOF-MS and ⁄ or LC-IT- or QTOF- 3 h, before being increased linearly to 1000 V over the next MS ⁄ MS in a Multiprobe II MassPREP Station 3 h and held at that voltage for another 3 h. In the final two (Waters ⁄ Micromass, Inc., Milford, MA). Routinely, the periods of 6 h each, the voltage was increased linearly to excised gel plugs were washed several times for 15 min, first 5000 V and held at that value. Upon completion of the with aqueous 50 mm ammonium bicarbonate (pH 7.8) and isoelectric focusing separations, the isoelectric focusing then with 50% (v ⁄ v) acetonitrile containing 50 mm ammo- strips were equilibrated with 50 mm Tris ⁄ HCl (pH 8.8) con- nium bicarbonate before dehydration for 5 min in acetoni- taining 6 m urea, 30% (v ⁄ v) glycerol, 2% (w ⁄ v) SDS and trile. The gel plugs were digested at 37 Cin25lL50mm 2% (w ⁄ v) dithiothreitol for 20 min with gentle stirring ammonium bicarbonate (pH 7.8) containing 6 lgÆmL)1 before the addition of 2.5% (w ⁄ v) iodoacetamide, and alkyl- trypsin (sequencing grade, modified, Promega) for 7 h. The ation of the samples for 20 min. Electrophoresis in the sec- resulting peptide digests were successively extracted in ond dimension was performed on Criterion 8.7 · 13.3 cm 20 lL 0.1% (v ⁄ v) formic acid in 2% (v ⁄ v) acetonitrile, and precast 8–16% (w ⁄ v) polyacrylamide gradient gels (Bio- with 0.1% (v ⁄ v) formic acid in 50% (v ⁄ v) acetonitrile. Rad) at 50 V for the first hour and at 200 V for the second For MALDI-TOF-MS, 1 lL aliquots of the peptide hour in a Protean Cell Criterion apparatus (Bio-Rad). extracts were mixed with 1 lL 2–5 mgÆmL)1 matrix (a-cyano-4-hydroxycinnamic acid; purified by recrystalliza- Gel staining, image analysis and spot intensity tion from ethanol) dissolved in 0.1% (w ⁄ v) trifluoroacetic determinations acid ⁄ 50% (v ⁄ v) acetonitrile before spotting 1 lL aliquots onto a 96-place stainless steel target plate. For LC-IT- or For analyses of the protein profiles, the gels were stained with QTOF-MS ⁄ MS, the digests prepared in the MassPREP sta- colloidal Coomassie Brilliant Blue G-250 (GE Healthcare), tion were taken to incipient dryness in a Speedvac for scanned with an Epson flatbed scanner, and the TIFF files immediate resuspension in LC loading buffer. generated using adobe photoshop 7.0 were subjected to spot detection, matching and quantitation using the pdquest soft- ware package (v. 7.2, Bio-Rad). For quantitative analyses, Protein identification by MS and database three-dimensional Gaussian images of the spots were gener- searching ated and their volumes were estimated after background sub- The MALDI-TOF-MS analyses were performed on a traction. The intensity of each spot was normalized to the Micromass MALDI Reflectron MS (Waters ⁄ Micromass). sum total of spots on the gel such that the integrated and The LC-IT-MS ⁄ MS and QTOF-MS ⁄ MS analyses were normalized intensity for each spot expressed as a percentage performed on an LCQ Ion Trap MS (Deca XP Plus, was calculated as the intensity divided by the sum of the Thermo-Finnigan, San Jose, CA) and an ABI QSTAR XL intensities of all the spots in the Gaussian image.

FEBS Journal 274 (2007) 4287–4305 ª 2007 The Authors Journal compilation ª 2007 FEBS 4301 Proteomics of yeast vacuolar lumen J.-E. Sarry et al. hybrid QTOF-MS ⁄ MS (Applied Biosystems ⁄ MDS Sciex, 1–2 mm from the opening of the ion transfer capillary and Foster City, CA), respectively. slightly off axis. Most mass spectra were acquired using the The MALDI-TOF-MS, which was equipped with a repetitive ‘triple-play’ sequence as recommended by the pulsed nitrogen laser (337 nm, pulse width 4 ns), was oper- manufacturer which consisted of a full scan event for ions ated in reflectron mode in the mass range 900–3500 Da at with m ⁄ z ratios of 400–1200, a zoom scan event acquired an accelerating voltage of 20 kV. Spectra were calibrated within an m ⁄ z window of 10 units centered on the chosen using a mixture of angiotensin II (1046.54 Da), angioten- ion, and an MS ⁄ S scan event. Ions were selected for the sin I (1296.52 Da), substance P (1347.74 Da), fibrinopeptide zoom scan and for the MS ⁄ MS scan automatically in a (1570.64 Da), (1759.05 Da) and ACTH (18–39) data-dependent manner, whereby ions of sequentially (2465.70 Da). For routine database searches, spectra from decreasing abundance were chosen and two scan events 200 laser shots at several positions in the target spot were were allowed for any given ion in a 3-s time window. For combined to generate a peptide mass fingerprint. To mini- complex samples, the zoom scan event was omitted. The mize errors, spectrum acquisition and annotation were done tolerance for the selection of the precursor ions ranged both automatically and manually. Typically, 15–40 tryptic from 1.5 to 3.0 m ⁄ z (low to high m ⁄ z). peptide masses per fingerprint were acquired. The mass The measured MS ⁄ MS spectra were matched with tryptic spectra were analyzed using mascot (v. 1.7) assuming a peptide amino acid sequences from a S. cerevisiae database. maximum of one missed cleavage per peptide and a mass- Raw MS ⁄ MS data files that had a minimum total ion cur- to-charge (m ⁄ z) tolerance of 50 p.p.m., and searched rent of 105 and contained 15 or more fragment ions were against the S. cerevisiae NCBI database version v16 selected. The tolerance window for the grouping of raw released 05 ⁄ 12 ⁄ 06. Acetylation of the N-terminus, alkyl- MS ⁄ MS data files into input files for the finnigan ation of cysteine by carbamidomethylation, oxidation of sequest ⁄ turbo sequest software (revision 2.0; Thermo- methionine, and ⁄ or the formation of pyroglutamate from Quest) was set to 1.4 amu. The sequest algorithm was used N-terminal glutamine residues were accepted as possible to identify and retrieve peptide sequences from the database peptide modifications. Matching was decided on the basis possessing at least one tryptic end with a theoretical mass of the number of peptide matches (6–31) and the coverage within 1.25 amu of that measured for the precursor ion and (typically > 27%), with consideration of both the scores a theoretical y- and b-ion profile bearing a high degree of and the predicted molecular masses and pI values of the similarity to the experimental MS ⁄ MS spectrum. The simi- polypeptides retrieved from the database versus the appar- larity between an experimental and theoretical MS ⁄ MS ent molecular mass and pI values of the sample on the spectrum, reported as the cross-correlation factor (Xcorr) 2D gel. and the difference between the unit-normalized Xcorr values The LC-IT-MS ⁄ MS was interfaced to a Famos autosam- of the first- and second-ranked hits (DCorr) provided crite- pler (LC-Packings, Sunnyvale, CA) which was used to sam- ria for the preliminary assignment of amino acid sequences ple the tryptic digests after their resuspension in 5% (v ⁄ v) based on the experimental MS ⁄ MS spectra. Sequences were acetonitrile, 0.1% (v ⁄ v) formic acid. The aqueous phase reported if the Xcorr values were equal to or greater than (A) (0.1% v ⁄ v formic acid in 5 : 95 acetonitrile ⁄ water) and 1.50, 2.25 and 3.00 for singly, doubly or triply charged pre- organic phase (B) (0.1% v ⁄ v formic acid in 80 : 20 aceto- cursor ions, respectively, and if the DCorr value exceeded nitrile ⁄ water) were delivered using an Ultimate HPLC 0.1 against the database background. (LC-Packings). After first injecting the samples onto a The QTOF-MS ⁄ MS was equipped with a nanoelectro- l-Precolumn (300 lm i.d. · 5 mm, 5 lm particle size, spray source (Protana XYZ manipulator) and operated in 100 A˚ pore diameter, PepMap C18; LC-Packings) which TOF MS and MS ⁄ MS m ⁄ z ranges of 300–1500 and was washed with solvent A at a flow rate of 0.25 lLÆmin)1 65–1500, respectively. The instrumental setting were config- for 4 min, a Switchos microcolumn switching module ured and the database searches were conducted as describe (LC-Packings) was used to direct the flow from waste to by Chen [64] and Sheffield et al. [65]. the analytical column (75 lm i.d. · 150 lm, 3 lm particle size, 100 A˚ pore diameter, PepMap C18; LC-Packings). Protein quantification The analytical column was developed with 5% B for the first 8 min, 5–50% B from 8 to 38 min, 50–95% B from 38 Protein concentrations were estimated by the Bradford to 39 min, 95% B from 39 to 49 min, 95–5% B from 49 to method using BSA as the calibration standard [66]. 50 min, and 5% B from 50 to 75 min. The solvents deliv- ered by the pump were split at a flow ratio of 1 : 1250 with the postsplitter flow rate set at 250 nLÆmin)1.An8lm Acknowledgements aperture fused silica needle (New Objective, Woburn, MA) This work was funded by National Science Foundation was used to introduce samples. ESI was accomplished by Grant No. MCB-0313461, United States Department applying a voltage difference of ± 1.3 kV across the fused of Energy Grant No. DE-FG02–91ER20055, a Penn silica needle. The aperture of the needle was positioned

4302 FEBS Journal 274 (2007) 4287–4305 ª 2007 The Authors Journal compilation ª 2007 FEBS J.-E. Sarry et al. Proteomics of yeast vacuolar lumen

University Research Foundation Grant and a Penn vacuolar glutathione S-conjugate pump. J Biol Chem Genomics Institute Seed Grant awarded to PAR, and 271, 6509–6517. carried out in the Rea Laboratory, University of Penn- 12 Stevens TH, Rothman JH, Payne GS & Schekman R slvania. The authors thank Dr Daniel Klionsky for (1986) Gene dosage-dependent secretion of yeast invaluable discussions, Dr Fevzi Daldal for unre- vacuolar carboxypeptidase Y. J Cell Biol 102, 1551– stricted access to his IEF and PDQuest resources, and 1557. two anonymous reviewers for excellent suggestions on 13 Jones EW (1991) Three proteolytic systems in the yeast how the manuscript might be improved. Saccharomyces cerevisiae. J Biol Chem 266, 7963–7966. 14 Jelinek-Kelly S, Akiyama T, Saunier B, Tkacz JS & Herscovics A (1985) Characterization of a specific References a-mannosidase involved in oligosaccharide processing in Saccharomyces cerevisiae. J Biol Chem 260, 2253– 1 Klionsky DJ & Ohsumi Y (1999) Cytoplasm to 2257. vacuole protein transport. Annu Rev Cell Dev Biol 15, 15 Yoshihisa T, Ohsumi Y & Anraku Y (1988) Solubiliza- 1–32. tion and purification of alpha-mannosidase, a marker 2 Kucharczyk R & Rytka J (2001) Saccharomyces enzyme of vacuolar membranes in Saccharomyces cere- cerevisiae – a model organism for the studies on visiae. J Biol Chem 263, 5158–5163. vacuolar transport. Acta Biochim Pol 48, 1025–1042. 16 Horst M, Knecht EC & Schu PV (1999) Import into 3 Khalfan WA & Klionsky DJ (2002) Molecular machin- and degradation of cytosolic proteins by isolated yeast ery required for autophagy and the cytoplasm to vacu- vacuoles. Mol Biol Cell 10, 2879–2889. ole targeting (Cvt) pathway in S. cerevisiae. Curr Opin 17 Marcusson EG, Horazdovsky BF, Cereghino JL, Cell Biol 14, 468–475. Gharakhanian E & Emr SD (1994) The sorting receptor 4 Goffeau A, Barrell BG, Bussey H, Davis RW, Dujon B, for yeast vacuolar carboxypeptidase Y is encoded by Feldmann H, Galibert F, Hoheisel JD, Jacq C, John- the VPS10 gene. Cell 77, 579–586. ston M et al. (1996) Life with 6000 genes. Science 274, 18 Oda MN, Scott SV, Hefner GA, Caffarelli AD & 563–567. Klionsky DJ (1996) Identification of a cytoplasm to vac- 5 Shimaoka T, Ohnishi M, Sazuka T, Mitsuhashi N, uole targeting determinant in aminopeptidase I. J Cell Hara-Nishimura I, Shimazaki K, Maeshima M, Yokota Biol 132, 999–1010. A, Tomizawa K et al. T (2004) Isolation of intact vacu- 19 Yasuhara T, Nakai T & Ohashi A (1994) Aminopepti- oles and proteomic analysis of tonoplast from suspen- dase Y, a new aminopeptidase from Saccharomyces sion-cultured cells of Arabidopsis thaliana. Plant Cell cerevisiae – purification, properties, localization, and Physiol 45, 672–683. processing by protease-b. J Biol Chem 269, 13644– 6 Szponarski W, Sommerer N, Boyer JC, Rossignol M & 13650. Gibrat R (2004) Large-scale characterization of integral 20 Ammerer G, Hunter CP, Rothman JH, Saari GC, Valls proteins from Arabidopsis vacuolar membrane by two- A & Stevens TH (1986) PEP4 gene of Saccharomyces dimensional liquid chromatography. Proteomics 4, 397– cerevisiae encodes proteinase A, a vacuolar enzyme 406. required for processing of vacuolar precursors. Mol Cell 7 Carter C, Pan S, Zouhar J, Avila EL, Girke T & Raik- Biol 6, 2490–2499. hel NV (2004) The vegetative vacuole proteome of 21 Kambouris NG, Burke DJ & Creutz CE (1992) Clon- Arabidopsis thaliana reveals predicted and unexpected ing and characterization of a cysteine proteinase from proteins. Plant Cell 16, 3285–3303. Saccharomyces cerevisiae. J Biol Chem 267, 21570– 8 Wiemken A, Schellenberg M & Urech K (1979) Vacu- 21576. oles: the sole compartments of digestive enzymes in 22 Enenkel C & Wolf DH (1993) BLH1 codes for a yeast yeast (Saccharomyces cerevisiae)? Arch Microbiol 123, thiol aminopeptidase, the equivalent of mammalian 23–35. bleomycin hydrolase. J Biol Chem 268, 7036–7043. 9 Roberts CJ, Raymond CK, Yamashiro CT & Stevens 23 Yokoyama R, Kawasaki H & Hirano H (2006) Identifi- TH (1991) Methods for studying the yeast vacuole. cation of yeast aspartyl aminopeptidase gene by purify- Methods Enzymol 194, 644–661. ing and characterizing its product from yeast cells. 10 Kim EJ, Zhen R & Rea PA (1994) Heterologous expres- FEBS Lett 273, 192–198. sion of plant vacuolar pyrophosphatase in yeast demon- 24 Zhang N, Ismail T, Wu J, Woodwark KC, Gardner strates sufficiency of substrate-binding subunit for DC, Walmsley RM & Oliver SG (1999) Disruption of proton transport. Proc Natl Acad Sci USA 91, 6128– six novel ORFs on the left arm of XII 6132. reveals one gene essential for vegetative growth of 11 Li Z-S, Szczypka M, Lu Y-P, Thiele DJ & Rea PA Saccharomyces cerevisiae. Yeast 15, 1287–1296. (1996) The yeast cadmium factor protein (YCF1) is a

FEBS Journal 274 (2007) 4287–4305 ª 2007 The Authors Journal compilation ª 2007 FEBS 4303 Proteomics of yeast vacuolar lumen J.-E. Sarry et al.

25 Chu S, DeRisi J, Eisen M, Mulholland J, Botstein D, 40 Lafourcade C, Galan J-M, Gloor Y, Haguenauer-Tsapis Brown PO & Herskowitz I (1998) The transcription pro- R & Peter M (2004) The GTPase-activating enzyme gram of sporulation in budding yeast. Science 282, 699– Gyp1p is required for recycling of internalized mem- 705. brane material by inactivation of the Rab ⁄ Ypt GTPase 26 Naitou M, Hagiwara H, Hanaoka F, Eki T & Muraka- Ypt1p. Mol Cel Biol 24, 3815–3826. mi Y (1997) Expression profiles of transcripts from 126 41 Gladfelter AS, Pringle JR & Lew DJ (2001) The septin open reading frames in the entire chromosome VI of cortex at the mother–bud neck. Curr Opin Microbiol 4, Saccharomyces cerevisiae by systematic northern analy- 681–689. ses. Yeast 13, 1275–1290. 42 Levin DE (2005) Cell wall integrity signaling in Saccha- 27 Tatukova RL, Galperin MY, Natale DA & Koonin EV romyces cerevisiae. Microbiol Mol Biol Rev 69, 262–291. (2000) The COG database: a tool for genome-scale anal- 43 Liu L, Zeng M, Hausladen A, Heitman J & Stamler JS ysis of protein functions and evolution. Nucleic Acids (2000) Protection from nitrosative stress by yeast fla- Res 28, 33–36. vohemoglobin. Proc Natl Acad Sci USA 97, 4672–4676. 28 Sickmann A, Reinders J, Wagner Y, Joppich C, Zahedi 44 Silles E, Mazon MJ, Gevaert K, Goethals M, Van- R, Meyer HE, Schonfisch B, Perschil I, Chacinska A, dekerckhove J, Leber R & Sandoval IV (2000) Target- Guiard B et al. (2003) The proteome of Saccharomyces ing of aminopeptidase I to the yeast vacuole is mediated cerevisiae mitochondria. Proc Natl Acad Sci USA 100, by Ssa1p, a cytosolic member of the 70-kDa stress pro- 13207–13212. tein family. J Biol Chem 275, 34054–34059. 29 Kleijnen MF, Kirkpatrick DS & Gygi SP (2007) The 45 Ishihara K, Yamamoto H, Mitsuhashi K, Nishikawa K, ubiquitin–proteasome system regulates membrane fusion Tsuboi S, Tsuji H & Nakajima N (2004) Purification of yeast vacuoles. EMBO J 26, 275–287. and characterization of a-keto amide reductase from 30 Berger AC, Vanderford TH, Gernert KM, Nichols JW, Saccharomyces cerevisiae. Biosci Biotechnol Biochem 68, Faundez V & Corbett AH (2005) Saccharomyces cerevi- 2306–2312. siae Npc2p is a functionally conserved homologue of 46 Hahne K, Haucke V & Schatz G (1994) Incomplete the human Niemann–Pick disease type C2 protein, arrest in the outer membrane sorts NADH-cytochrome

hNPC2. Eukaryotic Cell 4, 1851–1862. b5 reductase to two different submitochondrial compart- 31 Uttenweiler A, Schwarz H, Neumann H & Mayer A ments. Cell 79, 829–839. (2006) The vacuolar transporter chaperone (VTC) is 47 Chang YH & Smith JA (1989) Molecular cloning and required for microautophagy. Mol Biol Cell 18, 166– sequencing of genomic DNA encoding aminopeptidase I 175. from Saccharomyces cerevisiae. J Biol Chem 264, 6979– 32 Nwaka S, Mechler B & Holzer H (1996) Deletion of the 6983. ATH1 gene in Saccharomyces cerevisiae prevents growth 48 Yasuhara T, Nakai T & Ohashi A (1994) Aminopepti- on trehalose. FEBS Lett 386, 235–238. dase Y, a new aminopeptidase from Saccharomyces 33 Klionsky DJ & Emr SD (1989) Membrane protein sort- cerevisiae. Purification, properties, localization, and pro- ing: biosynthesis, transport and processing of yeast vac- cessing by protease B. J Biol Chem 269, 13644–13650. uolar alkaline phosphatase. EMBO J 8, 2241–2250. 49 Nishizawa M, Yasuhara T, Nakai T, Fujiki Y & Ohashi 34 Chiang H-L & Schekman R (1991) Regulated import A (1994) Molecular cloning of the aminopeptidase Y and degradation of a cytosolic protein in the yeast vacu- gene of Saccharomyces cerevisiae. Sequence analysis and ole. Nature 350, 313–318. gene disruption of a new aminopeptidase. J Biol Chem 35 Scott SV & Klionsky DJ (1997) Nonclassical protein 269, 13651–13655. sorting. Trends Cell Biol 7, 225–229. 50 Nebes VL & Jones EW (1991) Activation of the protein- 36 Klionsky DJ (1998) Nonclassical protein sorting to the ase B precursor of the yeast Saccharomyces cerevisiae by yeast vacuole. J Biol Chem 273, 10807–10810. autocatalysis and by an internal sequence. J Biol Chem 37 Trabalzini L, Paffetti A, Scaloni A, Talamo F, Ferro E, 266, 22851–22857. Coratza G, Bovalini L, Lusini P, Martelli P & Santucci 51 Hirsch HH, Schiffer HH & Wolf DH (1992) Biogenesis A (2003) Proteomic response to physiological fermenta- of the yeast vacuole (lysosome). Eur J Biochem 207, tion stresses in a wild-type wine strain of Saccharomyces 867–876. cerevisiae. Biochem J 370, 35–46. 52 Hasilik A & Tanner W (1978) Biosynthesis of the vacu- 38 Onodera J & Ohsumi Y (2004) Ald6p is a preferred tar- olar yeast glycoprotein carboxypeptidase Y. Conversion get for autophagy in yeast, Saccharomyces cerevisiae. of precursor into the enzyme. Eur J Biochem 85, 599– J Biol Chem 279, 16071–16076. 608. 39 Eitzen G, Wang L, Thirngren N & Wickner W (2002) 53 Valls LA, Hunter CP, Rothman JH & Stevens TH The Vtc proteins in vacuole fusion: coupling NSF activ- (1987) Protein sorting in yeast: the localization determi- ity to Vo trans-complex formation. J Cell Biol 158, nant of yeast vacuolar carboxypeptidase Y resides in 669–679. the propeptide. Cell 48, 887–897.

4304 FEBS Journal 274 (2007) 4287–4305 ª 2007 The Authors Journal compilation ª 2007 FEBS J.-E. Sarry et al. Proteomics of yeast vacuolar lumen

54 Sørensen SO, van den Hazel HB, Kielland-Brandt MC normal human red cells under physiological conditions. & Winther JR (1994) pH-dependent processing of yeast Proc Natl Acad Sci USA 76, 3382–3386. procarboxypeptidase Y by proteinase A in vivo and 63 Sarry J-E, Kuhn L, LeLay P, Garin J & Bourguignon J in vitro. Eur J Biochem 220, 19–27. (2006) Dynamics of Arabidopsis thaliana soluble prote- 55 Klionsky DJ, Banta LM & Emr SD (1988) Intracellular ome in response to different nutrient culture conditions. sorting and processing of a yeast vacuolar hydrolase: Electrophoresis 27, 495–507. proteinase A propeptide contains vacuolar targeting 64 Chen S (2006) Rapid protein identification using direct information. Mol Cell Biol 8, 2105–2116. infusion nanoelectrospray ionization mass spectrometry. 56 van den Hazel HB, Kielland-Brandt MC & Winther JR Proteomics 6, 16–25. (1993) The propeptide is required for in vivo formation 65 Sheffield J, Taylor N, Fauquet C & Chen S (2006) Cas- of stable active yeast proteinase A and can function sava root proteome: protein identification and differen- even when not covalently linked to the mature region. tial expression. Proteomics 6, 1588–1598. J Biol Chem 268, 18002–18007. 66 Bradford MM (1976) A rapid and sensitive method for 57 van den Hazel HB, Kielland-Brandt MC & Winther JR the quantitation of microgram quantities of protein (1995) Random substitution of large parts of the pro- utilizing the principle of protein–dye binding. Anal peptide of yeast proteinase A. J Biol Chem 270, 8602– Biochem 72, 248–254. 8609. 58 De DN (2000) Plant Cell Vacuoles. CSIRO Publishing, Collingwood, Australia. Supplementary material 59 Vitale A & Raikhel NV (1999) What do proteins need The following supplementary material is available to reach different vacuoles? Trends Plant Sci 4, 149–155. online: 60 Robinson JS, Klionsky DJ, Banta LM & Emr SD Table S1. Identities of 2-DE separated vacuolar lumi- (1988) Protein sorting in Saccharomyces cerevisiae: isola- nal proteins from S. cerevisiae SEY6210. tion of mutants defective in the delivery and processing This material is available as part of the online article of multiple vacuolar hydrolases. Mol Cell Biol 8, 4936– from http://www.blackwell-synergy.com 4948. Please note: Blackwell Publishing is not responsible 61 Szczypka MS, Wemmie JA, Moye-Rowley WS & Thiele for the content or functionality of any supplementary DJ (1994) A yeast metal resistance protein similar to materials supplied by the authors. Any queries (other human cystic fibrosis transmembrane conductance regu- than missing material) should be directed to the corre- lator (CFTR) and multidrug resistance-associated pro- sponding author for the article. tein. J Biol Chem 269, 22853–22857. 62 Kosower NS, Kosower EM, Newton GL & Ranney HM (1979) Bimane fluorescent labels: labeling of

FEBS Journal 274 (2007) 4287–4305 ª 2007 The Authors Journal compilation ª 2007 FEBS 4305