Ald6p is a preferred target for autophagy in yeast, Saccharomyces cerevisiae.

Macroautophagy is the process of intracellular bulk protein degradation induced by nutrient starvation and is generally considered to be a nonselective degradation of cytosolic enzymes and organelles. However, it remains a possibility that some proteins may be preferentially degraded by autophagy. In this study, we have performed a systematic analysis on the substrate selectivity of autophagy in yeast, Saccharomyces cerevisiae, using two-dimensional PAGE. We performed a differential screen on wild-type and Deltaatg7/apg7 autophagy-deficient cells and found that cytosolic acetaldehyde dehydrogenase (Ald6p) decreased under nitrogen starvation. As assessed by immunoblot, Ald6p was reduced by greater than 82% after 24 h of nitrogen starvation. This reduction was dependent on Atg/Apg proteins and vacuolar proteases but was not dependent on the proteasome or the cytoplasm to vacuole targetting (Cvt) pathway. Using pulse-chase and subcellular fractionation, we have also demonstrated that Ald6p was preferentially transported to vacuoles via autophagosomes. Deltaatg7 Deltaald6 double mutant cells were able to maintain higher rates of viability than Deltaatg7 cells under nitrogen starvation, and Ald6p-overexpressing cells were not able to maintain high rates of viability. Furthermore, the Ald6p(C306S) mutant, which lacks enzymatic activity, had viability rates similar to Deltaald6 cells. Ald6p enzymatic activity may be disadvantageous for survival under nitrogen starvation; therefore, yeast cells may preferentially eliminate Ald6p via autophagy.

Macroautophagy is the process of intracellular bulk protein degradation induced by nutrient starvation and is generally considered to be a nonselective degradation of cytosolic enzymes and organelles. However, it remains a possibility that some proteins may be preferentially degraded by autophagy. In this study, we have performed a systematic analysis on the substrate selectivity of autophagy in yeast, Saccharomyces cerevisiae, using two-dimensional PAGE. We performed a differential screen on wild-type and ⌬atg7/apg7 autophagy-deficient cells and found that cytosolic acetaldehyde dehydrogenase (Ald6p) decreased under nitrogen starvation. As assessed by immunoblot, Ald6p was reduced by greater than 82% after 24 h of nitrogen starvation. This reduction was dependent on Atg/Apg proteins and vacuolar proteases but was not dependent on the proteasome or the cytoplasm to vacuole targetting (Cvt) pathway. Using pulse-chase and subcellular fractionation, we have also demonstrated that Ald6p was preferentially transported to vacuoles via autophagosomes. ⌬atg7 ⌬ald6 double mutant cells were able to maintain higher rates of viability than ⌬atg7 cells under nitrogen starvation, and Ald6p-overexpressing cells were not able to maintain high rates of viability. Furthermore, the Ald6p C306S mutant, which lacks enzymatic activity, had viability rates similar to ⌬ald6 cells. Ald6p enzymatic activity may be disadvantageous for survival under nitrogen starvation; therefore, yeast cells may preferentially eliminate Ald6p via autophagy.
Cellular activities require the maintenance of a balance between the synthesis and degradation of proteins. Macroautophagy (hereafter referred to as autophagy) is an intracellular bulk degradation system that is well conserved in eukaryotes; autophagy transports cytoplasmic components to the lysosome/ vacuole for degradation (1). This degradation is a cellular response to starvation and also plays a role in the recycling of cytoplasmic components, which may be important for cellular remodeling, development, and differentiation (2)(3)(4). A total of 16 genes that are essential for autophagy and that are named APG and AUT (current nomenclature is ATG) (5) have been identified by genetic screens in yeast, Saccharomyces cerevi-siae. Much progress has been made in the functional analysis of these genes (6 -8). In eukaryotic cells, there is another major degradation system, the ubiquitin proteasome pathway, which mediates the selective ubiquitination and subsequent degradation of proteins by the proteasome. This pathway serves mainly to degrade short-lived proteins, such as transcription factors, cell cycle regulators, and defective proteins. However, more than 99% of cellular proteins are long-lived, and autophagic degradation contributes to the turnover of these proteins. In contrast to selective degradation by the ubiquitin proteasome pathway, autophagy is generally thought to be nonselective. Autophagy is initiated by the sequestration of cytoplasmic components in a double-membrane structure termed the autophagosome. Immunoelectron microscopy has shown that ribosomes and typical cytosolic marker enzymes, such as alcohol dehydrogenase (ADH) 1 and phosphoglycerate kinase (PGK), are present in the autophagosome and autophagic bodies at the same densities as in the cytosol (9). The measurement of the enzymatic activities of these proteins also supports this conclusion (10). 2 If degradation of long-lived proteins is exclusively mediated by autophagy, all proteins might be expected to have similar lifetimes. However, long-lived proteins have a variety of lifetimes; therefore, the autophagic pathway might have some selectivity. It is known that fructose-1,6-bisphosphatase (via the vacuolar import and degradation, the Vid pathway) (6,11) and the peroxisome (via pexophagy) (12) are selectively transported from the cytoplasm to the vacuole and degraded.
To investigate the possibility of selective autophagic degradation, we attempted to compare the amounts of each intracellular protein under growth and starvation conditions in yeast, S. cerevisiae. We performed a systematic analysis using two-dimensional PAGE and MALDI-TOF mass spectrometry to detect the autophagy-dependent degradation of intracellular proteins. During these analyses, we observed an interesting behavior of the Mg 2ϩ -and NADPH-dependent cytosolic acetaldehyde dehydrogenase (Ald6p), which catalyzes the conversion of acetaldehyde to acetate in the cytosol (acetaldehyde ϩ NADP ϩ 3 acetate ϩ NADPH) (13). The S. cerevisiae genome encodes five or more different members of the aldehyde dehydrogenase family. Ald4p is the major K ϩ -and NAD ϩ -dependent mitochondrial acetaldehyde dehydrogenase (14), and Ald5p is a minor K ϩ -dependent mitochondrial acetaldehyde dehydrogenase, which is induced when cells are grown in ethanol-containing medium (15). Ald2p and Ald3p are closely related cytosolic enzymes that are required for in vivo pantothenic acid biosynthesis via conversion of 3-aminopropanol to ␤-alanine (16). Ald4p and Ald6p function in the conversion of acetaldehyde to acetate, which is a key intermediate during fermentation of sugars and growth on ethanol and are consequently important for acetyl-CoA production. In contrast, Ald2p, Ald3p, and Ald5p may not contribute to the oxidation of acetaldehyde in vivo. Therefore, Ald6p is the only cytosolic acetaldehyde dehydrogenase in the yeast cell. We demonstrate here that Ald6p is degraded preferentially by autophagy and that reduction of Ald6p may improve viability rates under nitrogen starvation.

EXPERIMENTAL PROCEDURES
Yeast Strains and Media-The S. cerevisiae strains used in this study are listed in Table I. Standard techniques were used for yeast manipulation (17). Yeast cells were grown in YPD medium (1% yeast extract, 2% polypeptone, and 2% glucose) or SD ϩ CA medium (0.17% yeast nitrogen base without amino acids and ammonium sulfate, 0.5% casamino acid, 0.5% ammonium sulfate, and 2% glucose) supplemented with 0.002% adenine sulfate, 0.002% uracil, and 0.002% tryptophan if necessary. For nitrogen starvation, SD(ϪN) medium (0.17% yeast nitrogen base without ammonium sulfate and amino acids and with 2% glucose) was used.
Plasmid Construction-To create the glutathione S-transferase-Ald6p fusion construct (pJO1), the open reading frame of ALD6 (YPL061w) lacking the initiation codon (1.5 kb) was amplified by genomic PCR using the following primers: 5Ј-CGCGGATCCACTAAG-CTACACTTTGACACTGC-3Ј and 5Ј-CCGCTCGAGCAACTTAATTCT-GACAGCTTTTACTTC-3Ј. This strategy incorporated novel BamHI and XhoI sites into the resulting DNA fragment, which was then cloned into the BamHI and XhoI sites of pGEX-4T-1 (Amersham Biosciences) to yield pJO1. To create the Ald6p-GFP genome integration vector (pJO-402), novel XbaI sites were added to the terminator sequence (0.5 kb) of ALD6 by genomic PCR amplification using the following primers: 5Ј-G-CTCTAGATGTACCAACCTGCATTTCTTTC-3Ј and 5Ј-GCTCTAGACG-AAGAAGGATGTTATTATATG-3Ј. Novel XhoI and BamHI sites were added to the ALD6 promoter region (0.3 kb) and the ALD6 open reading frame lacking the stop codon (1.5 kb) by genomic PCR amplification using the following primers: 5Ј-CGCTCGAGCACCGACCATGTGGGC-AAATTC-3Ј and 5Ј-CGCGGATCCCAACTTAATTCTGACAGCTTTTAC-3Ј. BamHI sites were added to the open reading frame of modified GFP (S65T) lacking the initiation codon by PCR amplification using the following primers: 5Ј-CGCGGATCCGGTAAAGGAGAAGAACTTTTCA-CTGG-3Ј and 5Ј-CGGGATCCTTACTTGTATAGTTCATCCATG-3Ј. The resulting DNA fragments were cloned into the pRS306 integration vector (18) to yield pJO402. To create the Ald6p overexpression construct (pJO203), XhoI and BamHI sites were added to a sequence containing the ALD6 open reading frame (1.5 kb), the promoter region (0.3 kb), and the terminator sequence (0.5 kb) by genomic PCR amplification using the following primers: 5Ј-CCGCTCGAGCACCGACCATGTGGG-CAAATTC-3Ј and 5Ј-CGCGGATCCCGAAGAAGGATGTTATTATATG-ATCTC-3Ј. The resulting DNA fragment was cloned into the BamHI and XhoI sites of the pRS426 multicopy plasmid (18) to yield pJO203. A QuikChange TM site-directed mutagenesis kit (Stratagene) was used to create the Ald6p C306S mutant overexpression construct (pJO213). To generate pJO213, the pJO203 plasmid was amplified by PCR with the following primers: 5Ј-AGAACGCTGGTCAAATTTCTTCCTCTGGTT-3Ј and 5Ј-AACCAGAGGAAGAAATTTGACCAGCGTTCT-3Ј. The site of mutagenesis in pJO213 was confirmed by automated DNA sequencing.
Two-dimensional PAGE and Peptide Mass Fingerprinting-The lysates were prepared by breaking yeast cells with glass beads in lysis buffer containing 50 mM Tris-HCl (pH 8.0), 1 mM EDTA, 1 mM phenylmethanesulfonyl fluoride, and protease inhibitor mixture (Roche Applied Science). The lysates were centrifuged at 100,000 ϫ g for 1 h, and the supernatant was desalted with NAP-10 TM (Amersham Biosciences). The protein concentrations were determined using a BCA assay kit (Pierce), and 300 g of each lysate was applied to the gel. Isoelectric focusing was performed with IPGphor™ (Amersham Biosciences) and a 13 cm Immobiline TM DryStrip pH 4 -7 (Amersham Biosciences) as described (19). The gel strip was subjected to SDS-PAGE (12.5% acrylamide), and the gel was stained with Coomassie Brilliant Blue R-250. The protein spots were picked, washed with 100 mM ammonium bicarbonate, dehydrated with acetonitrile, and dried in an evaporator. The spots were digested in the gel with 0.5 mg/ml of trypsin (Promega) in 100 mM ammonium bicarbonate for 12 h at 30°C. The digested peptides were extracted from the gel with 10% formic acid and 50% acetonitrile and desalted with the ZipTip TM C-18 (Millipore). The samples were mixed with ␣-cyano-4-hydroxy-cinnamic acid (Fluka) in a 2:1 ratio and were analyzed by MALDI-TOF mass spectrometry, REFLEX III (Bruker). The proteins were identified by searching the ProFound data base (www.129.85.19.192/).
Antibodies-Ald6p-specific antibodies were prepared as follows. The pJO1 plasmid was transformed into Escherichia coli (DH5␣), and transformants were grown in LB medium containing 50 g/ml ampicillin to an A 600 of 0.6. Recombinant protein expression was induced with 1 mM isopropyl-␤-D-thiogalactopyranoside for an additional 6 h at 37°C. The recombinant protein was separated by SDS-PAGE and simultaneously stained with gel code (Pierce). The protein band was excised from the gel and eluted with an electric current. The eluted protein-dye complex was used to immunize rabbits. Anti-Ald4p/6p antibodies were purchased from Rockland. Anti-ADH antibodies have been described previously (9). Anti-PGK antibody was purchased from Molecular Probes. Anti-aminopeptidase I (API) antibodies were our laboratory stock.
Immunoblotting of Total Cell Lysates-Whole cell lysates were prepared by disrupting cells with glass beads in lysis buffer. SDS-PAGE and immunoblotting were performed as described (20).
Pulse-Chase Experiments-The cells were cultured in YPD medium to an A 600 of 1.0 at 30°C and were then washed twice and suspended in SD(ϪN) medium. The cells were pulse-labeled for 30 min by adding 1 MBq of [ 35 S]methionine/1 A 600 unit and chased by adding 0.004% methionine and 0.003% cysteine at 30°C. Immunoprecipitation was performed as described (21).
Light Microscopy-Fluorescence microscopy was performed using a Delta Vision microscope (Applied Precision) as described (22). Subcellular Fractionation and Proteinase K Protection Assay-Yeast cells were converted to spheroplasts by treatment with 25 unit/ml of Zymolyase-100T (Seikagaku Corporation). The spheroplast lysates were separated into cell fractions as described previously (21). To examine proteinase K sensitivity, each fraction without protease inhibitors was treated with 2 mg/ml proteinase K on ice for 30 min with or without 1% Triton X-100. The samples were precipitated with 10% trichloroacetic acid, washed once with cold acetone, resuspended in SDS sample buffer, and analyzed by SDS-PAGE and immunoblotting. Determination of Cell Viability-The cell viability was measured by phloxine B (final concentration 2 g/ml) stain, and fluorescence microscopy was measured with a blue filter. Brightly fluorescent cells were counted as dead cells (2).

Screen for Proteins Reduced under Nitrogen Starvation-We
investigated the expression profiles of soluble proteins before and after nitrogen starvation using two-dimensional PAGE. Using this method, we expected to be able to identify cellular proteins whose levels decreased during nitrogen starvation. The yeast cells were grown at 30°C in YPD medium, were harvested at mid logarithmic phase (A 600 ϭ 1.0), and were washed twice with starvation medium. The cells were then transferred to SD(ϪN) medium and incubated for 24 h. We chose a long stress period of 24 h to observe obvious differences in protein expression; importantly, most of the cells were still viable at this time point (2). Major protein spots on the gel were identified by the peptide mass fingerprinting method using MALDI-TOF mass spectrometry.
In both wild-type (SEY6210) and autophagy-defective ⌬atg7/apg7 (KVY118) yeast cells, most proteins showed little change after starvation (Fig. 1A, lanes 1-4). However, several proteins showed increased levels after starvation, including stress-induced protein, heat shock protein, and quenching enzyme of reactive oxygen species (data not shown). Cytosolic acetaldehyde dehydrogenase (Ald6p) exhibited the most apparent decrease during starvation in wild-type cells (Fig. 1A, lane  5). Therefore, we focused on this protein for further analysis.
Proteins Required for the Reduction of Ald6p under Nitrogen Starvation-Using immunoblot analyses, we attempted to determine which proteins are required for the reduction of Ald6p. In wild-type cells (SEY6210), the amount of Ald6p decreased in a nearly linear manner and was ultimately reduced to 18% of the original level after 24 h of starvation (Fig. 1B). In contrast, Ald6p levels decreased only slightly in ⌬atg7 mutant cells (KVY118). We next investigated whether the amount of Ald6p was reduced in various yeast strains that are defective in various steps of autophagy. ⌬atg7 (KVY118), ⌬atg17/apg17 (JOY617), and all ⌬atg/apg mutant cells tested showed a similar defect in the loss of Ald6p (parts shown in Fig. 1C, lanes  1-6). The decrease of Ald6p also required Ypt7p, a protein that is essential for the fusion of autophagosomes to vacuoles (20), and Pep4p, vacuolar proteinase A (Fig. 1C, lanes 1, 2, and  7-10).
The selective transport of vacuolar enzymes (via the Cvt pathway), such as API and ␣-mannosidase, is known to utilize all of the Apg/Atg proteins except Atg17p (6,23). Atg11p/Cvt9p and Atg19p/Cvt19p function only in the Cvt vesicle formation and do not play a role in autophagosome formation (24,25). In ⌬atg11 (JOY69) and ⌬atg19 mutant cells, Ald6p was reduced in a manner similar to that of wild-type cells under nitrogen starvation (Fig. 1C, lanes 1, 2, 11, and 12; data for ⌬atg19 not shown). As expected, another system of vacuolar transport, the Vid pathway (6,11,26), was not involved in this phenomenon (Fig. 1C, lanes 1, 2, 13, and 14). One mutant allele of the proteasome subunit PRE1 is pre1-1, which is frequently used for the following reasons: the pre1-1 mutation causes a defect in the degradation of short-lived proteins, ubiquitinated proteins (27,28) and N-end rule substrates (29,30) at 30°C. In pre1-1 mutant cells (WCG4 -11a), Ald6p was decreased similarly to wild-type cells (WCG4a) under nitrogen starvation, indicating that Ald6p is not a substrate for proteasome-mediated degradation (Fig. 1D). Taken together, these mutant studies indicate that the reduction of Ald6p requires all of the Atg/Apg proteins and the processes of vacuolar proteolysis. However, Atg/Cvt proteins, Vid proteins, and proteasomal degradation are not involved in this phenomenon.
Reduced Ald6p Levels Implied a Rapid Degradation under Nitrogen Starvation-We hypothesized that the decrease in Ald6p levels was the result of rapid degradation during nitrogen starvation. To examine this possibility, the kinetics of Ald6p degradation was measured by pulse-chase experiments. Wild-type (SEY6210) and ⌬atg7 (KVY118) cells were pulselabeled for 30 min with [ 35 S]methionine and chased with cold methionine and cysteine for 0, 3, 6, and 9 h. In wild-type cells, the Ald6p was rapidly degraded and was barely detectable after 6 h of chase (Fig. 2). In contrast, the degradation rate of Ald6p was clearly slower in ⌬atg7 mutant cells. In addition, ADH, a known nonselective marker of autophagy (9), did not show rapid degradation like Ald6p (Fig. 2). The reduction of Ald6p levels implied a rapid degradation dependent on Atg7p during nitrogen starvation. These results suggest that Ald6p is transported to the vacuole and degraded much more rapidly than typical cytosolic proteins.  2. Pulse-chase analysis of Ald6p and ADH during nitrogen starvation. Wild-type (WT, SEY6210) and ⌬atg7 (KVY118) cells were grown to A 600 ϭ 1.0, preincubated in SD(ϪN) medium for 1 h, pulselabeled with [ 35 S]methionine for 30 min, and chased in SD(ϪN) medium for 0, 3, 6, and 9 h. The lysates were prepared and subjected to immunoprecipitation with anti-Ald6p or anti-ADH serum. The proteins were eluted and analyzed by SDS-PAGE followed by autoradiography.

Visualization of Ald6p under Nitrogen Starvation-
The process of Ald6p vacuolar transport was visualized by expressing physiological levels of an Ald6p-GFP fusion protein from the authentic ALD6 promoter. Upon starvation, the vacuoles gradually became fluorescent. In addition, in ⌬pep4 cells (JOY6005), many bright dots, which were presumably autophagic bodies, were observed moving around in the vacuole (Fig.  3). In ⌬pep4 ⌬atg7 double mutant cells (JOY6006), no fluorescence was observed in the vacuoles, but rather, the cytosol was evenly stained (Fig. 3). Because Ald6p-GFP was transported to the vacuole in autophagic bodies during nitrogen starvation, we hypothesized that transport of Ald6p from the cytosol to the vacuole occurred via the autophagosome.
Ald6p Was Preferentially Transported to the Vacuole via the Autophagosome-We previously reported that ⌬ypt7 cells accumulate autophagosomes in the cytosol under nitrogen starvation (20). Using proform of API as a selective cargo marker of autophagosomes, Ishihara et al. (21) showed the low speed pellet (P13) fraction enriches the autophagosomes. So next we studied the behavior of Ald6p in ⌬ypt7 cells (KVY4). Under growing conditions proform of API was exclusively resided in the high speed supernatant (S100), but under nitrogen starvation conditions a significant portion was recovered in the P13 fraction as reported (Ref. 21 and Fig. 4A). Similarly Ald6p was recovered in the P13 fraction only under nitrogen starvation condition (Fig. 4A). This fraction completely diminished in ⌬ypt7 ⌬atg1/apg1 mutant (YAK1; Fig. 4B, lanes 5-8), indicating that a certain amount of Ald6p is in the autophagosomes. As shown in Fig. 4D, Ald6p and proform of API in P13 fraction were resistant to proteinase K treatment but were digested in the presence of 1% Triton X-100. This also supported the possibility that Ald6p is sequestered into autophagosomes.
We also quantified the amount of Ald6p in the P13 fraction. Proform of API forms one or a few large complex named the Cvt complex in the cytosol and is taken up by an autophagosome at once (31). PGK is shown to be distributed evenly in the autophagosome, autophagic bodies, and cytosol (9). As shown in Fig.  4C, Ald6p translocated to the P13 fraction much more efficiently than PGK (recovery in P13 fraction; Ald6p ϭ 38.2 Ϯ 2.1% n ϭ 5; PGK ϭ 14.9 Ϯ 1.5% n ϭ 5; Fig. 4C) but less than proform of API (67.2 Ϯ 5.9% n ϭ 5). Taken together, we concluded that Ald6p is preferentially sequestered into autophagosome, possibly in a manner different from the substrates for the Cvt pathway.
Ald6p Enzymatic Activity May Be Disadvantageous during Nitrogen Starvation-Next, we examined the physiological relevance of the preferential degradation of Ald6p during nitrogen starvation. Autophagy-defective mutants cannot maintain via-bility under long span nitrogen starvation (2). ⌬atg7 ⌬ald6 mutant (JOY676) cells also started to die after 2 days of nitrogen starvation, but its viability decreased more slowly than  (YAK1; B) cells. Spheroplast lysates were spun at 500 ϫ g for 5 min to remove unbroken cells. Total, total lysate. Total lysates were spun at 13,000 ϫ g for 15 min to separate the pellet (P13) fraction. The supernatant was centrifuged at 100,000 ϫ g for 1 h to generate a pellet (P100) fraction and supernatant (S100) fraction. Each fraction (0.1 A 600 unit) was analyzed by immunoblotting with antibodies to API, Ald6p, or PGK. Growing, growing in YPD medium (A 600 ϭ 1.0); SD(ϪN)12 h, nitrogen-starved in SD(ϪN) medium for 12 h; prAPI, proform of API. C, quantification of proform of API, Ald6p, and PGK in autophagosomeenrich fraction. The recovery of each protein in P13 fraction was calculated. These band intensities were determined using LAS-1000 system (Fujifilm). These data were the averages of five independent experiments. D, proteinase K protection assay of proform of API and Ald6p in the subcellular fractions of ⌬ypt7 cell lysates. Each fraction was treated with 2 mg/ml proteinase K with/or without 1% Triton X-100 on ice for 30 min and then analyzed by immunoblotting with antiserum to API or Ald6p (0.2 A 600 unit/lane). dAPI, the proteinase K digestion product of proform of API. that of ⌬atg7 mutant cells (KVY118; Fig. 5A). The viability of Atg ϩ ⌬ald6 cells (JOY66) also improved slightly from that of wild-type cells (Atg ϩ ALD6; SEY6210) under nitrogen starvation. However, disruption of mitochondrial acetaldehyde dehydrogenase (Ald4p), Atg ϩ ald4 (JOY64), and ⌬atg7 ⌬ald4 (JOY674) cells had no effect on the viabilities of wild-type (SEY6210) and ⌬atg7 (KVY118) cells, respectively (Fig. 5B). Furthermore, wild-type cells (Atg ϩ ) expressing Ald6p via multicopy plasmid showed a defect in the maintenance of viability during nitrogen starvation (Fig. 5C). These results indicate that abundant Ald6p causes the decrease of viability, and the absence of Ald6p improves viability under nitrogen starvation.
We next asked whether Ald6p enzymatic activity or the protein molecule itself is harmful to the cell under nitrogen starvation. To address it we constructed an inactive Ald6p mutant. Farres et al. (32) isolated recombinant ALDH2 C302S from rat liver mitochondrial class-2 aldehyde dehydrogenase (ALDH2). Rat ALDH2-Cys 302 is an active site residue whose thiol group binds to the aldehyde group of the substrate. Ald6p-Cys 306 , which corresponds to Rat ALDH2-Cys 302 , was changed to a serine residue by site-directed mutagenesis. It was confirmed that the mutant form of Ald6p lost completely NADP ϩand Mg 2ϩ -dependent acetaldehyde dehydrogenase activity (data not shown). Overexpression of Ald6p C306S in Atg ϩ and ⌬atg7 cells had no effect on viabilities of Atg ϩ ⌬ald6 and ⌬atg7 ⌬ald6, respectively (Fig. 5D). These results indicate that the acetaldehyde dehydrogenase activity of cytosolic Ald6p may have a disadvantageous effect on the survival of yeast cells during nitrogen starvation. DISCUSSION We surveyed the changes of soluble proteins before and after nitrogen starvation using two-dimensional PAGE. One protein, Ald6p, showed a clear reduction, which was dependent on Atg/ Apg proteins, under nitrogen starvation for 24 h (Fig. 1). Previous morphological studies have indicated that autophagy degrades about 2% of the cytosol/h in yeast (9,10). Scott et al. (33) showed by [ 35 S]methionine pulse-chase experiments that the rate of vacuolar delivery of cytosolic Pho8⌬60p by autophagy was 4%/h during the initial 6 h of nitrogen starvation. Autophagy proceeds linearly during the first 6 h of starvation and then gradually slows (33). We know that both diploid and haploid cells induce autophagy in sporulation medium, 2% potassium acetate (10). In a previous report, Betz and Weiser (34) showed that protein degradation in haploid cells occurred at a slower rate than in diploid cells in sporulation medium. Diploid cells degraded 2.5% of the cellular protein/h in sporulation medium (34). Taken together, these results indicate that most proteins should not decrease below 70% of their original levels because of autophagy, even after 24 h of starvation. In wild-type cells, the amount of Ald6p was reduced to 18% of the initial level after 24 h of nitrogen starvation (Fig. 1B). This large decrease in Ald6p levels reflects preferential autophagic degradation.
The result shown in Fig. 4C indicates that the specificity of Ald6p degradation may be achieved by a step of sequestration to the autophagosome. Suzuki et al. (22,31) indicated that the vacuolar targeting of the proform of API (via the Cvt pathway) required localization with the preautophagosomal structure, which plays a central role in autophagosome formation. In both ⌬atg11/cvt9 and ⌬atg19/cvt19 mutant cells, the proform of API localized to the cytosol away from the preautophagosomal structure and was not targeted to the vacuole (31,35). It was expected that Atg11p and Atg19p would be membrane receptors for the proform of API (24,25). However, Ald6p degradation was not dependent on Atg11p (Fig. 1C, lanes 1, 2, 11, and 12) and Atg19p (data not shown). During nitrogen starvation, the half-life (t1 ⁄2 ) of Ald6p was 100 min (Fig. 2), and the half-life of API was 30 min (6). These results indicate that Ald6p is not likely to be a cargo of the general Cvt pathway. One factor contributing to protein targeting is the existence of a membrane receptor; it is possible that the selective sequestration of Ald6p is mediated by a yet unknown molecule(s) on the autophagosome. Further studies of the molecular mechanisms underlying targeted autophagy are now in progress to investigate these possibilities.
To address the physiological significance of this selective degradation, we analyzed the viability of ⌬ald6 or ALD6 overexpressing cells. We have demonstrated that Ald6p enzymatic activity might be disadvantageous for the survival of yeast cells during nitrogen starvation (Fig. 5). Brejning and Jespersen (36) have previously reported that Ald6p levels increased during lag phase, the first hours after inoculation of the culture. Meaden et al. (13) reported that the growth of ⌬ald6 mutant cells is slower than that of wild-type cells in both YPD and synthetic medium. It is known that acetaldehyde dehydrogenase is closely related to lipid biosynthesis through the intermediary of acetyl-CoA synthase and fatty acid synthase in the cytosol. Because lipid biosynthesis is a critical process, the expression of Ald6p would be necessary during growth under nutrient conditions. Why is cytosolic Ald6p acetaldehyde dehydrogenase activity harmful under nitrogen starvation conditions? One possible explanations might be that Ald6p may disturb NADPH flux during nitrogen starvation. It is well known that glucose-6-phosphate dehydrogenase (Zwf1p; glucose-6-phosphate ϩ NADP ϩ 3 6-phosphogluconolactone ϩ NADPH) is the greatest contributor to the reduction of NADP ϩ in the yeast cell. Grabowska and Chelstowska (37) have recently demonstrated that ⌬ald6 ⌬zwf1 double mutant cells are not viable under normal growth conditions or under anaerobic growth conditions even in the presence of glutathione. It is suggested that Ald6p plays an important role in maintaining a high rate of NADPH/NADP ϩ cycling in the yeast cell. However, upon nitrogen starvation, both fatty acid and deoxyribonucleoside biosynthesis, which consume large amounts of NADPH, shut down immediately (38,39). We speculate that the reduction of NADP ϩ by Ald6p might be excessive in nitrogen-starved cells. An excessive amount of NADPH might inhibit the enzymatic activity of Zwf1p, which catalyzes the initial reaction of the pentose phosphate pathway. This pathway contributes to the synthesis of ribose-5-phosphate, which is an essential material for the generation of some amino acids and ribonucleotides (39). Ald4p, the mitochondrial acetaldehyde dehydrogenase, utilizes mainly NAD ϩ as a co-enzyme (14) and is induced during nitrogen starvation (38). In our experiments, ⌬atg7 ⌬ald4 mutant cells were not able to maintain high rates of viability like ⌬atg7 ⌬ald6 cells (Fig. 5B). It is likely that the down-regulation of Ald6p by preferential autophagic degradation may optimize NADPH/NADP ϩ levels in the cytosol. Thus, Ald6p may have a bilateral character: it is beneficial in growth under nutrient conditions but disadvantageous to survival under nitrogen starvation.
Here, we show that Ald6p is one example of a preferential substrate for autophagic degradation. Ald6p was the only major protein on the two-dimensional PAGE gel to decrease during starvation; however, it is still possible that other minor proteins behave like Ald6p. If we are able to find such proteins, it would help clarify the molecular mechanisms of selective autophagy and the physiological significance of the preferential degradation.