Functional characterization and localization of acetyl-CoA hydrolase, Ach1p, in Saccharomyces cerevisiae.

Acetyl-CoA hydrolase (Ach1p), catalyzing the hydrolysis of acetyl-CoA, is presumably involved in regulating intracellular acetyl-CoA or CoASH pools; however, its intracellular functions and distribution remain to be established. Using site-directed mutagenesis analysis, we demonstrated that the enzymatic activity of Ach1p is dependent upon its putative acetyl-CoA binding sites. The ach1 mutant causes a growth defect in acetate but not in other non-fermentable carbon sources, suggesting that Ach1p is not involved in mitochondrial biogenesis. Overexpression of Ach1p, but not constructs containing acetyl-CoA binding site mutations, in ach1-1 complemented the defect of acetate utilization. By subcellular fractionation, most of the Ach1p in yeast was distributed with mitochondria and little Ach1p in the cytoplasm. By immunofluorescence microscopy, we show that Ach1p and acetyl-CoA binding site-mutated constructs, but not its N-terminal deleted construct, are localized in mitochondria. Moreover, the onset of pseudohyphal development in homozygote ach1-1 diploids was abolished. We infer that Ach1p may be involved in a novel acetyl-CoA biogenesis and/or acetate utilization in mitochondria and thereby indirectly affect pseudohyphal development in yeast.

The concentration of acetyl-CoA in cells is primarily regulated by its rate of synthesis and its utilization in various metabolic pathways. In the yeast Saccharomyces cerevisiae, biosynthesis of acetyl-CoA is mainly achieved by the acetyl-CoA synthetase reaction, whereas oxidative decarboxylation by the mitochondrial pyruvate dehydrogenase complex appears to be of minor importance (reviewed in Ref. 1). Even under glycolytic growth conditions, S. cerevisiae converts pyruvate into acetate, catalyzed by the subsequent action of pyruvate decarboxylase, acetaldehyde dehydrogenase, and acetyl-CoA synthetase (2). In the presence of a fermentable carbon source, acetyl-CoA may be mainly used as a precursor of fatty acid and sterol biosynthesis. On the other hand, an additional pool of acetyl-CoA is required for the glyoxylate cycle (citrate synthase and malate synthase reactions) when cells grow with a non-fermentable substrate such as ethanol or acetate.
Acetyl-CoA hydrolase, catalyzing the hydrolysis of acetyl-CoA, was first identified in the pig heart (3), and subsequently the enzyme has been found in many mammalian tissues (4 -10). During the purification of yeast N␣-acetyltransferase, an endogenous "inhibitor" of acetyltransferase was purified and shown to be acetyl-CoA hydrolase (11,12). Acetyl-CoA hydrolase also inhibits purified rat brain pyruvate carboxylase (13) and [acyl-carrier-protein]acetyltransferase (14). It has been shown that the expression of acetyl-CoA hydrolase (ACH1) from S. cerevisiae is glucose-repressible (15) and subjected to cAMP-dependent repression (16). The function of Ach1p 1 in vivo is still speculative. Previously, we have shown that the ability of ach1 mutants to grow on acetate is impaired (17). ACH1 is highly homologous to the aarC gene of Acetobacter aceti (18) and the Neurospora crassa gene acu8 (19,20). An acu-8 mutant strain, characterized as acetate non-utilizing, shows strong growth inhibition by acetate but will use it as a carbon source at low concentrations (20). The acu-8 mutant was also shown to be deficient in acetyl-CoA hydrolase and to accumulate acetyl-CoA when supplied with acetate. As in Saccharomyces, the Neurospora enzyme is acetate-inducible. The arrC-defective mutant also showed an inability to assimilate acetic acid (18). However, in all three organisms, disruption of these genes yields strains that grow normally on ethanol (17,18,20). Possibly, the acetyl-CoA balance during growth on acetate is disturbed in such mutants. Whether or not acetyl-CoA hydrolase is involved in regulating the endogenous pool(s) of acetyl-CoA remains to be established. In this study, we took an initial step to characterize the biochemical property of Ach1p in vivo and determine its subcellular localization. We demonstrate that the enzymatic activity of Ach1p is dependent upon its putative nucleotide (CoA) binding sites and show that Ach1p is a mitochondrial enzyme. In addition, we provide initial evidence that Ach1p is involved in development of pseudohyphae but not in mitochondrial biogenesis.

Strains, Media, and Microbiological
Techniques-Yeast culture media were prepared as described by Sherman et al. (21). YPD contained 1% Bacto-yeast extract, 2% Bacto-peptone, and 2% glucose. SD contained 0.7% Difco yeast nitrogen base (without amino acids) and 2% glucose. Nutrients essential for auxotrophic strains were supplied at specified concentrations. For comparison of Ach1p expression in different carbon sources, synthetic media containing 5% glucose, 2% galactose, 2% glycerol, and 2% potassium acetate were used. Yeast cells were transformed by the lithium acetate method (22). Plasmids were constructed by standard protocols as described by Sambrook  SstI-XbaI site of the yeast expression vector pVT101U, a 2--based expression plasmid containing the ADH1 promoter (24). ACH1 constructs containing truncations of the N-terminal region (Ach1pdN, deletion of 1-64 amino acid residues), the first mutation of the putative CoA binding site (Ach1pSS, substitution of amino acids Gly-277 and  Gly-279 by Ser-277 and Ser-279), and the second mutation of the putative CoA binding site (Ach1pES, substitution of amino acids Gly-393 and Gly-395 by Glu-393 and Ser-395) were made. All mutations were generated by PCR-based mutagenesis. The sequences of the resulting constructs were verified by sequencing.
Expression and Purification of Recombinant Proteins and Polyclonal Antibody Production-The open reading frame of ACH1 was obtained by PCR, by the use of primers that incorporated unique NcoI and BamHI sites at the initiating methionine and 6 bp downstream of the stop codon, respectively. For the His-tagged Ach1p, a DNA fragment containing the ACH1 coding region was generated by amplifying of yeast genomic DNA with sequence-specific primers. The PCR product was purified and ligated to the expression vector pET15b (Novagen), yielding pET15bACH1. The His-tagged fusion protein was synthesized in BL21(DE3) Escherichia coli and purified on nickel-nitrilotriacetic acid resin (Qiagen, Chatsworth, CA) as described (25). Denatured, purified recombinant Ach1p isolated from an SDS-PAGE gel was used as antigen for raising polyclonal antibodies in rabbits essentially as described (25).
Preparation of Crude Yeast Lysates and Assay of Acetyl-CoA Hydrolase Activity-Crude yeast lysates were prepared, and acetyl-CoA hydrolase activity was determined by radioactive assay, as described previously (15). One unit of activity is defined as the amount of enzyme that hydrolyzes 1 nmol of [1-14 C]acetyl-CoA in 1 min.
Western Blot Analysis and Immunofluorescence Microscopy-Yeast total proteins were prepared and subjected to Western blot analysis as described previously (26). Cells were prepared for immunofluorescence staining as described by Huang et al. (25). Alexa 594-or Alexa 488conjugated anti-IgG antibodies (Molecular Probes, Eugene, OR) were used as secondary antibodies. H33258 was diluted in mounting solution for nucleic acid staining. Fluorescence microscopy was performed with a Nikon Microphot SA microscope.
The monoclonal anti-yeast mitochondria porin antibodies were purchased from Molecular Probes. The monoclonal anti-yeast ribosomal protein Rpl3p antibody (also called anti-TCM1 antibody) was a gift of Dr. J. Warner (Albert Einstein College of Medicine, Bronx, NY), and the polyclonal anti-Kar2p antibody was a gift from Dr. M. Rose (Princeton University, Princeton, NJ). Polyclonal anti-porin antibody was diluted 1:20,000 for Western blot analysis and 1:5000 for immunofluorescence staining. The polyclonal antibodies against yeast Arf1p were generated by the use of recombinant proteins from our laboratory (26).
Subcellular Fractionation-Yeasts grown in selective minimal medium or YPD medium were harvested by centrifugation and washed once with 10 mM NaN 3 , before Lyticase digestion of cell walls in a solution of 1.2 M sorbitol and 100 mM potassium phosphate, pH 6.5. Spheroplasts were suspended in buffer containing 0.1 M sorbitol, 20 mM HEPES (pH 7.4), 50 mM potassium acetate, and 1 mM EDTA with protease inhibitors and disrupted on ice with 20 strokes in a Dounce homogenizer. The lysate was centrifuged (450 ϫ g) to remove debris and unbroken cells. Cleared lysate (0.8 ml) was loaded on top of a manually generated six-step sucrose gradient (0.7 ml each of 60, 50, 40, 30, 20, and 10% sucrose in lysis buffer), which was then centrifuged at 170,000 ϫ g for 3 h in a Beckman SW55 rotor at 4°C. Proteins in samples (100 l) of fractions, collected manually from the top, were precipitated with 10% trichloroacetic acid, separated by SDS-PAGE, and analyzed by immunoblotting. Diluted antibodies against mitochondrial porin (1:500) (Molecular Probes), Kar2 (1:1000), Emp47 (1:5000), and Arf1p (1:5000) (26) were used to identify organelles.
Nycodenz Gradients-The medium formula, cell treatments, and fractionated centrifugation were performed as described (27)(28)(29)(30) and with some modifications. Wild-type yeast cells were cultured in 50 ml of synthetic medium containing 2% glucose at 30°C with shaking overnight. Then the overnight culture was harvested and transferred into 100 ml of oleic acid-containing medium (0.3% yeast extract, 0.3% peptone, 0.1% oleic acid, 0.2% Tween 40, and 0.5% potassium phosphate) to induce the formation of peroxisomes. After 24 h, the yeast cells were harvested and washed once with 0.1 M potassium phosphate buffer (pH 6.3). Then cells were suspended in 3 ml of spheroplast solution (1.2 M sorbitol, 0.1 M potassium phosphate, pH 6.3), 150 l of ␤-mercaptoethanol was added, and the cells were kept at room temperature for 15 min with gentle rocking. After spinning down and resuspending in 3 ml of fresh spheroplast solution containing 30 l of ␤-mercaptoethanol, 1,000 units of lyticase were added. The cells were mixed gently and incubated at 30°C for generation of spheroplasts. The spheroplasts were spun down, washed once in ice-cold spheroplast solution, and finally suspended in lysis buffer (0.6 M sorbitol, 5 mM MES, pH 6.0, 1 mM KCl, and 0.5 mM EDTA). The spheroplasts were lysed by passage through a 26-gauge needle 20 times and incubated on ice for 30 min. Unbroken cells and nuclei were removed by centrifugation at 1,500 ϫ g for 5 min. The supernatant was centrifuged at 15,000 rpm for 15 min. The crude organelle pellet, consisting mainly of mitochondria and peroxisomes, was gently resuspended in lysis buffer and centrifuged at low speed (600 ϫ g) to remove larger aggregates. The organelle suspension was loaded on a 14 -36% Nycodenz gradient to further fractionate mitochondria and peroxisomes. The gradient was centrifuged at 32,500 rpm, 4°C, for 3.5 h. After centrifugation, the sample was divided into 14 fractions from 14 to 36% Nycodenz gradients; then proteins were precipitated by 10% trichloroacetic acid and analyzed by Western blotting.
Isolation and Fractionation of Mitochondria-Yeast was grown in YPGal to early stationary phase. Mitochondria were isolated as described previously (31). Mitochondria were resuspended in lysis buffer to give an approximate final concentration of 10 mg of protein/ml. To isolate the mitochondrial intermembrane space, a suspension of mitochondria (10 -20 mg of protein/ml in 0.6 M sorbitol, 10 mM Tris, pH 7.4) was diluted with 5 volumes of 10 mM Tris, pH 7.4, to a final sorbitol concentration of 0.1 M. The suspension was incubated at 4°C with gentle rocking for 20 min. The "shocked" mitochondria were sedimented at 20,000 rpm in a Beckman SW55 Ti rotor for 20 min. The supernatant contains the contents of the intermembrane space; then proteins were precipitated by 10% trichloroacetic acid and analyzed by Western blotting. To isolate mitochondrial membrane and matrix, shocked mitochondria were resuspended in 10 mM Tris, pH 7.4, to a protein concentration of about 2 mg/ml with five strokes in a Dounce homogenizer and left on ice to allow further swelling of the mitochondrial matrix space. After 5 min, "shrinking buffer" (one-third of the suspension volume) containing 1.8 M sucrose, 8 mM ATP, 8 mM MgC1 2 , adjusted to pH 7.4 with KOH, was added. The suspension was mixed carefully by three strokes in the Dounce homogenizer and left on ice. After 5 min, the suspension was exposed to ultrasonic irradiation for 3 ϫ 5 s on ice. Total mitochondrial membranes were sedimented for 60 min at 35,000 rpm in a Beckman SW55 Ti rotor at 4°C. The supernatant represents the matrix fraction; then proteins were precipitated by 10% trichloroacetic acid and analyzed by Western blotting.
FIG. 1. Western blot analysis and enzyme activity of Ach1p and its mutants. As shown in A, wild-type (WT) yeast, achl mutant, and achl mutant with expressed Ach1p, Ach1pdN, Ach1pES, and Ach1pSS were cultured with synthetic medium containing 2% glucose. Acetyl-CoA hydrolase activity was determined by radioactive assay, as described previously (15). One unit of activity is defined as the amount of enzyme that hydrolyzes 1 nmol of [1-14 C]acetyl-CoA in 1 min. As shown in B, wild-type yeast was cultured with synthetic medium containing 5% glucose, 2% galactose, 2% glycerol, or 2% potassium acetate. Total proteins (ϳ20 g/lane) were separated by SDS-PAGE, stained with Coomassie Blue (upper panels), or subjected to Western blot analysis (lower panels). Ach1p were identified with specific antibody (1:5000). The yeast Arf1p was used as internal control. Positions of protein standards are indicated on the left.

RESULTS
Identification of Endogenous Ach1p-To characterize the ACH1 gene product, we prepared a rabbit antiserum against an E. coli synthesized recombinant full-length His-tagged Ach1p fusion protein. Among total cellular proteins, antibodies prepared against Ach1p reacted only with a protein of ϳ64 kDa, the expected size for Ach1p (Fig. 1). This protein was not detected in an ach1 mutant ( Fig. 1) or by the preimmune serum (not shown). Immunoblotting with this antiserum detected nanogram amounts of Ach1p (data not shown) as well as various mutant forms of Ach1p (Fig. 1A). As in previous RNA blot analysis (15), Ach1p was subjected to glucose-dependent repression (Fig. 1B).
Putative Nucleotide (CoA) Binding Site Is Required for Ach1p Activity-A data base search showed significant homologies among Ach1p and other CoA-transferases, including Schizosaccharomyces pombe ACH1 (SpACH1; 64% identity), N. crassa Acu8 (NcAcu8; 57% identity), E. coli ACH1-like (EcCat1, Gen-Bank TM accession U28377; 38% identity), A. aceti AarC (AarC; 39% identity), Clostridium kluyveri CAT1 (CkCat1, succinyl-CoA:coenzyme A transferase; 37% identity), and C. kluyveri CAT2 (CkCat2, butyryl-CoA-acetate Coenzyme A; 20% identity). The amino acid sequence of the CoA (ADP) binding site (GXGXX(G/A)) was reported from an analysis of the known three-dimensional structures of ADP binding ␤-␣-␤-folds (32). Fig. 2 shows that the conserved CoA (ADP) binding site (GXGXX(G/A)) from heterodimeric CoA transferases are present in the amino acid sequences of Ach1p and homologous CoA-transferases (33,34). To determine whether these putative nucleotide (CoA) binding sites are required for Ach1p activity, we generated two mutants, Ach1pSS and Ach1pES, by site-directed mutagenesis (as described under "Materials and Methods"). Wild-type Ach1p, Ach1pSS, and Ach1pES were expressed in the ach1 mutant, and their expression was confirmed by Western blotting with anti-Ach1p antibody (Fig. 1A). Enzyme assays of protein extracts from these strains confirmed that Ach1pSS and Ach1pES expressed in the ach1 mutant contained little (Ͻ5%) detectable acetyl-CoA hydrolyzing activity, whereas the untransformed wild type had normal enzyme activity (Fig. 1A). In addition, in comparison with wild-type yeast, overexpressing Ach1p in the ach1 mutant had ϳ3.8-fold of activity. Because of the low solubility of E. coli producing Ach1p recombinant proteins, we failed to isolate recombinant Ach1p and its mutant constructs to assess their enzymatic activity. Previously, we have shown that the ability of ach1 mutants to grow on acetate is impaired (17). We next tested whether overexpressed Ach1p and its mutant constructs have biological function in ach1 mutant yeast. Specific growth rates of tested strains (wild-type strain, ach1 strain, ach1 strains overexpressing Ach1p, Ach1pES, and Ach1pSS) were obtained by the growth of cells in the synthetic medium containing acetate, and A 600 values were determined at specific time intervals. Overexpressing Ach1p in ach1 mutant can grow on acetate medium; however, the ability of the ach1 mutant and strains overexpressing Ach1pES and Ach1pSS to grow on acetate is impaired (data not shown). These data suggest that Ach1pES and Ach1pSS lose their biological activity in vivo. Subcellular Localization of Endogenous Ach1p-To study the subcellular distribution of Ach1p, the total yeast spheroplast-homogenized lysate was fractionated by 30 -60% discontinuous sucrose gradient centrifugation. As shown in Fig. 3A, distribution of most of Ach1p was similar to that of the mitochondrial protein porin, although little Ach1p was found in cytoplasmic fractions. To determine further whether Ach1p may also be present in peroxisomes, a homogenate of oleategrown cells was first subjected to differential centrifugation to obtain an organellar pellet. This material was further fractionated by density gradient centrifugation on Nycodenz. Fig. 3B shows good resolution between mitochondria (porin marker) and peroxisomes (thiolase marker), and the distribution of Ach1p was similar to that of the mitochondrial protein porin. By immunofluorescence microscopy, endogenous Ach1p, similar to porin, appeared to be localized to mitochondria (Fig. 3C). Yeast mitochondria can form branched networks distributed evenly around the circumference of the cell in the peripheral cytoplasm. Abnormal mitochondrial morphology was not seen in ach1 mutant yeast. To further localize the Ach1p within the purified mitochondria, we analyzed the intermembrane space, matrix, and membrane fractions (Fig. 4). Ach1p cofractionated with Mge1p but not with cytochrome oxidase subunit IV and porin, indicating that it is localized to the mitochondrial matrix.
N terminus but Not Putative Nucleotide (CoA) Binding Sites of Ach1p Are Required for Ach1p Mitochondrial Localization- Most proteins targeted to the mitochondrial matrix contain a cleavable N-terminal presequence with basic and hydroxylated amino acids interspersed throughout their length (35,36). Although the N terminus of Ach1p contains no typical matrixtargeting sequence, we suspected that deletion of the N-terminal domain from Ach1p might interfere with its mitochondrial localization. After expression in the ach1 mutant, Ach1pdN, lacking 64 amino acids at the N terminus, was recovered in the least dense fractions of the lysate (data not shown). In cells, most of the Ach1pdN mutant was in the cytoplasmic region with a punctate distribution (Fig. 5), and the mitochondrial morphology was similar to that in wild-type cells (Fig. 5). In addition, fusion between the N terminus (64 amino acids) of Ach1p and GFP protein failed to be imported into mitochondria (data not shown).
We further tested whether overexpression of Ach1p or its mutant constructs (Ach1pES and Ach1pSS) in yeast might cause dominant-negative effects on mitochondrial morphology. By immunofluorescence microscopy, Ach1pES and Ach1pSS, like overexpressed Ach1p, were present in some tubular or spherical structures that also stained with anti-porin antibody (Fig. 5). Thus, N terminus, but not putative Co-A binding sites of Ach1p, was required for Ach1p mitochondrial localization.
Ach1p Is Not Required for Mitochondrial Biogenesis-Be-cause most of Ach1p is localized in the mitochondria, we next examined whether Ach1p can affect mitochondrial function. Yeast, when cultured in glycerol medium, requires mature active mitochondria for oxidative metabolism and growth. Yeast grown in glucose initially does not need active, mature mitochondria, and the mitochondria are not well developed before being switched from anaerobic glucose fermentation to aerobic ethanol oxidation. The wild-type strain, ach1 strain, and ach1 strains overexpressing Ach1p and its deleted or mutated constructs (Ach1pES, Ach1pSS, and Ach1pdN) were cultured overnight in synthetic medium containing glucose. Cells were harvested, suspended in double distilled H 2 O, subjected to serial dilution, and dropped onto glucose, acetate, glycerol, succinate, and ethanol media plates. Fig. 6 shows that all strains grew on the glucose, glycerol, succinate, and ethanol media plates; however, the ach1 mutant and the Ach1pES, Ach1pSS, and Ach1pdN expression strains did not grow in acetate medium plates. Moreover, the Ach1p-overexpressing cells grew as well as the wild-type yeast on acetate plates. These results suggested that Ach1p is not involved in cellular events in mitochondrial biogenesis, which is required for cells to grow in non-fermentable carbon sources. Ach1p Is Required for Pseudohyphal Formation-Diploid cells of the yeast S. cerevisiae undergo pseudohyphal differen- FIG. 3. Ach1p is localized to mitochondria. A, sucrose gradient centrifugation analyses of Ach1p. Wild-type yeast cell lysate was subjected to centrifugation in a discontinuous 30 -60% sucrose gradient. Fractions were collected, and proteins were separated by SDS-PAGE followed by immunoblot analysis by the use of antibodies against proteins indicated on the left of each panel. Kar2p, endoplasmic reticulum protein; Porin, mitochondrial protein; Arf1p, cytoplasmic protein. B, subcellular distribution of Ach1p in oleate-grown S. cerevisiae. The organelle pellet (P15) of oleate-grown cells was prepared by differential centrifugation and then analyzed by Nycodenz-gradient centrifugation. Fractions 1-14 correspond to the 14 -36% Nycodenz gradients. The Ach1 distribution was detected by a polyclonal anti-Ach1 antibody at a 1:5000 dilution. The locations of mitochondria and peroxisomes were detected by anti-mitochondrial porin antibody and anti-thiolase antibody at 1:1000 and 1:2000 dilutions, respectively. C, immunofluorescence staining of Ach1p. Wild-type (WT) and achl mutant were fixed with formaldehyde; spheroplasts were prepared and reacted with anti-Achlp antibody (1:1000) and anti-mitochondrial porin antibody (1: 50) followed by secondary antibodies. Nucleic acids were stained with H33258. tiation in response to nutrient limitation (37). We have characterized the connection between acetyl-CoA changes and pseudohyphal growth. Wild-type, ach1/ach1 mutant, and ach1/ ach1-overexpressing Ach1p yeast cells were grown in low ammonium sulfate (SLAD; 50 M) medium to characterize their pseudohyphal differentiation. Importantly, we found that cells lacking the Ach1p were completely defective in pseudohyphal differentiation, whereas ach1/ach1-overexpressing Ach1p restores pseudohyphal growth (Fig. 7). However, expression of Ach1p deleted or mutated constructs (Ach1pES, Ach1pSS, and Ach1pdN) failed to restore pseudohyphal growth (data not shown). DISCUSSION In this study, we show that putative conserved nucleotide (CoA) binding sites and N terminus of Ach1p require its enzyme activity. Our data also show that Ach1p is localized to the mitochondria, and the N terminus of Ach1p is required for its localization. Finally, we show that Ach1p is not involved in mitochondrial biogenesis but may be involved in pseudohyphal differentiation.
The amino acid sequences of Ach1p and homologous CoAtransferases contain two conserved CoA (ADP) binding sites (GXGXX(G/A)) from heterodimeric CoA transferases (Fig. 2) (9,33). We determined whether these putative nucleotide (CoA) binding sites are required for Ach1p activity. Enzyme assays confirmed that Ach1pSS and Ach1pES expressed in the ach1 mutant contained little detectable acetyl-CoA hydrolyzing activity. We also showed that utilization of acetate as carbon source by the ach1 mutant is impaired, and overexpression of Ach1pES, or Ach1pSS, cannot restore this activity. Our data indicate that one or more putative nucleotide (CoA) binding sites are required for Ach1p enzymatic activity.
The subcellular localization of a protein is an important characteristic with functional implications. Our data show that most of Ach1p is localized to the mitochondria, although a little Ach1p was also found in small punctate form distributed in the cytoplasm. The majority of mitochondrial matrix-targeting signals are cleaved upon import into the mitochondria (36). The matrix-located MTF1 protein in yeast, which is a transcriptionstimulating factor (38), is an exceptional case. This protein lacks a recognizable matrix-targeting sequence, and its import is reported to be independent of outer membrane receptors. How specificity of targeting is achieved in this case and whether there is an entirely separate pathway for importing this protein remain to be clarified. We demonstrate that Nterminal but not putative CoA binding sites of Ach1p are required for localization to mitochondria. Thus, translocation of Ach1p from the cytoplasm to the mitochondrial matrix may, like that of MTF1, require that the N-terminal sequence lacks a recognizable matrix-targeting signal. A recent report described that Ach1p has two different protein spots by twodimensional gel electrophoresis (39). These two Ach1ps have the same relative molecular weight but differ in their pI, suggesting that Ach1p might be modified post-translationally. However, we attempted, but failed, to confirm that there are two different protein spots of Ach1p by two-dimensional gel electrophoresis. 2 Carnitine acetyltransferase (CAT) is known to be present in mitochondria and peroxisomes of oleate-grown S. cerevisiae, and both proteins are encoded by the same gene, YCAT (27). We also speculated whether Ach1p in oleate-grown cells might have a different subcellular localization. Our data showed that the majority of Ach1p is present in mitochondria but not in peroxisomes. Thus, we concluded that Ach1p is a mitochondrial enzyme and may execute its physiologic function in the matrix space.
Contemporary knowledge of the structure and function of acetyl-CoA hydrolases (i.e. cytosolic (8,40) and mitochondrial (5)) is incomplete. Ach1p resembles the rat mitochondrial acetyl-CoA hydrolase, is not affected by ADP or ATP, and is inhibited by ␤NADH (9, 12). The FIG. 5. The N terminus, but not putative CoA binding sites of Ach1p, is required for localization to mitochondria. The yeast achl mutant cells were transformed with pVTl0lU-ACHl, pVTl0lU-ACHldN, pVTl0lU-ACH1ES, and pVTl0lU-ACH1SS. Cells were fixed with formaldehyde; spheroplasts were prepared and reacted with anti-Achlp antibody (1:1000) and anti-mitochondrial porin antibody (1:50) followed by secondary antibodies. Nucleic acids were stained with H33258. the mitochondrial acetyl-CoA hydrolase from hamster brown fat (6,12). In addition, the pH optima for Ach1p and rat brain mitochondrial acetyl-CoA hydrolase are identical (pH ϳ8) (5). Our data demonstrate that Ach1p is localized to mitochondria, consistent with the previous finding that the biochemical properties of Ach1p is similar to those of mitochondrial acetyl-CoA hydrolases.
Diploid cells of the yeast S. cerevisiae undergo pseudohyphal differentiation in response to nutrient limitation (37). Studies on the way in which nitrogen and carbon starvation induce invasive and filamentous growth suggest multiple regulatory points in each pathway and cross-talk between the pathways (41). Changing any single component may rearrange metabolic fluxes in a manner that is difficult to predict. New approaches to metabolic system modeling and design are likely to contribute to identifying the particular components that signal invasion and filamentation and offer predictions for how to regulate the intracellular activities of those components. Our study showed that Ach1p was required for pseudohyphal formation, suggesting a physiologic connection between acetyl-CoA changes and pseudohyphal growth.
In recent studies, sequence alignment of genes with similar regulation patterns revealed a putative regulatory promoter element (CCWTTSRNCCG) for the glyoxylate cycle (42). This specific element was present in seven genes, including CIT2 (citrate synthase in peroxisomal matrix), ICL1 (isocitrate lyase), MLS1 (malate synthase in peroxisomal matrix), MDH2 (malate dehydrogenase in peroxisomal matrix), CAT2 (carnitine acetyltransferase in peroxisomal matrix and mitochon-dria), ACR1 (succinate-fumarate transporter in mitochondrial inner membrane), and ACH1, which were derepressed on ethanol or acetate. Consistent with this observation, three glyoxylate cycle genes, ICL1, MLS1, and MDH2, showed the same regulation pattern as ACH1 (42). In addition, acu-8 mycelium exhibited no significant flux through the glyoxylate cycle 10 h after transfer to acetate. Thus, it is reasonable to speculate that Ach1p may be involved in the glyoxylate cycle. Recently, Lorenz and Fink (43) showed that live S. cerevisiae cells isolated from the phagolysosome are induced for genes of the glyoxylate cycle. These findings in fungi, in conjunction with reports that isocitrate lyase is both up-regulated and required for the virulence of Mycobacterium tuberculosis, demonstrate the wide ranging significance of the glyoxylate cycle in microbial pathogenesis. It will be interesting to learn whether Ach1p, similar to genes involved in the glyoxylate cycle, can be induced in phagolysosomes.
The role of acetyl-CoA hydrolases catalyzing the scission of the high energy thioester bond acetyl-CoA with no apparent metabolic advantage represents a biochemical conundrum. Because acetyl-CoA hydrolase is highly expressed in yeast when media contain acetate, we suspect that, under conditions when acetate is used as the main carbon source, a large amount of acetyl-CoA is generated but not effectively incorporated into the trichloroacetic acid or glyoxylate cycle. Such an excess of acetyl-CoA could lead to autoacetylation of proteins, as well as to the generation of toxic ketone bodies or other noxious metabolites. It is possible, albeit not yet established, that the intracellular level of acetyl-CoA could be regulated at a "safe" level by hydrolysis of excessive acetyl-CoA by acetyl-CoA hydrolase. Another interpretation is that the ACH1 enzyme, in vivo, is regulated by its associated factor, which can alter the Ach1p enzyme to be an acetyltransferase. It has been shown that the distinction between acyltransferases and thioesterases is quite narrow (44). Moreover, acetyltransferases, in the absence of the acetyl acceptor, can transfer the acetyl group from acetyl-CoA to water and act as hydrolases in vitro. Interestingly, a recent proteomic analysis showed that exposure of S. cerevisiae to sorbic acid in YEPD medium, pH 4.5, resulted in the up-regulation of 10 proteins, including Ach1p (45), suggesting that the induction of Ach1p may confer resistance to the inhibitory effects of sorbic acid. Furthermore, Ach1p was indicated to have high homologies (42% identity and 62% similarity) to C. kluyveri CAT1 (CkCat1, succinyl-CoA:coenzyme A transferase) and was suggested to be a succinyl-CoA:CoA transferase. However, the specific enzymatic activity of C. kluyveri CAT1 in recombinant E. coli clones was very low (9). We attempted, but failed, to prove that Ach1p has such enzymatic activity. 2 In conclusion, this study has confirmed that putative conserved nucleotide (CoA) binding sites of Ach1p are required for its enzyme activity in vivo. We also demonstrated that Ach1p is a mitochondrial enzyme, although its potential function in the glyoxylate cycle needs to be investigated further.
We also showed that Ach1p is not involved in mitochondrial biogenesis, and our data suggest that the metabolism of acetyl-CoA by Ach1p is involved indirectly in pseudohyphal differentiation. Although yeast can use acetate or ethanol as carbon source by converting them to acetyl-CoA in the metabolic pathway, it will be interesting to know how ach1 mutants could impair acetate but not ethanol utilization. The exact physiologic role of this mitochondrial Ach1p needs to be investigated further.