Expression and function of a mislocalized form of peroxisomal malate dehydrogenase (MDH3) in yeast.

The malate dehydrogenase isozyme MDH3 of Saccharomyces cerevisiae was found to be localized to peroxisomes by cellular fractionation and density gradient centrifugation. However, unlike other yeast peroxisomal enzymes that function in the glyoxylate pathway, MDH3 was found to be refractory to catabolite inactivation, i.e. to rapid inactivation and degradation following glucose addition. To examine the structural requirements for organellar localization, the Ser-Lys-Leu carboxyl-terminal tripeptide, a common motif for localization of peroxisomal proteins, was removed by mutagenesis of the MDH3 gene. This resulted in cytosolic localization of MDH3 in yeast transformants. To examine structural requirements for catabolite inactivation, a 12-residue amino-terminal extension from the yeast cytosolic MDH2 isozyme was added to the amino termini of the peroxisomal and mislocalized “cytosolic” forms of MDH3. This extension was previously shown to be essential for catabolite inactivation of MDH2 but failed to confer this property to MDH3. The mislocalized cytosolic forms of MDH3 were found to be catalytically active and competent for metabolic functions normally provided by MDH2.

The malate dehydrogenase isozyme MDH3 of Saccharomyces cerevisiae was found to be localized to peroxisomes by cellular fractionation and density gradient centrifugation. However, unlike other yeast peroxisomal enzymes that function in the glyoxylate pathway, MDH3 was found to be refractory to catabolite inactivation, i.e. to rapid inactivation and degradation following glucose addition. To examine the structural requirements for organellar localization, the Ser-Lys-Leu carboxyl-terminal tripeptide, a common motif for localization of peroxisomal proteins, was removed by mutagenesis of the MDH3 gene. This resulted in cytosolic localization of MDH3 in yeast transformants. To examine structural requirements for catabolite inactivation, a 12-residue amino-terminal extension from the yeast cytosolic MDH2 isozyme was added to the amino termini of the peroxisomal and mislocalized "cytosolic" forms of MDH3. This extension was previously shown to be essential for catabolite inactivation of MDH2 but failed to confer this property to MDH3. The mislocalized cytosolic forms of MDH3 were found to be catalytically active and competent for metabolic functions normally provided by MDH2.
Differentially compartmentalized isozymes of malate dehydrogenase in eucaryotic cells catalyze the NAD(H)-dependent interconversion of oxaloacetate and malate. In mammalian cells this reaction, catalyzed by mitochondrial and cytosolic isozymes, respectively, is a critical step in the tricarboxylic acid cycle and in gluconeogenesis. The two isozymes also participate in the malate/aspartate shuttle cycle, a mechanism for exchange of reducing equivalents between cellular compartments. In yeast and plant cells, a third isozyme localized in peroxisomes catalyzes a step in the glyoxylate pathway. This pathway allows formation of C 4 metabolites from C 2 precursors. The malate dehydrogenase isozyme family is therefore an ideal focus for analysis of structural features responsible for differential compartmentation and metabolic function.
To initiate molecular genetic studies, the three isozymes of malate dehydrogenase have been purified from Saccharomyces cerevisiae, and the corresponding genes have been cloned, sequenced, and disrupted (1)(2)(3). The isozymes are all homodimers, and they exhibit similar kinetic properties (3). The aligned amino acid sequences have residue identities ranging from 43 to 50%. Among conserved residues are those with catalytic functions or those that participate in cofactor binding (4,5). Salient differences include regions with putative functions in organellar targeting. The yeast mitochondrial isozyme (MDH1, subunit molecular weight ϭ 33,500), for example, has a 17-residue amino-terminal extension not present on the other isozymes; this extension is removed upon mitochondrial import (6). The peroxisomal isozyme (MDH3, subunit molecular weight ϭ 37,200) has a unique carboxyl-terminal tripeptide sequence, Ser-Lys-Leu. Similar SKL termini on other peroxisomal proteins have been found to be necessary and sufficient for organellar localization (7,8). The importance of this carboxyl terminus for localization of MDH3 is a focus of this work.
Cellular levels of all three malate dehydrogenase isozymes are low in yeast cells cultivated with glucose as a carbon source (2,3,9). For MDH1, this appears to be the result of glucose repression of gene expression as is the case for many mitochondrial proteins. Levels of MDH2, the cytosolic isozyme (subunit molecular weight ϭ 40,700), exhibit more dramatic changes. This enzyme, along with others with functions in gluconeogenesis or in the glyoxylate pathway in yeast (10 -12), is subject to rapid glucose-induced inactivation and degradation (13), a phenomenon termed catabolite inactivation. MDH2 has a 12-residue amino-terminal extension not present on aligned sequences of mature MDH1 or of MDH3 polypeptides, and we have shown that removal of this amino-terminal extension makes MDH2 refractory to catabolite inactivation (13). In this report, we investigate changes in MDH3 expression with glucose addition and test the sufficiency of the MDH2 aminoterminal extension for catabolite inactivation.
Disruption of the chromosomal MDH loci in yeast and analysis of resulting growth phenotypes have elucidated metabolic functions of each isozyme. Disruption of the MDH1 gene in a haploid strain produces an inability to grow with acetate as a carbon source (9). A similar growth phenotype is obtained with disruption of genes encoding several other tricarboxylic acid cycle enzymes including citrate synthase (14) and NAD-dependent isocitrate dehydrogenase (15) and is attributed to an energy deficiency with this carbon source. Disruption of the MDH2 gene produces an inability to grow with acetate or ethanol on minimal medium (2). Since such a mutant grows with C 2 carbon sources on rich medium, this phenotype indicates an auxotrophic requirement, perhaps for C 4 metabolites. Singh et al. (16) have reported that disruption of MDH3 eliminates growth with oleate as a carbon source, perhaps indicative of a deleterious effect on peroxisomal ␤-oxidation. These different phenotypes imply distinct and critical metabolic functions for each isozyme.
Because of the unfavorable equilibrium for formation of malate from oxaloacetate (⌬G°Ј Х ϩ7 kcal/mol), it has been proposed that malate dehydrogenase may physically associate with other specific enzymes within a metabolic pathway to ensure direct transfer of oxaloacetate (17). One of our goals in manipulating the yeast MDH isozymes is to assess the ability of isozymes, which catalyze the same reaction but which have structural dissimilarities, to function in alternative metabolic pathways and cellular compartments. In this report, we examine the ability of MDH3, normally found in peroxisomes, to function in the cytosol.

EXPERIMENTAL PROCEDURES
Growth Conditions and Yeast Strains-Yeast strains were cultivated in rich YP medium (1% yeast extract, 2% bactopeptone) or in minimal YNB medium (0.17% yeast nitrogen base, 0.5% ammonium sulfate, pH 6.0) containing supplements of 20 g/ml to supplement auxotrophic requirements for growth. The carbon source added to 2% was glucose, ethanol, or acetate. Growth rates in liquid cultures were monitored spectrophotometrically as absorbance at 600 nm. For glucose inactivation studies, strains were cultivated in YP medium plus 2% acetate and then transferred to YP medium containing 5% glucose as described previously (13). For cellular fractionation studies, yeast strains were precultivated in YNB medium containing 2% raffinose and then transferred to induction medium (YP medium with 2% ethanol and 0.1% oleate) and grown for at least 24 h to A 600 nm ϭ 1.0.
Yeast strains used in this study were the parental haploid strain S173-6B (MATa leu2-3, 112 his3-1 ura3-52 trp1-289; Ref. 18) and a collection of mutants derived from that strain containing disruption mutations in MDH loci. The collection was obtained by a genetic cross of a haploid strain containing deletion/HIS3 insertion mutations in MDH1 and MDH2 chromosomal genes (1, 2) with a haploid strain of the opposite mating type containing a deletion/LEU2 insertion mutation in the MDH3 gene (3). Haploid strains obtained following tetrad dissection were analyzed using protein immunoblots and Southern blots as described previously (1-3) to define segregation patterns for the disrupted loci. A complete collection of MDH strains was obtained; this includes the wild type, strains containing each single gene disruption, strains containing the three possible combinations of double disruptions, and a strain containing the triple gene disruption. The method of Ito et al. (19) was used for yeast transformations.
Plasmid Construction and Mutagenesis-Plasmid DNA isolation, subcloning, and bacterial transformations into Escherichia coli strain DH5␣FЈ were conducted as described by Maniatis et al. (20).
The yeast MDH3 gene on a 3.5-kilobase pair HindIII restriction fragment was previously subcloned into a Bluescribe plasmid (designated pBSMDH3; Ref. 3). This plasmid was used to prepare singlestranded DNA for oligonucleotide-directed mutagenesis. A 66-mer oligonucleotide (5Ј-GAATTGCGACTTTGACCGAATCTTGTTCTATGGA-TGGTGTAACTGAGTGAGGCATGTTTATGATTG-3Ј) with ends complementary to the 5Ј-coding region of MDH3 was used for "loop-in" mutagenesis to insert 12 codons from the amino-terminal coding region of MDH2 between the AUG translational initiation codon and the codon for the amino terminus of MDH3. The resulting plasmid was designated pBSMDH2/3. The Bluescribe plasmids, pBSMDH3 and pBSMDH2/3, were used for subsequent mutagenesis to delete three codons (Ser-Lys-Leu) from the carboxyl-terminal end of the MDH3 coding region. This was accomplished using a 20-mer oligonucleotide (5Ј-AATT-GAAGTAGCTCAAGAGTCTAGGATGAA-3Ј) for "loop-out" mutagenesis. The resulting plasmids were designated pBSMDH3⌬SKL and pBSMDH2/3⌬SKL.
Oligonucleotide-directed mutagenesis was conducted as described by Zoller and Smith (21). Bacterial transformants containing the altered sequences were identified by differential colony hybridization using as probes appropriate oligonucleotides, which were 5Ј-end-labeled with 32 P using polynucleotide kinase. Single-stranded DNA was prepared from bacterial colonies, and mutations were confirmed by nucleotide sequence analysis. The dideoxy method (22) was used for sequence analysis with oligonucleotide primers previously used to sequence the MDH3 gene (3). Following confirmation of nucleotide sequences, the MDH3 gene constructs on 3.5-kilobase pair HindIII fragments were cloned into pRS316, a yeast/bacterial shuttle plasmid containing a centromere cassette for single-copy expression in yeast and a URA3 gene for selection (23). The resulting plasmids were designated pRSMDH3, pRSMDH3⌬SKL, pRSMDH2/3, and pRSMDH2/3⌬SKL. Cellular Fractionation and Protein Analyses-For cellular fractionation, yeast cells were lysed, and cellular extracts were prepared by Dounce homogenization as described by Daum et al. (24). Soluble cytosolic fractions and organellar pellets were obtained by centrifugation at 12,000 ϫ g for 20 min. The pellets were resuspended and layered, without additional washing to preserve peroxisomal integrity, onto 20 ml of 15-37% linear Nycodenz gradients (25). The gradients were centrifuged at 4°C for 90 min at 36,000 rpm in a Beckman 50.2 Ti rotor. One-ml fractions were collected for enzyme assays and immunoblot analysis.
Spectrophotometic enzyme assays for malate dehydrogenase (9) and catalase (27) were conducted as described previously. Protein concentrations were determined by the Bradford method (28) using bovine serum albumin as the standard. Immunoblot analyses were performed using polyclonal antisera against MDH1 (9), MDH2 (2), MDH3 (3), and mitochondrial NADP-dependent isocitrate dehydrogenase (26) as described previously. The MDH3 antiserum was found to cross-react with MDH1 and was used in some immunoblot analyses to identify both isozymes in the same extract.

Peroxisomal Localization and Expression of MDH3-Previ-
ous studies established that the MDH3 isozyme is enriched in an organellar pellet following cellular fractionation of lysed yeast cells (3). As illustrated in Fig. 1A, all three malate dehydrogenase isozymes are immunochemically detectable and electrophoretically distinguishable in a whole cell lysate (lanes H) from our parental yeast strain cultivated with ethanol and oleate as carbon sources; this condition induces expression of FIG. 1. Peroxisomal localization of MDH3. Cellular fractionation was conducted as described under "Experimental Procedures" using the parental yeast strain S173-6B cultivated on rich YP medium with ethanol plus oleate as carbon sources. A, protein extracts from the whole cell homogenate (100 g; lanes H) and, following centrifugation, from the organellar pellet (25 g; lanes P) and soluble supernatant (100 g; lanes S) were electrophoresed and used for immunoblot analysis with antisera specific for MDH1, MDH2, or MDH3. The organellar pellet was further subjected to Nycodenz gradient centrifugation, and resulting fractions were used for immunoblot analysis (100 l/lane; panel B) with antisera specific for MDH3, MDH1, or mitochondrial NADP-specific isocitrate dehydrogenase (IDP1; Ref. 26), and for enzyme assays (C) for malate dehydrogenase and catalase. Activities are expressed relative to peaks detected within the gradient. The large increase in malate dehydrogenase activity in fractions at the top of the gradient is due to the abundance of mitochondrial MDH1. the malate dehydrogenases (29) and of catalase (30). MDH1 and MDH3, but not MDH2, are immunochemically detectable in an organellar pellet (lanes P) obtained as described under "Experimental Procedures." MDH2 is primarily found in the soluble cytosolic fraction (lanes S), which also contains detectable levels of MDH1 and MDH3, presumably due to organellar breakage during homogenization. To further separate mitochondrial and peroxisomal organelles, we have adapted the technique for Nycodenz gradient centrifugation described by Singh et al. (25). Under the conditions of centrifugation of the organellar pellet used here, peroxisomes, with catalase activity as a marker, are enriched in fractions within the Nycodenz gradient (Fig. 1C, fractions 8 -14), whereas mitochondria, with NADP-dependent isocitrate dehydrogenase as an immunochemical marker (26), remain at the top of the gradient (Fig.  1B, fraction 18). Immunoblot analysis demonstrates an enrichment of MDH3 in peroxisomal fractions and of MDH1 in mitochondrial fractions (Fig. 1B). Based on relative levels of activity in these and similar fractionation experiments conducted with parental and various MDH gene disruption strains, MDH3 accounts for approximately 10% of the total cellular malate dehydrogenase activity under these growth conditions.
Most yeast peroxisomal enzymes that function in the glyoxylate pathway, including malate synthase (11) and isocitrate lyase (12), are subject to rapid glucose-induced inactivation and degradation. We therefore investigated this phenomenon for MDH3. The parental strain was cultivated with acetate as a carbon source and then shifted to glucose medium as described under "Experimental Procedures." As shown in Fig. 2 and as previously reported (13), levels of MDH2 in whole cell extracts are rapidly depleted after glucose addition and are immunochemically undetectable after 60 min, whereas MDH1 levels are gradually diminished due to glucose repression of expression. Changes in levels of MDH3 in the same extracts are seen to more closely resemble those for MDH1 (also cf. Fig. 6A) and not for MDH2, suggesting that MDH3 is subject to glucose repression but is not subject to the phenomenon of rapid glucose-induced degradation. With steady state growth conditions, we have also found that immunochemical levels of MDH1 and MDH3 are detectable in glucose grown cells at levels 5-10-fold below acetate growth levels (3,6), whereas MDH2 is not immunochemically detectable in extracts from glucose-grown cells.
Construction and Expression of Mutant Forms of MDH3-To gain an understanding of the roles of compartmental localization and structure that contribute to the different effects of glucose on MDH3 and MDH2 expression, we constructed mutant forms of MDH3 based on sequence differences between the two isozymes illustrated in Fig. 3. Among the major structural differences is an amino-terminal extension of 12 amino acid residues on MDH2, which is not present on the aligned sequences for MDH3 and the mature form of MDH1. In a previous study, we constructed a mutant form of MDH2 lacking this 12-residue extension. The truncated enzyme was found to be active but no longer subject to catabolite inactivation (13). Another obvious structural difference is the carboxyl-terminal SKL tripeptide unique to MDH3. As mentioned above, this and similar tripeptides have been demonstrated to be necessary and sufficient for organellar localization of several peroxisomal proteins (7,8).
To assess the importance of these structural features for localization and turnover, the cloned MDH3 gene was used for oligonucleotide-directed mutagenesis as detailed under "Experimental Procedures" to "loop in" codons for the MDH2 aminoterminal extension and to "loop-out" codons for the carboxylterminal SKL tripeptide. Four variants of the MDH3 gene were constructed and cloned into pRS316 (23), a centromere-based shuttle plasmid carrying the yeast URA3 gene for selection. As shown in Fig. 4 For different experiments described in following sections, the plasmids carrying wild type and mutant forms of the MDH3 gene were transformed into haploid yeast strains containing various disruptions in the three MDH genomic loci. Construction of a complete collection of yeast MDH mutants by a genetic cross is described under "Experimental Procedures." For initial tests of expression, a strain containing genomic disruptions in all three MDH loci was used for transformation with the pRS plasmids. Ura ϩ transformants were isolated and grown on rich medium with ethanol as the carbon source; this carbon source has previously been shown to be permissive for growth of various yeast MDH mutants (1-3). MDH3 levels in whole cell extracts from representative transformants were examined by immunoblotting. As shown in Fig. 4, MDH3 polypeptides are all well expressed from the four plasmid-borne genes (lanes A-D), although levels of the wild type protein are approximately 2-fold higher than mutant enzyme levels. Total cellular MDH activities are also similar, with values ranging from 0.5 to 1.0 units/mg cellular protein as determined in three independent experiments. These levels are similar to those obtained from a yeast strain with an intact MDH3 locus and containing disruptions in MDH1 and MDH2 genes (3), suggesting that the plasmid-encoded wild-type and mutant forms of MDH3 are stably expressed and fully active. The size differences among the MDH3 polypeptides are also apparent. As shown in Fig. 4, polypeptides containing the 12-residue MDH2 amino-terminal extension (lanes C and D) migrate more slowly during SDS-polyacrylamide gel electrophoresis than those with authentic MDH3 amino termini (lanes A and B). The polypeptides containing carboxyl-terminal SKL deletions (lanes B and D) also migrate slightly more rapidly than their counterparts retaining the MDH3 carboxyl terminus (lanes A and C). The latter difference is more apparent in extracts from yeast transformants containing an MDH3 gene disruption but with an intact MDH1 locus (lanes AЈ and BЈ).
Cellular Localization and Activity of Mutant Forms of MDH3-To examine cellular localization of mutant forms of MDH3, yeast transformants containing all three MDH genomic gene disruptions and thus expressing only the plasmid-borne MDH3 genes (cf. Fig. 4, lanes A-D) were used for cellular fractionation. Organellar pellets and soluble cytosolic fractions were prepared, and the organellar pellets were resolved using Nycodenz gradient centrifugation as described under "Experimental Procedures." Fractions collected from the gradients were used for enzyme assays and immunoblot analysis. As demonstrated in Fig. 5, the authentic MDH3 enzyme (A) and the MDH2/3 fusion protein (C) are predominantly localized to organellar pellets, and both fractionate on the gradients as expected for proteins in peroxisomal organelles. In contrast, activity and immunoreactivity associated with the two mutant enzymes lacking the carboxyl-terminal SKL tripeptide (MDH3⌬SKL (B) and MDH2/3⌬SKL (D)) are primarily associated with soluble cellular fractions, and the proteins are not immunochemically detectable in peroxisomal gradient fractions. In these experiments, NADP-isocitrate dehydrogenase was used as an immunochemical control for localization and quantitative comparison of mitochondrial fractions in the Nycodenz gradients. The distribution of catalase activity as a peroxisomal marker in gradient fractions was also found in all gradients to be similar to that shown in Fig. 1C. These data indicate that the carboxyl-terminal SKL tripeptide is essential for peroxisomal localization of MDH3 and that the aminoterminal extension of MDH2 does not prevent organellar localization if the SKL tripeptide is present. Another important observation is that the MDH3 enzymes lacking the SKL tripeptide retain catalytic activity in the soluble cytosolic fraction (cf. Fig. 4 and related text).
Glucose Inactivation Studies of Cytosolic Forms of MDH3-To determine whether cytosolic forms of MDH3 are subject to glucose-induced degradation, the pRS plasmids encoding authentic MDH3 as a control and the MDH3⌬SKL and MDH2/3⌬SKL mutant enzymes were transformed into a yeast strain containing a single disruption of the MDH3 gene. The transformants were grown with acetate as a carbon source and then transferred to glucose medium as described above. Cellular levels of MDH3 and MDH1 were monitored by immunoblot analysis. As shown in Fig. 6, changes in levels of the cytosolic forms of MDH3 (MDH3⌬SKL (B) and MDH2/3⌬SKL (D) closely paralleled those of authentic MDH3 (A) and of MDH1 (A, B, and D), i.e. a gradual diminution due to glucose repression of expression. Neither cytosolic form demonstrated the pattern of rapid degradation associated with MDH2 (cf. Fig. 2). Thus, cytosolic localization alone is insufficient for glucose-induced degradation of MDH3. Also, although the amino-terminal ex-  Fig. 4. Each panel shows independent immunoblot analyses using antisera specific for MDH3 and for IDP1. The trace of MDH3 in fraction 18 of panel B is attributed to variable cytosolic contamination of the mitochondrial fraction due to omission of an organellar washing step as described under "Experimental Procedures." tension of MDH2 was previously shown to be essential for glucose-induced degradation of MDH2 (13), the addition of this extension to the amino terminus of a cytosolic form of MDH3 does not render this enzyme susceptible to rapid degradation.
Metabolic Function of Cytosolic Forms of MDH3-Mislocalization of catalytically active forms of yeast MDH3 to the cytosol and the availability of a mutant with a disruption of the MDH2 genomic locus affords a direct test for interchangeable functions of the malate dehydrogenase isozymes in specific metabolic pathways. The pRS plasmids carrying wild type and mutant MDH3 genes were therefore transformed into a yeast strain containing a single chromosomal disruption of the MDH2 gene. Resulting Ura ϩ transformants were then tested for the ability to utilize acetate or ethanol as a carbon source. The MDH2 disruption strain is able to grow on rich medium but unable to grow on minimal medium with either of these carbon sources (2). As illustrated in Fig. 7, transformation of this strain with plasmids encoding MDH3 (A) or MDH2/3 (C), the peroxisomal forms of this isozyme, does not alter the auxotrophic growth phenotype on ethanol plates. Transformation with plasmids encoding the cytosolic forms of MDH3 (MDH3⌬SKL (B) and MDH2/3⌬SKL (D)), in contrast, restores the ability of the MDH2 disruption strain to grow on minimal medium with ethanol. Growth rate studies with liquid cultures were also conducted; these confirmed that doubling times in minimal medium with ethanol as the carbon source are similar for transformants containing the cytosolic forms of MDH3 (8 and 8.5 h) and for the parental yeast strain (8 h), whereas no growth was detected over a 48-h period for transformants containing the peroxisomal forms of MDH3. From previous (3) and current estimates using extracts from various disruption strains grown on ethanol or acetate, the cellular malate dehydrogenase activity attributable to MDH2 is normally 3-5-fold greater than that attributable to MDH3. Thus, despite lower specific activities for the cytosolic forms of MDH3 as compared with that for authentic MDH2, the former can apparently function to correct metabolic lesions associated with loss of MDH2.

DISCUSSION
Two types of peroxisomal targeting signals, termed PTS1 and PTS2, have been identified and are described in recent reviews of organellar import and biogenesis (7,8). The SKL carboxyl-terminal tripeptide of the MDH3 isozyme of malate dehydrogenase in S. cerevisiae is a canonical type 1 PTS. In this report, we demonstrate that MDH3 is localized to microbody organelles isolated by density gradient centrifugation and that removal of the SKL tripeptide results in mislocalization of the enzyme to the soluble cytosolic fraction of cellular lysates. Interestingly, the plant homologue of yeast MDH3 contains a type 2 PTS, in this case an amino-terminal extension of 37 residues, which is proteolytically removed upon organellar import (31). The plant malate dehydrogenase extension has also been demonstrated to be essential for import (32). Distinct import mechanisms for proteins containing the two types of PTS motifs have been implicated by analyses of mutants with defects in peroxisome assembly (7,8). Thus, although the yeast and plant enzymes are structurally similar (43% residue identity; Ref. 3) and presumably functionally equivalent, it is likely that relatively recent evolutionary events involved acquisition of different import motifs.
The metabolic function of yeast and plant peroxisomal malate dehydrogenases is believed to be catalysis of a reaction in the glyoxylate cycle. This cycle is important during plant seedling germination for conversion of storage fatty acids to gluconeogenic precursors and in yeast allows growth with fatty acids or C 2 carbon sources such as acetate and ethanol. The report by Singh et al. (16) that disruption of the S. cerevisiae MDH3 gene results in an inability to grow with oleate as a carbon source supports a role in glyoxylate metabolism. However, we have been unable to duplicate this observation with our disruption mutants, suggesting there may be some strainspecific differences in this pathway.
The cytosolic form of MDH3 obtained by removal of the carboxyl-terminal SKL appears to retain full catalytic activity. This is not surprising given sequence homology with mitochondrial MDH1 and cytosolic MDH2, which lack this carboxyl terminus. Also, recent reports (33,34) suggest that proteins targeted to the peroxisomal matrix may attain a folded state and form oligomeric complexes prior to peroxisomal import. Thus, cytosolic formation of an active MDH3 dimer may be one step in the normal import pathway. Because the cytosolic form of MDH3 is catalytically active and stably expressed, we have been able to test the competence of this isozyme for replacement of MDH2 in yeast strains containing an MDH2 gene disruption. Cytosolic MDH3 was found to restore the ability of these strains to grow on C 2 carbon sources. Thus, despite some significant differences in physical properties of the two isozymes, e.g. respective pI values of 8.6 and 4.4 for MDH3 and MDH2 (3), function in gluconeogenesis is apparently not dependent on unique structural attributes of MDH2. However, cellular growth rate studies are a relatively gross measure of function, and conclusions about equivalence of metabolic contributions will require finer methods for measuring function in vivo as well as reciprocal compartmental exchanges of MDH isozymes. In related previous experiments, we attempted to obtain cytosolic localization of mitochondrial MDH1 by removal of the 17-residue amino-terminal targeting presequence. However, we found that this did not prevent mitochondrial import FIG. 6. Glucose-induced changes in levels of mutant forms of MDH3. Immunoblot analysis using antiserum for MDH3 and MDH1 was conducted using whole cell extracts (100 g) obtained from a yeast strain containing a chromosomal MDH3 disruption and transformed with pRSMDH3 (A), pRSMDH3⌬SKL (B), or pRSMDH2/3⌬SKL (D). Extracts were obtained from cells cultivated on YP medium with 2% acetate (lanes 0) and at times indicated (in minutes) following a shift to YP medium with 5% glucose. Panel designations correspond with sequence designations in Fig. 4.   FIG. 7. Growth phenotypes of MDH3 mutant strains. Yeast strains containing a chromosomal disruption of the MDH2 gene and transformed with pRSMDH3 (A), pRSMDH3⌬SKL (B), pRSMDH2/3 (C), or pRSMDH2/3⌬SKL (D) were streaked onto plates containing rich YP medium or minimal YNB medium with 2% ethanol as carbon source and grown, respectively, for 2 or 4 days at 30°C. in vivo (6). We are currently altering possible cryptic internal targeting signals with the intent of testing metabolic functions of cytosolic MDH1 for comparison with our results for MDH3.
We have found that, unlike gluconeogenic enzymes including MDH2 (13) and other glyoxylate pathway enzymes including malate synthase (11) and isocitrate lyase (12), MDH3 is not subject to rapid glucose-induced degradation. Cytosolic localization of MDH3 following removal of the SKL tripeptide does not increase susceptibility to the phenomenon. This result was not unexpected since we previously found that removal of the amino-terminal 12-residue extension of MDH2 produces an active enzyme refractory to glucose-induced inactivation and degradation. Thus, catalytic activity and cytosolic localization are insufficient attributes to target malate dehydrogenase for degradation. Also, addition of the 12-residue amino-terminal extension of MDH2 to peroxisomal or cytosolic forms of MDH3 has no effect on turnover, suggesting that, while this extension is essential for rapid turnover of MDH2 (13), there are other unique structural features of MDH2 that are recognized by components of the inactivation/turnover pathway. A final conclusion from current and previous results is that the aminoterminal extension of MDH2 is not critical for cellular location since its removal has no effect on cytosolic localization of MDH2 and its addition does not hinder peroxisomal localization of authentic MDH3. In future studies, we plan to extend these studies to correlate unique structural features of the malate dehydrogenase isozymes with differential expression and compartmental function.