IDP3 encodes a peroxisomal NADP-dependent isocitrate dehydrogenase required for the beta-oxidation of unsaturated fatty acids.

In Saccharomyces cerevisiae the metabolic degradation of saturated fatty acids is exclusively confined to peroxisomes. In addition to a functional beta-oxidation system, the degradation of unsaturated fatty acids requires auxiliary enzymes, including a Delta2, Delta3-enoyl-CoA isomerase and an NADPH-dependent 2,4-dienoyl-CoA reductase. We found both enzymes to be present in yeast peroxisomes. The impermeability of the peroxisomal membrane for pyrimidine nucleotides led to the question of how the NADPH needed by the reductase is regenerated in the peroxisomal lumen. We report the identification and functional analysis of the IDP3 gene product, which is a yeast peroxisomal NADP-dependent isocitrate dehydrogenase. The newly identified peroxisomal protein is homologous to the mitochondrial Idp1p and cytosolic Idp2p, which both are yeast NADP-dependent isocitrate dehydrogenases. Yeast cells lacking Idp3p grow normally on saturated fatty acids, but growth is impaired on unsaturated fatty acids, indicating that the peroxisomal Idp3p is involved in their metabolic utilization. The data presented are consistent with the assumption that peroxisomes of S. cerevisiae contain the enzyme equipment needed for the degradation of unsaturated fatty acids, including an NADP-dependent isocitrate dehydrogenase, a putative constituent of a peroxisomal NADPH-regenerating redox system.

Peroxisomes harbor variable metabolic pathways that differ depending on cell type, developmental stage, and food supply (1,2). In reference to the multiplicity of cellular functions and to the ability of cells to adjust the enzymatic equipment as well as the size and number of these organelles in response to the cellular demand, peroxisomes are appropriately called multipurpose organelles (3). The importance of peroxisomes for cellular function is especially emphasized by a number of inherited diseases in humans that are caused by peroxisomal dysfunction and usually have profound clinical consequences (4).
A typical metabolic pathway of peroxisomes is the ␤-oxidation of fatty acids (5,6). In fact, whereas the presence of a mitochondrial ␤-oxidation system is restricted to mammalian cells and a few protists (7), the fatty acid oxidation in peroxisomes is nearly ubiquitous among eukaryotic cells (7,8). The peroxisomal and the mitochondrial degradation of fatty acids is performed by functionally comparable but genetically distinct proteins (8,9). In fungi and plants, the degradation of fatty acids exclusively takes place in peroxisomes, and growth on fatty acids results in proliferation of peroxisomes accompanied by a massive induction of peroxisomal proteins including the ␤-oxidation enzymes (7,10).
In addition to the chain shortening ␤-oxidation system, the oxidation of unsaturated fatty acids requires auxiliary enzymes for the elimination of the double bonds (8). Degradation of unsaturated fatty acids with odd-numbered double bonds requires a ⌬2,⌬3-enoyl-CoA isomerase (Fig. 1B) (11). For the degradation of unsaturated fatty acid with even-numbered double bonds, an NADPH-dependent 2,4-dienoyl-CoA reductase is needed in addition to the isomerase (Fig. 1A) (12). Recently, a novel pathway for the degradation of unsaturated fatty acids with double bonds at odd-numbered carbon atoms has been described that also requires the NADPH-dependent reductase described above (Fig. 1C) (13,14). The successive reduction and isomerization of double bonds by these auxiliary enzymes results in the formation of intermediates that can reenter the ␤-oxidation spiral (8). The presence of both the ⌬2,⌬3-enoyl-CoA isomerase and the NADPH-dependent 2,4dienoyl-CoA reductase has been demonstrated in all peroxisomes studied so far (8). As the peroxisomal membrane has been suggested to be impermeable for small solutes (15), the requirement of the peroxisomal enoyl-CoA reductase for NADPH raises the question of the existence of an NADPH regenerating system in peroxisomes.
We applied a reverse genetic approach to identify proteins essential for peroxisome function in Saccharomyces cerevisiae.
Here we report the identification, characterization, and functional analysis of a peroxisomal NADP-dependent isocitrate dehydrogenase. Deficiency in this enzyme resulted in an impaired growth of S. cerevisiae on unsaturated fatty acids, whereas growth on saturated fatty acids was not affected. The peroxisomal Idp3p is suggested to be involved in the regeneration of the NADPH needed for the peroxisomal degradation of unsaturated fatty acids.
Yeast strains were grown at 30°C in YPD or in minimal medium (SD) as described previously (19). For oleic acid induction of peroxisome proliferation, cells were grown in SD medium to late log phase and then shifted into YNO (20) and incubated for 14 h. Necessary auxotrophic requirements were added according to Ausubel et al. (21). Whole yeast cell extracts were prepared by the method of Yaffe and Schatz (22). Enzymatic modification of DNA, fragment purification, and bacterial transformation were performed essentially as described by Ausubel et al. (21). Yeast transformations were carried out according to Gietz and Sugino (23).
Purification and Amino Acid Sequencing of Idp3p-High salt-extracted peroxisomal membranes were prepared from oleic acid-induced SKQ2N cells. Further separation of the peroxisomal membrane proteins was achieved by reverse-phase HPLC 1 according to Erdmann and Blobel (24).
For sequencing of Idp3p, the SDS samples of HPLC fractions containing the protein (fractions 42-46; see Fig. 2) were pooled and separated on a 12% SDS-polyacrylamide gel. Polypeptides were electrophoretically transferred onto a polyvinylidene difluoride membrane and visualized with 0.1% Amido Black in 10% acetic acid. Idp3p was excised, and Lys-C-derived peptides of the protein were separated by HPLC and subjected to sequence analysis on a gas phase sequenator according to Fernandez et al. (25). Protein sequence analysis was provided by the Rockefeller University Protein Sequencing Facility, which is supported in part by National Institutes of Health shared instrumentation grants and by funds provided by the U.S. Army and Navy for the purchase of equipment.
Isolation and Sequencing of IDP3-According to the obtained internal sequences of Idp3p, degenerated sense (5Ј-GCGAATTCA(C/T)C-CIAT(A/C/T)GT(A/G/C/T)GA(A/G)ATG-3Ј) and antisense (5Ј-TCTAAG-CTT(A/G/C/T)GCIAC(C/T)TC(A/G)TC(T/G/A)AT-3Ј) oligonucleotide primers, distinguishing between the IDP3 gene and the highly homologous IDP1 and IDP2 genes were designed. The corresponding genomic region of the IDP3 gene was amplified by the polymerase chain reaction with yeast genomic DNA (100 ng; Promega, Madison, WI) as template. The amplification product was isolated and subcloned into a derivative pBluescript SK(ϩ) using the Srf1 kit (Stratagene, La Jolla, CA), resulting in pSRF-IDP3. The authenticity of the insert was confirmed by DNA sequencing. A [ 32 P]dATP-labeled probe of 520 base pairs was generated by PCR with oligonucleotide primers iso6 (5Ј-GCCACTATAACAC-CCGATG-3Ј) and iso1 (5Ј-CGTACGTTATTTTTAAAGCCTG-3Ј) and the plasmid pSRF-IDP3 as template, using a random-primed labeling kit (Boehringer, Mannheim, Germany). High stringency hybridization according to Maniatis et al. (26) was performed to screen a YEp13-based yeast genomic library (27) that was kindly provided by M. Bolotin-Fukuhara. Two positive clones containing the IDP3 gene were isolated, and sequencing was directly performed on one of the plasmids, YEp13-IDP3, with an automatic sequencer (model 373A; Applied Biosystems, Weiterstadt, Germany), the DyeDesoxy terminator cycle sequencing kit (Applied Biosystems), and synthetic oligonucleotides. Both strands of the IDP3 gene were sequenced.
Cell Fractionation-Spheroplasting of yeast cells, homogenization, and differential centrifugation at 25,000 ϫ g of homogenates were performed as described previously (19).
For subfractionation by isopycnic sucrose density gradient centrifugation, cell lysates or organellar pellets were loaded onto linear 20 -53% sucrose density gradients (34). Centrifugation, fractionation of gradients, and preparation of the samples for SDS-PAGE were carried out as described (24).
Protease Protection-Peroxisomal peak fractions from a sucrose density gradient were pooled and diluted five times in gradient buffer (34). Peroxisomes were sedimented at 25,000 ϫ g for 30 min and subsequently resuspended in homogenization buffer (19) without protease inhibitors but supplemented with 50 mM KCl. Equal amounts of isolated peroxisomes were incubated with increasing amounts of proteinase K for 10 min on ice. Protease was inactivated immediately after the incubation with 4 mM PMSF, proteins were precipitated with trichloroacetic acid, and samples were processed for SDS-PAGE.

RESULTS
Isolation and Identification of Idp3p-Peroxisomes were isolated from oleic acid-induced S. cerevisiae cells and successively extracted by low and high salt treatments. The proteins of high salt extracted peroxisomes were solubilized with SDS and separated by HPLC and SDS-PAGE (Fig. 2). Lys-C-derived internal fragments of the 45-kDa protein marked in Fig. 2 were subjected to amino-terminal protein sequencing in preparation for DNA cloning and sequencing of the corresponding gene (see "Experimental Procedures"). The open reading frame of the isolated DNA fragment encoded a new protein with a calculated molecular mass of 48 kDa (Fig. 3) that later on also appeared in the yeast genome data base as open reading frame YNL009w. A search of the GenBank TM data base with the predicted amino acid sequence of IDP3 revealed the yeast genes IDP1 (17) and IDP2 (18) as close relatives of the newly identified gene, hereafter referred to as IDP3. The overall identity of Idp3p with Idp1p and Idp2p is 68 and 70%, respectively ( Fig. 4). Idp1p and Idp2p represent the two NADP-dependent isocitrate dehydrogenases reported for S. cerevisiae to date. Idp2p is localized in the yeast cytosol, and Idp1p is a mitochondrial isoenzyme that differs from the other two proteins by an N-terminal extension, which functions as a mitochondrial targeting signal (Fig. 4) (18,40). Idp3p lacks the mitochondrial targeting signal and instead is characterized by an additional nine amino acids at the extreme C terminus. These terminal amino acids of Idp3p comprise the tripeptide cysteine-lysineleucine (CKL; Fig. 3), a putative peroxisomal targeting signal 1 (PTS1) for S. cerevisiae (41,42). The prominent presence in the HPLC profile of peroxisomal proteins (Fig. 2), the sequence similarity to Idp1p and to Idp2p (Fig. 4), and the presence of the putative peroxisomal targeting signal (PTS1; Fig. 3) suggested that Idp3p might be a peroxisomal NADP-dependent isocitrate dehydrogenase.
IDP3 Is Induced upon Growth on Oleic Acid-Antibodies were generated against an internal fragment of Idp3p comprising amino acids 144 -210, which displays the lowest similarity of the protein to Idp1p and Idp2p (Fig. 4). A polypeptide with the predicted molecular mass for Idp3p (48 kDa) was detected in wild-type but not in idp3⌬ yeast extracts (Fig. 5A). Although binding of the antibodies to other proteins was observed under low stringency conditions, none of these bands disappeared in mutants lacking either Idp1p or Idp2p, indicating that the antibodies generated do not recognize Idp1p or Idp2p but are specific for Idp3p (Fig. 5A).
In S. cerevisiae, growth on oleic acid results in a massive proliferation of peroxisomes accompanied by the induction of peroxisomal enzymes involved in peroxisomal fatty acid degradation (7,10). The oleic acid induction is mediated by the transcription activator Pip2p, which binds to a well defined oleic acid-responsive element at the promoter of several peroxisomal proteins (43). A perfect consensus sequence for Pip2p binding is also present upstream of the IDP3 open reading frame (position Ϫ311 to Ϫ289; Fig. 3), and Idp3p was found to be highly inducible by oleic acid (Fig. 5B), supporting the assumption of Idp3p being involved in peroxisomal fatty acid degradation.
Peroxisomes of S. cerevisiae Contain an NADP-dependent Isocitrate Dehydrogenase-The subcellular localization of NADP-dependent isocitrate dehydrogenases in S. cerevisiae was first analyzed by differential centrifugation of cell homogenates from oleic acid-induced wild-type yeast cells. More than 60% of the Idp activity was found in the supernatant fraction, suggesting that the cytosolic isoform might be responsible for the majority of the endogenous enzyme activity (data not shown). Sedimented organelles were further fractionated by sucrose density gradient centrifugation. NADP-dependent isocitrate dehydrogenase activity was found in both the mitochondrial and the peroxisomal fractions (Fig. 6). The peroxisomal peak, however, comprised a smaller fraction of the total particular enzyme activity. To exclude the possibility that the activity found in the peroxisomal fraction was due to a mitochondrial contamination, the subcellular localization of the enzymes was also analyzed in a idp1⌬ mutant strain, lacking the mitochondrial NADP-dependent isocitrate dehydrogenase. After differential centrifugation of cell homogenates of the idp1⌬ mutant strain, about 15% of the total NADP-dependent isocitrate dehydrogenase activity was still localized to the organellar fraction. Subsequent sucrose density gradient fractionation confirmed the absence of the mitochondrial enzyme and demonstrated the activity to exclusively co-segregate with peroxisomal marker enzymes (Fig. 6), consistent with the presence of a peroxisomal isoenzyme of the NADP-dependent isocitrate dehydrogenases in S. cerevisiae.
The Peroxisomal NADP-dependent Isocitrate Dehydrogenase Activity Is Performed by Idp3p-As a first step to analyze whether the peroxisomal NADP-dependent isocitrate activity is due to Idp3p, we studied the subcellular localization of the protein. Immunological detection of Idp3p in fractions generated by differential centrifugation of yeast cell homogenates revealed that the protein is exclusively found in the organellar pellet (Fig. 7A). Immunoblot analysis of fractions gained by subsequent sucrose density gradient centrifugation of the organellar pellets demonstrated the protein to be exclusively localized in the peroxisomal fractions (Fig. 7B). In agreement with Idp3p being responsible for the peroxisomal NADP-dependent isocitrate dehydrogenase activity, the absence of this protein in idp3⌬ mutant cells correlated with the disappearance of the enzyme activity in the organellar pellet and in the peroxisomal fractions of sucrose density gradients (Fig. 7).
The sequence similarity of Idp3p to the two NADP-dependent isocitrate dehydrogenases suggested that the protein Idp3p itself is an NADP-dependent isocitrate dehydrogenase. However, to confirm that the lack of the peroxisomal NADP-dependent isocitrate dehydrogenase activity upon IDP3 deletion is not caused indirectly, the protein was heterologously expressed in E. coli, and the enzyme properties of isolated Idp3p were analyzed. Transformation of E. coli with a plasmid carrying the IDP3 gene under the control of the bacterial promoter resulted in a massive increase in NADP-dependent isocitrate dehydrogenase activity in bacterial lysates accompanied by the appearance of immunoreactive Idp3p (Fig. 8A). Taking advantage of the C-terminal histidine tag, the protein was purified to apparent homogeneity as judged by SDS-PAGE. The isolated Idp3p showed a specific NADP-dependent isocitrate dehydrogenase activity of 1654 nanokatals/mg. Since expression of Idp3p did result in the concomitant appearance of NADP-dependent isocitrate dehydrogenase activity in bacterial extracts and since the enzyme activity was retained by purified Idp3p, these data confirmed Idp3p to be the yeast peroxisomal NADP-dependent isocitrate dehydrogenase (Fig. 8B).
A set of kinetic properties of the Idp3p enzyme activity were studied with the recombinant yeast protein purified from E. coli extracts. The enzyme activity did strongly depend on the presence of NADP ϩ that could not be replaced by NAD ϩ (data not shown). The K m values for NADP ϩ and isocitrate were 0.02 and 0.05 mM, respectively (Fig. 9), and are in the range of those reported for the peroxisomal NADP-dependent isocitrate dehydrogenase of Candida tropicalis (0.016 mM for NADP and 0.11 mM for isocitrate) (44).
Idp3p Is Localized in the Peroxisomal Lumen-An organellar fraction isolated from spheroplasts of yeast wild-type cells was subjected to extraction by low salt, high salt, and carbonate at pH 11 according to Ref. 37. Idp3p was resistant to low salt extraction but was released by high salt and carbonate treatment of the organelles (Fig. 10A). These extraction properties distinguished Idp3p from two other peroxisomal proteins. Pex3p was resistant to all treatments, consistent with it being an integral membrane protein (34). As expected for a matrix protein, peroxisomal thiolase (Fox3p) (30) was extracted by all treatments. The extractability of Idp3p by carbonate treatment suggested that Idp3p does not span the peroxisomal membrane. Idp3p also does not seem to be tightly associated with the peroxisomal membrane, since part of the protein could be extracted with high salt. The extraction properties of Idp3p are similar to those observed for Pcs60p, a protein of the peroxisomal matrix that is also loosely associated with the peroxisomal membrane (32). To distinguish whether Idp3p is associated with the outer aspects of peroxisomes or whether the protein resides in the peroxisomal lumen, we analyzed the sensitivity of organellar Idp3p to externally added proteases. In the presence of detergents and proteases, all proteins were rapidly degraded. When detergents were present, degradation of proteins was also observed without the addition of protease, presumably due to the liberation of endogenous proteases (data not shown). However, in the absence of detergents, both the FIG. 6. Organellar localization of NADP-dependent isocitrate dehydrogenase isoenzymes in wild-type and idp1⌬ mutant cells. Organelles obtained by a 25,000 ϫ g centrifugation of cell homogenates from oleic acid-induced wild-type and idp1⌬ mutant cells were separated on isopycnic 20 -53% (w/w) sucrose density gradients. Fractions of 1 ml were collected from the bottom of the gradients. Relative amounts of the peroxisomal marker enzyme catalase and the mitochondrial fumarase as well as NADP-dependent isocitrate-dehydrogenase were monitored by activity measurements. Peroxisomes peaked in fraction 7 at a density of 1.21 g/ml, and mitochondria peaked in fraction 12 at a density of 1.18 g/ml. NADP-dependent isocitrate-dehydrogenase activity was detected in both peroxisomal and mitochondrial fractions of the wild-type lysate but only in peroxisomal fractions of the idp1⌬ lysate.

IDP3 Encodes a Peroxisomal Isocitrate Dehydrogenase
intraperoxisomal thiolase (Fox3p) and Idp3p were protected against added proteases (Fig. 10B). Under the same conditions, Pex14p, which is located at the cytosolic face of the peroxisomal membrane (33), was rapidly degraded. Taken together, these results are consistent with an intraperoxisomal localization of Idp3p.
Peroxisomal Targeting of Idp3p Depends on the Presence of the Three C-terminal Amino Acids-The amino acids CKL at the extreme C terminus of Idp3p fit the consensus for a yeast PTS1 (41,42). To analyze whether this putative PTS1 of Idp3p is functional, we analyzed the subcellular localization of a mutated Idp3p⌬CKL lacking the last three amino acids. Idp3p⌬CKL was expressed in an idp3⌬ strain, and localization of the protein was determined by subcellular fractionation of whole-cell homogenates on sucrose density gradients. Idp3p⌬CKL did not co-segregate with the peroxisomal markers but instead was exclusively found in the loading zone of the gradient, suggesting a cytosolic localization of the protein (data not shown). This result indicated that the last three amino acids of Idp3p are essential for the peroxisomal targeting of the protein. The presence of a functional PTS1 in Idp3p is in line with the observed protease resistance of the protein (Fig. 10B), since this signal sequence is known to target proteins to the peroxisomal matrix (41,45).
Idp3p Is Required for the Peroxisomal Degradation of Unsaturated Fatty Acids-In search for the function of Idp3p in peroxisomal metabolism, we tested the growth abilities of idp3⌬ cells on different carbon sources. Cells grew normally on medium containing glucose, glycerol, or stearate as a single carbon source (Fig. 11A). Also, on oleic acid plates, no significant growth differences between wild-type and idp3⌬ mutant cells were observed (Fig. 11B). In liquid oleic acid medium, however, the generation time of idp3⌬ mutant cells increased from 8 h as determined for the wild type to 12 h for the mutant (Fig. 11B). Because the only difference between stearic acid and oleic acid is the presence of one double bond in position 9, the observed growth defect suggested that the peroxisomal Idp3p might play a role in the degradation of unsaturated fatty acids. This assumption was further supported by the complete inability of cells lacking Idp3p to grow on petroselinic acid, an unsaturated fatty acid that contains a double bond at position 6 ( Fig. 11C). The observed growth defects on oleic acid and petroselinic acid medium were complemented upon transformation of the idp3⌬ mutant with the wild-type IDP3 gene (Fig.  11, B and C). These results confirmed that the impaired growth of idp3⌬ mutant cells on unsaturated fatty acids was indeed FIG. 7. Idp3p is localized in yeast peroxisomes and is required for the peroxisomal oxidative decarboxylation of isocitrate. A, immunoblot analysis and enzyme activity measurements of cell fractions that were obtained by differential centrifugation of cell-free extracts from oleic acid-induced idp1⌬ and idp1/idp3⌬ cells. Equal volumes of each fraction were immunologically analyzed for the presence of Idp3p. In parallel, the fractions were assayed for NADP-dependent isocitrate dehydrogenase activity. Idp3p was exclusively localized to the organellar fraction of idp1⌬ cells but was absent in idp1/idp3⌬ cells. In idp1⌬ cells, about 80% of the total enzyme activity was found in the soluble fraction. In idp1/idp3⌬ cells, the deficiency in Idp3p correlated with the disappearance of NADP-dependent isocitrate dehydrogenase activity from the organellar fraction. B, correlation of Idp3p presence and NADP-dependent isocitrate dehydrogenase activity in peroxisomal fractions obtained by isopycnic 20 -53% (w/w) sucrose density gradient centrifugation of organelles of the 25,000 ϫ g pellet from oleic acid-induced idp1⌬ and idp1/idp3⌬ cells. Equal volumes of each fraction were immunologically analyzed for the presence of Idp3p and thiolase. Relative amounts of NADP ϩ -dependent isocitrate dehydrogenase and peroxisomal marker enzymes catalase and mitochondrial fumarase were monitored by enzyme activity measurements. Peroxisomes peaked at a density of 1.21 g/ml, and mitochondria peaked at a density of 1.17 g/ml. Idp3p was found predominantly in the peroxisomal peak fractions of idp1⌬ cells but was absent in idp1/idp3⌬ cells. No NADP-dependent isocitrate dehydrogenase activity was detected in peroxisomes lacking the Idp3p. caused by the lack of Idp3p.
Yeast Peroxisomes Contain Auxiliary Enzymes Needed for the Degradation of Unsaturated Fatty Acids-The ability of S. cerevisiae to grow on unsaturated fatty acids as the single carbon source (Fig. 11), the presence of an NADP-dependent isocitrate dehydrogenase in peroxisomes (Fig. 6), and its suggested role of supplying NADPH for the degradation of unsaturated fatty acids encouraged us to search for auxiliary enzymes of this pathway. Both the ⌬2,⌬3-enoyl-CoA isomerase and the NADPdependent 2,4-dienoyl-CoA reductase activities were detected in whole-cell yeast lysates (data not shown). For subcellular localization of the activities, wild-type yeast homogenates were subjected to sucrose density gradient centrifugation, which did result in a clear separation of peroxisomes and mitochondria as judged by organelle-specific marker enzymes (Fig. 12). Both the ⌬2,⌬3-enoyl-CoA isomerase and the NADP-dependent 2,4-dienoyl-CoA reductase activities co-segregated with the peroxisomal marker catalase, demonstrating that both enzymes are localized in peroxisomes of S. cerevisiae (Fig. 12). These data suggest that peroxisomes of S. cerevisiae harbor the entire enzyme equipment needed for the utilization of unsaturated fatty acids, including an NADP-dependent isocitrate dehydrogenase, a putative component of an NADPH-regenerating system. DISCUSSION Here we report on the molecular identification and functional characterization of a peroxisomal NADP-dependent isocitrate dehydrogenase of S. cerevisiae. In line with a role in the peroxisomal metabolism of unsaturated fatty acids, the Idp3p has been demonstrated to be exclusively peroxisomal, and the protein was shown to be essential for the growth of S. cerevisiae on unsaturated fatty acids but dispensable for growth on saturated fatty acids. The supposed function of the protein in peroxisomal fatty acid metabolism is the regeneration of NADPH that is needed by the NADPH-dependent 2,4dienoyl-CoA reductase for the reductive elimination of double bonds of unsaturated fatty acids. This reductase and the ⌬2,⌬3enoyl-CoA isomerase, another auxiliary enzyme needed for the degradation of unsaturated fatty acids, have been localized to yeast peroxisomes (Fig. 12). The presence of these enzyme activities in peroxisomes has far reaching implications for our understanding of the peroxisomal metabolism and transport of metabolites across the peroxisomal membrane. The data presented are consistent with the assumption that peroxisomes of S. cerevisiae maintain the entire enzyme equipment needed for the degradation of unsaturated fatty acids, including an NADPdependent isocitrate dehydrogenase, a putative constituent of a peroxisomal NADPH-regenerating redox system.
Idp3p was isolated from peroxisomes of oleic acid-induced yeast cells (Fig. 2), and peptide sequence data of the protein were instrumental in cloning the corresponding gene from a genomic yeast library (Fig. 3). Idp3p is exclusively localized in peroxisomes, and consistent with its function in peroxisomal fatty acid metabolism, Idp3p was highly induced upon growth on oleic acid (Figs. 5-7). Protease protection data suggested that the protein resides in the peroxisomal lumen (Fig. 10), which is further supported by the observation that peroxisomal targeting of Idp3p depends on the presence of a C-terminal type 1 peroxisomal targeting signal (data not shown), known to target proteins from the cytosol across the peroxisomal membrane barrier into the peroxisomal matrix (41,45,46). That the peroxisomal Idp3p indeed is an NADP-dependent isocitrate dehydrogenase was confirmed by the characterization of the enzymatic properties of purified, recombinant Idp3p (Fig. 8).
Interestingly, an NADP-dependent isocitrate dehydrogenase has also been detected in peroxisomes of the n-alkane-utilizing yeast C. tropicalis (44).
Beside the peroxisomal Idp3p, three yeast isoenzymes of isocitrate dehydrogenase have been described that catalyze the oxidative decarboxylation of isocitrate to ␣-ketoglutarate. The NAD-specific mitochondrial isoenzyme is an octamer of two nonidentical subunits designated Idh1p and Idh2p (47,48) and is believed to catalyze a key regulation step in the tricarbonic FIG. 8. Idp3p is a peroxisomal NADP-dependent isocitrate dehydrogenase. A, cell homogenates from E. coli BL21(DE3) transformed with the pET vector or pET-IDP3 were analyzed for the presence of Idp3p and of NADP-dependent isocitrate dehydrogenase activity. Equal amounts of cell homogenates were subjected to immunoblot analysis with Idp3p antibodies and to enzyme measurements. Immunoreactive Idp3p was only detected in E. coli transformants expressing IDP3, and expression correlated with the appearance of specific NADP-dependent isocitrate dehydrogenase activity in the bacterial extracts. B, HIS 6 -tagged Idp3p from bacterial extracts of E. coli BL21(DE3) was purified by affinity chromatography on nickel-nitrilotriacetic acid resin (see "Experimental Procedures"). The block diagram shows the specific activity of NADP-dependent isocitrate dehydrogenase of equal portions of the bacterial homogenate (load), the column flow-through (FT), four steps of washing (washes 1-4), and the eluate. Equal portions of the fractions were processed for SDS-PAGE and Coomassie staining (a) and immunoblot analysis with antibodies against the Idp3p (b). As judged by SDS-PAGE and immunoblot analysis, purification of Idp3p was to apparent homogeneity, and the purified protein retained the NADP-dependent isocitrate dehydrogenase activity. nkat, nanokatal. acid cycle. Less clear are the functions of the two NADP-specific isoenzymes located in mitochondria and the cytosol (17,18). The glutamate auxotrophy upon deletion of both Idp1p and Idh1p suggest that both enzymes contribute to the anaplerotic supply of ␣-ketoglutarate for glutamate formation (40). Furthermore, as isocitrate and ␣-ketoglutarate can traverse the mitochondrial membrane via specific transporters (49), it has been suggested that the proteins may participate in an inter-compartmental exchange of reducing equivalents (18). This raises the question of whether Idp3p might play a comparable role in the peroxisomal metabolism. In the cytosol, NADPH is generated by, for instance, the pentose phosphate pathway. However, because of the impermeability of the peroxisomal membrane for pyrimidine nucleotides (15), the cytosolic NADPH pool cannot directly account for the peroxisomal need for NADPH. This emphasizes the necessity for an NADPHregenerating system in the peroxisomal lumen. Because the formation of ␣-ketoglutarate for the production of glutamate is primarily catalyzed by the yeast mitochondrial NAD-dependent and NADP-dependent isocitrate dehydrogenases (40), the most likely biological function of Idp3p is the regeneration of NADPH. The involvement of Idp3p in the intraperoxisomal regeneration of NADPH, which is necessary for the degradation of unsaturated fatty acids, is also more in agreement with the peroxisomal localization and with the oleic acid inducibility of the protein.
The requirement of the peroxisomal degradation of fatty acids with even-numbered double bonds for NADPH is well established (8,12). Degradation of these fatty acids in the ␤-oxidation spiral leads to 2,4-dienoyl-CoA intermediates that are reduced to 3-trans-enoyl-CoA in a redox reaction that requires NADPH and that is catalyzed by the peroxisomal NADPHdependent 2,4-dienoyl-CoA reductase. The resulting 3-transenoyl-CoA is subsequently isomerized to 2-trans-enoyl-CoA, which can be reintroduced into the ␤-oxidation spiral (Fig. 1).
The assumption that Idp3p provides the NADPH for this chain of reactions is supported by the observation that cells lacking the protein grow normally on stearate (C18:0) but have lost the ability to grow on petroselinic acid (⌬6-C18:1; Fig. 11).
Until recently, it was generally believed that unsaturated fatty acids with double bonds extending from odd-numbered carbon atoms are chain-shortened to 3-cis-enoyl-CoA esters, which after isomerization to 2-trans-enoyl-CoA are further degraded by the ␤-oxidation spiral (Fig. 1B) (11). According to this pathway, NADPH would not be needed for the metabolization of these unsaturated fatty acids (Fig. 1). However, Tserng and Jin (50) reported that in mammalian cells also the degradation FIG. 10. Subperoxisomal localization of Idp3p. A, extraction of peroxisomes. 25,000 ϫ g organelle pellets were prepared from oleic acid-induced wild-type cells and extracted by low salt, high salt, and carbonate treatment at pH 11.0. Extracted proteins were separated from the membranes by centrifugation. Equal proportions of pellet (P) and supernatant (S) fractions were analyzed by SDS-PAGE and Western blot analysis using specific antibodies against Idp3p, the integral membrane protein Pex3p (34), Pcs60p (32), and the peroxisomal matrix marker Fox3p (30). As Idp3p is hardly extracted by high salt treatment and totally extracted at pH 11, it behaves like Pcs60p, which is localized in the peroxisomal matrix but is also found loosely associated with the peroxisomal membrane (32). B, protease protection analysis of purified peroxisomes. A cell homogenate of wild-type cells was separated by sucrose density gradient centrifugation, and peroxisomal peak fractions were pooled. Equal amounts of the pooled peroxisomal fractions were incubated in the presence or absence of proteinase K for 10 min on ice. Samples were analyzed by SDS-PAGE and Western blot analysis using antibodies against Idp3p, Pex14p, and Fox3p. The resistance of Idp3p against externally added protease suggests that the protein is protected by the peroxisomal membrane and thus resides in the peroxisomal lumen. In contrast, Pex14p, a peripheral membrane protein localized at the cytosolic face of the peroxisomal membrane (33), is rapidly degraded under these conditions.
FIG. 11. The peroxisomal NADP-dependent isocitrate dehydrogenase Idp3p is required for growth on unsaturated fatty acids. Growth kinetics from wild-type and idp3⌬, idp3⌬ [IDP3], and pex7⌬ mutant cells on solid agar plates (upper panels) as well as wild-type and idp3⌬ cells on liquid media (lower panels) containing stearic acid (A), oleic acid (B), or petroselinic acid (C) as a single carbon source. The growth of cells lacking the Idp3p was severely affected on oleic acid media and completely impaired on petroselinic acid-containing media. Growth abilities of idp3⌬ mutant cells complemented with the IDP3 gene were indistinguishable from those of the wild type. of unsaturated fatty acids with double bonds extending from odd-numbered carbon atoms requires NADPH. This observation gained support by the exploration of a novel pathway for the reductive removal of odd-numbered double bonds of fatty acids (Fig. 1C) (13). According to this pathway, a ⌬3,5,⌬2,4dienoyl-CoA isomerase, together with the NADPH-dependent 2,4-dienoyl-CoA reductase and the ⌬3,⌬2-dienoyl-CoA isomerase facilitate the reduction of odd-numbered double bonds as illustrated in Fig. 1. Recently, it has been suggested that this novel pathway might also be responsible for the degradation of odd-numbered double bonded fatty acids in mammalian peroxisomes (8,14). In this respect, it is interesting to note that also yeast cells lacking Idp3p are less capable than the wild type of growing on oleic acid (⌬9-C18:1) as the single carbon source (Fig. 11).
The peroxisomal localization of the Idp3 leads to questions about the origin of the isocitrate and the fate of the ␣-ketoglutarate that is produced. The most simple explanation would be that ␣-ketoglutarate is exported directly in exchange for isocitrate as has been demonstrated for mitochondria (49). In principle, isocitrate could also form in peroxisomes from the citrate that is generated by the fusion of acetyl-CoA with oxalacetate, catalyzed by the peroxisomal citrate synthase (Cit2p) (15,51). However, despite efforts, an aconitase activity has not yet been detected in yeast peroxisomes, thus making the peroxisomal formation of isocitrate from citrate rather unlikely. The direct import of isocitrate from the cytosol into the peroxisomal lumen would predict the existence of a peroxisomal membrane transporter for isocitrate; however, experimental evidence for such a transporter is still missing. In general, our knowledge on the influx and efflux of peroxisomal metabolites and especially on the nature of the carriers involved is still rather limited. For S. cerevisiae, only two peroxisomal metabolite carriers have been described. The heterodimeric Pat1p/ Pat2p ABC-transporter has been suggested to participate in the peroxisomal import of acyl-CoA esters (52), and the peroxisomal carnitin acetyl transferase is involved in the export of the ␤-oxidation-derived acetyl-CoA (15,53).
The data presented here are consistent with the idea that peroxisomes of S. cerevisiae maintain the entire enzyme equipment needed for the degradation of unsaturated fatty acids, including an NADP-dependent isocitrate dehydrogenase, a putative constituent of a peroxisomal NADPH-regenerating redox system. The latter supports the notion of an involvement of peroxisomes in an intercompartmental exchange of reducing equivalents and predicts novel peroxisomal metabolite transporters as constituents of a redox shuttle across the peroxisomal membrane. FIG. 12. Yeast peroxisomes harbor the auxiliary enzymes 2,4dienoyl-CoA reductase and enoyl-CoA isomerase for the degradation of unsaturated fatty acids. Fractions were obtained by isopycnic 20 -53% (w/w) sucrose density gradient centrifugation of cell-free extracts from oleic acid-induced wild-type cells. Relative amounts of the peroxisomal marker enzyme catalase and mitochondrial fumarase as well as 2,4-dienoyl-CoA reductase and enoyl-CoA isomerase were monitored. Peroxisomes peaked in fraction 6 at a density of 1.21 g/ml, and mitochondria peaked in fraction 11 at a density of 1.18 g/ml. 2,4-Dienoyl-CoA reductase and enoyl-CoA isomerase activities co-segregated with the peroxisomal marker enzymes, suggesting that both enzymes are localized in peroxisomes.