The Saccharomyces cerevisiae Peroxisomal 2,4-Dienoyl-CoA Reductase Is Encoded by the Oleate-inducible GeneSPS19 *

β-Oxidation is compartmentalized in mammals into both mitochondria and peroxisomes. Fatty acids with double bonds at even-numbered positions require for their degradation the auxiliary enzyme 2,4-dienoyl-CoA reductase, and at least three isoforms, two mitochondrial and one peroxisomal, exist in the rat. TheSaccharomyces cerevisiae Sps19p is 34% similar to the human and rat mitochondrial reductases, and an SPS19deleted strain was unable to utilize petroselineate (cis-C18:1(6)) as the sole carbon source, but remained viable on oleate (cis-C18:1(9)). Sps19p was purified to homogeneity from oleate-induced cells and the homodimeric enzyme (native molecular weight 69,000) converted 2,4-hexadienoyl-CoA into 3-hexenoyl-CoA in an NADPH-dependent manner and therefore contained 2,4-dienoyl-CoA reductase activity. Antibodies raised against Sps19p decorated the peroxisomal matrix of oleate-induced cells.SPS19 shares with the sporulation-specificSPS18 a common promoter region that contains an oleate response element. This element unidirectionally regulates transcription of the reductase and is sufficient for oleate induction of a promoterless CYC1-lacZ reporter gene. SPS19 is dispensable for growth and sporulation on solid acetate and oleate media, but is essential for these processes to occur on petroselineate.

enoyl-CoA in bacteria (5,6). Reductases have previously been purified from Escherichia coli (7), from the yeasts Candida lipolytica (8) and Candida tropicalis (9), and from rat and bovine liver (10,11). Mammals possess at least two mitochondrial reductases (molecular weights 120,000 and 60,000) and a third peroxisomal one (12,13), and although the rat and human cDNAs for the M r 120,000 mitochondrial isoform have been cloned (14,15), the peroxisomal isoform has remained uncharacterized at the molecular level. A reductase-deficient E. coli mutant unable to grow on petroselineate (cis-C 18:1 (6) ) as the sole carbon source (16) further underscored the essential requirement for this enzyme, and a lethal inborn 2,4-dienoyl-CoA reductase deficiency in a human patient has been reported (17).
In Saccharomyces cerevisiae ␤-oxidation is solely confined to the peroxisomes (18). In cells grown on oleate, ␤-oxidation enzymes are induced and peroxisomes dramatically increase in number and size (19). The induction of a number of other S. cerevisiae genes encoding peroxisomal proteins such as catalase A (CTA1) and thiolase (FOX3) is mediated via a positive gene control sequence called the oleate response element (ORE 1 ; 20, 21). OREs have recently been shown to act as the binding target for the products of PIP2 (20) and OAF1 (21), and strains deleted at either locus were unable to grow on oleate (cis-C 18:1 (9) ) as the sole carbon source.
SPS19 was cloned together with the S. cerevisiae sporulation-specific gene SPS18, that was in turn isolated from a genomic yeast DNA library fused to a promoterless E. coli lacZ gene following a screen for sporulation-specific expression (22). Homozygous sps18/19⌬ deletants demonstrated a 4-fold reduction in sporulation, and the viable spores that did form failed to become resistant to ether and were more sensitive to lytic enzymes. Sequence analysis of the shared SPS18/19 promoter region showed that it contained potential OREs, and the deduced amino acid sequence of SPS19 revealed a carboxyl-terminal SKL peroxisome targeting signal (23,24). A search of the data bases for similarities to the Sps19p sequence indicated that it was 34% similar to the human and rat mitochondrial 2,4-dienoyl-CoA reductase. Since the roles of ␤-oxidation and peroxisomes during eukaryotic development and particularly during yeast sporulation are poorly understood, we were prompted to elucidate the properties and regulation of this gene product.
We show here that Sps19p is a peroxisomally localized 2,4dienoyl-CoA reductase and that the sps19⌬ deletant is unable to grow on petroselineate as the sole carbon source, although it remains viable on oleate. SPS19 is dispensable for ascosporo-genesis on solid acetate and oleate media but is essential for this process to occur on solid petroselineate medium. It is regulated by an ORE that is sufficient for heterologous expression of a promoterless CYC1-lacZ reporter gene under oleate induction conditions.

EXPERIMENTAL PROCEDURES
Strains and Plasmids-The S. cerevisiae strains and plasmids used are listed in Table I. E. coli strain DH10B was used for all plasmid amplifications and isolations.
Media and Growth Conditions-For RNA isolations, logarithmic cultures of the wild-type strain MF24-6x were shifted to YP medium (1% yeast extract, 2% peptone) containing the indicated carbon sources and grown for a further 16 h (20). For electron microscopy cells were propagated as described (25), and for ␤-galactosidase measurements cells were induced in oleate medium following a modified protocol (26). Stationary-phase haploid wild-type strains from overnight precultures consisting of YP medium and 5% glucose were transferred to 100-ml conical flasks containing 50 ml of YP medium containing both 0.2% oleate and 0.02% Tween 80 (adjusted to pH 7.0 with NaOH), 0.05% glucose, and 75 g/ml ampicillin to an absorbance of A 600 nm ϭ 0.2. The cultures were grown at 30°C with vigorous aeration at 170 rpm for the periods indicated. For 2,4-dienoyl-CoA reductase (Sps19p) purification, cell propagation was scaled up to 1-liter cultures in 5-liter conical flasks; the extreme flocculence of the strain used obviated the requirement for absorbance determination prior to transfer into oleate medium. Plates containing Tween 80 alone (Sigma), or Tween 80 with oleate (E. Merck AG, Darmstadt, Germany), or petroselineate (Sigma) (0.67% yeast nitrogen base with amino acids, 0.1% yeast extract, 0.5% potassium P i at pH 6.0, and 2% agar, autoclaved with 0.5% Tween 80 alone or with 0.5% Tween 80 and either 0.125% oleate or 0.125% petroselineate, respectively) were prepared by pouring a thin layer at a temperature below 55°C. These plates were used to assess utilization of the fatty acids (clear-zone formation) and for sporulation (by direct microscopic examination of cells) after about 7-day incubation.
Deletion of SPS19 -A 1.8-kb KpnI-XbaI SPS18/19 fragment from pJC18-2 (22) was cloned into pBluescript® SK(ϩ) (Stratagene, La Jolla CA) to yield pAG113. The 189-base pair ClaI-SphI fragment within SPS19 was then replaced with a 2-kb SmaI-SphI fragment containing the LEU2 gene from pJJ250 (27), resulting in the disruption plasmid pAG129. A 2.8-kb fragment produced by digesting pAG129 with ScaI and StuI was used for transformation ( Fig. 1B). SPS19 ORE:CYC1-lacZ Construct-The SalI-delineated SPS19O-RE1 and ORE2 oligonucleotides (Table I) defining the oleate response element within the shared SPS18/19 promoter region were annealed, and the double-stranded fragment was cloned into the XhoI site of the integrative vector pMF6 (28), a derivative of pLG669Z (29), to produce pAG244. Nucleotide sequencing revealed that the orientation of the ORE fragment with respect to the lacZ fusion boundary as being in the SPS18 direction, i.e. in the reverse orientation to SPS19, and integration of StuI-linearized pMF6 and pAG244 into the ura3 locus of BJ1991 resulted in strains yAG257 and yAG259, respectively.
Other Reporter Genes-The plasmids YIp357-19 and YIp357-19M1 were kindly donated by P. Yeoh (University of New South Wales, Sydney, Australia). Briefly, a 1.4-kb XbaI-SphI fragment containing the intergenic region and part of the coding regions of both SPS18 and SPS19 was excised from pUC18-KXC (23) and inserted into the corresponding sites in YIp357 (30) to create YIp357-19 (pAG454; SPS19-lacZ) as well as into M13mp19 for the substitution of the 3Ј ORE half-site sequence 5Ј-ACGCCGTGAG-3Ј with a unique XhoI site using site-directed mutagenesis. The substituted DNA was verified by nucleotide sequencing and cloned into YIp357 to create YIp357-19M1 (pAG23). Integration of the StuI-linearized plasmids containing the native and substituted DNA into the ura3 locus of BJ1991 resulted in strains yAG456 and yAG295, respectively. Cloning of the reverse orientation by inserting the 1.4-kb XbaI-SphI fragment isolated from pAG454 into the appropriate sites of YIp356R resulted in pAG534 (SPS18-lacZ). Similar integration of this plasmid yielded strain yAG561.
Enzyme Assays-2,4-Dienoyl-CoA reductase was assayed spectrophotometrically at 23°C as described (1). The assay mixture consisted of 0.1 M potassium P i (pH 7.4), 125 M NADPH, and 60 M 2,4-hexadienoyl-CoA (synthesized from trans-2,trans-4-hexadienoic acid via the a The numbers in superscript following the strains' designation refer to their parental genotype 1), e.g. MF24 -6ϫ 4 was derived from 4) GA1-8C. b Harvard University (Cambridge, MA). c Columbia University (New York). d University of New South Wales (Sydney, Australia). mixed anhydride system; (31)) as the substrate. Reductase activity was expressed as micromoles of substrate metabolized per min. ␤-Galactosidase activity was assayed in crude extracts prepared by breakage of cells with glass beads (32) and the values reported here, expressed as nanomoles of o-nitrophenyl-␤-D-galactopyranoside hydrolyzed per min/mg of protein, were the average of three experiments.
Purification Procedure-To purify Sps19p S. cerevisiae cells yAG162 harboring pJC18 (22) were grown in oleate medium for 3 days. The cells were collected by centrifugation, washed in 2 volumes of cold distilled water, and frozen at Ϫ70°C until required. All subsequent work was performed at 4°C. A cell pellet of 15 g wet weight was thawed, suspended in 150 ml of breakage buffer (50 mM sodium P i at pH 7.0, 0.4 M KCl, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 0.5 mM dithiothreitol), and a crude extract of cells disrupted with glass beads was prepared (33). A volume of 100 ml of the 6 ϫ 10 6 ϫ g ϫ min supernatant was diluted to 130 mM KCl, loaded onto a phosphocellulose P-11 column (2.5 ϫ 15 cm; Whatman) in equilibrium with 20 mM potassium P i (pH 7.0). After washing the column with 55 ml of the same buffer, the reductase activity was eluted with 20 mM potassium P i (pH 7.0) containing 0.5 M NaCl. Following the adjustment of the salt concentration to 0.7 M NaCl, the eluted reductase-containing fraction was loaded onto a Matrex gel red A column (1.5 x.10 cm; Amicon Corp, Lexington, MA) in equilibrium with 20 mM potassium P i (pH 7.0) containing 0.3 M NaCl. The column was washed with 55 ml of 20 mM potassium P i (pH 7.0) containing 10% glycerol and 0.7 M NaCl, and the reductase was eluted with 2.5 M NaCl in the same buffer. The buffer exchange to 50 mM sodium P i (pH 7.6) containing 10% glycerol and 1 mM dithiothreitol was carried out with a HiTrap™ desalting column (Pharmacia Biotech Inc.). The purification of Sps19p was completed by applying the sample to a SMART™-linked Resource™ S column (Pharmacia Biotech Inc.) in equilibrium with the buffer mentioned above, and the bound proteins were eluted with a linear gradient (20 ml from 0 -500 mM NaCl in the equilibration buffer).
Demonstration of End Product Accumulation-An aliquot containing 5 g of Resource™ S-purified protein was incubated with 810 l of a mixture that consisted of 0.01 M potassium P i (pH 7. Miscellaneous-The following procedures were performed according to published methods: nucleic acid manipulations, formaldehyde gel electrophoresis, blotting and hybridization (36), DNA fragment isolation (37), yeast transformation (38), verification of single plasmid integration (39), yeast RNA preparation for Northern analysis (40), determination of protein concentration (41), sodium dodecyl sulfate (SDS)polyacrylamide gel electrophoresis (42), separation of organelles using homogenized spheroplasts generated from oleate-induced cells (43), and nucleotide sequencing of pAG244 using the CYC1 oligonucleotide (44). Double labeling of the antigens Fox3p and Sps19p by repeating the antigen-antibody complex incubation sequence twice in tandem with the two antibodies raised in rabbits was performed as outlined (45,46). The use of FOX3 and ACT1 as probes has been described previously (20), and the 0.9-kb SspI-XbaI SPS18 and 0.5-kb SspI-SphI SPS19 fragments were obtained from pAG454. The [␣-32 P]dATP-labeled probes were generated with a random primer labeling kit (Prime-a-Gene™, Promega) according to the manufacturer's instructions. To remove any remaining minor contaminants, the purified reductase was applied onto a RPC™ reverse phase column (Pharmacia Biotech Inc.), and the eluted polypeptide was used for immunization of rabbits (47). The IgG fraction was purified from the antiserum using a protein A-Sepharose column (Pharmacia Biotech Inc.).

Sps19p Is Homologous to the Human and Rat 2,4-Dienoyl-
CoA Reductases-Analysis of the amino acid sequence of Sps19p revealed an overall 28% identity to the human mitochondrial 2,4-dienoyl-CoA reductase and 24% identity to the corresponding rat enzyme (Fig. 1A). Comparison of the NH 2 -terminal residues showed that unlike its mitochondrially targeted mammalian homologues, Sps19p lacked the appropriate leader sequence and instead contained a carboxyl-terminal SKL peroxisomal targeting signal (24). This suggested that Sps19p may be an enzyme involved in ␤-oxidation (23), since in yeast this process is restricted to the peroxisomes (18). The known 2,4-dienoyl-CoA reductases are NADPH-dependent, and the NAD(P) ϩ -binding site includes a highly conserved ␤ 1 ␣␤ 2 fold (48). The respective Sps19p sequence (Ala 30 to Glu 62 ) that came closest to this nucleotide-binding motif ( Fig.  1A; Ref. 49) deviated from the consensus sequence both at the second and third glycine residues. Sps19p vaguely resembled a reductase, and to obtain a clearer indication for its physiological requirement for the breakdown of unsaturated fatty acids, an appropriately deleted yeast strain was constructed (Fig. 1B).
Homozygous sps19⌬ Strains Do Not Grow or Sporulate on Petroselineate-To elucidate the potential participation of Sps19p in the metabolism of unsaturated lipids, wild-type and sps19⌬ strains were grown on plates containing fatty acids with double bonds at odd-numbered (oleate) and even-numbered (petroselineate) positions. In these solid media Tween 80 was added to help form the emulsion in the plate, but also to act as a relatively poor carbon source. Hence the strains could all grow on these plates ( Fig. 2) but zones of clearing indicate utilization of the additional fatty acid substrate (oleate or petroselineate).
We found that the deleted strain was unable to form clear zones and to sporulate at all on petroselineate, but did so on oleate (Fig. 2). This indicated that Sps19p was involved in the removal of double bonds at even-numbered positions. Transformation of the sps19⌬ strain with the multicopy plasmid pJC18 containing the intact SPS19 (22) complemented the mutant phenotype, since it restored its ability to utilize petroselineate.
Purification and Characterization of Sps19p-Pilot-scale production of Sps19p using pJC18 transformed cells yielded crude extracts that contained a high 2,4-dienoyl-CoA reductase activity (53 nmol ϫ min Ϫ1 ϫ mg Ϫ1 protein). Since reductase activity in the extracts from similarly propagated parental strains (without pJC18) was below the detection limit of the used assay, Sps19p could be identified throughout the ensuing purification process by enzymatic monitoring. Reductase activity was chromatographically purified from homogenized oleateinduced cultures using phosphocellulose P-11 (cation exchanger), Matrex gel red A (general dye ligand), and Resource™ S (cation exchanger) columns (Fig. 3). In the final preparation, an overall purification of about 50-fold was obtained with an approximate 26% yield, and this preparation showed a specific activity of 1.79 mol ϫ min Ϫ1 ϫ mg Ϫ1 protein for 2,4-hexadienoyl-CoA (Table II). SDS-polyacrylamide gel electrophoresis revealed a single protein band with an apparent molecular weight of 34,000, and size exclusion chromatography on a Su-perdex™ 200 HR 10/30 column (Pharmacia Biotech) yielded a native molecular weight of 69,000, indicating that the reductase was present as a homodimer. The monomer size so obtained agreed with the molecular weight of 31,400 calculated from the deduced amino acid sequence (23).
The amino-terminal residues of the protein were determined by microsequencing following reverse-phase chromatography and the resulting sequence (N-T-A-N-T-L-D-G-A-F-V-T-) conformed to the deduced residues 5-15 of Sps19p. Analysis of the SPS19 nucleotide sequence had revealed two potential translational start sites, each of which was followed by the TTA codon for Asn (23). Comparison of the data base sequences for Sps19p had shown a number of differences, most notably at position 2, where Asn (in GenBank™ number M90351) is replaced by Asp (in GenBank™ number X78898; Fig. 1A). The consensus for the N at position 5 is significant since asparagine is one of nine most abundant penultimate amino-terminal residues (4% of all Saccharomyces genes). Moreover, the finding that the Met at position 4 most likely acted as the signal for initiation of translation concurred with the consensus se-  (Fig. 4A). By generating a protein-free 2,4-hexadienoyl-CoA-depleted reductase reaction using purified Sps19p (Fig. 4B), and then adding the enzymes down-stream in the ␤-oxidation pathway, we were able to identify the end product of the enzyme (Fig. 4C).
The addition of 2-enoyl-CoA hydratase I to the depleted reductase filtrate in the presence of NAD ϩ and L-3-hydroxyacyl-CoA dehydrogenase failed to initiate the generation of 3-ketohexanoyl-CoA. However, subsequent addition of ⌬ 3 ,⌬ 2 -enoyl-CoA isomerase to the mixture drove the reaction forward, clearly demonstrating that 3-enoyl-CoA and not 2-enoyl-CoA was the product of the previous reaction. With the enzymatic activity and the reaction end product defined, completion of the characterization of Sps19p required that its location within the cell be determined biochemically and immunohistochemically.
Sps19p Is Localized to the Peroxisomes-The protein A-Sepharose purified IgG fraction from a rabbit injected with the purified Sps19p as antigen was applied to immunoblots containing crude extracts of ethanol-propagated and oleateinduced wild-type and sps19⌬ strains (Fig. 5A). A single band FIG. 1. Homology study, gene deletion, and ORE analysis of SPS19. A, Comparison of the deduced amino acid sequence of Sps19p with those of the human and rat 2,4-dienoyl-CoA reductases. Stars and points indicate amino acid identity and similarity, respectively; the first residue of each protein is shown in boldface; the dashes indicate the arrangement of the sequences for best fit; h in boldface indicates a hydrophobic amino acid residue. The salient features noted include the absence of the NH 2 -terminal leader sequence and the concomitant presence of a COOH-terminal SKL tripeptide in the yeast protein, as well as a degree of similarity with the mammalian peptides at the nucleotide-binding site of NADPH. The alignment was performed using CLUSTAL (58). B, construction of the SPS19 deletion. Restriction sites include: ClaI, (C), KpnI (K), ScaI (A), SphI (S), SspI (P), StuI (U), TthI (T), XbaI (X). A 189-base pair fragment within the SPS19 open reading frame was replaced by a 2-kb SmaI-SphI fragment containing the yeast LEU2 gene (not to scale). A 2.8-kb ScaI-StuI fragment was used to generate SPS19 deletants. C, schematic drawing of the ORE-containing region within the shared SPS18/19 promoter region. The ORE and the GC-rich region within the promoter are boxed. The XhoI site-substituted DNA in pAG23 is indicated by a bar below the boxes, whereas the ORE consensus is shown above. N6, any six nucleotides.
with the molecular weight of 34,000 was detected solely for the oleate-induced wild-type and not in the sps19⌬ strain (for which no band was detected under any conditions), thereby confirming the antibodies' monospecificity and potential utility in studying the subcellular localization of Sps19p. Subsequent fractionation studies of homogenized yeast spheroplasts demonstrated that Sps19p was exclusive to the organellar pellet (Fig. 5B). We then studied the subcellular location of Sps19p using immunoelectron microscopy and found that anti-Sps19p antibodies applied to sections of wild-type cells grown on oleate medium clearly decorated the peroxisomal matrix (Fig. 5, C  and D).
The SPS18/19 Promoter Region Is Unidirectionally Activated under Oleate Conditions-Oleate response elements, best described as palindromic CGG triplets spaced by 15-18 base pairs, mediate the transcriptional regulation of a number of peroxisomal protein-encoding genes, including POX1 (51), FOX3 (52), CTA1 (28), and PMP27 (53). The smallest element capable of relaying the fatty acid signal, a single ORE half-site (5Ј-CGGNNNTNA-3Ј), is sufficient for bi-directional induction (28). Previous nucleotide analysis of the shared SPS18/19 promoter region had identified potential ORE half-sites (Fig. 1C), two of which occurred as an appropriately spaced inverted repeat (23). We studied the ability of the intragenic region to initiate oleate induced bi-directional transcription by testing the expression of a lacZ reporter gene fused to either end of it (Fig. 6A). We noted that ␤-galactosidase expression by the SPS19-lacZ integrant under oleate conditions was 27-fold  higher compared with that expressed by the strain carrying SPS18-lacZ. We also observed that haploids were able to express SPS19 in a sporulation-independent manner. We investigated the transcription of the two genes under different carbon source conditions (Fig. 6B) and showed that while the profile of SPS19 was similar to that of oleate-inducible FOX3, no signal was detected on the Northern filter when radiolabeled SPS18 was applied. Hence, SPS19 transcription was likely to be mediated by an orientation-governed ORE, and identification of the responsive element, so as to demonstrate its sufficiency for oleate induction, would be facilitated using a reporter gene in which it was mutated.
The SPS19 ORE Is Sufficient for Oleate-dependent Initiation of Transcription-Of the three potential ORE half-sites present within the promoter, we chose to test the one that fully conformed to the 5Ј-CGGNNNTNA-3Ј consensus (Fig. 1C). A derivative of the SPS19-lacZ reporter gene with the distal end of the palindromic ORE replaced by an XhoI site was tested against the parent construct (Fig. 6C). The loss of all ␤-galactosidase activity in the strain carrying the mutated reporter clearly indicated that this was the sequence responsible for the oleate inducibility of SPS19. We then monitored the expression of a UAS-less CYC1-lacZ reporter gene into which the palindromic SPS19 ORE had been inserted in the direction of SPS18. The single integrant of the recombinant construct demonstrated a 22-fold increase in ␤-galactosidase expression com- FIG. 4. Sps19p is a 2,4-dienoyl-CoA reductase. A, the position of the enzyme in fatty acid breakdown (modified from Ref. 59). The ⌬ 3,5 ,⌬ 2,4 -dienoyl-CoA isomerase pathway (3,4) is not included in the figure due to the observations described in this paper. The yeast genes coding for the designated enzymes are as follows: I, POX1; II, SPS19; III, not identified; IV and V, FOX2 (Fox2p in yeast is a multifunctional enzyme catalyzing sequential 2-enoyl-CoA hydratase 2 and D-specific 3-hydroxyacyl-CoA dehydrogenase reactions via D-3-hydroxyacyl-CoA intermediates); VI, FOX3. The eukaryotic and bacterial reaction end products are indicated with a plain and dashed arrow, respectively. B, demonstration of reductase activity for Sps19p. Pure protein (arrow 1) was reacted with NADPH (arrow 2), and following the resetting of the spectrophotometer (*), 2,4-hexadienoyl-CoA substrate was added (arrow 3), and the oxidation of NADPH was monitored. C, the end product of the Sps19p reductase reaction is 3-enoyl-CoA. The previous proteinfree, substrate-depleted reaction (arrow 4) was reacted with the NAD ϩdependent, monofunctional mammalian mitochondrial enzyme L-specific 3-hydroxyacyl-CoA dehydrogenase (arrow 5), the monofunctional mammalian mitochondrial enzyme hydrating trans-2-enoyl-CoA to L-3hydroxyacyl-CoA, 2-enoyl-CoA hydratase I (arrow 6), and finally with ⌬ 3 ,⌬ 2 -enoyl-CoA isomerase (arrow 7), and the generation of the Mg 2ϩ 3-ketohexenoyl-CoA was monitored as described under "Experimental Procedures."  ). B, Sps19p is present only in the organellar pellet. The fractionation procedure used to isolate the peroxisomes and mitochondria was monitored using peroxisomal catalase A as an organellar marker (crude, crude spheroplast homogenate; sup, supernatant; org, organellar pellet). The sps19⌬ and wild-type (WT) strains used were yAG161 and yAG162, respectively. C, immunoelectron micrograph of an oleate-induced wild-type cell (BJ1991) doubly labeled with anti-Sps19p antibody (decorated with goat anti-rabbit IgG-gold complex, 10 nm) and anti- pared with the control (Fig. 6D). Thus, the SPS19 ORE was sufficient for oleate-dependent initiation of transcription. DISCUSSION This work describes the first molecular characterization of a peroxisomal 2,4-dienyol-CoA reductase. The S. cerevisiae enzyme (Sps19p) is compartmentalized differently from its rat and human mitochondrially targeted homologues (14,15), and its peroxisomal location is in line with the fact that in yeast ␤-oxidation is exclusively a peroxisomal process (18). Although the peroxisomal protein has been enzymatically and immunocytochemically demonstrated in rat liver (10,47,54), it has not yet been cloned. The relevance of reductase isoforms to human health has been underscored by the finding that an infant with a lethal 2,4-dienoyl-CoA reductase deficiency (postulated to be mitochondrial) retained only 40 and 17% of the normal level of activity in the liver and muscles, respectively. The residual activity was attributed to a peroxisomal (17) as well as a second functional mitochondrial isoform (13).
The deduced Sps19p amino acid residues deviate from the conserved NADPH-binding site in reductases, serving to broaden the consensus for this moiety (49). The E. coli 2,4dienoyl-CoA reductase consists of a single polypeptide chain with a molecular weight of 70,000 (6), and the Candida lipolytica enzyme (33,000 monomer) has been reported to consist of 10 -12 identical subunits (8). The mammalian reductases (molecular weight 120,000) act as homotetramers, whereas Sps19p (molecular weight 69,000) is a homodimer. The significant homology of the yeast protein to the mammalian homologues, together with the demonstration of its reductase activity and peroxisomal localization, places Sps19p as a closely related auxiliary enzyme to the one proposed to exist in humans. Work is currently underway to test whether the phenotype of the yeast mutant is complemented by the appropriately targeted human mitochondrial reductase.
The end products of mammalian 2,4-dienoyl-CoA reductases are always trans-3-enoyl-CoA esters (Fig. 4A), although they accept both trans-2,cis-4-, and trans-2,trans-4-dienoyl-CoA as substrates (6,10). The inability of sps19⌬ strains to grow on petroselineate plates demonstrated that the S. cerevisiae reductase is physiologically indispensable for the degradation of the trans-2,cis-4-dienoyl-CoA intermediate that arises from the degradation of this fatty acid (cis-6-octadecenoic acid). In our end product accumulation assay we had used trans-2,trans-4hexadienoyl-CoA. Although the formation of the 3-enoyl-CoA product for the S. cerevisiae reductase had clearly been demonstrated, there exists some uncertainty regarding the ester's chirality, and we reason that it is likely to be the trans isomer as it is in mammals.
We also provide here physiological data on the phenotype of the first organism devoid of a 2,4-dienoyl-CoA reductase. Although a reductase-deficient E. coli strain had been described previously, it retained 12% of the wild-type enzymatic activity and expressed this defective protein to apparently normal levels (16). Our observation that the deleted yeast strain was unable to utilize petroselineate as the sole carbon source, and that it remained viable on oleate, agrees with the data presented for the bacterial reductase mutant.
If the ⌬ 3,5 ,⌬ 2,4 -dienoyl-CoA isomerase pathway as demonstrated in rats is taken into account, then reductases are also required for the efficient breakdown of fatty acids with double bonds at odd-numbered positions (3,4). However, the likelihood that the observed normal growth of sps19⌬ cells on oleate may have been due to the action of a second, antigenically related 2,4-dienyol-CoA reductase isoform seems low. Anti-Sps19p antibodies capable of cross-reacting with the rat homologue from livers of clofibrate-treated animals (data not shown) did not react with crude extracts from the oleate-induced deletant. In addition, no isoforms are known to exist for other ␤-oxidation enzymes such as Pox1p (55), Fox2p (56), and Fox3p (57), where the missing activity in the respective mutants was not replaced by that of other putative isomers. Therefore, the capability of the sps19⌬ cells to grow on oleate suggested that dienoyl-CoA isomerase is not obligatory for the breakdown of unsaturated fatty acids with double bonds at odd-numbered positions.
The unidirectional transcription of SPS19 is mediated by an ORE within the SPS18/19 intergenic region. We identified the site responsible for oleate-induced expression of the reductase gene and demonstrated its sufficiency for oleate-mediated transcription. The shared SPS18/19 promoter region is capable of versatile regulation, where in diploids under sporulation conditions both genes are activated during meiosis, but SPS18 is induced to a 4-fold greater extent than SPS19. In contrast, under oleate conditions SPS19 is exclusively transcribed, and this induction is independent of ploidy.
FIG. 6. SPS19 is transcriptionally regulated by an ORE. A, SPS19-lacZ and SPS18-lacZ reporter activities following oleate induction. The yAG456 and yAG561 strains (19 and 18, respectively) were tested for ␤-galactosidase expression following induction in oleate medium for 18 h. Promoter expression was restricted to the SPS19 orientation. B, carbon source-dependent transcriptional activation of SPS19. The Northern blot containing RNA from the wild-type strain MF24-6x was probed with labeled SPS18, SPS19, FOX3, and the constitutively transcribed ACT1. The transcriptional profile of SPS19 resembled that of FOX3. No signal was obtained using SPS18. C, displacement of the SPS19 consensus ORE abolished ␤-galactosidase expression. The wildtype haploid strains harboring the native (yAG456, 19) and displaced ORE (yAG295, 19-M1) lacZ reporter genes were propagated under oleate medium conditions for 18 h. No expression was detected for the strain carrying the mutated plasmid. D, the SPS19 ORE was sufficient for oleate-dependent expression. The yAG259 strain carrying the recombinant SPS19 ORE:CYC1-lacZ reporter gene (ORE) was tested against the yAG257 control strain harboring the UAS-less reporter (pMF6 vector) following 19-h oleate induction. No significant expression was measured for the control strain.
Curiously, Sps19p is dispensable for the breakdown of oleate, although it is nonetheless induced by the presence of this fatty acid. We speculate that under physiological conditions yeast cells are unlikely to be confronted with a single lipid species as the sole carbon source, and hence when sensing the fatty acid signal cells will metabolically prime themselves to oxidize carbon double bonds at both odd-as well as at even-numbered positions. A deeper insight into how fatty acids induce Sps19p is required.
The yeast locus for the reductase had previously been isolated in a screen for sporulation genes, and the acetate medium-derived ascospores from the poorly sporulating, doubly deleted diploid strain (sps18/19⌬) were sensitive to diethyl ether treatment and to lytic enzymes in a manner suggesting a spore wall defect (23). We found no evidence for reductase requirement for solid oleate and acetate media-driven sporulation, nor for development of ether resistance in spores formed in liquid acetate medium (data not shown). SPS18 is therefore likely to play a key role in this developmental process, and its urgent characterization may also shed light on its potential involvement in lipid utilization.
There is evidence for both SPS19 transcript accumulation as well as production of reductase activity in cells transferred to acetate medium (data not shown), although functionally, the reductase is dispensable for sporulation in this medium type. There is however no ambiguity regarding the essential requirement of the reductase for ascosporogenesis using solid petroselineate medium, on which sps19⌬ mutants failed to grow, to form clear zones, and to sporulate. It remains unclear whether the deletant's incapacity for sporulation on this nonfermentable carbon source is due to its inability to obtain energy for completion of this process or due to the potential toxicity of the substrate, and further work is required to elucidate the role of this auxiliary enzyme in particular, and the roles of ␤-oxidation and peroxisomes in general, during fungal sporulation.