Δ3,5-Δ2,4-Dienoyl-CoA Isomerase from Rat Liver

rECH1, a recently identified rat cDNA (FitzPatrick, D. R., Germain-Lee, E., and Valle, D. (1995)Genomics 27, 457–466) encodes a polypeptide belonging to the hydratase/isomerase superfamily. We modeled the structure of rECH1 based on rat mitochondrial 2-enoyl-CoA hydratase 1. The model predicts that rECH1p has the hydratase fold in the core domain and two domains for interaction with other subunits. When we incubated 3,5,8,11,14-eicosapentaenoyl-CoA with purified rECH1p, the spectral data suggested a switching of the double bonds from the Δ3-Δ5 to the Δ2-Δ4 positions. This was confirmed by demonstrating that the product was a valid substrate for 2,4-dienoyl-CoA reductase. These results indicate that rECH1p is Δ3,5-Δ2,4-dienoyl-CoA isomerase. Subcellular fractionation and immunoelectron microscopy using antibodies to a synthetic polypeptide derived from the C terminus of rECH1p showed that rECH1p is located in the matrix of both mitochondria and peroxisomes in rat liver. Consistent with these observations, the 36,000-Da rECH1p has a potential N-terminal mitochondrial targeting signal as well as a C-terminal peroxisomal targeting signal type 1. Transport of the protein into the mitochondria with cleavage of the targeting signal results in a mature mitochondrial form with a molecular mass of 32,000 Da; transport to peroxisomes yields a protein of 36,000 Da.

Using a subtractive strategy, FitzPatrick et al. (15) recently isolated a peroxisome proliferator-induced rat liver cDNA that encodes a novel member of the hydratase/isomerase protein family with a type 1 peroxisomal targeting signal (SKL) at its C terminus. The gene was named rat enoyl-CoA hydratase (rECH1), referring to its possible activity. Here we show that the rECH1 protein (rECH1p) actually is a ⌬ 3,5 -⌬ 2,4 -dienoyl-CoA isomerase participating in the metabolism of unsaturated fatty acids with double bonds at the ⌬ 5 position. This is the first characterization of a dienoyl-CoA isomerase at the molecular level.
rECH1p was modeled based on the structure of the 2-enoyl-CoA hydratase 1 at 2.5 Å resolution (25) (Protein Data Bank (27)), ID number 1DUB. The sequence alignment was performed with the GCG (28) and Clustal W program packages. The final alignment was per-* This work was supported by grants from the Sigrid Juselius Foundation and from the Academy of Finland. A portion of this work was supported by an award from National Institute of Child Health and Human Development, National Institutes of Health Grant HD10981 to the Kennedy Kreiger Institute. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ formed using information from multiple sequence alignments and secondary structural information from the three-dimensional structures. The model was built with the program Insight II (Biosym Technologies, Inc., San Diego, CA). A side chain rotamer library was used to model amino acid substitutions. Deletions and insertions were modeled by searching a data base of high resolution structures. The search was performed for fragments of the required length and end point separation by using two residues at each end of the loop as anchor points. The fragments obtained were evaluated on the basis of three criteria: root mean square deviation from the anchor points, sequence similarity, and interference with the core region. There were three insertions and no deletions with respect to hydratase 1. The model was refined by energy minimization with the program Discover using Amber force field in a step-wise manner. First, hydrogen atoms were relaxed, then the borders of insertions and deletions were harmonically restrained, and the rest of the molecule was fixed. In the next step, the borders of insertions and deletions and the C ␣ atoms of the conserved regions and finally only the C ␣ atoms of the conserved regions were harmonically constrained.
Antibodies-We immunized rabbits with a synthetic peptide (CAM-EKKDSKSITFSKL-COOH) from the C-terminal end of rECH1p. An N-terminal cysteine residue was added to link the peptide to Imject maleimide-activated ovalbumin (Pierce). We also raised rabbit antibodies against porcine fumarase (EC 4.2.1.2) (Boehringer Mannheim) purified by reverse phase chromatography on a microreverse phase column C2/C18 PC 3.2/3 column (Pharmacia Biotech Inc.). The IgG fraction from these antisera was purified by ammonium sulfate precipitation to a saturation of 45% followed by size exclusion chromatography on Superdex 200 high resolution 10/30 column (Pharmacia) in equilibrium with 10 mM sodium phosphate, 2 mM potassium phosphate, 140 mM NaCl, 3 mM KCl, 5 mM ␤-mercaptoethanol, pH 7.4, (PBS) containing 0.02% sodium azide. Rabbit anti-peroxisomal multifunctional enzyme 1 antibody (10) and anti-rat ⌬ 3 -⌬ 2 -enoyl-CoA isomerase antibody (29) were obtained as described. Antibodies against the rat catalase were a gift from Dr. Stefan Alexson (Karolinska Institute, Huddinge, Sweden). The use of animals for immunization was approved by the university committee on animal experimentation.
Expression of Truncated rECH1 in Escherichia coli-We utilized rECH1 cDNA in Bluescript KSϪ (15) as a template for polymerase chain reaction amplification. The forward primer (5Ј-GACATATG-GCATACGAGTCTATTCAAGTAACATCTGC) corresponds to nucleotides 170 -196 with an additional 11 5Ј nucleotides providing an NdeI restriction site, an ATG initiation methionine codon, and a GCA (alanine) codon. 3Ј of this, the sequence continues with TAC encoding Tyr-54 of the full-length protein. The reverse primer (5Ј-CAGGATC-CCACTCAGAGCTTGGAGAAGGTGAT) corresponds to nucleotides 974 -997 and contains an additional BamHI restriction site on its 5Ј end. We cloned the amplified polymerase chain reaction fragment into pUC 18 plasmid with the Sure Clone ligation kit (Pharmacia). After digestion with NdeI and BamHI, the cDNA was ligated into a pET 3a expression vector (30). The protein was expressed in the BL21(DE3)LysS E. coli strain using a pET expression system (Novagen, Madison, WI) according to manufacturer instructions. Induction was carried out for 3 h at 35°C, and the cells were harvested by centrifugation, washed with PBS, and suspended in 0.01 volume of the same buffer. The cells were lysed by freezing in liquid nitrogen and thawing at 37°C. RNase, DNase, and lysozyme were added to final concentrations of 2, 20, and 100 g/ml, respectively, and the mixture was incubated at 35°C for 45 min. The lysate was centrifuged at 15,000 ϫ g for 15 min, and the supernatant was used for further experiments.
Subcellular Fractionation-Rat liver was homogenized in 0.25 M sucrose, 0.1% (v/v) ethanol, 1 mM EDTA, pH 7.2, and fractionated by differential centrifugation into nuclear, heavy mitochondrial, light mitochondrial, microsomal, and cytosolic fractions as described by Verheyden et al. (31). Peroxisomes were purified from the light mitochondrial fraction by centrifugation in a Nycodenz step gradient as described by Vanhove et al. (32). Mitochondria were isolated by centrifugation in a self-generating Percoll gradient (33). Thirty g of crude homogenate and 10 g of mitochondrial and peroxisomal fractions were separated in an SDS-polyacrylamide gel electrophoresis and blotted onto a nitrocellulose filter. Filters were used for immunodetection with a secondary antibody, goat anti-rabbit IgG-conjugated with horseradish peroxidase (Bio-Rad).
Purification of rECH1p-Fourteen ml of the lysate containing 52 mg of soluble proteins was applied to a DEAE-Sephacel column (2.5 ϫ 11 cm) in equilibrium with 20 mM potassium phosphate, pH 6.5, and 100 mM NaCl. The flow-through fraction containing rECH1p was dialyzed against 3 ϫ 500 ml of 20 mM MES, pH 6.0, and applied to a 1-ml Resource S column (Pharmacia) in equilibrium with the same buffer.
The bound rECH1p was eluted with a linear 20-ml gradient from 0 to 0.3 M NaCl in 20 mM MES, pH 6.0. During the purification procedure, the elution of rECH1p was detected by dot-blotting aliquots of fractions on a nitrocellulose filter and immunolabeling with anti-rECH1. The native molecular weight of rECH1p (20 g) was determined by size exclusion chromatography on a Superdex 200 high resolution 10/30 column (Pharmacia) in 200 mM phosphate buffer, pH 7.4, containing 1 mM EDTA.
Immunoelectron Microscopy-Livers were harvested from control and clofibrate-treated (fed with 0.5% (w/w) clofibrate-supplemented chow for two weeks) Sprague-Dawley rats after vascular perfusion fixation. For this procedure, the rats were heparinized (15 units of heparin/g of body weight), euthanized, and perfused through the left ventricle and ascending aorta at a hydrostatic pressure of 120 cm (37) for 2 min with PBS containing 2 mM EDTA, 1 mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluoride. A prefixative solution containing 8% paraformaldehyde and 20 mM ethylacetimidate in PBS was then administered for 3 min. After this, the main fixative solution containing 4% paraformaldehyde and 0.25% (v/v) glutaraldehyde in PBS was infused for 15 min. Ethylacetimidate was included in the prefixative solution for its described capacity for maintaining antigen accessibility and preservation of antigenicity (38). After perfusion fixation, 1-2-mm blocks of liver were further fixed for 2 h at room temperature by immersion in the same fixative solution. Finally the tissue blocks were washed three times in 7.5% (w/v) sucrose in PBS, where they were kept at 4°C until processed.
For thin sections, tissues were embedded in Micro-Bed resin (Electron Microscopy Sciences, Fort Washington, PA) according to manufacturer instructions, except that the tissues were infiltrated from 70% ethanol instead of completion of the dehydration protocol. The resin was polymerized at 4°C under UV light (365 nm) for 60 h.
For immunogold labeling, grids were incubated for 5 min with 50 mM glycine in PBS and for 15 min in blocking medium containing 1% bovine serum albumin and 1% fish skin gelatin in PBS (PBS-BSA-FSG). Primary labeling of the sections was made by incubating the grids for 2 h at room temperature with the primary antibody. Similar grids were incubated with the IgG fraction of rabbit serum containing antibodies to the rat peroxisomal peroxisomal MFE1 and mitochondrial ⌬ 2 -⌬ 4 -dienoyl-CoA reductase as quality controls of the specificity of labeling. The grids were washed in PBS followed by incubation for 1 h with protein A-gold, 20 nm (for rECH1p test sections) or 10 nm (for specificity controls) diluted 1:30 in PBS-BSA-FSG-0.1% Tween 20. The grids were then washed in 500-l drops of PBS containing 250 mM NaCl. The antigen-antibody complexes were fixed with 1% glutaraldehyde in PBS, and grids were finally washed in distilled water and counterstained with uranyl acetate and lead citrate.

Molecular Modeling and Structure Prediction of rECH1-
The sequence identity between rECH1p and enoyl-CoA hydratase 1 is modest (25%), but the residues and elements that are functionally and structurally important in hydratase 1 are conserved ( Fig. 1) (25). We modeled the rECH1p structure using the coordinates of hydratase 1 as a template. The alignment of rECH1 with the mature form of hydratase 1 (starting at Gly-30) (39) shows 48 additional N-terminal amino acid residues, three insertions, and three amino acid residues in the C terminus (SKL). The C-terminal extension could not be modeled. Regions containing a single amino acid residue insertion appear in the beginning and in the end of the trimerization domain 1, but the insertions do not affect the general scaffolding of the protein. There is a large insertion in the helical portion of the second part of the hydratase spiral (Fig. 2). This section was modeled as an extended ␣-helix based on structural requirements and sequence analysis. The corresponding region in hydratase 1 has a flexible loop, which might indicate allowance of greater variability in this region. Insertions were modeled using a library of loops generated with the Insight II program from known high resolution three-dimensional structures in the Protein Data Bank. The loop library was scanned for structural elements giving the best matches with the predicted secondary structure element in the inserted region after positioning the insert in the model. The final model has a typical globular structure according to several structural tests (40).
All the secondary structural elements forming the fold in enoyl-CoA hydratase 1 are conserved in rECH1p, and the insertions appear outside the central core of the enzyme. Especially Glu-144 and Glu-164, which are proposed to participate in the catalysis in hydratase 1, are also conserved in rECH1p (Asp-127 and Glu-147), with the exception that Glu-144 of hydratase 1 was replaced by Asp in rECH1p. The trimerization domains 1 and 2 of hydratase 1, which are involved in subunit interactions (25) are also conserved in rECH1p and do not contain insertions or deletions. We speculate, therefore, that like hydratase 1, rECH1p has a trimeric structural unit(s). The mature hydratase 1 is composed of dimers of two trimers forming a monohexameric enzyme.
Expression of rECH1 and Purification-We expressed rECH1p in E. coli as an N-terminal-truncated version beginning at Tyr-54. An additional N-terminal Met-Ala sequence was contributed by the 5Ј primer. The truncation site was chosen based on the following observations. The amino acid sequence similarity of rECH1p with the mature rat mitochondrial enoyl-CoA hydratase 1 (39), Pseudomonas sp. 4-chlorobenzoate dehalogenase (41) (Fig. 1), and the mature rat mitochondrial short-chain ⌬ 3 -⌬ 2 -enoyl-CoA isomerase (9) (data not shown) started near Tyr-54 of rECH1p. Furthermore, the model of rECH1p indicates that the N-terminal residues  are not involved in the folding of the isomerase/hydratase fold or the T1 and T2 domains and apparently do not contribute to the architecture of the substrate binding site.
Starting with 0.65 g of bacterial cells, which corresponds to 51.8 mg of soluble proteins, the purification protocol yielded 1.7 mg of purified rECH1p. In SDS-polyacrylamide gel electrophoresis analysis, only one band of 30,000 Da was observed, indicating that the protein was purified to apparent homogeneity. The observed mass of the polypeptide agreed well with the mass of the purified recombinant polypeptide (30,100 Da) as determined by matrix-assisted laser desorption/ionizationtime of flight (MALDI-TOF) (Kompakt MALDI III, Kratos Analytical, Manchester, UK) mass spectroscopy and with that deduced from the open reading frame of pET-rECH1 (30,539 Da). When applied to a Superdex 200 high resolution 10/30 column, the purified rECH1p eluted slightly before hydratase 1 (the native molecular mass 161,600 Da (39) purified from rat liver). The estimated molecular mass of rECH1p was 170,000 Da, indicating that it is a hexamer.
Search for Function of rECH1p-The functionally characterized members of the hydratase/isomerase superfamily utilize coenzyme A thioesters of carboxylic acids as substrates, and therefore, we hypothesized that rECH1p could also bind acyl-CoA esters. Some members of the family act on ⌬ 2 or ⌬ 3 double bonds of enoyl-CoA esters or catalyze hydration reactions. Because trans-2-eicosenoyl-CoA and 3,5,8,11,14-eicosapentenoyl-CoA fulfill the substrate criteria mentioned above and because they have known spectral characteristics in the near UV region (13,14) we chose them for screening substrates.
We generated the substrates by incubating arachidoyl-CoA and arachidonoyl-CoA with yeast acyl-CoA oxidase. Fig. 3A shows that the incubation with arachidoyl-CoA resulted in the increase of absorbance at 260 nm, indicating the formation of a double bond in the ⌬ 2 position of the acyl chain (accumulation of trans-2-eicosenoyl-CoA). The subsequent addition of rECH1p into this reaction mixture did not significantly change the absorption in the region scanned (240 -390 nm). When arachidonoyl-CoA was incubated with acyl-CoA oxidase (Fig. 3B), the initial increase of absorption was similar to the previous one (curve 2 in Fig. 3B). However, when the incubation was further continued for 10 min, an increase of absorption appeared at 240 nm. This increase can be explained by the intrinsic ⌬ 3 -⌬ 2 isomerase activity of yeast acyl-CoA oxidase (11), allowing the transfer of trans 2 double bonds to the ⌬ 3 position. In the case of 2,5,8,11,14-eicosapentenoyl-CoA, isomerization results in the formation of a conjugated ⌬ 3 -⌬ 5 double bond, which is known to absorb at 240 nm (13,14). The addition of rECH1p into this reaction mixture caused a rapid decrease of the ab-sorbance at 240 nm and a concomitant increase of the absorbance at 300 nm. This absorption spectrum is typical of ⌬ 2 -⌬ 4enoyl-CoA esters (35,42), which suggests that the conjugated ⌬ 3 -⌬ 5  the CoA ester. This being the case, the accumulated metabolite should be a substrate for 2,4-dienoyl-CoA reductase.
To investigate this possibility, we incubated arachidonoyl-CoA (Fig. 4A) in the presence of acyl-CoA oxidase and rECH1p and monitored the absorbance at 300 nm. When the proteinfree ultrafiltrate of the incubation was supplemented with NADPH, the subsequent addition of 2,4-dienoyl-CoA reductase was followed by a decrease of absorbance at 340 -385 nm (Fig.  4B). In control incubations with arachidoyl-CoA, these changes were not observed (Fig. 4, C and D), indicating that the substrate for 2,4-dienoyl-CoA reductase was produced only in the initial incubation with arachidonoyl-CoA. These results show that rECH1p transfers the conjugated double bonds from ⌬ 3,5 to ⌬ 2,4 positions of enoyl-CoAs, and thus, rECH1p is a ⌬ 3,5 -⌬ 2,4dienoyl-CoA isomerase.
Subcellular Localization-Two bands corresponding to molecular masses of 36,000 and 32,000 Da were detected by immunoblot analysis with anti-rECH1p of rat liver homogenate subjected to SDS-polyacrylamide gel electrophoresis (Fig. 5). When purified peroxisomes and mitochondria were immunoblotted, the 36,000-Da protein was associated with peroxisomes, whereas the 32,000-Da protein was enriched in the mitochondrial preparation. These data suggest that rECH1p in rat liver has dual subcellular location. To study this further, we performed immunoelectron microscopy utilizing the protein Acolloid gold-labeling technique. Gold particles localized the rECH1p to peroxisomes and mitochondria in both control and clofibrate-treated rats. To determine the relative amount of antigen in both organelles, we calculated the gold/area ratio (number of gold particles/m 2 , mean Ϯ S.D. in electron microscopy frames measured) for both organelles. The area outside the organelles was taken as background. In control (clofibratetreated) rats, the labeling density was 4.17 Ϯ 1.47 (31.2 Ϯ 13.0; p Ͻ 0.003 control versus clofibrate-treated) in mitochondria, 7.0 Ϯ 4.29 (25.7 Ϯ 9.2; p Ͻ 0.006) in peroxisomes, and 1.75 Ϯ 0.96 (0.75 Ϯ 0.96; difference not significant) in background (Fig. 6).
As controls, we used antibodies against peroxisomal MFE1 (peroxisomes) and short-chain ⌬ 3 -⌬ 2 -enoyl-CoA isomerase (mitochondria). In both cases, the antibodies recognized the ex- was added (arrow 1), and the absorption at 300 nm was monitored. After incubation for 20 min, the proteins were removed with a Millipore ultrafree polysulfone pressure-driven filters with a cut-off of 10,000 Da. Subsequently, NADPH was added into 800 l of filtrate, giving a final concentration of 0.11 mM. The oxidation of NADPH at A 340 -385 nm was monitored, and 0.01 unit of purified yeast 2,4-dienoyl-CoA reductase was added (arrow 2). Panels B and D present the incubation of filtrates from A and C, respectively. pected subcellular organelles, and the labeling density was increased in clofibrate-treated animals as compared with controls. These observations agree with previous results (29,43). DISCUSSION Published studies on the distribution of the ⌬ 3,5 -⌬ 2,4 -dienoyl-CoA isomerase in the rat including subcellular fractionation, immunoblotting, immunoelectron microscopy, enzyme activity measurements, and protein purification provide evidence that the dienoyl-CoA isomerase activity in liver is associated with both peroxisomes and mitochondria (11)(12)(13)(14)44). However, the amino acid sequences or molecular characteristics of the enzyme(s) are not known. The results of our present experiments indicate unambiguously that rECH1 encodes dienoyl-CoA isomerase. Overexpression of rECH1 in E. coli gives an enzymatically active protein, and purified recombinant protein transfers the conjugated ⌬ 3 -⌬ 5 double bond of 3,5,8,11,14-eicosapentenoyl-CoA to the ⌬ 2 -⌬ 4 position. We verified this reaction spectroscopically by following the changes of absorbance in the near UV region and by demonstrating that the product of the double-bond transfer was a substrate for 2,4-dienoyl-CoA reductase. Identification of rECH1p as a ⌬ 3,5 -⌬ 2,4 -dienoyl-CoA isomerase adds a new activity among the reactions catalyzed by the members of the hydratase/isomerase family.
At present, the only members of the hydratase/isomerase enzyme family with known structures are 2-enoyl-CoA hydratase 1 (25) and 4-chlorobenzoyl dehalogenase (26). The modeled subunit of rECH1 shows the presence of both trimerization domains similar to hydratase 1 and 4-chlorobenzoyl dehalogenase, suggesting that the enzyme forms a trimer. In line with this proposal, size exclusion chromatography gave a molecular mass of 170,000 Da. This indicates that in the native state rECH1p may be a hexamer similar to the rat mitochondrial 2-enoyl-CoA hydratase 1, which is a dimer of two trimers (25).
Prediction of the secondary structure elements of the rECH1p indicates the presence of the hydratase spiral (25) forming the CoA binding site and the pocket where catalysis takes place. It is also logical to assume that the basic folding of rECH1 would be similar to the other members of the same family. Interestingly, in the hydratase 1, the E144 (25), which acts as the catalytic base for activation of a water molecule in the hydratase, is replaced with Asp-127 in rECH1. The origin of the two immunodetected bands of molecular masses of 36,000 and 32,000 Da in the liver homogenates is intriguing. A logical explanation is provided by the results of immunoblot experiments with peroxisomes and mitochondria from rat liver. Because the band of 36,000 Da was detected in isolated peroxisomes and rECH1p has the known peroxisomal targeting signal type 1 (SKL) in its C terminus, it is possible that the product encoded by the entire open reading frame is targeted to the peroxisomal matrix and remains uncleaved. However, the first 40 amino acid residues of rECH1p resemble a mitochondrial matrix targeting signal as shown by Mitoprot II analysis (45). Thus, the band of 32,000 Da detected in tissue homogenates and in isolated mitochondria likely represents a processed mitochondrial isoform. The size difference between the 36,000 and 32,000 Da bands is consistent with the length of the predicted mitochondrial targeting signal. Thus, rECH1p is a new example of dual-organelle distribution of the same protein in mammalian cells. The immunoelectron microscopy of liver sections also provides further support for the dual distribution of rECH1p. An earlier example of this kind of intracellular targeting is the mammalian hydroxymethylglutaryl-CoA lyase (46), which also has a single mRNA translated into a polypeptide with both targeting signals. As a consequence, in mammalian tissues, a larger peroxisomal and a smaller mitochondrial hydroxymethylglutaryl-CoA lyase isoform is found. Similarly, in some species like rat and mouse, alanine:glyoxylate aminotransferase is located both in mitochondria and peroxisomes, whereas in humans it is mostly peroxisomal. In some patients with inherited deficiency of alanine:glyoxylate aminotransferase (primary hyperoxaluria type I), mutations alter the targeting of the protein so that most of it localizes to mitochondria rather than to peroxisomes (47). Interestingly, expression in Pichia pastoris of full-length rECH1 gave two bands with molecular weights corresponding to those observed in rat liver homogenates (data not shown).
The 32,000-Da size of the form of rECH1p that we localized to mitochondria is consistent with the subunit sizes previously reported for mitochondrial dienoyl-CoA isomerase (13). Furthermore, our observation that mitochondrial and peroxisomal dienoyl-CoA isomerases are actually isoforms of the same protein explains their similar kinetic properties (12). The relationship of rECH1p to the rat liver dienoyl-CoA isomerase activities of 55,000 Da (11) and 66,000 Da (12) is uncertain.
Recently Tserng et al. (48) carried out experiments applying a stable isotope-labeled substrate technique to assess the contribution of dienoyl-CoA isomerase-/dienoyl-CoA reductase-dependent pathway versus direct ␤-oxidation of cis-5-enoyl-CoA esters. C 10 substrate was metabolized completely via dienoyl-CoA isomerase-reductase pathway in both rat liver and heart mitochondria. When the chain length of the substrate was extended to C 14 , the isomerase/reductase pathway remained the major metabolic route, although its overall contribution to the metabolism decreased to 86 and 65% in liver and heart mitochondria, respectively. This experiment verified the role of isomerase/reductase pathway in rat mitochondria, and the authors also suggest that 2,4-dienoyl-CoA reductase is the ratelimiting enzyme in this pathway.
Thus far only one inborn error of the auxiliary enzymes of ␤-oxidation has been identified in humans, namely a single FIG. 6. Immunoelectron microscopic localization of rECH1 in rat liver. Liver tissues from a control (a, c, and e) or clofibrate-treated (b, d, and f) rat were immunolabeled. a and b, antibodies against rECH1. c and d, antibodies against peroxisomal multifunctional enzyme type 1 from rat liver. e and f, antibodies against mitochondrial ⌬ 3 -⌬ 2 -enoyl-CoA isomerase from rat liver. Gold particles conjugated to protein A were used to identify the attached antibodies. Mitochondria are marked with M, and peroxisomes are marked with a P. The bar is 0.25 m. patient with 2,4-dienoyl-CoA reductase deficiency (49). This paucity of clinical material might simply reflect the fact that until recently this enzyme system has been poorly appreciated, or it may be that affected individuals are not recognized.