Δ3,5,7,Δ2,4,6-Trienoyl-CoA Isomerase, a Novel Enzyme That Functions in the β-Oxidation of Polyunsaturated Fatty Acids with Conjugated Double Bonds*

The mitochondrial metabolism of unsaturated fatty acids with conjugated double bonds at odd-numbered positions,e.g. 9-cis,11-trans-octadecadienoic acid, was investigated. These fatty acids are substrates of β-oxidation in isolated rat liver mitochondria and hence are expected to yield 5,7-dienoyl-CoA intermediates. 5,7-Decadienoyl-CoA was used to study the degradation of these intermediates. After introduction of a 2-trans-double bond by acyl-CoA dehydrogenase or acyl-CoA oxidase, the resultant 2,5,7-decatrienoyl-CoA can either continue its pass through the β-oxidation cycle or be converted by Δ3,Δ2-enoyl-CoA isomerase to 3,5,7-decatrienoyl-CoA. The latter compound was isomerized by a novel enzyme, named Δ3,5,7,Δ2,4,6-trienoyl-CoA isomerase, to 2,4,6-decatrienoyl-CoA, which is a substrate of 2,4-dienoyl-CoA reductase (Wang, H.-Y. and Schulz, H. (1989)Biochem. J. 264, 47–52) and hence can be completely degraded via β-oxidation. Δ3,5,7,Δ2,4,6-Trienoyl-CoA isomerase was purified from pig heart to apparent homogeneity and found to be a component enzyme of Δ3,5,Δ2,4-dienoyl-CoA isomerase. Although the direct β-oxidation of 2,5,7-decatrienoyl-CoA seems to be the major pathway, the degradation via 2,4,6-trienoyl-CoA makes a significant contribution to the total β-oxidation of this intermediate.

sized by a Wittig reaction using the procedure of Maryanoff et al. (15). Twenty mmol of (4-carboxybutyl)triphenylphosphonium bromide in 10 ml of anhydrous tetrahydrofuran were combined with 42 mmol of lithium bis(trimethylsilyl)amide in 42 ml of tetrahydrofuran at 25°C and under N 2 with stirring. After 15 min, 16 mmol of trans-2-pentenal in 10 ml of tetrahydrofuran were added slowly. The color of the reaction changed from red to yellow. After 60 min, 50 ml of water were added to quench the reaction. The mixture was extracted with ether to remove the unreacted aldehyde. The aqueous solution was acidified with 10% HCl and extracted with ether. The combined ether extracts were extracted with water, dried over anhydrous Na 2 SO 4 , and concentrated to yield 0.82 g of 5,7-decadienoic acid (ϳ30% yield). The structure of the product was confirmed by mass spectrometry, which showed the expected [M ϩ NH 4 ϩ ] peak at 186. Enzyme and Protein Assays-Dienoyl-CoA isomerase was assayed spectrophotometrically by measuring the increase in absorbance at 300 nm on a Hitachi Model U-3000 spectrophotometer at 25°C as described by Luo et al. (10). A typical assay mixture contained 20 M 3,5-octadienoyl-CoA in 0.2 M potassium P i (pH 8.0). An extinction coefficient of 27,800 M Ϫ1 cm Ϫ1 was used to calculate rates. Trienoyl-CoA isomerase activities were determined by monitoring the absorbance increase at 337 nm. An extinction coefficient of 49,300 M Ϫ1 cm Ϫ1 was used to calculate rates (2). A typical assay mixture contained 20 M 3,5,7decatrienoyl-CoA in 0.2 M potassium P i (pH 8.0) and enzyme to give an absorbance change of 0.02/min. One unit of enzyme activity is defined as the amount of enzyme that catalyzes the conversion of 1 mol of substrate to product/min. Protein concentrations were determined as described by Bradford (16) with bovine liver albumin as the standard.
Purification of Trienoyl-CoA Isomerase from Pig Heart-Trienoyl-CoA isomerase was purified by the procedure developed for the purification of dienoyl-CoA isomerase (10). All operations were carried out at 4°C. One frozen pig heart (330 g) was minced and blended with 1 liter of 20 mM potassium P i (pH 8.3) containing 5 mM mercaptoethanol, 1 mM EDTA, 1 mM EGTA, 1 mM benzamidine, and 0.5 mM phenylmethylsulfonyl fluoride (buffer A). The resulting suspension was centrifuged at 6500 ϫ g for 20 min. Polyethylene glycol was added to the supernatant to achieve a final concentration of 10%. After keeping the suspension for 30 min, it was centrifuged at 6500 ϫ g for 20 min. The pellet was resuspended in 200 ml of buffer A containing 0.2 M KCl. After stirring it overnight, the suspension was centrifuged at 100,000 ϫ g for 1 h. The supernatant was diluted with 4 volumes of buffer A to lower the salt concentration and was applied to a Q-Sepharose column (2.5 ϫ 50 cm) previously equilibrated with buffer A. The column was extensively washed with buffer A and then eluted with 600 ml of buffer A containing 0.2 M KCl. Fractions with trienoyl-CoA isomerase activity were combined, and the pH was adjusted to 6.3 with 1 M KH 2 PO 4 . The mixture was applied to a hydroxylapatite column (2.5 ϫ 25 cm) equilibrated with 5 mM potassium P i (pH 6.3) containing 5 mM mercaptoethanol, 1 mM EDTA, 1 mM EGTA, 1 mM benzamidine, and 0.5 mM phenylmethylsulfonyl fluoride (buffer B). The column was washed with buffer B containing 0.5 M KCl and then was developed with a gradient made up of 300 ml of buffer B and 300 ml of buffer B containing 0.8 M potassium P i (pH 6.3). Fractions of 6 ml were collected, and the active fractions were combined and concentrated in an Amicon concentrator with a YM-10 membrane. After dialyzing overnight against 25 mM ethanolamine/acetic acid (pH 9.4) containing 5 mM mercaptoethanol, 1 mM EDTA, 1 mM benzamidine, 0.5 mM phenylmethylsulfonyl fluoride, and 20% glycerol (buffer C), the sample was applied to a chromatofocusing column (1.5 ϫ 25 cm) containing polybuffer exchanger 94 equilibrated with buffer C. The column was washed with buffer C and then developed with 12 column volumes of polybuffer 96 adjusted to pH 6.0 with acetic acid. Fractions of 5 ml were collected, and the active fractions were combined and concentrated. Thereafter, the sample was passed through a Sepharose CL-6B column (1.5 ϫ 90 cm) equilibrated with 10 mM potassium P i (pH 7.0) containing 1 mM EDTA, 5 mM mercaptoethanol, and 25% glycerol (buffer D). The active fractions were combined and loaded onto a reactive red 120 column (1.5 ϫ 13 cm) previously equilibrated with buffer D. After washing with buffer D, the column was eluted with 15 M 2-trans,4-trans-decadienoyl-CoA. Fractions of 2 ml were collected and assayed for both dienoyl-CoA and trienoyl-CoA isomerase activities. Active fractions were analyzed by SDS-polyacrylamide gel electrophoresis and stained with Coomassie Brilliant Blue. The intensities of the bands were determined by gel scanning with a densitometer.
Isolation of Mitochondria and Respiration Measurements-Rat liver mitochondria were isolated as described by Nedergaard and Cannon (17). For respiration measurements, 1.5 mg of rat liver mitochondria were incubated in 1.9 ml of a basal medium containing 20 mM Tris-HCl (pH 7.4), 4 mM potassium P i , 0.1 M KCl, 4 mM MgCl 2 , and 0.1 mM EGTA. To this mixture were added, in the indicated orders, bovine serum albumin (0.5 mg/ml), 0.5 mM L-carnitine, 0.5 mM L-malate, 1 mM ADP, and one of the following: 15 M linoleoyl-CoA, 15 M 9-cis,11-transoctadienoyl-CoA, 0.1 mM decanoic acid, or 0.1 mM 5,7-decadienoic acid. Rates of respiration were measured polarographically with a Clarketype oxygen electrode attached to a Yellow Springs-oxygraph.
Metabolism of 2,5,7-Decatrienoyl-CoA-The direct ␤-oxidation of 2,5,7-decatrienoyl-CoA was measured by incubating this compound at concentrations between 5 and 50 M in 0.2 M potassium P i (pH 8.0) with 1 mM NAD ϩ , 0.3 mM CoASH, and a soluble extract of rat liver mitochondria (4 g/ml). The formation of NADH was determined fluorometrically by excitation at 340 nm and by measuring the emission at 460 nm with a PTI spectrofluorometer. The conversion of 2,5,7-decatrienoyl-CoA to 3,5,7-decatrienoyl-CoA was coupled to the isomerization of the latter compound to the 2,4,6-isomer in the presence of trienoyl-CoA isomerase (0.25 units/ml). 2,5,7-Decatrienoyl-CoA, at concentrations between 5 and 60 M in 0.2 M potassium P i (pH 8.0), was incubated with a soluble extract of rat liver mitochondria (4 g/ml). The increase in absorbance at 337 nm was measured spectrophotometrically. Rates were calculated based on an extinction coefficient of 49,300 Immunoprecipitation-Dienoyl-CoA isomerase (2 g) in 50 mM potassium P i (pH 7) containing 1 mM benzamidine, 1 mM EDTA, and 5 mM mercaptoethanol was combined with various amounts of anti-dienoyl-CoA isomerase serum containing between 0 and 300 g of protein.
The total volume was 0.2 ml. This mixture was kept for 20 min at 25°C and then combined with 0.2 ml of a suspension containing 10% (w/v) S. aureus (Cowan strain). After an additional 10 min of incubation at 25°C, the mixture was centrifuged at 13,000 ϫ g for 3 min. The supernatant was transferred to clean tubes and kept at 4°C until assayed for dienoyl-CoA and trienoyl-CoA isomerase activities.
Analysis and Purification of Acyl-CoA Thioesters by HPLC-Acyl-CoAs were purified or analyzed by reverse-phase HPLC on a Waters Bondapak C 18 column (30 cm x 3.9 mm) attached to a Waters gradient HPLC system. The absorbance of the effluent was monitored at 254 nm. Separation of different acyl-CoA thioesters was achieved by linearly increasing the acetonitrile/H 2 O (9:1, v/v) content of the 50 mM ammonium phosphate buffer (pH 5.5) from 20 to 55% at a flow rate of 2 ml/min. When acyl-CoAs were purified, the desired fraction of the effluent was collected. After evaporation of acetonitrile under vacuum, the product was concentrated by use of a Sep-Pak C 18 cartridge.

Metabolism of 5,7-Decadienoyl-CoA-
The ␤-oxidation of polyunsaturated fatty acids with conjugated double bonds at odd-numbered positions is expected to produce 5,7-dienoyl-CoA intermediates. To study the metabolism of these intermediates, 5,7-decadienoyl-CoA was synthesized. The required 5,7-decadienoic acid was prepared from (4-carboxybutylidene)triphenylphosphorane and 2-trans-pentenal by a Wittig reaction (15). Since the synthetic procedure is predicted to yield mostly the trans-isomer of the newly formed double bond, the major product is expected to be 5-trans,7-trans-decadienoic acid. The CoA derivative of this acid was obtained in pure form after converting the acid to the CoA thioester and isolating the major product by HPLC (Fig. 1A).
In a preliminary experiment, the capacity of mitochondria to oxidize fatty acids with odd-numbered conjugated double bonds was assessed. The data presented in Table I demonstrate that such fatty acids supported the respiration of coupled rat liver mitochondria at rates that were slightly lower than those obtained with the corresponding fatty acids having either nonconjugated double bonds or no double bond at all. This result suggests the presence of a mitochondrial pathway for the ␤-oxidation of fatty acids that have odd-numbered conjugated double bonds.
The step-by-step degradation of 5,7-decadienoyl-CoA was studied spectrophotometrically by use of purified enzymes. The spectrum of 5,7-decadienoyl-CoA (shown in Fig. 2A, spectrum 1) is characterized by a major absorbance band centered around 230 nm and a shoulder at 260 nm. These absorbances are attributed to the diene and CoA chromophores, respec-tively. Treatment of 5,7-decadienoyl-CoA with acyl-CoA oxidase produced a 40% increase in the absorbance at 263 nm ( Fig.  2A, compare spectra 3 and 1). Such a change in absorbance agrees with the conversion of an acyl-CoA to 2-trans-enoyl-CoA. 5,7-Decadienoyl-CoA is also a substrate of medium-chain acyl-CoA dehydrogenase. In fact, it is a better substrate than is decanoyl-CoA (0.66 versus 0.45 units/mg). When 2,5,7-decatrienoyl-CoA was treated with enoyl-CoA isomerase, the absorbance at 260 nm increased, whereas the absorbance around 230 nm decreased ( Fig. 2A). These absorbance changes agree with the isomerization of a conjugated diene to a conjugated triene that would take place during the conversion of 2,5,7decatrienoyl-CoA to 3,5,7-decatrienoyl-CoA. The same absorbance changes were observed when 5,7-decadienoyl-CoA was reacted with acyl-CoA oxidase and peroxisomal multifunctional protein I, which harbors enoyl-CoA isomerase activity (Fig.  2B).
When the suspected 3,5,7-decatrienoyl-CoA was treated with a soluble extract of rat liver mitochondria or a partially purified preparation of dienoyl-CoA isomerase, an absorbance band close to 340 nm appeared, whereas the absorbance at 260 nm declined (Fig. 2C). The increase in the absorbance at 340 nm agrees with the formation of a 2,4,6-trienoyl-CoA, which was reported to have an absorbance maximum at 337 nm (2). Since 2,4,6-octatrienoyl-CoA was reported to be reduced by NADPHdependent 2,4-dienoyl-CoA reductase (2), this reaction was used to confirm the structure of the suspected 2,4,6-decatrien-oyl-CoA (Fig. 2, C, spectrum 5; and D, spectrum 1). As shown in Fig. 2D, the chromophore at 340 nm disappeared in a time-dependent manner when NADPH and 2,4-dienoyl-CoA reductase were added to the buffered solution of 2,4,6-decatrienoyl-CoA.
Further evidence for the identity of 2,4,6-decatrienoyl-CoA was obtained by its enzymatic conversion and HPLC analysis of the resultant products. For this purpose, synthetic 5,7-decadienoyl-CoA (Fig. 1A) was converted to 3,5,7-decatrienoyl-CoA by acyl-CoA oxidase and peroxisomal multifunctional protein I. The resultant two compounds (Fig. 1B) are assumed to be the 3-cis-isomer (minor peak) and 3-trans-isomer (major peak) of 3,5,7-decatrienoyl-CoA because both are converted to 2,4,6decatrienoyl-CoA by a partially purified preparation of dienoyl-CoA isomerase exhibiting trienoyl-CoA isomerase activity (Fig.  1C). The reduction of 2,4,6-decatrienoyl-CoA (Fig. 1D) by NADPH in the presence of 2,4-dienoyl-CoA reductase from Escherichia coli yielded a single product that was eluted at the same position as was the starting material (Fig. 1, compare D  and E). However, the addition of crotonase to the reaction product, presumed to be 2,6-decadienoyl-CoA because the E. coli reductase catalyzes the reduction of the 4,5-double bond, produced a more polar compound, most likely 3-hydroxydec-6enoyl-CoA (Fig. 1F). In contrast, the addition of crotonase to 2,4,6-decatrienoyl-CoA (Fig. 1D) did not produce a product of different polarity. This experiment supports the assigned structures of 2,4,6-decatrienoyl-CoA (Fig. 1D) and 2,6-decadienoyl-CoA (Fig. 1E) because 2-enoyl-CoA compounds are hydrated to a significant extent only when the 2-double bond is in conjugation with the thioester group, but not when it is part of a more extended chromophore (18).
Although 5,7-decadienoyl-CoA can be converted enzymatically to 2,4,6-decatrienoyl-CoA and further degraded after the   NADPH-dependent reduction of the latter intermediate, the operation of this pathway in mitochondria had not been demonstrated. Toward this end, the metabolism of 2,5,7-decatrienoyl-CoA by a soluble extract of rat liver mitochondria was investigated. Rates of the direct ␤-oxidation of 2,5,7-decatrienoyl-CoA in the presence of CoASH and NAD ϩ were determined by measuring fluorometrically the formation of NADH (Fig. 3,  curve A). The rates so obtained were compared with rates of 2,4,6-decatrienoyl-CoA formation determined spectrophotometrically at 340 nm in the absence of cofactors, but in the presence of excess trienoyl-CoA isomerase (Fig. 3, curve B). Under these conditions, the rate of the 2,5,7-trienoyl-CoA to 3,5,7-trienoyl-CoA isomerization is measured. Since this isomerization is, for all practical purposes, irreversible, it determines the fraction of 2,5,7-trienoyl-CoA that will be metabolized by way of the trienoyl-CoA isomerase-dependent pathway. The direct ␤-oxidation was found to be the favored pathway (Fig. 3). However, the degradation of 2,5,7-decatrienoyl-CoA via 2,4,6-decatrienoyl-CoA was found to make a significant contribution to the total metabolism of 2,5,7-decatrienoyl-CoA, accounting for almost 50% of the total flux at low and high substrate concentrations.
Identification and Characterization of Trienoyl-CoA Isomerase-The identification of trienoyl-CoA isomerase prompted its further characterization and purification. Since a partially purified preparation of dienoyl-CoA isomerase exhibited trienoyl-CoA isomerase activity, the relationship between these two enzyme activities was evaluated by an immunoprecipitation experiment. As is apparent from Fig. 4, the trienoyl-CoA isomerase activity was precipitated by antibodies raised against dienoyl-CoA isomerase. The two immunoprecipitation curves were close enough to suspect that the two enzymes might be associated with the same protein.
The relationship between trienoyl-CoA and dienoyl-CoA isomerases was further investigated by studying their behaviors during purification. The enzymes were purified from pig heart to minimize a possible interference by peroxisomal forms of these enzymes. The results of this purification effort, summarized in Table II, demonstrate the co-purification of the two enzymes. The activities of trienoyl-CoA and dienoyl-CoA isomerases remained inseparable throughout the procedure even though the dienoyl-CoA isomerase/trienoyl-CoA isomerase ratio changed from 35:1 to 143:1. The result of the last purification step, the elution of the purified enzyme from a reactive red 120 column, is shown in Fig. 5A. The elution of trienoyl-CoA isomerase from this column coincided with the appearance of dienoyl-CoA isomerase and was proportional to the amount of protein present in each fraction. Moreover, an analysis of individual column fractions by SDS-polyacrylamide gel electrophoresis revealed the presence of only one band (Fig.  5B) that corresponded to a protein with a molecular mass close to 32 kDa, which is the mass reported for dienoyl-CoA isomerase.
The co-purification of trienoyl-CoA and dienoyl-CoA isomerases as well as the co-immunoprecipitation of these two enzyme activities by antibodies raised against dienoyl-CoA isomerase strongly suggested that both enzymes reside on one protein.
The demonstration that one protein, as indicated by a single band on SDS-polyacrylamide gel electrophoresis, exhibited both trienoyl-CoA and dienoyl-CoA isomerase activities confirmed the conclusion about the association of both enzyme activities with one protein.
Dienoyl-CoA isomerase was first isolated from rat liver mi-tochondria and reported to have a subunit molecular mass of 32 kDa (10). Subsequently, peroxisomes were shown to contain a form of this enzyme that cross-reacted with antibodies raised against the mitochondrial enzyme (19). The molecular cloning of dienoyl-CoA isomerase yielded a cDNA sequence that strongly suggested peroxisomal as well as mitochondrial localizations of this enzyme (20). Evidence in support of the dual subcellular localization of this protein was obtained by immunoelectron microscopy. However, the precise terminal sequences of the mature forms of dienoyl-CoA isomerase detected in mitochondria and peroxisomes have not been reported.
Since the published cDNA sequence encodes a 36-kDa polypeptide, molecular masses Ͼ36 kDa reported for this enzyme (19,21) must be those of other polypeptides. Finally, expression of a fragment of the cDNA coding for dienoyl-CoA isomerase yielded an active protein that was crystallized and analyzed by x-ray diffraction (22). The crystal structure revealed an active-site pocket that is hydrophobic except for the side chains of three acidic residues. Two of these residues, Glu-196 and Asp-204, were proposed to facilitate the proton removal from carbon 2 and the proton addition to carbon 6,  respectively, of the substrate, 3,5-dienoyl-CoA. If 3,5,7-trienoyl-CoA binds to the same active site, the proton abstraction from carbon 2 could also be facilitated by . However, the residue involved in the protonation of carbon 8 remains to be identified.