Significance of the Reductase-dependent Pathway for the β-Oxidation of Unsaturated Fatty Acids with Odd-numbered Double Bonds MITOCHONDRIAL METABOLISM OF 2-TRANS-5-CIS-OCTADIENOYL-CoA

Abstract The β-oxidation of unsaturated fatty acids with odd-numbered double bonds proceeds by reduction of the double bond (reductase-dependent pathway) in addition to the well established isomerization of the double bond (isomerase-dependent pathway). The metabolic significance of the reductase-dependent pathway was assessed with 2-trans-5-cis-octadienoyl-CoA (2,5-octadienoyl-CoA) and its products, all of which are metabolites of α-linolenic acid. A kinetic evaluation of β-oxidation enzymes revealed that the presence of a 5-cis double bond in the substrate most adversely affected the activity of 3-ketoacyl-CoA thiolase although not enough to become rate-limiting. Concentration-dependent and time-dependent measurements indicated that most (80%) of 2,5-octadienoyl-CoA is metabolized via the isomerase-dependent pathway. The reason for the greater flux through the isomerase-dependent pathway is the higher activity of l-3-hydroxyacyl-CoA dehydrogenase as compared with Δ3,Δ2-enoyl-CoA isomerase. These two enzymes catalyze the rate-limiting steps in the isomerase-dependent and reductase-dependent pathways, respectively. Once 2,5-octadienoyl-CoA is converted to 3,5-octadienoyl-CoA (perhaps fortuitously because of the presence of Δ3,Δ2-enoyl-CoA isomerase), the only effective route for its degradation is via the reductase-dependent pathway. It is concluded that the reductase-dependent pathway assures the degradation of 3,5-dienoyl-CoA intermediates, thereby preventing the depletion of free coenzyme A and a likely impairment of mitochondrial oxidative function.

For the synthesis of 2-trans-5-cis-octadienoyl-CoA, 10 mol of 5-cisoctenoyl-CoA were incubated with 5 units of acyl-CoA oxidase from Arthrobacter species in 60 ml of 0.1 M KP i (pH 9.0) at room temperature. The conversion of 5-cis-octenoyl-CoA to 2-trans-5-cis-octadienoyl-CoA was monitored by analyzing aliquots of 50 l, withdrawn from the reaction mixture at different time intervals, by HPLC. Under the conditions described above, 3-trans-5-cis-octadienoyl-CoA was observed to be a minor product of the reaction. When 5-cis-octenoyl-CoA was completely converted to products, usually after 1 h, the reaction was terminated by adjusting the pH to 1.0 with concentrated HCl to avoid further isomerization of 2-trans-5-cis-octadienoyl-CoA to 3-trans-5-cisoctadienoyl-CoA. The protein was removed from the reaction mixture by filtration through a 0.22-m pore size membrane, and the pH was readjusted to 8.0 with 4.0 N KOH. To avoid the difficult separation of small amounts of 3,5-cis-octadienoyl-CoA from 2,5-cis-octadienoyl-CoA by HPLC, 0.5 unit of purified dienoyl-CoA isomerase from rat liver was added to the reaction mixture to isomerize 3,5-cis-octadienoyl-CoA to 2-trans-4-trans-octadienoyl-CoA, which can be separated more conveniently from 2,5-cis-octadienoyl-CoA by HPLC. After adjusting the pH to 1.0 with concentrated HCl, precipitated protein was removed from the reaction mixture by filtering the mixture through a 0.22-m pore size membrane. The reaction mixture was concentrated, after adjusting its pH to 3.0, by passing it through a Sep-Pak mini column and eluting it with methanol. Methanol was evaporated under a stream of N 2 , the resultant acyl-CoA was redissolved in 1 ml of H 2 O, and the pH was adjusted to 3.0 with concentrated HCl. 2-trans-5-cis-Octadienoyl-CoA was further purified by HPLC.
3,5-cis-Octadienoyl-CoA was synthesized by incubating 10 mol of 5-cis-octenoyl-CoA with 5 units of acyl-CoA oxidase from Arthrobacter species in 60 ml of 0.1 M KP i (pH 8.0) at room temperature. The conversion of 5-cis-octenoyl-CoA to 3,5-cis-octadienoyl-CoA was monitored by analyzing aliquots of 50 l, taken at different time intervals from the reaction mixture, by HPLC. When 5-cis-octenoyl-CoA was completely converted to products, the reaction was terminated by adjusting the pH to 1.0 with concentrated HCl, and the precipitated protein was removed by filtration through a 0.22-m pore size membrane after readjusting the pH to 5.0. The resultant acyl-CoA was concentrated by use of a SEP-Pak mini column as described above, and 3,5-cis-octadienoyl-CoA was purified by HPLC.
The synthesis of L-3-hydroxy-5-cis-octenoyl-CoA was accomplished by incubating 10 mol of 5-cis-octenoyl-CoA with 15 units of acyl-CoA oxidase from yeast and 5 units of bovine liver crotonase in 60 ml of 0.1 M KP i (pH 8.0) at room temperature. The conversion of 5-cis-octenoyl-CoA to L-3-hydroxy-5-cis-octenoyl-CoA was monitored by analyzing aliquots of 50 l, withdrawn from the reaction mixture at different time intervals, by HPLC. The reaction was terminated by adjusting the pH to 1.0 with concentrated HCl when no further increase in the formation of 3-hydroxy-5-cis-octenoyl-CoA was observed upon the addition of more enzymes.
For the purpose of synthesizing 3-keto-5-cis-octenoyl-CoA, the above reaction was carried out in the presence of 10 units of pig heart L-3hydroxyacyl-CoA dehydrogenase, 1.0 mM pyruvate, 1.0 mM NAD ϩ , and 15 units of lactate dehydrogenase. The conversion of L-3-hydroxy-5-cisoctenoyl-CoA to 3-keto-5-cis-octenoyl-CoA was monitored by analyzing aliquots of 50 l, withdrawn from the reaction mixture at different time intervals, by HPLC. The reaction was allowed to proceed for approximately 5 h. The reaction was terminated by adjusting the pH to 1.0 with concentrated HCl, and precipitated protein was removed by filtration through a 0.22-m pore size membrane. The resultant 3-keto-5-cisoctenoyl-CoA was concentrated by use of a Sep-Pak minicolumn as described above and purified by HPLC.
3-Ketooctanoyl-CoA was synthesized by incubating 5 mol of 2-octynoyl-CoA with 10 units of bovine liver crotonase in 5 ml of 50 mM MES (pH 6.0). The reaction was allowed to proceed for 45 min and was terminated by adjusting the pH to 1.0 with concentrated HCl. Precipitated protein was removed by filtration through a 0.22-m pore size membrane, and the pH was readjusted to 3.0. The product was purified by HPLC. Concentrations of all acyl-CoA thioesters were measured spectrophotometrically by quantifying CoASH with Ellman's reagent (16) after cleaving the thioesters bond with NH 2 OH at pH 7.0 (15).
Preparation of a Soluble Extract of Rat Liver Mitochondria-Rat liver mitochondria were isolated as described by Nedergaard and Cannon (17) and stored at Ϫ70°C. Mitochondria were precipitated by centrifugation and resuspended in an equal volume of 0.2 M KP i (pH 8.0) containing 20 mM 2-mercaptoethanol, 0.2% (v/v) hexamethylphosphoramide, 10 mM benzamidine, and pepstatin A (2 g/ml). The resuspended mitochondria were sonicated 5 times for 5 s each under cooling at 4°C and then were centrifuged at 100,000 ϫ g for 30 min. The supernatant represented the soluble extract of mitochondria.
Enzyme and Protein Assays-Purified crotonase from bovine liver or crotonase present in a soluble extract of rat liver mitochondria was assayed spectrophotometrically at 280 nm. A standard assay mixture contained 0.2 M KP i , (pH 8.0), 20 M 2-octenoyl-CoA or 2,5-octadienoyl-CoA, and enzyme to give an absorbance change of approximately 0.04/ min. A molar extinction coefficient of 5100 M Ϫ1 cm Ϫ1 was used to calculate rates. Purified 3-hydroxyacyl-CoA dehydrogenase or 3-hydroxyacyl-CoA dehydrogenase present in a soluble extract from rat liver mitochondria was assayed spectrophotometrically by measuring the formation of NADH at 340 nm. A standard assay mixture contained 0.2 M KP i , (pH 8.0), 20 M 2-octenoyl-CoA or 2,5-octadienoyl-CoA, 1 mM NAD ϩ , 0.3 mM CoASH, crotonase (0.17 unit), 3-ketoacyl-CoA thiolase (0.1 unit), and enzyme to give an absorbance change of approximately 0.04/min. When the mitochondrial extract served as an enzyme source, no purified thiolase was added to the assay mixture. A molar extinction coefficient of 6220 M Ϫ1 cm Ϫ1 was used to calculate rates. Purified 3-ketoacyl-CoA thiolase from pig heart or 3-ketoacyl-CoA thiolase present in a soluble extract from rat liver mitochondria was assayed by measuring the formation of product by HPLC. A standard assay contained 0.2 M KP i (pH 8.0), 0.3 mM CoASH, 20 M 3-ketooctanoyl-CoA, or 20 M 3-keto-5-octenoyl-CoA and enzyme to convert approximately 20% of the substrate to product during the 1-min incubation period. The reaction was terminated by adjusting the pH to 1.0 with concentrated HCl. After removal of precipitated protein by filtration through 0.22-m pore size membranes, the pH was readjusted to 5, and the samples were analyzed by HPLC. The products were quantified by use of standard curves established with either hexanoyl-CoA or an equilibrium mixture of 2-trans-hexenoyl-CoA and 3-hydroxyhexanoyl-CoA. With 3-keto-5octenoyl-CoA as the substrate, 3-cis-hexenoyl-CoA was formed when purified thiolase was used. However, 3-hydroxyhexanoyl-CoA and 2-trans-hexenoyl-CoA were the products when the mitochondrial extract served as the enzyme source. The thiolytic product of 3-ketooctanoyl-CoA was hexanoyl-CoA. Enoyl-CoA isomerase was assayed spectrophotometrically with the purified enzyme and 2,5-octadienoyl-CoA as the substrate. The increase in absorbance at 240 nm, due to the formation of 3,5-octadienoyl-CoA was recorded. An extinction coefficient of 18,800 M Ϫ1 cm Ϫ1 was used to calculate rates. With the same substrate but with the mitochondrial extract as an enzyme source, the increase in absorbance, due to the formation of 2,4-octadienoyl-CoA, was recorded. Purified dienoyl-CoA isomerase, which was added as a coupling enzyme in some measurements, was found to produce maximally a 20% increase in the rate. An extinction coefficient of 27,000 M Ϫ1 cm Ϫ1 was used to calculate rates. With purified enoyl-CoA isomerase and 3-octenoyl-CoA as the substrate, an increase in absorbance at 263 nm, due to the formation of 2-enoyl-CoA, was recorded. An extinction coefficient of 6700 M Ϫ1 cm Ϫ1 was used to calculate rates. A standard assay contained 0.2 M KP i (pH 8.0), 20 M 2,5-octadienoyl-CoA or 3-enoyl-CoA, and enzyme to produce an absorbance change of approximately 0.04/min. All spectrophotometric assays were performed an a Gilford, model 260, recording spectrophotometer at 25°C. Kinetic parameters (apparent V max and K m ) were obtained by nonlinear curve fitting using the Sigma plot program. One unit of enzyme activity is defined as the amount of enzyme that converts 1 mol of substrate to product per min. Protein concentrations were determined as described by Bradford (18) with bovine serum albumin as the standard.
Metabolic Studies-For rate measurements, various amounts of 2,5octadienoyl-CoA in 0.2 M KP i (pH 8.0) were incubated with a soluble extract of rat liver mitochondria. When the flux through the isomerasedependent branch was evaluated, 1 mM NAD ϩ and 0.3 mM CoASH were added, and the formation of NADH was recorded at 360 nm. An extinction coefficient of 4140 M Ϫ1 cm Ϫ1 was used to calculate rates. When rates of metabolism via the reductase-dependent pathway were determined, either no cofactors or 1 mM NAD ϩ , 0.3 mM CoASH, 1 mM pyruvate, and 1 unit of lactate dehydrogenase were present. The absorbance increase at 300 nm was recorded, which reflects the formation of 2,4-octadienoyl-CoA. An extinction coefficient of 27,000 M Ϫ1 cm Ϫ1 was used to calculate rates. The rate of 2,4-octadienoyl-CoA degradation was measured spectrophotometrically by recording the decrease of the absorbance at 300 nm due to the disappearance of the substrate. The assay mixture contained in 0.2 M KP i (pH 8.0), various amounts of 2,4-octadienoyl-CoA, 1 mM NAD ϩ , 0.3 mM CoASH, mitochondrial extract, and either 0 or 0.5 mM NADPH. Observed absorbance changes were corrected for changes detected in the absence of cofactors that most likely were due to the hydrolysis of the substrate. An extinction coefficient of 27,000 M Ϫ1 cm Ϫ1 was used to calculate rates. When the time-dependent formation of metabolites from 2,5-octadienoyl-CoA or 3,5-octadienoyl-CoA was studied, 20 M of either substrate in 0.2 M KP i (pH 8.0) was incubated with 1 mM NAD ϩ , 0.3 mM CoASH, 0.5 mM NADPH, and mitochondrial extract (0.1 mg/ml). Reactions were terminated after various periods of incubation by adjusting the pH to 1 with concentrated HCl. After readjusting the pH to 5, samples were clarified by filtration through 0.22-m pore size membranes and analyzed by HPLC. Standard curves established with purified acyl-CoAs were used to quantify the metabolites. In some experiments, a reconstituted ␤-oxidation system was used in place of the mitochondrial extract. Such incubation mixtures contained in 0.2 M KP i (pH 8.0) either 20 M 3,5-octadienoyl-CoA or 20 M 2,5-octadienoyl-CoA plus 1 mM NAD ϩ , 0.3 mM CoASH, enoyl-CoA isomerase (0.7 milliunit), enoyl-CoA hydratase (0.5 unit), 3-hydroxyacyl-CoA dehydrogenase (18 milliunits), and 3-ketoacyl-CoA thiolase (0.11 unit). Samples were processed and analyzed by HPLC as described above.
Analysis and Purification of Acyl-CoA Thioesters by HPLC-Acyl-CoA substrates and metabolites were purified and analyzed by reversephase HPLC on a Waters Bondapak C 18 column (30 cm ϫ 3.9 mm) attached to a Waters gradient HPLC system. The absorbance of the eluate was monitored at 254 nm. Separation of metabolites was achieved by first washing the column with 50 mM ammonium phosphate (pH 5.5) containing 5% of acetonitrile/water (9:1, v/v) for 15 min and then eluting acyl-CoAs by linearly increasing the organic phase from 5 to 50% in 30 min at a flow rate of 2 ml/min. When the kinetics of purified 3-ketoacyl-CoA thiolase with 3-keto-5-octenoyl-CoA as a substrate were studied, an isocratic elution was used for 7 min. Thereafter, a 5-50% gradient was developed within 13 min. With 3-ketooctanoyl-CoA as substrate, the separation of substrate and product was achieved by linearly increasing the methanol content of the 75 mM ammonium phosphate (pH 5.5) elution buffer from 30 to 60% in 30 min at a flow rate of 2.5 ml/min. When the sample contained only acyl-CoAs, the isocratic part of the elution program was omitted, and a linear gradient from 10 to 50% was used to achieve elution within 30 min at a flow rate of 2 ml/min.

Does the 5-cis Double Bond Affect the ␤-Oxidation of 5-Enoyl-
CoAs?-The ␤-oxidation of unsaturated fatty acids with oddnumbered double bonds yields 5-enoyl-CoA intermediates, which are converted to 2,5-dienoyl-CoA by acyl-CoA dehydrogenase (3). The focus of this study was the mitochondrial metabolism of 2,5-octadienoyl-CoA. The degradation of 2,5-octadienoyl-CoA, which is a metabolite of ␣-linolenic acid, involves only soluble enzymes that are located in the matrix of mitochondria. Hence, a soluble extract of mitochondria represents a suitable enzyme system to study the metabolism of this compound. The direct ␤-oxidation of 2,5-octadienoyl-CoA requires the sequential actions of crotonase, 3-hydroxyacyl-CoA dehydrogenase, and 3-ketoacyl-CoA thiolase (see Fig. 1, pathway A).
In an attempt to determine if the 5-cis double bond affects the rate at which 2,5-octadienoyl-CoA is metabolized via the isomerase-dependent pathway (Fig. 1, pathway A), the apparent K m and V max values were determined for these three ␤-oxidation enzymes using substrates with and without the 5-cis double bond. The results, shown in Table I, demonstrate that the 5-cis double bond affects the activities of the three enzymes differently. With enoyl-CoA hydratase, a 3-fold higher V max was observed when the substrate contained the 5-cis double bond, while the K m was virtually unaffected. The opposite observation was made with 3-hydroxyacyl-CoA dehydrogenase, which yielded a 4-fold higher K m value for the unsaturated as compared with the saturated substrate. However, the maximal velocity of this enzyme was unaffected by the presence of the 5-cis double bond. The most adverse effect was detected with 3-ketoacyl-CoA thiolase, which exhibited with the unsaturated substrate only one-sixth of the maximal velocity obtained with the saturated substrate. In addition, the K m value was 5-fold higher when the 5-cis double bond was present in the sub- strate. These data demonstrate that 5-cis-enoyl-CoA intermediates can be directly metabolized by ␤-oxidation, but perhaps more slowly than the corresponding saturated metabolites.
Metabolism of 2,5-Octadienoyl-CoA via the Isomerasedependent and Reductase-dependent Pathways-Since 2,5-octadienoyl-CoA can be metabolized by both the isomerase-dependent (Fig. 1A) and reductase-dependent (Fig. 1B) pathways, it was important to determine the relative fluxes via these two routes in mitochondria. For estimating the flux through the isomerase-dependent pathway, rates of NADH formation in the presence of NAD ϩ and CoASH but in the absence of NADPH were measured spectrophotometrically with a soluble extract of rat liver mitochondria (see Fig. 2A). Since initial velocities were determined, the simultaneous entry of substrate into the reductase-dependent pathway was assumed to have little or no effect on the rates of degradation via the isomerase-dependent pathway. Flux through the reductase-dependent pathway was determined by measuring spectrophotometrically the formation of 2,4-dienoyl-CoA at 300 nm (see Fig. 2B) in the absence of coenzymes. A comparison of the data shown in Fig. 2, curves A and B, indicates that 2,5-octadienoyl-CoA is metabolized mostly (80%) by the isomerase-dependent pathway and to a lesser extent (20%) via the reductase-dependent route. When the rate of 2,4-dienoyl-CoA formation was measured in the presence of NAD ϩ , CoASH, and pyruvate plus lactate dehydrogenase to reoxidize NADH, rates were lower by approximately 50% than those observed in the absence of coenzymes (compare curves B and C of Fig. 2). This decline in rates is attributed, at least in part, to the ␤-oxidation of 2,4-dienoyl-CoA via route C (see Fig. 1), which requires the involvement of enoyl-CoA isomerase and dienoyl-CoA isomerase, but not the reductase. The direct ␤-oxidation of 2,4-dienoyl-CoA, as outlined in Fig.  1C, is possible, albeit at a low rate, because this intermediate has the 2-trans-4-trans configuration, which makes it a substrate of the mitochondrial ␤-oxidation system in contrast to the 2-trans-4-cis-isomer, which is not directly oxidized (14). The time-dependent ␤-oxidation of 2,5-octadienoyl-CoA by a mitochondrial extract from rat liver mitochondria in the presence of NAD ϩ , CoASH, and NADPH was analyzed by HPLC (see Fig.  3). Striking was the almost instantaneous hydration of 2,5octadienoyl-CoA to 3-hydroxy-5-octenoyl-CoA, the first metabolite of the isomerase-dependent pathway (see Fig. 3A). Purified enoyl-CoA hydratase was used to determine the  3. HPLC analysis of metabolites formed from 2,5-octadienoyl-CoA by a soluble extract of rat liver mitochondria in the presence of NAD ؉ , CoASH, and NADPH. Shown are products detected after 5 s (A), 1 min (B), and 5 min (C). Peaks identified by use of authentic compound were as follows: 2,5-octadienoyl-CoA (⌬ 2,5 ); 3-hydroxy-5-octenoyl-CoA (3OH⌬ 5 ); butyryl-CoA (C 4 ); 3-keto-5-octenoyl-CoA (3keto-⌬ 5 ); 3,5-octadienoyl-CoA (⌬ 3,5 ); 2,4-octadienoyl-CoA (⌬ 2,4 ); and hexanoyl-CoA (C 6 ). For details, see "Experimental Procedures." ␤-Oxidation of Unsaturated Fatty Acids equilibrium ratio of 3-hydroxy-5-octenoyl-CoA to 2,5-octadienoyl-CoA. This ratio was found to be 6.4:1. One minute into the reaction, significant amounts of 3-keto-5-octenoyl-CoA and butyryl-CoA, the final product of the isomerase-dependent pathway, were detected (see Fig. 3B). At the same time, only small quantities of metabolites belonging to the reductase-dependent pathway had been formed (marked ⌬ 3,5 and ⌬ 2,4 in Fig. 3B). Hexanoyl-CoA, the final product of the reductase-dependent pathway, was detected toward the end of the reaction, when it constituted 18% of butyryl-CoA (see Fig. 3C). Since butyryl-CoA is also formed by the direct ␤-oxidation of 2,4-dienoyl-CoA, approximately 20% of 2,5-octadienoyl-CoA is metabolized by the reductase-dependent pathway. The kinetic parameters (apparent V max and K m ) of several ␤-oxidation enzymes that are present in a soluble extract of rat liver mitochondria were determined with substrates having 5-cis double bonds. As is apparent from the data shown in Table II, the activity of enoyl-CoA hydratase with 2,5-octadienoyl-CoA as a substrate is at least 50-fold higher than the activity of the second most active enzyme in this group of ␤-oxidation enzymes in rat liver mitochondria. The hydratase activity is sufficiently high to maintain an equilibrium or a near equilibrium situation between 2,5-octadienoyl-CoA and its product of hydration. As a consequence, the effective concentration of 2,5-octadienoyl-CoA in the isomerization reaction was only 14% of the expected concentration based on the flux through the pathway. Of the first three enzymes of pathway A (see Fig. 1), the least active one was 3-hydroxyacyl-CoA dehydrogenase. However, the specific activity of this enzyme was 6 times higher than the activity of enoyl-CoA isomerase, which is the enzyme that determines the rate at which intermediates enter the reductasedependent pathway (see Fig. 1B). The relative activities of the rate-determining enzymes of the isomerase-dependent and reductase-dependent pathway explain the observed greater flux through the former than through the latter pathway. If pathway A would be inhibited, then a larger percentage of 2,5-octadienoyl-CoA might enter pathway B. This hypothesis was tested by evaluating the effect of NADH on the fluxes through pathways A and B. Since NADH is a product inhibitor of 3-hydroxyacyl-CoA dehydrogenase, the flux through pathway A should be inhibited, and more of the intermediate might enter pathway B. The results shown in Fig. 4, specifically the reduced formation of butyryl-CoA, prove the predicted inhibition of the flux through pathway A. However, the formation of hexanoyl-CoA via pathway B also was reduced, although only slightly, so that almost equal amounts of butyryl-CoA and hexanoyl-CoA were formed when either 0.5 or 1 mM NADH were present in the incubation mixture (see Fig. 4, B and C).
Metabolism of 3,5-Octadienoyl-CoA-The isomerization of 2,5-octadienoyl-CoA to 3,5-octadienoyl-CoA reduces the flux through the isomerase-dependent pathway (pathway A) unless the isomerization is freely reversible. Since the equilibrium concentration of 3,5-octadienoyl-CoA was found to be at least 20 times higher than the concentration of the 2,5-isomer, the rate of the ⌬ 3,5 3⌬ 2,5 isomerization was expected to be slow. An analysis of metabolites formed from 3,5-octadienoyl-CoA by a soluble extract of rat liver mitochondria in the presence of NAD ϩ and CoASH revealed the rapid conversion of 3,5-octadienoyl-CoA to 2,4-octadienoyl-CoA (see Fig. 5A) and the slower appearance of butyryl-CoA (see Fig. 5, B and C). This pattern of metabolite formation is indicative of the degradation of 3,5octadienoyl-CoA via pathway C, shown in Fig. 1. The absence of metabolites characteristic of pathway A suggests that butyryl-CoA was not formed via pathway A after the conversion of 3,5-octadienoyl-CoA to 2,5-octadienoyl-CoA. When NADPH was present besides NAD ϩ and CoASH, hexanoyl-CoA and butyryl-CoA were formed in a ratio of 3:2 (see Fig. 5D). A similar ratio was observed when the rate of 2,4-octadienoyl-CoA utilization by a soluble extract of rat liver mitochondria was determined in the presence of either NAD ϩ , NADPH, and CoASH (B and C) or NAD ϩ and CoASH (C) (see Fig. 6). This experiment revealed that 2,4-octadienoyl-CoA can be ␤-oxidized directly (see Fig. 1, pathway C) in addition to being degraded after the NADPH-dependent reduction of one double bond (see Fig. 1, pathway B). The fluxes through the two pathways are of similar magnitude, with the reductive branch facilitating a slightly faster flow than the direct oxidation. Although the above mentioned results are indicative of the degradation of 3,5-octadienoyl-CoA via pathways B and C, a possible contribution of pathway A cannot be excluded. The possibility of a slow degradation of 3,5-octadienoyl-CoA via pathway A (see Fig. 1) was evaluated by use of a ␤-oxidation system reconstituted from purified enzymes. All soluble ␤-oxidation enzymes necessary to degrade 3,5-octadienoyl-CoA via pathway A were combined at activity levels observed in a soluble extract of rat liver mitochondria. The absence of dienoyl-CoA isomerase prevented the degradation via pathways B and C. When 2,5-octadienoyl-CoA was incubated with such a reconstituted ␤-oxidation system in the presence of NAD ϩ and CoASH, its rapid degradation via pathway A was observed, as indicated by the almost complete disappearance of the starting material within 1 min (see Fig. 7A). In contrast, 3,5-octadienoyl-CoA, when incubated under identical conditions, remained mostly unaltered after the same incubation time (see Fig. 7B).

␤-Oxidation of Unsaturated Fatty Acids
However, small amounts of butyryl-CoA and 3-keto-5-octenoyl-CoA so detected are indicative of a slow degradation of 3,5octadienoyl-CoA via pathway A.

DISCUSSION
The recognition that 2,5-dienoyl-CoAs, which are metabolites of unsaturated fatty acids with odd-numbered double bonds, can be degraded by the reductase-dependent pathway (see Fig. 1B) (3), raised the question as to whether the isomerasedependent pathway (see Fig. 1A) is operative at all. If it is, the relative fluxes through both pathways needed to be determined. Should the reductase-dependent pathway not be essential for the rapid ␤-oxidation of unsaturated fatty acids, then its metabolic function needs to be elucidated. The results of this study as well as data presented by Tserng et al. (7) demonstrate that 2,5-dienoyl-CoAs can be degraded via the isomerase-dependent pathway. A kinetic study of enoyl-CoA hydratase, 3-hydroxyacyl-CoA dehydrogenase, and 3-ketoacyl-CoA thiolase revealed that the presence of a 5-cis double bond in the  Fig. 1). C, degradation in the presence of NAD ϩ and CoASH via pathway C (see Fig. 1). For details, see "Experimental Procedures." that it refers to the ␤-oxidation of unsaturated fatty acid with odd-numbered double bonds via a sequence of reactions with 2,4-dienoyl-CoA as intermediate.