Intermediate channeling on the trifunctional beta-oxidation complex from pig heart mitochondria.

The kinetic properties of the purified trifunctional β-oxidation complex (TOC) from pig heart mitochondria were analyzed with the aim of elucidating the functional consequence of having three sequentially acting enzymes of β-oxidation associated in one complex. The kinetic parameters of TOC and of the component enzymes of TOC, long-chain enoyl-CoA hydratase, long-chain 3-hydroxyacyl-CoA dehydrogenase, and long-chain 3-ketoacyl-CoA thiolase, were determined with substrates having acyl chains with 16 carbon atoms. Quantification by high performance liquid chromatography of intermediates formed during the degradation of 2-trans-hexadecanoyl-CoA to myristoyl-CoA and acetyl-CoA by TOC revealed the accumulation of 3-hydroxyhexadecanoyl-CoA, whereas 3-ketohexadecanoyl-CoA was undetectable. The observed rates of NADH and acetyl-CoA formation were higher than the theoretical rates calculated by use of the kinetic parameters and measured concentrations of intermediates. When the sequence of reactions catalyzed by TOC was inhibited by acetyl-CoA, the steady-state concentration of the 3-hydroxyacyl-CoA intermediate was not affected, whereas a small amount of 3-ketohexadecanoyl-CoA was detected. The differences between observed and predicted reaction rates and between measured and expected concentrations of intermediates are best explained by the operation of a channeling mechanism. As a consequence of intermediate channeling between the active sites on the complex, more coenzyme A is available in the mitochondrial matrix and metabolites like 3-ketoacyl-CoA thioesters, which are strong inhibitors of several β-oxidation enzymes, do not accumulate.

The apparent absence of intermediates of fatty acid oxidation from mitochondria (1) prompted the suggestion that the soluble matrix enzymes, which catalyze the individual reactions of this pathway, may exist as a multienzyme complex (2,3). Subsequent studies with radioactively labeled fatty acids revealed the formation of low levels of chain shortened acyl-CoAs in respiring mitochondria (4 -8). In some studies only acyl-CoA intermediates were identified (4,5) that were produced by one or several full cycles of ␤-oxidation (referred to as saturated intermediates). Other investigators additionally identified even smaller amounts of 2-enoyl-CoA and 3-hydroxyacyl-CoA (6 -8). However, it is possible that intermediates or certain intermediates are formed only in damaged mitochondria. To avoid such artifacts, quantification of intermediates are best done with whole cells, which contain undamaged mitochondria. In a recent study of ␤-oxidation with human fibroblasts, only saturated short-chain and medium-chain intermediates were detected in the incubation medium (9). However, with fibroblasts that are deficient with respect to certain ␤-oxidation enzymes, different intermediate patterns were observed that were indicative of the enzyme defects. For example, a deficiency of long-chain 3-hydroxyacyl-CoA dehydrogenase resulted in the accumulation of long-chain saturated as well as 3-hydroxy intermediates. Thus, it seems that under normal conditions long-chain intermediates do not accumulate in fibroblasts and that of all possible short-chain and medium-chain intermediates only saturated ones escape from these cells, presumably because of their intramitochondrial accumulation. One likely reason for the absence of free intermediates is their channeling between active sites of enzymes catalyzing consecutive reactions. Such channeling would most likely occur if the enzymes of ␤-oxidation were organized as multienzyme complexes. Although this has not been demonstrated for the soluble matrix enzymes, it has been established for long-chain enoyl-CoA hydratase (EC 4.2.1.74), long-chain L-3-hydroxyacyl-CoA dehydrogenase, and long-chain 3-ketoacyl-CoA thiolase, which exist as a trifunctional ␤-oxidation complex (TOC) 1 in the inner mitochondrial membrane (10 -12). Since, additionally, very long-chain acyl-CoA dehydrogenase is an enzyme of the inner mitochondrial membrane (13), these four enzymes may act in concert with the result that long-chain intermediates are channeled between their active sites and do not accumulate in the mitochondrial matrix.
In this study, purified TOC was studied with the aim of detecting any channeling of intermediates. The results reveal that intermediates are either not formed or accumulate at levels insufficient to account for the rate of the overall reaction. This finding agrees with the existence of a channeling mechanism, which would explain the intermediate spectrum observed with whole cells. anhydride method as detailed by Fong and Schulz (15). 2-Hexadecynoyl-CoA, after purification by hydrophobic chromatography on octyl-Sepharose (16), was converted to 3-ketohexadecanoyl-CoA by incubating 1 mM 2-hexadecynoyl-CoA in 20 mM HEPES buffer (pH 7) with crotonase (40 units/ml) for 1 h at 25°C as described in principle by Thorpe (17). L-3-Hydroxyhexadecanoyl-CoA was prepared by incubating 1 mM 2-trans-hexadecenoyl-CoA in 0.1 M KP i (pH 7.6) with crotonase (10 units/ml) at 25°C for 30 min. All of the above substrates were purified by HPLC. Concentrations of most acyl-CoA thioesters were determined by measuring released CoASH by the method of Ellman (18) after quantitatively cleaving the thioester bond with 1 M hydroxylamine at pH 7. The concentration of 3-ketohexadecanoyl-CoA was determined by measuring its complete reduction by NADH at pH 7 in the presence of L-3-hydroxyacyl-CoA dehydrogenase. The concentration of purified L-3-hydroxyhexadecanoyl-CoA also was calculated based on its absorbance at 259 nm and an extinction coefficient of 15,400 M Ϫ1 cm Ϫ1 .
Enzyme Assays and Protein Determination-The three coupled reactions (overall reaction) catalyzed by TOC were assayed either by measuring spectrophotometrically the formation of NADH or by determining the concentration of acetyl-CoA or myristoyl-CoA by HPLC. A standard assay mixture contained in 1 ml of 0.1 KP i (pH 7.6) 1 mM NAD ϩ , 0.2 mM CoASH, 20 M 2-trans-hexadecenoyl-CoA, and 3 g of TOC. Long-chain enoyl-CoA hydratase of TOC was assayed by either the direct or indirect method. The direct method is based on the decrease in absorbance at 280 nm due to the hydration of 2-enoyl-CoA. A standard assay mixture contained in 0.1 M KP i (pH 7.6) 20 M 2-trans-hexadecenoyl-CoA and TOC to give an absorbance change of approximately 0.02 A/min. The molar extinction coefficient for calculating rates is 5,100 M Ϫ1 cm Ϫ1 (19). The indirect method is based on a coupled assay, in which L-3-hydroxyacyl-CoA formed by the hydration of 2-trans-enoyl-CoA is dehydrogenated and thiolytically cleaved by the combined actions of L-3-hydroxyacyl-CoA dehydrogenase and 3-ketoacyl-CoA thiolase in the presence of NAD ϩ plus CoASH. A standard assay mixture contained 0.1 M KP i (pH 7.6) 1 mM NAD ϩ , 0.2 mM CoASH, 20 M 2-trans-hexadecenoyl-CoA, pig heart L-3-hydroxyacyl-CoA dehydrogenase (2 units/ml), pig heart 3-ketoacyl-CoA thiolase (0.1 unit/ml), and TOC to give an absorbance change of approximately 0.02 A/min at 340 nm. Long-chain L-3-hydroxyacyl-CoA dehydrogenase of TOC was assayed by measuring the formation of NADH spectrophotometrically at 340 nm. A standard assay mixture contained 0.1 M KP i (pH 7.6) 1 mM NAD ϩ , 0.2 mM CoASH, 20 M L-3-hydroxyhexadecanoyl-CoA, pig heart 3-ketoacyl-CoA thiolase (0.1 unit/ml), and TOC to give an absorbance change of approximately 0.02 A/min. Long-chain 3-ketoacyl-CoA thiolase of TOC was assayed by determining the concentration of acetyl-CoA or myristoyl-CoA by HPLC. A standard assay mixture of 1 ml contained 0.1 M KP i (pH 7.6) 0.2 mM CoASH, 20 M 3-ketohexadecanoyl-CoA, and 3 g of TOC. After 45 s of incubation, the reaction was terminated by adjusting the pH to 1-2 with concentrated HCl. An aliquot (250 l) of the reaction was subjected to HPLC analysis to determine the concentration of acetyl-CoA or myristoyl-CoA. Kinetic parameters (K m , V max ) were determined by nonlinear curve-fitting using the SigmaPlot program. One unit of enzyme activity is defined as the amount of enzyme that catalyzed the conversion of 1 mol of substrate to product/min. Protein concentrations were determined as described by Wang and Smith (20).
HPLC Analyses-Prior to analysis by HPLC, reactions were terminated by adjusting the pH to 1-2 with concentrated HCl. Samples were filtered through 0.22-m (pore size) membranes, after which the pH was adjusted to 5.5 with KOH. The filtrates were applied to a Waters Bondapak C 18 reverse-phase column (30 cm ϫ 3.9 mm) attached to a Waters gradient HPLC system. The absorbance of the effluent was monitored at 254 nm. Separation was achieved by linearly increasing the acetonitrile/H 2 O (9/1) content of the 50 mM ammonium phosphate elution buffer (pH 5.5) from 5% to 75% in 40 min at a flow rate of 2 ml/min. For the purification of different synthetic acyl-CoA substrates, the acetonitrile/H 2 O (9/1) content of the buffer was linearly increased from 5% to 50% in 15 min. For the purification of acetyl-CoA, the acetonitrile/H 2 O (9/1) content was increased from 5% to 25% in 15 min. Concentrations of intermediates and products were determined by use of standard curves that were established with HPLC-purified acyl-CoA thioesters.

RESULTS
Kinetic Properties of TOC-The kinetic parameters (K m ,V max ) of TOC were determined with long-chain substrates at fixed coenzyme concentrations that were saturating (1 mM NAD ϩ ) or close to the level known to exist in respiring mito-chondria (0.2 mM CoASH) (21). The sequence of the three reactions catalyzed by TOC, referred to as overall reaction (see Fig. 1), was measured either by recording spectrophotometrically the reduction of NAD ϩ or by measuring the formation of acetyl-CoA by HPLC. Since both methods yielded the same results (data not shown), the more convenient spectrophotometric assay was used to determine the apparent K m and V max values for the overall reaction. In addition, the kinetic parameters of the individual reactions catalyzed by long-chain enoyl-CoA hydratase, long-chain 3-hydroxyacyl-CoA dehydrogenase, and long-chain 3-ketoacyl-CoA thiolase of TOC were determined. The results, shown in Table I, indicate that the dehydrogenase catalyzes the slowest reaction in the sequence and that the apparent K m values for 2-hexadecenoyl-CoA in the first and overall reactions are similar but lower than the K m values for the substrates of the second and third reactions. The K m value of 3.9 M for 3-hydroxyhexadecanoyl-CoA in the dehydrogenase reaction was obtained by assuming that the substrate was in equilibrium with its dehydration product, whereas the K m value of 5.5 M is based on the assumption that the substrate was not dehydrated.
Determination of Intermediates and Comparison of Actual and Calculated Reaction Rates-The substrate, intermediates, and products of the TOC-catalyzed overall reaction were separated by HPLC and quantified by use of standard curves established for each of the compounds with HPLC-purified acyl-CoA thioesters. As is apparent from Fig. 2A, only the first intermediate, 3-hydroxyhexadecanoyl-CoA (I 1 ) was detected besides the substrate (⌬ 2 C 16 ) and the two products myristoyl-CoA (C 14 ) and acetyl-CoA (C 2 ). The second intermediate, 3-ketohexadecanoyl-CoA (I 2 ), would have been detected, because it can be separated from all other compounds under the conditions used in this experiment (see Fig. 2B). The quantitative recovery of acyl-CoA thioesters was demonstrated by adding pentadecanoyl-CoA to the reaction mixture and determining its concentration by HPLC (data not shown). The concentration of intermediate I 1 , 3-hydroxyhexadecanoyl-CoA, was determined as a function of the incubation time. As can be seen from Fig.  3A, the concentration of I 1 increased during the first minute of the reaction but declined thereafter. The measured concentration of I 1 and the kinetic parameters of the dehydrogenase (see Table I) were used to calculate rates of NADH formation catalyzed by long-chain 3-hydroxyacyl-CoA dehydrogenase. The calculated formation of NADH, based on the concentration of free 3-hydroxyhexadecanoyl-CoA in the reaction mixture, is significantly lower than the observed formation of NADH (see Fig. 3B). Lines 1 and 2 in Fig. 3B represent the lower and upper limits, respectively, of the theoretical NADH formation based Since the observed rate of NADH formation was higher than either of the calculated rates, the effective concentration of I 1 must be higher than the concentration of I 1 in the bulk phase. Such condition could be achieved if I 1 were channeled from the active site of long-chain enoyl-CoA hydratase to that of longchain 3-hydroxyacyl-CoA dehydrogenase. The presence of I 1 in the bulk phase may be due to an excess capacity of the hydratase and a leaky channeling mechanism.
In contrast to the leaky channeling of I 1 , 3-ketohexadecanoyl-CoA (I 2 ) was not detected at any time during the course of the overall reaction under the same conditions at which free I 1 was formed. Since 0.5 M I 2 or less could have been detected, the observed formation of acetyl-CoA was compared with the calculated formation supported by 0.5 M and 0.1 M I 2 . As shown in Fig. 4, the observed rate is much higher than the expected rate in the presence of 0.5 M I 2 in the bulk phase.
This discrepancy between observed and predicted rates supports the hypothesis of intermediate channeling between the active sites of the dehydrogenase and thiolase of TOC.
Effect of Acetyl-CoA on the Formation of Intermediates and Products of the TOC-catalyzed Reaction Sequence-The product inhibition of the TOC-catalyzed overall reaction by acetyl-CoA was studied with the aim of evaluating the proposed intermediate channeling. Acetyl-CoA inhibited the formation of myristoyl-CoA from 2-hexadecenoyl-CoA by 50% when the acetyl-CoA concentration was raised from zero to 1 mM (see Fig.  5). However, the effect of this inhibitor on the concentration of intermediates was limited. Noteworthy is the accumulation of 0.5 M I 2 in the presence of 1 mM acetyl-CoA, whereas this intermediate was not detected in the absence of acetyl-CoA. A small amount of I 2 also accumulated when the rate of the thiolase-catalyzed reaction was reduced by lowering the con- a Overall reaction, the three coupled reactions catalyzed by TOC as measured by the formation of NADH; hydratase, long-chain enoyl-CoA hydratase; dehydrogenase, long-chain L-3-hydroxyacyl-CoA dehydrogenase; thiolase, long-chain 3-ketoacyl-CoA thiolase. For experimental details, see "Experimental Procedures." b Apparent K m and V max values are means of two determinations, which differed by 12% or less.
c The K m value of 5.5 M was obtained by assuming that the substrate was not dehydrated, whereas the K m value of 3.9 M is based on the assumption that the equilibrium of the dehydration/hydration was reached instantaneously. centration of CoASH from 0.2 mM to 0.05 mM (data not shown). However, the concentration of I 1 changed insignificantly even though the hydration of 2-hexadecenoyl-CoA had not reached the equilibrium. In fact, the observed concentration of I 1 was less than 50% of its equilibrium concentration. Acetyl-CoA, the product of the last of the three sequential reactions, was expected to inhibit thiolase. This assumption was proven to be correct by demonstrating that acetyl-CoA inhibited the thiolytic cleavage of 3-ketohexadecanoyl-CoA to myristoyl-CoA and acetyl-CoA (see Fig. 6A). Since 3-keto-acyl-CoAs, as for example acetoacetyl-CoA, are known inhibitors of 3-hydroxyacyl-CoA dehydrogenase (22), the effect of 3-ketohexadecanoyl-CoA on the reduction of NAD ϩ in the overall reaction was determined. As shown in Fig. 6B, 3-ketohexadecanoyl-CoA at low micromolar concentrations inhibited the dehydrogenation of 3-hydroxyhexadecanoyl-CoA, the slowest reaction in the reaction sequence. This inhibition could explain the decreased formation of NADH by acetyl-CoA in the overall reaction, because the latter compound inhibits thiolase with the result that I 2 accumulates, which in turn inhibits the dehydrogenase-catalyzed formation of NADH. However, if intermediate channeling occurs, the concentration of bound intermediate would most likely be higher than reflected by the concentration of free intermediate. Hence the degree of inhibition might be higher than could be accounted for by the concentration of the free intermediate. This seems to be the situation when the inhibition of the overall reaction by acetyl-CoA is analyzed. Although 1 mM acetyl-CoA caused an inhibition of the overall reaction by 50% (see Fig. 6A), the accumulation of 0.5 M I 2 only explains a 30% decrease of the dehydrogenase-catalyzed reaction (see Fig. 6B), which is limiting the overall reaction. The differences between the observed and predicted degrees of inhibition are attributed to intermediate channeling. DISCUSSION The recent characterization of several long-chain specific ␤-oxidation enzymes (10 -13) has made it necessary to modify the traditional view of how the enzymes of ␤-oxidation cooperate to completely degrade fatty acids. As schematically shown in Fig. 7, a set of four long-chain specific enzymes, located in the inner mitochondrial membrane and consisting of very longchain acyl-CoA dehydrogenase, long-chain enoyl-CoA hydratase, long-chain 3-hydroxyacyl-CoA dehydrogenase, and long-chain 3-ketoacyl-CoA thiolase, are presumed to catalyze the chain shortening of long-chain fatty acyl-CoAs. After one or several rounds of ␤-oxidation, the soluble matrix enzymes, consisting of short-chain, medium-chain and perhaps longchain acyl-CoA dehydrogenases in addition to enoyl-CoA hydratase (crotonase), 3-hydroxyacyl-CoA dehydrogenase, and 3-ketoacyl-CoA thiolase, take over to complete the degradation of acyl-CoAs. This revised hypothetical view of mitochondrial ␤-oxidation necessitates a reevaluation of the proposed control mechanism(s) of ␤-oxidation, especially of the energy-linked regulation in extra-hepatic tissues, e.g. in heart (23). Changes in the energy demand of a tissue oxidizing fatty acids were thought to cause changes in the concentrations of ␤-oxidation intermediates which in turn may control the activity of the pathway by regulating activities of key enzymes (23). However, the question of whether and if so, to which degree intermediates accumulate during ␤-oxidation in vivo has not been answered unambiguously.
The evidence presented here prompts the conclusion that long-chain intermediates of mitochondrial ␤-oxidation are channeled between the active sites of TOC. This conclusion does not exclude the possibility of intermediates dissociating from TOC under certain conditions and exiting from mitochondria and even cells. For example, 3-hydroxyacyl-CoAs have been reported to accumulate in isolated mitochondria and to exit from them as acylcarnitines, especially when the reoxidation of NADH is impaired (4 -8) or when long-chain 3-hydroxyacyl-CoA dehydrogenase is deficient (24). In contrast, the accumulation of 3-ketoacyl-CoAs has not been observed. The observed accumulation of a relatively large quantity of 3-hydroxyhexadecanoyl-CoA (I 1 ) in this study may be a consequence of the specific experimental set-up with 2-hexadecenoyl-CoA serving as a substrate instead of hexadecanoyl-CoA and with the activity of long-chain 3-hydroxyacyl-CoA hydratase being high compared to the other activities of TOC.
This study prompts the conclusion that the channeling of long-chain ␤-oxidation intermediates is most likely the underlying cause for their absence from the mitochondrial matrix or for their presence at very low levels. This situation also would explain why long-chain intermediates of ␤-oxidation were not detected in the extracellular fluid of fibroblasts unless an enzyme defect like long-chain 3-hydroxyacyl-CoA dehydrogenase deficiency impaired the further metabolism of long-chain intermediates (9). If experiments with isolated mitochondria and whole cells yield different results with respect to the formation of ␤-oxidation intermediates, the results obtained with whole cells may be more relevant to the in vivo situation because mitochondria present in whole cells, in contrast to isolated ones, are most likely undamaged. However, when extracellular or extra-mitochondrial levels of intermediates are measured, the question arises of whether they truly reflect the intramitochondrial concentrations of acyl-CoAs. In fact the available evidence indicates that this may not be the case (7).
In considering the metabolic consequences of the non-accumulation of ␤-oxidation intermediates, the availability of more free coenzyme A in the mitochondrial matrix is perhaps most important. If each intermediate of ␤-oxidation were present in the matrix only at a low micromolar concentration, a substantial amount of the available CoASH would be tied up by the 27 intermediates that are formed along the pathway from palmitoyl-CoA to acetyl-CoA. Moreover, if ␤-oxidation intermediates accumulate, they might inhibit various mitochondrial enzymes, especially those with binding sites for acyl-CoAs. In fact, 3-ketoacyl-CoA intermediates at nanomolar concentra-tions are effective inhibitors of acyl-CoA dehydrogenases (K i ϭ 80 M) (25,26). 3-Ketoacyl-CoAs also inhibit 3-hydroxyacyl-CoA dehydrogenase (27). Effective product inhibition has been observed with other enzymes of ␤-oxidation. For example, acyl-CoA dehydrogenases are strongly inhibited by 2-trans-enoyl-CoAs (25,26) and enoyl-CoA hydratase (crotonase) is inhibited by L-3-hydroxyhexadecanoyl-CoA with a K i of 0.35 M (19). Altogether, the available evidence suggests that the accumulation of intermediates would strongly inhibit the flux through the ␤-oxidation spiral. Such inhibition may be avoided by the absence of intermediates due to channeling.
Although it has been argued that channeling of intermediates provides a kinetic advantage to the ␤-oxidation system (28), the importance of this property is less certain as many metabolic pathways do not seem to be designed for optimal kinetic or energetic efficiency. However, the control of ␤-oxidation, especially of the energy-dependent regulation in extrahepatic tissues, may be greatly affected by intermediate channeling. In a pathway with intermediate channeling, the regulation of any reaction is expected to affect the whole system without any or significant changes in the concentration of intermediates. For example, the proposed regulation of 3-ketoacyl-CoA thiolase by the [acetyl-CoA]/[CoASH] ratio (23) could directly affect the activity of the first reaction of ␤-oxidation catalyzed by acyl-CoA dehydrogenase without 3-ketoacyl-CoA intermediates accumulating in the matrix and inhibiting acyl-CoA dehydrogenases (25).