Presence of three acyl-CoA oxidases in rat liver peroxisomes. An inducible fatty acyl-CoA oxidase, a noninducible fatty acyl-CoA oxidase, and a noninducible trihydroxycoprostanoyl-CoA oxidase.

Mammalian liver peroxisomes are capable of beta-oxidizing a variety of substrates including very long chain fatty acids and the side chains of the bile acid intermediates di- and trihydroxycoprostanic acid. The first enzyme of peroxisomal beta-oxidation is acyl-CoA oxidase. It remains unknown whether peroxisomes possess one or several acyl-CoA oxidases. Peroxisomal oxidases from rat liver were partially purified by (NH4)2SO4 precipitation and heat treatment, and the preparation was subjected to chromatofocusing, chromatography on hydroxylapatite and dye affinity matrices, and gel filtration. The column eluates were assayed for palmitoyl-CoA and trihydroxycoprostanoyl-CoA oxidase activities and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The results revealed the presence of three acyl-CoA oxidases: 1) a fatty acyl-CoA oxidase with a pI of 8.3 and an apparent molecular mass of 145 kDa. The enzyme consisted mainly of 52- and 22.5-kDa subunits and could be induced by clofibrate treatment; 2) a noninducible fatty acyl-CoA oxidase with a pI of 7.1 and an apparent molecular mass of 427 kDa. It consisted mainly, if not exclusively, of one polypeptide component of 71 kDa; and 3) a noninducile trihydroxycoprostanoyl-CoA oxidase with a pI of 7.1 and an apparent molecular mass of 139 kDa. It consisted mainly, if not exclusively, of one polypeptide component of 69 kDa. Our findings are probably related to the recent discovery of two species of acyl-CoA oxidase mRNA in rat liver (Miyazawa, S., Hayashi, H., Hijikata, M., Ishii, N., Furata, S., Kagamiyama, H., Osumi, T., and Hashimoto, T. (1987) J. Biol. Chem. 262, 8131-8137) and they probably also explain why in human peroxisomal beta-oxidation defects an accumulation of very long chain fatty acids is not always accompanied by an excretion of bile acid intermediates and vice versa.


Mammalian
liver peroxisomes are capable of @-ox& dizing a variety of substrates including very long chain fatty acids and the side chains of the bile acid intermediates di-and trihydroxycoprostanic acid. The first enzyme of peroxisomal @-oxidation is acyl-CoA oxidase. It remains unknown whether peroxisomes possess one or several acyl-CoA oxidases. Peroxisomal oxidases from rat liver were partially purified by (NH&SO., precipitation and heat treatment, and the preparation was subjected to chromatofocusing, chromatography on hydroxylapatite and dye affinity matrices, and gel filtration. The column eluates were assayed for palmitoyl-CoA and trihydroxycoprostanoyl-CoA oxidase activities and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The results revealed the presence of three acyl-CoA oxidases: 1) a fatty acyl-CoA oxidase with a p1 of 8. 3  ,&oxidation sequence is acyl-CoA oxidase. As the other enzymes of hepatic peroxiso-ma1 P-oxidation, the liver enzyme is induced lo-20-fold after treatment of rats with a group of structurally diverse compounds, called peroxisome proliferators (2). Acyl-CoA oxidase has been purified from livers of rats treated with peroxisome proliferators (3,4). The purified induced enzyme consists of three polypeptide components, A, B, and C with molecular masses of 72, 52, and 21 kDa, respectively. The molar ratio of the components was 1:5:5. All available evidence indicates that components B and C are formed in viuo from component A by post-translational proteolytic cleavage and that the enzyme exists as a mixture of AZ, B2C2, and ABC (5,6). The cDNA sequence and the deduced amino acid sequence of component A are known and the site at which A is cleaved in B and C has been identified (7,8). In addition, it has been found that rat liver contains two species of acyl-CoA oxidase mRNA, which have the same number of nucleotides and which differ in their nucleotide sequence only in a small region. They are produced by alternative splicing of the transcript of a single gene (7,8). One of the two mRNA species is more abundant than the other and its nucleotide sequence corresponds to the amino acid sequence of the purified induced enzyme. It is not known whether the less abundant species is also translated to a protein.
In rats treated with peroxisome proliferators, hepatic acyl-CoA oxidase activity is increased when the CoA esters of mono-and dicarboxylic fatty acids and of prostaglandins are used as the substrate but not when the CoA ester of trihydroxycoprostanic acid is used (9,10). This indicates that THC-CoA' oxidase is not inducible. The presence in liver of two species of acyl-CoA oxidase mRNA and of a noninducible THC-CoA oxidase makes several enzyme combinations possible: 1) presence of an inducible fatty acyl-CoA oxidase and a noninducible THC-CoA oxidase, which, in turn, may or may not have fatty acyl-CoA oxidase activity; 2) presence of two separate, inducible fatty acyl-CoA oxidases and a noninducible THC-CoA oxidase; 3) presence of an inducible fatty acyl-CoA oxidase, a noninducible fatty acyl-CoA oxidase, and a separate, noninducible THC-COA oxidase. The experiments described below indicate that the latter is the case.

Animals-Male
Wistar rats weighing 120-150 g were maintained on a standard laboratory diet for 2 weeks. Clofibrate-treated rats were kept on the same diet containing 0.3% (v/w) clofibrate. Partial Purification of Peroxisomal Oxidases-Rat liver and kidney were homogenized and the homogenates were fractionated by differential centrifugation as described previously (11) in order to prepare an "L"-fraction, enriched in lysosomes and peroxisomes. The Lfraction was suspended in 10 mM PPi buffer, pH 9 and coworkers (7) reported the deduced amino acid sequence of the translation products of the two acyl-CoA oxidase mRNA species present in rat liver. We calculated the p1 for each of the polypeptides and found 8.6 for the first sequence, which corresponds to the purified induced enzyme, an 8.1 for the second sequence. Since the difference seemed reasonably large, we decided to chromatofocus a partially purified preparation of peroxisomal oxidases. When a preparation from a normal rat liver was used, we observed two separate peaks of palmitoyl-CoA oxidase activity with a p1 of 8.31 and 7.13, respectively (Fig. 1A).' THC-CoA oxidase displayed a single peak that coincided with the palmitoyl-CoA oxidase peak at pH 7.1. When a preparation from a clofibrate-treated rat liver was focused, palmitoyl-CoA oxidase was no longer present in two peaks but a large single peak with a p1 of 8.35 appeared (Fig. 1B). The activity at pH 8.3 was much larger in the liver from treated than from control rats, demonstrating that this activity represents the inducible enzyme. Except for a small shoulder at pH 7.1-7.2, there was little indication of a second palmitoyl-CoA oxidase in the liver from clofibrate-treated rats. It is probable, however, that the second enzyme was masked by the trailing edge of the large peak of inducible activity. In any case, the data show that the palmitoyl-CoA oxidase found at pH 7.1 was not induced by clofibrate treatment. Equally importantly, the fact that the enzyme with the lower p1 was certainly not more abundant in the liver from treated rats, precludes the possibility that it was artifactually formed in vitro from the enzyme with the more alkaline p1 and most probably also explains why the enzyme with the lower p1 has been overlooked during earlier purification of the induced enzyme (3,4).
THC-CoA oxidase was found in the treated liver in the same pH region as in the normal liver. Its activity was not increased after clofibrate treatment, confirming that the enzyme is not induced (10).
No THC-CoA oxidase was found in kidney ( Fig. 1, C and  D). The palmitoyl-CoA oxidase with the more alkaline p1 was by far the major form of the enzyme in kidney from treated 'For the sake of simplicity, the pH of the fraction in which a protein is eluted from the chromatofocusing column, is considered here as the p1 of the protein. This is not entirely correct. Although a protein is eluted from the column when the pH of the surrounding buffer has reached the p1 value of the protein, the buffer travels faster through the column than the protein so that the pH of the fraction in which the protein is recovered, is somewhat lower than the true p1 value. Osumi et al. (3) found a p1 of 9.2 for the inducible (see further) fatty acyl-CoA oxidase after isoelectric focusing. as well as from untreated animals. It is not clear whether the small shoulder at pH 7.1-7.3 (Fig. 1C) is an indication for the presence in kidney of a small amount of the second palmitoyl-CoA oxidase.
The chromatofocusing experiments with the liver enzymes (Fig. 1, A and B), did not reveal whether THC-CoA oxidase and palmitoyl-Cob oxidase with the lower p1 are activities of a single enzyme or of two separate enzymes.
As illustrated in Fig. 2, A-D, the activities behaved completely differently and sometimes even oppositely in a number of experimental conditions that were tested in function of the eventual purification of the enzymes. These differences in behavior clearly demonstrate that the low p1 fatty acyl-CoA oxidase and THC-CoA oxidase are separate enzymes. This conclusion was supported by the elution pattern of the two activities, when pooled fractions containing the low p1 palmitoyl-CoA oxidase and THC-CoA oxidase from a chromatofocusing column, were applied to the dye affinity matrices green A and orange A. The pooled and concentrated fraction (0.5 ml) that was applied contained 10 milliunits of palmitoyl-CoA oxidase, 1.7 milliunits of THC-CoA oxidase and 0.2 mg of protein/ml. The columns were washed with 5 ml of PPi buffer and eluted with 6 ml of PPi buffer containing 1.5 M KCl. The percentages of applied palmitoyl-CoA oxidase, THC-CoA oxidase, and protein that were washed from the green A column were 0, 55, and 94%, respectively and those washed from the orange A column 37, 95, and lOO%, respectively. Those eluted from the green A column were 14,20, and lo%, respectively. No enzyme activities or proteins were detected in the eluate of the orange A column. Although the recoveries of palmitoyl-CoA oxidase were extremely low due to a possible adsorption to the columns, inactivation, or perhaps interference of KC1 with the assay (see the legend to Fig. 2), the differences again confirm that the oxidations of palmitoyl-CoA and of THC-CoA are not catalyzed by the same enzyme.
Three Acyl-CoA Oxidases Finally, a complete separation of THC-CoA oxidase from palmitoyl-CoA oxidase was obtained when pooled fractions containing THC-CoA oxidase and the low p1 fatty acyl-CoA oxidase from a chromatofocusing column, were applied to a hydroxylapatite column, which was eluted by means of a discontinuous phosphate gradient (Fig. 3). THC-CoA oxidase eluted at the lower phosphate concentrations, whereas palmitoyl-CoA oxidase eluted at the higher concentrations. Pooled fractions containing the inducible fatty acyl-CoA oxidase and those containing the noninducible fatty acyl-CoA oxidase and THC-CoA oxidase from a chromatofocusing column were also subjected to gel filtration.
The fractions eluted from the various columns were also analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Fig. 4 shows the results for the chromatofocusing experiment represented in Fig. 1A. In the control liver the inducible fatty acyl-CoA oxidase, which eluted around pH 8.3, consisted mainly of subunits with molecular masses of 52 and 22.5 kDa. A subunit with a molecular mass of 71 kDa was also present, but less abundant.
A similar pattern was observed for the induced enzyme after clofibrate treatment (data not shown), confirming earlier reports by other laboratories (3,4). The fractions, containing THC-CoA oxidase and the low p1 fatty acyl-CoA oxidase showed the presence of a polypeptide of approximately 70 kDa (Fig. 4). Only small amounts of the 52-and 22.5-kDa subunits were found, suggesting that a 70-kDa polypeptide is the main component of THC-CoA oxidase as well as of the low p1 fatty acyl-CoA oxidase. In kidney, where no THC-CoA oxidase was present and where little or no fatty acyl-CoA oxidase with low p1 was found (see Fractions eluting from a chromatofocusing column between pH 7.2 and 6.9 were pooled, dialyzed, and concentrated. A 0.5-ml aliquot, containing 17. 8   The fractions from the chromatofocusing column (same experiment as described in Fig. lA dantly the 71-kDa polypeptide as was the case for liver (data not shown). The fractions eluting from the green A or orange A columns (see above) contained a polypeptide with a molecular mass of approximately 70 kDa, whenever THC-CoA oxidase and/or palmitoyl-CoA oxidase activity was found (data not shown). Electrophoretic analysis of the fractions from the hydroxylapatite column confirmed that the main polypeptide component of the low p1 fatty acyl-CoA oxidase as well as of THC-CoA oxidase is a polypeptide of approximately 70 kDa (Fig. 5)." The molecular mass of the acyl-CoA oxidase component appears to be slightly larger than that of the THC-CoA oxidase component, 70.8 f 0.75 and 69.4 + 0.75 kDa, respectively (mean f S.D. of four experiments). Our results confirm earlier reports about the apparent molecular mass and subunit composition of the inducible fatty acyl-CoA oxidase (3,4). In addition, they indicate that the noninducible fatty acyl-CoA oxidase and THC-CoA oxidase are composed of six and two identical subunits, respectively. DISCUSSION Our experiments establish that rat liver contains three acyl-CoA oxidases: an inducible fatty acyl-CoA oxidase, which has already been purified (3,4), a noninducible fatty acyl-CoA oxidase, and a noninducible THC-CoA oxidase. The separation of the three enzymes, as described above, will allow for the determination of the exact substrate specificity of each enzyme.
It Same experiment as shown in Fig. 3. For experimental conditions, see legend to dases are the translation products of the two mRNA species recently discovered in rat liver (7,8). If this is the case, the interesting situation ensues that during induction the rate of transcription of the single acyl-CoA oxidase gene is increased but that, most probably, the concentration of only one of the two alternatively spliced mRNAs is increased. The fact that the mRNA species for the inducible enzyme is more abundant than the other species in livers from animals treated with peroxisome proliferators (7,8) would support the above contention. THC-CoA oxidase seems to be encoded by a separate gene. This conclusion is based on the smaller molecular mass of the polypeptide component and the sensitivity of the enzyme to N-ethylmaleimide.
No cysteine residue is present in the region where the two acyl-CoA oxidase sequences differ (7).
Experiments are currently underway in our laboratory to verify whether the two fatty acyl-CoA oxidases correspond to the two mRNA species described by the laboratory of Osumi and Hashimoto (7,8) and to determine which mRNA species increases its concentration on induction. Given that the two fatty acyl-CoA oxidases would correspond to the two mRNAs, an additional interesting question would emerge: why is the inducible enzyme readily cleaved and the noninducible one only slightly or not, despite the fact that both A polypeptides would have the same amino acid sequence at and in the vicinity of the cleavage site? Perhaps the limited difference in amino acid sequence at distance would render the cleavage site inaccessible. This limited difference in amino acid sequence and/or the fact that the 71-kDa subunit is not split, would then also seriously influence the quaternary structure of the enzyme (6 x 71-kDa subunits instead of two (cleaved) 71-kDa subunits).
Our observations are also of interest with respect to human pathology.
Children born with a deficiency of peroxisomal/3oxidation accumulate very long chain fatty acids in their tissues and excrete abnormal bile acids (di-and trihydroxycoprostanic acids instead of chenodeoxycholic and cholic acids), because peroxisomes are responsible for the P-oxidation of very long chain fatty acids and the carboxy side chain of bile acid intermediates (for a review, see Ref. 1). Recently, two cases of fatty acyl-CoA oxidase deficiency have been reported, in which there was an accumulation of very long chain fatty acids but no abnormal bile acids (16). This suggests that, as in rat liver, a separate THC-CoA oxidase is present in human liver. In addition, a patient is known who excretes abnormal bile acids, but who does not display an accumulation of very long chain fatty acids.4 This case suggests that a separate deficiency of THC-CoA oxidase can also occur. Whether the human liver possesses also two fatty acyl-CoA oxidases and what the consequences of a deficiency of one of the enzymes would be, remain to be studied.