Characterization of an Acyl-Coa Thioesterase that Functions as a Major Regulator of Peroxisomal Lipid Metabolism

Peroxisomes function in β-oxidation of very longand long-chain fatty acids, dicarboxylic fatty acids, bile acid intermediates, prostaglandins, leukotrienes, thromboxanes, pristanic acid and xenobiotic carboxylic acids. These lipids are mainly chain-shortened for excretion as the carboxylic acids or transported to mitochondria for further metabolism. Several of these carboxylic acids are slowly oxidized and may therefore sequester coenzyme A (CoASH). To prevent CoASH sequestration and to facilitate excretion of chain-shortened carboxylic acids, acyl-CoA thioesterases, which catalyze the hydrolysis of acyl-CoAs to the free acid and CoASH, may play important roles. We have here cloned and characterized a peroxisomal acyl-CoA thioesterase from mouse, named PTE-2, which was first isolated as a HIV-1 Nef binding protein in human (Liu et al. J. Biol. Chem. (1997) 272, 13779-13785, Watanabe et al. Biochem. Biophys. Res. Comm. (1997) 238, 234-239). PTE-2 is ubiquitously expressed and induced at mRNA level by treatment with the peroxisome proliferator WY-14,643 and fasting. Induction seen by these treatments was dependent on the peroxisome proliferator-activated receptor alpha (PPARα). Recombinant PTE-2 showed a broad chain-length specificity with acyl-CoAs from shortand medium-, to long-chain acylCoAs, and other substrates including trihydroxycoprostanoyl-CoA, hydroxymethylglutaryl-CoA and branched chain acyl-CoAs, all of which are present in peroxisomes. Highest activities were found with the CoA esters of primary bile acids choloyl-CoA and chenodeoxycholoyl-CoA as substrates. PTE-2 activity is inhibited by free CoASH, suggesting that intraperoxisomal free CoASH levels regulate the activity of this enzyme. The acyl-CoA specificity of recombinant PTE-2 closely resembles that of purified mouse liver peroxisomes, suggesting that PTE-2 is the major acyl-CoA thioesterase in peroxisomes. Addition of recombinant PTE-2 to incubations containing isolated mouse liver peroxisomes strongly inhibited bile


INTRODUCTION
Peroxisomes are cellular organelles present in all eukaryotic cells. They play an indispensable role in the metabolism of a variety of lipids including very long-chain fatty acids, dicarboxylic fatty acids, bile acids, prostaglandins, leukotrienes, thromboxanes, pristanic acid and xenobiotic fatty acids (for review, see (1,2)). The peroxisomal β-oxidation system contains two sets of enzymes, one of which is involved in the oxidation of branched chain fatty acids and intermediates in the hepatic bile-acid biosynthetic pathway, and consists of one or two branched-chain acyl-CoA oxidase(s), a D-specific bifunctional protein and the sterol carrier-like protein x (SCPx). The second pathway is involved in the oxidation of very long straight-chain fatty acids, CoA esters of prostaglandins, leukotrienes and thromboxanes (prostanoids) and dicarboxylic acids. This system is composed of a straight-chain acyl-CoA oxidase (ACO), an L-specific bifunctional protein and straight-chain 3-ketoacyl-CoA thiolase. However, these two pathways are not mutually exclusive (1). All enzymes in this latter pathway are induced by peroxisome proliferators, whereas the enzymes in the former β-oxidation pathway are not. Peroxisome proliferators are a group of structurally diverse compounds including hypolipidemic drugs, which induce peroxisome proliferation, hepatomegaly and cause hepatocarcinogenesis in rodent liver. These peroxisome proliferators, together with free fatty acids act as ligands for the peroxisome proliferator-activated receptor alpha (PPARα) (3,4). The PPARα is a nuclear receptor that plays a central role in fatty acid metabolism and induces the expression of many enzymes involved in peroxisomal and mitochondrial β-oxidation and ω-oxidation of fatty acids. Targeted disruption of the PPARα gene in mouse has established a key role for this receptor as a mediator of lipid metabolism (5)(6)(7)(8) and in the adaptive response to fasting (7)(8)(9)(10).
Prior to transport into peroxisomes and β-oxidation, all substrates must be activated to their CoA ester, which occurs at different cellular locations. Long-chain acyl-CoA synthetases are present in different subcellular membranes (11,12) and very longchain acyl-CoA synthetase in the peroxisomal membrane activates very long-and long-chain fatty acids and possibly branched-chain fatty acids to their CoA esters.
Dicarboxylic fatty acids, prostanoids and di-and trihydroxycoprostanic acid are activated to their CoA esters in the endoplasmic reticulum. These CoA esters are then transported into peroxisomes, possibly via ABC transporters. Peroxisomal βoxidation of very long-and long-chain fatty acyl-CoAs results in chain-shortening of these esters, and the chain-shortened products can be transported as carnitine esters to mitochondria for further degradation. However, peroxisomes appear to have important roles in β-oxidation of a number of xenobiotic carboxylic acids which may only be partially metabolized in peroxisomes (13,14). The β-oxidation of other CoA esters, such as prostanoids results in chain-shortening in peroxisomes, which are subsequently excreted in urine as the free carboxylic acid (for review see (15)).
Similarly the β-oxidation of dicarboxylic acids results in a chain-shortening in peroxisomes with the dicarboxylic acid being excreted in urine as the free acid or as glycine conjugates. Several of these carboxylic acids are only slowly β-oxidized, implying that intraperoxisomal CoASH may be sequestered to such an extent that it becomes limiting. In order to maintain appropriate CoASH levels and to facilitate excretion of chain-shortened carboxylic acids, acyl-CoA thioesterases are likely to play important roles (for review see (16)). The hydrolysis of CoA esters to the free acids requires the presence of an acyl-CoA thioesterase, an enzyme which functions to hydrolyze CoA esters to the free acid and CoASH. To date however, the specific thioesterase(s) active on CoA esters of prostanoids or dicarboxylic acids have not been identified.
Another important reaction that occurs in liver peroxisomes is the formation of bile acids. The primary bile acids, cholic acid and chenodeoxycholic acid, are formed by a number of enzymatic modifications of the cholesterol backbone by P-450 enzymes.
The di-or trihydroxycoprostanoyl-CoA (DHCA-CoA or THCA-CoA) formed undergoes a final β-oxidative cleavage of the side-chain in peroxisomes with the release of propionyl-CoA, to form chenodeoxycholoyl-CoA and choloyl-CoA, respectively (17,18). The peroxisomal bile acid-CoA:amino acid N -acyltransferase (BAAT) then catalyzes the conjugation of the CoA-activated bile acid to taurine or glycine prior to secretion from liver into bile. Recently, a multiorganellar bile acid-CoA thioesterase activity which hydrolyzes the bile acid-CoA esters to free bile acids and CoASH, has been characterized in human liver peroxisomes (19,20). This bile acid-CoA thioesterase activity described may compete with the BAAT enzyme for the bile acid-CoA substrate, thereby influencing intracellular levels of free and conjugated bile acids.
The existence of selective acyl-CoA thioesterases or a "broad range" acyl-CoA thioesterase could provide important control points in the oxidation of many peroxisomal substrates, and to regulate intracellular levels of CoA esters and CoASH. This may be especially important during times of high β-oxidation and fatty acid overload, to generate free CoASH necessary for fatty acid β-oxidation to proceed.
Acyl-CoA thioesterase activity has indeed been shown to be present in peroxisomes (21)(22)(23)(24) and isolated rat brown adipose tissue peroxisomes contain acyl-CoA thioesterase(s) active on short-and medium-chain acyl-CoAs, which are inhibited by CoASH (23). Rat liver peroxisomes showed broad acyl-CoA thioesterase activity on C 2 -C 22 acyl-CoAs and one enzyme was partially purified, which was shown to be most active on myristoyl-CoA (24). To date, several peroxisomal acyl-CoA thioesterases have been cloned from yeast, mouse and human. PTE-Ia and -Ib have been cloned from mouse, which belong to a novel family of Type-I acyl-CoA thioesterases, with related enzymes also in cytosol and mitochondria (25). Acyl-CoA thioesterases have been identified in yeast and human peroxisomes, named PTE1 (26). The human homologue of PTE1 was previously identified as hACTEIII/hTE, a protein that interacted with and activated the HIV-1 Nef protein (27,28). In the present study we have cloned the mouse homologue of PTE1/hACTEIII/hTE, which we have named mouse peroxisomal acyl-CoA thioesterase 2 (PTE-2). We have characterized this enzyme in detail, in order to examine its putative function in peroxisomal β-oxidation. Characterization of PTE-2 suggests that it is a major thioesterase with an array of functions in peroxisomal lipid metabolism.

Animals and treatments
Diurnal variation was investigated in adult male C57 BL/6 mice (B & K, Sollentuna, Sweden). The mice were maintained on a normal chow diet and sacrificed at the time-points indicated. The dark periods were between 18.00 and 06.00 hours. All other treatments were carried out on ten to twelve week old male wild-type or PPARα-null mice on a pure Sv/129 genetic background (derived from the original colony of mixed background mice) (5). These animals were housed in a temperature and light controlled environment. In fasting experiments, mice were maintained on a normal chow diet (R36 Lactamin, Vadstena, Sweden) prior to the start of the experiment and then transferred to new cages and fasted for 24 hours. Alternatively mice were treated with 0.1% WY- 14, in the diet for one week. All mice had access to water ad libitum. Animals were sacrificed by CO 2 asphyxiation followed by cervical dislocation, and weighed immediately. Tissues were then excised, weighed and frozen in liquid nitrogen. For subcellular fractionation experiments, livers from untreated wild-type mice were homogenized directly after sacrifice.

Northern blot analysis
Total RNA was isolated from mouse tissue samples using QuickPrep R Total RNA Extraction Kit (Amersham Pharmacia Biotech Sverige, Uppsala, Sweden) and Northern blot analysis was carried out as described (7). Blots were probed with the full-length cDNA for mouse PTE-2 or a probe for β-actin and were exposed to X-ray film at -70 o C.

Identification of human PTE-2 gene
Using human hACTEIII/hTE/hPTE1cDNA sequence (26)(27)(28) as search templates, a PAC clone containing approximately 86 kb of human genomic sequence was found to contain the gene encoding PTE-2. This human genomic sequence (Accession No. AL008726) was downloaded from NCBI ( http://www.ncbi.nlm.nih.gov) . The gene structure was compiled using Lasergene Software Package and 5' splice donor sites and 3' splice acceptor sites conformed to the general consensus sequences.

cDNA cloning and expression of PTE-2 in Escherichia coli
The sequence for hACTEIII/hTE/hPTE1 (26)(27)(28) was used to search the mouse EST database and several hits were obtained. The full-length cDNA sequence was compiled from overlapping EST-sequences. The mPTE-2 cDNA was amplified using the following primers: 5'-CATATGTCAGCGCCAGAGGGTCTG-3' and 5'- Interface software.

Determination of acyl-CoA thioesterase activity
Acyl-CoA thioesterase activity was measured spectrophotometrically at 412 nm with 5,5'-dithiobis(2-nitrobenzoic acid) (DTNB). The medium contained 200 mM potassium chloride, 10 mM Hepes and 0.05 mM DTNB (pH 7.4). An E 412 =13,600 m -1 cm -1 was used to calculate the activity. Since PTE-2 thioesterase activity was inhibited at substrate concentrations higher than 5-10 µM with acyl-CoAs longer than C 10 , bovine serum albumin (BSA) was added to a molar ratio of BSA/acyl-CoA of 4.5:1. The effect of CoASH, DTNB and p-chloromercuribenzoic acid (pCMB) on enzyme activity was measured at 232 nm in phosphate buffered saline. The enzyme kinetics were calculated using Sigma Plot enzyme kinetics programme.

Preparation and characterization of mouse liver subcellular fractions
Livers from wild-type and PPARα-null mice were homogenized as previously described (30) and bile acid-CoA thioesterase activity was measured as in (19). Livers from untreated wild-type mice were fractionated as described (19). The peroxisomeenriched fraction was further fractionated using a 15-45% Optiprep gradient to obtain highly purified peroxisomes. Protein concentrations were determined according to Bradford (31).

Bile acid-CoA:amino acid N -acyltransferase assay (BAAT)
The propanol and methanol and analyzed by HPLC using a Beckman ODS 5µ (4.6 mm x 25 cm) column with 20% 30 mM trifluoroacetic acid (adjusted to pH 2.9 with triethylamine) in methanol as mobile phase. The eluents were fractionated and assayed for 14 C-radioactivity by liquid scintillation counting.

PTE-2 cloning and sequence analysis
The cDNA isolated for PTE-2 encodes a protein of 320 amino acids, with a calculated

PTE-2 is localized in peroxisomes
The mouse PTE-2 contains a well-characterized consensus peroxisomal Type 1 targeting signal (PTS1) of -SKL (serine, lysine, leucine) at its C-terminal end, which has been shown to target proteins to peroxisomes (34). Previously, human PTE-2 has been shown to be localized in peroxisomes (26,35) and to test if mouse PTE-2 is peroxisomal, we cloned PTE-2 in-frame with GFP, which leaves the C-terminal -SKL sequence accessible. We transfected this vector into both control fibroblasts and fibroblasts from a Zellweger patient, which are unable to import peroxisomal matrix proteins. Using immunofluorescence microscopy for GFP detection, mPTE-2 showed a punctate pattern of expression in control fibroblasts, indicative of a peroxisomal localization (Fig. 3A). This localization was confirmed with a Tritc-labeled fluorescent secondary antibody to GFP (Fig. 3B). However, in Zellweger fibroblasts, transfection of mPTE-2 resulted in a diffuse GFP expression (Fig. 3D). The transfection of the Zellweger cells was confirmed using a Tritc-labeled fluorescent secondary antibody to GFP, which identified the transfected cells (Fig. 3E). Phase contrast microscopy is also shown in both cases ( Fig. 3C & 3F) as a control, indicating other untransfected cells. The strong staining present in the nucleus is Hoescht staining.

Recombinant expression and characterization of PTE-2
The cloning of the cDNA encoding PTE-2 into the NdeI site of the pET16B vector results in expression of the PTE-2 as a His-tagged fusion protein, to allow for purification using affinity chromatography. Following purification of PTE-2 on a HiTrap™ column, the purified protein was detected as a single band of approximately 36 kDa in mass on SDS-PAGE gel stained with Coomassie brilliant blue (Fig. 4A). An antibody towards the human PTE1 (a kind gift from Jacob Jones) cross-reacted with the mouse PTE-2, confirming the correct protein (data not shown).
The expressed protein was further analyzed by size-exclusion chromatography as described previously (36) 4D). PTE-2 activity was also inhibited by DTNB (IC 50 ≈150 µM) and pCMB (IC 50 ≈ 1 µM), two cysteine-reactive agents (data not shown), however to date a cysteine involved in the active site catalysis has not been identified.
Calculation of the V max and K m values showed that the K m for medium-to long-chain acyl-CoAs was in the order of 1.4-6.7 µM, with short chain acyl-CoAs ranging from 8-30 µM (Table I) The acyl-CoA thioesterase chain-length specificity was measured in isolated mouse peroxisomes (Fig. 5A). Activity was seen with straight-chain acyl-CoAs from chain length of C 2 up to C 20 , together with very high activity towards bile-acid CoA esters, THCA-CoA and branched-chain-CoA esters. There was a striking similarity between the acyl-CoA chain-length pattern for the recombinant PTE-2 and the chain length specificity in isolated peroxisomes (Fig. 5B). Superimposing the curves for the substrate specificities of isolated peroxisomes and that of PTE-2 showed that PTE-2 apparently catalyzes most of the activities seen in peroxisomes, with the only observable difference being the putative presence of an additional acyl-CoA thioesterase in peroxisomes, mainly catalyzing the hydrolysis of C 12 -C 14 -CoA.
Comparison of the specific activities of recombinant PTE-2 and isolated peroxisomes indicates that PTE-2 constitutes about 1% of total peroxisomal protein.

PTE-2 competes with bile acid-CoA:amino acid N -acyltransferase for bile acid-CoA
We examined the ability of recombinant PTE-2 to compete with BAAT for the bile acid-CoA substrate chenodeoxycholoyl-CoA. BAAT activity was measured in highly purified peroxisomes isolated from control mouse liver. Addition of recombinant PTE-2 caused an approximately 80% inhibition of BAAT activity (Table II). This data shows that PTE-2 can compete with BAAT for the bile acid-CoA substrate in vitro.
We have also tested the ability of recombinant mouse cytosolic acyl-CoA thioesterase I (mCTE-I) (37) to hydrolyze CA-CoA or CDCA-CoA. However, this enzyme showed no detectable activity with CA-CoA or CDCA-CoA, further emphasizing that the bile-acid CoA thioesterase activity of PTE-2 is specific (data not shown).

Choloyl-CoA thioesterase activity is induced in mouse liver in a PPAR -dependent manner
Since PTE-2 very efficiently hydrolyzes CoA esters of bile acids, and PPARα has been shown to be involved in bile acid biosynthesis (38), we investigated the possible regulation of bile acid-CoA thioesterase activity by feeding mice WY-14,643, a potent PPARα activator. In liver homogenates of wild-type mice, choloyl-CoA thioesterase activity was increased 3.5-fold following treatment with WY-14,643 (Fig. 6). However in PPARα-null mouse liver homogenates, this increase in choloyl-CoA thioesterase activity was not evident, showing that the WY-14,643 mediated induction of bile acid-CoA thioesterase activity in mouse liver was PPARα-dependent.

PTE-2 mRNA expression is PPAR -regulated
In view of the fact that PTE-2 can hydrolyze bile-acid CoA esters, together with the fact that choloyl-CoA thioesterase activity was shown to be induced in a PPARαdependent manner in mouse liver, we used the PPARα-null mouse model to examine regulation of the PTE-2 at mRNA level. The basal expression of PTE-2 in PPARα-null mice was approximately half that detected in wild-type animals.
Treatment of mice with a WY-14,643 containing diet for 1 week resulted in a very strong (> 10 fold) induction of PTE-2 mRNA (Fig. 7A, upper panel). However, this up-regulation was not evident in the PPARα-null mice also treated with WY-14,643, showing that the effect mediated on PTE-2 by peroxisome proliferators is dependent on the PPARα.
The role of the PPARα in the fasting-mediated induction of several genes has also been shown (7)(8)(9)(10)38). We examined the effect of fasting on PTE-2 mRNA levels in the PPARα-null mouse model, which resulted in a significant increase in PTE-2 mRNA in liver (Fig. 7A, lower panel). However, this induction was not seen in the PPARα-null mice that were similarly treated.

PTE-2 shows a diurnal regulation of expression
As PTE-2 is regulated by fasting and is therefore under nutritional regulation, we examined if this enzyme could also be under a diurnal regulation. Mice were fed a normal chow diet and were sacrificed every 4th hour commencing at 09.00h.
Quantitation of the mRNA signal showed that during the light period (between 06.00h and 18.00h), when animals are less active, the mRNA levels were increased, thus indicating an induction by fasting (Fig. 7B). During the dark period, when feeding mainly takes place (between 18.00h and 06.00h) the mRNA levels declined rapidly, indicating a rapid nutritional regulation in response to refeeding.

PTE-2 mRNA is ubiquitously expressed
The tissue expression of PTE-2 was examined using Northern blot analysis on several different tissues from mouse (Fig. 7C). PTE-2 was ubiquitously expressed as a 1.2 kb transcript in all tissues examined, which is similar to results obtained for expression in human tissues (27). Tissue expression was highest in kidney, liver and testis, with weaker expression evident in heart and muscle. In testis a second transcript was evident at approximately 2 kb, which may be due to splicing events.

DISCUSSION
In this study we have cloned and characterized the mouse peroxisomal acyl-CoA thioesterase 2, called PTE-2, which is localized in peroxisomes. This enzyme is the mouse homologue of human PTE1 (26) which was first identified by the yeast twohybrid system as hACTEIII and hTE, a HIV-1 Nef binding protein (27,28). There is a lot of confusion regarding the nomenclature of acyl-CoA thioesterases, and based on a recent attempt to accomplish a more uniform terminology (16), we propose the name PTE-2 for this acyl-CoA thioesterase, as the nomenclature PTE-I has already been assigned to two enzymes which are members of a novel multi-gene family, with other members in mitochondria and cytosol (25) (Table III). However, the rule applied to the nomenclature of yeast enzymes states that they remain known by the name applied when they were first identified, therefore this enzyme will remain known as PTE1. A PTE2 cloned as a putative peroxisomal enzyme (40) (Table III) shows 100% identity to MTE-I (mitochondrial acyl-CoA thioesterase) without its N- The HIV-1 Nef which interacts with the thioesterase PTE-2, is a cytosolic 27-kDa myristoylated protein that is required for high viral load and full pathological effect of HIV (41). It is thought that Nef activity triggers endocytosis of CD4 and major histocompatibility complex class-I molecules. Although the interaction of Nef with the acyl-CoA thioesterase is not fully understood, it has been shown that Nef contains sites critical for binding of the human thioesterase, which results in downregulation of CD4 (35). The binding of Nef to the thioesterase also targets Nef to peroxisomes in-vivo and increases thioesterase activity in-vitro (27). It was also shown that abolishment of the interaction of the thioesterase with Nef resulted in impairment of Nef biological functions (42). It then became evident that although

PTE-2 can have a multitude of functions in regulating peroxisomal lipid metabolism
In the present study we have cloned and characterized the mouse PTE-2 with respect to regulation of expression and kinetic activity.

PTE-2 is a PPAR -target gene that may regulate bile-acid formation
Bile acids are formed in liver peroxisomes by β-oxidative cleavage of the side-chain of DHCA-CoA or THCA-CoA to chenodeoxycholoyl-CoA and choloyl-CoA, respectively, with the concomitant production of propionyl-CoA. In the last step, bile acid-CoAs are conjugated to taurine or glycine, a reaction catalyzed by the BAAT enzyme that acts as an acyltransferase. In fact the best substrates for PTE-2 were found to be THCA-CoA, CA-CoA and CDCA-CoA. We have now established that PTE-2 is the bile acid-CoA thioesterase previously identified (19,20), suggesting that a major With the identification of the gene for the human PTE-2, a putative peroxisome proliferator-response element (PPRE) was identified at 438 bp upstream of the ATG start site. This site conforms well to the consensus DR1 (AGGTCAnAGGTCA) which has been shown to bind both PPAR/RXR heterodimers and hepatocyte nuclear factor 4α (HNF-4α). It will be of interest to identify if this site is functional in the human promoter in binding these transcription factors, thus leading to possible activation of human PTE-2 by either peroxisome proliferators or fatty acids (the ligands for PPARs (3,4)), or acyl-CoAs and CoA esters of peroxisome proliferators, (the agonists/antagonists for HNF-4α (53,54)).

Can PTE-2 be considered as an auxilliary -oxidation enzyme?
In CoA produced in peroxisomes from dicarboxylic and monocarboxylic acids and acetate production is increased by peroxisome proliferator treatment (61). The production of free acetate from acetyl-CoA requires the hydrolysis by an acyl-CoA thioesterase present in peroxisomes (23,24). Again, recombinant PTE-2 showed high activity towards acetyl-CoA, which constitutes a mechanism to prevent trapping of CoASH in the acetyl-CoA unit.
In summary, we have shown that PTE-2 acts as a general acyl-CoA thioesterase that is responsible for most of the thioesterase activity detected in isolated peroxisomes.    HMG, 3-hydroxy-3-methyl-glutaryl-CoA; AcAc, acetoacetyl-CoA. (D) Inhibition of PTE-2 activity by CoASH. Acyl-CoA thioesterase activity was measured at 232 nm as described in EXPERIMENTAL PROCEDURES. The activity was measured using octanoyl-CoA as substrate, to avoid the required addition of BSA, which will strongly interfere at 232 nm when measuring activity with longer acyl-CoAs. CoASH was added to the enzyme assay at concentrations indicated.   mice were maintained on a normal chow diet ad libitum and were sacrificed at the time-points indicated. The light period was between 06.00 and 18.00 and the dark period between 18.00 and 06.00. Northern blot analysis was carried out on 20 µg total RNA using α-32 P-labeled probes for PTE-2 and β-actin. Filters were exposed to X-ray film and signals were quantified using Image Master Software 3.0. The mean of PTE-2/actin mRNA ± range for two animals is shown. (C) Tissue expression of PTE-2 mRNA. Total RNA was isolated from various tissues of Sv129 male mice. Northern blot analysis was carried out on 20 µg total RNA using α-32 P-labeled cDNA probe for PTE-2 as described in EXPERIMENTAL PROCEDURES. Ethidium bromide (Et Br) staining of the gel shows even loading of samples.