The Human Bile Acid-CoA:Amino Acid N-Acyltransferase Functions in the Conjugation of Fatty Acids to Glycine*

Bile acid-CoA:amino acid N-acyltransferase (BACAT) catalyzes the conjugation of bile acids to glycine and taurine for excretion into bile. By use of site-directed mutagenesis and sequence comparisons, we have identified Cys-235, Asp-328, and His-362 as constituting a catalytic triad in human BACAT (hBACAT) and identifying BACAT as a member of the type I acyl-CoA thioesterase gene family. We therefore hypothesized that hBACAT may also hydrolyze fatty acyl-CoAs and/or conjugate fatty acids to glycine. We show here that recombinant hBACAT also can hydrolyze long- and very long-chain saturated acyl-CoAs (mainly C16:0–C26:0) and by mass spectrometry verified that hBACAT also conjugates fatty acids to glycine. Tissue expression studies showed strong expression of BACAT in liver, gallbladder, and the proximal and distal intestine. However, BACAT is also expressed in a variety of tissues unrelated to bile acid formation and transport, suggesting important functions also in the regulation of intracellular levels of very long-chain fatty acids. Green fluorescent protein localization experiments in human skin fibroblasts showed that the hBACAT enzyme is mainly cytosolic. Therefore, the cytosolic BACAT enzyme may play important roles in protection against toxicity by accumulation of unconjugated bile acids and non-esterified very long-chain fatty acids.

Bile acid (BA) 1 formation is the major pathway in mammals for the excretion of cholesterol (for review, see Ref. 1). BAs are synthesized from cholesterol in the liver and are conjugated to either glycine or taurine before secretion into the bile. This conjugation (or amidation) plays several important biological roles in that it promotes the secretion of BAs and cholesterol into bile and increases the detergent properties of BAs in the intestine, which facilitates lipid and vitamin absorption. BAs are deconjugated by bacteria in the intestine and recycled back to the liver for reconjugation. The initial steps in the biosynthesis of BAs involve oxidative modifications of the cholesterol backbone and side chain to form dihydroxycholestanoic acid and trihydroxycholestanoic acid (THCA). Dihydroxycholestanoic acid and THCA are activated to their corresponding CoA-esters followed by ␤-oxidative cleavage in peroxisomes to form chenodeoxycholoyl-CoA (CDCA-CoA) and choloyl-CoA (CA-CoA), respectively, which are substrates for conjugation to glycine or taurine (2)(3)(4)(5). This conjugation is catalyzed by the enzyme bile acid-CoA:amino acid N-acyltransferase (BACAT), an enzyme recently implicated in the inheritance of familial hypercholanemia (6). Recent data show that BACAT activity is present both in peroxisomes and in the cytosol (5), suggesting the existence of two BACAT enzymes. It has been proposed that whereas the peroxisomal enzyme conjugates de novo synthesized BAs, the cytosolic enzyme has a function in conjugating BAs recycled from the intestine to the liver. Further support for the existence of two pathways for conjugation of BAs comes from a recent study by Mihalik et al. (7). They show that human very long-chain acyl-CoA synthetase (VLCS), present in peroxisomes and the endoplasmic reticulum, primarily activates THCA, which is a precursor for de novo synthesis of bile acids. In contrast, a homologue of this enzyme, VLCS-H2, located in the endoplasmic reticulum and referred to as bile acid-CoA synthetase (BACS), activates mainly BAs (7). It was therefore suggested that BACS activates recycled BAs for conjugation by BACAT in the cytosol. However, to date only one BACAT enzyme has been identified and characterized. The enzyme has been purified from several species such as rat (8), bovine (9), domestic fowl (10), fish (11), and human (12), and partially from pig (13), canine (14), guinea pig, and rabbit (15). Molecular cloning of the human BACAT showed that the enzyme conjugates bile acids to both glycine and taurine (16).
We have recently identified and characterized a family of highly homologous acyl-CoA thioesterases, referred to as type I acyl-CoA thioesterases, with putative localizations in peroxisomes (PTE-Ia and PTE-Ib), mitochondria (MTE-I), and cytosol (CTE-I) (17). Data base searches and subsequent sequence alignments revealed that these acyl-CoA thioesterases show sequence homology only to BACAT from rat (18), mouse (19), and human (16), with a sequence identity of 40 -45% to the type I acyl-CoA thioesterases. Acyl-CoA thioesterases hydrolyze acyl-CoAs to non-esterified fatty acids and CoASH. By preserving a balance of acyl-CoA, free fatty acids, and CoASH in the cell, acyl-CoA thioesterases directly and indirectly, via gene regulation, influence numerous cellular processes involved in lipid metabolism, for example ␤-oxidation and esterification of fatty acids (for review, see Ref. 20). All members of this acyl-CoA thioesterase family contain a conserved serine-histidine-aspartic acid catalytic triad (21). Sequence alignments showed that BACAT also contains an almost identical catalytic triad, with one notable difference, the presence of a cysteine in place of the acyl-CoA thioesterase nucleophilic serine. Therefore it is likely that the cysteine is the nucleophilic residue in the BACAT enzyme and, possibly, that BACAT also amidates fatty acids.
In this study we have characterized the human BACAT enzyme, showing that in addition to catalyzing the conjugation of BAs, BACAT also conjugates fatty acids to glycine in vitro and that it can also act as a very long-chain acyl-CoA thioesterase. From transfection experiments of BACAT as a fusion protein with green fluorescent protein (GFP) we also show that BACAT is mainly cytosolic. As BACAT also shows a wider tissue distribution than anticipated previously, we propose that BACAT may also have functions other than conjugating bile acids to glycine.

EXPERIMENTAL PROCEDURES
Cloning and Expression of Human BACAT cDNA-The hBACAT cDNA encoding the entire open reading frame was amplified using the following primers: 5Ј-CATATGATCCAGTTGACAGCT-3Ј and 5Ј-CATATGGAGACATTCCGCCATG-3Ј with the addition of NdeI sites (indicated in bold). The full-length cDNA was amplified with One Step RNA PCR kit (AMV, Takara Biomedicals) using a template of human liver total RNA. RT-PCR was performed at 58°C for 30 min followed by 35 cycles of 94°C for 1 min, 52°C for 30 s, and 72°C for 4 min. The resultant PCR product was cloned into the pET16b vector (Novagen Inc.) and expressed as described elsewhere (21), except that the bacteria were induced with isopropyl-1-thio-␤-D-galactopyranoside at 30°C for 5 h.
Purification of Wild-type and Mutant hBACAT Protein-For initial studies on hBACAT, the bacterial pellets were thawed in phosphate buffer (20 mM potassium phosphate, 0.5 M NaCl, pH 7.4) containing 50 mM imidazole and solubilized by sonication with 10 ϫ 5-s pulses at 5-s intervals (XL2020 from Heat Systems). The bacterial suspension was centrifuged for 60 min at 35,000 ϫ g at 4°C. The supernatant was filtered through a 0.22-m filter, and His-tagged recombinant proteins were purified using HiTrap™ chelating columns (Amersham Biosciences). The recombinant protein was eluted stepwise with increasing concentrations of imidazole. For some experiments, the bacterial pellet was resuspended in Bugbuster protein extraction reagent (Novagen) according to the manufacturer's instructions. The supernatant was filtered as described above and the protein purified using HiTrap™ chelating columns. Because of precipitation of the protein when using 0.5 M NaCl, the columns were equilibrated with phosphate buffer containing 20 mM potassium phosphate, 0.1 M NaCl, and 10 mM imidazole, pH 7.4. The supernatant was mixed with the same buffer in a 1:1 ratio and loaded onto the column, and the protein was eluted stepwise with increasing imidazole concentrations.
Gel Electrophoresis and Protein Determination-Purified recombinant proteins or bacterial protein extracts were separated by SDS-PAGE on 10% polyacrylamide gels and stained with Coomassie Brilliant Blue. Protein concentration was determined according to Bradford (22).
Determination of hBACAT Activity-Bile acid-CoA thioesterase and conjugation activity of hBACAT was measured spectrophotometrically at 412 nm using 5,5Ј-dithiobis-(2-nitrobenzoic acid) (DTNB). The medium contained 200 mM potassium chloride, 10 mM HEPES, and 0.05 mM DTNB (pH 7.4). For conjugation measurements, 50 mM glycine or taurine was added to the cuvette. Medium and substrate were premixed, and the reaction was started with the addition of protein, which allowed for the testing of non-enzymatic hydrolysis of the CoA esters. An ⑀ 412 nm ϭ 13,600 M Ϫ1 cm Ϫ1 was used to calculate the activity. Kinetic analysis was carried out using the Sigma Plot Enzyme Kinetics program. Activity measurements on CA-CoA and N-acylglycines were carried out in duplicate on three different hBACAT protein preparations, and measurements for CDCA-CoA and THCA-CoA were carried out in duplicate on two hBACAT protein preparations. Data are shown as means Ϯ S.E. or range (n ϭ 2-3).
Electrospray Mass Spectrometry (ESMS) Analysis-Incubation mixtures were set up containing chenodeoxycholoyl-CoA (50 M) or arachidoyl-CoA (40 M), hBACAT enzyme (2-5 g), and glycine (50 mM) in 50 mM potassium phosphate buffer, pH 8. The mixtures were incubated for 48 h at room temperature after which they were purified using Sepac C18 columns. The Sepac C18 columns were equilibrated with 5-ml volumes of chloroform:methanol (2:1), followed by 95% methanol in water, and finally by water. The incubation mixtures were then loaded onto the column followed by a further washing step with 5 ml of water. The column was then eluted using 5 ml of methanol, and the eluate was dried under nitrogen. The samples containing products from reactions with chenodeoxycholoyl-CoA and arachidoyl-CoA were reconstituted in ϳ50 l of methanol or tetrahydrofurane-water (1:1 v/v), respectively, of which 4 -5 l was loaded into gold-coated glass capillaries. Mass spectrometry was performed on a Quattro Micro triple quadrupole mass spectrometer (Micromass, Wythenshawe, Manchester, UK) equipped with a nano-electrospray ion source. Mass spectra were acquired in the negative ion mode over a mass scan range of m/z 50 -1200 for 2 min at a scan rate of 4 s/scan.
Identification of Putative Candidates for the Catalytic Triad of Human BACAT-Fasta3 (23) (European Bioinformatics Institute server, www.ebi.ac.uk/fasta33) was used to generate a multiple sequence alignment between hBACAT and its homologous sequences in the Gen-Bank TM . Examination of the sequence alignment identified a possible cysteine-aspartic acid-histidine catalytic triad.
Generation and Expression of Human BACAT Mutants-Point mutations were introduced by PCR using the QuikChange™ site-directed mutagenesis kit (Stratagene, La Jolla, CA). The mutagenic sense oligonucleotides used are shown in Table I. PCR reactions for single amino acid mutations were run for 16 cycles of 30 s at 95°C and 1 min at 55°C followed by 15 min at 68°C. The mutant plasmids were sequenced using the ABI Prism Dye Terminator Ready-reaction kit (PerkinElmer Life Sciences) at Cybergene AB (Huddinge, Sweden). The mutants were expressed and purified, and activity was determined on bile acid-CoA substrates as described above for the wild-type hBACAT.
Localization of Human BACAT Using Green Fluorescent Fusion Protein-The human BACAT open reading frame was cloned into the pcDNA3.1/NT-GFP vector (Invitrogen), in-frame with the GFP at the N-terminal end. Site-directed mutagenesis was carried out using the QuikChange™ site-directed mutagenesis kit, using the primer 5Ј-TC-CAGATGTGACCAGTAAACTCTAAGAAGACTAGATATTCC-3Ј and its reverse complement primer to mutate the C-terminal -SQL to -SKL. The mutated nucleotide (C to A) is underlined in the primer sequence, and the nucleotide change was verified by sequencing. Human skin fibroblasts were grown as described (26). Fibroblasts were transfected with 10 g of hBACAT/NT-GFP plasmid and the hBACAT/NT-GFP -SKL mutant plasmid using the calcium phosphate method as described (26), but without staining of the nucleus with Hoechst 33342.

Recombinant Expression and Characterization of hBACAT-
For recombinant expression of hBACAT, we cloned the corresponding cDNA by PCR. Sequencing of the cDNA identified four nucleotide differences compared with the published sequence (16), of which two result in the following changes: Pro-8 3 Arg and Arg-20 3 Gln. The bile acid conjugation and thioesterase activities of the expressed hBACAT enzyme were tested using choloyl-CoA and chenodeoxycholoyl-CoA as substrates in the presence or absence of glycine and taurine. Nonenzymatic hydrolysis of the CoA esters did not occur (data not shown), whereas the thioesterase activity, measured in the absence of glycine, was about 164 nmol/min/mg protein with 50 M CA-CoA and about 160 nmol/min/mg with 50 M CDCA-CoA (Fig. 1A). The addition of glycine to the reaction increased the activity to about 762 nmol/min/mg with CA-CoA and to about 733 nmol/min/mg with CDCA-CoA. Thus, CA-CoA appears to be a slightly better substrate than CDCA-CoA for both the thioesterase and conjugation activities of hBACAT. Also, glycine is a slightly better substrate than taurine, at least when measured at saturating concentrations. In contrast, the hBA-CAT enzyme showed very low activity with THCA-CoA, both in the presence and absence of glycine. It is noteworthy that the thioesterase activity is ϳ20% of the conjugation activity with both CA-CoA and CDCA-CoA. Both the thioesterase and conjugation activities of hBACAT were confirmed by ESMS (Fig. 1B).
Identification of the Putative Catalytic Triad in hBACAT-From an examination of the sequence alignment of hBACAT and its homologues in the GenBank TM data base, a putative catalytic triad consisting of residues Cys-235, Asp-328, and His-362 was identified in this enzyme (Fig. 2). Following identification of the amino acids in the catalytic triad, the Cys-235 3 Ala, Asp-328 3 Ala, and His-362 3 Glu mutants were generated to test the effects of these amino acid substitutions. The corresponding proteins were expressed and purified to near homogeneity as judged by Coomassie Brilliant Blue staining of SDS-PAGE (data not shown) and were characterized enzymatically. In an experiment to measure the activity of the generated mutants, choloyl-CoA conjugation activity (measured in the presence of 50 mM glycine) of wild-type hBACAT was about 1041 nmol/min/mg protein (Table II). However, the activities of the mutated proteins Cys-235 3 Ala, Asp-328 3 Ala, and His-362 3 Gln were all less than 0.4% of the wild-type activity, strongly suggesting that Cys-235, Asp-328, and His-362 indeed constitute the active site amino acids of hBACAT. We also generated a Cys-235 3 Ser mutant, which converted the BACAT into an efficient bile acid-CoA thioesterase (data not shown).
hBACAT Has Both acyl-CoA Thioesterase and Glycine Conjugation Activities with Long-and Very Long-chain Saturated Acyl-CoAs in Vitro-Because the BACAT enzymes show a high degree of homology to the type I acyl-CoA thioesterases, we hypothesized that the BACAT enzyme may also be active on acyl-CoAs. Bile acids are much more bulky substrates than straight long-chain acyl-CoAs, and acyl-CoAs are therefore likely to be accommodated in the active site of the BACAT enzyme. The expressed hBACAT enzyme was therefore incubated with fatty acyl-CoAs (each at 10 M) in both the absence and presence of 50 mM glycine (Fig. 3). Indeed, the hBACAT enzyme showed thioesterase activity with saturated long-chain acyl-CoAs, with the highest activity seen with saturated C 16 -C 20 acyl-CoAs. Incubation of these acyl-CoAs with 50 mM glycine inconsistently stimulated the activity to some degree. However, a consistent and larger glycine-dependent stimulation of hBACAT activity was observed with the very long-chain acyl-CoA substrates (C 22:0 -C 26:0 ). Introduction of double bonds resulted in a strikingly lower activity (compare C 18:1 -CoA, C 18.2 -CoA, and C 20:4 -CoA and the corresponding saturated acyl-CoAs). To identify the reaction products in more detail, we incubated hBACAT with arachidoyl-CoA (C 20:0 -CoA) as a model substrate and analyzed the reaction products by ESMS. As shown in Fig. 4a, incubation of hBACAT with arachidoyl-CoA and glycine produced both arachidic acid and Narachidoylglycine, the glycine conjugate of arachidic acid. Incubation of hBACAT with arachidoyl-CoA in the absence of glycine produced only arachidic acid (Fig. 4b). We also confirmed the formation of N-stearoylglycine by BACAT using ESMS (data not shown). A more detailed kinetic analysis of hBACAT activity with arachidoyl-CoA resulted in a V max of 223 nmol/min/mg and a K m of 19.3 M when measured in the presence of glycine (Fig. 5). These results show that hBACAT, in addition to its bile acid conjugating activity, also possesses thioesterase and glycine conjugating activities with long-and very long-chain acyl-CoAs, at least in vitro. Although the activity is much lower with acyl-CoAs as compared with the activity with bile acids, it is possible that both the thioesterase and glycine conjugating activities could serve important functions in vivo.
Human and Mouse BACAT and BACS Show Wide Tissue Expression-The novel activity of the hBACAT enzyme to hydrolyze and glycine-conjugate long-and very long-chain acyl-CoAs suggests a wider function for this enzyme also in fatty acid metabolism. Recently it was shown that the hBACS enzyme activates both bile acids and very long-chain fatty acids to their corresponding CoA-esters, suggesting a functional link between hBACS and hBACAT (7). We therefore set out to re-investigate the tissue expression of both the BACS and BACAT enzymes by RT-PCR in human and mouse tissues. Because of restricted access to human tissues, expression of BACAT and BACS was investigated only in human liver, gallbladder mucosa, and pancreas (Fig. 6). Interestingly, both BACAT and BACS were expressed in these tissues, showing a reasonably co-ordinate expression. In addition, data base searches revealed a number of expressed sequence tag (EST) clones encoding BACAT expressed mainly in human liver, as expected, but also in skin, heart, lung, and pancreas. In mouse, BACAT was strongly expressed in liver, kidney, gallbladder, proximal intestine, and distal intestine, and a band of the correct size was also weakly detected in adrenal, heart, lung, brain, and muscle tissue (Fig. 6B). A data base search revealed two EST clones from the adrenal gland, one of which was obtained from the Riken Consortium (GenBank TM accession no. BB593308, Riken clone A3300001A17). This clone was sequenced and verified to encode mouse BACAT (mBACAT). In addition, ESTs were also found from mouse gallbladder, cerebellum, aorta, and thymus, demonstrating a wide tissue expression of mBACAT. In mouse, BACS was most strongly expressed in liver and gallbladder, but expression was also evident in kidney, proximal intestine, and adrenal tissue and weakly in heart and muscle. We also compared the expression of hBACAT in mouse and human liver, showing that BACAT is much more strongly expressed in mouse than in human (Fig. 6C).
Human BACAT Is Localized in Cytosol-The human and mouse BACAT enzymes contain a C-terminal variant, -SQL, of the well characterized consensus peroxisomal type I targeting signal (PTS1) of -SKL (serine, lysine, leucine), the latter which has been shown to target proteins to peroxisomes (27). To establish whether BACAT is localized in peroxisomes, we cloned the human BACAT in-frame with GFP at the N-terminal end, which leaves the C-terminal -SQL sequence accessible. The plasmid encoding the hBACAT-GFP fusion protein was  transfected into human skin fibroblasts, and by using immunofluorescence microscopy for detection of a TRITC-labeled anti-GFP antibody, hBACAT showed a diffuse GFP expression with little sign of a punctate pattern, indicative of a cytosolic localization (Fig. 7A). Similar results were obtained by transfection of the mouse BACAT-GFP fusion plasmid (data not shown). We also mutated the C-terminal -SQL to -SKL, the consensus PTS1 targeting signal, and transfected this NT-GFP construct into human skin fibroblasts. Immunofluorescence microscopy showed that the hBACAT-SKL mutant was translocated to peroxisomes based on the punctate pattern (Fig. 7B). Similar results were also obtained for the mouse BACAT-SKL mutant (data not shown).

Characterization of Recombinant
Human BACAT-In this study we cloned human BACAT, expressed the enzyme in bac-teria, and purified the protein to near homogeneity for structural and biochemical characterization. The functionality of the expressed protein was assessed by activity measurements with CA-CoA and CDCA-CoA. Because the spectrophotometric assay measures conjugation activity indirectly (as release of CoASH), this activity was verified by ESMS. It has previously been shown that the substrate specificity of BACAT relates to the acyl chain length. Although other bile acids such as lithocholoyl-CoA and deoxycholoyl-CoA were shown to be good substrates for the bovine BACAT enzyme, compounds such as acetyl-CoA, benzoyl-CoA, and phenylacetyl-CoA proved not to be substrates (28). Also, shortening or extending the normal 4-substituted pentanoic acid side chain of CA-CoA by one methylene group, to produce norcholoyl-CoA and homocholoyl-CoA, respectively, results in a severe decrease in activity (29). Our data consolidate this point, as THCA-CoA proved to be a very poor substrate in comparison with CA-CoA and CDCA-CoA. The low activity with THCA-CoA is consistent with the observed accumulation of THCA in the serum of Zellweger patients, in which bile acid synthesis is severely impaired (30).
Identification of a Catalytic Triad in Human BACAT-It was confirmed previously by site-directed mutagenesis experiments that the residue Cys-235 is essential for BACAT activity (31). However, sequence alignments of BACAT to the type I acyl-CoA thioesterases and its homologues in the data bases led to the identification of a putative catalytic triad in BACAT consisting of residues Cys-235, Asp-328, and His-362 (21). This catalytic triad was found to be conserved in human, mouse, and rat BACAT sequences, in which the cysteine was located in a conserved SXCXG motif. Mutation of the Cys-235, Asp-328, and His-362 residues leads to abolishment of BACAT activity, which strongly implies that BACAT is an enzyme containing a Cys-His-Asp catalytic triad and confirms it as a member of the ␣/␤-hydrolase family. While this manuscript was in preparation, another study was published also identifying the Cys-235, Asp-328, and His-362 as the catalytic triad amino acids in human BACAT (32).
Human BACAT Can Also Conjugate Long-and Very Longchain Fatty Acids to Glycine-Although BACAT and the type I acyl-CoA thioesterases are assumed to have completely different functions in the cell, the striking sequence and structural homology between these enzymes suggest that they probably derive from a common ancestor. The type I thioesterases contain a nucleophilic serine in the active site, where a cysteine is located in BACAT, and we have previously reported an unsuccessful attempt to engineer mouse cytosolic acyl-CoA thioesterase I (CTE-I) into an acyltransferase by mutation of the serine to a cysteine (21). However, the Ser-232 3 Cys mutant of CTE-I became very strongly acylated when incubated with palmitoyl-CoA, demonstrating that replacement of Ser-232 with a cysteine allows a stable acyl-enzyme intermediate to be formed. Therefore the lack of acyltransferase activity in the CTE-I Ser-232 3 Cys mutant is likely to be due to structural restraints that would not allow the CTE-I active site to accept bulky substrates and/or a second nucleophile. With this in mind, we investigated whether BACAT can also act as a thio-esterase and, by having an active site proven to accept bulky substrates, whether BACAT could also accommodate the less bulky fatty acyl-CoA substrates. Although BACAT contains a cysteine as the nucleophilic residue, the wild-type enzyme catalyzed the hydrolysis of CA-CoA and CDCA-CoA to the free bile acid and CoASH with an activity of ϳ20% of the conjugating activity. Interestingly, the BACAT enzyme also hydrolyzed saturated fatty acyl-CoAs of 16 -26 carbons in chain length. The activity with fatty acyl-CoAs was ϳ20% of that with CA-CoA. However, mono-, di-, and polyunsaturated acyl-CoAs were much poorer substrates than the corresponding saturated acyl-CoA. The addition of glycine had only a marginal stimulatory effect on the activity with 14 -18 carbon acyl-CoAs, whereas the addition of glycine clearly stimulated the activity with C20:0 -C26:0 acyl-CoAs, indicating that fatty acids may also be conjugated to glycine by the BACAT enzyme in vitro. Analysis of incubation products by ESMS showed that BACAT indeed has the ability to conjugate fatty acids to glycine. This acyl chain length specificity of BACAT suggests a possible function in the conjugation of long-and very long-chain saturated fatty acids to glycine.
The Human BACAT Enzyme Is Cytosolic-There has been some controversy during the past years concerning the subcellular localization of BACAT. Previous studies have reported BACAT activity in the lysosomal fraction of human liver (33), in the microsomal/peroxisomal fraction of rat liver homogenates (4,34) and guinea pigs (35), as well as in the soluble fraction of liver homogenates from rat (8). Experiments carried out on homogenates from frozen human liver showed BACAT to be cytosolic, although the idea was not excluded that the freezing and thawing of the liver may have caused lysis of the cellular organelles leading to a redistribution of BACAT activity from the peroxisomes to the cytosol (13). More recently it has been shown that BACAT activity is present both in cytosolic and peroxisomal fractions (5). Although both the human and mouse BACAT proteins contain a variant -SQL (16) of the type I peroxisomal targeting signal (PTS1) -SKL (36), our data show that the BACAT protein is mainly cytosolic. However, we cannot exclude the possibility of some very weak peroxisomal localization, which may not be visible due to the strong expres- sion in cytosol. Mutation of the -SQL sequence to -SKL results in the translocation of the protein from the cytosol, exclusively to peroxisomes. These results are consistent with recent findings showing that the C-terminal -SQL is only able to interact strongly with the PTS1 receptor if it is preceded by a basic amino acid at position Ϫ4 (37). A study on human catalase demonstrated that the last four amino acids (-KANL) are necessary and sufficient to target this protein to peroxisomes (38). Mutation of the lysine at position Ϫ4 to a non-basic residue was shown to abolish targeting to peroxisomes. Our data showing that the human BACAT (ending -TSQL) and mouse BACAT (ending -SSQL) enzymes are cytosolic further underscore the importance of a basic residue at position Ϫ4 in at least some variants of the -SKL consensus PTS1.
Possible Physiological Functions for the Cytosolic BACAT Enzyme-There are two pathways for the conjugation of bile acids to glycine or taurine in the cell. BACAT activity is present in peroxisomes for the amidation of de novo synthesized bile acids, whereas a cytosolic BACAT enzyme functions in amidation of recycled bile acids (5). The specific BACAT activity from human (5) and mouse liver 2 is much higher in the peroxisomal fraction compared with the cytosolic fraction. Based on the present finding that BACAT is mainly cytosolic, it is proposed that the function of this BACAT is in the conjugation of recycled bile acids, and it is predicted that there exists another, as yet unidentified, peroxisomal BACAT enzyme that conjugates de novo synthesized bile acids. Conjugation of bile acids is believed to occur mainly in the liver. In mouse, the cytosolic BACAT enzyme is strongly expressed in liver, gallbladder, and intestine (both proximal and distal), compatible with an important function in amidation of bile acids and protection of gastrointestinal mucosal cells from the accumulation of free bile acids. Recycled free bile acids need to be activated to the corresponding CoA-ester prior to conjugation to glycine by the BACAT enzyme. Recent studies have shown that bile acids are activated mainly by BACS, a microsomal bile acid-CoA synthetase, while THCA, a precursor of de novo bile acid synthesis in peroxisomes, is activated mainly by VLCS, located in the peroxisomal and endoplasmic reticulum membranes (7,39). Pircher et al. (40) have now shown that both BACAT and BACS are target genes of the farnesoid X receptor (FXR), the key nuclear receptor involved in the regulation of bile acid synthesis, consolidating their function in the conjugation of bile acids. Previous studies have shown the expression of BACS (24) and the mouse homologue of BACS, the very long-chain acyl-CoA synthetase-related protein (25), to be largely liver-specific, but our data also show expression in human gallbladder mucosa and pancreas. In mouse, BACS expression was strong in the liver and gallbladder, consistent with a function in the activation of bile acids. The novel activities of BACAT in hydrolysis of very long-chain acyl-CoAs and the conjugation to glycine suggest that BACAT may have more general functions in fatty acid metabolism than was believed previously and therefore would be expressed also in other tissues. Kwakye et al. showed the presence of both BACAT mRNA and immunoreactive protein (13) along with BACS activity (41) in rat kidney, which is to date the only reported extra-hepatic expression of BACAT. We reinvestigated tissue expression of BACAT in more detail using RT-PCR and found strong expression in liver, kidney, gallbladder, and proximal and distal intestine, with weak expression in FIG. 6. Tissue expression of human and mouse BACAT and BACS. A, RT-PCR for BACAT and BACS was carried out on 1 g of total RNA isolated from various human tissues. Lanes marked with an asterisk are samples from patients suffering from cholesterolosis. B, RT-PCR for BACAT and BACS was carried out on 1 g of total RNA isolated from various mouse tissues. C, comparison of hBACAT expression by RT-PCR in human and mouse liver. Blank reactions did not contain any template RNA. DNA ladder sizes are indicated on the left in bp. 7. hBACAT is localized mainly in cytosol. Human skin fibroblasts were transfected with a hBACAT-GFP plasmid (with the C-terminal sequence -SQL) and the same plasmid mutated in the Cterminal to -SKL. The cells were incubated with a rabbit GFP antibody and then with a Cy3-conjugated affinity-purified donkey anti-rabbit IgG and examined by immunofluorescence microscopy. A, wild-type hBACAT (-SQL); B, hBACAT-SKL mutant. adrenal gland, muscle, lung, and brain. Computer searches in data bases identified ESTs from liver, gallbladder, cerebellum, and adrenal gland (in line with the data shown in Fig. 6) but also in the aorta (four ESTs) and thymus (one EST). In human, expression was high in liver, as expected. We also analyzed four samples from gallstone patients, two of which showed cholesterolosis and two that did not. Expression of BACAT mRNA was at least as high in the gallbladder mucosa of two of the patients as in normal liver, although the different expression levels did not appear to correlate to the different phenotypes. Expression of BACAT was also evident in the pancreas. Searches in human data bases identified ESTs mainly in liver but also in lung, skin, heart, and pancreas. Our data also show that BACS is expressed in tissues other than liver, i.e. in the gallbladder mucosa and pancreas in human. Interestingly, BACS and VLCS also activate very long-chain fatty acids to their CoA-esters, suggesting that co-expression of the BACS and VLCS enzymes with the cytosolic BACAT may mediate conjugation of fatty acids. Although in vivo conjugates of medium-chain fatty acids to glycine have been detected in the blood of patients suffering from defects in fatty acid metabolism (42,43), only sparse information is available concerning the in vivo conjugation of long-chain fatty acids to glycine. Here we have shown that BACAT, an enzyme located in a wide number of tissues in the body, can catalyze the formation of N-acylglycines, which could provide a possible pathway for excretion of these fatty acids under conditions in which they might otherwise become toxic.

FIG.
In conclusion, we have identified Cys-235, Asp-328, and His-362 in human BACAT as the catalytic triad of amino acids and deduced that BACAT is member of the ␣/␤-hydrolase enzyme family. Furthermore, we show that BACAT is mainly a cytosolic enzyme, which probably functions in the conjugation of recycled bile acids. The novel thioesterase and glycine conjugating activities on long-and very long-chain fatty acyl-xCoA substrates, together with a wide tissue expression, suggests that BACAT may be involved in protection against the accumulation of free bile acids and very long-chain fatty acids throughout the body.