Participation of Two Members of the Very Long-chain Acyl-CoA Synthetase Family in Bile Acid Synthesis and Recycling*

Bile acids are synthesized de novo in the liver from cholesterol and conjugated to glycine or taurine via a complex series of reactions involving multiple or-ganelles. Bile acids secreted into the small intestine are efficiently reabsorbed and reutilized. Activation by thio-esterification to CoA is required at two points in bile acid metabolism. First, 3 (cid:1) ,7 (cid:1) ,12 (cid:1) -trihydroxy-5 (cid:2) -choles-tanoic acid, the 27-carbon precursor of cholic acid, must be activated to its CoA derivative before side chain cleavage via peroxisomal (cid:2) -oxidation. Second, reutilization of cholate and other C 24 bile acids requires reacti- vation prior to re-conjugation. We reported previously that homolog 2 of very long-chain acyl-CoA synthetase (VLCS) can activate cholate (Steinberg, S. J., Mihalik, S. J., Kim, D. G., Cuebas, D.

The synthesis of bile acids from cholesterol is a complex process requiring modifications of both the steroid nucleus and the aliphatic side chain (reviewed in Refs. 1 and 2). Cholic acid and chenodeoxycholic acid, the two primary bile acids in humans, are synthesized in the liver via the concerted action of enzymes located in the endoplasmic reticulum, cytosol, mitochondria, and peroxisomes. The immediate precursors of the C 24 bile acids cholate and chenodeoxycholate are the C 27 com-pounds 3␣,7␣,12␣-trihydroxy-5␤-cholestanoic acid (THCA) 1 and 3␣,7␣-dihydroxy-5␤-cholestanoic acid (DHCA), respectively (1,2). It is well established that chain shortening of the methyl-branched side chain of THCA and DHCA by ␤-oxidation takes place in peroxisomes (3). For ␤-oxidation to occur, THCA and DHCA must first be activated to their CoA thioesters. The products of chain shortening are choloyl-CoA and chenodeoxycholoyl-CoA and not the free bile acids (4).
The CoA thioesters of the primary bile acids are conjugated to the amino acid glycine or taurine in the hepatocyte before secretion into the bile canaliculi (1,2). After bile secretion into the intestine during digestion, the conjugated primary bile acids are subject to deconjugation, 7␣-dehydroxylation, and other modifications by the intestinal flora (1,5). 7␣-Dehydroxylation converts conjugated or unconjugated cholate to deoxycholate, and chenodeoxycholate to lithocholate. Deoxycholate and lithocholate are referred to as secondary bile acids. A substantial portion of the intestinal bile acid pool is reabsorbed by passive diffusion throughout the small bowel and by active transport in the distal ileum; all of these compounds return to the liver via the enterohepatic circulation for reutilization (6). Recycled unconjugated primary and secondary bile acids must then be re-conjugated to glycine or taurine before return to the bile. However, re-conjugation can occur only after the free acids are activated to their CoA derivatives.
We recently demonstrated that human very long-chain acyl-CoA synthetase homolog 2 (hVLCS-H2), a liver-specific member of the protein family that includes very long-chain acyl-CoA synthetases (VLCSs) and fatty acid transport proteins, is a choloyl-CoA synthetase (7). We hypothesized that a major function of this protein is to activate recycled bile acids prior to their re-conjugation. In this report, we provide further evidence for a role for hVLCS-H2 in activation of primary and secondary bile acids. Furthermore, we demonstrate that hVLCS and hVLCS-H2 are capable of activating the cholic acid precursor (THCA) and speculate on the complementary role of these proteins in bile acid synthesis. Based on the substrate specificity of hVLCS-H2, we will hereafter refer to this protein as human bile acyl-CoA synthetase (hBACS).
Synthesis of THCA and [26-14 C]THCA-⌬24-THCA was synthesized by the method of Xu and Cuebas (10). Unlabeled THCA was prepared by the catalytic hydrogenation of ⌬24-THCA using Adams' catalyst (Aldrich) suspended in methanol at 1-atmosphere pressure for 12 h. The product is a racemic mixture of (25S)-and (25R)-THCA stereoisomers. [26-14 C]THCA was prepared from THCA by a modification of the method of Tserng and Klein (11). THCA was formylated to protect the hydroxyl groups by heating 0.24 mmol in 1 ml of 90% formic acid containing 0.025 ml of 70% perchloric acid for 2 h at 55°C. After cooling to ϳ45°C and dropwise addition of acetic anhydride until massive bubbling occurred, the solution was further cooled to room temperature, and the formylated THCA was extracted with diethyl ether and dried (85% yield). The corresponding 25-chloronorcholane was then obtained by dissolving 0.2 mmol of the formylated THCA in 2.5 ml of dry benzene containing 0.4 mmol of lead tetraacetate and 0.4 mmol of LiCl and heating to 85°C for 3 h. At this time and again at 4 h, 0.2 mmol of LiCl was added. The reaction mixture was then cooled, centrifuged to remove precipitated material, washed with ice-cold 2% NaOH and ice-cold water, dried with MgSO 4 , and evaporated to dryness to afford 25-chloro-3␣,7␣,12␣-triformyloxynorcholane (75% yield). 40 mol of this product was dissolved in 1 ml of dry dimethyl sulfoxide containing 2.0 mCi of Na 14 CN (36 mol, 55 mCi/mmol) and heated for 32 h at 70°C. After cooling, 1 ml of 2 N NaOH was added, and the solution was heated to 90°C for 1 h to hydrolyze the formyl groups. The product (cyano-3␣,7␣,12␣-trihydroxy[26-14 C]norcholane) was then extracted with chloroform, dried, and purified by preparative thin-layer chromatography in cyclohexane/ethyl acetate/acetic acid (21:69:9; 1.55 mCi; 77% yield). Finally, the purified nitrile was dissolved in 2 ml of ethanol and 10% aqueous NaOH (2:1) and heated to 80°C for 40 h. After cooling, the reaction mixture was concentrated to one-third of its original volume to remove ethanol and restored again to its original volume by adding water. After acidification, the reaction product was extracted with chloroform. Analysis by thin-layer and gas-liquid chromatography showed that the reaction product was identical to THCA (1.22 mCi; 79% yield).
Cloning and Expression of VLCS and BACS Constructs-Full-length cDNAs encoding hVLCS and hBACS were cloned into the expression vector pcDNA3 (Invitrogen) as previously described (8,12). Cloning of N-myc-hBACS was also described previously (12).
For production of N-myc-hVLCS, the full-length open reading frame of hVLCS was amplified by PCR using oligonucleotide primers P0-17 (5Ј-CCCGTCGACCATGCTTTCCGCCATCTACA-3Ј, which incorporates a SalI restriction site) and P0-4 (5Ј-GTCTGCGGCCGCCCTCCAA-CAAGCTCAGTTTGC-3Ј, which incorporates a NotI site) with hVLCS in pcDNA3 as template. The product was ligated into the XhoI and NotI sites of N-myc-pcDNA3, a modified pcDNA3 vector that places the c-myc epitope in-frame at the amino terminus (a gift of Dr. S. Gould).
A similar approach was used to produce C-myc-hBACS. A 1801-bp fragment containing the start codon was excised from hBACS in pcDNA3 using KpnI and NheI. A 422-bp fragment of hBACS in which the stop codon was replaced by a HindIII restriction site was amplified by PCR using hBACS in pcDNA3 as template with oligonucleotide primers P2-9 (5Ј-CGCCTCGGGGACACCTTC-3Ј, forward) and P2-18 (5Ј-TTTAAGCTTGAGCCTCCAGGTTCCCTC-3Ј, reverse); this fragment was cut with NheI and HindIII, yielding a 270-bp fragment. These two pieces were ligated into the KpnI and HindIII sites of pcDNA3.1-myc-HisA, forming C-myc-hBACS.
Expressed sequence tag (EST) clones representing the murine homolog of hVLCS were identified by homology probing of the EST data base using rat VLCS cDNA sequence (13) as the query. An EST clone (GenBank TM accession number AA038113) containing the C-terminal 921 base pairs of the open reading frame plus 8 bp of the 3Ј-untranslated region was obtained from Genome Systems, Inc. The EST clone insert was moved to the mammalian cloning vector pcDNA3 using EcoRI and NotI. The 5Ј-untranslated region of rat VLCS (13) was used to design a forward PCR primer (5Ј-CGCGGGTACCGATCTCACCT-GACCCCAGACG-3Ј) with an engineered KpnI site and to amplify the 5Ј-end of mouse VLCS from liver total cDNA with a reverse primer (5Ј-tgccttcagtggatgcgtagaac-3Ј). The ϳ1100-bp PCR product encompassed a second KpnI site near the 3Ј-end. Both KpnI sites in the PCR fragment were used to clone the PCR fragment into a KpnI site near the 5Ј-end of EST AA038113 in pcDNA3, and the clone was re-sequenced. The full-length open reading frame of mouse VLCS was deposited in the GenBank TM /EBI Data Bank (accession number AF033031).
The completed mouse VLCS cDNA sequence was used to search the EST data base for related mouse ESTs. ESTs AA106121 and AA538033 both contained full-length cDNA encoding the murine homolog of hBACS. The insert of EST AA106129 was released by digestion with AccI and SpeI and was blunt-ended with Klenow enzyme. The fragment was then cloned into the EcoRV site of pcDNA3, and the clone was sequenced.
In most expression experiments, COS-1 cells (a gift of Dr. C. Thompson) were transfected with 10 g of plasmid DNA by electroporation as described previously (8). In some experiments, cells were transfected using GeneSHUTTLE-40 (Qbiogene, Inc.). 10 g of plasmid DNA was mixed with 30 l of GeneSHUTTLE-40 and incubated with COS-1 cells in serum-free medium for 8 h. Following addition of an equal volume of culture medium containing 20% fetal bovine serum, cell incubation was continued for 16 h. At this time, the culture medium was replaced with culture medium containing 10% fetal bovine serum, and the cells were harvested 72 h post-transfection. Washed cell pellets were resuspended in 0.25 M sucrose containing 10 mM Tris (pH 8.0) and 1 mM EDTA, frozen at Ϫ80°C for a minimum of 4 h, and thawed on ice for assay.
Indirect Immunofluorescence-COS-1 cells were transfected as described above with hVLCS or hBACS containing the c-myc epitope in-frame at either the N or C terminus of the protein. After 3 days, cells were fixed in 4% formaldehyde in phosphate-buffered saline and solubilized with either 1.0% Triton X-100 or 15 g/ml digitonin prior to incubation with primary and secondary antibodies as described previously (14).
Bile Acyl-CoA Synthetase and Fatty Acyl-CoA Synthetase Assays-Both isotopic and high-performance liquid chromatography (HPLC)based assays of bile acid activation were performed as previously described (7), with the following exceptions. The poorly soluble sodium salt of lithocholate was dried on the bottom of reaction tubes and was solubilized using 50 l of ␤-cyclodextrin (10 mg/ml) in 10 mM Tris (pH 8.0) by brief sonication prior to addition of other reaction components and enzyme. In the experiments shown in Figs. 1C and 3B, the gradient program for elution of bile acyl-CoA derivatives from the reverse-phase HPLC column (25-65% solvent B over 30 min), which deviated from the previously described method (25-50% solvent B over 25 min) (7). The unlabeled THCA used in these experiments also contained ⌬24-THCA (8%, w/w). In the HPLC tracing shown in Fig. 1D, peak 3 was tentatively identified as the CoA derivative of ⌬24-THCA. The retention time of this product was identical to that of ⌬24 THCA-CoA produced from the incubation of 24-hydroxy-THCA (varanoyl-CoA) with fibroblasts from patients with mutations in the 3-hydroxyacyl-CoA dehydrogenase domain of peroxisomal D-bifunctional protein. Tracings from patients with mutations in the enoyl hydratase portion of the gene had severe reductions in this product, as described previously (15). For inhibition experiments (see Fig. 3B), (25S)-THCA-CoA was quantitated using a Hewlett-Packard 3392A integrator. Activation of [1-14 C]palmitic acid by transfected COS-1 cells was measured as previously described (8).
In Situ Hybridization-The method of Giger et al. (16) was used, with modification. Plasmids containing full-length cDNA encoding the murine homologs of VLCS and BACS in pcDNA3 were linearized by cutting at a restriction site upstream of the initiator methionine codon (for antisense probes) or just downstream of the stop codon (for sense probes). Antisense RNA probes for VLCS (1865 bp) and BACS (2071 bp) were synthesized using digoxigenin-labeled UTP (Roche Molecular Biochemicals) and SP6 RNA polymerase (Ambion Inc.). The corresponding digoxigenin-labeled sense probes were synthesized for use as controls using T7 RNA polymerase (Ambion Inc.). Livers obtained from adult mice were quick-frozen in liquid nitrogen. Slices (8 m) were obtained using a Microm HM500M cryostat, collected on Superfrost Plus slides (Fisher), fixed for 30 min with 4% paraformaldehyde in phosphatebuffered saline, and acetylated for 10 min in 0.25% acetic anhydride in 0.1 M triethanolamine. After 8 h of prehybridization in 50% formamide containing 5ϫ Denhardt's solution (Sigma), 5ϫ SSC, and 0.25 mg/ml yeast tRNA (Invitrogen), slices were hybridized in the same buffer containing digoxigenin-labeled probes (200 ng/ml) for 16 h at 60°C. The slides were washed under high-stringency conditions (5ϫ SSC in 50% formamide for 5 min at 59°C, 2ϫ SSC for 1 min at 59°C, 0.2ϫ SSC for 30 min at 65°C, and 0.2% SSC for 5 min at room temperature). After a 1-h incubation in blocking reagent (Roche Molecular Biochemicals), slices were incubated for 1.5 h with alkaline phosphatase-conjugated anti-digoxigenin antibody (Roche Molecular Biochemicals). Color development with nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate was carried out for 16 h.

hBACS Catalyzes Activation of Primary and Secondary Bile
Acids-We previously showed that when expressed in COS-1 cells, hBACS (formerly called hVLCS-H2, a member of the VLCS/fatty acid transport protein family of acyl-CoA synthetases) catalyzes the activation of cholic acid to its CoA derivative (7). This observation was consistent with the tissue (liver) and subcellular (endoplasmic reticulum) distribution of this protein (17,18). We also reported that hBACS weakly activates the very long-chain fatty acid lignoceric acid (C 24:0 ; 20-40% of the activity measured with cholate) and was not able to activate the long-chain fatty acid palmitic acid (C 16:0 ) (7,12). To determine whether the other human primary bile acid, chenodeoxycholic acid, and the secondary bile acids deoxycholic acid and lithocholic acid are also substrates for hBACS, we incubated these compounds with ATP, MgCl 2 , CoA and freeze/thaw-disrupted COS-1 cells expressing this or related proteins. The products were separated by HPLC as previously described (7). When chenodeoxycholate was the substrate, a product with a retention time of 20.3 min was detected in cells expressing hBACS (Fig. 1A). No product was detected in cells transfected with hVLCS, hVLCS-H1, rat liver long-chain acyl-CoA synthetase-1 (rACS1), or the vector (pcDNA3) alone (Fig. 1A). We had previously reported that no product is formed in the absence of ATP, CoA, or the bile acid substrate (7) (data not shown). To substantiate the results of the HPLC-based assay, we incubated COS-1 cells expressing hBACS and related proteins with radiolabeled chenodeoxycholate (Fig. 2A). Again, only cells expressing hBACS were capable of catalyzing chenodeoxycholate activation. Thus, we conclude that chenodeoxycholate is also a substrate for hBACS.
Similar results were obtained using the HPLC-based assay when the secondary bile acids deoxycholate (Fig. 1B) and lithocholate (Fig. 1C) were substrates for activation by hBACS. The putative products deoxycholoyl-CoA and lithocholoyl-CoA eluted at 21.7 and 19.7 min, respectively. No product was detected with either substrate in cells expressing hVLCS, hVLCS-H1, or rACS1, consistent with the hypothesis that the liver-specific hBACS is a bile acyl-CoA synthetase capable of activating both primary and secondary bile acids.
Similar to the situation with primary and secondary bile acids, no product was detected when THCA was incubated with COS-1 cells expressing hVLCS-H1, rACS1, or the vector alone (Fig. 1D). In contrast, cells expressing hVLCS (which were not capable of activating the bile acids) yielded the same three product peaks as did cells expressing hBACS (Fig. 1D). This observation suggests that both hVLCS and hBACS are capable of activating THCA for its subsequent conversion to cholate.
THCA Inhibits hBACS-catalyzed Activation of Cholate-To elucidate the role(s) of hVLCS and hBACS in bile acid metabolism, we first investigated the ability of the biosynthetic precursor of cholate (THCA) to inhibit cholate activation. We incubated COS-1 cells expressing hBACS with radiolabeled cholate in the presence of increasing concentrations of THCA. A concentration of THCA (4 M) that was 20% of the cholate concentration (20 M) reduced cholate activation by nearly half (Fig. 3A). When THCA was present at the same (20 M) and at double (40 M) the cholate concentration, we observed 72 and 82% inhibition, respectively (Fig. 3A).
Cholate Inhibits Activation of THCA by hBACS, but Not by hVLCS-Because cholate is a substrate for hBACS, but not for hVLCS, we next examined the effect of unlabeled cholate on THCA activation by these two enzymes. When COS-1 cells expressing hBACS were incubated with 20 M THCA, formation of (25S)-THCA-CoA was reduced by 11, 21, and 39% when 4, 20, and 40 M cholate, respectively, were present in the assay (Fig. 3B). In contrast, with COS-1 cells expressing hVLCS, an enzyme that does not activate cholate, there was no apparent cholate inhibition of THCA activation (Fig. 3B).
Kinetic Studies of Cholate and Palmitate Activation-Kinetic studies were performed to establish further that hVLCS does not activate bile acids and that hBACS does not activate longchain fatty acids. COS-1 cells expressing hBACS exhibited normal Michaelis-Menten kinetics when cholate was the substrate (Fig. 4A); cells transfected with the vector alone had no activity with this substrate (data not shown). Under these conditions, the K m of hBACS for cholate was 2.8 M, and the V max was 3.8 nmol/20 min/mg of COS cell protein, as determined by nonlinear regression analysis. Similar experiments were performed using the substrate palmitic acid (C 16:0 ), a long-chain fatty acid that is the preferred substrate for hVLCS (8). Kinetic analysis is complicated by the fact that untransfected and vector-transfected COS-1 cells contain several endogenous acyl-CoA synthetases that can activate C 16:0 (see Fig.  4B); thus, endogenous activity must be subtracted from that of hBACS-transfected cells. At each of five C 16:0 concentrations tested (range of 2.5-40 M), there was no statistically significant increase in C 16:0 activation by hBACS-transfected cells compared with pcDNA3-transfected cells (p Ն 0.1, n ϭ 3; data not shown). It is therefore unlikely that C 16:0 is activated to a significant degree by hBACS in vivo.
The kinetics of C 16:0 and cholate activation by COS-1 cells expressing hVLCS were also examined. When C 16:0 was the substrate, both pcDNA3-and hVLCS-transfected COS-1 cells exhibited apparently normal Michaelis-Menten kinetics (Fig.  4B). Nonlinear regression analysis revealed a K m of 48 M and a V max of 167 nmol/20 min/mg of COS cell protein for vectortransfected cells and a K m of 32 M and a V max of 175 nmol/20 min/mg of COS cell protein for cells expressing hVLCS. The contribution of hVLCS to the overall C 16:0 -CoA synthetase activity was estimated by subtraction (Fig. 4B), and a K m of 12 M and a V max of 28 nmol/20 min/mg of COS cell protein were calculated. It must be emphasized that these derived values are, at best, crude estimates. In our previous study (7), we found that neither vector-transfected nor hVLCS-transfected COS-1 cells activate 20 M cholate to its CoA derivative. We now report that no detectable activation was observed at any of five concentrations of cholate tested (range of 2.5-40 M) (data not shown).
Hepatic Organization of VLCS-and BACS-expressing Cells-We hypothesize that the primary metabolic role of hBACS is bile acid activation. On the other hand, although hVLCS may play a role in bile acid synthesis from THCA, this enzyme can activate a wide variety of substrates, including long-, very long-, and branched-chain fatty acids (7,8). If these two synthetases have different functions, the intrahepatic localization of cells expressing them might differ. We used in situ hybridization to address this question in mouse liver. Using an antisense RNA probe synthesized from cDNA encoding the murine ortholog of hVLCS, we found that hepatocytes uniformly expressed this enzyme (Fig. 5, D and E). In contrast, hepatocytes expressing the murine homolog of hBACS mainly had a periportal distribution (Fig. 5, A and B). These observations are consistent with the differing proposed roles of the two acyl-CoA synthetases in bile acid and fatty acid metabolism.
Topographic Orientation of hVLCS and hBACS Expressed in COS-1 Cells-We previously reported that although hBACS is primarily a microsomal enzyme, hVLCS is localized to both peroxisomes and the endoplasmic reticulum (8,20). Other laboratories have reported that activation of both bile acids (17,18,21) and the bile acid precursors DHCA and THCA (22) occurs in microsomes. The fact that hBACS activates both bile acids and THCA is consistent with these reports. Our observation that hVLCS, with its dual subcellular location, can activate THCA, but not bile acids, raised questions about the precise role(s) of hVLCS and hBACS in bile acid synthesis.
An initial approach to answer this question was to examine the topographic orientation of hVLCS and hBACS in the endoplasmic reticulum membrane. For these studies, epitope tags were placed at either end of the expressed proteins. Full-length hVLCS cDNA clones incorporating an in-frame myc epitope at the N and C termini were constructed as outlined under "Experimental Procedures." Similar myc-tagged constructs of hBACS were prepared. COS-1 cells expressing these cDNAs were examined by indirect immunofluorescence using anti-myc antibody to determine the orientation of the expressed protein.
We previously reported that when COS-1 cells expressing hVLCS are permeabilized with Triton X-100, the protein is detected primarily in the endoplasmic reticulum, whereas the endogenous protein is also detected in peroxisomes (8,20). COS-1 cells transfected with either N-myc-hVLCS (Fig. 6A) or C-myc-hVLCS (Fig. 6E) also had a staining pattern consistent with expression in the endoplasmic reticulum. The location was confirmed by colocalization of immunofluorescence of hVLCS with that of the endoplasmic reticulum luminal resident protein calreticulin (Fig. 6, B and F). Also as reported previously (12), an endoplasmic reticulum N-myc-hBACS staining pattern was observed in Triton X-100-permeabilized COS-1 cells expressing this protein (Fig. 6I). Cells expressing C-myc-hBACS had an identical immunostaining pattern (Fig. 6M).
Permeabilization of cells with digitonin rather than Triton X-100 yielded a somewhat different result. The C-myc-hVLCS protein was detected in digitonin-permeabilized COS-1 cells transfected with its cDNA (Fig. 6G), but N-myc-hVLCS was not visualized (Fig. 6C). Under the conditions utilized, the luminal protein calreticulin could not be detected (Fig. 6, D and H). These data indicate that although the C terminus of hVLCS faces the cytoplasm, the N terminus faces the lumen of the endoplasmic reticulum. We previously reported that the C terminus of endogenous hVLCS is oriented toward the peroxisomal matrix and not the cytoplasm (20). In contrast, both the N-myc-hBACS (Fig. 6K) and C-myc-hBACS (Fig. 6O) proteins could be detected in digitonin-permeabilized cells (compare with calreticulin) (Fig. 6, L and P). These data suggest that hBACS may preferentially act on substrates present in the cytoplasm.

DISCUSSION
The synthesis of bile acids from cholesterol, conjugation with amino acids, secretion into the bile and subsequently into the FIG. 6. Topographic orientation of hVLCS and hBACS in endoplasmic reticulum membranes. COS-1 cells were transfected with full-length hVLCS or hBACS constructs containing the c-myc epitope in-frame at either the N or C terminus of the protein. Three days following transfection, cells were fixed and permeabilized with either Triton X-100 (exposing antigens facing both the cytoplasm and the endoplasmic reticulum lumen) or digitonin (exposing antigens facing the cytoplasm only). Double-label indirect immunofluorescence analyses using antibodies specific for the c-myc epitope and the endoplasmic reticulum luminal resident protein calreticulin were performed as described under "Experimental Procedures." The anti-myc antibody was detected by a fluorescein-conjugated secondary antibody (green), whereas the anti-calreticulin antibody was detected by a lissamine rhodamine-conjugated secondary antibody (red). A-H, cells expressing hVLCS; I-P, cells expressing hBACS. The position of the epitope tag (Nversus C-terminal) and the permeabilization method (Triton X-100 versus digitonin) are indicated on the right. Both C-myc-hBACS and Nmyc-hBACS were enzymatically active when expressed in COS-1 cells (ϳ50 and ϳ20% of untagged hBACS activity, respectively). N-myc-hVLCS and untagged hVLCS activated palmitic acid equally well, but the activity of C-myc-hVLCS was Ͻ20% than that of hVLCS. However, the results of topographic experiments with hVLCS (untagged and enzymatically active)-transfected cells and an antibody raised against the C terminus of this protein gave results identical to those obtained with C-myc-hVLCS and anti-myc antibody (7) (data not shown). intestinal tract, reabsorption from the intestine, and reutilization require the integrated activity of multiple enzymes, organelles, and tissues. At two points in the hepatic phase of bile acid metabolism, activation of the carboxylic acid side chain of the steroid to a CoA derivative is required. First, in the pathway of de novo synthesis from cholesterol, the C 27 precursors of cholate and chenodeoxycholate (THCA and DHCA, respectively) must be converted to THCA-CoA and DHCA-CoA for cleavage of three carbons from the side chain to occur. Second, to conjugate the C 24 bile acids cholate and chenodeoxycholate to glycine or taurine, their carboxylic acid side chains must first be activated to the corresponding CoA derivatives. The studies reported here begin to address the nature of the acyl-CoA synthetases that are involved in these two aspects of bile acid metabolism.
Based on the biochemical and morphological data presented here and our previous findings (7), we have developed a working hypothesis regarding the role(s) of hVLCS and hBACS in bile acid metabolism (Fig. 7). Conversion of cholesterol to THCA and DHCA requires several enzymatic steps that take place in the endoplasmic reticulum, cytosol, and mitochondria (1,2). Activation of THCA and DHCA, precursors of cholic acid and chenodeoxycholic acid, respectively, could be catalyzed by either hVLCS or hBACS in the endoplasmic reticulum or by hVLCS in peroxisomes. Once inside peroxisomes, the side chains of THCA-CoA and DHCA-CoA are shortened by three carbons via ␤-oxidation, yielding choloyl-CoA and chenodeoxycholoyl-CoA, respectively. The peroxisomal enzyme bile acyl-CoA:amino-acid N-acyltransferase (23) then catalyzes the formation of the glycine and taurine conjugates of cholic acid and chenodeoxycholic acid. Exiting the peroxisome, the conjugated bile acids are secreted into bile canaliculi and eventually reach the small intestine. Although a large proportion of the conjugated bile acids remain intact, a small quantity undergoes conversion in the bowel. Intestinal flora catalyze both deconjugation and dehydroxylation reactions (1,2). 7␣-Dehydroxylation of cholate and chenodeoxycholate yields the secondary bile acids deoxycholate and lithocholate, respectively. Intestinal metabolism thus yields a mixture of primary and secondary bile acids that includes both conjugated and free acids. Reabsorption of free bile acids occurs by passive diffusion throughout the small bowel (1,24). When these free acids reach the liver via the portal circulation, they must be reactivated and then conjugated before they are returned to the bile. hBACS is uniquely positioned in the endoplasmic reticulum membrane to catalyze the reactivation reaction. Bile acyl-CoA:amino-acid N-acyltransferase is found in the cytoplasm as well as the peroxisome (17,25). Following re-conjugation in the cytoplasm, primary and secondary bile acids can be secreted into the bile for reutilization. This hypothesis is consistent with that recently proposed by Solaas et al. (25), who investigated bile acyl-CoA:amino-acid N-acyltransferase in human liver.
Using subcellular fractions from rat liver, other investigators have shown that activation of THCA to its CoA derivative occurs primarily in microsomes (22,25). Our data indicate that both hVLCS and hBACS are capable of activating THCA and that both enzymes are found in the endoplasmic reticulum. Because hVLCS is also found in peroxisomes (8,20), it is not clear why Schepers et al. (22) and Solaas et al. (25) did not detect significant THCA activation in peroxisome-enriched liver fractions. Activation of THCA in peroxisomes rather than in microsomes is a more attractive hypothesis because CoA derivatives generally do not cross biological membranes. hVLCS is oriented in the peroxisomal membrane facing the matrix (20,26), and activation of THCA would use intraperoxisomal CoA-SH. Amidation of the ␤-oxidation product, choloyl-CoA, by peroxisomal bile acyl-CoA: amino-acid N-acyltransferase would regenerate free CoA-SH inside the organelle.
Alternatively, if activation of THCA takes place solely in the endoplasmic reticulum, the CoA-SH balance must somehow be maintained. A possible mechanism for maintaining a constant intraperoxisomal CoA-SH pool is one in which THCA-CoA entering the organelle is coupled to CoA-SH exiting. However, it is not clear at this time how either THCA or THCA-CoA enters peroxisomes. Furthermore, it remains to be determined whether the THCA activation reaction is catalyzed primarily by VLCS or by BACS, if this process does not occur in peroxisomes. Data reported by Schepers et al. (22) on pH dependence and Triton X-100 sensitivity indicate that in rat liver microsomes, the enzyme catalyzing THCA activation is distinct from that catalyzing cholate activation. This suggests that VLCS and not BACS might be the primary enzyme that activates THCA. On the other hand, Wheeler et al. (27) purified a bile acid:CoA ligase from rat liver (most likely rat BACS) that activated both chenodeoxycholate and THCA. We plan to address this issue in future studies by investigating THCA activation in subcellular fractions prepared from livers of mice in which the VLCS gene has been disrupted. 2 The data presented here suggest that hVLCS activates THCA more robustly than does hBACS and is not sensitive to inhibition by cholate. These results are consistent with the notion that hVLCS has a diverse role in metabolism, activating both fatty acids and THCA. In contrast, hBACS showed a greater preference for bile acids and their precursors as substrates, suggesting a more limited role for this enzyme (7,12). THCA inhibition of choloyl-CoA synthesis was stronger than cholate inhibition of THCA-CoA synthesis in hBACS-transfected cells. These data are consistent with those reported by Schepers et al. (22) for rat liver microsomes.
The peroxisomal pathway for chain shortening of the C 27 compounds THCA-CoA and DHCA-CoA to C 24 bile acids by ␤-oxidation is now well established (3). In humans, the first step is catalyzed by branched-chain acyl-CoA oxidase, the same enzyme that oxidizes ␣-methyl-branched fatty acids such as pristanic acid (28). In rodents, distinct enzymes carry out these two processes (29). D-bifunctional protein catalyzes the second (enoyl-CoA hydratase) and third (hydroxyacyl-CoA dehydrogenase) activities (30). The thiolase associated with the 58-kDa protein sterol carrier protein X is thought to carry out the final reaction, thiolytic cleavage and release of propionyl-CoA (31). It is noteworthy that the C 24 products of the thiolase reaction are the CoA derivatives of cholic acid and chenodeoxycholic acid and not the free acids. Furthermore, bile acyl-CoA:amino-acid N-acyltransferase, the conjugating enzyme, contains a peroxisomal targeting signal and can be detected in this organelle (23). Thus, peroxisomes possess the complete enzymatic machinery required to convert THCA and DHCA into conjugated bile acids.
In situ hybridization studies are consistent with the proposed roles of VLCS and BACS in bile acid and fatty acid metabolism. Hepatic lipogenic enzymes appear to have a more perivenous distribution (32,33), whereas the rate-limiting step in bile acid synthesis, cholesterol 7␣-hydroxylase, is found mainly in periportal hepatocytes (34). Expression of hBACS was also found to have a more periportal distribution. The enterohepatic circulation returns bile acids to the liver via the portal vein; thus, hBACS is well positioned to efficiently "scavenge" these bile acids, reactivating them for recycling. In contrast, hVLCS, which robustly activates fatty acids as well as THCA, had a more uniform distribution throughout the liver acinus.
The topographic orientation of hBACS, in which both the C and N termini face the cytoplasm, also makes it well suited for its proposed role in reactivation of bile acids returning to the hepatocytes via the enterohepatic circulation. Analysis of the hBACS amino acid sequence by the PHDhtm algorithm revealed that the protein contains no unequivocal transmembrane-spanning domains (35). The orientation of hVLCS in the endoplasmic reticulum membrane, with the N terminus facing the lumen and the C terminus facing the cytoplasm, is consistent with the model proposed by Smith et al. (26). These authors suggested that VLCS is an integral membrane protein in microsomes. Schepers et al. (22) were unable to release THCA-CoA synthetase activity from microsomes using high salt alone, consistent with the notion that VLCS and not BACS is primarily responsible for this synthetase activity.
In summary, in this report, we have described the ability of two members of the VLCS/fatty acid transport protein family to activate bile acids and their precursors. Based on these data, we presented a working hypothesis regarding the roles of these enzymes in bile acid metabolism. Solaas et al. (25) recently hypothesized that the peroxisomal conversion of THCA to choloyl-CoA is the rate-limiting step in production of conjugated bile acids via de novo synthesis. They further proposed that the rate-limiting step in the reutilization of bile acids returning to the liver via the enterohepatic circulation is the reactivation step (25). Future studies will enable us to test these hypotheses.