Bile acid glucuronidation by rat liver microsomes and cDNA-expressed UDP-glucuronosyltransferases.

Four rat UDP-glucuronosyltransferases (UGTs), UGT2B1, UGT2B2, UGT2B3 and UGT2B6, synthesized in COS-7 cells from appropriate cDNA clones were screened for activity towards a range of bile acids, neutral steroids and retinoic acid. For comparison, as well as optimization of enzymatic assays and product identification, rat liver microsomal preparations from Sprague-Dawley, Fischer 344 and phenobarbital-induced Fischer 344 male rats were also used as enzyme sources. Only two of the expressed proteins, UGT2B1 and UGT2B2, were active in bile acid glucuronidation. UGT2B1 exhibited a high substrate specificity for the carboxyl function of bile acids, whereas UGT2B2 demonstrated less specificity, accepting both hydroxyl and carboxyl functions of bile acids. The preferred substrates for both cloned enzymes were mono-hydroxylated bile acids, followed by di-hydroxylated 6-OH compounds. The levels of UGT activity were sufficient to allow for the identification of the biosynthesized products. The data presented here demonstrate that bile acid glucuronidation is carried out, at least in part, by members of the UGT2B subfamily. Similar results have been obtained previously for neutral steroid glucuronidation. UGT2B3 and UGT2B6 was not involved in BA glucuronidation; none of the cloned enzymes was active toward retinoic acid.

Dietary lipids coinduce xenobiotic metabolizing enzymes in rat liver.

We examined the role of dietary lipids in regulating the activities and amounts of epoxide hydrolase, UDP-glucuronosyltransferase and glutathione S-transferase in rat liver. Male Wistar rats were fed a fat-free (FF) diet or isocaloric control diet containing 5% corn oil (CO) or 5% fish oil (FO) for 3 weeks. The activities of these enzymes were approx. 2-fold higher in rats fed the FO diet vs. the FF diet. Intermediate levels of enzyme activity were found in rats fed the CO diet. Diet-induced differences in enzyme levels were shown by immunoblotting. The highest levels of epoxide hydrolase, UDP-glucuronosyltransferase and glutathione S-transferase were detected in rats fed the FO diet. The lowest levels of these enzymes were found in rats fed the FF diet. Intermediate levels of enzyme were found in rats fed the CO diet. Thus, diet-induced differences in enzyme activities were paralleled by changes in enzyme levels. Fatty acid analysis of microsomal lipids showed that the FF diet was associated with decreased levels of n-6 fatty acids vs. the CO diet. The FO diet resulted in increased levels of n-3 fatty acids vs. the other diets.

Characterization of UDP-glucuronic acid transport in rat liver microsomal vesicles with photoaffinity analogs.

The endoplasmic reticulum (ER) of rat liver contains several well characterized UDP-glucuronosyltransferases (UGTs), membrane-bound proteins of 50-54 kDa, and also less well identified UDP-glucosyltransferases, with nucleotide binding sites located on the lumenal surface. There is evidence that the substrates for these enzymes, UDP-glucuronic acid (UDP-GlcUA) and UDP-glucose (UDP-Glc), biosynthesized in the cytosol, are transported into the lumen of the ER via unknown mechanisms, the characteristics of which are poorly defined. A new approach for the study of the transport process has been devised using two active-site directed photoaffinity analogs, [beta-32P]5-azido-UDP-GlcUA and [beta-32P]5-azido-UDP-Glc. Photoincorporation of these probes into the lumenally oriented UGTs of intact rat liver microsomal vesicles was used as an indicator of transport. In intact vesicles, [32P]5N3UDP-GlcUA was efficiently incorporated into UGTs in a time, temperature and concentration dependent manner. In contrast, [32P]5N3UDP-Glc apparently was not transported effectively; maximal photolabeling of the 50-54 kDa proteins by this probe was dependent on detergent disruption of the vesicles. Vesicular uptake of and subsequent photolabeling of the 50-54 kDa proteins by [32P]5N3UDP-GlcUA were inhibited by UDP-GlcUA and 5N3UDP-GlcUA while UDP-Glc, 5N3UDP-Glc, UDP-xylose and UDP-N-acetylglucosamine were less inhibitory, suggesting a high degree of specificity for the uptake/photolabeling process. The anionic transport inhibitors DIDS and SITS inhibited [32P]5N3UDP-GlcUA photoincorporation into UGTs in intact vesicles, but also inhibited photolabeling of these and other enzymes in detergent disrupted vesicles. These data suggest the presence in rat liver microsomal vesicles of a specific, carrier-mediated transport process for UDP-GlcUA which is distinct from the mechanism of UDP-Glc transport.

Biosynthesis and chemical synthesis of carboxyl-linked glucuronide of lithocholic acid.

The glucuronidation of lithocholic acid (LA) by phenobarbital-induced male Fischer 344 rat liver microsomes supplemented with UDP-glucuronic acid was studied. A single radioactive metabolite was formed and its structure was determined by high pressure liquid chromatography/particle beam/mass spectrometry (HPLC/PB/MS), both with and without prior methylation and acetylation of the sample. The reaction product was rigorously identified as the 1-O-acyl-beta-D-glucuronide of LA by comparison with a chemically synthesized standard. The chemical synthesis of the acyl glucuronide of LA was accomplished via a condensation reaction using benzyl 2,3,4-tri-O-benzyl-D-glucopyranuronate. The latter compound was prepared in two steps from benzyl 2,3,4-tri-O-benzyl-1-O-methyl-alpha-D-glucopyranuronate via the 1-O-acetyl derivative. The stereoselective beta coupling of LA with 2,3,4-tri-O-benzyl-D-glucopyranuronate was achieved by the Mitsunobu reaction, in the presence of the free hydroxyl function of LA, using triphenylphosphine and diisopropyl azodicarboxylate in THF followed by preparative TLC. The benzylic ester and ether groups were cleaved by hydrogenation with Pd on charcoal as the catalyst. Positive identification of the glucuronide was established by HPLC/PB/MS and 1H-NMR spectra. No side products formed by acyl migration were detected, but the free acyl glucuronide underwent rapid transesterification in methanol.

Photoaffinity labeling for evaluation of uridinyl analogs as specific inhibitors of rat liver microsomal UDP-glucuronosyltransferases.

The UDP-glucuronosyltransferases (UGT) involved in glucuronidation of endogenous and exogenous toxic compounds transfer the glucuronic acid residue from UDP-glucuronic acid (UDP-GlcUA), to various acceptor groups. A series of compounds that contain N-acyl phenylaminoalcohol derivatives linked to uridine or isopropylideneuridine were tested as UGT inhibitors. The potency of these inhibitors was determined by studying their effect on the photoaffinity labeling of rat liver microsomal UGTs by two photoaffinity probes, [beta-32P]5-azido-UDP-glucuronic acid (5N3UDP-GlcUA) and [beta-32P]5-azido-UDP-glucose (5N3UDP-Glc) and on the enzymatic formation of the two glucuronide conjugates (3-O- and carboxyl-specific) of lithocholic acid. All but one of the compounds tested proved to have an inhibitory effect on UGTs, both in the photoaffinity labeling system and in the enzymatic glucuronidation assay. In the photoaffinity labeling system, the inhibitors containing the isopropylidene moiety were less effective than their unprotected derivatives; however, the protected forms were, with one exception, more potent inhibitors of enzymatic activity. The photoaffinity labeling of UGTs with [beta-32P]5N3UDP-Glc was more susceptible to inhibition by all derivatives than that with [beta-32P]5N3UDP-GlcUA. The effect of one inhibitor, PP50B, on the two enzymatic activities involved in LA glucuronidation was extensively tested. A double-reciprocal plot suggested a competitive inhibition for UDP-GlcUA with an apparent Ki of 35 microM for LA 3-O-glucuronide formation and 94 microM for the carboxyl-linked glucuronide of the same substrate.

Two kinetically-distinct components of UDP-glucuronic acid transport in rat liver endoplasmic reticulum.

Previous studies have documented the presence of protein-mediated transport of UDP-glucuronic acid (UDP-GlcUA) in rat liver endoplasmic reticulum (ER). Measurement of uptake at varying concentrations of high specific activity [beta-32P]UDP-GlcUA has revealed the presence of a two component UDP-GlcUA transporting system. Transport at low substrate concentrations occurred predominantly via a high affinity component (K(m) = 1.6 microM), whereas a low affinity component (K(m) = 38 microM) predominated at high substrate concentrations. The K(m) for the high affinity system is in agreement with that previously published, while the low affinity component is a new finding. The uptake of UDP-GlcUA was temperature-sensitive, time dependent, and saturable for both components. The high affinity transport was affected by trans-stimulation and cis-inhibition by UDP-N-acetylglucosamine (UDP-GlcNAc); however, the same concentrations of UDP-GlcNAc had less effect on the low affinity system. In order to further study the two transport components, various inhibitors of anion transport carriers were tested. The high affinity component was strongly inhibited by 4-acetamido-4'-isothiocyanatostilbene-2,2'-disulfonic acid (SITS) and furosemide, while the low affinity system was less sensitive to these reagents. Dose-dependent inhibition by 4,4'-diisothiocyanatostilbene-2,2'-disulfonic acid (DIDS) was found for both transport systems. Probenecid was found to be a weak inhibitor of both components of the UDP-GlcUA uptake. Finally, the major metabolite of 3'-azido-3'-deoxythymidine, 3'-azido-3'-deoxythymidine monophosphate (AZTMP), was able to inhibit the uptake of UDP-GlcUA by both components. The results indicate the presence of two carrier-mediated UDP-glucuronic acid transporting components in rat liver ER.

The anionic conjugates of bilirubin and bile acids stimulate ATP hydrolysis by S-(dinitrophenyl)glutathione ATPase of human erythrocyte.

These studies demonstrate that bilirubin-ditaurate (an analog of bilirubin-diglucuronide), lithocholic acid 3-O-sulfate, and lithocholic acid 3-O-glucuronide, which are believed to be transported from liver into bile through an active transport process stimulate ATP hydrolysis by purified dinitrophenylglutathione ATPase of human erythrocytes. The Km and Vmax values of the enzyme for these substrates are similar to those for dinitrophenylglutathione indicating the transport mechanisms for bilirubin conjugates, and anionic bile acid-conjugates from hepatocytes to bile and transport of GSH-conjugates from erythrocytes may be mediated by similar mechanisms.

Hepatic metabolism of 3-oxoandrost-4-ene-17 beta-carboxylic acid in the adult rat: formation of carboxyl-linked glucuronides both in vivo and in vitro.

The hepatic metabolism of 3-oxoandrost-4-ene-17 beta-carboxylic acid (etienic acid), a probable acidic catabolite of deoxycorticosterone, was investigated using rats prepared with an external biliary fistula. Metabolic products were identified by GC-MS after hydrolysis with beta-glucuronidase and by proton nuclear magnetic resonance after chromatographic purification of protected glucuronides. About 80% of the injected dose was secreted into bile in 20 hours. Three fully reduced etianic acids (3 alpha-hydroxy-5 alpha-, 3 beta-hydroxy-5 alpha-, 3 alpha-hydroxy-5 beta-androstan-17 beta-carboxylic acids) were identified as were several of their di- and trihydroxylated congeners. Glucuronides of these reduced and/or hydroxylated metabolites constituted over half of the recovered dose, with carboxyl-linked glucuronides predominating over 3-hydroxyl-linked glucuronides. The mode of glucuronidation correlated well with the ability of liver microsomes to form the corresponding compounds in vitro from the set of four 3,5-diastereomeric etianic acids.

Application of photoaffinity labeling with [11,12-3H]all-trans-retinoic acid to characterization of rat liver microsomal UDP-glucuronosyltransferase(s) with activity toward retinoic acid.

[3H]All-trans-retinoic acid has been shown to be an effective photoaffinity label for microsomal UDP-glucuronosyltransferases. Labeling of rat liver microsomal proteins with [3H]all-trans-retinoic acid and [32P]5-azido-UDP-glucuronic acid has shown that at least one protein in the 50-56 kDa mass range encompassing the UDP-glucuronosyltransferases photoincorporated both probes. The fraction of solubilized microsomal protein eluted from a UDP-hexanolamine affinity column with 50 microM UDP-glucuronic acid contained two protein bands, both of which photoincorporated [3H] all-trans-retinoic acid and were detected on Western blot by anti-UDP-glucuronosyltransferase antibodies. Enzymatic glucuronidation activity toward atRA in the same fraction was enriched five-fold over that of native or solubilized microsomes.

The pathogenesis of cholestasis.

How does cholestasis develop when there is no obvious obstruction, destruction, or malformation of bile ducts? Recent findings on bile formation and flow are reviewed, and two models of intrahepatic cholestasis are presented: one of rapid onset, which is bile acid induced, and the other of slow onset, which is estrogen induced. Potential clinical correlations are discussed.

Hyodeoxycholate-6-O-glucuronide cannot be quantitated with 3 alpha-hydroxysteroid dehydrogenase.

The reaction of the 6-hydroxylated bile acid, hyodeoxycholic acid, and its 6-O-glucuronide conjugate with 3 alpha-hydroxysteroid dehydrogenase was examined. A standard end-point assay and determination of the initial rates of reaction showed only minimal activity of the enzyme toward hyodeoxycholate-6-glucuronide in spite of the presence of a free 3 alpha-hydroxyl group. It was established that 6-hydroxylation itself did not significantly affect the enzyme reaction. It is concluded that the 6-glucuronide either blocks or hinders enzyme access to the 3-hydroxyl group.

Formation of three types of glucuronides of 6-hydroxy bile acids by rat liver microsomes.

The glucuronidation of 6-hydroxylated bile acids by rat liver microsomes was studied in vitro; for comparison, several major bile acids lacking a hydroxyl group in position 6 were also investigated. The highest reaction rates were found for lithocholic and deoxycholic acid (10.2 +/- 0.2 and 7.3 +/- 1.4 nmol/mg.min, respectively); our results for these substrates agree well with published values. Glucuronidation rates for the 6 beta-hydroxylated bile acids 3 alpha, 6 beta-dihydroxy-5 beta-cholanoate (murideoxycholate) and 3 alpha, 6 beta, 7 beta-trihydroxy-5 beta-cholanoate (beta-muricholate) were only slightly lower (3.7 +/- 0.3 and 3.6 +/- 0.3 nmol/mg.min). 6 alpha-Hydroxylated bile acids were glucuronidated at rates that were lower than those for their 6 beta-hydroxy counterparts. Rigorous product identification by high-field proton NMR of methyl/acetyl derivatives revealed that while bile acids lacking a 6-hydroxyl group gave rise exclusively to the typical 3-O-glucuronide, the presence of a hydroxyl group in position 6 led to the formation, in ratios depending on the substrate, of three types of conjugate: the 3-O-, the 6-O-, and the carboxyl-linked (acyl-) glucurnide. The latter is the first example of an acyl glucuronide of a bile acid of conventional (C24) size.

Hepatic metabolism of short-chain bile acids. Inversion of the 3-hydroxyl group of isoetianic acid (3 beta-hydroxy-5 beta-androstane-17 beta-carboxylic acid) by the adult rat.

The stereospecificity of mechanisms for hepatic transport of short-chain bile acids has been examined by following the hepatic metabolism and biliary secretion of 3 beta-hydroxy-5 beta-androstane-17 beta-carboxylic acid (isoetianic acid) administered in two different labeled forms to rats prepared with an external biliary fistula. While 93% of the administered [2,2,4,4-3H]isoetianic acid was recovered in bile after 20 h, only 18% of a similar dose of [3 alpha-3H]isoetianic acid was secreted in bile over the same time period. The recovered radioactivity of the latter compound was mainly associated with bile water. With the [2,2,4,4-3H]isoetianic acid, the bulk of the biliary isotope was determined to be in the form of two glucuronide conjugates. Spectral analysis identified these metabolites as the hydroxyl-linked (major) and carboxyl-linked (minor) beta-glucuronides, not of the 3 beta-hydroxy compound administered, but of 3 alpha-hydroxy-5 beta-androstane-17 beta-carboxylic acid (etianic acid), i.e., the products of hydroxyl group inversion. It is concluded that isoetianic acid is efficiently cleared from plasma and conjugated with glucuronic acid after its epimerization to etianic acid. The prevalent, but not complete, loss of the 3-tritium atom and the retention of the 2- and 4-tritium atoms probably indicates a 3-oxo-5 beta-androstane-17 beta-carboxylic acid intermediate with partial return of the label via a limited labeled pool of reduced nicotinamide cofactor.

Distinct forms of cytochrome P-450 are responsible for 6 beta-hydroxylation of bile acids and of neutral steroids.

Cytochrome P-450-dependent 6 beta-hydroxylation of bile acids in rat liver contributes to the synthesis of the quantitatively important pool of 6-hydroxylated bile acids, as well as to the detoxification of hydrophobic bile acids. The lithocholic acid 6 beta-hydroxylation reaction was investigated and compared with androstenedione 6 beta-hydroxylation. Differential responses of these two activities to inducers and inhibitors of microsomal P-450 enzymes, lack of mutual inhibition by the two substrates and differential inhibition by antibodies raised against several purified hepatic cytochromes P-450 were observed. From these results it was concluded that 6 beta-hydroxylation of lithocholic acid is catalysed by P-450 form(s) different from the subfamily IIIA cytochromes P-450 which are responsible for the bulk of microsomal androstenedione 6 beta-hydroxylation. Similar, but more tentative, results revealed that the 7 alpha-hydroxylation of lithocholic acid and of androstenedione may be also catalysed by distinct P-450 enzymes. The results indicate that cytochromes P-450 hydroxylating bile acids are distinct from analogous enzymes that carry out reactions of the same regio- and stereo-specificity on neutral steroids (steroid hormones). A comparison of pairs of cytochromes P-450 that catalyse the same reaction on closely related steroid molecules will help to define those structural elements in the proteins that determine the recognition of their respective substrates.

Urinary excretion of lithocholic acid and its conjugates by the bile duct-ligated rat.

The 3-O-glucuronide of lithocholic acid has been shown to be a potent cholestatic agent in rats. However, even after the onset of lithocholic acid glucuronide-induced cholestasis, little of the administered material was recovered in urine. To determine whether this phenomenon was related to the steroid moiety or the form of conjugation, small doses of radiolabeled lithocholic acid glucuronide, lithocholic acid, taurolithocholic acid and/or lithocholic acid sulfate were administered to rats with ligated bile ducts. Urinary excretion of isotope was followed for 24 hr and urinary metabolites of the administered compounds were identified by thin-layer chromatography. Lithocholic and taurolithocholic acids were slowly but relatively efficiently excreted in urine with 73% and 91% of the dose, respectively, recovered in urine over 24 hr. More than 80% of the label in urine from animals receiving these two compounds was in the form of taurine-conjugated beta-muricholic acid. In contrast, lithocholic acid 3-glucuronide and 3-sulfate were poorly excreted: 9% and 12% of the administered doses, respectively, were recovered in urine in 24 hr. Of the small amount of label in urine from rats given the glucuronide, 90% was identified as lithocholic and taurolithocholic acid glucuronides. When lithocholic acid sulfate was given, thin-layer chromatography of urine showed two peaks, which were tentatively identified as tauromurideoxycholic and taurolithocholic acid sulfates. More definitive identification was not possible because of the small amount of the administered dose excreted in urine in these forms.(ABSTRACT TRUNCATED AT 250 WORDS)

Membrane translocation and regulation of uridine diphosphate-glucuronic acid uptake in rat liver microsomal vesicles.

BACKGROUND/AIMS: Hepatic glucuronidation is quantitatively the most important conjugation reaction by which an array of endogenous compounds and xenobiotics undergo biotransformation and detoxification. The active site of the uridine diphosphate (UDP) glucuronosyltransferases, which catalyze glucuronidation reactions, has been postulated to reside in the lumen of the endoplasmic reticulum. The aim of this study was to characterize the process whereby UDP glucuronic acid (UDP-GlcUA), the cosubstrate for all glucuronidation reactions, is transported into microsomal vesicles. METHODS: The uptake process was analyzed using rapid filtration techniques, radiolabeled UDP-GlcUA, and rat liver microsomes. RESULTS: Uptake was saturable with respect to time and concentration, inhibited by 4,4'-diisothiocyanato-stilbene-2,2'-disulfonic acid and 4-acetamido-4'-isothio-cyanatostilbene-2-2'-disulfonic acid, and was osmotically sensitive. Transport was stimulated by Mg2+ and guanosine triphosphate (50 mumol/L) but not guanosine 5'-O-(3-thiotriphosphate) or adenosine triphosphate. Luminal UDP-N-acetylglucosamine (1 mmol/L) produced enhanced uptake of UDP-GlcUA (trans stimulation). In contrast to nucleotide sugar transport in the Golgi apparatus, trans uridine monophosphate and UDP did not alter UDP-GlcUA transport in microsomes, indicating distinct processes. CONCLUSIONS: These data provide unambiguous evidence for the existence of a unique, substrate-specific, regulated, carrier-mediated process that transports UDP-GlcUA into the lumen of hepatocyte microsomes. This transporter may regulate glucuronidation in vivo.

Characterization of human liver microsomal UDP-glycosyltransferases using photoaffinity analogs.

The photoaffinity analogs [beta-32P]5-azido-UDP-glucuronic acid ([32P]5N3UDP-GlcUA) and [beta-32P]5-azido-UDP-glucose ([32P]5N3UDP-Glc) were used to characterize UDP-glycosyltransferases of microsomes prepared from human liver. Photoincorporation of both probes into proteins in the 50- to 56-kdalton range, known to contain UDP-glucuronosyl transferases (UGTs), was concentration dependent, and photolabeled proteins were susceptible to trypsin digestion only in the presence of detergent. The latter was demonstrated by the appearance on Western blots of the trypsin-treated, detergent-disrupted microsomes of a protein band of slightly lower molecular mass than, and presumably derived from, the UGTs. However, a labeled cleavage product was found only in samples photolabeled with [32P]5N3UDP-GlcUA and not in those labeled with [32P]5N3UDP-Glc. In detergent-treated microsomes, all of the nucleotide sugars that were tested protected better against photoinsertion of [32P]5N3UDP-GlcUA than of [32P]5N3UDP-Glc, with UDP-glucose being the most effective, followed by UDP-GlcUA and UDP-galactose. The pattern of inhibition of a series of uridinyl analogs toward photolabeling by the two probes was quite different: one inhibitor that was ineffective in blocking photoincorporation of [32P]5N3UDP-GlcUA (L-DPASiU) was one of the most potent inhibitors of photolabeling with [32P]5N3UDP-Glc. A similar dichotomy was seen with several inhibitors in enzymatic assays measuring hyodeoxycholic acid 6-O glucuronidation and glucosidation activities; the most potent inhibitors of HDCA glucosidation were not as effective against glucuronidation. The results indicate a lumenal orientation for human microsomal UGTs and provide substantial evidence that two distinct enzyme systems are involved in 6-O glucuronidation and 6-O glucosidation of HDCA.

Metabolism of lithocholic acid in the rat: formation of lithocholic acid 3-O-glucuronide in vivo.

Milligram amounts of [3 beta-3H]lithocholic (3 alpha-hydroxy-5 beta-cholanoic) acid were administered by intravenous infusion to rats prepared with a biliary fistula. Analysis of sequential bile samples by thin-layer chromatography (TLC) demonstrated that lithocholic acid glucuronide was present in bile throughout the course of the experiments and that its secretion rate paralleled that of total isotope secretion. Initial confirmation of the identity of this metabolite was obtained by the recovery of labeled lithocholic acid after beta-glucuronidase hydrolysis of bile samples. For detailed analysis of biliary metabolites of [3H]lithocholic acid, pooled bile samples from infused rats were subjected to reversed-phase chromatography and four major labeled peaks were isolated. After complete deconjugation, the two major compounds in the combined first two peaks were identified as murideoxycholic (3 alpha, 6 beta-dihydroxy-5 beta-cholanoic) and beta-muricholic (3 alpha, 6 beta, 7 beta-trihydroxy-5 beta-cholanoic) acids and the third peak was identified as taurolithocholic acid. The major component of the fourth peak, after isolation, derivatization (to the methyl ester acetate), and purification by high pressure liquid chromatography (HPLC), was positively identified by proton nuclear magnetic resonance as lithocholic acid 3 alpha-O-(beta-D-glucuronide). These studies have shown, for the first time, that lithocholic acid glucuronide is a product of in vivo hepatic metabolism of lithocholic acid in the rat.

Glucuronidation of monohydroxylated short chain bile acids by human liver microsomes and purified human liver UDP-glucuronosyltransferases.

The glucuronidation of monohydroxylated bile acids and their analogs with a shortened side chain (short-chain bile acids) by human liver microsomes and by two UDP-glucuronosyltransferases purified therefrom has been studied in vitro. In microsomes, all 18 substrates tested underwent glucuronidation; the rate of reaction and the site of attachment of glucuronic acid (hydroxyl group in position 3, side chain carboxyl group, or various ratios of both products) were strongly dependent on the length of the side chain, the configuration of the 3-hydroxyl group, and the configuration of the A/B ring junction (5 alpha-H/5 beta-H). Two UDP-glucuronosyltransferases (UDPGTs) purified from human microsomes, designated "pl 7.4" and "pl 6.2" according to their behavior in chromatofocusing, accounted for the formation of hydroxyl-linked glucuronides of a different pair of bile acids each. The pl 7.4 human liver UDPGT catalyzed the glucuronidation of C20 and C22 with the 3 alpha-OH, 5 beta-H configuration, while the pl 6.2 human liver UDPGT catalyzed the glucuronidation of either 3-OH epimer of C21 and C24 acids with the 5 alpha-H configuration. The enzymes displayed a relatively high selectivity in that they did not accept any of the remaining 14 bile acids as substrates; none of the enzymes led to the formation of a carboxyl-linked glucuronide. In addition, purified human liver UDPGT did not catalyze the glucuronidation of cholate, deoxycholate, chenodeoxycholate, ursodeoxycholate, lithocholate, hyocholate, hyodeoxycholate, bilirubin, morphine, or 4-hydroxybiphenyl. The above results suggest that several bile-acid UDP-glucuronosyltransferases of high specificity exist in human liver.