Purification and Characterization of Hamster Liver Microsomal 7α-Hydroxycholesterol Dehydrogenase

While studying the bile acid synthetic pathway of hamsters, we discovered an NADP+-dependent liver microsomal 7α-hydroxycholesterol dehydrogenase (7α-HCD) activity that was not observed in rat liver microsomal fractions. The hamster liver microsomal 7α-HCD was purified to homogeneity using 2′,5′-ADP and cholic acid-agarose affinity chromatography. 7α-HCD displayed a molecular weight of approximately 34,000 on SDS-polyacrylamide gel electrophoresis; it is an intrinsic membrane protein of the hamster liver endoplasmic reticulum and exists as a multimeric aggregate in pure form. Partial N-terminal amino acid sequence analysis showed that 7α-HCD had high sequence similarity to human 11β-hydroxysteroid dehydrogenase (11β-HSD; 24/30 amino acid identity). TheK m values for corticosterone and 7α-hydroxycholesterol were 1.2 and 1.9 μm, respectively, for purified 7α-HCD; both reactions displayed identicalV max values (approximately 170 nmol/min/mg of protein). The IC50 of carbenoxolone, a competitive inhibitor of 11β-HSD, was 75 nm for 7α-hydroxycholesterol dehydrogenation and 210 nm for corticosterone dehydrogenation. The tissue-specific expression in hamster was as follows: adrenal ≥ liver > kidney > testis ≫ brain > lung. Microsomal 7α-HCD is uniquely expressed in hamster liver and to some extent in human liver but not in rat liver. Western blot analysis with two antibodies elicited against an N-terminal peptide of the human 11β-HSD and purified hamster liver 7α-HCD, respectively, suggested the presence of multiple forms of 7α-HCD in hamster liver, most likely due to the existence of a family of 11β-HSD proteins. Since 7-oxocholesterol is a potent inhibitor of cholesterol 7α-hydroxylase, alternative mechanisms for regulation of bile acid synthesis may exist in human and hamster liver due to production of this metabolite and its potential as an oxysterol.

The first regulated step in bile acid synthesis is hydroxylation of cholesterol to 7␣-hydroxycholesterol by cholesterol 7␣hydroxylase, the apparent rate limiting reaction in this pathway (1). Oxidation of the 3␤-ol group to an oxo group and rearrangement of the unsaturated bond at ⌬ 5 -position to ⌬ 4 followed by hydroxylation at either the C-12 or C-27 position converts the sterol to its active form essential for digestion of fatty nutrients. Recently, an alternate pathway to bile acids has been elucidated involving first 27-hydroxylation of cholesterol and subsequent oxidation at position C-3 and hydroxylation at position C-7 eventually leading to chenodeoxycholic acid (2).
A unique hamster liver microsomal 7␣-hydroxycholesterol dehydrogenase (7␣-HCD) 1 was discovered in our laboratory during the development of an HPLC method for assay of cholesterol 7␣-hydroxylase activity (3,4). The oxidized metabolite produced by 7␣-HCD, 7-oxocholesterol, is a bioactive sterol and potent competitive inhibitor of cholesterol 7␣-hydroxylase activity (approximately K i for 7-oxocholesterol of 7 M versus approximately K m for exogenous cholesterol of 100 M; Ref. 5). Although oxycholesterol derivatives are formed during chemical autoxidation of cholesterol, the presence of 7-oxo-bile acids observed in the blood of infants suggests that a unique metabolic pathway leading to 7-oxo-bile acids may exist in humans and other mammals (6). Individuals with the rare inherited disease cerebrotendinous xanthomatosis who lack cholesterol 27-hydroxylase activity display significant levels of 7-oxocholesterol in their blood (7). In addition, chronic salt loading of hypertensive baboons (8) or rhesus monkeys with hemorrhagic fever (9) also result in significant blood levels of 7-oxocholesterol, in addition to 7␣-and 7␤-hydroxycholesterol. Interestingly, intravenous infusion of 7-oxocholesterol into rats decreased bile acid secretion and increased cholesterol 7␣hydroxylase enzyme activity, protein content, and mRNA levels (10) to compensate for the lack of bile acid production. Therefore, hepatic 7␣-HCD and its products may participate in the regulation of bile acid synthesis.
A number of genes encoding proteins that oxidize hydroxyl groups of sterols have been identified to date (11,12). For example, unique 3␤-, 11␤-, and 17␤-hydroxysterol dehydrogenases have been described; each possesses a unique stereochemical reaction, pyridine nucleotide specificity, and tissuespecific localization. These short chain alcohol dehydrogenases found in the liver are thought to be involved in either the processing or termination of function of various bioactive sterols, such as bile acids, dehydroepiandrosterone, 5-ene-androsten-3␤,17␤-diol, androstenedione, cortisol, corticosterone, 17␣-hydroxypregnenolone, pregnenolone, and progesterone. However, no enzyme with 7␣-hydroxysterol dehydrogenase activity has been described to date. Since 7-oxocholesterol apparently could be formed by an enzyme-mediated process, we purified this enzyme to homogeneity from hamster liver to further characterize it. Using the purified enzyme, we analyzed the N-terminal amino acid sequence of the protein, studied its substrate specificity, and elicited a polycolonal antibody to assess the tissue-and speciesspecific expression of 7␣-HCD. Our results suggest that this enzyme, which exhibits high substrate specificity for corticosterone, cortisol, and 7␣-hydroxycholesterol in the hamster liver (and possibly humans), may be a member of the 11␤-hydroxysteroid dehydrogenase family. However, either it is not expressed in rat liver, or the rat 11␤-HSD does not catalyze oxidation of 7␣-or 7␤-hydroxylated sterols.

EXPERIMENTAL PROCEDURES
Materials-7␣-Hydroxycholesterol and 7-oxocholesterol were obtained from Steraloids Inc. (Wilton, NH). DEAE-cellulose (DE52) was obtained from Whatman. 2Ј,5Ј-ADP-agarose, cholic acid-agarose, dodecyl nonaoxyethylene glycol monoester (C 12 E 9 ), and other chemicals were purchased from Sigma in analytical or HPLC grade. An antibody elicited against an N-terminal peptide of human 11␤-hydroxysteroid dehydrogenase was obtained from Ian Mason (Department of Clinical Biochemistry, University of Edinburgh, Edinburgh, Scotland). Human liver microsomal fractions were obtained from F. P. Guengerich (Department of Biochemistry, Vanderbilt University School of Medicine, Nashville, TN).
Animals-Male Syrian hamsters (Charles River Breeding Laboratories, Inc., Wilmington, MA; Lak:LVG(SYR), 60 -90 g) and male Sprague-Dawley rats (Harlan Sprague-Dawley, Indianapolis, IN; Hsd:Sprague-Dawley, 180 -250 g) were fed laboratory chow ad libitum. Microsomal fractions were prepared from hamster, rat, human liver, and other tissues as described previously (13) and assayed immediately. Liver microsomal fractions in 10% glycerol were stored at Ϫ70°C for up to 3 months without loss of activity. Protein concentration was determined by the method of Lowry et al. (14) using bovine serum albumin as the standard. Male New Zealand White rabbits (Myrtles, Inc., Thompson Station, TN; 1-2 kg) were fed laboratory chow ad libitum during antibody production.
HPLC Assay of 7␣-Hydroxycholesterol Dehydrogenase Activity-Microsomal protein fractions were incubated at 37°C for 5 min with 0.1 M potassium phosphate buffer, pH 7.4, containing 0.1 mM EDTA and 20 mM cysteamine-HCl in a final volume of 1.0 ml. After the addition of NADP ϩ (final concentration 0.5 mM), 7␣-hydroxycholesterol was added in 2-propanol to the reaction mixture at a final concentration of 60 M to initiate the reaction. The solvent concentration was always less than 1% 2-propanol and had no effect on the enzyme activity. The reaction was terminated by adding an equal volume of methanol, and the reaction mixture was extracted three times with 5 ml of petroleum ether. The extract was analyzed by normal phase HPLC as described previously (3).
Fluorometric Assay of 7␣-Hydroxycholesterol Dehydrogenase Activity-Purified enzyme preparations were incubated at 37°C in a final volume of 2 ml with the same buffer used in the HPLC assay method. The reaction was initiated by adding exogenous 7␣-hydroxycholesterol or other steroids in 2-propanol. Enzyme activity was monitored by measuring the production of NADPH at excitation ϭ 340 nm and emission ϭ 460 nm for 2 min using a model 50B luminescence spectrometer LS 50B (Perkin-Elmer). The reaction mixtures containing NADP ϩ and 7␣-hydroxycholesterol alone were first monitored as control rates. Reaction rates were calculated before and after the addition of purified 7␣-HCD to give enzymatic rates of NADP ϩ reduction. For the rate measurements to obtain Michaelis-Menten parameters, the results were analyzed by nonlinear regression analysis of data obtained from triplicate assays at five different substrate concentrations.
Purification of 7␣-Hydroxycholesterol Dehydrogenase-The microsomal fraction from hamster liver was suspended at 6.3 mg/ml in 20 mM potassium phosphate buffer, pH 7.4, containing 10% (v/v) glycerol and 0.5 mM EDTA (elution buffer). The detergent C 12 E 9 was then added to obtain a final concentration of 0.3%. The suspension was stirred for 20 min and insoluble protein was sedimented at 135,000 ϫ g for 20 min. Polyethylene glycol 8000 (PEG) was added to the supernatant to attain a final concentration of 8%, the suspension was stirred for 10 min, and the mixture was sedimented at 10,000 ϫ g for 10 min. The supernatant was dialyzed with the original buffer overnight and then applied to a DE52-cellulose column (1 ϫ 22 cm) equilibrated with same buffer used for solubilization of protein from the microsomal membrane. The active fractions that were not retained by the DE52 matrix were pooled and applied to a 2Ј,5Ј-ADP-agarose column (1 ϫ 6 mm). After the enzyme bound to the column, the column was eluted first by 20 column volumes of 50 mM NaCl in elution buffer and then by gradient elution from 0 to 0.5 mM NAD ϩ in 10 column volumes of elution buffer to remove proteins not bound by the 2Ј,5Ј-ADP matrix with high affinity. Finally, the 7␣-HCD activity was eluted by gradient elution with 0 -0.2 mM NADP ϩ in 10 column volumes of elution buffer containing 0.03% C 12 E 9 . The active fractions from 2Ј,5Ј-ADP affinity column were pooled and applied to a cholic acid-agarose affinity column (1 ϫ 6 mm). 7␣-HCD activity was eluted by a cholic acid gradient from 0.2 to 0.8% (w/v) in elution buffer consisting of 20 mM potassium phosphate buffer, pH 7.4, containing 0.5 mM EDTA, 10% (v/v) glycerol, and 0.03% (v/v) C 12 E 9 . All procedures were performed at 4°C.
The purity of each fraction was determined by 12% SDS-PAGE using a Bio-Rad Mini Protein II apparatus. After fixing the protein with 10% ethanol and 5% acetic acid, the gel was soaked in a solution containing 3.4 mM K 2 Cr 2 O 7 and 3.2 mM HNO 3 , followed by staining with 12 mM silver nitrate. Protein bands were visualized with 0.28 M sodium carbonate and 1.85% paraformaldehyde.
Partial N-terminal Amino Acid Sequence Analysis-Purified enzyme from cholic acid affinity chromatography was subjected to 12% SDS-PAGE and transferred to a polyvinylidene difluoride membrane using transfer buffer, pH 11, containing 10 mM CAPS and 10% (v/v) methanol. Transferred protein was stained with 0.2% Coomassie Blue followed by destaining with 30% methanol and 10% acetic acid. The proteins of interest were manually cut from the membrane and sequenced using an Applied Biosystems 470A gas phase protein sequencer equipped with a 120A phenylthiohydantoin analyzer (Applied Biosystems, Foster City, CA) in the Protein Chemistry Core Laboratory (Department of Biochemistry and Molecular Biology, University of Louisville). The repetitive yield at each cycle was approximately 94%, and the protein sequencing was replicated five times with nearly perfect identity of sequence.
Production of Antibody-Purified enzyme preparations from the 2Ј,5Ј-ADP-agarose affinity column were subjected to 12% SDS-PAGE. After staining with 0.1% Coomassie Blue dye, the protein of interest was manually cut from the gel and extracted with 0.1 M Tris buffer, pH 8.0, containing 0.1% SDS. After dialysis against 5 mM NH 4 CO 3 and 0.05% SDS, the lyophilized protein was precipitated with 90% acetone containing 1 mM HCl, and the pellet was rinsed with chilled acetone prior to air drying. Antigen was mixed with RIBI emulsion (RIBI Immunochem Research, Inc., Hamilton, MT) and injected into a rabbit at multiple sites on the back. A second inoculation at 14 days was performed using the same amount of antigen mixed with 0.7 ml of Freund's Incomplete Adjuvant (Life Technologies, Inc.). Blood was obtained biweekly to determine specific antibody content. After clotting, the serum was heated and stored in the presence of 50 M phenylmethylsulfonyl fluoride at Ϫ70°C. Antibody against 7␣-HCD protein was detected by an enzyme-linked immunosorbent assay method.
Western Blot Analysis-Purified enzyme or microsomal protein samples were subjected to 12% SDS-PAGE. After transfer of protein to a nitrocellulose membrane, the membrane was blocked in a solution of 20 mM Tris-HCl buffer, pH 7.5, containing 0.5 M NaCl and 0.5% Tween 20 (0.1% TTBS) and washed with 0.5% TTBS. The membrane was incubated with a 1:2000 to 1:4000 dilution of the primary antibody, and after several washes the membrane was incubated with a 1:5000 dilution of second goat anti-rabbit globulin conjugated with horseradish peroxidase. Protein-immunoglobulin complex formation was detected by enhanced chemiluminescent light emission of luminol oxidation by horseradish peroxidase using Renaissance Western blotting chemiluminescence reagents (DuPont).

RESULTS
Purification of 7␣-Hydroxycholesterol Dehydrogenase-Since we noted that an abundant source of enzyme activity was the hamster liver microsomal protein fraction (4), we embarked upon the purification of the enzyme responsible for 7␣-HCD activity from that tissue fraction. We ascertained the concentration-dependent action for several detergents in solubilizing 7␣-HCD from microsomal membrane. Solubilization with C 12 E 9 yielded the highest amount of 7␣-HCD activity released among the detergents tested (data not shown). Polyethylene glycol 8000 (PEG, 8%) was chosen to precipitate nonspecific protein, since higher concentrations of PEG did not precipitate additional protein and interfered with the binding of 7␣-HCD in subsequent chromatography steps. 7␣-HCD was soluble at all PEG concentrations tested, while other proteins such as NADPH:cytochrome P450 oxidoreductase, cytochrome b 5 , and cytochrome P450 were precipitated at 8% PEG.
DE52 column chromatography was used to separate tightly bound proteins from 7␣-HCD (data not shown). 7␣-HCD activity eluted quickly from the column with a slightly browncolored protein; i.e. DE52 matrix did not bind 7␣-HCD activity, but a large amount of unrelated protein was retained on the column. Since 7␣-HCD was potently inhibited by 2Ј-AMP (4), we subsequently utilized 2Ј,5Ј-ADP affinity chromatography with a NADP ϩ gradient to elute the enzyme activity of interest. To remove minor contaminates, the eluate from the 2Ј,5Ј-ADP affinity column containing 7␣-HCD was subjected to cholic acid-agarose affinity chromatography. Table I shows the purification steps leading to nearly homogeneous protein, and Fig.  1 shows the analysis of the various purified fractions by 12% SDS-PAGE. There were always two proteins in the active fractions, which migrated as approximately 34-kDa proteins.
Determination of the N-terminal Sequence of Purified 7␣-HCD-In order to determine the N-terminal sequence of 7␣-HCD, purified 7␣-HCD prepared from the cholic acid affinity column was transferred to a polyvinylidene difluoride membrane and sequenced as described under "Experimental Procedures." Both the upper and lower bands shown in Fig. 1 were sequenced; the results are shown in Table II. These sequences have high sequence similarity to human type I 11␤-hydroxysteroid dehydrogenase (24/30 identical amino acid residues in the N termini), indicating that 7␣-HCD may belong to the 11␤-HSD family. Using various proteinases, we generated a number of oligopeptides, but all contained a sequence identical to the N-terminal portion of the larger protein (data not shown) Tissue-specific Expression of 7␣-HCD Enzyme Protein Correlates with 7␣-HCD Activity in Hamsters-A specific polyclonal antibody against 7␣-HCD was elicited in male New Zealand White rabbits as described under "Experimental Procedures." Tissue-specific 7␣-HCD expression was studied by Western blot analysis of hamster adrenal, liver, kidney, testis, brain, and lung microsomal fractions using the antiserum described above. The results are shown in Fig. 2. To confirm that the 34-kDa protein is indeed 7␣-HCD, the levels of expressed 7␣-HCD protein content were compared with 7␣-HCD activity measured by HPLC. The protein content determined by Western blotting correlated well with enzyme activity, supporting the assumption that the 34-kDa protein band(s) in the SDS-PAGE is/are indeed the active 7␣-HCD. The levels of expression of 7␣-HCD were as follows: adrenal Ն liver Ͼ kidney Ͼ testis Ͼ Ͼ brain Ͼ lung. Since the 7␣-HCD N-terminal amino acid sequence was similar to human 11␤-HSD, the tissuespecific expression of 7␣-HCD was compared with that of 11␤-HSD. A polyclonal antibody against a conserved N-terminal peptide of human 11␤-HSD that is closely related to those of 11␤-HSD from different species was used in a parallel experiment. The results obtained with this antibody demonstrated that the tissue-specific expression patterns of 11␤-HSD and 7␣-HCD are similar (Fig. 2). However, in hamster liver there were two bands detected by both antibodies. The apparent content of each band was different for the two antibodies against either 7␣-HCD or 11␤-HSD, suggesting that 7␣-HCD may be immunochemically related to human 11␤-HSD and that it is apparently expressed at higher levels in the liver than in the adrenal of hamster. N-terminal sequence analysis of the two bands showed that both protein species had identical N termini, also supporting the hypothesis that two forms of 11␤-HSD may exist in hamster liver microsomes.
7␣-HCD Is Expressed in Hamster and Human but Not Rat Liver-Expression of 7␣-HCD was also examined in a human hepatic tumor cell line and livers of hamster, human, and rat using the same strategy employed in the tissue-specific expression study. Compared with 11␤-HSD, 7␣-HCD was highly expressed in hamster and expressed to some extent in four human samples but not expressed in rat liver microsomal protein (Fig. 3). In the HepG2 human hepatoma cell line, an immuno-  Microsomal protein of each tissue (0.3 mg) was used to determine enzyme activity using the normal phase HPLC method described under "Experimental Procedures." The results are the average of duplicate assays. Microsomal protein from each tissue (20 g) was subjected to 12% SDS-PAGE, and Western blot analysis was performed as described under "Experimental Procedures" using antibodies to human type I 11␤-HSD and hamster 7␣-HCD. reactive protein with a slightly larger apparent molecular weight was detected. These data suggest that 7␣-HCD may be uniquely expressed in hamster and possibly human liver. The level of expressed 7␣-HCD protein in human liver did not parallel the levels of enzyme activity. However, the human liver microsomal preparations used in this study were prepared in 1992, and the age of the sample may account for the low enzyme activity compared with the levels of expressed 7␣-HCD protein. Rat liver microsomal fractions contain low levels of 7␣-HCD activity and had little or no 7␣-HCD-like immunoreactive protein, suggesting either that rat liver may not contain 7␣-HCD or that it may contain an alternative form of the enzyme. These results demonstrate that 7␣-HCD is immunochemically related to human 11␤-HSD and that it may be a member of the classical 11␤-HSD Type I (hepatocyte-specific isozyme) family.
Characterization of Purified 7␣-Hydroxycholesterol Dehydrogenase-The absolute absorbance spectrum of the purified enzyme fraction (240 -700 nm) revealed a single wavelength maximum at 278 nm (data not shown), indicating that the enzyme lacks chromophores, such as flavin or heme. When the purified 7␣-HCD from 2Ј,5Ј-ADP-agarose affinity chromatography was analyzed by 4 -30% nondenaturing gradient PAGE, there were two protein bands corresponding to apparent molecular masses of 440 and 240 kDa, respectively (data not shown). These results indicate that 7␣-HCD exists as aggregates in the absence of detergent, a characteristic of other intrinsic membrane proteins. Since no reverse reaction could be measured under any conditions employed (data not shown), purified 7␣-HCD apparently could not readily catalyze the reduction of oxo groups at either position 7 or 11 of C 27 or C 21 steroids, respectively. This conclusion was confirmed for 7␣-hydroxy-and 7-oxocholesterol by HPLC analysis in a separate experiment at an NADP ϩ /NADPH concentration of 500 and 50 M, respectively (data not shown).
Substrate Specificity-Since the literature suggests that 7-oxocholesterol is formed nonenzymatically (15), the goals of our studies were to determine which hydroxysterols are substrates and whether 7␣-hydroxycholesterol may be an endogenous substrate of 7␣-HCD. The purified enzyme obtained from 2Ј,5Ј-ADP affinity chromatography was used to perform substrate specificity studies, since the enzyme obtained from cholic acid affinity column chromatography displayed lower specific activity. Since 7␣-HCD uses NADP ϩ as a cofactor to oxidize the hydroxyl group to an oxo group in the sterol substrate, we utilized a fluorometric method to monitor the formation of NADPH to facilitate these studies; with pure protein, the rate of 7-oxocholesterol production was stoichiometric with NAD ϩ reduction (data not shown).
In this study, the concentrations of steroids and nucleotides utilized were 50 and 500 M, respectively. All of the concentrations were above the concentrations of endogenous substrate normally required for maximal hydroxysteroid dehydrogenase activity (11,12). As shown in Table III, the first three steroids were C 27 sterol derivatives. Among them, 7␣-hydroxycholesterol gave the highest rate of oxidation. However, the oxidation rate for 7␤-hydroxycholesterol was almost 70% of the rate for 7␣-hydroxycholesterol. Interestingly, cholesterol also could be oxidized, i.e. at approximately 15% of the rate seen for 7␣hydroxycholesterol. The major product observed by HPLC comigrated with authentic 3-oxocholesterol (data not shown). These data suggest that some C 27 hydroxysteroids are substrates for 7␣-HCD and that oxidation at the 7-position is not absolutely stereospecific.
The second class of compounds studied was C 24 bile acids. For these molecules, the hydroxyl group at the 3␣-position alone was oxidized at a rate similar to that for cholesterol. A substrate with a hydroxyl group at the 7␣-position was only about 1 ⁄4 as good as a substrate containing only a 3␣-hydroxyl group, such as cholesterol or lithocholate. If the substrate had a hydroxyl group at the 12␣-position, the rate was only about 1 ⁄12 that of a 3␣-hydroxyl group alone. No oxidation was de- 7␣-Hydroxycholesterol to 7␣-olcholestene-3-one The results shown are means of duplicate assays. b The results were obtained or confirmed by the normal phase HPLC method.
c ND, not detectable (Ͻ0.01 nmol of product produced/min/mg of protein). tected when the substrate contained a hydroxyl group at the 6␣-position. Substrates with two hydroxyl groups at the 7␣and 12␣-positions were not significantly oxidized. These data suggest that 7␣-HCD poorly oxidizes the 3␣-hydroxyl group of bile acid derivatives; hydroxyl groups at the 6␣-, 7␣-, and 12␣positions cannot be oxidized and also diminish dehydrogenation at the 3␣-position of bile acids.
The third group of compounds tested were C 21 and C 19 steroid hormones. Hydrocortisone is the natural substrate of 11␤-HSD in humans, and corticosterone is the natural substrate of 11␤-HSD in rodents. Since 7␣-HCD was obtained from hamster liver, it was not surprising that 7␣-HCD oxidized corticosterone more rapidly than hydrocortisone (Table III). The enzyme can also oxidize 11-dehydrocorticosterone-21-ol to some extent but could not oxidize 11-dehydrocorticosterone-17-ol. This indicates that oxidation is position-specific for additional hydroxyl groups on 11-ol derivatives. The fact that 4-pregnene-11␣-ol-3,20-dione was not oxidized indicates that oxidation at carbon 11 is also stereospecific (i.e. 11␣ versus 11␤ configuration). These data strongly suggest that 7␣-HCD is a member of the 11␤-HSD family. Compared with cholesterol and lithocholic acid, dehydroepiandrosterone, a 3␤-hydroxy C 19 sterol, was oxidized at a very low rate, demonstrating again that 7␣-HCD cannot easily oxidize hydroxyl groups at the 3␤-position for C 19 steroids as is the case for 3␤-HSD.
The last group of compounds shown in Table III normally are metabolized by dehydrogenases that utilize NAD ϩ instead of NADP ϩ to oxidize sterol substrate. Dehydroepiandrosterone is the natural substrate for 3␤-HSD, while 5-androstene-3␤,17␤diol is the natural substrate for 17␤-HSD. Both use NAD ϩ as an oxidation cofactor. 7␣-HCD did not oxidize either substrate with NAD ϩ . This demonstrates that 7␣-HCD does not catalyze reactions identical to those of the 3␤-or 17␤-HSD families. 7␣-HCD oxidizes hydroxyl groups at the 7-position using NAD ϩ as cofactor but not at the 3-position of 7␣-hydroxycholesterol. We previously showed that NAD ϩ can serve as an oxidizing substrate for this enzyme (4), but the K m for NAD ϩ is 70-fold larger than that for NADP ϩ and the V max is 1 ⁄3 as large, respectively. We also have performed experiments suggesting that other 7␣-hydroxylated steroids, such as 7␣-hydroxydehydroepiandrosterone, are also substrates for this enzyme. 2 These experiments confirmed the conclusion that 7␣-HCD is distinct from 3␤-hydroxy-⌬ 5 -C 27 -steroid oxidoreductase.
Michaelis-Menten Kinetic Constants for Selected Substrates of 7␣-HCD-To determine whether 7␣-hydroxycholesterol might be a physiological substrate for this enzyme, steadystate kinetic parameters for NADP ϩ reduction were determined. Since these reactions displayed simple Michaelis-Menten kinetics, the K m and V max of these sterols for 7␣-HCD were determined with enzyme obtained from the 2Ј,5Ј-ADP agarose column (Table IV). The K m of 7␣-and 7␤-hydroxycholesterol was nearly identical to that of corticosterone, and the V max values were also identical, indicating that both classes of sterols are metabolized at nearly identical maximal rates of substrate oxidation. When the V max /K m values were considered, corticosterone was found to be a slightly better substrate (approximately 1.8-fold) than 7␣-hydroxycholesterol.
Effect of an 11␤-HCD Inhibitor on 7␣-HCD Activity-To confirm the results of the kinetic study, an inhibition experiment using carbenoxolone, a specific competitive inhibitor of 11␤-HSD (16), was performed. In these experiments, the sterol substrate concentration was 4 M, a concentration approximately equal to the K m for both 7␣-hydroxycholesterol and corticosterone. The concentration of NADP ϩ and purified 7␣-HCD used were 0.5 mM and 19 g/ml, respectively. The results are shown in Fig. 4. The IC 50 of carbenoxolone for 7␣-hydroxycholesterol oxidation was approximately 75 nM, and for corticosterone oxidation it was approximately 210 nM. Although carbenoxolone is a slightly more potent inhibitor of 7␣-HCD activity and the V max /K m parameters for corticosterone and 7␣-hydroxycholesterol vary by 1.8-fold, 7␣-hydroxycholesterol may serve as a natural substrate for the 7␣-HCD found in the hamster liver microsomal fraction. DISCUSSION Purified 7␣-HCD had an apparent molecular mass of 34 kDa and preferred NADP ϩ as cofactor, like the liver-specific type I 2 J. Fitzpatrick, X. D. Lei, and R. A. Prough, unpublished results.
FIG. 4. Inhibition of 7␣-HCD-catalyzed oxidation of 7␣-hydroxycholesterol and corticosterone by carbenoxolone. Carbenoxolone in 100 mM potassium phosphate buffer, pH 7.4, was added to reaction mixtures at the indicated concentrations. Purified enzyme was incubated with inhibitor at 37°C for 5 min, and the reaction was initiated by the addition of either 7␣-hydroxycholesterol or corticosterone. Enzyme activity was assayed by the fluorometric method described under "Experimental Procedures." 11␤-HSD. NAD ϩ was about 50% as effective as a cofactor at a much higher concentration than needed for NADP ϩ (4). Partial N-terminal amino acid sequence analysis demonstrated that 7␣-HCD has high sequence similarity to human and rat type I 11␤-HSD in the N-terminal region, indicating that it may be related to type I 11␤-HSD. The 7␣-HCD dehydrogenase reaction appeared to be irreversible for the purified hamster enzyme, which was also observed by others for purified rat liver type I 11␤-HSD (17). In contrast to purified rat liver type I 11␤-HSD, membrane-bound (18) and expressed recombinant rat liver type I 11␤-HSD (17,19) have both reductase and dehydrogenase activity. The reason that purified 11␤-HSD and 7␣-HCD did not display reductase activity is unknown. Substrate specificity studies demonstrate that 7␣-HCD has substrates in common with 11␤-HSD but not with 3␤-HSD or 17␤-HSD. 7␣-HCD has high affinity for 7␣-hydroxycholesterol and corticosterone, indicating that both steroids may serve as endogenous substrate in vivo. Other bile acids and cholesterol derivatives could be oxidized at the 3␤-ol position, albeit at much lower rates of oxidation than 7␣-hydroxycholesterol. Experiments designed to determine the species-specific expression of 7␣-HCD using a polyclonal antibody demonstrated that 7␣-HCD was expressed in hamster and to some extent in human microsomes but not in rat liver microsomes. The tissuespecific expression of 7␣-HCD was as follows: adrenal Ն liver Ͼ kidney Ͼ testis Ͼ Ͼ brain Ͼ lung, which is similar to the pattern of expression of type I 11␤-HSD. These data indicate that 7␣-HCD may also participate in adrenal steroid hormone metabolism. The most surprising finding of this study is that the 7-oxocholesterol formation is catalyzed by a steroid dehydrogenase similar to 11␤-HSD type I. Two 11␤-HSD forms have been identified. Type I 11␤-HSD preferentially uses NADP ϩ and is present mainly in liver. It may be responsible for inactivating glucocorticoid hormones, i.e. cortisol or corticosterone, in that tissue.
The type II 11␤-HSD enzyme utilizes NAD ϩ instead of NADP ϩ as a cofactor and is present in rabbit kidney cortical collecting duct cells (20), sheep kidney (21), human placenta (22), and many human fetal tissues (23). Agarwal et al. (24) isolated a sheep kidney cDNA clone encoding Type II 11␤-HSD by expression screening using Xenopus oocytes. The cDNA (1.8 kilobases) encodes a protein of 427 amino acid residues (M r ϭ 46,700). Based on predicted amino acid sequence, it was only 20% sequence similar to the type I 11␤-HSD; regions that had identity included the putative pyridine nucleotide-and steroidbinding regions observed in many short chain alcohol dehydrogenases. The expressed enzyme functions as an NAD ϩ -dependent 11␤-dehydrogenase with apparent K m values of 15 nM for cortisol and 0.7 nM for corticosterone and displayed no detectable reductase activity. The cDNA hybridized to a 1.9-kilobase mRNA species in kidney and adrenal and was detectable in colon. There was no detectable hybridization of the cDNA probe to mRNA from liver, lung, testis, ovary, heart, stomach, small intestine, or skin. The corresponding gene in humans is a candidate for the syndrome of apparent mineralocorticoid excess. In humans, the cortisol metabolite cortisone is excreted into the urine (25), but in rats about 90% of corticosterone metabolite 11-dehydrocorticosterone is recovered in bile (26), reflecting primarily hepatic metabolism.
Glycyrrhetinic acid, a component of licorice, is a cyclic triterpene whose fused ring structure resembles that of glucocorticoids. This agent has been linked to a syndrome similar to mineralocorticoid excess in a Dutch population (16). The observation that glycyrrhetinic acid can decrease blood cholesterol levels (27) and increase bile secretion (28) indicates that isozyme(s) of 11␤-HSD may participate in or regulate bile acid synthesis. Our study demonstrates that 7␣-HCD is a putative type I 11␤-HSD in hamster and is inhibited by the glycyrrhetinic acid derivative, carbenoxolone. Since the product of 7␣-HCD, 7-oxocholesterol, is a potent inhibitor of cholesterol 7␣hydroxylase, inhibition of 7␣-HCD might diminish the formation of 7-oxocholesterol and thus activate cholesterol 7␣hydroxylase. This effect would enhance flux of the sterols through the bile acid synthesis pathway and possibly lower blood cholesterol. 7␣-HCD may also catalyze formation of a yet unknown hormone as a side product of the bile acid biosynthetic pathway in hamsters and possibly in humans.