Enzymatic Properties of a Member (AKR1C19) of the Aldo-Keto Reductase Family

June 2005 Notes Biol. Pharm. Bull. 28(6) 1075—1078 (2005) 1075

Shuhei ISHIKURA, Kenji HORIE, Masaharu SANAI, Kengo MATSUMOTO, and Akira HARA* Laboratory of Biochemistry, Gifu Pharmaceutical University; Mitahora-higashi, Gifu 502–8585, Japan. Received January 31, 2005; accepted March 1, 2005

A member (AKR1C19) of the aldo-keto reductase (AKR) superfamily, found by mouse genomic analysis, was shown to be highly expressed in the liver and gastrointestinal tract, but its function remains unknown. In this study, the recombinant AKR1C19 was expressed and purified to homogeneity. The enzyme was a 36-kDa monomer, and reduced a-dicarbonyl compounds such as camphorquinone and isatin using both NADH and NADPH as the coenzymes. Although apparent kinetic constants for the two coenzymes were similar, the NADPH-linked activity was potently inhibited by submillimolar concentrations of NAD, but the inhibition of the NADH-linked activity was not significant, suggesting that the enzyme exhibits the NADH-linked reductase activity in vivo. AKR1C19 slowly oxidized 3-hydroxyhexobarbital, S-indan-1-ol and cis-benzene dihydrodiol, but was inactive towards steroids, prostaglandins, monosaccharides, and other xenobiotic alcohols. In addition, the enzyme was inhibited only by dicumarol, lithocholic acid and genistein of various compounds tested. Thus, AKR1C19 possesses properties distinct from other members of the AKR superfamily, and may function as a reductase for endogenous isatin and xenobiotic a-dicarbonyl compounds in the liver and gastrointestinal tract. Key words aldo-keto reductase superfamily; AKR1C19; dual coenzyme specificity; isatin; 3-hydroxyhexobarbital dehydrogenase

The aldo-keto reductase (AKR) superfamily is a rapidly Dr. R. Takenoshita (University of Fukuoka, Japan). All other growing group of NAD(P)(H)-dependent oxidoreductases chemicals were of the highest grade that could be obtained that metabolize carbohydrates, steroids, prostaglandins, and commercially. other endogenous aldehydes and ketones, as well as xenobi- cDNA Isolation The cDNA for AKR1C19 was ampli- otic compounds.1) Currently there are more than a hundred fied by reverse transcription (RT)-PCR from the total RNA of known members of this superfamily classified into 14 fami- an 8-week-old male ICR mouse liver. The preparation of the lies.2) The largest family, AKR1, is subdivided into five total RNA and RT were carried out as described previously.7) subfamilies: AKR1A) mammalian aldehyde reductases; PCR was performed with Pfu DNA polymerase (Stratagene). AKR1B) mammalian aldose reductases; AKR1C) hydroxy- The sense primer (5-ATGAGTTCCAAACAGCAA) and steroid dehydrogenases (HSDs) and prostaglandin F syn- antisense primer (5-ACTAAAATTCATCAGAAAAG) cor- thases; AKR1D) D 4-3-ketosteroid-5b-reductases; and respond to positions 1—18 and 954—973, respectively, of AKR1E) mouse keto-reductase. The subfamilies are defined the sequence of a transcript of the AKR1C19 gene. The PCR by a 60% identity in amino acid sequence among subfam- products of 973 base pairs were ligated into pCR T7/CT- ily members, but the AKR1C subfamily includes some mem- TOPO vectors (Invitrogen), and the expression constructs bers that have not been studied as to their enzymatic proper- were transfected into Escherichia coli BL21 (DE3) pLysS ac- ties or functions. One such member is AKR1C19 that was cording to the protocol described by the manufacturer. The found by mouse genomic analysis and cDNA isolation from inserts of the cloned cDNAs were sequenced using a the liver and gastrointestinal tract.3) Although AKR1C19 is CEQ2000XL DNA sequencer (Beckman Coulter) to confirm postulated to be a 3(20)a-HSD-like enzyme based on its that the deduced amino acid sequences of the cDNAs are highest sequence identity (72%) with human 3(20)a-HSD identical to that of AKR1C19 reported by Vergnes et al.3) (AKR1C1) in the AKR1C subfamily,2) a mouse counterpart Expression and Purification of Recombinant Protein of AKR1C1 is shown to be an NADP(H)-dependent The E. coli cells were cultured in a LB medium containing 3a/3b/20a-HSD (AKR1C18),4,5) which shares a lower se- ampicillin (50 mg/ml) at 37 °C until the absorbance at 600 nm quence identity (65%) with AKR1C19. To determine the reached 0.5. Then isopropyl 1-thio-b-D-galactopyranoside functional relationship of AKR1C19 with AKR1C18, we ex- (1 mM) was added, the culture was continued for 24 h at amined the enzymatic properties of the recombinant 20 °C. The cells were collected and the extract was prepared AKR1C19. as described previously.7) The recombinant AKR1C19 was

puriﬁed at 4 °C by (NH4)2SO4 fractionation and subsequent MATERIALS AND METHODS three column chromatography steps. The enzyme fraction,

precipitated between 35 and 75% (NH4)2SO4 saturation, was Chemicals Prostaglandins were obtained from Cayman dialyzed against Buffer A (10 mM Tris–HCl, pH 7.5, 5 mM 2- Chemicals (Ann Arbor, MI, U.S.A.), steroids were from mercaptoethanol, 1 mM EDTA and 20% glycerol), and was Sigma Chemicals and Steraloids (Newport, RI, U.S.A.), and applied to a Sephadex G-100 column (370 cm) equilibrated resins for column chromatography were from Amersham with Buffer A. The enzyme fraction was applied to a Q- Biosciences (Piscataway, NJ, U.S.A.). trans-Benzene dihy- Sepharose column (220 cm) equilibrated with Buffer A. 6) drodiol was synthesized by the method of Platt and Oesch. This enzyme eluted with a linear gradient of 0—0.1 M NaCl a- and b-3-Hydroxyhexobarbitals (3HBs) were gifts from was dialyzed against Buffer A, and applied to a Red-

Sepharose column (1.55 cm) equilibrated with the same buffer. The column was washed with Buffer A containing 0.1 M NaCl, and the enzyme was eluted with Buffer A containing 2 mM NAD and 0.1 M NaCl. Assay of Enzyme Activity Reductase and dehydrogenase activities of AKR1C19 were assayed by measuring the rate of change in NAD(P)H absorbance (at 340 nm) and its fluorescence (at 455 nm with an excitation wavelength of 340 nm), respectively. The standard reaction mixture for the reductase activity consisted of 25 mM Tris–HCl buffer, pH 7.4, 0.1 mM NADH, substrate and enzyme, in a total volume of 2.0 ml. To detect low activities in the cell extract and enzyme preparations during the purification, 0.1 M potassium phosphate buffer, pH 6.0, was employed instead of the Tris– HCl buffer. Isatin (0.1 mM) was used as the substrate, unless otherwise noted. The dehydrogenase activity was determined in 25 mM Tris–HCl, pH 7.4, containing 1 mM NAD and an Fig. 1. SDS-PAGE of E. coli Cell Extracts and Purified Recombinant appropriate amount of alcohol substrate. One unit (U) of en- AKR1C19 zyme activity was defined as the amount that catalyzes the The gel (12.5%) was stained with 0.2% Coomassie Brilliant Blue. Lanes: a, molecular mass markers (their positions are indicate in kDa); b, the extract (20 mg) of E. coli reduction or formation of 1 mmol NADH per minute at cells transfected with the expression vector alone; c, the extract (20 mg) of E. coli cells transfected with the expression vector harboring AKR1C19 cDNA; and d, the purified 25 °C. The apparent Km and kcat values were determined over a range of five substrate concentrations at a saturating con- recombinant AKR1C19 (2 mg). centration of coenzyme by fitting the initial velocities to the

Michaelis–Menten equation. The kinetic constants and IC50 (inhibitor concentrations required for 50% inhibition) values are expressed as the means of two determinations. Protein concentration was determined by the method of Bradford8) using bovine serum albumin as the standard.

RESULTS AND DISCUSSION

Purification of Expressed AKR1C19 SDS-PAGE analysis of the E. coli cell extracts (Fig. 1) shows that the cells transfected with the expression plasmids harboring the AKR1C19 cDNA overexpresses a 36-kDa protein, which was not present in the control cells transfected with the vector alone. This cell extract exhibited NADH-linked reductase activity towards isatin (0.0032 U/mg). The recombinant AKR1C19 was purified by both detecting the 36-kDa protein Fig. 2. Effect of NAD on NADPH-Linked and NADH-Linked S-Cam- on SDS-PAGE and assaying the isatin reductase activity. The phorquinone Reductase Activities of AKR1C19 The activity was assayed with 0.1 mM NADPH () or 0.1 mM NADH () as the enzyme was eluted at a low molecular weight of approximately 35 kDa on the Sephadex G-100 chromatography step, coenzyme, and is expressed as a percentage relative to the activity without NAD . suggesting its monomeric nature. The final preparation with isatin activity of 0.53 U/mg was greater than 99% pure (Fig. NADP(H). The Km and kcat/Km values for NADH determined 1), and the purification yield was 26% (17 mg/l of cells). in the presence of 0.1 mM S-camphorquinone were 6.9 m M 1 1 pH Dependency The NADH-linked isatin reductase ac- and 3.3 min m M , respectively, and the respective values 1 1 tivity of the purified AKR1C19 was increased by decreasing for NADPH were 11 m M and 2.8 min m M . In the reverse the pH from 8.0 to 6.0 in 0.1 M phosphate buffers. The activ- reaction with 2 mM a-3HB as the substrate, NAD (Km 1 1 ity was also influenced by species and concentrations of the 32 m M and kcat/Km 0.61 min m M ) was a better coenzyme 1 1 assay buffer. For example, the activities assayed in 0.1 M than NADP (Km 153 m M and kcat/Km 0.017 min m M ). Tris–HCl buffers (pH 7.2—8.0) were higher (approximately Since the intracellular ratio of NAD to NADH is high in 9) 1.5-fold) than those assayed in 0.1 M phosphate buffers, and contrast to the low NADP /NADPH ratio, the effects of further increased to a maximum of 1.2-fold by lowering the NAD on the NADH-linked and NADPH-linked reductase buffer concentration to 25 mM. In the reverse reaction with activities of AKR1C19 were examined. The NADPH-linked NAD as the coenzyme, the enzyme oxidized a-3HB and reductase activity was potently inhibited by NAD even at a the maximal rate was observed at pH 7.5 when 25 mM Hepes low concentration of 0.1 mM, whereas the inhibition of the and Tris–HCl buffers were used. Therefore, the enzymatic NADH-linked activity was low, and about 50% of the activity 9) properties of AKR1C19 were examined at a physiological retained at a cellular NAD concentration of 2 mM (Fig. 2). pH of 7.4 using 25 mM Tris–HCl buffer. It is likely that the enzyme acts as both NADH-linked reduc- Coenzyme Specificity The recombinant AKR1C19 dis- tase and NAD-linked dehydrogenase in vivo. In addition, played dual coenzyme specificity for NAD(H) and the difference in NAD inhibition between the NADH- and June 2005 1077

Table1. Kinetic Constants for Substrates the AKR1C subfamily and sugars such as D/L-threitols, meso- erythritol, D/L-arabitols, ribitol, xylitol, inositol, dulcitol, D- Km kcat kcat/Km Substrate 1 1 1 sorbitol and D-xylose. (m M) (min ) (min m M ) Inhibitor Sensitivity The reductase activity of Reduction AKR1C19 towards 0.1 mM S-camphorquinone was inhibited S-Camphoquinone 3.9 49.0 12.6 by dicumarol, lithocholic acid and genistein (IC50 values 16-Ketoestrone 1.4 13.7 9.79 were 1.9, 5.7 and 56 m M, respectively). No significant inhibi- Acenaphthenequinone 4.4 16.7 3.80 Phenanethrenequinone 18 47.3 2.63 tion (less than 25%) was observed by 5 m M benzbromarone, 5-Bromoisatin 16 32.2 2.01 10 m M medroxyprogesterone acetate, 50 m M dexamethasone, R-Camphorquinone 8.8 16.6 1.89 hexestrol, diethylstilbestrol, indomethacin, phenolphthalein, Isatin 16 21.4 1.34 fulfenamic acid or 0.1 mM 1,10-phenanthroline (data not Dimethyl-2-oxoglutarate 38 12.9 0.34 shown), which are potent inhibitors of HSDs and/or 2,3-Hexanedione 46 6.54 0.14 2,3-Heptanedione 209 15.3 0.073 prostaglandin F synthases of the AKR1C subfamily. Thus, 2,3-Pentanedione 149 8.32 0.056 the inhibitor sensitivity is an additional characteristic of Oxidation AKR1C19. a-3HB 580 2.06 0.004 Roles of AKR1C19 The substrate specificity indicates b-3HB 820 1.28 0.002 that AKR1C19 is not a HSD, in contrast to its previously S-Indan-1-ol 1200 2.53 0.002 cis-Benzene dihydrodiol 6000 2.69 0.0004 postulated function as a HSD. Among the substrates of this enzyme, only isatin is an endogenous component, which is produced through indole supplied by metabolism of trypto- NADPH-linked activities suggests that the affinity for NADH phan and/or phenylalanine in intestinal bacteria.21) Since is much higher than that for NADPH, although the apparent isatin has been shown to inhibit monoamine oxidase and

Km values for the two coenzymes are similar. The preference acetylcholinesterase, and to exhibit various pharmacological for NAD(H) over NADP(H) may be due to the presence of actions,20,21) it must be metabolized into inactive metabolites. 5) glutamine and aspartic acid at positions 270 and 276, respec- In mice, NADP(H)-dependent AKR1C18, aflatoxin B1 alde- 22) 23) tively, which are conserved in NAD(H)-preferring members hyde reductase (AKR7A4), carbonyl reductases and L- of the AKR1C subfamily and are thought to interact with the xylulose reductase24) have been shown to reduce isatin. How- adenine ribose of the coenzyme.10) ever, AKR1C19 differs from the known isatin-reducing en- Substrate Specificity The purified recombinant zymes with respect to tissue distribution, in that its mRNA is AKR1C19 reduced several a-dicarbonyl compounds, and the specifically expressed in the gastrointestinal tract and liver.3) catalytic efficiency (kcat/Km) for alicyclic and aromatic sub- AKR1C19 may play a role in the reductive metabolism of strates was higher than that for aliphatic substrates such as isatin in the gastrointestinal tract. Alternatively, it can be in- dimethyl-2-oxoglutarate and 2,3-hexanedione (Table 1). volved in the gastrointestinal detoxification of chemically re- While isatin and 5-bromoisatin were good substrates, 1- active a-dicarbonyl compounds that are ingested with foods. phenylisatin, a substrate of AKR1C18,5) was not reduced. From the point of view of drug metabolism, it is interesting The enzyme showed less than 5% of the isatin reductase ac- that the Km values for a- and b-3HBs of AKR1C19 are com- tivity towards 1 mM of other a-dicarbonyl compounds parable to those of mouse liver 3HB dehydrogenase, which (methylglyoxal, diacetyl and 3,4-hexanedione) and mono- has been suggested to be identical to NADP -preferring carbonyl compounds (4-nitrobenzaldehyde, pyridine-4- 17b-HSD (AKR1C6).25) AKR1C19 is clearly different from aldehyde, 4-nitroacetophenone, 4-benzoylpyridine), which the previously known 3HB dehydrogenase with respect to the are efficiently reduced by AKR1C18,5) mouse 3(17)a- specificity for coenzymes and other substrates, and is a new HSD (AKR1C21),11) mouse 17b-HSD (AKR1C6),12,13) type of 3HB dehydrogenase. AKR1C1,14,15) and/or human 3a-HSDs (AKR1C2 and AKR1C4).14,16) In this respect, AKR1C19 also differs from REFERENCES other members of the AKR superfamily such as aldehyde re- 1) Jez J. M., Flynn T. G., Penning T. M., Biochem. Pharmacol., 54, 639— 17) ductase, aldose reductase and aflatoxin B1 aldehyde reduc- 647 (1997). tase14) with broad substrate specificity for a variety of car- 2) Hyndman D., Bauman D. R., Heredia V. V., Penning T. M., Chem. Biol. bonyl compounds. AKR1C19 was inactive towards various Interact., 143—144, 499—525 (2003). 3) Vergnes L., Phan J., Stolz A., Reue K., J. Lipid Res., 44, 503—511 ketosteroid substrates of the above HSDs, sugars (D-glycer- (2003). aldehyde, D-xylose, D-glucose and D-glucuronate) and prosta- 4) Ishida M., Chang K.-T., Hirabayashi K., Nishihara M., Takahashi M., glandin D2, a representative substrate of human and bovine J. Reprod. Dev., 45, 321—329 (1999). prostaglandin F synthases (AKR1C3 and AKR1C11).18,19) 5) Ishikura S., Nakajima S., Kaneko T., Shinitani S., Usami N., Ya- In the reverse reaction, AKR1C19 oxidized a- and b- mamoto I., Carbone V., El-Kabbani O., Hara A., “Enzymology and 3HBs, S-indan-1-ol and cis-benzene dihydrodiol, but the cat- Molecular Biology of Carbonyl Metabolism 12,” ed. by Weiner H., Purdue University Press, West Lafayette, 2005, in press. alytic efficiency for the substrates was lower compared to 6) Platt K. L., Oesch F., Synthesis, 7, 449—450 (1977). that in the reductase reaction. The enzyme was inactive to- 7) Nakanishi M., Deyashiki Y., Ohshima K., Hara A., Eur. J. Biochem., wards S-tetralol, R-indan-1-ol, trans-benzene dihydrodiol and 228, 381—387 (1995). 9a,11b-prostaglandin F , which are non-steroidal substrates 8) Bradford M. M., Anal. Biochem., 72, 248—254 (1976). 2 9) Williamson J. R., Corkey B.E., Method Enzymol., 13, 434—513 for the above HSDs and prostaglandin F synthases of the (1969). AKR1C subfamily. The enzyme was also inactive towards 10) Todaka T., Yamano S., Toki T., Arch. Biochem. Biophys., 374, 189— hydroxysteroid substrates of the above HSDs belonging to 197 (2000). 1078 Vol. 28, No. 6

11) Ishikura S., Usami N., Nakajima S., Kameyama A., Shiraishi H., Car- 19) Chen L.-Y., Watanabe K., Hayaishi O., Arch. Biochem. Biophys., 296, bone V., El-Kabbani O., Hara A., Biol. Pharm. Bull., 27, 1939—1945 17—26 (1992). (2004). 20) Medvedev A. E., Clow A., Sandler M., Glover, V., Biochem. Pharma- 12) Hara A., Inoue Y., Tanabe H., Nakagawa M., Nakayama T., Sawada col., 52, 385—391 (1996). H., Biochemistry (Life Sci. Adv.), 7, 121—126 (1988). 21) Glover V., Bhattacharya S. K., Chakrabarti A., Sandler M., Stress 13) Deyashiki Y., Ohshima K., Nakanishi M., Sato K., Matsuura K., Hara Med., 14, 225—229 (1998). A., J. Biol. Chem., 270, 10461—10467 (1995). 22) Hinshelwood A., McGarvie G., Ellis E., FEBS Lett., 523, 213—218 14) O’Connor T., Ireland L. S., Harrison D. J., Hayes J. D., Biochem. J., (2002). 343, 487—504 (1999). 23) Usami N., Kitahara K., Ishikura S., Nagano M., Sakai S., Hara A., Eur. 15) Higaki Y., Kamiya T., Usami N., Shintani S., Shiraishi H., Ishikura S., J. Biochem., 268, 5755—5763 (2001). Yamaomoto I., Hara A., Drug Metabol. Pharmacokin., 17, 348—356 24) Nakagawa J., Ishikura S., Asami J., Isaji T., Usami N., Hara A., Saku- (2002). rai T., Tsuritani K., Oda K., Takahashi M., Yoshimoto M., Otsuka N., 16) Deyashiki Y., Taniguchi H., Amano T., Nakayama T., Hara A., Sawada Kitamura K., J. Biol. Chem., 277, 17883—17891 (2002). H., Biochem. J., 282, 741—746 (1992). 25) Takenoshita R., Naitoh K., Toki S., “Enzymology and Molecular Biol- 17) Flynn T. G., Biochem. Pharmacol., 31, 2705—2712 (1982). ogy of Carbonyl Metabolism 2,” eds. by Weiner H., Flynn T. G., Alan 18) Mastuura K., Shiraishi H., Hara A., Sato K., Deyashiki Y., Ninomiya R. Liss, New York, 1989, pp. 279—291. M., Sakai S., J. Biochem. (Tokyo), 124, 940—946 (1998).