BAY 2402234

Rabbit dehydrogenase/reductase SDR family member 11 (DHRS11): Its identity with acetohexamide reductase with broad substrate specificity and inhibitor sensitivity, different from human DHRS11

Abstract

Human dehydrogenase/reductase SDR family member 11 (DHRS11) has been recently reported to be an NADP+-dependent 3(17)β-hydroxysteroid dehydrogenase, and its orthologs are predicted in genomic analyses of various animals. Among them, the amino acid sequence of predicted rabbit DHRS11 shares 92% identity with that of human DHRS11 and matches peptide sequences (composed of total 87 amino acids) of rabbit heart acetohexamide reductase (RHAR) previously reported. However, the physiological role of RHAR remains un- known, because its known substrates are only acetohexamide and 1,4-naphthoquinone. To elucidate whether the two rabbit enzymes are identical, we have isolated the cDNA for rabbit DHRS11, which was abundantly detected in the brain, heart, kidney and intestine by RT-PCR. The recombinant rabbit DHRS11 reduced acetohexamide and 1,4-naphthoquinone, and was inhibited by tolbutamide and phenobarbital (RHAR-specific inhibitors), de-
monstrating its identity with RHAR. Rabbit DHRS11 also reduced α-dicarbonyl compounds, aldehydes and aromatic ketones (acetylbenzenes and acetylpyridines), and exhibited 3(17)β-hydroxysteroid dehydrogenase activity. It was competitively inhibited not only by tolbutamide and phenobarbital, but also more potently by several non-steroidal anti-inflammatory drugs such as diclofenac and sulindac. The broad substrate specificity and inhibitor sensitivity were different from those of human DHRS11, which did not reduce aliphatic aldehydes and aromatic ketones despite its higher 3(17)β-hydroxysteroid dehydrogenase activity, and was insensitive to tolbutamide, phenobarbital and diclofenac. The site-directed mutagenesis of Thr163 and Val200 in human DHRS11 to the corresponding residues (Gly and Leu, respectively) in rabbit DHRS11 suggested that these re- sidues are pertinent to the differences in properties of rabbit and human DHRS11s.

1. Introduction

A 260-amino acid protein encoded in the DHRS11 gene is currently annotated as “dehydrogenase/reductase SDR family member 11 (DHRS11)” in the HUGO gene nomenclature database (http://www. genenames.org). DHRS11 belongs to the short-chain dehydrogenase/reductase (SDR) superfamily, which encompasses numerous enzymes that play roles in the metabolism of lipids, carbohydrates, vitamins, drugs and xenobiotics [1]. In this superfamily, DHRS11 is classified as SDR24C1 [1,2]. A crystal structure of tetrameric human DHRS11 complexed with NADP+ and acetic acid was previously deposited in the RCSB protein data bank (PDB-ID: 1XG5). The enzyme has been recently reported to be a reductive 17β-hydroxysteroid dehydrogenase (HSD) that reduces estrone, dehydroepiandrosterone (DHEA) and 17-keto-5α-androstanes using NADPH as coenzyme [3] (Fig. 1). Human DHRS11 also exhibits reductive 3β-HSD activity towards 3-keto-5β-androstanes, 3-keto-5β-pregnanes and 3-ketobile acids, and furthermore reduces α- dicarbonyls such as diacetyl and methylglyoxal. The DHRS11 mRNA is expressed widely in human tissues, most abundantly in the testis, small intestine, colon and kidney, and has been thought to function in me- tabolism of steroids and endogenous α-dicarbonyls. DHRS11 was highly expressed in some cancer cells [3] and cerebral melanoma metastases tissue samples [4].

Orthologs of human DHRS11 sharing more than 90% amino acid sequence identity are predicted in genomic analyses of other primates, horse, pig, cow, dog, rabbit, rat and mouse. Among them, the amino acid sequence of the predicted rabbit DHRS11 (SDR24C189) matches those of peptides derived from rabbit heart acetohexamide reductase (RHAR) [5] (Fig. 2). RHAR was originally purified as a 110 kDa-tetra- meric NADPH-dependent reductase that specifically catalyzes the re- duction of acetohexamide into hydroxyhexamide (Fig. 1) [6]. Subse- quently, the enzyme was reported to efficiently reduce 1,4- naphthoquinone (NQ) and 5-hydroxy-1,4-NQ [5], and to be uniquely inhibited by phenobarbital and tolbutamide [5,7]. The substrate spe- cificity and inhibitor sensitivity are different from those of rabbit monomeric carbonyl reductases [7,8] and peroxisomal tetrameric car- bonyl reductase [9,10] belonging to the SDR superfamily. Thus, role and endogenous substrates of RHAR remain unknown.

Fig. 1. 17β-HSD (a) and 3β-HSD (b) activities of human DHRS11 and acet- ohexamide reduction by RHAR (c).

Fig. 2. Alignment of amino acid sequences of rabbit DHRS11 (rDHRS11) and human DHRS11 (hDHRS11). The underlined residues in rDHRS11 are identical to those determined by peptide sequencing of RHAR. The hyphens in the hDHRS11 sequence represent identical residues to those of rDHRS11. The po- sitions of catalytic triad of the structurally evaluated SDR family enzymes and mutated residues of human DHRS11 are indicated with closed and open ar- rowheads, respectively.

In order to elucidate whether rabbit DHRS11 is identical to RHAR, we have cloned the cDNA for rabbit DHRS11 and studied the enzymatic properties of the recombinant enzyme. Besides its identity with RHAR, rabbit DHRS11 was found to reduce various carbonyl compounds in- cluding 3- and 17-ketosteroids, and to be inhibited by non-steroidal anti-inflammatory drugs (NSAIDs) and the above known RHAR in- hibitors. However, the broad substrate specificity and unique inhibitor sensitivity of rabbit DHRS11 differ from those of human DHRS11, de- spite their high amino acid sequence identity of 92% (Fig. 2). We also investigated the residues responsible for the differences in substrate specificity and inhibitor sensitivity between rabbit and human DHRS11s by molecular modelling and site-directed mutagenesis.

2. Materials and methods

2.1. Materials

Steroids were obtained from Sigma Chemicals (Perth, WA) and Steraloids (Newport, RI); prostaglandins were from Cayman Chemical (Ann Arbor, MI); and pCold I expression vector was from Takara Bio (Otsu, Japan). Hydroxyhexamide and 3-deoxyglucosone were kindly denoted by Dr. Y. Imamura (Kumamoto University, Japan) and Kaken Pharmaceutical Co. (Tokyo, Japan), respectively. Acetohexamide (Shionogi & Co., Osaka, Japan), ketoprofen (Hisamitsu Pharmaceuticals, Saga, Japan) and suprofen (Taiyo Yakuhin Co., Nagoya, Japan) were provided by the manufacturers. All other chemi- cals were of the highest grade that could be obtained commercially. The cDNA for human DHRS11 and its recombinant enzyme were prepared as described previously [3].

2.2. cDNA isolation and site-directed mutagenesis

The expression of the mRNA for rabbit DHRS11 in various tissues of a male Japanese white rabbit was examined by reverse transcription (RT)-PCR. The preparation of total RNA, RT, and DNA techniques fol- lowed the standard procedures described by Sambrook et al. [11]. PCR was performed with Pfu DNA polymerase and a pair of sense and an- tisense primers, which contain NdeI and SalI sites. The cDNA for rabbit
β-actin was also amplified as an internal control with the specific primers. The PCR products were separated by agarose gel electrophoresis, and stained with ethidium bromide. The primers for the amplification of the cDNA for rabbit DHRS11 were designed based on the mRNA for Oryctolagus cuniculus DHRS11 (accession no. XM_008271169) predicted from the rabbit genomic analysis, and their sequences and PCR condi- tions are summarized in Table S1 (Supplementary data). The PCR product from the total RNA sample of rabbit heart was purified, and ligated into the pCold I vectors that had been digested with NdeI and SalI. The insert of the cloned cDNA was sequenced by using a Beckman CEQ8000XL DNA sequencer, to confirm that the 260-amino acid se- quence of rabbit DHRS11 fused to the N-terminal 6-His tag is encoded. The nucleotide sequence of the cDNA was deposited in DDBJ database with the accession no. LC380589.

Mutagenesis was performed using a KOD-Plus-Mutagenesis Kit (Toyobo, Osaka, Japan) and the pCold I expression plasmid harboring the cDNA for human DHRS11 as the template according to the protocol described by the manufacturer. The sequences of the primers used for the mutagenesis of Thr163Gly and Val200Leu are shown in Table S1 (Supplementary data). The cDNA for the double mutant of Thr163Gly/ Val200Leu (TGVL) was prepared using the expression plasmid har- boring the cDNA for Thr163Gly mutant as the template. The coding regions of the cDNAs in the expression plasmids were sequenced in order to confirm the presence of the desired mutation and ensure that no other mutation had occurred.

2.3. Production of recombinant enzymes

Recombinant rabbit DHRS11 and mutants of human DHRS11 were expressed in Escherichia coli BL21 (DE3) pLysS cells (Invitrogen, Carlsbad, CA) transformed with the expression plasmids harboring the cDNA as described previously [12]. The enzymes were purified to homogeneity from the cell extract using a Ni Sepharose 6 Fast Flow resin (GE healthcare, Little Chalfont, UK) according to the manufac- turer’s manual, and their purities were analyzed by SDS-PAGE ac- cording to standard procedures. Protein concentration was determined by Bradford’s method using bovine serum albumin as the standard [13].

2.4. Assay of enzyme activity

Reductase and dehydrogenase activities were assayed by measuring the rate of change in NADPH absorbance (at 340 nm) and its fluores- cence emission (at 455 nm with an excitation wavelength of 340 nm), respectively, in a 2-ml reaction mixture containing 0.1 M potassium phosphate (pH 7.0), 0.1 mM NADPH or 0.5 mM NADP+, substrate and enzyme. The apparent Km and Vmax values were determined over a range of five substrate concentrations at a saturating concentration of coenzyme by fitting the initial velocities to the Michaelis-Menten equation,
v = Vmax [S] / (Km + [S]) where v is the initial velocity; [S] is the substrate concentration; and Vmax is the maximum velocity. The IC50 (inhibitor concentrations re- quired for 50% inhibition) values were determined in NADP+-linked oxidation of 50 μM 5-androstene-3β,17β-diol (for rabbit DHRS11) and 10 μM 5α-androstane-3β,17β-diol (for human DHRS11 [3]). The kinetic studies in the presence of three concentrations of an inhibitor were carried out in the hydroxysteroid oxidation by the enzymes. The in- hibition pattern was judged from Lineweaver-Burk double reciprocal plots of initial velocities versus substrate concentrations using the fol- lowing equation, 1 = Km ⎡1 + [I] ⎤ 1 + 1 with that predicted from the genomic analysis. The amino acid se- quence deduced from the cDNA was identical to those (composed of a total 87 residues) of 6 peptides derived from RHAR [5], with the ex- ception of two replacements (Trp38→Val and Phe67→Glu) (Fig. 2).

Fig. 3. RT-PCR analysis for expression of mRNA for rDHRS11 in male rabbit tissues: Brain (Br), lung (Lu), heart (He), stomach (St), liver (Li), adrenal gland (Ad), small intestine (SI), colon (Co) and testis (Te). The kidney was divided into the renal cortex (Rc) and renal medulla (Rm). The expression of mRNA for β-actin is shown as the control.

The purified recombinant rabbit DHRS11 showed high NADPH- linked reductase activity towards acetohexamide, and the Km and Vmax values were 0.71 mM and 3.5 U/mg, respectively (Table 1), which are comparable to those reported with RHAR [6]. It also efficiently reduced 1,4-naphthoquinone (NQ) and 5-hydroxy-1,4-NQ, but showed no ac- tivity towards menadione. The specificity for p-quinones also resembles
that reported with RHAR [5]. Furthermore, the acetohexamide reductase activity of rabbit DHRS11 was inhibited to 40 and 25% by where [I] is the inhibitor concentration and Ki is the inhibition con- stant. The Ki for the competitive inhibitor was determined from replots of the slopes (Km/Vmax) of the reciprocal plots versus inhibitor con- centrations ([I]), where a straight line of the replots was obtained and its [I]-axis intercept was equal to – Ki. The Km, Vmax and IC50 values of the wild-type enzymes are expressed as the means of two determina- tions, and Ki values and kinetic constants of the mutant enzymes re- present the means ± S.E. from three determinations. One unit (U) of enzyme activity was defined as the enzyme amount that catalyzes the oxidation or formation of 1 μmol of NADPH per min at 37 °C.

2.5. Molecular and homology modelling

The coordinates for human DHRS11 were obtained from the RCSB Protein Data Bank (PDB-ID: 1XG5). The docking study was performed using MF myPresto ver.3.2.0.33 (https://www.mypresto5.jp/en/), which is a graphical user interface (GUI) software of myPresto [14] and developed by FiatLux Corporation (Tokyo, Japan). We used the general AMBER force field, and the molecular topology files and 100 con- formers for the compound were generated by tplgeneL/myPresto and sievgene programs [15], respectively. For flexible docking, smooth re- action path generation method [16] was used.

A homology model of rabbit DHRS11 was created using the auto- mated protein structure homology-modelling server, SWISS-MODEL [17]. The ternary complex of human DHRS11 with NADP+ and acetic acid (PDB-ID: 1XG5) was selected as the template. The model was su- perimposed with the DHEA-docked model of human DHRS11 using PyMOL (DeLano Scientific, San Carlos, CA).

3. Results and discussion

3.1. cDNA cloning of rabbit DHRS11 and its identity with RHAR

The expression of DHRS11 mRNA in rabbit tissues was first ex- amined by RT-PCR using the forward and reverse primers (Table S1), which were complementary to the sequence of the Oryctolagus cuniculus DHRS11 mRNA (NCBI accession no. XM_008271169) predicted from the rabbit genomic analysis. The corresponding DNA band (685 base pairs) was detected in all the tissues, of which the brain, heart, kidney, intestine and colon showed high expression levels (Fig. 3). The full- length cDNA for rabbit DHRS11 was next isolated from the heart by RT- PCR using the primers (Table S1), and its nucleotide sequence coincided 1 mM phenobarbital and tolbutamide, respectively, which were re- ported to inhibit RHAR to similar extents [5–7]. In addition to matching the amino acid sequences, the above enzymatic properties of rabbit DHRS11 indicate the identity of rabbit DHRS11 with RHAR.

3.2. Substrate specificity for non-steroidal carbonyl compounds

The known substrates of RHAR are only acetohexamide, 1,4-NQs and DL-glyceraldehyde [5,6], whereas human DHRS11 reduces α-di- carbonyls and 3/17-ketosteroids [3]. Therefore, we examined the sub- strate specificity of rabbit DHRS11 for various nonsteroidal carbonyl compounds, whose structures are shown in Supplementary data, Fig. S1. Since acetohexamide is an acetophenone (AP) derivative that pos- sesses a long cyclohexylsulfonamide group at its p-position (Fig. 1), several AP derivatives were first tested as the substrates of rabbit DHRS11 (Table 1). The enzyme exhibited low activity towards AP and its derivatives substituted by nitro or carboxyl group at m- or p-position. Similarly, rabbit DHRS11 reduced 4- and 3-acetylpyridines, which structurally replace the benzene of AP with a pyridine. The catalytic efficiencies (Vmax/Km values) for the aromatic ketones were lower than that for acetohexamide, and the Km values for AP and acetylpyridines without any substituents were much higher than those for the AP de- rivatives and acetohexamide. This suggests that the m- or p-substituent on AP increase the catalytic activity of the enzyme. It should be noted that rabbit DHRS11 did not reduce aromatic ketones with long acyl groups (propiophenone, butyrophenone and 1-phenyl-1,3-butanedione) or with another aromatic ring (benzophenone, 4-benzoylpyridine, 2- phenylacetophenone, metyrapone, ketoprofen and suprofen) instead of the acetyl group of AP. The results suggest that the small acetyl group of the AP derivatives properly binds to the active site of rabbit DHRS11 and is reduced to the hydroxyethyl group. This is also supported by a previous finding that acetohexamide derivatives with longer acyl groups than the acetyl group did not reduce by RHAR [6].

We next examined aromatic aldehydes as the substrates, because the aldehyde group is smaller than the acetyl group of the above AP deri- vatives. Rabbit DHRS11 reduced aromatic aldehydes (pyridinecarbox- aldehydes and benzaldehydes substituted with a nitro, carboxy or chloro group) more highly than the above aromatic ketones. In addi- tion, aliphatic linear-chain aldehydes were reduced with low Vmax va- lues, in which the catalytic efficiency tended to decrease with de- creasing the chain lengths. As reported for RHAR [6], rabbit DHRS11 showed a very low activity towards 1 mM DL-glyceraldehyde, but the activity was increased at higher substrate concentrations, giving an extremely high Km value.

1,4-NQ and 5-hydroxy-1,4-NQ were the most excellent substrates, showing the highest Vmax/Km values, among substrates of rabbit DHRS11. In addition, rabbit DHRS11 efficiently reduced polycyclic aromatic hydrocarbon o-quinones (acenaphthenequinone and 9,10- phenanthrenequinone), although its reactivity towards a non-aromatic o-quinone, 1S-camphorquinone, was low. Furthermore, rabbit DHRS11 reduced various α-dicarbonyl compounds including endogenous dia- cetyl, isatin, methylglyoxal and 3-deoxyglucosone [18–21]. Among them, diacetyl and isatin were reduced with the highest Vmax values and the lowest Km values, respectively. The Km value for NADPH de- termined in the presence of a saturated concentration (10 mM) of dia- cetyl was 4.0 μM.

Since rabbit DHRS11 thus exhibited broad substrate specificity, we examined the reactivity of human DHRS11 towards these newly found carbonyl substrates of the rabbit enzyme. As the results are appended to Table 1, the substrate specificity of human DHRS11 was broader than that previously reported [3], but was clearly different from that of rabbit DHRS11 as follows. 1) Human DHRS11 showed no or little ac- tivity towards the aromatic ketones, 1,4-NQs and aliphatic aldehydes, and conversely reduced menadione that was not a substrate for rabbit DHRS11. 2) Aromatic aldehydes, o-quinones and α-dicarbonyl com- pounds were reduced by human DHRS11 with lower Vmax/Km values than those of rabbit DHRS11, except that isatin was a good endogenous substrate for both rabbit and human enzymes.

3.3. Substrate specificity for steroids and hydroxyhexamide oxidation

Rabbit DHRS11 reduced 3- and 17-ketosteroids at low rates, but its Km values for most steroidal substrates were as low as those for the quinones and isatin, good nonsteroidal substrates (Table 2). No significant activity was detected with 20-ketosteroids (5β-pregnan-3α/β- ol-20-ones and 5β-pregnane-3β,21-diol-20-one). In the reverse reaction using NADP+ as the coenzyme, rabbit DHRS11 oxidized 3β- and 17β- hydroxysteroids (Table 3), but did not show significant activity towards 3α-hydroxysteroids (lithocholic acid, 5α/β-pregnan-3α-ol-20-ones, 4- androsten-3α-ol-17-one and 5α/β-androstan-3α-ol-17-ones) and 17α- hydroxysteroids (17α-estradiol, epitestosterone and 5-androstene-
3β,17α-diol). The Km value for NADP+ determined in the presence of a saturated concentration (100 μM) of 5-androstene-3β,17β-diol was
4.9 μM. The steroid specificities in both reduction and oxidation are similar to that of human DHRS11, indicating the rabbit DHRS11 also acts as 3(17)β-HSD. However, there was a difference between the two enzymes: Rabbit DHRS11 showed lower Vmax/Km values for most of the steroidal substrates than human DHRS11.

Rabbit DHRS11 also oxidized hydroxyhexamide, a product of acetohexamide reduction (Fig. 1) [5,6], suggesting that the enzyme reversibly catalyzes the conversion of acetohexamide to hydro- xyhexamide. In contrast, human DHRS11 did not show the hydro- xyhexamide dehydrogenase activity.

3.4. Possible role of rabbit DHRS11 in carbonyl and steroid metabolism

The substrate specificity of rabbit DHRS11 suggests its role in me- tabolism of xenobiotic and endogenous carbonyl compounds. In the rabbits, monomeric carbonyl reductase [7,8,22,23], aldo-keto re- ductases (AKR1C29-AKR1C33) with 3α-, 17β- and/or 20α-HSD activity [24,25], aldehyde reductase [22,23], aldose reductase (AKR1B2) and aldose reductase-like protein (AKR1B19) [26] were previously char- acterized as cytosolic reductases for the carbonyl compounds. Although rabbit DHRS11 reduced only acetylbenzenes and acetylpyridines of aromatic ketones, its specificities for aldehydes and α-dicarbonyls are broad and similar to those of the above known enzymes. Among the α- dicarbonyl substrates, diacetyl, isatin, methylglyoxal and 3-deoxyglucosone are endogenous and reactive [18–21], and have been re- ported to be reduced by AKR1B2 and AKR1B19 [26]. The catalytic ef- ficiencies for diacetyl and isatin of rabbit DHRS11 are comparable or superior to those of AKR1B2 and AKR1B19, although those for other α-dicarbonyls (methylglyoxal and 3-deoxyglucosone) are lower. Rabbit DHRS11 may function as one of detoxification enzymes for the en- dogenous α-dicarbonyls.

Rabbit DHRS11 exhibited 3(17)β-HSD activity and its Km values for most of the steroidal substrates were lower than those for the non- steroidal carbonyl compounds. Like human DHRS11 [3], rabbit DHRS11 is probably involved in steroid metabolism. Previously,AKR1B19 was reported to exhibit reductive 3β-HSD activity towards 3- keto-5α/β-dihydrosteroids, and is abundantly expressed in rabbit lung, heart, kidney and adrenal gland [26]. Since the catalytic efficiencies for the 3-ketosteroids of AKR1B19 are higher than those of rabbit DHRS11, the importance of rabbit DHRS11 as a reductive 3β-HSD remains ob- scure. In contrast, its role as a reductive 17β-HSD for, particularly, DHEA and its sulfate, may be pivotal, because DHEA and its sulfate act directly as ligands for many hepatic nuclear receptors and G-protein- coupled receptors as well as neurosteroids [27,28], and 5-androstene- 3β,17β-diol, the DHEA metabolite by rabbit DHRS11, is an estrogen receptor ligand [29,30]. DHEA and its sulfate are not metabolized by previously characterized rabbit reductive 17β-HSDs (AKR1C29 – AKR1C33) [24,25 and our unpublished results]. The abundant expression of rabbit DHRS11 in rabbit brain further supports the role of the enzyme in controlling intracellular concentrations of neuroactive and estrogenic DHEA and its metabolites. On the other hand, the physio- logical relevance of the high expression of DHRS11 in rabbit heart re- mains obscure. Recently, rat DHRS11, together with several steroid metabolizing SDR enzymes, has been reported to be highly expressed in adult heart and suggested to be related to cardiomyocyte differentiation [31].

3.5. Inhibitor sensitivity

Known inhibitors of RHAR (i.e., rabbit DHRS11) are phenobarbital, tolbutamide and flavonoids [5–7], whereas those of human DHRS11 are carbenoxolone, flufenamic acid and flavonoids [3]. We examined the effects of these inhibitors on the 5-androstene-3β,17β-diol dehy- drogenase activity of rabbit DHRS11 (Table 4). As described in the identity of rabbit DHRS11 with RHAR, tolbutamide and phenobarbital inhibition was observed with primidone, which is distinct from phe- nobarbital only with respect to lacking C-2 oxygen on the barbiturate ring. The results suggest that the C-5 cyclic 6-membered ring and C-2 carbonyl group on the barbiturate ring are structural requisites for the binding to the enzyme.

Flavonoids, carbenoxolone and flufenamic acid also inhibited rabbit DHRS11. Since the IC50 values for carbenoxolone and flufenamic acid were higher than those of human DHRS11, other NSAIDs were tested as the inhibitors of the rabbit and human enzymes. Although no sig- nificant inhibition was observed by 100 μM flurbiprofen, ibuprofen, acetylsalicylic acid and phenylbutazone (data not shown), following
NSAIDs differently inhibited the two enzymes and were suggested to be divided into three types: 1) Diclofenac was a potent and selective in- hibitor of rabbit DHRS11. 2) Sulindac, zomepirac, indomethacin and fenoprofen inhibited rabbit DHRS11 more potently than human DHRS11. 3) Like flufenamic acid, ketoprofen and suprofen were more inhibitory to human DHRS11 than rabbit DHRS11.

Carbenoxolone and flufenamic acid are competitive inhibitors of human DHRS11, which is inhibited noncompetitively by flavonoids [3]. To compare the inhibitor sensitivity between rabbit and human DHRS11s, the inhibition patterns of the three-types of NSAIDs, tolbu- tamide, phenobarbital and cyclobarbital were examined in the NADP+- linked oxidation of 5-androstene-3β,17β-diol (for rabbit DHRS11) and 5α-androstane-3β,17β-diol (for human DHRS11 [3]). As the re- presentative results of sulindac are shown in Fig. 4, all the inhibitors
showed competitive inhibition with respect to the steroid substrates, suggesting that these structurally distinct inhibitors bind to the sub- strate-binding sites of the rabbit and human enzymes. The Ki values for the inhibitors are summarized in Table 5, which clearly demonstrates the difference in inhibitor sensitivity between the rabbit and human enzymes that was suggested by the IC50 determination. In addition to differences in substrate specificity between rabbit and human DHRS11s, the selective inhibition of rabbit DHRS11 by diclofenac, tolbutamide and barbiturates suggests that the substrate-binding sites of the rabbit and human enzymes are structurally distinct.

Fig. 4. Inhibition patterns of sulindac. The dehy- drogenase activities (mU/mL) towards 5-androstene- 3β,17β-diol (for rabbit DHRS11, A) and 5α-andros- tane-3β,17β-diol (for human DHRS11, B) in the absence and presence of sulindac were double-re- ciprocally plotted as a function of concentration of the substrate (S). Inset: the replot of the slopes of the double reciprocal plot versus sulindac concentra- tions: 0 (●), 1 (○), 2 (▲) and 3 (Δ) μM in (A); and 10 (●), 20 (▲) and 30 (Δ) μM in (B).

3.6. Alteration of substrate specificity and inhibitor sensitivity by mutagenesis

The above results suggested that the differences in substrate speci- ficity and inhibitor sensitivity between rabbit and human DHRS11s result from those in amino acids in their substrate-binding sites. Despite solving the crystal structure of human DHRS11 complexed with NADP+ and acetic acid (PDB-ID: 1XG5), its substrate-binding residues remain unclear. Therefore, we constructed a DHEA-docked model of human DHRS11 (Fig. 5), in which DHEA was surrounded by several residues including catalytically important Tyr166 and Ser151 that are shown by crystallographic studies of members of the SDR superfamily [1,2,32]. To clarify the differences in substrate-binding residues between rabbit DHRS11 and human DHRS11, we also prepared a homology model of rabbit DHRS11 based on the crystal structure of human DHRS11 that differs from rabbit DHRS11 by only 22 residues (Fig. 2). When the two models were superimposed (Fig. 5), five residues at positions 158, 160, 163, 200 and 224 were suggested to be different substrate-binding re- sidues between rabbit and human DHRS11s. Among them, Thr163 and Val200 are situated upper side (adjacent to the catalytic residue, Tyr166) and lower side, respectively, of the docked DHEA in human reductase activities were observed in both Thr163Gly and Val200Leu mutants, and the double TGVL mutant showed almost the same Km and Vmax values as those of rabbit DHRS11, suggesting that both Gly163 and Leu200 are necessary and sufficient for exerting this activity in rabbit DHRS11. In the reduction of menadione, the Thr163Gly mutation did not influence the kinetic constants, but the Val200Leu and double mutations abolished this reductase activity, suggesting that Leu200 is a critical residue that prevents the catalytically active binding of mena- dione to rabbit DHRS11.

Fig. 5. Superimposed active-site residues of DHEA-docked model of human DHRS11 and homology model of rabbit DHRS11. NADP+, DHEA and residues (within 4 Å from DHEA) of human DHRS11 are shown in yellow, green and blue, respectively. Among the residues, only five residues are different from those of rabbit DHRS11, which are shown in purple. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

We next compared the effects of the mutations on the kinetic con- stants for 3-pyridinecarboxaldehyde and steroids (5β-androstan-17β-ol- 3-one and 5-androstene-3β,17β-diol) that are reduced with high Vmax values by rabbit DHRS11 and human DHRS11, respectively (Table 6).
Compared to wild-type human DHRS11, the mutation of Thr163Gly, but not Val200Leu, increased the Vmax value for 3-pyridinecarbox- aldehyde by 5-fold, and the double mutation further elevated both the Vmax and Km values, which were similar to those of rabbit DHRS11. In the reduction of 5β-androstan-17β-ol-3-one, the mutations of Thr163Gly and Val200Leu induced opposite effects only on the Vmax
value, and the double TGVL mutation affected the Vmax and Km values (2-fold higher Vmax and almost the same Km compared to rabbit DHRS11). Similar effects of these mutations on the Vmax values were observed in the oxidation of 5-androstene-3β,17β-diol, although the kinetic constants of the double mutant were slightly higher than those of rabbit DHRS11. Collectively, the results of the mutations indicate that the two residues at positions 163 and 200 play critical roles in binding of small and nonsteroidal substrates and contribute to the dif- ference in the substrate specificity between rabbit and human DHRS11s, although additional residues may dedicate to binding of larger substrates (such as acetohexamide and steroids) to the enzymes. Furthermore, we examined the effects of the mutations on the po- tencies of rabbit DHRS11-specific (diclofenac, tolbutamide and phe- nobarbital) and potent (sulindac) inhibitors (Table 6). Between the two single mutations, only the Thr163Gly mutation changed human DHRS11 to a diclofenac-sensitive enzyme. The Val200Leu mutation induced high susceptibility to tolbutamide more significantly than the Thr163Gly mutation, and both single mutations potentiated the sensi- tivity to sulindac. Although the Ki values of the double TGVL mutant for the three inhibitors were similar to those of rabbit DHRS11, the con- tribution of the two residues at positions 163 and 200 to binding of the respective inhibitors may be different. In contrast to the above in- hibitors, phenobarbital was not inhibitory to the two single mutants and weakly inhibited the double mutant, suggesting that residue(s) other than Gly163 and Leu200 are involved in its binding to rabbit DHRS11.

4. Conclusion

The present study has revealed that rabbit DHRS11, an ortholog of human DHRS11, is an identical protein to RHAR and shows broad substrate specificities for aromatic ketones, quinones, aldehydes, α-dicarbonyl compounds and 3/17-ketosteroids, which raises the intri- guing possibility that the enzyme functions in the metabolism of xe- nobiotic and endogenous carbonyl compounds, as well as steroid me- tabolism. Additionally, we found that differences in substrate specificity and inhibitor sensitivity between the rabbit and human DHRS11s sharing high amino acid sequence identity, and result, at least in part, from residue differences at positions 163 and 200. The structural knowledge can promote development of selective inhibitors, which are useful tools to elucidate the physiological roles of human DHRS11 and its possible involvement in hormone-dependent cancers [3] in future studies.

Conflicts of interest

All authors declare that they have no conflicts of interest concerning this work.

Funding

This study was partially supported by JSPS Grant-in-Aid for Scientific Research (C) from the Japan Society for the Promotion of Science (MEXT KAKENHI Grant Number 17K11151).

Acknowledgement

We thank Dr. Akira Hara for insightful discussion and critical reading of the manuscript.

Transparency document

Transparency document related to this article can be found online at https://doi.org/10.1016/j.cbi.2019.03.026.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https:// doi.org/10.1016/j.cbi.2019.03.026.

References

[1] B. Persson, Y. Kallberg, J.E. Bray, E. Bruford, S.L. Dellaporta, A.D. Favia,
R.G. Duarte, H. Jörnvall, K.L. Kavanagh, N. Kedishvili, M. Kisiela, E. Maser,
R. Mindnich, S. Orchard, T.M. Penning, J.M. Thornton, J. Adamski, U. Oppermann, The SDR (short-chain dehydrogenase/reductase and related enzymes) nomen- clature initiative, Chem. Biol. Interact. 178 (2009) 94–98 https://doi: 10.1016/j.
cbi.2008.10.040.
[2] B. Persson, Y. Kallberg, Classification and nomenclature of the superfamily of short- chain dehydrogenases/reductases (SDRs), Chem. Biol. Interact. 202 (2013)
111–115 https://doi: 10.1016/j.cbi.2012.11.009.
[3] S. Endo, N. Miyagi, T. Matsunaga, A. Hara, A. Ikari, Human dehydrogenase/re- ductase (SDR family) member 11 is a novel type of 17β-hydroxysteroid dehy- drogenase, Biochem. Biophys. Res. Commun. 472 (2016) 231–236 https://doi: 10. 1016/j.bbrc.2016.01.190.
[4] N. Zila, A. Bileck, B. Muqaku, L. Janker, O.M. Eichhoff, P.F. Cheng, R. Dummer,
M.P. Levesque, C. Gerner, V. Paulitschke, Proteomics-based insights into mitogen- activated protein kinase inhibitor resistance of cerebral melanoma metastases, Clin. Proteonomics 15 (2018) 13 https://doi: 10.1186/s12014-018-9189-x.
[5] Y. Imamura, T. Koga, Y. Uriu, M. Otagiri, K. Satoh, A. Hara, Catalytic properties for naphthoquinones and partial primary structure of rabbit heart acetohexamide re- ductase, Biol. Pharm. Bull. 23 (2000) 155–158.
[6] Y. Imamura, A. Ryu, T. Koga, T. Higuchi, M. Otagiri, M. Nozawa, H. Akita,
Purification and catalytic properties of a novel acetohexamide-reducing enzyme from rabbit heart, J. Biochem. 119 (1996) 648–652.
[7] Y. Imamura, T. Koga, T. Migita, A. Ryu, M. Otagiri, Nozawa, H. Akita, Characterization of acetohexamide reductases purified from rabbit liver, kidney, and heart: structural requirements for substrates and inhibitors, J. Biochem. 121
(1997) 705–710.
[8] B. Gonzalez, A. Sapra, H. Rivera, W.D. Kaplan, B. Yam, G.L. Forrest, Cloning and expression of the cDNA encoding rabbit liver carbonyl reductase, Gene 154 (1995) 297–298.
[9] N. Usami, S. Ishikura, H. Abe, M. Nagano, M. Uebuchi, A. Kuniyasu, M. Otagiri,H. Nakayama, Y. Imamura, A. Hara, Cloning, expression and tissue distribution of a tetrameric form of pig carbonyl reductase, Chem. Biol. Interact. 143–144 (2003) 353–361.
[10] Y. Imamura, T. Migita, M. Otagiri, T. Choshi, S. Hibino, Purification and catalytic properties of a tetrameric carbonyl reductase from rabbit heart, J. Biochem. 125 (1999) 41–47.
[11] J. Sambrook, E.F. Fritsch, T. Maniatis, Molecular Cloning: A Laboratory Manual,
second ed., Cold Spring Harbor Laboratory Press, New York, 1989.
[12] S. Endo, T. Matsunaga, T. Kuragano, S. Ohno, Y. Kitade, K. Tajima, O. El-Kabbani,
A. Hara, Properties and tissue distribution of a novel aldo-keto reductase encoding in a rat gene (Akr1b10), Arch. Biochem. Biophys. 503 (2010) 230–237 https://doi: 10.1016/j.abb.2010.08.010.
[13] M.M. Bradford, A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding, Anal. Biochem. 72 (1976) 248–254.
[14] Y. Fukunishi, Y. Mikami, H. Nakamura, The filling potential method: A method for
estimating the free energy surface for protein−ligand docking, J. Phys. Chem. B 107 (2003) 13201–13210 https://doi: 10.1021/jp035478e.
[15] Y. Fukunishi, Y. Mikami, H. Nakamura, Similarities among receptor pockets and among compounds: analysis and application to in silico ligand screening, J. Mol. Graph. Model. 24 (2005) 34–45 https://doi:10.1016/j.jmgm.2005.04.004.
[16] H. Nguyen, T. Tran, Y. Fukunishi, J. Higo, H. Nakamura, L. Le, Computational study
of drug binding affinity to influenza A neuraminidase using smooth reaction path generation (SRPG) method, J. Chem. Inf. Model. 55 (2015) 1936–1943 https:// doi:10.1021/acs.jcim.5b00319.
[17] M. Biasini, S. Bienert, A. Waterhouse, K. Arnold, G. Studer, T. Schmidt, F. Kiefer,
T. Gallo Cassarino, M. Bertoni, L. Bordoli, T. Schwede, SWISS-MODEL: modelling protein tertiary and quaternary structure using evolutionary information, Nucleic Acids Res. 42 (2014) W252–W258 https://doi: 10.1093/nar/gku340.
[18] T. Shibamoto, Diacetyl: occurrence, analysis, and toxicity, J. Agric. Food Chem. 62
(2014) 4048–4053 https://doi: 10.1021/jf500615u.
[19] A. Medvedev, O. Buneeva, O. Gnedenko, P. Ershov, A. Ivanov, Isatin, an en- dogenous nonpeptide biofactor: a review of its molecular targets, mechanisms of actions, and their biomedical implications, Biofactors 44 (2018) 95–108 https://
doi: 10.1002/biof.1408.
[20] H. Esterbauer, R.J. Schaur, H. Zollner, Chemistry and biochemistry of 4-hydro- xynonenal, malonaldehyde and related aldehydes, Free Radic. Biol. Med. 11 (1991) 81–128.
[21] L.W. Kroh, T. Fiedler, J. Wagner, α-Dicarbonyl compounds–key intermediates for the formation of carbohydrate-based melanoidins, Ann. N. Y. Acad. Sci. 1126 (2008) 210–215 https://doi.org/10.1196/annals.1433.058.
[22] R.L. Felsted, D.R. Richter, D.M. Jones, N.R. Bachur, Isolation and characterization of rabbit liver xenobiotic carbonyl reductases, Biochem. Pharmacol. 29 (1980) 1503–1516.
[23] H. Sawada, A. Hara, T. Nakayama, F. Kato, Reductases for aromatic aldehydes and
ketones from rabbit liver. Purification and characterization, J. Biochem. 87 (1980) 1153–1165.
[24] S. Endo, T. Matsunaga, A. Matsumoto, Y. Arai, S. Ohno, O. El-Kabbani, K. Tajima,
Y. Bunai, S. Yamano, A. Hara, Y. Kitade, Rabbit 3-hydroxyhexobarbital dehy- drogenase is a NADPH-preferring reductase with broad substrate specificity for ketosteroids, prostaglandin D₂, and other endogenous and xenobiotic carbonyl
compounds, Biochem. Pharmacol. 86 (2013) 1366–1375 https://doi: 10.1016/j.
bcp.2013.08.024.
[25] S. Endo, T. Matsunaga, Y. Arai, A. Ikari, K. Tajima, O. El-Kabbani, S. Yamano,
A. Hara, Y. Kitade, Cloning and characterization of four rabbit aldo-keto reductases featuring broad substrate specificity for xenobiotic and endogenous carbonyl compounds: relationship with multiple forms of drug ketone reductases, Drug
Metab. Dispos. 42 (2014) 803–812 https://doi: 10.1124/dmd.113.056044.
[26] S. Endo, T. Matsunaga, S. Kumada, A. Fujimoto, S. Ohno, O. El-Kabbani, D. Hu,
N. Toyooka, J. Mano, K. Tajima, A. Hara, Characterization of rabbit aldose re- ductase-like protein with 3β-hydroxysteroid dehydrogenase activity, Arch. Biochem. Biophys. 527 (2012) 23–30 https://doi: 10.1016/j.abb.2012.07.012.
[27] R.A. Prough, B.J. Clark, C.M. Klinge, Novel mechanisms for DHEA action, J. Mol.
Endocrinol. 56 (2016) R139–R155 https://doi: 10.1530/JME-16-0013.
[28] L. Stárka, M. Dušková, M.S. Hill, Dehydroepiandrosterone: a neuroactive steroid, J. Steroid Biochem. Mol. Biol. 145 (2015) 254–260 https://doi: 10.1016/j.jsbmb.
2014.03.008.
[29] J. Poortman, J.A. Prenen, F. Schwarz, J.H. Thijssen, Interaction of delta-5-andros- tene-3β,17β-diol with estradiol and dihydrotestosterone receptors in human myo- metrial and mammary cancer tissue, J. Clin. Endocrinol. Metab. 40 (1975) 373–379 https://doi:10.1210/jcem-40-3-373.
[30] M.E. Baker, K.Y. Uhb, C. Chandsawangbhuwana, 3D models of human ERα and ERβ complexed with 5-androsten-3β,17β-diol, Steroids 77 (2012) 1192–1197 https:// doi: 10.1016/j.steroids.2012.07.014.
[31] E. Roussel, M.-C. Drolet, A.-M. Lavigne, M. Arsenault, J. Couet, Multiple short-chain dehydrogenases/reductases are regulated in pathological cardiac hypertrophy, FEBS Open Bio 8 (2018) 1624–1635 https://doi: 10.1002/2211-5463.12506.
[32] C. Filling, K.D. Berndt, J. Benach, S. Knapp, T. Prozorovski, E. Nordling,
R. Ladenstein, H. Jörnvall, U. Oppermann, Critical residues for structure and cat- alysis in short-chain dehydrogenases/reductases, J. Biol. Chem.BAY 2402234 277 (2002) 25677–25684 https://doi: 10.1074/jbc.M202160200.