Characterization of hamster NAD+-dependent 3(17)β-hydroxysteroid dehydrogenase belonging to the aldo-keto reductase 1C subfamily

Article abstract of DOI:10.1093/jb/mvv057

The cDNAs for morphine 6-dehydrogenase (AKR1C34) and its homologous aldo-keto reductase (AKR1C35) were cloned from golden hamster liver, and their enzymatic properties and tissue distribution were compared. AKR1C34 and AKR1C35 similarly oxidized various xenobiotic alicyclic alcohols using NAD⁺, but differed in their substrate specificity for hydroxysteroids and inhibitor sensitivity. While AKR1C34 showed 3α/17β/20α-hydroxysteroid dehydrogenase activities, AKR1C35 efficiently oxidized various 3β- and 17β-hydroxysteroids, including biologically active 3β-hydroxy-5α/β-dihydro-C₁₉/C₂₁-steroids, dehydroepiandrosterone and 17β-estradiol. AKR1C35 also differed from AKR1C34 in its high sensitivity to flavonoids, which inhibited competitively with respect to 17β-estradiol (K_i 0.11-0.69 μM). The mRNA for AKR1C35 was expressed liver-specific in male hamsters and ubiquitously in female hamsters, whereas the expression of the mRNA for AKR1C34 displayed opposite sexual dimorphism. Because AKR1C35 is the first 3(17)β-hydroxysteroid dehydrogenase in the AKR superfamily, we also investigated the molecular determinants for the 3β-hydroxysteroid dehydrogenase activity by replacement of Val54 and Cys310 in AKR1C35 with the corresponding residues in AKR1C34, Ala and Phe, respectively. The mutation of Val54Ala, but not Cys310Phe, significantly impaired this activity, suggesting that Val54 plays a critical role in recognition of the steroidal substrate.

Full text of DOI:10.1093/jb/mvv057

Journal of Biochemistry Advance Access published June 17, 2015
J. Biochem. 2015;1—10 doi:10.1093/jb/mvv057
Characterization of hamster NAD+-dependent
3(17)b-hydroxysteroid dehydrogenase belonging
to the aldo-keto reductase 1C subfamily
Received February 19, 2015; accepted April 28, 2015; published online May 22, 2015
Satoshi Endo1, Misato Noda1, Akira Ikari1,
Kenjiro Tatematsu2, Ossama El-Kabbani3,
Akira Hara4, Yukio Kitade4 and
spectrometry; M6DH, morphine 6-dehydrogenase;
RT, reverse transcription; SDR, short-chain
dehydrogenase/reductase; TLC, thin-layer
chromotography.
Toshiyuki Matsunaga1,*
1Laboratory of Biochemistry and 2Laboratory of Radiochemistry,
Gifu Pharmaceutical University, Gifu 501-1196, Japan; 3Nagoya
University Graduate School of Medicine, 65 Tsurumai-cho, Showa-ku,
Nagoya 466-8550, Japan; and 4Faculty of Engineering, Gifu
University, Gifu 501-1193, Japan
Hydroxysteroid dehydrogenases (HSDs) catalyze the
oxidoreduction between hydroxysteroids and ketoster-
oids using NAD+ or NADPH as the coenzyme, and
play important roles in the activation/inactivation of
all classes of steroid hormones (1). Mammalian HSDs
have been structurally divided into two protein super-
families, the short-chain dehydrogenase/reductases
(SDR) (2) and the aldo-keto reductases (AKR) (3).
The SDR superfamily members are microsomal
3a-HSD, Á5-3b-hydroxysteroid dehydrogenase/Á5-Á4
isomerase (Á5-3b-HSDI), 11b-HSD and most isozymes
of 17b-HSD that are distributed in various subcellular
fractions (4). In addition, hamster mitochondrial P26h
protein (5), human peroxisomal dehydrogenase/
reductase (SDR family) member 4 (6) and pig cytosolic
carbonyl reductase (7), belonging to this superfamily,
*Toshiyuki Matsunaga, Laboratory of Biochemistry, Gifu
Pharmaceutical University, Gifu 501-1196, Japan. Tel: +81-58-230-
8100, Fax: +81-58-230-8105, email: matsunagat@gifu-pu.ac.jp
The cDNAs for morphine 6-dehydrogenase (AKR1C34)
and its homologous aldo-keto reductase (AKR1C35)
were cloned from golden hamster liver, and their
enzymatic properties and tissue distribution were com-
pared. AKR1C34 and AKR1C35 similarly oxidized
various xenobiotic alicyclic alcohols using NAD+, but
differed in their substrate specificity for hydroxyster-
oids and inhibitor sensitivity. While AKR1C34 showed
3a/17b/20a-hydroxysteroid dehydrogenase activities,
AKR1C35 efficiently oxidized various 3b- and
17b-hydroxysteroids, including biologically active
3b-hydroxy-5a/b-dihydro-C19/C21-steroids, dehydroe-
piandrosterone and 17b-estradiol. AKR1C35 also dif-
fered from AKR1C34 in its high sensitivity to
flavonoids, which inhibited competitively with respect
to 17b-estradiol (Ki 0.11—0.69 kM). The mRNA for
AKR1C35 was expressed liver-specific in male ham-
sters and ubiquitously in female hamsters, whereas the
expression of the mRNA for AKR1C34 displayed
opposite sexual dimorphism. Because AKR1C35 is the
first 3(17)b-hydroxysteroid dehydrogenase in the AKR
superfamily, we also investigated the molecular deter-
minants for the 3b-hydroxysteroid dehydrogenase activ-
ity by replacement of Val54 and Cys310 in AKR1C35
with the corresponding residues in AKR1C34, Ala and
Phe, respectively. The mutation of Val54Ala, but not
Cys310Phe, significantly impaired this activity, sug-
gesting that Val54 plays a critical role in recognition
of the steroidal substrate.
exhibit
NAD+-dependent 3a-HSD, NADPH-
dependent reductive 3b-HSD, and 3a/3b/20b-HSD
activities, respectively. The AKR superfamily includes
cytosolic NADPH-dependent 3a-HSD, 17a-HSD,
17b-HSD and 20a-HSD, most of which exist in mul-
tiple forms and show overlapped steroid specificities
(3). For example, human 20a-HSD (AKR1C1), 3a-
HSD type-3 (AKR1C2), 17b-HSD type-5 (AKR1C3)
and 3a-HSD type-1 (AKR1C4) exhibit reductive 3a-,
17b- and 20a-HSD activities (8), and also non-
stereospecifically
reduce
5a-dihydrotestosterone
(DHT) into its 3a- and 3b-hydroxy metabolites (9).
In addition to the NADPH-dependent HSDs,
NAD+-dependent enzymes that oxidize 3a-, 17b-
and/or 20a-hydroxysteroids are identified in rats
(AKR1C16, AKR1C17 and AKR1C24) and mice
(AKR1C12 and AKR1C13) (10—13). Recently, rabbit
aldose reductase-like protein (AKR1B19) that reduces
various aldehydes has been reported to exhibit
NADPH-dependent reductive 3b-HSD activity to-
wards 3-keto-C19/C21-steroids (14). However, no
NAD+-dependent AKR with 3b- or 17a-HSD activity
has yet been identified.
Keywords: aldo-keto reductase/flavonoid/3(17)b-
hydroxysteroid dehydrogenase/morphine 6-dehydro-
genase/sexually dimorphic expression.
Abbreviations: AKR, aldo-keto reductase; DHT,
dihydrotestosterone; 5a,3a-DIOL, 5a-androstane-
3a,17b-diol; 5a,3b-DIOL, 5a-androstane-3b,17b-diol;
HSD, hydroxysteroid dehydrogenase; Á5-3b-HSDI,
Á5-3b-hydroxysteroid dehydrogenase/Á5-Á4 isomer-
ase; LC/MS, liquid chromatography/mass
We previously characterized three rabbit NAD+-
dependent HSDs (AKRs: 1C26, 1C27 and 1C28) that
belong to the AKR superfamily (15). The rabbit HSDs
show broad substrate specificity for 3a-, 17b- and/or
20a-hydroxysteroids and xenobiotic alcohols, and
ß The Authors 2015. Published by Oxford University Press on behalf of the Japanese Biochemical Society. All rights reserved
1

S. Endo et al.
Purification of enzymes
AKR1C26 is identified as morphine 6-dehydrogenase
(M6DH). M6DH with 17b-HSD activity belonging to
the AKR superfamily was purified from hamster liver
(16), but isolation of the corresponding complete
cDNA has not been reported previously. In this
study, we have isolated cDNAs for two AKRs that
share >86% amino acid sequence identity with the
partial sequence of hamster liver M6DH (16). The pro-
teins encoded in the genes have been assigned as
AKR1C34 and AKR1C35 in the AKR superfamily
(http://www.med.upenn.edu/akr/). Although AKR1C34
was identical to hamster liver M6DH, AKR1C35 was
found to be 3(17)b-HSD, which is the first NAD+-de-
pendent member with 3b-HSD activity in the AKR
superfamily. We describe the enzymatic properties
and tissue distribution of the two hamster AKRs. In
addition, the molecular determinant for 3(17)b-HSD
activity was investigated by site-directed mutagenesis
studies of AKR1C35.
The recombinant AKR1C35, its mutant enzymes, and AKR1C34
were expressed in E. coli BL21 (DE3) pLysS cells transformed with
the expression plasmids harbouring their cDNAs as described pre-
viously (19). The enzymes were purified from the cell extracts using a
nickel-charged Sepharose 6FF resin (GE Healthcare Ltd,
Amersham, UK) according to the manufacturer’s manual. The
enzyme fraction was concentrated by ultrafiltration and
dialyzed against 10 mM Tris-HCl buffer, pH 8.0, containing 5 mM
2-mercaptethanol, 0.5 mM EDTA and 20% (v/v) glycerol. Purity
was confirmed by sodium dodecyl sulfate-polyacrylamide gel electro-
phoresis (SDS-PAGE), and protein concentration was determined by
Bradford’s method using bovine serum albumin as the standard (20).
Assay of enzyme activity
The dehydrogenase activities for the enzymes were assayed by mea-
suring the rate of change in fluorescence (at 455 nm with an excita-
tion wavelength of 340 nm) or absorbance (at 340 nm) of NAD(P)H.
The standard reaction mixture consisted of 0.1 M potassium
phosphate buffer, pH 7.4, 1.0 mM NAD+
, 1.0 mM S-(+)-
1,2,3,4-tetrahydro-1-naphthol (S-tetralol) and enzyme, in a total
volume of 2.0 ml. The reductase activities were determined by mea-
suring the rate of change in NAD(P)H absorbance in the phosphate
buffer, pH 7.4, containing 0.1 mM NADH or NADPH and an
appropriate amount of carbonyl substrate. The steroids and other
substrates, which are hardly soluble in water, were dissolved in
methanol or 50% methanol, and added into the reaction mixture,
in which the final concentration of methanol was 52.5%. The con-
centration of methanol did not affect the activity of the enzymes.
One unit of enzyme activity was defined as the amount of enzyme
that catalyzes the formation or oxidation of 1 mmol NAD(P)H per
min at 25ꢀC.
The pH dependency of the enzyme activity was estimated with
0.1 M potassium phosphate (pH 5.8—8.0), Tris-HCl (pH 7.5—9.0)
and glycine-NaOH (pH 8.5—11) buffers. The apparent Km and kcat
values were determined over a range of five substrate concentrations
at a saturating concentration of coenzyme by fitting the initial velo-
cities to the Michaelis—Menten equation. The kinetic constants for
NAD(P)+ and NADH were determined in the presence of saturated
concentrations of S-tetralol and S-camphorquinone, respectively, as
substrates. The IC50 (inhibitor concentrations required for 50%
inhibition) values were determined with 20 mM S-tetralol
(for AKR1C34) or 5 mM 17b-estradiol (for AKR1C35) as the sub-
strate in the standard reaction mixture. The inhibitor constant, Ki,
was estimated from the Dixon plot and/or Cornish—Bowden plot of
the velocities obtained in the 17b-estradiol range (1—30ꢁKm) with
three concentrations of the inhibitor. The kinetic constants and IC50
are expressed as the means SD of three determinations.
Experimental Procedures
Materials
Steroids were obtained from Sigma Chemicals (Perth, WA) and
Steraloids (Newport, RI). A pCold I expression vector was pur-
chased from Takara (Kusatsu, Japan); Pfu DNA polymerase was
from Stratagene; KOD FX DNA polymerase was from Toyobo
(Osaka, Japan); and pCR2.1 plasmid, restriction enzymes, RACE
kit and Escherichia coli BL21 (DE3) pLysS were from Invitrogen
(Carlsbad, CA). trans-Benzene dihydrodiol was synthesized as
described (17). a- and b-3-Hydroxyhexobarbitals were kindly
denoted by Dr R. Takenoshita (Fukuoka University, Japan). All
other chemicals were of the highest grade that could be obtained
commercially.
cDNA isolation and site-directed mutagenesis
The cDNAs for AKR1C34 and AKR1C35 were isolated from the
total RNA preparation of a liver of male golden hamster (Japan
SLC Inc., Shizuoka, Japan) by reverse transcription (RT)-PCR.
The preparation of total RNA, RT and DNA techniques followed
the standard procedures described by Sambrook et al. (18). PCR was
performed with Pfu DNA polymerase and a pair of sense and anti-
sense primers, which contain NdeI and SalI sites. The primer
sequences and PCR conditions are summarized in Supplementary
Table S1. The PCR products were purified, digested with the two
restriction enzymes and ligated into the pCold I vectors that had
been digested with the two restriction enzymes. The insert of the
cloned cDNA was sequenced using a Beckman CEQ8000XL DNA
sequencer, and was confirmed to encode the 324-amino acid
sequences of AKR1C34 and AKR1C35 fused to the N-terminal
6-His tag. The 30-untranslated regions of the mRNAs for the two
enzymes were generated using the RACE kit and the gene-specific
primers. The fragments were subcloned into the pCR2.1 plasmids
and sequenced as described earlier. The sequences of the cDNAs for
AKR1C34 (1,192 bp) and AKR1C35 (1,203 bp) were deposited in
the DDBJ database with the accession numbers LC006027 and
LC006028, respectively.
Mutagenesis was performed using a QuickChange site-directed
mutagenesis kit (Stratagene) and the pCold I expression plasmid
harbouring the cDNA for AKR1C35 as the template according to
the protocol described by the manufacturer. The primer pair used for
the mutagenesis was composed of sense and antisense oligonucleo-
tides to give V54A and C310F mutant enzymes (Supplementary
Table S1). The cDNA for the double mutant of V54A/C310F was
prepared using the expression plasmid harbouring the cDNA for
V54A mutant as the template. The coding regions of the cDNAs
in the expression plasmids were sequenced to confirm the presence
of the desired mutation and ensure that no other mutation had
occurred.
Product identification
To identify products, reaction was conducted in a 2.0-ml system
containing 1.0 mM NAD+ or 0.2 mM NADH, substrate
(20—50 mM), enzyme (30—200 mg) and 0.1 M potassium phosphate,
pH 7.4. The substrate and products were extracted into 4 ml ethyl
acetate 20 min after the reaction started at 37ꢀC. The steroidal prod-
ucts were identified by thin-layer chromatography (TLC) and liquid
chromatography/mass spectrometry (LC/MS) using
a Chiralcel
OJ-H 5 mm column (21). The TLC chromatogram was developed
in chloroform—ethyl ether (9:1, v/v), chloroform—ethyl ether—acetic
acid (90:10:1, v/v) and chloroform—acetone—aqueous ammonia
(70:30:1, v/v). The products and co-chromatographed authentic ster-
oids were visualized by spraying with ethanol/H2SO4 (1:1, v/v)
solution and heating at 110ꢀC for 1 h. In LC/MS analysis,
3b-hydroxyprogesterone and progesterone were separated at a flow
rate of 0.5 ml/min using hexane/isopropanol (90:10) as the mobile
phase. The flow rate and mobile phase were 1.0 ml/min and hexane/
isopropanol (85:15), respectively, for the separation of 17b-estradiol
and estrone, and were 0.4 ml/min and hexane/isopropanol (95:5),
respectively for the separation of 5a-pregnan-20a-ol-3-one and 5a-
pregnane-3,20-dione. The steroids were detected by total ion moni-
toring mode or single ion monitoring of their expected molecular or
fragment ion (positive ion mode). The products were also identified
by comparing their retention times with those of their authentic
steroids.
2

Hamster 3(17)b-hydroxysteroid dehydrogenase
Tissue distribution analysis
pH optimum and coenzyme specificity
The total RNA samples were prepared from the tissues of 6-week-
old male and female hamsters (18). The total RNA samples were
subjected to RT-PCR using KOD FX DNA polymerase and the
primers were employed for the cloning of the cDNAs for
AKR1C34 and AKR1C35. The cDNA for hamster b-actin was
also amplified as an internal control with the specific primers (22).
The PCR products were separated by agarose gel electrophoresis and
revealed with ethidium bromide.
The purified preparations of the recombinant
AKR1C34 and AKR1C35 showed single 37-kDa pro-
tein bands on the SDS-PAGE analysis. The enzymes
exhibited NAD+-linked dehydrogenase activity to-
wards S-tetralol that is an excellent substrate for the
above NAD+-dependent HSDs of other species
(11—13, 15). The specific activities of AKR1C34 and
AKR1C35 determined at pH 7.4 were 0.039 and
0.23 U/mg, respectively. The pH optima for the activ-
ity of AKR1C34 and AKR1C35 were observed at
pH 10.0 and 9.0, respectively, where the enzyme activ-
ities were higher about 3- and 2-fold, respectively, than
those assayed at pH 7.4. As Gln270 and Glu276 that
are critical residues for the NAD+ specificity (10) were
conserved in AKR1C34 and AKR1C35 (Fig. 1), the
enzymes showed high preference for NAD+ compared
with NADP+. The kcat/Km values for NAD+ in the
S-tetralol oxidation by the two enzymes were
much higher than those for NADP+ (Table I). In the
reverse reaction, AKR1C34 and AKR1C35 reduced
S-camphorquinone using NADH as the coenzyme,
but their reductase activities determined with 0.1 mM
NADPH as the coenzyme were >17% of their
NADH-linked activities. Therefore, the substrate
specificity of the two enzymes was examined using
the preferable coenzyme NAD+ at a physiological
pH of 7.4.
Results
cDNA cloning and sequence comparison
The cDNAs for AKR1C34 and AKR1C35 were iso-
lated by RT-PCR from the total RNA sample of ham-
ster liver with primers that were designed based on the
sequences of mRNAs for Mesocricetus auratus AKR
family 1 member C13-like transcripts (accession no.
XM_005071750 and XM_005071747, respectively)
predicted from the genomic analysis in the NCBI data-
base. The nucleotide sequences of the coding regions of
the isolated cDNAs for AKR1C34 and AKR1C35
(each 972 bp) were identical to those of the transcripts
predicted from the genomic analysis, with the excep-
tion of a nucleotide at position 679 in AKR1C35
cDNA. Although this is deposited as an unknown
nucleotide in XM_005071747, it was found to be
T so that the encoded amino acid was Cys227
(Fig. 1). The amino acid sequence identity of
AKR1C34 and AKR1C35 was 88%. The sequence of
AKR1C34 was identical to those (composed a total of
303 residues) of peptides derived from the purified
hamster liver M6DH (16), with the exception of four
replacements (Ser34!Ala, Leu35!Met, His170!
Met and Val244!Gly). The two proteins exhibited
high sequence identity (83—87%) with NAD+-depend-
ent HSDs of rats (AKRs: 1C16, 1C17 and 1C24) and
mice (AKRs: 1C12 and 1C13), followed by
rabbit NAD+-dependent HSDs (AKRs: 1C26, 1C27
and 1C28; 70—72%), but shared lower sequence
Table I. Kinetic constants for coenzymes.
Km
Coenzyme (mM)
kcat
(minꢂ1
kcat/Km
(minꢂ1 mMꢂ1
)
Enzyme
)
AKR1C35 NADP+
NAD+
2,400 530 9.0 2.0 0.0038
23
17
3
2
6.8 0.3 0.30
6.2 0.9 0.36
NADHa
AKR1C34 NADP+
NAD+
1,380 450 5.5 0.2 0.004
13
13
2
2
1.5 0.2 0.12
10 0.76
NADHa
1
identity
(567%)
with
NADP(H)-dependent
aThe values for NADH were determined in the reduction of
human HSDs (AKRs: 1C1—1C4) and hamster 3-
hydroxyhexobarbital
HSD (23).
dehydrogenase/3a(17b)-
S-camphorquinone, whose concentrations were 0.1 mM and 40 mM
for AKR1C35 and AKR1C34, respectively.
Fig. 1 Alignment of amino acid sequences of AKRs (1C34 and 1C35). In the 1C34 sequence, the hyphens represent identical residues to those of
AKR1C35, and the underlined residues are identical to those determined by peptide sequencing of hamster liver M6DH. The determinants for
coenzyme specificity for NAD(H) are shown with asterisks. The positions of substrate-binding residues in AKRs (1C1, 1C2, 1C3, 1C5, 1C9 and
1C21) are indicated with arrowheads. The mutated residues of AKR1C35 are in boldface.
3

S. Endo et al.
Substrate specificity
5-androstene-3b,17b-diol and 5b-androstane-3b,17b-
diol were oxidized only by AKR1C35. The product
of 5a,3a-DIOL oxidation was identified as 5a-andro-
stan-3a-ol-17-one by TLC, whereas the oxidation of
5a,3b-DIOL produced 5a-DHT and 5a-androstan-
3b-ol-17-one as the major and minor products, respect-
ively, indicating that AKR1C35 mainly oxidizes the
3b-hydroxy group of 5a,3b-DIOL.
AKR1C34 and AKR1C35 oxidized non-steroidal ali-
cyclic and long-chain aliphatic alcohols (Table II),
whose structures are shown in Supplementary Fig.
S1. There were some differences in the kinetic con-
stants between the enzymes. AKR1C35 oxidized
b-3-hydroxyhexobarbital more efficiently than its
a-isomer, whereas AKR1C34 showed the opposite spe-
cificity for the two isomers. Among the substrates, an
endogenous aliphatic alcohol, geranylgeraniol, was the
most excellent substrate for AKR1C35, showing the
highest kcat/Km value that was also higher than that
of AKR1C34. In contrast, AKR1C34 exhibited lower
Km and higher catalytic efficiency (kcat/Km) for other
alicyclic and aliphatic alcohols than AKR1C35.
The two hamster AKRs showed marked differences
in the specificity for hydroxysteroids (Table III). The
most striking characteristic of AKR1C35 was the oxi-
dation of 3b-hydroxysteroids, which were not oxidized
by AKR1C34. The 3b-hydroxysteroid substrates
included 3b-hydroxyprogesterone, dehydroepiandros-
terone and 3b-hydroxy-5a/b-dihydro-C19/C21-steroids,
and the oxidized products of 3b-hydroxyprogesterone
and 5a-androstan-3b-ol-17-one were identified as pro-
gesterone and 5a-androstane-3,17-dione, respectively,
by LC/MS (Fig. 2) and TLC, respectively. The
enzyme did not show significant activity for
3b-hydroxy derivatives of 5-pregnenes (pregnenolone
and 17-hydroxypregnenolone), bile acids (isolitho-
cholic acid and 5a-cholic acid-3b,7b-diol) and
5a-cholestan-3b-ol. Although 17b-hydroxysteroids
were common substrates of the two AKRs, the high
reactivity of AKR1C35 to 17b-estradiol was re-
markable, as the kcat/Km value for 17b-estradiol
was the highest among the values for all
substrates. The product of 17b-estradiol oxi-
dation was identified as estrone (Fig. 2). Among 3,17-
dihydroxysteroids, 5a-androstane-3a,17b-diol (5a,
3a-DIOL), 5a-androstane-3b,17b-diol (5a,3b-DIOL),
As hamster liver M6DH exhibited 17b-HSD activity
(16), AKR1C34 oxidized 17b-hydroxysteroids other
than 17b-estradiol, 5a,3b-DIOL, 5-androstene-
3b,17b-diol,
5a,3a-DIOL
and
5b-androstane-
3b,17b-diol (Table III). In addition, AKR1C34
exhibited dehydrogenase activity towards several 3a-
and 20a-hydroxysteroids, which were not oxidized by
AKR1C35. The oxidation of 5a-pregnan-20a-ol-3-one
into the corresponding 20-ketosteroid was confirmed
by product analysis using LC/MS (Fig. 2). It should
be noted that AKR1C34 and AKR1C35 did not show
significant activity towards other 3a-hydroxysteroids
(5a-pregnan-3a-ol-20-one, 5a-androstan-3a-ol-17-one
and lithocholic acid), 17a-hydroxysteroids (17a-
estradiol and epitestosterone) and 20-hydroxysteroids
(5b-pregnan-20a-ol-3-one, 20b-hydroxyprogesterone
and 5a/b-pregnan-20b-ol-3-ones).
In the NADH-linked reverse reaction, AKR1C35
and AKR1C34 highly reduced non-steroidal
a-dicarbonyl compounds (9,10-phenanthrenequione,
isatin, 1-phenyl-1,2-propanedione, S-camphorquinone
and methylglyoxal), and also exhibited moderate activ-
ities towards aromatic aldehydes (4-nitrobenzaldehyde
and pyridine-3-aldehyde), a-tetralone and methyl-
glyoxal (Supplementary Table S2). The two enzymes
showed significant difference in the specificity for
ketosteroids. AKR1C34 showed low activity towards
5b-pregnane-3,20-dione but not 3- and 17-ketosteroids
(5a-androstane-3,17-dione, 5a-DHT, 5b-DHT and
estrone). In contrast, AKR1C35 did not reduce
5b-pregnane-3,20-dione and accepted the 3- and
Table II. Substrate specificity for non-steroidal alcohols.
AKR1C35
AKR1C34
a
Km
a
a
a
Km
a
a
Substrate
kcat
kcat/Km
kcat
kcat/Km
Alicyclic alcohols
b-3-Hydroxyhexobarbital
a-3-Hydroxyhexobarbital
1-Acenaphthenol
S-Tetralol
S-1-Indanol
R-Tetralol
trans-Benzene dihydrodiol
cis-Benzene dihydrodiol
2-Cyclohexen-1-ol
Cyclohexanol
13
30
148
61
181
95
1,000
630
569
5.8
2.4
17
6.3
6.0
0.45
2.7*
1.9
0.45
0.13
0.11
0.10
0.033
0.005
0.003*
0.003*
0.003*
40
14
23
1.5
1.2
28
11
9.7
26
74
1.5
1.1
1.3
1.5
1.7
1.3
2.0
1.5
2.4
0.038
0.078
0.057
1.0
1.4
0.047
0.18
0.15
0.093
0.007*
1.7
ns
0.54*
Aliphatic alcohols
Geranylgeraniol
b-Ionol
Farnesol
Geraniol
Nerol
1.2
9.7
11
76
162*
2.3
3.2
0.27
0.92
0.69*
2.0
0.32
0.025
0.012*
0.004*
1.0
0.46
1.3
1.1
1.4
1.4
0.46*
1.3
0.85
0.34
0.13
1.0
1.2
4.1
11
a
Km (mM), kcat (minꢂ1) and kcat/Km (minꢂ1mMꢂ1) were determined with 1.0 mM NAD+ as the coenzyme. The experimental errors were less
than 15%, except that those indicated by asterisks were 15—20%. ns, no significant activity (>0.05 minꢂ1).
4

Hamster 3(17)b-hydroxysteroid dehydrogenase
17-ketosteroids as substrates. The product of the
5a-androstane-3,17-dione reduction was identified as
5a-androstan-3b-ol-17-one by TLC.
kidney, adrenal gland and small intestine. In the female
hamster, the mRNA for AKR1C35 was expressed in
many tissues except for the colon. In contrast, opposite
sexual dimorphism was observed in the expression of
the mRNA for AKR1C34. In the male hamster, the
mRNA for AKR1C34 was expressed in many tissues,
of which the liver showed high expression. In the
female hamster, it was predominantly detected in the
liver.
Inhibitor sensitivity
AKR1C35 and AKR1C34 also differed in inhibitor
sensitivity (Table IV). The dehydrogenase activity of
AKR1C35 was inhibited by flavonoids, lithocholic
acid and zearalenone, which did not show potent
inhibitory effects on AKR1C34. AKR1C34 was select-
ively inhibited by phenolphthalein and hinokitiol,
which have been reported to be inhibitors of rabbit
M6DH (AKR1C26) (15). The inhibition patterns of
AKR1C35 by these inhibitors were all competitive
inhibitors with respect to the alcohol substrate,
17b-estradiol. The inhibitory flavonoids are
7-hydroxyflavones, with the exception of genistein
that is an isoflavone. Flavone without hydroxyl
group and daidzein, a des-5-hydroxy derivative of gen-
istein, did not inhibit at a high concentration of 20 mM,
suggesting that the 7-hydroxy group of flavones and
5-hydroxy group of isoflavones are structural requis-
ites for binding to the enzyme.
Effect of mutagenesis on steroid specificity of
AKR1C35
The most significant difference between AKR1C35 and
AKR1C34 was the oxidation of 3b-hydroxysteroids
and 17b-estradiol, despite their high sequence identity,
as described earlier. Previous crystallographic studies
of several members of the AKR1C subfamily (24—29)
show 16 substrate-binding residues, of which five resi-
dues at positions 54, 118, 227, 308 and 310 are different
between AKR1C35 and AKR1C34 (Fig. 1). Among
them, the residues 54 and 310 are situated inside (ad-
jacent to the catalytic residue, Tyr55) and outside, re-
spectively, of the active sites of known AKR crystal
structures. In addition, Val54 and Cys310 in
AKR1C35 are Ala and Phe, respectively, in
AKR1C34 and other AKRs (1C12, 1C13, 1C16,
1C17, 1C24, 1C26, 1C27 and 1C28) with NAD+-de-
pendent 3a- and/or 3a/17b-HSD activity. To examine
whether the two residues are responsible for the unique
steroid specificity of AKR1C35, we prepared two
single mutant AKR1C35s (V54A and C310F) and a
Tissue distribution
The tissue distribution of AKR1C35 in male and
female hamsters was assessed by RT-PCR, and
compared with that of AKR1C34 (Fig. 3). The expres-
sion of the mRNA for AKR1C35 was markedly
sex-different: In the male hamster, its expression was
almost liver-specific, and was much lower in the brain,
Table III. Substrate specificity for hydroxysteroids.
AKR1C35
AKR1C34
a
a
a
a
Km
a
a
Substrate
Km
kcat
kcat/Km
kcat
kcat/Km
3ꢀ-Hydroxysteroids
3a-Hydroxyprogesterone
5b-Pregnan-3a-ol-20-one
5b-Androstan-3a-ol-17-one
3ꢁ-Hydroxysteroids
ns
na
ns
0.5
1.0
0.5
0.35
0.12
0.058*
0.70
0.12
0.12*
3b-Hydroxyprogesterone
5a-Androstan-3b-ol-17-one
5a-Pregnan-3b-ol-20-one
Dehydroepiandrosterone
5b-Androstan-3b-ol-17-one
5b-Pregnan-3b-ol-20-one
5b-Pregnane-3b,21-diol-20-one
17ꢁ-Hydroxysteroids
17b-Estradiol
0.9
2.5
3.9
4.3
46
3.4*
1.6
3.8*
0.64
0.20
0.14
0.028*
0.025*
0.017*
ns
na
na
na
na
na
na
0.78
0.60
1.3*
0.17*
0.26
6.8
15
0.4
8.1
7.6
13
2.8
1.7
1.3
0.47
7.0
ns
Testosterone
5a-DHT
5b-DHT
0.21
0.17
0.036*
3.4
0.6
1.2
0.53
0.61
0.37
0.16
1.0
0.31
20ꢀ-Hydroxysteroids
20a-Hydroxyprogesterone
5a-Pregnan-20a-ol-3-one
3,17-Dihydroxysteroids
5a,3b-DIOL
5-Androstene-3b,17b-diol
5a,3a-DIOL
4-Androstene-3a,17b-diol
5b-Androstane-3a,17b-diol
5b-Androstane-3b,17b-diol
na
ns
0.5
1.4
0.37
0.78
0.74
0.56
4.5
3.3
7.2
8.1
3.9
15
3.4
1.9
2.6
2.6
0.72
2.2
0.76
0.58
0.36
0.33
0.18
0.15
ns
ns
ns
0.22*
0.73
ns
2.0
0.7
0.11*
1.0
a
Km (mM), kcat (minꢂ1) and kcat/Km (minꢂ1mMꢂ1) were determined with 1.0 mM NAD+ as the coenzyme. The experimental errors were
515%, except that those indicated by asterisks were 15—20%. ns, no significant activity (>0.05 minꢂ1). na, no activity.
5

S. Endo et al.
Fig. 2 LC/MS analysis of oxidized products of 3b-hydroxyprogesterone (3b-OHP), 17b-estradiol (E2) and 5a-pregnan-20a-ol-3-one (P-20a-ol). (A)
Oxidation of 3b-OHP by AKR1C34 (1C34), AKR1C35 (1C35) and its V54A mutant. 3b-OHP and progesterone (P) were monitored at their
total ions of m/z 317.4 and 299.4, respectively. (B) Oxidation of E2 by 1C34 and 1C35. E2 and estrone (E1) were monitored at their total ions of
m/z 271.1 and 255.1, respectively. (C) Oxidation of P-20a-ol by 1C34 and 1C35. P-20a-ol and 5a-pregnane-3,20-dione (P-dione) were monitored
at their fragment ions of m/z 319.2 and 317.2, respectively. The peaks other than the above reduced products and substrates are due to unknown
substances (*).
resulted in >20-fold decreases in kcat and kcat/Km
Table IV. Inhibitor sensitivity of AKR1C34 and AKR1C35.
values for steroidal substrates. The large impairment
of the 3b-HSD activity was also evident from the
insignificant formation of progesterone in the product
AKR1C35
AKR1C34
IC50 (mM)
a
Inhibitor
IC50 (mM)
Ki (mM)
analysis of the 3b-hydroxyprogesterone oxidation by
this mutant enzyme (Fig. 2). Compared with the
V54A mutation, the C310F mutation showed smaller
effects on the kinetic constants, with the exception of
its larger effect on the Km value for 5a,3b-DIOL. Like
wild-type AKR1C35, the C310F mutant enzyme
formed 5a-DHT and 5a-androstan-3b-ol-17-one as
the major and minor products of the oxidation of
5a,3b-DIOL. The double mutation of V54A/C310F
impaired the kcat and kcat/Km values for all steroidal
substrates more greatly than the single V54A muta-
tion. No significant dehydrogenase activities for 3a-
and 20a-hydroxysteroids were detected for the V54A
and double mutant enzymes, whereas the C310F
mutant enzyme showed 3a- and 20a-HSD activities
Flavonoids
7-OH flavone
Chrysin
Quercetin
Genistein
Galangin
Others
Lithocholic acid
Hexestrol
0.92 0.02
1.0 0.2
2.6 0.2
4.2 0.2
5.0 0.3
0.11 0.02
0.11 0.02
0.46 0.03
0.55 0.02
0.69 0.05
>20a
>20a
>20a
>20a
>20a
4.4 0.2
6.5 0.5
0.39 0.1
0.54 0.06
3.3 0.6
ndb
>20a
1.3 0.2
Zearalenone
Phenolphthalein
Hinokitiol
15
1
>50a
>50a
>50a
0.65 0.02
3.2 0.2
ndb
a
The inhibition by the indicated concentration of inhibitor was
536% inhibition.
nd, not determined.
b
towards
3a-hydroxyprogesterone
(Km = 0.8
0.1 mM, kcat = 1.2 0.1 minꢂ1), 5b-pregnan-3a-ol-
20-one (Km = 0.8 0.1 mM, kcat = 0.48 0.08 minꢂ1),
double V54A/C310F mutant enzyme, and compared
the effects of mutations on kinetic constants for 3b-
hydroxysteroids (3b-hydroxyprogesterone, 5a-andro-
stan-3b-ol-17-one and 5a,3b-DIOL), 17b-estradiol
and S-tetralol (Table V). The mutation of V54A
5a-pregnan-20a-ol-3-one
(Km = 0.2 0.03 mM,
kcat = 0.13 0.02 minꢂ1), and 20a-hydroxyprogester-
one (Km = 1.7 0.2 mM, kcat = 0.09 0.01 minꢂ1).
In contrast to the alteration in the steroid specificity,
6

Hamster 3(17)b-hydroxysteroid dehydrogenase
tissues,
in
which
AKR1C35
was
ubiqui-
tously distributed. Among other 17b-hydroxysteroids
and 3a,17b-dihydroxysteroids, the neurosteroid
5a,3a-DIOL (30) was most efficiently oxidized into
inactive 5a-androstan-3a-ol-17-one by AKR1C35. In
the brain of female hamsters, the enzyme might be
involved in inactivation of this neurosteroid.
The most striking difference of AKR1C35 from
AKR1C34 and the aforementioned rodent and rabbit
NAD+-dependent AKRs is its ability to oxidize vari-
ous 3b-hydroxy-5a/b-dihydrosteroids and 3b-hydroxy-
Á5-androstenes (Table III). In this respect, the enzyme
is also distinct from NAD+-preferring 3a(17b)-HSD
that was purified from hamster liver cytosol (31). It
has been known that NAD+-dependent isozymes of
rat and mouse Á5-3b-HSDIs exhibit 3b-HSD activity
towards 3b-hydroxy-5a-androstanes (32), and alco-
hol dehydrogenase isozymes such as human
ADH1C(g) and horse ADH1S oxidize 3b-hydroxy-
5b-dihydro-C19/C21-steroids (33). The steroid specifi-
city of AKR1C35 is broader than those of the isozymes
of Á5-3b-HSDI and alcohol dehydrogenase. In ham-
sters, such a 3b-HSD activity has not been examined
for NAD+-dependent isozymes (types I and II) of Á5-
3b-HSDI that are expressed in the adrenal gland and
gonads (34), but alcohol dehydrogenase BB-isoenzyme
that is mainly distributed in the liver oxidizes only
5b-androstan-3b-ol-17-one of the 3b-hydroxysteroids
(35). Although AKR1C35 did not efficiently oxidize
5b-androstan-3b-ol-17-one, it may act as a major
3b-HSD for other 3b-hydroxy-5a/b-dihydrosteroids
in this animal. The 3b-hydroxysteroid substrates of
AKR1C35 include ligands of estrogen recep-
tor b (5a,3b-DIOL and 5-androstene-3b,17b-diol)
(36, 37) and antagonists of g-aminobutyric acid
type-A receptor (5a/b-pregnan-3b-ol-20-ones and 5b-
pregnane-3b,21-diol-20-one) (38), of which 5b-
pregnan-3b-ol-20-one also induces erythropoiesis
(39). AKR1C35 may participate in the metabolism of
the active 3b-hydroxysteroids, particularly, in female
hamster tissues including the brain.
As described earlier, the ubiquitous expression of
AKR1C34 in male hamster tissues and AKR1C35 in
the female tissues is probably related to their roles in
controlling the intracellular concentrations of 5a-DHT
(for AKR1C34), 17b-estradiol and biologically active
3b-hydroxysteroids (for AKR1C35). In contrast,
AKR1C34 and AKR1C35 were expressed almost spe-
cifically and highly in the livers of female and male
hamsters, respectively. Although the relevance of the
female liver-specific expression of AKR1C34 to steroid
metabolism is unclear, the hepatic enzyme would con-
tribute to generation of progesterone from its metab-
olites 3a/20a-hydroxyprogesterones. The male-liver
specific expression of AKR1C35 is similar to that of
hamster Á5-3b-HSDI type III isozyme that catalyzes
the NADPH-dependent conversion of 5a-DHT into
5a,3b-DIOL (34). In this respect, AKR1C35 with the
opposite activity (converting 5a,3b-DIOL back to 5a-
DHT) may play a role in controlling the concentra-
tions of the two active steroids in male hamster liver
in concert with Á5-3b-HSDI type III isozyme.
Alternatively, AKR1C34 and AKR1C35 may act as
Fig. 3 RT-PCR analysis for expression of mRNAs for AKR1C35 and
AKR1C34 in male and female hamster tissues. Tissues: brain (Br),
lung (Lu), heart (He), stomach (St), liver (Li), kidney (Ki), adrenal
gland (Ad), small intestine (SI), colon (Co), testis (Te) and ovary
(Ov). The expression of mRNA for b-actin is shown as the control.
the three mutations did not significantly alter the kin-
etic constants for the small and non-steroidal sub-
strate, S-tetralol, except that the V54A and double
mutations caused 5-fold decreases in the kcat value.
Discussion
The present cDNA cloning and characterization of the
encoded proteins, AKR1C34 and AKR1C35, reveal
that AKR1C34 is NAD+-dependent M6DH that was
previously purified from hamster liver (16). The iden-
tity of AKR1C34 and M6DH is evidenced by the
match between their amino acid sequences and simi-
larity of their enzymatic properties including substrate
specificity for xenobiotic alcohols and 17b-hydroxys-
teroids and low inhibitor sensitivity to quercetin and
lithocholic acid. In addition, AKR1C34 resembles
rabbit M6DH (AKR1C26) (15) in their substrate spe-
cificity for endogenous isoprenyl alcohols (geranylger-
aniol and farnesol) and 3a/20a-hydroxysteroids.
However, the tissue expression of hamster AKR1C34
is different from ubiquitously expressed rabbit M6DH
(15 and S. Endo, unpublished results) in its sexual di-
morphism (ubiquitous in the males versus apparently
liver-specific in the females). Because AKR1C34
showed high catalytic efficiency for 5a-DHT, it may
function in metabolism of the active androgen espe-
cially in male hamster tissues. AKR1C34 also
showed lower Km and higher kcat/Km values for most
xenobiotic alcohols compared with AKR1C35, sug-
gesting that the enzyme is involved in the hepatic
metabolism of xenobiotic alcohols, as well as biologic-
ally active steroids.
AKR1C35, that was newly identified in this study, is
the first NAD+-dependent dehydrogenase with 3(17)b-
HSD activity in the AKR superfamily. Among the
substrates of AKR1C35, 17b-estradiol is the best sub-
strate showing the highest kcat/Km value, which is also
higher than those of NAD+-dependent 3a/17b/20a-
HSDs of rabbits (AKR1C27 and AKR1C28) (15),
rats (AKR1C16 and AKR1C24) (11, 12) and mice
(AKR1C12 and AKR1C13) (12, 13). This ability
most likely contributes to controlling the intracellular
concentration of the active estrogen in female hamster
7

S. Endo et al.
Table V. Alteration of kinetic constants for steroidal and non-steroidal substrates by mutagenesis.
Km
kcat
kcat/Km
Substrate
Enzyme
(kM)
Mu/WTa
(minꢂ1
)
Mu/WTa
(kMꢂ1 minꢂ1
)
Mu/WTa
b
3b-Hydroxyprogesterone
WT
V54A
0.9
1
7
0.8
6
1
4
1
2
1
0.4
0.2
1
1
5
1
6
1
1
3.4
0.18 0.02
1.2 0.3
0.078 0.010
1.6
0.074 0.009
2.4 0.6
0.053 0.007
3.4
0.069 0.009
2.7 0.4
0.018 0.003
2.8
0.20 0.03
0.31 0.06
0.088 0.009
6.3
1.4 0.3
5.6 0.7
1.1 0.2
1
3.8
0.028
1.7
0.014
0.64
0.012
1.0
0.014
0.76
0.035
3.4
0.0035
7.0
0.13
0.62
0.038
0.10
0.021
0.093
0.035
1
0.007
0.5
0.004
1
0.02
2
0.02
1
0.05
4
0.005
1
0.02
0.09
0.005
1
0.2
0.9
0.4
6.3 0.2
0.7 0.1
5.5 0.6
2.5
6.2 0.7
2.3 0.4
3.7 0.5
4.5
2.0 0.2
0.8 0.1
5.1 0.3
0.4
1.6 0.2
0.5 0.1
2.3 0.2
61
0.05
0.4
0.02
1
0.05
2
0.03
1
0.02
0.8
0.005
1
0.09
0.5
0.03
1
0.2
0.9
0.2
C310F
V54A/C310F
b
5a-Androstan-3b-ol-17-one
5a,3b-DIOL
WT
V54A
C310F
V54A/C310F
b
WT
V54A
C310F
V54A/C310F
b
17b-Estradiol
WT
V54A
C310F
V54A/C310F
b
S-Tetralol
WT
V54A
68
60
31
3
5
2
C310F
V54A/C310F
1
0.5
a
Ratio of mutant enzyme to wild-type (WT).
The values of WT are taken from Tables II and III.
b
major dehydrogenases for xenobiotic alcohols in the
livers of the females and males, respectively, because
the xenobiotic substrates include long-chain aliphatic
alcohols, constituents of plants ingested by animals.
The regulation mechanism of the expression of the
AKR1C35 and Á5-3b-HSDI isozyme genes in relation
to the sexual dimorphism needs to be established in
further metabolomic and molecular genetic studies.
7-Hydroxyflavones and an isoflavone, genistein,
were AKR1C35-specific inhibitors, suggesting that its
activity is affected by ingested flavonoids. NAD+-
dependent 3a/17b/20a-HSDs of rats (AKR1C24) (11)
and rabbits (AKR1C27 and AKR1C28) (15), and
specificity from AKR1C1 into AKR1C2 and vice versa
is achieved by replacement of the residue 54 between
Val in AKR1C1 and Leu in AKR1C2 (43, 44). In the
light of these studies, the present mutation of V54A
can be interpreted to alter the orientation of 3b- and
17b-hydroxy groups of the steroidal substrates in the
active site of AKR1C35. However, the mutations of
the residues in AKR1C35 to the corresponding resi-
dues in AKR1C34 did not convert it to a 3a/17b/
20a-HSD that is similar to AKR1C34. As the double
mutation further impaired the 3b- and 17b-HSD activ-
ities in contrast to the appearance of 3a- and 20a-HSD
activities by the single C310F mutation, the steroid
specificity of AKR1C35 may be dictated by a combin-
ation of Val54, Cys310 and other substrate-binding
residues, yet to be identified.
rabbit
ovarian
NADPH-dependent
20a-HSD
(AKR1C5) (40) were also sensitive to flavonoids, but
these enzymes are insensitive to isoflavones and
therefore differ from genestin-sensitive AKR1C35.
Genistein
also
inhibits
NAD+-dependent
Supplementary Data
Á5-3b-HSDI isozymes (41, 42) that are expressed in
steroidogenic glands (32). When the isoflavone is
ingested in the hamster, it may decrease not only the
synthesis of steroid hormones in the endocrine glands
by inhibiting the Á5-3b-HSDI isozymes, but also the
metabolism of 5a-DHT, 17b-estradiol and active
Supplementary Data are available at JB Online.
Funding
This work was partly funded by grant-in-aid for Scientific
Research (C) from the Japan Society for the Promotion of Science
(26460149).
3b-hydroxysteroids
by
inhibiting
AKR1C35.
Therefore, the isoflavone might not exert large effects
on concentrations of these active steroids in this
animal.
Conflict of Interest
None declared.
The results of the mutagenesis have revealed that
Val54 plays more critical role in recognition of 3b-
and 17b-hydroxysteroid substrates by AKR1C35
than Cys310. In the previous studies on 3b-HSD activ-
ity of human NADPH-dependent AKRs (1C1 and
1C2), the residue at position 54 is suggested to play
an important role in determining the stereospecific
reduction of 3-keto group of 5a-DHT into 3a- or
3b-hydroxy group (9). The conversion of the steroid
References
1. Penning, T.M. (1997) Molecular endocrinology of
hydroxysteroid dehydrogenases. Endocr. Rev. 18,
281—305
2. Kavanagh, K.L., Jornvall, H., Persson, B., and
Oppermann, U. (2008) The SDR superfamily: functional
8

Hamster 3(17)b-hydroxysteroid dehydrogenase
and structural diversity within a family of metabolic and
regulatory enzymes. Cell. Mol. Life Sci. 65, 3895—3906
3. Barski, O.A., Tipparaju, S.M., and Bhatnagar, A. (2008)
The aldo-keto reductase superfamily and its role in drug
metabolism and detoxification. Drug Metab. Rev. 40,
553—624
15. Endo, S., Matsunaga, T., Fujimoto, A., Kumada, S.,
Arai, Y, Miura, Y., Mikamo, H., El-Kabbani, O.,
Yamano, S., Iinuma, M., and Hara, A. (2012)
Characterization of a rabbit morphine 6-dehydrogenase
and two NAD+-dependent 3a(17b)-hydroxysteroid de-
hydrogenases. Arch. Biochem. Biophys. 529, 131—139
4. Wu, X., Lukacik, P., Kavanagh, K.L., and Oppermann,
U. (2007) SDR-type human hydroxysteroid dehydrogen-
ases involved in steroid hormone activation. Mol. Cell.
Endocrinol. 265-266, 71—76
5. Ishikura, S., Usami, N., Kitahara, K., Isaji, T., Oda, K.,
Nakagawa, J., and Hara, A. (2001) Enzymatic character-
istics and subcellular distribution of a short-chain
dehydrogenase/reductase family protein, P26h, in ham-
ster testis and epididymis. Biochemistry 40, 214—224
6. Matsunaga, T., Endo, S., Maeda, S., Ishikura, S.,
Tajima, K., Tanaka, N., Nakamura, KT., Imamura,
Y., and Hara, A. (2008) Characterization of human
DHRS4: an inducible short-chain dehydrogenase/reduc-
tase enzyme with 3a-hydroxysteroid dehydrogenase
activity. Arch. Biochem. Biophys. 477, 339—347
7. Nakajin, S., Takase, N., Ohno, S., Toyoshima, S., and
Baker, M.E. (1998) Mutation of tyrosine-194 and lysine-
198 in the catalytic site of pig 3a/b,20b-hydroxysteroid
dehydrogenase. Biochem. J. 334, 553—557
8. Penning, T.M., Burczynski, M.E., Jez, J.M., Hung, C.F.,
Lin, H.K., Ma, H., Moore, M., Palackal, N., and
Ratnam, K. (2000) Human 3a-hydroxysteroid dehydro-
genase isoforms (AKR1C1-AKR1C4) of the aldo-keto
reductase superfamily: functional plasticity and tissue
distribution reveals roles in the inactivation and forma-
tion of male and female sex hormones. Biochem. J. 351,
67—77
9. Steckelbroeck, S., Jin, Y., Gopishetty, S., Oyesanmi, B.,
and Penning, T.M. (2004) Human cytosolic 3a-
hydroxysteroid dehydrogenases of the aldo-keto reduc-
tase superfamily display significant 3b-hydroxysteroid
dehydrogenase activity: implications for steroid hormone
metabolism and action. J. Biol. Chem. 279, 10784—10795
16. Todaka, T., Yamano, S., and Toki, S. (2000) Purification
and characterization of NAD-dependent morphine
6-dehydrogenase from hamster liver cytosol, a new
member of the aldo-keto reductase superfamily. Arch.
Biochem. Biophys. 374, 189—197
17. Platt, K.L. and Oesch, F. (1977) An improved synthesis
of trans-5,6-dihydroxy-1,3- cyclohexadiene (trans-1,2-
dihydroxy-1,2-dihydrobenzene). Synthesis, 7, 449—450
18. Sambrook, J., Fritsch, E.F., and Maniatis, T. (1989)
Molecular Cloning: A Laboratory Manual, 2nd ed.
Cold Spring Harbor Laboratory, Cold Spring Harbor.
19. Endo, S., Matsunaga, T., Kuragano, T., Ohno, S.,
Kitade, Y., Tajima, K., El-Kabbani, O., and Hara, A.
(2010) Properties and tissue distribution of a novel aldo-
keto reductase encoding in a rat gene (Akr1b10). Arch.
Biochem. Biophys. 503, 230—237
20. Bradford, M.M. (1976) A rapid and sensitive method for
the quantitation of microgram quantities of protein uti-
lizing the principle of protein-dye binding. Anal.
Biochem. 72, 248—254
21. Ishikura, S., Usami, N., Nakajima, S., Kameyama, A.,
Shiraishi, H., Carbone, V., El-Kabbani, O., and Hara, A.
(2004) Characterization of two isoforms of mouse
3(17)a-hydroxysteroid dehydrogenases of the aldo-keto
reductase family. Biol. Pharm. Bull. 27, 1939—1945
22. Uehara, Y., Urata, H., Ideishi, M., Arakawa, K., and
Saku, K. (2002) Chymase inhibition suppresses high-
cholesterol diet-induced lipid accumulation in the ham-
ster aorta. Cardiovasc. Res. 55, 870—876
23. Takenoshita, R., Nomura, Y., and Toki, S. (2001)
Cloning and expression of cDNA encoding hamster
liver 3-hydroxyhexobarbital/17b(3a)-hydroxysteroid de-
hydrogenase 1. Chem. Biol. Interact. 130-132, 863—870
10. Sanai, M., Endo, S., Matsunaga, T., Ishikura, S.,
Tajima, K., El-Kabbani, O., and Hara, A. (2007) Rat
NAD+-dependent 3a-hydroxysteroid dehydrogenase
(AKR1C17): a member of the aldo-keto reductase
family highly expressed in kidney cytosol. Arch.
Biochem. Biophys. 464, 122—129
11. Ishikura, S., Matsumoto, K., Sanai, M., Horie, K.,
Matsunaga, T., Tajima, K., El-Kabbani, O., and Hara,
A. (2006) Molecular cloning of a novel type of rat cyto-
plasmic 17b-hydroxysteroid dehydrogenase distinct from
the type 5 isozyme. J. Biochem. 139, 1053—1063
12. Endo, S., Sanai, M., Horie, K., Matsunaga, T., Ishikura,
S., Tajima, K., El-Kabbani, O., and Hara, A. (2007)
Characterization of rat and mouse NAD+-dependent
3a/17b/20a-hydroxysteroid dehydrogenases and identifi-
cation of substrate specificity determinants by site-
directed mutagenesis. Arch. Biochem. Biophys. 467,
76—86
13. Endo, S., Matsumoto, K., Matsunaga, T., Ishikura, S.,
Tajima, K., El-Kabbani, O., and Hara, A. (2006)
Substrate specificity of a mouse aldo-keto reductase
(AKR1C12). Biol. Pharm. Bull. 29, 2488—2492
24. Penning, T.M. (1999) Molecular determinants of steroid
recognition and catalysis in aldo-keto reductases.
Lessons from 3a-hydroxysteroid dehydrogenase.
J. Steroid Biochem. Mol. Biol. 69, 211—225
25. Couture, J.F., Legrand, P., Cantin, L., Labrie, F.,
Luu-The, V., and Breton, R. (2004) Loop relaxation, a
mechanism that explains the reduced specificity of rabbit
20a-hydroxysteroid dehydrogenase, a member of the aldo-
keto reductase superfamily. J. Mol. Biol. 339, 89—102
26. Couture, J.F., Legrand, P., Cantin, L., Luu-The, V.,
Labrie, F., and Breton, R. (2003) Human 20a-hydroxys-
teroid dehydrogenase: crystallographic and site-directed
mutagenesis studies lead to the identification of an alter-
native binding site for C21-steroids. J. Mol. Biol. 331,
593—604
27. Jin, Y., Stayrook, S.E., Albert, R.H., Palackal, N.T.,
Penning, T.M., and Lewis, M. (2001) Crystal structure
of human type III 3a-hydroxysteroid dehydrogenase/
bile-acid binding protein complexed with NADP+ and
ursodeoxycholate. Biochemistry 40, 10161—10168
28. Qiu, W., Zhou, M., Labrie, F., and Lin, S.X. (2004)
Crystal structures of the multispecific 17b-hydroxyster-
oid dehydrogenase type 5: critical androgen regulation in
human peripheral tissues. Mol. Endocrinol. 18,
1798—1807
29. Faucher, F., Cantin, L., de Jesus-Tran, K.P., Lemieux,
M., Luu-The, V., Labrie, F., and Breton, R. (2007)
Mouse 17a-hydroxysteroid dehydrogenase (AKR1C21)
14. Endo, S., Matsunaga, T., Kumada, S., Fujimoto, A.,
Ohno, S., El-Kabbani, O., Hu, D., Toyooka, N.,
Mano, J., Tajima, K., and Hara, A. (2012)
Characterization of rabbit aldose reductase-like protein
with 3b-hydroxysteroid dehydrogenase activity. Arch.
Biochem. Biophys. 527, 23—30
9

S. Endo et al.
binds steroids differently from other aldo-keto reduc-
regulation
of
hormonal
stress
reactivity.
tases: Identification and characterization of amino acid
residues critical for substrate binding. J. Mol. Biol. 369,
525—540
J. Neuroendocrinol. 21, 351—358
38. Baker, M.E. (2013) What are the physiological estro-
gens? Steroids 78, 337—340
30. Reddy, D.S. (2010) Neurosteroids: endogenous role in
the human brain and therapeutic potentials. Prog.
Brain Res. 186, 113—137
31. Ohmura, M., Hara, A., Nakagawa, M., and Sawada, H.
(1990) Demonstration of 3a(17b)-hydroxysteroid de-
hydrogenase distinct from 3a-hydroxysteroid dehydro-
genase in hamster liver. Biochem. J. 266, 583—589
32. Simard, J., Ricketts, M.L., Gingras, S., Soucy, P., Feltus,
F.A., and Melner, M.H. (2005) Molecular biology of the
3b-hydroxysteroid dehydrogenase/Á5-Á4 isomerase gene
family. Endocr. Rev. 4, 525—582
33. Plapp, B.V. and Berst, K.B. (2003) Specificity of human
alcohol dehydrogenase 1C*2 (g2g2) for steroids and
simulation of the uncompetitive inhibition of ethanol me-
tabolism. Chem. Biol. Interact. 143-144, 183—193
34. Rogerson, F.M., Courtemanche, J., Fleury, A., Head,
J.R., LeHoux, J.G., and Mason, J.I. (1998)
Characterization of cDNAs encoding isoforms of ham-
ster 3b-hydroxysteroid dehydrogenase/Á5-Á4 isomerase.
J. Mol. Endocrinol. 20, 99—110
35. Keung, W.M. (1996) Isolation and characterization of
three alcohol dehydrogenase isozymes from Syrian
golden hamsters. Alcohol Clin. Exp. Res. 20, 213—220
36. Wang, M., He, Y., Eisenman, L.N., Fields, C., Zeng,
C.M., Mathews, J., Benz, A., Fu, T., Zorumski, E.,
Steinbach, J.H., Covey, D.F., Zorumski, C.F., and
Mennerick, S. (2002) 3b-hydroxypregnane steroids are
pregnenolone sulfate-like GABAA receptor antagonists.
J. Neurosci. 22, 3366—3375
37. Handa, R.J., Weiser, M.J., and Zuloaga, D.G. (2009) A
role for the androgen metabolite, 5a-androstane-3b,17b-
diol, in modulating oestrogen receptor b-mediated
39. Besa, E.C., Gorshein, D., Hait, W.A., and Gardner, F.H.
(1973) Effective erythropoiesis induced by 5b-pregnane-
3b-hydroxy-20-one in squirrel monkeys. J. Clin. Invest.
52, 2278—2282
40. Endo, S., Arai Y., Hara A., Kitade Y., Bunai Y., El-
Kabbani O., and Matsunaga T. (2013) Substrate
specificity and inhibitor sensitivity of rabbit 20a-
hydroxysteroid dehydrogenase. Biol. Pharm. Bull. 36,
1514—1518
41. Ohno, S., Matsumoto, N., Watanabe, M., and Nakajin,
S. (2004) Flavonoid inhibition of overexpressed human
3b-hydroxysteroid dehydrogenase type I. J. Steroid
Biochem. Mol. Biol. 88, 175—182
42. Hu, G.X., Zhao, B.H., Chu, Y.H., Zhou, H.Y.,
Akingbemi, B.T., Zheng, Z.Q., and Ge, R.S. (2010)
Effects of genistein and equol on human and rat
testicular
3b-hydroxysteroid
dehydrogenase
and
17b-hydroxysteroid dehydrogenase 3 activities. Asian J.
Androl. 4, 519—526
43. Matsuura, K., Deyashiki, Y., Sato, K., Ishida, N., Miwa,
G., and Hara, A. (1997) Identification of amino
acid residues responsible for differences in sub-
strate specificity and inhibitor sensitivity between
two human liver dihydrodiol dehydrogenase isoen-
zymes by site-directed mutagenesis. Biochem. J. 323,
61—64
44. Zhang, B., Zhu, D.W., Hu, X.J., Zhou, M., Shang, P., and
Lin, S.X. (2014) Human 3a-hydroxysteroid de-
hydrogenase type 3 (3a-HSD3): the V54L mutation
restricting the steroid alternative binding and enhancing
the 20a-HSD activity. J. Steroid Biochem. Mol. Biol. 141,
135—143
10