Structure-activity relationships for inhibition of human 5α-reductases by polyphenols

Article abstract of DOI:10.1016/S0006-2952(02)00848-1

The enzyme steroid 5α-reductase (EC 1.3.99.5) catalyzes the NADPH-dependent reduction of the double bond of a variety of 3-oxo-Δ⁴ steroids including the conversion of testosterone to 5α-dihydrotestosterone. In humans, 5α-reductase activity is critical for certain aspects of male sexual differentiation, and may be involved in the development of benign prostatic hyperplasia, alopecia, hirsutism, and prostate cancer. Certain natural products contain components that are inhibitors of 5α-reductase, such as the green tea catechin (-)-epigallocatechin gallate (EGCG). EGCG shows potent inhibition in cell-free but not in whole-cell assays of 5α-reductase. Replacement of the gallate ester in EGCG with long-chain fatty acids produced potent 5α-reductase inhibitors that were active in both cell-free and whole-cell assay systems. Other flavonoids that were potent inhibitors of the type 1 5α-reductase include myricetin, quercitin, baicalein, and fisetin. Biochanin A, daidzein, genistein, and kaempferol were much better inhibitors of the type 2 than the type 1 isozyme. Several other natural and synthetic polyphenolic compounds were more effective inhibitors of the type 1 than the type 2 isozyme, including alizarin, anthrarobin, gossypol, nordihydroguaiaretic acid, caffeic acid phenethyl ester, and octyl and dodecyl gallates. The presence of a catechol group was characteristic of almost all inhibitors that showed selectivity for the type 1 isozyme of 5α-reductase. Since some of these compounds are consumed as part of the normal diet or in supplements, they have the potential to inhibit 5α-reductase activity, which may be useful for the prevention or treatment of androgen-dependent disorders. However, these compounds also may adversely affect male sexual differentiation.

Full text of DOI:10.1016/S0006-2952(02)00848-1

Biochemical Pharmacology 63 ,2002) 1165±1176
Structure±activity relationships for inhibition of human
5a-reductases by polyphenols
Richard A. Hiipakka, Han-Zhong Zhang, Wei Dai, Qing Dai, Shutsung Liao*
Department of Biochemistry and Molecular Biology, The Ben May Institute for Cancer Research, and The Tang Center for
Herbal Medicine Research MC6027, University of Chicago, 5841 S. Maryland, Chicago, IL 60637, USA
Received 17 April 2001; accepted 1 August 2001
Abstract
The enzyme steroid 5a-reductase ,EC 1.3.99.5) catalyzes the NADPH-dependent reduction of the double bond of a variety of 3-oxo-D4
steroids including the conversion of testosterone to 5a-dihydrotestosterone. In humans, 5a-reductase activity is critical for certain aspects
of male sexual differentiation, and may be involved in the development of benign prostatic hyperplasia, alopecia, hirsutism, and prostate
cancer. Certain natural products contain components that are inhibitors of 5a-reductase, such as the green tea catechin ,À)-
epigallocatechin gallate ,EGCG). EGCG shows potent inhibition in cell-free but not in whole-cell assays of 5a-reductase. Replacement
of the gallate ester in EGCG with long-chain fatty acids produced potent 5a-reductase inhibitors that were active in both cell-free and
whole-cell assay systems. Other ¯avonoids that were potent inhibitors of the type 1 5a-reductase include myricetin, quercitin, baicalein,
and ®setin. Biochanin A, daidzein, genistein, and kaempferol were much better inhibitors of the type 2 than the type 1 isozyme. Several
other natural and synthetic polyphenolic compounds were more effective inhibitors of the type 1 than the type 2 isozyme, including
alizarin, anthrarobin, gossypol, nordihydroguaiaretic acid, caffeic acid phenethyl ester, and octyl and dodecyl gallates. The presence of a
catechol group was characteristic of almost all inhibitors that showed selectivity for the type 1 isozyme of 5a-reductase. Since some of
these compounds are consumed as part of the normal diet or in supplements, they have the potential to inhibit 5a-reductase activity, which
may be useful for the prevention or treatment of androgen-dependent disorders. However, these compounds also may adversely affect
male sexual differentiation. # 2002 Elsevier Science Inc. All rights reserved.
Keywords: 5a-Reductase; Polyphenols; Flavonoids; Green tea catechins; Epigallocatechin gallate; Prostate cancer
1. Introduction
optimum and low af®nity for testosterone ,Km > 1 mM),
while the type 2 isozyme has an acidic pH optimum and
high af®nity for testosterone ,Km < 10 nM) [8].
The microsomal enzyme steroid 5a-reductase ,EC
1.3.99.5) catalyzes the NADPH-dependent reduction of
the D4,5 double bond of a variety of 3-oxo-D4 steroids [1,2],
including the conversion of testosterone to 5a-dihydrotes-
tosterone, a process thought to amplify the androgenic
response, perhaps because of the higher af®nity of the
androgen receptor for 5a-dihydrotestosterone than for
testosterone [3]. Two different 5a-reductase isozymes have
been characterized in humans, monkeys, rats, and mice [4±
7]. The two human isozymes share approximately 50%
sequence identity and have different biochemical proper-
ties. For example, the type 1 isozyme has a broad basic pH
Studies of mice with genetically engineered 5a-reduc-
tase gene knockouts, as well as investigations of natural
5a-reductase de®ciencies in humans, have identi®ed some
of the roles this enzyme plays in different biological
processes [7]. Female mice de®cient in the type 1 5a-
reductase have impaired cervical ripening, leading to
defective parturition [9,10]. This defect may be due to
impaired catabolism of cervical progesterone, which is
also a substrate for 5a-reductase. Uterine type 1 5a-reduc-
tase is also important for fetal viability, because 5a-reduc-
tase activity limits fetal exposure to toxic levels of 17b-
estradiol by competing for substrate with aromatase, which
catalyzes the synthesis of 17b-estradiol from testosterone
[11]. In humans, activity of the type 2 isozyme is critical
for differentiation of the prostate and male external geni-
talia [4,12,13]. Based on studies of individuals with
* Corresponding author. Tel.: ‡1-773-702-6999; fax: ‡1-773-834-1770.
E-mail address: sliao@huggins.bsd.uchicago.edu ,S. Liao).
Abbreviations: DMEM, Dulbecco's modified Eagle's medium; EC,
,À)-epicatechin; EGC, ,À)-epigallocatechin; ECG, ,À)-epicatechin gal-
late; EGCG, ,À)-epigallocatechin gallate.
0006-2952/02/$ ± see front matter # 2002 Elsevier Science Inc. All rights reserved.
PII: S 0 0 0 6 - 2 9 5 2 , 0 2 ) 0 0 8 4 8 - 1

1166
R.A. Hiipakka et al. / Biochemical Pharmacology 63 72002) 1165±1176
inherited de®ciencies in 5a-reductase activity as well as
laboratory studies of affected tissues, 5a-reductase may
also have a role in the development of a variety of human
disorders including benign prostatic hyperplasia [14], acne
[15], alopecia [16,17], and hirsutism [18]. 5a-Reductase
also has been proposed to have a role in the development of
prostate cancer and possibly may be responsible for differ-
ences in prostate cancer mortality among different racial
groups [19]. A common missense mutation ,V89L) that
decreases the activity of the type 2 5a-reductase, is com-
mon among Asians, a group that has a lower mortality from
prostate cancer compared with African±American men and
non-Hispanic whites [20]. Finasteride, a synthetic 5a-
reductase inhibitor, is currently used to treat benign pro-
static hyperplasia [21] and alopecia [22], and it is also
being studied in clinical trials as a chemopreventative for
prostate cancer [23].
Diet has an important role in modulating cancer inci-
dence and mortality, and differences in diet may explain
geographical differences in prostate cancer mortality
[24,25]. Since androgens regulate the growth and function
of the normal prostate and prostate cancer [26,27], dietary
components capable of altering this growth signaling path-
way in the prostate may affect prostate cancer development
and progression. We have shown that green tea contains
phytochemicals called catechins that are inhibitors of 5a-
reductase [28]. It is not known whether green tea catechins
modulate androgenic activity in vivo in humans, but rats
injected with the green tea catechin EGCG have a variety
of endocrine changes, including lower serum testosterone
concentrations and smaller prostates than the controls [29].
Other natural product inhibitors of 5a-reductase that have
been identi®ed include polyunsaturated fatty acids, such as
g-linolenic acid [30]; the macrocyclic ellagitannins,
oenothein A and B [31,32]; the ¯avonoids and lignans,
genistein, formononentin, biochanin A, daidzein, coumes-
trol, equol, and enterolactone [33]; the bisnaphthoquinone,
impatienol [34]; and the isoprenylated ¯avone and stilbene,
artocarpin and chlorophorin [35]. Since many of these
inhibitors of 5a-reductase are polyphenols, we have inves-
tigated the ability of a variety of natural and synthetic
polyphenols to inhibit 5a-reductase to determine what
structural elements are important for potent inhibition of
5a-reductase by this class of compounds in both the cell-
free and whole-cell assay systems.
Other chemicals either were purchased from Sigma or
Aldrich or were synthesized in our laboratory as described
below. Caffeic acid phenethyl ester was synthesized as
described [36]. A variety of semi-synthetic derivatives of
the green tea catechin EGC were synthesized by esterifying
various aromatic and aliphatic acids to the 3-hydroxy
group of EGC ,Fig. 1). EGC derivatives were synthesized
by acetylating all six hydroxyl groups of EGC using acetic
anhydride in pyridine, followed by selective deacetylation
in Tris buffer at pH 8.2 to give the 3-monoacetate ,2).
Silylation of the phenolic hydroxyls and subsequent dea-
cetylation afforded pentasilylated epigallocatechin ,4)
[37]. Aryl esters ,5) of EGC were prepared by transester-
i®cation of the 3-hydroxyl with the appropriate methyl
esters [38]. Aliphatic esters ,7) of EGC were prepared by
reaction with the requisite acids in the presence of dicy-
clohexylcarbodiimide and dimethylaminopyridine [39].
Silyl protective groups were then removed from 5 and 7
with triethylamine trihydro¯uoride, providing the EGC
analogs. All compounds synthesized in our laboratory were
puri®ed by silica gel chromatography and were character-
ized by 1H-NMR. Spectroscopic data were consistent with
their structure, e.g. EGC-myristoleate in CD3OD: 6.42
,2H, s, Harom), 5.92 ,1H, d, J ˆ 2:28 Hz, Harom), 5.88
,1H, d, J ˆ 2:28 Hz, Harom), 5.36±5.28 ,3H, m, CH=CH,
H3), 4.92 ,1H, s, H2), 2.82 ,2H, t, J ˆ 7:48 Hz, CH2CO),
2.00 ,4H, m, CH2±C=C±CH2), 1.50±1.0 ,14H, m, CH2),
0.90 ,3H, t, J ˆ 7:1 Hz, CH3) ppm.
2.2. Expression of human 5a-reductase isozymes
and preparation of cell extracts
Cell lines expressing the different types of human 5a-
reductase were prepared as described previously [28]. In
brief, cDNAs for the human types 1 and 2 5a-reductases
were isolated from human prostate cDNA libraries and
subcloned into the retroviral expression vector pMV7 [40],
and high titer stocks of viruses containing the types 1 and 2
5a-reductase cDNAs were generated using the packaging
cells BOSC 23 293 [41]. Rat 1A cells [42,43] were infected
with retrovirus, and cells containing integrated retrovirus
were selected for resistance to G418 [44]. Cells expressing
either the type 1 or 2 human 5a-reductase were expanded
under G418 selection and, when con¯uent, were removed
from plates by trypsinization. Cells were collected by
centrifugation ,500 g for 10 min at 228), washed once with
culture medium containing 10% fetal bovine serum,
quickly frozen on dry ice, and stored at À908 until used
in the preparation of a microsomal fraction containing
5a-reductase.
2. Materials and methods
2.1. Materials
Rat 1A cells were chosen for expression of 5a-reductase
based on their low background of 5a-reductase activity
,1.8 pmol/min/106 cells) compared with a variety of other
cell types that were tested. 5a-Reductase activity ,type 1 or
2) in cells transduced with recombinant retroviruses con-
taining 5a-reductase cDNA was 28-fold greater ,52 pmol/
[4-14C]-Testosterone ,50±60 mCi/mmol) was a product
of Perkin-Elmer Life Sciences. Puri®ed catechins, ,À)-
epicatechin ,EC), ,À)-epigallocatechin ,EGC), ,À)-epica-
techin gallate ,ECG), and EGCG, were isolated from green
tea ,Camellia sinensis) in our laboratory as described [28].

R.A. Hiipakka et al. / Biochemical Pharmacology 63 72002) 1165±1176
1167
Fig. 1. Synthesis scheme for various aromatic and aliphatic ester derivatives of EGC. Ac: acetate; TBDMS: tert-butyldimethylsilane; TBDMSCl: tertiary-
butyldimethylsilyl chloride; DCC: dicyclohexylcarbodiimide; DMAP: dimethylpyridine; THF: tetrahydrofuran; R0: aliphatic group; Ar: aromatic group.
min/106 cells) than in cells transduced with control virus
lacking 5a-reductase cDNA. Since the activity of the
recombinant human 5a-reductase activity was so much
greater than endogenous rat 5a-reductase, our results
should re¯ect characteristics of the exogenous human
and not the endogenous rat enzyme.
,217 pmol/min per mg of protein; type 2) greater than
activity ,11.5 pmol/min per mg of protein) in microsomes
from cells transduced with a control virus.
2.3. Assay of 5a-reductase
For the preparation of microsomes containing 5a-reduc-
tase, cells were thawed on ice and then homogenized in
10 vol. of 0.32 M sucrose containing 20 mM potassium
phosphate, pH 7, 1 mM EDTA, 0.1 mM phenylmethylsul-
fonyl ¯uoride and 1 mg leupeptin/mL, using a Dounce
homogenizer and then sonication. The homogenate was
centrifuged at 1000 g for 10 min at 48 and the pellet was
discarded. The supernatant was centrifuged at 100,000 g
for 60 min at 48, and the resulting pellet ,microsomal
fraction) was resuspended at 10±20 mg protein/mL in
0.32 M sucrose, 20 mM potassium phosphate, pH 7,
1 mM EDTA, using a Dounce homogenizer. This micro-
somal fraction was frozen on dry ice and stored at À908
until used in assays of 5a-reductase. We assayed 5a-
reductase in various cellular fractions after differential
centrifugation of cell extracts. Of the total 5a-reductase
activity in cell extracts, 70±80% was present in the micro-
somal fraction, 10±20% in the nuclear fraction ,1000 g
pellet), and 2±6% in the cytosolic fraction. 5a-Reductase
activity in microsomes isolated from cells transduced with
virus containing 5a-reductase cDNA was 31-fold
,350 pmol/min per mg of protein; type 1) and 19-fold
The cell-free assay was based on the measurement of
5a-dihydrotestosterone production from testosterone in the
presence of microsomes prepared from cells containing
either the type 1 or 2 human 5a-reductase. The assay
mixture, in a ®nal volume of 0.25 mL, contained 3 mM
[4-14C]-testosterone, 0.1 mM NADPH, 100 mM potassium
phosphate, pH 7.0. Test compounds were prepared in water
or dimethyl sulfoxide, and 2.5 mL of appropriate dilutions
was added to a reaction mix prior to the addition of
enzyme. Controls contained similar concentrations of sol-
vents. The reaction was started by the addition of 25 mL of
microsomes containing 25 mg of protein to 225 mL of an
assay mixture at 378. The mixture was incubated at 378 for
1 hr, and the reaction was stopped by the addition of
0.5 mL of ethyl acetate and mixing for 1 min. After
centrifugation ,10,000 g for 5 min at 228), the organic
phase was removed, dried under vacuum, dissolved in
25 mL of ethyl acetate and applied to a silica gel 60
TLC plate, which was developed in a solvent system
consisting of methylene chloride:ethyl acetate:methanol
,85:15:3). Conversion of testosterone to 5a-reduced meta-
bolites was measured by scanning the TLC plate on an

1168
R.A. Hiipakka et al. / Biochemical Pharmacology 63 72002) 1165±1176
AMBIS radioanalytical scanner or scanning a phosphor
screen exposed to the TLC plate on a Storm 860 phos-
phoimager. 5a-Dihydrotestosterone was the predominant
metabolite ,>95%), with little or no conversion of testos-
terone to androstanediols, androstandione, 4-androstene-
dione, or other metabolites, and recovery of radioactivity
was from 90 to 95%.
data points. 5a-Dihydrotestosterone was the predominant
metabolite ,>95%), with little or no conversion of testos-
terone to androstanediols, androstandione, 4-androstene-
dione, or other metabolites, and recovery of radioactivity
was from 85 to 90%.
2.4. HPLC analysis
Inhibition of 5a-reductase activity was also measured in
intact cells expressing either human type 1 or 2 5a-reduc-
tase. Cells were plated at 50,000 per well in a 24-well plate
in 1 mL of Dulbecco's modi®ed Eagle's medium ,DMEM)
containing 50 U penicillin/mL, 50 mg streptomycin/mL,
and 10% fetal bovine serum and then placed in an incu-
bator with 5% CO2 for 18 hr at 378. The medium was then
changed to 0.5 mL of serum-free DMEM, and 5 mL of test
compound in water or dimethyl sulfoxide was added. Cells
were incubated for 1 hr at 378 before the addition of
[4-14C]-testosterone, at a ®nal concentration of 1.5 mM.
Then cells were incubated for 3 hr at 378, the medium was
removed, and radioactive steroids were extracted with
ethyl acetate. The amounts of labeled testosterone and
5a-dihydrotestosterone in extracts were determined by
TLC and scanning the TLC plate as described above.
The concentration of test compound inhibiting the conver-
sion of testosterone to 5a-dihydrotestosterone by 50%
The stability of EGCG at 378 in cell culture medium
,DMEM) equilibrated with 5% CO2 ,pH 7.4±7.5) or in
100 mM potassium phosphate buffer, pH 7.0, medium or
buffers that were used for whole-cell and cell-free assays of
5a-reductase, respectively, was determined by monitoring
EGCG concentrations by HPLC using a Novapak ,Waters)
C18 column ,3:9 mm  150 mm) and isocratic elution at
408 at a ¯ow rate of 1 mL/min with acetonitrile:ethyl
acetate:0.05% phosphoric acid in water ,12:2:86). EGCG
was detected by monitoring the absorbance at 275 nm.
3. Results
The structures of various compounds investigated in this
report are shown in Figs. 2 and 3. A comparison of the
abilities of the four major green tea catechins, EC, EGC,
ECG, and EGCG ,Fig. 2), to inhibit the types 1 and 2
,IC ) was determined by interpolation between appropriate
50
Fig. 2. Structures of flavonoids tested for inhibitory activity against 5a-reductase.

R.A. Hiipakka et al. / Biochemical Pharmacology 63 72002) 1165±1176
1169
Fig. 3. Structures of polyphenols tested for inhibitory activity against 5a-reductase.
isozymes of 5a-reductase in cell-free and whole-cell
assays is presented in Table 1. Using a cell-free assay,
ECG and EGCG were better inhibitors of 5a-reductase
than were EC and EGC, and the type 1 isozyme was more
sensitive to these inhibitors than the type 2 isozyme. Since
ECG and EGCG only differ structurally from EC and EGC
by the presence of a gallic acid ester on the 3-hydroxyl, the
gallate group is important for the enhanced ability of ECG

1170
R.A. Hiipakka et al. / Biochemical Pharmacology 63 72002) 1165±1176
Table 1
Inhibition of 5a-reductase isozymes by green tea catechinsa
Catechin
5a-Reductase
Cell-free assay IC ,mM)
50
Whole-cell assay IC50 ,mM)
Type 1
Type 2
Type 1
Type 2
EC
>100 ,14)
>100 ,15)
11 ,100)
15 ,99)
>100 ,4)
>100 ,3)
69 ,83)
74 ,74)
>100 ,0)
>100 ,15)
>100 ,0)
>100 ,6)
>100 ,1)
>100 ,1)
>100 ,0)
>100 ,0)
EGC
ECG
EGCG
a
IC50: concentration ,mM) of compound producing 50% inhibition of
5a-reductase activity. Values in parentheses are percent inhibition of 5a-
reductase activity in the presence of 100 mM concentration of the indicated
compound.
and EGCG to inhibit 5a-reductase. These green tea cate-
chins had little inhibitory activity against 5a-reductase in
whole cells. The lack of activity in whole cells may be due
to an inability of these catechins to cross the cell membrane
or to enzymatic or non-enzymatic changes in the structure
of these catechins in assays using whole-cell cultures. The
stability of EGCG in culture medium may be responsible,
in large part, for the lower activity of EGCG in the cell
culture assay, since the half-life of EGCG in culture
medium and phosphate buffer used for whole-cell and
cell-free 5a-reductase assays was 9:5 Æ 0:5 and
74:7 Æ 6:4 min ,mean Æ SEM, N ˆ 3), respectively. The
stability of EGCG in aqueous solution is highly dependent
on pH [45,46], and the difference in pH between culture
medium ,pH 7.5) and the phosphate buffer for cell-free
assays ,pH 7.0) may be responsible, in part, for this 8-fold
difference in stability. The half-life of EGCG in phosphate
buffered saline, pH 7.5, was determined to be 21:2 Æ
2:0 min.
Certain ¯avonoids, including EGCG, produce hydrogen
peroxide in aqueous solutions at physiological pH, possi-
bly through a superoxide intermediate [47,48]. To deter-
mine if these reactive oxygen species may have some role
in inhibition of 5a-reductase by EGCG, we added catalase
,25±250 mg/mL) or superoxide dismutase ,0.5±5 mg/mL)
to assay mixtures containing EGCG. However, addition of
these enzymes did not affect inhibition of 5a-reductase
type 1 or 2 by 20 or 100 mM EGCG. Therefore, peroxide
and superoxide do not appear to be responsible for inhibi-
tion of 5a-reductase by EGCG.
Fig. 4. Kinetic analysis of the inhibition of type 1 5a-reductase by EGCG.
Initial reaction velocities ,V) were determined for different substrate ,panel
A, testosterone; panel B, NADPH) concentrations and as a function of
EGCG concentration: 0 mM ,&), 5 mM ,^), and 10 mM ,*) in panel A
and 0 mM ,&), 20 mM ,^), and 30 mM ,*) in panel B.
EGC derivatives was synthesized and tested using the
cell-free and whole-cell assays ,Table 2). Modi®cation
of the hydroxyl groups of the gallate ester by methylation
or replacement of gallic acid with various aromatic groups
without phenolic groups did not improve inhibitory activity
in either the cell-free or whole-cell assay. The most sig-
ni®cant structural change leading to enhanced activity in
the whole-cell assay was introduction of an aliphatic acid
ester in place of the gallic acid ester of EGCG. EGC
derivatives with long-chain aliphatic acids were better
inhibitors than derivatives with short-chain aliphatic acids,
and derivatives with aliphatic acids with some degree of
unsaturation were better inhibitors than EGC derivatives
esteri®ed with saturated aliphatic acids. EGC esteri®ed to
either g-linolenic or myristoleic acid were potent inhibitors
of both 5a-reductases in whole cells with IC50 values of less
than 15 mM. We have reported previously that certain
polyunsaturated fatty acids, such as g-linolenic acid, are
We performed a kinetic analysis of the inhibition of type
1 5a-reductase by EGCG using the cell-free assay to
determine the mode of inhibition of EGCG [49]. EGCG
was a competitive inhibitor of the substrate NADPH and a
non-competitive inhibitor of the substrate testosterone
based on double-reciprocal plots of the kinetic data ,Fig. 4).
To determine what structural features of the gallate
group of EGCG were important for inhibitory activity
against 5a-reductase, and to determine whether structural
changes in or replacement of the gallate group could
enhance inhibitory activity in whole cells, a series of
inhibitors ,IC : 5±10 mM) of 5a-reductase [30]. Methyl
50
and cholesterol esters of g-linolenic acid were not potent
inhibitors of 5a-reductase ,ic50 > 100 mM) in cell-free and
whole-cell assays ,data not shown). Therefore, it is likely
that the enhanced inhibitory activity of EGC esteri®ed to

R.A. Hiipakka et al. / Biochemical Pharmacology 63 72002) 1165±1176
1171
Tab1e 2
Inhibition of 5a-reductase by various density of EGCa
Group esterified to the 3-hydroxyl of EGC
5a-Reductase
Cell-free assay IC ,mM)
50
Whole-cell assay IC ,mM)
50
Type 1
Type 2
Type 1
Type 2
H ,EGC)
62 ,61)
12 ,99)
31 ,94)
29 ,93)
29 ,97)
48 ,85)
>100 ,35)
47 ,90)
30 ,98)
59 ,95)
20 ,99)
>100 ,31)
25 ,97)
>100 ,30)
73 ,76)
>100 ,15)
>100 ,11)
>100 ,12)
NDb
>100 ,1)
>100 ,5)
>100 ,7)
ND
Gallate ,EGCG)
3,4,5-Trimethoxybenzoate
3,5-Dihydroxy-4-methylbenzoate
4-Hydroxy-3,5-dimethoxybenzoate
Benzoate
>100 ,20)
99 ,51)
>100 ,21)
>100 ,24)
>100 ,8)
>100 ,39)
78 ,74)
43 ,83)
ND
62 ,72)
ND
Acetate ,C2:0)
Caproate ,C6:0)
>100 ,8)
>100 ,10)
58 ,89)
28 ,97)
7 ,99)
>100 ,9)
>100 ,0)
72 ,83)
32 ,98)
8 ,98)
Caprylate ,C8:0)
Myristate ,C14:0)
Myristoleate ,C14:1)
Stearate ,C18:0)
71 ,84)
76 ,96)
>100 ,0)
63 ,93)
42 ,90)
8 ,99)
74 ,81)
14 ,98)
g-Linolenate ,C18:3)
a
IC50: concentration ,mM) of compound producing 50% inhibition of 5a-reductase activity. Values in parentheses are percent inhibition of 5a-reductase
activity in the presence of 100 mM concentration of the indicated compound. EGC and EGCG are presented for comparison and are the natural compounds
isolated from green tea.
b ND: not determined.
g-linolenic acid is due to the combined functionality of this
derivative and not simply due to hydrolysis of the ester
bond and release of free g-linolenic acid. Also, cellular
morphology, as determined by light microscopy, was not
altered when cells were incubated with EGC derivatives
containing g-linolenic or myristoleic acid esters; therefore,
inhibition of 5a-reductase in whole cells was not due to
gross changes in cell integrity.
to modi®cation of the 3-hydroxy group. Taxifolin, a ¯a-
vanone that is structurally similar to quercitin but lacking
the 2,3-double bond in the C-ring, was ineffective against
either isozyme ,ic50 > 100 mM). Biochanin A, kaemp-
ferol, genistein, and daidzein were more effective inhibi-
tors of the type 2 than the type 1 isozyme. With the
exception of kaempferol, a ¯avonol with a single B-ring
hydroxyl, these type 2 inhibitors are iso¯avones with single
hydroxyls on the B-ring. The inhibitory effects of biocha-
nin A, genistein, and daidzein on 5a-reductase have been
reported previously [33]. When tested for inhibitory activ-
ity in whole cells, most ¯avonoids showed little or no
To determine what other structural attributes were
important for inhibition of 5a-reductase by polyphenolic
compounds, we tested a variety of natural and synthetic
polyphenols for their ability to inhibit 5a-reductase iso-
zymes in both the cell-free and whole-cell assays. Several
naturally occurring ¯avonoids with structures related to the
tea catechins were tested ,Fig. 2, Table 3). Four ¯avonoids,
myricetin, quercitin, baicalein, and ®setin, had marked
,ic50 < 100 mM) activity and were more active against
the type 1 than the type 2 isozyme. The number and
position of B-ring hydroxyl groups appear to be important
for inhibitory activity against the type 1 5a-reductase. The
¯avonols quercitin, myricetin, and ®setin, with a catechol
or pyrogallol con®guration in the B-ring ,Fig. 2), had
greater inhibitory activity against the type 1 isozyme than
the ¯avonols chrysin, kaempferol, and morin that lack
hydroxyls in a catechol or pyrogallol con®guration
,Table 3). A comparison of the structures and inhibitory
activities of the ¯avanols EC and EGC and the ¯avonols
myricetin and quercitin highlights the importance of a 2,3-
double bond and a 4-keto group in the C-ring for enhanced
inhibitory activity. In contrast to quercitin, rutin, the 3-
rutinose glycoside of quercitin, was ineffective against
either isozyme ,ic50 > 100 mM). The inactivity of rutin
compared with quercitin may be due to the presence of the
bulky oligosaccharide rutinose causing steric hinderance or
Table 3
Inhibition of 5a-reductase isozymes by various natural polyphenolsa
Polyphenol
5a-Reductase
Cell-free assay IC ,mM)
50
Whole-cell assay IC50 ,mM)
Type 1
Type 2
Type 1
Type 2
Myricetin
Quercitin
Baicalein
Fisetin
23 ,96)
>100 ,31)
>400 ,14)
99 ,51)
>100 ,11)
>100 ,15)
>100 ,24)
>100 ,42)
64 ,64)
>100 ,13)
79 ,60)
>100 ,22)
NDb
>100 ,11)
>100 ,29)
>100 ,4)
>400 ,27)
5 ,93)
7 ,89)
20 ,85)
20 ,89)
ND
23 ,96)
29 ,79)
57 ,97)
>100 ,4)
17 ,74)
29 ,69)
Biochanin A
Daidzein
Kaempferol
Genistein
Morin
100 ,50)
>100 ,3)
>100 ,22)
>100 ,16)
>100 ,6)
>100 ,5)
>100 ,2)
>100 ,4)
12 ,62)
23 ,76)
>100 ,33)
>100 ,5)
>100 ,1)
>100 ,0)
Taxifolin
Chrysin
ND
ND
ND
ND
Rutin
ND
ND
a
IC50: concentration ,mM) of compound producing 50% inhibition of
5a-reductase activity. Values in parentheses are percent inhibition of
5a-reductase activity in the presence of 100 mM concentration of the
indicated compound.
b ND: not determined.

1172
R.A. Hiipakka et al. / Biochemical Pharmacology 63 72002) 1165±1176
Table 4
Inhibition of 5a-reductase isozymes by compounds containing catecholsa
Catechol
5a-Reductase
Cell-free assay IC ,mM)
50
Whole-cell assay IC ,mM)
50
Type 1
Type 2
50 ,97)
Type 1
Type 2
Anthrarobin
4 ,99)
7 ,98)
6 ,91)
NDb
>100 ,31)
ND
Bromopyrogallol red
Gossypol
84 ,58)
21,99)
7 ,99)
7 ,100)
ND
6 ,99)
ND
Pyrogallol red
Nordihydroguaiaretic acid
Caffeic acid phenethyl ester
Octyl gallate
15 ,97)
19 ,99)
25 ,97)
27 ,99)
30 ,81)
42 ,69)
43 ,88)
48 ,85)
70 ,60)
>100 ,13)
>100 ,7)
>100 ,7)
>100 ,5)
>100 ,5)
>100 ,0)
<100 ,2)
>100 ,27)
50 ,80)
19 ,99)
8 ,99)
7 ,99)
ND
22 ,99)
7 ,98)
18 ,94)
ND
>100 ,36)
58 ,90)
Purpurogallin
Hydroxydopamine
Dodecyl gallate
Pyrocatechol violet
Pyrogallol
>100 ,31)
>100 ,41)
>100 ,36)
>100 ,47)
>100 ,28)
>100 ,8)
>100 ,13)
>100 ,9)
>100 ,0)
>100 ,3)
>100 ,0)
>100 ,2)
ND
ND
3 ,99)
ND
7 ,98)
ND
>100 ,7)
ND
>100 ,15)
ND
Caffeic acid
Esculetin
Ellagic acid
ND
ND
ND
ND
Catechol
>100 ,9)
>100 ,0)
>100 ,5)
ND
>100 ,3)
>100 ,0)
>100 ,0)
ND
Methyl gallate
Propyl gallate
Fraxetin
a
IC50: concentration ,mM) of compound producing 50% inhibition of 5a-reductase activity. Values in parentheses are percent inhibition of 5a-reductase
activity in the presence of 100 mM concentration of the indicated compound.
b ND: not determined.
activity against the type 1 isoenzyme, perhaps indicating
limited penetration of these polyhydroxy compounds
across the cell membrane or enzymatic or non-enzymatic
changes in the structure of these compounds in assays
using whole-cell cultures. In contrast to the results with the
type 1 enzyme, four ¯avonoids, biochanin A, daidzein,
genistein, and kaempferol, had signi®cant inhibitory activ-
ities against the type 2 isozyme in the whole-cell assay. The
most active of these, biochanin A and daidzein, have only
two and three free hydroxyl groups, respectively. These
¯avonoids may be active in whole cells because they may
penetrate cells easier than other ¯avonoids that have more
hydroxyl groups. These ¯avonoids also may be less sus-
ceptible to modi®cation in cell cultures.
zymes. Alizarin ,1,2-dihydroxyanthraquinone) and pur-
purin ,1,2,4-trihydroxyanthraquinone), which contain a
catechol group on an anthraquinone backbone, were potent
inhibitors of 5a-reductase. The structural isomers of ali-
zarin, anthra¯avic acid ,2,6-dihydroxyanthraquinone),
anthraru®n ,1,5-dihydroxyanthraquinone), and quinizarin
,1,4-dihydroxyanthraquinone), which do not have cate-
chols in their structure, were weak inhibitors, highlighting
again the importance of catechols for inhibitory activity
against 5a-reductase. The p-quinone structure found in
alizarin may not be necessary for inhibitory activity, since
anthrarobin ,1,2,10-anthracenetriol) had inhibitory activity
Table 5
Inhibition of 5a-reductase isozymes by hydroxyanthraquinonesa
Since ¯avonoids that had catechol moieties were potent
inhibitors of the type 1 5a-reductase, a variety of naturally
occurring and synthetic compounds that have catechol
groups as part of their structure were tested for inhibitory
activity against 5a-reductase ,Tables 4 and 5, Fig. 3).
Anthraquinone 5a-Reductase
Cell-free assay IC ,mM)
50
Whole-cell assay IC ,mM)
50
Type 1
Type 2
Type 1
Type 2
Sixteen compounds had IC values below 100 mM and
50
Purpurin
2 ,95) >100 ,20)
3 ,95) >100 ,54)
30 ,91) >100 ,8)
40 ,67) >100 ,13)
>100 ,27) >100 ,22)
NDb
ND
®ve compounds had IC values below 10 mM. All were
50
Alizarin
6 ,75)
>100 ,22)
ND
>100 ,27)
>100 ,1)
ND
more active against the type 1 than the type 2 isozyme.
Several of these compounds, including alizarin, anthrar-
obin, dodecyl gallate, gossypol, octyl gallate, caffeic acid
phenethyl ester, and nordihydroguaiaretic acid also were
Alizarin red S
Anthrarufin
Anthraflavic
acid
ND
ND
Ouinizarin
>100 ,26) >100 ,7)
ND
ND
inhibitors of 5a-reductase in whole-cell assays with IC
50
a
IC50: concentration ,mM) of compound producing 50% inhibition of
5a-reductase activity. Values in parentheses are percent inhibition of 5a-
reductase activity in the presence of 100 mM concentration of the indicated
compound.
values of less than 20 mM. In the whole-cell assay, anthrar-
obin and alizarin were much more effective against the
type 1 than the type 2 isozyme, whereas the other ®ve
inhibitors were equally effective inhibitors of both iso-
b ND: not determined.

R.A. Hiipakka et al. / Biochemical Pharmacology 63 72002) 1165±1176
1173
equivalent to that of alizarin ,Tables 4 and 5). Alizarin red
S, which is structurally similar to alizarin, but has an
additional 3-sulfate group adjacent to the catechol, was
10±20 times less potent than alizarin ,Table 5). The
charged sulfate group on alizarin red S may interfere with
binding to 5a-reductase, a hydrophobic enzyme, as well as
interfere with the transport of this class of molecule across
the cell membrane. The difference in the activities of
caffeic acid ,ic50 > 100 mM) and caffeic acid phenethyl
ester ,ic50 ˆ 25 mM) ,Table 4) may have a similar expla-
nation. Gossypol, a potent inhibitor in both whole-cell and
cell-free assays, contains two catechol moieties. Both of
these groups could be contributing to the inhibitory activity
of this compound. The methyl and propyl esters of gallic
acid were much less potent inhibitors of 5a-reductase than
the octyl and dodecyl esters. The latter are more hydro-
phobic than the former and may interact more readily with
microsomal 5a-reductase because of their hydrophobic
nature. Dodecyl and octyl gallate were more potent inhi-
bitors in the whole-cell than in the cell-free assay ,Table 4).
The long fatty acid esters on dodecyl and octyl gallate may
enhance uptake of these compounds in whole cells, and
these compounds may concentrate in cell membranes
leading to inhibition of 5a-reductase. Hydroxydopamine
,3,4,5-trihydroxyphenethylamine), which is structurally
similar to the short-chain esters of gallic acid, but is
positively charged at physiological pH, had inhibitory
activity in the cell-free assay that was similar in potency
to that of dodecyl gallate. The compounds catechol ,1,
2-dihydroxybenzene), pyrogallol ,1,2,3-trihydroxyben-
zene), and gallic acid ,3,4,5-trihydroxybenzoic acid), which
havecatecholsintheirstructure, hadweakinhibitoryactivity
in both cell-free and whole-cell assay systems. Three dyes,
bromopyrogallol red ,50,500-dibromopyrogallolsulfoneph-
thalein), pyrocatechol violet ,pyrocatecholsulfonephtha-
lein), and pyrogallol red ,pyrogallolsulfonephthalein),
each containing catechol groups, were potent inhibitors
of 5a-reductase in the cell-free assay. In contrast to alirazin
red S, all three of these dyes are effective inhibitors, even
though they contain a charged sulfate group. This differ-
ence in inhibitory potency may be because the sulfate in
alirazin red S is adjacent to the catechol group, while the
sulfate group in the other dyes that are active inhibitors is
located some distance from the catechol group. Three
naturally occurring catechol-containing compounds, ellagic
acid, a condensation product of two gallic acid molecules,
and the coumarins, esculetin ,6,7-dihydroxycoumarin) and
fraxetin ,7,8-dihydroxy-6-methoxycoumarin), had little
activity ,ic50 > 100 mM) in the cell-free assay.
polyphenol, inhibits 5a-reductase, as well as to identify
other natural product inhibitors of this enzyme. Inhibition
studies were conducted using a cell-free assay, as well as
assays with intact cells. The latter assay may provide some
estimate of the potential of a particular compound to inhibit
5a-reductase activity in vivo. This study identi®ed several
natural products that were inhibitors of 5a-reductase. Since
someofthesecompoundswereeffective onwholecells, they
may be capable of modulating the activity of 5a-reductase
in vivo. Activity in vivo will ultimately be dependent upon
attaining pharmacologically active levels of these agents
in target tissues, which will depend on absorption, distribu-
tion, metabolism, and excretion of the agents.
There is the possibility that diets containing natural
inhibitors of 5a-reductase may adversely affect processes
dependent upon 5a-reductase activity, such as male sexual
differentiation. Mutations in the type 2 5a-reductase gene
have been linked to isolated cases of hypospadias, a
common developmental abnormality of the male repro-
ductive tract [50]. It also has been reported that a maternal
vegetarian diet during pregnancy is associated with
hypospadias [51]. The authors of this study speculated
that phytoestrogens in the maternal diet may have a
deleterious effect on the developing male reproductive
system. It also seems possible that natural or synthetic
dietary components that are potent inhibitors of 5a-reduc-
tase may have adverse effects on male sexual development
when consumed in high levels at critical periods during
pregnancy.
A variety of natural and synthetic polyphenolic com-
pounds were found to be inhibitors of 5a-reductase. Many
of these compounds were better inhibitors of the type 1
than the type 2 isozyme, while a few inhibited both
isozymes equally. Biochanin A, daidzein, genistein, and
kaempferol were the only polyphenols tested that were
better inhibitors of the type 2 than the type 1 isozyme. The
®rst three compounds are iso¯avones, while kaempferol is
a ¯avonol. These compounds are found in a variety of plant
dietary products including soybean-based products. Since
the type 2 isozyme of 5a-reductase has a critical role in
prostate development and is the predominant isozyme
present in the adult human prostate [52], diets rich in these
particular compounds have the potential to affect the
development and function of the prostate by modulating
the activity of 5a-reductase. Since excessive 5a-reductase
activity has been proposed to be a possible contributing
factor in prostate cancer development or progression [19],
the development and progression of prostate cancer may
also be affected by diets containing inhibitors of 5a-
reductase. Certain natural dietary agents may have the
ability to act as prostate cancer chemopreventative agents
by modulating 5a-reductase activity. If some of these
compounds are active in vivo, they may be important
candidates for prostate cancer chemopreventative agents,
either taken in pure or enriched formulations or as com-
ponents of natural dietary products. Differences in the
4. Discussion
We performed this structure±activity relationship study
of 5a-reductase inhibitors in an effort to provide some
insight into the mechanism by which EGCG, a green tea

1174
R.A. Hiipakka et al. / Biochemical Pharmacology 63 72002) 1165±1176
intake of dietary inhibitors of 5a-reductase also may be
responsible, in part, for geographical and racial differences
in prostate cancer mortality.
and pro-oxidant activities, and their ability to form com-
plexes with proteins [59]. Given our current understanding
of the mechanism of 5a-reductase, it does not appear likely
that metal ion complexation or anti- or pro-oxidant activity
would be responsible for inhibition of 5a-reductase by
EGCG and other polyphenols. The tea catechins ECG and
EGCG will form precipitates with soybean lipoxygenase
and yeast alcohol dehydrogenase [61]. We also have
observed that EGCG will rapidly precipitate certain pro-
teins, such as chicken egg white lysozyme ,unpublished
observation). The basis for this precipitation activity has
not been de®ned thoroughly, but it may be due to the ability
of certain polyphenols to form both numerous H-bonds
with a protein, as well as unselective association of the
aromatic nuclei of a polyphenol with certain amino acids,
especially prolines [59]. Types 1 and 2 human 5a-reduc-
tases contain 14 ,5.4%) and 17 ,6.7%) proline residues,
respectively; hence, these enzymes are not proline-rich
proteins. Also, only ®ve of these proline residues are in the
carboxyl-terminal half of the protein containing the puta-
tive NADPH-binding site.
Catechols can form o-quinones, which are known to
react covalently with both primary amines and sulfhydryls
[62]. There are reports that EGCG [63] and other green tea
catechins [64] react covalently with sulfhydryls. Both the
type 1 and 2 5a-reductases are inhibited by sulfhydryl
modifying agents, such as N-ethylmaleimide, 5,50-dithio-
bis,2-nitrobenzoic acid), 2,20-bispyridyldisul®de, p-hydro-
xymercuribenzoate, and mercuric chloride ,unpublished
observation). However, inclusion of 0.1±10 mM dithio-
threitol or b-mercaptoethanol in assays did not prevent
inhibition of 5a-reductase by EGCG ,unpublished obser-
vations). Therefore, it is unlikely that EGCG inhibited
5a-reductase by covalently modifying essential sulfhydryl
groups.
A consistent observation in this study was that poly-
phenolic inhibitors of the type 1 5a-reductase had a
catechol in their structure. Flavonoids that were better
inhibitors of the type 2 than the type 1 5a-reductase,
however, did not contain catechols. Although a catechol
group was necessary for potent inhibition of the type 1
isozyme by polyphenols, it was not always suf®cient. For
instance, the natural product ellagic acid contains two
catechol groups and was a weak ,ic50 > 100 mM) inhibitor
of 5a-reductase. The proximity of the catechols in ellagic
acid to other molecular groups may have steric effects, and
the highly constrained structure of ellagic acid may prevent
inter- and intra-molecular interactions necessary for inhi-
bition by catechol-containing compounds.
Inhibition of the type 1 5a-reductase by EGCG, which
contains two separate catechol/pyrogallol groups in its
structure, was competitive with the substrate NADPH.
Therefore, the catechol groups in EGCG may be interact-
ing with amino acid residues important for binding of this
cofactor by 5a-reductase. Several studies, based upon
characterization of naturally occurring mutants [53,54],
site-directed mutagenesis [55±57], and photoaf®nity label-
ing [58] of 5a-reductase, have identi®ed certain amino acid
residues that may have a role in substrate and cofactor
binding. NADPH-binding is altered by certain amino acid
changes in the carboxyl-terminal half of the protein, while
substrate ,testosterone) or inhibitor ,®nasteride) binding is
affected predominantly by changes in the amino-terminal
half of the protein, although some changes in the carboxyl-
terminal half also affect binding of substrate. Photoaf®nity
‡
labeling of the rat type 1 5a-reductase with 2-azido NADP
modi®es a portion of the carboxyl-terminal half of the
protein that is conserved among human and rat 5a-reduc-
tase isoforms [58]. Since EGCG was a competitive inhi-
bitor of NADPH, it may have inhibited the enzyme by
interactions with residues in the carboxyl-terminal portion
of the protein. Although the inhibition of 5a-reductase
by EGCG was determined to be competitive, we have
observed that inhibition of the type 1 5a-reductase by
100 mM EGCG could not be reversed by pelleting micro-
somes exposed to EGCG and then resuspending them in
new reaction buffer without EGCG. EGCG either must be
strongly bound to microsomes or must permanently alter
5a-reductase causing irreversible inhibition. Inhibition of
5a-reductase by EGCG also did not increase when micro-
somes containing the type 1 or 2 5a-reductase were
incubated with EGCG for 15±60 min prior to the start
of the assay ,unpublished observations).
Acknowledgments
This work was supported, in part, by grants from the
National Institute of Health and the Tang Foundation to
S.L.
References
[1] McGuire JS, Hollis VW, Tomkins GM. Some characteristics of the
microsomal steroid reductases ,5a) of rat liver. J Biol Chem
1960;235:3112±7.
[2] Russell DW, Wilson JD. Steroid 5a-reductase: two genes/two
enzymes. Ann Rev Biochem 1994;63:25±61.
[3] Liao S, Liang T, Fang S, Castaneda E, Shao T-C. Steroid structure
and androgenic activity. Specificities involved in the receptor binding
Natural polyphenols, such as EGCG and certain other
¯avonoids, have been shown to inhibit a variety of enzymes
[59,60]. Three properties of these compounds that may be
responsible for their biological activity are their ability to
form complexes with certain metal ions, their anti-oxidant
and nuclear retention of various androgens.
1973;248:6154±62.
J Biol Chem
[4] Andersson S, Berman DM, Jenkins EP, Russell DW. Deletion of
steroid 5a-reductase 2 gene in male pseudohermaphroditism. Nature
1991;354:159±61.

R.A. Hiipakka et al. / Biochemical Pharmacology 63 72002) 1165±1176
1175
[5] Normington K, Russell DW. Tissue distribution and kinetic
characteristics of rat steroid 5a-reductase isozymes. Evidence for
distinct physiological functions. J Biol Chem 1992;267:19548±54.
[6] Levy MA, Brandt M, Sheedy KM, Holt DA, Heaslip JI, Trill JJ, Ryan
PJ, Morris RA, Garrison LM, Bergsma DJ. Cloning, expression and
functional characterization of type 1 and type 2 steroid 5a-reductases
from Cynomolgus monkey: comparisons with human and rat
isoenzymes. J Steroid Biochem Mol Biol 1995;52:307±19.
[7] Mahendroo MS, Russell DW. Male and female isoenzymes of steroid
5a-reductase. Rev Reprod 1999;4:179±83.
The effects of castration on advanced carcinoma of the prostate
gland. Arch Surg 1941;43:209±23.
[28] Liao S, Hiipakka RA. Selective inhibition of steroid 5a-reductase
isozymes by tea epicatechin-3-gallate and epigallocatechin-3-gallate.
Biochem Biophys Res Commun 1995;214:833±8.
[29] Kao Y-H, Hiipakka RA, Liao S. Modulation of endocrine systems
and food intake by green tea epigallocatechin gallate. Endocrinology
2000;141:980±7.
[30] Liang T, Liao S. Inhibition of steroid 5a-reductase by specific
aliphatic unsaturated fatty acids. Biochem J 1992;285:557±62.
[31] Lesuisse D, Berjonneau J, Ciot C, Devaux P, Doucet B, Gourvest JF,
Khemis B, Lang C, Legrand R, Lowinski M, Maquin P, Parent A,
Schoot B, Teutsch G. Determination of oenothein B as the active 5a-
reductase inhibiting principle of the folk medicine Epilobium
parviflorum. J Nat Prod 1996;59:490±2.
[8] Thigpen AE, Cala KM, Russell DW. Characterization of Chinese
hamster ovary cell lines expressing human steroid 5a-reductase
isozymes. J Biol Chem 1993;268:17404±12.
[9] Mahendroo MS, Porter A, Russell DW, Word RA. The parturition
defect in steroid 5a-reductase type 1 knockout mice is due to
impaired cervical ripening. Mol Endocrinol 1999;13:981±92.
[10] Mahendroo MS, Cala KM, Russell DW. 5a-Reduced androgens play
a key role in murine parturition. Mol Endocrinol 1996;10:380±92.
[11] Mahendroo MS, Cala KM, Landrum CP, Russell DW. Fetal death in
mice lacking 5a-reductase type 1 caused by estrogen excess. Mol
Endocrinol 1997;11:917±27.
[32] Ducrey B, Marston A, Gohring S, Hartman RW, Hostettmann K.
Inhibition of 5a-reductase and aromatase by the ellagitannins
oenothein A and oenothein B from Epilobium species. Planta Med
1997;63:111±4.
[33] Evans BAJ, Griffiths K, Morton MS. Inhibition of 5a-reductase in
genital fibroblasts and prostate tissue by dietary lignans and
isoflavanoids. J Endocrinol 1995;147:295±302.
[12] Imperato-McGinley J, Guerrero L, Gautier T, Peterson RE. Steroid
5a-reductase deficiency in man: an inherited form of male
pseudohermaphroditism. Science 1974;186:1213±5.
[34] Ishiguro K, Oku H, Kato T. Testosterone 5a-reductase inhibitor
bisnapthoquinone derivative from Impatiens balsamina. Phytother
Res 2000;14:54±6.
[13] Wilson JD, Griffin JE, Russell DW. Steroid 5a-reductase
deficiency. Endocr Rev 1993;14:577±93.
2
[35] Shimizu K, Fukuda M, Kondo R, Sakai K. The 5a-reductase
inhibitory components from heartwood of Artocarpus incisus:
structure±activity investigations. Planta Med 2000;66:16±9.
[36] Grunberger D, Banerjee R, Eisinger K, Oltz EM, Efros EM, Caldwell
M, Estevez V, Nakanishi K. Preferential cytotoxicity on tumor cells
by caffeic acid phenethyl ester from propolis. Experientia
1988;44:230±2.
[14] Horton R. Benign prostatic hyperplasia: new insights.
Endocrinol Metab 1992;74:504A±C.
J Clin
[15] Sansone G, Reisner RM. Differential rates of conversion of
testosterone to dihydrotestosterone in acne and in normal human
skin: a possible pathogenic factor in acne. J Invest Dermatol
1971;56:366±72.
[16] Bingham KD, Shaw DA. The metabolism of testosterone by human
male scalp skin. J Endocrinol 1973;25:83±91.
[37] Plattner JJ, Gless RD, Rapoport H. Synthesis of some DE and CDE
ring analogs of camptothecin. J Am Chem Soc 1972;94:8613±5.
[17] Schweiker HU, Wilson JD. Regulation of human growth by steroid
hormones. I. Testosterone metabolism in isolated hairs. J Clin
Endocrinol Metab 1974;38:811±9.
[38] Meth-Cohn O. A simple, powerful and efficient method for
transesterification. J Chem Soc, Chem Commun 1986;1986:695±7.
[39] Nystrom JE, McCanna TD, Helquist P, Iyer RS. Short intramolecular
Diels±Adler approach to functionalized spiro[4.5]decances. Tetra-
hedron Lett 1985;26:5393±6.
[18] Kuttenn F, Mowszowicz I, Schaison G, Mauvais-Jarvis P. Androgen
production and skin metabolism in hirsutism.
1977;25:83±91.
J Endocrinol
[40] Kirschmeier PT, Housey GM, Johnson MD, Perkins AS, Weinstein
[19] Ross RK, Bernstein L, Lobo RA, Shimizu H, Stanczyk FZ, Pike MC,
Henderson BE. 5a-Reductase activity and risk of prostate cancer
among Japanese and US white and black males. Lancet
1992;339:887±9.
IB. Construction and characterization of a retroviral vector
demonstrating efficient expression of cloned cDNA sequences.
DNA 1988;7:219±25.
[41] Pear WS, Nolan GP, Scott ML, Baltimore D. Production of high-titer
helper-free retroviruses by transient transfection. Proc Natl Acad Sci
USA 1993;90:8392±6.
[20] Makridakis N, Ross RK, Pike MC, Chang L, Stanczyk FZ, Kolonel
LN, Shi C-Y, Yu MC, Henderson BE, Reichardt JKV. A prevalent
missense substitution that modulates activity of prostatic steroid 5a-
reductase. Cancer Res 1997;57:1020±2.
[42] Felsher DW, Bishop JM. Transient excess of MYC activity can elicit
genomic instability and tumorigenesis. Proc Natl Acad Sci USA
1999;96:3940±4.
[21] Stoner E. The Finasteride Study Group. The clinical effects of a 5a-
reductase inhibitor, finasteride, on benign prostatic hyperplasia. J
Urol 1992;147:1298±302.
[43] Topp WC. Normal rat cell lines deficient in nuclear thymidine kinase.
Virology 1981;113:408±11.
[22] Kaufman KD, Olsen EA, Whiting D, Savin R, DeVillez R, Bergfeld
W, Price VH, Van Neste D, Roberts JL, Hordinsky M, Shapiro J,
Binkowitz B, Gormley GJ. The Finasteride Male Pattern Hair Loss
Study Group. Finasteride in the treatment of men with androgenetic
alopecia. J Am Acad Dermatol 1998;39:578±89.
[44] Brown AMC, Scott MRD, Retroviral vectors. In: Glover DM, editor.
DNA cloning, a practical approach, vol. III. Oxford: IRL Press, 1987.
p. 189±212.
[45] Zhu Q-Y, Zhang A, Tsang D, Huang Y, Chen ZY. Stability of green
tea catechins. J Agric Food Chem 1997;45:4624±8.
[46] Zimeri J, Tong CH. Degradation kinetics of ,À)-epigallocatechin
gallate as a function of pH and dissolved oxygen in a liquid model
system. J Food Sci 1999;64:753±8.
[23] Brawley OW, Ford LG, Thompson I, Perlman JA, Kramer BS. 5a-
Reductase inhibition and prostate cancer prevention. Cancer
Epidemiol Biomarkers Prev 1994;3:177±82.
[24] Kolonel LN, Nomura AMY, Cooney RV. Dietary fat and prostate
cancer: current status. J Nat Cancer Inst 1999;91:414±28.
[25] Clinton SK, Giovannucci E. Diet, nutrition and prostate cancer. Ann
Rev Nutr 1998;18:413±40.
[47] Miura YH, Tomita I, Watanabe T, Hirayama T, Fukui S. Active
oxygens generation by flavanoids. Biol Pharm Bull 1998;21:93±6.
[48] Long LH, Clement MV, Halliwell B. Artifacts in cell culture: rapid
generation of hydrogen peroxide on addition of ,À)-epigallocatechin,
,À)-epigallocatechin gallate, ,‡)-catechin, and quercitin to com-
monly used cell culture media. Biochem Biophys Res Commun
2000;273:50±3.
[26] Huggins C, Hodges CV. Studies of prostatic cancer. I. The effect of
castration. Cancer Res 1941;1:293±7.
[27] Huggins C, Stevens RE, Hodges CV. Studies of prostatic cancer. II.

1176
R.A. Hiipakka et al. / Biochemical Pharmacology 63 72002) 1165±1176
[49] Dixon M, Webb EC, Enzymes. New York: Academic Press, 1964.
[50] Silver RI, Russell DW. 5a-Reductase type 2 mutations are present in
some boys with isolated hypospadias. J Urol 1999;162:1142±5.
cofactor binding and C-terminal regions. J Steroid Biochem Mol Biol
2001;77:177±82.
[58] Bhattacharyya AK, Chavan AJ, Haley BE, Taylor MF, Collins DC.
Identification of the NADP,H) binding site of rat liver microsomal
5a-reductase ,isozyme-1): purification of a photolabeled peptide
corresponding to the adenine binding domain. Biochemistry
1995;34:3663±9.
[51] North K, Golding J. The ALSPAC Study Team.
A maternal
vegetarian diet in pregnancy is associated with hypospadias. Br J
Urol Int 2000;85:107±13.
[52] Thigpen AE, Silver RI, Guileyardo JM, Casey ML, McConnell JD,
Russell DW. Tissue distribution and ontogeny of steroid 5a-reductase
isozyme expression. J Clin Invest 1993;92:903±10.
[59] Haslam E. Natural polyphenols ,vegetable tannins) as drugs: possible
modes of action. J Nat Prod 1996;59:205±15.
[53] Wigley WC, Prihoda JS, Mowszowicz I, Mendonca BB, New MI,
Wilson JD, Russell DW. Natural mutagenesis study of the human
steroid 5a-reductase 2 isozyme. Biochemistry 1994;33:1265±70.
[54] Makridakis NM, di Salle E, Reichardt JKV. Biochemical and
pharmacogenetic dissection of human steroid 5a-reductase type 2.
Pharmacogenetics 2000;10:407±13.
[60] Liao S, Kao Y-H, Hiipakka R, Green tea: biological and biochemical
basis for health benefits. In: Litwack G, editor. Vitamins and
hormones, vol. 62. New York: Academic Press, 2001. p. 1±94.
[61] Sekiya J, Kajiwara T, Monma T, Hatanaka A. Interaction of tea
catechins with proteins: formation of protein precipitate. Agric Biol
Chem 1984;48:1963±7.
[55] Thigpen AE, Russell DW. Four amino acid segment in steroid
5a-reductase 1 confers sensitivity to finasteride, a competitive
inhibitor. J Biol Chem 1992;267:8577±83.
[62] Bolton JL, Trush MA, Penning TA, Dryhurst G, Monks TJ. Role of
quinones in toxicology. Chem Res Toxicol 2000;13:136±60.
[63] Yasuda H, Arakawa T. Deodorizing mechanism of ,À)-epigalloca-
techin gallate against methyl mercaptan. Biosci Biotech Biochem
1995;59:1232±5.
[56] Baxter FO, Trivic S, Lee IR. Structure±function studies of human
5a-reductase type 2 using site directed mutagenesis. J Steroid
Biochem Mol Biol 2001;77:167±75.
[64] Kennedy DO, Matsumoto M, Kojima A, Matsui-Yuasa I. Cellular
thiol status and cell death in the effect of green tea polyphenols in
Ehrlich ascites tumor cells. Chem Biol Interact 1999;122:59±71.
[57] Bhattacharyya AK, Collins DC. Site-directed mutagenesis studies of
rat steroid 5a-reductase ,isozyme-1): mutation of residues in the