Synthesis, biological evaluation and molecular docking of avicequinone C analogues as potential steroid 5α-reductase inhibitors

Article abstract of DOI:10.1248/cpb.c16-00727

Avicequinone C (5a), a furanonaphthoquinone isolated from the Thai mangrove Avicennia marina has been shown previously to have interesting steroid 5α-reductase type 1 inhibitory activity. In this study, a series of avicequinone C analogues containing furanonaphthoquinone with different degrees of saturation and substituents at the furan ring were synthesized. The resulting synthetic avicequinone C and analogues (5a-f) along with some related compounds including 2,5-dihydroxy-1,4-benzoquinone (6) and natural naphthoquinones such as lawsone (7a) and lapachol (7b) were evaluated for their in vitro cell viability and steroid 5α-reductase type 1 inhibitory activities using the cultured cell line of human keratinocytes (HaCaT). This cell-based bioassay was performed based on a direct detection of the enzymatic product dihydrotestosterone (2) by using a non-radioactive high performance thin layer chromatography (HPTLC) method. Among the furanonaphthoquinones in this series, 5e having a propionic substituent at furan ring possessed approximately 22-fold more potent than the original isolated compound 5a. However, the compounds without furan motif such as 6, 7a and b could not inhibit the activity of steroid 5α-reductase. Molecular docking results of the in silico three-dimensional steroid 5α-reductase type 1-reduced nicotinamide adenine dinucleotide phosphate (NADPH) binary complex was performed via AutoDock Vina and it illustrated that the furanonaphthoquinone moiety and the substituent at furan ring might play a key role as pharmacophores for the steroid 5α-reductase inhibitory activity.

Full text of DOI:10.1248/cpb.c16-00727

Vol. 65, No. 3
Chem. Pharm. Bull. 65, 253–260 (2017)
253
Regular Article
Synthesis, Biological Evaluation and Molecular Docking of Avicequinone C
Analogues as Potential Steroid 5α-Reductase Inhibitors
Wiranpat Karnsomwan,a Ponsawan Netcharoensirisuk,a Thanyada Rungrotmongkol,b
,a
Wanchai De-Eknamkul,a and Supakarn Chamni*
a Department of Pharmacognosy and Pharmaceutical Botany, Faculty of Pharmaceutical Sciences, Chulalongkorn
University; 254 Phayathai Road, Pathumwan, Bangkok 10330, Thailand: and b Structural and Computational Biology
Research Group, Department of Biochemistry, and Ph.D. Program in Bioinformatics and Computational Biology,
Faculty of Science, Chulalongkorn University; 254 Phayathai Road, Pathumwan, Bangkok 10330, Thailand.
Received September 13, 2016; accepted December 25, 2016
Avicequinone C (5a), a furanonaphthoquinone isolated from the Thai mangrove Avicennia marina has
been shown previously to have interesting steroid 5α-reductase type 1 inhibitory activity. In this study, a
series of avicequinone C analogues containing furanonaphthoquinone with different degrees of saturation
and substituents at the furan ring were synthesized. The resulting synthetic avicequinone C and analogues
(5a–f) along with some related compounds including 2,5-dihydroxy-1,4-benzoquinone (6) and natural naph-
thoquinones such as lawsone (7a) and lapachol (7b) were evaluated for their in vitro cell viability and steroid
5α-reductase type 1 inhibitory activities using the cultured cell line of human keratinocytes (HaCaT). This
cell-based bioassay was performed based on a direct detection of the enzymatic product dihydrotestosterone
(2) by using a non-radioactive high performance thin layer chromatography (HPTLC) method. Among the
furanonaphthoquinones in this series, 5e having a propionic substituent at furan ring possessed approxi-
mately 22-fold more potent than the original isolated compound 5a. However, the compounds without furan
motif such as 6, 7a and b could not inhibit the activity of steroid 5α-reductase. Molecular docking results of
the in silico three-dimensional steroid 5α-reductase type 1-reduced nicotinamide adenine dinucleotide phos-
phate (NADPH) binary complex was performed via AutoDock Vina and it illustrated that the furanonaph-
thoquinone moiety and the substituent at furan ring might play a key role as pharmacophores for the steroid
5α-reductase inhibitory activity.
Key words furanonaphthoquinone; avicequinone C analogue; steroid 5α-reductase inhibitor; dihydrotestos-
terone; androgenic alopecia
Steroid 5α-reductase is a membrane bound enzyme in the on the discovery and development of new steroid 5α-reductase
oxidoreductase family, which controls the biological actions inhibitors for baldness treatment, including steroid and
in steroid metabolism.1,2) Presently, three isozymes of this en- non-steroid analogues.9,16) However, most of the reported
zyme are known and their amino acid sequence similarity has 5α-reductase inhibitors have not been designed specifically
been analyzed and compared.3) The steroid 5α-reductase type for the inhibition of steroid 5α-reductase at the catalytic site,
1 is mainly found in the scalp, non-genital skin, sebaceous due to the lack of a single X-ray crystal structure, which has
gland and liver, while steroid 5α-reductase type 2 is specifi- hindered a mechanistic understanding and delayed the discov-
cally locates in the prostate gland, testis, epididymis and scro- ery and development of an effective inhibitor. Thus, several
tum.4,5) Steroid 5α-reductase type 3, on the other hand, has in silico structures of steroid 5α-reductase have recently been
been reported in benign and malignant tissues.6) Physiologi- reported to use as a surrogate structure in the screening of
cally, the overexpression of steroid 5α-reductase, especially potential inhibitors of this enzyme.17–19)
types 1 and 2, is believed to affect the balance between testos-
Previously, it was reported that avicequinone C (5a), a
terone (1) and dihydrotestosterone (2), and such imbalance is furanonaphthoquinone, isolated from the methanolic extract
implicated in androgenic disorders, including prostate cancer, of Avicennia marina, could inhibit the activity of steroid
hirsutism and androgenic alopecia.7)
5α-reductase type 1 at a micromolar scale (IC50=38.8µM).
Androgenic alopecia is described as a form of scalp-hair This finding was based on the use of a newly developed non-
loss caused by decreasing the growth of hair, which may radioactive human hair dermal papilla cell-based assay, which
lead to baldness.8,9) This symptom can happen in both male directly detects dihydrotestosterone (DHT, 2), the enzymatic
and female, although the severity and frequency are greater product, by high performance thin layer chromatography
in men than women and in some specific ethnic groups. The (HPTLC).20,21) Even though the steroid 5α-reductase inhibi-
causes of androgenic alopecia are genetically related, causing tory activity of 5a was not as potent as dutasteride (4), the
some physical disorders that lead to the overproduction of discovery of a new chemical scaffold has provided a potential
5α-dihydrotestosterone.10) The most common prescribed drugs for the development of a new group of steroid 5α-reductase
for androgenic alopecia treatment are finasteride (3), which is inhibitors. Thus, we hypothesized that compounds having a
a steroid 5α-reductase type 2 inhibitor,11,12) and dutasteride (4), furanonaphthoquinone, similar to 5a and a more simplified
which is an effective inhibitor to both 5α-reductase types 1 structure, naphthoquinone, would be used as the potential ste-
and 213–15) (Chart 1). Currently, there have been several reports roid 5α-reductase inhibitors.
*To whom correspondence should be addressed. e-mail: supakarn.c@pharm.chula.ac.th
© 2017 The Pharmaceutical Society of Japan

254
Chem. Pharm. Bull.
Vol. 65, No. 3 (2017)
Chart 1. Inhibition of Steroid 5α-Reductase
homologous protein template of steroid 5α-reductase type 119)
was performed via AutoDock Vina25) to predict the biomo-
lecular interaction between the 5α-reductase-reduced nicotin-
amide adenine dinucleotide phosphate (NADPH) binary com-
plex and the respective furanonaphthoquinones. Herein, an
essential pharmacophore and inhibitory activity of a potential
steroid 5α-reductase inhibitor are discussed.
Results and Discussion
The synthetic protocol involved the two-step transforma-
tions. First, the reaction of one-pot Michael-O-alkylation
pertaining bromination and base-mediated cyclization was
employed to construct a furan ring.26–28) Second, the resulting
2-acetyl furanonaphthoquinone was subjected to methylation
via Grignard’s reaction to form an isopropyl alcohol side chain
(Chart 3). Under this strategy, lawsone (7a) was used as the
starting material and was treated with an effective Michael ac-
ceptor, such as methyl vinyl ketone (8a), pen-1-en-3-one (8b),
and methyl vinyl sulfone (8c) to obtain the desired furanon-
aphthoquinones 5a–f at a yield of 12–57%. During base-medi-
ated cyclization, refluxing gave furans 5b and e, while heating
Chart 2. Avicequinone C Analogues (5a–f), Quinone 6 and Naphtho-
quinones 7a and b as Potential Steroid 5α-Reductase Inhibitors
In this study, we focused on the synthesis and biological at 60°C furnished 5c as the major product. For the preparation
study of potential steroid 5α-reductase inhibitors from a se- of 5f, initial optimizations were examined with methyl vinyl
ries of avicequinone C analogues (5a–f) that contain different sulfone (8c), but only low yields were obtained. Next, the
degrees of saturation and substituents at the furan ring. The resulting furanonaphthoquinones 5b and c were reacted with
structural related compounds including 2,5-dihydroxy-1,4- methyl magnesium bromide to convert their acetyl group to
benzoquinone (6) along with natural naphthoquinones such isopropyl alcohol, which gave an acceptable yield. Under this
as lawsone (2-hydroxy-1,4-naphthoquinone, 7a) and lapachol condition, 5a was synthesized at a gram-scale, with a yield of
(2-hydroxy-3-(3-methylbut-2-enyl)naphthalene-1,4-dione, 7b) up to 19% yield.
were employed to investigate the assumption about the im-
The desired avicequinone C (5a), 2,3-dihydronaphtho[2,3-
portance of the furanonaphthoquinone moiety in the inhibi- b]furan analogues (5c and d), and furanonaphthoquinone
tion of steroid 5α-reductase type 1 (Chart 2). Avicequinone analogues (5b, e, and f) were characterized by standard
C and analogues (5a–f) were synthesized from lawsone (7a) spectroscopic analyses. Spectroscopic data of 5a, b, d, and f
by chemical transformations involving the two-step-one-pot were matched with the reported natural furanonaphthoquinone
Michael-O-alkylation followed by methylation. The resulting structures of avicequinone C,20) napabucasin,29) ( )-stenocar-
synthetic furanonaphthoquinones analogues 5a–f, quinone 6, poquinone B,30) and avicequinone B,31) respectively.
and naphthoquinones 7a and b were then evaluated for their in
Next, all synthetic avicequinone C analogues 5a–f, quinone
vitro cytotoxicity and steroid 5α-reductase inhibitory activities 6 and naphthoquinones 7a and b were evaluated for their in
using HaCaT, a human keratinocyte cell line. This cell line has vitro cytotoxicity and steroid 5α-reductase inhibitory activities
been shown to be a source of testosterone (1) metabolism22,23) using the HaCaT cell line. The results would verify the impor-
and to have steroid 5α-reductase type 1 expression.24) Thus, tance of the furanonaphthoquinone moiety, which could be a
the HaCaT cell line could serve as a general cell-based assay key pharmacophore for the inhibition of 5α-reductase activity.
for the study of steroid 5α-reductase inhibition. The steroid
As the results shown in Table 1, 5a and its synthetic fu-
5α-reductase inhibition was measured based on the amount of ranonaphthoquinone analogues (5b–f) exhibited an in vitro
dihydrotestosterone (2) produced, determined using HPTLC HaCaT cellular toxicity in the range of 2.66–155.79µM.
as previously described.20,21) Molecular docking using the The most potent cytotoxic compound in this series was 5b

Vol. 65, No. 3 (2017)
Chem. Pharm. Bull.
255
r.t.: room temperature.
Chart 3. Synthesis of Avicequinone C (5a) and Analogues (5b–f)
Table 1. HaCaT Cell-Based MTT Assay of Avicequinone C Analogues
(5a–f), Quinone 6 and Naphthoquinones 7a and ba)
Cytotoxicity
LC50 S.D. (µM)
Cytotoxicity
LC50 S.D. (µM)
Compound
Compound
5a
5b
5c
5d
5e
28.06 1.94
2.66 0.57
2.92 0.48
155.79 1.58
3.94 0.62
5f
6
25.09 2.37
>200
>200
7a
7b
>200
a) Cytotoxicities were examined at 24h.
Fig. 1. Preliminary Screening of Steroid 5α-Reductase Inhibitory Ac-
tivity with HPTLC Analysis of Avicequinone C Analogues
(LC50=2.66µM), which is a furanonaphthoquinone contain-
ing a conjugated structure with an acetyl group at the furan
ring. We observed that 5b was 10-fold more cytotoxic than 5a
(LC50=28.06µM), which contains an isopropanol substituent
T refers to testosterone (1) and DHT refers to dihydrotestosterone (2). 5a, d, and
f at 5µM; b, c, and e at 0.5µM; and dutasteride (DU, 4) at 20µM were employed as
positive control in this experiment.
at the furan ring. Additionally, 5b showed a broadly similar acetate for extracting the remaining 1, which was added as an
cytotoxic level to 5c (LC50=2.92µM) and 5e (LC50=3.94 µM), enzyme substrate and the resulting 2, the steroid 5α-reductase
which have a 2,3-dihydrofuran ring and a one-carbon lon- product. The steroid 5α-reductase activity was then evaluated
ger propionic group at furan ring, respectively. Without the based upon the amount of 2 by HPTLC using a mixture of
substituent at furan ring, 5f (LC50=25.09 µM) displayed equal cyclohexane–ethyl acetate–triethylamine (1.5:1:0.1, v/v/v) as
HaCaT cellular toxicity to 5a. The loss of both the electron the mobile phase. Then, 1 and 2 were visualized under 366nm
withdrawing group and conjugated system by having a 2,3-di- on a phosphoric acid staining TLC plate. We found that the
hydrofuran moiety and isopropanol substituent, 5d appeared furanonaphthoquinone analogues (5a–f) enabled to inhibit the
to tremendously reduce cytotoxicity of the compound. Inter- activity of steroid 5α-reductase type 1 (Fig. 1).
estingly, 5d showed the lowest cytotoxicity (LC50=155.79 µM)
Furthermore, the steroid 5α-reductase type 1 inhibitory ac-
among compounds in this series against HaCaT cell line. This tivity of furanonaphthoquinones 5a–f, along with quinone 6,
observation was similar to the cytotoxic results of 6, 7a, and synthetic precursor 7a (lawsone), and natural naphthoquinone
b; the quinone and naphthoquinones without the furan ring, 7b (lapachol) were evaluated for their IC50 profiles. HaCaT
which showed no toxicity at 200µM against the HaCaT cell cells were treated with various non-cytotoxic concentrations
line. In this way, we speculated that furanonaphthoquinone of 5a–f, 6, 7a, and b that were chosen based on cytotoxicity
moiety and the electron withdrawing substituent at the furan obtained at 24h (Table 1). As shown in Fig. 2 and Table 2,
ring might play a role in the cell viability against the HaCaT furanonaphthoquinones 5a–f exhibited the inhibitory activity
cell line.
in the range of 0.20–9.63µM. It was found that both furano-
Next, avicequinone C and synthetic analogues (5a–f) along naphthoquinones 5a and f displayed similar degree of inhibi-
with synthetic precursor (7a) and the non-furanonaphthoqui- tion (IC50=4.45 and 4.26µM, respectively). The less cytotoxic
none compounds (6 and 7b) were preliminary screened for 5d; the 2,3-dihydrofuran analog having isopropanol substitu-
their steroid 5α-reductase type 1 inhibitory activity using ent resemble to 5a, showed 2-fold less inhibitory activity
the HaCaT cells for steroid 5α-reductase type 1 expression to toward steroid 5α-reductase type 1 (IC50=9.63 µM), relative
convert testosterone (1) to dihydrotestrosteron (2) in conjuga- to 5a. The group of highly toxic analogues 5c and e, which
tion with the non-radioactive HPTLC method for the detec- contained the ketone substituents adjacent to 2,3-dihydrofuran
tion of 2.20) In this work, the HaCaT cells were cultured and and furan moieties possessed strong steroid 5α-reductase type
treated with all nine test compounds including 5a–f, 6, 7a, 1 inhibitory profile, approximately 4- and 22-fold more inhibi-
and b using the highest concentration, which no cell death tory activity (IC50=1.21 and 0.20µM), respectively, relative to
more than half population was observed. Aqueous medium 5a. The steroid 5α-reductase inhibitory activity of the most
of each experiment was collected and partitioned with ethyl potent cytotoxic 5b could not be examined due to the death of

256
Chem. Pharm. Bull.
Vol. 65, No. 3 (2017)
Fig. 2. Cell Viability and Steroid 5α-Reductase Enzyme Activity of Avicequinone C Analogues (5a–f), Quinone 6 and Naphthoquinones 7a and b
toward HaCaT Cell-Based Assay at 48h
type 1, the docking studies were performed. The in silico
three-dimensional (3D) 5α-reductase type 1-NADPH binary
complex was prepared based on the previously reported strate-
gy.19) The binding pattern and intermolecular interaction of all
furanonaphthoquinone analogues (5a–f) were predicted (Fig.
3). The furanonaphthoquinones 5a–e were resided closely
adjacent to NADPH and likely interacted with Tyr95 of the in
silico 5α-reductase type 1 as well as NADPH. Note that the
oxygen atom of the substituent at the furan ring was in cor-
respondence to the reported result of 4.19) The binding site of
furanonaphthoquinone 5f to the 5α-reductase type 1-NADPH
binary complex was slightly different from the others; how-
ever its oxygen atom on quinone motif was able to form a
hydrogen bond with Tyr95. The molecular docking results
Table 2. Steroid 5α-Reductase Inhibitory Activity (IC50) via HaCaT Cell-
Based Assay at 48h of Avicequinone C Analogues (5a–f), Quinone 6, and
Naphthoquinones 7a and ba)
5α-Reductase inhibitory
activity IC50 S.D. (µM)
5α-Reductase inhibitory
activity IC50 S.D. (µM)
Compound
Compound
5a
5b
5c
5d
5e
4.45 0.42
n.d.b)
5f
6
4.26 0.38
>200
>200
1.21 0.12
9.63 0.78
0.20 0.03
7a
7b
>200
a) Dutasteride (20µM) was employed as a control and provided complete inhibi-
tion. b) n.d. refers to not determined.
HaCaT cell line. Thus, the cell-free enzyme activity assay will emphasized the importance of the substituents at furan ring
be required for further study of 5b. Interestingly, the non-fu- and quinone moieties; however the structure–activity relation-
ranonaphthoquinones 6, 7a and b were found to be incapable ship might not be intensively concluded here. The further in
to inhibit the steroid 5α-reductase activity. According to the vitro study is required to identify the inhibitory mechanism of
results, it seems that compounds with furanonaphthoquinone furanonaphthoquinone toward steroid 5α-reductase inhibition.
such as 5a–f showed a promising 5α-reductase inhibitory
activity, which could potentially serve as a new 5α-reductase
Conclusion
inhibitor that will be further studied and developed for the
In conclusion, the synthesis of six furanonaphthoquinone
treatment of androgenic disorder such as prostate cancer, analogues (5a–f) was achieved. All synthetic furanonaphtho-
benign prostatic hyperplasia, acne, hirsutism, and androgenic quinones (5a–f), the commercial available quinone 6, naph-
alopecia.2)
thoquinone 7a, employed as a synthetic starting material in
To rational the structure–activity relationship of furanon- this work, and the known natural naphthoquinone 7b were
aphthoquinone analogues (5a–f) toward steroid 5α-reductase examined for their anti-proliferative and steroid 5α-reductase

Vol. 65, No. 3 (2017)
Chem. Pharm. Bull.
257
Fig. 3. Molecular Docking Results of Furanonaphthoquinone Analogues (5a–f) with the In Silico Three-Dimensional 5α-Reductase Type 1-NADPH
Binary Complex
type 1 inhibitory activity against the HaCaT cell line. The Room temperature was 25°C. Commercial solvents and re-
steroid 5α-reductase type 1 inhibitory activity was measured agents were used as received. Anhydrous solvents were dried
based on the production of 5α-dihydrotestosterone (2) via the over 4Å molecular sieves. All reactions were monitored by
non-radioactive HPTLC procedure. The series of furanonaph- TLC using aluminium silica gel 60 F254 (Merck). IR spectra
thoquinones (5a–f) showed interesting cellular toxicity and were measured on a Fourier transform (FT)-IR spectrometer,
enabled to inhibit the activity of steroid 5α-reductase type 1 PerkinElmer, Inc. (U.S.A.) 1H- and 13C-NMR spectra were
of the HaCaT cell line with the improved inhibitory activity, obtained on a Bruker Avance DPX-300 FT-NMR spectrom-
1
relative to avicequinone C (5a); the natural furanonaphthoqui- eter. H-NMR chemical shifts (δ) and coupling constants (J)
none isolated from Avicennia marina. To our surprise, 5e; the were given in ppm and Hz, respectively. Deuterated chloro-
1
most cytotoxic compound among all synthetic furanonaphtho- form (CDCl3) served as an internal standard for both H- and
quinones having a propionic substituent, exhibited approxi- 13C-NMR spectra at 7.26 and 77.0ppm, respectively. Mass
mately 22-fold more potent toward the steroid 5α-reductase spectra were recorded on a microTOF Bruker Daltonics mass
inhibitory activity, comparing to 5a. Considering the results spectrometer.
between the in vitro steroid 5α-reductase inhibitory activity
Synthesis of Avicequinone C (5a) To a stirred solution
and molecular docking of the in silico 3D steroid 5α-reductase of methyl vinyl ketone (8a) (2.5 equiv, 7.18mmol, 0.6mL)
type 1-NADPH binary complex, we concluded that the fura- in 6mL of pentane was added a solution of bromine (2.6
nonaphthoquinone moiety and the substituent at furan ring are equiv, 7.46mmol, 0.3mL) in 3mL of pentane at −15°C
the potential pharmacophores for the steroid 5α-reductase in- (NaCl mixed ice bath). The reaction was stirred for 10min.
hibitory activity. New chemical feature for 5α-reductase inhib- Then, all volatiles were evaporated under reduced pressure
itor displaying a stronger 5α-reductase inhibition than 5a was to yield the crude 1,2-dibromobutan-2-one, which was fur-
discovered. Further studies in molecular dynamics along with ther used in the next step without purification. Next, law-
in vitro experiments could provide insight information regard- sone (7a) (1 equiv, 2.87mmol, 500mg) was weighted into
ing the biological mechanisms toward the steroid 5α-reductase a 50-mL oven-dried round-bottomed flask and dissolved
metabolism. Finally, a new series of furanonaphthoquinone in 15mL of tetrahydrofuran (THF), followed by the addi-
will be developed as the promising steroid 5α-reductase in- tion of 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) (3.7 equiv,
hibitors for the treatment of androgenic disorders such as 10.62mmol, 1.6mL), giving a red solution. The reaction was
androgenic alopecia.
stirred at 0°C (ice bath) under a nitrogen atmosphere for
20min. A freshly prepared solution of 1,2-dibromobutan-2-one
in 3mL of THF was added and the ice bath was immediately
Experimental
Materials Reactions were performed in oven-dried glass- removed. The reaction mixture was gradually warmed up and
ware and magnetically stirred under an inert atmosphere using was then heated to reflux for 5h. Another portion of DBU
a syringe tube equipped with an argon or nitrogen balloon. (0.52 equiv, 1.49mmol, 0.2mL) was added and the reaction

258
Chem. Pharm. Bull.
Vol. 65, No. 3 (2017)
mixture was continuously refluxed and stirred for another 3h. 3093, 2949, 2860, 1737, 1682, 1652, 1631, 1591, 1395, 1375,
The reaction was then quenched by the addition of saturated 1361, 1242, 1195, 963, 722cm−1; HR-MS-ESI m/z 281.0223
aqueous ammonium chloride and extracted with chloroform ([M+K]+, calcd for C14H10O4K+ 281.0211). The resulting 5c
(25 mL×3 times). The organic extracts were combined, washed was subjected to methylation using a similar reaction condition
with water and saturated sodium chloride, dried over anhy- as the preparation of 5a, to give 2-(2-hydroxypropan-2-yl)-2,3-
drous Na2SO4, filtered and evaporated under reduced pressure dihydronaphtho[2,3-b]furan-4,9-dione (5d) as a yellow powder
1
to obtain the crude product. Purification of the crude reaction at a yield of 8.8mg (18%). H-NMR (CDCl3, 300MHz) δ: 8.08
by silica gel column chromatography eluting with dichloro- (1H, m, 5-H), 8.08 (1H, m, 8-H), 7.70 (1H, m, 6-H), 7.70 (1H,
methane–hexanes (4:1, v/v) provided 2-acetylnaphtho[2,3-b]- m, 7-H), 4.85 (1H, t, J=9.9Hz, 2-H), 3.17 (2H, d, J=9.9Hz,
furan-4,9-dione (5b) as a yellow powder at a yield of 420mg 3-H), 1.40 (3H, s, 11-Ha), 1.28 (3H, s, 11-Hb); 13C-NMR
(57%). 1H-NMR (CDCl3, 300MHz) δ: 8.26 (1H, m, 5-H), (CDCl3, 100MHz) δ: 182.2 (C-4), 177.7 (C-9), 159.9 (C-9a),
8.25 (1H, m, 8-H), 7.81 (1H, m, 6-H), 7.81 (1H, m, 7-H), 7.61 134.2 (C-4a), 133.0 (C-6), 133.0 (C-7), 131.5 (C-8a), 126.3
(1H, s, 3-H), 2.67 (3H, s, 11-H); 13C-NMR (CDCl3, 100MHz) (C-8), 126.1 (C-5), 125.0 (C-3a), 71.7 (C-10), 92.1 (C-2), 29.7
δ: 187.5 (C-10), 179.7 (C-4), 173.9 (C-9), 155.5 (C-9a), 152.9 (C-3), 25.8 (C-11a), 24.1 (C-11b); Characterization of carbonyl
(C-2), 134.4 (C-6), 134.3 (C-7), 133.1 (C-8a), 132.6 (C-4a), groups of quinone at C-4 and C-9 were based on reported
130.7 (C-3a), 127.4 (C-8), 127.2 (C-5), 112.3 (C-3), 26.8 (C-11); data27); IR (KBr) 3436, 2925, 2854, 1686, 1655, 1632, 1597,
IR (KBr) 3113, 3014, 2854, 1690, 1674, 1581, 1359, 1285, 1464, 1376, 1200, 1079, 970cm−1; HR-MS-ESI m/z 281.0793
1259, 1224, 1197, 977, 875, 717cm−1; high resolution (HR)- ([M+Na]+, calcd for C15H14O4Na+ 281.0784).
MS-electrospray ionization (ESI) m/z 263.0321 ([M+Na]+,
Synthesis of 5e The propionyl containing furanonaph-
calcd for C14H8O4Na+ 263.0315). The resulting 5b (1 equiv, thoquinone 5e was prepared by following the synthesis of
0.208mmol, 50mg) was weighted into a 50-mL oven-dried 5b. Pen-1-en-3-one (8b) was employed as the Michael-O-
round-bottomed flask and dissolved in 20mL of THF. The alkylation electrophile to give 2-propionylnaphtho[2,3-b]-
reaction mixture was stirred at room temperature under an furan-4,9-dione (5e) as a yellow powder at a yield of 259mg
argon atmosphere. Methyl magnesium bromide (3.0M in di- (34%). 1H-NMR (CDCl3, 300MHz) δ: 8.26 (1H, m, 5-H),
ethylether, 6 equiv, 1.249mmol, 0.4mL) was added to the 8.23 (1H, m, 8-H), 7.81 (1H, m, 6-H), 7.81 (1H, m, 7-H), 7.61
reaction mixture. The reaction mixture was stirred at room (1H, s, 3-H), 3.06 (2H, q, J=7.5, 14.7Hz, 11-H), 1.26 (3H, t,
temperature for 6h. After completion, the reaction mixture J=7.2Hz, 12-H); 13C-NMR (CDCl3, 100MHz) δ: 190.8 (C-10),
was quenched by the addition of aqueous hydrochloric acid 179.8 (C-4), 173.9 (C-9), 155.4 (C-9a), 152.7 (C-2), 134.4 (C-7),
(2 N, 15mL) and stirred at room temperature for 30min. The 134.3 (C-6), 133.1 (C-8a), 132.6 (C-4a), 130.7 (C-3a), 127.3
reaction mixture was then evaporated under reduced pressure (C-5), 127.3 (C-8), 112.0 (C-3), 32.6 (C-11), 7.4 (C-12); IR
until about 10mL of reaction mixture remained. The reaction (KBr) 3385, 3115, 2978, 2878, 1698, 1672, 1579, 1567, 1354,
mixture was added water (5mL), extracted with ethyl acetate 1219, 958, 721cm−1; HR-MS-ESI m/z 277.0464 ([M+Na]+,
(35 mL×3 times) and washed with saturated sodium chlo- calcd for C15H10O4Na+ 277.0471).
ride (20mL). The organic layers were collected, dried over
Synthesis of 5f Methyl vinyl sulfone (8c) (1.0 equiv,
anhydrous Na2SO4 and filtered. The reaction mixtures were 4.71mmol, 500mg) was weighted into a 50-mL oven-dried
evaporated to obtain the crude compound. Purification by round-bottomed flask and dissolved in dichloromethane
silica gel column chromatography using ethyl acetate–hexanes (10mL). The reaction mixture was stirred at room temperature
(3:7, v/v) as eluent provided avicequinone C (5a) as a brown under an argon atmosphere. Bromine (1.5 equiv, 7.07mmol,
1
solid at a yield of 17.2mg (34%). H-NMR (CDCl3, 300MHz) 0.2mL) was slowly added into the reaction mixture giving
δ: 8.21 (1H, m, 5-H), 8.18 (1H, m, 8-H), 7.75 (1H, m, 6-H), the dark orange solution. The reaction mixture was heated to
7.75 (1H, m, 7-H), 6.82 (1H, s, 3-H), 1.69 (3H, s, 11-Ha), 1.69 reflux for 6h. The reaction mixture was then concentrated to
(3H, s, 11-Hb); 13C-NMR (CDCl3, 100MHz) δ: 180.8 (C-4), yield a sticky residue, dissolved in THF (20mL) and cooled
173.4 (C-9), 167.9 (C-9a), 151.8 (C-2), 133.9 (C-7), 133.8 (C-6), at 0°C in an ice-bath under an argon atmosphere. DBU (1.5
133.0 (C-8a), 132.5 (C-4a), 131.3 (C-3a), 126.9 (C-5), 127.0 equiv, 7.07mmol, 1.1mL) was slowly added dropwise over
(C-8), 102.6 (C-3), 69.4 (C-10), 29.7 (C-11a), 28.8 (C-11b); IR 20min. The reaction mixture was stirred for 30min at 0°C
(KBr) 3532, 3388, 3178, 3121, 2986, 2926, 1682, 1665, 1586, in an ice-bath under an argon atmosphere. Lawsone (7a) (1.0
1376, 1231, 1165, 1154, 968, 954, 724cm−1; HR-MS-ESI m/z equiv, 4.71mmol, 820.2mg) was added and another portion
279.0635 ([M+Na]+, calcd for C15H12O4Na+ 279.0628).
of DBU (1.5 equiv, 7.07mmol, 1.1mL) was slowly added
Synthesis of 5c and d Dihydrofuran 5c was prepared by dropwise over 20min. The reaction mixture was stirred for
following the synthesis of 5b. The reaction temperature during 30min at 0°C in an ice-bath under an argon atmosphere. The
the DBU-mediated cyclization was controlled at 60°C to ob- ice-bath was then removed. The reaction was warmed up to
tain 2-acetyl-2,3-dihydronaphtho[2,3-b]furan-4,9-dione (5c) as room temperature and heated to reflux for 6h. The reaction
1
a yellow powder at a yield of 263mg (38%). H-NMR (CDCl3, was then concentrated under reduced pressure and the resi-
300MHz) δ: 8.08 (1H, m, 5-H), 8.08 (1H, m, 8-H), 7.72 (1H, due was dissolved in dichloromethane (100mL), washed with
m, 6-H), 7.72 (1H, m, 7-H), 5.27 (1H, t, J=9.6Hz, 2-H), 3.42 water (100mL) and saturated aqueous ammonium chloride
(2H, dd, J=2.1, 10.8Hz, 3-H), 2.39 (3H, s, 11-H); 13C-NMR (100mL). The organic layer was separated and the aqueous
(CDCl3, 100MHz) δ: 204.6 (C-10), 181.7 (C-4), 177.2 (C-9), layer was extracted with dichloromethane (50mL×3 times).
159.2 (C-9a), 134.4 (C-7), 133.3 (C-6), 132.8 (C-4a), 131.4 The combined organic layer was dried over anhydrous Na2SO4
(C-8a), 126.4 (C-8), 126.2 (C-5), 123.8 (C-3a), 87.1 (C-2), 30.1 and concentrated to obtain the crude product. The crude
(C-3), 26.5 (C-11); Characterization of carbonyl groups of qui- product was purified over silica gel column chromatography
none at C-4 and C-9 were based on reported data27); IR (KBr) using dichloromethane–hexanes (3:1, v/v) as eluent to provide

Vol. 65, No. 3 (2017)
Chem. Pharm. Bull.
259
naptho[2,3-b]furan-4,9-dione (5f) as a pale yellow solid at a mean standard deviations (S.D.) were obtained from three
1
yield of 69mg (12%). H-NMR (CDCl3, 300MHz) δ: 8.22 (1H, independent experiments. The IC50 values were calculated ac-
m, 5-H), 8.22 (1H, m, 8-H), 7.77 (1H, m, 6-H), 7.77 (1H, m, cording to the cell viability ratios.
7-H), 7.77 (1H, d, J=1.5Hz, 2-H), 7.01 (1H, d, J=1.5Hz, 3-H);
HaCaT-Based Steroid 5α-Reductase Inhibitory Assay
13C-NMR (CDCl3, 100MHz) δ: 180.6 (C-4), 173.6 (C-9), 152.7 HaCaT cells were seeded at a cell density of 2×105 cells/well
(C-9a), 148.6 (C-2), 132.5 (C-4a), 134.0 (C-7), 133.9 (C-6), onto 6-well plates, and allowed to adhere for 24h. Testos-
133.2 (C-8a), 130.5 (C-3a), 127.1 (C-8), 127.0 (C-5), 108.7 (C-3); terone (1, substrate) and 5α-dihydrotestosterone (2, product)
IR (KBr) 3142, 2853, 1683, 1585, 1566, 1478, 1365, 1206, 1182, were employed as the internal standard and enzyme–activity
952, 714cm−1; HR-MS-ESI m/z 221.0212 ([M+Na]+, calcd for control, respectively. Dutasteride (4) at 20µM was used as a
C12H6O3Na+ 221.0215). Spectroscopic data of 5f were matched positive control. The overall volume of each well was 2mL.
with reported data of avicequinone B.31)
DMEM (1mL) was added into each standard and control
Molecular Docking The 3D predicted structures of wells. A solution of testosterone (1, substrate) at 50µM (1mL)
5α-reductase type 1 was achieved by following the reported in DMEM containing 1% (v/v) DMSO was added in each
protocol.32) The homology modeling strategies involved au- test well. The final concentration of 1 was 25µM. Cells were
tomated homologous protein template searching by SWISS- treated with a serial solution at non-cytotoxicity of each test
MODEL. The isoprenylcysteine carboxyl methyltransferase compound that was dissolved in DMEM containing 1% (v/v)
(ICMT, PDB code: 4A2N) was automatically selected as a DMSO (1mL). Each test was repeated in triplicate. After in-
homologous protein template for building the in silico 3D cubating in humidified atmosphere of CO2 at 37°C for 48h,
structure of 5α-reductase type 1. Ramachandran plot for stan- the cell culture medium was collected and the remaining
dard in silico structure evaluation, active site identification viable cells were determined by MTT assay as previously
by docking study of finasteride (3) and dutasteride (4) using explained. The aqueous medium was extracted with ethyl
AutoDock Vina,25) transmembrane site prediction via Dis- acetate (1.5mL), vigorously shaken, centrifuged at 10000rpm
covery Studio 2.5, and docking of the reported 5α-reductase for 5min, and the 1mL of ethyl acetate layer was kept. The
inhibitors to confirmed the active site and protein-ligand in- extraction was repeated two times. The resulting ethyl acetate
teractions. The 3D structures of all studied compounds were extracts that consisted of the remaining 1 and the resulting 2
generated by the Open Babel software.33) Each compound were combined and air dried at room temperature. The crude
was then docked into the resulting in silico 5α-reductase type extract was then reconstituted with 20µL of methanol and
1-NADPH binary complex with 100 independent runs using 8µL of the resulting methanol solution was spotted onto a
AutoDock Vina.25) Then, each ligand was docked into the TLC silica gel 60 F254 aluminum plate using TLC sample
resulting in silico 5α-reductase type 1-NADPH binary com- loader (CAMAG Linomat 5, Switzerland). The TLC plate was
plex with the grid box near the position of NADPH (center developed using a mixture of cyclohexane–ethyl acetate–tri-
x=0.487, center y=36.43, and center z=96.802) and the grid methylamine (1.5:1:0.1, v/v/v) as a mobile phase. Then, the
box size is of 38×18×16 Å3. The docking calculations of developed TLC plate was stained with a solution of 42.5%
5α-reductase type 1 were performed with 100 runs and the phosphoric acid in ethanol and heated at 120°C for 20min
average results were analyzed.
for visualization. Visual detection of 5α-dihydrotestosterone
In Vitro Cell Viability 3-(4,5-Dimethylthiazol-2-yl)-2,5- (2) at 366nm was captured using the TLC scanner (CAMAG
diphenyltetrazolium Bromide (MTT) Assay HaCaT cells TLC Scanner 3, Switzerland). The band intensity of 2 was
were grown in Dulbecco’s modified Eagle’s medium (DMEM) measured by the Image J software and the inhibitory activity
containing 10% (v/v) fetal bovine serum (FBS) and 1% (v/v) was calculated according to the inhibition ratios. In order to
antibiotic–antimycotic solution at 37°C in humidified atmo- avoid false positive result, the cell viability of the treated cells
sphere of 5% CO2 in a T-75 flask. The cells were grown until was determined. The test compound at concentration showing
at 80% confluency, and then subcultured. Cells were seeded more than 50% cell viability was further considered for IC50
at an initial cell density of 5000 cells/well onto 96-well plates calculation. The IC50 values were calculated according to the
and allowed to adhere overnight. After 24h, the cells were inhibition ratios. The mean S.D. were obtained from three
individually treated with a series of synthetic avicequinone independent experiments.
C
and analogues (5a–f), 2,5-dihydroxy-1,4-benzoquinone
(6), and natural naphthoquinones including lawsone (7a) and
Acknowledgments This research has been supported by
lapachol (7b). Dutasteride (4) was used as a positive control. the Ratchadaphiseksomphot Endowment Fund 2013 of Chul-
Each test compound was prepared as the serial concentrations alongkorn University (CU-56-334-HR) to S.C., the PERDO’s
in DMEM containing 0.1% (v/v) dimethyl sulfoxide (DMSO). Center of Excellence on Medical Biotechnology (CEMB) pro-
DMEM was used as a negative control. Cells were incubated gram to W.D. and the Thailand Research Fund (IRG5780008)
at 37°C in atmosphere of 5% CO2 for 24h. Cell viability was to T.R. We are also grateful to the Cell-Based Drug and
measured by the MTT assay.34) In the presence of viable cells, Health Product Development Research Unit and Department
cells were incubated with 100µL of a solution of MTT in of Pharmacology and Physiology, Faculty of Pharmaceutical
DMEM medium at a concentration of 0.5mg/mL. The tests Sciences, Chulalongkorn University for technical assistance.
plates were incubated at 37°C under 5% CO2 atmosphere for
3h. Then, all liquid medium were removed. The resulting
Conflict of Interest The authors declare no conflict of
formazan crystals were dissolved in DMSO (100µL). The interest.
absorbance was measured at 540nm using a VICTOR3 multi-
label plate reader (PerkinElmer, Inc.). Experiments with trip-
Supplementary Materials The online version of this
licate data were performed to obtain mean cell viability. The article contains supplementary materials. See supplementary

260
Chem. Pharm. Bull.
Vol. 65, No. 3 (2017)
materials for 1H- and 13C-NMR spectra and HPTLC chro-
matograms of 5a–f.
S., Med. Chem. Res., 25, 1049–1056 (2016).
20) Jain R., Monthakantirat O., Tengamnuay P., De-Eknamkul W., Mol-
ecules, 19, 6809–6821 (2014).
21) Jain R., Monthakantirat O., Tengamnuay P., De-Eknamkul W., BMC
Complement. Altern. Med., 16, 18 (2016).
22) Seiffert K., Seltmann H., Fritsch M., Zouboulis C. C., Horm. Metab.
Res., 39, 141–148 (2007).
23) Izgi A., Turglu M., Akyildrum B., Celen S., Celik N., American
Journal of Dermatology and Venereology, 3, 23–29 (2014).
24) Thiboutot D., Knaggs H. G., Gilliland K., Lin G., Dermatology,
196, 38–42 (1998).
25) Trott O., Olson A. J., J. Comput. Chem., 31, 455–461 (2010).
26) Hagiwara H., Sato K., Nishino D., Hoshi T., Suzuki T., Ando M., J.
Chem. Soc., 2001, 2496–2957 (2001).
27) Jiang Z., Wang A., Hu H., Xu J., Hu Y., Li X., Ye Y., Wang J., Li Q.,
(2012) [The International Patent System].
28) Aswathanarayanappa C., Bheemappa E., Bodke Y. D., Bhovi V. K.,
Ningegowda R., Shivakumar M. C., Peethambar S. K., Telkar S.,
Med. Chem. Res., 22, 78–87 (2013).
29) Peraza-Sánchez S. R., Chávez D., Chai H. B., Shin Y. G., Garcia R.,
Mejía M., Fairchild C. R., Lane K. E., Menendez A. T., Farnsworth
N. R., Cordell G. A., Pezzuto J. M., Kinghorn A. D., J. Nat. Prod.,
63, 492–495 (2000).
30) Ribeiro C. M. R., de Souza P. P., Ferreira L. D. M., Pereira S. L.,
Martins I. d. S., Epifanio R. d. A., Costa-Lotufo L. V., Jimenez
P. C., Pessoa C., de Moraes M. O., Lett. Org. Chem., 8, 347–351
(2011).
References
1) Russell D. W., Wilson J. D., Annu. Rev. Biochem., 63, 25–61 (1994).
2) Azzouni F., Godoy A., Li Y., Mohler J., Adv. Urol., 2012, 1–18
(2012).
3) Jain R., De-Eknamkul W., Expert Opin. Ther. Targets, 18, 787–806
(2014).
4) Titus M. A., Gregory C. W., Ford O. H. 3rd, Schell M. J., Maygar-
den S. J., Mohler J. L., Clin. Cancer Res., 11, 4365–4371 (2005).
5) Jain R., De-Eknamkul W., Expert Opin. Ther. Targets, 18, 787–806
(2014).
6) Godoy A., Kawinski E., Li Y., Oka D., Alexiev B., Azzouni F.,
Titus M. A., Mohler J. L., Prostate, 71, 1033–1046 (2011).
7) Marks L. S., Adv. Urol., 6, S11–S21 (2004).
8) Randall V. A., The endocrine control of hair follicle. Hair growth
and disorder. Springer-Verlag Berlin Heidelberg, Germany, 2008.
9) Kelly Y., Blanco A., Tosti A., Drugs, 76, 1349–1364 (2016).
10) Birch M. P., Messenger A. G., EJD, 11, 309–314 (2001).
11) Steiner J. F., Clin. Pharmacokinet., 30, 16–27 (1996).
12) Mella J. M., Perret M. C., Manzotti M., Catalano H. N., Guyatt G.,
Arch. Dermatol., 146, 1141–1150 (2010).
13) Olsen E. A., Hordinsky M., Whiting D., Stough D., Hobbs S., Ellis
M. L., Wilson T., Rittmaster R. S., Dutasteride Alopecia Research
Team, J. Am. Acad. Dermatol., 55, 1014–1023 (2006).
14) Eun H. C., Kwon O. S., Yeon J. H., Shin H. S., Kim B. Y., Ro B. I.,
Cho H. K., Sim W. Y., Lew B. L., Lee W. S., Park H. Y., Hong S. P.,
Ji J. H., J. Am. Acad. Dermatol., 63, 252–258 (2010).
15) Gubelin Harcha W., Barboza Martínez J., Tsai T.-F., Katsuoka K.,
Kawashima M., Tsuboi R., Barnes A., Ferron-Brady G., Chetty D.,
J. Am. Acad. Dermatol., 70, 489–498.e3 (2014).
16) Aggarwal S., Thareja S., Verma A., Bhardwaj T. R., Kumar M.,
Steroids, 75, 109–153 (2010).
17) Yao Z., Xu Y., Zhang M., Jiang S., Nicklaus M. C., Liao C., Bioorg.
Med. Chem. Lett., 21, 475–478 (2011).
31) Tseng C.-H., Chen Y.-L., Yang S.-H., Peng S., Cheng C.-M., Han
C.-H., Lin S.-R., Tzeng C.-C., Bioorg. Med. Chem., 18, 5172–5182
(2010).
32) Karnsomwan W., Rungrotmongkol T., De-Eknamkul W., Chamni
S., Med. Chem. Res., 25, 1049–1056 (2016).
33) O’Boyle N. M., Banck M., James C. A., Morley C., Vandermeersch
T., Hutchison G. R., J. Chemometr., 3, 33 (2011).
34) Riss T. L. M., R. A., Niles, A. L., Benink, H. A., Worzella, T. J.,
Minor, L. “Assay Guidance Manual [Internet],” ed. by Sittampalam
G. S., Coussens N. P., Nelson H., Eli Lilly & Company and the Na-
tional Center for Advancing Translational Sciences, Bethesda, MD,
2004.
18) Min-Rui O., Jun-Qian L., Chinese Journal of Structure Chemistry,
31, 1618–1626 (2012).
19) Karnsomwan W., Rungrotmongkol T., De-Eknamkul W., Chamni