Functionalized aurones as inducers of NAD(P)H:quinone oxidoreductase 1 that activate AhR/XRE and Nrf2/ARE signaling pathways: Synthesis, evaluation and SAR

Article abstract of DOI:10.1016/j.ejmech.2010.03.023

The chemopreventive potential of functionalized aurones and related compounds as inducers of NAD(P)H:quinone oxidoreductase 1 (NQO1, EC 1.6.99.2) are described. Several 4,6-dimethoxy and 5-hydroxyaurones induced NQO1 activity of Hepa1c1c7 cells by 2-fold at submicromolar concentrations, making these the most potent inducers to be identified from this class. Mechanistically, induction of NQO1 was mediated by the activation of AhR/XRE and Nrf2/ARE pathways, indicating that aurones may be mixed activators of NQO1 induction or agents capable of exploiting the proposed cross-talk between the AhR and Nrf2 gene batteries. QSAR analysis by partial least squares projection to latent structures (PLS) identified size parameters, in particular those associated with non-polar surface areas, as an important determinant of induction activity. These were largely determined by the substitution on rings A and B. A stereoelectronic role for the exocyclic double bond as reflected in the E _LUMO term was also identified. The electrophilicity of the double bond or its effect on the conformation of the target compound are possible key features for induction activity.

Full text of DOI:10.1016/j.ejmech.2010.03.023

European Journal of Medicinal Chemistry 45 (2010) 2957e2971
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Original article
Functionalized aurones as inducers of NAD(P)H:quinone oxidoreductase
1 that activate AhR/XRE and Nrf2/ARE signaling pathways:
Synthesis, evaluation and SAR
*
Chong-Yew Lee, Eng-Hui Chew, Mei-Lin Go
Department of Pharmacy, National University of Singapore, 18 Science Drive 4, Singapore 117543, Singapore
a r t i c l e i n f o
a b s t r a c t
Article history:
The chemopreventive potential of functionalized aurones and related compounds as inducers of NAD(P)
H:quinone oxidoreductase 1 (NQO1, EC 1.6.99.2) are described. Several 4,6-dimethoxy and 5-hydrox-
yaurones induced NQO1 activity of Hepa1c1c7 cells by 2-fold at submicromolar concentrations, making
these the most potent inducers to be identified from this class. Mechanistically, induction of NQO1 was
mediated by the activation of AhR/XRE and Nrf2/ARE pathways, indicating that aurones may be mixed
activators of NQO1 induction or agents capable of exploiting the proposed cross-talk between the AhR
and Nrf2 gene batteries. QSAR analysis by partial least squares projection to latent structures (PLS)
identified size parameters, in particular those associated with non-polar surface areas, as an important
determinant of induction activity. These were largely determined by the substitution on rings A and B.
A stereoelectronic role for the exocyclic double bond as reflected in the ELUMO term was also identified.
The electrophilicity of the double bond or its effect on the conformation of the target compound are
possible key features for induction activity.
Received 18 December 2009
Received in revised form
15 March 2010
Accepted 16 March 2010
Available online 25 March 2010
Keywords:
Aurones
Chemoprevention
NAD(P)H:quinone oxidoreductase 1
Phase II enzyme inducers
Ó 2010 Elsevier Masson SAS. All rights reserved.
1. Introduction
the tumor suppressor p53 in response to DNA-damaging stimuli
[11]. The induction of NQO1 in vivo and in vitro is correlated to the
Cancer chemoprevention is a strategy aimed at reducing the risk
of cancer by avoiding exposure to cancer-causing agents, enhancing
host defense mechanisms, lifestyle adjustments and supplemen-
tation with chemopreventive agents that inhibit, reverse or block
tumorigenesis [1e3]. Many chemopreventive agents are dietary
phytochemicals such as sulforaphane from broccoli, glucosinolate-
derived indoles found in cruciferous vegetables and structurally
diverse flavonoids that are present in fruits, vegetables and plant-
derived beverages [4e6]. These bioactive compounds selectively
modulate the expression levels or biological activities of key
proteins involved in various cell signaling cascades, particularly
enzymes involved in the detoxification of xenobiotics [7e10]. The
enzyme NAD(P)H:quinone oxidoreductase 1 (NQO1, EC 1.6.99.2)
has received particular attention in this regard because the induc-
tion of NQO1 provides the cell with multiple layers of protection
against environmental insults [8]. Briefly, NQO1 detoxifies highly
reactive quinones to quinols without generating reactive semi-
quinones and maintains endogenous lipid soluble antioxidants in
their reduced and active forms. It has also been reported to stabilize
induction of other protective phase II enzymes and this has led to
its reputation as a reasonable biomarker for evaluating the che-
mopreventive potential of target compounds [12].
Aurones (2-benzylidenebenzofuran-3(2H)-ones) are a class of
flavonoids found in fruits and flowers where they function as
phytoalexins against infections and contribute to the yellow
pigmentation of plant parts [13]. Our interest in aurones centers on
their potential to enhance NQO1 activity and hence function as
chemopreventive agents. Jang and co-workers have reported two
aurones (sulfuretin and 6,40-dihydroxy-30-methoxyaurone, Fig. 1)
isolated from the seeds of the Tonka Bean (Dipteryx odorata) that
induced NQO1 activity in murine hepatoma cells [14]. Structurally,
aurones possess features that are associated with NQO1 induction
activity, the most prominent of which is the presence of a Michael
acceptor moiety. In addition, the olefinic double bond of this moiety
is exocyclic in aurones. Talalay et al. noted greater NQO1 induction
potencies among Michael acceptors with exocyclic double bonds
(for example, 2-methylene-4-butyrolactone and 3-methylene-2-
norbornanone) as compared to others such as furan-2(5H)-one
where the double bond is endocyclic [15]. They proposed that the
exocyclic location of the double bond enhanced the electrophilic
reactivity of the Michael acceptor moiety, resulting in greater NQO1
induction activity. If this motif is indeed associated with greater
* Corresponding author. Tel.: þ65 65162654; fax: þ65 67791554.
E-mail address: phagoml@nus.edu.sg (M.-L. Go).
0223-5234/$ e see front matter Ó 2010 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.ejmech.2010.03.023

2958
C.-Y. Lee et al. / European Journal of Medicinal Chemistry 45 (2010) 2957e2971
greater cell growth inhibitory properties than their corresponding
OH
OR
4,5,6-trimethoxy regioisomers [25]. Investigations on the inhibitory
effects of aurones on the tyrosinase enzyme in human melanocytes
showed that the most active compounds had hydroxyl groups on ring
A as compared with less active compounds that had rings A that were
either unsubstituted or had only one hydroxyl group [26]. Based on
these observations, we synthesized five series of aurones with
different ring A substituents. These are 4,6-dimethoxyaurones (Series
1), 4,6-dihydroxyaurones (Series 2), monohydroxylated aurones
(Series 3 and 4) and aurones without substituents on ring A (Series 5).
The choice of methoxy and hydroxy groups was based on the preva-
lence of these groups in naturally occurring aurones [27]. Isomeric
isoaurones (Series 6, 7) inwhich the carbonyl function is embedded as
a lactone, unlike the ketone carbonyl in aurones, were also synthe-
sized. Series 8 were dehydroaurones in which the exocyclic double
bond is reduced. The structures of compounds in Series 1e8 are given
in Table 1. In addition, the substitution pattern on ring B was varied to
include groups with different lipophilicity and electron donating/
withdrawing properties [28].
HO
O
O
Fig. 1. Structures of sulfuretin (R ¼ H) and 6,40-dihydroxy-30-methoxyaurone
(R ¼ CH3).
reactivity, then aurones which are unique among flavonoids in
having an exocyclic double bond, should be outstanding inducers.
On the other hand, only modest induction potencies were found for
sulfuretin and 6,40-dihydroxy-30-methoxyaurone [14]. It may be
argued that the NQO1 induction potential of aurones has not been
fully exploited and that structural modifications of the scaffold
would result in more potent members.
The mechanism of NQO1 induction by aurones is also of interest.
In a previous study, flavonoids were classified as bifunctional
inducers because they induced both Phase I and Phase II metabolic
enzymes by a mechanism that involves activation of the aryl
hydrocarbon receptor (AhR)-xenobiotic-response element (XRE)
pathway [15,16]. Briefly, the flavonoid binds to cytosolic AhR
leading to its translocation to the nucleus where it causes tran-
scriptional activation of Phase I enzymes like cytochrome P450-1A1
(CYP1A1) and Phase II enzymes (NQO1, glutathione S-transferase
among others) [10,17,18]. Phase II enzymes play a key role in
detoxification whereas Phase I enzymes (CYP1A1) activate poly-
cyclic aromatic hydrocarbons to ultimate carcinogens [19]. Thus
inducers that affect both Phase I and II enzymes are generally
considered to be less promising chemopreventive agents compared
to monofunctional inducers that selectively activate Phase II
enzymes [20]. Mechanistically, monofunctional inducers (which
are usually electrophiles) interact with cysteine residues in Kelch-
like erythroid cell derived protein 1 (Keap1), a binding partner of
nuclear factor-erythroid related factor 2 (Nrf2) that is involved in
regulating its transcriptional activity by sequestering the cytosolic
Nrf2 and targeting it for proteosomal degradation. Consequently,
increased accumulation of nuclear Nrf2 accelerates transcription of
Phase II enzymes through activation of the antioxidant response
element (ARE) of the relevant genes [21]. However, recent findings
showed that several flavonoids behave as mixed AhR/Nrf2 activa-
tors and induced Phase II enzymes by activating both ARE- and
XRE-signaling pathways [10,22]. The few mixed activators identi-
fied to date are structurally diverse, such as flavonoids (quercetin,
luteolin, chrysin), 1,2-dithiol-3-thiones and oltipraz [10]. It has
been argued that by acting on dual signaling pathways, mixed
activators bring about a greater chemopreventive effect than
mono- or bifunctional inducers [23,24].
Table 1
Structures of compounds in Series 1e8.
Compound R1
%
Yield
Compound R1
%
Yield
Series 1
4'
3'
B
2'
R1
H3CO
O
C
A
O
OCH3
1-1
1-2
1-3
1-4
1-5
1-6
1-7
1-8
H
2
40
90
60
46
98
54
43
15
1-15
40-F
80
67
46
38
37
65
21
20
0-Cl
1-16
1-17
1-18
1-19
1-20
1-21
1-22
40-CN
40-CF3
30-Cl
40-Cl
20-OH
30-OH
40-OH
20, 30-
(OH)2
20, 40-
(OH)2
40-NO2
20-CH3
40-CH3
40-N(CH3)2
40-N-methylpiperazinyl
1-9
20
1-23
20-pyridinyla
20
1-10
1-11
1-12
1-13
20-OCH3 94
30- OCH3 20
40- OCH3 74
1-24
1-25
1-26
1-27b
30-pyridinyla
27
30
51
56
40-pyridinyla
30-quinolinyla
4,6-Dimethoxybenzofuran-3
(2H)-one
20-F
58
1-14
30-F
37
Series 2
In this study, a series of functionalized aurones were investi-
gated for their ability to induce NQO1 in Hepa1c1c7 cells. We
showed here that the more potent aurones induced NQO1 by
activating both the AhR/XRE and Nrf2/ARE pathways and discussed
the structureeactivity relationships associated with their NQO1
induction activities.
2-1
2-2
2-3
2-4
2-5
2-6
2-7
2-8
H
15
44
23
16
31
43
62
2-9
30-OCH3
68
31
20
65
28
15
32
88
20-Cl
30-Cl
40-Cl
20-OH
30-OH
40-OH
2-10
2-11
2-12
2-13
2-14
2-15
2-16c
40-OCH3
40-CN
2. Chemical considerations
40-CH3
20-F
2.1. Design of target compounds
30-OCH3, 40-OH
30, 40, 50-(OCH3)3
4,6-Dihydroxybenzofuran-3
(2H)-one
Anecdotal evidence suggests that changes in the substitution on
ring Aof theauronescaffoldhaveamarkedeffectonbiologicalactivity.
Lawrence and co-workers noted that 5,6,7-trimethoxyaurones had
20-OCH3 35

C.-Y. Lee et al. / European Journal of Medicinal Chemistry 45 (2010) 2957e2971
2959
Table 1 (continued)
7-3
7-4
7-5
30-OH
20
18
32
7-8
7-9
7-10
7-11
40-Cl
20-F
30-F
40-F
22
20
31
55
40-OH
Compound R1
%
Yield
Compound R1
%
Yield
20-OCH3
Series 3
Compound
R1
%Yield
Series 8
3-1
3-2
3-3
3-4
3-5
3-6
H
25
55
22
58
47
60
3-7
3-8
3-9
3-10
3-11
3-12
40-OH
60
50
22
28
50
65
20-Cl
30-Cl
40-Cl
20-OH
30-OH
20-OCH3
30- OCH3
40- OCH3
40-CN
40-CH3
8-1
8-2
H
20-OH
50
35
Series 4
a
Ring B is replaced by heterocyclic ring.
b
c
No R1 substituent. Benzylidene side chain is omitted.
No R1 substituent. Benzylidene side chain is omitted.
2.2. Synthetic approaches
87 compounds were synthesized in this investigation, of which
56 have not been previously reported in the SciFinder ScholarÔ
(Chemical Abstract Service). The synthesis of the 4,6-dimethox-
yaurones (Series 1) is shown in Scheme 1. Briefly, 3,5-dimethox-
yphenol was reacted with chloroacetic acid in the presence of
sodium hydride to give 3,5-dimethoxyphenoxyacetic acid (1-28).
The acidic side chain was then cyclized by a Friedel-Craft acylation
reaction to give the benzofuran-3(2H)-one (1-27) which reacted
with various aldehydes in a base-catalyzed aldol reaction to give
the target compounds [25].
4-1
4-2
4-3
4-4
4-5
4-6
H
38
33
65
67
43
30
4-7
4-8
4-9
4-10
4-11
4-12
40-OH
53
62
50
24
76
60
20-Cl
30-Cl
40-Cl
20-OH
30-OH
20-OCH3
30- OCH3
40- OCH3
40-CN
40-CH3
Compound
R1
%Yield
Series 5
In the case of the 4,6-dihydroxyaurones (Series 2), phlor-
oglucinol was reacted with chloroacetonitrile via a Hoesch acyla-
tion to give an iminium salt which was hydrolyzed under acidic
conditions to give the acetophenone (2-17) [26,29]. Base-catalyzed
cyclization of 2-17 gave 4,6-dihydroxybenzofuran-3(2H)-one (2-16)
which was condensed with various benzaldehydes by microwave
heating (110 ꢀC, 10e15 min) in sealed tubes (Scheme 2). This
procedure significantly shortened reaction times and made it
unnecessary to protect the phenolic hydroxyl groups prior to aldol
condensation. However, 2-13 and 2-14 could not be obtained by
this method and their synthesis was achieved by conventional
heating and protection of the phenolic hydroxyl groups.
5-1
5-2
20-OH
20-Cl
65
31
Series 6
A similar route was followed for the monohydroxylated aurones
of Series 3 and 4 (Scheme 3). Starting from either 20, 50-dihydrox-
yacetophenone or 20, 40-dihydroxyacetophenone,
a-bromination
with cupric bromide gave the brominated acetophenones 3-13 and
4-13, followed by base-catalyzed cyclization to give the hydrox-
ybenzofuran-3(2H)-ones (3-14, 4-14) [30]. Aldol condensation with
substituted benzaldehydes gave the desired aurones in yields
ranging from 22 to 76%.
6-1
6-2
6-3
6-4
6-5
H
47
12
25
48
31
20-OH
20-OCH3
20-Cl
40-Cl
The 4,6-dimethoxyisoaurones of Series 6 were synthesized by
the Kindler-Willgerodt condensation of 3,5-dimethoxy-2-hydrox-
yacetophenone with morpholine and sulphur to form a thio-
morpholide adduct (Scheme 4) [31]. Hydrolysis yielded the
corresponding phenylacetic acid (6-6) and lactonization of the acid
side chain in the presence of phosphorus oxychloride gave the
benzofuran-2(3H)-one (6-7). The latter was reacted with various
substituted benzaldehydes in acetic anhydride to give the desired
isoaurones.
Compound
R1
%Yield
Compound
R1
%Yield
Series 7
For Series 5 and 7, straightforward condensation of the commer-
cially available benzofuran-3(2H)-one with substituted benzalde-
hydes under aldol conditions gave the desired products. Series 8
7-1
7-2
H
20-OH
25
65
7-6
7-7
40-OCH3
20-Cl
32
15

2960
C.-Y. Lee et al. / European Journal of Medicinal Chemistry 45 (2010) 2957e2971
R'
H3CO
O
COOH
H3CO
H3CO
O
H3CO
OH
O
a
c
b
O
OCH3
1-28
O
OCH3
OCH3
Aurones 1-1 to 1-26
OCH3
1-27
Scheme 1. Reagents and conditions: (a) Chloroacetic acid, NaH, DMF, rt, 12 h. (b) Polyphosphoric acid, 90 ꢀC, 8 h. (c) Substituted benzaldehyde, KOH 50% in water, MeOH, rt, 1e3 h.
comprised of two dehydroaurones which were obtained by hydro-
genation of the 4,6-dimethoxyaurones 1-1 and 1-5 in a Parr hydro-
genator with palladium on charcoal (10%) as catalyst.
Phenolic hydroxyl groups on reacting benzaldehydes were
protected as tetrahydropyranyl ethers which were subsequently
removed by acid hydrolysis of the final condensation product. The
protectionedeprotection step was not necessary for benzaldehydes
with ortho-hydroxyl groups.
(1-10, 1-13, 1-15) were screened at 1
affected cell viability at 5 M. In all, CD values of 31 compounds and
two established NQO1 inducers, sulforaphane (a monofunctional
inducer) [36] and -naphthoflavone (BNF, a bifunctional inducer)
mM because they adversely
m
b
[7] were determined (Table 2).
The induction activities of the test compounds revealed some
interesting structureeactivity trends. An important observation
was the inactivity of compounds that did not have an intact olefin.
These were the dehydroaurones (8-1, 8-2) in which the double
bond was reduced and compounds (1-27, 2-16) where the benzy-
lidene side chain was removed. These findings underscored an
important role for the exocyclic double bond and the Michael
acceptor moiety of which it is a part for induction activity. Another
notable observation is the strong representation of 4,6-dimethox-
yaurones (Series 1) among the active compounds in Table 2. Series 1
had the largest number of compounds (8) with submicromolar CD
values as well as the most potent inducers (1-5, 1-13, 1-14) iden-
tified in this investigation. Replacement of the methoxy groups
with hydroxyl resulted in less potent inducers, as seen from Series 2
to 4. The poor activities of 5-1 and 5-2 clearly signaled the need for
maintaining substitution on ring A.
2.3. Stereochemical considerations
The exocyclic double bond in aurones and isoaurones may be E
(trans) or Z (cis). These configurational isomers have characteristic
1H and 13C chemical shifts in the NMR spectra. The chemical shift of
the olefinic (b) proton is deshielded in the E isomer of aurones and
appears downfield (ca 7.01 ppm) compared to the Z isomer (ca
6.70 ppm) [29,32]. Both 1H and 13C NMR data of Series 1e5 sup-
ported the assignment of Z-stereochemistry and this was further
confirmed by the x-ray structures of 1-10 and 3-10 which showed
the presence of the exocyclic double bond in the thermodynami-
cally more stable Z configuration [33,34].
Lactones were reported to be poor NQO1 inducers, possibly due
to hydrolytic instability or the weaker electron withdrawing effect
of the carbonyl group when it is part of a lactone [15]. As such, the
isoaurones were anticipated to be poor inducers and this was duly
observed. The isoaurones of Series 6 and 7 were only modest
Unlike the aurones, analysis of the NMR spectra of the iso-
aurones pointed to the presence of a mixture of E and Z isomers,
with the E isomer present in larger amounts (80e90%). The
chemical shifts of the olefinic C
to ring B were used to distinguish between the Z and E isomers.
The deshielding by the carbonyl group caused the olefinic C H of
bH and the ortho-H atoms attached
inducers of NQO1 with CD values of 3e25 mM. We did not observe
b
hydrolytic instability of the isoaurones during their syntheses or
biological evaluation. Thus the diminished electrophilicity of the
exocyclic bond in the isoaurones may be a contributing factor. The
4,6-dimethoxyisoaurones (Series 6) were significantly poorer
inducers than their aurone counterparts (Series 1), but the 5-
hydroxyisoaurones (Series 7) were comparable to the 5-hydrox-
yaurones (Series 3). Both Series 3 and 7 were notable in having
E-isoaurones to shift downfield by 0.04e0.34 ppm compared with
their Z counterparts [35]. In addition, a downfield shift (by
0.49e0.77 ppm) was observed for H atoms attached to ortho
positions (if any) of ring B of Z-isoaurones. The predominance of
the E isomer was established by comparing the integrated areas of
the singlet peak assigned to CbH of each isomer.
several potent inducers (CD values 0.47e3.9 mM) which may reflect
3. Results
a special preference for 5-OH on ring A. A few substituents on ring B
were consistently associated with good activity across the different
Series. These were halogens (in Series 1, 3, 4, 6), hydroxyl (Series 1,
7) and methoxy (Series 1, 6). There was also a preference for the
groups to be located at 20 or 30 on ring B.
3.1. Induction of NQO1 in Hepa1c1c7 cells
The compounds were screened for induction of NQO1 activity in
the murine hepatoma (Hepa1c1c7) cells. Screening was carried out
at 5
2-fold or more at either concentration were short-listed for the
determination of CD (concentration at which compound
increased NQO1 activity by 2-fold). Results of the initial screen are
given in Supplementary Information (Table 1). Three compounds
mM or 25 mM and compounds that increased NQO1 activity by
3.2. Induction of NQO1 in mutant murine hepatoma BPrc1 cells
a
The mutant Hepa1c1c7 cells (BPrc1) contain mutations in the
Ah-receptor mediated pathway and do not respond to compounds
R1
HO
OH
HO
OH
HO
HO
OH
HO
O
O
Cl
Cl
c
a
b
d
+
OH NH2 Cl-
O
O
OH
O
OH
OH
OH
2-17
2-16
Aurones 2-1 to 2-15
Scheme 2. Reagents and conditions: (a) Chloroacetonitrile, HCl, ZnCl2, Et2O, 0 ꢀC / rt, 24 h. (b) 1 N HCl, 100 ꢀC, 1 h. (c) NaOAc, MeOH, reflux, 2 h. (d) Substituted benzaldehyde, KOH
50% in water, MeOH, microwave heating, 110 ꢀC, 10e15 min.

C.-Y. Lee et al. / European Journal of Medicinal Chemistry 45 (2010) 2957e2971
2961
R1
OH
O
O
OH
O
O
R
R
R
R
Br
a
c
b
O
O
R = 5-OH 3-13
R = 4-OH 4-13
R = 5-OH
R = 4-OH
R = 5-OH 3-1 to 3-12
R = 4-OH 4-1 to 4-12
R = 5-OH 3-14
R = 4-OH 4-14
Scheme 3. Reagents and conditions: (a) CuBr2, CHCl3-ethyl acetate, reflux, 8 h. (b) KOH 50%/H2O, methanol, reflux, 2 h. (c) Substituted benzaldehyde, KOH 50%/H2O, methanol,
60 ꢀC, 1 h.
that induce NQO1 activity by this route. When investigated on the
BPrc1 cells, none of the compounds in Table 2 induced NQO1
activity which shows that the AhR/XRE signaling pathway was
involved in the induction activity of these compounds. Some
representative figures are given in Fig. 2. Sulforaphane induced
increase luciferase activity. As seen from Fig. 3, significant increases
in luciferase activity were observed for 1-1, 1-5, 2-5 and sulfor-
aphane. In contrast, aurones 5-1 and 8-2 which were inactive as
NQO1 inducers did not increase luciferase activity. Aurone 2-5
significantly increased luciferase activity only at a higher concen-
NQO1 activity in the Bprc1 cells (CD ¼ 0.34
m
M ꢁ 0.01) to almost the
tration of 25
mM unlike 1-1 and 1-5 which caused significant
same degree as the Hepa1c1c7 cells (CD ¼ 0.21
m
M ꢁ 0.02). On the
increases at both 10 and 25 m
M. However, none of the aurones
other hand, BNF, like the aurones, did not induce NQO1 activity in
increased luciferase activity to the same degree as sulforaphane,
even though some of them (1-1, 1-5) were as active as sulforaphane
Bprc1 cells.
in inducing NQO1 activity (CD values of 0.2e0.3 mM). The discrep-
3.3. Effect on EROD activity in Hepa1c1c7 cells
ancy between the CD values of the aurones and their effects on the
luciferase activities of the transfected cells suggests that while
Nrf2/ARE signaling may be essential for NQO1 induction, other
pathways may play a more dominant role.
To further probe the involvement of the Nrf2/ARE pathway in
the NQO1 induction properties of these aurones, we proceeded to
investigate their effects on Nrf2 protein levels. Hepa1c1c7 cells
To provide further insight on the induction activity of the test
compounds, we evaluated their effects on CYP1A1 whose tran-
scription is regulated by the Ah-receptor. CYP1A1 activity in Hep-
a1c1c7 cells was determined by the ethoxyresorufin-O-deethylase
(EROD) assay which monitors the formation of fluorescent resor-
ufin when 7-ethoxyresorufin is de-ethylated by CYP1A1. Inducers
of CYP1A1 increase resorufin levels and the ratio of its fluorescence
in treated versus untreated control cells. The CYP1A1 induction
were treated with the test aurone or sulforaphane at 10 mM for 6 h,
after which cell lysates were prepared and probed for Nrf2
immunoreactivity (Fig. 4). Densitometric quantitation of the visu-
alized bands of Nrf2 protein revealed by Western analysis showed
that only lysates from cells treated with MG132 (a proteosome
inhibitor which increases Nrf2 levels), sulforaphane and aurones 1-
1, 1-5, and 2-5 had significantly higher levels of Nrf2 proteins than
control.
ratios were determined at a fixed concentration (1, 5, 10 or 25 mM)
of the test compound (Table 2). Most of the compounds were tested
at concentrations that exceeded their CD values for NQO1 induc-
tion. As expected, sulforaphane did not induce CYP1A1 activity
while BNF, a known bifunctional inducer, showed strong induction
activity (ratio of 4.25 at 1
m
M). Interestingly, none of the aurones
The ARE regulatory sequences controlled by Nrf2 are respon-
sible for the transcriptional activation of several genes besides
NQO1. Here, we probed the effect of the aurones on the functional
activity of two proteins whose genes are regulated by the tran-
scriptional activation of ARE consensus sequences by Nrf2. They
had values of the same magnitude as BNF. Nearly two-third had
values that were less than 2, indicating modest induction of
CYP1A1 activity even though they were tested at higher concen-
trations in this assay. These findings raised questions as to whether
the aurones only affected the AhR/XRE signaling pathway to bring
about NQO1 induction.
are
g-glutamylcysteine ligase (GCL) [38] and thioredoxin reduc-
tase (TrxR, E.C. 1.8.1.9) [39] GCL is the rate limiting enzyme
involved in the de novo synthesis of GSH. Activation of the Nrf2/
ARE pathway would bring about up-regulation of GCL and an
increase in cellular GSH levels. GSH levels in Hepa1c1c7 cells
were determined with Ellman’s reagent after incubation (24 h)
3.4. Effect on the Nrf2/ARE signaling pathway
To measure the potential of selected aurones to induce ARE-
mediated gene expression, Hepa1c1c7 cells were transiently
transfected with a firefly luciferase reporter gene whose expression
is regulated by an ARE from the human NQO1 gene (pGL3-ARE)
[37]. Compounds that increase NQO1 gene transcription would
with selected aurones (10 mM of 1-1, 1-5, 2-5, 5-1, and 8-2). As
shown in Fig. 5, 1-1, 1-5 and 2-5 significantly increased GSH
content compared to untreated cells but this was not observed for
5-1 and 8-2.
H3CO
d
H3CO
c
O
H3CO
OH
COOH
O
H3CO
OH
H3CO
OH
O
O
O
N
b
OCH3
a
OCH3
6-7
OCH3
6-6
OCH3
S
OCH3O
R1
6-1 to 6-5
Scheme 4. Reagents and conditions: (a) Sulphur, morpholine, 150 ꢀC, 24 h. (b) 2 M NaOH, reflux, 6 h. (c) POCl3, dichloroethane, rt, 16 h. (d) Substituted benzaldehyde, Et3N, Ac2O,
60 ꢀC, 6e12 h.

2962
C.-Y. Lee et al. / European Journal of Medicinal Chemistry 45 (2010) 2957e2971
Table 2
CD values, fold induction of CYP1A1 and ELUMO of test compounds.
Series
Compound
R0 on ring B
CD (
m
M)a
Fold induction of CYP1A1b
M)
ELUMO (eV)c
1-1
1-2
1-3
1-5
1-6
1-8
1-10
1-13
1-14
1-15
1-23
1-25
H
0.31 ꢁ 0.02
0.40 ꢁ 0.07
0.61 ꢁ 0.05
0.18 ꢁ 0.05
9.8 ꢁ 1.3
2.04 ꢁ 0.15 (10
m
ꢂ1.888
ꢂ2.047
ꢂ2.083
ꢂ1.779
ꢂ1.876
ꢂ1.746
ꢂ1.754
ꢂ1.970
ꢂ2.025
ꢂ1.934
ꢂ2.037
ꢂ2.228
20Cl
1.55 ꢁ 0.09 (5
m
m
M)
M)
30Cl
1.18 ꢁ 0.26 (5
Series 1,
20OH
1.87 ꢁ 0.22 (10
1.16 ꢁ 0.05 (25
0.97 ꢁ 0.04 (10
m
m
m
M)
M)
M)
30OH
20, 30- (OH)2
20OMe
20F
3.37 ꢁ 0.15
0.31 ꢁ 0.02
0.16 ꢁ 0.10
0.15 ꢁ 0.03
0.68 ꢁ 0.20
3.8 ꢁ 0.5
1.56 ꢁ 0.13 (1
m
M)
NDd
30F
40F
1.48 ꢁ 0.02 (5
NDd
m
M)
pyridine-2-yle
pyridine-4-yle
0.98 ꢁ 0.04 (10
m
m
M)
M)
8.5 ꢁ 0.4
1.73 ꢁ 0.23 (25
Series 2,
2-2
2-5
2-7
2-9
2-13
2-14
20Cl
1.7 ꢁ 0.1
3.0 ꢁ 0.26
24.3 ꢁ 1.7
4.4 ꢁ 0.2
5.5 ꢁ 1.1
16.5 ꢁ 1.2
1.18 ꢁ 0.11 (10
1.10 ꢁ 0.17 (10
1.05 ꢁ 0.06 (25
1.13 ꢁ 0.07 (10
1.00 ꢁ 0.04 (10
1.06 ꢁ 0.02 (25
m
m
m
m
m
m
M)
M)
M)
M)
M)
M)
ꢂ2.135
ꢂ1.860
ꢂ1.843
ꢂ1.920
ꢂ2.057
ꢂ1.816
20OH
40OH
30OMe
20F
30OMe, 40OH
Series 3,
3-1
3-3
3-5
3-8
H
3.9 ꢁ 0.4
0.96 ꢁ 0.17
2.4 ꢁ 0.4
2.36 ꢁ 0.09 (10
2.95 ꢁ 0.24 (10
2.01 ꢁ 0.22 (10
1.68 ꢁ 0.20 (25
m
m
m
m
M)
M)
M)
M)
ꢂ2.299
ꢂ2.493
ꢂ2.174
ꢂ2.146
30Cl
20OH
20OMe
0.47 ꢁ 0.10
Series 4,
4-3
4-4
4-5
4-6
30Cl
2.8 ꢁ 0.1
7.8 ꢁ 1.2
9.4 ꢁ 1.1
11.2 ꢁ 1.5
1.05 ꢁ 0.03 (10
1.24 ꢁ 0.09 (25
1.07 ꢁ 0.03 (25
1.08 ꢁ 0.03 (25
m
m
m
m
M)
M)
M)
M)
ꢂ2.338
ꢂ2.309
ꢂ2.023
ꢂ2.129
40Cl
20OH
30OH
Series 5,
5-1
20OH
z50f
1.40 ꢁ 0.60 (5
m
M)
ꢂ2.176
Series 6,
6-1
6-3
6-4
H
25.2 ꢁ 1.8
11.6 ꢁ 1.2
6.4 ꢁ 1.2
1.04 ꢁ 0.08 (25
1.03 ꢁ 0.05 (25
0.97 ꢁ 0.09 (10
m
m
m
M)
M)
M)
ꢂ1.874
ꢂ1.715
ꢂ1.997
20OMe
20Cl
Series 7,
7-2
7-8
20OH
40Cl
2.8 ꢁ 0.5
3.4 ꢁ 0.2
1.12 ꢁ 0.08 (10
1.04 ꢁ 0.10 (10
m
m
M)
M)
ꢂ2.118
ꢂ2.436
Series 8,
8-2
20OH
z50f
NDd
ꢂ1.007

C.-Y. Lee et al. / European Journal of Medicinal Chemistry 45 (2010) 2957e2971
2963
Table 2 (continued)
Series
Compound
R0 on ring B
CD (m
M)a
Fold induction of CYP1A1b
ELUMO (eV)c
Sulphoraphaneg
0.21 ꢁ 0.02
0.92 ꢁ 0.04 (1
4.25 ꢁ 0.15 (1
m
m
M)
M)
NDd
BNFg
0.012 ꢁ 0.05
NDd
a
Only compounds that increased NQO1 activity in Hepa1c1c7 cells by at least 2 fold at 5 mM or 25 mM were evaluated for CD (concentration required to increase NQO1
activity by 2-fold). Mean ꢁ SD for n ¼ 3
b
Fold induction of CYP1A1 activity in Hepa1c1c7 cells as determined by the EROD assay. Mean ꢁ SD for n ¼ 3. Concentration at which compound was tested is given in
parenthesis.
c
d
e
f
ELUMO values were determined using SPARTAN 2006 on energy minimized structures. Details are given in Section 5.13.
Not determined.
Ring B is replaced by a pyridine ring.
5-1 and 8-2 did not induce NQO1 activity by 2-fold at the highest concentration tested (25
m
M). CD was estimated to be z50
mM.
g
CD values were comparable to those reported by Dinkova-Kostova et al. [36] using the same method.
Next, aurones 1-1, 1-5, 2-5, 5-1, 8-2 and sulforaphane were
investigated for their effects on TrxR, an antioxidant enzyme whose
transcription is activated by the Nrf2 pathway [39]. We found that
(PLS) and the genetic algorithm (GA) approach to provide a quan-
titative analysis of structureeactivity relationships [41,42]. PLS
involves the identification of principal components that can
concurrently explain the variance in the descriptor set as well as
maximize the correlation to the dependent variable (ꢂlog CD or
pCD). GA mimics molecularevolution byengagingcycles of variation
and selection as driving forces for optimization. Seventeen
descriptors related to size, electronic properties and lipophilicity
were collected for all compounds but analysis was limited to those
compoundswith CD values (Table 2) aswellas 5-1 and 8-2. The latter
were included to provide a balance between strong inducers and
weaker members (5-1, 8-2), as well as to ensure that all series were
represented in the analysis.
The thirty-three compounds characterized by 17 descriptors
were analyzed by PLS using pCD as dependent variable. A satis-
factory one-component model (p < 0.05) that could account for
36% (R2Y) and predict 26% (Q2) of the variation in activity was
obtained. The model was improved by omitting the 4,6-dime-
thoxyisoaurones of Series 6 (n ¼ 3) because their large residual
values (pCDobserved ꢂ pCDpredicted ꢃ 0.9) suggested they were
outliers. The resulting one-component model showed improved
1-1, 1-5 and 2-5 significantly induced TrxR activity at 10 mM after
24 h incubation while the inactive NQO1 inducers 5-1 and 8-2
failed to do so under similar conditions (Fig. 6). Sulforaphane
caused a modest but significant induction of TrxR activity (1.3 fold).
A similar level of TrxR induction was reported for sulforaphane at
12
mM in HepG2 cells [40].
To determine if the induction of TrxR activity was accompanied by
an upregulation in TrxR protein expression, lysates from the TrxR
activity assays were analyzed by Western blotting for TrxR immu-
noreactivityand the blotswere quantified by densitometry. As shown
in Fig. 7, TrxR levels from cells treated with 1-1, 1-5 and sulforaphane
were significantly increased compared to untreated cells.
3.5. Quantitative structureeactivity relationship (QSAR) of NQO1
induction activities of aurones and isoaurones
The induction activities of the aurones and related compounds
were analyzed by partial least squares projection to latent structures
Sulforaphane
Beta-naphthoflavone
5
10
Hepa
Hepa
Bprc1
4
3
8
6
4
2
1
2
Bprc1
0
1E-3 0.01
0.1
1
0.01
0.1
1
Concentration (µM)
Concentration (µM)
1-5
1-14
Hepa
5
4
3
2
1
12
10
8
Hepa
Bprc1
6
4
Bprc1
10
2
0
0.01
0.1
1
0.01
0.1
1
Concentration (µM)
Concentration (µM)
Fig. 2. NQO1 induction in Hepa1c1c7 and BpRc1 cell lines by sulforaphane, beta-naphthoflavone, 1-5 and 1-14.

2964
C.-Y. Lee et al. / European Journal of Medicinal Chemistry 45 (2010) 2957e2971
60
6
5
4
3
2
1
0
*
10 μM
25 μM
*
*
50
*
40
30
20
* *
*
*
*
*
Control
1-1
1-5
2-5
5-1
8-2
SUL
10
0
Compound
Fig. 3. Induction of luciferase activity of Hepa1c1c7 cells transiently transfected with
pGL3-ARE firefly luciferase reporter plasmid and incubated with test compound
control
1-1
1-5
2-5
5-1
8-2
SUL
(10 mM or 25 mM) for 24 h. The fold increase in luciferase activity is represented as the
mean and SD of 3 independent replicates of cell transfection and exposure. (*) indi-
Compound
cates a response significantly different from the control (p < 0.05, independent T-test).
Fig. 5. GSH levels (expressed as nmol per total mg protein) in Hepa1c1c7 cells incu-
bated with aurones and sulforaphane (SUL) at 10
m
M, 24 h. Values are mean ꢁ SEM
(n ¼ 3). * Indicates significant elevation of GSH compared to control cells (p< 0.05,
statistics (R2Y ¼ 0.586, Q2 ¼ 0.467). The contribution to activity by
the various descriptors is depicted in the coefficient plot (Fig. 8).
This figure shows that activity was correlated to size (area, volume),
non-polar surfaces (HAS-HYD, ASA) and lipophilicity (Log P) and
inversely related to polar surface areas (TPSA, VSA-POL) and to
a lesser extent, the number of H bond donors (HD) and acceptors
(HA). Electronic descriptors like HOMO, LUMO and dipole moment
had negligible effects on activity. Thus, the PLS model identified
potent inducers to be those with smaller total polar surface areas,
larger VDW areas/volumes and larger hydrophobic VDW surface
areas.
To assess the validity of the model, the 30 compounds were
randomly divided into a training set (n ¼ 16) and a test set (n ¼ 14).
A model was generated with the training set using the same
descriptors as before and it was used to predict the activity of the
test set. A reasonable error of prediction (RMSEP) of 0.382 was
obtained for the test set.
1-way ANOVA with Dunnett as posthoc test).
Fixed length:
pCD ¼ 4:74 ꢂ 2:10ELUMO ꢂ 2:52 Log P
ꢂ 0:66½VDW surface area of H bond acceptorꢄ
þ 0:08½Hydrophobic VDW surface areaꢄ
(1)
RMSE ¼ 0.487, R2 ¼ 0.522, R2 (adjusted) ¼ 0.454, AIC (Akaike
information criterion) ¼ 59.19, LOF (Lack of Fit) ¼ 0.414, F ¼ 7.64
Variable length:
pCD ¼ ꢂ4:75 ꢂ 1:55ELUMO
þ 0:014½Water accessible surface areaꢄ
(2)
RMSE ¼ 0.552, R2 ¼ 0.387, R2 (adjusted) ¼ 0.346, AIC ¼ 62.40,
LOF ¼ 0.394, F ¼ 9.45.
Next, the induction activities of the 33 compounds were
analyzed by GA using the same descriptors. Models were generated
by the fixed and variable length approaches. The best equation from
each approach is given in Equations (1) and (2). Variable length GA
was preferred because it gave a simpler model with fewer
descriptors (Equation (2)).
Validation was carried out using the same training and test sets
described for PLS, except that the Series 6 compounds were not
omitted and one extra member (6-1) was included in the training
set (n ¼ 17). Equations were generated with variable length GA and
leave-one-out validation. As GA uses an “evolutionary” approach to
generate models, different equations were obtained for each run.
These equations differed in the type of descriptors identified but
consistently gave a high weight to the ELUMO term. Equation (3) is
an example:
A
6 hours
MG132 Control 1-1 1-5
2-5
5-1
8-2
SUL
Nrf2
β-actin
2.5
*
2
*
*
B
*
1.5
*
0.6
*
*
0.5
0.4
0.3
0.2
0.1
0
*
*
1
0.5
0
MG132 control 1-1
1-5
2-5
5-1
8-2
SUL
SUL
control
1-1
1-5
2-5
5-1
8-2
Fig. 4. Immunoblot analysis of Nrf2 protein. (A) Western blot analysis of Hepa1c1c7
cell lysates probed with Nrf2 antibody. Test compounds (10 M), MG132 (25 M) and
sulforaphane (SUL, 10 M) were incubated with Hepa1c1c7 cells for 6 h. MG132 is
m
m
Compound
m
a proteosome inhibitor and causes the accumulation of Nrf2. Immunoblots shown
were representative of 3 experiments. (B) Quantitation of immunoblots by image
densitometry. (*) indicates that the intensity of the blot is greater than that of control
(p < 0.05, independent T-test).
Fig. 6. Induction of TrxR activity by test compounds and sulforaphane (SUL) at 10 mM,
24 h incubation in Hepa1c1c7 cells. Fold induction is measured by the ratio of absor-
bances of treated cells versus control cells. Mean ꢁ SEM (n ¼ 3). * indicates significant
induction (p < 0.01, 1-way ANOVA with Dunnett as posthoc test).

C.-Y. Lee et al. / European Journal of Medicinal Chemistry 45 (2010) 2957e2971
2965
Control
1-1
1-5
8-2
SUL
variation in ELUMO values of the other compounds did not
adequately reflect the variation in their CD values (Table 2). For
example, 5-1 which was the other poor NQO1 inducer (CD esti-
2-5
5-1
TrxR
mated to be 50 mM) included in the analysis, had an ELUMO
β-actin
(ꢂ2.17 eV) that was comparable to other potent inducers (2-2, 3-
3, 3-5). Thus, it may be that the low ELUMO value of 8-2 coupled
with its negligible NQO1 induction activity influenced the
outcome of the GA analysis. To test this possibility, the analysis
was repeated without 8-2, and with other conditions unchanged.
Indeed, we found that the resulting model no longer included
ELUMO as a descriptor (Equation (4)). It would seem that the
usefulness of the ELUMO term is limited to compounds that
possess or lack the exocyclic double bond but it is not an
adequate gauge of induction activity among compounds with
intact double bonds.
0.6
0.5
0.4
0.3
0.2
0.1
0
*
*
*
control
1-1
1-5
2-5
5-1
8-2
SUL
Fig. 7. Immunoblot analysis of TrxR protein. (A) Western blot analysis of Hepa1c1c7
cell lysates probed with the TrxR antibody. Test compounds (10 M) and sulforaphane
(SUL, 10 M) were incubated with Hepa1c1c7 cells for 24 h. Immunoblots shown were
m
Variable length:
m
representative of 3 experiments. (B) Quantitation of immunoblots by image densi-
tometry. (*) indicates that the intensity of the blot is greater than that of control
(p < 0.05, independent T-test).
pCD ¼ ꢂ1:91 þ 0:031½Water accessible surface areaꢄ
ꢂ 0:030½van der Waals areaꢄ
(4)
RMSE ¼ 0.542, R2 ¼ 0.364, R2 (adjusted) ¼ 0.320, AIC ¼ 59.63,
LOF ¼ 0.384, F ¼ 8.31.
pCD ¼ ꢂ5:50 ꢂ 1:94ELUMO
þ 0:023½water accessible surface areaꢄ
ꢂ 0:59½molar refractivityꢄ
(3)
4. Discussion
RMSE ¼ 0.527, RMSE (test set) ¼ 0.567, R2 ¼ 0.487, R2 (adjusted) ¼
0.377, AIC ¼ 37.99, LOF ¼ 0.624, F ¼ 4.42.
Prior to the present investigation, only two aurones have been
reported to induce NQO1 activity in Hepa1c1c7 cells [14]. Both
compounds were 6-hydroxyaurones and induced NQO1 to
Unlike PLS, the analysis by GA identified ELUMO as a key
contributor to activity. ELUMO represents the energy required for
a
molecule in the dilute gas phase to accept an electron.
a moderate degree (CD 6e20 mM). We have shown that appropriate
Compounds with lower ELUMO energies have greater affinities for
electrons and are stronger oxidizing agents. An inverse relationship
was predicted between activity (pCD) and ELUMO which meant that
potent NQO1 inducers were compounds that had low lying ELUMO
and were ready acceptors of electrons.
functionalization of the aurone template resulted in compounds
with markedly improved NQO1 induction properties. Several
potent inducers with submicromolar CD values were found among
the 4,6-dimethoxyaurones highlighting the potential of this
template for future lead optimization.
When the ELUMO values of the compounds were compared, it
was noted that 8-2 had the lowest LUMO energy and that the
The aurones are proposed to activate both the AhR/XRE and
Keap1/Nrf2/ARE pathways to bring about induction of NQO1. The
0.1
-0.0
-0.1
-0.2
Var ID (Primary)
SIMCA-P 11 - 11/28/2008 4:26:52 PM
Fig. 8. Coefficient Plot generated by PLS on 30 compounds and 17 descriptors (R2Y ¼ 0.586, Q2 ¼ 0.467). Descriptors are represented by columns and the height of each
column is directly related to its contribution to activity (pCD). A positive coefficient indicates a direct correlation to activity and a negative coefficient indicates an inverse
correlation. MW ¼ molecular weight; Log P ¼ log P (octanol/water); Rot B ¼ number of rotatable bonds; DM ¼ dipole moment; HOMO4 ¼ EHOMO; LUMO4 ¼ ELUMO; APOL ¼ sum
of atom polarizations; MR ¼ molar refractivity; HA ¼ number of H bond acceptors; HD ¼ number of H bond donors; VSA-HA ¼ van der Waals’ surface areas of
H bond
acceptors; HAS-HYD ¼ Hydrophobic van der Waals’ surface area; ASA ¼ water accessible surface area; TPSA ¼ topological polar surface area; VDW-Area and VDW-Vol ¼ van der
Waals’ area and volume; VSA-POL ¼ Polar van der Waals’ surface area.

2966
C.-Y. Lee et al. / European Journal of Medicinal Chemistry 45 (2010) 2957e2971
involvement of the AhR/XRE pathway was seen from (i) the failure
of the target compounds to induce NQO1 activity in mutant Bprc1
cells that lacked functional Ah-receptors and (ii) their ability to
induce CYP1A1 activity in Hepa1c1c7 cells. The involvement of the
Nrf2/ARE pathway was seen from the ability of selected aurones (1-
1, 1-5) to (i) increase the luciferase activity of Hepa1c1c7 cells
transiently transfected with a reporter gene under expression
regulation of an ARE from the human NQO1 gene, (ii) increase Nrf2
levels in Hepa1c1c7 cells and (iii) induce TrxR expression (with
increase in enzyme activity) and increase GSH content in
Hepa1c1c7 cells. The failure of aurones (5-1 and 8-2) that were
poor NQO1 inducers to affect these processes under similar
conditions lend further support to the contribution of Nrf2/ARE
signaling to the NQO1 induction process.
The involvement of two pathways for the induction of NQO1
raises the question as to whether aurones affected both routes to
the same degree. The present findings suggest a greater role for the
AhR/XRE pathway as seen from the failure of the aurones to induce
NQO1 activity in mutant Hepa1c1c7 cells that were defective in
AhR signaling. Moreover, the more potent aurones (1-1, 1-5) caused
only modest increases in luciferase activity which was regulated by
Nrf2/ARE signaling. A larger increase, closer to that observed
for sulforaphane under similar conditions, would be expected if
these aurones induced NQO1 activity primarily or exclusively by
the Nrf2/ARE pathway. Additional investigations using reporter cell
lines that are stably transfected with a firefly luciferase reporter
gene under expression regulation of an ARE or XRE-containing
sequence from relevant genes would provide a clearer assessment
of the relative contributions of these two pathways. Using these
stably transfected cell lines, Lee-Hilz and co-workers demonstrated
that at physiologically relevant concentrations, some flavonoids
(quercetin, fisetin and some hydroxylated flavones) induced ARE-
mediated gene expression to a greater degree than XRE-mediated
gene expression [22].
The present findings support a mixed activator status for
aurones, not unlike that reported for other flavonoids like quer-
cetin and chrysin [7,27]. It has been argued that mixed activators
have greater chemopreventive potential than mono- or bifunc-
tional inducers. Pretreatment of human colon LS-174 cells by
a combination of indolo[3,2-b]carbazole (a bifunctional inducer)
and sulforaphane (a monofunctional inducer) provided substantial
protection against the genotoxic effects of benzo[a]pyrene, and
this protection was greater than that achieved by either agent
alone [43]. Mixed activators chrysin [23] or BNF [24] were stronger
inducers of the Phase 2 enzyme UGT1A6 (controlled by both AhR
and Nrf2 activated pathways) than TCDD, a selective AhR activator.
Alternatively, the results may reflect the coordinated regulation of
the AhR and Nrf2 gene batteries that has been expounded by
others [44,45]. Proposed links between the two pathways may
involve Nrf2 as a target gene of AhR or an indirect activation of
Nrf2 by CYP1A1 generated ROS/electrophiles or a direct interac-
tion between AhR/XRE and Nrf2/ARE signaling due to their close
proximities in the regulatory regions of NQO1 [44]. Aurones may
be agents capable of exploiting the cross-talk between the AhR
and Nrf2 gene batteries by mechanisms that remain to be
understood.
double bond was reduced (Series 8) or omitted (1-27, 2-16). This
feature was not identified by any of the descriptors in the PLS
analysis but it was captured by the ELUMO term in the GA approach.
This raises the question as to what aspect of the double bond is
actually reflected by ELUMO. By its definition, this should be
the electron affinity (electrophilicity) of the double bond and the
Michael acceptor of which it is a part. We have shown that the
aurones activated the Nrf2/ARE pathway and accordingly, their
reactivities as inducers should be strongly linked to the electron
affinity of the Michael acceptor moieties. However this proposal is
framed with several caveats. First, we were not able to demonstrate
the reaction of Series 1 aurones with electron rich thiol reagents
like GSH and N-acetylcysteine (unpublished results). Flavonoids
also failed to show electrophilic activity but they produced elec-
trophilic metabolites that could react with glutathione or DNA
[46,47]. Second, other QSAR studies on NQO1 induction activity
recognized EHOMO and not ELUMO as a key descriptor [46,48,49]. In
these reports, compounds that were good donors of electrons (high
EHOMO) were more potent inducers. The role of EHOMO was
explained by the involvement of active metabolites that mediated
the actual biological response. Lee-Hilz and co-workers postulated
that flavonoids with a catechol moiety were oxidized to quinones
or semiquinones which were the reactive species [46]. In a study of
diphenols, a 2-step mechanism of NQO1 induction was proposed
involving an initial oxidation of the diphenol inducer to its quinone
and the subsequent oxidation of reactive thiol groups in a protein
involved in NQO1 induction by these quinones [49]. EHOMO of these
compounds would then reflect the ease with which electron
donation occurs to give these active species. As the 1,4- or 1,2-
diphenolic motif involved in quinone formation is not found in the
present series of compounds (except 1-8), a correlation to EHOMO
would not be anticipated.
Besides electrophilicity, the ELUMO term could reflect the role of
olefinic double bond in influencing the conformation of the target
compound. Reduction of the double bond causes a loss in planarity
and greater conformational flexibility for ring B. This feature may
not be adequately captured by the existing descriptors used in the
QSAR analysis and ELUMO may indirectly reflect the conformational
change. The importance of conformational planarity was high-
lighted by the failure of the non-planar flavonoid taxifolin to acti-
vate ARE-mediated gene expression compared with other planar
flavonoids like quercetin and fisetin [50]. The explanation was that
interaction with Keap1 or other effector/receptor protein involves
a stereospecific interaction which was diminished in a non-planar
compound.
The structureeactivity relationships deduced in this study
would contribute to lead optimization of aurones to give more
potent chemopreventive agents. It also provides insight into the
structural requirements of agents that activate both the AhR/XRE
and Nrf2/ARE signaling pathways. Planarity and size are important
components for interaction with the AhR [51] and these features
are evident among the aurones/isoaurones. The reactivity of
the Michael reaction acceptor is a prominent feature of NQO1
inducers that activate the Nrf2/ARE pathway but the present results
do not conclusively support the involvement of this moiety as an
electrophile.
The structural features associated with NQO1 induction activity
was investigated in some detail. The strong representation of
In conclusion, this study has highlighted several desirable
features of aurones as inducers of chemopreventive Phase II
enzymes, namely their potencies as inducers of NQO1 as seen from
the submicromolar CD values of several 4,6-dimethoxyaurones and
their ability to activate both AhR and Nrf2 gene/protein batteries.
An analysis of structureeactivity trends highlighted a stereo-
electronic role for the exocyclic double bond (for conformational
planarity and/or electrophilicity) and a substituted ring A to
maintain steric bulk and lipophilicity.
potent inducers from the 4,6-dimethoxyaurones of Series
1
compared to the other series emphasized the need for a substituted
ring A, preferably one with bulky and lipophilic groups. This was
corroborated by PLS which identified non-polar size parameters
like van der Waals’ area/volume and hydrophobic surface area as
important contributors to activity. Another crucial feature was the
exocyclic double bond as seen from the loss of activity when the

C.-Y. Lee et al. / European Journal of Medicinal Chemistry 45 (2010) 2957e2971
2967
5. Experimental
ethyl acetate (3 ꢅ 20 ml). The organic phase was washed with
brine, dried over Na2SO4, and evaporated in vacuo to yield crude
products which were purified by column chromatography using
hexane:ethyl acetate (9:1) as eluting solvents.
5.1. General details
Reagents (synthetic grade or better) were obtained from Sig-
maeAldrich Chemical Company Inc (Singapore) and used without
further purification. Melting points were determined on a Gallen-
kamp melting point apparatus and reported as uncorrected values.
Nominal mass spectra were captured on an LCQ Finnigan MAT
equipped with a chemical ionization (APCI) probe and m/z values
for the molecular ion were reported. HRMS were taken using a Q-
TOF Premier (Waters Corp., Milford, USA). 1H and 13C NMR spectra
were determined on a Bruker Spectrospin 300 Ultrashield spec-
trometer and referenced to TMS (for 1H NMR spectra). Merck silica
60 F254 sheets and Merck silica gel (0.040e0.063 mm) were used
for thin layer chromatography (TLC) and flash chromatography
respectively. Purity of final compounds were verified by combus-
tion analysis (C,H) on a Perkin Elmer PRE-2400 Elemental Analyzer
or by high pressure liquid chromatography (compound 6-2).
Spectroscopic and other analytical data of final compounds are
given in Supplementary Information.
5.4. General procedure for the synthesis of series 3
5.4.1. 5-Hydroxybenzofuran-3(2H)-one (3-14)
The method of King and Ostrum [30] was followed. Briefly,
a solution of 20, 50-dihydroxyacetophenone (16.4 mmol) in ethyl
acetate (20 ml) and chloroform (20 ml) was added to CuBr2
(33 mmol) contained in a round bottom flask under a blanket of N2.
The reaction mixture was refluxed (8 h) with vigorous stirring after
which reduced copper (I) bromide was filtered off and the filtrate
evaporated in vacuo to give 2-bromo-1-(2,5-dihydroxyphenyl)
ethanone (3-13) as an amber oil which was used without further
purification. Crude 3-13 was dissolved in methanol (10 ml), treated
with 0.5 ml of KOH (50% in distilled water) and refluxed (2 h). The
solvent was removed in vacuo and the residue purified by column
chromatography (hexane:ethyl acetate ¼ 8:1) to give 3-14 as a light
yellow solid, mp 133e135 ꢀC (lit [52] mp 152e153 ꢀC) 55% yield. 1H
NMR (DMSO-d6, 300 MHz):
d
4.74 (s, 2H), 6.86 (d, J ¼ 2.3 Hz, H4),
5.2. General procedure for the synthesis of series 1
7.12 (d, J ¼ 9 Hz, H7), 7.18 (dd, J1 ¼ 2.3 Hz, J2 ¼ 11 Hz, H6); 13C NMR
(DMSO-d6, 75 MHz):
d 75.1, 106.3, 114.1, 121.1, 126.7, 152.3, 167.2,
3,5-Dimethoxyphenoxyacetic acid (1-28) and 4,6-dimethox-
ybenzofuran-3(2H)-one (1-27) were synthesized as described
earlier [25,34]. To a solution of 1-27 (100 mg, 0.52 mmol) in
methanol (10 ml) was added the substituted benzaldehyde
(0.76 mmol), followed by a solution of KOH (500 mg, 8.92 mmol) in
distilled water (1 ml). The solution was stirred at room temperature
for 1e3 h. The desired compound was obtained as a precipitate,
removed by suction filtration, washed with cold methanol and
crystallized in ethanol or methanol to give the purified aurones. In
cases where the compounds did not precipitate, the reaction
mixture was quenched with dilute HCl and extracted with ethyl
acetate (3 ꢅ 20 ml). The organic phase was washed with brine,
dried over MgSO4, and evaporated in vacuo to yield crude products
which were purified by column chromatography with hexane:ethyl
acetate as eluting solvents.
199.9; MS (APCI) m/z [M þ 1]þ 151.2.
5.4.2. Condensation of 3-14 with substituted benzaldehydes
The procedure described in Section 5.3.2 was followed except
that 3-14 was reacted with the benzaldehyde by heating in an oil
bath at 60 ꢀC for 1 h.
5.5. General procedure for the synthesis of series 4
5.5.1. 6-Hydroxybenzofuran-3(2H)-one (4-14)
The procedure described in Section 5.4.1 was followed except
that 20, 40-dihydroxyacetophenone was the starting material. The
intermediate 2-bromo-1-(2,4-dihydroxyphenyl) ethanone (4-13)
was also used for the next step without purification. 4-14 was
obtained as a light yellow solid in 21% yield. Mp 242e244 ꢀC (lit.
[53] mp 245 ꢀC). 1H NMR (DMSO-d6, 300 MHz):
d 4.70 (s, 2H), 6.50
5.3. General procedure for the synthesis of series 2
(d, J ¼ 1.5 Hz, H7), 6.61 (dd, J1 ¼1.9 Hz, J2 ¼ 10 Hz, H5) 7.48 (d,
J ¼ 8.3 Hz, H4), 10.9 (br s, eOH); 13C NMR (DMSO-d6, 75 MHz):
5.3.1. 20, 40, 60-Trihydroxy-2-chloroacetophenone (2-17) [29]
2-17 was prepared according to the method of Beney et al. [29].
It was obtained in 39% yield, mp 208e210 ꢀC; 1H NMR (DMSO-d6,
d 80.8, 101.0, 108.1, 115.6, 131.1, 163.2, 166.3, 198.4; MS (APCI) m/z
[M þ 1]þ 151.1
300 MHz):
d
4.96 (s, CH2), 5.81 (s, 2H),10.5 (s,1H),12.0 (s, 2H, eOH).
5.5.2. Condensation of 4-14 with substituted benzaldehydes
The procedure described in Section 5.4.2 was followed.
5.3.2. 4, 6-Dihydroxybenzofuran-3(2H)-one (2-16) [29]
2-17 (5 mmol) was dissolved in methanol (30 ml), treated with
NaOAc (14 mmol) and refluxed (2 h). The solvent was removed in
vacuo to give a residue that was diluted in water, extracted with
diethyl ether, washed with brine and dried with Na2SO4. On
removal of the solvent under reduced pressure, 2-16 was obtained
as a brown solid in 88% yield, Mp 211e212 ꢀC (lit. [29] mp
5.6. General procedure for the synthesis of series 5
Benzofuran-3(2H)-one was purchased and reacted with
2-hydroxybenzaldehyde or 2-chlorobenzaldehyde as described in
Section 5.4.2 to give the desired products.
210e212 ꢀC). 1H NMR (Methanol-d4, 300 MHz):
d
4.59 (s, 2H), 5.91
5.7. General procedure for the synthesis of series 6
(d, J ¼ 1.5 Hz, H5), 5.97 (d, J ¼ 1.9 Hz, H7); 13C NMR (Methanol-d4,
75 MHz):
d
74.8, 90.1, 96.2, 99.6, 102.6, 157.5, 167.6, 175.6, 193.9; MS
5.7.1. 4, 6-Dimethoxy-2-hydroxyphenylacetic acid (6-6) [31]
(APCI) m/z [M þ 1]þ 167.2.
Briefly, 4,6-dimethoxy-2-hydroxyacetophenone (1 g, 5 mmol)
was heated with sulphur (0.24 g, 7.5 mmol) and morpholine
(1.2 ml, 7.5 mmol) at 150 ꢀC, 24 h. The reaction mixture was taken
into dichloromethane, and washed successively with brine and
dilute HCl. On removal of solvent, the residue was crystallized in
ethanol to give the thiomorpholide which was hydrolyzed by
refluxing (6 h) in 2 M NaOH. The solution was then neutralized with
1 N HCl, extracted with diethyl ether, the organic layer washed with
5.3.3. Condensation of 2-16 with substituted benzaldehydes
To a solution of 2-16 in methanol (10 ml/mmol) was added the
substituted benzaldehyde (0.76 mmol), followed by a solution of
KOH 50% in water (1.5 ml/mmol) and heated in a dedicated
microwave reactor (Biotage Initiator) at 110 ꢀC for 10e15 min. The
reaction mixture was quenched with dilute HCl and extracted with

2968
C.-Y. Lee et al. / European Journal of Medicinal Chemistry 45 (2010) 2957e2971
brine and dried with Na2SO4. On removal of solvent in vacuo, 6-6
was obtained in 16% yield. White crystals, mp 140e142 ꢀC (lit [31]
d 3.69 (s, CH2),
3.76 (s, OCH3), 3.79 (s, OCH3), 6.10 (s, 1H), 6.11 (s, 1H).
benzaldehyde (2-fluorobenzaldehyde or 3-methoxy-4-hydrox-
ybenzaldehyde). At the end of the reaction, the protecting groups
were removed by adding ethereal HCl to acidify the methanolic
solution and heating at 60 ꢀC for another 2 h. The reaction mixture
was diluted with water and extracted with ethyl acetate (3 ꢅ 20 ml),
after which the organic phase was washed with brine, dried over
Na2SO4, and evaporated in vacuo to yield the crude product which
was purified by column chromatography as described earlier.
mp 140e140.5 ꢀC); 1H NMR (CDCl3, 300 MHz):
5.7.2. 4, 6-Dimethoxybenzofuran-2(3H)-one (6-7) [31]
A mixture of 6-6 (0.5 g, 2.35 mmol) and phosphorus oxychloride
(2 ml) in dichloroethane (20 ml) was stirred at room temperature,
18 h, after which the reaction mixture was diluted with distilled
water and extracted with dichloromethane. The organic layer was
washed successively with dilute NaHCO3, water and brine, and
dried with MgSO4. The solvent was removed in vacuo and the
residue was purified by column chromatography with hexane:ethyl
acetate (5:1) to give 6-7 as a white solid in 80% yield, mp
5.11. Protection and deprotection of phenolic hydroxyl groups on 3-
hydroxybenzaldehyde, 4-hydroxybenzaldehyde, 2, 3-
dihydroxybenzaldehyde and 2,4-dihydroxybenzaldehyde
The benzaldehyde (3 mmol), pyridinium p-toluenesulphonate
(50 mg, 0.2 mmol) and 3,4-dihydro-2H-pyran (673 mg, 8 mmol)
were dissolvedindichloromethane (10 ml) and stirredfor 4 h atroom
temperature.Thereactionmixturewasthenwashedwith1 MNa2CO3
(3 ꢅ 20 ml), dried over anhydrous Na2SO4 and the solvent removed in
vacuo to give the crude tetrahydropyranyl ether as oil. It was used
without purification for the condensation with the benzofuranone as
described in earlier paragraphs. At the end of the reaction, depro-
tection was carried out by acidifying the reaction mixture with 4 M
HCl and stirring for 4 h at room temperature. The reaction mixture
was extracted with ethyl acetate and worked up as described earlier.
155e156 ꢀC (lit.[31] mp 154 ꢀC) 1H NMR (CDCl3, 300 MHz):
d 3.60
(s, 2H), 3.81 (s, OCH3), 3.82 (s, OCH3), 6.22 (d, J ¼ 1.8 Hz,1H), 6.32 (d,
J ¼ 1.8 Hz, 1H); 13C NMR (CDCl3, 75 MHz):
d 30.8, 55.5, 55.7, 89.4,
94.2, 102.3, 155.8, 156.1, 161.6, 174.8; MS (APCI) m/z [M þ 1]þ 194.7.
5.7.3. Condensation of 6-7 with substituted benzaldehydes
A solution of 6-7 (50 mg, 0.25 mmol) in acetic anhydride (5 ml)
was added with 1.5 ml of triethylamine and the substituted benz-
aldehyde (0.5 mmol). The mixture was heated at 60 ꢀC for 6e12 h.
TLC was used to monitor if reaction was completed. Thereafter, the
mixture was diluted with water, refluxed (20 min), cooled and
extracted with dichloromethane (3 ꢅ 20 ml). The organic layer was
then washed, dried with MgSO4 and removed under reduced
pressure to give the crude product which was purified by column
chromatography with hexane:ethyl acetate 8:1.
5.12. Biological evaluation
5.12.1. General details
Glutathione (reduced), 5-sulfosalicylic acid, 5,50-dithiobis-(2-
nitrobenzoic acid) (DTNB), bovine insulin, yeast glutathione
reductase, sulforaphane, beta-naphthoflavone (BNF), yeast glucose-
6-phosphate dehydrogenase, menadione, dicoumarol, 7-ethoxyr-
esorufin and salicylamide were purchased from SigmaeAldrich,
Singapore. Human thioredoxin protein was obtained from Imco
Corporation (Stockholm, Sweden, www.imcocorp.se). Bradford dye
concentrate was from Bio-Rad Laboratories (Hercules, CA, USA).
Chemicals and reagents for cell-based assays and cell culture were
purchased from SigmaeAldrich Pte Ltd, Singapore. Hepa1c1c7 and
Bprc1 cell lines were obtained from the American Type Culture
Collection (Rockville, MD, USA). Primaryantibodies against Nrf2 and
TrxR were purchased from Santa Cruz Biotechnology Inc. (Santa
5.8. General procedure for the synthesis of series 7
Commercially purchased 5-hydroxy-2-benzofuran-2-(3H)-one
(100 mg, 0.67 mmol) was dissolved in ethanol (10 ml), treated with
1.5 ml of ethanolic KOH (4 g in 100 ml) and the substituted benz-
aldehyde (1.3 mmol). The reaction mixture was stirred at room
temperature for 12 hours after which the mixture was diluted with
dilute HCl and extracted with ethyl acetate (3 ꢅ 30 ml). The organic
fractions were washed with brine and dried with Na2SO4. The
solvent was evaporated to yield the crude product which was
purified by column chromatography with hexane:ethyl acetate
(5:1) as eluting solvents.
Cruz, CA, USA). Antibody against b-actin was obtained from Sig-
maeAldrich, Singapore. Dual Luciferase Assay SystemÒ and pRL-
CMV vector were purchased from Promega Pte Ltd, Singapore. Test
compounds were dissolved in DMSO and were diluted with medium
so that the final concentration of DMSO did not exceed 0.5%.
5.9. General procedure for the synthesis of series 8 (8-1, 8-2)
1-1 or 1-5 (80 mg, 0.3 mmol) was dissolved in methanol (50 ml)
and 10 mg of 10%Pd/C was added under a blanket of N2. Reaction
with H2 (40 psi) was carried out in a Parr hydrogenator with
agitation for 24 h. At the end of the reaction, the catalyst was
removed by filtration and the filtrate was evaporated under
reduced pressure to give the crude product as an oily residue. The
desired product was obtained after purification by column chro-
matography with hexane:ethyl acetate (5:1) as eluting solvents.
5.12.2. Cell culture
Mouse hepatoma Hepa1c1c7 and Bprc1 cell lines were grown in
75 cm3 flasks in a humidified, 5% CO2 incubator at 37 ꢀC. Growth
medium was
Corp., CA, USA) supplemented with 100 U/ml penicillin G sodium,
100 g/ml streptomycin sulfate, and 10% fetal bovine serum
a-minimal essential medium (a-MEM, Invitrogen
m
(Hyclone, UT, USA). Fetal bovine serum was treated with 1 g/l of
activated charcoal and heated at 55 ꢀC for 1.5 h before addition to
medium. Cells were sub-cultured when they reached 80e90%
confluency and passage numbers 4e20 were used for experiments.
5.10. Protection and deprotection of phenolic hydroxyl groups of
2-16 [26]
A solution of 2-16 (3 mmol) and N, N-diisopropylethylamine
(2 equiv) in anhydrous DMF was cooled to 0 ꢀC. 2-Methoxyethox-
ymethyl chloride (MEM Cl, 2 equiv) was added dropwise and the
reaction mixture was stirred at room temperature, 30 min. The
mixture was then diluted with water and extracted with ethyl
acetate (3 ꢅ 60 ml). The organic fractions were washed, dried with
Na2SO4 and evaporated in vacuo to yield the product was an oil. It
was used without further purification for condensation with the
5.12.3. Procedure for NQO1 induction assay [54]
Cells were grown for 24 h in 96-well microtitre plates at
a density of 104 cells per well and subsequently incubated with test
compounds for 48 h, after which the media was decanted and the
cells lysed by incubation (10 min) with a solution (50 ml per well) of
0.8% w/v digitonin and 2 mM EDTA, after which the lysed cells were
gently agitated (10 min) prior to the addition of a freshly prepared

C.-Y. Lee et al. / European Journal of Medicinal Chemistry 45 (2010) 2957e2971
2969
reaction cocktail of the following composition: 0.5 M TriseCl (pH
7.4) (2.5 ml), bovine serum albumin (33 mg), 150 mM glucose-6-
wells. Each concentration of test compound was evaluated on 3
separate occasions.
phosphate (330
glucose-6-phosphate dehydrogenase (100 units), MTT (15 mg) and
deionized water to give a final volume of 50 ml. Menadione (1 l of
50 mM stock in acetonitrile per ml reaction mixture) was added
just before the reaction mixture (200 l per well) was dispensed
ml), 7.5 mM FAD (33 ml), 50 mM NADP (90 ml), yeast
5.12.6. ARE-dependent luciferase reporter assay
m
ARE reporter assays were carried out using a pGL3-ARE firefly
luciferase reporter plasmid constructed by Dhakshinamoorthy and
Porter [37]. Briefly, Hepa1c1c7 cells were cultured in 12-well plates.
At approximately 50% confluency, the cells were co- transfected with
m
into the wells. Five minutes after the addition of the reaction
mixture, the reaction was quenched by the addition of 0.3 mM
pGL3-ARE (1.2 mg) and pRL-CMV (10 ng, Promega Pte Ltd, Singapore)
dicoumarol in 0.5% DMSO and 5 mM K3PO4 pH 7.4 (50
m
l per well).
as an internal control for transfection efficiency, using GeneJuiceÒ
transfection reagent (Merck KGaA, Darmstadt, Germany) that was
premixedinserum-freemedium. After24 h, thetransfectionmedium
wasreplaced byculture mediumand thecells were incubatedfor12 h
in the presence of the test compound. Firefly and renilla luciferase
activities were assayed using the Dual Luciferase Assay SystemÒ from
Promega in accordance to the recommended protocol. Means and
standard deviations were based on three independent replicates of
cell transfection experiments and subsequent exposure.
A blue color due to the formation of formazan was observed in each
well and its absorbance was measured at 590 nm on a microplate
reader (Infinite M200, Tecan AG, Switzerland). Blank wells were
similar to treated wells except for the exclusion of cells and control
wells were similar to treated wells except for the absence of test
compound. The content of DMSO in blank and control wells were
0.5% v/v. NQO1 induction activity of test compound at a given
concentration was determined from the equation: Degree of
induction ¼ Atest compound ꢂ Ablank/Acontrol ꢂ Ablank where
A
is
absorbance of formazan measured at 590 nm.
5.12.7. Effect of test compounds on total GSH content of Hepa1c1c7
cells [57]
Sulforaphane and BNF were determined under similar condi-
tions as positive controls. The results were plotted (degree of
induction versus concentration) with OriginPro 7.5 (OriginLab
Corp., MA) and the concentration of test compound required to
increase the basal NQO1 activity by two fold (CD) was determined.
CD was reported as mean ꢁ SD from 3 separate determinations.
Hepa1c1c7 cells were grown in 10-cm diameter tissue culture
plates at a density of 106 cells/plate for 24 h. The cells were then
incubated with test compounds (final concentration of 10 mM) for
24 h. Each compound was tested in duplicate on separate plates,
with one plate used for GSH determination and the other for protein
determination. After 24 h, the cells were trypsinized and the media
containing detached and trypsinized cells were pooled and pelleted
by centrifugation (320 g, 5 min). The pellets were washed with ice-
cold PBS and treated with 3% w/v 5-sulfosalicylic acid (1 ml). The
contents were vortexed and centrifuged (105 g, 5 min) to remove
precipitated proteins. Total GSH content was assayed by following
the change in absorbance at 412 nm over 5 min in a cuvette volume
of 1 ml containing a final concentration of 0.6 mM DTNB, 0.2 mM
5.12.4. Determination of 7-ethoxyresorufin-O-deethylase (EROD)
activity in Hepa1c1c7 cells [55]
Hepa1c1c7 cells were grown at a density of 104 cells per well in
a 96-well plate for 24 h. The cells were incubated with test compound
for 48 h, after which the mediumwas removed, the cells washed with
200
m
l of PBS and then incubated (37 ꢀC, 40 min) with 5
mM 7-
ethoxyresorufin and 2 mM salicylamide in 200
ml of medium. The
fluorescence of resorufin was measured at lexcitation of 530 nm and
lemission of 590 nm on a fluorometer. The readings of empty wells (no
cells, “blank”) and wells with Hepa1c1c7 cells in medium containing
0.5% DMSO but without test compound (“control”) were also deter-
mined. BNF and sulforaphane were used as controls.
NADPH,10 mg/ml GSH reductase and 20 ml of sample in 0.1 M sodium
phosphate-5 mM EDTA buffer. The content of GSH in the superna-
tant was determined from a calibration plot constructed using five
GSH concentrations (0.125e1.5 nmol/200 ml 0.01 M HCl). For the
determination of protein content, the cells were lysed with digitonin
and 2 mM EDTA as described in Section 5.12.3 and subjected to the
Bradford method [58].
CYP1A1 induction activity was given by the expression:
Degree of induction ¼ FCellsþtest compound ꢂ FBlank=FControl
ꢂ FBlank
5.12.8. Preparation of cell lysates for the TrxR assay
and western blots
Hepa1c1c7 cells were grown in 10-cm diameter tissue culture
plates at a density of 1 to 2 ꢅ105 cells/ml for 24 h. The cells were
incubatedwith testcompounds(10 m
M) for24 h(5% CO2, 37 ꢀC) after
5.12.5. Determination of the cytotoxicity of test compounds by the
microculture tetrazolium (MTT) assay [56]
The MTT assay was performed to obtain the cytotoxicity profiles
of the test compounds. Briefly, Hepa1c1c7 cells were grown in 96-
well plates at 104 cells/well for 24 h. The cells were then incubated
which they were trypsinized and pelleted by centrifugation (320 g,
5 min). Cell pellets were lysed in a freshly prepared lysis solution
made up of 0.8% w/v digitonin, 2 mM EDTA, and protease inhibitor
cocktail (Roche Diagnostics, Mannheim, Germany). The lysate was
incubated at 37 ꢀC for 10 min and then gently agitated for another
10 minutes prior to collection of supernatant from cell debris by
centrifugation. Total protein concentration of the lysate was then
determined by the Bradford method.
with 5 or 25
decanted. 100
well and the plate was incubated at 37 ꢀC for 3 hours. The contents
of the wells were decanted and 150 l of DMSO was added into
mM test compounds for 48 h after which the media was
ml of MTT solution (0.5 mg/ml) was added into each
m
each well to dissolve the purple formazan crystals. Cell viability was
measured by the expression:
5.12.9. Effect of test compounds on thioredoxin reductase (TrxR)
activity
TrxR activity was measured in 96-well plates using an end point
insulin reduction assay as described by Arner and Holmgren [59].
hꢀ
ꢁ.
Cell survival ð%Þ ¼
Acellsþtest compound ꢂ Ablank
Briefly, 25
of 50 l containing 85 mM HEPES (pH 7.6), 0.3 mM insulin, 660
NADPH, 2.5 mM EDTA, and 5
mg of cell lysate was incubated in a final reaction volume
i
m
mM
ꢅ ðAuntreated cells ꢂ AblankÞ ꢅ 100
m
M human thioredoxin for 20 min at
37 ꢀC. Background controls involved incubating cell lysates under
similar conditions except without human thioredoxin. The reaction
where A is the absorbance of formazan measured at 590 nm in the
test (Acells þ test compound), control (Auntreated cells) or blank (Ablank
)

2970
C.-Y. Lee et al. / European Journal of Medicinal Chemistry 45 (2010) 2957e2971
was then quenched by adding 250
ml of 1 mM DTNB in 6 M
independent T-test (SPSS 15.0 for Windows, Chicago, IL). p val-
guanidine hydrochloride - 200 mM TriseHCl, pH 8.0 solution. The
free thiols generated from the insulin reaction were reacted with
DTNB to give the reduced product TNB, the absorbance of which is
measured at 412 nm. TrxR activity was determined by subtracting
absorbance due to the background control from that obtained with
the treated lysates.
ues < 0.05 were considered statistically significant.
Acknowledgements
This work was supported by National University of Singapore
Academic Research Fund R 148 000 084 112 (to GML). LCY grate-
fully acknowledged the financial support (research scholarship)
from Ministry of Education, Republic of Singapore and National
University of Singapore. The authors thank Dr Elias Arner and Dr
Qing Cheng of the Medical Nobel Institute for Biochemistry, Kar-
olinska Institute for their kind provision of human thioredoxin and
Dr Alan Porter of the Institute of Molecular and Cell Biology,
Singapore for providing pGL3-ARE firefly luciferase reporter
plasmid.
5.12.10. Western blot of Nrf2 and TrxR
Cell lysates (50 mg total protein) were fractionated on 15% SDS-
PAGE gel, transferred to nitrocellulose membranes (Bio-Rad Labo-
ratories Pte Ltd, Singapore), and blocked with 10% non-fat milk in
Tris-buffered saline containing 0.05% Tween 20. The membranes
were blotted with primary antibodies at room temperature for
3e5 h and then incubated with horseradish peroxidase-conjugated
secondary antibodies for 1 h. Protein bands were visualized by
enhanced chemiluminescence detection reagents (Amersham
Pharmacia Biotechnology, UK) and quantified by densitometric
measurements with the ImageJ Program (National Institute of
Health, Bethesda, USA). At least 3 independent determinations
were made for each compound. Actin was used as a control to
ensure equal protein loading.
Appendix. Supplementary information
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.ejmech.2010.03.023.
References
5.13. Molecular modeling
[1] M.B. Sporn, K.T. Liby, Nat. Clin. Pract. Oncol. 2 (2005) 518e525.
[2] N.J. Hail, M. Cortes, E.N. Drake, J.E. Spallholz, Free Rad. Biol. Med. 45 (2008)
97e110.
[3] C. Chen, A.N.T. Kong, Trends Pharmacol. Sci. 26 (2005) 318e326.
[4] Y.J. Surh, Nature Rev. Cancer 3 (2003) 768e780.
[5] B.B. Aggarwal, H. Ichikawa, Cell Cycle 4 (2005) 1201e1215.
[6] A. Scalbert, C. Manach, C. Morqand, C. Remesy, L. Jimenez, Crit. Rev. Food Sci.
Nutr. 45 (2005) 287e306.
[7] S. Yannai, A.J. Day, G. Williamson, M.J.C. Rhodes, Food Chem. Toxicol. 36
(1998) 623e630.
[8] P. Nioi, J.D. Hayes, Mutation Res. 555 (2004) 149e171.
[9] A. Giudice, M. Montella, Bioessays 28 (2006) 169e181.
[10] C. Kohle, K.W. Bock, Biochem. Pharmacol. 72 (2006) 795e805.
[11] G. Asher, J. Lotem, B. Cohen, L. Sachs, Y. Shaul, Proc. Natl. Acad. Sci. USA 30
(2001) 1188e1193.
[12] M. Cuendet, C.P. Oteham, R.C. Moon, J.M. Pezzuto, J. Nat. Prod. 69 (2006)
460e463.
[13] P.W. Pare, N. Dmitrieve, T.J. Mabry, Phytochemistry 30 (1991) 1133e1135.
[14] D.S. Jang, E.J. Park, M. Hawthorne, J.S. Vigo, J.G. Graham, F. Cabieses, B.
D. Santarsiero, A.D. Mesecar, H.H.S. Fong, R.G. Mehta, J.M. Pezzuto, A.
D. Kinghorn, J. Nat. Prod. 66 (2003) 583e587.
[15] P. Talalay, M.J. De Long, H.J. Prochaska, Proc. Natl. Acad. Sci. USA 85 (1988)
8261e8265.
[16] H.J. Prochaska, P. Talalay, Cancer Res. 48 (1988) 4776e4782.
[17] D.W. Nebert, A.L. Roe, M.Z. Dieter, W.A. Solis, Y. Yang, T.P. Dalton, Biochem.
Pharmacol. 59 (2000) 65e85.
[18] J.P. Whitlock, Annu. Rev. Pharmacol. Toxicol. 39 (1999) 103e125.
[19] F.P. Guengerich, T. Shimada, Mutat Res. 400 (1998) 201e213.
[20] P. Talalay, Chemical protection against cancer by induction of electrophile
detoxication (Phase II) enzymes. in: V. Steel, G.D. Stoner, C.W. Boone, G.J. Kelloff
(Eds.), Cellular and Molecular Targets for Chemoprevention. CRC Press, Boca
Raton, FL, 1992, pp. 193e203.
[21] W. Li, A.N. Kong, Mol. Carcinogenesis 48 (2009) 91e104.
[22] Y.Y. Lee-Hilz, S. ter Borg, W.J.H. van Berkel, I.M.C.M. Rietjens, J.M. Aarts,
Toxicol. In vitro 22 (2008) 921e926.
[23] A. Galijatovic, U.K. Walle, T. Walle, Pharmaceut. Res. 17 (2000) 21e26.
[24] B.S. Bock-Hennig, C. Kohle, K. Nill, K.W. Bock, Biochem. Pharmacol. 63 (2002)
123e128.
[25] N.J. Lawrence, D. Rennison, A.T. McGown, J.A. Hadfield, Bioorg. Med. Chem.
Lett. 13 (2003) 3759e3763.
[26] S. Okombi, D. Rival, S. Bonnet, A. Mariotte, E. Perrier, A. Boumendjel, J. Med.
Chem. 49 (2006) 329e333.
[27] N.C. Veitch, R.J. Grayer, Nat. Prod. Rep 25 (2008) 555e611.
[28] N.P. Craig, J. Med. Chem. 14 (1971) 680e684.
Aurones (Series 1e5) and isoaurones (Series 6, 7) were drawn in
the Z and E configuration respectively and geometry minimized
using the Hamiltonian forcefield MMFF94x in MOE (Chemical
Computing Group, Montreal, Canada). 8-1 and 8-2 were arbitrarily
represented as R isomers because similar descriptor values were
obtained for either R or S isomer. The following molecular
descriptors were collected for the compounds using the QuaSAR
module in MOE:molecular weight (MW), lipophilicity (Log P in
octanol/water), number of rotatable bonds (Rot B), dipole moment
(DM), molar refractivity (MR), atom polarization (APOL), number of
hydrogen bond acceptors (HA), and hydrogen bond donors (HD),
topological polar surface area (TPSA), water accessible surface area
(ASA), van der Waals surface area of hydrogen bond acceptors
(VSA-HA), van der Waals area (VDW-Area) and volume (VDW-Vol),
hydrophobic van der Waals surface area (HAS-HYD), polar van der
Waals surface area (VDW-Pol). HOMO and LUMO energies (HOMO4
and LUMO4) were calculated using SPARTAN 2006 for Linux server
(Wavefunctions Inc., Irvine, CA). Structures initially minimized by
MMFF94x in MOE were minimized again using semi-empirical
PM3, followed by restricted Hartree-Fock, and a density functional
theory (DFT) B3LYP functional using a 6e31G* basis set. EHOMO and
ELUMO energies of the final minimized structures were reported. PLS
was run on SIMCA-P 11 (Umetrics AB, Umea, Sweden) with default
settings. Water accessible surface area, van der Waals area, polar
van der Waals surface area and topological polar surface area were
transformed to their logarithmic values for this analysis. QSAR by
genetic algorithm was carried out with the QuaSAR-Evolution
module on MOE with default settings (population 100, generation
50,000, mutation probability 0.5, eugenic factor 100, initial length
4, operant density 4). In the fixed length approach, a limit of 4
descriptors was specified per model. No restriction was placed on
the number of descriptors identified in the variable length
approach. 100 equations were generated for each run and the 1st
equation obtained by either approach was selected as the “best”
equation.
[29] C. Beney, A. Mariotte, A. Boumendjel, Heterocycles 55 (2001) 967e972.
[30] L.C. King, G.K. Ostrum, J. Org. Chem. 29 (1964) 3459e3461.
[31] J. Gripenberg, B. Juselius, Acta Chem. Scand. 8 (1954) 734e737.
[32] K. Thakkar, M. Cushman, J. Org. Chem. 60 (1995) 6499e6510.
[33] A. Ur-Rahman, M.I. Choudhary, S. Hayat, A.M. Kahn, A. Ahmed, Chem. Pharm.
Bull. 49 (2001) 105e107.
5.14. Statistical analysis
[34] H.M. Sim, C.Y. Lee, P.L.R. Ee, M.L. Go, Eur. J. Pharm. Sci. 35 (2008) 293e306.
[35] S. Venkateswarlu, G.K. Panchagnula, M.B. Guraiah, G.V. Subbaraju, Tetrahe-
dron 62 (2006) 9855e9860.
Data were analyzed for statistically significant differences
using one-way ANOVA followed by Dunnett’s posthoc test, or

C.-Y. Lee et al. / European Journal of Medicinal Chemistry 45 (2010) 2957e2971
2971
[36] A.T. Dinkova-Kostova, K.T. Liby, K.K. Stephenson, W.D. Holtzclaw, X. Gao,
N. Suh, C. Williams, R. Risingsong, T. Honda, G.W. Gribble, M.B. Sporn,
P. Talalay, Proc. Natl. Acad. Sci. USA 102 (2005) 4584e4589.
[37] S. Dhakshinamoorthy, A.G. Porter, J. Biol. Chem. 279 (2004) 20096e20107.
[38] A.C. Wild, R.T. Mulcahy, Free Radic. Res. 32 (2000) 281e301.
[39] Z.H. Chen, Y. Saito, Y. Yoshida, A. Sekine, N. Noguchi, E. Niki, J. Biol. Chem. 280
(2005) 41921e41927.
[49] R.V. Bensasson, V. Zoete, A.T. Dinkova-Kostova, P. Talalay, Chem. Res. Toxicol.
21 (2008) 805e821.
[50] A.M. Boerboom, M. Vermeulen, H. van der Woude, B.I. Bremer, Y.Y. Lee-Hilz,
E. Kampman, P.J. van Bladeren, I.M. Rietjens, J.M. Aarts, Biochem. Pharmacol.
72 (2006) 217e226.
[51] I.J. Lee, K.S. Jeong, B.J. Roberts, A.T. Kallarakal, P. Fernandez-Salguero,
F.J. Gonzalez, B.J. Song, Mol. Pharmacol. 49 (1996) 980e988.
[52] M.C. Kloetzel, R.P. Dayton, B.Y. Abadir, J. Org. Chem. 20 (1955) 38e49.
[53] J.S.H. Davies, P.A. McCrea, W.L. Norris, G.R. Ramage, J. Chem. Soc. (1950)
3206e3213.
[40] J. Zhang, V. Svehlikova, Y. Bao, A.F. Howie, G.J. Beckett, G. Williamson, Carci-
nogenesis 24 (2003) 497e503.
[41] D. Livingstone, DataAnalysis forChemists. OxfordUniversityPress,Oxford, 1995.
[42] D. Rogers, A.J. Hopfinger, J. Chem. Inf. Comput. Sci. 34 (1994) 854e866.
[43] C. Bonnesen, I.M. Eggleston, J.D. Hayes, Cancer Res. 61 (2001) 6120e6130.
[44] C. Kohle, K.W. Bock, Biochem. Pharmacol. 73 (2007) 1853e1862.
[45] W. Miao, L. Hu, P.J. Scrivens, G. Batist, J. Biol. Chem. 280 (2005) 20340e20348.
[46] Y.Y. Lee-Hilz, A.M. Boerboom, A.H. Westphal, W.J. Berkel, J.M. Aarts,
I.M. Rietjens, Chem. Res. Toxicol. 19 (2006) 1499e1505.
[47] H. van der Woude, M.G. Boersma, G.M. Alink, J. Vervoort, I.M. Rietjens, Chem.
Biol. Interact. 160 (2006) 193e203.
[48] V. Zoete, M. Rougee, A.T. Dinkova-Kostova, P. Talalay, R.V. Bensasson, Free
Rad. Biol. Med. 36 (2004) 1418e1423.
[54] J.W. Fahey, A.T. Dinkova-Kostova, K.K. Stephenson, P. Talalay, Methods
Enzymol. 382 (2004) 243e258.
[55] A. Marchand, R. Barouki, M. Garlatti, Mol. Pharmacol. 65 (2004) 1029e1037.
[56] T. Mosmann, J. Immunol. Methods 65 (1983) 55.
[57] F. Tietze, Anal. Biochem. 27 (1969) 502e522.
[58] M.M. Bradford, Anal. Biochem. 72 (1976) 248e254.
[59] E.S. Arner, A. Holmgren, Measurement of thioredoxin and thioredoxin
reductase. in: M.D. Maines, L.G. Costa, D.J. Reed, S. Sassa, I.G. Sipes (Eds.),
Current Protocols in Toxicology. John Wiley & Sons, New York, 2000 Suppl. 5,
pp. 7.4.1e7.4.14.