Efficient NQO1 substrates are potent and selective anticancer agents

Article abstract of DOI:10.1021/cb4005832

A major goal of personalized medicine in oncology is the identification of drugs with predictable efficacy based on a specific trait of the cancer cell, as has been demonstrated with gleevec (presence of Bcr-Abl protein), herceptin (Her2 overexpression), and iressa (presence of a specific EGFR mutation). This is a challenging task, as it requires identifying a cellular component that is altered in cancer, but not normal cells, and discovering a compound that specifically interacts with it. The enzyme NQO1 is a potential target for personalized medicine, as it is overexpressed in many solid tumors. In normal cells NQO1 is inducibly expressed, and its major role is to detoxify quinones via bioreduction; however, certain quinones become more toxic after reduction by NQO1, and these compounds have potential as selective anticancer agents. Several quinones of this type have been reported, including mitomycin C, RH1, EO9, streptonigrin, β-lapachone, and deoxynyboquinone (DNQ). However, no unified picture has emerged from these studies, and the key question regarding the relationship between NQO1 processing and anticancer activity remains unanswered. Here, we directly compare these quinones as substrates for NQO1 in vitro, and for their ability to kill cancer cells in culture in an NQO1-dependent manner. We show that DNQ is a superior NQO1 substrate, and we use computationally guided design to create DNQ analogues that have a spectrum of activities with NQO1. Assessment of these compounds definitively establishes a strong relationship between in vitro NQO1 processing and induction of cancer cell death and suggests these compounds are outstanding candidates for selective anticancer therapy.

Full text of DOI:10.1021/cb4005832

Articles
pubs.acs.org/acschemicalbiology
Efficient NQO1 Substrates are Potent and Selective Anticancer
Agents
Elizabeth I. Parkinson, Joseph S. Bair, Megan Cismesia, and Paul J. Hergenrother*
Department of Chemistry, Roger Adams Laboratory, University of Illinois at Urbana−Champaign, Urbana, Illinois 61801, United
States
S
* Supporting Information
ABSTRACT: A major goal of personalized medicine in
oncology is the identification of drugs with predictable efficacy
based on a specific trait of the cancer cell, as has been
demonstrated with gleevec (presence of Bcr-Abl protein),
herceptin (Her2 overexpression), and iressa (presence of a
specific EGFR mutation). This is a challenging task, as it
requires identifying a cellular component that is altered in
cancer, but not normal cells, and discovering a compound that
specifically interacts with it. The enzyme NQO1 is a potential
target for personalized medicine, as it is overexpressed in many
solid tumors. In normal cells NQO1 is inducibly expressed, and
its major role is to detoxify quinones via bioreduction; however,
certain quinones become more toxic after reduction by NQO1,
and these compounds have potential as selective anticancer
agents. Several quinones of this type have been reported, including mitomycin C, RH1, EO9, streptonigrin, β-lapachone, and
deoxynyboquinone (DNQ). However, no unified picture has emerged from these studies, and the key question regarding the
relationship between NQO1 processing and anticancer activity remains unanswered. Here, we directly compare these quinones
as substrates for NQO1 in vitro, and for their ability to kill cancer cells in culture in an NQO1-dependent manner. We show that
DNQ is a superior NQO1 substrate, and we use computationally guided design to create DNQ analogues that have a spectrum of
activities with NQO1. Assessment of these compounds definitively establishes a strong relationship between in vitro NQO1
processing and induction of cancer cell death and suggests these compounds are outstanding candidates for selective anticancer
therapy.
QO1 (NAD(P)H:quinone oxidoreductase-1) is a dimeric
flavoprotein that catalyzes the 2-electron reduction of a
little-to-no NQO1 in normal tissues,9−13 compounds bio-
reduced by NQO1 to unstable hydroquinones could be potent
and selective anticancer agents.2 Several examples of NQO1-
activated compounds are reported in the literature, including
the alkylators mitomycin C (MMC),8,16,17 RH1,18−20 and
EO917,21,22 and the redox cycling compounds streptonigrin
(STN)20,23 and β-lapachone12,24 (β-lap, Figure 1B). However,
none of these compounds have been able to demonstrate the
full potential of an NQO1-activated anticancer agent either due
to activation by other reductases (MMC,2,25,26 RH1,27,28
STN29), detoxification by NQO1 under select conditions
(EO9,30,31 MMC31,32), short in vivo half-lives (RH1,33 EO9,34
β-lap35), or severe toxicity (STN36).
N
variety of quinones, as well as quinoneimines, azoaromatic, and
nitroaromatic compounds.1−3 NQO1 catalyzes this reduction
via a ping-pong mechanism in which either NADH or NADPH
can be utilized as the electron source.4 Typically, this reduction
detoxifies quinones, resulting in the formation of a relatively
stable hydroquinone that is subsequently conjugated (to
glutathione, sulfate, or glucose) and eliminated.5,6 This process
prevents 1-electron reductases (e.g., cytochrome P450 reduc-
tase and cytochrome b5 reductase) from converting the
quinone to a highly reactive semiquinone that undergoes
reduction−oxidation (redox) cycling, which would indiscrim-
inately produce high levels of toxic reactive oxygen species
(ROS) (Figure 1A).1,7
While NQO1 is usually a detoxifying enzyme, its reduction
can turn certain quinones into potent cell death-inducing
compounds. This occurs when the bioreduction of a quinone
by NQO1 results in the formation of an unstable hydroquinone
capable of either alkylating DNA or undergoing redox cycling,
generating ROS (Figure 1A).2,8 Due to the dramatic elevation
of NQO1 in many cancers (including lung,9−12 colon,9,11
pancreatic,13−15 and breast cancer9,10) and the expression of
We recently identified deoxynyboquinone (DNQ, Figure 1B)
as a potent anticancer compound and developed an efficient,
modular synthesis of this small molecule that allowed for
further biological characterization.37 These studies show that
cells treated with DNQ rapidly generate ROS,38 and this ROS
Received: April 12, 2013
Accepted: August 12, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/cb4005832 | ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
Figure 1. (A) Reduction pathways and subsequent redox cycling of quinones with 1-electron reductases (such as cytochrome P450 reductase and
cytochrome b5 reductase) and 2-electron reductases (such as NQO1). (B) Cytotoxic quinones reported to be bioactivated by NQO1.
is the main cause of cancer cell death.37 The mechanism by
which DNQ induces ROS formation has been determined to be
RESULTS AND DISCUSSION
■
Quinones as NQO1 Substrates In Vitro. The ability of
NQO1 to process various quinones (DNQ, β-lap, STN, RH1,
and MMC) in vitro was assessed. In this assay, the quinone is
via NQO1-mediated reduction followed by spontaneous
reoxidation to the quinone (Figure 1A).38 Additionally, DNQ
was shown to have potent activity in vivo against an aggressive
coincubated with both NQO1 as well as an excess of NADH,
and oxidation of NADH is followed by the decrease in
absorbance at 340 nm. DNQ, β-lap, and STN were all
substrates for NQO1, and each utilized greater than 1 equiv. of
NADH over the course of the assay, demonstrating the ability
of these quinones to redox cycle (Supporting Information (SI)
Figure 1). From these data, Michaelis−Menten curves were
generated, and apparent catalytic efficiencies were calculated
(apparent because they also reflect the kinetics of redox cycling
for each compound).
As shown in Figure 2, DNQ is a highly efficient substrate and
redox cycler, with an apparent catalytic efficiency that
approaches the diffusion controlled limit (kcat/KM = 6.2 ×
107 M−1s−1). DNQ is processed over 9 times faster than the
next best compound, β-lap (kcat/KM = 0.67 × 107 M−1·s−1), and
24 times faster than STN (kcat/KM = 0.26 × 107 M−1·s−1). RH1
and MMC are extremely poor substrates for NQO1 in this
assay with observed activity less than 100 μmol/min/μmol.
Quinones versus Cancer Cells in Culture. DNQ, β-lap,
STN, RH1, and MMC were also investigated for their potency
against cancer cells that overexpress NQO1. The lung
adenocarcinoma cell line A549 has robust expression of
NQO1,38 and we measured the NQO1 activity in the cell
orthotopic lung carcinoma model in athymic mice.38
Critically, the relationship between NQO1 processing and
quinone anticancer activity is not well understood. A study that
investigates the anticancer capacity of a single structural class of
quinones in relation to their ability to be processed by NQO1
would greatly increase this understanding. Such a study would
allow for the development of optimized compounds and would
aid in tracking any on-target and off-target toxicities in vivo.
The addition of DNQ to the spectrum of NQO1 substrates
now affords the opportunity to fully interrogate this relation-
ship between NQO1, quinones, and cancer cell death. Herein is
described the head-to-head assessment of MMC, RH1, STN, β-
lap, and DNQ. Based on the superiority of DNQ in these
experiments, computational guidance was used to design
multiple DNQ derivatives that were predicted to vary in their
ability to be processed by NQO1. These compounds were then
synthesized and assessed for their ability to be substrates for
NQO1 in vitro, and for their NQO1-dependent anticancer
activity in cell culture. These experiments strongly suggest that
highly efficient NQO1 substrates make outstanding anticancer
agents.
B
dx.doi.org/10.1021/cb4005832 | ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
tested against A549 cells in the presence of the NQO1
inhibitors dicoumarol (DIC, 25 μM, Figure 3A) and ES936
(100 nM, Figure 3B). DIC is a competitive inhibitor of NQO1
(KI = 1−10 nM) that interacts with the NAD(P)H binding
site.39 ES936 is a potent mechanism based inhibitor which
alkylates key tyrosine residues within the NQO1 active site.40
While neither compound has perfect specificity for
NQO1,40−42 both DIC and ES936 are widely used in studies
of NQO1-mediated cell death,12,24,38,43,44 as incubation of cells
with these inhibitors is effective in blocking the enzymatic
activity of NQO1.15,44−46 As shown in Figure 3A and B,
coincubation with DIC and ES936 dramatically protects A549
cells from DNQ-mediated cell death, shifting the IC50 53-fold
and >170-fold, respectively (the fold is the ratio of the IC50 of
cotreatment with quinone and inhibitor to the IC50 of
treatment with only quinone, and a higher ratio indicates
greater protection and greater NQO1 selectivity). DIC and
ES936 also protect cells from β-lap-induced cell death, shifting
the IC50 10-fold and 6-fold (Figure 3A and B). DIC has little-
to-no effect on STN, RH1, or MMC-induced cell death,
suggesting that only DNQ and β-lap kill in this assay by an
NQO1-dependent mechanism. Analogous results were ob-
served with an NQO1-overexpressing breast cancer cell line,
MCF-7 (NQO1 activity = 1900 nmol/min/μg protein), as
shown in SI Figure 2. A small but statistically significant
Figure 2. Michaelis−Menten curves for DNQ, β-lap, and STN with
NQO1. Virtually no activity (<100 μmol/min/μmol) is observed with
RH1 and MMC at these concentrations (data not shown).
lysate at 2700 nmol/min/μg protein (compared to <10 nmol/
min/μg for cells that do not express NQO1;38 see SI for
details). A549 cells were exposed to each of the five quinones
for 2 h, and cell death was assessed at 72 h. As shown by the
logistical dose response curves (Figure 3A), the five quinones
show considerably different potency in their ability to induce
cell death, with DNQ and RH1 being the most potent (IC50
0.06 μM and 0.03 μM, respectively). The quinones were also
≈
Figure 3. (A) Cell death curves of A549 cells treated for 2 h with DNQ, β-lap, streptonigrin (STN), mitomycin C (MMC), and RH1 in the presence
and absence of the NQO1 inhibitors dicoumarol (DIC, 25 μM) and (B) ES936 (100 nM) and (C) against the isogenic pancreatic cell lines
MiaPaCa-2 NS (nonsense shRNA) and shNQO1 (shRNA against NQO1). IC50 values and fold protections (Fold) for each treatment are listed
below the graphs with standard error (n ≥ 3). The fold protection is (IC50 of quinone with DIC/IC50 of quinone alone).*p < 0.05, **p < 0.01, ***p <
0.001; black stars are F-test for the curves, and purple stars are paired t tests on the IC50 values.
C
dx.doi.org/10.1021/cb4005832 | ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
Figure 4. Cell death curves for isogenic cell line pairs with varying levels of NQO1 treated with DNQ. (A) MDA-MB-231 transfected with empty
vector (231−) and the gene for NQO1 (231+), (B) H596 transfected with empty vector (H596−) and the gene for NQO1 (H596+), and (C) the
normal lung fibroblast cell line IMR90. IC50s and fold protections (Fold) for each treatment are listed below the graphs with standard error (n ≥ 3).
*p < 0.05, **p < 0.01, ***p < 0.001; black stars are F-test for the curves, and purple stars are paired t tests on the IC50 values.
Figure 5. Modeling of DNQ and derivatives in the NQO1 active site. (A) The π-stacking and hydrogen bonding interactions among DNQ (blue),
the cofactor FAD (green), and Tyr-126 and 128 (yellow) with NQO1. (B) DNQ fits deeply into the active site pocket (d = hydride donor−acceptor
distance). (C) Derivative IB-DNQ (compound 2, difference from DNQ highlighted in orange) also fits deeply into the active site. (D) Derivative 3
does not fit as deeply into the active site and has a longer hydride donor−acceptor distance.
difference in IC50 values was observed for STN, RH1, and
MMC with ES936.
These compounds were then examined utilizing the
lines. Previously, the role of NQO1 in cell death mediated by
STN, RH1, and MMC has been unclear. This head-to-head
assessment allows for the conclusion that cell death induced by
these compounds, at least in A549, MCF-7, and MiaPaCa-2
cells, is not significantly NQO1 dependent. Instead, other one-
and two-electron reductases are likely responsible for their
activity (STN, cytochrome P450 reductase and xanthine
dehydrogenase;29 RH1, cytochrome P450 reductase and
NQO2;18,27,48 MMC, cytochrome P450 reductase2,25).
Selectivity of DNQ for Cancer Cells That Express
NQO1. Between 4.4% and 20.3% of people (percentage varies
based on ethnicity) carry an NQO1*2 polymorphism that
causes them to produce inactive NQO1, which is rapidly
ubiquitinated and degraded.2 We utilized cancer cell lines
derived from such individuals (MDA-MB-231 and H596) and
examined both the wt (NQO1−) as well as isogenic cells that
had been stably transfected with the gene for NQO1
(NQO1+)12,38 (Western blots and activities can be found in
SI Figure 3). As can be seen in Figure 4A and B, the NQO1−
pancreatic cell line MiaPaCa-2, which expresses high levels of
NQO1 (1130
80 nmol/min/μg) along with an isogenic
shRNA knockdown cell line (NQO1 activity = 87 3 nmol/
min/μg).47 Again, we found that the cell line with high NQO1
activity (MiaPaCa2 NS, Figure 3C) was very sensitive to DNQ.
The knockdown cell line (MiaPaCa2 shNQO1) was less
sensitive but was still susceptible to DNQ. This result is in
agreement with the effect previously observed for β-lap in
which MiaPaCa-2 cells with ≥90 nmol/min/μg showed similar
sensitivities while cell lines with <90 nmol/min/μg were
significantly less susceptible to β-lap.47 Again, there was no
significant difference in the IC50 values of the cell lines treated
with either MMC or STN, and RH1 was actually more potent
in the cells with diminished levels of NQO1 (Figure 3C). This
data suggests that the anticancer activity of MMC, STN, and
RH1 is not dependent on NQO1 bioactivation in these cell
D
dx.doi.org/10.1021/cb4005832 | ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
Figure 6. (A) DNQ derivatives. Below each structure is the catalytic efficiency (kcat/KM) with NQO1 and IC50 against A549 cells (first value) and
MCF-7 cells (second value) in culture. (B) Correlation between in vitro processing of DNQ analogues by NQO1 (kcat/KM in M−1·s−1) with their
ability to induce death (IC50 in μM) in A549 cells and (C) MCF-7 cells. Red points are DNQ and IB-DNQ.
cells were not sensitive to DNQ while the NQO1+ cells were,
supporting the necessity of NQO1 for the cell death inducing
activity of DNQ. Finally, we examined the noncancerous lung
fibroblast cell line IMR90; this cell line has no detectable
NQO1 protein by Western blot, and no NQO1 activity is
detectable in its cell lysate (SI Figure 3F). As shown in Figure
4C, these cells are completely insensitive to DNQ, further
supporting the specificity of this strategy for cancerous over
normal tissues. Overall, these data strongly support that DNQ
is selectively activated by NQO1 and that this activation is the
major factor in the ability of DNQ to induce death in cancer
cells.
Design of DNQ Derivatives. A suite of compounds that
are structurally related, but processed by NQO1 at different
rates, would be valuable toward establishing the relationship
between in vitro NQO1 processing and anticancer activity.
Thus, due to the remarkable processing of DNQ by NQO1,
and the clear NQO1-dependent activity of DNQ against cancer
E
dx.doi.org/10.1021/cb4005832 | ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
a
Scheme 1. General Synthesis Used for the Construction of DNQ Derivatives 2−19, 21−23
a
These compounds were prepared in an average yield of 12% from the aryl bispinacol borane.
Scheme 2. Alternative Deprotection Conditions Used for Preparation DNQ Derivatives 20, 24−26
cells in culture, derivatives of this compound were designed and
synthesized, and their ability to be substrates of NQO1 was
assessed.
compounds with bulky substitutions in place of the N-methyl
such as isobutyl-DNQ (IB-DNQ, compound 2) still fit deep
into the active site and maintain π-stacking with FAD and
hydrogen bonding with Tyr-126 and Tyr-128 (Figure 5C,
hydride donor−acceptor distance of 4.41 Å), suggesting they
would be excellent substrates for NQO1. Alternatively,
compounds with multiple substitutions (e.g., 3) did not fit
deeply into the active site and thus did not π-stack or make the
appropriate hydrogen bonds (Figure 5D, hydride donor−
acceptor distance of 5.65 Å), suggesting that these compounds
would be poorer substrates for NQO1. Overall, these
computational results were encouraging, with many derivatives
fully entering the NQO1 active site, and others being at least
partially excluded. A collection of DNQ derivatives was
therefore synthesized to provide structurally related compounds
predicted to be differentially processed by NQO1.
Synthesis of DNQ Derivatives. With the modeling data in
hand, a diverse set of 25 DNQ derivatives (compounds 2−26,
Figure 6A) that were expected to have a range of NQO1
activities was synthesized, and this was accomplished using a
modified version of the original DNQ synthesis (Scheme 1).
Compounds 2−19, and 21−23 were synthesized via the four
step protocol outlined in Scheme 1, with isolated yields
comparing favorably to those from the original synthesis
(average yield over 4 steps = 12%, see Supporting Information
for complete characterization).37
The NQO1-mediated bioreduction of quinones initially
involves a hydride transfer from NAD(P)H to the associated
FAD cofactor, which subsequently transfers the hydride to the
substrate.49 Crystallographic data with known substrates such
as duroquinone show that NQO1 substrates typically π-stack
with the isoalloxazine ring of the FAD.49−51 Additionally,
depending on the substrate, the quinone oxygens can hydrogen
bond with Tyr-126, Tyr-128, or His-161. Therefore, we
hypothesized that DNQ and its active derivatives would π-
stack and hydrogen bond similarly to these known substrates.
To address this hypothesis computationally, DNQ and its
derivatives were docked into the active site of NQO1 (PDB
1DXO)50 using the Molecular Operating Environment.52
Previously, a correlation was observed between NQO1 catalytic
efficiency and the docked distance between the substrate
(carbon α to quinone carbonyl) and the FAD cofactor
(nitrogen which transfers the hydride), with compounds that
have a shorter predicted “hydride donor−acceptor distance”
being better NQO1 substrates.53 Therefore, we analyzed this
key distance parameter for DNQ and representative derivatives.
DNQ was found to form π-stacking interactions with the
isoalloxazine ring of the associated FAD cofactor and hydrogen
bonds with Tyr 126 and 128 (Figure 5A), very similar to
duroquinone.50 Additionally, the hydride donor−acceptor
distance for DNQ was found to be 3.88 Å (Figure 5B).
Many DNQ derivatives, depending on the position of
substitution, fit well into the active site. For example,
Although some of the amides bore very bulky substituents,
we were pleased to find that the Suzuki−Miyaura cross-
couplings and Buchwald−Hartwig amidations all proceeded
uniformly well. However, a handful of intermediates, specifically
F
dx.doi.org/10.1021/cb4005832 | ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
those en route to the isopropyl (20), cyclopropyl (24),
cyclooctyl (25), and alcohol (26) derivatives, were not
amenable to deprotection under the harsh hydrobromic acid
conditions. In these cases, an alternative deprotection strategy
was utilized via treatment with trifluoroacetic acid followed by
BBr3 to provide the penultimate phenols; these compounds
were subsequently oxidized under standard conditions
(salcomine and O2) to give the desired products (Scheme 2).
Evaluation of DNQ Derivatives as NQO1 Substrates. In
order to compare the newly constructed derivatives to DNQ,
the in vitro catalytic efficiency of NQO1 with these derivatives
as substrates was first assessed. All compounds were found to
be substrates of NQO1, with a range of catalytic efficiencies,
consistent with the in silico modeling; the structure of the DNQ
derivatives together with their catalytic efficiencies with NQO1
are listed in Figure 6. Derivatives with multiple substitutions,
such as 3, 5, and 6, are less efficient substrates, processed 30,
11, and 15 times less efficiently than DNQ, respectively. In
contrast, substitution of the N-methyl (R1 in Scheme 1) with
alkyl groups has little effect on the catalytic efficiency, with even
bulky compounds such as 21−24 being processed very
efficiently by NQO1. The computational modeling suggests
that extensions at this position project out of the active site
while still allowing the quinone core access to the deep pocket
(for example, Figure 5C). However, substrates bearing
extremely bulky or lengthy substitution at R1 (such as 13 or
25) were less efficient NQO1 substrates.
Evaluation of DNQ Derivatives versus Cancer Cells in
Culture. In an effort to probe the relationship of NQO1
processing with the ability to induce death of NQO1-expressing
cancer cells, the collection of DNQ analogues was assessed
versus A549 and MCF-7 cancer cells in culture (DNQ potency
in these cells is shown in Figure 3 and SI Figure 2). Each
derivative was assessed as before: cells were incubated with
compound for 2 h followed by analysis of cell death at 72 h and
IC50 values were calculated from logistical dose response curves
(values are listed in Figure 6A). Graphing the NQO1 catalytic
efficiency versus the IC50 value for A549 cells (Figure 6B) and
MCF-7 cells (Figure 6C) reveals a clear trend: compounds that
are efficient NQO1 substrates are also potent cancer cell death
inducers.
Additionally, cotreatment with DIC (Figure 7A−D) or
ES936 (SI Figure 4A) significantly protects A549 cells from
compounds that were good NQO1 substrates (e.g., 2 and 23)
while having little to no effect on compounds that were poor
substrates (e.g., 3 and 13). This is seen most clearly in the fold
protection. As would be expected, cell death induced by
compounds that are better substrates for NQO1 (higher kcat/
Km) is prevented to a greater extent by dicoumarol (higher fold
protection) (Figure 7E). Similar results were observed for
MCF-7 cells (SI Figure 5). This specificity was further
confirmed by analysis of select derivatives with isogenic cell
line pairs (MDA-MB-231 , H596 , and MIA PaCa-2 NS and
shNQO1) and the normal cell line IMR90. The isogenic cell
lines showed a similar pattern to that observed with the
chemical inhibitors (SI Figure 4B−D) and the susceptibility of
normal fibroblasts remained unchanged upon the addition of
dicoumarol (SI Figure 4E).
Figure 7. Cell death curves for (A) 2, (B) 3, (C) 13, and (D) 23 in the
presence and absence of the NQO1 inhibitor DIC (25 μM). *p < 0.05,
**p < 0.01, ***p < 0.001; black stars are F-test for the curves and blue
stars are paired t tests on the IC50 values. (E) Correlation of catalytic
efficiency with the fold protection by DIC in A549 cells. Red points
are DNQ and IB-DNQ.
success. NQO1 is an excellent target in this regard, given its
well documented overexpression in many solid tumors coupled
with existing diagnostic tests for NQO1 polymorphism and
overexpression.2,54,55 Development of NQO1 substrates as
anticancer drugs has lagged behind, however, partly due to lack
of head-to-head comparison testing of existing compounds and
undefined criteria for compound development. A key question
in any drug development program relates to what drug−target
combination is optimal for biological activity. Although such
studies typically take place in the context of enzyme inhibition,
we set out to address this in an unusual setting, the activation of
CONCLUSION
■
Predictable efficacy is a major goal of anticancer drug discovery.
Such “personalized” strategies enable preselection of patients
based on molecular markers, increasing the odds of therapeutic
G
dx.doi.org/10.1021/cb4005832 | ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
a small molecule by an enzyme. To do this required a set of
compounds that had an identical core structure (to preserve
key active site interactions) and a range of activities with the
enzyme. Thus, after determining the superiority of DNQ to
other reported NQO1 substrates, multiple derivatives of DNQ
were synthesized, and their NQO1 processing in vitro and
cancer cell death induction in cell culture were assessed.
Our study provides several critical pieces of data that are of
fundamental importance for NQO1 activation as an anticancer
strategy, and of translational importance for the development of
DNQ/derivatives as antitumor drugs:
the key features of potent NQO1 substrates are identified. This
information will be important for the development of DNQ
and derivatives as potent and selective anticancer agents.
METHODS
■
In Vitro NQO1 Assay. Deoxynyboquinone, β-lapachone, and other
quinones (0.1−100 μM) were monitored as NQO1 substrates using
an NADH recycling assay24 and recombinant NQO1 (Sigma, St.
Louis, MO), in which NADH oxidation to NAD+ was monitored by
absorbance (A340 nm) on a SpectraMax Plus 384 (Molecular Devices,
Sunnyvale, CA). Compounds in DMSO stock (2 μL of 10X stock per
well) were added to a 96 well plate. NADH (400 μM) and NQO1 (1.4
μg/mL) in 50 mM potassium phosphate buffer (pH = 7.4) were added
to each well (198 μL). Data was recorded at 2 s intervals for 5 min.
Changes in absorbance were converted to changes in NADH
concentration using a calibration curve for NADH. NADH oxidation
rates were compared with reactions lacking compound or containing
dicoumarol (10 μM). Initial velocities were calculated and data
expressed as dicoumarol-inhibited relative units (μmol NADH
oxidized/min·μmol protein). Initial velocities were calculated for a
variety of concentrations and Michaelis−Menten curves were
generated using KaleidaGraph4 (Syngergy Software, Reading, PA).
Cell Lines and Culture. Cell lines were obtained from either
ATCC or the Boothman lab (UT Southwestern). A549 cells were
grown in RPMI 1640 medium with 10% fetal bovine serum, 100 U/
mL penicillin, and 100 μg/mL streptomycin. MCF-7 and IMR90 cells
were grown in EMEM with 10% fetal bovine serum, 100 U/mL
penicillin, and 100 μg/mL streptomycin. MDA-MB-231 cells were
grown in RPMI 1640 medium with 10% fetal bovine serum, 100 U/
mL penicillin, and 100 μg/mL streptomycin, and 1 μg/mL puromycin
to select for transfected cells. H596 and MIA Paca-2 cells were grown
in DMEM with 10% fetal bovine serum, 100 U/mL penicillin, and 100
μg/mL streptomycin, and 1 μg/mL puromycin to select for transfected
cells. Cells were cultured at 37 °C in a 5% CO2−95% air humidified
atmosphere.
(1) A strong correlation between NQO1 bioreduction and
anticancer activity exists: compounds that are out-
standing NQO1 substrates are also outstanding anti-
cancer compounds, as clearly shown in Figure 6B and C.
Notably, as compounds become less efficient NQO1
substrates, the NQO1-mediated contribution to the cell
death diminishes (as assessed by fold-protection by DIC,
Figure 7E). Therefore, increasing NQO1 processing
enhances both the potency and the selectivity of
candidate drugs; thus, this parameter should be
maximized for the development of DNQ and analogues
as anticancer agents.
(2) The existing NQO1 X-ray structures, coupled with
computational docking, can predict the propensity of a
compound to be a substrate for NQO1. This information
is expected to be very useful in the design of next-
generation compounds, as needed.
(3) This set of DNQ derivatives (Figure 6A) allows for the
first definition of a structure−activity relationship for this
important compound. The data show that replacement of
the N-methyl (R1 in Scheme 1) with other alkyl chains is
largely tolerated, as are small (e.g., methyl to ethyl) single
substitutions at R2 and R3. However, when groups larger
than methyl are placed at R2 and R3 (structure in Scheme
1), potency is generally reduced ∼10-fold in both assays
(e.g., compound 18). Importantly, we have identified a
significant number of DNQ derivatives that are both
outstanding NQO1 substrates and potent and selective
anticancer agents with activity similar to DNQ (e.g., 2, 9,
21, 23), providing a cadre of potent NQO1 substrates for
translational development.
(4) With DNQ and many of its derivatives, catalytic
processing by NQO1 is observed that is close to the
diffusion controlled limit. This is also apparent from
Figure 6B and C, in which the best compounds group
together in the lower right portion of these graphs,
suggesting that it will be difficult to identify derivatives
with better catalytic efficiency or more potent cell culture
activity. Instead, DNQ and its most active derivatives are
likely optimal compounds for the exploration of NQO1-
mediated bioreductive activation as an anticancer
strategy.
Cytotoxicity Assays. Cells were seeded at 2000 (A549, IMR90,
MDA-MB-231, H596, and MIA PaCa-2) or 5000 (MCF-7) cells per
well in 96 well plates and allowed to attach overnight. Cells were then
treated with various concentrations of compound (0.3 nM to 100 μM
unless solubility prevented this; for less soluble compounds 0.03 nM
to 10 μM was used). When investigating the effect of dicoumarol, cells
were cotreated with vehicle or 25 μM dicoumarol and the compound
of interest for 2 h with at least five technical replicates. When
determining the effect of ES936, cells were pretreated for 1 h with 100
nM ES936 and then cotreated with 100 nM ES936 and compound of
interest for 2 h with at least five technical replicates. In either case,
drug-free medium was then added and cells were allowed to grow for
48 to 72 h until control cells reached ∼100% confluence. Viability was
then assessed using the sulforhodamine B (SRB) assay.56 Cells were
fixed with 3% trichloroacetic acid for 1h and then dyed with 0.057%
SRB in 1% acetic acid. The dye was solubilized in 10 mM Tris base pH
10.4 and absorption (A510 nm) was read on a SpectraMax Plus 384
(Molecular Devices, Sunnyvale, CA). Percent death was calculated by
subtracting background from all wells and setting 0% death to DMSO-
treated controls. Percent death was plotted against concentration and
this data was fitted to a Logistic dose response curve (y = (A1 − A2)/
(1 + (x/x0)p) + A2, where A1 = initial value, A2 = final value, x0 =
center, p = power) using Origin 9.0.0 (Northampton, MA). Statistical
significance was determined using an F-test to compare the cell death
curves or a paired t test to compare the IC50 values. All data are
averages of at least three independent replicates.
NQO1-mediated activation is a promising strategy for
selective anticancer treatment, with selectivity arising from
the low/no expression of NQO1 in normal tissues and its high
expression in many solid tumors. The ability of DNQ to redox
cycle with NQO1 allows for outstanding potency, as one
molecule of DNQ can be processed by NQO1 to generate
dozens of molecules of ROS. The data reported herein
demonstrate that substrate processing by NQO1 in vitro is
directly tied to anticancer activity in cell culture. In addition,
Western Blot Analysis. Cells were grown to 70−90% confluency
at which point they were trypsinized and harvested. After
centrifugation and washing, the cells were lysed with RIPA lysis
buffer containing protease inhibitor, and cell debris was removed by
centrifugation (16000g for 5 min). Protein concentration was
determined by the Pierce BCA Assay (Thermo Scientific, Rockford,
IL) and whole cell lysate (20−40 μg) was resolved by 4−20% SDS-
PAGE gel electrophoresis at 120 V for 90 min, after which proteins
H
dx.doi.org/10.1021/cb4005832 | ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
were transferred onto nitrocellulose membranes (45 V for 2 h) and
blocked in 5% milk TBST overnight at 4 °C. The membranes were
blotted for molecules of interest with anti-NQO1 (1:2000 in 2% milk
TBST) and antiactin (1:1000 in 2% milk TBST) overnight at 4 °C.
The bound primary antibodies were detected using appropriate
secondary HRP conjugated antibodies (1:20000 for NQO1 and
1:10000 for actin in TBST) for 1 h at RT and visualized by ECL
autoradiography. The membranes were stripped in acidic methanol
and reprobed as necessary.
NQO1 Activity Assay. Cells were grown to 70−90% confluency in
T-75 flasks and harvested by trypsinization. Supernatants were
prepared as previously described.24 Cells were washed twice in ice-
cold PBS and resuspended in 500 μL of PBS, pH 7.4 containing 10
μg/μL aprotinin. Cell suspensions were sonicated on ice four times
using 3 s pulses and then centrifuged at 14000g for 20 min. Protein
concentration was determined using the Pierce BCA Assay (Thermo
Scientific, Rockford, IL). The NQO1 activity assay was then
performed as previously described.12,24 Samples containing 0, 10, 25,
and 50 μg protein were analyzed. Supernatant was added to a 1-mL
quartz cuvette. To this was added the reaction medium containing 77
μM cytochrome c (Sigma), 0.14% bovine serum albumin, 200 μM
NADH as the electron donor, and 10 μM menadione as the
intermediate electron acceptor in Tris-HCl buffer (50 mM, pH 7.5).
The initial rate of change of absorbance (A550 nm) was read on a
SpectraMax Plus 384 (Molecular Devices, Sunnyvale, CA), and the
extinction coefficient for cytochrome c (21.1 mM−1 cm−1) was used to
determine the change in concentration. These experiments were
repeated in the presence of 25 μM dicoumarol. NQO1 activity was
calculated as the dicoumarol inhibited oxidoreductase activity. At least
three independent replicates were performed.
Science Foundation predoctoral fellow, and an ACS Medicinal
Chemistry Division Predoctoral Fellow.
REFERENCES
■
(1) Siegel, D., Reigan, P., and Ross, D. One- and two-electron-
mediated reduction of quinones: Enzymology and toxicological
implications. In Advances in Bioactivation Research; Springer: New
York, 2008; Vol. IX, pp 1−29.
(2) Danson, S., Ward, T. H., Butler, J., and Ranson, M. (2004) DT-
diaphorase: A target for new anticancer drugs. Cancer Treat. Rev. 30,
437−449.
(3) Begleiter, A., and Fourie, J. (2004) Induction of NQO1 in cancer
cells. Methods Enzymol. 382, 320−351.
(4) Deller, S., Macheroux, P., and Sollner, S. (2008) Flavin-
dependent quinone reductases. Cell. Mol. Life Sci. 65, 141−160.
(5) Guo, W., Reigan, P., Siegel, D., and Ross, D. (2008) Enzymatic
reduction and glutathione conjugation of benzoquinone ansamycin
heat shock protein 90 inhibitors: Relevance for toxicity and
mechanism of action. Drug Metab. Dispos. 36, 2050−2057.
(6) Long, D. J., 2nd, Waikel, R. L., Wang, X. J., Roop, D. R., and
Jaiswal, A. K. (2001) NAD(P)H:quinone oxidoreductase 1 deficiency
and increased susceptibility to 7,12-dimethylbenz[a]-anthracene-
induced carcinogenesis in mouse skin. J. Natl. Cancer Inst. 93,
1166−1170.
(7) Siegel, D., Gustafson, D. L., Dehn, D. L., Han, J. Y., Boonchoong,
P., Berliner, L. J., and Ross, D. (2004) NAD(P)H:quinone
oxidoreductase 1: Role as a superoxide scavenger. Mol. Pharmacol.
65, 1238−1247.
(8) Siegel, D., Yan, C., and Ross, D. (2012) NAD(P)H:quinone
oxidoreductase 1 (NQO1) in the sensitivity and resistance to
antitumor quinones. Biochem. Pharmacol. 83, 1033−1040.
(9) Schlager, J. J., and Powis, G. (1990) Cytosolic NAD(P)H:
(quinone-acceptor)oxidoreductase in human normal and tumor tissue:
Effects of cigarette smoking and alcohol. Int. J. Cancer 45, 403−409.
(10) Marin, A., Lopez de Cerain, A., Hamilton, E., Lewis, A. D.,
Martinez-Penuela, J. M., Idoate, M. A., and Bello, J. (1997) DT-
diaphorase and cytochrome B5 reductase in human lung and breast
tumors. Br. J. Cancer 76, 923−929.
(11) Smitskamp-Wilms, E., Giaccone, G., Pinedo, H. M., van der
Laan, B. F., and Peters, G. J. (1995) DT-diaphorase activity in normal
and neoplastic human tissues; an indicator for sensitivity to
bioreductive agents? Br. J. Cancer 72, 917−921.
(12) Bey, E. A., Bentle, M. S., Reinicke, K. E., Dong, Y., Yang, C. R.,
Girard, L., Minna, J. D., Bornmann, W. G., Gao, J., and Boothman, D.
A. (2007) An NQO1- and PARP-1-mediated cell death pathway
induced in non-small-cell lung cancer cells by β-lapachone. Proc. Natl.
Acad. Sci. U.S.A. 104, 11832−11837.
(13) Lewis, A. M., Ough, M., Hinkhouse, M. M., Tsao, M. S.,
Oberley, L. W., and Cullen, J. J. (2005) Targeting NAD(P)H:quinone
oxidoreductase (NQO1) in pancreatic cancer. Mol. Carcinog. 43, 215−
224.
(14) Awadallah, N. S., Dehn, D., Shah, R. J., Russell Nash, S., Chen,
Y. K., Ross, D., Bentz, J. S., and Shroyer, K. R. (2008) NQO1
expression in pancreatic cancer and its potential use as a biomarker.
Appl. Immunohistochem. Mol. Morphol. 16, 24−31.
(15) Cullen, J. J., Hinkhouse, M. M., Grady, M., Gaut, A. W., Liu, J.,
Zhang, Y. P., Weydert, C. J., Domann, F. E., and Oberley, L. W. (2003)
Dicumarol inhibition of NADPH:quinone oxidoreductase induces
growth inhibition of pancreatic cancer via a superoxide-mediated
mechanism. Cancer Res. 63, 5513−5520.
(16) Siegel, D., Beall, H., Senekowitsch, C., Kasai, M., Arai, H.,
Gibson, N. W., and Ross, D. (1992) Bioreductive activation of
mitomycin C by DT-diaphorase. Biochemistry 31, 7879−7885.
(17) Fitzsimmons, S. A., Workman, P., Grever, M., Paull, K.,
Camalier, R., and Lewis, A. D. (1996) Reductase enzyme expression
across the national cancer institute tumor cell line panel: Correlation
with sensitivity to mitomycin C and EO9. J. Natl. Cancer Inst. 88, 259−
269.
Molecular Modeling of DNQ in NQO1. DNQ (or derivative)
was built, and a 10 Å water layer was built around the molecule. The
DNQ structure was then energy minimized using MOE52 with a
MMFF94x forcefield using gas phase calculations and a cutoff of 0.01.
Charges were then fixed using an MMFF94 forcefield. The NQO1
structure was downloaded from the PDB (1DXO).50 One of the
homodimers was extracted and protonated. DNQ was then modeled
into the protein active site, using the site of duroquinone to identify
the active site. It was docked using the Dock program in MOE, which
uses Triangle Matching for the placement of the small molecule and
London dG for rescoring of the placement of the small molecule.
Using LigX, the top configuration was protonated and energy was
minimized.
Synthesis of DNQ Derivatives. Synthesis of derivatives was
accomplished using a modified version of the original synthesis of
DNQ.37 Details are reported in Supporting Information.
ASSOCIATED CONTENT
■
S
* Supporting Information
Supporting figures, experimental protocols, and spectral data.
This material is available free of charge via the Internet at
http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: hergenro@uiuc.edu.
Notes
The authors declare the following competing financial
interest(s): The University of Illinois has filed patents on
DNQ and its derivatives.
ACKNOWLEDGMENTS
■
The MDA-MB-231, H596, and MIA PaCa-2 cell lines were
kindly provided by D. A. Boothman (UT Southwestern). We
thank Hani I. Kuttab for assistance in synthesis. This work was
supported by the University of Illinois. E.I.P. is a National
I
dx.doi.org/10.1021/cb4005832 | ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
(18) Nemeikaite-Ceniene, A., Sarlauskas, J., Anusevicius, Z.,
Nivinskas, H., and Cenas, N. (2003) Cytotoxicity of RH1 and related
aziridinylbenzoquinones: involvement of activation by NAD(P)-
H:quinone oxidoreductase (NQO1) and oxidative stress. Arch.
Biochem. Biophys. 416, 110−118.
(19) Dehn, D. L., Inayat-Hussain, S. H., and Ross, D. (2005) RH1
induces cellular damage in an NAD(P)H:quinone oxidoreductase 1-
dependent manner: Relationship between DNA cross-linking, cell
cycle perturbations, and apoptosis. J. Pharmacol. Exp. Ther. 313, 771−
779.
(20) Dehn, D. L., Winski, S. L., and Ross, D. (2004) Development of
a new isogenic cell-xenograft system for evaluation of NAD(P)-
H:quinone oxidoreductase-directed antitumor quinones: Evaluation of
the activity of RH1. Clin. Cancer Res. 10, 3147−3155.
(21) Walton, M. I., Bibby, M. C., Double, J. A., Plumb, J. A., and
Workman, P. (1992) DT-diaphorase activity correlates with sensitivity
to the indoloquinone EO9 in mouse and human colon carcinomas.
Eur. J. Cancer 28A, 1597−1600.
(22) Plumb, J. A., Gerritsen, M., Milroy, R., Thomson, P., and
Workman, P. (1994) Relative importance of DT-diaphorase and
hypoxia in the bioactivation of EO9 by human lung tumor cell lines.
Int. J. Radiat., Oncol., Biol., Phys. 29, 295−299.
(23) Beall, H. D., Liu, Y., Siegel, D., Bolton, E. M., Gibson, N. W.,
and Ross, D. (1996) Role of NAD(P)H:quinone oxidoreductase (DT-
diaphorase) in cytotoxicity and induction of DNA damage by
streptonigrin. Biochem. Pharmacol. 51, 645−652.
(24) Pink, J. J., Planchon, S. M., Tagliarino, C., Varnes, M. E., Siegel,
D., and Boothman, D. A. (2000) NAD(P)H:Quinone oxidoreductase
activity is the principal determinant of β-lapachone cytotoxicity. J. Biol.
Chem. 275, 5416−5424.
(25) Ross, D., Beall, H. D., Siegel, D., Traver, R. D., and Gustafson,
D. L. (1996) Enzymology of bioreductive drug activation. Br. J. Cancer
Suppl. 27, S1−8.
(26) Keyes, S. R., Fracasso, P. M., Heimbrook, D. C., Rockwell, S.,
Sligar, S. G., and Sartorelli, A. C. (1984) Role of NADPH:cytochrome
c reductase and DT-diaphorase in the biotransformation of mitomycin
C1. Cancer Res. 44, 5638−5643.
(27) Yan, C., Kepa, J. K., Siegel, D., Stratford, I. J., and Ross, D.
(2008) Dissecting the role of multiple reductases in bioactivation and
cytotoxicity of the antitumor agent 2,5-diaziridinyl-3-(hydroxymethyl)-
6-methyl-1,4-benzoquinone (RH1). Mol. Pharmacol. 74, 1657−1665.
(28) Tudor, G., Alley, M., Nelson, C. M., Huang, R., Covell, D. G.,
Gutierrez, P., and Sausville, E. A. (2005) Cytotoxicity of RH1:
NAD(P)H:quinone acceptor oxidoreductase (NQO1)-independent
oxidative stress and apoptosis induction. Anticancer Drugs 16, 381−
391.
(29) Yee, S. B., and Pritsos, C. A. (1997) Comparison of oxygen
radical generation from the reductive activation of doxorubicin,
streptonigrin, and menadione by xanthine oxidase and xanthine
dehydrogenase. Arch. Biochem. Biophys. 347, 235−241.
(30) Plumb, J. A., Gerritsen, M., and Workman, P. (1994) DT-
diaphorase protects cells from the hypoxic cytotoxicity of indoloqui-
none EO9. Br. J. Cancer 70, 1136−1143.
(31) Plumb, J. A., and Workman, P. (1994) Unusually marked
hypoxic sensitization to indoloquinone EO9 and mitomycin C in a
human colon-tumour cell line that lacks DT-diaphorase activity. Int. J.
Cancer 56, 134−139.
(32) Gutierrez, P. L. (2000) The role of NAD(P)H oxidoreductase
(DT-Diaphorase) in the bioactivation of quinone-containing anti-
tumor agents: A review. Free Radical Biol. Med. 29, 263−275.
(33) Danson, S. J., Johnson, P., Ward, T. H., Dawson, M., Denneny,
O., Dickinson, G., Aarons, L., Watson, A., Jowle, D., Cummings, J.,
Robson, L., Halbert, G., Dive, C., and Ranson, M. (2011) Phase I
pharmacokinetic and pharmacodynamic study of the bioreductive drug
RH1. Ann. Oncol. 22, 1653−1660.
(35) Miao, X. S., Song, P., Savage, R. E., Zhong, C., Yang, R. Y., Kizer,
D., Wu, H., Volckova, E., Ashwell, M. A., Supko, J. G., He, X., and
Chan, T. C. (2008) Identification of the in vitro metabolites of 3,4-
dihydro-2,2-dimethyl-2H-naphthol[1,2-b]pyran-5,6-dione (ARQ 501;
β-lapachone) in whole blood. Drug Metab. Dispos. 36, 641−648.
(36) Smith, G. M., Gordon, J. A., Sewell, I. A., and Ellis, H. (1967) A
trial of streptonigrin in the treatment of advanced malignant disease.
Br. J. Cancer 21, 295−301.
(37) Bair, J. S., Palchaudhuri, R., and Hergenrother, P. J. (2010)
Chemistry and biology of deoxynyboquinone, a potent inducer of
cancer cell death. J. Am. Chem. Soc. 132, 5469−5478.
(38) Huang, X., Dong, Y., Bey, E. A., Kilgore, J. A., Bair, J. S., Li, L. S.,
Patel, M., Parkinson, E. I., Wang, Y., Williams, N. S., Gao, J.,
Hergenrother, P. J., and Boothman, D. A. (2012) An NQO1 substrate
with potent antitumor activity that selectively kills by PARP1-induced
programmed necrosis. Cancer Res. 72, 3038−3047.
(39) Asher, G., Dym, O., Tsvetkov, P., Adler, J., and Shaul, Y. (2006)
The crystal structure of NAD(P)H quinone oxidoreductase 1 in
complex with its potent inhibitor dicoumarol. Biochemistry 45, 6372−
6378.
(40) Winski, S. L., Faig, M., Bianchet, M. A., Siegel, D., Swann, E.,
Fung, K., Duncan, M. W., Moody, C. J., Amzel, L. M., and Ross, D.
(2001) Characterization of a mechanism-based inhibitor of NAD(P)-
H:quinone oxidoreductase 1 by biochemical, X-ray crystallographic,
and mass spectrometric approaches. Biochemistry 40, 15135−15142.
(41) Yan, C., Shieh, B., Reigan, P., Zhang, Z., Colucci, M. A.,
Chilloux, A., Newsome, J. J., Siegel, D., Chan, D., Moody, C. J., and
Ross, D. (2009) Potent activity of indolequinones against human
pancreatic cancer: identification of thioredoxin reductase as a potential
target. Mol. Pharmacol. 76, 163−172.
(42) Gonzalez-Aragon, D., Alcain, F. J., Ariza, J., Jodar, L., Barbarroja,
N., Lopez-Pedrera, C., and Villalba, J. M. (2010) ES936 stimulates
DNA synthesis in HeLa cells independently on NAD(P)H:quinone
oxidoreductase 1 inhibition, through a mechanism involving p38
MAPK. Chem. Biol. Interact. 186, 174−183.
(43) Keyes, S. R., Rockwell, S., and Sartorelli, A. C. (1989)
Modification of the metabolism and cytotoxicity of bioreductive
alkylating agents by dicoumarol in aerobic and hypoxic murine tumor
cells. Cancer Res. 49, 3310−3313.
(44) Dehn, D. L., Siegel, D., Swann, E., Moody, C. J., and Ross, D.
(2003) Biochemical, cytotoxic, and genotoxic effects of ES936, a
mechanism-based inhibitor of NAD(P)H:quinone oxidoreductase 1, in
cellular systems. Mol. Pharmacol. 64, 714−720.
(45) Hosoda, S., Nakamura, W., and Hayashi, K. (1974) Properties
and reaction mechanism of DT diaphorase from rat liver. J. Biol. Chem.
249, 6416−6423.
(46) De Haan, L. H., Boerboom, A. M., Rietjens, I. M., van Capelle,
D., De Ruijter, A. J., Jaiswal, A. K., and Aarts, J. M. (2002) A
physiological threshold for protection against menadione toxicity by
human NAD(P)H:quinone oxidoreductase (NQO1) in Chinese
hamster ovary (CHO) cells. Biochem. Pharmacol. 64, 1597−1603.
(47) Li, L. S., Bey, E. A., Dong, Y., Meng, J., Patra, B., Yan, J., Xie, X.
J., Brekken, R. A., Barnett, C. C., Bornmann, W. G., Gao, J., and
Boothman, D. A. (2011) Modulating endogenous NQO1 levels
identifies key regulatory mechanisms of action of β-lapachone for
pancreatic cancer therapy. Clin. Cancer Res. 17, 275−285.
(48) Hasinoff, B. B., and Begleiter, A. (2006) The reductive
activation of the antitumor drug RH1 to its semiquinone free radical
by NADPH cytochrome P450 reductase and by HCT116 human
colon cancer cells. Free Radical Res. 40, 974−978.
(49) Li, R., Bianchet, M. A., Talalay, P., and Amzel, L. M. (1995) The
three-dimensional structure of NAD(P)H:quinone reductase, a
flavoprotein involved in cancer chemoprotection and chemotherapy:
mechanism of the two-electron reduction. Proc. Natl. Acad. Sci. U.S.A.
92, 8846−8850.
(50) Faig, M., Bianchet, M. A., Talalay, P., Chen, S., Winski, S., Ross,
D., and Amzel, L. M. (2000) Structures of recombinant human and
mouse NAD(P)H:quinone oxidoreductases: Species comparison and
(34) Loadman, P. M., Bibby, M. C., and Phillips, R. M. (2002)
Pharmacological approach towards the development of indolequinone
bioreductive drugs based on the clinically inactive agent EO9. Br. J.
Pharmacol. 137, 701−709.
J
dx.doi.org/10.1021/cb4005832 | ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
structural changes with substrate binding and release. Proc. Natl. Acad.
Sci. U.S.A. 97, 3177−3182.
(51) Faig, M., Bianchet, M. A., Winski, S., Hargreaves, R., Moody, C.
J., Hudnott, A. R., Ross, D., and Amzel, L. M. (2001) Structure-based
development of anticancer drugs: Complexes of NAD(P)H:quinone
oxidoreductase 1 with chemotherapeutic quinones. Structure 9, 659−
667.
(52) Vilar, S., Cozza, G., and Moro, S. (2008) Medicinal chemistry
and the molecular operating environment (MOE): Application of
QSAR and molecular docking to drug discovery. Curr. Top. Med.
Chem. 8, 1555−1572.
(53) Mendoza, M. F., Hollabaugh, N. M., Hettiarachchi, S. U., and
McCarley, R. L. (2012) Human NAD(P)H:quinone oxidoreductase
type I (hNQO1) activation of quinone propionic acid trigger groups.
Biochemistry 51, 8014−8026.
(54) Eickelmann, P., Schulz, W. A., Rohde, D., Schmitz-Drager, B.,
and Sies, H. (1994) Loss of heterozygosity at the NAD(P)H:quinone
oxidoreductase locus associated with increased resistance against
mitomycin C in a human bladder carcinoma cell line. Biol. Chem.
Hoppe-Seyler 375, 439−445.
(55) Kelsey, K. T., Ross, D., Traver, R. D., Christiani, D. C., Zuo, Z.
F., Spitz, M. R., Wang, M., Xu, X., Lee, B. K., Schwartz, B. S., and
Wiencke, J. K. (1997) Ethnic variation in the prevalence of a common
NAD(P)H quinone oxidoreductase polymorphism and its implications
for anti-cancer chemotherapy. Br. J. Cancer 76, 852−854.
(56) Vichai, V., and Kirtikara, K. (2006) Sulforhodamine B
colorimetric assay for cytotoxicity screening. Nat. Protocols 1, 1112−
1116.
K
dx.doi.org/10.1021/cb4005832 | ACS Chem. Biol. XXXX, XXX, XXX−XXX