Redox-activated, hypoxia-selective DNA cleavage by quinoxaline 1,4-di-N-oxide

Article abstract of DOI:10.1016/S0968-0896(01)00163-8

Quinoxaline 1,4-dioxide (4) is the historical prototype for modern heterocyclic N-oxide antitumor agen s such as 3-amino-1,2,4-benzotriazine 1,4-dioxide (tirapazamine, 1) and 3-amino-2-quinoxalinecarbonitrile 1,4-dioxide (11). Early experiments in bacterial cell lines suggested that enzymatic, single-electron reduction of quinoxaline 1,4-dioxides under low-oxygen (hypoxic) conditions leads to DNA damage. Here the ability of quinoxaline 1,4-dioxide to cleave DNA has been explicitly characerized using in vitro assays. The hypoxia-selective DNA-cleaving properties of 4 reported here may provide a chemical basis for understanding the cytotoxic and mutagenic activities of various quinoxaline 1,4-dioxide antibiotics. Copyright

Full text of DOI:10.1016/S0968-0896(01)00163-8

Bioorganic & Medicinal Chemistry 9 (2001) 2395–2401
Redox-Activated, Hypoxia-Selective DNA
Cleavage by Quinoxaline 1,4-di-N-Oxide
Brian Ganley, Goutam Chowdhury, Jennifer Bhansali,
J. Scott Daniels and Kent S. Gates*
Departments of Chemistry and Biochemistry, University of Missouri–Columbia, Columbia, MO 65211, USA
Received 27 March 2001; accepted 2 May 2001
Abstract—Quinoxaline 1,4-dioxide (4) is the historical prototype for modern heterocyclic N-oxide antitumor agents such as
3-amino-1,2,4-benzotriazine 1,4-dioxide (tirapazamine, 1) and 3-amino-2-quinoxalinecarbonitrile 1,4-dioxide (11). Early experi-
ments in bacterial cell lines suggested that enzymatic, single-electron reduction of quinoxaline 1,4-dioxides under low-oxygen
(hypoxic) conditions leads to DNA damage. Here the ability of quinoxaline 1,4-dioxide to cleave DNA has been explicitly char-
acterized using in vitro assays. The hypoxia-selective DNA-cleaving properties of 4 reported here may provide a chemical basis for
understanding the cytotoxic and mutagenic activities of various quinoxaline 1,4-dioxide antibiotics. # 2001 Elsevier Science Ltd.
All rights reserved.
Introduction
review of the literature reveals that the historical proto-
Several heterocyclic N-oxides have recently attracted
attention as potential antitumor agents.1ꢀ4 Most prom-
inent among these compounds is 3-amino-1,2,4-benzo-
triazine 1,4-dioxide (1, tirapazamine, tirazone, SR4233,
WIN59075), a heterocyclic di-N-oxide that is selectively
toxic to oxygen-poor (hypoxic) tumor cells.1,5 This
clinically promising antitumor agent is enzymatically
reduced in vivo to yield a radical intermediate (2) that
causes cytotoxic DNA strand breaks (as shown in
Scheme 1).6ꢀ13 It remains uncertain whether DNA clea-
vage is mediated directly by the drug radical 2 or by
hydroxyl radical derived from homolytic fragmentation
of 2 (Scheme 1), though recent studies have provided
support for the involvement of hydroxyl radical.9,14 The
selective toxicity of 1 toward hypoxic cells appears to
stem from the rapid deactivation of the reductively
activated form of the drug (2) by reaction with mol-
ecular oxygen (Scheme 1).3,15 Recent studies have iden-
tified a dual role for tirapazamine in DNA damage.10ꢀ12
In addition to initiating the formation of DNA radicals,
tirapazamine can transfer oxygen atoms from its
N-oxide functional groups to these radicals, converting
them to base-labile strand cleavage sites.10,11
type for this general class of medicinal agents is prob-
ably quinoxaline 1,4-di-N-oxide (4). The antibiotic
activity of quinoxaline 1,4-dioxides was first reported in
1943.16 In the early 1970’s, it was noted that the anti-
biotic activity of 4, like that observed much later for 1,
is enhanced under hypoxic conditions.17 Subsequent
studies indicated that DNA may be an important bio-
logical target for the quinoxaline 1,4-dioxides. Based in
part upon the examination of the toxicity of quinoxaline
1,4-dioxide against Escherichia coli KL399 cells com-
pared to that in DNA repair deficient strains, Suter and
coworkers concluded that redox-activated 4 ‘‘modifies
the integrity and structure’’ of cellular DNA under
hypoxic conditions.18 Consistent with its suspected
ability to damage DNA, the compound was found to be
mutagenic in bacteria and yeast and is carcinogenic in
rats.19ꢀ21 The mutagenicity of 4 is enhanced under
hypoxic conditions.22 The groundbreaking biological
experiments of Suter et al. showed that the compound is
rapidly reduced in vivo to yield radical intermediates
and they suggested that the hypoxia-selective antibiotic
activity of 4 involves enzymatic reduction of the com-
pound to a crucial oxygen-sensitive radical intermediate
(5, Scheme 2).18
Although compound 1 is the first redox-activated het-
erocyclic N-oxide to find clinical use in humans, a
In order to better understand the biological action of
quinoxaline 1,4-dioxides and to determine whether the
chemical properties of this compound are analogous to
the more extensively studied di-N-oxides such as
*Corresponding author. Tel.: +1-573-882-6763; fax: +1-573-882-
2754; e-mail: gatesk@missouri.edu
0968-0896/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(01)00163-8

2396
B. Ganley et al. / Bioorg. Med. Chem. 9 (2001) 2395–2401
Scheme 1.
Scheme 2.
tirapazamine (1), we have investigated whether com-
pound 4 can serve as a redox-activated DNA-cleaving
agent in vitro. The results described here provide the
first direct evidence that quinoxaline 1,4-dioxide is a
hypoxia-selective, redox-activated DNA-cleaving agent.
Our results provide support for the hypothesis18 that
DNA cleavage by 4 requires enzymatic one-electron
reduction of the compound to an activated, oxygen-
sensitive intermediate (5, Scheme 2) and may offer a
chemical basis for understanding the cytotoxic and
mutagenic activities of this class of compounds.
We find that 4, in conjunction with the xanthine/xan-
thine oxidase system, causes direct single-strand breaks
in DNA (Fig. 1). In the absence of the enzymatic redu-
cing system, 4 causes no detectable DNA cleavage. The
mono-N-oxide (6), which is the major metabolite
resulting from reductive activation of 4 (see below),
does not cleave DNA alone, in the presence of xanthine
oxidase, or in the presence of the complete xanthine/
xanthine oxidase reducing system. Interestingly, under
these reaction conditions, compound 4, in the presence
of xanthine oxidase, but without the enzyme substrate
xanthine, consistently yields small, but significant
amounts of DNA strand cleavage. Although this finding
initially seems inconsistent with the mechanism pro-
posed for DNA cleavage in Scheme 2, an explanation
consistent with the proposed mechanism is provided in
the final section of Results below.
Results
Redox-activated cleavage of supercoiled plasmid DNA
by quinoxaline 1,4-dioxide (4)
The results of previous biological studies of quinoxaline
1,4-di-N-oxides,18 and current understanding of the
analogous heterocyclic di-N-oxide antitumor agent tir-
apazamine (1),7ꢀ9 suggested that one-electron reductive
activation of 4 might initiate DNA cleavage by these
compounds under low-oxygen (hypoxic) conditions. In
the present studies, we employed a xanthine/xanthine
oxidase enzyme system as a one-electron reducing agent
for the activation of compound 4. This enzyme system
has been used similarly in previous studies with the het-
erocyclic N-oxide tirapazamine7,9 and other bioreduc-
tively-activated DNA-damaging agents.23ꢀ25 We initially
examined the ability of 4 to function as a redox-acti-
vated DNA-cleaving agent using a plasmid-based sys-
tem. In this assay, DNA strand scission is readily
measured by observing the conversion of supercoiled
(form I) plasmid DNA to the open circular form (form
II) resulting from nicking of the DNA backbone.
Assays were prepared in an inert atmosphere glove bag
and the solutions freeze–pump–thaw degassed or
purged with inert gas to remove molecular oxygen. Care
was taken to shield samples from excessive exposure to
light because 4 and other N-oxides are known to be
capable of mediating photochemical DNA cleavage.26,27
Figure 1. Cleavage of supercoiled plasmid DNA (pGL-2 Basic) by
quinoxaline 1,4-dioxide (4) in the presence of xanthine/xanthine
oxidase as an activating system. All reactions contain DNA (600 ng),
sodium phosphate buffer (50 mM, pH 7.0), catalase (100 mg/mL),
superoxidase dismutase (10 mg/mL) and desferal (1 mM) and were
prepared and incubated under anaerobic conditions (except lane 9).
After incubation for 15 h at 24 ꢁC the reactions were analyzed by
agarose gel electrophoresis. The value in the parenthesis after each
lane description represents the mean number of strand breaks per
plasmid molecule (S) calculated using the equation S=ꢀln f1 where f1
is the fraction of plasmid present as form I.49 Lane 1, DNA alone
(0.23ꢂ0.02); lane 2, xanthine (500 mM)/xanthine oxidase (0.4 U/mL)
(0.25ꢂ0.01); lane 3, 4 (500ꢂmM)+xanthine/xanthine oxidase (1.06ꢂ
0.05); lane 4, 4+xanthine/xanthine oxidase+methanol (500 mM)
(0.39ꢂ0.02); lane 5, 4+xanthine/xanthine oxidase+ethanol (500 mM)
(0.45ꢂ0.03); lane 6, 4+xanthine/xanthine oxidase+t-butanol
(500 mM) (0.41ꢂ0.04); lane 7, 4+xanthine-xanthine oxidase+DMSO
(500 mM) (0.35ꢂ0.02); lane 8, 4+xanthine/xanthine oxidase+
mannitol (500 mM) (0.34ꢂ0.01); lane 9, 4 (500ꢂmM)+xanthine/
xanthine oxidase+air (0.22ꢂ0.02); lane 10, 4+xanthine oxidase
(0.42ꢂ0.07); lane 11, 4 only (0.22ꢂ0.01); lane 12, 6 (500 mM)+
xanthine/xanthine oxidase (0.28ꢂ0.02).

B. Ganley et al. / Bioorg. Med. Chem. 9 (2001) 2395–2401
2397
Redox-activated DNA cleavage by 4 is significantly
inhibited by addition of radical scavenging agents such
as ethanol, methanol, mannitol and DMSO (Fig. 1). We
have previously shown that these additives do not inhi-
bit reductive metabolism of N-oxides by the xanthine/
xanthine oxidase enzyme system;9 thus, the observed
inhibition of DNA cleavage is attributed to scavenging
of DNA-cleaving radicals by these additives. DNA
cleavage by 4 is almost completely inhibited under
aerobic conditions.
reverse-phase HPLC retention time with authentic syn-
thetic material (by co-injection), and was ultimately
confirmed by LC/MS and LC/MS/MS experiments. In
keeping with the idea that compound 6 is produced by
the same chemical process that causes DNA cleavage,
we find that the formation of 6, like DNA-strand scis-
sion, is inhibited under aerobic conditions. Our findings
mesh with those of Suter and coworkers18 who sug-
gested, based on UV–vis data, that 6 was produced by
metabolism of 4 in bacterial cells. The reductive meta-
bolism characterized here for compound 4, resulting in
deoxygenation of a heterocyclic N-oxide functional
group, is analogous to that observed previously for the
triazine N-oxide, tirapazamine (1).
Side-by-side assays clearly show that redox-activated
DNA cleavage by 4 is less efficient than that by the
antitumor agent tirapazamine (1). Inspection of the data
shown in Figure 2 reveals that higher concentrations of
4 are required to achieve DNA-cleavage efficiencies
comparable to that afforded by 1. In addition, it is
interesting to note that the naturally-occurring N-oxide
antibiotic, 2-carboxyquinoxaline 1,4-oxide (7)21 can also
function as a redox-activated DNA-cleaving agent with
properties analogous to 4 (Fig. 3).
Figure 2. Cleavage of plasmid DNA by various concentrations of 1 or
4 in the presence of xanthine/xanthine oxidase as a reducing system.
All assays were prepared under hypoxic conditions as described in the
Experimental. Supercoiled plasmid DNA (600 ng) was incubated with
4 (100–500 mM ) or1 (25–225 mM), xanthine (500 mM), xanthine oxidase
(0.4 U/mL), catalase (100 mg/mL), superoxidase dismutase (10 mg/mL),
sodium phosphate buffer (50 mM, pH 7.0) and desferal (1 mM) under
anaerobic conditions at room temperature for 15 h, followed by agar-
ose gel electrophoretic analysis. Strand breaks per plasmid DNA
molecule (S) was calculated using the equation S=ꢀln f1 where f1 is
the fraction of plasmid present as form I.
Sequence-specificity of redox-activated DNA cleavage by
quinoxaline 1,4-dioxide (4)
In order to further characterize DNA strand scission by
4, we examined the sequence specificity of redox-acti-
vated cleavage of a 50-32P-labeled 145-base pair restric-
tion fragment under hypoxic conditions. Polyacrylamide
sequencing gel analysis reveals that cleavage by 4 occurs
at every base with little sequence specificity (Fig. 4).
Redox-activated DNA cleavage by 4 closely resembles
that observed for the prototypical sequence-indepen-
dent DNA-cleaving system, iron EDTA (Fig. 4).28
Figure 3. Cleavage of supercoiled plasmid DNA (pGL-2 Basic) by
2-quinoxaline carboxalic acid 1,4-dioxide (7) in the presence of xan-
thine-xanthine oxidase as an activating system. All reactions contain
DNA (1 mg), sodium phosphate buffer (50 mM, pH 7.0), catalase
(200 mg/mL) and desferal (1 mM) in a total volume of 100 mL and were
prepared and incubated under anaerobic conditions. After incubation
for 16 h at 37 ꢁC the reactions were analyzed by agarose gel electro-
phoresis. The value in the parenthesis after each lane description
represents the mean number of strand breaks per plasmid molecule (S)
calculated using the equation S=ꢀln f1 where f1 is the fraction of
plasmid present as form I.49 Lane 1, DNA alone (0.21ꢂ0.01); lane 2,
xanthine (500 mM)+xanthine oxidase (0.4 U/mL) (0.22ꢂ0.05); lane 3,
Identification of quinoxaline N-oxide (6) as a major
product resulting from hypoxic metabolism of
quinoxaline 1,4-dioxide (4)
Identification of the products resulting from reductive
metabolism of 4 under anaerobic conditions is crucial
for understanding the chemical mechanism by which
this compound effects DNA strand scission. Consistent
with the mechanism proposed in Scheme 2, we find that
treatment of 4 with xanthine/xanthine oxidase under
anaerobic conditions yields the mono-N-oxide 6 as the
major product (Fig. 5). The identity of the major meta-
bolite 6 was initially ascertained by comparison of its
7
(500 mM) (0.26ꢂ0.03); Lane 4,
7 (500 mM)+xanthine/xanthine
oxidase (1.41ꢂ0.03); Lane 5, 7 (500 mM)+xanthine/xanthine oxidase
+mannitol (100 mM) (0.38ꢂ0.004); lane 6, 7 (500 mM)+xanthine/
xanthine oxidase+DMSO (100 mM) (0.36ꢂ0.02); lane 7, 7 (500mM)
+xanthine/xanthine oxidase+ethanol (100 mM) (0.57ꢂ0.02); lane 8,
7 (500 mM)+xanthine/xanthine oxidase+methanol (100 mM) (0.75ꢂ
0.02); lane 9,
7 (500 mM)+xanthine-/xanthine oxidase+t-butanol
(100 mM) (0.62ꢂ0.01).

2398
B. Ganley et al. / Bioorg. Med. Chem. 9 (2001) 2395–2401
Incubation of 4 with xanthine/xanthine oxidase under
anaerobic conditions affords small amounts of two
products in addition to the major product 6. Further
experiments reveal that these minor products are
formed by the action of xanthine oxidase on the pri-
mary metabolite 6 (no xanthine substrate required).
This was demonstrated by the finding that incubation of
the mono-N-oxide (6) with xanthine oxidase under
anaerobic conditions produces the two metabolites (ꢃ6
and ꢃ8.5 min retention times in the reverse phase
HPLC chromatogram shown in Fig. 5). This result
suggests that the mono-N-oxide (6) can be oxidized by
xanthine oxidase, presumably undergoing conversion to
products such as 2(1H)-quinoxalinone 1-oxide (8) or
2(1H)-quinoxalinone 4-oxide (9). Xanthine oxidase is
known to catalyze the oxidation of a wide variety of
nitrogen heterocycles;29ꢀ31 however, to the best of our
knowledge, this is the first example of a heterocyclic
N-oxide serving as a substrate for this enzyme.
pound eluting at ꢃ8.5 min was identified in LC/MS and
LC/MS/MS experiments as 2-hydroxyquinoxaline (10)
by comparison with authentic material. These observa-
tions are best explained by a scenario in which xanthine
oxidase oxidizes 6 to the 2(1H)-quinoxalinone oxide (8
or 9) and the reduced form of the enzyme generated in
this initial reaction can subsequently reduce the 2(1H)-
quinoxalinone oxide (8 or 9) to 2-hydroxyquinoxaline
(10). The xanthine oxidase-mediated isomerization of 6
to 10 likely occurs via two separate reactions (not in a
single enzyme-catalyzed event where the same molecule
is oxidized and subsequently reduced at the active site).
The stepwise nature of this transformation is supported
by the finding that, when molecular oxygen is present as
an alternate electron acceptor (aerobic conditions), the
2(1H)-quinoxalinone oxide metabolite 8/9 is still pro-
duced, but its subsequent reduction to 2-hydroxy-
quinoxaline (10) is inhibited. Evidence that the
metabolite 8, 9 and 10 do not play a significant role in
DNA damage by 4 is provided by the observation that
incubation of 6 (which produces these compounds) with
the xanthine/xanthine oxidase system does not yield
significant amounts of DNA damage (Fig. 1).
Analysis of the reaction mixture by LC/MS confirms
that the compound eluting at ꢃ6 min has a mass con-
sistent with the structures
8 and 9 (m/z 204,
M+H+CH3CN; ESI, positive ion mode). It remains
unclear as to which regioisomer is produced. The com-
Discussion
Our work provides direct evidence that quinoxaline 1,4-
dioxide (4) is a redox-activated, hypoxia-selective DNA-
cleaving agent. In conjunction with the xanthine/xan-
thine oxidase one-electron reducing system, compound
4 cleaves DNA under hypoxic conditions. Redox-acti-
vated DNA cleavage by 4 is completely inhibited under
aerobic conditions. Hypoxic metabolism of 4 by xan-
thine/xanthine oxidase produces the two-electron
reduced mono-N-oxide 6 as the major product. The
metabolite 6 does not show significant DNA-cleaving
activity either alone or in the presence of the xanthine/
xanthine oxidase system. The dependence of DNA
cleavage on reductive activation, combined with the
observed inactivity of the two-electron reduced meta-
bolite 6, strongly supports the notion18 that the one-
electron reduced species (5) is a necessary intermediate
in this DNA-damage process.
The action of 4 on DNA in the assays reported here
yields direct strand breaks, consistent with the involve-
ment of radical species.32,33 In contrast, alkylating
agents typically yield alkaline-labile lesions. In addition,
strand scission is inhibited by commonly used radical
scavenging agents such as methanol, ethanol, mannitol
and dimethylsulfoxide.34 The observation that DNA
cleavage occurs with almost no sequence specificity is
also consistent with a radical-mediated cleavage process
involving hydrogen atom abstraction from the deoxyr-
ibose backbone of DNA.32,33 Inhibition of DNA clea-
vage under aerobic conditions probably results from
molecular oxygen-mediated oxidation of the activated
species (5) back to the starting material (4) (Scheme 2).
Figure 4. Cleavage of a 50-32P-labeled DNA fragment by reductively-
activated quinoxaline 1,4-dioxide (4). In a typical assay, a 145-base
pair, 50-labeled restriction fragment (ꢃ100,000 cpm) in a deoxy-
genated solution of phosphate buffer (20 mM, pH 7), desferal
(0.4 mM) and catalase (0.2 units/mL) was treated with quinoxaline 1,4-
dioxide (4, 5.5 mM), xanthine (35 mM) and xanthine oxidase (1.8 units/
mL), followed by 20% denaturing polyacrylamide sequencing gel
analysis as described in the Experimental. Lane 1, DNA alone; lane 2,
DNA+4+xanthine/xanthine oxidase; lane 3, DNA+Fe/EDTA
cleaving system; lane 4, Maxam–Gilbert G-reaction.
While the exact identity of the DNA-cleaving radical
resulting from redox activation of 4 remains uncertain,
there are two obvious possibilities to consider. First, the

B. Ganley et al. / Bioorg. Med. Chem. 9 (2001) 2395–2401
2399
quinoxaline 1,4-dioxide radical (5) may directly abstract
hydrogen atoms from the DNA backbone, followed by
elimination of water to yield the metabolite 6. Alter-
natively, the quinoxaline 1,4-dioxide radical (5) may
fragment to form the known DNA-cleaving agent
hydroxyl radical and the metabolite 6 (as shown in
Scheme 2). Analogous to our previous discussion of
DNA cleavage by tirapazamine,9 we favor the mechan-
ism involving production of hydroxyl radical for several
reasons. First, fragmentation reactions of this type are
thermodynamically favorable and are well precedented
in the chemical literature.35ꢀ40 In addition, for the spe-
cific case of redox-activated DNA damage by 4, the
involvement of hydroxyl radical receives further cir-
cumstantial support from early experiments18 showing
that the systems involved in the repair of DNA damage
by this agent in E. coli are the same as those required to
repair damage by X-rays (hydroxyl radical is the prin-
ciple DNA-damaging agent produced by X-rays).41,42
Finally, the lack of sequence specificity observed for
redox-activated DNA cleavage by 4 in our experiments
(Fig. 3) is consistent with involvement of a small, highly
reactive species such as hydroxyl radical. The observed
cleavage by 4 is similar in nature to that by the iron–EDTA
system (lane 3, Fig. 4) that is known to generate hydroxyl
radical (or an oxidant of very similar reactivity).28
Thus, traces of compound 6 formed spontaneously in
assay mixtures will be oxidized by xanthine oxidase and
the reduced enzyme will then reductively activate a
molecule of 4, which, in turn, decomposes to yield
another equivalent of 6. These reactions yield a cycle, in
which the product resulting from the reductive activa-
tion of 4 can serve as ‘fuel’ for further xanthine oxidase-
mediated activation of 4. This is a slow process relative
to activation of 4 by xanthine/xanthine oxidase (data
not shown), but given sufficiently long incubation times,
the spontaneous generation of traces of the mono-N-
oxide 6 from 4 can ultimately lead to significant DNA
cleavage in assays containing only 4 and xanthine
oxidase (no xanthine).
The hypoxia-selective antibiotic and mutagenic activity
of quinoxaline 1,4-dioxides has been known for over 20
years. Our studies provide some additional under-
standing of the chemical processes that may be respon-
sible for the biological activities of these compounds.
Recently, several substituted quinoxaline 1,4-dioxides
such as 11 with promising hypoxia-selective antitumor
activity (comparable to tirapazamine, 1) have been
reported. Although the ability of these compounds to
damage DNA has not been investigated, the results
presented here for the parent quinoxaline 1,4-dioxide
heterocycle 4 and for the carboxy-substituted analogue
7 clearly suggest that substituted di-N-oxides such as 11
can act as redox-activated, hypoxia-selective DNA-dama-
ging agents under physiologically relevant conditions.4
We are currently conducting further investigations on the
DNA-damaging properties of substituted quinoxaline
1,4-dioxides.
The finding that quinoxaline N-oxide (6) can be oxi-
dized by xanthine oxidase offers a reasonable explana-
tion for the small amounts of DNA cleavage produced
in the control lane containing xanthine oxidase and 4
(Fig. 1). The di-N-oxide 4 (like many N-oxides)43,44 is
prone to thermal or photochemical N-deoxygenation.
Figure 5. HPLC analysis of the products formed by xanthine/xanthine oxidase-mediated metabolism of quinoxaline 1,4-dioxide. The figure depicts
C18-reverse phase HPLC analysis of the products formed during hypoxic metabolism of quinoxaline 1,4 dioxide (4, 1 mM) in 50 mM sodium
phosphate buffer (pH 7.0) containing 500 mMxanthine, 0.2 mg/mL catalase, and 0.4 units/mL xanthine oxidase as described in the Experimental.

2400
B. Ganley et al. / Bioorg. Med. Chem. 9 (2001) 2395–2401
Experimental
Alpha Innotech IS-1000 digital imaging system. The
values reported are not corrected for differential stain-
ing of form I and form II DNA by ethidium bromide.
DNA cleavage assays containing radical scavengers
were performed as described above with the exception
that radical scavengers like methanol, ethanol, t-buta-
nol, DMSO or mannitol (500 mM) were added to the
reaction mixture before addition of xanthine oxidase.
Superoxide dismutase and catalase were added to all
assay mixtures to minimize background DNA damage
that might be caused by the conversion of remaining
traces of molecular oxygen into superoxide radical.
Materials
Materials were purchased from the following suppliers
and were of the highest purity available: xanthine,
mannitol, DMSO, sodium phosphate, and quinoxaline,
2-hydroxyquinoxaline, Aldrich Chemical Co. (Milwaukee,
WI); sodium acetate, tris(hydroxymethyl)aminomethane
(Tris), glycerol, and N,N0-methylenebisacrylamide, Sigma
Chemical Co. (St Louis, MO); hydrogen peroxide,
HPLC grade solvents (dichloromethane, acetonitrile,
ethyl acetate, dimethylformamide, methanol) and t-
butanol, Fisher (Pittsburgh, PA); T4 polynucleotide
kinase, calf intestinal phosphatase, DNA polymerase I,
EcoR I, Nhe I, New England Biolabs (Beverly, MA); 20-
deoxyadenosine triphosphate, Pharmacia (Piscataway,
NJ); xanthine oxidase, catalase, pBR322 plasmid DNA,
acrylamide, and ethidium bromide, Roche Molecular
Biochemicals (Indianapolis, IN); Seakem ME agarose,
FMC; ethanol; McCormick Distilling Co. (Brookfiled,
CN); urea, xylene cyanol, bromophenol blue, and
sodium dodecyl sulfate (SDS), United States Biochem-
ical (Cleveland, OH); desferal was a generous gift from
Ciba-Geigy Co.; 50-[g-32P]dATP, New England Nuclear-
DuPont (Wilmington, DE). Compounds 4 and 6 were
prepared via the oxidation of quinoxaline with trioxal-
orhenium using the general method of Sharpless and
coworkers.45 Compound 7 was prepared by selenium
dioxide-mediated oxidation of 2-methylquinoxaline 1,4-
dioxide following the method of Elina and Magidson.46
The starting 2-methylquinoxaline 1,4-dioxide for this
synthesis was prepared by the reaction of benzofuroxan
with acetone.47
Cleavage of a 50-32P-labeled 145-base pair restriction
fragment by reductively-activated quinoxaline 1,4-dioxide
A 50-labeled 145-base pair BamH I-Nhe I fragment of
pBR322 was prepared using standard techniques.48
Cleavage of this restriction fragment by quinoxaline 1,4-
dioxide was carried out in PyrexTM glass tubes, freeze–
pump–thaw degassed (3ꢄ) and sealed with a propane
torch. In a typical cleavage reaction, a degassed solution
(final volume 48 mL) containing the 50-labeled restriction
fragment (ꢃ1 million cpm), quinoxaline 1,4-dioxide
(5.5 mM), phosphate buffer (20 mM, pH 7), desferal (0.4
mM), xanthine (35 mM) was freeze–pump–thaw degas-
sed (3ꢄ) in a Pyrex tube, torch sealed and transferred to
an inert atmosphere glove box. Reactions were initiated
by breaking the Pyrex tubes open in the glove box, fol-
lowed by addition of an argon-degassed solution of
xanthine oxidase and catalase. The final concentrations
of xanthine oxidase and catalase were 1.8 and 0.2 units/
mL, respectively. Reactions were transferred to 0.5-
dram glass screw-top vials and incubated overnight at
37 ꢁC. Reactions were ethanol precipitated, briefly dried
under vacuum, and redissolved in formamide loading
buffer. Reactions were then heated for 5 min at 90 ꢁC
and cooled immediately in ice water for 5 min. The
resulting solution was then loaded onto a 20% dena-
turing polyacrylamide gel (acrylamide/bisacrylamide
19:1, 0.4 mm thick, containing 7.5 Murea) in aliquots
containing 100,000 cpm. The gel was electrophoresed at
800 V for 30 min, and then at 1800 V for 2 h in 1ꢄ TBE
buffer. Following electrophoresis, radioactivity on the
gel was visualized using Fuji RX X-ray film or phos-
phorimage analysis. Note: these gel electrophoresis
conditions employed here do not resolve the expected
phosphate and phosphoglycolate end products expected
from oxidative damage by hydroxyl radical.28
Comparison reactions in which the DNA fragment was
cleaved using an iron–EDTA system were performed as
described by Pogozelski and Tullius.28
Cleavage of supercoiled plasmid DNA
In a typical assay, supercoiled plasmid DNA (600 ng)
was incubated with 4 (100–500 mM ) or1 (25–225 mM),
xanthine (500 mM), xanthine oxidase (0.4 U/mL), cata-
lase (100 mg/mL), superoxidase dismutase (10 mg/mL),
sodium phosphate buffer (50 mM, pH 7.0) and desferal
(1 mM) in a total volume of 30 mL. A stock solution of
xanthine was prepared by dissolving the compound in
40% NaOH. Individual components of the DNA clea-
vage reactions, except DNA and the enzymes, were
degassed prior to use by three freeze-pump-thaw cycles
in Pyrex tubes and then sealed under vacuum. Sealed
tubes were scored, transferred to a glove bag purged
with argon, opened and degassed solutions used to pre-
pare individual reactions. Enzymes and DNA were
diluted with degassed water in the glove bag to prepare
working stock solutions. Reactions were initiated by
adding xanthine oxidase and the microentrifuge tube
was then wrapped with aluminum foil to prevent expo-
sure to light and the mixture incubated in the glove bag
for 15 h at room temperature (24 ꢁC). Following incu-
bation, the reactions were stopped by adding 3 mL of
50% glycerol loading buffer and the resulting mixture
loaded onto a 0.9% agarose gel. The gel was electro-
phoresed for approximately 3 h at 80 V in 1ꢄTAE buf-
fer and then stained in a solution of aqueous ethidium
bromide (0.3 mg/mL) for 1–2 h. DNA was visualized by
UV-transillumination and quantitated by using an
Detection of the products formed by xanthine/xanthine
oxidase-mediated metabolism of quinoxaline 1,4-dioxide
using HPLC
A typical anaerobic reduction experiment contained 1 mM
quinoxaline 1,4 dioxide (4) 50mMsodium phosphate
buffer (pH 7.0), 500 mMxanthine, 0.2mg/mL catalase, and
0.4 units/mL xanthine oxidase. Solutions were degassed
and reactions performed under inert atmosphere as
described above for the plasmid assays. Reactions were

B. Ganley et al. / Bioorg. Med. Chem. 9 (2001) 2395–2401
2401
incubated for 24 h at 37 ꢁC for 24 and analyzed by
HPLC on a MicrosorbTM-MV C-18 reverse-phase ana-
lytical column (100 A spheres, 5 mMpore size, 25 cm
length, 4.6 mm id, Varian) eluted with an isocratic sol-
vent mixture composed of 74% water, 25% methanol,
and 1% acetic acid at a flow rate of 0.9 mL/min. Pro-
ducts were observed by monitoring absorbance at
240 nm. The identities of the deoxygenated product (6)
and 2-hydroxyquinoxaline (10) were initially determined
by comparison of their retention times to that of
authentic standards and confirmed by coinjection
experiments.
13. Siim, B. G.; van Zijl, P. L.; Brown, J. M. Br. J. Cancer
1996, 73, 952.
14. Patterson, L. H.; Taiwo, F. A. Biochem. Pharmacol. 2000,
60, 1933.
15. Elwell, J. H.; Siim, B. G.; Evans, J. W.; Brown, J. M.
Biochem. Pharmacol. 1997, 54, 249.
16. McIlwain, H. J. Chem. Soc. (Part I) 1943, 322.
17. Hennessey, T. D.; Edwards, J. R. Vet. Rec 1972, 90, 187.
18. Suter, W.; Rosselet, A.; Knusel, F. Antimicrob. Agents
Chemother. 1978, 13, 770.
19. Tucker, M. J. J. Natl. Cancer Inst. 1975, 55, 137.
20. Voogd, C. E.; van der Stel, J. J.; Jacobs, J. J. J. A. A.
Mutat. Res. 1980, 78, 233.
21. Nunoshiba, T.; Nishioka, H. Mutat. Res. 1989, 217, 203.
22. Beutin, L.; Preller, E.; Kowalski, B. Antimicrob. Agents
Chemother. 1981, 20, 336.
23. O’Connor, C. J., McLennan, D. J., Sutton, B. M., Denny,
W. A. and Wilson, W. R. J. Chem. Soc. Perkin 2 1991
24. Chatterjee, M.; Rokita, S. E. J. Am. Chem. Soc. 1994, 116,
1690.
25. Taatjes, D. J.; Gaudiano, G.; Resing, K.; Koch, T. H. J.
Med. Chem. 1996, 39, 4135.
26. Negishi, K.; Arao, Y.; Oka, N.; Hayatsu, H. Nucleic Acids
Res. Symp. Ser. 1980, s105, 8.
27. Daniels, J. S.; Chatterji, T.; MacGillivray, L. R.; Gates,
K. S. J. Org. Chem. 1998, 63, 10027.
28. Pogozelski, W. K.; McNeese, T. J.; Tullius, T. D. J. Am.
Chem. Soc. 1995, 117, 6428.
29. Krenitsky, T. A.; Neil, S. M.; Elion, G. B.; Hitchings,
G. H. Arch. Biochem. Biophys. 1972, 150, 585.
30. McCormack, J. J.; Allen, B. A.; Hodnett, C. N. J. Het-
erocyclic Chem. 1978, 15, 1249.
LC/MS analysis of the products formed by xanthine/
xanthine oxidase-mediated metabolism of quinoxaline
1,4-dioxide
Reactions were performed as described above for HPLC
analysis. LC/MS and LC/MS/MS experiments were
carried out on a Finnigan TSQ 7000 triple quadrupole
instrument interfaced to a ThermoSeparations liquid
chromatograph (TSP4000). HPLC separation was per-
formed as described above except using an isocratic
solvent mixture composed of 75% water, 25% aceto-
nitrile, and 0.1% phosphoric acid at a flow rate of
1.5 mL/min. Positive ion electrospray was used as the
means of ionization and collision-induced dissociation
(CID) involved argon gas (ꢃ2 mtorr). Other instrument
settings included a capillary voltage of 4.5 kV, a capil-
lary temperature of 350 ꢁC, and a source temperature of
75 ꢁC.
31. Rastelli, G.; Costantino, L.; Albasini, A. J. Am. Chem.
Soc. 1997, 119, 3007.
32. Pogozelski, W. K.; Tullius, T. D. Chem. Rev. 1998, 98, 1089.
33. Pratviel, G.; Bernadou, J.; Meunier, B. Angew. Chem., Int.
Ed. Eng. 1995, 34, 746.
34. Halliwell, B.; Gutteridge, J. M. C. Methods Enzymol.
1990, 186, 1.
Acknowledgement
The authors gratefully acknowledge support of this work
by the American Cancer Society (RPG-00-028-01).
35. Barton, D. H. R.; Crich, D.; Motherwell, W. B. Tetra-
hedron 1985, 41, 3901.
36. Barton, D. H. R.; Jasberenyi, J. C.; Morrell, A. I. Tetra-
hedron Lett. 1991, 32, 311.
37. Adam, W.; Ballmaier, D.; Epe, B.; Grimm, G. N.; Saha-
Moller, C. R. Angew. Chem., Int. Ed. Eng. 1995, 34, 2156.
38. Aveline, B. M.; Kochevar, I. E.; Redmond, R. W. J. Am.
Chem. Soc. 1996, 118, 289.
39. Theodorakis, E. A.; Wilcoxen, K. M. J. Chem. Soc. Chem.
Comm 1996, 1927.
40. Nagai, K.; Carter, B. J.; Xu, J.; Hecht, S. M. J. Am.
Chem. Soc. 1991, 113, 5099.
41. Hall, E. J. Int. J. Radiation Oncol. Biol. Phys 1994, 30, 1019.
42. Armal, P. R.; Strniste, G. F.; Wallace, S. S. Radiation Res.
1977, 69, 328.
43. Albini, A.; Pietra, S. Heterocyclic N-Oxides; Albini, A.,
Pietra, S., Eds.; CRC Boca Raton, FL, 1991, pp. 120–132.
44. Kurasawa, Y.; Takada, A.; Kim, H.S J. Heterocyclic
Chem. 1995, 32, 1085.
45. Coperet, C.; Adolfson, H.; Khuong, T.-A. V.; Yudin,
A. K.; Sharpless, K.B J. Org. Chem. 1998, 63, 1740.
46. Elina, A. S.; Magidson, O.Y J. Gen. Chem. USSR Eng.
1955, 25, 145.
References and Notes
1. Brown, J. M. Cancer Res. 1999, 59, 5863.
2. Priyadarsini, K. I.; Dennis, M. F.; Naylor, M. A.; Stratford,
M. R. L.; Wardman, P. J. Am. Chem. Soc. 1996, 118, 5648.
3. Wardman, P.; Priyadarsini, K. I.; Dennis, M. F.; Everett,
S. A.; Naylor, M. A.; Patel, K. B.; Stratford, I. J.; Stratford,
M. R. L.; Tracy, M. Br. J. Cancer 1996, 74, S70.
4. Monge, A.; Palop, J. A.; de Cerain, A. L.; Senador, V.;
Martinez-Crespo, F. J.; Sainz, Y.; Narro, S.; Garcia, E.; de
Miguel, C.; Gonzalez, M.; Hamilton, E.; Barker, A. J.; Clarke,
E. D.; Greenhow, D. T. J. Med. Chem. 1995, 38, 1786.
5. Brown, J. M. Br. J. Cancer 1993, 67, 1163.
6. Patterson, A. V.; Saunders, M. P.; Chinje, E. C.; Patterson,
L. H.; Stratford, I. J. Anti-Cancer Drug Des 1998, 13, 541.
7. Laderoute, K. L.; Wardman, P.; Rauth, M. Biochem.
Pharmacol. 1988, 37, 1487.
8. Fitzsimmons, S. A.; Lewis, A. D.; Riley, R. J.; Workman,
P. Carcinogenesis 1994, 15, 1503.
9. Daniels, J. S.; Gates, K. S. J. Am. Chem. Soc. 1996, 118, 3380.
10. Daniels, J. S.; Gates, K. S.; Tronche, C.; Greenberg,
M . M C. hem. Res. Toxicol. 1998, 11, 1254.
11. Hwang, J.-T.; Greenberg, M. M.; Fuchs, T.; Gates, K. S.
Biochemistry 1999, 38, 14248.
12. Jones, G. D. D.; Weinfeld, M. Cancer Res. 1996, 56, 1584.
47. Ley, K.; Seng, F. Synthesis 1975, 415.
48. Sambrook, J.; Fritsch, E. F.; Maniatis, T.Molecular
Cloning: A Lab Manual; Cold Spring Harbor: Cold Spring,
NY, 1989.
49. Hertzberg, R. P.; Dervan, P. B. Biochemistry 1984, 23, 3934.