Photoinduced DNA cleavage by atomic oxygen precursors in aqueous solutions

Article abstract of DOI:10.1039/c3ra41597j

Reactive oxygen species are known to induce DNA strand cleavage and have been explored as treatments for cancer. The development of aqueous-soluble dibenzothiophene-S-oxide (DBTO) derivatives has made it possible to investigate the mechanism of DNA cleavage by these photoactivatable precursors of atomic oxygen. In addition to the release of atomic oxygen, DBTO can also undergo other processes such as α-cleavage. An objective of this work was to establish whether the extent of strand scission could be attributed to a direct reaction between atomic oxygen and DNA. To accomplish this aim, the extent of strand cleavage upon irradiation of three different DBTO derivatives was measured by the conversion of circular pUC19 plasmid (Form I) to nicked (Form II) as monitored by gel electrophoresis. The interaction of the sulfoxides with DNA was systematically studied by optical melt and fluorescence anisotropy experiments. Thiols are susceptible to rapid oxidation by atomic oxygen, and thus, glutathione was used as a ROS scavenger to determine if DNA cleavage was induced by the release of atomic oxygen. The results from these experiments indicated atomic oxygen was at least partially responsible for the observed strand scission. The Royal Society of Chemistry 2013.

Full text of DOI:10.1039/c3ra41597j

RSC Advances
View Article Online
View Journal | View Issue
PAPER
Photoinduced DNA cleavage by atomic oxygen
precursors in aqueous solutions3
Cite this: RSC Advances, 2013, 3,
12390
James Korang, Ismaila Emahi, Whitney R. Grither, Sara M. Baumann, Dana A. Baum*
and Ryan D. McCulla*
Reactive oxygen species are known to induce DNA strand cleavage and have been explored as treatments
for cancer. The development of aqueous-soluble dibenzothiophene-S-oxide (DBTO) derivatives has made it
possible to investigate the mechanism of DNA cleavage by these photoactivatable precursors of atomic
oxygen. In addition to the release of atomic oxygen, DBTO can also undergo other processes such as
a-cleavage. An objective of this work was to establish whether the extent of strand scission could be
attributed to a direct reaction between atomic oxygen and DNA. To accomplish this aim, the extent of
strand cleavage upon irradiation of three different DBTO derivatives was measured by the conversion of
circular pUC19 plasmid (Form I) to nicked (Form II) as monitored by gel electrophoresis. The interaction of
the sulfoxides with DNA was systematically studied by optical melt and fluorescence anisotropy
experiments. Thiols are susceptible to rapid oxidation by atomic oxygen, and thus, glutathione was used
as a ROS scavenger to determine if DNA cleavage was induced by the release of atomic oxygen. The results
from these experiments indicated atomic oxygen was at least partially responsible for the observed strand
scission.
Received 7th September 2012,
Accepted 10th May 2013
DOI: 10.1039/c3ra41597j
www.rsc.org/advances
generation, the reactivity of O(3P) in solution had not received
much attention until DBTO was discovered as an efficient,
Introduction
photoactivatable precursor.10–15 Greer and coworkers reported
the ability of DBTO to photoinduce the cleavage of plasmid
DNA.16 In their work, it was observed that DNA strand cleavage
occurred during irradiation in acetonitrile/water solutions in
the absence of molecular oxygen. It was speculated that the
photoinitiated DNA strand cleavage by DBTO occurred
through O(3P)-mediated process. This suggested DBTO had
potential as a DNA photocleavage agent that does not require
molecular oxygen. Recently, the first aqueous-soluble DBTO
derivatives were prepared and a considerable increase in the
quantum yield of photodeoxygenation was observed.17 The
increase in quantum yield suggests these derivatives could be
efficient DNA photocleavage agents since the extent of DNA
cleavage was expected to depend on the quantity of oxidant
generated during the photodeoxygenation.
Despite the known harmful physiological effects of reactive
oxygen species (ROS), judicious application of ROS has
resulted in many advances in medicine.1–3 For example, the
combination of light, photosensitizers, and molecular oxygen
to generate singlet oxygen (1O2) has been employed in
photodynamic therapy (PDT), which is an important treatment
option for cancer.4–6 In PDT, the ability of ROS to oxidatively
cleave DNA is used to destroy tumor cells. The hypoxic nature
of tumor cells hinders PDT by limiting the photochemical
generation of singlet oxygen, which limits the effectiveness of
PDT treatments.7 In this sense, other light assisted treatments
that do not require the presence of molecular oxygen for the
generation of ROS would be advantageous.8,9 An ideal
photoagent for hypoxic tumors would be (1) effective in the
absence of molecular oxygen, (2) stable until irradiated, and (3)
able to intercalate into DNA. Therefore, a two-component
system involving light and a prodrug that can intercalate and
generate ROS without the presence of molecular oxygen would
be advantageous.
Dibenzothiophene-S-oxide (DBTO) undergoes deoxygena-
tion upon irradiation, generating atomic oxygen [O(3P)] as
shown in Scheme 1.10–12 Due to inconvenient means of
Department of Chemistry, Saint Louis University, 3501 Laclede Avenue, St. Louis,
MO, USA. E-mail: rmccull2@slu.edu; dbaum1@slu.edu
Electronic supplementary information (ESI) available: UV absorption,
3
fluorescence titration. See DOI: 10.1039/c3ra41597j
Scheme 1
12390 | RSC Adv., 2013, 3, 12390–12397
This journal is ß The Royal Society of Chemistry 2013

View Article Online
RSC Advances
Paper
was determined by measuring the octanol water coefficient
Kow by the shake flask method.17,18 The Kow for the compound
3 in water was determined to be less than 0.001. 1H NMR (400
MHz, DMSO): d 8.132 (d, J = 1.2, 2H), 8.053 (d, J = 8.0, 2H),
7.803 (dd, J = 8.0, J = 1.4, 2H), 5.210 (s, 2H) 13C NMR (400 MHz,
D2O) d 147.63, 145.17, 137.20, 128.14, 127.63, 120.30 HRMS
(FAB): m/z found 358.935, calcd 359.943 (M 2 H)+, M =
Scheme 2
C
12H8O7S3.
Irradiations
To investigate this hypothesis, the capacity of water-soluble
DBTO derivatives, shown in Scheme 2, to photoinduce DNA
cleavage was investigated. In order to accomplish this aim,
5-oxodibenzothiophene-2,8-disulfonic acid (1) was synthesized
and its photochemistry compared to 4,6-hydroxymethyldi-
benzothiophene S-oxide (2) and 2,8-hydroxymethyldiben-
zothiophene S-oxide (3). The role of O(3P) in the DNA
photocleavage induced by these three sulfoxides and the
noncovalent interactions between the sulfoxides and the
plasmid DNA were investigated.
The quantum yield measurements were carried out with a 75
W Xe arc lamp focused on a monochromator for wavelength
selection. Slit widths allowed ¡6 nm of linear dispersion from
the set wavelength. Samples (4.8 mL) in a 1 cm quartz cell were
placed in a permanently mounted cell holder such that all of
the exiting light hit the sample without further focusing. For
all direct quantum yield measurements, the concentration of
the starting material was confirmed as sufficient to obtain a
minimum absorbance of 2 at the selected wavelength, and the
samples were irradiated until approximately 15% conversion
of the starting material was reached. Analysis of the reaction
mixtures at various time points was performed with an HPLC.
Photolysis of azoxybenzene to yield the rearranged product,
o-hydroxyazobenzene, was used as the actinometer.19
Materials and methods
Reagents
pUC19 plasmid DNA
Commercial materials were obtained from Sigma-Aldrich (St.
Louis, MO) or Fischer Scientific (Fair Lawn, NJ) and used
without modification, except as noted. Sulfoxides 2 and 3 were
prepared as described previously.17 The sulfoxide 1 was
prepared by the following procedure.
Plasmid pUC19 DNA was transformed into chemically compe-
tent E. coli (TOP10 chemically competent E. coli, Invitrogen,
Carlsbad, CA) following the manufacturer’s protocol. Plasmid
DNA was isolated from cultured cells using a PerfectPrep Spin
Mini kit (5PRIME, Gaithersburg, MD), followed by ethanol
precipitation. The concentration of the DNA stock was
Dibenzothiophene-2,8-disulfonic acid (1S)
quantified by UV absorbance using
Spectrophotometer (Thermo Scientific).
a
NanoDrop
Chlorosulfonic acid (0.5 ml, 7.5 mmol) was added to 40 ml
dichloromethane (CH2Cl2) at 0 uC. Dibenzothiophene (0.50 g,
2.7 mmol) was added gradually over 15 min. The reaction was
stirred for 30 min on ice. The reaction mixture was allowed to
warm to room temperature and then allowed to stir for an
additional 10 min. The reaction mixture was then placed on
ice and the precipitated products were filtered and washed
with CH2Cl2. The precipitate was dissolved in methanol and
passed through silica plug with methanol serving as the eluent
to remove salts. The product was recrystallized from water/
acetonitrile twice to yield dibenzothiophene-2,8-disulfonic
acid (0.19 g, 20% yield). 1H NMR (400 MHz, DMSO-d): d
8.397 (d, 2H, J = 1.8), 7.995 (d, J = 8.2, 2H), 7.775 (dd, J = 8.2, J =
1.8, 2H), 7.460 (s, 2H) 13C NMR (400 MHz, D2O) d 142.19,
138.99, 133.98, 123.94, 123.13, 118.75 HRMS (FAB): m/z found
342.940, calcd 342.948 (M 2 H)+, M = C12H8O6S3.
DNA photocleavage studies
An aqueous solution containing a mixture of 0.1 mM sulfoxide
and 40.0 ng ml21 pUC19 plasmid DNA was irradiated in fused-
silica test tubes at room temperature. A Luzchem LZC-4C
photoreactor with broadly emitting fluorescent bulbs centered
at 350 nm was used as the irradiation source. All samples were
purged with argon or allowed to stand in air for 1 h in the dark
prior to irradiation.
Atomic oxygen inhibition studies
A 1X PBS [11.9 mM phosphates (Na2HPO4 and KH2PO4), 137
mM NaCl, 2.7 mM KCl, pH 7.0] buffer solution containing 0.1
mM sulfoxide and varying glutathione (GSH) concentrations
was irradiated together with 40.0 ng ml21 pUC19 plasmid DNA
in fused-silica test tubes in a Luzchem LZC-4C photoreactor
using broadly emitting fluorescent bulbs centered at 350 nm.
5-oxodibenzothiophene-2,8-disulfonic acid (1)
Compound 1S (0.19 g, 0.54 mmol) was dissolved in 95 : 5
methanol/water mixture and cooled to 242 uC.
Dimethyldioxirane (DMDO, 0.8 M, 1 eq.) was added dropwise
and the reaction monitored by HPLC. When the starting
material was totally consumed (y15 min) the reaction was
quenched by removing the solvent under reduced pressure.
The mixture was purified by preparative HPLC using HASIL
100 C18 5 mm column (0.11 g, 57% yield). Solubility in water
DNA photocleavage analysis
After photolysis, 20 mL of the irradiated solution was analyzed
by agarose gel electrophoresis. Products were separated on a
1% agarose gel using 1X TAE (40 mM Tris-acetate, 1 mM
EDTA, pH 8.0) running buffer. The conversion of uncleaved
supercoiled pUC19 plasmid DNA (Form I) to its nicked (Form
II) was visualized by ethidium bromide staining. Once
This journal is ß The Royal Society of Chemistry 2013
RSC Adv., 2013, 3, 12390–12397 | 12391

View Article Online
Paper
RSC Advances
electrophoresis had been done, all the stained gels were
photographed using a FOTODYNE Gel Documentation system
equipped with a FOTO/UV 26 dual light transilluminator. The
intensity of the bands were analyzed using the ImageQuantTL
software package (GE Healthcare). The band percentage
corresponding to the amount of Form I or Form II was
obtained.
dibenzothiophene using chlorosulfonic acid, and the resulting
DBT was oxidized to DBTO with DMDO. Hydrophilicity was
gauged by the octanol/water coefficient Kow which was
obtained experimentally using the shake-flask method.21
A
solution of 8-octanol (5 mL) and water (5 mL) were allowed to
come to a saturation equilibrium overnight. 2–3 mmol of the
desired compound was added to the solution in a capped test
tube. The solution was continuously inverted over the period
of 1 h and then centrifuged to eliminate emulsion. The octanol
and water layers were then separated, and concentration of
compound in each layer was assessed by HPLC.17 The value of
Kow for the sulfoxides were less than 0.001, 2.55 and 1.88 for 1,
2 and 3 respectively. These values indicates the aqueous
solubility of 1 is significantly higher than 2 and 3. In general,
these DBTO derivatives are very soluble in water compared to
DBTO with Kow = 78, which was also determined by the shake-
flask method.
Optical melting studies
The ability of sulfoxides 1–3 to bind to DNA was studied by
conducting optical melting experiments using the sulfides 1S–3S
since they are not photoactive. Four different DNA duplexes were
used for the melt studies. Each duplex consisted of 12 base pairs
with constant stem sequences of GGTGXXXXGTGG/
CCACZZZZCACC on either side of a varied 4 base-pair sequence.
Each DNA duplex was melted in the presence and absence of the
sulfides to determine changes in the melting temperature (DTM).
The melt experiments were performed in melt buffer (1 M NaCl, 20
mM sodium cacodylate, 0.5 mM EDTA, pH 7.0). HPLC-purified
DNA oligonucleotides were obtained from Integrated DNA
Technologies (IDT, Coralville, IA) and used to prepare the DNA
duplexes. The 1 : 1 ratio of 50 mM sulfide and DNA duplex were
placed in quartz cuvettes (0.1 cm). The absorbance at 260 nm was
continuously monitored for the DNA duplex solution in the
absence and presence of the sulfide whilst increasing the
temperature from 25 uC to 95 uC at a rate of 0.5 uC min21 using
a Beckman Coulter DU 800 spectrophotometer equipped with a
Beckman-Coulter high performance temperature controller. Data
analysis was performed with MeltWin1 3.5 software, which was
used to calculate the first derivative of the DA260/DT vs.
temperature. The melting temperature, TM, value for each melting
isotherm was determined by finding the maximum of the first
derivative plot obtained by plotting DA260/DT vs. temperature.20
Quantum yield and photoproducts
Measurements of the quantum yields of photodeoxygenation
were performed in water at room temperature. The resulting
photoproducts for 1–3 are shown in Scheme 3. The photo-
deoxygenation of sulfoxides 2 and 3 to generate O(3P) were
investigated previously.17 Irradiation of 2 and 3 was reported
to produce other photoproducts in addition to the expected
deoxygenation photoproducts 2S and 3S, respectively. These
other photoproducts 4 and 5 for 2 and 6 and 7 for 3 were found
to be as a result of oxidation of the hydroxymethyl substituent
used to improve the aqueous solubility. Self-oxidation of the
side-chains of 2 and 3 necessarily reduced the amount of O(3P)
being released. The sulfonic acid-substituted precursor 1 was
expected to have improved solubility and be resistant to
oxidation by O(3P). As expected, the irradiation of 1 gave only a
single photoproduct 1S. The quantum yields for the consump-
tion of 1 and the formation of 1S were determined using
azoxybenzene as the actinometer.19 Quantum yields for the
photodeoxygenation reactions of 1–3 are shown in Table 1. The
difference between the quantum yields for deoxygenation of 1
and the formation of 1S were not significant, which was
Fluorescence anisotropy studies
To further investigate the ability of the sulfoxide compounds
1–3 to bind to DNA, fluorescence anisotropy measurements
were carried out using a Fluorolog 3 fluorometer (Horiba
Scientific, Edison, NJ). For each compound, a 100 mM solution
of the compound was made in 150 mL of 1X PBS buffer. This
solution was transferred into a cuvette and the anisotropy
value recorded. Additional solutions containing 100 mM of the
sulfoxide compound with different DNA concentrations (5, 10,
25, and 50 ng ml21) were prepared similarly and anisotropy
values measured. All anisotropy measurements were done at
an excitation wavelength of 290 nm and an emission
wavelength of 387 nm. The excitation bandpass was set to 19
and signals were taken at 9800 counts per second (cps).
Measurements were taken three times at each DNA concentra-
tion, with an integration time of 10 s, and the resulting
anisotropy values were averaged.
consistent with the expectation that
1 would be less
susceptible to oxidation. The significant difference between
the quantum yields of deoxygenation for sulfoxides 2 and 3
Results and discussion
Photoactivatable O(3P) precursors
The sulfoxides 2 and 3 were synthesized as reported in
literature.17 Sulfoxide 1 was synthesized by sulfonation of
Scheme 3
12392 | RSC Adv., 2013, 3, 12390–12397
This journal is ß The Royal Society of Chemistry 2013

View Article Online
RSC Advances
Paper
Table 1 Quantum yields for the photodeoxygenation of DBTO derivatives 1–3a
DNA photocleavage
The ability of 1–3 to photochemically cleave DNA was
investigated in aqueous media at room temperature. The
photoinduced DNA cleavage by DBTO derivatives was mon-
itored using gel electrophoresis by measuring the conversion
of supercoiled pUC19 plasmid DNA (Form I) to nicked circular
DNA (Form II), which occurs upon a single strand scission. In
these experiments, aqueous solutions containing supercoiled
plasmid DNA and the sulfoxide of interest were irradiated at
room temperature in a Luzchem LZC-4C photoreactor using
broadly emitting fluorescent bulbs centered at 350 nm.
Irradiations were performed under both aerobic and anaerobic
conditions. From a representative experiment, Fig. 1 depicts
the gels obtained for the irradiation of pUC19 plasmid DNA
and a sulfoxide (1–3) under aerobic conditions. The figure
clearly shows a substantial decrease in the band percent of
Form I from 90% to between 40–15% for all three DBTO
derivatives. Similar results were obtained under anaerobic
condition. Anaerobic conditions were achieved by degassing
the samples through argon sparging for 15 min. The decrease
in percent band for Form I confirmed the ability of 1–3 to
photoinduce DNA cleavage in aqueous solutions at low
concentrations. The extent of strand scission increases with
time and was consistent with other DNA cleavage agents.24,25
In control experiments, no significant increase in Form II was
observed when 1–3 were replaced by their corresponding
sulfides (1S–3S).
For all the experiments, a small percentage of nicked
circular DNA was present even prior to photolysis.
Additionally, a small (,10%) variable increase in the percent
of Form II was observed in control samples that were
irradiated but contained no DBTO derivatives. A means to
account for these variables was needed. Thus, the following
normalization procedure was adopted: the normalized percent
change of Form II (N) in control and sample was calculated by
dividing the change in band percent before and after
irradiation (%FormII+hn 2%FormII2hn and dividing by the
amount of Form I prior to irradiation (%FormI2hn) as shown
in eqn (1).
Sulfoxides
l (nm)
W2DBTO
W+DBT
1
1
254
294
330
330
330
0.009 ¡ 0.003b
0.018 ¡ 0.001
0.022 ¡ 0.009
0.075 ¡ 0.020
0.061 ¡ 0.001
0.004 ¡ 0.002
0.011 ¡ 0.005
0.019 ¡ 0.001
0.016 ¡ 0.003
0.016 ¡ 0.009
1
2c
3c
a
Quantum yields for the deoxygenation of the sulfoxide (2DBTO)
and the formation of the deoxygenated product (+DBT) determined
b
in aqueous solution. Error reported as 95% confidence interval.
c
Data taken from ref. 17.
compared with the quantum yields of formation of 2S and 3S
was consistent with a fraction of the oxidant being trapped by
the side chains.
The quantum yields for the DBTO derivatives in water were
higher than that reported for DBTO (W+DBT y 0.003) in organic
media.14,15 In organic solvents, the posited primary deox-
ygenation mechanism for DBTO is the unimolecular S–O bond
cleavage leading to the sulfide and O(3P) from the first singlet
excited state of DBTO.11,15,17,22,23 The increase in quantum
yield of photodeoxygenation in low or neutral pH water was
suggested to arise from the increased polarity of the solvent.17
The polar solvent was postulated to stabilize charge separation
during the cleavage of S–O bond, leading formally to an oxygen
atom radical anion and a DBT radical cation, prior to favorable
back electron transfer (BET) leading to the formation of O(3P)
and the sulfide. Thus, the significant increase in deoxygena-
tion quantum yields in aqueous media was suggested to result
from the water assisting in separating the charges formed
during bond cleavage compared to organic solvents. The
differences in deoxygenation quantum yields (W2DBTO
)
between 1–3 are consistent with the expected substituent
effects for the proposed mechanism. As shown in Scheme 4,
an increase of a positive charge on sulfur of the dibenzothio-
phene moiety would be expected to delocalize to the 2 and 8
positions. The electron withdrawing sulfonic acid substituent
would destabilize the dibenzothiophene cation radical and
concomitantly lower the quantum yield of deoxygenation.
Likewise, the hydroxymethyl substituents would be expected to
stabilize the positive charge, which is consistent with the
higher quantum yields for deoxygenation 2 and 3 compared to
1.
ꢀ
ꢁ
%Form IIzhn{%Form II{hn
N~
|100
(1)
%Form I{hn
DNA Cleavage = Nsample 2 Ncontrol
(2)
The extent of DNA cleavage that could be attributed to the
irradiation of 1–3 obtained from the difference between
sample (with 1–3) and control (no 1–3) using eqn (2). The
physical and photochemical properties of the 1–3 are
compared to the extent of DNA cleavage caused by the
irradiation of these DBTO derivatives is reported in Table 2.
The DNA cleavage values reported in Table 2 were obtained
from a minimum of five experiments represented by Fig. 1. For
1–3, the DNA cleavage values in aerobic and anaerobic
conditions ranged from 46% to 61% with 95% confidence
intervals of approximately ¡5%. Thus, significant differences
Scheme 4
This journal is ß The Royal Society of Chemistry 2013
RSC Adv., 2013, 3, 12390–12397 | 12393

View Article Online
Paper
RSC Advances
Fig. 1 Time-dependent conversion of supercoiled plasmid pUC19 DNA (Form I) to nicked circular DNA (Form II) in aerobic condition monitored with gel
electrophoresis in the presence of (a) compound 1 (b) compound 2 and (c) compound 3 at 0, 300, 600 and 900 s in aqueous media. The concentration of the
sulfoxides was kept at 0.1 mM and the plasmid DNA was 40.0 ng ml21. The gels were imaged and analyzed using ImageQuantTL from which the band percent of
Form I to Form II were obtained.
between the values of DNA cleavage photoinduced by 1–3 in
anaerobic and aerobic conditions were very small. While the
data weakly indicated 3 was the most effective at photoindu-
cing DNA cleavage, a more general conclusion that the extent
of photoinduced DNA cleavage was similar for all three DBTO
derivatives seems more prudent in the absence of more
experiments with other DNA examples.
Hydrophilicity, as gauged by Kow, increased significantly for
1 compared to 2 and 3. However, the extent of photoinduced
plasmid DNA cleavage by 1–3 derivatives remained the same.
Thus, no correlation between the hydrophilicity of the O(3P)-
precursors and the extent of photoinduced DNA cleavage was
observed. If all of the oxidant generated led to DNA cleavage,
was expected that some of the oxidant would be lost through
internal oxidation of the substituents. Thus, the quantum
yield for the formation of the sulfide (W+DBT) can be used to
represent the amount of escaping oxidant. Both the extent of
DNA cleavage and the quantum yield of sulfides formation
were similar for 1–3, which indicated only the escaped O(3P)
led to DNA cleavage. The quantum yields of sulfide formation
were higher under anaerobic compared to aerobic condition
for the three sulfoxides. Larger quantum yields in anaerobic
conditions were expected since molecular oxygen is an
efficient quencher of both singlet and triplet excited states.
However, it should be noted that no indication of singlet
oxygen generation was observed previously.17 Molecular
oxygen is known to enhance intersystem crossing, which
would inhibit deoxygenation, without singlet oxygen forma-
tion.15,26 Thus, the similarities in DNA cleavage under both
aerobic and anaerobic conditions might be accounted for by
the generation of ozone in aerobic condition. The generation
of ozone from O(3P) and O2 is a well-known process and ozone
was observed during the photodeoxygenation previously.17
To assess if the observed DNA cleavage was the result of
direct oxidation by O(3P), the photoinduced plasmid DNA
cleavage was performed in the presence of the glutathione
(GSH). Glutathione is a common ROS scavenging antioxidant,
then the quantum yield of photodeoxygenation (W2DBTO
)
would be expected to correlate with the extent of DNA
cleavage. However, the quantum yield for the photodeoxygena-
tion of the sulfoxides 2 and 3 in aerobic conditions (0.049 and
0.027, respectively) were higher than for 1 (0.011). Likewise,
significantly different quantum yields of photodeoxygenation
for 1, 2 and 3, (0.022, 0.075 and 0.061 respectively) were
observed in anaerobic conditions, and yet, similar amounts of
DNA cleavage were observed upon irradiation with 1–3. Thus,
no correlation between W2DBTO and the extent of photo-
induced DNA cleavage by DBTO was observed. For 2 and 3, it
Table 2 Summary of results of solubility, quantum yields and DNA photocleavage of sulfoxides 1–3 in aqueous mediaa
W2DBTO
W+DBT
DNA Cleavagec
Aerobic
b
b
Sulfoxide
Kow
Aerobic
Anaerobic
Aerobic
Anaerobic
Anaerobic
1
2
3
0.001
1.88
2.55
0.011 ¡ 0.006
0.049 ¡ 0.001
0.027 ¡ 0.006d
0.022 ¡ 0.009
0.075 ¡ 0.002
0.061 ¡ 0.001
0.009 ¡ 0.006
0.008 ¡ 0.003
0.009 ¡ 0.005
0.019 ¡ 0.001
0.016 ¡ 0.003
0.016 ¡ 0.009
54 ¡ 6
46 ¡ 3
58 ¡ 6
47 ¡ 7
58 ¡ 2
61 ¡ 6
a
b
The solubility of the sulfoxides was determined using octanol/water partition coefficient Kow
.
Quantum yields for the consumption of the
sulfoxide (2DBTO) and formation of the deoxygenation product (+DBT) were determined in aqueous solution in both aerobic and anaerobic
c
d
conditions at 330 nm. The DNA cleavage data was obtained by normalizing the test result with the control experiment. 95% confidence
interval for at least 5 experiments.
12394 | RSC Adv., 2013, 3, 12390–12397
This journal is ß The Royal Society of Chemistry 2013

View Article Online
RSC Advances
Paper
inhibition of photoinduced plasmid DNA cleavage by GSH for
1–3 under both aerobic and anaerobic conditions are shown. A
decrease in DNA cleavage was observed with increasing GSH
concentration suggesting an escaping oxidant from 1–3 was
involved in the photoinduced plasmid DNA cleavage. However,
at GSH concentrations higher than 1.2 mM, there was no
further significant inhibition of DNA cleavage, which indicated
other reaction mechanisms of 1–3 photolysis apart from that
involving O(3P) may also induce DNA strand scission. The
products of GSH oxidation may lead to DNA cleavage, however,
no DNA cleavage was observed in control experiments using
GSH alone in buffered solutions. Another possibility was that
noncovalent interaction between the DNA and dibenzothio-
phene S-oxide derivatives prevented GSH from intercepting the
oxidant. In order to determine whether
a noncovalent
interaction plays a role in photoinduced cleavage of DNA by
1–3, we directed our efforts to investigating the interactions
between the sulfoxides and DNA.
DNA binding studies
The photoinduced DNA cleavage by dibenzothiophene S-oxide
derivatives correlated with the generation of O(3P) as measured
by the formation of the sulfide (W+DBT). However, the inability
to completely inhibit DNA cleavage at high GSH concentration
was possibly due to noncovalent interaction between the DNA
and the DBTO derivatives, which would allow O(3P) to react
rapidly with the bound DNA thereby preventing scavenging by
GSH. Molecules that intercalate with DNA are often planar
aromatic heterocycles. Thus, the ability of 1–3 to intercalate
with DNA was investigated. Investigation of the intercalation
of the DBTO derivatives was performed using optical melt
studies. DNA melt studies provide insight into the extent a
compound helps stabilize a DNA duplex through non-covalent
binding. Specifically, DTM gives information as to how well the
compound stacks within the base pairs. A positive DTM value
for DNA duplex in the presence of a molecule indicates the
molecule helps stabilize the DNA duplex by binding noncova-
lently to the DNA duplex. In the optical melt experiment, the
corresponding sulfides (1S–3S) of 1–3 were used to prevent
interference from deoxygenation of the sulfoxide during the
melt studies. Four different DNA duplexes in melt buffer
solution were used for the melt studies. As shown in Table 3,
most of the DTM’s were within experimental error of zero,
indicating there was no intercalation between the DNA
duplexes and the sulfides analogs.
Fig. 2 Glutathione (GSH) inhibition of photoinduced plasmid pUC19 DNA
cleavage under both (a) anaerobic and (b) aerobic conditions is shown in bars:
sulfoxide 1 (dark grey), 2 (white) and 3 (light grey) was performed in 1X PBS
buffer to maintain a pH of 7.0 during irradiation. The concentration of
glutathione was increase from 0.1 mM to 1. 2 mM whilst keeping the
concentration of 1–3 and plasmid DNA constant at 0.1 mM and 40 ng ml21. The
anaerobic condition was obtained by bubbling Argon in the solution for 15 min
in the quartz test-tube. 95% confidence level for at least 5 experiments is
reported as the error.
and thiols were shown to be susceptible to oxidation by
O(3P).17,27,28 Thus, GSH was expected to act as a competitive
trap for O(3P), and thus, increasing concentrations of GSH
were expected to inhibit the photoinduced plasmid DNA
cleavage by DBTO derivatives if O(3P) was directly involved.
The photoinduced DNA cleavage experiment was performed in
the presence of GSH in 1X PBS buffer under both aerobic and
anaerobic conditions. The buffer was used to prevent DNA
cleavage which occurred upon irradiation if GSH was added
without a buffer. In aerobic conditions, the combination of O2
with O(3P) to form ozone was expected and under anaerobic
condition, the only oxidative species expected was O(3P).17 The
ratio of the DBTO derivative and GSH concentration was varied
from 1 : 1 to 1 : 12. In Fig. 2, bar graphs of the relative
Table 3 DTM of the sulfide-duplex equilibria
DNA 12-mera
1Sb
2Sb
3Sb
AAAA/TTTT
ATAT/TATA
GGGG/CCCC
GCGC/CGCG
20.87 ¡ 0.33
0.32 ¡ 0.46
20.27 ¡ 0.49
20.01 ¡ 2.36
20.15 ¡ 0.58
20.12 ¡ 1.41
20.13 ¡ 0.85
20.26 ¡ 2.17
20.07 ¡ 0.60
20.18 ¡ 1.28
0.26 ¡ 1.43
21.31 ¡ 2.63
a
Nucleotides (X)4/(Z)4 within (GGTG(X)4GTGG)/d(CCAC(Z)4(CACC)
b
overall 12-mer DNA duplex. The corresponding sulfide of the
sulfoxides were use to prevent photodeoxygenation during the melt
studies.
This journal is ß The Royal Society of Chemistry 2013
RSC Adv., 2013, 3, 12390–12397 | 12395

View Article Online
Paper
To further confirm that the three sulfoxides do not
intercalate with DNA, UV absorption and fluorescence titra-
tions were used to determine if DNA binding was possible.29,30
The spectra of the sample would be expected to change as the
amount of plasmid pUC19 increased if the sulfoxides were
intercalating. The data from both fluorescence and absor-
bance titration experiments of 1–3 with plasmid pUC19 are
RSC Advances
Aromatic sulfoxides are known to undergo a-cleavage of C–S
bond leading to biradicals.32 Studies by other researchers have
proposed the biradical as a means of DNA cleavage through
H-abstraction.33 It is unknown if the addition of GSH would
significantly inhibit DNA cleavage by these biradicals, and
therefore, it is possible to speculate that the biradical from the
a-cleavage of the DBTO derivatives 1, 2, and 3 may be involved
in the observed DNA cleavage.
reported in the ESI. The increasing concentration of plasmid
3
DNA in both the UV absorption and fluorescence titrations
resulted in no change in the spectra of the sample, again
indicating the sulfoxides do not intercalate into DNA.
Conclusions
While the optical melt and titrations experiments indicated
there was no intercalation of 1–3, other modes of noncovalent
binding were still possible. To rule out other noncovalent
interactions, fluorescence anisotropy was employed to study if
1–3 interact with the pUC19 plasmid. Fluorescence anisotropy
is based on the concept that when a molecule binds to another
molecule which is fluorescent, the binding event causes a
change in the anisotropy of the fluorescing molecule. The
technique has been used largely to study protein–DNA binding
interactions, in which case the DNA is usually labeled with a
fluorophore.31 Binding of the protein to the DNA causes an
increase in the size of the protein–DNA complex. This
subsequently leads to an increased rotational correlation time
and therefore anisotropy of the complex compared to that of
the free or unbound DNA. In this study, binding of the
sulfoxides to the DNA is monitored by change in anisotropy of
the sulfoxide compounds. As shown in Fig. 3, there is no
significant change in the anisotropy of the sulfoxides 1, 2, and
3 with increasing concentration of DNA, suggesting that they
do not bind. This ruled out the possibility of a noncovalent
interaction preventing the inhibition of DNA cleavage by GSH.
This suggests another mechanism in addition to direct
cleavage of DNA by O(3P) must account for the some of the
observed cleavage in the presence of a ROS scavenger.
In summary, we find that the aqueous soluble DBTO
derivatives effectively cleave DNA photochemically in aqueous
media. The quantum yields for photodeoxygenation of the
these sulfoxides are higher compared to DBTO in organic
solvents. The extent of DNA photocleavage correlates well with
the quantum yields for the formation of the corresponding
sulfides, which indicates DNA cleavage was induced by an
escaped oxidant from the photodeoxygenation of DBTO
derivatives. The inhibition of DNA photocleavage in the
presence of glutathione indicated that atomic oxygen O(3P)
plays a role in the DNA cleavage abilities of 1–3. However,
another mechanism in addition to the release of O(3P) must
account for some of the cleavage since high concentrations of
glutathione could not completely inhibit DNA cleavage. The
three aqueous soluble dibenzothiophene-S-oxides do not bind
to DNA as confirmed by UV melt and fluorescence anisotropy
studies.
Acknowledgements
The authors thank Dr. Brent Znosko and the Znosko research
group for assistance with the optical melt studies and Dr.
Tomasz Heyduk for assistance with the fluoresence anisotropy
studies. Portions of this material is based upon work
supported by the National Science Foundation under CHE-
1255270. The authors thank the National Science Foundation
for support under grant CHE-0963363 for renovations to the
research laboratories in Monsanto Hall. This work was
additionally supported by donors to the Herman Frasch
Foundation.
References
1 K. Bedard and K. H. Krause, Physiol. Rev., 2007, 87,
245–313.
2 A. L. Garner, C. M. St. Croix, B. R. Pitt, G. D. Leikauf,
S. Ando and K. Koide, Nat. Chem., 2009, 1, 316–321.
3 K. M. Kadish, Z. P. Ou, J. G. Shao, C. P. Gros, J. M. Barbe,
F. Jerome, F. Bolze, F. Burdet and R. Guilard, Inorg. Chem.,
2002, 41, 3990–4005.
4 S. V. Kovalenko and I. V. Alabugin, Chem. Commun., 2005,
1444–1446.
5 J. P. Celli, B. Q. Spring, I. Rizvi, C. L. Evans, K. S. Samkoe,
S. Verma, B. W. Pogue and T. Hasan, Chem. Rev., 2010, 110,
2795–2838.
Fig. 3 Fluorescence anisotropy studies of sulfoxides interaction with pUC19
plasmid DNA. Anisotropy was monitored at excitation wavelength of 290 nm
and emission wavelength of 387 nm for all three compounds. Measurements
were taken three times at each DNA concentration, with an integration time of
10 s, and the resulting anisotropy values were averaged. Standard deviations
are approximately ¡0.004 (or less) anisotropy unit.
12396 | RSC Adv., 2013, 3, 12390–12397
This journal is ß The Royal Society of Chemistry 2013

View Article Online
RSC Advances
Paper
6 Y. Huang, Y. Zhang, J. Zhang, D. W. Zhang, Q. S. Lu, J.
L. Liu, S. Y. Chen, H. H. Lin and X. Q. Yu, Org. Biomol.
Chem., 2009, 7, 2278–2285.
7 B. Breiner, K. Kaya, S. Roy, W. Y. Yang and I. V. Alabugin,
Org. Biomol. Chem., 2012, 10, 3974–3987.
8 J. S. Daniels and K. S. Gates, J. Am. Chem. Soc., 1996, 118,
3380–3385.
9 J. S. Daniels, T. Chatterji, L. R. MacGillivray and K. S. Gates,
J. Org. Chem., 1998, 63, 10027–10030.
10 A. B. Thomas and A. Greer, J. Org. Chem., 2003, 68,
1886–1891.
11 M. Nag and W. S. Jenks, J. Org. Chem., 2004, 69, 8177–8182.
12 R. D. McCulla and W. S. Jenks, J. Am. Chem. Soc., 2004, 126,
16058–16065.
13 J. C. Scaiano, G. Bucher, M. Barra, D. Weldon and R. Sinta,
J. Photochem. Photobiol., A, 1996, 102, 7–11.
14 Z. H. Wan and W. S. Jenks, J. Am. Chem. Soc., 1995, 117,
2667–2668.
15 D. D. Gregory, Z. H. Wan and W. S. Jenks, J. Am. Chem. Soc.,
1997, 119, 94–102.
16 O. R. Wauchope, S. Shakya, N. Sawwan, J. F. Liebman and
A. Greer, J. Sulfur Chem., 2007, 28, 11–16.
17 J. Korang, W. R. Grither and R. D. McCulla, J. Am. Chem.
Soc., 2010, 132, 4466–4476.
19 S. P. Vaish, J. LaMarre and N. J. Bunce, Photochem. and
Photobiol., 1984, 39, 531.
20 P. L. Vanegas, T. S. Horwitz and B. M. Znosko, Biochemistry,
2012, 51, 2192–2198.
21 A. Leo, C. Hnasch and D. Elkins, Chem. Rev., 1971, 71,
525–616.
22 K. Kumazoe, K. Arima, S. Mataka, D. J. Walton and
T. Thiemann, J. Chem. Res., 2003, 60–61.
23 E. Lucien and A. Greer, J. Org. Chem., 2001, 66, 4576–4579.
24 M. M. Gonzalez, M. Pellon-Maison, M. A. Ales-Gandolfo, M.
R. Gonzalez-Baro, R. Erra-Balsells and F. M. Cabrerizo, Org.
Biomol. Chem., 2010, 8, 2543–2552.
25 C. A. Terry, L. Gude, M. J. Fernandez, A. Lorente and K.
B. Grant, Bioorg. Med. Chem. Lett., 2011, 21, 1047–1051.
26 C. S. Parmenter and J. D. Rau, J. Chem. Phys., 1969, 51,
2242–2246.
27 K. Bahrami, Tetrahedron Lett., 2006, 47, 2009–2012.
28 J. F. Turrens, J. Physiol., 2003, 552, 335–344.
29 M. Aslanoglu and N. Oge, Turk. J. Chem., 2005, 29, 477–485.
30 B. Rajendar, Y. Sato, S. Nishizawa and N. Teramae, Bioorg.
Med. Chem. Lett., 2007, 17, 3682–3685.
31 T. Heyduk and J. C. Lee, Proc. Natl. Acad. Sci. U. S. A., 1990,
87, 1744–1748.
32 Y. Guo and W. S. Jenks, J. Org. Chem., 1997, 62, 857–864.
33 A. G. Myers, S. B. Cohen and B. M. Kwon, J. Am. Chem. Soc.,
1994, 116, 1670–1682.
18 R. Kowalewski and P. Margaretha, Helv. Chim. Acta, 1993,
76, 1251–1257.
This journal is ß The Royal Society of Chemistry 2013
RSC Adv., 2013, 3, 12390–12397 | 12397