DNA-polyamine cross-links generated upon one electron oxidation of DNA

Article abstract of DOI:10.1021/tx500063d

The possibility to induce the formation of covalent cross-links between polyamines and guanine following one electron oxidation of double stranded DNA has been evaluated. For such a purpose, a strategy has been developed to chemically synthesize the polyamine-C8-guanine adducts, and efforts have been made to characterize them. Then, an analytical method, based on HPLC separation coupled through electrospray ionization to tandem mass spectrometry, has been setup for their detection and quantification. Using such a sensitive approach, we have demonstrated that polyamine-C8-guanine adducts could be produced with significant yields in double stranded DNA following a one-electron oxidation reaction induced by photosensitization. These adducts, involving either putrescine, spermine, or spermidine, are generated by the nucleophilic addition of primary amino groups of polyamines onto the C8 position of the guanine radical cation. Our data demonstrate that such a nucleophilic addition of polyamines is much more efficient than the addition of a water molecule that leads to 8-oxo-7,8-dihydroguanine formation.

Full text of DOI:10.1021/tx500063d

Article
pubs.acs.org/crt
DNA-Polyamine Cross-Links Generated upon One Electron Oxidation
of DNA
Stephanie Silerme, Laure Bobyk, Marisa Taverna-Porro,† Camille Cuier, Christine Saint-Pierre,
́
and Jean-Luc Ravanat*
Laboratoire Les
17 rue des Martyrs, 38000 Grenoble ced
́
ions des Acides Nuclei
́
ques, Universite
ex 9, France
́
Grenoble Alpes, INAC-SCIB, F-38000 Grenoble, France, CEA, INAC-SCIB,
́
S
* Supporting Information
ABSTRACT: The possibility to induce the formation of covalent
cross-links between polyamines and guanine following one
electron oxidation of double stranded DNA has been evaluated.
For such a purpose, a strategy has been developed to chemically
synthesize the polyamine-C8-guanine adducts, and efforts have
been made to characterize them. Then, an analytical method,
based on HPLC separation coupled through electrospray
ionization to tandem mass spectrometry, has been setup for
their detection and quantification. Using such a sensitive
approach, we have demonstrated that polyamine-C8-guanine
adducts could be produced with significant yields in double
stranded DNA following a one-electron oxidation reaction
induced by photosensitization. These adducts, involving either putrescine, spermine, or spermidine, are generated by the
nucleophilic addition of primary amino groups of polyamines onto the C8 position of the guanine radical cation. Our data
demonstrate that such a nucleophilic addition of polyamines is much more efficient than the addition of a water molecule that
leads to 8-oxo-7,8-dihydroguanine formation.
8-oxo-7,8-dihydroguanine, through the addition of a water
molecule onto C8 of guanine and subsequent oxidation of the
latter generated radical.8 This hydration reaction could also be
followed by a competitive reduction, with an opening of the
imidazole ring, producing 2,6-diamino-4-hydroxy-5-formamido-
pyrimidine.6 Other guanine decomposition products have been
identified and are certainly generated through an initial
nucleophilic addition at C8, involving either C5′-OH group,9
C5′-amino residues,10−12 or even methanol as recently
reviewed.13
Such a nucleophilic addition of amino or alcoholic residues
onto C8 of guanine is of particular relevance in the issue of
DNA−protein cross-links. In 2006, Perrier et al. described the
formation of a covalent cross-link between a lysine moiety and
2′-deoxyguanosine (dGuo), after photosensitization of a TGT
oligonucleotide in the presence of a trilysine peptide.14 The
cross-link involves the nucleophilic addition of the ε amino
group of the central amino acid residue of the peptide onto the
C8 position of deprotonated guanine radical cation G(-H)•.
Such a reaction has also been implicated in the formation of
intrastrand cross-links through the addition of a neighboring
thymine to G(-H)•.15,16 According to these observations, we
hypothesized that other nucleophilic molecules would be able
INTRODUCTION
■
The DNA molecule is constantly exposed to various kinds of
genotoxic agents, including among others reactive oxygen
species (ROS) produced during oxidative stress.1 It is well
established that normal cellular metabolism is a source of ROS,
which may interact with DNA, leading to chemical
modifications including single and double strand breaks or
base and sugar lesions.2 These oxidative damages could also be
generated after exposure to ionizing radiation, UV light, or to
different exogenous chemical agents and are known to be
involved in mutagenesis and carcinogenesis.2,3
For instance, photosensitization reactions are known to be
able to induce oxidative DNA lesions through either the
transient formation of singlet oxygen (Type II photosensitiza-
tion reaction) or following a one electron oxidation reaction
(Type I).4 The latter mechanism is one of the main processes
involved in the formation of oxidized base lesions.3 Guanine is
the preferential target of this reaction since it exhibits the
lowest ionization potential among DNA components.5 Various
agents and free radicals are able to abstract one electron from
guanine,6 leading to the formation of a guanine radical cation
G•+. The latter could also be produced following an electron
transfer reaction from guanine to a radical cation initially
generated on other DNA bases.7
Several nucleophilic addition reactions on the C8 position of
the guanine radical cation have been already reported. The first
described mechanism involves the formation of the well-known
Received: February 24, 2014
Published: May 5, 2014
© 2014 American Chemical Society
1011
dx.doi.org/10.1021/tx500063d | Chem. Res. Toxicol. 2014, 27, 1011−1018

Chemical Research in Toxicology
Article
Thermo Scientific equipped with a Thermo Accela Pump, an Accela
autosampler, and an Accela photodiode array detector. The data were
processed using Xcalibur software. HPLC purification was performed
using an Agilent 1200 Series LC system. 1H NMR spectra were
recorded on Bruker Avance DMX 200 and Bruker Avance 500
spectrometers at 295 K in 400 μL of either CD3OD (isotope purity
>99.8%) or D2O (isotope purity: 99.95%).
Chemical Synthesis. 8-Bromo-2′-deoxyguanosine (8-Br-dGuo,
1). dGuo (4 g, 15 mmol) was suspended in a mixture of water (40 mL)
and acetonitrile (160 mL). N-Bromo-succinimide (4g, 22.5 mmol) was
added at 0 °C in three portions. The reaction mixture was stirred for
30 min at room temperature and subsequently concentrated under
reduced pressure. This material was then dissolved in acetone (80 mL)
and stirred for 2 additional hours at room temperature. The resulting
to add to C8 of guanine upon one electron oxidation of double
stranded (ds) DNA.
Thus, in the present work, attempts have been made to
determine if small polyamine molecules could also be involved
in such a reaction. Polyamines are organic cations localized in
the nucleus of all eukaryotic cells17 at millimolar concentration
ranges. They are biosynthesized by enzymatic decarboxylation
of the amino acids ornithine and arginine by ornithine
decarboxylase (ODC).18 It has been proposed that due to
their affinity for anionic sites of nucleic acids, the polyamines’
main role is to stabilize and condense DNA.19 They are also
involved in several essential metabolic processes, such as
regulation of RNA transcription, specific gene expression, and
the progression of the cell cycle.18 They play a role in cell
growth and cancer, and an increase in their cellular
concentration correlates with a loss of cell proliferation control
and could consequently lead to neoplesia.18 Furthermore,
polyamines are present in a high concentration in several
biological fluids in patients suffering from different types of
cancer.20 The assessment of their levels in urine, blood, or
cerebrospinal liquid can be informative as a marker of tumor
growth in treated patients. It has been used for the prediction of
relapse as well as in the follow-up of chemotherapy
efficiency.21,22 For these reasons, polyamines are a potential
target in cancer treatment, and various inhibitors of their
biosynthesis have been developed. The most studied is α-
difluoromethylornithine, which is an irreversible inhibitor of
ODC. Polyamine oxidase, an enzyme involved in the
catabolism of polyamines, is also considered as an interesting
outlook in the development of antineoplasic drugs.23 Finally,
recent studies have highlighted the interest of using a low
polyamine diet in addition to chemotherapy.24 Therefore, the
link between polyamines and cancer has been widely described.
Nevertheless, the exact mechanism in which the upregulation of
their metabolism could be implicated in carcinogenesis is still
unknown.
precipitate was collected by vacuum filtration with a Buchner funnel.
̈
The filtrate was rinsed twice with cold acetone and dried under
vacuum to provide 3.85 g (11.1 mmol, 74%) of a slightly pink powder.
This compound was used for the next reaction steps without further
purification. 1H NMR (D2O, 200 MHz, δ): 2.13 (ddd, 1H, J = 2.3, 6.5,
13.6 Hz, H2′′), 3.07 (ddd, 1H, J = 6.3, 8.1, 13.6 Hz, H2′), 3.64 (dd,
1H, J = 4.2, 12.0 Hz, H5′′), 3.75 (dd, 1H, J = 3.4, 12.0 Hz, H5′), 3.89−
3.95 (m, 1H, H4′), 4.48−4.54 (m, 1H, H3′), 6.30 (dd, 1H, J = 6.5, 8.1
Hz, H1′).
8-((4-Aminobutyl)amino)-2′-deoxyguanosine (8-put-dGuo, 2). 8-
Br-dGuo (1) (69 mg, 0,2 mmol) and 1,4 diaminobutane (100 μL, 1
mmol) were dissolved in DMSO (1.2 mL). Triethylamine (135 μL, 1
mmol) was added, and the reaction mixture was heated for 15 h at 110
°C. According to HPLC-UV monitoring, the substitution reaction
afforded 25% of 8-put-dGuo. The crude material was diluted 10 times
in water, neutralized with HCl, and then purified by C18 reverse phase
HPLC. The elution was performed using a C18 column (Interchim
uptisphere 5 μm ODB 250·4.6 mm), at a flow rate of 1 mL/min, with
0.1% trifluoroacetic acid as the mobile phase (eluent A) and a linear
gradient of acetonitrile (eluent B, 0−14% over 21 min). The product
which eluted at 15 min (k′ = 3.41) was collected, concentrated, and
lyophilized. 1H NMR (D2O, 500 MHz, δ): 1.70−1.78 (m, 4H,
HputB,C), 2.26 (ddd, 1H, J = 1.8, 6.6, 14.5 Hz, H2′′), 2.77 (ddd, 1H, J =
6.6, 8.4, 14.5 Hz, H2′), 3.02−3.08 (m, 2H, H putD), 3.34−3.42 (m, 2H,
HputA), 3.88 (dd, 1H, J = 2.3, 12.4 Hz, H5′′), 3.92 (dd, 1H, J = 2.1,
12.4 Hz, H5′), 4.12−4.14 (m, 1H, H4′), 4.63−4.65 (m, 1H, H3′), 6.31
+
The aim of our study is to investigate the possible formation
of adducts between polyamines and 2′-deoxyguanosine in
dsDNA following a one electron oxidation reaction. For such a
purpose, we have designed a chemical synthesis of the
potentially generated DNA adducts involving either putrescine,
spermine, or spermidine. In addition, we have optimized a
HPLC-MS/MS method to quantify these adducts in DNA
subsequent to enzymatic digestion into corresponding free
nucleosides. Using such a strategy, we have demonstrated that
polyamine-C8-guanine adducts could be generated following a
type I photosensitization reaction involving riboflavin and UVA,
a well-known approach to generate guanine radical cations25 in
dsDNA.
(dd, 1H, J = 6.6, 8.4 Hz, H1′). MALDI-HRMS m/z for C14H24N7O4
([M + H]+): calcd 354.1890; found 354.1883.
8-((3-((4-Aminobutyl)amino)propyl)amino)-2′-deoxyguanosine
and 8-((4-((3-aminopropyl)amino)butyl)amino)-2′-deoxyguano-
sine (8-spd-dGuo 3a and 3b). The synthesis was performed following
the same procedure as for compound 2, using 1 equiv of 8-Br-dGuo, 5
equiv of spermidine, and 5 equiv of triethylamine. The crude material
was purified by HPLC, as described for compound 2, with a gradient
of eluent B from 0 to 12% over 20 min. The two products 3a and 3b
which eluted at 14.3 min (k′ = 3.21) and 15.7 min (k′ = 3.62),
respectively, were collected. The two isomers exhibit identical 1H
NMR spectral features: 1H NMR (D2O, 500 MHz, δ): 1.76−1.80 (m,
3H, Hspd), 2.06−2.12 (m, 3H, Hspd), 2.34 (ddd, 1H, J = 2.2, 6.1, 14.3
Hz, H2′′), 2.74 (ddd, 1H, J = 6.5, 9.1, 14.3 Hz, H2′), 3.09−3.12 (m,
4H, Hspd), 3.14−3.18 (m, 2H, HspdG), 3.44−3.47 (m, 2H, HspdA), 3.88
(dd, 1H, J = 2.8, 12.4 Hz, H5′′), 3.94 (dd, 1H, J = 2.1, 12.4 Hz, H5′),
4.19−4.20 (m, 1H, H4′), 4.63−4.66 (m, 1H, H3′), 6.42 (dd, 1H, J =
6.1, 9.1 Hz, H1′). MALDI-HRMS m/z for C17H31N8O4+ ([M + H]+):
calcd 411.2468; found 411.2534.
EXPERIMENTAL PROCEDURES
■
Chemicals. N-Bromosuccinimide, 1,4-diaminobutane (putrescine),
spermidine, spermine, triethylamine, and solvents for organic synthesis
were purchased from Sigma-Aldrich (St. Louis, MO). 2′-Deoxygua-
nosine (dGuo) and 2′-deoxythymidine were obtained from Pharma
8-((3-((4-((3-Aminopropyl)amino)butyl)amino)propyl)amino)-2′-
deoxyguanosine (8-spm-dGuo, 4). The synthesis was performed
following the same procedure as that for compound 2, using 1 equiv of
8-Br-dGuo, 5 equiv of spermine, and 5 equiv of triethylamine. The
crude material was purified by HPLC, as described for compound 2,
with a gradient of eluent B from 0 to 14% over 21 min. The product
Waldhof (Dusseldorf, Germany), HPLC grade acetonitrile was
̈
purchased from VWR BDH Prolabo (Fontenay-sous-Bois, France)
and trifluoroacetic acid and heptafluorobutyric acid from Sigma-
Aldrich. Enzymes for DNA digestion (nuclease P1, phospodiesterase I
and II, and alkaline phosphatase) and calf thymus DNA were
purchased from Sigma-Aldrich. NMR solvents were obtained from
Euriso-top (Gif-sur-Yvette, France).
1
which eluted at 13.8 min (k′ = 3.06) was collected. H NMR (D2O,
500 MHz, δ): 1.36−1.39 (m, 4H, HspmE,F), 1.64−1.72 (m, 4H, HspmB,I),
1.94 (ddd, 1H, J = 2.0, 6.2, 14.3 Hz, H2′′), 2.35 (ddd, 4H, J = 6.3, 8.6,
14.3 Hz, H2′), 2.68−2.72 (m, 8H, HspmC,D,G,H), 2.74−2.77 (m, 2H,
HspmJ), 3.09−3.13 (m, 2H, HspmA), 3.48 (dd, J = 2.6, 12.4 Hz, H5′′),
Instrumentation. The HPLC-MS/MS analyses were performed
with a triple quadrupole mass spectrometer TSQ Quantum Ultra from
1012
dx.doi.org/10.1021/tx500063d | Chem. Res. Toxicol. 2014, 27, 1011−1018

Chemical Research in Toxicology
Article
3.53 (dd, J = 2.4, 12.4 Hz, H5′), 3.77−3.79 (m, 1H, H4′), 4.23−4.25
(m, 1H, H3′), 5.99 (dd, J = 6.2, 8.6 Hz, H1′). MALDI-HRMS m/z for
The synthesis consists first in the bromination of dGuo onto
the C8 position using N-bromosuccinimide,29 followed by the
nucleophilic substitution of generated 8-Br-dGuo by the three
polyamines under alkaline conditions as described previously
for aromatic amines.30 According to HPLC analysis, the
estimated yield for the reaction was about 25%, whatever the
nature of the polyamine.
+
C20H38N9O4 ([M + H]+): calcd 468.3047; found 468.3094.
UV Characterization of Polyamine-dGuo Adducts. The
concentration of the polyamine solutions was determined by
1
quantitative H NMR and used for the determination of the molar
extinction coefficients of the different adducts. For the 1H NMR
quantification, the relaxation time t1 of the anomeric proton of C8
polyamine adducts was measured between 2 and 3 s. Thus, acquisition
was performed during 512 scans in D2O; the relaxation delay was set at
12 s. The concentration was calculated by comparing the integrals of
the anomeric protons of the polyamine adducts to those of 2′-
deoxythymidine in a calibrated solution (obtained by UV 3,0068 mM,
ε = 9318 L·mol−1·cm−1). A capillary containing a reference solution of
3-(trimethylsilyl)propionic acid sodium (Merck, Darmstadt, Germany)
was inserted into each NMR tube in order to normalize integrations.
The molar extinction coefficient was determined by measuring UV
absorbance of the calibrated polyamine-dGuo adducts and applying
the Beer−Lambert law.
The three adducts (8-put-dGuo, 8-spd-dGuo, and 8-spm-
dGuo, Figure 1) were purified using C18 reverse phase HPLC.
UVA Irradiation and Digestion of Isolated DNA Samples. An
aqueous solution of calf thymus DNA (0.1 mg/mL) containing NaCl
(0.2 M), riboflavin (40 μM), and 500 μM of each of the three
polyamines was irradiated at room temperature with a 500 W halogen
lamp placed at a distance of 20 cm. The solution was kept saturated
with oxygen by continuous air bubbling and maintained at room
temperature by circulating water. Samples of 50 μL (5 μg of DNA)
were collected at 0, 1, 2, 5, and 10 min. DNA was precipitated by
adding 125 μL of 100% cold ethanol and 10 μL of NaCl (4 M), the
DNA pellet was rinsed 3 times with 70% ethanol, dissolved in 50 μL of
water, and digested as follows:26 to each sample of irradiated DNA
(0.1 μg/μL, 50 μL), 0.025 U of phosphodiesterase II, 2.5 U of DNase
II, 0.5 U of nuclease P1 in 2.5 μL of buffer P1 (300 mM ammonium
acetate and 1 mM ZnSO4, pH 5.3), and 2.5 μL of buffer MNSPDE
(200 mM succinic acid, 100 mM CaCl2, pH 6) were added. The
samples were incubated for 2 h at 37 °C. Then, 0.015 U of
phosphodiesterase I was added, together with 6 μL of buffer Palk 10×
(500 mM Tris and 1 mM EDTA, pH 8) and 2 U of alkaline
phosphatase. The given solutions were incubated again for 2 h at 37
°C. Then, 3.5 μL of hydrochloric acid 0.1 N was added. The resulting
nucleoside mixture was then analyzed by HPLC-MS/MS.
HPLC-MS/MS Analysis. The separation of hydrolyzed nucleosides
was achieved using a ODB uptisphere 3 μm 150 × 2.1 mm column
from Interchim under linear gradient conditions as previously reported
for measuring other DNA lesions.27 The flow rate was set at 0.2 mL/
min with 0.1% heptafluorobutyric acid in water as the mobile phase
(eluent A) and acetonitrile from 0 to 45% (eluent B) over 30 min.
Unmodified nucleosides were detected using a UV detector set at 280
nm. The multiple reactions monitoring mode (MRM) was used in
order to detect two specific transitions for each adduct: 354.3 → 221.2
and 354.3 → 238.5 for 8-put-dGuo; 468.0 → 207.2 and 468.0 → 352.4
for 8-spm-dGuo; 411.5 → 207.2 and 411.5 → 221.2 for 8-spd-dGuo.
Collision energies and ionization parameters have been optimized by
infusing standard solutions of each adduct as reported for 8-
oxodGuo.28 For each series of samples, calibrated solutions of a
mixture of normal nucleosides, 8-oxodGuo, or polyamine-dGuo
adducts were first injected to check the sensitivity of the MS/MS
system. Then, after an analysis of a maximum 12 samples, these
standards were reinjected to make sure that no variation in the
sensitivity of the detector occurred with time. The levels of normal
nucleosides and lesions were calculated by external calibration for both
UV and MS/MS detection, respectively, and results were expressed as
the number of DNA lesions (8-oxodGuo or polyamine-dGuo adducts)
per million normal nucleosides.
Figure 1. Chemical structures of C8 polyamine-dGuo adducts.
Interestingly, the substitution of 8-Br-dGuo by spermidine led
to two possible isomers of 8-spd-dGuo (3a and 3b) with similar
yields. The two isomers were isolated by HPLC purification
and identified according to their fragmentation pattern (vide
infra). Solutions of the three adducts were calibrated by NMR
and were used as external standards for the HPLC-MS/MS
determination of the amounts of adducts generated in dsDNA
upon one electron mediated oxidation.
Characterization of Polyamine-dGuo Adducts. Struc-
tures of the three synthesized adducts have first been confirmed
by 1H NMR. Besides hydrogen atoms of 2′-deoxyribose, on 1H
NMR spectra one can observe additional signals attributed to
methylene protons of the polyamine moiety. The lack of the
signal corresponding to the H8 proton of the guanine base
confirmed that the polyamine residues are bound to the C8
atom of guanine. Furthermore, the hydrogen atoms in the α
position of the NH-guanine bond and those at the end of the
polyamine chain are not equivalent confirming that one
primary amino group is involved in the cross-link with C8 of
guanine.
UV spectra exhibit a maximum absorption at around 260 nm
(Figure S1, Supporting Information) and a shoulder at around
300 nm. A bathochromic shift is observed for the adducts
compared to dGuo as was already observed for 8-lysine-2′-
deoxyguanosine14 and for 8-oxodGuo. Molar extinction
coefficients, determined by NMR quantification, have been
found to be higher than that of the nonsubstituted 2′-
deoxyguanosine and to slightly depend on the length of the
polyamine chain, as shown in Table 1.
Mass spectra of the adducts in positive ionization mode
(Figure 2) exhibit intense molecular ions at m/z 354, 411, and
468 for 8-put-dGuo, 8-spd-dGuo, and 8-spm-dGuo, respec-
tively. These ions correspond to the protonated molecules [M
+ H+] (molecular weight = 353, 410, and 467, respectively).
RESULTS
■
Synthesis of C8 Polyamine-2′-deoxyguanosine Ad-
ducts. Adducts between dGuo and polyamines (putrescine,
spermidine, and spermine) have been prepared according to
the strategy described in Scheme S1 (Supporting Information).
1013
dx.doi.org/10.1021/tx500063d | Chem. Res. Toxicol. 2014, 27, 1011−1018

Chemical Research in Toxicology
Article
Fragmentation spectra obtained following collision induced
dissociation of the pseudomolecular ion exhibit intense ions
resulting from the loss of the sugar moiety (m/z 238, 295, and
352) as classically observed for 2′-deoxyribonucleosides.31 In
addition, other ions resulting from fragmentations occurring
along the polyamine chain in the α position of amino groups
(m/z 221 for 8-put-dGuo; 278 and 207 for 8-spd-dGuo 3a; 278
and 221 for 8-spd-dGuo 3b; and 278, 224, and 207 for 8-spm-
dGuo) are also observed. Interestingly, the two adducts 3a and
3b isolated after the substitution of 8-Br-dGuo with spermidine
exhibit a different fragmentation pattern of the spermidine
moiety, which is explained by the lack of symmetry of the
polyamine residue, that, following addition of one or the other
amino residue, gives rise to two distinct isomers.
Table 1. UV Spectral Features of 8-Polyamine-dGuo Adducts
compd name
λmax (nm)
ε (λmax) (M−1·cm−1
)
8-put-dGuo (2)
260
258
258
16 930
19 590
20 820
8-spd-dGuo (3a, 3b)
8-spm-dGuo (4)
Development of an HPLC-MS/MS Method of Quanti-
fication. An analytical method to detect the three adducts was
then developed, using HPLC coupled through electrospray
ionization to tandem mass spectrometry. We used the so-called
MRM mode in order to have a sensitive, specific, and
quantitative detection method. Optimization of collision
energies was performed for specific fragmentations of each
adduct, as reported in Table 2. For the simultaneous
Table 2. Main Transitions Used for the Detection of the
Different 8-Polyamine-dGuo Adducts in the Multiple
Reaction Monitoring (MRM) Mode (Positive Ionization)
main transitions
relative
molecular precursor
fragment
intensity
(%)
adduct
weight
ion (m/z) ions (m/z)
8-put-dGuo (2)
353.4
354.3
354.3
411.5
411.5
411.5
221.2
100
6
238.5
8-spd-dGuo (3a, 3b)
8-sp-dGuo (4)
410.5
467.6
207.2 (3a)
221.3 (3b)
50
88
100
295.2 (3a
and 3b)
468.0
468.0
207.2
352.4
100
23
measurement of the three adducts we have defined common
ionization parameters and optimized HPLC conditions for their
separation. Strong acids, such as trifluoroacetic acid (TFA) or
heptafluorobutyric acid (HBFA), were tested as the mobile
phase in order to improve the chromatographic behavior of
these molecules which contain primary and secondary amino
groups.
To improve the specificity of detection, two main transitions
were monitored simultaneously for each lesion (Figure 3),
whereas normal nucleosides were quantified on line by UV
detection set at 280 nm. The quantification was performed by
considering the most intense transition for each adducts, and
results were normalized compared to the amount of
unmodified nucleosides. The second transition served as a
qualifier (Table 2). For 8-spd-dGuo, the relative amounts of the
two isomers were similar, and thus, the reported amounts of
generated 8-spd-dGuo adducts were the sum of the two 3a and
3b isomers. The limit of detection of the assay was close to 1
pmol injected for each adduct. In parallel, 8-oxodGuo was also
measured using the 284 → 168 MRM transition as already
reported.28
Figure 2. Collision-induced dissociation mass spectra of the [M + H]+
pseudomolecular ion of the different 8-polyamine-dGuo adducts
(positive ionization mode). A, 8-put-dGuo (2); B, 8−spd-dGuo 3a; C,
8-spd-dGuo 3b; and D, 8-spm-dGuo (4).
8-Polyamine-dGuo Adduct Formation in Isolated
DNA. In order to investigate the formation of 8-polyamine-
2′-dGuo adducts, calf thymus dsDNA was exposed to UVA
1014
dx.doi.org/10.1021/tx500063d | Chem. Res. Toxicol. 2014, 27, 1011−1018

Chemical Research in Toxicology
Article
performed in the presence of the different polyamines at a
concentration of 500 μM, which is below the concentration
found in the nucleus of eukaryotic cells. Kinetics of formation
of the different adducts are presented in Figure 5, for 8-put-
Figure 3. Typical HPLC chromatograms obtained for detection by
HPLC-MS/MS of 8-polyamine-dGuo adducts. Chromatograms
represent the analysis of a mixture of the 4 polyamine dGuo adducts,
corresponding to the injection of 1 pmol of each adduct.
light in the presence of riboflavin. Such a photosensitization
reaction is known to produce a high amount of 8-oxodGuo
through the transient formation of a guanine radical cation.8
Thus, as expected under these conditions an increase in the
level of 8-oxodGuo was observed when increasing the
irradiation time. Levels as high as 5000 8-oxodGuo per million
nucleosides were obtained following 10 min of irradiation
(Figure 4). This photosensitization reaction was then
Figure 5. Kinetics of formation of 8-put-dGuo, 8-spd-dGuo, and 8-
spm-dGuo in dsDNA exposed to riboflavin and UVA-mediated one
electron oxidation in the presence of the corresponding polyamine at
500 μM. Data represent the average and standard deviation of at least
three independent determinations and are expressed as the number of
lesions per million normal nucleosides.
dGuo, 8-spd-dGuo, and 8-spm-dGuo. After 10 min of
irradiation, yields of formation reach around 700, 150, and
2000 lesions per million normal nucleosides for 8-put-dGuo, 8-
spd-dGuo, and 8-spm-dGuo, respectively. The kinetics of
formation of adducts was almost linear with the irradiation time
at least for short irradiation periods; however, a saturation
occurred for long irradiation periods, as reported already for 8-
oxodGuo and 8-lys-dGuo. Interestingly, when irradiation was
performed in the presence of the polyamines, the yield of
Figure 4. Kinetics of riboflavin and UVA-mediated formation of 8-
oxodGuo in dsDNA. Data represent the average and standard
deviation of at least three independent analyses and are expressed as
the number of 8-oxodGuo per million normal nucleosides.
1015
dx.doi.org/10.1021/tx500063d | Chem. Res. Toxicol. 2014, 27, 1011−1018

Chemical Research in Toxicology
Article
formation of 8-oxodGuo was dramatically reduced (Figure 4
compared to Figure 5). Under these conditions, 8-oxodGuo
formation never exceeded 20 lesions per million nucleosides.
Thus, in the presence of the polyamine, the yield of formation
of 8-oxodGuo was reduced by about 2 orders of magnitude, and
the yield of polyamine-dGuo adducts was significantly higher
than that of 8-oxodGuo, regardless of the polyamine’s identity.
It is notable that spermine was more prone to generate adducts
compared to putrescine and spermidine.
We also noticed that spermine is more reactive than the other
polyamines. Interestingly, we observed that formation of
adducts is accompanied by a decrease in 8-oxodGuo formation.
Indeed, the irradiation in the presence of a high concentration
of polyamines (500 μM) was found to generate more
polyamine adducts than 8-oxodGuo, especially in the case of
spermine. These results highlight the competition between
water and the polyamines as nucleophiles. However, the
decrease in 8-oxodGuo formation is not completely compen-
sated for by polyamine adduct formation, and thus the sum of
damages (polyamine adducts + 8-oxodGuo, Figure 5) that is
produced in the presence of the polyamines is lower than the
yield of 8-oxodGuo (Figure 4) produced in the absence of any
polyamine. This observation adds evidence to the radio-
protective ability of polyamines in inhibiting the formation of 8-
oxodGuo, as reported by Douki et al.34
DISCUSSION
■
The one electron induced oxidation of DNA is a complex
process leading to the formation of numerous oxidative base
damages, especially involving the guanine moiety. In this work,
attempts have been made to determine the ability of
polyamines, ubiquitous molecules found in contact with DNA
in all eukaryotic cells, to react with the guanine radical cation,
and thus to induce the formation of polyamine−guanine cross-
links.
As a result, we hypothesize that the decrease in 8-oxodGuo
formation could be explained by two processes: first, by the
ability of polyamines to condense DNA35 and in this way to
protect it from oxidation; and second by the competitive
nucleophilic addition of polyamines onto the guanine radical
cation that reduces the yield of 8-oxodGuo formation but in the
meantime produces other DNA adducts. Regarding this latter
point, our data indicate that the addition of the amino group of
polyamines is much more efficient than that of water. Indeed, in
the presence of spermine at submillimolar concentration, the
yield of 8-spm-dGuo formation is about 2 orders of magnitude
higher than that of 8-oxodGuo. Interestingly, it has been
demonstrated35 that the B-structure of dsDNA is preserved
even in the presence of high concentrations of polyamines, and
thus, the decrease in 8-oxodGuo formation could not be
attributed to a change in DNA structure but rather due to
condensation. Moreover in the above-mentioned work, it has
been shown that spermine and spermidine undergo hydrogen
bonding interactions (or at least alter the existing interaction)
with guanine N7, and this may explain why nucleophilic
addition of one amino group of the polyamines with C8 of G+•
in dsDNA is more efficient than the addition of a water
molecule to produce 8-oxodGuo. Extrapolating these observa-
tions to cells and taking into consideration that concentration
of polyamines in the nuclei of eukaryotic cells is within the
millimolar range, DNA-polyamine adducts might be generated
in higher yield than 8-oxodGuo and thus may represent suitable
biomarkers of oxidative stress. In addition, the high proportion
of generated polyamine adducts compared to 8-oxodGuo could
also be partly explained by the fact that these polyamines are
positively charged at neutral pH and that this may favor their
interaction with negatively charged DNA, promoting the
formation of polyamine adducts. In that respect, we noticed
that when dsDNA is photosensitized in 10 mM phosphate
buffer, the yield of polyamine adduct formation is significantly
reduced (data not shown) but is still higher than that of 8-
oxodGuo. We have also noticed that the formation of
polyamine-dGuo adducts is linear for short irradiation times
but then reaches a plateau when a high amount of adduct is
present in dsDNA (Figure 5). This could be certainly explained
by the sensitivity of such C8 adducts to undergo secondary
decomposition reactions as reported previously for 8-oxodGuo
and also for the lysine−guanine adduct. Secondary decom-
position reactions of the C8-polyamine-dGuo adducts probably
generate spiro-related derivatives as reported for 8-oxodG-
uo36,37 through a second nucleophilic addition of polyamine at
According to the literature, we anticipated that nucleophilic
addition of polyamines could occur at the C8 position of G•+ or
rather its deprotonated G(-H)• form that should predominate
at neutral pH to generate C8-guanine adducts. Thus, we
developed a chemical synthesis of the different possible
adducts, involving a nucleophilic addition of the amine onto
8-bromo-2′-deoxyguanosine, as reported previously for aro-
matic amines. The structure of the generated adducts was
confirmed by the mean of mass spectrometry, UV, and NMR.
NMR was also used to prepare calibrated solutions in order to
be able to obtain quantitative data. Interestingly, spermidine
being an asymmetric molecule, reaction of either the amino-
propyl or aminobutyl moiety with Br-dGuo could produce the
two different isomers 3a and 3b, respectively (Figure 1). These
1
two isomers were undistinguishable by H NMR but gave
different fragmentation patterns. Indeed, when the polyamine is
linked by the aminobutyl moiety, which is the case for
putrescine, an intense ion at m/z = 221 is observed. When the
cross-link involves the aminopropyl moiety, as observed for 8-
spm-dGuo, an ion at 207 is detected. These ions which arose
from the fragmentation of the glycosidic bond of the nucleoside
and of the polyamine residue as described in Figure 3 allowed
us to characterize the 8-spd-dGuo (3a) adduct as 8-((3-((4-
aminobutyl)amino)propyl)amino)-2′-deoxyguanosine and 3b
as 8-((4-((3-aminopropyl)amino)butyl)amino)-2′-deoxyguano-
sine. However, since these two adducts are generated in a
similar yield (not shown), reported data represent the sum of
the two adducts 3a and 3b, named 8-spd-dGuo (3).
The quantification of C8 polyamine-dGuo was performed by
HPLC coupled to tandem mass sepectrometry in the classical
MRM mode32 that provides a very low limit of detection
(about 20 fmol injected for 8-oxodGuo33). To improve
specificity, two transitions were used for each lesion (Table
2). Quantification was performed by external calibration, and
results were expressed as the number of lesions per million
normal nucleosides quantified by UV detection set at 260 nm.
We performed type I photooxidation of dsDNA in the
presence of polyamines, in order to determine if 8-polyamine-
dGuo adducts are generated upon a one electron oxidation
reaction. Under these conditions, we detected a significant
formation of 8-put-dGuo, 8-spd-dGuo, and 8-spm-dGuo using
the developed analytical HPLC-MS/MS method. Our results
indicate that the three polyamines, putrescine, spermine, and
spermidine, are able to react with the guanine radical cation.
1016
dx.doi.org/10.1021/tx500063d | Chem. Res. Toxicol. 2014, 27, 1011−1018

Chemical Research in Toxicology
Article
Present Address
C5. Additional work has to be performed to confirm that
hypothesis.
†(M.T.-P.) Department of Chemistry, Johns Hopkins Uni-
versity, 3400 North Charles Street, Baltimore, MD 21218.
Additional work will be performed to search for the
formation of these new DNA lesions in irradiated cells.
However, the limit of detection of our developed method for 8-
polyamine-dGuo detection is around 1 pmol injected. Such a
sensitivity is good enough to monitor the formation of these
lesions in isolated irradiated DNA because the yield of cross-
link formation with polyamines is relatively high. Nevertheless,
in biological environments, the yield of formation of DNA
lesions is generally much lower, and thus the investigation of
polyamine adducts at the cellular level would require a higher
sensitivity, with a limit of quantification of only a few fmol. It
should be noted that the observed limit of detection for the
polyamine−guanine adducts studied here is mostly explained
by the nonclassical chromatographic behavior of this kind of
molecule. The presence of a primary amino group in these
adducts causes the compounds to elute as a broad peak, this
effect being particularly important for low injected amounts
(below 1 pmol under our conditions). Thus, efforts will be
made to improve the sensitivity of the method, and for such a
purpose, attempts to derivatize the primary free amino group
are in progress.
Funding
This work was partly supported by Labex PRIMES (ANR-11-
LABX-0063), COST action CM1201 “Biomimetic Radical
Chemistry,” and TODNAL project (ANR-09-PIRI-0022-01).
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The contributions of Pierre-Alain Bayle (LRM/SCIB/CEA-
Grenoble) to the NMR analysis and Hassan Kanso to the
preliminary experiments are gratefully acknowledged.
ABBREVIATIONS
■
ROS, reactive oxygen species; dGuo, 2′deoxyguanosine; G+•,
guanine radical cation; G(-H)•, deprotonated guanine radical
cation; 8-Br-dGuo, 8-bromo-2′-deoxyguanosine; 8-oxodGuo, 8-
oxo-7,8-dihydro-2′-deoxyguanosine; 8-put-dGuo (2), 8-((4-
aminobutyl)amino)-2′-deoxyguanosine; 8-spd-dGuo, 8-((3-
((4-aminobutyl)amino)propyl)amino)-2′-deoxyguanosine (3a)
and 8-((4-((3-aminopropyl)amino)butyl)amino)-2′-deoxygua-
nosine (3b); 8-spm-dGuo (4), 8-((3-((4-((3-aminopropyl)-
amino)butyl)amino)propyl)amino)-2′-deoxyguanosine; ODC,
ornithine decarboxylase; dsDNA, double stranded DNA;
HPLC-MS/MS, HPLC coupled through electrospray ionization
to tandem mass spectrometry
CONCLUSIONS
■
This work demonstrates that polyamines are able to react onto
the C8 position of guanine following one electron oxidation
and that such a process generates polyamine−guanine cross-
links. In the meantime, we demonstrated that the presence of
polyamines inhibits the formation of 8-oxodGuo. This study
suggests that besides the ability of polyamines to protect DNA
from oxidation by condensation, the observed decrease of 8-
oxodGuo formation in the presence of these molecules could
be partially explained by the competitive formation of
polyamine adducts. Thus, the observed protection effect of
polyamine might be considered with caution.
REFERENCES
■
(1) Lindahl, T. (1993) Instability and decay of the primary structure
of DNA. Nature 362, 709−715.
(2) Cooke, M. S., Evans, M. D., Dizdaroglu, M., and Lunec, J. (2003)
Oxidative DNA damage: mechanisms, mutation, and disease. FASEB J.
17, 1195−1214.
Further work is in progress to improve the sensitivity of the
HPLC-MS/MS method for the detection of such polyamine
adducts, in order to be able to investigate their formation within
cells. In addition, attempts will be made to determine their
biological consequences and to identify repair mechanisms
involved in the elimination of these specific DNA lesions.
The investigation of DNA-polyamine cross-links also
represents a convenient model for the study of other more
complex lesions involving nucleophilic addition onto C8 of the
guanine radical cation or its deprotonated form. A similar
strategy will be applied to search for cross-links between
guanine and several amino acids, such as lysine, arginine, and
serine. This outlook could be promising in the understanding of
structural aspects of DNA−protein cross-links.38
(3) Kanvah, S., Joseph, J., Schuster, G. B., Barnett, R. N., Cleveland,
C. L., and Landman, U. (2010) Oxidation of DNA: damage to
nucleobases. Acc. Chem. Res. 43, 280−287.
(4) Foote, C. S. (1991) Definition of type I and type II
photosensitized oxidation. Photochem. Photobiol. 54, 659.
(5) Pratviel, G., and Meunier, B. (2006) Guanine oxidation: one- and
two-electron reactions. Chem.Eur. J. 12, 6018−6030.
(6) Cadet, J., Douki, T., and Ravanat, J. L. (2008) Oxidatively
generated damage to the guanine moiety of DNA: Mechanistic aspects
and formation in cells. Acc. Chem. Res. 41, 1075−1083.
(7) Giese, B. (2002) Electron transfer in DNA. Curr. Opin. Chem.
Biol. 6, 612−618.
(8) Kasai, H., Yamaizumi, Z., Berger, M., and Cadet, J. (1992)
Photosensitized formation of 7,8-dihydro-8-oxo-2′-deoxyguanosine (8-
hydroxy-2′-deoxyguanosine) in DNA by riboflavin: a non singlet
oxygen mediated mechanism. J. Am. Chem. Soc. 114, 9692−9694.
(9) Buchko, G. W., Cadet, J., Ravanat, J.-L., and Labataille, P. (1993)
Isolation and characterization of a new product produced by ionizing
irradiation and type I photosensitization of 2′-deoxyguanosine in
oxygen-saturated aqueous solution: (2S)-2,5′-anhydro-1-(2-deoxy-α-
D-erythro-pentofuranosyl)-5-guanidinylidene-2-hydroxy-4-oxoimida-
zolidine. Int. J. Radiat. Biol. 63, 669−676.
ASSOCIATED CONTENT
■
S
* Supporting Information
Conditions for exact mass determination, general procedure for
the chemical synthesis of the polyamine-C8 guanine adducts,
UV absorption spectra of 8-put-dGuo compared to dGuo, and
1H NMR spectra of 8-spm-dGuo. This material is available free
of charge via the Internet at http://pubs.acs.org.
(10) Morin, B., and Cadet, J. (1995) Type I benzophenone-mediated
nucleophilic reaction of 5′-amino-2′,5′-dideoxyguanosine. A model
system for the investigation of photosensitized formation of DNA-
protein cross-links. Chem. Res. Toxicol. 8, 792−799.
(11) Morin, B., and Cadet, J. (1995) Chemical aspects of the
benzophenone-photosensitized formation of two lysine-2′-deoxygua-
nosine cross-links. J. Am. Chem. Soc. 117, 12408−12415.
AUTHOR INFORMATION
■
Corresponding Author
*Tel: +33 4 38 78 47 97. E-mail: jravanat@cea.fr.
1017
dx.doi.org/10.1021/tx500063d | Chem. Res. Toxicol. 2014, 27, 1011−1018

Chemical Research in Toxicology
Article
(12) Morin, B., and Cadet, J. (1994) Benzophenone photo-
sensitization of 2′-deoxyguanosine: characterization of the 2R and 2S
diastereoisomers of 1-(2-deoxy-β-D-erythro-pentofuranosyl)- 2-me-
thoxy-4,5-imidazolidinedione. A model system for the investigation
of photosensitized formation of DNA-protein crosslinks. Photochem.
Photobiol. 60, 102−109.
(13) Cadet, J., Douki, T., and Ravanat, J.-L. (2006) One-electron
oxidation of DNA and inflammation processes. Nat. Chem. Biol. 2,
348−349.
(14) Perrier, S., Hau, J., Gasparutto, D., Cadet, J., Favier, A., and
Ravanat, J. L. (2006) Characterization of lysine-guanine cross-links
upon one-electron oxidation of a guanine-containing oligonucleotide
in the presence of a trilysine peptide. J. Am. Chem. Soc. 128, 5703−
5710.
(15) Madugundu, G. S., Wagner, J. R., Cadet, J., Kropachev, K., Yun,
B. H., Geacintov, N. E., and Shafirovich, V. (2013) Generation of
guanine-thymine cross-links in human cells by one-electron oxidation
mechanisms. Chem. Res. Toxicol. 26, 1031−1033.
(16) Rokhlenko, Y., Geacintov, N. E., and Shafirovich, V. (2012)
Lifetimes and reaction pathways of guanine radical cations and neutral
guanine radicals in an oligonucleotide in aqueous solutions. J. Am.
Chem. Soc. 134, 4955−4962.
(17) Sarhan, S., and Seiler, N. (1989) On the subcellular localization
of the polyamines. Biol. Chem. Hoppe. Seyler 370, 1279−1284.
(18) Gerner, E. W., and Meyskens, F. L. (2004) Polyamines and
cancer: Old molecules, new understanding. Nat. Rev. Cancer 4, 781−
792.
(19) Saminathan, M., Thomas, T., Shirahata, A., Pillai, C. K., and
Thomas, T. J. (2002) Polyamine structural effects on the induction
and stabilization of liquid crystalline DNA: potential applications to
DNA packaging, gene therapy and polyamine therapeutics. Nucleic
Acids Res. 30, 3722−3731.
(20) Bachrach, U. (2004) Polyamines and cancer: Minireview article.
Amino Acids 26, 307−309.
(29) Munzel, M., Szeibert, C., Glas, A. F., Globisch, D., and Carell, T.
(2011) Discovery and synthesis of new UV-induced intrastrand C(4−
8)G and G(8−4)C photolesions. J. Am. Chem. Soc. 133, 5186−5189.
(30) Johnson, F., Huang, C. Y., and Yu, P. L. (1994) Synthetic and
oxidative studies on 8-(arylamino)-2′-deoxyguanosine and -guanosine
derivatives. Environ. Health Perspect. 102 (Suppl 6), 143−149.
(31) Regulusqq, P., Spessotto, S., Gateau, M., Cadet, J., Favier, A.,
and Ravanat, J.-L. (2004) Detection of new radiation-induced DNA
lesions by liquid chromatography coupled to tandem mass
spectrometry. Rapid Commun. Mass Spectrom. 18, 2223−2228.
(32) Cadet, J., and Weinfeld, M. (1993) Detecting DNA damage.
Anal. Chem. 65, 675A−682A.
(33) Ravanat, J. L., Duretz, B., Guiller, A., Douki, T., and Cadet, J.
(1998) Isotope dilution high-performance liquid chromatography-
electrospray tandem mass spectrometry assay for the measurement of
8-oxo-7,8-dihydro-2′-deoxyguanosine in biological samples. J. Chro-
matogr., B 715, 349−356.
(34) Douki, T., Bretonniere, Y., and Cadet, J. (2000) Protection
against radiation-induced degradation of DNA bases by polyamines.
Radiat. Res. 153, 29−35.
(35) Deng, H., Bloomfield, V. A., Benevides, J. M., and Thomas, G. J.,
Jr. (2000) Structural basis of polyamine-DNA recognition: spermidine
and spermine interactions with genomic B-DNAs of different GC
content probed by Raman spectroscopy. Nucleic Acids Res. 28, 3379−
3385.
(36) Johansen, M. E., Muller, J. G., Xu, X., and Burrows, C. J. (2005)
Oxidatively induced DNA-protein cross-linking between single-
stranded binding protein and oligodeoxynucleotides containing 8-
oxo-7,8-dihydro-2′-deoxyguanosine. Biochemistry 44, 5660−5671.
(37) Xu, X., Muller, J. G., Ye, Y., and Burrows, C. J. (2008) DNA-
protein cross-links between guanine and lysine depend on the
mechanism of oxidation for formation of C5 vs C8 guanosine adducts.
J. Am. Chem. Soc. 130, 703−709.
(38) Barker, S., Weinfeld, M., and Murray, D. (2005) DNA-protein
crosslinks: their induction, repair, and biological consequences. Mutat.
Res. 589, 111−135.
(21) Moulinoux, J. P., Quemener, V., Havouis, R., Guille, F., Martin,
C., and Seiler, N. (1991) Accumulation of polyamine analogs in red
blood cells: a potential index of tumor proliferation rate. Anticancer
Res. 11, 2143−2146.
(22) Moulinoux, J. P., Quemener, V., and Khan, N. A. (1991)
Biological significance of circulating polyamines in oncology. Cell. Mol.
Biol. 37, 773−783.
(23) Pegg, A. E., and Hu, R. H. (1995) Effect of polyamine analogues
and inhibition of polyamine oxidase on spermidine/spermine N1-
acetyltransferase activity and cell proliferation. Cancer Lett. 95, 247−
252.
(24) Demoor-Goldschmidt, C., and Raynard, B. (2009) How can we
integrate nutritional support in medical oncology? Bull. Cancer 96,
665−675.
(25) Kasai, H., Yamaizumi, Z., Berger, M., and Cadet, J. (1992)
Photosensitized formation of 7,8-dihydro-8-oxo-2′-deoxyguanosine (8-
hydroxy-2′-deoxyguanosine) in DNA by riboflavin - a nonsinglet
oxygen mediated reaction. J. Am. Chem. Soc. 114, 9692−9694.
(26) Ravanat, J.-L., Douki, T., Duez, P., Gremaud, E., Herbert, K.,
Hofer, T., Lasserre, L., Saint-Pierre, C., Favier, A., and Cadet, J. (2002)
Cellular background level of 8-oxo-7,8-dihydro-2′-deoxyguanosine: an
isotope based method to evaluate artefactual oxidation of DNA during
its extraction and subsequent work-up. Carcinogenesis 23, 1911−1918.
(27) Frelon, S., Douki, T., Ravanat, J.-L., Pouget, J.-P., Tornabene, C.,
and Cadet, J. (2000) High performance liquid chromatography -
tandem mass spectrometry for the measurement of radiation-induced
base damage to isolated and cellular DNA. Chem. Res. Toxicol. 13,
1002−1010.
(28) Ravanat, J.-L., Duretz, B., Guiller, A., Douki, T., and Cadet, J.
(1998) Isotope dilution high-performance liquid chromatography-
electrospray tandem mass spectrometry assay for the measurement of
8-oxo-7,8-dihydro-2′-deoxyguanosine in biological samples. J. Chro-
matogr., B 715, 349−356.
1018
dx.doi.org/10.1021/tx500063d | Chem. Res. Toxicol. 2014, 27, 1011−1018