High-intensity UV laser photolysis of DNA and purine 2'- deoxyribonucleosides: Formation of 8-oxopurine damage and oligonucleotide strand cleavage as revealed by HPLC and Gel electrophoresis studies

Article abstract of DOI:10.1021/ja971728r

Emphasis was placed in this work on the measurement of purine oxidation products generated upon nano- and picosecond UV laser biphotonic photolysis of 2'-deoxyadenosine, 2'-deoxyguanosine, calf thymus DNA, and a synthetic duplex oligonucleotide (37-mer) in aerated aqueous solutions. The overall formation of 8-oxo-7,8-d hydro- 2'-deoxyguanosine and 8-oxo-7,8-dihydro-2' - deoxyadenosme was determined using a HPLC-electrochemical assay. Denaturing gel electrophoresis analysis in association with a formamidopyrimidineNA glycosylase treatment was applied to reveal the sites recognized by this DNA repair enzyme. Both 8-oxo-7,8-dihydro-2'-deoxyguanosine and 8-oxo-7,8 dihydro-2'-deoxyadenosine were shown to be minor decomposition products of the related nucleoside purine radical cations in agreement with earlier observations. Interestingly, a dramatic increase in the yield of both photoproducts, this applying particularly to 8-oxo-7,8-dihydro-2'- deoxyguanosine, was observe in DNA. It should be noted that the yield of 8- oxo-7,8-dihydro-2'-deoxyguanosine was about 3-fold lower in heat-denatured DNA than in double-stranded DNA. These observations provide strong support to the significant involvement of base stacking and probably DNA solvatation in the chemical reactions of the purine radical cations. Other interesting information dealt with the similarity in the level of 8-oxo-7,8-dihydro-2'- deoxyguanosine and the number of formanidopyrimidine- DNA glycosytase sensitive guanine lesions. This strongly suggests that the latter formidopyrimidine-DNA glycosyiase pine nucleoside is the major DNA photodamaged product recognized by the DNA repair glycosylase. Another striking feature is the almost 10-fold decrease in the saturation dose E(s) for the two-quantum ionization of the guanine base in double-stranded DNA as compared to that observed for free 2'-deoxyguanosine. This can be explained by either an enhancement of the quantum yield of photoionization from the intermediate excited state in DNA (φ₂) and/or hole migration with preferential trapping by guanine residues.

Full text of DOI:10.1021/ja971728r

J. Am. Chem. Soc. 1997, 119, 11373-11380
11373
High-Intensity UV Laser Photolysis of DNA and Purine
2′-Deoxyribonucleosides: Formation of 8-Oxopurine Damage
and Oligonucleotide Strand Cleavage as Revealed by HPLC and
Gel Electrophoresis Studies
D. Angelov,† A. Spassky,‡ M. Berger,§ and J. Cadet*,§
Contribution from the Institute of Solid State Physics, Bulgarian Academy of Sciences,
1784 Sofia, Bulgaria, De´partement de Recherche Fondamentale sur la Matie`re Condense´e,
SCIB/Laboratoire “Le´sions des Acides Nucle´iques”, CEA/Grenoble, 17, rue des Martyrs,
F-38054 Grenoble, Cedex 9, France, and UniVersite´ Rene´ Descartes, UA 400, 45,
rue des Saints Pe`res, F-75270 Paris, Cedex, France
ReceiVed May 27, 1997X
Abstract: Emphasis was placed in this work on the measurement of purine oxidation products generated upon nano-
and picosecond UV laser biphotonic photolysis of 2′-deoxyadenosine, 2′-deoxyguanosine, calf thymus DNA, and a
synthetic duplex oligonucleotide (37-mer) in aerated aqueous solutions. The overall formation of 8-oxo-7,8-dihydro-
2′-deoxyguanosine and 8-oxo-7,8-dihydro-2′-deoxyadenosine was determined using a HPLC-electrochemical detection
assay. Denaturing gel electrophoresis analysis in association with a formamidopyrimidine-DNA glycosylase treatment
was applied to reveal the sites recognized by this DNA repair enzyme. Both 8-oxo-7,8-dihydro-2′-deoxyguanosine
and 8-oxo-7,8-dihydro-2′-deoxyadenosine were shown to be minor decomposition products of the related nucleoside
purine radical cations in agreement with earlier observations. Interestingly, a dramatic increase in the yield of both
photoproducts, this applying particularly to 8-oxo-7,8-dihydro-2′-deoxyguanosine, was observed in DNA. It should
be noted that the yield of 8-oxo-7,8-dihydro-2′-deoxyguanosine was about 3-fold lower in heat-denatured DNA than
in double-stranded DNA. These observations provide strong support to the significant involvement of base stacking
and probably DNA solvatation in the chemical reactions of the purine radical cations. Other interesting information
dealt with the similarity in the level of 8-oxo-7,8-dihydro-2′-deoxyguanosine and the number of formamidopyrimidine-
DNA glycosylase sensitive guanine lesions. This strongly suggests that the latter formamidopyrimidine-DNA
glycosylase purine nucleoside is the major DNA photodamaged product recognized by the DNA repair glycosylase.
Another striking feature is the almost 10-fold decrease in the saturation dose Es for the two-quantum ionization of
the guanine base in double-stranded DNA as compared to that observed for free 2′-deoxyguanosine. This can be
explained by either an enhancement of the quantum yield of photoionization from the intermediate excited state in
DNA (æ2) and/or hole migration with preferential trapping by guanine residues.
Introduction
pyrimidine radical cation of dThd, dCyd, and 5-methyl-2′-
deoxycytidine was mostly inferred from type I photosensitization
studies involving 2-methyl-1,4-naphthoquinone as the photo-
dynamic agent.3-6 Significant progress in the determination of
the major oxidation products deriving from the purine radical
cations is also associated with photosensitization studies. Thus,
it was found that riboflavin-mediated photosensitization of DNA
leads to the significant formation of 8-oxo-7,8-dihydro-2′-
deoxyguanosine (8-oxo-dGuo) through a type I mechanism.7
Interestingly, 8-oxo-dGuo is not generated, at least in significant
amounts, in free dGuo and short single-stranded oligonucleotides
due to a fast deprotonation of the transient guanine radical
cation.8,9 The main oxidation product which arises from the
latter pathway was characterized as an oxazolone derivative,10,11
It is now well established that exposure of DNA and related
model compounds to either 248- or 266-nm high-intensity laser
pulses gives rise to pyrimidine and purine radical cations as
the result of biphotonic excitation.1,2 It is also assumed that
the latter reactive transient radicals are converted into stable
oxidation products via two competitive pathways involving
deprotonation and hydration.3 Only limited information is
available on the stable decomposition products generated by
biphotonic laser photolysis of DNA components. One major
exception deals with a preliminary report on the characterization
of the main UV laser-induced photoproducts of thymidine
(dThd) and 2′-deoxycytidine (dCyd).4 It should be mentioned
that detailed information on the chemical reactions of the
(4) Angelov, D. A.; Berger, M.; Cadet, J.; Getoff, N.; Keskinova, E.;
Solar, S. Radiat. Phys. Chem. 1991, 37, 717-727.
* To whom the correspondence should be addressed. Telephone: (33)-
4-76-88-49-87. FAX: (33)-4-76-88-50-90. E-mail: cadet@drfmc.ceng.cea.fr.
† Bulgarian Academy of Sciences.
(5) Cadet, J.; P. Vigny P. In Bioorganic Photochemistry; Morrison, H.,
Ed.; John Wiley & Sons: New York, 1990; Vol. 1, pp 1-272.
(6) (a) Decarroz, C.; Wagner, J. R.; van Lier, J. E.; Murali Krishna, C.;
Riesz, P.; Cadet, J. Int. J. Radiat. Biol. 1986, 50, 491-505. (b) Wagner, J.
R.; van Lier J. E.; Decarroz, C.; Berger, M.; Cadet, J. Methods Enzymol.
1990, 186B, 502-511. (c) Wagner, J. R; van Lier, J. E.; Berger, M.; Cadet,
J. J. Am. Chem. Soc. 1994, 116, 2235-2242. (d) Bienvenu, C.; Wagner, J.
R.; Cadet, J. J. Am. Chem. Soc. 1996, 118, 11406-11411.
(7) Kasai, H.; Yamaizumi, Z.; Berger, M.; Cadet, J. J. Am. Chem. Soc.
1992, 114, 9692-9694.
‡ Universite´ Rene´ Descartes.
§ CEA/Grenoble.
X Abstract published in AdVance ACS Abstracts, October 15, 1997.
(1) Nikogosyan, D. N. Int. J. Radiat. Biol. 1990, 37, 233-299.
(2) Go¨rner, H. J. Photochem. Photobiol. B 1994, 26, 117-139.
(3) (a) Cadet, J. In DNA Adducts: Identification and Biological
Significance; Hemminki, K., Dipple, A., Shuker, D. E. G., Segerba¨ck, D.,
Barsch, H., Eds.; IARC Scientific Publications: Lyon, France, 1994; Vol.
125, pp 245-276. (b) Cadet, J.; Berger, M.; Douki, T.; Ravanat, J.-L. ReV.
Physiol. Biochem. Pharmacol. 1997, 131, 1-87.
(8) Steenken, S. Chem. ReV. 1989, 89, 503-520.
(9) O’Neill, P.; Fielden, M. AdV. Radiat. Biol. 1993, 17, 53-120.
S0002-7863(97)01728-9 CCC: $14.00 © 1997 American Chemical Society

11374 J. Am. Chem. Soc., Vol. 119, No. 47, 1997
AngeloV et al.
a highly alkali-labile compound.12 It was shown on the basis
of sequencing gel electrophoresis analysis that exposure of DNA
to high-intensity laser pulses results in the formation of alkali-
labile lesions13,14 and DNA polymerase stops15 at guanine sites.
However, there is still a paucity of information on the chemical
structure of the oxidative damage occurring upon photoioniza-
tion of purine bases. One major exception deals with a recent
study on the 248-nm excimer laser-mediated photoionization
of oligonucleotides and free 2′-deoxyguanosine.16 It was shown
that 8-oxo-dGuo and unknown piperidine-labile guanine lesions
are generated under these conditions. A more indirect evidence
for the formation of 8-oxo-dGuo was provided by the measure-
ment of formamidopyrimidine-DNA glycosylase (Fpg)-sensi-
tive sites within 193-nm irradiated DNA fragments.17 It was
also reported that either mono- or biphotonic laser photolysis
of DNA gives rise to strand cleavages. However, the process
is at the best inefficient with respect to the formation of oxidative
base damage and usual pyrimidine photoproducts including
cyclobutadipyrimidines and pyrimidine (6-4) pyrimidone pho-
toadducts.1,16,18
Emphasis was placed in the present work on the measurement
of the quantum yield of photoinduced 8-oxo-dGuo and 8-oxo-
dAdo in different purine substrates including 2′-deoxyguanosine
(dGuo), 2′-deoxyadenosine (dAdo), calf thymus DNA, and a
defined double-stranded oligonucleotide (37-mer). This was
achieved using two different approaches including a HPLC-
electrochemical detection assay and a denaturing gel electro-
phoresis method associated with a formamidopyrimidine-DNA
glycosylase protein19 treatment. The observed features are
discussed in terms of efficiency of ionization of purine DNA
samples and chemical reactivity of the related radical cations.
Figure 1. Sequence of the 9G and 4G oligonucleotides.
deoxyribonucleotides (37-mer) were prepared on an Applied Biosystem
synthesizer and deprotected according to the manufacturer’s procedure
(for the sequence, see Figure 1). A final purification was achieved on
a high-resolution 15% polyacrylamide/7 M urea denaturing gel.
Oligonucleotides were 5′-end labeled using [γ-32P]ATP (Amersham)
and T4 polynucleotide kinase (Boehringer). The resulting radiolabeled
oligonucleotide was purified by filtration on a Sephadex G-50 column
and subsequently annealed with the corresponding complementary
strand. Nonlabeled oligonucleotides, used for HPLC analysis, were
annealed with the complementary DNA fragment in a 1:1 ratio.
Controls were made on a nondenaturing 15% electrophoretic gel. The
purified Fpg protein was a kind gift of Dr Serge Boiteux (CEA/
Fontenay-aux-Roses, France). All other chemicals used were from the
highest purity grade.
Laser Apparatus. The fourth harmonic (λ ) 266 nm) of two
different types of Nd:YAG lasers was used as the source of UV
radiation. The first one was a passively mode locked laser delivering
single 30-ps pulses with an energy maximum of 8 mJ at 266 nm. The
second far-UV source was a Q-switched laser providing 15-ns duration
pulses with more than 30 mJ of energy maximum at 266 nm. The
energy of the laser radiation was measured with a model Rj7100 (Laser
Precision) calibrated pyroelectric energy meter ((5%) using the Fresnel
reflections from a fused silica beam splitter positioned behind the
sample. The beam fluence was varied by either using a focusing (f )
40 cm) quartz lens or changing the laser amplifier voltage together
with detuning the harmonic generation crystals. The laser pulse dose
was determined by using the measured energy and by assuming regular
distribution of the energy across the laser beam. The diameter of the
beam was determined by a photodiode array camera.
Laser Irradiation. Irradiation of dAdo, dGuo, and calf thymus
DNA was carried out in 0.5-cm spectrophotometer fused silica cuvettes
(optically thick solutions) under continuous stirring. The solutions of
2 mM purine nucleosides were prepared in bidistilled water (pH 6.5)
whereas calf thymus DNA (1 mM concentration in nucleotides) was
irradiated in 50 mM phosphate buffer (pH 7.6). For the different laser
intensities, the total irradiation energy was selected in the linear range
of the quantum yield response. The 37 bp DNA fragments, dissolved
in TE buffer (10 mM Tris (pH 7.6), 1 mM EDTA, 40 mM NaCl),
were irradiated as 10-µL aliquots in small 0.65-mL siliconized
Eppendorf tubes (optically thin layer). The laser beam diameter was
reduced to 0.25 cm (S ) 0.05 cm2) using a circular diaphragm to match
with the area of the irradiated sample. [32P]-5′-Labeled DNA (0.02
pmol) was irradiated with one pulse at the maximum fluence of 0.1
J‚cm-2 (I ) 6.6 MW‚cm-2). In parallel the number of pulses at lower
intensities was increased to a total irradiation dose producing similar
DNA degradation (below 10-15%) for all intensities used. For HPLC
analysis, 1.83 µg of nonlabeled 37 bp DNA was dissolved in 50 µL of
TE buffer and aliquots of 10 µL were transferred prior to irradiation
into five Eppendorf tubes to ensure an optically thin irradiation layersD
) 0.133 in the 0.2-mm irradiation path length (assuming ꢀDNA ) 6000
M-1 cm-1 at 266 nm). All five samples were exposed to two pulses
having energies of 5.6 mJ each (I ) 7.5 MW‚cm-2). After irradiation,
the aliquots were transferred in one tube and lyophilized prior to HPLC
analysis.
HPLC-EC Determination of 8-Oxo-dGuo and 8-Oxo-dAdo in
Photoirradiated Aqueous Solutions of Purine 2′-Deoxyribonucleo-
sides. After laser irradiation, the solutions of dAdo and dGuo were
evaporated to dryness. The resulting residues were dissolved in about
200 µL of water prior to HPLC analysis for their content in 8-oxo-
dAdo and 8-oxo-dGuo. The HPLC system which was used for this
purpose consisted of a LKB model 2150 pump (Pharmacia) equipped
with a Rheodyne loop injector, Model 7125. The electrochemical
detection was achieved using a dual LC-4B/17 AT amperometer
detector (Bioanalytical System). The separation of 8-oxo-dAdo and
8-oxo-dGuo was carried out on an analytical octadecylsilyl silica gel
column (4.6 mm i.d. × 250 mm, 5 µm) from Interchim. The eluent
was a mixture of 0.05 M citrate buffer (pH 5) and methanol (80:20,
v/v). The oxidation potential was set up at +850 mV for the detection
Experimental Section
Materials. 2′-Deoxyadenosine (dAdo), 2′-deoxyguanosine (dGuo),
and calf thymus DNA were purchased from Sigma. Proteins were
removed from calf thymus DNA by three successive phenol-
chloroform extractions. 8-Oxo-dGuo and 8-oxo-dAdo were prepared
according to slight modifications of reported procedures.20,21 Oligo-
(10) Cadet, J.; Berger, M.; Buchko, G. W.; Joshi, P. C.; Raoul, S.;
Ravanat, J.-L. J. Am. Chem. Soc. 1994, 116, 7403-7404.
(11) Buchko, G. W.; Cadet, J.; Morin, B.; Weinfeld, M. Nucleic Acids
Res. 1995, 19, 3954-3961.
(12) Raoul, S.; Berger, M.; Buchko, G. W.; Joshi, P. C.; Morin, B.;
Weinfeld, M.; Cadet J. J. Chem. Soc., Perkin Trans. 2 1996, 371-381.
(13) Kovalsky, O. G.; Panyutin, I. G.; Budowsky, E. I. Photochem.
Photobiol. 1990, 52, 509-517.
(14) Angelov, D.; Berger, M.; Cadet, J.; Marion, C.; Spassky, A. Trends
Photochem. Photobiol. 1994, 3, 643-663.
(15) Panyutin, I.; Kovalsky, G. O. I.; Budowsky, E. I. FEBS Lett. 1989,
258, 274-276.
(16) Cullis, P. M.; Malone M. E.; Merson-Davies, L. A. J. Am. Chem.
Soc. 1996, 118, 2775-2781.
(17) Melvin, T.; Cunniffe, S.; Papworth, D.; Roldan-Arjona, T.; O’Neill,
P. Photochem. Photobiol. 1997, 65, 660-665.
(18) (a) Croke, D. T.; Blau, W.; OhUigin, C.; Kelly, J. M.; McConnell,
D. J. Photochem. Photobiol. 1988, 47, 527-536. (b) Kochevar, I. E.;
Buckley, L. A. Photochem. Photobiol. 1990, 51, 527-532. (c) Bothe, E.;
Go¨rner, H.; Opitz, J.; Schulte-Frohlinde, D.; Saddiqi, A.; Wala, M.
Photochem. Photobiol. 1990, 52, 949-959. (d) Masnic, T. W.; Minton, K.
Photochem. Photobiol. 1991, 54, 99-107. (e) Gurzadyan, G. G.; Go¨rner,
H. Photochem. Photobiol. 1993, 58, 71-80. (f) Melvin, T.; Plumb, M. A.;
Botchway, S. W.; O’Neill, P.; Parker, A. W. Photochem. Photobiol. 1995,
61, 584-591. (g) Sies, H.; Schulz, W. A.; Steenken, S. J. Photochem.
Photobiol. B 1996, 32, 97-102. (h) Melvin, T.; Botchway, S. W.; Parker,
A. W.; O’Neill, P. J. Am. Chem. Soc. 1996, 118, 10031-10036.
(19) (a) Tchou, J.; Kasai, H.; Shibutani, S.; Chung, M.-H.; Laval, J.;
Grollman, A. P.; Nishimura, S. Proc. Natl. Acad. Sci. U.S.A. 1991, 88,
4690-4694. (b) Boiteux, S. J. Photochem. Photobiol. B 1993, 19, 87-96.
(c) Tchou, J.; Grollman, A. P. J. Biol. Chem. 1995, 270, 11671-11677.
(20) Lin, T.-S.; Cheng, J. C.; Ishiguro, K.; Sartorelli, C. J. Med. Chem.
1985, 28, 1194-1197.
(21) Berger, M.; de Hazen, M.; Nejjari, A.; Fournier, J.; Guignard, H.;
Pezerat, H.; Cadet, J. Carcinogenesis 1993, 14, 41-46.

UV Laser Photolysis of DNA and Related Compounds
J. Am. Chem. Soc., Vol. 119, No. 47, 1997 11375
of both 8-oxo-dAdo and 8-oxo-dGuo.22 However, the detection of
8-oxo-dGuo could be accomplished in a more specific way, using a
lower value (650 mV) of oxidation potential.23 The quantitative
determination of unmodified dAdo and dGuo was achieved by UV
measurement using a Merck Model L 4000 variable spectrometer set
up at 254 nm. Quantification of normal and oxidized nucleosides was
made using calibration curves.
and guanine bases within DNA at 266 nm (ꢀ ) 6000 M-1 cm-1), Etot
is the total irradiation fluence expressed in photons‚cm-2, L ) 37 is
the number of bases in the 5′-labeled oligonucleotide, and RT and RL
are the total radioactivity deposited and the radioactivity in the ladder
sequence, respectively.
Results
Quantitative Measurement of 8-Oxo-dAdo and 8-Oxo-dGuo in
Calf Thymus DNA and Duplex Oligodeoxyribonucleotide. Irradiated
DNA samples were ethanol precipitated and redissolved in 100 µL of
water for enzymatic digestion. This was achieved by adding to the
latter solution 10 µL of 10 X (300 mM sodium acetate (pH 5.3), 10
mM zinc sulfate) and 5 µL of P1 nuclease ready to use (Boehringer).
After incubation for 2 h at 37 °C, 12 µL of 10 X (2 M NH4HCO3, pH
9) and 2 µL of alkaline phosphatase (Boehringer) were added. The
resulting hydrolysate was analyzed for their content in 8-oxo-dAdo,
8-oxo-dGuo, and normal nucleosides by applying the above-described
HPLC assays.
Generation of 8-Oxo-dAdo and 8-Oxo-dGuo in Isolated
dAdo, dGuo, and Calf Thymus DNA. The quantum yields
for the generation of 8-oxo-dAdo and 8-oxo-dGuo versus the
laser pulse intensity upon exposure of dAdo and dGuo to nano-
and picosecond pulses are reported in Figure 2. The solid lines
represent theoretical fits of the experimental points with the
corresponding expressions27 describing the biphotonic photolysis
in an optical thick layer. Consideration of the slope of the
curves which show an initial linear increase followed by a
saturation unambiguously demonstrates the biphotonic character
of the photoprocess involved. The values of the saturation
fluences Is are similar to those already reported for chromophore
loss,28 namely several tens of MW‚cm-2 and several GW‚cm-2
in the case of nanosecond and picosecond photolysis, respec-
tively. The saturated (maximum) values of the quantum yields
QYmax are very low and similar for both photoproducts.
However, under the picosecond mode of irradiation, the quantum
yields are about 5 times higher. Upon addition of an •OH radical
scavenger (5% MeOH), significant (3-5-fold) decreases in the
yields of both 8-oxo-dAdo and 8-oxo-dGuo are observed under
Denaturing Gel Electrophoresis. The laser-irradiated [32P]-5′-
labeled 37 bp DNA fragments were incubated with 5 ng of Fpg protein
for 45 min at 37 °C. Then, the solution was lyophilized and the
resulting dry residue resuspended in formamide loading buffer. DNA
fragments were heat denatured during 5 min at 95 °C and subsequently
run on a 15% denaturing polyacrylamide mini gel. The same procedure,
but without Fpg treatment, was applied for measuring frank DNA strand
breaks. Following electrophoresis, gels were dried and autoradio-
graphed by using a Phosphorimager apparatus (Molecular Dynamics).
The overall radioactivity deposited in one hand and the radioactivity
in the ladder sequence on the other hand were determined by volume
integration of rectangles using the Image Quant software version 3.1
(Molecular Dynamics).
•
both irradiation conditions. It is likely that the OH radicals
Determination of the Quantum Yields. The quantum yield (QY)
for the formation of a product P may be defined in two ways.24
Convention 1: QYtot(P) ) number of products P per total number of
photons absorbed. Convention 2: QY(P) ) number of products P per
number of photons absorbed only by the precursors of P. Total means
photons absorbed by all molecules present in the solution. Here QYs
were determined according to convention 2, with one exception that
concerns the gel electrophoresis experiments, where QYtot values are
also reported. In the case of irradiation of optical thick samples (dAdo,
dGuo, and calf thymus DNA), the total number of incident and absorbed
photons are identical. To determine QY(8-oxo-dGuo) and QY(8-oxo-
dAdo), the contents of dGuo (22%) and dAdo (28%) in calf thymus
DNA were taken into account. In addition, a correction was made for
the difference in molar extinction of dAdo and dGuo, assuming by
analogy with free purine nucleosides, that ꢀdAdo ) 1.4ꢀdGuo. In the case
of irradiation of optically thin layers, the quantum yield QY is
practically impossible to be directly determined. Therefore, we
determined the quantum efficiency (QE(tot)), which is the amount of
photoproduct per (total) quanta absorbed under linear absorption
approximation (for the definition, see refs 25 and 26). For the
experiments dealing with the HPLC analysis of 37 bp DNA fragments,
the total number of absorbed photons under linear absorption ap-
proximation was determined using the steady-state absorbance value,
D ) 0.133, in the 0.2-cm irradiation path length of the Eppendorf tube.
Thus, the total number of photons absorbed by 1.83 µg of DNA was
2 × 1016, whereas the number of photons absorbed by the guanine
residues was 0.35 × 1016. For the experiments involving [32P]-labeled
DNA and gel electrophoresis analysis, the quantum efficiency was
calculated using the formula
which are involved in the formation of 8-oxo-dAdo and 8-oxo-
dGuo arise from water photosensitization.1,29 Interestingly,
several major changes were noted in the photochemistry of
naked DNA (Figure 3) by comparison with isolated nucleosides
(Figure 2) (see also Table 1). The quantum yields QYmax for
8-oxo-dGuo in DNA increase about 50 times (without MeOH)
and 200 times (with MeOH) compared to those observed for
free nucleosides. The quantum yield for the formation of 8-oxo-
dAdo in DNA is about 2-3 times lower than that for 8-oxo-
dGuo. Interestingly, within the experimental error range, the
•
presence of OH radical scavengers (either 2-5% MeOH or
tBuOH) has almost no influence on the formation of both 8-oxo-
dAdo and 8-oxo-dGuo in DNA. The saturation fluence of the
quantum yield, which is defined as the fluence corresponding
to 0.63QYmax, for 8-oxo-dGuo in DNA is almost 1 order of
magnitude lower than in free dGuo. In addition, QY(8-oxo-
dGuo) was found to be about 2.5 times lower in single-stranded
(heat-denatured) DNA than in double-stranded DNA (Table 1).
Generation of 8-Oxo-dGuo in 37 bp Duplex DNA. It was
found that the irradiation of DNA (1.83 µg) with two pulses (I
) 7.5 MW‚cm-2) leads to the formation of 5.7 ( 0.25 ng (or
1.2 × 1013 molecules) of 8-oxo-dGuo. Taking into account the
number of photons absorbed (see the Experimental Section),
the corresponding values for the quantum efficiency are QEtot(8-
oxo-dGuo) ) 6 × 10-4 (total photon absorbed) and QE(8-oxo-
dGuo) ) 3.4 × 10-3 (photons absorbed by guanine residues
only). Although the HPLC-electrochemical assay is very
sensitive, a relatively large amount of purified DNA fragments
(a few micrograms per analysis) is required for each determi-
nation. To obtain the intensity dependence of the quantum yield,
a more sensitive assay was used. This involved enzymatic
incubation of irradiated DNA fragments with the DNA repair
QEtot ) RL/(RTLσ1Etot)
(1)
where σ1 ) 2.3 × 10-17 cm2 is the absorption cross section of DNA
(22) Berger, M.; Anselmino, C.; Mouret, J.-F.; Cadet, J. J. Liq.
Chromatogr. 1990, 13, 929-940.
(23) Floyd, R. A.; Watson, J. J.; Wong, P. K.; Altmiller, D. H.; Rickard,
R. C. Free Rad. Res. Commun. 1986, 1, 163-172.
(24) Nikogosyan, D. N. Laser Chem. 1987, 7, 29-34.
(25) Kovalsky, O. I.; Budowsky, E. I. Photochem. Photobiol. 1990, 6,
659-665.
(26) Shaquiri, Z.; Keskinova, E.; Spassky, A.; Angelov, D. Photochem.
Photobiol. 1997, 65, 517-521.
(27) Nikogosyan, D. N.; Letokhov, V. S. RiV. NuoVo Cimento 1983, 6,
1-74.
(28) Nikogosyan, D. N.; Angelov, D. A.; Oraevsky, A. A. Photochem.
Photobiol. 1982, 35, 627-635.
(29) Oraevsky, A. A.; Nikogosyan; D. N. Chem. Phys. 1985, 100, 429-
445.

11376 J. Am. Chem. Soc., Vol. 119, No. 47, 1997
AngeloV et al.
Figure 2. Laser pulse intensity (dose) dependence of the quantum yield QY on the generation of 8-oxo-dAdo (a, c) and 8-oxo-dGua (b, d) under
nanosecond (a, b) and picosecond (c, d) laser photolysis of free dAdo and dGuo, respectively (optical thick layer, c ) 2 mM). Circles: water.
Squares: 5% methanol. Solid curves represent the best fits between the experimental data and eqs 16 and 22 taken from the ref 27.
Fpg protein followed by denaturing gel electrophoresis analysis
(Figure 4). The latter assay takes advantage of the ability for
the Fpg to recognize and quantitatively excise guanine lesions
including 8-oxo-7,8-dihydroguanine and the related formamido-
pyrimidine derivative.19,30 In addition, comparison of the present
data with those obtained by HPLC analysis allowed the
determination of the relative amount of 8-oxo-dGuo that is
recognized by the Fpg protein with respect to other putative
guanine lesions.
The 5′-labeled 9G oligonucleotide annealed with the comple-
mentary nonlabeled 4G strand was exposed to different laser
fluences. High-resolution sequencing gel analysis revealed that
only lesions located at guanine are cleaved by the Fpg repair
enzyme.31 The oxazolone compound, another major transfor-
mation product of the guanine radical cation,3,10 is not recog-
nized by the Fpg protein. However, the latter oxidative guanine
lesion which is piperidine labile is likely to be generated upon
UV laser photolysis of DNA.31 It should be remembered that
both Fpg-sensitive lesions and immediate strand breaks (sb’s)
are revealed simultaneously by denaturing gel analysis. There-
fore, the irradiated samples were split into two aliquots; the
first one was directly analyzed to establish QEtot(sb), whereas
the second aliquot was treated with the Fpg protein to determine
Qtot(Fpg+sb). Thus, the quantum efficiency of guanine lesions
QEtot(Fpg) recognized by the Fpg enzyme was obtained by
subtracting QEtot(sb) from QEtot(Fpg+sb). The determination
of QE(Fpg) (convention 2) implies taking into account the fact
that there are eight guanine residues among the 37 bases in the
ladder sequence. Consequently QE(Fpg) ) QEtot(Fpg) × 37/
8.
the curves that both immediate strand breaks together with Fpg-
sensitive lesions originate from biphotonic excitation. The solid
lines were obtained by least-squares approximation of the
experimental points using eq 2 (see the Discussion and refs 27
and 28). It was found that QE(Fpg) ) 3.5 × 10-3 ( 5% for a
fluence of 0.1 J‚cm-2
.
The value of the quantum efficiency (QE(Fpg) ) 2.4 × 10-3
)
for the opposite 4G strand after exposure to one pulse (E ) 0.1
J cm-2) was determined in a similar manner. Taking into
account these values, for the 9G/4G duplex, QE(Fpg) was found
to be 3.2 × 10-3
. The maximum values of the quantum
efficiencies extrapolated from the curves (QEmax) as well as the
values at E ) 0.1 J‚cm-2 are reported in Table 1. It is important
to note that QE(Fpg) and QE(8-oxo-dGuo), which were
determined by electrophoresis and HPLC measurements, re-
spectively, exhibit similar values within the experimental error
(( 5%) for each determination (Table 1). Therefore, it may
be concluded that the biphotonically generated guanine lesion
recognized by the Fpg protein is 8-oxo-dGuo. Consequently,
8-oxo-dGuo is the only major Fpg-sensitive lesions of the
guanine radical cation in double-stranded DNA.
It should be noted that no Fpg-sensitive sites were detected
at any adenine residue,31 although 8-oxo-dAdo is generated in
a relatively high yield in duplex DNA upon exposure to UV
laser irradiation. This is in agreement with previous results
which showed that 8-oxo-7,8-dihydroadenine is not a substrate
for the glycosylase activity of the Fpg protein.30
DNA Strand Breakage. The quantum efficiency of DNA
strand breaks was determined by direct gel electrophoresis
analysis (Figure 3). For the strand 9G it was established that
QE(sb) ) 4 × 10-4 (E ) 0.1 J‚cm-2) and QEmax(sb) ) 6.3 ×
10-4 by extrapolation of the curve. For the opposite 4G strand,
QE(sb) and QEmax(sb) were found to be 3.3 × 10-4 and 5.1 ×
10-4, respectively (data not shown). Thus, the averaged values
The fluence-response curves of the quantum efficiencies are
reported in Figure 5. It is obvious by considering the slopes of
(30) (a) Tchou, J.; Bodepudi, V.; Shibutani, S.; Antoshechkin, I.; Miller,
J.; Grollman; A. P.; Johnson, F. J. Biol. Chem. 1994, 269, 15318-15324.
(b) Ravanat, J.-L.; Berger, M.; Boiteux, S.; Laval, J.; Cadet, J. J. Chim.
Phys. 1993, 90, 871-879.
for the 9G/4G duplex are QE(sb) ) 3.6 × 10-4 (E ) 0.1 J‚cm-2
)
(31) Spassky, A.; Angelov, D. Biochemistry 1997, 36, 6571-6576.
and QEmax(sb) ) 5.7 × 10-4 (Table 1). It is important to note

UV Laser Photolysis of DNA and Related Compounds
J. Am. Chem. Soc., Vol. 119, No. 47, 1997 11377
only 25% of 8-oxo-dGuo and 8-oxo-dAdo are formed through
the intermediary of the corresponding radical cations. In fact,
the bulk of the two latter oxidized nucleosides (about 75%) is
likely to be generated through initial addition of •OH radical at
C8. It should be remembered that, under the experimental
•
conditions used (optically thick samples), generation of OH
radicals through direct two-photon absorption of water is
negligible upon both nanosecond and picosecond photolysis.1,27
•
Interestingly, the relative contribution of OH radicals to the
quantum yields was similar irrespective of pico- and nanosecond
irradiation conditions and of the laser intensity. These observa-
tions constitute additional arguments to support the hypothesis
that •OH radicals originate from biphotonic excitation of dAdo
and dGuo, but not from direct two-photon photolysis of water.
Generation of •OH radicals upon two-step excitation of organic
macrocyclic compounds to an energy level exceeding 6.5-6.8
eV was already established by spin-trapping34 and scavenging
experiments.29 This was presumably attributed to both ioniza-
tion and homolytic dissociation of photosensitized water mol-
ecules. Hydroxyl radicals are known to react efficiently with
purine DNA components giving rise to the corresponding
8-oxopurine nucleosides.3,23 It may be concluded that the
•
contribution of OH radicals to the biphotonic photolysis of
monomeric purine 2′-deoxyribonucleosides, in terms of quantum
yield, lies within the 10-5 range for nanosecond photolysis. An
about 5-fold higher value is noted for picosecond irradiation.
The chemical reactivity of radical cations does not depend
on how they are formed, and in particular of the pulse duration.
Therefore, the highest value of the saturated quantum yield
observed under picosecond photolysis is indicative of a higher
efficiency of formation of the purine radical cation under these
conditions. This demonstrates that biphotonic ionization of
purines either as free nucleosides or within DNA is more
efficient under picosecond excitation than upon nanosecond
irradiation. This is not surprising, since it is well established
that two-step excitation involves different intermediate states
according to the photolysis conditions used. S1 and T1 excited
states are generated upon exposure of purine components to
pico- and nanosecond UV laser pulses, respectively.27,28 An
important matter is to determine what is the proportion of purine
radical cations that is converted into 8-oxo-dGuo and 8-oxo-
dAdo. It should be noted that there are no available data on
the quantum yield of biphotonic ionization of monomeric dAdo
and dGuo. However, the lower limit for the quantum yield for
chromophore loss of free adenine and guanine upon nanosecond
Figure 3. Laser pulse intensity (dose) dependence of the quantum yield
QY on the generation of 8-oxo-dGuo upon nanosecond (a) and
picosecond (b) laser photolysis of calf thymus DNA (optical thick layer,
c ) 1 mM). The solid curves represent the best fits between the
experimental data and eqs 16 and 22 taken from the ref 27.
that the saturation fluence Es for 8-oxo-dGuo is about 2-fold
lower than that for the averaged value of quantum yield for
DNA strand breakage (see Figure 5).
Discussion
Monomeric Purine 2′-Deoxyribonucleosides. The two
major pathways of the initial transformation of the transient
purine radical cations in aqueous solutions at room temperature
is deprotonation and water addition.3,5,7 It was inferred from
time-resolved spectroscopy analysis that the transient radical
cations of dGuo and dAdo are mainly transformed through
deprotonation at neutral pH to the related oxidizing oxyl and
aminyl purine radicals, respectively.5,9,32 By analogy, this
behavior was often generalized to purine bases within single-
or double-stranded DNA, although direct experimental evidence
was not provided.18f,33 It should be noted that in contrast to
our present data, it was recently mentioned that exposure of
free dGuo to high-intensity 248 nm laser radiation leads to the
formation of 8-oxo-dGuo16 with a yield similar to that expected
in DNA. However, neither quantitative data for 8-oxo-dGuo
formation from dGuo nor quantum yield values for 8-oxo-dGuo
within DNA were reported. In the present study, it was found
that 8-oxo-dGuo and 8-oxo-dAdo which are the main hydration
products of the radical cation of dGuo and dAdo, respectively,
are generated as minor photoproducts with quantum yields in
the 10-5 range. The use of •OH radical scavenger showed that
laser photolysis was estimated to be about (1-2) × 10-3 28
.
Therefore, for free purine 2′-deoxyribonucleosides, the formation
of 8-oxo-dGuo and 8-oxo-dAdo represents only a few percent
of the overall decomposition products of the corresponding
radical cations. Since 8-oxo-dGuo and 8-oxo-dAdo are formed
through hydration of the related purine radical cation,3,7,33 it
may be concluded that this pathway is quite inefficient with
respect to the competitive deprotonation mechanism. This is
in agreement with previous observations which showed that the
formation of 8-oxo-dGuo is, at the best, a minor process upon
one-electron oxidation of dGuo by either type of photosensitizer7
•-
or SO4 inorganic radical.8 Under the former conditions, the
oxazolone nucleoside whose formation arises from the depro-
tonation of the dGuo radical cation10-12 is the major decomposi-
tion product.
Purine Lesions in DNA. In contrast to what was observed
for free dGuo, 8-oxo-dGuo is the predominant decomposition
product of the guanine radical cation within DNA. This is
(32) Candeias, L. P; Steenken, S. J. Am. Chem. Soc. 1992, 114, 699-
704.
(33) Candeias, L. P.; O’Neill, P.; Jones, G. D.; Steenken, S. Int. J. Radiat.
Biol. 1992, 61, 15-20.
(34) Gantchev, Ts.; Angelov, D. A.; Keskinova, E.; Kaltchev, M. Lasers
Life Sci. 1991, 4, 37-51.

11378 J. Am. Chem. Soc., Vol. 119, No. 47, 1997
AngeloV et al.
Table 1. Quantum Yields and Efficiencies of Purine Lesions and DNA Damage upon Nano- and Picosecond 266-nm Photolysis of DNA and
Purine 2′-Deoxynucleosides
substrate
dAdo
solvent (buffer)
H2O
lesion
lasera
QY (×10-4
)
QYmax (×10-4
)
method
8-oxo-dAdo
ns
ps
0.73
3.5
0.92
5.8
HPLC-EC
dGuo
dAdo
dGuo
H2O
8-oxo-dGuo
8-oxo-dAdo
8-oxo-dGuo
ns
ps
ns
ps
ns
ps
0.72
3.3
0.16
0.9
0.17
0.82
14
0.98
5.7
0.2
1.4
0.22
1.2
HPLC-EC
HPLC-EC
HPLC-EC
H2O + 10% MeOH
H2O + 10% MeOH
ds DNA
ds DNA
phosphate
phosphate
8-oxo-dAdo
8-oxo-dGuo
nsb
nsb
psc
nsb
nsd
nsd
nsd
HPLC-EC
HPLC-EC
35
45
180
150
14
ss DNA
phosphate
TE
TE
8-oxo-dGuo
8-oxo-dGuo
Fpg
HPLC-EC
HPLC-EC
PAGE
ds 9G/4G
ds 9G/4G
ds 9G/4G
34e
32e
38e
TE
ssb
3.6e
5.7e
PAGE
a The experimental error is typically less than 10%. Irradiation conditions: ns, τp ) 15 ns, E ) 0.64 J‚cm2; ps, τp ) 30 ps, E ) 0.24 J‚cm2; ns,b
τp ) 15 ns, E ) 0.1 J‚cm2; ps,c τp ) 30 ps, E ) 0.1 J‚cm2; ns,d τp ) 15 ns, E ) 0.1 J‚cm2. e QE (for the definition, see the Experimental Section.
Figure 4. Denaturing gel electrophoresis of 37 bp DNA (5′-end labeled
9G strand). Nonirradiated samples: lines 1 and 3. Exposed samples to
one pulse with an overall dose of 0.1 J‚cm-2: lines 2 and 4. Lines 1
and 2: non-Fpg-treated samples. Lines 3 and 4: Fpg-treated samples.
The two strong bands in the ladder sequence on line 4 belong to the
two GGG runs.
mostly inferred from a study in which riboflavin-mediated type
I photosensitization was shown to efficiently generate 8-oxo-
dGuo.7 It should be added that the oxazolone photoproduct is
also generated within DNA, but in a relatively lower yield than
Figure 5. Laser pulse dose dependence on the quantum efficiency:
(a) QEtot, total photon absorbed by DNA; (b) QE, photons absorbed
by guanines only. Under nanosecond laser photolysis of the 37 bp DNA
8-oxo-dGuo.3,10 However, the formation of both oxazolone and
8-oxo-dGuo has not been yet measured within DNA upon
exposure to either nano- or picosecond laser photolysis. The
(a and b) and free guanine (insert c). Single strand breaks: open circles.
observed increase in the yield of 8-oxo-dGuo in DNA cannot
Fpg-sensitive lesions together with single strand breaks: open squares.
be unambiguously attributed to only an increase in the efficiency
of the hydration pathway of the guanine radical cation. An
alternative explanation could be a more efficient one-electron
oxidation of the guanine residues within DNA. To answer these
questions, the quantum yield for the formation of 8-oxo-dGuo
in both calf thymus DNA and a 37 bp oligonucleotide was
determined by HPLC-EC and further compared with the results
obtained on free dGuo. Interestingly, a 100-fold increase in
the yield of 8-oxo-dGuo was observed in both DNA samples
as compared with free dGuo. It should be added that the value
of the quantum yield for 8-oxo-dGuo in DNA is similar to that
of ionization (hydrated electron generation) under the same pulse
Fpg-sensitive lesions only: open diamonds. Chromophore loss QEcl
of free guanine: insert c. The curve c was synthesized using the values
of æ1 and æ2σ2 from ref 28 and eq 2. The solid curves are the best fits
between the experimental points and eq 2.
that hydration of the guanine radical cation within DNA is a
predominant process in contrast to that was observed for free
dGuo. This is a direct demonstration of the crucial role played
by base stacking on the chemical transformation of the primary
guanine radical cation. However, an enhancement of Q(8-oxo-
dGuo), although less important, is also observed in single-
stranded DNA. This may be explained by the occurrence of a
partial base stacking. The high efficiency of hydration of the
guanine radical cation in DNA may be related to solvatation
effects. It is likely that the deprotonation of the guanine radical
cation in DNA is not as efficient as in free dGuo. Interestingly,
recent femtosecond time resolution studies showed that the
disappearance of the radical cations, at least for thymine, is likely
dose with Q(e-aq) ) 4.1 × 10-3 18c
.
A comparison of the yields
of 8-oxo-dGuo and oxazolone in the same 37 bp duplex and in
other synthetic DNA fragments showed that the former is in
average twice higher.31 Consequently, it may be concluded that
8-oxo-dGuo is a major decomposition product of the guanine
radical cation in double-stranded DNA. Therefore, it appears

UV Laser Photolysis of DNA and Related Compounds
J. Am. Chem. Soc., Vol. 119, No. 47, 1997 11379
to be completed in less than 1 ps.36 This would suggest that
the primary transformation processes are very fast and inac-
cessible to usual laser flash-photolysis techniques.
are the overall and the pulse irradiation doses, respectively,
expressed in photons‚cm-2. Taking into account both, the value
of æ1σ1 lying in the 5 × 10-19 to 5 × 10-20 cm2 range1,27,28
and the slope of the curves, approximated according to eq 2, it
was found that æ2σ2 . æ1σ1. In other words, the saturation of
the quantum efficiency is due to the saturation of the second
transition. With a good approximation, the following relations
were extrapolated from eq 1:
Another major observation deals with the efficient formation
of 8-oxo-dAdo in calf thymus DNA upon laser photolysis,
showing that guanine residues are not exclusive ultimate targets
of photoionization. This strongly suggests that the hole trapping
ability of guanines within DNA, as proposed on the basis of a
microsecond time-resolution spectroscopy analysis,18f,32,33 and
gel sequencing studies18a is, at least, overestimated. The
presently reported data show also that the primary chemical
reactions of the purine radical cations are strongly dependent
on the structure in which the purine moiety is inserted.
Consequently, microsecond time scale transient absorption
spectra corresponding to secondary radicals arising from photo-
ionization of either free nucleosides or single- or double-stranded
DNA are likely to represent different mixtures of intermediates.
It is noteworthy that conclusions based on such comparisons
should be considered with care.18f,32,33
QEmax ≈ Ræ1; Es ≈ 1.5/æ2σ2
(3)
Thus, it is clear that the saturation dose Es and the maximum
quantum efficiency of the radical cation QEmax are determined
by æ2 and æ1, respectively. Consequently, the lower value of
Es (0.045 J‚cm-2) that was observed for the average of guanines
in DNA (Figure 5b), as compared to isolated dGuo (0.3-0.35
J‚cm-2 28
(Figure 5c, insert), is due to a higher value of æ2,
)
assuming that σ2 does not change considerably. An important
decrease in the saturation intensity for the biphotonic photolysis
in double-stranded DNA as compared to single-stranded DNA
and free nucleobases was already observed under picosecond
laser photolysis.37 This was attributed to a cooperative excita-
tion due to energy migration over a rather long (150-200 bp)
distance. However, the present data for nanosecond photolysis
of the 37 bp DNA fragment are incompatible with such a
mechanism that should require a longer energy transfer distance
than the number of bases in the DNA fragment used. It is
noteworthy that the cumulative energy of the two absorbed
quanta largely exceeds the electron ejection threshold for all
four types of nucleobases1,18c and they are all readily ionizable.
The reason for the decrease in the saturation dose in DNA may
be attributed either to an enhanced value of the ionization
probability from highly excited state or to a positive charge
(hole) migration38 with a preferential trapping by guanine
residues. The consequence of the former process should be also
an enhancement of æ2 for the electron donating bases and a
concomitant decrease of æ2 for the electron accepting bases.
However, the considerable decrease of Es for guanine residues
in DNA, if explained only by hole transfer, should require a
longer hole migration distances than that available in the 37 bp
oligonucleotide. In addition, only a 2-fold higher Es value was
observed for guanine lesions as compared to the averaged value
for DNA strand breaks at all sites (see below and Figure 5).
Therefore, the present data are compatible with the involvement
of both effects: probably a short-range (1-3 base distance) hole
migration to guanines and an enhanced ionization ability of all
bases in DNA as compared to free nucleosides.
Another major observation was inferred from the comparative
quantitative study of the formation of 8-oxo-dGuo and Fpg-
sensitive sites within the 37 bp duplex DNA. It was found that
QE(8-oxo-dGuo) ) QE(Fpg). Therefore, it may be concluded
that the overwhelming transformation product of the guanine
radical cation within DNA that is cleaved by the Fpg protein is
8-oxo-dGuo. The formation of other lesions from the guanine
radical cation, including the formamidopyrimidine derivative
of guanine and apurinic sites which may be also recognized by
the Fpg protein,19 should be at the best minor processes. Thus,
Fpg treatment followed by gel electrophoresis constitutes an
almost quantitative tool for determining 8-oxo-d-Guo in photo-
ionized DNA.
The saturation intensity in DNA is about 1 order of magnitude
lower than in monomeric 2′-deoxyguanosine. The slope of the
intensity response curve of the quantum efficiency for nano-
second biphotonic photolysis depends on the laser pulse dose
(energy per unit area). However, this is independent of the laser
pulse shape.27,28 Therefore, it is preferable to analyze the
saturation characteristics in terms of saturation pulse dose (Es),
corresponding to QEs ) 0.63QEmax,instead of saturation inten-
sity. It is important to note that there is an almost 10-fold
decrease of Es in the case of DNA, (Figures 2 and 3). The
discussion will be restricted to the data for nanosecond
photolysis in optical thin layer which are more easily to be
processed (see Figure 5 and refs 27 and 28). The relationship
between the quantum efficiency of a biphotonic photochemical
process and the laser fluence is given by the expression
R
σ1Etot
e-æ σ E
e-æ σ E
1
1
2 2
QE )
1 -
-
(2)
æ1σ1
æ2σ2
DNA Strand Breakage. The quantum efficiency for strand
breakage (QE(sb) ) 3.6 × 10-4) in the 37 bp DNA duplex
upon exposure to a 0.1 J‚cm-2 fluence was inferred from
electrophoresis analysis. In addition, the value for QEmax(sb)
was estimated to be 5.7 × 10-4. The literature is abundant with
values of quantum yields (efficiencies) for strand breakage of
nucleic acids under nanosecond laser photolysis which are
scattered within the 10-5-10-4 range:
1 -
1 -
[
]
æ2σ2
æ1σ1
where R is the probability of the radical cation to form the
product (R < 1) and σ1 and σ2 are the absorption cross sections
of the ground singlet S0 and excited triplet T1 states, respectively.
In addition, æ1 and æ2 are the quantum yields for T1 excitation
and for photoionization from T1, respectively. In the absence
of energy and charge-transfer processes, æ1 coincides with the
intersystem crossing yield whereas æ2 corresponds to the
ionization probability from the highly excited state. Etot and E
(37) Nikogosyan, D. N.; Oraevsky, A. A.; Letokhov, V. S. Chem. Phys.
1985, 97, 31-34.
(38) (a) Candeias, L. P.; Steenken, S. J. Am. Chem. Soc. 1993, 115,
2437-2440. (b) Hall, D. B.; Holmlin, R. E.; Barton, J. K. Nature 1996,
382, 731-735. (c) Arkin, M. R.; Stemp, E. D. A.; Holmlin, R. E.; Barton,
J. K.; Ho¨rmann, A.; Olson, E. J. C.; Barbara, P. F. Science 1996, 273, 475-
480. (d) Arkin, M. R.; Stemp, E. D. A.; Turro, C.; Turro, N. J.; Barton, J.
K. J. Am. Chem. Soc. 1996, 118, 2267-2274.
(35) Cadet, J.; Berger, M., Buchko, G. W.; Ravanat, J.-L.; Kasai, H. In
Frontiers in Photobiology; Shima, A., Ichahashi, M., Fujiwara, Y., Takebe,
H., Eds.; Elsevier Science: Amsterdam, 1993; pp 49-54.
(36) Reuther, A.; Nikogosyan, D.; Laubereau, A. J. Chem. Phys. 1996,
100, 5570-5577.

11380 J. Am. Chem. Soc., Vol. 119, No. 47, 1997
AngeloV et al.
(i) plasmid assay: 8 × 10-6 39
;
4.5 × 10-4 17
;
1-2 × 10-4 18d
.
Conclusions
(ii) analytical centrifugation: 1.7 × 10-4 in TMV RNA.40
Evidence is provided in this work that exposure of calf thymus
DNA and a synthetic duplex oligonucleotide (37-mer) to nano-
and picosecond UV laser pulses gives rise to 8-oxo-7,8-dihydro-
2′-deoxyguanosine and to a lesser extent to 8-oxo-7,8-dihydro-
2′-deoxyadenosine. However, it should be noted that the two
latter purine oxidation products are generated from the corre-
sponding 2′-deoxyribonucleosides in a yield which is about 2
orders of magnitude lower than in DNA. This illustrates the
preponderant role played by the DNA conformation features
and more likely base stacking in promoting the hydroxylation
of the C8 position of both purine moieties. A reasonable
mechanism for the formation of both oxidation products is initial
one-electron oxidation of the purine rings which is followed at
least partly by hydration. The significant formation of 8-oxo-
dAdo within double-stranded DNA would indicate that adenine
may be a secondary sink for positive hole transfer according to
the sequence context. Work is in progress by using defined
sequence oligonucleotides to further evaluate this possibility.
(iii) low-angle light scattering: 6 × 10-5 at E ) 0.1 J‚cm-2
2,18c
in double-stranded and single-stranded calf thymus DNA.
The low yield of strand breakage and the lack of significant
base specificity are characteristic of biphotonic photolysis of
DNA.15,16,31 However, somewhat higher QE(sb) values are
observed using denaturing gel electrophoresis analysis. There-
fore, it may be anticipated that, under the latter conditions, heat
labile base lesions may lead to the formation of additional DNA
strand breaks. Consequently, the values of QE(sb) presently
reported should be considered as a higher estimate, since they
could be attributed not only to immediate strand breaks but also
to heat unstable base lesions. In this respect it should be
reminded that direct frank breaks were found to be generated
at the best in very low yields within DNA upon exposure to
either 248-nm16,18d or 193-nm18h laser pulses. Finally, the
•
involvement of OH radicals under nanosecond biphotonic
photolysis, although restricted within the 10-5 quantum yield
range, was unambiguously demonstrated. The highly oxygen
reactive species are likely to explain, at least partially, the
formation of frank DNA strand breaks. However, it should be
Acknowledgment. The work was supported by the National
Science Foundation (Bulgaria) (Contract F-513), French Min-
istry of Science and Research (ACC-SV No. 8-MESR 1995),
and the CNRS (Bulgarian Academy of Sciences Collaboration
Program III.1). D.A. acknowledges the University of Paris VI
for providing a PAST fellowship. This paper is dedicated to
Professor Waldemar Adam on the occasion of his 6oth birthday.
•
noted that addition of OH scavengers was not found to exert
any significant protection against strand cleavage.2,18c This may
be explained, at least partly, by proximity effects.
(39) Zavilgelsky, G. B.; Gurzadian, G. G.; Nikogosyan, D. Photochem.
Photobiol. 1984, 38, 175-187.
(40) Dobdov, E. I.; Arbieva, Z. Kh.; Timofeeva, E. K.; Esenaliev, E.
O.; Oraevsky, A. A.; Nikogosyan, D. N. Photochem. Photobiol. 1989, 49,
595-598.
JA971728R