Reactions of psoralen radical cations with biological substratest

Article abstract of DOI:10.1562/0031-8655(2000)072<0155:ROPRCW>2.0.CO;2

The reactions of several psoralen and coumarin radical cations with biological substrates such as nucleotides, amino acids and alkenes that serve as models for unsaturated fatty acids have been examined. The radical cations were generated by laser photoionization of the parent psoralen or coumarin in aqueous buffer in most cases. Easily oxidized substrates such as tyrosine, tryptophan and guanosine monophosphate react with the 8-methoxypsoralen and several methoxy-substituted coumarin radical cations with rate constants in excess of 2 × 10⁹ M^-1 s^-1. In each case reaction occurs via electron transfer, as demonstrated by the observation of quencher-derived radical cations or radicals by transient absorption spectroscopy. For other substrates such as histidine, methionine and adenosine monophosphate the measured rate constants are significantly slower and vary with the oxidation potential of both the parent psoralen or coumarin and the quencher, again indicative of electron transfer reactivity. Most of the alkenes studied also react with the psoralen or coumarin radical cations via electron transfer, although there is some evidence for addition for linoleic acid. Product studies carried out using both lamp and laser irradiation in the presence of deoxyguanosine as a radical cation trap lead to the formation of characteristic base-derived Type-I (electron transfer) products. This lends support to our previous hypothesis that photoionization occurs via a monophotonic process and is thus relevant to conditions used in clinical phototherapeutic applications of psoralens. The results demonstrate the relevance of electron transfer chemistry to the use of psoralens and related compounds as photoactivated drugs.

Full text of DOI:10.1562/0031-8655(2000)072<0155:ROPRCW>2.0.CO;2

Reactions of Psoralen Radical Cations with Biological Substrates¶
Author(s): Paul D. Wood, Anisa Mnyusiwalla, Lie Chen, and Linda J. Johnston
Source: Photochemistry and Photobiology, 72(2):155-162.
Published By: American Society for Photobiology
https://doi.org/10.1562/0031-8655(2000)072<0155:ROPRCW>2.0.CO;2
URL: http://www.bioone.org/doi/full/10.1562/0031-8655%282000%29072%3C0155%3AROPRCW
%3E2.0.CO%3B2
BioOne (www.bioone.org) is a nonprofit, online aggregation of core research in the biological, ecological, and
environmental sciences. BioOne provides a sustainable online platform for over 170 journals and books published
by nonprofit societies, associations, museums, institutions, and presses.
Your use of this PDF, the BioOne Web site, and all posted and associated content indicates your acceptance of
BioOne’s Terms of Use, available at www.bioone.org/page/terms_of_use.
Usage of BioOne content is strictly limited to personal, educational, and non-commercial use. Commercial inquiries
or rights and permissions requests should be directed to the individual publisher as copyright holder.
BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers, academic institutions, research
libraries, and research funders in the common goal of maximizing access to critical research.

Photochemistry and Photobiology, 2000, 72(2): 155–162
Reactions of Psoralen Radical Cations with Biological Substrates†¶
Paul D. Wood, Anisa Mnyusiwalla, Lie Chen and Linda J. Johnston*
Steacie Institute for Molecular Sciences, National Research Council Canada, Ottawa, ON, Canada
Received 27 March 2000; accepted 24 April 2000
ABSTRACT
for the treatment of dermatological disorders (1–5). More
recently there has been substantial interest in the develop-
ment of new psoralen derivatives for other therapeutic ap-
plications such as ex vivo treatment of cutaneous T-cell lym-
phoma and virus inactivation for purification of blood prod-
ucts. The phototherapeutic effects of psoralens such as the
commonly used 8-methoxypsoralen (8-MOP) are believed to
result from intercalation of the drug between adjacent base
pairs in the DNA duplex followed by two successive pho-
tocycloaddition reactions that generate a psoralen diadduct
that crosslinks the two DNA strands. The latter leads to in-
hibition of replication and hence cell proliferation. Despite
widespread clinical application of PUVA therapy, there is
evidence that its long-term use may be associated with a
variety of harmful side effects, including carginogenicity,
mutagenicity and immune system modulation (2,6,7).
In addition to their DNA-specific chemistry, psoralens and
related compounds have been shown to generate singlet ox-
ygen (8–12), to bind to proteins (13–17) and to undergo
photocycloaddition reactions with unsaturated fatty acids
(18,19). DNA adducts in which the psoralen is attached to
the sugar rather than to the base have also been characterized
(20). Thus it is clear that the UVA irradiation of psoralens
can initiate a complex series of reactions that depend not
only on the overall reactivity of the psoralen excited state
but also on the initial site of localization of the psoralen in
the cell (e.g. membrane solubilized vs DNA-intercalated).
Understanding the basis for either the phototherapeutic or
harmful effects of PUVA therapy will require a detailed ex-
amination of the overall reactivity of the psoralen excited
state and the reactive species formed from it as well as
knowledge of the localization of the substrate. In this context
it should be noted that the photophysical properties of pso-
ralens and coumarins are strongly medium dependent
(21,22). Further it has been shown that approximately 80%
of bound 8-MOP detected after UVA irradiation of mouse
skin is localized in protein and lipid fractions (23).
The reactions of several psoralen and coumarin radical
cations with biological substrates such as nucleotides,
amino acids and alkenes that serve as models for unsat-
urated fatty acids have been examined. The radical cat-
ions were generated by laser photoionization of the par-
ent psoralen or coumarin in aqueous buffer in most cas-
es. Easily oxidized substrates such as tyrosine, trypto-
phan and guanosine monophosphate react with the
8-methoxypsoralen and several methoxy-substituted cou-
marin radical cations with rate constants in excess of 2
Ϫ
1
Ϫ1
؋
109 M s . In each case reaction occurs via electron
transfer, as demonstrated by the observation of quench-
er-derived radical cations or radicals by transient ab-
sorption spectroscopy. For other substrates such as his-
tidine, methionine and adenosine monophosphate the
measured rate constants are significantly slower and vary
with the oxidation potential of both the parent psoralen
or coumarin and the quencher, again indicative of elec-
tron transfer reactivity. Most of the alkenes studied also
react with the psoralen or coumarin radical cations via
electron transfer, although there is some evidence for ad-
dition for linoleic acid. Product studies carried out using
both lamp and laser irradiation in the presence of deox-
yguanosine as a radical cation trap lead to the formation
of characteristic base-derived Type-I (electron transfer)
products. This lends support to our previous hypothesis
that photoionization occurs via a monophotonic process
and is thus relevant to conditions used in clinical pho-
totherapeutic applications of psoralens. The results dem-
onstrate the relevance of electron transfer chemistry to
the use of psoralens and related compounds as photoac-
tivated drugs.
INTRODUCTION
The combination of psoralens and UVA irradiation (PUVA
therapy)‡ has been widely used over the last two decades
Recent work from our laboratory (24,25) and from other
groups (26–31) has investigated the potential role of electron
transfer chemistry, and in particular radical cations, in pso-
ralen photochemistry. We have demonstrated that the 8-
¶Posted on the web on 11 July 2000.
†Issued as NRCC 43838.
*To whom correspondence should be addressed at: Steacie Institute
for Molecular Sciences, National Research Council Canada, Ot-
tawa, ON K1A 0R6 Canada. Fax: 613-952-0068;
e-mail: linda.johnston@nrc.ca
deoxyguanosine; GMP, guanosine 5Ј-monophosphate; HPLC,
high-performance liquid chromatography; LC/MS, liquid chro-
matography/mass spectrometry; 7-MC, 7-methoxycoumarin; 8-
MOP, 8-methoxypsoralen; PUVA, psoralen plus UVA irradiation;
‡Abbreviations: AMP, adenosine 5Ј-monophosphate; CMP, cytidine
5Ј-monophosphate; 6,7-DMC, 6,7-dimethoxycoumarin; dGuo,
SCE, saturated calomel electrode; TMP, thymidine 5
Ј-monophos-
᭧
2000 American Society for Photobiology 0031-8655/00
$5.00
ϩ0.00
phate; 4,5 ,8-TMP, 4,5 ,8-trimethylpsoralen.
Ј
Ј
155

156 Paul D. Wood et al.
MOP radical cation is generated via 355 nm laser photoion-
ization in aqueous solution under conditions that are com-
parable to those employed in PUVA therapy (24). The rad-
ical cation yield showed a linear dependence on the laser
energy, indicating that photoionization was a monophotonic
process with a quantum yield of 0.015. This raises the pos-
sibility that the direct formation of radical cations could play
a role in PUVA therapy, particularly since we have also
•ϩ
shown that 8-MOP reacts with guanosine monophosphate
Ϫ
1
Ϫ1
(GMP) with a rate constant of 2.5
ϫ
109 M
s
to yield a
base-derived radical cation. We subsequently demonstrated
that a number of structurally related psoralens and coumarins
undergo photoionization with varying efficiencies and were
able to rationalize these differences on the basis of the oxi-
dation potentials and excited singlet state properties of the
substrates (25). Of particular interest is the fact that several
dimethoxycoumarins undergo reasonably efficient photoion-
ization (quantum yields between 0.1 and 0.2) in aqueous
buffer.
The initial studies described above have prompted us to
carry out a number of additional experiments that are aimed
at providing further evidence for the role of electron transfer
chemistry in PUVA therapy. Firstly we have examined the
reactivity of several representative psoralen and coumarin
radical cations with biological substrates such as nucleotides,
nucleosides, amino acids and models of membrane compo-
nents. The results demonstrate that psoralen and coumarin
radical cations react rapidly with a range of substrates, pri-
marily via electron transfer to regenerate the initial psoralen
or coumarin and yield products resulting from the one-elec-
tron oxidized quencher. This suggests that radical cations
formed via either photoionization or photoinduced electron
transfer initiated by endogenous photosensitizers will play a
role in PUVA therapy. Secondly, we have carried out laser
and lamp product studies that provide further evidence for
the monophotonic nature of psoralen and coumarin photo-
ionization.
Figure 1. Transient absorption spectra obtained following direct
excitation at 355 nm of oxygen-purged aqueous phosphate buffer
(pH
3.9
and 2.1
ϭ
␮
7) containing 8-MOP (top, delays of 0.2
s [ ] and 6.8 s [ ]) and 7-MC (bottom, delays of 0.5
s [ ]).
␮
s [
●
], 0.9
␮
␮
s [
s [
⅙
●]
],
Ⅲ
␮
␮
▫
ࡗ
those of authentic mixtures prepared by irradiation of benzophenone
and dGuo under similar conditions. Liquid chromatography/mass
spectrometry (LC/MS) data were recorded on a VQ Quatro elec-
trospray mass spectrometer coupled to a HP 1090 liquid chromato-
graph.
RESULTS
Reactivity of psoralen and coumarin radical cations
MATERIALS AND METHODS
The photochemical generation and characterization of pso-
ralen and coumarin radical cations using nanosecond flash
photolysis have been reported in detail (24,25). For most of
the experiments herein, the radical cations were generated
by 355 nm laser excitation of the substrate in aqueous phos-
phate buffer at pH 7. In several cases either the quencher or
substrate was insufficiently soluble in aqueous buffer and the
experiments were carried out in aqueous acetonitrile. As typ-
ical examples, Fig. 1 shows the spectra obtained following
355 nm excitation of 8-MOP and 7-methoxycoumarin (7-
MC) in oxygen-saturated aqueous solution. In both cases the
spectra show a characteristic band between 550 and 650 nm
due to the radical cation. The spectra at early time show
Chemicals. All of the compounds employed were commercial sam-
ples (Sigma-Aldrich, Oakville, ON, Canada) and were used as re-
ceived, unless otherwise noted. Chloranil was recrystallized from
ethanol and dienes were purified by passage through a short alumina
column before use. All solvents were of the highest available com-
mercial grade (Omnisolv).
Laser flash photolysis. The nanosecond laser system has been
described in detail elsewhere (32). The excitation source was either
a Lumonics HY750 Nd:YAG laser (355 nm, 10 ns/40 mJ pulses) or
a Nd:YAG–pumped Lumonics HD500 dye laser (420 nm, 10 ns/12
mJ pulses). Samples were contained in a 7
were prepared with absorbances of 0.45
ϫ
Ϯ
7 mm2 quartz cell and
0.05 (typically 0.2 mM
psoralen or coumarin) at the excitation wavelength unless otherwise
stated. Solutions were prepared in either aqueous potassium phos-
phate buffer (pH 7.0), acetonitrile or 1:1 mixtures of the two. UV–
visible absorption spectra were recorded on a Varian CARY 3 spec-
trophotometer. Unless otherwise noted, solutions were purged with
oxygen for a minimum of 2 min per mL before laser irradiation.
Product studies. Samples of 6,7-dimethoxycoumarin (6,7-DMC)
in aqueous buffer were irradiated both in the presence and absence
of deoxyguanosine (dGuo) (4–5 mM) using either 350 nm Rayonette
lamps or 355 nm laser excitation. Irradiated samples were analyzed
by high-performance liquid chromatography (HPLC) on a Hewlett–
Packard HP 1090 liquid chromatograph using a reverse phase (C18)
column and eluting with methanol/buffer mixtures. Products were
identified by comparison of retention times and mass spec data with
additional absorptions at
ϳ500 nm and below 400 nm for 7-
MC and 8-MOP, respectively, that are assigned to triplet
absorptions (24). Note that the radical cations also have an
absorption band below 400 nm that is clearly visible at lon-
ger delays. Oxygen-purged samples were used for most ex-
periments to minimize contributions to the spectrum from
either triplet or solvated electron signals (Ͼ700 nm).
Previous results have shown that the 8-MOP radical cation
reacts with GMP with a rate constant that approaches the

Photochemistry and Photobiology, 2000, 71(2) 157
Table 1. Rate constants for reaction of psoralen and coumarin radical cations with nucleotides in aqueous phosphate buffer at pH 7.
Oxidation potentials for the substrates are given in parentheses (V vs SCE, in acetonitrile [25])
1
1
kq (MϪ sϪ
)
7-MC
(1.91)
5,7-DMC
(1.73)
8-MOP
(1.67)
6,7-DMC
(1.53)
Nucleotide
GMP
AMP
CMP
TMP
Guanosine
dGuO
4.1
3.3
ϫ
ϫ
ϫ
ϫ
ϫ
—
109
2.9
1.4
ϫ
ϫ
ϫ
ϫ
—
—
109
108
107
107
2.5
3.4
ϫ
ϫ
ϫ
ϫ
ϫ
109
107
107
107
109
2.5
ϫ
ϫ
ϫ
109
107
107
109
107
107
109
Ͻ
Ͻ
1
1
Ͻ
Ͻ
2
1
Ͻ1
1
Ͻ1.5
Ͻ
Ͻ1.5
Ͻ ϫ 107
2.5
2.8
2.6
2.8
ϫ
ϫ
109
109
—
diffusion controlled limit (24). By contrast, adenosine mono-
phosphate (AMP) reacted approximately two orders of mag-
nitude more slowly and the pyrimidine mononucleotides re-
stants were the same within experimental error for both the
nucleotide and the nucleoside. Although several other pso-
ralens are not sufficiently soluble in aqueous solution for
Ϫ
1
Ϫ1
acted too slowly (
Ͻ
107 M s ) to be measured under our
photoionization experiments, the radical cation of 4,5
Ј,8-tri-
conditions. The decreased reactivity with increasing oxida-
tion potential of the bases and the transient spectra obtained
in the presence of GMP indicated that reaction is via electron
transfer to regenerate the psoralen radical cation and give
the based derived radical cation which then deprotonates
(Eq. 1)
methylpsoralen (4,5 ,8-TMP) can be generated in 1:1 aque-
Ј
ous acetonitrile; for comparison with the data presented in
Table 1, this radical cation reacts with GMP with a similar
Ϫ
1
Ϫ1
rate constant (4.2
ϫ
109 M s ).
kobs k0 kq[Q]
ϭ
ϩ
(2)
GMP
Ϫ
Hϩ
Spectra measured after reaction of the 6,7-DMC radical
cation with GMP are shown in Fig. 2. Quenching of the
initial radical cation (shown in the insert of Fig. 2) is ac-
companied by the formation of a weakly absorbing transient
in the 300–400 nm region and a tailing absorption that ex-
tends into the visible. These results are similar to those ob-
tained earlier for 8-MOP, and are consistent with an electron
transfer mechanism to generate a guanosine-derived radical
(33). The guanosine radical cation has a pKa of 3.9 and de-
•ϩ
•ϩ
•
8-MOP
→
GMP
→
GMP
(1)
The reaction of three methoxy-substituted coumarin radi-
cal cations with the same four mononucleotides has now
been examined. The rate constants (kq) were obtained from
the slopes of plots of rate constants for decay (kobs) of the
radical cations at their respective maxima as a function of
the concentration of quencher (Eq. 2). The data are sum-
•ϩ
marized in Table 1. As observed previously for 8-MOP ,
Ϫ
1
protonates with a rate constant of
ϳ3 ϫ
105 s in neutral
each of the radical cations was unreactive toward the pyrim-
idines. The rate constants for GMP were all in excess of 2
aqueous solution (34,35). Spectra measured after quenching
Ϫ
1
Ϫ1
109 M
s
whereas those for AMP ranged from 3.3
ϫ
•ϩ
ϫ
of 6,7-DMC with AMP did not indicate the formation of
Ϫ
1
Ϫ
1
Ϫ
1
Ϫ1
109 M
s
to
Ͻ
1
ϫ
107 M
5,7-DMC
For comparison purposes, rate constants for reaction of both
s
and decreased in the fol-
any additional transients between 340–700 nm.
lowing order: 7-MC
Ͼ
Ͼ 8-MOP Ͼ 6,7-DMC.
In a similar set of experiments to those described above,
the reactivity of several psoralen and coumarin radical cat-
ions with a series of amino acids was measured; the data are
compiled in Table 2. The easily oxidized amino acids such
as tryptophan and tyrosine reacted with each radical cation
•ϩ
•ϩ
8-MOP and 6,7-DMC were also measured with guanosine
and/or dGuo in aqueous solution. In each case the rate con-
Table 2. Rate constants for reaction of psoralen and coumarin rad-
ical cations with amino acids in aqueous phosphate buffer at pH 7
1
1
kq (MϪ sϪ
)
Amino acid*
7-MC
8-MOP
6,7-DMC
Tyrosine
Tryptophan
Cysteine†
3.2
ϫ
ϫ
ϫ
ϫ
ϫ
ϫ
ϫ
109
3.1
2.1
1.7
7.1
9.7
ϫ
ϫ
ϫ
ϫ
ϫ
ϫ
ϫ
109
2.1
3.0
1.2
1.0
ϫ
ϫ
ϫ
ϫ
ϫ
ϫ
ϫ
109
109
108
107
107
107
109
4.2
3.1
1.4
2.2
109
108
109
109
107
109
109
108
107
106
107
109
Histidine
Methionine
Phenylalanine
Urocanic acid
Ͻ
Ͻ
1
1
Ͻ2
Ͻ1
4.5
5.4
1.5
1
1
Figure 2. Transient absorption spectra obtained following direct
* Upper limits of Ͻ1 ϫ
107 MϪ sϪ were estimated for reaction of
excitation at 355 nm of oxygen-purged aqueous phosphate buffer
both 8-MOP and 6,7-DMC radical cations with serine, threonine,
arginine and lysine.
3
(pH
of 3.8
shows the spectrum of the 6,7-DMC radical cation in aqueous buffer.
ϭ
7) containing 6,7-DMC and GMP (5.0
ϫ
10Ϫ M) at delays
␮
s ( ), 8.6 s ( ), 15.8 s ( ) and 35.0
⅜
␮
ⅷ
␮
᭝
␮
s (
᭡). The insert
† Reaction was pH dependent; the observed rate constant for reac-
1
1
tion of 7-MC with cysteine at pH 8 is 1.1
ϫ .
109 MϪ sϪ

158 Paul D. Wood et al.
Figure 4. Transient absorption spectra obtained following direct
excitation at 355 nm of oxygen-purged aqueous phosphate buffer
(pH
of 3.6
ϭ
7) containing 6,7-DMC and tyrosine (2.7
s ( ), 11.6 s ( ) and 26.8 s ( ).
ϫ
10Ϫ3 M) at delays
␮
⅜
␮
ⅷ
␮
᭡
is consistent with the reported spectrum for the tryptophan
radical cation that deprotonates with a rate constant of 1
ϫ
Ϫ
1
Ϫ1
106 M
s
in neutral aqueous solution (37,38). The spec-
trum measured in neutral solution is in good agreement with
the spectrum of the tryptophan radical with at 520 nm
␭
max
•ϩ
(37,38). Spectra measured after reaction of 6,7-DMC with
tyrosine are also consistent with an electron transfer reaction
as shown in Fig. 4. Quenching of the radical cation generates
Figure 3. (a) Transient absorption spectra obtained following direct
a characteristic phenoxyl-radical spectrum with
␭
at 415
max
excitation at 355 nm of oxygen-purged aqueous phosphate buffer
nm (39,40). Deprotonation of the initial tyrosine radical cat-
ion is expected to be rapid on the timescale of our experi-
ments in aqueous solution since measured rate constants for
deprotonation of various substituted phenol radical cations
3
(pH
delays of 4.4
Transient absorption spectra obtained following direct excitation at
355 nm of oxygen-purged aqueous phosphate buffer (pH 7) con-
10Ϫ M) and 5% trifluoracetic
ϭ
7) containing 6,7-DMC and tryptophan (4.0
ϫ
10Ϫ M) at
␮
s ( ), 12.4 s ( ), 34.8 s ( ) and 69.2
⅜
␮
ⅷ
␮
᭝
␮s (᭡). (b)
ϭ
3
Ϫ
1
Ϫ1
by water are in the range of 108 M
s
(41). Spectra mea-
taining 6,7-DMC, tryptophan (4.0
acid at delays of 4.4 s ( ), 11.6
).
ϫ
␮
⅜
␮s (
ⅷ), 30.0
␮
s (᭝) and 68.2 ␮s
sured in the presence of histidine, urocanic acid or cysteine
did not show any evidence for formation of transients in the
350–700 nm region.
(᭡
The reactivity of the same radical cations toward a number
of alkenes that serve as models for unsaturated lipid chains
in bilayer membranes was also examined. Some of these
reactions were studied in acetonitrile due to solubility limi-
tations in aqueous solution. In these cases the radical cations
were generated by photosensitized electron transfer (420 nm
excitation) using triplet chloranil as described previously
(25). The data for several alkenes and dienes are summarized
in Table 3 and demonstrate that, as expected on the basis of
the results described above, the radical cations react with
easily oxidized substrates such as 1,4-diphenylbutadiene,
2,5-dimethyl-2,4-hexadiene and 2,3-dimethyl-2-butene with
rate constants that are at or close to the diffusion controlled
limit. 8,10-Dodecadienol and sorbic acid reacted consider-
ably more slowly with rate constants following the trend (7-
Ϫ
1
Ϫ1
with rate constants in excess of 2
ϫ
109 M
s
in aqueous
buffer. By contrast, rate constants for histidine ranged from
•ϩ
Ϫ
1
Ϫ
1
•ϩ
1.4
ϫ
109 for 7-MC to 1.0
ϫ
107 M
s
for 6,7-DMC .
Rate constants for cysteine were also significantly less than
diffusion controlled, although in this case there was little
dependence on the radical cation structure. However, it is
likely that the measured rate constants reflect reaction with
both cysteine and its thiolate anion, given that the pKa for
cysteine is 8.37 (36). Consistent with this the measured rate
constant for 7-MC increases substantially when the pH is
Ϫ
1
Ϫ1
increased (3.1
ϫ ϫ
108 vs 1.1 109 M s at pH 7 and 8).
No change in the kinetics for radical cation decay was ob-
served for other amino acids such as phenylalanine, serine,
threonine, arginine and lysine at the highest concentrations
used in our experiments (generally the maximum solubility);
in these cases one can only estimate an upper limit for the
quenching rate constant. Urocanic acid, a histidine metabo-
lite, also reacted with rate constants that approached the dif-
fusion controlled limit.
Spectra were measured in the presence of a number of
amino acid quenchers in order to provide evidence for the
quenching mechanism. For example, Fig. 3 shows spectra
obtained after reaction of 6,7-DMC with tryptophan in
aqueous buffer (a, pH 7) and in the presence of 5% trifluo-
MC
Ͼ 8-MOP Ͼ 6,7-DMC) observed above with amino
acids and nucleotides. By contrast, linoleic acid, a noncon-
jugated diene, and a similar monoalkene, oleic acid, reacted
•ϩ
•ϩ
with both 8-MOP and 7-MC with similar rate constants
Ϫ
1
Ϫ1
(ϳ6 ϫ
107 M s ).
Product studies
•ϩ
6,7-DMC was irradiated under a variety of conditions using
both lamp and laser irradiation in order to provide further
support for our earlier hypothesis of monophotonic photo-
roacetic acid (b). The latter spectrum has ␭max at 570 nm and

Photochemistry and Photobiology, 2000, 71(2) 159
Table 3. Rate constants for reaction of psoralen and coumarin radical cations with alkenes in acetonitrile, aqueous phosphate buffer or a
1:1 mixture of the two
1
1
kq (MϪ sϪ
)
Alkene*
Solvent
CH3CN
CH3CN
CH3CN
Aq buffer
Aq CH3CN
Aq CH3CN
Aq CH3CN
7-MC
—
8-MOP
6,7-DMC
1,4-Diphenylbutadiene
2,5-Dimethyl-2,4-hexadiene
2,3-Dimethylbutene
8,10-Dodecadienol
Sorbic acid
2.0
ϫ
ϫ
ϫ
ϫ
ϫ
ϫ
ϫ
1010
1.4
2.6
ϫ
ϫ
—
ϫ
ϫ
ϫ
—
1010
109
1.3
ϫ
ϫ
ϫ
ϫ
ϫ
ϫ
1010
3.9
4.4
1
109
109
107
106
107
107
4.7
1.4
4.6
6.9
4.7
109
109
108
107
107
ϳ7
Ͻ5
Ͻ5
106
106
106
Ͻ5
Linoleic acid
Oleic acid
6.5
5.2
* Oxidation potentials of 1.56 and 1.33 (V vs SCE in acetonitrile) have been reported for 2,3-dimethylbutene and 2,5-dimethyl-2,4-hexadiene
(53). The oxidation potential for 8,10-dodecadienol should be similar to that of 2,4-hexadiene (1.82 V), while sorbic acid will be slightly
higher (0.1–0.2 V) as a result of the additional carboxylic acid substituent. Linoleic and oleic acids will be similar to simple dialkylethy-
lenes, for which oxidation potentials of
ϳ2.2 V are typical.
ionization. This substrate was selected because it has the
highest quantum yield for photoionization amongst the pso-
ralens and coumarins that we have examined. Since the di-
rect irradiation of coumarins and psoralens has previously
been shown to give photodimerization products (42,43), we
anticipated that in the absence of a quencher the radical cat-
ion, if formed, would add to the precursor to give a similar
mixture of dimers. Note that our previous work had shown
that several psoralen and coumarin radical cations add to
dGuo-radical products (Type-I) that are formed following
deprotonation of the initial radical cation. Addition of oxy-
gen and then water to the initial guanosine radical, followed
by extensive rearrangement gives both oxoimidazolidine 1
and imidazolone 2; the latter can undergo addition of water
to give oxazolone 3, which may be in equilibrium with its
ring-opened guanidino acid form (44). These characteristic
Type-I products of dGuo have now been characterized by
several groups in both nucleosides and DNA fragments and
have been used as a diagnostic test for the formation of
electron transfer initiated chemistry (48,49).
Ϫ
1
their precursors with rate constants in the range of 107 M
Ϫ
1
s
(24,25). Therefore, the irradiations were carried out in
the presence of 4–5 mM dGuo which is sufficient to effi-
ciently trap the radical cation and to limit reaction of the
radical cation (Ͻ1%) with the precursor coumarin. Irradia-
tions were carried out in oxygen-saturated solution, which
minimizes any involvement of triplet excited states.
The photosensitized electron transfer chemistry of dGuo
has been studied in detail by Cadet and coworkers using both
benzophenone and riboflavin as sensitizers in aqueous so-
lution (44–47). Scheme 1 summarizes the characteristic
Irradiation of 6,7-DMC in oxygen-saturated aqueous buff-
er was carried out with either the 355 nm laser or with 350
nm Rayonet lamps and products were analyzed by HPLC
and LC/MS. Similar irradiations of benzophenone in the
presence of dGuo served as standards for comparison of the
retention times and mass spectral data for the base-derived
products. Both lamp and laser irradiations showed that each
of the three characteristic Type-I products was formed, in
addition to several other products, some of which were the
same as those observed upon irradiation of 6,7-DMC alone.
There appeared to be slightly less of the dGuo radical cat-
ion–derived products in the lamp irradiation, although this
apparent difference has not been quantified. Nevertheless the
observation of the expected products in both lamp and laser
irradiations is consistent with a one-photon ionization pro-
cess. It is important to note that these experiments are all
carried out in the presence of oxygen, which prevents any
electron transfer–sensitized chemistry involving triplet 6,7-
DMC. Further confirmation that triplet chemistry is not in-
volved is provided by the fact that the same products were
observed in the presence of sorbic acid as a triplet quencher.
DISCUSSION
The results described above demonstrate that psoralen and
coumarin radical cations react with a variety of biological
substrates, including the purine bases, amino acids and di-
enes. There are a number of factors that indicate that reaction
of the initial radical cation occurs via an electron transfer
reaction that regenerates the psoralen or coumarin and ini-
tiates one-electron oxidation of the quencher. For example,
in a number of cases the product quencher radical cation or
Scheme 1.

160 Paul D. Wood et al.
its deprotonated form can be detected directly by transient
absorption measurements. These include reactions with
GMP, tryptophan and tyrosine, for which the observation of
transients assigned to the oxidized quencher provides un-
equivocal evidence for an electron transfer reaction. For oth-
er quenchers it was not possible to provide direct transient
evidence for the formation of the oxidized quencher. For
example, the AMP radical cation is quite acidic with an es-
tential of the psoralen or coumarins. In fact the rate constants
vary by only a factor of two for the various radical cations,
as compared to approximately two orders of magnitude for
histidine, despite the fact that both amino acids react at sig-
nificantly less than the diffusion limited rate. This can be
rationalized by the fact that the measured rate constants for
cysteine depend on the pH, in line with previous observa-
•ϩ
tions for reactions of 8-MOP with glutathione, a cysteine-
timated pKa of
Ͻ
1 and unlikely to be detectable in our ex-
containing tripeptide (24). For example, the rate constant for
reaction of 7-MC with cysteine increases by a factor of 4
•ϩ
periments (50,51). The nitrogen-centered radical formed by
deprotonation of AMP has an absorption maximum at
•ϩ
when the pH is changed from 7 to 8. The reported pKa of
8.37 (36) for cysteine can be used to calculate that approx-
imately 4.3% of the cysteine is present as the thiolate ion at
pH 7. If one assumes that only the thiolate ion is reactive,
then one can calculate corrected rate constants of 7, 4 and
ϳ330 nm, and may be obscured by ground state psoralen
bleaching or absorption.
The trends in reactivity as a function of the oxidation po-
tential of both the psoralen (or coumarin) and the quenchers
are also consistent with electron transfer chemistry. For
quenchers such as AMP, histidine and several of the alkenes
the measured rate constants depend on the oxidation poten-
tial of the substrate with the measured rate constants de-
Ϫ
1
Ϫ
1
•ϩ
•ϩ
•ϩ
3
ϫ
109 M
s
for 7-MC , 8-MOP and 6,7-DMC , re-
spectively, at pH 7. None of the more easily oxidized amino
Ϫ
1
Ϫ1
acids have rate constants in excess of
ϳ
4
ϫ
•ϩ
109 M
s
in
aqueous solution, suggesting that for 7-MC both the thiol-
ate anion and cysteine may be reactive. This is also consis-
tent with the fact that methionine (in which the cysteine –
CH2SH group has been replaced by CH2CH2SCH3) reacts at
least two orders of magnitude faster with 7-MC than with
either 8-MOP or 6,7-DMC radical cations.
creasing for the series 7-MC
Ͼ 5,7-DMC Ͼ 8-MOP Ͼ 6,7-
DMC. The oxidation potentials for these substrates decrease
from 1.91 for 7-MC to 1.53 for 6,7-DMC (Table 1, [25]).
Furthermore, for any given quencher the rate constant for 7-
MC, which has the highest oxidation potential of the pso-
ralens or coumarins that have been studied in this work, is
always faster than those for the other radical cations, unless
all are already at or close to the diffusion controlled limit.
Similarly, the radical cations are unreactive toward the
quenchers that are more difficult to oxidize such as many of
the amino acids and the pyrimidine bases. The substantially
lower reactivity of pyrimidines over purines is in line with
literature data for the oxidation potentials which show that
The observed reactivity of the psoralen and coumarin rad-
ical cations with nucleotides and amino acids may be rele-
vant to several literature reports on psoralen photochemistry.
For example, competing cycloaddition and DNA strand
breaks have been observed following irradiation of DNA-
intercalated 8-MOP (54). Electron transfer reactions may
provide a mechanism for initiation of damage that ultimately
leads to strand breaks. Similarly, recent studies of the reac-
tion of 8-MOP with T7 RNA polymerase, a transcription
protein, indicate that the covalent photobinding of psoralen
to the protein involves exclusive formation of a psoralen–
tyrosine adduct (14), a result consistent with an electron
transfer mechanism. Reports of membrane damage can also
be explained by electron transfer–initiated alkene oxidation
or addition chemistry, although photocycloaddition reactions
with unsaturated membrane components would equally well
explain the results (18,19,55).
Our earlier observations of monophotonic photoionization
of several psoralens and coumarins coupled with the present
results demonstrating the high reactivity of psoralen radical
cations with various biological substrates clearly illustrate
the potential relevance of electron transfer chemistry in
PUVA therapy. The product studies described herein add
further credence to our earlier suggestion of one-photon ion-
ization of psoralens and coumarins. However, irrespective of
whether or not photoionization plays a significant role in in
vivo psoralen chemistry, electron transfer involving endog-
enous sensitizers (e.g. quinones) may provide an alternate
route for generating psoralen radical cations during PUVA
therapy. Another important consideration is the fact that the
psoralen radical cations are likely to react via reduction to
the parent compound with the concomitant formation of a
one-electron oxidized nucleotide, protein or membrane com-
ponent. This regeneration of the active drug is in contrast to
the photocycloaddition chemistry of psoralens in which each
photochemical event leads to depletion of the photoactivated
drug. In considering the relevance of electron transfer chem-
guanosine and adenosine are
ϳ0.4 and 0.3 V easier to reduce
than thymidine (52). Thus the evidence for electron transfer
quenching is unequivocal, although it should be noted that
one cannot rule out the possibility of minor amounts of other
reactions contributing to the overall reactivity.
There are two exceptions to the general observation that
the reactivity of the psoralen and coumarin radical cations
increases with increasing oxidation potential. One is provid-
ed by the two alkenes, linoleic and oleic acid, for which
comparable rate constants have been measured for both 7-
•ϩ
•ϩ
MC and 8-MOP . As noted in Table 3, the oxidation po-
tential for both linoleic and oleic acid should be in excess
of 2.2 V vs saturated calomel electrode (SCE), based on
reported oxidation potentials for simpler dialkyl-substituted
ethylenes (53). Thus one expects electron transfer to be sig-
nificantly slower for linoleic and oleic acid than for sorbic
•ϩ
acid and 8,10-dodecadienol. The rate constants for 7-MC
follow this trend but those for 8-MOP do not; the anoma-
lously high rate constants for reaction of 8-MOP with both
linoleic and oleic acids must include contributions from re-
actions other than electron transfer, with addition to the al-
kene being the most probable alternative. One might expect
•ϩ
•ϩ
•ϩ
addition reactions to be faster for 8-MOP than for the cou-
marin radical cations based on our earlier observation of
substantially slower rates of addition to their precursors for
7-MC and 6,7-DMC radical cations than for 8-MOP or
4,5Ј,8-TMP (25).
Cysteine is the second example for which the rate con-
stants do not follow the expected trend with oxidation po-

Photochemistry and Photobiology, 2000, 71(2) 161
psoralen with a natural protein. Photochem. Photobiol. 65, 937–
944.
istry to PUVA therapy it is clearly important to investigate
the behavior of psoralens in media that are more relevant to
the cellular environment. To this end we have initiated a
detailed study of the electron transfer chemistry of psoralens
in heterogeneous media such as micelles and vesicles. Pre-
liminary results in micelles indicate that in some cases the
photoionization efficiency can be significantly increased by
incorporating the psoralen in a micelle (L. J. Johnston, P. D.
Wood, L. Chen, A. Mnyusiwalla, unpublished). The micelle-
solubilized radical cations are still quite reactive toward both
micelle bound and aqueous quenchers, albeit with substantial
changes in their kinetic behavior as compared to homoge-
neous solution, clearly demonstrating the potential of the en-
vironment to control the overall chemistry. In summary, it
is obvious that electron transfer chemistry may be of more
general significance than has previously been considered and
may play a role in either the phototherapeutic or phototoxic
effects of psoralens. This is also a consideration for the de-
sign of new psoralen derivatives with improved photother-
apeutic applications.
15. Sastry, S. S., B. M. Ross and P. P’arraga (1997) Cross-linking
of DNA-binding proteins to DNA with psoralen and psoralen
furan-side monoadducts. J. Biol. Chem. 272, 3715–3723.
16. Schmitt, I. M., S. Chimenti and F. P. Gasparro (1995) Psoralen–
protein photochemistry—a forgotten field. J. Photochem. Pho-
tobiol. B: Biol. 27, 101–107.
17. Bordin, F., F. Carlassare, L. Busulini and F. Baccichetti (1993)
Furocoumarin sensitization induces DNA-protein cross-links.
Photochem. Photobiol. 58, 133–136.
18. Dall’Acqua, F. and P. Martelli (1991) Photosensitizing action
of furocoumarins on membrane components and consequent in-
tracellular events. J. Photochem. Photobiol. B: Biol. 8, 235–254.
19. Caffieri, S., A. Daga, D. Vedaldi and F. Dall’Acqua (1988) Pho-
toaddition of angelicin to linoleic acid methyl ester. J. Photo-
chem. Photobiol. B: Biol. 2, 515–521.
20. Voituriez, L. and J. Cadet (1999) Isolation and characterizationn
of two furan-side photoadducts of 7-methylpyrido[3,4-
c]psoralen to the sugar moiety of 2Ј-deoxyadenosine. Photo-
chem. Photobiol. 70, 152–158.
21. Sousa, C. and T. Sa e Melo (1996) Microenvironment effects
on the excited state properties of psoralens: a clue to their pho-
tobiological activity. Photochem. Photobiol. 63, 182–186.
22. Becker, R., S. Chakravorti, C. A. Gartner and M. de Graca Mi-
guel (1993) Photosensitizers: comprehensive photophysics/pho-
tochemistry and theory of coumarins, chromones, their homo-
logues and thione analogues. J. Chem. Soc. Faraday Trans. 89,
1007–1019.
Acknowledgements We thank Malgoshia Daroszewska for help
with the LC and LC/MS experiments.
23. van Henegouwen, G. M. J. B., E. T. Wijn, S. A. Schoonder-
woerd and F. Dall’Acqua (1989) A method for the determina-
tion of PUVA-induced in vivo irreversible binding of 8-meth-
oxyposralen to epidermal lipids, proteins and DNA/RNA. J.
Photochem. Photobiol. B: Biol. 3, 631–635.
24. Wood, P. D. and L. J. Johnston (1997) Generation and charac-
terization of psoralen radical cations. Photochem. Photobiol. 66,
642–648.
25. Wood, P. D. and L. J. Johnston (1998) Photoionization and
photosensitized electron transfer reactions of psoralens and cou-
marins. J. Phys. Chem. 102, 5585–5591.
26. Kang, H. K., E. J. Shin and S. C. Shim (1992) Transient ab-
sorption spectra and quenching of coumarin excited states by
nucleic acid bases. J. Photochem. Photobiol. B: Biol. 13, 19–
28.
27. Chen, T., M. S. Platz, M. Robert, J.-M. Saveant, A. Marcinek,
J. Rogowski, J. Gebicki, Z. Zhu and T. Bally (1997) Electron
transfer chemistry of psoralen and coumarin derivatives by
means of radiolytic and electrochemical experiments. J. Phys.
Chem. A 101, 2124–2130.
28. Rai, S., C. Kasturi, J. Grayzar, M. S. Platz, R. P. Goodrich, N.
R. Yerram, V. Wong and B. H. Tay-Goodrich (1993) Dramatic
improvements in viral inactivation with brominated psoralens,
naphthalenes and anthracenes. Photochem. Photobiol. 58, 59–
65.
29. Shaquiri, Z., E. Keskinova, A. Spassky and D. Angelov (1997)
A picosecond flash photolysis study of the biphotonic ionization
of psoralen derivatives and ethidium bromide. Photochem. Pho-
tobiol. 65, 517–521.
30. Sloper, R. W., T. G. Truscott and E. J. Land (1979) The triplet
state of 8-methoxypsoralen. Photochem. Photobiol. 29, 1025–
1029.
REFERENCES
1. Gasparro, F. P. (1988) Psoralen DNA Photobiology, Vols. 1 and
2. CRC Press, Boca Raton.
2. Gonzalez, E. (1995) PUVA for psoriasis. Dermatol. Clinics 13,
851–866.
3. Goodrich, R. P., N. R. Yerram, B. H. Tay-Goodrich, P. Forster,
M. S. Platz, C. Kasturi, S. C. Park, J. N. Aebischer, S. Rai and
L. Kulaga (1994) Selective inactivation of viruses in the pres-
ence of human platelets: UV sensitization with psoralen deriv-
atives. Proc. Natl. Acad. Sci. USA 91, 5552–5556.
4. Pathak, M. A. and T. B. Fitzpatrick (1992) The evolution of
photochemotherapy with psoralens and UVA (PUVA): 200 BC
to 1992 AD. J. Photochem. Photobiol. B: Biol. 14, 3–22.
5. Bensasson, R. V., E. J. Land and T. G. Truscott (1993) Excited
States and Free Radicals in Biology and Medicine. Oxford Uni-
versity Press, Oxford.
6. Stern, R. S., K. T. Nichols and L. H. Vakeva (1997) Malignant
melanoma in patients treated for psoriasis with methoxsalen
(psoralen) and ultraviolet A radiation (PUVA). N. Engl. J. Med.
336, 1041–1045.
7. Young, A. R. (1990) Photocarcinogenicity of psoralens used in
PUVA treatment: present status in mouse and man. J. Photo-
chem. Photobiol. B: Biol. 6, 237–247.
8. Poppe, W. and L. I. Grossweiner (1975) Photodynamic sensi-
tization by 8-methoxypsoralen via the singlet oxygen mecha-
nism. Photochem. Photobiol. 22, 217–219.
9. De Mol, N. J. and G. M. J. Beijersbergen van Henegouwen
(1979) Formation of singlet molecular oxygen by 8-methoxyp-
soralen. Photochem. Photobiol. 30, 331–335.
10. Blan, Q. A. and L. I. Grossweiner (1987) Singlet oxygen gen-
eration by furocoumarins: effect of DNA and liposomes. Pho-
tochem. Photobiol. 45, 177–183.
31. Adam, W., K. Mielke, C. R. Saha-Moller, M. Moller, H. S.
Hutterer, F. W. Schneider, D. Ballmaier, B. Epe, F. F. Gasparro,
X. Chen and J. Kagan (1997) Photochemical and photobiolog-
ical studies of a furonaphthopyranone as a benzo-spaced pso-
ralen analog in cell-free and cellular DNA. Photochem. Photo-
biol. 66, 46–54.
11. Giles, A. J., W. Wamer and A. Kornhauser (1985) In vivo pro-
tective effect of ␤-carotene against psoralen phototoxicity. Pho-
tochem. Photobiol. 41, 661–666.
12. Bordin, F., M. T. Conconi and A. Capozzi (1987) Certain singlet
oxygen quenchers affect the photoreaction between 8-MOP and
DNA. Photochem. Photobiol. 46, 301–304.
32. Kazanis, S., A. Azarani and L. J. Johnston (1991) Diffuse re-
flectance laser flash photolysis studies of reactions of triplet ben-
zophenone with hydrogen donors on silica. J. Phys. Chem. 95,
4430–4435.
13. Sastry, S. S., H. P. Spielmann, Q. S. Hoang, A. M. Phillips, A.
Sancar and J. E. Hearst (1993) Laser-induced protein-DNA
cross-links via psoralen furanside monoadducts. Biochemistry
32, 5526–5538.
33. Candeias, L. P. and S. Steenken (1992) Ionization of purine
nucleosides and nucleotides and their components by 193-nm
14. Sastry, S. S. (1997) Isolation and partial characterization of a
novel psoralen–tyrosine photoconjugate from a photoreaction of

162 Paul D. Wood et al.
laser photolysis in aqueous solution: Model studies for oxidative
damage to DNA. J. Am. Chem. Soc. 114, 699–704.
34. Gut, I. G., P. D. Wood and R. W. Redmond (1996) Interaction
of triplet photosensitizers with nucleotides and DNA in aqueous
solution at room temperature. J. Am. Chem. Soc. 118, 2366–
2373.
35. Candeias, L. P. and S. Steenken (1989) Structure and acid-base
properties of one-electron oxidized deoxyguanosine, guanosine,
and 1-methylguanosine. J. Am. Chem. Soc. 111, 1094–1099.
36. Dawson, R. M. C., D. C. Elliott, W. H. Elliott and K. M. Jones
(1986) Data for Biochemical Research. Clarendon Press, Ox-
ford.
37. Jovanovic, S. V. and S. Steenken (1992) Substituent effects on
the spectral, acid-base and redox properties of indolyl radicals:
a pulse radiolysis study. J. Phys. Chem. 96, 6674–6679.
38. McGimpsey, W. G. and H. Gorner (1996) Photoionization of
indole, N-methylindole and tryptophan in aqueous solution upon
excitation at 193 nm. Photochem. Photobiol. 64, 501–509.
39. Feitelson, J., E. Hayon and A. Treinin (1973) Photoionization
of phenols in water. Effects of light intensity, oxygen, pH, and
temperature. J. Am. Chem. Soc. 95, 1025–1029.
40. Feitelson, J. and E. Hayon (1975) Electron ejection and electron
capture by phenolic compounds. J. Phys. Chem. 77, 10–15.
41. Gadosy, T. A., D. Shukla and L. J. Johnston (1999) Generation,
characterization and deprotonation of phenol radical cations. J.
Phys. Chem. 103, 8834–8839.
46. Buchko, G. W., J. Cadet, J.-L. Ravanat and P. Labataille (1993)
Isolation and characterization of a new product produced by
ionizing irradiation and type I photosensitization of 2Ј-deoxy-
guanosine in oxygen-saturated aqueous solution. Int. J. Radiat.
Biol. 63, 669–676.
47. Raoul, S., M. Berger, G. W. Buchko, P. C. Joshi, B. Morin, M.
Weinfeld and J. Cadet (1996) 1H, 13C and 15N nuclear magnetic
resonance analysis and chemical features of the two main radical
oxidation products of 2Ј-deoxyguanosine: oxazolone and imi-
dazolone nucleosides. J. Chem. Soc. Perkin Trans. 2, 371–381.
48. Adam, W., C. R. Saha-Moller and A. Schonberger (1997) Type
I and type II photosensitized oxidative modificatin of 2-deoxy-
guanosine (dGuo) by triplet excited ketones generated thermally
from the 1,2-dioxetane HTMD. J. Am. Chem. Soc. 119, 719–
723.
49. Kino, K., I. Saito and H. Sugiyama (1998) Product analysis of
GG-specific photooxidation of DNA via electron transfer; 2-
aminoimidazolone as a major guanine oxidation product. J. Am.
Chem. Soc. 120, 7373–7374.
50. Steenken, S. (1989) Purine bases, nucleosides and nucleotides:
Aqueous solution redox chemistry and transformation reactions
of their radical cations and eϪ and OH adducts. Chem. Rev. 89,
503–520.
51. Viera, A. J. S. C. and S. Steenken (1987) Pattern of OH radical
reaction with 6- and 9-substituted purines. Effect of substituents
on the rates and activation parameters of the unimolecular trans-
formation reactions of two isomeric OH adducts. J. Phys. Chem.
91, 4138–4144.
52. Steenken, S. and S. V. Jovanovic (1997) How easily oxidizable
is DNA? One-electron reduction potentials of adenosine and
guanosine radicals in aqueous solution. J. Am. Chem. Soc. 119,
617–618.
53. Schepp, N. P. and L. J. Johnston (1996) Reactivity of radical
cations. Effect of radical cation and alkene structure on the ab-
solute rate constants of radical cation mediated cycloaddition
reactions. J. Am. Chem. Soc. 118, 2872–2881.
54. Kagan, J., X. Chen, T. P. Wang and P. Forlot (1992) Psoralens
cleave pBR322 DNA under ultraviolet radiation. Photochem.
Photobiol. 56, 185–193.
55. Sousa, C., T. Sa e Melo, J.-C. Maziere and R. Santus (1998)
PUVA-induced cell mortality in NCTC 2544 keratinocytes: is
it related to the microenvironmental properties of the excited
states of psoralens? Photochem. Photobiol. 67, 561–564.
42. Shim, S. C., B. M. Jeong and Y. H. Pack (1992) Photodimeri-
zation of 5,7-dimethoxycoumarin. Bull. Korean Chem. Soc. 13,
684–688.
43. Gervais, J. and F. C. De Schryver (1975) Photochemistry of
some furo (3,2-␥)-coumarin and 2,3-dihydrofuro (3,2-␥)-cou-
marin derivatives. Photochem. Photobiol. 21, 71–75.
44. Gasparutto, D., J.-L. Ravanat, O. Gerot and J. Cadet (1998)
Characterization and chemical stability of photooxidized oligo-
nucleotides that contain 2,2-diamino-4-[(2-deoxy-␤-D-erythro-
pentofuranosyl)amino]-5(2H)-oxazolone. J. Am. Chem. Soc.
120, 10 283–10 286.
45. Cadet, J., M. Berger, G. W. Buchko, P. C. Joshi, S. Raoul and
J.-L. Ravant (1994) 2,2-Diamino-4-[3,5-di-O-acetyl-2-deoxy-
-erythro-pentofuranosyl]-5-(2H)-oxazolone: a novel and pre-
dominant radical oxidation product of 3 ,5 -di-O-acetyl-2 -deox-
yguanosine. J. Am. Chem. Soc. 116, 7403–7404.
␤-
D
Ј
Ј
Ј