Photochemistry of N-isopropoxy-substituted 2(1H)-pyridone and 4-p-tolylthiazole-2(3H)-thione: Alkoxyl-radical release (spin-trapping, EPR, and transient spectroscopy) and its significance in the photooxidative induction of DNA strand breaks

Article abstract of DOI:10.1021/jo025856w

UVA-irradiation of the photo-Fenton reagents N-isopropoxypyridone 2b and N-isopropoxythiazole2(3H)-thione 3b releases radicals which induce strand breaks. Transient spectroscopy establishes N-O bond scission [Φ_N-O = (75 ± 8)% for 2b and (65 ± 7)% 3b] as the dominating primary photochemical process to afford the DNA-damaging radicals. Product studies and laser-flash experiments reveal that the thiazolethione 3b leads primarily to the disulfide 5, from which through C-S bond breakage, the bithiazyl 6, the thiazole 7, and the isothiocyanate 8 are derived. Upon irradiation of pyridone 2b (300 nm) in aqueous media, a mixture of isopropoxyl and 2-hydroxyprop2-yl radicals is formed, as confirmed by trapping with 5,5-dimethyl-1-pyrroline N-oxide (DMPO) and EPR spectroscopy. In contrast, the photolysis of the thiazolethione 3b (350 nm) affords exclusively the DMPO adducts of the isopropoxyl radicals. Control experiments disclose that the thiazolethione-derived photoproduct disulfide 5, or the intermediary thiyl radicals B, scavenge the carbon-centered 2-hydroxyprop-2-yl radicals, which are generated from the isopropoxyl radicals by hydrogen shift. With supercoiled pBR 322 DNA in a 60:40 mixture of H₂O-MeCN, the pyridone 2b and the thiazolethione 3b display moderate strand-break activity (17% open-circular DNA for 2b and 12% for 3b). In pure water, however, the pyridone 2b photoinduces substantially more DNA cleavage (32% open-circular DNA), which is attributed to the peroxyl radicals generated from the 2-hydroxyprop-2-yl radicals by oxygen trapping. The lower strand-break activity of the thiazolethione 3b derives presumably from isopropoxyl radicals, because only these are detected in the photolysis of this photo-Fenton reagent.

Full text of DOI:10.1021/jo025856w

P h otoch em istr y of N-Isop r op oxy-Su bstitu ted 2(1H)-P yr id on e a n d
4-p-Tolylth ia zole-2(3H)-th ion e: Alk oxyl-Ra d ica l Relea se
(Sp in -Tr a p p in g, EP R, a n d Tr a n sien t Sp ectr oscop y) a n d Its
Sign ifica n ce in th e P h otooxid a tive In d u ction of DNA Str a n d
Br ea k s
Waldemar Adam,*,† J ens Hartung,† Hideki Okamoto,† Stefan Marquardt,† Werner M. Nau,‡
Uwe Pischel,‡ Chantu R. Saha-Mo¨ller,† and Kristina Sˇpehar†
Institut fu¨r Organische Chemie, Universita¨t Wu¨rzburg, Am Hubland, D-97074 Wu¨rzburg, Germany, and
Institut fu¨r Physikalische Chemie, Universita¨t Basel, Klingelbergstrasse 80, CH-4056 Basel, Switzerland
adam@chemie.uni-wuerzburg.de
Received April 25, 2002
UVA-irradiation of the photo-Fenton reagents N-isopropoxypyridone 2b and N-isopropoxythiazole-
2(3H)-thione 3b releases radicals which induce strand breaks. Transient spectroscopy establishes
N-O bond scission [ΦN-O ) (75 ( 8)% for 2b and (65 ( 7)% 3b] as the dominating primary
photochemical process to afford the DNA-damaging radicals. Product studies and laser-flash
experiments reveal that the thiazolethione 3b leads primarily to the disulfide 5, from which through
C-S bond breakage, the bithiazyl 6, the thiazole 7, and the isothiocyanate 8 are derived. Upon
irradiation of pyridone 2b (300 nm) in aqueous media, a mixture of isopropoxyl and 2-hydroxyprop-
2-yl radicals is formed, as confirmed by trapping with 5,5-dimethyl-1-pyrroline N-oxide (DMPO)
and EPR spectroscopy. In contrast, the photolysis of the thiazolethione 3b (350 nm) affords
exclusively the DMPO adducts of the isopropoxyl radicals. Control experiments disclose that the
thiazolethione-derived photoproduct disulfide 5, or the intermediary thiyl radicals B, scavenge the
carbon-centered 2-hydroxyprop-2-yl radicals, which are generated from the isopropoxyl radicals
by hydrogen shift. With supercoiled pBR 322 DNA in a 60:40 mixture of H2O-MeCN, the pyridone
2b and the thiazolethione 3b display moderate strand-break activity (17% open-circular DNA for
2b and 12% for 3b). In pure water, however, the pyridone 2b photoinduces substantially more
DNA cleavage (32% open-circular DNA), which is attributed to the peroxyl radicals generated from
the 2-hydroxyprop-2-yl radicals by oxygen trapping. The lower strand-break activity of the
thiazolethione 3b derives presumably from isopropoxyl radicals, because only these are detected
in the photolysis of this photo-Fenton reagent.
In tr od u ction
decades as a consequence of exposure to the UV radiation
in sunlight, it is essential to elucidate the mechanism of
these biologically relevant photooxidative events to de-
velop effective methods for prevention. For such studies,
the reactive oxygen species need to be generated selec-
tively under as mild conditions as possible to elucidate
their chemical behavior and interaction with biomol-
ecules.
In the case of hydroxyl and alkoxyl radicals, the Fenton
reaction has been used for their generation. This reaction
plays an important role in cellular systems, since lipid
hydroperoxides serve as precursors for lipid alkoxyl
radicals, which have been shown to induce strand breaks
in DNA.6 However, the Fenton reaction, despite its
biological relevance, is not a selective route to oxyl
radicals, since alternative redox processes with biomol-
ecules may compete.7 More recently, photo-Fenton re-
agents have been developed, which release oxyl radicals
In the photooxidative damage of biomolecules, the
reactive oxygen species, hydroxyl, alkoxyl, and peroxyl
radicals as well as singlet oxygen, have been held
responsible for the chemical transformations.1 In par-
ticular, oxidative modifications of cellular DNA by such
potent oxidants are implicated in aging,2,3 mutagenesis,
and carcinogenesis4 and have been classified under the
general phenomenon of oxidative stress.5 In view of the
steady increase of skin cancer incidences during the past
* To whom correspondence should be sent. Tel.: + 49 931 888 5340.
Telefax: + 49 931 888 4756. Internet: http://www.organik.chemie.
uni-wuerzburg.de.
† Universita¨t Wu¨rzburg.
‡ Universita¨t Basel.
(1) Scharffetter-Kochanek, K.; Wlaschek, M.; Brenneisen, P.; Schauen,
M.; Blaudschun, R.; Wenk, J . Biol. Chem. 1997, 378, 1247-1257.
(2) Harman, D. J . Gerontol. 1985, 11, 298-300.
(3) Sohal, R. S. Aging Milano 1993, 5, 3-17.
(4) Halliwell, B.; Gutteridge, J . M. C. Free Radicals in Biology and
Medicine; Oxford University Press: Oxford, 1989.
(5) Sies, H. (ed.) Oxidative Stress, Oxidants and Antioxidants;
Academic Press: New York, 1991.
(6) Yang, M.-H.; Schaich, K. M. Free Radic. Biol. Med. 1996, 20,
225-236.
(7) Walling, C. Acc. Chem. Res. 1998, 31, 155-157.
10.1021/jo025856w CCC: $22.00 © 2002 American Chemical Society
Published on Web 07/19/2002
J . Org. Chem. 2002, 67, 6041-6049
6041

Adam et al.
on irradiation with UVA or visible light.8-12 As a model
system for alkoxyl radicals, the tert-butoxyl radical has
been chosen, for which DNA intercalating13 and DNA
binding14,15 reagents have been developed. Strand-break
formation and base oxidation have been observed as
damage of DNA induced by the alkoxyl radicals.
For a photo-Fenton reagent to act selectively, it is
indispensable that neither it itself nor its photoproducts
display significant oxidative damages on the biological
targets through photosensitization (type I and type II
photooxidation).16 Among such photo-Fenton reagents,
pyridinethiones 1 have been extensively investigated as
photochemicalsources ofhydroxyland alkoxylradicals.17-19
F IGURE 1. Electronic absorption spectra of pyridone 2b and
thiazolethione 3b in a H2O-MeCN mixture (60:40).
SCHEME 1. P h otop r od u cts of th e P yr id on e 2b
They cause strand breaks and oxidative base modification
in DNA upon photolysis;18 however, the DNA oxidation
by the pyridinethiones 1 is complicated by the fact that
their photoproducts may possess significant damaging
activity by photosensitization.19,20
Recently, we have reported that N-hydroxypyridone
2a 21 and N-hydroxythiazolthione 3c22 serve as selective
photochemical hydroxyl-radical sources for photobiologi-
cal studies: On UVA irradiation, they induce strand
breaks and base oxidation in DNA by the released
hydroxyl radicals, while their photoproducts display
negligible oxidative activity. These advantages motivated
us to utilize the chromophores in the N-isopropoxy
derivatives of the pyridone 2b and thiazolethione 3b to
generate photochemically isopropoxyl radicals. Of par-
ticular interest was to assess the efficacy and selectivity
of these alkoxyl radical sources for the oxidative damage
of DNA.
In the present work, we report the details of the
photochemistry of the N-isopropoxy derivatives of pyri-
done 2b and the thiazolethione 3b in terms of product
studies. The intermediary radicals have been detected
by means of laser-flash spectroscopy and EPR-spectral
analysis of the spin-trapped adducts derived from 5,5-
dimethyl-1-pyrroline N-oxide (DMPO). The efficiency of
photoinduced strand-break formation in DNA has been
determined for these alkoxyl-radical sources.
Resu lts
The pyridone 2b23,24 and the thiazolethione 3b25 were
synthesized according to the literature procedures. Py-
ridone 2b is soluble in water, whereas the thiazolethione
3b requires MeCN as cosolvent (40%) for solubilization
in aqueous media. To enable direct comparison, both
substrates were investigated in a H2O-MeCN mixture
(60:40).
The absorption spectra are shown in Figure 1. The
pyridone 2b possesses an absorption maximum at λmax
300 nm (log ꢀ ) 3.73), whereas for the thiazolethione 3b
the λmax appears at 318 nm (log ꢀ ) 4.15), with tailing to
360 nm. These absorption characteristics indicate that
the thiazolethione 3b absorbs UVA radiation more ef-
ficiently and out to a longer wavelength than the pyri-
done 2b.
P h otod ecom p osition of th e P h oto-F en ton Re-
a gen ts 2b a n d 3b. P h otop r od u cts of th e P yr id on e
2b. To elucidate the photochemical behavior of the
pyridone 2b, a product analysis was essential. The
photolysis (300 nm) of the pyridone 2b (Scheme 1) led to
the parent pyridone 4a (22%) and the isopropoxypyridone
4b (11%); the rest was made up by undefined, intractable
(8) Saito, I.; Takayama, M.; Matsuura, T.; Matsugo, S.; Kawanishi,
S. J . Am. Chem. Soc. 1990, 112, 883-884.
(9) Boivin, J .; Cre´pon, E.; Zard, S. Z. Bull. Soc. Chim. Fr. 1992, 129,
145-150.
(10) Adam, W.; Cadet, J .; Dall’Acqua, F.; Epe, B.; Frank, S.;
Ramaiah, D.; Saha-Mo¨ller, C. R.; Vedaldi, D. Angew. Chem., Int. Ed.
Engl. 1995, 34, 107-110.
(11) Adam, W.; Ballmaier, D.; Epe, B.; Grimm, G. N.; Saha-Mo¨ller,
C. R. Angew. Chem., Int. Ed. Engl. 1995, 34, 2156-2158.
(12) Epe, B.; Ballmaier, D.; Adam, W.; Grimm, G. N.; Saha-Mo¨ller,
C. R. Nucleic Acids Res. 1996, 24, 1625-1631.
(13) Adam, W.; Arnold, M. A.; Grimm, G. N.; Saha-Mo¨ller, C. R.;
Dall’Acqua, F.; Miolo, G.; Vedaldi, D. Photochem. Photobiol. 1998, 68,
511-518.
(14) Adam, W.; Grimm, G. N.; Saha-Mo¨ller, C. R.; Dall’Acqua, F.;
Miolo, G.; Vedaldi, D. Chem. Res. Toxicol. 1998, 11, 1089-1097.
(15) Mahler, H. C.; Schulz, I.; Adam, W.; Grimm, G. N.; Saha-Mo¨ller,
C. R.; Epe, B. Mutat. Res. 2001, 461, 289-299.
(16) Foote, C. S. Photochem. Photobiol. 1991, 54, 659.
(17) Barton, D. H. R.; Crich, D.; Motherwell, W. B. Chem. Commun.
1983, 939-941.
(18) Adam, W.; Grimm, G. N.; Saha-Mo¨ller, C. R. Free Radic. Biol.
Med. 1998, 24, 234-238.
(19) Adam, W.; Grimm, G. N.; Marquardt, S.; Saha-Mo¨ller, C. R. J .
Am. Chem. Soc. 1999, 121, 1179-1185.
(20) Aveline, B. M.; Kochevar, I. E.; Redmond, R. W. J . Am. Chem.
Soc. 1996, 118, 10113-10123.
(21) Adam, W.; Marquardt, S.; Saha-Mo¨ller, C. R. Photochem.
Photobiol. 1999, 70, 287-291.
(22) Adam, W.; Hartung, J .; Okamoto, H.; Saha-Mo¨ller, C. R.;
Sˇpehar, K. Photochem. Photobiol. 2000, 72, 619-624.
(23) Newbold, G. T.; Spring, F. S. J . Chem. Soc. 1948, 1864-1866.
(24) Raban, M.; Kost, D. J . Org. Chem. 1972, 37, 499-501.
6042 J . Org. Chem., Vol. 67, No. 17, 2002

Alkoxyl-Radical Release
F IGURE 2. Transient spectra observed in the laser-flash photolysis (308 nm) for pyridone 2b in water (a), and for the
thiazolethiones 3b (b) and 3a (c) and disulfide 5 (d) in MeCN; for the thiazolethiones 3b (b) and 3a (c), the transient spectra
display the traces for the first and the second laser shots.
SCHEME 2. P h otop r od u cts of th e Th ia zoleth ion e
3b
Also a high-resolution mass spectrum supports the
assigned structure 8.
Tr a n sien t Sp ectr oscop y. The transient spectrum of
the laser-flash photolysis (308 nm) of the pyridone 2b is
shown in Figure 2a. It displays a transient UV band with
λmax 390 nm, which corresponds to the pyridyloxyl radical
A, since it matches the reported spectrum for the
authentic species.26
The flash photolysis of the thiazolethione 3b afforded
a transient absorption with maxima at 355 and 470 nm
(Figure 2b, solid curve). The same transient spectrum
was observed for the hydroxy derivative 3a (Figure 2c,
solid curve). In view of these facts, the transient spectra
(Figures 2b and 2c, solid curves) have been assigned to
the thiyl radical B, generated through photochemical
N-O bond scission. The absorption band at 470 nm was
not affected by molecular oxygen, but was diminished by
2-propanol and 1,3-cyclohexadiene (1,3-CHD). Such chemi-
cal behavior is characteristic for aromatic thiyl radi-
cals,20,27,28 which possess absorption maxima at 300-390
and 460-500 nm, as reported for the thiyl radicals
derived from thiophenols and thionaphthols.29
On further flash photolysis (second laser shot) of the
thiazolethione 3b , a new transient was observed with
absorption maxima at 375 and 500 nm (Figure 2b, dotted
curve). The same transient was detected on photolysis
of the photoproduct of the thiazolethione 3b, namely the
disulfide 5 (Figure 2d, dotted curve). Since both absorp-
tion bands decayed faster in the presence of molecular
oxygen, this new radical species cannot be sulfur-centered
and is presumably the carbon-centered radical C, formed
by C-S bond breakage in the disulfide 5. On one hand,
higher-molecular-weight material. The structure of the
rearrangement product 4b was determined by 1H and 13C
NMR spectroscopy (especially by HMBC and HMQC
methods) and by its mass spectrum.
P h otop r od u cts of th e Th ia zoleth ion e 3b. When the
thiazolethione 3b was irradiated (>350 nm) in a H2O-
MeCN mixture (15:85), the disulfide 5 (10%), the bithi-
azyl 6 (5%), and the thiazole 7 (10%) were isolated and
characterized (Scheme 2), as previously described for the
photolysis of the N-hydroxythiazole-2(3H)-thione deriva-
tive (3c).22 The disulfide 5 and the bithiazyl 6 resisted
all attempts of chromatographic separation, and conse-
quently they were characterized directly in their mixture
1
by H NMR and mass spectra and are in close analogy
to their p-chlorophenyl derivatives.22
Additionally, the unknown isothiocyanate 8 was formed
in the photoreaction of the N-isopropoxythiazolethione
3b as the main photoproduct (27%). With only 52% of
the products accounted for, presumably much material
was lost on workup due to the hydrolysis of such labile
ring-opened products as the isothiocyanate 8. The pres-
ence of the isothiocyanate functionality was confirmed
by a strong IR absorption band (2074 cm-1), the full
(25) Hartung, J .; Kneuer, R.; Schwarz, M.; Svoboda, I.; Fuess, H.
Eur. J . Org. Chem. 1999, 97-106.
(26) Aveline, B. M.; Kochevar, I. E.; Redmond, R. W. J . Am. Chem.
Soc. 1996, 118, 10124-10133.
(27) Aveline, B. M.; Kochevar, I. E.; Redmond, R. W. J . Am. Chem.
Soc. 1995, 117, 9699-9708.
(28) Ito, O.; Matsuda, M. J . Am. Chem. Soc. 1981, 103, 5871-5874.
(29) Hermann, R.; Dey, G. R.; Naumov, S.; Brede, O. Phys. Chem.
Chem. Phys. 2000, 2, 1213-1220.
1
structure was established by H and 13C NMR spectra
on the basis of HMQC, HMBC, and NOESY techniques.
J . Org. Chem, Vol. 67, No. 17, 2002 6043

Adam et al.
TABLE 1. Qu a n tu m Yield s for th e P h otod ecom p osition
(Φd eca ) a n d for th e N-O Bon d Scission (ΦN-Ob) of th e
P yr id on e 2b a n d th e Th ia zoleth ion e 3b
substrate
solvent
Φdec
ΦN-O
2b
2b
3b
H2O-MeCN (60:40)
H2O
H2O-MeCN (60:40)
0.36 ( 0.02
0.32 ( 0.01
0.60 ( 0.05
-
0.75 ( 0.08c
0.65 ( 0.07d
a
Potassium ferrioxalate actinometer was used to determine the
photon fluence of the light sources. The error limits are based on
at least two independent runs; the pyridone 2b was irradiated with
a black-light lamp (312 nm), the thiazolethione 3b with the Xe
b
lamp of a fluorescence photometer (308 nm). Determined by
transient spectroscopy through actinometry with the triplet-
triplet absorption of benzophenone as reference (see Supporting
Information). c Calculated with ꢀA ) 680 M-1 cm-1 at 390 nm (ref
d
26). Estimated by assuming that the extinction coefficient of the
transient B (ꢀB) is that of the pyridyl-2-thiyl radical [ΦPyS ) 600
M-1 cm-1 at 460 nm (ref 20)].
significantly overlap with that of the starting pyridone
2b. Hence, the conversion of pyridone 2b was followed
by HPLC analysis, coupled with UV detection.
The quantum yields for the photodecomposition of
pyridone 2b and thiazolethione 3b are listed in Table 1.
F IGURE 3. Time profile of the UV spectra during the
irradiation (308 nm, 20 °C) of the thiazolethione3b; the inset
displays the decrease of the absorbance at λmax 318 nm.
These data disclose that the thiazolethione 3b (Φdec
)
carbon-centered radicals are efficiently trapped by mo-
lecular oxygen;30 on the other hand, products 6 and 7
(Scheme 2) demand that such a heteroaryl radical
intervenes.
0.60 ( 0.05) is about twice more photolabile than the
pyridone 2b (Φdec ) 0.36 ( 0.02). For the pyridone 2b,
the change of the solvent from the H2O-MeCN mixture
(60:40) to water did not significantly influence the
quantum yield of photodecomposition.
Qu a n tu m Yield s for th e P h otod ecom p osition . The
efficiencies of the photochemical degradation of pyridone
2b and thiazolethione 3b reflect their photoreactivity and
the upper-limit efficiency of radical generation from these
photo-Fenton reagents. For this purpose, the quantum
yields of photodecomposition were determined for the
compounds 2b and 3b in order to compare their photo-
reactivity. The pyridone 2b and the thiazolethione 3b
were irradiated at 308 or 312 nm, potassium ferrioxalate
was used as actinometer.31 For the quantification of the
substrate conversion, a detailed examination of their
absorption spectra as a function of time was necessary.
As shown in Figure 3, the absorbance of the thiazole-
thione 3b at λmax 318 nm decreased during irradiation;
the change displays a first-order dependence on the
irradiation time up to 10 min (Figure 3, inset). Although
the photoproducts of the thiazolethione 3b may absorb
light at 318 nm and possibly serve as photosensitizer,
the linear logarithmic diminution of the absorbance and
the isosbestic point at 290 nm indicate that the absorp-
tion at 318 nm is not falsified by photoproduct absorption.
Therefore, the time profile reflects reliably the consump-
tion of the starting thiazolethione 3b during short
irradiation times and this was used to quantify its
conversion upon irradiation.32 In contrast, in the case of
the pyridone 2b, the absorption by its photoproducts
Since the quantum yield of the decomposition is only
an upper-limit of the effective radical release, it was of
interest to determine the quantum yield of the N-O bond
scission ΦN-O, which is equal to ΦOiPr, the efficiency of
the production of isopropoxyl radicals. By means of
transient absorption spectroscopy (see Supporting Infor-
mation),26,33 the quantum yield of radical release from
the pyridone 2b in aqueous media was determined to be
(75 ( 8)%, for which the reported value for ꢀA (680 M-1
26
cm-1
)
at 390 nm in water was used.
For the thiazolethione 3b, the quantum yield of the
N-O bond scission could only be estimated, because the
extinction coefficient ꢀB of the thiyl radical B is not
known. With the assumption that its extinction coef-
ficient is the same as for the related pyridyl-2-thiyl
radical (ꢀPyS ) 600 M-1 cm-1),20 the quantum yield of
radical release from the thiazolethione 3b was estimated
to be (65 ( 7)%.
Ra d ica l Tr a p p in g by DMP O in th e P h otolysis of
th e P h oto-F en ton Rea gen ts 2b a n d 3b. Dep en d en ce
of Ra d ica l Tr a p p in g on th e Ra d ica l Sou r ce a n d th e
Solven t. To assess whether the pyridone 2b and the
thiazolethione 3b release alkoxyl radicals upon irradia-
tion, spin-trapping experiments with DMPO were per-
formed, followed by EPR-spectral detection (Figure 4).
Indeed, the photolysis in benzene led to isopropoxyl
radicals for both photo-Fenton reagents, as manifested
by their detection in the form of the spin-trapping adduct
with DMPO (Figures 4a and 4b). The g values and the
hyperfine coupling constants are consistent with the
reported ones for the isopropoxyl-radical spin adduct
DMPO-OiPr.18 In a mixture of H2O-MeCN (60:40),
again only the DMPO-OiPr was observed for the pho-
tolysis of the thiazolethione 3b (Figure 4c). In contrast,
in the case of the pyridone 2b, also EPR signals assign-
(30) Maillard, B.; Ingold, K. U.; Scaiano, J . C. J . Am. Chem. Soc.
1983, 105, 5, 5095-5099.
(31) Hatchard, C. G.; Parker, C. A. Proc. R. Soc. (London) 1956,
A235, 518-536.
(32) While this clean photochemical behavior may contradict the
results of the laser-flash experiments (Figure 2), in which substantial
secondary photolysis of the disulfide 5 product has been observed (see
Scheme 5), it should be kept in mind that in the laser-flash experiments
a much higher photon density applies, such that significantly more
disulfide 5 is formed through recombination of the thiyl radical B, and
hence photocleavage of the latter is much more pronounced than in
the steady-state photolysis (Figure 3).
6044 J . Org. Chem., Vol. 67, No. 17, 2002

Alkoxyl-Radical Release
F IGURE 4. EPR spectra recorded immediately after irradiation of the thiazolethione 3b (350 nm, 10 °C, 5 min) or the pyridone
2b (300 nm, 10 °C, 5 min) in various solvents and at different DMPO concentrations; DMPO-OiPr (b), DMPO-CMe2OH (0).
able to a carbon-centered radical appeared in addition
to those of DMPO-OiPr (Figure 4d). When the pyridone
2b was irradiated in pure water instead of a H2O-MeCN
mixture (60:40), the carbon-centered radical was detected
exclusively (Figure 4i).
Dep en d en ce of th e Ra d ica l Tr a p p in g on th e
DMP O Con cen tr a tion . To assess how effective the
radical trapping by DMPO may compete with the trans-
formation of the isopropoxyl radicals generated in the
photoreactions of the substrates 2b and 3b, their pho-
tolysis was conducted as a function of the DMPO con-
centration in a 60:40 H2O-MeCN mixture. In the case
of the thiazolethione 3b, its photolysis afforded exclu-
sively isopropoxyl radicals (Figure 4g) even at 9 mM of
DMPO. In contrast, in the photolysis of the pyridone 2b,
a decrease of the DMPO concentration increased sub-
stantially the fraction of the carbon-centered radicals in
the EPR spectra (Figure 4d, 4f, 4h).
An a lysis of th e EP R Sp ectr a . Generally, for primary
and secondary alkoxyl radicals, hydrogen shift dominates
â scission.34-36 For example, it is known that in aqueous
solution, the isopropoxyl radical readily undergoes a
formal 1,2-H shift to form the 2-hydroxyprop-2-yl radical
J . Org. Chem, Vol. 67, No. 17, 2002 6045

Adam et al.
F IGURE 5. Experimental and simulated EPR spectra of DMPO spin adducts with the isopropoxyl and 2-hydroxypropyl-2-yl
radicals: (a) DMPO-CMe2OH detected in the photolysis of acetone (480 mM) in 2-propanol (600 mM) and DMPO (90 mM) in a
H2O-MeCN mixture (60:40); (b) DMPO-OiPr detected in the photolysis of thiazolethione 3b (3.00 mM) and DMPO (27 mM); (c)
a mixture of DMPO-CMe2OH and DMPO-OiPr detected in the photolysis of pyridone 2b (3.00 mM) and DMPO (27 mM); Figs.-
(d) and (e) are the simulated spectra of DMPO-CMe2OH and DMPO-OiPr computed with the parameters acquired from the
experimental spectra (a and b); (f) the superposition of the simulated spectra (d and e) in the ratio of 1:3.
SCHEME 3. Ch em ica l Tr a n sfor m a tion of th e
Isop r op oxyl Ra d ica l
5 photoproduct was investigated by means of control
experiments. When the pyridone 2b was irradiated (300
nm) in the presence of the thiazolethione 3b or the
disulfide 9, exclusively isopropoxyl radicals were detected
as the spin-trapping adduct with DMPO (Scheme 4).
Therefore, it is concluded that the formation of the
DMPO adduct of the 2-hydroxyprop-2-yl radical was
strongly supressed by the photoproducts of the thi-
azolethione 3b, namely by the disulfide 5, as inferred
from the disulfide 9 as surrogate.
Str a n d -Br ea k F or m a tion in DNA. To assess the
DNA-photocleaving efficacy of the pyridone 2b and the
thiazolethione 3b, their photoinduced strand-break for-
mation in supercoiled pBR 322 DNA was examined. As
shown in Figure 6, both the pyridone 2b and the
thiazolethione 3b display moderate DNA cleavage, that
is, (17 ( 1)% open-circular DNA for 2b and (12 ( 1)%
for 3b (for convenience the blank was set to zero).
Addition of 2-propanol (10 vol %) suppressed almost
completely the strand-break formation by pyridone 2b
and thiazolethione 3b (Figure 6), which signifies that
radicals are responsible for the DNA damage. Also the
solvent effect on the DNA-cleaving activity was examined
for pyridone 2b in a H2O-MeCN mixture (60:40) and in
H2O. As shown in Figure 6, the irradiation of pyridone
2b in H2O gave about twice more open-circular DNA
[(32 ( 2)%] than in the 60:40 H2O-MeCN mixture
[(17 ( 1)%].
(Scheme 3).34 Thus, the observed carbon-centered radical
trapped during the photolysis of pyridone 2b was sus-
pected to be the 2-hydroxyprop-2-yl radical. To confirm
the identity of the carbon-centered radical as the corre-
sponding radical adduct of DMPO (DMPO-CMe2OH), it
was independently prepared by the photolysis of acetone
in the presence of 2-propanol and DMPO. Figure 5(a)
shows the resulting EPR spectrum of the DMPO-CMe2-
OH adduct, which is consistent with the reported one.37
Both EPR spectra, that of DMPO-CMe2OH generated
authentically (Figure 5a) and that of DMPO-OiPr ob-
served during the photolysis of the thiazolethione 3b
(Figure 5b), were simulated as shown in Figures 5d and
5e. The superposition of these two simulated spectra in
a 1:3 ratio (Figure 5f) reproduced exactly the experimen-
tal spectrum measured for the photolysis of pyridone 2b
and DMPO in a 60:40 mixture of H2O and MeCN (Figure
5c). Thus, the carbon-centered radical trapped by DMPO
during the irradiation of pyridone 2b was unequivocally
assigned to the 2-hydroxyprop-2-yl radical.
Ra d ica l Sca ven gin g by th e Th ia zoleth ion e 3b a n d
th e Disu lfid e 9. To clarify why the 2-hydroxyprop-2-yl
radicals derived from the isopropoxyl radicals were
observed only in the photolysis of the pyridone 2b, but
not in the case of the thiazolethione 3b, possible radical
scavenging by the thiazolethione 3b and/or its disulfide
Discu ssion
The EPR experiments (Figures 4 and 5) provide
unequivocal evidence that both the pyridone 2b and the
thiazolethione 3b generate isopropoxyl radicals upon
photoexcitation.
The most significant difference in the photochemical
properties between these two photo-Fenton reagents is
that in addition to the isopropoxyl radicals, the pyridone
(33) Bensasson, R. V.; Land, E. J . Trans. Faraday Soc. 1971, 67,
1904-1915.
(34) Gilbert, B. C.; Holmes, R. G. G.; Norman, R. O. C. J . Chem.
Res. 1977(M), S1, 0101-0109.
6046 J . Org. Chem., Vol. 67, No. 17, 2002

Alkoxyl-Radical Release
SCHEME 4. Effect of th e Th ia zoleth ion e 3b (1.0 m M) a n d th e Disu lfid e 9 (<0.5 m M) on th e EP R Sp ectr a of
P yr id on e 2b (1.0 m M) u p on P h otolysis in th e P r esen ce of DMP O (27 m M)
9, the 2-hydroxyprop-2-yl radical was not detected.
Evidently, in the photolysis of the thiazolethione 3b, the
carbon-centered radical is efficiently trapped presumably
by the disulfide product 5, but also intermediary thiyl
radicals may serve this purpose. Consequently, to provide
mechanistic insight on this point, the photochemistry of
the two photo-Fenton reagents 2b and 3b needs to be
considered.
The photochemical pathways shown in Scheme 5 are
proposed for the photoproducts derived from the thi-
azolethione 3b. The formation of the disulfide 5, the
bithiazyl 6, and the thiazole 7 may be explained analo-
gously to the reported photoreaction of N-hydroxy-
thiazolethione 3c.22 Dimerization of the thiyl radical B
gives the disulfide 5 (Scheme 5, path a), which on further
photolysis affords radicals C and D through the C-S
bond cleavage, a photoreactivity which was confirmed by
the transient spectrum of the disulfide 5 upon laser flash
photolysis (Figure 2d).38,39 The radical C, which may also
result from sulfur extrusion of the radical D,38 abstracts
F IGURE 6. Strand-break formation for the photolysis (312
nm, 0 °C, 30 min) of thiazolethione 3b (4.00 mM) or pyridone
a hydrogen atom to give the thiazole 7 or dimerizes to
2b (4.00 mM) in the presence of pBR 322 DNA (10 µg/mL) in
the bithiazyl 6. The unexpected isothiocyanate 8 may
result from ring-opening of the thiyl radical B to the thiyl
KH2PO4 buffer (5 mM, pH 7.4) and MeCN (60:40).
radical E (Scheme 5, path b) and subsequent radical
coupling between the thiyl radicals C and E.
2b affords also the carbon-centered 2-hydroxyprop-2-yl
In regard to the trapping of the 2-hydroxyprop-2-yl
radical, especially in aqueous media. The 2-hydroxyprop-
radical, formed by transformation of the released iso-
2-yl radicals results from H shift of the isopropoxyl
propoxyl radical, a likely scavenger is the disulfide 5
radicals (Scheme 3); such transformation is known for
which may afford hemithioacetal 10 upon SH2 reaction.40
secondary alkoxyl radicals and is facilitated in water as
Furthermore, also the thiyl radical B or the perthiyl
radical D may serve as trapping agent.
The photoproducts of the pyridone 2b are accounted
reaction medium.34,36 However, the 2-hydroxyprop-2-yl
radical was not detected in the photolysis of the thi-
azolethione 3b even at low DMPO concentration (Figure
4g), although isopropoxyl radicals have been generated.
Presumably, the 2-hydroxyprop-2-yl radical, if formed
for on the basis of the reported photoreactions of N-oxyl
pyridones (Scheme 6).41 Upon photoexcitation, N-O
homolysis and subsequent hydrogen-atom abstraction by
through the transformation of the released isopropoxyl
the pyridyloxyl radical A, followed by combination of the
radical, are scavenged in the photoreaction of the thi-
radical pair, provides the observed products 4a ,b. The
azolethione 3b. Indeed, control experiments with the
high value of 75% for the quantum yield of the N-O bond
scission ΦN-O reveals that this cleavage is the dominant
pyridone 2b confirmed (Scheme 4) that for the photolysis
in the presence of the thiazolethione 3b or the disulfide
(35) von Sonntag, C.; Schuchmann, H. P. Angew. Chem. 1991, 103,
1255-1279.
(36) Konya, K. G.; Paul, T.; Lin, S.; Lusztyk, J .; Ingold, K. U. J .
Am. Chem. Soc. 2000, 122, 7518-7527.
(37) Chiu, T.-M.; Siemiarczuk, A.; Wong, S. K.; Bolton, J . R. J . Phys.
Chem. 1985, 89, 3343-3347.
(38) Abdou, W. M.; Sidky, M. M.; Wamhoff, H. Z. Naturforsch. 1987,
42b, 1153-1158.
J . Org. Chem, Vol. 67, No. 17, 2002 6047

Adam et al.
SCHEME 5. P r op osed Mech a n ism for th e P h otop r od u cts of th e Th ia zoleth ion e 3b P h otolysis
SCHEME 6. P r op osed Mech a n ism for th e
F or m a tion of th e P h otop r od u cts in th e P h otolysis
of th e P yr id on e 2b
activity (Figure 6). In the presence of 2-propanol (10 vol
%), the strand-break formation is suppressed, which
confirms that radical species are generated in the pho-
tolysis of both photo-Fenton reagents and are responsible
for the DNA damage.
The pertinent question relates to which radical is
responsible for the DNA damage. For the photolysis of
pyridone 2b in aqueous medium, the 2-hydroxyprop-2-
yl radical was exclusively detected as the spin-trapping
adduct with DMPO (Figure 4i). Under these conditions,
a 2-fold higher DNA-cleaving activity was observed
compared to that in a 60:40 H2O-MeCN mixture (Figure
6). Since these DNA oxidations are run under atmo-
spheric conditions, presumably the 2-hydroxyprop-2-yl
radicals couple with molecular oxygen to afford the more
active alkylperoxyl radicals as oxidant.42 Therefore, the
photoinduced DNA-damaging activity of the pyridone 2b
in water may be attributed to the damaging species
formed in the radical cascade of the isopropoxyl radicals,
with the involvement of molecular oxygen. Also in a
mixture of H2O-MeCN (60:40), the radical cascade is
operative because the 2-hydroxyprop-2-yl radical was
detected in the photolysis of the pyridone 2b under these
conditions (Figures 4d, 4f, 4h).
primary photochemical step in the decomposition path-
way of the pyridone 2b. The Norrish-Type II reaction is
of minor importance. Neither the pyridone 2b itself nor
its photoproducts 4a,b are expected to trap the 2-hydroxy-
prop-2-yl radical, and therefore the latter carbon-centered
radicals may be observed in the photolysis of the pyridone
2b, in addition to the isopropoxyl radical, by DMPO
trapping and EPR-spectral detection.
Since the quantum yield of radical release (ΦN-O) is
about twice (Table 1) that of the pyridone conversion
(Φdec), it is concluded that recombination of the pyridyl-
oxyl radical A with the isopropoxyl radical back to the
pyridone 2b is an efficient pathway. Consequently, the
In the case of the photolysis of thiazolethione 3b, in
contrast to the pyridone 2b, only isopropoxyl radicals
were detected under all conditions (Figures 4c, 4e, 4g),
since the 2-hydroxyprop-2-yl radical is efficiently scav-
enged in situ (Scheme 4). The moderate DNA-damaging
activity of the thiazolethione 3b is, thus, due to the
isopropoxyl radical that is generated in the photolysis of
Φ
dec quantum yield is a better measure of alkoxyl radicals
effective for oxidation purposes.
With the difference in the photochemical behavior of
the photo-Fenton reagents 2b and 3b mechanistically
rationalized, its consequence on the DNA-cleaving activ-
ity was of relevance. In a mixture of H2O-MeCN
(60:40), both reagents display moderate DNA-damaging
6048 J . Org. Chem., Vol. 67, No. 17, 2002

Alkoxyl-Radical Release
this photo-Fenton reagent. Clearly, the thiazolethione 3b
is a less effective photochemical DNA-cleaving agent than
the pyridone 2b, despite the fact that the thiazolethione
3b (Φdec ) 0.60) is about twice more efficiently photolyzed
than the pyridone 2b (Φdec ) 0.36), see Table 1.
radicals. The reason for this advantage over the pyridone
2b is in-situ trapping of any potentially generated carbon-
centered radicals by N-alkoxythiazolethione-derived pho-
toproducts such as the disulfide 5.
Ack n ow led gm en t. Generous financial support by
the Deutsche Forschungsgemeinschaft (SFB 172 “Mole-
kulare Mechanismen kanzerogener Prima¨rvera¨nderun-
gen”) and by the Fonds der Chemischen Industrie is
appreciated. The authors are grateful to the SC-NMR
Laboratory of Okayama University for the 1H (500 MHz)
and 13C (125 MHz) NMR measurements. H. Okamoto
thanks the Ministry of Education, Science, Sports and
Culture of J apan for a postdoctoral fellowship from
October 1999 to September 2000.
In summary, the isopropoxyl radicals themselves,
generated initially from the photo-Fenton reagents 2b
and 3b on photolysis, rearrange readily in aqueous media
by H shift to the 2-hydroxyprop-2-yl radicals, which on
coupling with molecular oxygen provide the peroxyl
radicals as DNA-damaging species.42 Most significant for
the purpose of assessing the potential deleterious effects
of alkoxyl radicals in photobiological applications, of the
two photo-Fenton reagents examined herein, the thi-
azolethione 3b serves as a clean source of isopropoxyl
Su p p or tin g In for m a tion Ava ila ble: Experimental de-
tails for the synthetic work, for the EPR and transient-
spectroscopy, and for the determination of strand breaks in
pBR 322 DNA. This material is available free of charge via
the Internet at http://pubs.acs.org.
(39) Gilbert, A.; Baggott, J . In Essentials of Molecular Photochem-
istry; Blackwell Scientific Publication: Oxford, 1991; pp 477-479.
(40) Crich, D.; Quintero, L. Chem. Rev. 1989, 89, 1413-1432.
(41) Katritzky, A. R.; Chapman, A. V.; Cook, M. J .; Millet, G. H. J .
Chem. Soc., Perkin Trans. 1 1980, 2743-2754.
(42) Mere´nyi, G.; Lind, J .; Engman, L. J . Chem. Soc., Perkin Trans.
2 1994, 2551-2553.
J O025856W
J . Org. Chem, Vol. 67, No. 17, 2002 6049