A comparative repair study of thymine- and uracil-photodimers with model compounds and a photolyase repair enzyme

Article abstract of DOI:10.1002/(SICI)1521-3765(20000103)6:1<62::AID-CHEM62>3.0.CO;2-7

Cyclobutane uridine and thymidine dimers with cis-syn-structure are DNA lesions, which are efficiently repaired in many species by DNA photolyases. The essential step of the repair reaction is a light driven electron transfer from a reduced FAD cofactor (

Full text of DOI:10.1002/(SICI)1521-3765(20000103)6:1<62::AID-CHEM62>3.0.CO;2-7

FULL PAPER
A Comparative Repair Study of Thymine- and Uracil-Photodimers with
Model Compounds and a Photolyase Repair Enzyme
Jens Butenandt,[a] Robert Epple,[a] Ernst-Udo Wallenborn,[b] Andre P. M. Eker,[c]
Volker Gramlich,[d] and Thomas Carell*[a]
Abstract: Cyclobutane uridine and thy-
midine dimers with cis-syn-structure are
DNA lesions, which are efficiently re-
paired in many species by DNA photo-
lyases. The essential step of the repair
reaction is a light driven electron trans-
order to investigate the cleavage vulner-
ability of thymine and uracil cyclobu-
tane photodimers outside the protein
environment, two model compounds,
containing a thymine or a uracil dimer
and a covalently connected flavin, were
prepared and comparatively investigat-
ed. Cleavage investigations under inter-
nal competition conditions revealed, in
contrast to all previous findings, faster
repair of the sterically less encumbered
uracil dimer. Stereoelectronic effects are
offered as a possible explanation. Ab
initio calculations and X-ray crystal
structure data reveal a different cyclo-
butane ring pucker of the uracil dimer,
which leads to a better overlap of the p*-
fer from
a
reduced FAD cofactor
C(4) O(4)-orbital
with
the
s*-
(FADH ) to the dimer lesion, which
splits spontaneously into the monomers.
Repair studies with UV-light damaged
DNA revealed significant rate differ-
ences for the various dimer lesions. In
particular the effect of the almost
eclipsed positioned methyl groups at
the thymidine cyclobutane dimer moiety
on the splitting rates is unknown. In
C(5) C(5')-orbital. Enzymatic studies
with a DNA photolyase (A. nidulans)
and oligonucleotides, which contain ei-
ther a uridine or a thymidine dimer
analogue, showed comparable repair
efficiencies for both dimer lesions. Un-
der internal competition conditions sig-
nificantly faster repair of uridine dimers
is observed.
Keywords: DNA photolyases
´
DNA repair ´ DNA structures ´
oligonucleotides ´ UV photolesions
DNA lesions, which may cause cell death and the degener-
ation of cells into tumor cells.[4±6] DNA photolyases are flavin-
dependent DNA-repair enzymes, which revert cyclobutane
dimers in the genome of many organisms. The basis of the
repair reaction is a light driven electron transfer from the
FADH cofactor to the dimer, which cleaves spontaneously as
its radical anion.[7, 8] The depletion of the ozone layer[9] and the
current threat of increasing skin cancer risks[10] as a result of
rising UV-irradiation levels is fueling the interest in the
mechanism of the light induced enzymatic repair reaction. In
this context, all parameters that 1) determine the efficient
molecular recognition of DNA lesions and 2) affect the
cleavage rates of the various photolesions, are of particular
interest. Both factors determine the repair rate and conse-
quently the chance of survival of organisms living exposed to
sunlight. We[11] and others[12] recently showed that cis-syn-
configured photolesions are much more vulnerable towards
electron transfer induced cleavage compared with trans-syn-
or trans-anti-photodimers. Investigations of Sancar and co-
workers revealed faster enzymatic repair of thymidine dimers
compared with uridine dimers. It was suggested that the two
almost eclipsed positioned methyl groups at the thymidine
cyclobutane ring destabilize the dimer leading to an increased
cleavage vulnerability.[13, 14] Recent quantum chemical studies
Introduction
UV irradiation of cells causes the formation of cis-syn-
cyclobutane pyrimidine dimers through a [2p‡2p] cyclo-
addition of two pyrimidines, located next to each other in the
DNA double strand.[1±3] The main lesions are thymidine and
cytidine dimers. Cytidine-containing dimers undergo a rapid
deamination with a half-life time of approximately 6 h to yield
the corresponding uridine-containing photoproducts.[4] The
uridine and thymidine photoproducts are pre-mutagenic
[a] Dr. T. Carell, Dr. J. Butenandt, Dr. R. Epple
Laboratorium für Organische Chemie, ETH-Zürich
Universitätstrasse 16, CH-8092 Zürich (Switzerland)
Fax : (‡ 41)1-632-1109
E-mail: tcarell@org.chem.ethz.ch
[b] Dr. E.-U. Wallenborn
Laboratorium für Physikalische Chemie, ETH-Zürich
Universitätstrasse 16, CH-8092 Zürich (Switzerland)
[c] Dr. A. P. M. Eker
Erasmus University Rotterdam
Department of Cell Biology and Genetics
P.O. Box 1738, NL-3000 DR Rotterdam (The Netherlands)
[d] Prof. Dr. V. Gramlich
Laboratorium für Kristallographie, ETH-Zürich
Universitätstrasse 16, CH-8092 Zürich (Switzerland)
62
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0601-0062 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 1

62±72
of cyclobutane uracil and thy-
mine dimers support the desta-
bilizing effect of the two methyl
groups, but question significant
changes of the dimer reactivi-
ty.[15, 16] Ultrafast spectroscopic
studies by Michel-Beyerle and
calculations by Rösch and co-
1. Reduction
2. Light
O
N
O
N
O
N
O
N
R
R
R
R
NH
HN
NH
HN
3. Reoxidation
O
O
O
O
O
O
O
O
O
N
O
N
N
N
N
N
N
N
N
H
H
H
N
H
N
N
O
O
workers led to the conclusion
that the different repair rates
might be caused by a different
binding geometry of uridine
and thymidine photodimers
within the enzyme active site.
1: R = H
2: R = Me
3: R = H
4: R = Me
5: R = H
6: R = Me
Scheme 1. Light induced cleavage reaction of the two cyclobutane uracil and thymine dimer model compounds 1
and 2 into the reaction products 3 ± 6.
Different binding geometries
could determine the repair rate
A. nidulans photolyase[19] and uridine and thymidine dimer-
containing oligonucleotides.[20, 21] This allowed the determi-
nation of the repair rate for both photolesions and enabled to
gain initial insight into the molecular recognition of both
DNA lesions.
by influencing the electron transfer processes between the
FADH cofactor and the corresponding lesion.[17, 18]
In order to investigate the molecular reasons for the
different repair efficiencies in detail, we studied the cleavage
rates of uracil and thymine dimers inside and outside the
photolyase active site. To this end, the two model compounds
1 and 2 (Scheme 1) were prepared and investigated. In
particular, their ability to cleave, in the reduced form, upon
exposure to light into 3 ± 6 was measured. The obtained
cleavage data were compared with competitive and non-
competitive enzymatic repair data obtained with the
Results and Discussion
Synthesis of the model compounds 1 and 2: The preparation
of the cis-syn-uracil model compound 1 was performed as
described recently[11] from the cis-syn-uracil dimer diacid
7,[22±24] flavin 8 and pentylamine 9 (see Scheme 4). For the
synthesis of the cis-syn-thymine model compound 2, N(1)-
carboxymethyl-thymine (10) was esterified with benzyl alco-
hol, as depicted in Scheme 2, to give the thymine benzyl ester
11. The ester 11 was irradiated in acetone in a standard Pyrex-
photochemical apparatus with a medium pressure mercury
lamp (l > 300 nm) as the light source. During the irradiation a
mixture of all four possible isomers 12 ± 15 was formed. These
were separated by using a combination of selective precip-
itation, chromatography, and recrystallization as described in
the experimental part. Starting material 11 (12 g) yielded the
desired cis-syn-compound 12 (on average 300 mg) and the
trans-syn-dimer 13 (approx. 4.5 g).
The X-ray crystal structure analysis (Figure 1) of all four
thymine photodimers 12 ± 15 allowed the correct structural
assignment. Subsequent hydrogenolytic cleavage of the
benzyl esters in both compounds 12 and 13 gave the cis-syn-
and trans-syn-thymine dimer dicarboxylic acids 16 and 17,
required for this study, in excellent yields.
The flavin ethylamine building block 8 was prepared as
depicted in Scheme 3. ipso-Substitution with dinitrodimethyl-
benzene 18[25] and mono-Boc-protected ethylene diamine 19
furnished the nitroaniline 20 after chromatography and
recrystallization from ethanol in 67%.[26, 27] Compound 19
was prepared in one step from ethylene diamine and Boc2O.[28]
Hydrogenolytic reduction of 20 to 21 and condensation of 21
with alloxan and boric acid in acetic acid, following a general
flavin synthesis protocol of Kuhn and co-workers,[29, 30]
furnished the flavin compound 22 (80%). Alkylation at
N(3) in 22 with ethyl iodide and Cs2CO3 in dimethylforma-
mide (DMF) gave the flavin derivative 23 after chromatog-
Abstract in German: Cyclobutan Uridine- und Thymidindi-
mere mit cis-syn-Struktur sind DNA-Schäden, die in vielen
Spezies durch DNA-Photolyasen effizient repariert werden.
Der entscheidende Schritt der Reparaturreaktion ist ein licht-
getriebener Elektronentransfer von einem reduzierten FAD-
Cofaktor (FADH ) auf das Dimer, welches als Radikalanion
spontan monomerisiert. Reparaturstudien mit UV-geschädigter
DNA ergaben signifikante Unterschiede in den Reparaturraten
der verschiedenen Cyclobutan-Dimeren. Unbekannt ist vor
allem der Effekt, der zwei beinahe ekliptisch angeordneten
Methylgruppen des Thymidin Dimers auf die Reparaturrate.
Um die Spaltungseffizienz von Cyclobutan Uracil- und
Thymindimeren zu untersuchen, wurden zwei Modellverbin-
dungen dargestellt. Eine vergleichende Spaltungs- (Reparatur)
Studie unter internen Konkurrenzbedingungen ergab, im
Gegensatz zu allen bisherigen Messungen, eine wesentlich
schnellere Reparatur des sterisch weniger belasteten Uracildi-
mers. Als Erklärung schlagen wir stereoelektronische Gründe
vor. So zeigen ab initio Berechnungen und Röntgenkristall-
strukturdaten eine etwas andere Verzerrung des Uracil-Cyclo-
butanrings. Dies führt zu einer besseren Überlappung des p*-
C(4) O(4) Orbitals mit dem s*-C(5) C(5') Orbital. Enzyma-
tische Untersuchungen mit einer DNA-Photolyase (A. nidu-
lans) und mit Oligonukleotiden, die entweder ein Uridin- oder
ein Thymidindimeranalogon enthalten, zeigen vergleichbare
Reparaturraten für beide Dimere in der aktiven Tasche. Unter
internen Konkurrenzbedingungen wird im Gegensatz hierzu,
eine signifikant schnellere Reparatur der Uridindimeren beob-
achtet.
Chem. Eur. J. 2000, 6, No. 1
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0601-0063 $ 17.50+.50/0
63

T. Carell et al.
FULL PAPER
COOBn
O
COOBn
O
N
O
N
O
N
O
N
O
N
O
O
N
O
N
N
O
N
O
O
NH
ii
HN
NH
HN
NH
NH
i
HN
HN
NH
NH
O
O
O
O
O
N
O
O
O
COOBn
ROOC
COOR
ROOC
COOR
COOH
BnOOC
BnOOC
15
14
11
10
12: R = Bn
16: R = H
13: R = Bn
17: R = H
iii
iii
Scheme 2. Synthesis of the cyclobutane thymine dimer building blocks 16 and 17 starting from 10 required for the preparation of the model compounds 2 and
24. i) BnOH, carbonyldiimidazole (CDI), DMF, RT; ii) hnÄ > 300 nm, acetone, RT; iii) 10% Pd/C catalyst, H2, HOAc, RT.
Figure 1. X-ray crystal structures of all four thymine photodimers 12 ± 15. ORTEP plots of the molecular structures. Arbitrary numbering. Displacement
ellipsoids are shown at the 50% probability level.
O
cleavage rates (Scheme 1) of
NHR1
BocHN
NHBoc
O
O
the uracil and the thymine pho-
todimer-containing model com-
pounds 1 and 2 was performed
under strictly identical condi-
tions. To this end 1 and 2 were
dissolved in various solvents
(10 5 m or 10 6 m). A one to
one mixture was prepared and
filled into a quartz cuvette stop-
pered with a rubber septum and
equipped with a magnetic stir-
rer. A small amount of triethyl-
NH
NH2
NO2
NO2
N
O
19
N
N
N
O
NH
H
N
R2
i
R
iii
O
18
22: R1 = Boc, R2 = H
23: R1 = Boc, R2 = Et
8: R1 = H, R2 = Et
20: R = NO2
21: R = NH2
ii
iv
v
Scheme 3. Synthesis of the aminoethyl substituted flavin building block 8. i) pyridine, 908C, 24h; ii) 10% Pd/C
catalyst, H2, HOAc, RT; iii) HOAc, B(OH)3, RT; iv) EtI, Cs2CO3, DMF, RT; v) TFA, H2O.
amine (25 mL) was added to ensure deprotonation of the
reduced flavin cofactor. The solution was purged with nitro-
gen for 20 min in order to establish anaerobic conditions. To
reduce the flavin chromophore, 20 mL of an aqueous sodium
dithionite solution (0.05m) was added. The assay solution was
stirred and irradiated with monochromatic light (l ˆ 366 nm).
During the irradiation experiment, five to ten 50 mL samples
were removed from the assay solution with a micro syringe
after defined time intervals. The samples were filled into small
vials and subsequently vigorously shaken to reoxidize the
flavin cofactor. This helps to stop the cleavage reaction, while
the sample is analyzed. All samples were analyzed by
reversed-phase HPLC in order to quantify the amount of
the model compounds 1 and 2 and of the expected photosplit
products 3 and 4.[11] A representative set of HPLC chromato-
grams showing the simultaneous cleavage of both model
compounds 1 and 2 into 3 and 4 is depicted in Figure 2.
raphy in 71% yield. Cleavage of the tert-butyloxycarbonyl
(Boc)-protecting group furnished the flavin ethylamine 8
(97%) in an overall yield of 38%.
The preparation of the model compounds 1, 2, and 24
(Scheme 4) required activation of the carboxylic acids of 7, 16,
and 17 with benzotriazol-1-yl-oxy-tris-(dimethylamino)-phos-
phonium hexafluorophosphate (BOP)[31] in DMF. After
addition of one equivalent of the flavin ethylamine 8 and
diisopropylethylamine (DIEA), a large excess of pentylamine
9 was added to each reaction mixture. The model compounds
1, 2, and 24 (prepared for comparison reasons) were obtained
after chromatography on silica gel-H in approximately 30%
yield as yellow powders.
Cleavage studies with the uracil dimer- and thymine dimer-
containing model compounds 1 and 2: Analysis of the
64
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0601-0064 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 1

Thymine- and Uracil-Photodimers
62±72
2 ± 3 times more rapidly compared
with the corresponding thymine
dimer. In order to determine the
quantum yields for the reactions,
the number of photons emitted by
the light source into the sample
was determined by ferrioxalate
actinometry.[32, 33] Together with
the absorption of the samples,
which were determined by UV/
O
N
O
N
R
R
HN
NH
O
N
O
N
O
N
O
N
O
O
O
O
R
R
O
N
HN
NH
HN
NH
or
N
N
N
N
H
H
O
O
O
O
N
O
HOOC
COOH HOOC
COOH
1: R = H
17
16: R = Me
7: R = H
2: R = Me
i
or
NH2
Vis-spectroscopy
(eflavin,red,366nm
ˆ
O
N
O
N
NH2
1
4900 Lmol 1 cm ), this value al-
lowed the calculation of the quan-
tum yield F (F ˆ number of react-
ed molecules per number of absor-
bed photons). For the uracil dimer
a quantum yield of approximately
4.5% was determined in ethylene
glycol as the solvent. The thymine
dimer cleaved significantly slower
with a quantum yield of about
2.1% (Figure 4).
HN
NH
N
N
O
O
+
O
O
O
O
N
N
O
N
8
9
N
N
N
N
H
H
N
O
24
Scheme 4. Synthesis of the flavin-containing photolyase model compounds 1, 2 and 24. i) BOP, TEA, DMF,
RT.
Figure 3. Time-dependent formation of the cleaved photoproducts 3 and 4
during the irradiation of a mixture containing 1 and 2. Irradiation was
Figure 2. Series of reversed-phase HPLC chromatograms obtained during
the light induced conversion of the model compounds 1 and 2 into the
cleaved products 3 and 4 (ethylene glycol). The chromatograms show 1 and
2 eluting at approximately 18 min and 21 min, respectively. The cleavage
products 3 and 4 elute at approximately 6 min and 8 min, respectively.
Conditions: Lichrosphere column (250 Â 4) 100/5. Gradient: water/meth-
anol from 70:30 to 10:90 over 30 min. Detection at 450 nm.
&
*
performed at 366 nm in ethylene glycol at 258C. : 3, : 4.
Evidently, the two model compounds 1 and 2, with
retention times of 18 min (1) and 21 min (2), react cleanly to
the photosplit products 3 and 4 with retention times of 6 min
and 8 min, respectively. The quantitative cleavage rate of both
model compounds was obtained from plots of the yields of 3
and 4 against the irradiation time. One representative
diagram is shown in Figure 3. The plot shows, in contrast to
the current belief, a significantly faster cleavage of the
sterically less strained uracil dimer model compound 1. The
cyclobutane uracil dimer was found to cleave by a factor of
Figure 4. Bar diagrams showing the quantum yields (F) for the cleavage
process of 1 and 2 reacting to 3 and 4 in various solvents. DMF ˆ dime-
thylformamide, Etgly ˆ ethylene glycol, EtOH ˆ ethanol, EtOAc ˆ ethyl
acetate. Irradiation was performed at l ˆ 366 nm at 258C. Dark grey bars:
uracil model compound 1. Light grey bars: thymine model compound 2.
Chem. Eur. J. 2000, 6, No. 1
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0601-0065 $ 17.50+.50/0
65

T. Carell et al.
FULL PAPER
In order to exclude that the surprising cleavage rate
difference is influenced by the chosen medium, all measure-
ments were repeated in a variety of solvents including water,
ethanol, and ethyl acetate (plus 25 mL triethylamine each).[34]
Since the sodium dithionite solution is not miscible with most
of these solvents, we reduced the flavin unit by catalytic
hydrogenation. To achieve complete reduction, a small
amount of Pd/BaSO4 was added into the cuvette as the
hydrogenation catalyst. The cuvettes were again purged with
nitrogen. Then a stream of hydrogen was passed through the
solution. Complete reduction and deprotonation of the flavin
cofactor was monitored by UV spectroscopy.[11] The assay
solutions were subsequently irradiated at 366 nm and ana-
lyzed as described above. The obtained cleavage quantum
yields are compiled in Figure 4. We observe moderately
reduced cleavage efficiencies in nonpolar solvents, in agree-
ment with earlier data.[11] The fastest cleavage is detected in
pure water at pH > 7. No cleavage takes place at pH 3 in
agreement with the absolute necessity to deprotonate the
reduced flavin chromophore (pKa 6.3).[11, 35] Most important,
however, is the observation that the uracil dimer-containing
model compound 1 cleaves in all investigated solvents by a
factor of approximately 2 more efficient. The data obtained in
these experiments clearly disprove the current opinion that
the two ªeclipsedº positioned methyl groups at the thymine
dimer cyclobutane ring enforce faster cleavage. In contrast,
the sterically less encumbered uracil dimer possesses the
higher reductive cleavage vulnerability.
Figure 5. Time-dependent formation of the cleaved photoproduct 4 during
the irradiation of 2 and 24. Irradiation was performed at l ˆ 366 nm in
&
*
ethylene glycol at 258C. : 4 formed upon irradiation of 2. : 4 formed
upon irradiation of 24.
For comparison we also investigated the cleavage rate of
the trans-syn-thymine dimer model compound 24. The result
of this study is depicted in Figure 5. Clearly evident is the
reduced cleavage rate of the trans-syn-dimer (factor of 2)
compared with the cis-syn-configured thymine dimer in model
compound 2. This result is in excellent agreement with data
from earlier studies, which showed that trans-syn-configured
dimers are more stable towards reductive ring opening.[11] This
result underlines that the cleavage rate is strongly influenced
by the substitution pattern and the configuration of the dimer
unit.
33 and 34 and the single crystal structure of the thymidine
dimer analogue 34 is depicted in Figure 6. In order to
investigate the enzymatic repair reaction, the synthesized
oligonucleotides were dissolved in 500 mL of a photoreacti-
vation buffer (see Experimental Section). The solutions were
filled into a quartz cuvette equipped with a magnetic stirrer.
After addition of 50 mL of an enzyme stock solution, the
cuvette contained 5 Â 10 6 m, 10 6 m, or 10 7 m DNA, and 2.5 Â
10 8 m or 5 Â 10 8 m enzyme. The corresponding solution was
stirred and irradiated with monochromatic light at 435 nm.
20 mL samples were removed from the assay solution after
defined time intervals. Each sample was diluted with 10 mL of
0.1m acetic acid, to ensure that the cleavage reaction remains
stopped during sample analysis. The samples were subse-
quently vigorously stirred and subjected to reversed-phase or
ion-exchange HPLC analysis (nucleosil column 120/3 or
nucleogel SAX). This allowed us to quantify the amount of
damage-containing oligonucleotide and of enzymatically
repaired product.
Enzymatic studies: In order to determine the enzymatic
cleavage rates of uridine and thymidine DNA photolesions,
we prepared a series of lesion-containing oligonucleotides and
investigated their repair efficiencies (Scheme 5). The oligo-
nucleotides 25 ± 28, containing uridine or thymidine photo-
lesion analogues in different base sequence contexts, which
are expected to be repaired resulting in 29 ± 32, were prepared
by using DNA-lesion phosphoramidite building blocks
and solid-phase oligonucleotide
i
synthesis. As the lesion ana-
logues we used the recently
25: 5'-d(CGACGT=TGCAGC)-3'
29: 5'-d(CGACGTTGCAGC)-3'
30: 5'-d(CGACGUUGCAGC)-3'
31: 5'-d(CGTATTTATTCTGC)-3'
32: 5'-d(CGTATUUATTCTGC)-3'
i
i
26: 5'-d(CGACGU=UGCAGC)-3'
described formacetal-linked ur-
idine and thymidine dimers,
which are readily available in
larger quantities. Both are
27: 5'-d(CGTATT=TATTCTGC)-3'
28: 5'-d(CGTATU=UATTCTGC)-3'
i
excellent
photolyase
sub-
Scheme 5. Depiction of the prepared oligonucleotides 25 ± 32. 25 ± 28 contain either a uridine dimer analogue (26
and 28) or a thymidine analogue (25 and 27). Depiction of the repair reaction giving the repaired oligonucleotides
29 ± 32. i) A. nidulans photolyase, hnÄ ˆ 435 nm.
strates.[20, 21] The molecular
structures of the building blocks
66
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0601-0066 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 1

Thymine- and Uracil-Photodimers
62±72
investigated under direct competition conditions. To this end,
uridine and the thymidine dimer-containing oligonucleotides
were mixed [(25‡28), (26‡27), (25‡26), (27‡28)] and used
as a substrate mixture in the enzymatic assay. The assay was
performed as described above. A fine-tuned HPLC gradient
enabled the separation of all four oligonucleotides (the two
oligonucleotide reactants and products). Under these com-
petitive conditions each oligonucleotide functions as a com-
petitive inhibitor for the other, which allowed us to examine
how efficiently both substrates are recognized by the repair
enzyme. The repair data (t1/2 values) obtained from these
measurements are listed in Table 2. In the competition
experiments we observed consistently significantly faster
repair of the uridine type photolesion. The results show that
if the enzyme is confronted with both photolesions at the same
time, it repairs the uridine dimer 50 ± 70% more efficiently
compared with the thymidine photolesion.
Figure 6. Depiction of the molecular structures of the uridine and
thymidine photolesion analogues 33 and 34 used for the enzymatic studies.
X-ray crystal structure of the thymidine dimer 34. For the crystal structure
of 33 see ref. [21]. ORTEP plot of the molecular structure. Arbitrary
numbering. Displacement ellipsoids are shown at the 50% probability
level.
Table 2. Competitive repair investigation of the oligonucleotides 25 ± 28 with
the A. nidulans photolyase. Irradiation at 435 nm. Buffer: 100mm NaCl, 10mm
KH2PO4, 5mm DTT, pH 7.0. t1/2 is the time at which 50% of the damaged
oligonucleotide is repaired.
The product peak was integrated and the value plotted
against the photoreactivation time. The time point at which
50% of the damaged oligonucleotide was repaired was
determined (t1/2). These t1/2 values are listed in Table 1. The
ˆ
ˆ
Substrate cDNA [m]
cenzyme [m] t1/2 (T T) [min] t1/2 (U U) [min]
6
8
25‡28[c]
26‡27[c]
27‡28[d]
25‡26[d]
0.5 Â 10 each 5 Â 10
4.0 Æ 0.5
5.5 Æ 0.5
21 Æ 1
2.7 Æ 0.5[a]
3.8 Æ 0.5[a]
13 Æ 1[b]
6
8
Table 1. Repair investigation with the oligonucleotides 25 ± 28 and the A.
nidulans photolyase. Irradiation at 435 nm. Buffer: 100mm NaCl, 10mm
KH2PO4, 5mm DTT, pH 7. 0. t1/2 is the time at which 50% of the damaged
oligonucleotide is repaired.
0.5 Â 10 each 5 Â 10
6
8
8
2.5 Â 10 each 2.5 Â 10
6
2.5 Â 10 each 2.5 Â 10
24 Æ 1
15 Æ 1[b]
Bandwidth: [a] 1.7 nm. [b] 5.1 nm. [c] Analysis performed using ion-exchange
chromatography. [d] Analysis performed using reversed-phase chromatogra-
phy.
Substrate
cDNA [m]
cenzyme [m]
t1/2 [min]
7
8
25[c]
25[c]
25[d]
26[c]
26[c]
26[d]
27[c]
27[c]
27[d]
28[c]
28[c]
28[d]
10
10
5 Â 10
2.2 Æ 0.3[a]
11 Æ 1[a]
6
8
5 Â 10
6
8
5 Â 10
2.5 Â 10
23 Æ 1[b]
2.1 Æ 0.2[a]
12 Æ 1[a]
The noncompetitive enzymatic studies clearly show
that the two ªeclipsedº methyl groups at the thymidine
cyclobutane moiety do not enforce increased cleavage. Both
photolesions are repaired with similar rates. The data show
that once bound by the enzyme the lesion cleavage rate is
identical.
The competitively determined data show that the microbial
photolyase[37, 38] used possesses an active site that allows more
efficient binding of uridine-type photolesions. The data
support the hypothesis that the binding geometry within the
active site and not the substitution pattern of the cyclobutane
moiety determines the higher enzymatic repair rate of uridine
type DNA lesions. This result is in good agreement with
recent ultrashort time spectroscopic data of Michel-Beyerle[18]
and calculation of Rösch[17], which suggest that the two
thymidine methyl groups hinder the thymidine lesion to
penetrate as deep as the uridine substrate into the active
site.[18]
7
8
10
5 Â 10
6
8
10
5 Â 10
6
8
5 Â 10
2.5 Â 10
21 Æ 1[b]
1.8 Æ 0.2[a]
12 Æ 1[a]
7
8
10
5 Â 10
6
8
10
5 Â 10
6
8
5 Â 10
2.5 Â 10
14 Æ 1[b]
1.7 Æ 0.2[a]
10 Æ 1[a]
7
8
10
5 Â 10
6
8
10
5 Â 10
6
8
5 Â 10
2.5 Â 10
13 Æ 1[b]
Bandwidth: [a] 1.7 nm, [b] 5.1 nm. [c] Analysis performed using ion-
exchange chromatography. [d] Analysis performed using reversed-phase
chromatography.
t1/2 data show that all four oligonucleotides 25 ± 28 are repaired
with almost identical efficiency. Throughout all measure-
ments, we determined a similar repair rate for uridine and
thymidine dimer-containing oligonucleotides. The observed
repair differences are small and close to the error limit of our
assay. Since the oligonucleotide concentration was chosen to
be far higher (0.1 ± 5mm) compared with the expected Km
values (Kd in the nm range),[36] we can assume that the enzyme
operates under saturation conditions. All experiments were
performed additionally at three different substrate concen-
trations (0.1mm, 1mm, and 5mm). We always obtained com-
parable results. In summary, the data show that uridine and
thymidine photodimers are enzymatically cleaved with similar
repair efficiencies.
Calculations: In order to investigate how the structure of the
dimer influences the cleavage rate ab initio calculations were
performed[11] with the X-ray crystal structural data of 12 and
13 as input geometries for the uracil and thymine dimers in
their neutral and radical anion states. The goal of the
calculations was to differentiate how structural factors and
the binding situation of the dimer lesion in the active site
modulate the repair efficiency. To keep the calculations
feasible and allow comparison with the uridine and thymidine
In order to obtain more data about the enzymatic reaction,
the repair of the damaged oligonucleotides 25 ± 28 were
Chem. Eur. J. 2000, 6, No. 1
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0601-0067 $ 17.50+.50/0
67

T. Carell et al.
FULL PAPER
lesion analogues, only truncated dimers (without the carboxy-
methyl substituents) were retained from the X-ray structures
(Figure 1). The positions of the hydrogen atoms were
subsequently determined by partial optimization, constrain-
ing all internal coordinates involving heavy atoms. The
currently accepted cleavage scenario includes the donation
Conclusion
Previous investigations of the repair of thymidine and uridine
DNA lesions by DNA photolyases revealed faster repair of
the thymidine dimer lesion. Suggested explanations include:
1) a possible additional destabilization of the dimer due to the
two almost eclipsed positioned methyl groups at the thymine
dimer cyclobutane ring and 2) a different positioning of the
thymidine dimer in the photolyase active site, which might
effect the electron transfer processes between the dimer
lesion and the FADH cofactor. In this study, the investiga-
tion of the intrinsic cleavage rates of cyclobutane uracil and
thymine dimers was performed with the model compounds 1
and 2. The obtained data clearly show a higher instability of
the uracil dimer radical anion. The uracil dimer radical anion
was found to cleave by a factor of 2 ± 3 faster under our
conditions. This observation is independent of the solvent
polarity. The presented measurements show that the two
ªeclipsedº positioned thymine methyl groups do not increase
the lability of the cyclobutane thymine dimer radical anion,
which is in good agreement with recent calculations.[15, 16]
Based on our own ab initio calculations of the Mayer bond
orders[41] of the neutral dimers and their radical anions, we
suggest that the cleavage rates are influenced by stereo-
electronic parameters. We believe that the better overlap
ˆ
of an electron first into the p*-orbital of the C(4) O(4)
double bond and subsequent delocalization of electron
density into the s*-C(5) C(5')-bond orbital[15, 39], which
reduces the activation barrier for the bond cleavage.[40] To
approximate the bond order decrease that occurs after the
electron uptake, Mayer bond orders[41] of the C(5) C(5') bond
were calculated in the neutral and radical anion photodimers.
The results of these calculation are listed in Table 3. The data
show a decrease of the C(5) C(5')-bond order after electron
Table 3. Calculated UHF/6 ± 31G* Mayer bond orders[41] for the neutral
dimers and their radical anions based on truncated X-ray crystal structures
of the dibenzyl esters 12 and 13 and for the uridine and thymidine lesion
analogues 33 and 34 as input geometries (hydrogen atoms optimized on the
RHF/3 ± 21G level). The calculations were performed with the GAME-
SS(US) package.[48]
Compound
C(5) C(5') Neutral dimer C(5) C(5') Anionic dimer
cis-syn-uracil
cis-syn-thymine
trans-syn-thymine 0.951
cis-syn-uridine 0.946
cis-syn-thymidine 0.940
0.948
0.943
0.775
0.835
0.828
0.777
0.807
ˆ
between the p*-orbital of the C(4) O(4) double bond and the
s*-C(5) C(5')-bond orbital in the uracil dimer enhances its
reductive cleavage vulnerability. Similar results were calcu-
lated for uridine and thymidine photolesions. These dimers
were incorporated into oligonucleotides and the repair was
investigated with the A. nidulans photolyase. Surprisingly we
found that both dimer lesion are repaired with almost
identical efficiency under substrate saturation conditions. In
direct competition experiments, however, significantly faster
repair of the uridine-lesion occured, which indicates that
cyclobutane uridine photolesions are recognized by the
enzyme with higher affinity.
donation to all dimers. This bond order decrease is signifi-
cantly stronger for the cis-syn-uracil dimers compared with
the cis-syn- and the trans-syn-thymine dimers. The calculated
data are in good agreement with the experimentally observed
increased stability of both thymine dimer radical anions in the
model compounds. The result shows that the higher reactivity
of the uracil dimer is influenced by the structure of the dimer
unit, which determines the overlap of the participating
orbitals.
Future studies are now required in order to learn how the
DNA sequence and the DNA structure modulates the repair
rate. Exact binding constants for the different substrates need
to be determined. Data from such experiments could provide
further information about the lesion ªflippingº process,[20, 28, 44]
which seems to be involved in the molecular recognition of
photolesions in a DNA-double strand.[45, 46]
In order to obtain information about the enzymatic splitting
vulnerability of thymidine and uridine dimer, similar calcu-
lations (Table 3) were performed for the cis-syn-thymidine
and the cis-syn-uridine photodimers, which were used for the
enzymatic repair studies. Here again, all calculations were
performed by using truncated and partially optimized X-ray
crystal structures (Figure 6). The obtained Mayer bond orders
suggest again faster uridine dimer repair. This, however, is not
observed. Both dimers were found to be repairedÐunder
noncompetitive conditionsÐwith almost identical efficiency
within the accuracy limit of our assay. We explain the
discrepancy between the calculated cleavage vulnerabilities
and the enzymatic rate data with a possible structural
alteration of the dimers if they are placed in the DNA double
strand and are bound in the active site. This hypothesis is
supported by the observation that the critical cyclobutane
dimer ring pucker changes from CB -observed in all X-ray
Experimental Section
General methods: All materials were obtained from commercial suppliers
and were used without further purification. Solvents of technical quality
were distilled prior to use. The aqueous buffers were prepared by using
deionized distilled water. For reactions under an inert gas atmosphere,
nitrogen of standard quality was used. For analytical thin-layer chroma-
tography, precoated glass silica gel plates (Merck 60F254) were used. The
amino compounds were stained with a ninhydrin solution. Flash chroma-
tography was performed on silica gel (Merck 0.040 ± 0.063 mm) and silica
gel-H (Fluka, 0.005 ± 0.040 mm). Melting points are uncorrected and were
determined on a Büchi Smp 20. IR spectra were recorded on a Perkin ±
Elmer 1600 and FT-IR with KBr pellets or CHCl3 solutions. UV spectra
were recorded on a Varian Cary 5 UV/Vis spectrophotometer in 1 cm
quartz cuvettes. Fluorescence spectra and irradiation experiments were
‡
crystal structures of the dimer monomers-[21, 42] to CB in a
DNA duplex environment.[43]
68
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0601-0068 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 1

Thymine- and Uracil-Photodimers
62±72
performed with a Spex 1680 0.22 m double Spectrometer in 1 cm quartz
cuvettes. NMR spectra were recorded on a Varian Gemini 200 (200 MHz
(1H), 50 MHz (13C)), Varian Gemini 300 (300 MHz (1H), 75 MHz (13C)) and
a Bruker AM500 (500 MHz (1H), 125 MHz (13C)). The chemical shift (d) is
reported in ppm downfield from tetramethylsilane (TMS, d ˆ 0). Alter-
natively, the resonance of residual solvent protons were used as the
reference. EI-mass spectra, FAB-mass spectra, MALDI-TOF-mass spectra
and ESI-mass spectra were measured by the staff of the mass spectrometry
facilities of the ETH Zürich on a Hitachi ± Perkin ± Elmer VG TRIBRID
with 70 eV ionization energy (EI), on a ZAB-2 SEQ with 3-nitrobenzyl
alcohol as the matrix (FAB), on a Bruker Reflex spectrometer and on a
Finnigan TSQ 7000. Elemental analysis were performed by the micro-
analysis laboratory of the ETH Zürich. HPLC chromatograms were
obtained with a Knauer HPLC instrument (HPLC pumps 64, Knauer
variable wavelength UV detector, Knauer degasser) with HPLC grade
solvents.
room temperature until a clear solution was obtained. Chloroacetic acid
(15.00 g, 159 mmol) was added and the reaction mixture was heated to
reflux for 90 min. The cold reaction mixture was acidified with conc. HCl
until pH 1 was reached and kept at 48C overnight. The crystallized product
10 was filtered off and dried in vacuo over P2O5 (12.50 g, 86%). M.p.: 285-
2878C (Lit.: 2878C[47]); 1H NMR (200 MHz, [D6]DMSO): d ˆ 4.42 (s, 3H),
5.20 (s, 2H), 5.62 (d, J ˆ 7.9 Hz, 1H), 7.60 (d, J ˆ 7.9 Hz, 1H), 11.33 (s, 1H).
Benzyl-1,2,3,4-tetrahydro-5-methyl-2,4-dioxopyrimidine-1-acetate
(11):
1,2,3,4-Tetrahydro-5-methyl-2,4-dioxopyrimidine-1-acetic acid (10) (10 g,
54.3 mmol) and 1,1'-carbonylbis[1H]-diimidazole (12 g, 74.0 mmol) were
stirred in DMF (80 mL) at room temperature. After 15 minutes, benzyl
alcohol (8 g, 76.1 mmol) was added and the reaction mixture was stirred for
35 h at RT. The reaction mixtures was concentrated in vacuo and the
residual material was suspended in H2O (50 mL) and filtered off.
Recrystallization of 11 from MeOH yielded 11 as colorless plates (14.5 g,
97%). Rf (CHCl3/MeOH 10:1): 0.65; m.p.: 179 ± 1818C; IR (KBr): 3156m,
3043m, 2944m, 1737s, 1693s, 1650s, 1458m, 1420m, 1387m, 1360m, 1224s,
1150m, 958m, 894w, 865w, 793w, 757s; 1H NMR (400 MHz, [D6]DMSO):
d ˆ 1.76 (s, 3H), 4.55 (s, 2H), 5.19 (s, 2H), 7.38 (m, 5H), 7.53 (s, 1H), 11.41
(s, 1H); 13C NMR (100 MHz, [D6]DMSO): d ˆ 11.79, 48.38, 66.34, 108.56,
127.86 (2C), 128.12, 128.38 (2C), 135.46, 141.45, 150.89, 164.21, 168.11; MS
X-ray crystal-structure data of 12: Colorless needles; C28H28N4O8, Mr ˆ
548.5; monoclinic, space group P2(1)/n: Dc ˆ 1.403 g cm 3, Z ˆ 4, a ˆ
15.003 (9), b ˆ 6.258(5), c ˆ 28.09(4) Š, a ˆ 90.00, b ˆ 99.08 (9), g ˆ 90.00,
Vˆ 2596 (5) Š3, CuKa (l ˆ 1.54178 Š) radiation, 2203 reflections collected,
Tˆ 293 K. The crystal structure was solved by direct methods (SHELXTL
PLUS) and refined by full-matrix least-squares analysis by using exper-
imental weights (heavy atoms anisotropic; H-atoms riding model, fixed
isotropic). Final (R(F) ˆ 0.0411, wR(F 2) ˆ 0.1063) for 362 variable and 2099
observed reflections (F > 3.0 s(F)).
‡
(EI): m/z (%): 274 (19) [M] , 167 (17), 139 (57), 91 (100), 65 (7), 41 (9), 28
(12); C14H14N2O4 (274.28): calcd: C 61.31, H 5.14, N 10.21, found: C 61.25, H
5.34, N 10.18.
Irradiation of 11: The benzyl ester 11 (2.00 g, 7.3 mmol) was dissolved in
acetone (200 mL) and the solution was degassed for 15 min in an ultrasonic
bath and another 15 min outside the ultrasonic bath by purging nitrogen
through the solution. The solution was irradiated for 3 h with a 150 W Hg
medium pressure lamp in a standard photochemical apparatus (Pyrex).
During the irradiation the solution was purged with nitrogen. The reaction
mixture was filtered, evaporated to dryness in vacuo and the product was
precipitated with acetone and filtrated. The residual solid material
contained mainly 13 and 15. The filtrate was subjected to column
chromatography (silica gel 60, d ˆ 3 cm, L ˆ 18 cm, CH2Cl2/acetone
7:1 !5:1) to yield the residual amounts of the isomers 13 and 15 well
separated as white powders and a mixture of the isomers 12 and 14.
Separation of the mixture, containing 12 and 14, was achieved by column
chromatography (silica gel 60, d ˆ 3 cm, L ˆ 18 cm CHCl3/methanol 15:1).
Recrystallization of the isomers 13 and 15 from HCOOH/H2O and of 12
and 14 from methanol yielded all four isomers in analytically pure form.
X-ray crystal-structure data of 13: Colorless needles; C28H28N4O8, Mr ˆ
548.5; monoclinic, space group C2/c: Dc ˆ 1.373 g cm 3, Z ˆ 4, a ˆ 18.174
(13), b ˆ 8.282 (3), c ˆ 18.732 (11) Š, a ˆ 90.00, b ˆ 109.78 (5), g ˆ 90.00,
Vˆ 2653 (3) Š3, MoKa (l ˆ 0.71073 Š) radiation, 2413 reflections collect-
ed, Tˆ 293 K. The crystal structure was solved by direct methods
(SHELXTL PLUS) and refined by full-matrix least-squares analysis by
using experimental weights (heavy atoms anisotropic; H-atoms riding
model, fixed isotropic). Final (R(F) ˆ 0.0492, wR(F 2) ˆ 0.1319) for 194
variable and 2337 observed reflections (F > 3.0s(F)).
X-Ray crystal-structure data of 14: Colorless needles; C28H28N4O8, Mr ˆ
548.5; monoclinic, space group P2(1)/c: Dc ˆ 1.383 g cm 3, Z ˆ 4, a ˆ 13.032
(13), b ˆ 16.156 (11), c ˆ 12.639(8) Š, a ˆ 90.00, b ˆ 98.02 (9), g ˆ 90.00,
Vˆ 2635 (4) Š3, MoKa (l ˆ 0.71073 Š) radiation, 3652 collected reflec-
tions, Tˆ 293 K. The crystal structure was solved by direct methods
(SHELXTL PLUS) and refined by full-matrix least-squares analysis by
using experimental weights (heavy atoms anisotropic; H-atoms riding
model, fixed isotropic). Final (R(F) ˆ 0.040, wR(F 2) ˆ 0.097) for 362
variable and 3463 observed reflections (F > 3.0s(F)).
Dibenzyl-cis-[4a]-transoid-[4a,4b]-cis-[4b]-dodecahydro-9,11-dimethyl-
2,4,6,8-tetraoxocyclobuta[1,2-d:3,4-d']dipyrimidine-1,5-diacetate
(15):
200 mg (35%); Rf (CHCl3/MeOH 10:1) ˆ 0.73; m.p.: 242 ± 2448C; IR
(KBr): 3429m, 3233m, 3067w, 1744m, 1711s, 1683s, 1473m, 1389m, 1356w,
1283m, 1256w, 1224m, 982w, 756w, 735w; 1H NMR (400 MHz, [D6]DMSO):
d ˆ 1.25 (s, 6H), 3.61 (d, J ˆ 17.5 Hz, 2H), 3.95 (s, 2H), 4.44 (d, J ˆ 17.5 Hz,
2H), 5.13 (d, J ˆ 12.4 Hz, 2H), 5.18 (d, J ˆ 12.4 Hz, 2H), 7.37 (s, 10H), 10.78
(s, 2H); 13C NMR (100 MHz, [D6]DMSO): d ˆ 17.91, 44.67, 47.13, 61.38,
66.26, 127.91 (2C), 128.09, 128.34, 135.50, 151.46, 168.02, 172.63; MS (pos.
X-ray crystal-structure data of 15: Colorless needles; C28H28N4O8, Mr ˆ
548.5; triclinic, space group P1: Dc ˆ 1.172 g cm 3, Z ˆ 1, a ˆ 7.611 (3),
b ˆ 7.980 (4), c ˆ 12.368(7) Š, a ˆ 73.56 (4), b ˆ 87.47 (4), g ˆ 67.45 (3), Vˆ
663.6 (6) Š3, MoKa (l ˆ 0.71073 Š) radiation, 2536 collected reflections,
Tˆ 293 K. The crystal structure was solved by direct methods (SHELXTL
PLUS) and refined by full-matrix least-squares analysis by using exper-
imental weights (heavy atoms anisotropic; H-atoms riding model, fixed
isotropic). Final (R(F) ˆ 0.042, wR(F 2) ˆ 0.1202) for 182 variable and 2345
observed reflections (F > 3.0 s(F)).
‡
FAB): m/z (%): 549 (10, [M ‡ H] ), 491 (10), 338 (28); C28H28N4O8
(548.56): calcd C 61.31, H 5.14, N 10.21; found: C 61.28, H 5.32, N 10.16.
X-ray crystal-structure data of 34: Colorless needles; C21H30N4O11, Mr ˆ
514.49; orthorhombic space group P2(1)2(1)2(1): Dc ˆ 1.510 g cm 3, Z ˆ 4,
a ˆ 9.787 (4), b ˆ 13.127 (4), c ˆ 17.620 (5) Š, a ˆ 90.00, b ˆ 90.00, g ˆ 90.00,
Vˆ 2263.7 (13) Š3, MoKa (l ˆ 0.71073 Š) radiation, 4576 collected reflec-
tions, Tˆ 293 K. The crystal structure was solved by direct methods
(SHELXTL PLUS) and refined by full-matrix least-squares analysis by
using experimental weights (heavy atoms anisotropic; H-atoms riding
model, fixed isotropic). Final (R(F) ˆ 0.0312, wR(F 2) ˆ 0.0756) for 446
variable and 4576 observed reflections (F > 3.0s(F)).
Dibenzyl-cis-[4a]-transoid-[4a,4b]-cis-[4b]-dodecahydro-11,12-dimethyl-
2,4,5,7-tetraoxocyclobuta[1,2-d:4,3-d']dipyrimidine-1,8-diacetate
(13):
750 mg (38%); Rf (CHCl3/MeOH 10:1) ˆ 0.58; m.p.: 204 ± 2068C; IR
(KBr): 3430m, 3178w, 3065w, 2856w, 1745s, 1700s, 1687s, 1478s, 1388w,
1378w, 1358m, 1284m, 1208s, 956w, 754w, 698w; 1H NMR (400 MHz,
[D6]DMSO): d ˆ 1.23 (s, 6H), 3.98 (s, 2H), 4.09 (d, J ˆ 17.7 Hz, 2H), 4.25 (d,
J ˆ 17.7 Hz, 2H), 5.11 (d, J ˆ 12.5 Hz, 2H), 5.15 (d, J ˆ 12.5 Hz, 2H), 7.36
(m, 10H), 10.65 (s, 2H); 13C NMR (100 MHz, [D6]DMSO): d ˆ 20.82,
45.66, 48.17, 63.32, 66.12, 127.90, 128.08, 128.35, 135.51, 151.31, 169.17,
‡
170.73; MS (pos. FAB): m/z (%): 549 (100, [M ‡ H] ), 338 (55);
Crystallographic data (excluding structure factors) for the structures
reported in this paper have been deposited with the Cambridge Crystallo-
graphic Data Centre as supplementary publication no. CCDC-119735 (12),
CCDC-119738 (13), CCDC-119734 (14), CCDC-119737 (15), CCDC-
119736 (34). Copies of the data can be obtained free of charge on
application to CCDC, 12 Union Road, Cambridge CB21EZ, UK (Fax:
(‡44)1223-336-033; E-mail: deposit@ccdc.cam.ac.uk).
C28H28N4O8 ‡ 0.5H2O (557.56): calcd C 60.32, H 5.24, N 10.05; found: C
60.59, H 5.30, N 10.34.
Dibenzyl-cis-[4a]-cisoid-[4a,4b]-cis-[4b]-dodecahydro-9,11-dimethyl-
2,4,6,8-tetraoxocyclobuta[1,2-d:3,4-d']dipyrimidine-1,5-diacetate
(14):
50 mg (2.5%); Rf (CHCl3/MeOH 10:1) ˆ 0.55; m.p.: 215 ± 2178C; IR
(KBr): 3433m, 3078w, 2967w, 1744m, 1710s, 1633w, 1475m, 1456w, 1389w,
1356w, 1300w, 1244m, 1194m, 961w, 756w, 700w; 1H NMR (400 MHz,
[D6]DMSO): d ˆ 1.43 (s, 6H), 3.74 (s, 2H), 3.84 (d, J ˆ 17.6 Hz, 2H), 4.48 (d,
J ˆ 17.6 Hz, 2H), 5.14 (d, J ˆ 12.4 Hz, 2H), 5.19 (d, J ˆ 12.4 Hz, 2H), 7.38
1,2,3,4-Tetrahydro-5-methyl-2,4-dioxopyrimidine-1-acetic acid (10):[47]
Thymine (10.00 g, 79.3 mmol) was suspended in H2O (150 mL). 50 mL of
an aqueous KOH-solution (3.6m) was added. The mixture was stirred at
Chem. Eur. J. 2000, 6, No. 1
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0601-0069 $ 17.50+.50/0
69

T. Carell et al.
FULL PAPER
(m, 10H), 10.57 (s, 2H); 13C NMR (100 MHz, [D6]DMSO): d ˆ 20.82, 47.56,
(t, J ˆ 5.8 Hz, 1H); 13C NMR (100 MHz, CDCl3): d ˆ 17.94, 20.08, 28.08
47.87, 61.49, 66.13, 127.89 (2C), 128.07, 128.34 (2C), 135.57, 151.01, 168.58,
(3C), 38.82, 42.23, 77.70, 114.48, 124.01, 125.35, 128.86, 143.81, 147.31,
‡
‡
170.71; MS (pos. FAB): m/z (%): 549 (100, [M ‡ H] ), 491 (8), 338 (45);
155.86; MS (EI): m/z (%): 309 (3, [M] ), 179 (100), 57 (47); C15H23N3O4
C28H28N4O8 (548.56): calcd C 61.31, H 5.14, N 10.21; found: C 61.03, H 5.24,
N 10.34.
(309.37): calcd C 58.24, H 7.49, N 13.58; found: C 58.35, H 7.52, N 13.52.
10-[2-({[(tert-Butyl)oxy]carbonyl}amino)ethyl]-7,8-dimethyl-[3H,10H]-
benzo[g]pteridine-2,4-dione (22): A suspension of 10% Pd/C catalyst
(100 mg) in acetic acid (10 mL) was slowly added to a solution of the nitro
compound 20 (3.0 g, 9.70 mmol) in methanol (150 mL). The suspension was
stirred for 7 h at room temperature in an H2 atmosphere. The reaction
mixture containing 21 was filtered through Celite. To the filtrate was added
alloxan monohydrate [2,4,5,6(1H,3H)-pyrimidinetetrone] (8 g, 50 mmol)
and boric acid (14 g, 230 mmol) and the reaction mixture was stirred for
12 h at room temperature. The reaction mixture was diluted with CHCl3
(500 mL) and extracted with water (3 Â 250 mL). The organic phase was
separated, dried with MgSO4, and concentrated in vacuo. The product 22
was obtained after column chromatography (silica gel-H, d ˆ 3 cm, L ˆ
15 cm, CHCl3/MeOH 15:1 !10:1 and recrystallization from ethanol/1,4-
dioxane as orange-colored needles (3.02 g, 80%). Rf (CHCl3/MeOH
10:1) ˆ 0.46; m.p.: 212 ± 2148C; IR (CHCl3): 3456w, 3378w, 3007w, 1712m,
1683m, 1581m, 1548s, 1506m, 1458w, 1394w, 1368w, 1344w, 1294w, 1272w,
1164m, 867w; 1H NMR (400 MHz, [D6]DMSO): d ˆ 1.23 (s, 9H), 2.40 (s,
3H), 2.49 (s, 3H), 3.41 (q, J ˆ 6.0 Hz, 2H), 4.64 (t, J ˆ 6.0 Hz, 2H), 6.96 (t,
J ˆ 6.0 Hz, 1H), 7.83 (s, 1H), 7.89 (s, 1H), 8.31 (s, 1H); 13C NMR (100 MHz,
[D6]DMSO): d ˆ 18.67, 20.75, 27.94 (3C), 36.82, 44.13, 79.13, 116.17, 130.91,
131.44, 133.85, 135.62, 136.82, 146.26, 150.35, 155.46, 155.76, 159.86; MS
Dibenzyl-cis-[4a]-cisoid-[4a,4b]-cis-[4b]-dodecahydro-11,12-dimethyl-
2,4,5,7-tetraoxocyclobuta[1,2-d:4,3-d']dipyrimidine-1,8-diacetate
(12):
40 mg (2%); Rf (CHCl3/MeOH 10:1) ˆ 0.50; m.p.: 146 ± 1488C; IR
(KBr): 3411w, 3217w, 3078w, 2956w, 1744s, 1709s, 1475m, 1391w, 1361w,
1287m, 1194m, 965w, 754w, 698w; 1H NMR (400 MHz, [D6]DMSO): d ˆ
1.31 (s, 6H), 3.92 (d, J ˆ 17.4 Hz, 2H), 3.99 (s, 2H), 4.29 (d, J ˆ 17.4 Hz, 2H),
5.12 (d, J ˆ 12.5 Hz, 2H), 5.18 (d, J ˆ 12.5 Hz, 2H), 7.36 (s, 10H), 10.53 (s,
2H); 13C NMR (100 MHz, [D6]DMSO): d ˆ 18.18, 46.09, 47.15, 59.34, 66.07,
127.82 (2C), 128.05, 128.33 (2C), 135.56, 152.17, 168.54, 170.34; MS (pos.
‡
FAB): m/z (%): 549 (100, [M ‡ H] ), 338 (18); C28H28N4O8 (548.56): calcd
C 61.31, H 5.14, N 10.21; found: C 61.37, H 5.28, N 10.15.
cis-[4a]-transoid-[4a,4b]-cis-[4b]-Dodecahydro-2,4,5,7-tetraoxo-cyclobu-
ta[1,2-d:4,3-d']dipyrimidine-1,8-diacetic acid (17): The diester 13 (300 mg,
0.547 mmol) was dissolved in acetic acid (20 mL). After the addition of
10% Pd/C catalyst (10 mg) the suspension was stirred in an H2 atmosphere
for 2 h. The reaction mixture was filtered through Celite. The Celite pad
was washed with hot acetic acid twice. The combined filtrates were
concentrated in vacuo. The product precipitated upon addition of diethyl
ether to the residual oily material and was filtered off, washed with acetone
and dried in vacuo to yield a colorless powder (185 mg, 92%). M.p.: 243 ±
2458C; IR (KBr): 3600 ± 2800m, 3400w, 3231w, 3067w, 2867w, 1761m, 1693s,
1483s, 1385m, 1302m, 1239m, 1160m, 883w, 822w, 759w; 1H NMR
(400 MHz, [D6]DMSO): d ˆ 1.27 (s, 6H), 3.89 (s, 2H), 3.97 (d, J ˆ
17.6 Hz, 2H), 4.04 (d, J ˆ 17.6 Hz, 2H), 10.55 (s, 2H), 11 ± 13 (brs, 2H);
13C NMR (100 MHz, [D6]DMSO): d ˆ 21.02, 45.60, 48.40, 63.69, 151.24,
‡
(EI): m/z (%): 385 (1, [M] ), 312 (10), 256 (9), 243 (100), 84 (28), 49 (31);
C19H23N5O4 (385.42): calcd C 59.21, H 6.01, N 18.17; found: C 59.35, H 6.12,
N 18.29.
10-[2-({[(tert-Butyl)oxy]carbonyl}amino)ethyl]-7,8-dimethyl-3-ethyl-
[3H,10H]-benzo[g]pteridine-2,4-dione (23): Ethyl iodide (1 mL,
12.37 mmol) were dropwise added to a suspension of the flavin derivative
22 (500 mg, 1.3 mmol) and Cs2CO3 (0.6 g, 1.83 mmol) in dry (4 Š mol
sieves) DMF (50 mL). The reaction mixture was stirred at room temper-
ature for 3 h. The mixture was diluted with CHCl3 (200 mL) and extracted
with water (3 Â 80 mL). The organic phase was separated, dried with
MgSO4, filtered, and evaporated in vacuo. The product 23 was obtained
after flash chromatography on silica gel-H (d ˆ 3 cm, L ˆ 18 cm, CHCl3/
MeOH 20:1) and recrystallization from MeOH/H2O as yellow needles
(380 mg, 71%). Rf (CHCl3/MeOH 20:1) ˆ 0.41; m.p.: 227 ± 2298C; IR
(CHCl3): 3398w, 2977w, 2933w, 1706m, 1693m, 1644s, 1583s, 1545m, 1447w,
1433w, 1337m, 1311w, 1294w, 1272w, 1250m, 1200w, 1176m, 1017w, 994w,
‡
170.63, 170.76; MS (pos. FAB): m/z (%): 369 (20, [M ‡ H] ), 338 (100), 273
(24), 165 (30); C14H16N4O8‡0.5CH3COOH (398.33): calcd C 45.23, H 4.55,
N 14.07; found: C 44.94, H 4.87, N 14.05.
cis-[4a]-cisoid-[4a,4b]-cis-[4b]-dodecahydro-2,4,5,7-tetraoxo-cyclobuta-
[1,2-d:4,3-d']dipyrimidine-1,8-diacetic acid (16): The diester 12 (250 mg,
0.456 mmol) was dissolved in acetic acid (20 mL). After the addition of
10% Pd/C catalyst (10 mg) the solution was stirred in an H2 atmosphere for
2 h. The reaction mixture was filtered through Celite and the Celite pad was
washed with hot acetic acid. The combined acetic acid solutions were
concentrated in vacuo. Addition of diethylether to the oily residual
material yielded 16 as a colorless powder. The product was filtered off,
washed with acetone and dried in vacuo (168 mg, 100%). M.p.: 229 ±
2318C; IR (KBr): 3600 ± 2800m, 3433w, 3222w, 3067w, 1703s, 1486s,
1396m, 1290m, 1219m, 1194m, 800w, 772w, 756w; 1H NMR (400 MHz,
[D6]DMSO): d ˆ 1.34 (s, 6H), 3.65 (d, J ˆ 17.3 Hz, 2H), 3.94 (s, 2H), 4.16 (d,
J ˆ 17.3 Hz, 2H), 10.41 (s, 2H), 12.5 ± 13.5 (brs, 2H); 13C NMR (100 MHz,
[D6]DMSO): d ˆ 18.23, 46.19, 47.07, 59.22, 152.12, 170.00, 170.51; MS (pos.
1
933w, 889w, 806q, 772w; H NMR (400 MHz, [D6]DMSO): d ˆ 1.16 (t, J ˆ
7.0 Hz, 3H), 1.22 (s, 9H), 2.40 (s, 3H), 2.50 (s, 3H), 3.41 (q, J ˆ 5.7 Hz, 2H),
3.93 (t, J ˆ 7.0 Hz, 2H), 4.66 (t, J ˆ 5.7 Hz, 2H), 6.98 (t, J ˆ 5.7 Hz, 1H), 7.86
(s, 1H), 7.93 (s, 1H); 13C NMR (100 MHz, [D6]DMSO): d ˆ 12.83, 18.63,
20.71, 27.86 (3C), 35.82, 36.90, 43.95, 77.75, 116.10, 130.87, 131.33, 134.10,
135.69, 135.86, 146.40, 148.87, 154.40, 155.69, 159.09; MS (FAB): m/z (%):
‡
414 (100, [M ‡ 1] ), 358 (6), 340 (8), 271 (36); C21H27N5O4 ‡ 2ꢁ3H2O
‡
FAB): m/z (%): 369 (64, [M ‡ H] ), 338 (100).
(425.49): calcd C 59.28, H 6.71, N 16.46; found: C 59.21, H 6.65, N 16.29.
N-tert-Butyloxycarbonyl-ethylene diamine (19):[28] A solution of Boc2O
(30 g, 0.14 mol) in 1,4-dioxane (100 mL) was added dropwise within 2 h at
room temperature to a solution of ethylene diamine (25.0 g, 0.42 mol) in
1,4-dioxane (150 mL). The reaction mixture was stirred for another 20 h at
room temperature. The reaction mixture was filtered and the filtrate was
concentrated in vacuo. The residual oil was distilled and the product 19 was
isolated at 1058C (20 Torr) and obtained as a colorless oil (13 g, 62%) and
used without further purification. 1H NMR (300 MHz, CDCl3): d ˆ 1.19 (s,
2H), 1.38 (s, 9H), 2.72 (t, J ˆ 6.0 Hz, 2H), 3.10 (m, 2H), 5.20 (brs, 1H).
10-(2-Aminoethyl)-7,8-dimethyl-3-ethyl-[3H,10H]-benzo[g]pteridine-2,4-
dione, TFA salt (8): A solution of the Boc-protected flavin derivative 23
(437 mg, 1.03 mmol) in TFA/water (10 mL 95:5) was stirred for 3 h at room
temperature. The reaction solution was evaporated in vacuo and the
product was precipitated through addition of diethyl ether. Compound 8
was obtained as the TFA salt in form of a yellow powder (427 mg, 97%). IR
(CHCl3): 3444m, 1713s, 1615s, 1580s, 1542s, 1458m, 1347m, 1250s, 1193m,
1022w, 929w, 878w, 806w, 773w; 1H NMR (300 MHz, [D6]DMSO): d ˆ 1.16
(t, J ˆ 6.9 Hz, 3H), 2.42 (s, 3H), 2.50 (s, 3H), 3.20 (m, 2H), 3.94 (q, J ˆ
6.9 Hz, 2H), 4.91 (t, J ˆ 6.0 Hz, 2H), 7.98 (s, 1H), 8.05 (s, 1H), 8.17 (m, 3H);
13C NMR (75 MHz, [D6]DMSO): d ˆ 12.82, 18.63, 20.41, 35.93, 35.99, 41.11,
116.05, 130.68, 131.28, 134.30, 136.20, 136.46, 147.31, 149.63, 154.68, 159.25;
N-[2-({[(tert-Butyl)oxy]carbonyl}amino)ethyl]-4,5-dimethyl-2-nitro-ani-
line (20): The mono-Boc protected ethylene diamine (19) (6.3 g,
39.3 mmol) was dissolved in pyridine (15 mL). This solution was slowly
‡
MS (FAB): m/z (%): 314 (100, [M CF3COO] ), 271 (15), 232 (7).
added to
a solution of 1,2-dinitro-4,5-dimethylbenzene (18) (5 g,
25.5 mmol) in pyridine (50 mL). The reaction mixture was stirred for 24 h
at 908C and subsequently concentrated in vacuo. The residual material was
recrystallized from ethanol. This yielded 20 as intensively red-colored
needles (5.3 g, 67%). Rf (toluene/EtOAc 10:1) ˆ 0.33; m.p.: 120 ± 1228C;
IR (KBr): 3378m, 3356m, 2978w, 2933w, 1683s, 1639m, 1572m, 1528s,
1511m, 1461w, 1400w, 1367w, 1344w, 1311w, 1294w, 1289w, 1250w, 1222m,
1194s, 1172m, 1150s, 1033w, 1000w, 861w; 1H NMR (400 MHz, CDCl3): d ˆ
1.37 (s, 9H), 2.14 (s, 3H), 2.25 (s, 3H), 3.19 (q, J ˆ 6.0 Hz, 2H), 3.38 (m, J ˆ
5.8 Hz, 2H), 6.93 (s, 1H), 7.01 (t, J ˆ 5.3 Hz, 1H), 7.82 (s, 1H), 8.09
Model compound 2: A solution of the cis-syn-thymine dimer dicarboxylic
acid 16 (55 mg, 0.15 mmol) and an excess of BOP (300 mg, 0.646 mmol)
were dissolved in DMF (4 mL) and stirred at room temperature for 10 min.
After the addition of a solution of 8 ´ TFA (50 mg, 0.17 mmol) in DMF
(2 mL) and of ten drops of triethylamine, the reaction was stirred for 2 h at
room temperature. Then pentyl alcohol 9 (40 mg, 0.38 mmol) was added to
the reaction mixture, and stirring was continued for another 4 h at room
temperature. The reaction mixture was diluted with water (100 mL) and
extracted with CHCl3 (3 Â 200 mL). The combined organic layers were
70
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0601-0070 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 1

Thymine- and Uracil-Photodimers
62±72
separated, dried with MgSO4, filtrated, and concentrated in vacuo. The
crude reaction product was dissolved in a small amount of methanol and
precipitated through addition of diethyl ether. The product was purified by
silica gel column chromatography (silica gel-H, d ˆ 3 cm, L ˆ 20 cm,
CHCl3/MeOH 7.5:1 !5:1. 2 was obtained as a yellow powder (38 mg,
35%). Rf (CHCl3/MeOH 10:1): 0.22; m.p. : 240 ± 2428C; IR (KBr): 3430m,
3211m, 3078w, 2933w, 2856w, 1704s, 1650s, 1584m, 1548s, 1465m, 1389w,
DNA-photolyase (25nm or 50nm) were prepared and filled into a quartz
cuvette equipped with a magnetic stirrer. The enzyme was added in the
dark. The assay solutions were mixed and irradiated at 435 nm (A.
nidulans). For each time point 20 mL were removed and 0.1m HOAc
(10 mL) was added in order to keep the enzyme reaction stopped. The
samples were analyzed by reversed-phase HPLC or ion-exchange chro-
matography. Reversed phase: Nucleosil 120-3, C18 (250 Â 4 mm),
1
1
1282m, 1228m, 1017w, 928w, 806w; H NMR (500 MHz, [D6]DMSO): d ˆ
0.7 mLmin
,
A: 20mm NH4OAc, B: 50% MeCN, 50%
A
(v/v),
0.86 (t, J ˆ 7.1 Hz, 3H), 1.15 (t, J ˆ 7.0 Hz, 3H), 1.26 (m, 4H), 1.31 (s, 3H),
1.36 (s, 3H), 1.37 (m, 2H), 2.42 (s, 3H), 2.52 (s, 3H), 3.05 (m, 2H), 3.34 (m,
1H), 3.49 (m, 2H), 3.79 (d, J ˆ 5.2 Hz, 1H), 3.87 (d, J ˆ 5.2 Hz, 1H), 3. 91
(m, 1H), 3.93 (q, J ˆ 7.1 Hz, 2H), 4.15 (d, J ˆ 16.5 Hz, 1H), 4.20 (d, J ˆ
16.5 Hz, 1H), 4.61 (m, 1H), 4.73 (m, 1H), 7.92 (s, 1H), 7.93 (t, J ˆ 5.6 Hz,
1H), 7.97 (s, 1H), 8.22 (t, J ˆ 6.0 Hz, 1H), 10.36 (s, 1H), 10.39 (s, 1H);
13C NMR (125 MHz, [D6]DMSO): d ˆ 12.82, 13.80, 18.06, 18.21, 18.66,
20.70, 21.72, 28.46, 28.66, 35.76, 35.88, 38.44, 42.91, 46.34, 46.57, 47.77, 47.84,
59.30, 59.61, 115.99, 130.87, 131.03, 134.09, 135.87, 136.15, 146.70, 149.05,
151.96, 151.97, 154.61, 159.07, 166.96, 168.32, 170.47, 170.53; MS (pos. FAB):
0 !20 min: 6 !16% B. Retention times: 25: 7.7 min, 26: 6.8 min, 27:
13.3 min, 28: 12.2 min, 29: 10.8 min, 30: 9.8 min, 31: 17.6 min, 32: 14.6 min.
Ion-exchange chromatography: Nucleogel SAX 1000-8/77, 2 mLmin 1, A:
0.2m NaCl, 10mm Na2HPO4, pH 11.4, B: 1m NaCl, 10mm Na2HPO4,
pH 11.4, 0 !6 min: 100% A, 6 !23 min: 0 !51% B. Retention times: 25:
14.8 min, 26: 15.0 min, 27: 18.0 min, 28: 18.0 min, 29: 17.1 min, 30: 17.2 min,
31: 19.6 min, 32: 19.5 min
‡
m/z (%): 733 (100, [M ‡ H] ), 507 (13); HR-MS (pos. FAB) C35H44N10O8:
Acknowledgment
‡
calcd for [M ‡ H] : 733.3422; found: 733.3426.
Model compound 24: Compound 24 was prepared similar to compound 2.
As the dimer carboxylic acid the trans-syn-compound 17 (55 mg,
0.15 mmol) was used. The crude product 24 was dissolved in a small
amount of methanol and precipitated through addition of diethyl ether.
Compound 24 was obtained after column chromatography on silica gel
(silica gel-H, d ˆ 3 cm, L ˆ 20 cm, CHCl3/MeOH 7:1) as a yellow powder
(35 mg, 32%). Rf (CHCl3/MeOH 10:1) ˆ 0.28; m.p.: 200 ± 2028C; IR (KBr):
3430m, 3200m, 3078w, 2956w, 2922w, 2856w, 1700s, 1650s, 1578m, 1548s,
1473m, 1406w, 1372w, 1355m, 1289m, 1224m, 1150w, 1094w, 933w; 1H NMR
(500 MHz, [D6]DMSO): d ˆ 0.86 (t, J ˆ 7.1 Hz, 3H), 1.16 (t, J ˆ 7.1 Hz, 3H),
1.20 (s, 3H), 1.23 (m, 4H), 1.27 (s, 3H), 1.38 (m, 2H), 2.41 (s, 3H), 2.51 (s,
3H), 3.02 (m, 2H), 3.48 (q, J ˆ 6.9 Hz, 2H), 3.65 (d, J ˆ 8.2 Hz, 1H), 3.73 (d,
J ˆ 16.6 Hz, 1H), 3.85 (d, J ˆ 16.6 Hz, 1H), 3.87 (d, J ˆ 8.2 Hz, 1H), 3.93 (q,
J ˆ 7.1 Hz, 2H), 3.97 (d, J ˆ 16.5 Hz, 1H), 4.00 (d, J ˆ 16.5 Hz, 1H), 4.63 (t,
J ˆ 6.9 Hz, 2H), 7.87 (t, J ˆ 5.8 Hz, 1H), 7.92 (s, 1H), 7.93 (s, 1H), 8.25 (t,
J ˆ 6.0 Hz, 1H), 10.46 (s, 1H), 10.50 (s, 1H); 13C NMR (125 MHz,
[D6]DMSO): d ˆ 12.82, 13.80, 18.71, 20.70, 20.83, 21.16, 21.73, 28.47, 28.59,
35.45, 35.83, 38.52, 42.93, 45.38, 45.40, 45.59, 49.13, 63.77, 63.87, 115.98,
130.86, 131.03, 134.03, 135.93, 136.05, 146.69, 148.76, 151.22, 151.44, 154.52,
159.07, 167.63, 168.98, 170.73, 170.79; MS (pos. FAB): m/z (%): 733 (26,
This work was supported by the Schweizer Nationalfond, by the ETH
Zürich and the Boehringer Ingelheim Fonds (Doctoral Scholarship to J.B.).
We thank Prof. F. Diederich for the continuing generous support and
Prof. U. Wild for enabling us to perform the GAMESS(US) calculations
(E.-U.W.). We thank M. Leduc for a careful inspection of the manuscript.
[1] P. F. Heelis, R. F. Hartman, S. D. Rose, Chem. Soc. Rev. 1995, 289 ± 297.
[2] S.-T. Kim, A. Sancar, Photochem. Photobiol. 1993, 57, 895 ± 904.
[3] T. Carell, Angew. Chem. 1995, 107, 2697 ± 2700; Angew. Chem. Int. Ed.
Engl. 1995, 34, 2491 ± 2494.
[4] J.-S. Taylor, Acc. Chem. Res. 1994, 27, 76 ± 82.
[5] J.-S. Taylor, J. Chem. Educ. 1990, 67, 835 ± 841.
[6] T. Carell, Chimia 1995, 49, 365 ± 373.
[7] A. Sancar, Biochemistry 1994, 33, 2 ± 9.
[8] E. C. Friedberg, G. C. Walker, W. Siede, DNA repair and mutagenesis,
ASM Press, Washington, D.C. 1995.
[9] M. Prather, P. Midgley, F. S. Rowland, R. Stolarski, Nature 1996, 381,
551 ± 554.
[10] H. Slaper, G. J. M. Velders, J. S. Daniel, F. R. deGruijl, J. C. van der -
Leun, Nature 1996, 384, 256 ± 258.
[11] R. Epple, E.-U. Wallenborn, T. Carell, J. Am. Chem. Soc. 1997, 119,
7440 ± 7451.
‡
[M ‡ H] ), 507 (100), 480 (10), 338 (9); HR-MS (pos. FAB) C35H44N10O8:
‡
calcd for [M ‡ H] : 733.3422; found: 733.3423.
Oligonucleotide synthesis: Oligonucleotide synthesis was performed on a
Pharmacia-Gene-Plus synthesizer connected to a Olivetti-M 300 personal
computer. Synthesis of the oligonucleotides 25 ± 32 was performed by using
a modified 1.3 mmol cycle and Pac-amidites. Solvents and solutions were
made according to the manufacturers protocol. The phosphoramidite (0.1m
in MeCN) and 1H-tetrazole (0.5m in MeCN) solution were equal in
concentrations to those used for the synthesis of natural oligodeoxynu-
cleotides. Average coupling yields monitored by on-line trityl assay were
generally in the range of 95 ± 99%. All syntheses were run in the trityl-on
mode, resulting DMTr-25 ± DMTr-32. For the deprotection and cleavage of
DMTr-26 and DMTR-28 from the support, the solid support was
suspended in sat. NH3/MeOH (1 mL, MeOH dried over 3 Š molecular
sieves) and kept at 258C with occasional shaking for one day. For the
deprotection and purification of all other oligonucleotides, the solid
support was suspended in conc. aqueous NH3/MeOH ˆ 3:1 (2 mL) and left
for 5 h at 558C to effect deprotection and cleavage. After addition of 1m
NEt3/HOAc (2 mL) to each solution, the support was filtered off and
washed with water (3 Â 0.5 mL). The solution was concentrated to 1 mL
and the oligonucleotides were purified by HPLC (C18-RP, A: 0.1m NEt3/
HOAc, B: 80% MeCN, 25 !40% B in 40 min). After removal of the
solvent in vacuo, all oligonucleotides were detritylated by addition of 50%
HCOOH/H2O (3 mL). After 5 min the formic acid was removed in vacuo
and 25 ± 32 were purified by HPLC (C18-RP, 5 !10% B in 5 min, 10 !25%
B in 30 min). Laser-desorption mass spectra (negative ions detected): calcd
for 25: 3595.5, found: 3596.1; calcd for 26: 3567.4, found: 3567.9; calcd for
27: 4168.8, found: 4168.3; calcd for 28: 4140.8, found: 4140.0; calcd for 29:
3595.5, found: 3596.4; calcd for 31: 4168.8, found: 4170.5.
[12] I. D. Podmore, P. F. Heelis, M. C. R. Symons, A. Pezeshk, Chem.
Commun. 1994, 1005 ± 1006.
[13] S.-T. Kim, A. Sancar, Biochemistry 1991, 30, 8623 ± 8630.
[14] D. L. Svoboda, C. A. Smith, J.-S. Taylor, A. Sancar, J. Biol. Chem.
1993, 268, 10684 ± 10700.
[15] A. A. Voityuk, N. Rösch, J. Phys. Chem. A 1997, 101, 8335 ± 8338.
[16] A. A. Voityuk, M. E. Michel-Beyerle, N. Rösch, J. Am. Chem. Soc.
1996, 118, 9750 ± 9758.
[17] J. Hahn, M.-E. Michel-Beyerle, N. Rösch, J. Phys. Chem. B 1999, 103,
2001 ± 2007.
[18] T. Langenbacher, X. Zhao, G. Bieser, P. F. Heelis, A. Sancar, M. E.
Michel-Beyerle, J. Am. Chem. Soc. 1997, 119, 10532 ± 10536. This
study was performed with the E. coli enyzme and dinucleotide
cyclobutane substrates. For calculations see: D. B. Sander, O. Wrest, J.
Am. Chem. Soc. 1999, 121, 5127 ± 5134.
[19] A. P. M. Eker, P. Kooiman, J. K. C. Hessels, A. Yasui, J. Biol. Chem.
1990, 265, 8009 ± 8015.
[20] J. Butenandt, L. T. Burgdorf, T. Carell, Angew. Chem. 1999, 111, 718 ±
721; Angew. Chem. Int. Ed. 1999, 38, 708 ± 711.
[21] J. Butenandt, A. P. M. Eker, T. Carell, Chem. Eur. J. 1998, 4, 642 ± 653.
The thymidine dimer was prepared as described for the uridine dimer.
All details and the analytical data will be reported in due course.
[22] T. Carell, R. Epple, V. Gramlich, Helv. Chim. Acta 1997, 80, 2191 ±
2203.
[23] A. G. Cochran, R. Sugasawara, P. G. Schultz, J. Am. Chem. Soc. 1988,
117, 5453 ± 5461.
Enzymatic studies: 500 mL solutions containing the damaged oligonucleo-
tide (1 ± 5 mmol), 0.1m NaCl, 10mm KH2PO4 pH 7.0, and 5mm DTT and
[24] J. R. Jacobsen, A. G. Cochran, J. C. Stephans, D. S. King, P. G. Schultz,
J. Am. Chem. Soc. 1995, 118, 5453 ± 5461.
Chem. Eur. J. 2000, 6, No. 1
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0601-0071 $ 17.50+.50/0
71

T. Carell et al.
FULL PAPER
[25] R. Kuhn, W. v. Klaveren, Ber. Dtsch. Chem. Ges. 1938, 71, 779 ± 780.
[26] T. Carell, J. Butenandt, Angew. Chem. 1997, 109, 1590 ± 1593; Angew.
Chem. Int. Ed. Engl. 1997, 36 1461 ± 1464.
[27] T. Carell, H. Schmid, M. Reinhard, J. Org. Chem. 1998, 63, 8741 ± 8747.
[28] C. Petry, Dissertation, Ruprecht-Karls-Universität Heidelberg, 1993,
p. 186.
[39] R. F. Hartman, J. R. van Camp, S. D. Rose, J. Org. Chem. 1987, 52,
2684 ± 2689.
[40] S.-T. Kim, S. D. Rose, J. Phys. Org. Chem. 1990, 3, 581 ± 586.
[41] I. Mayer, Int. J. Quant. Chem. 1986, 29, 477 ± 483.
[42] F. E. Hruska, L. Voituriez, A. Grand, J. Cadet, Biopolymers 1986, 25,
1399 ± 1417.
[29] R. Kuhn, F. Weygand, Ber. Dtsch. Chem. Ges. 1934, 67, 1409 ± 1413.
[30] R. Kuhn, F. Weygand, Ber. Dtsch. Chem. Ges. 1935, 68, 1282 ± 1288.
[31] B. Castro, G. Evin, C. Delve, R. Seyer, Synthesis 1977, 413 ± 469.
[32] C. G. Hatchard, C. A. Parker, Proc. R. Soc. A. 1956, 235, 518 ± 536.
[33] H. G. O. Becker, Einführung in die Photochemie, Deutscher Verlag
der Wissenschaften, Berlin, 1991.
[34] C. Reichardt, Solvents and Solvent Effects in Organic Chemistry,
VCH, Weinheim, 1988.
[35] R. F. Hartman, S. D. Rose, J. Am. Chem. Soc. 1992, 114, 3559 ± 3560.
[36] I. Husain, A. Sancar, Nucleic Acids Res. 1987, 15, 1109 ± 1120.
[37] H.-W. Park, S.-T. Kim, A. Sancar, J. Deisenhofer, Science 1995, 268,
1866 ± 1872.
[43] K. McAteer, Y. Jing, J. Kao, J.-S. Taylor, M. A. Kennedy, J. Mol. Biol.
1998, 282, 1013 ± 1032.
[44] B. J. V. Berg, G. B. Sancar, J. Biol. Chem. 1998, 273, 20276 ± 20284.
[45] R. J. Roberts, Cell 1995, 82, 9 ± 12.
[46] G. L. Verdine, S. D. Bruner, Chem. Biol. 1997, 4, 329 ± 334.
[47] A. S. Jones, P. Lewis, S. F. Withers, Tetrahedron 1973, 29, 2293.
[48] M. W. Schmidt, K. K. Baldridge, J. A. Boatz, S. T. Elbert, M. S.
Gordon, J. H. Jensen, S. Koseki, N. Matsunaga, K. A. Nguyen, S. J. Su,
T. L. Windus, M. Dupuis, J. A. Montgomery, J. Comput. Chem. 1994,
14, 1347 ± 1363.
[38] T. Tamada, K. Kitadokoro, Y. Higuchi, K. Inaka, A. Yasui, P. E.
de Ruiter, A. P. M. Eker, K. Miki, Nature Struct. Biol. 1997, 11, 887 ±
891.
Received: April 26, 1999 [F 1745]
72
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0601-0072 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 1