Crystal structures and repair studies reveal the identity and the base-pairing properties of the UV-induced spore photoproduct DNA lesion

Article abstract of DOI:10.1002/chem.201100177

UV light is one of the major causes of DNA damage. In spore DNA, due to an unusual packing of the genetic material, a special spore photoproduct lesion (SP lesion) is formed, which is repaired by the enzyme spore photoproduct lyase (Spl), a radical S-adenosylmethionine (SAM) enzyme. We report here the synthesis and DNA incorporation of a DNA SP lesion analogue lacking the phosphodiester backbone. The oligonucleotides were used for repair studies and they were cocrystallized with a polymerase enzyme as a template to clarify the configuration of the SP lesion and to provide information about the base-pairing properties of the lesion. The structural analysis together with repair studies allowed us to clarify the identity of the preferentially repaired lesion diastereoisomer.

Full text of DOI:10.1002/chem.201100177

FULL PAPER
DOI: 10.1002/chem.201100177
Crystal Structures and Repair Studies Reveal the Identity and the Base-
Pairing Properties of the UV-Induced Spore Photoproduct DNA Lesion
Korbinian Heil,[a] Andrea Christa Kneuttinger,[a] Sabine Schneider,[b]
Ulrike Lischke,[a] and Thomas Carell*[a]
In memory of Athel Beckwith
Abstract: UV light is one of the major
causes of DNA damage. In spore
DNA, due to an unusual packing of
the genetic material, a special spore
photoproduct lesion (SP lesion) is
formed, which is repaired by the
enzyme spore photoproduct lyase
(Spl), a radical S-adenosylmethionine
(SAM) enzyme. We report here the
synthesis and DNA incorporation of a
DNA SP lesion analogue lacking the
phosphodiester backbone. The oligonu-
cleotides were used for repair studies
and they were cocrystallized with a
polymerase enzyme as a template to
clarify the configuration of the SP
lesion and to provide information
about the base-pairing properties of
the lesion. The structural analysis to-
gether with repair studies allowed us to
clarify the identity of the preferentially
repaired lesion diastereoisomer.
Keywords: DNA damage
·
en-
zymes · oligonucleotides · spore
photoproduct · UV irradiation
Introduction
product lyases (Spl) form, next to CPD photolyases and (6–
4) photolyases, a third enzyme family, which is able to di-
rectly repair mutagenic UV lesions. Repair by the Spl is a
radical process, which is initiated by a protein-bound iron–
sulfur cluster and which requires the cofactor SAM.[5] The
mechanism that leads to the formation of the spore lesion is
still not completely understood. It is believed that triplet
states are involved and it is clear that the photosensitizer di-
picolinic acid and dry conditions favor formation of the
lesion.[6] So far, formation of the lesion was only observed in
thymidylyl–(3’–5’)–thymidine (TpT) sequences, in which, by
an unknown process, one of the methyl groups present at
one thymine reacts at the C5 position of the second thymine
base (Scheme 1).[7] This reaction creates a new stereocenter
Genetic material is strongly protected in bacterial spores,
which were designed by nature to store DNA material over
long periods of time.[1] Spores can protect the genetic infor-
mation for over thousands of years, which make them the
most sophisticated and durable information storage systems
on earth. DNA packed inside a spore is also more resistant
towards UV light, which is otherwise an agent that constant-
ly damages DNA on earth in light-exposed habitats.[2] De-
spite this unusual UV-light resistance, it is known that spore
DNA is also at some point damaged by UV light. In con-
trast to normal DNA, however, where UV light causes the
formation of cyclobutane pyrimidine dimers (CPDs) and
pyrimidineACHTUNGTRENNUNG(6–4)pyrimidones ((6–4) lesions), in spores an un-
usual lesion accumulates, which is called the spore photo-
product lesion.[3] This lesion hinders the ability of spores to
restore gene function upon germination. The spore lesions
are repaired by special, oxygen-sensitive, radical S-adenosyl-
methionine (SAM) enzymes, which are able to locate the le-
sions and to perform a direct repair.[4] These spore photo-
Scheme 1. Depiction of the two possible spore photoproducts with 5S
and 5R configuration.
[a] Dipl.-Chem. K. Heil, A. C. Kneuttinger, Dipl.-Biol. U. Lischke,
Prof. Dr. T. Carell
Center for Integrated Protein Science (CiPSM) at the
Department of Chemistry, Ludwig-Maximilians-University
Butenandtstrasse 5–13, 81377 Munich (Germany)
E-mail: thomas.carell@cup.uni-muenchen.de
at C5. In addition, the new bond can form between the C5
carbon atom of the 5’dT base and the 3’dT-methyl group
(5’!3’) or vice versa. This leads in principle to four different
spore lesions. Due to the geometrical constraints imposed
by the B-duplex structure, reaction in the 5’!3’ direction,
however, must give the 5R-configured lesion, whereas reac-
tion in the 3’!5’ direction should provide the 5S-configured
compound (Scheme 1) making these two diastereoisomers
[b] Dr. S. Schneider
Department of Chemistry
Technical University of Munich
Lichtenbergstrasse 4, 85747 Garching, Munich (Germany)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201100177.
Chem. Eur. J. 2011, 17, 9651 – 9657
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
9651

the only possible spore photoproduct (SP) lesions formed in
the duplex.
Recently, we reported the synthesis of 3’!5’ lesion ana-
logues with 5S and 5R configurations as phosphoramidite
building blocks and showed the preparation of DNA du-
plexes containing both compounds.[8] The 5S isomer was
indeed found to create the more stable duplex proving that
the 3’!5’ reaction would form preferentially the 5S-config-
ured lesion. Initial studies with a dinucleotide building block
provided evidence that the Spl is able to repair the 5S com-
pound.[9] Recent seminal studies by the groups of Broderick
and Bardet showed, however, that the SP lesion forms ex-
clusively in the 3’!5’ direction with the 5R configuration
and that the 5R compound is the preferred substrate for the
repair enzyme.[10] We report here the development of a syn-
thesis of the 5S- and 5R-configured 5’!3’ reaction products
as phosphoramidite building blocks, their incorporation into
oligonucleotides, and repair studies, now with full oligonu-
cleotides. Our prepared lesions are again analogues that
lack the central phosphodiester group to abbreviate the long
and challenging synthesis of the natural spore lesion, as well
as to ease the analysis of the repair reaction, which gives
with our analogues an easy-to-detect strand break. Finally,
we report crystal structures of DNA duplexes containing the
5S and 5R 5’!3’ reaction products using a DNA polymerase
as a crystallization template. The structure provides insight
into the base-pairing properties of both compounds, most
importantly for the biologically relevant 5R-configured 5’!
3’ molecule.
Results and Discussion
Scheme 2. Synthesis of the phosphoramidite building blocks 7 and 8.
a) tert-Butylchlorodiphenylsilane (TBDPSCl), 4-dimethylaminopyridine
(DMAP), 75%; b) tert-butyldimethylsilyl chloride (TBSCl), imidazole,
99%; c) 2-(trimethylsilyl)ethoxymethyl chloride (SEMCl), N,N-diisopro-
pylethylamine (DIEA), 61%; d) N-bromosuccinimide (NBS), azobisiso-
butyronitrile (AIBN), 50%; e) H2, Rh/Al2O3, 98%; f) TBSCl, DMAP,
79%; g) TBDPSCl, imidazole, 99%; h) SEMCl, DIEA, 80%; i) lithium
diisopropylamide (LDA) then 2, ꢀ788C to RT; j) toluene sulfonic acid
(pTsOH), 508C; k) normal-phase (np) HPLC separation of diastereoiso-
mers (ds) 5 (8%) and 6 (5%) over 3 steps; l) 4,4’-dimethoxytriphenyl-
methyl chloride (DMTCl), DMAP, 66%; m) 2-cyanoethyl N,N-diisopro-
pylchlorophosphoramidite (CEDCl), DIEA, 66%; n) DMTCl, DMAP,
64%; o) CEDCl, DIEA, 44%.
The synthesis of the two phosphoramidites is depicted in
Scheme 2. The starting point of the synthesis is thymidine 1,
which was firstly TBDPS-protected at the primary OH
group, secondly TBS-protected at the secondary OH groups,
thirdly SEM-protected at the ring nitrogen, and finally bro-
minated to give the fully protected allyl bromide 2. The
second building block was prepared from compound 1 by
firstly hydrogenation, secondly TBS protection, and TBDPS
protection of the primary and secondary OH groups, respec-
tively, followed by final SEM protection of the ring nitrogen
to give compound 3 (Scheme 2). Enolate–allyl bromide cou-
pling furnished the coupled dinucleotide 4 as a diasterotopic
mixture.[11] Separation of the two diastereoisomers was pos-
sible after TBS-group cleavage to give compound 5 and 6
after isocratic normal-phase HPLC on a silica-gel column
with ethyl acetate (38%), n-heptane (58%), and a small per-
centage of isopropanol (4%). The isopropanol helps to sep-
arate residual compound 3 from the reaction mixture. Com-
pounds 5 and 6 were subsequently protected by using DMT
and the secondary OH groups were converted to the corre-
sponding phosphoramidites 7 and 8. Incorporation of both
compounds into DNA was performed using standard phos-
phoramidite DNA synthesis with slightly modified condi-
tions, in which only the coupling times were doubled. To
remove the SEM groups after DNA synthesis, the solid-
phase material containing the oligonucleotides was treated
with SnCl4 in CH2Cl2 for 1 h at room temperature. This is a
critical step because of the general sensitivity of nucleic
acids towards acids and also Lewis acids. After removal of
the deprotection solution, the SEM-deprotected oligonucle-
otides were cleaved from the resin with concomitant cleav-
age of the standard nucleobase protecting groups with NH3
in ethanol/water. The crude oligonucleotides were subse-
quently treated with 3·HF-Et3N to remove the TBDPS
groups. The DNA was finally purified by reversed-phase
HPLC using
a
water/acetonitrile gradient with an
9652
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 9651 – 9657

Spore Photoproduct DNA Lesion
FULL PAPER
Et3N·HOAc buffer. The integrity of the oligonucleotides
was investigated with MALDI-TOF mass spectrometry. The
sequences of the prepared oligonucleotides together with
the calculated and found molecular weights are depicted in
Table 1.
Table 1. Depiction of the prepared oligonucleotides ODN1 and ODN2
using compounds 7 and 8. ODN3 and ODN4 are the counter strands
used for crystallization and repair assays, respectively. ODN5 and ODN6
are the expected product strands of the repair reaction.
Number Sequence (5’!3’)
N
N
ODN 1
ODN 2
ODN 3
ODN 4
ODN 5
ODN 6
AGGGT
AGGGTACHTUNGTRENNUNG
ATGCGACCAACCCT
GACCAACCCT
TGGTC
G
3036
3036
4194
2957
1495
1544
3035
3036
4195
2957
1496
1544
AGGGT
To unequivocally assign the stereochemistry at the C5 po-
sition of the spore lesion analogues and to investigate the
pairing properties of both lesions in the duplex, we used the
possibility to crystallize DNA in the complex with the DNA
polymerase from Geobacillusstearothermophilus (Bacillus
stearothermophilus, B. st. Pol I).[12] For this study we placed
the lesion not in the active site of the polymerase but in the
double-stranded region to study the lesion in a B-duplex en-
vironment. Indeed, upon combining a solution of the spore
lesion analogue-containing DNA duplexes ODN1:ODN3
and ODN2:ODN3 (with the lesion positioned 5 bases away
from the active site) and a solution of the B. st. Pol I protein
gave, under suitable conditions (45–50% (NH4)2SO4, 3.5–
4.5% 2-methyl-2,4-pentanediol (MPD), and 100 mm 2-(N-
morpholino)ethanesulfonic acid (MES), pH 5.8, hanging-
drop vapor diffusion method), crystals that diffracted X-rays
to a resolution of 2.15 (5R) and 2.80 ꢁ (5S). The absolute
stereoconfiguration, determined with this crystallographic
method is in agreement with the recent papers from Bardet
and Broderick,[10b,c] thereby showing that older assignments
led to erroneous conclusions.[11]
Figure 1. Depiction of the crystal structure of the 5’!3’ 5R spore photo-
product analogue-containing duplex in complex with B. st. Pol I (pdb en-
try 2y1j). a) B. st. Pol I in complex with the 5R-lesion-containing DNA.
b) Enhanced representation of the 5R spore photoproduct with overlaid
electron density. c) Overlay of the lesion (red) and of an undamaged
DNA segment (blue). d) Overlay of undamaged DNA and the spore pho-
toproduct-containing DNA bound to the enzyme.
bonds are at an angle of 208. The glycosidic bonds in a di-
pyrimidine sequence are almost parallel (ꢁ58 for 1NJY),
making the angle between the glycosidic bonds a good indi-
cator of backbone distorsion. Interesting is a comparison
with the structures of the other UV-induced lesions shown
in Figure 2. Cyclobutane pyrimidine dimer (CPD) lesions,
Figure 1a shows a schematic representation of the struc-
ture. Figure 1b provides a magnification of the lesion, to-
gether with an example of the electron density. The result of
the structure with the 5R lesion is that this diastereoisomer
fits quite well into the DNA duplex and that it forms two
almost perfect Watson–Crick base pairs with the two oppo-
site adenines with hydrogen-bond lengths between 3.1 and
2.9 ꢁ (see Figure S1 in the Supporting Information). It is
also notable that the missing phosphodiester backbone does
not seem to disturb the structure. Overlay of our structure
with a structure of B. st Pol I in complex with undamaged
DNA (pdb entry 1NJY) shows that the phosphate of the un-
damaged oligonucleotide fits perfectly between the 3’ OH
group of the thymidine and the 5’ hydroxyl group of the 5,6-
dihydrothymidine part of the lesion (Figure 1c).[13] The two
heterocycles of the 5R spore lesion are at an angle of ap-
proximately 308 to form the new bond and the glycosidic
such as the TACTHUGNETRNNU(G CPD)T lesion (Figure 2b), are considered to
disturb the duplex structure only slightly. Here, however,
the two 5,6-dihydrothymine rings are arranged at an angle
of about 508 and the glycosidic bonds at about 308 (pdb en-
try 3MR3, see Figures S2 and S3 in the Supporting Informa-
tion).[14] For the other two major UV-induced lesions, the
(6–4) lesions (Figure 2c, pdb: 3CVU) and the Dewar lesions
(Figure 2d, pdb: 2WQ6), the structural changes are much
more dramatic (see Figure 2).[15] In comparison to all these
UV-induced lesion structures, the result of our study is that
formation of the 5’!3’ 5R spore lesion induces by far the
smallest changes. This result is supported by a complete
overlay of the 5R lesion-containing DNA with undamaged
DNA (Figure 2d). The overall structural perturbation
caused by the 5’!3’ 5R lesion is only minimal.
The argument that the structural perturbations are small
because the duplex is stiffly held by the enzyme in the B
Chem. Eur. J. 2011, 17, 9651 – 9657
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9653

T. Carell et al.
Figure 2. Crystal structures of the four major photo lesions: a) spore pho-
toproduct (pdb entry 2yli), b) T (6–4)T
(CPD)T (3MR3),[14] c) T
(3CVU),[15b] and d) T(Dew)C* (2WQ6).[15a] Note that the phosphodiester
A
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
present in the SP lesion structure (a) was inserted by modeling (shown
with 50% transparency).
structure was investigated with the duplex containing the
5’!3’ 5S lesion. The structure of the B. st. Pol I in complex
with the duplex containing the 5S lesion analogue is provid-
ed in Figure 3a. A magnification of the lesion and a part of
the electron density is shown in Figure 3b. The overlay with
a complex of undamaged oligonucleotides bound to B. st.
Pol I is presented in Figures 3c and d (pdb: 1NJY). In this
structure we detect that the 5,6-dihydrothymidine is fully
turned around the bond between the C5-methyl group and
the C5 atom even if bound to the enzyme, thereby proving
that the structural analysis is able to report duplex perturba-
tions. We were unable to model a hypothetical phospho-
diester group between the two free hydroxyl groups of the
sliced backbone due to the tremendous distortion, which
shows that this 5S analogue is unable to fit into the duplex.
It is a hypothetical possible diastereoisomer of the spore
lesion, which cannot form due to the duplex environment in
which the lesion formation process has to take place. How-
ever one should keep in mind that the DNA is bound to a
protein and that the duplex distortion may change in solu-
tion.
To investigate whether the oligonucleotide containing the
5R-configured 5’!3’ lesion is accepted as a substrate as ex-
pected by the Spl enzyme we treated the two single strands
ODN 1 and ODN 2 individually with the SplG enzyme from
Geobacillusstearothermophilus at 378C and analyzed the re-
action mixture by HPLC.[16] The lack of a phosphodiester
backbone enabled us to detect the repair reaction through
the simple analysis of a strand break that occurs only when
repair takes place. The two thymidines are no longer con-
nected after repair. Two shorter DNA product strands are
Figure 3. Depiction of the crystal structure of the 5S spore photoproduct
analogue-containing duplex in complex with B. st. Pol I (pdb entry 2y1i).
a) Molecular architecture of B. st. Pol I in complex with the 5S lesion-
containing DNA. The enzyme is represented as surface (gray). The DNA
(green) is shown as a stick model with the lesion highlighted in pink. The
phosphate backbone is shown as a sketch. b) Enhanced representation of
the 5S spore photoproduct with overlaid electron density. c) Alignment
of the complex with undamaged DNA bound to B. st. Pol I (1NJY). The
3’ thymidine together with its sugar is completely turned. d) Overall over-
lay of the undamaged DNA (1NJY) and the spore photoproduct-contain-
ing DNA. Normal Watson–Crick base paring is impossible for the lesion.
formed (ODN 5 and ODN 6, see Table 1). Figure 4a shows
the results obtained for the 5R isomer. Line (I) depicts the
situation of the repair assay after addition of the enzyme
and SAM under anaerobic conditions. The starting material
ODN 1 disappeared and two new peaks are visible. To
prove that these new peaks are formed by ODN 5 and
ODN 6, we prepared a mixture of synthetic ODN 1, ODN 5,
and ODN 6 and analyzed the mixture of these three oligo-
nucleotides under the same HPLC conditions. The data,
shown in line (II), are identical to those obtained from the
original enzyme assay. Finally a co-injection experiment
with the assay solution taken after complete repair was per-
formed (line (III)) showing again that the newly formed
peaks are caused by ODN 5 and ODN 6. ODN 1 containing
the 5R-configured 5’!3’ lesion product is as a result effi-
ciently repaired.
Figure 4b shows the results obtained with ODN 2, which
contained the incorporated 5’! 3’ 5S lesion analogue. The
HPLC chromatogram of assay (I) shows only the starting
9654
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 9651 – 9657

Spore Photoproduct DNA Lesion
FULL PAPER
Conclusion
We present the synthesis of both diastereoisomers (5R and
5S) of the 5’!3’ spore photoproduct as phosphoramidites
with an open backbone. These phosphoramidites enabled
the incorporation of the lesion analogues into oligonucleo-
tides by using standard solid-phase synthesis in combination
with an elaborate multistep oligonucleotide-deprotection
protocol. Together with previous synthetic results from our
laboratory, which show incorporation of the backbone-
opened 3’!5’ 5S and 5R lesion analogues, the whole set of
the 4 possible spore lesions can now be prepared and insert-
ed into DNA.[8] Repair studies show that the 5’!3’ 5R ana-
logue, incorporated in oligonucleotides, is efficiently re-
paired. Crystallization of a DNA duplex containing the 5’!
3’ 5R compound in complex with a polymerase, which was
used only as a crystallization matrix, with the compound sit-
uated in a duplex-binding region of the protein, shows that
the lesion disturbs the duplex only slightly, in comparison
with double-stranded DNA bound to the same protein. In
fact, in comparison with structures of the other main UV-in-
duced lesions, (CPD, (6–4), and Dewar) the 5’!3’ 5R spore
lesion induces only minor structural perturbations. In the
lesion, both thymines form almost perfect Watson–Crick hy-
drogen bonds to the opposite adenine base-pairing partners.
The structure obtained here allows modeling of a phospho-
diester group between the two 3’ and 5’ OH groups present
on the sugar molecules, which shows that the lack of these
groups affect the overall structure only slightly. In the lesion
the two thymine heterocycles are stiffly connected by means
of a covalent bridge and embedded in the B-duplex environ-
ment. The 5’!3’ 5S lesion analogue in contrast disturbs the
duplex significantly and more importantly it is impossible to
model a phosphodiester bridge between the two 2-deoxyri-
bose units, thereby showing that this hypothetical compound
does not fit into the duplex, in full agreement with the idea
that the structural constraints of the B-duplex prohibit for-
mation of the 5S spore photoproduct. Together with our
previous study of the two 5S/R 3’!5’ compounds, we can
conclude that indeed only the two spore lesions 5S 3’!5’
and 5R 5’!3’ can form in the duplex. In the 5’!3’ case it is
the 5R compound that is efficiently repaired. For the analy-
sis of the spore lesion formation and repair processes it is
important to analyze critically which reaction products 5’!
3’ versus 3’!5’ one analyzes. This has also affected a previ-
ous repair analysis by us, in which the different reaction di-
rections were not sufficiently considered.[9b,23]
Figure 4. Depiction of the HPLC analysis of the repair assay at 378C.
The spore photoproduct-containing DNA was incubated together with
the spore photoproduct lyase. a) Depiction of the results obtained with
the 5R analogue-containing ODN 1. (I) shows the reaction mixture.
Here, only the oligonucleotides ODN 5 and ODN 6 are visible and the
lesion is consequently fully repaired. Line (II) shows a mixture of synthe-
sized ODN 1, ODN 5, and ODN 6. Line (III) shows the co-injection of
the mixture (II) to the assay solution. No new peaks, aside from ODN 1,
appear, proving that the peaks in (I) are the repaired strands ODN 5 and
ODN 6. b) Depiction of the repair results obtained with ODN 2 contain-
ing the 5S analogue. In (I) only the starting material (ODN 2) is visible.
No repair occurs. Line (II) shows a mixture of synthesized ODN 2,
ODN 5, and ODN 6. Line (III) shows the co-injection of the mixture (II)
to the assay solution. No new peaks, aside from ODN 5 and ODN 6
appear.
material ODN 2. The chromatograms of the mixture of syn-
thesized ODN 5, ODN 6, and ODN 2 shown in (II) and the
co-injection of this mixture and the assay solution, provided
in (III), prove that the 5S lesion is not repaired.
Since DNA is naturally double stranded, the assays were
repeated with the lesion-containing duplexes ODN1:ODN4
and ODN2:ODN4. The temperature was reduced to 208C
to avoid melting of the duplex during the repair reaction
and increased to 608C during the HPLC analysis to ensure
single-stranded oligonucleotides for the analysis (see Fig-
ure S4 in the Supporting Information). Again, repair was ob-
served for the 5’!3’ 5R lesion, whereas the 5’!3’ 5S lesion
showed no reaction. Due to the unsuitable temperature for
a thermophilic enzyme and the lower kinetics, caused by the
duplex, the reaction proceeded with only about 70% yield.
Experimental Section
General: The Escherichia coli strain TunerACTHNUTRGNE(UNG DE3)pLysS (Novagen) and
the pDEST17 expression vector (Invitrogen) were used for overexpres-
sion of SplG. The BugBuster 10X protein-extraction reagent was pur-
chased from Novagen and the nickel–nitrilotriacetic acid (Ni–NTA) su-
perflow resin was from Qiagen. S-Adenosylmethionine was obtained
from Sigma–Aldrich and used without further purification. Other chemi-
cals and solvents were purchased from ABCR, Alfa Aesar, Acros, Fluka,
Chem. Eur. J. 2011, 17, 9651 – 9657
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9655

T. Carell et al.
Sigma–Aldrich, or TCI in the qualities puriss., p.a., or purum. Dry sol-
vents (<50 ppm H2O) were obtained from Fluka and Acros. All reac-
tions employing dry solvents were performed under an inert atmosphere
(N2). Technical-grade solvents were distilled prior to use for column chro-
matography and liquid–liquid extractions on a rotary evaporator (Heidol-
phLaborota 4000). Reaction products were dried under high vacuum
(10 mbar). Aqueous solutions were dried on a SpeedVac plus CS110A or
SPD 111V from Savant or lyophilized (Christ ALPHA 2-4). Thin-layer
chromatography (TLC) was performed with aluminum plates (silica gel
60 F254, 10ꢂ5 cm). Substances were visualized by illumination with UV
light (l=254 nm). ESI-MS was performed on a Finnigan LTQ Fourier
transform ion cyclotron resonance (FTICR) spectrometer. MALDI-TOF
was performed on a BrukerAutoflex II. NMR spectra were recorded on
the following spectrometers: Varian Oxford 200, Bruker AC 300, Varian
XL 400, and Bruker AMX 600.
coding for an N-terminal His6 tag. For protein expression, transformed
TunerACTHNUGRTENUNG(DE3)pLysS (Novagen) cells were grown in Luria Bertani medium
supplemented with carbenicillin (100 mgmLꢀ1) and FeIII–citrate (100 mm)
at 378C. Protein expression was induced with 1 mm isopropyl b-d-thioga-
lactopyranoside (IPTG) at 228C (OD600 =0.6–0.7). After 1 h at 228C, the
culture was cooled to 48C and incubated overnight for 13–15 h. Cells
were harvested by centrifugation and transferred into a glovebox and all
following steps were performed under anaerobic conditions. Cell lysis
was carried out using 10ꢂ BugBuster (Novagen) in 50 mm HEPES,
300 mm NaCl, 0.5% Tween-20, and 5 mm b-mercaptoethanol, pH 8.0. The
protein was purified by nickel-affinity chromatography in
a batch
method. The column was washed with 50 mm HEPES, 300 mm NaCl,
15 mm imidazole, 0.5% Tween-20, and 5 mm b-mercaptoethanol, pH 8.0,
and the protein was eluted with 50 mm HEPES, 300 mm NaCl, 200 mm
imidazole, and 5 mm b-mercaptoethanol, pH 8.0. The protein sample was
concentrated in 50 mm HEPES (pH 8.0), 300 mm NaCl, and 5 mm dithio-
threitol (DTT) by centrifugal filter devices and stored at 48C and used
for the repair assays within one day without further purification.
Oligonucleotide synthesis and deprotection: The experiments were per-
formed on an Expedite 8909 nucleic acid synthesis system (PerSeptive-
Biosystems) using standard DNA synthesis conditions. Phosphoramidites
for dA, dC, dG, dT, and CPG carriers were obtained from Amersham,
Glen Research, or PE Biosystems. SEM-protecting groups were removed
by rinsing the 0.2 mmol cartridges with 1m SnCl4 in CH2Cl2 (20 mL) over
1 h at room temperature under dry conditions, followed by rinsing with
CH2Cl2 (20 mL). The oligonucleotides were removed from the resin with
concomitant cleavage of the standard nucleobase protecting groups by
treatment with concentrated NH3 in water/ethanol 3:1 (1 mL) at room
temperature for 18 h. The solution was decanted from the resin and dried
in a speed vacuum. For cleavage of the TBDPS groups the residue was
heated at 658C in a mixture of anhydrous DMSO (100 mL) and triethyla-
mine trihydrofluoride (TEA·3HF) (125 mL) for 1 h. After precipitation in
butanol and HPLC purification the oligonucleotides containing the lesion
analogues were obtained in yields of about 15%, with respect to the
resin loading.
Assay for DNA repair: To ensure full reduction, prior to the DNA repair
assay, the enzyme was incubated for at least 1 h with 4 mm Na dithionite.
To investigate DNA repair, 6 nmol of SplG were incubated with 2 nmol
single-stranded lesion-containing DNA. The reaction mixture further
contained 3 mm Na dithionite, 5 mm dithiothreitol (DTT), and 1 mm
SAM. After 14–16 h at 37 or 208C, respectively, the protein was removed
by precipitation and centrifugation and the supernatant was stored at
ꢀ208C until HPLC analysis.
Accession numbers: Atomic coordinates and structure factors (PDB en-
tries 2yli; 2ylj) of the protein in complex with DNA have been deposited
in the PDB at the EBI Macromolecular Structure Database (http://
www.ebi.ac.uk/pdbe).
HPLC: Separation of the diastereoisomers was performed on a Merck-
Hitachi system (L-7400 UV detector, L-7480 fluorescence detector, L-
7100 pump) using a VP 250/40 Nucleodur 100-5 column from Macherey–
Nagel (isocratic ethyl acetate/heptanes/isopropanol). Purification and
analysis of the ODNs was performed on a Waters system (Alliance 2695
with PDA 2996; preparative HPLC: 1525EF with 2484 UV detector)
with VP 250/10 Nucleodur 100-5 C18 ec and VP 250/4 Nucleodur 100-5
C18 ec columns from Macherey–Nagel using a gradient of 0.1m triethyla-
mine/acetic acid in water and 80% acetonitrile.
Acknowledgements
We thank the Excellence Cluster CiPSM and SFB 749 for financial sup-
port and the beamline staff at the SLS for setting up the beamlines and
for helpful advice.
[1] C. Desnous, D. Guillaume, P. Clivio, Chem. Rev. 2010, 110, 1213–
1232.
HPLC-ESI-MS: The samples (100 mL injection volume) were analyzed
by HPLC-ESI-MS on a Thermo Finnigan LTQ Orbitrap XL and were
subjected to chromatography with a Dionex Ultimate 3000 HPLC system
using a gradient of 2 mm ammonium formate in water and 80% acetoni-
trile over an Uptisphere120-3HDO column from Interchim.
[2] a) T. Carell, Angew. Chem. 1995, 107, 2697–2700; Angew. Chem.
Int. Ed. Engl. 1995, 34, 2491–2494; b) A. Sancar, Chem. Rev. 2003,
103, 2203–2237; c) J. Cadet, E. Sage, T. Douki, Mutat. Res. Fund.
Mol. Mech. Mutagen. 2005, 571, 3–17; d) J.-S. Taylor in DNA
Damage Recognition (Eds.: W. Siede, Y. Wah Kow, P. W. Doetsch),
CRC Press, 2005; e) K. Heil, D. Pearson, T. Carell, Chem. Soc. Rev.,
DOI: DOI: 10.1039/C000407N.
[3] J. E. Donnellan, Jr., R. B. Setlow, Science 1965, 149, 308–310.
[4] a) N. Munakata, C. S. Rupert, J. Bacteriol. 1972, 111, 192–198; b) P.
Fajardo-Cavazos, C. Salazar, W. L. Nicholson, J. Bacteriol. 1993, 175,
1735–1744; c) P. Setlow, Annu. Rev. Microbiol. 1995, 49, 29–54;
d) R. Rebeil, W. L. Nicholson, Proc. Natl. Acad. Sci. USA 2001, 98,
9038–9043; e) T. A. Slieman, R. Rebeil, W. L. Nicholson, J. Bacter-
iol. 2000, 182, 6412–6417; f) J. M. Buis, J. Cheek, E. Kalliri, J. B.
Broderick, J. Biol. Chem. 2006, 281, 25994–26003.
[5] a) A. Chandor-Proust, O. Berteau, T. Douki, D. Gasparutto, S. Ol-
lagnier-de-Choudens, M. Fontecave, M. Atta, J. Biol. Chem. 2008,
283, 36361–36368; b) J. Cheek, J. B. Broderick, J. Am. Chem. Soc.
2002, 124, 2860–2861; c) R. A. Mehl, T. P. Begley, Org. Lett. 1999, 1,
1065–1066.
[6] a) A. Chandor, O. Berteau, T. Douki, D. Gasparutto, Y. Sanakis, S.
Ollagnier-de-Choudens, M. Atta, M. Fontecave, J. Biol. Chem. 2006,
281, 26922–26931; b) A. Moysan, A. Viari, P. Vigny, L. Voituriez, J.
Cadet, E. Moustacchi, E. Sage, Biochemistry 1991, 30, 7080–7088.
[7] a) G. Lin, L. Li, Angew. Chem. 2010, 122, 10122–10125; Angew.
Chem. Int. Ed. 2010, 49, 9926–9929; b) J. Cadet, P. Vigny in Bioor-
B. st. Pol I purification, crystallization, and structure determination: B. st.
Pol I was overproduced, purified, and cocrystallized with the lesion-con-
taining duplexes by using published procedures.[12] Data were collected at
the PX III (Swiss Light Source (SLS), Villigen, Switzerland) and pro-
cessed with the programs XDS[17] and SCALA.[18] Structure solution was
carried out by molecular replacement with PHASER[19] using the coordi-
nates of PDB code 1U45. To reduce model bias, prior to model building
in COOT[20] the temperature factors were reset to the Wilson B factor
and a simulated annealing omit map, with the area around the lesion re-
moved, was calculated with PHENIX.[21] Restrained refinement was car-
ried out in REFMAC5.[22] All structural super positions and structural
figures were prepared with Pymol (Delano Scientific). Data processing
and refinement statistics are summarized in Table S5 in the Supporting
Information.
Expression and purification of recombinant SP lyase: The gene coding
for the SplG (UniProt accession ADU94823) from Geobacillusstearother-
mophilus DSM No. 22 was amplified using the polymerase chain reaction
(PCR) from genomic DNA using the primers 5’-CACCAT-
GAAACCGTTTGTGCCAAAACTT-3’ and 5’-TTACGTAAAATACTG-
CACTTGGG-3’ (Metabion). The PCR product was cloned using the
Gateway system (Invitrogen) first into the entry vector pENTR/TEV/d-
TOPO and further transferred into the destination vector pDEST17
9656
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 9651 – 9657

Spore Photoproduct DNA Lesion
FULL PAPER
ganic Photochemistry, Vol. 1, Photochemistry and the Nucleic Acids
(Ed.: H. Morrison), Wiley, New York, 1990.
Barends, A. F. Glas, M. J. Cryle, T. Domratcheva, S. Schneider, I.
Schlichting, T. Carell, Angew. Chem. Int. Ed. 2008, 47, 10076–10080.
[16] J. C. Pieck, U. Hennecke, A. J. Pierik, M. G. Friedel, T. Carell, J.
Biol. Chem. 2006, 281, 36317–36326.
[17] W. Kabsch, J. Appl. Crystallogr. 1993, 26, 795–800.
[18] Collaborative Computational Project Number 4, Acta Crystallogr.
Sect. D 1994, 50, 760–763.
[8] E. Bꢃrckstꢃmmer, T. Carell, Chem. Commun. 2008, 4037–4039.
[9] a) M. G. Friedel, J. C. Pieck, J. Klages, C. Dauth, H. Kessler, T.
Carell, Chem. Eur. J. 2006, 12, 6081–6094; b) M. G. Friedel, O. Ber-
teau, J. C. Pieck, M. Atta, S. Ollagnier-de-Choudens, M. Fontecave,
T. Carell, Chem. Commun. 2006, 445–447.
[10] a) S. C. Silver, T. Chandra, E. Zilinskas, S. Ghose, W. E. Broderick,
J. B. Broderick, J. Biol. Inorg. Chem. 2010, 15, 943–955; b) T. Chan-
dra, S. C. Silver, E. Zilinskas, E. M. Shepard, W. E. Broderick, J. B.
Broderick, J. Am. Chem. Soc. 2009, 131, 2420–2421; c) C. Mantel,
A. Chandor, D. Gasparutto, T. Douki, M. Atta, M. Fontecave, P. A.
Bayle, J. M. Mouesca, M. Bardet, J. Am. Chem. Soc. 2008, 130,
16978–16984.
[11] S. J. Kim, C. Lester, T. P. Begley, J. Org. Chem. 1995, 60, 6256–6257.
[12] M. Munzel, L. Lercher, M. Mꢃller, T. Carell, Nucl. Acids Res. 2010,
38, e192.
[19] A. J. McCoy, R. W. Grosse-Kunstleve, L. C. Storoni, R. J. Read,
Acta Crystallogr. Sect. D 2005, 61, 458–464.
[20] P. Emsley, B. Lohkamp, W. G. Scott, K. Cowtan, Acta Crystallogr.
Sect. D 2010, 66, 486–501.
[21] P. D. Adams, R. W. Grosse-Kunstleve, L. W. Hung, T. R. Ioerger,
A. J. McCoy, N. W. Moriarty, R. J. Read, J. C. Sacchettini, N. K.
Sauter, T. C. Terwilliger, Acta Crystallogr. Sect. D 2002, 58, 1948–
1954.
[22] G. N. Murshudov, A. A. Vagin, E. J. Dodson, Acta Crystallogr. Sect.
D 1997, 53, 240–255.
[13] S. J. Johnson, L. S. Beese, Cell 2004, 116, 803–816.
[14] C. Biertꢃmpfel, Y. Zhao, Y. Kondo, S. Ramon-Maiques, M. Gregory,
J. Y. Lee, C. Masutani, A. R. Lehmann, F. Hanaoka, W. Yang,
Nature 2010, 465, 1044–1048.
[15] a) A. F. Glas, E. Kaya, S. Schneider, K. Heil, D. Fazio, M. J. Maul, T.
Carell, J. Am. Chem. Soc. 2010, 132, 3254–3255; b) M. J. Maul, T. R.
[23] See also, G. Lin, C.-H. Chen, M. Pink J. Pu, L. Li, Chem. Eur. J.
2011, 17, 9658–9668.
Received: January 18, 2011
Revised: May 5, 2011
Published online: July 20, 2011
Chem. Eur. J. 2011, 17, 9651 – 9657
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9657