Indications of 5′ to 3′ Interbase Electron Transfer as the First Step of Pyrimidine Dimer Formation Probed by a Dinucleotide Analog

Article abstract of DOI:10.1002/chem.201700045

Pyrimidine dimers are the most common DNA lesions generated under UV radiation. To reveal the molecular mechanisms behind their formation, it is of significance to reveal the roles of each pyrimidine residue. We thus replaced the 5′-pyrimidine residue wit

Full text of DOI:10.1002/chem.201700045

A Journal of
Accepted Article
Title: Indications of 5' to 3' Interbase Electron Transfer as the First Step
of Pyrimidine Dimer Formation Probed by a Dinucleotide Analog
Authors: Yajun Jian, Egle Maximowitsch, Degang Liu, Surya Adhikari,
Lei Li, and Tatiana Domratcheva
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201700045
Link to VoR: http://dx.doi.org/10.1002/chem.201700045
Supported by

Chemistry - A European Journal
10.1002/chem.201700045
FULL PAPER
Indications of 5′ to 3′ Interbase Electron Transfer as the First Step
of Pyrimidine Dimer Formation Probed by a Dinucleotide Analog
†
[a, b]
† [c]
[a]
[a,d]
[c]
Yajun Jian,
Egle Maximowitsch, Surya Adhikari, Lei Li,*
and Tatiana Domratcheva*
Abstract: Pyrimidine dimers are the most common DNA lesions
generated under UV radiation. To reveal the molecular mechanisms
behind their formation, it is of significance to reveal the roles of each
pyrimidine residue. We thus replaced the 5'-pyrimidine residue with
a photochemically inert xylene moiety (X). The electron-rich X can
be readily oxidized but not reduced defining the direction of
interbase electron transfer (ET). Irradiation of the XpT dinucleotide
under 254 nm UV light generates two major photoproducts: a
pyrimidine (6-4) pyrimidone (6-4PP) analog and an analog of the so-
called spore photoproduct (SP). Both products are formed via
reaction at C4=O of the photo-excited 3'-T, which indicates that
excitation of a single “driver” residue is sufficient to trigger pyrimidine
dimerization. Our quantum-chemical calculations demonstrated that
photo-excited 3'-T accepts an electron from 5′-X. The resulting
charge-separated radical pair lowers its energy upon formation of
interbase covalent bonds eventually yielding 6-4PP and SP.
leading to the formation of Dewar valence photoproduct (Dewar
[
3]
PP) via a ring rearrangement reaction. Moreover, in bacterial
endospores where the genomic DNA is suggested to adopt an
A-like conformation, UV radiation results in another thymidine
dimer, 5-thyminyl-5,6-dihydrothymine, which is commonly
[4]
referred to as the spore photoproduct (SP). It is generally
believed that SP is mainly formed in spores, whereas CPD, 6-
4PP and Dewar PP are formed in a wide range of species. In
humans, mutations induced by pyrimidine dimers are considered
[5]
as the dominant cause for the occurrence of skin cancer. It is
therefore of significance to understand the pyrimidine
photochemistry in detail.
O
O
N
O
N
O
N
O
N
HN
O
HN
NH
(A)
HN
NH
(B)
NH
O
N
O
O
UV
O
O
UV
N
O
O
O
O
O
O
5'
RO
O−
P
3'
OR
O−
P
3'
OR
5'
OR
3'
O O
P− O
RO
O
O
RO
5
O
O
O
'
O
O
O
CPD
TpT
Oxetane
Introduction
(
C) UV
O
O
N
H
O
H
As the genetic information carrier, DNA is under constant
environmental stresses, such as UV radiation from the solar light.
Among the four 2′-deoxyribonucleotides, thymidine is the most
HN
OH
O
N
HN
H
O
NH
N
O
N
O
N
O
O
O
O
O−
O
P
O
3'
OR
3'
OR
[
1]
5
RO
'
5'
RO
O
P
O
UV sensitive one followed by 2′-deoxycytidine. UV radiation of
pyrimidine residues leads to dimerization reactions, with
cyclobutane pyrimidine dimers (CPDs) and pyrimidine (6-4)
pyrimidone photoproducts (6-4PPs) as the two most common
O
−
O
O
6-4PP
UV
[
1a, 1c-e]
O
N
photoproducts (Figure 1).
The pyrimidone moiety in 6-4PP
O
O
N
O
O
[2]
HN
NH
HN
OH
N
O
N
HN
NH
absorbs at 326 nm, that can be excited by UVA/UVB light,
N
O
O
N
N
O
O
O
O
O
O
O
3'
OR
O
O
3'
OR
O−
O
P
O
O
5'
RO
5'
RO
3'
OR
5'
P− O
O
P
O
O
RO
O
−
O
O
O
[a]
Dr. Yajun Jian, Dr. Surya Adhikari, Prof. Lei Li
Department of Chemistry and Chemical Biology
Indiana University-Purdue University Indianapolis (IUPUI)
Dewar PP
SP
402 North Blackford Street, Indianapolis, Indiana, 46202, United
States
E-mail: lilei@iupui.edu
Dr. Yajun Jian
School of Chemistry & Chemical Engineering
Shaanxi Normal University (SNNU)
No. 620, West Chang'an Avenue, Xi'an, Shaanxi, 710119, P. R.
China
Figure 1. Thymidine dimerization reactions under UV radiation. (A) Formation
of CPD. (B) Formation of 6-4PP via a putative oxetane intermediate. The
formed 6-4PP can further isomerize into Dewar PP under UVA/UVB light. (C)
Formation of SP which is initiated by an H-atom abstraction mediated by a
diradical species formed at the C5=C6 bond followed by radical recombination.
[b]
[c]
[d]
Egle Maximowitsch, Dr. Tatiana Domratcheva
Department of Biomolecular Mechanisms
Max-Planck Institute for Medical Research
Jahnstrasse 29, 69120 Heidelberg, Germany
E-mail: Tatjana.Domratcheva@mpimf-heidelberg.mpg.de
Prof. Lei Li
Department of Dermatology
Indiana University School of Medicine
Indianapolis, Indiana 46202, United States
Despite the fact that CPDs, 6-4PPs and SP have been
[
6]
discovered for half a century or so,
issues still remain
regarding their formation mechanisms. The formation of CPDs is
indicated to occur via a [2 + 2] cyclophotoaddition reaction. Both
pyrimidine singlet and triplet excited states are found to be able
[
7]
to induce CPD formation. Between these two states, the triplet
states represent a minor reaction channel and only contribute to
†
<
–
10% of CPDs formed in single-stranded DNA under UVC (100
The authors contributed equally to this work.
[
8]
290 nm) and UVB (290 – 320 nm) irradiation. Surprisingly,
Supporting information for this article is given via a link at the end of
the document.
the long-wavelength UVA light (320 – 400 nm) was also found to
induce CPD formation potentially through triplet energy
[
1e, 9]
transfer.
CPDs were found to be generated after finishing
chemiexcitation
exposure to UVA light presumably by
a
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201700045
FULL PAPER
[
10]
mechanism.
6-4PP is indicated to occur via an oxetane or
Previously, we replaced the 3'-T with a toluene (To) or a meta-
azetidine intermediate formed between the two involving
pyrimidine residues after their photo-excitation by UVC/UVB
light; such an intermediate then thermally decomposes into the
xylene (X) group, the resulting dinucleotide TpTo or TpX
[
22]
supports CPD formation under UVC light (Scheme 1). In this
report, we retained the 3'-T and replaced the 5'-T by an X to
obtain an XpT dinucleotide to shed light on the 6-4PP formation
because under UV light, the xylene oxidation is preferred in XpT
resulting a pair of 5′-X cation and 3′-T anion radicals. The XpT
photoreaction indeed affords a 6-4PP analog as a major product,
providing direct experimental evidence that the photo-excited 3′-
residue alone is enough to trigger the 6-4PP formation and the
reaction proceeds via a 5′ to 3′ ET process. The reaction also
yields a SP analog; while no CPD analogs are obtained. Our
quantum-chemical calculations further demonstrate that the
interbase ET from the 5′-X to the 3′-T is responsible for the
formation of both photoproducts; the 6-4PP and SP formation
may thus share some similar mechanistic scheme.
[
1a]
corresponding 6-4PP.
UVA light is unable to induce 6-4PP
formation, indicating that 6-4PP is formed via pyrimidine singlet
[1d, 9a]
excited states.
SP formation is achieved via a diradical
mediated H-atom abstraction followed by radical recombination
[
11]
mechanism.
photosensitizer, UVA light can also induce SP formation,
suggesting that thymine triplet excited state is likely responsible
In the presence of pyridopsoralens as
[
12]
[
13]
for the SP photochemistry.
A recent computational study with dinucleotide TpT implies
that the excitation of the 3'-T alone may be sufficient for the 6-
[
14]
4PP photochemistry.
The 3'-T is excited to the triplet state
displaying an elongated C4=O bond via a fast singlet−triplet
crossing, which then reacts with the C5=C6 bond of the 5'-T to
[
14]
form 6-4PP.
This result contradicts with the previous finding
that only the singlet excited state is involved in the 6-4PP Results
[
1d]
formation; the reason for this discrepancy is unclear. Further
computational studies suggest the 6-4PP photochemistry is
triggered by the so-called charge-transfer states featuring
electron transfer (ET) from the 5'-base to the 3'-base in the TpT
XpT
1
(A)
(B)
2
[
7, 15]
30 min
*
and TpC sequences.
However, given that the two pyrimidine
1
2
*
XpT
residues likely possess similar redox potentials, ET in both
directions, from 5′ to 3′ and from 3′ to 5′, should be possible.
Therefore, there is an urgent need to identify an experimental
system by replacing a pyrimidine with a photo-inert but ET-active
moiety to test these computational results above.
1
5 min
0 min
1
Because pyrimidine photoreactions are largely governed by
[
16]
the base stacking conformation,
the replacement should
5
min
possess a similar shape and stacking interaction to a natural
pyrimidine. Toluene and its derivatives like 2,4-difluorotoluene
0
5
10
15
min
20 25
30
0
5
10
15
min
20 25
30
[
17]
are close steric analogues of thymine;
they can be
incorporated into DNA by polymerases as if they were
Figure 2. HPLC chromatograms monitored at 230 nm (A) and 320 nm (B) for
the XpT photoreaction initiated under 254 nm UV light in 10 mM pH 7.4
phosphate buffer at ambient temperature. XpT was irradiated for 5, 10, 15 and
30 minutes; two thermally stable products 1 and 2 were generated as the
major species, both of which absorb at 320 nm. The XpT photoreaction also
generated several minor species. Using mass spectrometry, the compounds
donated with * appear to be water adducts, the structures of which were not
pursued in the current study.
[
17c, 18]
thymines.
The incorporations do not lead to drastic DNA
[19]
conformational changes, as proved by NMR spectroscopic
and crystallographic studies.
[20]
However, these residues
possess rather different electronic structures from that of
thymine due to their higher symmetries, absence of the nπ*
[1a, 21]
states and lack of tautomeric structures,
making them
rather photo-inert. At the same time, their aromatic rings are
electron-rich, that facilitates electron donation but impedes
electron acceptance, thus determining the direction for any
possible interbase ET. These properties render toluene or its
derivatives as a perfect thymine substitute to reveal mechanistic
details in the 6-4PP photochemistry.
Product formation in XpT photoreaction. UV radiation of the
XpT solution resulted in two major products 1 and 2 that are
thermally stable at ambient temperature (Figure 2). The
-4
quantum yields of 1 and 2 were determined as (3.2 ± 1.0) × 10
-4
O
and (3.3 ± 1.0) × 10 respectively. ESI-MS analyses at the
negative ion mode revealed that 1 possesses an m/z value of
3
2
O
5
4
5
H
3
HN
2
1
HN
5
6
4
5
–
5'
HO
6
525.17 for the [M – H] species, which is the same to that of
O
N
6
O
5'
HO
1
hv
O
O
N
[23]
6
XpT. Formation of these species is consistent with the general
H
O-
O
O
O
OH
O
O-
property of pyrimidine photoreactions where pyrimidine residues
P
3'
O
O
OH
O
P
dimerize into species with identical molecular weight to the
3'
O
–
unreacted dinucleotides.
5
2
exhibits a [M – H] signal of
TpX
CPD (TpX)
[
23]
07.16, which is 18 amu less than that of XpT, suggesting an
elimination of a water molecule during its formation. Several
minor products were also generated during the course of the
reaction, which are likely XpT water adducts as indicated by
Scheme 1. TpX photoreaction yielding a CPD analog.
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201700045
FULL PAPER
mass spectrometry (+ 18 amu relative to XpT in regular water
and + 20 amu in O labeled water).
that they are still attached to unsaturated carbons. A singlet
signal of 4.05 ppm was found for H6a. Comparing with signals
found in XpT, these H NMR signals revealed in 1 move toward
1
8
1
1
1
1
0
0
0
.8
.5
.2
.9
.6
.3
2.4
2.0
1.6
1.2
0.8
0.4
1.2
1.0
0.8
0.6
0.4
0.2
the upfield region, indicating a loss of aromaticity at the xylene
ring. The biggest chemical shift change was found for H6a (>
XpT
(1)
(2)
3ppm), which, together with the chemical shift change for the
methyl group 8a, indicates that the reaction likely occurs at the
C5a-C6a bond. Thus, similar to the CPD formation in TpX
2
00
250
300
350 400
200
250
300
350 400
200
250
300
350 400
[22]
photoreactions,
the xylene ring in XpT also undergoes a
Wavelength (nm)
Wavelength (nm)
Wavelength (nm)
photo-addition reaction.
Figure 3. UV-vis spectra of XpT, 1 and 2 in aqueous solution at pH 7.0. 1 and
exhibit absorption bands at 320 and 317 nm respectively, indicating the
existence of a pyrimidone ring in these two products.
1
2
Table 1. Summary of the H NMR spectra for non-deoxyribose protons in XpT,
1 and 2.
a
Chemical Shift (ppm)
XpT
6.96
6.92
7.33
7.75
2.23
2.27
1.89
1
2
7.24
7.29
6.77
8.25
H3a
H4a
H6a
H6b
5.60
5.38
4.05
7.66
1.70
1.32
2.25
Characterization of 1 and 2 via UV-vis spectroscopy. Also as
shown in Figure 2, when monitored at 320 nm on HPLC, both 1
and 2 exhibit significant absorption while XpT barely absorbs.
We obtained the UV-vis spectra for XpT, 1 and 2 (Figure 3) to
shed light on their possible chemical structures. XpT does not
exhibit any absorption peak at > 300 nm, while both 1 and 2
absorb around 320 nm. Because an absorption band around
-CH
(8a) (–CH
CH (7b)
3
(7a)
2.27
-CH
3
2
- for 2)
4.00, 4.25
1.71
-
3
a. The NMR spectra of XpT can be found in the Supporting Information.
8
a
4
a
8
a
4
a
8
a
4
a
3a
3
20 nm is typically induced by the pyrimidone moiety formed in
3
a
3
a
OH
5
a
6a
[
2]
5
6
a
5
a
2
a
2
a
2
a
6
HO 7a
the 6-4PP species, the UV-vis spectra indicate that such a
moiety may be formed in both products.
a
4b
HO
7
a
a
HO
7
a
4b
7
b
1
a
7
b
1
a
O
4b 7b
1
a
5b
O
A
N
2b
O
A
O
A
H
5b
6b
N
HN
2b 6b
N
O
N
5
b
2
b
N
O
O
O
O
6
b
O
O
O
O
O
O
O
P
P
O
B
P
O
B
O
-
-
O
-
O
B
O
OH
OH
OH
XpT
(1)
(2)
Figure 5. Chemical structures of XpT, 1 and 2 indicated by UV-vis and NMR
spectroscopy. All non-deoxyribose carbons are numbered.
The methyl group 7b (2.25 ppm) and H6b (7.66 ppm) at the
3′-thymine ring are still associated with unsaturated carbons. In
the HMBC spectrum, H6a and C5b exhibit a strong coupling
signal (J
C4b.
3
), indicating a direct connection between C6a and
The chemical shifts of the four carbons on the 3'-
[
23-24]
nucleobase of 1 (also in 2 as shown below) are nearly identical
1
3
to those at the pyrimidone ring in the C NMR spectrum of the
[
23, 25]
natural 6-4PP.
Moreover, the 3′-residue does not adopt a
Dewar structure, in which H6b exhibits a chemical shift of 5.28
ppm and C6b exhibits a chemical shift of 72.3 ppm due to the
[
3]
saturation at this hydrocarbon. Therefore, we conclude that 1
possesses a 3'-pyrimidone ring and is a 6-4PP analog (Figure 5),
agreeing with the UV-vis data.
1
Figure 4. Zoom-in-view of the H NMR spectra of XpT, 1 and 2 showing
signals for protons associated with non-deoxyribose carbons. D
2
O was used
In the ROESY spectrum, H6a interacts with H1a' and H4a' but
not with H2a' and H3a' at the 2′-deoxyribose, implying that the
as the NMR solvent for 2; MeOH-d was used for the other two compounds.
4
5'-nucleoside adopts a syn conformation. H6b exhibits strong
interactions with H2'b and H3'b, suggesting an anti conformation
Characterization of 1 via NMR spectroscopy. We next turned
to NMR spectroscopy for structural analysis. The H NMR
[
26]
1
for the 3'-nucleoside. Similar interactions were observed in the
[
2, 24, 26b]
natural 6-4PP TpT.
However, since the xylene ring has
spectrum of 1 shows all three methyl groups at 7a, 7b, and 8a
are retained (Figure 4). For hydrogen atoms on the 5′-xylene
ring, the two doublet signals at 5.60 and 5.38 ppm exhibit the
same JH-H coupling constant; they were assigned as H3a and
H4a respectively. The relatively downfield chemical shifts imply
no carbonyl moiety at C4a, the absolute configuration at C5a is
S not R found in the natural 6-4PP; both C5a and C6a in 1 adopt
an S configuration.
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201700045
FULL PAPER
1
Characterization of 2 via NMR spectroscopy. In the H NMR
spectrum, one of the methyl moieties (8a) disappeared and new
signals composed of a pair of doublets with a large coupling
constant (JH-H = 15.5 Hz) emerged at 4.00 and 4.25 ppm (Figure
rise to the longest-wavelength absorption band corresponds to
1
1
the ππ* excited thymine, the T* state. In XpT, the T* state at
109.9 kcal/mol (corresponds to 260 nm) is likely to be populated
in our experiments using the 254-nm excitation-energy source.
4, Table 1), indicating the formation of a methylene group at C8a. At 121.3 kcal/mol in XpT and at 132.8 kcal/mol in TpT,
Moreover, the H8a protons exhibit strong coupling interactions
delocalized states that correspond to electron transfer (ET) from
1
+
–
1
+ –
(
J
2
) with C4b and C5a, indicating C8a is directly linked to C4b.
the 5′ base to the 3′ base, the (X T ) and (T T ) states, were
found. The variation of the ET energy in the two compounds is
consistent with the ionization potentials of the xylene and
thymine electronic systems (Table 3). In TpT, we also identified
a delocalized state that corresponds to ET from 3'-T to 5'-T, the
1
The xylene ring in 2 remains intact, as implied by the H NMR
signals similar to those found in XpT. H6b (8.25 ppm) is still
located on an aromatic ring, which is consistent with the
formation of a 3′-pyrimidone moiety. Therefore, the NMR data
supports the structure of 2 shown in Figure 5.
1
– +
(T T ) state, with the energy 136.9 kcal/mol.
In the ROESY spectrum, H6a exhibits crosspeaks with H2a'
H3a' and H5a' at the 2′-deoxyribose, suggesting that the 5'-
,
[
26]
nucleotide adopts an anti conformation.
In contrast, H6b
interacts with H1b' H4b' and H5b', but not with H2b' and H3b′,
,
suggesting a syn conformation for the 3'-pyrimidone. Therefore,
the two nucleotides in 2 exhibit an anti-syn combination, different
from the syn-anti combination found in 1.
Table 2. Singlet-state excitation energies (kcal/mol) and oscillator strength (in
brackets) in the XpT and TpT dinucleotides computed with the XMCQDPT2-
PCM method at the B3LYP-D3-optimized geometries.
1
1
+
-
1
+
-
1
- +
Dinucleotide
T*
(X T ) or (T T )
(T T )
a
XpT
109.9 (0.416)
121.3 (0.045)
132.8(0.011)
b
TpT
5’: 108.3 (0.331)
136.9 (0.003)
3
’: 110.5 (0.538)
Figure 6. Interbase ET in the excited XpT dinucleotide. (A). Ground-state- and
excited-state structures optimized using the TD-B3LYP-D3 and UB3LYP-D3
a, Results obtained with XMCQDPT2-SA5-CASSCF(6,4)-PCM ; b, results
obtained with XMCQDPT2-SA6-CASSCF(4,4)-PCM. Complete data are
presented in the Supporting Information Table S1.
methods.
A dinucleotide fragment comprising the two bases is shown.
Selected distances (Å) are indicated. (B). XMCQDPT2-SA9-CASSCF(6,4)-
PCM energies at the optimized geometries. At each geometry, the energies of
nine states (five singlets and four triplets) were computed (Supporting
Information Table S2). The five lowest-energy states are included in the graph.
Black and red lines connect the energies corresponding to the same electronic
structure labelled in Italic font. The energy corresponding to the optimized
geometry labelled in bold font is indicated with the filled symbol. For each
geometry, vertical excitation energies are indicated with open symbols
connected by a vertical line. The electronic structure of the states was
identified from the electron density analysis (Supporting Information Figure S3).
Table 3. Electron affinities and ionization potentials (kcal/mol) of xylene (X),
thymine (T) and triplet thymine ( T*) estimated using the (U)B3LYP-D3-PCM
method.
3
Base
Electron affinity
Ionization potential
134.3
X
T
4.6
30.6
137.8
3
a
a
T*
98.2
70.2
3
a, The energies of T* differ from the energies of T by the adiabatic excitation
energy of T* equal to 67.6 kcal/mol.
3
Energies of excited states in stacked XpT and TpT. To
elucidate how 1 and 2 are formed, we performed quantum-
chemical calculations. We considered a stacked geometry of the
XpT dinucleotide with both sugar residues adapting the C3'-endo
configuration, which is rather similar to the stacked geometry of
the TpT dinucleotide fragment in B-DNA. The B-DNA-like
conformation has the lowest energy among several tested
conformations of XpT according to our B3LYP-D3-PCM
computations (Supporting Information Figure S1). The high
tendency of the XpT dinucleotide to form a stacked conformation
was confirmed by our MD simulations (Supporting Information
Figure S2). At the optimized stacked structures of TpT and XpT,
we performed calculations of the excited-state energies and
oscillator strengths with the XMCQDPT2 method supplemented
with the equilibrium PCM treatment of the water solvent (below
referred to as XMCQDPT2-PCM). The electronic structure of the
states was identified by considering transition electron densities
Excited-state geometries of XpT. The computed excitation
spectrum suggests population of the singlet-excited thymine
1
3
Xp T*, triplet-excited thymine Xp T* and also charge-separated
+
–
radical pair X pT upon photoexcitation of XpT. Using the TD-
B3LYP-D3 calculations, the optimized excited-state geometries
1
+
–
Xp T* and X pT were obtained; whereas using the U-B3LYP-D3
3
calculations, the Xp T* geometry was obtained (Figure 6A).
[
14]
Because of the low-energy nπ* state of thymine, the optimized
1
1
1
Xp T* geometry corresponds to the nπ*/ T* state crossing
1
1
rather than to the true T*-state energy minimum. The Xp T* and
3
Xp T* geometries indicate that the ππ* excited thymine adapts a
non-planar geometry in the singlet and triplet electronic
+
-
configurations. In contrast, the radical cation and anion X and T
+
-
fragments of X pT assume planar geometries. The computed
+
-
net charges are +0.8 and –1.2 on X and T , respectively. In
1
3
Xp T* and Xp T*, similar to the ground-state XpT, the net
–
charges on X and T are −0.1 and –0.3, respectively. In the T
(
Supporting Information Figure S3). In the dinucleotide, the
radical anion, the C4b=O4b bond becomes elongated to 1.27 Å;
whereas the elongation of the C4b=O4b bond is less
lowest-energy state with a considerable oscillator strength giving
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201700045
FULL PAPER
pronounced in the ππ*-excited thymine than in the radical anion.
Furthermore, the interbase distances become shorter in X pT .
In particular, the distance between C5a and O4b decreases
support to the identified structures of 1 and 2, we present the
computed absorption spectra of these compounds (Figure 7).
The geometries of 1 and 2 were optimized using the B3LYP-D3
method. Then, excitation spectra were computed with the
XMCQDPT2-PCM method (Supporting Information Table S4).
We note a remarkable agreement of the computed spectra
(Figure 7) with the experimental spectra (Figure 3).
+
–
(
favorable for the formation of 1), and the distance between the
methyl group bond to C8a and O4b decreases (favorable for the
formation of 2).
Energies and electronic coupling controlling interbase ET.
At the optimized excited-state geometries, we compared the
energies of the singlet and triplet states computed with the
1
XMCQDPT2-PCM method (Figure 6B). The Xp T* geometry in
1
the T* state has an energy of 106.6 kcal/mol. This energy is
higher than the energy of the true energy-minimum because of
1
the incomplete geometry optimization; the T*-energy of 102.4
3
kcal/mol at the triplet geometry Xp T* is probably closer to the
1
energy of the true T*-state energy minimum. The charge-
+
–
1
+
–
3
+ –
separated radical pair X pT in the (X T ) and (X T ) states has
energies 107.5 and 106.9 kcal/mol, respectively, which is only
slightly higher than the energy of the singlet-excited thymine.
Figure 7. Computed UV-vis spectra of XpT and its major photoproducts 1 and
2. Excited-state computations were performed with the XMCQDPT2-PCM
method at the B3LYP-D3-optimized geometries. The excitation energies and
oscillator strengths were convoluted with Gaussian functions of 16-nm FWHM.
Complete data are listed in the Supporting Information Table S4.
3
3
The energy of Xp T* geometry in the T* state is only 72.4
kcal/mol. This low energy indicates that formation of the charge-
+
–
separated radical pair X pT is rather unfavorable in the triplet
3
3
T* state starting from the Xp T* geometry. In contrast, radical-
1
pair formation in the singlet state starting from the Xp T*
geometry is energetically possible, especially upon further
stabilization of the afforded radical-pair complex. The energy of
the X pT optimized geometry being lower than the vertical
excitation energy of thymine at the XpT ground-state geometry
is also favorable for interbase ET in the singlet-excited state.
Because of the low oscillator strengths, direct population of
+
–
1
+
–
+
–
the (X T ) state leading to the formation of the radical pair X pT .
upon photoexcitation is less probable as compared to the
population of the excited thymine. However, excited thymine
may undergo an interbase ET reaction, accepting an electron
from xylene. Indeed, electron affinity increases and the
ionization potential decreases in the excited thymine (Table 3).
Thus, thymine photoexcitation promotes ET in both directions in
the TpT dinucleotide. In XpT, excited thymine is coupled to
xylene, which is likely to serve as electron donor, therefore ET
from 5'-X to 3'-T is more favorable than ET from 3'-T to 5'-X.
In addition to the favorable energies, the rate of interbase ET
1
depends on the amount of electronic coupling between the T*
1
+ –
and (X T ) states. We estimated the electronic coupling from
the excitation energies, state dipole moments and transition
[
27]
dipole moments according to the Mulliken-Hush method
Supporting Information Table S3). The obtained coupling of the
Figure 8. Formation of 1 and 2 form XpT. (A). Chemical structures of the XpT
reactant and ground-state intermediates Oxetane and Spore-Int. (B) Key
structures along the reaction pathways according to the (U)B3LYP-D3
calculations (the respective energies are indicated in the Supporting
Information Figure S4). Only a fragment of the dinucleotide comprising the two
(
1
1
+
–
1
T* and (X T ) states at the Xp T* geometry equals 69 meV.
The coupling computed for the nπ*-excited thymine with the
1
+ –
(
X T ) state of 17 meV is also substantial, indicating that
bases is shown. Selected distances (Å) are indicated. The CoIn
1 2
and CoIn
population of the nπ*-state thymine does not abolish interbase
ET. The estimated amounts of coupling correspond to the
maximum ET rate constant (assuming the activation energy is
structures were optimized with the SA2-CASSCF(12,12) method.
[
28]
-1
1
-1
zero)
of 15 ps ( T* state) and 1 ps (nπ* state). For
Photochemical reaction coordinates in the ground- and
triplet states. To elucidate how 1 and 2 are formed, we
performed calculations of the respective reaction coordinates in
the ground closed-shell (CS) state and in the first excited triplet
state, using the (U)B3LYP-D3 method. These rather
straightforward computations enabled us to identify the
1
1
+ –
comparison, the electronic coupling of the T* and (T T ) states
at the ground-state TpT geometry is 45 meV corresponding to
-1
the maximum ET rate constant of 6 ps .
Computed absorption spectra of XpT and its photoproducts.
To validate our computational procedure and to provide further
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201700045
FULL PAPER
structures of transition states and intermediates that are involved
in the formation of interbase covalent bonds (Supporting
Information Figure S4). We found that 1 and 2 are formed via
the Oxetane and Spore-Int intermediates, respectively (Figure
further shortening of the C6a−C4b and C8a−C4b distances
leads to the formation of Oxetane and Spore-Int. We note that
1
the T* state shows a similar energy trend as the CS state,
which indicates that formation of Oxetane and Spore-Int in the
1
8
A). Two high-energy saddle-points TS
1
and TS
2
were found in
T* state requires high activation energies. Thus, thymine
the ground state (Figure 8B). The TS
1
geometry features an
photoexcitation as such does not catalyze formation of the
photoproducts. Instead, excited thymine undergoes interbase ET
which catalyzes the formation of the photoproducts.
already-formed C5a−O4b bond and a 2.30-Å distance between
C6a and C4b, corresponding to an activated (pre-formed) bond.
Further shortening of the C6a−C4b distance leads to Oxetane.
2
The TS geometry features a formed O4b−H bond whereas the
distance between C8a and C4b is found at 2.81 Å, which further
decreases in Spore-Int. The Oxetane pathway was found to be
remarkably similar in XpT and TpT, even though the Oxetane
energy is lower in TpT (Supporting Information Figure S5).
3
In the triplet-state T*, formation of the interbase covalent
bonds occurs in a step-wise fashion via a biradical intermediate.
Formation of the C5a−O4b and O4b−H bonds leads to the
3
3
biradical intermediates Int
1
and Int
2
, respectively (Figure 8B). In
3
3
Int and Int , the spin density becomes delocalized over the
1 2
two bases, indicating an interbase ET process. Formation of the
intermediates requires more than 10 kcal/mol activation energy
(
Supporting Information Figure S4), which is in agreement with
[
13-14, 29]
the previous findings for TpT.
Subsequent formation of
the C6a−C4b and C8a−C4b bonds results in recombination of
the biradical intermediates, eventually yielding Oxetane and
Spore-Int in the ground state.
Excited-state energies along the photochemical reaction
coordinates. To demonstrate the key role of 5' to 3' interbase
ET in the XpT photochemistry, we compared the energies of the
1
3
1
+
–
3
+ –
T*, T*, (X T ) and (X T ) states along the reaction
coordinates. The singlet-state route (Figure 9A) starts from the
+
-
Figure 9. Excited-state energies at the key structures from the photoreaction
coordinates. Only states of interest are included in the graph. (A) Energy
change along the singlet-excited state route. (B) Energy change along the
triplet-state route. The energy calculations and presentation style are similar to
those of Figure 6B. The geometries at which the energies were computed are
presented in Figure 8B and Supporting Information Figure S4. Geometries (1)
and (2) correspond to the energy maxima of the relaxed energy scans
computed with the TD-B3LYP-D3 method presented in the Supporting
Information Figure S6. All energies are listed in the Supporting Information
Table S5.
radical-pair reactant geometry X pT (107.5 kcal/mol) with the
shortened C5a−O4b and H−O4b distances. Further decrease of
the distances and formation of the respective covalent bonds
lowers the energy of the (X T ) state. At the same time, the
energy of the CS ground state significantly increases, leading to
conical intersections (CoIn) characterized by the degenerate CS
and (X T ) states. To characterize energy barriers separating
the X pT minimum and the CoIn structures, we computed
relaxed energy scans in the (X T ) state by gradually
decreasing the C5a−O4b and H−O4b distances and optimizing
all other geometry parameters using the TD-B3LYP-D3 method
1
+ –
1
+ –
+
–
1
+ –
The triplet route (Figure 9B) starts from the triplet reactant
3
(
Supporting Information Figure S6). Two energy barriers,
Xp T* (72.4 kcal/mol), which is formed from the initially
1
indicated as (1) and (2), were found (Figure 9A). The energies at
the geometries comprising the scans were recomputed with the
XMCQDPT2-PCM method. At this level of theory, only structure
populated T* state via an intersystem crossing (ISC) involving a
nπ* excited state. Interbase ET in the triplet state leads to the
formation of Int and Int (73.2 and 69.8 kcal/mol, respectively).
1 2
The energy of the interbase ET state (X T ) decreases upon
formation of the C5a−O4b and O4b−H bonds, so it becomes the
lowest-energy triplet state at Int
calculations demonstrated that both singlet and triplet routes
involve interbase ET from 5'-X to 3'-T. The considerable
[
14]
3
3
3
+ -
(
2) remains an energy barrier, whereas the relative energy of
1
+ –
structure (1) in the (X T ) state decreases, indicating
a
+
–
3
3
barrierless path from the X pT minimum to the CoIn
The minimum-energy CoIn geometries (Figure 8B) were
obtained using the CASSCF geometry optimization. CoIn
mediating formation of Oxetane is structurally very similar to an
analogous structure previously reported for two thymines.
CoIn corresponding to the Spore-Int formation is the first
structure of such a type characterized in our study for the first
time. CoIn and CoIn structurally resemble the ground-state TS
and TS geometries. From the CoIn and CoIn geometries,
1
structure.
1
2
and Int . Thus, the
1
3
3
+ –
T*− (X T ) excitation energy and reordering of the states at the
[
30]
3
3
3
Xp T*, Int
1 2
and Int geometries clearly indicate the existence of
3
3
3
energy barriers separating Xp T* from Int and Int . Therefore,
formation of the C5a−O4b and O4b−H covalent bonds may be
delayed in the triplet state as compared to the singlet-excited
1 2
state. At the Int and Int geometries, the small singlet-triplet
1 2
2
1
2
1
3
3
2
1
2
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201700045
FULL PAPER
energy difference is favorable for biradical recombination via ISC. residue where the pyrimidine anion eliminates a bromide before
However, recombination before the C6a−C4b and C8a−C4b
bonds are formed may result in decay towards XpT instead of
formation of Oxetane and Spore-Int, indicating that the
photoproducts are unlikely to form through the triplet route.
Rearrangement of the Oxetane intermediate to 6-4PP via
protonation at O4b. Despite the extensive effort to look for the
Oxetane and Spore-Int intermediates, neither species was
observed in our experiments. Previously, the assistance of the
solvent water molecules has been suggested to accelerate
Oxetane rearrangement to the final 6-4PP photoproduct via
proton transfer from N3bH to O4b. Here we extend this earlier
hypothesis by considering the rearrangement of Oxetane via
protonation of the O4b position which leads to the 6-4PP isomer
in which O4bH and N3bH are simultaneously protonated. We
compared protonation in the XpT Oxetane and a stable thietane
using the B3LYP-D3 calculations of the respective reaction
coordinates (Supporting Information Figure S7). Depending on
the nature of the proton donor, opening of the Oxetane/thietane
the resulting radical reacts with the purine cation to form the
crosslinking bond. Interbase ET in a DNA single strand (dT)18
occurs on a sub-picosecond timescale, and the resulting radical
[
34]
cation and anion species decay on
timescale. Moreover, interbase ET in the TpT step may also
a
100 picosecond
[35]
[
7, 15a]
play
a
key role in the 6-4PP formation.
However,
experimental evidence for directionality of the ET is not easy to
obtain due to the similar redox potential between the Ts making
both 5′ to 3′ and 3′ to 5′ interbase ET possible. The substitution
of a thymine by a xylene residue resulting in dinucleotides TpX
and XpT removes this ambiguity. UV irradiation of the TpX
[
31]
[
22]
complex in which 3′ to 5′ ET occurs only generates CPD. In
contrast, in the XpT studied here, 5′ to 3′ ET occurs that leads to
6-4PP formation and completely abolishes the CPD
photochemistry. Therefore, our system provides experimental
evidence that 6-4PP photochemistry is initiated by the 5′ to 3′ ET
process.
Using
extensive
quantum-chemical
calculations
we
ring by O4b/S4b protonation can be energetically favorable. In
demonstrated that the initially formed charge-separated radical
pair further lowers its energy and subsequently recombines by
formation of the interbase covalent bonds. In fact, the distinct
chemical structure of the two bases in the 6-4PP and SP-like
photoproducts is determined by interactions of the 3’-base
radical anion and 5’-base radical cation. Under favorable
reaction conditions in duplex DNA, the interbase ET leading to
6-4PP and SP formation may compete with other excitation-
decay pathways such as interstrand proton transfer resulting
from the H-bonding interactions between Watson-Crick base
pairs implied by recent excited state dynamics studies in duplex
+
particular, when the NH
4
molecule is considered as a proton
donor, the O4b/S4b protonation is practically barrierless. The
resulting protonated 6-4PP isomer is by 24 kcal/mol lower in
+
energy than the Oxetane-NH
4
complex and by 8.7 kcal/mol than
+
the theitane-NH
4
complex. This result indicates that O4b is
more basic than S4b. Rearrangement by protonation, thus, is
more favorable for Oxetane than for thietane, consistent with the
[
32]
known increased stability of thietane.
[
33a, d, e]
Discussion
DNA.
Formation of various pyrimidine dimers is long known to be
[
36]
wavelength-dependent. The short-wavelength UVC light leads
to the formation of all three dimers. In contrast, under the longer-
wavelength UVB light, 6-4PP formation is drastically reduced;
Pyrimidine dimers are the major DNA photoproducts under
solar irradiation in living organisms. Both residues in
a
bipyrimidine sequence can be photo-excited; whether these
residues contribute equally to the dimerization reaction however
is unclear. In the present contribution, we demonstrated that
after replacing the 5'-residue by a meta-xylene, the resulting XpT
dinucleotide after thymine photoexcitation readily supports 6-
[
1e, 9b, 37]
while no 6-4PP can be formed under UVA irradiation.
Moreover, in the presence of triplet-state photosensitizers, both
CPD and SP can be formed under UVA light; again no 6-4PP
[
12]
was detected.
It is therefore concluded that both singlet and
triplet excited states can contribute to CPD and SP formation,
while only singlet excited states are responsible for 6-4PP
4
PP photochemistry. We previously demonstrated that upon
replacement of the 3'-pyrimidine with a toluenyl (To) or meta-
xylenyl moiety (X) CPD analogs are formed as the dominant
photoproducts under UV irradiation. In the formation of SP in
the TpT photoreaction,
abstracts an hydrogen from the methyl group at the 3'-
thymine.
one of the two involving pyrimidine residues being sufficient to
drive the dimerization to completion is likely a general model for
all three types of pyrimidine photoreactions.
[
1d]
formation.
Our calculations demonstrated that photochemical
[
22]
formation of interbase covalent bonds proceeding through the
triplet and singlet routes may differ drastically in timescales,
formation of intermediates and quantum yields. In particular, the
quantum yield of the triplet route might be rather low because
radical recombination before the C6a−O4a bond is formed leads
to the recovery of the initial dinucleotide rather than to 6-4PP
formation.
[
11]
the photo-excited 5'-thymine residue
[
13]
Taken together, that initial photoexcitation of only
The oxetane intermediate is widely accepted as an
intermediate preceding 6-4PP formation when the 3'-residue is a
thymidine. By replacing the 3'-thymidine with a 4-thiothymidine in
Furthermore, our present results have indicated that the 6-
PP formation is triggered by the acceptance of an electron by
4
the “driver” 3'-thymine residue from the “passenger” 5'-thymine
residue. It is known that UV radiation may trigger intrastrand ET
in DNA resulting in a pair of transient cation and anion
[
32]
a dinucleotide context,
a semi-stable thietane species was
generated under 360 nm UVA irradiation, which was under
equilibrium with the corresponding 6-4PP analog. Given the
similar structures between thietane and oxetane, it is reasonable
that the thymidine photoreaction may as well proceed via an
[33]
radicals. The transient radicals are indicated to be responsible
for the formation of intrastrand crosslink products involving a 5-
bromouracil or 5-bromocytosine and
a neighboring purine
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201700045
FULL PAPER
oxetane intermediate to form 6-4PP. However, although oxetane
intermediates via the Paternò-Büchi reaction are relatively
Such a reaction mechanism may be similar to the formation of
SP, where the C5=C6 bond at the excited 5′-thymine abstracts
an H atom from the 3′-thymine residue before the resulting
[
38]
common in organic photochemistry, the lifetime of the oxetane
intermediate generated during the 6-4PP formation is estimated
[11]
thymine allyl radical recombines with the C5 radical. The so-
formed Spore-Int intermediate, likewise Oxetane, does not
accumulate and instead undergoes elimination of a water
molecule to form photoproduct 2. As we did not detect any
stable reaction intermediate during the formation of 2, it is
probable that the Spore-Int intermediate, like Oxetane, is
unstable towards protonation or water-assisted proton transfer.
Therefore, besides 6-4PP formation, our work here presents the
first indication that ET between the two bases may be
responsible for SP-type photochemistry.
[
39]
to be ~ 4 milliseconds in the dinucleotide TpT photoreaction.
Our calculations indicate that oxetane rearrangement to 6-4PP
may involve protonation of the basic O4b position. The
protonation mechanism also offers an explanation for the higher
[
32]
stability of the thietane intermediate,
as the S4b position is
less basic as compared to O4b. In line with our finding, an
equilibrium between thietane and the 6-4PP isomer is known to
[
32]
be pH dependent.
Increasing pH may lead to N3b
deprotonation, and therefore, complete displacement of the
equilibrium toward the 6-4PP isomer, as observed for
Interestingly, a photoproduct with a structure similar to 2 was
also observed in the previous thymidylyl (3'-5')-4-thiothymidine
[
32]
thietane.
Under pH conditions when N3bH is protonated,
[
32]
protonation/deprotonation of S4b results in an equilibrium
photoreaction,
suggesting that the photo-excited 4-
[
32]
between the thietane and 6-4PP isomers. As our calculations
suggest, protonation of O4b is more energetically favorable as
compared to S4b, hence the equilibrium should be completely
displaced to the protonated 6-4PP form in the case of oxetane,
possibly explaining the fact that oxetane is not observed upon
XpT irradiation. Nonetheless, deprotonation of O4b orchestrated
with protonation of N3b by a photolyase enzyme has been
tentatively proposed as a crucial step in the reversal of 6-4PP to
oxetane and then to the intact nucleobases in some
thiothymidine can also initiate the H-atom abstraction process
from the methyl group of the 5'-thymine base. The fact that both
XpT and thymidylyl (3'-5')-4-thiothymidine exhibit the SP-type
photoreaction indicates that such a route may be a general
mechanism in pyrimidine photochemistry. As dinucleotides
generally exhibit similar stacking patterns with those steps in the
duplex DNA, the observed formation of 2 raises an interesting
question: Can this C4=O mediated SP-type photoreaction occur
in a duplex oligonucleotide and in the genomic DNA of solar
radiated cells?
[
40]
organisms.
The repair of CPD/6-4PP by a respective photolyase is
initiated by intermolecular ET resulting in a radical anion specie
[
40-41]
of the photoproduct.
It is thus of interest for us to compare
Conclusions
the ET-mediated 6-4PP formation with the photolyase mediated
pyrimidine-dimer repair. Photolyases adopt a flavin adenine
dinucleotide cofactor that absorbs visible light and in the excited
state pumps an electron into the 5'-residue of the CPD/6-4PP to
convert it to an anion radical to initiate the dimer repair
In summary, we report the synthesis and photoreaction of a
dinucleotide analog XpT, in which a photo-inert xylene residue is
incorporated into the 5'-end as a steric analog of thymine and as
an electron donor. Irradiation of the XpT dinucleotide under 254
nm UV light generates two major photoproducts: a pyrimidine (6-
[
40-42]
process.
Such a trend is inverted as compared to the 6-4PP
formation shown in our XpT system here, in which the transient
anion radical is located at the 3'-residue. Apparently, interbase
bond formation and cleavage in 6-4PP may be triggered by
electron donation to the different bases, 3' and 5' respectively.
The transient 5'-anion radical formed in photolyase then
undergoes the repair possibly by the formation of a transient
4
) pyrimidone (6-4PP) analog and an analog of the so-called
spore photoproduct (SP). Both products are formed via reaction
at the C4=O of the photo-excited T, which, together with our
previous studies on the formation of cyclobutane pyrimidine
dimers (CPDs) and SP, indicates that excitation of a single
“
driver” residue is sufficient to trigger pyrimidine dimerization and
[
42-43]
oxetane intermediate at the enzyme active site
or by direct
this is likely a general model for all three types of pyrimidine
photoreactions.
[
44]
transfer of the O4bH group from the 5'-base to the 3'-base.
Although our conclusions on 6-4PP mechanism are reached
by using a xylene as a thymine analog, the results are certainly
consistent with the 6-4PP photochemistry of natural pyrimidine
residues. The key structures and energy trends controlling this
photochemistry identified in our present computational study for
the XpT dinucleotide are similar to those found previously for the
Our observations that the photoreaction of TpX does not yield
any 6-4PP but that of XpT does indicate that formation of 6-4PP
and SP requires 5′ to 3′ interbase ET (but not 3′ to 5′) as the
critical step for both detected photoreactions. Using quantum-
chemical calculations we demonstrated that indeed 5′ to 3′
interbase ET catalyzes the formation of the interbase covalent
bonds in 6-4PP and SP. The photochemistry of the XpT
dinucleotide analog thus provides the first unambiguous
evidence that the 6-4PP formation is triggered by the
acceptance of an electron by the “driver” 3'-thymine residue from
the “passenger” 5'-thymine residue. Our work also provides the
first indication that interbase ET may be responsible for
triggering the SP-type photochemistry.
[
7, 14, 29-30]
TpT and TpC dinucleotides.
Besides the generation of 6-4PP analog 1, the formation of
SP analog 2 in the XpT photoreaction is also significant to the
understanding of pyrimidine photochemistry triggered by
interbase ET. In this case, the C4b−O4b group of the 3′-thymine
radical anion abstracts a proton form the methyl-group of the 5'-
radical cation after electron donation. Subsequent radical
recombination results in the formation of the C5a−C4b bond.
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201700045
FULL PAPER
Our quantum-chemical calculations imply that opening of the
oxetane ring may be facilitated by protonation of the O4b atom.
Toluene-based thymine analogs like xylene are proved to
exhibit similar base-stacking interactions as a thymine. Further,
our calculations identified structures controlling the XpT
photochemistry that are rather similar to those found in the
previous computational studies of the TpT and TpC
dinucleotides. Hence, even though our conclusion on the role of
Preparation of XpT photoproducts in a large scale. To characterize
the XpT photoproducts 1 and 2, the reaction was repeated at a bigger
scale. Briefly, 200 mg XpT was dissolved in 10 mM pH 7.4 phosphate
buffer to a final concentration of 2 mM. The solution was placed on ice
and the photoreaction was allowed to proceed for 2 hours. The products
were purified by semi-preparative HPLC.
concentrated, and desalted by reinjection into HPLC and using H
1
and 2 were collected,
O and
2
acetonitrile as mobile phases. The reaction afforded 2.4 mg 1 (1.2 %
yield) and 5.1 mg 2 (2.6% yield) as white solids. The desalted products
were then characterized by MS/MS and NMR spectroscopy.
5′ to 3′ ET in pyrimidine dimerization is reached by using a
xylene residue, the result is certainly consistent with the
photochemistry of natural pyrimidine residues.
HPLC analyses. HPLC analyses were performed at room temperature
using
UV/Visible detector and a Waters XBridge OST C18 column (2.5 µm,
.6×50 mm). The reaction was monitored at 230 nm and 320 nm
a Waters (Milford, MA) breeze HPLC system with a 2489
4
Experimental and Computational Section
simultaneously. The HPLC column was first equilibrated with 99% mobile
phase A (10 mM ammonium acetate aqueous solution, pH 7.0) and 1%
of mobile phase B (acetonitrile). Compounds were eluted with an
ascending gradient (1% to 20% in 30 min) of mobile phase B at a flow
rate of 1 mL/min. Under this gradient, 1 was eluted at 8.3 min, 2 at 14.2
min, and XpT at 23.9 min.
Materials and general methods. All solvents and chemicals were of
analytical grade and purchased from Sigma-Aldrich, Fisher, or VWR and
used without further purification. The 254 nm UV irradiation was provided
by a Spectroline germicidal UV sterilizing lamp XX-15G, which generates
unfiltered UV light ranging from 180 nm to 280 nm and peaking at 254
nm. The 302 nm UV irradiation is provided by using a UVP XX-15M lamp,
which generates unfiltered UV light ranging from 280 nm to 360 nm and
peaking at 302 nm. NMR spectra were obtained via a Bruker 500 MHz
Fourier transform NMR spectrometer. Mass spectrometric (MS) analyses
were conducted via electrospray ionization (ESI) employing an Agilent
Product purification by semi-preparative HPLC was also performed at
room temperature with the same Waters HPLC system and a Waters
TM
XBridge OST C18 column (2.5 µm particle size, 10 × 50 mm). 10 mM
2
pH 7.0 ammonium acetate in H O was used as mobile phase A and
acetonitrile was used as mobile phase B. The column was equilibrated
with 99% mobile phase A and 1% of mobile phase B. Compounds were
eluted with an ascending gradient (1 - 5% in 20 min followed by 5 - 20%
in another 20 min) of mobile phase B at a flow rate of 4.73 mL/min.
Under this gradient, 1 was eluted at 12.1 min, 2 at 26.0 min, and XpT at
1100-6130A LC-MS with an ion-trap mass analyzer. High resolution MS
and tandem mass spectrometry (MS/MS) analyses were performed using
an Agilent 1200-6520 capillary LC-Q-TOF MS spectrometer. The MS
data were acquired via the “Agilent MassHunter Workstation Data
Acquisition (B.03.00)” software and analyzed via “Qualitative Analysis of
MassHunter Acquisition Data (B.03.00)” software.
34.0 min.
Product analyses via tandem mass spectrometry (MS/MS). MS/MS
analyses of the reaction products were conducted using an Agilent 6520
Accurate Mass Q-TOF LC/MS spectrometer with an Agilent ZORBAX
Bonus-RP column (5 µm particle size, 250 × 4.6 mm i.d.). The column
was equilibrated in mobile phase A (5 mM ammonium acetate in 99%
water and 1% acetonitrile, pH 6.5), while compounds were eluted with an
ascending gradient (0 - 15% in 30 min followed by 15 - 35% in 20 min) of
mobile phase B (5 mM ammonium acetate in 50% methanol and 50%
acetonitrile) at a flow rate of 0.5 mL/min. Under this gradient, 1 was
Synthesis of XpT. The thymidine isostere (X) containing a meta-xylene
moiety was synthesized via a procedure reported previously by our
[
22]
group.
phosphodiester moiety, obtaining the dinucleotide analog XpT (Scheme
). Detailed synthesis and characterization of XpT can be found in the
We then linked X to a thymidine residue through a
2
supporting information.
XpT photoreaction under 254 nm UV light.
XpT was dissolved in
200 µL 10 mM pH 7.4 phosphate buffer to a final concentration of 2 mM.
eluted at 24.1 min,
2 at 37.2 min, and XpT at 42.0 min. Mass
The resulting solution was then irradiated under unfiltered 254 nm UVC
light on ice with the samples ~ 5 cm away from the lamp. At various times,
spectrometry was conducted under both positive and negative ion modes.
1
0 µL of the solution was aliquoted out for later HPLC analysis. Two
Ground-
and
excited-state
calculations.
Quantum-chemical
major products 1 and 2 were observed when monitored at 260 and 320
nm by a UV director on HPLC.
calculations were performed for XpT and TpT dinucleotides with
deprotonated phosphate (total charge −1). Geometry optimization was
performed with density-functional theory methods with the B3LYP
[
46]
[47]
XpT photoreaction under 302 nm UV light. The XpT was repeated
under unfiltered 302 nm UVB light under identical conditions and the
reaction products analyzed by HPLC. No products were observed.
functional supplemented by the D3 correction for dispersion energy.
In the ground state and lowest-energy triplet state, geometries of the
stationary points (minima and saddle points) were first optimized and
then confirmed by the normal vibrational mode analysis in harmonic
approximation. From each computed saddle point, intrinsic reaction
coordinates (irc) were computed in both directions following the
imaginary-frequency mode. The irc calculations for the formation of 1 and
2 converged to similar structures of the ground-state and triplet-state
reactant corresponding to a stacked conformation of XpT. In the first
excited singlet state, geometry optimization was performed with the TD-
B3LYP-D3 method, starting from the ground-state stacked conformation
of XpT. The relaxed energy scans for the C5a-O4b and Ha-O4b bond
formation were computed in the first excited state corresponding to the
charge-separated radical-pair state with the TD-B3LYP-D3 method.
Quantum yield determination. The quantum yields for products 1 and 2
were determined using a modified literature method.
[
45]
Briefly, 2 ml of
aqueous solution containing equal amount of dinucleotides TpT and XpT
was placed in a small beaker and stirred continuously under the 254 nm
UV light. After 10 minutes, the photoreactions were stopped and the
products formed in these two reactions analysed by HPLC. Under the
similar reaction conditions, the quantum yield of 6-4PP TpT was
determined as (2.0 ± 0.7) × 10 . The quantum yields of 1 and 2 were
determined as (3.2 ± 1.0) × 10 and (3.3 ± 1.0) × 10 , respectively by
comparing with that of 6-4PP TpT.
[
23]
-3 [45]
-4
-4
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201700045
FULL PAPER
[
5] a) D. E. Brash, Photochem. Photobiol. 2015, 91, 15-26; b) E. D. Pleasance,
Single-point excited-state calculations were performed with the
multireference multiconfigurational second-order perturbation theory
R. K. Cheetham, P. J. Stephens, D. J. McBride, S. J. Humphray, C. D.
Greenman, I. Varela, M.-L. Lin, G. R. Ordonez, G. R. Bignell, K. Ye, J. Alipaz,
M. J. Bauer, D. Beare, A. Butler, R. J. Carter, L. Chen, A. J. Cox, S. Edkins, P.
I. Kokko-Gonzales, N. A. Gormley, R. J. Grocock, C. D. Haudenschild, M. M.
Hims, T. James, M. Jia, Z. Kingsbury, C. Leroy, J. Marshall, A. Menzies, L. J.
Mudie, Z. Ning, T. Royce, O. B. Schulz-Trieglaff, A. Spiridou, L. A. Stebbings,
L. Szajkowski, J. Teague, D. Williamson, L. Chin, M. T. Ross, P. J. Campbell,
D. R. Bentley, P. A. Futreal and M. R. Stratton, Nature 2010, 463, 191-196.
[6] a) H. E. Johns, M. Delbruck and S. A. Rapaport, J. Mol. Biol. 1962, 4, 104-
114; b) A. J. Varghese and S. Y. Wang, Science 1967, 156, 955-957; c) J. E.
Donnellan, Jr. and R. B. Setlow, Science 1965, 149, 308-310.
[
48]
methods XMCQDPT2
with the zeroth-order CASSCF wave function.
The XMCQDPT2 calculations were combined with the equilibrium
[
49]
dielectric polarizable continuum model (DPCM) of water. The CASSCF
active space was composed from the frontier molecular orbitals (MOs)
localized on xylene and thymine rings (Supporting Information Figures S8
and S9). CASSCF energy averaging over all included states (singlet
states or both singlet and triplet states) with equal weights was
performed. The XMCQDPT2 calculations were carried out with intruder-
state avoidance energy shift (edshft) of 0.02 hartree. Optimization of the
conical intersections geometries was carried out with the SA2-
CASSCF(12,12) method (12 electrons in 12 MOs with energy-averaging
[
7] A. Banyasz, T. Douki, R. Improta, T. Gustavsson, D. Onidas, I. Vaya, M.
Perron and D. Markovitsi, J. Am. Chem. Soc. 2012, 134, 14834-14845.
[8] W. J. Schreier, P. Gilch and W. Zinth, Annu. Rev. Phys. Chem. 2015, 66,
497-519.
[9] a) A. Banyasz, I. Vaya, P. Changenet-Barret, T. Gustavsson, T. Douki and
D. Markovitsi, J. Am. Chem. Soc. 2011, 133, 5163-5165; b) S. Mouret, C.
Baudouin, M. Charveron, A. Favier, J. Cadet and T. Douki, Proc. Natl. Acad.
Sci. 2006, 103, 13765-13770.
[10] S. Premi, S. Wallisch, C. M. Mano, A. B. Weiner, A. Bacchiocchi, K.
Wakamatsu, E. J. Bechara, R. Halaban, T. Douki and D. E. Brash, Science
2
[
50]
over 2 states).
The Mulliken atomic charges derived from the zeroth-
order QDPT electron densities were used to determine the net charges
on the bases. In the XpT and TpT dinucleotides, electronic coupling and
maximum ET rate constants were evaluated using the generalized
[
27]
[28]
Mulliken-Hush method and semi-classical Marcus theory.
015, 347, 842-847.
[11] a) D. M. Ames, G. Lin, Y. Jian, J. Cadet and L. Li, J. Org. Chem. 2014, 79,
Thymine and xylene methylated at position N1/C1 were used in
calculations of the redox energies. Geometry optimization and energy
calculations were performed with the (U)B3LYP-D3-PCM method.
Electron affinities and ionization potentials were computed as adiabatic
energy differences between the radical ion and neutral closed-shell (CS)
or triplet species. A positive electron affinity corresponds to the radical
anion being lower in energy than the neutral CS or triplet species. A
positive ionization potential corresponds to the radical cation being higher
in energy than the neutral CS or triplet species.
4843-4851; b) G. Lin and L. Li, Angew. Chem. Int. Ed. 2010, 49, 9926-9929; c)
E. C. Hayes, Y. Jian, L. Li and S. Stoll, J. Phys. Chem. B 2016, 120, 10923-
10931; d) Y. Jian, D. M. Ames, H. Ouyang and L. Li, Org. Lett. 2015, 17, 824-
827.
[
12] A. Moysan, A. Viari, P. Vigny, L. Voituriez, J. Cadet, E. Moustacchi and E.
Sage, Biochemistry 1991, 30, 7080-7088.
13] Q. Du, H. Zhao, D. Song, K. Liu and H. Su, J. Phys. Chem. B 2012, 116,
1117-11123.
14] A. Giussani, L. Serrano-Andrés, M. Merchán, D. Roca-Sanjuán and M.
[
1
[
Garavelli, J. Phys. Chem. B 2013, 117, 1999-2004.
[15] a) R. Improta, J. Phys. Chem. B 2012, 116, 14261-14274; b) L. Esposito,
A. Banyasz, T. Douki, M. Perron, D. Markovitsi and R. Improta, J. Am. Chem.
Soc. 2014, 136, 10838-10841; c) A. Banyasz, L. Esposito, T. Douki, M. Perron,
C. Lepori, R. Improta and D. Markovitsi, J. Phys. Chem. B 2016, 120, 4232-
4242; d) D. Markovitsi, Photochem. Photobiol. 2016, 92, 45-51.
The standard 6-31G** basis set was used throughout. All calculations
[
51]
were performed with Firefly 8.1 software
the US GAMESS code.
which is partially based on
[
52]
[
16] a) W. J. Schreier, T. E. Schrader, F. O. Koller, P. Gilch, C. E. Crespo-
Hernandez, V. N. Swaminathan, T. Carell, W. Zinth and B. Kohler, Science
007, 315, 625-629; b) T. Ostrowski, J.-C. Maurizot, M.-T. Adeline, J.-L.
2
Fourrey and P. Clivio, J. Org. Chem. 2003, 68, 6502-6510; c) C. Moriou, M.
Thomas, M.-T. Adeline, M.-T. Martin, A. Chiaroni, S. Pochet, J.-L. Fourrey, A.
Favre and P. Clivio, J. Org. Chem. 2007, 72, 43-50.
Acknowledgements
[17] a) B. A. Schweitzer and E. T. Kool, J. Am. Chem. Soc. 1995, 117, 1863-
1
2
872; b) K. M. Guckian and E. T. Kool, Angew. Chem. Int. Ed. 1997, 36, 2825-
828; c) E. T. Kool and H. O. Sintim, Chem. Commun. 2006, 3665-3675.
The authors thank the National Science Foundation (CHE
454184 to L.L) and the Minerva Program of the Max-Planck
1
[18] a) S. Moran, R. X.-F. Ren and E. T. Kool, Proc. Natl. Acad. Sci. 1997, 94,
10506-10511; b) S. Moran, R. X. F. Ren, S. Rumney and E. T. Kool, J. Am.
Society (to T.D.) for financial support. The NMR and MS facilities
at IUPUI are supported by National Science Foundation MRI
grants CHE-0619254 and DBI-0821661, respectively. T.D. and
E.M. are very grateful to Chris Roome (MPImF Heidelberg) for
excellent support of high performance computing.
Chem. Soc. 1997, 119, 2056-2057; c) E. T. Kool, Annu. Rev. Biochem. 2002,
71, 191-219.
[19] a) K. M. Guckian, T. R. Krugh and E. T. Kool, Nat. Struct. Biol. 1998, 5,
954-959; b) K. M. Guckian, T. R. Krugh and E. T. Kool, J. Am. Chem. Soc.
2000, 122, 6841-6847.
[
[
[
20] P. S. Pallan and M. Egli, J. Am. Chem. Soc. 2009, 131, 12548-12549.
21] P. R. Callis, Annu. Rev. Phys. Chem. 1983, 34, 329-357.
22] a) D. Liu and L. Li, RSC Adv. 2013, 3, 19545-19550; b) D. Liu, Y. Zhou, J.
Pu and L. Li, Chem. Eur. J. 2012, 18, 7823-7833.
Keywords: thymine dimer • DNA photochemistry • electron
transfer • pyrimidine (6-4) pyrimidone photoproduct • spore
photoproduct
[23] See supporting information.
[24] S. Iwai, M. Shimizu, H. Kamiya and E. Ohtsuka, J. Am. Chem. Soc. 1996,
118, 7642-7643.
[
[
25] G. Lin, Y. Jian, H. Ouyang and L. Li, Org. Lett. 2014, 16, 5076-5079.
26] a) J.-S. Taylor, D. S. Garrett and M. J. Wang, Biopolymers 1988, 27,
[1] a) J. Cadet and P. Vigny in Photochemistry and the nucleic acids, Vol. 1
1571-1593; b) L. S. Kan, L. Voituriez and J. Cadet, Biochemistry 1988, 27,
5796-5803.
(Ed. H. Morrison), Wiley, New York, 1990, pp. 1-272; b) J. Cadet, E. Sage and
T. Douki, Mutat. Res-Fund. Mol. M. 2005, 571, 3-17; c) J. Cadet, A. Grand and
T. Douki in Solar UV radiation-induced DNA bipyrimidine photoproducts:
formation and mechanistic insights, Vol. 356 Eds.: M. Barbatti, A. C. Borin and
S. Ullrich), Springer International Publishing, 2015, pp. 249-275; d) J. Cadet, S.
Mouret, J. L. Ravanat and T. Douki, Photochem. Photobiol. 2012, 88, 1048-
[27] C.-P. Hsu, Acc. Chem. Res. 2009, 42, 509-518.
[28] R. A. Marcus and N. Sutin, BBA Rev. Bioenergetics 1985, 811, 265-322.
[29] a) Z. b. Yang, R. b. Zhang and L. A. Eriksson, Phys. Chem. Chem. Phys.
2011, 13, 8961-8966; b) Z. b. Yang, L. A. Eriksson and R. b. Zhang, J. Phys.
Chem. B 2011, 115, 9681-9686.
[30] L. Blancafort and A. Migani, J. Am. Chem. Soc. 2007, 129, 14540-14541.
[31] V. Labet, N. Jorge, C. Morell, T. Douki, A. Grand, J. Cadet and L. A.
Eriksson, Photochem. Photobiol. Sci. 2013, 12, 1509-1516.
[32] P. Clivio, J. L. Fourrey, J. Gasche and A. Favre, J. Am. Chem. Soc. 1991,
113, 5481-5483.
[33] a) Y. Zhang, K. de La Harpe, A. A. Beckstead, L. Martínez-Fernández, R.
Improta and B. Kohler, J. Phys. Chem. Lett. 2016, 7, 950-954; b) Y. Zhang, J.
Dood, A. A. Beckstead, X.-B. Li, K. V. Nguyen, C. J. Burrows, R. Improta and
B. Kohler, Proc. Natl. Acad. Sci. 2014, 111, 11612-11617; c) D. B. Bucher, B.
1065; e) J. Cadet, S. Courdavault, J. L. Ravanat and T. Douki, Pure Appl.
Chem. 2005, 77, 947-961.
[2] R. E. Rycyna and J. L. Alderfer, Nucleic Acids Res. 1985, 13, 5949-5963.
[3] J. S. Taylor and M. P. Cohrs, J. Am. Chem. Soc. 1987, 109, 2834-2835.
[4] a) C. l. Desnous, D. Guillaume and P. Clivio, Chem. Rev. 2010, 110, 1213-
1
232; b) P. Setlow, Trends Microbiol. 2007, 15, 172-180; c) P. Setlow, J. Appl.
Microbiol. 2006, 101, 514-525; d) P. Setlow and L. Li, Photochem. Photobiol.
015, 91, 1263-1290.
2
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201700045
FULL PAPER
M. Pilles, T. Carell and W. Zinth, Proc. Natl. Acad. Sci. 2014, 111, 4369-4374;
d) Y. Zhang, K. de La Harpe, A. A. Beckstead, R. Improta and B. Kohler, J.
Am. Chem. Soc. 2015, 137, 7059-7062; e) Y. Zhang, X. B. Li, A. M. Fleming, J.
Dood, A. A. Beckstead, A. M. Orendt, C. J. Burrows and B. Kohler, J. Am.
Chem. Soc. 2016, 138, 7395-7401.
[44] a) T. Domratcheva and I. Schlichting, J. Am. Chem. Soc. 2009, 131,
17793-17799; b) T. Domratcheva, J. Am. Chem. Soc. 2011, 133, 18172-18182.
[45] Y. K. Law, J. Azadi, C. E. Crespo-Hernández, E. Olmon and B. Kohler,
Biophys. J. 2008, 94, 3590-3600.
[46] A. D. Becke, J. Chem. Phys. 1993, 98, 5648-5652.
[
34] a) Y. Zeng and Y. Wang, Nucleic Acids Res. 2006, 34, 6521-6529; b) H.
Hong and Y. Wang, J. Am. Chem. Soc. 2005, 127, 13969-13977.
35] B. M. Pilles, D. B. Bucher, L. Liu, P. Gilch, W. Zinth and W. J. Schreier,
Chem. Commun. 2014, 50, 15623-15626.
36] T. Matsunaga, K. Hieda and O. Nikaido, Photochem. Photobiol. 1991, 54,
03-410.
37] T. Douki, A. Reynaud-Angelin, J. Cadet and E. Sage, Biochemistry 2003,
2, 9221-9226.
[47] S. Grimme, J. Antony, S. Ehrlich and H. Krieg, J. Chem. Phys. 2010, 132,
154104.
[48] A. A. Granovsky, J. Chem. Phys. 2011, 134, 214113.
[49] a) R. Cammi and J. Tomasi, J. Comput. Chem. 1995, 16, 1449-1458; b) B.
Mennucci and J. Tomasi, J. Chem. Phys. 1997, 106, 5151-5158.
[50] A. A. Granovsky, J. Chem. Phys. 2015, 143, 231101.
[
[
4
[
4
[51]
Alex
A.
Granovsky,
Firefly
version
8,
www
[
[
[
[
[
38] M. Auria and R. Racioppi, Molecules 2013, 18, 11384.
39] S. Marguet and D. Markovitsi, J. Am. Chem. Soc. 2005, 127, 5780-5781.
40] A. Sancar, Chem. Rev. 2003, 103, 2203-2238.
41] A. Sancar, J. Biol. Chem. 2008, 283, 32153-32157.
42] Y. Wang, P. P. Gaspar and J.-S. Taylor, J. Am. Chem. Soc. 2000, 122,
http://classic.chem.msu.su/gran/firefly/index.html, access date March 2017.
[52] 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 and J. A. Montgomery, J. Comput. Chem. 1993, 14, 1347-1363.
5510-5519.
[43] a) X. Zhao, J. Liu, D. S. Hsu, S. Zhao, J.-S. Taylor and A. Sancar, J. Biol.
Chem. 1997, 272, 32580-32590; b) S. T. Kim, K. Malhotra, C. A. Smith, J. S.
Taylor and A. Sancar, J. Biol. Chem. 1994, 269, 8535-8540.
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201700045
FULL PAPER
Entry for the Table of Contents (Please choose one layout)
Layout 2:
FULL PAPER
†
[a, b]
† [c]
Yajun Jian,
Egle Maximowitsch,
[
a]
[a]
Degang Liu, Surya Adhikari, Lei Li,*
[
a,d]
[c]
and Tatiana Domratcheva*
Indications of 5′ to 3′ Interbase
Electron Transfer as the First Step of
Pyrimidine Dimer Formation Probed
by a Dinucleotide Analog
Excited thymine
O
e
O
HN
−
N
OH
N
2
54 nm
+
HN
N
O
H
N
+
N
O
O
N
O
O
O
O
O
O
O
O
O
5'
OH
OH
3'
5'
OH
O
P
O−
O
O
O
O
5'
OH
OH
3'
5
'
OH
3'
O
P
O−
O
O
P
O−
OH
3'
O
OH
O
P
O−
O
O
XpT
Radical cation / anion
6-4PP analog
SP
analog
This article is protected by copyright. All rights reserved.