Expanding the horizon of the thymine Isostere biochemistry: Unique cyclobutane dimers formed by photoreaction between a thymine and a toluene residue in the dinucleotide framework

Article abstract of DOI:10.1002/chem.201200816

Substituted toluenyl groups are considered as close isosteres of the thymine residue. They can be recognized by DNA polymerases as if they were thymine. These toluene derivatives are generally inert toward radical additions, including the [2+2] photocycloadditions, due to the stable structure of the aromatic ring and are usually used as solvents for radical reactions. Surprisingly, after incorporating toluene into the dinucleotide framework, we found that the UV excited thymine residue readily dimerizes with the toluenyl moiety through a [2+2] photoaddition reaction. Furthermore, the reaction site on the toluenyl moiety is not the C5=C6 bond, as commonly observed in cyclobutane pyrimidine dimers, but the C4=C5 or C3=C4 instead. Such a reaction pattern suggests that in the stacked structure, it is one of these bonds, not the C5=C6, that is close to the thymine C5=C6 bond. A similar structural feature is found in DNA duplex with a thymine replaced by a 2,4-difluorotoluene. Our results argue that although the substituted toluenyl moieties closely mimic the size and shape of the thymine residue, their more hydrophobic nature determines that they stack on DNA bases differently from the natural thymine residue and likely cause local conformational changes in duplex DNA.

Full text of DOI:10.1002/chem.201200816

FULL PAPER
DOI: 10.1002/chem.201200816
Expanding the Horizon of the Thymine Isostere Biochemistry: Unique
Cyclobutane Dimers Formed by Photoreaction between a Thymine and a
Toluene Residue in the Dinucleotide Framework
Degang Liu,[a] Yan Zhou,[a] Jingzhi Pu,*[a] and Lei Li*[a, b]
Abstract: Substituted toluenyl groups
are considered as close isosteres of the
thymine residue. They can be recog-
nized by DNA polymerases as if they
were thymine. These toluene deriva-
tives are generally inert toward radical
additions, including the [2+2] photo-
cycloadditions, due to the stable struc-
ture of the aromatic ring and are usual-
ly used as solvents for radical reactions.
Surprisingly, after incorporating tolu-
ene into the dinucleotide framework,
we found that the UV excited thymine
residue readily dimerizes with the tol-
uenyl moiety through a [2+2] photo-
addition reaction. Furthermore, the re-
action site on the toluenyl moiety is
not the C5=C6 bond, as commonly ob-
served in cyclobutane pyrimidine
dimers, but the C4=C5 or C3=C4 in-
stead. Such a reaction pattern suggests
that in the stacked structure, it is one
of these bonds, not the C5=C6, that is
close to the thymine C5=C6 bond. A
similar structural feature is found in
DNA duplex with a thymine replaced
by a 2,4-difluorotoluene. Our results
argue that although the substituted tol-
uenyl moieties closely mimic the size
and shape of the thymine residue, their
more hydrophobic nature determines
that they stack on DNA bases differ-
ently from the natural thymine residue
and likely cause local conformational
changes in duplex DNA.
Keywords: cyclobutane pyrimidine
dimers · DNA · photochemistry ·
stereochemistry · thymine
Introduction
Watson–Crick hydrogen bonding formed between nucleo-
bases AT or GC is widely regarded as one of the basic inter-
actions in life. It enables the formation of the double-strand-
ed oligonucleotide and becomes the foundation for genetic
information storage, direct replication, transcription and
translation. Kool et al. replaced thymine with a non-polar
2,4-difluorotoluene residue (DFTo, Scheme 1), which pos-
sesses a nearly identical size and shape to thymine,[1] and
found that DNA polymerases readily incorporate it into
DNA as if it were thymine.[1c,2] This result argues that DNA
polymerases are governed largely by the steric effects, not
the hydrogen bonds between the corresponding nucleobases
in making DNA copies.[2c,3] Two recent X-ray crystallograph-
ic studies, however, indicate that DFTo still forms hydrogen
bonds with adenine,[4] even though the bonding interactions
Scheme 1. Thymidine and its steric analogues.
might not be as strong as those between the natural A and
T.
[a] Dr. D. Liu, Dr. Y. Zhou, Prof. Dr. J. Pu, Prof. Dr. L. Li
Department of Chemistry and Chemical Biology
Indiana University–Purdue University Indianapolis (IUPUI)
402 N. Blackford St., Indianapolis, IN 46202 (USA)
Fax : (+1)317-274-4701
Thymine (T) is known for its rich photochemical reactivi-
ty and is the most UV sensitive nucleobase.[5] In B-form
DNA, thymine dimerizes with a neighboring T; this results
in the cyclobutane pyrimidine dimer (CPD, or T< >T) as
the major photoproduct.[5] In contrast, toluene is suggested
to possess different molecular orbital diagrams from thy-
mine and is rather inert under UV irradiation.[5,6] This is
why it is commonly used as the solvent for radical reactions,
including many [2+2] photo-cycloadditions.[7] T< >T is
considered to form through a [2+2] photoaddition at the
singlet excited state of thymine,[8] as implied by its ultrafast
E-mail: lilei@iupui.edu
jpu@iupui.edu
[b] Prof. Dr. L. Li
Department of Biochemistry and Molecular Biology
Indiana University School of Medicine
635 Barnhill Drive, Indianapolis, Indiana 46202 (USA)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201200816.
Chem. Eur. J. 2012, 18, 7823 – 7833
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7823

formation rate.[9] Our previous research showed that the ex-
cited 5’-T behaves like a di-radical and conducts the H atom
abstraction during the spore photoproduct (SP) forma-
tion.[10] We thus wondered if the excited thymine can induce
the [2+2] cycloaddition to a rather inert toluenyl moiety in-
corporated into the DNA framework.
Results and Discussion
After radiating 1 (1 mm) under 254 nm UV light in a frozen
1:1 glycerol/water glass in liquid N2 for 15 min, analysis of
the reaction by HPLC revealed formation of two products 2
and 3 (Figure 1a). The reaction yields were found to be
7.8% for 2 and 4.6% for 3. ESI-MS analysis displayed
a [M+H]+ signal of 513.2 for both 2 and 3, which is the
same as that of 1. Such a result is typically observed in the
Toluene is approximately 0.15 ꢁ smaller than thymine.
Like DFTo, it is recognized by DNA polymerase as a thy-
mine. The transcription efficiency, however, decreases by ap-
proximately 2–3 orders of magnitude due to its smaller
size.[11] Compared with DFTo, the synthesis of 2-deoxy-1-(3-
toluenyl)-b-d-ribofuranose (To, Scheme 1) is much easier.[12]
We thus chose To instead of DFTo for our proof-of-concept
photochemical studies. Dinucleotide TpT mimics the photo-
chemistry occurring in duplex DNA due to the similar thy-
mine stacking interaction;[5,13] this results in the cis–syn T<
>T as the dominating photoproduct. As the toluenyl moiety
exhibits an even stronger stacking interaction with DNA
bases than thymine due to the enhanced desolvation
effect,[14] the dinucleotide analogue TpTo (1; Scheme 2) is
likely to adopt a stacked structure in aqueous solution. We
wondered if such a TpTo compound mimics the rich TpT
photochemistry.
Figure 1. a) The HPLC chromatograph of the CPD analogues 2 and 3
formed from the irradiation of TpTo (1) at 77 K. b) The HPLC chroma-
tograph of the CPD analogue 5 formed from the TpMeTo (4) photoreac-
tion at 77 K (* denotes an impurity in 1). No other product with the
same mass was detected above the basal level in either reaction.
thymine dimerization reac-
tion,[5] and suggests that a simi-
lar reaction might have oc-
curred in the irradiation of 1.
Both 2 and 3 were also isolated
in the TpTo photoreaction in
H2O at 298 K (Figure S1 in the
Supporting
Information).[12b]
The yields were 3.2 and 3.3%
for 2 and 3, respectively. These
lower yields are likely due to
the enhanced thermal motion
at ambient temperature, which
results in fewer TpTo molecules
with the stacked reactive struc-
ture.
Compared with the TpT pho-
toreaction, where the C5=C6 is
the reaction site for T< >T
formation, all six C=C bonds of
the toluenyl group can in
theory react with the C5=C6 of
T should the right conformation
be adopted; this results in six
possible ortho T< >To re-
gioisomers
Among these products, the
symmetric structure of the
(Scheme 2).[6b]
Scheme 2. Photoreaction of 1 could yield six possible ortho CPD regioisomers (note: all these CPDs are drawn
as the cis–syn isomers. The cis–syn conformations for products 2 and 3 are later confirmed by NOE spectrosco-
py, as discussed in the main text).
7824
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 7823 – 7833

Thymine Isostere Biochemistry
FULL PAPER
phenyl ring determines that reflected by the C1
G
two carbons associated with the C=C bonds, which are
within the three-bond distance to the CH3–To to enable
a strong 1H–13C coupling interaction. Taken together, our
data strongly suggest that 2c is the TpTo photo-dimerization
product 2.
six possible isomers can be divided into three groups. It is
logical to assume that only one of the three reactions in
Scheme 2 occurred in irradiation of 1. The fact that only two
products, 2 and 3, were isolated further supports this as-
sumption. Which reaction has occurred here?
1
Two methyl signals were observed in the H NMR spectra
1
The H NMR spectrum of 2 exhibited two methyl singlet
of 3. The signal at 1.60 ppm was assigned to CH3–T and that
at 1.73 ppm to the CH3–To. Such assignments were con-
firmed by the NOSEY spectrum as the CH3–T showed cou-
pling interaction with the two H atoms on the cyclobutane
ring, whereas the CH3–To did not exhibit any coupling to
signals at 1.35 and 1.49 ppm; these correspond to CH3 To
À
and CH3 T, respectively. These two signals were distinguish-
1
ed as the later signal exhibits a long-range H–13C coupling
interaction with the thymine C4=O carbon in HMBC NMR
spectroscopy.[12b] Three 1H NMR signals were observed in
the low-field region; this suggests that three protons are as-
sociated with the double bonds. These signals include one
singlet (s, 5.55 ppm), one doublet (d, 5.59 ppm) and one dd
signal (5.36 ppm; Figure 2a). Among the six possible struc-
tures, only 3b and 2c support such a coupling pattern. In ad-
dition, H6–T exhibits a singlet peak at 3.81 ppm, whereas
H4–To (2c) or H2–To (3b) exhibit a doublet signal at
2.58 ppm. The dd signal at 5.41 ppm was assigned to the H1’
on the thymine 2-deoxyribose.
1
the ring protons. Two H NMR signals were obtained in the
low magnetic field (Figure 2b), a doublet at 5.76 ppm and
a singlet at 5.52 ppm; this suggests that two protons are as-
sociated with the C=C bonds. Among the six possible prod-
ucts in Scheme 2, only 3a and 3c are considered possible for
such an NMR pattern. The 3a structure suggests that a dd
signal should be observed for the H2–To, however, the
signal observed at 3.66 ppm exhibited a multiple splitting
pattern, which can only be caused by the H3–To in 3c. Such
a structural assignment is further supported by the COSY
spectrum.[12b] The doublet at
3.10 ppm was assigned to H4–
To as it exhibited a strong cou-
pling interaction with both
methyl groups in the NOSEY
spectrum.[12b] The doublet at
4.54 ppm was ascribed to the
H6–T. The dd signal at
5.80 ppm was assigned to the
H1’ on the 2-deoxyribose of the
thymine residue. Compared
with those in 2, some of the
1H NMR signals of 3 shifted
toward the low magnetic field;
these shifts are likely due to the
change of the NMR solvent.
The fact that 2 and 3 corre-
spond to products 2c and 3c,
respectively, suggests that it is
the reaction C, not A, which is
commonly observed in pyrimi-
dine dimerization reaction, that
occurred in the TpTo photore-
action.
Such CPD structures are very
rare in pyrimidine photoreac-
tions. To confirm the structural
assignments above, we intro-
duced a methyl group at the C2
Figure 2. Zoom-in view of the 1H NMR signals associated with the protons on the To or MeTo ring as well as
the protons on the cyclobutane ring in the resulting CPD photoproducts. a) Compound 2 in [D6]DMSO; b) 3
in D2O; c) 5 in D2O. The NMR spectroscopy solvents were chosen to achieve the optimal signal resolution, es-
pecially for the protons associated with the unsaturated region on the To ring. Full NMR spectra for these
CPD species can be found in the Supporting Information.
To distinguish which compound is the true product, we
further examined the HMBC NMR spectrum of 2. Three
strong 1H–13C couplings were observed within the three-
bond distance between the CH3–To protons and the three
neighboring carbons, respectively. Among these three car-
bons, one is in the unsaturated region and two in the satura-
ted.[12b] This can only be fulfilled by 2c as 3b contains only
position of the toluenyl moiety, to yield a dinucleotide ana-
logue TpMeTo (4) and studied its photoreaction. MeTo
(Scheme 1) was chosen for multiple purposes: 1) to decrease
the number of H atoms on the toluenyl ring to simplify the
NMR spectroscopy analysis of the resulting T< >MeTo
complex. 2) To test if the special CPD configuration identi-
fied above is induced by the unique interaction between the
Chem. Eur. J. 2012, 18, 7823 – 7833
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7825

L. Li, J. Pu et al.
T and the To rings. The new methyl moiety introduced lo-
cates at the periphery of the phenyl ring and is expected to
have little effect on the ring–ring interaction. 3) To simplify
the reaction pattern. As the 2-deoxyribose is connected to
the C1 position of the To ring, a methyl group at C2 will not
affect the formation of 5, the analogue of 2 (Scheme 3),
tional theory. These products were optimized by Gaussi-
ACTHNUTRGNEUNG
an 03[16] at the B3LYP/6-31+G(d,p) level[17] by using
a TpDFTo conformation previously determined in duplex
DNA at an intrahelical position[18] as the initial structural
template. Additional models based on an extrahelical con-
formation found in a ternary complex of DNA–photolyase
that contains a flipped CPD-analogue[19] were also construct-
ed. Optimizations initiated from these models yield almost
identical geometries for each compound except the orienta-
tion of 5’-OH on the thymine 2-deoxyribose due to the dif-
ferent initial orientation of the phosphate group in the two
starting structures. Thus, our discussion below focuses on
the results of the TpDFTo based models.
The computation shows that at the B3LYP/6-31+GACHTUNGTRENNUNG(d,p)
level, 5 is 5.47 kcalmolÀ1 more stable than 6 in the gas
phase. The cyclobutane ring in 5 is very flat with the dihe-
À
À
À
dral angle C5 C6 C5 C4 being 1.78; this results in a config-
uration that allows the Tring and the MeTo ring to overlap
very well (Figure 3a). In contrast, the corresponding dihe-
Scheme 3. Photoreaction of 4 only produces CPD analogue 5. Formation
of 6 is inhibited due to the steric hindrance caused by the 2-CH3 moiety.
should reaction c in Scheme 2 occur in the irradiation of 4.
However, steric hindrance between the methyl group and
the 2-deoxyribose should prevent the formation of 6, the an-
alogue of 3 (Scheme 3). Thus, only 5 is expected in the pho-
toreaction of 4. Again, the synthesis of MeTo is straightfor-
ward, which makes the preparation of 4 much easier.
Indeed, irradiation of 4 generated only one new product
with a yield of 16.5% (Figure 1b). NMR spectroscopy con-
firmed that the product adopts the same structure as 5. As
Figure 3. The stacking conformation of the T and MeTo rings in: a) 5,
and b) 6 the structures of which were computationally optimized at the
B3LYP/6-31+GACTHNUGRTNEUNG(d,p) level by using TpDFTo conformation previously de-
termined in duplex DNA as the initial template. The poor stacking inter-
action in 6 is likely due to the steric hindrance of the 2-CH3 moiety on
the toluenyl ring.
1
shown in Figure 2, two H signals were revealed at the low
magnetic field in the NMR spectrum, a doublet at 5.21 ppm
and a singlet at 5.60 ppm; these were assigned to the H3
and H6 atoms of the To group, respectively. The singlet at
4.10 ppm was assigned to the H6–T and the doublet at
2.78 ppm to the H4–To. Such assignments were further sup-
ported by the COSY/HSQC/HMBC spectroscopic analyses,
the details for which are available in the Supporting Infor-
mation. Again, the triplet signal at 5.74 ppm was due to the
H1’ on the thymine 2-deoxyribose.[15] The fact that the same
reaction pattern was obtained in irradiations of both 1 and 4
confirms that the stacking conformation between T and To
is mainly determined by the ring–ring interaction; introduc-
tion of the methyl substitute has little effect to this interac-
tion.
À
À
À
dral angle (C5 C6 C3 C4) is 17.48 in 6; this suggests that
the cyclobutane ring is twisted and the interaction between
the T and To rings is weakened (Figure 3b). The ring defor-
mation is due to the steric repulsion from the 2-CH3 of
MeTo; this leads the cyclobutane ring to adopt a twisted
structure, making it too unstable to form. A similar steric re-
pulsion is expected in 4 should 6 form, making few mole-
cules of 4 to adopt the conformation favoring 6 formation.
This subsequently causes the conformers favoring 5 to in-
crease, as implied by the improved yield of 5 relative to
those of 2 and 3.
To confirm that the lack of formation of 6 during the
UVC irradiation of 4 is due to the steric hindrance caused
by the 2-methyl group on the toluenyl ring, stabilities of 5
and 6 were examined computationally by using density func-
All these CPD products adopt the cis–syn configuration,
as revealed by their 2D-NOESY spectra.[12b] As illustrated
by the spectrum of 2 in Figure 4, the H6–T and H4–To pro-
tons were found to associate with both methyl groups; this
7826
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 7823 – 7833

Thymine Isostere Biochemistry
FULL PAPER
The photo-formation of a CPD analogue between a thy-
mine and a toluene residue is surprising. The substituted tol-
uene anions have been extensively studied as computational
models to facilitate the understanding of the pyrimidine
electronic properties. Neutral molecules like toluene, how-
ever, were suggested to possess rather different molecular
orbital diagrams due to their higher symmetries, less p elec-
trons, fewer nonbonding electronic transitions as well as
lack of tautomeric structure,[5,6] making them rather inert
under UV radiation. That is why toluene is often selected as
the solvent for photo-cycloaddition reactions.[7b,c] Formations
of CPD analogues 2, 3 and 5 must be due to the photochem-
ically excited thymine residue. In previous mechanistic stud-
ies of the spore photoproduct formation, we proved that the
excited thymine C5=C6 behaves like a di-radical and con-
ducts typical radical reactions, such as the H atom abstrac-
tion.[10] The formation of T< >To is likely due to the radical
mediated [2+2] cycloaddition, and further illustrates the di-
radical nature of the excited thymine C5=C6 bond.
Figure 4. Zoomed-in view of the ROESY spectra of 2 in [D6]DMSO. The
protons associated with the cyclobutane ring are labeled and the interac-
tions among these protons are highlighted in circles.
suggests that these protons and the methyl groups locate on
the same side of the cyclobutane plane. This observation
also provides evidence that the two CH3 moieties and the
two H protons are at the same side of the plane. The syn
conformation of these four groups dictates the thymine ring
to be cis to the toluenyl plane. Similar CH3–H6 interactions
were observed in the NOESY spectrum of cis–syn T< >
T.[20] In contrast, the CH3 in the trans–syn T< >T only inter-
acts with the H6 proton on the same thymine.[20,21] Taken to-
gether, the NMR spectroscopy data prove 2 to be a CPD an-
alogue with a cis–syn configuration.
It is worth pointing out that in dinucleotide TpT photo-
chemistry, two cis–syn and two trans–syn diastereomers, are
possible in theory.[20b] The first cis–syn diastereomer adopts
a structure similar to 2a (Scheme 2); its formation requires
both thymidines to adopt the anti N-glycosidic conforma-
tions. The second cis–syn isomer, the four chiral centers on
which the cyclobutane ring adopts opposite stereogeome-
tries, requires the syn–syn N-glycosidic conformations in-
stead and is never observed in TpT photochemistry.[20b,22] X-
ray structure reveals the DFTo nucleoside to adopt an anti
glycosidic conformation,[23] the conformation remains once it
is incorporated into the duplex framework as revealed by
the NMR spectroscopy studies.[18] These observations sug-
The fact that toluene is often used as a solvent for photo-
reaction without causing obvious side reactions[7] suggests
that the stacked structure in the dinucleotide TpTo frame-
work must play a key role in minimizing the entropic effect
and providing the required template to enable the occur-
rence of the [2+2] photoaddition reaction. As revealed by
Kool et al., the desolvation effect is largely responsible for
the strong stacking association of an aromatic ring to the
DNA bases.[14] Disturbing this stacking interaction by con-
ducting the TpTo photoreaction in methanol or dry film to-
tally abolished the T< >To formation. Thymine photo-di-
merization occurs on the picosecond timescale.[9] This ultra-
fast nature determines that the original DNA conformation
controls the outcome of the photoreaction.[9,13,25] The differ-
ent CPDs generated by TpT and TpTo photoreactions sug-
gest that they are likely to adopt different stacking struc-
tures. In TpT, the two C5=C6 bonds must be close to each
other and dimerize to form the CPD commonly observed in
pyrimidine photochemistry. In contrast, although in theory
all six bonds of the To ring are reactive, that the cross-link-
ing reaction involves the C4=C5 in the formation of 2 im-
plies that this bond is in the vicinity of the thymine C5=C6
bond.
To test this hypothesis, we turned to the duplex DNAs
that contain either a TpT step or a TpDFTo sequence. It
was suggested that the base stacking interaction is similar in
single- and double-stranded DNA contexts; the DNA con-
text does not affect the CPD formation.[9,26] A recent study
further proves that the vertical base stacking, not the base-
paring interaction, is the dominating force in stabilizing the
double helical DNA.[27] Our DFT calculation also suggests
that the TpTo conformation remains the same whether it is
at intrahelical or extrahelical position. Although the struc-
tural information for the dinucleotides is not available, the
stacking interaction revealed in duplex DNA could help ex-
plain the reaction pattern observed in dinucleotide photo-
chemistry.
À
gest that the replacement of the natural C N glycosidic
À
bond by a C C bond results in little conformational change.
For the To residue, the syn and anti conformation makes
little difference energetically due to the lack of substitution
at the 2-position. The 2-MeTo is likely to resemble DFTo
and adopt the anti glycosidic conformation. As the anti con-
formation is dominant in thymidine,[24] the syn–syn N-glyco-
sidic conformations are not expected in either compound
1 or 4. Consequently, although our NMR spectroscopy data
do not allow us to completely rule out the possible forma-
tion of type II cis–syn diastereomer, the resulting CPD spe-
cies 2, 3 and 5 are most likely to possess the conformations
similar to the type I cis–syn species observed in the TpT
photochemistry.
Chem. Eur. J. 2012, 18, 7823 – 7833
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7827

L. Li, J. Pu et al.
Indeed, analyses of the DFTo containing DNA structure
indicate that the thymine C5=C6 bond is very close to the
C4=C5 bond of the toluenyl moiety in duplex DNA. As re-
vealed by an NMR spectroscopy study,[18] the C5=C6 bond
of T aligns perfectly with the C4=C5, not the C5=C6 bond
of DFTo, in a duplex DNA dodecamer.[18] In a recent X-ray
structure, although the C5=C6 bond of T is no longer paral-
lel with the C4=C5 of DFTo, the distance between these two
bonds is still about 0.5 ꢁ shorter than that between the two
C5=C6 bonds.[4a] In contrast, the distance between the two
thymine C5=C6 bonds is approximately 0.5 ꢁ shorter than
It is generally accepted that the DNA conformation con-
trols the outcome of the DNA photoreaction.[9,13,25] Howev-
er, the exact thymine conformation and factors controlling
that conformation to facilitate a specific thymine dimer for-
mation are largely unclear. The DNA tertiary structure
could impact the relative positions of the thymines, resulting
in the formation of “unusual” CPD product(s). This conclu-
sion can be illustrated by the formation of interstrand cis–
anti cyclobutane dimer in a 14-mer duplex DNA[33] as well
as the cis–syn/cis–anti thymine dimers formed between Ts
located at different positions of G-quadruplex loops.[34] The
conformation can also be affected by modifications on
either 2-deoxyribose[13a] or nucleobase.[35] A recent study
used locked nucleic acids (LNA) to constrain the sugar at
the C3’-endo conformation, and obtained the cis–syn T< >
T as the exclusive thymine photoproduct.[13a] Acetylation of
the 4-NH2 moiety in cytosine resulted in a trans–syn C< >T
isomer with the syn–anti conformation in a highly stereose-
lective manner due to the steric hindrance induced by the
acetyl modification.[35] Our report here further reveals that
by simply changing the hydrophobicity of the bases, the
DNA conformation can also be altered. The stacked TpTo
structure is likely approximately 608 away from the TpT
conformation; this results in the formation of the regular
cis–syn T< >T. Such a conformational change alters the
outcome of thymine photochemistry, as demonstrated by the
formation of new CPD analogues 2, 3, and 5 reported here.
À
that between the C5=C6 of the 5’-T and C3 C4 of the 3’-T
in the typical B-form DNA, as revealed by both NMR spec-
troscopy[28] and X-ray crystallographic studies.[29]
Such results support our findings in the TpTo photochemi-
cal studies, and suggest that the thymine substitution by
DFTo might cause an additional twist to the DNA double-
helical turns. This twist could have been largely overlooked
in previous thymine isostere biochemical studies as attention
was mainly focused on the maintenance of the duplex struc-
ture after the thymine substitution. Little structural differ-
ence was suggested after the substitution as compared with
thymine and adenine in the same duplex sequence.[1c,18,30]
Our results indicate that although the substituted toluene
closely mimics the size and shape of thymine, its stronger
hydrophobic nature can change the DNA local structure,
which can be probed by the structure-sensitive DNA photo-
chemistry.
The unique TpTo configuration discussed above clearly
favors the formation of 2. The To ring might flip and stack
on the Tring using the other side of the aromatic plane.
Such a conformation brings the C3=C4 of To to the neigh-
borhood of the thymine C5=C6, producing 3 under UV irra-
diation. Although both TpTo conformers are possible, the
lower yield of 3 (4.6% for 3 vs. 7.8% for 2) suggests that
the TpTo conformers favoring the formation of 2 outnumber
those favoring that of 3 by twofold at 77 K. Such a difference
in TpTo population is ascribed to the hydrophobic nature of
the methyl group, as a potential Me–Me interaction is ex-
pected in the TpTo conformers before 2 forms. In contrast,
both methyl groups are exposed to water in those favoring
the formation of 3.
The methyl–methyl interaction has been suggested to be
the major stabilizing force in the TT steps of duplex DNA,
which could contribute up to 15% of the total stacking
energy.[31] Our data provide the first experimental evidence
for the importance of this interaction. Such a stabilization
effect is best reflected by reactions at low temperature. As
the pyrimidine–pyrimidine stacking interaction was estimat-
ed to stabilize DNA by approximately 2 kcalmolÀ1,[32] the
stabilization effect by the Me–Me interaction is negligible at
298 K, as reflected by the roughly equal yields of 2 and 3 in
the solution reactions (Figure S1 in the Supporting Informa-
tion).[12b] The sum of the yields for 2 and 3 roughly equals
that of 5 (16.5%) for reactions at 77 K; this suggests that
the overall populations in TpTo complexes favoring the
CPD formation are similar.
Conclusion
In summary, we have synthesized a dinucleotide TpTo and
studied its photoreaction under 254 nm UV irradiation. Al-
though toluene is largely inert under UV light, it dimerizes
with a thymine residue through a [2+2] photoreaction prob-
ably due to the presence of the dinucleotide framework,
which minimizes the entropic effect and provides the re-
quired reaction template. The TpTo photochemistry yields
two cis–syn T< >To species as the major photoproducts.
More importantly, different from the normal CPDs formed
between the two C5=C6 bonds in pyrimidine residues, the
CPD species in TpTo forms between the C5=C6 of T and
the C3=C4 or C4=C5 bond of To. Considering the ultrafast
reaction rate in DNA photoreaction, the dinucleotide con-
formation must remain during the dimerization reaction.
Thus, the stacked structure in TpTo is likely to be 608 away
from the photoactive conformation in TpT. These data sug-
gest that although the substituted toluene closely mimics the
size and shape of thymine, its stronger hydrophobic nature
might result in a subtle DNA conformational change.
The natural DNA adopts a right-handed helical structure,
which results in an approximately 308 turn between the ad-
jacent bases. To make the C5=C6 of T close to the C4=C5
bond of To, an additional base rotation that can be as small
as ꢀ308 is needed. We believe that such a minor rotational
movement induced by the To substitution is reasonable as it
is unlikely that the more hydrophobic toluenyl moiety
7828
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 7823 – 7833

Thymine Isostere Biochemistry
FULL PAPER
with CH2Cl2 (30 mL). The combined organic layers were washed with
a saturated aqueous solution of NaHCO3 and brine, subsequently dried
over anhydrous MgSO4 and the solvent was removed under reduced pres-
sure. The product was purified by silica gel column chromatography
(hexane/ethyl acetate=16:1) to yield an a/b mixture (a/b=1.6:1) as a col-
orless oil (26.42 g, 98%).
would result in no structural change at all after it replaces
the thymine residue in duplex DNA. At the same time, this
minor rotation largely retains the DNA duplex structure,
which likely explains why such a conformational change
were present in the previous NMR spectroscopy[18] and X-
ray structures[29b] of the DFTo containing duplex DNA but
was never clearly reported in these previous studies. These
structures predict that a similar photoreaction should still
occur in the context of duplex DNA. We will repeat the
TpTo photochemistry in double-strand DNA to confirm this
prediction in our future research. Further work to reveal the
mechanistic details of the TpTo photochemistry (for in-
stance, singlet vs. triplet excited state) as well as the correla-
tion between the thymine conformational change due to the
base rotation and thymine photochemistry is also in prog-
ress.
The resulting isomer mixture was then dissolved in CH2Cl2 (529 mL) and
benzenesulfonic acid (16.57 g, mmol) as well as TFA (100 mL) were
added. The solution was refluxed for 24 h and Et3N (331 mL) was added
to neutralize the acids. The solvents were removed by rotary evaporation
and the resulting residue was purified by column chromatography
(hexane/ethyl acetate=16:1) to afford the b isomer compound 7a
(Scheme 4) as a colorless oil (15.39 g, 58%). 1H NMR (CDCl3): d=2.13
Experimental Section
General methods: All reagent grade chemicals were purchased from
Sigma, Fisher, or VWR and were used without further purification
except the 3,5-bis(4-chlorobenzoyl)-2-deoxy-a-d-ribofuranosyl chloride,
which was purchased from Shanghai Hanhong Chemical Co., Ltd.
(Shanghai, P.R. China). All reactions were carried out by using oven- or
flame-dried glassware under a nitrogen atmosphere in distilled solvents.
Dichloromethane and pyridine were distilled over calcium hydride. Pu-
rification of reaction products was carried out by flash chromatography
by using silica gel (Dynamic Adsorbents Inc, 32–63 mm). For TLC analy-
sis, precoated plates (w/h F254, Dynamic Adsorbents, Inc., 0.25 mm
thick) were used. 1H and 13C NMR spectra were obtained on a Bruker
500 MHz NMR Fourier transform spectrometer. NMR spectra were re-
corded in sample solutions in deuterated chloroform (CDCl3) with resid-
ual chloroform (d=77.0 ppm for 13C NMR) and TMS (d=0 ppm for
1H NMR), deuterated methanol (d=3.31 ppm for 1H NMR and d=
49.0 ppm for 13C NMR), deuterated DMSO (d=2.50 ppm for 1H NMR
and d=39.52 ppm for 13C NMR). The chemical shifts in the NMR spectra
are reported in parts per million (ppm). Mass (MS) analysis was obtained
by using Agilent 1100 series LC/MSD system with electrospray ionization
(ESI). The TpTo and TpMeTo photoreactions were carried out by using
a Spectroline germicidal UV sterilizing lamp (Dual-tube, 15 W, intensity:
1550 uWcmÀ2) with the samples about 5 inches from the lamp.
DFT calculations on the stability of CPD analogues 5 and 6: The initial
structures of the cyclobutane pyrimidine dimer (CPD) were constructed
by using residues 5 and 6 of chain A, corresponding to the T< >T in the
intrahelical position of a duplex DNA in structure 1BW7 (PDB ID), and
residues 7 and 12 of chain I, corresponding to the T< >T in the extrahel-
ical 1TEZ (PDB ID). In 1TEZ, the intradimer contains a formacetal
linker instead of the phosphate in the structure. The potential CPD prod-
ucts 5 and 6 were optimized from these two T< >T structures revealed
by the PDB files. The structures of 5 and 6 were optimized with G03[16]
at the Austin Model 1 (AM1) level first, then at HF/6-31G(d) and
Scheme 4. Synthesis of the dinucleotide TpTo (1).
B3LYP/6-31+GACHTUNGTRENNUNG
(d,p) level.[17]
Synthesis of 7a: Magnesium (3.24 g, 126 mmol) was activated by iodine
etching in a Schlenk flask while being stirred without solvent. THF
(354 mL) and 3-bromotoluene (22.6 g, 123 mmol) were then added and
the mixture was stirred at 508C until all the magnesium was consumed.
The resulting solution was cooled to 08C before copper iodide (13.1 g,
68.7 mmol) was added. The reaction mixture was allowed to warm to
208C and stirred until all the copper iodide was dissolved (~30 min). The
solution was heated to 408C, 2-deoxy-a-d-erythro-pentofuranosyl chlo-
ride bis(4-chloride benzoate) (25 g, 58.1 mmol) was added and the mix-
ture was allowed to stir for 1 h at 408C. The reaction was stopped by the
addition of a 10% aqueous solution of NH4Cl followed by extraction
(m, 1H), 2.21 (s, 3H), 2.31 (s, 3H), 2.56 (dd, J=4.55, 13.45 Hz, 1H), 4.51
(dd, J=2.0, 4.0 Hz, 1H), 4.66 (t, J=5.0 Hz, 2H), 5.22 (dd, J=5.0,
11.0 Hz, 1H), 5.59 (d, J=6.5 Hz, 1H), 7.10 (d, J=7.5 Hz, 1H), 7.24 (m,
4H), 7.40 (d, J=8.5 Hz, 2H), 7.46 (d, J=8.6 Hz, 2H), 8.00 ppm (m, 4H);
13C NMR (CDCl3): d=21.3, 41.5, 65.0, 77.6, 80.8, 82.7, 122.9, 126.4, 128.1,
128.2, 128.4, 128.7, 128.8, 131.0, 138.2, 139.6, 139.9, 140.3, 165.2,
165.4 ppm; ESI-MS (positive mode) calcd for C26H23Cl2O5: 485.1
AHCTUNGTRENNUNG
[M+H+], found 485.0.
Synthesis of 8a: Compound 7a (1.92 g, 3.935 mmol) was dissolved in
methanol (44 mL) and NaOMe (0.66 g, 12.24 mmol) was added immedi-
Chem. Eur. J. 2012, 18, 7823 – 7833
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7829

L. Li, J. Pu et al.
ately. The mixture was stirred for 3 h at room temperature and NH4Cl
(0.66 g, 12.24 mmol) was added to quench the reaction. The solvent was
removed by rotary evaporation and the residue was purified by column
chromatography (CH2Cl2/MeOH=20:1) to afford the product as a color-
tography (CH2Cl2/Et3N/MeOH=30:1.5:1) to afford compound 14 as
a white solid (3.84 g, 87%). 1H NMR (CDCl3): d=1.34 (s, 1H), 2.39 (m,
1H), 2.59 (m, 1H), 3.39 (dd, J=2.5, 10.5 Hz, 1H), 3.48 (dd, J=2.5,
10.5 Hz, 1H), 3.78 (s, 6H), 4.27 (d, J=2.0 Hz, 1H), 5.01 (m, 1H), 6.46
(dd, J=5.5, 8.0 Hz, 1H), 6.81 (s, 2H), 6.83 (s, 2H), 7.22 (m, 1H), 7.28 (m,
6H), 7.40 (d, J=7.0 Hz, 2H), 7.60 ppm (s, 1H); 13C NMR (CDCl3): d=
9.2, 45.5, 55.1, 63.4, 74.0, 84.3, 85.1, 86.8, 111.0, 113.1, 120.4, 120.5, 122.9,
126.9, 127.8, 128.1, 129.1, 130.0, 135.2, 135.3, 135.6, 144.2, 150.6, 158.5,
164.0 ppm.
1
less oil (0.88 g, 82%). H NMR (CDCl3): d=1.94 (m, 1H), 2.16 (m, 1H),
2.32 (s, 3H), 3.67 (d, J=5.5 Hz, 2H), 3.96 (dd, J=2.5, 5.5 Hz, 1H), 4.30
(m, 1H), 5.09 (dd, J=5.5, 10.5 Hz, 1H), 7.08 (m, 3H), 7.20 ppm (t, J=
7.5 Hz, 1H); 13C NMR (CDCl3): d=21.4, 43.3, 63.2, 73.5, 80.2, 87.3,
123.1, 126.7, 128.4, 128.5, 138.1, 140.9 ppm; ESI-MS (positive mode)
calcd for C12H16NaO3: 231.1 [M+H+], found 231.0.
Synthesis of compound 15a: Compounds 11a (297 mg, 0.488 mmol) and
14 (111 mg, 0.444 mmol) were dried and dissolved in anhydrous pyridine
(10 mL); PivCl (0.43 mL, 3.11 mmol) was added dropwise to this solu-
tion. The reaction mixture was stirred at room temperature for 20 min,
quenched by addition of water (1.3 mL) and oxidized with I2 (345 mg,
1.36 mmol). The solution was further stirred for 1 h before Na2S2O4 was
added to quench the reaction. Et3N (5.5 mL) was then added to facilitate
the formation of the triethylammonium salt with the phosphate moiety in
15a, thus enhancing its solubility in organic solvents. After removing the
solvents by rotary evaporation, the resulting residue was dissolved in
CH2Cl2. The solution was washed with NaHCO3 (aq., saturated) three
times, the solvent was evaporated and the resulting compound was puri-
fied by column chromatography (CH2Cl2/Et3N/MeOH=20:1:1) to afford
a triethylammonium salt of compound 15a as a white solid (0.433 g,
56.6%). 1H NMR (CDCl3): d=3.78 (s, 6H), 4.02 (m, 2H), 4.15 (d, J=
1.5 Hz, 1H), 4.32 (d, J=1.0 Hz, 1H), 5.01 (m, 1H), 5.25 (d, J=5.5 Hz,
1H), 5.35 (s, 1H), 6.81 (d, J=8.5 Hz, 4H), 7.05 (s, 1H), 7.15 (m, 3H),
7.24 (m, 7H), 7.37 (d, J=7.5 Hz, 2H), 7.58 ppm (s, 1H); 13C NMR
(CDCl3): d=10.9, 20.5, 20.8, 38.8, 40.3, 52.5, 53.1, 54.6, 63.4, 65.1, 76.0,
79.8, 83.0, 83.1, 83.9, 86.3, 110.4, 112.5, 122.5, 126.2, 126.4, 127.3, 127.6,
127.9, 129.5, 134.6, 134.8, 135.2, 137.2, 140.0, 143.7, 149.9, 158.0, 163.3,
169.8 ppm.
Synthesis of 9a: Compound 8a (0.88 g, 4.25 mmol) was dissolved in pyri-
dine (8 mL) and Et3N (0.77 mL) and DMAP (0.104 g, 0.85 mmol) was
added to this solution. After being stirred for 3 min at room temperature,
4,4’-dimethoxytrityl chloride (DMTrCl; 1.87 g, 5.52 mmol) was added at
room temperature. After being stirred for 6 h, MeOH (4 mL) was added
to quench the reaction and the solvent was removed by rotary evapora-
tion. The resulting residue was purified by column chromatography
(hexane/ethyl acetate=4:1) to afford compound 9a as a white solid
1
(1.504 g, 70%). H NMR (CDCl3): d=2.18 (m, 1H), 2.29 (s, 3H), 2.40 (d,
J=3.5 Hz, 1H), 3.28 (d, J=5.0 Hz, 1H), 3.31 (d, J=4.5 Hz, 1H), 4.06 (d,
J=2.5 Hz, 1H), 4.38 (d, J=2.0 Hz, 1H), 5.13 (dd, J=5.5, 10.0 Hz, 4H),
6.80 (dd, J=3.0, 12.0 Hz, 4H), 7.06 (d, J=7.5 Hz, 1H), 7.20 (m, 3H),
7.25 (m, 3H), 7.36 (m, 4H), 7.47 ppm (d, J=8.5 Hz, 2H); 13C NMR
(CDCl3): d=21.3, 43.9, 55.1, 64.5, 74.5, 80.0, 86.1, 86.3, 113.0, 123.1,
126.6, 126.7, 127.7, 128.1, 130.0, 136.0, 137.8, 141.6, 144.8, 158.3 ppm.
Synthesis of 11a: Compound 9a (1.5 g, 2.95 mmol) was dissolved in pyri-
dine (10 mL) and DMAP (37 mg, 0.30 mmol) and Ac2O (0.42 mL,
4.43 mmol) were added to this solution at room temperature. After being
stirred for 6 h, H2O (5 mL) was added to quench the reaction and the sol-
vent removed by rotary evaporation. The resulting residue was dissolved
in CH2Cl2 (50 mL) and washed with NH4Cl (aq., saturated) and NaHCO3
(aq., saturated). Then the solvent was evaporated and the resulting resi-
due purified by column chromatography to afford compound 10a as
a white solid (1.57 g, 96%). The product was used directly in the next
step of synthesis without further purification.
Synthesis of 1: The triethylammonium salt of compound 15a (0.43 g,
0.251 mmol) was dissolved in 20 mL CH2Cl2 solution of TFA (3%). After
10 min, the solvent was removed by rotary evaporation. The residue was
subsequently dissolved in MeOH (13 mL) and NH3·H2O (38 mL) was
added to this solution. The mixture was stirred for 12 h, the solvents
were removed by rotary evaporation and the crude product was stirred
with the cation exchange resin in water for 3 h to exchange the corre-
sponding cations to proton. After filtration to remove the resin, the sol-
vents were evaporated and the crude product was purified with RP-
HPLC. The eluent was collected and solvents were removed to afford
Compound 10a (1.57 g, 1.09 mmol) was dissolved in a solution of TFA
(3%) in dichloromethane (40 mL) at room temperature. After 30 min,
the red solution was evaporated to dryness under reduced pressure. The
crude material was redissolved in CH2Cl2 and purified by silica gel chro-
matography (hexane/ethyl acetate=4:1). Compound 11a was obtained as
a white foam after evaporation of the solvent (0.483 g, 68%). 1H NMR
(CDCl3): d=2.09 (m, 1H), 2.13 (s, 3H), 2.34 (m, 1H), 2.36 (s, 3H), 3.83
(dd, J=4.5, 11.5 Hz, 1H), 3.87 (dd, J=4.0, 12.0 Hz, 1H), 4.08 (m, 1H),
5.06 (dd, J=5.0, 11.0 Hz, 1H), 5.22 (dd, J=2.0, 6.0 Hz, 1H), 7.15 (m,
3H), 7.26 ppm (m, 1H); 13C NMR (CDCl3): d=21.1, 21.4, 41.0, 63.3,
76.8, 80.4, 85.5, 123.1, 126.7, 128.5, 128.8, 138.2, 140.1, 171.0 ppm; ESI-
MS (positive mode) calcd for C14H19O4: 251.1 [M+H+], found 251.1.
1
compound 1 as a colorless solid (76 mg, 58.9%). H NMR (CD3OD): d=
1.86 (s, 3H), 2.02 (m, 1H), 2.19 (m, 2H), 2.31 (s, 3H), 2.49 (m, 1H), 3.76
(t, J=2.5 Hz, 2H), 3.98 (dd, J=5.0, 10.0 Hz, 1H), 4.02 (dd, J=5.0,
10.0 Hz, 1H), 4.05 (m, 1H), 4.14 (dd, J=3.0, 5.5 Hz, 1H), 4.42 (dd, J=
1.5, 4.0 Hz, 1H), 4.84 (dd, J=3.5, 6.5 Hz, 1H), 6.27 (dd, J=6.0, 7.5 Hz,
1H), 7.04 (d, J=6.0 Hz, 1H), 7.17 (m, 3H), 7.78 ppm (s, 1H); 13C NMR
(CD3OD): d=21.5, 22.7, 40.0, 44.4, 62.9, 64.4, 67.2, 74.4, 76.9, 81.8, 86.3,
87.5, 111.6, 124.4, 128.0, 129.3, 138.2, 139.1, 142.9, 152.4, 166.4,
178.2 ppm; ESI-MS (positive mode) calcd for C22H30N2O10P: 513.2 [M+
H+], found 513.2.
Synthesis of 13: Thymidine (4.00 g, 16.6 mmol; 12) was dissolved in pyri-
dine (150 mL) and DMTrCl (6.75 g, 20.0 mmol) was added at room tem-
perature. After being stirred for 6 h, MeOH (4.00 g, 125 mmol) was
added to quench the reaction and the solvents were evaporated under
a reduced pressure. The residue was purified by column chromatography
(CH2Cl2/MeOH=40:1) to afford compound 13 as a white solid (7.8 g,
86%).[5] 1H NMR (CDCl3): d=1.45 (s, 3H), 2.31 (m, 1H), 2.43 (m, 1H),
3.36 (dd, J=3.0, 10.5 Hz, 1H), 3.45 (dd, J=3.0, 10.5 Hz, 1H), 3.76 (s,
6H), 4.09 (d, J=2.0 Hz, 1H), 4.57 (s, 1H), 6.44 (t, J=7.0 Hz, 1H), 6.82
(d, J=8.5 Hz, 4H), 7.21 (m, 1H), 7.27 (m, 7H), 7.39 (J=7.5 Hz, 2H),
7.61 ppm (s, 1H); 13C NMR (CDCl3): d=11.7, 40.9, 53.4, 55.2, 63.6, 72.4,
84.8, 86.3, 86.8, 111.2, 113.2, 127.0, 127.9, 128.1, 130.0, 135.3, 135.4, 135.7,
144.3, 149.4, 150.7, 158.6, 164.0 ppm.
Synthesis of 7b: Magnesium (0.98 g, 40.8 mmol) was activated by iodine
etching in a Schlenk flask while being stirred without solvent. THF
(60 mL) and 3-bromo-4-methyl-toluene (7.400 g, 40 mmol) were then
added and the mixture was stirred at 508C until all the magnesium was
consumed. Then THF (150 mL) was added. The resulting solution was
cooled to 08C before copper iodide (3.960 g, 20.7 mmol) was added. The
reaction mixture was allowed to warm to 208C and stirred until all the
copper iodide was dissolved (~30 min). The solution was heated to 408C,
2-deoxy-a-d-erythro-pentofuranosyl chloride bis(4-chloride benzoate)
(7.570 mg, 17.6 mmol) was added and the mixture was allowed to stir for
2 h at 408C. The reaction was stopped by the addition of a 10% aqueous
solution of NH4Cl followed by extraction with CH2Cl2 (30 mL). The com-
bined organic layers were washed with a saturated aqueous solution of
NaHCO3 and brine, subsequently dried over anhydrous MgSO4 and sol-
vent was removed under reduced pressure. The product was purified by
silica gel column chromatography (hexane/ethyl acetate=30:1) to yield
compound 7b (Scheme 5) as a colorless oil (5.300 g, 55%). 1H NMR
Synthesis of 14: Diphenyl phosphite (4.00 g, 11.1 mmol) was added to
a solution of 13 (4.00 g, 11.1 mmol) in pyridine (5 mL). The reaction was
quenched after 15 min by addition of a mixture of water–triethylamine
(1:1 v/v, 2 mL) and left standing for 15 min. The solvent was removed by
rotary evaporation and the residue was partitioned between CH2Cl2
(50 mL) and 5% aq. NaHCO3 (20 mL). The organic layer was washed
twice with 5% aq. NaHCO3 (20 mL), dried over Na2SO4 and finally
evaporated to yield an oil. The product was purified by column chroma-
7830
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 7823 – 7833

Thymine Isostere Biochemistry
FULL PAPER
Synthesis of 9b: Compound 8b (2.110 g, 9.51 mmol) was dissolved in pyr-
idine (18 mL) and Et3N (1.72 mL) and DMAP (0.233 g, 1.91 mmol) was
added to this solution. After being stirred for 3 min at room temperature,
4,4’-dimethoxytrityl chloride (DMTrCl; 4.187 g, 12.36 mmol) was added
at room temperature. After being stirred for 6 h, MeOH (9 mL) was
added to quench the reaction and the solvent was removed by rotary
evaporation. The resulting residue was purified by column chromatogra-
phy (hexane/ethyl acetate=4:1) to afford compound 9b as a yellow solid
(4.548 g, 91%). 1H NMR (CDCl3): d=1.95 (m, 1H), 2.23 (s, 3H), 2.25
(m, 1H), 2.27 (s, 3H), 3.36 (m, 2H), 3.77 (s, 6H), 4.03 (d, J=3.0 Hz,
1H), 4.41 (t, J=3.0 Hz, 1H), 5.32 (dd, J=5.7, 9.6 Hz, 1H), 6.82 (d, J=
8.8 Hz, 4H), 6.96 (d, J=7.7 Hz, 1H), 7.01 (d, J=7.7 Hz, 1H), 7.20 (t, J=
7.4 Hz, 1H), 7.27 (dd, J=7.4, 8.6 Hz, 2H), 7.37 (m, 4H), 7.41 (s, 1H),
7.49 ppm (d, J=8.1 Hz, 2H); 13C NMR (CDCl3): d=18.7, 21.1, 42.6, 46.1,
55.2, 64.3, 74.7, 76.8, 85.7, 86.2, 113.1, 125.6, 126.8, 127.7, 127.8, 128.2,
130.0, 130.1, 131.3, 135.5, 136.1, 139.9, 144.9, 158.4 ppm.
Synthesis of 11b: Compound 9b (4.546 g, 2.95 mmol) was dissolved in
pyridine (10 mL) and DMAP (37 mg, 0.30 mmol) and Ac2O (0.42 mL,
4.43 mmol) were added to this solution at room temperature. After being
stirred for 6 h, H2O (5 mL) was added to quench the reaction and the sol-
vent was removed by rotary evaporation. The resulting residue was dis-
solved in CH2Cl2 (50 mL) and washed with NH4Cl (aq., saturated) and
NaHCO3 (aq., saturated). Then the solvent was evaporated and the re-
sulting residue was purified by column chromatography to afford com-
pound 10b as a white solid (1.66 g, 99%). The product was used directly
in the next step of synthesis without further purification.
Compound 10b (1.66 g, 2.94 mmol) was dissolved in a solution of TFA
(3%) in dichloromethane (40 mL) at room temperature. After 30 min,
the red solution was evaporated to dryness under reduced pressure. The
crude material was redissolved in CH2Cl2 and purified by silica gel chro-
matography (hexane/ethyl acetate=4:1). Compound 11b was obtained as
a white solid after evaporation of the solvent (0.565 g, 73%). 1H NMR
(CDCl3): d=2.00 (m, 1H), 2.13 (s, 3H), 2.30 (s, 3H), 2.32 (s, 3H), 2.36
(m, 1H), 3.88 (m, 2H), 4.06 (dd, J=4.4, 7.2 Hz, 1H), 5.22 (m, 2H), 7.00
(d, J=7.8 Hz, 1H), 7.04 (d, J=7.7 Hz, 1H), 7.26 ppm (s, 1H); 13C NMR
(CDCl3): d=18.7, 21.1, 39.7, 63.2, 85.1, 125.2, 128.3, 130.3, 131.8, 135.7,
137.9, 171.1 ppm. ESI-MS (positive mode) calcd for C15H21O4: 265.1
AHCTUNGTRENNUNG
[M+H+], found 265.0.
Synthesis of compound 15b: Compound 11b (400 mg, 1.52 mmol) and
compound 14 (1.010 mg, 1.67 mmol) were dried and dissolved in anhy-
drous pyridine (34 mL); PivCl (1.48 mL, 10.70 mmol) was added drop-
wise to this solution. The reaction mixture was stirred at room tempera-
ture for 20 min, quenched by addition of water (1.3 mL) and oxidized
with I2 (561 mg, 2.21 mmol). The solution was further stirred for 1 h
before Na2S2O4 was added to quench the reaction. Et3N (5.5 mL) was
then added to facilitate the formation of the triethylammonium salt with
the phosphate moiety in 15b, thus enhancing its solubility in organic sol-
vents. After removing the solvents by rotary evaporation, the resulting
residue was dissolved in CH2Cl2. The solution was washed with NaHCO3
(aq., saturated) for three times, the solvent was evaporated and the re-
sulting compound was purified by column chromatography (CH2Cl2/
Et3N/MeOH=20:1:1) to afford a triethylammonium salt of compound
15b as a white solid (0.412 g, 31%). 1H NMR (CDCl3): d=2.07 (s, 3H),
2.24 (s, 6H), 2.71 (m, 1H), 3.40 (d, J=2.4 Hz, 2H), 3.76 (s, 3H), 3.77 (s,
3H), 4.00 (m, 1H), 4.07 (m, 1H), 4.17 (m, 1H), 4.33 (d, J=1.8 Hz, 1H),
5.06 (t, J=5.8 Hz, 1H), 5.15 (dd, J=4.9, 11.0 Hz, 1H), 5.30 (b, 1H), 6.48
(dd, J=5.4, 8.9 Hz, 1H), 6.80 (d, J=8.9 Hz, 4H), 6.93 (d, J=7.6 Hz, 1H),
6.98 (d, J=7.7 Hz, 1H), 7.25 (m, 8H), 7.37 (d, J=7.6 Hz, 2H), 7.59 ppm
(s, 1H); 13C NMR (CDCl3): d=10.4, 11.4, 18.6, 21.0, 21.1, 39.3, 45.8, 55.1,
84.5, 86.8, 110.9, 113.1, 125.6, 126.9, 127.8, 127.9, 128.2, 129.9, 130.1,
131.4, 135.2, 135.4, 135.8, 138.4, 144.3, 150.4, 158.5, 163.9, 170.3 ppm.
Scheme 5. Synthesis of the dinucleotide TpMeTo (4).
(CDCl3): d=2.22 (m, 1H), 2.29 (s, 3H), 2.51 (dd, J=4.1, 14.0 Hz, 1H),
4.50 (m, 1H), 4.72 (m, 2H), 5.41 (m, 1H), 5.60 (t, J=2.9 Hz, 1H), 6.99
(d, J=7.7 Hz, 1H), 7.03 (d, J=7.6 Hz, 1H), 7.33 (s, 1H), 7.39 (d, J=
8.4 Hz, 2H), 7.46 (d, J=8.5 Hz, 2H), 8.01 ppm (m, 4H); 13C NMR
(CDCl3): d=18.6, 21.0, 40.2, 64.9, 77.6, 77.7, 82.4, 125.2, 128.1, 128.2,
128.3, 128.8, 128.9, 130.2, 131.0, 131.2, 135.8, 138.3, 139.6, 139.9, 165.2,
165.4 ppm; ESI-MS (positive mode) calcd for C27H24Cl2O5Na: 521.1
ACHTUNGTRENNUNG
[M+Na+], found 521.0.
Synthesis of 8b: Compound 7b (5.300 g, 10.6 mmol) was dissolved in
methanol (118 mL) and NaOMe (1.775 g, 32.9 mmol) was added immedi-
ately. The mixture was stirred for 3 h at room temperature and NH4Cl
(1.775 g, 32.6 mmol) was added to quench the reaction. The solvent was
removed by rotary evaporation and the residue was purified by column
chromatography (CH2Cl2/MeOH=20:1) to afford the product as a color-
less oil (2.110 g, 90%). 1H NMR (CD3OD): d=1.79 (m, 1H), 2.21 (m,
1H), 2.26 (s, 3H), 2.28 (s, 3H), 3.70 (m, 2H), 3.92 (m, 1H), 4.31 (m, 1H),
5.30 (dd, J=5.4, 10.5 Hz, 1H), 6.94 (d, J=7.8 Hz, 1H), 6.98 (d, J=
7.7 Hz, 1H), 7.34 ppm (s, 1H); 13C NMR (CD3OD): d=18.7, 21.2, 43.4,
64.0, 74.4, 78.3, 88.7, 126.6, 128.8, 131.0, 132.6, 136.6, 140.8 ppm; ESI-MS
(positive mode) calcd for C13H19O3: 223.1 [M+H+], found 223.1.
Synthesis of 4: The triethylammonium salt of compound 15b (0.412 g,
0.251 mmol) was dissolved in a CH2Cl2 (30 mL) solution of TFA (3%).
After 10 min, the solvent was removed by rotary evaporation. The resi-
due was subsequently dissolved in MeOH (7 mL) and NH3·H2O (20 mL)
was added to this solution. The mixture was stirred for 12 h, the solvents
were removed by rotary evaporation and the crude product was stirred
with the cation exchange resin in water for 3 h to exchange the corre-
Chem. Eur. J. 2012, 18, 7823 – 7833
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7831

L. Li, J. Pu et al.
sponding cation to proton. After filtration to remove the resin, the sol-
vents were evaporated and the crude product was purified with RP-
HPLC. The eluent was collected and solvents were removed to afford
compound 4 as a colorless solid (0.166 mg, 65%). 1H NMR (CD3OD):
d=1.85 (s, 3H), 1.89 (m, 1H), 2.23 (m, 1H), 2.26 (m, 1H), 2.26 (s, 3H),
2.27 (s, 3H), 2.51 (m, 1H), 3.74 (d, J=2.9 Hz, 2H), 4.07 (d, J=3.1 Hz,
1H), 4.19 (m, 3H), 4.37 (t, J=2.9 Hz, 1H), 4.99 (t, J=6.1 Hz, 1H), 5.32
(dd, J=5.5, 10.3 Hz, 1H), 6.26 (dd, J=5.9, 8.2 Hz, 1H), 6.92 (d, J=
7.6 Hz, 1H), 6.97 (d, J=7.7 Hz, 1H), 7.32 (s, 1H), 7.74 ppm (s, 1H);
13C NMR (CD3OD): d=12.5, 18.8, 21.3, 39.6, 43.2, 62.6, 68.6, 74.1, 78.6,
79.1, 86.1, 86.4, 87.3, 111.7, 126.6, 128.9, 131.1, 132.8, 136.6, 137.9, 140.4,
452.3, 166.3 ppm; ESI-MS (positive mode) calcd for C23H32N2O10P: 527.2
[M+H+], found 527.1.
a 40ꢂ50 cm plate under liquid nitrogen. The frozen solution was then ex-
posed for 15 min to UVC (254 nm) light. The solid was subsequently
warmed to room temperature, frozen again in liquid N2 and exposed for
15 min to UVC (254 nm) light. This process was repeated six times and
the resulting products were purified by reverse phase HPLC in the gradi-
ent mode by using ammonium acetate, pH 6.8, and acetonitrile as sol-
vents. The eluted 4 was recycled for another round of photoreaction.
1
Compound 5: H NMR (D2O): d=1.44 (s, 3H), 1.58 (s, 3H), 1.76 (s, 3H),
1.95 (dd, J=6.5, 13.4 Hz, 1H), 2.22 (m, 1H), 2.51 (m, 2H), 2.78 (d, J=
4.8 Hz, 1H), 3.75 (dd, J=6.1, 12.5 Hz, 1H), 3.83 (dd, J=3.3, 12.6 Hz,
1H), 3.93 (dd, J=4.8, 11.1 Hz, 1H), 3.98 (m, 2H), 4.04 (d, J=11.1 Hz,
1H), 4.10 (s, 1H), 4.47 (m, 2H), 4.89 (t, J=7.5 Hz, 1H), 5.21 (d, J=
5.0 Hz, 1H), 5.60 (s, 1H), 5.74 ppm (t, J=5.8 Hz, 1H); 13C NMR (D2O):
d=18.5, 24.2, 28.4, 37.7, 39.8, 41.4, 48.2, 48.5, 60.9, 65.5, 65.6, 71.2, 73.9,
75.8, 82.7, 84.1, 84.3, 118.7, 123.5, 130.4, 138.5, 152.6, 174.8 ppm; ESI-MS
(positive mode) calcd for C23H32N2O10P: 527.2 [M+H+], found 527.1.
Photoreaction of compound 1 and 4: Irradiation of 1 or 4 (1 mm) was
conducted in aqueous solution at pH 7 at ambient temperature. The reac-
tion was repeated at 77 K. To ensure the formation of a transparent glass,
an equal volume of glycerol was added to the aqueous solution before
being frozen. The resulting glass was then exposed for 15 min to the
UVC (254 nm) light. The ice was subsequently warmed to room tempera-
ture and analyzed by HPLC.
Preparation of 2 and 3 in glycerol/H2O for NMR spectroscopy: An aque-
ous solution of 1 (10 mm; 20 mL) was titrated with NaOH (0.1m) to
pH 7.0. After removal of water under vacuum, the resulting sodium salt
was dissolved in glycerol/H2O (1:1, 25 mL, 4.36 mgmLÀ1). The solution
was frozen in a 40ꢂ50 cm plate under liquid nitrogen. The frozen solu-
tion was then exposed for 15 min to the UVC (254 nm) light. The solid
was subsequently warmed to room temperature, frozen again in liquid N2
and exposed for 15 min to UVC (254 nm) light. This process was repeat-
ed six times and the resulting products were purified by reverse phase
HPLC in the gradient mode by using ammonium acetate, pH 6.8, and
acetonitrile as solvents. The eluted 1 was recycled for another round of
photoreaction. Compound 2: 1H NMR (D2O): d=1.47 (s, 3H), 1.63 (s,
3H), 1.96 (m, 1H), 2.19 (m, 1H), 2.52 (m, 1H), 2.63 (m, 1H), 2.87 (d, J=
5.5 Hz, 1H), 3.75 (dd, J=6.0, 12.5 Hz, 1H), 3.84 (dd, J=3.5, 12.5 Hz,
1H), 3.93 (m, 1H), 3.97 (dd, J=2.5, 6.0 Hz, 1H), 4.00 (m, 1H), 4.03 (dd,
J=2.0, 11.0 Hz, 1H), 4.11 (d, J=0.5 Hz, 1H), 4.47 (m, 2H), 4.56 (t, J=
7.5 Hz, 1H), 5.51 (dd, J=5.0, 9.0 Hz, 1H), 5.52 (d, J=5.0 Hz, 1H), 5.74
(t, J=6.0 Hz, 1H), 5.81 ppm (d, J=9.0 Hz, 1H); 13C NMR (D2O): d=
21.3, 21.6, 24.3, 28.6, 37.7, 39.9, 40.3, 48.7, 48.8, 60.9, 65.8, 66.0, 71.3, 73.9,
79.0, 82.5, 82.6, 84.3, 84.4, 121.9, 123.5, 124.1, 137.0, 152.5, 174.5 ppm;
1H NMR (CD3OD): d=1.47 (s, 3H), 1.61 (s, 3H), 1.83 (m, 1H), 2.10 (m,
1H), 2.36 (m, 1H), 2.71 (d, J=5.2 Hz, 1H), 2.85 (m., 1H), 3.77 (m, 1H),
3.83 (m, 2H), 3.93 (dd, J=3.8, 12.5 Hz, 1H), 4.01 (m, 1H), 4.04 (s, 1H),
4.06 (m, 1H), 4.36 (m, 1H), 4.43 (t, J=7.2 Hz, 1H), 4.55 (m, 1H), 5.50
(dd, J=5.2, 9.7 Hz, 1H), 5.66 ppm (m, 3H); 1H NMR ([D6]DMSO): d=
1.35 (s, 3H), 1.49 (s, 3H), 1.60 (m, 1H), 1.98 (m, 1H), 2.07 (m, 1H), 2.54
(m, 1H), 2.58 (d, J=5.2 Hz, 1H), 3.52 (m, 1H), 3.63 (m, 3H), 3.76 (m,
2H), 3.81 (s, 1H), 4.13 (dd, J=5.2, 10.9 Hz, 1H), 4.29 (t, J=6.8 Hz, 1H),
4.41 (m, 1H), 5.10 (b, 1H), 5.36 (dd, J=5.1, 9.8 Hz, 1H), 5.41 (dd, J=
2.7, 7.9 Hz, 1H), 5.55 (s, 1H), 5.59 ppm (d, J=9.8 Hz, 1H); 13C NMR
([D6]DMSO): d=26.2, 29.8, 39.6, 39.8, 42.2, 47.4, 48.4, 60.5, 64.5, 65.7,
69.4, 72.2, 77.5, 81.7, 83.6, 84.5, 121.1, 123.0, 123.6, 136.0, 150.1,
170.7 ppm; ESI-MS (positive mode) calcd for C22H30N2O10P: 513.2 [M+
Acknowledgements
The authors thank Professors Eric Long and Kyungsoo Oh at IUPUI for
helpful discussions. This work was supported by the National Institutes of
Health (ES017177 to L.L.), the IUCRG grant (L.L.) and the IUPUI
startup fund (L.L. and J.P.). The NMR spectroscopy and MS facilities are
supported by the National Science Foundation (CHE-0619254, DBI-
0821661).
[1] a) B. A. Schweitzer, E. T. Kool, J. Am. Chem. Soc. 1995, 117, 1863–
1872; b) E. T. Kool, Acc. Chem. Res. 2002, 35, 936–943; c) E. T.
Kool, H. O. Sintim, Chem. Commun. 2006, 3665–3675.
[2] a) S. Moran, R. X. Ren, E. T. Kool, Proc. Natl. Acad. Sci. USA 1997,
94, 10506–10511; b) S. Moran, R. X. Ren, S. Rumney, E. T. Kool, J.
Am. Chem. Soc. 1997, 119, 2056–2057; c) E. T. Kool, Annu. Rev.
Biochem. 2002, 71, 191–219.
[3] a) T. A. Kunkel, R. Bebenek, Annu. Rev. Biochem. 2000, 69, 497–
529; b) M. Strerath, J. Cramer, T. Restle, A. Marx, J. Am. Chem.
Soc. 2002, 124, 11230–11231; c) M. F. Goodman, Proc. Natl. Acad.
Sci. USA 1997, 94, 10493–10495.
[4] a) P. S. Pallan, M. Egli, J. Am. Chem. Soc. 2009, 131, 12548–12549;
b) S. Xia, W. H. Konigsberg, J. Wang, J. Am. Chem. Soc. 2011, 133,
10003–10005.
[5] J. Cadet, P. Vigny, Photochemistry and the Nucleic Acids, Wiley,
New York, 1990, p. 437.
[6] a) P. R. Callis, Ann. Rev. Phys. Chem. 1983, 34, 329–357; b) U.
Streit, C. G. Bochet, Beilstein J. Org. Chem. 2011, 7, 525–542.
[7] a) G. Litwinienko, A. L. J. Beckwith, K. U. Ingold, Chem. Soc. Rev.
2011, 40, 2157–2163; b) F. Mueller, J. Mattay, Chem. Rev. 1993, 93,
99–117; c) J. Svoboda, B. Konig, Chem. Rev. 2006, 106, 5413–5430.
[8] a) D. Roca-Sanjuꢃn, G. Olaso-Gonzalez, I. Gonzaꢄlez-Ramirez, L.
Serrano-Andres, M. Merchan, J. Am. Chem. Soc. 2008, 130, 10768–
10779; b) M. Boggio-Pasqua, G. Groenhof, L. V. Schafer, H. Grub-
muller, M. A. Robb, J. Am. Chem. Soc. 2007, 129, 10996–10997;
c) L. Blancafort, A. Migani, J. Am. Chem. Soc. 2007, 129, 14540–
14541; d) W. J. Schreier, J. Kubon, N. Regner, K. Haiser, T. E.
Schrader, W. Zinth, P. Clivio, P. Gilch, J. Am. Chem. Soc. 2009, 131,
5038–5039.
1
H+], found 513.2. Compound 3: H NMR (D2O): d=1.60 (s, 3H), 1.73 (s,
3H), 1.95 (m, 1H), 2.27 (m, 1H), 2.52 (m, 1H), 2.57 (m, 1H), 3.10 (d, J=
11.0 Hz, 1H), 3.66 (b, 1H), 3.74 (dd, J=5.0, 12.5 Hz, 1H), 3.80 (dd, J=
6.5, 11.0 Hz, 1H), 3.90 (dd, J=2.5, 12.5 Hz, 1H), 4.00 (m, 1H), 4.01 (d,
J=11.0 Hz, 1H), 4.11 (m, 1H), 4.34 (t, J=7.5 Hz, 1H), 4.42 (dd, J=3.5,
8.0 Hz, 1H), 4.54 (d, J=8.0 Hz, 1H), 4.64 (t, J=7.0 Hz, 1H), 5.52 (s,
1H), 5.76 (d, J=4.0 Hz, 1H), 5.80 ppm (dd, J=2.5, 6.0 Hz, 1H);
13C NMR (D2O): d=21.5, 24.8, 37.7, 38.4, 38.7, 45.8, 52.1, 57.6, 60.4, 65.6,
71.1, 72.2, 78.9, 82.5, 83.9, 84.3, 112.0, 118.7, 134.9, 140.2, 152.7,
175.1 ppm; ESI-MS (positive mode) calcd for C22H30N2O10P: 513.2
[9] W. J. Schreier, T. E. Schrader, F. O. Koller, P. Gilch, C. E. Crespo-
Hernandez, V. N. Swaminathan, T. Carell, W. Zinth, B. Kohler, Sci-
ence 2007, 315, 625–629.
[10] G. Lin, L. Li, Angew. Chem. 2010, 122, 10122–10125; Angew. Chem.
Int. Ed. 2010, 49, 9926–9929.
ACHTUNGTRENNUNG
[M+H+], found 513.2.
Preparation of 5 in glycerol/H2O for NMR analysis: An aqueous solution
of 4 (10 mm; 20 mL) was titrated with NaOH (0.1m) to pH 7.0. After re-
moval of water under vacuum, the resulting sodium salt was dissolved in
glycerol/H2O (1:1, 25 mL, 4.36 mgmLÀ1). The solution was frozen in
[11] T. W. Kim, J. C. Delaney, J. M. Essigmann, E. T. Kool, Proc. Natl.
Acad. Sci. USA 2005, 102, 15803–15808.
[12] a) B. A. Schweitzer, E. T. Kool, J. Org. Chem. 1994, 59, 7238–7242;
b) see the Supporting Information.
7832
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 7823 – 7833

Thymine Isostere Biochemistry
FULL PAPER
[13] a) C. Desnous, B. R. Babu, C. Moriou, J. U. O. Mayo, A. Favre, J.
Wengel, P. Clivio, J. Am. Chem. Soc. 2008, 130, 30–31; b) C.
Moriou, M. Thomas, M. T. Adeline, M. T. Martin, A. Chiaroni, S.
Pochet, J.-L. Fourrey, A. Favre, P. Clivio, J. Org. Chem. 2007, 72,
43–50.
[14] K. M. Guckian, B. A. Schweitzer, R. X. F. Ren, C. J. Sheils, D. C.
Tahmassebi, E. T. Kool, J. Am. Chem. Soc. 2000, 122, 2213–2222.
[15] Note: the triplet signal observed suggests that in D2O, the H2’ and
H2’’ protons become equal. In contast, a dd signal was observed for
5 in D4-methanol, similar to those exhibited by 2 and 3 for this H1’
(see the Supporting Information).
[24] W. Saenger, Principles of Nucleic Acid Structure, Springer, New
York, 1984, p. 556.
[25] T. Ostrowski, J. C. Maurizot, M. T. Adeline, J. L. Fourrey, P. Clivio,
J. Org. Chem. 2003, 68, 6502–6510.
[26] C. E. Crespo-Hernꢅndez, B. Cohen, B. Kohler, Nature 2005, 436,
1141–1144.
[27] P. Yakovchuk, E. Protozanova, M. D. Frank-Kamenetskii, Nucleic
Acids Res. 2006, 34, 564–574.
[28] a) J.-C. Xuan, I. T. Weber, Nucleic Acids Res. 1992, 20, 5457–5464;
b) S.-G. Kim, L.-J. Lin, B. R. Reid, Biochemistry 1992, 31, 3564–
3574.
[16] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A.
Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A.
Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F.
Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M.
Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima,
et al. Gaussian 03, Revision C.02, Gaussian, Inc., Wallingford CT,
2004.
[17] P. J. Stephens, F. J. Devlin, C. F. Chabalowski, M. J. Frisch, J. Phys.
Chem. 1994, 98, 11623–11627.
[18] K. M. Guckian, T. R. Krugh, E. T. Kool, Nat. Struct. Biol. 1998, 5,
954–959.
[29] a) H. R. Drew, R. M. Wing, T. Takano, C. Broka, S. Tanaka, K. Ita-
kura, R. E. Dickerson, Proc. Natl. Acad. Sci. USA 1981, 78, 2179–
2183; b) F. D. Lewis, X. Liu, Y. Wu, S. E. Miller, M. R. Wasielewski,
R. L. Letsinger, R. Sanishvili, A. Joachimiak, V. Tereshko, M. Egli,
J. Am. Chem. Soc. 1999, 121, 9905–9906; c) X. Shui, L. McFail-
Isom, G. G. Hu, L. D. Williams, Biochemistry 1998, 37, 8341–8355.
[30] K. M. Guckian, T. R. Krugh, E. T. Kool, J. Am. Chem. Soc. 2000,
122, 6841–6847.
[31] V. R. Cooper, T. Thonhauser, A. Puzder, E. Schroder, B. I. Lundqv-
ist, D. C. Langreth, J. Am. Chem. Soc. 2008, 130, 1304–1308.
[32] J. Norberg, L. Nilsson, J. Am. Chem. Soc. 1995, 117, 10832–10840.
[33] D. G. T. Su, J. L. F. Kao, M. L. Gross, J.-S. A. Taylor, J. Am. Chem.
Soc. 2008, 130, 11328–11337.
[19] A. Mees, T. Klar, P. Gnau, U. Hennecke, A. P. M. Eker, T. Carell,
L.-O. Essen, Science 2004, 306, 1789–1793.
[20] a) J. Kemmink, R. Boelens, R. Kaptein, Eur. Biophys. J. 1987, 14,
293–299; b) F. E. Hruska, D. J. Wood, K. K. Ogilvie, J. L. Charlton,
Can. J. Chem. 1975, 53, 1193–1203.
[21] J. L. F. Kao, S. Nadji, J. S. Taylor, Chem. Res. Toxicol. 1993, 6, 561–
567.
[34] D. G. T. Su, H. Fang, M. L. Gross, J.-S. A. Taylor, Proc. Natl. Acad.
Sci. USA 2009, 106, 12861–12866.
[35] J. Yamamoto, K. Nishiguchi, K. Manabe, C. Masutani, F. Hanaoka,
S. Iwai, Nucleic Acids Res. 2011, 39, 1165–1175.
[36] M. J. S. Dewar, E. G. Zoebisch, E. F. Healy, J. J. P. Stewart, J. Am.
Chem. Soc. 1985, 107, 3902–3909.
[22] T. M. G. Koning, J. J. G. van Soest, R. Kaptein, Eur. J. Biochem.
1991, 195, 29–40.
[23] K. M. Guckian, E. T. Kool, Angew. Chem. 1997, 109, 2942–2945;
Angew. Chem. Int. Ed. Engl. 1997, 36, 2825–2828.
Received: March 11, 2012
Published online: May 15, 2012
Chem. Eur. J. 2012, 18, 7823 – 7833
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7833