Additional base-pair formation in DNA duplexes by a double-headed nucleotide

Article abstract of DOI:10.1002/chem.201103467

We have designed and synthesised a double-headed nucleotide that presents two nucleobases in the interior of a dsDNA duplex. This nucleotide recognises and forms Watson-Crick base pairs with two complementary adenosines in a Watson-Crick framework. Furthermore, with judicious positioning in complementary strands, the nucleotide recognises itself through the formation of a T:T base pair. Thus, two novel nucleic acid motifs can be defined by using our double-headed nucleotide. Both motifs were characterised by UV melting experiments, CD and NMR spectroscopy and molecular dynamics simulations. Both motifs leave the thermostability of the native dsDNA duplex largely unaltered. Molecular dynamics calculations showed that the double-headed nucleotides are accommodated in the dsDNA by entirely local perturbations and that the modified duplexes retain an overall B-type geometry with the dsDNA unwound by around 25 or 60°, respectively, in each of the modified motifs. Both motifs can be accommodated twice in a dsDNA duplex without incurring any loss of stability and extrapolating from this observation and the results of modelling, it is conceivable that both can be multiplied several times within a dsDNA duplex. These new motifs extend the DNA recognition repertoire and may form the basis for a complete series of double-headed nucleotides based on all 16 base combinations of the four natural nucleobases. In addition, both motifs can be used in the design of nanoscale DNA structures in which a specific duplex twist is required. Copyright

Full text of DOI:10.1002/chem.201103467

FULL PAPER
DOI: 10.1002/chem.201103467
Additional Base-Pair Formation in DNA Duplexes by a Double-Headed
Nucleotide
Charlotte S. Madsen,[a] Sarah Witzke,[a] Pawan Kumar,[a] Kushuma Negi,[a]
Pawan K. Sharma,[b] Michael Petersen,*[a] and Poul Nielsen*[a]
Abstract: We have designed and syn-
thesised a double-headed nucleotide
that presents two nucleobases in the in-
terior of a dsDNA duplex. This nucleo-
tide recognises and forms Watson–
Crick base pairs with two complemen-
and molecular dynamics simulations.
Both motifs leave the thermostability
of the native dsDNA duplex largely
unaltered. Molecular dynamics calcula-
tions showed that the double-headed
nucleotides are accommodated in the
dsDNA by entirely local perturbations
and that the modified duplexes retain
an overall B-type geometry with the
dsDNA unwound by around 25 or 608,
respectively, in each of the modified
motifs. Both motifs can be accommo-
dated twice in a dsDNA duplex with-
out incurring any loss of stability and
extrapolating from this observation and
the results of modelling, it is conceiva-
ble that both can be multiplied several
times within a dsDNA duplex. These
new motifs extend the DNA recogni-
tion repertoire and may form the basis
for a complete series of double-headed
nucleotides based on all 16 base combi-
nations of the four natural nucleobases.
In addition, both motifs can be used in
the design of nanoscale DNA struc-
tures in which a specific duplex twist is
required.
tary adenosines in
a Watson–Crick
framework. Furthermore, with judi-
cious positioning in complementary
strands, the nucleotide recognises itself
through the formation of a T:T base
pair. Thus, two novel nucleic acid
motifs can be defined by using our
double-headed nucleotide. Both motifs
were characterised by UV melting ex-
periments, CD and NMR spectroscopy
Keywords: DNA
recognition
·
DNA structures · nucleosides · mo-
lecular dynamics · base-pairing
Introduction
dsDNA duplex constitutes an excellent skeleton for supra-
molecular design.[2–4] For instance, duplex formation can be
used in the assembly of aromatic chromophores and aromat-
ic stacking can allow several consecutive artificial aromatic
moieties to be accommodated in the centre of the duplex
when attached to the natural backbone.[3] Furthermore,
DNA duplexes have been assembled in large frameworks,
like the so-called DNA origami,[4] and three-dimensional su-
perstructures, like an octahedron, as well as in nanomechan-
ical devices.[2] Ideally, the nanoscale engineer should have
a toolbox of DNA motifs with specific and well-character-
ised structural parameters like twist and bend to use in the
design of large DNA structures. For example, T:T mismatch-
es have been used to increase the flexibility of the dsDNA
duplex in the construction of circular dsDNA structures.[5]
Thus, to expand the repertoire of nanoscale DNA structures,
specific ꢀngstrçm-scale building blocks are needed. This
has prompted us and others to initiate studies of recognition
patterns both inside and on the surface of the DNA duplex
with a view to engineering new nucleic acid structural
motifs and functions by using double-headed nucleotides.[6–8]
As an example, we have synthesised the double-headed
nucleotide 1 with a methylene-linked thymine at the pro-S
position of the C-5’ carbon (Figure 1).[7] This position directs
the thymine into the minor groove and a systematic search
showed that it can interact with the thymine from another
nucleotide 1 if this is placed in the (À3)-position of the com-
plementary DNA strand.[7] However, the two thymines inter-
Nature uses the double-helical Watson–Crick DNA struc-
ture for the storage of genetic information. In a beautifully
simple scheme, thymine forms base pairs with adenine and
cytosine with guanine through two or three hydrogen bonds.
The four possible Watson–Crick base pairs are structurally
isosteric and stack one on top of another in a spiral-like ar-
rangement in the centre of the DNA duplex and on the out-
side sits the invariant sugar–phosphate backbone. The two
DNA strands are held together by van der Waals forces
from intra- and interstrand stacking and by hydrogen bond-
ing between complementary bases. This creates the iconic
dsDNA duplex with a twist of 368 between adjacent base
pairs and 10 base pairs per full helical turn.[1] With its regu-
lar structure and easily programmable formation, the
[a] Dr. C. S. Madsen, S. Witzke, Dr. P. Kumar, Dr. K. Negi,
Prof. Dr. M. Petersen, Prof. Dr. P. Nielsen
Nucleic Acid Center, Department of Physics and Chemistry
University of Southern Denmark, Campusvej 55
5230 Odense M (Denmark)
Fax : (+45)66158780
E-mail: pon@ifk.sdu.dk
[b] Prof. Dr. P. K. Sharma
Department of Chemistry, Kurukshetra University
Kurukshetra-136 119 (India)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201103467.
Chem. Eur. J. 2012, 00, 0 – 0
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
1
&
ÞÞ
These are not the final page numbers!

3
alkylation. The HMBC NMR spectrum showed JCH cou-
pling between the 2’-methylene protons and C-2 as well as
C-6 in the thymine and thus confirmed the N1-alkylation. In-
troduction of the uracil at C-1 was achieved by Vorbrꢂggen
coupling to afford a 5:1 mixture of the nucleoside 5 and its
a anomer in a yield of 64%. This indicates that the thymine
C-2 carbonyl does not give efficient anchimeric assistance
and the preference for the b nucleoside is probably due to
steric effects. The HMBC NMR spectrum shows coupling
between 1’-H and C-2 as well as C-6 in the uracil, thereby
confirming the formation of the double-headed nucleo-
side.[11,12] Separation of the two anomers was not possible at
this stage, but after subsequent simultaneous cleavage of all
the protecting groups in a hydrogenation reaction to give 6
followed by 5’-O-DMT protection, 7 was obtained as the de-
sired b anomer as a pure compound in a yield of 52% over
the two steps. Owing to the hydrophilicity of 6, it was best
not to purify 6. However, this was performed on a small
scale for a full NMR analysis. The ROESY NMR spectrum
of 7 confirms the b configuration as ROE contacts are ob-
served between 6-H (uracil) and 2’-H, between 1’-H and 4’-
H, and between 1’-H and the methylene protons. Finally,
phosphitylation gave the desired phosphoramidite 8 in
a yield of 83%, which corresponds to an overall yield of
20% from compound 3. This building block was then incor-
porated into oligonucleotides by standard automated solid-
phase DNA synthesis and pairs of complementary oligonu-
cleotides containing the modification, henceforth denoted
UT (Figure 1), were prepared.
Figure 1. Double-headed nucleotides 1 and 2.
act through stacking on the duplex surface in the minor
groove, which has no or little interaction specificity.[7] We
therefore turned our attention to nucleotide 2 in which the
thyminylmethylene is added at the C-2’ carbon (Figure 1)
giving the additional nucleobase the potential to enter into
the interior of a duplex to participate in base-pairing and
the transfer of information.
Results and Discussion
Synthesis: Nucleotide 2 was synthesised starting from d-
ribose, which in six known steps was converted into methyl
3,5-bis-O-(p-chlorobenzyl)-2-deoxy-2-hydroxymethyl-a-d-ri-
bofuranose (3; Scheme 1).[9] The thymine was the first nucle-
obase to be introduced by Mitsunobu reaction to give com-
pound 4. An N3-protected thymine[10] was used to ensure N1-
UT recognition: We first investigated the impact of a single
introduction of the double-headed nucleotide 2, UT, into
a duplex and subsequently the ability of the double-headed
nucleotide to recognise a complementary 5’-AA-3’ pair by
UV melting curve analysis (Table 1). For direct comparison
with UT, we used a 2’-deoxyuridine instead of a thymidine in
the unmodified sequences.[11] We noted that the incorpora-
tion of a UT instead of a U led to a decrease in duplex sta-
Table 1. Hybridisation data for duplexes that contain UT.
Entry
Duplex
Tm [8C][a]
5’-d(GCTCACUCTCCCA)
3’-d(CGAGTGAGAGGGT)
5’-d(GCTCACUT CTCCCA)
3’-d(CGAGTG A GAGGGT)
5’-d(GCTCACUTCTCCCA)
3’-d(CGAGTGAAGAGGGT)
5’-d(GCTCACUTCTCCCA)
3’-d(CGAGTGAXGAGGGT)
5’-d(GCTCAC UT CTCCCA)
3’-d(CGAGTGAAGAGGGT)
5’-d(GCTCAC UT CTCCCA)
3’-d(CGAGTGAXGAGGGT)
5’-d(GCTCAC U CTCCCA)
3’-d(CGAGTGAAGAGGGT)
1
52.0
2
3
4
5
6
7
45.0
53.0
43.0/41.0/46.5
X=T/C/G
51.0
46.0/43.0/50.0
X=T/C/G
Scheme 1. Synthesis of the double-headed nucleoside. Reagents and con-
ditions: a) N3-(benzyloxymethyl)thymine,[10] Ph3P, DEAD, THF (74%);
b) uracil, BSA, TMS-OTf, CH3CN (64%); c) H2, Pd(OH)2/C, MeOH;
d) DMT-Cl, pyridine (52%, two steps); e) NCACTHUNTRGENN(UG CH2)2OP(Cl)NACHUTGNTREN(NUGN iPr)2,
DIPEA, DCE (83%). DMT=4,4’-dimethoxytrityl, BOM=benzyloxy-
43.0
[a] Melting temperatures (Tm) were obtained from the maxima of the
first derivatives of the melting curves (A260 vs. temperature) recorded in
a medium salt buffer (Na2HPO4 (2.5 mm), NaH2PO4 (5.0 mm), NaCl
(100 mm), EDTA (0.1 mm), pH 7.0) using 1.5 mm strand concentrations.
methyl.
&
2
&
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Base-Pair Formation in DNA Duplexes
FULL PAPER
Table 2. Hybridisation data for duplexes with various zipper motifs of
UT.
bility of 78C (Table 1, entries 1 and 2). In contrast, when tar-
geting an AA dinucleotide step with the UT nucleotide, the
thermostability of the native DNA duplex was barely affect-
ed with a drop of only 28C (Table 1, entries 3 and 5). The in-
corporation of a UT instead of a U in a duplex with an A
bulge led to an increase in stability of 88C (Table 1, entries 5
and 7). This strongly indicates that both pyrimidines of the
UT form base pairs with both adenines in a 5’-UT:3’-AA
motif. We then tested the fidelity of the recognition with
a series of single nucleotide mismatches (Table 1, entries 4
and 6). In a Watson–Crick duplex, a T-X mismatch de-
creased the stability by around 6–128C with the T-G mis-
match being tolerated best (Table 1, entries 3 and 4). In the
context of the UT nucleotide, mismatches decreased the sta-
bility by around 1–88C (Table 1, entries 5 and 6), the rela-
tive stabilities of the mismatches following those of the
Watson–Crick duplex.
Entry
Duplex
Tm [8C][a]
A
X/Y=
T/T UT/T T/UT UT/UT dUT/dTU
5’-d(CGCATATXCGC)
3’-d(GCGYATAAGCG)
5’-d(CGCATAXTCGC)
3’-d(GCGYATAAGCG)
5’-d(CGCATATXCGC)
3’-d(GCGTAYAAGCG)
5’-d(CGCATAXTCGC)
3’-d(GCGTAYAAGCG)
5’-d(CGCAXATTCGC)
3’-d(GCGYATAAGCG)
5’-d(CGCAXATTCGC)
3’-d(GCGTAYAAGCG)
1 (À4)
45.0 34.0 30.0
45.0 38.0 30.0
45.0 34.0 38.0
45.0 38.0 38.0
45.0 39.0 30.0
45.0 39.0 38.0
n.t.[b]
n.t.[b]
39.0
42.0
41.0
45.5
23.0
23.5
32.0
33.0
33.5
34.0
2 (À3)
3 (À2)
4 (À1)
5 (À1)
6 (+1)
[a] See Table 1. [b] No transition was observed above 108C.
We then employed NMR spectroscopy to obtain atomic
details of the recognition. In particular, the observation of
imino proton signals in the H NMR spectra is direct proof
of hydrogen bonding as the imino protons then are protect-
ed from exchange with water. In a 10-mer duplex that con-
tains the 5’-UT:3’-AA motif, we observed eight imino signals
at 58C (Figure 2). The terminal base pairs were not ob-
temperatures of 30–398C, corresponding to decreases in sta-
bility of 6–158C compared with the unmodified duplex (the
T/T column). This fits with the decrease in Tm of 78C ob-
served in the first sequence context (Table 1). It is noticea-
ble that the largest decreases were observed for modifica-
tions towards the 3’-ends of the duplexes. Subsequently, vari-
ous zipper orientations of two double-headed nucleotides
were investigated (the UT/UT column). This created either
unstable duplexes (Table 2, entries 1 and 2) or duplexes that
are much less destabilised than would have been expected
by adding the thermal penalties for each of the correspond-
ing single-modified duplexes (Table 2, entries 3–6). For
these four entries, the Tm is higher for the zipper duplexes
than for each of the corresponding single-modified duplexes.
This clearly demonstrates a pronounced stabilising effect of
the second incorporation of a UT. We observed the most fa-
vourable interaction in a (+1) zipper in which the second
modified nucleotide is positioned in the opposite strand
shifted one position towards the 5’-end. This 5’-UTA:3’-AUT
motif exhibits an unchanged thermostability when compared
with the unmodified duplex (45.58C compared with 458C,
Table 2, entry 6). This should be compared with the decrease
in stability of 6–78C for a duplex with just one UT nucleo-
tide replacing a T. Note that compared with the 5’-UTA:3’-
ATU motif, the 5’-UTA:3’-AUT motif exhibits a striking in-
crease of around 118C in duplex stability (45.58C compared
with 348C, Table 2, entry 6). From these data it is clear that
the methylene attachment of the thymine base at the C-2’
carbon is strongly preferred in comparison with the usual at-
tachment at the sugar-phosphate backbone for thymine–thy-
mine recognition.
1
Figure 2. The imino regions of 1D 1H NMR spectra recorded at 58C
in 9:1 H2O:D2O. a) 5’-d(GCATAUTCAG):3’-d(CGTATAAGTC) and
b) 5’-d(CGCUTAGCG)2.
served due to end-fraying of the duplex. Two of the imino
signals reside in the usual chemical shift range for G:C
Watson–Crick pairs and six in the range for A:T and A:U
pairs. The imino resonances were assigned on the basis of
their NOE connectivities and two resonances could unam-
biguously be assigned to the UT nucleotide. Thus, the
double-headed nucleotide indeed forms two Watson–Crick
base pairs, as suggested by the UV melting data.
1
The imino region of the 1D H NMR spectrum of an 8-
mer self-complementary duplex with a central 5’-UTA:3’-
AUT motif displays only two imino resonances at 58C, both
of these in the G:C Watson–Crick base-pair region
(Figure 2). The observation of two G:C pairs shows that
a duplex is formed, but the lack of signals from A:U pairs
and the putative T:T pair show that they are broadened by
exchange. As the resonance from the methyl group of the
thymine base is not observed either, it is likely that the
Next, we investigated the ability of the UT nucleotide for
self-recognition by systematically varying the relative posi-
tions of UT nucleotides in both strands of
a duplex
(Table 2). We first tested the corresponding duplexes with
a single incorporation of UT instead of an unmodified T (the
UT/T and T/UT columns of Table 2), which led to melting
Chem. Eur. J. 2012, 00, 0 – 0
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
3
&
ÞÞ
These are not the final page numbers!

M. Petersen, P. Nielsen et al.
Table 3. Hybridisation data for duplexes with multiple incorporations of
UT.
broadening of lines is due to chemical exchange (see
below).
Entry
Duplex
Tm [8C][a]
The (À2) and (À1) zipper orientations of the two UT nu-
cleotides also demonstrate remarkably stable duplexes
(Table 2, entries 3–5, UT/UT column). This indicates the for-
mation of duplex systems that are combined motifs with the
additional thymines forming base pairs with complementary
adenines leaving either a central U:U mismatch pair in the
5’-AUT:3’-UTA motif (the (À1) zipper) or a U:T mismatch
pair in the 5’-ATUT:3’-UTAA motif (the (À2) zipper). Du-
plexes with unmodified UT dinucleotides at the same posi-
tions as the UTs (Table 2, the dUT/dUT column) exhibit no-
tably lower values of Tm for entries 3–5, which proves that
the compressed nature of the backbone leaves a much more
stable structure than the corresponding duplexes with the
same bases and mismatches on a natural backbone.
5’-d(CGCGUTUTCGCG)
3’-d(GCGCAAAAGCGC)
5’-d(CGCG UT UT CGCG)
3’-d(GCGCAAAAGCGC)
5’-d(CGCGUTAACGCG)
3’-d(GCGCAATUGCGC)
5’-d(CGCG UT AACGCG)
3’-d(GCGCAAUT GCGC)
5’-d(CGCGUTAUTACGCG)
3’-d(GCGCATUATUGCGC)
5’-d(CGCGUT AUT A CGCG)
3’-d(GCGC AUT AUT GCGC)
1
45.0
2
3
4
5
6
39.0
42.0[b]
43.0[b]
n.t.[b,c]
43.5[b]
[a] See Table 1. [b] A melting transition at 70–808C was also observed.[13]
[c] No duplex transition was observed above 108C.[13]
Having established that the UT nucleotide efficiently rec-
ognises an AA dinucleotide step and itself in a (+1) zipper,
the 5’-UTA:3’-AUT motif, we then examined if these struc-
tural motifs could be multiplied in a duplex. First, we noted
that two juxtaposed UT nucleotides recognise AAAA and
form a stable duplex albeit with a decrease in thermal stabil-
ity of 68C (Table 3, entries 1 and 2). On the other hand,
with two alternating UT nucleotides, and thus two adjoining
5’-UT:3’-AA motifs in opposite strands, the thermal stability
is slightly increased in comparison with the unmodified
duplex (Table 3, entries 3 and 4). Secondly, the presence of
two 5’-UTA:3’-AUT motifs with two T:T base pairs (two
(+1) zippers) in a duplex is also favourable, as proved by
a melting temperature of 43.58C; this compares with an un-
modified sequence showing no duplex transition at all
(Table 3, entries 5 and 6).[13] Thus, double incorporations of
two UT motifs are feasible in dsDNA duplexes.
pyrimidine mismatches are connected to the sugar-phos-
phate backbone through the normal deoxyribo linkage or
through the C-2’ linkage of a UT nucleotide.
Structure of the 5’-UT:3’-AA motif: To gain a deeper insight
into the duplex structures we performed molecular dynamics
(MD) simulations of duplexes with one or two 5’-UT:3’-AA
motifs (that is, Table 1, entry 5, a single incorporation, and
Table 3, entries 2 and 4, double incorporation into either the
same strand or different strands, respectively). All the simu-
lations were stable as shown by pertinent parameters, such
as energy, pressure, temperature and rmsds (root mean
square deviations), and the DNA duplexes remain stable
over 50 ns (Figure 3). In all cases, the UT nucleotide formed
U:A and T:A Watson–Crick base pairs with hydrogen-bond
occupancies close to those of normal Watson–Crick A:T
base pairs (see Tables S5–S7 in the Supporting Information).
This is in full accord with our experimental NMR data in
which the imino protons of both the uracil and thymine
bases are protected from exchange by hydrogen bonding. In
CD spectroscopy: As a first step to gauge the global struc-
tural impact of the UT nucleotide, we recorded CD spectra
of all the duplexes studied. These spectra in general display
features characteristic of B-type
duplexes with a negative band
at around 250 nm and positive
bands at around 220 and
280 nm (see the Supporting In-
formation).[14] The CD spectra
of fully Watson–Crick base-
paired duplexes, whether paired
by using natural bases or with
5’-UT:3’-AA motif(s), are rather
similar. For the series of zipper
motifs (Table 2), there is a con-
spicuous difference between the
(+1) zipper, that is, the 5’-
UTA:3’-UTA motif, and the
other zipper motifs in the inten-
sity of the band at around
275 nm. This shows, as noted
above, that it entails a structural
Figure 3. Representative snapshots from MD simulations of duplexes. a) Close up of the 5’-UT:3’-AA motif
(Table 1, entry 5). b) Close up of the 5’-UTA:3’-AUT motif (Table 2, entry 6, UT/UT). Multiple incorporations
of UT motifs. c) Two alternating 5’-UT:3’-AA motifs (Table 3, entry 4). d) Two adjoining 5’-UT:3’-AA motifs
(Table 3, entry 2). e) Two adjoining 5’-UTA:3’-AUT motifs (Table 3, entry 6). In (a) and (b), atoms are coloured
according to element type and hydrogen bonds are shown as dashed green lines. In (c), (d) and (e), nucleobas-
es are coloured cyan and the sugar-phosphate backbone blue except for UT nucleotides, which are shown in
difference whether pyrimidine– red.
&
4
&
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Base-Pair Formation in DNA Duplexes
FULL PAPER
addition, the NMR resonance positions indicate that two
Watson–Crick base pairs are formed. Taken together, these
experimental observations validate our computational struc-
ture. The incorporation of UT nucleotides entails an unwind-
ing of the helix to accommodate the additional T:A base
pair in the base-pair stack. With a single incorporation, the
helix is unwound by around 258, whereas double incorpora-
tions lead to an unwinding of around 50 and 708, depending
on whether the UT nucleotides are located in the same
strand or in opposite strands, respectively. The change in
duplex geometry is neatly absorbed by the sugar-phosphate
backbone that extends in the UT steps and snakes up in the
AA steps. These changes are accomplished by small changes
in the sugar puckers and changes in the e and z backbone
angles (see Tables S4 and S5 in the Supporting Information).
The changes in geometry are limited to the base pairs imme-
diately adjacent to the extra T:A pair, whereas the remain-
der of the duplex is unperturbed. To present the adenine
bases recognised by the UT nucleotides in a geometry suita-
ble for base-pairing, the glycosidic angles of the adenosines
are changed into high anti conformations (see Table S5). For
the double incorporations of the UT nucleotide, our simula-
tions qualitatively support the duplex stabilities obtained by
UV melting in as much as the duplex with two adjacent UT
modifications qualitatively appears to be strained, whereas
the duplex in which the two UT modifications of the 5’-
UT:3’-AA motifs are dispersed between the two strands
looks strain-free and regular (Figure 3). The backbone ge-
ometry is severely altered in these duplexes in comparison
with the unmodified DNA duplex (see Tables S4, S6 and S7
and Figures S7 and S8). However, when the two UT nucleo-
tides are placed in separate strands, the short UT backbone
is compensated by the juxtaposed long AA backbone in
Figure 4. Representative snapshots from MD simulations of duplexes
with quadruple 5-UT:3-AA motifs. a) Four adjoining 5’-UT:3’-AA motifs:
5’-d(CGCGUTUTUTUTCGCG):3’-d(CGCGAAAAAAAACGCG),
b) four alternating 5’-UT:3’-AA motifs: 5’-d(CGCGUTAAUTAACGCG)2,
c) top view of (a) and d) top view of (b). Nucleobases are coloured cyan
and the sugar-phosphate backbone blue except for UT nucleotides, which
are shown in red. Traces through the sugar-phosphate backbones are
shown with green ribbons.
servations correlate with the UV denaturation data, which
show that alternating 5’-UT:3’-AA motifs are accommodated
better than adjacent ones.
Conformation of the 5’-UTA:3’-UTA motif: As we lack de-
finitive proof of the T:T pair in the 5’-UTA:3’-AUT motif
from NMR spectroscopy, we used the molecular mechanics
Poisson–Boltzmann surface area (MM-PBSA) approach[15]
to evaluate the difference in Gibbs free energy for two dif-
ferent conformations of this motif with either the two thy-
mine bases stacking and forming a T:T pair or the two thy-
mine bases excluded completely from the base-pair stack by
using the duplex in Table 2, entry 6. In the latter simulation,
the two A:U pairs stack (see Figure S5 in the Supporting In-
formation). Nevertheless, the conformation with the T:T
pair is 17 kcalmolÀ1 more stable than that in which the thy-
mines are excluded from the stack (Table S3). The main
contribution to the favourable free energy comes from the
force-field term and hence from stacking and hydrogen
bonding. This, together with the melting temperatures of
Table 2, entry 6 and Table 3, entry 6, very strongly suggests
that the T:T pair is formed.
To validate the method, we performed a similar MM-
PBSA analysis on the 5’-UT:3’-AA motif and compared the
conformation in which the thymine base is rotated towards
the solvent with a fully stacked and base-paired conforma-
tion. If the thymine base is rotated towards the solvent, then
the uracil base is also prevented from stacking and is rotated
towards the solvent and stacks with the thymine base (see
Figure S5 in the Supporting Information). The conformation
with full stacking is favoured over the destacked conforma-
a
completely symmetrical manner. Besides leading to
a more regular duplex geometry, the hydrogen-bond occu-
pancies are also higher in this duplex than when the two UT
nucleotides are located in the same strand (see Tables S6
and S7).
To gauge the feasibility of larger tracts of modified
duplex, we carried out two further MD simulations of du-
plexes with quadruple UT incorporations in either the same
strand or alternating between the two strands. The trends
observed for double incorporations are exacerbated with
four UT nucleotides. If placed sequentially in one strand to
recognise eight adenines, the duplex adopts a quite distorted
structure in which the UT strand is elongated, almost linear,
and its nucleobases stack almost straight on top of each
other with the A strand winding around it (Figure 4). As
a consequence, the duplex has a very peculiar shape with
a large diameter. However, when the four UT nucleotides al-
ternate between the strands, the alternate extensions and
curls of the backbone compensate each other and the global
helical structure appears regular despite the large unwinding
(Figure 4). Hence, even with the incorporation of four UT
nucleotides, the duplex looks regular with a typical diameter
and the base pairs spanning the full width of the duplex as
in unmodified B-type DNA duplexes. These structural ob-
Chem. Eur. J. 2012, 00, 0 – 0
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
5
&
ÞÞ
These are not the final page numbers!

M. Petersen, P. Nielsen et al.
tion by a free energy of 12 kcalmolÀ1 (see Table S3). Thus,
for this motif, the MM-PBSA results are in full accord with
our NMR data, which show that both UT bases form
Watson–Crick base pairs with their cognate adenines. This
in turn validates the MM-PBSA results obtained for the 5’-
UTA:3’-AUT motif.
Conclusion
We have provided the first double-coding nucleotide 2 (UT)
that recognises nucleobases in the base-pair ladder in the
centre of a Watson–Crick duplex. Hybridisation data indi-
cate surprisingly stable duplexes when 5’-UT:3’-AA or 5’-
UTA:3’-AUT motifs are incorporated. In addition, multiple
incorporations of each of the motifs are also possible with-
out paying any penalty in duplex stability. Our MD simula-
tions have shown that the DNA sugar-phosphate backbone
accommodates the modified nucleotide by stretching or curl-
ing up as is needed. Based on the MD simulations, and the
fact that one Watson–Crick base pair is accommodated (the
T:A in the 5’-UT:3’-AA motif), we expect that all four base
pairs, that is, all 16 base combinations based on the double-
headed nucleotide structure 2, can be accommodated in this
way. The two structural motifs presented increase the range
of dsDNA motifs with a well-defined structure and may be
used as building blocks in nanoscale DNA scaffolds in
which full control of helical winding is desired. Furthermore,
with the double-headed nucleotides, a new framework for
transferring nucleic acid based molecular information can
be envisioned. Our future work will at first be directed to-
wards synthesising other base combinations and investigat-
ing the suitability of the double-headed nucleotides in poly-
merase reactions.
Structure of the 5’-UTA:3’-AUT motif: We also performed
MD simulations with a varying number of 5’-UTA:3’-AUT
motifs (Table 1, entry 6, with a single motif, and Table 3,
entry 6, with double motifs) and a duplex with quadruple
motifs (see the Supporting Information). The simulations
were monitored as above and were all stable on a timescale
of 50 ns. In all cases, the two A:U Watson–Crick base pairs
are formed with an additional T:T pair formed by the thy-
mine bases of the UT nucleotides although the hydrogen-
bond occupancy is slightly lower than observed for usual
Watson–Crick base pairs (see Tables S8–S9 in the Support-
ing Information). The backbone is extended by the interca-
lated T:T pair in both strands so the duplexes are length-
ened and unwound by around 308 in each step giving the 5’-
UTA:3’-AUT motif a very ladder-like appearance (Figure 3).
Again, the alterations are entirely local and apart from the
5’-UTA:3’-AUT element, the duplex is largely unaffected by
the modifications. The extension of the sugar-phosphate
backbone is achieved in much the same way as in the 5’-
UT:3’-AA motif (see Tables S5–S9 and Figures S8 and S9).
Notably, again the glycosidic angles of some adenosines are
found in the high anti range to present the adenine base for
base-pairing in the best possible manner.
Experimental Section
The central T:T pair can be formed with two different hy-
drogen-bonding schemes, the 3-H of one thymine bonding
to either the 2-O or 4-O of the other one. In one case, in the
double motif, we observed a change between the two
schemes. Thus, in a simulation of the T:T pairs over a total
period of 350 ns, we observed one shift between the two dif-
ferent T:T hydrogen-bonding possibilities, which corrobo-
rates our notion that NMR resonances are broadened by
fast chemical exchange.
In duplexes with multiple 5’-UTA:3’-AUT motifs, the AU
step between the 5’-UTA:3’-AUT motifs is also slightly un-
wound (by ꢀ108), giving such duplexes a very open appear-
ance with the nucleobases protruding somewhat into the
major groove (Figures 3 and S6 in the Supporting Informa-
tion). The 5’-UTA:3’-AUT motifs are twisted by around 228
relative to each other and the duplexes bend by around 108
into the minor groove at each 5’-UTA:3’-AUT motif. Extrap-
olating from the experimental melting temperatures of du-
plexes with one and two 5’-UTA:3’-AUT motifs and the re-
sults of our modelling, it appears quite likely that duplexes
incorporating several 5’-UTA:3’-AUT motifs are experimen-
tally feasible.
General: Reactions were performed under an atmosphere of nitrogen
when anhydrous solvents were used. Microwave-heated reactions were
performed on an EmrysTM Creator microwave oven. Column chromatog-
raphy was carried out in glass columns on using silica gel 60 (0.040–
0.063 mm). Flash column chromatography was performed on a Biotage
SP4TM Purification System. NMR spectra were recorded at 400 MHz for
1H NMR, 101 MHz for 13C NMR and 162 MHz for 31P NMR spectrosco-
py. The d values are reported in ppm relative to tetramethylsilane as in-
ternal standard (for 1H and 13C NMR) or relative to 85% H3PO4 as exter-
nal standard (for 31P NMR). Assignments of the NMR spectra are based
on 2D correlation spectroscopy (COSY, HSQC and HMBC spectra) and
follow standard carbohydrate and nucleoside style, that is, the carbon
atom next to a nucleobase is assigned C1’, etc. High-resolution ESI mass
spectra were recorded in positive-ion mode.
Synthesis of methyl 2-C-[N3-(benzyloxymethyl)thymin-1-yl]methyl-3,5-
bis-O-(4-chlorobenzyl)-2-deoxy-a-d-ribofuranose (4): To a solution of
compound 3[9] (700 mg, 1.64 mmol) in anhydrous THF (16 mL) was
added Ph3P (862 mg, 3.3 mmol), N3-(benzyloxymethyl)thymine[10]
(808 mg, 3.3 mmol) and a 40% solution of DEAD in toluene (1.5 mL,
3.3 mmol). The stirred mixture was heated with microwave irradiation at
1008C for 5 min and then allowed to cool. A saturated aqueous solution
of NaHCO3 (15 mL) was added and the mixture was extracted with
CH2Cl2 (3ꢃ20 mL). The combined organic layers were dried (Na2SO4)
and concentrated under reduced pressure. The residue was purified by
flash column chromatography (0–20% EtOAc/petroleum ether) to give
the product 4 (793 mg, 74%) as an oil. Rf =0.4 (50% EtOAc/petroleum
ether); 1H NMR (400 MHz, CDCl3, 258C, TMS): d=7.41–7.13 (m, 13H,
Ph), 6.98 (s, 1H, 6-H), 5.49 (s, 2H, NCH2O), 4.83 (d, J=4.9 Hz, 1H, 1’-
H), 4.71 (s, 2H, NCH2OCH2Ph), 4.57 (d, J=12.4 Hz, 1H, 3’-OCH2Ph),
4.48 (s, 2H, 5’-OCH2Ph), 4.34 (d, J=12.4 Hz, 1H, 3’-OCH2Ph), 4.30 (m,
1H, 4’-H), 4.00 (dd, J=13.8, 6.5 Hz, 1H, 2’-CH2a), 3.90 (dd, J=7.4,
1.9 Hz, 1H, 3’-H), 3.82 (dd, J=13.8, 8.5 Hz, 1H, 2’-CH2b), 3.47–3.39 (m,
&
6
&
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Base-Pair Formation in DNA Duplexes
FULL PAPER
4H, 5’-Ha, OCH3), 3.32 (dd, J=9.9, 6.3 Hz, 1H, 5’-Hb), 2.79 (m, 1H, 2’-
H), 1.81 ppm (s, 3H, CH3); 13C NMR (101 MHz, CDCl3, 258C, TMS): d=
163.84 (C-4), 151.66 (C-2), 140.59 (C-6), 138.06, 136.18, 136.14, 133.77,
133.66, 129.14, 129.05, 128.70, 128.66, 128.28, 127.63 (Ph), 109.21 (C-5),
104.49 (C-1’), 83.28 (C-4’), 79.23 (C-3’), 72.83 (5’-OCH2Ph), 72.31
(NCH2OCH2Ph), 71.13 (3’-OCH2Ph), 70.63 (NCH2O), 70.44 (C-5’), 55.19
(OCH3), 45.63 (C-2’), 45.37 (2’-CH2), 12.92 ppm (CH3); HRMS (ESI): m/
z: calcd for C34H36Cl2N2O7Na+: 677.1792 [M+Na]+; found 677.1783.
158.05 (DMT), 150.72 (C-2(T)), 150.43 (C-2(U)), 144.37 (DMT), 141.36
(C-6(T)), 139.71 (C-6(U)), 135.23, 134.93, 129.63, 127.89, 127.60, 126.69,
113.21 (DMT), 108.02 (C-5(T)), 101.84 (C-5(U)), 85.99 (C-1’), 85.64
(C(Ph)3), 85.15 (C-4’), 71.45 (C-3’), 63.75 (C-5’), 54.97 (OCH3), 45.45 (C-
2’), 44.29 (2’-CH2), 11.73 ppm (CH3); HRMS (ESI): m/z: calcd for
C36H36N4O9Na+: 691.2375 [M+Na]+; found: 691.2389.
Synthesis of 5’-O-(4,4’-dimethoxytrityl)-3’-O-[P-(b-cyanoethoxy)-N,N-di-
AHCTUNGTERGiNNUN sopropylaminophosphinyl]-2’-deoxy-2’-C-(thymin-1-yl)methyluridine (8):
Synthesis of 2’-C-[N3-(benzyloxymethyl)thymin-1-ylmethyl]-3’,5’-bis-O-
(4-chlorobenzyl)-2’-deoxyuridine (5): A mixture of compound 4 (559 mg,
0.85 mmol) and uracil (192 mg, 1.7 mmol) was co-evaporated with anhy-
drous CH3CN (2ꢃ4 mL) and then redissolved in the same solvent
(9 mL). N,O-Bis(trimethylsilyl) acetamide (BSA; 940 mL, 3.85 mmol) was
added and the mixture was stirred at reflux for 45 min. The stirred homo-
geneous solution was cooled to room temperature and a solution of tri-
methylsilyl triflate (386 mL, 2.1 mmol) in anhydrous CH3CN (2 mL) was
slowly added. The reaction mixture was stirred at room temperature for
20 h. The reaction was quenched by the addition of a saturated aqueous
solution of NaHCO3 (10 mL) and the mixture was stirred for 5 min.
Brine (10 mL) was added and the mixture was extracted with CH2Cl2 (3ꢃ
30 mL). The combined organic layers were dried (Na2SO4) and concen-
trated under reduced pressure. The residue was purified by column chro-
matography (0–5% iPrOH/CHCl3) to give the double-headed nucleoside
5 (402 mg, 64%, 5:1 ratio of epimers) as an oil. Rf =0.5 (10% iPrOH/
CHCl3); major isomer: 1H NMR (400 MHz, CDCl3): d=8.75 (s, 1H,
NH), 7.56 (d, J=7.8 Hz, 1H, 6-H(U)), 7.39–7.16 (m, 13H, Ph), 6.80 (s,
1H, 6-H(T)), 6.22 (d, J=9.0 Hz, 1H, 1’-H), 5.49 (m, 1H, 5-H(U)), 5.43
(s, 2H, NCH2O), 4.72–4.55 (m, 4H, NCH2OCH2Ph, 5’-OCH2Ph), 4.49 (s,
2H, 3’-OCH2Ph), 4.31 (m, 1H, 4’-H), 4.06 (d, J=5.5 Hz, 1H, 3’-H), 3.92
(dd, J=14.0, 7.9 Hz, 1H, 2’-CH2a), 3.82 (dd, J=14.0, 6.7 Hz, 1H, 2’-
CH2b), 3.71 (dd, J=10.3, 3.5 Hz, 1H, 5’-Ha), 3.54 (dd, J=10.3, 2.1 Hz,
2H, 5’-Hb), 3.11 (m, 1H, 2’-H), 1.78 ppm (s, 3H, CH3); major isomer:
13C NMR (101 MHz, CDCl3): d=163.55 (C-4(T)), 162.42 (C-4(U)),
151.40 (C-2(T)), 150.53 (C-2(U)), 139.93 (C-6(T)), 139.34 (C-6(U)),
137.87, 135.36, 135.11, 134.35, 134.13, 129.15, 129.02, 128.99, 128.93,
128.80, 128.30, 127.70, 127.68 (Ph), 109.67 (C-5(T)), 103.20 (C-5(U)),
86.68 (C-1’), 81.95 (C-4’), 80.07 (C-3’), 72.95 (5’-OCH2Ph), 72.20
(NCH2OCH2Ph), 70.91 (3’-OCH2Ph), 70.46 (NCH2O), 46.47 (2’-CH2),
45.98 (C-2’), 12.79 ppm (CH3); HRMS (ESI): m/z: calcd for
C37H36Cl2N4O8Na+ 757.1803 [M+Na]+; found: 757.1812.
Compound 7 (175 mg, 0.26 mmol) was co-evaporated with anhydrous 1,2-
dichloroethane (2ꢃ3 mL) and redissolved in the same solvent (2 mL).
DIPEA (912 mL, 5.2 mmol) was added and the solution was stirred at
room temperature. N,N-Diisopropylamino-b-cyanoethylphosphinochlor-
ite (175 mL, 0.786 mmol) was added and the reaction mixture was stirred
at room temperature for 3 h. Anhydrous EtOH (3 mL) was added and
the mixture was concentrated under reduced pressure. The residue was
purified by column chromatography (1% pyridine/0–5% MeOH/CH2Cl2)
to give the phosphoramidite 8 (188 mg, 83%) as a white foam. Rf = 0.6
(20% MeOH/CH2Cl2); 31P NMR (162 MHz, CDCl3): d=150.32,
150.23 ppm; HRMS (ESI): m/z: calcd for C45H53N6O10PNa+: 891.3453
[M+Na]+; found 891.3468.
Spectral data for 2’-deoxy-2’-C-(thymin-1-yl)methyluridine (6): Com-
pound
6 was purified by column chromatography (0–12% MeOH/
CH2Cl2) on a small scale for a full NMR analysis. 1H NMR (400 MHz,
DMSO): d=11.32 (brs, 1H, NH), 11.21 (brs, 1H, NH), 7.77 (d, J=
8.2 Hz, 1H, 6-H(U)), 7.36 (s, 1H, 6-H(T)), 6.02 (d, J=9.2 Hz, 1H, 1’-H),
5.62 (d, J=8.2, 1H, 5-H(U)), 5.58 (d, J=4.8 Hz, 1H, 3’-OH), 5.06 (t, J=
5.2 Hz, 1H, 5’-OH), 4.18 (t, J=4.8 Hz, 1H, 3’-H), 3.90–3.85 (m, 2H, 4’-H,
2’-CH2a), 3.75 (m, 1H, 2’-CH2b), 3.54–3.52 (m, 2H, 5’-H), 2.70 (m, 1H, 2’-
H), 1.69 ppm (s, 3H, CH3); 13C NMR (101 MHz, DMSO): d=163.98 (C-
4(T)), 162.66 (C-4(U)), 150.73 (C-2(T)), 150.57 (C-2(U)), 141.38 (C-
6(T)), 140.08 (C-6(U)), 108.07 (C-5(T)), 102.20 (C-5(U)), 87.01 (C-4’),
85.42 (C-1’), 71.23 (C-3’), 61.45 (C-5’), 45.70 (C-2’), 43.79 (2’-CH2),
11.74 ppm (CH3).
Synthesis of oligo-deoxynucleotides: The oligonucleotides (ONs) were
synthesised on an automated DNA synthesiser following the phosphor-
ACHUTNGRENUaNG mACHUTGNTRENiNUGN dite approach on a 0.2 mmol scale (LCAA-CPG support) by using the
modified phosphoramidite 8 as well as the corresponding commercial 2-
cyanoethyl phosphoramidites of the natural 2’-deoxynucleosides, includ-
ing 2’-deoxyuridine. The synthesis followed the regular protocol for the
DNA synthesiser, including extended coupling time (20 min) and 0.25m
pyridinium chloride (625 equiv) in CH3CN as the activator for 8. The
coupling yields for all the unmodified 2-cyanoethyl phosphoramidites
were >96% and for 8, >75%. The 5’-O-DMT-ON oligonucleotides were
removed from the solid support by treatment with concentrated aqueous
ammonia at 558C for 12 h. The ONs were purified by reversed-phase
HPLC by using a Waters 600 system with an Xterra MS C18 (10 mm,
7.8ꢃ10 mm) pre-column and an Xterra MS C18 (10 mm, 7.8ꢃ150 mm)
column. Buffer A: 0.05m triethylammonium acetate, pH 7.4. Buffer B:
75% MeCN/H2O (3:1, v/v). Program: 2 min 100% A, 100–30% A over
38 min, 10 min 100% B, 10 min 100% A. All oligonucleotides were de-
Synthesis of 5’-O-(4,4’-dimethoxytrityl)-2’-deoxy-2’-C-(thymin-1-yl)
ACHTUNGTNERmNUNG eth-
ACHTUNGTRENNUNGyluridine (7): Pd(OH)2/C (30 mg) was added to a stirred solution of com-
pound 5 (330 mg, 0.45 mmol) in MeOH (10 mL) and the suspension was
degassed with argon for 30 min and then with H2 for 1 h. The mixture
was then stirred at room temperature under an atmosphere of H2 for
16 h. The suspension was filtered through a layer of Celite, which was
washed with EtOH (30 mL). The combined filtrates were concentrated
under reduced pressure to give the crude unprotected double-headed nu-
cleoside 6 (231 mg) as an oil. Rf =0.3 (12% MeOH/CH2Cl2); HRMS
(MALDI-TOF): m/z: calcd for C15H18N4O7Na+: 389.1068 [M+Na]+;
found: 389.1015. The NMR data for a purified sample of 6 is given
below.
AHCTUNGERTGtNNUN ritylated by treatment with 80% aqueous acetic acid (100 mL) for 20 min
at room temperature and subsequently quenched with water (100 mL).
Sodium perchlorate (5m, 15 mL), sodium acetate (3m, 15 mL) and abs.
ethanol (1 mL) were then added and allowed to precipitate at À188C for
12–16 h. Finally, the ONs were washed with cold abs. ethanol (2ꢃ
300 mL), solvent residues were removed by heating (558C) under a flow
of nitrogen and the ONs were redissolved in water. The purity and con-
stitution of the synthesised ONs were verified by ion-exchange chroma-
tography and MALDI-MS analysis recorded in positive ion mode on
a PerSeptive Voyager STR spectrometer with 3-hydroxypicolinic acid as
matrix.
Without further purification, the crude residue was co-evaporated with
anhydrous pyridine (5 mL) and then redissolved in the same solvent
(10 mL). DMT-Cl (278 mg, 0.82 mmol) was added and the reaction mix-
ture was stirred at room temperature for 6 h. The solution was concen-
trated under reduced pressure and the residue was purified by column
chromatography (1% pyridine/0–12% MeOH/CH2Cl2) to give compound
7 (175 mg, 52% over two steps) as a white solid. Rf =0.3 (20% MeOH/
CH2Cl2); 1H NMR (400 MHz, DMSO): d=11.37 (d, J=2.0 Hz, 1H,
NH(U)), 11.30 (s, 1H, NH(T)), 7.53 (d, J=8.2 Hz, 1H, 6-H(U)), 7.38–
7.31 (m, 5H, 6-H(T), DMT), 7.26–7.18 (m, 4H, DMT), 6.91 (d, J=
8.9 Hz, 4H, DMT), 6.06 (d, J=9.1 Hz, 1H, 1’-H), 5.67 (d, J=4.7 Hz, 1H,
3’-OH), 5.23 (dd, J=8.1, 2.0 Hz, 1H, 5-H(U)), 4.28 (t, J=5.1 Hz, 1H, 3’-
H), 4.03–3.94 (m, 2H, 4’-H, 2’-CH2a), 3.81–3.69 (m, 7H, 2’-CH2b, 2
OCH3), 3.22 (dd, J=10.5, 4.1 Hz, 1H, 5’-Ha), 3.12 (dd, J=10.5, 3.3 Hz,
1H, 5’-Hb), 2.90 (m, 1H, 2’-H), 1.69 ppm (d, J=0.8 Hz, 3H, CH3);
13C NMR (101 MHz, DMSO): d=164.01 (C-4(T)), 162.43 (C-4(U)),
Thermal denaturation studies: The ONs were thoroughly mixed in
medium salt buffer (0.1 mm EDTA, 100 mm NaCl adjusted to pH 7.0 by
10 mm NaH2PO4/5 mm Na2HPO4). ON concentrations of 1.5 mm were de-
termined by using the extinction coefficient stated at www.idtdna.com.
For the double-headed nucleotide 2, the extinction coefficient was as-
sumed equal to two thymines. The samples were denatured by heating to
808C and subsequent cooling to 108C. Quartz optical cells with a path
Chem. Eur. J. 2012, 00, 0 – 0
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
7
&
ÞÞ
These are not the final page numbers!

M. Petersen, P. Nielsen et al.
length of 10.0 mm were used. UV absorption spectra (A260 vs. tempera-
ture) were recorded with Varian Cary 3E UV/VIS spectrometer
ed NPT MD at 300 K with a time step of 2 ps. The reference pressure
was 1 atm, the barostat coupling parameter 2.0 ps and the thermostat
coupling parameter 1.0 ps. Production runs were carried out for 50 ns
with parameters as for the last step of the equilibration except for the
pressure and temperature coupling parameters which were both 5.0 ps.
a
equipped with a Varian Cary temperature controller. The temperature of
the denaturation experiments ranged from 10–758C, performed with
a temperature ramp of 1.08CminÀ1. Thermal denaturation temperatures
(Tm) were determined as the maximum of the first derivatives of the ab-
sorbance-versus-temperature curve; the values reported herein are an
average of at least two measurements within Æ0.58C. The melting curves
were found to be reversible.
Analysis of trajectories: Hydrogen bonds, sugar-puckering, torsion angles
and rmsds were determined with Carnal from the AMBER program
suite. Hydrogen bonds were considered formed when the distance
AHCTUNGTREGbNNUN etween the donor and acceptor heavy atoms was below 3.5 ꢀ and the
À
CD spectroscopy: The ONs were thoroughly mixed in medium salt
buffer (0.1 mm EDTA, 100 mm NaCl adjusted to pH 7.0 by 10 mm
NaH2PO4/5 mm Na2HPO4), heated to 908C and cooled to room tempera-
ture. Quartz optical cells with a path length of 5.0 mm were used. The
CD spectra (200–350 nm) were recorded on a JASCO J-815 CD spec-
trometer as an average of 10 scans in steps of 2.0 nm with a scan speed of
100 nmminÀ1 and at 3.0 mm concentrations of each strand.
D H···A angle smaller than 308. Helix parameters were analysed with
Curves5.1 by using a single helix axis for the double helix.[25] For cases in
which end-fraying was observed, the terminal base pairs were omitted
from the analysis.
Free-energy calculations: To evaluate the difference in free energy be-
tween conformations of the 5’-UTA:3’-AUT motif with either the thy-
mines stacking and hydrogen bonding in a T:T pair or both turned to-
wards the solvent, we performed a further 50 ns simulation on entry 9
from Table S4 in the Supporting Information but with both UT thymine
bases destacked and turned towards the surrounding solvent. In a similar
manner, we performed a 50 ns simulation on entry 5 from Table S4 with
the thymine base of the UT nucleotide turned towards the solvent rather
than being involved in Watson–Crick hydrogen bonding with its cognate
adenine. In the MM-PBSA approach, the relevant energies were aver-
aged over 2251 snapshots, one every 20 ps from the last 45 ns of the tra-
jectories, stripped for explicit water molecules and ions. To calculate the
gas-phase energy and the solvation free energy, the perl script
mm pbsa.pl in the AMBER9 suite was used. The gas-phase energy was
found from the force-field energies by using a dielectric constant of 1.0.
The polar contribution to the solvation free energy was found by solving
the linear Poisson–Boltzmann (PB) equation with 1000 steps of iteration.
The ionic strength was set to 100 mm, the internal dielectric constant to
1.0 and the external dielectric constant to 80.0. The non-polar solvation
free energy was calculated by using the solvent-accessible surface area
(SASA), a surface tension of 0.0072 kcalꢀÀ2 molÀ1 and no correcting
offset. The entropic contribution was estimated by using the nmode pro-
gram in AMBER9 assuming the rigid rotor and harmonic approximation
to the normal modes at 300 K. To calculate the entropy, the stripped
snapshots were prepared in the same way as described above. These
structures were minimised for 50000 steps or until the root-mean-square
of the elements of the gradient vector was 10À4 kcalmolÀ1 ꢀÀ1 by using
a distance-dependent dielectric constant of 4r and a non-bonded cutoff of
99 ꢀ and then minimised further by the Newton–Raphson method up to
100 steps until the root-mean-square of the elements of the gradient
Molecular modelling
Setting up the UT library file: The modified nucleoside was built in the
xleap module of AMBER with the two nucleobases stacking and the ge-
ometry optimised at the HF/6-31G* level of theory by using Gaussi-
an 03.[16] The backbone conformation of the optimised nucleoside re-
mained in a duplex-like geometry through the geometry optimisation.
Atomic charges were calculated by using the RESP methodology[17] with
the native charges from the force field of Cornell and co-workers for the
backbone and sugar atoms except for C-1’, C-2’, 1’-H and 2’-H.[18,19] In
this manner we were consistent with the charge derivation of the native
atomic charges. Atom types and atomic charges for UT are shown in
Table S3 in the Supporting Information.
Setting up the deoxy-U library file: The methyl group of thymidine was
changed to a hydrogen (atom type HA) in xleap and atomic charges
were calculated in a manner consistent with the charge derivation in
AMBER, as described by Kollman and co-workers.[19]
Building UT-modified duplexes: All starting coordinates were generated
by using idealised B-DNA geometries. For duplexes containing UT resi-
dues, helices with two thymidines at the UT site(s) were built initially and
the resulting pdb file edited by deleting the second thymidine(s) and the
methyl group of the first thymidine(s) and changing the name of the UT
unit appropriately to be recognised by the LEaP module of AMBER.
For duplexes containing deoxy-U residues, helices with a thymidine at
the dU were built, the methyl group deleted and the residue name
changed appropriately. In LEaP, net-neutralising Na+ ions were added
and the whole system was surrounded by a truncated octahedron of
TIP3P waters with a minimum distance of 10.0 ꢀ from the helix to the
edges of the box.
vector was below 10À5 kcalmolÀ1 ꢀÀ1
.
MD simulations: All MD simulations were carried out with the
AMBER9 program suite[20,21] by using the ff99[22] and ff99-bsc0[23] param-
eters for nucleic acids and ions. The TIP3P water model was used.[24] For
UT-modified systems, an initial energy minimisation was performed with
only the UT residue and the 3’-following nucleotide free to move. Har-
monic positional restraints with a force constant of 500 kcalmolÀ1 ꢀÀ2
were applied to the remainder of the system. This energy minimisation
relieved the strain on the modified backbone. Afterwards, harmonic posi-
tional restraints with a force constant of 500 kcalmolÀ1 ꢀÀ2 were applied
to the entire duplex. The system was minimised for 500 steps of steepest
descent and 500 steps of conjugate gradient minimisation. Finally, a fur-
ther 1000 steps of steepest descent and 1500 steps of conjugate gradient
minimisation were carried out with the entire system free to move.
Acknowledgements
The project was supported by The Danish National Research Founda-
tion, The Danish Research Agencyꢄs Programme for Young Researchers,
The Danish Council for Independent Research (Natural Sciences (FNU)
and Technology and Production Sciences (FTP)), the Villum Kann Ras-
mussen Foundation and the Danish Center for Scientific Computing.
In the MD equilibration, the SHAKE algorithm was applied with a 1 fs
time step. The non-bond cutoff was 9 ꢀ and the particle mesh Ewald
method with default parameters was used to calculate long-range electro-
static interactions. The temperature of the system was raised from 0 to
300 K over 20 ps at constant volume by using the Berendsen thermostat
with a 0.2 ps coupling parameter and by applying harmonic positional re-
straints with a force constant of 10 kcalmolÀ1 ꢀÀ2 to the DNA molecules.
Centre-of-mass movement was removed every thousandth step. Subse-
quently, a 10 ps MD simulation was carried out at 300 K at constant
volume and temperature with parameters as in the first round of equili-
bration. The final step of the equilibration consisted of 200 ps unrestrain-
[1] W. Saenger, Principles of Nucleic Acid Structure, Springer, New
York 1984.
[2] a) M. Endo, H. Sugiyama, ChemBioChem 2009, 10, 2420–2443;
b) N. C. Seeman, Annu. Rev. Biochem. 2010, 79, 65–87.
[3] a) V. L. Malinovskii, D. Wenger, R. Hꢅner, Chem. Soc. Rev. 2010,
39, 410–422; b) T. J. Bandy, A. Brewer, J. R. Burns, G. Marth, T.
Nguyen, E. Stulz, Chem. Soc. Rev. 2011, 40, 138–148.
[4] T. Tørring, N. V. Voigt, J. Nangreave, H. Yan, K. V. Gothelf, Chem.
Soc. Rev. 2011, 40, 5636–5646.
[5] R. Kumar, A. El-Sagheer, J. Tumpane, P. Lincoln, L. M. Wilhelms-
son, T. J. Brown, J. Am. Chem. Soc. 2007, 129, 6859–6864.
&
8
&
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Base-Pair Formation in DNA Duplexes
FULL PAPER
[6] a) S. L. Pedersen, P. Nielsen, Org. Biomol. Chem. 2005, 3, 3570–
3575; b) M. S. Christensen, A. Bond, P. Nielsen, Org. Biomol. Chem.
2008, 6, 81–91; c) A. S. Jørgensen, K. I. Shaikh, G. Enderlin, E.
Ivarsen, S. Kumar, P. Nielsen, Org. Biomol. Chem. 2011, 9, 1381–
1388.
[7] a) M. S. Christensen, C. M. Madsen and P. Nielsen, Org. Biomol.
Chem. 2007, 5, 1586–1594; b) C. Andersen, P. K. Sharma, M. S.
Christensen, S. I. Steffansen, C. M. Madsen, P. Nielsen, Org. Biomol.
Chem. 2008, 6, 3983–3988; c) K. I. Shaikh, C. S. Madsen, L. J. Niel-
sen, A. S. Jørgensen, H. Nielsen, M. Petersen, P. Nielsen, Chem.
Eur. J. 2010, 16, 12904–12919.
[8] a) C. Switzer, T. P. Prakash, Y. Ahn, Bioorg. Med. Chem. Lett. 1996,
6, 815–818; b) T. Wu, M. Froeyen, G. Schepers, K. Mullens, J. Ro-
zenski, R. Busson, A. Van Aershot, P. Herdewijn, Org. Lett. 2004, 6,
51–54; c) T. Wu, K. Nauwelaerts, A. Van Aershot, M. Froeyen, E.
Lescrinier, P. Herdewijn, J. Org. Chem. 2006, 71, 5423–5431; d) T.
Umemoto, J. Wengel, A. S. Madsen, Org. Biomol. Chem. 2009, 7,
1793–1797; e) P. Kielkowski, R. Pohl, M. Hocek, J. Org. Chem.
2011, 76, 3457–3462.
[9] N.-S. Li, J. A. Piccirilli, J. Org. Chem. 2004, 69, 4751–4759.
[10] O. R. Ludek, C. Meier, Eur. J. Org. Chem. 2006, 941–946.
[11] Uracil was chosen as the nucleobase for C-1’ instead of thymine be-
cause a linear synthetic strategy starting from uridine was originally
approached[12] and to allow direct comparison with other double-
headed uridine analogues.[6]
[16] Gaussian 03, Revision D.02, M. J. Frisch, G. W. Trucks, H. B. Schle-
gel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgom-
ery, Jr., T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S.
Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani,
N. Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K.
Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,
O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian,
J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E.
Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W.
Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J.
Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C.
Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari,
J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cio-
slowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaro-
mi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng,
A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W.
Chen, M. W. Wong, C. Gonzalez, J. A. Pople, Gaussian, Inc., Wall-
ingford CT, 2004.
[17] C. I. Bayly, P. Cieplak, W. D. Cornell, P. A. Kollman, J. Phys. Chem.
1993, 97, 10269–10280.
[18] W. D. Cornell, P. Cieplak, C. I. Bayly, I. R. Gould, K. M. Merz,
D. M. Ferguson, D. C. Spellmeyer, T. Fox, J. W. Caldwell, P. A. Koll-
man, J. Am. Chem. Soc. 1995, 117, 5179–5197.
[19] P. Cieplak, W. D. Cornell, C. Bayly, P. A. Kollman, J. Comput.
Chem. 1995, 16, 1357–1377.
[12] A preliminary synthesis of the nucleoside has been presented in
a conference proceeding: C. Andersen, S. L. Pedersen, P. Nielsen,
Nucleosides Nucleotides & Nucleic Acids 2007, 26, 1435–1438.
[13] The self-complementary sequences (Table 3, entries 3–6) showed
both duplex and hairpin transitions; the latter melting at 70–808C.
The presence of duplex transitions was proved by performing melt-
ing experiments with two-strand concentrations (1.5 and 0.375 mm).
The transitions assigned to the duplexes were reduced by 1.5–2.58C
upon lowering the strand concentration whereas the hairpin transi-
tions did not change (see Table S1 in the Supporting Information).
In general, the modified sequences demonstrated sharper duplex
transitions than the unmodified ones and in the case of Table 3,
entry 5, only a hairpin was detected.
[20] D. A. Pearlman, D. A. Case, J. W. Caldwell, W. S. Ross, T. E. Cheat-
ham III, S. DeBolt, D. M. Ferguson, G. L. Seibel, P. A. Kollman,
Comp. Phys. Commun. 1995, 91, 1–41.
[21] D. A. Case, T. E. Cheatham, T. Darden, H. Gohlke, R. Luo, K. M.
Merz, A. Onufriev, C. Simmerling, B. Wang, R. J. Woods, J. Comput.
Chem. 2005, 26, 1668–1688.
[22] J. M. Wang, P. Cieplak, P. A. Kollman, J. Comput. Chem. 2000, 21,
1049–1074.
[23] A. Pꢆrez, I. Marchan, D. Svozil, J. Sponer, T. E. Cheatham, C. A.
Laughton, M. Orozco, Biophys. J. 2007, 92, 3817–3829.
[24] W. L. Jorgensen, J. Chandrasekhar, J. D. Madura, R. W. Impey,
M. L. Klein, J. Chem. Phys. 1983, 79, 926–935.
[25] a) R. Lavery, H. Sklenar, J. Biomol. Struct. Dyn. 1988, 6, 63–91;
b) R. Lavery, H. Sklenar, J. Biomol. Struct. Dyn. 1989, 7, 655–667.
[14] A. Rodger, B. Nordꢆn, Circular Dichroism and Linear Dichroism,
Oxford University Press, Oxford, 1997.
[15] a) P. A. Kollman, I. Massova, C. Reyes, B. Kuhn, S. Huo, L. Chong,
M. Lee, T. Lee, Y. Duan, W. Wang, Acc. Chem. Res. 2000, 33, 889–
897; b) J. Srinivasan, T. E. Cheatham, P. Cieplak, P. A. Kollman,
D. A. Case, J. Am. Chem. Soc. 1998, 120, 9401–9409.
Received: November 4, 2011
Revised: March 13, 2012
Published online: && &&, 0000
Chem. Eur. J. 2012, 00, 0 – 0
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
9
&
ÞÞ
These are not the final page numbers!

M. Petersen, P. Nielsen et al.
DNA Structures
C. S. Madsen, S. Witzke, P. Kumar,
K. Negi, P. K. Sharma, M. Petersen,*
P. Nielsen* ..................... &&&& — &&&&
Additional Base-Pair Formation in
DNA Duplexes by a Double-Headed
Nucleotide
Two heads are better than one! We
show that a double-headed nucleotide
that presents two nucleobases in
a DNA duplex recognises and forms
Watson–Crick base pairs with two
complementary adenines (see figure).
Furthermore, with judicious position-
ing, the nucleotide recognises itself
through the formation of a T:T pair.
These nucleic acid motifs extend the
DNA recognition repertoire and may
form the basis for a complete series of
double-headed nucleotides based on
all 16 base combinations of the four
natural nucleobases.
&
10
&
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!