Base-Pairing Properties of Double-Headed Nucleotides

Article abstract of DOI:10.1002/chem.201901077

Nucleotides that contain two nucleobases (double-headed nucleotides) have the potential to condense the information of two separate nucleotides into one. This presupposes that both bases must successfully pair with a cognate strand. Here, double-headed nucleotides that feature cytosine, guanine, thymine, adenine, hypoxanthine, and diaminopurine linked to the C2′-position of an arabinose scaffold were developed and examined in full detail. These monomeric units were efficiently prepared by convergent synthesis and incorporated into DNA oligonucleotides by means of the automated phosphoramidite method. Their pairing efficiency was assessed by UV-based melting-temperature analysis in several contexts and extensive molecular dynamics studies. Altogether, the results show that these double-headed nucleotides have a well-defined structure and invariably behave as functional dinucleotide mimics in DNA duplexes.

Full text of DOI:10.1002/chem.201901077

A Journal of
Accepted Article
Title: Base Pairing Properties of Double-Headed Nucleotides
Authors: Mick Hornum, Julie Stendevad, Pawan Kumar Sharma,
Pawan Kumar, Rasmus Blaaholm Nielsen, Michael Petersen,
and Poul Nielsen
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201901077
Link to VoR: http://dx.doi.org/10.1002/chem.201901077
Supported by

10.1002/chem.201901077
Chemistry - A European Journal
FULL PAPER
Base-Pairing Properties of Double-Headed Nucleotides
Mick Hornum,[a] Julie Stendevad,[a] Pawan K. Sharma,[b] Pawan Kumar,[a] Rasmus B. Nielsen,[a] Mi-
chael Petersen[a] and Poul Nielsen*[a]
Dedication ((optional))
Abstract: Nucleotides that contain two nucleobases (double-headed
nucleotides) have the potential to condense the information of two
separate nucleotides into one. This presupposes that both bases must
successfully pair with a cognate strand. Herein, we develop and ex-
amine in full detail double-headed nucleotides that feature cytosine,
guanine, thymine, adenine, hypoxanthine and diaminopurine linked to
the C2′-position of an arabinose scaffold. These monomeric units
were efficiently prepared via convergent synthesis and incorporated
into DNA oligonucleotides by means of the automated phospho-
ramidite method. Their pairing efficiency were assessed by UV-based
melting temperature analysis in several contexts and extensive mo-
lecular dynamics studies. Altogether, our results show that these dou-
ble-headed nucleotides have a well-defined structure and invariably
behave as functional dinucleotide mimics in DNA duplexes.
uptake, since the same molecular information can be delivered
using a shorter sequence and decreased polyanionic charge.
The C2′-exo position of β-nucleosides emerges as the most
reasonable site for introducing an additional base and enable dual
Watson–Crick contacts.18 In our preceding studies, we have
gauged the potential of three designs (double-headed β-ᴅ-ri-
bosides,25 β-ᴅ-deoxyribosides26,27 and β-ᴅ-arabinosides;24,28,29
see Figure 1) in terms of their synthetic feasibility and dinucleotide
potential. For convenience, uracil has invariably been used as the
C1′-linked base.
Introduction
Figure 1. Three double-headed nucleoside variants. B = nucleobases.
Besides its renowned role in heredity, DNA has emerged as a
valuable tool for a wide range of purposes. Well-studied examples
include therapeutics,1–3 diagnostics,4 genetic engineering,5 foren-
sics,6 phylogenetics,7 data storage,8,9 nanostructures,10,11 nano-
mechanical devices,12,13 etc. In most cases, the versatility of DNA
arises from its predictable, yet malleable, structure and program-
mable assembly. In addition, chemically modified DNA is a very
mature field of research, and thousands of synthetic derivatives
have been developed to suit specific needs, such as tailored
physical properties,14 enhanced drug pharmacokinetics15 or in the
desire to construct genetic systems based on alternative chemical
platforms.16,17 A unique type of DNA modification is the so-called
double-headed nucleotides, which are single nucleotides holding
two nucleobases.18–23 If both nucleobases are oriented towards
the duplex core for partaking in base pairing, the double-headed
nucleotide is turned into a functional dinucleotide. Such bifurcated
structures can function to condense the genetic information in
DNA by comprising two Watson–Crick base pairs per each nucle-
otide unit.24 This design can be used to e.g. improve cellular
Incorporation of the double-headed ribo structure turned out to
destabilize DNA duplexes25 indicating that the C2′-linked nucleo-
base is not well-accommodated. Contrarily, the deoxyribo struc-
ture displayed potent recognition of complementary sequences
indicative of dinucleotide behavior.26,27 This highlighted the im-
portance of using a shorter linker and/or the importance of the
sugar puckering. Nevertheless, this building block could only be
obtained following a long synthetic route with poor yielding steps,
which foiled its development. In the present work, we focus on the
arabino structure, which is essentially isosteric to the deoxyribo
structure, yet we have shown that it can be obtained in remarkably
improved yields over fewer steps.24,28 Favorably, the 2′-OH group
is shielded and does not necessitate a protecting group for oligo-
nucleotide synthesis.28
In this report, we take advantage of the synthetic ease of-
fered by the arabino design to gain a deeper insight into double-
headed nucleotides’ base-pairing efficiency compared to natural
dinucleotides. Accordingly, we set out to investigate a full set of
canonical and modified DNA bases linked to the C2′-position (Fig-
ure 2). Additionally, we accompaniment the experimental work
with atomistic molecular dynamics computations for assessing
the double-headed nucleotides’ binding mode and impact on the
helix’s tertiary structure. This study presents the first synthesis of
double-headed nucleotides containing hypoxanthine (H) and 2,6-
diaminopurine (D). As the nucleobase of inosine found in the wob-
ble position of many tRNAs, H has often been exploited as a uni-
versal base in various applications such as hybridization
probes30–32 or primers.33 In the light of its promiscuous base pair-
ing, we rationalized that H could potentially serve as an indicator
for better understanding how the additional nucleobase is posi-
tioned in the base stack. The weakly fluorescent D, a selective
[a]
Dr. M. Hornum, J. Stendevad, Dr. P. Kumar, R. B. Nielsen, Dr. M.
Petersen, Dr. P. Nielsen
Department of Physics, Chemistry and Pharmacy
University of Southern Denmark
Campusvej 55, DK-5230 Odense M
E-mail: pouln@sdu.dk
[b]
Dr. P. K. Sharma
Department of Chemistry
Kurukshetra University
Kurukshetra, Haryana 136119, India
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/chem.201901077
Chemistry - A European Journal
FULL PAPER
Figure 2. DNA building blocks of this study.
analog of A, has often been introduced to improve binding affinity
by potentially allowing three hydrogen bonds to T/U,34,35 rendering
it useful in e.g. primers36 or hybridization probes.37 However, this
enthalpic gain is sometimes offset by the entropic penalty caused
by disruption of the spine of hydration around the minor groove of
DNA duplexes.38 Also, the practical use of D in duplexes is ham-
pered by its susceptibility to depurination during oligonucleotide
synthesis.39 For the present case, depurination of the second nu-
cleobase is irrelevant, and we conjectured that D could be advan-
tageous in improving duplex stability (when paired-up with T) and
recognition fidelity. In this regard, the extra 2-amino group com-
pared to A can serve as a useful indicator of base pairing through
Watson–Crick faces.
the latter adhered strongly to the nucleoside. Excess Et3N·3HF
was neatly decomposed into volatiles (acetylene, TMSF and
Et3N) when treated with trimethylsilylacetylene. Finally, the 5′ and
3′ hydroxyls were routinely dimethoxytritylated and phosphitylated.
Due to solvation difficulties with 5D in pyridine and thus reduced
reactivity, a DMSO/2,6-lutidine solvent system45 was used to sig-
nificantly improve the rate of tritylation. In this way, all nucleoside
phosphoramidites 7 were successfully obtained in overall yields
of 27% (UT), 40% (UC), 25% (UA), 30% (UG), 40% (UH) and 24%
(UD) starting from 3.
All double-headed phosphoramidites 7 were successfully in-
corporated into oligonucleotides on an automated solid-phase
DNA synthesizer when activated with 1H-tetrazole. Coupling effi-
ciencies exceeded 95% for all monomers. After completion of the
oligonucleotide synthesis, the O-allyl protecting groups on UG and
UH were successfully removed by treating the solid-support bound
oligonucleotides with a mixture of Pd(PPh3)4 and Et2NH·HCO3 in
CH2Cl2 following a protocol by Eschenmoser et al.44 All oligonu-
cleotides were deprotected and cleaved from the solid support by
treatment with ammonium hydroxide for 24 h at rt. In the case of
UD, we observed slow deprotection of the benzoyl protecting
groups as expected.46 Incubation in ammonium hydroxide at
65 °C47 or a methylamine/ ammonia mixture48 caused substantial
depyrimidination of the UD monomer in our case (see the Support-
ing Information p. 13 for discussion). Instead, the use of anhy-
drous ammonia in methanol and incubation at 65 °C for 14 days
successfully gave the fully deprotected oligonucleotides. The final
crude oligonucleotides were purified by ion-exchange chromatog-
raphy followed by desalting to provide the final oligonucleotides in
high yield and purity.
Results
Synthesis
The double-headed nucleosides (Figure 2) were constructed ac-
cording to our previously described methodology,24,28 and incor-
porated into oligonucleotides using conventional phosphoramidite
chemistry. The syntheses of the phosphoramidite building blocks
are illustrated in Scheme 1. The mild TEMPO-based40,41 oxidation
of 3′,5′-O-(tetraisopropyldisiloxane-1,3-diyl)uridine 1 followed by
Corey–Chaykovsky epoxidation42 of the 2′-keto-uridine intermedi-
ate 2 were used to obtain the 2′-spiro-epoxy nucleoside 3. Both
reactions proceed stereospecifically and in high yields. This 2′-
spiroepoxy nucleoside is the central branch point for the genera-
tion of all the double-headed nucleotides; the epoxide is readily
opened by nucleophilic bases under alkaline conditions. All six
nucleobases are alkylated with good-to-exclusive N-1 (for pyrim-
idines) and N-9 (for purines) selectivity. Moreover, selectivity was
greatly enhanced using protecting groups that were anyway re-
quired for oligonucleotide synthesis. Protection of the exocyclic
amines were realized with standard benzoyl (C, A and D) and iso-
butyryl (G) groups. In addition, the G and H bases were found to
require installation of 6-O allyl groups43,44 to reduce the nucleo-
philicity of the N-3 atom, which appeared to otherwise react intra-
molecularly with the P(III) center once activated during the cou-
pling step of the oligonucleotide synthesis (see the Supporting In-
formation p. 12 for details and discussion). The isolated yields for
the introduction of the second nucleobases varied from 53% (4T)
to 83% (4H). After formation of the double-headed structure, the
silyl ethers were removed using TBAF to obtain the corresponding
sugar-deprotected double-headed nucleotides 5. In the case of
4D, the use of Et3N·3HF was superior to TBAF, since residues of
11-mer duplexes containing a twelfth base pair
To gauge the quality of the double-headed nucleotides’ base pair-
ing, they were first analyzed in undecamers containing an A, T
and U rich core flanked by CGC/GCG trinucleotides to minimize
fraying effects (Figure 3). When any two of the double-headed
nucleotides are placed in a so-called +1 interstrand zipper ar-
rangement, the two C2′-linked nucleobases come in spatial con-
tact, and a twelfth base pair may ensue (Figure 3A). On basis of
the six different C2′-linked nucleobases, this twelfth base pair
(X·Y) can assume a total of 62 = 36 possible combinations. The
influence of this twelfth base pair was studied by melting temper-
ature (Tm) experiments and compared with a reference duplex
without the C2′-linked bases (Figure 3B, U = 2′-deoxyuridine), as
well as duplexes where the double-headed nucleotides are
This article is protected by copyright. All rights reserved.

10.1002/chem.201901077
Chemistry - A European Journal
FULL PAPER
Scheme 1. Reagents and conditions: (a) TEMPO, IBD, AcOH, CH2Cl2, rt; (b) (CH3)3SOI, NaH, THF, DMSO, rt; (c) Bases B–H, and NaHMDS, THF, 55 °C (4C), or
NaH, DMF, 35 °C (4G,H), or NaH, DMF, 110 °C (4T,D), or KHMDS, THF, 55 °C (4A); (d) TBAF, THF, rt (5C,G,T,A,H), or Et3N·3HF, THF, rt (5D); (e) DMTCl,
pyridine, rt (6C,G,T,A,H), or DMTCl, 2,6-lutidine, DMSO, rt (6D); (f) 2-Cyanoethyl N,N-diisopropylchlorophosphoramidite, iPr2NEt, CH2Cl2, rt; (g) Incorporation into
DNA oligonucleotides. Deprotection conditions: Conc. aq. NH3, 24 h, rt (UT,UA,UC), or Et2NH·HCO3, Pd(PPh3)4, PPh3, CH2Cl2, 50 °C, then conc. aq. NH3, 24 h, rt
(UG,UH), or conc. NH3 in MeOH, 14 days, 65 °C (UD).
A
B
C
Y
5’
3’
5’
5’
3’
3’
T
C
G
A
H
D
Y
G C
C G
G C
A T
A T
U A
X Y
A U
T A
G C
C G
G C
G
C
G C
C G
G C
A T
A T
U A
X Y
A U
T A
G C
C G
G C
C
T
C
G
A
T
C
42.5 41.5 48.0 50.5 43.5 52.5
42.5 41.5 48.0 50.5 43.5 52.5
G
C
T
G
A
A
U
A
T
G
C
G
T
C
G
A
34.0 33.5 38.0 44.5
34.0 33.5 38.0 44.5
44.5 47.0 54.5 42.5 45.5 43.5
44.5 47.0 54.5 42.5 45.5 43.5
T
33.0 37.5 49.0 32.5
33.0 37.5 49.0 32.5
G
A
H
D
48.5 58.0 40.5 45.0 41.5 41.0
48.5 58.0 40.5 45.0 41.5 41.0
A
U
A
C
G
C
44.0
X
X
37.5 49.0 30.5 35.0
37.5 49.0 30.5 35.0
51.5 43.0 45.0 40.5 46.5 42.0
51.5 43.0 45.0 40.5 46.5 42.0
45.0 33.5 36.5 33.0
45.0 33.5 36.5 33.0
43.0 46.0 40.0 49.0 38.5 36.5
43.0 46.0 40.0 49.0 38.5 36.5
52.5 44.0 36.5 41.5 37.0 37.5
52.5 44.0 36.5 41.5 37.0 37.5
11-mer
reference
12-mer
reference
Figure 3. 11-mer DNA duplexes containing a twelfth base pair (X·Y) arising from the C2′-linked bases (A), and the corresponding regular 11- and 12-mer regular
DNA duplexes (B,C). Matrices comprise the measured Tm values (°C) at 1.5 µM concentrations of each DNA strand in a medium salt buffer (2.5 mM Na2HPO4, 5.0
mM NaH2PO4, 100 mM NaCl, 0.1 mM EDTA) at pH 7.0.
swapped for their corresponding natural deoxydinucleotide pairs
(Figure 3C).
in contact (Tm = 54.5–58.0 °C) or C2′-linked T and A/D in contact
(Tm = 50.5–52.5 °C). It follows that intercalation of additional base
pairs in the duplex is highly thermostabilizing (up to +14 °C), and
Watson–Crick base combinations are favored. Non-Watson–
Crick pairs generally have a neutral or destabilizing effect relative
to the 11-mer reference duplex except for the three classical wob-
ble pairs H·A, H·C and G·T (which are weakly stabilizing by +1.5
to +5.0 °C) as well as the C·C pair (+3.0 °C).
Remarkably, the modified 11-mer duplexes with matched
Watson–Crick base pairing were also notably more stable (+5.5
to +9.0 °C) than the corresponding genuine 12-mer duplexes
(compare Figure 3A with 3C). Specifically, the Tm value is in-
creased by 5.5–9.0 °C and 6.0–6.5 °C when the twelfth base pair
is a C2′-linked G·C or T·A pair, respectively. Hence, it is very
For the Tm experiments, DNA duplexes were formed by an-
nealing the constituent oligonucleotides in neutral medium salt
buffer using 1.5 μM concentrations of each strand. The duplexes’
Tm values were derived from the 260 nm UV absorption melting
curves. We observed consistency between duplicate melting
curve recordings, and annealing curves were all within 1.0 °C of
the corresponding melting curves. The measured Tm values
(rounded to nearest half) are listed in the matrices of Figure 3. As
indicated, the unmodified 11-mer reference has an experimental
Tm value of 44.0 °C (Figure 3B). The 36 different modified 11-mer
duplexes have Tm values in the range of 36.5–58.0 °C (Figure 3A).
The most stable duplexes ensued by placing C2′-linked C and G
This article is protected by copyright. All rights reserved.

10.1002/chem.201901077
Chemistry - A European Journal
FULL PAPER
beneficial for the DNA duplex to convert one of the standard base
pairs into a C2′-linked base pair. This net stabilizing effect can be
rationalized in terms of reduced torsional freedom and electro-
static repulsion of the helix backbone as the number of phos-
phates is reduced. Indeed, the Tm vales of all the modified 11-mer
duplexes―and not only the fully Watson–Crick matched du-
plexes―are elevated non-specifically by 5.5–11.5 °C when com-
pared to the 12-mer reference. The base pairing fidelity of the
twelfth pair is in general slightly compromised by the double-
headed nucleotides. The most tolerated non-Watson–Crick base
pair is G·T with a discrimination of only 2.5–3.0 °C relative to the
A·T matched pair (Figure 3A), while the discrimination is 6.5–
7.5 °C in the native duplex (Figure 3C). Relative to the G·C pair,
the G·T is discriminated by 6.5–9.5 °C, but 11.0–11.5 °C in the
native duplexes. The least tolerated are the purine-purine base
pairs G·G, H·H, D·D, D·H, G·D and G·H in the modified duplex.
The most discriminating residues are UD and UC with discrimina-
tions of no less than 8.5 °C and 7.5 °C, respectively. UD is partic-
ularly interesting, since it demonstrates much better base pairing
specificity than UA, and it binds more strongly to UT (by 1–2 °C).
Conversely, UH is the most promiscuous base, as expected, with
a slight binding preference for UA (Tm = 46.5–49.0 °C), and least
preference for other purines UG, UD and UH (Tm = 36.5–41.5 °C).
Masked as a T analog, H discriminates A over G better than T
(5.0–9.0 vs 2.5–3.0 °C), since it cannot partake in wobble pairing
with G.
48.5 °C. In addition, the stabilities of two duplexes containing the
mismatched base pairs T·G and A·G are slightly increased. These
two combinations causes inferior discrimination of G. Specifically,
the T·G and A·G mismatches are discriminated by only 1.5 °C
(down from 6.5 °C) and 1.0 °C (down from 4.5 °C) relative to
matched duplexes with T·A and A·T. Less important reductions in
the mismatch discrimination are seen in two cases, i.e. the T·T
mismatch relative to the T·A match, which is discriminated by
6.0 °C (down from 9.5 °C), as well as the mismatch A·C relative
to the matched A·T, which is discriminated by 5.0 °C (down from
8.5 °C). Altogether, UT and UA clearly behave as dinucleotide
mimics with predilection for Watson–Crick base pairing, however,
mismatches are moderately tolerated.
In the unmodified duplex with H and D (Figure 4B), the H·C
and D·T pairs are the thermally most stable configurations with Tm
values of 49 °C and 50 °C, respectively. Thus, H and D behave
mostly as G and A analogs, respectively, in standard settings. Yet,
H display selectivities of only 3.0–5.5 °C, thereby confirming its
degeneracy. In the duplex with UH (Figure 4A), a slight preference
for base pairing to A (Tm = 48 °C) over C (Tm = 47 °C) was ob-
served, which means that a C2′-linked H more take the role of a
T analog rather than a G analog. Consequently, the relative bind-
ing preference to A versus C is reversed compared to the unmod-
ified duplex, where H has preferential binding to C (Tm = 49 °C)
and binding to A is somewhat less stable (Tm = 46 °C). In addition,
H·G and H·T mismatches (relative to H·A) are better discrimi-
nated with UH (5.0 °C) than 5′-UH (2.0–2.5 °C). In the duplex with
UD, preferential base pairing to T is dominant (Tm = 49.5 °C) like
in the reference duplex, and it confirms that C2′-linked D is a po-
tent A analog. The improvement of +1.0 °C in the Tm compared to
UA matches the typical thermal gain of replacing an A∙T pair with
D∙T,35 although this gain is not observed in the unmodified duplex.
Notably, the improved base-pairing fidelity of UD compared to UA
is distinct. Decreases in the Tm of 5.5–9.5 °C are observed for du-
plexes mismatches opposite to the C2′-linked D, and mismatches
are thus less tolerated than for UA (1.0–8.0 °C).
Since UA and UD carry two complementary bases on the
same unit, we also examined their abilities to pair with themselves
in what is essentially 14 bp duplexes made from two 13-mer oli-
gonucleotides (Figure 4C). In our preliminary report,28 we found
that this UA·UA pair increases the Tm by 4.5 °C relative to the reg-
ular 14-mer duplex containing 5′-UA:5′-UA (Tm = 51.0 °C), and
6.5–7.5 °C compared to the corresponding singly-modified du-
plexes.28 It follows that UA binds to itself much more strongly than
to the natural dinucleotide. Nonetheless, our present results indi-
cate that UD is markedly inferior to UA in this regard. In fact, the
UD·UD pair was found to increase the duplex stability by only
1.0 °C (Tm = 52.0 °C, Figure 4C) compared to the reference du-
plex. This is in the light of the six potential hydrogen bonds in
UD·UD instead of just four in UA·UA. The explanation for this dis-
crepancy might in fact come from the reinforcement of base pair-
ing, which reduces the flexibility of the DNA or disrupts the hydra-
tion of the minor groove. However, it is still of note that the duplex
with self-pairing UD is more stable than the unmodified duplex with
the same number of regular base pairs (by +1 °C), and addition-
ally increases the duplex by 3–4 °C compared to the constituent
Double-headed nucleotides behaving as condensed dinucle-
otides
Next, to explore the actual dinucleotide behavior of the double-
headed nucleotides (UC, UG, UA, UT, UD and UH), they were placed
across two natural nucleotides. For this study, we turned to DNA
duplexes of 14 base pairs with a single perturbation site (Figure
4). An asymmetric, nonpalindromic sequence was chosen to di-
minish loop formation or mispairing. The measured Tm values
from this study are displayed on the horizontal axes in Figure 4.
Here, combinations of X:Y are arranged vertically; sorted in de-
scending order of Tm values. In the modified duplex (Figure 4A),
X corresponds to a C2′-linked nucleobase whereas in the unmod-
ified duplex (Figure 4B), X takes up a normal 2′-deoxyribonucleo-
tide. Thus, the specific base-pairing properties of each of the six
double-headed nucleotides can be compared directly with that of
the corresponding genuine dinucleotides
As seen in the unmodified duplex (Figure 4B), the base pair-
ing of X·Y is dictated by Watson–Crick pairing (C·G, T·A) as ex-
pected. Likewise, the C2′-linked nucleobases favor Watson–Crick
base combinations (Figure 4A). Notably, the duplex Tm values re-
lating to UC are all lifted by 1 °C in relation to the natural contexts,
thus retaining the same level of mismatch discrimination (14–
16 °C). In the case of UG, the Tm value of the fully-matched duplex
is intact (Tm = 55.0 °C) compared to the native duplexes, and the
base-pairing specificity is improved by 1.0–2.0 °C. Accordingly,
both UC and UG appear as flawless analogs of 5’-UC and 5’-UG
dinucleotides. While UT and UA also favor Watson–Crick base
pairing, these fully-matched duplexes are slightly destabilized in
comparison to 5′-UT or 5′-UA (compare Figure 4A with 4B for X =
T,A); the Tm values are lowered from 51.0–51.5 °C to 48.0–
This article is protected by copyright. All rights reserved.

10.1002/chem.201901077
Chemistry - A European Journal
FULL PAPER
X
Y
X
Y
A
B
C
C
C
C
C
G
G
G
A
T
C
C
T
C
C
C
C
G
G
G
A
T
C
C
T
55.0
55.0
54.0
55.0
47.0
46.0
44.5
51.5
45.0
42.0
40.5
51.0
46.5
42.5
42.0
46.0
49.0
41.0
41.0
39.0
40.0
40.0
38.0
5’
3’
5’
3’
5’
3’
G C
C G
T A
C G
A T
C G
U A
X Y
C G
T A
C G
C G
C G
A T
G C
C G
T A
C G
A T
C G
U A
X Y
C G
T A
C G
C G
C G
A T
G C
C G
T A
C G
A T
C G
U X
45.0
44.5
43.5
G G
G G
G A
X
G
T
T
T
T
A
A
A
G
T
C
T
T
T
T
T
A
A
G
T
C
T
48.0
46.5
42.0
39.0
48.5
47.5
43.5
40.5
48.0
47.0
43.0
43.0
49.5
A
D
55.5
52.0
X U
0
10
20
30
40
50
60
C G
T A
C G
C G
C G
A T
Tm [°C]
A G
A G
A
A
H
H C
H G
H
D
D
D G
D A
C
A
A
A
A
H
H C
H G
H
D
D
D G
D A
C
A
A
44.0
43.5
T
T
C
T
T
C
14-mer
reference
50.0
44.5
42.0
41.0
44.0
43.0
40.0
0
10
20
30
40
50
60
0
10
20
30
40
50
60
Tm [°C]
Tm [°C]
Figure 4. Double-headed nucleotides (UX) behaving as condensed dinucleotides in 14-mer DNA across different combinations (3′-AY) of natural dinucleotides (A),
and the corresponding native DNA duplexes (B). In addition, data for two duplexes containing self-pairing UA and UD nucleotides are shown to the right (C). Bars
representing data from the mismatched base pairs (X·Y) are shown in lighter gray color. Tm values were measured in medium salt buffer (2.5 mM Na2HPO4, 5.0
mM NaH2PO4, 100 mM NaCl, 0.1 mM EDTA) at pH 7.0 containing 1.5 µM concentrations of each DNA strand.
singly-modified duplexes containing UD:3′-AU (Tm = 48.0 °C) and
5′-UA:UD (Tm = 49.0 °C) (data not shown).
backbone conformational parameters (as defined by the 3DNA
software package;50 see Supplementary Figure S33) were moni-
tored during the trajectories to quantitatively probe the conforma-
tional behavior of duplexes. All simulated structures remained in
canonical B-DNA form over 500 ns. Incorporation of the double-
headed nucleotides led to no visual perturbation of the helixes
outside of the three central base pairs U7·A22, X·Y and C9·G20.
The distributions of helix parameters, sugar puckering and pseu-
dorotational angles for the six central base pairs during the MD
simulations are shown in the probability mass functions of Sup-
plementary Figures S34–S51. Supplementary Tables S2–S19
summarize the corresponding translational and rotational mode
values for an easier overview.
First, the MD simulations of duplexes 1–5, where the C2′-
linked nucleobases are situated across matching bases, were an-
alyzed with a primary focus on establishing the binding geometry
and kinematics. The average geometry of the three central base
pair residues are shown in Figure 6 (duplexes 1–4) and Supple-
mentary Figure S28 (duplex 5). In all cases, standard base pairing
schemes were maintained throughout the simulations, i.e. stand-
ard Watson–Crick contacts in the case of UA, UT and UD (duplexes
1–3), and standard H·A and H·C wobble contacts in the case of
UH (duplexes 4 and 5). In fact, the calculated hydrogen bond oc-
cupancies are indifferentiable from the native duplexes (see Sup-
plementary Figure S26). Interestingly, the modified and native
base pairs display geometric similarity (isostericity), meaning that
the base positions and the interstrand distances are very similar.
In all cases, the X8·Y21 base pair of the modified duplexes is neatly
Base-pairing properties of double-headed nucleotides at the
atomic level
To gain deeper structural insights into the dinucleotide behavior
of the double-headed nucleotides, we selected a subset of the
constructed 14-bp DNA duplexes for atomistic molecular dynam-
ics (MD) simulations. Understanding the degree of helix perturba-
tion in combination with the mode of base pairing enabled us to
better rationalize the observed trends in the Tm values. Accord-
ingly, we focused our in-silico exploration on the nine duplexes
that contain UA, UT, UD and UH in their capacities as compressed
dinucleotides across matching bases (i.e. UA:3′-AT, UT:3′-AA,
UD:3′-AT, UH:3′-AA and UH:3′-AC) or across a G mismatch (i.e.
UA:3′-AG, UT:3′-AG, UD:3′-AG and UH:3′-AG), respectively. The
selected duplexes and numbering of nucleobases are indicated in
Figure 5. The double-headed nucleotides are located at base
numbers 7 (U7) and 8 (X), and are situated across A22 and Y21,
respectively. To compare the modified duplexes with idealized B-
DNA helix parameters, the corresponding unmodified controls
were also subjected to MD simulations. The duplexes were first
constructed in canonical B-DNA conformations with explicit rep-
resentation of water and ions. The MD simulations were per-
formed using the AMBER14 package.49 After full equilibration of
the systems, the duplexes were subjected to 500 ns of Langevin
dynamics. To validate the results, two parallel trajectories with dif-
ferent initial velocities were performed. Base pair, base step and
This article is protected by copyright. All rights reserved.

10.1002/chem.201901077
Chemistry - A European Journal
FULL PAPER
5’
3’
5’
3’
5’
1
2
3
4
5
G C 28
C G 27
1
2
G C 28
C G 27
T
A
26
C G 25
24
3
4
5
T
A
26
C G 25
T
A 24
Duplex
X
Y
№
1
2
3
4
5
6
7
8
9
A
T
D
H
H
A
T
T
A
T
A
C
G
G
G
G
A
T
6
7
8
9
C G 23
U A 22
X Y 21
C G 20
6
7
8
9
C G 23
U A 22
X Y 21
C G 20
5’
100
115
10
T
A
19
10
T
A
19
D
H
11 C G 18
12 C G 17
13 C G 16
11 C G 18
12 C G 17
13 C G 16
3’
3’
14
A
T
15
14
A
T
15
Duplex 1‒9
Duplex 1*‒9*
3’
Figure 5. Sequences of the nine DNA duplexes subjected to molecular dy-
namics studies, and the nucleobase numbering scheme used in analysis.
Figure 7. Superposition of the backbone of the 5′-UTC (blue) and 5′-UT (red) di-
nucleotide step from the MD simulations of duplex 2 and 2*, respectively. Bases
are omitted for clarity. The internal hydrogen bond interaction of the arabinose
sugar is shown as a dashed line. The illustration indicates the projected angle
between the two successive C3′-P and P-C4′ vectors. The structure is representa-
tive for all simulated double-headed nucleotides.
placed in the stack of nucleobases with ordinary B-DNA rise val-
ues of 3.0–3.5 Å. It follows that incorporation of the compressed
dinucleotide mimic does not reduce the helical pitch; rather the
geometry change is compensated by certain adjustments of the
backbone torsion angles in the modified strand (Figure 7).
deoxyuridine residue in the native setting (χ ~ –135°). To achieve
the optimal geometry for base pairing, the nucleotide that is pre-
sented to the C2′-linked base (i.e. Y21) also adopts glycosidic an-
gles of –75 to –105° (versus –110 to –135° in native settings).
While the helix width is unaffected by the presence of the
double-headed nucleotides, they do have a few substantial ef-
fects on the stacking geometry of the central X8·Y21 pair and the
immediately adjacent base steps. A striking feature is the reduced
twist of the UX step (base step 7-8), which with an average value
of +10° is invariably lower than the standard B-DNA twist of ap-
prox +34° that is observed in the native duplexes (Figure 8A). In
addition to the reduced helical twist, specific changes in the
buckle (Figure 8B) and propeller (Figure 8C) parameters of the
UX:3′-AY segment were observed. While the native U7·A22 and
Specifically, the segment of the 3′-side of the UX nucleotides
is elongated to accommodate the extra nucleobase. This happens
by changes to the ε, ζ and β+1 torsion angles. The double-headed
nucleotides adopt very fixed geometries throughout the simula-
tions. Notably, they feature a robust hydrogen bond from the C2′-
hydroxyl group to the intra-residual O5′ atom (indicated with a
dashed line in Figure 7). This contact may help to stabilize the
sugar conformation in a rigid C2′-endo puckering mode (P ~ 150°),
causing the C2′-linked base to be laterally oriented into the base
stack. While this hydrogen bond is not seen in stand-ard arabino
nucleic acids,51 it has been reported in force field simulations of
other C2′-substituted arabino nucleic acids.52 Also common to all
the double-headed nucleotides is a relatively high anti orientation
(χ ~ –105°) of the glycosidic bond compared to the 2′-
Duplex 1
Duplex 2
Duplex 3
Duplex 4
Duplex 1*
Duplex 2*
Duplex 3*
Duplex 4*
Figure 6. Perpendicular-to-the-helical view of the central 5′-d(U7X8C9)/(A22Y21G20)-5′ motifs of duplexes 1–4 obtained from the MD simulations (top row). The corre-
sponding native duplexes are marked with an asterisk (*). The models are generated by averaging MD frames over the course of the full trajectory (500 ns or 250k
frames) in which all nucleobases are robustly positioned in the helical stack. Potential H-bonds are shown as dotted blue lines. The double-headed nucleotides are
colored―UA (red), UT (green), UD (blue), and UH (yellow)―and the corresponding native dinucleotides are colored accordingly.
This article is protected by copyright. All rights reserved.

10.1002/chem.201901077
Chemistry - A European Journal
FULL PAPER
Duplexes
Duplexes
Duplexes
A
B
C
1
2
3
4
5
1*
2*
3*
4*
5*
1
2
3
4
5
1*
2*
3*
4*
5*
1
2
3
4
5
1*
2*
3*
4*
5*
30
50
30
40
30
20
10
15
0
15
0
-15
-30
-15
-30
50
0
C₄·G₂₅ A₅·T₂₄ C₆·G₂₃ U₇·A₂₂ X₈·Y₂₁ C₉·G₂₀ T₁₀·A₁₉
15
C₄A₅ A₅C₆ C₆U₇ U₇X₈ X₈C₉ C₉T₁₀ T₁₀C₁₁
C₄·G₂₅ A₅·T₂₄ C₆·G₂₃ U₇·A₂₂ X₈·Y₂₁ C₉·G₂₀ T₁₀·A₁₉
15
40
30
20
10
0
Base-pair
Base-step
Base-pair
0
-15
-30
0
-15
-30
C₄·G₂₅ A₅·T₂₄ C₆·G₂₃ U₇·A₂₂ X₈·Y₂₁ C₉·G₂₀ T₁₀·A₁₉
C₄·G₂₅ A₅·T₂₄ C₆·G₂₃ U₇·A₂₂ X₈·Y₂₁ C₉·G₂₀ T₁₀·A₁₉
C₄A₅ A₅C₆ C₆U₇ U₇X₈ X₈C₉ C₉T₁₀ T₁₀C₁₁
Base-pair
Base-pair
Base-step
Figure 8. Twist, buckle and propeller angles (°) at central base steps/pairs of the 14-mer duplexes 1–5 (top charts) and 1*–5* (bottom charts). Duplex 1 (UA:3′-AT),
duplex 2 (UT:3′-AA), duplex 3 (UA:3′-AT), duplex 4 (UH:3′-AA) and duplex 5 (UH:3′-AC). The values are computationally determined mean values based on every fifth
frame within the 500 ns simulations. Error bars indicate standard deviations (±σ). The pictorial diagrams are reproduced from the 3DNA protocol.53 The major groove
is at positive x.
X8·Y21 base pairs are co-planar along the helix axis (i.e. has a
buckle of 0°) and have an average propeller of approx –10°, the
double-headed nucleotides cause buckling of the X8·Y21 pair and
change the propeller of the U7·A22 pair. These structural perturba-
tions are larger in the cases where Y21 is a pyrimidine (duplexes
1, 3 and 5) than when it is a purine (duplex 2 and 4). In the former,
Y21 adopts pseudorotational angles of 45–75° (C4′-exo) instead
of the native C2′-endo (135–150°) pucker (see the Supporting In-
formation), and the X8·Y21 pair buckles by approx +15° (Figure
8B). In addition, the U7·A22 pair takes positive propeller values
(+10°, Figure 8C). In contrast, when Y21 is a purine, it adopts a
Southern pseudorotational angle of ~180° (23T pucker), and buck-
ling and propeller-twisting of the base pairs are negligible. For the
other base pair and base-step parameters (opening, roll, slide,
shift, stagger, etc.) there is largely consensus between the un-
modified and modified duplexes, and any disparities are only vis-
ible in fine details.
assess possible explanations, we closely inspected the 500 ns
MD simulations of duplexes 6–9 from Figure 5.
In the simulations, structural perturbations were confined to
the region of the UX:3′-AG mismatch leaving the global helix struc-
ture unaffected. In contrast to the matched duplexes 1–5, the mis-
matched duplexes 6–9 displayed more conformational variability.
Specifically, the mismatched duplexes adopted five distinct con-
formations a–e that are schematized in Table 1 along with the
percentage occupancy of each conformation. The edge-to-edge
Table 1. Different conformations a–e of the mismatched duplexes.
a
b
c
d
e
Occupancy of conformations
Duplex Run
a
b
c
d
e
6
7
8
9
i
ii
i
ii
i
ii
i
ii
56%
-
31%
13%
-
Structural studies of double-headed nucleotides’ mispairing
to G
84%
-
16%
-
-
100%†
51%
-
-
-
-
-
-
-
-
-
49%
Having established the double-headed nucleotides’ behavior as
functional dinucleotides across matching bases, we turned our at-
tention toward aberrant base pairing with G. As demonstrated by
Tm readings, the C2′-linked A and T bases of UA and UT distin-
guish a mismatched G nucleotide somewhat poorly. Enhanced fi-
delity against G is attained by using UD and UH, respectively, as
proxies. Several factors could contribute to the higher stability of
the UA:3′-AG and UT:3′-AG structures, including reduced fre-
quency of base-flipping events as A·G and T·G mispairs are bet-
ter accommodated by the double-headed design, or the C2′-
linked A and T bases engage in favorable interactions with G that
are conformationally improbable for classical base pairs. To
95%
-
-
-
-
-
3%
15%
-
-
-
2%
14%
-
-
-
71%
63%
37%
33%
-
67%‡
100%
6*‒9* i/ii
Schematic representations of the different geometries a–e of the central five
base pairs during the two MD simulations of the modified duplexes 6–9. The
table shows the occupancy of conformations a–e during each of the two trajec-
tories (runs): (a) all-in-stack conformation; (b) flipping of C9; (c) flipping of G21
;
(d) flipping of both C9 & G21 (cross-strand stacking); (e) interstrand H-bonding
between neighboring base planes. †Wobble T·G contributes 75% to the con-
formation, the remaining 25% being T·G pairing where T8(O4) connects to both
G21(N1-H) and G21(C2-NH2). ‡Hoogsteen configuration makes up 15% of con-
formation a.
This article is protected by copyright. All rights reserved.

10.1002/chem.201901077
Chemistry - A European Journal
FULL PAPER
conformation (a) with all bases in stack dominated the trajectories
(i.e. occupancies of 51–100%). Base-flipping of C9 and/or G21 (b–
d), or “twisted” interstrand neighboring base pairing (e) were ob-
served to varying degrees (2–49%). These alternative confor-
mations were not observed for the unmodified duplexes 6*–9*.
Representative snapshots of duplexes 6–9 in the different confor-
mations (a–e) are given in Supplementary Figures S29–S32. The
MD simulations show that the C2′-linked A, T, D and H bases pair
up with G similarly to the native nucleotides, however, the confor-
mational rigidity of the double-headed nucleotides causes alter-
native base dynamics (see the Supporting Information p. 49 for
discussion). That is, the double-headed nucleotides cannot easily
absorb the structural stress of the mismatch, which is propagated
into neighboring residues instead. It follows that G-mismatches
are less likely to retain canonical Watson–Crick structure when
double-headed nucleotides are involved. However, our observa-
tions do not explicitly explain why this behavior translates into a
roughly +1 °C gain in the Tm for UA:3′-AG, UT:3′-AG and UD:3′-AG
relative to the natural dinucleotides; no unusual geometry seems
to account for this slight increase.
With its favorable DNA binding properties, double-headed
nucleotides represent an interesting candidate worthy of further
exploration. The molecular modeling offers insight into underlying
driving forces of base pairing in atomic detail, and into how dou-
ble-headed nucleotides are housed in DNA duplexes. Knowing
such information about new DNA architectures is useful for as-
sessing enzyme compatibility. In this regard, it is of note that dou-
ble-headed nucleotides do not visibly affect the helix diameter,
groove widths, or minor groove faces, which are well-known ele-
ments of polymerase activity.54 Whilst we have established the
eligibility and feasibility of double-headed nucleotides as building
blocks in duplexes, attention to their functional roles and biologi-
cal interactions is needed. On a long term, we hope that double-
headed nucleotides can unlock an untapped potential in nucleic
acid-based drugs or DNA nanotechnology, or lead to a new para-
digm in nucleic acid communication. The investigation of double-
headed nucleotides’ functional properties will be the subject of a
following communication.
Conclusions
Discussion
In this study, we have developed a full set of double-headed nu-
cleotides containing six C2′-linked nucleobases (A, T, C, G, D and
H). With the aim to shed light on the base-pairing properties of
this interesting new class of xenobiotic nucleotides, we investi-
gated their fidelity of binding in several motifs, and we utilized mo-
lecular modeling to obtain structural insight into their base-pairing
geometry. We found that the double-headed nucleotides adopt
C2′-endo sugar puckers and fit very well into the geometry of the
B-form duplex. The two bases of the double-headed nucleotides
were found to stack very rigidly with a smaller twist than the cor-
responding native base steps. Double-headed nucleotides hybrid-
ize to complementary targets neatly with their Watson–Crick
faces using a similar profile as natural DNA. And like normal ba-
ses, the C2′-linked bases associate preferentially with G bases,
giving rise to the strongest Watson–Crick base pair and the most
stable mismatches. C is the most discriminating base in double-
headed nucleotides forming the strongest Watson–Crick base
pair and the weakest mismatches. Double-headed nucleotides
containing D and H were found to strongly bind to T and A, re-
spectively, and neatly discriminate against G; they could be a use-
ful tool in designs where fidelity is important. All in all, the results
reported herein provide new insights into the potential of inserting
double-headed nucleotides in DNA, and provide valuable infor-
mation for upcoming designs of new double-headed nucleotides.
The results presented herein now allow for a generalized descrip-
tion of the association properties of double-headed nucleotides.
For the first time, we provide exclusive evidence (experimental
and computational) to support our hypothesis that both bases of
the double-headed nucleotides in fact communicate efficiently
with a target strand. Indeed, the arabino design appears seamless
for this purpose. Our results also support the view that double-
headed nucleotides, like natural nucleotides, preferentially accept
complementary bases at the Watson–Crick face of the nucleo-
bases and not at the Hoogsteen face. As demonstrated, double-
headed nucleotides can associate with either natural dinucleo-
tides or other double-headed nucleotides in the opposite strand.
In the latter, the backbone easily adapts to the situation where two
modified 11-mer oligonucleotides hybridize to contain 12 base
pairs. Hereby, the duplex carries a larger number of Watson–
Crick base pairs per phosphate unit, and information can there-
fore be delivered using a shorter sequence. Notably, such du-
plexes have higher thermodynamic stability (+5.5 to +9.0 °C) rel-
ative to the natural dodecamer. We attribute this stabilization to
reduced repulsive Coulombic interactions between the phos-
phates, and an entropic gain due to the reduced flexibility of the
sugar–phosphate backbone. To what extent differential solvation
of the backbone affects the entropy term is unknown so far. When
double-headed nucleotides are used to replace natural dinucleo-
tides, more energetically neutral effects (–2.5 to +2.0 °C) are ob-
served within the Watson–Crick pairing regime. Nevertheless,
such duplexes contain fewer phosphates without perturbing the
overall duplex geometry. This feature may be advantageous for
e.g. improving membrane permeability whilst retaining biological
function. In this regard, we have recently shown that two or three
double-headed nucleotides can be incorporated in the same DNA
strand without compromising target binding.29
Acknowledgements
The authors gratefully acknowledge support for this work by the
Danish National Research Foundation, the Danish Council for In-
dependent Research | Natural Sciences (FNU). Computation/sim-
ulation for the work described in this paper was supported by the
Danish e-Infrastructure Cooperation (DeIC) National HPC Centre,
SDU.
This article is protected by copyright. All rights reserved.

10.1002/chem.201901077
Chemistry - A European Journal
FULL PAPER
[29] K. Beck, C. Reslow-Jacobsen, M. Hornum, C. Henriksen, P. Nielsen, Bio-
org. Med. Chem. Lett. 2019, 29, 740–743.
Keywords: oligonucleotides • double-headed nucleotides • 2,6-
diaminopurine • hypoxanthine
[30] F. H. Martin, M. M. Castro, F. Aboul-ela, I. Tinoco, Nucleic Acids Res.
1985, 13, 8927–8938.
[1]
[2]
[3]
[4]
C. Chen, Z. Yang, X. Tang, Med. Res. Rev. 2018, 38, 829–869.
W. B. Wan, P. P. Seth, J. Med. Chem. 2016, 59, 9645–9667.
G. F. Deleavey, M. J. Damha, Chem. Biol. 2012, 19, 937–954.
T. R. Gingeras, R. Higuchi, L. J. Kricka, Y. M. D. Lo, C. T. Wittwer, Clin.
Chem. 2005, 51, 661–671.
[31] A. V. Aerschot, B. Peeters, H. Vanderhaeghe, Nucleosides Nucleotides
1987, 6, 437–439.
[32] F. Seela, Y. Chen, Nucleic Acids Res. 1995, 23, 2499–2505.
[33] H. Liu, R. Nichols, BioTechniques 1994, 16, 24–26.
[34] F. B. Howard, H. T. Miles, Biochemistry 1984, 23, 6723–6732.
[35] S. Gryaznov, R. G. Schultz, Tetrahedron Lett. 1994, 35, 2489–2492.
[36] M. J. Cairns, T. Thomas, C. E. Beltran, D. Tillett, BMC Genomics 2009,
10, 344.
[5]
J. D. Watson, A. A. Caudy, R. M. Myers, J. A. Witkowski, Recombinant
DNA: Genes and Genomes—A Short Course, W. H. Freeman & Co.,
New York, 2007.
[6]
[7]
[8]
[9]
J. M. Butler, Forensic DNA Typing: Biology, Technology, and Genetics
of STR Markers, Elsevier Academic Press, New York, 2005.
E. O. Wiley, B. S. Lieberman, Phylogenetics: Theory and Practice of Phy-
logenetic Systematics, John Wiley & Sons, Hoboken, 2011.
N. Goldman, P. Bertone, S. Chen, C. Dessimoz, E. M. LeProust, B. Sipos,
E. Birney, Nature 2013, 494, 77–80.
[37] U. B. Voss, A. Chollet, A. D. B. Malcolm, Biochem. Soc. Trans. 1989, 17,
913–913.
[38] F. B. Howard, C. Chen, J. S. Cohen, H. T. Miles, Biochem. Biophys. Res.
Commun. 1984, 118, 848–853.
[39] T. Brown, A. G. Craig, Nucleosides Nucleotides 1989, 8, 1051–1051.
[40] A. De Mico, R. Margarita, L. Parlanti, A. Vescovi, G. Piancatelli, J. Org.
Chem. 1997, 62, 6974–6977.
G. M. Church, Y. Gao, S. Kosuri, Science 2012, 337, 1628–1628.
[10] N. C. Seeman, Annu. Rev. Biochem. 2010, 79, 65–87.
[11] O. I. Wilner, I. Willner, Chem. Rev. 2012, 112, 2528–2556.
[12] N. C. Seeman, Trends Biochem. Sci. 2005, 30, 119–125.
[13] A. Kuzuya, Y. Ohya, Acc. Chem. Res. 2014, 47, 1742–1749.
[14] W. Xu, K. M. Chan, E. T. Kool, Nat. Chem. 2017, 9, 1043–1055.
[15] R. S. Geary, D. Norris, R. Yu, C. F. Bennett, Adv. Drug Deliv. Rev. 2015,
87, 46–51.
[41] S. Lemaire, I. Houpis, R. Wechselberger, J. Langens, W. A. A. Vermeu-
len, N. Smets, U. Nettekoven, Y. Wang, T. Xiao, H. Qu, J. Org. Chem.
2011, 76, 297–300.
[42] E. J. Corey, M. Chaykovsky, J. Am. Chem. Soc. 1965, 87, 1353–1364.
[43] Y. Hayakawa, M. Hirose, R. Noyori, J. Org. Chem. 1993, 58, 5551–5555.
[44] S. Pitsch, S. Wendeborn, R. Krishnamurthy, A. Holzner, M. Minton, M.
Bolli, C. Miculca, N. Windhab, R. Micura, M. Stanek, B. Jaun, A. Eschen-
moser, Helv. Chim. Acta 2003, 86, 4270–4363.
[16] V. B. Pinheiro, A. I. Taylor, C. Cozens, M. Abramov, M. Renders, S.
Zhang, J. C. Chaput, J. Wengel, S.-Y. Peak-Chew, S. H. McLaughlin, P.
Herdewijn, P. Holliger, Science 2012, 336, 341–344.
[45] P. Børsting, K. E. Nielsen, P. Nielsen, Org. Biomol. Chem. 2005, 3,
2183–2190.
[17] A. W. Feldman, F. E. Romesberg, Acc. Chem. Res. 2018, 51, 394–403.
[18] P. K. Sharma, P. Kumar, P. Nielsen, Aust. J. Chem. 2016, 69, 1094–
1101.
[46] A. Canol, M. F. Goodman, R. Eritja, Nucleosides Nucleotides 1994, 13,
501–509.
[47] C. Rosenbohm, D. S. Pedersen, M. Frieden, F. R. Jensen, S. Arent, S.
Larsen, T. Koch, Bioorg. Med. Chem. 2004, 12, 2385–2396.
[48] V. Boudou, K. Rothenbacher, A. V. Aerschot, P. Herdewijn, Nucleosides
Nucleotides 1999, 18, 1429–1431.
[19] P. Kielkowski, H. Cahová, R. Pohl, M. Hocek, Bioorg. Med. Chem. 2016,
24, 1268–1276.
[20] D. S. Chatzileontiadou, V. Parmenopoulou, S. Manta, A. L. Kantsadi, P.
Kylindri, M. Griniezaki, F. Kontopoulou, A. Telopoulou, H. Prokova, D.
Panagopoulos, E. Boix, N. A. Balatsos, D. Komiotis, D. D. Leonidas,
Bioorg. Chem. 2015, 63, 152–165.
[49] D. A. Case, V. Babin, J. T. Berryman, R. M. Betz, Q. Cai, D. S. Cerutti,
T. E. Cheatham III, T. A. Darden, R. E. Duke, H. Gohlke, A. W. Goetz, S.
Gusarov, N. Homeyer, P. Janowski, J. Kaus, I. Kolossváry, A. Kovalenko,
T.S. Lee, S. LeGrand, T. Luchko, R. Luo, B. Madej, K. M. Merz, F.
Paesani, D. R. Roe, A. Roitberg, C. Sagui, R. Salomon-Ferrer, G. Seabra,
C. L. Simmerling, W. Smith, J. Swails, R. C. Walker, J. Wang, R. M. Wolf,
X. Wu, P. A. Kollman. Amber14; University of California, San Francisco,
2014.
[21] S. L. Pedersen, P. Nielsen, Org. Biomol. Chem. 2005, 3, 3570-3575.
[22] T. Umemoto, J. Wengel, A. S. Madsen, Org. Biomol. Chem. 2009, 7,
1793–1797.
[23] T. Wu, K. Nauwelaerts, A. van Aerschot, M. Froeyen, E. Lescrinier, P.
Herdewijn, J. Org. Chem. 2006, 71, 5423–5431.
[24] M. Hornum, P. K. Sharma, C. Reslow-Jacobsen, P. Kumar, M. Petersen,
P. Nielsen, Chem. Commun. 2017, 53, 9717–9720.
[50] X.-J. Lu, W. K. Olson, Nucleic Acids Res. 2003, 31, 5108–5121.
[51] N. Martín-Pintado, M. Yahyaee-Anzahaee, R. Campos-Olivas, A. M. No-
ronha, C. J. Wilds, M. J. Damha, C. González, Nucleic Acids Res., 2012,
40, 9329–9339.
[25] P. Kumar, C. S. Madsen, P. Nielsen, Bioorg. Med. Chem. Lett. 2013, 23,
6847–6850.
[26] C. S. Madsen, S. Witzke, P. Kumar, K. Negi, P. K. Sharma, M. Petersen,
P. Nielsen, Chem. Eur. J. 2012, 18, 7434–7442.
[52] D. Venkateswarlu, D. M. Ferguson, J. Am. Chem. Soc. 1999, 121, 5609–
5610.
[27] P. Kumar, P. K. Sharma, C. S. Madsen, M. Petersen, P. Nielsen, Chem-
BioChem 2013, 14, 1072–1074.
[53] X.-J. Lu, W. K. Olson, Nat. Protoc. 2008, 3, 1213–1227.
[54] J. C. Morales, E. T. Kool, J. Am. Chem. Soc. 1999, 121, 2323–2324.
[28] P. Kumar, P. K. Sharma, P. Nielsen, J. Org. Chem. 2014, 79, 11534–
11540.
This article is protected by copyright. All rights reserved.

10.1002/chem.201901077
Chemistry - A European Journal
FULL PAPER
Entry for the Table of Contents
FULL PAPER
Mick Hornum, Julie Stendevad, Pawan
K. Sharma, Pawan Kumar, Rasmus B.
Nielsen, Michael Petersen, and Poul
Nielsen*
Page No. – Page No.
Base-pairing properties of double-
headed nucleotides
Doubling the information of a nucleotide: Double-headed nucleotides constructed
on arabinose sugars exhibit excellent dinucleotide nature for all combinations of nu-
cleobases; both with respect to maintaining helix geometry and base pairing mode.
Accordingly, this study lays the groundwork for condensing genetic information in
DNA onto shorter strands.
This article is protected by copyright. All rights reserved.