Intermolecular 'cross-torque': The N4-cytosine propargyl residue is rotated to the 'CH'-edge as a result of Watson-Crick interaction

Article abstract of DOI:10.1093/nar/gkv285

Propargyl groups are attractive functional groups for labeling purposes, as they allow CuAAC-mediated bioconjugation. Their size minimally exceeds that of a methyl group, the latter being frequent in natural nucleotide modifications. To understand under which circumstances propargyl-containing oligodeoxynucleotides preserve base pairing, we focused on the exocyclic amine of cytidine. Residues attached to the exocyclic N4 may orient away from or toward the Watson-Crick face, ensuing dramatic alteration of base pairing properties. ROESY-NMR experiments suggest a uniform orientation toward the Watson-Crick face of N⁴-propargyl residues in derivatives of both deoxycytidine and 5-methyl-deoxycytidine. In oligodeoxynucleotides, however, UV-melting indicated that N⁴-propargyl-deoxycytidine undergoes standard base pairing. This implies a rotation of the propargyl moiety toward the 'CH'-edge as a result of base pairing on the Watson-Crick face. In oligonucleotides containing the corresponding 5-methyl-deoxycytidine derivative, dramatically reduced melting temperatures indicate impaired Watson-Crick base pairing. This was attributed to a steric clash of the propargyl moiety with the 5-methyl group, which prevents back rotation to the 'CH'-edge, consequently preventing Watson-Crick geometry. Our results emphasize the tendency of an opposing nucleic acid strand to mechanically rotate single N⁴-substituents to make way for Watson-Crick base pairing, providing no steric hindrance is present on the 'CH'-edge.

Full text of DOI:10.1093/nar/gkv285

Published online 30 April 2015
Nucleic Acids Research, 2015, Vol. 43, No. 11 5275–5283
doi: 10.1093/nar/gkv285
Intermolecular ‘cross-torque’: the N4-cytosine
propargyl residue is rotated to the ‘CH’-edge as a
result of Watson–Crick interaction
Olwen Domingo1, Isabell Hellmuth1, Andres Ja¨schke2, Christoph Kreutz3 and Mark Helm1,*
1Department of Pharmaceutical Chemistry, Institute of Pharmacy and Biochemistry, Johannes Gutenberg University,
55128 Mainz, Rhineland-Palatinate, Germany, 2Institute of Pharmacy and Molecular Biotechnology, Heidelberg
University, 69120 Heidelberg, Baden-Wuerttemberg, Germany and 3Institute of Organic Chemistry, University of
Innsbruck, 6020 Innsbruck, Tyrol, Austria
Received November 06, 2014; Revised March 23, 2015; Accepted March 24, 2015
ABSTRACT
INTRODUCTION
The chemical modification of nucleic acid monomers has
become common practice in a variety of research fields
for gaining access to artificially enhanced DNA and RNA
properties. In particular, [3+2] cycloaddition reactions be-
tween azides and alkynes (CuAAC and SPAAC), together
with other forms of click reactions, are amongst the fa-
vorites of the bioconjugation reaction types and require
the incorporation of functionalities onto the nucleic acid
monomers (1–6).
Propargyl groups are attractive functional groups for
labeling purposes, as they allow CuAAC-mediated
bioconjugation. Their size minimally exceeds that
of a methyl group, the latter being frequent in
natural nucleotide modifications. To understand
under which circumstances propargyl-containing
oligodeoxynucleotides preserve base pairing, we fo-
cused on the exocyclic amine of cytidine. Residues
attached to the exocyclic N4 may orient away
from or toward the Watson–Crick face, ensu-
ing dramatic alteration of base pairing properties.
ROESY-NMR experiments suggest a uniform orien-
tation toward the Watson–Crick face of N4-propargyl
residues in derivatives of both deoxycytidine and
5-methyl-deoxycytidine. In oligodeoxynucleotides,
however, UV-melting indicated that N4-propargyl-
deoxycytidine undergoes standard base pairing. This
implies a rotation of the propargyl moiety toward
the ‘CH’-edge as a result of base pairing on the
Watson–Crick face. In oligonucleotides containing
the corresponding 5-methyl-deoxycytidine deriva-
tive, dramatically reduced melting temperatures indi-
cate impaired Watson–Crick base pairing. This was
attributed to a steric clash of the propargyl moi-
ety with the 5-methyl group, which prevents back
rotation to the ‘CH’-edge, consequently preventing
Watson–Crick geometry. Our results emphasize the
tendency of an opposing nucleic acid strand to me-
chanically rotate single N4-substituents to make way
for Watson–Crick base pairing, providing no steric
hindrance is present on the ‘CH’-edge.
The exocyclic amine groups in nucleobases are interesting
targets for the attachment of small groups, such as termi-
nal alkynes. Terminal alkynes, because of their small size,
can be expected to serve as a tag, which marks a specific
point for later bioconjugation via click labeling, while ex-
acting minimal perturbance until that time. Since function-
alization is likely to affect the hydrogen bonding edges of
the nucleoside, it is important to consider a potential in-
fluence on base pairing, and, as a consequence, also on
helical stability. Taking a clue from naturally occurring
modifications of exocyclic amines, we observed that N4-
methylcytidine (m4C) and N6-methyladenosine (m6A) oc-
cur in Watson–Crick helices of both DNA and RNA (7–
11), where presumably the methyl group would be rotated
toward the Hoogsteen edge/‘CH’-edge, leaving the second
hydrogen on the Watson–Crick free to interaction in hy-
drogen bonding. By generating a crystal structure of a hex-
americ duplex, Cervi et al. (12) have shown that this is in-
deed the case for m4C. In contrast, double methylation in
N4,N4-dimethylcytidine (m42C), N6,N6-dimethyladenosine
(m62A) and N2,N2-dimethylguanosine (m22G) is incompati-
ble with standard Watson–Crick pairing, although N2,N2-
dimethylguanosine (m22G) in certain RNAs is thought to in-
stead form a ‘wobble’-pair with uridine (13), which avoids
implication of the exocyclic N2 in hydrogen bonding. More
bulky, albeit single modifications of the exocyclic N6 of
*To whom correspondence should be addressed. Tel: +49 6131 39 25731; Fax: +49 6131 39 20373; Email: mhelm@uni-mainz.de
ꢀ
C
The Author(s) 2015. Published by Oxford University Press on behalf of Nucleic Acids Research.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which
permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.

5276 Nucleic Acids Research, 2015, Vol. 43, No. 11
pulsion of the 5-methyl group and the flexible ligand at the
4-position. This allowed detection of the 5-methyl group by
virtue of restricted rotation around the C4–N4 bond in the
pyrimidine base, leading to a preferential cis-conformation
of the substitution on N4, thus causing interference with
Watson–Crick base pairing within a sequence context.
Since the same effect would be expected of our N4-
propargyl substituent, we conducted a comparative study
on the nucleoside level by NMR, and in an oligonucleotide
context by UV-melting curves on modified cytidines with
and without 5-CH3-modification. Our results show that the
5-CH3-modification destabilizes both A- and B-type helices
when present in combination with even a single substituent
on position N4 of cytidine. In the absence of that methyl
group however, the propargyl group is well tolerated, thus
identifying a position for DNA tagging with minimal struc-
tural interference.
Figure 1. Hypothetical orientation of N4-substituents. A single modifica-
tion (R) on the exocyclic amine of cytosine may, in principle, orient toward
the Watson–Crick edge, or be twisted toward the ‘CH’-edge, thus enabling
normal Watson–Crick base pairing within a double helix. For an addi-
tional methyl group on position 5 of the pyrimidine ring, a steric clash
might impede this feature.
adenosine (t6A, i6A), occur in single stranded regions of
tRNA (14), as do single methylations of N1/N3 on the
Watson–Crick face (m1A, m1G, m3C, m3U), which are also
incompatible with standard base pairing (15).
MATERIALS AND METHODS
Chemical reagents and solvents for monomer and oligonu-
cleotide syntheses and purification were obtained from
Sigma–Aldrich (Munich, Germany). Thin layer chromatog-
raphy was done on pre-coated Polygram Sil G/UV254 silica
gel plates, obtained from Macherey-Nagel (Dueren, Ger-
many). Column chromatography was performed on either
silica gel 60 (mesh 230–400 mesh) or neutral aluminium
oxide 90 (70–230 mesh) from Merck (Darmstadt, Ger-
many). Mass spectra of the modified monomers were ei-
ther recorded on a Finningan MAT 95 or a Micromass
LCT mass spectrometer. A Bruker AC 300 MHz was used
A previous effort to introduce clickable units onto nucle-
osides was focused on N2-propargylated guanosine in RNA
(16). Earlier reports on alkylation of the exocyclic guano-
sine amine had suggested minimal impact on duplex sta-
bility for single substitutions, and had also reported that
significant differences in melting temperature of modified
duplexes were only evident upon a double methylation (17–
19). In agreement with this finding, our results showed that
mono-labeling on the sugar edge of guanosine via its exo-
cyclic amine had little impact on structure and RNAi ac-
tivity of an siRNA labeled on the 5ꢁ of the antisense strand
(16). However, data on the impact of propargyl groups on
exocyclic amines inside helices are still amiss.
Literature on substituents on the cytidinic exocyclic
amine suggested the N4 of cytidines as a viable target for
propargylation. While various reports do not show any in-
dications of destabilization of nucleic acid duplexes con-
taining N4-methylcytidine (20–23), investigations on the nu-
cleoside level by the laboratories of von Hippel (24) and
Becker (25), respectively, described a preferred orientation
of the exocyclic amine substituent in the single nucleoside
toward the Watson–Crick edge, mainly due to restricted ro-
tation around the C4–N4 bond. In their studies, von Hippel
(24) and Becker (25), however, have shown that a methyl
group on the exocyclic amine of cytidine has no partic-
ular effect on Watson–Crick base pairing in the context
of an oligodeoxynucleotide. This suggests that a small N4-
substituent will preferentially orient toward the Watson–
Crick edge, but that hybridization into duplexes entails
enough free energy to compensate for the minor penalty
required to reorient the substituent toward the Hoogsteen
edge (Figure 1).
1
for measuring H and 13C NMRs. 31P NMRs of the final
phosphoramidites were measured on an Avance III HD 400
MHz. 2D phase sensitive ROESY NMR measurements of
the deprotected derivatives were done on a Bruker Avance
300 MHz, with a mixing time of 250 ms and relaxation de-
lay of 2 s. Spectral width was 10 ppm in both dimensions.
An AVATAR 330 FT-IR (Thermo Nicolet) was used for in-
frared measurements.
An Agilent 1100 series was used for oligonucleotide pu-
rification on a DNAPac PA200 column (250 × 4 mm),
eluting with a gradient system of increasing concentrations
of NaClO4 in Tris buffer (pH 8). Details of HPLC pu-
rification can be found in the supporting information sec-
tion. Masses of the final oligonucleotides were confirmed
on a Bruker BIFLEX III by means of matrix-assisted laser
desorption/ionization time-of flight (MALDI-TOF) mass
spectrometry.
Nucleoside modification and phosphoramidite synthesis
Additional details to the synthesis procedures can be found
in the supplementary section. For the synthesis of both the
5-CH3 containing and the non-methyl containing propar-
gylated cytidine derivatives, the same reaction route was
followed starting from either thymidine or 2ꢁ-deoxyuridine,
respectively. The synthesis procedure toward the phospho-
ramidite building blocks is depicted in Scheme 1 below.
Hence, we decided to investigate the preferred orienta-
tion of an N4-propargyl group in cytidines. Our approach
to this problem was based on the notion that no natu-
rally occurring double modifications of cytidine are known
that combine additions to both the ‘CH’-edge and the exo-
cyclic N4. Even synthetic work on dual modifications of this
type is, to our knowledge, restricted to Carell’s 5-methyl-
2´-deoxycytidine-O-allylhydroxylamine adduct (26), which
displayed a highly interesting property, namely a steric re-
Tritylation reaction. Starting from a pyridine solution of
thymidine and 2ꢁ-deoxyuridine respectively, 0.5 equiv 4-
dimethylamino pyridine (DMAP) was added, to which a

Nucleic Acids Research, 2015, Vol. 43, No. 11 5277
Aminolysis with propargylamine. Twenty equivalents of
propargylamine was added to a solution of the respective
sulfonyl esters in dry dioxane. The reaction was allowed to
proceed at ambient temperature overnight. The products
were purified by means of column chromatography and re-
sulted in 10% and 33% yield for 4a and 4b.
Desilylation. 4a and 4b were dissolved in anhydrous
tetrahydrofuran and placed under argon. 1.6 equiv tetra-
butylammonium fluoride (TBAF) was added to each so-
lution. The desilylation reaction was complete after 1 h at
ambient temperature. Ethyl acetate was added to the mix-
ture and extraction from NaHCO3 was performed. Column
chromatography with aluminium oxide resulted in products
5a and 5b in yields of 42% and 91%, respectively.
Phosphoramidite
synthesis. The
phosphoramidites
were formed by treating compounds 5a and 5b with 5
equiv N,N-diisopropylethylamine and cyanoethyl-N,N-
diisopropylaminechlorophosphoramidite (CEP-Cl) in an-
hydrous dichloromethane and under an argon atmosphere.
The reaction mixtures were stirred at ambient temperature
for 2 h, after which the solutions were extracted with 5%
aqueous NaHCO3. After column chromatography, the
products 7a and 7b were isolated in 52% and 60% yield,
respectively.
Detritylation for ROESY NMR experiments. Compounds
5a and 5b were detritylated by treatment with 1.5 equiv
trifluoroacetic acid (TFA). Both detritylated compounds
(7a/7b) were collected in 70% yield after purification by col-
umn chromatography.
Scheme 1. Synthesis of the respective phosphoramidite building
blocks, 6a and 6b, together with the deprotected modified nucle-
osides for NMR analysis, 7a and 7b. Reagents and solvents: (a)
DMT-Cl (4,4ꢁ-dimethoxytrityl chloride)/pyridine, (b) TBDMS-Cl
(tert-butyldimethylsilyl chloride)/THF, (c) TPS-Cl (triisopropylben-
zenesulfonyl chloride)/DCM, (d) propargylamine/dioxane, (e) TBAF
(tetrabutylammonium fluoride)/THF, (f) CEP-Cl (2-cyanoethyl-N,
N-diisopropylaminechlorophosphoramidite)/DCM and (g) TFA
(trifluoroacetic acid)/DCM.
ROESY-NMR spectroscopy. The NMR samples were pre-
pared by dissolving the nucleosides in 9/1 H2O/D2O. NMR
data were acquired on a Bruker Avance II+ instrument op-
1
erating at 14.1 T. H NMR spectra were acquired using a
solution of dimethoxytrityl chloride (DMT-Cl) in pyridine
was slowly dropped. Following a reaction time of ∼12 h,
methanol was added to quench the reaction and the crude
product was purified by means of column chromatography.
The tritylated products, 1a and 2a, were obtained in 94%
and 83% yield, respectively.
double-pulsed field gradient spin-echo (DPFGSE) pulse se-
quence (27). For 7b NOESY spectra were acquired at 273,
278, 283 and 288 K with mixing times of 10, 25, 50, 75, 125,
200, 400, 750 and 1000 ms. The size of the data matrices
for each spectrum was 2048 × 256 complex data points, the
number of scans was 8 and the interscan delay was 1.5 s,
yielding a total measuring time of ∼6 h at each tempera-
ture.
TBDMS-protection. The respective tritylated com-
pounds, 1a and 2a, were treated with 4.4 equiv imidazole
and 4 equiv tert-butyldimethylsilyl chloride (TBDMS-Cl)
in dichloromethane at ambient temperature. After extrac-
tion from NaHCO3 and brine, the products were isolated
in 87% and 97% yield.
Determination of exchange kinetics. Spectral processing
and peak integration were performed using NMRPipe and
the NMRDraw software package (28). All subsequent steps
were performed using in-house written software written in
Matlab (The MathWorks, www.mathworks.com) according
to an earlier published protocol (29). Errors in the extracted
rate constants were determined by Monte Carlo analysis,
where peak intensities were randomly modulated according
to the signal to noise levels in the 2D correlation maps.
Incorporation of the sulfonyl ester. The deoxyribose-
protected products were both treated with DMAP and
triethylamine in dichloromethane, after which 1.2 equiv
triisopropylbenzenesulfonyl chloride (TPS-Cl) was added.
The reaction was allowed to stir at ambient temperature
overnight under an argon atmosphere. The reaction mix-
tures were dried in vacuo and column chromatography re-
sulted in the pure sulfonyl esters, 3a and 3b, in 50% and 25%
yield, respectively.
Arrhenius analysis. The temperature dependence of the
rate constants for 7b obtained from the exchange data was
obtained by linear regression of ln(kAB) or ln(kBA) versus
1/T, respectively, yielding the activation barriers for the s-
cis to s-trans isomerization step.

5278 Nucleic Acids Research, 2015, Vol. 43, No. 11
Table 1. Sequence of the synthesized oligodeoxynucleotides, together with
the positions of modifications for each of the two cytosine derivatives.
(Goettingen, Germany). For the oligonucleotides that con-
tain the modifications at positions 10 and 11 from the 5´-
end (ODN10.1, ODN11.1, ODN10.2 and ODN11.2 in Ta-
ble 1) hybridizations were carried out with DNA strands
that contain an abasic site directly opposite the modifica-
tion. As control, the same duplexes were formed with the
unmodified DNA strand (ODN0).
Click functionalization
All click reactions were performed in aqueous solu-
tions containing 5% (v/v) DMSO. The solutions were
buffered to pH 8 with NaH2PO4 (100 mM) and con-
tained 10 ␮M oligonucleotides, 100 ␮M azido functional-
ized Atto 647N dye (ATTO-TEC, Siegen, Germany), 0.5
mM CuSO4·5H2O, 2.5 mM tris-[4-(3-hydroxypropyl)-(1,
2, 3)triazolyl-1-methyl]amine (TPTA) and 5 mM sodium
ascorbate. The reaction mixtures were agitated under light
protection at 25 ◦C for 2 h.
Oligodeoxynucleotide synthesis and purification
Starting from preloaded CPG (Proligo), seven 22mer
oligodeoxynucleotides (ODNs) for each of the two cytosine
modifications, each carrying the respective derivative at a
varying position (see Table 1), were custom-synthesized on
a 1 ␮mol scale on an Expedite 8909 DNA/RNA synthe-
sizer (ABI/PerSeptiveBiosystems). A 22mer with the same
sequence, however lacking the modifications, was synthe-
sized in a similar fashion (ODN0), to serve as reference.
After cleavage from the solid support, the DNA strands
were purified by means of anion exchange chromatogra-
phy, eluting with NaClO4 in Tris buffer (pH 8). The samples
were all gel filtered on NAPTM columns (GE Healthcare)
to remove the elution salts. The molecular masses of all
synthesized oligonucleotides were confirmed by means of
MALDI-TOF mass spectrometry. Examples of these mea-
surements are included in the supporting information sec-
tion.
Electrophoretic mobility shift assay
Click reactions of single and double strands were analyzed
by non-denaturing PAGE. Fifty picomoles of single- and
25 pmol of double-stranded oligonucleotides (based on the
measurement of the sense strand) were loaded onto a 15%
non-denaturing polyacrylamide gel containing 1× TBE
(compounds for nondenaturing PAGE from Carl Roth).
PAGE was performed in 1× TBE buffer at temperatures
◦
< 30 C, in order to prevent strand separation. Gels were
post-stained for 20 min with Stains-All (Sigma–Aldrich,
Munich, Germany) and destained overnight in 75% iso-
propanol. Detection was carried out on a Typhoon 9400
(GE Healthcare, Munich, Germany), before and after stain-
ing, using 633 nm for excitation. Emission signals were
recorded at 670 nm.
Hybridization and temperature dependent UV-absorption
melting experiments
RESULTS
All complementary DNA and RNA strands were obtained
from IBA (Goettingen, Germany). Using the in-house syn-
thesized, unmodified DNA (ODN0) as reference, the mod-
ified ODNs were each hybridized to the complementary
DNA strand. The hybridization experiments were carried
out in 1× phosphate buffered saline (pH 7.4), with the two
complementary strands in a 1:1 ratio, to result in a final du-
plex concentration of 5 ␮M. The strands were first incu-
bated at 90 ◦C for 3 min and duplex formation was allowed
at 37 ◦C over 1 h.
To investigate the orientation of a propargyl group on the
exocyclic N4 of cytidines on both the nucleoside level and
in an ODN context, we devised a synthetic route to a com-
mon precursor, from which both the free nucleoside and the
phosphoramidite required for ODN synthesis could each be
obtained in a single reaction step.
Synthesis of modified nucleosides and corresponding phos-
phoramidites
The duplexes were diluted to a final concentration of
0.5 ␮M in degassed, RNase-free buffer, containing 10 mM
NaH2PO4 (pH 8) and NaCl (45 mM), to a final volume of
800 ␮L. Melting curves were recorded at 260 nm, a path-
Starting from deoxyuridine and thymidine, respectively, the
key step conversion into cytidines was pursued by a route
employing early incorporation of the 5ꢁ-O-dimethoxytrityl
(DMT) group (30), as displayed in Scheme 1. This was fol-
lowed by a temporary protection of the 3ꢁ-OH functional-
ity with tert-butyldimethylsilyl (TBDMS), to allow the or-
thogonal incorporation of an aryl sulfonate ester, i.e. tri-
isopropylbenzenesulfonyl (TPS), onto position four of the
nucleobase (30). An interesting observation was the for-
mation of a 2-substituted derivative of up to 15% in re-
lation to the main, 4-substituted product, as determined
◦
length of 10 mm and with a heating rate of 0.4 C/min, a
slit of 2 nm and a response of 0.2 s. All measurements were
carried out at least three times in the temperature ranges
between 20 and 85 ◦C.
To determine the effect of the modification on differ-
ent helix conformations, the same hybridization and UV-
melting experiments were repeated with duplexes formed
with complementary RNA strands, also obtained from IBA
1
by H NMR. This side product was separated from the

Nucleic Acids Research, 2015, Vol. 43, No. 11 5279
main product by means of column chromatography, before
the subsequent substitution of TPS with propargylamine
in an aminolysis reaction. This resulted in the formation
of the two respective cytidine derivatives, now containing
the key functional group on the exocyclic amine. Depro-
tection of the 3ꢁ-hydroxyl functionality led to 5a/5b, which
yielded the free nucleosides 7a/7b after cleavage of the 5ꢁ-
DMT group. Thin layer chromatography of the reaction
mixture during detritylation of the two cytosine deriva-
tives indicated the formation of both the tritylated form
of thymidine/2ꢁ-deoxyuridine (1a/1b in supplementary sec-
tion), as well as their untritylated forms, i.e. thymidine/2ꢁ-
deoxyuridine. These results suggest a likely deamination of
the propargyl residue under acidic conditions, which even-
tually resulted in a yield of only 70% for both 7a and 7b.
5a/5b were also used to obtain the respective phospho-
ramidites, 6a/6b after standard 2ꢁ-phosphitylation.
Orientation of the propargyl moiety on the nucleoside level
The orientation of the N4-propargyl moiety on the nu-
cleoside level was evaluated via ROESY NMR in wa-
ter (H2O/D2O mixture). The ROESY technique is espe-
cially suitable for the detection of spatial proximity between
atoms and thus also the degree of rotation around a certain
bond via the so-called Nuclear Overhauser Effect (NOE)
(24,31). Several potential NOEs could in principle be diag-
nostic of either an s-trans or an s-cis conformation, namely
between N4H or H8 and either H5/CH3 or H6. While,
no NOEs were found for H6 signals in any compound, a
CH3-N4H NOE signal in compound 7a was in evidence
(expanded in Figure 2a), while no CH3-H8 peak was de-
tectable (highlighted in Figure 2a). Similarly, in compound
7b, an H5-N4H peak is in evidence (expanded in Figure 2b),
but no H5–H8 signal (highlighted in Figure 2b). These pro-
vide strong evidence for s-cis being the predominant confor-
mation in both compounds. However, in the ROESY spec-
trum (Figure 2b) of 7b acquired at 278 K we found cross
peaks for H5 and H6 arising from an exchange process be-
tween the s-cis and s-trans conformation. Based on the anal-
ysis of the observed signal pattern of the higher populated
(90%) species, conformation A represents the s-cis confor-
mation, whereas conformation B (populated to 10%) very
likely refers to the s-trans conformation. We then addressed
the exchange kinetics of the process by acquiring EXSY
spectra with varying mixing times, ranging from 10 to 1000
ms at 273, 278, 283 and 288 K (Figure 2c). We could only
use the exchange data of H6 for the determination of the ex-
change kinetics, as resonance overlap impaired the analysis
of the kinetic rates using H5. The temperature dependence
of the rate constants between conformation A and B for 7b
was used to calculate the activation barriers (Ea(AB) = 27
kcal mol−1 and Ea(BA) = 19 kcal mol−1) for the s-cis to s-
trans isomerization step, implying a ꢀH of 7.81 kcal mol−1
between both conformations (Figure 2d, details of the Ar-
rhenius analysis in Supplementary Figure S14 of the sup-
porting information section). Comparison of the relative
population of both states at various temperatures placed the
difference in Gibbs free energy (ꢀG) around 1.1–1.4 kcal
Figure 2. ROESY NMR measurements of the two respective deprotected
cytosine derivatives (7a and 7b in Scheme 1), either (A) with or (B) with-
out a methyl group on position 5. In both cases, the propargyl preferen-
tially adapts an s-cis-conformation, but a minor population of s-trans-
conformation is also found in 7b by inspection of chemical exchange in
the H6 signal (C), which was identified in depth at various temperatures
to derive activation energy and thermodynamic parameters for the equi-
librium between both conformations, as shown in (D).
These findings are in agreement with above mentioned
investigations on N4-methylcytidine on the nucleotide level
(24,25), which report a preference for the cis-conformation
as well. Given the numerous reports on N4-methylcytidine
in helical structures of oligonucleotides (10–12,20–23), we
hypothesized that N4-propargylcytidine might behave simi-
larly and, in the structural context of a Watson–Crick helix,
might alter the orientation of the propargyl group, direct-
ing it away from the Watson–Crick face. The free energy re-
quired would presumably stem from hybridization. We have
therefore set out to test this hypothesis.
The propargyl group within a double helix. To verify the
above mentioned hypothesis in an oligonucleotide con-
text, both cytidine derivatives were incorporated into DNA
strands. An unmodified oligodeoxynucleotide (ODN0) as
shown in Table 1 was synthesized as reference. Modi-
fied oligonucleotides contained propargylated cytidines re-
mol−1
.

5280 Nucleic Acids Research, 2015, Vol. 43, No. 11
(32), we find that the double modification of propargyl-
N4 and methyl-C5 results in a pronounced and position-
specific negative effect on duplex stability (ODNx.1 series).
While modifications at either extremity show no or weak ef-
fects, helical destabilization is extremely pronounced in the
center of the duplex, with the strongest decrease of 11 ◦C ob-
served in a DNA–RNA hybrid of ODN12.1 with presumed
A-form helix.
While the results of the ODNx.2 series are compatible
with a minor energy penalty paid for the propargyl group
rotating away from the Watson–Crick face to allow stan-
dard base pairing, the drastic decrease in the center of the
ODNx.1 series implies major interference with standard
Watson–Crick interactions. Presumably, the rotation of the
propargyl group away from the Watson–Crick face and to-
ward the Hoogsteen edge/‘CH’-edge is prevented by the
bulk of the methyl group on C5, and the propargyl group is
still pointed to sterically interfere with Watson–Crick pair-
ing, as depicted in Figure 3.
To verify this hypothesis, we recorded melting curves of
duplexes of selected oligonucleotides of the ODNx.1 series,
with DNA strands featuring abasic sites opposite the se-
lected modified positions: ODN10.1 (or 10.2, respectively)
was hybridized to COMP10 and ODN11.1/11.2 were hy-
bridized to COMP11 (Figure 4). These correspond to du-
plexes lacking the opposing guanosine, which leaves no
Watson–Crick partner, but provides space to alleviate a
steric clash caused by the propargyl group. Melting curves
are shown in Supplementary Figure S16 (supporting infor-
mation), and derivatives of the melting curves are given
in Figure 4. Figure 4a compares melting of duplexes of
unmodified ODN0 and either of the two abasic ODNs to
the reference DNA duplex, clearly showing that a Watson–
Crick gap in the middle of the helix destabilizes by ∼10 ◦C in
melting temperature for the 10.x position, and by 6 ◦C for
the 11.x position. Thus, oligodeoxynucleotides containing
the singly modified cytidine derivative, i.e. ODNsx.2, form
duplexes with comparable stability to normal duplexes.
However, for the doubly modified cytidine oligodeoxynu-
cleotides, melting behavior mimicked that of duplexes that
have no base pair at either position 10 or 11. This under-
scores the suspicion that the methyl containing cytidine
derivatives cannot enter into Watson–Crick base pairs. Of
note, numerous stable alternative base pairs have been de-
scribed in A- and B-form helices, which may employ the
Hoogsteen edge (33) or which do not rely on hydrogen
bonds, but exclusively on the stabilizing contribution of
base stacking (34,35). Our melting data only give occasional
suggestions of the impact of base stacking in variegated se-
quence context. For example, the melting temperatures of
various constructs change depending on the position of the
modified cytidine (compare duplexes containing ODN10.2
versus ODN11.2 and ODN10.1 versus 10.2 in Figure 4b,
c). However, our understanding of these contexts is not ad-
vanced enough for a more detailed interpretation.
Figure 3. Proposed Watson–Crick base pairing for N4-substituted cytidine
derivatives both in the absence and presence of a methyl group on position
5. The 5-CH3 likely prevents rotation of the propargyl group away from the
Watson–Crick edge during base pairing, thus abolishing such H-bonding
completely.
placing individual single cytidines within the DNA se-
quence. The nomenclature applied here uses the posi-
tion of substitution counted from the 5ꢁ, and the type
of modification, where 1 corresponds to the 5-methyl-
N4-propargylaminyl-2´-deoxycytidine and 2 corresponds to
N4-propargylaminyl-2´-deoxycytidine (Table 1).
The DNA strand corresponds in sequence to the pas-
senger strand of an anti-eGFP siRNA which we had used
in previous studies (16). To investigate the stability of
helices containing propargylated cytidines, the respective
oligodeoxynucleotides were thus hybridized (i) to an olig-
oribonucleotide identical to the antisense strand of the cor-
responding siRNA, and (ii) to an oligodeoxynucleotide of
identical sequence. Determination of melting temperature
would thus inform on perturbed stability in (i) the A-type
helices of RNA-DNA hybrids and (ii) B-type DNA-DNA
helices, respectively. Of note, the resulting duplexes are of
classical siRNA design with a two-nucleotide overhang on
each 3´-end, such that the substitution of A21 against a
modified cytidine cannot be expected to disrupt a base pair.
Results of UV melting curves, as depicted in Table 2,
show relatively minor variations (maximum decrease of 3
◦C) between the unmodified duplexes and duplexes of all
oligodeoxynucleotides of the ODNx.2 series, i.e. those con-
taining propargylated cytidines, but no methyl-C5. How-
ever, where the introduction of a single methyl-C5 has been
reported to increase melting temperatures of DNA helices
In summary, the melting data lead to the conclusion that
the monosubstituted N4-propargyl cytidines integrates well
into a helix with melting temperatures only slightly be-
low unmodified duplexes. In contrast, an additional CH3-
substituent, away from the Watson–Crick face on the ‘CH’-
edge, leads to drastic and position-specific drops in helix

Nucleic Acids Research, 2015, Vol. 43, No. 11 5281
Figure 4. Comparison of melting temperatures obtained between duplexes both with and without abasic sites, presented as first derivatives of temperature
dependent UV-absorption melting curves for duplexes formed between (A) the unmodified strand and two complementary strands, each carrying an abasic
site at a different internal position (dark and light blue), as well as one complementary strand lacking an abasic site (black). Similar duplexes were formed
with the modified strands: (B) m5C-derivative and (C) cytidine derivative lacking the methyl group on position 5.
stability, which is comparable to duplexes in which base
pairing at the respective site is ablated by introduction of
an abasic site.
then identified the unlabeled species as well. The results,
which are depicted as an overlay in Figure 5, show that the
cytidines are only marginally less efficiently labeled than a
commercial oligonucleotide containing a hexynyl group as
‘clickable’ moiety. Despite presumed structural differences
of the duplexes, labeling efficiency was not significantly af-
fected by hybridization. Further experiments under milder
conditions (not shown), resulted in lower labeling yields, but
still did not reveal any significant differences between the
differentially modified cytosines (7a or 7b) that might be
interpreted in terms of reduced accessibility of the alkynyl
group in any of the nucleic acids.
In order to test the ‘click’-ability of the propargyl-
containing oligodeoxynucleotides, we applied ODN10.1
and ODN10.2, and their double strands (ds) with an anti-
sense siRNA strand (as) respectively, in a Cu(I)-catalyzed-
azide-alkyne-[3+2]-cycloaddition (CuAAC) reaction with
the fluorescent dye Atto 647N azide. After removal of resid-
ual dye, the nucleic acids were separated by non-denaturing
PAGE, which allowed to distinguish between double and
single stranded DNA. Also a significant shift in mobility of
the clicked species allowed an assessment of the click effi-
ciency. Fluorescence of the attached dyes was imaged be-
fore staining with Stains-All and renewed imaging, which

5282 Nucleic Acids Research, 2015, Vol. 43, No. 11
Table 2. Melting temperatures obtained for duplexes formed between the synthesized DNA strands and complementary DNA and RNA strands, respec-
tively, where * indicates the position of modification.
The melting temperatures are shown as: Tm Standard error of the mean (SEM), as was calculated from the mean of three measurements for the range
20–85 ◦C for each duplex.
for both B-form helices (DNA/DNA) and A-form helices
(DNA/RNA), this small penalty is reflected in melting tem-
peratures lowered by as little as one or two degrees through-
out all positions within the helix (Table 2, left two columns).
The melting data lead to the conclusion that the mono-
substituted N4-propargyl cytidines integrate well into a he-
lix, which implies rotating the substituent to the ‘CH’-edge
to allow base pairing. In contrast, the same mechanism is
not possible when the ‘CH’-edge is blocked by a 5-methyl
group, consequently leading to disruption of Watson–Crick
base pairing, with a destabilizing effect on helices that is
most pronounced at center positions of the helix. Figura-
tively spoken, the propargyl group is torqued into a position
where it interferes with standard base pairing by the methyl
group.
In conclusion, we can confirm the idea that a single alky-
Figure 5. Comparison of the click efficiency of the single stranded (ss),
lation of the exocyclic amine of nucleosides may be toler-
alkyne modified oligonucleotides MH662 (commercial, left), ODN10.1,
ated during duplex formation (17–20), and extend this find-
ing to the propargyl residue on the N4 of cytidines. For the
purpose of further nucleic acid functionalization, the termi-
nal alkyne of our propargyl functionality remains a handy
handle for click applications. Such applications may involve
CuAAC-mediated labeling after transfection of oligonu-
cleotides into cells to investigate subcellular distribution,
as has been exemplifed on the nucleoside level by Jao and
Salic, using ethynyluridine (36). Further studies may in-
volve structural and biological effects of triazole linkages
as formed upon CuAAC as, for example, published by Beal
et al. (37).
ODN10.2, as compared to their double strands (ds), formed with an anti-
sense siRNA (as). Atto647N azide was used as azido functionalized dye.
For analysis of click efficiency, native polyacrylamide gel electrophoresis
was followed by fluorescence scanning at different wavelengths to distin-
guish unclicked and clicked (*) oligonucleotides (see blue bands). Excita-
tion, before (blue) and after staining (red) with Stains-All, was done at 633
nm. Emission signals were recorded at 670 nm.
DISCUSSION
A single substituent on the N4 of cytosine has the possi-
bility to adopt either a cis- or a trans-conformation, de-
pending on surrounding influences. According to previous
studies of methyl groups as substituents, the limited rota-
tion around the C4–N4 bond in cytosine likely leads to a
cis-coordinated structure (24), which, in our case, would di-
rect the N4-functionality toward the Watson–Crick edge, at
least on the nucleoside level. Our NOESY experiments con-
firm this for a propargyl substituent on the nucleoside level.
It has been speculated (25) that in the context of an
ODN, hybridization to a complementary strand might pro-
vide the free energy needed to rotate the substituent to the
‘CH’-edge to allow Watson–Crick base pairing. We find that
SUPPLEMENTARY DATA
Supplementary Data are available at NAR Online.
ACKNOWLEDGEMENTS
We thank Dr Sven Aldenkortt for his invaluable support
in preparing the supplementary data. The authors would
like to thank Mr Heiko Rudy at the IPMB (University of

Nucleic Acids Research, 2015, Vol. 43, No. 11 5283
17. Ho¨bartner,C., Mittendorfer,H., Breuker,K. and Micura,R. (2004)
Triggering of RNA secondary structures by a functionalized
nucleobase. Angew. Chem. Int. Ed. Engl., 43, 3922–3925.
18. Pallan,P.S., Kreutz,C., Bosio,S., Micura,R. and Egli,M. (2008)
Effects of N2,N2-dimethylguanosine on RNA structure and stability:
crystal structure of an RNA duplex with tandem m2, 2G:A pairs.
RNA, 14, 2125–2135.
19. Rife,J.P., Cheng,C.S., Moore,P.B. and Strobel,S.A. (1998)
N2-methylguanosine is iso-energetic with guanosine in RNA duplexes
and GNRA tetraloops. Nucleic Acids Res., 26, 3640–3644.
20. Guennewig,B., Stoltz,M., Menzi,M., Dogar,A.M. and Hall,J. (2012)
Properties of N(4)-methylated cytidines in miRNA mimics. Nucleic
Acids Ther., 22, 109–116.
Heidelberg) for MALDI-TOF measurements of the synthe-
sized oligonucleotides.
FUNDING
Johannes Gutenberg-University Mainz. Funding for open
access charge: Johannes Gutenberg-University Mainz.
Conflict of interest statement. None declared.
21. Micura,R., Pils,W., Ho¨bartner,C., Grubmayr,K., Ebert,M.-O. and
Jaun,B. (2001) Methylation of the nucleobases in RNA
oligonucleotides mediates duplex-hairpin conversion. Nucleic Acids
Res., 29, 3997–4005.
22. Allerson,C.R., Chen,S.L. and Verdine,G.L. (1997) A chemical
method for site-specific modification of RNA: the convertible
nucleoside approach. J. Am. Chem. Soc., 7863, 7423–7433.
23. Grasby,J.A., Singh,M., Karn,J. and Gait,M.J. (1995) Synthesis and
applications of oligonucleotides containing N4-methylcytidine.
Nucleosides Nucleotides Nucleic Acids, 14, 1129.
24. Engel and von,Hippel (1974) Effect of methylation on nucleic acid
conformations: Studies at the monomer level. Biochemistry, 13, 4143.
25. Shoup,R.R., Miles,H.T. and Becker,E.D. (1972) Restricted rotation
about the exocyclic carbon-nitrogen bond in cytosine derivatives. J.
Phys. Chem., 76, 64–70.
26. Mu¨nzel,M., Lercher,L., Mu¨ller,M. and Carell,T. (2010) Chemical
discrimination between dC and 5MedC via their hydroxylamine
adducts. Nucleic Acids Res., 38, e192.
27. Hwang,T.L. and Shaka,A.J. (1995) Water suppression that works.
Excitation sculpting using arbitrary wave-forms and pulsed-field
gradients. J. Magn. Reson. Ser. A, 112, 275–279.
28. Delaglio,F., Grzesiek,S., Vuister,G.W., Zhu,G., Pfeifer,J. and Bax,A.
(1995) NMRPipe: a multidimensional spectral processing system
based on UNIX pipes. J. Biomol. NMR, 6, 277–293.
29. Spitzer,R., Kloiber,K., Tollinger,M. and Kreutz,C. (2011) Kinetics of
DNA refolding from longitudinal exchange NMR spectroscopy.
Chembiochem, 12, 2007–2010.
30. Hansen,A.S., Thalhammer,A., El-Sagheer,A.H., Brown,T. and
Schofield,C.J. (2011) Improved synthesis of
REFERENCES
1. Gramlich,P.M.E., Warncke,S., Gierlich,J. and Carell,T. (2008)
Click-click-click: single to triple modification of DNA. Angew. Chem.
Int. Ed. Engl., 47, 3442–3444.
2. Lallana,E., Riguera,R. and Fernandez-Megia,E. (2011) Reliable and
efficient procedures for the conjugation of biomolecules through
Huisgen azide-alkyne cycloadditions. Angew. Chem. Int. Ed. Engl.,
50, 8794–8804.
3. Neef,A.B. and Luedtke,N.W. (2014) An azide-modified nucleoside for
metabolic labeling of DNA. Chembiochem, 15, 789–793.
4. El-Sagheer,A.H. and Brown,T. (2010) Click chemistry with DNA.
Chem. Soc. Rev., 39, 1388–1405.
5. Schoch,J., Staudt,M., Samanta,A., Wiessler,M. and Ja¨schke,A.
(2012) Site-specific one-pot dual labeling of DNA by orthogonal
cycloaddition chemistry. Bioconjug. Chem., 23, 1382–1386.
6. Holstein,J.M., Schulz,D. and Rentmeister,A. (2014) Bioorthogonal
site-specific labeling of the 5ꢁ-cap structure in eukaryotic mRNAs.
Chem. Commun. Cambridge Engl., 50, 4478–4481.
7. Valinluck,V., Liu,P., Burdzy,A., Ryu,J. and Sowers,L.C. (2002)
Influence of local duplex stability and on uracil recognition by
mismatch-specific uracil-DNA glycosylase (Mug) N6-methyladenine.
Chem.Res.Toxicol., 15, 1595–1601.
8. Ehrlich,M., Wilson,G.G., Kuo,K.C. and Gehrke,C.W. (1987)
N4-Methylcytosine. J. Bacteriol., 169, 939–943.
9. Wei,C., Gershowitz,A. and Moss,B. (1975) Methylated nucleotides
block 5´-terminus of HeLa cell messenger RNA. Cell, 4, 379–386.
10. Bart,A., van Passel,M.W.J., van Amsterdam,K. and van der Ende,A.
(2005) Direct detection of methylation in genomic DNA. Nucleic
Acids Res., 33, e124.
11. Klimasauskas,S., Steponaviiene,D., Maneliene,Z., Petrusyte,M. and
Butkus,V. (1990) M. Smal is an N4-methylcytosine specific
DNA-methylase. Nucleic Acids Res., 18, 6607–6609.
12. Cervi,A.R., Guy,A., Leonard,G.A., Teoule,R. and Hunter,W.N.
(1993) The crystal structure of N4-methylcytosine-guanosine pairs in
the synthetic hexanucleotide d(CGCGmCG). Nucleic Acids Res, 21,
5623–5629.
5-hydroxymethyl-2´-deoxycytidine phosphoramidite using a
2´-deoxyuridine to 2´-deoxycytidine conversion without temporary
protecting groups. Bioorg. Med. Chem. Lett., 21, 1181–1184.
31. Fazakerley,G.V., Kraszewski,A., Teoule,R. and Guschbauer,W.
(1987) NMR and CD studies on an oligonucleotide containing
N4-methylcytosine. Nucleic Acids Res., 15, 2191–2201.
32. Xodo,L.E., Manzini,G., Quadrifoglio,F., van der Marel,G.A. and van
Boom,J.H. (1991) Effect of 5-methylcytosine on the stability of
triple-stranded DNA–a thermodynamic study. Nucleic Acids Res., 19,
5625–5631.
13. Urbonavicius,J., Armengaud,J. and Grosjean,H. (2006) Identity
elements required for enzymatic formation of N2,
N2-dimethylguanosine from N2-monomethylated derivative and its
possible role in avoiding alternative conformations in archaeal tRNA.
J. Mol. Biol., 357, 387–399.
33. Lescoute,A., Leontis,N.B., Massire,C. and Westhof,E. (2005)
Recurrent structural RNA motifs, Isostericity matrices and sequence
alignments. Nucleic Acids Res., 33, 2395–2409.
14. Machnicka,M.A., Milanowska,K., Oglou,O.O., Purta,E.,
Kurkowska,M., Olchowik,A., Januszewski,W., Kalinowski,S.,
Dunin-Horkawicz,S., Rother,K.M. et al. (2013) MODOMICS: A
database of RNA modification pathways – 2013 update. Nucleic
Acids Res., 41, 1–6.
34. Kool,E.T. (2001) Hydrogen bonding, base stacking, and steric effects
in dna replication. Annu. Rev. Biophys. Biomol. Struct., 30, 1–22.
35. Wojciechowski,F. and Leumann,C.J. (2011) Alternative DNA
base-pairs: from efforts to expand the genetic code to potential
material applications. Chem. Soc. Rev., 40, 5669.
15. Helm,M., Giege,R. and Florentz,C. (1999) A Watson–Crick
base-pair-disrupting methyl group (m1A9) is sufficient for cloverleaf
folding of human mitochondrial tRNA Lys. Biochemistry, 38,
13338–13346.
16. Seidu-Larry,S., Krieg,B., Hirsch,M., Helm,M. and Domingo,O.
(2012) A modified guanosine phosphoramidite for click
functionalization of RNA on the sugar edge. Chem. Commun. Camb.
Engl., 48, 11014–11016.
36. Jao,C.Y. and Salic,A. (2008) Exploring RNA transcription and
turnover in vivo by using click chemistry. Proc. Natl. Acad. Sci.
U.S.A., 105, 15779–15784.
37. Phelps,K.J., Ibarra-Soza,J.M., Tran,K., Fisher,A.J. and Beal,P.A.
(2014) Click modification of RNA at adenosine: structure and
reactivity of 7-ethynyl- and 7-triazolyl-8-aza-7-deazaadenosine in
RNA. ACS Chem. Biol., 15, 1780–1787.