Effect of N3 Modifications on the Affinity of Spin Label c for Abasic Sites in Duplex DNA

Article abstract of DOI:10.1002/cbic.201100728

Noncovalent site-directed spin labeling (NC-SDSL) of abasic sites in duplex DNAs with the spin label c, a cytosine analogue, is a promising approach for spin-labeling nucleic acids for EPR spectroscopy. In an attempt to increase the affinity of c for abasic sites, several N3 derivatives were prepared, and their binding affinities were determined by EPR spectroscopy. Most of the N3 substituents had a detrimental effect on binding. The triazole-linked polyethylene-glycol derivative (12a) showed a 15-fold decrease in affinity, whereas the binding affinities of ethyl azido (8b) and hydroxyl (8c) derivatives were five- to sixfold lower. The spin-labeled nucleoside C showed only a twofold decrease, thus binding better than 8c, even though it contains the larger 2′-deoxyribose substituent at N3 instead of a 2-hydroxyethyl group. N3 derivatives that contained the basic ethyl amino (9) or ethyl guanidino (10) substituents had both higher binding affinity and solubility, attributed to their cationic charge at neutral pH. Compounds 9 and 10 are promising candidates for NC-SDSL of nucleic acids, for distance measurements by pulsed EPR spectroscopy.

Full text of DOI:10.1002/cbic.201100728

DOI: 10.1002/cbic.201100728
Effect of N3 Modifications on the Affinity of Spin Label Å
for Abasic Sites in Duplex DNA
Sandip A. Shelke and Snorri Th. Sigurdsson*[a]
Noncovalent site-directed spin labeling (NC-SDSL) of abasic
sites in duplex DNAs with the spin label Å, a cytosine analogue,
is a promising approach for spin-labeling nucleic acids for EPR
spectroscopy. In an attempt to increase the affinity of Å for
abasic sites, several N3 derivatives were prepared, and their
binding affinities were determined by EPR spectroscopy. Most
of the N3 substituents had a detrimental effect on binding.
The triazole-linked polyethylene-glycol derivative (12a)
showed a 15-fold decrease in affinity, whereas the binding af-
finities of ethyl azido (8b) and hydroxyl (8c) derivatives were
five- to sixfold lower. The spin-labeled nucleoside C¸ showed
only a twofold decrease, thus binding better than 8c, even
though it contains the larger 2’-deoxyribose substituent at N3
instead of a 2-hydroxyethyl group. N3 derivatives that con-
tained the basic ethyl amino (9) or ethyl guanidino (10) sub-
stituents had both higher binding affinity and solubility, attrib-
uted to their cationic charge at neutral pH. Compounds 9 and
10 are promising candidates for NC-SDSL of nucleic acids, for
distance measurements by pulsed EPR spectroscopy.
Introduction
Electron paramagnetic resonance (EPR) spectroscopy is a valua-
ble technique for obtaining information about the structures
and conformational dynamics of nucleic acids under biological-
ly relevant conditions.[1] Structural information can be obtained
from direct measurement of the distances between paramag-
netic centers. Continuous wave (CW) EPR is useful for measur-
ing short to intermediate distances (5–20 ꢀ),[2] while pulsed
electron–electron double resonance (PELDOR, also known as
double electron–electron resonance (DEER)) can be used to
measure long-range distances (20–80 ꢀ).[1a,d,3] EPR also detects
spin label motion; this can be used to extract information on
the dynamics and local structure of specific sites.[1b,e,f,4]
The study of nucleic acids by EPR spectroscopy requires in-
corporation of reporter groups that contain unpaired electrons
(spin labels). Persistent aminoxyl radicals (nitroxides) are usual-
ly the spin labels of choice. To be informative, the spin labels
need to be linked to specific sites by using site-directed spin
labeling (SDSL).[1b,e,g,5] SDSL of nucleic acids is usually accom-
plished by one of two methods: incorporation of spin labels
during the oligonucleotide synthesis, or by post-synthesis spin
labeling, where the label is attached through modification of a
uniquely reactive functional group.[6] Both approaches are
based on attachment of the spin label by covalent bonds.
We have previously reported a noncovalent SDSL (NC-SDSL)
strategy for nucleic acids by utilizing ligand–receptor interac-
tions. In this approach, the rigid spin label Å, an analogue of
cytosine (C), binds to an abasic site in duplex DNA through
hydrogen bonding to a guanine (G) on the complementary
strand (Scheme 1A) and p-stacking interactions with the flank-
ing nucleotides.[7] CW-EPR spectroscopy revealed that Å was
specifically and fully bound to the abasic site at À308C.[7] This
NC-SDSL approach is promising for distance measurement in
nucleic acids by PELDOR, as this is performed with frozen solu-
Scheme 1. A) Base pairing between guanine G and the spin-labeled nucleo-
side C¸ (or nucleobase Å). B) Structures of the abasic sites used in this study.
AP=apurinic or apyrimidinic site. B=nucleobase.
tions. Sample preparation is straight-forward as DNA that con-
tains abasic sites can be readily prepared by automated solid-
phase synthesis with commercially available phosphoramidite;
spin-labeled samples for EPR studies are prepared simply by
mixing DNA and the spin label prior to EPR measurements.
Recently, we have shown that NC-SDSL with Å is highly se-
quence dependent.[8] For some flanking sequences, incomplete
binding was observed at À308C. At temperatures lower than
[a] S. A. Shelke, Prof. Dr. S. T. Sigurdsson
Science Institute, University of Iceland
Dunhaga 3, 107 Reykjavꢀk (Iceland)
E-mail: snorrisi@hi.is
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cbic.201100728.
684
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2012, 13, 684 – 690

Spin-Label Ligands
À308C, the spin label has limited solubility in aqueous solu-
tion, even with 30% ethylene glycol (generally used as a cryo-
protectant in pulsed-EPR studies) and a small amount of
DMSO. In an attempt to find spin labels that have better solu-
bility and higher affinity for abasic sites, we prepared several
N3 derivatives of the spin label Å. We show that both the ethyl
amino and ethyl guanidine derivatives of Å have higher bind-
ing affinity for abasic sites in duplex DNA and higher solubility
in aqueous solutions than Å.
Results and Discussion
The binding of Å to an abasic site in duplex DNA is strongly
governed by hydrogen bonding to an orphan base on the
complementary strand[7] and stacking interactions with the
base pairs flanking the abasic site.[8] Having identified the
structural components of the abasic site that are important for
noncovalent binding to Å, we turned our attention to increas-
ing the binding affinity through structural modifications of Å. A
logical site for modification is N3 because of its position rela-
tive to the structural boundaries of the abasic site pocket and
because N3 derivatives can be readily synthesized. We have
previously installed nonpolar alkyl substituents at N3 of Å, and
we have shown that increased alkyl chain length decreases
binding affinity and solubility.[8] In this work we examined
other structural variations, including aromatics and various
polar functional groups. Of particular interest was the incorpo-
ration of basic functional groups, such as amino and guanidine
groups, that are protonated under physiological conditions
and were expected to increase the binding affinity of the spin
label to the negatively charged DNA. Alkyne and azide deriva-
tives of Å were also prepared for conjugation to various ligands
by using the CuI-catalyzed Huisgen–Meldal–Sharpless [3+2] cy-
cloaddition reaction.[9]
Scheme 2. Regioselective N1 modification of 5-bromouracil. a) HMDS,
TMS·Cl; b) BrCH2CH2Br or BrCH2CꢀCH, I2, 40–75%; c) NaN3, 80%; d) K2CO3,
30%.
Syntheses of spin label derivatives
The syntheses of the spin label derivatives began with regiose-
lective alkylation of 5-bromouracil at the N1 position by one of
two methods, depending on the alkyl halide (Scheme 2). For
preparation of compounds 2 and 2a (Scheme 2A), the alkyla-
tion was performed by a one-pot, two-step reaction: silylation
with 1,1,1,3,3,3-hexamethyldisilazane (HMDS), followed by
treatment with the corresponding alkyl halide in the presence
of a catalytic amount of iodine to yield the N1-modified 5-bro-
mouracil (Scheme 2A).[7–8] However, this method could not be
used for the preparation of 2c, presumably because of the
lower reactivity of the hydroxyl-protected 2-bromoethanol 3.
Instead, 2c was synthesized by reacting 5-bromouracil with 3
in the presence of K2CO3 in DMSO (Scheme 2B).[8] The aliphatic
bromide of 2 was replaced with an azide by treatment with
NaN3 in DMSO to yield 2b.[10] The N1 derivatives of 5-bromour-
acil (2a–c) were treated with 2,4,6-triisopropylbenzenesulfonyl
chloride (TPS-Cl), in the presence of triethylamine and a cata-
lytic amount of 4-dimethylaminopyridine (DMAP), to yield the
O4-sulphonyl-activated compounds 4a–c (Scheme 3). Com-
pounds 4a–c were reacted with amino phenol 5[11] to give con-
Scheme 3. Syntheses of N3 derivatives of Å containing polar functional
group side-chains. a: R=CH2CꢀCH; b: R=CH2CH2N3; 2c, 4c, 6c:
R=CH2CH2O·TBDMS; 7c, 8c: R=CH2CH2OH). a) TPS-Cl, DMAP, Et3N, 30–70%;
b) Et3N, 50–70%; c) CsF, 50%; d) mCPBA, 50–65%; e) PPh3, 80%; f) DIPEA,
50%.
ChemBioChem 2012, 13, 684 – 690
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
685

S. Sigurdsson et al.
jugates 6a–c and subsequently phenoxazine derivatives 7a–c
upon cyclization with cesium fluoride. For 6c, cesium fluoride
treatment also removed the TBDMS group to yield the desired
hydroxyl compound 7c. Oxidation of the aliphatic amines of
7a–c to nitroxides with meta-chloroperbenzoic acid (mCPBA)
afforded spin labels 8a–c.
The azide group of nitroxide 8b was reduced by using Stau-
dinger conditions[12] to give spin label 9, which was subse-
quently guanidinylated to yield spin label derivative 10.[13] Spin
label 8a (containing a terminal acetylene) was prepared with
the intention of conjugating various ligands to spin label Å
through a CuI-catalyzed 1,3-dipolar cycloaddition reaction. To
this end, an azide containing a poly(ethylene glycol) chain
(11 a) and 1-azido-2-bromoethane (11 b) were conjugated to
8a to yield triazole-containing spin labels 12a and 12b, re-
spectively (Scheme 4A).
Binding affinity of Å derivatives monitored by CW-EPR
spectroscopy
To quantify the effect of N3 substituents on the binding affinity
of Å, the derivatives were individually incubated with a 14-mer
duplex DNA that contained an abasic site (F, Scheme 1B and
Figure 1), and the EPR spectra were recorded in a phosphate
buffer (pH 7) containing 30% ethylene glycol and 2% DMSO.
The spectra were recorded at several temperatures between 0
and À308C. At 08C, about 30 to 60% binding was observed,
whereas the labels were mostly bound at À308C (data not
shown). The EPR spectra recorded at À108C were chosen for
determination of the dissociation constants (see the Support-
ing Information).[7] The data show that there was considerable
variation in binding affinity within the family of N3 derivatives
(Figure 1). The propargyl derivative 8a had a similar affinity to
Scheme 4. A) Modification of spin label 8a by a CuI-catalyzed Huisgen–
Meldal–Sharpless [3+2] cycloaddition. B) Synthesis of an N3-benzyl deriva-
tive of spin label Å.
that of Å, while the affinity of propyl derivative 13 was nearly
three times weaker. The higher binding affinity of 8a com-
pared with that of 13 might be attributable to the smaller size
and linear geometry of the terminal acetylene (thus, more
easily accommodated in the abasic pocket than the sp3-hybrid-
Figure 1. Structures of N3 derivatives of the spin label Å (top). Dissociation constants (bottom) upon binding to a DNA duplex containing the abasic site F as
determined from EPR data collected at À108C. The structure of C¸ is shown in Scheme 1.
686
www.chembiochem.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2012, 13, 684 – 690

Spin-Label Ligands
ized carbon chain of the propyl group of 13). It was somewhat
surprising that derivatives 8b and 8c had about two times
lower affinity than the propyl derivative 13. They have similar
sizes, but 8b and 8c have polar groups at the ends of their
chains; these could interact with the phosphate backbone or
the solvent.
triazole ring. Neither 12b nor benzyl derivative 15 (prepared in
one step from an intermediate (14) in the synthesis of Å,[7]
Scheme 4B) was sufficiently soluble to enable determination of
their binding affinities. We also evaluated the spin-labeled
nucleoside C¸,[11] which contains a 2-deoxyribose at the N3 posi-
tion (Scheme 1). Interestingly, C¸ bound with only twofold
lower affinity than Å in spite of having a rather bulky group at
the N3 position. A molecular model of C¸ noncovalently bound
to an abasic site in duplex DNA shows that the abasic linker is
oriented outwards, which enables the sugar ring of C¸ to be
nicely accommodated at the abasic site (Figure 2B). Further-
more, the 3’- and 5’-hydroxyl groups of the sugar form hydro-
gen bonds with the non-bridging oxygen atoms of the phos-
phodiesters.
The highest binding affinity was observed for the amino (9)
and the guanidino (10) spin label derivatives: about twofold
higher than Å. The enhanced binding affinities of 9 and 10
were most likely the result of ionic interactions between the
negatively charged phosphate backbone and the positively
charged amino and guanidine groups, both of which are pro-
tonated at pH 7. In an attempt to gain an insight into how the
sugar-phosphate backbone at the abasic site could accommo-
date the spin-label side chains, a molecular model of spin label
9 in a 1:1 complex with a 14-mer DNA duplex containing an
abasic site was generated (Figure 2A). In the model, the spin
label is stacked in the abasic pocket, where it forms three hy-
drogen bonds to the guanine on the complementary strand,
while the abasic sugar is flipped outwards. The tethered am-
monium cation was oriented towards the sugar-phosphate
backbone where it formed two hydrogen bonds with non-
bridging oxygen atoms of two phosphodiesters.
Effect of abasic site linker: C3 versus F
We have previously shown that the affinities of N3-alkyl deriva-
tives was similar for an abasic site containing the tetrahydro-
furan analogue F and for one containing the propane-1,3-diol-
derived C3 linker, an acyclic analogue of 2’-deoxyribofuranose
(Scheme 1B).[8] Because of the large variation in the structures
of the side chains in 8–13, we decided to evaluate their bind-
ing to a DNA duplex containing C3. Only a minor variation in
binding affinity of the spin labels was observed for the two
abasic-site linkers and most of the derivatives bound with
slightly better affinity for F than for C3 (data not shown). The
larger size and lower flexibility of F clearly does not have a det-
rimental effect on binding when compared with C3. This is pre-
sumably due to the fact that the tetrahydrofuran ring can
rotate outwards to provide access for the N3 side chains.
A few of the derivatives contain rings at the N3 position.
The poly(ethylene glycol) chain of the triazole derivative 12a
was installed to increase the solubility of the spin label in
aqueous solutions. However, the binding affinity of 12a was
about 15-fold lower than that of Å. This dramatically lower
binding of 12a could be due to the hydrophilicity of the poly-
(ethylene glycol) chain, which increases the affinity of the label
for the polar solvent over the hydrophobic abasic site or the
Conclusion
Several N3 derivatives of spin label Å were synthesized, and
their binding affinities for an abasic site in duplex DNA were
quantified by EPR spectroscopy. All but two of the derivatives
had lower affinity for the abasic site than did Å. However, with
the exception of poly(ethylene glycol) derivative 12a, the var-
iation in binding affinities was less than fivefold. The size,
shape, and polarity of the side chains influence the binding af-
finity. For example, the N3-ethyl alcohol derivative 8c had
more than twofold lower affinity than C¸ ; this could be due to
a better fit of the 2’-deoxyribose in the abasic pocket, as indi-
cated by molecular modeling. The amino (9) and guanidino
(10) derivatives showed enhanced binding and increased solu-
bility, presumably because of their positive charge at neutral
pH. The apparent high binding affinities of derivatives 9 and
10 makes them promising candidates for distance measure-
ments by pulsed EPR. Studies into this will be reported in due
course.
Figure 2. Stereo views of molecular models of A) spin label 9 and B) spin-la-
beled nucleoside C¸ (black) bound noncovalently to duplex DNA (gray) that
contains an abasic site. Hydrogen bonds between the ammonium group of
9 or the hydroxyls on the 2’-deoxyribose of C¸ with non-bridging oxygen
atoms of the phosphate backbone are shown as dotted lines. During the
energy minimization, the sugar-phosphate backbone around the abasic site
and the spin label side chains were allowed to move, while the rest of the
structure was constrained.
Experimental Section
General: Compounds 3,[14] 5,[11] 11a,[15] 11 b,[16] 13,[8] 14[7] and C¸[11]
were prepared according to previously reported procedures. All re-
actions were carried out in oven-dried reaction flasks under an
ChemBioChem 2012, 13, 684 – 690
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
687

S. Sigurdsson et al.
argon atmosphere. Reactions were monitored by thin layer chro-
matography (TLC), performed on glass-backed TLC plates with an
extra hard layer (Kieselgel 60 F254, 250 mm; Silicycle, Quebec,
Canada), and compounds were visualized with UV light. Dichloro-
methane, pyridine, and acetonitrile were freshly distilled over calci-
um hydride prior to use. All commercial reagents were purchased
from Sigma–Aldrich and used without further purification. Flash
column chromatography was carried out with silica gel (230–400
mesh, 60 ꢀ) as the stationary phase (Silicycle). The petroleum ether
used for chromatography was from distillation between 60 and
suspension of 5-bromouracil (5 g, 26.18 mmol) in 1,2-dichloro-
ethane (1,2-DCE, 25 mL). The reaction mixture was refluxed for 5 h,
after which it became clear; it was cooled to 608C, and the solvent
was removed in vacuo to yield a colorless oil. The residue was dis-
solved in DMF (25 mL) and treated with an alkyl halide (dibromo-
ethane (9.022 mL, 104.7 mmol) for 2, propargyl bromide (4.51 mL,
52.35 mmol) for 2a)) and a catalytic amount of I2 (0.067 g,
0.26 mmol) at 258C. The reaction mixtures were refluxed for 12 h,
cooled to 258C, diluted with H2O (25 mL), and extracted with
CH2Cl2. The combined organic layers were washed with a saturated
aqueous solution of NaHCO3 and brine, dried over anhydrous
Na2SO4, and concentrated in vacuo. The crude mixtures were puri-
fied by flash column chromatography (MeOH (3%) in CH2Cl2) to
1
908C. H and 13C NMR spectra were recorded on a Bruker Avance
400 MHz spectrometer, and coupling constants are reported in
Hertz. The chemical shifts are reported in ppm relative to the resid-
1
1
ual proton signal (for H NMR) and the carbon signal (for 13C NMR)
yield white solids. 2: Yield 40%; H NMR ([D6]DMSO): d=11.85 (s,
of the deuterated solvents: [D6]DMSO (2.50 ppm), CDCl3
(7.26 ppm), [D4]MeOH (4.84 and 3.31 ppm) for H NMR; [D6]DMSO
1H; NH), 8.24 (s, 1H; CH), 4.07 (t, J=6.4 Hz, 2H; CH2), 3.71 (t, J=
6.4 Hz, 2H; CH2); 13C NMR ([D6]DMSO): d=159.57, 150.19, 145.33,
94.54, 48.89, 40.15, 30.30; HR-ESI-MS: m/z 318.8688 [M+Na]+, calcd
1
(39.52 ppm), CDCl3 (77.0 ppm), [D4]MeOH (49.05 ppm) for 13C NMR.
Molecular masses of organic compounds were determined by HR-
ESI-MS (MicroTof-Q, Bruker). The CW-EPR spectra were recorded on
a MiniScope MS200 spectrometer (modulation frequency100 kHz,
modulation amplitude1.0 G, microwave power 2.0 mW; Magnet-
tech, Berlin, Germany). Temperature was regulated by an M01 tem-
perature controller (error Æ0.58C; Magnettech).
1
for C6H6Br2N2O2 295.8696. 2a: Yield 76%. H NMR ([D6]DMSO): d=
8.26 (s, 1H; CH), 4.50 (d, J=2.5 Hz, 2H; CH2), 3.44 (t, J=2.4 Hz, 1H;
CH); 13C NMR ([D6]DMSO): d=159.47, 149.71, 144.13, 95.28, 78.18,
76.10, 37.11; HR-ESI-MS: m/z 228.9780 [M+H]+, calcd for
C7H5BrN2O2 227.9534.
Compound 2b: NaN3 (0.6 g, 10 mmol) was added to a solution of
2 (1.6 g, 5.37 mmol) in DMSO (20 mL), and the resulting mixture
was stirred for 24 h at 258C. The reaction mixture was diluted with
water (60 mL) in an ice bath and extracted with CH2Cl2 (3ꢂ20 mL).
The combined organic layers were washed with water and brine
and dried over anhydrous Na2SO4. The solvent was removed in
vacuo to afford 2b as a white solid, which was used without fur-
DNA synthesis and purification: Deoxyoligonucleotides (either
unmodified DNA or containing an F or C3 abasic site) were pre-
pared by phosphoramidite chemistry on an automated ASM800
DNA synthesizer (Biosset, Novosibirsk, Russia) by using a trityl-off
protocol and phosphoramidites with standard protecting groups
on a 1.0 mmol scale (1000 ꢀ CPG columns). All commercial phos-
phoramidites, CPG columns and chemical reagents for DNA synthe-
sis were purchased from ChemGenes Corporation (Wilmington,
MA). The C3 phosphoramidite was prepared according to a previ-
ously reported procedure.[17] After deprotection of oligonucleotides
and cleavage from the solid support by incubation in concentrated
aqueous ammonia solution at 558C for 8 h, oligomers were puri-
fied by denaturing polyacrylamide (20%) gel electrophoresis
(DPAGE). The oligonucleotides were visualized under UV light, and
the bands were excised from the gel, crushed and eluted from gel
in Tris buffer (Tris (10 mm, pH 7.5), NaCl (250 mm), Na2EDTA
(1 mm)). The DNA elution solutions were filtered through 0.45 mm
polyethersulfone membrane (Whatman) and desalted in a Sep-Pak
cartridge (Waters Corporation). Solvent was removed in a SpeedVac
(Thermo Fisher), and DNA was dissolved in deionized and sterilized
water (200 mL). UV absorbance at 260 nm was used to calculate the
concentrations of oligonucleotides according to Beer’s law. Extinc-
tion coefficients were determined by using the UV WinLab oligo-
nucleotide calculator (V2.85.04; PerkinElmer).
1
ther purification. Yield 80%; H NMR ([D6]DMSO): d=11.83 (s, 1H;
NH), 8.22 (s, 1H; CH), 3.87 (t, J=7.6 Hz, 2H; CH2), 3.60 (t, J=5.6 Hz,
2H; CH2); 13C NMR ([D6]DMSO): d=159.57, 150.28, 145.30, 94.70,
48.82, 46.93, 41.85; HR-ESI-MS: m/z 281.9597 [M+Na]+, calcd for
C6H6BrN5O2 258.9705.
Compound 2c: K2CO3 (0.5 g, 3.61 mmol) and 3 (0.44 g, 1.83 mmol)
were added to a solution of 5-bromouracil (0.5 g, 2.61 mmol) in
DMSO (10 mL). After the resulting mixture had been stirred for 2 h
at 258C, the reaction mixture was cooled to 108C, diluted with
H2O (50 mL) and extracted with CH2Cl2. The combined organic
layers were washed with saturated NaHCO3 and brine, dried over
anhydrous Na2SO4, and concentrated. The crude product was puri-
fied by flash column chromatography (MeOH (2%) in CH2Cl2) to
yield 2c as a white solid. Yield 30%; 1H NMR (CDCl3): d=8.57 (s,
1H; NH), 7.65 (s, 1H; CH), 3.88 (m, 2H; CH2), 3.81 (m, 2H; CH2),
0.88 (s, 9H; 3CH3), 0.03 (s, 6H; 2CH3); 13C NMR (CDCl3): d=159.45,
150.05, 146.04, 95.28, 60.98, 51.24, 25.95, 18.30, À4.40; HR-ESI-MS:
m/z 371.0397 [M+Na]+, calcd for C12H21BrN2O3Si 348.0505.
Molecular modeling: Molecular modeling was carried out with B-
form DNA duplex generated in Spartan 10 (Wavefunction, Irvine,
CA) with default parameters. The abasic site was generated by de-
leting the corresponding cytosine (C) base and a hydrogen atom
at the anomeric carbon of the 2’-deoxysugar. The spin label Å was
manually docked into the abasic pocket so that it formed three hy-
drogen bonds with the G on the opposing strand. The energy min-
imization to obtain equilibrium geometry at ground state was per-
formed by using molecular mechanics (MMFF). The sugar-phos-
phate backbone around the abasic site and the spin-label side
chains were allowed to move while the rest of the helix was con-
strained. The energy-minimized models were exported as PDB files
and visualized in PyMOL (DeLano Scientific LLC).
Compounds 4a–c: TPS-Cl (2.12 g, 7.0 mmol), DMAP (42.7 mg,
0.35 mmol), and Et3N (1.9 mL, 14 mmol) were added to solutions of
2a–c (3.5 mmol) in CH2Cl2 (25 mL) at 08C. The resulting reaction
mixtures were stirred for 6 h at 08C, diluted with CH2Cl2 (20 mL),
and washed sequentially with H2O (2ꢂ20 mL), saturated NaHCO3
(15 mL), and brine (15 mL). The organic layers were dried over
anhydrous Na2SO4 and concentrated. The residues were purified by
flash column chromatography (EtOAc (10%) in petroleum ether) to
yield compounds 4a–c as white solids. 4a: Yield 40%; 1H NMR
(CDCl3): d=8.16 (s, 1H; CH), 7.20 (s, 2H; 2Ar-CH), 4.59 (d, J=2.6 Hz,
2H; CH2), 4.25–4.32 (m, 2H; 2CH), 2.88–2.92 (m, 1H; CH), 2.65 (t,
J=2.6 Hz, 1H; CH), 1.29 (d, J=6.7 Hz, 12H; 4CH3), 1.25 (d, J=
6.9 Hz, 6H; 2CH3); 13C NMR (CDCl3): d=163.21, 155.01, 152.60,
151.72, 147.69, 130.27, 124.29, 87.02, 78.43, 74.95, 39.45, 34.42,
29.79, 24.64, 23.57; HR-ESI-MS: m/z 517.0767 [M+Na]+, calcd for
Compounds 2 and 2a: HMDS (16.45 mL, 78.53 mmol) and trime-
thylsilyl chloride (TMS-Cl, 1.56 mL, 13.08 mmol) were added to a
688
www.chembiochem.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2012, 13, 684 – 690

Spin-Label Ligands
1
C22H27BrN2O4S 494.0875. 4b: Yield 33%; H NMR (CDCl3): d=7.80 (s,
1H; CH), 7.21 (s, 2H; 2Ar-CH), 4.24–4.31 (m, 2H; 2CH), 3.88 (dd, J=
6.2, 4.3 Hz, 2H; CH2), 3.70 (dd, J=6.2, 4.3 Hz, 2H; CH2), 2.87–2.92
(m, 1H; CH), 1.29 (d, J=6.8 Hz, 12H; 4CH3), 1.25 (d, J=6.9 Hz, 6H;
2CH3); 13C NMR (CDCl3): d=163.33, 155.01, 152.95, 151.69, 150.42,
130.22, 124.27, 86.51, 50.72, 48.75, 34.39, 29.79, 24.62, 23.54; HR-
ESI-MS: m/z 548.0914 [M+Na]+, calcd for C21H28BrN5O4S 525.1045.
4c: Yield 68%; 1H NMR (CDCl3): d=7.82 (s, 1H; CH), 7.20 (s, 2H;
2Ar-CH), 4.29–4.32 (m, 2H; 2CH), 3.91 (m, 2H; CH2), 3.79 (m, 2H;
CH2), 2.88–2.92 (m, 1H; CH), 1.29 (d, J=6.7 Hz, 12H; 4CH2), 1.25 (d,
J=6.9 Hz, 6H; 2CH3), 0.82 (s, 9H; 3CH3), À0.09 (s, 6H; 2CH3);
13C NMR (CDCl3): d=154.83, 153.10, 151.67, 151.50, 130.43, 124.22,
100.12, 85.37, 59.99, 53.06, 34.42, 29.78, 25.85, 24.66, 23.60, 18.17,
À5.57; HR-ESI-MS: m/z 637.1737 [M+Na]+, calcd for
C27H43BrN2O5SSi 614.1845.
Spin labels 8a–c, 15: Solutions of 7a–c and 14 (0.054 mmol) in
CH2Cl2 (10 mL) were treated with a solution of mCPBA (11.25 mg,
0.065 mmol) in CH2Cl2 (2 mL) at 08C. The reaction mixtures were
stirred for 6 h at 08C, and solvent was removed in vacuo. The resi-
dues were purified by preparative TLC (MeOH (8%) in CH2Cl2) to
1
afford 8a–c as pale yellow solids. 8a: Yield 65%; H NMR (CDCl3):
d=7.05 (s, 1H; CH), 4.72 (brs, 2H; CH2), 2.55 (s, 1H; CH), 1.29 (s,
1H); HR-ESI-MS: m/z 352.1521 [M+H]+, calcd for C19H19N4O3
351.1457. 8b: Yield 60%; 1H NMR (CDCl3/CD3OD 95:5): d=8.09
(brs, 1H; CH), 6.69 (brs, 1H; CH), 3.90 (brs, 2H; CH2), 3.72 (brs, 2H;
CH2), 1.26 (brs, 4H); HR-ESI-MS: m/z 405.1568 [M+Na]+, calcd for
C18H20N7O3 382.1628. 8c: Yield 50%; 1H NMR (CDCl3/CD3OD 95:5):
d=8.23 (brs, 1H), 8.10 (s, 1H; CH), 1.26 (s, 1H); HR-ESI-MS: m/z
380.1481 [M + Na]+, calcd for C18H21N4O4 357.1563. 15: Yield 68%;
1H NMR (CDCl3): d=7.34–7.42 (m, 4H; Ar-CH), 4. 96 (brs, 1H; CH),
1.26 (brs, 2H); HR-ESI-MS: m/z 426.1641 [M+Na]+, calcd for
C23H23N4O3 403.1770. Note: Because of the paramagnetic nature of
nitroxides, the NMR spectra of spin labels show significant broad-
ening of the signals; thus some peaks (particularly of nuclei close
to the radical) are not seen in the spectra.
Compounds 6a–c: Amino phenol 5 (0.130 g, 0.63 mmol) and Et3N
(0.14 mL, 1.0 mmol) were added to solutions of 4a–c (0.57 mmol)
in CH2Cl2 (10 mL). The resulting reaction mixtures were stirred at
258C for 48 h in the dark, solvent was removed in vacuo, and the
residues were purified by flash column chromatography (MeOH
(2%) in CH2Cl2 containing NH3 (1%)) to yield 6a–c as pale yellow
solids. 6a: Yield 70%; 1H NMR ([D6]DMSO): d=8.27 (s, 1H; CH),
7.72 (s, 1H; Ar-CH), 6.66 (s, 1H; Ar-CH), 6.58 (s, 1H; Ar-CH), 4.53 (d,
J=2.4 Hz, 2H; CH2), 3.4 (t, J=2.4 Hz, 1H; CH), 1.34 (s, 12H; 4CH3);
13C NMR ([D6]DMSO): d=153.27, 145.51, 116.25, 107.80, 99.51,
87.22, 78.82, 75.75, 38.12, 32.04, 31.78; HR-ESI-MS: m/z 417.0930
Spin label 9: PPh3 (41 mg, 0.156 mmol) was added to a solution of
8b (40 mg, 0.104 mmol) in dry THF (2 mL), and the reaction mix-
ture was stirred for 30 min at 258C prior to addition of H2O (2.1 mL,
0.11 mmol). The resulting reaction mixture was stirred for 12 h at
258C and concentrated in vacuo. The residue was purified by prep-
arative TLC (MeOH (10%) in CH2Cl2) to yield 9 as a pale yellow
solid. Yield 80%; 1H NMR (CDCl3/CD3OD 95:5): d=8.01 (brs, 1H;
CH), 4.61 (brs, 1H; CH), 1.18 (brs, 4H; CH3); HR-ESI-MS: m/z
357.1811 [M+H]+, calcd for C18H22N5O3 356.1723.
[M+H]+, calcd for C19H21BrN4O2 416.0848. 6b: Yield 55%; H NMR
1
(CDCl3/CD3OD 95:5): d=8.03 (s, 1H; CH), 7.57 (s, 1H; Ar-CH), 6.57
(s, 1H; Ar-CH), 3.87–3.81 (m, 2H; CH2), 3.70–3.63 (m, 2H; CH2), 1.44
(s, 6H; 2CH3), 1.37 (s, 6H; 2CH3); 13C NMR (CDCl3/CD3OD 95:5): d=
157.34, 155.32, 147.47, 145.59, 125.76, 114.47, 108.37, 88.93, 63.67,
63.32, 30.97, 30.82, 29.68; HR-ESI-MS: m/z 448.1066 [M+H]+, calcd
Spin label 10: A solution of 9 (20 mg, 0.056 mmol) in DMF (2 mL)
was treated with 1H-pyrazole-1-carboxamidine hydrochloride
(10 mg, 0.067 mmol) and diisopropylethyl amine (15 mL,
0.084 mmol). The resulting reaction mixture was stirred for 48 h at
258C, then concentrated, and the residue was purified by prepara-
tive TLC (MeOH (20%) in CH2Cl2, ammonia (2%)) to yield 10 as a
pale yellow solid. Yield 50%; 1H NMR (CDCl3/CD3OD/[D6]DMSO
90:8:2): d=7.91 (brs, 1H; CH), 7.03 (brs, 1H; CH), 3.78 (brs, 2H;
CH2), 3.34 (brs, 2H; CH2); HR-ESI-MS: m/z 399.2010 [M+H]+, calcd
for C19H24N7O3 398.1941.
1
for C18H22BrN7O2 447.1018. 6c: Yield 52%; H NMR (CDCl3): d=7.70
(s, 1H; CH), 6.99 (s, 1H; Ar-CH), 6.82 (s, 1H; Ar-CH), 3.95 (d, J=
4.1 Hz, 2H; CH2), 3.85 (d, J=3.5 Hz, 2H; CH2) 1.43 (s, 12H; 4CH3),
0.88 (s, 9H; 3CH3), 0.01 (s, 6H; 2CH3); 13C NMR (CDCl3): d=149.66,
149.46, 141.58, 116.07, 114.09, 62.79, 62.55, 60.85, 52.34, 32.22,
32.12, 31.89, 25.94, 18.24, À5.43; HR-ESI-MS: m/z 537.1878 [M+H]+,
calcd for C24H37BrN4O3Si: 536.1818.
Spin labels 12a, b: A solution of 8a (10 mg, 0.0284 mmol) in ace-
tone was treated with either 11 a[15] or 11 b[16] (0.034 mmol) in the
presence of a catalytic amount of CuI (1 mg, 0.0028 mmol). The re-
sulting reaction mixtures were stirred under reflux for 12 h. After
cooling, the reaction mixtures were filtered, and the filtrates were
concentrated in vacuo. The residues were purified by preparative
TLC (MeOH (10%) in CH2Cl2) to afford the spin labels as pale
Compounds 7a–c: CsF (0.56 g, 3.72 mmol) was added to solutions
of 6a–c (0.37 mmol) in EtOH (10 mL), and the resulting reaction
mixtures were stirred at 858C for three days. Solvent was evaporat-
ed, and the crude products were purified by flash column chroma-
tography (MeOH (5%) in CH2Cl2 containing NH3 (1%)) to give cy-
clized compounds 7a–c as pale yellow solids. 7a: Yield 50%;
1H NMR (95:5 CDCl3:[D6]DMSO): d=7.08 (s, 1H; CH), 6.69 (s, 1H;
Ar-CH), 6.44 (s, 1H; Ar-CH), 4.50 (s, 2H; CH2), 2.48 (s, 1H; CH), 1.38
(s, 6H; 2CH3), 1.37 (s, 6H; 2CH3); 13C NMR (95:5 CDCl3:[D6]DMSO):
d=154.44, 144.79, 143.94, 142.33, 128.68, 126.03, 123.58, 110.33,
108.61, 75.73, 63.00, 38.17, 31.46, 31.39; HR-ESI-MS: m/z 337.1605
[M+H]+, calcd for C19H20N4O2 336.1586. 7b: Yield 50%; 1H NMR
(CDCl3): d=7.54 (s, 1H; CH), 6.83 (s, 1H; Ar-CH), 6.45 (s, 1H; Ar-
CH), 3.79–3.81 (m, 2H; CH2), 3.66–3.69 (m, 2H; CH2), 1.45 (s, 6H;
2CH3), 1.42 (s, 6H; 2CH3) 13C NMR (CDCl3): d=155.29, 154.53,
142.11, 127.98, 126.32, 125.73, 111.59, 108.23, 49.82, 49.73, 31.72,
29.83; HR-ESI-MS: m/z 368.1836 [M+H]+, calcd for C18H21N7O2
1
yellow solids. 12a: Yield 50%; H NMR (CDCl3): d=8.11 (s, 1H; CH),
7.93 (brs, 1H; CH), 6.96 (brs, 1H; CH), 5.06 (brs, 2H; CH2), 4.55
(brs, 2H; CH2), 3.62–3.89 (m, 16H; 8CH2), 1.27 (brs, 2H; CH3); HR-
ESI-MS: m/z 593.2568 [M+Na]+, calcd for C27H36N7O7 570.2676.
12b: Yield 45%; 1H NMR (CDCl3): d=7.91 (s, 1H; CH), 7.02 (brs,
1H; CH), 4.97 (brs, 2H; CH2), 4.71 (brs, 2H; CH2), 3.72 (brs, 2H;
CH2), 1.20 (brs, 3H; CH3); HR-ESI-MS: m/z 503.1134 [M+2H]+, calcd
for C21H23BrN7O3 500.1046.
1
367.1757. 7c: Yield 25%; H NMR ([D6]DMSO): d=7.27 (s, 1H; CH),
Acknowledgements
6.62 (s, 1H; Ar-CH), 6.52 (s, 1H; Ar-CH), 3.64 (brs, 2H; CH2), 3.56
(brs, 2H; CH2), 1.27 (s, 12H; 4CH3); 13C NMR ([D6]DMSO): d=153.48,
144.20, 143.51, 141.66, 125.94, 109.41, 108.24, 69.77, 61.92, 61.87,
58.64, 51.04, 40.14, 31.89, 31.72; HR-ESI-MS: m/z 343.1765 [M+H]+,
calcd for C18H22N4O3: 342.1652.
We thank the Icelandic Research Fund (110035021) and the Uni-
versity of Iceland Research Fund for financial support. S.A.S. ac-
knowledges a doctoral fellowship from the University of Iceland
ChemBioChem 2012, 13, 684 – 690
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
689

S. Sigurdsson et al.
[5] a) C. Altenbach, S. L. Flitsch, H. G. Khorana, W. L. Hubbell, Biochemistry
1989, 28, 7806–7812; b) W. L. Hubbell, A. Gross, R. Langen, M. A. Liet-
zow, Curr. Opin. Struct. Biol. 1998, 8, 649–656.
Research Fund. We thank Dr. Sigridur Jonsdottir for help with
NMR and MS analyses.
[6] a) S. A. Shelke, S. Th. Sigurdsson, Eur. J. Org. Chem. 2012; DOI: 10.1002/
ejoc.201101434; b) S. A. Shelke, S. T. Sigurdsson in Spin-Labels and Intrin-
sic Paramagnetic Centers in the Biosciences: Structural Information from
Distance Measurements (Ed.: C. Timmel), Springer, 2012; DOI: 10.1007/
430_2011_62.
[7] S. A. Shelke, S. Th. Sigurdsson, Angew. Chem. 2010, 122, 8156–8158;
Angew. Chem. Int. Ed. 2010, 49, 7984–7986.
[8] S. A. Shelke, S. Th. Sigurdsson, Nucleic Acids Res. 2011; DOI: 10.1093/
nar/gkr1210.
[9] V. V. Rostovtsev, L. G. Green, V. V. Fokin, K. B. Sharpless, Angew. Chem.
2002, 114, 2708–2711; Angew. Chem. Int. Ed. 2002, 41, 2596–2599.
[10] S. G. Alvarez, M. T. Alvarez, Synthesis 1997, 413–414.
[11] N. Barhate, P. Cekan, A. P. Massey, S. Th. Sigurdsson, Angew. Chem. 2007,
119, 2709–2712; Angew. Chem. Int. Ed. 2007, 46, 2655–2658.
[12] B. Frisch, C. Boeckler, F. Schuber, Bioconjugate Chem. 1996, 7, 180–186.
[13] M. S. Bernatowicz, Y. Wu, G. R. Matsueda, J. Org. Chem. 1992, 57, 2497–
2502.
Keywords: aminoxyl radicals · DNA · EPR spectroscopy ·
nitroxide · SDSL
[1] a) O. Schiemann, T. F. Prisner, Q. Rev. Biophys. 2007, 40, 1–53; b) G. Z.
Sowa, P. Z. Qin, Prog. Nucleic Acid Res. Mol. Biol. 2008, 82, 147–197; c) L.
Hunsicker-Wang, M. Vogt, V. J. DeRose, Methods Enzymol. 2009, 468,
335–367; d) O. Schiemann, Methods Enzymol. 2009, 469, 329–351;
e) X. J. Zhang, P. Cekan, S. Th. Sigurdsson, P. Z. Qin, Methods Enzymol.
2009, 469, 303–328; f) P. Z. Qin, P. Nguyen, WIREs RNA 2012 3, 62–72;
g) S. Th. Sigurdsson, Pure. Appl. Chem. 2011, 83, 677–686.
[2] a) J. C. Macosko, M. S. Pio, I. Tinoco, Jr., Y.-K. Shin, RNA 1999, 5, 1158–
1166; b) N.-K. Kim, A. Murali, V. J. DeRose, Chem. Biol. 2004, 11, 939–
948.
[3] a) T. Prisner, M. Rohrer, F. MacMillan, Annu. Rev. Phys. Chem. 2001, 52,
279–313; b) G. Jeschke, Y. Polyhach, Phys. Chem. Chem. Phys. 2007, 9,
1895–1910; c) G. W. Reginsson, O. Schiemann, Biochem. Soc. Trans.
2011, 39, 128–139.
[14] K. E. Borbas, P. Mroz, M. R. Hamblin, J. S. Lindsey, Bioconjugate Chem.
2006, 17, 638–653.
[4] a) D. E. Budil, S. V. Kolaczkowski, A. Perry, C. Varaprasad, F. Johnson, P. R.
Strauss, Biophys. J. 2000, 78, 430–438; b) T. E. Edwards, S. Th. Sigurds-
son, Biochemistry 2002, 41, 14843–14847; c) P. Z. Qin, K. Hideg, J.
Feigon, W. L. Hubbell, Biochemistry 2003, 42, 6772–6783; d) T. E. Ed-
wards, S. Th. Sigurdsson, Biochemistry 2005, 44, 12870–12878; e) P. Z.
Qin, J. Feigon, W. L. Hubbell, J. Mol. Biol. 2005, 351, 1–8; f) P. Cekan,
S. Th. Sigurdsson, J. Am. Chem. Soc. 2009, 131, 18054–18056; g) A. M.
Popova, T. Kꢃlai, K. Hideg, P. Z. Qin, Biochemistry 2009, 48, 8540–8550;
h) A. Marko, V. Denysenkov, D. Margraf, P. Cekan, O. Schiemann, S. Th.
Sigurdsson, T. F. Prisner, J. Am. Chem. Soc. 2011, 133, 13375–13379.
[15] C. R. Bertozzi, M. D. Bednarski, J. Org. Chem. 1991, 56, 4326–4329.
[16] B. Carboni, M. Vaultier, R. Carriꢄ, Tetrahedron 1987, 43, 1799–1810.
[17] a) F. Seela, K. Kaiser, Nucleic Acids Res. 1987, 15, 3113–3129; b) A. Ono,
C. N. Chen, L. S. Kan, Biochemistry 1991, 30, 9914–9922.
Received: November 17, 2011
Published online on March 8, 2012
690
www.chembiochem.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2012, 13, 684 – 690