Noncovalent and site-directed spin labeling of nucleic acids

Article abstract of DOI:10.1002/anie.201002637

Spin label parking: The rigid spin label c (magenta), which is an analogue of cytidine, binds site-specifically to abasic sites in duplex DNAs, through hydrogen bonding to guanine (blue) and π stacking with the flanking base pairs. EPR spectroscopy shows

Full text of DOI:10.1002/anie.201002637

Communications
DOI: 10.1002/anie.201002637
Noncovalent Spin Labeling
Noncovalent and Site-Directed Spin Labeling of Nucleic Acids**
Sandip A. Shelke and Snorri Th. Sigurdsson*
Electron paramagnetic resonance (EPR) spectroscopy is
widely used to study free radicals or paramagnetic centers
associated with biopolymers.[1] With the advent of pulsed
EPR methods, which allow accurate distance measurements
between 20 and 80 ꢀ, structures of biopolymers have
increasingly been interrogated by this technique.[2] Some of
the advantages of EPR spectroscopy over other structural
techniques are its sensitivity, that it is not restricted by
molecular size, and that measurements can be performed
under biological conditions.[1b] However, stable free radicals,
such as nitroxide spin labels, must be incorporated into the
biopolymers prior to EPR studies.
In site-directed spin labeling (SDSL), spin labels are
covalently attached to the biopolymers at a specific site of
interest.[3] For nucleic acids there have been two main
strategies for SDSL. First, spin labels have been incorporated
during automated oligonucleotide synthesis by employing
spin-labeled phosphoramidite building blocks.[3b,c,e] This
approach has the advantage that very sophisticated and
structurally complex labels can be incorporated at specific
sites. However, the synthetic challenges of spin-labeled
phosphoramidites can be considerable.[3e] Furthermore, spin
labels can be partially reduced upon exposure to the reagents
used in the automated synthesis of oligonucleotides.[4] The
second SDSL approach is post-synthetic modification of the
biopolymer.[3b,c,e] Here, a spin-labeling reagent is incubated
with an oligonucleotide that contains a reactive functional
group at a specific site. Post-synthetic labeling is in general
less labor intensive than the phosphoramidite strategy, but
drawbacks include incomplete labeling and side reactions of
the spin label with inherent functional groups of the nucleic
acids, such as the exocyclic amino groups of the nucleobases.
Both strategies usually require purification of the spin-labeled
material, which can be nontrivial. Here we report a new and
straightforward SDSL protocol for nucleic acids that is based
on noncovalent labeling.
Figure 1. a) Base-pairing scheme of spin labels C¸ and ꢀ with G. dR is
2’-deoxyribose. b) Structure of an abasic site in DNA.
rigidity of C¸ enabled precise distance measurements by EPR,
determination of the relative angular orientation between two
spin labels,[6] and has been used to study DNA dynamics and
folding.[4c,5,7] The strategy for noncovalent labeling was to
disconnect the glycosidic bond of C¸ to give an abasic site (F)
and the free spin-labeled base ꢀ (Figure 1). The spin label
would bind in the abasic site through receptor–ligand
interactions involving hydrogen bonding and p-stacking
interactions.[8]
The synthesis of spin label ꢀ started with regioselective
alkylation of 5-bromouracil at N1 by a one-pot, two-step
procedure using HMDS and benzyl bromide in the presence
of a catalytic amount of iodine to obtain compound 2
(Scheme 1).[9] Activation of 2 by conversion to the O4-
sulfonylated derivative,[10] followed by coupling with isoindol
amino phenol derivative 4 yielded conjugate 5.[5] Subsequent
ring closure, facilitated by cesium fluoride, yielded phenox-
azine derivative 6.[11] Removal of the N1-protecting benzyl
group by boron tribromide[12] and oxidation of the amine to a
nitroxide[13] with mCPBA gave spin label ꢀ.
The new approach utilizes a nitroxide that is structurally
related to the rigid spin label C¸ .[5] The spin label C¸ is an
analogue of cytidine (C), with a nitroxide-bearing isoindol
moiety fused to cytosine by an oxazine linkage, and forms a
stable Watson–Crick base pair with guanine (Figure 1). The
The EPR spectrum of ꢀ in an aqueous solution containing
ethylene glycol (30%) shows three narrow lines that broaden
on reducing the temperature from 0 to À308C (Figure 2, left),
due to slower tumbling of ꢀ in solution.[3c,e] On mixing a DNA
duplex containing an abasic site with ꢀ, a slow-moving
component appears in the EPR spectrum (shown by arrows,
Figure 2 middle), indicating binding of the spin label to the
abasic site. On further cooling, the extent of spin-label
binding increased, and at À308C the narrow lines (the fast-
motion component of the spectrum) had completely disap-
peared, consistent with the spin label being fully bound. For
comparison, EPR spectra of a covalently C¸ -labeled 14-mer
were recorded under identical conditions (Figure 2, right).
The mobility of the spin label that is covalently linked to the
dsDNA is the same as that of the slow-moving component in
the sample containing ꢀ and the abasic DNA (Figure 2,
[*] S. A. Shelke, Prof. Dr. S. Th. Sigurdsson
Science Institute, University of Iceland
Dunhaga 3, 107 Reykjavik (Iceland)
Fax: (+354)552-8911
E-mail: snorrisi@hi.is
[**] This work was supported by the Icelandic Research Fund
(080041022) and the University of Iceland Research Fund, including
a doctoral fellowship to S.A.S.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201002637.
7984
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 7984 –7986

Angewandte
Chemie
bind nearly two equivalents of spin label (Supporting
Information).
Experiments were carried out to verify that the spin label
was binding directly to the abasic site, rather than nonspecifi-
cally to the DNA, for example, by intercalation or groove
binding. First, spin label ꢀ was titrated into a solution
containing a fixed concentration of abasic DNA and EPR
spectra were recorded at À308C (Figure 3a). At 0.5 and
Scheme 1. Synthesis of spin label ꢀ. Bn=benzyl, DMAP=4-dimethyla-
minopyridine, HMDS=1,1,1,3,3,3-hexamethyldisilazane, mCPBA=
meta-chloroperbenzoic acid, TMS=trimethylsilyl, TPS=2,4,6-triisopro-
pylbenzenesulfonyl.
Figure 3. a) EPR spectra of 0.5 (left), 1 (middle), and 2 equiv (right) of
spin label ꢀ in the presence of a dsDNA (100 mm) containing an
abasic site, recorded at À308C with the same number of scans and
phase-corrected. The DNA sequence is the same as for the middle
panel in Figure 2. b) EPR spectra of spin label ꢀ (100 mm) in the
presence of an unmodified dsDNA (100 mm) containing the same
sequence as in a) except that the abasic site has been replaced by C.
The EPR spectra were recorded in phosphate buffer (pH 7.0) contain-
ing 30% ethylene glycol and 2% DMSO.
1.0 equiv of ꢀ, the spin label was fully bound, whereas both the
bound and excess free spin label were observed for 2.0 equiv
of ꢀ. Thus, the spin label binds readily until the abasic binding
site has been saturated. Second, the spin label was incubated
with an unmodified DNA duplex of the same sequence, that
is, a sequence that did not contain an abasic site. No binding
was observed down to À208C, but at À308C a small amount
(< 5%) of binding was observed (Figure 3b), and indicates
minor nonspecific binding in the absence of an abasic site.
Third, a duplex containing a G overhang did not show
increased nonspecific binding (Supporting Information),
although the spin label can, in principle, bind to the unpaired
G through Watson–Crick pairing while stacking at the end of
the duplex. This experiment clearly shows that, in addition to
hydrogen bonding, efficient binding requires p stacking of
both faces of the spin label, as provided by an abasic site in a
duplex region.
The specificity of spin-label binding was also evaluated by
determining the change in dissociation constant Kd for abasic
DNA on addition of a large excess of unmodified DNA;[14] Kd
was determined to be (6.95 Æ 0.8) mm at À108C but (3.7 Æ
0.3) mm (Kd’) in the presence of a tenfold excess of unmodified
DNA. High specificity of binding is reflected in the small
change in the dissociation constant.
Figure 2. EPR spectra of spin label ꢀ (left), of ꢀ (200 mm) in the
presence of abasic DNA (400 mm) (middle), and of a C¸-labeled 14-mer
(right). The EPR spectra were recorded in phosphate buffer (pH 7.0)
containing 30% ethylene glycol and 2% DMSO at different temper-
atures, with the same number of scans, phase corrected and aligned
to the height of central peak. F denotes an abasic site.
To probe specific interactions of the spin label within the
abasic site, the structure of the abasic site was systematically
varied. Specifically, the contribution of the base opposite the
abasic site, which was expected to provide hydrogen-bonding
sites for the label, to the binding affinity was evaluated. To
middle) at all temperatures, and this provides further
evidence for noncovalent binding of ꢀ to the abasic DNA. A
duplex DNA containing two abasic sites was also shown to
Angew. Chem. Int. Ed. 2010, 49, 7984 –7986
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7985

Communications
this end, binding of ꢀ to four different DNA duplexes,
Keywords: DNA · EPR spectroscopy · nucleic acids ·
receptor–ligand interactions · spin labels
.
containing either A, C, G, or Topposite to the abasic site, was
determined by EPR spectroscopy at À108C. Visual inspection
of the spectra showed significant differences in extent of
binding; as expected, the spin label bound most efficiently
when it was paired with G (Supporting Information). A plot
of the dissociation constants of spin-label binding as a
function of sequence shows a six- to eightfold higher affinity
of the spin label with G than with A, C, or T (Figure 4). The
[1] a) L. Hunsicker-Wang, M. Vogt, V. J. DeRose, Methods Enzy-
mol. 2009, 468, 335 – 367; b) E. J. Hustedt, A. H. Beth, Annu.
Rev. Biophys. Biomol. Struct. 1999, 28, 129 – 153; c) T. Prisner,
M. Rohrer, F. MacMillan, Annu. Rev. Phys. Chem. 2001, 52, 279 –
313; d) M. Ubbink, J. Worrall, G. Canters, E. Groenen, M.
Huber, Annu. Rev. Biophys. Biomol. Struct. 2002, 31, 393 – 422.
[2] a) O. Schiemann, Methods Enzymol. 2009, 469, 329 – 351; b) O.
Schiemann, T. F. Prisner, Q. Rev. Biophys. 2007, 40, 1 – 53;
c) Y. D. Tsvetkov, Y. A. Grishin, Instrum. Exp. Tech. 2009, 52,
615 – 636.
[3] a) W. L. Hubbell, D. S. Cafiso, C. Altenbach, Nat. Struct. Biol.
2000, 7, 735 – 739; b) O. Schiemann, N. Piton, J. Plackmeyer,
B. E. Bode, T. F. Prisner, J. W. Engels, Nat. Protoc. 2007, 2, 904 –
923; c) G. Z. Sowa, P. Z. Qin, Prog. Nucleic Acid Res. Mol. Biol.
2008, 82, 147 – 197; d) J. P. Klare, H. J. Steinhoff, Photosynth. Res.
2009, 102, 377 – 390; e) X. J. Zhang, P. Cekan, S. T. Sigurdsson,
P. Z. Qin, Methods Enzymol. 2009, 469, 303 – 328.
[4] a) P. M. Gannett, E. Darian, J. Powell, E. M. Johnson, C.
Mundoma, N. L. Greenbaum, C. M. Ramsey, N. S. Dalal, D. E.
Budil, Nucleic Acids Res. 2002, 30, 5328 – 5337; b) N. Piton, Y. G.
Mu, G. Stock, T. F. Prisner, O. Schiemann, J. W. Engels, Nucleic
Acids Res. 2007, 35, 3128 – 3143; c) P. Cekan, A. L. Smith, N.
Barhate, B. H. Robinson, S. T. Sigurdsson, Nucleic Acids Res.
2008, 36, 5946 – 5954.
Figure 4. Dissociation constants of ꢀ binding to dsDNA as a function
of the nucleotide opposite to the abasic site (denoted F), determined
at À108C (400 mm abasic dsDNA, 200 mm ꢀ in a phosphate buffer,
pH 7.0, containing 30% ethylene glycol and 2% DMSO). The DNA
sequence is shown on top.
[5] N. Barhate, P. Cekan, A. P. Massey, S. T. Sigurdsson, Angew.
Chem. 2007, 119, 2709 – 2712; Angew. Chem. Int. Ed. 2007, 46,
2655 – 2658.
[6] O. Schiemann, P. Cekan, D. Margraf, T. F. Prisner, S. T.
Sigurdsson, Angew. Chem. 2009, 121, 3342 – 3345; Angew.
Chem. Int. Ed. 2009, 48, 3292 – 3295.
spin label binds with similar affinity to A and C and slightly
higher affinity to T, in agreement with previously reported
stability data for C mismatches.[15] This data is, therefore,
supportive of a successful design: the spin label binds
specifically in the abasic site, where it forms hydrogen
bonds with G opposite to the abasic site.
[7] P. Cekan, E. O. Jonsson, S. T. Sigurdsson, Nucleic Acids Res.
2009, 37, 3990 – 3995.
In summary, we have developed a general and straightfor-
ward approach for site-directed spin labeling of nucleic acids
using receptor–ligand interactions. Binding sites can simply
be carved out at specific locations in the nucleic acid by
incorporation of abasic sites, as long as the labeling site is in a
duplex region. This strategy has clear advantages over the
traditional spin-labeling methods using spin-labeled phos-
phoramidites or post-synthetic labeling, as it enables labeling
by simply adding the spin label to a solution of commercially
available nucleic acids that contain abasic sites. Since the
design of ꢀ is based on the rigid spin label C¸ , it should have the
same spectroscopic properties and facilitate extraction of
both accurate distances and relative orientations by EPR
spectroscopy. Although the noncovalent, site-directed spin-
labeling (NC-SDSL) method requires cooling the sample to
À308C to achieve complete binding, it can be readily used for
structural studies by pulsed EPR spectroscopic techniques
such as DEER or PELDOR, which are all carried out in
frozen solutions. Easy access to site-specifically labeled
nucleic acids will facilitate their structural studies by pulsed
EPR spectroscopy, and applications along these lines will be
reported in due course.
[8] a) K. Fukui, K. Tanaka, Nucleic Acids Res. 1996, 24, 3962 – 3967;
b) A. Peracchi, L. Beigelman, N. Usman, D. Herschlag, Proc.
Natl. Acad. Sci. USA 1996, 93, 11522 – 11527; c) J. Stivers,
Nucleic Acids Res. 1998, 26, 3837 – 3844; d) J. Lhomme, J.
Constant, M. Demeunynck, Biopolymers 1999, 52, 65 – 83; e) F.
Thomas, J. Michon, J. Lhomme, Biochemistry 1999, 38, 1930 –
1937; f) B. Rajendar, S. Nishizawa, N. Teramae, Org. Biomol.
Chem. 2008, 6, 670 – 673; g) N. Sankaran, Y. Sato, F. Sato, B.
Rajendar, K. Morita, T. Seino, S. Nishizawa, N. Teramae, J. Phys.
Chem. B 2009, 113, 1522 – 1529; h) K. Maekawa, S. Nakazawa,
H. Atsumi, D. Shiomi, K. Sato, M. Kitagawa, T. Takui, K.
Nakatani, Chem. Commun. 2010, 46, 1247 – 1249.
[9] a) H. Aoyama, Bull. Chem. Soc. Jpn. 1987, 60, 2073 – 2077; b) V.
Malik, P. Singh, S. Kumar, Tetrahedron 2005, 61, 4009 – 4014.
[10] C. Hꢁbartner, S. K. Silverman, Angew. Chem. 2005, 117, 7471 –
7475; Angew. Chem. Int. Ed. 2005, 44, 7305 – 7309.
[11] K. Rajeev, M. Maier, E. Lesnik, M. Manoharan, Org. Lett. 2002,
4, 4395 – 4398.
[12] N. G. Kundu, R. P. Hertzberg, S. J. Hannon, Tetrahedron Lett.
1980, 21, 1109 – 1112.
[13] T. R. Miller, S. C. Alley, A. W. Reese, M. S. Solomon, W. V.
McCallister, C. Mailer, B. H. Robinson, P. B. Hopkins, J. Am.
Chem. Soc. 1995, 117, 9377 – 9378.
[14] V. Bonnard, S. Azoulay, A. Di Giorgio, N. Patino, Chem.
Commun. 2009, 2302 – 2304.
[15] F. Aboul-ela, D. Koh, I. J. Tinoco, F. Martin, Nucleic Acids Res.
1985, 13, 4811 – 4824.
Received: May 2, 2010
Revised: July 7, 2010
Published online: September 15, 2010
7986 www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 7984 –7986