Redesigned tetrads with altered hydrogen bonding patterns enable programming of quadruplex topologies

Article abstract of DOI:10.1039/b805227a

Guanosine analogs assemble to unusual nucleobase tetrads, restricting possible conformations of four-stranded DNAs. The Royal Society of Chemistry.

Full text of DOI:10.1039/b805227a

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Redesigned tetrads with altered hydrogen bonding patterns enable
programming of quadruplex topologiesw
¨
Armin Benz and Jorg S. Hartig*
Received (in Cambridge, UK) 27th March 2008, Accepted 18th July 2008
First published as an Advance Article on the web 29th July 2008
DOI: 10.1039/b805227a
Guanosine analogs assemble to unusual nucleobase tetrads,
restricting possible conformations of four-stranded DNAs.
incompatible with guanosines with respect to tetrad forma-
tion, it should be possible to program quadruplex topologies
for a given modified sequence. By exchanging a terminal
guanine tetrad with a tetrad that contains opposing arrange-
ments of two 8-oxoguanines and two xanthines, the formation
of eight hydrogen bonds should still be possible. In compar-
ison to the guanine tetrad, the orientation of the hydrogen
bonds is inverted at two positions; see Fig. 1A, black dots.
Nevertheless, since the same number of hydrogen bonds is
possible, quadruplexes containing 8-oxoguanosine/xanthosine
tetrads should display similar stabilities compared to their
Guanine-rich nucleic acids are prone to fold into four-
stranded structures called quadruplexes. Within such struc-
tures, four guanines engage in hydrogen bonds building a
guanine tetrad. In quadruplex structures, such tetrads are
stacked to form stable four-stranded secondary structures.1,2
While the core element, the guanine tetrad, is similar in all
known guanine quadruplexes, the orientation of the partici-
pating strands differs significantly from one quadruplex to
another and results in highly polymorphic folds.1,2 While
RNA quadruplexes prefer conformations with all-parallel
oriented strands due to the disfavored syn-conformation of
the glycosidic bond,3 DNA quadruplexes display a variety of
different topologies. Several factors affect DNA quadruplex
conformations. The quadruplex topology is largely determined
by the composition of the loops and the number of stacked
G-tetrads.4 For example, sequences with short loops favor
parallel orientation.5,6 For a given sequence, the conformation
can strongly depend on the presence and type of cation.
Probably the most characterized four-stranded sequence is
the hexameric repeat found in vertebrate telomeres including
those of humans.7 Four repeats are necessary to form an
intramolecular quadruplex structure. Interestingly, several
different topologies have been found.1,2 For example, a crystal
structure in the presence of K+ shows an all-parallel orienta-
tion, the so called propeller conformation.8 In addition, anti-
parallel orientation has been reported in Na+-containing
solutions.9,10 More recently, in the presence of K+ further
structures were observed that contain mixed arrangements of
both parallel and anti-parallel strand orientation.2,10,11 Circu-
lar dichroism is a powerful tool to distinguish different topol-
ogies. The all-parallel conformation displays a maximum of
circular dichroism at 265 nm, while folds containing anti-
parallel orientations are characterized by a maximum at
290 nm.12
In the many different conformations adopted by quadru-
plexes, the terminal tetrads are built from different guanines.
Upon exchanging guanosines with isostructural elements
Fig. 1 Programmable quadruplex topologies. A: Comparison of
tetrads composed of guanines (left) and 8-oxoguanines pairing with
xanthines. The tetrads differ in the orientation of two hydrogen bonds
(black dots). B: Two possible topologies of quadruplexes build by the
human telomeric sequence. Left: anti-parallel topology, right: parallel
topology. The loop connectivity is shown in dashed lines. C: Sequences
characterized in this study. The positions of the 8-oxoguanosine (O)
and xanthosine (X) modifications should make it possible to selectively
stabilize one topology compared to the other.
University of Konstanz, Department of Chemistry and Konstanz
Research School Chemical Biology (KoRS-CB), Universitatsstr. 10,
¨
78457 Konstanz, Germany. E-mail: joerg.hartig@uni-konstanz.de;
Fax: +49-7531-885140; Tel: +49-7531-884575
w Electronic supplementary information (ESI) available: Details of
experimental procedures, analytical data, sequences, and further de-
pictions of modified and mixed tetrads. See DOI: 10.1039/b805227a
ꢀc
This journal is The Royal Society of Chemistry 2008
4010 | Chem. Commun., 2008, 4010–4012

View Article Online
natural counterpart. On the other hand, variation of the
position of the guanosine analogs within the human telomeric
repeat should make it possible to shift the preference of the
quadruplex conformation towards a distinct topology. In
parallel quadruplexes different guanosines participate in one
tetrad compared to anti-parallel topologies. When designing a
sequence that should fold into an anti-parallel topology, the
alternative, parallel fold contains two tetrads each lacking two
hydrogen bonds and vice versa (for a depiction of such
destabilizing tetrads see ESIw). Hence, it should be possible
to force the human telomeric sequence to adopt one fold or the
other by programming hydrogen bonding patterns within
tetrads composed of nucleobases other than guanosine.
The presented strategy should prove valuable if a specific
conformation of a given sequence needs to be studied. For
example, compounds binding to quadruplexes could be eval-
uated for their potential to differentiate between certain con-
formations. Recently described compounds differentiating
between quadruplex structures have the potential to specifi-
cally target a quadruplex of interest leaving other naturally
occurring G-rich sequences unaffected.13 In addition, the
approach enables the functional characterization of individual
quadruplex conformations.
In order to test our hypothesis, we have synthesized four
sequences related to the human telomeric repeat that each
contain two 8-oxoguanosine and two xanthosine modifications
instead of guanosine, see Fig. 1C. Double NPE-protected,
DMT-protected deoxyriboxanthosine phosphoramidite was
synthesized according to published procedures.14 8-Oxo-
deoxyguanosine phosphoramidite was commercially available
and was incorporated together with the deoxyxanthosine
building blocks into sequences HT-ap, HT-ap (inv), HT-p,
and HT-p (inv) using standard solid phase oligonucleotide
chemistry. The oligonucleotides were deprotected by treatment
with 1 M DBU followed by conventional ammonia cleavage
and deprotection at 37 1C.15 The oligonucleotides were purified
by PAGE and subsequently analyzed by ESI-MS as well as
analytical PAGE; for details of oligonucleotide synthesis see
ESI.w
Next, circular dichroism (CD) spectra were recorded in
order to determine the predominant topology of the modified
G-rich sequences. As a reference, an unmodified DNA strand
composed of the vertebrate telomeric repeat (HT-c, see
Fig. 1C) was analyzed. We chose potassium-containing buffer
conditions that are known to favor the presence of an equili-
brium of both parallel and anti-parallel topologies of the
telomeric sequence.2,10,11 Indeed, the unmodified sequence
displayed a maximum at 290 nm together with a pronounced
shoulder at shorter wavelengths indicative of a mixture of anti-
parallel and parallel topologies, see Fig. 2A, black line. Inter-
estingly, the modified sequence HT-p containing a tetrad
program that should favor the parallel topology displayed a
CD spectrum resembling a parallel fold whilst the second
sequence HT-ap designed to adopt an anti-parallel conforma-
tion showed a clear shift to the anti-parallel fold. As a
confirmation, the sequences HT-ap (inv) and HT-p (inv) were
also characterized. The sequences differ by inversion of the
respective 8-oxo-G and xanthosine positions, see Fig. 1C. The
sequences should as well make it possible to program anti-
Fig. 2 Circular dichroism studies. A: CD spectra of the sequences
HT-c: black; HT-ap: red; HT-ap (inv): orange; HT-p: blue; HT-p (inv):
green; recorded in 10 mM Tris, pH 7.5, 25 mM KCl at 25 1C. B:
Thermal denaturation studies followed by CD spectroscopy. Sequence
HT-c was observed at 265 nm (gray) and 290 nm (black), HT-p
sequences at 265 nm and HT-ap sequences at 290 nm. Buffer condi-
tions were similar to 2A, data were normalized for better comparison.
parallel and parallel topologies. Indeed, the inverted sequences
display similar CD spectra compared to HT-ap and HT-p,
respectively. Hence, by redesigning the hydrogen bonding
pattern of one tetrad by exchanging four guanines for two
pairs of xanthines and 8-oxoguanines we succeeded in pro-
gramming the overall topology of the quadruplex fold. In
order to prove that indeed intramolecular quadruplexes are
formed when a synthetic tetrad is present, we carried out gel
shift experiments. All modified sequences display a band that
corresponds to monomolecular quadruplexes although higher-
order aggregates are visible as well (see ESIw).
The CD spectra show a clear shift to the expected topo-
logies. Nevertheless, especially the spectra of the anti-parallel
sequences still display a slight shoulder around 290 nm
corresponding to the disfavored topology. Hence it seems that
there is still a minor fraction of the alternative fold present. In
order to assess whether the redesigned tetrads interfere with
the overall stability of the quadruplex, we carried out CD
melting experiments. For comparison, the unmodified se-
quence HT-c was used. We observed the temperature-depen-
dent decrease of the CD signal at both 265 and 290 nm. This
makes it possible to assess the stabilities of the anti-parallel as
well as the parallel fold of the unmodified sequence, see
Fig. 2B. Both folds showed significant stabilities represented
by melting points of 49 1C for the anti-parallel conformation
and of 55.5 1C for the parallel fold. In contrast, when the
ꢀc
This journal is The Royal Society of Chemistry 2008
Chem. Commun., 2008, 4010–4012 | 4011

View Article Online
stability of the xanthine/8-oxoguanine-tetrad-containing
quadruplexes was determined, both the parallel folds
(observed at 265 nm) and the anti-parallel folds (observed at
290 nm) showed significantly lowered melting temperatures of
40 1C or below, see Fig. 2B. Although the incorporation of an
8-oxoguanine/xanthine-tetrad makes it possible to favor one
quadruplex topology over the other, incorporation of the
unnatural tetrad slightly destabilizes the quadruplex structure
compared to a conventional tetrad made up from guanines.
In conclusion, the use of 8-oxoguanosine in combination
with xanthosine makes it possible to program the overall
topology of quadruplexes by redesigning constructive hydro-
gen bonding patterns. Notably, the strategy should even make
it possible to program topologies very specifically. For exam-
ple, the basket, chair, and dogear conformations of intra-
molecular quadruplexes differ in the guanosines participating
in the external tetrads and hence should be individually
programmable using this approach. The present study proves
the viability of the concept whereas more elaborate structural
studies will be necessary in order to differentiate between the
mentioned conformations.
sequences,22 it is remarkable that the difference of four out
of in total 24 possible hydrogen bonds (eight per tetrad) is able
to tip the folding equilibrium between anti-parallel and paral-
lel quadruplexes to one side or the other (folding of the non-
programmed conformation results in the loss of two hydrogen
bonds in two tetrads together with potential steric hindrance in
one tetrad, see ESIw). Besides stabilizing factors such as
nucleobase stacking and hydrogen bonding, electrostatic in-
teractions play an important role, particularly in quadruplex
structures.23 Hence, it will be interesting to characterize the
behavior of the modified sequences with respect to different
concentrations and types of cations. Anyway, only small
differences in energetic stability of anti-parallel and parallel
telomeric sequences have been found which could explain the
viability of the presented approach.5,11
We thank Elmar Weinhold for sparking discussions, Astrid
Joachimi and Vicki Lee for excellent technical assistance. JSH
gratefully acknowledges the VolkswagenStiftung for funding a
Lichtenberg-Professorship.
Notes and references
A variety of nucleobase analogs participating in quadruplex
structures have been studied before. For example, Mergny and
co-workers have characterized several guanosine derivatives in
the context of tetrameric quadruplexes, finding mostly negative
effects on stability and assembly kinetics of the four-stranded
structures.16 In addition, 8-bromoguanosine-modified G-rich
sequences showed altered quadruplex structures.17,18 Incor-
poration of adenosine analogs has been reported to result in
changed conformations in tetrameric quadruplexes.19 Incor-
poration of ribonucleosides into DNA quadruplexes has been
shown to restrict the possible conformations to parallel strand
orientations. The likely reason for this observation is the
preference of the anti-conformation of the ribonucleoside
glycosidic bond.3 In contrast, substitutions in the 8-position
favor the syn-conformation.20 Nevertheless, the anti-conforma-
tion of the glycosidic bond is still possible with guanosine
analogs modified at C8.17,21 The parallel conformation of the
natural telomeric sequence in its K+ form contains only anti-
conformations of the glycosidic bond.8
The basket-type Na+ form of the natural telomeric se-
quence displays alternating anti–syn–anti–syn conformations
of guanines 1, 6, 7, and 12 of the first tetrad referring to the
numbering in Fig. 1.9,10 Since 8-oxo-G favors the syn-con-
formation, one out of the two 8-oxo-Gs is in the disfavored
anti-conformation in both HT-ap as well as HT-ap (inv). In
the parallel form, both 8-oxo-Gs would adopt the disfavored
glycosidic conformation. This would represent an explanation
for the decreased stabilities of all modified quadruplexes
compared to the natural sequence as shown in Fig. 2B. Never-
theless, the question of whether the structures of the modified,
programmed sequences adopt the exact conformations of the
natural sequences can only be addressed with more sophisti-
cated structural studies. For a detailed review of the glycosidic
conformations in vertebrate telomeric quadruplexes see ref. 2.
In addition, the presented findings are interesting with
respect to the contribution of hydrogen bonds to the overall
stability of quadruplex sequences. Since stacking interactions
1 S. Burge, G. N. Parkinson, P. Hazel, A. K. Todd and S. Neidle,
Nucleic Acids Res., 2006, 34, 5402.
2 D. J. Patel, A. T. Phan and V. Kuryavyi, Nucleic Acids Res., 2007,
35, 7429.
3 C. F. Tang and R. H. Shafer, J. Am. Chem. Soc., 2006, 128, 5966.
4 P. A. Rachwal, T. Brown and K. R. Fox, Biochemistry, 2007, 46,
3036; A. Bugaut and S. Balasubramanian, Biochemistry, 2008, 47,
689.
5 P. Hazel, J. Huppert, S. Balasubramanian and S. Neidle, J. Am.
Chem. Soc., 2004, 126, 16405.
6 P. A. Rachwal, T. Brown and K. R. Fox, FEBS Lett., 2007, 581,
1657.
7 R. K. Moyzis, J. M. Buckingham, L. S. Cram, M. Dani, L. L.
Deaven, M. D. Jones, J. Meyne, R. L. Ratliff and J. R. Wu, Proc.
Natl. Acad. Sci. U. S. A., 1988, 85, 6622.
8 G. N. Parkinson, M. P. Lee and S. Neidle, Nature, 2002, 417, 876.
9 Y. Wang and D. J. Patel, Structure, 1993, 1, 263.
10 A. T. Phan, V. Kuryavyi, K. N. Luu and D. J. Patel, Nucleic Acids
Res., 2007, 35, 6517.
11 L. Ying, J. J. Green, H. Li, D. Klenerman and S. Balasubramanian,
Proc. Natl. Acad. Sci. U. S. A., 2003, 100, 14629.
12 P. Balagurumoorthy and S. K. Brahmachari, J. Biol. Chem., 1994,
269, 21858; D. M. Gray, J. D. Wen, C. W. Gray, R. Repges, C.
Repges, G. Raabe and J. Fleischhauer, Chirality, 2008, 20, 431; S.
Paramasivan, I. Rujan and P. H. Bolton, Methods, 2007, 43, 324.
13 Z. A. Waller, P. S. Shirude, R. Rodriguez and S. Balasubramanian,
Chem. Commun., 2008, 1467; A. Bugaut, K. Jantos, J. L. Wietor,
R. Rodriguez, J. K. M. Sanders and S. Balasubramanian, Angew.
Chem., Int. Ed., 2008, 47, 2677.
14 R. Eritja, D. M. Horowitz, P. A. Walker, J. P. Ziehler-Martin, M.
S. Boosalis, M. F. Goodman, K. Itakura and B. E. Kaplan, Nucleic
Acids Res., 1986, 14, 8135; A. Vanaerschot, M. Mag, P. Herdewijn
and H. Vanderhaeghe, Nucleosides Nucleotides, 1989, 8, 159.
15 H. Sigmund and W. Pfleiderer, Helv. Chim. Acta, 2003, 86, 2299.
16 J. Gros, F. Rosu, S. Amrane, A. De Cian, V. Gabelica, L. Lacroix
and J.-L. Mergny, Nucleic Acids Res., 2007, 35, 3064.
17 V. Esposito, A. Randazzo, G. Piccialli, L. Petraccone, C. Giancola
and L. Mayol, Org. Biomol. Chem., 2004, 2, 313; Y. Xu, Y.
Noguchi and H. Sugiyama, Bioorg. Med. Chem., 2006, 14, 5584.
18 L. Petraccone, I. Duro, A. Randazzo, A. Virno, L. Mayol and C.
Giancola, Nucleosides, Nucleotides Nucleic Acids, 2007, 26, 669.
19 L. Petraccone, I. Duro, E. Erra, A. Randazzo, A. Virno and C.
Giancola, Nucleosides, Nucleotides Nucleic Acids, 2007, 26, 675.
20 G. W. Hsu, M. Ober, T. Carell and L. S. Beese, Nature, 2004, 431, 217.
21 A. Virgilio, V. Esposito, A. Randazzo, L. Mayol and A. Galeone,
Nucleic Acids Res., 2005, 33, 6188.
22 E. T. Kool, Acc. Chem. Res., 2002, 35, 936.
23 I. V. Smirnov and R. H. Shafer, Biopolymers, 2007, 85, 91.
likely play
a
major role in stabilizing four-stranded
ꢀc
This journal is The Royal Society of Chemistry 2008
4012 | Chem. Commun., 2008, 4010–4012