The use of a peptidic scaffold for the formation of stable guanine tetrads: Control of a H-bonded pattern in water

Article abstract of DOI:10.1002/chem.201003556

With a little help from a scaffold: The pre-organization of guanine units onto a peptidic scaffold allows a remarkable stabilization of the G-quartet motif (see graphic). Indeed, the quartet motif was found to be stable even in water without addition of s

Full text of DOI:10.1002/chem.201003556

COMMUNICATION
DOI: 10.1002/chem.201003556
The Use of a Peptidic Scaffold for the Formation of Stable Guanine Tetrads:
Control of a H-bonded Pattern in Water
Pierre Murat, Bꢀatrice Gennaro, Julian Garcia, Nicolas Spinelli, Pascal Dumy, and
Eric Defrancq*[a]
Controlling the formation of hydrogen bonds is crucial for
the development of supramolecular architectures. In this
context, biological systems are the source of inspiration of
supramolecular research for designing two- and three-di-
mensional nanostructures of increasing complexity. Among
all biological macromolecules, DNA is certainly the most
promising candidate. Indeed, DNA-based materials are rele-
vant examples of structures for which the self-assembly of
hydrogen-bonded nucleobases is used to create highly or-
dered and complex architectures.[1] In particular, guanine de-
rivatives are known to self-assemble through hydrogen
bonding into various discrete architectures, including dimers,
ribbons, and quartets.[2] The G-quartet formed by the copla-
nar arrangement of four Hoogsteen-paired guanine bases
represents a relevant model for studies of molecular self-as-
sembly and noncovalent synthesis. Stacked guanine tetrads
are found in particular DNA secondary structures, such as
G-quadruplex DNA, which is hypothesized to have conse-
quences in several biological processes.[3] Furthermore, gua-
nine-based structures are of interest for applications in mo-
lecular electronics.[4]
One challenging goal is to organize small molecules, such
as the guanine building blocks, into well-defined assemblies.
Self-assembly of guanines has been mainly exploited by
using lipophilic systems to create smart supramolecular ar-
chitectures in organic solvents. Indeed, water is a highly
competitive solvent that disfavors hydrogen bonds between
small molecules. Cations also play an important role in the
structuring of assemblies based on the G-quartet motif.[5]
For these reasons, self-assembly of guanines in water are
mainly carried out in high concentrations of guanines and
salts. Rivera and co-workers have recently studied the self-
assembly of a 2’-deoxyguanosine derivative, which was
modified at the C8 position by a m-acetylphenyl group and
at the 2’-position by a dimethylamino moiety. At neutral
pH, the positively charged ammonium groups render the
molecule soluble in water. Interestingly, they demonstrated
that the molecule is able to self-assemble in water as a
result of the formation of a hexadecamer.[6] They attributed
the enhanced stability to a combination of factors, including
the formation of extra H bonds and increased p-stacking in-
teractions. Nevertheless, the studies were performed by
using 1m KCl, which emphasizes the difficulties of creating
hydrogen bonds between small molecules in water.
For stabilization of only one G-quartet tetrad, the use of a
scaffold has been proposed. Pioneered by Davis and co-
workers, guanosine moieties were anchored to a calixarene
to constrain the nucleobases in the assembled conforma-
tion.[7] More recently Sherman and Nikan reported the use
of a cavitand scaffold and showed a higher and cation-inde-
pendent stability due to the pre-organization of the guanine
units.[8] However, due to the hydrophobic nature of the scaf-
fold, the studies were carried out in organic solvent.[9]
Herein, we report on a new system that is able to form a
stable G-tetrad in water without the addition of any cations.
The method is based on the use of a cyclodecapeptide scaf-
fold introduced by Mutter for the design of template-assem-
bled synthetic proteins.[10] This scaffold has been used re-
cently for the efficient formation of G-quadruplexes.[11]
Compounds 3 and 4 have been designed and synthesized
by Huisgen 1,3-dipolar cycloaddition (Scheme 1). Interest-
ingly, the resulting triazole linkers are shown to actively par-
ticipate in the structuring of the G-quartet, as described
below.[12] Compound 3 was prepared from the cyclodecapep-
tide scaffold 1, which was functionalized with alkyne moiet-
ies by incorporating propargylglycine-modified amino acids
during the solid-phase peptide synthesis. The coupling with
5’-azido-2’,3’-O-isopropylideneguanosine (2)[13] was achieved
by 1,3-dipolar cycloaddition, which was catalyzed by CuI, af-
fording the desired compound in 79% yield after purifica-
tion. The solubility of 3 in water was too low for NMR stud-
ies and the isopropylidene residues were therefore removed
under acidic conditions (TFA/H20, 50:50, v/v) to obtain 4 in
69% yield. Compound 4 was found to be fully soluble in
water (see the Supporting Information).
[a] P. Murat, B. Gennaro, Prof. J. Garcia, Dr. N. Spinelli, Prof. P. Dumy,
Prof. E. Defrancq
1H NMR analysis of starting-unit 2 was carried out in
DMSO (Figure 1a) and revealed a well-resolved spectrum
with the chemical shift of the imino proton (NH) at d=
10.69 ppm, which is the classical value found for free gua-
nine in solution. 1H NMR analysis was also performed in
water with guanosine, because 2 is not soluble under these
Dꢀpartement de Chimie Molꢀculaire UMR CNRS 5250
Universitꢀ Joseph Fourier BP 53
38041 Grenoble cedex 9 (France)
Fax : (+33)456520803
E-mail: eric.defrancq@ujf-grenoble.fr
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201003556.
Chem. Eur. J. 2011, 17, 5791 – 5795
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5791

in water. For compounds 3 and 4, NMR analyses were car-
ried out in a mixture of H2O/CD3CN and in water, respec-
tively. Prior to NMR-sample preparation, the compounds
were thoroughly desalted by size-exclusion chromatography
(see the Supporting Information). Interestingly, the imino
proton is detected and resonates at d=11.33 ppm for 3,
whereas a duplication of the imino proton signal, at d=
11.01 and 10.70 ppm, is observed for 4 (Figure 1d). The ob-
servation of imino signals in water as well as the higher
chemical-shift values in comparison with 2 indicate the exis-
tence of hydrogen bonds and suggest that the guanine bases
1
are organized. Furthermore, H NMR analysis was carried
out in H2O/CD3CN by varying the temperature from À10 to
508C to access the stability of the structured system. The
disappearance of the signal corresponding to imino protons
was observed when the temperature reached 308C (see also
thermal denaturation studies by CD analysis). All these data
demonstrates that the scaffold allows efficient formation of
a stable, structured guanosine motif in an aqueous solution.
A detailed NMR conformational study by using TOCSY,
double-quantum-filter (DQF)-COSY, and NOESY tech-
niques was then performed (see the Supporting Informa-
tion).
2D-NOESY experiments, which were used for the full as-
signment of protons chemical shifts of compound 3 (see
Figure 11–14 and Table 1 in the Supporting Information),
showed a number of correlations between the protons of the
guanosines moieties and the protons of the peptidic scaffold.
In particular, strong correlations between the H8 and H1’
for 3 are observed; this is indicative of a syn conformation
along the glycosidic bond (Figure 2).[14] This conformation is
known to prevent the formation of structures, such as gua-
nine ribbons,[15] and thus validates the formation of the gua-
nine quartet under these conditions. Note that the imino
proton signal is slightly split. Two NOE correlations, be-
tween H8 and H4’ for two guanosine residues and between
NHPra (amido proton of the peptidic scaffold, see the Sup-
porting Information for the nomenclature) and H4’ for two
Scheme 1. Synthesis of compounds 3 and 4. 1) Cu0, phosphate-buffered
saline (PBS)/DMSO/tBuOH (50:40:10; 10 mm PBS, 0.3m NaCl, pH 8.0),
558C; 2) trifluoroacetic acid (TFA)/H2O (50:50). Pra is the l-propargyl-
glycine-modified amino acid.
Figure 1. 500 MHz 1H NMR spectra of a) compound 2 in [D6]DMSO at
RT; b) guanosine in H2O/D2O (90:10 v/v) at 108C; c) compound 3 in
H2O/CD3CN (50:50 v/v) at 108C, and d) compound 4 in H2O/D2O (90:10
v/v) at 108C.
condition. As anticipated, the imino proton is not observed
(Figure 1b) due to the rapid exchange with the solvent, thus
demonstrating the absence of structuring of guanosine units
Figure 2. Parts of the NOESY spectra showing the correlations between
the H8 and H1’ protons for a) 3 in H2O/CD3CN (50:50 v/v) at À108C
and b) 4 in H2O/D2O (90:10 v/v) at 108C.
5792
www.chemeurj.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 5791 – 5795

Formation of a Single Synthetic G-Tetrad
COMMUNICATION
other guanosine residues, are observed. This evidences that
the four guanosines are nonequivalent; two of them are or-
ganized in the plan of the peptide (i.e., are close to the pep-
tidic backbone that produces the above-mentioned NOE
correlation) and two of them are outside of the peptidic
backbone structure. This causes the diastereotopy of the
imino protons of the G-quartet and thus explains the split of
scaffold (data not shown). In particular, the appearance of
four signals corresponding to the H8 and H1’ correlations in
comparison with 3 is consistent with the co-existence of two
products (Figure 2b). This latter observation is also indica-
tive of a syn conformation along the glycosidic bond for
both products. To differentiate the two co-existing products
in solution, diffusion experiments (DOSY) were performed
(see the Supporting Information). This method has been
used recently for determining the molecular weight of unim-
olecular quadruplexes and the stoichiometry of nucleic-acid
quadruplex folds.[18]
1
the imino signal in the H NMR spectrum.
Molecular modeling by using the different NOE con-
straints (see the Supporting Information) was therefore per-
formed and the so-obtained structure was found quite rigid
(Figure 3a,b). It is also remarkable that the triazole linkers
Two main products were distinguished and their hydrody-
namic radii (Rh) were estimated (see the Supporting Infor-
mation).[19] The ratio Rh1/Rh2 was around 1.7, which indicates
the co-existence of two products in solution, one being ap-
proximately twice as big as the other one. Thus compound 4
was proposed to form a dimer in solution. The imino region
1
of the H NMR spectrum of compound 4 can be assigned as
follows: the two peaks at d=11.01 ppm correspond to the
imino protons of the monomer and the two peaks at d=
10.70 ppm to the imino protons of the dimer. The small dif-
ference should result from the shielding of the imino proton
by the two stacked G-tetrads in the dimer. Because a single
set of signals is observed for the dimer, the two G-quartets
are homotopic and the dimer displays a D4 symmetry.
Molecular modeling of dimer 4 was thus performed using
the observations described above. NOE correlations be-
tween H8 and H4’, and between NHPra and H4’ are no
longer observed in the NOESY spectrum of compound 4
(data not shown), suggesting a more flexible structure.
These constraints were thus removed from molecular-model-
ing study. A head-to-head dimer was proposed and the so-
obtained, minimized structure is depicted in Figure 3c.[20]
The G-quartet motifs are stabilized by the formation of the
dimer through the stacking of two tetrads. The distance be-
tween the two G-quartets is similar to stacked tetrads found
in G-quadruplexes (around 3.5 ꢂ). However, we were
unable to observe NOE correlations between the two moiet-
ies.
Figure 3. Molecular modeling using NOE constraints: a) Top view (the
peptidic scaffold is shown as a thinner structure to improve clarity),
b) side view of monomer 3, and c) side view of the proposed dimer 4, by
using NOE constraints from 3. The gray ribbon indicates the peptidic
backbone of the scaffold.
participate in the structuring of the compound; the aromatic
part of the guanosine is stacked on a tetrad formed by the
four triazole residues. Indeed, the distance between the G-
quartet and the triazole tetrad has been estimated to 3.6 ꢂ,
which is similar to the distance between two G-quartets in
DNA quadruplexes.[16]
To verify the structure and evaluate the stability of the
different architectures, circular-dichroism (CD) analyses
were performed. This technique has been widely employed
to study biological structures, notably DNA G-quadruplex-
es.[21] Spectra were recorded in the same conditions as for
the NMR studies. Note also that the additional peptidic
scaffold and triazole chromophores could contribute to the
observed CD bands.[22] The spectrum of compound 3 dis-
plays a minimum at l=230 nm, corresponding mainly to the
b-sheet conformation adopted by the cyclodecapeptide, and
a maximum at 281 nm with a shoulder at 265 nm (Figure 4a,
solid line), which are in agreement with a guanine tetrad.
Compound 4 shows a rather different spectrum. At 108C, in
pure water, the CD spectrum of 4 displays a minimum at
240 nm and two maxima at 260 and 288.5 nm, respectively
(Figure 4b, solid line). The global shape of the CD spectrum
is consistent with stacked tetrads.[21] Thermal denaturation
was then carried out to investigate the stability of both ar-
Cations, such as K+ or NH4+, are known to contribute to
the stability of G-quadruplex structures.[5] Thus, 1H NMR
studies were also carried out in presence of these cations.
No significant changes of the spectra could be observed
upon addition of NH4Cl. In particular, imino protons reso-
nate at the same chemical shift as in the absence of cations
(see the Supporting Information). Upon addition of KCl, a
slightly different chemical shift is observed for the imino
protons; this indicates an interaction of K+ cations with the
guanines tetrads. Such behavior can be also observed for G-
quartets stabilized in long G-wire tracks[17a] and for G-quad-
ruplex motifs prepared in 40% polyethyleneglycol-crowding
medium.[17b]
1
Compound 4 displays a more complex H NMR spectrum
in which most of the signals are split (Figure 1d). 2D experi-
ments suggest a mixture of two products, as evidenced by
the NOE correlations between NH and Ha of the peptidic
Chem. Eur. J. 2011, 17, 5791 – 5795
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5793

E. Defrancq et al.
Acknowledgements
This work was supported by the Centre National de la Recherche Scienti-
fique (CNRS) and the Ministꢃre de la Recherche. The authors are grate-
ful to the NanoBio program for access to the synthesis platforms. The
Cluster de Recherche Chimie de la Rꢀgion Rhꢄne-Alpes is acknowl-
edged for the financial support.
Keywords: cyclodecapeptide
·
G-quartets
·
hydrogen
bonds · scaffolds · supramolecular chemistry
[1] a) S. Sivakova, S. J. Rowan, Chem. Soc. Rev. 2005, 34, 9–21; b) J. L.
Sessler, C. M. Lawrence, J. Jayawickramarajah, Chem. Soc. Rev.
2007, 36, 314–325; c) U. Feldkamp, B. Saccꢅ, C. M. Niemeyer,
Angew. Chem. 2009, 121, 6110–6114; Angew. Chem. Int. Ed. 2009,
48, 5996–6000, and references therein.
Figure 4. CD spectra of a) compound 3 in H2O/CH3CN (50:50 v/v) and
b) compound 4 in H2O. The spectra were recorded at 10 (solid line) and
908C (dashed line).
[2] a) J. T. Davis, Angew. Chem. 2004, 116, 684–716; Angew. Chem. Int.
Ed. 2004, 43, 668–698; b) J. T. Davis, G. P. Spada, Chem. Soc. Rev.
2007, 36, 296–313; c) A. Likhitsup, R. J. Deeth, S. Ottob, A. Marsh,
Org. Biomol. Chem. 2009, 7, 2093–2103.
[3] a) S. Neidle, FEBS J. 2010, 277, 1118–1125; b) A. De Cian, L. La-
croix, C. Douarre, N. Temine-Smaali, C. Trentesaux, J.-F. Riou, J.-L.
Mergny, Biochimie 2008, 90, 131–155; c) L. Oganesian, J. Karlseder,
J. Cell Sci. 2009, 122, 4013–4025.
[4] R. Rinaldi, G. Maruccio, A. Biasco, V. Arima, R. Cingolani, T.
Giorgi, S. Masiero, G. P. Spada, G. Gottarelli, Nanotechnology 2002,
13, 398–403.
[5] a) G. Wu, A. Wong, Z. Gan, J. T. Davis, J. Am. Chem. Soc. 2003,
125, 7182–7183; b) X. Shi, K. M. Mullaugh, J. C. Fettinger, Y. Jiang,
S. A. Hofstadler, J. T. Davis, J. Am. Chem. Soc. 2003, 125, 10830–
10841; c) A. Ciesielski, S. Lena, S. Masiero, G. P. Spada, P. Samori,
Angew. Chem. 2010, 122, 2007–2010; Angew. Chem. Int. Ed. 2010,
49, 1963–1966.
[6] a) J. E. Betancourt, J. M. Rivera, J. Am. Chem. Soc. 2009, 131,
16666–16668; b) M. Garcıa-Arriaga, G. Hobley, J. M. Rivera, J. Am.
Chem. Soc. 2008, 130, 10492–10493.
[7] F. W. Kotch, V. Sidorov, Y. F. Lam, K. J. Kayser, H. Li, M. S. Kauch-
er, J. T. Davis, J. Am. Chem. Soc. 2003, 125, 15140–15150.
[8] a) M. Nikan, J. C. Sherman, Angew. Chem. 2008, 120, 4978–4980;
Angew. Chem. Int. Ed. 2008, 47, 4900–4902; b) M. Nikan, J. C. Sher-
man, J. Org. Chem. 2009, 74, 5211–5218.
[9] NMR studies should be carried out at a low temperature (À408C)
in CDCl3, in which the guanine-linked cavitand is able to form a G-
quartet. On the other hand, no self-assembly could be observed in
DMSO.
[10] M. Mutter, P. Dumy, P. Garrouste, C. Lehman, M. Mathieu, C. Peg-
gion, S. Peluso, A. Razaname, G. Tuchscherer, Angew. Chem. 1996,
108, 1588–1591; Angew. Chem. Int. Ed. Engl. 1996, 35, 1482–1485.
[11] a) P. Murat, D. Cressend, N. Spinelli, A. Van der Heyden, P. Labbꢀ,
P. Dumy, E. Defrancq, ChemBioChem 2008, 9, 2588–2591; b) P.
Murat, R. Bonnet, A. Van der Heyden, N. Spinelli, P. Labbꢀ, D.
Monchaud, M.-P. Teulade-Fichou, P. Dumy, E. Defrancq, Chem.
Eur. J. 2010, 16, 6106–6114.
chitectures. The CD spectra were found to be quite similar
for compound 3 at the different temperatures (Figure 4a de-
picts the CD spectra at 10 and 908C). A transition could be
observed during a melting experiment and the thermal de-
naturation of the so-formed tetrad was evaluated at 298C.
These results are consistent with the different NMR obser-
vations (see above). On the contrary, dramatic changes were
observed for compound 4. Indeed, the intensity of the signal
at 260 nm strongly decreases upon heating to 908C. At high
temperatures the spectrum becomes quite similar to the one
of compound 3.[23] This is again in agreement with the for-
mation of a dimer for compound 4, which on heating disso-
ciates to the monomer. Note that the head-to-head dimer
structure is proposed although CD data is more consistent
with a head-to-tail dimer.[22] However, in this case, the pres-
ence of the triazole moiety might contribute to the CD
signal.[22,24]
In summary, we have designed and synthesized two origi-
nal architectures based on a scaffold assistance to form G-
quartets in water without the addition of cations.[25] In addi-
tion to the entropically driven formation of the G-quartet,
the strategy takes advantage of the triazole linkers for fur-
ther stabilization of the G-quartet motif. We are currently
interested in studying the behavior of such architectures in
presence of monovalent and divalent cations. Moreover,
preliminary studies have shown that the use of compound 3
in the presence of free guanosine monophosphate (GMP)
allows the nucleation of G-quartet aggregates. The designed
architectures display a free amino moiety on the lower face
of the cyclodecapeptide scaffold; this moiety can be used for
the functionalization of a surface and the growth control of
quartet fibers on surfaces. These approaches may allow stud-
ies and understanding of complex phenomena, such as the
self-assembly of nucleobases in aqueous media.
[12] A system with a more flexible linker between the scaffold and the
guanine units has been prepared. Unfortunately, we were unable to
observe the formation of a G-quartet in water (see the Supporting
Information for the structure of this compound). This emphasizes
the unique role of the triazole linkers in the structure of the G-quar-
tet.
[13] M. G. Stout, M. J. Robins, R. K. Olsen, R. K. Robins, J. Med. Chem.
1969, 12, 658–662.
[14] D. J. Patel, S. A. Kozlowski, A. Nordheim, A. Rich, Proc. Natl.
Acad. Sci. USA 1982, 79, 1413–1417.
[15] J. L. Sessler, M. Sathiosatham, K. Doerr, V. V. Lynch, K. A.
Abboud, Angew. Chem. 2000, 112, 1356–1359; Angew. Chem. Int.
Ed. 2000, 39, 1300–1303.
5794
www.chemeurj.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 5791 – 5795

Formation of a Single Synthetic G-Tetrad
COMMUNICATION
[16] a) S. L. Forman, J. C. Fettinger, S. Pieraccini, G. Gottarelli, J. T.
Davis, J. Am. Chem. Soc. 2000, 122, 4060–4067; b) X. Liu, I. C. M.
Kwan, S. Wang, G. Wu, Org. Lett. 2006, 8, 3685–3688.
[17] a) N. Borovok, T. Molotsky, J. Ghabboun, D. Porath, A. Kotlyar,
Anal. Biochem. 2008, 374, 71–78; b) Z.-Y Kan, Y. Yao, P. Wang, X.-
H Li, Y. Hao, Z. Tan, Angew. Chem. 2006, 118, 1659–1662; Angew.
Chem. Int. Ed. 2006, 45, 1629–1632.
[18] a) A. Ambrus, D. Yang, Anal. Biochem. 2007, 367, 56–67; b) P.
Groves, M. Webba da Silva, Chem. Eur. J. 2010, 16, 6451–6453.
[19] D. K. Wilkins, S. B. Grimshaw, V. Receveur, C. M. Dobson, J. A.
Jones, L. J. Smith, Biochemistry 1999, 38, 16424–16431.
[20] Due to a mixture of two products (i.e., the monomer and the
dimer), the complete structural determination of compound 4 was
not possible. Some crucial NOE correlations, such as H8 and H1’
correlations, were observed. Thus, the dimer model was built by
using the same NOE constraints as for 3, except for the mentioned
correlations. Some minimization steps (using VA09A conjugate gra-
dient and steepest descent) were performed to place the two tetrads
in a head-to-head position allowing a D4 symmetry for the system.
[21] S. Masiero, R. Trotta, S. Pieraccini, S. De Tito, R. Perone, A. Ran-
dazzo, G. P. Spada, Org. Biomol. Chem. 2010, 8, 2683–2692.
[22] Further discussions on additional chromophores can be found in
Shermanꢆs works (see ref. [8]).
[23] The slight differences could be attributed to the ribose moiety. In
fact, in compound 3 the ribose pucker is constrained due to the iso-
propylidene on the 2’,3’-hydroxyl groups, whereas in compound 4,
the sugar ring is more flexible.
[24] The example of the head-to-head dimer shown by A. Randazzo and
co-workers (see ref. [22]) displays, in CD spectroscopy, two minima
at 240 and 290 nm, respectively, and a maximum at 260 nm. In our
case we observed a minimum at 240 nm and a maximum at 260 nm.
The absence of minimum at 290 nm can be attributed to the pres-
ence of the triazole moiety, which absorbs light in this wavelength
range.
[25] Note that during synthesis of the compounds, cations-containing
buffers were used. Thus, it is reasonable to think that one cation is
still present in the centre of the G-quartet, as previously reported.
Received: December 9, 2010
Published online: April 18, 2011
Chem. Eur. J. 2011, 17, 5791 – 5795
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5795