An aquatic host-guest complex between a supramolecular G-quadruplex and the anticancer drug doxorubicin

Article abstract of DOI:10.1039/c2ob25913c

We describe the synthesis of a fluorescent deoxyguanosine derivative that co-assembles (in water) with an unlabeled analogue into a heteromeric supramolecular G-quadruplex, which forms a host-guest complex with doxorubicin as evidenced by FRET experiments.

Full text of DOI:10.1039/c2ob25913c

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
Organic &
Biomolecular
Chemistry
Cite this: Org. Biomol. Chem., 2012, 10, 7562
www.rsc.org/obc
PAPER
An aquatic host–guest complex between a supramolecular G-quadruplex and
the anticancer drug doxorubicin†
José M. Rivera,* Mariana Martín-Hidalgo and Jean C. Rivera-Ríos
Received 12th May 2012, Accepted 1st August 2012
DOI: 10.1039/c2ob25913c
We describe the synthesis of a fluorescent deoxyguanosine derivative that co-assembles (in water) with an
unlabeled analogue into a heteromeric supramolecular G-quadruplex, which forms a host–guest complex
with doxorubicin as evidenced by FRET experiments.
Biological systems from cells to whole organisms remain an
open frontier for supramolecular chemistry.1 Developing effec-
tive strategies for interfacing the naturally occurring supramole-
cules with their synthetic counterparts poses both great
challenges and opportunities to improve human health. Two key
elements for moving forward are the invention of biocompatible
systems with the concomitant development of suitable character-
ization strategies. For example, while NMR still represents an
indispensable tool for the characterization of most supramolecu-
lar assemblies in solution, the current trend towards increasingly
complex supramolecules requires adopting techniques more
commonly used in biochemistry and molecular biology.
The Förster Resonance Energy Transfer (FRET) technique has
been used extensively in biophysical studies with a wide variety
of biological complexes (e.g., DNA–protein, protein–protein)
and for the development of biosensors.2 In particular, FRET has
been used to study the thermal stability3 of the folding–unfold-
ing of oligonucleotide based (e.g., DNA/RNA) G-quadruplexes
(OGQs)4 made from G-rich sequences in the presence of suitable
cations. It has also been used to study their function as host
molecules when binding to proteins,5 and even small molecule
ligands.6
In recent years we have developed 8-aryl-2′-deoxyguanosine
(8ArG) derivatives that self-assemble into precise supramolecu-
lar G-quadruplexes (SGQs) in both organic and aquatic media.9
SGQs are coaxially stacked planar G-tetramers whose formation
is promoted by cations such as potassium.10 Although NMR
spectroscopy continues to be essential for studying SGQs, the
need to study these systems at lower concentrations and/or in
complex environments (e.g., cells, organisms) moved us to
develop alternative characterization strategies. Here we report the
results of our first attempts towards this goal where a fluorescent
8ArG derivative (1) and the anticancer drug doxorubicin (DOX)
act, respectively, as a donor–acceptor pair in a FRET process
(Fig. 1).
Although the use of FRET remains somewhat limited in the
field of supramolecular chemistry, it has, nonetheless, gained
increasing prominence after the pioneering reports by the Rebek
group.7,8 They validated the versatility of the technique for
studying the kinetics and thermodynamics of self-assembled
heteromeric supramolecular capsules in the presence of suitable
guests. These studies relied on the use of labeled assembling
subunits (e.g., calixarenes, resorcinarenes) with appropriate
donor–acceptor pairs and/or by using suitable fluorescent guests.
Fig. 1 Top: Kekulé structures of the hydrophilic dG derivatives 1
(labeled with a coumarin and represented with a blue rectangle) and 2
(unlabeled and represented with a grey rectangle). Bottom: Cartoon
depictions of one of the putative heteromeric hexadecamers (hSGQ),
obtained after mixing 1 and 2 in the presence of KCl, and its resulting
host–guest complex with the anti-cancer drug doxorubicin
(hSGQ·DOX; the cartoon depicts only one of the possible configur-
ations for the complex).
Department of Chemistry, University of Puerto Rico Río Piedras
Campus, San Juan, PR 00931, USA. E-mail: jmrivortz@mac.com
†Electronic supplementary information (ESI) available: Experimental
details, NMR spectra and characterization for new compounds,
photophysical characterization and emission data. See DOI:
10.1039/c2ob25913c
7562 | Org. Biomol. Chem., 2012, 10, 7562–7565
This journal is © The Royal Society of Chemistry 2012

View Article Online
Our strategy relies on the use of the coumarin-labeled 8-(m-
acetylphenyl)-2′-dG derivative 1 as the donor component. We
have shown9b that the m-carbonylphenyl group is an essential
element for the high fidelity formation of hexadecameric SGQs.
Of the potential fluorophores available, we selected the coumarin
moiety because its small size was expected to minimize non-
specific interactions and poor aqueous solubility after its conju-
gation with the 8ArG fragment. Furthermore, in order to
maximize the hydrophilicity, we chose to co-assemble 1 with the
unlabeled analogue 2 to form heteromeric supramolecular
G-quadruplexes (hSGQ) (Fig. 1) with a putative statistical distri-
bution of both subunits. The solvent-exposed (outer) G-tetrads
provide a suitable platform to interact with planar aromatic mole-
cules, a fact that has been exploited in the development of
ligands for OGQs.4,6,11 Specifically, the Neidle group have
reported that the anthraquinone-containing drug daunomycin
interacts with OGQs via end-stacking interactions.11 We selected
the closely related anthraquinone DOX because of the latter pre-
cedent and also because its absorbance spectrum (acting as an
acceptor) complements very well the emission of the coumarin
moiety in 1 (donor), which would enable the desired FRET
studies (ESI, Fig. S12–S13†). Molecular modeling reveals that
the donor and acceptor groups could coexist at a distance below
the limit required (∼10 nm) to observe FRET even in a
configuration (one of the multiple possible) where the coumarin
moiety and the DOX are at the opposite ends of the hSGQ
(Fig. 2c and ESI, Fig. S24†).11b
acid afforded the alkynyl-containing 5, which was converted to 6
after deprotection and subsequent esterification with 6-bromo-
hexanoic acid. Coupling 6 and 3-azido-7-hydroxycoumarin,12
using the copper-catalyzed Huisgen 1,3-dipolar cycloaddition
(“click”) reaction, yields compound 7. Finally, displacement of
the bromine with NaN3 followed by a second click reaction furn-
ished the desired fluorescent derivative 1 (Scheme 1).
Self-assembly studies were performed by mixing dye 1
(1 equiv.) with 2 (15 equiv.)9b in a lithium phosphate buffer
1
(LiPB). As expected, the H NMR of 1 shows no detectable dis-
crete assembly since lithium only promotes the formation of
loosely bound aggregates (LBA) or thermodynamically unstable
SGQs (Fig. 2a) (500 MHz, 10% D2O in LiPB, pH 7.4,
298.2 K).13 Nevertheless, the simple addition of the anti-cancer
drug DOX (1 equiv.) seemed to be enough to induce the stabiliz-
ation of some LBA and even a small amount of hexadecamers
as evidenced by the appearance of signature peaks (e.g., double
sets of N1H peaks; Fig. 2b). Furthermore, those signature pairs
of peaks increase significantly (to a total of 65%; Fig. 2c) after
the addition of KCl (1 M), indicating the formation of robust and
discrete hexadecamers, as previously reported for 2 under
similar conditions.9b
Photophysical characterization of 1 and DOX confirmed that
the two compounds could serve as a donor–acceptor pair, due to
the excellent overlap between the emission of the former
(477 nm) and the absorbance of the latter (490 nm) (ESI,
Fig. S12–S13†). FRET titration measurements were performed
to study the changes in the emission intensity of the donor (1)
following the addition of the acceptor (DOX). Incremental
The synthesis of 1 begins with the selective protection of the
5′ hydroxyl group of 3 to generate the tert-butyldimethyl silyl
ether 4. Esterification of the 3′ hydroxyl group with 5-hexynoic
1
Fig. 2 Partial H NMR spectra (500 MHz, 10% D2O in LiPB, pH 7.4,
298.2 K) of a 5 mM sample with: (a) a mixture of 0.06 equiv. of 1 and
0.94 equiv. of 2; (b) same as (a) but with 0.06 equiv. of DOX, where the
N1H peaks (corresponding to the formation of a hexadecamer) begin to
show; (c) same as (b) but with 1 M KCl showing a clearly defined
double set of N1H signals. The right panel illustrates the equilibrium
between (top) monomeric subunits of 1 and 2 and loosely bound aggre-
gates (LBA); and (bottom) one of the potential complexes between
DOX and the various possible heteromeric supramolecular G-quadru-
plexes (hSGQ·DOX).
Scheme 1 Synthesis of coumarin labeled dG derivative 1.
Org. Biomol. Chem., 2012, 10, 7562–7565 | 7563
This journal is © The Royal Society of Chemistry 2012

View Article Online
Fig. 4 Melting profiles for hSGQ·DOX (LiPB pH 7.4) as determined
by (a) variable temperature FRET experiments, monitoring the donor
emission at increasing temperatures. The reported Tm is the inflection
point in the melting profile as determined by the minimum in the first
derivative plot (dashed red line). (b) Overlay of the DSC endotherm for
hSGQ (grey curve) and for hSGQ·DOX (red curve), showing an
enhancement in the thermal stability of the hexadecamer when interact-
ing with the drug.
Fig. 3 Overlaid excitation spectra (590 nm) of 11215 (5 mM mixture in
a LiPB pH 7.4) with increasing amounts of the acceptor DOX showing
the quenching of the former (donor) with the concomitant increase in
intensity for the latter (acceptor) as indicated by the trend arrows; (a)
sample without and (b) with 1 M KCl.
addition of DOX (0.0625–2 equiv.) to samples with KCl
(hSGQ·DOX = [11215·3K+]·DOX) and without potassium
(LBA·DOX) (5 mM, LiPB pH 7.4) showed a linear quenching
of the donor emission upon excitation of both samples at 390 nm
(ESI, Fig. S15–S16†). The excitation spectra (emission at
590 nm) demonstrate a stronger quenching for hSGQ·DOX
when compared to LBA·DOX (Fig. 3). A plot of the excitation
intensity of 1 (390 nm) as a function of the concentration of
DOX showed a more linear decrease (R2 = 0.98) and a steeper
slope for the sample with KCl than for the sample without it
(R2 = 0.90) (ESI, Graph S4†). A more linear quenching suggests
that the addition of DOX is the main cause for the quenching of
1 while a less linear quenching is suggestive of other factors
such as non-specific aggregation. The steeper slope in the pres-
ence of KCl hints at a more efficient energy transfer as would be
expected for a well-defined supramolecular complex like
hSGQ·DOX.
Variable temperature (VT) FRET experiments with
hSGQ·DOX resulted in a sigmoidal melting profile with a calcu-
lated melting temperature (Tm) of 72.5 °C (Fig. 4a).5 In contrast,
similar measurements in the absence of KCl lead to more
gradual emission increments of the donor with no significant
inflection points (ESI, Fig. S18†). This suggests that the mole-
cules of DOX promote the formation of random aggregates in
addition to the small amounts of the hexadecamer detected in
the NMR experiments (Fig. 2b), even in the absence of KCl.
Addition of the potassium salt, as also seen with NMR measure-
ments, displaces the equilibrium further towards the high fidelity
formation of hexadecamers (Fig. 2, right panel).
Fig. 5 Normalized fluorescence emission intensity (471 nm) of 8
(0.312 mM) in either LiPB (red circles) or LiPB with additional 1 M
KCl (purple diamonds) as a function of the equivalents of added DOX
(excited at 390 nm). The corresponding solutions of 8 mixed with 1 and
2 (5 mM) show significantly lower emissions in both LiPB (black
squares) and LiPB with additional 1 M KCl (brown triangles).
Control experiments were performed in order to evaluate the
quenching due to concentration effects. The coumarin derivative
8 (ESI, Fig. S19†)14 has UV/Vis and fluorescence spectra similar
to 1 but does not self-assemble because it lacks the guanine
moiety. Mixing it with DOX under identical conditions to those
7564 | Org. Biomol. Chem., 2012, 10, 7562–7565
This journal is © The Royal Society of Chemistry 2012

View Article Online
used for the aforementioned FRET experiments shows a slight
quenching of 8 (Fig. 5) with a slight further quenching after the
addition of KCl (ESI, Fig. S20†). In addition, variable tempera-
ture experiments for this mixture do not show significant
changes in donor emission over the relevant temperature range
(ESI, Fig. S21†). These observations reveal that a simple
mixture of the donor and acceptor groups is not enough for
efficient energy transfer, thus highlighting the essential role of
the guanine moiety.
Differential Scanning Calorimetry (DSC) experiments for
hSGQs reveal an endothermic peak corresponding to a Tm of
61 °C, while similar measurements for the complex hSGQ·DOX
indicate a significant increase in the melting temperature
(ΔTm = +9 °C; Fig. 4b).15 DSC measurements lead to a value for
the Tm having an excellent correlation with the one determined
by the VT FRET experiments. These results underscore that the
drug is interacting with and stabilizing the hSGQ in a manner
reminiscent to that reported for OGQs with small molecule
ligands.4,6,11
We have recently reported the discovery of an amphiphilic
hexadecameric SGQ that showed the phenomenon of thermally
induced self-assembly known as Lower Critical Solution Temp-
erature (LCST).16a At temperatures above the LCST this SGQ
formed microglobules that encapsulated DOX, which we demon-
strated by sedimentation experiments and fluorescence micro-
scopy.16b Up until now, we had little evidence to demonstrate the
direct interactions between DOX and the SGQs within the
microglobules. The results presented here fill that gap, and also
open the door to further studies with other biomedically impor-
tant small molecules or macromolecular guests (e.g., DNA/
RNA, proteins), in a manner similar to nucleic acid based supra-
molecular receptors.5 The results of such studies will be reported
in due course.
Notes and references
1 J. W. Steed and J. L. Atwood, Supramolecular Chemistry, John Wiley &
Sons, LTD, Chichester, 2000.
2 J. R. Lakowicz, Principles of Fluorescence Spectroscopy, Springer,
New York, 2006.
3 P. A. Rachwal and K. R. Fox, Methods, 2007, 43, 291–301.
4 S. Neidle and S. Balasubramanian, Quadruplex Nucleic Acids, The Royal
Society of Chemistry, Cambridge, 2006.
5 D. Margulies and A. D. Hamilton, Angew. Chem., Int. Ed., 2009, 48,
1771–1774.
6 A. Benz, V. Singh, T. U. Mayer and J. S. Hartig, ChemBioChem, 2011,
12, 1422–1426.
7 R. K. Castellano, S. L. Craig, C. Nuckolls and J. Rebek Jr., J. Am. Chem.
Soc., 2000, 122, 7876–7882.
8 E. S. Barrett, T. J. Dale and J. Rebek Jr., J. Am. Chem. Soc., 2007, 129,
3818–3819.
9 SGQs refer exclusively to homomeric assemblies as the ones described
in: (a) V. Gubala, J. E. Betancourt and J. M. Rivera, Org. Lett., 2004, 6,
4735–4738; (b) M. García-Arriaga, G. Hobley and J. M. Rivera, J. Am.
Chem. Soc., 2008, 130, 10492–10493; (c) J. E. Betancourt and
J. M. Rivera, Org. Lett., 2008, 10, 2287–2290; (d) M. d. C. Rivera-
Sánchez, I. Andujar-de-Sanctis, M. Garcia-Arriaga, V. Gubala, G. Hobley
and J. M. Rivera, J. Am. Chem. Soc., 2009, 131, 10403–10405;
(e) L. R. Rivera, J. E. Betancourt and J. M. Rivera, Langmuir, 2011, 24,
1409–1414.
10 (a) J. T. Davis and G. P. Spada, Chem. Soc. Rev., 2007, 36, 296–313;
(b) J. T. Davis, Angew. Chem., Int. Ed., 2004, 43, 668–698.
11 (a) S. Neidle, Therapeutic Applications of Quadruplex Nucleic Acids,
Elsevier, 2012, pp. 67–91; (b) G. R. Clark, P. D. Pytel, C. J. Squire and
S. Neidle, J. Am. Chem. Soc., 2003, 125, 4066–4067. The crystal struc-
ture of the complex daunomycin-d(TGGGGT) was used as the reference
to model DOX interacting with [11215·3K+]; see ESI, Fig. S24† for
further details.
12 K. Sivakumar, F. Xie, B. M. Cash, S. Long, H. N. Barnhill and Q. Wang,
Org. Lett., 2004, 24, 4603–4606.
13 Lithium phosphate buffer was prepared instead of the standard sodium
based PBS, because it is known that small cations such as lithium pre-
clude the formation of stable GQs. N. V. Hud and J. Plavec, in Quadru-
plex Nucleic Acids, ed. S. Neidle and S. Balasubramanian,
RSCPublishing, Cambridge, 2006, pp. 100–130.
14 V. Hong, S. I. Presolski, C. Ma and M. G. Finn, Angew. Chem., Int. Ed.,
2009, 48, 9879–9883.
This research was financially supported by the NIH-SCoRE
(Grant No. 5SC1GM093994-02). MMH thanks the NSF-IF-
N-EPSCoR (01A-0701525) and the Alfred P. Sloan Foundation
for graduate fellowships. JCR thanks the NIH-MARC (5 T34
GM007821) program for an undergraduate fellowship.
15 ΔTm = Tm ([11215·3K+]·DOX) − Tm (11215·3K+).
16 (a) J. E. Betancourt and J. M. Rivera, J. Am. Chem. Soc., 2009, 131,
16666–16668; (b) J. E. Betancourt, C. Subramani, J. L. Serrano-Velez,
E. Rosa-Molinar, V. M. Rotello and J. M. Rivera, Chem. Commun., 2010,
46, 8537–8539.
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 7562–7565 | 7565