Loop-mediated fluorescent probes for selective discrimination of parallel and antiparallel G-Quadruplexes

Article abstract of DOI:10.1016/j.bmc.2021.116077

Herein we report simple pyridinium (1–3) and quinolinium (4) salts for the selective recognition of G-quadruplexes (G4s). Among them, the probe 1, interestingly, selectively discriminated parallel (c-KIT-1, c-KIT-2, c-MYC) G4s from anti-parallel/hybrid (22AG, HRAS-1, BOM-17, TBA) G4s at pH 7.2, through a switch on response in the far-red window. Significant changes in the absorption (broad 575 nm → sharp 505 nm) and emission of probe 1 at 620 nm, attributed to selective interaction with parallel G4s, resulted in complete disaggregation-induced monomer emission. Symmetrical push/pull molecular confinements across the styryl units in probe 1 enhanced the intramolecular charge transfer (ICT) by restricting the free rotation of C[dbnd]C units in the presence of sterically less hindered and highly accessible G4 surface/bottom tetrads in the parallel G4s, which is relatively lower extent in antiparallel/hybrid G4s. We confirm that the disaggregation of probe 1 was very effective in the presence of parallel G4–forming ODNs, due to the presence of highly available free surface area, resulting in additional π-stacking interactions. The selective sensing capabilities of probe 1 were analyzed using UV–Vis spectroscopy, fluorescence spectroscopy, molecular dynamics (MD)–based simulation studies, and ¹H NMR spectroscopy. This study should afford insights for the future design of selective compounds targeting parallel G4s.

Full text of DOI:10.1016/j.bmc.2021.116077

Bioorg. Med. Chem. 35 (2021) 116077
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Loop-mediated fluorescent probes for selective discrimination of parallel
and antiparallel G-Quadruplexes
Anup Panditha, Upendra Nagarajacharib, Ravi Kumara Guralamatta Siddappaa, Sungjin Leec,
,
*
Chin–Ju Parkc, Krishnaveni Sannathammegowdab, Young Jun Seoa
a Department of Chemistry, Jeonbuk National University, Jeonju 54896, Republic of Korea
b Department of Physics, University of Mysore, Karnataka 570006, India
c Department of Chemistry, Gwangju Institute of Science and Technology, Gwangju 61005, Republic of Korea
A R T I C L E I N F O
A B S T R A C T
Keywords:
Herein we report simple pyridinium (1–3) and quinolinium (4) salts for the selective recognition of G-quad-
ruplexes (G4s). Among them, the probe 1, interestingly, selectively discriminated parallel (c-KIT-1, c-KIT-2, c-
MYC) G4s from anti-parallel/hybrid (22AG, HRAS-1, BOM-17, TBA) G4s at pH 7.2, through a switch on response
in the far-red window. Significant changes in the absorption (broad 575 nm → sharp 505 nm) and emission of
probe 1 at 620 nm, attributed to selective interaction with parallel G4s, resulted in complete disaggregation-
G-Quadruplexes
Intramolecular charge transfer
Molecular confinements
Disaggregation
Molecular dynamics
induced monomer emission. Symmetrical push/pull molecular confinements across the styryl units in probe 1
–
enhanced the intramolecular charge transfer (ICT) by restricting the free rotation of C C units in the presence of
–
sterically less hindered and highly accessible G4 surface/bottom tetrads in the parallel G4s, which is relatively
lower extent in antiparallel/hybrid G4s. We confirm that the disaggregation of probe 1 was very effective in the
presence of parallel G4–forming ODNs, due to the presence of highly available free surface area, resulting in
additional π-stacking interactions. The selective sensing capabilities of probe 1 were analyzed using UV–Vis
spectroscopy, fluorescence spectroscopy, molecular dynamics (MD)–based simulation studies, and 1H NMR
spectroscopy. This study should afford insights for the future design of selective compounds targeting parallel
G4s.
1. Introduction
such probes have displayed topology-specific variations in their optical
properties under physiological conditions.8 More commonly, the sensing
G-quadruplexes (G4s) can be classified as parallel, antiparallel, or
hybrid types depending on the topologies arising from their dynamic
conformations.1 Among them, parallel G4s are generally present in
several oncogene promoter regions, including c-MYC, VEGF, and KRAS,
forming intramolecular parallel G4s.2,3 Considering their biological
significance (e.g., in cell proliferation; their transcription factor regula-
tory activities; and their natural existence in genomes), intramolecular
G4s have been investigated more extensively under specific condi-
tions4,5 than have been corresponding intermolecular G4s. Various
fluorescent probes have been designed for the selective recognition of
G4s over duplex [double-stranded (ds)] and single-stranded (ss) DNAs,
with some of them functioning under in vitro conditions.6
strategies have relied on conjugating G4-stabilizing ligands to conven-
tional fluorophores9,10 or attaching quencher-free probes (e.g., squary-
lium,11,12 thiazolium,13,14 and pyridinium15,16 units), accompanied by
electron donating or withdrawing groups in both symmetrical and un-
symmetrical arrangements, to induce effective push/pull effects.
Considering the diverse topologies of G4 structures, various coumarin/
anthracene, naphthalene diimide, and squaraine-based fluorescent ma-
terials have been developed recently to recognize parallel G4s in
vitro.17,19
Although many fluorescent probes have been reported for the
recognition of G4s, the structural similarity between parallel and anti-
parallel/hybrid topologies has made the development of probes target-
ing only parallel G4s still challenging, especially in the far-red window.
Based on the concept of conventional aggregation-induced quench-
ing, and considering the variety of homo-supramolecular assemblies
Nevertheless, only a few papers have reported topology-oriented
selectivities (discrimination between parallel and antiparallel/hybrid
topologies) using in vitro models.7 Interestingly, a limited number of
* Corresponding author.
E-mail address: yseo@jbnu.ac.kr (Y.J. Seo).
https://doi.org/10.1016/j.bmc.2021.116077
Received 1 January 2021; Received in revised form 4 February 2021; Accepted 9 February 2021
Available online 16 February 2021
0968-0896/© 2021 Elsevier Ltd. All rights reserved.

A. Pandith et al.
Bioorganic & Medicinal Chemistry 35 (2021) 116077
that are typically formed under physiological conditions through
cooperative binding between monomer units, we designed probe 1
featuring dual functionalities (phenolic OH and NEt2 units as electron
donors; a pyridinium core as an electron acceptor) to undergo a shift
from an intramolecular mode of charge transfer to an intermolecular
charge transfer process upon aggregation. Upon consideration of the
conformations of the sugar and nucleobase units in each part of the G4
tetrad, especially those in the loop portion, which generally discriminate
the parallel and antiparallel G4 topology, we designed our molecular
materials (Scheme 1 probes 1–4) to have bent(crescent)/linear shapes,
moderate flexibility, and various degrees of hydrophobicity. We vali-
dated the sensing capabilities of probe 1 through studies using UV–Vis
spectroscopy, fluorescence spectroscopy, 1H NMR spectroscopy, and
molecular dynamics (MD)–based simulations.
between the electron-deficient [Py]+ units of probe 1 and the electron-
rich phenoxide units of probe 1 in the aggregates (ions pairs) led to a
strong intermolecular electron coupling–aided charge transfer
phenomenon.
Dynamic light scattering (DLS) revealed the formation of molecular
aggregates having an average size of 6362.6 nm in H2O, consistent with
our hypothesis (Fig. S4). Excitation-dependent studies revealed shifts in
the aggregation-induced exciplex emission wavelength, supporting the
presence of nanometer-to-micrometer-sized particles of 1 in H2O
(Fig. S5).
The ground state optimized geometry of 1, calculated using DFT,
revealed that the highest occupied molecular orbitals (HOMOs) and
lowest unoccupied molecular orbitals (LUMOs) were distributed in the
diethylaminosalicylaldehyde and pyridinium units, respectively, sup-
porting a push/pull-based ICT process. The energy-minimized structure
of probe 1 in the ground state geometry had a crescent shape (bent
2. Results and discussion
conformation). The obtained band gap energy of –2.42 eV (ΔE = EHOMO
–
Probes 1 and 2 were synthesized from lutidine and 2-methylpyridine
scaffolds, respectively, by forming their respective 2,6-dimethylpyridi-
nium and 2-methylpyridinium salts and then performing simple Knoe-
venagel condensations with N,N-diethylaminosalicylaldehyde in the
presence of a catalytic amount piperidine. Probe 3 was synthesized
through the condensation of 4-picolylamine and 1,8-naphthalic anhy-
dride in EtOH and subsequent reaction with 9,10-dicholoromethylan-
thracene in MeCN. Similarly, probe 4 was prepared from the reaction
of quinoline with 9,10-dicholoromethylanthracene. The synthesized
probes were characterized using NMR spectroscopy and mass spec-
trometry (see the ESI).
ELUMO) was concordant with the UV–Vis electronic transitions of 1 in
H2O (ca. 510 nm). Similarly, the excited state geometry band gap energy
(ΔE = ꢀ 2.01 eV) was in good agreement with the emission maxima at
620 nm (Fig. S6).
Having investigated the photophysical properties of our probe 1 in
organic and aqueous conditions and considering the significance of
push/pull systems having a bent molecular architecture21, we used op-
tical methods to examine the interactions of 1 with ODNs having various
canonical and non-canonical structure–forming sequences under phys-
iological conditions (Table 1).
UV–Vis spectra of the probe 1 in the presence of 2.0 eq. of parallel
G4-forming ODN sequences [e.g., c-MYC (Pu27/22 nt), c-KIT-1, c-KIT-2]
featured a sharp monomer band near 510 nm with complete disap-
pearance of the aggregation (broad) band, suggesting efficient disas-
sembly of 1 and the formation of a highly stable complex with a specific
molecular confinement (Fig. 1a, b and Fig. S7). Notably, the unfolded
versions of the G4-forming ODNs c-MYC (pu27) in H2O and 22AG in
H2O did not induce any significant degrees of monomer band formation,
implying that the probe 1 could recognize parallel G4s only in their
folded forms, rather than in their linear/unfolded forms. In contrast, the
presence of hybrid and antiparallel-forming G4s [viz., 22AG (K+/Na+),
TBA, HRAS-1, BOM17] led to minor decreases in aggregation (disas-
sembly of colloids), but it did not induce the appearance of the sharp
monomer band in the UV–Vis spectra (Fig. 1b). Similarly, the addition of
2.0 eq. of various ODNs, ssDNA, dsDNA, and hairpin and TWJ ODNs did
not result in any significant changes in the broad aggregation band
(from > 630 to 750 nm) along with monomer region (Fig. 1b). We
suspect that the stronger interactions of the probe 1 with the parallel G4s
resulted in sharp increases in the intensity of the monomer bands near
507 nm in the UV–Vis absorption spectra. The relative absorbance ratios
Photophysical studies of probe 1 in various solvents revealed that it
displayed environment-dependent absorption band. Monomer band
were exhibited near 510 nm in organic solvent and this band were
decreased in the H2O or H2O with KCl (aggregation state) (Fig S1a).
Fluorescence spectra exhibited monomer emission near 615 nm in
organic solvent and this emission were decreased and new band were
increased at 730 nm in H2O or H2O with KCl; we assumed that this is
related with aggregation-induced exciplex emission (Fig S1b,
Table ES1).
We examined the aggregation properties of probe 1 in H2O because
we will use this for probe the G-quadruplex structure in buffer condition.
Upon addition of various volume-percentages of H2O to a 1 µM solution
of probe 1 in MeCN, we observed concordant decreases in the molar
absorptivity with significant broadening in the range 500–760 nm in the
UV–Vis spectra (Fig. S2). Such broadening, due to the formation of
colloids, is typical of Mie scattering in solution.20 In addition, we also
measured fluorescence spectra of probe 1 (1 μM) in H2O with various
volume-percentages of MeCN. We, firstly, observed quenching emission
near 615 nm (±7.5 nm) and exciplex emission at 730 nm in water (ag-
gregation state) and it exhibited increased monomer emission near 615
nm (±7.5 nm) and decreased exciplex emission near 730 nm with
increased ratio of MeCN (Fig. S3). We assume that the quenching in
monomer emission in aggregation state may be originated from stacking
A
507 nm/A630 nm for the parallel G4s were sufficiently high to distinguish
them from the rest of the nucleic acids under physiological conditions.
To investigate the selectivity of these interactions, we recorded
fluorescence spectra of the solutions in buffered H2O supplemented with
100 mM KCl. Interestingly, among the various G4-forming ODNs, only
the parallel G4-forming ODNs [i.e., c-MYC Pu27/c-MYC (22nt), c-KIT-1,
c-KIT-2] provided higher selectivities and sensitivities upon interactions
with the probe 1 (Fig. 2 and Fig. S8a) through “switch on” (>12-fold)
responses in the far-red region (ca. 620 nm). For the antiparallel G4-
forming ODNs and unfolded G4-forming ODNs [c-MYC (pu27) in H2O,
22AG in H2O], the fluorescence enhancement was weaker (<2.0-fold)
for the emission centered at 620 nm, concordant with the UV–Vis
spectral behavior (Fig. S8b). In contrast, DLS studies of probe 1 with
parallel ODNs showed aggregates size in up to 600 nm which was ~ 10
times smaller than the probe 1 aggregations itself in similar conditions
(Fig. S4). On the other hand antiparallel, hybrid G4s and dsDNA se-
quences showed relatively larger particle sizes, compared to that of
probe 1 aggregations.
Scheme 1. Structures of our pyridinium (1–3)- and quinolinium (4)-
Obtained results clearly suggested that, Probe-1 disaggregation was
based probes.
2

A. Pandith et al.
Bioorganic & Medicinal Chemistry 35 (2021) 116077
Table 1
ODNs used in this study.
ODN
Sequence 5’→ 3’
Topology/structure
Molar extinction coefficient (L
molꢀ 1 cmꢀ 1
279,900
)
c-MYC (Pu
27)
TGGGGAGGGTGGGGAGGGTGGGGAAGG
Intramolecular parallel
c-MYC (22)
c-KIT-1
c-KIT-2
TBA
TGAGGGTGGGTAGGGTGGGTAA
AGGGAGGGCGCTGGGAGGAGGG
CCCGGGCGGGCGCGAGGGAGGGGAGG
GGTTGGTGTGGTTGG
Intramolecular parallel
Parallel
228,700
226,700
253,400
143,300
Parallel
Intramolecular anti-
parallel
22AG (K+/
AGGGTTAGGGTTAGGGTTAGGG
Intramolecular anti-
parallel/hybrid
Intramolecular parallel
Intramolecular anti-
parallel
228,500
Na+)
ODN-2/G30
HRAS-1
GGGGGGGGGGGGGGGGGGGGGGGGGGGGGG
TCGGGTTGCGGGCGCAGGGCACGGGCG
304,400
250,400
BOM17
Hairpin
G-Triplex
TWJ
GGTTAGGTTAGGTTAGG
Anti-parallel
174,600
249,700
ACGTGCCACGATTCAACGUGGCACAG
Not specified
AGGGTTAGGGT and TAGGGT
Not specified
114,800 & 61,700
CGC AAG CGA CAG GAA CCT CGA GGA ATT CAA CCA CCG GAC G GCA GGC TAG GAC GGA TCC
CTC GAG GTT CCT GTC GCT TGC G
Not specified
757,800
ssDNA
dsDNA
A30
CCAGTTCGTAGTAACCC
Single-stranded
Double-stranded
Not specified
Not specified
Not specified
160,900
GGGTTACTACGAACTGG & CCAGTTCGTAGTAACCC
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
TTTTTTTTTTTTTTTTTTTTTTTTTTTTTT
CCCCCCCCCCCCCCCCCCCCCCCCCCCCCC
167,400 & 160,900
363,400
T30
243,600
C30
216,200
significant in the case of parallel G4s forming ODNs thereby causing
strong intramolecular charge transfer as well as restricted C=C rotations
resulted significant changes emission intensity. It is worth to mention
that, intermolecularly induced (tightly packed aggregates) exciplex peak
was disappeared with most of the ODNs, unambiguously supported the
1•ODNs complex was predominant species, arose from coulombic force
of interactions. Such interactions greatly perturbed the tightly packed
molecular orientations of aggregates which resulted in disappearance of
exciplex peaks in emission spectra (Fig. S9).
order of 104 Mꢀ 1 (Table ES2). These results support the notion of the
preferential binding capability of 1 toward parallel G4s over other ca-
nonical and non-canonical forms of ODNs, even in the presence of an
excess concentration.
Next, we recorded 1H NMR spectra of the c-MYC (22 nt) ODN and the
probe 1 to examine their interactions. Because the wild-type c-MYC (Pu
27) sequence forms multiple heterogeneous structures, we recorded 1H
NMR spectra using the c-MYC (22 nt) ODN, which uniformly forms a
propeller-type parallel structure. Considering the similar binding modes
of probe 1 with c-MYC (Pu27) and c-MYC (22 nt), as supported by
UV–Vis and fluorescence spectral studies in H2O at pH 7.2 supplemented
with 100 mM KCl, we recorded the 1H NMR spectra. Upon increasing the
concentration of the probe 1 from 0 to 6 eq., we observed peak broad-
ening when the probe content ranged from 0.5 to 2.0 eq. and, thereafter,
the signal intensity (spinning) diminished slightly upon reaching 6.0 eq.
We assume that the signal broadening originated from a bound–un-
bound state of bonding between the probe 1 and the c-MYC (22 nt) ODN.
The observed signals decreased sharply until the ratio of 1 to c-MYC (22
nt) reached 2:1, validating the stoichiometry obtained from the Job’s
plot (Fig. 5b). Intensity decay profiles revealed prominent changes in
surface and bottom G tetrads proton upon addition of various concen-
We also examined the effect of various biologically relevant cations
and anions and miscellaneous molecules with probe 1. Fluorescence
studies with various biologically relevant cations and anions [i.e., Fe3+
,
Mn2+, Zn2+, Cu2+, PO43–, SO42–, HSO4–, HP2O37– (PPi), Cl–, OAc–, OBz–] and
miscellaneous molecules [glycine (Gly), phenylalanine (PhA), lysine
(Lys), serine (Srn), proline (Prl), glutamic acid (Glu)] revealed no effects
on the fluorescence behavior of Probe 1 in TRIS-HCl–buffered H2O at pH
7.2 in the presence of 100 mM KCl (Fig. S8b).
We performed UV–Vis spectroscopic titration studies to evaluate the
cooperative binding capabilities of the parallel G4 ODNs with the probe
1 (Fig. 3a). Upon addition of aliquots of c-MYC (Pu27, parallel G-
quadruplex) from 0 to 5 µM, Mie scattering effects were diminished with
concomitant increases in the intensity of the monomer absorption band,
supporting disaggregation effects; these effects were invariant after
trations of probe 1 (0–800 μM), revealed stacking was preferred over
groove binding (Fig. 5b). In contrast the upon addition of probe 1 imino
signals from the guanine residues, especially G7, G9, and G13, were
perturbed significantly (shifted up-field) suggesting that there were
stacking modes on either side of the loop, but we did not observe any
significant changes in chemical shift for the signals of G8, G12, G17, and
G21 in the groove (Fig. 5a).
approximately 4.0 μM had been added (Fig. 3b). In contrast, fluores-
cence titration studies supported a good binding constant (Ka) of 6.75 ×
106 Mꢀ 1 (±1.2%) calculated using a non-linear independent binding site
model based on fitting Eq.(1) using the Levenberg–Marquardt fitting
routine (Origin Professional 2020). Similarly, the association constants
for c-KIT-1 and c-KIT-2 were 5.1 × 106 and 4.4 × 106 Mꢀ 1, respectively
(Figs. S10 & 11). We also examined the interaction between probe 1 and
parallel G4 ODNs using Job’s plots method. Job’s plots of the probe 1
with the parallel G4–forming ODNs c-MYC and c-KIT-1/c-KIT-2 sup-
ported their 1:2 and 1:1 binding stoichiometries, respectively (Fig. 4).
In contrast, fluorescence titration studies of probe 1 with antiparallel
and hybrid G4s revealed that relatively higher concentrations of the
ODNs were required to attain saturation (maximum intensity near 620
nm) (Fig. S12). The higher complexity in biomolecular interactions and
minor differences in binding process (micro-environments) generally
results in various types of non-radiative relaxation process. Due to these
reasons binding models were fitted in independent site models which
generally does not co-relate the emission intensity and binding constants
during calculations22. The observed association constants were in the
To evaluate the vital interactions between the probe 1 and the
various G4 topologies, we performed molecular simulations using par-
allel, hybrid, and anti-parallel G4s. Considering the preferential binding
toward parallel G4s, we selected c-MYC (22 nt) PDBID 2L7V for anal-
ysis. To validate the behavior of hybrid and anti-parallel G4s, we
selected PDBID 2MB3 and 22AG PDBID 143D for simulation studies. We
used molecular docking to identify the possible binding modes (poses)
and interaction sites. In addition to the obtained binding poses, we
validated the stabilities through MD-based simulations by calculating
the binding free energies based on MM/PBSA protocols.
Accordingly, molecular docking studies and trajectory based visu-
alization of the G4:probe 1 complexes in both 1:1 (each at 5’ and 3’
ends) and 1:2 (5’ and 3’ ends) stoichiometric relations revealed the two
preferential binding sites for the parallel G4–forming c-MYC (22 nt)
3

A. Pandith et al.
Bioorganic & Medicinal Chemistry 35 (2021) 116077
Figure 2. Histogram of relative fluorescence enhancement of probe 1 (1 μM) at
λem 620 nm, in the presence of various canonical and non-canonical ODNs in
H2O (TRIS-HCl, pH 7.2) in the presence of 100 mM KCl/NaCl upon excitation at
λex = 490 nm. Where I and I0 are the intensity of the probe 1 at λem 620 nm, in
the presence and absence of analytes. Error<±2.17%
Figure 1. (a) UV–Vis spectra of probe 1 (2 μM) in the absence and presence of
various G4s and non-G4s forming ODNs in TRIS-HCl–buffered H2O at pH 7.2
supplemented with 100 mM KCl. (b) Probe 1 (2 μM) absorbance ratios (A507 nm/
A630 nm) measured in the absence and presence of various ODNs or miscella-
neous biomolecules. Note: The ODN 22AG was supplemented with 100 mM
NaCl to form hybrid G4s. Wherever “H2O” is specified, only the buffered me-
dium was maintained to examine the interactions of 1 with the unfolded G4-
forming ODNs. Errors: <±1.25%.
Figure 3. (a) UV–Vis spectra of probe 1 (2.0
MYC (Pu27). (b) Absorbance of probe 1 plotted with respect to the concen-
tration of c-MYC (Pu27). (c, d) Fluorescence spectral titration of probe 1 (1 M)
with the c-MYC (Pu27) ODN (0–5 M) in TRIS-HCl–buffered H2O at pH 7.2
supplemented with 100 mM KCl. Note: UV–Vis and fluorescence spectral ti-
trations were performed from 0 to 5 M to minimize errors in fitting the
titration data of c-MYC (Pu27) (irregular pattern observed from 0 to 0.3 M);
data taken from 0.3 to 5 M (saturation point) are presented in the figure.
μM) in the presence of 0–5 μM c-
μ
μ
(PDBID 2L7V) without changing the quartet structure of the G4 (Fig. 6
and Fig. S13). This finding implies that probe 1 interacted through the
end stacking mode with c-MYC (22 nt). The substituents across the styryl
units in probe 1 remained in a trans mode; flipping was not observed.
Both the 5’ end and 3’ end docking studies revealed the end stacking
mode of the entire surface of probe 1. The calculated binding free energy
component for last 50 ns of the G4:probe 1 (1:2) complex was
–2048.735 kJ molꢀ 1 (Table ES3); this value is two times higher than that
of single ligand interactions, supporting the notion that two-ligand in-
teractions were thermodynamically more feasible.23 Probe-1 binding
process and its stabilities were monitored RMSD (root mean square
deviations) values for the 1:2 G4:probe 1 complex converges through
equilibration (i.e high stability) also individual nucleotide contributions
revealed that the surface and bottom quartet nucleotides and the loop
nucleotides interacted more effectively without harnessing the spurious
secondary structure (Fig. S14).
μ
μ
μ
binding mode (binding free energy: –1180.107 kJ molꢀ 1) was less
feasible than the 5’ end stacking (binding energy: –1231.429 kJ molꢀ 1).
The contributions of the individual nucleotides in pose 1 (groove bind-
ing mode) and pose 2 (5’ end stacking mode) also revealed that the
groove and loop nucleotides were involved in binding process (Figs. S15
& S16).
The antiparallel G4–forming 22AG (PDBID 143D) also had one-
ligand interactions with four different modes of binding (Figs. 7c–f).
Among them, pose 3 was relatively unstable (Fig. 7e); it was omitted in
from the binding free energy calculations. According to those calcula-
tions, pose 1 (groove binding), pose 2 (3’ end stacking), and pose 4 (5’
end stacking) had free binding energies of –939.008, –926.698, and
–874.11 kJ molꢀ 1, respectively. These binding energies are all lower
than those of the parallel and hybrid G4 topologies (Table ES3). The
In contrast, the hybrid G4 (PDBID 2 MB3) underwent groove binding
as well as 5’ end stacking (Figs. 7a and 7b), but only one mode of binding
was stable (either groove binding or 5’ end stacking). The groove
4

A. Pandith et al.
Bioorganic & Medicinal Chemistry 35 (2021) 116077
Figure 7. (a, b) Docking of probe 1 to 2 MB3, revealing two bindings poses: (a)
2 MB3-Pose 1: Groove binding; (b) 2 MB3-Pose 2: stacked at 5’ end. (c–f)
Docking of probe 1 to 143D, revealing four binding poses: (c) Pose 1: groove
binding; (d) Pose 2: 3’ end stacking; (e) Pose 3: groove binding; and (f) Pose 4:
5’ end stacking.
Figure 4. Job’s plots of probe 1 (1 μM) and c-MYC (Pu27, 1 µM), c-KIT-1 (1
μ
M), and c-KIT-2 (1 μM) at pH 7.2 (TRIS-HCl) in the presence of 100 mM KCl.
Note: λex = 490 nm; λem = 620 nm. Errors: <±3.55%.
individual nucleotide contributions in poses 1, 2, and 4 of the probe 1:
antiparallel G4 complex revealed that these interactions were hindered
because of the sterically crowded environment arising from the diagonal
and lateral loops (Figs. S17, S18 & S19).
Based on the optical, NMR spectroscopic, and MD simulation studies,
we speculate that the presence of lateral loops on the periphery led to
relatively sterically less crowded surfaces of the G4 tetrads on either side
(top and bottom) of the parallel G4s, resulting in highly stable (probe
1)2•G4 complexes.24,25 This arrangement would lead to the appearance
of a perfect monomer emission centered near 620 nm. In antiparallel and
hybrid G4s, however, the loops were oriented toward the upper and
lower sides of the tetrad, providing additional steric bulk that resulted in
weaker π-stacking and, therefore, less stable (probe-1)2•G4 complex-
es.26–28 The presence of crescent-shaped and flat molecular architectures
and the formation of highly stable ions pairs of 1 in the H2O medium,
additionally supported by peripheral NEt2 substitutions, tended to
enhance the selectivity by enhancing conventional noncovalent in-
teractions (i.e.,
π -stacking, multiple hydrogen bonding, and coulombic
interactions).
To validate our hypothesis of molecular geometry–oriented
probe•G4 interactions, we examined the behavior of the single-armed
probe 2 featuring only one styryl unit. The UV–Vis spectra of the
probe 2 featured its value of λmax at 468 nm, without any broadening in
the presence of KCl, even at 100 mM, unlike the broadening observed for
probe 1 (Fig. S20a) with very low emission behavior in H2O (Fig. S20b).
This finding suggests that greater hydrophilicity tends to enhance the
solubility, thereby forming smaller colloids. We examined the interac-
tion of probe 2 with different ODN and different types of G4 sequences.
We didn’t observed significant changes in absorption spectra and also
we could not discriminate the parallel G-quadruplex sequences from
other sequences using fluorescence spectra (Fig. S21a, Fig S21b,
Fig. S22). The resulting association constants for probe 2 toward the
G4s, obtained through fluorescence spectral titration, were on the order
of 104 to 105 Mꢀ 1 (Fig. S23). We speculate that the smaller size of probe
2 meant that it could bind parallel, antiparallel, and hybrid G4s without
inducing steric hindrance on either side (top and bottom) of the G4s.29
To evaluate the significance of the styryl and pyridinium units, we
examined probe 3, which features a more flexible molecular geometry
and greater hydrophobicity, for its opto-analytical capabilities under
Figure 5. (Top) 1H NMR spectra (Bruker Avance II 700 MHz, equipped with a
cryogenic probe; 298 K) of 0.3 mM c-MYC (22nt) recorded in the presence of
various concentrations of probe 1 in D2O at pH 7.2 (20 mM TRIS-HCl) sup-
plemented with 100 mM KCl (1% DMSO). (a) Values of Δδ for various signals
from the c-MYC (22 nt) ODN; the dashed line indicates an average + 1 standard
deviation; Δδ = δH at [probe 1/G4] = 1 –δH of free G4s; G8, G17, G13, and G22
have been excluded from analysis because of signal overlap. (b) Normalized
intensity upon increasing the concentration of probe 1 (0–800 μM).
Figure 6. (a, b) Docking studies of probe 1 to the parallel G4–forming c-MYC
22 nt (PDBID 2L7V) revealing two bindings modes (poses): (a) 2L7V-Pose 1:
stacked at 5’ end (probe 1 is depicted in a sphere (space-filling) representation;
DNA with ribbon backbones and filled base and sugar rings), (b) 2L7V-Pose 2:
stacked at 3’ end (Pose 2). (c) Probe 1:DNA modeled in a 2:1 stoichiometry
(2L7V-Pose3) for MD simulations.
similar conditions. Considering the significance of π-stacking on a G
tetrad (top and bottom sides of the G4s), we incorporated a hydrophobic
anthracene moiety at the 4-position of each pyridinium unit to enhance
the probability of stacking. In addition, to ensure greater flexibility, we
positioned alkyl groups on either side of the anthracene moiety in a
symmetrical manner (at the 9- and 10-positions). With such a design, we
5

A. Pandith et al.
Bioorganic & Medicinal Chemistry 35 (2021) 116077
expected two different types of
π
-stacking could occur upon interacting
bottom stacking mode very effectively. Thus, we believe that the
disaggregation of 1 was very effective in the presence of parallel G4-
forming ODNs, due to the presence of highly available free surface
area, resulting in more highly negative values for the change in free
energy (ΔG) of formation upon complexation (Fig. 8).
with G4s. Probe 3 did not, however, exhibit any selective discrimination
of dsDNA over G4s at pH 7.2 (TRIS-HCl) supplemented with 100 mM KCl
in either UV–Vis or fluorescence spectral studies (Fig. S24). The presence
of the pyridinium core and the freely rotatable CH2 groups at the 9- and
10-positions of the anthracene moiety may tend to increase the flexi-
bility and the hydrophobicity, both of which did not favor selective in-
teractions with the G4s. Furthermore, the symmetric zigzag
arrangement of the naphthalimide units, resulting from the two freely
In contrast the model compounds such as 2, 3 and 4 recognition
capabilities were varied according to the molecular geometry, hydro-
philicity and hydrophobicity’s supports the crescent shaped molecular
architecture with optimized hydrophobic and hydrophilic properties are
essential factors for the discrimination of G4s with various topology
(Table ES5). In addition to that, size of the probe 1, was found appro-
priately fit to stack on surface and bottom tetrads, helps to inhibit the
rotatable methyl groups, on either side of the axially dispersed
(anthracene core) enhanced the steric hindrance (Fig. S25).
π-clouds
To evaluate the effect of the pyridinium units, we incorporated
quinolinium moieties in probe 4 and monitored its opto-analytical ca-
pabilities. Again, probe 4 did not display any selectivity toward nucleic
acids when tested under conditions similar to those used for probe 1
(Fig. S26). The greater hydrophobicity and rigidity of probe 4 presum-
ably resulted in an imbalance among the coulombic and other non-
covalent interactions, leading to inadequate space for probe–ODN
interactions and, therefore, the photophysical properties were invariant
when interacting with the various nucleic acids under the test
conditions.
–
free rotation of styryl (C C) units resulted in highly rigid confinements
–
tend to enhance the intramolecular charge transfer process at maximal
extent. Due to these reasons, high intensity monomer emission was
observed only in the case of parallel G4s (freely available tetrads for π-π
stacking), than the conventional groove binding or intercalation mode
(Fig. 9). Additionally probe 1 parallel G4s discrimination over antipar-
allel/hybrid G4s were comparatively significant from recently reported
far red emissive molecules.45–50
3. Experimental section
CD Spectra and Melting Temperatures
3.1. Synthesis of probes
We did not observe any significant changes in the CD spectra of the
parallel G4–forming sequences [i.e., c-MYC (pu27)51, c-KIT-1, c-KIT-2]
or the antiparallel and hybrid sequences in the absence and presence of
the probe 1 (Fig. S27). Nevertheless, the CD spectra in the 400–700 nm
region in the absence and presence of 1 revealed that the helicity
changed in the region near 500 nm (Figs.S27a–d).Furthermore, the
normalized melting temperatures, measured using the absorbance at
295 nm, revealed a slight increase in the values of Tm (2.2–2.8 ◦C) when
the probe 1 interacted with the parallel G4s; we did not observe such
changes when 1 interacted with the antiparallel G4s. Thus, the probe 1
appears to not significantly stabilize the secondary structures on the
surfaces of the parallel G4s (Fig. S28, Table ES4).
Probe 1: A solution of lutidine (2.00 g, 18.7 mmol) and MeI (2.32 g,
37.3 mmol) in CH2Cl2 (50 mL) was stirred under a N2 atmosphere at
room temperature for 5 h, monitoring through TLC. The creamy white
solid was filtered off, washed with ether (3 × 15 mL), and dried under
vacuum to give 1,2,6-trimethylpyridinium iodide (89%); 1H NMR [400
MHz, CDCl3, δ (ppm)] 2.97 (s, 6H), 4.03 (s, 3H), 7.74 (d, J = 7.90 Hz,
2H), 8.18 (t, J = 7.90 Hz, 1H); 13C NMR (100 MHz, CDCl3) δ 23.4, 42.6,
128.0, 144.2, 156.1; ESI-MS [C8H12N]+, m/z 122.10. A mixture of 1,2,6-
trimethylpyridinium iodide (2.00 g, 8.03 mmol), 4-(diethylamino)sali-
cylaldehyde (3.49 g, 18.1 mmol), and piperidine (catalytic amount) was
heated under reflux in toluene (100 mL) under N2 in a Dean–Stark
apparatus for 16 h, monitoring through TLC. The mixture was concen-
trated to dryness and the solid residue washed with Et2O. A mixture of
the brownish solid (I– salt) and KPF6 (17.2 g, 93.3 mmol) in MeCN (125
mL) was heated under reflux for 14 h and then evaporated to dryness.
The dark-pink semisolid was dissolved in CH2Cl2 and filtered through a
Celite-545 pad. The eluate was further purified through flash column
chromatography (SiO2; EtOAc/CH2Cl2/hexane/MeOH, 6:2:1:1) to
furnish 1 as a dark-pink amorphous solid (350 mg, 70%); 1H NMR [400
MHz, DMSO‑d6, δ (ppm)] 1.12 (t, J = 6.8 Hz, 12H), 3.37 (q, J = 6.8 Hz,
8H), 4.07 (s, 3H), 6.19 (d, J = 2.4 Hz, 2H), 6.29 (dd, J = 5.2, 2.4 Hz, 2H),
7.22 (d, J = 15.6 Hz, 2H), 7.55 (d, J = 8.8 Hz, 2H), 7.74 (d, J = 15.6 Hz,
The crescent shape and flat molecular configuration of probe 1 can
lead to highly stable ion pair–based aggregates in buffered H2O, addi-
tionally stabilized by 100 mM KCl because of common ion effects in
aqueous media. Those aggregates dissolved specifically (disaggregated)
only in the presence of parallel G4–forming ODNs, giving clear solu-
tions. The high selectivity and sensitivity of probe 1 toward the c-MYC
(Pu27) ODN presumably arose because the four flanking guanine resi-
dues on the 5’ end and the purine bases (AAGG) on the 3’ end could
further stabilize the hydrophobic interior of the G4s through additional
π
-stacking interactions.30,31 For the antiparallel and hybrid G4s, how-
ever, the lateral, diagonal, and v-shaped loops tended to increase the
steric hinderance when stacked on either side of the G4 tetrads (MD
studies).
2H), 7.90 (d, J = 8.0 Hz, 2H), 8.05 (t, J = 8.0 Hz, 1H), 10.07 (s, 2H); 13
C
NMR [100 MHz, DMSO‑d6, δ (ppm)] 13.1, 44.4, 97.7, 104.8, 110.7,
From the results obtained from the 1H NMR spectral studies, we
suspect that the guanine residues on the top and bottom sides of the c-
MYC (22 nt) G4s were significantly perturbed, because those tetrads
were open for ligand interactions without any steric factors. Accord-
ingly, the signals of the G7, G9, G13, and G20 nucleobases underwent
significant changes upon interacting with probe 1. The presence of py-
rimidine units in the loops, as well as in the flanking ends, typically leads
to less steric hindrance than does the presence of larger purine residues
(near to top and bottom tetrads of the G4s); for this reason, the changes
in chemical shift were not similar on the two sides of the tetrads.32,33
Additionally, MD studies also supported our hypothesis that parallel G4s
typically form more stable conformations, due to propeller loops,
without harnessing any spurious secondary structure.
Based on these results, we believe that the parallel G4s (c-MYC, c-
KIT-1, c-KIT-2) typically feature propeller loops in a regular or unusual
mode (sometimes acting as lateral loops). Such orientations result in
more flexible conformations in the loop, tending to enhance the surface/
Figure 8. Plausible sensing mechanisms of the pyridinium- and quinolinium-
based probes 1–4.
6

A. Pandith et al.
Bioorganic & Medicinal Chemistry 35 (2021) 116077
evaporation under N2 at room temperature. The crystals that formed
were collected through filtration to furnish 3 as a yellowish powder
(50%); 1H NMR [400 MHz, DMSO‑d6, δ (ppm)] 5.47 (s, 4H), 7.03 (s, 4H),
7.74–7.76 (m, 4H), 7.91 (t, J = 7.6 Hz, 5H), 8.07 (d, J = 6.4 Hz, 4H),
8.49 (d, J = 7.2 Hz, 4H), 8.54–8.57 (m, 7H), 8.76 (d, J = 6.4 Hz, 4H); 13
C
NMR [100 MHz, DMSO‑d6, δ (ppm)] 47.9, 60.2, 127.2, 128.9, 129.9,
131.0, 131.2, 131.4, 132.5, 133.0, 133.4, 134.9, 136.2, 136.6, 140.1,
148.9, 149.1, 162.8, 168.9; HR-FAB MS: calcd for {[C52H36N4O4] +
Cl–}+, m/z 815.2427, found 815.2427.
Probe 4: A solution of 9,10-bisdichloromethylanthracene (2.00 g,
7.27 mmol) and quinoline (2.15 g, 18.2 mmol) in MeCN was heated
under reflux under N2 for 12 h. The mixture was concentrated under
vacuum and then the solid residue dissolved in THF/EtOAc (6:4, v/v).
The solution was filtered through a pad of Celite 545 and then evapo-
rated to dryness. The solid residue was washed with cold THF/Et2O (7:3,
v/v) to furnish 4 as a dark-yellow powder (56%). 1H NMR [400 MHz,
DMSO‑d6, δ (ppm)] δ 7.36 (s, 4H), 7.67 (dd, J = 8.0, 2.8 Hz, 4H),
7.87–7.90 (m, 2H), 8.23–8.26 (m, 2H), 8.47 (dd, J = 8.0, 2.4 Hz, 4H),
8.54 (dt, J = 7.2, 1.2 Hz, 2H), 8.65 (d, J = 5.6 Hz, 4H), 9.29 (d, J = 7.2
Hz, 2H), 9.35 (d, J = 6.8 Hz, 2H); 13C NMR [100 MHz, DMSO‑d6, δ
(ppm)] 53.3, 120.5, 122.8, 125.3, 125.6, 128.7, 130.4, 130.8, 131.1,
132.4, 136.1, 139.6, 147.2, 148.2; HR-FAB MS: calcd for {[C34H26N2] +
Cl–}+, m/z 497.1785, found 497.1783.
Figure 9. Plausible mode of interactions of probe 1 with various types of G4
ODNs according to their loop orientations (propeller, diagonal, and lateral) in
parallel and antiparallel topologies. Respective G4 structures were adopted
from the references noted in the Experimental section.
112.0, 120.3, 131.4, 138.6, 141.4, 151.2, 154.6, 159.3; HR-MS: calcd for
[C30H38N3O2]+ [M]+ m/z 472.2964, found 472.2961.
Probe 2: A solution of 2-methylpyridine (2.00 g, 21.5 mmol) and MeI
(2.00 g, 32.2 mmol) was stirred in dry CH2Cl2 (50 mL) under a N2 at-
mosphere at room temperature. Upon completion of the reaction (TLC),
the mixture was evaporated to dryness. The solid residue was washed
thoroughly with cold EtOAc/Et2O (2:8, v/v) several times and then
dried under high vacuum to furnish 2-methyl-N-methylpyridinium iodide
(2a) as a creamy white solid (94%); 1H NMR [400 MHz, DMSO‑d6, δ
(ppm)] 2.80 (s, 3H), 4.25 (s, 3H), 7.96 (t, J = 6.8 Hz, 1H), 8.06 (d, J =
8.0 Hz, 1H), 8.48 (t, J = 7.6 Hz, 1H), 9.00 (d, J = 6.0 Hz, 1H); 13C NMR
[100 MHz, DMSO‑d6, δ (ppm)] 20.5, 46.0, 125.7, 129.6, 145.4, 146.5,
156.4; ESI-MS+: m/z 108.20. A solution of 2a (0.500 g, 4.62 mmol), 4-
(diethylamino)salicylaldehyde (0.940 g, 4.85 mmol), and piperidine
(catalytic amount) in dry toluene (50 mL) was heated under reflux under
N2 in a Dean–Stark apparatus for 16 h. Upon completion of the reaction
(TLC), the solvent was evaporated to dryness. The dark-red powdery
residue and KPF6 (2.13 g, 11.6 mmol) were dissolved in ethylene
dichloride (50 mL) and heated under reflux under N2 for 12 h. The
mixture was filtered through a pad of Celite 545 and evaporated. The
dark-pink powdery residue was purified through flash silica column
chromatography (SiO2; EtOAc/CH2Cl2/hexane/MeOH, 6:1:2:1) to
furnish 2 as a dark-red solid (60%); 1H NMR [400 MHz, DMSO‑d6, δ
(ppm)] 1.13 (t, J = 6.99 Hz, 6H), 3.38 (q, J = 7.00 Hz, 4), 4.20 (s, 3H),
6.20 (d, J = 2.33 Hz, 1H), 6.33 (dd, J = 8.97, 2.32 Hz, 1H), 7.22 (d, J =
15.65 Hz, 1H), 7.59 (d, J = 8.80 Hz, 2H), 7.97 (d, J = 16.65 Hz, 1H),
8.24 (d, J = 15.72 Hz, 1H), 8.35 (d, J = 8.22 Hz, 1H), 8.68 (d, J = 6.15
Hz, 1H), 10.23 (s, 1H); 13C NMR [100 MHz, DMSO‑d6, δ (ppm)] 13.1,
44.2, 44.5, 45.8, 97.5, 105.0, 109.4, 110.7, 122.4, 123.4, 132.1, 140.5,
143.1, 145.4, 151.8, 154.4, 159.9; HR-MS: calcd for [C18H23N2O]+ [M]+
m/z 283.1810, found 283.1808.
3.2. Computational studies
Density functional theory (DFT)-based molecular simulations were
performed according to previously reported procedures.34 Energy-
minimized structures of the probes were determined in H2O as an im-
plicit medium (SM8 model). The xyz coordinates of the thus-obtained
DFT-based geometries were used for molecular docking studies and
subsequent MD studies with G4 structures.
Molecular modeling and MD simulations were performed initially
with G4s of various topologies, retrieved from the Protein Data Bank
(PDB) ID source: c-MYC (22 nt, Parallel) PDB ID 2L7V,35 PDB ID 2 MB3
(hybrid),36 and 22AG (antiparallel) PDB ID 143D.37 To identify the
binding sites and conformations of the ligands to the G4s, various to-
pological molecular docking studies were performed using Autodoc 4.2
by choosing a sufficiently large grid box to cover the whole G4.38 The
Lamarckian genetic algorithm was used for docking with 250,000 en-
ergy evaluations from an initial population of 150 randomly placed in-
dividuals having a mutation rate of 0.02 along with a maximum number
of 27,000 generations. A crossover rate of 0.8 and 300 iterations of local
search were used. Finally, 10 independent docking runs were performed
for each G4 topology.
Further, MD simulations were performed for the modeled G4-ligand
complexes through all-atom MD simulations using the GROMACS-
2018.6 package.39 The force field parameters and atomic charges for
the ligands were derived from the generalized AMBER force field2
(GAFF2) and bcc charges using the ANTECHAMBER module of the
AMBERTOOLS20 package.40 Topologies for DNA were generated from
AMBER99SB-ILDN force field. The modeled complex was placed in the
triclinic box with a minimum distance of 1.2 nm from the box edges
under the periodic boundary conditions. The solvent was filled using the
TIP3P water model and the total system was neutralized by replacing
Na+ and Cl– with solvent molecules. Next, energy minimization was
performed for 50,000 steps using the steepest descent algorithm with the
energy tolerance of 1000 kJ molꢀ 1 nmꢀ 1. NVT (number of volume
temperature) and NPT (number of pressure temperature) equilibration
simulations of 2 ns were performed by restraining the DNA-ligand
complex to attain a temperature and pressure of the simulated system
of 300 K and 1 bar, respectively, using v-rescale and the Parinello-
Rahman barostat. Finally, 50-ns production simulations were per-
formed using the NPT ensemble. Long-range interactions were handled
using the particle mesh Ewald (PME) method. Hydrogen bonds were
constrained with the LINCS algorithm. Binding energy calculations were
Probe 3: A solution of 1,8-naphthalic anhydride (2.00 g, 10.1 mmol)
and 4-(aminomethyl)pyridine (1.18 g, 11.6 mmol) in anhydrous EtOH
(50 mL) was heated under reflux for 8 h. Upon completion of the reac-
tion (TLC), the mixture was cooled to room temperature and left for 10 h
to furnish creamy-white needle-shaped crystals, which were washed
with EtOH/H2O (2:8) at 10 ◦C and dried under vacuum to give 2-(4-
pyridylmethyl)-1H-benz[de]isoquinoline-1,3(2H)-dione (3a, 88%). 1H
NMR [400 MHz, CDCl3, δ (ppm)] 5.31 (s, 2H), 7.31 (dd, J = 1.2, 3.2 Hz,
2H), 7.69–7.73 (m, 2H), 8.18 (dd, J = 7.6, 0.8 Hz, 2H), 8.47 (dd, J = 2.8,
1.6 Hz, 2H), 8.55–8.57 (m, 2H); 13C NMR [100 MHz, CDCl3, δ (ppm)]
42.6, 122.3, 123.3, 127.1, 128.3, 131.7(d), 134.5, 145.9, 150.0, 164.2;
API+-MS [C18H12N2O2 + H]+, m/z 289.18. A mixture of 3a (2.00 g, 6.94
mmol) and 9,10-bisdichloromethylanthracene44 (0.960 g, 3.47 mmol) in
dry MeCN was heated under reflux under N2 for 14 h. Upon completion
of the reaction (TLC), the solvent was evaporated to dryness. The solid
residue was dissolved in THF/EDC/MeOH (2:3:5) and subjected to slow
7

A. Pandith et al.
Bioorganic & Medicinal Chemistry 35 (2021) 116077
performed using the g mmpbsa module.41 All analyses were performed
using built-in GROMACS tools. Visualizations and image rendering were
performed using visual molecular dynamics (VMD)42 and Chimera43
packages. Gnuplot was used for drawing graphs.
4
5
6
Konig SLB, Evans AC, Huppert JL. BioMol Concepts. 2010;1:197–213.
Murat P, Balasubramanian S. Curr Opin Gen Dev. 2014;25:22–29.
Suseela Y, Narayanaswamy V, Pratihar S, Govindaraju T. Chem Soc Rev. 2018;47:
1098–1131.
7
Yuan JH, Shao W, Chen S-B, Huang Z-S, Tan J-H. Biochem Biophys Res Commun. 2020;
531:18–24.
8
9
Pandith A, Siddappa RG, Seo YJ. J Photochem Photobiol C: Rev. 2019;40:81–116.
Vummidi BR, Alzeer J, Luedtke NW. ChemBioChem. 2013;14:540–558.
4. Conclusion
10 Ma D-L, Zhang Z, Wang M, Lu L, Zhong H-J, Leung C-H. Chem Biol. 2015;22:
812–828.
We have demonstrated that probe 1, a simple pyridinium-based salt,
allows the identification of parallel G4s over antiparallel and hybrid G4
topologies as well as other non-canonical/canonical forms of DNA (i.e.,
ssDNA, dsDNA, Poly G, triplex, TWJ). Probe 1 recognized the tested
parallel G4s with a selective switch on response in the far-red emission
region, with excellent selectivity and sensitivity. We analyzed the se-
lective sensing capabilities of probe 1 using UV–Vis spectroscopy,
fluorescence spectroscopy, 1H NMR spectroscopy, and molecular dy-
namics (MD)–based simulation studies. We confirm that the disaggre-
gation of probe 1 was very effective in the presence of parallel
G4–forming ODNs, due to the presence of highly available free surface
11 Grande V, Shen C-A, Deiana M, et al. Chem Sci. 2018;9:8375–8381.
12 Zhang X, Wei Y, Bing T, et al. Sci Rep. 2017;7:4766.
13 Zhang S, Sun H, Wang L, et al. Nucleic Acids Res. 2018;46:7522–7532.
14 Chen Y, Yan S, Yuan L, et al. Org Chem Front. 2014;1:267–270.
15 Tseng T-Y, Wang Z-F, Chien C-H, Chang T-C. Nucleic Acids Res. 2013;41:
10605–10618.
16 Huang W-C, Tseng T-Y, Chen Y-T, et al. Nucleic Acids Res. 2015;43:10102–10113.
17 He H-Z, Chan D-S-H, Leung C-H, Ma D-L. Nucleic Acids Res. 2013;41:4345–4359.
18 Kumar R, Chand K, Bhowmik S, et al. Nucleic Acids Res. 2020;48:1108–1119.
19 Yang C, Hu R, Li Q, et al. ACS Omega. 2018;3:10487–10492.
20 Horn D, Rieger J. Angew Chem Intl Ed. 2001;40:4330–4361.
´
ˇ
21 Klikar M, Solanke P, Tydlitat J, Bures F. The Chemical Record. 2016;16:1886–1905.
22 Jin B, Zhang X, Zheng W, et al. Anal Chem. 2014;86:943–952.
23 Chen S-B, Wu W-B, Hu M-H, et al. Chem Commun. 2014;50:12173–12176.
24 Yang Q, Xiang J, Yang S, et al. Nucleic Acids Res. 2010;38:1022–1033.
25 Zhang L, Cheng Er J, Li X, et al. Chem Commun. 2015;51:7386–7389.
26 Asamitsu S, Bando T, Sugiyama H. Chem Eur J. 2019;25:417–430.
27 Harikrishna S, Kotaru S, Pradeepkumar PI. Mol BioSyst. 2017;13:1458–1468.
28 Deiana M, Chand K, Jamroskovic J, Obi I, Chorell E, Sabouri N. Angew Chem Intl Ed.
2020;132:906–912.
area, resulting in additional π-stacking interactions. In addition, we have
postulated our hypothesis based on systematic analysis of various model
compounds 2, 3 and 4 by changing, molecular geometry, hydrophilicity
and hydrophobicity under similar experimental conditions. Thus, simple
molecules and cost-effective strategies can be used to identify G4s with
appreciably good topological selectivity. Studying rationally designed
probes appears to be excellent initial platform for designing novel G4-
tracking functional materials and their therapeutics.
29 Xie X, Reznichenko O, Chaput L, Martin P, Teulade-Fichou M-P, Granzhan A. Chem
Eur J. 2018;24:12638–12651.
30 Bhat J, Chatterjee S. RSC Adv. 2016;6:36667–36680.
31 Bessi I, Wirmer-Bartoschek J, Dash J, Schwalbe H. Modern Magnetic Reson. 2018;13:
2189–2210.
32 Le DD, Antonio MD, Chana LKM, Balasubramanian S. Chem Commun. 2015;51:
8048–8050.
Declaration of Competing Interest
33 Dash J, Shirude PS, Hsu S-T-D, Balasubramanian S. J Am Chem Soc. 2018;130:
15950–15956.
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
34 Pandith A, Seo YJ. J Inorg Biochem. 2020;203, 110867.
35 Dai J, Carver M, Hurley LH, Yang D. J Am Chem Soc. 2011;133:17673–17680.
36 Chung WJ, Heddi B, Tera M, Iida K, Nagasawa K, Phan AT. J Am Chem Soc. 2013;135:
13495–13501.
37 Wang Y, Patel DJ. Structure. 1993;1:263–282.
Acknowledgement
38 Morris GM, Huey R, Lindstrom W, et al. J Comput Chem. 2009;30:2785–2791.
39 Abraham MJ, Murtola T, Schulz R, et al. SoftwareX. 2015;1:19–25.
¨
40 Salomon-Ferrer R, Gotz AW, Poole D, Le Grand S, Walker RC. J Chem Theory Comput.
This study was supported by the Basic Science Research Program
2013;9:3878–3888.
through
the
National
Research
Foundation
of
Korea
41 Kumari R, Kumar R, Consortium OSDD, Lynn AJ. Chem Inform Model. 2014;54:
1951–1962.
(2017R1A2B4002398), funded by the Republic of Korea.
42 Humphrey W, Dalke A, Schulten K. J Mol Graphics. 1996;14:33–38.
43 Pettersen EF, Goddard TD, Huang CC, et al. J Comput Chem. 2004;25:1605–1612.
44 Pandith A, Uddin N, Choi C-H, Kim H-S. Sens Actuators B. 2017;247:840–849.
45 Doria F, Nadai M, Zuffo M, Perrone R, Freccero M, Richter SN. Chem Commun. 2017;
53:2268–2271.
Appendix A. Supplementary material
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.bmc.2021.116077.
46 Yu L, Yang Q, Xiang J, Tang Y. Anal Methods. 2015;7:5483–5489.
47 Zhang F, Li G, Lv F-L, Jian G-B, Wang H-X, Wang M-Q, Li S. Tet Lett. 2018;59:
3272–3278.
References
48 You D, Liu L, Yang Q, Wu X, Li S, Li A. Dyes Pigm. 2020;176, 108222.
49 Gao F, Cao S, Sun W, Long S, Fan J, Peng X. Dyes Pigm. 2019;171, 107749.
50 Wang X-L, Zhang J-T, Sun X, Yang D-W, Tang Y-L. Dyes Pigm. 2021;185, 108882.
51 Calabrese DR, Chen C, Leon EC, et al. Nat Commun. 2018;9:4229. https://doi.org/
10.1038/s41467-018-06315.
1
2
3
Yang D, Okamoto K. Future Med Chem. 2010;2:619–646.
Manna S, Srivatsan SG. RSC Adv. 2018;8:25673–25694.
Largy E, Granzhan A, Hamon F, Verga D. Top Curr Chem. 2013;330:111–177.
8