The influence of positional isomerism on G-quadruplex binding and anti-proliferative activity of tetra-substituted naphthalene diimide compounds

Article abstract of DOI:10.1016/j.bmc.2013.05.027

The synthesis together with biophysical and biological evaluation of a series of tetra-substituted naphthalene diimide (ND) compounds, are presented. These compounds are positional isomers of a recently-described series of quadruplex-binding ND derivatives, in which the two N-methyl-piperidine-alkyl side-chains have now been interchanged with the positions of side-chains bearing a range of end-groups. Molecular dynamics simulations of a pair of positional isomers are in accord with the quadruplex stabilization and biological data for these compounds. Analysis of structure-activity data indicates that for compounds where the side-chains are not of equivalent length then the positional isomers described here tend to have improved cell proliferation potency and in some instances, superior quadruplex stabilization ability.

Full text of DOI:10.1016/j.bmc.2013.05.027

Accepted Manuscript
The influence of positional isomerism on G-quadruplex binding and anti-pro‐
liferative activity of tetra-substituted naphthalene diimide compounds
Sheila Mpima, Stephan A. Ohnmacht, Maria Barletta, Jarmila Husby, Luke C.
Pett, Mekala Gunaratnam, Stephen T. Hilton, Stephen Neidle
PII:
S0968-0896(13)00463-X
http://dx.doi.org/10.1016/j.bmc.2013.05.027
BMC 10853
DOI:
Reference:
To appear in:
Bioorganic & Medicinal Chemistry
Please cite this article as: Mpima, S., Ohnmacht, S.A., Barletta, M., Husby, J., Pett, L.C., Gunaratnam, M., Hilton,
S.T., Neidle, S., The influence of positional isomerism on G-quadruplex binding and anti-proliferative activity of
tetra-substituted naphthalene diimide compounds, Bioorganic & Medicinal Chemistry (2013), doi: http://dx.doi.org/
10.1016/j.bmc.2013.05.027
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

2
Graphical Abstract
The influence of positional isomerism on G-
quadruplex binding and anti-proliferative
activity in tetra-substituted naphthalene
diimides
Leave this area blank for abstract info.
Sheila Mpima, Stephan A. Ohnmacht, Maria Barletta, Jarmila Husby, Luke C. Pett, Mekala Gunaratnam,
Stephen T. Hilton and Stephen Neidle
The School of Pharmacy, University College London, London WC1N 1AX, UK

3
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com
The influence of positional isomerism on G-quadruplex binding and anti-proliferative
activity of tetra-substituted naphthalene diimide compounds
Sheila Mpima, Stephan A. Ohnmacht, Maria Barletta, Jarmila Husby, Luke C. Pett, Mekala Gunaratnam,
Stephen T. Hilton and Stephen Neidle*
The School of Pharmacy, University College London, 29-39 Brunswick Square, London WC1N 1AX, UK
ARTICLE INFO
ABSTRACT
Article history:
Received
Received in revised form
Accepted
The synthesis together with biophysical and biological evaluation of a series of tetra-substituted
naphthalenediimide (ND) compounds, are presented. These compounds are positional isomers of
a recently-described series of quadruplex-binding ND derivatives, in which the two N-methyl-
piperidine-alkyl side-chains have now been interchanged with the positions of side-chains
bearing a range of end-groups. Molecular dynamics simulations of a pair of positional isomers
are in accord with the quadruplex stabilization and biological data for these compounds.
Analysis of structure-activity data indicates that for compounds where the side-chains are not of
equivalent length then the positional isomers described here tend to have improved cell
proliferation potency and in some instances, superior quadruplex stabilization ability.
Available online
Keywords:
G-quadruplex
naphthalenediimide
pancreatic cancer
molecular dynamics
2013 Elsevier Ltd. All rights reserved.
have significant therapeutic consequences.8 In addition, there
is evidence that this can be accompanied by a DNA damage
1. Introduction
response, augmenting the consequences for affected cells.9
Quadruplex DNAs and RNAs are four-stranded structures
formed by the stabilization of repetitive G (guanine) tracts in
duplex or single-stranded nucleic acids by means of Hoogsteen
hydrogen bonds to form stable G-quartets.1-3 The targeting of
quadruplex DNA elements in telomeres and in oncogene
promoter regions4,5 is emerging as a useful approach to
anticancer therapy, especially now that the existence of
quadruplex DNAs as structural nucleic acid motifs distinct from
double-stranded DNA has been validated at the genomic level.6,7
The concept has several distinct consequences, in particular.
The task of designing small molecules that have selectivity for
quadruplexes in general over duplex DNA affinity, has been
effectively explored,10,11a with a large number of small molecules
binding G-quadruplexes now described in the literature. The
majority are synthetic compounds sharing the common features
of a planar heteroaromatic core together with one or more
cationic side-arms. Some have metallo- containing groups.11b
However, the goal of designing ligands to target a single
quadruplex is altogether more demanding, not least because only
a small fraction of the potential quadruplex landscape has been
examined. In reality this may not be a therapeutic disadvantage as
a quadruplex-focused poly-targeting approach may be especially
useful for those genetically highly aberrant solid cancers where
multiple oncogene pathways may be dis-regulated.

Telomeric DNA length is maintained in the majority of
human cancers by means of the reverse transcriptase activity
of the telomerase enzyme complex, whose expression is up-
regulated, contrasting with the lack of telomerase expression
in normal somatic cells. Single-stranded telomeric DNA is
the substrate for the enzyme and quadruplex formation or
ligand-induced stabilization results in telomerase inhibition,
with consequential telomere shortening, cellular senescence
and selective cell death as a consequence of DNA damage
induction.4
A number of naphthalene diimide compounds have been
described, some with high quadruplex affinity and cellular
potency.12-17 In general these have straight forward chemical
accessibility and potential for ready analogue development. We
have previously disclosed several series of tetra-substituted
naphthalene diimide compounds,18-25 a number of which have
exceptional telomeric quadruplex-binding ability and associated
potent cellular proliferation properties in a panel of cancer cell
lines. Most recently we have used structure-based design
methods to optimize the pharmacological properties of several
analogues with four side-chains each terminating in a N-methyl-
piperazine group.25 Those with di-N-methyl-piperazine, di-

The existence of quadruplex-forming sequences within
promoter regions, and close to transcription start sites, that
can be stabilized by appropriate small molecules, has been
demonstrated for a number of oncogenes and other genes
involved in cellular growth proliferation, such as c-myc, c-kit,
HIF, VEGF and HSP90. Quadruplex stabilization leads to
inhibition of transcription for that particular gene, which may

4
morpholine end-groups have superior (by up to 10-fold) cellular
potency, notably against a panel of pancreatic and non-small-cell
lung cancer cell lines.
with its less bulky and less rigid framework, compared to the
cyclic 6-membered N-methyl-piperazine moiety. This result is
particularly interesting as it suggests that having an oxygen atom
(without a further available nitrogen atom, as in the morpholino-
analogues) within the side chain, can provide significant
stabilization in the FRET assay, although the cellular assay
results suggest that the biological profile of this compound is
sub-optimal (see below).
This paper reports on a study to examine the effects on
quadruplex binding and cell growth inhibition, of interchanging
the positions of the extended side-chain substituents around the
naphthalene diimide core, for these most recent active
compounds,25 in order to ensure that these features have been
optimized. Thus, earlier compounds 1a-5a are compared with
their isomers 1-5, whose synthesis and evaluation is reported
here. Three quadruplex sequences have been used as examples of
diverse quadruplex types and because of their potential
involvement in the mode of action of tetra-substituted
2.3 Short term SRB and telomere length assays
All compounds were screened for their anti-proliferative
activities against several cancer cell lines (breast, renal (x2),
pancreatic (x2), lung) in a short-term 96 hr sulphorhodamine
(SRB) assay (Table 2). A general trend is apparent, that the
majority of the new positional isomers, compounds 1-5, are
generally more active in these assays than their previously-
synthesized counterparts (compounds 1a-5a). It is notable that
the change in substituent positions has improved the biological
profile of compounds 4aand 5a, with their positional isomers,
compounds 4and 5, having sub-micromolar potencies in the
majority of cancer cell lines in the panel. The trend of particular
activity in the pancreatic and lung carcinoma cell lines MIA
PaCa2 and A549, observed for compounds in the earlier 1a-5a
series,25 is still apparent for the positional isomers. Data is now
reported as well (where available) for the chemo-resistant
PANC1 pancreatic carcinoma28 cell line, which shows that the
isomer of the previously most active compound 2a, is five-fold
more active in PANC1 than in MIA PaCa2 cells, with a large
therapeutic window compared to its activity in the normal human
fibroblast cell line WI38.
naphthalene diimides;18-25
a well-studied human telomeric
sequence and two promoter quadruplexes from the HSP90
gene.22,26 The senescent cellular responses to the lead compound
identified in a recent study (2a) suggests that it targets telomeres
and elicits and DNA damage/stress response following telomeric
quadruplex interaction.25 Computational molecular modeling has
been used to rationalize the effects of positional isomerism, and
is based on the recent co-crystal structures of several naphthalene
diimide analogues with human telomeric quadruplexes.23, 25
2. Results and Discussion
2.1 Chemistry
Synthesis of compounds 1-5 followed our recently-developed
general procedure.25 Dibromoisocyanuric acid was reacted with
anhydride (I) in sulfuric acid to give the bis-brominated
intermediate (II), followed by two sequential microwave
reactions with selected primary amines, which afforded the crude
tetra-substituted naphthalene diimides. Purification on a C18
reversed phase preparative HPLC system produced the final
compounds.
Possible effects on telomere length were also examined. MIA
PaCa2 cells were exposed to compounds 2 and 2a at various
concentrations in the range 1-7 nM for a period of three to five
weeks. Cells were treated twice a week and counted at the end of
each week. Cell samples were taken at weekly intervals and
Southern blot analysis was performed on extracted DNA. No
changes in telomere length were found with either compound at
any of the time points (blots not shown here).
2.2 Fluorescence Resonance Energy Transfer (FRET)
studies
The extent of G-quadruplex stabilization by compounds 1-5
and 1a-5a, as measured with
a high-throughput FRET
2.4 Molecular modeling studies
(Fluorescence Resonance Energy Transfer) technique,27 has been
used to compare these naphthalene diimide derivatives with their
positional isomers (Table 1). In general differences in melting
temperatures (ΔTm values) between individual isomers are
greatest when the respective side chains are of different linker
length. Both compounds in the pair 2 and 2a have almost
identical melting temperatures to each other (and to all the
quadruplexes) for all three quadruplexes; it is notable that in both
compounds the side chains contain three -CH2- groups. The
stabilizing effects of compounds 3, 3a and 3b are also closely
similar for quadruplex binding. However differences in duplex
binding are apparent, with the two compounds 3 and 3a showing
Molecular modeling studies were performed with the most
potent positional isomers, the compounds, 2 and 2a. Figure 2
shows a schematic view of these two simulated quadruplex (G4)
ligand complexes; (a) the original X-ray structure of the 21-mer
complex with compound 2a (PDB id 3UYH), and (b) the
modeled complex with compound 2 and the 21-mer quadruplex,
based on this X-ray structure, obtained by interchanging the N-
methyl piperazine and morpholino side chains attached to the
naphthalene diimide core. The core itself remained unchanged,
which is clearly visible in the structural alignment (c) of the two
complexes. An overview of the 50 ns MD simulations is given in
Table 3. RMSD values for the G4 backbone atoms as a function
of the simulation time were used as a measure of stabilization of
the two models (G4-2a, G4-2), comparing both initial reference,
and time-averaged structures of the models (with the latter
providing superior insight into the structures reaching the
plateau). The trajectories of the G4-ligand complexes were
stabilized at around ∼3.1 Å (and around ∼1.1 Å with respect to
the time-averaged structure obtained from the MD) for the G4-2a
complex at around ~3.0 Å (and at around ~1.9 Å for the time-
averaged structure) for the G4-2 complex. This suggests that 2a
has improved G4-stabilizing properties comparing to its isomer,
compound 2. The trajectory of the native G4 21-mer, which was
some duplex DNA stabilization even at
concentration.
a low DNA
Significant differences in melting temperatures between pairs
of positional isomers are apparent when side chains are less than
two carbon atoms long. Thus comparing pairs of compounds 4
and 4a, and 5 and 5a,the melting temperature differences are up
to 21C at a 1 μM quadruplex concentration (4 vs 4a with the
F21T sequence) and up to 26C at 2 μM (5vs 5a with the
HSP90A sequence). Differences in melting temperatures of up to
7C were observed between1 and 1a, even though their side
chains contain equal numbers of -CH2-groups. These particular
differences may be due to the acyclic nature of the OMe- group,

5
used as a control structure, was stabilized at around ~2 Å (and
around ~1.2 Å respectively).
proliferative activity and higher ΔTm values across the three
quadruplexes, compared to compound 5a. The same pattern of
behaviour was not observed for compounds 4 and 4a, suggesting
that the one-carbon atom difference between the furanose and
pyranose rings in the two pairs of compounds is sufficient to
produce differences in quadruplex stabilization and biological
response, assuming that the latter is a consequence of the former.
This suggests a concept of ‘virtual overall elongation’ by one
side chain atom. Having the longer side-chain group (in this case
n=3 N-methyl-piperazine) connected via the diimide
functionality possibly provides an overall net increase in
quadruplex stabilization as the optimal chain length may not
necessarily be n=3, but actually is between n=2 and n=3, which is
better achieved by having the shorter side chains in the optimal
position to interact with the quadruplex structure (Figure 4).
Cluster analysis was applied to the large amount of data
generated in order to provide a statistical description of the
dynamics of the G4-ligand complexes. This analysis used the
1920 frames ( ~48 ns) extracted from the MD trajectories at time
interval of 25 ps for the matrix construction, with a 2.0 Å RMSD
cut-off applied for the neighbour search. The resulting clusters
are then represented by middle structures, the conformations
sampled during the simulation at 300 K. While the G4-2a
complex was represented by one cluster only (i.e ~ 100% of the
conformational space sampled), the first seven clusters were
required to cover the total ensemble of sampled conformational
space for the G4-2 complex, and even this only included85% of
the space. Cluster 1 was always the predominant conformer;
Figure 3 shows the predominant conformations which were
adopted by these two G4-ligand complexes through the
simulation time, with respect to their starting structures in Figure
2 (prior to the MD simulations).
The molecular dynamics simulation studies on the telomeric
quadruplex complexes with the positional isomers 2 and 2a have
revealed that an anti-clockwise rotation of 90° in the G-quartet
plane for compound 2, so that it ends up in the same position as
the experimental pose for compound 2a (Figure 3e). This is in
accord with the very similar quadruplex stabilization data from
the FRET assay (Table 2) as well as closely similar inhibition of
cell proliferation. This equivalence, not yet evaluated by
simulation, appears to be the case only when the side-chains are
all of equal length, and possibly with the end-groups having
similar steric features, as in the N-methyl-piperazine and
morpholine groups of 2 and 2a.
In Figure 3(a) the original conformation adopted by
compound 2a in the X-ray structure (yellow) has only slightly
changed through the 50 ns MD run, as represented by one
predominant conformation (green). One morpholino side chain
has moved from its position, becoming closer to the quadruplex
loops (the rmsd between the two structures is ~ 1.8 Å); Figures 3
(b-d). In the case of the complex with compound 2, progressive
anti-clockwise rotation of the ligand takes place until the ligand
has moved by 90° from its starting structure position. This is
represented by seven predominant clusters (sampling ~85%
conformational space). Only the three most significant
changes/cluster representations are shown here, which follow
ligand rotation as the simulation progresses. Calculation of
binding free energies for 2 and 2a has given very similar values.
This study has shown that the positional isomer (compound 2)
of the previously-reported25 lead compound 2a, does not show
any advantages in terms of either quadruplex binding or
inhibition of cell proliferation, at least in the lines evaluated to
date, and the former compound is currently being evaluated in
tumor xenografts. On the other hand, the present study has
revealed that the positional isomers of more asymmetrically-
shaped compounds such as 3a and 4a are significantly more
potent and thus may provide new directions for lead compound
generation. The fact that clogP values are identical for both
members of an isomer pair indicates that differences in cellular
behavior are not due to uptake differences but to more subtle
differences in binding to their cellular targets.
3. Conclusions
We have generated a small focused library of tetra-substituted
naphthalene diimides that are positional isomers of the
previously-disclosed series,25 using a short three-step synthesis
utilizing microwave irradiation. Several of the isomers showed
very potent stabilization of several representative quadruplexes,
with ΔTm values of up to 34°C at a 1 μM concentration. They
compare well with the behaviour of their positional isomers, with
one new compound showing superior stabilizing abilities for all
three quadruplexes compared to the behavior of its previously-
reported isomer, together with negligible duplex stabilization.
This compound (5), with two pyranose end-groups, is also active
in the cancer cell line panel, in contrast to its positional isomer.
Several of these compounds are potent inhibitors of cell growth
in pancreatic and lung cancer cell lines, with IC50 values for two
compounds (2 and 2a) against the chemo-resistant PANC1 line,
of 2-3 nM. Long-term telomere length studies showed no change
in mean telomere length, indicating that inhibition of telomerase
is not involved in the anti-proliferative response to these agents.
Further studies will aim to elucidate the mode of action of these
compounds in responsive cell lines, following the observations
that compound 2a produces senescence in affected cells,
accompanied by changes in the expression of a number of
cellular stress/DNA damage response genes.
4. Experimental procedures
4.1 Chemistry
All chemicals, reagents and solvents were purchased from
Sigma-Aldrich, Alfa Aesar and Lancaster Synthesis and used
without further purification. Solvents were supplied by Aldrich,
VWR and Fisher scientific. Microwave reactions were performed
in an Initiator microwave (Biotage, Sweden). Column
chromatography was performed using BDH silica gel (BDH
153325P). HPLC analysis was carried out with a Gilson
apparatus combining a 322 PUMP and an Agilent 1100 SERIES
detector, using a C18 5μ (100 x 4.6 mm) column (41622271 (W),
YMC, Japan), at a flow of 1 mL/min. Preparative HPLC was
carried out with a Gilson apparatus combining a 322 PUMP and
a UV/VIS-155 detector with detection at 254 nm, using a C18 5μ
(100 x 20 mm) column (201022272) (W), YMC, Japan, at a flow
of 20 mL/min. Water and methanol with 0.1 % formic acid were
used as solvents for HPLC. NMR spectra were recorded at 400
MHz Bruker spectrometer in CDCl3 (with 0.05 % TMS,
Cambridge Isotope Laboratories, USA). NMR spectra were
analyzed with MestReC 4.5.6.0 with chemical shifts using TMS
as a standard (δ = 0.00 ppm). NMR multiplicity abbreviations are
s (singlet), bs (broad singlet), d (doublet), t (triplet), 4q (quartet),
In general the ΔTm values for quadruplex stabilization overall
do not correlate with the IC50 anti-proliferative data. However in
one instance (compound 5), it appears that the longer side chain
amino functionality attached to the 6-membered diimide moieties
of the naphthalene diimide core, results in superior anti-

6
5q (quintet), and m (multiplet). Coupling constants J are reported
as observed in Hertz (Hz). High Resolution Mass spectra
(HRMS) were measured on a Micromass Q-TTOF Ultima Global
tandem mass spectrometer run under electrospray ionisation
(ESI), and processed using the MassLab 3.2 software. The
general procedure for the synthesis of compounds 1-6 has been
previously published and is available online.25 Compounds 1a-5a
have been synthesized according to this previously published
procedure.
(CDCl3, 100MHz) δ 166.2 (2x quat), 163.0 (2x quat), 149.4 (2x
quat), 125.8 (2x quat), 121.2 (2x quat), 118.4 (2x quat), 102.0 (2x
quat), 67.6 (4x CH2), 55.1 (2x CH3), 53.2 (2x CH2), 50.3 (4x
CH2), 49.1 (2xCH2), 43.7 (2x CH2), 38.4 (2x NMe), 35.2 (4x
CH2), 30.9 (4x CH2). HRMS (ES+) calculated for C42H60N8O6
[M+H]+ 773.4714, found 773.4713.
Analytical data for 1a-5a has previously been reported (see
reference 25, with assigned compound numbers 2a, 2c, 2d, 2e
and 2h)
Compound 1
1H NMR (CDCl3 400MHz)δ9.43 (2H, t,J= 5.5Hz, 2x NH), 8.20
(2H, s, ArH), 4.26 (4H, t, J= 7.4Hz, 2x CH2), 3.66 (4H, q, J=6.4,
4.0Hz, 2x CH2), 3.59 (4H, t, J=5.8Hz, 2x CH2), 3.42 (6H, s, 2x
OMe), 2.54(18H, m, 9x CH2), 2.25 (6H, s, 2x CH3), 2.07 (4H, m,
2x CH2), 1.99-1.92 (6H, m,3xCH2), 13C NMR (CDCl3 100 MHz)
δ 166.2 (2x quat), 163.1 (2x quat), 149.3 (2x quat), 125.9 (2x
quat), 121.2 (2x quat), 118.3 (2x CH), 102.1 (2x quat),70.1 (4x
CH2), 58.8 (4x CH2), 56.0 (2x CH2), 55.1 (4x CH2), 52.9 (4x
CH2), 45.9 (2x CH3), 40.5 (2x CH2), 38.8 (2x CH2), 29.5 (2x
CH3), 25.2 (2x CH2).HRMS (ES+) calculated for C38H56N8O6
[M+H]+721.4069, found 721.4401.
4.2 FRET assay
The following oligonucleotide sequences, all purchased from
Eurofins,
were
used:
F21T:
(5'-FAM-GGG
TTAGGGTTAGGGTTAGGG-TAMRA-3'), Hsp90A: (5'-FAM-
GGGCCAAAGGGAAGGGGTGGG-TAMRA-3'), Hsp90B: (5'-
FAM-GGGCGGGCCAAAGGGAAGGGG-TAMRA-3')T-Loop:
(5'-FAM-TATAGCTATATTTTTTTATAGCTATA-TAMRA-
3'). TAMRA (6-carboxytetramethylrhodamine) is the acceptor
fluorophore, and FAM (6-carboxyfluorescein) is the donor
fluorophore. From 50 μM stock solutions, 400 nM solutions in
FRET buffer (60 mM potassium cacodylate pH 7.4) were
prepared. The nucleotides were annealed by heating the samples
to 95 °C for 10 min and allowing them to cool down to RT within
3 h. 10 mM solutions of the compounds in deionised water were
prepared and diluted to double of the required concentrations
with FRET buffer. In RT-PCR 96 well plates (MJ Research,
Waltham, MA), each well was loaded with 50 μL of nucleotide
solution and 50 μL of drug solution.
Drug concentrations of 0.1, 0.5, 1, 2, 3, 4 and 5 were used,
and every drug concentration was repeated three times.
Measurements were made on a DNA Opticon Engine (MJ
Research) with excitation at 450 – 495 nm and detection at 515 –
545 nm. The fluorescence was read at intervals of 0.5 °C in the
range 30 – 100 °C. Before each reading the temperature was held
constant for 30 s. The raw data were processed using Origin
(Version 7.0, OriginLab Corp.). The graphs were smoothed using
a 10-point running average and normalized. The melting
temperatures were obtained by determining the maxima of the
first derivative of the smooth melting curves. The ΔTm value is
the melting temperature difference between the oligonucleotide
with ligand and the negative control native sequence in the
absence of ligand.
Compound 2
1H NMR(CDCl3 400MHz) δ 9.43 (2H, t, J=5.5Hz, 2x NH), 8.20
(2H, s, ArH), 4.26 (4H, t, J=7.5Hz, 2x CH2), 3.80(8H, t, J=4.7Hz
, 4x CH2), 3.64 (4H, q, J= 6.5, 2.0Hz, 2x CH2), 2.97 (8H, s, 4x
CH2), 2.79 (8H, s, 4x CH2), 2.68-2.60(22H, m, 11x CH2),2.05-
1.95(8H, m, 4x CH2).13C NMR(CDCl3100MHz) δ166.1 (2x
quat), 163.1 (2x quat), 149.2 (2x quat), 125.8 (2x quat), 121.3 (2x
quat), 118.5 (2x CH), 101.9 (2x quat), 66.6 (4x CH2), 56.1 (2x
CH3), 55.1 (2x CH2), 53.6 (4x CH2), 53.3 (4x CH2), 50.3 (4x
CH2), 43.8 (2x CH2), 41.3 (2x CH2), 38.4 (2x CH2), 26.0 (2x
CH2), 24.6 (2x CH2).HRMS (ES+) calculated for C44H66N10O6 [M
+ H]+ 831.5203, found 831.5203.
Compound 3
1H NMR (CDCl3 400MHz)δ 9.42 (2H, t, J=5.5Hz, 2x NH),
8.17(2H, s, ArH), 4.35(4H, t, J=7.00Hz, 2x CH2), 3.76 (8H, t,
J=5.5Hz, 4x CH2), 3.59 (4H, dd, J=12.5,6.6Hz, 2x CH2) 2.77
(18H, m, 9x CH2), 2.52 (20H, m, 2x CH3, 7 x CH2) 2.06-1.97(4H,
m, 2x CH2). 13C NMR (CDCl3100MHz) δ 166.0 (2x quat), 163.1
(2x quat), 149.3 (2x quat), 125.8 (2x quat), 121.3 (2x quat), 118.5
(2x quat), 101.9 (2x quat),66.8 (4x CH2), 56.2 (2x CH3), 55.2 (2x
CH2), 53.9 (4x CH2), 53.8 (2x CH2), 51.4 (2x CH2), 44.2 (2x
CH2), 41.3 (2x CH2), 37.3 (2x CH2), 34.4 (2x CH), 26.2 (2x
CH2). HRMS (ES+) calculated forC42H62N10O6 [M+H]+ 803.4934,
found 803.4940.
Compound 4
4.3 Cell culture
1H NMR (CDCl3 400MHz)δ 9.54(2H, t, J=5.3Hz, 2x NH),
8.11(2H, s, ArH), 4.30-4.23 (6H, m, 3x CH2), 4.04-3.99 (2H, m,
2x CH), 3.91-3.85 (2H, m, 2x CH), 3.71-3.66 (2H, m, 2x CH),
3.56-3.50 (2H, m, 2x CH), 2.80-2.50 (14H, m, 7x CH2), 2.61(4H,
t, J=7.1Hz, 2x CH2), 2.50 (6H, s, 2x NMe), 2.21-2.13 (2H, m, 2x
CH), 2.07-1.94 (8H, m, 4x CH2), 1.80-1.72(2H, m, 2x CH). 13C
NMR (CDCl3 100MHz)δ 166.0 (2x quat) , 163.0 (2x quat), 149.2
(2x quat), 125.3 (2x quat), 121.2 (2x quat), 118.4 (2x quat), 102.0
(2x quat), 68.5 (2x CH2), 55.2 (2x CH3), 53.4 (2x CH2), 50.1 (4x
CH2), 47.3 (2x CH2), 43.9 (2x CH2), 38.5 (2x CH), 29.3 (2x
CH2), 25.9 (4x CH2), 24.6 (2x CH2).HRMS (ES+) calculated for
C44H64N8O6 [ M + H]+745.4378, found 745.4401.
The cell lines MCF7, A549, MIA PaCa2, PANC1, RCC4,
786-O and WI38, purchased from American Type Cell Culture
(ATCC), were maintained in monolayer culture in 75 cm2 flasks
(TPP, Switzerland) under a humidified 5% CO2 atmosphere at 37
°C. Incubations were also done under these conditions, unless
specified otherwise. For the cell lines MCF7 and A549,
Dulbecco’s MEM medium was used (GIBCO 21969, Invitrogen,
UK) supplemented with L-glutamine (2 mM, GIBCO 25030,
Invitrogen, UK), essential amino acids (1 %, GIBCO 11140,
Invitrogen, UK), foetal calf serum (10 %, S1810, Biosera, UK)
and hydrocortisone (0.5 μg/mL, Acros Organics, UK). For MIA
PaCa2 and PANC1 cells, Dulbecco’s MEM medium was also
used, supplemented with L-glutamine (2 mM) and foetal calf
serum (10 %). MEM medium (M2279, Sigma, UK) with added
L-glutamine (2 mM), essential amino acids (1 %) and foetal calf
serum (10 %) was used for the WI-38cell line. RPMI
medium1640 (GIBCO 31870, Invitrogen, UK), supplemented
with L-glutamine (2 mM), foetal calf serum (10 %) and sodium
Compound 5
1H NMR (CDCl3, 400MHz) δ 9.46 (2H, t, J= 5.6Hz, 2x NH),
8.13 (2H, s, 2x ArH), 4.23 (4H,t, J= 7.3Hz, 2x CH2), 4.04 (4H,
dd, J= 11.5, 3.4 Hz,2x CH2), 3.48-3.41 (8H, m, 4x CH2), 2.87-
2.79 (8H, m, 4x CH2), 2.64 (8H, m, 4x CH2), 2.55(4H, t, J=
7.2Hz, 2x CH2), 1.92 (6H, s, NMe), 1.96-1.87 (6H, m, 2 x CH3),
1.55 (4H, d, J= 12.8 Hz,2x CH2), 1.45 (4H, m, 2x CH2).13C NMR

7
pyruvate (1 mM, 11360-039, Invitrogen, UK) was used for the
786-Ocell line. To passage the cells, they were washed with PBS
(GIBCO 14040, Invitrogen, UK), treated with trypsine (GIBCO
25300, Invitrogen, UK), and re-seeded into fresh medium,
resulting in an initial cell density of approximately 1x104 cells/
other parameters for the small-molecule fragments were obtained
via the ACPYPE tool, employing the ANTECHAMBER module
of the AMBER32 program with the GAFF force field. All MD
protocols were kept identical for consistency of the results.
Explicit solvent simulations were performed at T=300K with a
time constant for coupling of 0.1 ps under the control of a
velocity rescaling thermostat, and isotropic constant-pressure
boundary conditions controlled by the Parinello-Rahman
algorithm of pressure coupling. Long-range electrostatics were
calculated using the PME algorithm with grid spacing of 1.17 Å,
and the LINCS algorithm was employed to constrain all bonds.
Non-bonded van der Waals interactions were treated in terms of
Lennard-Jones 12-6 potential with a 10.0 Å cutoff. The solute
was soaked in a triclinic box of TIP3P water with a minimal
clearance of 20.0 Å between periodic images for the starting
configurations. Additionally, positively-charged K+ counter-ions
were included in the systems to neutralize the negative net-charge
on the DNA backbone. In each of the MD runs, there were two
temperature-coupling groups; DNA with the structural K+ ions
(and ligand, when present), and water with counter-ions. The
systems were then subjected to 5,000 steps of potential energy
minimization, followed by 200 ps of molecular dynamics at
200K while keeping the solutes constrained, and a further 100 ps
of MD during which the systems were slowly heated to 300K and
further equilibrated prior to unconstrained 50 ns production-level
MD trajectory calculations. The time-step applied was 2.0 fs with
coordinates saved every 5.0 ps. All MD simulations were
computed on in-house Linux 64-bit Intel Core-i7 workstations,
with efficient parallel scaling and double-precision calculations
to prevent any energy conservation and stability issues.
Trajectories were analyzed employing the programs in the
GROMACS 4.5.3 suite package and visualized by means of the
VMD33 and PyMol34 programs. The initial 2 ns were rejected for
the analysis of the MD trajectories.
mL medium. Cells were counted using
a
Neubauer
haemocytometer (Assistant, Germany) by microscopy on a
suspension of cells obtained by washing with PBS, trypsinisation,
centrifugation at 8 °C at 8000 rpm for 3 min, and re-suspension
in fresh medium.
4.4 Sulforhodamine B (SRB) assay
Cells were counted and diluted to the required concentration in
20 mL medium. For the MCF7, A549, MIA PaCa2 and PANC1
cell lines, 1000 - 4000 cells with 160 μL media were seeded into
each well of a 96 well plate (Nunc, Denmark). After incubation
for 24 hrs, the compounds to be tested, dissolved in 40 μL of
medium were added in a range of concentrations, and the cells
incubated for 96 hrs. The medium was then removed and the
cells fixed by incubation with TCA (10 %, Sigma-Aldrich, UK)
in water for 30 min at 4 °C. After removal of the TCA, the cells
were washed with deionised water five times and dried at 60 °C
for 1 hr. The cells were then incubated with sulforhodamine B
(80 μL, 0.4 % in 1 % acetic acid, Acros Organics, UK) for 15
min at RT. The SRB was removed, the wells washed with 1 %
acetic acid (200 μL), and dried at 60 °C for 1 hr. Tris-base (100
μL, 10 mM, Acros Organics, UK) solution was added to each
well, and the plates were gently shaken for 5 min. The
absorbance at 540 nm was measured with a plate reader
(Spectrostar Omega, BMG Labtech, Germany). The data were
normalized to the value of 100 for the control experiment
(untreated cells), and the IC50 values were obtained by
interpolation from a plot with Origin (Version 7.0, OriginLab
Corp.), as the concentration required for a reduction in
absorbance intensity of 50%.
Cluster analysis methods previously described35 were applied
here, to identify clusters of structures in a trajectory by means of
the GROMOS agglomerative clustering algorithm (as
implemented in the GROMACS clustering utility), with a rmsd
cutoff distance of 2.0 Å applied for two structures considered to
be neighbors.
4.5 Computational studies
The crystal structure25 of an intramolecular human telomeric
DNA G-quadruplex complexed with the naphthalene diimide
compound 2a (pdb id 3UYH; 1.95 Å resolution) was used as a
starting point for the two quadruplex-ligand complex molecular
dynamics simulations. (Although reported as a 22-mer the
flanking 5' terminal residue was not clearly observed in the
crystal structure, hence the truncated 21-mer was used here).
Consecutive K+ ions vertically aligned within the central core of
the quadruplex mid-way between each G-quartet were retained at
their respective crystallographic positions, while the K+ ion
outside of the central core of the G-tetrad (at the 5'-site) was
removed. The ligand, compound 2a, was located at the 3'-
terminal G-quartet site of the 21-mer G-quadruplex.
The structure of the positional isomer, compound 2, was
obtained by interchanging the side-chains using the Discovery
Studio Visualizer program(www.accelrys.com). The overall net
charge of both ligands was assigned as +2, and geometry was
optimized by a short cycle of energy minimization employing the
fast Dreiding force field within the Accelrys Discovery Studio.
In total, three 50 ns molecular dynamics (MD) simulations were
performed; (1) on the 21-mer native quadruplex29 (as a
reference), (2) and (3) on 21-mer complexes of 2 and 2a (with the
ligand binding poses in both cases corresponding to the crystal
structure). All full-atom simulations were performed with the
GROMACS v 4.5.3 program,30 employing the parmbsc0 force
field31 previously ported into GROMACS. The topologies and
The MM/PB(GB)SA method36 was used to compute the
relative free energies of binding, employing the thermodynamic
cycle that combines molecular mechanical (MM) energies with
the implicit solvent methods. This method takes advantage of
multiple snapshots from a trajectory, to provide an average of
energies. The change in free energy for the molecules upon
complex formation was calculated (for each of the snapshots) as
the difference in free energy between their bound and unbound
states, as given in equation 1, free energy of each term was then
calculated according to equation 2:
∆Gbind = G(G4-LIG complex) - G(G4) - G(LIG)
∆G = ∆EMM + ∆GSOL - T∆S
1
2
where the molecular mechanics energy (EMM) term is a sum of
the internal energy (bonds, angles and dihedrals), electrostatic
energy and van der Waals term; while the GSOL term accounts for
the solvation energy, comprising both polar and nonpolar
component. The polar part of the solvation term accounts for the
electrostatic contribution to solvation and was calculated using
both the Poisson-Boltzmann (PB) model, and the generalized-
Born (GB) model37 (igb=2; model GBOBC1) with rescaled
effective Born radii, accounting for interstitional spaces between
atom spheres. Solvent probe radius, dielectric constants grid-
spacing parameters were kept at default values. The entropy term

8
(T∆S) was not included in our simulations as different
conformations and binding poses of a ligand were explored. The
free-energy calculations on both quadruplex complexes were
performed employing the single-trajectory approach over 48 ns,
representing 1920 frames of the simulation.
Acknowledgments and notes
Sheila Mpima and Stephan A. Ohnmacht contributed equally to this
work. We are grateful for support to Cancer Research UK (Program
Grant No. C129/A4489 to SN), and the Pancreatic Cancer Research
Fund (project grant to SN).
20. Gunaratnam, M.; Swank, S.; Haider, S. M.; Galesa, K.; Reszka, A.
P.; Beltran, M.; Cuenca, F.; Fletcher, J. A.; Neidle, S. J. Med.
Chem. 2009, 52, 3774-3783.
21. Hampel, S. M.; Sidibe, A.; Gunaratnam, M.; Riou, J.-F.; Neidle, S.
Bioorg. Med. Chem. Lett. 2010, 20, 6459-6463.
References
1. Burge, S.; Parkinson, G. N.; Hazel, P.; Todd, A. K.; Neidle, S.
Nucleic Acids Res. 2006, 34, 5402-5415.
2. Phan, A. T. FEBS J. 2010, 277, 1107-1117.
3. Bochman, M. L.; Paeschke, K.; Zakian, V. A. Nat. Rev. Genetics
2012, 13, 770-780.
4. De Cian, A..; Lacroix, L..; Douarre, C.; Temime-Smaali, N.;
Trentesaux, C., Riou, J.-F.; Mergny, J.-L. Biochimie 2008, 90,
131-155.
5. Balasubramanian, S.; Hurley, L. H.; Neidle, S. Nat. Rev. Drug
Discovery 2011, 10, 261-275.
22. Gunaratnam, M.; de la Fuente, M.; Hampel, S. M.; Todd, A. K.;
Reszka, A. P.; Schatzlein, A.; Neidle, S. Bioorg. Med. Chem.
2011, 19, 7151-7157.
23. Collie, G. W.; Promontorio, R.; Hampel, S. M.; Micco, M.;
Neidle, S.; Parkinson, G. N. J. Am. Chem. Soc. 2012, 134, 2723-
2731.
24. Hampel, S. M.; Pepe, A.; Greulich-Bode, K. M.; Malhotra, S. V.;
Reszka, A. P.; Veith, S.; Boukamp, P.; Neidle, S. Molec.
Pharmacol. 2013, 83, 470-480.
25. Micco, M.; Collie, G. W.; Dale, A. G.; Ohnmacht, S. A.; Pazitna,
I.; Gunaratnam, M.; Reszka, A. P.; Neidle, S. J. Med. Chem. 2013,
DOI: 10.1021/jm301899y.
6. Müller, S.; Kumari, S.; Rodriguez, R.; Balasubramanian S. Nat.
Chem. 2010, 2, 1095-1098.
7.
Biffi, G.; Tannahill, D; McCafferty, J.; Balasubramanian, S. Nat.
Chem. 2013, 5, 182-186.
26. Ohnmacht, S. A.; Micco, M.; Petrucci, V.; Todd, A. K.; Reszka,
A. P.; Gunaratnam, M.; Carvalho, M. A.; Zloh, M.; Neidle, S.
Bioorg. Med. Chem. Lett.2012, 22, 5930-5935.
27. Guyen, B.; Schultes, C. M.; Hazel, P.; Mann J.; Neidle, S. Org.
Biomol. Chem .2004, 2, 981-988.
28. Fryer, R. A.; Barlett, B.; Galustian, C.; Dalgleish, A. G.
Anticancer Res. 2011, 31, 3747-3756.
29. Parkinson, G. N.; Lee, M. P. H.; Neidle, S. Nature 2002, 417, 876-
880.
30. Pronk, S.; Páll, S.; Schulz, R.; Larsson, P.; Bjelkmar, P.;
Apostolov, R.; Shirts, M. R.; Smith, J. C.; Kasson, P. M.; van der
Spoel, D.; Hess, B.;Lindahl, E. Bioinformatics 2013, 29, 845-854.
31. Pérez, A.; Marchán, I.; Svozil, D.; Sponer, J.; Cheatham, T. E.;
Laughton, C. A.; Orozco, M. Biophysical J. 2007, 92, 3817-3829.
32. Case, D. A.; Darden, T. A.; Cheatham, T. E. III.; Simmerling, C.
L; Wang, J.; Duke, R. E.; Luo, R.; Walker, R .C.; Zhang, W.;
Merz, K. M.; Roberts, B.; Hayik, S.; Roitberg, A.; Seabra, G.;
Swails, J.; Goetz, A. W.; Kolossváry, I.; Wong, K. F.; Paesani, F.;
Vanicek, J.; Wolf, R. M.; Liu, J.; Wu, X.; Brozell, S. R.;
Steinbrecher, T.; Gohlke, H.; Cai, Q.; Ye, X.; Wang, J.; Hsieh,
M.-J.; Cui, G.; Roe, D. R.; Mathews, D. H.; Seetin, M. G.;
Salomon-Ferrer, R.; Sagui, C.; Babin, V.; Luchko, T.; Gusarov,
S.; Kovalenko, A.; Kollman, P. A. (2012), AMBER 12, University
of California, San Francisco, USA.http://ambermd.org
33. Humphrey, W.; Dalke, A.; Schulten, K. J. Mol. Graphics 1996,
14, 33-38.
8. Siddiqui-Jain, A.; Grand, C. L.; Bearss, D. J.; Hurley, L. H. Proc.
Natl. Acad. Sci. USA 2002, 99, 11593-11598.
9. Rodriguez, R.; Miller, K. M.; Forment, J. V.; Bradshaw, C. R.;
Nikan, M.; Britton, S.; Oelschlaegel, T.; Xhemalce B.;
Balasubramanian, S.; Jackson, S. P. Nat. Chem. Biol. 2012, 8,
301-310.
10. Neidle, S. Therapeutic Applications of Quadruplex Nucleic Acids.
Academic Press, San Diego, 2011.
11. (a) Monchaud, D.; Teulade-Fichou, M.-P. Org. Biomol. Chem.
2008, 6, 627-636. (b) Georgiades, S. N.; Abd Karim, N. H.;
Suntharalingam, K.; Vilar, R. Angew. Chem. Int. Ed. Engl. 2010
49, 4020-4034; Ralph, S. F. Curr. Top. Med. Chem. 2011, 11,
572-590.
12. Di Antonio, M.; Doria, F.; Richter, S. N.; Bertipaglia, C.; Mella,
M.; Sissi, C.; Palumbo, M.; Freccero, M. J. Am. Chem. Soc. 2009,
131, 13132-13141
13. Nadai, M.; Doria, F.; Di Antonio, M.; Sattin, G.; Germani, L.;
Percivalle, C.; Palumbo, M.; Richter, S. N.; Freccero, M.
Biochimie 2011, 93, 1328-1340
14. Peduto, A.; Pagano, B.; Petronzi, C.; Massa, A.; Esposito, V.;
Virgilio, A.; Paduano, F.; Trapasso, F.; Fiorito, F.; Florio, S.;
Giancola, C.; Galeone, A.; Filosa, R. Bioorg. Med. Chem. 2011,
19, 6419-6429.
15. McKnight, R. E.; Reisenauer, E.; Pintado, M. V.; Polasani, S. R.;
Dixon, D. W. Bioorg. Med. Chem. Lett. 2011, 21, 4288-4291.
16. Doria, F.; Nadai, M.; Folini, M.; Di Antonio, M.; Germani, L.;
Percivalle, C.; Sissi, C.; Zaffaroni, N.; Alcaro, S.; Artese, A.;
Richter, S. N.; Freccero, M. Org. Biomol. Chem. 2012, 10, 2798-
2806.
17. Doria, F.; Nadai, M.; Sattin, G.; Pasotti, L.; Richter, S. N.;
Freccero, M. Org. Biomol. Chem. 2012, 10, 3830-3840.
18. Cuenca, F.; Greciano, O.; Gunaratnam, M.; Haider, S.; Munnur,
D.; Nanjunda, R.; Wilson, W. D.; Neidle S. Bioorg. Med. Chem.
Lett. 2008, 18, 1668-1673.
34. DeLano, W. L. 2008, The PyMOL Molecular Graphics System.
DeLano
Scientific
LLC,
Palo
Alto,
CA,
USA.
http://www.pymol.org
35. Husby, J.; Todd, A. K.; Haider, S. M.; Zinzalla, G.; Thurston, D.
E.; Neidle, S. J. Chem. Inf. Model. 2012, 52, 1179-1192.
36. Gohlke, H.; Case, D. A. J. Comput. Chem. 2004, 25,238-250.
37. Onufriev, A.; Bashford, D.;Case, D. A. Proteins 2004, 55, 383-
394.
19. Gunaratnam, M.; Cuenca, F.; Neidle S. Naphthalene diimide
compounds interacting with G-quadruplex regions in DNA.
PCT/GB08/51131.

9
Table 1. Tm values (C) for a human telomeric quadruplex sequence (F21T), two HSP90 quadruplex sequences, and a representative
duplex DNA sequence (T loop), at 1 and 2 μM ligand concentration (with 60 mM K+ at pH 7.4). Esds are  0.5C, estimated from
multiple readings. Data for compounds 1a-5a has been taken from reference 25.
Compound
F21T
Hsp90
A
Hsp90
B
T
loop
1μM 2 μM
1 μM
22
29
34
33
33
34
32
21
27
2
2 μM
24
32
36
36
36
36
36
23
30
27
<2
1 μM
22
24
30
29
28
29
31
13
21
3
2 μM 1 μM
2 μM
0
0
9
9
8
1
12
0
0
1
1a
2
2a
3
3a
3b
4
4a
5
<2
<2
29
27
26
27
28
<2
23
<2
<2
16
22
34
32
31
31
33
15
26
22
<2
26
29
34
33
32
34
34
17
28
23
<2
0
0
5
5
2
0
8
0
0
0
0
0
0
5a
<2
<2
Table 2. Growth inhibition data for compounds 1-5 and 1a-5a for a panel of cancer cell lines and a normal human fibroblast line (WI-
38), measured by the SRB assay for 96 h exposure and expressed as IC50 values in μM. Data for compounds 1a-5a has been previously
reported,25 apart from the response in the PANC1 cell line The panel of cancer carcinoma cell lines used includes: A549 (lung), RCC4
and 786-0 (renal), MIA PaCa2 and PANC1(pancreatic), and MCF7 (breast). Data for the two most active compounds are highlighted in
mauve.
Compound
RCC4
786-O
MCF7
MIA
A549
PANC1
WI38
PaCa2
1
1a
2
7.60
±0.01
10.51
±0.14
0.61
±0.07
0.56
16.10
±0.63
7.17
±0.41
0.44
±0.07 ±0.001
0.32 0.070
±0.01 ±0.007
0.16 0.021
17.50
±0.98
5.62
±0.15
0.014
7.05
±0.83
2.79
5.75
±0.51
2.54
±0.01
0.007
±0.001
0.019
2.50
±0.34
n/a
>25
3.32
±0.50
0.30
±0.023
0.23
±0.01
0.18
±0.002
0.61
±0.02
2.46
±0.09
0.050
±0.002
0.010
±0.01
0.005
±0.001
0.04
0.002
±0.001
0.003
2a
3
±0.05
0.30
±0.005 ±0.0005
0.02
±0.001
1.55
±0.02
0.023
±0.006
0.16
±0.002
5.57
±0.87
0.46
±0.01
0.005
±0.001
n/a
±0.016 ±0.021 ±0.001
3a
3b
4
1.75
±0.18
0.28
±0.06
3.00
±0.17
>10
0.63
±0.06
n/a
0.17
±0.01
0.03
±0.01
0.56
±0.01
0.01
n/a
±0.01
0.095
±0.003
5.65
±0.93
0.137
±0.031
2.50
±0.02
2.3
±0.08
>10
0.39
±0.02 ±0.034
>10
1.50
±0.17
n/a
4a
5
>10
2.20
0.29
0.51
1.10
±0.13
n/a
2.8
±0.015 ±0.002 ±0.017
8.38
±0.5
±0.23
12.65
±0.11
5a
1.20
±0.03
3.12
±0.13
2.92
±0.01
±0.01
Table 3. Summary of the simulations and MD simulation statistics together with calculated binding free energies for the lead compound
2 and its positional isomer 2a.
Simulation system
Timescale
RMSD [Â]
Cluster analysis*
ΔGbind [kcal/mol]
G4 vs. initial G4 vs. avg
MM/PBSA
MM/GBSA
-38.72
-37.59
-
G4-2a
G4-2
50 ns
50 ns
50 ns
3.09
2.99
1.98
1.09
1.94
1.24
1 cluster/ 100%
-50.48
-45.91
-
7 clusters/ 84%
-
Native G4 21-mer

10
Figure legends
Scheme. Outline of the synthesis of tetra-substituted naphthalene diimides.
Figure 1. The group of tetra-substituted naphthalene diimide compounds reported and discussed in this study.
Figure 2. Structures of the two G4-ligand complexes, (a) G4-2a, (b) G4-2, and (c) structural alignment of the two complexes prior to
the MD simulations. The DNA structure is shown in cartoon representation, and the ligands, 2 (purple) and 2a (yellow) are in stick
representation. Oxygen atoms are colored red, and nitrogen atoms are blue.
Figure 3. The predominant conformations adopted by the two G4-ligand complexes through the simulation time, with respect to their
starting structures (prior to the MD simulations). (a) The G4-2a complex at the start of the simulation (yellow), and the sole
predominant conformer formed during the simulation (green). (b-d) Show an anti-clockwise rotation of 90° in the G-quartet plane for
compound 2 (cyan) with respect to its starting-structure/conformation (purple). (b) The first 20-25 ns of the simulation time with the
starting rotation highlighted by a red arrow; (c) at ~35 ns, a significant rotation of the ligand occurs, as shown, and (d) from ~45 ns
onwards, ligand rotation is practically complete, with the morpholino side-chains replacing the N-methyl piperazine ones, and vice
versa. (e) Shows the final rotated conformation of compound 2 (cyan) superimposed onto the initial, reference structure of compound
2a (yellow). The cluster representations are always aligned and superimposed onto the starting structures, in order to retain the
perspective of the conformational change. The DNA of the reference (ie initial) structure/conformation is colored orange, while the
individual cluster representations are colored white (G4-DNA only).
Figure 4. Visualization of the differences in side-chain linker length, based on the substitution patterns around the ND scaffold -
suggesting an increased circular shape (as opposed to elliptical) for the most biologically potent analogues, with equal-length side-
chains.
Scheme

11
Figure 1

12
Figure 2
Figure 3

13
Figure 4