A platinum(ii) phenylphenanthroimidazole with an extended side-chain exhibits slow dissociation from a c-kit G-quadruplex motif

Article abstract of DOI:10.1002/chem.201301590

A series of three platinum(II) phenanthroimidazoles each containing a protonable side-chain appended from the phenyl moiety through copper(I)-catalyzed azide-alkyne cycloaddition (CuAAC) were evaluated for their capacities to bind to human telomere, c-Myc, and c-Kit derived G-quadruplexes. The side-chain has been optimized to enable a multivalent binding mode to G-quadruplex motifs, which would potentially result in selective targeting. Molecular modeling, high-throughput fluorescence intercalator displacement (HT-FID) assays, and surface plasmon resonance (SPR) studies demonstrate that complex 2 exhibits significantly slower dissociation rates compared to platinum phenanthroimidazoles without side-chains and other reported G-quadruplex binders. Complex 2 showed little cytotoxicity in HeLa and A172 cancer cell lines, consistent with the fact that it does not follow a telomere-targeting pathway. Preliminary mRNA analysis shows that 2 specifically interacts with the ckit promoter region. Overall, this study validates 2 as a useful molecular probe for c-Kit related cancer pathways. Bound to impress: A series of three platinum(II) phenanthroimidazoles each containing a protonable side-chain appended from the phenyl moiety through copper(I)-catalyzed azide-alkyne cycloaddition (CuAAC), was evaluated for their capacity to bind to human telomere, c-Myc, and c-Kit derived G-quadruplexes (see scheme). The presence of the side-chain enables a multivalent binding mode to G-quadruplex motifs. Copyright

Full text of DOI:10.1002/chem.201301590

DOI: 10.1002/chem.201301590
A Platinum(II) Phenylphenanthroimidazole with an Extended Side-Chain
Exhibits Slow Dissociation from a c-Kit G-Quadruplex Motif
[
a]
[a]
[a]
[b]
Katherine J. Castor, Zhaomin Liu, Johans Fakhoury, Mark A. Hancock,
[
a]
[a]
[a]
Anthony Mittermaier, Nicolas Moitessier, and Hanadi F. Sleiman*
Abstract: A series of three platinu-
m(II) phenanthroimidazoles each con-
taining a protonable side-chain ap-
pended from the phenyl moiety
through copper(I)-catalyzed azide–
alkyne cycloaddition (CuAAC) were
evaluated for their capacities to bind to
human telomere, c-Myc, and c-Kit de-
rived G-quadruplexes. The side-chain
has been optimized to enable a multiva-
lent binding mode to G-quadruplex
motifs, which would potentially result
in selective targeting. Molecular mod-
eling, high-throughput fluorescence in-
phenanthroimidazoles without side-
chains and other reported G-quadru-
plex binders. Complex 2 showed little
cytotoxicity in HeLa and A172 cancer
cell lines, consistent with the fact that
it does not follow a telomere-targeting
pathway. Preliminary mRNA analysis
shows that 2 specifically interacts with
the ckit promoter region. Overall, this
study validates 2 as a useful molecular
probe for c-Kit related cancer path-
ways.
tercalator
displacement
(HT-FID)
assays, and surface plasmon resonance
(SPR) studies demonstrate that com-
plex 2 exhibits significantly slower dis-
sociation rates compared to platinum
Keywords: cancer therapeutic
·
click chemistry · drug design · G-
quadruplexes · platinum(II)
Introduction
rather than all, while minimizing binding to the more abun-
dant genomic duplex DNA.
Guanine quadruplexes, non-canonical secondary structures
formed in guanine (G)-rich DNA have emerged as impor-
Of particular interest for chemotherapeutic targets is the
c-KIT protein, a transmembrane growth factor receptor
with tyrosine kinase activity that belongs to the PDGF/CSF-
[1]
tant targets for selective antitumor therapy. These higher-
order DNA structures result from the Hoogsteen hydrogen
bonding of the guanine units of G-rich DNA sequences into
stacked tetramers stabilized by monovalent cations, namely
[8]
1 family. The c-KIT protein, encoded by the human proto-
oncogene c-Kit stimulates cell proliferation and survival
[9]
when activated by its ligand stem cell factor. Inhibition of
+
+
[2]
[10]
K
and Na .
G-quadruplex-forming sequences exist
this protein is an important target in cancer therapeutics
[3]
throughout the human genome and recent studies have
shown that stabilizing these structures with small molecules
can bring about antitumor effects by several mechanisms.
These mechanisms include the inhibition of the cancer cell-
immortalizing enzyme telomerase by sequestration of the te-
since mutated forms of c-KIT are expressed in a variety of
[11]
cancers including gastrointestinal tumors (GIST), systemic
[12]
mastocytosis, subsets of acute and chronic myeloid leuke-
[13]
[14]
mia
and melanoma.
Although c-KIT inhibition has
been realized with the small molecule inhibitor imatinib
[4]
[15]
lomere substrate and the regulation of gene expression,
(Gleevec) in GIST tumors, drug resistance has developed
[5]
[16]
particularly in cancer-related genes including c-Myc, c-
in many cases due to secondary mutations.
Even when
[
6]
[7]
Kit, and vascular endothelial growth factor (VEGF)
a secondary drug, such as sunitinib, has been used to combat
imatinib-resistant mutations, studies have found that: 1) the
drug–drug interactions cause greater toxicity, or 2) the
tumors often become quiescent and can resume prolifera-
among others. Since different DNA sequences create struc-
tural diversity among G-quadruplex polymorphs, an impor-
tant therapeutic goal is to selectively target one polymorph
[17]
tion when treatment is stopped.
For these reasons, new
molecular probes for the c-KIT-based pathway need to be
developed.
One such class of probes consists of molecules that regu-
late the expression of the oncogene coding for c-KIT rather
than inhibiting the protein itself. In this manner, the prob-
lems of drug resistance and toxic combination therapies
could be avoided since the protein would not be present in
the cells. Such an approach has been explored with the c-
[
a] K. J. Castor, Z. Liu, Dr. J. Fakhoury, Prof. A. Mittermaier,
Prof. N. Moitessier, Prof. H. F. Sleiman
McGill University Department of Chemistry
8
01 Sherbrooke West, Montreal, Quebec, H3A 0B8 (Canada)
E-mail: hanadi.sleiman@mcgill.ca
[
b] Dr. M. A. Hancock
McGill University SPR Facility, 3773 University Street
Montreal, Quebec, H3A 2B4 (Canada)
[5b,c,e,f,18]
Myc oncogene,
in which case drugs were used to
induce and stabilize G-quadruplex DNA motifs within the
promoter nuclease hypersensitivity element (NHE) III1—
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201301590.
17836
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 17836 – 17845

FULL PAPER
a G-rich region that is responsible for up to 90% of c-Myc
transcription. Analogous regulation of c-Kit is plausible,
action with the c-Kit promoter region, however, this needs
to be further studied with cell lines that over-express the c-
KIT protein. Thus, complex 2 is a useful probe for examin-
ing the effects of increasing the selectivity and residence
times for the c-Kit promoter on gene expression, and the po-
tential use of these attributes for effective antitumor thera-
pies.
[19]
given that two distinct G-quadruplexes can form in the pro-
[10,20]
moter region of the c-Kit oncogene.
Indeed, three
classes of small molecules have been shown to down-regu-
[6]
late c-KIT expression in GIST cells.
When attempting to influence biological processes like
DNA transcription, many ligands can benefit from slow dis-
sociation rates once they complex with their targets. One
well-characterized example is actinomycin, an effective sup-
pressor of transcription which forms a stable complex with
double-stranded DNA to inhibit RNA synthesis by RNA
Results and Discussion
Synthesis of platinum(II) phenanthroimidazoles: Using our
[21]
polymerase.
In an analogous manner, one can envision
previously established platinum ethylenediamine phenan-
[24,25]
that regulation of c-Kit transcription can be accomplished
with a molecule that effectively complexes with G-quadru-
plex DNA and exhibits slow dissociation (i.e., multivalent
interactions including intercalation, p-stacking, groove bind-
ing, conformational change, etc.).
throimidazole core,
we now appended side-chains from
the phenyl moiety. We strategically chose alkylamino side-
chains since these and similar substituents have been shown
to increase binding strength to G-quadruplexes (via interac-
[26]
tion with the grooves, loops, and backbone). In addition,
they each contain a biocompatible 1,4-substituted 1,2,3-tria-
Although the rational design of DNA-based inhibitors has
been successful in the creation of many duplex bisintercala-
[27]
zole moiety,
a synthetic building block that has been
[
22]
tors,
e.g., large p-surface area, major and minor grooves, loops,
G-quadruplexes have topologically distinct features
shown to help discriminate between G-quadruplex grooves
[28]
(
as well as increase binding strength. The synthesis is facile
and starts from economical, commercially available starting
materials. The side-chains were synthesized by a simple sub-
stitution reaction or amide-bond formation (where appropri-
ate) followed by a Boc-deprotection before azidation. The
azides were then coupled by copper-catalyzed cycloaddition
with 4-ethynylbenzaldehyde under microwave irradiation
for 15 min. The phenanthroimidazole units were synthesized
through the condensation of 1,10-phenanthroline-5,6-dione
and the side-chain aldehydes. Subsequently, these ligands
were treated with potassium tetrachloroplatinate, to yield
etc.) that allow them to be differentiated from duplex DNA.
Of the numerous G-quadruplex binders reported in the liter-
[23]
ature, two characteristics likely account for their increased
affinity for quadruplex over duplex DNA. First, the scaffolds
are electron-poor, predominately through the use of an aro-
matic p-surface, and preferably large enough to hinder the
moleculesꢁ abilities to intercalate into duplex DNA. Second,
the incorporation of protonable or positively charged side-
chains allows the ligands to penetrate into the grooves of
the quadruplexes and to interact with the loops. To date,
however, little is known about the residence time of G-
quadruplex binders with biologically relevant promoter G-
quadruplex motifs.
[Pt(phenanthroimidazole)Cl ], followed by treatment with
2
ethylenediamine (en) and precipitation with ammonium
hexafluorophosphate,
resulting
in
complexes
1–3
We have previously developed a progression of platinu-
m(II)-based phenanthroimidazole complexes capable of
(Scheme 1). The complexes (as well as intermediates) were
1
13
characterized by H and C NMR spectroscopy, and HR-
MS. In the case of 2, the final complex was purified by
HPLC for cell-based assays.
[24]
binding to both tetramolecular and human-telomere de-
[25]
rived intramolecular G-quadruplexes.
We have shown
that a simple modification to the phenanthroimidazole scaf-
fold significantly increases binding to the basket-polymorph
of human telomeric (h-telo) G-quadruplexes. Herein, we
present the design, synthesis, and evaluation (biophysical
and biological) of a series of platinum(II) ethylenediamine
phenylphenanthroimidazoles specifically engineered for
Oligonucleotides: To evaluate the binding specificity of each
complex for different G-quadruplex polymorphs, we chose
three G-quadruplex DNA sequences for use in a high-
throughput fluorescence intercalator displacement (HT-
FID) and circular dichroism (CD) experiments: 22AG cor-
responds to the h-telo repeat and is polymorphic depending
on whether sodium or potassium ions are in solution (denot-
ed hereafter as 22AG-Na and 22AG-K, respectively); cmyc
and ckit1 correspond to human proto-oncogene promoter
sequences (Figure 1). A duplex control sequence, ds26-mer,
was obtained with a self-complementary sequence. For sur-
face plasmon resonance (SPR) experiments, we used biotin-
ylated dT versions of 22AG, cmyc, and ckit1 with a 4-thy-
mine spacer between the 5’-biotin-dT and the sequence
start; a CG-rich hairpin (HP) was also used as a duplex con-
trol (see Table S1 in the Supporting Information for se-
quence details).
[21]
slow dissociation from target G-quadruplexes (Scheme 1).
One of the complexes highlighted in the present study, 2, ex-
hibited high selectivity for one of the G-quadruplexes found
in the c-Kit promoter (ckit1). We modeled this phenomenon
and rationalized slow dissociation due to chelation/multiva-
lent interactions spread across the entire molecule, and we
demonstrated that the slow dissociation of 2 from ckit1 is
a major component of its binding mechanism. Complex 2
does not cause cytotoxicity at high concentrations in two te-
lomerase-positive cell lines, consistent with the fact that it
does not follow a telomere-targeting pathway. Preliminary
mRNA analysis in A172 cells gave evidence of specific inter-
Chem. Eur. J. 2013, 19, 17836 – 17845
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
17837

H. F. Sleiman et al.
Scheme 1. Synthesis of 1–3; a series of platinum(II) ethylenediamine phenylphenanthroimidazoles with side-chains appended from the phenyl moiety
through copper-catalyzed azide–alkyne cycloaddition (CuAAC). Conditions: a) NaH/THF, RT; b) TFA/DCM, RT, 1 h; c) EDC/HOAt in DMF, RT;
d) HCl/ethyl acetate; e) NaNO
2
and NaN
3
in 17% HCl/EtOH; f) CuSO4//ascorbate in DCM/ACN, microwave, 15 min, 1008C; g) NH
PF
4
OAc in glacial
acetic acid, reflux, 24 h; h) K PtCl
2
4
in H O/DMSO; i) ethylenediamine in EtOH, then NH
2
4
6
.
Molecular modeling: Due to precedents set by the bisin-
To investigate the binding mode of 2 on the G-quadruplex
in greater detail, we used a combination of molecular dock-
ing and molecular dynamics (MD) simulations; the former
allowed a large search of the conformational space, whereas
the latter allowed the highly flexible G-quadruplex system
to freely interact with 2 and account for the effect of water
[22f,29]
tercalators of Nordꢂn and co-workers
and the threading
[21]
nature of actinomycin, we hypothesized that the presence
of the side-chain would enable a multivalent binding mode
and subsequently, longer residence times on the G-quadru-
plex.
17838
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 17836 – 17845

Platinum(II) Phenylphenanthroimidazole
FULL PAPER
Figure 2. Snapshots of the MD simulations of 2 and G-quadruplexes.
+
+
Figure 1. Schematic drawings of the DNA G-quadruplex motifs formed
a,b) 22AG in Na and K crystal, respectively; c) cmyc, and d) ckit1.
+
II
by the chosen sequences for: a) human telomeric basket in Na solution
Complex 2 is shown in cyan with the Pt atom shown as an orange ball.
+
+
+
(
22AG-Na); b) human telomeric 3+1 hybrid in K solution (22AG-K);
G-quadruplex structures are in gray, native metal atoms Na and K are
shown in blue and purple balls, respectively. The backbone phosphates
involved in hydrogen bonds with 2 are shown.
c) c-Myc promoter (cmyc) and; d) c-Kit promoter (ckit1). The gray and
white parallelograms represent syn and anti orientations, respectively, of
the guanine bases contained within the tetrads.
[
30]
and counter ions. This previously reported protocol began
with the generation of a large number of potential binding
modes (referred to as poses) through the use of our in-
slow process of binding in vitro, the kinetics of ligand bind-
ing cannot be quantified. Our efforts to predict the free
energy of binding through MD simulations combined with
MM/PBSA have not yet been conclusive since accurately
computing the entropy contribution to these binding pro-
cesses is challenging. Thus, we should be cautious when re-
lating predictions with the in vitro binding assays.
[31]
house FITTED docking program. We selected a few rep-
resentative poses through clustering and visual inspection as
well as FITTED energy and subjected those poses to MD
simulations using AMBER.
The platinum complexes 1–3 were designed such that mul-
tiple interactions across the G-quadruplex surface would si-
multaneously take place. As shown in Figure 2, our molecu-
lar modeling study suggested that multiple interactions are
likely occurring. For 2, the large aromatic phenanthroimida-
zole unit as well as the triazole and phenyl rings, are p-
stacked with two guanine bases of the top G-tetrad layer.
This stacking orients the platinum atom near the phosphate
backbone and enables electrostatic interactions with these
negatively charged phosphate groups. In addition, strong
ionic hydrogen bonds are predicted to form between phos-
phate groups and both ends of the platinum complex, that
is, between the ethylenediamine ligand and the ammonium
group of the side-arm. Although the ionic hydrogen bond
between the side-arm and phosphate group is enthalpically
favored, it is entropically disfavored. In fact, this hydrogen
bond is fairly labile throughout MD simulations. However,
this flexibility enables the formation of such ionic hydrogen
bonds with multiple phosphate groups in the neighborhood
HT-FID assay: Given that our modeling predictions sup-
ported our multivalent binding hypothesis, we wanted to
assess how this would affect binding affinity. Initially, the
binding of 1–3 (and S4 [(bpy)Pten], as a negative control) to
the h-telo, cmyc, and ckit1 G-quadruplexes was assessed by
a
fluorescence intercalator displacement assay (FID),
a robust and versatile method developed by Teulade-Fichou
[32]
and co-workers. The assay relies upon the competitive dis-
placement of the intercalated thiazole orange dye from the
DNA targets by 1–3, with a concomitant decrease in fluores-
cence intensity. We adapted the high-throughput FID (HT-
[33]
FID) method so that it could be performed by a Biomek
FX liquid handler and a SAGIAN core robot (Beckman;
see Figures S6 and S7 in the Supporting Information for
plate layout). Binding affinities were evaluated from the re-
sulting isotherms where the concentration of platinum com-
plex required to give a 50% decrease in dye fluorescence is
the DC50 (Table 1).
(
Figure 2 vs. Figures S2–S5 in the Supporting Information).
FID assays are typically incubated with short 5–10 min pe-
riods before the plates are read as the displacement and
The G-quadruplex residues involved in the binding vary
among structures. Although the G-quadruplexes may fold
differently with the first G-tetrad layer being covered by
a loop (Figure 2a), or fully (Figure 2b) or partially (Fig-
ure 2c, d) exposed to the aqueous medium, the predicted
binding modes of 2 always retain a similar interaction pat-
tern. Since the simulation timeframe is short as compared to
[33]
binding process is believed to be fast. However, we hy-
pothesized that the displacement and binding process need
longer incubation periods due to the addition of the side-
chains on 1–3. Using both short (10 min) and long (90 min)
incubation periods, we determined DC50 values (Table 1).
The 90 min incubation assay resulted in lower DC50 values
Chem. Eur. J. 2013, 19, 17836 – 17845
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
17839

H. F. Sleiman et al.
Table 1. DC50 values [mm] for 1–3 (10 min/90 min incubations), as deter-
mined by HT-FID assay.
plexes (see Figure S10 in the Supporting Information for
structures) could specifically bind to the immobilized DNA
sequences, whereas diluent alone (i.e., running buffer con-
[
a,b]
Sequence
Complex 1
Complex 2
Complex 3
ꢀ
1
2
2
2AG-Na
2AG-K
1.45/1.09
1.50/1.14
1.74/0.645
>2.50/0.544
2.12/1.76
0.931/0.663
1.29/0.890
1.76/1.32
>2.50/0.282
2.06/1.68
>2.50/2.00
>2.50/1.75
>2.50/1.01
>2.50/0.985
>2.50/2.24
taining 1 mgmL CM dextran to reduce minor nonspecific
binding of the complexes to SA-coated sensors) or our nega-
tive control compound ([(bpy)Pten]; Figure S10 in the Sup-
cmyc
ckit1
ds26-mer
[25]
porting Information (S4), described elsewhere ) generated
no signal responses (Figure 3a–c). With matching DNA sur-
face densities and equimolar injections of complex, binding
to duplex DNA was approximately 50% of quadruplex re-
sponses (i.e., different number of binding sites). In both
cases, all complexes bound with fast association kinetics and
signal responses exceeded theoretical Rmax predictions (i.e.,
stoichiometry greater than 1:1, similar to previous reports in
[
[
a] Averages of four repeats performed by the robot with errors ꢂ5%.
b] As a negative control, our 10 min and 90 min incubations (present
[
25]
study) with S4 yielded negligible binding to all sequences (data not
shown).
across the board with the most dramatic difference involving
2
and the ckit1 G-quadruplex (see Figure S8 in the Support-
[34a]
ing Information for binding isotherms). Notably, this DC50
of 0.282 mm is the lowest value that we have achieved with
our series of platinum phenanthroimidazole complexes to
date.
It is difficult to compare the DC50 values directly between
sequences since the binding affinity of the thiazole orange
dye for each G-quadruplex motif varies greatly. Thus, we de-
veloped a method to compare data sets according to overall
the literature ). Most importantly, the non-sidechain-con-
taining complexes (Figure S10 in the Supporting Informa-
tion) exhibited rapid dissociation kinetics compared to the
significantly slower dissociation rate for 2 (Figure 3d–f).
Using lower density surfaces and increased injection
times, a second set of SPR titrations revealed saturable,
dose-dependent binding of 2 to duplex (data not shown) and
quadruplex DNA (Figure 3d–f). Although the fast-on, slow-
off kinetics deviated from a simple 1:1 binding model in all
cases, affinities were determined based upon equilibrium re-
sponses in each titration series (Figure S11 in the Supporting
Information). On average, the binding of 2 to the quadru-
affinity. Equilibrium association constants (K ) for the com-
A
plexes were extracted from the FID fluorescence isotherms,
using the concentrations and affinity of thiazole orange for
[33]
the different G-quadruplexes (see the Experimental Sec-
tion for details).
plex sequences (K =0.58ꢁ0.1 mm) was approximately ten-
D
After fitting of the data to Equation (7) shown in the Ex-
perimental Section, and determining the corresponding
equilibrium dissociation constants (K =1/K ), we obtained
fold stronger compared to duplex DNA (K =5.0ꢁ1.0 mm,
D
data not shown). Although submicromolar quadruplex bind-
ꢀ
2
ꢀ1
ers with dissociation rate constants in the 10
s
range
D
A
the values in Table 2. Notably, the 0.282 mm DC50 for 2 with
ckit1 translates into a KD of approximately 73 nm. Complex
have been previously reported (e.g., 2Mn-TMPyP+h-
[34a]
[34b]
telo;
telo;
TMPyP4+ckit2;
Se2SAP+cmyc=0.62 mm KD and 2.9ꢃ10
quinazoline derivative+h-
[
34d]
ꢀ3 ꢀ1
[35]
2
also demonstrates the best selectivity between G-quadru-
s
k ;
d
s
ꢀ
3
ꢀ1
plex DNA and duplex DNA compared to the other two
complexes, and is on par with the best reported ckit1 bind-
corrole 4+G4 quadruplex=0.52 mm K and 2.2ꢃ10
D
[36]
kd ), 2 exhibits similar overall affinity; but the dissociation
[
6,28e]
ꢀ3 ꢀ1
ꢀ3 ꢀ1
ers.
rates (22AG-K~1.7ꢃ10 s ; cmyc~1.0ꢃ10 s ; ckit1~
ꢀ
3
ꢀ1
1
.2ꢃ10 s ) are slower than other reported quadruplex
Table 2. Apparent K
determined by HT-FID assay.
D
values [mm] for 1–3 (10 min/90 min incubations), as
[a]
binders. When complex 2 binds to the ckit1 G-quadruplex,
ꢀ
3
ꢀ1
the increased stability of this interaction (i.e., slow 10
s
Sequence
Complex 1
Complex 2
Complex 3
dissociation rate) may translate into desirable pharmacoki-
netics (i.e., a long residence time means the compound oc-
cupies the therapeutic target longer).
2
2
2AG-Na
2AG-K
2.63/0.633
1.49/0.568
1.28/0.102
51.1/0.262
1.30/0.324
1.12/0.366
1.69/0.621
2.30/1.01
1.00/0.0726
8.41/1.10
5.79/4.12
2.42/0.792
4.28/0.134
0.492/0.559
6.47/0.390
cmyc
ckit1
ds26-mer
In examining the PDB files of the different quadruplexes,
a major difference between the structure of ckit1 and the
other polymorphs can be observed (Figure 2). One can see
a “clamp-like” shape of the ckit1 polymorph (Figure 2d)
around the predicted binding pocket of complex 2. The
slow-off binding kinetics of complex 2 to ckit1 may be
a result of the need of this “clamp” to unstack/unbind
before release of 2 from the binding pocket.
[
a] Averages of four repeats performed by the robot with errors ꢂ5%.
SPR assays: To complement our FID data, we performed
label-free, real-time SPR, which has emerged as a powerful
technique to examine the binding specificity and kinetics of
compounds for duplex versus G-quadruplex DNA. Similar
to these previously reported methods, we used a potassium-
containing running buffer to focus upon 1–3 binding to the
[
34]
MTS assay: To assess the antiproliferative activities of 2, it
was evaluated in an MTS [(3-(4,5-dimethylthiazol-2-yl)-5-(3-
carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium)]
2
2AG-K, cmyc, and ckit1 G-quadruplexes; to determine se-
[37]
lectivity between G-quadruplex and duplex DNA, we used
a CG-rich hairpin (HP). Preliminary, fixed-concentration
screening revealed that several of our platinum-based com-
assay that evaluates mitochondrial metabolic activity.
Complex 2 was incubated for 72 h with two cell lines: HeLa
[38]
(non-c-KIT expressing human cervical cancer ) and A172
17840
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 17836 – 17845

Platinum(II) Phenylphenanthroimidazole
FULL PAPER
2
interacts with the c-Kit pro-
moter region and induces
a change in mRNA transcrip-
tion in cellulo. To determine
this, basal levels of c-Kit
mRNA in HeLa and A172 cell
lines were first assessed in the
absence of 2. Basal levels of
mRNA were found only in the
A172 cell line. The cell lines
were incubated with 10 mm of 2,
S4, and S7 (as negative con-
trols) for 48 h and the c-Kit
mRNA levels were measured
relative to 18S ribosomal RNA
(
for internal quantification con-
trol). Figure 4 shows that the
treatment of A172 cells with 2
leads to a slight qualitative in-
crease in c-Kit mRNA at the
examined dose compared to
DMSO and the negative con-
trols.
We speculate that this small
increase could result from one
of two mechanisms; either 1) 2
causes unfolding of the c-Kit
quadruplex motif, or 2) the
binding of 2 to the c-Kit quad-
ruplex causes opening up of
other repressor elements in the
promoter region that were pre-
viously sequestered. To rule out
Figure 3. Representative SPR screening of compounds (5 mm each) binding to 450 RU immobilized: a) 22AG-
K, b) cmyc, and c) duplex DNA at 25 mLmin (5 min association+10 min dissociation). Lines: light gray, near
the x axis, S4 negative control; black, S5 [(3-CLIP)Pten]; medium gray, S6 [(3,5-CLIP)Pten]; dashed, complex
ꢀ1
2
. Representative SPR titrations of 2 (0–5 mm; twofold dilution series) binding to 300 RU immobilized: the first possibility, namely un-
d) 22AG-K, e) cmyc, and f) ckit1 quadruplex DNA at 25 mLmin (10 min association+20 min dissociation). folding of the c-Kit quadruplex
by 2, CD titrations were per-
ꢀ
1
[
8]
(
c-KIT expressing human glioblastoma ). In tandem, we
formed in three different buffer conditions at 378C. If un-
folding of the quadruplex occurred, the positive CD peak at
260 nm signifying the parallel ckit1 folding would decrease
with added 2. Both the 10 and 90 min incubations show neg-
ligible changes in the CD traces in all conditions indicating
also tested a positive control ([(FLIP)Pten]; Figure S13 in
the Supporting Information (S7)) which we had shown to ef-
ficiently bind the h-telo sequence, and a negative control
[25]
(
(
[(bpy)Pten];
Figure S10 in the Supporting Information
S4)). We were unable to determine an IC50 for S4, consis-
tent with its lack of cytotoxicity under our conditions. Com-
plex 2 also had very high IC50 values (>100 mm) for HeLa
and A172 cell lines, whereas S7 had a low mm IC50 value (~
10 mm; Figure S12 in the Supporting Information for images
of cells after incubation). Due to the high IC50 values for
these telomerase-positive cell lines, we hypothesize that 2
does not follow the same telomere-targeting pathway as pos-
[25]
Figure 4. mRNA levels of c-Kit after treatment with the platinum com-
plexes. mRNA levels of c-Kit were assessed in A172 and HeLa cells.
Cells were cultured for 72 h in the presence of S4, 2, and S7 (10 mm).
RNA was collected and reverse transcribed to cDNA. PCR amplification
was performed for c-Kit and GAPDH. Top: A172 cells (glioblastoma) ex-
hibited an increased level of c-Kit mRNA in the presence of 10 mm com-
plex 2. HeLa cells were unaffected and basal level of c-Kit remained un-
changed. Bottom: 18S rRNA was used as a control. Levels of 18S re-
mained unchanged across the board. DMSO was used as the untreated
control.
itive control S7.
within the cell lysate by a simple HPLC detection method
Figure S14 in the Supporting Information).
We confirmed that 2 is indeed found
(
Measurement of mRNA levels: Given that 2 has stronger
binding affinity for ckit1 over h-telo, and due to its absence
of cytotoxicity at high concentrations, we conducted a pre-
liminary semiquantitative mRNA analysis to assess whether
Chem. Eur. J. 2013, 19, 17836 – 17845
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
17841

H. F. Sleiman et al.
that no unfolding of the structure occurs with up to fivefold
excess of 2 (Figure S15 in the Supporting Information). Ad-
ditionally, we conducted CD melting experiments to further
support the stabilization of the ckit1 quadruplex by 2. Com-
plex 2 was incubated overnight, at 48C in a 4:1 ratio of 2/G-
quadruplex and the sample was subsequently melted. The
Tm values were compared to the G-quadruplex alone and
the DTm value was determined to be 7.3ꢁ0.28C. The curve
becomes much broader above 708C, such that two transi-
tions may be evident, and does not fully plateau at 958C
Conclusion
We have generated phenylphenanthroimidazole platinu-
m(II) complexes with side-chains specifically designed to in-
teract with G-quadruplex DNA. This study demonstrates
that these compounds can interact in a multivalent manner
and result in fast-on, slow-off binding kinetics. Spreading
the hydrogen-bond and p-stacking interactions across the
binding surface allows 2 to exhibit, to the best of our knowl-
edge, the slowest reported rate of dissociation from a G-
quadruplex target. We are currently working on elucidating
the mechanism of interaction using cells that over-express
the c-KIT protein, and exploring the use of 2 in combination
therapies. Thus, complex 2 is a useful probe for examining
the effects of increasing the selectivity to the c-Kit promoter
and lengthening the drug residence time on gene expression,
and the potential use of these attributes for effective antitu-
mor therapies.
(
Figure S16 and Table S2 in the Supporting Information).
This is most likely due to the greater stability imparted by 2.
These results lead us to speculate that the binding of 2 in
cellulo may allow for a rearrangement of the promoter
region that exposes enhancer elements or sequesters re-
[39]
pressor elements. One must note here that these prelimi-
nary studies were carried out in A172 cells that only express
low levels of c-Kit mRNA; these results may be different
with cell lines (such as HGC-27) that over-express c-Kit
[40]
mRNA, or with other doses of the platinum complex.
Experimental Section
Western blot analysis: We were also interested to detect
protein levels after incubation with the platinum complexes.
To this end, we used two cell lines, one that expresses the c-
KIT protein, A172, and one that does not, HeLa. We readily
detected the presence of mRNA of c-Kit in A172, however,
as expected we did not observe detectable levels of c-Kit
mRNA in HeLa cells. We were further interested in quanti-
fying the levels of c-KIT protein in A172 cells after being
exposed to the complexes. We thus performed Western blot
analysis; quantification levels relative to the untreated
DMSO control are shown in Figure 5. We observed that
after exposure of A172 to 10 mm 2 for 72 h, levels of c-KIT
were increased, whereas we did not observe any change in
c-KIT levels when cells were exposed to both the untreated
DMSO control and S4. Interestingly, we did see a decrease
when S7 was used (Figure 5). This could possibly be due to
the different mechanisms of action/interaction of S7 and 2
with different areas of the promoter region, as it contains
two guanine-rich tracts capable of forming G-quadruplex-
Molecular modeling: The G-quadruplex structures used herein were re-
trieved from the Protein Data Bank (PDB): the first two structures with
[
42]
[43]
PDB codes 143D and 3T5E were related to 22AG in sodium and po-
[
44]
tassium solutions, respectively; another two structures (1XAV
O3M ) represented the cmyc and ckit1 structure, respectively. Muta-
and
[20a]
2
tions in 1XAV were manually converted back to the wild-type sequence
in order to be consistent with our experiments. The platinum complexes
were optimized at the B3LYP/LCV3P** level of density functional
theory (DFT); atomic charges of the complexes were obtained from ESP
fitting for later use in docking and MD simulations. All DFT computa-
tions were performed with Jaguar 7.0 (Schrodinger LLC). An in-house
[
53]
drug discovery program suite forecaster including prepare, process,
smart and fitted was used to prepare structures and carry out the dock-
ing experiments. In order to probe the entire structure for potential bind-
ing locations, the complete G-quadruplex structure was set as the “bind-
ing site”. fitted uses a matching-algorithm-enhanced genetic algorithm
applied to a set (i.e., population) of conformations. In the present study,
100 docking runs were performed with the initial population of 200 poses
and the matching algorithm turned off. The final docked poses were then
clustered based on their similarity (root-mean-square deviation, RMSD,
below 2 ꢄ). The most likely pose of each cluster was selected as repre-
sentative based on the fitted energy score.
[
10,20b]
es,
in addition to being an inherently complex gene
[45]
The AMBER10 package was used for this MD simulation study. The
[41]
[
46]
promoter. Thus, the interaction of the platinum complexes
with the individual promoter elements will most likely lead
to different responses.
parm99.dat force field was used for G-quadruplex and GAFF parame-
[
47]
II
[30a]
ters (including previously developed Pt force field parameters ) for
the platinum complexes. G-quadruplex structures and side-arm com-
plexes were combined into one system by the LEaP module; additional
+
+
counter ions, such as Na for 143D-complex and K for the other three
systems, were added to neutralize the highly negative systems. The sys-
tems were soaked into periodic truncated octahedron box (10 ꢄ) with
[
48]
TIP3P water molecules. During the following simulations, the Particle
[
49]
Mesh Ewald method was used to treat long-range electrostatic interac-
tions.
The position of water molecules were first optimized by a combination of
steepest descent and conjugate gradient minimization steps with the G-
quadruplex and side-arm complex structures being frozen. The whole
system was then relaxed through additional steepest descent and conju-
gate gradient minimization steps. The temperature of the relaxed systems
was then slowly increased over 20 ps by applying Langevin dynamics
Figure 5. Western blot analysis of A172 cells after 72 h incubation with
ꢀ1
1
0 mm S4, 2, and S7. Comparison of the bands for 2 to DMSO/S4 show
method with collision frequency of 1.0 ps , with a time-step of 1.6 fs and
[
50]
a slight increase in c-KIT levels (1.38 vs. 1/0.93). b-Actin was used as
the SHAKE algorithm
applied to the bonds containing hydrogen
a loading control and was unchanged after incubation with the ligands.
atoms. This was followed by 100 ps equilibration at 300 K performed on
17842
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 17836 – 17845

Platinum(II) Phenylphenanthroimidazole
FULL PAPER
each of the systems at constant pressure and temperature. Data were
then collected over 8 ns. Stability of the systems was assessed by comput-
ing the deviation over the course of the simulations) and representative
binding modes were selected based on the binding energy calculated
mined by solving the third-order polynomial obtained from the equality:
½
Mꢃ
T
¼ ½Mꢃ þ ½Lꢃ
T
f
L
þ ½Xꢃ
T
f
X
ð5Þ
[
51]
from MM/PBSA.
where [L]
T
is the total concentration of ligand and f
L
is the fraction of
HT-FID assay: G-quadruplex DNA sequences 22AG, cmyc, ckit1, and
ds26-mer were diluted from concentrated stock solutions in Milli-Q
water to 250 mm in sodium cacodylate buffer (10 mm; pH 7.4) with
ligand bound to G-quadruplex. This yields:
K
L
½Mꢃ
K
X
½Mꢃ
½
Mꢃ þ ½Lꢃ
T
þ ½Xꢃ
T
ꢀ ½Mꢃ
T
¼ 0
ð6Þ
1
00 mm added KCl (except 22AG, which was evaluated in both the potas-
1
þ K ½Mꢃ
L
1 þ K ½Mꢃ
X
sium containing buffer and the 10 mm sodium cacodylate buffer (pH 7.4)
with 100 mm added NaCl). G-quadruplex sequences were heated to 958C
for 10 min and then rapidly cooled to room temperature. For ds26-mer,
the sequence was heated to 958C and then slowly cooled to room tem-
perature over 7 h. Each sequence was stored overnight, at 48C. Then, the
sequences were diluted to 0.5 mm and 1.0 mm thiazole orange dye in
DMSO was added. In the case of ds26-mer 1.5 mm thiazole orange was
added. All robotics methods were performed on a Biomek FX liquid
handler and a SAGIAN core robot (Beckman) at the CIAN core facility
which can be expanded to give the following polynomial in [M]:
3
2
K
L
K
X
½Mꢃ þ ðK
L
þ K
X
þ K
L
K
X
ð½Lꢃ
T
þ ½Xꢃ
T
ꢀ ½Mꢃ
T
ÞÞ½Mꢃ þ
ð7Þ
ð1 þ K ½LꢃT þ K ½XꢃT ꢀ ðKL þ K Þ½M ꢃÞ½Mꢃ ꢀ ½Mꢃ ¼ 0
L
X
X
T
T
The positive real root of Equation (7) is equal to [M] and was deter-
mined numerically. The total amounts of G-quadruplex, dye, and ligand
([M] , [X] , and [L] ) are known for each step of the titration. The affini-
(
Department of Biology, McGill University). Briefly, buffer and com-
plexes were aspirated from dedicated reservoirs in a two-step procedure
buffer first, then complex for a total volume of 100 mL) into black 96-
T
T
T
ties of the thiazole orange dye (K ) for different G-quadruplex isoforms
X
(
have been experimentally determined by Teulade-Fichou and co-workers.
well plates (Fluoroplates, Nunc). The complexes were tested in quadru-
plicate (that is, two complexes per plate divided between rows D and E)
where the appropriate buffer was first dispensed in decreasing volumes
across the rows (100 to 0 mL), followed by each complex (5 mm stock solu-
tion in appropriate buffer) dispensed in increasing volumes across the
rows (0 to 100 mL). Then, such that the final volume was 200 mL, 100 mL
of 0.5 mm sequence plus thiazole orange was transferred to every well
using the 96-multichannel pipetting head. The final concentrations of
complex from column 1 to column 12 were 0, 0.125, 0.25, 0.375, 0.5,
The values of the ligand binding affinity constant (K ), and the initial
A
and final fluorescence values (I and I , respectively) were varied to mini-
0
1
mize the sum of squared deviations between calculated and experimental
FID data points using a Simplex algorithm (see Figure S9 in the Support-
ing Information for examples of fits).
SPR assay: SPR measurements were performed on research-grade strep-
tavidin-coated (SA) sensor chips (XanTec Bioanalytics GmbH, Muenster,
Germany) at 258C using filtered (0.2 mm) and degassed HBS-KT running
buffer (10 mm HEPES, pH 7.4, 150 mm KCl, 3 mm EDTA, 0.05% (v/v)
Tween-20). Protein-grade detergents (Tween-20, Empigen) were from
Anatrace (Maumee, USA) and Pierce gentle elution (PGE) buffer was
from Thermo Scientific (Illinois, USA); all other chemicals were reagent
grade quality. Concentrated compound stocks were prepared in 100%
DMSO, quantified using their molar extinction coefficients at 320 nm,
and stored at 48C. As recommended by the manufacturer, the SA sensors
were preconditioned with 1m NaCl in 50 mm NaOH (three 1 min pulses
0
.625, 0.75, 1.0, 1.25, 1.5, 2.0 and 2.5 mm (see the Supporting Information
for details). After 6 min of agitation on a plate-shaker, the plates were al-
lowed to incubate at room temperature for approximately 84 min before
the fluorescence was measured. Fluorescence (F) was measured on
a plate reader (DTX 880, Beckman) equipped with 485ꢁ20 and 535ꢁ
3
5 nm filters for excitation and emission of thiazole orange, respectively.
Percentage displacement (PD) was calculated by PD=100ꢀ
0
ACHTUNGTNRENUNG[ (F/F )ꢃ
1
00], in which F is F without addition of platinum complexes. The con-
0
ꢀ1
ꢀ1
at 50 mLmin ) before the capture of DNA at 10 mLmin (15 mm stocks
diluted to 100 nm in running buffer containing 0.5m KCl). To minimize
nonspecific binding to the SA-coated sensors, the compounds were serial-
centration of added complex was then plotted against the PD, and
a DC50 value was determined by the concentration of platinum complex
needed to achieve a 50% displacement of thiazole orange.
ꢀ1
ly diluted in running buffer (<0.1% DMSO final) containing 1 mgmL
Affinity constants for the complexes (K
fluorescence isotherms by fitting the following equation using in-house
MATLAB scripts:
A
) were extracted from the FID
CM dextran (BioChemika #27560, Fluka/Sigma–Aldrich, Missouri,
USA). To assess binding specificity in multicycle “KINJECT” mode, di-
luted compounds were injected over reference (SA-only) and DNA-im-
ꢀ
1
mobilized (450 RU) surfaces at 25 mLmin (5 min association+10 min
dissociation; 5 mm fixed). Between sample injections, the sensors were re-
I
0
þ ðI
1
ꢀI
0
Þf
B
ð1Þ
ꢀ
1
generated at 50 mLmin using two 30 s pulses of PGE buffer containing
0.05% (v/v) Triton-X100 (solution I) or Empigen (solution II), followed
by EXTRACLEAN and RINSE procedures. To assess dose-dependent
binding in multicycle KINJECT mode, diluted compounds were titrated
where I
0
and I
1
are the fluorescence intensities in the absence and with
saturating amounts of ligand, respectively; f
plex bound to thiazole orange at any given ligand concentration:
B
is the fraction of G-quadru-
(
0–5 mm, twofold dilution series) over lower-density DNA surfaces
½
MXꢃ
(300 RU) and were regenerated in a similar manner. Data were doubled-
referenced and represent duplicate injections acquired from at least two
independent trials. For each replicate series, a buffer blank was injected
first, the highest titrant concentration second, and serial dilutions fol-
lowed (from the lowest to the highest concentration); comparing respons-
es between the two highest titrant injections verified consistent DNA sur-
face activity throughout each assay. To estimate apparent equilibrium dis-
f
B
¼
ð2Þ
½
Mꢃ
T
in which [MX] is the concentration of dye-bound G-quadruplex and [M]
is the total concentration of all free and bound forms of G-quadruplex.
MX] was calculated according to:
T
[
½
MXꢃ ¼ ½Xꢃ
T
f
X
ð3Þ
D
sociation constants (K ), steady-state binding responses (Req; average
RU at the end of the association phase) were plotted as a function of
complex concentration (C) and then subjected to nonlinear regression
where f
X
is the fraction of dye molecules bound to G-quadruplex at
is the total concentration of both
was calculated as:
a given ligand concentration and [X]
free and bound dye: f
T
(
“steady-state affinity” model, BIAevaluation v4.1 software). Titration
series were also analyzed using the “Fit k /k separate” tool to estimate
the individual dissociation rate constants (k ; evaluated in the early por-
tion of the dissociation phase to exclude rebinding effects). Theoretical
binding maxima were predicted using the following equation: Rmax
MWA/MWL)(R )(n) where Rmax is the maximal binding response (RU)
X
a
d
d
K
X
½Mꢃ
f
X
¼ 1
ð4Þ
þ K ½Mꢃ
X
=
(
L
where K
X
is the affinity equilibrium constant for thiazole orange binding
at saturating compound concentrations; MWA is the molecular weight
(Da) of the compound injected in solution; MWL is the molecular
and [M] is the concentration of unbound G-quadruplex. [M] was deter-
Chem. Eur. J. 2013, 19, 17836 – 17845
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
17843

H. F. Sleiman et al.
weight (Da) of the DNA fragment immobilized; R
L
is the amount (RU)
Acknowledgements
of DNA immobilized; n is the predicted binding stoichiometry (e.g., 1:1).
G-quadruplex sequences 22AG, cmyc, and ckit1, and duplex HP were
formed identically to those used for CD melting in potassium conditions.
This work was supported by a Canadian Institute for Health Research
(CIHR) Grant MOP-89978 to H.S. and N.M.K. J.C. was supported by the
CIHR Strategic Training Initiative in Chemical Biology and the McGill
CIHR Training in Drug Development studentships. A.M. is supported by
the National Science and Engineering Research Council (Canada), and
the Groupe de Recherche Axꢂ sur la Structure des Protꢂines (GRASP).
The authors thank Dr. Guillaume Lesage from McGill Cell Imaging and
Analysis Network (CIAN) for conducting the robotics assays and
Mr. Mitchell Huot and Dr. Samuel Sewall for use of the Biotage Initiator
microwave. The McGill SPR Facility thanks the Canada Foundation for
Innovation (CFI) and Canadian Institutes of Health Research (CIHR
Research Resource Grants to Sheldon Biotechnology Centre) for infra-
structure support.
MTS assay: The cytotoxicities of complexes 2, S4, and S7 were assessed
using the CellTiter96 kit from Promega according to the manufacturerꢁs
instructions. Briefly, HeLa cells (human cervical cancer) and A172 cells
(
human glioblastoma) were seeded at a density of 5000 cells per well in
a 96-well plate in DMEM media (Invitrogen) supplemented with 10%
FBS and antibiotics and antimycotic (ABAM). Platinum complex stock
solutions were initially diluted in growth medium to yield final concentra-
tions ranging from 1.6 to 100 mm. Cells were incubated for 72 h in 5%
CO
2
at 378C. After the incubation period, the MTS reagent was added to
at 378C. Subsequent-
each well and further incubated for 3 h in 5% CO
2
ly, 96-well plates were allowed to equilibrate at room temperature and
the absorbance was read at 490 nm using a BioTek Epoch microplate
reader. All quantifications were done using GraphPad Prism 5 software.
[
[
[
1] A. De Cian, L. Lacroix, C. Douarre, N. Temime-Smaali, C. Trente-
saux, J. F. Riou, J. L. Mergny, Biochimie 2008, 90, 131–155.
2] A. N. Lane, J. B. Chaires, R. D. Gray, J. O. Trent, Nucleic Acids Res.
Measurement of mRNA levels: Whole-cell mRNA was extracted and
cDNA strands were reverse transcribed using the SSIII reverse transcrip-
tion kit (Invitrogen) according to the manufacturerꢁs recommendations.
We then assessed the presence of mRNA of c-Kit by PCR with the fol-
lowing primers: forward primer: 5’-CAG GCA ACG TTG ACT ATC
AGT-3’ and reverse primer: 5’-ATT CTC AGA CTT GGG ATA ATC-
2
008, 36, 5482–5515.
3] J. L. Huppert, S. Balasubramanian, Nucleic Acids Res. 2006, 35,
06–413.
4
[
[
4] S. Neidle, FEBS J. 2010, 277, 1118–1125.
[
52]
3
’. PCR conditions used were: 948C for 45 s, 558C for 30 s, and 728C
5] a) A. Siddiqui-Jain, C. L. Grand, D. J. Bearss, L. H. Hurley, Proc.
Natl. Acad. Sci. USA 2002, 99, 11593–11598; b) P. Wu, D. L. Ma,
C. H. Leung, S. C. Yan, N. Zhu, R. Abagyan, C. M. Che, Chem. Eur.
J. 2009, 15, 13008–13021; c) T. Agarwal, S. Roy, T. K. Chakraborty,
S. Maiti, Biochemistry 2010, 49, 8388–8397; d) T. Shalaby, A. O. Von
Bueren, M. L. Hꢅrlimann, G. Fiaschetti, D. Castelletti, T. Masayuki,
K. Nagasawa, A. Arcaro, I. Jelesarov, K. Shin-ya, M. Grotzer, Mol.
Cancer Ther. 2010, 9, 167–179; e) P. Wang, C. H. Leung, D. L. Ma,
S. C. Yan, C. M. Che, Chem. Eur. J. 2010, 16, 6900–6911; f) T. M.
Ou, J. Lin, Y. J. Lu, J. Q. Hou, J. H. Tan, S. H. Chen, Z. Li, Y. P. Li,
D. Li, L. Q. Gu, Z. S. Huang, J. Med. Chem. 2011, 54, 5671–5679.
6] a) M. Bejugam, S. Sewitz, P. S. Shirude, R. Rodriguez, R. Shahid, S.
Balasubramanian, J. Am. Chem. Soc. 2007, 129, 12926–12927; b) M.
Bejugam, M. Gunaratnam, S. Mꢅller, D. A. Sanders, S. Sewitz, J. A.
Fletcher, S. Neidle, S. Balasubramanian, ACS Med. Chem. Lett.
for 60 s for 35 cycles. PCR of 18S rRNA with the following forward pri-
mers: 5’-GTA ACC CGT TGA ACC CCA TTC-3’ and reverse primer:
5
’-CCA TCC AAT CGG TAG TAG CG-3’, were used as internal con-
trol.
Western blot analysis: For Western blot analysis we seeded 250000 cells
in a 6-well plate. The next day, each complex (2, S4, and S7) was added
at 10 mm final concentration. Cells and ligands were incubated for 72 h.
Samples were lyzed with cell lysis buffer (NEB #9803). Protein concen-
trations were assessed by the Bradford assay and equal amounts (50 mg)
of protein were loaded and resolved by SDS-PAGE. Gel contents were
transferred to a PVDF membrane and Western blot analysis was carried
out. Blocking was performed with TBST (Tris-buffered saline with
[
0
.05% Tween 20)+5% nonfat milk for c-KIT and TBST+3% BSA for
b-actin. Anti-c-KIT (Santa Cruz, SC-168) was incubated, overnight, in
blocking solution at a concentration of 1:100. Anti-b-actin (Abcam,
ab8226) was incubated, overnight, in blocking solution at a dilution
2
010, 1, 306–310; c) K. I. McLuckie, Z. A. Waller, D. A. Sanders, D.
Alves, R. Rodriguez, J. Dash, G. J. McKenzie, A. R. Venkitaraman,
S. Balasubramanian, J. Am. Chem. Soc. 2011, 133, 2658–2663.
7] a) T. Taka, K. Joonlasak, L. Huang, T. Randall Lee, S. W. T. Chang,
W. Tuntiwechapikul, Bioorg. Med. Chem. Lett. 2012, 22, 518–522;
b) D. Sun, W. J. Liu, K. Guo, J. J. Rusche, S. Ebbinghaus, V. Go-
khale, L. H. Hurley, Mol. Cancer Ther. 2008, 7, 880–889.
8] Y. Yarden, W. J. Kuang, T. Yang-Feng, L. Coussens, S. Munemitsu,
T. J. Dull, E. Chen, J. Schlessinger, U. Francke, A. Ullrich, EMBO J.
1987, 6, 3341–3351.
1
:1000. Blots were washed four times with 1ꢃTBST. Anti-mouse and
[
[
anti-rabbit goat polyclonal HRP-conjugated antibodies were used to
detect b-actin and c-KIT, respectively. Secondary antibodies were diluted
at 1:5000 in appropriate blocking buffers. Blots were washed four times
with 1ꢃTBST and bands were detected with Clarity Western ECL (Bio-
Rad, 170-5060).
CD titrations: The ckit1 G-quadruplex was formed (50 mm) in: 1) 10 mm
K
2
HPO
KCl, and 3) 0.906 mm Na
10.34 mm NaCl by heating to 958C for 10 min and then cooling rapidly
4
/10 mm KH
2
PO
4
, 2) 10 mm K
2
HPO
4
/10 mm KH
2
PO
4
plus 100 mm
[9] T. J. Shaw, E. J. Keszthelyi, A. M. Tonary, M. Cada, B. C. Vanderhy-
den, Exp. Cell Res. 2002, 273, 95–106.
10] S. Rankin, A. P. Reszka, J. Huppert, M. Zloh, G. N. Parkinson, A. K.
Todd, S. Ladame, S. Balasubramanian, S. Neidle, J. Am. Chem. Soc.
2
HPO /NaH
4
2
PO
4
plus 5.33 mm KCl and
[
1
to room temperature. Samples were stored overnight, at 48C before di-
luting to 10 mm in the respective buffers. For the titration, 0–5 mL stock
solution of 2 (2 mm) in DMSO were added in series with either a 10 or
2
005, 127, 10584–10589.
[
[
11] M. Sattler, R. Salgia, Leuk. Res. 2004, 28 Suppl 1, S11–S20.
12] A. S. Yavuz, P. E. Lipsky, S. Yavuz, D. D. Metcalfe, C. Akin, Blood
2002, 100, 661–665.
9
3
0 min incubation period before being scanned. Spectra were scanned at
78C from 350–230 nm at 100 nmmin with a bandwidth of 1 nm and
ꢀ1
[
[
[
13] B. D. Smolich, Blood 2001, 97, 1413–1421.
a response time of 2 s.
14] M. A. Postow, R. D. Carvajal, Cancer J. 2012, 18, 137–141.
15] P. Rutkowski, Z. Nowecki, P. Nyckowski, W. Dziewirski, U. Grzesia-
kowska, A. Nasierowska-Guttmejer, M. Krawczyk, W. Ruka, J.
Surg. Oncol. 2006, 93, 304–311.
16] C. R. Antonescu, P. Besmer, T. Guo, K. Arkun, G. Hom, B. Koryo-
towski, M. A. Leversha, P. D. Jeffrey, D. Desantis, S. Singer, M. F.
Brennan, R. G. Maki, R. P. DeMatteo, Clin. Cancer Res. 2005, 11,
4182–4190.
CD melting assay: Each G-quadruplex was formed (15 mm) in either
1
0 mm K
2
HPO
4
/10 mm KH
2
PO
4
with or without 100 mm KCl depending
/10 mm NaH PO with 100 mm NaCl,
on the sequence, or 10 mm Na
where appropriate, by being heated to 958C for 10 min and then cooled
rapidly to room temperature. Samples were incubated with 4 equivalents
2
HPO
4
2
4
[
(
60 mm) complex 2, overnight, at 48C before melting. For the cmyc and
ckit1 sequences, there was no added KCl. Melting experiments were con-
ducted from 10–958C with a heating rate of 18Cmin while monitoring
ꢀ1
[17] L. K. Ashman, R. Griffith, Expert Opin. Invest. Drugs 2013, 22,
the wavelength at 295 nm.
103–115.
17844
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 17836 – 17845

Platinum(II) Phenylphenanthroimidazole
FULL PAPER
[
18] a) P. V. Boddupally, S. Hahn, C. Beman, B. De, T. A. Brooks, V. Go-
khale, L. H. Hurley, J. Med. Chem. 2012, 55, 6076–6086; b) D. S. H.
Chan, H. Yang, M. H. T. Kwan, Z. Cheng, P. Lee, L. P. Bai, Z. H.
Jiang, C. Y. Wong, W. F. Fong, C. H. Leung, D. L. Ma, Biochimie
[33] E. Largy, F. Hamon, M. P. Teulade-Fichou, Anal. Bioanal. Chem.
2011, 400, 3419–3427.
[34] a) I. M. Dixon, F. Lopez, J. P. Esteve, A. M. Tejera, M. A. Blasco, G.
Pratviel, B. Meunier, ChemBioChem 2005, 6, 123–132; b) S.
Manaye, R. Eritja, A. Avino, J. Jaumot, R. Gargallo, Biochim Bio-
phys Acta 2012, 1820, 1987–1996; c) R. Nanjunda, E. A. Owens, L.
Mickelson, S. Alyabyev, N. Kilpatrick, S. Wang, M. Henary, W. D.
Wilson, Bioorg. Med. Chem. 2012, 20, 7002–7011; d) Z. Li, J. H.
Tan, J. H. He, Y. Long, T. M. Ou, D. Li, L. Q. Gu, Z. S. Huang, Eur.
J. Med. Chem. 2012, 47, 299–311; e) K. Suntharalingam, A. J. White,
R. Vilar, Inorg. Chem. 2009, 48, 9427–9435.
2
011, 93, 1055–1064.
19] M. Cooney, G. Czernuszewicz, E. Postel, S. Flint, M. Hogan, Science
988, 241, 456–459.
[
[
1
20] a) A. T. Phan, V. Kuryavyi, S. Burge, S. Neidle, D. J. Patel, J. Am.
Chem. Soc. 2007, 129, 4386–4392; b) H. Fernando, A. P. Reszka, J.
Huppert, S. Ladame, S. Rankin, A. R. Venkitaraman, S. Neidle, S.
Balasubramanian, Biochemistry 2006, 45, 7854–7860.
[
[
21] W. Mꢅller, D. M. Crothers, J. Mol. Biol. 1968, 35, 251–290.
22] a) M. E. Peek, L. A. Lipscomb, J. A. Bertrand, Q. Gao, B. P.
Roques, C. Garbay-Jaureguiberry, L. D. Williams, Biochemistry
[35] J. Seenisamy, S. Bashyam, V. Gokhale, H. Vankayalapati, D. Sun, A.
Siddiqui-Jain, N. Streiner, K. Shin-ya, E. White, W. D. Wilson, L. H.
Hurley, J. Am. Chem. Soc. 2005, 127, 2944–2959.
1
994, 33, 3794–3800; b) J. B. Chaires, F. Leng, T. Przewloka, I. Fokt,
Y. H. Ling, R. Perez-Soler, W. Priebe, J. Med. Chem. 1997, 40, 261–
66; c) G. G. Hu, X. Shui, F. Leng, W. Priebe, J. B. Chaires, L. D.
[36] B. Fu, D. Zhang, X. Weng, M. Zhang, M. Heng, M. Yuzhi, X. Zhou,
Chem. Eur. J. 2008, 14, 9431–9441.
[37] a) K. Berg, L. Zhai, M. Chen, A. Kharazmi, T. C. Owen, Parasitol.
Res. 1994, 80, 235–239; b) G. Malich, B. Markovic, C. Winder, Toxi-
cology 1997, 124, 179–192.
2
Williams, Biochemistry 1997, 36, 5940–5946; d) H. Robinson, W.
Priebe, J. B. Chaires, A. H. Wang, Biochemistry 1997, 36, 8663–
8
1
1
670; e) F. Leng, W. Priebe, J. B. Chaires, Biochemistry 1998, 37,
743–1753; f) B. ꢆnfelt, P. Lincoln, B. Nordꢂn, J. Am. Chem. Soc.
999, 121, 10846–10847; g) L. A. Howell, R. Gulam, A. Mueller,
[38] J. R. Caceres-Cortes, J. A. Alvarado-Moreno, K. Waga, R. Rangel-
Corona, A. Monroy-Garcia, L. Rocha-Zavaleta, J. Urdiales-Ramos,
B. Weiss-Steider, A. Haman, P. Hugo, R. Brousseau, T. Hoang,
Cancer Res. 2001, 61, 6281–6289.
[39] G. R. Vandenbark, Y. Chen, E. Friday, K. Pavlik, B. Anthony, C.
deCastro, R. E. Kaufman, Cell Growth Differ. 1996, 7, 1383–1392.
[40] Z. A. Waller, S. A. Sewitz, S. T. Hsu, S. Balasubramanian, J. Am.
Chem. Soc. 2009, 131, 12628–12633.
M. A. OꢁConnell, M. Searcey, Bioorg. Med. Chem. Lett. 2010, 20,
956–6959.
23] a) D. Monchaud, M. P. Teulade-Fichou, Org. Biomol. Chem. 2008, 6,
27–636; b) S. N. Georgiades, N. H. Abd Karim, K. Suntharalingam,
6
[
6
R. Vilar, Angew. Chem. 2010, 122, 4114–4128; Angew. Chem. Int.
Ed. 2010, 49, 4020–4034.
[41] G. R. Vandenbark, Y. Chen, E. Friday, K. Pavlik, B. Anthony, C.
deCastro, R. E. Kaufman, Cell Growth Differ. 1996, 7, 1383–1392.
[42] Y. Wang, D. J. Patel, Structure 1993, 1, 263–282.
[
24] R. Kieltyka, J. Fakhoury, N. Moitessier, H. F. Sleiman, Chem. Eur. J.
2
008, 14, 1145–1154.
[
25] K. J. Castor, J. Mancini, J. Fakhoury, N. Weill, R. Kieltyka, P. Engle-
bienne, N. Avakyan, A. Mittermaier, C. Autexier, N. Moitessier,
H. F. Sleiman, ChemMedChem 2012, 7, 85–94.
[43] G. W. Collie, R. Promontorio, S. M. Hampel, M. Micco, S. Neidle,
G. N. Parkinson, J. Am. Chem. Soc. 2012, 134, 2723–2731.
[44] A. Ambrus, D. Chen, J. Dai, R. A. Jones, D. Yang, Biochemistry
2005, 44, 2048–2058.
[45] T. A. D. D. A. Case, T. E. Cheatham III, C. L. Simmerling, J. Wang,
R. E. Duke, R. Luo, M. Crowley, C. Walker Ross, W. Zhang, K. M.
Merz, B. Wang, S. Hayik, A. Roitberg, G. Seabra, I. Kolossvꢇry,
K. F. Wong, F. Paesani, J. Vanicek, X. Wu, S. R. Brozell, T. Stein-
brecher, H. Gohlke, L. Yang, C. Tan, J. Mongan, V. Hornak, G. Cui,
D. H. Mathews, M. G. Seetin, C. Sagui, V. Babin, P. A. Kollman,
University of California, San Francisco, 2008.
[46] a) W. D. Cornell, P. Cieplak, C. I. Bayly, I. R. Gould, K. M. Merz,
D. M. Ferguson, D. C. Spellmeyer, T. Fox, J. W. Caldwell, P. A. Koll-
man, J. Am. Chem. Soc. 1995, 117, 5179–5197; b) T. E. Cheatha-
m III, P. Cieplak, P. A. Kollman, J. Biomol. Struct. Dyn. 1999, 16,
845–862.
[
26] a) V. Casagrande, E. Salvati, A. Alvino, A. Bianco, A. Ciammaichel-
la, C. D’Angelo, L. Ginnari-Satriani, A. M. Serrilli, S. Iachettini, C.
Leonetti, S. Neidle, G. Ortaggi, M. Porru, A. Rizzo, M. Franceschin,
A. Biroccio, J. Med. Chem. 2011, 54, 1140–1156; b) H. Bertrand, A.
Granzhan, D. Monchaud, N. Saettel, R. Guillot, S. Clifford, A.
Guedin, J. L. Mergny, M. P. Teulade-Fichou, Chem. Eur. J. 2011, 17,
4
529–4539; c) W. C. Drewe, R. Nanjunda, M. Gunaratnam, M. Bel-
tran, G. N. Parkinson, A. P. Reszka, W. D. Wilson, S. Neidle, J. Med.
Chem. 2008, 51, 7751–7767; d) C. Pivetta, L. Lucatello, A. P. Krap-
cho, B. Gatto, M. Palumbo, C. Sissi, Bioorg. Med. Chem. 2008, 16,
9
331–9339; e) D. D’Ambrosio, P. Reichenbach, E. Micheli, A.
Alvino, M. Franceschin, M. Savino, J. Lingner, Biochimie 2012, 94,
54–863.
8
[
[
27] J. E. Moses, A. D. Moorhouse, Chem. Soc. Rev. 2007, 36, 1249–1262.
28] a) A. D. Moorhouse, A. M. Santos, M. Gunaratnam, M. Moore, S.
Neidle, J. E. Moses, J. Am. Chem. Soc. 2006, 128, 15972–15973;
b) W. C. Drewe, S. Neidle, Chem. Commun. 2008, 5295–5297;
c) J. E. Moses, D. J. Ritson, F. Zhang, C. M. Lombardo, S. Haider, N.
Oldham, S. Neidle, Org. Biomol. Chem. 2010, 8, 2926–2930; d) S.
Sparapani, S. M. Haider, F. Doria, M. Gunaratnam, S. Neidle, J. Am.
Chem. Soc. 2010, 132, 12263–12272; e) J. Dash, Z. A. Waller, G. D.
Pantos, S. Balasubramanian, Chem. Eur. J. 2011, 17, 4571–4581.
29] B. ꢆnfelt, P. Lincoln, B. Nordꢂn, J. Am. Chem. Soc. 2001, 123,
[47] a) J. Wang, W. Wang, P. A. Kollman, D. A. Case, J. Mol. Graph.
Model. 2006, 25, 247–260; b) J. Wang, R. M. Wolf, J. W. Caldwell,
P. A. Kollman, D. A. Case, J. Comput. Chem. 2004, 25, 1157–1174.
[48] W. L. Jorgensen, J. Chandrasekhar, J. D. Madura, R. W. Impey,
M. L. Klein, J. Chem. Phys. 1983, 79, 926–935.
[49] T. Darden, D. York, L. Pedersen, J. Chem. Phys. 1993, 98, 10089–
10092.
[50] W. F. van Gunsteren, H. J. C. Berendsen, Mol. Phys. 1977, 34, 1311–
1327.
[51] P. A. Kollman, I. Massova, C. Reyes, B. Kuhn, S. Huo, L. Chong, M.
Lee, T. Lee, Y. Duan, W. Wang, O. Donini, P. Cieplak, J. Srinivasan,
D. A. Case, T. E. Cheatham III, Acc. Chem. Res. 2000, 33, 889–897.
[52] J. Li, J. Quirt, H. Q. Do, K. Lyte, F. Fellows, C. G. Goodyer, R.
Wang, Am. J. Physiol. Endocrinol. Metab. 2007, 293, E475–E483.
[53] E. Therrien, P. Englebienne, A. G. Arrowsmith, R. Mendoza-San-
chez, C. R. Corbeil, N. Weill, V. Campagna-Slater, N. Moitessier, J.
Chem. Inf. Model. 2012, 52, 210–224.
[
[
3
630–3637.
30] a) R. Kieltyka, P. Englebienne, J. Fakhoury, C. Autexier, N. Moitess-
ier, H. F. Sleiman, J. Am. Chem. Soc. 2008, 130, 10040–10041; b) R.
Kieltyka, P. Englebienne, N. Moitessier, H. Sleiman, Methods Mol.
Biol. 2010, 608, 223–255.
[
[
31] C. R. Corbeil, P. Englebienne, N. Moitessier, J. Chem. Inf. Model
2
007, 47, 435–449.
32] a) D. Monchaud, C. Allain, M. P. Teulade-Fichou, Bioorg. Med.
Chem. Lett. 2006, 16, 4842–4845; b) D. Monchaud, C. Allain, H.
Bertrand, N. Smargiasso, F. Rosu, V. Gabelica, A. De Cian, J. L.
Mergny, M. P. Teulade-Fichou, Biochimie 2008, 90, 1207–1223.
Received: April 25, 2013
Revised: September 16, 2013
Published online: November 18, 2013
Chem. Eur. J. 2013, 19, 17836 – 17845
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
17845