Tridentate N-donor palladium(II) complexes as efficient coordinating quadruplex DNA binders

Article abstract of DOI:10.1002/chem.201102300

Fifteen complexes of palladium, platinum, and copper, featuring five different N-donor tridentate (terpyridine-like) ligands, were prepared with the aim of testing their G-quadruplex-DNA binding properties. The fluorescence resonance energy transfer melti

Full text of DOI:10.1002/chem.201102300

DOI: 10.1002/chem.201102300
Tridentate N-Donor Palladium(II) Complexes as Efficient Coordinating
Quadruplex DNA Binders
Eric Largy,[a] Florian Hamon,[a] Frꢀdꢀric Rosu,[b] Valꢀrie Gabelica,[b] Edwin De Pauw,[b]
Aurore Guꢀdin,[c] Jean-Louis Mergny,[c] and Marie-Paule Teulade-Fichou*[a]
Abstract: Fifteen complexes of palladi-
um, platinum, and copper, featuring
five different N-donor tridentate (ter-
pyridine-like) ligands, were prepared
with the aim of testing their G-quadru-
plex–DNA binding properties. The
fluorescence resonance energy transfer
melting assay indicated a striking posi-
tive effect of palladium on G-quadru-
plex DNA stabilization compared with
platinum and copper, as well as an in-
fluence of the structure of the organic
ligand. Putative binding modes (non-
coordinative p stacking and base coor-
dination) of palladium and platinum
complexes were investigated by ESI-
MS and UV/Vis spectroscopy experi-
ments, which all revealed a greater
ability of palladium complexes to coor-
dinate DNA bases. In contrast, plati-
num compounds tend to predominantly
bind to quadruplex DNA in their aqua
form by noncoordinative interactions.
dine, tMebip=2,2’-(4-p-tolylpyridine-
2,6-diyl)bis(1-methyl-1H-benzo[d]imid-
AHCTUNGTRENNaGUN zole)) coordinate efficiently G-quad-
ruplex structures at room temperature
in less than 1 h, and are more efficient
than their platinum counterparts for in-
hibiting the growth of cancer cells. Al-
together, these results demonstrate that
both the affinity for G-quadruplex
DNA and the binding mode of metal
complexes can be modulated by modi-
fying either the metal or the organic
ligand.
Remarkably, complexes of [Pd
ACHTUNGTRNE(NUNG ttpy)]
and [Pd(tMebip)] (ttpy=tolylterpyri-
AHCTUNGTRENNUNG
Keywords: DNA · FRET · G-quad-
ruplexes · mass spectrometry · tran-
sition metals
Introduction
translation, etc.), and thereby, could be involved in the regu-
lation of gene expression.[3] The synthesis of small molecules
capable of targeting G4 DNA is a rapidly expanding field:
these compounds could act pharmacologically through se-
quence/structure specificity and allow better understanding
of the biological role(s) of quadruplexes.[4] Small molecules
usually bind to G4 DNA through p-stacking interactions
with external G-quartets, but some also interact with the
loops[5] and grooves.[6]
Since the discovery more than 40 years ago that cisplatin
derivatives could act as antitumor agents,[7] an extensive
number of metal complexes aimed at binding DNA have
been synthesized. Predictably, metal complexes have also
emerged in the G4 DNA field and display various binding
modes (p stacking, metalation of bases, cleavage).[8] The ad-
vantages of using metal complexes are numerous. Different
metal cations adopt diverse geometries for a given ligand,
while introducing positive charges. Moreover, the metal can
withdraw electrons from the ligand, thus making it more
suitable for p stacking with electron-donor bases, such as
guanines. Terpyridine (tpy) derivatives with a tolyl moiety
(ttpy) or an extended aromatic surface (dibenzoterpy
(BisQ); Scheme 1) have been successfully used to prepare
transition-metal complexes (Pt2+, Cu2+) that have shown
high binding affinity for the human telomeric G4.[9] In par-
Guanine (G)-rich DNA and RNA sequences have the abili-
ty to fold into four-stranded helicoidal structures called G-
quadruplexes (abbreviated to G4s). A large number of se-
quences identified by bioinformatics studies[1] may form var-
ious quadruplex structures, which share the G-quartet as a
common monomeric motif, but differ in the loop arrange-
ments.[2] A G-quartet is the association of four guanines by
Hoogsteen hydrogen bonds in a coplanar fashion. This motif
self-stacks through p–p aromatic forces and by sandwiching
alkaline cations (Na+, K+), which participate greatly in the
stability of the quadruplex structure. It is now widely as-
sumed that G4 DNA may interfere with various biological
events related to the transfer and maintenance of genetic in-
formation (replication, transcription, telomeric functions,
[a] E. Largy, Dr. F. Hamon, Dr. M.-P. Teulade-Fichou
UMR176 CNRS, Institut Curie, Centre de Recherche,
Centre Universitaire, 91405 Orsay (France)
Fax : (+33)169-07-53-81
E-mail: marie-paule.teulade-fichou@curie.fr
[b] Dr. F. Rosu, Dr. V. Gabelica, Prof. E. De Pauw
Mass Spectrometry Laboratory, Chemistry Institute B6C
Universitꢀ de Liꢁge, 4000 Liꢁge (Belgium)
[c] A. Guꢀdin, Dr. J.-L. Mergny
ticular, [PtACTHNUTRGNEUG(N ttpy)] selectively platinates adenine bases located
Universitꢀ de Bordeaux, Inserm U869, IECB,
ARNA laboratory. 2 Rue Robert Escarpit, 33607 Pessac (France)
in the loops of this quadruplex structure.[10] Square-planar
platinum(II) complexes, with two exchangeable ligands, such
as cisplatin,[11] or one, such as tpy complexes,[12] can coordi-
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201102300.
13274
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 13274 – 13283

FULL PAPER
s,[21a] including luminescent platinum complexes.[22] In the
case of the tpy complexes, a tolyl group in the para position
of the central pyridine ring dramatically increased the quad-
ruplex binding affinity.
We therefore envisaged the preparation of the tolyl deriv-
ative of the Mebip ligand (tMebip; 2,2’-(4-p-tolylpyridine-
2,6-diyl)bis(1-methyl-1H-benzo[d]imidazole)), which com-
bined the structural features of ttpy and benzimidazole scaf-
folds, to prepare a metal complex with improved G4-recog-
nition properties. Similarly, the tolyl moiety was introduced
into the dibenzoterpyridine (BisQ) scaffold used in previous
studies,[10] thereby affording a second ligand (tBisQ) featur-
ing a tolylpyridine core and an extended aromatic surface.
The five ligands derived from the tpy model (Scheme 1)
were combined with the three metallic cations (Pd2+, Cu2+
,
Scheme 1. Structures of tridentate N-donor ligand described herein.
Pt2+). Binding affinities, binding modes, and binding kinetics
of the five metal complex families for G4 DNA were investi-
gated by fluorescence resonance energy transfer (FRET)
melting,[23] electrospray mass spectrometry experiments,[24]
and time-dependent UV/Vis absorbance measurements.
nate to DNA bases (with a strong preference for guanine
N7) after a hydrolysis step (aquation), which is the rate-lim-
iting step and highly dependent on the chelating ability of
the ligand coordinated to the metal.[13] Palladium(II) and
platinum(II) are soft Lewis acids and share the same
square-planar geometry.[14] They display similar characteris-
tics when coordinated to N-donor polydentate ligands, but
interestingly, the key difference rests on the ligand-exchange
kinetics, in particular, during the aquation reaction. Hydrol-
ysis rates are much faster for palladium than for platinum
complexes (105 times faster according to instrumental meth-
ods,[15] and 106 times estimated by ab initio studies[16]). This
results in kinetic instability of Pd–DNA complexes relative
to their Pt counterparts, which has limited the use of Pd de-
rivatives for DNA targeting and for biomedical applications.
Nevertheless, because this property can be modulated by
the nature of the heterocyclic ligand surrounding the metal,
interest is now gradually shifting towards palladium and
other transition-metal complexes for biomedical applica-
tions.[17]
Results and Discussion
Synthesis of complexes: tMebip was prepared in three steps
from chelidamic acid, with an overall yield of 45%
(Scheme 2). Formation of benzimidazoles was achieved by
thermal cyclization of chelidamic acid and N-methyl-1,2-
phenylenediamine in the presence of polyphosphoric acid,
which was used as a solvent.[25] Position 4 of the pyridine
was then brominated by phosphorus oxybromide,[26] allowing
Based on the reasons outlined above, we speculated that
the introduction of palladium to our metal tpy G4 binders
could modulate both noncoordinative binding and the coor-
dination capability of quadruplex structures. Therefore, we
launched a program aimed at modulating both the ligand
surface and the coordinating metallic cation. It is worth
noting that palladium complexes devoted to quadruplex
DNA recognition have already been synthesized by Vilar
et al. but the series investigated did not display strong bind-
ing affinities for the targeted DNA.[18]
Bis(N-methylbenzimidazolyl)pyridine (Mebip) is a triden-
tate nitrogen ligand often used in inorganic chemistry as an
analogue of tpy, but compared with the latter, it is a moder-
ate s donor and also a p acceptor.[19] The larger aromatic
surface, relative to tpy, seems to be more suited to p overlap
with a G-quartet.[4b] For instance, the benzimidazole motif is
a purine mimic that may stack efficiently on DNA bases[20]
and has already been used to construct quadruplex-DNA li-
gands.[21] In addition, the bis(benzimidazole)pyridine core
was recently shown to lead to efficient quadruplex ligand-
Scheme 2. Synthesis of tMebip complexes. Reagents and conditions:
a) N-methylbenzene-1,2-diamine, polyphosphoric acid (PPA), 2158C,
17 h; b) POBr3, 1408C, 17 h; c) p-tolylboronic acid, [Pd
THF/water, reflux, 24 h; 45% yield over three steps. d) K2PtCl4, nitrome-
thane, reflux, 24 h, 15% yield; e) [Pd(cod)Cl2] (cod=1,5-cyclooctadiene),
(NO3)2, CH2Cl2, acetonitrile, 48C,
ACHTNUGERTN(NUGN PPh3)4], K2CO3,
AHCTUNGTRENNUNG
DMF, 508C, 24 h, 48% yield; f) CuACHTNUGTRENNUNG
48 h, 42% yield.
Chem. Eur. J. 2011, 17, 13274 – 13283
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
13275

M.-P. Teulade-Fichou et al.
the introduction of the tolyl group through a Suzuki cou-
pling.[27] The final product was easily obtained because it
precipitated at room temperature in the THF/water mixture
used as the solvent. This synthesis can be performed on the
gram scale and requires few purification steps.
BisQ was prepared by following procedure described in
the literature[10] based on a double Friedlꢃnder condensation
between two equivalents of 2-aminobenzaldehyde and 2,6-
acetylpyridine (Scheme 3). 2-Aminobenzaldehyde is not
The key intermediate, 2-acetylquinoline (5), was obtained
by esterification of quinaldic acid, followed by Claisen con-
densation, and then a saponification/decarboxylation step.
The ligand was finally obtained in 54% yield by a double
Krçhnke reaction, using a double condensation with aque-
ous ammonia on 2-acetylquinoline, ammonium acetate, and
4-methylbenzaldehyde under harsh conditions in a sealed
tube at 1008C for 24 h.[30]
Palladium, platinum, and copper complexes of BisQ,
tBisQ, and tMebip were prepared by adapting procedures
described in the literature.[9,31] These complexes were ob-
tained in lower yields than those with tpy and ttpy, presuma-
bly because of their larger aromatic surface, which delocaliz-
es the nitrogen lone-pair electrons.
FRET melting assay: The binding performances of the free
ligands and their corresponding complexes were first evalu-
ated by FRET melting assays. This well-known assay is
based on monitoring the stability induced by binding of a
ligand to a fluorescently labeled quadruplex structure, such
as the human telomeric sequence F21T (FAM-G3ACHTNUTRGEN(UGN T2AG3)3-
Tamra) or other G4-forming sequences (F21RT, F25CebT,
F21CTAT, FmycT, Fkit1T, Fkit2T; see the Experimental
Section and the Supporting Information).[23] The stabiliza-
tion, which was measured by a FRET effect, was expressed
as the increase in melting temperature of the labeled oligo-
nucleotide (DT1/2) induced by the ligand. Addition of an un-
labeled DNA competitor (for example, duplex DNA) ena-
bles one to evaluate the selectivity for the targeted quadru-
plex structure.
The FRET melting data, summarized in Figure 1, indicate
little or no stabilizing effect of the free organic ligands
(BisQ, tBisQ, and tMebip), thereby confirming the crucial
effect of the metallic cation, as previously observed.[9] Re-
markably, compared with the corresponding Pt2+ and Cu2+
complexes, Pd2+ derivatives are clearly the most efficient at
Scheme 3. Synthesis of BisQ complexes. Reagents and conditions: a) iron
powder, 0.1n HCl (aq), 958C, 30 min; b) 2,6-diacetylpyridine, potassium
hydroxide, 958C, 1 h; 75% yield over two steps in a one-pot procedure;
c) K2PtCl4, nitromethane, reflux, 24 h, 21% yield; d) [PdACHTUNGTRENNUNG
508C, 24 h, 84% yield; e) CuACHTGNUTRENNUNG
yield.
stable,[28] and thus, it was generated in situ by reduction of
2-nitrobenzaldehyde.[29] An overall yield of 75% was ob-
tained for the two-step, one-pot process.
The synthesis of tBisQ was less straightforward and was
achieved in 4 steps (Scheme 4) with an overall yield of 27%.
Scheme 4. Synthesis of tBisQ complexes. Reagents and conditions: a) sul-
furic acid, MeOH, reflux, 16 h, 89% yield; b) tBuOK, ethyl acetate, RT,
15 min, 79% yield; c) HCl, dioxane, 1008C, 16 h, 70% yield; d) 4-methyl-
benzaldehyde, KOH, aqueous ammonia, EtOH, 1008C, sealed tube, 24 h,
Figure 1. FRET melting stabilization (DT1/2 in 8C) of the human telomer-
ic sequence F21T (0.2 mm) in lithium cacodylate buffer (10 mm), KCl
(10 mm), and LiCl (90 mm) in the presence of metal complexes or the
corresponding free ligand (5 equiv). Ligand: white, Cu2+: diagonal hatch-
ing, Pt2+: horizontal hatching, Pd2+: black.
54% yield; e) K
[PtCl4], nitromethane, reflux, 24 h, 94% yield; f) [Pd-
(NO3)2, CH2Cl2, acetoni-
G
trile, 65%.
13276
www.chemeurj.org
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 13274 – 13283

Palladium(II) Complexes Coordinating Quadruplex DNA
FULL PAPER
stabilizing human telomeric G4. The tpy series represents a
striking example of the pronounced effect of Pd2+. [Pd
ACTHNUTRGNE(NUG tpy)]
induces strong stabilization of the quadruplex structure
(DT1/2 = +17.38C), whereas the two other tpy complexes ex-
hibit a negligible effect (DT1/2 < +28C). Similarly, stabiliza-
tion in the tMebip series is particularly impressive: DT1/2 for
the Pd2+ complex almost reaches the limit of the test with a
DT1/2 value of around +358C, whereas DT1/2 values of
around 10–128C are recorded for the Cu2+ and Pt2+ coun-
terparts. The same trend was observed in the ttpy series, but
to a lesser extent (DT1/2 = +20.3, +12.9, +6.38C for Pd, Pt,
and Cu derivatives, respectively). Conversely, the effect of
palladium was much less striking in the BisQ and tBisQ
series, for which the three complexes showed similar per-
formances (7–108C<DT1/2 <148C). The hindrance of the ar-
omatic surface of these ligands is likely to counterbalance
the effect of the metallic cations, suggesting that in those
cases the interaction is dominated by the aromatic ligand.
On the other hand, the interaction is clearly governed by
the nature of the metallic cation in the three other series
(tpy, ttpy, and tMebip). The beneficial effects of cation and
Figure 3. FRET melting stabilization (DT1/2 in 8C) of the labeled G4
DNA human telomeric sequence F21T (0.2 mm) in lithium cacodylate
buffer (10 mm), KCl (10 mm) + LiCl (90 mm), in the presence of increas-
ing concentrations of competitor duplex DNA ds26 (0, 15, and 50 equiv,
black to light gray), by 5 (left) and 2 equiv (right) of tMebip complexes.
not cause a further decrease in stabilization (Figure 3). The
same trend was observed for all Pd2+ complexes, irrespec-
tive of the ligand (see Figure S1 in the Supporting Informa-
tion), but not for the Pt2+ and Cu2+ complexes. Neverthe-
ligand are clearly additive in the case of [PdACHTNUTRGNE(UNG tMebip)],
which appears to be the best candidate of the fifteen com-
plexes studied.
Therefore, the tMebip series was similarly assayed for a
wider spread of quadruplex-forming sequences of biological
interest. The results shown in Figure 2 confirm the stronger
less, [PdACTHNGUTERNN(UG tMebip)] retains a significant effect (+178C) under
these conditions of harsh competition (50 equiv of duplex),
and thus, remains the best candidate. The same experiment
conducted at a lower concentration of [PdACTHNUTRGNE(NUG tMebip)] indicat-
ed that the decrease was abolished under these conditions
of lower compound/DNA ratio, thereby suggesting that the
excess compound may participate “nonspecifically” in the
stabilization observed (Figure 3, left part).
This unusual behavior raises questions about the nature
of the interaction and how to rationalize the effect of the
palladium relative to the other cations, in particular, plati-
num. It could be hypothesized that a portion of the complex
was bound externally, thus being easily redistributed on the
competitor. In addition, metal coordination to DNA bases
may occur during the experiment favored by the tempera-
ture increase. The DT1/2 value measured could thus result
both from noncoordinative binding and from in situ coordi-
nation. To investigate further the nature of the interaction,
in the case of Pd2+ and Pt2+ complexes, complementary ex-
periments were conducted for the ttpy and tMebip families
by using isothermal methods, namely, ESI-MS and UV/Vis
spectroscopy.
Figure 2. FRET melting stabilization (DT1/2 in 8C) of the G4s F21T,
F21RT, F25CebT, F21CTAT, FmycT, Fkit1T, Fkit2T (from black to light
gray) and of the duplex FdxT (white) (0.2 mm) in lithium cacodylate
buffer (10 mm), KCl (10 mm), and LiCl (90 mm) in the presence of
tMebip complexes. [DNA]=0.2 mm; [complexes]=1 mm.
ESI-MS: Mass spectrometry was used to characterize the
amount of ligand bound to duplexes and G4 structures.[24]
Furthermore, analysis of the masses of the complexes
formed between [Pt(ttpy or tMebip)]/[Pd(ttpy or tMebip)]
and DNA revealed a particular binding mode. Figure 4
shows a magnification of the complexes detected with the
stabilization of almost all G4s by [PdACTHNUTRGNEUNG(tMebip)], compared
with copper and platinum derivatives. Importantly, very low
binding was observed when using the doubly labeled duplex
(FdxT, 2.98C), revealing the strong preference of the com-
pound for G4 structures. Notably, high quadruplex selectivi-
ty was also observed for the Cu2+ and Pt2+ complexes. The
tetramolecular G4 Q1, dACHTNUTRGENNG[U TGGGGT]4. Not shown is the
selectivity of [Pd
(tMebip)], with regards to duplex structure,
signal of the free Q1 at charge state 5À, with an average m/z
of 1499.67. In the region corresponding to the complexes,
three species can be distinguished: the association between
DNA and the metal complex [Pt(ttpy or tMebip)]/[Pd(ttpy
or tMebip)], the association between DNA and the aqua
was also evaluated by competitive FRET melting using
F21T and a 26 bp duplex (ds26) as a competitor. A high de-
crease was observed when adding the first dose (3 mm) of
the duplex competitor, whereas the second dose (10 mm) did
Chem. Eur. J. 2011, 17, 13274 – 13283
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
13277

M.-P. Teulade-Fichou et al.
tions of the metal complexes with DNA (Scheme 5B and
C), whereas species lacking Cl or OH are interpreted as co-
ordinative associations involving coordination of the metal
to DNA (Scheme 5D).
Experiments conducted after incubation of the Pt com-
plexes with DNA for one day showed that the fraction of
complex directly coordinated to DNA increased very slowly
with time. For palladium complexes, however, hydrolysis
and subsequent metalation of the DNA occur at a much
faster rates than those for the platinum complexes and after
incubation for 1 h most of the complexes are coordinated to
DNA (Figure 4C and D). The aqua form is minor and the
chloride form is absent. Because no ammonium cation is re-
leased from the G4 structure, opening of the G-quartet is
unlikely to occur, and furthermore, this should be prevented
by p stacking of the aromatic ligand. Thus, the guanines of
the quadruplex core should be protected from metalation,[32]
since their N7 position is engaged in Hoogsteen hydrogen
bonding. Consequently, we hypothesized that complexes co-
ordinated to thymines surrounding the G-quartet (Figure S2
in the Supporting Information). Indeed, Q1 possesses eight
accessible thymines, the N3 of which is able to coordinate
Pt2+ and Pd2+ cations when not engaged in base pairing.[33]
Mass spectrometry therefore allows the amount of metal
complex bound coordinately and noncoordinately to the
DNA to be quantified separately. Figure 5 summarizes the
Figure 4. Nature of the 1:1 (DNA/ligand) complex of various metal deriv-
atives ( [Pt(ttpy)] (A), [Pt(tMebip)] (B), [Pd(ttpy)] (C), [Pd(tMebip)]
G
E
N
ACHTUNGTRENNUNG
(D)) with [dTGGGGT]4 obtained 1 h after mixing. For platinum com-
plexes (A and B) the binding of the aqua form (coordination of the Pt
with OH) to Q1 is predominant. Minor species are complexed with the
chloride form and direct coordination with DNA (neither OH nor Cl
bound to the metallic cation). For palladium derivatives (C and D),
direct coordination to the DNA is the major form, the aqua form is
minor, and the chloride form is undetected.
form of the metal complex [PtOH(ttpy or tMebip)]/
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
In all cases, the free complex is in the chlorido form, as
shown by high-resolution mass spectrometry (HRMS) analy-
sis. For the two platinum complexes (Figure 4A and B), the
major form bound to the DNA is, however, the aqua form
(ClÀ is replaced by OHÀ), but a few chlorido complexes as-
sociated with DNA are still observed. Another minor form
is the complex directly bound to DNA, for which we con-
clude that one DNA base or residue occupies the fourth co-
ordination site of Pt. The species with the chlorido and aqua
forms are therefore interpreted as noncoordinative associa-
Figure 5. Amounts of metal complexes ([Pt
[Pd(ttpy)] (C), [Pd(tMebip)] (D)) bound to different DNA G4s (Q1=
[dTGGGGT]4, Q2=human telomeric sequence d(GGGTTA)3GGG) and
ACHUTNGTRENNUG(ttpy)] (A), [PtACHTUNGTREN(NGUN tMebip)] (B),
A
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
duplexes (D1=(dCGCGGGCCCGCG)2, D2=(dCGCGAATTCGCG)2,
D3=(dCGTAAATTTACG)2). The concentration was 4.8 mm of DNA
and 8 mm of compound. The black bars correspond to the fraction of
ligand directly coordinated to DNA and the grey bars correspond to the
fraction of noncoordinately bound complexes (aqua and chlorido forms)
1 h after mixing.
quantification experiments with different quadruplexes (Q2
is the 21-nt human telomeric sequence and D1–D3 are
12 bp duplexes with decreasing GC content). Globally, the
amount of compound bound to DNA was higher for ttpy
complexes (Figure 5A and C) than for the tMebip counter-
parts (Figure 5B and D). The platinum derivative of
Scheme 5. Binding modes of [MACTHNURTGNEU(GN tMebip)] (M=Pt, Pd) observed by ESI-
MS and time-dependent UV/Vis absorbance measurements. Complexes
(A) can quickly p stack on a G-quartet (chlorido (B) and hydrolyzed (C)
forms) and/or coordinate to DNA bases (D).
13278
www.chemeurj.org
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 13274 – 13283

Palladium(II) Complexes Coordinating Quadruplex DNA
FULL PAPER
tMebip, in particular, appears to be very poorly active (Fig-
ure 5B). However, the ratios of complexes bound noncoor-
dinately and coordinately differ strongly: noncoordinating
binding is largely predominant for Pt derivatives, whereas
coordinate binding predominates for Pd derivatives. This is
consistent with the faster coordination rate of the Pd series
and confirms the formation of stable adducts with DNA. In
all cases, no clear structural selectivity appears because both
quadruplexes (Q1,2) and duplexes (D1,D2) are bound at a
more or less comparable level. Nonetheless, binding to du-
plexes is strikingly dependent on the GC content, since AT-
rich duplex D3 is very poorly bound, especially for palladi-
um derivatives (Figure 5C and D). These observations are
consistent with the preferential coordination of Pt2+ and
Pd2+ to the N7 position of guanines,[34] and with the fast co-
ordination rate of Pd complexes. Moreover, these tend to
demonstrate that the palladium complexes bind externally
in the major groove, where the N7 positions are accessible,
and hence, do not insert into the duplex structure. Finally,
the two Pd derivatives coordinate to the telomeric quadru-
plex Q2 more or less to the same extent (around 1 mm Fig-
ure 4C and D), but coordinative binding is much more prev-
alent for the ttpy complex (Figure 4C). In these cases, pref-
erential coordination of the metal to the adenines present in
the loops may be expected, as previously shown for Pt deriv-
atives.[10]
Time-dependent UV/Vis absorbance spectra: In an attempt
to gain further insight into the coordination capability of the
two series, the kinetics of coordination to DNA were fol-
lowed by UV/Vis spectroscopy measurements, as previously
described for [Pt
trum of [Pd(tMebip)] has a broad band with two local
(tpy)] complexes.[35] The absorption spec-
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
maxima at 330 and 400 nm, corresponding to the ligand-cen-
tered band (Figure 6). In the tail of this band, the weak con-
Figure 6. Change in absorption spectra during the interaction of [Pd-
AHCTUNGTREG(NNUN tMebip)] (10 mm) with 22AG (10 mm) in 10 mm lithium cacodylate,
pH 7.2 buffer, and 100 mm KCl. Complex alone: solid line (solid arrow
shows DNA addition). Successive times are 0, 50, 80, 160, 240 (black to
light gray dashed lines), 320, 415, 1320, 3000 min (black to light gray
short dashed lines). Dashed arrows indicate the kinetic evolution.
In summary, the apparent affinity and structural selectivi-
ty should thus be interpreted in light of these two possible
binding interactions, which may occur consecutively and/or
concomitantly. Thus, if we assume that palladium complexes
are stacked on one external quartet and locked into the
structure by coordination to loop bases, this dual interaction
should strongly stabilize the quadruplex structure. Conse-
quently, it is not surprising that the greatest stabilization of
quadruplexes is observed with palladium derivatives upon
FRET melting because a large fraction of the complexes
should be coordinately bound. However, the mass spectrom-
etry results highlight that a comparison of FRET melting
data for platinum and palladium complexes must be inter-
preted carefully, since the nature of the interaction (non-
coordinative vs. coordinative) with DNA evolves with time
and the ratio of noncoordinative versus coordinative binding
can be modified by the temperature increase. Additionally,
coordinative and noncoordinative binding modes should
contribute differently to the apparent stabilization of G4s,
depending on their respective fraction and DHbinding. For the
same reasons, the binding of palladium derivatives to duplex
DNA is difficult to evaluate because 1) it is strongly se-
quence dependant and seems to occur without insertion into
the duplex; 2) it is not necessarily a stabilizing interac-
tion,[34b] and hence, may not be easily detectable by FRET
melting; and 3) the fast coordination kinetics of palladium
may favor coordination to duplexes under the ESI-MS con-
ditions (a DNA concentration that is 24-fold higher than
that in FRET melting and no competition with quadruplex-
es).
tribution lying in the visible region is assigned to the metal-
to-ligand charge-transfer band (MLCT).[36] Indeed this con-
tribution cannot be the result of a metal centered d–d transi-
tion because it appears at a wavelength higher than 400 nm
and is characterized by a low intensity.[37] The addition of te-
lomeric quadruplex DNA [22AG, 5’-AG3ACHTNUTRGEN(GNU T2AG3)3-3’]
(1 equiv) resulted in moderate hypochromism, suggesting p
stacking of the complex, presumably on a G4. After this ini-
tial step, the spectrum further evolved over time with promi-
nent hyperchromism of the ligand band, displaying a maxi-
mum at 330 nm. Concomittantly, the MLCT band increased
in intensity and was slightly redshifted, while one isosbestic
point common to all curves remained at 425 nm. This indi-
cates the existence of interconversion between two species
and agrees with coordination of the complex to DNA, as
evidenced by the ESI-MS data. Indeed, the substitution of
an electron-withdrawing atom (chlorine) with a neutral
donor (nitrogen lone pair from a DNA base) can increase
the electron density of the metal center, and therefore, de-
crease the MLCT energy.[37–38] Finally, two other isosbestic
points appeared successively (395 nm, plain lines (Figure 6,
then 380 nm, short dashed lines), suggesting the existence of
two reactions with dissimilar rate constants, which may cor-
respond to two differently coordinated species.[39] In the
case of [PdACTHNUTRGNEUG(N ttpy)], the addition of G4 DNA also resulted in
a rapid change of the absorption spectrum of the complex
(Figure S3 in the Supporting Information). A significant hy-
pochromism of the ligand band also suggested direct stack-
ing of the complex on G4. The time evolution of the system
Chem. Eur. J. 2011, 17, 13274 – 13283
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
13279

M.-P. Teulade-Fichou et al.
was comparable to that of [Pd
in the absorbance of the ligand together with significant
bathochromism of the MLCT. Overall, the UV data for [Pd-
G
ordinated DNA bases to a large extent in a short time. No-
tably, mass spectrometry revealed that, after incubation for
1 h at room temperature, a major fraction of the complexes
involved palladated DNA. In addition, the palladium com-
plexes may coordinate to thymines in the vicinity of the G4
core when no purine base is accessible. Overall, we demon-
strated that tpy-like metal complexes could be finely tuned,
by varying the nature of both metal and ligand, to target G4
by a multiple interaction based on noncoordinative p stack-
ing and coordination to residues surrounding terminal G4s.
This particular interaction may open up perspective for spe-
cifically targeting G4 structures exhibiting proper loop con-
formations and/or dynamics.
ACHTUNGTRENNUNG(tMebip)] and [PdACHTUNGTRENNUNG(ttpy)] agree with the mass spectral data
and suggest an interaction with quadruplex DNA through
first stacking over a G4 and subsequently by fast coordina-
tion (on a minute timescale).
In contrast, no detectable immediate change occurred
after the addition of 22AG to [Pt
Supporting Information). Then the spectrum evolved in a
similar fashion to that of [Pd(tMebip)], but to a much lower
extent, with almost no change monitored in the MLCT
band. Conversely, addition of 22AG to [Pt(ttpy)] led to
AHCTUNGTRENNUNG
rapid hypochromism followed by an increase in absorbance
over time, along with the appearance of two simultaneous
isosbestic points (350 and 415 nm; Figure S5 in the Support-
Experimental Section
ing Information). Similar to [PtACTHNUGTRNEUNG(tMebip)], no significant
modification of the MLCT was detected.
General: 1H and 13C NMR spectra were recorded at 258C on a Bruker
Avance 300 spectrometer using tetramethylsilane (TMS) as an internal
standard. Deuterated CDCl3 and [D6]DMSO were purchased from SDS.
The following abbreviations are used: singlet (s), doublet (d), triplet (t)
and multiplet (m). Low-resolution ESI mass spectra were recorded on a
micromass ZQ 2000 (Waters) instrument. High-resolution ESI mass spec-
tra and elemental analyses were provided by the Institut de Chimie des
Substances Naturelles (I.C.S.N., Gif-sur-Yvette, France). TLC analysis
was carried out on silica gel (Merck 60F-254) plates with visualization at
254 and 366 nm. Preparative flash chromatography was carried out with
Merck silica gel (Si 60, 40–63 mm). Reagents and chemicals were pur-
chased from Sigma-Aldich, Acros, or Alfa-Aesar unless otherwise stated.
Solvents were purchased from SDS. Melting points were recorded on a
Kofler melting point apparatus and are uncorrected. UV/Vis spectra
were recorded on a Secoman Uvikon XL spectrophotometer and fluores-
cence melting curves were recorded on a Stratageme Mx 3005P real-time
PCR machine. Oligonucleotides purified by reversed-phase HPLC were
purchased from Eurogentec (Belgium).
In vitro cytotoxicity: Finally, we evaluated the effect of
tMebip and ttpy complexes on the growth of several cancer
cell lines. For both families, Pd complexes were highly effi-
cient (Table S1 in the Supporting Information). In particu-
lar, [PdACHTUNGTRENNUNG(ttpy)] strongly inhibited the proliferation of the
three tested cell lines (KB, A549, and MCF7) with IC50
values in the nanomolar range (65–115 nm; Table S2 in the
Supporting Information). This palladium complex was, in
this regard, more efficient than [PtACHTNUTRGNEG(UN ttpy)], which has IC50
values in the micromolar range.[10] Whether the higher cyto-
toxicity of palladium derivatives is related to differences in
the DNA interaction observed in vitro is not known and re-
quires further investigations. Nevertheless, these results indi-
cate that the novel compounds penetrate the cell membrane
and exhibit a promising drug-like potential.
Preparation of palladium complexes: A solution of dichloro(1,5-cyclooc-
tadiene)palladium (1.2 equiv) in a minimum amount of CH2Cl2 was
added to a solution of ligand (1 equiv) in a minimum amount of DMF.
The resulting yellow solution was stirred for 2–3 d at room temperature
under argon and protected from light. The resulting suspension was fil-
tered. The solid was washed with DMF, CH2Cl2, and then diethyl ether.
The powder was then dried under vacuum.
Conclusion
Using straightforward synthetic pathways, we prepared fif-
teen metal complexes, featuring three different metals,
namely, copper(II), platinum(II), and palladium(II). Five
structurally related tridentate N-donor ligands were used,
including well known tpy and ttpy, and the already studied
dibenzoterpyridine BisQ. To enhance the capability of metal
complexes binding G4 DNA, two larger ligands (tBisQ and
tMebip) were prepared. Unexpectedly, FRET melting ex-
periments revealed, for the first time, a significant positive
effect of palladium on the stabilization properties of the
complexes for G4 DNA. We hypothesize that this striking
difference between platinum and palladium complexes is
due to a difference in the nature of the DNA binding inter-
action despite the fact that they share the same square-
planar geometry. Consequently, ESI-MS and UV/Vis spec-
troscopy experiments were carried out and showed that,
under the conditions examined, platinum complexes tended
to predominantly bind G4 DNA in their aqua form by non-
coordinative interactions, whereas palladium complexes co-
Preparation of platinum complexes: Platinum complexes with tpy and
ttpy were prepared as previously described.[9] Other complexes were pre-
pared as follows:[31c] Potassium platinum(IV) chloride (1 equiv) and
sodium tetrafluoroborate (2 equiv) were added to a solution of ligand in
dry nitromethane. The mixture was heated at reflux for 48 h under argon
and protected from light. The NaCl precipitate was removed by hot fil-
tration. The product was precipitated by addition of diethyl ether (when
necessary), filtered through a membrane (Schleicher & Schuell, 1 mm),
washed with CH2Cl2 and diethyl ether, then dried under vacuum.
Preparation of copper complexes:
A solution of copper nitrate
(1.1 equiv) in a minimum amount of anhydrous acetonitrile was carefully
added dropwise to a solution of the ligand (1 equiv) in a minimum
amount of CH2Cl2 to form two immiscible layers. The biphasic solution
was kept at 48C until complete discoloration of the upper phase (typical-
ly 1–3 days). Green needles were filtered off; washed carefully with ace-
tonitrile, CH2Cl2, and diethyl ether; then dried under vacuum.
Synthesis of 1: The procedure for the synthesis of 1 was adapted from lit-
erature.[25] Chelidamic acid (3 g, 15 mmol) and N-methylphenylenedia-
mine (3.4 mL, 30 mmol) in polyphosphoric acid (55 g) were added to a
250 mL round-bottomed flask to afford a red viscous solution that was
stirred under argon at 2508C. After 48 h, the solution was poured (hot)
into cold water (300 mL) under vigorous stirring. The blue solid was col-
13280
www.chemeurj.org
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 13274 – 13283

Palladium(II) Complexes Coordinating Quadruplex DNA
FULL PAPER
lected by filtration, taken up into aqueous Na2CO3 (10%) at 1108C for
15 min, then filtered off. The solid was suspended in water (300 mL) and
the pH was adjusted to four. The foamy solid was collected again by fil-
tration. It was then recrystallized from hot DMSO by adding water until
the solution became cloudy. After the solution cooled, compound 1 was
recovered as white needles (4.4 g, 12.38 mmol, 83%). M.p. 1358C
(decomp); 1H NMR (300 MHz, [D6]DMSO): d=11.40 (brs 1H), 7.84–
7.58 (m, 2H), 7.47–7.12 (m, 4H), 4.25 ppm (s, 6H); 13C NMR (75 MHz,
[D6]DMSO): d=165.5, 150.2, 149.5, 140.7, 136.7, 123.7, 123.0, 119.0,
113.1, 111.2, 32.7 ppm.
LRMS: m/z (%): 527 (100) [MÀCl]+, 555 (80) [MÀCl+CN]+, 563.9 (5)
[M+H+]; HRMS: m/z calcd for C23H15N3NaClPt: 586.0500; found:
586.0519.
[CuACTHNUGRTENUNG(BisQ)]: Green needles (52%); LRMS: m/z (%): 458.04 (100)
[M+H+], 396.0 (80) [MÀ2NO3]+; HRMS: m/z calcd for
C23H15N4O363Cu: 458.0440; found: 458.0435.
Synthesis of 3: In a 250 mL flask, quinaldic acid (4.12 g, 23.79 mmol) was
dissolved in MeOH (40 mL). Concentrated H2SO4 (1 mL) was added and
the mixture was heated at reflux for 16 h. After cooling to RT, the mix-
ture was neutralized with a saturated aqueous solution of NaHCO3 and
extracted with CH2Cl2 (3ꢄ100 mL). The organic phases were combined,
dried with MgSO4, and evaporated to afford 3 as a white solid (3.97 g,
21.21 mmol, 89%). M.p. 838C (81–838C lit. [40] ); 1H NMR (300 MHz,
CDCl3): d=8.32 (d, J=8.3 Hz, 2H), 8.21 (d, J=8.5 Hz, 1H), 7.89 (d, J=
8.2 Hz, 1H), 7.80 (t, J=7.7 Hz, 1H), 7.66 (t, J=7.5 Hz, 1H), 4.09 ppm (s,
3H); 13C NMR (75 MHz, CDCl3): d=166.0, 147.9, 147.5, 137.33, 130.7,
130.3, 129.4, 128.6, 127.6, 121.0, 52.2 ppm; LRMS: m/z (%): 188.2 (100)
[M+H+].
Synthesis of 4:[40] Solid tBuOK (1.559 g, 13.89 mmol) was slowly added to
a solution of 3 (2 g, 10.68 mmol) in ethyl acetate (35 mL). The mixture
was stirred for 15 min at RT and then quenched with H2O (60 mL). The
organic layer was separated and the aqueous phase was extracted with
EtOAc (3ꢄ50 mL). The organic phases were combined, dried with
MgSO4, and evaporated. The residue was purified by flash chromatogra-
phy (EtOAc/cyclohexane, 1:4) to afford 4 as an off-white paste (2.06 g,
8.47 mmol, 79%). 1H NMR (300 MHz, CDCl3): d=8.27 (d, J=8.3 Hz,
1H), 8.16 (t, J=7.2 Hz, 2H), 7.88 (d, J=8.2 Hz, 1H), 7.79 (t, J=7.2 Hz,
1H), 7.64 (t, J=7.5 Hz, 1H), 4.36 (s, 2H), 4.22 (q, J=7.2 Hz, 2H),
1.26 ppm (t, J=7.2 Hz, 3H).
Synthesis of 2: Compound 2 was prepared from 1 as described in the lit-
erature.[26]
Synthesis of tMebip: Compound 2 (100 mg, 0.24 mmol), 4-tolylboronic
acid (32.6 mg, 0.24 mmol), potassium acetate (99 mg, 0.72 mmol), and tet-
rakis(triphenylphosphine)palladium (17.3 mg, 0.015 mmol) were added to
a degassed solution of THF (80 mL) and water (35 mL) to give a brown
solution that was heated at reflux under argon for 24 h. The reaction mix-
ture was allowed to cool slowly to room temperature. Light brown nee-
dles were filtered off. Recrystallization from THF/water afforded pure
tMebip (66.8 mg, 65% yield). M.p. 2158C (dec); 1H NMR (300 MHz,
[D6]DMSO): d=8.64 (s, 2H), 7.89 (s, 2H), 7.76 (dd, J=24.2, 7.9 Hz, 4H),
7.51–7.23 (m, 6H; Ar), 4.30 (s, 6H), 2.42 ppm (s, 3H); 13C NMR
(75 MHz, CDCl3): d=150.7, 150.7, 150.2, 142.8, 140.1, 137.4, 134.2, 130.1,
127.4, 123.8, 123.1, 123.0, 120.4, 110.2, 32.8, 21.5 ppm; HRMS: m/z calcd
for C28H23N5Na: 452.185; found: 452.1869.
[PdACHTUNGTRENNUNG
(tMebip)]: Green–yellow powder (48%); 1H NMR (300 MHz,
[D6]DMSO): d=8.43 (s, 2H), 8.17 (d, J=8.0 Hz, 2H), 7.84 (d, J=8.2 Hz,
2H), 7.59 (d, J=8.0 Hz, 2H), 7.35 (m, 6H), 4.19 (s, 6H), 2.53 ppm (s,
3H); LRMS: m/z: 571 [M+H+]; HRMS: m/z calcd for C28H23N5ClPd:
574.0676; found: 570.0504.
Synthesis of 5:[40] In a 250 mL flask, compound 4 (2.06 g, 7.95 mmol) was
dissolved in dioxane (20 mL). HCl (1m, 20 mL) was added and the solu-
tion was stirred at 1008C for 16 h, then it was concentrated under re-
duced pressure. The residual aqueous phase was extracted with EtOAc
(3ꢄ50 mL). The organic phases were combined, washed with a saturated
aqueous solution of NaHCO3, dried with MgSO4, and evaporated. The
residue was purified by chromatography on a small silica gel column to
[PtACHTUNGTRENNUNG
(tMebip)]: Dark orange powder (16%); 1H NMR (300 MHz,
[D6]DMSO): d=8.21 (s, 2H), 8.06 (d, J=7.5 Hz, 2H), 7.67–7.45 (m, J=
7.9 Hz, 4H), 7.39 (dt, J=24.9 Hz, 4H), 7.14 (d, 2H), 4.12 (s, 6H),
2.50 ppm (s, 3H); LRMS: m/z: 661.2 [M+H+]; HRMS: m/z calcd for
C28H23N535ClPt: 659.1290; found: 659.1319.
[CuACHTUNGTRENNUNG(tMebip)]: Green needles (42%); LRMS (in presence of formic
1
afford 5 as a white solid (954 mg, 5.57 mmol, 70%). M.p. 498C; H NMR
acid): m/z: 553.9 [M+H+], 537.0 [MÀNO3 +formic acid], 526.9
[MÀNO3 +2OH], 491.9 [MÀNO3]; HRMS: m/z calcd for C28H23N6O3Cu:
554.1128; found: 554.1112.
(300 MHz, CDCl3): d=8.27 (d, J=8.5 Hz, 1H), 8.20 (d, J=8.5 Hz, 1H),
8.13 (d, J=8.5 Hz, 1H), 7.87 (d, J=8.1 Hz, 1H), 7.79 (t, J=7.7 Hz, 1H),
7.65 (t, J=7.5 Hz, 1H), 2.88 (s, 3H) ppm; 13C NMR (75 MHz, CDCl3):
d=200.7, 153.2, 147.2, 136.9, 130.6, 130.0, 129.6, 128.6, 127.7, 118.0,
25.6 ppm; LRMS: m/z: 194.2 [M+Na+].
Synthesis of BisQ: Iron powder (1120 mg, 20.05 mmol) was added to a
solution o-nitrobenzaldehyde (303 mg, 2.005 mmol) in ethanol (10 mL)
followed by 0.1n aqueous hydrochloric acid (4 mL, 0.400 mmol). The re-
sulting mixture was vigorously stirred at 958C (oil bath). When TLC
analysis showed complete reduction of o-nitrobenzaldehyde (after ap-
proximately 2 h 30 min), 2,6-diacetylpyridine (135 mg, 1.003 mmol) and
potassium hydroxide (337 mg, 6.02 mmol) were added. The reaction mix-
ture was stirred at 958C for 16 h, then cooled to RT, diluted with CH2Cl2
(100 mL), and filtered through a pad of Celite. The solution was concen-
trated and the residue was purified by column chromatography on silica
gel (gradient CH2Cl2/MeOH: 100/0 to 90/10) to afford BisQ as a white
solid (250 mg, 0.750 mmol, 75%). M.p. 216–2188C; 1H NMR (300 MHz,
CDCl3): d=8.86 (d, J=8.6 Hz, 2H), 8.78 (d, J=7.8 Hz, 2H), 8.36 (d, J=
8.7 Hz, 2H), 8.22 (d, J=8.4 Hz, 2H), 8.08 (t, J=7.8 Hz, 1H), 7.90 (d, J=
8.1 Hz, 2H), 7.76 (t, J=7.5 Hz, 2H), 7.58 ppm (t, J=7.4 Hz, 2H);
13C NMR (75 MHz, CDCl3): d=156.3, 155.7, 148.1, 138.1, 136.9, 130.0,
129.7, 128.5, 127.8, 126.9, 122.2, 119.2 ppm;. LRMS: m/z: 334.2 [M+H+];
HRMS: m/z calcd for C23H16N3: 334.1344; found: 334.1350.
Synthesis of tBisQ: Compound 5 (100 mg, 0.584 mmol) was added in a
solution of 4-methylbenzaldehyde (35.1 mg, 0.292 mmol), potassium hy-
droxyde (32,8 mg, 0,584 mmol), and aqueous ammonia (28%; 0.73 mL,
33.7 mmol) in EtOH (5 mL). The mixture was heated for 24 h at 1008C
into a sealed tube. After the mixture had been cooled to RT, a white pre-
cipitate was filtered, washed with MeOH and Et2O, and dried to afford
the first batch of product. The filtrate was concentrated and the residue
was purified by column chromatography on silica gel (cyclohexane/
CH2Cl2 from 50/50 to 0/100) to afford a second batch of product. Purifi-
cation afforded tBisQ as a white powder (67 mg, 0.059 mmol, 54%). M.p.
1
244–2468C; H NMR (300 MHz, CDCl3): d=9.02 (s, 2H), 8.89 (d, J=8.6,
2H), 8.36 (d, J=8.6, 2H), 8.25 (d, J=8.4, 2H), 7.97–7.83 (m, 4H), 7.77
(t, J=7.7, 2H), 7.59 (t, J=7.5, 2H), 7.39 (d, J=7.9, 2H), 2.48 ppm (s,
3H); 13C NMR (75 MHz, CDCl3): d=156.5, 156.2, 150.4, 148.1, 139.2,
136.8, 135.9, 130.0, 129.9, 129.7, 128.5, 127.4, 126.9, 119.9, 119.5,
21.5 ppm; LRMS: m/z: 424.1 [M+H]+; HRMS: m/z calcd for
C30H21N3Na: 446.1633; found: 446.1623.
[PdACHTUNGTRENNUNG
(BisQ)]: Brown–yellow powder (84%); 1H NMR (300 MHz,
[PdACHTNUTGRNEUNG
(tBisQ)]: Yellow powder (40%); 1H NMR (300 MHz, [D6]DMSO):
[D6]DMSO): d=9.40 (s, 1H), 9.04 (s, 1H), 8.81 (d, J=16.1 Hz, 2H), 8.70
(d, J=7.6 Hz, 2H), 8.57 (d, J=7.1 Hz, 1H), 8.16 (dd, J=17.3, 8.0 Hz,
2H), 8.04 (s, 2H), 7.83 (s, 2H), 7.66 ppm (s, 2H); LRMS: m/z (%): 465
(50) [MÀCl+CN]+, 439 (100) [MÀCl]+, 474 (50) [M]+; HRMS: m/z
calcd for C23H16N3ClPd: 475.0068; found: 475.0061.
(BisQ)]: Dark orange powder (21%); 1H NMR (300 MHz,
[PtACHTUNGTRENNUNG
[D6]DMSO): d=8.90 (d, J=8.4 Hz, 2H), 8.76 (d, J=7.5 Hz, 2H), 8.62 (d,
J=8.6 Hz, 2H), 8.26 (t, J=7.4 Hz, 1H), 8.17 (d, J=8.2 Hz, 2H), 8.10 (d,
J=7.8 Hz, 2H), 7.87 (t, J=7.0 Hz, 2H), 7.70 ppm (d, J=7.1 Hz, 2H);
d=9.38 (s, 1H), 9.00–8.85 (m, 4H), 8.56 (d, J=7.2 Hz, 2H), 8.16–7.98
(m, 4H), 7.86 (t, J=8.7 Hz, 4H), 7.66 (t, J=7.5 Hz, 2H), 7.41 (d, J=
7.9 Hz, 2H), 2.43 ppm (s, 3H); LRMS (in presence of acetonitrile): m/z:
566.0 [M+H+], 456.1 [MÀCl+CN+H+]; HRMS: m/z calcd for
C30H22N3ClPd: 565.0537; found: 565.0537.
(tBisQ)]: Red powder (94%); 1H NMR (300 MHz, [D6]DMSO): d=
[PtACHTNUTGRNEUNG
9.30 (d, J=8.6 Hz, 2H), 8.81 (s, 4H), 8.64 (d, J=8.2 Hz, 2H), 7.95 (d, J=
7.5 Hz, 4H), 7.77 (dt, J=14.6, 7.0 Hz, 4H), 7.19 (d, J=7.7 Hz, 2H),
Chem. Eur. J. 2011, 17, 13274 – 13283
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
13281

M.-P. Teulade-Fichou et al.
2.34 ppm (s, 3H); LRMS (in presence of acetonitrile): m/z: 654.1 [M+H+
]; HRMS: m/z calcd for C30H21N335Cl196Pt: 654.1073; found: 654.1124.
ture plates in medium (200 mL) and treated 24 h later with compounds
dissolved in DMSO using a Biomek 3000 (Beckman) instrument. Con-
trols received the same volume of DMSO (1% final volume). After 72 h
exposure, MTS reagent (Celltiter 96Aqueous One solution, Promega)
was added and incubated for 3 h at 378C: the absorbance was monitored
at 490 nm and results were expressed as the inhibition of cell prolifera-
tion calculated as the ratio {[1-(OD490 treated/OD490 control)]ꢄ100} in
triplicate experiments. For IC50 determinations (50% inhibition of cell
proliferation), cells were incubated for 72 h by following the same proto-
col with compound concentrations ranging from 0.5 nm to 10 mm in sepa-
rate duplicate experiments. The antiproliferative activities of complexes
were determined against a panel of human tumor cell lines. First percen-
tages of inhibition of both compounds were evaluated at a concentration
[CuACHTUNGTRENNUNG
(tBisQ)]: Green needles (65%); LRMS: m/z (%): 548.1 (15) [M+H+
],486.1 (100) [MÀ2NO3]+; HRMS: m/z calcd for C30H21N363Cu: 486.1031;
found: 486.1038.
[PdACHTUNGTRENNUNG
(tpy)]: Yellow powder (99%); m.p. >2608C (decomp); 1H NMR
(300 MHz, [D6]DMSO): d=8.79–8.58 (m, 7H), 8.47 (t, J=7.8 Hz, 2H),
7.88 (t, J=6.5 Hz, 2H) ppm; LRMS (in presence of acetonitrile): m/z:
376 [M+H+], 365.0 [MÀCl+CN+H+]; HRMS: m/z calcd for
C15H11N335Cl108Pd: 375.9680; found: 375.9668.
[PdACHTUNGTRENNUNG
(ttpy)]: Orange powder (78%); m.p. >2608C (decomp); 1H NMR
(300 MHz, [D6]DMSO): d=8.95 (d, J=5.4 Hz, 2H), 8.86 (d, J=7.8 Hz,
2H), 8.71 (d, J=5.4 Hz, 2H), 8.50 (d, J=8.0 Hz, 2H), 8.13 (d, J=8.1 Hz,
2H), 7.89 (d, J=6.8 Hz, 2H), 7.49 (d, J=8.2 Hz, 2H), 2.45 (s, 3H) ppm;
LRMS (in presence of acetonitrile): m/z: 466.0 [M+H+], 455.1
[MÀCl+CN+H+]; HRMS: m/z calcd for C22H17N335ClPd: 464.0146;
found: 464.0168.
of 10À5 mÀ1 the ligand. [Pd
ligand and IC50 values were determined.
ACHTUNGTRENN(UNG ttpy)] appeared to be a much more potent
FRET melting: FRET melting assays were performed with oligonucleo-
tides that mimic the human telomeric sequence, as well as other quadru-
plex-forming oligonucleotides (25Ceb, 21CTA, c-Myc, c-Kit1, c-Kit2),
and the control duplex ds26, equipped with FRET partners at each ex-
tremity (see sequences given in the Supporting Information). Measure-
ments were made with excitation at 492 nm and detection at 516 nm in a
10 mm lithium cacodylate pH 7.2 buffer supplemented with 10 mm KCl
and 90 mm LiCl.
Acknowledgements
We gratefully acknowledge the generous financial support from the
Centre National de la Recherche Scientifique and the Institut Curie
(joint Ph.D. fellowship to E.L.), the Agence Nationale de la Recherche
(ANR) for financial support to F.H., J.L.M., and M.P.T.F. (ANR-09-
BLAN-0355 “G4Toolbox”), the Conseil rꢀgional dꢅAquitaine and Associ-
ation pour la Recherche sur le Cancer (ARC) (to J.L.M.), and the Fonds
de la Recherche Scientifique-FNRS (V.G. is a FNRS research associate;
F.R. is a FNRS scientific collaborator at the research associate level). We
sincerely thank Dr. Sophie Bombard for fruitful discussions.
ESI mass spectrometry: ESI-MS experiments were performed on a Solar-
iX 9.4T FTICR mass spectrometer (Bruker Daltonics, Bremen, DE). The
electrospray ion source was used in negative ion mode with a capillary
voltage of À3.1 kV. The source parameters were tuned so as to minimize
collisional activation (source pressure 3.1 mbar, skimmer at À20 V). The
instrument was externally calibrated with sodium iodide (<1 ppm accura-
cy). High-resolution mass spectra and comparison with theoretical isotop-
ic patterns were performed to unambiguously assign the nature of the ob-
[1] a) J. L. Huppert, S. Balasubramanian, Nucleic Acids Res. 2007, 35,
406–413; b) A. K. Todd, Methods 2007, 43, 246–251.
[2] M. Gellert, M. N. Lipsett, D. R. Davies, Proc. Natl. Acad. Sci. USA
1962, 48, 2013–2018.
[3] H. J. Lipps, D. Rhodes, Trends Cell Biol. 2009, 19, 414–422.
[4] a) S. Balasubramanian, S. Neidle, Curr. Opin. Chem. Biol. 2009, 13,
345–353; b) D. Monchaud, M.-P. Teulade-Fichou, Org. Biomol.
Chem. 2008, 6, 627–636.
served
(3644.45 Da),
complexes.
Oligodeoxynucleotides
d-CGTAAATTTACG
(3646.44 Da), d-
d-CGCGAATTCGCG
CGCGGGCCCGCG (3678.40 Da), d-TGGGGT (1863.26 Da), and d-
(GGGTTA)3GGG (6653.35 Da) were purchased from Eurogentec (An-
gleur, Belgium) and used without further purification. Duplex and quad-
ruplex solutions were prepared in ammonium acetate (150 mm) according
to previous reports.[24] The quantification of free DNA and complexes
was performed on the most abundant charge state (5À).
[5] N. H. Campbell, M. Patel, A. B. Tofa, R. Ghosh, G. N. Parkinson, S.
Neidle, Biochemistry 2009, 48, 1675–1680.
Molecular modeling: Optimized structures of OH-Pt/[Pd
ACHTUNGTREN(NUNG tMebip)] and
[6] a) S. Cosconati, L. Marinelli, R. Trotta, A. Virno, L. Mayol, E. Nov-
ellino, A. J. Olson, A. Randazzo, J. Am. Chem. Soc. 2009, 131,
16336–16337; b) S. Cosconati, L. Marinelli, R. Trotta, A. Virno, S.
De Tito, R. Romagnoli, B. Pagano, V. Limongelli, C. Giancola, P. G.
Baraldi, L. Mayol, E. Novellino, A. Randazzo, J. Am. Chem. Soc.
2010, 132, 6425–6433; c) Q. Li, J. Xiang, X. Li, L. Chen, X. Xu, Y.
Tang, Q. Zhou, L. Li, H. Zhang, H. Sun, A. Guan, Q. Yang, S. Yang,
G. Xu, Biochimie 2009, 91, 811–819.
[7] B. Rosenberg, L. vanCamp, J. E. Trosko, V. H. Mansour, Nature
1969, 222, 385–386.
[8] S. N. Georgiades, N. H. Abd Karim, K. Suntharalingam, R. Vilar,
Angew. Chem. 2010, 122, 4114–4128; Angew. Chem. Int. Ed. 2010,
49, 4020–4034.
Thymine-Pt/[Pd(tMebip)] were obtained by using DFT with the hybrid
functional B3LYP and the 6–1G(d) basis set. The LANL2DZ effective
core potential was used for Pt and Pd. The molecular model of the com-
ACHTUNGTRENNUNG
plex [PdACHTUNGTRENNUNG(tMebip)] to the quadruplex (TGGGGT)4 (Q1) was obtained as
follows: The starting structure of Q1 was obtained after 10 ns of in vacuo
molecular dynamics by using a modified AMBER parm99 force field.[41]
The ligand was manually docked to the 3’ side of the G4 on top of the
last G tetrad. One of the thymines of Q1 was coordinated to [Pd-
ACHTUNGTRENNUNG(tMebip)] and the complex was geometrically optimized prior to 1 ns mo-
lecular dynamics calculations.
UV kinetics: Absorbance spectra were recorded on an Uvikon XL Seco-
mam spectrophotometer. Absorbance spectra were recorded from 300 to
500 nm with 10 mm ligand in 10 mm lithium cacodylate pH 7.2 buffer sup-
plemented with 100 mm KCl. Local maxima were determined by using a
second derivative method with quadratic Savitsky-Golay smoothing.
After addition of 1 equiv of quadruplex-folded telomeric DNA, absorb-
ance spectra were recorded over time.
[9] H. Bertrand, D. Monchaud, A. D. Cian, R. Guillot, J.-L. Mergny,
M.-P. Teulade-Fichou, Org. Biomol. Chem. 2007, 5, 2555–2559.
[10] H. Bertrand, S. Bombard, D. Monchaud, E. Talbot, A. Guedin, J.-L.
Mergny, R. Grunert, P. J. Bednarski, M.-P. Teulade-Fichou, Org.
Biomol. Chem. 2009, 7, 2864–2871.
In vitro cytotoxicity: The human cell lines originated from ATCC, except
when otherwise stated. The human cell lines KB (nasopharynx epider-
moid carcinoma) were grown in D-MEM medium supplemented with
10% fetal calf serum (InVitrogen) in the presence of 100 UImLÀ1 peni-
cillin, 100 mgmLÀ1 streptomycin, and 1.5 mgmLÀ1 fungizone in a 75 mL
flask under 5% CO2, whereas MCF7 cells (breast adenocarcinoma) were
grown in RPMI medium. Resistant MCF7 cells were obtained by pro-
longed treatment with adriamycin. Cells were plated in 96-well tissue cul-
[11] a) E. Wong, C. M. Giandomenico, Chem. Rev. 1999, 99, 2451–2466;
b) E. R. Jamieson, S. J. Lippard, Chem. Rev. 1999, 99, 2467–2498;
c) J. Kozelka, F. Legendre, F. Reeder, J.-C. Chottard, Coord. Chem.
Rev. 1999, 190–192, 61–82.
[12] I. Eryazici, C. N. Moorefield, G. R. Newkome, Chem. Rev. 2008, 108,
1834–1895.
[13] J. L. Butour, S. Wimmer, F. Wimmer, P. Castan, Chem.-Biol. Interact.
1997, 104, 165–178.
13282
www.chemeurj.org
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 13274 – 13283

Palladium(II) Complexes Coordinating Quadruplex DNA
FULL PAPER
[14] a) M. Cusumano, A. Giannetto, J. Inorg. Biochem. 1997, 65, 137–
144; b) Y. Ma, C. S. Day, U. Bierbach, J. Inorg. Biochem. 2005, 99,
2013–2023.
[15] a) M. Shoukry, H. Hohmann, R. van Eldik, Inorg. Chim. Acta 1992,
198–200, 187–192; b) A. Garoufis, S. K. Hadjikakou, N. Hadjiliadis,
Coord. Chem. Rev. 2009, 253, 1384–1397.
[28] a) J. R. B. Gomes, J. Phys. Chem. A 2009, 113, 1628–1634; b) L.
Ukhin, E. Belov, Russ. Chem. Bull. 2008, 57, 418–421.
[29] A.-H. Li, E. Ahmed, X. Chen, M. Cox, A. P. Crew, H.-Q. Dong, M.
Jin, L. Ma, B. Panicker, K. W. Siu, A. G. Steinig, K. M. Stolz,
P. A. R. Tavares, B. Volk, Q. Weng, D. Werner, M. J. Mulvihill, Org.
Biomol. Chem. 2007, 5, 61–64.
[16] J. V. Burda, M. Zeizinger, J. Leszczynski, J. Chem. Phys. 2004, 120,
1253–1262.
[17] A. Garoufis, S. K. Hadjikakou, N. Hadjiliadis in Metallotherapeutic
Drugs and Metal-Based Diagnostic Agents: The Use of Metals in
Medicine (Eds.: M. Gielen, E. R. T. Tiekink), Wiley, New York,
2005, pp. 399–419.
[30] F. Krçhnke, Synthesis 1976, 1976, 1–24.
[31] a) S. D. Cummings, Coord. Chem. Rev. 2009, 253, 449–478; b) Q.-Z.
Yang, L.-Z. Wu, Z.-X. Wu, L.-P. Zhang, C.-H. Tung, Inorg. Chem.
2002, 41, 5653–5655; c) X. Liu, E. J. L. McInnes, C. A. Kilner, M.
Thornton-Pett, M. A. Halcrow, Polyhedron 2001, 20, 2889–2900.
[32] S. Redon, S. Bombard, M.-A. Elizondo-Riojas, J.-C. Chottard, Bio-
chemistry 2001, 40, 8463–8470.
[33] a) J. Duguid, V. A. Bloomfield, J. Benevides, G. J. Thomas, Biophys.
J. 1993, 65, 1916–1928; b) J. Ruiz, J. Lorenzo, L. Sanglas, N. Cutillas,
C. Vicente, M. D. Villa, F. X. Avilꢀs, G. Lꢆpez, V. Moreno, J. Pꢀrez,
D. Bautista, Inorg. Chem. 2006, 45, 6347–6360; c) A. Messere, E.
Fabbri, M. Borgatti, R. Gambari, B. D. Blasio, C. Pedone, A. Roma-
nelli, J. Inorg. Biochem. 2007, 101, 254–260.
[34] a) M.-H. Baik, R. A. Friesner, S. J. Lippard, J. Am. Chem. Soc. 2003,
125, 14082–14092; b) J. K. Barton in Bioinorganic Chemistry (Eds.:
I. Bertini, H. B. Gray, S. J. Lippard, J. S. Valentine), University Press,
Mill Valley, 1994, pp. 455–504.
[35] C. S. Peyratout, T. K. Aldridge, D. K. Crites, D. R. McMillin, Inorg.
Chem. 1995, 34, 4484–4489.
[18] J. E. Reed, A. J. P. White, S. Neidle, R. Vilar, Dalton Trans. 2009,
2558–2568.
[19] T. Yutaka, S. Obara, S. Ogawa, K. Nozaki, N. Ikeda, T. Ohno, Y.
Ishii, K. Sakai, M.-a. Haga, Inorg. Chem. 2005, 44, 4737–4746.
[20] B. M. OꢅNeill, J. E. Ratto, K. L. Good, D. C. Tahmassebi, S. A. Hel-
quist, J. C. Morales, E. T. Kool, J. Org. Chem. 2002, 67, 5869–5875.
[21] a) G. Li, J. Huang, M. Zhang, Y. Zhou, D. Zhang, Z. Wu, S. Wang,
X. Weng, X. Zhou, G. Yang, Chem. Commun. 2008, 4564–4566;
b) J. Huang, G. Li, Z. Wu, Z. Song, Y. Zhou, L. Shuai, X. Weng, X.
Zhou, G. Yang, Chem. Commun. 2009, 902–904; c) A. K. Jain, V. V.
Reddy, A. Paul, M. K, S. Bhattacharya, Biochemistry 2009, 48,
10693–10704; d) S. Bhattacharya, P. Chaudhuri, A. K. Jain, A. Paul,
Bioconjugate Chem. 2010, 21, 1148–1159.
[22] P. Wang, C.-H. Leung, D.-L. Ma, S.-C. Yan, C.-M. Che, Chem. Eur.
J. 2010, 16, 6900–6911.
[36] U. S. Schubert, H. Hofmeier, G. R. Newkome, Modern Terpyridine
Chemistry, Wiley-VCH, Weinheim, 2006, pp. 37–68.
[23] a) J. L. Mergny, J. C. Maurizot, ChemBioChem 2001, 2, 124–132;
b) A. De Cian, L. Guittat, M. Kaiser, B. Sacca, S. Amrane, A. Bour-
doncle, P. Alberti, M. P. Teulade-Fichou, L. Lacroix, J. L. Mergny,
Methods 2007, 42, 183–195.
[24] F. Rosu, E. De Pauw, V. Gabelica, Biochimie 2008, 90, 1074–1087.
[25] P. Froidevaux, J. M. Harrowfield, A. N. Sobolev, Inorg. Chem. 2000,
39, 4678–4687.
[37] W. Zhang, C. Bensimon, R. J. Crutchley, Inorg. Chem. 1993, 32,
5808–5812.
[38] D. Chatterjee, J. Chem. Sci. 1999, 111, 437–442.
[39] R. A. Reinhardt, J. S. Coe, Inorg. Chim. Acta 1969, 3, 438–440.
[40] B. K. Chan, M. A. Ciufolini, J. Org. Chem. 2007, 72, 8489–8495.
[41] V. Gabelica, F. Rosu, M. Witt, G. Baykut, E. De Pauw, Rapid
Commun. Mass Spectrom. 2005, 19, 201–208.
[26] P. K. Iyer, J. B. Beck, C. Weder, S. J. Rowan, Chem. Commun. 2005,
319–321.
[27] D. Knapton, M. Burnworth, S. J. Rowan, C. Weder, Angew. Chem.
2006, 118, 5957–5961; Angew. Chem. Int. Ed. 2006, 45, 5825–5829.
Received: July 26, 2011
Published online: October 18, 2011
Chem. Eur. J. 2011, 17, 13274 – 13283
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
13283