Tuning G-quadruplex vs double-stranded DNA recognition in regioisomeric lysyl-peptidyl-anthraquinone conjugates

Article abstract of DOI:10.1021/bc200389w

Anthraquinone is a versatile scaffold to provide effective DNA binders. This planar system can be easily conjugated to protonable side chains: the nature of the lateral groups and their positions around the tricyclic moiety largely affect the DNA recognition process in terms of binding affinity and mode, as well as sequence and structure of the target nucleic acid. Starting from an anthracenedione system symmetrically functionalized with N-terminal lysyl residues, we incremented the length of side chains by introducing a Gly, Ala, or Phe spacer, characterized by different flexibility, lipophilicity, and bulkiness. Moreover, 2,6, 2,7, 1,8, and 1,5 regioisomers were examined to yield a small bis(lysyl-peptidyl) anthracenedione library. By merging spectroscopic, enzymatic, and cellular results, we showed that the proper combination of a basic aminoacid (Lys) with a more hydrophobic residue (Phe) can provide selective G-quadruplex recognition, in particular when side chains are located at positions 2,6 or 2,7. In fact, while these derivatives effectively bind G-quadruplex structures, they behave at the same time as rather poor double-stranded DNA intercalators. As a result, the Lys-Phe substituted anthraquinones are poorly cytotoxic but still able to promote a senescence mechanism in cancer cells. This combination of chemical and biological properties foresees potentially valuable applications in anticancer medicinal chemistry.

Full text of DOI:10.1021/bc200389w

ARTICLE
pubs.acs.org/bc
Tuning G-Quadruplex vs Double-Stranded DNA Recognition in
Regioisomeric Lysyl-Peptidyl-Anthraquinone Conjugates
Giuseppe Zagotto, Antonio Ricci, Elena Vasquez, Andrea Sandoli, Silvia Benedetti, Manlio Palumbo, and
Claudia Sissi*
Department of Pharmaceutical Sciences, University of Padova, Via Marzolo 5, 35131 Padova, Italy
S Supporting Information
b
ABSTRACT: Anthraquinone is a versatile scaffold to provide
effective DNA binders. This planar system can be easily con-
jugated to protonable side chains: the nature of the lateral
groups and their positions around the tricyclic moiety largely
affect the DNA recognition process in terms of binding affinity
and mode, as well as sequence and structure of the target nucleic
acid. Starting from an anthracenedione system symmetrically
functionalized with N-terminal lysyl residues, we incremented
the length of side chains by introducing a Gly, Ala, or Phe
spacer, characterized by different flexibility, lipophilicity, and
bulkiness. Moreover, 2,6, 2,7, 1,8, and 1,5 regioisomers were examined to yield a small bis(lysyl-peptidyl) anthracenedione library.
By merging spectroscopic, enzymatic, and cellular results, we showed that the proper combination of a basic aminoacid (Lys) with a
more hydrophobic residue (Phe) can provide selective G-quadruplex recognition, in particular when side chains are located at
positions 2,6 or 2,7. In fact, while these derivatives effectively bind G-quadruplex structures, they behave at the same time as rather
poor double-stranded DNA intercalators. As a result, the Lys-Phe substituted anthraquinones are poorly cytotoxic but still able to
promote a senescence mechanism in cancer cells. This combination of chemical and biological properties foresees potentially
valuable applications in anticancer medicinal chemistry.
’ INTRODUCTION
which is inaccessible to the RNA template component of
telomerase.8
Telomeres are specialized nucleoprotein structures localized
at the end of chromosomes. They protect genomic information
from degradation, fusion, and recombination and allow the
DNA damage response machinery to discriminate between
double-strand breaks and natural chromosome ends.1 Telomeric
DNA comprises several kilobases of tandem repeats of a
G-rich DNA sequence (TTAGGG in humans) and are char-
acterized by a single-stranded 30 end (ca. 150ꢀ200 nucleotides).
However, its length is progressively reduced in somatic cells
due to the “end replication” problem.2 When telomeres reach a
critical length, cells enter a replicative senescence state and
division no longer occurs.3 On the contrary, over 80% of cancer
cells maintain telomere integrity due to an up-regulation of a
reverse transcriptase called telomerase.4 This enzyme catalyzes
the synthesis of telomeric DNA repeats, stabilizing telomeres
length and contributing to cellular immortalization and
oncogenesis.5
Due to its involvement in cell immortalization, telomerase
represents a promising anticancer target.6 Various strategies
have been devised to directly target the enzyme or associated
proteins. An alternative approach comprises derivatives directed
toward the 30 single-stranded end of telomeric DNA.7 In this
case, small molecules are used to induce nucleic acid folding
into a noncanonical structure called the G-quadruplex,
In the past decades, a large number of small molecules have
been reported to efficiently bind G-quadruplex DNA and to
inhibit telomerase.9ꢀ11 Most of them are based on a polycyclic
aromatic core that stacks by πꢀπ interactions over the terminal
planar G-tetrads of a quadruplex. In addition, cationic side chains
can interact with the negatively charged phosphate backbones in
G-quadruplexes.
Properly substituted anthraquinones were early identified as
potential effective G-quadruplex binders.12ꢀ14 In particular, the
substitution pattern with protonable side chains was shown to
play a major role in driving selectivity toward double-stranded or
G-quadruplex structures. In previous studies, we reported the
modulation of G-quadruplex recognition by anthracenedione
derivatives symmetrically disubstituted at 2,6 or 2,7 positions
with one or two aminoacidic residues.15,16 We showed the impor-
tance of the size of the planar surface, which was affected by rever-
sal of the direction of the amide linkage connecting the anthra-
cenedione to the side chains. Additionally, a significant role of
ionic interactions was confirmed, since N-terminal protonable
Received: July 19, 2011
Revised:
August 26, 2011
Published: September 12, 2011
r
2011 American Chemical Society
2126
dx.doi.org/10.1021/bc200389w Bioconjugate Chem. 2011, 22, 2126–2135
|

Bioconjugate Chemistry
ARTICLE
aminoacids (Lys, Arg) largely incremented G-quadruplex binding.
Clearly, charged interactions simultaneously promote binding to
double-stranded DNA too. However, taking into account the
different geometries of the two DNA structures, it should be
possible to optimize the relative position of the charged groups
with reference to the aromatic core to preferentially drive the
ligand toward one DNA fold.
Here, we synthesized and analyzed a small library of
disubstituted peptidyl anthraquinones characterized by a lysyl
N-terminal residue. We modulated its distance from the planar
moiety by introducing mono- and dipeptidyl spacers, character-
ized by increasing steric hindrance and hydrophobicity (Gly, Ala,
Phe). Additionally, the relative position of the two protonable
groups was varied by considering the 2,6, 2,7, 1,8, and 1,5
regioisomers. These structural features should allow identifying
the peptidyl anthraquinone requirements that grant selectivity
for G-quadruplex vs duplex DNA by establishing distinct patterns
of hydrogen bonding, hydrophobic, and charge interactions. In
particular, preferential binding to G-quadruplex folded telomeric
DNA would likely produce telomerase inhibition and cancer cell
death without greatly affecting somatic cells.
pouring it into cold Et2O (250 mL). The resulting precipitate was
collected by centrifugation, washed with HCl 0.1 M (3 ꢁ 50 mL),
isopropanol (20 mL), and Et2O (50 mL), and dried under
vacuum.
General Deprotection Procedure of the Fmoc Group.
A solution of the desired AA-AQ derivative (10 mmol) in
piperidine/DMF 10% (1 mL) was stirred at RT for 20 min and
then poured into cold Et2O (20 mL). The resulting precipitate
was collected by centrifugation, washed with Et2O (3 ꢁ 20 mL),
MeOH (1 ꢁ 20 mL), and Et2O (2 ꢁ 30 mL), and finally dried
under vacuum.
General Procedure for Peptide Coupling via TBTU, HOBt,
andDIPEA.A solutionofthe desired Fmoc-aminoacid(40mmol)
in anhydrous DMF (5 mL) was introduced in a round-bottomed
flask containing TBTU (42 mmol), HOBt (42 mmol), and
DIPEA (128 mmol); the resulting solution was stirred for
15 min at RT under N2 atmosphere. A solution of the desired
AA-AQ (10 mmol) in anhydrous DMF (5 mL) was added and
the mixture stirred at RT for 4 h under N2 atmosphere. The
mixture was poured in Et2O (50 mL), and the resulting
precipitate was collected by centrifugation, washed with Et2O
(3 ꢁ 50 mL), MeOH (2 ꢁ 10 mL), and Et2O (50 mL), and
finally dried under vacuum.
General Deprotection Procedure of the Boc Group. A
solution of the desired Boc-AA-AQ (10 mmol) in TFA/H2O
95% (5 mL) was stirred at RT for 20 min and then poured
into Et2O (30 mL). The resulting precipitate was collected by
centrifugation, washed with Et2O (3 ꢁ 30 mL), and dried under
vacuum.
Synthetic Oligonucleotides. The labeled human telomeric
sequence l-Tel23 50-Dabcyl-AGGGTTAGGGTTAGGGT-
TAGGGT-FAM 30 was synthesized and purified by ATDBIO
(Southampton, UK). All other oligonucleotides were provided
by Eurogentec (Belgium) (Tel22, AGGGTTAGGGTTAGGGT-
TAGGG; dsUP, 50-FAM- GTGAGATACCGACAGAAG;
dsDOWN CTTCTGTCGGTATCTCAC-Dabcyl 30; 1GGG,
TACAGATAGTTAGGGTTAGACTTA; 2GGG, TACAGATAG-
TTAGGGTTAGGGTTA; TS AATCCGTCGAGCAGAGTT;
ACX: GTGCCCTTACCCTTACCCTTACCCTAA; TSNT:
AATCCGTCGAGCAGAGTTAAAAGGCCGAGAAGCGAT;
NT: ATCGCTTCTCGGCCTTTT; Tup:TGAGGATCCGC-
CTGGACAGCATGG-30; Tdown: GTCGAATTCTCGGCGA-
GAAGCAGG).
Calf thymus DNA (ctDNA) was purchased from Sigma
Aldrich (USA) and used with no further purification.
Fluorescence Melting Assay. Fluorescence melting curves
were determined in a Roche LightCycler (λecc 488 nm, λem
520 nm) in a total reaction volume of 20 μL containing 0.25 μM
of l-Tel23 and increasing ligands concentrations in LiP buffer
(10 mM LiOH, 50 mM KCl, pH 7.4 with H3PO4). In a typical
experiment, the oligonucleotide was first denatured by heating to
95° at a rate of 0.1 °C sꢀ1, maintained at 95 °C for 5 min, and
then annealed by cooling to 30 °C at a rate of 0.1 °C sꢀ1. Then,
samples were maintained at 30° for 5 min before being slowly
heated to 95 °C (1 °C/min) and annealed at a rate of 1 °C/min.
When double-stranded DNA was used, dsUP and dsDOWN were
mixed at equimolar concentrations, heated to 95 °C for 5 min,
and then cooled to room temperature overnight before use.
Recordings were taken during both the annealing and melt-
ing steps.
’ EXPERIMENTAL PROCEDURES
Synthesis. All water- or air-sensitive reactions were carried out
in oven-dried glassware under a nitrogen atmosphere with dry
solvents, unless otherwise noted. All commercial reagents were
purchased from Sigma-Aldrich and were used without further
purification. The solvents were purchased from Fluka, WMC,
and Carlo Erba; they were anhydrified as follows: N,N-Dimethyl-
formamide (DMF) was distilled under reduced pressure over
BaO; diisopropylethylamine (DIPEA) and piperidine were dis-
1
tilled over CaH2. H NMR spectra were recorded on a Bruker
Avance AMX 300 spectrometer. Chemical shifts (δ) are expressed
in parts per million relative to tetramethylsilane (ppm), and the
following abbreviations were used for the multiplicities: s =
singlet, d = doublet, dd = doublet doublets, t = triplet, dt =
doublet triplets, tt = triplet triplets, q = quartet, m = multiplet,
bm = broad multiplet, and bs = broad singlet. Mass spectrometric
data were obtained using a Mariner API-TOF (Perseptive Bio-
systems Inc., Framingham, MA 01701, USA). Purities (g95%)
of all tested compounds were established by HPLC using a
Varian Pro Star equipped with an UVꢀvis detector (BioRad) and
a Phenomenex C18 (4.6 ꢁ 250 mm, 5 μm) column. The mobile
phase results from combination of (A) H2O/TFA (99:1) and
(B) MeOH/TFA (99:1). A gradient from 5% to 60% of solution
B over 20 min followed by 5 min at the latter percentage was
used; the flow rate was set at 1.0 mL/min. The signals were
detected with a UVꢀvis detector at 254 nm. Within the experi-
mental part, AQ stands for anthraquinone; in particular, AA-AQ
stands for amino acid substituted anthraquinone.
General Procedure for the Synthesis of the Acyl Chloride
Derivative of an Amino Acid. SOCl2 (1 mol) was added to a
solution of the desired Fmoc-amino acid (100 mmol) in 50 mL of
DCM and heated at reflux under N2 atmosphere for 4 h. The
solvent was then removed by distillation and the crude white
solid dried under high vacuum.
General Procedure for Peptide Coupling via Acyl Chlor-
ide. The acyl chloride derivative of the desired Fmoc-amino acid
(100 mmol) was added to a solution of the required diamino-
anthraquinone (25 mmol) in 15 mL of DMF. The solution thus
obtained was stirred at RT under N2 atmosphere for 3 h before
Tm values were determined from the first derivatives of the
melting profiles using the Roche LightCycler software. Each curve
2127
dx.doi.org/10.1021/bc200389w |Bioconjugate Chem. 2011, 22, 2126–2135

Bioconjugate Chemistry
ARTICLE
Scheme 1a
a Reagents: (i) AA1-Cl, DMF, RT, N2, 3 h. (ii) Piperidine/DMF 20%, RT, 20 min. (iii) AA2,3,4, TBTU, HOBt, DIPEA, RT, N2, DMF, 4 h. (iv) TFA/H2O
90%, RT, 20 min.
was repeated at least three times and errors were (0.4 °C.
ΔTm were calculated by subtracting the Tm value recorded in the
presence of the ligand from the corresponding value in the abse-
nce of ligand.
Circular Dichroism Measurements. Circular dichroism
spectra from 220 to 450 nm were recorded using 10 mm path
length cells on a Jasco J 810 spectropolarimeter equipped with a
NESLAB temperature controller and interfaced to a PC 100 in
10 mM Tris-HCl, 50 mM KCl pH 7.4. Before data acquisition,
Tel22 solutions (4 μM) were heated at 95 °C for 5 min and
left to cool overnight at room temperature. The reported
spectrum of each sample represents the average of 3 scans
recorded with 1 nm step resolution. Observed ellipticities were
converted to mean residue ellipticity [θ] = deg ꢁ cm2 ꢁ d molꢀ1
(Mol. Ellip.).
Electromobility Shift Assay. End-labeled single-stranded
oligonucleotides were obtained by incubating the oligonucleo-
tides with T4 polynucleotide Kinase and [γ-32P] ATP for 30 min
at 37 °C. The enzyme was then removed by extraction with
phenol/chloroform/isoamylic alcohol (25:24:1). A mixture of
purified labeled and unlabeled oligonucleotides (total final con-
centration 10 μM) was heated to 95 °C for 5 min in 10 mM Tris,
1 mM EDTA, 100 mM KCl, pH 8.0 buffer and allowed to cool
overnight at room temperature. The folding of the starting
material was monitored by native 16% polyacrylamide gel elec-
trophoresis in 0.5ꢁ TBE (44.5 mM Tris base, 44.5 mM boric
2128
dx.doi.org/10.1021/bc200389w |Bioconjugate Chem. 2011, 22, 2126–2135

Bioconjugate Chemistry
ARTICLE
acid and 1 mM Na2EDTA) added of KCl (100 mM in buffer and
20 mM in gels).
Scheme 2
Resolved bands on dried gels were visualized and quantified on
a PhosphorImager (Amersham).
Taq Polymerase Assay. Compounds were assayed with Taq
polymerase reaction by using pBR322 (2.5 ng) as DNA template
and appropriate primer sequences (Tup and Tdown, 0.5 μM each)
to amplify the 964ꢀ1064 plasmid sequence by PCR. The
reaction was carried out in a Perkin-Elmer thermocycler per-
forming 25 cycles of 30 s at 94 °C, 30 s at 65 °C, and 30 s at 72 °C.
The reaction products were resolved on a 2% agarose gel in 1ꢁ
TBE (89 mM Tris base, 89 mM boric acid, and 2 mM Na2EDTA)
and stained by ethidium bromide.
Table 1. Schematic Structures of the Synthesized 9,10-An-
thracenedione-Aminoacyl Conjugates
side chains
Telomerase Activity Assay (TRAP Assay). Telomerase activ-
ity was assayed using a modified telomere repeat amplification
protocol (TRAP) assay. As a source of telomerase, the total cell
lysates derived from human melanoma cell line HeLa were used.
Protein concentration of the lysates was assayed using Bio-Rad
protein assay kit using BSA as standard. A 50 μL TRAP reaction
mix (50 μM of dNTPs, 0.2 μg of primer TS, 0.1 μg of return
primer ACX, 500 ng of protein extract, 2 U Taq polymerase) was
prepared in the presence/absence of increasing drug concentra-
tion in 20 mM Tris-HCl, pH 8.3, 68 mM KCl, 1.5 mM MgCl2,
1 mM EGTA, 0.05% v/v Tween-20. An internal control template
(0.01 mmol TSNT) with its return primer (1 ng NT) were added
to the reaction mixture. Then, the telomerase elongation step has
been performed (30 min at 30 °C) followed by a PCR amplifica-
tion step (30 cycles of 30 s at 37 °C and 30 s at 58 °C). Telo-
merase products were resolved by 10% polyacrylamide gel
electrophoresis in TBE 0.5ꢁ and visualized by staining with
Sybr Green I.
MTT Assays. HeLa (human hepitelial) cell line was maintai-
ned in DMEM medium supplemented with 10% heat-inactivated
fetal calf serum, 50 U/mL of penicillin G, and 50 μg/mL of
streptomycin, at 37 °C in humidified atmosphere and 5% of CO2.
To evaluate toxic profiles of the potential antitelomerase
compounds, MTT assays were performed as described: cells
were plated in 96 well plates at 10 000 cells/well, and cultured
overnight. Afterward, compounds were added in triplicate and
plates were incubated in the presence of the drug for 72 h. At the
end of this period, MTT was added to a final concentration
of 0.8 mg/mL, and two additional hours of incubation were
performed. After that, medium was aspirated carefully and 150
μL of DMSO were added per well. Soluble formazan salts were
homogenized by manual pipetting and absorbance at 540 nm
was read. Curves consisted of 8 serial dilutions in triplicate in
each case, and results were analyzed as sigmoidal doseꢀresponse
curves.
Cell Senescence Assay. HeLa cells were plated in 6-well
plates at 10 000 cells/well and cultured overnight. Afterward,
compounds were added in triplicate (2.5 μM and 0.5 μM for
C-16 and D-16, respectively) and plates were incubated in the
presence of the drug for 240 h. Every 72 h, the medium was
changed and new drug was added. At the end of the treatment,
medium was removed and cells were washed once with 1 mL of
PBS and fixed with 1 mL of 2% formaldehyde, 0.2% glutaralde-
hyde in PBS for 15 min at room temperature. After another
washing step with PBS, cells were stained with 1 mL of 37.2 mM
citric acid/sodium phosphate (pH 6.0), 140 mM NaCl, 1.8 mM
MgCl2, 5mMK3Fe(CN)6, 5mMK4Fe(CN)6 3H2O, and 1 mg/mL
X-gal in dimethylformamide. Plates were incubated overnight at
37 °C and then photographed with a light microscope.
regioisomer
2,6
compound
AA1
AA2
AA3
A-13
A-14
A-15
A-16
B-13
B-14
B-15
B-16
C-13
C-14
C-16
D-13
D-14
D-16
E-13
E-14
E-16
Gly
Lys
Ala
Gly
Phe
Lys
Ala
Gly
Phe
Lys
Ala
Phe
Lys
Ala
Phe
Lys
Ala
Phe
Gly
Lys
Lys
Lys
Gly
Gly
2,6
βAla
βAla
βAla
βAla
βAla
βAla
βAla
βAla
βAla
βAla
βAla
βAla
βAla
Lys
Lys
Lys
2,7
1,5
1,8
Lys
Lys
Lys
Lys
Lys
Lys
’ RESULTS
Chemistry. Starting materials for the synthesis of all deriva-
tives were the appropriate diamino-substituted anthraquinones
(AQ). While 2,6-diamino and 1,5-diamino AQ are commercially
available, the 2,7-diamino and 1,8-diamino AQ were obtained via
reduction of the corresponding dinitro derivatives.17 In particu-
lar, the 2,7-dinitro AQ scaffold was obtained through a one-pot
nitrationꢀoxidation of commercial anthrone, as previously
described.16 The common synthetic route for all the disubstitut-
ed AQs is summarized in Scheme 1. Due to the low reactivity
of the aryl amino groups, the first aminoacidic coupling (AA1)
was efficiently carried out using the appropriate acyl chloride
derivative (AA-Cl). These substrates were prepared treating
the selected AA (Fmoc-glycine or Fmoc-β-alanine) with an
excess of thionyl chloride in CH2Cl2 (Scheme 2). A first Fmoc
deprotection of the amino groups, followed by salification with
trifluoroacetic acid (TFA) led to the first group of compounds
(A-2, B-2, C-2, D-2, and E-2). The salification step was
performed to favor the solubility of the anthraquinone derivatives
in water.
These compounds were used as starting blocks to synthesize
several derivatives which can be divided into 5 related families: A
and B where a glycine or a β-alanine was the AA1 at positions 2
and 6, respectively; C, D, and E which correspond to the 2,7, 1,5,
and 1,8 regioisomers of compound B.
2129
dx.doi.org/10.1021/bc200389w |Bioconjugate Chem. 2011, 22, 2126–2135

Bioconjugate Chemistry
ARTICLE
Figure 1. Variation of the melting temperature (ΔTm) of a G-quadruplex folded telomeric sequence (A, 0.25 μM) or double-stranded DNA
(B, 0.25 μM) upon addition of 1 μM selected anthraquinones in 50 mM KCl. Errors were (0.4 °C.
The novel derivatives were prepared by standard peptide
coupling of the free amino AA1-AQ and the desired Fmoc- or
Boc-protected AA2 derivatives using TBTU, HOBt, and DIPEA as
coupling reagents (Scheme 2). In the first case, basic deprotection
led to the free amino derivatives, which were coupled with a third
amino acid or salified with TFA to afford more water-soluble
derivatives. Inthe case of the Boc-protected compounds, the acidic
condition used to remove the Boc protective group directly led the
desired trifluoroacetic salts derivatives. The side chains composi-
tion of the synthesized anthraquinones is reported in Table 1.
Fluorescence Melting Assays. All test derivatives were
evaluated for their ability to bind and stabilize G-quadruplex
structures by fluorescence melting measurements.
We used an oligonucleotide corresponding to the human
telomeric sequence labeled at the 50-end with a quencher
(dabcyl) and at the 30-end with a fluorophore (fluorescein).
When the oligonucleotide folds into an intramolecular G-quad-
ruplex, these groups are in close proximity, and as a consequence,
the fluorescence is quenched. When DNA melts, the strands fall
far apart and a large increase in fluorescence occurs. This system
allows us to easily monitor DNA folding/unfolding processes
upon temperature variation.18,19 Drug binding to DNA affects
nucleic acid stability, hence its melting temperature (Tm). Since a
correlation exists between the observed Tm shift (ΔTm) and the
DNA binding affinity of structurally related binders, this allows us
to build a relative G-quadruplex affinity ranking order.20 Thus,
the melting profile of the oligonucleotide (Tm in the absence of
ligand 57.2 °C) was recorded in the presence of increasing
anthracenedione concentration (0ꢀ20 μM, Supporting In-
formation). The ΔTm values determined for all derivatives at 1
μM concentration are included in Table 3. The behavior of the
test ligands toward telomeric G-quadruplex is generally con-
served among the 5 tested families (Figure 1A). In particular, it
emerged that the introduction of an amino acid between AA1 and
the charged residue sensibly reduces the G-quadruplex stabilizing
efficiency. This effect is more significant when shifting from Gly
or Ala to become maximal with Phe. This suggests that the flexi-
bility and bulkiness of the linker is relevant for G-quadruplex
interaction, likely favoring the proper localization of the positive
charge on the polyanionic target.
Fluorescence melting analysis was used also to evaluate the
preference of the new derivatives for G-quadruplex structure vs
duplex DNA arrangement by using a random double-stranded
sequence (50-GTGAGATACCGACAGAAG, Tm 52.8 °C). The
ΔTm value determined for all derivatives at 1 μM concentration
are included in Table 3 and plotted in Figure 1B. All tested com-
pounds stabilize double-stranded DNA to a considerably lesser
extent than G-quadruplex. Interestingly, dsDNA recognition is
more prominent for derivatives of the D and E families that bear
substituents at positions 1,5 or 1,8. Comparison of melting data
on different DNA folding underlines that for derivatives contain-
ing the Phe-Lys dipeptide (compounds with side chain -16)
relevant thermal stabilization occurs only with the G-quadruplex
structure, which suggests remarkable selectivity in the drugꢀ
DNA recognition process.
For a direct comparison of the ligand-mediated DNA stabili-
zation on the two tested folds, melting profiles of the telomeric
sequence were recorded in the presence of variable amounts
of its complementary strand (Figure 2).21 The results confirmed
that the Phe-Lys containing analogues are the most selective
for G-quadruplex structures. Indeed, they are the only
derivatives able to increment the melting temperature of the
quadruplex form when the double-stranded arrangement
is highly favored (presence of an excess of complementary
C-rich strand).
Circular Dichroism. The peculiar signature of the CD spec-
trum of the G-quadruplex folded telomeric sequence Tel22 is
well-characterized. In potassium-containing buffer, it shows a
positive band at 295 nm with a shoulder at 260 nm, which
corresponds to a mixed type (antiparallel and parallel) strand
arrangement.22 Tel22 dichroic response was extensively mod-
ified in the presence of increasing concentrations of our ligands,
thus confirming that they are effective binders (Figure 3). In
particular, CD titrations confirmed a strong binding affinity for
all tested compounds to G-quadruplex: saturation occurred at
comparable drug/DNA ratio and indicated the binding of two
anthraquinones per G-quadruplex unit (Figure 3C). In analogy
to most of the known G-quadruplex interacting agents, this
supports the stacking of one ligand on both terminal tetrads of
the G-quadruplex.
2130
dx.doi.org/10.1021/bc200389w |Bioconjugate Chem. 2011, 22, 2126–2135

Bioconjugate Chemistry
ARTICLE
Figure 2. First derivatives of the melting profiles of a G-quadruplex folded telomeric sequence (0.25 μM) recorded in the presence of an excess
(20-fold) of its complementary strand in 50 mM KCl. Arrows 1 and 2 indicate the shifts of the duplex and quadruplex melting transitions, respectively.
Panels A and B refer to addition of C-16 and A-13, respectively.
Figure 3. Circular dichroism spectra of Tel22 (4 μM) recorded in the presence/absence of anthraquinone derivatives in 10 mM Tris, 50 mM KCl, pH
7.4, 25 °C. Arrows indicate the changes induced by increasing concentrations of C-13 (A) and A-16 (B). The variation of the fraction of bound Tel22
upon increasing ligand/DNA ratio (r) determined by CD titrations is described in panel C.
Generally, an increment in the positive band at 295 nm is
observed suggesting a stabilization of the antiparallel components
of the G-quadruplex nucleic acid structure. Interestingly, the Phe-
Lys substituted derivatives (compounds with side chain ꢀ16)
induced the most notable variations at 260 nm. Since the tested
derivatives showed comparable absorption spectra and, addition-
ally, no significant induced CD in the drug absorption range,
we are confident in assigning the observed differences primarily
to slightly different binding modes.
we still found odd behaviors for derivatives containing the Phe-Lys
sequence (Figure 4B). In general (see, for example, the series -13,
Figure 4A), an increment of the DNA dichroic intensity at
275 nm occurred which sustains intercalation into the double
helix. On the contrary, all Phe-Lys containing derivatives showed
a reduction in the DNA band intensities at both 275 and 245 nm
(Figure 4B). Although this confirms a DNAꢀligand interaction,
it suggests that intercalation is probably not the principal binding
mechanism. Generally, DNA intercalation of anthraquinone
derivatives disubstituted at positions 1,5, 2,6, and 2,7 occurs by
a threading mechanism with the planar system being oriented
more or less perpendicularly to the base pair longest dimension
When similar measurements were performed using a double-
stranded DNA (we used calf thymus DNA, ctDNA, due to its
similarity to the human nucleic acid composition) as the target,
2131
dx.doi.org/10.1021/bc200389w |Bioconjugate Chem. 2011, 22, 2126–2135

Bioconjugate Chemistry
ARTICLE
Figure 4. Circular dichroism spectra of ctDNA (88 μM) recorded in the presence of increasing concentrations of anthraquinone derivatives in 10 mM
Tris, 50 mM KCl, pH 7.4, 25 °C.
Figure 5. Induction of G-quadruplex structure by selected ligands in 10 mM Tris, 50 mM KCl, pH 7.4, 25 °C. Panel A: variation of the electrophoretic
mobility of the unfolded oligonucleotide 2GGG (10 μM) upon addition of D-16 due to its conversion into a dimeric G-quadruplex.
A marker (L) for the monomeric (M), dimeric (D), or tetrameric (T) form is included. Panel B: plot of % of G-quadruplex forms promoted by
increasing ligands concentrations.
and one side chain located in each DNA groove.23 We can infer
that the presence of the bulky Phe residue prevents threading of
the side chains between the base pairs. As a result, an external
binding mode promoted by the presence of highly charged
residues in the side chains can be favored. This is particularly
true for B-16, since it converts dsDNA into an insoluble compact
form also at low ligand/DNA ratio.
Electromobility Shift Assay. Electromobility shift assays were
performed to evaluate the ability of tested molecules to induce
G-quadruplex folding of the human telomeric sequence (Figure 5).
In these experiments, we used an oligonucleotide which con-
tains two repeats of the human telomeric sequence TTAGGG
(2GGG). This sequence can form intermolecular (dimeric or
tetrameric) G-quadruplex structures that migrate more slowly
than the nonstructured single-stranded DNA (Figure 5A). None
of our tested derivatives promoted the formation of tetrameric
G-quadruplexes. However, some of them were able to promote
induction of a dimeric G-quadruplex arrangement of the 2GGG
sequence depending upon side chain composition.
Negligible conversion of oligomer 2GGG to an intermolecular
G-quadruplex structure was observed when an amino acid is
interposed between the AA1 and the N-terminal Lys residue
(compare compounds -13 with homologues -14; Figure 5B).
Interestingly, thepresenceofaPhe(compoundswithside chain-16)
that was observed to comparably reduce the thermal stability of
the DNAꢀligand complex, allows retention of effective G-quad-
ruplex induction, the most successful being derivative D-16 with
side chains located at positions 1,5. These data further support a
peculiar behavior associated to the presence of a Phe residue in
the side chains.
Enzymatic Inhibition and Cell Toxicity Assays. To assess
possible interference of test drugs with enzymatic processes
involving DNA, we examined the effects of peptidyl anthraqui-
nones upon DNA amplification and elongation mediated by
Taq polymerase and telomerase, respectively. In addition, we
determined the cytotoxic activity of our compounds by MTT
assay on HeLa human cancer cells. The results are summarized in
Table 2.
2132
dx.doi.org/10.1021/bc200389w |Bioconjugate Chem. 2011, 22, 2126–2135

Bioconjugate Chemistry
ARTICLE
Table 2. Taq Polymerase Inhibition, Telomerase Inhibition
and Cell Cytotoxicity (MTT on HeLa cell line) Produced by
the Test Anthraquinone-Aminoacyl Conjugates
IC50 (μM)
compound
Taq polymerasea
telomerasea
HeLa
A-13
A-14
A-15
A-16
B-13
B-14
B-15
B-16
C-13
C-14
C-16
D-13
D-14
D-16
E-13
E-14
E-16
5
30
15
15
10
22
>40
20
2.5
5
0.8
1.9
0.9
7.0
0.8
6.5
3.8
7.1
1.5
3
27 ( 1.2
25 ( 1.5
57 ( 1.1
41 ( 1.8
14 ( 2.0
28 ( 2.7
49 ( 2.3
57 ( 2.7
31 ( 2.6
23 ( 5.0
18 ( 3.0
2.6 ( 0.3
27 ( 3.3
2.5 ( 0.2
1.2 ( 0.3
1.8 ( 0.1
6.0 ( 0.2
Figure 6. Correlation between Telomerase elongation inhibition
(IC50) and thermal stabilization (ΔTm/Tm) of telomeric G-quadruplex
measured for all tested derivatives.
20
2.0
5
10
n.o.b
2.5
4
related to G-quadruplex binding thermodynamics as shown by
the almost linear dependence of IC50 for telomerase inhibition
from ΔTm/Tm, where Tm is the melting temperature in the pre-
sence of saturating amounts of the drug (Figure 6).15 On the
contrary, the G-quadruplex affinity data are poorly related to cell
cytotoxicity.
15
5
n.o.
18
>40
>40
20
a Experimental error (4%. b n.o., not observed up to 40 μM.
In turn, Taq inhibition correlates with double-stranded DNA
stabilization and, to a lesser extent, to cell cytotoxicity. In this
case, the most active compounds are 1,5 and 1,8 derivatives that,
indeed, were confirmed to be the most efficient DNA inter-
calators. We can infer that a short-term drug exposure causes an
interference with processing of the nucleic acid that ultimately is
translated into a cytotoxic signal.
Table 3. Increase of the Melting Temperature (ΔTm °C)
Produced by Tested Derivatives (1 μM) on a G-Quadruplex
Folded Telomeric Sequence or Double-Stranded DNA in the
Presence of 50 mM KCla
compound
G-quadruplex telomeric sequence
double-stranded DNA
Accordingly, poor dsDNA binders like B-16 and C-16 are
poorly cytotoxic. Interestingly, however, both latter anthraqui-
nones stimulated β-galactosidase activity in HeLa cells treated for
240 h at subcytotoxic concentrations without appreciably affect-
ing cell growth (Figure 7), which is consistent with an induction
of delayed senescence. However, when the more cytotoxic 1,5
regioisomer (D-16) was used in the same conditions, senescence
induction was clearly coupled to a significant drop in cell growth.
These results further emphasize the peculiar behavior of the Phe
containing derivatives also at a cellular level.
A-13
A-14
A-15
A-16
B-13
B-14
B-15
B-16
C-13
C-14
C-16
D-13
D-14
D-16
E-13
E-14
E-16
15.6
11.9
11.0
7.1
2.6
1.8
4.0
0.9
2.5
2.3
3.6
0.3
1.4
1.0
0.1
4.6
4.5
1.2
4.5
2.2
0.1
13.7
7.4
8.2
7.4
9.5
5.2
3.9
’ DISCUSSION
20.1
14.1
15.7
17.6
12.1
4.3
The DNA binding properties of 9,10-anthracenedione deri-
vatives have been extensively studied in the past,25 highlighting
that the nature of the side chains as well as their relative position
are critical factors to address their DNA binding mode.12,14
Conceivably, the presence of positively charged groups in the
side chains greatly favors the binding process although generally
regardless of the sequence/structure they bind to. However, pro-
vided that different DNA arrangements locate phosphate groups
at different relative distances, in principle, charged interactions
could be used for selective recognition if the binders’ chains are
appropriately positioned with reference to the macromolecule.
Here, we showed for all test derivatives that increasing the dis-
tance of the lysyl residue from the anthracenedione ring system
invariantly decreases G-quadruplex affinity. This is probably due
to the fact that the positive charges are, on average, located more
a DNA concentration was 0.25 μM, experimental error (0.4°C.
For all active derivatives, inhibition of telomerase-mediated
DNA elongation occurred generally at lower concentrations in
comparison to DNA amplification. From these data, selective
telomerase inhibition by our ligands cannot be firmly established
from the present results, since competition between telomerase
elongation and Taq polymerase amplification can occur.24 In any
event, the observed modulation of telomerase elongation is well-
2133
dx.doi.org/10.1021/bc200389w |Bioconjugate Chem. 2011, 22, 2126–2135

Bioconjugate Chemistry
ARTICLE
Figure 7. β-Galactosidase activity of HeLa cell when untreated (A) or treated with C-16 (B) or D-16 (C) for 144 h.
distant from the anionic sites of DNA. Hence, compounds
having longer side chains with clear additional flexibility
(possibly rendered less effective by the rigid peptide bonds)
cannot get closer to the target counterions than the starting
congeners. Apparently, introduction of the bulky Phe residue is
even more detrimental than the presence of a glycine or alanine
linker. However, Phe-containing derivatives turn out to be
selective for G-quadruplex because, unlike other congeners, they
hardly bind to double-stranded DNA, yet they are still able to
stack on top of the terminal tetrads without requiring significant
DNA distortion.
’ ASSOCIATED CONTENT
S
Supporting Information.
Structural properties, syn-
b
thetic procedures, and analytical data (NMR, HRMS, and
combustion) for new compounds. This material is available free
of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
*E-mail: claudia.sissi@unipd.it, tel.: +39-049 8275711.
Shifting side chains from position 2,6 or 2,7 (families A, B,
and C) to 1,5 (family D) or, to a lesser extent, to 1,8 (family E)
helps increasing G-quadruplex stabilization, but at the same
time, it promotes effective binding to dsDNA. Interestingly, this
behavior is distinct from the one we reported for anthracene
derivatives of the bisanthrene family.26 In fact, among those
anthracene derivatives, the shift of the side chains from 2,6 to
1,5 positions slightly incremented G-quadruplex recognition
but remarkably disfavored double-stranded DNA binding. This
suggests that the presence of the carbonyl groups causes different
orientations of the anthracenedione vs anthracene ring when
bound to G-quadruplex DNA.
As far as the biochemical and biological results, we found
pronounced cell growth inhibition when the double-stranded
DNA recognition process is efficient. On the other hand, the
onset of senescence in cancer cells is particularly evident with the
more telomerase-sensitive derivatives, which produce selective
G-quadruplex stabilization. In conclusion, although we did not
succeed in effectively improving G-quadruplex binding, we were
able to remarkably uncouple quadruplex vs duplex affinity, which
gives useful hints for future design of compounds exhibiting
selectivity and efficient tetraplex-inducing properties at the
same time.
’ ACKNOWLEDGMENT
This work was supported by AIRC (Associazione Italiana per
la Ricerca sul Cancro) grant # 5286.
’ ABBREVIATIONS:
AA, amino acid; AQ, anthraquinones; β-Ala, β-alanine; TBTU,
O-(benzotriazol-1-yl)-N,N,N0,N0-tetramethyluronium tetrafluoro-
borate; HOBt, hydroxybenzotriazole; DIPEA, diisopropylethyla-
mine; TFA, trifluoroacetic acid; DMF, dimethylformamide; EGTA,
ethylene glycol-bis(β-aminoethyl)-N,N,N0,N0-tetraacetic acid;
DMSO, dimethylsulfoxide; FAM, fluorescein; Tm, melting tem-
perature; CD, circular dichroism; MTT, (3-(4,5-dimethylthiazol-
2-yl)-2,5-diphenyltetrazolium; PCR, polymerase chain reaction;
TRAP, telomerase repeat amplification protocol
’ REFERENCES
(1) O’Sullivan, R. J., and Karlseder, J. (2010) Telomeres: protecting
chromosomes against genome instability. Nat. Rev. Mol. Cell Biol. 11, 171–81.
(2) Watson, J. D. (1972) Origin of concatemeric T7 DNA. Nat. New
Biol. 239, 197–201.
2134
dx.doi.org/10.1021/bc200389w |Bioconjugate Chem. 2011, 22, 2126–2135

Bioconjugate Chemistry
ARTICLE
(3) Shay, J. W., Wright, W. E., and Werbin, H. (1991) Defining the
molecular mechanisms of human cell immortalization. Biochim. Biophys.
Acta 1072, 1–7.
(4) Shay, J. W., and Bacchetti, S. (1997) A survey of telomerase
activity in human cancer. Eur. J. Cancer 33, 787–91.
(5) Greider, C. W., and Blackburn, E. H. (1985) Identification of a
specific telomere terminal transferase activity in Tetrahymena extracts.
Cell 43, 405–13.
mode for anthracene-9,10-dione antitumor drugs. Biochemistry 31,
11632–40.
(24) De Cian, A., Cristofari, G., Reichenbach, P., De Lemos, E.,
Monchaud, D., Teulade-Fichou, M. P., Shin-Ya, K., Lacroix, L., Lingner,
J., and Mergny, J. L. (2007) Reevaluation of telomerase inhibition by
quadruplex ligands and their mechanisms of action. Proc. Natl. Acad. Sci.
U. S. A. 104, 17347–52.
(25) Lown, J. W. (1993) Anthracycline and anthraquinone antic-
ancer agents: current status and recent developments. Pharmacol. Ther.
60, 185–214.
(26) Folini, M., Pivetta, C., Zagotto, G., De Marco, C., Palumbo, M.,
Zaffaroni, N., and Sissi, C. (2010) Remarkable interference with telomeric
function by a G-quadruplex selective bisantrene regioisomer. Biochem.
Pharmacol. 79, 1781–90.
(6) Folini, M., Gandellini, P., and Zaffaroni, N. (2009) Targeting the
telosome: therapeutic implications. Biochim. Biophys. Acta 1792, 309–16.
(7) Neidle, S. (2010) Human telomeric G-quadruplex: the current
status of telomeric G-quadruplexes as therapeutic targets in human
cancer. Febs J. 277, 1118–25.
(8) Davis, J. (2004) G-quartets 40 years later: from 50-GMP to
molecular biology and supramolecular chemistry. Angew. Chem. 43,
668–698.
(9) De Cian, A., Lacroix, L., Douarre, C., Temime-Smaali, N.,
Trentesaux, C., Riou, J. F., and Mergny, J. L. (2008) Targeting telomeres
and telomerase. Biochimie 90, 131–155.
(10) Neidle, S. (2009) The structures of quadruplex nucleic acids
and their drug complexes. Curr. Opin. Struct. Biol. 19, 239–50.
(11) Monchaud, D., and Teulade-Fichou, M. P. (2008) A hitchhiker’s
guide to G-quadruplex ligands. Org. Biomol. Chem. 6, 627–36.
(12) Perry, P. J., Gowan, S. M., Reszka, A. P., Polucci, P., Jenkins,
T. C., Kelland, L. R., and Neidle, S. (1998) 1,4- and 2,6-disubstituted
amidoanthracene-9,10-dione derivatives as inhibitors of human telomer-
ase. J. Med. Chem. 41, 3253–60.
(13) Perry, P. J., Read, M. A., Davies, R. T., Gowan, S. M., Reszka,
A. P., Wood, A. A., Kelland, L. R., and Neidle, S. (1999) 2,7-Disubsti-
tuted amidofluorenone derivatives as inhibitors of human telomerase.
J. Med. Chem. 42, 2679–84.
(14) Perry, P. J., Reszka, A. P., Wood, A. A., Read, M. A., Gowan,
S. M., Dosanjh, H. S., Trent, J. O., Jenkins, T. C., Kelland, L. R., and
Neidle, S. (1998) Human telomerase inhibition by regioisomeric
disubstituted amidoanthracene-9,10-diones. J. Med. Chem. 41, 4873–84.
(15) Zagotto, G., Sissi, C., Lucatello, L., Pivetta, C., Cadamuro, S. A.,
Fox, K. R., Neidle, S., and Palumbo, M. (2008) Aminoacyl-anthraqui-
none conjugates as telomerase inhibitors: synthesis, biophysical and
biological evaluation. J. Med. Chem. 51, 5566–74.
(16) Zagotto, G., Sissi, C., Moro, S., Dal Ben, D., Parkinson, G. N.,
Fox, K. R., Neidle, S., and Palumbo, M. (2008) Amide bond direction
modulates G-quadruplex recognition and telomerase inhibition by 2,6
and 2,7 bis-substituted anthracenedione derivatives. Bioorg. Med. Chem.
16, 354–61.
(17) Dahan, A., Ashkenazi, T., Kuznetsov, V., Makievski, S., Drug, E.,
Fadeev, L., Bramson, M., Schokoroy, S., Rozenshine-Kemelmakher, E.,
and Gozin, M. (2007) Synthesis and evaluation of a pseudocyclic
tristhiourea-based anion host. J. Org. Chem. 72, 2289–96.
(18) De Cian, A., Guittat, L., Kaiser, M., Sacca, B., Amrane, S.,
Bourdoncle, A., Alberti, P., Teulade-Fichou, M. P., Lacroix, L., and
Mergny, J. L. (2007) Fluorescence-based melting assays for studying
quadruplex ligands. Methods 42, 183–95.
(19) Risitano, A., and Fox, K. R. (2003) Stability of intramolecular
DNA quadruplexes: comparison with DNA duplexes. Biochemistry 42,
6507–13.
(20) Crothers, D. M. (1971) Statistical thermodynamics of nucleic
acid melting transitions with coupled binding equilibria. Biopolymers
10, 2147–60.
(21) Darby, R. A., Sollogoub, M., McKeen, C., Brown, L., Risitano, A.,
Brown, N., Barton, C., Brown, T., and Fox, K. R. (2002) High throughput
measurement of duplex, triplex and quadruplex melting curves using
molecular beacons and a LightCycler. Nucleic Acids Res. 30, e39.
(22) Ambrus, A., Chen, D., Dai, J., Bialis, T., Jones, R. A., and Yang,
D. (2006) Human telomeric sequence forms a hybrid-type intramole-
cular G-quadruplex structure with mixed parallel/antiparallel strands in
potassium solution. Nucleic Acids Res. 34, 2723–35.
(23) Tanious, F. A., Jenkins, T. C., Neidle, S., and Wilson, W. D.
(1992) Substituent position dictates the intercalative DNA-binding
2135
dx.doi.org/10.1021/bc200389w |Bioconjugate Chem. 2011, 22, 2126–2135