Synthesis of a non-cationic, water-soluble perylenetetracarboxylic diimide and its interactions with G-quadruplex-forming DNA

Article abstract of DOI:10.1016/j.bmc.2006.09.075

A number of N,N′-disubstituted perylenetetracarboxylic diimides have been reported to bind effectively to DNA that adopts G-quadruplex motifs. In some cases, this binding may actively drive the transition from single-strand DNA to the quadruplex form. The perylenediimides in the reported cases all have amine-containing side chains, which are thought to interact with the grooves of the quadruplex and help dictate the selectivity of these compounds for quadruplex versus duplex DNA. We synthesized a polyethyleneglycol-swallowtailed (PEG-tailed) perylenediimide that is water-soluble even though it is uncharged. Binding to duplex and quadruplex DNA of this perylenediimide was studied by fluorescence quenching titrations under a variety of salt conditions, and the compound's effect on quadruplex formation was studied by non-denaturing gel electrophoresis. Our results indicate that while the molecule binds to single-stranded DNA quite effectively and with selectivity, it does not drive the transition of the DNA to the tetrameric quadruplex structure, supporting the idea that charge neutralization is a key component of perylene compounds that stabilize tetrameric quadruplexes.

Full text of DOI:10.1016/j.bmc.2006.09.075

Bioorganic & Medicinal Chemistry 15 (2007) 186–193
Synthesis of a non-cationic, water-soluble perylenetetracarboxylic
diimide and its interactions with G-quadruplex-forming DNA
Ramakrishna Samudrala, Xu Zhang, Randy M. Wadkins and Daniell Lewis Mattern*
Department of Chemistry and Biochemistry, Box 1848, University of Mississippi University, MS 38677, USA
Received 27 July 2006; revised 27 September 2006; accepted 29 September 2006
Available online 10 October 2006
Abstract—A number of N,N0-disubstituted perylenetetracarboxylic diimides have been reported to bind effectively to DNA that
adopts G-quadruplex motifs. In some cases, this binding may actively drive the transition from single-strand DNA to the quadru-
plex form. The perylenediimides in the reported cases all have amine-containing side chains, which are thought to interact with the
grooves of the quadruplex and help dictate the selectivity of these compounds for quadruplex versus duplex DNA. We synthesized a
polyethyleneglycol-swallowtailed (PEG-tailed) perylenediimide that is water-soluble even though it is uncharged. Binding to duplex
and quadruplex DNA of this perylenediimide was studied by fluorescence quenching titrations under a variety of salt conditions,
and the compound’s effect on quadruplex formation was studied by non-denaturing gel electrophoresis. Our results indicate that
while the molecule binds to single-stranded DNA quite effectively and with selectivity, it does not drive the transition of the
DNA to the tetrameric quadruplex structure, supporting the idea that charge neutralization is a key component of perylene
compounds that stabilize tetrameric quadruplexes.
ꢀ 2006 Elsevier Ltd. All rights reserved.
1. Introduction
single strands.11 A number of other perylenediimides
bind to quadruplex DNA, and their affinities are con-
trolled by the structures of the side chains attached to
the imide nitrogens. The side chains examined have all
contained cationic amino moieties under the conditions
studied, and these groups are thought to control speci-
ficity for quadruplex versus duplex DNA binding.12,13
We are unaware of any reports on non-cationic peryl-
enes binding to quadruplex DNA, although the com-
pound PIPER2 (2), containing a 2-pyridino side chain,
has only a small fractional charge at pH 7 and above.12
Disinterest in uncharged perylenediimides is likely due
to the extremely poor solubility of most substituted per-
ylenediimides in aqueous solutions.
It is now well established that guanine-rich sequences of
DNA can adopt secondary structures quite distinct from
typical Watson-Crick duplex DNA. These secondary
structures include ‘frayed wires’1 and quadruplex
DNA.2,3 Quadruplexes are of particular interest in can-
cer biology as they are thought to play a role in the sta-
bilization of telomeric DNA and its binding to the
enzyme telomerase, which is active in a number of tumor
types.4,5 Consequently, agents that target telomeric
quadruplexes are thought to be potential antitumor
agents.6,7 More recently, quadruplexes have also been
identified in promoters of specific genes, notably
c-MYC,8–10 and agents that bind quadruplex DNA
can affect transcription of genes regulated by these
promoters.
Recently, Wescott and Mattern14 described the synthesis
of a series of perylenediimides substituted with a 19-car-
bon ‘swallowtail’ (a long alkyl group connected at the
middle of its chain15) on one of the imide nitrogens,
and a variety of one-electron donor groups on the other
nitrogen. These materials were prepared to test their
ability to rectify electrical current when aligned in a
monolayer16 utilizing the electron-acceptor character
of perylenediimides. Swallowtails (as in 3) are crucial
for imparting workable solubility to perylenediimides
in nonpolar solvents. In polar solvents like alcohol or
water, however, the solubility of alkyl-swallowtailed
Although a number of different small molecules are
known to interact with quadruplex DNA, the most of-
ten studied are those that are derivatives of porphyrin
and perylene. The perylenediimide PIPER (1) has been
shown to promote formation of quadruplex DNA from
*
Corresponding author. Tel.: +1 662 915 5335; fax: +1 662 915
7300; e-mail: mattern@olemiss.edu
0968-0896/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2006.09.075

R. Samudrala et al. / Bioorg. Med. Chem. 15 (2007) 186–193
187
perylenediimides is negligible. We anticipated that devel-
NH2
O O
NH2
opment of a hydrophilic swallowtail would be advanta-
geous for eventual control of rectifier orientation.
Perylenediimides that contain long polyethylene glycol
(PEG) swallowtails17 or shorter PEG segments separat-
ed from the imide nitrogen by three-carbon spacers like
O
O
O
O
O
O
O
18 have good water solubility, and suggested that peryl-
O O
4
enediimide 5 with two PEG swallowtails would be solu-
ble in polar solvents. We report here the synthesis of 5
and its aqueous solubility. With 5 in hand, its interac-
tion with DNA in aqueous solutions could be readily
investigated, allowing observation of the effect of un-
charged perylenediimides on the formation of quadru-
plex DNA, which is the focus of this report. To our
knowledge, 5 is the first water-soluble perylenediimide
that does not contain an amino-substituted side chain
to be examined for its interaction with DNA.
8
9
O
Zn(OAc)2
imidazole, Δ
Zn(OAc)2
imidazole, Δ
7
Scheme 1.
genation using a large excess of Pd/C. (Similar attempts
to prepare 9 using single t-Boc23 and CBZ24 protection
of the amino group were unsuccessful.) The resulting
amine (9) was condensed with 7 in molten imidazole with
zinc acetate25 to give the crude target 5 (Scheme 2).
Purification of the doubly PEG-swallowtailed 5 was
performed by exploiting the basicity of imidazole with
an acidic ion exchange chromatographic separation.
After rinsing the column with HCl (aq), elution of
the reaction mixture (acidified with AcOH) with meth-
anol afforded 5 free from imidazole. This material
could be further purified by column chromatography
(basic alumina, 0.4–0.8% MeOH in EtOAc). Com-
pound 5 was found to be highly soluble in water,
methanol, chloroform, and ethyl acetate, giving deep
red solutions in each. Less solubility was observed in
diethyl ether (Et2O/H2O partition ratio by UV–visible
absorbance = 1:7) and no solubility was observed in
hexane. Because of the high solubility of 5 in water
Ph-CH2-Br
NH2
N
K2CO3
EtOH
88%
HO
O
OH
HO OH
10
11
+
Ts-Cl
pyridine
O
O
O
13
s
OH
12
NaH
85%
2. Results and discussion
2.1. Synthesis
H2, Pd/C
81%
9
N
Perylenediimides are constructed by reacting perylenetet-
racarboxylic dianhydride (7) with an appropriate 1ꢁ
amine19 (Scheme 1). The polyether 1ꢁ amine 10-amino-
2,5,8,12,15,18-hexaoxanonadecane (9) was therefore a
key intermediate for the synthesis of 5. The reaction of
serinol (10) with benzyl bromide and K2CO3 in ethanol20
gave dibenzylserinol (11),21 which was treated with NaH
and the PEG tosylate 13.22 The resulting double William-
son alkylation22 gave the protected PEG-amine 14.
Deprotection to 9 was accomplished by catalytic hydro-
7
O O
Zn(OAc)2
imidazole, Δ
28%
O
O
O O
5
14
Scheme 2.

188
R. Samudrala et al. / Bioorg. Med. Chem. 15 (2007) 186–193
Table 1. UV–visible absorption maxima and molar absorptivities of 5
in various solvents
and other solvents, and the small amounts accessible
by synthesis, we were unable to saturate enough solu-
tion to determine its maximal solubility.
k1
k2
k3
e (kmax)
Et2O
452
457
458
472
482
486
489
501
517
521
525
539
6.5 · 104
6.8 · 104
8.2 · 104
2.4 · 104
We also prepared 6, a perylenediimide with two short
unbranched PEG tails, to compare with 5. The Mitsun-
obu reaction of alcohol 15 with phthalimide, DIAD, and
PPh3 gave N-(3,6,9-trioxadecyl)phthalimide (16), which
afforded 17 upon hydrazinolysis, as described by Dombi
et al.26 Coupling of 17 with 7 gave CHCl3-soluble 6.17
Assuming a molar absorptivity comparable to that of
5 (vide infra), we estimate the solubility of 6 in water
as 13 mM. However, 6 was substantially less soluble in
buffers containing 100 mM KCl, and precipitation of
this compound in buffers made it undesirable for studies
of its DNA-binding abilities (Scheme 3).
MeOH
CHCl3
H2O
kmax values (nm) are in bold.
Et2O, CH3OH, and H2O), as shown in Figure 1 and
Table 1. The spectra were similar in the organic
solvents, resembling that of alkyl-swallowtailed peryl-
enediimides like 3 in CHCl3.27 However, solvatochro-
mism was observed for 5 in water, as shown by a
decreased extinction coefficient and a bathochromic
shift of absorption peaks. Furthermore, the relative
intensities of the absorption peaks changed, with the
0 ! 1 transition (at 501 nm) becoming kmax. Similar
behavior has been reported for tethered oligomers of
perylenediimides as well as for concentrated solutions
of perylenediimides,28,29 suggesting that 5 in water
forms aggregates at concentrations ꢀ1 mM where
solutions of perylenediimides in nonpolar solvents
are monomeric.28,29 As aggregation of perylenedii-
mides is important for quadruplex stabilization,11,13
the pH-independent aggregation may prove important
for the interaction of non-cationic pereylenediimides
with DNA. The UV–visible spectra of 6 in chloroform
and water (Fig. 1) are similar to those of 5, although
the absorptions in water are not as well resolved. Both
compounds also exhibited intense fluorescence emis-
sion (described below for 5).
The solubility of 5 and 6 allowed us to compare their
UV–visible spectra in a range of solvents (CHCl3,
O
O
O
+
O
15
OH
HN
O
DIAD, PPh
46%
O
3
O
O
O
N
O
16
NH NH
50%
2
2
7
6
O
O
O
NH
Zn(OAc)
2
2
2.2. Titration curves
imidazole, Δ
17
77%
We investigated the ability of 5 to bind to DNA. For
comparison, we used PIPER11 (1; a kind gift of Sean
Scheme 3.
Figure 1. (A) UV–visible absorption spectra of 5 in various solvents: H2O, 43 lM; methanol (MeOH), 42 lM; diethyl ether (Et2O), 49 lM; CHCl3,
38 lM. (B) UV–visible absorption spectra of 6 in CHCl3 and H2O. The sample in water is saturated (ca. 13 mM).

R. Samudrala et al. / Bioorg. Med. Chem. 15 (2007) 186–193
189
M. Kerwin, University of Texas). The two synthetic
DNA sequences used were the G-rich ssDNAs taken
from the human telomeric repeat (2HTR, 32-bases)
and from the nuclease hypersensitive region of the c-
MYC gene (NHE-27, 27-bases).
DNA over the concentration range used in our
experiments. Significant aggregation was noted to oc-
cur at concentrations >5 lM with both compounds
(data not shown) but at the concentrations used for
the titration curves with DNA (34–168 nM) no signif-
icant aggregation was observed.
Fluorescence emission spectra for 5 were obtained
using a Spex Fluoromax spectrophotometer. Excita-
tion light of 450 nm was used. Fluorescence emission
spectra in the absence and presence of increasing con-
centrations of NHE-27 are shown in Figure 2. We
performed a simple fluorescence quenching assay for
both PIPER and compound 5 in the absence of
Clearly, both PIPER and 5 bind to G-rich DNA
sequences. However, the dissociation constants for
the two ligands are significantly different. The quench-
ing of fluorescence at 545 nm was used to construct
the titration curves shown in Figure 3. As can be seen
in Figure 3B and D, PIPER binds the single-strand
NHE-27 with a Kd of 5.5 2.1 · 10ꢁ9 M with an
apparent 7 binding sites per strand. The affinity for
5
is almost 200-fold less, with
a
Kd of
A
3.0
9.8 0.3 · 10ꢁ7 M, and all curves for 5 could be fit
with a single binding site per strand. The difference
between the two illustrates the effect of the cationic
charge of PIPER in stabilizing the ligand–DNA
complex. It also may reflect an element of steric inter-
ference from the swallowtails of 5, which likely reduce
affinity. However, the affinity of 5 is well within the
range of small, uncharged molecules that bind DNA,
such as actinomycin D.
a
2.5
2.0
1.5
1.0
0.5
PIPER binds readily to the ssDNA 2HTR as well
(Fig. 3A). However, 5 does not bind this DNA to an
appreciable extent (Fig. 3C) and, in fact, binds the
G- and C-rich ssDNAs with low affinity equivalent to
that obtained with the dsDNA formed from both
strands (we estimate the Kd as >20 · 10ꢁ6 M). This is
in contrast to PIPER, which shows a clear higher affinity
for the G-rich ssDNA as well as the dsDNA formed
from the two strands (Fig. 3A).
b
0.0
500
550
600
650
wavelength (nm)
B
1.0
a
Curiously, 5 binds the dsDNA of NHE-27 weakly as
compared to the G-rich NHE-27 strand alone. These
means that, although the binding affinity for NHE-27
by 5 is moderate compared to cationic perylene analogs,
it possesses marked selectivity toward ssDNA. This may
prove useful in generating other water-soluble perylenes
with greater binding affinity.
0.8
0.6
0.4
0.2
0.0
In contrast, PIPER binds with slightly higher affinity
to dsDNAs of both NHE-27 and 2HTR than to the
ssDNAs alone. This is likely due to the ease of
intercalation into a double-stranded structure versus
a quadruplex structure since at pH 8 the PIPER is
less aggregated and favors dsDNA.30 Fitting the
binding curves for PIPER binding to dsDNA further
suggested a simple intercalation model. The value
for the number of binding sites per duplex had to
be very large to effectively fit the data, and this val-
ue was fixed at 27 or 32 for NHE-27 and 2HTR,
respectively. This implies approximately one base
pair per binding site on the DNA, which is what
would be expected with simple intercalation into
the DNA structures (assuming no site exclusion).
Using these values of n, the values of Kd for PIPER
binding to double strand NHE-27 and 2HTR were
b
550
wavelength (nm)
500
600
650
Figure 2. Binding of PIPER and 5 to single-strand DNA. The
fluorescence emission spectra of (A) PIPER and (B) 5 are shown in
Tris–EDTA buffer (pH 8.0) containing 100 mM KCl. The fluorescence
of the perylene compounds is quenched upon binding to DNA. In both
panels, curve
a indicates initial fluorescence of the compounds
(excitation at 450 nm) and curve b represents the final, fully bound
molecules. Intervening spectra indicate the effects of addition of the G-
rich ssDNA NHE-27. The quenching of fluorescence upon binding to
all forms of DNA was used to determine the binding affinities of the
two molecules, as shown in Figure 3.
6.6
respectively.
1.5 · 10ꢁ9 M
and
3.0 0.3 · 10ꢁ8 M,

190
R. Samudrala et al. / Bioorg. Med. Chem. 15 (2007) 186–193
Figure 3. Affinities of PIPER and 5 to DNA. From fluorescence quenching experiments (Fig. 2), titration curves were constructed for binding of
perylene analogs to different forms of DNA. In all panels, symbols are (•) G-rich ssDNA, (ꢂ) complementary C-rich ssDNA, and (h) the double
stranded duplex DNA made from the two ssDNAs. Panels are: (A) PIPER binding to 2HTR, (B) PIPER binding to NHE-27, (C) 5 binding to
2HTR, and (D) 5 binding to NHE-27. Lines indicate curve fits to the simple binding model given in Eq. 1. The resulting dissociation constants (Kd)
and binding sites per DNA sequence (n) are given in the text.
2.3. Non-denaturing gel electrophoresis
3. Conclusion
The perylene PIPER has been shown to drive the transi-
tion from single stranded DNA to quadruplex
DNA.11–13 We conducted similar experiments with 5
using a non-denaturing 16% polyacrylamide gel with
TBE (containing KCl) as the running buffer. Concentra-
tions of 5 were chosen according to the dissociation con-
stants to give >90% of DNA with bound ligand at the
highest concentration. As can be seen in Figure 4, while
PIPER easily induces dimer and tetramer formation
with increasing concentrations, compound 5 does not.
PIPER can induce tetramer formation in both NHE-
27 and 2HTR, and while 5 binds to the G-rich ssDNA
NHE-27, it does not convert it to a dimer or tetramer.
Hence, the binding seen with NHE-27 must be to the
ssDNA alone. This raises the interesting possibility that
uncharged perylenes bind to DNA by a mechanism that
does not involve quadruplex stabilization. It will be
interesting to observe how this affects telomerase activity
on DNAs containing the NHE-27 sequence.
In this paper, we present the synthesis and characteriza-
tion of an uncharged, water-soluble perylene compound
that binds with selectivity to G-rich DNA capable of
forming quadruplex structures. To our knowledge, all
previously characterized perylene analogs that bind to
DNA have contained side chains with substituted amino
moieties, and with the exception of PIPER2 at pH > 7,12
all have been cationic. In fact, the cationic charges are
related to the selectivity of the compounds for ssDNA
versus dsDNA.11,13 As perylene analogs become depro-
tonated, they aggregate and selectively interact with
G-rich ssDNA.
Our compound 5 is non-cationic and water soluble in its
own right and shows preferential binding to the ssDNA
NHE-27. While the binding affinity is lower for 5 than
for the related PIPER molecule, it is still well within
the range of biologically important small molecules that
bind to DNA. Further, the lack of charge on 5 may

R. Samudrala et al. / Bioorg. Med. Chem. 15 (2007) 186–193
191
scaffold structure for development of ssDNA-specific
molecules that may be useful for affecting transcription
of a number of genes.31 Exploration of the telomerase
inhibition activity of compound 5 is underway.
A
B
Piper
Piper
5
5
T
D
4. Experimental
4.1. 2-(N,N-Dibenzylamino)-1,3-propanediol (11)
Serinol (10, 2.00 g, 22.0 mmol) and K2CO3 (8.36 g,
60.5 mmol) were dissolved in 40 mL EtOH and stirred
for 30 min. Afterwards, benzyl bromide (8.34 g,
48.8 mmol) was added dropwise and the reaction mix-
ture was allowed to stir for two days at room temper-
ature. The mixture was then filtered to remove
undissolved K2CO3 and the filtrate was concentrated
by rotary evaporation to give a white solid. This was
treated with hot benzene, and the soluble portion was
purified by crystallization from 25 mL of benzene and
30 mL of hexane to give white needles (5.25 g, 88%),
mp 81–82 ꢁC, Rf (silica gel, 1:1 hexane/EtOAc) 0.23.
1H NMR (CDCl3) d 2.31 (s, 2H), 3.06 (m, 1H), 3.67
(m, 2H), 3.79 (m, 6H), 7.33 (m, 10H). 13C NMR
S
M
1 2 3 4 5 6 7 8 9 1011 1 2 34 5 6 7 8 910 11
Figure 4. Induction of quadruplex structures by perylene analogs. (A)
shows results for 2HTR, and (B) for NHE-27. In both panels, lane 1
indicates the G-rich ssDNA alone, while lanes 2-6 indicate increasing
concentrations of PIPER and lanes 7–11 indicate increasing concen-
trations of compound 5. In (A), the PIPER:2HTR ratios were 3, 6, 9,
12, and 15, and the 5:2HTR ratios were 4.2, 8.4, 12.6, 21, and 30. In
(B), the PIPER:NHE-27 ratios were 2.5, 5, 7.5, 10, and 12.5, and the
5:NHE-27 ratios were 37, 75, 112, 149, and 186. S, M, D, and T
indicate unstructured ssDNA, monomeric folded ssDNA, dimeric
quadruplex, and tetrameric quadruplex, respectively. The gels and
running buffer contained 100 mM KCl.
(CDCl3 = 77.16 ppm)
d 54.1, 60.07, 60.14, 127.4,
128.7, 129.1, 139.4. Anal. Calcd for C17H21NO2: C,
75.25; H, 7.80; N, 5.16. Found: C, 75.02; H, 7.94; N,
5.04.
4.2. N,N-Dibenzyl-2,5,8,12,15,18-hexaoxa-10-nonadec-
anamine (14)
Table 2. Summary of dissociation constants (Kd) and apparent
binding sites (n) for perylenetetracarboxylic diimides binding to DNA
NaH (1.3 g of a 60% dispersion in mineral oil,
32 mmol) was washed 3 · with 1 mL portions of hex-
ane. Then 50 mL of THF was added to the NaH and
the mixture was stirred to make a suspension. A solu-
tion of 11 (3.10 g, 11.6 mmol) in 15 mL THF was then
added dropwise under nitrogen and, after 30 min of
stirring, a solution of 2-methoxyethoxyethyl tosylate32
(13, 6.90 g, 25.2 mmol) in 20 mL THF was added drop-
wise. The reaction mixture was refluxed overnight,
cooled, and treated carefully with MeOH to destroy ex-
cess NaH. The solvent was removed under reduced
pressure and the resulting brown oil was extracted 4·
with 20 mL portions of hot ether. The combined ether
fractions were concentrated by rotary evaporation to
give 4.6 g (85%) of a yellowish liquid, Rf (EtOAc)
DNA
Kd (M)
n (sites
/DNA)
PIPER
2HTR
Complementary 2HTR 1.0 0.4 · 10ꢁ8
Double-strand 2HTR
3.4 1.8 · 10ꢁ9
6
1
3.0 0.3 · 10ꢁ8 32
5.5 2.1 · 10ꢁ9
NHE-27
Complementary NHE-27 nd
Double-strand NHE-27 6.6 1.5 · 10ꢁ9 27
7
nd
Compound 5 2HTR
>20 · 10ꢁ6
nd
nd
nd
Complementary 2HTR >20 · 10ꢁ6
Double-strand 2HTR
>20 · 10ꢁ6
9.8 0.3 · 10ꢁ7
NHE-27
Complementary NHE-27 >20 · 10ꢁ6
1
nd
1
Double-strand NHE-27 14.8 2.1 · 10ꢁ6
1
0.47. H NMR (CDCl3) d 3.1 (m, 1H), 3.38 (s, 3H),
3.56 (m, 20H), 3.78 (s, 4H), 7.19 (m, 2H), 7.27 (m,
4H), 7.38 (m 4H). 13C NMR (CDCl3) d 55.2, 56.2,
59.1, 70.4, 70.58, 70.63, 70.7, 72.0, 126.7, 128.1,
128.7, 140.8. Anal. Calcd for C27H41NO6 1/2 H2O: C,
66.92; H, 8.73; N, 2.89. Found: C, 67.21; H, 8.63; N,
2.94.
serve to make the molecule more easily accumulated
into cells. However, compound 5 clearly does not induce
formation of quadruplex DNA in the NHE-27 G-rich
strand. This is in agreement with data for the weakly
charged PIPER2, which also does not induce quadru-
plex formation. This in turn lends support to the
hypothesis that the interactions of the cationic side
chains with the phosphate groups of DNA are impor-
tant for stabilization of quadruplexes by PIPER and
related molecules.13 Compound 5 does show excellent
selectivity for G-rich ssDNA strands, despite its inability
to stabilize quadruplex structures. Hence, it may form a
4.3. 2,5,8,12,15,18-Hexaoxa-10-nonadecanamine (9)
N,N-Dibenzyl-2,5,8,12,15,18-hexaoxa-10-nonadecan-
amine (14, 4.00 g, 8.41 mmol) was dissolved in 110 mL
MeOH and 2.5 g of 10% Pd/C catalyst was added. The
mixture was hydrogenated overnight in a Parr shaker
at 55 psi H2. The reaction mixture was then filtered
and the solvent was removed under reduced pressure

192
R. Samudrala et al. / Bioorg. Med. Chem. 15 (2007) 186–193
to give 2.01 g of an oil (81%), Rf (72:18:10 CH2Cl2/
NH2OH(aq)/MeOH) 0.51. H NMR (CDCl3) d 2.05 (s,
2HTR: 50-AATCCGTCGAGCAGAGTTAGGGTTA
GGGTTAG-30
1
2H), 3.18 (m, 1H), 3.35 (m, 8H), 3.50 (m, 6H), 3.61
(m, 12H). 13C NMR (CDCl3) d 59.1, 70.48, 70.51,
70.56, 70.62, 71.9, 72.0. IR (thin film) 3522, 3364,
3290, 2891 cmꢁ1. Anal. Calcd for C13H29NO6: C,
52.86; H, 9.90; N, 4.74. Found: C, 52.73; H, 9.93; N,
4.51.
NHE-27: 50-TGGGGAGGGTGGGGAGGGTGGGG
AAGG-30
Fluorescence quenching (Fig. 2) was used to construct
the titration curves shown in Figure 3. Analogs were
mixed with DNA for 2 min at 25 ꢁC and fluorescence
spectra were recorded. For PIPER, 3 mL of buffer con-
taining 100 mM KCl, 1 mM EDTA, and 33.8 nM PIP-
ER were titrated with G-rich ssDNA 2HTR (2.04 lM
strands), the complementary C-rich ssDNA (2.04 lM
strands), or duplex from mixing of the two strands
(2.65 lM duplex). A similar experiment used NHE-27
(3.36 lM strands), its C-rich complementary strand
(2.7 lM strands), or duplex (1.42 lM duplex). Because
of potential protonation effects with PIPER, the buffer
solutions were either 10 mM Tris–HCl (pH 8.0) or
200 mM MES (pH 6.5). No difference was observed
between either pH.
4.4. N,N0-Di-[10-(2,5,8,12,15,18-hexaoxanonadecyl)]per-
ylene-3,4,9,10-bis(dicarboximide) (5)
2,5,8,12,15,18-Hexaoxa-l0-nonadecanamine (9, 200 mg,
0.68 mmol), 3,4,9,10-perylenetetracarboxylic dianhy-
dride (7, 133 mg, 0.34 mmol), imidazole (2.56 g), and
Zn(OAc)2 (0.03 g , 0.15 mmol) were heated at 165 ꢁC
overnight. The reaction mixture solidified upon cooling
and was dissolved in MeOH. A few drops of acetic acid
were added and the mixture was stirred for 15 min and
then eluted through an Amberlite IR-120 (+) ion ex-
change column (pre-washed with water, then 2 N HCl)
to remove imidazole. Lastly, imidazole-free material
was obtained by column chromatography on basic alu-
mina (0.4–0.8% MeOH in EtOAc) to give 0.09 g (28%)
of deep-red solid, mp 143–144 ꢁC, Rf (2:1 MeOH/
EtOAc) 0.70. Further elution gave 0.07 g (22%) of less
For compound 5, a 2.2 mL solution of 168 nM 5 in
Tris–HCl buffer (10 mM) containing 100 mM KCl
and 1 mM EDTA was titrated with G-rich ssDNA
2HTR (496 lM strands), the complementary C-rich
ssDNA (530 lM strands), or duplex from mixing of
the two strands (496 lM duplex). A similar experi-
ment used NHE-27 (437 lM strands), its C-rich
complementary strand (580 lM strands), or duplex
(218 lM duplex).
1
pure material. H NMR (CDCl3) d 3.28 (s, 12H), 3.42
(m, 8H), 3.56 (m, 8H), 3.61 (m, 12H), 3.72 (m, 4H),
3.99 (m, 4H), 4.19 (m, 4H), 5.72 (m, 2H), 8.58 (m,
4H), 8.63 (m, 4H). 13C NMR (CDCl3) d 52.2, 59.1,
69.4, 70.5, 70.62, 70.63, 72.0, 123.2, 123.6 (b), 126.4,
129.7, 131.7 (b), 134.6, 164.0. IR (KBr) 1700,
1654 cmꢁ1. Anal. Calcd for C50H62N2O16 1/2H2O: C,
62.82; H, 6.64; N, 2.93. Found: C, 62.91; H, 6.46; N,
2.93.
Data for both compounds were recorded as ꢁDF/Fo,
where ꢁDF is the difference in fluorescence at each
DNA concentration from Fo, the initial fluorescence of
the drug. Fluorescence changes were fitted to the simple
non-interacting site model of mass action. This model
assumes a ligand (L) binds to a DNA site (D) according
to:
4.5. N,N0-Di-[10-(3,6,9-trioxadecyl)]perylene-3,4,9,10-
bis(dicarboximide) (6)
3,6,9-Trioxadecylamine (90 mg, 0.55 mmol), 7 (120 mg,
0.31 mmol), imidazole (1 g), and Zn(OAc)2 (catalytic
amount) were combined in a reaction flask and heated
to 160 ꢁC for 18 h. The reaction mixture was then cooled
to rt, stirred for 10 min with 50 mL of water, and
extracted with CHCl3 (3· 50 mL). The combined ex-
tracts were washed with 5% HCl (aq) to remove residual
imidazole, dried over MgSO4, and concentrated by rota-
ry evaporation to give 200 mg of crude product. This
was recrystallized from hexane/CHCl3 to give 145 mg
L þ D ꢀ LD
ð1Þ
with a dissociation constant Kd = ([L][D])/[LD]. The val-
ues [L], [D], and [LD] are molar concentrations of the
free ligand, free DNA sites, and ligand-bound DNA
sites, respectively. The number of sites per DNA strand
is given by:
½Dꢃ ¼ n½Dꢃ ;
ð2Þ
o
where [D]o is the concentration of DNA in strands or
duplex and n is the number of sites per DNA. Fitting
of data to the above model was done using the non-lin-
ear least squares function of Kaleidagraph (Synergy
Software, Reading, PA) with Kd and ꢁDFmax/Fo as
adjustable parameters. Titrations were performed in
triplicate (all data are indicated in Fig. 3), and the
computed values of the parameters are reported in Ta-
ble 2 for mean standard deviations of these triplicate
determinations. For compound 5, the number of sites
per strand was fitted with n = 1. For PIPER, multiple
binding sites were evident (n ꢄ 1), and n was fixed at
32 and 27 for double strand 2HTR and NHE-27,
respectively.
1
(77%) of 6, Rf (9:1 CHCl3/MeOH) 0.58; H NMR in
agreement with the literature.17 13C NMR (CDCl3) d
39.5, 59.2, 68.0, 70.3, 70.7, 70.8, 72.0, 123.2, 123.3,
126.4, 129.4, 131.5, 134.6, 163.5. IR (KBr) 1691,
1654 cmꢁ1
.
4.6. Titration curves
The DNAs used were synthesized by solid phase meth-
ods and obtained from Midland Certified Reagent Com-
pany (Midland, TX). The sequences, previously shown
to be converted to quadruplex structures by small
molecules,12,33 were:

R. Samudrala et al. / Bioorg. Med. Chem. 15 (2007) 186–193
193
4. Kim, N. W.; Piatyszek, M. A.; Prowse, K. R.; Harley,
C. B.; West, M. D.; Ho, P. L. C.; Coviello, G. M.;
Wright, W. E.; Weinrich, S. L.; Shay, J. W. Science
1994, 266, 2011.
5. Sharma, H. W.; Maltese, J.-Y.; Zhu, X.; Kaiser, H. E.;
Narayanan, R. Anticancer Res. 1996, 16, 511.
6. Hurley, L. H.; Wheelhouse, R. T.; Sun, D.; Kerwin, S. M.;
Salazar, M.; Fedoroff, O. Y.; Han, F. X.; Han, J.; Izbicka,
E.; Von Hoff, D. D. Pharmacol. Therapeu. 2000, 85, 141.
7. Hurley, L. H. Biochem. Soc. Trans. 2001, 29, 692.
8. Seenisamy, J.; Rezler, E. M.; Powell, T. J.; Tye, D.;
Gokhale, V.; Joshi, C. S.; Siddiqui-Jain, A.; Hurley, L. H.
J. Am. Chem. Soc. 2004, 126, 8702.
4.7. Gel electrophoresis
The effects of PIPER and 5 in converting G-rich ssDNA
to quadruplex structures were analyzed by non-denatur-
ing gel electrophoresis (Fig. 4). A 16% polyacrylamide
gel was used with TBE (pH 8.0) containing 100 mM
KCl as the running buffer. Samples were prepared by
first end-labeling 2HTR and NHE-27 with 32P using
T4 polynucleotide kinase, with purification of labeled
DNA by spin column separation. Samples of 2HTR
(1 pmol) or NHE-27 (0.9 pmol) were mixed with increas-
ing concentrations of PIPER or 5 such that, according
to the titration curves obtained, greater than 90% of
the DNA had perylene analog bound to it at the highest
concentration used. Ratios of PIPER:2HTR used were
0, 3, 6, 9, 12, and 15 in a 10 lL final volume. Ratios
of PIPER:NHE-27 used were 0, 2.5, 5, 7.5, 10, and
12.5 in a 10 lL final volume. Ratios of 5: 2HTR used
were 4.2, 8.4, 12.6, 21, and 30 in a 10 lL final volume.
Ratios of 5: NHE-27 used were 37, 75, 112, 149, and
186 in a 10 lL final volume. Samples were incubated
at 23 2 ꢁC for 4 h, then loaded onto the gel and run
at 60 V at 25 ꢁC. The resulting gel mobilities were deter-
mined by radiography.
9. Siddiqui-Jain, A.; Grand, C. L.; Bearss, D. J.; Hurley, L.
H. Proc. Natl. Acad Sci. U.S.A. 2002, 99, 11593.
10. Grand, C. L.; Han, H.; Munoz, R. M.; Weitman, S.; Von
Hoff, D. D.; Hurley, L. H.; Bearss, D. J. Mol. Cancer
Ther. 2002, 1, 565.
11. Kern, J. T.; Kerwin, S. M. Bioorg. Med. Chem. Lett. 2002,
12, 3395.
12. Rossetti, L.; Franceschin, M.; Bianco, A.; Ortaggi, G.;
Savino, M. Bioorg. Med. Chem. Lett. 2002, 12, 2527.
13. Rossetti, L.; Franceschin, M.; Schirripa, S.; Bianco, A.;
Ortaggi, G.; Savino, M. Bioorg. Med. Chem. Lett. 2005,
15, 413.
14. Wescott, L. D.; Mattern, D. L. J. Org. Chem. 2003, 68,
10058.
15. Demmig, S.; Langhals, H. Chem. Ber. 1988, 121, 225.
16. Metzger, R. M. Chem. Rev. 2003, 103, 3803.
17. Williams, M. E.; Murray, R. W. Chem. Mater. 1998, 10,
3603.
Acknowledgments
We thank the National Science Foundation, Grants #
DMR-0099674 and DBI-0421319, for financial sup-
port. The authors also thank Dr. Sean M. Kerwin,
University of Texas, for providing us PIPER for these
studies.
18. Cormier, R.; Gregg, B. Chem. Mater. 1998, 10, 1309.
19. Langhals, H. Heterocycles 1995, 40, 477.
20. Beaulieu, P. L.; Wernic, D. J. Org. Chem. 1996, 61, 3635.
21. Chandrasekhar, S.; Mohanty, P. K.; Harikishan, K.;
Sasmal, P. K. Org. Lett. 1999, 1, 877.
22. The reverse Williamson is not feasible because the
ditosylate of 11 is not stable. Weber, K.; Kuklinski, S.;
Gmeiner, P. Org. Lett. 2000, 2, 647.
Supplementary data
23. Benoist, E.; Loussouarn, A.; Remaud, P.; Chatal, J.-F.;
Gestin, J.-F. Synthesis 1998, 1113.
24. Harada, H.; Morie, T.; Hirokawa, Y.; Kato, S. Chem.
Pharm. Bull. 1996, 44, 2205.
Procedures for preparation of compounds 13, 16, and
17. 1H NMR and 13C NMR spectra for 5. Supplementa-
ry data associated with this article can be found, in the
online version, at doi:10.1016/j.bmc.2006.09.075.
25. Langhals, H. Chem. Ber. 1985, 118, 4641.
26. Dombi, K. L.; Griesang, N.; Richert, C. Synthesis 2002, 816.
27. Langhals, H.; Jona, W. Chem. Eur. J. 1998, 4, 2110.
28. Wang, W.; Li, L.-S.; Helms, G.; Zhou, H.-H.; Li, A. D. Q.
J. Am. Chem. Soc. 2003, 125, 1120.
References and notes
29. Li, A. D. Q.; Wang, W.; Wang, L.-Q. Chem. Eur. J. 2003,
9, 4594.
30. Kerwin, S. M.; Chen, G.; Kern, J. T.; Thomas, P. W.
Bioorg. Med. Chem. Lett. 2002, 12, 447.
1. Protozanova, E.; Macgregor, R. B. Biochemistry 1996, 35,
16638.
31. Wadkins, R. M. Curr. Med. Chem. 2000, 7, 1.
32. Weber, E. Liebigs Ann. Chem. 1983, 770.
33. Rangan, A.; Federoff, O. Y.; Hurley, L. H. J. Biol. Chem.
2001, 276, 4640.
2. Williamson, J. R. Annu. Rev. Biophys. Biomol. Struct.
1994, 23, 703.
3. Guschlbauer, W.; Chantot, J.-F.; Thiele, D. J. Biomol.
Struct. Dyn. 1990, 8, 491.