Bisintercalating Threading Diacridines: Relationships between DNA Binding, Cytotoxicity, and Cell Cycle Arrest

Article abstract of DOI:10.1021/jm030253d

We have synthesized a series of bis(9-aminoacridine-4-carboxamides) linked via the 9-position with neutral flexible alkyl chains, charged flexible polyamine chains, and a semirigid charged piperazine-containing chain. The carboxamide side chains comprise N,N-dimethylaminoethyl and ethylmorpholino groups. The compounds are designed to bisintercalate into DNA by a threading mode, in which the side chains are intended to form hydrogen-bonding contacts with the O6/N7 atoms of guanine in the major groove, and the linkers are intended to lie in the minor groove. By this means, we anticipate that they will dissociate slowly from DNA, and be cytotoxic as a consequence of template inhibition of transcription. The dimers remove and reverse the supercoiling of closed circular DNA with helix unwinding angles ranging from 26° to 46°, confirming bifunctional intercalation in all cases, and the DNA complexes of representative members dissociate many orders of magnitude more slowly than simple aminoacridines. Cytotoxicity for human leukemic CCRF-CEM cells was determined, the most active agents having IC₅₀ values of 35-50 nM in a range extending over 20-fold, with neither the dimethylaminoethyl nor the ethylmorpholino series being intrinsically more toxic. In common with established transcription inhibitors, the morpholino series, with one exception, have no effect on cell cycle distribution in randomly dividing CCRF-CEM populations. By contrast, the dimethylaminoethyl series, with two exceptions, cause G2/M arrest in the manner of topoisomerase poisons, consistent with possible involvement of topoisomerases in their mode of action. Thus, the cellular response to these bisintercalating threading agents is complex and appears to be determined by both their side chain and linker structures. There are no simple relationships between structure, cytotoxicity, and cell cycle arrest, and the origins of this complexity are unclear given that the compounds bind to DNA by a common mechanism.

Full text of DOI:10.1021/jm030253d

5790
J . Med. Chem. 2003, 46, 5790-5802
Bisin ter ca la tin g Th r ea d in g Dia cr id in es: Rela tion sh ip s betw een DNA Bin d in g,
Cytotoxicity, a n d Cell Cycle Ar r est
Laurence P. G. Wakelin,*,† Xianyong Bu,† Alexandra Eleftheriou,† Alpesh Parmar,† Charbel Hayek,† and
Bernard W. Stewart‡,§
School of Medical Sciences, and the School of Women’s and Children’s Health, University of New South Wales, Sydney 2052,
New South Wales, Australia, and Public Health Unit, South East Sydney Area Health Service, Sydney 2031,
New South Wales, Australia
Received May 27, 2003
We have synthesized a series of bis(9-aminoacridine-4-carboxamides) linked via the 9-position
with neutral flexible alkyl chains, charged flexible polyamine chains, and a semirigid charged
piperazine-containing chain. The carboxamide side chains comprise N,N-dimethylaminoethyl
and ethylmorpholino groups. The compounds are designed to bisintercalate into DNA by a
threading mode, in which the side chains are intended to form hydrogen-bonding contacts with
the O6/N7 atoms of guanine in the major groove, and the linkers are intended to lie in the
minor groove. By this means, we anticipate that they will dissociate slowly from DNA, and be
cytotoxic as a consequence of template inhibition of transcription. The dimers remove and
reverse the supercoiling of closed circular DNA with helix unwinding angles ranging from 26°
to 46°, confirming bifunctional intercalation in all cases, and the DNA complexes of representa-
tive members dissociate many orders of magnitude more slowly than simple aminoacridines.
Cytotoxicity for human leukemic CCRF-CEM cells was determined, the most active agents
having IC50 values of 35-50 nM in a range extending over 20-fold, with neither the
dimethylaminoethyl nor the ethylmorpholino series being intrinsically more toxic. In common
with established transcription inhibitors, the morpholino series, with one exception, have no
effect on cell cycle distribution in randomly dividing CCRF-CEM populations. By contrast,
the dimethylaminoethyl series, with two exceptions, cause G2/M arrest in the manner of
topoisomerase poisons, consistent with possible involvement of topoisomerases in their mode
of action. Thus, the cellular response to these bisintercalating threading agents is complex
and appears to be determined by both their side chain and linker structures. There are no
simple relationships between structure, cytotoxicity, and cell cycle arrest, and the origins of
this complexity are unclear given that the compounds bind to DNA by a common mechanism.
In tr od u ction
stituents that binds by a threading mode.10 In the
nogalamycin-DNA complex, one sugar ring lies in the
DNA minor groove and the other makes hydrogen-
bonding interactions with the O6/N7 atoms of guanine
in the major groove. Actinomycin D is used in the
treatment of Wilm’s tumor and childhood sarcomas, and
echinomycin has been widely assessed in phase II
clinical trials.11,12
DNA-intercalating agents inhibit cell growth by two
well-established mechanisms: poisoning topoisomeras-
es1-3 and inhibiting transcription.4 Clinically useful
topoisomerase poisons include the anthracyclines doxo-
rubicin and daunomycin, the anthracenedione mitox-
antrone, and the 9-anilinoacridine amsacrine.1-3 Trap-
ping of transient topoisomerase-DNA cleavable complexes
by these drugs leads to clastogenesis, cell cycle arrest,
and apoptosis.1-3 Related experimental cytotoxins based
on acridine-4-carboxamide chromophores, such as N-[2-
(dimethylamino)ethyl]acridine-4-carboxamide, known as
DACA, as well as its 9-amino derivatives, also trap
topoisomerase cleavable complexes,5,6 and DACA is in
phase II clinical trials.7 Intercalating transcription
inhibitors are exemplified by the antitumor antibiotics
actinomycin D, echinomycin, and nogalamycin.8-10 Ac-
tinomycin D is a monofunctional intercalator,8 echino-
mycin is a bisintercalating agent,9 and nogalamycin is
a monofunctional intercalator bearing two sugar sub-
A characteristic shared by the transcription inhibitors
discussed is that they bind to DNA reversibly with
modest affinities in the range 105-107 M-1, placing
bulky groups in the minor groove covering four to six
basepairs, and form long-lived complexes that impede
the progression of RNA polymerases.8-10,13-16 This abil-
ity to form slowly dissociating DNA complexes, with
half-lives of hundreds of seconds, distinguishes the
transcription inhibitors from the topoisomerase poisons,
which dissociate from DNA rapidly in the millisecond
to second time range.17-19 While intercalating topo-
isomerase poisons have been widely investigated as
potential antitumor agents,1-3 scant attention has been
paid to the similar development of template transcrip-
tion inhibitors, despite their impressive experimental
antitumor activity (see citations in refs 4, 8-10, and 13-
16). This is, in part, because of the difficulty of endowing
chemically accessible compounds with the capacity to
* To whom correspondence should be addressed. Phone: +61-2-9385
2563. Fax: +61-2-9385 1059. E-mail: l.Wakelin@unsw.edu.au.
† School of Medical Sciences, University of New South Wales.
‡ School of Women’s and Children’s Health, University of New South
Wales.
§ Public Health Unit, South East Sydney Area Health Service.
10.1021/jm030253d CCC: $25.00 © 2003 American Chemical Society
Published on Web 11/20/2003

Bisintercalating DNA-Threading Agents
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 26 5791
1
F igu r e 1. Structure of the threading diacridines studied. -N(CH2CH2)2N- denotes piperazine. Data taken from ref 27.
form long-lived reversible complexes that occupy ap-
proximately half a turn of the helix in the minor groove,
down which RNA polymerases track. Here, we describe
our initial attempts to develop agents intended to meet
these requirements and to be cytotoxic as a consequence
of inhibiting transcription.
solid tumor activity of intercalating agents is limited
by their capacity to penetrate into tumor masses, and
this is in part determined by their DNA affinity.23 This
feature distinguishes our work from that of Chaires and
colleagues, who have sought to enhance the activity of
the anthracycline topoisomerase II poisons by develop-
ing nonthreading dimers with exceptionally high DNA-
binding constants.24 In collaboration with Iverson and
others, Chaires has also described a series of highly
cationic, peptide-linked, polyintercalating agents based
on naphthalenetetracarboxylic diimides, which appear
to thread their interchromophore linkers through the
DNA helix.25 These agents are active against bacteria
but are not toxic to mammalian cell lines,26 presumably
because of transport and/or metabolic stability prob-
lems. Given that the bisthreading compounds described
here are designed to dissociate slowly from DNA, with
the purpose of inhibiting transcription, and yet are
composed of elements known to poison topoisomerases,
we further anticipate that their biological properties
might reflect this dual character. Such dual activity,
should it occur, may prove beneficial in circumventing
the development of drug resistance selected by “single
character” agents.
We have sought to determine the extent to which the
interaction of the bis(9-aminoacridine-4-carboxamides)
with DNA is anticipated by molecular modeling, and
whether relative cytotoxic activity, as well as the mode
of cell killing, can be related to variations in linker and
side chain structure. To establish their capacity to
bisintercalate, we measured their helix unwinding
angles as determined by the ability to remove and
reverse the supercoiling of covalently closed circular
DNA.27 Poisoning of topoisomerase II activity is almost
invariably accompanied by cell cycle arrest at the G2/M
boundary28,29 and is often followed by cell death: a
desirable scenario so far as the response of malignant
cells to putative cytotoxic drugs is concerned.30,31 In the
present work, we assessed cell cycle distribution by flow
cytometry after treatment of the human lymphoblastoid
T-cell line CCRF-CEM. Such analysis also provides an
indication of cell death, by dint of the accumulation of
Our approach exploits a structural motif that com-
bines the bisintercalation of echinomycin with the
threading mode of nogalamycin. The novel agents are
dimers of 9-aminoacridine-4-carboxamide joined via the
9-amino group by various linkers, and bear one of two
different 4-carboxamide threading side chain “hooks”
(Figure 1). Recently, by crystallographic analysis, we
have established that nonthreading versions of the
9-aminoacridine-4-carboxamides with the same side
chains as used here, N,N-dimethylaminoethyl and eth-
ylmorpholino groups, bind to DNA with the side chains
lying in the major groove, forming hydrogen-bonding
interactions with the O6 and N7 atoms of guanine in a
manner similar to the amino sugar of nogalamycin.21,22
Thus, we anticipate that the bisthreading agents will
bind with their carboxamide side chains making analo-
gous bonding contacts with guanines in the major
groove, and that their linkers, comprising neutral
flexible alkyl chains, charged flexible polyamine chains,
and a semirigid charged piperazine-containing chain,
will come to lie in the minor groove. By binding in this
manner, the side chains are intended to behave as
“barbed spears” that do not excessively hinder interca-
lation as they thread through the helix but that make
withdrawal of the intercalated chromophores difficult
as a consequence of their guanine interactions. In
previous work with monofunctional threading agents of
this kind, based on 9-anilinoacridine-4-carboxamides,
we have demonstrated a substantial increase in complex
lifetime, without enhancing affinity, as a result of this
binding mode.20 This uncoupling of thermodynamics and
kinetics is a consequence of the slower rate of intercala-
tion associated with the requirement to thread the side
chain through the DNA duplex20. The ability to dissoci-
ate slowly, yet bind with modest affinity, is an important
element in the design of the threading dimers, since

5792 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 26
Sch em e 1. Synthesis of Acridone-4-carboxylic acid
Wakelin et al.
Sch em e 3. Synthesis of 9-Aminoacridine-4-carboxamide
dimers
Sch em e 2. Synthesis of
9-Aminoacridine-4-carboxamides
Sch em e 4. Synthesis of o-Aminoisophthalic acid
Syn t h esis of 9-Oxoa cr id a n -4-ca r b oxylic Acid .
The acridone-4-carboxylic acid was prepared via J our-
dan-Ullman reaction of o-iodoisophthalic acid with
aniline to give an N-phenylanthranilic acid, which was
then cyclized using polyphosphorous acid, yielding
9-oxoacridan-4-carboxylic acid as described by Rewcastle
et al.36 and shown in Scheme 1. This approach was
adopted because it can be readily modified to yield
5-substitued acridone-4-carboxylic acids by the use of
o-substituted anilines: an objective of future work, since
substitutions in the 5-position of 9-amino-DACA are
known to slow DNA dissociation rates as a result of
enhanced chromophore-base pair stacking interac-
tions.37 o-Iodoisophthalic acid was synthesized from
o-aminoisophthalic acid by diazotization and substitu-
tion with iodide according to standard literature meth-
ods. o-Aminoisophthalic acid is frequently prepared by
oxidation of 2-nitro-1,3-xylene, followed by reduction of
the nitro group with tin in hydrochloric acid.38 However,
in our hands this method gave poor yields and the
workup procedure was difficult. Accordingly, we sought
to develop a more efficient route to o-aminoisophthalic
acid in the current work. Since oxidation of isatin with
alkaline hydrogen peroxide is known to give o-ami-
nobenzoic acid in good yield,39 we reasoned that o-
aminoisophthalic acid could be made by similar treat-
ment of isatin-7-carboxylic acid, which itself can be
efficiently prepared by the method of Deady et al.40 And,
indeed, we found that oxidation of isatin-7-carboxylic
acid with alkaline hydrogen peroxide (Scheme 4) does
give o-aminoisophthalic acid in high yield. The advan-
tages of this method include the lack of a need to purify
intermediates and the final product, convenience of
handling and workup procedures, and high yields.
cells having less than G1 DNA content.32 For the
purposes of providing comparative data with type-
specific examples of topoisomerase poisons and inter-
calating transcription inhibitors, we also evaluated the
cell cycle effects of camptothecin (topoisomerase I),33
amsacrine, amsacrine-4-carboxamide and mitoxantrone
(topoisomerase II),34,35 actinomycin D, echinomycin, and
nogalamycin. We find that the novel diacridines are
indeed bisintercalating threading agents, and prelimi-
nary data indicate that the DNA complexes of selected
members dissociate many orders of magnitude more
slowly than those of their monofunctional counterparts.
Their cytotoxicity varies markedly with structure. As
expected, the known topoisomerase poisons invariably
caused G2/M arrest in the cell population tested,
whereas the established transcription inhibitors were
without effect on cell cycle progression. With the excep-
tion of the octamethylene- and piperazine-linked dimers,
the dimethylaminoethyl series of diacridines arrest cells
in G2/M, suggesting the involvement of topoisomerase
poisoning. However, the cycle signature of the former
compounds, which are among the most potent of the
dimethylaminoethyl dimers, is more akin to that of the
transcription inhibitors. Similarly, with the exception
of the hexamethylene-linked member, the morpholino
dimers fail to perturb the cell cycle distribution, strongly
suggesting the lack of involvement of topoisomerase
poisoning in their mode of action. Thus, there is a
complex cellular response to challenge across the series
of threading dimers, the determinants of which are
unclear given that the compounds studied appear to
bind to DNA by a common mechanism.
Syn th esis of 9-Ch lor oa cr id in e-4-ca r boxa m id es.
9-Chloro-[N-(2-dimethylamino)ethyl]acridine-4-carboxa-
mide and 9-chloro-N-[2-(4-morpholinyl)ethyl]-4-acridi-
necarboxamide were prepared from 9-oxo-acridan-4-
carboxylic acid via 9-chloroacridine-4-carbonyl chloride
as described by Atwell et al.41 Thus, amidation of the
acid chloride with an excess of N,N-dimethylethylene-
diamine or N-morpholinoethyleneamine in anhydrous
dichloromethane gave the respective target 9-chloro-
acridine-4-carboxamides (Scheme 2). The 9-aminoacri-
dine-4-carboxamide monomers were prepared from their
Resu lts
Ch em istr y. Our approach to the synthesis of bis(9-
aminoacridine-4-carboxamides) is outlined in Schemes
1-3. In short, we first prepared acridone-4-carboxylic
acid, converted it to the 9-chloroacridine-4-acyl chloride,
amidated with the carboxamide side chain of choice, and
finally coupled two monomers by nucleophilic reaction
with bifunctional R,ω-diamines.

Bisintercalating DNA-Threading Agents
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 26 5793
Sch em e 5. Synthesis of 1,4-Bis(aminoethyl)piperazine
has 38 negative supercoils at the ionic strength used
here,47,48 and given that removal of each supercoil
requires the Watson-Crick duplex to be unwound by
360°, the total amount of helix unwinding per DNA
molecule at equivalence is 13 680°. Knowing that pBR
322 has 4362 basepairs, it is a simple matter to calculate
the helix unwinding angle per bound ligand molecule
from the measured equivalence binding ratios for each
of the dimers studied, and these values are given in
Table 1. With regard to the dimethylaminoethyl series,
helix unwinding angles range from 26° to 46°, which is
1.5-2.7-fold greater than that of the parental monomer,
9-amino-DACA, and confirm that the compounds bind
with both of their chromophores simultaneously inter-
calated (see ref 27). In results to be published elsewhere,
the findings for the C6 and C8 DMAE dimers reported
here have been corroborated using viscometric titrations
with a different plasmid DNA (see ref 27).
corresponding 9-chloro compounds by reacting with
ammonia gas in hot phenol solution (Scheme 2).
Lin k er Design a n d Dia cr id in e Syn th esis. With
the exception of the C6 compounds (see Figure 1), the
linkers were designed to enable the diacridine chro-
mophores to extend over a distance of 10-11 Å and thus
span two basepairs in the bisintercalated complex: a
geometry known to provide stable bisintercalation.27
This structural requirement necessitates a linker length
of at least seven atoms (Figure 1). Bisintercalation by
hexamethylene-linked diacridines is believed to involve
a single basepair sandwich complex,27 although two
basepair sandwich models in which the DNA helix is
kinked or bent have been mooted.27 The linkers used
here comprise neutral alkyl chains, flexible polyamine
chains providing one or two cationic charges at physi-
ological pH, and a semirigid polyamine yielding two net
charges for the linker. Most of the R,ω-diamines used
are available commercially. The tetraamine required for
the C6N2 linker was purchased as its tetrahydrochlo-
ride salt, basified to the free base, distilled azeotropically
with benzene under nitrogen to give the anhydrous
agent, and used directly to avoid decomposition. 1,4-
Bis(aminoethyl)piperazine, required for the C2pipC2
linker, was prepared from piperazine, as outlined in
Scheme 5. Piperazine was reacted with chloroethaneni-
trile in the presence of sodium carbonate to give
piperazino-1,4-bis(ethanenitrile) in quantitative yield.42
Subsequent reduction of the bis-nitrile with LiAlH4 in
N-ethylmorpholine solution afforded the desired bis-
amine.43 The diacridines were finally coupled by follow-
ing a modified procedure described by Denny et al.,44
as shown in Scheme 3. The 9-chloroacridine-4-carboxa-
mides were dissolved in phenol and reacted with half a
molecular equivalent of the R,ω-diamine linkers. After
heating at 125 °C for 2 h, and suitable workup, the
diacridine free bases were recovered from chloroform
and purified by flash chromatography and recrystalli-
zation. Their corresponding hydrochloride salts were
prepared by treating the free base in methanol with dry
hydrogen chloride, followed by precipitation with ethyl
acetate.
The measured helix-unwinding angles reveal linker-
dependent variations in duplex winding in the DNA
complexes of the dimethylaminoethyl dimers, the un-
winding angles for the C8 and C3NC3 linkers being
particularly large and that for the C2pipC2 linker being
noticeably low. These findings imply the existence of
specific interactions between the linker and the DNA
base pairs that modify the fine structure of the com-
plexes. A low unwinding angle for simple nonthreading
piperazine-containing bisintercalating diacridines has
been noted before,27 which suggests particularly strong
interactions between the piperazine ring and the DNA
bases. Figure 2 also reveals that the intrinsic electro-
phoretic mobility of the closed circular DNA diminishes
at binding ratios above the equivalence point, where it
acquires ligand-induced right-handed supercoils, which
is no doubt a consequence of the increase in DNA
contour length that accompanies intercalation (see ref
27). This phenomenon is also reflected in the response
of the nicked circular species, which behaves hydrody-
namically and electrophoretically in a similar way to
linear DNA. Thus, the mobility of the nicked circular
DNA decreases linearly with increased ligand binding
(plot not shown), tending to reach a constant value at
about one ligand molecule bound per four basepairs, a
value we take as indicative of the stoichiometry of
bisintercalative binding. Generally, this plot has a
gradient of -1, an exception being the result for C3NC3
DMAE, where the slope is -1.5. Taken as a whole, the
results of the circular DNA experiments suggest that
the dimethylaminoethyl dimers, with the exception of
the C6 analogue, bisintercalate forming a “two basepair
sandwich” complex, in which the binding site extends
over a total of four basepairs, as would be expected given
the 10-11 Å length of their linkers (Figure 1). The helix
unwinding angle for the C6 DMAE dimer is consistent
with a bisthreading mode of binding, but its linker, at
8.8 Å, is too short on the basis of modeling to allow
bisintercalation without severe bending of the DNA
helix in the complex. Alternatively, it may bisintercalate
sandwiching only one basepair, thereby violating the
neighbor-exclusion principle.27 Either way, the structure
of its DNA complex, while bisintercalated, is necessarily
different from those of its congeners.
DNA Bin d in g. To establish whether the diacridines
bind to DNA by a bisintercalating threading mode, we
investigated their capacity to remove and reverse the
supercoiling of closed circular pBR 322 DNA as assessed
by electrophoretic mobility measurements on agarose
gels.27,45,46 Figure 2 shows typical titration curves for
the C8 DMAE, C6N2 DMAE, and C3NC3 morpholino
dimers and reveals equivalence binding ratios, where
all the supercoiling has been removed and the closed
and nicked circular species comigrate, in the range
0.075-0.12 ligand molecules per base pair. Similar
titration curves were found for all the compounds
studied, and equilibrium dialysis measurements indi-
cate that binding is effectively quantitative at equiva-
lence in these low ionic strength experimental conditions
(cation concentration of 30 mM). Since pBR 322 DNA
The lower basicity of the morpholino group, whose pK
lies between 7 and 8,41 leads to only partially charged

5794 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 26
Wakelin et al.
F igu r e 2. Effects of the C8 DMAE, C6N2 DMAE, and C3NC3 morpholino threading diacridines on the supercoiling of pBR 322
DNA. Measurements were made in 40 mM Tris-acetate buffer, pH 7.5 on 0.8% agarose gels electrophoresed for 75 min at 5 V/cm.
Drug to DNA ratios in lanes 1-16 for C8 DMAE are 0, 0.01, 0.02, 0.03, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.10, 0.15, 0.20, 0.25,
0.30, and 0. Drug to DNA ratio in lanes 1-16 for C2N6 DMAE and C3NC3 morpholino are 0, 0.04, 0.06, 0.08, 0.10, 0.12, 0.14,
0.16, 0.18, 0.20, 0.25, 0.30, 0.35, 0.40, 0.50, and 0.
side chains for the morpholino dimers at pH 7.5. One
consequence of this is that the circular DNA complexes
of the alkyl- and C2pipC2-linked morpholino dimers
appeared to partly dissociate on agarose gels under the
standard experimental conditions used here. Accord-
ingly, the helix unwinding angles for these three
compounds were determined at pH 6.1, where their
DNA complexes are stable to electrophoresis. In con-
trast, the remaining morpholino dimers, linked by
charged linear amines, appeared stable to electrophore-
sis at the higher pH. The results of the superhelical
titrations for the morpholino compounds are included
in Table 1 and also confirm this series of novel dia-
cridines as bisintercalating agents that, with the excep-
tion of the C6 dimer, have helix-unwinding angles
ranging from 27° to 33°. However, in distinction to the
findings with the dimethylaminoethyl series, the mea-
sured unwinding angles are largely independent of
linker structure, which implies that the acridine chro-
mophore and its pendant side chain is the most domi-
nant determinant of helical winding in bisintercalated
complexes of the morpholino dimers. Notwithstanding
this generalization, the piperazine-linked compound
remains an exception, with a low unwinding angle,
indistinguishable from that of its dimethylaminoethyl
analogue. The C6 dimer has an unwinding angle closer
to the value of its monofunctional parent, suggesting
either a highly strained bisintercalated complex or that
the ligand binds in a mixed monofunctional-bifunctional
mode.
In preliminary kinetic studies using the surfactant
sequestration method (see refs 15-20) to measure the
rates of dissociation of calf thymus DNA complexes of
selected threading dimers and actinomycin, we found

Bisintercalating DNA-Threading Agents
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 26 5795
Ta ble 1. Cytotoxicity and Helix-Unwinding Angles for the
Adding a single positive charge to the linker of the
bisthreading agent slows average dissociation rates
approximately 2-fold in both the dimethylaminoethyl
and morpholino series (Table 2), and a second positive
charge slows dissociation yet further (Table 2). These
preliminary findings, when compared with complex
lifetimes in the range 0.1-0.4 s for 9-amino-DACA and
the C8-linked nonthreading diacridine,17,44 illustrate the
profound slowing effect imparted by the threading
mechanism when set in a bisintercalating binding mode.
Compounds Studiedh
helix
cytotoxicity equivalence unwinding
IC50 (nM) binding ratio angle, deg
agent
Topoisomerase Poisons
3.6 ( 1.3
mitoxantrone
amsacrine
18a
21b
25c
NA
130 ( 39
amsacrine-4-carboxamide 120 ( 20
camptothecin
5.3 ( 1.5
NA
NA
Transcription Inhibitors
1.6 ( 1.0
actinomycin D
echinomycin
nogalamycin
chromomycin
18d
48e
12d
NA
Cytotoxicity. The relationships between cytotoxicity
and structure for the threading diacridines were ex-
plored using a conventional growth inhibition assay
with CCRF-CEM human leukemia cells. For the pur-
poses of comparison, we also investigated the cytotox-
icity of the topoisomerase II poisons mitoxantrone,
amsacrine, and amsacrinecarboxamide;1-3,20,34,35 the
topoisomerase I poison camptothecin;33 the intercalating
transcription inhibitors actinomycin D, echinomycin,
and nogalamycin;8-10,13-16 and the minor groove binding
transcription inhibitor chromomycin.49 Among the to-
poisomerase poisons, the anilinoacridines amsacrine
and amsacrinecarboxamide have similar IC50 values in
the range 120-130 nM, whereas mitoxantrone and
camptothecin are some 25-fold more potent (Table 1).
With the exception of nogalamycin, whose IC50 is 59 nM,
the transcription inhibitors exhibited IC50 values of
approximately 1 nM. In respect of the monofunctional
parents of the threading dimers, the toxicity of 9-amino-
DACA is indistinguishable from that of the anili-
noacridines, which accords with its known ability to
poison topoisomerase II,5,6 whereas the morpholino
analogue is 30-fold less potent, consistent with previous
findings in other cell lines.41 In the dimethylaminoethyl
dimer series, the C6-, C8-, C3NC3-, and C2pipC2-linked
compounds are more potent than 9-amino-DACA, by
about 2-3-fold on average, whereas the C3NC4 and
C6N2 dimers are 2- and 5-fold less active, respectively.
Thus, for the dimethylaminoethyl side chain, dimeriza-
tion can enhance or diminish biological activity, depend-
ing on the nature of the linker. In contrast, dimerization
in the morpholino series consistently enhances cytotox-
icity, in the case of the C3NC3- and C3NC4-linked
dimers, bringing potency into the same range as that
exhibited by their dimethylaminoethyl homologues.
Thus, the most active morpholino dimers are some 50-
80-fold more potent than their parental monomer. For
both the dimethylaminoethyl and the morpholino dimers,
varying the linker alters cytotoxicity by more than an
order of magnitude. The C3NC3 linker confers high
potency on both groups of compounds tested, but
otherwise different groups, for example, C8 and C2pipC2
in the case of the dimethylaminoethyl dimers, are also
associated with high cytotoxicity. Overall, neither the
dimethylaminoethyl series nor the morpholino series of
dimers is conspicuously more toxic than the other.
0.5 ( 0.1
59 ( 19
3.5 ( 0.7
DMAE Diacridines
DMAE monomer
C6 DMAE
C8 DMAE
C3NC3 DMAE
C3NC4 DMAE
C6N2 DMAE
C2pipC2 DMAE
120 ( 25
17f
87 ( 6
0.090 ( 0.010
0.075 ( 0.005
0.070 ( 0.005
0.110 ( 0.010
0.090 ( 0.050
0.125 ( 0.010
36 ( 3
43 ( 2
46 ( 2
30 ( 2
36 ( 2
26 ( 3
35 ( 5
50 ( 14
250 ( 36
590 ( 55
63 ( 2
Morpholino Diacridines
morpholino monomer
C6 morpholino
C8 morpholino
C3NC3 morpholino
C3NC4 morpholino
C6N2 morpholino
C2pipC2 morpholino
3800 ( 60
18f
710 ( 110
810 ( 180
47 ( 10
0.145 ( 0.015
22 ( 2g
30 ( 2g
29 ( 3
33 ( 3
33 ( 4
27 ( 2g
0.110 ( 0.010
0.112 ( 0.014
0.100 ( 0.010
0.100 ( 0.015
0.120 ( 0.010
75 ( 25
710 ( 61
290 ( 150
a
b
Data taken from ref 53. Data taken from ref 54. c Data taken
from ref 20. Data taken from ref 55. e Data taken from ref 9.
d
f Data taken from ref 41. Measurements made at pH 6.1. Cy-
totoxicity was measured after 72 h incubation with CCRF-CEM
human leukemia cells. Helix unwinding angles were measured by
agarose gel electrophoresis using pBR322 covalently closed circular
DNA at pH 7.5. Equivalence binding ratios are expressed in
basepair units. NA, not applicable.
g
h
Ta ble 2. Dissociation Kinetics for Calf Thymus DNA
Complexes of Selected Agents Studieda
dissociation time constants, s
compd
C8 DMAE
C3NC3 DMAE
C3NC3 morpholino
C6N2 DMAE
τ1
τ2
τ3
5
6
5
6
7
17
32
30
33
22
68
53
130
88
180
220
790
C2PipC2 DMAE
actinomycin
19
a
A 900 µL portion of a DNA-ligand complex solution compris-
ing 555 µM DNA in base pairs and 55 µM ligand in 0.1 SHE buffer
was mixed with 100 µL of 200 mM SDS solution in 0.1 SHE buffer
in a 1 mL semi-micro-quartz cuvette at 23 °C. Data were collected
on a Cary 50 spectrophotometer at wavelengths of 423 nm for the
bis(9-aminoacridine-4-carboxamides) and 430 nm for actinomycin
D. Reaction progress curves were deconvoluted into three expo-
nentials using the Levenberg-Marquardt algorithm in the Origin
v 7.0 software package.
that the complexes dissociate by a complex mechanism
involving four discernible steps, three of which can be
captured by a manual mixing technique (Table 2). The
rapid, missing component accounts for about one-third
of the total reaction in buffer of ionic strength 0.1 at 23
°C. Under these conditions, actinomycin dissociates from
DNA with resolvable time constants of 19, 69, and 790
s, and the C8 DMAE diacridine dissociates with time
constants of 5, 17, and 53 s. Taking into consideration
the relative amplitudes of the individual transients
(data not shown), the weighted mean dissociation rate
of the DNA complex of the C8 DMAE dimer is ap-
proximately 6-fold faster than that of actinomycin.
G2/M Ar r est. Flow cytometric analysis was specifi-
cally undertaken to assess the impact of the threading
dimers on cell cycle progression. However, the respective
analyses also provide information, from detection of cells
having less than G1 DNA content, on occurrence of cell
death, and this matter will be considered separately.
Flow cytometry was performed after incubation for 24
and 48 h in the presence of agents at concentrations

5796 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 26
Wakelin et al.
Ta ble 3. Effects of Compounds Studied on Proportion (%) of
CCRF-CEM Cells at G2/M Determined by Flow Cytometry
response is clearly distinguished from that obtained
using the transcription inhibitors actinomycin D, chro-
momycin, echinomycin, and nogalamycin (Table 3), for
which there was no evidence for accumulation of cells
in G2/M. With regard to the threading dimethylamino-
ethyl dimers, the parental monomer, 9-amino-DACA,
consistently caused G2/M arrest, but at generally higher
doses than required for the other topoisomerase poisons
(Table 3). With respect to the novel agents, two response
patterns were observed: either G2/M arrest or no effect
on cell cycle distribution. Typical flow cytometric profiles
are shown in Figure 1 of the Supporting Information,
and the data for all the agents studied are summarized
in Table 3. With the exception of the C8 and C2pipC2
members, which appear to have no impact on cell cycle
distribution, the dimethylaminoethyl dimers caused
G2/M arrest, at least when they were assessed at high
concentrations (Figure 1a of the Supporting Informa-
tion, Table 3). The results at lesser concentrations
typically indicate an increase in the proportion of cells
at G2/M, but not to the extent of reaching the arbitrary
level adopted on the basis of the established topoi-
somerase poisons. In every case, increasing the concen-
tration of the agent enhanced, or at least maintained,
the size of the G2/M peak, in a manner similar to that
found for 9-amino-DACA. By contrast, the behavior of
the morpholino-based agents is generally quite different.
The monomer and all the dimers, with the exception of
the C6-linked member, showed no discernible effect on
cell cycle distribution under the conditions employed in
these studies (Figure 1b of the Supporting Information,
Table 3). Accumulation of cells in G2/M was observed
using the C6 morpholino dimer at the highest concen-
tration tested.
24 h
48 h
G2/M
b
compound
1× 2× 10× 1× 2× 10× arrestc
Topoisomerase Poisons
mitoxantrone
amsacrine
amsacrine-4-carboxamide 83 73
32 53 83 24 32 50 yes: 2×
51 47 57 35
d
35 42
31 47
52 yes: 1×
d
d
d
d
yes: 1×
yes: 2×
camptothecin
29 48
Transcription Inhibitorsa
actinomycin D
21 20 28 20 22 11 no
DMAE Diacridines
DMAE monomer
C6 DMAE
C8 DMAE
C3NC3 DMAE
C3NC4 DMAE
C6N2 DMAE
C2pipC2 DMAE
20 30 61 18 22 53 yes: 10×
17 34 34 12 43 39 yes: 2×
20 23 29 17 21 21 no
20 19 44 19 20 46 yes: 10×
19 30 52 14 25 81 yes: 10×
19 25 61 16 16 69 yes: 10×
18 18 20 21 19 20 no
Morpholino Diacridines
morpholino monomer
C6 morpholino
C8 morpholino
C3NC3 morpholino
C3NC4 morpholino
C6N2 morpholino
C2pipC2 morpholino
17 17 21 10 18 18 no
25 29 38 23 25 44 yes: 10×
19 19 24 14 17 24 no
18 15 13 12 12 13 no
17 15 17 11 13 11 no
24 27 24 24 27 23 no
27 24 28 23 22 23 no
a
Data for transcription inhibitors echinomycin D, nogalamycin,
and chromomycin have been omitted since the findings were as
b
shown for actinomycin D (no variation from control). Cells were
incubated for 24 or 48 h in the presence of the compound shown
at concentrations equal to the IC50 (see Table 1) or at two or 10
times that concentration, control preparations being left untreated.
No consistent differences between 24 and 48 h control values were
evident (total ) 44 determinations), the G2/M proportion being
almost always in the range 17-22% and never more than 25%.
c Occurrence of G2/M arrest was arbitrarily determined on the
b
basis of the G2/M proportion being 40% or more. Occurrence of
cell death at a level sufficient to preclude identification and
computation of any G2/M peak in the flow cytometric profile.
Kin etics of Cell Dea th . In the course of the flow
cytometric analyses summarized in Table 3, we were
able to discern accumulation of cells having less than
G1 DNA content in some instances. Although the flow
cytometric studies involved equitoxic concentrations
based on 72 h incubation, it was evident that the
appearance of dead cells at earlier times was nonran-
domly associated with particular agents. By the “sub-
G1 DNA peak” criterion, death of the entire cell popu-
lation after incubation for 24 or 48 h was only observed
with the recognized topoisomerase poisons amsacrine,
amsacrine-4-carboxamide, and camptothecin. These
results were as expected, insofar as the effect was
observed at the highest concentration employed (10
times), and was more marked after 48 h than at 24 h.
“Early death” results obtained with the transcription
inhibitors chromomycin, echinomycin, actinomycin D,
and nogalamycin were similar to those found for the
threading dimers and mitoxantrone. In these cases, with
the exception of the amine-linked morpholino dimers,
some cell death was evident in all instances, at least
using the highest concentration of the agent after
incubation for 48 h, but DNA profiles were never
distorted beyond recognition. The amine-linked mor-
pholino dimers, along with 9-amino-DACA, appear not
to induce evident cell death following exposure of CCRF-
CEM cells for up to 48 h.
ranging from the IC50 (based on 72 h incubation),
upward to two, five, or 10 times that level. It was readily
apparent, with respect to both the established drugs
employed for control purposes and also for the novel
agents under test, that when treatment-related alter-
ation of the flow cytometric profile did occur, the pattern
was consistent. Such change invariably involved an
increase in the proportion of cells at G2/M, and when
observed, accumulation was typically time- and concen-
tration-dependent within the limitations of the param-
eter selected for study. Thus, to facilitate data presen-
tation, we have focused on the proportion of cells at G2/
M, respective values being adopted after reviewing each
profile and confirming that the numeric value computed
is identifiable with an evident “peak”. If no such peak
was apparent, a scenario most often synonymous with
cell death and complete loss of the typical profile, no
value was adopted for “percentage of cells in G2/M”.
Otherwise, G2/M arrest was deemed to have occurred
if the result was more than 40%, which compares with
a value of about 21% consistently observed in prepara-
tions of cells not subject to drug treatment.
The topoisomerase poisons mitoxantrone and camp-
tothecin and the amsacrine-related agents all caused
G2/M arrest by the standard adopted (see Figure 1c of
the Supporting Information and Table 3). These agents
exhibit the effect at relatively low concentrations, typi-
cally one or two times the IC50 values, and G2/M arrest
is readily evident after incubation for 24 h. This
Discu ssion
DNA Bin d in g a n d Cytotoxicity. In this work we
have designed and synthesized a series of bis(9-ami-

Bisintercalating DNA-Threading Agents
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 26 5797
noacridine-4-carboxamides) and shown that they bisin-
tercalate into DNA via a threading mode, whereby the
carboxamide side chains come to lie in one groove of the
DNA and the linker lies in the other. Preliminary
kinetic studies with selected members indicate that the
dimers dissociate orders of magnitude more slowly from
DNA than simple nonthreading diacridines, as postu-
lated, which implies the involvement of the carboxamide
side chains in interactions with the DNA bases. Ad-
ditional NMR or crystallographic studies are required
to establish that these interactions are with guanines,
as anticipated. Notwithstanding the fact that the thread-
ing dimers share a common mode of binding, the
relationships between cytotoxicity, ligand structure, and
helix unwinding are complex, allowing few, if any,
generalizations to be made across the group as a whole.
Rather, the various subgroups must be considered.
Thus, among members of the dimethylaminoethyl se-
ries, the most active compounds are the C8-, C3NC3-,
and C2pipC2-linked dimers which are two to three times
more potent than 9-amino-DACA and, respectively, have
a neutral, a flexible singly charged, and a semirigid
doubly charged linker, indicating biological efficacy
across a wide range of linker types (Table 1). However,
that subtle structural features are important is made
clear by the finding that adding a single methylene
group to the C3NC3 linker, to give C3NC4, lowers
toxicity 5-fold and removing the ethyl cross-link from
C2pipC2 to give the flexible C6N2 linker suppresses it
by an order of magnitude. In addition, dichotomies are
found when comparing helix-winding effects with cyto-
toxicity data, which are best illustrated by the observa-
tion that the C8 and C3NC3 dimers have the highest
unwinding angles, 43° and 46°, whereas C2pipC2 has
the lowest, 26°. When considering the morpholino series,
we found, as noted by Atwell et al.,41 that the parental
monomer is much less cytotoxic than 9-amino-DACA.
However, when incorporated into threading dimers, the
cytotoxicity of the resulting bismorpholino compounds
is substantially enhanced in a manner dependent on
linker structure, the singly charged amine linkers
C3NC3 and C3NC4 having potencies comparable to
those of the most active members of the dimethylami-
noethyl series. Adding an additional charge, making the
linker neutral, or making it more rigid all compromise
cytotoxicity (Table 1). Resolving these difficulties in
understanding the structure-activity relationships for
the bis(9-aminoacridine-4-carboxamides) is made more
complex by the finding that the dimers engage apoptosis
by at least two distinct mechanisms, as discussed below.
hand and caspase activation as initiating apoptosis on
the other have not been defined3. Critical to such
pathways is disruption of cell cycle progression and the
adequacy, or otherwise, of cell cycle checkpoints.50 It is
logical, therefore, that processes associated with topo-
isomerase function, cell cycle arrest, and apoptosis
induction, be specifically viewed as focal points for
cancer drug development.51 The observations recorded
here, however, indicate that direct and causal relation-
ships between various steps triggered by drug-target
interaction are not invariably observed. While the
sequence drug-target interaction, cell cycle arrest, and
cell death could be delineated in relation to the recog-
nized topoisomerase poisons used, this sequence varied
and/or was only partially exhibited by the bis(9-ami-
noacridine-4-carboxamides) and the established tran-
scription inhibitors.
In common with etoposide,32 G2/M arrest of CCRF-
CEM cells was readily evident within 24 h of exposure
to mitoxantrone, amsacrine, amsacrine-4-carboxamide,
and camptothecin, a pattern of response that distin-
guished these agents from the known transcription
inhibitors used (Table 3). Although G2/M arrest is
commonly observed in response to a range of DNA-
damaging drugs, the poisoning of topoisomerase II is
well-recognized to precipitate such an event,30 which
accords with our findings. In contrast, the recognized
transcription inhibitors and the bisintercalating mor-
pholino dimer series (with the exception of the C6-
member, whose DNA complex, as illustrated by its low
helix unwinding angle, is somewhat different to those
of its longer homologues) fail to arrest cells in G2/M.
This necessarily means these agents engage apoptosis
by a different mechanism than the topoisomerase
poisons, and this finding may be taken to imply that
the morpholino dimers, by analogy to the transcription
template inhibitors and consistent with their design
principles, are cytotoxic as a consequence of inhibiting
transcription. The situation with the dimethylamino-
ethyl dimers and cycle progression is more complex.
With the exception of the C8- and the C2pipC2-linked
DMAE ligands, these agents arrest cells in G2/M in a
manner similar to the recognized topoisomerase poisons,
which suggests the involvement of topoisomerase activ-
ity in their mechanism of action. By analogy, this
finding also implies that the C8 DMAE and C2pipC2
DMAE dimers engage apoptosis through a transcrip-
tion-initiated pathway, in a manner similar to the
morpholino series. To substantiate this conclusion we
have begun studies of the effects of the bis(9-aminoacri-
dine-4-carboxamides) on global gene expression using
DNA microarrays. The preliminary data are limited but
indicate that, as a class, these agents profoundly perturb
gene expression, with the concentration of many tran-
scripts being diminished, while the concentration of
others is enhanced. Clearly, further studies are needed,
both with the established template inhibitors and the
most toxic of the threading dimers, to elucidate rela-
tionships between ligand structure, gene expression,
topoisomerase poisoning, cycle arrest, and apoptosis.
G2/M Ar r est a n d th e Mech a n ism of Cell Dea th .
The mechanism of action of cytotoxic drugs most com-
monly involves perturbations to DNA synthesis, tran-
scription, or mitosis, the disruption of any of these
processes being sufficient to induce apoptosis in sus-
ceptible cells. The requisite signaling pathway following
formation of the critical drug-target complex has
proven difficult to elucidate for cytotoxins in general and
specially so for transcription template inhibitors such
as actinomycin D, because these drugs are often used
as “tools” for dissecting such pathways by establishing
a requirement for transcription in particular contexts.
Likewise, the pathways between disruption of topo-
isomerase function and associated DNA damage on one
Con clu sion s
In summary, our findings show that both the dim-
ethylaminoethyl and morpholino dimer series bisinter-

5798 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 26
Wakelin et al.
without further purification, where portions of it were added
over 40 min (total, 15.0 g, 67.5 mmol) to a stirred solution of
concentrated sulfuric acid (75 g) at 60-65 °C. The resulting
mixture was heated at 95 °C for 1 h and then poured into ice-
water after being cooled. The precipitate that formed was
filtered off and dissolved in 1 M sodium hydroxide, and
insoluble material was removed by filtration. The filtrate was
then acidified with concentrated hydrochloric acid to pH 2, and
the solid that precipitated filtered off, washed with cold water,
and dried. Isatin-7-carboxylic acid was obtained as an orange
solid (10.1 g, 78%), mp 274-275 °C, (lit.52 mp 276-277 °C).
o-Am in oisop h th a lic Acid . Isatin-7-carboxylic acid (9.55
g, 50 mmol) was treated with 3% hydrogen peroxide (200 mL)
and 10% sodium hydroxide (200 mL). The resulting mixture
was stirred at room temperature for 2 h and then acidified
with concentrated hydrochloric acid to pH 2. The precipitate
that formed was collected by filtration, washed with cold water,
and dried. The product was obtained as a yellow solid (6.60 g,
73%), mp 226-227 °C (from acetic acid), (lit.38 mp 227-228
calate into DNA by a threading mode, in which they
place the 4-carboxamide side chain in one groove, and
the interchromophore linker in the other. The com-
pounds appear to have binding sites extending over four
basepairs in which the middle two are sandwiched by
the acridine chromophores. This binding mode results
in a substantial slowing of dissociation rates, as indi-
cated by preliminary kinetic findings, so that the
lifetimes of the bisintercalated complexes begin to
approach those of known template transcription inhibi-
tors. However, despite common structural features,
cytoxicity varies markedly within both the dimethyl-
aminoethyl and the morpholino series, depending on
linker type, although potency is similar for the most
active compounds in each class. The latter, with IC50
values around 50 nM, are C8 DMAE, C3NC3 DMAE,
C2pipC2 DMAE, C3NC3 morpholino, and C3NC4 mor-
pholino. With the exception of C3NC3 DMAE, none of
these compounds, in common with established tran-
scription inhibitors, perturb cell cycle progression, from
which we infer that their mechanism of action does not
involve poisoning of topoisomerases. By contrast, C3NC3
DMAE, like established topoisomerase poisons, causes
cells to arrest in G2/M, which we take to imply at least
partial involvement of topoisomerase poisoning in its
mode of action. Thus, specifically in relation to the most
cytotoxic compounds, the common mode of bisinterca-
lation evokes complex cellular responses that vary
according to side chain and linker structure. A detailed
explanation for the manner in which particular struc-
tural features determine response patterns and ultimate
toxicity is not available. However, the extent to which
apparently minor structural change perturbs toxicity
encourages evaluation of more members of this class of
agent in order to explore more fully their pharmacologi-
cal potential.
1
°C). H NMR (DMSO-d6): 6.57 (t, 1H, J ) 7.8 Hz, H-5), 8.00
(d, 2H, J ) 7.8 Hz, H-4 and H-6).
o-Iod oisop h th a lic Acid . o-Iodoisophthalic acid was pre-
pared according to literature procedures,40 mp 239-240 °C
1
(lit.40 mp 240-241 °C). H NMR (DMSO-d6): 7.06 (t, 1H, J )
7.9 Hz, H-5), 7.89 (d, 2H, J ) 7.9 Hz, H-4 and H-6), 11.51 (br
s, 2H, 2CO2H).
2-(P h en yla m in o)isop h th a lic Acid . This compound was
made by J ourdan-Ullman reaction of the o-iodoisophthalic
acid and aniline as previously described,40 mp 250-251 °C
(lit.40 mp 247-248 °C). 1H NMR (DMSO-d6): 6.85-6.90 (m,
3H, 3ArH), 6.99 (t, H, J ) 7.9 Hz, ArH), 7.18 (t, 2H, J ) 7.7
Hz, 2ArH), 7.91 (d, 2H, J ) 7.9 Hz, 2ArH), 9.62 (br s, 1H, NH),
12.94 (br s, 2H, 2CO2H).
9-Oxoa cr id a n -4-ca r boxylic Acid . Cyclization of the above
2-(phenylamino)isophthalic acid in polyphosphoric acid at 120
°C gave the title product as a yellow solid, mp >300 °C (lit.
mp 324-325 °C). 1H NMR (DMSO-d6): 7.29-7.36 (m, 2H,
2ArH), 7.74-7.76 (m, 2H, 2ArH), 8.22 (d, H, J ) 8.3 Hz, ArH),
8.42 (d, 1H, J ) 7.1 Hz, ArH), 8.51 (d, 1H, J ) 7.9 Hz, ArH),
11.93 (s, 1H, NH).
9-Ch lor oa cr id in e-4-ca r bon yl Ch lor id e. This intermedi-
ate was prepared from 9-oxoacridan-4-carboxylic acid by
following the literature procedure of Atwell et al.,41 and was
used in the next step without further purification.
Exp er im en ta l Section
N-[2-(N-Mor p h olin o)eth yl]-9-ch lor oa cr id in e-4-ca r box-
a m id e. An ice-cold solution of 4-aminoethylmorpholine (4.88
g, 37.5 mmol) in dichloromethane (10 mL) was added over 5
min to a stirred suspension of the above carbonyl chloride (7.5
mmol) in dichloromethane (10 mL) cooled on ice. The resulting
solution was stirred at room temperature for 2 h, washed twice
with 10% sodium carbonate and then with brine, and finally
dried over sodium sulfate. Removal of the solvent in vacuo gave
the product as a yellow solid (2.73 g), which was recrystallized
from benzene/petroleum ether as a yellow powder (2.27 g,
82%), mp 123-124 °C. 1H NMR: 2.63-2.66 (m, 4H, 2CH2),
2.79 (t, 2H, J ) 6.2 Hz, CH2), 3.80-3.86 (m, 6H, 3CH2), 7.69-
7.79 (m, 2H, 2ArH), 7.89 (dt, J ) 6.8, 1.5 Hz, ArH), 8.34 (d,
1H, J ) 8.5 Hz, ArH), 8.47 (d, 1H, J ) 8.3 Hz, ArH), 8.63 (dd,
1H, J ) 8.5, 1.5 Hz, ArH), 9.01 (dd, 1H, J ) 6.8, 1.5 Hz, ArH),
11.83 (br s, 1H, CONH).
Ch em istr y. Melting points were measured on a Reichert
“thermopan” microscope hot-stage apparatus up to 300 °C and
are recorded uncorrected. Microanalyses were performed by
the Microanalytical Unit, Australian National University,
Canberra, Australia. NMR spectra were recorded on a Bruker
AM-300 spectrometer operating at 300.17 and 75.48 MHz for
1H and 13C, respectively. Chemical shifts are reported as δ
values in ppm relative to internal tetramethylsilane. The
solvent was deuterated chloroform, unless otherwise specified.
DEPT-135 was used to identify proton-bound carbon atoms
in 13C NMR spectra. MALDI-TOF mass spectra were obtained
using a hydroxycinnamic acid matrix on a PerSeptive Biosys-
tems Voyager mass spectrometer. All solvents were laboratory
grade and used without further purification unless otherwise
stated. Phenol was redistilled and kept under nitrogen in the
dark. Dichloromethane was stored over anhydrous calcium
chloride, refluxed for 4 h, and then distilled and stored over
4A molecular sieves. All chemical reagents were used as
obtained from commercial suppliers without further purifica-
tion unless otherwise indicated.
Isa tin -7-ca r boxylic Acid . A solution of methyl anthra-
nilate (31.5 g, 0.21 mol), chloral hydrate (35.0 g, 0.21 mol),
and hydroxylamine hydrochloride (28.6 g, 0.41 mol) in con-
centrated sulfuric acid (25 g) and water (1.4 l) was heated at
95 °C with stirring for 10 min and then kept at 4 °C for 16 h.
The precipitate that formed on cooling was collected by
filtration, washed with cold water, and dried. Methyl 2-
{[(hydroxyimino)acetyl]amino}benzoate was obtained as a
cream-colored solid (31.2 g, 67% yield), mp 174-175 °C (lit.52
mp 180 °C). This intermediate was used in the next step
N-[2-(N-Dim eth yla m in o)eth yl]-9-ch lor oa cr id in e-4-ca r -
boxa m id e. The title compound was prepared from the above
carbonyl chloride and N,N-dimethylethylenediamine, in the
same manner as N-[2-(N-morpholino)ethyl]-9-chloroacridine-
4-carboxamide and was used in following steps without further
1
purification. H NMR: 1.34 (s, 6H, N(CH3)2), 1.61 (t, 2H, J )
5.7 Hz, CH2), 2.68 (q, 2H, J ) 5.7 Hz, CH2), 6.54-6.62 (m,
2H, 2ArH), 6.73 (t, 1H, J ) 6.4 Hz, ArH), 7.08 (d, 1H, J ) 8.2
Hz, ArH), 7.27 (d, 1H, J ) 8.5 Hz, ArH), 7.42 (dd, 1H, J ) 8.2,
1.5 Hz, ArH), 7.86 (dd, 1H, J ) 6.4, 1.1 Hz, ArH), 10.76 (s,
1H, CONH).
N-[2-(N-Mor p h olin o)eth yl]-9-a m in oa cr id in e-4-ca r box-
a m id e. Dry ammonia gas was bubbled through a solution of
N-[2-(N-morpholino)ethyl]-9-chloroacridine-4-carboxamide (0.37

Bisintercalating DNA-Threading Agents
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 26 5799
67.58; H, 7.56; N, 14.80. C42H50N8O‚2.5H2O requires: C, 67.81;
g, 1.0 mmol) in anhydrous phenol (3 g) at 115 °C for 30 min.
After cooling, the reaction mixture was stirred with 20%
sodium hydroxide (20 mL), and the resulting suspension was
extracted with chloroform. The organic extract was then
washed with saturated sodium carbonate and dried over
sodium sulfate. Evaporation of the solvent gave the 9-amino
compound, which was recrystallized from benzene/petroleum
1
H, 7.45; N, 15.06. H NMR: 1.35-1.37 (m, 4H, 2CH2), 1.67-
1.69 (m, 4H, 2CH2), 2.39 (s, 12H, 2N(CH3)2), 2.67 (t, 4H, J )
6.2 Hz, 2CH2), 3.70-3.73 (m, 8H, 4CH2), 5.38 (br s, 2H, 2NH),
7.27-7.32 (m, 4H, 4ArH), 7.63 (t, 2H, J ) 7.5 Hz, 2ArH), 7.97-
8.02 (m, 4H, 4ArH), 8.12 (d, 2H, J ) 8.2 Hz, 2ArH), 8,78 (br s,
2H, 2ArH), 12.52 (br s, 2H, 2CONH). MALDI-MS: 699.0 (M
+ 1). Its tetrahydrochloride salt was obtained as a yellow solid,
mp 210-212 °C.
C3NC3-m or p h olin o Dim er . This was prepared as for the
C8-morpholino dimer, from N-[2-(N-morpholino)ethyl]-9-chlo-
roacridine-4-carboxamide and N-(3-aminopropyl)-1,3-propanedi-
amine, and then flash chromatographed on alumina (ethyl
acetate/diisopropylamine, v/v ) 10/1 and then methanol/
diisopropylamine, v/v ) 5/1) as an orange solid (52% yield),
mp 127-129 °C. Found: C, 65.23; H, 6.88; N, 14.9. C46H55N9O4‚
2.5H2O requires: C, 65.53; H, 7.17; N, 14.99. 1H NMR: 1.93-
1.96 (m, 4H, 2CH2), 2.60-2.63 (m, 8H, 4CH2), 2.73-2.77 (m,
4H, 2CH2), 2.87-2.90 (m, 4H, 2CH2), 3.78-3.81 (m, 12H,
6CH2), 7.23-7.25 (m, 4H, 4ArH), 7.61 (m, 2H, 2ArH), 8.06-
8.21 (m, 6H, 6ArH), 8.79 (m, 2H, 2 ArH), 12.52 (br s, 2H,
2CONH). MALDI-MS: 797.9 (M + 1). Its pentahydrochloride
salt was obtained as a yellow solid, mp 226-229 °C.
1
ether as a yellow solid in 96% yield. H NMR: 2.62-2.65 (m,
4H, 2CH2), 2.78 (t, 2H, J ) 6.0 Hz, CH2), 3.79-3.83 (m, 6H,
3CH2), 6.05 (br s, 2H, NH2), 7.31-7.41 (m, 2H, 2ArH), 7.70 (t,
1H, J ) 7.5 Hz, ArH), 7.93 (d, 1H, J ) 8.5 Hz, ArH), 8.01-
8.05 (m, 2H, 2ArH), 8.82 (d, 1H, J ) 6.8 Hz, ArH), 12.61 (br s,
1H, CONH). The di-hydrochloride salt of this product was
prepared by treating its methanol solution with hydrogen
chloride gas followed by precipitation with ethyl acetate to give
a yellow solid, mp 283-284 °C, (lit.41 mp 282-284 °C).
N-[2-(N-Dim eth yla m in o)eth yl]-9-a m in oa cr id in e-4-ca r -
boxa m id e (9-a m in o-DACA). The title compound was pre-
pared in 77% yield from its corresponding 9-chloro precursor
as a yellow solid, by the procedure used to make the mor-
pholino derivative described above. Its dihydrochloride salt was
obtained as a yellow solid, mp 288-290 °C (lit.41 mp 292-293
°C).
C8-DMAE Dim er : Gen er a l P r oced u r e for Syn th esis of
Dia cr id in es. A mixture of N-[2-(N-dimethylamino)ethyl]-9-
chloroacridine-4-carboxamide (0.35 g, 1.0 mmol), 1,8-diami-
nooctane (72 mg, 0.50 mmol), and phenol was heated at 125
°C with stirring for 2 h. After cooling, the reaction mixture
was added to a large excess of diethyl ether and the organic
solution decanted. The residue was washed with diethyl ether
and dissolved in 0.5 M hydrochloric acid, and insoluble
material was removed by filtration. The filtrate was then
basified with 10% sodium hydroxide and extracted twice with
chloroform. The combined organic extracts were washed with
10% sodium carbonate and brine and dried over sodium
sulfate. Evaporation of the solvent gave the product as an
orange oil which solidified gradually. The solid was recrystal-
lized from methanol and then flash chromatographed on
alumina (ethyl acetate then methanol/chloroform (v/v ) 1/1)
to yield a yellow solid in 55% yield, mp 92 °C. Found: C, 70.94;
H, 7.39; N, 14.80. C44H54N8O‚H2O requires: C, 70.94; H, 7.58;
N, 15.04. 1H NMR: 1.25-1.32 (m, 8H, 4CH2), 1.69 (m, 4H,
2CH2), 2.40 (s, 12H, 2N(CH3)2), 2.68 (t, 4H, J ) 6.1 Hz, 2CH2),
3.75 (q, 8H, J ) 6.1 Hz, 4 CH2), 5.29 (br s, 2H, 2NH), 7.33-
7.41 (m, 4H, 4ArH), 7.68 (t, 2H, J ) 7.5 Hz, 2ArH), 8.04 (d,
4H, J ) 8.7 Hz, 4ArH), 8.17 (d, 2H, J ) 8.5 Hz, 2ArH), 8.85
(d, 2H, J ) 6.8 Hz, 2ArH), 12.53 (br s, 2H, 2CONH). 13C
NMR: 26.5 (2CH2), 28.9 (2CH2), 31.5 (2CH2), 37.9 (2CH2), 45.4
(2N(CH3)2), 50.9 (2CH2), 58.3 (2CH2), 115.6 (2C), 116.4 (2C),
122.0 (2CH), 122.3 (2CH), 123.3 (2CH), 126.4 (2CH), 128.3
(2C), 129.5 (2CH), 130.4 (2CH), 134.4 (2CH), 147.2 (2C), 147.9
(2C), 152.4 (2C), 166.4 (2C). MALDI-MS: 727.0 (M + 1). Its
tetrahydrochloride salt was obtained as a yellow solid, mp
186-188 °C. The following diacridines were prepared in a
similar way.
C6-m or p h olin o Dim er . This was prepared as for the C8-
DMAE dimer, from N-[2-(N-morpholino)ethyl]-9-chloroacri-
dine-4-carboxamide and hexamethylenediamine, and then
recrystalized from ethyl acetate as a yellow solid (75% yield),
mp 200-202 °C. Found: C, 69.17; H, 6.74; N, 13.99. C46H54N8O4‚
H2O requires: C, 68.98; H, 7.05; N, 13.99. 1H NMR: 1.45-
1.47 (m, 4H, 2CH2), 1.76-1.78 (m, 4H, 2CH2), 2.61-2.63 (m,
8H, 4CH2), 2.76 (t, 4H, J ) 6.2 Hz, 2CH2), 3.79-3.81 (m, 16H,
8CH2), 5.21 (s, 2H, 2NH), 7.36-7.45 (m, 4H, 4ArH), 7.70 (t,
2H, J ) 7.5 Hz, 2ArH), 8.03-8.19 (m, 6H, 6ArH), 8.87 (d, 2H,
J ) 6.8 Hz, 2ArH), 12.44 (s, 2H, 2CONH). MALDI-MS: 782.9
(M + 1). Its tetrahydrochloride salt was obtained as a yellow
powder, mp 217-219 °C.
C3NC3-DMAE Dim er . This was prepared as for the
C3NC3-morpholino dimer, from N-[2-(N-dimethylamino)ethyl]-
9-chloroacridine-4-carboxamide and N-(3-aminopropyl)-1,3-
propane-diamine, and then flash chromatographed on alumina
(ethyl acetate/diisopropylamine, v/v ) 10/1 and then methanol/
diisopropylamine, v/v ) 5/1). The product obtained after
column chromatography was still impure and was further
treated with hot acetonitrile followed by cooling on ice. The
dimer was now obtained as a sticky oil, which gradually
solidified as a yellow solid (37% yield), mp 103-105 °C. 1H
NMR: 1.97-2.02 (m, 4H, 2CH2), 2.40 (s, 12H, 2N(CH3)2), 2.71
(t, 4H, J ) 5.9 Hz, 2CH2), 2.87-2.91 (m, 4H, 2CH2), 3.70-
3.74 (m, 4H, 2CH2), 3.92-3.97 (m, 4H, 2CH2), 5.42 (br s, 3H,
3NH), 71.6-7.21 (m, 4H, 4ArH), 7.56 (t, 2H, J ) 7.4 Hz, 2ArH),
7.80 (d, 2H, J ) 8.3 Hz, 2ArH), 8.09 (d, 2H, J ) 8.3 Hz, 2ArH),
8.22 (d, 2H, J ) 7.9 Hz, 2ArH), 8.53-8.56 (m, 2H, 2ArH), 11.80
(br s, 2H, 2CONH). MALDI-MS: 713.7 (M + 1). The mi-
croanalysis was carried out on its pentahydrochloride salt, a
yellow solid, mp 246-249 °C. Found: C, 51.42; H, 6.86; N,
12.90. C42H51N9O2‚5HCl‚5H2O requires: C, 51.15; H, 6.74; N,
12.78.
C4NC3-m or p h olin o Dim er . This was prepared as for the
C8-morpholino dimer, from N-[2-(N-morpholino)ethyl]-9-chlo-
roacridine-4-carboxamide and spermidine, and then purified
by column chromatography (alumina, ethyl acetate then
methanol/chloroform). The product was obtained as a yellow
solid in 62% yield, mp 82-83 °C. 1H NMR: 1.67-1.72 (m, 2H,
CH2), 1.84 (t, 4H, J ) 6.0 Hz, 2CH2), 2.59 (br s, 8H, 4CH2),
2.67-2.75 (m, 6H, 3CH2), 2.88 (t-like, 2H, CH2), 3.76-3.84 (m,
14H, 7CH2), 3.96 (t-like, 2H, CH2), 7.18 (t, 2H, J ) 7.7 Hz,
2ArH), 7.26-7.31 (m, 2H, 2ArH), 7.55-7.65 (m, 2H, 2 ArH),
7.97-8.10 (m, 4H, 4 ArH), 8.17 (d, 2H, J ) 8.1 Hz, 2 ArH),
8.70-8.78 (m, 2H, 2ArH), 12.48 (br s, 2H, 2CONH). MALDI-
MS: 811.6 (M + 1). The microanalysis was carried out on its
pentahydrochlorde salt, a yellow solid, mp 232-234 °C.
Found: C, 49.79; H, 6.58; N, 11.02. C47H57N9O4‚5HCl‚8H2O
requires: C, 49.59; H, 6.90; N, 11.07.
C4NC3-DMAE Dim er . This was prepared as for the
C4NC3-morpholino dimer, from N-[2-(N-dimethylamino)ethyl]-
9-chloroacridine-4-carboxamide and spermidine. The product
obtained after column chromatography (alumina, ethyl acetate
then methanol/chloroform) was still not pure in this case, but
it was found that the impurities could be extracted with warm
ethyl acetate. The pure product was finally obtained as a
yellow solid (in 41% yield) after recrystalization from ethyl
acetate/methanol, mp 143-145 °C. Found: C, 67.97; H, 7.31;
N, 16.28. C43H53N9O2‚2H2O requires: C, 67.60; H, 7.52; N,
C6-DMAE Dim er . This was prepared as for the C6-DMAE
dimer, from N-[2-(N-dimethylamino)ethyl]-9-chloroacridine-4-
carboxamide and hexamethylenediamine, and then flash chro-
matographed on alumina (ethyl acetate then methanol/
chloroform) and recrystallized from ethyl acetate/petroleum
ether as a yellow solid (60% yield), mp 98-100 °C. Found: C,
1
16.50. H NMR: 1.69-1.74 (m, 2H, CH2), 1.82-1.89 (m, 4H,
2CH2), 2.40 (s, 12H, 2N(CH3)2), 2.67-2.72 (m, 6H, 3CH2), 2.91
(t, 2H, J ) 5.8 Hz, CH2), 3.73-3.76 (m, 4H, 2CH2), 3.85 (t,
2H, J ) 6.2 Hz, CH2), 4.01 (t, 2H, J ) 5.8 Hz, CH2), 7.20-

5800 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 26
Wakelin et al.
7.36 (m, 4H, 4ArH), 7.57-7.66 (m, 2H, 2ArH), 7.92-8.23 (m,
6H, 6ArH), 8.70-8.76 (m, 2H, 2ArH), 12.43 (br s, 2H, 2CONH).
MALDI-MS: 728.0 (M + 1). Its pentahydrochloride salt was
obtained as a pale yellow solid, mp 239-242 °C.
C48H58N10O4‚2H2O requires: C, 65.88; H, 7.14; N, 16.01. 1H
NMR: 2.62-2.80 (m, 24H, 12CH2), 3.80-3.83 (m, 12H, 6CH2),
3.93-3.97 (m, 4H, 2CH2), 6.69 (br s, 2H, 2NH), 7.38-7.48 (m,
4H, 4ArH), 7.72 (t, 2H, J ) 7.5 Hz, 2ArH), 8.13-8.21 (m, 4H,
4ArH), 8.31 (d, 2H, J ) 8.4 Hz, 2ArH), 8.87 (d, 2H, J ) 7.1
Hz, 2ArH), 12.48 (br s, 2H, 2CONH). MALDI-MS: 839.3 (M
+ 1). Its hexahydrochloride salt was obtained as a yellow solid,
mp 271 °C (dec).
C6N2-m or p h olin o Dim er . This was prepared as for the
C8-morpholino dimer, from N-[2-(N-morpholino)ethyl]-9-chlo-
roacridine-4-carboxamide and anhydrous triethylenetetramine
(obtained by azeotropic distillation with benzene for 16 h). The
crude product obtained after normal workup was flash chro-
matographed on alumina gel (ethyl acetate and then methanol/
chloroform) and then recrystallized from ethyl acetate as a
C2p ip C2-DMAE Dim er . This was prepared as for the C8-
DMAE dimer, from N-[2-(N-dimethylamino)ethyl]-9-chloro-
acridine-4-carboxamide and 1,4-bis(2-aminoethyl)piperazine,
and purified by column chromatography (alumina, ethyl
acetate/diisopropylamine (v:v ) 50:1) and then methanol/
chloroform). The free base was obtained as a yellow solid (in
48% yield), which become dark on standing in the air. 1H
NMR: 2.42 (s, 12H, 2N(CH3)2), 2.68-2.76 (m, 16H, 8CH2), 3.77
(q, 4H, J ) 5.7 Hz, 2CH2), 3.92 (t, 4H, J ) 5.7 Hz, 2CH2), 6.68
(br s, 2H, 2NH), 7.38-7.45 (m, 4H, 4ArH), 7.70 (t, 2H, J ) 7.4
Hz, 2ArH), 8.05 (d, 2H, J ) 8.6 Hz, 2ArH), 8.16 (d, 2H, J )
8.6 Hz, 2ArH), 8.28 (dd, 2H, J ) 8.5 Hz, 1.1 Hz, 2ArH), 8.85
(d, 2H, J ) 6.8 Hz, 2ArH), 12.53 (br s, 2H, 2CONH). MALDI-
MS: 755.5 (M + 1). The microanalysis was carried out on its
hexahydrochloride salt, a yellow solid, mp 217-218 °C.
Found: C, 49.52; H, 6.45; N, 13.45. C44H54N10O2‚6HCl‚5H2O
requires: C, 49.68; H, 6.63; N, 13.17.
1
yellow solid (in 34% yield), mp 152-154 °C. H NMR: 2.61-
2.66 (m, 12H, 6CH2), 2.77 (t, 4H, J ) 6.1 Hz, 2CH2), 3.79-
3.82 (m, 16H, 8CH2), 3.93-3.97 (m, 4H, 2CH2), 7.38-7.45 (m,
4H, 4ArH), 7.72 (t, 2H, J ) 7.5 Hz, 2ArH), 8.13 (d, 2H, J )
8.6 Hz, 2ArH), 8.19 (d, 2H, J ) 8.6 Hz, 2ArH), 8.31 (dd, 2H, J
) 8.6, 1.1 Hz, 2ArH), 8.86 (d, 2H, J ) 6.8 Hz, 2ArH), 12.49
(br s, 2H, 2CONH). MALDI-MS: 812.8 (M + 1). The mi-
croanalysis was carried out on its hexahydrochlorde salt, a
yellow solid, mp 220-222 °C. Found: C, 47.77; H, 6.90; N,
12.25. C46H56N10O4‚6HCl‚7H2O requires: C, 47.71; H, 6.62; N,
12.10.
C6N2-DMAE Dim er . This was prepared as for the C8-
DMAE dimer, from N-[2-(N-dimethylamino)ethyl]-9-chloro-
acridine-4-carboxamide and triethylenetetramine. The product
obtained after column chromatography on alumina (ethyl
acetate then methanol/chloroform) was still impure; recrys-
tallization from a variety of solvents was not successful. The
clean free base was finally obtained by rebasification of its
hydrochloride salt, which was purified by recrystallization from
methanol/ethyl acetate, as a yellow solid (in 49% yield), mp
DNA Helix Un w in d in g Assa y. Covalently closed circular
plasmid DNA, pBR 322, isolated from E. coli (dam+, dcm+),
was purchased from Progen Industries Ltd as a 0.5 mg per
mL solution in 10 mM Tris-HCl buffer, pH 7.5. It was found
to be 90% in the supercoiled form by measurement of the
relative intensities of ethidium-stained bands for the closed
circular and nicked circular forms on agarose gels. Ethidium
bromide was purchased from Amresco Ltd. and used without
further purification. The agarose used was DNA grade as
supplied by Progen Industries Ltd. Ultrapure grade Tris base
[tris(hydroxymethyl)aminomethane] was purchased from Am-
resco Ltd and analytical grade glacial acetic was from Scharlau
Ltd. Drug solutions were prepared in 40 mM TAE buffer (40
mM tris(hydroxymethyl)aminomethane, 30 mM glacial acetic
acid, 1 mM EDTA, pH 7.5) at the required concentrations;
hence 5 µL of solution was mixed with an equal volume of a
30 µg per mL solution of pBR 322 DNA (150 ng of DNA, 23.6
µM in base pairs) in TAE buffer pH 7.5. Complexes were
incubated in the dark for 1 h at room temperature prior to
application to the agarose gel. Each titration was repeated at
least three times, and the precision in the value of the
unwinding angles was determined by the width of the equiva-
lence points. To prepare the gel, 1.2 g of agarose was dissolved
in 150 mL of TAE buffer, giving a concentration of 0.8% by
heating to about 55 °C. The solution was cooled to 30 °C and
poured onto a gel plate to set, and the plate was placed in an
electrophoresis apparatus and submerged in TAE buffer to a
depth of 1 mm. Following incubation of the DNA and drug, 3
µL of loading buffer [0.25% (w/v) bromophenol blue and 30%
(w/v) glycerol] was mixed with the complex samples, and the
samples were loaded onto the gel. Electrophoresis was con-
tinued for 1 h and 15 min at 74 V (5 V/cm) at room
temperature. After electrophoresis, the gel was stained for 30
min by placing it in a TAE bath (250 mL) containing 1 µg per
mL of ethidium bromide, and the fluorescence emission
contrast was enhanced by destaining in a TAE bath for an
additional 20 min. The gel was transilluminated by UV light
and fluorescence emission visualized by a CCD camera coupled
to a Bio-Rad Gel Doc 2000 apparatus. Bio-Rad Multianalyst
1.20 software was used to capture the image as TIFF files,
which were later manipulated with PhotoShop Adobe 6.0 for
graphical presentation. On occasions, helix-unwinding experi-
ments were performed in buffer containing 40 mM Bistris (2-
[bis(2-hydroxyethyl)amino]-2-(hydroxymethyl)propane-1,3-diol-
(hydroxymethyl )methane), 30 mM glacial acetic acid, 1 mM
EDTA, pH 6.1.
1
78-79 °C. H NMR: 2.39 (s, 12H, 2N(CH3)2), 2.67 (t-like, 4H,
2CH2), 2.80-2.83 (m, 4H, 2CH2), 2.93-2.96 (m, 4H, 2CH2),
3.75 (t-like, 4H, 2CH2), 3.83-3.86 (m, 4H, 2CH2), 4.78 (br s,
3H, 3NH), 7.28-7.33 (m, 4H, 4ArH), 7.64 (t, 2H, J ) 7.5 Hz,
2ArH), 7.99 (d, 2H, J ) 7.5 Hz, 2ArH), 8.08 (d, 2H, J ) 7.9
Hz, 2ArH), 8.21 (d, 2H, J ) 7.9 Hz, 2ArH), 8.80-8.83 (m, 2H,
2ArH), 12.55 (br s, 2H, 2CONH). MALDI-MS: 729.1 (M + 1).
The microanalysis was carried out on its hexahydrochlorde
salt, a yellow solid, mp 290 °C (dec). Found: C, 48.02; H, 6.86;
N, 13.40. C42H52N10O2‚6HCl‚6H2O requires: C, 47.78; H, 6.68;
N, 13.26.
P r ep a r a tion of 1,4-Bis(2-a m in oeth yl)p ip er a zin e. An-
hydrous sodium carbonate (10.6 g, 0.10 mol) was added, in a
single portion, to a solution of piperazine (2.15 g, 25 mmol)
and chloroacetonitrile (4.53 g, 60 mmol) in absolute ethanol
(100 mL). The whole was then refluxed with stirring for 4 h.
The hot mixture was filtered and the insoluble part washed
with hot ethanol and filtered again. The combined filtrate was
condensed to about 40 mL and then cooled on ice. 1,4-Bis-
(ethanenitrile)piperazine crystallized from the solution as
white needles (3.47 g, 85%), mp 165-166 °C (lit.42 mp 165 °C).
1H NMR: 2.66 (s, 8H, 4CH2), 3.52 (s, 4H, 2CH2). To a mixture
of the above bis(ethane)nitrile compound (1.64 g, 10 mmol) in
N-ethylmorpholine (25 mL), was added, in portions, lithium
aluminum hydride (0.91 g, 24 mmol) over 30 min under a
nitrogen atmosphere. The resulting mixture was refluxed with
stirring for 1 h under nitrogen and then cooled to room
temperature. The insoluble material was filtered off, diethyl
ether (80 mL) was added to the filtrate, and the whole was
stored at 4 °C overnight. The oil which appeared was decanted,
washed repeatedly with diethyl ether, and dried in a vacuum
desiccator. 1,4-Bis(2-aminoethyl)piperazine was then obtained
as a viscous oil and was used in the following steps without
further purification.1H NMR: 2.36-2.46 (m, 12H, 4CH2+2NH2),
2.77 (t, 4H, J ) 6.2 Hz, 2CH2).
C2p ip C2-m or p h olin o Dim er . This was prepared as for
the C8-morpholino dimer, from N-[2-(N-morpholino)ethyl]-9-
chloroacridine-4-carboxamide and 1,4-bis(2-aminoethyl)pip-
erazine. The product obtained after column chromatography
on alumina (ethyl acetate and then methanol/chloroform) was
still not pure and was recrystallized from ethyl acetate/
methanol to afford the pure dimer as a pale yellow solid (31%
yield), mp 200-201 °C. Found: C, 65.97; H, 6.84; N, 16.15.
DNA Dissocia tion Kin etics. Actinomycin D and calf
thymus DNA, sodium salt type I, were purchased from Sigma-
Aldrich, and used without further purification. A stock solution

Bisintercalating DNA-Threading Agents
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 26 5801
of 1 M SHE buffer containing 0.2 M 4-(2-hydroxyethyl)-1-
piperazine-ethanesulfonic acid, 10 mM EDTA, and 1 M sodium
chloride was adjusted to pH 7.0 using NaOH. A 100-fold
dilution of this buffer, 0.01 SHE, was used for dissolving and
dialysing DNA. The experimental buffer used for spectropho-
tometric and kinetic measurements, designated 0.1 SHE, was
prepared by adding sufficient solid NaCl to 0.01 SHE buffer
to bring the NaCl concentration to 0.1 M. The calf thymus
DNA was sonicated to reduce its viscosity, so as to facilitate
efficient mixing with SDS solutions. Kinetic measurements
were performed using a Cary 50 spectrophotometer utilizing
the Cary Win UV Kinetics software (Varian Inc.). A 900 µL
portion of a DNA-ligand complex solution comprising 555 µM
DNA and 55 µM ligand in 0.1 SHE buffer was mixed with 100
µL of 200 mM SDS solution in 0.1 SHE buffer in a 1 mL semi-
microquartz cuvette at 23 °C. Thus the final concentrations
were 500 µM in DNA, 50 µM in ligand, and 20 mM in SDS.
The ligand-to-DNA ratio before dissociation was 0.1, so that
there was one drug molecule bound per turn of the duplex. To
ensure rapid mixing, the solutions were simultaneously in-
jected into the cuvette using Gilson pipets. The DNA-ligand
complex had been previously equilibrated at room temperature
for 1 h. Data were collected at the wavelength where the
difference between the absorbance of the DNA-bound ligand
and the ligand in SDS micelles was greatest, which for the
bis(9-aminoacridine-4-carboxamides) was 423 nm and for
actinomycin D was 430 nm. A total of 2000 data points were
collected for each kinetic run at three different sampling
frequencies, distributed proportionately according to the mag-
nitude of the changing absorbance signal. The final equilib-
rium absorbance was also experimentally determined. Total
sweep times were in the range 420-3600 s. The primary
absorbance data were manipulated to express measured
absorbances as a fraction of the total equilibrium absorbance
change, so as to cast them as a dimensionless fraction of
reaction. Such progress curves were then analyzed for multiple
exponential components by using a nonlinear least squares
facility in the Origin v 7.0 software package (OriginLab). The
curves were deconvoluted into one to three exponentials using
the Levenberg-Marquardt algorithm employing analytical
partial differentials to find the curve of best fit, as determined
by minimizing Gaussian standard errors of the residuals.
Residual plots were generated to enable a visual inspection of
the goodness of fit. Final data are the average of five separate
kinetic runs.
were then washed in phosphate-buffered saline and fixed in
ice-cold ethanol for at least 2 h before analysis, where after
they were stained with propidium iodide as outlined in Current
Protocols in Cytometry 2, and analyzed on a Becton-Dickson
FACSort, using CellQuest software. These measurements
provided cell cycle profiles as determined by assessing total
DNA content and were used to evaluate the proportion of cells
in the various stages of the cycle in the presence of the drugs
after 24 and 48 h of exposure.
Ab b r evia t ion s. DACA, N-[2-(dimethylamino)ethyl]acri-
dine-4-carboxamide; 9-amino-DACA, 9-amino-[N-(2-dimethyl-
amino)ethyl]acridine-4-carboxamide; DMAE, N,N-dimethyl-
aminoethyl; SDS, sodium dodecyl sulfate.
Ack n ow led gm en t. This work was supported by the
Australian National Health and Medical Research
Council (L.P.G.W., B.W.S.).
Su p p or tin g In for m a tion Ava ila ble: Flow cytometric
profiles. This material is available free of charge via the
Internet at http://pubs.acs.org.
Refer en ces
(1) Liu, L. F. DNA topoisomerase poisons as antitumor drugs. Annu.
Rev. Biochem. 1989, 58, 351-75.
(2) Malonne, H.; Atassi, G. DNA topoisomerase targeting drugs:
Mechanisms of action and perspectives. Anticancer Drugs 1997,
8, 811-22.
(3) Kellner, U.; Sehested, M.; J ensen, P. B.; Gieseler, F.; Rudolph,
P. Culprit and victim -- DNA topoisomerase II. Lancet Oncol.
2002, 3, 235-43.
(4) Verweij, J .; den Hartigh, J .; Pinedo, H. M. Antitumor antibiotics.
In Cancer Chemotherapy: Principles and Practice; Chabner, B.
A, Collins, J . M., Eds.; J . B. Lippincott Co.: Philidelphia, 1990;
pp 382-96.
(5) Denny, W. A.; Roos, I. A. G.; Wakelin, L. P. G. Interrelations
between antitumour activity, DNA breakage, and DNA binding
kinetics for 9-aminoacridinecarboxamide antitumor agents. Anti-
Cancer Drug Des. 1986, 1, 141-7.
(6) Bridewell, D. J .; Finlay, G. J .; Baguley, B. C. Topoisomerase I/II
selectivity among derivatives of N-[2-(dimethylamino)ethyl]-
acridine-4-carboxamide (DACA). Anti-Cancer Drug Des. 2001,
16, 317-24.
(7) Caponigro, F.; Dittrich, C.; Sorensen, J . B.; Schellens, J . H.;
Duffaud, F.; Paz Ares, L.; Lacombe, D.; de Balincourt, C.;
Fumoleau, P. Phase II study of XR 5000, an inhibitor of
topoisomerases I and II, in advanced colorectal cancer. Eur. J .
Cancer 2002, 38, 70-4.
(8) Muller, W.; Crothers, D. M. Studies of the binding of actinomycin
and related compounds to DNA. J . Mol. Biol. 1968, 35, 251-90.
(9) Waring, M. J .; Wakelin, L. P. G. Echinomycin: A bifunctional
intercalating antibiotic. Nature 1974, 25, 653-657.
(10) Searle, M. S.; Hall, J . G.; Denny, W. A.; Wakelin, L. P. G. NMR
studies of the interaction of the antibiotic nogalamycin with the
hexadeoxyribonucleotide duplex d(5′-GCATGC)2. Biochemistry
1988, 27, 4340-4349.
Cell Lin e. CCRF-CEM, a human lymphoblastoid (T-cell
leukemia) cell line32 was obtained from Dr. Maria Kavallaris,
Childrens Cancer Research Institute, Sydney, Australia, and
maintained in RPMI medium 1640 (Gibco BRL) supplemented
with 2 mM l-glutamine, 100 units/mL of penicillin/streptomy-
cin, and 10% foetal bovine serum (Trace) at 37 °C, in the
presence of 5% CO2.
(11) Farber, S.; D’Angio, G.; Evans, A.; Mitus, A. Clinical studies of
actinomycin
D with special reference to Wilms’ tumor in
children. J . Urol. 2002, 168, 2560-2.
Gr ow th In h ibition Assa y. Cells were seeded at a density
of 1 × 105 cells/mL and adjusted to a final volume of 1.6 mL
with RPMI medium in 24-well plates. The test agent, dissolved
at 1 mM concentration in deionized water, was diluted in
phosphate-buffered saline to 400 µL at a range of concentra-
tions before addition to the cell suspension. Drug-treated cells
were incubated for at least four rounds of division and counted
on a Coulter Counter after 72 h. Each drug concentration was
measured in duplicate, the whole growth inhibition curve being
repeated at least three times. Control cells were treated in a
similar way, save they received the addition of 400 µL of
medium instead of drug. The concentration of drug required
to inhibit growth by 50% compared to the control, the IC50
value, was obtained using a sigmoidal plotting function
provided in the GraphPad Prism software.
(12) Gradishar, W. J .; Vogelzang, N. J .; Kilton, L. J .; Leibach, S. J .;
Rademaker, A. W.; French, S.; Benson, A. B., III. A phase II
clinical trial of echinomycin in metastatic soft tissue sarcoma.
An Illinois Cancer Center Study. Invest. New Drugs 1995, 13,
171-4.
(13) Kamitori, S.; Takusagawa, F. Crystal structure of the 2:1
complex between d(GAAGCTTC)2 and the anticancer drug
actinomycin D. J . Mol. Biol. 1992, 225, 445-56.
(14) Ughetto, G.; Wang, A. H.; Quigley, G. J .; van der Marel, G. A.;
van Boom, J . H.; Rich, A. A. Comparison of the structure of
echinomycin and triostin A complexed to a DNA fragment.
Nucleic Acids Res. 1985, 13, 2305-23.
(15) Fox, K. R.; Brassett, C.; Waring, M. J . Kinetics of dissociation
of nogalamycin from DNA: Comparison with other anthracycline
antibiotics. Biochim. Biophys. Acta 1985, 840, 383-92.
(16) Fox, K. R.; Wakelin, L. P. G.; Waring, M. J . Kinetics of the
interaction between echinomycin and DNA. Biochemistry 1981,
20, 5768-5779.
F low Cytom etr y. Cells were seeded at a density of 1 ×
105 cells/mL in 25 mL flasks, and incubated for 24 h before
addition of drugs. Drug samples dissolved in phosphate-
buffered saline were then added at 1, 2, 5, and 10 times the
measured IC50 concentrations, and the cells were incubated
for a further 24 and 48 h. Suitable control cells, without the
addition of drugs, were prepared in a similar manner. The cells
(17) Wakelin, L. P. G.; Atwell, G. J .; Rewcastle, G. W.; Denny, W. A.
Relationships between DNA binding kinetics and biological
activity for the 9-aminoacridine-4-carboxamide class of antitu-
mour agent. J . Med. Chem. 1987, 30, 855-861.
(18) Denny, W. A.; Wakelin, L. P. G.. Kinetics of the binding of
mitoxantrone, ametantrone and analogues to DNA: Relationship
to binding mode and antitumour activity. Anti-Cancer Drug Des.
1990, 5, 189-200.

5802 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 26
Wakelin et al.
(19) Wakelin, L. P. G.; Adams, A.; Denny, W. A. Kinetic studies of
the binding of acridinecarboxamide topoisomerase poisons to
DNA: Implications for the mode of binding of ligands with
uncharged chromophores. J . Med. Chem. 2002, 45, 894-901.
(20) Wakelin, L. P. G.; Chetcuti, P.; Denny, W. A. Kinetic and
equilibrium studies of amsacrine-4-carboxamides: A class of
asymmetrical DNA intercalating agent which must bind by
threading through the DNA helix. J . Med. Chem. 1990, 33,
2039-2044.
(21) Adams, A.; Collyer, C.; Guss, M.; Denny, W. A.; Wakelin, L. P.
G. Crystal structure of the topoisomerase II poison 9-amino-[N-
(2-dimethylamino)ethyl]acridine-4-carboxamide bound to the
DNA hexanucleotide d(CGTACG)2. Biochemistry 1999, 38, 9221-
9233.
(22) Adams, A.; Guss, J . M.; Denny, W. A.; Wakelin, L. P. G. Crystal
Structure of 9-amino[N-(2-morpholino)ethyl]acridine-4-carboxa-
mide bound to d(CGTACG)2: Implications for Structure Activity
Relationships of Acridinecarboxamide Topoisomerase Poisons.
Nucleic Acids Res. 2002, 30, 719-725.
(23) Denny, W. A. DNA-intercalating ligands as anti-cancer drugs:
Prospects for future design. Anti-Cancer Drug Des. 1989, 4, 241-
263.
(24) Leng, F,; Priebe, W.; Chaires, J . B. Ultratight DNA binding of a
new bisintercalating anthracycline antibiotic. Biochemistry 1998,
37, 1743-53.
(25) Murr, M. M.; Harting, M. T.; Guelev, V.; Ren, J .; Chaires, J . B.;
Iverson, B. L. An octakis-intercalating molecule. Bioorg. Med.
Chem. 2001, 9, 1141-8.
(26) Miller, C. T.; Weragoda, R.; Izbicka, E.; Iverson, B. L. The
synthesis and screening of 1,4,5,8-naphthalenetetracarboxylic
diimide-peptide conjugates with antibacterial activity. Bioorg.
Med. Chem. 2001, 9, 2015-24.
(27) Wakelin, L. P. G. Polyfunctional DNA intercalating agents. Med.
Res. Rev. 1986, 6, 275-340.
(28) Chow, K.-C.; Ross, W. E. Topoisomerase-specific drug sensitivity
in relation in cell cycle progression. Mol. Cell. Biol. 1987, 7,
3119-23.
(29) Sleiman, R. J .; Catchpoole, D. R.; Stewart, B. W. Drug-induced
death of leukaemic cells after G2/M arrest: Higher order DNA
fragmentation as an indicator of mechanism. Br. J . Cancer 1998,
77, 40-50.
(30) Kohn, K. W.; J ackman, J .; O’Connor, P. M. Cell cycle control
and cancer chemotherapy. J . Cell Biol. 1994, 54, 440-52.
(31) Hannun, Y. A. Apoptosis and the dilemma of cancer chemo-
therapy. Blood 1997, 89, 1845-53.
(32) Sleiman, R. J .; Stewart, B. W. Early caspase activation in
leukemic cells subject to etoposide-induced G2-M arrest: Evi-
dence of commitment to apoptosis rather than mitotic cell death.
Clin. Cancer Res. 2000, 6, 3756-65.
(33) Del Bino, G.; Skierski, J . S.; Darzynkiewicz, Z. Diverse effects
of camptothecin, an inhibitor of topoisomerase I, on the cell cycle
of lymphocytic (L1210, MOLT-4) and myelogenous (HL-60, KG1)
leukemic cells. Cancer Res. 1990, 50, 5746-50.
(34) Rene, B.; Fosse, P.; Khelifa, T.; J acquemin-Sablon, A.; Bailly,
C. Cytotoxicity and interaction of amsacrine derivatives with
topoisomerase II: Role of the 1′ substitute on the aniline nucleus.
Bull. Cancer 1997, 84, 941-8.
(35) Arlin, Z. A. Mitoxantrone and amsacrine: Two important agents
for the treatment of acute myelogenous leukemia (AML) and
acute lymphoblastic leukemia (ALL). Bone Marrow Transplant
1989, 4 Suppl 1, 57-9.
(38) D’Rozario, A. P.; Greig, D. J .; Hudson, R. F.; Williams, A.;
Stepwise Proton Transfer in the Acid-catalysed Hydrolysis of
3,1-Benzoxazin-4-ones: Electrostatic or Hydrogen-bond Stabi-
lisation of the Conjugate Acid. J . Chem. Soc., Perkin Trans. 2
1981, 590-596.
(39) Sumpter, W. C.; J ones, W. F. The Nitration of Isatin. J . Am.
Chem. Soc. 1943, 65, 1802-1803.
(40) Deady, L. W.; Kaye, A. J .; Finlay, G. J .; Baguley, B. C.; Denny,
W. A. Synthesis and antitumor properties of N-[2-(dimethylami-
no)ethyl]carboxamide derivatives of fused tetracyclic quinolines
and quinoxalines: A new class of putative topoisomerase inhibi-
tors. J . Med. Chem. 1997, 40, 2040-6.
(41) Atwell, G. J .; Cain, B. F.; Baguley, B. C.; Finlay, G. J .; Denny,
W. A. Potential antitumor agents. 43. Synthesis and biological
activity of dibasic 9-aminoacridine-4-carboxamides, a new class
of antitumor agent. J . Med. Chem. 1984, 27, 1481-85.
(42) Adelson, D. E.; Polland, C. B. Derivatives of Piperazine. IV.
Reactions with Derivatives of Monochloroacetic Acid. J . Am.
Chem. Soc. 1935, 57, 1280-81.
(43) Chiari, B.; Piovesana, O.; Tarantelli, T.; Zanazzi, P. F. Exchange
interaction in multinuclear transition-metal complexes. 5.
Through-bond exchange coupling in Cu2A(CH3COO)2.2CH3OH
(A2- ) anion of N,N′-bis(2-((o-hydroxybenzhydrylidene)-amino)-
ethyl)piperazine). Inorg. Chem. 1984, 23, 2542-47.
(44) Denny, W. A.; Atwell, G. J .; Baguley, B. C.; Wakelin, L. P. G.
Potential antitumour agents. Part 44. Synthesis and antitumour
activity of new classes of diacridines: The importance of linker
chain rigidity for DNA-binding kinetics and antitumor activity.
J . Med. Chem. 1985, 28, 1568-1574.
(45) Nielsen, P. E.; Zhen, W. P.; Henriksen, U.; Buchardt, O.
Sequence-influenced interactions of oligoacridines with DNA
detected by retarded gel electrophoretic migrations. Biochemistry
1988, 27, 67-73.
(46) Chen, S. F.; Bhenrens, D. L.; Czerniak, P. M.; Dexter, D. L.;
Dusak, B. J .; Fredericks, J . R.; Gale, K. C.; Gross, J . L.;
Krishenbaum, M. R.; McRipley, L. M.; Patten, A. D.; Perella, F.
W.; Seltz, S. P.; Stafford, M. P.; Sun, J . H.; Sun, T.; Wuonola,
M. A.; von Hoff, D. D. XB596, a promising bis-naphthalimide
anti-cancer agent. Anticancer Drugs 1993, 4, 447-457.
(47) Fritzsche, H.; Triebel, H.; Chaires, J . B.; Dattagupta, N.;
Crothers, D. M. Studies on the interaction of anthracycline
antibiotics and deoxyribonucleic acid: Geometry of intercalation
of iremycin and daunomycin. Biochemistry 1982, 21, 3940-3946.
(48) Bauer, W. R. Structure and reactions of closed duplex DNA.
Annu. Rev. Biophys. Bioeng. 1978, 7, 287-313.
(49) Gao, X. L.; Mirau, P.; Patel, D. J . Structure refinement of the
chromomycin dimer-DNA oligomer complex in solution. J . Mol.
Biol. 1992, 223, 259-79.
(50) Blagosklonny, M. V.; Pardee, A. B. Exploiting cancer cell cycling
for selective protection of normal cells. Cancer Res. 2001, 61,
4301-5.
(51) Solary, E.; Droin, N.; Sordet, O.; Rebe, C.; Filomenko, R.;
Wotawa, A.; Plenchette, S.; Ducoroy, P. Cell death pathways as
targets for anticancer drugs. In Anticancer Drug Development;
Baguley, B. C., Kerr, D. J ., Eds.; Academic Press: San Diego,
2002; pp 55-76.
(52) Waldmann, H. Isatincarboxylic acid. J . Prakt. Chem. 1937, 147,
338-43.
(53) Lown, J . W.; Morgan, A. R.; Yen, S. F.; Wang, Y. H.; Wilson, W.
D. Characteristics of the binding of the anticancer agents
mitoxantrone and ametantrone and related structures to deoxy-
ribonucleic acids. Biochemistry 1985, 24, 4028-35.
(36) Rewcastle, G. W.; Denny, W. A. The synthesis of substituted
9-oxoacridan-4-carboxylic acids; Part 2. The use of 2-Iodoiso-
phthalic acid in the J ourdan-Ullmann Reaction. Synthesis 1985,
217-220.
(37) Adams, A.; Guss, J . M.; Collyer, C. A.; Denny, W. A.; Prakash,
A. S.; Wakelin, L. P. G. Acridinecarboxamide topoisomerase
poisons: Structural and kinetic studies of the DNA complexes
of 5-substituted 9-amino-(N-(2-dimethylamino)ethyl)-acridine-
4-carboxamides. Mol. Pharmacol. 2000, 58, 649-658.
(54) Waring, M. J . DNA-binding characteristics of acridinylmethane-
sulphonanilide drugs: Comparison with antitumour properties.
Eur. J . Cancer 1976, 12, 995-1001.
(55) Waring, M. J . Variation of the supercoils in closed circular DNA
by binding of antibiotics and drugs: Evidence for molecular
models involving intercalation. J . Mol. Biol. 1970, 54, 247-279.
J M030253D