The preparation and characterisation of cyclam/anthraquinone macrocyle/intercalator complexes and their interactions with DNA

Article abstract of DOI:10.1039/b302123h

A new series of macrocycle/intercalator adducts have been prepared from cyclam (1,4,8,11-tetraazacyclotetradecane) and the 1- and 2-substituted anthraquinones (1-[(2-aminoethyl)amino]anthracene-9,10-dione (1C2mac), 1-[(3-amino-propyl)amino]anthracene-9,10-dione (1C3mac), 2-[(3-aminopropyl)amino]anthracene-9,10-dione (2C3mac)). The copper complexes of two of these adducts were prepared and the crystal structure of the complex of 1C2mac ([Cu(1C2mac)(CH₃CN)₂](PF₆)₂.0.2H ₂O) was determined as the hexafluorophosphate salt. The equilibrium constants for the binding to DNA of 1C2-mac, with and without the copper added, were found by UV titrations to be 4.7 × 10³ M^-1 and 6.2 × 10³ M^-1, respectively. Reaction with plasmid DNA allowed comparison between the effects of the cyclam/anthraquinone adducts, their copper complexes and the intercalators alone. Addition of the macrocycle increased the extent and effect of DNA intercalation and addition of copper increased the effect still further. The adducts with the longer side chains caused substantially more unwinding of the DNA. None of the adducts inhibited cleavage at dGpG or dGpC sites by restriction enzymes. Molecular modelling was used to investigate the effects of the side chain for a number of possible DNA binding modes. These models reveal that intercalation is not significantly interfered with by the presence of the macrocycle, irrespective of the position of attachment. The increased unwinding caused by adducts with the longer side chain is therefore most likely to be due to specific interactions between the macrocycle and the DNA.

Full text of DOI:10.1039/b302123h

View Article Online / Journal Homepage / Table of Contents for this issue
The preparation and characterisation of cyclam/anthraquinone
macrocyle/intercalator complexes and their interactions with
DNA†
Leanne T. Ellis,a David F. Perkins,b Peter Turnera and Trevor W. Hambley*a
a Centre for Heavy Metals Research, School of Chemistry, University of Sydney, NSW 2006,
Australia. E-mail: t.hambley@chem.usyd.edu.au
b Institute for Biomolecular Sciences, University of Wollongong, Wollongong, NSW 2522,
Australia
Received 24th February 2003, Accepted 21st May 2003
First published as an Advance Article on the web 5th June 2003
A new series of macrocycle/intercalator adducts have been prepared from cyclam (1,4,8,11-tetraazacyclotetradecane)
and the 1- and 2-substituted anthraquinones (1-[(2-aminoethyl)amino]anthracene-9,10-dione (1C2mac), 1-[(3-amino-
propyl)amino]anthracene-9,10-dione (1C3mac), 2-[(3-aminopropyl)amino]anthracene-9,10-dione (2C3mac)). The
copper complexes of two of these adducts were prepared and the crystal structure of the complex of 1C2mac
([Cu(1C2mac)(CH3CN)2](PF6)2.0.2H2O) was determined as the hexafluorophosphate salt. The equilibrium constants
for the binding to DNA of 1C2-mac, with and without the copper added, were found by UV titrations to be 4.7 × 103
M
Ϫ1 and 6.2 × 103 MϪ1, respectively. Reaction with plasmid DNA allowed comparison between the effects of the
cyclam/anthraquinone adducts, their copper complexes and the intercalators alone. Addition of the macrocycle
increased the extent and effect of DNA intercalation and addition of copper increased the effect still further. The
adducts with the longer side chains caused substantially more unwinding of the DNA. None of the adducts
inhibited cleavage at dGpG or dGpC sites by restriction enzymes. Molecular modelling was used to investigate
the effects of the side chain for a number of possible DNA binding modes. These models reveal that intercalation
is not significantly interfered with by the presence of the macrocycle, irrespective of the position of attachment.
The increased unwinding caused by adducts with the longer side chain is therefore most likely to be due to specific
interactions between the macrocycle and the DNA.
Introduction
The localisation of macrocyclic metal complexes on DNA is
of interest for a number of reasons. If the metal ion is a radio-
nuclide, increased DNA damage will result,1 if it is a metal that
facilitates phosphate hydrolysis, then DNA cleavage may result2
and if the metal binds to DNA bases, then sequence selective
binding can occur.3 Intercalators bind rapidly and reversibly
Scheme 1
to DNA and have been shown to increase the rate and extent
Experimental
of DNA binding of a variety of groups including metal
complexes.3–6 Kimura and colleagues have described a series
of macrocyclic complexes linked to intercalators and have
shown that these interact in a sequence selective manner with
DNA.3,7–12 They have also shown that the sequence selectivity
depends on the metal ion, the nature of the intercalator and
the nature and position of the link between the macrocycle
and the intercalator.3,7–12 In all of these studies the macrocycle
employed was the 12-membered ring, cyclen (1,4,7,10-tetraaza-
cyclododecane), a ligand that forces the metal to lie out of the
coordination plane promoting interactions with the DNA.
The 14-membered tetraazamacrocycle cyclam (1,4,8,11-tetra-
azacyclotetradecane) can tightly bind a wide range of metal
ions including radionuclides, phosphate-hydrolysis-promoting
metal ions and metals that can bind to DNA. Therefore, we
were interested in attaching cyclam to an intercalator and
establishing how this influenced interactions with DNA. Here
we describe the preparation of a series of cyclam/anthra-
quinone adducts (Scheme 1), the copper() complexes of two
of these and experimental and molecular mechanics modelling
studies of the interactions between the adducts and DNA.
Instrumentation
Mass spectra were obtained in the appropriate solvents (flow
rate 100 µL per minute) on a Finnigan LCQ MS Detector
and presented as an electrospray ionisation low resolution mass
spectrum using Finnigan LCQ data processing software (ver-
sion 12.0). An ESI spray voltage of 4 kV was applied and a
nitrogen sheath gas pressure of 60 psi with a heated capillary
temperature of 200 ЊC were used. Diffuse reflectance infrared
transform spectra (DRIFTS) were recorded in a potassium
bromide background on a Bio-Rad FTS-40 spectrometer using
Win-IR software. 1H NMR spectra were collected on a Bruker
AC200F 200 MHz spectrometer and referenced to TMS or
solvent used. 195Pt NMR spectroscopy was carried out by
Jacques Nemorin of The University of Sydney on a Bruker
Avance 600 MHz spectrometer, referenced externally to
K2PtCl4 in D2O at Ϫ1624 ppm. Elemental analyses were carried
out by Reet Bergman at the Microanalytical Unit of the
Research School of Chemistry at the Australian National
University, Canberra.
1-[(Hydroxyethyl)amino]anthracene-9,10-dione, 1C2OH
† Electronic supplementary information (ESI) available: Figs. S1–8; gel
electrophoresis patterns, 1H NMR spectra, and overlap interactions
between the macrocycle and DNA. See http://www.rsc.org/suppdata/dt/
b3/b302123h/
1-Chloroanthracene-9,10-dione (15 g, 63 mmol) was heated in
butanol (150 mL). When the solution became clear yellow (at
∼70 ЊC), ethan-1-ol-2-amine (15 mL, 0.255 mol) was added
D a l t o n T r a n s . , 2 0 0 3 , 2 7 2 8 – 2 7 3 6
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3
2728

View Article Online
quickly and the solution refluxed for 3 h. The solution went
from yellow to deep red, was cooled to room temperature and
concentrated to about 60 mL. The reaction mixture was
allowed to cool to room temperature and then cooled to Ϫ20 ЊC
overnight. The solid product was removed by filtration,
recrystallised from ethyl acetate and then column chromato-
graphed (silica gel 60, particle size 0.040–0.063 nm, 3 × 25 cm).
Impurities were eluted with petroleum ether–ethyl acetate
(1 : 2), then chloroform–ethanol (9 : 1) was added to elute the
pure red product. The solvent was removed by rotary evap-
oration and the red solid placed in an oven to dry (100 ЊC) (yield
75%). The NMR spectrum (CDCl3) was in accord with the
proposed structure.
(1 equivalent), Cs2CO3 (1 equivalent) and a catalytic amount
of KI were added and the reaction mixture was refluxed for 2 h.
The reactions were monitored by TLC (silica gel, CHCl3). The
reaction mixture was cooled to room temperature and the
excess cyclam that precipitated was filtered off. The filtrate was
washed with aqueous sodium hydroxide (5%, 3 × 5 mL), water
(2 × 5 mL) and dried with magnesium sulfate. The solution was
rotary evaporated to dryness and the yellow solid collected.
The solid was dissolved in chloroform and column chromato-
graphed on silica gel 60 (particle size 0.040–0.063 nm, 4 × 20
cm, CHCl3–EtOH–NH4OH, 70 : 25 : 5). The last band eluted
was the desired product.
2Cmac: cyclam (0.5 g, 2.5 × 10Ϫ4 mol), 2-bromomethyl-
anthracene-9,10-dione (0.015 g, 5 × 10Ϫ5 mol, Sigma-Aldrich),
yield 72%. MS (EI): 421, [M ϩ Hϩ]. The NMR spectrum is
in accord with the proposed structure, (ESI,† Fig. S3).
C25H32N4O2; found: C, 71.0; H, 7.4; N, 12.9%; calc.: C, 71.4; H,
7.7; N 13.3%.
1-[2-(Bromoethyl)amino]anthracene-9,10-dione, 1C2Br
1C2OH (0.85 g, 3.2 mmol) was dissolved in dry dichloro-
methane (45 mL) and 1 equivalent of dry pyridine (111 µL) was
added. The solution was heated to 40 ЊC, 1 equivalent of thionyl
bromide (257 µL) was added and the solution was refluxed for
5 h. Progress of the reaction was monitored by TLC (silica
gel, CHCl3) and an extra 50 µL of thionyl bromide was added
twice to drive it to completion. The reaction mixture was then
cooled to room temperature and water was carefully added to
decompose and remove excess thionyl bromide. The organic
phase was washed with water twice and dried with magnesium
sulfate, filtered and rotary evaporated to dryness. The product
was dissolved in minimal chloroform and loaded onto a silica
gel column (4 × 40 cm) and eluted with chloroform. The first
major band was collected, evaporated to dryness and dried in
an oven at 110 ЊC for 2 h (yield 78%). The NMR spectrum
(CHCl3) was in accord with the proposed structure.
1C2mac: cyclam (0.254 g, 1.3 × 10Ϫ3 mol), 1C2Br (0.084 g,
2.54 × 10Ϫ5 mol), yield 87%. MS (EI): 450.3, [M ϩ Hϩ]. The
NMR spectrum is in accord with the proposed structure (ESI,†
Fig. S4). C26H37N5O22ϩ, 2HCO3Ϫ.2H2O; MW 609.6, found: C,
53.4; H, 6.8; N, 11.3%; calc.: C, 55.2; H, 7.1; N 11.4%.
1C3mac: cyclam (0.254 g, 1.3 × 10Ϫ3 mol), 1C3Br (0.088 g,
2.54 × 10Ϫ5 mol), yield 37%, MS (EI): 464.4, [M ϩ Hϩ]. The
NMR spectrum is in accord with the proposed structure.
C27H39N5O22ϩ, 2HCO3Ϫ.2H2O; MW 623.7, found: C, 57.8; H,
6.9; N, 11.2%; calc.: C, 55.8; H, 7.3; N, 11.2%.
2C3mac: cyclam (0.058 g, 3.0 × 10Ϫ4 mol), 2C3Br (0.020 g,
5.8 × 10Ϫ5 mol), yield 63%, MS (EI): 464.4 [M ϩ Hϩ].
Syntheses of copper complexes
2-[(Hydroxypropyl)amino]anthracene-9,10-dione, 2C3OH
General procedure. The appropriate cyclam/anthraquinone
(16.0 mmol) was dissolved in a minimum of dimethylform-
amide and water (∼10 mL) was added. Copper() acetate was
dissolved in the minimum amount of water and added to the
solution. The pH was checked and if necessary was adjusted to
be 5.5 with either hydrochloric acid (1 M) or sodium hydroxide
(1 M). The reaction was allowed to proceed at room temper-
ature and monitored by TLC (silica gel, CHCl3–EtOH–NH3,
70 : 25 : 5). Further aliquots of copper acetate were added every
hour until the spot due to the starting material disappeared.
Dichloromethane was added to the reaction mixture followed
by a saturated solution of sodium hexafluorophosphate result-
ing in movement of the red product into the dichloromethane
layer. The organic layer was washed twice with water (with
a few drops of saturated sodium hexafluorophosphate added
to the water layer if necessary to keep the red product in the
dichloromethane layer). The solvent was removed by rotary
evaporation and TLC (silica gel, CHCl3–EtOH–NH3 70 : 25 : 5)
showed only a clean spot on the baseline. The solid was
dissolved in acetonitrile and isopropyl ether was diffused into
a refrigerated solution to purify the product and to obtain
X-ray quality crystals.
2-Chloroanthracene-9,10-dione (5 g, 21 mmol) and propan-1-
ol-3-amine (10 mL, excess) were refluxed in butanol (50 mL)
under nitrogen overnight. The reaction was monitored by TLC
(silica gel, chloroform). The reaction mixture was cooled to
room temperature and then cooled to Ϫ20 ЊC overnight, filtered
and the solid obtained was recrystallised from ethyl acetate. The
product was column chromatographed (silica gel 60, particle
size 0.040–0.063 nm, 3 × 35 cm) with chloroform as the eluent
until the desired product started eluting. The product was then
eluted with 50 : 50 chloroform–ethanol. The solvent was
removed by rotary evaporation and the product dried in an
oven at 110 ЊC (yield 35%). The NMR spectrum (d6-DMF) was
in accord with the proposed structure.
2-[3-(Bromopropyl)amino]anthracene-9,10-dione, 2C3Br
2C3OH (0.1 g, 35.6 mmol) was dissolved in dry dichloro-
methane (10 mL) and 1 equivalent of dry pyridine (29 µL) was
added. The solution was heated to 40 ЊC, 1 equivalent of thionyl
bromide (28 µL) was added and the solution was then refluxed
for 5 h. Progress of the reaction was monitored by TLC (silica
gel, chloroform) and an extra 10 µL of thionyl bromide was
added twice to drive it to completion. The reaction mixture was
then cooled to room temperature and water was carefully added
to decompose and remove excess thionyl bromide. The organic
phase was washed with water twice and dried with magnesium
sulfate, filtered and rotary evaporated to dryness. The product
was dissolved in minimal chloroform and loaded onto a silica
gel column (1.5 × 20 cm) and eluted with chloroform. The first
major band was collected, evaporated to dryness and dried in
an oven at 110 ЊC for 1 h (yield 41%). The NMR spectrum
(CHCl3) was in accord with the proposed structure.
Cu–2Cmac: yield 27%. MS (EI): 482.1 [M ϩHϩ], 241.7
[2ϩ charge]. [Cu(2Cmac)](PF6)2, CuC25H32N4O2(PF6)2; MW
774.1, found: C, 38.3; H, 4.6; N, 7.3%; calc.: C, 38.8; H, 4.2; N,
7.2%.
Cu–1C2mac: yield 39%. MS (EI): 511.2, [M ϩHϩ], 256
[2ϩ charge]. [Cu(1C2mac)(CH3CN)2](PF6)2ؒ0.2H2O, CuC30H41-
N7O2(PF6)2ؒ2CH3CN; MW 885.2, found: C, 40.54; H, 4.78; N,
10.99%; calc.: C, 40.70; H, 4.67; N, 11.07%.
X-Ray crystallography
Crystals obtained by the vapour diffusion of isopropyl ether
into an acetonitrile solution of the Cu–1C2mac complex
initially formed too quickly, so toluene was added to the
isopropyl ether and the solution refrigerated to slow the crystal
formation and produce higher quality crystals. A crystal was
Macrocycle/intercalator syntheses
General
procedure.
1,4,8,11-Tetraazacyclotetradecane
(cyclam, 5 equivalents, Sigma-Aldrich) was dissolved in hot
toluene (10 mL), the appropriate brominated anthraquinone
D a l t o n T r a n s . , 2 0 0 3 , 2 7 2 8 – 2 7 3 6
2729

View Article Online
attached to a thin glass fibre and mounted on a Bruker SMART
1000 CCD diffractometer employing graphite-monochromated
Mo-Kα radiation generated from a sealed tube. Cell constants
at 150 ЊC were obtained from least-squares refinement against
5864 reflections located in a 2θ range of 2.39 to 28.00Њ. Data
were collected to 28.00Њ and the full sphere coverage was >99%.
An empirical absorption correction, determined with SAD-
ABS,13 was applied to the data and the data integration and
reduction were undertaken with SAINT and XPREP.14 The
structure was solved by direct methods with SHELXS-86,15 and
refined using teXsan.16 Hydrogen atoms were included at
calculated sites and non-hydrogen atoms were refined aniso-
tropically. The two contributors to the disordered anthra-
quinone moiety were each refined with a 50% occupancy.
Neutral atom scattering factors were taken from International
Tables.17 Anomalous dispersion effects were included in Fc;18
the values for ∆f Ј and ∆f Љ were those of Creagh and McAuley.19
The values for the mass attenuation coefficients are those of
Creagh and Hubbell.20 All other calculations were performed
using the teXsan crystallographic software package of Molec-
ular Structure Corporation.16 ORTEP21 plots (30% thermal
ellipsoids) of the complex are shown in Fig. 1(a) and (b).
H2O) were added to each tube. The volume in each tube was
adjusted to 20 µL with H2O and DMF to give a final solution
with 20% DMF (v/v). The tubes were vortexed and pulse
centrifuged to facilitate mixing and the reaction was allowed to
proceed by incubation at 37 ЊC for 3 h. The samples were frozen
to prevent any further reaction from taking place. A concen-
tration range of 0–100 µM was used for the initial study of each
compound and the results of these studies were then used to
select the appropriate range.
Where relevant, a restriction enzyme and buffer were added.
For BamH1, X10 restriction buffer B (3 µL, 10 mM Tris-HCl, 5
mM MgCl2, 100 mM NaCl, 1 mM 2-mercaptoethanol, pH 8.0,
Roche Molecular Biochemicals) and BamH1 (0.5 µL, activity
11 U/µL, excess, Roche Molecular Biochemicals) were added.
For PvuII, restriction buffer M (3 µL, 10 mM Tris-HCl, 10 mM
MgCl2, 50 mM NaCl, 1 mM dithioerythritol, pH 7.5, Roche
Molecular Biochemicals) and PvuII (0.5 µL, activity 10 U/µL,
excess, Roche Molecular Biochemicals) were added. Water was
added to give a final volume in each tube of 30 µL. The tubes
were incubated at 37 ЊC for 1 h to allow any DNA digestion to
take place, then rapidly cooled to 0 ЊC to prevent any further
reaction from taking place.
Tracking dye (6 µL, 0.25% Bromophenol Blue, 40% sucrose
in water) was added to each tube and 20 µL portions were
loaded on an agarose gel and electrophoretically chromato-
graphed. The agarose gels (Molecular biology certified agarose,
ultra pure grade, BIO-RAD, made up to 1.5% w/v) were
prepared using TBE buffer (90 mM Tris, 90 mM boric acid,
2 mM EDTA, adjusted to pH 8.0 with 10 M HCl) containing
ethidium bromide (0.5 g mLϪ1, Sigma). The agarose gel was
immersed in TBE buffer solution with ethidium bromide (0.5 µg
mLϪ1). DNA size marker (1 µL in 5 µL tracking dye, λ DNA/
HindIII digestion, Roche Molecular Biochemicals) was used as
a standard and loaded onto the first lane of each gel. Electro-
phoresis was carried out at approximately 25 V for 18 h (Gene
power supply GPS 200/400). For the post-staining experiments,
the agarose gels were prepared in the same manner with the
exclusion of the ethidium bromide in the agarose gel and TBE
buffer. After the electrophoresis was complete, the gels were
stained in TBE buffer (90 mM Tris, 90 mM boric acid, 2 mM
EDTA, adjusted to pH 8.0 with 10 M HCl) containing ethidium
bromide (0.5 mg mLϪ1) for 2 h and then rinsed with water. Each
gel was photographed under UV light through a clear and an
orange filter (Wratten 2A and Wratten 22 Kodak filter) using
Polaroid type 55 film (Foto Riesel). The photos were scanned
with a Hewlett Packard ScanJet 6100C.
Crystal data for Cu–1C2mac. Formula CuC30H42N7O2.2P2F12;
Mr = 888.12, triclinic, space group P1, a = 8.304(2), b = 9.899(3),
c = 22.647(6) Å, α = 95.644(5), β = 95.614(4), γ = 105.800(4)Њ, V
= 1767.5(7) Å3, Dc = 1.669 g cm Ϫ3, Z = 2, µ(Mo-Kα) = 0.814
mmϪ1, λ(Mo-Kα) = 0.71063 Å, T = 150 K, F(000) = 912
¯
electrons, 2θmax = 28.32, N = 8116, No = 5864 (I ≥ 2.5σ(I )), Nvar
=
660, max. shift = 0.08σ, R [Σ(||Fo| Ϫ |Fc||)/Σ|Fo|] = 0.050, Rw
[Σ(w1/2||Fo| Ϫ |Fc||)/Σw1/2|Fo|] = 0.037 and w = 1/σ2, ρmax,min = 0.50,
Ϫ0.54 e ÅϪ3
.
CCDC reference number 204635.
See http://www.rsc.org/suppdata/dt/b3/b302123h/ for crystal-
lographic data in CIF or other electronic format.
UV Spectrophotometric titrations
UV spectroscopy was carried out on a HP 8452A Diode Array
Spectrometer with HP 89531A MS-DOS-UV/VIS operating
software. Stirring and temperature were controlled with a
8909A Peltier temperature control accessory. The spectro-
photometric titration was performed using similar conditions
to those used by Kikuta et al.12 Salmon sperm DNA (Sigma)
was dissolved in autoclaved, milli-Q H2O and filtered (MFS 13,
cellulose acetate membrane, pore 0.02 µm, Millipore). The
DNA concentration was determined spectrophotometrically,
ε260 = 6700 MϪ1 cmϪ1 (per nucleotide).22 Stock DNA was made
up to a final concentration of 1 mM (nucleobases) in HEPES
buffer (Sigma-Aldrich) (10 mM, pH 8.0, with I = 0.1 (NaNO3)).
The compounds (1C2mac and Cu–1C2mac, 20 µM) were made
up in HEPES (10 mM, pH 8.0, with I = 0.1 (NaNO3)). Stock
DNA was added in increasing amounts (0–200 µM) to the
cyclam/anthraquinone adduct or Cu2ϩ complex (2 mL) in
a 1 cm quartz cuvette with continuous stirring at 25 ЊC.
The absorption at λmax was measured after each addition. The
calculations for the binding constant were as described by
Arounaguiri and Maiya23 and Kikuta et al.12
Molecular modelling
The AMBER force field parameters were used for the nucleic
acid interactions.25 The models were constructed using the
HyperChem programme.26 The double stranded B-DNA
(5Ј-CCCCGGGG-3Ј) was generated using the nucleic acid
feature in the database of HyperChem and the intercalation
site opened using distance constraints. The models were geo-
metrically minimised using Fletcher-Reeves conjugate-gradient
minimisation. Convergence was defined when the gradient of
the average root mean square (RMS) shift reached 0.4 kJ ÅϪ1
molϪ1. Models that contained metals were set up in Hyper-
Chem without the metal present to produce a starting model,
then transferred into the MOMEC-SG programme,27 and
minimised with the metal present. The calculations performed
in MOMEC-SG used the Newton–Raphson method for energy
minimisation.
Plasmid binding
Plasmid binding experiments with DMF soluble compounds
were adapted from methods described by Broomhead et al.24
Fresh stocks of the water insoluble compounds (1 mg) were
prepared by dissolving them in high purity DMF (2 mL) with
sonication and slight warming if necessary. The solution was
diluted with water (3 mL) and millipore filtered. Appropriate
dilutions were made and the required volumes of solutions were
added to the Eppendorf tubes to achieve a set of concen-
trations in the range 0–100 µM. Sterilised buffer (4 µL, 5 mM
phosphate, pH 7.4), and PBR322 DNA (5 µL, 0.083 µg/µL
Results and discussion
Macrocycle/intercalator syntheses
The procedure for the combination of the cyclam and the
anthraquinone was adapted from a method for the addition of
D a l t o n T r a n s . , 2 0 0 3 , 2 7 2 8 – 2 7 3 6
2730

View Article Online
naphthalene to cyclam reported by Engeser et al.28 and involved
the reaction between cyclam and an anthraquinone with a
brominated side chain. Addition of Cs2CO3 and KI to catalyse
the reaction was found to be necessary. Initially, the method
was trialled using cyclam and 2-bromoanthraquinone and
when this was shown to be successful, it was extended to alkyl-
aminoanthraquinones with brominated side chains prepared
using methods reported previously. The procedure is outlined
in Scheme 2. Thus, 1C2OH and 2C3OH were prepared in a
manner analogous to that described by Katzhendler et al.29 for
the preparation of 1C3OH. The preparation of 1C3Br has
been reported by Gibson et al.30 and the preparation of 1C2Br
and 2C3Br by the same method was successful. These were used
to prepare 1C2mac, 1C3mac and 2C3mac using the procedure
described by Engeser et al.28 Our attempts to complete the
series by preparing 2C2mac were frustrated by the failure of
all attempts to prepare the intermediate 2C2Br from 2C2OH
despite numerous variations in the reaction conditions. There
have been no other reports of 2C2Br in the literature and it may
be that reaction of 2C2OH gives rise only to other products.
Fig. 1 ORTEP plots for Cu–1C2mac, 30% probability ellipsoids.
Scheme 2
the Cu–N(secondary amine) bond lengths is 2.014 Å, similar to
the 2.021 Å in [Cu(cyclam)(MeCN)2]2ϩ 31
but the bond to the
Syntheses of copper complexes
,
tertiary amine carrying the anthraquinone moiety is substanti-
ally longer at 2.061(3) Å. There is an intramolecular hydrogen
bond between an oxygen atom of the anthraquinone and the
hydrogen of the amine group in the side chain in both con-
formations of the Cu–1C2mac which is similar to that in
the structure of a related anthraquinone moiety, 1-[2-(diethyl-
amino)ethylamino)]anthraquinone.32 There is a short contact
(2.11 Å) between the H(C8) and H(N5) atoms of the anthra-
quinone moiety and the macrocycle moiety in one of the
anthraquinone orientations (Fig. 1(a)), but no other significant
interactions between the macrocyclic and intercalator groups
are observed. The next shortest distances are 3.68 Å between
O(2) and C(9) (Fig. 1(a)) and 3.55 Å between O(1Ј) and C(28)
(Fig. 1(b)). Table 1 provides a comparisons of bond distances,
angles and hydrogen bonds of Cu–1C2mac with other similar
structures.
The procedure for the synthesis of the copper complexes of the
cyclam/anthraquinones was adapted from that reported by
Jones-Wilson et al.25 and was tested using the C2mac complex.
The cyclam/anthraquinone adducts were oily red products that
were difficult to remove from the reaction flask. Therefore,
chloroform was added to the entire sample and a portion was
pipetted into another flask. The solvent was then removed and
the amount transferred was weighed and used in the reaction.
Formation of the copper complex was achieved by adding
aliquots of copper() acetate until no trace of the free ligand
was observable in a TLC. Purification of the complexes was
readily achieved by diffusion of ether into an acetonitrile
solution of the crude product.
Crystal structure
The crystal structure of Cu–1C2mac confirms the synthesis of
the desired product and ORTEP21 plots (Fig. 1), show this
product with two acetonitrile groups coordinated in the axial
sites of the copper. There are two conformations of the adduct
in the crystal lattice, with the anthraquinone in different
orientations and with 50% occupancies (Fig. 1(a) and (b)). In
both orientations, the rings of adjacent molecules in the lattice
are stacked parallel to one another. The two orientations are
approximately perpendicular to one another and atoms C(14),
C(15), C(21) and C(22) of adjacent molecules overlap. This
gives rise to anomalies in the bond lengths making detailed
discussion of the geometry of the anthraquinone inappropriate.
The copper–macrocycle moiety did not undergo this disorder.
The interatomic distances and angles are similar to those from
related structures of each of the two moieties. The average of
UV spectrophotometric titrations
UV spectrophotometric titrations were carried out to compare
the affinities of 1C2mac and its copper complex for salmon
sperm DNA by determining their apparent binding constants
(Kap (MϪ1)). The observed decrease in the intensity of the
absorption for each of the compounds on the addition of DNA
was used to determine the binding constants using similar
conditions to those described by Kikuta et al.12
The absorbance was measured at a maximum of 314 nm for
1C2mac (Fig. 2(a)) and 316 nm for Cu–1C2mac (Fig. 3(a)) over
a range of salmon sperm DNA concentrations. There was
a steady decrease in the intensity of the absorbance with the
addition of each increment of salmon sperm DNA (0–200
D a l t o n T r a n s . , 2 0 0 3 , 2 7 2 8 – 2 7 3 6
2731

View Article Online
Table 1 Comparisons of bond distances, angles and hydrogen bonds of Cu–1C2mac with other similar structures
Cu–1C2mac
[Cu(cyclam)(MeCN)2]2ϩ 31
1-[2-(Diethylamino)ethylamino)]anthraquinone32
Cu–N(1)
Cu–N(2)
Cu–N(3)
Cu–N(4)
Cu–N(MeCN)
Cu–N(MeCN)
2.008(3)
2.025(3)
2.009(3)
2.061(3)
2.485(3)
2.717(3)
2.020(4)
2.023(4)
2.575(5)
N(1)–Cu–N(2)
N(2)–Cu–N(3)
N(3)–Cu–N(4)
N(4)–Cu–N(1)
N(1)–Cu–N(6)
N(2)–Cu–N(6)
N(3)–Cu–N(6)
N(4)–Cu–N(6)
N(1)–Cu–N(7)
N(2)–Cu–N(7)
N(3)–Cu–N(7)
N(4)–Cu–N(7)
H(N5) ؒ ؒ ؒ O(2)
N(5)–O(2)
93.4(1)
86.1(1)
93.8(1)
86.7(1)
91.26(9)
82.89(9)
89.38(9)
95.0(1)
87.9(1)
90.0(1)
91.4(1)
92.1(1)
1.823
94.0(2)
86.0(2)
91.6(2)
87.3(2)
88.4(2)
92.7(2)
1.862(3)
2.638(3)
2.614
1.875
2.602
H(N5Ј) ؒ ؒ ؒ O(1Ј)
N(5Ј)–O(1Ј)
Distances are in angstroms (Å) and angles are in degrees (Њ). Estimated standard deviations in the least significant figure are given in parentheses.
Fig. 2 (a) UV spectrophotometric titration curves and (b) half-
reciprocal plot for 1C2mac.
Fig. 3 (a) UV spectrometric titration curves and (b) half-reciprocal
mM). The apparent binding constants (Kap) were determined
plot for Cu–1C2mac.
from the half-reciprocal plots (Figs. 2(b) and 3(b)) and were
found to be 4.7 × 103 MϪ1 for 1C2mac and 6.2 × 103 MϪ1 for
reported binding constants for an anthraquinone attached to
the 12-membered macrocycle via a one-carbon chain (9,10-
anthraquinon-2-yl)methyl-1,4,7,10-tetraazacyclododecane, Kap
Cu–1C2mac. These values, and the change on addition of Cu2ϩ
,
are comparable with those found by Kikuta et al.12 who
D a l t o n T r a n s . , 2 0 0 3 , 2 7 2 8 – 2 7 3 6
2732

View Article Online
Table 2 Concentrations and Rt values (complexes per base pair) at
which supercoiled DNA was unwound to an open-circular
conformation
that the unwound supercoiled DNA recoiled in the opposite
direction after this point leading to a slight increase in
its mobility while the mobility of the nicked open circular
DNA remained the same. The latter interpretation is more
consistent with the behaviour of the DNA when exposed to
1C3mac and 2C3mac and with the effect of 1C2mac on linear
DNA.
Compound
Concentration/µM
Rt
1C2mac
Cu–1C2mac
1C3mac
21
2
4
0.69
0.07
0.13
The results for 1C3mac (Fig. 4(b)) are remarkably different in
that the supercoiled DNA becomes open circular when the
concentration of 1C3mac reaches approximately 4 µM, four-
fold lower than the concentration of 1C2mac needed to achieve
maximum unwinding. After this point, the mobility of the
supercoiled DNA increased until it tapered off at a concen-
tration of 21 µM. The open circular nicked DNA decreased in
mobility slightly with increasing 1C3mac concentration until
21 µM was reached, after which it tapered off. The gel electro-
phoresis pattern for 2C3mac treated DNA (ESI,† Fig. S2) shows
similar results to those seen for 1C3mac in that the supercoiled
DNA became open circular when the concentration of 2C3mac
reached approximately 4 µM. After this point, the mobility of
the supercoiled DNA increased until it tapered off at a 2C3mac
concentration of 15 µM. The open circular nicked DNA
decreased in mobility slightly with increasing 2C3mac concen-
tration until 15 µM was reached.
Comparing the plasmid binding assays for an anthraquinone
with an unmodified amine side chain (2C2, ESI,† Fig. S5) with
those for the cyclam/anthraquinone adducts (Table 2) reveals
that addition of the macrocycle has a strong influence on the
binding of the intercalator. The anthraquinones by themselves
produce very little change in the mobility or unwinding of
the supercoiled DNA probably because of the high lability of
the intercalation of the simple anthraquinones preventing
any permanent unwinding of the DNA. In contrast, 1C2mac
unwinds the supercoiled DNA and it becomes open circular at
an Rt of 0.69 indicating that the macrocycle is enhancing the
binding of the anthraquinone. This enhanced binding could be
due to hydrogen bonding associations between the macrocycle
and the DNA and/or to the macrocycle being dipositively
charged at neutral pH. Also possible is the presence of a cation
within the cyclam, (e.g. Csϩ, Kϩ and Mg2ϩ) which would also
result in a charged species capable of binding electrostatically.
Daunomycin, which is an excellent intercalator, fully unwinds
the supercoiled DNA at lower concentrations (Rt = 0.17–0.34)
than does 1C2mac (Rt = 0.69) which probably reflects differ-
ences in the strength of binding to the DNA. The 1C3mac and
2C3mac compounds (ESI,† Figs. S1 and S2) fully unwound the
supercoiled DNA to become open circular at an Rt of 0.13
indicating that these compounds are more efficient at unwind-
ing the DNA than is 1C2mac. This is probably due to the longer
2C3mac
Daunomycin
4
0.13
0.17–0.34
5–10
= 5.9 × 103 MϪ1) and its Cu2ϩ complex (Kap = 8.9 × 103 MϪ1).
The binding constant was three times greater for the Zn2ϩ con-
taining macrocycle (Kap = 2.8 × 104 MϪ1), suggesting that the
metal plays an important role in determining the affinity of the
complex for DNA.
Plasmid binding assays
The DNA binding of the cyclam/anthraquinone adducts and
their metal complexes was compared by reacting them with
plasmid DNA and separating the products using gel electro-
phoresis. The results are summarised in Table 2. Binding at the
GpG and GpC sites was probed using the restriction enzymes
BamH1 and PvuII which cleave at the 5Ј-GGATCC-3Ј and
5Ј-CAGCTG-3Ј sequences, respectively.
Fig. 4(a) shows the gel for PBR322 DNA treated with
1C2mac. As the concentration of 1C2mac was increased, the
supercoiled DNA became increasingly unwound as shown by a
decrease in its mobility. The supercoiled DNA became open
circular when the concentration of 1C2mac reached approx-
imately 21 µM, representing an Rt value of 0.69. After this
point, it is difficult to determine with certainty which band of
DNA corresponds to the nicked open circular form and which
band corresponds to the unwound supercoiled form due to
their proximity on the gel. One interpretation of the gel in this
region is that the mobility of the unwound supercoiled DNA
decreased further so that it is slightly less than that of the
nicked open circular DNA while the mobility of the open circu-
lar nicked DNA increased slightly. Another interpretation is
chain length connecting the two moieties providing
a
macrocycle moiety better placed for interaction with the DNA
phosphate backbone.
The addition of either BamH1 (which cuts at a site con-
taining GpG) or PvuII (which cuts at a site containing GpC)
restriction enzymes to supercoiled DNA results in restriction to
linear DNA at all concentrations of 1C2mac, 1C3mac or
2C3mac (Fig. 5). This indicates that the cyclam/anthraquinone
adducts do not inhibit the restriction of the DNA at either of
these two sites. This may be due to the complex being present
at the site, but not interfering with the restriction enzyme.
However, given the strong stabilising interaction between the
macrocycle and the DNA, it is more likely to be due to the
complex binding at other sites. Daunomycin, similarly, causes
no inhibition of the restriction by BamH1 over the con-
centration range examined. However, at high concentrations,
daunomycin inhibits some restriction by PvuII (ESI,† Fig. S6)
whereas there is no inhibition of restriction by PvuII at high
concentrations of any of the cyclam/anthraquinone adducts.
This could be due to a greater binding strength of daunomycin
or the targeting of the PvuII site resulting in the inhibition of
Fig. 4 Gel electrophoresis patterns of PBR322 DNA treated with (a)
1C2mac and (b) 1C3mac. The lanes (from left to right) correspond to
the marker, control-DNA only, then concentrations of 1, 4, 7, 11, 15,
18, 21, 25, 29, 32 and 36 µM of intercalator incubated with DNA,
followed by the marker.
D a l t o n T r a n s . , 2 0 0 3 , 2 7 2 8 – 2 7 3 6
2733

View Article Online
Fig. 5 Gel electrophoresis pattern of PBR322 DNA treated with
1C2mac and BamH1 restriction enzyme. The lanes (from left to right)
correspond to the marker, control-DNA only, control-DNA and
BamH1 only, then concentrations of 1, 4, 7, 11, 15, 18, 21, 25, 29, 32
and 36 µM of 1C2mac incubated with DNA and BamH1, followed by
the marker.
Fig. 7 Gel electrophoresis pattern of PBR322 DNA treated with Cu–
1C2mac and BamH1 restriction enzyme. The lanes (from left to right)
correspond to the marker, control-DNA only, control-DNA and
BamH1 only, then concentrations of 1, 5, 10, 15, 20, 25, 30, 35, 40, 45
and 50 µM of Cu–1C2mac incubated with DNA and BamH1, followed
by the marker.
restriction. There is also a possibility that the daunomycin
interferes with PvuII at high concentrations.
may be less labile because of a stronger interaction with DNA
resulting in the slower progress of the bulkier complex/DNA
adduct through the electrophoresis gel. The 2ϩ charge of
the copper within the macrocycle may also strengthen the
interaction of the complex with DNA. However, it is notable
that the UV-titrations gave similar values for the binding
constants of 1C2mac and its copper complex. Therefore, the
changes in mobility and unwinding are likely to be specific
effects of the copper ion or the complex on these properties.
It is clear that adding a macrocyclic moiety (with or without
1C2mac causes no variation in the rate of migration of the
linear DNA consistent with the effects on supercoiled DNA
being almost entirely due to unwinding. 1C3 and 2C3 cause a
slight decrease in mobility after the addition of 1 µM of
complex until 21 µM of complex has been added, after which
the decrease tapers off. This decrease is probably due to the
positive charge on the macrocyclic moiety reducing the attrac-
tion to the electrode and/or to the increased size of the DNA
adduct relative to the DNA alone.
Complexation of copper to the 1C2mac adduct ligand
substantially modifies the binding of the complex (Fig. 6) as
a
metal present) greatly enhances the binding of the
intercalating moiety. There is significant disruption to the DNA
arising from the combination of an intercalator uncoiling
the DNA and a macrocyclic moiety reducing the intercalators
lability and itself interacting with the DNA.
Molecular modelling
Molecular modelling was used to investigate the interaction
with DNA of the copper complexes of the anthraquinone:
macrocycle adducts. The effect of the position of the side chain
linking the anthraquinone to the macrocycle on interactions
between the DNA and the macrocycle was examined by
modelling the binding of 1C2mac and 2C2mac. Models were
investigated with the long axis of the anthraquinone either
perpendicular to the bases or parallel to them and with the
macrocycle lying in either the minor groove or the major
groove. In each case, only a single enantiomer was modelled
because detailed investigation of the interactions between the
macrocycle and the DNA are beyond the scope of the current
study.
Fig. 8 shows Cu–1C2mac bound to DNA with the anthra-
quinone moiety intercalating from the major groove. For either
orientation of the anthraquinone (parallel or perpendicular),
the metal containing macrocycle can sit unhindered in the
major or minor groove. Fig. S7† shows the overlap of the
anthraquinone moiety of the cyclam/1-anthraquinone com-
plexes with the pair of bases above and below the plane, with
the anthraquinone in different orientations. When the 1-anthra-
quinone is inserted parallel to the DNA bases, slight rotation
is required to accommodate the side chain and macrocyclic
moiety in the minor groove (ESI,† Fig. S7(a)). The cyclam
causes little steric hindrance when the side chain protrudes into
the major groove (ESI,† Fig. S7(b)). When the 1-anthraquinone
is inserted perpendicular to the DNA bases with the side chain
in the minor groove (ESI,† Fig. S7(c)), it slides out into the
minor groove to reduce the steric interactions with the DNA.
When the side chain is placed in the major groove (ESI,†
Fig. S7(d)), there is a rotation away from the perpendicular
orientation to accommodate the side chain. There is significant
overlap between the DNA bases and the anthraquinone when
its long axis is parallel to the long axis of the DNA bases, but
Fig. 6 Gel electrophoresis pattern of PBR322 DNA treated with
Cu–1C2mac. The lanes (from left to right) correspond to the marker,
control-DNA only, then concentrations of 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15,
20 and 25 µM of Cu–1C2mac incubated with DNA, followed by the
marker.
indicated by the lower concentration required to fully unwind
the supercoiled DNA (2 µM, Rt = 0.07). This may be due to the
charge of the copper() complex, but there is also the possibil-
ity of nucleophilic sites on the DNA binding in the axial
positions of the copper resulting in greatly increased binding
and/or unwinding. Metal binding to DNA has been reported
previously for Zn2ϩ bound in a cyclen/anthraquinone ligand.12
The addition of the restriction enzymes BamH1 and PvuII in
the presence of Cu–1C2mac adduct resulted in the restriction
of the DNA to its linear form over the full range of concen-
trations tested (Fig. 7). This indicates that the Cu–1C2mac
complex does not inhibit restriction at these sites. However,
Cu–1C2mac caused more unwinding than daunomycin, so it is
possible that, even though Cu–1C2mac causes less inhibition of
the PvuII digestion enzyme, that it binds more strongly at other
sequences.
There was a slight decrease in the mobility of linear DNA
up to a concentration of 10 µM of Cu–1C2mac (Rt = 0.33)
probably due to the DNA becoming saturated with the
complex. This contrasts with the effect of 1C2mac which causes
no change in the mobility of linear DNA over the con-
centration range examined. This difference could be due to a
combination of factors such as changes in the mass, charge or
degree of unwinding of the DNA. The Cu–1C2mac complex
D a l t o n T r a n s . , 2 0 0 3 , 2 7 2 8 – 2 7 3 6
2734

View Article Online
Table 3 Comparisons between the modelling fetures of the 1- and 2-anthraquinone:macrocycles
Side chain position
Orientation compared to DNA bases
Groove
Energy/kJ molϪ1
Cavity/Å
Unwinding angle/Њ
1
1
1
1
2
2
2
2
Perpendicular
Perpendicular
Parallel
Minor
Major
Minor
Major
Minor
Major
Minor
Major
Ϫ2799
Ϫ2797
Ϫ2814
Ϫ2787
Ϫ2799
Ϫ2780
Ϫ2814
Ϫ2796
6.7
6.6
6.6
6.8
6.9
6.5
6.4
6.4
19.5
16.7
15.1
16.4
10.9
13.4
18.8
21.5
Parallel
Perpendicular
Perpendicular
Parallel
Parallel
there is less overlap when it is perpendicular to the DNA base
axis (ESI,† Fig. S7(c) and (d)).
Fig. S8† shows the overlap of the anthraquinone moiety of
the cyclam/2-anthraquinone complex, Cu–2C2mac, with the
pair of bases above and below it. When the 2-anthraquinone is
inserted perpendicular to the DNA bases, with the side chain in
either the major or minor groove (ESI,† Fig. S8(a) and (b)),
there is no change needed in the orientation to accommodate
the side chain and the macrocycle. When the 2-anthraquinone is
inserted parallel to the DNA bases with the side chain in the
minor groove (ESI,† Fig. S8(c)), the anthraquinone shifts
sideways to reduce the steric interactions between the DNA and
the side chain. When the side chain is placed in the major
groove (ESI,† Fig. S8(d)), only slight rotation away from the
perpendicular orientation is necessary to accommodate this
side chain. There is significant overlap between the DNA bases
and the anthraquinone when the long axis of the 2-anthra-
quinone is orientated parallel to the long axis of the DNA
bases, but there is less overlap when it sits in a position
perpendicular to the long axis of the DNA bases (ESI,† Fig.
S8(c) and (d)).
All of the intercalative modes modelled, regardless of the
position or length of the side chains or the orientation of
the anthraquinone within the DNA, appear to be viable due to
the minimal disturbance to the DNA stacking and the signifi-
cant amount of overlap between the DNA bases and the
anthraquinone rings. Some of the features of the models,
including energies, size of the cavity created with intercalation
and the amount of DNA unwinding are listed in Table 3.
The energies are similar and cannot be used to distinguish
between the different binding orientations. The cavity sizes are
all similar and are in accord with reported values for other
intercalating moieties of approximately 6.8 Å.33 The unwinding
angles vary between 13.4 and 21.5 Њ, with no systematic trends
evident.
The models of the Cu–1C2mac with the 1-anthraquinone
moiety intercalated in an 8-mer showed that the complex was
better accommodated when the 1-anthraquinone was in the
parallel position (especially when in the major groove) with
little steric hindrance between the complex and the DNA. Even
though the position of the side chain was better accommodated
in the major groove due to the larger space available, the elec-
trostatic interactions between the cationic groups in the side
chains of the intercalators and the negative electrostatic
potential of the minor groove34 as well as the possibility of
close van der Waals contacts within the minor groove are
expected to favour binding from this groove. The anthra-
quinones with the side chains in the 2-position were easily
accommodated in the DNA either parallel or perpendicular to
the DNA bases.
It is clear from the models that the chain length does not
influence the ability of the cyclam/anthraquinone to intercalate.
Therefore, the effects of chain length on the plasmid binding
properties described above, are probably indicative of specific
interactions between the macrocycle and the DNA that are
facilitated by the longer linker.
Fig. 8 Molecular models of metal-1C2mac with the anthraquinone
moiety intercalated within the DNA in either (a) the parallel position or
(b) the perpendicular position, with the metal-macrocycle sitting out in
the major groove.
D a l t o n T r a n s . , 2 0 0 3 , 2 7 2 8 – 2 7 3 6
2735

View Article Online
15 G. M. Sheldrick, in Crystallographic Computing 3 ed. G. M.
Conclusions
Sheldrick, C. Kruger and R. Goddard, Oxford University Press,
1986.
16 teXsan, Version 1.8 ed., Molecular Structure Corporation, The
Woodlands, TX, 1992–1997.
17 International Tables for X-ray Crystallography, Kynoch Press,
Birmingham, vol. 4, 1974.
18 J. A. Ibers and W. C. Hamilton, Acta Crystallogr., 1964, 17, 781.
19 D. C. Creagh and W. J. McAuley, International Tables for
Crystallography, Vol. C, Table 4.2.6.8, Kluwer Academic Publishers,
Boston, MA, 1992.
The anthraquinone/cyclam adducts and their metal complexes
represent exciting new compounds that have excellent DNA
binding properties. There is the potential to produce many
variations of this adduct by varying such things as the inter-
calating moiety, the nature of the metal ion, macrocycle size,
composition and/or donor atom types, side chain position,
length and/or composition as well as the number and/or
positions of the intercalators on the macrocycle. These can all
be varied to optimise the adduct properties and achieve desired
behaviours.
20 D. C. Creagh and J. H. Hubbell, International Tables for
Crystallography, Vol. C, Table 4.2.4.3, Kluwer Academic Publishers,
Boston, MA, 1992.
21 C. K. Johnson, ORTEP-II,
A FORTRAN Thermal-Ellipsoid
Plot Program, Oak Ridge National Laboratory, Oak Ridge, TN,
1976.
22 R. D. Wells, J. E. Larson, R. C. Grant, B. E. Shortle and
C. R. Cantor, J. Mol. Biol., 1970, 54, 465.
References
1 S. Jurisson, D. Berning, W. Jia and D. Ma, Chem. Rev., 1993, 93,
1137.
2 E. Kimura and E. Kikuta, J. Biol. Inorg. Chem., 2000, 5, 139.
3 E. Kimura, T. Ikeda and M. Shionoya, Pure Appl. Chem., 1997, 69,
2187.
4 D. Gibson, N. Mansur and K. F. Gean, J. Inorg. Biochem., 1995, 58,
79.
5 J. Whittaker, W. D. McFadyen, G. Wickham, L. P. G. Wakelin and
V. Murray, Nucleic Acids Res., 1998, 26, 3933.
6 L. C. Perrin, P. D. Prenzler, C. Cullinane, D. R. Phillips,
W. A. Denny and W. D. McFadyen, J. Inorg. Biochem., 2000, 81, 111.
7 M. Shionoya, T. Ikeda, E. Kimura and M. Shiro, J. Am. Chem. Soc.,
1994, 116, 3848.
23 S. Arounaguiri and B. G. Maiya, Inorg. Chem., 1996, 35, 4267.
24 J. A. Broomhead, L. M. Rendina and L. K. Webster, J. Inorg.
Biochem., 1993, 49, 221.
25 T. M. Jones-Wilson, K. A. Deal, C. J. Anderson, D. W. McCarthy,
Z. Kovacs, R. J. Motekaitis, A. D. Sherry, A. E. Martell and
M. J. Welch, Nucl. Med. Biol., 1998, 25, 523.
26 HyperCube, in HYPERCHEM, Release 4.5 for Windows, 1995.
27 T. W. Hambley, in MOMECSG, Programme for Strain Energy
Minimisation, 1996.
28 M. Engeser, L. Fabbrizzi, M. Licchelli and D. Sacchi, Chem.
Commun., 1999, 1191.
8 E. Kimura, M. Kikuchi, H. Kitamura and T. Koike, Chem. Eur. J.,
1999, 5, 3113.
9 E. Kikuta, M. Murata, N. Katsube, T. Koike and E. Kimura, J. Am.
Chem. Soc., 1999, 121, 5426.
10 S. Aoki and E. Kimura, J. Am. Chem. Soc., 2000, 122, 4542.
11 E. Kimura, H. Kitamura, K. Ohtani and T. Loike, J. Am. Chem.
Soc., 2000, 122, 4668.
29 J. Katzhendler, K.-F. Gean, G. Bar-ad, Z. Tashma, R. Ben-Shoshan,
I. Ringel, U. Bachrach and A. Ramu, Eur. J. Med. Chem., 1989, 24,
23.
30 D. Gibson, I. Binyamin, M. Haj, I. Ringel, A. Ramu and
J. Katzhendler, Eur. J. Med. Chem., 1997, 32, 823.
31 M. J. Scott and R. H. Holm, J. Am. Chem. Soc., 1994, 116,
11357.
12 E. Kikuta, R. Matsubara, N. Katsube, T. Koike and E. Kimura,
J. Inorg. Biochem., 2000, 82, 239.
13 G. M. Sheldrick, SADABS. Empirical absorption correction
program for area detector data, 1996.
32 P. Almond, S. D. I. Cutbush, S. A. Islam, R. Kuroda and S. Neidle,
Acta Crystallogr., Sect. C, 1983, 39, 627.
33 H. M. Berman and P. R. Young, Annu. Rev. Biophys. Bioeng., 1981,
10, 87.
14 Bruker SMART, SAINT, XPREP, Bruker Analytical X-ray
Instruments, Inc., Madison, WI, USA, 1995.
34 G. M. Blackburn and M. J. Gait, Nucleic Acids in Chemistry and
Biology, IRL Press at Oxford University Press, Oxford, 1990.
D a l t o n T r a n s . , 2 0 0 3 , 2 7 2 8 – 2 7 3 6
2736