Solution chemistry of 1,15-bis(N,N-dimethyl)-5,11-dioxo-8-(N-benzyl)-1,4,8, 12,15-pentaazapentadecane with metal ions of biological interest - Insights toward active metal ion containing therapeutics and diagnostic agents

Article abstract of DOI:10.1039/b602977a

The equilibrium constants of Cu(ii), Zn(ii), Ca(ii) and Gd(iii) with 1,15-bis(N,N-dimethyl)-5,11-dioxo-8-(N-benzyl)-1,4,8,12,15-pentaazapentadecane (L^a) have been studied at 25 °C and an ionic strength of 0.15 mol dm^-3. Copper forms more stable complexes with L^a than the other metal ions investigated. This is probably due to the ease with which Cu(ii) deprotonates the nitrogen donor atoms of the amide groups. UV/Vis spectrophotometric data indicate tetradentate binding of the ligand towards copper in [CuL^aH_-1] and pentadentate binding in [CuL ^aH_-2]. Octanol-water partition coefficients of Cu(ii)-L^a complexes indicate that although these species are largely hydrophilic, approximately 5.62% of the [CuL^aH_-1] complex goes into the organic phase. This percentage may promote dermal absorption of copper with a calculated penetration rate of 3.75 × 10^-4 mm h^-1. The [CuL^aH_-1] species which predominates at pH 7.4 is a poor mimic of native copper-zinc superoxide dismutase. Blood-plasma simulation studies predict that L^a is unable to increase the low molecular mass copper fraction in vivo. This has been confirmed by biodistribution patterns, which are similar to those of ⁶⁴CuCl ₂. The Royal Society of Chemistry 2006.

Full text of DOI:10.1039/b602977a

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
Solution chemistry of 1,15-bis(N,N-dimethyl)-5,11-dioxo-8-(N-benzyl)-
,4,8,12,15-pentaazapentadecane with metal ions of biological
1
interest—Insights toward active metal ion containing therapeutics and
diagnostic agents†‡
a
b
a
b
Thembelani E. Nomkoko, Graham E. Jackson,* Bandile S. Nakani and Roger Hunter
Received 27th February 2006, Accepted 26th May 2006
First published as an Advance Article on the web 8th June 2006
DOI: 10.1039/b602977a
The equilibrium constants of Cu(II), Zn(II), Ca(II) and Gd(III) with 1,15-bis(N,N-dimethyl)-5,11-
a
◦
dioxo-8-(N-benzyl)-1,4,8,12,15-pentaazapentadecane (L ) have been studied at 25 C and an ionic
−
3
a
strength of 0.15 mol dm . Copper forms more stable complexes with L than the other metal ions
investigated. This is probably due to the ease with which Cu(II) deprotonates the nitrogen donor atoms
of the amide groups. UV/Vis spectrophotometric data indicate tetradentate binding of the ligand
towards copper in [CuL H−1] and pentadentate binding in [CuL H−2]. Octanol–water partition
coefficients of Cu(II)–L complexes indicate that although these species are largely hydrophilic,
a
a
a
a
approximately 5.62% of the [CuL H−1] complex goes into the organic phase. This percentage may
−
4
−1
promote dermal absorption of copper with a calculated penetration rate of 3.75 × 10 mm h . The
a
[
CuL H−1] species which predominates at pH 7.4 is a poor mimic of native copper–zinc superoxide
a
dismutase. Blood-plasma simulation studies predict that L is unable to increase the low molecular
mass copper fraction in vivo. This has been confirmed by biodistribution patterns, which are similar to
6
4
those of CuCl .
2
Introduction
dissociated in vivo. This led to the synthesis of a number of
triamine–diamide ligands containing a central amine with benzyl-
The disease rheumatoid arthritis (RA) is an inflammatory
process of auto-immune nature characterised by an invasive
synovial hyperplasia which leads to polyarticular destruction.
Chemotherapy in the form of anti-inflammatory and immuno-
suppressive drugs becomes the obvious choice at this stage of
disease progression. Complicating such efforts, however, is the
recognition that it is often not a single sequence that is ideally
targeted, but rather an ensemble of physiological imbalances.
Furthermore, these drugs are rather unsatisfactory being poorly
effective and having considerable and sometimes, severe and life-
threatening side effects. In spite of these challenges, the most
interesting contributions in recent years on RA focus on the
7
or phenylethyl-substituents. The presence of the central amine
was initially thought to provide a fifth, strongly coordinating,
donor atom for copper(II) resulting in an enhanced thermody-
namic stability of the complexes. The benzyl- and phenylethyl-
groups at this central amine, on the other hand, were envisaged
to increase the lipophilicity of these complex species. The po-
tentiometric results revealed that these triamine–diamide ligands
coordinated weakly to copper in spite of the presence of the central
3
,4
3,4
7
benzyl amine. The low formation constants of the copper(II)–
triamine diamide complexes were due to the strain imposed by
having four contiguous five-membered rings that destabilised these
systems. Blood-plasma mobilising indices indicated the copper
species were more likely to dissociate in plasma in spite of the
3,4
treatment of this difficult and disabling disease. There is vast
evidence implicating the involvement of copper in this disease.
5
,6
7
morphological alterations of the ligands. Low octanol–water
In attempts to search for more effective therapies for RA,
partition coefficients of these copper complexes indicated that
they are likely to be of little utility as topical agents in RA
chemotherapy despite the presence of benzyl- and phenylethyl-
7
Jackson et al. investigated complexes of copper(II) with 3,5,9,12-
tetraazatetradecanedioic acid (H
2
ttda) and 3,6,9-triazaundecane-
dioic acid (H dtda). It was observed that the resulting copper(II)–
2
7
groups.
carboxylate complexes were too stable in vivo and were excreted
intact in urine. A subsequent study on diamine–diamide systems
revealed that these ligands bonded weakly to copper while animal
experiments indicated that the corresponding copper complexes
As part of our ongoing research on the development of safer
low molecular mass copper complexes as anti-arthritic agents, we
report in this paper the synthesis of 1,15-bis(N,N-dimethyl)-5,11-
7
a
dioxo-8-(N-benzyl)-1,4,8,12,15-pentaazapentadecane (L ). The
formation constants of this ligand with copper(II), zinc(II), cal-
cium(II) and gadolinium(III) have also been determined. Zn(II)
and Ca(II) are potential competitors of Cu(II) in blood plasma
while the affinity of the ligand for all three divalent metal ions
should be taken into account in the evaluation of gadolinium
complexes as contrast agents in magnetic resonance imaging
a
Department of Chemistry, University of Transkei, P/Bag X1, Unitra, 5117
Umtata, South Africa
8,9
b
Department of Chemistry, University of Cape Town, P/Bag, 7700 Ron-
debosch, South Africa. E-mail: jackson@science.uct.ac.za; Fax: +27 021
6
897499; Tel: +27 021 6502531
1,2
†
Copper Anti-inflammatory Drugs in Rheumatoid Arthritis. Part 8.
This journal is © The Royal Society of Chemistry 2006
Dalton Trans., 2006, 4029–4038 | 4029

View Article Online
1,2
a
(
MRI). L was designed to incorporate a methylene group on
each arm of the benzylamine moiety to decrease the strain of
having four contiguous five-membered rings by substituting the
middle chelate rings with two six-membered rings. The interest
a
7
in L , despite the aforementioned discouraging results, also stems
10
from its resemblance to oligopeptides, which are believed to be
involved in a variety of physiological functions and the fact that
receptors expressed on cell surfaces are able to bind these small
peptides with great affinity.
UV/Vis spectrophotometry was used to infer the structures of
the copper species in solution. As a measure of the lipophilicity of
the copper complexes, octanol–water partition coefficients were
investigated while the nitroblue tetraazolium (NBT) assay was
used to determine the superoxide dismutase mimetic activity of
the copper complex present at pH 7.4. Computer simulation of
blood plasma was used to predict the in vivo speciation of copper
a
1–4
Fig. 1 Schematic diagram of L and L
.
a
in the presence of L while these predictions were checked by
biodistribution experiments on mice. These data were also used
to calculate the in vivo speciation of Gd(III)/L , a possible MRI
contrast agent.
involving a mixture of the all s-cis and all s-trans diastereoisomers
(one set of resonances for each), or as the ligand having one
arm in the s-cis configuration and the other in the s-trans. Of
the two possibilities, the latter was considered to be the more
likely in view of the expected high-energy of the all cis form.
A mixture of all possibilities of diastereoisomers was precluded
since more resonances would have been expected in the spectrum.
Although not a quantitative indication due to relaxation times,
the observed 75 : 25 ratio would thus indicate an approximate 50 :
50 mixture of trans–trans to cis–trans. The appearance of such
a high proportion of the higher energy cis-moiety may indicate
preferences in the transition state of the second substitution
a
Results and discussion
Ligand synthesis
The ligand was prepared by a procedure similar to that reported
7
before for similar ligands. Refluxing one mole equivalent of ben-
zylamine with two mole equivalents of ethyl 3-bromopropionate
in the presence of acetonitrile and Hunig’s base (diisopropylethy-
step involving interaction, e.g., hydrogen-bonding, with the trans-
◦
13
lamine) for 30 min at 120 C in a cycloaddition tube afforded a mix-
amide already in place. Doubling of resonances in the
C
12
ture of oily ethyl-3-benzylaminopropionate and diethyl-4-aza-N-
benzylheptanedioate, which could be separated by column chro-
matography. Further reaction via reflux of one mole equivalent of
pure diethyl-4-aza-N-benzylheptanedioate with 2.5 mole equiv-
alents of N,N-dimethylethylenediamine in a cycloaddition tube
NMR spectra of amides has been noted previously. Similarly,
13
13
Bergeron et al. have reported similar complex C-NMR spectra
for desferrioxamine B as a result of E- and Z-configurations.
However, this behaviour was not observed for the previously
studied N,N -bis(2-hydroxyiminopropionyl) propane-1,3-diamine
ligand reported by us, nor for dioxocyclomacrocycles, 2,6-dioxo-
ꢀ
◦
2
at 80 C for 24 h yielded 1,15-bis(N,N-dimethyl)-5,11-dioxo-8-
a
(
N-benzyl)-1,4,8,12,15-pentaazapentadecane (L ). The molecular
1,4,7,10,13-pentaazacyclopentadecane and 2,6-dioxo-1,4,7,11,14-
pentaazacycloheptadecane, investigated by Steenland and co-
a
1
structure of L is depicted in Fig. 1. The combination of H-NMR
and C-NMR data together with Fast-Atom Bombardment Mass
1
3
14
workers.
7
Spectroscopy (FAB-MS in positive mode; m/z envelope centred at
In contrast to the earlier syntheses of related ligand systems,
a
+
a
3
92.54259 [(HL ) ] with correct isotope distribution is indicative
the compound L was obtained in a mono-hydrochloride form
a
+
of the formation of the protonated ligand [C21
H
38
N
5
O
2
for (HL ) ]
(C21
H
38
N
5
2
O Cl) as a pure diastereoisomer suitable for analysis and
with 98.98% total abundance. The calculated molecular mass of
potentiometric titrations. The presence of the mono-protonated
a
pure ligand (L ) formulated as C21
species C21 are 391.29473 and 392.30255 Da, respectively,
based on atomic weights given by Williams and Fleming.
H
37
N
5
O
2
and the protonated
ligand was revealed by an m/z signal centred at 392.54259
a
+
H
38
N
5
O
2
[HL ] (C21
H
38
N
5
2
O ). This resulted in the elimination of the
11
hydrochloridation step involving the bubbling of dried HCl gas
a
a
Satisfactory micro-analysis could not be obtained probably due to
the hygroscopic nature of this compound. It should, however, be
over the L solution. The higher molecular weight of L compared
to its previously synthesized analogues and the high basicity of
its terminal nitrogen donor atoms is probably the reason for
the compound being obtained as the solid mono-hydrochloride
when crystallised from dichloromethane. The dihydrochloride
form of the ligand was not obtained probably because of the
close proximity of the protonated and unprotonated terminal
tertiary amino nitrogen donor atoms to each other or a possible
hydrogen bond between them. The purity of the ligand was also
confirmed by acid–base glass electrode potentiometric titrations
mentioned that all carbon signals (d in ppm) as indicated below as
C
C=O (172.11, 172.32), Ph (127.10, 127.17, 128.16, 128.30, 128.40,
1
5
28.91, 138.46, 139.47), ArCH
8.00), COCH (49.91, 51.25), ((CH
) (36.68, 36.72), BnNCH (33.87, 35.87), were observed
2
(58.42, 60.18), CONHCH
2
(57.97,
2
3
)
2
NCH )) (45.07, 45.21),
2
N(CH
3
)
2
2
to exist in pairs, indicating the presence of diastereoisomers
around the amide bonds, with s-cis defined as involving a cis-
relationship between the alkyl substituents attached to the amide
◦
−3
group and s-trans, a trans-one. In view of C
2
-symmetry in the
at 25 C and I = 0.15 mol dm and was found to be greater than
ligand, the two sets of resonances could be interpreted as either
95% pure.
4
030 | Dalton Trans., 2006, 4029–4038
This journal is © The Royal Society of Chemistry 2006

View Article Online
14
Potentiometry
et al. have recently observed a similar electron withdrawing
property of an amide group in the dioxomacrocylic analogues
In any solution containing a metal ion, M, ligand, L and proton, H,
a
of L . The benzylamine is also 2–4 log units more basic than the
a complex whose general formula denoted by M
and the associated equilibrium constant is then designated by the
symbol bpqr. For convenience the species M is sometimes
p
L
q
H
r
is predicted
1
4
7
a
central amine of L –L (Fig. 1). L contains an extra methylene
group on each side of the amine. This has two effects. Firstly, it
increases the distance between the protonated amine groups which
decreases the electrostatic repulsion. Secondly, the inductive effect
of the methylene group increases the basicity of this central amine.
We believe that the basic nitrogen sites in the ligand should not be
affected by conformational inter-conversion discussed above and
hence identical protonation constants should be observed for the
cis- and trans-conformers.
p
q r
L H
denoted by the three stoichiometric coefficients p, q and r given
in the order M, L, H, e.g. 11−1 denotes the species MLH−1. A
negative value for r refers to proton removal or hydroxide-ion
addition during formation of the complex. Thermodynamically
these two processes are equivalent and cannot be distinguished
potentiometrically. The log bpqr’s are given in Table 1. For
simplicity, the charges are omitted in these species.
Complex formation and deprotonation function
Protonation constants
a
To take full advantage of the large amount of data generated, com-
L contains three basic sites with pKa1 = 8.823, pKa2 = 8.431 and
16
plex formation (Z -bar) and deprotonation (Q-bar) functions
M
pKa3 = 6.898. pKa1 and pKa2 correspond to proton additions to the
terminal tertiary amines while pKa3 refers to the protonation of
the central less basic benzylamine. The difference between the first
and second protonation constants is 0.39 log units, which indicates
that proton additions to these sites are independent of each other.
The central amine is some 2 log units less basic than the terminal
amines because of the electronic withdrawing benzyl group and
electrostatic repulsion from the protonated terminal amines. The
terminal amino groups of L are both one order of magnitude less
basic than those of N,N,N ,N -(tetramethyl)ethylenediamine and
other tertiary amines. This is due to the electron withdrawing
effect of the amide groups at the beta(b)-positions. Steenland
were evaluated. The former measures the number of ligands bound
per metal while the latter indicates the number of protons released
on metal-ion complexation. The two functions were calculated
from the free and total concentrations of the participating
components as well as the evaluated log bpqr’s and were then used
as criteria for testing the validity of the potentiometric model
proposed in the data analysis. The complexation chemistry of
a
a
the Cu(II)–L system is characterised by exclusive formation of
ꢀ
ꢀ
mononuclear species of 1 : 1 M : L stoichiometry. This is indicated
15
by the levelling off of Z -bar at a value of 1 (Fig. 2). The
M
formation of either hydroxo and/or mixed-hydroxo complexes
also characterises this system as evident from the fanning back
a
a
+
of the Z -bar at pL values less than 4.
Table 1 Formation constants, log bpqr, for L with H , Cu(II), Zn(II) and
M
◦
−3
−
Ca(II) and Gd(III) studied at 25 C and I = 0.15 mol dm (Cl ). rpqr
denotes the standard deviation in log bpqr; R is the Hamiltonian R-factor
and n is the number of titration points. The general formula of a complex
is M
H
p
q r
L H
Metal
p
q
r
log bpqr
r
pqr
n
pH range
R
H
+
H
0
0
0
1
1
1
1
2
3
8.823 0.010 943 2.2–10.8
17.254 0.008
24.152 0.012
0.012
Cu(II)
Zn(II)
Ca(II)
Gd(III)
1
1
1
1
1
1
1
1
1
0
15.020 0.022 389 2.2–11.0
8.911 0.018
0.034
−1
−0.896 0.039
−2
−8.677 0.042
1
1
2
1
1
1
1
1
2
0
−2
19.920 0.045 697 2.4–10.5
3.462 0.147
−9.018 0.044
0.030
0.022
0.065
−2 −13.201 0.027
a
M
Fig. 2 Experimental and theoretical Z -bar curves for the Cu(II)–L
1
1
1
1
1
1
1
1
2
1
0
−1
19.636 0.034 674 2.4–11.0
system. M : L ratios 1 : 3 (᭜), 1 : 4 (᭹), 1 : 5 (ꢀ) and 1 : 6 (ꢁ) were used.
The model given in Table 1 was used to calculate the theoretical line. For
clarity only a few points have been plotted.
10.808 0.116
2.424 0.046
−8.857 0.052
1
2
1
1
1
1
2
1
1
1
2
−2
−2
21.333 0.037 509 2.3–11.0
0.305 0.070
−9.695 0.045
The Q-bar function (Fig. 3), on the other hand, rises to an
average value of 2.7 in the small pH range of 5.5–6.5. This is an
indication that approximately three protons have been released due
to complexation. Ideally, 2 and 3 protons should be released due to
the exclusive formation of 111 and 110 complexes, respectively. The
obtained average Q-bar = 2.7 is due to the simultaneous formation
of these two species as revealed by the species distribution curves.
We have recently observed similar complexation patterns for Cu(II)
−3 −19.343 0.036
−4 −30.924 0.074
a
−3
−3
Ligand and metal concentration range 3.0–8.0 × 10 mol dm and 2.0–
−
3
−3
6
1
.0 × 10 mol dm , respectively with metal-to-ligand ratios of 1 : 3, 1 : 4,
: 5 and 1 : 6. For metal-to-ligand ratios of 1 : 1, 1 : 2 and 2 : 1, precipitation
was observed.
This journal is © The Royal Society of Chemistry 2006
Dalton Trans., 2006, 4029–4038 | 4031

View Article Online
Deprotonation of ML occurs with a pK
a
(logb110 − logb11−1
)
of 8.0. This value is close to the hydrolysis constant of an axially
2
+
18
coordinated water molecule in [Cu(H
2
O)
6
]
(7.87 ) and the pK
a
of the terminal amine. It is unlikely that the second amide proton
would be lost while the terminal amine remained protonated,
which would be necessary to achieve the MLH−1 stoichiometry.
Similarly the pK
a
of MLH−1 to form MLH−2, is 9.8. Since this is
above the pK of the terminal amine, the proton must come from a
a
coordinated water molecule or the second amide. It is not possible
to decide, on potentiometric grounds, the source of these protons.
14
Steenland et al. have postulated that the amide is deprotonated
a
in the Cu(II) complexes of the macrocyclic analogues of L
a
Although copper forms stable complexes with L , the reported
constants are 1–1.5 log units lower than those of the previously
1
–4
7
investigated triamine–diamide ligand systems (L , Fig. 1).
a
Fig. 3 Experimental and theoretical Q-bar and n-bar curves for the
Furthermore, the increased basicity of the central amine in L
is not reflected in the Cu(II) formation constants. This may be
due to the formation of two less stable contiguous six-membered
rings on deprotonation and coordination of the two amide groups.
a
Cu(II)–L system. M : L ratios 1 : 3 (᭜), 1 : 4 (᭹),1 : 5 (ꢀ) and 1 :
6
(ꢁ) were used. The model given in Table 1 was used to calculate the
theoretical line. For clarity only a few points have been plotted.
14
Steenland et al. have reported log b110’s of 8.5 and 10.1 for Cu(II)
1,2
a
with similar ligands. The n-bar function measures the average
number of protons that would be bound to the ligand in the
absence of the metal ion. It is worth noting that the Q-bar never
rises above the n-bar curve below pH 8.0 (Fig. 3) because of the
presence of protonated nitrogen donor atom(s) on the ligand in the
acidic to weakly basic conditions. An excellent agreement between
with the dioxomacrocycle analogues of L . These are close to the
value reported in this study for the same species. However, our
measured constants for MLH−1 and MLH−2 are 4–5 log units lower
than those for these same ligands (log b11−1 = 4.6 and 3.3, log b11−2
14
of −4.3 and −5.9). An explanation for this is the macrocyclic
effect and the strain introduced by having two contiguous six-
membered rings.
the observed and calculated Z
ratios and satisfactory low standard deviations in the measured
b’s and Hamilton R -factor (Table 1) give confidence to the
M
-bar and Q-bar for varying M : L
It should be mentioned that, in order to bind to the metal
ion, the ligand will have to undergo a conformational trans–cis
H
1,2
proposed complexation model. Chemical speciation is a function
of pH and varies as the acidity or basicity of the solution is
changed. Furthermore, ligands bind to metal ions to satisfy specific
conformational requirements. For that reason and just like the
protonation equilibria, we did not expect the presence of the cis-
and trans-conformers to lead to different complexation patterns.
isomerisation. We have witnessed previously that this leads to
slow kinetics in metal-ion complexation, a phenomenon that was
observed in the present systems.
19
He Bu et al. have recently reported Cu(II) and Zn(II) com-
plexation equilibria with two mono- and dioxocyclam ligands
containing 2-pyridylmethyl groups. Contrary to expectations,
copper(II) did not facilitate the deprotonation of the amide groups.
In fact, the crystal structure of the ML complex showed that
the metal ion coordinated to the ligand through the appended
pyridylmethyl nitrogens and the backbone amino and carbonyl
oxygen atoms. This could be because the pyridyl nitrogens are
much more basic than the amide nitrogens. Similarly, Santos and
Formation constants
The formation constants for the MLH, ML, MLH−1 and MLH−2
a
complexes of the Cu(II)–L system are given in Table 1. The
−
3
a
calculated species distribution of a Cu(II) (0.001 mol dm )–L
−
3
20
(
0.003 mol dm ) solution indicates that the MLH and ML form
coworkers have also shown that if an hydroxamate group is
simultaneously at pH 4.6, where the former reaches a maximum
value of 42% at pH 6.0 while the metal is 85% present as the latter
species at pH 7.4. At this physiological pH, the MLH−1 species
constitutes the remaining 15% and reaches a maximum value of
attached to the dioxocyclam ring, Cu(II) coordinates to the oxygen
of the hydroxamate rather than the amide nitrogen. The pH driven
translocation of Cu(II) from one ligand compartment (containing
predominantly nitrogen donor atoms) to another (which contains
amide groups) is also evidence of the difficulty associated with
the deprotonation of the amide moiety. Another notable feature of
the dioxomacrocycle ligands is the presence of MLH−1 as a minor
species. This is understandable for the macrocylic ligands, where
decreased flexibility of the ligands causes both amide nitrogens to
21
6
3% at pH 8.6 while the MLH−2 predominates above pH 9.0.
The MLH species is believed to be formed by coordination of
the benzylamine and one of the terminal tertiary amines. Since
the pH at which this species forms is well below the pK of the
a
terminal amine, if it were not coordinated, it would be protonated.
There appears to be little or no stabilisation of this species due
to the binding of the neutral carbonyl oxygen(s) to the Cu(II), a
phenomenon that has been observed previously by Kholeif and
coordinate simultaneously and preferentially form MLH−2
.
Zinc(II), calcium(II) and gadolinium(III) form much weaker
a
complexes than copper(II) with L , with protonated (MLH
2
and
17
Anderegg. Deprotonation of MLH occurs with a pK
a
of 6.1
MLH) and hydroxo and/or mixed-hydroxo species predominat-
ing. In fact, the predominance of the protonated species is an
indication of the weak complexes formed by these metal ions with
the ligand. We do not believe that these metal ions are capable of
deprotonating the amide moiety, a phenomenon which has also
(
log b111 − log b110) and leads to formation of the ML species.
The proton that is lost could come from the protonated terminal
amine or from the amide nitrogen if either were to coordinate to
the metal i.e. metal-assisted deprotonation.
4
032 | Dalton Trans., 2006, 4029–4038
This journal is © The Royal Society of Chemistry 2006

View Article Online
1
9
−1
3
−1
been observed by He Bu et al. for Zn(II) with amide containing
ligands. The reported hydroxo and/or mixed-hydroxo complexes
are thought to result from the ionisation of coordinated water
molecules. The aqueous chemistry of Gd(III) parallels in many
ways that of Ca(II). This is not surprising in view of the fact that
(e = 40 mol dm cm ). This blue shift reflects the increased
a
ligand field splitting of L as opposed to H
of the ML complex, there is a further 20 nm blue shift (kmax
2
O. On formation
=
−
1
3
−1
580 nm; e = 115 mol dm cm ). We attribute this blue shift to
a change in coordination from a carbonyl oxygen to an amide
22
26,27
both metal ions have nearly the same ionic radii and are “hard”
nitrogen.
There is also a substantial increase in the molar
23
in terms of the HSAB theory and hence prefer hard negatively
charged oxygen donor atoms. In fact, the Gd(III) ion has been
used to probe the structural and functional behaviour of Ca(II) in
extinction coefficient of this species. This is a result of increased
tetragonal distortion and in fact, the extinction coefficient is
typical of square planar Cu(II). This observation supports the
conclusions drawn from the potentiometric results.
24
in vivo systems.
The MLH−1 species also has a kmax of 580 nm and a e of
−
1
3
−1
UV-visible spectrophotometry
120 mol dm cm . Since the spectrum of this species is very
similar to that of the ML complex, we conclude that there has
been little change in the coordination environment of the Cu(II).
This is difficult to rationalise as the loss of this proton has to
be metal assisted. One explanation is that the terminal amine
deprotonates, coordinates to the metal and the complex rearranges
so that the carbonyl oxygen moves to an axial position. Equatorial
coordination of the amine should cause a blue shift while axial
coordination results in a red shift. In this case, the two opposing
effects would cancel.
a
During solution studies of the Cu(II)–L system, distinct colour
changes were noted as the titrations progressed. This is because
the speciation of the solution changes with pH and the different
complexes have different absorption spectra. Since the elec-
tronic spectra reflect the coordination geometry of the complex,
UV/visible spectra were recorded for this system in the pH range
2
.5 to 11.0. At pH values below 3.54, an absorption band at 780 nm
predominated. The absorption spectrum at pH 5.5 consisted of
several peaks, one at 780 nm and others as broad humps between
−
1
The MLH−2 species has a kmax of 540 nm (e = 110 mol
6
40–700 nm. In the pH range 6.0–9.0, these multiple, broad bands
3
−1
dm cm ) and we attribute this 40 nm shift to coordination of
the last amide. From the potentiometric data, it was not possible
to decide if the proton was lost from the terminal amine or
a coordinated water molecule. Square-pyramidal and trigonal
merged into a single band which increased in intensity as the pH
was increased. Above pH 9, the intensity of the band decreased.
In order to gain more insight into the coordination geometries
of the complexes, individual species spectra were calculated by
deconvolution of the UV/Vis data. The resulting spectra for
the individual species are shown in Fig. 4. The deconvolution is
based on the metal-ion speciation which is calculated from the
potentiometric model. This speciation is incorporated into an
expanded Beer–Lambert Law which is then solved for the molar
absorptivity of the different species at each wavelength. Since
each wavelength is treated independently, the fact that smooth
spectra are obtained for each species lends confidence to the
2
+
bipyramidal Cu complexes of N
5
coordination usually exhibit
1,2
28,29
30
two bands at 728 and 900 nm.
Jubert et al. have observed a
broad band around 700 nm which they attributed to six coordinate
CuN species. The axial coordination of amino groups in these
6
2
+
Cu complexes causes a red shift in kmax of approximately 100 nm
compared to the absorption maxima observed in this study.
31
−
Nakani and Jackson have shown the coordination of OH in an
axial position leads to a red shift of 50–90 nm. Hence we conclude
that an axial water has not hydrolysed but rather that both amides
have coordinated.
2
+
thermodynamic model. The spectrum of [Cu(H
also calculated during the analysis is used as a check on the process.
2
O) ] , which is
6
Molecular mechanics
Molecular mechanics is a useful tool which can be used to
calculate the strain energy introduced into a ligand by virtue
of its coordination geometry. When used to its full extent, it is
possible to calculate the free energy of complexation. However,
these calculations require a large amount of computing power
since the transformation has to be integrated over time. For this
reason, we have calculated the internal strain energy introduced
when a ligand adopts a particular conformation and coordinates
to a metal ion. This strain energy would be one of the energy
contributions to the overall stability of the complex. The chelate
effect, as an entropic contribution, and the ligand field stabilisation
energy of the metal ion, are two other energy contributions,
which are important, but are not taken into account in these
calculations.
Fig. 4 Absorption spectra of the individual copper species.
Fig. 5shows five possible isomers for the different species formed
a
The broad absorption band observed for these complexes
by the Cu(II)–L system. Table 2 lists the different contributions to
2
2
2
2
envelopes the expected spin-allowed A1g← B1g, B2g← B1g and
← B1g transitions. Coordination of Cu(II) results in a 180 nm
blue shift in kmax; [Cu(H O) has a kmax of 780 nm and e <
0 mol dm cm while the MLH species has kmax of 600 nm
the strain energy. Structures 5a and 5b are two possible geometries
for the MLH species. In these structures, two of the amines
and both carbonyl groups are coordinated to the metal. The
terminal amine is protonated and not coordinated to the metal
2
2
25
E
g
2
+
2
6
]
−
1
3
−1
2
This journal is © The Royal Society of Chemistry 2006
Dalton Trans., 2006, 4029–4038 | 4033

View Article Online
Table 2 Total internal energy (EInt), bond stretching, angle bending and
torsion angle energies (kcal mol ) for different conformers presented in
Fig. 5
amine to coordinate to the metal ion. The internal strain energy of
this complex is 87.7 kcal mol , which could account for the low
stability of this complex.
−
1
−1
Two different geometries are presented for the MLH−2 species.
These differ only in the axial position. In Fig. 5e, the amide
nitrogen is axial while in Fig. 5f the central amine is axial. The
strain energy of these two conformations is very similar, however,
the total potential energy (including Coulombic interactions) of
Structure
Energy
a
b
c
d
e
f
Internal
Bond
Angle
57.6
11.3
25.0
19.1
77.7
16.3
28.7
31.9
44.5
10.5
26.7
6.9
87.7
13.0
47.0
25.5
73.7
11.9
44.7
16.2
72.5
7.3
48.4
15.8
−1
−1
5e is −110.2 kcal mol while that of 5f is −71.8 kcal mol . On
Torsion
this basis we would favour structure 5f. This structure is also more
aesthetically pleasing.
1
-Octanol–water partition coefficient studies
Available evidence indicates that compounds containing molecu-
−
1
lar weights less than 600 Da (g mol ) are ideally suited for passive
3
2,33
diffusion across the skin and other biomembranes.
has been found to depend on the lipophilicity of the drug and an
-octanol–water partition coefficient of a compound in the form
of log Poct/aq is often used by medicinal chemists as a measure
This process
1
32,33
of lipophilicity.
We have determined octanol–water partition
coefficients of the copper complexes in order to ascertain if they
can be applied as topical agents in the alleviation of the anti-
inflammatory process in RA. Fig. 6 shows the log Poct/aq values as
a
a function of pH for the Cu(II)–L system. A notable feature is
that the partition coefficients are always <1, i.e. the complexes are
all hydrophilic.
Fig. 5 Optimised molecular mechanics conformations of the different
a
possible structures for the Cu(II)–L system. For clarity protons have been
retained where they are necessary for the stoichiometry of the complex,
otherwise they, together with the benzyl substituent, have been removed.
a
ion. Surprisingly, structure 5b, which has 4 coordinating atoms
in a plane, has the larger strain energy. The main contribution
to the total internal energy comes from the torsion strain and
the main contribution to this energy comes from a distortion of
the planar amide geometry. The high kmax (600 nm) and low emax
are consistent with both structures although we would suggest
structure 5b, being more tetragonally distorted, should have a
higher extinction coefficient.
Fig. 5c is the most likely structure of the ML species. One amide
is coordinated and one terminal amine is protonated. The planar
arrangement of the ligand is consistent with kmax for this species.
This structure has the lowest torsion angle strain and the least
distortion of the amide planarity.
Fig. 6 Log Poct/aq of Cu(II)–L system plotted against pH. Each point is
the mean of three independent observations.
a
From the speciation of the Cu(II)–L system it is possible to
estimate the partition coefficient of the individual species. Thus
MLH has a log Poct/aq of −2.0, ML −1.79, MLH−1 −1.25 and
MLH−2 −1.25. These partition coefficients are all greater than
those determined previously for related ligands (Table 3). Leo
et al. have observed that incorporation of two methylene groups
into a ligand backbone adds approximately one log unit to
the partition coefficient. However, comparing L and L where
two methylene groups have been added, there has been a 100
fold increase in the partition coefficients. Increased metal ion
coordination should increase the lipophilicity of the complex
by decreasing the number of available nucleophilic sites capable
of hydrogen bonding. The partition coefficient results therefore
32
a
2
Only one structure (5d) is presented for the MLH−1 species in
which the terminal amine has lost a proton in coordinating to
the metal ion. This structure has a 7 membered chelate, which is
too large for any chelate effect, but is necessary for the terminal
a
support the proposed structures for the Cu(II)–L species.
4
034 | Dalton Trans., 2006, 4029–4038
This journal is © The Royal Society of Chemistry 2006

View Article Online
a
−3
Table 3 Octanol–water partition coefficients (log P) for the different
In fact, at a total L concentration of 1 mol dm , the Cu(II) p.m.i.
is less than 0.01. This is due to the weak complexes formed by the
metal ion with this ligand compared to the previously investigated
a
7
Cu(II) species with L and related ligands
Ligand
1,2
dioximes.
In order to validate the speciation calculations, in vivo biodis-
tribution was measured in mice using Cu(II) as a radiolabel. The
results are given in Table 4.
One of the noticeable features about the in vivo behaviour of both
Cu[CuL H−1] and Cu [CuCl
a
1
2
3
4
Species
L
L
L
L
L
6
4
MLH
ML
MLH−1
MLH−2
−2.00
−1.79
−1.25
−1.25
−6.3
−6.6
−5.3
−5.2
−5.6
−4.4
−3.7
−4.1
−4.5
−4.3
−4.7
−4.4
−4.3
−3.8
−3.5
−4.4
64
a
64
2
] is their rapid clearance from the
blood circulatory system. In fact, less than 1% of the total injected
1
,2,40
dose remains in the blood after 6 h. We have recently
reported
The log Poct/aq = −1.25 for the 11−1 complex indicates that
6
4
similar results for Cu[CuLH−1] complexes of dioxime ligands.
5
.62% of the complex partitions into the organic phase. Michaels
6
4
a
34
More than 60% of both Cu[CuL H−1] and the control is taken
up by the liver 6 h post-injection and drops to 30–40% at 24 h. The
high activity in the liver is expected since copper metabolism occurs
et al. have developed a predictive model for dermal absorption in
which partition coefficient of a permeate is used to predict maximal
transdermal flux. From the partition coefficient measurements, a
6
4
a
−
5
−1
a
here. The slightly lower activity in this organ for Cu[CuL H−1]
transdermal flux of 3.75 × 10 cm h is predicted for [CuL H−1].
compared to the control is compensated for by the high amount
of activity excreted in the urine, 8% and 25% after 6 and 24 h,
This flux is higher than that of a copper complex previously
patented for topical treatment of RA.
35
a
respectively. Overall, the biodistribution behaviour of [CuL H−1] is
similar to that of the control supporting the speciation calculations
which suggests that the complex would dissociate in vivo.
a
Superoxide dismutase mimetic activity of the [CuL H−1] complex
Copper–zinc superoxide dismutase (CuZnSOD) is an enzyme that
catalyses the dismutation of the super oxide anion radical to
Conclusion
36,37
molecular oxygen, water and/or hydrogen peroxide in vivo.
Recently, in vitro experiments have indicated that low molecular
a
Cu(II) forms more stable complexes with L than Zn(II), Ca(II)
and Gd(III) probably due to the ease with which the former
metal ion deprotonates the amide groups of the ligand. The
38
mass copper(II) complexes also possess SOD activity. Because
a
of this property, the SOD activity of the [CuL H−1] complex
36–38
a
was investigated using the NBT assay.
The concentration of
weak complexation of Gd(III) and L preclude its use as an MRI
the complex required to reduce diformazan formation by 50%
contrast agent. UV/visible spectrophotometric data indicate an
−
3
38
a
a
(
IC50) was measured as 11.7 lmol dm . Roberts and Robinson
−1
MN type of coordination of L towards Cu(II) in both [CuL H ]
4
−
3
a
have reported an IC50 value of 0.011 lmol dm for the native
−
and [CuL H ] complexes leaving one nitrogen donor atom
2
a
a
CuZnSOD. The high IC50 value exhibited by the [CuL H−1
]
−
protonated in the former species. Although, the [CuL H ] species,
1
complex is indicative of its low SOD activity when compared
predominant above pH 7.4 and is reasonably lipophilic, it is a
poor mimic of the native SOD. Its Cu(II) complexes are predicted
a
to the native enzyme. Although [CuL H−1] is a poor mimic of
−
•
CuZnSOD, it is capable of catalysing the dismutation of O
a factor of 5–9 better than previously investigated complexes.
There seems to be a correlation between the SOD activity and
2
by
to dissociate in blood-plasma and biodistribution experiments
1
,2
a
indeed confirmed that the [CuL H ] complex is too weak to
−
1
remain stable in vivo. Studies are underway to develop a method
38
the low thermodynamic stability. The low stability of good SOD
to characterise the nature of the species and/or metabolites of
−
•
a
mimics may be due to the presence of available sites for O
2
the [CuL H ] complex excreted in the urine while simultaneously
−
1
a
coordination and dismutation. For [CuL H−1], one coordination
modifying the ligand to increase the thermodynamic stability of
the resulting Cu(II) species.
site is occupied by a water molecule and dismutation may be due
to its replacement by the superoxide anion radical. More recently,
+
it has been observed that the reduced Cu
2
Zn(II) SOD is more
2
Experimental
39
stable than its oxidised analogue. The tetrahedral nature of the
active site of both the reduced and oxidised states in the enzyme
may be viewed as coordinatively unsaturated in sharp contrast to
Materials
a
Solvents such as dichloromethane, hexane, methanol, ethanol,
ethyl acetate, acetonitrile and triethylamine were dried and freshly
distilled using the normal procedures prior to use. All other
reagents were commercially available and of analytical grade and
were used without further purification unless noted otherwise.
Reactions and the progress of chromatographic separations were
monitored by thin layer chromatography (TLC). These were con-
ducted on silica-gel 60 F254 precoated TLC plates. The compounds
were visualised on these plates by illumination with 254 nm UV
light or staining with iodine vapour. Column chromatography was
performed using silica-gel 60, 230–400 mesh from Merck. Melting
points were measured using a Reichert Thermovar melting point
the high field provided by L .
Biodistribution experiments on mice
In order to assess the ability of the ligand to mobilise endogenously
available metal ion(s) in general, and copper(II) in particular,
a plasma mobilising index (p.m.i.) is used. P.m.i. is defined as
9
the ratio of the concentration of low molecular mass complex
species in the presence of the ligand (drug) and in its ab-
sence. Unlike 3,3,9,9-tetramethyl-4,8-diazaundecane-2,10-dione
5
ꢀ
dioxime (L ) and N,N -bis(2-hydroxyiminopropionyl) propane-
6
1,2
1
,3-diamine (L ), speciation calculations using the blood-plasma
model indicate that L is devoid of any copper mobilising ability.
a
1
13
apparatus and are uncorrected. Proton ( H) and C NMR proton
This journal is © The Royal Society of Chemistry 2006
Dalton Trans., 2006, 4029–4038 | 4035

View Article Online
a
Table 4 Percentage injected dose (%ID) per organ and per gram tissue for [CuL H−1] species
Complex
a
+
[
CuL H−1
]
2
[CuCl ]
Organ
Liver
Time/h
6
%ID/organ
%ID/g
%ID/organ
%ID/g
64.93 ± 0.10
28.21 ± 4.24
—
46.43 ± 3.97
22.44 ± 3.19
1.96 ± 0.02
3.45 ± 0.86
0.82 ± 0.15
1.94 ± 0.27
3.48 ± 0.76
3.30 ± 0.70
0.70 ± 0.10
1.64 ± 0.32
3.96 ± 0.46
7.82 ± 1.30
9.34 ± 0.31
13.75 ± 1.96
6.19 ± 0.39
8.06 ± 1.73
0.72 ± 0.01
2.14 ± 0.50
4.55 ± 0.35
4.41 ± 1.01
2.32 ± 0.34
5.57 ± 2.08
0.36 ± 0.08
1.05 ± 0.17
—
73.50 ± 1.89
37.14 ± 9.20
—
57.75 ± 5.14
29.66 ± 5.42
2.19 ± 0.18
4.61 ± 1.55
0.61 ± 0.06
2.00 ± 0.44
2.85 ± 0.67
3.60 ± 1.22
0.83 ± 0.16
1.94 ± 0.80
2.73 ± 0.24
10.49 ± 1.04
4.95 ± 0.67
10.50 ± 5.70
4.27 ± 0.08
11.44 ± 2.38
0.56 ± 0.01
2.03 ± 0.40
4.01 ± 0.45
4.95 ± 0.75
2.33 ± 0.26
5.89 ± 1.76
0.21 ± 0.02
1.03 ± 0.20
—
2
2
2
2
2
2
2
2
2
2
4
6
4
6
4
6
4
6
4
6
4
6
4
6
4
6
4
6
4
6
4
Blood
Carcass
Tail
—
—
7.98 ± 0.46
18.87 ± 4.30
1.26 ± 0.05
1.41 ± 0.29
0.11 ± 0.01
0.14 ± 0.05
0.50 ± 0.001
0.83 ± 0.13
1.36 ± 0.07
1.44 ± 0.35
1.91 ± 0.02
2.14 ± 0.79
1.74 ± 0.06
4.38 ± 1.39
11.77 ± 0.72
16.18 ± 1.58
0.25 ± 0.01
0.52 ± 0.19
0.08 ± 0.001
0.37 ± 0.22
7.67 ± 2.48
25.16 ± 2.11
4.86 ± 0.80
21.87 ± 3.88
1.28 ± 0.06
1.79 ± 0.53
0.08 ± 0.02
0.15 ± 0.05
0.42 ± 0.06
1.07 ± 0.43
0.85 ± 0.02
1.51 ± 0.83
1.47 ± 0.04
3.57 ± 0.59
1.21 ± 0.11
4.16 ± 0.64
10.10 ± 0.52
15.84 ± 2.01
0.23 ± 0.02
0.61 ± 0.17
0.07 ± 0.005
0.43 ± 0.12
5.16 ± 1.41
11.16 ± 0.92
Muscle
Heart
Lung
Kidney
Head
Intestines
Spleen
Brain
2
6
2
4
6
4
Urine
2
—
—
decoupled spectra were obtained on a VXR-200 or MERCURY
and (ii) diethyl 4-aza-N-benzylheptanedioate, (2.464 g, 79.2%),
1
3
00 or UNITY 400 MHz Varian spectrometers. H NMR chemical
shifts were referenced to TMS (0 ppm) using residual protio
solvent peaks as internal standards (singlets at d 7.24 for
). The mass spectrum of L was obtained on a VG 70-
1.22 (6H, t, CH
3
), 2.50 (4H, CH
2
CO), 2.85 (4H, t, BnN(CH
2
)
2
),
3.53 (2H, s, ArCH
C NMR (75.45 MHz, CDCl
2
), 4.14 (4H, q, OCH
2
–), 7.30 (5H, m, Ar–H);
1
3
H
3
), 14.08 (CH ), 32.71 (BnNCH
3
2
),
a
CDCl
3
49.17 (CH
2
CO), 58.23 (CH
2
O), 60.18 (Ar–CH N), 126.89, 128.09,
2
SEQ mass spectrometer. Micro-analyses were performed on an
Elemental Analyser EA 1108 CHNS-O (Fisons Instruments) and
are reported as an average of two combustions.
128.57 and 139.05 (Ar–C’s), 172.38 (C=O).
To 2.22 g (0.0072 mol) of diethyl 4-aza-N-benzylheptanedioate
3
was added 1.60 g (2 cm , 0.018 mol) of N,N-dimethyle-
thylenediamine in a cycloaddition tube. The reaction mixture
◦
was heated under reflux at 80 C for 24 h and the excess
Synthesis of 1,15-bis(N,N-dimethyl)-5,11-dioxo-8-(N-benzyl)-
a
N,N-dimethylethylenediamine was then removed on a vacuum
pump. The resulting residue was then subjected to column
chromatography with dichloromethane–methanol–triethylamine,
9 : 5 : 1 ratio (v/v) as the eluent to afford L (2.464 g, 0.0063 mol,
87% yield). L was dissolved in dichloromethane followed by
1,4,8,12,15-pentaazapentadecane (L )
3
2
.4 cm (3.38 g, 0.02 mol) of ethyl 3-bromopropionate was
3
a
added drop wise with continuous stirring to 1.02 cm (1.00 g,
a
0
.009 mol) of benzylamine in acetonitrile and diisopropy-
lethylamine for 30 min in a cycloaddition tube. The re-
action mixture was heated under reflux at 120 C, yield-
ing an oily ethyl-3-benzylaminopropionate and diethyl-4-aza-N-
filtration to remove impurities and the excess solvent was removed
on a rotary evaporator. Subsequent drying on a vacuum pump
◦
a
◦
1
furnished L in 90% yield, mp 198–200 C. H NMR (300 MHz,
CDCl ), 2.22 (12H, s, N(CH , 2.40 (4H, m, (CH NCH ),
2.76 (4H, m, (CH CO), 2.88 (4H, m, BnN(CH ), 3.25 (4H,
m, CONHCH ), 3.80 (2H, s, ArCH N), 7.29, (7H, m, ArH and
NH); C NMR (100.6 MHz, CDCl
36.68, 36.72 N(CH ), 45.07, 45.21 ((CH
(COCH ), 57.97, 58.00 (CONHCH ), 58.42, 60.18 (ArCH
benzylheptanedioate by TLC with R
f
values of 0.5 and 0.8, re-
3
3
)
2
3
)
2
2
spectively, in a 4 : 1 (v/v) hexane–ethyl acetate mixture. Following
removal of solvent on a rotary evaporator, the hydrobromide salt
of diisopropyl ethylamine was dissolved in a saturated solution
2
)
2 2
2
2
1
3
3
), 33.87, 35.87 (BnNCH
NCH
2
),
),
of Na
2
CO
3
and the organic components of the mixture were
3
)
2
3
)
2
2
), 49.91, 51.25
extracted with ethyl acetate. The combined organic layers were
dried over anhydrous magnesium sulfate, filtered and concentrated
on a rotary evaporator. Column chromatography using the same
2
2
2
127.10, 127.17, 128.16, 128.30, 128.40, 128.91, 138.46, 139.47 (Ar–
C’s for cis- and trans-forms), 172.11, 172.32 (C=O); FAB-MS gave
a
+
eluent as for TLC afforded: (i) ethyl 3-benzylaminopropionate,
m/z values, Found (Calculated, % Total abundance), [HL ] =
1
(
0.3055 g, 20.4%); H NMR (200 MHz, CDCl
3
) 1.25 (3H, t, CH
3
),
392.54259 for C21H N O (392.30255, 98.98%), 393.29959 for
38 5 2
1
5
13
1
3
.85 (1H, s, NH), 2.45 (2H, t, CH
2
CO), 2.85 (2H, t, CH
2
N),
C
21
H
38
N
4
NO
2
(393.29958, 1.42%), 393.30591 for C20 CH38
N
5
O
2
ꢀ
13
,79 (2H, s, Ar–CH ) 4.25 (2H, q, OCH
2
2
), 7.30 (5H, m, Ar–H s);
(393.30590, 18.19%), 394.30926 for C19
C
2
H
38
N
5
O (394.30926,
2
4
036 | Dalton Trans., 2006, 4029–4038
This journal is © The Royal Society of Chemistry 2006

View Article Online
a
+
a
2
2
.04%), and [NaL ] = 414.55 for NaC21
H
37
N
5
O
2
(414.28473,
of the [CuL H−1] complex were added prior to the addition
of xanthine oxidase. The absorbance readings at 560 nm were
obtained at 30-s intervals for a total period of 15 min for both
the control and experimental solutions. For each set of readings
corresponding to a control and particular concentrations of SOD
mimic, the absorbance was plotted against time and the slope
of the resulting plots were computed. The % inhibition of the
5.00%). Anal. Found (Calc. for C21
H, 8.68 (9.52); and N,15.81 (17.89).
H
37
N
5
O
2
): C, 63.20 (64.42);
Preparation of solutions
All solutions were prepared in distilled, deionised water which
had been boiled to remove dissolved CO and cooled in a
container protected by a “Carbosorb” (BDH) carbon dioxide
trap. These solutions were made up to an ionic strength of
.15 mol dm with sodium chloride to mimic blood-plasma
2
reduction of NBT to blue formazan was calculated and IC50
similarly determined from a plot of % inhibition versus
the concentration of the [CuL H−1] complex.
Molecular mechanics calculations were performed on a Sync-
Master 1100p computer running the Accelrys Biosym/MSI
software package. The copper complexes were constructed using
the Builder module and the geometry optimised using the esff
forcefield under Discover_3. All calculations were performed in
vacuum.
36–38
values
1,2
a
−
3
0
conditions. Reagents used were commercially available and of
analytical grade. These were used without further purification.
45
−
3
NaOH and HCl solutions (0.1 mol dm ) were prepared from
Merck Titrisol ampoules and stored in plastic bottles. The HCl
and carbonate free-NaOH solutions were standardised by titration
with potassium hydrogen phthalate (KHP) and sodium tetraborate
decahydrate (Borax), respectively, using the standard methods of
Vogel. The NaOH solutions were further standardised against
standard HCl solutions. These acid–base titrations were also used
to check the carbonate content of the NaOH titrant solution
using the Gran method. These standardised solutions were
used within a short period of time after preparation and were
discarded whenever there were signs of carbonate contamination.
The titration procedure has been described before where the
ligand was weighed directly into the titration vessel and the
64
Cu was prepared by neutron irradiation of 1.7–2.5 mg copper
II) oxide (99.9999%, Aldrich Chem. Co. Milw., WI) for 24 h at
(
41
14
−2 −1
a thermal neutron flux of 1.0 × 10 n cm
s
in the Safari-1
Research Reactor. The targets were dissolved in 1.0 ml of 1.0 M
HCl and diluted to 3.0 ml with 1.0 M HCl. Typical specific
42
−1
activities of ∼600 MBq mg target material were obtained.
Animal studies were approved by the Research Animals Ethics
Committee of the University of Cape Town (permission number
1
0
1/060) and performed in compliance with generally accepted
guidelines governing such work. The authority to possess and
use the radioactive nuclide Cu was granted by the university’s
radiation protection and health safety committee in conjunction
withthe Department of Health(authoritynumber 33/3/01/0327).
The procedure for labelling of injectates [CuL H−1] and CuCl
has been described previously. Biodistribution was measured
electrode system calibrated in situ. A value of 13.73 was used
64
for pK
the ESTA
w
. The potentiometric titration data were analysed using
43,44
suite of computer programs.
a
2
UV-visible spectroscopy
1,2
3
following intravenous injection via the tail vein of 4–6 week old,
female, Balb/C mice. Radio-TLC showed that the [CuL H−1] was
UV-visible electronic spectra of a 3.00 cm aqueous solution
a
a
containing a 1 : 3 ratio of Cu(II) : L , in a cuvette, were taken
over the pH range 2.00–11.00. Small amounts of 0.1 mol dm
NaOH and HCl were used to adjust the pH during the titra-
−
3
>90% radio-labelled.
6
4
Caution: Cu decays by positron emissions, which annihilate
to produce tissue penetrating high energy gamma radiation. It is
tion. The temperature of the solutions was kept at 25.0 ±
6
4
◦
advisable to use a low activity of Cu and perform all experiments
behind lead bricks inside a fume-cupboard.
0
.1 C. Spectrophotometric measurements were recorded in the
wavelength range 340–800 nm using a Unicam SP 1700 Ultraviolet
Spectrophotometer with slit and band widths set at 0.4 mm and
1
.2 nm, respectively.
Acknowledgements
Blood-plasma simulation, partition coefficients, superoxide
dismutase mimic activities, biodistribution studies and molecular
mechanics calculations
The authors would like to thank the National Research Founda-
tion (NRF) and the Councils of the Universities of Trankei and
Cape Town for financial support and a bursary to T. E. N. We are
also grateful to Dr Jan R. Zeevaart and Dr Werner K. A. Louw of
the South African Nuclear Energy Cooperation (Ltd.) at Pretoria
Blood-plasma simulation was performed using the ECCLES
8
program and the database of May et al. to which the formation
6
4
9
for supplying us with CuCl . The technical assistance with animal
2
constants measured in this study had been incorporated. Partition
◦
experiments from J. Boniaszczuk (Nuclear Medicine Department-
Groote-Schuur Hospital), J. Visser (Animal Unit), T. Jackson and
S. Odisitse is gratefully acknowledged.
coefficients were measured at 25 C and an ionic strength of
−
3
−
32,33
0
.15 mol dm (Cl ) using the shake flask method
as detailed
1
in a previous paper. The superoxide dismutase mimetic activity of
CuL H−1] was determined using the NBT assay as described by
Fridovich and Auclair and Voisin to assay tissue homogenates.
The procedure and concentrations of the various species have been
published before. Briefly, calculated amounts of xanthine, EDTA,
a
[
36
37
References
1
1 Part 6,T. E. Nomkoko, G. E. Jackson, B. S. Nakani, W. K. A. Louw
NBT, catalase were added to a 0.05 M phosphate buffer and this
constituted a blank solution. In a control experiment, in addition
to the above components, measured amounts of xanthine oxidase
and buffer were added, while in a test solution varying amounts
and J. R. Zeevaart, Dalton Trans., 2004, 741.
2
3
Part 7,E. T. Nomkoko, G. E. Jackson and B. S. Nakani, Dalton Trans.,
2
004, 1432.
R. W. Kinne, E. P. Kinne and F. Emmrich, Biochim. Biophys. Acta,
1997, 360, 109.
This journal is © The Royal Society of Chemistry 2006
Dalton Trans., 2006, 4029–4038 | 4037

View Article Online
4
5
Mechanisms and Models in Rheumatoid Arthritis, ed. B. Henderson,
J. C. W. Edwards and E. R. Pettipher, 1995, vol. 25, p. 67.
J. E. Weder, C. T. Dillon, T. W. Hambley, B. J. Kennedy, P. A. Lay, J. R.
Biffin, H. L. Regtop and N. M. Davies, Coord. Chem. Rev., 2002, 232,
21 V. Amendola, L. Fabbrizzi, C. Mangano, P. Pallavicini, A. Perotti and
A. Taglietti, J. Chem. Soc., Dalton Trans., 2000, 185.
22 R. D. Shannon, Acta Crystallogr., Sect. A, 1976, 32, 751.
23 R. G. Pearson, J. Am. Chem. Soc., 1985, 107, 6801.
24 A. E. Martell and R. D. Hancock, Metal Complexes in Aqueous
Solutions, Plenum Press, New York, 1996, p. 1.
25 A. B. P. Lever, Studies in physical and theoretical chemistry-Inorganic
Electronic Spectroscopy, Elsevier, Amsterdam, 2nd edn, 1984, p.
554.
26 E. J. Billo, Inorg. Nucl. Chem. Lett., 1974, 10, 613.
27 C. R. Hartzell and F. R. N. Gurd, J. Biol. Chem., 1969, 244, 147.
28 A. Jancso, I. Torok, L. Korecz, A. Rockenbauer and T. Gadja, J. Chem.
Soc., Dalton Trans., 2002, 2601.
9
5.
6
7
C. T. Dillon, T. W. Hambley, B. J. Kennedy, P. A. Lay, Q. Zhou, N. M.
Davies, J. R. Biffin and H. L. Regtop, Chem. Res. Toxicol., 2003, 16,
2
8.
G. E. Jackson, L. M. Gama, A. Voy e´ and M. J. Kelly, J. Inorg. Biochem.,
2
000, 79, 147; G. E. Jackson, P. W. Linder and A. Voye, J. Chem. Soc.,
Dalton Trans., 1996, 4605; G. E. Jackson and M. J. Kelly, J. Chem. Soc.,
Dalton Trans., 1990, 1889; G. E. Jackson and M. J. Kelly, J. Chem. Soc.,
Dalton Trans., 1989, 2429.
8
9
P. M. May, P. W. Linder and D. R. Williams, J. Chem. Soc., Dalton
Trans., 1977, 548.
29 E. A. Enyedy, H. Csoka, I. Lazar, G. Micera, E. Garriba and E. Farks,
J. Chem. Soc., Dalton Trans., 2002, 2632.
30 C. Jubert, A. Mohamadou, C. Gerard, S. Brandes, A. Tabard and J.-P.
Barbier, J. Chem. Soc., Dalton Trans., 2002, 2660.
31 B. S. Nakani and G. E. Jackson, J. Chem. Soc., Dalton Trans., 1996,
1373.
32 A. Leo, C. Hansch and D. Elkins, Chem. Rev., 1971, 71, 525.
33 T. X. Xiang and B. D. Anderson, J. Membr. Biol., 1994, 140, 111.
34 J. E. Shaw, M. Prevo, R. Gale, Su Yum, II, in Physiology, Biochemistry
and Molecular Biology of the Skin, Oxford University Press, Oxford,
2nd edn, 1991, ch. 55, p. 1447; A. S. Michaels, P. S. L. Wong, R. Prather
and R. M. Gale, AIChE J., 1975, 21, 1073.
G. E. Jackson, P. M. May and D. R. Williams, J. Inorg. Nucl. Chem.,
1
978, 40, 1227.
1
1
1
0 A. Signore, A. Annovazzi, M. Chianelli, F. Corsetti, C. Van de Wiele,
R. N. Watherhouse and F. Scopinaro, Eur. J. Nucl. Med., 2001, 28,
1
555.
1 D. H. Williams and I. Fleming, Spectroscopic Methods in Organic
Chemistry, McGraw-Hill Book Company (UK) Ltd., 2nd edn, 1973, p.
1
50.
2 P. Dumy, M. Keller, D. E. Ryan, B. Rohwedder, T. Wohr and M. Mutter,
J. Am. Chem. Soc., 1997, 119, 918; P. Li, X. G. Chen, E. Shulin and S. A.
Asher, J. Am. Chem. Soc., 1997, 119, 1116; V. P. Manea, K. J. Wilson
and J. R. Cable, J. Am. Chem. Soc., 1997, 119, 2033; C. S. Fischer and
G. Fischer, J. Am. Chem. Soc., 2001, 123, 6227.
3 R. J. Bergeron, J. S. McManis, O. Phanstiel, IV and J. R. Timothy-
Vinson, J. Org. Chem., 1995, 60, 109.
35 R. Cassimjee, F. Cassimjee, A. H. Cassimjee and Feroz Cassimjee,
S. Afr. Pat., ZA9204061, 1993.
36 I. Fridovich, In CRC Handbook of Molecular and Oxygen Radical
Research, ed. R. A. Greenwald, CRC Press, 1985, p. 53.
37 C. Auclair and E. Voisin, in CRC Handbook of Molecular and Oxygen
Radical Research, ed. R. A. Greenwald, CRC Press, 1985, p. 125.
38 N. A. Roberts and P. A. Robinson, Br. J. Rheumatol., 1985, 24, 128.
39 J. A. Roe, A. Butler, D. M. Scholler, J. S. Valentine, L. Marky and K. J.
Breslauer, Biochemistry, 1988, 27, 950.
40 E. T. Nomkoko, G. E. Jackson and B. S. Nakani, Inorg. Chem.
Commun., 2003, 6, 335.
41 A. I. Vogel, Vogel’s Textbook of Quantitative Inorganic Analysis,
Longman Group, London, 4th edn, 1978, 236, [Revised by: J. Bassett,
R. C. Denney, G. H. Jeffery and J. Mendham].
1
1
4 M. W. A. Steenland, I. Dierck, G. G. Herman, B. Devreese, W. Lippens,
J. Van Beeumen and A. M. Goeminne, J. Chem. Soc., Dalton Trans.,
1
997, 3879.
1
5 L. D. Pettit, IUPAC-Stability Constants Database Squery Version 1.11,
Academic Software, Sourby Old Farm, Timble, Otley, Yorks, LS21
2
PW, UK, 1993, (licenced to P. Linder and G. Jackson, UCT, South
Africa).
1
1
1
6 P. M. May, K. Murray and D. R. Williams, Talanta, 1985, 35, 825.
7 S. Kholeif and G. Anderegg, Inorg. Chim. Acta, 1998, 275–276, 37.
8 C. F. Baes, Jr. and R. E. Mesmer, The Hydrolysis of Cations, Wiley-
Interscience, New York, 1976, p. 268.
9 X. H. Bu, D. L. An, X. C. Cao, R. H. Zhang, T. Clifford and E. Kimura,
J. Chem. Soc., Dalton Trans., 1998, 2247.
0 M.-A. Santos, M. Gaspar and M. T. Amorim, Inorg. Chim. Acta, 1999,
42 G. Gran, Analyst, 1952, 77, 661.
43 K. Murray and P. M. May, ESTA (Equilibrium Simulation for Titration
Analysis) Manual, University of Wales Institute of Science and
Technology, Cardiff, 1984.
1
2
44 P. M. May, K. Murray and D. R. Williams, Talanta, 1988, 35, 927.
45 InsightII, User Manual, 2005 .
2
84, 20.
4
038 | Dalton Trans., 2006, 4029–4038
This journal is © The Royal Society of Chemistry 2006