Biologically active thiosemicarbazone Fe chelators and their reactions with ferrioxamine B and ferric EDTA; A kinetic study

Article abstract of DOI:10.1039/c1dt11685a

The Fe^III abstraction from Fe^III/DFO and Fe ^III/EDTA complex systems by thiosemicarbazone ligands derived from 2-acetylpyridine has been studied from a kinetico-mechanistic perspective at relevant pH conditions and at varying temperatures and buffer solutions. The reactions have been found to be extremely dependent on the dominant E/Z isomeric form of the TSC ligands present in the reaction medium. Consequently the isomerisation processes occurring on the free ligands have also been monitored under equivalent conditions. The isomerisation process is found to be acid dependent, despite the absence of protonation under the conditions used, and presumably proceeds via an azo-type tautomer of the ligand. In all cases the existence of outer-sphere interaction processes has been established, both promoting the reactions and producing dead-end complexes. The better oriented forms of the ligands (EZ thiolate) have been found to react faster with the [Fe(HDFO)]⁺ complex, although for mono-N⁴ substituted thiosemicarbazones the process is retarded by the formation of a dead-end outer-sphere complex. A comparison with the abstraction of Fe^III from [Fe(EDTA)(H₂O)]^- has also been conducted with significant differences in the kinetic features that implicate keystone outer-sphere interactions which dominate reactivity, even with isomeric forms that are not the best suited for direct complexation.

Full text of DOI:10.1039/c1dt11685a

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
Dalton
Transactions
Cite this: Dalton Trans., 2012, 41, 2122
www.rsc.org/dalton
PAPER
Biologically active thiosemicarbazone Fe chelators and their reactions with
ferrioxamine B and ferric EDTA; a kinetic study†
Paul V. Bernhardt,a Manuel Mart´ınez,*b Carlos Rodr´ıguezb and Marta Vazquezb
Received 6th September 2011, Accepted 7th November 2011
DOI: 10.1039/c1dt11685a
The FeIII abstraction from FeIII/DFO and FeIII/EDTA complex systems by thiosemicarbazone ligands
derived from 2-acetylpyridine has been studied from a kinetico-mechanistic perspective at relevant pH
conditions and at varying temperatures and buffer solutions. The reactions have been found to be
extremely dependent on the dominant E/Z isomeric form of the TSC ligands present in the reaction
medium. Consequently the isomerisation processes occurring on the free ligands have also been
monitored under equivalent conditions. The isomerisation process is found to be acid dependent,
despite the absence of protonation under the conditions used, and presumably proceeds via an azo-type
tautomer of the ligand. In all cases the existence of outer-sphere interaction processes has been
established, both promoting the reactions and producing dead-end complexes. The better oriented
forms of the ligands (EZ thiolate) have been found to react faster with the [Fe(HDFO)]+ complex,
although for mono-N4 substituted thiosemicarbazones the process is retarded by the formation of a
dead-end outer-sphere complex. A comparison with the abstraction of FeIII from [Fe(EDTA)(H2O)]-
has also been conducted with significant differences in the kinetic features that implicate keystone
outer-sphere interactions which dominate reactivity, even with isomeric forms that are not the best
suited for direct complexation.
chemistry (eqn (1)) as their Fe complex, leading to intracellular
oxidative stress.
Introduction
Heterocyclic thiosemicarbazones (TSC) are potentially tridentate
[FeIIL2] + H2O2 → [FeIIIL2]+ + OH∑ + OH-
(1)
chelating agents bearing both N and S donor centres including an
aromatic heterocycle, usually pyridine. Although TSC complexes
of many metal ions have been reported,1 their iron coordination
chemistry in particular is biologically important.2,3 In previous
papers it has been shown that TSC ligands derived from di-2-
pyridyl ketone,4 2-benzoylpyridine5 and 2-acetylpyridine6 exhibit
remarkable anti-proliferative activity against a variety of tumor
cells and also that they are well tolerated by healthy cells.
There has also been considerable interest in the related TSC
Triapine (3-amino-2-pyridinecarbaldehyde thiosemicarbazone)7,8
and this compound has progressed to the clinic as a potential
anticancer drug.9,10 Triapine inhibits ribonucleotide reductase (the
rate limiting process in DNA replication) and its action is linked
with the depletion of the iron within the cell and also to redox
activity, as the iron complex of Triapine is more active than the
free ligand.2,11,12 So it has emerged that iron is an important target
and that the ligands appear to act through a dual mechanism that
involves sequestering intracellular Fe and also catalysing Fenton
The complexity and importance of intracellular Fe chelation by
small molecules should not be underestimated from an inorganic
mechanistic point of view.13 In this respect some studies on Fe(III)
speciation between different biologically relevant molecules have
already been conducted.14,15 The Fe chelator desferrioxamine B
(DFO) has been the ‘gold standard’ in therapeutic Fe chelation
for the treatment of severe transfusion-linked Fe overload for
more than 30 years. The effectiveness of DFO as an Fe chelator
relies upon the high stability of its Fe(III) complex, ferrioxamine
B (Scheme 1), in vivo, so any biologically active chelator that
is capable of removing Fe from ferrioxamine B must have
sufficient thermodynamic stability to also be competitive for
labile intracellular Fe. Iron uptake from ferrioxamine B has been
observed by some of the thiosemicarbazone ligands discussed in
this paper.16–18
An often neglected issue in TSC coordination chemistry is
the possibility of E/Z isomerism about the C N double bond
of the Schiff base. No systematic study has been reported on
the effect of E/Z isomerisation on complexation behaviour in a
biologically relevant medium. In this paper we report the study of
the solution behaviour of a series of 2-acetylpyridine TSC ligands
(HApT, HAp4mT, HAp4pT and HAp44mT; Scheme 1), which
show significant changes in their isomeric distribution in different
aSchool of Chemistry and Molecular Biosciences, University of Queensland,
Brisbane, 4072, AUSTRALIA
bDepartament de Qu´ımica Inorga`nica, Universitat de Barcelona, Mart´ı i
Franque`s 1-11, E-08028, Barcelona, SPAIN
† Electronic supplementary information (ESI) available. See DOI:
10.1039/c1dt11685a
2122 | Dalton Trans., 2012, 41, 2122–2130
This journal is
The Royal Society of Chemistry 2012
©

View Article Online
Scheme 1
media. The reaction mechanism operating for isomerisation has
been studied and the actuation of proton-assisted outer-sphere
interactions has been established for the process. The existence
of this solvent-assisted recognition is indicative of what may be
occurring during the complexation processes in vivo.
isomeric distribution and interconversion at biologically relevant
pH values (where all ligands exist as neutral molecules).4 In all
cases the solutions were made up in water/DMSO mixtures (from
3.3 : 1 to 0.65 : 1 ratios depending on the ligand) to ensure total
solubility of the compounds. The stability and time evolution
of this E/Z isomeric distribution was also studied, this being
especially relevant due to the rather slow kinetics of iron removal
from ferrioxamine B and [Fe(EDTA)(H2O)]- by the TSC ligands
in Scheme 1 (see next section).16
The deprotonated EZ thiolate isomer is the one present in
most of the metal complexes of TSC ligands (see Scheme 1),25,26
especially those bearing a pyridine thiosemicarbazone tridentate
NNS donor set.6,27 Complexation is accompanied by deproto-
nation (pKa in the 9.5–11 range for the TSC free ligand).4,17
The same thiolate tautomeric form can also be present in the
free ligand, being stabilised in solution by hydrogen bonding
between the same set of donors.20 For N4-disubstituted TSC
ligands the E thione and Z thione tautomers are disfavoured
(there being no N4-H present for H-bonding e.g. HAp44mT), and
instead the unusual EZ thiolate zwitterionic tautomer (Scheme 2)
is found in TSC ligands derived from 2-acetylpyridine.6 For
ligands bearing at least one proton on N4 (e.g. HApT, HAp4mT,
HAp4pT) intramolecular H-bonding stabilises both the E thione
and Z thione forms; N1 ◊ ◊ ◊ HN3 for the Z form and N2 ◊ ◊ ◊ HN4
in all cases.6,19 The corresponding tautomers, ZE thiol and ZZ
thiol, are expected to be unimportant once the N1 ◊ ◊ ◊ HN3 H-
bonding is lost, despite the possible existence of a new N2 ◊ ◊ ◊ HS
interaction. Summarising, even though six different forms of the
The iron abstraction from ferrioxamine B (Scheme 1) has also
been studied kinetically with the same HAp4mT and HAp44mT
TSC ligands at different pH and temperatures. The results
indicate that the more favourably oriented isomer/tautomer of
each TSC ligand reacts faster with ferrioxamine B. Nevertheless,
a proton-assisted mechanism is also found for the process,
indicating the involvement of important outer-sphere recognition
reactions en route to the final product. Dead-end outer-sphere
complexes appear important for incorrectly oriented isomeric
forms of the partially terminal amine substituted HAp4mT
ligand. Finally the study has been extended to include reactions
with [Fe(EDTA)(H2O)]- (Scheme 1) with the most active TSC
HAp44mT ligand. Again, the existence of outer-sphere recogni-
tion reactions between the TSC ligands and the iron complex has
been established.
Results and discussion
Ligand isomeric distribution in solution
Given the large amount of literature data concerning the possible
isomers of tridentate heterocyclic TSC ligands in this study2,17,19–24
(see Scheme 2 for definitions) we have investigated the E/Z
This journal is
The Royal Society of Chemistry 2012
Dalton Trans., 2012, 41, 2122–2130 | 2123
©

View Article Online
Table 1 Isomeric and tautomeric distribution (%) as from Scheme 2 for the different forms possible for the thiosemicarbazone ligands indicated in
Scheme 1
Ligand
Solvent
E thione
EZ thiolate
Z thione
HApT
HAp4mT
DMSO(d6)
95
90
85
90
90
80
90
85
85
90
75
85
90
55
55
50
55
25
5
—
—
—
—
—
10
—
—
—
—
20
—
—
15
5
DMSO(d6)
10
15
10
10
10
10
15
15
10
5
15
10
30
40
40
40
20
D2O/DMSO(d6)
DMSO(d6)
HAp4pT
D2O/DMSO(d6)
after 4 h at 60 ◦
C
H2O/DMSO(d6), NaClO4 0.30 M
H2O/DMSO(d6), (Bu4P)Br 0.30 M
H2O/DMSO(d6), pHa = 5.4 (PIPPS)
H2O/DMSO(d6), pHa = 5.5 (MES)
after 24 h at 60 ◦
C
H2O/DMSO(d6), pHa = 6.5 (MES)
H2O/DMSO(d6), pHa = 7.5 (TRIS)
DMSO(d6)
HAp44mT
D2O/DMSO(d6)
after 24 h at 50 ◦
C
10
5
55
H2O/DMSO(d6), pHa = 5.5 (MES)
after 24 h at 60 ◦
C
a As measured in the buffer aqueous component of the solvent mixture.
the slow E ꢀ Z process should make these isomeric forms fully
distinguishable also.28–30
1H NMR of freshly prepared DMSO-d6 solutions of
HApT, HAp4mT and HAp4pT confirmed, as reported in the
literature,6,19,21,22 that the E thione isomer is dominant. The key
resonances for the HApT, HAp4mT and HAp4pT ligands are
found at 10.3, 10.3 and 10.7 ppm (N3H) and 8.1, 8.6 and 10.2 ppm
(N4H), respectively. Minor downfield signals were associated with
the presence of small amounts of EZ thiolate tautomers; 14.1, 14.2
and 14.6 ppm for the N2H proton for the same ligands, respectively
(Table 1). For the disubstituted HAp44mT ligand the E thione, EZ
thiolate and Z thione forms were all present (9.6 ppm and 15.4 ppm
for the N3H proton of the E thione and Z thione forms respectively,
and 14.7 ppm for the N2H proton of the EZ thiolate form).20 The
relative proportions are in fair agreement with the literature,6,19–21
but only partially support the H-bonding considerations discussed
above.
On dissolution of the ligands in aqueous DMSO (with different
DMSO contents, see above) only minor changes are seen in the
thione/thiolate tautomeric distribution (Table 1), as determined
from the associated 1H NMR signals of the aromatic (non-water
exchangeable) protons. The thione/thiolate ratios in Table 1 were
time-independent in this solvent. However, the relative amount of
thiolate tautomer increased slightly at more acidic pH values.17
The E/Z isomeric ratio of these solutions was found to be time
dependent. The amount of the Z form increases gradually to a
final thermodynamic ratio which is in much better agreement with
the distribution expected from the hydrogen bonding interactions
discussed above.17 Clearly the E/Z isomeric equilibration reaction
appears to be promoted by protons.
Scheme 2
TSC ligands studied could be present in solution, only some
of them are expected to be observable once thermodynamic
equilibrium conditions are attained. For the HApT, HAp4mT and
HAp4pT (each bearing at least one H-atom on N4), the preferred
forms should be E thione/EE thiol and Z thione/ZE thiol, while
for the fully substituted HAp44mT the EZ thiolate is expected to
be dominant. The rigidity caused by the H-bonding interactions
In view of the marked changes in the final equilibrium
distribution of the different forms occurring in aqueous acidic
solutions of these ligands (relative to pure DMSO), a detailed
kinetic study of the E ꢀ Z isomerisation process (kE→Z+kZ→E
)
was pursued at varying temperature, pressure, pH and ionic
strength conditions. Time-resolved UV-Vis spectroscopy of the
TSC ligands in water/DMSO indicates a neat single step process
makes the tautomeric forms distinguishable by H NMR,6 while
1
2124 | Dalton Trans., 2012, 41, 2122–2130
This journal is
The Royal Society of Chemistry 2012
©

View Article Online
Table 2 Relevant kinetic and activation parameters (pH = 5.5, 0.08 M MES) for the E ꢀZ isomerisation process observed for the HAp44mT, HAp4mT
and HAp4pT TSC ligands under different conditions; [TSC] = (1–3)¥10-5 M, I = 0.27–1.0 M NaClO4
Ligand
T/◦C
104¥kobs/s-1
DH‡/kJ mol-1
DS‡/J K-1mol-1
DV‡/cm3 mol-1
[H+]-dependence
HAp44mT
HAp4mT
HAp4pT
60
60
62
2.1a
78
not measured
122
2
-85
7
1.2 0.4
not measured
not measured
k = (61 4)¥[H+]
not measured
0.073a
0.093b
not measured
16 15
5
k = (0.8 0.1)¥[H+]
a 23% DMSO; b 62% DMSO.
Fig. 1 a) UV-Vis spectral changes obtained on a 2.5 ¥ 10-5 M aqueous (23% DMSO) solution of HAp44mT at 40 ◦C (pH = 5.5, MES); total time 24 h.
b) Plot of the rate constants observed for the same E ꢀ Z isomerisation process versus [H+] (T = 62 ◦C, I = 0.27 M (NaClO4), MES 0.08 M).
(Fig. 1a) that produces well-behaved first order absorbance/time
kinetic traces. Table 2 collects the relevant kinetic and activation
parameters for the process; the values for the observed rate
constants under the conditions used for the study are collected
in Table S1, ESI.†
nism has been recently proposed for hydrazone rotary switches
(Scheme 3).33
From the data in Table S1, ESI,† it is clear that the equilibration
rate constant (kE→Z+kZ→E) is independent, within experimental
error, of the ionic strength and the buffer concentrations used in
the study, although it is definitively accelerated by increasing [H+]
of the solution (Fig. 1b). The latter agrees with the fact that the
E ꢀ Z isomerisation process is only observed when slightly acidic
or polar solutions are used (Table 1),17 and effectively indicates
that the process must occur via a proton interaction with the
TSC in the transition state. The pH range investigated here is
well above the published pyridyl ring protonation constants of the
HApT series (ca. pKa 3.5),17 so no significant protonated TSC
will be present in the low pH region of this study. The proton
interaction should enable the above mentioned loss of double bond
Scheme 3
C
N2 character and produce an increased rotation reaction rate,
as observed in other E ꢀ Z or related isomerisation processes.28–31
The proton interaction in the transition state evidently should be
linked with any of the available neighbouring nitrogen donors. A
simple interaction with the pyridine N1 or thioamide N3 donors
would inductively stabilise a Cd-–Nd+ charge separation of the
double bond generated on a transition state;31,32 the alternative
direct interaction with the imine N2 donor would stabilise a Cd+–
Nd- charge separation in the same manner. Nevertheless, the most
plausible explanation for the process involves the acid catalysed
azo N2–N3 tautomerisation. The tautomerisation produces an
increase of the negative charge on the pyridine nitrogen (N1) that
can be easily stabilised with external protons; a similar mecha-
The fact that all the coordination compounds prepared from
these TSC compounds show the EZ thiolate tridentate arrange-
ment, despite the presence of other forms of the ligand in solution,
can thus be rationalised. Attachment of the metal ion to the most
accessible pyridyl N1 donor is likely as a prelude to eventual
tridentate coordination in a sort of zipper motion, as found
for other types of complexation processes;34–37 on doing so the
facile conversion of the isomeric mixture to the final preferred EZ
thiolate arrangement will be attained (Scheme 3).
Interestingly, the thermal and pressure activation parameters
for the isomerisation reaction indicated in Table 2 are very
different between the two pyridine thiosemicarbazone ligands
This journal is
The Royal Society of Chemistry 2012
Dalton Trans., 2012, 41, 2122–2130 | 2125
©

View Article Online
studied at variable temperature. For the HAp44mT ligand, the
thermal activation parameters are similar to those found for
E ꢀ Z isomerisation processes involving a loss of double bond
character by coordination and charge separation stabilisation via a
rotational transition state.29,30,32,33 Although the value determined
for DV‡ is rather surprising (expected to be negative in any of the
two highly polar solvents used, or in their mixtures), the reactivity
indicated in Scheme 3 can easily explain the opposite trends of
the entropy and volume of activation. The existence of a hydrogen
bonded solvent aggregation due to the presence of a protonated
nitrogen in the isomerising species has been consistently held
responsible for these opposite trends.38
For the HAp4pT ligand the process is ca. two orders of
magnitude slower (as the value for HAp4mT), and the value
for DH‡ is much more positive while that for DS‡ is practically
zero (DV‡ has not been determined for this very slow process).
The data clearly indicates some changes in the mechanistic
characteristics of the reaction, which is also shown by the ca.
two orders of magnitude difference in the [H+]-dependence of the
rate constant. Although an alternative direct inversion process,
as indicated for comprehensive studies on azo-derivatives,31,32,39,40
would agree with the measured thermal activation parameters for
the HAp4pT molecule, the [H+]-dependence of the rate constant
data in Table 2 does not. The partially substituted character of the
N4 centre should be related to the overall change in reactivity;
the very stable dominant initial E thione form of the ligand
encumbers the formation of the azo transition state indicated
in Scheme 3 for the E ꢀ Z isomerisation via N4–N2 hydrogen
bonding.
Despite their structural similarity, there was variation in
reactivity across the series of TSC ligands. The unsubstituted
ligand HApT showed no significant reactivity with [Fe(HDFO)]+
relative to HAp4pT, HAp4mT and HAp44mT on the same
time-scale. Nevertheless, for the HAp4pT ligand the spectral
changes indicated that, even in the slightly acidic pH range
used, reduction of [Fe(Ap4pT)2]+ takes place, which is only
prevented at higher acidities (pH < 5.5). This observation is
consistent with [Fe(Ap4pT)2]+ having the highest FeIII/II redox
potential of this series;6 no further studies were conducted on
the [Fe(HDFO)]+/HAp4pT system as a consequence of this
behaviour. For the HAp4mT and HAp44mT systems the reactivity
observed leads to final UV-Vis spectra that agree with those
expected from the characterized [FeIII(Ap4XT)2]+ complexes. The
bulk process observed by UV-Vis is complex and shows a kinetic
profile that depends on the chosen TSC ligand (Fig. 2).
No dependence of the spectral changes on the concentration of
TSC was found, indicating a non-equilibrium reaction under the
conditions of the study. In this respect the recent stability constants
evaluated for similar systems agree with the complete formation of
the [FeIII(Ap4XT)2]+ complexes under the conditions of the study.17
The data are found independent of ionic strength in the 0.1–
0.6 M range. Similarly no dependence on the concentration (0.01–
0.16 M) of the buffer was found. The overall transmetallation
reaction is faster than the isomerisation process (see above),
thus indicating that the reaction is occurring on the initial non-
equilibrated isomeric distribution of the TSC ligands used in the
study.
For both systems studied (>20-fold excess of ligand HAp4mT or
HAp44mT), the changes in the initial spectrum of [Fe(HDFO)]+ in
the visible region (where the ligand isomerisation processes do not
interfere) can be modelled by a set of two first order reactions (once
the induction period observed for the HAp4mT ligand system has
finished). The values of pseudo-first order rate constants derived
(kobs1 and kobs2, Table S1, ESI†) follow the same trends with respect
to the ligand concentration for the two systems. While the values
of kobs1 increase linearly with the concentration of the TSC with
no significant intercept, the values for kobs2 are independent of the
concentration of the ligand; Fig. 3a and Fig. S1a, ESI,† collect
these respective data for the HAp44mT and HAp4mT ligand
systems. From the slopes of the plots of kobs1 versus [TSC] a value
TSC iron uptake from ferrioxamine B
In view of preliminary observations that indicate that iron is
extracted from ferrioxamine B13,41 by similar TSC ligands,16 and
the fact that the process is so slow that equilibration studies are
not practical, the kinetics of the iron uptake from ferrioxamine B
by the ligands indicated in Scheme 1 was pursued at pH values
between 5.5 and 7.5. Lower values were considered biologically
irrelevant and higher (pH 8.0–9.0) values produced [FeII(ApXT)2]
complexes (presumably due to the ferric complexes oxidising water
being facilitated by alkaline pH).4,6,42
Fig. 2 a) Spectral changes upon the reaction of an aqueous solution of FeIII/DFO with HAp44mT; b) Absorbance versus time spectral changes obtained
from aqueous solutions of FeIII/DFO upon addition of the TSC ligands studied (note the different time-scales). [[Fe(HDFO)]+] = 5 ¥ 10-5 M, [TSC] =
5 ¥ 10-3 M, pH = 5.5, T = 60 ◦C. 23% DMSO in aqueous solution.
2126 | Dalton Trans., 2012, 41, 2122–2130
This journal is
The Royal Society of Chemistry 2012
©

View Article Online
Fig. 3 a) Dependence on the concentration of HAp44mT for the faster process observed on its reaction with [Fe(HDFO)]+ at 55 ◦C; b) pH profile of
the values derived from kfast and kslow for the same reaction. 23% DMSO in aqueous solution.
reactive in comparison with the major [Fe(HDFO)]+ component
of the system in this pH range; the alternative existence of an
Table 3 Kinetic data for the two processes observed upon reaction of
different TSC ligands with [Fe(HDFO)]+ at various pH values. The values
of kslow indicated correspond to the average for the [TSC] used. 23% DMSO
outer-sphere complex between the neutral TSC ligand and the
in aqueous solution
[Fe(HDFO)]+ species, very sensitive to [H+] as in the studies of the
Ligand
T/◦C
pH
103¥kfast/M-1s-1
105¥kslow/s-1
previous section, can also explain the data much more reasonably.
As for the values of kslow, they are smaller than those found
for the E ꢀ Z isomerisation process indicated in the previous
section (Table 2) (and this isomerisation cannot be observed at the
wavelengths shown in Fig. 2a), which definitively agrees with the
slow step observed being the final chelation of the TSC ligands.
Summarising, the substitution of DFO by the incoming TSC
ligand seems to be governed by the sequential entry, as found
for similar systems,45 and partial coordination (kfast) followed by
the complete tridentate coordination (kslow) of one of the TSC
ligands. After the first initial coordination, DFO dissociates from
the ternary FeIII/TSC/DFO complex and fast entry of a second
TSC ligand completes the reaction.
In this respect, the dramatic differences in the values of the
rate constants determined for the two studied ligands could
be tentatively attributed to the different amounts of the EZ
thiolate (Scheme 2) form present in solution for the two ligand
systems; this form being the reactive one for formation of the
final [Fe(TSC)2]+ complexes. Effectively, a definite decrease in the
reaction rate constant for the fast step (kobs1) is obtained for the
reaction of [Fe(HDFO)]+ with HAp44mT at pH 5.5 after isomeric
equilibration (according to the data collected in the previous
section), which is fully compensated once the change from a 40%
to a 20% EZ thiolate form of the ligand is considered (Table 1,
Fig. S2, ESI†). Nevertheless, this can only explain a factor of
three in the rate constant increase from HAp4mT to HAp44mT,
and by no means the induction period observed for the spectral
changes on HAp4mT transmetallation (Fig. 2b, 2.5 h at 60 ◦C
and pH = 5.5). Given the fact that the basicities of the pyridine
rings in HAp4mT and HAp44mT are similar, and should not lead
to any significant differences in reactivity according to the facts
indicated in the previous section,46,47 the degree of substitution on
the N4 atom, has to be considered as responsible. The fast initial
formation of a very favourable outer-sphere aggregate of HAp4mT
with the [Fe(HDFO)]+ complex, that does not evolve directly to the
substituted complex (i.e. a dead-end species), can easily explain the
HAp44mT
55
5.5
6.5
7.4
7.5
5.5
5.5
6.0
6.5
7.5
240
73
44
37
580 50
26
16
10
9
4.0 0.5
3.5 0.5
2.4 0.1
2.4 0.1
5.5 0.5
0.030a
2
3
5
60
70
HAp4mT
4
1
1
not determined
0.060a
5.0 0.1
0.050a
a Large errors are involved for such a slow process.
of kfast can be derived, which shows a decrease upon increasing
the pH value of the medium, as indicated in Fig. 3b and Fig. S1b,
ESI.† The values of kobs2 = kslow are seen to be independent of pH
in the same plots. Table 3 summarises the data collected for these
systems as a function of the different variables used.
It is clear that while the value of kfast should correspond to a
coordination of the ligand to FeIII ion, the process associated with
kslow does not correspond to a ligand concentration dependent
entry of a TSC to the coordination sphere. An intramolecular
process involving the final thiosemicarbazone chelation appears
responsible for this second slower step; the alternative DFO final
dissociation would be expected to show an increasing rate on
acidity. The two-fold increase of kobs1 obtained on increasing the
DMSO content (not observed for kobs2), agrees with the formation
of a relevant {[Fe(HDFO)]+;TSC0} outer-sphere complex prior to
the initial substitution step, favoured in less polar medium.
The pH profile (Fig. 3b and Fig. S1b, ESI†) for the values of kfast
indicate that a significant acceleration of the process is observed on
increasing the acidity of the medium. This is in good agreement
with what had been observed for the exchange process between
[Fe(HDFO)]+ and ferrichrome A,14 and was associated with a
partial protonation of the FeIII/DFO system (despite a pKa value
for [Fe(HDFO)]+ of ca. 1.0),43 thus labilising the iron centre.44
If this is so, the [Fe(H2DFO)]2+ species should be extremely
This journal is
The Royal Society of Chemistry 2012
Dalton Trans., 2012, 41, 2122–2130 | 2127
©

View Article Online
Table 4 Relevant kinetic (average for 3–5 runs at different TSC concentration at 60 ◦C) and thermal activation data for the processes observed upon
reaction of HAp44mT with [Fe(EDTA)(H2O)]-. 40% DMSO in aqueous solution
pH
104¥333kobs1/s-1
104¥333kobs2/s-1
DH‡/kJ mol-1
DS‡/J K-1mol-1
105¥333kobs3/s-1
DH‡/kJ mol-1
DS‡/J K-1mol-1
5.5
6.5
2.1 (2.1)a
1.6
1.8
48
5
-176 15
not determined
1.7
1.2
79
1
-101
1
0.18 (0.18)a
not determined
not determined
not determined
a In brackets are the values of the E ꢀ Z isomerisation process indicated in Table 2.
induction period observed. The decomposition of this aggregate
to produce the final [Fe(Ap4mT)2]+ would be consequently slower
than for the species directly involving the N1 donor in iron capture.
In this respect, the fact that the induction period is observed to
increase with acidity seems to indicate that an involvement of a
solvent- and proton-assisted outer-sphere complexation is present
in this reactivity pattern.
pH, and have spectral changes which correspond to the formation
of the final substituted complex. The relevant kinetic and thermal
activation parameters collected in Table 4 clearly indicate that
these two (kobs2 and kobs3) processes monitored correspond to an
associative process, which is in good agreement with a simple
chelation sequence from a FeIII/EDTA/TSC ternary system to
produce the final [Fe(Ap44mT)2]+ complex. Furthermore, the
simple substitution processes on the [Fe(EDTA)(H2O)]- complex
has been established as having a well defined dissociative activation
character.48,49
TSC iron uptake from FeIII/EDTA
In view of the data determined in the previous section and
the historical importance of the EDTA competition studies to
assess the quality of a chelator system,41 the kinetics of iron
removal from ferric EDTA was pursued. The reactions were
studied with the HAp44mT ligand at pH values of 5.5 and
6.5, where the precursor complex exists as a heptacoordinate
[Fe(EDTA)(H2O)]- pentagonal bipyramid (Scheme 1)48,49 and the
TSC ligands exist dominantly as neutral molecules.4 Under these
conditions no FeII/TSC signature at 640 nm was observed,4,6,42
which confirms the fact the TSC ligand removes FeIII directly
from the EDTA complex without any redox chemistry; the final
UV-Vis spectra showed in all cases the characteristics of the
[Fe(Ap44mT)2]+ complex. The time resolved spectral changes
are very complex, nevertheless, the fitting of the full spectral
changes in the 400–700 nm range using SPECFIT50 produces
three identifiable reaction steps with well defined spectra and
concentration profiles (Fig. 4). One of the reaction rate constants,
as well as the spectral changes associated, corresponds within
error limits to the previously determined E ꢀ Z isomerisation
process, kobs from Table 2, while the other two, kobs2 and kobs3, do
not show any dependence on the concentration of the ligand or
From the data it is clear that the mechanism involved in the
[Fe(EDTA)(H2O)]- → [Fe(Ap44mT)2]+ conversion process is very
different from that observed for the FeIII/DFO system. While for
the substitution of [Fe(HDFO)]+ the reaction clearly involves the
EZ thiolate form of the HAp44mT ligand, this is not the case for
the [Fe(EDTA)(H2O)]- substitution process where an initial build-
up (Fig. 4) of the Z thione isomer is observed before the reaction
occurs. Experiments similar to those conducted in the previous
section were run on E ꢀ Z equilibrated HAp44mT solutions at pH
5.5 and 6.5 and results confirmed that the reaction sequence for the
HAp44mT iron complexation from [Fe(EDTA)(H2O)]- is much
more complex than expected. At pH 6.5 only two consecutive first-
order kinetic steps are observed, as expected from a build up of
the Z thione isomer of HAp44mT. Furthermore, the rate constants
agree reasonably well with those measured for kobs2 and kobs3 on the
initial non-equilibrated reaction mixture. At pH 5.5, nevertheless,
the results are more surprising, but highly revealing. After a
fast reaction (occurring on conventional thermostating time), the
kinetic profiles are equivalent to those measured in standard
experiments. The nature of the fast reaction involves spectral
changes that are the reverse of the observed E ꢀ Z isomerisation
reaction; the UV-Vis spectrum of the reaction mixture immediately
after the complex [Fe(EDTA)(H2O)]- is added agrees with that
measured for the initial non-equilibrated HAp44mT solution.
It seems thus that, under these pH conditions, the reaction
of [Fe(EDTA)(H2O)]- with the Z thione form present in the
equilibrated reaction mixture of the HAp44mT ligand (see Table
1) competes unfavourably with a catalysed back isomerisation
to the initial E thione and EZ thiolate tautomer mixture of
the ligand. This catalytic effect has to be clearly related to the
formation of an acid-assisted outer-sphere complex between the
iron [Fe(EDTA)(H2O)]- complex and the Z thione form of the
TSC ligand, in line with the data reported in the first part of this
section. Similar experiments carried out at different pH from 4.5
to 6.5 and at room temperature indicated that this reversion to the
initial mixture of the E thione and EZ thiolate forms is effectively
faster at higher acidities (from 2–3 h to 24 h at pH 4.5 and 6.5
Fig.
4 UV-Vis spectral changes observed for the reactions of
[Fe(EDTA)(H2O)]- with HAp44mT. Inset: Kinetically fitted concentration
profile of the different species for the same reaction. [HAp44mT] = 6 ¥ 10-4
M, [Fe(EDTA)(H2O)]- = 2.2 ¥ 10-5 M, pH = 5.5, T = 70 ◦C, 40% DMSO
in aqueous solution.
◦
respectively at 20 C). Thus, the different behaviour observed at
pH 5.5 and 6.5 for the iron abstraction from the FeIII/EDTA
system by thermodynamically equilibrated isomeric mixtures of
2128 | Dalton Trans., 2012, 41, 2122–2130
This journal is
The Royal Society of Chemistry 2012
©

View Article Online
HAp44mT is related to the fact that, at pH = 5.5, the catalysed
reversion to the E thione and EZ thiolate forms of the ligand
competes favourably with the substitution of EDTA by the Z
thione form of the HAp44mT ligand.
electrode. UV-Vis spectra were recorded on Cary50, HP8453 or
J&M TIDAS instruments.
Kinetics
All the solutions to be used for monitoring the kinetic processes
were prepared by adding the calculated amounts of the different
ligands dissolved in water or DMSO, according to their solubility
in the final media and ionic strength, to stock buffered aqueous
solutions. In all cases the maximum volume percentage of DMSO
was 65%. The final ionic strength was set normally at 0.27 M with
NaClO4, although changes from 0.15 to 1.0 M do not produce any
significant difference in the processes observed.
All the kinetic measurements for the different systems carried
out at ambient pressure were followed by UV-Vis spectroscopy
in the full 300–800 nm range inside the thermostatic multicell
compartments of a Cary50 or HP8453 instrument. For the
monitoring of reactions at variable pressure, an already described
pressurising setup51–53 connected to an J&M TIDAS instrument
has been used. For the isomerisation studies the values of the
concentrations of the thiosemicarbazone ligands were set within
the (1–3)¥10-5 M range, while for all the substitution systems the
limiting [FeIII] values were within the (1–5)¥10-5 M range, unless
stated, and the uptaking ligands in excess were in at least 10 or
20-fold flooding factor (depending on the stoichiometry of the
final complex). The general kinetic technique has already been
described.54–56 The time-dependence of the spectral changes were
fitted to single (kobs1) or double (kobs1 plus kobs2) or triple (kobs1 plus
Conclusions
The freshly prepared thiosemicarbazones of the HApT family
used in this study show a dominant presence of the isomeric
E thione and EZ thiolate tautomers that equilibrates in protic
medium to a mixture with increased presence of the Z thione
form. The relative thione/thiolate tautomeric ratios are dependent
on the degree of substitution on the N4 terminal amino group.
The E ꢀ Z isomerisation reaction has been kinetically studied at
variable temperatures and pressures and indicates an acid-assisted
isomerisation mechanism.
The kinetic studies carried out for the iron abstraction from the
FeIII/DFO system with the fully substituted HAp44mT ligand
indicate a dominant activity of its EZ thiolate form (the one
matching the conformation of the final bis-thiosemicarbazone
complexes), which is acid-assisted via the formation of an outer-
sphere complex. When the mono-N4-methylated ligand HAp4mT
is used instead, the above mentioned outer-sphere complexation
leading to a dead-end complex implies a definite induction
period for the formation of the final [Fe(Ap4mT)2]+ complex with
consequently longer reaction times.
Unexpectedly, the HAp44mT iron uptake from the FeIII/EDTA
system is much more complex than that observed for the
FeIII/DFO. In this case the substitution-active TSC ligand isomer
is found to be the Z thione form; the importance of outer-sphere
hydrogen bonding in these reactions is a key point, as in the
previous cases. The [Fe(EDTA)(H2O)]- complex is seen to catalyse
the recovery of the E thione/EZ thiolate non-equilibrated mixture
after a previously conducted thermodynamic equilibration.
k
obs2 plus kobs3) exponential dependence by the use of the SPECFIT
software.50 For the simpler processes, fitting was conducted on a
single wavelength basis in the range where no interference of other
secondary reactions was observed; in these cases no dependence of
the value of the observed rate constants with the wavelength was
observed. Nevertheless, for systems where the process was more
complex, the fitting of the full spectral changes was also conducted
with good results. Table S1, ESI† collects all the values of kobs
determined for the different systems as a function of temperature
and pH. All post run fittings were carried out by least-squares
methods using standard commercial software. The typical error
limits for the values of kobs are within 15% of the value indicated
in Table S1, ESI.†
Experimental
Compounds
All the thiosemicarbazone ligands used in this study were prepared
according to literature methods.4 [Fe(EDTA)(H2O)]- is commer-
cially available and was used without any further treatment;
[Fe(HDFO)]+ (ferrioxamine B) was also prepared according to
the literature.14
The buffer solutions used in this study were selected according
to their inability to act as competing ligands for the FeIII complexes
in this study as well as their pKa values. MES, PIPPS, TRIS, CAPS
have been used in all cases at concentrations that were capable of
buffering all the relevant species in the medium. The final pH was
adjusted by adding perchloric acid or sodium hydroxide to the
final solutions. Their final ionic strength was always adjusted to
0.25–0.30 M with NaClO4 unless stated.
Acknowledgements
We acknowledge financial support from the Ministerio de Ciencia
e Innovacio´n, grants PCI-2006-a7-0530 and CTQ2009-14443-
C02-02, and the Australian Research Council, grants DP1096029
and LX0776107.
References
1 D. X. West, A. E. Liberta, S. Padhye, R. C. Chikate, P. Sonawane, A.
Kumbhar and R. G. Yerande, Coord. Chem. Rev., 1993, 123, 49–71.
2 Y. Yu, D. S. Kalinowski, Z. Kovacevic, A. R. Siafakas, P. J. Jansson,
C. Stefani, D. B. Lovejoy, P. C. Sharpe, P. V. Bernhardt and D. R.
Richardson, J. Med. Chem., 2009, 52, 5271–5294.
3 Z. Kovacevic, D. S. Kalinowski, D. B. Lovejoy, Y. Yu, Y. S. Rahmanto,
P. C. Sharpe, P. V. Bernhardt and D. R. Richardson, Curr. Top. Med.
Chem., 2011, 11, 483–499.
Instruments
1H NMR spectra were recorded on a Varian VNMRS 400 MHz
instrument at the Serveis Cient´ıfico-Te`cnics de la Universitat de
Barcelona. pH values were measured on a Crison pH & ION-
Meter GLP 22+ instrument using a standard glass combined
4 D. R. Richardson, P. C. Sharpe, D. B. Lovejoy, D. Senaratne, D. S.
Kalinowski, M. Islam and P. V. Bernhardt, J. Med. Chem., 2006, 49,
6510–6521.
This journal is
The Royal Society of Chemistry 2012
Dalton Trans., 2012, 41, 2122–2130 | 2129
©

View Article Online
5 D. S. Kalinowski, P. C. Sharpe, P. V. Bernhardt and D. R. Richardson,
J. Med. Chem., 2007, 50, 6212–6225.
6 D. R. Richardson, D. S. Kalinowski, V. Richardson, P. C. Sharpe, D.
B. Lovejoy, M. Islam and P. V. Bernhardt, J. Med. Chem., 2009, 52,
1459–1470.
7 R. A. Finch, M. C. Liu, A. H. Cory, J. G. Cory and A. C. Sartorelli,
Adv. Enzyme Regul., 1999, 39, 3–12.
8 R. A. Finch, M. Liu, S. P. Grill, W. C. Rose, R. Loomis, K. M. Vasquez,
Y. Cheng and A. C. Sartorelli, Biochem. Pharmacol., 2000, 59, 983–
991.
30 I. Favier, M. Go´mez, J. Granell, M. Mart´ınez, M. Font-Bard´ıa and X.
Solans, Dalton Trans., 2005, 123–132.
31 J. Garcia-Amoro´s, M. Mart´ınez, H. Finkelman and D. Velasco, J. Phys.
Chem. A, 2010, 114, 1287–1293.
32 T. Asano and T. Okada, J. Org. Chem., 1986, 51, 4454–4458.
33 S. M. Landge, E. Tkatchouk, D. Ben´ıtez, D. A. Lanfranchi, M. Elhabiri,
W. A. Goddard and I. Aprahamian, J. Am. Chem. Soc., 2011, 133,
9812–9823.
34 J. Esteban, P. Hirva, P. Lahuerta and M. Mart´ınez, Inorg. Chem., 2006,
45, 8776–8784.
9 J. J. Knox, S. J. Hotte, C. Kollmannsberger, E. Winquist, B. Fisher and
E. A. Eisenhauer, Invest. New Drugs, 2007, 25, 471–477.
10 B. Ma, B. C. Goh, E. H. Tan, K. C. Lam, R. Soo, S. S. Leong, L. Z.
Wang, F. Mo, A. T. Chan, B. Zee and T. Mok, Invest. New Drugs, 2008,
26, 169–173.
11 C. E. Cooper, G. R. Lynagh, K. P. Hoyes, R. C. Hider, R. Cammack
and J. B. Porter, J. Biol. Chem., 1996, 271, 20291–20299.
12 D. S. Kalinowski and D. R. Richardson, Pharmacol. Rev., 2005, 57,
547–583.
35 A. G. Algarra, M. G. Basallote, C. E. Castillo, M. P. Clares, A. Ferrer,
E. Garc´ıa-Espan˜a, J. M. Llinares, M. A. Ma´n˜ez and C. Soriano, Inorg.
Chem., 2009, 48, 902–914 .
36 A. G. Algarra, M. G. Basallote, R. Belda, S. Blasco, C. E. Castillo, J.
M. Llinares, E. Garc´ıa-Espan˜a, L. Gil, M. A. Ma´n˜ez, C. Soriano and
B. Verdejo, Eur. J. Inorg. Chem., 2008, 62–75.
37 C. E. Castillo, M. A. Ma´n˜ez, J. Gonza´lez, J. M. Llinares, H. R. Jime´nez,
M. G. Basallote and E. Garc´ıa-Espan˜a, Chem. Commun., 2010, 46,
6081–6083.
13 A. L. Crumbliss and J. M. Harrington, Adv. Inorg. Chem., 2009, 61,
38 G. Aullo´n, P. V. Bernhardt, F. Bozoglia´n, M. Font-Bard´ıa, B. P.
Macpherson, M. Mart´ınez, C. Rodr´ıguez and X. Solans, Inorg. Chem.,
2006, 45, 8551–8562.
179–250.
14 T. P. Tufano and K. N. Raymond, J. Am. Chem. Soc., 1981, 103, 6617–
6624.
39 R. van Eldik, T. Asano and W. J. Le Noble, Chem. Rev., 1989, 89,
15 I. Turcot, A. Stintzi, J. Xu and K. N. Raymond, JBIC, J. Biol. Inorg.
549–688.
Chem., 2000, 5, 634–641.
40 H. Yamataka, S. C. Ammal, T. Asano and Y. Ohga, Bull. Chem. Soc.
Jpn., 2005, 78, 1851–1855.
41 K. Matsumoto, T. Ozawa, K. Jitsukawa and H. Masuda, Inorg. Chem.,
2004, 43, 8538–8546.
42 E. A. Enyedy, N. V. Nagy, E. Zsigo, C. R. Kowol, V. B. Arion, B. K.
Keppler and T. Kiss, Eur. J. Inorg. Chem., 2010, 1717–1728.
43 G. Schwarzenbach and K. Schwarzenbach, Helv. Chim. Acta, 1963, 46,
1390–1400.
16 P. V. Bernhardt, P. C. Sharpe, M. Islam, D. B. Lovejoy, D. S. Kalinowski
and D. R. Richardson, J. Med. Chem., 2009, 52, 407–415.
17 E. A. Enyedy, M. F. Primik, C. R. Kowol, V. B. Arion, T. Kiss and B.
K. Keppler, Dalton Trans., 2011, 40, 5895–5905.
18 G. Serratrice, F. Biaso, J. L. Pierre, S. Blanc and A. M. Albrecht-Gary,
Eur. J. Inorg. Chem., 2007, 3681–3685.
19 E. Bermejo, A. Castin˜eiras, R. Dom´ınguez, R. Carballo, C. Maichle-
Mo¨ssmer, J. Stra¨hle and D. X. West, Z. Anorg. Allg. Chem., 1999, 625,
961–968.
44 B. Monzyk and A. L. Crumbliss, J. Am. Chem. Soc., 1982, 104, 4921–
4929.
20 C. R. Kowol, R. Eichinger, M. A. Jakupec, M. Galanski, V. B. Arion
and B. K. Keppler, J. Inorg. Biochem., 2007, 101, 1946–1957.
21 E. Bermejo, R. Carballo, A. Castin˜eiras, R. Dom´ınguez, A. E. Liberta,
C. Maichle-Mo¨ssmer, M. M. Salberg and D. X. West, Eur. J. Inorg.
Chem., 1999, 965–973.
22 A. Kumbhar, P. Sonawane, S. Padhye, D. X. West and R. Butcher, J.
Chem. Crystallogr., 1997, 27, 533–539.
45 B. Monzyk and A. L. Crumbliss, J. Inorg. Biochem., 1983, 19, 19–39.
46 S. Blasco, B. Verdejo, M. P. Clares, C. E. Castillo, A. G. Algarra, J.
Latorre, M. A. Ma´n˜ez, M. G. Basallote, C. Soriano and E. Garc´ıa-
Espan˜a, Inorg. Chem., 2010, 49, 7016–7027.
47 B. Verdejo, A. Ferrer, S. Blasco, C. E. Castillo, J. Gonza´lez, J. Latorre,
M. A. Ma´n˜ez, M. G. Basallote, C. Soriano and E. Garc´ıa-Espan˜a,
Inorg. Chem., 2007, 46, 5707–5719.
23 K. Nomiya, K. Sekino, M. Ishikawa, A. Honda, M. Yokoyama, N. C.
Kasuga, H. Yokoyama, S. Nakano and K. Onodera, J. Inorg. Biochem.,
2004, 98, 601–615.
24 D. X. West, G. A. Bain, R. J. Butcher, J. P. Jasinski, Y. Li, R. Y.
Pozdniakiv, J. Valde´s-Mart´ınez, R. A. Toscano and S. Herna´ndez-
Ortega, Polyhedron, 1996, 15, 665–674.
25 R. Pedrido, M. J. Romero, A. M. Gonza´lez-Noya, M. R. Bermejo,
M. Mart´ınez-Calvo and G. Zaragoza, Inorg. Chem., 2009, 48, 10862–
10864.
26 I. Garc´ıa Santos, U. Abram, R. Alberto, E. Va´zquez Lo´pez and A.
Sa´nchez, Inorg. Chem., 2004, 43, 1834–1836.
27 E. A. Enyedy, N. V. Nagy, E. Zsigo´, C. R. Kowol, V. B. Arion, B. K.
Keppler and T. Kiss, Eur. J. Inorg. Chem., 2010, 1717–1728.
28 K. Baba, H. Ono, E. Itoh, S. Itoh, K. Noda, T. Usui, K. Ishihara, M.
Inamo, H. D. Takagi and T. Asano, Chem.–Eur. J., 2006, 12, 5328–5333.
29 M. Crespo, M. Font-Bard´ıa, J. Granell, M. Mart´ınez and X. Solans,
Dalton Trans., 2003, 3763–3769.
48 M. Dellert-Ritter and R. van Eldik, J. Chem. Soc., Dalton Trans., 1992,
1037–1044.
49 A. Brausam and R. van Eldik, Inorg. Chem., 2004, 43, 5351–5359.
50 R. A. Binstead, A. D. Zuberbuhler and B. Jung, SPECFIT32. [3.0.34],
2005, Spectrum Software Associates.
51 R. van Eldik, in Inorganic High Pressure Chemistry, ed. R. van Eldik,
Elsevier, 1986, Chapter 1, pp. 1–68.
52 B. P. Macpherson, B. M. Alzoubi, P. V. Bernhardt, M. Mart´ınez, P.
Tregloan and R. van Eldik, Dalton Trans., 2005, 1459–1467.
53 F. Estevan, G. Gonza´lez, P. Lahuerta, M. Mart´ınez, E. Peris and R.
van Eldik, J. Chem. Soc., Dalton Trans., 1996, 1045–1050.
54 P. V. Bernhardt, M. Mart´ınez, C. Rodr´ıguez and M. Vazquez, Inorg.
Chem., 2011, 50, 1429–1440.
55 C. A. Bell, P. V. Bernhardt, L. R. Gahan, M. Mart´ınez, M. J. Monteiro,
C. Rodr´ıguez and C. A. Sharrad, Chem.–Eur. J., 2010, 16, 3166–3175.
56 P. V. Bernhardt, C. Gallego and M. Mart´ınez, Organometallics, 2000,
19, 4862–4869.
2130 | Dalton Trans., 2012, 41, 2122–2130
This journal is
The Royal Society of Chemistry 2012
©