Complex formation of ICL670 and related ligands with FeIII and FeII

FULL PAPER

Complex Formation of ICL670 and Related Ligands with Fe^IIIand Fe^II

Stefan Steinhauser,^[a]Uwe Heinz,^[a]Mark Bartholomä,^[a]Thomas Weyhermüller,^[b]

Hanspeter Nick,^[c]and Kaspar Hegetschweiler*^[a]

Keywords: Iron / Ligand design / Chelates / N,O ligands / Redox chemistry

Complex formation of 4-[3,5-bis(2-hydroxyphenyl)-1,2,4-tria-

zol-1-yl]benzoic acid (ICL670, H₃L^x), 4-[3,5-bis(2-hydroxy- and the [Fe^III(L^x)₂]³⁻/[Fe^II(L^x)₂]⁴⁻couples differs significantly.

consequence, the redox chemistry of the [Fe^III(L^x)]/[Fe^II(L^x)]⁻

phenyl)-1,2,4-triazol-1-yl]benzosulfonic acid (H₃L^y), and 3,5-

bis(2-hydroxyphenyl)-1-phenyl-1,2,4-triazole (H₂L^z) with

Fe³⁺and Fe²⁺was investigated in H₂O and in H₂O/DMSO

mixtures by potentiometry, spectrophotometry and cyclic vol-

tammetry. ICL670 has previously been considered as a prom-

ising drug for an oral treatment of iron overload. In this pa-

per, the stability and redox properties of the various Fe^IIand

Fe^IIIcomplexes were elucidated with a particular focus on

their potential involvement in the generation of oxidative

stress. The overall stability constants of [Fe^III(L^x)] and

[Fe^III(L^x)₂]³⁻is a very weak oxidizing agent (E_1/2is approxim-

ately −0.6 V versus NHE) and reduction of [Fe^III(L^x)₂]³⁻is not

anticipated under physiological conditions. The reduction

potential of the [Fe^III(L^x)]/[Fe^II(L^x)]⁻couple is considerably

less negative and was estimated to be +0.1 V (versus NHE).

The possible roles of the various Fe complexes as catalysts

for the Fenton reaction in biological media are discussed. The

crystal structures of H₃L^x, Na[Fe(L^z)₂]·4EtOH, Na[Al(L^z)₂]·

4EtOH, and [Cu(L^z)(pyridine)]₂were investigated by single-

crystal X-ray diffraction, and the possible influence of the

particular steric requirements of these ligands on the stability

of the metal complexes has been analyzed.

[Fe^III(L^x)₂]³⁻(25 °C, 0.1

M KCl in H₂O) are log β₁= 22.0 and

log β₂= 36.9, respectively. The affinity of these ligands for

Fe²⁺is remarkably poor. In particular, the 1:2 complexes

( Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,

[Fe^II(L^x)₂]⁴⁻and [Fe^II(L^y)₂]⁴⁻were found to be less stable. As a Germany, 2004)

1. Introduction

ally; however, its application does suffer from some side ef-

fects.^[3]The compound 4-[3,5-bis(2-hydroxyphenyl)-1,2,4-

triazol-1-yl]benzoic acid (ICL670, deferasirox) is also an

orally active iron chelator.^[4]In previous animal studies, it

proved to be more effective than parenteral DFO,^[5]ICL670

is presently in Phase III of clinical trials.

Iron, an essential bioelement, is toxic when present in

excess.^[1]Iron overload is a particularly grave condition, be-

cause the human body lacks a physiological mechanism for

actively excreting iron. There are a variety of different

causes for the excessive loading of iron in a person; an im-

portant one is regular blood transfusions, necessary in the

treatment of hematological defects such as β-Thalassaemia.

Consequently, transfusion therapy must be combined with

the administration of an iron-binding agent, transforming

excess body iron into a chelated, excretable form. Presently,

the microbial siderophore desferrioxamine (desferal, DFO)

is used for this purpose (Scheme 1). This ligand, however,

has some drawbacks. DFO is orally inactive and has a

rather short plasma half-life. The need for new orally active

iron chelators has been recognized for more than 30 years.^[2]

Deferiprone (L1), a hydroxy-pyridinone, can be taken or-

[a]

Anorganische Chemie, Universität des Saarlandes,

Postfach 15 11 50, 66041 Saarbrücken, Germany

Fax: ϩ49-681-302-2663

Scheme 1

E-mail: hegetschweiler@mx.uni-saarland.de

[b]

Max-Planck-Institut für Bioanorganische Chemie,

Postfach 10 13 65, 45413 Mülheim an der Ruhr, Germany

[c]

Novartis Institutes for BioMedical Research, WKL-136.P.72,

A potential problem associated with iron chelation is the

generation of oxidative stress.^[6]Reactive oxygen species

(ROS), and in particular the OH^·radical, have a consider-

Klybeckstrasse 141, 4002 Basel, Switzerland

Supporting information for this article is available on the

WWW under http://ww w .eurjic.org or from the author.

Eur. J. Inorg. Chem. 2004, 4177Ϫ4192

DOI: 10.1002/ejic.200400363

 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

4177

S. Steinhauser, U. Heinz, M. Bartholomä, T. Weyhermüller, H. Nick, K. Hegetschweiler

FULL PAPER

able potential to damage a variety of cellular components.^[7]the ferrous form, appear to be inactive in terms of oxygen

There has been Ϫ and still is Ϫ a debate on the precise toxicity, whereas aminopolycarboxylates such as NTA^3Ϫ

,

mechanism of oxidative stress.^[8Ϫ11]It has been proposed EDTA^4Ϫor DTPA^5Ϫefficiently catalyze the production of

that OH^·is generated by an Fe-catalyzed process; Fe^IIIis single and double strand breaks of DNA and of modified

reduced by O₂^·Ϫto Fe^II, which in turn reduces H₂O₂to OH^·nucleobases.^[14]The reduction potential of possible Fe^IIIL_n/

and OH^Ϫ.^[7]It has also been argued that the level of O₂in Fe^IIL_mspecies is thus a crucial parameter when developing

the cell would be much higher than that of H₂O₂, and that new iron chelators for therapeutic use (Figure 1).^[16]

the initial step of oxidative stress could be assigned to the

In the present investigation, we address the problem of

reaction of Fe^IIwith O₂rather than H₂O₂.^[12]Moreover, the the possible redox activity of FeϪICL670 complexes. For

reduction of Fe^IIIto Fe^IIis not necessarily mediated by this purpose, we have studied the various species that form

O₂^·Ϫ, but could occur through the use of abundant biologi- in solution, using a variety of physicochemical methods.

cal reducing agents such as ascorbate and thiols.^[7]

Some preliminary results have already been published.^[17]

Although the precise mechanism of oxidative stress is not Herein, we present a comprehensive report on the complex

completely understood, it is generally accepted that iron formation of ICL670 (H₃L^x) and the related ligands H₃L^y

can catalyze the production of ROS in terms of redox cyc- and H₂L^z(Scheme 1) with Fe^IIand Fe^{III [19Ϫ21]}

.

ling, when it is available in

a labile, redox-active

form.^{[6Ϫ9,13Ϫ15]}It is noteworthy that either a very high or a

very low redox potential of an iron complex prevents oxi-

dative stress. Ligands such as desferrioxamine or 1,10-phen-

2. Results

anthroline, having either a high preference for the ferric or 2.1. Solid-State Structures

2.1.1. Free Ligands

The unsubstituted derivative H₂L^zwas first described by

Ryabukhin.^[21]We recently reported its crystal structure and

documented the importance of the OϪH···N hydrogen-

bonding interactions between the phenolic hydroxy groups

and N1 or N3 of the triazole ring.^[22]Both OϪH groups of

H₂L^zare involved in intramolecular hydrogen bonds for-

ming six-membered ring structures. OϪH···N hydrogen

bonding was also observed in the carboxylic acid derivative

H₃L^x(ICL670), but in this ligand, only one of the hydroxy

groups formed an intramolecular hydrogen bond (Figure 2).

The other OϪH···N bond was intermolecular, leading to

infinite zigzag chains parallel to the crystallographic z-axis.

These chains, which are further stabilized by some stacking

interactions, are interconnected by the well-known pairing

of the carboxylic acid moieties. The resulting two-dimen-

sional sheets are oriented parallel to the y-z plane. Only one

of the phenolic rings is roughly coplanar with the central

triazole ring (the torsional angle is 2.3 °), while the other

Figure 1. The possibility for redox cycling induced by different

Fe^III/Fe^IIcomplexes in biological media according to ref.^[16]; the

redox potentials are listed for pH 7, they were taken from refs.^[16,18]

·Ϫ

and from this work (FeϪICL670); the O₂/O₂potential of Ϫ0.16

V refers to equal concentrations of dissolved O₂and O₂^·Ϫ, and not

to the standard state (1  O₂^·Ϫ, 1 atm O₂); DFO ϭ desferrioxamine,

ent.bct ϭ enterobactine, phen ϭ 1,10-phenanthroline

Figure 2. Molecular structure of H₃L^x(ICL670); the benzoic acid moiety and one phenol moiety of two neighboring H₃L^xmolecules are

also shown in order to illustrate the various types of intra- and intermolecular hydrogen bonding; atomic displacement ellipsoids are

drawn at the 30% probability level; H(ϪO) hydrogen atoms are shown as spheres of arbitrary size; H(ϪC) hydrogen atoms are omitted

for clarity. Bond lengths and angles are as expected

4178

 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org

Eur. J. Inorg. Chem. 2004, 4177Ϫ4192

Complex Formation of ICL670 and Related Ligands with Fe^IIIand Fe^II

FULL PAPER

phenolic ring with a torsional angle of 77.7° stands almost again not planar. The torsional angles between the phenol-

perpendicular to the triazole moiety. The benzoic acid moi- ate rings and the central triazole unit are 14.1° and 23.5°

ety is tilted slightly out of the triazole plane by an angle (for the isostructural [Al(L^z)₂]^Ϫ, the corresponding values

of 13.9°.

are 14.0° and 19.6°).^[24]The sodium counterion is associ-

ated with one phenoxo group of each ligand. Four EtOH

moieties complete the roughly octahedral NaO₆coordi-

nation environment.

2.1.2. Metal Complexes

There have been several unsuccessful attempts to crys-

tallize metal complexes with H₃L^xand H₃L^y. However,

complex formation with H₂L^zyielded, a variety of crystal-

line materials. Single crystals of Na[Fe(L^z)₂]·4EtOH were

found to be very fragile and disintegrated rapidly in air (loss

of EtOH). Moreover, the crystal structure of this complex

proved to be severely disordered. Refinement of a data set

collected at room temperature resulted in rather poor agree-

ment factors.^[23]This disorder could partially be resolved in

a low-temperature experiment (100 K). Two superimposed

conformers in a 2:1 ratio were located (Figure 3), and some

constraints had to be used to model the CϪC distances in

the peripheral phenyl ring realistically. The [Fe(L^z)₂]^Ϫanion

was located on a crystallographic C₂axis. This resulted in

additional disorder of the sodium counterion, which was

found in proximity of this axis (the disorder remained when

the data was refined in the lower-symmetry space group

C/c). The structure of Na[Al(L^z)₂]·4EtOH is isotypic, show-

ing the same type of disorder as observed for the Fe com-

plex.^[24]

The combination of solutions of CuBr₂, H₂L^z, and a base

resulted in the precipitation of Cu(L^z). A polymeric chain

structure has been proposed by Ryabukhin for this com-

pound.^[21]The green solid redissolved in boiling pyridine,

and addition of MeOH and water resulted in the deposition

of green crystals with the composition [Cu(L^z)(pyridine)]₂

(Figure 4). The dimeric complex is centrosymmetric; the

two Cu^IIcations have a slightly distorted square-pyramidal

coordination geometry. The basal plane is defined by the

OϪNϪO donor set of L^zand the pyridine N donor trans

to the triazole nitrogen atom. Two [Cu(L^z)(pyridine)] units

are paired by mutual bridging through a coordinated

phenoxo group. This phenoxo bridge represents the apex of

the CuN₂O₃pyramid. The CuϪO bond lengths fall in ex-

pected ranges.^[26]However, it is interesting to note that the

˚

two CuϪN lengths (1.935 and 2.052 A) differ significantly,

the distance to the pyridine donor being longer. The bridg-

ing interaction of the phenoxo group is asymmetric with a

long CuϪO distance to the donor in apical position. The

two phenoxo moieties are again twisted with torsional

angles of 19.2° and 24.0°. The structure is stabilized by

some stacking interactions between the triazole and the

pyridine moieties. As expected, the dimeric complex shows

antiferromagnetic coupling between the two Cu^IIcenters

with an S ϭ 0 ground state.

[Fe(L^z)₂]^Ϫadopts the expected bis-structure with a meri-

dional coordination mode of the two L^zentities, a distorted

octahedral coordination environment, and FeϪO distances

at the short end of the range expected for high-spin Fe^{III [25]}

.

˚

The FeϪN distance of 2.092(5) A is also remarkably short.

It is worth noting that the ligand in this complex is once

Figure 3. Molecular structure of the [Fe(L^z)₂]^Ϫanion, showing the major and minor components of the disordered structure; the atomic

˚

displacement ellipsoids are drawn at the 30% probability level; hydrogen atoms are omitted for clarity; selected bond lengths [A] and

angles (°): Fe1ϪO11 1.958(4), Fe1ϪO1 2.002(4), Fe1ϪN1 2.092(5), O11ϪFe1ϪO11Ј 92.4(3), O11ϪFe1ϪO1 166.5(2), O11ϪFe1ϪO1Ј

90.6(2), O1ϪFe1ϪO1Ј 89.7(2), O11ϪFe1ϪN1 84.1(2), O11ϪFe1ϪN1Ј 101.0(2), O1ϪFe1ϪN1 82.4(2), O1ϪFe1ϪN1Ј 92.5(2),

N1ϪFe1ϪN1 172.7(3)

Eur. J. Inorg. Chem. 2004, 4177Ϫ4192

www.eurjic.org

 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

4179

S. Steinhauser, U. Heinz, M. Bartholomä, T. Weyhermüller, H. Nick, K. Hegetschweiler

FULL PAPER

Figure 5. The dependence of the pK_a,2and pK_a,3of H₃L^yon

x_DMSO; circles indicate values with x_DMSOՆ0.06, squares refer the

experimental values in pure water; the fitted line illustrates the re-

liability of the extrapolation procedure (x_DMSOǞ 0) with an accu-

racy of the x_DMSOϭ 0 values of about Ϯ0.1

H₃L^x, H₃L^yand H₂L^z. Moreover, the influence of the inert

electrolyte (KCl or KNO₃) appears to be negligible, and the

interpolation procedure yielded reliable estimates for the

pK_as in H₂O, with an accuracy of about Ϯ0.1. For an ap-

propriate interpretation of these data, it is important to

note that the ionic product (apparent K_wϭ Q_cϭ [H^ϩ] ϫ

[OH^Ϫ]) in the H₂O/DMSO system is considerably smaller

than in pure H₂O.^[27]As a consequence, the pH scale for

Figure 4. Molecular structure of the dinuclear [Cu(L^z)(pyridine)]₂

complex; the atomic displacement ellipsoids are drawn at the 30%

probability level; hydrogen atoms are omitted for clarity; selected

˚

bond lengths [A] and angles (°): Cu1ϪO112 1.922(6), Cu1ϪN11

1.935(6), Cu1ϪO142 1.946(5), Cu1ϪN1P 2.052(6), Cu1ϪO142Ј

2.360(5), O112ϪCu1ϪN11 89.6(2), O112ϪCu1ϪO142 176.2(2),

N11ϪCu1ϪO142

N11ϪCu1ϪO142Ј

O112ϪCu1ϪN1P

86.8(2),

96.5(2),

92.7(3),

O112ϪCu1ϪO142Ј

O142ϪCu1ϪO142Ј

N11ϪCu1ϪN1P

96.9(2),

84.8(2),

166.3(3),

x

_DMSOϭ 0.2 is extended to a range of 0Ϫ15.6, and neutral

O142ϪCu1ϪN1P 90.4(3), N1PϪCu1ϪO142Ј 96.7(2)

solutions have a pH value of about 7.8.

The observed pK_avalues can be grouped in three categor-

ies:

2.2. pK_aValues of the Free Ligands

The pK_as were measured by potentiometric and spectro-

(a) pK_aϽ2. Such low values could not be determined

photometric titrations (25 °C, 0.1  KCl/KNO₃). In their by conventional potentiometric methods and were therefore

fully protonated forms all three derivatives show rather low measured spectrophotometrically. All the investigated li-

solubility in H₂O. Therefore, some of the deprotonation gands showed significant spectral changes in the range 0 Ͻ

1

studies were performed in H₂O/DMSO solutions with 0.06 pH Ͻ 2. A H NMR spectroscopic study of H₃L^xshowed

Յ x_DMSOՅ 0.20 (x_DMSOis the mol fraction of DMSO). considerable shifts for the hydrogen atoms of the two phe-

The dependence of the pK_as on x_DMSOwas found to be nol rings, whereas the resonances for the C₆H₄ϪCOO hy-

roughly linear (Figure 5), and some of the missing values drogen atoms (which appeared as a characteristic AAЈXXЈ

for pure H₂O could be estimated by extrapolation: spin system) remained virtually constant. Obviously, these

pK_a,i(x Ͼ 0) ϭ pK_a,i(x ϭ 0) ϩ a ϫ x, with 4.7 Ͻ a Ͻ changes must be attributed to protonation of a triazole ni-

7.4. Inspection of these data (Table 1) showed no significant trogen atom (for the unsubstituted 1,2,4-triazole, a pK_aof

1

differences for pK_a,2and pK_a,3between the three ligands 2.19 is reported in the literature).^[28]According to the H

Table 1. pK_avalues (25 °C, 0.1  KCl or KNO₃) of H₃L^x, H₃L^y, and H₂L^zin H₂O and in H₂O/DMSO solutions with a mol fraction

x_DMSOϭ 0.2; uncertainties (3σ) are given in parentheses

H₃L^x(ICL670)

x_DMSOϭ 0.20

H₃L^y

H₂L^z

[a]

pK_a,1

pK_a,2

pK_a,3

4.61(2)^{[b] [c]}

10.12(2)^[b][c]

12.08(2)^[b][c]

4.61(4)^{[d] [e]}

10.00(3)^[d][e]

12.12(4)^[d][e]

Ͻ2

10.10(4)^[d][e]

12.15(2)^[d][e]

10.12(1)^[c][d]

10.0(1)^[d][e]

12.2(1)^[d][e]

12.21(3)^[c][d]

H₂O only

Ϫ^[f]

8.80(1)^[d][h]

10.61(1)^[d][h]

3.7(1)^[d][g]

9.0(1)^[d][g]

10.6(1)^[d][g]

Ͻ2^[c][d]

1.1(1)^[d][e]

[a]

pK_a,1

pK_a,2

pK_a,3

8.74(1)^[c][d]

10.63(1)^[c][d]

8.76(3)^[d][e]

10.67(3)^[d][e]

8.6^[d][g]

Ϫ^[f]

10.7^[d][g]

^[a]pK_a,iϭ ϪlogK_a,i, K_a,iϭ [H_3ϪiL] ϫ [H] ϫ [H_4ϪiL]^Ϫ1

.

^[b]0.1  KNO₃. ^[c]pH titration of H_mL with KOH. ^[d]0.1  KCl. ^[e]Spectrophoto-

[f]

[g]

[h]

metric titration. Insufficient solubility. Extrapolated for x_DMSOǞ 0. pH titration of L^{x 3Ϫ}with HCl.

4180

 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org

Eur. J. Inorg. Chem. 2004, 4177Ϫ4192

Complex Formation of ICL670 and Related Ligands with Fe^IIIand Fe^II

FULL PAPER

NMR spectroscopic characteristics, protonation of the ni- and spectrophotometric methods. The overall stability con-

trogen atom is also observed for the sulfonic acid derivative stants, which are defined according to: mFe^3ϩϩ nL^uϪ

ϩ

Ǟ

ǟ

H₃L^y. Consequently, H₃L^ymust be present predominantly qH^ϩ

Fe_mL_nH_q^3mϪunϩq, β_mnqϭ [Fe_mL_nH_q^3mϪunϩq] ϫ

in a zwitterionic form in strongly acidic aqueous media [Fe^{3ϩ Ϫm}

]

ϫ [L^{uϪ Ϫn}

]

ϫ [H^ϩ]^Ϫq, (L^x, L^y: u ϭ 3, L^z: u ϭ 2),

(Scheme S1, Supporting Information).

are summarized in Table 2.

(b) 3 Ͻ pK_aϽ 5 was only observed for H₃L^xand corre-

sponds to the deprotonation of the carboxylic acid group.

(c) pK_aϾ8. These values are characteristic of the depro-

tonation of the phenolic OH groups. Although the two OH

groups are structurally inequivalent, it appears highly prob-

able that they are deprotonated simultaneously, and that the

observed pK_as represent averages of the individual (micro-

scopic) acidity constants.^[29]

In acidic solution (pH Ͻ3.5), formation of a 1:1 complex

was verified by JobЈs plots and was further confirmed by

spectrophotometric batch titrations for all three ligands

(Figures S5ϪS7, Supporting Information). Below pH 4,

ICL670 forms a protonated complex [Fe(HL^x)]^ϩ(indicated

as 111 in Scheme 2). This species could be deprotonated

to [Fe(L^x)] (110). We propose that the additional proton is

bonded to the peripheral carboxylate group of the ligand.

This assignment is consistent with the observation that a

corresponding 111 species does not form with the sulfonic

2.3. Complex Formation with Fe^III

The formation of Fe^IIIcomplexes with H₃L^x, H₃L^yand acid H₃L^yand the phenyl derivative H₂L^z. Clearly, the per-

H₂L^zwas also studied in H₂O/DMSO by potentiometric ipheral sulfonate group of H₃L^yis not sufficiently basic

Table 2. Overall formation constants (log β_mnq) of Fe^IIIcomplexes with H₃L^x, H₃L^y, and H₂L^z(25 °C, 0.1  KCl or KNO₃) in H₂O or

in H₂O/DMSO solutions with a mol fraction x_DMSOϭ 0.20; uncertainties (3σ) are given in parentheses

[a]

log β_mnq

H₃L^x(ICL670)

H₃L^y

H₂L^z

mnq

x_DMSOϭ 0.20

111

110

122

121

120

27.5(1)^[b][c]

23.33(3)^[d][e], 23.5(1)^[b][c]

48.7(1)^[d][e]

22.50(1)^[b][c]

22.85(1)^[b][c]

44.36(1)^[d][e], 44.5(1)^[b][c]

38.56(2)^[d][e], 38.6(1)^[b][c]

43.05(6)^[b][c]

37.46(9)^[b][c]

not determined^[f]

H₂O only

111

110

122

121

120

24.3^[b][g]

22.0^[b][g]

43.4^[b][g][h]

41.2^[b][g]

36.9^[b][g]

21.30(4)^[b][c], 21.2^[b][g]

not determined^[f]

40.89(3)^[b][e], 40.79(4)^[b][c], 40.9^[b][g]

35.95(5)^[b][e], 36.06(4)^[b][c], 35.8^[b][g]

not determined^[f]

[a]

[b]

[c]

[d]

[e]

β_mnqϭ [Fe_mL_nH_q] ϫ [Fe]^Ϫmϫ [L]^Ϫnϫ [H]^Ϫq

.

0.1  KCl.

Spectrophotometric titration.

0.1  KNO₃.

Potentiometric

[f]

[g]

[h]

titration. Insufficient solubility. Extrapolated for x_DMSOǞ 0.

This species does not appear in the equilibrium in pure H₂O to a

significant extent.

Scheme 2

Eur. J. Inorg. Chem. 2004, 4177Ϫ4192

www.eurjic.org

 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

4181

S. Steinhauser, U. Heinz, M. Bartholomä, T. Weyhermüller, H. Nick, K. Hegetschweiler

FULL PAPER

to allow protonation in aqueous solution. Moreover, clusive evidence for the nature of this complex is obtained

[Fe(HL^x)]^ϩand [Fe(L^x)] have very similar optical spectra from an analysis of the Vis spectroscopic data, which estab-

(Figure 6b). Since the observed bands in the Vis spectra are lished that the spectra of [Fe(L^x)(HL^x)]^2Ϫ(121) and

mainly charge-transfer transitions, protonation of the per- [Fe(L^x)₂]^3Ϫ(120) are distinctly different (Figure 6b). The

ipheral carboxylate group does not modify the chromo- situation is similar for [Fe(L^y)(HL^y)]^2Ϫand [Fe(L^y)₂]^3Ϫ(see

phore of this complex to a significant extent.

Figure S8, Supporting Information). Obviously, the two

complexes must have different coordination spheres, and we

therefore propose a bidentate coordination mode for one of

the ligands in the 121 complex with the triazole nitrogen

and one of the phenolic oxygen donors as the only coordi-

nating atoms (Scheme 2). The second phenolic oxygen

donor is protonated and probably hydrogen bonded to the

free triazole nitrogen atom (as observed for the free ligands,

see Figure 2). Further evidence for this particular structure

is provided by FAB^ϩMS (measured in EtOH). The spec-

trum shows the formation of the species [Fe(L^z)₂(EtOH)]^ϩ

with rather high intensity. The binding of one (and only

one) EtOH molecule can be nicely explained, if the complex

is formulated as [Fe(L^z)(HL^z)OEt]^ϩ, with a bidentate HL

ligand and the sixth position in the coordination sphere oc-

cupied by an OEt^Ϫmoiety.

The spectra of the monoprotonated [Fe(L^x)(HL^x)]^2Ϫand

the doubly protonated [Fe(HL^x)₂]^Ϫdo not differ to any sig-

nificant extent, and we therefore assign the reaction of

[Fe(L^x)(HL^x)]^2Ϫwith H^ϩto a simple protonation of one of

the peripheral carboxylate groups. The similarity of the two

spectra, together with a rather low abundance of these spec-

ies, prevents the determination of the formation constant

β₁₂₂of [Fe(HL^x)₂]^Ϫby spectrophotometric methods. The

value for log β₁₂₂could only be established by potentio-

Figure 6. Determination of the formation constants log β_mnqof

[Fe_m(L^x)_nH_q]^3mϪ3nϩq; a) the sample spectra of the spectrophoto-

metric titration with total [Fe]:total [L^x] ϭ 1:3 in the range 3.2 Ͻ

pH Ͻ 6.6 (x_DMSOϭ 0.2); b) calculated spectra of the individual

[Fe_m(L^x)_nH_q]^3mϪ3nϩqspecies indicated as mnq; the spectra of Fe^3ϩ

(ϭ 100) and of [Fe(HL^x)]^ϩ(ϭ 111) were measured separately and

were imported without refinement

At higher pH, complex formation with H₃L^xand H₃L^ymetric titrations. In agreement with this assignment, the

was explored by continuous titrations (Figure 6). An ad- doubly protonated [Fe(HL^y)₂]^Ϫand [Fe(HL^z)₂]^ϩspecies

ditional ligand is bound to the metal center above pH 4, were not observed.

and a series of 1:2 complexes are formed. Experiments with

A comparison of the various log β values (Table 2) shows

H₂L^zcould not be evaluated due to precipitation. Forma- the following trends:

tion of a reddish-violet solid was always observed, even if a

(a) Dependence on DMSO concentration. The sulfonic acid

very low concentration of total Fe (10^Ϫ4) and a large derivative has significantly enhanced solubility, and the entire

excess of ligand (total L^z: total Fe Ͼ 4) was used. The pre- range 0 Յ x_DMSOՅ 0.2 could thus be investigated. A steady

cipitation began after addition of about 1.2 equivalents of increase of log β was generally noted for an increase in the

base, and the precipitate redissolved completely above pH mol fraction of DMSO. The difference ∆log β ϭ log

6. Obviously, a bis complex of zero charge is formed with β_x(DMSO)ϭ0.2Ϫ log β_x(DMSO)ϭ0is 1.2 for [Fe(L^y)], and 1.5

a tentative composition [Fe(L^z)(HL^z)] (indicated as 121 in for [Fe(L^y)₂]^3Ϫ. Complex formation of H₃L^xwith Fe^3ϩwas

Scheme 2). Formation of a similar 121 intermediate was performed in solutions with a mol fraction x_DMSOof 0.22,

also established for the sulfonic acid H₃L^yby potentio- 0.20 and 0.18. The extrapolation x_DMSOǞ 0 yielded ∆log

metric and spectrophotometric methods. However, due to β₁₁₀ϭ 1.3 and ∆log β₁₂₀ϭ 1.6. The similar behavior of the

the anionic nature of [Fe(L^y)(HL^y)]^2Ϫ, an insoluble solid two ligands H₃L^xand H₃L^yand the linear dependence of

did not form. The 121 complex deprotonated around pH log β on x as observed for H₃L^yjustify application of the

5Ϫ6 to form [Fe(L^y)₂]^3Ϫ(ϭ 120), which was then present extrapolation procedure as described above for the pK_aval-

as the sole species up to pH 12. The carboxylic acid H₃L^xues (Figure 5). As a result, we note that Fe^IIIϪL^xcomplexes

behaved similarly, although an additional doubly pro- with a peripheral COOH group are much stronger acids in

tonated bis complex [Fe(HL^x)₂]^Ϫ(ϭ 122) was detected as a H₂O than in H₂O/DMSO mixtures. Corresponding pK_aval-

minor species in the region 4 Ͻ pH Ͻ 5.

ues derived from the overall formation constants (Table 2)

For the mono complex, the reaction [Fe(L^x)] ϩ H^ϩ

Ǟ

are 2.3/2.2 (H₂O) and 4.2/4.3 (x_DMSOϭ 0.2) for [Fe(HL^x)]^ϩ/

[Fe(HL^x)]^ϩwas attributed to a simple protonation of the [Fe(HL^x)₂]^Ϫ. We attribute this effect to the different solvation

peripheral carboxylate group (vide supra). However, for the properties of the two media. Consequently, such protonated

bis complexes, a different type of reaction must be con- species appear in a substantial amount in the H₂O/DMSO

sidered, since [Fe(L^y)₂]^3Ϫand [Fe(L^z)₂]^Ϫhave no peripheral mixture (Figure 7a), whereas in pure H₂O, their formation is

sites available for protonation around pH 5 (Figure 7). Con- of minor relevance or not significant at all (Figure 7b).

4182

 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org

Eur. J. Inorg. Chem. 2004, 4177Ϫ4192

Complex Formation of ICL670 and Related Ligands with Fe^IIIand Fe^II

FULL PAPER

Figure 7. Species distribution plots for equilibrated solutions of Fe^IIIϪH₃L systems; a) H₃L^x, x_DMSOϭ 0.2; b) H₃L^x, x_DMSOϭ 0.0; c)

H₃L^y, x_DMSOϭ 0.2; d) H₃L^y, x_DMSOϭ 0.0; total Fe and total ligand concentrations are 1.0 m, and 2.0 m, respectively; only metal-

containing species are shown (total Fe ϭ 100%); the equilibrium constants listed in Tables 1 and 2 were used for the calculations

(b) Comparison of the different ligands. The stability of ent to be oxidized even in the complete absence of oxygen.

[FeL] with x_DMSOϭ 0.2 decreases in the order H₃L^x

Ͼ

Corresponding CV measurements (see section 2.5.2) con-

H₂L^zϾ H₃L^y. However, the differences are relatively minor. firmed that [Fe^IIL₂]^4Ϫis able to reduce H₂O (formation of

In general, the iron binding properties of H₃L^xand H₃L^yH₂) in neutral solution. It was therefore not possible to de-

are indeed closely related: the pFe values^[30]for H₃L^xand termine any formation constant for the ferrous 1:2 complex

H₃L^y(total [L] ϭ 10 µM, total Fe ϭ 1 µM, pH ϭ 7.4, H₂O, with this method. The potentiometric measurements could,

I ϭ 0.1 ) are 21.5 and 20.6, respectively.

however, be used to investigate the formation of the 1:1

complex [Fe^II(L^y)]^Ϫ, according to the reaction Fe^2ϩ

ϩ

2.4. Complex Formation with Fe^II

H₃L^{y Ǟ}[Fe(L^y)]^Ϫϩ 3 H^ϩ. The overall formation constant

ǟ

β₁₁₀(referring to the reaction Fe^2ϩϩ L^{3Ϫ Ǟ}[FeL]^Ϫ) was

The formation of Fe^IIcomplexes with H₃L^ywas again

studied in H₂O/DMSO mixtures (x_DMSOϭ 0.20). However,

a quantitative determination of the formation constants

proved extremely difficult, because the ferrous complexes

reacted as strong reducing agents and were readily oxidized

to the corresponding Fe^IIIspecies. This oxidation process

could easily be observed by the intense red or purple color

of the Fe^IIIcomplexes (the Fe^IIcomplexes are virtually col-

orless). It was thus of utmost importance to remove any

ǟ

determined in H₂O/DMSO (x_DMSOϭ 0.2) to be log β₁₁₀

ϭ

11.5(2). As expected, this number is substantially lower

than the corresponding value for Fe^3ϩ(22.5, see Table 2).

2.5. Redox Properties of the Fe^II/Fe^III؊H₃L System

2.5.1. Redox Chemistry in Acidic Solutions

Below pH 3, the 1:1 complexes [Fe(HL^x)]^ϩand [Fe(L^y)]

traces of oxygen from the sample solutions. A series of are predominantly formed, even when the ligand is present

potentiometric titrations was performed under Ar to deter- in excess (Figure 7). The redox characteristics of these two

mine the corresponding formation constants. However, it complexes have been investigated thoroughly by means of

was not possible to suppress the oxidatively induced side cyclic voltammetry. The two complexes behave rather simi-

reactions completely. In particular, the 1:2 complex larly, and we will focus here on the behavior of [Fe(HL^x)]^ϩ

[Fe^II(L^y)₂]^4Ϫwas found to be a strong enough reducing ag- (Figure 8) and present the results for [Fe(L^y)] in an abbrevi-

Eur. J. Inorg. Chem. 2004, 4177Ϫ4192

www.eurjic.org

 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

4183

S. Steinhauser, U. Heinz, M. Bartholomä, T. Weyhermüller, H. Nick, K. Hegetschweiler

FULL PAPER

ated manner. A Pt working electrode was used exclusively Supporting Information). The cathodic peak current was

in acidic solution, and all experiments were performed in a found to be linearly dependent on the square root of the

H₂O/DMSO medium (x_DMSOϭ 0.2) to avoid precipitation scan rate, as expected for a diffusion controlled reaction

of the insoluble H₃L. The use of DMSO as a particularly (Figure S12, Supporting Information). The peak potential

useful solvent in electrochemistry is well-established.^[31]

of the reduction slightly shifted to less positive values with

increasing scan rate, and this is consistent with an irrevers-

ible process. As a result of the complete decomposition of

the complex, the reverse scans showed a half wave indicative

of the reoxidation of the free Fe^2ϩto Fe^3ϩat about 400 mV

(versus Ag/AgCl) according to:

Fe^2ϩǞ Fe^3ϩϩ e^Ϫ

(3)

This assignment is further supported by a characteristic

time dependence, which was observed in subsequent scans.

If a sufficiently slow scan rate was used (Ͻ 50 mV/s), en-

ough time was available for the re-formation of the ferric

complex after reoxidation of Fe^2ϩto Fe^3ϩ, according to:

Fe^3ϩϩ H₃L^xǞ [Fe^III(HL^x)]^ϩϩ 2 H^ϩ

(4)

Consequently, the second scan was identical to the first

scan. Further scans did not result in any change in the be-

havior. However, if a higher scan rate of about 50Ϫ500 mV/

s was applied, [Fe^III(HL^x)]^ϩwas only partially recovered

[Equation (4)], and in the following scan, the solution still

contained some free Fe^3ϩ, which was reduced to Fe^2ϩ

:

Fe^3ϩϩ e^ϪǞ Fe^2ϩ

(5)

Figure 8. Cyclic voltammograms of [Fe(HL^x)]^ϩ(pH ϭ 2.9, total

[Fe] ϭ 2.5 m, total [L^x] ϭ 5 m, x_DMSOϭ 0.2 using a Pt working

electrode at scan rates as indicated

The above reaction [Equation (5)] resulted in the appear-

ance of a new reduction wave at around 300 mV (versus

Ag/AgCl), and as a consequence, two reduction waves [see

Equation (1) and Equation (5)] were now observable. The

ratio of their intensity varied characteristically with scan

rate. Increasing scan rates resulted in an increase in the sup-

pression of Equation (4), and therefore in an increase in the

amount of free Fe^3ϩand a decrease in the amount of

[Fe^III(HL^x)]^ϩ. Consequently, the intensity of the peak at

300 mV increased with increasing scan rate, whereas that of

the peak at around Ϫ100 mV (its counterpart) diminished.

At a scan rate higher than 500 mV/s, the amount of com-

plex formed after the first cycle was less than 50%, and

the quasi-reversible pattern for the free Fe^2ϩ/Fe^3ϩcouple

[Equation (3) and Equation (5)] was observed. This time-

dependence was found to be in good agreement with the

kinetics of [Fe(HL^x)]^ϩformation, which was established

previously by stopped flow experiments.^[17]

Reduction of [Fe(HL^x)]^ϩresulted in the formation of the

corresponding Fe^IIcomplex according to the reaction:

[Fe^III(HL^x)]^ϩϩ e^ϪǞ [Fe^II(HL^x)]

(1)

However, the ferrous complex is not stable and decom-

poses immediately and completely according to the reac-

tion:

[Fe^II(HL^x)] ϩ 2 H^ϩǞ Fe^2ϩϩ H₃L^x

(2)

The electron uptake and complex decomposition de-

scribed by the above reactions [Equation (1) and Equa-

tion (2)] was indicated by a characteristic irreversible peak

in the range ϩ100 mV to Ϫ150 mV (versus Ag/AgCl). As

indicated by Equation (2), the decomposition is strongly pH

dependent. The peak shifted to less positive values at higher

pH with a slope of 109 mV/pH ([Fe(L^y)]: 129 mV/pH), indi-

If the pH was lowered to Յ2, the acidity of this medium

was high enough to lead to the partial decomposition of

[Fe^III(HL^x)]^ϩ(Figure 7):

cating a transfer of two protons per electron (Figure S13, [Fe^III(HL^x)]^ϩϩ 2 H^ϩǞ Fe^3ϩϩ H₃L^x

(6)

4184  2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.eurjic.org

Eur. J. Inorg. Chem. 2004, 4177Ϫ4192

Complex Formation of ICL670 and Related Ligands with Fe^IIIand Fe^II

Consequently, two reduction waves were observed al-

FULL PAPER

For ICL670 (H₃L^x), the reduction of [Fe^III(L)₂]^3Ϫ[Equa-

ready in the first scan, and it was not possible to suppress tion (7)] was also investigated in H₂O, and a very similar

the peak at 300 mV by sufficiently slow scan rates.

behavior was noted. The observed reduction potential was

determined as Ϫ0.58 V (versus NHE). Excess ligand was

used as buffer in our aqueous experiments, to stabilize the

pH in the range 11Ϫ12 (the free ligand is electrochemically

inactive). Moreover, the excess ligand also inevitably served

to suppress the decomposition of [Fe^IIL₂]^4Ϫ[Equation (8)].

2.5.2. Redox Chemistry in Neutral and Alkaline Solutions

Above pH 6, 1:2 complexes are predominantly formed,

and the relevant reduction reaction is thus:

[Fe^III(L)₂]^3Ϫϩ e^ϪǞ [Fe^II(L)₂]^4Ϫ

(7)

3. Discussion

3.1. Structure and Selectivity

However, CV experiments with an Au, Pt or glassy car-

bon working electrode did not show any redox activity for

[Fe^III(L)₂]^3Ϫ. Obviously, [Fe^II(L)₂]^4Ϫis a stronger reducing

agent than H₂(see section 2.4), and the failure to observe

any redox activity is due to the production of hydrogen

prior to the reduction of the complex. An Hg (hanging

drop) electrode was thus used to circumvent the problem of

hydrogen formation, and the quasi-reversible reduction of

[Fe^III(L)₂]^3Ϫ[Equation (7)] could then be detected above

pH 12 (Figure 9). The peak separations ∆E_Pwere 93 mV

and 102 mV, and the reduction potentials for the

[Fe^III/II(L)₂]^3Ϫ/4Ϫcouple were Ϫ0.92 V and Ϫ0.89 V for

H₃L^xand H₃L^y, respectively (x_DMSOϭ 0.2, vs. Ag/AgCl).

As expected, this value does not depend on pH (Figure

S14). Plots of the cathodic peak current vs. the square root

of scan rate established a purely diffusion-controlled pro-

cess (Figure S12). At pH Ͻ12, the two waves of to this

redox process were still visible; however, additional peaks

were observed in the reverse scan at more positive values

(Figure S15). We attribute these characteristics to a partial

dissociation of the ferrous complex [Equation (8)]:

It has been suggested that the application of ICL670 in

the therapy of iron overload could be problematic due to

the lack of sufficient selectivity.^[2]In particular, the presence

of an sp²nitrogen atom would generate a donor set with a

significant tendency to bind divalent biometal cations such

as Zn^2ϩ. Although the affinity of high-spin Fe^IIIfor six

negative oxygen donors generally appears to be over-

estimated,^[32]it is true that an increasing number of nitrogen

donors would enhance the stability of metal complexes with

late divalent transition metal cations (Ni^2ϩ, Cu^2ϩ, Zn^2ϩ),

because an sp²nitrogen donor is somewhat softer than a

negative oxygen donor.^[33,34]However, the metal selectivity

of a complexing agent cannot be explained simply in terms

of the hardness of the donor set. Additional steric factors

of the ligand must also be considered,^[33]and it appears that

in discussing the metal-ion selectivity of ICL670, these

steric aspects have not received the necessary attention in

some of the previous reports.

With regard to the particular steric properties of ICL670,

it is important to recognize that an entirely planar LϪM

unit with an OϪMϪO angle of 180° (Scheme 3a) would

˚

imply an MϪO bond length of 1.60 A and an MϪN bond

˚

length of 1.90 A. Only very small cations are capable of

[Fe^IIL₂]^4Ϫϩ nH^ϩǞ [Fe^IIL]^Ϫϩ H_nL^nϪ3

(8)

forming such a short MϪO bond. It has been shown by

Hancock that ligands forming six-membered chelate rings

have a distinct preference for small metal centers, whereas

five-membered chelate rings are more suited for larger cat-

ions.^[35]Clearly, with its rigid structure and its constraint of

two six-membered chelate rings, ICL670 will have a strong

preference to bind small metal cations. As a consequence,

all complexes of ICL670 with di- and trivalent transition-

metal cations must be regarded as strained, and this strain

increases significantly with increasing ionic size. This result

is of importance with regard to selectivity of this ligand in

biological systems. It is noteworthy that most of the com-

peting biometal cations such as Zn^2ϩare larger than Fe^3ϩ

,

and binding of these elements to ICL670 is thus clearly dis-

criminated on the basis of steric considerations. The bond

strain in such complexes can be absorbed to a large extent

by angle deformation, and the MϪO distance can be wid-

ened by out-of-plane twisting of the two phenyl rings

(Scheme 3b). As already noted in the results section (2.1.2),

this type of twist has generally been observed in the crystal

structures of such metal complexes, and the corresponding

twist angle increases, as expected, with increasing ionic ra-

Figure 9. Cyclic voltammogram of [Fe(L^x)₂]^3Ϫ(pH ϭ 12.7, total

[Fe] ϭ 2.5 m, total [L^x] ϭ 5 m, x_DMSOϭ 0.2 at a Hg (hanging

drop) working electrode with a scan rate of 50 mV/s

Eur. J. Inorg. Chem. 2004, 4177Ϫ4192

www.eurjic.org

 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

4185

S. Steinhauser, U. Heinz, M. Bartholomä, T. Weyhermüller, H. Nick, K. Hegetschweiler

FULL PAPER

dius. However, the rotation of the two phenoxo rings inevi- of one of the coordinated phenoxo groups. Such a detach-

tably results in a compression of the MϪN bond. Corre- ment of a basic donor atom by protonation is well-estab-

sponding MϪN distances observed in the crystal structures lished for strained chelate rings.^[37]

are indeed rather short. Again, the compression is partially

In the bis complexes of ICL670, the discussion of struc-

compensated by deformation of the cisϪOϪMϪN and ture-stability correlations requires the analysis of additional

transϪOϪMϪO angles (Scheme 3c).

interligand interactions. For this purpose, it is necessary to

briefly discuss the different stereoisomers that are possible

for such an ML₂complex. In the simple ML unit, the twist

of the two phenyl rings can either occur in a clockwise or

counter-clockwise manner relative to the central triazole

unit (Scheme 4a). By analogy with the usual nomenclature,

the O···O line will be used in this work to refer to the two

forms as the λ or δ enantiomer, and the combination in a

bis complex would give either a λδ or a δδ (or λλ) form

(Scheme 4b). The two forms adopt considerably different

geometries. In the λδ isomer, the two ligands are oriented

perpendicularly, whereas the δδ isomer has a rather flat

shape with an almost parallel orientation of the four phenyl

rings that allow for some π-stacking to occur. Due to the

nonequivalence of the two noncoordinating nitrogen atoms,

the symmetry of the λδ isomer is C₁. For the δδ isomer, the

addition of a substituent to one of the nitrogen atoms re-

sults either in a cisoid or transoid structure. Both of these

have C₂symmetry. [Fe^III(L^z)₂]^Ϫand [Al^III(L^z)₂]^ϪϪ as ob-

served in the crystalline compounds Ϫ correspond to the C₂

isomer with the transoid configuration being predominant

(Figure 3). Inspection of the different structures reveals that

for all possible isomers, repulsive interligand interactions are

relatively minor. The exclusive formation of the λλ/δδ form

in the Fe^IIIand Al^IIIcomplexes may be attributed to the

possible stabilization by intramolecular π-stacking and also

to the better accessibility of the Na^ϩcounterion for its

bonding to two phenoxo oxygen atoms (indicated by arrows

in Scheme 4b).

Scheme 3

The favored binding of small metal cations is nicely illus-

trated by the stability constants of [M(L^x)] complexes with

Mg^2ϩ(log K ϭ 7.6) and Ca^2ϩ(log K ϭ 5.5).^[17]These val-

ues should be compared with values for other ligands that

have a mixed O,N donor set such as NTA^3Ϫ(log K_MgNTAϭ

5.5, log K_CaNTAϭ 6.3) or EDTA (log K_MgEDTAϭ 8.8, log

K

_CaEDTAϭ 10.7).^[36]The two aminopolycarboxylate ligands

exclusively form five-membered chelate rings, and as a

consequence, the complexes with the larger cation (Ca^2ϩ

)

are more stable. For ICL670, the reverse trend is observed,

and the complexes with the smaller Mg^2ϩare about two

orders of magnitude more stable. Consequently, replace-

ment of the ferric iron by the larger Zn^2ϩion, according to

[Fe(L^x)₂]^3Ϫϩ Zn^{2ϩ Ǟ}[Zn(L^x)₂]^4Ϫϩ Fe^3ϩ, is strongly dis-

ǟ

favored. The equilibrium constant for this reaction is 7.9 ϫ

10^{Ϫ22 [17]}

The significant stress, even in the ferric complex,

.

also explains the rather unexpected formation of the pro-

tonated species [Fe(L)(HL)]^2Ϫ(L ϭ L^x, L^y) or [Fe(L)(HL)]

(L ϭ L^z) that was observed for all three ligands. Obviously,

the transformation to a bidentate coordination mode arises

Scheme 4. a) The two enantiomeric ML conformations; b) the pos-

because of the release of strain induced by the detachment sible ML₂forms (λ,δ and δ,δ)

4186

 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org

Eur. J. Inorg. Chem. 2004, 4177Ϫ4192

Complex Formation of ICL670 and Related Ligands with Fe^IIIand Fe^II

FULL PAPER

µM).^[4]As expected, ICL670 displays optimum activity in

neutral and slightly alkaline solutions (6 Ͻ pH Ͻ 10).^[38]

High concentrations of H^ϩor OH^Ϫfavor either competitive

protonation of the ligand or render the solid hydroxide very

insoluble. At physiological conditions, more than 80% of

the ligand is still loaded with Fe^IIIas [FeL₂]^3Ϫeven at low

total L concentrations such as 1 µM, and the fraction of

the 1:1 complex is less than 3%, even if the level of total L

is 10 nM. The protonated 1:2 complex is not of importance

at physiological conditions as well.

3.2. Species Distribution and Redox Properties of the

Fe^III؊ICL670 System

As outlined in the introduction, the abundance and redox

potential of every single component must be considered

when discussing the problem of possible redox cycling. As

a tridentate ligand, ICL670 can form ML and ML₂com-

plexes, and both species must be regarded as possible redox

catalysts. A conventional species distribution diagram for

the ICL670ϪFe^3ϩsystem is shown in Figure 7 (see a and

b). However, in this type of representation, the amount of

total Fe in solution is arbitrary. In contrast, Figure 10 illus-

trates a situation in which the different ferric complexes are

in equilibrium with solid FeOOH. The concentration of free

Fe^3ϩis then mediated by the solubility product of this solid

phase, and the total amount of dissolved Fe is a direct

measure of the efficacy of a ligand for Fe^IIIchelation at a

given pH.^[38]This type of representation more closely re-

flects the condition in an iron-overloaded patient. Figure 10

shows the amount of dissolved Fe for total ligand concen-

trations of 100 µM, 1 µM, and 10 nM (the corresponding

value in the serum of an iron-overloaded patient during oral

treatment with ICL670 falls in the range 10 µMϪ100

The stability of ferrous [Fe^II(L^y)]^Ϫ(log β₁₁₀ϭ 11.5) can

be compared with the corresponding value log β₁₁₀ϭ 13.3

reported for [Zn(L^x)]^Ϫ.^[17]The ratio of these values corre-

sponds closely to the expectations from a linear free energy

relation ∆G^Zn/∆G^Feϭ constant.^[33]For a variety of ligands

having similar, mixed N,O donor sets and a related charge

or denticity, the ratio log β₁₁₀([Zn^IIL])/log β₁₁₀([Fe^IIL]) is

1.1Ϫ1.2 (Table 3). The corresponding ratio of 1.16 for log

β([Zn^II(L^x)]^Ϫ)/log β([Fe^II(L^y)]^Ϫ) corroborates the potentio-

metric results of the present investigation.^[39]The value for

β₁₂₀for the ferrous complex could not be determined di-

rectly due to instantaneous oxidation of [FeL₂]^4Ϫ. An anal-

ogous free energy relation as described for log β₁₁₀gave log

β₁₂₀ϭ 14.8 (x_DMSOϭ 0.2, see Table 3). With the use of an

appropriate correction for x_DMSOǞ 0.0 (Table 2), we esti-

mate log β₁₂₀ϭ 13.5 for pure water. This value is in good

agreement with log β₁₂₀ϭ 14.0 calculated from the re-

duction potential and from log β₁₂₀ϭ 36.9 for [Fe^IIIL₂]^3Ϫ

(Scheme 5).

Table 3. Linear free energy relation for Zn^2ϩand Fe^2ϩcomplex for-

mation

Zn^2ϩ

Fe^2ϩ

log β (Zn^2ϩ)/log β (Fe^2ϩ

)

log β₁₁₀log β₁₂₀log β₁₁₀log β₁₂₀

log β₁₁₀

1.11

1.17

1.21

1.16

log β₁₂₀

1.15

1.19

1.20

1.17

dipic^2Ϫ[a]6.35^[b]

11.88^[b]5.71^[b]

10.36^[b]

11.98^[c]

9.82^[d]

14.0^[h]

nta^{3Ϫ [a]}10.45^[c]14.24^[c]8.90^[c]

ida^{2Ϫ [a]}6.58^[d]

11.8^[d]

16.0^[f]

5.45^[d]

11.5^[g]

L^3Ϫ

13.3^[e]

[a]

Abbreviation of ligands: nitrilotriacetic acid (H₃nta), iminodia-

cetic acid (H₂ida), dipicolinic acid (H₂dipic). ^[b]From ref.^[36], 20 °C,

I ϭ 0.1 . From reference ^[36], 25 °C, I ϭ 0.1 . From ref.^[36]

,

[c]

[d]

25 °C, I ϭ 1.0 . From ref.^[17], L ϭ L^x, x_DMSOϭ 0.20, 25 °C,

[e]

[f]

I ϭ 0.1 .

Ref.^[17]listed the value log β₂ϭ 17.5 for L ϭ L^x,

x

_DMSOϭ 0.20. This value was adjusted for pure H₂O (x_DMSOϭ 0)

[g]

by adding an increment of Ϫ1.5 (see Table 2). This work, L ϭ

L^y, x_DMSOϭ 0.20, 25 °C, I ϭ 0.1 .^[h]L ϭ L^x, estimated for H₂O

(25 °C, I ϭ 0.1 ) from the reduction potential E_1/2(FeL₂) ϭ Ϫ0.58

V and log β₁₂₀(Fe^3ϩ) ϭ 36.9 (see Scheme 5 and Table 2).

The low stability of [Fe^IIL₂]^4Ϫ(L ϭ L^x, L^y) is an import-

ant result from this investigation. This complex is not only

a very strong reducing agent, but also dissociates very easily

to give the 1:1 complex. The individual association constant

for the second ligand has a value of log K₂ഠ 4, and the

Figure 10. Sequestering ability of ICL670 in H₂O (0.1  KCl, 25

°C) as a function of pH; the equilibrium concentration of the vari-

pH-dependent constant according to Equation (8), n ϭ 2,

4Ϫ Ϫ1

K ϭ [Fe^IIL^Ϫ] ϫ [H₂L^Ϫ] ϫ [Fe^IIL₂

]

ϫ [H^ϩ]^Ϫ2, is log

ous species is shown as a fraction of total L concentration (ϭ L_tot

)

with (a) L_totϭ 100 µM, (b) L_totϭ 1 µM, and (c) L_totϭ 10 nM;

the solubility product K_S0(FeOOH) ϭ 10^Ϫ39⁴corresponds to the

solubility of ferrihydrite, (ref.^[1]); the curves for iron-containing

species are in bold

K ϭ Ϫ15.4. At pH 7, this equilibrium is clearly shifted to

the right, and [Fe^IIL₂]^4Ϫis only stable at a relatively high

ligand concentration and high pH. The very low stability of

Eur. J. Inorg. Chem. 2004, 4177Ϫ4192

www.eurjic.org

 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

4187

S. Steinhauser, U. Heinz, M. Bartholomä, T. Weyhermüller, H. Nick, K. Hegetschweiler

FULL PAPER

present stage, it is not possible to derive any conclusive in-

formation about the nature of such species.

The formation of polynuclear species is another possibil-

ity for the stabilization of the 1:1 complex at the ferrous

stage.^[2]As already noted by Ryabukhin, the reaction of

H₂L^zwith divalent transition metal cations resulted in the

precipitation of insoluble polymers.^[21]A bis-bidentate co-

ordination mode (binding of a first metal cation to N1 and

O2, and a second cation to N3 and O1, see Figure 2) was

proposed for the resulting {M(L^z)}_nstructure. In turn, each

metal center would be chelated by two ligand entities. For

H₃L^x, the corresponding oligomers [{Fe^II(L^x)}_n]^nϪwould

carry a negative charge at physiological pH, and they could

therefore stay in solution. However, the smooth reaction of

Cu(L^z) with pyridine (Figure 4) demonstrates that such ag-

gregates depolymerize readily in the presence of an ad-

ditional coordinating agent.

With regard to the role of Fe^IIIϪICL670 complexes as

possible catalysts for redox cycling, one can clearly state

that the bis complex [Equation (7)] is redox inactive in any

physiological medium. If partial dissociation and formation

of the ternary [Fe^II(L^x)(A)]^(1ϩm)Ϫspecies at the ferrous

stage is considered, then the result is unchanged, as long as

the ligating agent A^mϪfurther stabilizes the ferric form.

Scheme 5. Synopsis of complex stability and redox chemistry for

iron complexes of H₃L^xin H₂O (0.1  KCl, 25 °C); (a) this work

(section 2.4), corrected for pure H₂O; (b) from ref.^[18b]; (c) this work

(see Table 2); (d) calculated as E_1/2(FeL) ϭ E⁰(Fe) Ϫ 0.059V ϫ (log

This is indeed the case for many biological low-molecular-

weight ligands such as citrate or ATP.

Although the reduction potential of [Fe^IIIL] cannot be

measured directly due to immediate decomposition of the

complex after reduction in acidic solution (2.5.1), it can be

calculated from the corresponding formation constants β₁₁₀

III

II

III

β₁Ϫ log β₁^II); (e) calculated as log β₂ϭ log β₂Ϫ [E⁰(Fe) Ϫ

E_1/2(FeL₂)]/0.059V; (f) this work (section 2.5.2)

the ferrous bis complex is attributed (a) to an unfavorable for [Fe^IIL]^Ϫand [Fe^IIIL]. The estimated E_1/2for the 1:1

electrostatic repulsion between the two trianionic ligand complex is ϩ0.07 V (versus NHE), a value which is in good

entities, and (b) to the distinct selectivity of this ligand for agreement with the irreversible cathodic wave (Figure 8) ob-

small metal ions. As already discussed in section 3.1, all served in acidic media in the range from ϩ100 mV to Ϫ150

complexes of ICL670 with di- and trivalent transition metal mV (versus Ag/AgCl). Clearly, [Fe^IIL]^Ϫis a significantly

cations must be regarded as significantly strained, and this weaker reducing agent than [Fe^IIL₂]^4Ϫ. The redox potential

strain increases with increasing ionic size. The addition of of the 1:1 complex falls in a similar range to that of the

one electron to the ferric complex results in an increase in Fe^III/IIϪEDTA couple (Figure 1).

the effective ionic radius from 0.645 A (high-spin Fe^III) to

˚

II [40]

˚

0.780 A (high-spin Fe ).

In addition to the complete dissociation of one ligand

from [Fe^IIL₂]^4Ϫ, the formation of a species with a partially

coordinated and partially protonated ligand (e.g.

4. Conclusions

A survey of the various formation constants and corre-

[Fe^II(L^x)(HL^x)]^3Ϫ) must be taken into account. Moreover, sponding redox potentials of the Fe^IIand Fe^IIIcomplexes

it should be noted that [Fe^IIL]^Ϫis susceptible to the binding with ICL670 (H₃L^x) in H₂O is summarized in Scheme 5.

of another ligating agent A^mϪwhich may be present in the The other ligands (H₃L^yand H₂L^z) behave very similarly.

medium. Formation of such

a

ternary complex Our study confirmed that these ligands are highly selective

[Fe^II(L^x)(A)]^(1ϩm)Ϫappears to be particularly likely if this for Fe^III, and that the affinity for Fe^IIis, in general, low.

ligating agent could stabilize the ferrous form. The simple, However, there is a distinct difference between the 1:1 and

pH-independent reduction potential [Equation (7)] must 1:2 complexes. [Fe^II(L^x)]^Ϫis sufficiently stable to be formed

therefore be determined at a sufficiently high pH.^[41]It is at moderate ligand concentrations around pH 6, whereas

noteworthy that a cyclic voltammogram of the bis complex [Fe^II(L^x)₂]^4Ϫis generally not stable in aqueous solution. In

recorded at pH 12.6 showed a quasi-reversible electron view of a possible role in redox catalysis, reduction of

transfer, whereas at lower pH, additional peaks were ob- [Fe^III(L^x)₂]^3Ϫis not possible in a physiological medium, and

served for the anodic wave at more positive potentials (Fig- the 1:2 complex is thus redox inactive. On the other hand,

ure S15). This observation is a clear indication that ad- the 1:1 complex [Fe^II(L^x)]^Ϫmust be regarded as a somewhat

ditional species (protonation products of [Fe^IIL₂]^4Ϫor ter- milder reducing agent, and its formation from the ferric

nary mixed ferrous complexes) are formed. However, at the species could well be anticipated in the presence of a bio-

4188

 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org

Eur. J. Inorg. Chem. 2004, 4177Ϫ4192

Complex Formation of ICL670 and Related Ligands with Fe^IIIand Fe^II

FULL PAPER

·Ϫ

4-[3,5-Bis(2-hydroxyphenyl)-1,2,4-triazol-1-yl]benzoic

Acid

(H₃L^x):^[19]4-Hydrazino-benzoic acid (1.75 g, 11.5 mmol) and NEt₃

(1.16 g, 11.5 mmol) were dissolved in boiling EtOH (80 mL). To

the clear solution was added 2-(2-Hydroxyphenyl)-4H-3,1-

benzoxazin-4-one (2.50 g, 10.45 mmol), and the reaction mixture

was refluxed for an additional 2 h. The solution was then allowed

to cool to room temperature, and H₂O was added until some per-

turbation (i.e. a first sign of precipitation) was observed. The mix-

ture was concentrated to a total volume of 50% under reduced

pressure and aqueous 6  HCl (40 mL) was added. The resulting

logical electron donor. In particular, reduction by O₂

could occur (Figure 1), and reoxidation by H₂O₂would also

be possible. However, the formation of [Fe^III(L^x)] is only

relevant in acidic solutions. At pH 7Ϫ8, a ligand redistri-

bution reaction, followed by precipitation of solid iron hy-

droxide according to: 2[Fe^III(L^x)] ϩ 3OH^ϪǞ FeOOH(s) ϩ

[Fe(L^x)₂]^3Ϫϩ H₂O, would be favored (Figure 10). There-

fore, it appears rather unlikely that any possible enhance-

ment of oxidative stress (Fenton chemistry etc.) needs to be

considered under physiological conditions. This is clearly solid was filtered and dried for 24 h in vacuo (3.11 g, 8.34 mmol,

different at lower pH values, where [Fe^III(L^x)₂]^3Ϫwill par-

80%). C₂₁H₁₅N₃O₄(373.37): calcd. C 67.56, H 4.05, N 11.25; found

1

tially dissociate to [Fe^III(L^x)], and this species can be re-

C 67.76, H 3.85, N 11.14. H NMR ([D₆]DMSO): δ ϭ 6.90 (d, 1

duced to [Fe^II(L^x)]^Ϫ. As a matter of fact, the partial dis-

H), 7.00Ϫ7.04 (m, 3 H), 7.37 (m, 2 H), 7.55 (d, 1 H), 7.58 (d, 2

H), 8.02 (d, 2 H), 8.08 (d, 1 H) 10.05 (s, OH), 10.81 (s, OH), 13.20

sociation of Fe^IIIcomplexes in slightly acidic solutions ap-

(broad, OH) ppm. ¹³C NMR ([D₆]DMSO): δ ϭ 113.6, 114.4, 116.2

pears to be a general problem for di- and tridentate iron

(CH), 117.0 (CH), 119.4 (CH), 119.6 (CH), 123.2 (2 CH), 126.7

chelators such as L1 or ICL670.^[42]This is, of course, differ-

(CH), 130.2 (2 CH), 130.5, 131.0 (CH), 131.4 (CH), 132.5, 141.2

ent for ligands with a higher denticity such as DFO, where

(CH), 152.0, 155.2, 156.4, 159.9, 166.4 ppm. UV/Vis (H₂O/DMSO,

x_DMSOϭ 0.2, pH Ͻ2, 25 °C): λ_max(ε) ϭ 284 nm (1.5 ϫ 10⁴), 296

(sh). FAB^ϩ-MS: m/z ϭ 374 (100%, [H₄L^x]^ϩ). Single crystals of

H₃L^xwere grown by slow evaporation of a MeOH solution at room

temperature. The solid was allowed to stand exposed to the air to

form the monohydrate H₃L^x·H₂O. C₂₁H₁₇N₃O₅(391.39): calcd. C

64.45 H 4.38, N 10.74; found C 64.11, H 4.60, N 10.44.

an ‘‘all-or-nothing’’ coordination mode prevents possible

formation of partially coordinated, ‘‘dangerous’’ species.

Experimental Section

Materials, Instrumentation and Analyses: The chemicals used for

the synthetic work were commercially available products of reagent

grade quality and were used as obtained. H₂O was distilled twice

(quartz apparatus). DMSO was from Fluka (puriss., p.a.). MeOH

and EtOH (reagent grade) were purified by distillation prior to use.

For the potentiometric and spectrophotometric titrations, metal

salts of highest available quality (Ͼ99.95%) were used. FeCl₃(anhy-

drous) was from Fluka, (NH₄)₂Fe(SO₄)₂·6H₂O (Ͼ99.997%) from

Aldrich. ¹H and ¹³C{¹H} NMR spectra were measured in

[D₆]DMSO (28 °C) with a Bruker DRX 500 spectrometer (reson-

ance frequencies: 500.13 MHz for ¹H and 125.9 MHz for ¹³C).

Chemical shifts (in ppm) are given relative to tetramethylsilane as

an internal standard (ϭ 0 ppm). EI^ϩ-MS were measured on a

Varian MAT 311 spectrometer (Dr. Graf, Universität des Saar-

landes), and FAB^ϩ-MS were recorded on a VG ZAB-VSEQ spec-

trometer (Dr. W. Amrein, R. Häfliger, ETH Zürich). Samples for

FAB were dissolved in EtOH with 3-nitrobenzyl alcohol as a ma-

trix. C,H,N analyses were performed by H. Feuerhake (Universität

des Saarlandes).

4-[3,5-Bis(2-hydroxyphenyl)-1,2,4-triazol-1-yl]benzosulfonic

Acid

(H₃L^y):^[20]4-Hydrazino-benzosulfonic acid (2.4 g, 12.75 mmol) was

suspended in EtOH (300 mL). The solid was dissolved by the ad-

dition of one equivalent of NEt₃(1.29 g, 12.75 mmol). The solution

was refluxed, and 2-(2-hydroxyphenyl)-4H-3,1-benzoxazin-4-one

(3.00 g, 12.5 mmol) was added. Heating was continued for an ad-

ditional 20 h. The clear solution was then allowed to cool to room

temperature, and 6  aqueous HCl (60 mL) was added. The re-

sulting acidic solution was evaporated under reduced pressure until

a white solid started to precipitate. A further portion of 6  aque-

ous HCl (60 mL) was added. The solid was then separated by fil-

tration and washed with a small portion of cold H₂O, and dried in

vacuo over P₄O₁₀for 24 h (4.37 g, 10.67 mmol, 85%). C₂₀H₁₅N₃O₅S

(409.42): calcd. C 58.67, H 3.69, N 10.26; found C 58.99, H 3.90,

1

N 9.88. H NMR ([D₆]DMSO): δ ϭ 6.91 (d, 1 H), 6.95Ϫ7.01 (m,

2 H), 7.05 (d, 1 H), 7.36Ϫ7.40 (m, 2 H), 7.46 (d, 2 H), 7.51 (d, 1

H), 7.72 (d, 2 H), 8.02 (d, 1 H) ppm. ¹³C NMR (DEPT90,

[D₆]DMSO): δ ϭ 113.3, 113.4, 116.4 (CH), 117.1 (CH), 119.5 (CH),

119.8 (CH), 123.4 (2 CH), 126.7 (2 CH), 127.4 (CH), 131.2 (CH),

131.9 (CH), 132.9 (CH), 137.7, 148.1, 151.6, 155.5, 156.4, 158.4

ppm. UV/Vis (H₂O; pH Ͻ2; 25 °C): λ_max(ε) ϭ 255 (1.4 ϫ 10⁴);

295 nm (1.2 ϫ 10⁴). FAB^ϩ-MS: m/z ϭ 410 (100) [H₄L^y]^ϩ. The

solid was allowed to stand exposed to the air, which resulted in the

formation of the monohydrate H₃L^y·H₂O. C₂₀H₁₇N₃O₆S (427.44):

calcd. C 56.20, H 4.01, N 9.83; found C 56.02, H 4.24, N 9.61.

2-(2-Hydroxyphenyl)-benzo-4H-[1,3]-oxazin-4-one:^[43]Salicylic acid

(24.9 g, 0.18 mol), salicylamide (20.6 g, 0.15 mol) and pyridine (1.5

mL) were refluxed in xylene (30 mL). Thionyl chloride (23.7 mL,

0.33 mol) was added with vigorous stirring over a period of 4 h.

An intense evolution of SO₂and HCl was noted. At the end of the

addition, the product started to crystallize. Stirring was continued

for an additional 30 min, and the xylene was removed by distil-

lation at reduced pressure. The resulting solid residue was sus-

pended in EtOH (60 mL) and acetic acid (1.5 mL). The mixture

was heated gently and then allowed to cool to 20 °C. The precipi-

3,5-Bis(2-hydroxyphenyl)-1-phenyl-1,2,4-triazole (H₂L^z):^[43]Phenyl-

hydrazine hydrochloride (1.45 g, 10.03 mmol) was dissolved in boil-

ing EtOH (80 mL). 2-(2-hydroxyphenyl)-4H-3,1-benzoxazin-4-

tate was filtered and recrystallized from 2-methoxyethanol (70 mL). one (2.0 g, 8.36 mmol) was added, and the resulting clear yellow

Yield: yellow needles (19.78 g, 82.7 mmol, 55%). The product

decomposed slowly when exposed to moisture, forming bis(salicyl)-

solution was heated under reflux for an additional 2 h. The yellow

color disappeared, and the colorless solution was allowed to cool

imide. C₁₄H₉NO₃(239.2): calcd. C 70.29, H 3.79, N 5.85; found C to room temperature. Addition of aqueous 1  HCl (40 mL) re-

1

70.80, H 3.88, N 5.70. H NMR ([D₆]DMSO): δ ϭ 7.09 (m, 2 H),

7.62 (m; 2 H), 7.78 (d, 1 H), 7.94 (t, 1 H), 8.07 (d, 1 H), 8.20 (d, 1

H), 8.78 (OH) ppm. ¹³C NMR ([D₆]DMSO): δ ϭ 111.4, 117.4,

sulted in precipitation of a white solid, which was filtered off and

dried for 12 h in vacuo. Recrystallization from hot MeOH yielded

a crystalline material that was suitable for single-crystal X-ray

117.8, 117.9, 119.5, 126.7, 127.2, 129.0, 136.0, 136.7, 153.8, 161.8, analysis. Yield: (1.72 g, 5.23 mmol, 62.5%). C₂₀H₁₅N₃O₂(329.36):

163.4, 164.8 ppm. UV/Vis (H₂O, 25 °C): λ_maxϭ 371, 323 nm. calcd. C 72.94, H 4.59, N 12.76; found C 72.61, H 4.36, N 12.64.

Eur. J. Inorg. Chem. 2004, 4177Ϫ4192

www.eurjic.org

 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

4189

S. Steinhauser, U. Heinz, M. Bartholomä, T. Weyhermüller, H. Nick, K. Hegetschweiler

FULL PAPER

¹H NMR ([D₆]DMSO): δ ϭ 6.89 (d, 1 H), 6.94 (t, 1 H), 6.98Ϫ7.04 of pyrogallol to remove traces of O₂. The other titrations were per-

(m, 2 H), 7.37 (t, 2 H), 7.42Ϫ7.48 (m, 6 H), 7.85 (d, 1 H), 9.88 (s, formed under N₂, which was passed through a 0.1  aqueous KCl

OH), 10.71 (s, OH) ppm. ¹³C NMR ([D₆]DMSO): δ ϭ 113.7, 114.5,

solution. The apparent ionization constant of the different H₂O/

116.0, 116.9, 119.1, 119.4, 123.6 (2 CH), 126.5, 128.5, 129.0 (2 CH), DMSO media was determined by multiple acidϪbase titrations.

130.8, 131.1, 132.1, 137.7, 151.6, 155.3, 156.2, 159.4 ppm. UV/Vis The pK_as of the ligands were measured using solid, analytically

(H₂O/DMSO, x_DMSOϭ 0.2; pH Ͻ7; 25 °C): λ_max(ε) ϭ 255 (1.5 ϫ pure samples of H₃L^x, H₃L^y, and H₂L^z. The formation constants

10⁴); 300 nm (1.2 ϫ 10⁴). MS-EI^ϩ: m/z ϭ 329.4 (100) [H₂L^z]^ϩ.

of the ferric complexes were determined using stock solutions of

the ligands (0.01 ), which were standardized by acidϪbase ti-

trations, and stock solutions of Fe^III, which were standardized

photometrically as tris-phenanthroline-Fe^II.^[47]Additional acid

(HCl or HNO₃) was added to the Fe^IIIstock solutions to prevent

partial hydrolysis prior to the titration. Complex formation with

Fe^IIIwas found to be sluggish, and additional back titrations were

performed to ensure complete equilibration. A measuring time of

800Ϫ920 s per point proved necessary to avoid a significant hyster-

esis. Total ligand and total Fe concentrations ranged from 5ϫ10^Ϫ4

to 2ϫ10^Ϫ3.^[48]

Na[Fe(L^z)₂]·2H₂O·C₂H₅OH: FeCl₃(324 mg, 2 mmol) dissolved in

abs. EtOH (15 mL) was added to a solution of H₂L^z(660 mg,

2 mmol) in ethanol (15 mL). The resulting solution turned deep

purple, indicating the formation of the 1:1 complex [Fe(L^z)]^ϩ. In a

separate reaction, solid Na was added to abs. EtOH to satu-

ration.This NaOEt solution (1 mL) was then layered with the

Fe^IIIϪL^xsolution (0.5 mL). After some hours, the deposition of

deep red crystals was observed at the boundary of the two phases.

It is important that the crystals remained in this intermediate me-

dium and do not sink into the NaOEt solution. They were sub-

sequently removed and immediately cooled to 100 K for the X-ray

diffraction study. The single crystal analysis established the compo-

sition as Na[Fe(L^z)₂]·4C₂H₅OH. However, the crystal disintegrated

rapidly and the compound that formed had the composition

Na[Fe(L^z)]·2H₂O·C₂H₅OH. C₄₂H₃₆FeN₆NaO₇(815.63): calcd. C

61.85, H 4.45, N 10.30; found C 62.49, H 4.15, N 10.02. FAB^ϩ-

Spectrophotometric Titrations: The titration cell used for potent-

iometry was equipped with an immersion probe (HELLMA),

which was connected to a diode array spectrophotometer (J&M,

TIDAS-UV/NIR/100Ϫ1). A PC was used to trigger the recording

of a spectrum just prior to the addition of each new aliquot of

titrant.^[45]For the pK_adetermination of H₃L^x, H₃L^y, and H₂L^z, the

total ligand concentration was about 10^Ϫ5Ϫ10^Ϫ4, and absorption

data in the range 250 Ͻ λ Ͻ 400 nm were collected. For the forma-

tion constants of the Fe^IIIcomplexes, total [Fe] was in the range

9.1ϫ10^Ϫ5Ϫ3.5ϫ10^Ϫ4, and total [L] in the range

9.1ϫ10^Ϫ5Ϫ9.7ϫ10^Ϫ4. The pK_a,1of H₃L^yand the formation con-

stants of [Fe(HL^x)]^ϩ, [Fe(L^y)], and [Fe(L^z)]^ϩwere determined by a

batch method with 10Ϫ40 individually sealed 10-mL samples in

the range 1 Ͻ pH Ͻ 2 (25.0 °C) allowing an equilibration time

of 24 h.^[48]

MS: m/z

ϭ

709.5Ϫ714.5 (20%, superposition of [Fe(L^z)₂]^ϩ,

[Fe(L^z)(HL^z)]^ϩ, [Fe(HL^z)₂]^ϩ), 731.5Ϫ736.5 (100%, superposition of

[NaFe(L^z)₂]^ϩand [NaFe(L^z)(HL^z)]^ϩ), 754.5Ϫ758.5 (79%,

[Fe(L^z)₂(EtOH)]^ϩ).

Na[Al(L^z)₂]·H₂O·C₂H₅OH: The procedure reported for the prep-

aration of the corresponding Fe^IIIcomplex was used. Single crys-

tals of composition Na[Al(L^z)₂]·4C₂H₅OH were obtained. They

disintegrated rapidly to form Na[Al(L^z)₂]·H₂O·C₂H₅OH.

C₄₂H₃₄AlN₆NaO₆(768.75): calcd. C 65.62, H 4.46, N 10.93; found

C 65.73, H 4.20, N 10.56.

Calculations of Equilibrium Constants: The computer programs

HYPERQUAD^[49]and SPECFIT^[50]were used to evaluate the

potentiometric and spectrophotometric data, respectively. All equi-

librium constants were calculated as concentration constants, and

pH was defined as Ϫlog [H^ϩ]. For pure aqueous solutions, a pK_w

of 13.78 was used.^[36]In H₂O/DMSO solutions, the apparent ioniz-

ation constant Q_cϭ [H^ϩ]ϫ[OH^Ϫ] depended linearly on x_DMSOin

the range 0.1 Յ x_DMSOՅ 0.2 with pQ_cϭ 14.65 (x_DMSOϭ 0.10)

and pQ_cϭ 15.59 (x_DMSOϭ 0.20). These values are in good agree-

ment with those reported in the literature.^[27]A graphical represen-

tation of pQ_cversus x_DMSOis shown in the supporting information

(Figure S2). The pK_w(or pQ_c) and the total concentrations of the

reactants were always treated as fixed values. Additionally, the pro-

tonation constants of the ligands were fixed when refining forma-

Cu(L^z):^[21]CuBr₂(223 mg, 1 mmol) was added to a solution of

H₂L^z(329 mg, 1 mmol) in EtOH (10 mL). The mixture was stirred

until a green solid precipitated. The precipitation was completed

by adding a concentrated solution of NaOMe in MeOH. Yield

290 mg (742 µmol, 74%). C₂₀H₁₃CuN₃O₂(390.89): calcd. C 61.46,

H 3.35, N 10.75; found C 61.96, H 3.35, N 10.73.

[Cu(L^z)(pyridine)]₂: The procedure, reported by Uhlemann and co-

workers^[44]for the preparation of a Cu^IIcomplex with a related

tridentate [O,N,O]-ligand, was used with some minor modifi-

cations: Solid Cu(L^z) (290 mg, 742 µmol) was dissolved in boiling

pyridine (20 mL). This solution (2 mL) was mixed with MeOH (1

mL). H₂O (100 µL) was then layered on the resulting green solu-

tion. Dark green crystals of [Cu(L^z)(pyridine)]₂suitable for single-

crystal X-ray analysis grew within 24 h at 4 °C.

tion

constants

of

metal-containing

species.

Because

ϩ

Fe(NH₄)₂(SO₄)₂·6H₂O was used as a Fe^IIsource, the pK_aof NH₄

was determined to be 8.90 (x_DMSOϭ 0.2). However, the formation

of [Fe^II(L^y)]^Ϫ, which was investigated in the range of 5.9 Ͻ pH Ͻ

6.9, did not interfere with the deprotonation of NH₄^ϩ. For the

evaluation of the spectrophotometric titrations, the spectrum of

free Fe^3ϩwas measured separately for each H₂O/DMSO ratio, and

was then imported and treated as fixed during refinement. Only

absorption data with λ Ͼ400 nm were used for the evaluation of

metal complex formation, and the free ligands and their pro-

tonation products could therefore be considered as colorless (see

Figures S3 and S4).

Potentiometric Measurements: All titrations were carried out at an

ionic strength of 0.1  (KCl or KNO₃) using a Metrohm 665 piston

burette, a Metrohm 6.0222.100 glass electrode with an incorporated

Ag/AgCl reference (aqueous), and a Metrohm 713 pH/mV meter.

The pH meter and the piston burette were controlled by a PC.^[45]

The samples (50 mL) were placed in a double-jacketed beaker, and

the temperature was maintained at 25.0 Ϯ 0.1 °C using a Lauda

ecoline103 thermostat. KOH (0.10 , Merck Titrisol) having the

same H₂O/DMSO molar ratio as the corresponding sample solu-

tion was used as titrant. The H₂O/DMSO samples were prepared

using published values of the partial molar volumes of the two

components (the H₂O/DMSO mixtures show a significant devi-

ation from ideal behavior).^[46]Titrations of the Fe^IIcomplexes were

Electrochemistry: Cyclic voltammograms were recorded in H₂O or

H₂O/DMSO (x_DMSOϭ 0.2) using 0.1  KCl as supporting electro-

lyte at ambient temperature (23 Ϯ 3 °C). A BAS C2 cell equipped

performed under Ar, which was scrubbed with an alkaline solution with a BAS 100B/W 2 potentiostat, a Pt counter electrode, and an

4190  2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.eurjic.org

Eur. J. Inorg. Chem. 2004, 4177Ϫ4192

Complex Formation of ICL670 and Related Ligands with Fe^IIIand Fe^II

FULL PAPER

Table 4. Crystallographic data for ICL670 (H₃L^x), Na[Fe(L^z)₂]·4EtOH, and [Cu(L^z)(pyridine)]₂

ICL670

Na[Fe(L^z)₂]·4EtOH

[Cu(L^z)(pyridine)]₂

Empirical formula

Molecular weight

Crystal system

Space group

C₂₁H₁₅N₃O₄

373.36

monoclinic

P2₁/c (No. 14)

8.900(2)

26.946(5)

7.558(2)

C₄₈H₅₀FeN₆NaO₈

917.78

monoclinic

C2/c (No. 15)

23.933(4)

8.7399(12)

24.166(4)

C₅₀H₃₆Cu₂N₈O₄

939.95

monoclinic

P2₁/c (No. 14)

10.944(2)

10.476(2)

18.419(4)

˚

a [A]

˚

b [A]

˚

c [A]

β [°]

94.77(3)

1806.3(7

4

107.84(3)

4811.8(13)

4

100.05(3)

2079.3(7)

2

3

˚

V [A ]

Z

T [K]

215(2)

1.373

0.097

100(2)

1.267

0.380

293(2)

1.501

1.082

D_calcd.[g cm^Ϫ3

]

µ [mm^Ϫ1

]

Crystal size [mm]

θ_min, θ_max

Trans (min., max)

Data set

Total/Unique data

Parameter/Restraints

R_1,wR₂, [I Ͼ 2σ(I)]

0.3 ϫ 0.2 ϫ 0.2

2.42, 24.13

Ϫ

Ϫ10/10; Ϫ30/30; Ϫ8/8

11218, 2810

313/0

0.0468, 0.1033

0.0755, 0.1133

0.227/Ϫ0.218

0.35 ϫ 0.18 ϫ 0.18

1.79, 24.00

0.5567, 0.9625

Ϫ28/24; Ϫ10/9; Ϫ27/26

9450, 3738

269/9

0.0870, 0.2352

0.1456, 0.2832

1.240/Ϫ0.550

0.08 ϫ 0.05 ϫ 0.02

2.25, 23.99

Ϫ

Ϫ11/11; Ϫ11/11; Ϫ20/21

12786, 3077

289/0

0.0599, 0.1143

0.1595, 0.1450

0.375/Ϫ0.442

R_1,wR₂, (all data)

Max. peak/hole [e/ A ]

3

˚

Ag/AgCl reference electrode was used. The total iron concentration

refined isotropically with U_isoϭ 1.3 ϫ U_eqof the oxygen atom. All

was 2.5 m. Measurements at pH Ͻ4 (E ϾϪ200 mV vs. Ag/AgCl)

other H atom positions of the two metal complexes were calculated

were performed using a Pt working electrode. For the high pH (riding model). Isotropic displacement parameters were used with

samples (pH Ͼ6, E ϽϪ400 mV vs. Ag/AgCl) a hanging Hg drop

working electrode was used.

U_isoϭ 1.5 ϫ U_eq(C) for methyl groups and U_isoϭ 1.2 ϫ U_eq(C)

for all other hydrogen atoms. The H atom positions of H₃L^xcould

all be located. They were successfully refined with variable isotropic

displacement parameters.

Crystal Structure Determination:^[51]X-ray diffraction data of H₃L^x

(215 K), Na[Fe(L^z)₂]·4EtOH (293 K),^[23]and [Cu(L^z)(pyridine)]₂

(293 K) were collected on a STOE IPDS diffractometer. Data col-

lection of Na[Al(L^z)₂]·4EtOH was performed on

a Siemens

Acknowledgments

SMART CCD diffractometer at 293(2) K. Additionally, a data set

of the Fe complex was collected at 100(2) K on the SMART CCD

diffractometer. Graphite monochromated Mo-K_αradiation (λ ϭ

X-ray diffraction data for H₃L^x(ICL670), Na[Fe(L^z)₂]·4EtOH

(293 K) and [Cu(L^z)(pyridine)]₂were recorded by Dr. Volker Huch

(Saarbrücken). The structure of [Cu(L^z)(pyridine)]₂was solved and

refined by Dr. Jürgen Sander (Saarbrücken). Some of the CV meas-

urements were performed by Michael Krämer (Saarbrücken). We

thank Prof. G. Schwitzgebel (Saarbrücken, retired) for helpful

discussions in interpreting the CV measurements, and Dr. Peter

Osvath (Melbourne) for valuable advice.

˚

0.71073 A) was used for all structures. Crystallographic data for

H₃L^x, the Fe complex at 100 K, and the Cu complex are summar-

ized in Table 4. The Al complex proved to be isotypic with the Fe

complex; however, the reflections were rather weak, and the data

could not be refined satisfactorily.^[24]All data sets were corrected

for Lorentz and polarization effects, and an absorption correction

was applied to the data of the Fe complex collected at 100 K. Mini-

mal and maximal transmission factors were 0.5567 and 0.9625. All

structures were solved by direct methods (SHELXS-97) and refined

by full-matrix least-squares calculations on F²(SHELXL-97).^[52]

Anisotropic displacement parameters were refined for all non-hy-

drogen atoms of H₃L^xand [Cu(L^z)(pyridine)]₂. In the low-tempera-

ture structure of the Fe complex, the [Fe(L^z)₂]^Ϫanion was modeled

as a superposition of two orientations, having the Fe center, the

phenoxo rings, and the CϪNϪC moieties of the triazole rings in

common. For the two outer nitrogen atoms (N2 and N3) and the

peripheral phenyl ring, a major (66%) and a minor (34%) compo-

nent were located. Non-hydrogen atoms of the nondisordered part

and of the major component (except N2 and N3) of the anion, as

well as the Na position and the C and O positions of a nondis-

ordered EtOH molecule were refined with anisotropic displacement

parameters. All C and N atoms of the minor component, N2 and

N3 of the major component, and the C and O position of a second,

disordered EtOH molecule were refined isotropically. The hydroxy

hydrogen atom H(80) of the nondisordered EtOH was located and

[1]

R. Crichton, Inorganic Biochemistry of Iron Metabolism Ϫ

From Molecular Mechanisms to Clinical Consequences, 2nd

Edition, John Wiley, Chichester, 2001.

[2] [2a]

Z. D. Liu, R. C. Hider, Coord. Chem. Rev. 2002, 232, 151.

[2b]

Z. D. Liu, R. C. Hider, Med. Res. Rev. 2002, 22, 26.

[3]

R. C. Hider, Z. D. Liu, Curr. Med. Chem. 2003, 10, 1051.

E. Nisbet-Brown, N. F. Olivieri, P. J. Giardina, R. W. Grady,

[4]

E. J. Neufeld, R. Sechaud, A. J. Krebs-Brown, J. R. Anderson,

D. Alberti, K. C. Sizer, D. G. Nathan, Lancet 2003, 361, 1597.

[5] [5a]

C. Hershko, A. M. Konijn, H. P. Nick, W. Breuer, Z. I.

[5b]

Cabantchik, G. Link, Blood 2001, 97, 1115.

H. Nick, P.

Acklin, R. Lattmann, P. Buehlmayer, S. Hauffe, J. Schupp, D.

Alberti, Curr. Med. Chem. 2003, 10, 1065.

R. S. Britton, K. L. Leicester, B. R. Bacon, Int. J. Hematol.

2002, 76, 219.

[6]

[7]

[8]

[9]

S. I. Liochev, Met. Ions Biol. Syst. 1999, 36, 1.

M. J. Burkitt, Prog. React. Kinet. Mech. 2003, 28, 75.

H. B. Dunford, Coord. Chem. Rev. 2002, 233Ϫ234, 311.

[10] [10a]

[10b]

W. H. Koppenol, Redox Rep. 2001, 6, 229.

ochev, I. Fridovich, Redox Rep. 2002, 7, 55.

S. I. Li-

Eur. J. Inorg. Chem. 2004, 4177Ϫ4192

www.eurjic.org

 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

4191

S. Steinhauser, U. Heinz, M. Bartholomä, T. Weyhermüller, H. Nick, K. Hegetschweiler

FULL PAPER

[11]

[29]

W. H. Koppenol, Redox Rep. 2001, 6, 339.

A. E. Martell, R. D. Hancock, Determination and Use of Sta-

[12]

bility Constants, 2nd Edition, VCH Publishers, New York,

1992.

S. Y. Qian, G. R. Buettner, Free Radic. Biol. Med. 1999, 26,

1447.

[30]

[13]

W. R. Harris, K. N. Raymond, F. L. Weitl, J. Am. Chem. Soc.

1981, 103, 2667.

S. J. Stohs, D. Bagchi, Free Radic. Biol. Med. 1995, 18, 321.

[14] [14a]

[14b]

S. Toyokuni, Free Radic. Biol. Med. 1996, 20, 553.

Meneghini, Free Radic. Biol. Med. 1997, 23, 783.

R.

[31] [31a]

[31b]

J. N. Butler, J. Electroanal. Chem. 1967, 14, 89.

D.

[15]

T. Sawyer, J. L. Roberts, Jr., Experimental Electrochemistry for

Chemists, John Wiley, 1974, New York.

B. P. Esposito, W. Breuer, P. Sirankapracha, P. Pootrakul, C.

Hershko, Z. I. Cabantchik, Blood 2003, 102, 2670.

M. Fontecave, J. L. Pierre, Bull. Soc. Chim. Fr. 1993, 130, 77.

U. Heinz, K. Hegetschweiler, P. Acklin, B. Faller, R. Lattmann,

H. P. Schnebli, Angew. Chem. 1999, 111, 2733; Angew. Chem.

Int. Ed. 1999, 38, 2568.

[32] [32a]

[16]

[17]

K. Hegetschweiler, M. Ghisletta, L. Hausherr-Primo, T.

Kradolfer, H. W. Schmalle, V. Gramlich, Inorg. Chem. 1995,

34, 1950. ^[32b]K. Hegetschweiler, M. Weber, V. Huch, M. Veith,

H. W. Schmalle, A. Linden, R. J. Geue, P. Osvath, A. M. Sarge-

son, A. C. Willis, W. Angst, Inorg. Chem. 1997, 36, 4121.

A. E. Martell, R. D. Hancock, Metal Complexes in Aqueous

Solutions, Plenum Press, New York, 1996.

[18] [18a]

[18b]

A.-F. Miller, Compr. Coord. Chem. II 2004, 8, 479.

D.

[33]

[34]

R. Lide (Ed.), CRC Handbook of Chemistry and Physics 83rd

Edition 2002Ϫ2003, CRC-Press, Washington D. C. 2002.

R. Lattmann, P. Acklin (Novartis Pharma AG), PCT Int. Appl.

WO 9749395 A1 1997 [Chem. Abstr. 1998, 128, 114953e].

F. Bachmann, J. Dannacher, G. Schlingloff, M. Hazenkamp,

G. Haensler, K. Hegetschweiler, U. Heinz, (Ciba Specialty

AG), PCT Int. Appl. WO 2003053986 A1 2003 [Chem. Abstr.

2003 139, 87009].

Y. I. Ryabukhin, N. V. Shibaeva, A. S. Kuzharov, V. G. Korob-

kova, A. V. Khokhlov, A. D. Garnovskii, Koord. Khim. 1987,

13, 869; English Translation: Soviet J. Coord. Chem. 1987, 13,

493.

R. G. Pearson, Chemical Hardness, WILEY-VCH, New York,

1997.

[19]

[20]

[35] [35a]

[35b]

R. D. Hancock, Prog. Inorg. Chem. 1989, 37, 187.

D. Hancock, Acc. Chem. Res. 1990, 23, 253.

R.

[36]

R. M. Smith, A. E. Martell, R. J. Motekaitis, Critically Selec-

ted Stability Constants of Metal Complexes, NIST Standard

Reference Database 46; Version 7.0; Gaithersburg, MD, USA

2003.

P. Paoletti, R. Walser, A. Vacca, G. Schwarzenbach, Helv.

Chim. Acta 1971, 54, 243.

[21]

[37]

[38]

[39]

K. Hegetschweiler, Chem. Soc. Rev. 1999, 28, 239.

The minor differences between the two ligands L^xand L^ycan

be ignored for these rather general considerations.

R. D. Shannon, Acta Crystallogr., Sect. A 1976, 32, 751.

It can be assumed that a hypothetical protonated complex

[Fe^II(L^x)(HL^x)]^3Ϫwould in any case be more acidic than the

[22]

[23]

U. Heinz, J. Sander, K. Hegetschweiler, Z. Kristallogr. New

Cryst. Struct. 2001, 216, 113.

Crystal data: C₄₈H₅₀FeN₆NaO₈, a ϭ 24.138(5), b ϭ 8.855(2),

c ϭ 24.597(5), β ϭ 107.77(3), V ϭ 5006.6(18), Z ϭ 4, T ϭ

293(2) K, space group C2/c, reflections: 15286 collected, 3741

unique, 2689 observed with I Ͼ 2σ(I). R₁(observed data) ϭ

0.1224, wR₂(all data) ϭ 0.3476.

Crystal data: C₄₈H₅₀AlN₆NaO₈, a ϭ 23.922(9), b ϭ 8.712(3),

c ϭ 24.168(10), β ϭ 107.26(3), V ϭ 4810.0(32), Z ϭ 4, T ϭ

293(2) K, space group C2/c, reflections: 7299 collected, 2578

unique, 1232 observed with I Ͼ 2σ(I). R₁(observed data) ϭ

0.1402, wR₂(all data) ϭ 0.4545.

[40]

[41]

free ligand HL^2Ϫ

.

[42]

[43]

[44]

[45]

A. El-Jammal, D. M. Templeton, Inorg. Chim. Acta 1996,

245, 199.

Yu. I. Ryabukhin, L. N. Faleeva, V. G. Korobkova, Chem.

Heter. Com. (Engl. Trans.) 1983, 19, 332.

E. Ludwig, U. Schilde, E. Uhlemann, H. Hartl, I. Brüdgam,

Z. Anorg. Allg. Chem. 1996, 622, 701.

Messlabor, a computer program for the control of potentiometric

[24]

[25] [25a]

Z. Travnicek, Z. Sindelar, J. Marek, Z. Kristallogr. New

[25b]

Cryst. Struct. 1997, 212, 125.

I. Vencato, A. Neves, B. R.

and

spectrophotometric

titrations,

Scientec

GmbH,

Vincent, C. Erasmus-Buhr, W. Haase, Acta Crystallogr., Sect.

Saarbrücken, 2002.

C 1994, 50, 386. ^[25c]B. S. Snyder, G. S. Patterson, A. J. Abrah-

[46]

[47]

[48]

J. T. W. Lai, F. W. Lau, D. Robb, P. Westh, G. Nielsen, C.

Trandum, A. Hvidt, Y. Koga, J. Solution Chem. 1995, 24, 89.

A. I. Vogel, A Textbook of Quantitative Inorganic Analysis, 5th

edn, Longmans, London, 1987.

[25d]

amson, R. H. Holm, J. Am. Chem. Soc. 1989, 111, 5214.

T. Weyhermüller, T. K. Paine, E. Bothe, E. Bill, P. Chaudhuri,

Inorg. Chim. Acta 2002, 337, 344. ^[25e]H. Aneetha, K. Panneer-

selvam, T.-F. Liao, T.-H. Lu, C.-S. Chung, J. Chem. Soc., Dal-

ton Trans. 1999, 2689.

Further details of the titration experiments (Tables S1ϪS13

and Figures S3, S4, S6ϪS8) are provided as Supporting Infor-

mation.

[26] [26a]

M. T. Garland, J. Y. Le Marouille, E. Spodine, Acta Crys-

[26b]

[49]

tallogr., Sect. C 1985, 41, 855.

P. Lemoine, D. Nguyen-

P. Gans, A. Sabatini, A. Vacca, Talanta 1996, 43, 1739.

[50] [50a]

Huy, B. Viossat, J.-M. Leger, A. Tomas, Acta Crystallogr., Sect.

C 1999, 55, 2068. ^[26c]R. M. Buchanan, C. Wilson-Blumenberg,

C. Trapp, S. K. Larsen, D. L. Greene, C. G. Pierpont, Inorg.

Chem. 1986, 25, 3070. ^[26d]N. Matsumoto, M. Mimura, Y. Sun-

R. A. Binstead, B. Jung, A. D. Zuberbühler, SPECFIT/32

Version 3.0, Spectrum Software Associates, Marlborough, MA

[50b]

01752, USA, 2000.

H. Gampp, M. Maeder, C. J. Meyer,

A. D. Zuberbühler, Talanta 1985, 32, 95.

[51]

atsuki, S. Eguchi, Y. Mizuguchi, H. Miyasaka, T. Nakashima,

Crystallographic data (excluding structure factors) for

[26e]

Na[Fe(L^z)₂]·4EtOH (293 K), [Cu(L^z)(pyridine)]₂, H₃L^x

Bull. Chem. Soc., Jpn. 1997, 70, 2461.

Y. Sunatsuki, M.

Nakamura, N. Matsumoto, F. Kai, Bull. Chem. Soc., Jpn. 1997,

(ICL670),

Na[Fe(L^z)₂]·4EtOH

(100 K),

and

70, 1851. ^[26f]C. A. Otter, S. M. Couchman, J. C. Jeffery, K. L.

Na[Al(L^z)₂]·4EtOH (293 K) have been deposited with the Cam-

bridge Crystallographic Data Center as supplementary publi-

cation nos. CCDC 237449Ϫ237453. Copies of the data can be

obtained free of charge on application to CCDC, 12 Union

Road, Cambridge CB2 1EZ, UK [Fax (internat.): ϩ44-1223-

336033; E-mail: deposit@ccdc.cam.ac.uk].

V. Mann, E. Psillakis, M. D. Ward, Inorg. Chim. Acta 1998,

[26g]

278, 178.

Y. Sunatsuki, T. Matsuo, M. Nakamura, F. Kai,

N. Matsumoto, J.-P. Tuchagues, Bull. Chem. Soc., Jpn. 1998,

[26h]

71, 2611.

A. Bavoso, L. Menabue, M. Saladini, M. Sola,

Inorg. Chim. Acta 1996, 244, 207. ^[26i]P. Lemoine, A. Mazurier,

I. Billy, D. Nguyen-Huy, B. Viossat, Z. Kristallogr. New Cryst.

Struct. 2000, 215, 523.

[52] [52a]

G. M. Sheldrick, SHELXS-97 Program for Crystal Struc-

ture Solution, Göttingen, 1990.

[52b]

G. M. Sheldrick,

[27] [27a]

E. M. Woolley, L. G. Hepler, Anal. Chem. 1972, 44, 1520.

P. Fiordiponti, F. Rallo, F. Rodante, Z. Phys. Chem. 1974,

SHELXL-97 Program for Crystal Structure Refinement,

Göttingen, 1997.

[27b]

88, 149.

Received May 4, 2004

[28]

T. Eicher, S. Hauptmann, The Chemistry of Heterocycles, 2nd

edn, WILEY-VCH, Weinheim, 2003.

Early View Article

Published Online September 13, 2004

4192

 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org

Eur. J. Inorg. Chem. 2004, 4177Ϫ4192

Article Doi

Complex formation of ICL670 and related ligands with FeIII and FeII

DOI: 10.1002/ejic.200400363

Source and publish data:

Authors:

Article abstract of DOI:10.1002/ejic.200400363

Full text of DOI:10.1002/ejic.200400363

Products guided by the article

R&D Labs maybe for 110981-89-0

Relevant to this article

Hot Product