N-2-(Azol-1(2)-yl)ethyliminodiacetic acids: A novel series of Gd(III) chelators as T2 relaxation agents for magnetic resonance imaging

Article abstract of DOI:10.1016/S0968-0896(98)00273-9

The synthesis, physicochemical properties, and toxicological implications of a novel series of N-2-(azol-1(2)-yl)ethyliminodiacetic acids, useful as contrast agents for magnetic resonance imaging are reported. Compounds were prepared by alkylation of methyl iminodiacetate with N-2-bromoethylazoles and subsequent hydrolysis. Stability constants of the corresponding Gd(III) complexes and T₁ and T₂ relaxivities were determined and interpreted in terms of optimized geometries obtained by semiempirical PM3 calculations. Compounds show increased T₂ relaxivity and decreased toxicity in vitro as compared to EDTA-Gd(III) complexes. Copyright (C) 1999 Elsevier Science Ltd.

Full text of DOI:10.1016/S0968-0896(98)00273-9

Bioorganic & Medicinal Chemistry 7 (1999) 517±527
N-2-(Azol-1(2)-yl)ethyliminodiacetic Acids: A Novel Series of
Gd(III) Chelators as T2 Relaxation Agents for Magnetic
Resonance Imaging
Pilar Lopez,a Christa G. Seipelt,a Patrick Merkling,b Laszlo Sturz,b Jose Alvarez,c
Andreas Dolle,b Manfred D. Zeidler,b Sebastian Cerdanc and Paloma Ballesterosa,*
a
Â
Â
Â
Dpto. Quõmica Organica y Biologõa, Facultad de Ciencias, UNED, Senda del Rey s/n, 28040, Madrid, Spain
bInstitut fur Physikalische Chemie, Rheinisch-Westfalische Technische Hochschule, D-52056 Aachen, Germany
È
c
È
Instituto de Investigaciones Biomedicas CSIC, c/ Arturo Duperier 4, 28029 Madrid, Spain
Â
Received 7 August 1998; accepted 6 November 1998
AbstractÐThe synthesis, physicochemical properties, and toxicological implications of a novel series of N-2-(azol-1(2)-yl)ethylimi-
nodiacetic acids, useful as contrast agents for magnetic resonance imaging are reported. Compounds were prepared by alkylation of
methyl iminodiacetate with N-2-bromoethylazoles and subsequent hydrolysis. Stability constants of the corresponding Gd(III)
complexes and T1 and T2 relaxivities were determined and interpreted in terms of optimized geometries obtained by semiempirical
PM3 calculations. Compounds show increased T2 relaxivity and decreased toxicity in vitro as compared to EDTA-Gd(III) complexes.
# 1999 Elsevier Science Ltd. All rights reserved.
Introduction
This study reports the syntheses, physicochemical char-
acterization, and toxicological properties of a novel
series of chelating agents containing heterocyclic rings.
Magnetic resonance imaging (MRI) as routinely
employed in clinical settings often requires the use of
paramagnetic ion chelates as contrast agents.1±3 Gd(III)
remains as the optimal paramagnetic ion for this pur-
pose because of its high electronic spin (S=7/2), rela-
tively slow electronic relaxation rate and exchange labile
hydration sphere. However, free Gd(III) is toxic in vitro
and in vivo, a circumstance which has prompted
research for chelating agents able to minimize its toxi-
city and optimize its relaxivity properties. Chelating
agents developed previously for this purpose included a
series of poliaminopolicarboxylic acid derivatives and a
family of macrocyclic ligands.2,4 Four of these com-
pounds have already been approved for clinical use;
Gd-diethylenetriaminepentaacetic acid (Gd-DTPA2 ),
Gd-1,4,7,10-tetra-azacyclododecane-1,4,7,10-tetraacetic
acid (Gd-DOTA2 ), Gd-10-(2-hydroxy-propyl)-1,4,7,
10-tetraazacyclododecane-1,4,7-triacetic acid (Gd-
HPDO3A), and Gd-diethyelenetriamine-N,N0,N00-tri-
acetic acid-N,N00bis(methylamide) (Gd-H3DTPA-BMA).
However, the design of new chelate molecules to
improve relaxivity, decrease toxicity, achieve higher
organ selectivity or optimize image quality in speci®c
applications constitutes an active area of research.5±7
Chelating properties of complexones are well docu-
mented.8 The presence of one or several moieties of
iminodiacetic acid (IDA) allow these compounds to
form tridentate or higher order coordination complexes
with almost all cations, giving chelate rings. Complex-
ones have been used in the quantitative analyses of
metal ions, the complexation of heavy metals in nuclear
medicine9 and more recently as contrast agents in mag-
netic resonance imaging.10,11 In general, chemical struc-
ture of the known complexones includes carboxy,
hydroxy or amino groups in addition to one or more
iminodiacetic moieties. However, in spite of the high
capability of heterocyclic rings as nitrogen-donor
ligands,12±14 few complexones have incorporated yet
such rings in their structure. Some 2,20-bipyridine tetra-
acetic acid derivatives were proposed as luminescent
makers for bioanity assays or more recently, even as
contrast agents for MRI.15±18 Our previous experience
on azole-containing ligands19±21 prompted us to further
explore the role of nitrogen based heterocycles as addi-
tional ligands in complexone structures.
We synthesized complexones 1±4 and evaluated their
physicochemical and toxicological properties using IDA
or ethylenediamine-N,N,N0,N0-tetraacetic acid (EDTA)
*Corresponding author. Fax: +34-1-398-6697.
0968-0896/99/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(98)00273-9

Â
P. Lopez et al. / Bioorg. Med. Chem. 7 (1999) 517±527
518
of 1,2-dibromoethane using liquid±liquid phase transfer
catalysis. Although compounds 8 and 9 have been
described previously23,24 we studied in more detail this
reaction to optimize conditions preventing the formation
of undesirable side vinyl- and bis-derivatives (Table 1).
1
Bis-azolylethanes were identi®ed in the H NMR spec-
tra of the reaction crudes by comparison with the pre-
viously reported resonances.25 In the case of indazole 7,
only the 2-substituted isomer of the vinyl derivative was
formed in this reaction. Subsequent alkylation with N-
2-bromoethylazoles to obtain the corresponding methyl
N-2-(azol-1(2)yl) ethyliminodiacetates 12±15, was per-
formed using two moles of methyl iminodiacetate26
to trap the HBr formed in the alkylation process.27
The methyl N-2-(azol-1(2)yl) ethyliminodiacetate hydro-
bromide was recycled by prior basi®cation with
Na2CO3. The free complexones 1±4 and 16 were
obtained by basic or acid hydrolysis. Structures of the
diesters and the diacid salts were established by ele-
Chart 1.
mental analyses and spectroscopic methods. 1H and 13
C
NMR spectra were assigned on the basis of previous
data.19±21
as well characterized reference compounds. The Gd(III)
complexes of 3 and 4 display improved T2 relaxivity and
lower toxicity in vitro than EDTA-Gd(III). These cir-
cumstances make them potentially useful in MRI appli-
cations requiring T2 enhancement as the imaging of ¯ow
and perfusion.22
Heterocyclic complexones as magnetic resonance
relaxation agents
Figure 1 shows the e€ects of increasing Gd(III) con-
centrations on the linewidth and chemical shift of the
water signal in the absence and presence of compounds
1±4 or EDTA. Increasing concentrations of Gd(III)
caused increased line broadening and down®eld shifts of
the water signal. For the same concentration of com-
plexones and Gd(III), the water linewidth was broader
with complexones 1±4 than in EDTA. These results
suggested that the Gd(III) complexes of compounds 1±4
Results
Syntheses of heterocyclic complexones
N-2-Bromoethylazoles 8±11 (Scheme 1) were prepared
by reaction of the corresponding azoles 5±7 with an excess
Scheme 1.

Â
P. Lopez et al. / Bioorg. Med. Chem. 7 (1999) 517±527
519
Table 1. Results obtained in the reaction of azoles 5±7 with 1,2-
dibromoethane
Gd(III) complexes of compounds 3 and 4 may become
useful contrast agents for MRI applications requiring
T2 weighting. This aspect is illustrated in Figure 2.
Azole Reaction Bromoethylazole Vinylazole Bis-azolylethane
conditions
(%)
(%)
(%)
The ®gure shows a magnetic resonance image taken
from a phantom consisting of three di€erent capillaries
containing saline, the Gd(III)-4 complex and Gd(III)-
EDTA complex, respectively. The left panel shows a
proton density image in which the three capillaries
depict similar intensity. The right panel shows a repre-
sentative T2-weighted image in which the capillary con-
taining the complex Gd(III)-4 appears as the darkest,
con®rming the stronger T2 weighting e€ect than
Gd(III)-EDTA.
5
6
7
24 h/rt
1 h/re¯ux
24 h/rt
8: 90a; 72b
9: 60a; 42b
10: 55a; 32b
11: 34a; 18b
10: 58a; 35b
11: 27a; 15b
Tracesa
10a
Tracesa
30a
11a
Nonea
1 h/re¯ux
15a
Nonea
aDetermined from 1H NMR spectra of the reaction crude.
bIsolated yield.
perturb the magnetic relaxation properties of water
more e€ectively than the corresponding Gd-EDTA
complexes.
Themodynamic stability constants for Gd(III) and Ca2+
complexation
The e€ects of Gd(III) complexes of compounds 1±4 on
water relaxation are investigated in more detail in
Table 2. This table summarizes the results of T1 and T2
measurements and corresponding longitudinal and
transversal water relaxivities for the Gd(III) complexes
of compounds 1±4, EDTA and IDA. Heterocyclic
complexones depicted higher T1 relaxivity than IDA
and slightly lower than EDTA. T2 relaxivity increased
in the order 1<2<3<4, con®rming the e€ects of these
complexes on the water linewidth. Interestingly, trans-
versal relaxivity of complexones 3 and 4 was twice that
of EDTA. Taken together, these results indicate that
Apparent stability constants for Gd(III) complexes
from compounds 1±4, EDTA and IDA were investigated
spectrophotometrically by competition with the metal-
lochromic indicator ArsenazoIII (ArsIII)28 (Fig. 3).
Titrations of ArsIII with Gd(III) depicted typical hyper-
bolic saturation curves, yielding an apparent stability
constant of 2Â105 M for the Gd(III)-ArsIII complex
(0.15 M ionic strength, pH 6.7, 22 ꢀC). For the same
total amount of ArsIII and Gd(III) added, the presence
of 10 mM IDA, 10 mM compounds 1±4 or 1 mM EDTA
in the titration mixture caused a reduction in the amount
of Gd(III) bound to ArsIII. This reduction reveals the
Figure 1. E€ects of increasing Gd(III) concentrations on some magnetic properties of the water resonance in the presence and absence of com-
1
pleꢀxones 1±4 or EDTA. H NMR spectra were acquired as indicated in Experimental. 25 mM compounds 1±4 or EDTA in D2O (99.9 D, pD 8±9,
.
22 C, 1 mL ®nal volume) were titered with GdCl3 6H2O. A: Linewidth at half height. B: Chemical shift referred to external TSP.
Table 2. Inverse T1 and T2 values and longitudinal (R1) and transverse (R2) relaxivities of EDTA, IDA, and compounds 1±4
1
1
Compound
Á 1/T1
R1/M 1 s
Á 1/T2
R2/M 1 s
EDTA
IDA
1
2
3
4
3.542‹0.037
1.392‹0.036
2.284‹0.033
3.282‹0.123
2.885‹0.35
2.941‹0.031
7083‹145 (6)
2784‹99 (3)
4569‹112 (6)
6565‹312 (6)
5770‹128 (6)
5883‹121 (6)
6.313‹0.232
2.584‹0.273
4.097‹0.289
7.440‹0.567
11.651‹0.523
13.892‹1.511
12627‹590 (6)
5168‹597 (3)
8193‹660 (6)
14880‹1283 (12)
23302‹1278 (6)
27784‹3300 (12)

Â
P. Lopez et al. / Bioorg. Med. Chem. 7 (1999) 517±527
520
Figure 2. Spin echo MR images (360.1 MHz) of three capillaries (1 mm diameter) containing 0.5 mM Gd(III) and 1 mM EDTA in 155 mM NaCl
(top), 0.5 mM Gd(III) and 10 mM compound 4 in 155 mM NaCl (bottom left) and 155 mM NaCl (bottom right). 100 mM Tris/HCl was used in all
1
cases to bu€er pH to 8.5. The three capillaries were placed in a 5 mm H NMR tube. MR images were acquired as indicated in Experimental. (A)
Proton density image, TR: 15 s, TE: 2 ms. (B) T2 weighted image, TR: 15 s, TE: 43 ms. Slice thickness: 1 mm. Pixel size: 78 mÂ78 m.
Figure 3. Titrations (25 ꢀC, pH 6.7, 0.1 M NH4Cl) of ArsIII (0.1 mM) with Gd(III) in the absence and presence of compounds 1±4 (10 mM), EDTA
(1 mM), or IDA (10 mM). The concentration of the Gd(III)-ArsIII complex (GdAs) was measured spectrophotometrically as indicated in Experi-
mental. The symbols refer to titrations of the di€erent complexones. Continuous lines show the results of simulations performed with optimized
values of the corresponding stability constants (K0).
amount of Gd(III) bound to the complexone and per-
mits calculation of the stability constant of the Gd(III)
complexone complex. Di€erent titration behavior was
observed in IDA, EDTA and complexones 1±4. EDTA
presented a much higher anity constant for Gd(III)
than ArsIII causing no ArsIII-Gd(III) complex forma-
tion when the concentration of EDTA was higher than
ArsIII. In contrast, the IDA titration curve was similar
to that of ArsIII revealing no signi®cant competition for
Gd(III) between these two ligands. Complexones 3 and
4 showed an intermediate behavior between IDA and
EDTA. In these cases, a sigmoidal binding pattern
appeared, revealing the presence of two di€erent binding
modes. At low concentrations of Gd(III), complexones

Â
P. Lopez et al. / Bioorg. Med. Chem. 7 (1999) 517±527
521
3 and 4 depicted a similar anity for Gd(III) than
EDTA and only a small amount of Gd(III)-ArsIII
complex was formed. However, at concentrations of
Gd(III) higher than 50 mM, a lower anity binding
mode similar to that of IDA appeared in the complex-
ones, as revealed by the increased concentration of
Gd(III)-ArsIII complex. The high anity binding mode
most probably represents a tetradentate coordination
chelate of complexones 1±4, similar to the pentadentate
Gd(III)-EDTA complex at pH 7.0. The low anity
binding mode represents most probably a tridentate
complex similar to that of Gd(III)-IDA. Thus, com-
plexones 1±4 present di€erent Gd(III) binding proper-
ties depending on the relative concentrations of Gd(III)
to complexone. Computer simulations permitted the
estimation of values for the stability constants (K0) of
the di€erent complexes formed. Optimal simulations of
the titration curves were obtained when values for the
stability constants of the high anity binding modes of
1 and 4 were 105 M, a very similar value to that
obtained for EDTA at the same pH (106 M). Similarly,
simulations of binding isotherms revealed values for the
low anity binding modes of: IDA=26.1 M; 1=
48.8 M; 2=60.6 M; 3=40.8 M; 4=40.0 M.
Table 3. Heat of formation (Hf, kcal/mol), e€ective charges (e), and
bond orders of calculated geometries of complexes from compounds
1±4, EDTA, and IDA with Mg2+, Ca2+, and Gd(III)
Complex
Hf
Metal e€ective
chargea
Metal bond
order
1-Mg
2-Mg
3-Mg
4-Mg
IDA-Mg
EDTA-Mg
187
0.9
0.9
0.8
0.9
1.2
0.7
2.6
2.6
2.6
2.6
2.3
2.7
193
162
151
234
449
1-Ca
2-Ca
3-Ca
4-Ca
IDA-Ca
EDTA-Ca
292
325
295
288
-286
607
2.5
2.5
2.5
2.5
3.0
2.3
1.5
1.5
1.5
1.5
1.2
1.6
1-Gd
2-Gd
3-Gd
4-Gd
IDA-Gd
EDTA-Gd
220
247
227
215
182
699
2.5
2.5
2.4
2.5
3.1
2.6
3.2
3.2
3.3
3.2
2.7
3.2
aMetal e€ective charge=Mulliken charge+electrostatic charge.
induced shifts (LIS) in these compounds were similar to
those obtained in EDTA. Since the magnitude of LIS
values is related to the distance of a proton to the
paramagnetic center,29 these results con®rm that the
azole moiety is involved in the complexation as indi-
cated by the theoretical calculations. In contrast, shifts
induced in Ca2+ titrations in the protons from the azole
moiety were negligible.
Additionally, stability constants for Ca2+ of com-
pounds 1±4, IDA and EDTA were determined poten-
tiometrically with a Ca2+ selective electrode. Log K0
values were (pH 9.0); 1 (3.35), 2 (2.77), 3 (3.75), 4 (3.04),
IDA (2.24), EDTA (7.12). Thus, compounds 1±4 and
IDA present similar Ca2+ binding properties revealing
that the azole moiety is not e€ectively involved in Ca2+
complexation.
Toxicity assays
Calculations of molecular geometry
Figure 5 shows the results of toxicity tests performed
with cultures of H35 hepatoma cells. Compounds 1±4,
IDA, and EDTA were added to the cultures in increas-
ing concentrations of 10 mM, 100 mM, and 1 mM in the
absence and presence of Gd(III). Toxicity was assayed
as the release of LDH to the medium after incubations
of the cells with the complexones for increasing periods
of time up to 6 h. In the absence of Gd(III) all com-
pounds had similar toxicity with the exception of 1. In
the presence of Gd(III), the toxicity of compounds 2±4
was not increased.
We performed semiempirical (PM3) calculations of the
structure of chelates from compounds 1±4, EDTA, and
IDA with three di€erent cations: Ca2+, Mg2+, and
Gd(III). The geometry of the complex is determined by
the cation, originating in all cases a cage structure in
which the cation is buried. Calculated bond orders and
e€ective charges on the metal for 1±4, IDA, and EDTA
are given in Table 3. Figure 4 depicts an illustrative
example of the optimized geometries from the Gd(III)
complexes of compounds 3 (panel A) and 4 (panel B).
Low values of metal e€ective charge and high values of
metal bond orders infer high anity for the metal.
Taking into account the bond order and the e€ective
charge on the cation, the following metal chelation
Discussion
capacities can be deduced from the calculations: Mg2+
:
Although the physical theories of relaxivity in com-
plexes of paramagnetic ions used as contrast agents in
MRI are considerably developed, less is known on the
relation between chemical structure of the ligand and
relaxivity. The present study provides a useful frame to
address this topic. Measurements of relaxation times
indicate that heterocyclic complexones 3 and 4 origi-
nate paramagnetic Gd(III)-complexes with signi®cantly
improved T2 relaxivity compared with the more classical
complexones like EDTA, DTPA, or DOTA, which do
not contain heterocycles.4 This is a remarkable ®nding
since the classical Gd(III) complexes are known as T1
relaxivity enhancers.
IDA<1=2=3=4<EDTA; Ca2+: IDA<1=2=3=4
<EDTA; Gd(III): IDA<1=2=EDTA<3=4. Nota-
bly, these calculated results matched well with the sta-
bility constants determined experimentally for Ca2+
and Gd(III).
The participation of the azole ring in the complexation
was com®rmed experimentally. 1H NMR spectra of
compounds 3 and 4 obtained with increasing Gd(III)
concentrations showed that protons from the azole
moiety and the iminodiacetic part were shifted simi-
larly by the paramagnetic ion. Indeed, the lanthanide

Â
P. Lopez et al. / Bioorg. Med. Chem. 7 (1999) 517±527
522
Figure 4. Semiempirical PM3 calculations of the optimized geometries of Gd(III) complexes from compounds 3 (A) and 4 (B). E€ect of increasing
Gd(III) concentrations on the chemical shifts of the protons from compounds 3 (C) and 4 (D). Titrations were performed as in Figure 1. Ád indi-
cates down®eld shift induced by Gd(III).
Figure 5. Toxicity assays of compounds 1±4, IDA, and EDTA in the absence (left) and presence (right) of Gd(III). Acute toxicity was determined in
vitro measuring the release of LDH to the medium in cultures of H35 cells incubated for 1 h with increasing concentrations of the compounds. LDH
release is expresed as a percentage of total releasable LDH. When present, Gd(III) was added as 10% of the total concentration of compounds 1±4
and IDA or equimolar concentration of EDTA.
T1 or T2 relaxivity in paramagnetic complexes is
thought to depend on the electronic charge distribution
around the paramagnetic ion.30 T1 relaxivity enhance-
ment in Gd(III) complexes was proposed to depend on
the isotropic charge distribution around Gd(III). In
contrast, improved T2 relaxivity in paramagnetic com-
plexes of EDTA, DTPA or DOTA could be obtained
only when Gd(III) was substituted by other lanthanides
with an unsymmetrical charge distribution such as
Dy(III).31 The present study shows that T2 relaxivity
can also be induced in Gd(III) complexes when an
appropriate unsymmetrical ligand is used. The substitu-
tion in 3 and 4 of one iminodiacetic moiety of EDTA by
a benzazole ring results in an unsymmetrical structure
containing four chemically di€erent donor sites: two
carboxylic groups and a tertiary amino group from the
iminodiacetic part and the benzazole. MO calculations
support these interpretations showing that the species
containing the azole ring as donor, folds over creating
an unsymmetrical cavity in which the metal is placed

Â
P. Lopez et al. / Bioorg. Med. Chem. 7 (1999) 517±527
523
(Fig. 4). Cage structures are typical of many chelates
and a symmetrical cage structure has been resolved by
X-ray di€raction of several Ln(III)-EDTA chelates.32
interesting ®nding since previous Gd(III) complexes
were considered mainly as T1 enhancers. 1H NMR
measurements in solution combined with MO calcula-
tions indicate that the structure of these complexes is
governed by the metal, giving cage structures where the
paramagnetic ion is inserted. The toxicity in vitro of the
Gd(III) complexes of these heterocyclic complexones is
lower than that of Gd(III)-EDTA.
Relaxivity e€ects are the consequence of additive inner
sphere and outer sphere e€ects. These e€ects involve
magnetic relaxation of the water molecules in direct
contact with the paramagnetic ion and of those water
molecules further away in the solution, respectively.30 It
is possible to analyze the T1 or T2 relaxivity e€ects in
complexones 3 and 4 in terms of these inner or outer
sphere contributions. Expressions describing the inner
sphere e€ects on T1 and T2 relaxivity are:
Experimental
General
Melting points were obtained on a microscope hot stage
and are uncorrected. Elemental analyses were per-
formed with a Perkin±Elmer 240 apparatus. Mass spec-
tra were carried out (GC/mass spectrometer Schimadzu
QP-5000) at 70 eV. IR spectra were determined on a
Philips PU-9700 spectrophotometer. NMR spectra were
1
‰CŠq
1
ˆ
T1is 55:6 T1M ‡ tM
1
‰CŠq
ˆ
ꢀdo†2tM
T2is 55:6
1
recorded with a Bruker AC-200 (200.13 MHz for H,
1
and 50.33 MHz for 13C). H and 13C chemical shifts (d)
in CDCl3 are given from internal tetramethylsilane, H
1
where [C] is the molar concentration of paramagnetic
ion, q the number of water ligands per ion, T1M is the
longitudinal relaxation time of water protons bound to
the metal ion and in rapid exchange with the solvent, tM
is the residence time of these water molecules in the
complex and do is the chemical shift e€ect.30 Thus,
inner sphere T1 and T2 relaxivities depend on the num-
ber of water molecules in direct contact with the ion and
their residence time, but T2 relaxivity depends addition-
ally on a chemical shift e€ect. In contrast, theory indi-
cates that outer sphere contributions to T1 and T2
relaxivity do not depend on the number of water mole-
cules coordinated to the metal ion or on the chemical
shift e€ects.30 For a nine or 10 possible coordination
sites in free Gd(III), high anity binding modes of the
Gd(III) complexes of 3 and 4 form tetradentate com-
plexes (Fig. 4), leaving up to ®ve or six coordination
sites for water in the ®rst coordination sphere. This
number of water molecules is larger than the three water
molecules bound to Gd(III)-EDTA complex or the sin-
gle water molecule bound to the Gd(III)-DTPA com-
plex. Thus, increased T2 relaxivity in Gd(III) complexes
of 3 and 4 appears to be caused mainly by an inner
sphere e€ect due to the increased number of water
molecules in the ®rst coordination sphere of the Gd(III).
We propose that hydrogens from the various water
molecules experience di€erent deshilding e€ects. Thus,
the resulting chemical shifts vary for every water mole-
cule depending from the balance of the electronic con-
tributions of metal and ligand. These ligand induced
shifts are anysotropic due to the di€erent location of the
azole moiety and the iminodiacetic acid part in the
d in D2O are given from external 3-trimethylsilyl(tetra-
deutero)propionic acid sodium salt, and 13C d in D2O
are given from external DMSO-d6, with an accuracy of
‹0.01 for 1H and ‹0.1 ppm for 13C. The residual water
1
signal in H NMR spectra in D2O solutions was sup-
pressed when necessary using a 1 s (low power, 0.5
watts) presaturating pulse applied with the decoupler.
1H±1H coupling constants (J) are accurate to ‹0.2 Hz
1
for H NMR spectra. TLC chromatography was per-
formed on DC-Alufolien/Kieselgel 60 F254 (Merck,
0.2 mm) and column chromatography through silica gel
Merck 60 (70±230 mesh). Methyl iminodiacetate hydro-
chloride was obtained from iminodiacetodinitrile,26 and
basi®ed with solid Na2CO3 in the minimum amount of
water prior to use. D2O (99.9 D) was purchased from
Appollo Scienti®c (Stockport, Great Britain). The rest
of the products were purchased from Aldrich.
Synthesis of 1(2)-(2-bromoethyl)azoles. General method.
A one-necked round-bottomed ¯ask was ®tted with a
re¯ux condenser and with a magnetic stirrer. The ¯ask
was charged with 40% NaOH, the azole, tetrabutyl-
ammonium bromide (TBAB), and 1,2-dibromoethane
(3:1:1/40:10 molar ratio). The reaction mixture was left
to stir under the conditions shown in Table 1. The
organic phase was then separated, dried and evaporated
in vacuo. The residue was puri®ed by column chroma-
tography and/or distillation.
1-(2-Bromoethyl)pyrazole (8). (From 29.4 to 117.6 mmol
of azole). bp0.01 38±40 ꢀC. IR (®lm): 3110, 3020, 2970,
2920, 1510, 1445, 1410, 1395, 1290, 1260, 1215, 1125,
1090, 1050, 965, 915, 880, 755 cm 1. MS: m/z 176
(M+1, 11%), 174 (M 1, 11%), 95 (31%), 81 (41%), 69
(8%), 68 (100%), 54 (15%), 53 (17%), 52 (8%). 1H
NMR (CDCl3, d):23 7.56 (d, 1H, 3J34=1.6 Hz, H3), 7.46
2
complex, a€ecting the ꢁdo† term and therefore T2.
Conclusions
3
We have synthesized a novel series of complexones
containing azole rings, N-2-(azol-1(2)-yl)ethyliminodia-
cetic acids. The new complexones show improved T2
relaxivity than Gd(III) complexes of EDTA. This is an
(d, 1H, J45=2.1 Hz, H5), 6.26 (apparent triplet, 1H,
H4), 4.51 (t, 2H, 3J=6.4 Hz, CH2-N), 3.72 (t, 2H,
3J=6.4 Hz, CH2-Br); 13C NMR (CDCl3, d): 139.8 (d,
1JC3H3=185.2 Hz, C3), 129.7 (d, 1JC5H5=186.4 Hz, C5),

Â
P. Lopez et al. / Bioorg. Med. Chem. 7 (1999) 517±527
524
1
1
105.1 (d, JC4H4=177.0 Hz, C4), 52.9 (t, J=141.6 Hz,
1
methyl iminodiacetate (1:2 molar ratio). The mixture
was heated with stirring in an oil bath at 110 ꢀC for 2.5
to 4.5 h. After cooling at room temperature the mixture
was extracted with CH2Cl2. Organic solvent was eva-
porated in vacuo and the residue puri®ed by column
chromatography and/or distillation.
CH2-N), 30.2 (t, J=153.6 Hz, CH2-Br).
1-(2-Bromoethyl)-3,5-dimethylpyrazole (9). (For 14.0 mmol
of azole). Chromatographic eluent: ethyl acetate. bp0.05
:
46±48 ꢀC Lit.24 bp0.4 89±90 ꢀC. IR (®lm): 3130, 3030,
2960, 2920, 2870, 1550, 1455, 1420, 1385, 1300, 1260,
1215, 1145, 1030, 975, 880, 780 cm 1. MS: m/z 204
(M+1, 11%), 203 (M, 1%), 202 (M 1, 11%), 123
(4%), 109 (43%), 97 (6%), 96 (100%), 95 (35%), 82
Methyl N-(2-pyrazol-1-yl)ethyliminodiacetate (12). (From
2.24 to 6.7 mmol of 1-(2-bromoethyl)pyrazole). Reac-
tion time: 2.5 h. Chromatographic eluent: hexane:etha-
nol, 8:2. bp0.01 128±130 ꢀC. Yield: 55%. IR (®lm): 3140,
3120, 3000, 2960, 2840, 1740, 1510, 1435, 1395, 1365,
1280, 1205, 1180, 1145, 1090, 1060, 1040, 1010, 880,
760 cm 1. MS: m/z: 256 (M+1, 4%), 255 (M, 13%), 196
(28%), 188 (6%), 187 (50%), 174 (55%), 146 (40%), 128
(70%), 116 (27%), 100 (17%), 95 (17%), 94 (14%), 81
(18%), 69 (14%), 68 (37%), 59 (18%), 56 (25%), 54
1
(6%), 81 (6%), 68 (11%), 55 (6%). H NMR (CDCl3,
3
d): 5.79 (s, 1H, H4), 4.31 (t, 2H, J=6.7 Hz, CH2-N),
3
3.68 (t, 2H, J=6.7 Hz, CH2-Br), 2.27 (s, 3H, CH3-5-),
2.21 (s, 2 H, CH3-3-); 13C NMR (CDCl3, d): 148.0 (s,
C3), 139.2 (s, C5), 104.9 (d, 1JC4H4=172.6 Hz, C4), 49.3
(t, 1J=140.7 Hz, CH2-N), 30.2 (t, 1J=153.4 Hz, CH2-Br),
1
1
13.3 (q, J=126.9 Hz, CH3-3-), 10.8 (q, J=128.5 Hz,
CH3-5-).
1
(18%), 45 (100%). H NMR (CDCl3, d): 7.53 (d, 1H,
3
3J34=1.8 Hz, H3), 7.50 (d, 1H, J45=2.0 Hz, H5), 6.22
1-(2-Bromoethyl)indazole (10). (From 4.24 to 17.0 mmol
of azole). Chromatographic eluent: hexane:ethyl ace-
tate, 6:4. IR (®lm): 3080, 3060, 2980, 2940, 1645, 1615,
1500, 1465, 1435, 1420, 1315, 1300, 1275, 1230, 1210,
1160, 1010, 910, 850, 835, 755, 745 cm 1. MS: m/z 226
(M+1, 17%), 225 (M, 1%), 224 (M 1, 17%), 132
(9%), 131 (100%), 118 (16%), 104 (11%), 103 (15%), 89
(apparent triplet, 1H, H4), 4.24 (t, 2H, 3J=6.2 Hz, CH2-
Azole), 3.19 (t, 2H, 3J=6.2 Hz, CH2-N-); 13C NMR
1
(CDCl3, d): 171.6 (s, 2C, CO), 139.3 (d, JC3H3
186.9 Hz, C3), 130.0 (d, JC5H5=190.9 Hz, C5), 105.1
=
1
(d, JC4H4=176.3 Hz, C4), 55.3 (t, 2C, 1J=136.5 Hz,
1
1
CH2-CO-), 54.7 (t, J=135.8 Hz, CH2-N-), 51.5 (q, 2C,
1
1J=147.2 Hz, CH3), 51.2 (t, J=139.6 Hz, CH2-Azole).
(8%), 77 (21%), 63 (17%), 51 (9%). H NMR (CDCl3,
5
Picrate: mp 91±93 ꢀC. Anal. calcd for C17H20N6O11: C,
42.34; H, 4.07; N, 17.08. Found: C, 42.31; H, 4.10; N,
17.11.
1
d): 8.04 (d, 1H, J37=0.8 Hz, H3), 7.78 (ddd, 1H,
3J45=8.0 Hz, J46=0.9 Hz, J47=0.9 Hz, H4), 7.53±7.38
4
5
3
(m, 2H, H6 y H7), 7.20 (ddd, 1H, J45=8.0 Hz,
3J56=6.2 Hz, 4J57=1.8 Hz, H5), 4.76 (t, 2H, 3J=6.8 Hz,
CH2-N), 3.79 (t, 2H, J=6.8 Hz, CH2-Br); 13C NMR
Methyl N-2-(3,5-dimethylpyrazol-1-yl)ethyliminodiacet-
ate (13). (From 2.46 to 5.55 mmol of 1-(2-bromoethyl)-
3,5-dimethylpyrazole). Reaction time: 3.5 h. bp0.01 140±
142 ꢀC. Yield: 65%. IR (Film): 3130, 3000, 2960, 2930,
2870, 2220, 1740, 1550, 1435, 1385, 1260, 1200, 1180,
1170, 1120, 1050, 1020, 915, 775, 730 cm 1. MS: m/z:
283 (M, 5%), 224 (13%), 187 (28%), 174 (27%), 146
(35%), 128 (55%), 124 (7%), 116 (19%), 114 (6%), 109
(7%), 100 (9%), 97 (19%), 96 (16%), 82 (14%), 68
(6%), 59 (9%), 56 (12%), 54 (7%), 45 (100%). 1H
NMR (CDCl3, d): 5.74 (s, 1H, H4), 4.07 (t, 2H,
3J=6.8 Hz, CH2-Azole), 3.67 (s, 6H, CH3), 3.47 (s, 4H,
3
1
(CDCl3, d): 139.5 (s, C7a), 133.7 (d, JC3H3=189.9 Hz,
C3), 126.3 (d, 1JC6H6=165.7 Hz, C6), 123.6 (s, C3a), 120.9
1
1
(d, JC5H5=154.9 Hz, C5), 120.6 (d, JC4H4=165.4 Hz,
C4), 49.7 (t, 1J=141.4 Hz, CH2-N), 29.4 (t, 1J=153.5 Hz,
CH2-Br).
2-(2-Bromoethyl)indazole (11). (From 4.24 to 17.0 mmol
of azole). Chromatographic eluent: hexane:ethyl ace-
tate, 6:4. IR (®lm): 3120, 3060, 2970, 2920, 1625, 1510,
1470, 1425, 1380, 1350, 1310, 1265, 1160, 1140, 785,
760 cm 1. MS: m/z: 226 (M+1, 11%), 225 (M, 1%), 224
(M 1, 11%), 119 (9%), 118 (100%), 103 (4%), 91
(9%), 89 (9%), 77 (10%), 63 (16%), 51 (5%). 1H NMR
3
CH2-CO-), 3.09 (t, 2H, J=6.8 Hz, CH2-N-), 2.24 (s,
3H, CH3-5-), 2.18 (s, 3H, CH3-3); 13C NMR (CDCl3, d):
171.2 (s, 2C, CO), 147.0 (s, C3), 139.1 (s, C5), 104.4 (d,
5
1
(CDCl3, d): 8.00 (d, 1H, J37=0.8 Hz, H3), 7.70 (dddd,
1H, J67=7.3 Hz, J57=0.9 Hz, J47=1.0 Hz, J37=
1JC4H4=172.1 Hz, C4), 55.1 (t, 2C, J=136.4 Hz, CH2-
3
4
5
5
CO-), 54.2 (t, 1J=136.6 Hz, CH2-N-), 51.1 (q, 2C,
3
4
1
0.8 Hz, H7), 7.66 (ddd, 1H, J45=7.3 Hz, J46=1.0 Hz,
3
1J=142.1 Hz, CH3), 47.5 (t, J=138.8 Hz, CH2-Azole),
5J47=1.0 Hz; H4), 7.30 (ddd, 1H, J67=7.3 Hz, J56=
13.0 (q, J=123.4 Hz, CH3-3-), 10.5 (q, J=127.9 Hz,
CH3-5-). Picrate: mp: 83±85 ꢀC. Anal. calcd for C19H24
N6O11: C, 44.53; H, 4.69; N, 16.41. Found: C, 44.58; H,
4.64; N, 16.54.
3
1
1
4
3
6.7 Hz, J46=1.0 Hz, H6), 7.10 (ddd, 1H, J45=7.3 Hz,
3J56=6.7 Hz, 4J57=0.9 Hz, H5), 4.76 (t, 2H, 3J=6.3 Hz,
CH2-N), 3.79 (t, 2H, J=6.3 Hz, CH2-Br); 13C NMR
3
1
(CDCl3, d): 149.0 (s, C7a), 126.1 (d, JC6H6=160.6 Hz,
=
C6), 123.6 (d, 1JC3H3=189.9 Hz, C3), 121.6 (d, 1JC5H5
Methyl N-2-(indazol-1-yl)ethyliminodiacetate (14). (For
6 mmol of 1-(2-bromoethyl)indazole). Reaction time:
4.5 h. bp0.01 160±162 ꢀC. Yield: 56%. IR (®lm): 3060,
3000, 2950, 2850, 1740, 1610, 1505, 1470, 1435, 1315,
1200, 1180, 1160, 1120, 1085, 1010, 910, 830, 755,
740 cm 1. MS: m/z: 305 (M, 6%), 246 (11%), 187
(23%), 175 (6%), 174 (72%), 146 (39%), 131 (6%), 128
(17%), 118 (11%), 116 (20%), 115 (6%), 94 (10%), 91
(5%), 77 (14%), 56 (8%), 51 (5%), 45 (100%). 1H
1
159.7 Hz, C5), 121.2 (s, C3a), 120.1 (d, JC4H4
1
=
156.4 Hz, C4), 117.0 (d, JC7H7=167.0 Hz, C7), 54.5 (t,
1J=142.6 Hz, CH2-N), 29.4 (t, J=155.6 Hz, CH2-Br).
1
Synthesis of methyl N-2-(azol-1(2)-yl)ethyliminodiacetates.
General procedure. A 1-necked round-bottomed ¯ask
was ®tted with a re¯ux condenser attached to a drying
tube (CaCl2) and with a magnetic stirrer. The ¯ask was
charged with the corresponding bromoethylazole and
1
5
NMR (CDCl3, d): 7.99 (d, H, J37=0.8 Hz, H3), 7.71

Â
P. Lopez et al. / Bioorg. Med. Chem. 7 (1999) 517±527
525
(ddd, 1H, 3J45=8.1 Hz, 4J46=1.0 Hz, 5J47=1.0 Hz, H4),
7.50 (dddd, 1H, 3J67=8.1 Hz, 4J57=0.8 Hz, 5J47=1.0Hz,
corresponding ester and 0.6% NaOH (1:2 molar ratio).
The mixture was left at room temperature for 24 h.
After cooling water was removed in vacuo to yield
quantitatively the disodium salt.
3
3
5J37=0.8 Hz, H7), 7.37 (ddd, 1H, J67=8.1 Hz, J56=
4
3
6.8 Hz, J46=1.0 Hz, H6), 7.13 (ddd, 1H, J45=8.1 Hz,
4
3J56=6.8 Hz, J57=0.8 Hz, H5) 4.54 (t, 2H, J=6.8 Hz,
CH2-azole), 3.65 (s, 6H, CH3), 3.49 (s, 4H, CO-CH2-N),
3.26 (t, 2H, J=6.8 Hz, CH2-N-); C NMR (CDCl3, d):
171.1 (s, 2C, CO), 139.4 (s, C7a), 132.7 (d, JC3H3=
3
N-(2-Pyrazol-1-yl)ethyliminodiacetic sodium salt (1).
(From 4 to 40 mmol of diester). IR (KBr): 3130, 3100,
2960, 2940, 2870, 2840, 1730, 1600, 1505, 1425, 1395,
1345, 1300, 1205, 1180, 1140, 1115, 1090, 1055, 1030,
3
13
1
1
189.3 Hz, C3), 125.8 (d, JC6H6=160.7 Hz, C6), 123.6
1
1
(s, C3a), 120.5 (d, JC5H5=161.3 Hz, C5), 120.1 (d,
905, 870, 775, 740 cm 1. H NMR (D2O, d): 7.63 (dd,
1JC4H4=160.9 Hz, C4), 108.9 (d, 1JC7H7=163.6 Hz, C7),
55.1 (t, 2C, 1J=137.8 Hz, CH2-CO-), 53.5 (t, 1J=
135.8 Hz, CH2-N-), 51.1 (q, 2C, 1J=147.0 Hz, CH3),
47.7 (t, 1J=139.2 Hz, CH2-azole). Picrate: mp 105.5±
107.5 ꢀC. Anal. calcd for C21H22N6O11: C, 47.19; H,
4.20; N, 15.73. Found: C, 46.80; H, 4.07; N, 15.31.
1H, J45=1.8 Hz, J35=0.6 Hz, H5), 7.49 (dd, 1H,
4
3
4
3J34=1.5 Hz, J35=0.6 Hz, H3), 6.26 (apparent triplet,
3
1H, H4), 4.21 (t, 2H, J=6.7 Hz, CH2-azole), 3.09 (s,
3
4H, CH2-CO-), 2.98 (t, 2 H, J=6.7 Hz, CH2-N-).
N-2-(3,5-Dimethylpyrazol-1-yl)ethyliminodiacetic sodium
salt (2). (From 0.353 to 0.706 mmol of diester). IR
(KBr): 2980, 2950, 2920, 2880, 2830, 1600, 1550, 1420,
1350, 1300, 1260, 1230, 1200, 1140, 1095, 1040, 1010,
Methyl N-2-(indazol-2-yl)ethyliminodiacetate (15). (For
3.32 mmol of 2-(2-bromoethyl)indazole). Kugelrorh
ot0.01: 250 ꢀC. Yield: 60%. IR (®lm): 3120, 3060, 3000,
2960, 1735, 1620, 1505, 1430, 1375, 1355, 1200, 1180,
1160, 1055, 1010, 980, 905, 785, 760, 745 cm 1. MS: m/z:
305 (M, 10%), 246 (11%), 187 (31%), 174 (26%), 146
(28%), 128 (57%), 119 (8%), 118 (36%), 116 (13%), 114
(10%), 100 (9%), 94 (6%), 91 (6%), 89 (6%), 77 (9%),
59 (10%), 56 (7%), 54 (5%), 45 (100%). 1H NMR
1
1000, 920, 940, 900, 825, 770 cm 1. H NMR (D2O, d):
3
5.85 (s, 1H, H4), 4.04 (t, 2H, J=7.5 Hz, CH2-azole),
3.16 (s, 4H, CH2-CO-), 2.90 (t, 2H, 3J=7.5 Hz, CH2-N-),
2.17 (s, 3H, CH3-5-), 2.07 (s, 3H CH3-3-).
N-2-(Indazol-1-yl)ethyliminodiacetatic sodium salt (3).
(For 0.492 mmol of diester). IR (KBr): 3090, 3050,
2940, 2920, 2870, 2830, 1600, 1460, 1430, 1350, 1330,
1315, 1300, 1250, 1210, 1180, 1140, 1070, 1005, 940,
5
(CDCl3, d): 8.11 (d, 1H, J37=0.9 Hz, H3), 7.69 (dddd,
1H, J67=7.7 Hz, J57=1.0 Hz, J47=1.2 Hz, J37=
3
4
5
5
3
4
1
0.9 Hz, H7), 7.65 (ddd, 1H, J45=6.3 Hz, J46=1.2 Hz,
3
905, 880, 830, 795, 740 cm 1. H NMR (D2O, d): 8.02
5J47=1.2 Hz, H4) 7.27 (ddd, 1H, J56=8.3 Hz; J67=
(d, 1H, J37=1.0 Hz, H3), 7.75 (ddd, 1H, J45=8.1 Hz,
5
3
5
3
4
3
3
7.7 Hz, J46=1.2 Hz, H6), 7.06 (ddd, 1H, J56=8.3 Hz,
3J45=6.3 Hz, 4J57=1.0 Hz, H5), 4.52 (t, 2H, 3J=6.1 Hz,
CH2-azole), 3.65 (s, 6H, CH3), 3.46 (s, 4H, CH2-CO-),
3.35 (t, 2H, 3J=6.1 Hz, CH2-N-); 13C NMR (CDCl3, d):
171.5 (s, 2C, CO), 148.7 (s, C7a), 125.7 (d, JC6H6=
1
165.2 Hz, C6), 124.0 (d, JC3H3=190.3 Hz, C3), 121.5
4J46=1.1 Hz, J47=1.0 Hz, H4), 7.57 (dddd, 1H, J67=
4
5
5
7.8 Hz, J57=0.9 Hz, J47=1.0 Hz, J37=1.0 Hz, H7),
3
3
4
7.43 (ddd, 1H, J67=7.8 Hz; J56=6.8 Hz, J46=1.1 Hz,
4
3
3
H6), 7.14 (ddd, 1H, J45=8.1 Hz; J56=6.8 Hz, J57=
3
1
0.9 Hz, H5) 4.46 (t, 2H, J=7.0 Hz, CH2-azole), 3.14 (s,
4H, CH2-CO-), 3.04 (t, 2H, 3J=7.0 Hz, CH2-N-).
1
(s, C3a), 121.2 (d, JC4H4=159.4 Hz, C4), 120.1 (d,
1
1JC5H5=165.2 Hz, C5), 117.0 (d, JC7H7=160 Hz, C7),
N-2-(Indazol-2-yl)ethyliminodiacetic sodium salt (4).
(For 0.492 mmol of diester). IR (KBr): 3110, 3070,
2950, 2920, 2870, 2830, 1600, 1510, 1430, 1350, 1300,
1225, 1160, 1140, 1045, 1005, 980, 940, 905, 840, 775,
755 cm 1. 1H NMR (D2O, d): 8.29 (d, 1H, 5J37=0.9 Hz,
55.5 (t, 2C, 1J=137.6 Hz, CH2-CO-), 55.0 (t, 1J=
1
136.5 Hz, CH2-N-), 52.7 (t, J=141.2 Hz, CH2-azole),
51.4 (q, 2C, 1J=147.2 Hz, CH3). Picrate: mp 141±
142 ꢀC. Anal. calcd for C21H22N6O11: C, 47.19; H, 4.20;
N, 15.73. Found: C, 47.40; H, 4.62; N, 15.61.
3
4
5
H3), 7.71 (ddd, 1H, J45=8.4 Hz, J46=1.1 Hz, J47=
1.0 Hz, H4), 7.57 (dddd, 1H, 3J67=8.7 Hz, 4J57=1.0 Hz,
5J47=1.0 Hz, J37=0.9 Hz, H7), 7.31 (ddd, 1H, J67=
8.7 Hz, J56=6.6 Hz, J46=1.1 Hz, H6), 7.08 (ddd,
1H, J45=8.4 Hz, J56=6.6 Hz, J57=1.0 Hz, H5), 4.74
(t, 2H, 3J=6.8 Hz, CH2-azole), 4.51 (t, 3J=6.8 Hz,
CH2-N).
5
3
N-(2-Pyrazol-1-yl)ethyliminodiacetic acid hydrochloride
(16). A 1-necked round-bottomed ¯ask was ®tted with
a re¯ux condenser and with a magnetic stirrer. The
¯ask was charged with compound 12 (0.7 mmol) and 2
N HCl (1:18 molar ratio). The solution was heated in an
oil bath at 100 ꢀC for 4 h. After cooling at room tem-
perature the solvent was removed in vacuo and the
residue dried over P2O5 in vacuo. Compound 12 was
recrystallized from ethanol:diethyl ether. mp 142.5±
3
4
3
3
4
1
Determination of relaxivities. H NMR relaxation times
T1 and T2 (25 ꢀC, pH 8.8) of the water protons were
measured at 8.4 Tesla in a Bruker AM-360 NMR spec-
trometer. T1 or T2 values were determined by the inver-
sion recovery method or the spin echo sequence using
not less than 15 di€erent t values. At least six di€erent
measurements of T1 or T2 were performed in every
sample and the results were expressed as mean‹
standard deviation. Typically, 10 mM compounds 1±4
and IDA or 1 mM EDTA were dissolved in 100 mM
Tris/HCl, 155 mM NaCl containing or not 0.5 mM
Gd(III). Relaxivities R1(2) were calculated according to
the expression:
144.5 ꢀC. H NMR (D2O, d): 7.70 (d, 1H, J45=2.4 Hz,
1
3
3
H5), 7.61 (d, 1H, J34=2.2 Hz, H3), 6.31 (apparent tri-
3
plet, 1H, H4), 4.53 (t, 2H, J=6.0 Hz, CH2-azole), 3.70
(t, 2H, 3J=6.0 Hz, CH2-N-). Anal. calcd for C9H14
ClN3O4: C, 40.98; H, 5.35; N, 15.93. Found: C, 41.07;
H, 5.28; N, 15.50.
Basic hydrolysis. General method. A 1-necked round-
bottomed ¯ask was ®tted with a re¯ux condenser and
with a magnetic stirrer. The ¯ask was charged with the

Â
P. Lopez et al. / Bioorg. Med. Chem. 7 (1999) 517±527
526
R1ꢀ2† ˆ Á1=T1ꢀ2†=‰MŠ
determined potentiometrically with a Ca2+ selective
electrode (Ingold, Switzerland). Measurements of free
Ca2+ concentrations were performed in solutions con-
taining 20 mM Tris/HCl bu€er (pH 9.0), 100 mM
NH4Cl and 1 mM CaCl2 increasing concentrations of
the di€erent compounds.
where, for every complexone, Á is the di€erence in
relaxation rates 1/T1(2) of the water protons in the pre-
sence and absence of Gd(III), and [M] is the molar
concentration of Gd(III).
Magnetic resonance imaging. 1H MRI was performed at
8.4 T on a Bruker AM-360 spectrometer equipped with
a microscopic imaging accessory. A commercial 5 mm
1H microimaging probe was used. A spin echo sequence
(90 t 180 acquire) consisting of 90ꢀ nonselective
pulse (8 ms) and a slice selective 180ꢀ (4 ms) refocussing
pulse in the presence of a slice gradient (5 gauss/cm) was
used. The echo was collected in the presence of a read
gradient (15 gauss/cm) and phase encoding was induced
with 64 increases of the phase gradient (from 10.7 to
+10.7 gauss/cm). Pixel resolution (x, y, z) was 78 m,
78 m, 1 mm.
Theoretical calculations. Semiempirical studies of chela-
tion with di€erent cations were performed using the
Spartan 4.1.1. program implemented in a Silicon Gra-
phics platform. The method used Parametrization
Model 3 (including parameters for transition metals
(PM3tm) in gas phase because parameters for metals in
water were not available). The studies assume a basic
pH 7 to 11 at which EDTA chelates totally in the
deprotonated form (Y4 ) and in the monoprotonated
form (HY3 ).
Toxicity assays. Toxicity was assayed by measuring the
release to the medium of lactic dehydrogenase (LDH)
from cultures of H35 hepatoma cells. Cells were grown
until con¯uence in DMEM medium containing 10%
fetal calf serum. Cultures were incubated in the presence
of increasing concentrations of compounds 1±4, IDA,
or EDTA in the absence and presence of Gd(III). When
present, Gd(III) was added to a ®nal amount repre-
senting 10% of the total complexone concentration. The
amount of LDH released to the medium at 1, 2, 3, and
6 h was measured spectrophotometrically (340 nm)
using an incubation mixture containing 50 mM phos-
phate bu€er (pH 7.2), 10 mM pyruvate and 0.5 mM
NADH. At the end of the experiment cells were broken
by freeze thawing three times and the total releasable
LDH determined. The results obtained at every time
point were expressed as percentage of the total relea-
sable LDH after cell destruction.
Determination of apparent stability constants. Apparent
stability constants K0 (25 ꢀC, pH 6.7, 0.1 M NH4Cl) for
Gd(III) of ArsIII and the di€erent complexones were
determined spectrophotometrically at 680 nm monitor-
ing the increase in concentration of the ArsIII-Gd(III)
4
complex (e=1.888 10 M 1 cm 1). Solutions contained
.
20 mM HEPES (pH 6.7), 100 mM NH4Cl, 0.1 mM
ArsIII, titrating the Gd(III) concentration up to 200 mM
in 5 mM steps. Binding simulations and non linear least
squares regressions were programmed using the Mathe-
matica v3.0 program (Wolfram Research Inc.) imple-
mented on a MacIntosh Power PC platform using the
models and equations described by Wells.33 To obtain
the ArsIII-Gd(III) apparent stability constant (K01),
binding isotherms were ®tted to a quadratic equation
of the type x2+j1x+j2=0, where j1= ([ArsIII]t+
[Gd(III)]t+K01) and j2=[ArsIII]t[Gd]t. Apparent sta-
bility constants (K0) for Gd(III) of compounds 1±4,
IDA, and EDTA were determined subsequently by
competition with ArsIII. Brie¯y, the concentration of
ArsIII-Gd(III) complex was determined in solutions
prepared as above but containing 10 mM IDA, 10 mM
compounds 1±4 or 1 mM EDTA. The sigmoidal beha-
vior obtained in the complexones was best described
considering two binding sites for Gd(III). One high
anity binding site competitive with ArsIII (K02) and
one low anity binding site non competitive with ArsIII
(K03). The following simultaneous equilibria satisfy the
behavior of this system: x/(a x y)(b x)=K01, y/
(a x y)(c y)=K02, z/(a z)(d z)=K03; where x, y or z
are the equilibrium concentrations of ArsIII-Gd(III)
complex, high anity binding sites-Gd(III) complexes
or low anity binding sites Gd(III) complexes, respec-
tively. The total concentration of Gd(III) is denoted by
a (0<a<200 mM), b refers to the total concentration of
ArsIII (100 mM), c is the total concentration of com-
plexone high anity binding sites and d the total con-
centration of low anity binding site. For the known or
estimated values of a, b, c, d, and K01 the system can be
solved for x, y, and z adjusting iteratively the values of
K02 and K03 until the calculated value of x best ®ts the
experimental data. The apparent stability constants of
complexones 1±4, IDA, and EDTA for Ca2+ were
Acknowledgements
This work was supported in part by grants D.G.I.C.Y.T
(PB-93-0037, PB-94-0011 and Acciones Integradas
HA95-14B and HA96-99B), D.G.E.S. PB96-011-C01,
PB96-0864-CO2-02), Community of Madrid (AC-07/
105/96, AC-08.1/0023/97). PL received F.P.U. (Forma-
cion de Profesorado Universitario) fellowship from the
Spanish Ministry of Education and Science. CGS
received postdoctoral fellowships from the Deutsche
Forschungsgemeinschaft and Community of Madrid.
References
1. Tweedle, M. F.; Brittain, H. G.; Eckelman, W. C.;
Gaughan, G. T.; Hagan, J. J.; Wedeking, P. W.; Runge, V. M.
In Magnetic Resonance Imaging, 2nd ed.; Partain, C. L., Price,
R. R., Patton, J. A., Kulkarni, M. V., James, A. E. W. B.,
Eds.; Saunders: Philadelphia, 1988; pp 793±809.
2. Peters, J. A.; Huskens, J.; Raber, D. J. Progress NMR
Spectroscopy 1996, 28, 283.
3. Wood, M. L.; Hardy, P. A. J. Magn. Reson. Imaging 1993,
3, 149.
4. Lau€er, R. B. Chem. Rev. 1987, 87, 901.
5. Geraldes, C. F. G. C.; Sherry, A. D.; Lazar, I.; Miseta, A.;
Bogner, P.; Berenyi, E.; Sumegi, B.; Keifer, G. E.; McMillan,
K.; Maton, F.; Muller, R. N. Mag. Res. Med. 1993, 30, 696.

Â
P. Lopez et al. / Bioorg. Med. Chem. 7 (1999) 517±527
527
6. Petre, C.; Ni, Y.; Marchal, G.; Jie, Y.; Wevers, M.; Lau€er,
R. B.; Baert, A. L. Mag. Res. Med. 1995, 35, 532.
7. Saeed, M.; Wendland, M. F.; Masui, T.; Higgins, C. B.
Mag. Res. Med. 1994, 31, 31.
Jimenez, J. A.; Cano, M.; Heras, J. V.; Pinilla, E.; Monge, A.
Organometallics, 1994, 13, 289.
21. Lopez, M. C.; Claramunt, R. M.; Ballesteros, P. Hetero-
cycles 1994, 37, 891.
8. Anderegg, G. In Comprehensive Coordination Chemistry;
Wilkinson, G., Gillard, R. D., McCleverty J. A., Eds.;
Pergamon: New York, 1987; Vol. 2, pp 777±792.
9. Jones, C. J. In Comprehensive Coordination Chemistry;
Wilkinson, G., Gillard, R. D., McCleverty J. A., Eds.; Perga-
mon: New York, 1987, Vol. 6, pp 881±1009.
10. Andrew, E. R.; Byder, G.; Griths, J.; Iles, R.; Styles, P.
Clinical Magnetic Resonance Imaging and Spectroscopy. John
Wiley & Sons: New York, 1990.
11. Berners-Price, S. J.; Sadler, P. J. Coord. Chem. Rev. 1996,
151, 1.
12. Steel, P. G. Coord. Chem. Rev. 1990, 106, 227.
13. Sorrell, T. N. Tetrahedron 1989, 45, 3.
22. Rosen, B. R.; Belliveau, J. W.; Nevea, J. M.; Brady, T.
Mag. Res. Med. 1990, 14, 249.
23. Canty, A. J.; Honeymnan, R. T. J. Organomet. Chem.
1990, 387, 247.
24. Ochi, H.; Miyasaka,T.; Arakawa, K. Yakugaku Zasshi
1978, 98, 165. Chem. Abst., 1978, 89: 24213k.
25. Torres, J.; Lavandera, J. L.; Cabildo, P.; Claramunt, R.
M.; Elguero, J. J. Heterocycl. Chem. 1988, 25, 771.
26. Koelsch, C. F.; Robinson, F. M. J. Org. Chem. 1956, 21,
1211.
27. Schwarzenbach, G.; Anderegg, G.; Schneider, W.; Senn,
H. Helv. Chim. Acta 1955, 38, 1147.
28. Scarpa, A. Methods Enzymol. 1979, 56, 301.
29. Wenzel, T. NMR Shift Reagents; CRC: Boca Raton,
1987.
14. Sadimenko, A. P.; Basson, S. S. Complexes of pyrazoles.
Coord. Chem. Rev. 1996, 147, 247.
15. Mukkala, V. M.; Kankare, J. Helv. Chim. Acta 1992, 75,
1578.
16. Mukkala, V. M.; Kwiatkowski, J.; Kankare, J.; Takalo,
H. Helv. Chim. Acta, 1993, 76, 893.
30. Koenig, S. H.; Brown, R. D. Relaxivity of MRI magnetic
contrast agents. Concepts and principles in Handbook of metal-
ligand interactions in biological ¯uids; 1995; Vol. 2, Part 4, pp
1093±1108.
17. Rulo€, R.; Arnold, K.; Beyer, L.; Dietze, F.; Grunder,
W.; Wagner, M.; Hoyer, E. Z. Anorg. Allg. Chem. 1995, 621,
807.
18. Wagner, M.; Rulo€, R.; Hoyer, E.; Grunder, W. Z. Nat-
urforch. 1997, 52c, 508.
19. Lopez Gallego-Preciado, M. C.; Ballesteros, P.; Clar-
amunt, R. M.; Cano, M.; Heras, J. V.; Pinilla, E.; Monge, A.
J. Organomet. Chem. 1993, 450, 237.
31. Desreux, J. F.; Jacques, V. Role of metal-ligand interac-
tions in the design of MRI contrast agents. Concepts and prin-
ciples in Handbook of metal-ligand interactions in biological
¯uids; 1995, Vol. 2, Part 4, pp 1109±1118.
32. Hoard, J. L.; Lee, B.; Lind, M. D. J. Am. Chem. Soc.
1965, 87, 1612.
33. Wells, J. W. Analysis and interpretation of binding at
equilibrium. In Receptor Ligand Interactions. A Practical
Approach; Hulme, E. C., Ed.; IRL: Oxford, 1992; pp 289±395.
20. Ballesteros, P.; Lopez, C.; Lopez, C.; Claramunt, R. M.;