The Water-Exchange Rate in Neutral Heptadentate DO3A-Like GdIII Complexes: Effect of the Basicity at the Macrocyclic Nitrogen Site

Article abstract of DOI:10.1002/ejic.200300356

The rate of exchange of the two water molecules coordinated to a Gd ^III centre in neutral complexes with heptadentate DO3A derivatives is modulated by the basicity of the macrocyclic nitrogen bearing the pendant group: a lower basicity results

Full text of DOI:10.1002/ejic.200300356

SHORT COMMUNICATION
The Water-Exchange Rate in Neutral Heptadentate DO3A-Like GdIII
Complexes: Effect of the Basicity at the Macrocyclic Nitrogen Site
Enzo Terreno,[a] Patrizia Boniforte,[a] Mauro Botta,[b] Franco Fedeli,[c] Luciano Milone,[a]
Armando Mortillaro,[c] and Silvio Aime*[a]
Keywords: Gadolinium / Water exchange / Imaging agents / Kinetics
The rate of exchange of the two water molecules coordinated
to a GdIII centre in neutral complexes with heptadentate
DO3A derivatives is modulated by the basicity of the macro-
cyclic nitrogen bearing the pendant group: a lower basicity
results in a slower water-exchange rate.
( Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2003)
Introduction
and metal-bound water protons) and, therefore, kex values
of about two orders of magnitude smaller than the optimal
values required for GdIII-based agents are necessary for en-
suring an efficient saturation transfer.
It is worth remembering that kex refers to the exchange
of the water protons coordinated to the metal centre and
not to the exchange of the whole water molecule (kex-w).
Nevertheless, at pH values around neutrality the proton ex-
change is extremely slow (about 103 sϪ1) when compared
with the typical kex-w values measured for LnIII chelates.
Thus, it is reasonable to assume that at physiological pH
The exchange rate (kex) of the metal-coordinated water
protons in lanthanide() complexes plays a fundamental
role in determining their efficiency as contrast agents (CA)
for Magnetic Resonance Imaging (MRI) applications.[1] In
the case of GdIII-based agents the attainment of high re-
laxivity values is strongly dependent on kex. In fact, the
paramagnetic relaxation of the water protons for a given
GdIII chelate reaches the maximum efficiency when kex is
significantly larger than the longitudinal relaxation rate of
the water protons bound to the GdIII ion, i.e. kex ϾϾ R1M
.
values kex corresponds to kex-w
.
To this regard, it has been calculated that a kex value of
about 5 ϫ 107 sϪ1 at 298 K allows the attainment of the
highest relaxivity (at 0.47 T) for a large-sized GdIII-based
system.[2] Conversely, a novel class of paramagnetic CAs
based on saturation transfer, the so-called CEST (Chemical
Exchange Saturation Transfer) agents, require much lower
kex values.[3,4] In these LnIII chelates (with Ln ϶ Gd) the
contrast in the MR image arises from the decrease of the
NMR signal intensity of the bulk water protons following
the irradiation of the water protons coordinated to the LnIII
ion. Here, the condition ∆ω Ͼ kex has to be fulfilled (∆ω is
the chemical shift separation, in rad·Hz, between the free
In the case of nine-coordinate GdIII complexes, it has
been established that the water-exchange process occurs
through a dissociative pathway in which the transition state
is represented by an eight-coordinate species.[5] Thus, the
exchange rate is related to the free-energy difference, ∆G#,
between the transition and the ground states. Several fac-
tors contribute to the ∆G# value. ∆G# is usually dominated
by the enthalpic term, ∆H#, which mainly depends on the
strength of the LnϪO(water) bond. The highly electrostatic
nature of this bond accounts for the marked dependence
of the water-exchange rate on the overall electric charge in
isostructural LnIII chelates. In fact, relatively fast and ex-
ceedingly slow exchange rates have been observed for
anionic and cationic GdIII complexes, respectively.[1] More-
over, the ∆H# term may be affected either by the steric
crowding at the water coordination site or by the structural
properties of the second hydration sphere of the metal
ion.[6Ϫ8] The large number of parameters that contribute to
determine the water-exchange process renders the assess-
ment of the relative role played by each single contribution
quite difficult. For this reason, the evaluation of the kex val-
ues for GdIII complexes differing in only one main struc-
[a]
`
Dipartimento di Chimica I.F.M., Universita di Torino
Via P. Giuria 7, 10125 Torino, Italy
Fax: (internat.) ϩ39-011/670-7855
E-mail: silvio.aime@unito.it
[b]
[c]
`
Dipartimento di Scienze e Tecnologie Avanzate, Universita del
Piemonte Orientale ‘‘Amedeo Avogadro’’
Corso Borsalino 54, 15100 Alessandria, Italy
Fax: (internat.) ϩ39-0131/287-431
E-mail: mauro.botta@unipmn.it
Laboratorio Integrato di Metodologie Avanzate, Bioindustry
Park del Canavese
Via Ribes 5, Colleretto Giacosa (TO), Italy
Fax: (internat.) ϩ39-0125/538-346
3530
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/ejic.200300356
Eur. J. Inorg. Chem. 2003, 3530Ϫ3533

Neutral Heptadentate DO3A-Like GdIII Complexes
SHORT COMMUNICATION
tural or electronic feature may be particularly useful. On owing to the acid-catalysed demetallation process, whereas
this basis, it was deemed of interest to investigate the depen- at higher pH values the relaxivity decreases owing to the
dence of the water-exchange rate on the basicity of the formation of ternary adducts with anionic species like car-
macrocyclic nitrogen bearing the non-coordinating sub- bonate (from dissolved CO2) and hydroxide anions which
stituent in two neutral heptadentate DO3A-like GdIII che- displace the metal-bound water molecule(s). This behav-
lates. To this purpose, the ligands reported below, and their iour, coupled with the relatively high r1 values {8.0 and 8.5
corresponding EuIII, GdIII and YbIII complexes, have been sϪ1·mϪ1 at 298 K and pH ϭ 7 for [Gd(NO2PhDO3A)]
synthesised. In spite of a very similar structure, the two and [Gd(NH2PhDO3A)], respectively} provides a strong in-
LnIII chelates have a noticeable difference in the electronic dication that the hydration sphere of the two complexes is
availability at the macrocyclic nitrogen bearing the phenyl closely analogous to that of the parent [Gd(DO3A)] com-
group [σpara(NO2) ϭ 0.78, σpara(NH2) ϭ Ϫ0.66].
plex,[10] i.e. characterised by the presence of two water mol-
ecules coordinated to the metal ion. Nevertheless, it is likely
that the complexes are present in solution as a mixture of
species which may differ in the conformation of the donor
atoms wrapped around the metal centre (∆/Λ and δδδδ/
λλλλ isomerism)[11] or in the hydration state of the metal
ion (equilibrium between nine- and eight-coordinate iso-
mers).[10,12]
In principle, each species may have a different kex value[13]
and thus the isomeric population of the two complexes has
to be checked.
1
Figure 2 shows a comparison of the H NMR high reso-
lution spectra of the EuIII chelates. In spite of the com-
plexity of the spectra, it is evident that they are very similar
in terms of chemical shift range and number and relative
intensity of the resonances. Analogous conclusions can be
drawn by analysing the 1H NMR spectra of the correspond-
ing YbIII complexes.
Results and Discussion
The synthesis of the ligands (see Exp. Sect.) was carried
out by reacting 1,4,7,10-tetraazacyclododecane (TAZA)
with p-nitrofluorobenzene (molar ratio 3:1) in acetonitrile.
The purified mono-alkylated TAZA derivative was then
exhaustively carboxymethylated with bromoacetic acid at
pH 10. The NH2PhDO3A ligand (pale brown) was ob-
tained upon hydrogenation (Pd/C) of the nitro derivative.
The synthesis of the LnIII complexes was carried out in
water at pH 7 by mixing equimolar amounts of the ligand
and the corresponding LnIII chloride.
Figure 1 shows the pH dependence of the relaxivity, r1,
for the two GdIII complexes at 0.235 T (corresponding to a
proton Larmor frequency of 10 MHz) and 298 K.
1
Figure 2. H NMR spectra at 7.05 T, pH 7 and 275 K in D2O of
[Eu(NH2PhDO3A)] (top) and [Eu(NO2PhDO3A)] (bottom)
Further confirmation that the isostructurality between
the two LnIII chelates is retained for the GdIII complexes
was gained by the analysis of their NMRD (Nuclear Mag-
netic Relaxation Dispersion) profiles (Figure 3). Again, the
two complexes display a very similar magnetic-field depen-
dence of their r1 values (measured at pH 7 and 298 K) even
if the relaxivity of the p-nitro derivative is slightly lower
over the entire range of magnetic-field strengths investi-
gated.
Figure 1. pH dependence of relaxivity for [Gd(NO2PhDO3A] (᭿)
and [Gd(NH2PhDO3A] (ᮀ) at 298 K and 10 MHz
Qualitatively, both chelates behave similarly to other neu-
Finally, the exchange rate of the water molecules coordi-
tral heptadentate DO3A-like complexes.[9] r1 is constant in nated to GdIII has been determined through the well-estab-
the pH interval 5 and 8; at lower pH values r1 increases lished NMR method based on the analysis of the tempera-
Eur. J. Inorg. Chem. 2003, 3530Ϫ3533
www.eurjic.org
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3531

S. Aime et al.
SHORT COMMUNICATION
Table 1. Best-fit parameters (at 298 K) obtained from the analysis
of 17O-R2p versus
T
profiles and NMRD profiles of
[Gd(NO2PhDO3A)] and [Gd(NH2PhDO3A)] at pH 7
[Gd(NO2PhDO3A)]
[Gd(NH2PhDO3A)]
NMRD
NMRD[a] 17O-R2p
17O-R2p
[b]
[c]
[c]
kex (ϫ 106sϪ1
)
7.4 Ϯ 1.2
33.8 Ϯ 1.6
17.6 Ϯ 3.8
36.2 Ϯ 2.8
[d]
[d]
∆HM (kJ·molϪ1
)
∆2 (ϫ 1019 sϪ2
τV (ϫ 10Ϫ12 s)
τR (ϫ 10Ϫ12 s)
)
5.8 Ϯ 0.5 11.6 Ϯ 0.8 5.7 Ϯ 0.4 9.8 Ϯ 0.4
22.0 Ϯ 0.3 13.0 Ϯ 0.6 20.3 Ϯ 0.7 15.0 Ϯ 0.7
77.0 Ϯ 0.4
[d]
[d]
81.1 Ϯ 0.3
[a]
The fitting of the NMRD profiles has been carried out by keep-
˚
˚
ing fixed the following parameters: r ϭ 3.1 A, q ϭ 2, a ϭ 3.8 A,
D ϭ 2.25 ϫ 10Ϫ5 cm2·sϪ1 [b] The fitting of the temperature depen-
.
dence of 17O-R2p has been performed by keeping fixed the follow-
Figure 3. 1H NMRD profiles for [Gd(NO2PhDO3A)] (᭿) and
[Gd(NH2PhDO3A)] (ᮀ) at pH 7 and 298 K
-
ing parameters: ∆Hv ϭ 5.0 kJ·molϪ1, (A/h ) ϭ Ϫ3.8 ϫ 106, q ϭ 2.
[c] This parameter has been fixed to the value obtained by the fitting
of the data obtained with the other experimental technique. [d] This
parameter cannot be obtained from the fitting.
ture dependence of the paramagnetic contribution to the
transverse relaxation rate of the 17O-water nuclei. Figure 4
reports the data obtained for the two complexes.
limitations of the currently available theory for describing
adequately the magnetic-field dependence of the electronic
relaxation times, but, fortunately, this concern has only a
small effect on the estimation of kex. Therefore, the relax-
ometric analysis indicates that the water exchange for the
p-amino derivative occurs faster than the p-nitro complex.
It seems reasonable that such a difference could be ac-
counted for in terms of the different electronic properties
of the two p-substituents on the aromatic ring. The strong
electron-withdrawing character of the p-nitro group makes
the GdϪN bond much more polarised, thus increasing the
positive charge at the metal centre. This effect results in
a stabilisation of the nine-coordinate ground state of the
complex, thus slowing down the rate of exchange of the
coordinated water molecules. The opposite electronic effect
of the p-amino group leads to a destabilisation of the
GdIIIϪO(water) bond and to a faster exchange rate.
Figure
4.
Temperature
dependence
of
17O-R2p
for
[Gd(NO2PhDO3A)] (᭿) and [Gd(NH2PhDO3A)] (ᮀ) at pH 7 and
7.05 T ([GdL] ϭ 10 m)
Interestingly, the water-exchange regime reported for
[Gd(DO3A)] (11.0 ϫ 106 sϪ1 [10] is intermediate as expected
)
on the ground of the basicity of the secondary amino group.
The results reported here might also help to rationalise
the fast water-exchange rates (ca. 5.0 ϫ 107 sϪ1 at 298 K)
reported in the literature for several N-benzyl-DO3A de-
rivatives[14] and support the hypothesis forwarded for ex-
plaining the relatively fast exchange rate observed in a cat-
ionic N-methyl DO3A derivative.[15] In fact, in these com-
plexes a destabilisation of the GdIIIϪO(water) bond is ex-
pected on the basis of the inductive effect (ϪI) of the alkyl
group on the macrocyclic nitrogen.
The profiles appear rather different, with the top of the
bell-shaped curve shifted to lower temperature for the p-
amino derivative. This behaviour suggests that kex is signifi-
cantly influenced by the nature of the substituent at the
macrocyclic nitrogen.
A quantitative analysis of the 17O NMR spectroscopic
data, according to the available theory,[5] confirms this ef-
fect. The result of the best-fitting procedure, on the assump-
tion that the two labile water molecules have the same ex-
change rate, is reported in Table 1, along with the results
obtained from the analysis of the NMRD profiles carried
out according to the usual inner- and outer-sphere model.[1]
No relevant differences were obtained, within the same
experimental data set, for the parameters controlling the
Conclusion
In conclusion, the experimental evidence reported in this
electronic relaxation times ∆2 and τV, which represent the communication supports the view that the exchange rate of
mean energy of the transient zero-field splitting (ZFS) and the water molecules coordinated to the GdIII ion in a nine-
its time modulation, respectively. Nevertheless, different ∆2 coordinate DO3A-like arrangement may be successfully
and τV values were obtained depending on the experimental modulated through a proper choice of the electronic
technique. It is likely that this discrepancy is due to the properties of the substituent on the macrocyclic nitrogen.
3532
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2003, 3530Ϫ3533

Neutral Heptadentate DO3A-Like GdIII Complexes
SHORT COMMUNICATION
ding an equimolar amount of LnCl3 to the selected ligand. The pH
was monitored during the synthesis in order to keep it in the 6Ϫ8
range. The final concentration of the metal chelates was checked
by Evans’ method.16
However, whether the two water molecules exchange at the
same rate and/or the substituent effect described herein af-
fects the dissociation rate at the two coordination sites to a
different extent still remain undefined.
NMR Measurements: The high resolution NMR spectra and the
17O-R2p measurements were performed on a Bruker Avance 300
spectrometer (Karlsruhe, Germany) operating at 7.05 T (proton
Larmor frequency of 300 MHz).
Experimental Section
Synthesis of NO2PhDO3A: This ligand was synthesised in two
steps. First,
a suspension of 1,4,7,10-tetraazacyclododecane,
The water-proton longitudinal relaxation rates at 0.235 T (proton
Larmor frequency of 10 MHz) as well as the NMRD profiles in
the 0.000235Ϫ0.235 T range (0.01Ϫ10 MHz) were obtained from
a Stelar Fast Field Cycling Relaxometer (Mede (Pv), Italy). The
temperature was controlled with a Stelar VTC-91 air-flow heater
equipped with a copper-constantan thermocouple (uncertainty of
Ϯ0.1 °C). The NMRD profile was completed with data obtained
at 20 MHz and 90 MHz by using a Stelar Spinmaster and a JEOL
EX-90 NMR spectrometer, respectively.
TAZA, (35.05 g, 0.2035 mol) in acetonitrile (500 mL) was dissolved
whilst stirring at 60 °C by adding 50 mL of water. To this solution
was then added dropwise 50 mL of a solution of acetonitrile con-
taining 1-fluoro-4-nitrobenzene (9.57 g, 0.0678 mol). After 8 h at
60 °C the suspension was cooled down and filtered. The crude solid
was dissolved in 1 L of boiling water and the resulting solution
(obtained by filtration) was first diluted 1:1 with water, loaded onto
a chromatographic column (Amberlite XAD 1600) and then
eluted with water in order to separate the unchanged TAZA. The
1-(4-nitrophenyl)-1,4,7,10-tetraazacyclododecane ligand (14.08 g;
0.048 mol, 70.8% yield) was obtained by eluting the column with
methanol and subsequent evaporation of the solvent.
Acknowledgments
Financial support from MIUR and Bracco Imaging S.p.A (Milano,
Italy) is gratefully acknowledged. This work was carried out under
the framework of the COST-D18 action. A. M. is recipient of a
Research Fellowship from the Fondazione Internazionale di
Ricerca in Medicina Sperimentale, Torino, Italy.
In the second step, 2-bromoacetic acid (22.23 g, 0.16 mol) was ad-
ded whilst stirring to an aqueous solution of 1-(4-nitrophenyl)-
1,4,7,10-tetraazaciclododecane (11.73 g, 0.04 mol in 50 mL of
water). The pH of the solution was brought to 10 by adding 2 
NaOH. The solution was heated at 50 °C for 6 h by keeping the
pH constant. The reaction mixture was cooled down, filtered and
the filtrate was acidified at pH 3 with HCl. This solution was
loaded onto a chromatographic column (Amberlite XAD 1600)
and then eluted first with water, in order to separate the bromide
ions, and then with methanol. After evaporation of the solvent the
NO2PhDO3A ligand was obtained in the acid form as a yellow
[1]
The Chemistry of Contrast Agents in Medical Magnetic Reson-
ance Imaging (Eds.: A. E. Merbach; E. Toth), 2001, John
Wiley & Sons, Chichester (UK), Chapters 2 and 5.
S. Aime, M. Botta, M. Fasano, E. Terreno, Chem. Soc. Rev.
[2]
1998, 27, 19Ϫ29.
S. Zhang, P. Winter, K. Wu, A. D. Sherry, J. Am. Chem. Soc.
[3]
1
solid (16.38 g, 0.035 mol, 87.6% yield). H NMR (D2O, tBuOH as
2001, 123, 1517Ϫ1518.
S. Aime, D. Delli Castelli, E. Terreno, Angew. Chem. Int. Ed.
[4]
internal reference, 298 K): δ ϭ 8.0 (d, 2 H), 6.65 (d, 2 H), 3.7 (s, 4
H), 3.45 (s, 4 H), 3.35 (s, 2 H), 3Ϫ3.2 (m, 12 H) ppm. 13C{1H}
NMR (D2O, tBuOH as internal reference, 298 K): δ ϭ 173.2, 169.2,
153.9, 142.7, 124.2, 122.3, 55.1, 51.6, 50.9, 50.0, 47.7, 46.9 ppm.
MS (MALDI-TOF): calcd. for C20H29N5O8 467.47 uma; found
468.4 [MHϩ].
2002, 41, 4334Ϫ4336.
H. D. Powell, O. M. Ni Dhubhghaill, D. Pubanz, L. Helm, Y.
[5]
Lebedev, W. Schlaepfer, A. E. Merbach, J. Am. Chem. Soc.
1996, 118, 9333Ϫ9346.
D. Pubanz, G. Gonzalez, D. H. Powell, A. E. Merbach, Inorg.
[6]
Chem. 1995, 34, 4447Ϫ4453.
S. Aime, M. Botta, S. Geninatti Crich, G. B. Giovenzana, R.
[7]
Synthesis of NH2PhDO3A: Ammonia solution (26%) was added to
a suspension of NO2PhDO3A in water (7.70 g; 0.02 mol in 50 mL
of water) until the ligand dissolution was complete. Then, 0.7 g of
10% Pd/C, previously treated with 5 mL of water, was added to the
solution and the catalytic hydrogenation was carried out at room
temperature and pressure with hydrogen (1180 mL, 0.05 mol) pro-
duced electrochemically (Hydrogen generator HG 600; CLAIND).
After 4 h the suspension was filtered in order to remove the cata-
lyst. The filtrate (brown colour) was lyophilised and then redis-
solved in water acidified (pH 2.3) with HCl. 0.5 g of Carbopuron
was then added in order to partially decolourise and stabilise the
product. After filtration, the solution was re-lyophilised and the
NH2PhDO3A ligand was obtained in the acid form as a dark-
brown solid (8.16 g; 0.016 mol, 80% yield). 1H NMR (D2O, tBuOH
as internal reference, 298 K): δ ϭ 7.15 (d, 2 H), 6.7 (d, 2 H), 3.6
(s, 4 H), 3.4 (s, 4 H), 3.3 (s, 2 H), 3.27 (s, 8 H), 3.1 (s, 4 H) ppm.
13C{1H} NMR (D2O, tBuOH as internal reference, 298 K): δ ϭ
176.8, 169.1, 143.6, 138.9, 124.8, 116.0, 55.7, 54.2, 51.0, 50.2, 49.5,
47.9 ppm. MS (MALDI-TOF): calcd. for C20H31N5O6 437.5 uma;
found 438.3 [MHϩ].
Pagliarin, M. Piccinini, M. Sisti, E. Terreno, J. Biol. Inorg.
Chem. 1997, 2, 470Ϫ479.
[8]
M. Botta, Eur. J. Inorg. Chem. 2000, 399Ϫ407.
[9]
S. Aime, M. Botta, S. Geninatti Crich, G. Giovenzana, R. Pag-
liarin, M. Sisti, E. Terreno, Magn. Res. Chem. 1998, 36,
S200ϪS208.
E. Toth, O. M. Ni Dhubhghaill, G. Besson, L. Helm, A. E.
Merbach, Magn. Res. Chem. 1999, 37, 701Ϫ708.
P. Caravan, J. J. Ellison, T. J. McMurry, R. B. Lauffer, Chem.
Rev. 1999, 99, 2293Ϫ2352.
F. A. Dunand, S. Aime, S. G. Crich, G. B. Giovenzana, A. E.
Merbach, Magn. Res. Chem. 2002, 40, 87Ϫ92.
S. Aime, A. Barge, M. Botta, D. Parker, A. S. De Sousa, Angew.
Chem. Int. Ed. 1998, 37, 2673Ϫ2675.
S. Aime, E. Gianolio, E. Terreno, G. B. Giovenzana, R. Paglia-
rin, M. Sisti, G. Palmisano, M. Botta, M. P. Lowe, D. Parker,
[10]
[11]
[12]
[13]
[14]
J. Biol. Inorg. Chem. 2000, 5, 488Ϫ497.
M. Botta, S. Aime, A. Barge, G. Bobba, R. S. Dickins, D.
Parker, E. Terreno, Chem. Eur. J. 2003, 9, 2102Ϫ2109.
D. M. Corsi, C. Platas-Iglesias, H. van Bekkum, J. A. Peters,
[15]
[16]
Magn. Reson. Chem. 2001, 39, 723Ϫ726.
Received June 9, 2003
Synthesis of the LnIII Complexes: The LnIII complexes (Ln ϭ Gd,
Early View Article
Eu and Yb) were synthesised in water at room temperature by ad-
Published Online August 14, 2003
Eur. J. Inorg. Chem. 2003, 3530Ϫ3533
www.eurjic.org
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3533