Thermodynamic and structural properties of Gd3+ complexes with functionalized macrocyclic ligands based upon 1,4,7,10-tetraazacyclododecane

Article abstract of DOI:10.1039/a909098c

The co-ordination properties toward Na⁺, Ca²⁺, Zn²⁺, Cu²⁺ and Gd³⁺ of six ligands containing pendant functionalities attached to the common 1,4,7,10-tetraazacyclododecane macrocyclic ring have been studied by means of potentiometric and microcalorimetric methods. Stability constants for the l,4,7,10-tetra(methylcarbamoylmethyl) ligand (DTMA), and for the new 10-(2-hydroxypropyl) 1,4,7-tri(2-hydroxymethylacetic acid) (HPDO3A-3HM) and 1,4,7-tri(2-hydroxymethylacetic acid) ligand (DO3A-3HM) are reported for the first time, while analogous data for the remaining ligands have been redetermined under the same experimental conditions (0.1 mol dm^-3 NMe₄Cl, 298.1 ±0.1 K). The synthesis of DTMA is also described. The stability of the GdL complexes follows the order DOTA (1,4,7,10-tetraacetic acid) ≈ HPDO3A [10-(2-hydroxypropyl) 1,4,7-triacetic acid] > DO3A (1,4,7-triacetic acid) > HPDO3A-3HM > DO3A-3HM > DTMA, although considering both complex formation and ligand basicity the binding selectivity of these ligands at physiological pH follows the order HPDO3A > DOTA > HPDO3A-3HM > DO3A-3HM > DO3A > DTMA. The thermodynamic resistance of the Gd³⁺ complexes, at a plasma concentration of clinical administrations, toward demetallation in the presence of important components of blood plasma, such as Na⁺, Ca²⁺, Cu²⁺, Zn₂₊, phosphate and carbonate, has been evaluated: the results are consistent with the previous trend of binding selectivity. The formation of intermediate complexed species formed in the first minutes of the complexation reactions has been studied from a thermodynamic point of view. The crystal structure of [Gd(DTMA)(H₂O)][ClO₄]₃-NaClO₄3H2O) solved by X-ray analysis, displays the Gd³⁺ ion in a 9-co-ordinated environment defined by four nitrogens of the tetraazamacrocyclic moiety, eight amidic oxygen atoms of the side arms and by an oxygen atom of a water molecule. The Royal Society of Chemistry 2000.

Full text of DOI:10.1039/a909098c

View Article Online / Journal Homepage / Table of Contents for this issue
Thermodynamic and structural properties of Gd3؉ complexes
with functionalized macrocyclic ligands based upon 1,4,7,10-
tetraazacyclododecane
Antonio Bianchi,*a Luisella Calabi,*b Claudia Giorgi,a Pietro Losi,b Palma Mariani,a
Paola Paoli,c Patrizia Rossi,c Barbara Valtancolia and Mario Virtuanib
a Department of Chemistry, University of Florence, Via Maragliano 75/77, 50144 Florence,
Italy. E-mail: bianchi@chim1.unifi.it
b Research and Development Division, Bracco SpA, Via E. Folli 50, 20134 Milan, Italy
c Department of Energetics, University of Florence, Florence, Italy
Received 17th November 1999, Accepted 12th January 2000
Published on the Web 7th February 2000
The co-ordination properties toward Naϩ, Ca2ϩ, Zn2ϩ, Cu2ϩ and Gd3ϩ of six ligands containing pendant
functionalities attached to the common 1,4,7,10-tetraazacyclododecane macrocyclic ring have been studied by means
of potentiometric and microcalorimetric methods. Stability constants for the 1,4,7,10-tetra(methylcarbamoylmethyl)
ligand (DTMA), and for the new 10-(2-hydroxypropyl) 1,4,7-tri(2-hydroxymethylacetic acid) (HPDO3A-3HM) and
1,4,7-tri(2-hydroxymethylacetic acid) ligand (DO3A-3HM) are reported for the first time, while analogous data for
the remaining ligands have been redetermined under the same experimental conditions (0.1 mol dmϪ3 NMe4Cl,
298.1 0.1 K). The synthesis of DTMA is also described. The stability of the GdL complexes follows the order
DOTA (1,4,7,10-tetraacetic acid) ≈ HPDO3A [10-(2-hydroxypropyl) 1,4,7-triacetic acid] > DO3A (1,4,7-triacetic
acid) > HPDO3A-3HM > DO3A-3HM > DTMA, although considering both complex formation and ligand
basicity the binding selectivity of these ligands at physiological pH follows the order HPDO3A > DOTA >
HPDO3A-3HM > DO3A-3HM > DO3A > DTMA. The thermodynamic resistance of the Gd3ϩ complexes, at a
plasma concentration of clinical administrations, toward demetallation in the presence of important components of
blood plasma, such as Naϩ, Ca2ϩ, Cu2ϩ, Zn2ϩ, phosphate and carbonate, has been evaluated: the results are consistent
with the previous trend of binding selectivity. The formation of intermediate complexed species formed in the first
minutes of the complexation reactions has been studied from a thermodynamic point of view. The crystal structure
of [Gd(DTMA)(H2O)][ClO4]3ؒNaClO4ؒ3H2O, solved by X-ray analysis, displays the Gd3ϩ ion in a 9-co-ordinated
environment defined by four nitrogens of the tetraazamacrocyclic moiety, eight amidic oxygen atoms of the side
arms and by an oxygen atom of a water molecule.
rapid exchange with the bulk solvent, in order efficiently to
Introduction
enhance the relaxation of solvent protons.1
Over the last decade there has been an increasing interest in the
chemistry of Gd3ϩ complexes because they have proved useful
in diagnostic medicine as contrast agents for in vivo magnetic
resonance imaging (MRI).1–5 This technique furnishes images
of anatomical districts which are topological representations of
the 1H NMR parameters, and hence the presence of the highly
paramagnetic metal center, resulting in a large enhancement of
water proton relaxation rate, improves the resolution of tissues.
Unfortunately, the applicability of Gd3ϩ complexes as con-
trast agents is strongly limited by their toxicity. The toxic effect
of these metal complexes principally stems, apart from their
intrinsic toxicity, from the metal ion and the ligand released
before complex excretion. Numerous competitive equilibria,
such as ligand protonation, Ca2ϩ, Zn2ϩ, Cu2ϩ complexation and
precipitation of Gd3ϩ salts, to mention the principal ones, may
contribute to the in vivo dissociation of similar Gd3ϩ complexes.
To avoid complex dissociation, Gd3ϩ chelates characterized
by high thermodynamic stability and possibly kinetic inertness
are commonly sought.1 A convenient approach to obtain Gd3ϩ
complexes with high thermodynamic stability has been the
association of the high stability conferred to metal complexes
by macrocyclic ligands with the neutralization of metal
ion charge provided by negatively charged pendant groups.5
Furthermore, macrocyclic ligands form metal complexes which
are commonly very inert toward dissociation.
Ligands such as DOTA, a twelve-membered tetraazamacro-
cycle bearing four carboxymethyl (acetate) pendant groups, and
HPDO3A, the DOTA analogue containing only three acetate
groups (Chart 1), combine these properties and have been
successfully employed in the preparation of sufficiently safe
contrast agents. Their Gd3ϩ complexes are presently in use in
diagnostic MRI with the brand names DOTAREMTM and
PROHANCETM, respectively.
The enhancement of water proton relaxation rate by Gd3ϩ
complexes is ascribed to three principal phenomena, the first
consisting in the exchange of co-ordinated water molecules
with water molecules of the bulk solvent (inner-sphere contribu-
tion), the second involving the diffusion of water molecules in
proximity of the paramagnetic center (outer-sphere contribu-
tion), and the last determined by prototropic exchange of
mobile protons of the complex (prototropic contribution).1,6
The contribution of prototropic exchange is commonly very
difficult to assess, because this process is much slower than the
exchange of co-ordinated water molecules.7 The exchange life-
time of Gd3ϩ-co-ordinated water molecules, however, increases
with decreasing negative charge of the Gd3ϩ complex, allow-
ing the evaluation of the prototropic contribution.6,8 For this
reason Gd3ϩ complexes with neutral ligands have been con-
sidered. In a recent communication, dealing with the tripositive
Gd3ϩ complex of DTMA (Chart 1), a distinct evaluation of
water and prototropic exchange rates for a Gd3ϩ-co-ordinated
solvent molecule was reported for the first time, evidencing that
In addition the complex should present at least a water
molecule in the first co-ordination sphere of the metal ion, in
DOI: 10.1039/a909098c
J. Chem. Soc., Dalton Trans., 2000, 697–705
This journal is © The Royal Society of Chemistry 2000
697

View Article Online
crystallization water by Karl Fischer’s method, mass spec-
troscopy (ESI), 1H and 13C NMR, and HPLC, which verified the
stoichiometry and purity. The lack of residual Naϩ in the sam-
ples was indicated by means of atomic absorption analysis.
Synthesis of DTMA
The synthesis of 1,4,7,10-tetrakis(methylcarbamoylmethyl)-
1,4,7,10-tetraazacyclododecane (DTMA) was carried out by
reaction of cyclen (1,4,7,10-tetraazacyclododecane) with 2-
bromo-N-methylacetamide, according to a procedure adopted
for similar ligands.13
2-Bromo-N-methylacetamide. (CAUTION: this compound is
an irritant and must be handled with extreme care.) A solution
of BrCH2COBr (34.7 cm3, 400 mmol) in 1,2-dichloroethane (50
cm3) was added over a period of 90 minutes, with vigorous
stirring, to an aqueous solution of MeNH2 (400 mmol, 34.7
cm3) and NaOH (400 mmol, 16 g) at 10 ЊC. After the addition
was complete, the suspension was maintained at this temper-
ature for 50 minutes and then allowed to warm to room
temperature. After filtration the organic phase was dried over
anhydrous sodium sulfate, filtered and evaporated under
reduced pressure to give a yellowish oil, which was dissolved in
aqueous NaOH solution and extracted with diethyl ether. The
ether solution was dried over anhydrous sodium sulfate, filtered
and evaporated to dryness to give 24.5 g (40.3%) of a white
solid. Calc. (found) for C3H6BrNO: C, 23.71 (23.5); H, 3.98
Chart 1
1
(4.0); N, 9.21 (9.0%). H NMR (CDCl3): δ 6.88 (s, NH, 1 H),
3.83 (s, CH3, 3 H) and 2.80 (d, CH2, 2 H). 13C NMR (CDCl3):
equivalent proton relaxivities can be achieved by exploiting
δ 166.31 (s, CO, 1C), 28.94 (s, CH3, 1C) and 26.83 (s, CH2, 1C).
different routes.9
ϩ
ϩ
᎐
MS: 58.1 m/z ([O᎐CNHMe] ) and 72.0 ([CH ᎐COHNHMe] ).
Preliminary results concerning the speciation of the Gd3ϩ
/
᎐
᎐
2
DTMA system and the crystal structure of the [Gd(DTMA)]3ϩ
complex were reported in a successive communication from our
laboratories, in which the pH dependence of relaxivity was
interpreted in terms of the species formed in solution at
equilibrium.10 The use of the uncharged ligand, however, while
allowing one to assess the contribution of prototropic
exchange, leads to a Gd3ϩ complex of reduced stability.
In the present paper we report the results of a thermo-
dynamic study performed to evidence how the loss of complex
stability determined by the reduction of negative charge of lig-
ands can be compensated by appropriate modifications of the
ligand structures. To this purpose the six ligands reported in
Chart 1 have been considered; two of them, HPDO3A-3HM
and DO3A-3HM, were designed and synthesized intentionally.
The mechanism of Gd3ϩ complexation by similar ligands is
commonly composed of different stages leading to the
formation of intermediate species which undergo a slow
reorganization to give the final complexes. In order to get a
better understanding of the complexation process, we have
studied the formation of such intermediate complexes with the
considered ligands, by determining their stability constants and
the relevant enthalpic and entropic parameters. This is the first
microcalorimetric investigation on similar systems.
1,4,7,10-Tetrakis(methylcarbamoylmethyl)-1,4,7,10-tetra-
azacyclododecane (DTMA). 1,4,7,10-Tetraazacyclododecane
(Fluka) (8.08 mmol, 1.4 g) was suspended in 200 cm3 of
anhydrous CH3CN containing 80.8 mmol (14.2 cm3) of diiso-
propylethylamine. To this suspension was added a solution of
2-bromo-N-methylacetamide (6.3 g, 41.45 mmol) in 150 cm3
CH3CN over a period of 5 h, at room temperature. The resulting
solution was concentrated under reduced pressure and diethyl
ether added to precipitate a first portion of pure DTMA, as
a white solid, which was filtered off and dried in vacuum. The
ether solution was evaporated to dryness and the crude product
dissolved in the minimum quantity of chloroform and chrom-
atographed (flash chromatography) on silica gel, eluting with a
CHCl3–MeOH–NH3(30%) (6:3:0.5). The portions of DTMA
were combined to give 2.7 g (66%) of pure compound. Calc.
(found) for C5H10N2O: C, 52.61 (52.5); H, 8.83 (8.8); N, 24.54
(24.4%). 1H NMR (CDCl3): δ 4.96 (s, NH, 1 H), 3.17 (s,
CH2CO, 8 H), 2.83 (s, CH3, 12 H) and 2.77 (s, CH2CH2, 16 H).
13C NMR (CDCl3): δ 175.79 (s, CO, 4C), 60.53 (s, CH2, 4C),
55.58 (s, CH2CH2, 8C) and 27.68 (s, CH3, 4C). MS: m/z 457.57
([N ϩ H]ϩ).
Synthesis of [Gd(DTMA)(H2O)][ClO4]3ؒNaClO4ؒ3H2O
Finally, in the present paper we have also considered the
competitive complexation equilibria with endogenous species,
such as Naϩ, Ca2ϩ, Zn2ϩ and Cu2ϩ, and precipitation of gado-
linium() phosphate, carbonate and hydroxide to estimate the
thermodynamic resistance of these metal complexes towards
demetallation processes promoted by principal components of
blood plasma.
An aqueous solution (50 cm3) of [Gd(DTMA)]3ϩ was prepared
by adding GdCl3 (0.1 mmol) to a hot solution of DTMA (0.1
mmol). Crystals of [Gd(DTMA)(H2O)][ClO4]3ؒNaClO4ؒ3H2O
suitable for X-ray analysis were obtained by slow evaporation at
room temperature of this solution after addition of NaClO4
(0.5 g). Calc. (found) for C20H48Cl4GdN8NaO24: C, 21.71 (21.6);
H, 4.37 (4.3); N, 10.12 (10.1)%.
Experimental
Crystal structure determination of [Gd(DTMA)(H2O)][ClO4]3ؒ
NaClO4ؒ3H2O
The compounds DOTA, DO3A and HPDO3A were syn-
thesized as described.5c,11 The synthesis of the new ligands
HPDO3A-3HM and DO3A-3HM has been reported.12 All
these ligands were obtained in acidic form and characterized
by means of elemental analysis, including determination of
Crystal data. C20H48Cl4GdN8NaO24, M = 1106.70, triclinic,
a = 11.558(6), b = 13.251(9), c = 14.713(9) Å, α = 94.29(9), β =
109.65(8), γ = 102.97(6)Њ, U = 2040(2) Å3, T = 298 K, space
698
J. Chem. Soc., Dalton Trans., 2000, 697–705

View Article Online
group P1, (no. 2), Z = 2, µ(Mo-Kα) = 1.993 mmϪ1. 7410 Reflec-
thermostat which was maintained at the chosen temperature
¯
tions measured (up to 2θ = 50), 6377 with I > 2σ(I), corrected
for Lorentz-polarization effects, absorption correction by the
Walker and Stuart method.14 Direct methods (SIR 9215), full-
matrix least squares (SHELXL 9316), atomic scattering factors
and anomalous dispersion corrections for all atoms from ref.
17. Anisotropic displacement parameters for all the non-hydro-
gen atoms, hydrogen atoms of the macrocycle in calculated
positions, with an overall thermal parameter factor refined up
to 0.053(4) Å2. The Fourier difference map did not allow us to
localize the hydrogen atoms of the Gd co-ordinated and crystal-
lization water molecules. Final R[I > 2σ(I)] was 0.0496.
CCDC reference number 186/1797.
within 2 × 10Ϫ4 K. The microcalorimeter was checked by
determining the enthalpy of reaction of strong base (Me4-
NOH) with strong acid (HCl) solutions. The value obtained,
Ϫ13.55(5) kcal molϪ1, was in agreement with the literature
values.21 Further checks were performed by determining the
enthalpies of protonation of ethylenediamine.
In a typical experiment, a Me4NOH solution (0.1 mol dmϪ3
,
addition volumes 15 µl) was added to acidic solutions (1.5 cm3)
of the ligands (5 × 10Ϫ3 mol dmϪ3), containing Gd3ϩ (4 × 10Ϫ3
mol dmϪ3) in the complexation experiments. Further complex-
ation experiments were performed for each system by titrating
ligand solution with Gd3ϩ solutions. Corrections for the heats
of dilution were applied. The corresponding enthalpies of reac-
tion were determined from the calorimetric data by means of a
least squares fitting using the AAAL program.22 No confident
results were obtained for the fifth protonation step of DO3A
and hence the experiments performed with this ligand were
limited to the formation of the first four protonated species.
See http://www.rsc.org/suppdata/dt/a9/a909098c/ for crystal-
lographic files in .cif format.
Potentiometric measurements
All pH-metric measurements (pH = Ϫlog [Hϩ]) employed for
the determination of protonation and complexation constants
were carried out in 0.1 mol dmϪ3 NMe4Cl solutions at 298.1
0.1 K, by using equipment and the methodology that has
been described.18 The combined Ingold 405 S7/120 electrode
was calibrated as a hydrogen concentration probe by titrating
known amounts of HCl with CO2-free NaOH solutions and
determining the equivalence point by Gran’s method19 which
allows one to determine the standard potential EЊ and the ionic
product of water (pKw = 13.83(1) at 298.1 0.1 K in 0.1 mol
dmϪ3 NMe4Cl). Empirical corrections were applied for the
non-Nernstian response of the glass electrode below pH 2.5
and above pH 10.5.
Speciation calculation for multi-equilibrium systems
The amounts of all species formed in the presence of Gd3ϩ
complexes at a concentration of clinical administrations (2.5 ×
10Ϫ5 mol dmϪ3) and Naϩ (0.15 mol dmϪ3), Ca2ϩ (2.5 × 10Ϫ5 mol
dmϪ3), Cu2ϩ (1.8 × 10Ϫ5 mol dmϪ3), Zn2ϩ (1.5 × 10Ϫ5 mol dmϪ3),
phosphate (3.8 × 10Ϫ4 mol dmϪ3) and carbonate (2.5 × 10Ϫ5
mol dmϪ3) in blood plasma concentrations were calculated as a
function of pH by means of the computer program HYSS,23
including in the calculations the protonation constants of the
ligands, the stability constants of the Gd3ϩ, Naϩ, Ca2ϩ, Cu2ϩ
,
At least three measurements (about 100 data points each one)
were performed for each system in the pH ranges 2.0–12 for
ligand protonation and 2.5–10.5 for complexation experiments.
The ligand concentration [L] was about 1 × 10Ϫ3 mol dmϪ3 in
all measurements while the metal ion concentration in the
complexation experiments was 4[L] for Naϩ and 0.8[L]–2[L] for
and Zn2ϩ complexes, the solubility products of gadolinium()
phosphate, carbonate and hydroxide, and the protonation con-
stants of phosphate and carbonate, all determined at 298.1 K.
Results and discussion
Description of the structure
Ca2ϩ, Zn2ϩ and Cu2ϩ
.
Complexation reactions of Gd3ϩ with the present ligands are
slow; about four weeks are necessary to achieve the final equi-
librium. For this reason out-of-cell experiments were performed
in which 30 individual solutions (40 for DTMA), in the
approximate pH range 2.5–6 (2.5–9 for DTMA), corresponding
to single points of conventional titrations were stored in a
thermostat at 298.1 0.1 K, and their pH was periodically con-
trolled to ensure the achievement of equilibrium conditions.
For all ligands, DTMA excluded, the use of EDTA as an
auxiliary ligand was necessary due to the very high value of the
equilibrium constants to be determined. The variation of [Hϩ]
with time clearly evidences a two step complexation reaction, in
which an intermediate complex is formed in the first 25–50
minutes after mixing of ligand and metal ion, which develops
towards the final, more stable complexes in a few weeks. The
corresponding emf data were used to determine the equi-
librium constants corresponding to both intermediate and final
complex species. In the complexation experiments ligands
(including EDTA) and metal ion concentrations were 1 × 10Ϫ3
Crystals of [Gd(DTMA)(H2O)][ClO4]3ؒNaClO4ؒ3H2O
are
made of [Gd(DTMA)(H2O)]3ϩ cations, sodium ions, per-
chlorate counter ions and crystallization water molecules. The
macrocyclic ligand acts as octadentate. The metal ion is co-
ordinated by the four nitrogens of the tetraazamacrocyclic
moiety and by the amidic oxygen atoms provided by the side
arms. The ninth position of the co-ordination sphere is occu-
pied by an oxygen atom of a water molecule (Fig. 1). The Gd–N
and the Gd–O amidic bond distances spanning in the ranges
2.621(5)–2.649(6) and 2.351(5)–2.455(5) Å, respectively, as well
as the Gd–O (water) distance (2.461(5) Å) (Table 1) are compar-
able with those previously reported for analogous Gd3ϩ com-
plexes.25 The resulting polyhedron, which can be described as
a distorted square antiprism,26 capped by the water oxygen,
corresponds to the usual co-ordination geometry of Gd3ϩ
complexes with ligands based on the 1,4,7,10-tetraazacyclo-
dodecane structure.25 Table 2 reports the dihedral angle values
calculated for [Gd(DTMA)(H2O)]3ϩ together with those
corresponding to an ideal (C4v) monocapped square antiprism
co-ordination environment. The four nitrogens and the four
oxygen donor atoms, respectively, define two almost parallel
mean planes (1.2(1)Њ). The oxygen atom of the Gd3ϩ-co-
ordinated water molecule (O5) is ca. 1.7 Å out of the O1–O2–
O3–O4 mean plane, on the opposite side of the metal ion, the
Gd–O5 bond being almost perpendicular to this plane (87Њ). As
already found,5e,25b–d the metal ion is closer to the oxygen
donors’ mean plane than to the plane described by the four
nitrogens (0.74 vs. 1.59 Å).
mol dmϪ3
.
The computer program HYPERQUAD20 was used to calcu-
late the equilibrium constants from emf data.
Microcalorimetric measurements
The enthalpy changes for ligand protonation and metal ion
complexation were determined in 0.1 mol dmϪ3 NMe4Cl solu-
tions by means of microcalorimetric titrations performed with
an automated system composed of a Thermometric AB ther-
mal activity monitor (model 2277) equipped with a perfusion-
titration device and a Hamilton Pump (model Microlab M)
coupled with a 0.250 cm3 gas-tight Hamilton syringe (model
1750 LT). The measuring vessel was housed in a 25 dm3 water
The tetraaza ring has a [3333] C corners conformation,27 with
the four amidic chains in a head-to-tail relative arrangement
(Fig. 1, lower). The four amidic nitrogen atoms have a signifi-
cant sp2 character, indicating that the usual π conjugation takes
J. Chem. Soc., Dalton Trans., 2000, 697–705
699

View Article Online
Table 1 Primary co-ordination sphere bond distances (Å) and angles
(Њ) for [Gd(DTMA)(H2O)]3ϩ
Table 2 Dihedral angles (Њ) for [Gd(DTMA)(H2O)]3ϩ
Idealized angle
for a C4v
geometry
Gd(1)–O(2)
Gd(1)–O(3)
Gd(1)–O(4)
Gd(1)–O(1)
Gd(1)–O(5)
2.351(5)
2.351(4)
2.393(5)
2.455(5)
2.461(6)
Gd(1)–N(4)
Gd(1)–N(1)
Gd(1)–N(3)
Gd(1)–N(2)
2.621(6)
2.626(5)
2.648(5)
2.649(5)
Face 1
Face 2
Angle/Њ
N3–O2–O3
N4–O3–O4
N3–N4–O3
N2–N3–O2
N1–N3–N4
N1–O1–O4
N2–O1–O2
N1–N2–O1
N1–N4–O4
N1–N2–N3
155.2(2)
157.6(2)
142.1(2)
145.1
163.5
163.5
138.2
138.2
0.0
O(2)–Gd(1)–O(3)
O(2)–Gd(1)–O(4)
O(3)–Gd(1)–O(4)
O(2)–Gd(1)–O(1)
O(3)–Gd(1)–O(1)
O(4)–Gd(1)–O(1)
O(2)–Gd(1)–O(5)
O(3)–Gd(1)–O(5)
O(4)–Gd(1)–O(5)
O(1)–Gd(1)–O(5)
O(2)–Gd(1)–N(4)
O(3)–Gd(1)–N(4)
O(4)–Gd(1)–N(4)
O(1)–Gd(1)–N(4)
O(5)–Gd(1)–N(4)
O(2)–Gd(1)–N(1)
O(3)–Gd(1)–N(1)
O(4)–Gd(1)–N(1)
O(1)–Gd(1)–N(1)
84.1(2)
142.3(2)
83.7(2)
81.8(2)
145.1(2)
88.3(2)
73.4(2)
74.1(2)
69.0(2)
71.3(2)
142.5(2)
74.0(2)
65.7(2)
132.3(2)
126.5(2)
131.1(2)
142.4(2)
74.5(2)
65.4(2)
O(5)–Gd(1)–N(1)
N(4)–Gd(1)–N(1)
O(2)–Gd(1)–N(3)
O(3)–Gd(1)–N(3)
O(4)–Gd(1)–N(3)
O(1)–Gd(1)–N(3)
O(5)–Gd(1)–N(3)
N(4)–Gd(1)–N(3)
123.1(2)
69.2(2)
74.9(2)
66.1(2)
130.4(2)
138.3(2)
130.9(2)
68.5(2)
0.3(2)
ion occurs. In addition the water oxygen O5 is 3.483(7) Å from a
symmetry related O5 atom.
Finally a lot of hydrogen-bond contacts (<3.0 Å) involving
the crystallization water molecules, the oxygen atoms of the
perchlorate anions and various hydrogen and nitrogen atoms of
the ligand are present.
N(1)–Gd(1)–N(3) 106.0(2)
O(2)–Gd(1)–N(2)
O(3)–Gd(1)–N(2)
O(4)–Gd(1)–N(2)
O(1)–Gd(1)–N(2)
O(5)–Gd(1)–N(2)
N(4)–Gd(1)–N(2) 105.6(2)
N(1)–Gd(1)–N(2)
N(3)–Gd(1)–N(2)
66.7(2)
130.7(2)
142.6(2)
70.9(2)
The coprecipitated Naϩ ion is six-co-ordinated by four
perchlorate oxygens and two crystallization water molecules,
the Na–O bond lengths ranging from 2.413(9) to 2.80(1) Å.
127.8(2)
68.6(2)
68.3(2)
Ligand protonation studies in solution
The protonation constants of the six ligands considered in this
study, determined in 0.1 mol dmϪ3 NMe4Cl solutions at
298.1 0.1 K, are listed in Table 3. The log K values obtained
for DOTA, DO3A and HPDO3A follow the same trend
reported by other authors,5d,28,29 the first two protonation steps
taking place above neutral pH and the remaining ones below
neutral pH. In the case of DOTA our log K values are in good
agreement with previously reported data28 obtained under the
same experimental conditions, although a small difference in
the first protonation constant is observed. In the case of DO3A
all our protonation constants, but the first one, well compare
with previous values.29 The literature value of the first proton-
ation constant (log K = 11.59) is considerably lower than that
we found (12.46(1)). Several measurements in very alkaline
solutions (up to pH 12) were performed to verify our value
after accurate calibration of the glass electrode in such media
(see Experimental section). Significant differences were also
found for the protonation constants of HPDO3A with
respect to the literature values (log Ki = 11.96, 9.43, 4.30,
3.26).29 Also in this case, considerable experimental work
was performed to ensure the purity of the ligand employed
and the accuracy of the potentiometric measurements (see
Experimental section).
The compound DTMA displays a largely reduced ability to
bind Hϩ ions with respect to the other ligands, evidencing the
effect of the negatively charged groups on the basicity of such
ligands. Actually, carboxylate groups, in addition to their
intrinsic basicity, can assist protonation of ligand amine groups
via hydrogen bonding.
Finally HPDO3A-3HM and DO3A-3HM are characterized,
in each protonation step, by a moderately lower basicity with
respect to the other negatively charged ligands, which can be
ascribed to some impediment in the stabilization of the
protonated species via intramolecular hydrogen bonding,
determined by the increased ligand rigidity produced by the
bulky pendant arms.
Interesting information about the ligand protonation reac-
tions is furnished by the enthalpy and entropy changes deter-
mined for these proton transfer processes. As shown in Table 3,
protonations of the considered ligands are promoted by both
favorable enthalpic (∆HЊ < 0) and entropic (T∆SЊ > 0) contri-
butions, with the unique exceptions of the entropic changes for
the last protonation step of DTMA and DO3A-3HM. Con-
cerning the enthalpic contributions, it is well known that amine
protonation is a markedly exothermic reaction, while proton-
ation of acetate groups is an almost athermic reaction.21
Accordingly, considering the enthalpy changes listed in Table 3,
Fig. 1 ORTEP24 drawings of the [Gd(DTMA)(H2O)]3ϩ cation.
place in each amidic functional group thus enhancing the donor
capability of the carbonylic oxygens.
Concerning the packing, it is worth noting that the oxygen of
the Gd3ϩ-co-ordinated water molecule is very close (2.92(1) Å)
to an oxygen atom of a perchlorate ion, suggesting that a strong
hydrogen bond between [Gd(DTMA)(H2O)]3ϩ and a counter
700
J. Chem. Soc., Dalton Trans., 2000, 697–705

View Article Online
Table 3 Thermodynamic parameters for ligand protonation reactions, determined in 0.1 mol dmϪ3 NMe4Cl solutions at 298.1 0.1 K. ∆HЊ and
T∆SЊ in kcal mol Ϫ1
DOTA
DO3A
HPDO3A
Reaction
log K
Ϫ∆HЊ
T∆SЊ
log K
Ϫ∆HЊ
T∆SЊ
log K
Ϫ∆HЊ
T∆SЊ
H ϩ L
HLa
H2L
H3L
H4L
H5L
11.74(4)b
9.76(4)
4.68(4)
4.11(4)
2.37(5)
11.9(2)
9.5(2)
2.5(2)
3.0(2)
1.7(2)
4.1(2)
3.8(2)
3.9(2)
2.6(2)
1.5(2)
12.46(1)
9.49(1)
4.26(1)
3.51(1)
1.97(1)
8.1(1)
10.0(2)
1.8(2)
8.5(1)
2.9(2)
3.9(2)
11.17(1)
9.33(1)
4.99(1)
3.80(1)
2.84(1)
11.8(1)
9.8(1)
4.9(2)
4.1(2)
1.6(2)
3.4(1)
2.9(1)
1.9(2)
1.1(2)
2.3(2)
H ϩ HL
H ϩ H2L
H ϩ H3L
H ϩ H4L
2.7(2)
2.1(2)
c
c
DTMA
HPDO3A-3HM
DO3A-3HM
Reaction
log K
Ϫ∆HЊ
T∆SЊ
log K
Ϫ∆HЊ
T∆SЊ
log K
Ϫ∆HЊ
T∆SЊ
H ϩ L
HL
9.27(1)
5.55(2)
1.56(7)
11.04(4)
6.78(4)
4.0(3)
1.60(4)
0.79(4)
Ϫ1.9(3)
10.68(2)
7.81(3)
4.14(4)
3.37(5)
5.1(1)
5.8(1)
3.2(1)
1.9(1)
9.5(1)
4.8(1)
2.4(1)
2.7(1)
10.93(1)
7.08(2)
4.04(2)
3.49(2)
9.5(2)
5.1(2)
2.2(3)
5.1(3)
5.4(2)
4.5(2)
3.3(3)
H ϩ HL
H ϩ H2L
H ϩ H3L
H2L
H3L
H4L
Ϫ0.3(3)
a Charges omitted. b Values in parentheses are standard deviations in the last significant figure. c Not determined.
it is clear that the first two protonations occur at the amine
Gd3؉ Complexation in solution
groups of the ligands, although the acetate residues are capable
of contributing to the stability of such protonated species by
hydrogen bonding and charge–charge interaction. In the case of
DTMA also the third protonation takes place necessarily on an
amine group in accordance with the exothermic character of
the reaction. Conversely, for the other ligands most of the fol-
lowing protonation reactions are considerably less exothermic
indicating a principal involvement of carboxylate groups in the
last protonation steps. Such a protonation sequence (the first
two protons binding nitrogen atoms and the successive ones
binding carboxylate groups) was clearly evidenced for DOTA
by means of NMR studies in solution.30 Accordingly, the
crystal structure of DO3AؒH2SO4 displayed only two out of the
five acidic protons bound to nitrogen atoms, the remaining
three Hϩ ions being linked to carboxylate groups.29 The struc-
ture also revealed that two transannular nitrogen atoms are
involved in ligand protonation one of which is the secondary
one.
Only in the case of HPDO3A the third and the fourth pro-
tonations are rather exothermic suggesting that for this ligand
a significant participation of amine groups in the binding of Hϩ
ions occurs up to the fourth protonation step. This behavior is
rather surprising for a polyaminopolycarboxylic ligand based
upon 1,4,7,10-tetraazacyclododecane, resembling the features
of analogous ligands containing larger tetraazamacrocyclic
moieties.30b Nevertheless, also the fourth protonation step of
DO3A-3HM is significantly more exothermic than expected on
the basis of the general trend of decreasing enthalpic contribu-
tions with increasing number of protons on the ligands, in
agreement with a third proton binding to the tetraazamacro-
cyclic moiety at the fourth protonation stage.
The favorable entropic contributions to the protonation of
these ligands are mainly determined by desolvation processes
due to the charge neutralization occurring upon proton binding
by the ligand. On the other hand the increasing stiffening of
these cyclic molecules, occurring upon successive ligand proton-
ation, determined by the presence of charged ammonium
groups in close proximity to each other and by the formation
of intramolecular hydrogen bonds involving the ammonium
(amine) groups and the carboxylate (carboxylic) residues,
results in a loss of entropy which is reflected by the lower
entropic contributions observed in the last protonation steps. In
this sense, DTMA represents a particular case since it does not
contain charged functionalities. In such a case desolvation
processes are less important, and the entropic contributions to
DTMA protonation are very small, becoming unfavorable
(T∆SЊ < 0) for the last protonation step.
The stability constants of the Gd3ϩ complexes with DTMA,
HPDO3A-3HM and DO3A-3HM are reported in Table 4
together with those for DOTA, DO3A and HPDO3A we have
redetermined by means of our method, under our experimental
conditions. Very good agreement is found with a literature value
(log K = 24.6)5a of the stability constant of [Gd(DOTA)]Ϫ
(other values obtained at 25 ЊC and 0.1 mol dmϪ3 ionic strength
are log K = 24.0,5d 25.330a) while moderately higher values for
the complexes Gd(DO3A) and Gd(HPDO3A) (log K =
22.02(2), 24.5(1); Table 4) were found with respect to the
literature values (log K = 21.0, 23.8).28
As can be seen from Table 4, the stability of the Gd3ϩ
complexes is strongly affected by the nature of the pendant
functionalities linked to the common tetraazamacrocyclic ring.
The complex of lower stability is formed by DTMA, the lig-
and bearing four uncharged amidic groups. According to the
structure of [Gd(DTMA)(H2O)]3ϩ observed in the solid state
(Fig. 1), it seems likely that also in solution the metal ion is co-
ordinated by the four nitrogen atoms of the macrocyclic moiety
and by the four amidic oxygen atoms provided by the side arms,
the usual nine-co-ordination being completed by a solvent
water molecule. Starting from neutral pH this metal bound
water molecule dissociates (pKa = 7.9(1)) giving rise to the
hydroxo complex [Gd(DTMA)(OH)]2ϩ, while in acidic solu-
tions (pH < 5) [Gd(DTMA)]3ϩ undergoes protonation pro-
ducing the species [Gd(HDTMA)]4ϩ. Owing to the modest
co-ordinating ability of DTMA toward Gd3ϩ, unco-ordinated
metal ion is present in acidic media up to pH 4.
Considerably more stable complexes are formed by the lig-
ands bearing charged pendant functionalities. Among these,
the fully deprotonated forms of DOTA and HPDO3A form the
most stable complexes, the stability constants of [Gd(DOTA)]Ϫ
and Gd(HPDO3A) being equal within the experimental errors,
in spite of the lower negative charge of HPDO3A. On the other
hand, Gd(DO3A) displays a significantly lower stability than
that of Gd(HPDO3A) (Table 4), although both ligands (DO3A
and HPDO3A) have the same negative charge in the com-
plexes. Evidently, the presence of three negative charges instead
of four on the ligand does not influence the stability of the
Gd3ϩ complex, while the loss of a donor atom has a consid-
erable effect in lowering the complex stability. Actually, both
DOTA and HPDO3A bind Gd3ϩ producing a N4O4 co-
ordination environment, similar to that displayed by DTMA
(Fig. 1), while a N4O3 set of donors is furnished by DO3A, as
demonstrated in the solid state by crystal structures of the
complexes.25c,d,28
J. Chem. Soc., Dalton Trans., 2000, 697–705
701

View Article Online
Table 4 Stability constants of Gd3ϩ complexes determined in 0.1 mol dmϪ3 NMe4Cl solutions at 298.1 0.1 K
log K
Reaction
DOTA
DO3A
HPDO3A
24.5(1)
DTMA
HPDO3A-3HM
19.4(1)
DO3A-3HM
Gd ϩ L
GdL ϩ H
GdL ϩ OH
GdLa
Gd(HL)
GdL(OH)
24.67(7)b
22.02(2)
12.8(1)
3.4(1)
5.8(1)
18.82(9)
2.63(1)
a Charges omitted. b Values in parentheses are standard deviations in the last significant figure.
Fig. 2 Percentages of Gd3ϩ bound to DOTA and to HPDO3A
calculated as a function of pH for equimolar concentrations (1 × 10Ϫ3
mol dmϪ3) of reactants.
Fig. 3 Percentages of Gd3ϩ bound to DTMA and to DO3A-3HM
calculated as a function of pH for equimolar concentrations (1 × 10Ϫ3
mol dmϪ3) of reactants.
The compound HPDO3A-3HM is ideally obtained by insert-
ing three CH2OH groups on the side arms of HPDO3A bearing
the carboxylate residues. Hence, with respect to HPDO3A,
HPDO3A-3HM contains three additional oxygen atoms which
might act as donor atoms toward Gd3ϩ, enhancing the complex
stability. Conversely, the stability of the Gd(HPDO3A-3HM)
(log K = 19.4) complex is considerably lower that that displayed
by Gd(HPDO3A) (log K = 24.5). This is due, most likely, to a
greater steric hindrance of the hydroxypropanoic function-
alities with respect to the acetic ones, which determines a more
strained, less stable, co-ordination environment of the metal
ion. On the other hand, HPDO3A-3HM contains a surplus
of donor atoms and elimination of the hydroxypropyl group,
to give DO3A-3HM, does not result in a significant loss of
complex stability (log K = 18.82 for Gd(DO3A-3HM)), as
otherwise observed upon eliminating the hydroxypropyl group
of HPDO3A to give DO3A.
It is interesting that in spite of the higher stability constants
of the Gd3ϩ complex displayed by DO3A, HPDO3A-3HM and
DO3A-3HM demonstrate a greater ability in the binding of the
metal ion in the pH range considered (2–8). This is due to the
higher basicity of DO3A, which determines a stronger competi-
tion between metal ion complexation and ligand protonation.
On the other hand, predominant binding of Gd3ϩ by DO3A is
expected in very alkaline solutions where ligand protonation
does not compete with Gd3ϩ complexation.
Demetallation of Gd3؉ complexes by endogenous agents
As mentioned in the introduction, the toxicity of Gd3ϩ com-
plexes is principally determined by the release of the highly
toxic Gd3ϩ ion. In order to forecast the fate of such compounds
in the case of in vivo administration, some computer models of
blood plasma have been developed in an attempt to simulate all
the equilibria that are expected to take place in this medium. To
define a good model of blood plasma for a gadolinium com-
plex, all species contained in blood plasma which are known to
bind Gd3ϩ, to form different complexes with the examined lig-
and, and produce insoluble compounds with the components
of the complex, or the complex itself, should be considered.
Different models considering as many equilibria as possible32,33
or adopting simplified sets of equilibria34 have been proposed.
Nevertheless, the potential applicability of such simulation
efforts is presently limited by two main factors: (i) the absence
of complete sets of equilibrium data consistent with the physio-
logical conditions; (ii) the influence of the physiological regu-
latory systems that oppose, in vivo, the equilibrium shifts
predicted by calculation based upon a static model. In addition
such thermodynamic models do not consider the kinetic aspects
connected with the demetallation and/or transmetallation reac-
tions. From this point of view Gd3ϩ complexes with functional-
ized macrocyclic ligands are favorably sluggish, and predictions
based on such thermodynamic models are generally pessimistic
considering the time required for complex excretion.
As far as the selectivity of the present ligands in the binding
of Gd3ϩ in aqueous solution is considered, it is interesting that
the simple analysis of the complex stability constants does not
lead to correct interpretations. Recently, it has been shown that
a useful method to determine thermodynamic selectivity trends
in the co-ordination properties of couples of ligands consists in
the calculation of the percentages of metal bound to each lig-
and, as a function of pH, for solutions containing the ligands
and the metal ion in equimolar amounts.31 A similar calcu-
lation, performed for the systems Gd3ϩ/DOTA/HPDO3A (all
components 1 × 10Ϫ3 mol dmϪ3) by using the complex stability
constants and the ligand protonation constants reported in
Tables 4 and 3, respectively, furnishes the diagram in Fig. 2,
evidencing a greater propensity of HPDO3A to bind Gd3ϩ. By
performing similar calculations for all possible couples of
ligands we obtain the trend of binding selectivity HPDO3A >
DOTA > HPDO3A-3HM > DO3A-3HM > DO3A > DTMA
over all the pH range considered (2–8). Only for the Gd3ϩ
/
DTMA/DO3A-3HM system (Fig. 3) inversion of selectivity
between the two ligands is observed in acidic solution: while for
pH > 3 Gd3ϩ is preferentially bound to DO3A-3HM, at lower
pH the DTMA complexes prevail due to the formation of the
monoprotonated [Gd(HDTMA)]4ϩ species.
In the present work, in order to evaluate the amounts of
Gd3ϩ released, at equilibrium, by the considered complexes in
solutions containing important components of blood plasma,
702
J. Chem. Soc., Dalton Trans., 2000, 697–705

View Article Online
Table 5 Stability constants of Naϩ, Ca2ϩ, Zn2ϩ and Cu2ϩ complexes determined in 0.1 mol dmϪ3 NMe4Cl solutions at 298.1 0.1 K
log K
Reaction
DOTA
DO3A
HPDO3A
DTMA
HPDO3A-3HM
DO3A-3HM
Na ϩ L
Ca ϩ L
CaL ϩ H
Ca(HL) ϩ H
CaL ϩ 2H
Ca(H2L) ϩ H
CaL ϩ OH
CaL ϩ Ca
Zn ϩ L
NaLa
CaL
Ca(HL)
Ca(H2L)
Ca(H2L)
Ca(H3L)
CaL(OH)
Ca2L
ZnL
Zn(HL)
Zn(H2L)
4.03(4)b
17.219(8)
3.80(4)
3.77(9)
7.57(9)
2.20(2)
13.39(2)
3.19(3)
14.06(1)
5.38(2)
2.50(1)
11.09(4)
2.58(3)
8.84(2)
5.80(7)
5.54(7)
11.34(7)
9.47(3)
3.9(1)
11.36(5)
3.8(1)
2.9(8)
5.94(7)
3.70(5)
17.16(2)
4.23(2)
20.52(1)
4.35(1)
3.49(1)
7.84(1)
21.87(3)
3.27(2)
2.81(6)
6.08(6)
2.39(7)
20.41(8)
4.92(7)
3.13(8)
8.05(8)
13.00(3)
17.20(1)
3.51(1)
ZnL ϩ H
Zn(HL) ϩ H
ZnL ϩ 2H
ZnL ϩ OH
ZnL(OH) ϩ OH
Zn(H2L)
ZnL(OH)
3.75(5)
3.71(4)
3.1(1)
8.44(6)
ZnL(OH)2
Cu ϩ L
CuL ϩ H
Cu(HL) ϩ H
CuL ϩ 2H
CuL ϩ OH
CuL(OH) ϩ OH
CuL
Cu(HL)
Cu(H2L)
22.44(9)
4.31(9)
3.52(9)
7.83(9)
21.65(2)
3.86(2)
24.19(3)
4.61(4)
14.12(3)
19.32(3)
4.11(3)
19.31(5)
3.49(5)
2.23(8)
Cu(H2L)
CuL(OH)
5.46(4)
3.32(5)
8.18(8)
CuL(OH)2
a Charges omitted. b Values in parentheses are standard deviations in the last significant figure.
such as Naϩ, Ca2ϩ, Cu2ϩ, Zn2ϩ, phosphate and carbonate, the
concentrations of all species formed in the presence of Gd3ϩ
complexes at a plasma concentration of clinical administrations
(2.5 × 10Ϫ5 mol dmϪ3) were calculated as a function of pH
by means of the computer program HYSS.23 To perform
such calculations, we have determined the stability constants
of the Naϩ, Ca2ϩ, Cu2ϩ, and Zn2ϩ complexes with DTMA,
HPDO3A-3HM and DO3A-3HM, and redetermined under
our experimental conditions the stability constants of the com-
plexes formed by DOTA, DO3A, and HPDO3A with the same
metal ions (Table 5).
ligands should be in monoprotonated form and the metal ion is
expected to be co-ordinated by the carboxylate groups and at
least one nitrogen atom of the ligand, followed by its slow
reorganization.5f,35
In a recent communication36 some of us reported that by
means of potentiometric (pH-metric) measurements it is
possible to evidence different stages in these complexation
reactions of Gd3ϩ with DOTA and DO3A. Now we have
extended this study to HPDO3A, DTMA, HPDO3A-3HM and
DO3A-3HM. The additions of Gd3ϩ to solutions containing
protonated forms of these ligands produces a rapid jump of
Hϩ concentration due to the release of protons by the ligand
during metal ion complexation. After this almost instantan-
eous process the solution pH evolves in about 25–50 minutes
towards a seemingly constant value, although three to four weeks
are necessary to attain the final equilibrium corresponding to
the complex stability constants in Table 4. Potentiometric data
obtained for the 25–50 minutes equilibration evidenced that, at
this stage, both monoprotonated and unprotonated intermedi-
ate complexes are present in solution for all ligands, but for
HPDO3A and DO3A-3HM which do not form monoproto-
nated complexes, while also diprotonated species are formed
with HPDO3A-3HM and DO3A-3HM, allowing the determin-
ation of the relevant stability constants listed in Table 6. In
order to get insight into the nature of the metal to ligand inter-
action at this intermediate stage, we have also measured the
enthalpy changes associated to the formation of these com-
plexes (Table 6).
According to the calculations, the percentages of Gd3ϩ com-
plex which are retained at physiological pH 7.4 are 93.9% for
HPDO3A, 93.8% for DOTA, 92.4% for HPDO3A-3HM,
89.1% for DO3A-3HM, 87.8% for DO3A and 84.8% for
DTMA. The released Gd3ϩ precipitates in the form of GdPO4,
while gadolinium carbonate and hydroxide are not formed. On
the other hand the released ligand is almost totally involved in
the co-ordination to Cu2ϩ, Zn2ϩ and Ca2ϩ
.
It is to be stressed here that such calculations do not pretend
to produce a biospeciation in presence of similar Gd3ϩ com-
plexes but, in spite of the use of equilibrium data obtained at
298.1 K and the ignored equilibria involving further endogen-
ous agents, such as amino acids, citrate and human serum
albumin, they furnish a reasonable estimation of the thermo-
dynamic resistance of these metal complexes towards demetal-
lation processes promoted by principal components of blood
plasma. In addition such calculations can be performed by
means of a readily available computer program (HYSS).23
It is interesting that the calculated percentages of undissoci-
ated Gd3ϩ complexes are consistent with the trend of selectivity,
displayed by these ligands in the binding of Gd3ϩ at physio-
logical pH, discussed in the previous section. This means that
the thermodynamic properties of these complex systems (com-
plex stability and ligand basicity) are of major importance in
determining the applicability of such complexes as contrast
agents.
As far as the unprotonated complexes are considered, the
results obtained indicate that, with the exception of DTMA,
these intermediate species are very stable, the difference in
stability between the intermediate and the final ones being
rather small. In energetic terms, 80–90% of the free energy gain
(∆GЊ = ϪRT log K) associated with the attainment of the final
complexation stages (93% in the case of HPDO3A) is produced
in the first minutes of these long processes lasting over three
weeks. A similar behavior suggests that a good co-operativity
between carboxylate and amine groups in metal ion chelation
occurs also in the formation of the intermediate complexes.
Nevertheless, important differences in the complexation
modes are revealed by the enthalpic contributions to com-
plexation. Actually the formation of such Gd3ϩ intermediate
complexes is promoted by favorable enthalpic contributions
Formation of intermediate species in complexation reactions
It has been shown by kinetic studies that complexation of Gd3ϩ
by polyaminopolycarboxylic macrocyclic ligands proceeds
through the formation of intermediate complexes, in which the
J. Chem. Soc., Dalton Trans., 2000, 697–705
703

View Article Online
Table 6 Thermodynamic parameters for intermediate equilibria of Gd3ϩ complexation determined in 0.1 mol dmϪ3 NMe4Cl solutions at
298.1 0.1 K
DOTA
DO3A
HPDO3A
Reaction
log K*
Ϫ∆HЊ*
T∆SЊ*
log K*
Ϫ∆HЊ*
T∆SЊ*
log K*
Ϫ∆HЊ*
T∆SЊ*
Gd ϩ L
Gd ϩ L ϩ H
GdL ϩ H
GdLa
Gd(HL)
Gd(HL)
21.2(1)b
24.5(1)
3.3(1)
6.8(2)
10.2(1)
3.4(3)
22(1)
23.2(4)
1.0(4)
19.6(1)
22.5(1)
2.9(2)
4.64(8)
11.24(8)
6.6(1)
22.1(2)
19.5(3)
Ϫ2.6(4)
22.9(1)
11.2(3)
20.0(3)
DTMA
HPDO3A-3HM
DO3A-3HM
Gd ϩ L
GdL
8.6(1)
13.2(1)
10.1(2)
2.9(3)
1.6(2)
15.1(3)
15.2(1)
19.8(1)
22.8(1)
4.6(1)
2.9(1)
7.6(1)
Ϫ3.0(1)
23.7(1)
23.9(1)
22.5(2)
0.2(1)
Ϫ1.4(2)
Ϫ1.2(2)
15.5(1)
22.1(1)
Ϫ0.1(1)
21.2(1)
Gd ϩ L ϩ H
Gd ϩ L ϩ 2H
GdL ϩ H
Gd(HL) ϩ H
GdL ϩ 2H
Gd(HL)
Gd(H2L)
3.1(1)
8.6(2)
6.1(1)
5.5(2)
11.6(2)
10.2(1)
19.9(1)
Gd(HL)
Gd(H2L)
Gd(H2L)
4.6(1)
Ϫ7.2(3)
13.5(3)
6.6(1)
10.3(1)
Ϫ1.3(1)
a Charges omitted. b Values in parentheses are standard deviations in the last significant figure.
(∆HЊ < 0) in the case of DOTA, DO3A, HPDO3A and
DTMA, an unfavorable contribution (∆HЊ > 0) is found for
HPDO3A-3HM and no enthalpic contribution is observed
with DO3A-3HM.
this case protonation occurs at free amine nitrogens. Hence no
more than one amine nitrogen seems to be unco-ordinated in
these intermediate complexes.
Presumably, the ammonium groups of these protonated
species adopt an exo conformation in order to minimize the
electrostatic repulsion with the tricharged cation. Deproton-
ation of these groups and subsequent interconversion to an
endo conformation is, most likely, the process determining the
final, very slow equilibration step.
Similarly to the case of ligand protonation, also metal ion
complexation involves the formation of binding interactions
(co-ordinative bonds) and desolvation of the reacting species.
From an enthalpic point of view these processes furnish oppos-
ite contributions, the first one being favorable and second
unfavorable. The extent of desolvation primarily depends on
the neutralization of charge occurring upon interaction of the
metal ion with the donor atoms, and hence the negatively
charged acetate groups result in a larger release of solvent
molecules than the uncharged amine groups. On the other
hand, a large desolvation corresponds to a very favorable
entropic contribution. This is the reason why metal ion co-
ordination by carboxylates is promoted by the entropic term,
the enthalpic one being unfavorable or almost negligible, while
amine groups give rise to complexation reactions with favorable
enthalpy changes.21
Based on these considerations, the thermodynamic param-
eters reported in Table 4 are strongly indicative of the fact that
in the intermediate Gd3ϩ complexes with HPDO3A-3HM and
DO3A-3HM the ligand carboxylate groups give the principal
contribution to complex stability, while for the remaining lig-
ands co-ordination of amine nitrogens is more important. As
already said, the intermediate complexes of HPDO3A-3HM
and DO3A-3HM can bind up to two protons. The enthalpy
changes obtained for these protonation processes (Table 6) are
typical of unco-ordinated amine groups, indicating that at least
two nitrogen atoms of the macrocyclic backbone are not
involved in the co-ordination. Most likely, at this preliminary
complexation stage, HPDO3A-3HM and DO3A-3HM hold the
metal ion in the region of the pendant functionalities, thanks to
the presence of a large number of donor atoms, and probably
only one amine group takes part in complexation. Considering
the protonation properties of the twelve-membered tetra-
azamacrocyclic moiety it seems improbable that a greater
number of amine groups can participate.
Conclusion
The information obtained in this study for the complexation in
vitro of Gd3ϩ with ligands based upon the macrocyclic structure
of 1,4,7,10-tetraazacyclododecane, and containing pendant
functionalities, evidences several aspects which may be useful
for the preparation of gadolinium-containing MRI contrast
agents. (i) More stable complexes are formed by the ligands
bearing negatively charged functionalities, although the com-
plex stability is not strictly connected with the overall ligand
charge. (ii) The co-ordination selectivity of these ligands
towards Gd3ϩ does not follow the trend of complex stability,
principally due to the different ligand basicities, and inversion
of selectivity between ligands may occur upon changing the
solution pH. (iii) Conversely, the thermodynamic resistance of
the studied Gd3ϩ complexes towards demetallation processes
promoted by the principal components of blood plasma, such
as Naϩ, Ca2ϩ, Cu2ϩ, Zn2ϩ, phosphate and carbonate, at physio-
logical pH 7.4, is consistent with the trend of complex stability.
(iv) Complexation of Gd3ϩ by these ligands takes place through
a multi-step process lasting several weeks, but most of the free
energy gain associated with the attainment of the final, more
stable, complex species is produced in the first minutes of the
process.
Acknowledgements
We thank the Italian Ministero dell’Universtià e della Ricerca
Scientifica e Tecnologica (MURST) for financial support
(COFIN98).
Conversely, the intermediate Gd3ϩ complexes with the
remaining ligands are expected to have a structure, more similar
to that of the final species. These complexes display a consider-
ably lower tendency to bind protons in solution with respect to
the analogous species formed by HPDO3A-3HM and DO3A-
3HM (Table 6). The compound HPDO3A does not undergo
protonation in the pH range considered (pH > 2.5), while for
DOTA and DO3A only monoprotonated complexes are formed
in very acidic solutions, and the favorable enthalpy changes
(∆HЊ < 0) for such protonation reactions indicate that also in
References
1 R. B. Lauffer, Chem. Rev., 1987, 87, 901.
2 A. D. Sherry, J. Less-Common Met., 1989, 149, 133.
3 K. Kumar and M. F. Tweedle, Pure Appl. Chem., 1993, 65, 515.
4 J. Petersein, S. Saini and R. Weisslader, MRI Clin. N. Am., 1996, 4,
53.
5 (a) W. P. Cacheris, S. K. Nickle and A. D. Sherry, Inorg. Chem.,
1987, 26, 958; (b) E. Brucher and A. D. Sherry, Inorg. Chem., 1990,
704
J. Chem. Soc., Dalton Trans., 2000, 697–705

View Article Online
29, 1555; (c) D. D. Dischino, E. Delaney, J. E. Emswiler, G. T.
Gaughan, J. S. Prasad, S. K. Srivastava and M. F. Tweedle, Inorg.
Chem., 1991, 30, 1265; (d) E. T. Clarke and A. E. Martell, Inorg.
Chim. Acta, 1991, 190, 37; (e) S. Aime, P. L. Anelli, M. Botta,
F. Fedeli, M. Grandi, P. Paoli and F. Uggeri, Inorg. Chem., 1992, 31,
2422; ( f ) X. Wang, T. Jin, V. Combin, A. Lopez-Mut, E. Merciny
and J. F. Desreux, Inorg. Chem., 1992, 31, 1095; (g) K. Kumar,
T. Jin, X. Wang, J. F. Desreux and M. F. Tweedle, Inorg. Chem.,
1994, 33, 3823; (h) S. Aime, M. Botta, M. Fasano, S. Paoletti,
P. L. Anelli, F. Uggeri and M. Virtuani, Inorg. Chem., 1994, 33,
4707; (i) F. Uggeri, S. Aime, P. L. Anelli, M. Botta, M. Brocchetta,
C. de Haën, G. Ermondi, M. Grandi and P. Paoli, Inorg. Chem.,
1995, 34, 633; (j) S. Aime, M. Botta, D. Parker and J. A. G.
Williams, J. Chem. Soc., Dalton Trans, 1996, 17; (k) S. Aime,
M. Botta, D. Parker and J. A. G. Williams, J. Chem. Soc., Dalton
Trans., 1996, 3613; (l) J. Huskens, D. A. Torres, Z. Kovacs, J. P.
André, C. F. G. C. Geraldes and A. D. Sherry, Inorg. Chem., 1997,
36, 1495.
20 P. Gans, A. Sabatini and A. Vacca, Talanta, 1996, 43, 1739.
21 R. M. Smith and A. E. Martell, NIST Critical Stability Constants
Database, version 2, 1995.
22 A. Vacca, AAAL program, Department of Chemistry, University
of Florence, 1997.
23 L. Alderighi, P. Gans, A. Ienco, D. Peters, A. Sabatini and A. Vacca,
Coord. Chem. Rev., 1999, 184, 311. Program available on Web
(http://www.chim1.unifi.it/group/vacsab/hyss.htm).
24 C. K. Johnson, ORTEP II, Report ORNL-5138, Oak Ridge
National Laboratory, Oak Ridge, TN, 1973.
25 (a) M. S. Konings, W. C. Dow, D. B. Love, K. N. Raymond, S. C.
Quay and S. M. Rocklage, Inorg. Chem., 1990, 29, 1488; (b) S. I.
Kang, R. S. Ranganathan, J. E. Emswiler, K. Kumar, J. Z.
Gougoutas, M. F. Malley and M. F. Tweedle, Inorg. Chem., 1993,
32, 2912; (c) C. A. Chang, L. C. Francesconi, M. F. Malley,
K. Kumar, J. Z. Gougoutas, M. F. Tweedle, D. W. Lee and L. J.
Wilson, Inorg. Chem., 1993, 32, 3501; (d) J.-P. Dubost, J.-M. Leger,
M.-H. Langlois, D. Meyer and M. Schaefer, C.R. Acad. Sci., Ser. II,
1991, 312, 349.
6 S. Meiboon, J. Chem. Phys., 1961, 28, 283; J. A. Peters, J. Huskens
and D. J. Raber, Prog. Nucl. Magn. Reson. Spectrosc., 1996, 28,
283.
26 I. J. Guggenberger and E. L. Muetterties, J. Am. Chem. Soc., 1976,
98, 7221.
7 U. Frey, A. E. Merbach and D. H. Powell, in Dynamics of Solutions
and Fluids Mixtures by NMR, ed. J.-J. Delpuech, John Wiley & Sons
Ltd, Chichester, 1995.
8 G. Gonzàles, H. D. Powell, V. Tissières and A. E. Merbach, J. Phys.
Chem., 1994, 98, 53.
9 S. Aime, A. Barge, M. Botta, D. Parker and A. S. De Sousa, J. Am.
Chem. Soc., 1997, 119, 4767.
10 L. Alderighi, A. Bianchi, L. Calabi, C. Giorgi, P. Losi, P. Paoli,
P. Rossi, B. Valtancoli and M. Virtuani, Eur. J. Inorg. Chem., 1998,
1581.
11 M. F. Tweedle, G. Gaughan and J. J. Hagan, U.S. Pat., 4, 885,363,
1989; J. F. Desreux, Inorg. Chem., 1980, 19, 1319; G. Hernandez,
M. F. Tweedle and R. G. Bryant, Inorg. Chem., 1990, 29, 5109.
12 F. Uggeri, F. Fedeli, A. Maiocchi, M. Franzini and M. Virtuani,
Eur. pat., application no. 0872479, 1998.
13 D. Parker and J. A. G. Williams, J. Chem. Soc., Perkin Trans. 2,
1995, 1305; A. Beeby, D. Parker and J. A. G. Williams, J. Chem.
Soc., Perkin Trans. 2, 1996, 1565; D. Parker and J. A. G. Williams,
J. Chem. Soc., Perkin Trans. 2, 1996, 1581.
27 I. Bernal, Stereochemical and Stereophysical Behaviour of Macro-
cycles, Elsevier, Amsterdam, 1987, p. 87.
28 R. Delgado and J. J. Frausto da Silva, Talanta, 1982, 29, 815.
29 K. Kumar, C. A. Chang, L. C. Francesconi, D. D. Dischino, M. F.
Malley, J. Z. Gougoutas and M. F. Tweedle, Inorg. Chem., 1994, 33,
3567.
30 (a) J. F. Desreux, E. Merciny and M. F. Loncin, Inorg. Chem., 1981,
20, 987; (b) J. R. Ascenso, R. Delgado and J. J. R. Fraústo da Silva,
J. Chem. Soc., Perkin Trans. 2, 1985, 781.
31 A. Andrés, J. Aragó, A. Bencini, A. Bianchi, A. Domenech, V. Fusi,
E. Garcia-España, P. Paoletti and J. A. Ramírez, Inorg. Chem., 1993,
32, 3418; A. Bencini, A. Bianchi, E. Garcia-España, V. Fusi,
M. Micheloni, P. Paoletti, J. A. Ramírez, A. Rodriguez and
B. Valtancoli, J. Chem. Soc., Perkin Trans. 2, 1992, 1059; A. Bianchi
and E. Garcia-España, J. Chem. Educ., 1999, 76, 1727.
32 M. C. Gennaro, S. Aime, E. Santucci, M. Causà and C. De Stefano,
Anal. Chim. Acta, 1990, 233, 85.
33 G. E. Jackson, S. Wynchank and M. Woudenberg, Magn. Reson.
Med., 1990, 16, 57.
14 N. Walker and D. D. Stuart, Acta Crystallogr., Sect. A, 1983, 158.
15 A. Altomare, G. Cascarano, C. Giacovazzo, A. Guagliardi, M. C.
Burla, G. Polidori and M. Camalli, J. Appl. Crystallogr., 1994, 27,
435.
16 G. M. Sheldrick, SHELXL-93, University of Göttingen, 1993.
17 International Tables for X-Ray Crystallography, Kynoch Press,
Birmingham, 1974, vol. IV.
34 W. P. Cacheris, S. C. Quay and S. M. Rocklage, Magn. Reson. Imag.,
1990, 8, 467.
35 K. Kumar, C. A. Chang and M. F. Tweedle, Inorg. Chem., 1993,
32, 587; K. Kumar and M. F. Tweedle, Inorg. Chem., 1993, 32, 4193.
36 A. Bianchi, L. Calabi, L. Ferrini, P. Losi, F. Uggeri and B.
Valtancoli, Inorg. Chim. Acta, 1996, 249, 13.
18 A. Bianchi, L. Bologni, P. Dapporto, M. Micheloni and P. Paoletti,
Inorg. Chem., 1984, 23, 1201.
19 G. Gran, Analyst (London), 1952, 77, 661.
Paper a909098c
J. Chem. Soc., Dalton Trans., 2000, 697–705
705