Relaxometric and solution NMR structural studies on ditopic lanthanide(iii) complexes of a phosphinate analogue of DOTA with a fast rate of water exchange

Article abstract of DOI:10.1039/b518147j

A novel "ditopic" ligand containing two monophosphinate triacetate DOTA-like units linked by a thiourea bridge has been synthesized and its complexes with Ln³⁺ ions (Ln = Y, Eu, Gd, Dy) investigated by NMR spectroscopy and relaxometry. The pres

Full text of DOI:10.1039/b518147j

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
Relaxometric and solution NMR structural studies on ditopic lanthanide(III)
complexes of a phosphinate analogue of DOTA with a fast rate of water
exchange†
Jakub Rudovsky´,a Mauro Botta,*b Petr Hermann,*a Avtandil Koridzec and Silvio Aimed
Received 21st December 2005, Accepted 17th March 2006
First published as an Advance Article on the web 29th March 2006
DOI: 10.1039/b518147j
A novel “ditopic” ligand containing two monophosphinate triacetate DOTA-like units linked by a
thiourea bridge has been synthesized and its complexes with Ln3+ ions (Ln = Y, Eu, Gd, Dy)
investigated by NMR spectroscopy and relaxometry. The presence of one water molecule in the first
coordination sphere has been determined by the measurement of the dysprosium(III)-induced 17O NMR
shifts. The 1H and 31P NMR spectra of the EuIII derivative indicate a higher abundance of the
fast-exchanging twisted square antiprismatic (m) isomer than the isomeric square antiprismatic (M;
m/M = 3 : 2) complex. The analysis of the 89Y and 13C T1 NMR relaxation times in the GdIII/YIII
˚
mixed complex have provided useful structural information. Values of ca. 6.3 and 8.2 A for the Gd · · · Y
and Gd · · · C distances, respectively, have been estimated which indicate a rather compact solution
structure. This result finds support in the value of the relaxivity whose increase (at 20 MHz and 298 K)
on passing from the monomeric (5.7 s−1 mM−1) to the ditopic complex (8.2 s−1 mM−1) can be attributed
to the doubling of the inner-sphere term following the doubling of the molecular size. The structural
and dynamic relaxivity-controlling parameters were assessed by a simultaneous fitting of the variable
temperature 17O NMR and 1H NMRD relaxometric data. The mean water residence lifetime (298sM) has
been found to be 53 ns, one of the shortest values reported for ditopic complexes. The reorientational
correlation time is two times longer (298sR = 183 ps) than the corresponding value of the parent
monomeric GdIII complex, thus supporting the view of a limited degree of internal rotation. The
possible influence of magnetic Gd–Gd coupling has been excluded by a comparison of the 1H NMRD
profiles of the homodinuclear GdIII/GdIII and the heterodinuclear GdIII/YIII complexes.
of paramagnetic contrast agents (CA). These compounds function
Introduction
by catalyzing the proton magnetic relaxation times of nearby
During the last two decades, magnetic resonance imaging (MRI)
water protons in the tissues where they are distributed, thus
has became one of the most powerful diagnostic methods in
increasing the contrast between healthy and pathological regions.
medicine. It has been proved to be particularly suitable for the
imaging of soft tissues with an excellent spatial resolution. The
intrinsic contrast of MRI images originates from local variation
in concentration and nuclear magnetic relaxation times of water
protons. This contrast is further enhanced by the administration
Nowadays, more than 30% of the MRI investigations make use
of CA. The most common class of T1–CA is represented by the
complexes of the highly paramagnetic Gd3+ ion, where the central
ion is encapsulated in an organic ligand leaving just one or two
coordination sites available for water molecule(s). It is the chemical
exchange of the coordinated water molecule(s) with the bulk water
that transfers the paramagnetic relaxation to the surrounding
tissues giving rise to the contrast enhancement.
aDepartment of Inorganic Chemistry, Charles University, Hlavova 2030, 128
40, Prague 2, Czech Republic. E-mail: petrh@natur.cuni.cz; Fax: +420
22195-1253; Tel: +420 22195-1263
The extensive investigations carried out in the last 10–15 years
have provided chemists with a good understanding of the
relationship between the structural properties of the GdIII complex
and the relaxivity (r1). This is a parameter that measures the
efficacy of a metal-based CA and is defined as the increase
in the 1H relaxation rate of water protons in the presence of
a 1 mM concentration of the paramagnetic complex.1–4 The
relaxivity values predicted by theoretical models of paramagnetic
relaxation are much higher (≥100 s−1 mM−1) than those typical
of the commercial CA (ca. 4–5 s−1 mM−1).5 It is also well known
that, in order to attain the high relaxivity values required by the
new, emerging applications of MRI, e.g. in molecular imaging,6–8
several structural and dynamic parameters of the gadolinium(III)
chelates need to be optimized. The most important factors
bDepartment of Environmental and Life Sciences, University of Eastern
Piedmont “A. Avogadro”, Via Bellini 25/G, I-15100, Alessandria, Italy.
E-mail: mauro.botta@mfn.unipmn.it; Fax: +39 0131-360250; Tel: +39
0131-360253
cDepartment of Chemistry, I. Javakhishvili Tbilisi State University, 3
Chavchavadze ave., 380028, Tbilisi, Georgia
dDepartment of Chemistry I. F. M., University of Torino, Via P. Giuria 7,
I-10125, Torino, Italy
† Electronic supplementary information (ESI) available: Fig. S1: GdCl3
relaxometric titration of the title ligand. Fig. S2: DyIII-induced17O chemical
shift measurement. Fig. S3: 1H NMR spectrum of the EuIII–EuIII complex.
Fig. S4: Dependence of 1H relaxivity on pH. Fig. S5: 13C NMR spectra of
the free ligand and its GdIII–YIII complex. Table S1: The complete set of all
possible isomers of the dinuclear complexes. Table S2: 13C relaxation times
of the GdIII–YIII complex. Full set of SBM and SC equations used for
fitting of relaxometric data. See DOI: 10.1039/b518147j
This journal is
The Royal Society of Chemistry 2006
Dalton Trans., 2006, 2323–2333 | 2323
©

View Article Online
are certainly the residence lifetime of the coordinated water
molecule(s), sM, the rotational correlation time of the Gd–Hw
vector, sR, and the electron-spin relaxation parameters. Ideally, a
GdIII complex should reorient in solution at a rate of the order of
≈109 s−1, possess a coordinated water molecule exchanging at a
rate (kex = 1/sM) of ≈108 s−1, and be characterized by electronic
relaxation times as long as possible. Different strategies have
been devised in order to accelerate the slow water exchange rate
of the commercial CA and obtain gadolinium(III) chelates with sM
values in the optimal range.9–15 Recently, we have demonstrated
that the introduction of a phosphorous acid moiety into a
DOTA-like macrocyclic ligand structure (H4DOTA = 1,4,7,10-
tetraazacyclododecane-1,4,7,10-tetracetic acid; Chart 1) induces
a pronounced increase in the rate of water exchange in the
corresponding gadolinium(III) chelate.12–14 Furthermore, the re-
placement of a carboxylate with a phosphorous acid group
favours the formation of hydrogen-bonded, second-sphere water
molecules that contribute to the relaxivity of the GdIII complex,16
and allows for bifunctionality that is essential for further cova-
lent conjugation. These properties have been recently exploited
for the preparation of high relaxivity PAMAM [poly(amido
amine)] conjugates with monophosphinate DOTA-like units,
DO3A-PABn (H4DO3A-PABn = 1,4,7,10-tetraazacyclododecane-
4,7,10 - triacetic - 1 - {methyl[(4 - aminophenyl)methyl)]phosphinic
acid}; Chart 1).17
to have a high thermodynamic and kinetic stability in addition to
exchanging water rapidly.
Results and discussion
Synthesis and characterization
The desired ligand was first isolated as a byproduct from the
conjugation reaction used in the preparation of the PAMAM den-
drimeric conjugates with the corresponding monomeric DO3A-
PABn ligand.17 It is worth noting that a “dimerization” reaction
of this type invariably takes place during the conjugation of
isothiocyanates to any amine (e.g. PAMAM dendrimers) as a
side-reaction. The isothiocyanate group undergoes a hydrolysis
to the parent amine even under mild basic conditions (pH 8–9).
The amine subsequently reacts with the isothiocyanate group of
another molecule to produce the symmetrical dimeric structure
through the thiourea bridge. Bearing this in mind, we set up a
controlled coupling reaction between the amine (DO3A-PABn) and
its corresponding isothiocyanate derivative (DO3A-PNCS; Chart 1)
inorder topreparelarger amounts oftheditopicligandCS(DO3A-
PNBn)2. Both components were mixed in water and the pH was
adjusted to 9–10 with 2 M KOH. The mixture was then stirred at
room temperature and the course of the reaction was monitored by
TLC until the monomeric component disappeared (8 hours). The
crude product was then purified by chromatography on Amberlite
CG50 cation-exchange resin in H+-cycle. The desired CS(DO3A-
PNBn)2 ligand was finally characterized by NMR and ES/MS and
considered as a trihydrate according to the elemental analysis.
Different procedures were used for the preparation of the
The fast molecular reorientation of the commercial CA in
solution is responsible of their low relaxivity values at the imaging
fields. Clearly, every attempt at enhancing the relaxivity needs to
be focused primarily on the increase of sR either by increasing
the size of the complexes or by attaching the complexes to
macromolecular substrates. A relatively simple strategy is the
preparation of dimeric or multimeric metal chelates. However, up
to now, only few examples of this class of GdIII complex have been
reported, based on the DTPA- or DOTA-like basic structures.
The most investigated compounds have been the dinuclear18–22
and multinuclear23 complexes of mostly amide derivatives of
DOTA characterized by a rather slow rate of water exchange.
Only very recently, a few nice examples of multimeric complexes
exhibiting a high relaxivity due to the presence of two coordinated
water molecules have been investigated.24 Unfortunately, these
complexes are rather thermodynamically and kinetically unstable
due to either the relatively low number of donor atoms or the
open-chain structure of the ligand.
lanthanide(III) complexes. The Eu2–CS(DO3A-PNBn
) and Y2–
2
CS(DO3A-PNBn)2 complexes were prepared by mixing an aqueous
solution of the ligand with a ca. 10% molar excess of the
appropriate lanthanide(III) chloride. The pH was then adjusted
to 7 and the cloudy solution was stirred and heated at 60 ◦C
overnight until it became clear. Then, the Chelex 20 chelating
resin in potassium-cycle was added in a batch form to remove
the excess free Ln3+ ions. In the preparation of the corresponding
GdIII complex, the Chelex 20 resin was used in the column form.
The presence of free lanthanide(III) ions was checked by the usual
xylenol orange test.26
The synthesis of the mixed GdIII/YIII complex followed a
completely different scheme. The monomeric Gd–DO3A-PABn
complex, prepared according to the published procedure,14a was
mixed with DO3A-PNCS at pH 9–10 to obtain a mono-GdIII
complex of the ditopic ligand. Then, a 50% molar excess of
yttrium(III) chloride was added to the solution. To the mixture,
stirred at 60–70 ◦C and pH ∼7 overnight, the Chelex 20 was finally
added in a batch form. The excess of salts from the solution of the
GdIII/YIII ditopic complex was removed by ultrafiltration.
In this paper, we present a detailed NMR and relaxometric study
of lanthanide(III) complexes with a new ditopic ligand (Chart 1)
of general formula [Ln2(CS(DO3A-PNBn)2)(H2O)2]2−. The two
monomeric units are connected by a thiourea moiety commonly
used for the conjugation of small ligands to biomolecules.25
Because of the favourable properties of the parent monomeric Gd-
DO3A-PABn complex, the related novel ditopic complex is expected
Chart 1 Chemical structures of the ligands.
This journal is
2324 | Dalton Trans., 2006, 2323–2333
The Royal Society of Chemistry 2006
©

View Article Online
In order to confirm the complete formation of the ditopic ligand
and the formation of the [Ln2(CS(DO3A-PNBn)2)(H2O)2]2− com-
plexes, we focused on the gadolinium derivative and performed a
relaxometric titration of the ligand with GdCl3. To an aqueous
solution of the ligand (5 mM) increasing amounts of gadolin-
ium(III) were added and the water proton longitudinal relaxation
rate (R1) measured, at 400 MHz and 298 K. The plot of R1
versus the concentration of Gd3+ gives a straight line whose slope
corresponds to the relaxivity of the ditopic complex (Fig. S1†).
The relaxivity has a value of 5.7 s−1 mM−1 (per Gd; 400 MHz,
298 K) which clearly indicates the presence of one coordinated
water molecule (q = 1) in the inner coordination sphere of each
metal ion, as expected. The addition of an excess of GdCl3 causes a
steep increase in the value of R1 because of the different relaxivity
of the free Gd3+ (11.0 s−1 mM−1, under identical experimental
conditions). This change is observed at a 2 : 1 Gd/L molar
ratio, thus confirming the complete formation of the ditopic
complex. The value of the hydration number q was independently
assessed by measuring the water 17O NMR chemical shift as
The spectral region of the axial protons of the macrocyclic ring
(ca. 10–40 ppm) is typically used for the determination of the
number of isomeric species.29 These protons, being very close to
the pseudo-C4 magnetic axis of the complex, are the most shifted.
Three sets of four signals are detected in the 1H NMR spectra of
Eu2–CS(DO3A-PNBn)2 between273(Fig.S3†)and298 K(Fig. 1(a)),
a function of the concentration of the Dy2–CS(DO3A-PNBn
)
2
complex in the range 5–70 mM (Fig. S2†). The dysprosium(III)
induced 17O contact shifts of bound water molecules are, to a
very good approximation, independent of structure. Then, these
shifts are directly proportional to the value of q.27 From the slope
(−38.5 ppm M−1 at 298 K) of the straight line obtained by plotting
the 17O shifts as a function of the concentration of Dy3+ ion, we
can safely attribute a single bound water to each metal ion in the
ditopic complex.28
NMR solution studies of the [Eu2(CS(DO3A-PNBn)2)(H2O)2]2−
complex
Further insight into the structural properties of the ditopic
1
complexes in aqueous solution was obtained from the H and
31P NMR spectra of the EuIII complex. This metal ion was chosen
because its NMR spectra are characterized by a good compromise
between a limited paramagnetically induced line broadening and
relatively large isotropic paramagnetic shifts. In addition, the
information gained is likely to be valid also for the corresponding
complexes of gadolinium(III), due to the proximity of the two
elements in the Ln series.
It is well known that the LnIII complexes of DOTA-like ligands
exist in solution as a mixture of coordination isomers differing
in the relative conformation of the macrocyclic ring (kkkk or
dddd) and in the orientation of the four pendant arms (K or
D). One isomer has a mono-capped twisted square antiprismatic
geometry (TSAP; m isomer) and the other a mono-capped square
antiprismatic geometry (SAP; M isomer). Generally, the proton
NMR spectrum of the M isomer covers a significantly broader
range of chemical shifts than the corresponding m isomer.29
Furthermore, the m-type coordination isomer exhibits a much
faster rate of water exchange than the M-type complex.11,30 In
our case, the ligand contains a prochiral phosphorus atom in the
phosphinic acid pendant arm and then the number of isomeric
species in an aqueous solution of the LnIII complexes is expected
to increase, because of the presence of R and S epimers on the
phosphorus atom.31–33 In the case of the parent Eu–DO3A-PABn
complex, the solution1H NMR spectrum showed the presence of
three isomers: m1, m2 and M1.14a
Fig. 1 1H NMR spectrum at 298 K (a) and 31P NMR spectra of
[Eu2(CS(DO3A-PNBn)2)(H2O)2]2− at 273 (b), 298 (c) and 363 K (d). All
the spectra were recorded at 9.1 T and pH 7.
This journal is
The Royal Society of Chemistry 2006
Dalton Trans., 2006, 2323–2333 | 2325
©

View Article Online
that correspond to three isomeric species. The four resonances
at low field (28.6–38.6 ppm) are easily assigned to an M-type
isomer (MR(S)), whereas the two set of signals in the range 10.5–
conditions. Detailed information on the relaxometric properties of
the GdIII complex is provided by the measurement of the magnetic-
field dependence of the 1H relaxivity (nuclear magnetic relaxation
dispersion, NMRD, profiles) and the temperature dependence of
the 17O transverse relaxation rate, R2, of the solvent water. We
22.1 ppm can be attributed to two m-type isomers: mR(S) and mS(R)
.
The relative area of corresponding resonances is correlated to the
1
∼
relative population of the three isomers: mR(S) : MR(S) : mS(R) 54 : 38
have measured the H NMRD profiles, at 298 and 310 K, from
=
: 8. The overall m : M ratio, estimated from the spectrum recorded
0.01 to 70 MHz (Fig. 2(b)) and the 17O data at 2.12 T in the range
273–353 K (Fig. 2(a)). The experimental data were analysed by a
simultaneous fitting procedure to the well-established Solomon–
Bloembergen–Morgan (SBM) equations for the inner-sphere (IS)
paramagnetic relaxation and to Freed’s model for the outer-sphere
(OS) relaxivity. The 17O data were interpreted on the basis of the
Swift–Connick equations (for full equations, see ESI†).1,2,5,35
at 273 K, was found to be 6 : 4. This value is quite similar to that
found for the parent monomeric complex, Eu–DO3A-PABn
much higher than for the Eu–DOTA complex.34
,
14a and
The 31P NMR spectra were more informative. The spectrum
at 298 K (Fig. 1(c)) is rather complex and shows four groups of
signals in the range 75–95 ppm. The relative area of the four sets
of resonances, calculated after deconvolution of the signals, is ca.
55 : 37 : 6 : 2, from high to low field. The strict correspondence
of these ratios with those calculated in the proton spectra was the
basis of the assignment of Fig. 1(c). By lowering the temperature
to 273 K (Fig. 1(b)), the resonances sharpen up and their number
increases to 13. On the other hand, the spectrum at 363 K shows
only a single, broad peak (Dm1/2 = 500 Hz) centered at 81 ppm
(Fig. 1(d)). Clearly, a number of isomeric species exists in solution
that mutually interconvert over a broad temperature range. The
large number of isomers evidenced by the 31P NMR spectrum at
273 K can be rationalized by assuming that both monomeric units
in the ditopic complex behave independently and that each unit
may assume four isomeric forms: mR, mS, MR and MS. It follows
that the ditopic complexes may exist in solution as a mixture
of ten different isomeric species that could originate up to 16
distinct resonances in the 31P NMR spectra (Table S1, ESI†). This
means that, at low temperature, nearly all the possible isomeric
species are present in the aqueous solution of [Eu2(CS(DO3A-
PNBn)2)(H2O)2]2−. The isomers may exchange through both arm
rotation and inversion of the macrocyclic ring and, as a result, a
single exchange-averaged signal is detected in the high temperature
limiting spectrum.
The differences between the 31P and the 1H NMR spectra could
be explained by three different reasons: (i) the phosphorus atom
is the origin of the R/S chirality and is then more sensitive to
R and S isomerism than the axial protons of the macrocyclic
ring; (ii) since both phosphinate groups are connected by the same
thiourea bridge, they are more sensitive to configurational changes
occurring at the opposite end of the ditopic complex; (iii) the
europium(III)-induced 31P NMR isotropic shifts are sensibly larger
than the proton ones and this results in a better dispersion of the
resonances.
Fig. 2 Relaxometric data of [Gd2(CS(DO3A-PNBn)2)(H2O)2]2−: (a) 17O
transverse relaxation rates normalized to 1 mM concentration, R2,r at 2.12
T; (b) 1H NMRD profiles at 298 (open circles) and 310 K (filled circles).
1H and 17O relaxometry
The longitudinal proton relaxivity of the ditopic [Gd2(CS(DO3A-
PNBn)2)(H2O)2]2− complex is 8.2 s−1 mM−1, at 20 MHz and
298 K. This value represents an increase of ca. 44% over the relax-
ivity of the corresponding monomeric complex and it is sensibly
higher than those reported for related monoaqua–GdIII dimers.
Given the large number (18) of parameters involved, it is
customary to fix some of them to reasonable or typical values
during the fitting procedure. The hyperfine coupling constant,
Ě
A/h, modulating the Gd–O scalar interaction was fixed to the
value found for the related monomeric Gd–DO3A-PABn complex
(−2.9 × 106 rad s−1).14a The gadolinium–water oxygen distance
(RGdO) and the gadolinium–water proton distance (RGdH) were
1
Furthermore, the H relaxivity is pH independent in the range
2.5 to 12 (Fig. S4†) in close analogy with the observation made
for the monomeric parent [Gd(DO3A-PABn)(H2O)]− complex.14a
The constant value of r1 over a wide pH range indicates that
significant structural changes (coordination, hydration or isomer-
ization equilibria) do not occur on passing from alkaline to acidic
˚
fixed to 2.6 and 3.1 A, respectively. These values are slightly
longer than those typically used and reflect the presence of the
bulky phosphorus acid group that increases the steric interaction
with the coordinated water molecule.14b,13,36 The outer-sphere
2326 | Dalton Trans., 2006, 2323–2333
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
1
Table 1 Results of the multiparametrical simultaneous fitting of H NMRD and 17O NMR data. The calculated values are compared with the data
published for selected related complexes (structures of the ligands are shown in Charts 1 and 2)
Gadolinium(III) complex of
c
c
d
Parameter
CS(DO3A-PNBn
)
pip(DO3A)2
bisoxa(DO3A)2
pX(DTTA)2
[Fe(tpy-DTTA)2]e
OHECf
DO3A-PABng
DOTAc
2
310r1a/s−1 mM−1
D2/1020 s−2
298sM/ns
6.1
0.23 0.02
53
39
183
37
15.9 0.2
1
5.8
0.17
670
34.2
171
20.4
19
4.9
0.21
714
38.5
106
21.7
15
12.8
0.16
111
45.4
278
20.9
34
15.7
0.28
196
39.6
410
23
3.5
0.92
2500
30.9
208
17.3
11
4.2
0.25
16.2
20.6
88
29
11.2
1
3.8
0.16
244
49.8
77
16.1
11
1
—
−3.7
2.38
—
—
—
—
—
—
—
5
4
DHM/kJ mol−1
298sRb/ps
3
1
Er/kJ
298sv/ps
18.6
1
Ev/kJ
1
4
1
1
8
4.6
—
dg2 /10−2
8
0.3
2.8
−4.2
2.5
3.1
∼8
3.5
—
—
—
—
9.5
−3.7
2.5
3.1
12.5
3.5
—
—
—
—
—
Ě L
Ah−1/106 rad s−1
−2.89
−3.8
2.5
3.1
∼8
3.5
—
—
—
—
−3.6
2.5
3.1
—
3.5
—
—
—
—
−3.7
2.5
3.1
6.5
4.5
—
—
—
—
−2.89
2.6
3.1
—
3.6
15
9
3.6
1
˚
R
R
R
GdO/A
2.6
˚
GdH/A
3.1
6.3
3.6
10
23
3.5
1
˚
GdGd/A
˚
A/A
DHMss/kJ
298
s
/ps
3
Rss
˚
Rss/A
298
s
/ns
Mss
q
qss
1
1
1
—
1
—
2
—
2
—
1
—
1
1
1
—
a Values at 20 MHz. b The sR values reported here represent rotational correlation times of the Gd–H vector (sR(H)). Where only sR(O) values were
published, the corresponding values of sR were calculated from sR(O) and the sR(H)/sR(O) ratio. c Ref. 18. d Ref. 24d. e Ref. 24c. f Ref. 39. g Ref. 14a.
contribution is controlled by the distance of shortest approach,
A, of the bulk water molecules to the gadolinium(III) ion and
the relative diffusion coefficient, of the water molecules and the
more than one order of magnitude shorter with respect to the
ditopic complexes with q = 1 and shorter by a factor of 2–4 with
respect to the multimeric complexes with q = 2. On the other
hand, the exchange lifetime of [Gd2(CS(DO3A-PNBn)2)(H2O)2]2− is
more than three times longer than that for the related monomeric
complex:14a 53 and 16 ns, respectively, at 25 ◦C. This could be an
effect associated with a different structure and/or dynamics of the
second hydration sphere for the monomeric and ditopic complexes.
In fact, an hydrophilic amino group in DO3A-PABn is replaced
by a hydrophobic thiourea bridge in CS(DO3A-PNBn)2. On the
other hand, the different water accessibility of the two complexes
as a result of their different solution structures (vide infra)
cannot be excluded. In any case, the water exchange rate of the
gadolinium dimer is rather close to the optimal range of values (ca.
20–40 ns) predicted by the theory and thus fast enough so as not
to limit the relaxivity. At the physiological temperature and at the
frequencies of clinical use (20–60 MHz), the relaxivity of the GdIII
complexes of medium-low molecular weight is mostly controlled
by their rotational dynamics. An inspection at the rotational
correlation times (298sR) of the ditopic complexes of Table 1 shows
that the value found for [Gd2(CS(DO3A-PNBn)2)(H2O)2]2− (183 ps)
is significantly longer than that of Gd2–bisoxa(DO3A)2 (106 ps),18
a complex featuring a rather flexible oxoethylene linker, and very
similar to that of [Gd2(pip(DO3A)2)(H2O)2] (171 ps),18 a complex
characterized by the presence of a rigid cyclic bridging moiety.
Moreover, 298sR of 183 ns is just about two times longer than
that found for the parent Gd–DO3A-PABn complex (88 ps).14a So,
the twofold increase in molecular weight on passing from the
monomer to the dimer is closely mirrored by a twofold decrease
in the molecular tumbling rate, thus suggesting for the ditopic
complex a compact solution structure and a restricted internal
motion about the thiourea spacer. This can be seen in Fig. 3
where the different contributions to the relaxivity are plotted
˚
complex, DGdH. The distance A was fixed to 3.6 A, whereas the
value of DGdH was calculated according to the semi-empirical
Hindman’s equation.37 However, the analysis of the data only
in terms of the standard inner- and outer-sphere contributions
was far from being satisfactory. In particular, the experimental
NMRD profiles could not be reasonably reproduced. As for the
corresponding monomeric derivative, also a contribution to the
proton relaxivity from the water molecules hydrogen-bonded to
the polar groups of the ligand had to be taken into account (the
second-sphere contribution). This is often observed in the case
of phosphorus-containing groups that are highly hydrophilic and
able to participate in tight hydrogen-bonding interactions with
a number of water molecules near the paramagnetic center.16
This contribution depends on the number of water molecules in
the second coordination sphere, qss, their average distance from
the metal center, Rss, their mean residence lifetime, sMss, and the
rotational correlation time of the vector joining the Gd3+ ion to the
protons of these water molecules, sRss. The hydration number qss
was set to 1, in accordance with the choice made for the monomeric
298
complex,14a the residence lifetime
˚
s
Mss
was fixed to the very short
value of 1 ns,38 and Rss to 3.5 A. By considering the second-sphere
contribution (SS) to the relaxivity the experimental data could be
nicely reproduced (Fig. 2) following a well established procedure
(for full equations, see ESI†).14a
The best-fit parameters are summarized in Table 1 and com-
pared with those published for Gd–DOTA, Gd–DO3A-PABn and
a few selected multimeric complexes. Perhaps the most striking
characteristics of Gd2–CS(DO3A-PNBn)2 is the short value of the
residence lifetime of the coordinated water molecule as compared
to the other complexes listed in Table 1. The value of 298sM is
This journal is
The Royal Society of Chemistry 2006
Dalton Trans., 2006, 2323–2333 | 2327
©

View Article Online
Finally, we checked the possible influence of intramolec-
ular Gd3+–Gd3+ interactions in the ditopic complex on the
relaxivity. Merbach at al. have recently reported an enhance-
ment of the electron spin relaxation arising from this ef-
fect for the dinuclear gadolinium(III) complex of the macro-
cyclic ligand OHEC (H8OHEC = octaazacyclohexacosane-
1,4,7,10,14,17,20,23-octaacetate; Chart 2) where the distance
39
˚
between the paramagnetic ions is about 6.5 A. A more rapid
electronic relaxation might affect the relaxivity and limit its
enhancement obtained by increasing the rotational correlation
time. In the absence of EPR data,24c,40 we checked the possi-
ble influence of magnetic coupling between the Gd3+ ions by
1
comparing the H NMRD profiles of the homo- and heterod-
inuclear [Gd2(CS(DO3A-PNBn)2)(H2O)2]2− and [GdY(CS(DO3A-
PNBn)2)(H2O)2]2− complexes, at 298 K and 310 K (Fig. 4). The
Fig. 3 Comparison of proton relaxivities of Gd2–(CS(DO3A-PNBn)2 and
Gd–DO3A-PABn complexes (25 ◦C, 20 MHz). The overall values are
depicted in the form of stacked bars representing the sum of the inner-
(IS), second- (SS) and outer-sphere (OS) contributions.
for the monomeric and ditopic complexes. The IS term for the
dimer is two times larger than for the monomer, whereas the OS
contribution is about the same for the two complexes.
To the best of our knowledge, the 1H relaxivity of
[Gd2(CS(DO3A-PNBn)2)(H2O)2]2− (8.2 s−1 mM−1, at 298 K and 20
MHz) is the highest so far reported for low-molecular weight
dinuclear complexes with q = 1. The reasons for this are the
optimization of the water exchange rate, the lack of fast local
motions and the presence of a sizeable contribution of the second
hydration sphere. The SS term makes a contribution of ca. 7%
to the relaxivity of both the mono- and dinuclear gadolinium
complexes (Fig. 3).
Fig. 4 1H NMRD profiles of Gd2–(CS(DO3A-PNBn)2 (filled symbols) and
GdY–(CS(DO3A-PNBn
310 K (squares).
) (open symbols) complexes at 298 (circles) and
2
Chart 2 Chemical structures of the ligands discussed in the text and in Table 1.
2328 | Dalton Trans., 2006, 2323–2333 This journal is The Royal Society of Chemistry 2006
©

View Article Online
relaxivity of the two complexes is nearly identical at both temper-
atures and over the entire range of magnetic fields investigated.
It follows that the intramolecular interactions involving the Gd3+
ions can be neglected since they do not have any influence on the
relaxivity of the ditopic complex.
˚
distances of 6.3 and 6.2 A, respectively. This result confirms the
hypothesis that the dimer adopts, in aqueous solution, a rather
compact structure with the two terminal DOTA-like subunits close
to each other.
Additional information is provided by the 13C T1 relax-
ation times of the heteronuclear GdIII/YIII complex. The 13C NMR
spectrum (Fig. 6) can be partially assigned by comparison with
the spectrum of the pure ligand (Fig. S5†) owing to the lack of
pseudocontact contribution of the Gd3+ ion to the lanthanide-
induced shift (LIS).41 Three groups of signals can be distinguished:
(i) CH2 carbons of the macrocycle ring and of the pendant
arms, (ii) aromatic carbons and (iii) carbonyl and thiocarbonyl
carbons. As the carbon atoms of the gadolinium-containing
NMR solution study of the homonuclear YIII/YIII and
heteronuclear GdIII/YIII complexes
The long sR value suggests for the dimer a rather isotropic tumbling
motion without a large contribution from local rotation about the
bridge connecting the two paramagnetic units. This opens up the
problem of assessing the solution structure of the complex.
From molecular modeling, the possible range of Ln · · · Ln
˚
macrocyclic subunit are at a distance shorter than 3.5 A from
˚
distances in solution can be easily determined: from 21 A for a
the paramagnetic centre, their resonances are not visible in the
NMR spectrum because of the exceedingly large line broadening.
Thus, the spectrum of Fig. 6 contains only signals corresponding
to carbon atoms of the yttrium-containing subunit. The 13C
T1 relaxation times were measured with the standard inversion
recovery technique using a p/2 pulse of 12 ls and their mean
˚
totally stretched conformation to 5.5 A for the most compact
structure. In order to get some information on the actual structure
of the ditopic complex in aqueous solution, we prepared the mixed
GdIII/YIII species and used the 89Y isotope as a reporting NMR
nucleus and the gadolinium(III) ion as a paramagnetic probe.
˚
value was about 0.015 s. From this an average distance of 8.2 A
from the Gd3+ ion can be calculated from eqn (1) (for a complete
list of relaxation times see Tables S1 and S2†).
Fig. 5 89Y NMR spectrum of [GdY(CS(DO3A-PNBn)2)(H2O)2]2− at 9.1 T
and 298 K.
The 89Y NMR spectrum of the heteronuclear GdIII/YIII com-
plex, at 9.1 T and 298 K, consists of two peaks separated by
ca. 49 ppm in the relative ratio of 4 : 5 (Fig. 5). Since an
unambiguous assignment of the resonances is not possible, the
signals were tentatively attributed on the basis of the comparison
with the proton NMR spectrum of the EuIII complex and on the
expectation that the population of the m-type isomeric species
increases across the lanthanide series.34 So, the signals at 126.3 and
175.4 ppm were assigned to the M and m isomers, respectively.
Likely, the broadening at the bottom of these resonances may
be indicative of contributions arising from the occurrence of
various isomeric species as discussed above. Then, we measured
the Gd3+-induced decrease of the T1 relaxation times of the two
89Y signals in order to estimate the Gd · · · Y distance. The 89Y
longitudinal relaxation times are primarily dominated by the
dipolar interaction, as the through-bond (contact) contribution
is negligible because of the large distance separating the two metal
ions. With the further assumption that the motion in solution is
isotropic, the following simplified Solomon–Blombergen equation
can be used:41
Fig. 6 13C NMR spectra of [GdY(CS(DO3A-PNBn)2)(H2O)2]2− at 9.1 T,
298 K and pH 7. The assignment is based on the comparison with the 13
NMR spectrum of the free CS(DO3A-PNBn)2 ligand.
C
Besides the approximations involved in the procedure used, it
is worth noting that the calculated interatomic distances only
represent mean values over all the isomeric species that are present
in aqueous solution. In any case, the distance data obtained point
to a structure where the two metal centres are spatially close,
in agreement with the results of the analysis of the relaxometric
data. An approximate solution structure of the ditopic complex
can be obtained by molecular modeling simulations using as
c 2I l2 b2
ꢀ
ꢁ
2
1
2
5
l0
eff
˚
constraints the Gd · · · Y distance of 6.25 A and the average
=
sR
(1)
6
T1,p
4p
r
˚
Gd · · · C distance of 8.2 A. The only result that could satisfy
In eqn (1), l0/4p represents the magnetic permeability of a
vacuum (1 × 10−7 H m−1), c I is the gyromagnetic ratio for 89Y
(1.32 × 107), leff is the effective magnetic moment of Gd3+ (7.94 lB),
b is the Bohr magneton, r is the metal–metal distance and sR
the rotational correlation time of the complex (183 ps, from
NMRD and 17O data). For the two 89Y NMR peaks we measured
relaxation times of 0.088 and 0.076 s that correspond to Gd · · · Y
both conditions involves a face-to-face arrangement of the two
O4-planes of the coordination polyhedrons (Fig. 7). In such a
geometry the dimer shows a rather spherical shape which accounts
for the prevalent isotropic reorientational dynamics in solution
and a limited presence of independent rotational motions of
the two terminal subunits. This conformational geometry might
be stabilized by the presence of hydrogen bonding interactions
This journal is The Royal Society of Chemistry 2006
Dalton Trans., 2006, 2323–2333 | 2329
©

View Article Online
known to be endowed with a rate of exchange of the coordinated
water molecule, kex, faster than that of the related M-type isomer.
In fact, the rate of exchange found is the highest measured for
298
dinuclear GdIII complexes:
k
ex
= (1.9
0.2) × 107 s−1. The
high relaxivity of the [Gd2(CS(DO3A-PNBn)2)(H2O)2]2− complex is
explained by the presence of a sizeable contribution from water
molecules hydrogen-bonded to the phosphinate groups and by
a reorientational motion of the complex largely isotropic. This
latter property is highly relevant as most of the ditopic complexes
reported are characterized by a relaxivity limited by the presence
of a relatively fast local rotation about the linker joining the two
terminal subunits. The relatively rigid and compact structure of
the ditopic complexes is explained on the basis of a favoured
conformation in solution where the two terminal macrocyclic
chelates are facing each other with the metal ions at a distance
˚
of ca. 6.3 A. This conformation is stabilized by a network of
hydrogen-bonding interactions involving the phosphinate groups
and the water molecules.
The high water solubility, stability, rate of water exchange
and relaxivity make the new dinuclear [Gd2(CS(DO3A-
PNBn)2)(H2O)2]2− complex a promising candidate for the prepa-
ration of multimeric (e.g. dendrimeric) conjugates for molecular
imaging applications.
Experimental
Fig. 7 Proposed solution structure of the [GdY(CS(DO3A-PNBn)2)-
(H2O)2]2− complex in solution obtained from molecular modeling. The
colours of the atoms are the following: Y (green), Gd (magenta), O (red),
N (blue), P (orange), S (yellow) and C (gray). The network represents
the solvent accessibility surface of the molecule. The coordinated water
molecules were included into the simulation but they are omitted in the
picture for clarity. The ionic radii of the Ln3+ ions are not to scale and are
just a guide to the eyes.
Ligand synthesis
Reagents and solvents were purchased from commercial sources
with the highest quality grade and were used as received. The
starting ligand DO3A-PABn·3H2O was synthesized according to
the literature procedure.14a TLC was performed on Kieselgel 60
F245 aluminium sheets with fluorescence indicator (Merck) in
iPrOH/acetic acid/water 6 : 1 : 8 mixture with detection using UV
irradiation, ninhydrin spray or iodine vapour. Elemental analyses
were performed at the Institute of Macromolecular Chemistry of
the Czech Academy of Science (Prague). The 1H, 13C and 31P NMR
spectra were recorded on a Varian UNITYINOVA 400 spectrometer.
ES/MS spectra were run on Bruker ESQUIRE 3000 equipment
with ion-trap detection in positive or negative modes.
involving the phosphinate oxygen atoms. In fact, a rich network of
hydrogen bonds has been observed in the solid state structure of
the lutetium(III) complex of the monophosphonic acid analogue
of DOTA (H5DO3A-P, Chart 2),36 where pairs of complexes
are orientated face-to-face in an arrangement quite similar to
that proposed for the ditopic species. In addition, the solvent
accessibility surfaces calculated using the water molecule as a
1,4,7,10-Tetraazacyclododecane-4,7,10-triacetice-1-{methyl[(4-
isothiocyanatophenyl)methyl]phosphinic acid} (DO3A-PNCS). The
aminobenzylic derivative DO3A-PABn hydrate (0.5 g, 0.856 mmol)
was dissolved in 5 ml of water in a 20 ml plastic (PE) vessel.
The pH value was then adjusted to 2–3 (detected by universal
pH-test paper) by addition of concentrated hydrochloric acid.
In the next step, thiophosgene (88 ll, 1.03 mmol, purity ∼90%)
in tetrachloromethane (5 ml) solution was added and the vessel
was quickly capped and sealed with parafilm. The mixture was
vigorously shaken in a horizontal position overnight. The course
of the reaction was followed by TLC with ninhydrin detection
until the violet-red spot of the starting amine (Rf = 0.3) had
disappeared. The water phase was extracted with CCl4 (2 × 5 ml)
and Et2O (2 × 5 ml) and evaporated to dryness yielding crude
isothiocyanate DO3A-PNCS in the form of a solidified glass. This
glass was ground to a powder and the excess hydrochloric acid
in the powder was removed by vacuum-pumping (2.7 kPa, 5 h,
50 ◦C) to afford the target product (0.55 g, yield 90%). This form
of the isothiocyanate was used in further reactions or it was
˚
probe (1.4 A in diameter) indicate that both coordination centers
are easily accessible to the water molecules. This result is also
in good agreement with the fast rate of water exchange of the
coordinated water molecule.
Conclusion
We have synthesized a new “ditopic” phosphorus-containing lig-
and, CS(DO3A-PNBn)2, characterized by two terminal macrocyclic
subunits joined with a thiourea linker commonly used for the
preparation of high-molecular-weight or targeted MRI CA. In
the corresponding lanthanide(III) complexes, the two metal ions
are nine-coordinate with one inner-sphere water molecule. In
aqueous solution the complexes are present as a mixture of
stereoisomers mutually interconverting at high temperature. At
room temperature, the 1H NMR spectrum of the EuIII derivative
indicates the prevalence of isomeric species with a twisted square
antiprismatic geometry (m-type isomer). This isomeric form is
2330 | Dalton Trans., 2006, 2323–2333
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
purified by flash chromatography on a short silica gel column
with iPrOH/acetic acid/water 6 : 1 : 8 elution for characterization
purposes. TLC (UV, iodine vapour): Rf = 0.45. Found: C, 38.01;
H, 5.93; N, 10.10; S, 4.54. C23H34N5O8PS·4HCl (M = 717.4)
requires C, 38.51; H, 5.34; N, 9.76; S, 4.47%. NMR dH (400 MHz,
D2O, 25 ◦C; tBuOH): 2.87–3.90 (24H, m, ring and pendant
–CH2–); 7.20 (2H, m, aryl); 7.23 (2H, m, aryl); dC (100.6 MHz,
D2O, 25 ◦C; tBuOH): 53.6 (1C, d, 1JCP 79.8, P–CH2–C6H4); 54.9,
KOH solution and re-adjusted to this value until the pH value
was stable. The resulting cloudy mixture was stirred at 60 ◦C for
12 h. To remove the excess lanthanide(III) ions, the mixture was
concentrated using a vacuum evaporator and chromatographed
on a short column (Gd) or just stirred with a small batch (Y, Eu)
of chelating resin Chelex 20 (Fluka) in K+-form. The pure complex
was eluted with water in one fraction. The fraction was evaporated
under vacuum to dryness leaving a glassy/powder solid. For high-
resolution NMR measurements (Y2- and Eu2-complexes), 20 mM
solutions were prepared by dissolving about 30 mg of the solid
complex in D2O (1 ml) followed by filtration through a 20 lm
syringe filter into an NMR tube.26
1
55.8, 56.7 (6C, –CH2– ring); 57.5 (1C, d, JCP 82, N–CH2–P);
58.3 (2C, –CH2–COOH); 58.9 (2C, –CH2–); 59.7 (–CH2–COOH);
131.1 (2C, d, JCP 2.6, aryl CH); 134.5 (1C, d, JCP 3.8, aryl C); 136.2
= =
(2C, d, JCP 5.3, aryl CH); 136.8 (1C, bs, –N C S); 139.3 (1C, aryl
C); 173.6 (1C, –COOH); 179.7 (2C, –COOH); dP (161.9 MHz,
[Gd(DO3A-PABn)(H2O)]− complex solution. The DO3A-PABn
hydrate (0.49 g, 0.839 mmol) was dissolved in water (5 ml) and
GdCl3·6H2O (0.468 g, 1.26 mmol, 1.5 eq.) was added. The pH
value was then adjusted to 7 w◦ith KOH solution and the mixture
was stirred and heated at 60 C overnight. The excess of Gd3+
was removed on a cation exchange column (Amberlite CG50 in
H+-cycle) with water elution: the lanthanide(III) salts were eluted
first in acidic fractions whereas the pure complex was found in the
following neutral fractions. ESI/MS: m/z (Gd isotopic pattern)
681.1, 681.8, 682.5, 683.2, 685.1 (M)−, C22H36N5O8PGd requires
685.8 (for an average Gd atomic mass).
D2O, 90 ◦C; ext. 85% H3PO4): 29.8 (bs). ESI/MS: m/z 572.4
(M + H)+, C23H34N5O8PS requires 572.6; 361.3 (M −CH2
−
PNCS)+, C15H27N4O6 requires 360.4.
CS(DO3A-PNBn)2. A portion of the DO3A-PABn hydrate (0.5 g,
0.856 mmol) was dissolved in 5 ml of water. Subsequently, the
DO3A-PNCS hydrochloride (0.675 g, 0.942 mmol, 1.1 eq.) was
added and the pH was adjusted to 9–10 by addition of aqueous
1.5 M KOH. This reaction mixture was then stirred overnight
at RT. The crude product was purified by chromatography on
cationic exchange resin Amberlite CG50 (100–200 mesh) in H+-
form with water elution. In the first fractions, only hydrochloric
acid was eluted, whereas the pure product was eluted in the later
fractions. At the end, traces of unreacted DO3A-PABn were washed
out. Product-containing fractions were unified, dissolved in a
minimal amount of water (5 ml) and the solution added dropwise
to a large excess of anhydrous EtOH (500 ml). The white precipitate
was left to mature overnight, filtered off and dried under vacuum.
The pure product was obtained as a white solid (0.844 g, 83%).
TLC (UV, iodine vapour): Rf = 0. Found: C, 45.53; H, 6.54; N,
10.73; S, 4.00. C23H35N5O8PS·3H2O (M = 1154) requires C, 45.98;
H, 6.35; N, 10.38; S, 3.96%. NMR dH (400 MHz, D2O, 25 ◦C;
tBuOH): 3.03 (4H, d, P–CH2–C6H4, JHP 16.4); 3.16–3.50 (36H, m,
ring CH2 + N–CH2–P); 3.56 (4H, N–CH2COOH); 3.71 (8H, N–
CH2COOH); 7.31 (4H, m, aryl); 7.36 (4H, m, aryl); dC (100.6 MHz,
[Gd(CS(DO3A-PNBn)2)(H2O)2]2−
complex
solution. The
[Gd(DO3A-PABn)(H2O)]− complex14a (0.5 g, 0.728 mmol) was
dissolved in water (5 ml) and the DO3A-PNCS hydrochloride
(0.548 g, 0.764 mmol, 1.05 eq.) was added. The solution pH
value was immediately adjusted to 9–10 with diluted KOH
solution and the solution was stirred overnight at RT. The
reaction was monitored by TLC and isothiocyanate was added in
small portions over an interval of 6 h until the Gd–DO3A-PABn
TLC spot had disappeared. This solution of the complex was
used directly for further reactions. ESI/MS: m/z (Gd isotopic
pattern) 1253 (M)−, C45H70N10O16P2SGd requires 1257; 1292
(M + K)−, C45H70N10O16P2SGdK requires 1296 (for average Gd
atom mass).42
[GdY(CS(DO3A-PNBn)2)(H2O)2]2− complex. To the reaction
mixture containing the [Gd(CS(DO3A-PNBn)2)(H2O)]2− complex
obtained in the previous reaction, solid YCl3 hydrate (0.43 g, 1.5
eq. based on the amount of DO3A-PNCS) was added stepwise. In
each step, the pH value was re-adjusted to 7 by addition of aq.
KOH. The resulting mixture was subsequently heated (60 ◦C) and
stirred for 12 h. Then, the solution was left to cool down and a
batch of Chelex 20 resin was added (100 mg per 5 ml). After stirring
the mixture overnight at RT, all of the solids were filtered off and
the clear solution was ultrafiltered on an Amicon cell through
a YC-05 membrane (Mw-cut-off ∼500) at 4 atm to remove any
inorganic salts. The overall filtrate volume was about 500 ml. The
supernatant containing the mixed GdIII/YIII complex and traces of
the bis-YIII complex was evaporated under vacuum to yield a glassy
powdered solid. ESI/MS: m/z 669.6 (M)2−, C45H70N10O16P2SGdY
requires 670.
◦
D2O, 45 C; EtOH): 41.6 (2C, d, JCP 86.2; –P–CH2–C6H4); 53.1,
52.6 (16C, ring –CH2–); 53.9 (2C, d, JCP 85.5, N–CH2–P); 57.7
(2C, N–CH2–COOH); 58.1 (4C, N–CH2–COOH); 129.2 (4C, aryl
CH); 133.2 (4C, d, JCP 4.6, aryl CH); 136.1 (2C, d, JCP 7.2, aryl C);
138.4 (2C, aryl C); 173.6 (4C, –COOH); 178.3 (2C, –COOH); 182.3
(1C, NH–C(S)–N); dP (161.9 MHz, D2O, 90 ◦C; ext. 85% H3PO4):
32.7 (bs). ESI/MS: m/z 1099.8 (M + H)+, C23H35N5O8PS requires
1102.1; 361.3 (DO3A − CH2)+, C15H27N4O6 requires 360.4.
Synthesis of lanthanide(III) complexes and NMR sample
preparation
Lanthanide(III) chloride hydrates were purchased from Aldrich,
Strem or Alfa. If not stated otherwise, all NMR measurements
were carried out in deuterium oxide (99.95% or 99.98% D) received
from Chemtrade (Germany).
[Ln2(CS(DO3A-PNBn)2(H2O)2]2− complexes in the solid state
(Ln = Y, Eu, Gd). The ligand CS(DO3A-PNBn)2 hydrate (0.1 g,
0.086 mmol) was dissolved in water (2 ml). Then, appropriate
lanthanide(III) chloride hydrate (1.1 eq.; 0.95 mmol) was added
and the pH of the solution was slowly adjusted to 7 with dilute
89Y NMR measurements
For the 89Y measurements, a VT-low-frequency 10 mm probe of the
Varian 400 MHz system was utilized. The spectrometer was tuned
and the spectra referenced to a 1 M aqueous solution of Y(NO3)3
This journal is
The Royal Society of Chemistry 2006
Dalton Trans., 2006, 2323–2333 | 2331
©

View Article Online
(dY = 0 ppm) containing 1 mol% GdCl3. For the measurement
of the mixed YIII/GdIII complex, a 0.5 M aqueous solution in
a 10 mm tube (overall volume about 2 ml) was utilized. The
89Y T1 relaxation times were measured by the standard inversion
recovery technique with 0.5 s repetition time (d1) and pulse delay
(d2) incremented in the range 0.0003–0.154 s to 9 exponentially
sampled points. For each increment about 6000 transients were
accumulated. Measurements were done at 25 ◦C and pH 7.
and MIUR (FIRB). M. B., P. H. and J. R. gratefully acknowledge
the invaluable support of a NATO travel grant (No. PST.CLG
980045). The work was carried out within the frame of COST
D18 and the EU-supported NoE projects EMIL (No. LSHC-
2004-503569) and DiMI (No. LSHB-2005-512146).
References and notes
1 The Chemistry of Contrast Agents in Medical Magnetic Resonance
1H relaxation measurements and 1H NMRD profiles
´
Imaging, ed. A. E. Merbach and E. To´th, Wiley, Chichester, UK, 2001.
2 Topics in Current Chemistry, Springer Verlag, Heidelberg, 2002, vol.
The water proton NMRD profiles were measured at 25 and 37 ◦C
with a Stelar Spinmaster Spectrometer FFC-2000 (Mede, Pv,
Italy) on about 1 mmol gadolinium solutions in non-deuterated
water. The exact concentration of the solutions was determined by
measurement of the bulk magnetic susceptibility shifts of a tBuOH
signal.43 The 1H T1 relaxation times were acquired by the standard
inversion recovery method with typical 90◦ pulse width of 3.5 ls,
16 experiments of 4 scans. The reproducibility of the T1 data
was 5%. The temperature was controlled with a Stelar VTC-
91 airflow heater equipped with a calibrated copper-constantan
thermocouple (uncertainty of 0.1 ◦C). The NMRD profiles were
measured over a range of magnetic fields from 0.00024 to 1.6 T
(corresponding to 0.01–70 MHz proton Larmor frequencies).
221.
3 P. Caravan, J. J. Ellison, T. J. McMurry and R. B. Laufer, Chem. Rev.,
1999, 99, 2293.
4 S. Aime, M. Botta and E. Terreno, Adv. Inorg. Chem., 2005, 57, 173.
5 S. Aime, M. Botta, M. Fasano and E. Terreno, Acc. Chem. Res., 1999,
32, 941.
6 S. Aime, C. Cabella, S. Colombatto, S. G. Crich, E. Gianolio and F.
Maggioni, J. Magn. Reson. Imag., 2002, 16, 394.
7 M. Woods, Z. Kovacs and A. D. Sherry, J. Supramol. Chem., 2002, 2,
1.
8 V. J. Venditto, C. A. S. Regino and M. W. Brechbiel, Mol. Pharm., 2005,
2, 302.
´
´
9 (a) S. Laus, R. Ruloff, E. To´th and A. E. Merbach, Chem. Eur. J., 2003,
9, 3555; (b) R. Ruloff, E. To´th, R. Scopelliti, R. Tripier, H. Handel and
A. E. Merbach, Chem. Commun., 2002, 2630.
´
10 Z. Ja´szbere´nyi, A. Sour, E. To´th, M. Benmelouka and A. E. Merbach,
Dalton Trans., 2005, 2713.
17O relaxation measurements
11 (a) M. Woods, Z. Kovacs, S. Zhang and A. D. Sherry, Angew. Chem.,
Int. Ed., 2003, 42, 5889; (b) M. Woods, M. Botta, S. Avedano, J. Wang
and A. D. Sherry, Dalton Trans., 2005, 3829.
Variable-temperature 17O NMR relaxation measurement were
performed on a JEOL EX-90 (2.1 T, 12.2 MHz) spectrometer
with external locking system (D2O). Experimental settings were:
spectral width 10 000 Hz, pulse width 7 ls (90◦), acquisition time
10 ms, 2048 scans and no sample spinning. The complex solutions
were enriched by addition of H217O (6%, Yeda, Israel) to give ca.
0.2% 17O concentration. Transversal 17O NMR relaxation rates,
R2, were calculated from the line-width at half-height of the 1◦7O
signals. The 17O R2 data were measured from 0–90 ◦C with 10 C
increments. In order to stabilize the temperature, the samples
were maintained in the probe for at least 10 min prior to the
measurements being taken.
12 (a) P. Lebdusˇkova´, J. Kotek, P. Hermann, L. V. Elst, R. N. Muller,
I. Lukesˇ and J. A. Peters, Bioconjugate Chem., 2004, 15, 881; (b) J.
Kotek, P. Lebduskova, P. Hermann, L. Vander Elst, R. N. Muller, T.
Maschmeyer, I. Lukesˇ and J. A. Peters, Chem. Eur. J., 2003, 9, 5899.
13 J. Rudovsky´, P. C´ıgler, J. Kotek, P. Hermann, P. Vojt´ısˇek, I. Lukesˇ, J. A.
Peters, L. V. Elst and R. N. Muller, Chem. Eur. J., 2005, 11, 2373.
14 (a) J. Rudovsky´, J. Kotek, P. Hermann, I. Lukesˇ, V. Mainero and S.
Aime, Org. Biomol. Chem., 2005, 3, 112; (b) J. Kotek, J. Rudovsky´, P.
Hermann and I. Lukesˇ, Inorg. Chem., 2006, 45, 3097.
15 M. Pola´sˇek, J. Rudovsky´, P. Hermann, I. Lukesˇ, L. V. Elst and R. N.
Muller, Chem. Commun., 2004, 2602.
ˇ
´
16 M. Botta, Eur. J. Inorg. Chem., 2000, 399.
17 J. Rudovsky´, P. Hermann, M. Botta, S. Aime and I. Lukesˇ, Chem.
Commun., 2005, 2390.
18 D. H. Powell, O. M. N. Dhubhghaill, D. Pubanz, L. Helm, Y. S. Lebedev,
W. Schlaepfer and A. E. Merbach, J. Am. Chem. Soc., 1996, 118,
9333.
Best fitting procedure
´
19 E. To´th, S. Vauthey, D. Pubanz and A. E. Merbach, Inorg. Chem., 1996,
The analysis of the 1H NMRD and 17O NMR T2 data were
performed with Micromath Scientist44 fitting routines based on
SBM, Freed and Swift–Connick equations (refer to ESI†).1,2,5,35
All the data were pre-processed and post-processed in Microcal
Origin.45
35, 3375.
20 E. Zitha-Bovens, L. Vander Elst, R. N. Muller, H. van Bekkum and
J. A. Peters, Eur. J. Inorg. Chem., 2001, 3101.
21 T.-M. Lee, T.-H. Cheng, M.-H. Ou, C. A. Chang, G.-C. Liu and Y.-M.
Wang, Magn. Reson. Chem., 2004, 42, 329.
22 (a) W. H. Li, S. E. Fraser and T. J. Meade, J. Am. Chem. Soc., 1999, 121,
1413; (b) W. H. Li, G. Parigi, M. Fragai, C. Luchinat and T. J. Meade,
Inorg. Chem., 2002, 41, 1418.
23 V. Comblin, D. Gilsoul, M. Hermann, V. Humblet, V. Jacques, M.
Mesbahi, C. Sauvage and J. F. Desreux, Coord. Chem. Rev., 1999, 186,
451.
Molecular modeling and graphical export
Molecular modeling was carried out using the Hypercube Hyper-
chem 6.01 program package with the MM+ force-field taking into
account only bond dipoles.46 The final data were then exported
to MSI WebLab ViewerPro 4.0 where the solvent accessibility
surfaces were also calculated.47
24 (a) R. Ruloff, G. van Koten and A. E. Merbach, Chem. Commun., 2004,
´
842; (b) J. B. Livramento, E. To´th, A. Sour, A. Borel, A. E. Merbach
and R. Ruloff, Angew. Chem., Int. Ed., 2005, 44, 1480; (c) J. Costa, R.
Ruloff, L. Burai, L. Helm and A. E. Merbach, J. Am. Chem. Soc., 2005,
´
127, 5147; (d) J. Costa, E. To´th, L. Helm and A. E. Merbach, Inorg.
Chem., 2005, 44, 4747; (e) J. B. Livramento, A. Sour, A. Borel, A. E.
´
Merbach and E. To´th, Chem. Eur. J., 2006, 12, 989.
Acknowledgements
25 S. Liu and D. S. Edwards, Bioconjugate Chem., 2001, 12, 7.
26 All lanthanide-containing NMR samples were tested for the presence
of free lanthanide(III) by dropping the test solution into a urotropine
buffered water solution of xylenol orange (pH 5). If the solution
remained orange the test was negative.
The authors are thankful to Vojteˇch Kub´ıcˇek and Dr David Sy´kora
for the measurement of the mass spectra. We are grateful to the
ˇ
GACR (grant No. 203/03/0168), GAUK (423/2004/B-CH/PrF)
2332 | Dalton Trans., 2006, 2323–2333
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
27 M. C. Alpoin, A. M. Urbano, C. F. G. C. Geraldes and J. A. Peters,
J. Chem. Soc., Dalton Trans., 1992, 463.
35 J. H. Freed, J. Chem. Phys., 1978, 68, 4034.
36 P. Vojt´ısˇek, P. C´ıgler, J. Kotek, J. Rudovsky´, P. Hermann and I. Lukesˇ,
Inorg. Chem., 2005, 44, 5591.
28 For the DIS experiment, a series of complexes with variable concentra-
tions of Dy3+ and 10% excess of the ligand was prepared.
37 J. C. Hindman, J. Chem. Phys., 1974, 60, 4488.
29 (a) S. Aime, L. Barbero, M. Botta and G. Ermondi, Inorg. Chem., 1992,
31, 4291; (b) S. Hoeft and K. Roth, Chem. Ber., 1993, 126, 869.
30 (a) S. Aime, A. Barge, M. Botta, A. S. De Sousa and D. Parker, Angew.
Chem., Int. Ed., 1998, 37, 2673; (b) M. Woods, S. Aime, M. Botta,
J. A. K. Howard, J. M. Moloney, M. Navet, D. Parker, M. Port and O.
Rousseaux, J. Am. Chem. Soc., 2000, 122, 9781; (c) F. A. Dunand, S.
Aime and A. E. Merbach, J. Am. Chem. Soc., 2000, 122, 1506; (d) F. A.
Dunand, R. S. Dickins, D. Parker and A. E. Merbach, Chem. Eur. J.,
2001, 7, 5160; (e) S. Zhang, Z. Kovacs, S. Burgess, S. Aime, E. Terreno
and A. D. Sherry, Chem. Eur. J., 2001, 7, 288.
38 (a) P. Caravan, M. T. Greenfield, X. Li and A. D. Sherry, Inorg. Chem.,
2001, 40, 6580; (b) S. Aime, M. Botta, E. Terreno, P. L. Anelli and F.
Uggeri, Magn. Reson. Med., 1993, 30, 583.
39 G. M. Nicolle, F. Yerly, D. Imbert, U. Bo¨ttger, J.-C. Bu¨nzli and A. E.
Merbach, Chem. Eur. J., 2003, 9, 5453.
´
40 E. To´th, L. Helm and A. E. Merbach, in The Chemistry of Contrast
Agents in Medical Magnetic Resonance Imaging, ed. A. E. Merbach
´
and E. To´th, Wiley, Chichester, UK, 2001, pp. 45–119.
41 J. A. Peters, J. Huskens and D. J. Raber, Prog. Nucl. Magn. Reson.
Spectrosc., 1996, 28, 283.
31 (a) S. Aime, A. S. Batsanov, M. Botta, R. S. Dickins, S. Falkner, C. E.
Foster, A. Harrison, J. A. K. Howard, J. M. Moloney, T. J. Norman, D.
Parker and J. A. G. Williams, J. Chem. Soc., Dalton Trans., 1997, 3623;
(b) S. Aime, A. S. Batsanov, M. Botta, J. A. K. Howard, D. Parker, K.
Senanayake and J. A. G. Williams, Inorg. Chem., 1994, 33, 4696.
32 W. D. Kim, G. E. Kiefer, J. Huskens and A. D. Sherry, Inorg. Chem.,
1997, 36, 4128.
42 In the ES/MS spectrum, there were many minor peaks corresponding
to (M + nNa)− and (M + nK)− with a fine splitting due to the natural
isotopic abundance of the gadolinium isotopes.
43 D. M. Corsi, C. Platas-Iglesias, H. van Bekkum and J. A. Peters, Magn.
Reson. Chem., 2001, 39, 723.
44 Scientist for Windows version 2.01, Micromath Inc., Salt Lake City, UT,
1995.
ˇ
33 (a) J. Rohovec, P. Vojt´ısˇek, P. Hermann, J. Mosinger, Z. Za´k and I.
45 Origin version 6.0, MicroCal Software Inc., Northampton, MA,
1999.
Lukesˇ, J. Chem. Soc., Dalton Trans., 1999, 3585; (b) J. Rohovec, P.
Vojt´ısˇek, I. Lukesˇ, P. Hermann and J. Ludv´ık, J. Chem. Soc., Dalton
Trans., 2000, 141.
46 HyperChem Release 7.5 for Windows, Molecular Modeling Syste´m,
Hypercube Inc., Ontario, Canada, 2002.
34 S. Aime, M. Botta, M. Fasano, M. P. M. Marques, C. F. G. C. Geraldes,
D. Pubanz and A. E. Merbach, Inorg. Chem., 1997, 36, 2059.
47 MSI WebLabViewerPro 4.0 Preview, http://www.msi.com,
2000.
This journal is
The Royal Society of Chemistry 2006
Dalton Trans., 2006, 2323–2333 | 2333
©