Design principles and theory of paramagnetic fluorine-labelled lanthanide complexes as probes for19F magnetic resonance: a proof-of-concept study

Article abstract of DOI:10.1002/chem.200902300

The synthesis and spectroscopic properties of a series of CF ₃-labelled lanthanide(III) complexes (Ln = Gd, Tb, Dy, Ho, Er, Tm) with amidesubstituted ligands based on 1,4,7,10tetraazacyclododecane are described. The theoretical contributions of

Full text of DOI:10.1002/chem.200902300

DOI: 10.1002/chem.200902300
Design Principles and Theory of Paramagnetic Fluorine-Labelled Lanthanide
Complexes as Probes for 19F Magnetic Resonance: A Proof-of-Concept Study
Kirsten H. Chalmers,[a] Elena De Luca,[a] Naomi H. M. Hogg,[a] Alan M. Kenwright,[a]
Ilya Kuprov,*[b] David Parker,*[a] Mauro Botta,[c] J. Ian Wilson,[d] and
Andrew M. Blamire[d]
Abstract: The synthesis and spectro-
scopic properties of a series of CF3-la-
belled lanthanideACHTUNGTRENNUNG(III) complexes (Ln=
for a given LnIII ion. Selected com-
plexes exhibit pH-dependent chemical
shift behaviour, and a pKa of 7.0 was
determined in one example based on
the holmium complex of an ortho-
spectroscopy experiments are reported,
aided by identification of salient low-
energy conformers by using density
functional theory. The study of fluorine
relaxation rates, over a field range of
4.7 to 16.5 T allowed precise computa-
tion of the distance between the LnIII
ion and the CF3 reporter group by
using global fitting methods. The sensi-
tivity benefits of using such paramag-
netic fluorinated probes in 19F NMR
spectroscopic studies are quantified in
preliminary spectroscopic and imaging
experiments with respect to a diamag-
Gd, Tb, Dy, Ho, Er, Tm) with amide-
substituted ligands based on 1,4,7,10-
tetraazacyclododecane are described.
The theoretical contributions of the 19F
magnetic relaxation processes in these
systems are critically assessed and se-
lected volumetric plots are presented.
These plots allow an accurate estima-
tion of the increase in the rates of lon-
gitudinal and transverse relaxation as a
function of the distance between the
LnIII ion and the fluorine nucleus, the
applied magnetic field, and the re-rota-
tional correlation time of the complex,
cyano
DO3A-monoamide
ligand,
which allowed the pH to be assessed
by measuring the difference in chemi-
cal shift (varying by over 14 ppm) be-
tween two 19F resonances. Relaxation
analyses of variable-temperature and
variable-field 19F, 17O and 1H NMR
Keywords: fluorine
agents lanthanides
·
imaging
magnetic
netic yttriumACTHNUTRGENUGN(III) analogue.
·
·
properties · NMR spectroscopy
Introduction
Over the past twenty years, magnetic resonance imaging
(MRI) has become one of the most important diagnostic
tools in clinical practice. MRI instrumentation operating at
1.4 to 9.4 T is widely available and has been used with great
success for solving problems as diverse as cancer diagnos-
tics,[1] catalyst design[2] and the identification of Egyptian
mummies.[3] The advent of MRI contrast agents based on
super-paramagnetic iron oxide particles[4] and GdIII com-
plexes[5] has enhanced the quality of information that can be
obtained from medical MRI scans by allowing certain tis-
sues or regions of the body to be selectively highlighted.
Contrast remains a difficult issue though; standard instru-
[a] K. H. Chalmers, E. D. Luca, N. H. M. Hogg, Dr. A. M. Kenwright,
Prof. D. Parker
Department of Chemistry
Durham University
South Road, Durham, DH1 3LE (UK)
E-mail: david.parker@dur.ac.uk
[b] Dr. I. Kuprov
Oxford e-Research Centre
7 Keble Road, Oxford, OX1 3QG (UK)
E-mail: ilya.kuprov@oerc.ox.ac.uk
[c] Prof. M. Botta
Dipartimento di Scienze dellꢀ Ambiente e della Vita
Universitꢁ del Piemonte Orientale, “Amedeo Avogadro”
Viale Teresa Michel 11, 15100, Alessandria (Italy)
1
ments use H nuclei, and the concentration and relaxation
rate of protons in certain tissues is frequently similar.[6]
It has long been recognised that standard MRI instru-
[d] Dr. J. I. Wilson, Prof. A. M. Blamire
Newcastle Magnetic Resonance Centre
Newcastle University, Campus for Ageing and Vitality
Newcastle-upon-Tyne, NE4 5PL (UK)
1
ments can be easily re-tuned from H to 19F nuclei, which
have very similar magnetic properties.[7] This has stimulated
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200902300.
1
the development of binuclear H/19F imaging probes and re-
134
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 134 – 148

FULL PAPER
search into fluorinated materials suitable for detection in
vivo.[8,9] The published works range from quantitative NMR
spectroscopic tracking of the metabolism of fluorinated
drugs[10] to imaging studies of organ inflammation processes
by using chemically inert perfluorinated compounds as con-
trast agents.[11] Perfluorocarbons have also been used to
signal changes in local oxygen pressure as a consequence of
their interaction with this paramagnetic species in vivo.[9b]
Several reports have addressed functional 19F MRI probes
in which the observed signal reports a change in the chemi-
cal state of the probe. These include systems designed to
detect an irreversible chemical transformation of the fluori-
nated probe, for example, a cleavage reaction catalysed by
an enzyme.[12] The limiting features of this approach include
the irreversibility of the chemical transformation and the
modest change in the observed 19F shift, typically dꢀ2 ppm
for pH-independent systems.[9b,13] A more promising ap-
proach involves a probe that undergoes a continuous and re-
versible transformation as a function of a given parameter
so that the observed changes in chemical shift can be cali-
brated to report, for example, the local pH.[7] Using the
chemical shift (rather than signal intensity) as a reporter
also avoids the problems associated with variations in probe
concentration.[12b–e]
An alternative approach that we introduced,[13,14] involves
carefully positioning a CF3 reporter group within 7 ꢃ of a
paramagnetic centre (a lanthanideACHTUNTRGNEUNG(III) ion) located in the
same complex. With different lanthanide ions, the resulting
complex can function either as a fluorine magnetic reso-
nance probe (with moderate relaxation enhancers, such as
Tb3+, Dy3+, Ho3+, Er3+ or Tm3+ [19] or as a 1H MRI contrast
)
agent with Gd3+ for complexes that possess a fast-exchang-
ing water molecule. This dual nature allows two distinct
types of experiment: 1) Dual-imaging studies, in which 19F
1
images are merged with morphologically matched H images
to allow the precise anatomic localisation of the 19F signal
intensity[8c,11c] and 2) fluorine-only chemical shift or intensity
imaging to track the physical transport and chemical trans-
formation of the fluorinated probe. In the latter case, the
positioning of the CF3 group needs to be carefully consid-
ered because the lanthanide-induced pseudocontact shift is a
sensitive function of molecular geometry.[19a]
Herein, we define the theoretical framework to this ap-
proach, and in this proof-of-concept study describe the syn-
thesis and spectroscopic properties of a series of lanthanide
(III) complexes (Ln=Gd, Tb, Dy, Ho, Er, Tm) of macrocy-
clic ligands based on 1,4,7,10-tetraazacyclododecane (cyclen,
Figure 1). We report the relaxation analysis of 19F NMR, 17O
NMR and 1H NMRD data, aided by conformational analysis
using density functional theory. The charge neutral com-
As in most magnetic resonance techniques, sensitivity is a
key issue with 19F NMR spectroscopy and MRI. The obvious
strategy for improving the sen-
sitivity is to increase the local
concentration of the probe. Al-
ternatively, the longitudinal 19F
relaxation can be accelerated
by introducing
a proximate
paramagnetic centre (e.g.,
Ln3+ or Mn2+).[13,14] Relaxa-
tion enhancements of at least
two orders of magnitude are
frequently achieved, which
may give rise to an order of
magnitude increase in signal
intensity. In practice, a balance
needs to be struck between the
benefits of relaxation enhance-
ment and the reduced detec-
tion sensitivity associated with
broader spectral lines, given
that transverse relaxation is
also enhanced.[13b]
Figure 1. Structures of the ligands.
The standard method of
generating paramagnetic relax-
ation enhancement (originally developed to assist 13C and
15N NMR spectroscopy[15]) involves adding a paramagnetic
agent to the sample. At low probe concentrations, however,
the improvement is quite modest[16] because the encounter
probability between the fluorinated reporter group and the
paramagnetic molecule is quite low. Better results are ach-
ieved if the contact time is increased by, for example, ion-
pairing[17] or reversible coordination.[18]
plexes selected should exhibit sufficient kinetic and thermo-
dynamic stability to allow their safe usage in vivo.[1,20,21] The
sensitivity benefits of using such paramagnetic fluorinated
probes are assessed in spectroscopic and preliminary
19F MRI experiments, as compared with a diamagnetic
yttriumACHTNUTRGNE(NUG III) complex.
Chem. Eur. J. 2010, 16, 134 – 148
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
135

D. Parker, I. Kuprov et al.
Fluorine relaxation in paramagnetic systems: The five non-
negligible spin relaxation processes for 19F nuclei** in non-
viscous solutions of paramagnetic molecules are due to the
stochastic modulation of chemical shift anisotropy (CSA),
the inter-nuclear dipole–dipole (DD) interaction, the elec-
tron–nucleus contact interaction, the electron–nucleus DD
interaction and its special case known as Curie relaxa-
tion.[22–26] Typical imaging fields are around 3.0 T, placing the
systems in question firmly in the domain of perturbative
spin relaxation theories,[27–31] of which the Bloch–Redfield–
Wangsness (BRW) theory[28–30] is the one most widely used.
We shall also reluctantly bow to tradition and use the iso-
tropic tumbling approximation in the treatment below;
while the more sophisticated lattice models could be desira-
ble on theoretical grounds,[32–39] the Lorentzian spectral den-
sity does appear to work for the small molecules reported
here.
additional interaction modulation mechanism (by the elec-
tron relaxation for which T1e =T2e is assumed) has to be in-
cluded to give[25]
ꢀ
ꢁ
ꢂ
ꢃ
2
m0 gF2 ge2mB2 SðS þ 1Þ
3tRþe
7tRþe
2
ð4Þ
R1 ¼
R2 ¼
þ
1 þ w2Ft2Rþe 1 þ w2e t2Rþe
15 4p
r6
ꢀ
ꢁ
ꢂ
ꢃ
1
m0 gF2 ge2mB2 SðSþ1Þ
3tRþe
13tRþe
2
4tRþe
þ
þ
1 þ w2Ft2Rþe 1 þ w2et2Rþe
15 4p
r6
ꢂ1
in which tR+e =(tꢂR1 +T ). Although the point dipole ap-
1e
proximation is not usually valid in the electron–nuclear case,
the compact nature of the lanthanide f orbitals and the long
(over 5 ꢃ) electron–nucleus separation in complexes exam-
ined here make it reasonable.
The electron–nucleus contact relaxation terms are en-
countered in systems in which conformational mobility or
electron relaxation lead to stochastic modulation of the elec-
tron–nucleus contact interaction.[37]
The BRW theory expressions for the CSA and inter-nu-
clear DD relaxation rates (ignoring the dynamic frequency
ˆ
ˆ
shift, referring to the FZ state for R1 and the Fꢁ states for
R2) are given by[35,36]
SðS þ 1Þa2HFC 2te
ð5Þ
R1 ¼
R2 ¼
3
1 þ w2et2e
2
15
tR
1 þ w2Ft2R
R1 ¼ D2CSAB20g2F
ð1Þ
ꢀ
ꢁ
SðS þ 1Þa2HFC
te
ꢀ
ꢁ
te þ
3
1 þ w2et2e
1
45
3tR
1 þ w2Ft2R
R2 ¼ D2CSAB20gF2 4tR þ
in which aHFC is the isotropic hyperfine coupling constant,
and the correlation time, te, for its stochastic modulation is
equal to the electron relaxation time. It should be noted
that Equations (4) and (5) become more complicated if
T1e ¼T2e.
ꢀ
ꢂ
ꢃ
2
2
2
X
2
1
m0
gi gFꢀh
3tR
1 þ w2F t2R
6tR
R1 ¼
þ
2
ri6F
1 þ ðwF þ wiÞ t2R
10 4p
i
ð2Þ
ꢁ
tR
1 þ ðwF ꢂ wiÞt2R
þ
The final relaxation mechanism, Curie relaxation, is due
to the stochastic modulation of the dipole–dipole interaction
between the nucleus and the static electron magnetic
moment that results from the partial polarisation of the
electron by the applied magnetic field.[25] From the point of
view of the nucleus, the resulting interaction is algebraically
equivalent to CSA, which means that the expressions for the
resulting relaxation rates, in their spectral density part, are
structurally similar to those in Equation (1).
ꢀ
ꢂ
ꢃ
2
2
2
X
2
1
m0
gi gFꢀh
3tR
6tR
R2 ¼
4tR þ
þ
ri6F
1 þ w2Ft2R 1 þ w2i t2R
20 4p
i
ꢁ
6tR
tR
þ
þ
2
2
1 þ ðwF þ wiÞ t2R 1 þ ðwF ꢂ wiÞ t2R
in which the i index runs over the nearby nuclei, B0 is the
magnetic induction, gF is the magnetogyric ratio of the fluo-
rine nuclei, tR is the rotational correlation time and D2CSA is
the second invariant of the chemical shielding tensor s with
eigenvalues {sX, sY, sZ}, as follows:
ꢂ
ꢃ
2
2
m0 w2Fg4eS2ðS þ 1Þ
3tR
2
ð6Þ
R1 ¼
R2 ¼
2
1 þ w2Ft2R
ð3kTÞ r6
5 4p
(
ꢀ
ꢁ
ꢂ
ꢃ
2
s2X þ s2Y þ s2Z ꢂ sXsY ꢂ sXsZ ꢂ sYsZ ðgeneral caseÞ
1
5
m0 w2Fg4eS2ðS þ 1Þ
3tR
2
D2CSA
¼
4tR þ
2
1 þ w2Ft2R
ð3kTÞ r6
4p
2
ðs ꢂ s Þ
ðaxial caseÞ
jj
?
ð3Þ
The relatively small ligand-field splitting of lanthanide f
orbitals frequently leads to more than one term being popu-
lated at room temperature, which means that the total elec-
in which s =sZ and s =sX =sY in the axial case.
k
?
The electron–nuclear dipolar terms are structurally similar
to the dipolar terms in Equation (2), with minor changes;
the fluorine Zeeman frequency (wF) can be ignored com-
pared with the much larger electron frequency (we) and an
tron angular momentum SACTHUNRGTNE(NUG S+1) is an effective quantity. The
following (experimentally determined[31]) parameter will be
used in the treatment below:
D
E
2
m2eff ¼ g2em2B
S
ð7Þ
^
[**] The description below applies to most other spin-1= nuclei.
2
136
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 134 – 148

Paramagnetic Fluorine-Labelled Lanthanide Complexes
FULL PAPER
It is important to note that the five processes above do
not exhaust the list of relaxation phenomena in fluorine-
containing paramagnetic systems. Although the processes
listed above do, in most cases, suffice, a complete descrip-
tion can only be claimed if the full relaxation super-operator
treatment is performed[39,40] and the system stays within the
validity range of the BRW theory.
19F relaxation model for lanthanide-based MRI contrast
agents: This section deals with the specific case of 19F relaxa-
tion in the lanthanide complexes reported herein. The CSA
and inter-nuclear DD contributions may safely be ignored
for two reasons. Firstly, on theoretical grounds, the 19F CSA
in the trifluoromethyl group is quite small[41,42] and inter-nu-
clear DD interaction is swamped by the much stronger elec-
tron–nuclear interactions. Secondly, from experimental ob-
servations, the fluorine T1 time in diamagnetic Y3+ and La3+
complexes is about 1 s, so CSA and inter-nuclear DD cannot
account for more than around 1 Hz in the measured para-
magnetic relaxation rates, which are all of the order of
100 Hz. The contact mechanism may also be ruled out on
theoretical grounds; a DFT calculation of isotropic hyper-
fine coupling between the 19F nucleus positioned over 5 ꢃ
away from the f-type unpaired electron expectedly results in
a zero value.
Figure 2. Volumetric plot showing the variation in R1 with applied field
Bo, mean distance r from the paramagnetic centre and the effective rota-
tional correlation time, tR; the analysis is based on Equation (8) and as-
sumes meff =10 BM with te =0.2 ps.
The two remaining mechanisms (electron–nucleus DD
and Curie relaxation) do contribute significantly. The result-
ing relaxation rates are:
ꢀ
ꢁ
ꢂ
ꢃ
2
m0 g2Fmeff
7tRþe
3trþe
2
2
R1 ¼
þ
1 þ w2et2Rþe 1 þ w2Ft2Rþe
15 4p
r6
ð8Þ
ꢂ
ꢃ
w2Fm4eff
3tR
1 þ w2Ft2R
2
2
m0
þ
2
ð3kTÞ r6
5 4p
ꢀ
ꢁ
ꢂ
ꢃ
2
m0 g2Fmeff
3tRþe
13trþe
Figure 3. Volumetric plot showing the variation in R2 with Bo, mean dis-
tance r from the paramagnetic centre and tR; the analysis is based on
Equation (8) and assumes a value of 10 BM for meff with te =0.2 ps.
2
1
R2 ¼
4tRþe
þ
þ
1 þ w2et2Rþe 1 þ w2Ft2Rþe
15 4p
r6
ꢀ
ꢁ
ꢂ
ꢃ
2
1
m0
w2Fm4eff
3tR
1 þ w2Ft2R
þ
4tR þ
2
ð3kTÞ r6
5 4p
typical ranges of distances, fields and correlation times, the
longitudinal 19F relaxation rate has a maximum around
w2Ft2R =1 and the transverse relaxation rate rises monoto-
nously as the correlation time is increased, courtesy of the
J(0) term in the corresponding equation.
A lanthanide complex of the type described herein (Ln=
Tb, Dy, Ho, Er and Tm) possesses a tR value of the order of
250 ps and normally adopts a common mono-capped square
anti-prismatic coordination geometry in solution.[5,13b] At a
field of 3.0 T and at a distance of between 5.5 and 6.5 ꢃ, R1
and R2 values of about 100 Hz are expected. At a field of
9.4 T or above, greater line broadening is predicted for com-
plexes of Dy, Ho and Tb and, to a lesser extent, for Er and
Tm.
While it is possible to make further simplifications to the
above equations (extreme narrowing is sometimes assumed),
we have collected sufficient data to skip further approxima-
tions and perform a direct global fit to Equations (8). As
demonstrated previously,[13] due to the presence of inter-
mediate timescale conformational exchange, R1 is a more re-
liable relaxation rate measure for our systems. We have,
therefore, only used the R1 parameter for the relaxation
analyses discussed here.
Volumetric plots of the 19F relaxation rates resulting from
Equations (8) are given in Figures 2 and 3, based on an ide-
alised Ln3+ ion with a magnetic moment, meff, of 10.0 Bohr
magnetons (viz. Tb 9.7; Dy and Ho 10.6; Er 9.6; Tm
7.6 BM) and an electronic relaxation time of 0.20 ps. Due to
the steep distance dependence of both the dipolar and Curie
terms, the nuclear relaxation rates diminish fairly rapidly as
the lanthanide–fluorine distance is increased. Within the
Complex synthesis: A series of macrocyclic ligands based on
1,4,7,10-tetraazacyclododecane, L1–L7 (Figure 1) has been
prepared, in which an ortho-trifluoromethyl group is incor-
porated into the phenyl ring of the amide substituent. A
Chem. Eur. J. 2010, 16, 134 – 148
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
137

D. Parker, I. Kuprov et al.
DFT geometry optimisation (vide infra) gave distances of
between 5 and 7 ꢃ between the CF3 group and the para-
magnetic centre. Ligand L1 was selected as a model complex
for use in vitro, and is only for the purposes of comparison.
The lower stability of the lanthanide complexes of such a
tetraamide ligand precludes their safe usage in vivo. In its
complexes there may be four equivalent CF3 groups, which
would lead to better 19F NMR spectroscopic sensitivity.
ces in a 1:1:1:1 ratio. The diamagnetic complex [Y(L1)]3+
gave rise to one resolved resonance, although at 656 MHz a
minor species could be discerned as a shoulder. Previous
crystallographic analyses of structurally related tetraamide
complexes have confirmed the presence of 9-coordinate sys-
tems with one coordinated water molecule that adopt a
mono-capped C4 symmetric square antiprismatic geome-
try.[45,46] The sense of the 19F pseudocontact shift (relative to
the diamagnetic [Y(L1)]3+ complex) follows the sign of the
Bleaney coefficients;[19a] these are negative for Tb, Dy and
Ho and positive for Er, Tm and Yb. This behaviour is con-
sistent with the adoption of a common solution structure
across this series, with the CF3 groups at a similar distance
from the metal ion and subtending a similar angle to the
principal axis of the complex.
The lanthanideACHTUNGTRENNUNG(III) complexes of monoamide derivatives
of DO3A, for example, L2–L7, (and certain phosphinate ana-
logues) exhibit high kinetic and thermodynamic stability
with respect to metal dissociation, and hence are appropri-
ate systems for consideration for use in vivo.[13,36,43a] They
also form several isomeric species in solution,[43b] determined
by the relative energy of stereoisomers with differing N-C-
C-O (typically ꢁ308) and N-C-C-N (ꢁ608) dihedral angles.
Evidently, it is desirable that one major 19F resonance is ob-
served, associated with the preferential formation of a low-
energy stereoisomer.
For the Tb, Ho, Er and Tm complexes of ligand L2–L5
and L7, similar spectral features were noted. Immediately
following dissolution of the complex in water, one main spe-
1
cies (ꢃ60%) was observed by using 19F and H NMR spec-
Ligand L1 was prepared by alkylation of cyclen with the
a-haloamide 1 (MeCN, K2CO3), whereas ligands L2–L7 were
obtained by stepwise alkylation of readily available ester 2,
followed by de-protection (CF3CO2H/CH2Cl2/208C). Com-
plex formation and purification followed standard methods
troscopy (pD=5.4, 295 K), together with up to eight minor
species that varied in relative proportion according to the
nature of the LnIII ion (see the Supporting Information).
The 19F resonance in the corresponding GdIII complexes was
observed as a broad (ꢄ1000 Hz at 295 K, 188 MHz) reso-
nance, with a chemical shift very close to that of the analo-
by using [LnACHTUNGTRENNUNG
(OAc)3] in aqueous media (for L2–L7) and a
LnIII trifluoromethylsulfonate salt in MeCN for L1a.
gous diamagnetic Y
(III) complex (d=ꢂ62.5 ppm), and with
a longitudinal relaxation rate, R1, of 1400 Hz (188 MHz,
pD=5.4, 295 K).
Spectroscopic properties: 19F NMR spectroscopic data for
the complexes are reported in Table 1. For the [Ln(L1)]3+
complexes,[5,44] one major species was observed in every
case, with its fraction falling in a sequence that echoed the
ionic radius change, as follows: Tm (87%)>Er (85%)>Ho
(75%)>Tb (70%). In each case, a second species lacking
C4 symmetry was observed, which appeared as four resonan-
Complexes of the ortho-substituted ligands (L4b, L6a/6b),
for example, ortho-ester L6a and derived carboxylate L6b, be-
haved very differently. For each LnIII complex of L6a, two
species separated by d=49.7 (Tb), 33.4 (Ho), 20.0 (Er) and
53.6 ppm (Tm) were observed in a 1:1 ratio. For [LnACHTUNGTRENNUNG
(L6b)],
the two species were present in a 3:1 ratio with very similar
19F shifts to the ester series (Table 1) but with much greater
linewidths. The europium emission spectra for the ester and
acid complexes were identical in form and were the same at
pH 5 and 8. Measurements of the radiative lifetime of the
europium excited state, k, were made in H2O and D2O to
allow assessment of the complex hydration state. Values of k
Table 1. 19F NMR chemical shift data [ppm] for lanthanide
ACTHUNTGRENNG(U III) com-
plexes (295 K, pD=5.4, Ln=Tb, Ho, Er, Tm, Dy).[a,b]
Complex
[Ln(L1)][c,d]
[Ln(L2)][e,f]
Tb
ꢂ53.9
ꢂ51.9
Ho
ꢂ59.0
ꢂ64.2
Er
ꢂ63.5
ꢂ64.8
Tm
ꢂ65.1
ꢂ77.4
Dy
for [Eu
G
(L6a)] were 1.84 (H2O) and 0.54 msꢂ1 (D2O); for [Eu-
ACHTUNGTRENNUNG
G
G
N
ACHTUNGTRENNUNG
(L6b)] the corresponding values were 1.70 and 0.60 msꢂ1
.
[Ln
[Ln
[Ln
[Ln
N
ꢂ64.3
ꢂ65.0
This data is consistent with europium hydration states, q, of
1.2 and 0.9, respectively. The common hydration state and
the absence of change in the form and relative intensity of
the Eu emission spectra strongly suggests that a common
metal coordination environment is adopted in solution in
each case, with no evidence for carboxylate ligation to Eu.
A variable-temperature 1H NMR spectroscopic study of
A
G
G
G
[Ln(L5)]
ꢂ50.1
ꢂ89.0
ꢂ55.1
ꢂ82.9
ꢂ64.7
ꢂ49.7
ꢂ77.5
ꢂ38.0
ꢂ58.3
[Ln
ACHTUNGTRENNUNG
A
[Ln
A
A
both [TmACHTNUTRGENNGU ACHTUNGTRENNUGN
(L6a)] and [Tm(L6a)]ꢂ was carried out at pH 5
[Ln(L7)][f]
ꢂ66.7
(ꢂ40.6)
(D2O insert, 470 MHz) over a temperature range of 283 to
323 K. For the ester complex, the observed linewidth of
each resonance increased with increasing temperature. At
293 K the linewidth was 70 Hz and at 323 K each resonance
had a linewidth of 600 Hz, but there was less than a d=
[a] The ligands each have a chemical shift of d=(ꢂ61.6ꢁ0.5) ppm under
these conditions. [b] The corresponding GdIII complexes give rise to very
broad resonances at dF =(ꢂ62ꢁ2) ppm, w1/2 =3000 Hz at 4.7 T. [c] In
80% CD3OD/D2O. [d] Only one species was observed at d=ꢂ62.5 ppm.
[e] The major species was observed at d=ꢂ61.0 ppm. [f] Values in paren-
thesis refer to the second most abundant species (ratios are 1:1 for [Ln-
5 ppm change in the chemical shift. For [Tm
resonances shifted and broadened as T increased and a coa-
(L6b)]ꢂ, the two
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG ACHTUNGTRENNUNG
(L4b] and for [Ln(L6a/L6b)], the ratios are 1:1 and 3:1).
138
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 134 – 148

Paramagnetic Fluorine-Labelled Lanthanide Complexes
FULL PAPER
lescence phenomenon was observed at 320 K. The major
resonance was observed at d=ꢂ84 ppm (w1/2 ꢄ1800 Hz).
Evidently, these conformers (Dn=24,500 Hz) are in dynamic
exchange on the NMR timescale, with an activation energy
of the order of 50 kJmo1ꢂ1, based on a simple Eyring analy-
sis. Further information on this process and the nature of
these isomers was obtained in computational DFT studies
(vide infra).
NMR timescale, only one signal was usually observed,
which corresponds to the weighted average of the chemical
shifts of the amide and its conjugate base. The pKa associat-
ed with deprotonation of the amide NH, follows the se-
quence
o-CN<p-NO2 ꢄp-CN<p-CO2Et<o-CO2Et<p-
ꢂ
CO2ꢂ !o-CO2 . This order reflects the combined effects of
the electronegativity of the substituent and Coulombic re-
pulsion, and correlates well with published pKa values for
the ionisation of related substituted phenols.
Protonation equilibria: The nature of the substituents of the
aromatic ring in ligands L1–L7 determines the pKa of the
amide hydrogen and thereby the pH sensitivity of the ob-
served NMR chemical shift. Complexes of L1, L2 and L7,
which have H or an OH group in the para position, showed
no variation in the 19F chemical shift at pH values between
3.5 and 8.0, consistent with a pKa of ꢃ9.5. For each of the
LnIII complexes of L3–L6, the chemical shift of the CF3
group changed markedly between pH 6 and 10 (Table 2 and
Figure 4). Because the acid–base equilibrium is fast on the
The two major isomers of complexes of the ortho-cyano
ligand, [Ln
sociated with amide NH deprotonation at around pH 7; for
example, the pKa for [Ho
(L4b)] is 7.0 (Figure 5). The chemi-
(L4b)], exhibit a large change in chemical shift as-
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
Table 2. Chemical shift [ppm] and linewidth [Hz] data (295 K, 188 MHz,
H2O, D2O capillary lock, 1 mm) for LnIII complexes of L5 and L4a (pKa =
(7.77ꢁ0.03) and (7.81ꢁ0.03), respectively).
Complex
pH 5
pH 9
dF
w1/2
dF
w1/2
[Tb(L5)]
[Ho(L5)]
[Er(L5)]
[Tm(L5)]
ꢂ50.2
ꢂ55.4
ꢂ64.7
ꢂ77.5
ꢂ50.1
ꢂ40.6
ꢂ84.6
ꢂ55.4
ꢂ48.8
ꢂ78.1
ꢂ64.1
ꢂ49.6
ꢂ67.6
ꢂ78.2
ꢂ39.0
ꢂ88.1
73
39
63
45
63
ꢂ26.4
366
340
120
277
150
190
ꢂ36.8
ꢂ74.5
ꢂ89.6
ꢂ32.8
ꢂ3.9
[Tb
[Tb
U
500
ꢂ73.1
ꢂ42.5
ꢂ25.6
ꢂ68.8
ꢂ72.2
ꢂ53.1
ꢂ81.6
ꢂ87.6
ꢂ39.9
ꢂ109.0
[Ho
[Ho
U
45
420
125
195
[Er
[Er
U
39
425
95
190
[Tm
[Tm
U
36
420
99
195
Figure 5. Top: Variation in dF with pH for each isomer of [Ho
N
)ACHTUNGTRENNUNG(OH2)]
~
*
(
perature). Bottom: The difference in chemical shift between the isomers
due to the variation in pH. Conditions: 295 K, H2O with D2O capillary,
0.1m NaCl; pKa 7.0.
cal shift of each isomer exhibits a different sensitivity to pH,
so that the shift difference between the two species allows a
simple means of determining the pH. This is a very attrac-
tive method for determining pH because no chemical shift
referencing is required and the pKa value of 7.0 is very well
suited to studies in biological media.
In certain cases (e.g., [Ho(L5)]ꢂ), the 19F NMR linewidth
of the conjugate base is significantly broader, by about a
factor of eight, than the protonated amide [HoACTHNUTRGNEUNG
(L5H)]. At
higher temperatures, the linewidth decreased considerably.
Such behaviour is consistent with the presence of an inter-
mediate timescale conformational exchange in the deproton-
Figure 4. Variation in dF with pH for each isomer of [Ho
N
ACHTUNGTRENNUNG
(Ho3+ (major form): and Ho3+ (minor form): ), and [Tm
ACHTUNGTRENNUNG
&
*
(Tm3+ (major form):
!
~
H2O with D2O capillary, 0.1m NaCl.
Chem. Eur. J. 2010, 16, 134 – 148
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
139

D. Parker, I. Kuprov et al.
ated complex. This tendency was most pronounced for
[Ho(L5)] and [Tb(L5)], and was less apparent with [Er(L5)]
or [Tm(L5)], (see the Supporting Information). For the
Relaxation analysis: 19F longitudinal relaxation rates were
measured by using the inversion-recovery technique (with-
out proton decoupling) in dilute (1.0 mm) solutions in D2O
(pD=5.4; D2O used for lock) at 295 K. Inversion-recovery
type function was fitted to the resulting data by using Lev-
enberg–Marquardt minimisation of the non-linear least
squares error functional. The resulting relaxation rates
(Tables 3 and 4) follow the order: Dy>Tb>Ho>Er>Tm,
consistent with the presence of m2eff and me4ff terms in Equa-
para-substituted analogues, [LnACHTUNGTRNEUNG
(L3a)] (p-CO2Et) and
[Ln(L4)] (p-CN), somewhat different behaviour was ob-
served for the variation in the linewidth with pH. In these
cases, the linewidth reached a maximum at the pKa and de-
creased to a lower value as the pH was raised further. The
exchange broadening observed in these cases presumably re-
lates to the effect of chemical
exchange between the amide
and its conjugate base.
Table 3. 19F longitudinal relaxation rates for [Ln(L1)]Cl3 and [Ln
ACHTUNGTRENNUNG
Ln
[Ln(L1)]3+
4.7 T
ACHTUNGTRENNUNG
Two different chemical ex-
change processes can be con-
sidered to account for the ex-
tensive line broadening, most
clearly exemplified with the
conjugate base for [Ho(L5)]ꢂ
and [Tb(L5)]ꢂ. These involve
either a lanthanide re-coordi-
nation from O to N or a cis–
trans isomerisation around the
carbon–nitrogen double bond.
We reported in a preliminary
communication that extended
DFT calculations could be
used[13b] to assess the likeli-
hood of these processes. Lan-
thanide re-coordination was
ruled out because the calculat-
ed energy difference between
O- and N-coordinated consti-
9.4 T
185ꢁ4
11.7 T
224ꢁ3
16.5 T
282ꢁ8
4.7 T
9.4 T
11.7 T
16.5 T
149
100
250
147
323
179
565
192
Tb
Ho
Er
115ꢁ2
91ꢁ1
233ꢁ2
333ꢁ4
588ꢁ15
313ꢁ8
313ꢁ6
161ꢁ3
294
84.0ꢁ2.0
45.8ꢁ0.3
33.0ꢁ0.6
192.1ꢁ1.5
108.7ꢁ1.0
59.7ꢁ1.3
249.6ꢁ2.6
143.3ꢁ0.3
76.2ꢁ0.4
340.8ꢁ7.5
199.5ꢁ0.7
102.3ꢁ0.4
71ꢁ1
59.0ꢁ0.5
45ꢁ1
53
137ꢁ2
152.0ꢁ1.2
88ꢁ1
100
172ꢁ3
208.5ꢁ2.5
110ꢁ2
135
Tm
37
56
68
152
[a] For [Y(L1)]Cl3, R1 =0.78ꢁ0.03 sꢂ1 and R2 =2.9ꢁ0.2 sꢂ1. [b] Values for [Ln
(L6a)] refer to the two major con-
ACHTUNGTRENNUNG
formers that are in slow exchange on the NMR time-scale. [c] For [LnACHTNUTGRNEUNG
(L6b)], the two major species observed
are in intermediate exchange so that the same or similar R1 values are observed, notably at lower field.
Table 4. 19F relaxation parameters [Hz] for [Ln(L7)] (295 K, pD=5.4, D2O).[a]
magnetic
field [T]
Tb[b,c]
R1
Dy[b,c]
R1
Ho[b,c,d]
R1
Er[b,c]
R1
Tm[b,c]
R2
R2
R2
R2
R1
R2
4.7
9.4
11.7
16.5
74ꢁ6
124
206
271
407
89ꢁ8
156
355
543
740
45.5ꢁ2.5
124ꢁ10
87
34ꢁ2
120
198
251
317
22.9ꢁ0.5
47ꢁ7
53
89
112
168
133ꢁ7
162ꢁ11
201ꢁ13
286ꢁ13
192
267
441
75ꢁ3
102.5ꢁ0.7
144ꢁ2
161.7ꢁ0.5
211.1ꢁ0.5
169ꢁ19
57.5ꢁ0.5
74.4ꢁ0.5
238.7ꢁ1.6
[a] In these complexes, the CF3 group is estimated to be (6.9ꢁ0.8) ꢃ from the LnIII ion if a global fitting analy-
sis of all data sets is used, or (6.24ꢁ0.02) ꢃ, if literature meff data is used in the calculation (see text). [b] The
estimated standard deviations for R1 are given. [c] R2 values were estimated as (pw1/2) for a Lorentzian line fit.
tutional
isomers
was
62 kJmolꢂ1 in favour of the O-
bound isomer. The calculated
energy difference between the
cis and trans isomers was
18 kJmolꢂ1, and calculated ac-
tivation energies for the for-
[d] The R1 data at 4.7 T for related HoIII complexes was similar: [Ho(L2)] 48 sꢂ1, [Ho
N
ACHTUNGTRENNUNG
(L3b)]ꢂ
50 sꢂ1, [Ho(L4)] 42 sꢂ1, [Ho(L5)] 40 sꢂ1. Similar trends were observed for the Tb
plexes of L2–L5.
G
N
ACHTUNGTRENNUNG
ward and backward isomerisation reactions were computed
to be 27 and 45 kJmolꢂ1, respectively. Therefore, the isomer-
isation process depicted in Scheme 1 offers a reasonable ex-
planation, with preferred conformers placing the bulky CF3
group away from the amide oxygen.
tions (8). For a given lanthanide, the relaxation rates drop in
the sequence [Ln
(L6a)]>[Ln(L1)]3+ >[Ln(L7)], which sug-
gests that the distance between the CF3 group and the para-
magnetic centre also decreases in this order. For the para-
substituted analogues of [Ln(L7)], that is, [Ln(L2)–Ln(L5)],
the measured relaxation rates were found to be very similar.
AHCTUNGTRENNUNG
ꢂ
To extract the average Ln F
lengths and rotational correla-
tion times from the field de-
pendence of the longitudinal
relaxation rate, a global non-
linear least-squares fit was per-
formed for Equation (8). It
ꢂ
was assumed that the Ln F
length and the rotational cor-
relation time remain the same
in the {Dy, Tb, Ho, Er, Tm} se-
quence (which makes r and tR
Scheme 1. Protonation and conformational equilibria in [Ln(L5)].
140
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 134 – 148

Paramagnetic Fluorine-Labelled Lanthanide Complexes
FULL PAPER
global variables). This is reasonable because the ionic radius
variation in this series is minor.[47] The values of meff and te
were kept local for every lanthanide. While each individual
fit is ambiguous, the global fit (Figure 6) has a single, well-
species was found to be 5.7ꢁ0.1 ꢃ and 6.3ꢁ0.2 ꢃ with a tR
value of 243ꢁ23 ps (Figure 7a). DFT calculations on this
system gave values for the distance (linear average over the
three fluorine atoms) of 5.6 ꢃ and 6.2 ꢃ. A tR value of
300 ps was estimated by using the Stokes–Einstein equation
Figure 6. Variation in R1 with wF for [Ln(L7)
C
~
!
&
^
*
(
), Ho ( ), Er ( ) and Tm ( ); 295 K, 1 mm). The solid lines represent
the global least-squares fit to Equation (8). The steeper rise of the Ho
curve as a function of magnetic field is due to its shorter electron relaxa-
tion time compared with the other lanthanides.
defined weighted least-squares minimum that results in tR =
270ꢁ10 ps, r=6.3ꢁ0.7 ꢃ and the effective magnetic mo-
ments in Table 5. A Stokes–Einstein estimate by using DFT
molecular geometry in D2O at 298 K gives tR =272 ps
Table 5. Experimental values of meff [BM].
Ln
Values from refs. [5,19]
Measured
Tb
Dy
Ho
Er
9.7
10.6
10.6
9.6
9.8ꢁ1.8
10.6ꢁ1.9
10.4ꢁ1.9
9.1ꢁ1.6
7.6ꢁ1.4
Tm
7.6
ꢂ
(hrmoli=6.18 ꢃ), with a DFT-calculated Ln F length of hri=
6.95 ꢃ. Slightly lower values of tR that had been reported to
1
come from H NMRD spectroscopic analyses of structurally
related complexes[48] are probably the consequence of the
fact that 1H NMRD spectroscopy implicitly examines the
motion of the Ln–(water proton) vector, whereas the analy-
sis above refers to the Ln–F vector. Another possible factor
is the fact that the viscosity of D2O is 20% higher than H2O.
Similar global fitting analyses of the R1 data at different
fields (Table 3) were also carried out for the Tb, Ho, Er and
ꢂ
ACTHNUTRGNEUNG
Tm complexes of L1 and L6a. For [Ln(L1)(OH2)]3+, the Ln
F length was estimated to be 6.1ꢁ0.2 ꢃ, with a rotational
correlation time of 255ꢁ20 ps. For the ortho-ester com-
Figure 7. Variation in R1 with wF for the two major isomers of a) [ErACHTUNGTRENNUNG
(L6a)-
(L6a)
ACHTUNGTREN(UGNN L-COOEt)-B) and b) [TbACHTUNTGERNNUGN ACHTUNGTRENNUNG(OH2)]
&
*
G
U
&
plexes, [LnACHTUNGTRENNUNG ACHTUNGTRENNUNG
(L6a)(OH2)], 19F NMR spectroscopy revealed the
presence of two isomeric species in solution in approximate-
ly 1:1 ratio (not changing significantly between the different
lanthanides examined). The value of the distance, r, for each
*
( : Tb
(L-COOEt)-A, : Tb
N
solid lines represent the global least-squares fit to Equation (8). c) Calcu-
lated volumetric plot showing the variation in R1 with Bo and meff for r=
5.7 and 6.3 ꢃ. The rotational correlation time, tR is fixed at 250 ps and
te =0.2 ps is assumed; the analysis is based on Equation (8).
Chem. Eur. J. 2010, 16, 134 – 148
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
141

D. Parker, I. Kuprov et al.
from the DFT molecular volume (effective molecular radius
of 6.6 ꢃ).
Using the approach defined above for the volumetric
plots (Figures 2 and 3), the variation of R1 with meff and Bo
can be calculated by using the experimental distance values
of 5.7 and 6.3 ꢃ. This analysis, based on tR =250 ps and te =
0.2 ps, generates two surfaces (Figure 7b), which highlight
the steeper increase in R1 at higher fields for the Er and es-
pecially the Ho complex.
It appears that, in designing a fluorinated paramagnetic
probe, to avoid very high relaxation rates (and hence much
broader lines) the CF3 group should be no closer than 5.5 ꢃ
from the Ln ion in Ho and Er complexes, and by inference
the Tb/Dy analogues. Thulium complexes would be more
useful if higher magnetic fields are used. On the other hand,
Tb, Dy and Ho complexes relax the CF3 group efficiently at
a distance of about 6 ꢃ and over the field range of 3 to 7 T
and should, therefore, allow rapid acquisition of signal in-
tensity without undue line broadening and the associated
sensitivity issues.
Figure 8. DFT-calculated structures showing the two low-energy isomers
of [Ln
(L6a)
N
ꢂ
d=ꢂ38.0 and ꢂ91.6 ppm compared with d=ꢂ62 ppm for
the diamagnetic yttrium complex. Given that the Bleaney
coefficients[19a] for Ho and Tm are ꢂ39 and +53 (relative to
Dy at ꢂ100), the sense and size of the observed shifts agrees
well with the expected values and lends support to the hy-
pothesis of the two low-energy conformers identified by the
DFT calculation (Figure 8).
Conformational analysis: The DFT calculations were per-
formed by using the Gaussian 03[55] package for La3+ and
Y3+ complexes; the structures of complexes with other lan-
thanides are likely to be nearly identical to the Y3+ com-
plex. The Y3+ results were very similar to those obtained
with La3+, and given the similarity (ꢁ0.02 ꢃ) of the ionic
radii of Ho3+, Y3+ and Er3+, only the Y3+ complex compu-
tations are described herein.
Relaxation properties of gadoliniumACTHNUTRGNEUNG(III) complexes: The ga-
dolinium complexes of these ligands can also be used as
conventional contrast agents for proton MRI by enhancing
the rate of relaxation of the bulk water signal. Therefore,
the proton relaxation behaviour of selected examples was
also characterised. It should be noted that only the mono-
amide derivatives are likely to be useful in vivo because the
tetraamide examples reported here possess limited water
solubility and have much lower kinetic stability with respect
to metal dissociation, precluding their safe use in vivo.
19F NMR spectra reported for [Ln
(L6a)] and [Ln(L6b)] and
ACHUTGTNRENNUG CAHTUNGTRENNUGN
for the ortho-cyano series [Ln
AHCTUNGTRENNUNG
of comparable intensity but with very different chemical
shifts, which suggests a conformational equilibrium between
two states with very different positions of the CF3 group
with respect to the pseudocontact shift field of the complex.
The observed slow rate of chemical exchange also indicates
a large activation energy for the transition between the two
conformers.
In aqueous solution, mono-aqua gadoliniumACTHNUTRGENUG(N III) com-
plexes of a variety of acyclic and macrocyclic octadentate li-
gands efficiently catalyse the relaxation of the bulk water
1H NMR signal.[5] The measured relaxation rate of water,
R1, in the presence of a gadolinium complex is made up of
diamagnetic and paramagnetic terms, of which the latter set
comprises contributions from inner (is), second-sphere (ss)
and outer-sphere water (os) molecules, [Eqns (9)–(11)]. The
increment of the paramagnetic contribution to the water
proton relaxation rate per unit concentration is termed the
relaxivity, r1p, of the complex.
A very likely candidate for the conformational transition
in question is the variation of the C-N-C-C dihedral angle
that results in the phenyl ring flip and consequent reposi-
tioning of the CF3 group relative to the metal centre. To
confirm this hypothesis, a relaxed potential energy scan was
performed with respect to the C-C-N-C dihedral angle for
the YIII analogue of complexes involving ligands L4–L6a. In
each case, two distinct minima were indeed identified
(Figure 8), separated by an energy barrier of >20 kJmolꢂ1
and featuring very different positions of the CF3 reporter
group with respect to the metal centre. Taking the metal–
water–oxygen vector to be the principal axis, the CF3 groups
are oriented at angles of ꢂ1148 and +448, respectively, in
each isomer. In the axial PCS case, these angles would give
rise to opposite pseudocontact shifts. The calculated Y2,0
spherical harmonic terms for the angles of ꢂ114 and +448
are ꢂ0.50 and +0.55. For the Ho complex, 19F chemical
shifts of d=ꢂ82.9 and ꢂ49.5 ppm were measured (Table 1).
For the thulium analogue, the corresponding values were
R1 ¼ R1d þ R1p ¼ R1d þ R1is þ R1ss þ Ro1s
r1p ¼ R1p=½GdðLÞꢅ
ð9Þ
ð10Þ
½GdðLÞꢅq
is
1p
ð11Þ
R
¼
55:6ðT1m þ tmÞ
For low-molecular-weight Gd complexes, the inner-sphere
contribution often dominates and is approximated by Equa-
tion (11), in which tm is the mean water residence lifetime
(at Gd), q is the number of coordinated water molecules
142
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 134 – 148

Paramagnetic Fluorine-Labelled Lanthanide Complexes
FULL PAPER
and T1m is the longitudinal relaxation time of the coordinat-
ed water protons. If water exchange is slow (tm >T1m), then
the measured relaxivity may tend to a lower limit in which
only the second-sphere and outer-sphere terms contrib-
ute.[17,49–51]
in marked contrast to structurally related tetraamide com-
plexes, such as [Gd(L8)
ACTHNUTRGEN(UNG OH2)]X3, for which the measured
relaxivity was 2.3 mmꢂ1 sꢂ1 (20 MHz, 298 K). They exhibit a
markedly different temperature dependence (Figure 10) as-
sociated with a very long water exchange lifetime (tm
ꢄ100 ms) that is independent of the nature of X (X=Cl, Br,
In the case of [Gd(L1)
ACTHNUGTRENNUG(OH2)]Cl3, the limited water solu-
bility of this complex precluded accurate measurements of
the temperature dependence of the transverse relaxation
rate of the 17O water signal. For each of the other complexes
examined, variable-temperature 17O NMR measurements
were undertaken with 19 to 32 mm solutions of the complex
in 17O-enriched water and the data were analysed by using
standard Swift–Connick methodology[52] to give estimates of
the water exchange lifetime, tm. Values of tm for [Gd(L2)-
A
ACHTUNGTRENNUNG CAHTUNGTERN(NUGN OH2)] fell in the range of 0.5 to
(L6a/L6b)
AHCTUNGTRENNUNG
water exchange rate was extracted by analysing the temper-
ature dependence of r1p (as detailed in reference [42]) in as-
sociation with classical analyses of the field dependence of
relaxivity (0.01 to 70 MHz (Figure 9), based on the Solo-
mon–Bloembergen–Morgan equations that rationalise the
time dependence of the water proton–GdIII dipolar coupling
with the added consideration of the second-sphere contribu-
tion.[5,17] It is appreciated that these models have their limi-
tations, but the analysis used here adopts this approach to
allow comparison with similarly derived literature data.
Figure 10. Comparison of the variation in the relaxivity (r1p) of [Gd(L1)-
8
&
*
G
(
) with [Gd(L )
U
butions have been highlighted in the latter case due to slow water ex-
change (tm of ca. 100 ms).
I, OAc, NO3).[49] The long tm values for [Gd(L8)
ACTHNUGTRNEUGN(OH2)]X3
are associated with a high activation enthalpy (Table 6) of
the order of 120 kJmolꢂ1. This contrasts with the value of
30 kJmolꢂ1 estimated for [Gd(L1)(OH)2]Cl3. Such behaviour
is consistent with the presence of a highly hydrophobic envi-
The proton relaxivity of [GdACHTNUGTRENNUG AHCTUNGTREN(NGUN OH2)]Cl3 was
(L1a)
6.5 mmꢂ1 sꢂ1 (20 MHz, 298 K) and decreased as a function of
temperature, (Figure 9), consistent with relatively fast water
exchange at the Gd3+ centre (tm =3.5 ms). This behaviour is
ACHTUNGTRENNUNG
ronment around the coordinated water in [Gd(L8)(OH2)]3+
salts, that suppresses the water interchange process. Evident-
ly there must be more open structure in the
[Gd(L1)(OH)2]3+ complex, with
well-defined second-
a
a
sphere of hydration that assists the water interchange at ga-
dolinium.
The relaxivity of [Gd(L2) (L6a/L6b)
ACHTUGNTRENN(NUG OH2)] and [GdACHUTNRTGENNUNG ACHTUNGTRENNUNG(OH2)]
was independent of pH (ꢁ7%) over the pH range of 2.5 to
11.5 and did not vary significantly over a 24 h time period,
consistent with high kinetic stability with respect to acid-cat-
alysed or base-promoted dissociation pathways. There were
subtle differences in the relaxation parameters for the ester,
[Gd
AHCTUNGTRENNUNG
(L6a) (L6b)-
ACHTUNGTRENUNNG ACHTUNRTGEG(NNUN OH2)] and the derived carboxylate [GdACHTUNGTRENNUNG
(OH2)]ꢂ (Table 6). The relaxivity and tr values of the ester
complex were slightly higher and the water exchange life-
time was slightly lower. These variations are likely to reflect
small changes in the second sphere of hydration. For exam-
ple, in [Gd(L2) (L6a)
ACHTUGNTRENNN(GU OH2)] and [GdACHTUNREGTNNUNG ACHTNUGTERNN(UGN OH2)] the amide NH
can act as a hydrogen-bond donor to a proximate water
molecule. This tendency is likely to be suppressed for [Gd-
ACHUTNGRENNUG CAHTUNGTRENNUGN
(L6b)(OH2)]ꢂ because it can adopt a low-energy conforma-
tion that involves an intramolecular hydrogen bond between
the carboxylate anion and the amide NH proton.
Sensitivity enhancement in spectroscopic and preliminary
19F imaging studies: A comparative study was undertaken to
assess the gain in signal intensity for Tb, Ho and Er com-
plexes of L3 over the diamagnetic Y analogue. Equimolar
solutions were used and spectral signal intensity recorded at
Figure 9. Variation in the relaxivity (r1p) of [Gd(L1
ACTHNUGTRNEG(UN OH2)]Cl3 with a) tem-
perature and b) field showing the estimated contributions of inner- and
outer-sphere waters.
Chem. Eur. J. 2010, 16, 134 – 148
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
143

D. Parker, I. Kuprov et al.
1
Table 6. Summary of selected relaxation parameters (298 K) derived from analysis of variable-temperature H relaxivity, 17O NMR spectroscopy and fit-
ting of NMRD profiles.[a]
[b]
[Gd(L1)
(OH2)]Cl3
[Gd(L8)
(OH2)]I3
[Gd(L2)
E
[Gd
U
G
[GdACHTNGUTRENNUG ACHTUNGTRENNUNG
(L6b)(OH2)]ꢂ
r1p [mmꢂ1 sꢂ1
tm [ms]
]
6.5
3.5
118
13
4.0
30.1
16
2.3
101
103
8
1.5
117
14
5.5
0.90
88
5.4
0.57
4.95
0.93
81
13
6.3
33.7
18
2.0
1
tR [ps][c]
tr [ps]
81
15
6.3
38.9
18
14
D2 ꢄ1019 [sꢂ2
]
5.9
48.0
18
2.0
1
DHm [kJmolꢂ1
]
]
]
DHr [kJmolꢂ1
DHv [kJmolꢂ1
2.0
1
5.0
1
3.0
1
q[d]
DHd [kJmolꢂ1
]
ꢂ22
ꢂ25
ꢂ24
ꢂ25
ꢂ25
q’’
8
0
2
2
2
r’’ [ꢃ]
4.4
n/a
4.4
4.3
4.5
ꢂ
[a] The Gd H distances for the coordinated water molecule (r) and the outer-sphere water molecules (a) were fixed to the standard values of 3.0 and
4.0 ꢃ. The value of 2.24ꢄ10ꢂ5 cm2 sꢂ1 (298 K) was used for the relative diffusion coefficient between solute and solvent molecules. [b] The chloride salt
of [Gd(L8)] gave identical q, relaxivity, tm and tR values.[49] [c] Values of tR derived by analysis of NMRD profiles or r1p/T variations will tend to underes-
[3,13b]
ꢂ
timate tR
because they implicitly analyse the motion of the Gd OH2 moiety, which may not correlate well with the overall tumbling motion of the
complex. Values of tR derived from analysis of the relaxation times of 2H/13C-labelled ligands always give higher estimates.[48] [d] The values of q were
confirmed by analysis of emission lifetime data[53] for analogous Tb complexes, e.g., [Tb(L1)]Cl3, kH O =0.57, kD O =0.31, q=1.0 (ꢁ0.1).
2
2
9.4 T with an acquisition time of three times the measured
T1 value, and spectral data was acquired for 20 min in each
case. The values obtained (Table 7) reveal the expected
gains in signal intensity, which are of the order of a factor of
10.
Table 7. Assessment of 19F NMR spectroscopic sensitivity.[a]
Acquisition
time[b] [s]
Number of
transients
Relative signal
intensity[c,d]
[Y(L2)]
2.346
0.034
0.029
0.075
496
32688
38464
15536
1
25
7
Figure 11. Gradient echo 19F images of a) [Ho(L3)] and b) [Y(L3)] sam-
ples (4 mm). Samples were imaged by using an Ernst angle excitation of a
2 mm thick slice, in-plane resolution of 0.5ꢄ0.7 mm, TE =1.56 ms, TR =
20 ms and 768 transients.
[Tb(L2)]
[Ho(L2)]
[Er(L2)]
10
[a] For 2 mm samples at 188 MHz and 293 K. [b] The acquisition time
was set as 3ꢄT1, with no delay time; in spectral processing the line
broadening was set as 50% of the observed linewidth in each case.
[c] Relative signal intensity was scaled in proportion to the mole fraction
of the major isomer. [d] With no line-broadening function applied, the
order of signal intensity (as above) was 1:44:33:29.
It is likely that the full theoretical enhancement was not
observed in these imaging experiments due to several fac-
tors. At low TR/T1 ratios, a steady-state transverse magneti-
sation can be established which contributes constructively to
image intensity particularly for [Y(L3)], hence yielding a
greater SNR than expected. Loss of differential sensitivity is
expected for [Ho(L3)] due to greater R2 decay during the
echo. Phase evolution of the signal arising from the second
most abundant isomer of [Ho(L3)] during the echo time
(2210 Hz offset) led to partial destructive cancellation of the
imaging signal, reducing the observed signal to 0.83 of the
expected strength that could be recovered through optimisa-
tion of TE for in-phase imaging of both isomers. The experi-
mental measured sensitivity enhancement is then more relia-
bly estimated as 2.9. Finally, the theoretical enhancement is
calculated based on uniform Ernst angle excitation across
the entire sample; the small surface coil used for these stud-
ies provided high basic imaging sensitivity but a non-uni-
form flip angle that may differentially affect sensitivity.
A preliminary imaging study was carried out at 7 T. R1
values were estimated to be 77.8 sꢂ1 and 0.8 sꢂ1 for the Ho
and Y complexes of L3, which leads to Ernst angles[54] of 788
and 138, respectively, for a 20 ms repetition rate, TR. Fluo-
rine MR images of equimolar solutions of the samples were
collected (Figure 11) with 768 averages and a total data col-
lection time of 16.5 min. The signal-to-noise ratio in the Ho
complex was 10.7 compared with 4.4 in the yttrium complex,
which gives a sensitivity improvement of 2.4 compared with
the theoretical sensitivity difference of 5.7 for Ernst angle
imaging under these conditions. Although the minimum rep-
etition time could be reduced further, the theoretical sensi-
tivity increase in moving to an even shorter TR is small.
Therefore, scans were collected with a TR of 20 ms as a typi-
cal value that could be used for multi-slice measurements in
vivo.
144
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 134 – 148

Paramagnetic Fluorine-Labelled Lanthanide Complexes
FULL PAPER
Conclusions
pKa of 7.1. Such behaviour augurs well for future chemical-
shift imaging studies in which the observed parameter is in-
dependent of probe concentration. Finally, and by analogy
with Gd contrast agent research, several paramagnetic fluo-
rinated complexes could be conjugated to a biocompatible
vector, the nature of which determines conjugate biodistri-
bution and clearance rate. Examples based on L3b are indi-
cative.
The creation of paramagnetic fluorinated probes based on
kinetically robust macrocyclic complexes of lanthanideACTHNUTRGNEU(GN III)
ions allows increases in sensitivity to be gained in both MRS
and MRI experiments. Placing one (or more) trifluorometh-
yl groups between 5 and 7 ꢃ from the paramagnetic centre
in a complex gives rise to one major fluorine resonance
signal associated with formation of a preferred isomer in so-
lution. This means that longitudinal relaxation rate enhance-
ments (versus diamagnetic analogues) of the order of 30 to
300 can be gained over a field range of 3 to 9.4 T by using
Tm, Er, Ho and Tb complexes. This allows faster acquisition
of data in a given time period and leads to an order of mag-
nitude enhancement in signal intensity, for spectroscopy at
least.
Each of these aspects suggest a promising future for 19F
magnetic resonance studies with such paramagnetic com-
plexes, by using the switchable 1H/19F probes that can be
procured for current magnetic resonance instrumentation.
Experimental Section
The first detailed analysis of the interplay between the ap-
plied field, the selection of LnIII ion and the rotational corre-
lation time allows predictions to be made about the design
and selection of such probes. For the cases examined herein,
complexes of Ho, Tb and Dy are likely to give rise to large
increases in R1 at lower field (<4.7 T) without incurring too
great a loss in sensitivity through concomitant increases in
R2 and the observed linewidth. At fields of 7 T and above,
complexes of Tm and Er may be more useful, to avoid the
steeper increase in R2 associated with the quadratic depend-
ence on meff (Curie broadening).
Details of instrumentation, experimental procedures, analytical data and
NMR spectral analyses are given in the Supporting Information.
DFT calculations: The DFT calculations were performed by using the
Gaussian 03[55] package for La3+ and Y3+ complexes; the structures of
complexes of other lanthanides, including those of Y3+, are likely to be
nearly identical. Importantly, however, the use of diamagnetic La3+ and
Y3+ for DFT calculations avoids a host of largely unresolved theoretical
issues with spin-orbit coupling and zero-field splitting in open-shell lan-
thanides. The Y3+ results were found to be very similar to those obtained
with La3+, and given the similarity (ꢁ0.02 ꢃ) of the ionic radii of Ho3+
,
Y3+ and Er3+, only Y3+ complex computations are described herein.
Gaussian 03[55] logs and checkpoints are available from IK upon request.
By permuting the lanthanide ion, 1H/19F dual imaging
studies may be undertaken by using complexes of GdIII for
the proton MRI work. Each complex will possess identical
biodistributions in vivo, allowing the possibility of measuring
local probe concentration in 19F studies. This may be partic-
Molecular geometries were optimised in vacuo by using spin-restricted
B3LYP exchange-correlation functional with a compound basis set (cc-
pVDZ for CHNOFS, Stuttgart ECP28MWB for Ln and WGBS for Y).
Saddle points were located by using QST2 and QST3 methods. Hessians
were computed and intrinsic reaction coordinates traced in both direc-
tions to ensure that the located saddle points were first-order saddles cor-
responding to the process under consideration.
19F NMR spectroscopic relaxation analysis: 19F NMR spectroscopic longi-
tudinal relaxation times were measured in dilute solutions in D2O (typi-
cally 1 mm) at 295 K by using the inversion–recovery technique, without
proton decoupling, by using Varian spectrometers operating at magnetic
inductions corresponding to fluorine frequencies of 188, 376, 470 and
658 MHz (Mercury-200, Mercury-400, Inova-500, VNMRS-700), with
chemical shifts reported relative to fluorotrichloromethane. The resulting
free induction decays were subjected to backward linear prediction, opti-
mal exponential weighting, zero filling, Fourier transform, phasing and
baseline correction (by polynomial fitting to signal-free spectrum areas).
The signals were integrated by using Lorentzian line fitting. Inversion–re-
covery type function was fitted to the resulting data by using Levenberg–
Marquardt minimisation of the non-linear least squares error functional.
1
ularly useful with responsive H MRI probes, in which mod-
ulation of the relaxivity of the GdIII complex is a function of
both probe concentration and a given parameter, such as
pH, pM or pX.
An advantage of 19F magnetic resonance studies is the ex-
quisite sensitivity of the 19F shift of the reporter group to its
local chemical environment. This sensitivity is enhanced in
the paramagnetic complexes reported herein. By careful
consideration of the location of the CF3 reporter group, that
is, its polar coordinates, it is possible to devise systems in
which the shift non-equivalence of conformers of the same
complex is over d=50 ppm. A chemical transformation may
alter the relative population of such isomers and allow the
change to be monitored by studying the variation in the
ratio of the two isomers. The differing behaviour of the
ortho-substituted ester/carboxylate and ortho-cyano com-
Proton relaxometric studies: The water proton 1/T1 longitudinal relaxa-
tion rates (20 MHz, 258C) were measured by using a Stelar Spinmaster
Spectrometer (Mede, Pv, Italy) with aqueous solutions of the complexes
(0.5–2 mm). The choice of this lower field allows better comparison with
earlier literature data. For the T1 determinations, the standard inversion–
recovery method was used with a typical 908 pulse width of 3.5 ms, 16 ex-
periments of 4 scans. The reproducibility of the T1 data was estimated to
be ꢁ1%. The temperature was controlled by using a Stelar VTC-91 air-
flow heater equipped with a copper–constantan thermocouple (uncer-
tainty of 0.1ꢁ8C). The proton 1/T1 NMRD profiles were measured by
using a fast field-cycling Stelar Spinmaster FFC relaxometer over a con-
tinuum of magnetic field strengths from 0.00024 to 0.5 T (corresponding
to 0.01–20 MHz proton Larmor frequencies). The relaxometer operates
under computer control with an absolute uncertainty in 1/T1 of ꢁ1%.
Data points from 0.47 (20 MHz) to 1.7 T (70 MHz) were collected by
using a Stelar Spinmaster spectrometer operating at variable fields.
plexes [Ln
ple.
ACHTUNGTRENNUNG ACHTUNGTRENNUGN
(L6a/L6b)] and [Ln(L4b)] affords a proof of princi-
In addition, responsive systems that give rise to a modula-
tion of chemical shift of over d=20 ppm have been defined
to, for example, signal pH variation. The behaviour of the
ortho-cyano complexes based on L4b is most attractive in
this respect because the difference in the 19F chemical shift
observed between the two major isomers is a very sensitive
function of pH, varying by d=14 ppm for [HoACTHNUTRGNEUNG
(L4b)] with a
Chem. Eur. J. 2010, 16, 134 – 148
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
145

D. Parker, I. Kuprov et al.
17O NMR spectroscopy: Variable-temperature 17O NMR spectra were re-
corded by using a JEOL ECP-400 (9.4 T) spectrometer equipped with a
5 mm probe and a standard temperature control unit. Aqueous solutions
of the complexes (ꢄ20–30 mm) containing 2.8% of the 17O isotope
(Cambridge Isotope) were used. The observed transverse relaxation rates
subsequently left to boil under reflux for 3 h. Once cooled, the solvent
was removed under reduced pressure and the complex purified by alumi-
na chromatography (eluent: gradient, (5% MeOH/CH2Cl2!20%
MeOH/CH2Cl2) (0.07 g, 77%). 19F NMR (376 MHz, D2O): d=ꢂ62.8 ppm
(CF3); MS (ESI): m/z: 701.3 [MꢂH]ꢂ, 725.2 [M+Na]+, 741.2 [M+K]+;
HRMS (ESI): m/z calcd for C23H28O7N5F3155Gd: 698.1172; found:
698.1169.
(R2obs) were calculated from the signal width at half height (Dn1/2): Ro2bs
pꢄDn1/2
=
.
The following complexes were prepared as described for [L2Gd]:
19F MRI: Fluorine-19 imaging data were collected by using a 7 T Varian
Unity Inova micro-imaging system (Varian Inc., Palo Alto, California,
USA) equipped with broadband capability, actively shielded gradients
(180 ms rise time to 400 mT/m) and a purpose-built two-turn circular 19F
surface coil (i.d. 12 mm, 281 MHz 19F frequency), which was used for
both excitation and reception of the signal. Aqueous samples of the
[Y(L3)] and [Er(L3)] complexes (200ml, 4 mm) were prepared, placed in
5 mm NMR tubes and positioned on the axis of the coil. Conventional
proton MRI was used to localise and shim the samples based on the
proton water signal. Following 19F pulse calibration, R1 values were deter-
mined for each sample by using a saturation recovery sequence (35 ms
pulse, 22 experiments, 128 scans) to calculate Ernst angle pulses for opti-
mal imaging sensitivity. Fluorine MR images were then collected by
using a RF spoiled, gradient echo imaging sequence with a repetition
time (TR) of 20 ms, an echo time of (TE) 1.56 ms, a 0.5 ms sinc pulse for
selection of a 6 mm thick slice, a 48ꢄ48 mm field of view, a 32ꢄ32
matrix and a 50 kHz receiver bandwidth. Relative imaging sensitivity for
the two complexes was determined by region-of-interest analysis of the
signal-to-noise ratio in each image.
[Ho(L2)]: 1H NMR (200 MHz, D2O, pD =5.5) partial: d=ꢂ93.5, ꢂ79.3,
ꢂ69.1, ꢂ58.23, ꢂ47.4, ꢂ36.0, ꢂ32.1, ꢂ12.4, ꢂ1.3, ꢂ1.1, 4.6, 7.6, 11.0, 12.9,
13.9, 14.9, 18.7, 23.0, 28.9, 44.2, 55.3, 86.3, 92.5 ppm; 19F NMR (376 MHz,
D2O): d=ꢂ64.2 (CF3, major species), ꢂ62.7, ꢂ62.1, ꢂ60.5, ꢂ57.9, ꢂ56.2,
ꢂ54.2, ꢂ50.8 ppm (CF3, minor species); MS (ESI): m/z: 732.1 [M+Na]+,
708.2 [MꢂH]ꢂ; HRMS (ESI); m/z calcd for C23H28O7N5F3165Ho:
708.1250; found: 708.1250.
[Tm(L2)]: 1H NMR (500 MHz, D2O, pD =5.5) partial: d=ꢂ246.3,
ꢂ215.5, ꢂ208.7, ꢂ150.7, ꢂ142.6, ꢂ119.6, ꢂ111.4, ꢂ92.6, ꢂ80.5, ꢂ71.9,
ꢂ25.3, ꢂ18.3, 18.9, 24.6, 35.9, 37.6, 40.0, 45.7, 49.3, 64.4, 79.3, 320.6, 325.7,
340.2, 380.0 ppm; 19F NMR (188 MHz, D2O): d=ꢂ75.9 (CF3, major spe-
cies), ꢂ88.1, ꢂ87.3, ꢂ80.77, ꢂ77.4, ꢂ66.8, ꢂ49.5 ppm (CF3, minor spe-
cies); MS (ESI): m/z: 736.2 [M+Na]+, 712.2 [MꢂH]ꢂ; HRMS (ESI): m/z
calcd for C23H28O7N5F3169Tm: 712.1289; found: 708.1286.
[Er(L2)]: 1H NMR (500 MHz, D2O, pD =5.5) partial: d=ꢂ101.7, ꢂ89.9,
ꢂ84.4, ꢂ80.4, ꢂ71.4, ꢂ66.5, ꢂ57.6, ꢂ56.0, ꢂ38.1, ꢂ32.1, ꢂ29.5, ꢂ8.0,
ꢂ2.9, 12.8, 14.6, 15.9, 18.7, 21.9, 33.4, 128.9, 136.4, 145.0, 153.9,
164.0 ppm; 19F NMR (188 MHz, D2O): d=ꢂ64.8 (CF3, major species),
ꢂ76.2, ꢂ70.4, ꢂ69.2, ꢂ62.7, ꢂ61.8, ꢂ58.3, ꢂ54.8 ppm (CF3, minor spe-
cies); MS (ESI): m/z: 709.3 [MꢂH]ꢂ, 751.2 [M+K]+; HRMS (ESI): m/z
calcd for C23H28O7N5166ErF3: 709.1249; found: 709.1240.
Preparation of L2 and its lanthanide
ACHTUNGTREN(NGNU III) complexes
ACHTUNGTRENNUNG{4,7-Bis(tert-butoxycarbonylmethyl)-10-[(2-trifluoromethylphenylcarba-
moyl)methyl]-1,4,7,10-tetraazacyclododec-1-yl}acetic acid tert-butyl ester:
2-Chloro-N-(2-trifluoromethylphenyl)acetamide (0.167 g, 0.91 mmol) was
[Tb(L2)]: 19F NMR (188 MHz, D2O): d=ꢂ51.9 (CF3, major species),
ꢂ60.2, ꢂ44.9, ꢂ39.8, ꢂ36.7 ppm (CF3, minor species); MS (ESI): m/z:
742.1 [M+Na]+, 702.2 [MꢂH]ꢂ; HRMS (ESI): m/z calcd for
added to
a stirred solution of 1,4,7-tris(tert-butoxycarbonylmethyl)-
1,4,7,10-tetraazacyclododecane (0.30 g, 0.58 mmol), KI (ꢄ10 mg, as cata-
lyst) and K2CO3 (0.80 g, 0.58 mmol) in anhydrous CH3CN (20 mL) at
858C under argon. The mixture was left to boil under reflux for 15 h, and
gave a pale orange solution and a white precipitate. The precipitate was
removed by filtration and the residue washed with CH2Cl2 (2ꢄ30 mL).
The solvent was removed under reduced pressure and the resulting oil
was purified by silica gel column chromatography (eluent: gradient,
100% CH2Cl2 !5% CH3OH/CH2Cl2) to give a pale brown oil (0.25 g,
C23H28O7N5F3159Tb: 702.1198; found: 702.1189; t
2.78 ms; qTb =0.75 (ꢁ0.1).
ACHUTGTNRNE(NUG H2O) 1.76 ms, ttACHTUNGTRENNUNG(D2O)
[Y(L2)]: 19F NMR (188 MHz, D2O): d=ꢂ62.0 ppm; MS (ESI): m/z: 656.2
[M+Na]+, 632.2
[MꢂH]ꢂ;
HRMS
(ESI):
m/z
calcd
for
C23H29F3N5O7Na89Y: 656.0969; found: 656.0970.
1
61%). H NMR (200 MHz, CDCl3): d=1.43, (27H; CH3), 2.02–3.66 (very
br, 24H; CH2 ring and CH2CO), 7.28 (d, J=8.0 Hz, 1H; ArH), 7.43 (d,
J=8.0, 1H; ArH), 7.56 (m, 2H; ArH), 9.97 ppm (s, 1H; NHCO);
13C NMR (125 MHz, CDCl3): d=28.11 (CH3), 49.04 (CH2 ring), 53.68
Acknowledgements
1
(CH2 ring), 55.89 (CH2CO), 56.72 (CH2CO), 82.13 (CCH3), 123.84 (q, J-
2
(C,F)=274 Hz, CF3), 126.13 (q, J
E
N
We thank ESF COST Action D38 and the EC networks of excellence
EMIL and DiMI for support. This work was supported by the EPSRC
(EP/F065205/1, EP/H003789/1) and the NSCCS.
5 Hz, Ar), 129.40 (Ar), 132.35 (Ar), 134.50 (Ar), 135.58 (Ar), 172.29
(CO), 172.73 (CO), 173.16 ppm (CO); 19F NMR (200 MHz, CDCl3): d=
ꢂ60.80 ppm (CF3); MS (ESI): m/z: 716.3 [M+H]+; HRMS (ESI): m/z
calcd for C35H56O7N5F3Na: 738.4024; found: 738.4026.
ACHTUNGTRENNUNG{4,7-Bis(carboxymethyl)-10-[(2-trifluoromethylphenylcarbamoyl)methyl]-
[1] P. L. Choyke, A. J. Dwyer, M. V. Knopp, J. Magn. Reson. Imaging
2003, 17, 509; A. R. Padhani, J. E. Husband, Clin. Radiol. 2001, 56,
607; M. V. Knopp, F. L. Giesel, H. Marcos, H. von-Tengg-Kobligk,
P. L. Choyke, Top. Magn. Reson. Imaging 2001, 12, 301.
[2] J. A. Bergwerff, A. A. Lysova, L. Espinosa-Alonso, I. V. Koptyug,
B. M. Weckhuysen, Chem. Eur. J. 2009, 15, 2363; I. V. Koptyug,
A. V. Khomichev, A. A. Lysova, R. Z. Sagdeev, J. Am. Chem. Soc.
2008, 130, 10452.
[3] S. J. Karlik, R. Bartha, K. Kennedy, R. Chhem, AJR 2007, 189, 105;
F. J. Ruhli, H. von Waldburg, S. Nielles-Vallespin, T. Boni, P. Speier,
JAMA 2007, 298, 2618.
[4] C. Corot, P. Robert, J-M. Idee, M. Port, Adv. Drug Delivery Rev.
2006, 58, 1471; A. K. Gupta, M. Gupta, Biomaterials 2005, 26, 3995;
H-J. Weinmann, W. Ebert, B. Misselwitz, H. Schmitt-Willich, Eur. J.
Radiol. 2003, 46, 33.
1,4,7,10-tetraazacyclododec-1-yl}acetic acid: Trifluoroacetic acid (3 mL)
was added to a solution of {4,7-bis(tert-butoxycarbonylmethyl)-10-[(2-tri-
fluoromethylphenylcarbamoyl)methyl]-1,4,7,10-tetraazacyclododec-1-yl}-
acetic acid tert-butyl ester (0.23 g, 0.32 mmol) in CH2Cl2 (5 mL). The so-
lution was stirred at RT for 2 h. The solvent was removed under reduced
pressure and the resulting solid repeatedly washed with CH2Cl2 (5ꢄ
5 mL) to give the product as a trifluoroacetate salt. The residue was dis-
solved in H2O (5 mL) and left to stir for 2 h with anion exchange resin
(DOWEX 1ꢄ8 200–400 Mesh, pre-treated with 1m HCl) in water to give
the chloride salt. The resin was removed by filtration and the solvent re-
moved under reduced pressure to give a light yellow oil (0.13 g, 74%).
1H NMR (400 MHz, CDCl3): d=3.09–3.92 (very br, 16H; CH2 ring),
7.25–7.94 ppm (brm, 4H; ArH); 19F NMR (200 MHz, CDCl3): d=
ꢂ62.31 ppm (CF3); MS (ES+) m/z: 570.4 [M+Na]+.
[Gd(L2)]:
moyl)methyl]-1,4,7,10-tetraazacyclododec-1-yl}acetic
0.13 mmol) was dissolved in H2O (3 mL) and the pH adjusted to ꢄ6.
[Gd(OAc)3] (0.05 g, 0.14 mmol) was added to the solution, which was
{4,7-Bis(carboxymethyl)-10-[(2-trifluoromethylphenylcarba-
acid (0.07 g,
[5] A. E. Merbach, E. Toth, E. eds. ꢅ The Chemistry of Contrast Agents
in Medical Magnetic Resonance Imaging’, Wiley, New York, 2001; P.
Caravan, J. J. Ellison, T. J. McMurry, R. B. Lauffer, Chem. Rev. 1999,
99, 2293; P. Caravan, Chem. Soc. Rev. 2006, 35, 512; P. S. Tofts, J.
ACHTUNGTRENNUNG
146
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 134 – 148

Paramagnetic Fluorine-Labelled Lanthanide Complexes
FULL PAPER
Magn. Reson. Imaging 1997, 7, 91; S. Aime, M. Botta, E. Terreno,
Adv. Drug Delivery Rev. 2005, 57, 173.
[6] A. R. Padhani, J. Magn. Reson. Imaging 2002, 16, 407; A. R. Padha-
ni, J. E. Husband, Clin. Radiol. 2001, 56, 607.
[20] a) A. Harrison, C. A. Walker, K. Pereira, L. Royle, D. Parker, K.
Pulukkody, T. J. Norman, Magn. Reson. Imaging 1993, 11, 761; b) K.
Pulukkody, T. J. Norman, D. Parker, L. Royle, C. J. Broan, J. Chem.
Soc. Perkin Trans. 2 1993, 605; c) D. Parker, Chem. Soc. Rev. 1990,
19, 271; d) C. J. Broan, J. P. L. Cox, A. S. Craig, R. Kataky, A. Harri-
son, D. Parker, A. M. Randall, G. Ferguson, J. Chem. Soc. Perkin
Trans. 2 1991, 87; e) A. Harrison, C. Walker, J. P. L. Cox, K. J. Jan-
kowski, J. Sansom, M. A. W. Eaton, N. R. A. Beeley, A. T. Millican,
D. Parker, Int. J. Nucl. Med. Biol. 1991, 18, 469.
[7] M. Higuchi, Y. Matsuba, K. Sato, T. C. Saido, Nat. Neurosci. 2005, 8,
527; W. G. Schreiber, B. Eberle, S. Laukemper-Ostendorf, K. Mark-
staller, N. Weiler, A. Scholz, K. Burger, C. P. Heussel, M. Thelen, H-
U. Kauczor, Magn. Reson. Med. 2001, 45, 605; R. C. Samaratunga,
R. G. Pratt, Y. Zhu, R. J. Massoth, S. R. Thomas, Med. Phys. 1999,
21, 697; V. D. Mehta, P. V. Kulkani, R. P. Mason, A. Constantinescu,
S. Aravind, N. Gamer, P. P. Antich, FEBS Lett. 1994, 349, 234; C. J.
Deutsch, J. S. Taylor, Biophys. J. 1989, 55, 799.
[8] a) M. Xia, V. D. Kodibagkar, H. Liu, R. P. Mason, Phys. Med. Biol.
2006, 51, 45; b) R. Weissleder, M. J. Pittet, Nature 2008, 452, 580;
c) J. M. Janjic, M. Srinivas, D. K. K. Kadayakkara, D. T. Ahrens, J.
Am. Chem. Soc. 2008, 130, 2832; d) J. W. Bulte, Nat. Biotechnol.
2005, 23, 945.
[9] a) E. T. Ahrens, R. Flores, H. Y. Xu, P. A. Morel, Nat. Biotechnol.
2005, 23, 983; b) Y. Yu, V. D. Kodibagkar, W. Cui, R. P. Mason,
Curr. Med. Chem. 2005, 12, 819; c) J. Maki, C. Masuda, S. Morikawa,
M. Morita, T. Innbushi, Y. Matsusue, H. Taguchi, I. Tooyamin, Bio-
materials 2007, 28, 434; d) J. C. Jackson, J. T. Hammill, R. A. Mehl,
J. Am. Chem. Soc. 2007, 129, 1160; e) H. R. Kalbitzer, G. Rohr, E.
Nowak, R. S. Goody, W. Kuhn, H. Zimmerman, NMR Biomed.
1992, 5, 347; f) J. W. Peng, J. Magn. Reson. 2001, 153, 32; g) H.
Khan, I. Kuprov, T. D. Craggs, P. J. Hore, S. E. Jackson, J. Am.
Chem. Soc. 2006, 128, 10729.
[10] a) L. D. Stegman, A. Rehemtulla, B. Beattie, A. Kievit, T. S. Law-
rence, R. G. Blasberg, J. G. Tjuvajev, B. D. Ross, Proc. Natl. Acad.
Sci. USA 1999, 96, 9821; b) W. E. Hull, R. E. Port, R. Herrmann, B.
Britsch, W. Kunz, Cancer Res. 1988, 48, 1680; c) W. Wolf, M. J. Al-
bright, M. S. Silver, H. Weber, U. Reichardt, R. Sauer, Mag. Reson.
Imaging 1987, 5, 165.
[11] a) N. J. Baldwin, Y. Wang, T. C. Ng, Magn. Reson. Imaging 1996, 14,
275; b) W. Du, A. M. Nystrom, L. Zhang, K. T. Powell, Y. Li, C.
Cheng, S. A. Wickline, K. C. Wooley, Biomacromolecules 2008, 9,
2826; c) C. Flogel, Z. Ding, H. Harding, S. Jander, G. Reichmann, C.
Jacoby, R. Schubert, J. Schrader, J. Circulation 2008, 118, 140;
d) K. C. Partlow, J. Chen, J. A. Brant, A. M. Neubauer, T. E. Meyer-
rose, M. H. Creer, J. A. Nolta, S. D. Caruthers, G. M. Lanza, S. A.
Wickline, FASEB J. 2007, 21, 1647.
[12] a) V. D. Kodibagkar, J. Yu, L. Liu, H. P. Hetherington, R. P. Mason,
Magn. Reson. Imaging 2006, 24, 959; b) S. Mizukami, R. Takikawa,
F. Sugihara, Y. Heri, H. Tochio, M. Waelchli, M. Shirakawa, K. Ki-
kuchi, J. Am. Chem. Soc. 2008, 130, 794; c) M. Oishi, S. Sumitani, Y.
Nagasaki, Bioconjugate Chem. 2007, 18, 1379; d) K. Tanaka, N. Kita-
mura, K. Naka, Y. Chujo, Chem. Commun. 2008, 6176; e) J. X. Yu,
P. Otten, Z. Y. Ma, W. N. Cui, L. Liu, R. P. Mason, Bioconjugate
Chem. 2004, 15, 1334.
[21] a) P. Wedeking, K. Kumar, M. F. Tweedle, Magn. Reson. Imaging
1992, 10, 641; b) M. F. Tweedle, P. Wedeking, K. Kumar, Invest.
Radiol. 1995, 30, 372.
[22] L. Helm, Progr. NMR Spec. 2006, 49, 45–64.
[23] R. Sharp, L. Lohr, J. Miller, Progr. NMR Spec. 2001, 38, 115–158.
[24] J. Kowalewski, L. Nordenski##ld, N. Benetis, P-O. Westlund, Progr.
NMR Spec. 1985, 17, 141–185.
[25] M. Gueron, J. Magn. Reson. 1975, 19, 58–66.
[26] J. Kowalewski, C. Luchinat, T. Nilsson, G. Parigi, J. Phys. Chem. A
2002, 106, 7376–7382.
[27] J. H. Freed, J. Chem. Phys. 1968, 49, 376-&.
[28] R. Kubo, Hyperfine Interact. 1981, 8, 731–738.
[29] A. G. Redfield, IBM J. Res. Dev. 1957, 1, 19–31.
[30] R. K. Wangsness, F. Bloch, Phys. Rev. 1953, 89, 728–739.
[31] M. Goldman, J. Magn. Reson. 2001, 149, 160–187.
[32] E. Meirovitch, Y. E. Shapiro, A. Polimeno, J. H. Freed, J. Phys.
Chem. B 2007, 111, 12865–12875.
[33] E. Meirovitch, Y. E. Shapiro, A. Polimeno, J. H. Freed, J. Phys.
Chem. A 2006, 110, 8366–8396.
[34] E. Meirovitch, Y. E. Shapiro, Z. C. Liang, J. H. Freed, J. Phys.
Chem. B 2003, 107, 9898–9904.
[35] A. Abragam, The principles of nuclear magnetism; Clarendon Press:
Oxford, 1961.
[36] M. Goldman, J. Magn. Reson. 1984, 60, 437–452.
[37] N. Bloembergen, The Journal of Chemical Physics 1957, 27, 572–
573.
[38] J. A. Peters, J. Huskens, D. J. Raber, Progr. NMR Spec. 1996, 28,
283–350.
[39] I. Kuprov, T. D. Craggs, S. E. Jackson, P. J. Hore, J. Am. Chem. Soc.
2007, 129, 9004–9013.
[40] I. Kuprov, N. Wagner-Rundell, P. J. Hore, J. Magn. Reson. 2007, 184,
196–206.
[41] S. L. Grage, U. H. N. Durr, S. Afonin, P. K. Mikhailiuk, I. V. Komar-
ov, A. S. Ulrich, J. Magn. Reson. 2008, 191, 16–23.
[42] C. S. Yannoni, B. P. Dailey, G. P. Ceasar, J. Chem. Phys. 1971, 54,
4020.
[43] a) S. Amin, D. A. Voss, W. D. Horrocks, C. H. Lake, M. R. Churchill,
J. R. Morrow, Inorg. Chem. 1995, 34, 3294; b) S. Aime, A. Barge, M.
Botta, D. Parker, A. S. de Sousa, J. Am. Chem. Soc. 1997, 119, 4767.
[44] R. S. Dickins, J. A. K. Howard, J. M. Moloney, D. Parker, R. D. Pea-
cock, Angew. Chem. 1997, 109, 541; Angew. Chem. Int. Ed. Engl.
1997, 36, 521.
[45] a) R. S. Dickins, J. A. K. Howard, C. L. Maupin, J. M. Moloney, D.
Parker, J. P. Riehl, G. Siligardi, J. A. G. Williams, Chem. Eur. J. 1999,
5, 1095; b) S. Aime, A. Barge, M. Botta, J. A. K. Howard, J. M. Mo-
loney, D. Parker, A. S. de Sousa, M. Woods, J. Am. Chem. Soc. 1999,
121, 5762.
[13] a) P. K. Senanayake, A. M. Kenwright, D. Parker, S. van der Hoorn,
Chem. Commun. 2007, 2923; b) A. M. Kenwright, I. Kuprov, E.
De Luca, D. Parker, S. U. Pandya, P. K. Senanayake, D. G. Smith,
Chem. Commun. 2008, 2514.
[14] D. Parker, P. K. Senanayake, Int. Pat. Appl. WO2008/132433 (PCT/
GB 2008 1001284).
[15] R. H. Carr, J. Hernalsteen, J. Appl. Polym. Sci. 1994, 52, 1015.
[16] H. Lee, R. R. Price, G. E. Holburn, C. L. Partain, M. D. Adams,
W. P. Cacheris, J. Magn. Reson. Imaging 1994, 4, 609.
[17] a) S. Aime, A. Barge, M. Botta, D. Parker, A. S. de Sousa, J. Am.
Chem. Soc. 1997, 119, 4767; b) S. Aime, A. Barge, J. I. Bruce, M.
Botta, J. A. K. Howard, J. M. Moloney, D. Parker, A. S. de Sousa,
M. Woods, J. Am. Chem. Soc. 1999, 121, 5762.
[18] E. Terreno, M. Botta, P. Boniforte, C. Bracco, L. Milone, B. Mandi-
no, F. Uggeri, S. Aime, Chem. Eur. J. 2005, 11, 5531.
[19] a) B. Bleaney, J. Magn. Reson. 1972, 8, 91; b) I. Bertini, C. Luchinat,
G. Parigi, Progr. Nucl. Magn. Reson. Spectrosc. 2002, 40, 249; c) I.
Bertini, I. Capozzi, C. Luchinat, G. Nicastro, Z. C. Xia, J. Phys.
Chem. 1993, 97, 6351.
[46] a) S. Aime, P. L. Anelli, M. Botta, F. Fedeli, M. Grandi, P. Paoli, F.
Uggeri, Inorg. Chem. 1992, 31, 2422; b) S. Aime, M. Botta, D.
Parker, J. A. G. Williams, J. Chem. Soc. Dalton Trans. 1995, 2259.
[47] R. D. Shannon, Acta Crystallog. Sect. A: Cryst. Phys. Diffr. Theor.
Gen. Crystallogr. 1976, 32, 751.
[48] H. Lammers, F. Maton, D. Pubanz, M. W. Van Laren, H. Van Bek-
kum, A. E. Merbach, R. N. Muller, J. A. Peters, Inorg. Chem. 1997,
36, 2527.
[49] A. L. Thompson, D. Parker, D. A. Fulton, J. A. K. Howard, S. U.
Pandya, H. Puschmann, K. Senanayake, P. A. Stenson, A. Badari,
M. Botta, S. Avedano, S. Aime, Dalton Trans. 2006, 5606.
[50] D. Parker, Chem. Soc. Rev. 2004, 33, 156; S. Aime, M. Botta, D.
Parker, J. A. G. Williams, J. Chem. Soc. Dalton Trans. 1995, 2259; S.
Chem. Eur. J. 2010, 16, 134 – 148
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
147

D. Parker, I. Kuprov et al.
Aime, M. Botta, D. Parker, J. A. G. Williams, J. Chem. Soc. Dalton
Trans. 1996, 17.
O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian,
J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E.
Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W.
Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J.
Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C.
Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari,
J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cio-
slowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaro-
mi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng,
A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W.
Chen, M. W. Wong, C. Gonzalez, J. A. Pople, Gaussian, Inc., Wall-
ingford CT, 2004
[51] S. Aime, A. S. Batsanov, M. Botta, J. A. K. Howard, D. Parker, K.
Senanayake, J. A. G. Williams, Inorg. Chem. 1994, 33, 4696.
[52] T. J. Swift, R. J. Connick, J. Chem. Phys. 1962, 37, 307.
[53] A. Beeby, I. M. Clarkson, R. S. Dickins, S. Faulkner, D. Parker, L.
Royle, A. S. de Sousa, J. A. G. Williams, M. Woods, J. Chem. Soc.
Perkin Trans. 2 1999, 493.
[54] R. R. Ernst, W. A. Anderson, Rev. Sci. Instrum. 1966, 37, 93.
[55] Gaussian 03, Revision E.01, M. J. Frisch, G. W. Trucks, H. B. Schle-
gel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgom-
ACHTUNGTRENNUNGery, Jr., T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S.
Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani,
N. Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K.
Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,
Received: August 19, 2009
Published online: December 2, 2009
148
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 134 – 148