Dimer formation of GdDO3A-arylsulfonamide complexes causes loss of pH-dependency of relaxivity

Article abstract of DOI:10.1039/c7dt02985c

Gadolinium(iii) complexes with pH-dependent relaxivities have been proposed as responsive magnetic resonance imaging (MRI) contrast agents (CA) for mapping of pH value in living subjects. The latter is clinically relevant because hypoxia-induced reduction of interstitial pH (acidosis) is a hallmark of tumor progression and resistance against chemotherapy. In order to obtain versatile building blocks for integration of a pH-responsive MRI-CA functionality into larger molecular assemblies, such as bioconjugates, micelles or nanoparticles, we equipped the structural motif GdDO3A-ethylene(arylsulfonic acid) with additional carboxylic acid moieties in the aromatic para-position. Two novel compounds were characterized concerning their pH-dependent relaxivity as well as by ¹⁷O NMR and ¹H NMRD, augmented by determination of luminescence lifetimes of the respective Eu(iii) complexes and structural modelling by density functional theory (TPSSh/LCRECP/6-31G(d)). Unexpected involvement of the peripheral carboxylates into metal ion complexation effected self-assembly of the lanthanide(iii) complexes, resulting in dimeric species comprising two lanthanide ions, two symmetrically bridging ligands, and one coordinated water molecule per Gd(iii) (q = 1). These structures are stable even at low concentrations and in presence of competing anions like phosphate. The pH-sensitive sulfonamide moieties are not involved into Gd(iii) coordination, resulting in virtually constant relaxivities of r₁ = 6.7 mM^-1 s^-1 (298 K, 20 MHz) over the biologically relevant pH range (4 to 9). Since further functionalisation on the peripheral carboxylates would effectively inhibit dimer formation, the compounds are nonetheless suited for the initially envisaged field of application.

Full text of DOI:10.1039/c7dt02985c

Dalton
Transactions
View Article Online
View Journal
PAPER
Dimer formation of GdDO3A-arylsulfonamide
complexes causes loss of pH-dependency of
Cite this: DOI: 10.1039/c7dt02985c
relaxivity†
Anja Wacker,a Fabio Carniato,b Carlos Platas-Iglesias, c David Esteban-Gomez,c
a
Hans-Jürgen Wester,a Lorenzo Tei *b and Johannes Notni
*
Gadolinium(III) complexes with pH-dependent relaxivities have been proposed as responsive magnetic
resonance imaging (MRI) contrast agents (CA) for mapping of pH value in living subjects. The latter is
clinically relevant because hypoxia-induced reduction of interstitial pH (acidosis) is a hallmark of tumor
progression and resistance against chemotherapy. In order to obtain versatile building blocks for inte-
gration of a pH-responsive MRI-CA functionality into larger molecular assemblies, such as bioconjugates,
micelles or nanoparticles, we equipped the structural motif GdDO3A-ethylene(arylsulfonic acid) with
additional carboxylic acid moieties in the aromatic para-position. Two novel compounds were character-
ized concerning their pH-dependent relaxivity as well as by 17O NMR and 1H NMRD, augmented by deter-
mination of luminescence lifetimes of the respective Eu(III) complexes and structural modelling by density
functional theory (TPSSh/LCRECP/6-31G(d)). Unexpected involvement of the peripheral carboxylates into
metal ion complexation effected self-assembly of the lanthanide(III) complexes, resulting in dimeric
species comprising two lanthanide ions, two symmetrically bridging ligands, and one coordinated water
molecule per Gd(III) (q = 1). These structures are stable even at low concentrations and in presence of
competing anions like phosphate. The pH-sensitive sulfonamide moieties are not involved into Gd(III)
coordination, resulting in virtually constant relaxivities of r1 = 6.7 mM−1 s−1 (298 K, 20 MHz) over the bio-
logically relevant pH range (4 to 9). Since further functionalisation on the peripheral carboxylates would
effectively inhibit dimer formation, the compounds are nonetheless suited for the initially envisaged field
of application.
Received 11th August 2017,
Accepted 6th November 2017
DOI: 10.1039/c7dt02985c
rsc.li/dalton
ation, for example, strongly paramagnetic GdIII complexes
which reduce T1 of protons in their vicinity. Moreover, the
Introduction
Magnetic resonance imaging (MRI) is a well-established, portfolio of MRI applications ranges beyond imaging of mor-
powerful tool in non-invasive routine clinical diagnostics. The phology. Noninvasive mapping of certain physiological para-
most important advantages over computed tomography (CT) meters can be realized, for example, of tissue perfusion by
are absence of ionizing radiation and a superior soft tissue means of special pulse sequences (arterial spin labeling, ASL)
contrast.1 Image contrast is derived from substantially or using contrast agents (dynamic contrast enhanced MRI,
different T1 or T2 proton relaxation times in various tissues. DCE-MRI). MR imaging of complex biological functions can be
The contrast can further be enhanced by application of MRI achieved with even more sophisticated approaches, such as
contrast agents (CA), that is, compounds affecting proton relax- blood oxygenation level dependent (BOLD) contrast MRI2
which exploits the natural differences in magnetic suscepti-
bility of diamagnetic low-spin FeII and paramagnetic high-spin
aLehrstuhl für Pharmazeutische Radiochemie, Technische Universität München,
FeII in oxygenized and non-oxygenized haemoglobin,3 respect-
Walther-Meißner-Strasse 3, D-85748 Garching, Germany.
ively. It is primarily used for mapping of brain activity via
E-mail: johannes.notni@tum.de
bDipartimento di Scienze e Innovazione Tecnologica, Università del Piemonte
Orientale “Amedeo Avogadro”, Viale T. Michel 11, I-15121 Alessandria, Italy.
E-mail: lorenzo.tei@uniupo.it
cCentro de Investigacións Científicas Avanzadas (CICA) and Departamento de
Química Fundamental, Facultade de Ciencias, Universidade da Coruña,
E-15071, A Coruña, Galicia, Spain
measurement of neuronal oxygen consumption, a technique
referred to as functional MRI (fMRI).4
From a chemical point of view, BOLD-contrast imaging per-
fectly showcases a powerful strategy in MRI-CA design. It uti-
lizes the FeII–porphyrin complex haem as a naturally abundant
CA whose relaxivity is dependent from, or responds to, an
environmental parameter, namely, oxygen consumption. An
†Electronic supplementary information (ESI) available. See DOI: 10.1039/
c7dt02985c
This journal is © The Royal Society of Chemistry 2017
Dalton Trans.

View Article Online
Paper
Dalton Transactions
entire class of artificial ‘(bio)responsive’,5 ‘activatable’,6 or
Along these lines, we reasoned that DO3A-SA ligands with
‘smart’7 MRI-CA is essentially founded on the same idea. additional functional groups for conjugation should be highly
Many of these approaches are based on GdIII complexes8 useful building blocks, allowing for facile inclusion of a pH-
whose relaxivity is modulated by changes in environmental responsive MRI-CA functionality into larger constructs, such as
pH, oxygen partial pressure,9 enzyme activity,10 or the presence bioconjugates or nanoparticles. We therefore synthesized and
of other metal cations.11 Therefore, they can ‘report’ changes evaluated two respective ligands, featuring additional car-
in their physical or chemical environment and are thus, in boxylic acid moieties not intended to support metal ion com-
principle, useful for high-resolution 3D-mapping of said para- plexation. Surprisingly, the corresponding LnIII complexes
meters by MRI systems.
exhibited completely different profiles of pH-dependent relax-
On a molecular level, said modulation of relaxivity is ivity compared to those of the parent ligand, pointing at a
usually a result of a parameter-dependent change of structural remarkable influence of the distant functional groups, which
features of the GdIII complex, affecting the mode of interaction we therefore elucidated in detail.
and magnetic coupling between the paramagnetic GdIII ion
and the surrounding water molecules, for example, the
number of GdIII-coordinated ‘inner-sphere’ water molecules q,
their exchange lifetime τM, or the reorientational correlation
Results and discussion
In the following, ligands and complexes are sometimes
denoted without indicating charges, protonation states, and
coordinated water molecules. In these cases, it is referred to
any possible species, owing to the fact that complexes with
unknown or hypothetical composition are to be discussed.
time or ‘tumbling speed’ τR. In terms of pH-sensitive CA, the
first reported example was a GdIII complex of the tetrapho-
sphonate-DOTAM derivative Gd-DOTA-4Amp.12 For this
complex, a variation of pH modulates its relaxivity by changing
the protonation state of the pendant phosphonates which, in
turn, affects τM by altering the exchange rate of relaxed water
protons with protons of the bulk.13 In contrast, pH-depen-
dency of the relaxivity of Gd-DO3A complexes with one
β-sulfonamidoethyl pendant arm (GdDO3A-SA)14–16 is rooted
in a pH-dependent number of GdIII-bound aqua ligands q.
While at acidic pH, the sulfonamide is N-protonated and thus
unable to bind to GdIII, leaving room for more than one aqua
ligand (q ∼ 2) at the metal centre, coordination of the deproto-
nated sulfonamide nitrogen at pH > 7 replaces all Gd-bound
water, resulting in q = 0 and thus, a markedly lower relaxivity
Synthesis of ligands and LnIII-complexes
The ligands DO3A-SA-COOH (5) and DO3A-SA-(CH2)2-COOH (7)
were synthesized following a procedure adapted from the litera-
ture (Scheme 1).23 Briefly, the secondary amine of DO3A(tBu)3
was functionalised with N-Cbz-bromoethylamine in the pres-
ence of potassium carbonate in almost quantitative yield. The
carboxybenzyl protecting group was removed by hydrogenolysis
and the resulting primary amine reacted with either p-chloro-
sulfonyldihydrocinnamic acid or 4-(chlorosulfonyl)benzoic
acid. Removal of the tert-butyl protecting groups yielded the
ligands 5 and 7.
(Δr1 ∼ 5.5 mM−1 −1). Such GdDO3A-SA-like complexes have
s
been modified to bind to HSA17 or encapsulated in lipo-
somes18 aiming to further increase the extent of pH-dependent
variation of relaxivity. The corresponding EuDO3A-SA
complexes have been decorated with photosensitizers, and the
resulting constructs were successfully applied for measure-
ment of changes of intracellular pH, e.g., in lysosomes or the
endoplasmic reticulum.19 Structurally related Ln-DO3A
systems, featuring aminoethyl or diphenylphosphinamide
moieties instead of sulfonamides, showed similar pH-
dependent relaxivities, although at other pH values owing to
different pKa values of the protonated donor sites.20–22
Since complexation kinetics were reported to be slow,16
LnIII complexation was done in water at pH 6.5 and r.t. for
three days. Excess LnIII was removed by precipitation at pH 10,
followed by centrifugation and recovery of the soluble
complexes.
pH-Dependent modulation of relaxivities
The GdIII complexes of 5 and 7 showed the same relaxivity, r1 =
6.7 mM−1 s−1 at 298 K, 20 MHz and pH 7.0, which is 60% and
12% higher than that of GdDOTA and GdDO3A (r1 = 4.2 and
Scheme 1 Synthesis of ethylene(arylsulfonamide)-pendant DO3A ligands with peripheral carboxylic acids for conjugation.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2017

View Article Online
Dalton Transactions
Paper
6.0 mM−1 s−1, respectively, measured at the same experimental corresponding typical r1 values around 2. Apart from that, the
conditions).24 pH-Dependency of relaxivities was determined observed increase of r1 below pH 4 is fully in line with expec-
for dilute aqueous solutions over the range of pH 2 to 12. We tations, caused by protonation of the primary carboxylate
expected that owing to a similar mechanism of deprotonation donors and subsequent dissociation of the complex.27
and GdIII coordination of the sulfonamide associated with
Based on these observations, we assume that the Gd-5 and
higher pH, a relaxivity shift should be observed for Gd-5 and Gd-7 species feature a GdIII centre coordinated to four carboxy-
Gd-7 around neutral pH, just like for the GdIII complex of the late donors, including the carboxylates of the aromatic p-sub-
parent ligand DO3A-SA. In addition, we reasoned that for Gd-5 stituents, and a distant sulfonamide which replaces the peri-
and Gd-7, the pH value could furthermore affect the water pheral carboxylate upon deprotonation at pH > 10. While this
exchange rate τM and the rotational correlation time τR due to rationale explains the observed phenomena in terms of relaxiv-
a change in protonation state and hydrogen bonding pro- ities, the structure of the postulated species remains unclear.
perties of the distant carboxylate, resulting in an additional, Considering the steric situation, it appeared unlikely that
minor pH-dependent modulation of r1.
these complexes could be monomers, pointing at presence of
However, we were surprised that the relaxivities of Gd-5 and LmMn aggregates in solution. It is worth to note that the depen-
Gd-7 remained virtually constant between pH 4 and 10 (see dence of R1 with the concentration of Gd-7 (from 1.7 to
Fig. 1), particularly because it was confirmed earlier that the 0.2 mM, see ESI, Fig. S1†) is linear, highlighting that these
sulfonamides in arylsulfonamide-based GdDO3A-SA systems aggregates are formed already at very low concentrations.
invariantly exhibit pKa values around 7, while being influenced Finally, they are also stable in presence of phosphate buffer at
only moderately by aromatic p-substituents.14,15 Typically, an 10 mM concentration, which does not influence the measured
off-ligation of the sulfonamide pendant arm at acidic pH r1 value.
results in a bis-hydrated GdDO3A-like complex (q ∼ 2) with r1
values above 7–8 mM−1 s−1, while at pH > 7, deprotonation of the
Luminescence measurements
sulfonamide and subsequent coordination to the metal centre
yields r1 values characteristic for q = 0 (2–2.5 mM−1 s−1).15
Instead, Gd-5 and Gd-7 showed a peculiar behavior (Fig. 1),
indicating that the sulfonamide might not be coordinated to
the metal center between pH 4 and 9.
Furthermore, a ∼50% decrease in r1 between pH 10 and 12
was observed. Such behaviour is usually not displayed by
responsive CAs based on sulfonamides. In this context, it has
been reported earlier that the pKa values of arylsulfonamides
range between 10 and 12 unless located in close proximity to a
strongly lewis-acidic metal centre like ZnII, whereupon pKa is
lowered to values around 7.25,26 Hence, in case the sulfon-
amides in Gd-5 and Gd-7 are kept at distance of the GdIII
centre, their pKa should indeed remain around 10. Once
deprotonated, GdIII coordination ultimately occurs, resulting
in the expected monomeric complexes with q = 0 and the
In order to determine the hydration number and obtain
further insight into the metal coordination environment, the
emission spectra and lifetimes of 1 mM H2O and D2O solu-
tions of [EuDO3A-SA-(CH2)2-COO]− (Eu-7) were recorded.
A sizeable emission was observed by direct excitation of the
7
5L6 → F0 transition of EuIII at 395 nm.28 The emission spec-
trum displays the typical 5D0 → 7FJ transitions of EuIII (J = 0–4,
see Fig. 2). The shape of the 5D0
→
7FJ emission bands is
similar to those reported for the complex of DOTA and related
systems.29 We however noticed a higher ΔJ = 2/ΔJ = 1 intensity
ratio with respect to DOTA derivatives showing C4 symmetry,
which can be attributed to a lower degree of symmetry of the
EuIII coordination environment in Eu-7. The 5D0
→
7F1 tran-
sition presents three components with splittings of 126 and
223 cm−1 with respect to the highest energy component,
Fig. 1 pH dependent relaxivities (r1) of Gd-5 (circles) and Gd-7
Fig. 2 Emission spectrum of the Eu-7 complex, recorded for a 1 mM
(squares). Data were recorded at 21 MHz and 298 K.
solution under excitation at 395 nm (pH = 7.2).
This journal is © The Royal Society of Chemistry 2017
Dalton Trans.

View Article Online
Paper
Dalton Transactions
caused either by low local symmetry of the EuIII site or by pres-
ence of two isomers in solution (i.e., SAP and TSAP conformers
of DOTA-like complexes).30
Lifetimes of the 5D0 excited states of EuIII were measured in
solutions of the complexes in H2O and D2O to assess the
hydration number of the complex. The determined values,
τH O = 0.54 ms and τD O = 2.02 ms, correspond to hydration
2
2
numbers (q) of 1.3 or 1.2, calculated according to the methods
of Parker31 and Horrocks,32 respectively. These values indicate
presence of a single water molecule in the first coordination
sphere of EuIII, which is in line with the observed relaxivities
as well as 1H NMRD and 17O NMR data (see below). Further
evidence is given by an intermediate emission lifetime
recorded in H2O (0.54 ms), being longer than those reported
for pH-responsive DO3A-sulfonamides at low pH (typically
0.42–0.45 ms, corresponding to q = 2) and shorter than those
measured at high pH (typically 0.7–0.9 ms, corresponding to
q = 0).14,15 The emission lifetime recorded in H2O solution
remains constant (0.54 ms) in the concentration range 1.6 ×
10−5–1.0 × 10−3 M, suggesting that the aggregates are stable in
this concentration range.
Fig. 3 1H NMRD profiles of Gd-5 (solid circles) and Gd-7 (open tri-
angles). Data were recorded at 298 K and pH 6.8 and 6.7, respectively.
The solid lines represent the best fitting results of the experimental data
points with the parameters in Table 1.
Relaxometric measurements
Frequency-dependent relaxivities, that is, Nuclear Magnetic
Relaxation Dispersion (NMRD) profiles, of Gd-5 and Gd-7 were
determined for proton larmor frequencies ranging from 0.01
to 70 MHz (B = 2.34 × 10−4 − 1.64 T). These provide additional
structural information because amplitude and shape of NMRD
profiles are determined by contributions of inner-sphere (IS) and
outer-sphere (OS) water to r1. The IS term is determined by q, the
water exchange rate kex = 1/τM, τR, and the electronic relaxation
times (T1,2e) of Gd3+. The OS term depends on T1,2e, the relative
diffusion coefficient between the complex and the water
molecules (D), and their distance of closest approach, a. Since
relaxivities of Gd-5 and Gd-7 are pH-independent between pH 4
and 10 (see above), NMRD profiles were measured for 298 K and
neutral pH (Fig. 3). Almost identical datasets were obtained,
resembling those of typical monohydrated low molecular weight
CAs.33 A plateau region at low field strength is followed by a
single dispersion at approx. 4–8 MHz.
More accurate and quantitative information on the kinetics
of water exchange were obtained by measuring the temperature
dependency of the solvent 17O NMR reduced transverse relax-
ation rates ln(1/T2r) and shifts Δωr at 11.75 T for a 21.4 mM
solution of Gd-7 at neutral pH (Fig. 4). Temperature depen-
dency of 1/T2r data provides evidence for a fast water-exchange
regime over the entire temperature range, while that of the
reduced chemical shifts (Δωr) is in line with 1/T2r data. An
increase of 1/T2r with decreasing temperature indicates a faster
water exchange rate (1/τM) compared to Gd-DTPA or Gd-
DOTA.34 While the respective curves for these complexes show
a maximum at 315 K, the maximum for Gd-7 is observed at
280 K or below. Further evidence for fast IS water exchange
dynamics is obtained from the NMRD profiles of Gd-7 at three
different temperatures (Fig. 5). For any given proton Larmor
frequency, r1 values consistently decrease with rising tempera-
Fig. 4 Reduced transverse 17O relaxation rates (above) and chemical
shifts (below) measured at 11.74 T (pH 6.6) for Gd-7. The solid lines
correspond to the fits of the data as described in the text.
ture. In particular, r1 at 20 MHz and neutral pH decreases
from 9.0 mM−1 s−1 to 5.0 mM−1 s−1 at 283 K and 310 K,
respectively. This behaviour is characteristic for small chelates,
bearing a coordinated water molecule in fast exchange with
the bulk, and a τM short enough not to limit the relaxivity.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2017

View Article Online
Dalton Transactions
Paper
Table 1 Best-fit parameters obtained from the analysis of the 1/T1 1H
NMRD profiles collected at 298 K and 17O NMR data
Parameter
Gd-7
40
42.3 1.2
Gd-5
Gd2Pip(DO3A)2 GdDOTA
298τM/ns
8
40
—
—
667
244
ΔH ]M/kJ mol−1
34.2
−3.7
49.8
−3.8
77
A/ħ/106 rad s−1 −3.1 0.1
298τR/ps
145.6 2.2
44.5 2.4
1.53 0.11 1.71 0.10 1.7
145.1 2.0 171
43.2 2.2
298τV/ps
19
11
1.6
Δ2/1019 s−2
The outer sphere contribution to relaxivity was estimated by using
standard values for the distance of closest approach, a (4 Å).
linked by a piperidine heterocycle.34 The monomer GdDOTA
and the dimeric complex exhibit τR of 77 ps and 171 ps,
respectively, the latter value being more similar to that of Gd-5
and Gd-7, pointing at a dimeric structure for these species.
The higher τR value is furthermore in good accordance with a
broad hump of r1 between 20 and 70 MHz shown by the
NMRD profile of Gd-7 at 283 K, pointing at low rotational
dynamics of this species. From a fit to this NMRD profile, a τR
value of 210 ps is obtained for T = 283 K, supporting the
notion of presence a dimer. In order to further confirm the
correctness of the analysis, we performed the fitting also using
q = 2 and a best-fit was obtained only using a τR value of 68 ps,
lower than the value obtained for GdDOTA (see ESI, Fig. S2†).
Fig. 5 1H NMRD profiles of Gd-7 at 283 K (open triangles), 298 K (solid
circles) and 310 K (open diamonds) (pH 6.66).
The experimental data, 1H NMRD and 17O NMR, were fitted
simultaneously, according to the established theory of para-
magnetic relaxation, expressed by the Solomon–Bloembergen–
Morgan (SBM)35 and Freed’s36 equations for the inner- and
outer-sphere proton relaxation mechanisms, respectively, and
of the Swift–Connick theory for 17O relaxation.37 The 17O R2
Hypothetical structures
data depend on a series of parameters such as T1,2e, the hyper- In summary, the relaxometric data strongly suggest that at
fine Gd–17Owater coupling constant AO/ħ, τM, and q. neutral pH, Gd-5 and Gd-7 are actually forming dimeric struc-
Information on q and AO/ħ are derived from the temperature tures, [Gd2(5)2] and [Gd2(7)2]. As indicated above, the observed pH
dependence of Δω. Additional relevant parameters are those dependency profile of r1 allows for the conclusion that between
associated with the electronic relaxation times T1,2e, that is, pH 4 and 10, dimerisation occurs via coordination of the peri-
the trace of the square of the zero-field splitting tensor, Δ2; the pheral carboxylate of one complex monomer with the metal
correlation time describing the modulation of the zero-field centre of another one, while the sulfonamide is not involved in
splitting, τV, and its activation energy, EV; the enthalpy of acti- GdIII coordination. These species are apparently of remarkable
vation for the water exchange process, ΔH ]M. The data fitting stability because they are invariantly present under conditions
was optimized by using fixed values for some of the relaxation typically promoting formation of monomers, that is, at low GdIII
parameters: q = 1, a = 4.0 Å, 298D = 2.24 × 10−5 cm2 s−1, the dis- concentration (<0.5 mM) or in 10 mM phosphate buffer.
tance between the metal ion and the protons of the bound
As a further argument in favour of dimerization, a linear
water molecule, rGdH = 3.1 Å.
graph of τR versus molecular weight comparing [Gd2(5)2] and
A good fit was obtained and the resulting structural and [Gd2(7)2] with other examples of monomeric and dimeric Gd-
dynamic molecular parameters describing the magnetic inter- complexes is shown in Fig. 6. Our hypothesis is in line with
action of the solvent nuclei and the metal center are listed in previous studies, showing that carboxylate units can mediate
Table 1. In particular, the residence lifetime of the bound the formation of stable trinuclear LaIII entities in aqueous
water molecule was confirmed to be as short as 40 ns (298kex
=
solution40 as well as the aggregation of dinuclear GdIII com-
25 × 106 s−1). It is reasonable to assume a similar value for the plexes with DO3A ligands linked by rigid spacers.41,42 Finally,
other parent chelate, Gd-5. Notably, a reorientational corre- our EuIII luminescence measurements indicated presence of
lation time of 145 ps was calculated for both complexes, which one EuIII-bound water per metal centre in Eu-7, which is also
is much longer than that observed for monomeric GdIII-che- in line with the observed relaxivities and NMRD profiles.
lates with similar molecular weight. In particular, GdDO3A-SA However, it should be noted that GdDO3A-SA derivatives
bearing a methoxyphenyl group, a complex with similar mole- reported by Lowe et al.15 contain three peripheral carboxylate
cular weight to Gd-5 and Gd-7, showed a τR of 94 ps.18 In order groups that could, in principle, also interact with the Gd-
to illustrate the implications of this result, Table 1 also lists centre of another monomeric unit and form aggregates, thus
the relaxometric parameters for GdDOTA and Gd2Pip(DO3A)2, preventing a pH-dependent relaxivity. Actually, in that case,
a dimeric GdIII complex featuring two DOTA-monoamide units aggregation did not occur and pH-dependency was main-
This journal is © The Royal Society of Chemistry 2017
Dalton Trans.

View Article Online
Paper
Dalton Transactions
Fig. 6 Plot of the rotational correlation times of selected polyamino-
carboxylate Gd-complexes as a function of molecular weight. The cor-
relation coefficient R of the best fit line is 0.99. Data taken from ref. 24,
38 and 39 and this work.
Fig. 7 Geometries of the TSAP conformers of [Gd2(5)2(H2O)2] (top) and
[Gd2(7)2(H2O)2] (bottom), obtained from DFT calculations at the TPSSh/
LCRECP/6-31G(d) level. For the corresponding SAP conformers see the
ESI, Fig. S1.†
tained; we presume that the shorter length of the spacer with
respect to the one present in case of Gd-5 and Gd-7 may have
prevented dimerization.
To gain additional insight into the structure of the com-
1
plexes in solution, we recorded the H NMR spectrum of Eu-7
the four nitrogen atoms of the macrocycle, while the upper
plane is delineated by three oxygen atoms of the acetate
groups and one oxygen atom of the carboxylate group of the
sulfonamide arms. The calculated Gd–O distances involving
the carboxylate group of the sulfonamide pendant fall within
the range 2.45–2.52 Å, being somewhat longer than the dis-
tances between the metal ion and the oxygen atoms of acetate
groups (2.36–2.42 Å).
(see ESI, Fig. S3†). Unfortunately, the spectrum shows broad
featureless signals likely related to exchange processes present
in solution. This may be related to the presence of aggregates
in solution and/or to the exchange between SAP and TSAP
isomers. One can notice very broad signals in the range
25–30 ppm that are characteristic of the most shifted signals
of the SAP isomer.
Finally, we recorded the 1H DOSY NMR spectrum (500 MHz,
D2O) of diamagnetic analogues La-5, and compared the results to
LaDTPA (see ESI, Fig. S4†). The spectrum of La-5 shows very
broad signals owing to presence of dimers, similar to obser-
vations made for Eu-7. The different translational diffusion
coefficients (D) of the two compounds, (2.2 0.2) × 10−10 m2 s−1
for La-5 and (3.80 0.65) × 10−10 m2 s−1 for LaDTPA can be
ascribed to the different molecular weight, thus strongly support-
ing the notion of presence of La-5 in a dimeric form.
Altogether, our solution data provide strong evidence for a
composition of [Gd2(L)2(H2O)2] for complexes of both ligands
(L = 5 or 7), while it is most reasonable to assume symmetrical
structures with bridging ligands L and one metal-bound water
per GdIII. Based on this presumption, possible geometries for
both species were elucidated by means of density functional
theory (DFT) calculations.
The relative free energies of the TSAP and SAP isomers favor
the TSAP forms of [Gd2(7)2(H2O)2] by only 0.26 kcal mol−1
,
while for [Gd2(5)2(H2O)2], the two isomers exhibit virtually iden-
tical free energies. Thus, we assume that two diastereoisomers
are present in solution, slightly favoring the TSAP isomer. Since
the latter type of conformer is generally characterized by water
exchange rates that are about one order of magnitude higher
than those of the corresponding SAP forms, the fast exchange
determined by the analysis of the 1H NMRD and 17O NMR data
is well explained, which furthermore corresponds to the coordi-
nation of the carboxylate group of the sulfonamide arm on the
O4 plane of the coordination polyhedron, leaving the
coordinated water molecule at the labile capping position.43
Due to the different length of the aromatic substituents, varying
Gd–Gd distances of 13.8 and 15.3
Å are found for
[Gd2(5)2(H2O)2] and [Gd2(7)2(H2O)2], respectively.
DFT-calculated structures
Geometry optimisations of the [Gd2(L)2(H2O)2] systems pro-
vided energy minima that correspond to the well-known
Conclusion
square-antiprismatic (SAP) and twisted-square antiprismatic In order to obtain building blocks for inclusion of a pH-depen-
(TSAP) isomers of DOTA-like complexes (Fig. 7). As expected, dent MRI contrast agent functionality into larger constructs,
the lower plane of the coordination polyhedron is defined by such as bioconjugates, nanoparticles, or micelles, we syn-
Dalton Trans.
This journal is © The Royal Society of Chemistry 2017

View Article Online
Dalton Transactions
Paper
thesised derivatives of the established pH-responsive CA motif Ligand syntheses
Gd(DO3A-SA), bearing additional carboxylic acid moieties on
the sulfonamide-pendant arm of the DO3A backbone. Our
DO3A(tBu)3-(CH2)2-NH-Cbz (2). DO3A(tBu)3 (1, 1.00 g,
findings illustrate that such a seemingly innocent, peripheral 1.94 mmol) was dissolved in 20 mL dry MeCN and K2CO3
functionalisation of these pH-responsive MRI-CA actually has (806 mg, 5.83 mmol) was added. N-Cbz-Bromoethylamine
a dramatic influence on the complex structure in solution, (501 mg, 1.94 mmol) was dissolved in 5 mL dry MeCN and
essentially resulting in a complete loss of pH-dependency of added dropwise over 1 h. The resulting solution was stirred at
relaxivity in the biologically relevant pH range (4 to 9). Despite room temperature overnight. The crude product was purified
the steric position of the anchor groups was designed not to by HPLC (30–68% B in 10 min, flow rate 10 mL min−1, elution
interfere with the GdIII coordination environment, experi- time = 7.5 min). The collected fractions were lyophilized and
mental evidence supports the hypothesis of dimer formation the product obtained as a white solid as the mono-trifluoroace-
involving the distant carboxylates, thus preventing GdIII- tate salt (1.4 g, 1.74 mmol, 89.6%). Analytical HPLC (20–100%
coordination of the deprotonated sulfonamide at neutral pH. B in 15 min): tR = 9.9 min. MS (ESI+): C36H62N5O8 m/z: calcd
Most notably, these dimers are very stable even at low concen- 692.9 [M + H]+; found 692.8 [M + H]+, 714.7 [M + Na]+.
trations and in presence of coordinating anions such as phos-
DO3A(tBu)3-(CH2)2-NH2 (3). Under nitrogen atmosphere, 2
phate, and exhibit larger relaxivities than comparable (1.40 g, 1.74 mmol) was dissolved in 60 mL EtOH and Pd/C
GdDOTA- or GdDO3A-monomers, owing to larger rotational (140 mg, 10 wt%) was added. The flask was saturated with
correlation times and a favorable ratio of square-antiprismatic hydrogen and the suspension stirred at room temperature
and twisted square-antiprismatic conformers of the GdIII overnight. The crude product was purified over Celite 545
coordination environment. Hence, such ligands, resulting in (particle size 0.01–0.04 mm, preconditioning and elution with
dimeric or multimeric GdIII complexes, might inspire future MeOH). The product was obtained as a colourless, highly
design of high-relaxivity GdIII-based MRI-CA by exploiting for- viscous liquid (0.94 g, 1.68 mmol, 96.5%). Analytical HPLC
mation of supramolecular structures by self-assembly. (10–100% B in 15 min): tR = 6.8 min. MS (ESI+): C28H56N5O6
Notwithstanding this, dimerisation is unlikely to occur after m/z: calcd 558.8 [M + H]+; found 558.8 [M + H]+, 580.7 [M + Na]+;
further functionalisation on the peripheral carboxylates, result- 1H NMR (400 MHz, 298 K, DMSO-d6): δ (ppm) = 1.42–1.45 (m,
ing in non-coordinating or bulky moieties at this position. 27 H), 2.89–3.46 (m, 20 H), 3.5–3.77 (m, 6 H), 8.28 (bs, 2 H).
Hence, our ligands are nonetheless to be considered suitable
DO3A(tBu)3-SA-(CH2)2-COOH (6). 3 (100 mg, 179 μmol) was
for the initially intended purpose, warranting further appli- dissolved in 3 mL dry MeCN. DIPEA (116 mg, 156 μL,
cation in this context.
896 μmol) was added and the solution cooled on ice.
p-Chlorosulfonyldihydrocinnamic acid (45 mg, 179 μmol) was
dissolved in 1 mL dry MeCN and added dropwise over 1 h. The
resulting solution was slowly warmed to room temperature
and stirred overnight. The reaction was quenched by addition
of 2 mL H2O and the product purified by HPLC (25–55% in
10 min, tR = 8.9 min). The product was obtained as a colour-
Experimental
Materials and methods
1H NMR spectra were recorded on a Bruker AV400 at 298 K and less solid (50 mg, 65 μmol, 36.2%). Analytical HPLC (10–100%
a Bruker AV500cr equipped with a cryogenic probe. 1H NMR B in 15 min): tR = 8.2 min. MS (ESI+): C37H64N5O10S m/z: calcd
DOSY spectra were recorded on a Bruker Avance III spectro- 770.0 [M + H]+; found 770.8 [M + H]+, 792.6 [M + Na]+; 1H NMR
meter operating at 11.7 T. HPLC analyses were done using a (400 MHz, 298 K, DMSO-d6): δ (ppm) = 1.35–1.42 (m, 18 H),
3
Nucleosil 100-5C18 (125 × 4.6 mm, CS Chromatographie) 1.44–1.50 (m, 9 H), 2.41 (t, 2 H, JHH = 7.0 Hz), 2.82 (t, 2 H,
column on a Shimadzu HPLC System, equipped with a two- 3JHH = 7.0 Hz), 2.86–3.18 (m, 12 H), 3.32–3.44 (m, 8 H),
3
channel UV-VIS detector. For identification of the compounds 3.45–3.61 (m, 6 H), 7.14 (d, 2 H, JHH = 8.2 Hz), 7.50 (d, 2 H,
absorbances at 214 nm and 254 nm were recorded. Mass 3JHH = 8.2 Hz), 8.09 (t, 1 H, 3JHH = 5.3 Hz).
spectra were recorded on
a
Varian 500-MS IT Mass
DO3A-SA-(CH2)2-COOH (7). tert-Butyl protecting groups of 6
Spectrometer. All synthesised compounds were purified on a (50 mg, 65 μmol) were removed by reaction with 2 mL DCM/
J’Sphere ODS-H80 column (150 × 20 mm, YMC HPLC Column) TFA (1 : 3) overnight. Then, the solvent was removed under a
using either a Sykam S 1121 HPLC System equipped with a stream of nitrogen and the crude product purified by HPLC
DAD detector (Sykam S3210 UV/VIS Detector) or a semipre- (3–20% in 10 min, tR = 4–6 min). The product was obtained as
parative Shimadzu LC-20AT HPLC System, equipped with a a colourless solid (27 mg, 45 μmol, 69.4%). Analytical HPLC
SPD-20A UV-VIS detector. Solvents for analytical and prepara- (3–20% B in 15 min): tR = 5.3 min. MS (ESI+): C25H40N5O10S
tive HPLC runs were H2O with 0.1% trifluoroacetoc acid (TFA) m/z: calcd 602.7 [M + H]+; found 602.6 [M + H]+, 624.4 [M + Na]+.
(A) and MeCN with 0.1% TFA (B). Analytical runs to character-
DO3A(tBu)3-SA-COOH (4). 3 (300 mg, 539 μmol) was dis-
ize the GdIII complexes were done using solvents without the solved in 5 mL dry MeCN. DIPEA (348 mg, 469 μL, 2.69 mmol)
addition of TFA. Chemicals and solvents (HPLC grade) were was added and the solution cooled on ice. 4-(Chlorosulfonyl)
purchased from various commercial suppliers and used as benzoic acid (119 mg, 539 μmol) was dissolved in 2 mL dry
received.
MeCN and added dropwise over 1.5 h. The resulting solution
This journal is © The Royal Society of Chemistry 2017
Dalton Trans.

View Article Online
Paper
Dalton Transactions
was slowly warmed to room temperature and stirred overnight. by addition of 0.1 M NaOH and 0.1 M HCl, respectively. 17O
The reaction was quenched by addition of 2 mL H2O and the NMR measurements were recorded on a Bruker Avance III
product purified by HPLC (30–50% in 10 min, tR = 7.2 min). spectrometer (11.7 T) equipped with a 5 mm probe. The ana-
The product was obtained as a colourless solid (73.5 mg, lysis was carried out using a GdIII complex solution of approx.
99.1 μmol, 18.4%). Analytical HPLC (10–100% B in 15 min): tR = 20 mM concentration, containing 2.0% of the 17O isotope. The
7.9 min. MS (ESI+): C35H60N5O10S m/z: calcd 742.9 [M + H]+; exact GdIII concentrations were determined accordingly to
found 742.9 [M + H]+, 764.6 [M + Na]+; 1H NMR (400 MHz, Corsi et al. from the lanthanide(III)-induced shift of the tert-
298 K, DMSO-d6): δ (ppm) = 1.36–1.42 (m, 18 H), 1.47–1.52 (m, BuOH signal.44
9 H), 2.86–3.13 (m, 10 H), 3.34–3.56 (m, 16 H), 7.68 (d, 2 H,
3JHH = 8.2 Hz), 7.82 (d, 2 H, 3JHH = 8.2 Hz), 8.82–8.86 (m, 1 H).
Luminescence measurements
DO3A-SA-COOH (5). tert-Butyl protecting groups of 4 (74 mg,
99 μmol) were removed by reaction with 2 mL DCM/TFA (1 : 3)
overnight. The crude product was precipitated in Et2O and pur-
ified by HPLC (3–20% in 10 min, tR = 3–6 min). The product
was obtained as a colourless solid (47 mg, 83 μmol, 83.4%).
Analytical HPLC (3–20% B in 15 min): tR = 3.7 min. MS (ESI+):
C23H36N5O10S m/z: calcd 574.6 [M + H]+; found 574.8 [M + H]+,
596.7 [M + Na]+.
Emission and excitation spectra were measured on a Horiba
FluoroMax Plus-P spectrofluorometer equipped with a 150 W
ozone-free xenon arc lamp and a R928P photon counting emis-
sion detector, as well as a photodiode reference detector for
monitoring lamp output. Luminescence decays were measured
on the same instrument working in the phosphorescence
mode using a xenon flash lamp.
Density functional theory (DFT) calculations
Synthesis of the LnIII-complexes
All calculations were carried out using the TPSSh exchange–
correlation functional,45 which belongs to the hybrid meta-
GGA family, and the Gaussian 09 package (Revision D.01).46
Geometry optimizations of the [Gd2(L)2(H2O)]2 systems were
performed using the large-core relativistic effective core poten-
tial (LCRECP) of Dolg et al., which includes 46 + 4f7 electrons
of Gd3+ in the core, and the associated (7s6p5d)/[5s4p3d]-GTO
valence basis set.47 The standard 6-31G(d) basis set was used
for all other atoms. Bulk solvent effects (water) were considered
with the polarizable continuum model (PCM). We selected the
integral equation formalism (IEFPCM),48 together with univer-
sal force field radii (UFF)49 scaled by a factor of 1.1, to define
the solute cavities.
The two ligands were dissolved at ca. 25 mM concentration in
Tracepur® water. Gd, Eu or La-chloride hexahydrate (0.5 M
aqueous solutions) was added in a slight excess (1.1 eq.). The
pH was gradually adjusted to 6.5 and the solution stirred at
room temperature for three days. The pH was then increased
to 10 to precipitate excess LnIII as hydroxide. The solution was
centrifuged (3000 rpm, 3 min, r.t.) and the supernatant filtered
through a 0.2 μm filter. The pH was re-adjusted to 7 and the
solvent removed in vacuo. The LnIII-complexes were isolated as
colourless solids in quantitative yields.
Gd-7: Analytical HPLC (3–20% B in 15 min): tR = 3.6 min.
MS (ESI+): C25H36GdN5O10S m/z: calcd 756.0 [M + H]+; found
756.3 [M + H]+, 779.3 [M + Na]+.
Eu-7: MS (ESI+): C25H36EuN5O10S m/z: calcd 750.6 [M + H]+;
found 750.3 [M + H]+, 773.3 [M + Na]+.
Conflicts of interest
Gd-5: Analytical HPLC (3–20% B in 15 min): tR = 3.7 min.
MS (ESI+): C23H31GdN5O10S m/z: calcd 727.8 [M + H]+; found
728.4 [M + H]+, 750.3 [M + Na]+.
There are no conflicts to declare.
La-5: Analytical HPLC (3 min at 0% B, then 0–100% B in
15 min): tR = 9.7 min. MS (ESI+): C23H31LaN5O10S m/z: calcd
709.5 [M + H]+; found 709.4 [M + H]+.
Acknowledgements
Lorenzo Tei sincerely acknowledges the Alexander von
Humboldt Foundation (AvH) for a Research fellowship. The
authors also thank Dr. Claudio Cassino for the help in acqui-
sition of 1H NMR DOSY spectra, and the Deutsche
Forschungsgemeinschaft (CRC 824) for support.
1H NMR relaxometric measurements
1H NMRD profiles were recorded using a fast-field-cycling
Stelar SmarTracer relaxometer at magnetic field strengths
between 0.00024 and 0.25
T
(which correspond to
0.01–10 MHz proton Larmor frequencies). The relaxometer
operated under computer control with an absolute uncertainty
in 1/T1 of 1%. Additional data points in the range of
20–70 MHz were obtained using a Stelar Relaxometer and a
Bruker WP80 NMR electromagnet adapted to variable-field
measurements (20–70 MHz proton Larmor frequency). 1H T1
relaxation times at high magnetic fields were acquired with an
inversion recovery experiment with a mean 90° pulse width of
3.5 μs. The pH dependent relaxivities were measured at
20 MHz, 298 K in aqueous solution. pH values were adjusted
References
1 P. Caravan, C. T. Farrar, L. Frullano and R. Uppal, Contrast
Media Mol. Imaging, 2009, 4, 89–100.
2 K. R. Thulborn, J. C. Waterton, M. Matthews and
G. K. Radda, Biochim. Biophys. Acta, 1982, 714, 265–270.
3 L. Pauling and C. D. Coryell, Proc. Natl. Acad. Sci. U. S. A.,
1936, 22, 210–216.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2017

View Article Online
Dalton Transactions
Paper
4 M. E. Raichle, Proc. Natl. Acad. Sci. U. S. A., 1998, 95, 765– 28 K. Binnemans, Coord. Chem. Rev., 2015, 295, 1–45.
772.
29 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–
9792.
30 R. S. Dickins, D. Parker, J. I. Bruce and D. J. Tozer, Dalton
Trans., 2003, 1264–1271.
5 G. Angelovski, Angew. Chem., Int. Ed., 2016, 55, 7038–7046.
6 M. Carril, J. Mater. Chem. B, 2017, 5, 4332–4347.
7 C. T. Yang and K. H. Chuang, MedChemComm, 2012, 3,
552–565.
8 (a) S. M. J. van Duijnhoven, M. S. Robillard, S. Langereis
and H. Grüll, Contrast Media Mol. Imaging, 2015, 10, 282– 31 A. Beeby, I. M. Clarkson, R. S. Dickins, S. Faulkner,
308; (b) D. V. Hingorani, A. S. Bernstein and M. D. Pagel,
Contrast Media Mol. Imaging, 2015, 10, 245–265; (c) C. Shen
and E. New, Curr. Opin. Chem. Biol., 2013, 17, 158–166.
D. Parker, L. Royle, A. S. de Sousa, J. A. G. Williams and
M. Woods, J. Chem. Soc., Perkin Trans. 2, 1999, 493–
503.
9 C. Tu, R. Nagao and A. Y. Louie, Angew. Chem., Int. Ed., 32 R. M. Supkowski and W. D. Horrocks, Jr., Inorg. Chim. Acta,
2009, 48, 6547–6551. 2002, 340, 44–48.
10 A. Razgulin, N. Ma and J. Rao, Chem. Soc. Rev., 2011, 40, 33 S. Aime, M. Botta, M. Fasano and E. Terreno, Chem. Soc.
4186–4216. Rev., 1998, 27, 19–29.
11 E. L. Que and C. J. Chang, Chem. Soc. Rev., 2010, 39, 51–60. 34 D. H. Powell, O. M. N. Dhubhghaill, D. Pubanz, L. Helm,
12 S. Zhang, K. Wu and A. D. Sherry, Angew. Chem., 1999, 111,
3382–3384.
Y. S. Lebedev, W. Schlaepfer and A. E. Merbach, J. Am.
Chem. Soc., 1996, 118, 9333–9346.
13 F. K. Kálmán, M. Woods, P. Caravan, P. Jurek, M. Spiller, 35 N. Bloembergen and L. O. Morgan, J. Chem. Phys., 1961, 34,
G. Tircsó, R. Király, E. Brücher and A. D. Sherry, Inorg.
Chem., 2007, 46, 5260–5270.
14 M. P. Lowe and D. Parker, Chem. Commun., 2000, 707–708.
15 M. P. Lowe, D. Parker, O. Reany, S. Aime, M. Botta,
842–850.
36 J. H. Freed, J. Chem. Phys., 1978, 68, 4034–4037.
37 T. J. Swift and R. E. J. Connick, J. Chem. Phys., 1962, 37,
307–320.
G. Castellano, E. Gianolio and R. Pagliarin, J. Am. Chem. 38 F. Kielar, L. Tei, E. Terreno and M. Botta, J. Am. Chem. Soc.,
Soc., 2001, 123, 7601–7609. 2010, 132, 7836–7837.
16 A. Takács, R. Napolitano, M. Purgel, A. C. Bényei, 39 L. Tei, G. Gugliotta, G. Gambino, M. Fekete and M. Botta,
L. Zékány, E. Brücher, I. Tóth, Z. Baranyai and S. Aime,
Inorg. Chem., 2014, 53, 2858–2872.
17 L. Moriggi, M. A. Yaseen, L. Helm and P. Caravan, Chem. –
Eur. J., 2012, 18, 3675–3686.
Isr. J. Chem., 2017, 57, 887–895.
40 A. Rodriguez-Rodriguez, Z. Garda, E. Ruscsak, D. Esteban-
Gomez, A. de Blas, T. Rodriguez-Blas, L. M. P. Lima,
M. Beyler, R. Tripier, G. Tircso and C. Platas-Iglesias,
Dalton Trans., 2015, 44, 5017–5031.
18 E. Gianolio, S. Porto, R. Napolitano, S. Baroni,
G. B. Giovenzana and S. Aime, Inorg. Chem., 2012, 51, 41 J. Costa, E. Balogh, V. Turcry, R. Tripier, M. Le Baccon,
7210–7217.
F. Chuburu, H. Handel, L. Helm, E. Tóth and
19 (a) R. Pal and D. Parker, Org. Biomol. Chem., 2008, 6, 1020–
A. E. Merbach, Chem. – Eur. J., 2006, 12, 6841–6851.
1033; (b) D. G. Smith, B. K. McMahon, R. Pal and 42 M. Regueiro-Figueroa, A. Nonat, G. A. Rolla, D. Esteban-
D. Parker, Chem. Commun., 2012, 48, 8520–8522;
(c) B. K. McMahon, R. Pal and D. Parker, Chem. Commun.,
2013, 49, 5363–5365.
Gomez, A. de Blas, T. Rodriguez-Blas, L. J. Charbonniere,
M. Botta and C. Platas-Iglesias, Chem. – Eur. J., 2013, 19,
11696–11706.
20 G. B. Giovenzana, R. Negri, G. A. Rolla and L. Tei, 43 A. Rodriguez-Rodriguez, M. Regueiro-Figueroa, D. Esteban-
Eur. J. Inorg. Chem., 2012, 2035–2039.
Gomez, T. Rodriguez-Blas, V. Patinec, R. Tripier, G. Tircso,
F. Carniato, M. Botta and C. Platas-Iglesias, Chem. – Eur. J.,
2017, 23, 1110–1117.
21 Z. Baranyai, G. A. Rolla, R. Negri, A. Forgács,
G. B. Giovenzana and L. Tei, Chem. – Eur. J., 2014, 20,
2933–2944.
22 M. Giardiello, M. Botta and M. P. Lowe, Inorg. Chem., 2013,
52, 14264–14269.
23 E. Gianolio, R. Napolitano, F. Fedeli, F. Arena and S. Aime,
Chem. Commun., 2009, 6044–6046.
24 P. Caravan, J. J. Ellison, T. J. McMurry and R. B. Lauffer,
Chem. Rev., 1999, 99, 2293–2352.
25 T. Koike, E. Kimura, I. Nakamura, Y. Hashimoto and
M. Shiro, J. Am. Chem. Soc., 1992, 114, 7338–7345.
26 E. Kimura and T. Koike, Chem. Soc. Rev., 1998, 27, 179–
184.
44 D. M. Corsi, C. Platas-Iglesias, H. van Bekkum and
J. A. Peters, Magn. Reson. Chem., 2001, 39, 723–726.
45 J. M. Tao, J. P. Perdew, V. N. Staroverov and G. E. Scuseria,
Phys. Rev. Lett., 2003, 91, 146401.
46 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone,
B. Mennucci, G. A. Petersson, et al., Gaussian 09, Revision
D.01, Gaussian, Inc., Wallingford CT, 2009.
47 M. Dolg, H. Stoll, A. Savin and H. Preuss, Theor. Chim. Acta,
1989, 75, 173–194.
48 J. Tomasi, B. Mennucci and R. Cammi, Chem. Rev., 2005,
105, 2999–3093.
27 S. Aime, M. Botta, S. G. Crich, G. Giovenzana, R. Pagliarin,
M. Sisti and E. Terreno, Magn. Reson. Chem., 1998, 36, 49 A. K. Rappe, C. J. Casewit, K. S. Colwell and W. A. Goddard,
S200–S208.
J. Am. Chem. Soc., 1992, 114, 10024.
This journal is © The Royal Society of Chemistry 2017
Dalton Trans.