Aryl-phosphonate lanthanide complexes and their fluorinated derivatives: Investigation of their unusual relaxometric behavior and potential application as dual frequency 1H/19F MRI probes

FULL PAPER

Imaging Agents

M. P. Placidi, M. Botta, F. K. Kꢀlmꢀn,

G. E. Hagberg, Z. Baranyai,

A. Krenzer, A. K. Rogerson, I. Tꢁth,

N. K. Logothetis,

G. Angelovski* ............... &&&& — &&&&

Sealed in a dual: The unusually high

relaxivities of a series of macrocyclic

lanthanide complexes were investi-

gated, revealing a combination of

rapid water exchange rates with a sig-

nificant influence of water in the

second sphere. Fluorine-containing

1

complexes could be visualized by H

Aryl-Phosphonate Lanthanide

Complexes and Their Fluorinated

Derivatives: Investigation of Their

Unusual Relaxometric Behavior and

Potential Application as Dual

and ¹⁹F MR phantom imaging (see

figure); these platforms are promising

candidates for dual frequency MRI.

1

Frequency H/¹⁹F MRI Probes

Chem. Eur. J. 2013, 00, 0 – 0

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

&

1

&

ÞÞ

These are not the final page numbers!

DOI: 10.1002/chem.201300763

Aryl-Phosphonate Lanthanide Complexes and Their Fluorinated Derivatives:

Investigation of Their Unusual Relaxometric Behavior and Potential

1

Application as Dual Frequency H/¹⁹F MRI Probes

Matteo P. Placidi,^[a]Mauro Botta,^[b]Ferenc K. Kꢀlmꢀn,^[c]Gisela E. Hagberg,^{[a, d]}

Zsolt Baranyai,^[c]Andreas Krenzer,^[e]Alexandria K. Rogerson,^[f]Imre Tꢁth,^[c]

Nikos K. Logothetis,^{[a, g]}and Goran Angelovski*^[a]

Abstract: A series of low molecular

lower field strengths (20 MHz), two of

the complexes displayed a peak in r₁

(21.7 and 16.3 mm^ꢀ1s^ꢀ1) caused by an

agglomeration, that can be disrupted

through the addition of phosphate

anions. NMR spectroscopy revealed

that at least two species are present in

solution interconverting through an in-

tramolecular binding process. Two

complexes provided a suitable signal in

¹⁹F NMR spectroscopy and through the

selection of optimized imaging parame-

ters, phantom images were obtained in

a MRI scanner at concentrations as

low as 1 mm. The developed probes

could be visualized through both ¹H

and ¹⁹F MRI, showing their capability

to function as dual frequency MRI con-

trast agents.

weight lanthanide complexes were de-

1

veloped that have high H longitudinal

relaxivities (r₁) and the potential to be

1

used as dual frequency H and ¹⁹F MR

probes. Their behavior was investigated

in more detail through relaxometry,

pH-potentiometry, luminescence, and

multinuclear NMR spectroscopy. Fit-

ting of the ¹H NMRD and ¹⁷O NMR

Keywords: fluorine · gadolinium ·

imaging agents · NMR spectrosco-

py · relaxivity

profiles demonstrated

a very short

water residence lifetime (<10 ns) and

an appreciable second sphere effect. At

Introduction

Molecular imaging involves the detection of specific metab-

olites or biochemical processes at the cellular level, enabling

the early detection of diseases and dysfunctions by examin-

ing the underlying causes.^[1]Of the many different imaging

modalities considered, magnetic resonance imaging (MRI)

shows great promise, providing noninvasive anatomical

images with high spatiotemporal resolution at unlimited

depths of tissue penetration. Its diagnostic ability relies on

the differences in the longitudinal (T₁) and transverse (T₂)

relaxation times of healthy and diseased tissue. Paramagnet-

ic complexes are applied to improve the image quality re-

ducing the relaxation times of the tissues in which they lo-

calize.^[2]The vast majority administered in clinical settings

are T₁agents based on chelates of Gd³⁺, chosen because of

its favorable magnetic properties. The efficacy of a contrast

agent is measured in terms of longitudinal relaxivity (r₁), de-

fined as the increase in the observed longitudinal relaxation

rate of the surrounding water protons normalized to its con-

centration. The relaxivity arises from the time-dependent

magnetic coupling between the paramagnetic metal ion and

the protons of the surroundings water molecules. These can

be grouped in three main categories: the water molecules

directly coordinated to the metal ion (inner sphere; IS),

those diffusing in close proximity (outer sphere; OS), and

those interacting with the ligand structure (second sphere;

SS) for example, through hydrogen bonding.

[a] Dr. M. P. Placidi, Dr. G. E. Hagberg, Prof. Dr. N. K. Logothetis,

Priv.-Doz. Dr. G. Angelovski

Department for Physiology of Cognitive Processes

Max Planck Institute for Biological Cybernetics

72076 Tuebingen (Germany)

E-mail: goran.angelovski@tuebingen.mpg.de

[b] Prof. Dr. M. Botta

Dipartimento di Scienze e Innovazione Tecnologica

University of Piemonte Orientale “Amedeo Avogadro”

15121 Alessandria (Italy)

[c] Dr. F. K. Kꢁlmꢁn, Dr. Z. Baranyai, Prof. Dr. I. Tꢂth

Department of Inorganic and Analytical Chemistry

University of Debrecen, 4010 Debrecen (Hungary)

[d] Dr. G. E. Hagberg

Biomedical Magnetic Resonance, Department of Radiology

Tuebingen University Hospital, 72076 Tuebingen (Germany)

[e] A. Krenzer

Department of Chemistry, Faculty of Science

University of Tuebingen, 72076 Tuebingen (Germany)

[f] A. K. Rogerson

The School of Chemistry

University of Manchester, Manchester M139PL (UK)

[g] Prof. Dr. N. K. Logothetis

Department of Imaging Science and Biomedical Engineering

University of Manchester, Manchester M139PT (UK)

Supporting information for this article is available on the WWW

under http://dx.doi.org/10.1002/chem.201300763.

&

2

&

www.chemeurj.org

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 0000, 00, 0 – 0

ÝÝ

These are not the final page numbers!

Dual Frequency MRI Probes

FULL PAPER

Further progress in this direction has been hindered by

the inherent drawbacks associated with these agents—the

high detection limit and uncertainty regarding quantification

and localization. To lower the detection limit, their effec-

tiveness can be improved by the development of high-relax-

ivity agents. This requires tuning of the microparameters

that govern their inner sphere relaxivity, as described by the

Solomon–Bloembergen–Morgan treatment of paramagnetic

relaxation.^[3]These include slowing the rotation of the che-

late (t_R), decreasing the bound water residence lifetime

(t_M), or increasing the number of directly coordinated water

molecules (q).^[4]Portions of the ligand can also be used to

enhance the population of water molecules in the second

sphere, leading to an increase of r₁.^[5]Replacing one of the

carboxylate arms of GdDOTA with a methyl phosphonate

as in GdDO3AP, led to a reduction of t_Mand the presence

of two water molecules in the second sphere, thus causing

an increase in r₁.^[6]A number of studies in the past two dec-

ades have focused on the development of effective, high-re-

laxivity agents.^[7]

Agents able to satisfy the demands of the two MRI meth-

ods and produce both strong ¹H and ¹⁹F signals are rare.

They include either a mixture of molecules encapsulated in

micelles^[14]or self assembled structures.^[15]While these sys-

tems are effective, there is a desire to produce small, single

molecule agents with optimal dual frequency signals, by

combining a high-relaxivity agent with a ¹⁹F containing

group which can take advantage of the PRE effect due to

the proximity of the paramagnetic ion.

We have previously developed a series of macrocyclic li-

gands appended with an aryl-phosphonate moiety, suited for

both MR and optical imaging.^[16]The relaxivities of these

probes at high magnetic field strengths (7 T) were greater

than expected for low weight Gd³⁺complexes, more than

twice that of commercial agents. In addition to understand-

ing the factors influencing their behavior, we sought to pre-

pare and investigate the properties of four new agents, each

containing a CF₃moiety incorporated at different positions

(Figure 1). The distance and orientation of the ¹⁹F nucleus

The indirect detection of these agents results in difficulties

in interpreting their exact location and concentration, as

other intrinsic processes of the surrounding tissue also affect

the image contrast for example, hemorrhaging, inflammation

or lesions.^[8]To overcome this issue significant research has

been devoted to developing agents with a different MR

active nucleus where the measured signal is directly related

to the concentration of the substance, as in the case of radi-

oisotope imaging agents. ¹⁹F is ideally suited to this role as

the amount of fluorine in the soft tissues is insignificant re-

sulting in no interfering background signals. Also, it has a

1

very similar gyromagnetic ratio to H, allowing the use of

the same experimental setup to record both signals from

1

each subject. By combining the ¹⁹F image and the H ana-

tomical scan, much more detailed information can be ob-

tained. This has already been exploited to examine melano-

ma angiogenesis or quantify and monitor the fates of im-

planted stem and dendritic cells.^[9]Most investigations have

used perfluorinated nanoemulsions, however these systems

have inherently long relaxation times (the order of seconds),

which results in longer scan times. Also, some contain multi-

ple fluorine resonances that lead to a low signal intensity,

thus requiring high concentrations to obtain a suitable

signal. To overcome these issues, multiple fluorinated units

have been incorporated on macromolecules such as poly-

mers or dendrimers, though their high molecular weight re-

stricts them to the vasculature.^[10]Alternatively, using a para-

magnetic relaxation enhancement (PRE) effect, a paramag-

netic ion can be incorporated in close proximity to ¹⁹F, thus

reducing the relaxation times by several orders of magni-

tude, increasing the signal-to-noise ratio by enabling more

scans within shorter periods of time.^[11]In this direction re-

searchers have developed systems that are able to detect

enzyme activity,^[12]or a series of lanthanide chelates with

CF₃groups at optimized geometries to obtain an effective

signal detected at mm concentrations.^[13]

Figure 1. Structures of ligands L^1–8, DOTA, and DO3AP.

with respect to the Ln³⁺ion is of critical importance in de-

termining the strength of the ¹⁹F signal.^[17]Therefore, we

coupled the fluorinated aryl moiety to the macrocycle using

both a rigid and flexible linking unit, varying the position of

the CF₃group on the aromatic ring. Furthermore we sought

to analyze the best candidates in MRI phantoms, through

the use of new imaging parameters with the ultimate goal of

1

developing a platform for dual frequency H/¹⁹F MR imag-

ing.

Chem. Eur. J. 2013, 00, 0 – 0

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

&

3

&

ÞÞ

These are not the final page numbers!

G. Angelovski et al.

Results and Discussion

displayed a sharp peak at 395 nm of similar intensity, corre-

sponding to an f–f transition originating from the metal ion.

Excitation at either wavelength resulted in the same Eu³⁺

emission spectra centered at 617 nm. The observation of this

additional excitation peak indicates a reduced efficiency of

the sensitized emission process relative to the Tb³⁺com-

plexes. Given the electron-rich nature of the chromophore,

it is likely that the Eu³⁺emissive state is being partially de-

activated by a low-lying ligand to metal charge transfer

state, previously observed in other Eu³⁺phenol-containing

complexes.^[18]The nature of the other substituents on the ar-

omatic ring influences this quenching process to differing

extents, such that in EuL^1–4sensitized emission is inefficient

and in EuL^5–8it is not observed.

Ligand synthesis: The synthesis of the four new ligands L^1–4

followed a similar route to L^5–8(Figure 2),^[16]the main differ-

ence involved the construction of the multifunctional aro-

matic moiety. The MR active CF₃group was incorporated

either ortho or para to the phenol ether, hence the phospho-

nate was introduced by treating 2- or 4-triflouromethyl

phenol (1a,b) with diethyl phosphite to give 2a and 2b, re-

spectively. A Phospho-Fries rearrangement using LDA at

ꢀ788C, induced an ortho migration of the phosphonate re-

sulting in 3a and 3b. 1,3-dibromopropane and 1,4-bis(bro-

momethyl)benzene were monoalkylated with both 3a and

3b to give the bromides 4a–d. Here, the higher reactivity of

the xylyl-derivative necessitated its addition in a large

excess under high dilution conditions. The tris tert-butyl

ester derivative of cyclen was then treated with 4a–d and

the ligand precursors 5a–d were isolated using column chro-

matography on alumina, in high yields. The two different

ester protecting groups were cleaved sequentially, first using

bromotrimethylsilane in N,N-dimethylformamide and then

trifluoroacetic acid in dichloromethane. ¹H NMR spectro-

scopy and mass spectrometry confirmed the removal of the

protecting groups and each ligand (L^1–4) was purified using

RP-HPLC (see Tables S1–S3 in the Supporting Information

for the conditions) and lyophilized.

Table 1. Luminescence lifetimes^[a]and hydration numbers of EuL^1–8and

[b]

TbL^1–8

.

Complex

Tb³⁺

Eu³⁺

tD₂O

tH₂O

q

tD₂O

tH₂O

q

[ms]^[c]

[ms]^[d]

L¹

L²

L³

L⁴

L⁵

L⁶

L⁷

L⁸

3.07

3.28

2.78

2.83

3.40

3.23

3.10

3.04

1.65

1.83

1.63

1.60

2.01

1.97

1.90

2.20

1.1

0.9

1.1

0.7

0.3

1.92

1.90

1.73

1.77

1.73

0.80

0.97

0.76

0.54

0.60

0.64

0.58

0.36

0.47

0.37

1.3

1.1

0.9

1.1

1.5

1.0

1.3

The lanthanide complexes were prepared through the

slow addition of an equimolar amount of the corresponding

chloride salt to an aqueous solution of the ligand at 608C,

while maintaining the pH at 6–7 using sodium hydroxide.

Once the free ligands could no longer be detected through

mass spectrometry, the excess lanthanide ions were removed

by precipitation from solution at pH 10, followed by treat-

ment of the supernatant with Chelex. The absence of free

lanthanide ions was confirmed using the xylenol orange test.

[a] Lifetimes quoted are subject to a ꢁ10% error. [b] Data for TbL^5–8

from ref. [16]. [c] l_ex=285–300 nm, l_em=545 nm. [d] l_ex=395 nm, l_ex

=

617 nm.

When considering the available amino carboxylate donor

set, q is expected to be either 1 or 2 depending on if the

phosphonate is bound to the metal ion or not. This can be

determined from the luminescence lifetimes recorded in

H₂O and D₂O (MOPS, pH 7.4), and by applying the modi-

fied Horrocksꢃ equation (Table 1).^[19]The lifetimes were

measured either through excitation of the chromophore

(Tb³⁺, l_ex=285–300 nm) or by direct excitation of the metal

ion (Eu³⁺, l_ex=395 nm). For EuL^1–8the hydration numbers

ranged from 1–1.5 suggesting that the phosphonate is con-

nected to the metal ion. This is supported by the splitting

pattern of the DJ=1 manifold and the peak at 620 nm, ob-

Lanthanide luminescence: The luminescence properties of

EuL^1–4and TbL^1–4were examined by recording the excita-

tion and emission spectra (Figures S1–S4 in the Supporting

Information). For each of the Tb³⁺systems a strong ligand

based transition could be observed in the UV region and ex-

citation of this resulted in intense green emission centered

at 545 nm. In addition to the ligand excitation band, EuL^1–4

Figure 2. Synthetic scheme for L^1–4: a) HP(O)

d) tris-tBu-DO3A, K₂CO₃, CH₃CN or DMF, 708C; e) Me₃SiBr, DMF; f) TFA, CH₂Cl₂.

(OEt)₂, Et₃N, CCl₄; b) LDA, THF, ꢀ788C; c) dibromopropane or p-dibromoxylene, K₂CO₃, DMF, 608C;

&

4

&

www.chemeurj.org

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 0000, 00, 0 – 0

ÝÝ

These are not the final page numbers!

Dual Frequency MRI Probes

FULL PAPER

Table 2. Relaxometric parameters for GdL^1–8obtained from the fit of the ¹H NMRD profiles and ¹⁷O NMR

data.

served in the emission spec-

tra.^[20]The noninteger values

#

Gd³⁺

chelates

r₁

[mm^ꢀ1s^ꢀ1

D²

[10¹⁹s^ꢀ2

t_V

t_R

t_M

[ns]^[b]

DH_M

A

A/ꢀh

may represent

a mixture of

[a]

[b]

G

]

E

]

[Ps]

[kJmol^ꢀ1

]

[10⁶rads^ꢀ1

]

mono- and bis-hydrated species

(with the phosphonate in bound

and unbound positions respec-

tively). It is also possible that

the higher q values of EuL⁶and

EuL⁸are a result of the stron-

ger electron donating nature of

the phenol substituent, enhanc-

ing the charge transfer quench-

ing pathway and providing an

overestimation of q. TbL^1–7are

all monohydrated, the slightly

L¹

7.6

7.4

1.6 (3.0)^[c]

2.1

40 (27)^[c]

34

164 (148)^[c]

143

16 (15)^[c]

8

33.8 (31.6)^[c]

10

ꢀ3.7 (ꢀ3.7)^[c]

L²

–

L³

21.7 (8.4)^[d]

1.3 (2.6)^[d]

2.1 (2.4)^[d]

2.0

49 (29)^[d]

40 (32)^[d]

35

568 (166)^[d]

430 (168)^[d]

155

10 (12)^[d]

10 (11)^[d]

8.5

– (17.2)^[d]

– (18.1)^[d]

26.0

– (ꢀ3.5)^[d]

– (ꢀ3.6)^[d]

ꢀ3.5

L⁴

16.3 (8.6)^[d]

L⁵

7.7

7.9

6.9

6.3

4.7

L⁶

2.0 (4.1)^[c]

2.6

2.5

39 (24)^[c]

27

154 (142)^[c]

131

108

8.1 (8.3)^[c]

10

7.2

25.6 (25.3)^[c]

–

34.1

ꢀ3.5 (ꢀ3.5)^[c]

–

L⁷

L⁸

ꢀ3.4

DOTA^[e]

1.6

11

77

243

49.8

ꢀ3.6

[a] Recorded at 20 MHz and 258C. [b] Values obtained from the fitting of VT ¹⁷O NMR spectra. [c] Values ob-

tained by including the second sphere contribution in the fitting of the ¹H NMRD profile. [d] Values obtained

1

from fitting of the H NMRD profile in 200 mm PBS. [e] Ref. [23b].

lower q values are likely

a

result of the smaller ionic radii of Tb³⁺compared to Eu³⁺

.

ting of the NMRD profiles using the standard equations

that describe inner and outer sphere relaxivity, a series of

microparameters can be obtained (Table 2).^[4a]For the data

analysis it is necessary to make a reasonable estimate of

some of the parameters: q was fixed to 1; the Gd-inner

sphere water proton distance r=3.0 ꢄ, the Gd-outer sphere

water–proton distance a=4.0 ꢄ, the diffusion coefficient

D=2.24ꢅ10^ꢀ5cm²s^ꢀ1(258C). For each system increasing

the temperature from 25 to 378C, decreased r₁over the

entire frequency range investigated (0.01–70 MHz), indicat-

ing a stronger influence of t_Rover t_M. GdL^1–2,5–8displayed

typical behavior for monometallic chelates with a single dis-

persion at 1–7 MHz. The highest r₁values were observed at

The main exception is TbL⁸with a q value close to 0, which

seems reasonable given that the relaxivity of the GdL⁸is

lower than the other Gd³⁺complexes (see below).

Relaxometric characterization: The longitudinal proton re-

laxivities of GdL^1–8were determined at 20 MHz, 258C and

neutral pH by measuring the longitudinal relaxation times,

with two sets of values observed; 6.3–7.9 mm^ꢀ1s^ꢀ1for GdL^1–

and 16.3–21.7 mm^ꢀ1s^ꢀ1for GdL^3,4(Table 2). In the first

2,5–8

group there is an increase in r₁with increasing molecular

weight, in line with a reduction in the tumbling rate (Fig-

ure S5 in the Supporting Information). GdL¹and GdL⁶have

slightly elevated values given their hydration numbers (q=

1), indicating an additional contribution to r₁, likely due to

the presence of a significant number of second sphere water

molecules, previously observed for other phosphonate con-

taining systems.^[21]The elevated values of GdL^3,4are similar

to those found in slow tumbling macromolecular systems.^[22]

The r₁was measured as a function of pH (4–9), with each

complex displaying constant values (Figures S6–S9 in the

Supporting Information) indicating the hydration state and

hence the coordination environment is not significantly af-

fected at this pH range. As demonstrated previously for

GdL^5–8, the carbonate anions present at high pH do not in-

terfere by displacing the inner sphere water molecules. At

pH 3, a steady increase was observed indicating the partial

dissociation of the complex and ultimately the release of the

metal ion. The temperature dependence of r₁was also meas-

ured for three of the complexes (GdL¹, GdL³, and GdL⁴) re-

sulting in an exponential increase in r₁with decreasing tem-

perature—a typical behavior for rapidly exchanging systems

(Figures S10 and S11 in the Supporting Information). Here,

the water exchange rates (k_ex=1/t_M) are much shorter than

the longitudinal relaxation time of the bound water (T_1M

and hence will not limit the overall relaxivities.

)

NMRD profiles: The relaxometric behavior of GdL^1–8was

investigated in more detail by recording nuclear magnetic

resonance dispersion (NMRD) profiles (Figure 3 and Figur-

es S12–S15 in the Supporting Information).^[23]Through fit-

4

Figure 3. a) ¹H NMRD profiles of GdL³( ) and GdL

(

~

) in water at

&

5

^

*

258C; b) GdL in water at 25 ( ) and 378C ( ).

Chem. Eur. J. 2013, 00, 0 – 0

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

&

5

&

ÞÞ

These are not the final page numbers!

G. Angelovski et al.

lower magnetic field strengths, with similar values to

GdDOTA.^[4a,24]

This can be explained by the presence of a non-negligible

contribution from second sphere water molecules, not un-

common in the case of Gd³⁺chelates functionalized with

hydrophilic pendant groups. In this case we reanalyzed the

data by considering three additional parameters: t_R’ (r^S₁^S; in

the range 20–80 ps) and q’/r’ (r^S₁^S). A very good fit was ob-

tained by considering the contribution of two water mole-

cules (q’=2), located at a distance of 3.8 ꢄ from the metal

ion with a rotational correlation time of approximately 30–

35 ps (Figures S12 and S14 in the Supporting Information).

Clearly, this type of analysis is mainly qualitative as a net-

work of water molecules located at different distances from

the metal ion and with different lifetimes contribute collec-

tively to the overall relaxivity. However, it provides a rea-

sonable explanation for the unusual values of relaxivity and

more accurate relaxation parameters. In fact, all the NMRD

profiles have been analyzed assuming either the presence of

a single water molecule (q=1) in the inner sphere of the

metal ion and the occurrence of contributions to r₁from

inner- and outer sphere components only. One method to

check if q has been assigned correctly is to plot the calculat-

ed rotational correlation time, t_R, versus the molecular

weight of the complexes (Figure 4). According to the

Over this range r₁is known to be strongly influenced by

the electronic relaxation rates (1/T_ie), described by the trace

of the square of the zero-field splitting tensor, D²and the

correlation time describing its modulation, t_V. The D²values

are much closer to GdDOTA than GdDO3A,^[23b,25]support-

ing the predominance of a q=1 species and hence the coor-

dination of the phosphonate arm. At higher field strengths

r₁decreases, coinciding with the region where the t_Rstrong-

ly influences r₁, a result of the rapid tumbling rate of the

complex, though the t_Rvalues obtained for GdL^1–2,5–8are

twice that of GdDOTA at 150 ps.^[26]The linear increase of r₁

with molecular weight observed with most of these com-

pounds, combined with the occurrence of q=1 and the t_R

values calculated suggest that the phosphonate arm is most

likely bound to the metal center. If the phosphonate arm re-

mained unbound, the flexible solution state structure would

result in lower t_Rvalues for molecules of this molecular

weight.

The large peak in r₁measured at 30–40 MHz in the

NMRD profiles of GdL^3,4is indicative of a slow tumbling

species, observed previously for many high molecular weight

systems.^[27]The t_Rof 568 and 430 ps were obtained for

GdL^3,4, however, due to the presence of more than one spe-

cies in solution (see below) these values can only be consid-

ered in a qualitative sense. Given the available donor set

and the rigid nature of the xylyl linking unit, it appears that

two or more complexes are associating together to form

slower diffusing dimers or oligomers. This effect has been

observed previously by Merbach et al., who found a series

of xylyl-linked macrocyclic complexes to agglomerate in sol-

ution, providing high t_Rand r₁values.^[28]

This likely involves hydrophobic p–p stacking and/or hy-

drogen bonding interactions between the aromatic cores and

the macrocyclic binding sites respectively. To test this hy-

pothesis we recorded the NMRD profiles of GdL^3,4and

GdL¹(as a control) in a solution of PBS containing either a

standard or excess amount of phosphate anions (previously

shown to break down such aggregates).^[28]Should these in-

teractions be disrupted, then the form of the profiles would

alter and the peak in the slow tumbling region should de-

crease or disappear. A reduction in the peak height can

clearly be observed in the slow tumbling region in the pres-

ence of 10 mm phosphate (Figures S16 and S17 in the Sup-

porting Information), with the change being more promi-

nent for GdL⁴. The lesser effect on GdL³suggests that the

agglomeration is much stronger, which is reasonable given

that its r₁in the absence of PBS is higher. On addition of

200 mm phosphate the peak completely disappears, with the

form and intensity of the profiles resembling that of the

other monomeric complexes. Accordingly, the best-fit pa-

rameters assume standard values (Table 2). We previously

observed that the r₁of GdL¹and GdL⁶appear sensibly

higher than typical values for monohydrated Gd³⁺com-

plexes of similar molecular size (Figure S5 in the Supporting

Information).

Figure 4. Rotational correlation time, t_R, versus molecular weight for

GdDOTA and for the complexes discussed in the present work. Open

symbols refer to the t_Rvalues after refining the NMRD profiles with in-

clusion of the second sphere relaxivity (GdL^1,6).

Debye–Stokes equation, a linear relationship should be

found for rigid spherical molecules with negligible internal

motion. It is clear that a good linear correlation is obtained

for GdL^2–5. For GdL^1,6the calculated t_Rare too large, but

fall on the straight line after consideration of the second

sphere contribution. On the other hand, for GdL^7,8the cal-

culated t_Rappear to be underestimated and this is likely the

result of the compensation for a lower effective q value.

Both GdL^7,8values follow the Debye–Stokes relationship

more closely if q is set to approximately 0.9. Another possi-

ble explanation for the deviation from linearity in the case

of GdL^7,8may be the presence of an anisotropic molecular

rotation which results in a lower “effective” correlation time

t_R.

&

6

&

www.chemeurj.org

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 0000, 00, 0 – 0

ÝÝ

These are not the final page numbers!

Dual Frequency MRI Probes

FULL PAPER

Table 3. Protonation constants of L¹, L³, and L⁸(0.15m NaCl, 258C).

Variable temperature ¹⁷O NMR: Variation of t_Mbetween 5–

50 ns provided the same degree of fitting for the NMRD

profiles. Therefore, the transverse relaxation rates (R_2p) and

shift (Dw) of the ¹⁷O nucleus were recorded as a function of

temperature for a selection of complexes (Figure 5 and Fig-

L¹

L³

L⁸

DOTA^[a]

H

logK₁

logK₂

logK₃

logK₄

logK₅

logK₆

logK₇

9.89 (1)

9.41 (2)

10.67 (1)

10.06 (1)

9.64 (1)

7.72 (2)

4.25 (3)

3.52 (2)

2.00 (1)

47.86

9.14

9.21

4.48

4.03

1.99

–

9.88 (2)

7.09 (4)

4.25 (5)

3.61 (4)

1.79 (4)

–

9.26 (2)

7.17 (3)

4.53 (4)

3.37 (4)

2.15 (4)

–

SlogK_i^H

36.50

35.88

28.85

[a] Ref. [32].

Protonation equilibria: The protonation constants of L^1,3,8

(logK_i^H) were determined by pH-potentiometric titrations

(Table 3, the definitions and equations used for the evalua-

tion of the equilibrium data are summarized in the Support-

ing Information). The logK_i^Hvalues of these ligands differ

considerably from those of DOTA, which can be explained

in part by the different protonation sequences. The protona-

tion Scheme of DOTA has been studied in detail, where the

first two take place at the transannular macrocyclic amines

and the remaining three correspond to the carboxylate

arms.^[30]The protonation sequences of L^1,3are similar to

each other, the first and second steps correspond to the two

ring amines, the third, fourth, fifth, and sixth protonations

occur at the phosphonate and the carboxylate groups, re-

spectively. At very acidic conditions the phosphonate group

is expected to undergo a further protonation,^[31]which could

not be accurately determined by direct pH-potentiometric

titrations. For L⁸, the first protonation takes place at the

phenol, the second and third are related to the macrocyclic

amines. The fourth protonation of L⁸involves the phospho-

nate group and the remaining three take place at the car-

boxylate groups.

Figure 5. Reduced transverse: a) ¹⁷O relaxation rates, and b) ¹⁷O chemical

shifts of an aqueous GdL⁵(19.2 mm, 11.75 T, neutral pH).

ures S18–S22 in the Supporting Information) and t_Mcould

be accurately determined by application of the Swift–Con-

nick equations.^[29]Again, some parameters were fixed to

standard values: q=1, the activation energy of the correla-

tion time for the modulation of the transient zero-field-split-

ting tensor E_V=1.5 kJmol^ꢀ1and the activation energy of the

rotational correlation time E_R=18 kJmol^ꢀ1. In addition, for

D²and t_Vthe values derived from the fit of the NMRD pro-

files were used. The variable parameters were: the mean

The logK_i^Hvalues of the amines in the macrocyclic ring

for L^1,8are significantly higher than those of L³and DOTA,

H

to the extent that the logK₂of the former are higher than

H

the logK₁values of the latter. The higher amine logK_i^H

values of L^1,8may be due to the following effects: 1) the

electron donating and withdrawing effects of the alkyl (L^1,8)

versus aromatic (L³) linkers on the ring N-atom, which in-

creases the basicity, 2) a stronger complex formation be-

tween the fully deprotonated ligands (L³) and sodium ion

which reduce the proton affinity of the ring N-atoms, 3) a

greater degree of electrostatic repulsion between the proto-

nated ring N-atoms (L³), and 4) the potential barrier for the

formation of a H-bonding interaction between the protonat-

ed nitrogen atom and the carboxylate group increasing the

basicity of amines (L^1,8). The logK_i^Hof the phosphonate

group occurs at a range of values similar to that measured

previously for other aryl-phosphonate derivatives^[31]and the

logK_i^Hvalues of the carboxylic groups are comparable in all

cases.

residence lifetime of the inner sphere water molecule t_M, its

#

enthalpy of activation DH_M, and the scalar Gd ¹⁷O_wcou-

pling constant A/ꢀh.

ꢀ

An exponential increase in R_2pis observed with decreas-

ing temperature for each system demonstrating their rapid

exchange of water with the surrounding bulk, to the extent

that the peak in R_2pwhich corresponds to the crossover be-

tween the slow and fast exchange cannot be observed. A t_M

of approximately 8 ns was obtained for each system studied

(Table 2) in the range observed previously for other phos-

phonate appended macrocyclic systems. It is likely that the

presence of the bulky phosphonate group is causing steric

crowding around the metal ion, increasing the rate of ex-

change.^[6,21b,25]

Complexation properties: It is well established that di- and

trivalent metal ions form very stable complexes with DOTA

and its derivatives, due to the coordination cage formed by

Chem. Eur. J. 2013, 00, 0 – 0

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

&

7

&

ÞÞ

These are not the final page numbers!

G. Angelovski et al.

Table 4. Stability and protonation constants of the Ca²⁺, Zn²⁺, Cu²⁺and

the four ring nitrogen and four carboxylate oxygen donor

atoms. The stability constants of the Ln³⁺complexes L^1,3,8

(which possess four N- and three carboxylate O-donor

atoms) are expected to be lower than those of DOTA. How-

ever, given the solution-state behavior described in the pre-

vious sections, the influence of the aryl-phosphonate pend-

ant arm on the coordination with the metal center cannot

be ruled out. The formation of Ln³⁺complexes with DOTA

and its derivatives is slow at the pH range 3–6, because the

entrance of the Ln³⁺ion into the cage formed by the donor

atoms is hindered. The two macrocyclic nitrogens are proto-

nated forming an intermediate *LnH₂DOTA in the first step

of the complexation (the stability constant of the intermedi-

Gd³⁺complexes of L¹, L³and L⁸(0.15m NaCl, 258C).

L¹

L³

L⁸

DOTA

CaL

CaHL

CaH₂L

ZnL

10.26 (4)

7.32 (3)

–

18.66 (4)

7.25 (3)

3.86 (2)

2.99 (1)

–

22.56 (2)

7.11 (1)

4.07 (1)

2.63 (2)

1.74 (2)

–

18.3 (1)

6.86 (2)

–

10.07 (6)

7.12 (5)

–

16.98 (2)

7.11 (2)

3.99 (2)

2.97 (2)

9.43 (2)

13.84 (1)

3.90 (2)

–

17.35 (1)

4.16 (2)

3.40 (1)

2.65 (1)

–

10.63 (2)

8.18 (1)

17.95 (3)

10.55 (3)

8.13 (2)

4.10 (1)

2.96 (1)

21.75 (3)

10.50 (1)

8.02 (3)

4.27 (4)

2.63 (3)

2.12 (4)

19.0 (1)

10.15 (2)

6.92 (4)

–

ZnHL

ZnH₂L

ZnH₃L

ZnH₄L

CuL^[a]

CuHL

CuH₂L

CuH₃L

CuH₄L

CuH₅L

GdL^[c]

GdHL

GdH₂L

*GdH₃L^[g]

*GdH₄L^[g]

–

21.17 (4)

6.96 (3)

3.84 (4)

2.21 (4)

22.00^[b]

4.08

3.41

0.83

–

ate can be defined as *K_{LnH L}=[*LnH₂L]/ACHTNUTRGENNUG ACHTUTGNREN[NNGU H₂L]). This

[Ln³⁺]

1.80 (4)

2

–

slowly deprotonates with the rate determining step involving

the loss of the second proton^[33](here the stability constants

were determined using the “out-of-cell” technique, because

of the slow rate of complex formation). The use of separate

samples in these measurements leads to larger errors in

logK_GdDOTAresulting in a wide range of values (22.1–28.0) in

the literature.^[34]These large discrepancies may also be at-

tributed the use of very different measurement techniques

employed (pH-potentiometry, direct spectrophotometry,

competition with Arsenaso III, spectrofluorometry, and solu-

bility). Therefore, to accurately determine the stability con-

stants of the Gd³⁺complexes, the proton relaxation rates of

Gd³⁺(0.002m) in the presence of L^1,8(0.002m) were meas-

ured over the pH range 2.0–3.4 (Figure S23 in the Support-

ing Information).

24.7^[d]/22.4^[e]

–

*5.9 (Eu³⁺

)

[f]

5.25 (8)

–

5.51 (5)

[a] Determined by VIS-spectrophotometry; the ionic strength was not

constant. [b] Ref. [32]. [c] Determined by relaxometry (0.15m NaCl,

258C). [d] Ref. [35] (0.1m NaCl, 258C). [e] Ref. [36] (0.1m NaCl 258C).

[f] Stability constant of the EuACHTUNGTRNEUNG

(H₂DOTA) intermediate.^[33a,37][g] Stability

constant of intermediate (0.15m NaCl, 258C).

(CN) of the metal ions and the lower denticity of L^1,3,8. The

coordination requirements of Zn²⁺and Cu²⁺(CN 6), can be

fully saturated by the 7 donor atoms. Moreover, it is possible

that L^1,8have a more suitable cavity size for Zn²⁺and Cu²⁺

,

In addition to free Gd³⁺and GdL^1,8, the intermediates

*GdH₄L¹and *GdH₃L⁸are also present at low concentra-

tions (4–20%). These intermediates contribute significantly

to r₁due to the presence of 4–5 water molecules occupying

their inner-sphere. In *GdH₄L⁸two diagonal ring nitrogens,

the phenolate and phosphonate groups are protonated and

hence unable to coordinate, in *GdH₃L¹two diagonal ring

nitrogens and the phosphonate group are protonated. The

stability constants of GdL^1,8were calculated by using the r₁

resulting in higher stability constants. Interestingly, the

lower denticity of CuL¹resulting in a higher logK_MLvalue

(when compared to CuDOTA) has also been observed with

CuDO3A (logK_CuDO3A=26.49, 0.5m KNO₃, 258C).^[38]Since

the basicity of the N-donor atoms of L¹, DO3A and DOTA

ligands are very similar, it can be assumed that the higher

stability of the CuL¹complex is caused by the optimal coor-

dination geometry of the macrocyclic donor atoms for the

Cu²⁺, due to the presence of the aromatic substituent on the

ring N-atom.

of the Gd ion, GdH₂L⁸, GdHL¹, and the intermediates

3þ

aq:

*GdH₄L⁸and *GdH₃L¹, which were determined in separate

Overall the stability constants of the Ca²⁺, Zn²⁺, Cu²⁺

,

experiments (see Table S4 in the Supporting Information for

the r₁of each intermediate). The protonation constants were

also determined by pH-potentiometric titrations for GdL^1,8,

whereas the precipitation of GdH₂L³prevented an accurate

protonation constant determination.

and Gd³⁺-complexes indicate that the presence of the aryl-

phosphonate pendant arm and its potential interaction with

the metal center does not significantly improve the stability

of the complex. Indeed, the protonation of the Ca²⁺, Zn²⁺

,

and Cu²⁺complexes formed with L^1,3,8takes place at a simi-

lar or slightly higher pH than that of the corresponding li-

gands (Table 4), thereby indicating the absence of an inter-

action between the aryl-phosphonate group and the metal

ions. This is also seen in the further protonation constants of

the Zn²⁺and Cu²⁺complexes of L^1,3,8that involve the non-

coordinated carboxylate pendant arms. However, the Gd³⁺

complexes behave differently—all protonation constants (of

the phosphonate group in GdL¹or both, the phenol and the

phosphonate group in GdL⁸) are lower than those of the re-

spective ligands L^1,8(Tables 3 and 4). The more acidic

nature of the phosphonate group in a Ln³⁺complex has also

occurred in other phosphonate appended systems in which

The stability and protonation constants of Ca²⁺, Zn²⁺

,

and Cu²⁺complexes of L^1,3,8were investigated by pH-poten-

tiometric titrations (Table 4), and the stability constants of

CuL^1,3,8were determined by spectrophotometry (Figure S24;

the absorption spectra and the equations used for the evalu-

ation of the equilibrium data are summarized in the Sup-

porting Information). The stability constants of the Ca²⁺

and Gd³⁺-complexes of L^1,3,8are lower than those of the cor-

responding DOTA complexes, whereas the values for the

Zn²⁺and Cu²⁺complexes are comparable or higher. The

smaller logK_MLvalues of the Ca²⁺and Gd³⁺complexes can

be explained in terms of the higher coordination number

&

8

&

www.chemeurj.org

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 0000, 00, 0 – 0

ÝÝ

These are not the final page numbers!

Dual Frequency MRI Probes

FULL PAPER

this group interacts with Gd³⁺

.

These observations indi-

The proton assisted dissociation process most likely in-

volves a proton transfer step from the phosphonate side arm

to the nitrogen atom of the macrocyclic ring, followed by

the loss of Gd³⁺. The results of the best fit of the k_dvalues

to Equations (2) and (3) are shown in Table 5 (see Figur-

es S25–S27 in the Supporting Information for the depend-

ence of k_don [H⁺]).

[36,39]

cate that a similar process takes place in GdL^1,8, supporting

the above differences in stability constants of the investigat-

ed metal complexes, and a result of the higher CN of Gd³⁺

.

Dissociation kinetics: To further assess the potential stability

of a Gd³⁺complex in vivo, the kinetic inertness of GdL^1,3,8

was investigated. The dissociation of the metal com-

plexes in vivo can occur either when endogenous

metal ions (Ca²⁺, Zn²⁺or Cu²⁺) compete with the

Gd³⁺for the ligand or more rarely, by competition

of endogenous ligands for Gd³⁺. The effect of a

tenfold excess of Cu²⁺was investigated by spectro-

photometry in the pH range 3–5 and the kinetic in-

ertness of the metal complexes was characterized

by the rate constants of the exchange reactions with

the endogenous metal ion. The excess Cu²⁺ensures

the reaction can be regarded as pseudo-first-order

and the rate of the reaction can be expressed by Equa-

tion (1), where k_dis the pseudo-first-order rate constant and

[GdL]_totis the total concentration of the Gd³⁺complex.

Table 5. The rate constants of the dissociation of GdL¹, GdL³, and GdL⁸(258C, 0.15m

NaCl).

GdL¹

GdL³

GdL⁸

GdDTPA^[a]

GdDOTA^[b]

k₀[s^ꢀ1

k₁[s^ꢀ1m^ꢀ1

K_H[m^ꢀ1

]

G

–

5.2ꢁ0.3

–

0.58

100

6.7ꢅ10^ꢀ11

1.8ꢅ10^ꢀ6

14

]

8.5ꢁ0.2

4.0ꢁ0.1

]

–

190ꢁ62

[a] Ref. [39]. [b] Ref. [40].

During the fitting process the k₀values of the decomplex-

ation of GdH₂L⁸and GdHL³, as well as the protonation con-

stants of the intermediates K_{GdH L}₈and K_{GdH L}₃were found to

3

2

be negligible (values are comparable with their standard de-

viation). The proton assisted dissociation rate of the investi-

gated complexes is under one order of magnitude higher

than that of GdDTPA and 2–3 times slower than GdDTPA-

BMA,^[41]MRI contrast agents frequently used in medical

practice. However, it occurs several orders of magnitude

quicker than that of GdDOTA and GdDO3A.^[42]Previously

the substitution of the carboxylate with a phosphonate

group increased the rate of proton-assisted dissociation, as

the phosphonate groups are readily protonated even in

bound form, facilitating the proton transfer to the macrocy-

clic nitrogen.^[43]The presence of the phosphonate moiety in

the investigated ligands may help explain this lower kinetic

inertness, given that the interaction between the phospho-

nate group and Gd³⁺does not improve the thermodynamic

stability of the complex (Table 4). It is possible that addi-

tional effects such as steric crowding and/or intermolecular

dimerization are contributing to the destabilization of the

complexes. However, the investigated complexes are suita-

ble for further in vivo testing, as prototype contrast agents

due to their similar kinetic inertness to agents such as

GdDTPA.

d½GdLꢂ

ð1Þ

ꢀ

¼ k_d½GdLꢂ_tot

dt

Over the pH range 3–5, GdL⁸exists predominantly in the

diprotonated form, whereas GdL¹forms a monoprotonated

species. The equilibrium stability and protonation constants

of GdL³could not be determined owing to its low solubility,

therefore it was assumed that only the phosphonate moiety

is protonated. The decomplexation reactions of the Gd³⁺

complexes can occur via spontaneous, proton assisted and

metal assisted pathways (Scheme S1 in the Supporting Infor-

mation). The k_dvalues of the metal exchanging reactions in-

crease linearly with the H⁺concentration. It has been found

that the role of the exchanging metal ion is negligible in this

initial process, so the dissociation of the complexes can take

place via spontaneous and proton assisted pathways, fol-

lowed by the fast reaction between the free ligand and the

exchanging metal ion. Taking into account these possible

pathways and considering that [GdL⁸]_tot=[GdH₂L⁸]+

[GdH₃L⁸] and K_{GdH L}₈=[GdH₃L⁸]/

G

3

more [GdL^1,3]_tot=[GdHL^1,3]+[GdH₂L^1,3] and K_{GdH L}_1,3

ACHTUNGTERNUNNG =

ACHTUNGTRENNUNG

[GdH₂L^1,3]/[GdHL^1,3][H⁺], the rate constants characterizing

NMR spectroscopy: Further information about the coordi-

nation environment of Ln-DOTA- or DO3A-type com-

plexes can be obtained from NMR spectroscopy. The pres-

ence of a paramagnetic ion is expected to alter the width

and position of the ³¹P NMR peak, the extent of which is

not only dependent on the lanthanide itself but also the dis-

tance and orientation of the NMR active nuclei with respect

to this ion.^[44]Eu³⁺complexes are frequently used to deter-

mine the nature of the species present in solution, reflected

in the number and position of the peaks. Given its proximity

in the lanthanide series, Eu³⁺is often used as a surrogate to

provide information that cannot be obtained directly from

the Gd³⁺analogues. For EuL^1–8peaks in two regions at 10

the spontaneous and proton assisted dissociation can be

given by Equations (2) and (3), where k₀characterizes the

spontaneous decomplexation and k₁is the fist-order rate

constant of the proton assisted dissociation of GdH₂L⁸and

GdHL^1,3.

k₀þk₁½H^þꢂ

k_d¼

ð2Þ

ð3Þ

1þK^H_{GdH L}₈½H^þꢂ

2

k₀þk₁½H^þꢂ

1þK^H_GdHL_1,3½H^þꢂ

Chem. Eur. J. 2013, 00, 0 – 0

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

&

9

&

ÞÞ

These are not the final page numbers!

G. Angelovski et al.

to 15 and ꢀ90 to ꢀ150 ppm were observed (Figure 6 and

Figures S28–S31 in the Supporting Information), the latter

being significantly broader. This suggests that the phospho-

nate experiences two environments: one in close proximity

to the paramagnetic ion, the other further away. We envis-

age a reversible binding process with different proportions

of the phosphonate group free or coordinated to Eu³⁺

.

When considering the form of the NMRD profiles and the

t_Rof GdL^1,2,5–8(Table 2), it appears that this occurs in an in-

tramolecular fashion. Given the agglomeration observed for

GdL^3,4, the exact nature of the process is more difficult to

determine.

Figure 7. ¹H NMR spectra of EuL³(258C, D₂O).

ble with the MR imaging pulse sequences. There are many

factors that influence the number and proportion of isomers

of lanthanide complexes in solution and hence the number

of ¹⁹F signals, including the choice of the lanthanide ion.

Therefore, for each ligand L^1–4a series of 4 different com-

plexes (Eu³⁺, Gd³⁺, Tb³⁺, and Yb³⁺) were made and their

¹⁹F NMR spectra were recorded. 14 of the complexes con-

tained multiple peaks rendering them unsuitable, as this

would reduce the overall signal intensity in MRI. However,

for GdL⁴and YbL⁴a single broad signal was observed (Fig-

ure S36 in the Supporting Information) with T₁, T₂times of

3.1, 2.0 and 272.8, 4.2 ms respectively, allowing them to be

examined further in MRI phantom measurements (see

below).

Figure 6. ³¹P NMR spectra of EuL³(258C, D₂O, 1 % H₃PO₄external

standard).

Lanthanide complexes of DOTA and its derivatives typi-

cally exist in solution as two interconverting stereoisomers:

square antiprismatic (SAP) and twisted square antiprismatic

(TSAP) geometries.^[45]The isomers differ in terms of the

angle between the two planes containing the four oxygen

and four nitrogen coordinating atoms, that sit above and

below the lanthanide ion (ca. 40 and 258 respectively).^[46]

The agglomeration observed in the relaxometric analysis

was examined using the diamagnetic Y³⁺complexes of L^1,3,

prepared as representatives of discrete and aggregating spe-

1

cies respectively. H DOSY NMR spectroscopy was used to

determine the number of species in solution and the diffu-

1

sion rate (D) as a function of the H chemical shift (Figur-

1

The H NMR spectra of EuL^1–8displayed a series of broad-

es S37–S39 in the Supporting Information). D is known to

be dependent on the size and shape of a given species in sol-

ution^[49]and it is expected that the formation of agglomer-

ates will result in a reduction of this value. The aromatic

region was examined in more detail as it contained fewer

and better resolved signals for easier analysis, compared to

the broader signals of the aliphatic protons undergoing

rapid exchange processes. As expected, only one species

could be observed for YL¹with a diffusion coefficient of 3ꢅ

10^ꢀ10m²s^ꢀ1, matching that of a monomacrocyclic complex of

similar molecular weight (determined by PGSE ¹H NMR

experiments).^[50]For YL³at least two species were observed

with diffusion coefficients ranging from 1.5–2.3ꢅ10^ꢀ10m²s^ꢀ1,

both diffusing at a slower rate than YL¹. It is likely that the

slower diffusing species correspond to an intermolecular

dimer and oligomer, supported by the behavior demonstrat-

ed in the NMRD profile. The DOSY NMR spectrum of YL³

was then re-measured in the presence of 200 mm PBS to ex-

amine the disruption of the agglomeration process (Figur-

es S40 and S41 in the Supporting Information). Here a sim-

plification of the spectrum can be observed with a reduction

ened signals over a range of ꢀ40 to +40 ppm, characteristic

of complexes of this ion (Figure 7 and Figures S32–S35 in

the Supporting Information). Given the low symmetry of

this system and DO3A derivatives in general, a signal for

each proton is expected.^[47]Attention is often focused on the

axial protons of the macrocyclic ring as they experience the

greatest shift to a region with fewer signals, enabling an

easier interpretation of the species present. For EuL^3,6–8

only the SAP isomer was detected with four axial peaks in

the region 20–50 ppm.^[48]For the remaining complexes at

least two species were observed corresponding to both SAP

1

and TSAP isomers. When comparing the respective H and

³¹P NMR spectra, a single peak is present in the bound

region (ꢀ90 to 150 ppm in ³¹P NMR) when only the SAP

isomer is present, whereas two or more signals are observed

in this region when both SAP and TSAP isomers are detect-

ed.

For application in ¹⁹F MRI the complex must give a suita-

ble signal, ideally a single resonance in the corresponding

¹⁹F NMR spectrum with T₁and T₂relaxation times compati-

&

10

&

www.chemeurj.org

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 0000, 00, 0 – 0

ÝÝ

These are not the final page numbers!

Dual Frequency MRI Probes

FULL PAPER

in the number of signals, especially in the aromatic region

with a diffusion coefficient of 2.0ꢅ10^ꢀ10m²s^ꢀ1. This indicates

a significant disruption of the aggregate, to a form that

more closely resembles YL¹.

tom images with an SNR above 3 in a 1 h acquisition time

using the FISP sequence. The SNR of the 0.5 mm vial was

3–4 after 15 min (data not shown), representing the detec-

tion limit for the dual signal originating from this complex

(Figure 9).

¹H and ¹⁹F MRI phantom measurements: The imaging capa-

bilities of these complexes were assessed in vitro, by record-

1

ing H MR phantom images of GdL⁴at various concentra-

tions in MOPS buffered solution (10–80 mm), using a 7 T

horizontal animal bore MRI scanner. As expected the pres-

ence of the complex caused a significant brightening of the

image contrast, due to a reduction in the T₁time of the sur-

rounding water. In parallel,

a series of dilutions of

GdDOTA were also recorded at the same concentrations,

and throughout this range GdL⁴is more readily observable,

which is expected given its higher r₁value (Figure S42 in the

Supporting Information). ¹⁹F MR images were recorded for

both YbL⁴and GdL⁴which displayed a single broad signal

(see above), with the same MRI scanner. For each complex

a series of four buffered solutions at different concentrations

(1–5 mm) were prepared to determine the detection limit of

each system in vitro, which is related to the T₁and T₂times

of both complexes. Using a FLASH sequence signals from

both complexes could be observed, providing the images

with a total collection time of 60 min. This sequence proved

most effective for YbL⁴yielding a maximal signal-to-noise

ratio (SNR) of 14 at 5 mm with similar SNR for the highest

GdL⁴concentration (Figures 8a and b, respectively). Em-

ploying an alternative fast imaging FISP sequence,^[51]

a

much stronger signal for GdL⁴could be obtained. In only

15 min a SNR similar to that from the FLASH sequence

was achieved, with concentrations as low as 1 mm detected,

providing a maximal SNR of 9 (Figure 8c).

Figure 9. ¹H/¹⁹F MR phantom images of GdL⁴, recorded using an inver-

sion recovery sequence and R₁mapping by nonlinear least square fitting

(¹H, top) and the FISP (¹⁹F, bottom) sequence. The numbers in the

¹H MRI represent concentrations in mm.

Conclusion

Figure 8. ¹⁹F MRI phantom images of: a) YbL⁴using FLASH, b) GdL⁴

recorded using FLASH, and c) FISP sequences; complex concentrations:

a) 1.0 (bottom), 2.0 (right), 3.5 (left) and 5.0 mm (top); b,c) 1.0 (right),

2.0 (bottom), 3.5 (left), and 5.0 mm (top).

Various physicochemical characteristics of a series of aryl-

phosphonate macrocyclic complexes were investigated in

more detail to understand the origins of their high r₁and ex-

plore the potential for their use as dual frequency MRI con-

trast agents. Despite being monohydrated complexes, the

enhanced r₁appears to be the result of a comparatively slow

rotation combined with a significant contribution from

water molecules residing in the second hydration sphere.

Furthermore, the fast rate of water exchange of these com-

plexes will not limit their relaxivity at low temperatures or

represent a limiting factor of r₁in the case of their macro-

molecular conjugates.^[52]Additionally, two of the complexes

Since GdL⁴also has the potential to act as a dual frequen-

cy probe, further MR images were recorded at a lower

range of concentrations (50–1000 mm). As expected, the

1

shortest H T₁time and hence the brightest image resulted

from the highest concentration of the complex, with an in-

crease in T₁with decreasing concentration. By altering the

pulse sequence and tuning the receiver coil, complex con-

centrations as low as 0.1 mm could be observed in ¹⁹F phan-

Chem. Eur. J. 2013, 00, 0 – 0

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

&

11

&

ÞÞ

These are not the final page numbers!

G. Angelovski et al.

(282 MHz, CDCl₃,): d=ꢀ61.76 ppm; ³¹P{¹H} NMR (122 MHz, CDCl₃):

d=ꢀ7.92 ppm, ESI-MS (pos.): m/z: 299 [M+H]⁺, 321 [M+Na]⁺, 337

[M+K]⁺; HRMS (EI): m/z: calcd for C₁₁H₁₄O₄F₃P₁: 298.05818; found:

298.06175.

displayed a peak in the ¹H NMRD profiles, at magnetic

field strengths typically associated with slow tumbling mac-

romolecular assemblies. This suggests that an agglomeration

process is occurring, which was confirmed by

¹H DOSY NMR spectroscopy. However, the agglomeration

can be disrupted by the addition of phosphate ions returning

the systems to their monomeric state. ³¹P NMR spectroscopy

revealed that the phosphonate arm was reversibly binding

to the metal ion creating two distinct environments in solu-

tion. The pH-potentiometry showed the expected protona-

tion pattern of the ligands, whereas a reduction in the proto-

nation constant for the phosphonate groups upon formation

of the corresponding Gd³⁺complexes supported the idea of

the binding interaction. Furthermore, the complexes show a

similar demetallization rate to that of GdDTPA rendering

them suitable for further in vivo characterization for exam-

ple, in preclinical studies. The possibility of one of the sys-

tems to be used as an MRI contrast agent was investigated

through phantom measurements, where the ¹⁹F signals from

both the Yb³⁺and Gd³⁺complexes could be readily ob-

served. Moreover, the Gd³⁺complex showed the prospect

Diethyl 2-hydroxy-3-(trifluoromethyl) phenylphosphonate (3a): A flask

containing the phosphate ester 2a (4.30 g, 14.42 mmol) was purged of air

and kept under dinitrogen, dry tetrahydrofuran (30 mL) was added and

the solution was cooled to ꢀ788C. A solution of lithium diisopropyla-

mine also under a dinitrogen atmosphere was prepared over 30 min by

dissolving diisopropylamine (10.18 mL, 72.15 mmol) in tetrahyrofuran

(20 mL) and cooling this to ꢀ788C, followed by the addition of n-butyl

lithium (45.10 mL, 1.6m, 72.15 mmol). This was then transferred via can-

nula to the phosphate solution, which changed the solution to a caramel

color. After 3 h saturated ammonium solution was added to quench the

reaction and the solution was warmed to room temperature. The product

was extracted into diethyl ether; the organic phase was washed with satu-

rated sodium chloride solution and dried over sodium sulphate. The sol-

vent was removed under reduced pressure and the crude product was pu-

rified using column chromatography (silica gel, from 9:1 to 3:7 hexane/

ethyl acetate). On removal of the solvent 3a was isolated as an orange

1

liquid (3.34 g, 78%). Tlc: R_f=0.69 (silica, EtOAc/hexane=3:7); H NMR

(300 MHz, CDCl₃,): d=1.31 (t, 6H, ³J_H,H=7.0 Hz; POCH₂CH₃), 3.99–

4.24 (m, 4H; POCH₂CH₃), 6.94–7.00 (m, 1H; ArH), 7.52–7.59 (m, 1H;

ArH), 7.71 (d, 1H, ³J_H,H=7.7 Hz; ArH), 11.02 ppm (s, 1H; PhOH);

¹³C{¹H} NMR (75 MHz, CDCl₃,): d=15.98 (d, J_{C, P}=6.6 Hz; OCH₂CH₃),

1

of serving as a dual frequency H/¹⁹F probe, being detected

63.11 (d,

J_C,P=5.0 Hz; OCH₂CH₃), 109.82 (ArCH), 112.23 (ArCH),

118.77 (ArCH), 125.00 (ArC), 132.35 (q, J_C,F=2.2 Hz; CCF₃), 135.52

(ArC), 160.00 ppm (ArC); ¹⁹F{¹H} NMR (282 MHz, CDCl₃): d=

ꢀ63.22 ppm; ³¹P{¹H} NMR (122 MHz, CDCl₃): d=20.55 ppm; ESI-MS

(pos.): m/z: 299 [M+H]⁺, 321 [M+Na]⁺, 337 [M+K]⁺, (neg.): m/z: 297

[MꢀH]⁺; HRMS (EI): m/z: calcd for C₁₁H₁₄O₄F₃P₁: 298.05818; found:

298.05505.

at concentrations as low as 0.1 mm. This study demonstrates

the potential for these agents to be developed into multi-

functional MRI probes that induce MR signals at two fre-

quencies and provide a plenitude of critical information

1

gathered simultaneously from both H and ¹⁹F MRI.

Diethyl 2-(3-bromopropoxy)-3-(trifluoromethyl) phenylphosphonate

(4a): The phenol 3a (2.0 g, 6.71 mmol) was dissolved in anhydrous dime-

thylformamide (20 mL) and potassium carbonate (1.85 g, 13.41 mmol)

was added, this was heated to 608C for 1 hour under a dinitrogen atmos-

Experimental Section

phere, changing the solution to

a brown color. Dibromopropane

(2.05 mL, 20.19 mmol) was added dropwise and the mixture was heated

for a further 17 h. After cooling to room temperature the inorganic salts

were removed by filtration and dichloromethane was added. The organic

phase was washed with saturated sodium chloride solution and dried

over sodium sulphate, and the solvent was removed under reduced pres-

sure. The crude product was purified using column chromatography

(silica gel, from 9:1 to 4:6 hexane/ethyl acetate). Removal of the solvents

yielded 4a as a light yellow liquid (2.26 g, 81%). Tlc: R_f=0.49 (silica,

EtOAc/hexane=1:1); ¹H NMR (300 MHz, CDCl₃): d=1.36 (t, 6H,

³J_H,H=7.0 Hz; POCH₂CH₃), 2.43 (quin, 2H, ³J_H,H=6.4 Hz; BrCH₂CH₂),

3.63 (t, 2H, ³J_H,H=7.0 Hz; BrCH₂CH₂CH₂), 4.12–4.25 (m, 4H;

POCH₂CH₃), 4.30 (t, 2H, ³J_H,H=5.9 Hz; BrCH₂), 7.28–7.33 (m, 1H;

ArH), 7.80 (d, 1H, ³J_H,H=7.0 Hz; ArH), 7.99–8.07 ppm (m, 1H; ArH);

¹³C{¹H} NMR (75 MHz, CDCl₃): d=16.10, 29.33, 33.36, 62.40, 74.89,

121.10, 123.85, 124.86, 126.39, 131.62, 138.47, 159.28 ppm; ¹⁹F{¹H} NMR

(282 MHz, CDCl₃) d=ꢀ60.40 ppm; ³¹P{¹H} NMR (122 MHz, CDCl₃): d=

14.38 ppm; ESI-MS (pos.): m/z: 420 [M+H]⁺, 442 [M+Na]⁺; HRMS

(EI): m/z: calcd for C₁₄H₂₀O₄F₃P₁Br₁: 420.013596; found: 420.01163.

Synthesis: All chemicals were purchased from commercial sources and

used without further purification, with the exception of THF, which was

distilled prior to use. Tris tert-butyl-DO3A was prepared according to

previously published procedures.^[53]Column chromatography was per-

formed using silica gel 60 (70–230 mesh ASTM) or aluminum oxide 90

active basic from Merck (Germany). Reversed-phase high-performance

liquid chromatography was performed on a Varian PrepStar Instrument

(Australia), equipped with PrepStar SD-1 pump heads. Analytical RP-

HPLC was performed on an Atlantis C18 column 4.6 mm ꢅ 150 mm, par-

ticle size 5 mm, using method A (Table S1 in the Supporting Information).

Preparative RP-HPLC was performed on an Atlantis C18 column 19 mm

ꢅ 150 mm, particle size 5 mm, Waters corporation (USA) using method B

and C (Tables S2 and S3 in the Supporting Information).

Diethyl 2-(trifluoromethyl)phenyl phosphate (2a): 2-(Trifluoromethyl)

phenol 1a (5.00 g, 30.8 mmol) was dissolved in carbon tetrachloride

(30 mL) and this was cooled to 08C, triethylamine (5.37 mL, 38.6 mmol)

and diethyl phosphite (4.93 mL, 38.6 mmol) were added. After warming

the solution to room temperature a white precipitate formed and this

was left to stir at this temperature overnight. Water was added followed

by dichloromethane and the organic phase was washed with HCl (1m),

saturated sodium chloride solution, dried over sodium sulphate and the

solvent was removed under reduced pressure. The crude product was pu-

rified using column chromatography (silica gel, from 9:1 to 1:1 hexane/

ethyl acetate). Removal of the solvent yielded 2a as a light yellow oil

(8.89 g, 97%). Tlc: R_f=0.32 (silica, hexane/ethyl acetate=7:3); ¹H NMR

(300 MHz, CDCl₃): d=1.33 (td, J_H,H=7.2, 1.1 Hz, 6H; OCH₂CH₃), 3.99–

4.24 (m, 4H; OCH₂CH₃), 7.20 (t, J_H,H=7.9 Hz, 1H; ArH), 7.54–7.59 (m,

2H; ArH), 7.60 ppm (m, 2H; ArH); ¹³C{¹H} NMR (75 MHz, CDCl₃): d=

15.80 (d, J_C,P=6.6 Hz; OCH₂CH₃), 64.89 (d, J_C,P=6.6 Hz; OCH₂CH₃),

Tri-tert-butyl 2,2’,2’’-(10-(3-(2-(diethoxyphosphoryl)-6-(trifluoromethyl)-

phenoxy)propyl)-1,4,7,10-tetraazacyclododecane-1,4,7-triyl)triacetate

(5a): Tris tert-butyl-DO3A (2.80 g, 4.70 mmol), the phosphonate arm 4a

(2.96 g, 7.05 mmol) and potassium carbonate (1.95 g, 14.10 mmol) were

dissolved in acetonitrile (40 mL) and the mixture was heated to 708C

under a dinitrogen atmosphere. After 48 h the solution was cooled to

room temperature, the inorganic salts were removed by filtration and the

solvent was evaporated under reduced pressure. The resulting yellow oil

was dissolved in dichloromethane and washed with saturated sodium

chloride solution, dried over sodium sulphate and the solvent evaporated.

The crude product was purified using column chromatography (alumina

from 100% dichloromethane to 96:4 dichloromethane/methanol). On re-

moval of the solvents 5a was obtained as a yellow solid (3.19 g, 80%).

120.21

ACHTUNGTRENNUNG(ArCH), 121.17 (ArC), 124.19 (ArCH), 124.72 (q, J_C,F=5.5 Hz;

CCF₃), 128.33 (ArCH), 133.35 (ArCH), 148.55 ppm (ArC); ¹⁹F{¹H} NMR

&

12

&

www.chemeurj.org

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 0000, 00, 0 – 0

ÝÝ

These are not the final page numbers!

Dual Frequency MRI Probes

FULL PAPER

Tlc: R_f=0.53 (alumina, dichloromethane/methanol=92.5:7.5); ¹H NMR

(300 MHz, CDCl₃): d=1.35 (t, 6H, J_H,H=7.0 Hz; POCH₂CH₃), 1.43–1.45

fied using column chromatography (silica gel, from 9:1 to 1:9 hexane/

ethyl acetate). Removal of the solvent yielded 2b as a light yellow oil

(8.09 g, 88%). Tlc: R_f=0.47 (silica, hexane/ethyl acetate=7:3); ¹H NMR

(300 MHz, CDCl₃): d=1.35 (td, J_H,H=7.2, 0.9 Hz, 6H; OCH₂CH₃); 4.17–

4.27 (m, 4H; OCH₂CH₃), 7.33 (dd, J_H,H=8.9 Hz, 2H; ArH), 7.60 ppm

(dd, J_H,H=8.7 Hz, 2H; ArH); ¹³C{¹H} NMR (75 MHz, CDCl₃): d=15.87

(d, J_C,P=6.6 Hz; OCH₂CH₃), 64.76 (d, J_C,P=6.1 Hz; OCH₂CH₃), 120.2

(ArCH), 121.95 (ArCH), 125.55 (ArC), 126.97 (q, J_C,F=3.9 Hz; CCF₃),

3

(m, 27H; CCH₃), 1.96–4.25 (brm, 34H; CH₂ring, NCH₂COO,

NCH₂CH₂CH₂O, POCH₂CH₃), 7.30–7.33 (m, 1H; ArH), 7.79 (d, 1H,

³J_H,H=7.7 Hz; ArH), 7.94–8.04 ppm (m, 1H; ArH); ¹³C{¹H} NMR

(75 MHz, CDCl₃): d=15.89, 27.35, 49.73, 50.24, 51.67, 51.77, 55.31, 56.20,

62.00, 74.91, 80.19, 81.96, 82.32, 123.57, 123.75, 125.99, 131.31, 137.98,

159.19, 170.50, 172.15, 172.99 ppm; ¹⁹F{¹H} NMR (282 MHz, CDCl₃): d=

ꢀ60.24, ꢀ60.27, ꢀ60.29, ꢀ60.36 ppm; ³¹P{¹H} NMR (122 MHz, CDCl₃):

d=13.45, 14.05 ppm; ESI-MS (pos.): m/z: 853.5 [M+H]⁺, 875.4 [M+Na]⁺

; HRMS (FT-ICR): m/z: calcd for C₄₀H₆₉F₃N₄O₁₀P₁: 853.46979 [M+H]⁺;

found: 853.46590.

153.21 ppmACTHUNGTRENNUNG

³¹P{¹H} NMR (122 MHz, CDCl₃): d=ꢀ6.77 ppm; ESI-MS (pos.): m/z:

299 [M+H]⁺, 321 [M+Na]⁺, 337 [M+K]⁺; HRMS (EI): m/z: calcd for

C₁₁H₁₄O₄F₃P₁: 298.05818; found: 298.06124.

2,2’,2’’-(10-(3-(2-Phosphono-6-(trifluoromethyl)phenoxy)propyl)-1,4,7,10-

tetraazacyclododecane-1,4,7-triyl)triacetic acid (L¹): The ester 5a (1.2 g,

1.41 mmol) was dissolved in dimethylformamide (3 mL), this was cooled

to 08C in an ice bath and bromotrimethylsilane (1.86 mL, 1.41 mmol)

was added dropwise to the solution. This was warmed to room tempera-

ture and stirred for 18 h. The solvent was evaporated under reduced pres-

sure and the removal of the ethyl groups was confirmed through electro-

spray mass spectrometry. A mixture of dichloromethane and trifluoroace-

tic acid (1:1, 10 mL) was added to the crude orange oil and the mixture

was stirred at room temperature for 48 h. Following removal of the sol-

vents a brown oil was obtained and electrospray mass spectrometry again

confirmed the removal of the tert-butyl groups. The crude product was

purified through reverse phase HPLC. On freeze drying L¹was isolated

as a flocculent white solid that quickly became a yellow oil due to its hy-

Diethyl 2-hydroxy-5-(trifluoromethyl) phenylphosphonate (3b): A flask

containing the phosphate ester 2b (8.00 g, 26.80 mmol) was purged of air

and kept under dinitrogen, dry tetrahydrofuran (30 mL was added and

the solution was cooled to ꢀ788C. A solution of lithium diisopropyla-

mine also under a dinitrogen atmosphere, was prepared over 30 min by

dissolving diisopropylamine (11.36 mL, 80.40 mmol) in tetrahyrofuran

(20 mL) and cooling this to ꢀ788C, followed by the addition of n-butyl

lithium (50.25 mL, 1.6m, 80.4 mmol). This was then transferred via cannu-

la to the phosphate solution, which changed the solution to a caramel

color. After 3 h saturated ammonium solution was added to quench the

reaction and the solution was warmed to room temperature. The product

was extracted into diethyl ether; the organic phase was washed with satu-

rated sodium chloride solution and dried over sodium sulphate. The sol-

vent was removed under reduced pressure and the crude product was pu-

rified using column chromatography (silica gel, from 9:1 to 3:7 hexane/

ethyl acetate). On removal of the solvent 3b was isolated as an orange

oil (5.71 g, 71%). Tlc: R_f=0.50 (silica, EtOAc/hexane=3:7); ¹H NMR

(300 MHz, CDCl₃,): d=1.27 (t, 6H, ³J_H,H=7.2 Hz; POCH₂CH₃), 3.94–

4.18 (m, 4H; POCH₂CH₃), 6.94–6.99 (m 1H; ArH), 7.53–7.59 (m, 2H;

ArH), 10.59 ppm (s, 1H; PhOH); ¹³C{¹H} NMR (75 MHz, CDCl₃,): d=

16.04 (d, J_C,P=6.6 Hz; OCH₂CH₃), 63.13 (d, J_C,P=5.0 Hz; OCH₂CH₃),

108.35 (ArCH), 110.77 (ArCH), 118.26 (ArCH), 121.98 (ArC), 129.15

(CCF₃), 131.81 (ArC), 164.50 ppm (ArC); ¹⁹F{¹H} NMR (282 MHz,

CDCl₃): d=ꢀ61.93 ppm; ³¹P{¹H} NMR (122 MHz, CDCl₃): d=

20.18 ppm; ESI-MS (pos.): m/z: 299 [M+H]⁺, 321 [M+Na]⁺, 337

[M+K]⁺, (neg.): m/z: 297 [MꢀH]^ꢀ. HRMS (EI): m/z: calcd for

C₁₁H₁₄O₄F₃P₁: 298.05818; found: 298.05822.

1

groscopic nature (0.746 g, 82%). H NMR (300 MHz, D₂O): d=2.24–4.20

(brm, 28H; CH₂ring, NCH₂COO, NCH₂CH₂CH₂O,), 7.23 (m, 1H;

ArH), 7.68 (d, ³J_H,H=7.7 Hz, 1H; ArH), 7.83–7.93 ppm (m, 1H; ArH),

¹³C{¹H} NMR (75 MHz, D₂O): d=15.75, 23.34, 31.29, 36.80, 47.92, 48.18,

48.29, 49.63, 51.36, 52.44, 53.37, 56.34, 72.81, 121.82, 123.95, 130.21,

132.49, 137.72, 157.29, 164.78, 169.69, 173.06, 174.45 ppm; ¹⁹F{¹H} NMR

(282 MHz, D₂O): d=ꢀ60.10, ꢀ60.13, ꢀ60.19, ꢀ60.22 ppm; ³¹P{¹H} NMR

(122 MHz, D₂O): d=7.39, 9.47 ppm; ESI-MS (pos.): m/z: 629.3 [M+H]⁺,

657.3 [M+Na]⁺; HRMS (FT-ICR): m/z: calcd for C₂₄H₃₅F₃N₄O₁₀P₁:

627.20484 [MꢀH]^ꢀ; found: 627.20454.

2,2’,2’’-(10-(3-(2-Phosphono-6-(trifluoromethyl)phenoxy)propyl)-1,4,7,10-

tetraazacyclododecane-1,4,7-triyl)triacetic acid lanthanide complex

(LnL¹): All the complexes were prepared according to the following pro-

cedure: The ligand (L^1–4) and 1 equivalent of the lanthanide chloride

were dissolved in water. The solution was heated to 708C and the pH

was raised from 2–3 to 5–6 using sodium hydroxide (1m) solution. This

temperature and pH was maintained over a period of 24 h. After cooling

to room temperature the excess lanthanide ions were removed by raising

the pH to 10 and treating the supernatant with chelex-100. The xylenol

orange test was used to confirm the absence of any free lanthanide ions

in solution. The water was removed under reduced pressure to yield a

yellow solid.

YL¹: ESI-MS (neg.): m/z: 713.1 [MꢀH]^ꢀ. EuL¹: ¹H NMR (300 MHz,

D₂O): d=ꢀ20.70, ꢀ17.43, ꢀ15.31, ꢀ13.54, ꢀ11.68, ꢀ10.71, ꢀ8.58, ꢀ7.25,

ꢀ5.40, ꢀ2.74, ꢀ0.088, 7.34, 8.49, 13.45, 19.73, 25.66, 28.84, 32.20, 39.19,

46.01 ppm, ¹⁹F{¹H} NMR (282 MHz, D₂O, 38 8C): d=ꢀ59.43,

ꢀ59.68 ppm; ³¹P{¹H} NMR (122 MHz, CDCl₃, 258C): d=ꢀ75.22,

6.33 ppm; ESI-MS (neg.): m/z: 776.8 [MꢀH]^ꢀ. GdL¹: ¹⁹F{¹H} NMR

(282 MHz, D₂O, 38 8C): d=ꢀ62.27, ꢀ62.53 ppm; ESI-MS (neg.): m/z:

782.1 [MꢀH]^ꢀ; r₁7.12 mm^ꢀ1s^ꢀ1(pH 7.4, MOPS, 300 MHz). TbL¹:

¹⁹F{¹H} NMR (282 MHz, CDCl₃, 258C): d=ꢀ62.73 ppm; ESI-MS (neg.):

m/z: 783.1 [MꢀH]^ꢀ. YbL¹: ¹⁹F{¹H} NMR (282 MHz, CDCl₃, 258C): d=

ꢀ60.03, ꢀ60.08, ꢀ63.11 ppm; ESI-MS (neg.): m/z: 797.8 [MꢀH]^ꢀ.

Diethyl 4-(trifluoromethyl) phenyl phosphate (2b): 4-(Trifluoromethyl)-

phenol 1b (5.00 g, 30.8 mmol) was dissolved in carbon tetrachloride

(30 mL) and this was cooled to 08C, triethylamine (5.37 mL, 38.6 mmol)

and diethyl phosphite (4.93 mL, 38.6 mmol) were added. After warming

the solution to room temperature a white precipitate formed and this

was left to stir at this temperature overnight. Water was added followed

by dichloromethane and the organic phase was washed with 1m HCl, sa-

turated sodium chloride solution, dried over sodium sulphate and the sol-

vent was removed under reduced pressure. The crude product was puri-

Diethyl 2-(3-bromopropoxy)-5-(trifluoromethyl) phenylphosphonate

(4b): The phenol 3b (2.0 g, 6.71 mmol) was dissolved in anhydrous dime-

thylformamide (20 mL) and potassium carbonate (1.85 g, 13.41 mmol)

was added, this was heated to 608C for 1 hour under a dinitrogen atmos-

phere, changing the solution to

a brown color. Dibromopropane

(2.05 mL, 20.19 mmol) was added dropwise and the mixture was heated

for a further 17 h. After cooling to room temperature the inorganic salts

were removed by filtration and dichloromethane was added. The organic

phase was washed with saturated sodium chloride solution, dried over

sodium sulphate and the solvent was removed under reduced pressure.

The crude product was purified using column chromatography (silica gel,

from 9:1 hexane/ethyl acetate to 100% ethyl acetate). Removal of the

solvents yielded 4b as a light yellow oil (1.82 g, 65%). Tlc: R_f=0.29

(silica, EtOAc/hexane=1:1); ¹H NMR (300 MHz, CDCl₃): d=1.33 (t,

6H, ³J_H,H=7.0 Hz; POCH₂CH₃), 2.36 (quin, 2H, ³J_H,H=5.9 Hz;

BrCH₂CH₂); 3.73 (t, 2H, ³J_H,H=6.2 Hz; BrCH₂CH₂CH₂), 4.04–4.27 (m,

3

6H, POCH₂CH₃; BrCH₂), 7.00–7.05 (m, 1H; ArH), 7.76 (d, 1H, J_H,H

=

8.7 Hz; ArH), 8.12 ppm (d, 1H, J=15.3 Hz; ArH); ¹³C{¹H} NMR

(75 MHz, CDCl₃): d=16.14, 29.52, 31.69, 62.14, 65.71, 69.18, 111.70,

112.11, 117.97, 118.61, 122.44, 125.50, 131.39, 132.45, 162.38 ppm;

¹⁹F{¹H} NMR (282 MHz, CDCl₃): d=ꢀ61.78 ppm; ³¹P{¹H} NMR

(122 MHz, CDCl₃); d=14.46, 14.85 ppm; ESI-MS (pos.): m/z: 420

[M+H]⁺, 442 [M+Na]⁺; HRMS (EI): m/z: for C₁₄H₂₀O₄F₃P₁Br₁:

420.013596; found: 420.01622.

Tri-tert-butyl 2,2’,2’’-(10-(3-(2-(diethoxyphosphoryl)-4-(trifluoromethyl)-

phenoxy)propyl)-1,4,7,10-tetraazacyclododecane-1,4,7-triyl)triacetate

(5b): Tris tert-butyl-DO3A (1.51 g, 2.54 mmol), the phosphonate arm 4b

(1.62 g, 3.86 mmol) and potassium carbonate (1.05 g, 7.61 mmol) were

dissolved in acetonitrile (20 mL) and the mixture was heated to 708C

Chem. Eur. J. 2013, 00, 0 – 0

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

&

13

&

ÞÞ

These are not the final page numbers!

G. Angelovski et al.

under a dinitrogen atmosphere. After 48 h the solution was cooled to

room temperature, the inorganic salts were removed by filtration and the

solvent was evaporated under reduced pressure. The resulting yellow oil

was dissolved in dichloromethane and washed with saturated sodium

chloride solution, dried over sodium sulphate and the solvent evaporated.

The crude product was purified using column chromatography (alumina

from 100% dichloromethane to 96:4 dichloromethane/methanol). On re-

moval of the solvents 5b was isolated as a yellow solid (1.89 g, 87%).

Tlc: R_f=0.33 (alumina, dichloromethane/methanol=92.5:7.5); ¹H NMR

acetate to 100% ethyl acetate). Removal of the solvent yielded 4c as a

yellow oil. (4.05 g, 68%). Tlc: R_f=0.35 (silica, EtOAc/hexane 1:1);

3

¹H NMR (300 MHz, CDCl₃): d=1.30 (t, 6H, J_H,H=7.0 Hz; POCH₂CH₃),

4.09–4.22 (m, 4H; POCH₂CH₃), 4.52 (s, 2H; CCH₂O), 5.25 (s, 2H;

CCH₂Br), 7.30–7.44 (m, 3H; ArH), 7.58 (d, 1H, ³J_H,H=8.31 Hz; ArH),

7.85 (d, 1H, ³J_H,H=7.00 Hz; ArH), 8.03–8.16 ppm (m, 1H; ArH);

¹³C{¹H} NMR (75 MHz, CDCl₃): d=16.23, 33.27, 62.66, 121.47, 124.23,

125.63, 127.29, 128.84, 131.80, 137.03, 137.50, 138.53, 159.16, 160.71 ppm;

¹⁹F{¹H} NMR (282 MHz, CDCl₃): d=ꢀ60.12, ꢀ60.10 ppm; ³¹P{¹H} NMR

(122 MHz, CDCl₃): d=14.01, 16.02 ppm; ESI-MS (pos.): m/z: 482

[M+H]⁺; HRMS (EI): m/z: calcd for C₁₉H₂₁O₄F₃P₁Br₁: 482.029246;

found: 482.02574.

3

(300 MHz, CDCl₃): d=1.30 (t, 6H, J_HH=7.0 Hz; POCH₂CH₃), 1.40–1.42

(m, 27H; CCH₃), 1.99–4.30 (brm, 34H; CH₂ring, NCH₂COO,

NCH₂CH₂CH₂O, POCH₂CH₃), 7.12–7.17 (m, 1H; ArH), 7.72–7.74 (m,

1H; ArH), 8.00 ppm (d, 1H, J=15.11 Hz; ArH); ¹³C{¹H} NMR (75 MHz,

CDCl₃): d=16.11, 24.75, 27.55, 49.62, 50.23, 51.79, 55.35, 61.89, 67.06,

80.29, 82.32, 111.92, 115.79, 118.29, 121.98, 131.25, 132.25, 162.42, 170.62,

172.30, 173.37 ppm; ¹⁹F{¹H} NMR (282 MHz, CDCl₃): d=ꢀ61.77,

ꢀ61.65 ppm; ³¹P{¹H} NMR (122 MHz, CDCl₃): d=14.54, 15.05 ppm; ESI-

MS (pos.): m/z: 853.5 [M+H]⁺; HRMS (FT-ICR): m/z: calcd for

C₄₀H₆₉F₃N₄O₁₀P₁: 853.46979 [M+H]⁺; found: 853.46965.

Tri-tert-butyl 2,2’,2’’-(10-(4-((2-(diethoxyphosphoryl)-6-(trifluoromethyl)-

phenoxy)methyl)benzyl)-1,4,7,10-tetraazacyclododecane-1,4,7-triyl)triace-

tate (5c): Tris tert-butyl-DO3A (3.98 g, 6.69 mmol), 4c (4.0 g, 8.36 mmol)

and potassium carbonate (2.77 g, 20.06 mmol) were added to dimethylfor-

mamide (20 mL) and the mixture was heated to 608C under a dinitrogen

atmosphere for 48 h. The solvent was removed under reduced pressure,

and dichloromethane was added. Following the removal of the inorganic

salts the organic phase was washed with saturated sodium chloride solu-

tion and dried over sodium sulphate. Removal of the solvent yielded an

orange oil. The crude product was purified using column chromatography

(alumina from 100% dichloromethane to 93:7 dichloromethane/metha-

nol). Removal of the solvents yielded 5c as a yellow solid (4.77 g, 78%).

Tlc: R_f=0.25 (alumina, dichloromethane/methanol=95:5); ¹H NMR

(300 MHz, CDCl₃); d=1.20–1.40 (m, 33H; POCH₂CH₃, CCH₃), 2.16–

4.10 (brm, 34H; CH₂ring, NCH₂COO, NCH₂CH₂CH₂O, POCH₂CH₃),

5.16 (s, 2H; CCH₂N), 7.25–7.50 (m, 5H; ArH), 7.76–7.79 (m, 1H; ArH),

7.97–8.02 ppm (m, 1H; ArH); ¹³C{¹H} NMR (75 MHz, CDCl₃): d=14.90,

15.96, 27.57, 49.64, 51.18, 55.33, 33.66, 59.19, 62.29, 65.42, 80.26, 82.04,

82.50, 126.57, 126.86, 128.00, 128.26, 129.63, 131.49, 135.90, 136.84, 138.25,

158.74, 170.72, 172.16, 173.16 ppm; ¹⁹F{¹H} NMR (282 MHz, CDCl₃): d=

ꢀ60.19, ꢀ60.16, ꢀ60.11, ꢀ60.02 ppm; ³¹P{¹H} NMR (122 MHz, CDCl₃):

d=13.92, 14.20 ppm; ESI-MS (pos.): m/z: 915.5 [M+H]⁺, 937.5

[M+Na]⁺; HRMS (FT-ICR): m/z: calcd for C₄₅H₇₁F₃N₄O₁₀P₁: 915.48544

[M+H]⁺; found: 915.48555.

2,2’,2’’-(10-(3-(2-Phosphono-4-(trifluoromethyl)phenoxy)propyl)-1,4,7,10-

tetraazacyclododecane-1,4,7-triyl)triacetic acid (L²): The ester 5b

(0.948 g, 1.11 mmol) was dissolved in dimethylformamide (5 mL), this

was cooled to 08C in an ice bath and bromotrimethylsilane (0.733 mL,

5.56 mmol) was added dropwise to the solution. This was warmed to

room temperature and stirred for 18 h. The solvent was evaporated

under reduced pressure and the removal of the ethyl groups was con-

firmed through electrospray mass spectrometry. A mixture of dichloro-

methane and trifluoroacetic acid (1:1, 10 mL) was added to the crude

orange oil and the mixture was stirred at room temperature for 48 h. Fol-

lowing removal of the solvents a brown oil was obtained and electrospray

mass spectrometry again confirmed the removal of the tert-butyl groups.

The crude product was purified through reverse phase HPLC. On freeze

drying the L²was isolated as a flocculent white solid that quickly became

a yellow oil due to its hygroscopic nature (0.664 g, 95%). ¹H NMR

(300 MHz, D₂O): d=2.21–4.12 (brm, 28H; CH₂ring, NCH₂COO,

NCH₂CH₂CH₂O,), 7.03–7.87 ppm (m, 3H; ArH); ¹³C{¹H} NMR (75 MHz,

D₂O): d=22.69, 31.31, 36.80, 48.42, 49.79, 51.41, 53.31, 56.03, 65.23,

110.57, 111.89, 114.44, 118.32, 118.87, 121.75, 122.43, 124.45, 126.02,

129.84, 161.78, 162.27, 162.74, 164.74, 169.64, 172.99, 174.48, 174.45 ppm;

¹⁹F{¹H} NMR (282 MHz, D₂O): d=ꢀ61.18 ppm; ³¹P{¹H} NMR (122 MHz,

D₂O): d=8.78 ppm, ESI-MS (pos.): m/z: 629.3 [M+H]⁺, 657.3 [M+Na]⁺;

HRMS (FT-ICR): m/z: calcd for C₂₄H₃₅F₃N₄O₁₀P₁: 627.20484 [MꢀH]^ꢀ;

found: 627.20466.

2,2’,2’’-(10-(4-((2-Phosphono-6-(trifluoromethyl)phenoxy)methyl)benzyl)-

1,4,7,10-tetraazacyclododecane-1,4,7-triyl)triacetic acid (L³): The ester 5c

(1.0 g, 1.09 mmol) was dissolved in dimethylformamide (5 mL) and

cooled to 08C in an ice bath, bromotrimethylsilane (0.72 mL, 5.46 mmol)

was added slowly and the solution was stirred at room temperature for

18 h. The absence of the ethyl esters was confirmed by electrospray mass

spectrometry. A mixture of dichloromethane and trifluoroacetic acid (1:1,

10 mL) was added to the crude orange oil and the mixture was stirred at

room temperature for 48 h. Following removal of the solvents a brown

oil was obtained and electrospray mass spectrometry again confirmed the

removal of the tert-butyl groups. The crude product was purified through

reverse phase HPLC. On freeze drying L³was isolated as a flocculent

white solid that quickly became a yellow oil due to its hygroscopic nature

(0.661 g, 88%).¹H NMR (300 MHz, D₂O): d=2.75–3.85 (brm, 28H; CH₂

ring, NCH₂COO, NCH₂CH₂CH₂O,), 7.07–7.88 (m, 3H; ArH) ppm;

¹³C{¹H} NMR (75 MHz, D₂O): d=29.57, 48.18, 49.40, 54.66, 56.88, 77.15,

121.86, 124.30, 125.47, 129.37, 130.83, 133.12, 138.13, 157.43, 172.05 ppm;

¹⁹F{¹H} NMR (282 MHz, D₂O): d=ꢀ62.23, ꢀ59.91 ppm; ³¹P{¹H} NMR

(122 MHz, D₂O): d=7.36, 12.10 ppm; ESI-MS (neg.): m/z: 689.1

[MꢀH]^ꢀ; HRMS (FT-ICR): m/z: calcd for C₂₉H₃₇F₃N₄O₁₀P₁: 689.22049

[MꢀH]^ꢀ; found: 689.22093.

2,2’,2’’-(10-(3-(2-Phosphono-4-(trifluoromethyl)phenoxy)propyl)-1,4,7,10-

tetraazacyclododecane-1,4,7-triyl)triacetic acid lanthanide complex

(LnL²)

EuL²: ¹H NMR (300 MHz, D₂O): d=ꢀ20.88, ꢀ19.15, ꢀ17.51, ꢀ16.26,

ꢀ14.15, ꢀ12.13, ꢀ10.97, ꢀ7.99, ꢀ6.16, ꢀ3.66, ꢀ2.41, ꢀ1.44, 0.86, 1.35,

1.83, 2.89, 3.56, 6.83, 10.39, 13.28, 20.79, 26.85, 28.49, 30.60, 42.34 ppm;

¹⁹F{¹H} NMR (282 MHz, D₂O, 38 8C): d=ꢀ64.15 ppm; ³¹P{¹H} NMR

(122 MHz, CDCl₃, 258C): d=ꢀ130.02, 9.28 ppm; ESI-MS (neg.): m/z:

776.8 [MꢀH]^ꢀ. GdL²: ¹⁹F{¹H} NMR (282 MHz, D₂O, 38 8C): d=ꢀ62.27,

ꢀ62.53 ppm; ESI-MS (neg.): m/z: 782.0 [MꢀH]^ꢀ; r₁6.64 mm^ꢀ1s^ꢀ1(pH 7.4,

MOPS, 300 MHz). TbL²: ¹⁹F{¹H} NMR (282 MHz, CDCl₃, 258C): d=

ꢀ62.73 ppm; ESI-MS (neg.): m/z: 783.0 [MꢀH]^ꢀ. YbL²: ¹⁹F{¹H} NMR

(282 MHz, CDCl₃, 258C): d=ꢀ60.65, ꢀ76.63 ppm; ESI-MS (neg.): m/z:

797.8 [MꢀH]^ꢀ.

Diethyl (2-((4-(bromomethyl) benzyl) oxy)-3-(trifluoromethyl) phenyl)

phosphonate (4c): The phenol 3a (3.70 g, 12.41 mmol) was dissolved in

anhydrous dimethylformamide (100 mL) and potassium carbonate was

added (3.43 g, 24.82 mmol) and this was heated to 608C under a dinitro-

gen atmosphere. After 1 hour this was added dropwise to a solution of

1,4-bis (bromomethyl) benzene (32.8 g, 124 mmol) in dimethylformamide

(50 mL) and the mixture was heated for 18 h. The solvent was removed

under reduced pressure, and diethyl ether was added precipitating most

of the bromide starting material from solution which was removed by fil-

tration. Evaporation of diethyl ether yielded an orange solid which was

purified using column chromatography (silica gel, from 9:1 hexane/ethyl

2,2’,2’’-(10-(4-((2-Phosphono-6-(trifluoromethyl)phenoxy)methyl)benzyl)-

1,4,7,10-tetraazacyclododecane-1,4,7-triyl)triacetic acid lanthanide com-

plex (LnL³)

YL³: ESI-MS (neg.): m/z: 774.9 [MꢀH]^ꢀ. EuL³¹H NMR (300 MHz,

D₂O, 25 8C): d=ꢀ25.24, ꢀ21.68, ꢀ19.67, ꢀ18.74, ꢀ14.00, ꢀ13.53, ꢀ13.09,

ꢀ12.22, ꢀ11.74, ꢀ7.56, ꢀ4.90, ꢀ3.51, ꢀ2.09, 1.13, 3.29, 6.91, 7.52, 8.14,

9.22, 12.83, 14.93, 26.45, 30.99, 47.50 ppm; ¹⁹F{¹H} NMR (282 MHz, D₂O,

258C): d=ꢀ64.83, ꢀ63.81 ppm; ³¹P{¹H} NMR (122 MHz, CDCl₃, 258C):

d=ꢀ118.42, 13.59 ppm; ESI-MS (neg.): m/z: 838.8 [MꢀH]^ꢀ. GdL³:

¹⁹F{¹H} NMR (282 MHz, D₂O, 25 8C): d=ꢀ64.38, ꢀ64.27 ppm; ESI-MS

(neg.): m/z: 844.1 [MꢀH]^ꢀ, 866.1 [Mꢀ2H+Na]^ꢀ; r₁6.12 mm^ꢀ1s^ꢀ1(pH 7.4,

&

14

&

www.chemeurj.org

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 0000, 00, 0 – 0

ÝÝ

These are not the final page numbers!

Dual Frequency MRI Probes

FULL PAPER

MOPS, 300 MHz). TbL³: ¹⁹F{¹H} NMR (282 MHz, D₂O, 25 8C): d=

ꢀ62.25 ppm; ESI-MS (neg.): m/z: 845.1 [MꢀH]^ꢀ. YbL³: ¹⁹F{¹H} NMR

(282 MHz, D₂O, 25 8C): d=ꢀ55.69, ꢀ58.63, ꢀ61.60 ppm; ESI-MS (neg.):

m/z: 859.8 [MꢀH]^ꢀ.

2H, CCH₂N), 7.07 (t, ³J_H,H=6.23 Hz, 1H; ArH), 7.43–7.51 (m, 4H;

ArH), 7.67 (d, ³J_H,H=7.9 Hz, 1H; ArH), 7.96 ppm (d, ³J_H,H=14.5 Hz,

1H; ArH); ¹³C{¹H} NMR (75 MHz, D₂O): d=28.72, 28.79, 51.15, 53.36,

54.98, 56.37, 58.64, 71.14, 83.59, 85.79, 114.70, 114.82, 120.56, 123.09,

123.69, 123.90, 127.28, 129.19, 131.93, 132.51, 133.00, 139.92, 163.78,

166.77, 172.52 ppm; ¹⁹F{¹H} NMR (282 MHz, D₂O): d=ꢀ61.24 ppm;

³¹P{¹H} NMR (122 MHz, D₂O): d=8.61 ppm; ESI-MS (neg.): m/z: 689.1

[MꢀH]^ꢀ; HRMS (FT-ICR): m/z calcd for C₂₉H₃₇F₃N₄O₁₀P₁: 689.22049

[MꢀH]^ꢀ; found: 689.22028.

Diethyl (2-((4-(bromomethyl) benzyl) oxy)-5-(trifluoromethyl) phenyl)

phosphonate (4d): The phenol 3b (2.80 g, 9.39 mmol) was dissolved in

anhydrous dimethylformamide (100 mL) and potassium carbonate was

added (2.60 g, 18.78 mmol) and this was heated to 608C under a dinitro-

gen atmosphere. After 1 hour this was added dropwise to a solution of

1,4-bis (bromomethyl) benzene (24.79 g, 93.92 mmol) in dimethylforma-

mide (50 mL) and the mixture was heated for 18 h. The solvent was re-

moved under reduced pressure, and diethyl ether was added precipitating

most of the bromide starting material from solution which was removed

by filtration. Evaporation of diethyl ether yielded an orange solid which

was purified using column chromatography (silica gel, from 9:1 hexane/

ethyl acetate to 100% ethyl acetate). Removal of the solvent yielded 4d

as a white solid. (3.37 g, 75%). Tlc: R_f=0.42 (silica, EtOAc/hexane=

1:1); ¹H NMR (300 MHz, CDCl₃); d=1.28 (t, 6H, ³J_H,H=7.2 Hz;

POCH₂CH₃), 4.02–4.23(m, 4H; POCH₂CH₃), 4.48 (s, 2H; CCH₂O), 5.22

(s, 2H; CCH₂Br), 7.00–7.05 (m, 1H; ArH), 7.41 (d, 1H, ³J_H,H=8.3 Hz;

ArH), 7.48 (d, 1H, ³J_H,H=8.3 Hz; ArH), 7.70 (d, 1H, ³J_H,H=8.7 Hz;

ArH), 8.12 ppm (d, 1H, ³J_H,H=15.3 Hz; ArH), ¹³C{¹H} NMR (75 MHz,

CDCl₃): d=16.22, 32.86, 62.38, 70.00, 112.26, 116.92, 119.43, 121.98,

122.60, 123.05, 123.24, 127.30, 129.21, 131.27, 132.43, 135.80, 137.69,

2,2’,2’’-(10-(4-((2-Phosphono-4-(trifluoromethyl)phenoxy)methyl)benzyl)-

1,4,7,10-tetraazacyclododecane-1,4,7-triyl)triacetic acid lanthanide com-

plex (LnL⁴)

EuL⁴: ¹H NMR (300 MHz, D₂O): d=ꢀ20.91, ꢀ18.48, ꢀ14.40, ꢀ12.06,

ꢀ10.38, ꢀ9.81, ꢀ8.58, ꢀ7.63, ꢀ4.55, ꢀ3.89, ꢀ2.48, ꢀ0.08, 1.25, 7.33, 9.73,

10.56, 15.22, 27.07, 21.93, 24.93, 26.15, 27.65, 31.58, 33.56, 37.31, 39.63,

42.20, 48.05 ppm; ¹⁹F{¹H} NMR (282 MHz, D₂O, 25 8C): d=ꢀ60.71,

ꢀ61.08 ppm; ³¹P{¹H} NMR (122 MHz, CDCl₃, 258C): d=ꢀ139.50,

ꢀ112.34, 11.93 ppm; ESI-MS (neg.): m/z: 838.8 [MꢀH]^ꢀ. GdL⁴:

¹⁹F{¹H} NMR (282 MHz, D₂O, 25 8C): d=ꢀ60.45 ppm; ESI-MS (neg.): m/

z: 844.1 [MꢀH]^ꢀ; r₁6.12 mm^ꢀ1s^ꢀ1(pH 7.4, MOPS, 300 MHz). TbL⁴:

¹⁹F{¹H} NMR (282 MHz, D₂O, 25 8C) ꢀ60.98, ꢀ63.15 ppm; ESI-MS

(neg.): m/z: 845.2 [MꢀH]^ꢀ. YbL⁴: ¹⁹F{¹H} NMR (282 MHz, D₂O, 25 8C):

d=ꢀ60.54 ppm; ESI-MS (neg.): m/z: 859.8 [MꢀH]^ꢀ.

NMR spectroscopy and mass spectrometry: ¹H, ¹³C {¹H}, ³¹P {¹H} and

¹⁹F {¹H} NMR spectra, ¹H and ¹⁹F T₁and T₂times were recorded on a

Bruker Avance III 300 MHz ꢆMicrobayꢃ spectrometer at 258C Bruker

(Germany). ¹⁹F T₁and T₂times were measured using inversion recovery

and Carr-Purcell-Meiboom-Gill sequences respectively, on 5 mm solution

of complex in MOPS buffer (0.1m, pH 7.4). The exact concentrations of

the complexes were determined by measurement of bulk magnetic sus-

ceptibility shifts of a tBuOH signal.^[54]Variable pH relaxivities were de-

termined through inversion recovery measurements, where the pH was

adjusted using solid LiOH or TsOH to minimize the change in concentra-

tion. Variable-temperature ¹⁷O NMR measurements were recorded on a

Bruker Avance III (11.7 T) spectrometer equipped with a 5 mm probe

and standard temperature control units. Aqueous solutions of the com-

plexes (5–20 mm) containing 2.0% of the ¹⁷O isotope (Cambridge Iso-

tope) were used. The observed transverse relaxation rates were calculat-

ed from the signal width at half-height. ¹H DOSY measurements were

carried out nonspinning on a Varian VNMRS 500 MHz spectrometer at

296 K using the Oneshot DOSY pulse sequence. Data were acquired

with an array of 16 gradient amplitudes ranging from 3.0 G cm^ꢀ1to

27.0 G cm^ꢀ1in equal steps of gradient squared, using 16 transients, 65536

complex data points, a total diffusion encoding gradient of 2 ms and a dif-

fusion time of 0.05 s. DOSY spectra were constructed using correction

for the effects of pulsed field gradient nonuniformity. ESI-LRMS were

performed on an ion trap SL 1100 system Agilent (Germany). FT-ICR-

MS were performed on a Bruker FT-ICR Apex II spectrometer Agilent

(Germany). HR-EI-MS were performed on a MAT Sektorfeld mass spec-

trometer Finnigan (Germany).

162.22 ppm;

¹⁹F{¹H} NMR

(282 MHz,

CDCl₃)

d=ꢀ61.72 ppm;

³¹P{¹H} NMR (122 MHz, CDCl₃): d=14.47 ppm; ESI-MS (pos.): m/z: 482

[M+H]⁺; HRMS (EI): m/z: calcd for C₁₉H₂₁O₄F₃P₁Br₁: 482.029246;

found: 482.02514.

Tri-tert-butyl 2,2’,2’’-(10-(4-((2-(diethoxyphosphoryl)-4-(trifluoromethyl)-

phenoxy)methyl)benzyl)-1,4,7,10-tetraazacyclododecane-1,4,7-triyl)triace-

tate (5d): Tris tert-butyl-DO3A (2.20 g, 3.69 mmol), 4d (2.67 g,

5.54 mmol) and potassium carbonate (1.53 g, 11.08 mmol) were added to

dimethylformamide (20 mL) and the mixture was heated to 608C under a

dinitrogen atmosphere for 48 h. The solvent was removed under reduced

pressure, and dichloromethane was added. Following the removal of the

inorganic salts the organic phase was washed with saturated sodium

chloride solution and dried over sodium sulphate. Removal of the solvent

yielded an orange oil. The crude product was purified using column chro-

matography (alumina from 100% dichloromethane to 93:7 dichlorome-

thane/methanol). Removal of the solvents yielded 5d as a yellow solid

(1.91 g, 91%). Tlc: R_f=0.25 (alumina, dichloromethane/methanol=

95:5); ¹H NMR (300 MHz, CDCl₃): d=1.18–1.39 (m, 33H, POCH₂CH₃;

CCH₃), 1.97–4.14 (brm, 34H; CH₂ring, NCH₂COO, NCH₂CH₂CH₂O,

POCH₂CH₃), 5.17 (s, 2H; CCH₂Br), 6.98–7.04 (m, 1H; ArH), 7.27–7.44

(m, 4H; ArH), 7.66 (t, 1H, ³J_H,H=7.18 Hz; ArH), 8.05 ppm (d, 1H,

³J_H,H=15.30 Hz; ArH); ¹³C{¹H} NMR (75 MHz, CDCl₃): d=16.13, 27.78,

29.44, 49.46, 51.71, 51.91, 51.97, 55.57, 55.91, 56.32, 58.86, 59.62, 62.22,

69.98, 70.36, 80.39, 82.23, 82.64, 112.35, 116.63, 119.94, 122.23, 126.81,

129.06, 130.26, 131.21, 134.97, 136.84, 139.81, 162.25, 171.00, 172.38,

173.34 ppm; ¹⁹F{¹H} NMR (282 MHz, CDCl₃): d=ꢀ61.72, ꢀ61.69 ppm;

³¹P{¹H} NMR (122 MHz, CDCl₃): d=14.57 ppm; ESI-MS (pos.): m/z:

915.5 [M+H]⁺, 937.5 [M+Na]⁺; HRMS (FT-ICR): m/z: calcd for

C₄₅H₇₁F₃N₄O₁₀P₁: 915.48544 [M+H]⁺; found: 915.48533.

Luminescence spectroscopy: Steady-state and time resolved measure-

ments were performed on a QuantaMasterTM 3 PH fluorescence spec-

trometer from Photon Technology International, Inc. (USA). Details re-

garding the luminescence measurements for TbL^5–8are given in the sup-

plementary information of the previous work.^[16]The steady state meas-

urements were performed in H₂O (258C, pH 7.4, MOPS) at a concentra-

2,2’,2’’-(10-(4-((2-Phosphono-4-(trifluoromethyl)phenoxy)methyl)benzyl)-

1,4,7,10-tetraazacyclododecane-1,4,7-triyl)triacetic acid (L⁴): The ester 5d

(0.427 g, 0.467 mmol) was dissolved in dimethylformamide (5 mL) and

cooled to 08C in an ice bath, bromotrimethylsilane (0.647 mL,

4.91 mmol) was added slowly and the solution was stirred at room tem-

perature for 18 h. The absence of the ethyl esters was confirmed by elec-

trospray mass spectrometry. A mixture of dichloromethane and trifluoro-

acetic acid (1:1, 10 mL) was added to the crude orange oil and the mix-

ture was stirred at room temperature for 24 h. Following removal of the

solvents a brown oil was obtained and electrospray mass spectrometry

again confirmed the removal of the tert-butyl groups. The crude product

was purified through reverse phase HPLC. On freeze drying L⁴was iso-

lated as a flocculent white solid that quickly became a yellow oil due to

its hygroscopic nature (0.253 g, 79%). ¹H NMR (300 MHz, D₂O): d=

2.96–4.16 (brm, 28H; CH₂ring, NCH₂COO, NCH₂CH₂CH₂O,), 5.25 (s,

tion of 250 mm for EuL^1–450 mm for TbL^1–4, and 5 mm for EuL^5–8

, .

Excitation and emission slits were set to 2 nm for EuL^1–8and TbL^1–4, da-

tasets are an average of 10 scans. The time resolved measurements were

performed in H₂O and D₂O (258C, pH 7.4, MOPS) at a concentration of

50 mm for TbL^1–4and 5 mm for EuL^1–8. Excitation and emission slits were

set to 15 and 5 nm band pass respectively, with 10 ms resolution. Datasets

are recorded with a 100 ms delay and are an average of 15 scans. Each re-

ported value is the mean of three independent measurements and ob-

tained curves are fitted to the first order exponential decay with R²>

0.99.

Relaxometry: The proton 1/T₁NMRD profiles were measured on a fast

field-cycling Stelar SmartTracer relaxometer over a continuum of mag-

Chem. Eur. J. 2013, 00, 0 – 0

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

&

15

&

ÞÞ

These are not the final page numbers!

G. Angelovski et al.

netic field strengths from 0.00024 to 0.25 T (corresponding to 0.01–

10 MHz proton Larmor frequencies). The relaxometer operates under

computer control with an absolute uncertainty in 1/T₁of ꢁ1%. Addition-

al data points in the range 15–70 and 500 MHz were obtained on a

Bruker WP80 NMR electromagnet adapted to variable-field measure-

ments (15–80 MHz proton Larmor frequency) Stelar Relaxometer and

Bruker Avance III 500 MHz (11.7 T) spectrometer respectively. The ¹H

T₁relaxation times were acquired by the standard inversion recovery

method with typical 908 pulse width of 3.5 ms, 16 experiments of 4 scans.

The reproducibility of the T₁data was ꢁ5%. The temperature was con-

trolled with a Stelar VTC-91 airflow heater equipped with a calibrated

copper-constantan thermocouple (uncertainty of ꢁ0.18C).

method proposed by Irving et al.^[56]A 0.01m HCl (0.15m NaCl solution

was titrated with 0.2m NaOH and the difference between the measured

and calculated pH values was used to calculate the [H⁺] from the pH

values determined in the titration experiments. For the calculation of the

equilibrium constants the PSEQUAD program was used.^[57]

Kinetic measurements: The rates of the metal exchange reactions of the

Gd³⁺complexes of L^1,3,8with Cu²⁺ion were studied by UV-spectropho-

tometry, following the formation of the Cu²⁺complexes at 300 nm with

the use of 1.0 cm cells and Cary 1E spectrophotometer. The concentra-

tion of the complexes was 1ꢅ10^ꢀ4m, whereas the concentration of the

Cu²⁺was 10-times higher, in order to guarantee the pseudo-first-order

conditions. Some reactions were performed by using 40-fold excess of the

Cu²⁺ion to get information on the role of the exchanging metal ion,

which was found to be negligible. The temperature was maintained at

258C and the ionic strength of the solutions was kept constant, 0.15m

NaCl. The exchange rates were studied in the pH range 3.0–5.0. For

keeping the pH values constant, 1,4-dimethylpiperazine buffer (25 mm)

was used. The pseudo-first-order rate constants (k_d) were calculated by

fitting the absorbance data with the use of the following equation:

Equilibrium measurements: The chemicals used in the equilibrium meas-

urement experiments were of the highest analytical grade. For the prepa-

ration of the LnCl₃solutions, Ln₂O₃(Fluka, 99.9%) was dissolved in

6.0m HCl and the excess acid was evaporated. The concentration of the

LnCl₃, CaCl₂, CuCl₂, and ZnCl₂solutions were determined by complexo-

metric titration with standardized Na₂H₂EDTA and xylenol orange

(LnCl₃), murexide (CuCl₂, ZnCl₂), and Patton & Reeder (CaCl₂) as indi-

cators. The concentration and the protonation constants of L^1,3,8were de-

termined by pH-potentiometric titrations in the presence and absence of

a large (40-fold) excess of CaCl₂. All the equilibrium measurements were

carried out at constant 0.15m NaCl ionic strength at 258C. The protona-

tion and stability constants of the Ca²⁺, Zn²⁺, Cu²⁺, and Gd³⁺complexes

of L^1,3,8, CaDOTA, ZnDOTA, as well as the stability constants of the in-

termediate complexes, *GdH₃L¹and *GdH₄L⁸were also determined by

means of pH-potentiometric titration. The stability constants for Ca²⁺

and Zn²⁺complexes of DOTA are lower from those previously reported

due to the lower first protonation constant related to the formation of

the relatively stable NaDOTA complex.^[55]The protonation constants of

the Gd³⁺complexes were determined in the pH range 3–12 avoiding

their dissociation at pH<3,0. The best fitting was obtained by using the

model which includes the formation of ML and MH_1ꢀ5L species in equili-

brium depending on the investigated system. In all cases the metal-to-

ligand concentration ratios were 1:1. The concentration of the ligands

was generally 2–3 mm. For the calculation of the protonation and stability

constants mL base–pH data pairs (80–100), obtained in the pH range of

1.7–12, were used. Because of the slow formation rates of the macrocyclic

Gd³⁺complexes “out-of-cell” technique was used to determine their sta-

bility constants. 8 samples were prepared for the systems Gd-L^1,8at a

2 mm concentration between pH 2–4. After 1 month equilibration time

the pH and relaxivity values of the samples were measured. For the cal-

culation of the stability constants of the Gd³⁺complexes, the protonation

constant of the ligands and the complexes as well as the stability con-

stants of the intermediates were also considered. The determination of

the stability constants of Cu²⁺complexes were carried out by following

the formation of Cu²⁺complexes with spectrophotometry in very acidic

solutions (c_H+=0.01–1,0m). The concentration of Cu²⁺and the ligands in

the samples were 1 mm (6 samples were prepared for each system). The

total H⁺concentration in the samples was adjusted by the addition of a

calculated amount of 3m HCl. The ionic strength was not constant in

these samples. For the equilibrium calculations, the protonation constants

of the Cu²⁺complexes, the molar absorptivities of the CuCl₂and the

CuL^1,3,8complexes with different protonation states were used. The

molar absorptivities of CuL^1,3,8complexes and those of protonated forms

were determined at 11 wavelengths between 575–775 nm by recording

the spectra of 1.0–4.0ꢅ10^ꢀ3m solutions of CuL^1,3,8complexes in the pH

range 1.5–6.0 (0.15m NaCl, 258C). After 1 month equilibration time, the

absorbance values of the samples were measured also at 11 wavelengths,

as before. The spectrophotometric measurements were made by using

Cary 1E spectrophotometer at 258C. The optical length was 10 mm. The

pH-potentiometric titrations were carried out with a Methrohm 785DMP

Titrino titration workstation with the use of a Metrohm-6.0233.100 com-

bined electrode. The titrated solution (10.00 mL) was thermostated at

258C. The samples were stirred and to avoid the effect of CO₂, N₂gas

was bubbled through the solutions. The titrations were made in the pH

range 1.7–12. For the calibration of the pH meter, KH-phthalate

(pH 4.005) and borax (pH 9.177) buffers were used. The concentrations

of the H⁺ions were calculated from the measured pH values using the

_dt

A_t¼ ðA₀ꢀA_eÞe^ꢀkþA_e

ð4Þ

where A_t, A₀, and A_ewere the absorbance at time t, at the start and at

equilibrium of the reactions, respectively. The calculations were per-

formed by using the computer program Micromath Scientist, version 2.0

(Salt Lake City, UT, USA).

MR phantom imaging: ¹H phantom MR images were recorded on a

series of Eppendorf tubes containing different dilutions of GdL⁴and

DOTAREM (GdDOTA) in 150 mL of MOPS buffered H₂O (pH 7.4,

0.1m) at 0, 10, 20, 40, and 80 mm. The images were recorded at 218C

1

using 7 T Bruker BioSpec 70/30 USR Preclinical MRI system using a H/

volume transmit–receive coil tuned to the proton resonance (300.3 MHz).

The map was obtained from a set of inversion-recovery SE measure-

ments (adiabatic 1808 pulse with inversion time TI=23, 50, 150, 250, 500,

750, 1000, 1500, 2000, 2500, 3500, 5000 ms, TE/TR=10.7/6000 ms, flip

angle=90/1808, effective spectral width 50 kHz, matrix size=300ꢅ300,

voxel size 0.2ꢅ0.2ꢅ2 mm³, FOV: 6ꢅ6 cm). The included T₁maps show

the results obtained by voxel wise nonlinear least-squares fitting of the

proton signal, using the Levenberg–Marquardt steepest descent algorithm

implemented in MATLAB, version 7.10 (MathWorks Inc., Natick, USA).

¹⁹F MR phantom images were recorded on a series of Eppendorf tubes

containing different dilutions of GdL⁴and YbL⁴in 150 mL of MOPS buf-

fered H₂O (pH 7.4, 0.1m) at 1.0, 2.0, 3.5, and 5.0 mm, on the above scan-

ner equipped with a BGA-9S gradient insert using a dual ¹H/¹⁹F single

loop surface coil (282.56 MHz for ¹⁹F). For YbL^{4 19}F images were ac-

quired by a FLASH sequence with TE/TR=3.5/45 ms, flip angle=308,

effective spectral width 10 kHz, 2500 averages, 60 min acquisition time,

matrix size=32ꢅ32, voxel size 1ꢅ1ꢅ5 mm³, FOV: 3.2ꢅ3.2 cm. For GdL⁴

the parameters used for FLASH were: TE/TR=2.67/7.48 ms, flip angle=

84.58, effective spectral width 17.5 kHz, 15000 averages, 60 min acquisi-

tion time, matrix size=32ꢅ32, voxel size 1ꢅ1ꢅ5 mm³, FOV: 3.2ꢅ3.2 cm.

For GdL⁴¹⁹F MRI was also made using a FISP sequence with TE/TR=

1.2/2.3 ms, flip angle=608, effective spectral width 40 kHz, 5000 averages,

60 min acquisition time, matrix size=64ꢅ64, voxel size 1ꢅ1ꢅ5 mm³,

FOV: 6.4ꢅ6.4 cm. For the lower concentrations of GdL⁴shown in

Figure 9, the sequence parameters for FISP were TE/TR=0.9/1.8 ms, flip

angle=888, effective spectral width 29.8 kHz, 44000 averages, 60 min ac-

quisition time, matrix size=64ꢅ64, voxel size 1ꢅ1ꢅ5 mm³, FOV: 6.4ꢅ

6.4 cm. A region-of-interest was placed outside the phantoms to deter-

mine the Rayleigh corrected noise level that was used for calculating the

1

SNR of the ¹⁹F MR images. The H relaxation rates were quantified from

a set of single slice spin echo MR image using the same double tuned

coil as for ¹⁹F MRI (Figure 9) with a TE/TR=10.7/6000 ms, flip angle=

908, effective spectral width 50 kHz, 1 average, 20 min acquisition time,

matrix size=320ꢅ320, voxel size 0.1ꢅ0.1ꢅ2 mm³, FOV: 3.2ꢅ3.2 cm. A

slab-selective (6 mm thick) 1808 inversion pulse was used with the follow-

ing inversion times: 23/50/150/250/500/750/1000/1500/2000/2500/3500/

5000 ms. Nonlinear least-squares fitting of the modulus data was per-

formed voxel wise according to: S=S₀(1ꢀa·e^ꢀR2·TI) using the double-fit-

&

16

&

www.chemeurj.org

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 0000, 00, 0 – 0

ÝÝ

These are not the final page numbers!

Dual Frequency MRI Probes

FULL PAPER

ting algorithm minimizing chi-square.^[58]The a factor is 2 if a perfect in-

version is obtained. Across the imaged slice this factor was: 2ꢁ0.03.

MRI data were reconstructed on the scanner and analyzed with home-

written MATLAB code (v. 7.10, Mathworks Inc., Natick, USA).

[17] K. H. Chalmers, M. Botta, D. Parker, Dalton Trans. 2011 , 40 , 90 4 –

913.

[18] a) J. C. G. Bunzli, P. Froidevaux, J. M. Harrowfield, Inorg. Chem.

1993 , 32 , 330 6 –3 311; b) W. D. Horrocks, J. P. Bolender, W. D. Smith,

R. M. Supkowski, J. Am. Chem. Soc. 1997, 119, 5972–5973.

[19] 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 , 49 3 –5 03.

Acknowledgements

[20] P. Atkinson, B. S. Murray, D. Parker, Org. Biomol. Chem. 2006 , 4 ,

316 6 –3 171.

We are thankful to Dr. Sara Di Pinto and Dr. Claudio Cassino for their

help with relaxometric measurements, and Dr. Louise Natrajan and

Dr. Ilgar Mamedov for helpful discussions. Financial support from the

Max-Planck Society and the German Research Foundation (DFG, grant

No. AN 716/2-1) is gratefully acknowledged. The European COST

Action D38 “Metal-Based Systems for Molecular Imaging Applications”,

TD1004 “Theranostics Imaging and Therapy: An Action To Develop

Novel Nanosized Systems For Imaging-Guided Drug Delivery”, CM1006

“EUFEN: European F-Element Network”, the Hungarian Scientific Re-

search Foundation (OTKA PD-83253 and K-84291), the TꢇMOP-

4.2.2A-11/1/KONV-2012-0043 and TꢇMOP-4.2.2/B-10/1-2010-0024 proj-

ects implemented through the New Hungary Development Plan, co-fi-

nanced by the European Social Fund and the European Regional Devel-

opment Fund, and the Italy–Hungary Intergovernmental S&T Coopera-

tion Program (HU11MO2-TET_10-1-2011-0202) are also acknowledged.

[21] a) S. Aime, M. Botta, D. Parker, J. A. G. Williams, J. Chem. Soc.

Dalton Trans. 1996 , 1 7 –2 3; b) J. Rudovsky, P. Cigler, J. Kotek, P.

Hermann, P. Vojtisek, I. Lukes, J. A. Peters, L. Vander Elst, R. N.

Muller, Chem. Eur. J. 2005, 11, 2373–2384; c) J. Rudovsky, M.

Botta, P. Hermann, A. Koridze, S. Aime, Dalton Trans. 2006, 2323–

2333.

[22] a) S. Aime, M. Botta, E. Garino, S. G. Crich, G. Giovenzana, R. Pa-

gliarin, G. Palmisano, M. Sisti, Chem. Eur. J. 2000 , 6 , 260 9 –2 617;

b) J. B. Livramento, A. Sour, A. Borel, A. E. Merbach, V. Toth,

Chem. Eur. J. 2006 , 12 , 98 9 –1 003.

[23] a) S. H. Koenig, R. D. Brown, Prog. Nucl. Magn. Reson. Spectrosc.

1990 , 22 , 48 7 –5 67; b) D. H. Powell, O. M. Ni Dhubhghaill, D.

Pubanz, L. Helm, Y. S. Lebedev, W. Schlaepfer, A. E. Merbach, J.

Am. Chem. Soc. 1996 , 118 , 933 3 –9 346.

[24] S. Aime, M. Botta, M. Fasano, E. Terreno, Chem. Soc. Rev. 1998 , 27 ,

1 9 –2 9.

[25] S. Aime, M. Botta, S. G. Crich, G. Giovenzana, R. Pagliarin, M.

Sisti, E. Terreno, Magn. Reson. Chem. 1998 , 36 , S20 0 –S 208.

[26] D. Messeri, M. P. Lowe, D. Parker, M. Botta, Chem. Commun. 2001 ,

274 2 –2 743.

[27] a) A. Datta, J. M. Hooker, M. Botta, M. B. Francis, S. Aime, K. N.

Raymond, J. Am. Chem. Soc. 2008 , 130 , 254 6 –2 552; b) Z. Jꢁszberꢈ-

nyi, L. Moriggi, P. Schmidt, C. Weidensteiner, R. Kneuer, A. E. Mer-

bach, L. Helm, E. Toth, J. Biol. Inorg. Chem. 2007 , 12 , 40 6 –4 20;

c) G. M. Nicolle, E. Toth, H. Schmitt-Willich, B. Raduchel, A. E.

Merbach, Chem. Eur. J. 2002 , 8 , 104 0 –1 048.

[28] J. Costa, E. Balogh, V. Turcry, R. Tripier, M. Le Baccon, F. Chuburu,

H. Handel, L. Helm, E. Toth, A. E. Merbach, Chem. Eur. J. 2006 ,

12 , 684 1 –6 851.

[29] a) T. J. Swift, R. E. Connick, J. Chem. Phys. 1962 , 37 , 30 7 –3 20; b) K.

Micskei, L. Helm, E. Brucher, A. E. Merbach, Inorg. Chem. 1993 ,

32 , 384 4 –3 850.

[30] J. F. Desreux, E. Merciny, M. F. Loncin, Inorg. Chem. 1981 , 20 , 98 7 –

991.

[31] R. G. Franz, AAPS PharmSciTech 2001, 3, E10.

[32] Z. Baranyai, Z. Palinkas, F. Uggeri, E. Brꢉcher, Eur. J. Inorg. Chem.

2010 , 194 8 –1 956.

[33] W. P. Cacheris, S. K. Nickle, A. D. Sherry, Inorg. Chem. 1987 , 26 ,

95 8 –9 60.

[1] J. K. Willmann, N. van Bruggen, L. M. Dinkelborg, S. S. Gambhir,

Nat. Rev. Drug Discovery 2008 , 7 , 59 1 –6 07.

[2] M. M. J. Modo, J. W. M. Bulte, Molecular and Cellular MR Imaging,

CRC, Boca Raton, 2007, p.421.

[3] N. Bloembergen, L. O. Morgan, J. Chem. Phys. 1961 , 34 , 842.

[4] a) A. E. Merbach, E. Toth, The Chemistry of Contrast Agents in

Medical Magnetic Resonance Imaging, Wiley, Chichester, 2001; b) P.

Caravan, J. J. Ellison, T. J. McMurry, R. B. Lauffer, Chem. Rev. 1999 ,

99 , 229 3 –2 352.

[5] M. Botta, Eur. J. Inorg. Chem. 2000 , 39 9 –4 07.

[6] P. Lebduskova, P. Hermann, L. Helm, E. Toth, J. Kotek, K. Binne-

mans, J. Rudovsky, I. Lukes, A. E. Merbach, Dalton Trans. 2007,

493–501.

[7] a) P. Caravan, Chem. Soc. Rev. 2006 , 35 , 51 2 –5 23; b) E. J. Werner,

A. Datta, C. J. Jocher, K. N. Raymond, Angew. Chem. 2008 , 120 ,

869 6 –8 709; Angew. Chem. Int. Ed. 2008 , 47 , 856 8 –8 580.

[8] M. Srinivas, A. Heerschap, E. T. Ahrens, C. G. Figdor, I. J. M. d.

Vries, Trends Biotechnol. 2010 , 28 , 36 3 –3 70.

[9] a) J. Ruiz-Cabello, P. Walczak, D. A. Kedziorek, V. P. Chacko, A. H.

Schmieder, S. A. Wickline, G. M. Lanza, J. W. M. Bulte, Magn.

Reson. Med. 2008 , 60 , 150 6 –1 511; b) E. T. Ahrens, R. Flores, H. Y.

Xu, P. A. Morel, Nat. Biotechnol. 2005 , 23 , 98 3 –9 87.

[10] a) K. J. Thurecht, I. Blakey, H. Peng, O. Squires, S. Hsu, C. Alexand-

er, A. K. Whittaker, J. Am. Chem. Soc. 2010 , 132 , 533 6 –5 337; b) M.

Ogawa, S. Nitahara, H. Aoki, S. Ito, M. Narazaki, T. Matsuda, Mac-

romol. Chem. Phys. 2010 , 211 , 160 2 –1 609.

[11] a) L. Helm, Prog. Nucl. Magn. Reson. Spectrosc. 2006 , 49 , 4 5 –6 4;

b) P. Harvey, I. Kuprov, D. Parker, Eur. J. Inorg. Chem. 2012 , 201 5 –

2022.

[12] a) S. Mizukami, R. Takikawa, F. Sugihara, Y. Hori, H. Tochio, M.

Walchli, M. Shirakawa, K. Kikuchi, J. Am. Chem. Soc. 2008 , 130 ,

79 4 –7 95; b) S. Mizukami, H. Matsushita, R. Takikawa, F. Sugihara,

M. Shirakawa, K. Kikuchi, Chem. Sci. 2011 , 2 , 115 1 –1 155.

[13] K. H. Chalmers, A. M. Kenwright, D. Parker, A. M. Blamire, Magn.

Reson. Med. 2011 , 66 , 93 1 –9 36.

[14] a) S. A. Anderson, R. K. Rader, W. F. Westlin, C. Null, D. Jackson,

G. M. Lanza, S. A. Wickline, J. J. Kotyk, Magn. Reson. Med. 2000 ,

44 , 43 3 –4 39; b) K. C. Partlow, J. Chen, J. A. Brant, A. M. Neubauer,

T. E. Meyerrose, M. H. Creer, J. A. Nolta, S. D. Caruthers, G. M.

Lanza, S. A. Wickline, FASEB J. 2007 , 21 , 164 7 –1 654.

´

[34] I. Mamedov, P. Tꢁborsky, P. Lubal, S. Laurent, L. V. Elst, H. A.

Mayer, N. K. Logothetis, G. Angelovski, Eur. J. Inorg. Chem. 2009 ,

329 8 –3 306.

[35] a) M. Forsterova, I. Svobodova, P. Lubal, P. Taborsky, J. Kotek, P.

Hermann, I. Lukes, Dalton Trans. 2007, 535–549; b) S. L. Wu, W. D.

Horrocks, Inorg. Chem. 1995 , 34 , 372 4 –3 732.

[36] a) J. Moreau, E. Guillon, J.-C. Pierrard, J. Rimbault, M. Port, M.

Aplincourt, Chem. Eur. J. 2004 , 10 , 521 8 –5 232; b) E. Toth, E.

Brucher, I. Lazar, I. Toth, Inorg. Chem. 1994 , 33 , 407 0 –4 076.

[37] A. Bianchi, L. Calabi, F. Corana, S. Fontana, P. Losi, A. Maiocchi,

L. Paleari, B. Valtancoli, Coord. Chem. Rev. 2000 , 204 , 30 9 –3 93.

[38] H.-Z. Cai, T. A. Kaden, Helv. Chim. Acta 1994 , 77 , 38 3 –3 98.

[39] J. Kotek, F. K. Kalman, P. Hermann, E. Brucher, K. Binnemans, I.

Lukes, Eur. J. Inorg. Chem. 2006 , 197 6 –1 986.

[40] Z. Baranyai, Z. Pꢁlinkꢁs, F. Uggeri, A. Maiocchi, S. Aime, E. Brꢉch-

er, Chem. Eur. J. 2012 , 18 , 1642 6 –1 6435.

[41] L. Sarka, L. Burai, R. Kirꢁly, L. Zꢈkꢁny, E. Brꢉcher, J. Inorg. Bio-

chem. 2002 , 91 , 32 0 –3 26.

[15] E. Gianolio, R. Napolitano, F. Fedeli, F. Arena, S. Aime, Chem.

Commun. 2009 , 604 4 –6 046.

[16] M. P. Placidi, J. Engelmann, L. S. Natrajan, N. K. Logothetis, G. An-

gelovski, Chem. Commun. 2011 , 47 , 1153 4 –1 1536.

[42] K. Kumar, T. Jin, X. Wang, J. F. Desreux, M. F. Tweedle, Inorg.

Chem. 1994 , 33 , 382 3 –3 829.

Chem. Eur. J. 2013, 00, 0 – 0

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

&

17

&

ÞÞ

These are not the final page numbers!

G. Angelovski et al.

[43] F. K. Kꢁlmꢁn, Z. Baranyai, I. Tꢂth, I. Bꢁnyai, R. Kirꢁly, E. Brꢉcher,

S. Aime, X. Sun, A. D. Sherry, Z. Kovꢁcs, Inorg. Chem. 2008 , 47 ,

385 1 –3 862.

[44] a) J. A. Peters, J. Huskens, D. J. Raber, Prog. Nucl. Magn. Reson.

Spectrosc. 1996 , 28 , 28 3 –3 50; b) D. Parker, R. S. Dickins, H. Pusch-

mann, C. Crossland, J. A. K. Howard, Chem. Rev. 2002 , 102 , 197 7 –

2010.

[45] C. Platas-Iglesias, Eur. J. Inorg. Chem. 2012 , 202 3 –2 033.

[46] a) S. Aime, M. Botta, G. Ermondi, Inorg. Chem. 1992 , 31 , 429 1 –

4299; b) S. Aime, M. Botta, M. Fasano, M. P. M. Marques, C. F. G. C.

Geraldes, D. Pubanz, A. E. Merbach, Inorg. Chem. 1997 , 36 , 205 9 –

2068.

[47] a) M. Jauregui, W. S. Perry, C. Allain, L. R. Vidler, M. C. Willis,

A. M. Kenwright, J. S. Snaith, G. J. Stasiuk, M. P. Lowe, S. Faulkner,

Dalton Trans. 2009 , 628 3 –6 285; b) J. P. Andrꢈ, C. F. G. C. Geraldes,

J. A. Martins, A. E. Merbach, M. I. M. Prata, A. C. Santos, J. J. P.

de Lima, E. Toth, Chem. Eur. J. 2004, 10, 5804–5816.

[48] M. Woods, S. Aime, M. Botta, J. A. K. Howard, J. M. Moloney, M.

Navet, D. Parker, M. Port, O. Rousseaux, J. Am. Chem. Soc. 2000 ,

122 , 978 1 –9 792.

[49] C. S. Johnson, Prog. Nucl. Magn. Reson. Spectrosc. 1999 , 34 , 20 3 –

256.

[50] J. Henig, I. Mamedov, P. Fouskova, E. Tꢂth, N. K. Logothetis, G.

Angelovski, H. A. Mayer, Inorg. Chem. 2011 , 50 , 647 2 –6 481.

[51] R. Freeman, H. D. W. Hill, J. Magn. Reson. 1971, 4, 366–383.

[52] M. Botta, L. Tei, Eur. J. Inorg. Chem. 2012 , 194 5 –1 960.

[53] A. Dadabhoy, S. Faulkner, P. G. Sammes, J. Chem. Soc. Perkin

Trans. 2 2002 , 34 8 –3 57.

[54] D. M. Corsi, C. Platas-Iglesias, H. van Bekkum, J. A. Peters, Magn.

Reson. Chem. 2001 , 39 , 72 3 –7 26.

[55] R. Delgado, J. J. R. F. Dasilva, Talanta 1982 , 29 , 81 5 –8 22.

[56] H. M. Irving, M. G. Miles, L. D. Pettit, Anal. Chim. Acta 1967 , 38 ,

47 5 –4 88.

[57] D. J. Leggett, Computational Methods for the Determination of For-

mation Constants, Plenum, New York, 1985.

[58] D. R. Messroghli, A. Rudolph, H. Abdel-Aty, R. Wassmuth, T.

Kuhne, R. Dietz, J. Schulz-Menger, BMC Med. Imaging 2010 , 10 ,

16.

Received: February 26, 2013

Published online: && &&, 0000

&

18

&

www.chemeurj.org

ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 0000, 00, 0 – 0

ÝÝ

These are not the final page numbers!

Article Doi

Aryl-phosphonate lanthanide complexes and their fluorinated derivatives: Investigation of their unusual relaxometric behavior and potential application as dual frequency 1H/19F MRI probes

DOI: 10.1002/chem.201300763

Source and publish data:

Authors:

Article abstract of DOI:10.1002/chem.201300763

Full text of DOI:10.1002/chem.201300763

Products guided by the article

R&D Labs maybe for 1454305-47-5

Relevant to this article

Hot Product