Physicochemical characterization of the dimeric lanthanide complexes [en{Ln(DO3A)(H2O)}2] and [pi{Ln(DTTA)(H 2O)}2]2-: A variable-temperature 17O NMR study

Article abstract of DOI:10.1002/mrc.1315

The Gd(III) complexes of the two dimeric ligands [en(DO3A)₂] {N,N′-bis[1,4,7-tris(carboxymethyl)-1,4,7,10-tetraazacyclododecan-10-yl- methylcarbonyl]-N,N′-ethylenediamme} and [pi(DTTA)₂] ^8- [bisdiethylene-triaminepentaacetic acid (trans-1,2- cyclohexanediamine)] were synthesized and characterized. The ¹⁷O NMR chemical shift of H₂O induced by [en{Dy(DO3A)}₂] and [pi{Dy(DTTA)}₂]^2- at pH 6.80 proved the presence of 2.1 and 2.2 inner-sphere water molecules, respectively. Water proton spin-lattice relaxation rates for [en{Gd(DO3A)(H₂O)}₂] and [pi{Gd(DTTA)(H₂O)}₂]^2- at 37.0 ± 0.1°C and 20 MHz are 3.60 ± 0.05 and 5.25 ± 0.05 mM ^-1 s^-1 per Gd, respectively. The EPR transverse electronic relaxation rate and ¹⁷O NMR transverse relaxation time for the exchange lifetime of the coordinated H₂O molecule and the ²H NMR longitudinal relaxation rate of the deuterated diamagnetic lanthanum complex for the rotational correlation time were thoroughly investigated, and the results were compared with those reported previously for other lanthanide(III) complexes. The exchange lifetimes for [en{Gd(DO3A)(H ₂O)}₂] (769 ± 10 ns) and [pi{Gd(DTTA)(H ₂O)}₂]^2- (910 ± 10 ns) are significantly higher than those of [Gd(DOTA)(H₂O)]^- (243 ns) and [Gd(DTPA)(H₂O)]^2- (303 ns) complexes. The rotational correlation times for [en{Gd(DO3A)(H₂O)}₂] (150 ± 11 ps) and [pi{Gd(DTTA)(H₂O)}₂]^2- (130 ± 12 ps) are slightly greater than those of [Gd(DOTA)(H₂O)] ^- (77 ps) and [Gd(DTPA)(H₂O)]^2- (58 ps) complexes. The marked increase in relaxivity (r₁) of [en{Gd(DO3A)(H₂O)}₂] and [pi{Gd(DTTA)(H ₂O)}₂]^2- result mainly from their longer rotational correlation time and higher molecular weight. Copyright

Full text of DOI:10.1002/mrc.1315

MAGNETIC RESONANCE IN CHEMISTRY
Magn. Reson. Chem. 2004; 42: 329–336
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/mrc.1315
Physicochemical characterization of the dimeric
lanthanide complexes [enfLn(DO3A)(H O)g ] and
2
2
2
−
[
pifLn(DTTA)(H O)g ] : a variable-temperature
2
2
1
7
O NMR study
1
1
1
2
3
Tzu-Ming Lee, Tsan-Hwang Cheng, Ming-Hung Ou, C. Allen Chang, Gin-Chung Liu and
Yun-Ming Wang1
∗
1
School of Medicinal and Applied Chemistry, Kaohsiung Medical University, 100 Shih-Chuan 1st Road, Kaohsiung 807, Taiwan
Department and Institute of Biological Science and Technology, National Chiao Tung University, Hsinchu 300, Taiwan
Department of Radiology, Kaohsiung Medical University, 100 Shih-Chuan 1st Road, Kaohsiung 807, Taiwan
2
3
Received 16 June 2003; Revised 18 September 2003; Accepted 23 September 2003
ꢀ
The Gd(III) complexes of the two dimeric ligands [en(DO3A)
2
] fN,N -bis[1,4,7-tris(carboxymethyl)-1,4,7,10-
ꢀ
8−
tetraazacyclododecan-10-yl-methylcarbonyl]-N,N -ethylenediamineg and [pi(DTTA)
2
]
[bisdiethylene-
1
7
triaminepentaacetic acid (trans-1,2-cyclohexanediamine)] were synthesized and characterized. The
O
2
−
NMR chemical shift of H
2
O induced by [enfDy(DO3A)g
2
] and [pifDy(DTTA)g
2
]
at pH 6.80 proved
the presence of 2.1 and 2.2 inner-sphere water molecules, respectively. Water proton spin–lattice
2
−
◦
relaxation rates for [enfGd(DO3A)(H
2
O)g
2
] and [pifGd(DTTA)(H
2
O)g
2
]
at 37.0 ± 0.1 C and 20 MHz
−1
−1
are 3.60 ± 0.05 and 5.25 ± 0.05 mM
s
per Gd, respectively. The EPR transverse electronic relax-
1
7
ation rate and O NMR transverse relaxation time for the exchange lifetime of the coordinated H
2
O
2
molecule and the H NMR longitudinal relaxation rate of the deuterated diamagnetic lanthanum
complex for the rotational correlation time were thoroughly investigated, and the results were com-
pared with those reported previously for other lanthanide(III) complexes. The exchange lifetimes for
2
−
[
enfGd(DO3A)(H
2
O)g
2
] (769 ± 10 ns) and [pifGd(DTTA)(H
2
O)g
2
]
(910 ± 10 ns) are significantly higher
−
2−
than those of [Gd(DOTA)(H
2
O)] (243 ns) and [Gd(DTPA)(H
2
O)] (303 ns) complexes. The rotational
2
−
correlation times for [enfGd(DO3A)(H
slightly greater than those of [Gd(DOTA)(H
marked increase in relaxivity (r
2
O)g
2
] (150 ± 11 ps) and [pifGd(DTTA)(H
2
2−
O)g
2
]
(130 ± 12 ps) are
−
2
O)] (77 ps) and [Gd(DTPA)(H O)] (58 ps) complexes. The
2
2
−
1
) of [enfGd(DO3A)(H
2
O)g
2
] and [pifGd(DTTA)(H
2
O)g
2
]
result mainly
from their longer rotational correlation time and higher molecular weight. Copyright  2004 John Wiley &
Sons, Ltd.
KEYWORDS: NMR; EPR; H NMR; H NMR; 17O NMR; Gd(III) complexes; paramagnetic complexes; proton relaxation
1
2
INTRODUCTION
particularly those with enhanced relaxivity and improved
tissue targeting functions. The main barrier to this endeavor
is the low sensitivity of MRI contrast agents when they are
coupled with the low flux of most biochemical processes. This
problem can be addressed by developing contrast agents that
Several lanthanide [e.g. Gd(III)] and transition metal [e.g.
Mn(II) and Fe(III)] complexes of polyaminocarboxylates are
either commercially available or in clinical trials for use
as magnetic resonance imaging (MRI) contrast agents. The
MR signal of body fluids can be altered by the presence of
paramagnetic water relaxation agents to result in enhanced
image contrast. The general design criteria for safe and
efficacious MRI contrast agents have been reviewed by a
1
possess higher spin–lattice relaxivity ꢀr ꢁ. The dimeric MRI
contrast agents possess two gadolinium(III) ions and will
increase relaxivity.
,3 in which
This has been confirmed for H BOꢀDO3Aꢁ
6 2
the Gd(III) complex incorporates several desirable features.
As the ligand is a DOTA [DOTA D 1,4,7,10-tetraaza-1,4,7,10-
tetrakis(carboxymethyl)cyclododecane] derivative, we can
expect high kinetic inertness and thermodynamic stability,
and as a neutral complex it is preferable from the point
of view of application (a less painful injection because of
1
,2
number of investigators. However, they remain areas for
further research and development of MRI contrast agents,
Ł
Correspondence to: Yun-Ming Wang, School of Medicinal and
Applied Chemistry, Kaohsiung Medical University, 100
Shih-Chuan 1st Road, Kaohsiung 807, Taiwan.
E-mail: ymwang@kmu.edu.tw
3C
lower osmolality). The fact that two Gd ions are bound
to one molecule allows for smaller injection volumes for the
Contract/grant sponsor: National Science Council of the Republic
of China; Contract/grant number: NSC 91-2623-7009-002-NU.
ꢀ
1
same total amount in mmol Gd kg body weight. Finally,
Copyright  2004 John Wiley & Sons, Ltd.

330
T.-M. Lee et al.
H COOC
COOCH3
HOOC
COOH
3
SOCl2
N
H
N
H
N
H
N
H
MeOH
EDDA
HOOC
COOH
N
N
N
H COOC
COOCH3
3
+
N
H
N
H
HN
HOOC
DO3A
HOOC
COOH
HOOC
COOH
N
N
N
N
K CO
2
3
MeOH
N
N
O
N
N
NH
NH
HOOC
COOH
O
2
Scheme 1. Synthesis of en(DO3A)2.
the increased molecular weight and volume may result in
a longer rotational correlation time, and thus in a higher
20 h, solvent was removed from the fraction containing
the product by rotary evaporation. The residue was dried
under vacuum and a white powder was obtained (4.76 g,
proton relaxivity.
ꢀ
]6 and
1
To extend these observations, [enꢀDO3Aꢁ
2
80.21%). H NMR (D O), υ (ppm): 3.48 ꢀNCH CH Nꢁ, 3.75
2
2
2
8
ꢀ
[
piꢀDTTAꢁ
2
]
, the derivatives of DO3A (1,4,7,10-tetraaza-
ꢀNCH COOCH ꢁ, 4.03 ꢀNCH COOCH ꢁ.
2
3
2
3
cyclododecane-1,4,7-triaacetic acid) and DTPA [1,1,4,7,7-
pentakis(carboxymethyl)-1,4,7-triazaheptane] were synthe-
sized. The number of inner-sphere water molecules was
determined from the 17O NMR chemical shift of the water
as a function of Dy(III) concentration. The water proton
en(DO3A) (2) (Scheme 1).
DO3A (1.27 g, 3.66 mmol) and anhydrous methanol (120 ml)
were mixed in a single-necked flask and the pH of the
2
spin–lattice relaxivity r
1
of the [enfGdꢀDO3AꢁꢀH
2
Oꢁg
2
] and
solution was adjusted to 11.59 with NH OH. A water-bath
was used to keep the temperature at 40–50 °C. Compound
1 (0.49 g, 2.4 mmol) was dissolved in anhydrous methanol
4
ꢀ
]2 complexes at various temperatures
[
pifGdꢀDTTAꢁꢀH
2
Oꢁg
2
1
7
and pH values are described. The EPR and O NMR trans-
verse relaxation rate data were analyzed together in a
simultaneous multiple-parameter least-squares fitting pro-
(
240 ml) and slowly added to a single-necked flask which
contained DO3A and K CO (10 g). After 48 h, the reaction
2
3
2
cedure to determine the water residence lifetime. H NMR
spectroscopy was used to determine the rotational correla-
tion time.
solution was concentrated under reduced pressure to a pale
yellow oil. The residue was dissolved in 50 ml of distilled
water and made alkaline to pH 11.0 with ammonia solution.
The solution was then applied to an AG1-X8 anion-exchange
ꢀ
resin column (200–400 mesh, HCO
2
form, 60 ml of resin
EXPERIMENTAL
and 3.0 cm column diameter). After passing through an
anion-exchange resin column, the product eluted in the
0.035 M formic acid fraction. Solvent was removed from the
fraction containing the product by rotary evaporation and
co-evaporated five times with 200 ml of water to remove the
formic acid. The residue was dried in vacuum and a pale
yellow hygroscopic powder was obtained (1.374 g, 88%)
Materials
The acid forms of the free ligand, DO3A, were synthesized
and characterized in accordance with the literature.
4
Ethylenediaminediacetic acid dimethyl ester (1).
To a solution of EDDA (ethylenediaminediacetic acid,
1
3
2
8.41 mmol) in MeOH (150 ml) warmed to 40–50 °C was
2
C NMR (50 MHz, D O) (ppm): 179.2, 178.0, 173.0, 166.7,
added SOCl (8.24 ml, 113.64 mmol) dropwise slowly. After
2
58.6, 57.4, 54.2, 53.1, 51.8, 50.8, 47.1, 46.8, 45.0, 42.6. Anal.
Copyright  2004 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2004; 42: 329–336

Variable-temperature 17O NMR of dimeric lanthanide complexes 331
HOOC
HOOC
COOH
COOH
Pyridine
Acetic Anhydride
Acetonitrile
N
N
N
COOH
DTPA
O
O
O
H N
NH2
2
O
N
N
N
K CO
2
3
O
O
COOH
3
HOOC
HOOC
COOH
HOOC
COOH
COOH
N
N
N
N
N
N
N
H
COOH
N
H
COOH
O
O
4
Scheme 2. Synthesis of pi(DTTA)2.
Calculated for C35
H
64
N
10
O
15: C, 48.60; H, 7.46; N, 16.19. 24.04. Anal. Calculated for C34H N O18: C, 47.22; H, 6.53; N,
56 8
Found: C, 48.55; H, 7.40; N, 16.10%.
12.96. Found: C, 47.16; H, 6.50; N, 12.64%.
0
00
Diethylenetriamine-N -acetic acid-N,N -dianhydride (3).
Pyridine (50 ml) and acetonitrile (50 ml) warmed to 50 °C
General
GdCl
3
Ð6H
2
O (99.9%), DyCl Ð6H
2
O (99.9%) and LaCl
3
Ð7H
2
O
3
were mixed with a solution of DTPA (50 mmol, 19.7 g) in
acetic anhydride (318 mmol, 32.4 g). After 24 h, solvent was
removed from the fraction and the residue was washed with
acetic anhydride and diethyl ether. The solid was then dried
(99.9%) were obtained from Aldrich and used without further
3
C
3C
3C
purification. The concentrations of Gd , Dy and La
were determined by chelatometric titration with EDTA using
xylenol orange as indicator. All other reagents used for the
synthesis of the ligand were purchased from commercial
sources unless noted otherwise. H and C NMR spectra
and elemental analyses were used to confirm the composition
under vacuum and a white powder was obtained (16.6 g,
1
1
13
9
2%). H NMR (200 MHz, DMSO-d
6
), υ (ppm): 3.71 (s, 8H,
), 3.30 (s, 2H, central NCH CO ), 2.72 (t,
N), 2.59 (t, 4H, NCH CH N).
terminal NCH
H, NCH CH
2
CO
2
2
2
1
7
4
2
2
2
2
of the products. O-enriched water (20.1%) was purchased
from Isotec.
[
pi(DTTA) ] (4) (Scheme 2).
2
Compound 3 (2.3 g, 6.88 mmol), K
100 ml) were mixed in a single-necked flask and trans-
,2-cyclohexanediamine (0.33 ml, 2.75 mmol) was slowly
2
CO
3
(10 g) and DMSO
Deuteration
(
1
The lanthanum complexes were synthesized by reaction of
La O with ligands in water and precipitated by addition
2 3
added. After 12 h, the solution was fractioned and filtered
by ultrafiltration using ultrafiltration membranes YM3
of acetone. Deuteration of lanthanum complexes at the ˛-
position with respect to carboxylate groups was performed
5
(
2
diameter 25 mm, MW cut-off D 3000) and YM1 (diameter
5 mm, MW cut-off D 1000). Solvent was removed by
rotary evaporation. The residue was dried under vacuum
using the procedure described by Wheeler and Legg.
Deuteration was confirmed by H NMR spectroscopy.
1
1
3
and a pale-yellow oil was obtained (1.37 g, 43%). C NMR
100 MHz, D O), υ (ppm): 172.91, 170.80, 166.81, 166.72, 56.64,
6.22, 54.40, 52.93, 52.44, 52.34, 53.03, 50.67, 50.38, 49.89, 31.28,
Complexation
The Dy(III) and Gd(III) complexes were prepared by mixing
(
2
5
solutions of hydrated LnCl
3
(10 mM) and ligand (10 mM) in a
Copyright  2004 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2004; 42: 329–336

332
T.-M. Lee et al.
2
: 1 ratio. The pH was maintained at 7.50 with 1.0 M NaOH.
0
2
4
6
8
Complex formation was instantaneous at room temperature.
The solution was then evaporated under reduced pressure
and the residue dried overnight at 60 °C.
–
–
–
–
Proton T measurements
The samples were prepared by dissolving a measured
amount of the Gd(III) chelates in water at pH 6.80 using
1
0
the buffer solution (0.10 M) PIPES (PIPES D piperazine-N,N -
bis-2-ethanesulfonic acid)–NaOH. The buffer solution was
used to maintain a constant ionic strength (i.e. 0.10M). The
0
6
.10 M buffer was sufficient to maintain the solution pH at
.80. The buffered Gd(III) chelate solutions were all allowed
to equilibrate for at least 2 h. The pH of these solutions
was determined immediately before relaxation time ꢀT ꢁ
1
measurements.
Relaxation times of aqueous solutions of gadolinium(III)
6
ꢀ
8ꢀ
complexes with [enꢀDO3Aꢁ
2
]
and [piꢀDTTAꢁ
2
]
were
– 10
2–
measured to determine the relaxivity r
1
. All measurements
[pi{Dy(DTTA)(H O)} ]
2 2
were made at 20 MHz as a function of temperature on a
Bruker Minispec NMS-120 NMR spectrometer. The samples
were contained in 5 mm glass tubes. The spectrometer was
tuned and calibrated before each measurement. The values
[en{Dy(DO3A)(H2O)}2]
DyCl3
–
12
0
10
20
30
Conc (mM)
of T
1
were measured from eight data points generated by
Figure 1. Dy(III)-induced water 17O NMR shift versus Dy(III)
chelate concentration in D2O at 25.0 š 0.1 °C.
an inversion–recovery pulse sequence. The slope of plots of
1
mM
/T
1
ꢀ
versus the concentration of Gd(III) complex gives r
1
in
1
ꢀ1
s .
unit was used to stabilize the temperature. The value of
the transverse relaxation rate was obtained by evaluating
EPR measurements
17
the linewidth at half-height ꢀꢂ1/2ꢁ of the water O signal
EPR spectra were recorded at the X-band (0.34 T) using
a Bruker ER 200D-SRC spectrometer operated in the
continuous-wave mode. The samples were contained in the
17
ꢀ
R
2
D ꢃꢂ1/2ꢁ. Solutions containing 2.6% of the O isotope
were used.
1
mm glass tubes. The cavity temperature was stabilized
2H NMR
using electronic temperature control of the gas flowing
through the cavity. Temperature was verified by substituting
a thermometer for the sample tube. Measurements were
made from 273 to 363 K. The peak-to-peak linewidth was
measured from the recorded spectra using the instrument’s
software.
The rotational correction time values of [enfLaꢀDO3Aꢁg
2
]
2ꢀ
2
and [pifLaꢀDTTAꢁg
2
]
were determined by H NMR spec-
3C
troscopy. The samples were prepared by dissolving the La
complexes in D
2
O at pH 6.80. The measurement was carried
out in a 10 mm o.d. tube on a Varian Gemini-400 (9.4 T) spec-
trometer equipped with a broadband probe and measured
by a substitution technique as described elsewhere.
5
17O NMR
The hydration numbers of [enfDyꢀDO3Aꢁg
2
]
and
RESULTS AND DISCUSSION
ꢀ
[
pifDyꢀDTTAꢁg
2
]2 were determined using the method
6
17
Dy(III)-induced water 17O NMR shifts
described by Alpoim et al. The O NMR spectra were
recorded on a Varian Gemini-400 spectrometer at 25 °C.
Figure 1 shows the Dy(III)-induced water 17O NMR shifts
versus Dy(III) chelate concentration for solutions of
1
7
17
Induced O shift (d.i.s: dysprosium(III) induced O NMR
water chemical shift) measurements were made using D O as
2
ꢀ
2
DyCl , [enfDyꢀDO3Aꢁg ] and [pifDyꢀDTTAꢁg ] in D O
2
2
2
3
an external standard. Dy(III) chelate solutions were prepared
by combining solutions of Dy(III) and ligand in a 2 : 1 ratio,
and a stoichiometric amount of standardized NaOH was
added so that the complex was fully formed. Six solutions
of various dysprosium(III) concentrations were prepared by
serial dilution of the stock solution.
Measurement of the 17O transverse relaxation rate was
carried out with a Varian Gemini-300 (7.05 T, 40.65 MHz)
spectrometer, equipped with a 5 mm probe, by using an
at 25 °C. The slopes obtained for [enfDyꢀDO3Aꢁg ] and
2
ꢀ1
2
ꢀ
2
[pifDyꢀDTTAꢁg ] at pH 6.80 are ꢀ102.6 ppm mM ꢀr D
2
ꢀ
1
2
0.9764ꢁ and ꢀ107.8 ppm mM ꢀr D 0.9762ꢁ. On the other
ꢀ
1
2
hand, the slope for DyCl is ꢀ382.8 ppm mM ꢀr D 0.999ꢁ,
3
and eight hydration numbers have been proposed for the
7
–9
dysprosium(III) ion.
Therefore, [enfDyꢀDO3Aꢁg ] and
2
2ꢀ
[pifDyꢀDTTAꢁg ] complexes contain 2.1 and 2.2 inner-
2
sphere water molecules, respectively, at pH 6.80. The actual
number of inner sphere water molecules coordinated to the
2
ꢀ
external D
2
O lock. Experimental settings were spectral width
metal center for [enfDyꢀDO3Aꢁg
2
] and [pifDyꢀDTTAꢁg
2
]
1
0 000 Hz, pulse width 7 µs, acquisition time 10 ms and no
is one per Dy(III) ion. This result is similar to those for
1
0
sample spinning. A Varian VT-J103 temperature control
[pipfGdꢀDO3AꢁꢀH
2
Oꢁg
2
] and [bisoxafGdꢀDO3AꢁꢀH
2
Oꢁg
2
].
Copyright  2004 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2004; 42: 329–336

Variable-temperature 17O NMR of dimeric lanthanide complexes 333
Relaxometric studies of the gadolinium(III)
complexes
18
The longitudinal relaxivity r
Oꢁg ] and [pifGdꢀDTTAꢁꢀH
per Gd and 5.25 mM
7.0 š 0.1 °C and 20 MHz, respectively. The
of [enfGdꢀDO3AꢁꢀH Oꢁg ] is significantly higher than
1
values of [enfGdꢀDO3Aꢁ
2
ꢀ
ꢀ1 ꢀ1
16
ꢀ
H
2
2
2
Oꢁg
per Gd at pH 6.80,
value
2
]
are 3.60 mM s
ꢀ
1
ꢀ1
s
3
r
1
1
1
1
4
2
0
8
6
4
2
2
2
ꢀ
ꢀ1 ꢀ1
those of [GdꢀDOTAꢁ]
(3.38 mM
s
per Gd,
7.0 °C)11 and [BOfGdꢀDO3AꢁꢀH
ꢀ1
ꢀ1
3
2
Oꢁg
2
]
(4.61 mM
s
3
per Gd, 37.0 š 0.1 °C)
but lower than those of
ꢀ
1
ꢀ1
12
[
pipfGdꢀDO3AꢁꢀH
2
Oꢁg
2
] (5.8 mM
Oꢁg
s
per Gd, 40.0 °C)
ꢀ
1
ꢀ1
and [bisoxafGdꢀDO3AꢁꢀH
2
2
]
(4.9 mM
s
per Gd,
1
value of
is significantly higher than that
1
2
4
[
0.0 °C). Also, the longitudinal relaxivity r
2
ꢀ
pifGdꢀDTTAꢁꢀH
2
Oꢁg
2
]
2
ꢀ
of the monomer [GdꢀDTPAꢁ]
.
r
The origin of paramagnetic relaxation enhancement is
generally divided into two components, inner-sphere and
1
3
outer-sphere:
ꢀ
1/T
i
ꢁ
p
D ꢀ1/T
i
ꢁ
innerꢀsphere C ꢀ1/T
i
ꢁ
outerꢀsphere
i D 1, 2 ꢀ1ꢁ
Inner-sphere relaxation refers to relaxation enhancement of
a solvent molecule directly coordinated to the paramagnetic
ion, and outer-sphere relaxation refers to relaxation enhance-
ment of solvent molecules in the second coordination sphere
and beyond (i.e. bulk solvent). The inner-sphere relaxation
2–
[pi{Gd(DTPA)(H O)} ]
2 2
[
en{Gd(DO3A)(H O)} ]
2
2
0
2
1
4
contribution is obtained with the equation
73
293
313
333
353
373
T (K)
is
r
D Cq/[55.6ꢀT1M C ꢄ
M
ꢁ]
ꢀ2ꢁ
1
p
Figure 2. Temperature dependence of the relaxivity for
enfGdꢀDO3AꢁꢀH2Oꢁg2] and [pifGdꢀDTTAꢁꢀH2Oꢁg2] at pH
.80 and 20 MHz.
2
ꢀ
[
6
where C is the molar concentration of the gadolinium(III)
complex, q is the number of water molecules bound to metal
ion, T1M is the longitudinal relaxation time of the bound water
is the residence lifetime of the bound water.
Because of the inverse temperature dependence of T1M and
2
ꢀ
2
98
r of [enfGdꢀDO3AꢁꢀH Oꢁg ] and [pifGdꢀDTTAꢁꢀH Oꢁg ]
,
protons and ꢄ
M
1
7
2
2
2
2
1
2
O and H NMR spectra were used to determine the values
2
98
of ꢄM and ꢄR.
ꢄ
M
, two cases can be considered: (1) fast water-exchange
is
ꢀ
T1M × ꢄ
M
ꢁ, r1 increases as temperature decreases; (2) slow
p
Water-exchange lifetime studies of Gd(III)
complexes
The measured peak-to-peak line widths, Hpp, of the
derivative spectrum can be related to the overall transverse
electronic relaxation rate, 1/T2e, vi a` Eqn (5), where g
is
water-exchange ꢀT1M − ꢄ
M
ꢁ, r1 decreases as temperature
p
decreases. Figure 2 displays a monoexponential decrease of
observed relaxivity with increasing temperature in the range
2
78–343 K. This is characteristic of fast chemical exchange
behavior, occurring when the ꢄ
molecule is much shorter than T1M of the bound water proton.
In fact, Eqn (3) can express T1M
L
is the
2
98
M
of the coordinated water
3C
1
6
isotropic Land e´ g factor (g
L
D 2.0 for Gd ):
p
1
5
:
1
g
L
ꢆ
B
ꢃ
3
D
Hpp
ꢀ5ꢁ
T
2e
h
ꢀ
ꢁ
2
2
2
1
2 ꢅ g SꢀS C 1ꢁˇ
3ꢄC1
7ꢄC2
H
The temperature dependence of transverse electronic
relaxation rates at the X-band (0.34 T) at pH 6.80
for 50 mM solution of [enfGdꢀDO3AꢁꢀH Oꢁg ] and
D
C
ꢀ3ꢁ
ꢀ4ꢁ
6
H
2
2
2
2
C2
T
1M
15
1
r
1 C ω ꢄ
1 C ω ꢄ
H
C1
S
1
1
1
2
2
D ꢄ C ꢄ C ꢄ
[pifGdꢀDTTAꢁꢀH
Oꢁg
]2 are shown in Figs 3 and 4. The
ꢀ
ꢄ
Ci
R
M
Si
2
2
1
7
data were fitted simultaneously with the following O NMR
results. Analysis of the temperature dependence of the
transverse relaxation rate for the O water nuclei is the
3
C
where S is the electron spin quantum number (7/2 for Gd ),
is the proton nuclear magnetogyric ratio, ˇ is the Bohr
magneton, g is the Land e´ factor for the free electron, r is
the distance between the metal ion and the bound water
protons, ω and ω are the respective proton and electron
ꢅ
H
17
H
most accurate method for evaluating the exchange lifetime
of the water molecules directly coordinated to the metal in
3C
17
H
S
a paramagnetic Gd chelate. According to the Swift and
Larmor frequencies and ꢄCi ꢀi D 1, 2ꢁ is the correlation time
of the modulation of the dipolar electron-proton coupling.
The overall correlation time ꢄCi receives contributions from
(the electronic relaxation time of the metal
ion) [Eqn (4)]. To understand how ꢄ
14
O
Connick theory, the paramagnetic contribution ꢀR ꢁ to the
observed transverse relaxation rate is given by
2
p
2
2
O
O
ꢀ1
O
O
M
2
2
98
298
Cq
5.6
R
C ꢀꢄ ꢁ
R
ꢀ1
C ω
ꢄ
M
, ꢄ
R
and ꢄ
S
O
O
M
ꢀ1 2M
M
2M
R
D
ꢀꢄ ꢁ
ꢀ6ꢁ
2
p
ꢂ
ꢃ
2
2
98
298
O
M
O
M
O
M
5
M
and ꢄ
R
influence
R
C ꢀꢄ ꢁ
C ω
2
Copyright  2004 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2004; 42: 329–336

334
T.-M. Lee et al.
O
2
2
2
2
2
2
1
5
R
is expressed by
M
2
ꢀ
ꢁ
ꢀ
ꢁ
2
1
3
A
h¯
ꢄ
e2
O
R
D
SꢀS C 1ꢁ
ꢄ
e1
C
ꢀ7ꢁ
ꢀ8ꢁ
2
M
2
2
1 C ω ꢄ
4
3
2
1
0
s
e2
and
ꢀ
1
ꢀ
1
O
M
ꢀ1
C Tie
ꢄ
D ꢄ
ei
where S is the electronic spin quantum number [7/2 for
Gd(III)], A/ h¯ is the Gd– O scalar coupling constant and ꢄei
1
7
T
ꢀ
i D 1, 2ꢁ represents the correlation time of the processes
modulating the scalar interaction. This modulation may
occur through both the longitudinal and the transverse
average electronic relaxation times (T1e and T2e) and the
O
mean residence lifetime ꢀꢄ ꢁ of the water molecule at the
paramagnetic site.
The temperature dependence of R is determined by
the temperature effect on ꢄ , ꢄ
modulation of the zero field splitting interaction) and ωM
according to
M
O
2M
O
(the correlation time for
M
v
O
9
2
73
293
313
333
ꢄ
ꢀ
ꢁꢅ
ꢀ
1ꢁ298.15
ꢀ
ꢄ
j
T
H
R
j
1
1
T
T (K)
ꢀ1
ꢀ
ꢄ
j
ꢁ
D
D
exp
ꢀ
ꢀ9ꢁ
T
2
98.15
298.15
Figure 3. Temperature dependence of transverse electronic
relaxation rates at the X-band (0.34 T) and pH 6.80 for a 50 mM
solution of [enfGdꢀDO3AꢁꢀH2Oꢁg2].
g
L
ꢆ
B
SꢀS C 1ꢁB A
O

ω
ꢀ10ꢁ
M
B
3k T
h¯
where the subscript j refers to the different correlation times,
H is the activation enthalpy for the corresponding dynamic
process, B is the applied magnetic field strength and k is the
B
2
2
2
2
2
2
5
4
3
2
1
j
Boltzmann constant.
The water-exchange rates for [enfGdꢀDO3AꢁꢀH
2
Oꢁg
2
]
2
ꢀ
and [pifGdꢀDTTAꢁꢀH
2
Oꢁg
2
]
were obtained by measuring
1
7
O
the O NMR transverse relaxation rate ꢀR ꢁ as a function
2
p
of temperature. The data and its best simulation according
1
6,18
to Eqns (5)–(10)
are shown in Figs 5 and 6. As there
are a large number of parameters to be determined
1
7
in the quantitative analysis of the O NMR transverse
O
T
relaxation rate ꢀR ꢁ versus T profiles, it is convenient
2
p
to fix some of them. On this basis, in addition to the
6
ꢀ1
values of q and A/ h¯ ꢀꢀ3.8 ð 10 rad s ꢁ, the value of H
M
ꢀ
1 10
is fixed at 30 kJ mol
.
The parameters which provide
the best fit of the data for [enfGdꢀDO3AꢁꢀH
2
Oꢁg
2
] and
2
ꢀ
[
pifGdꢀDTTAꢁꢀH
2
Oꢁg
2
]
are listed in Table 1. By varying
O
the temperatures over a wide range, R is dominated by
2
p
1
M
/ꢄ in the slow kinetic region at low temperatures and
is dominated by 1/ꢄei in the fast kinetic region at high
temperature.
As shown in Table 1, the water-exchange lifetime
0
2
298
ꢄ
M
of [enfGdꢀDO3AꢁꢀH
2
Oꢁg
2
]
(769 š 10 ns) is simi-
73
293
313
333
10
lar to those of [pipfGdꢀDO3AꢁꢀH
2
Oꢁg
2
] (666 ns) and
T (K)
1
2
[
[
bisoxafGdꢀDO3AꢁꢀH
2
Oꢁg
2
] (714 ns) but higher than that of
Figure 4. Temperature dependence of transverse electronic
2ꢀ
18
GdꢀDOTAꢁꢀH
2
Oꢁ] (243 ns). The higher water-exchange
relaxation rates at the X-band (0.34 T) and pH 6.80 for a 50 mM
solution of [pifGdꢀDTTAꢁꢀH2Oꢁg2]2ꢀ
lifetimes for [enfGdꢀDO3AꢁꢀH
2
Oꢁg
2
], [pipfGdꢀDO3Aꢁ
.
ꢀ
H
2
Oꢁg
2
] and [bisoxafGdꢀDO3AꢁꢀH
2
Oꢁg
2
] is perhaps due to
the decreased number of carboxylate moieties which bind to
the Gd(III) ion, so that the ligand is pulled less tightly around
the metal center and is therefore less crowded around the
where R O2 M represents the 17O transverse relaxation rate of
the coordinated water molecule and ω the chemical shift
difference between the coordinated and bulk water O NMR
resonances.
O
M
1
7
10
298
water binding site. The water-exchange lifetime ꢄ
M
of
2
ꢀ
[pifGdꢀDTTAꢁꢀH
2
Oꢁg
2
]
(910 š 10 ns) is significantly higher
Copyright  2004 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2004; 42: 329–336

Variable-temperature 17O NMR of dimeric lanthanide complexes 335
Table 1. Kinetic and NMR parameters obtained from the simultaneous fit of 17O NMR and EPR data for Gd(III)
complexes
Complex
Gd3
2
ꢀ2
20
ꢄv298 (ps)
ꢄM298 (ms)
ꢄR298 (ps)
ꢀ1
ꢁ
 ꢀs ð 10 ꢁ
H ꢀkJ mol
Ca
1.19
0.16
0.46
0.17 š 0.01
0.21 š 0.02
0.49 š 0.02
0.90 š 0.01
7.3
11
25
19 š 2
15 š 1
17 š 1
14 š 2
1.2
243
303
666
41
77
58
171 š 12
106 š 14
105 š 11
130 š 12
15.3
49.8
51.6
34.2 š 1.8
38.5 š 1.8
35.2 š 1.0
45.0 š 1.8
ꢀ
a
[
[
[
[
[
[
Gd(DOTA)ꢀH2Oꢁ]
Gd(DTPA)ꢀH2Oꢁ]
2
ꢀb
b
pipfGd(DO3A)ꢀH2Oꢁg2]
b
bisoxafGd(DO3A)ꢀH2Oꢁg2]
enfGd(DO3A)ꢀH2Oꢁg2]
714
769 š 10
910 š 10
2
ꢀ
pifGd(DTTA)ꢀH2Oꢁg2]
a
Data from Ref. 11.
Data from Ref. 15.
b
2
2
2
2
2
1
1
1
1
1
800
600
400
200
000
800
600
400
200
2500
2000
1
1
500
000
p
o
p
o
R
R
500
000
2
80
290
300
310
320
330
340
350
0
2
T (K)
75
285
295
305
315
325
Figure 5. Temperature dependence of the transverse water
T (K)
1
7O relaxation rate at 7.05 T and pH 6.80 for a 50 mM solution
Figure 6. Temperature dependence of the transverse water
O relaxation rate at 7.05 T and pH 6.80 for a 50 mM solution
of [pifGdꢀDTTAꢁꢀH2Oꢁg2] . The line represents the
simultaneous least-squares fit to all data points as described in
the text.
of [enfGd (DO3A)ꢀH2Oꢁg2].
17
2
ꢀ
ꢀ
than that of [GdꢀDTPAꢁꢀH
2
Oꢁ]2 (303 ns).10 As described in
1
9
298
the literature, a ꢄ
M
of 1000 ns creates a situation wherein
the exchange rate of water is a significant limiting factor
2
98
determining relaxivity ꢀr
increase in these multidentate chelates, making a major con-
tribution even for dimers. Therefore, the difference in r is
not only one between rigid and non-rigid linkages, but is
1 1 R
ꢁ when r increases due to the ꢄ
3
atom, its value being ¾170 kHz in the case of an sp C-
atom. The measurement was performed on diamagnetic
1
lanthanum(III) complexes deuterated in the ˛-position to the
2
98
carboxylate groups. The values of ꢄ
R
for La(III) complexes
2
98
18
also due to ꢄ
R
in the higher molecular weight molecule.
6ꢀ
6ꢀ
6ꢀ
1
2
11
with en(DO3A)
,
pip(DO3A)2
,
bisoxa(DO3A)2
,
2
8
ꢀ
10
11
[
pi(DTTA)2] , DTPA and DOTA at 310 K are given in
Rotational correlation time studies of La(III)
complexes
2
98
Table 1. The ꢄ
R
values for Gd(III) dimers are significantly
2
ꢀ
In diamagnetic molecules, the relaxation rate of the 2
10
higher than those for Gd(III) monomers, [Laꢀꢀ H ꢁDOTAꢁ]
H
1
1
2
2ꢀ
11
(
77 ps) and [Laꢀꢀ H10ꢁDTPAꢁ] (58 ps). Thus, the change
nucleus is predominantly determined by the quadrupolar
5
in the molecular weight in these complexes alters ꢄR sig-
nificantly. On the other hand, the lower rotational cor-
mechanism, which is strictly intramolecular and modulated
by the sole rotation of the molecule. For fast-tumbling
systems, the relaxation rate is thus directly related to the
2
98
2
relation time ꢀꢄ
R
2
ꢁ values for [enfLaꢀꢀ H10ꢁ(DO3A)ꢁg
2
]
2
98
and [bisoxafLaꢀꢀ H ꢁ(DO3A)ꢁg ] compared with that of
rotational correlation time ꢀꢄ
R
ꢁ:
10
2
2
18
ꢀ
ꢁ
2
[pipfLaꢀꢀ H10ꢁ(DO3A)ꢁg
2
]
indicates that the more flexible
1
3
8
e2qQ
h
linker between the two macrocyclic chelating moieties in
R
1
D T
D
ꢄ
R
ꢀ11ꢁ
6ꢀ
6ꢀ
1
en(DO3A)2 and bisoxa(DO3A)2 decreases the ꢄ
R
value
2
where the quadrupolar coupling constant ꢀe qQ/hꢁ depends
on the hybridization state of the C-atom carrying the
and causes the lower relaxivity ꢀr ꢁ of the Gd(III) complexes.
1
2
H
The relationship between r
1
and the molecular weight of
Copyright  2004 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2004; 42: 329–336

336
T.-M. Lee et al.
8
7
6
5
4
3
2
1
0
0
200
400
600
800
1000
1200
1400
1600
Molecular Weight
Figure 7. Correlation of molecular weight of gadolinium(III) complexes with HP-DO3A (ꢀ), PA-DO3A (ꢁ), B22F (
ž
), B22 (°), TU1 (ð),
6ꢀ
8ꢀ
TU2 (ꢂ), en(DO3A)2 (ꢃ) and [pi(DTTA) ] (C) with relaxivity.
2
Gd(III) monomer and dimer complexes19 is shown in Fig. 7.
The results show that there is a strong correlation between
REFERENCES
1
. Reichert DE, Welch MJ. Coord. Chem. Rev. 2001; 212: 111.
r
1
and molecular weight, which means that the relaxivity
´
2. T o´ th E, Burai L, Merbach AE. Coord. Chem. Rev. 2001; 216–217:
value of the monomer generally increases with molecular
weight. However, in order to maximize the relaxivity gain,
the linking group and molecular weight must be taken into
account for Gd(III) dimer complexes.
363.
´
3
. T o´ th E, Vauthey S, Pubanz D, Merbach AE. Inorg. Chem. 1996;
5: 3375.
3
4
. Dischino DD, Delaney EJ, Emswiler JE, Gaughan GA, Prasad JS,
Srivastava SK, Tweedle MF. Inorg. Chem. 1991; 30: 1265.
. Wheeler WD, Legg JJ. Inorg. Chem. 1985; 24: 1292.
. Alpoim MC, Urbano AM, Geraldes CFGC, Peters JA. J. Chem.
Soc., Dalton Trans. 1992; 463.
5
6
CONCLUSION
7
. Kowall T, Foglia F, Helm L, Merbach AE. J. Am. Chem. Soc. 1995;
1
17: 3790.
1
7
From analysis of the O NMR relaxometric properties,
8. Cossy C, Helm L, Powell DH, Merbach AE. New J. Chem. 1995;
19: 27.
. Cossy C, Barnes AC, Enderby JE, Merbach AE. J. Chem. Phys.
2
98
the larger water-exchange lifetime ꢄ
M
for [enfGdꢀDO3Aꢁ
9
ꢀ
H
2
Oꢁg
2
] is perhaps due to a decrease in the number of
1989; 90: 3254.
carboxylate moieties, so that the ligand is pulled less tightly
around the metal center and there is less crowding around
the water binding site. We have demonstrated relaxivity
s ꢁ enhancement through the incorporation
of rigidifying elements in the linkers or increasing the
molecular weight of chelates. The water proton spin–lattice
1
0. Powell DH, Ni Dhubhghaill OM, Pubanz D, Helm L,
Lebedev YS, Schlaepfer W, Merbach AE. J. Am. Chem. Soc. 1996;
118: 9333.
ꢀ
1
ꢀ1
1
1
1
1. Micskei K, Helm L, Br u¨ cher E, Merbach AE. Inorg. Chem. 1993;
ꢀ
r
1
/Gd mM
3
2: 3844.
2. Carvalho J, Watson AD, Fellmann JD, Koo MD. US Patent 5 650
33, 1988.
1
2
ꢀ
relaxivity of [pifGdꢀDTTAꢁꢀH
2
Oꢁg
2
]
is higher than that for
3. Caravan P, Ellison JJ, McMurry TJ, Lauffer RB. Chem. Rev. 1999;
99: 2293.
4. Swift TJ, Connick RE. J. Chem. Phys. 1962; 37: 307.
5. Aime S, Gianolio E, Terreno E, Giovenzana GB, Pagliarin R,
Sisti M, Palmisano G, Botta M, Lowe MP, Parker D. J. Biol. Inorg.
Chem. 2000; 5: 488.
2
ꢀ
the monomer [Gd(DTPA)] owing to its longer rotational
correlation time and greater molecular weight. Furthermore,
approaches aimed at enhancing relaxivity by modulating
1
1
2
98
the water-exchange lifetime, ꢄ
M
, will be important for the
future development of molecular MRI contrast agents used
in imaging biochemical processes.
16. Reuben J. J. Phys. Chem. 1971; 75: 3164.
17. Aime S, Botta M, Crich SG, Giovenzana G, Pagliarin R, Sisti M,
Terreno E. Magn. Reson. Chem. 1998; 36: S200.
1
8. Aime S, Botta M, Fasano M, Terreno E. Acc. Chem. Res. 1999; 32:
41.
9
Acknowledgement
We are grateful to the National Science Council of the Republic of
China for financial support under Contract No. NSC 91-2623-7009-
1
9. Ranganathan RS, Fernandez ME, Kang SI, Nunn AD, Ratsep PC,
Pillai KMR, Zhang X, Tweedle MF. Invest. Radio. 1998; 11: 779.
002-NU.
Copyright  2004 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2004; 42: 329–336