Separation and characterization of the two diastereomers for [Gd(DTPA-bz-NH2)(H2O)]2-, a common synthon in macromolecular MRI contrast agents: Their water exchange and isomerization kinetics

Article abstract of DOI:10.1021/ic048645d

Chiral, bifunctional poly(amino carboxylate) ligands are commonly used for the synthesis of macromolecular, Gd^IIIbased MRI contrast agents, prepared in the objective of increasing relaxivity or delivering the paramagnetic Gd^III to a specific site (targeting). Complex formation with such ligands results in two diastereomeric forms for the complex which can be separated by HPLC. We demonstrated that the diastereomer ratio for Ln ^III DTPA derivatives (～60:40) remains constant throughout the lanthanide series, in contrast to Ln^III EPTPA derivatives, where it varies as a function of the cation size with a maximum for the middle lanthanides (DTPA^5- = diethylenetriaminepentaacetate; EPTPA ^5- = ethylenepropylenetriaminepentaacetate). The interconversion of the two diastereomers, studied by HPLC, is a proton-catalyzed process (/cobs = /d[H+]). It is relatively fast for [Gd(EPTPA-bz-NH₂)(H ₂O)]^2- but slow enough for [Gd(DTPA-bz-NH ₂)(H₂O)]^2- to allow investigation of pure individual isomers (isomerization rate constants are k₁ = (3.03 + 0.07) ′ 10⁴ and 11.6 ±0.5 s^-1 M^-1 for [Gd(EPTPA-bz-NH₂)(H₂O)]^2- and [Gd(DTPA-bz-NH₂)-(H₂O)]², respectively). Individual water exchange rates have been determined for both diastereomers of [Gd-(DTPA-bz-NH₂)(H₂O)]² by a variable-temperature ¹⁷O NMR study. Similarly to Ln^III EPTPA derivatives, k_ex values differ by a factor of 2 (k _ex⁹⁸ = (5.7 ± 0.2) × 10⁶ and (3.1 ± 0.1) × 10^{6 s-1}). This variance in the exchange rate has no consequence on the proton relaxivity of the two diastereomers, since it is solely limited by fast rotation. However, such difference in k_ex will affect proton relaxivity when these diastereomers are linked to a slowly rotating macromolecule. Once the rotation is optimized, slow water exchange will limit relaxivity; thus, a factor of 2 in the exchange rate can lead to a remarkably different relaxivity for the diastereomer complexes. These results have implications for future development of Gd^III-based, macromolecular MRI contrast agents, since the use of chiral bifunctional ligands in their synthesis inevitably generates diastereomeric complexes.

Full text of DOI:10.1021/ic048645d

Inorg. Chem. 2005, 44, 3561−3568
Separation and Characterization of the Two Diastereomers for
-
[Gd(DTPA-bz-NH2)(H2O)]2 , a Common Synthon in Macromolecular MRI
Contrast Agents: Their Water Exchange and Isomerization Kinetics
La´szlo´ Burai,† va To´th,* Ange´lique Sour,† and Andre´ E. Merbach†
EÄ
Laboratoire de Chimie Inorganique et Bioinorganique, Ecole Polytechnique Fe´de´rale de
Lausanne, EPFL-BCH, CH-1015 Lausanne, Switzerland
Received September 27, 2004
Chiral, bifunctional poly(amino carboxylate) ligands are commonly used for the synthesis of macromolecular, GdIII-
based MRI contrast agents, prepared in the objective of increasing relaxivity or delivering the paramagnetic GdIII
to a specific site (targeting). Complex formation with such ligands results in two diastereomeric forms for the
complex which can be separated by HPLC. We demonstrated that the diastereomer ratio for LnIII DTPA derivatives
(
∼
60:40) remains constant throughout the lanthanide series, in contrast to LnIII EPTPA derivatives, where it varies
-
as a function of the cation size with a maximum for the middle lanthanides (DTPA5
) diethylenetriaminepentaacetate;
EPTPA5
) ethylenepropylenetriaminepentaacetate). The interconversion of the two diastereomers, studied by
-
-
HPLC, is a proton-catalyzed process (kobs
enough for [Gd(DTPA-bz-NH2)(H2O)]2- to allow investigation of pure individual isomers (isomerization rate constants
)
k1[H+]). It is relatively fast for [Gd(EPTPA-bz-NH2)(H2O)]2 but slow
1
1
-
are k1
) (3.03 ± 0.07) × ±
104 and 11.6 0.5 s- M- for [Gd(EPTPA-bz-NH2)(H2O)]2 and [Gd(DTPA-bz-NH2)-
(H2O)]2-, respectively). Individual water exchange rates have been determined for both diastereomers of [Gd-
(DTPA-bz-NH2)(H2O)]2 by a variable-temperature 17O NMR study. Similarly to LnIII EPTPA derivatives, kex values
-
298
differ by a factor of 2 (kex
) (5.7 ± 0.2) × ± 0.1) ×
106 and (3.1 106 s-1). This variance in the exchange rate
has no consequence on the proton relaxivity of the two diastereomers, since it is solely limited by fast rotation.
However, such difference in kex will affect proton relaxivity when these diastereomers are linked to a slowly rotating
macromolecule. Once the rotation is optimized, slow water exchange will limit relaxivity; thus, a factor of 2 in the
exchange rate can lead to a remarkably different relaxivity for the diastereomer complexes. These results have
implications for future development of GdIII-based, macromolecular MRI contrast agents, since the use of chiral
bifunctional ligands in their synthesis inevitably generates diastereomeric complexes.
Introduction
agents in magnetic resonance imaging (MRI).3 In medical
applications involving metal chelates, DTPA5- is one of the
most commonly used ligands. This is mainly explained by
the remarkable thermodynamic stability of the DTPA
complexes, one of the highest among linear chelates.
[Gd(DTPA)(H2O)]2- was the first MRI contrast agent
approved for clinical application (1988) and probably still
remains the most widespread.
During the past decades, metal complexes have acquired
an important role in medical diagnosis and therapy. Com-
plexes of γ-emitting metal ions such as 99mTc, 111In, or 67Ga
are useful tools for obtaining diagnostic information,1 while
those of particle-emitting isotopes such as the â-emitter 90
Y
or 186Re can be used in internal radiotherapy.2 Paramagnetic
GdIII complexes are extensively applied as contrast enhancing
The attachment of a “reactive” pending arm to DTPA or
other small ligands leads to bifunctional agents and allows
* Author to whom correspondence should be addressed. E-mail:
eva.jakabtoth@epfl.ch. Tel: +41 21 693 9878. Fax: +41 21 693 9875.
† E-mail: laszlo.burai@epfl.ch (L.B.); angelique.sour@epfl.ch (A.S.);
andre.merbach@epfl.ch (A.E.M.).
(1) (a) Schwochau, K. Angew. Chem., Int. Ed. Engl. 1994, 33, 2258. (b)
McMurry, T. C.; Brechbiel, M.; Wu, C.; Gansow, O. A. Bioconjugate
Chem. 1993, 4, 236. (c) Craig, A. S.; Parker, D.; Adams, H.; Bailey,
N. A. J. Chem. Soc., Chem. Commun. 1989, 1793.
(2) (a) Moi, A. K.; Meares, C. F.; DeNardo, S. J. J. Am. Chem. Soc. 1988,
110, 6266. (b) Heppeler, A.; Froidevaux, S.; Ma¨cke, H. R.; Jermann,
E.; Be´he´, M.; Powell, P.; Hennig, M. Chem.sEur. J. 1999, 5, 1974.
(3) To´th, EÄ .; Helm, L.; Merbach, A. E. Relaxivity of Gadolinium(III)
Complexes: Theory and Mechanism. In The Chemistry of Contrast
Agents in Medical Magnetic Resonance Imaging; To´th, EÄ ., Merbach,
A. E., Eds.; Wiley: Chichester, U.K., 2001; pp 45-120.
10.1021/ic048645d CCC: $30.25
Published on Web 04/15/2005
© 2005 American Chemical Society
Inorganic Chemistry, Vol. 44, No. 10, 2005 3561

Burai et al.
Chart 1
agent. Diastereomers having different structure can exhibit
different water exchange rate. So far, no data are known for
the water exchange of individual isomers of chiral [Gd(DTPA)-
(H2O)]2- derivatives. Recently in a preliminary communica-
tion on [Ln(EPTPA-bz-NH2)(H2O)]2- and [Ln(EPTPA-bz-
NO2)(H2O)]2- diastereomers, we have reported that the
diastereomeric ratio (93:7) is independent of the pending arm,
whereas it remarkably changes along the lanthanide series
(Chart 1).8 For the first time, we have determined the
individual water exchange rates for the two diastereomers
of these complexes. In the case of both [Ln(EPTPA-bz-NH2)-
(H2O)]2- and [Ln(EPTPA-bz-NO2)(H2O)]2-, the water ex-
change was found to be twice as fast for the minor than for
the major isomer. The Gd(EPTPA) derivatives can be
regarded as potential building blocks for macromolecular
MRI contrast agents, since their water exchange is faster thus
closer to optimal than that for typical commercial GdIII
complexes.9 The water exchange rate plays a much more
important role for slowly rotating macromolecular complexes
than for small chelates, as the relaxivity of macromolecules
can be largely limited by slow exchange, whereas, for small
complexes, it is (almost) exclusively fast rotation that limits
relaxivity. For the development of highly efficient macro-
molecular contrast agents one should be able to fine-tune
the water exchange rate so that it falls into an optimal and
relatively restricted range. Consequently, differences such
as that found for the [Gd(EPTPA-bz-NH2)(H2O)]2- and [Gd-
(EPTPA-bz-NO2)(H2O)]2- diastereomers can be of great
importance. The diastereomer with a more optimal water
exchange rate bound to a macromolecule will exhibit higher
relaxivity, provided the isomerization rate is slow enough
to avoid the interconversion of the isomer during the time
of a medical MRI investigation.
covalent binding to macromolecules.4 Macromolecules such
as monoclonal antibodies carrying one or several complexes
can vehicle e.g. the radioactive metal to a specific tissue in
selective cancer therapy. In MRI, large molecules loaded with
GdIII have an improved efficiency due to their increased size,
or they can also allow for targeted imaging.5
The isothiocyanato functional group is a very common
linker for attaching small ligands such as DTPA to macro-
molecules. In the coupling reaction, the p-aminobenzyl
derivative of the ligand is the starting compound. The -NH2
of the p-aminobenzyl pending arm is transformed to an
isothiocyanato function prior to the reaction with the
macromolecule. These bifunctional, p-aminobenzyl deriva-
tives are commercially available for a variety of ligands, and
they are indeed widely used in the synthesis of macromol-
ecules for biomedical applications. The presence of the
p-aminobenzyl (or other) pending group in the 4-position of
DTPA (Chart 1) generates a chiral center. On coordination
of the central nitrogen (6-position) to the metal ion, a second
chiral center arises which leads to two diastereomers (4S6S,
4R6R and 4S6R, 4R6S). Diastereomers may have different
physicochemical properties and can be separated by chro-
matographic methods. HPLC separation has been performed
in the case of some Ln(DTPA) derivatives. For [Y(DTPA-
bz-NH2)(H2O)]2-, Cummins et al. reported a 60:40 diaster-
eomeric ratio as determined by HPLC.6 Similar values (63:
37) were found by NMR spectroscopy as well. A detailed
HPLC study was reported for the Gd3+ complex of another
chiral DTPA derivative, [Gd(S-EOB-DTPA)(H2O)]2-, which
is a new liver-specific MRI contrast agent (Primovist,
Schering AG).7 Interestingly, the diastereomeric ratio was
very similar, 65:35. After preparative HPLC separation of
the [Gd(S-EOB-DTPA)(H2O)]2- diastereomers, the kinetics
of their interconversion was also studied. Under physiological
conditions it is very slow; several days are needed to return
to equilibrium.
The objective of this study was 2-fold: (i) to establish
the isomerization kinetics of diastereomers for [Ln(DTPA-
bz-NH2)(H2O)]2- and Ln(EPTPA-bz-NH2)(H2O)]2-; (ii) to
determine the individual parameters influencing relaxivity,
in particular water exchange rate, for both [Gd(DTPA-bz-
NH2)(H2O)]2- diastereomers. To this end, the diastereomers
of the lanthanide complexes have been separated by HPLC
and their reconversion studied as a function of the pH. On
the individual diastereomers of [Gd(DTPA-bz-NH2)(H2O)]2-
separated by HPLC we have measured variable-temperature
transverse and longitudinal 17O relaxation rates and chemical
shifts, multiple field water proton relaxivities, and EPR
spectra.
Experimental Section
With regard to MRI contrast agent application, the water
exchange rate of the GdIII complex is one of the crucial
parameters to influence relaxivity, thus efficiency of the
Materials. H5DTPA-bz-NH2 was purchased from Macrocyclics
and used without further purification. H5EPTPA-bz-NH2 (3) was
synthesized according to Scheme 1. Compound 110 (5.59 g, 6.79
mmol) was dissolved under argon atmosphere in MeOH (120 mL).
Pd/C (10%; 0.285 g) was added, and the mixture was cooled to 0
°C. H2 gas was bubbled into the solution for 24 h. The mixture
(4) Meares, C. F. M.; Wensel, T. G. Acc. Chem. Res. 1984, 17, 202.
(5) Konda, S. D.; Aref, M; Brechbiel, M.; Wiener, E. C. InVest. Radiol.
2000, 35, 50.
(6) Cummins, C. H.; Rutter, E. W.; Fordyce, W. A., Jr. Bioconjugate
Chem. 1991, 2, 180.
(7) Schmitt-Willich, H.; Brehm, M.; Ewers, Ch. L J.; Michl, G.; Mu¨ller-
Fahrnow, A.; Petrov, O.; Platzek, J.; Radu¨chel, B.; Su¨lzle, D. Inorg.
Chem. 1999, 38, 1134.
(8) Burai, L.; To´th, EÄ .; Merbach, A. E. Chem. Commun. 2003, 2680.
(9) Laus, S.; Ruloff, R.; To´th, EÄ .; Merbach, A. E. Chem.sEur. J. 2003,
9, 3555.
(10) Corson, D. T.; Meares, C. F. Bioconjugate Chem. 2001, 11, 292.
3562 Inorganic Chemistry, Vol. 44, No. 10, 2005

[Gd(DTPA-bz-NH2)(H2O)]2- Diastereomers
Scheme 1. Synthesis of H5EPTPA-bz-NH2 (3)
was then stirred for 2 h at room temperature. Celite was added,
and the mixture was filtered through a pad of Celite to remove the
catalyst. The filtrate was evaporated to dryness, and the resulting
yellow oil 2 (4.98 g, 92%) was used without further purification.
1H NMR (CDCl3, 400 MHz): δ ) 1.41 (s, 9H), 1.44 (s, 18H),
1.45 (s, 18H), 1.52 (m, 2H), 2.28-2.50 (m, 8H), 3.03 (m, 1H),
3.28 (s, 2H), 3.37 (s, 4H), 3.43 (s br, 4H), 6.57 (d, 2H, J ) 7.8
Hz), 6.98 (d, 2H, J ) 7.8 Hz). MS (ESI): m/z 793.3, [M + H]+.
Compound 2 (4.8 g, 6.05 mmol) was dissolved in TFA (65 mL)
and stirred at room temperature for 24 h. The solvent was then
evaporated to dryness, and the resulting crusty yellow powder 3
(5.86 g, 99%) was obtained without further purification and stored
buffer). The collected fractions of the minor species were kept at
25 °C, and 25 µL of those was reinjected at appropriate time
intervals.
Method 3. A semipreparative HPLC method was developed for
the separation of [Gd(DTPA-bz-NH2)(H2O)]2- isomers to be used
in the 17O NMR and 1H NMRD measurements. A 600 µL volume
of 0.027 M [Gd(DTPA-bz-NH2)(H2O)]2- solution was injected in
5 portions to the Waters XTerra Prep RP18 column (10 µm, 19 ×
150 mm). The eluent was 0.04 M NH4HCO3 with 10 mL/min flow
rate. Two fractions for the isomers A and B were collected
(approximately 150 mL). The eluates were concentrated to 20 mL
by evaporation (35 °C, 30 mbar), and then 40 mL of methanol was
added and evaporated to dryness. This was repeated twice with 20
mL methanol to completely remove NH4HCO3. The final products
were redissolved in 300 µL of 0.1 M TRIS buffer and used for the
1
at -18 °C. H NMR (D2O, 400 MHz): δ ) 2.20 (m, 2H), 2.74
(dd, J ) 8 Hz, J ) 13.8 Hz, 1H), 3.08 (dd, J ) 13.6 Hz, J ) 6.8
Hz, 1H), 3.30-3.57 (m, 11H), 4.07 (s, 1H), 4.13 (s, 1H), 4.18 (s,
4H), 7.36 (d, J ) 8.4 Hz, 2H), 7.42 (d, J ) 8.4 Hz, 2H). MS
(ESI): m/z 513.2, [M + H]+.
1
17O NMR and H NMRD experiments. The purity of the isomers
was checked by method 1 and found to be 98.8% and 98.9% for A
and B, respectively. The concentration of the samples was
determined from the peak area of A and B using the [Gd(DTPA-
bz-NH2)(H2O)]2- calibration series. In certain cases, the GdIII
concentration of the samples was also verified by ICP measure-
ments.
All other reagents used were obtained from commercial sources.
Ln(ClO4)3 stock solutions were prepared by dissolving Ln2O3 under
reflux in a slight excess of HClO4 (Merck, pa, 70%) in double-
distilled water, followed by filtration. The concentration of Ln-
(ClO4)3 solutions was determined by titration with a Na2H2EDTA
solution using xylenol orange as indicator and urotropin for pH
regulation. For the preparation of the LnL complexes, weighed
quantities of solid ligands were dissolved in buffer solution (PIPES
or TRIS, 0.01-0.1 M), and then a weighed amount of Ln(ClO4)3
stock solution was added dropwise to form the chelate complexes.
The pH of the solutions was adjusted with 1 M NaOH, HClO4,
or HCl solutions, and the pH was measured with a combined
glass electrode (C14/02-SC, Moeller Scientific Glass Instruments,
Effretikon, Switzerland) connected to a Metrohm 692 pH/ion-meter
calibrated with Metrohm buffer solutions.
17O NMR Measurements. For the variable-temperature studies,
the [Gd(DTPA-bz-NH2)(H2O)]2- solutions were filled with a
syringe into glass spheres which were fitted into 10 mm NMR tubes.
Glass spheres are used to eliminate susceptibility effects on the
chemical shift.11 The 1/T1 and 1/T2 relaxation rates and chemical
shifts were measured with regard to a 0.1 M TRIS aqueous solution
(pH ) 9.0) as external reference. To improve sensitivity, 17O-
enriched water (10% H217O, Yeda (Israel)) was added to the
solutions to yield in 1-2% 17O enrichment. The 17O NMR
measurements were performed on a Bruker ARX-400 spectrometer
at 9.4 T and 54.2 MHz. Bulk water longitudinal relaxation rates,
1/T1, were obtained by the inversion recovery method,12 and
transverse relaxation rates, 1/T2, by the Carr-Purcell-Meiboom-
Gill spin-echo technique.13 The least-squares fit of 17O NMR data
was performed with the program Scientist for Windows by
Micromath, version 2.0. The reported errors correspond to one
standard deviation obtained by the statistical analysis.
1H NMRD Measurements. The 1/T1 nuclear magnetic relaxation
dispersion (NMRD) profiles of the solvent protons at 25 °C were
obtained on a Spinmaster FFC fast field cycling NMR relaxometer
(Stelar), covering a continuum of magnetic fields from 2.35 × 10-4
to 0.47 T (corresponding to a proton Larmor frequency range 0.02-
20 MHz). Higher frequency measurements were performed on a
50, 100, and 200 MHz magnet, connected to an Avance-200
console, and at 400 MHz on a Bruker ARX-400 spectrometer.
EPR. EPR spectra have been recorded at Q-band (35 GHz; at
ambient temperature) on Elexsys E500 and at W-band (95 GHz; T
) 286 K) on Elexsys E680 Bruker spectrometers.
HPLC Methods. All HPLC experiments were performed on a
Waters 600 instrument equipped with a Waters 2487 dual-
wavelength UV-vis detector and a Waters Fraction Collector II.
The detection wavelengths were 240 and 290 nm.
Method 1. The diastereomer ratios for [Ln(DTPA-bz-NH2)-
(H2O)]2- complexes were determined on Waters XTerra RP18 (5
µm, 4.6 × 150 mm) column using buffered aqueous eluent (10
mM TRIS, pH ) 8, 3% acetonitrile) with 0.5 mL/min flow rate.
Method 2. In the study of the isomerization kinetics of [Gd-
(DTPA-bz-NH2)(H2O)]2-, a YMC Hydrosphere C18 column (5 µm,
4.6 × 150 mm) was used. The eluent was water containing 10 mM
PIPES buffer (pH 8) and 5% acetonitrile with 0.5 mL/min flow
rate. The 10 µL portions of 4 mM [Gd(DTPA-bz-NH2)(H2O)]2-
solution were separated collecting two fractions of pure isomers A
and B (2 mL each). Five samples of each isomer (A and B) were
prepared by setting the pH to 7.86, 7.27, 6.68, 6.11, and 5.60 and
kept at 25 °C. The 25 µL portions of the samples were injected
after given periods of time, and the quantity of the two isomers
was determined.
The isomerization kinetics of [Gd(EPTPA-bz-NH2)(H2O)]2- was
established by modifying method 1 (2% acetonitrile and 1.0 mL/
min flow rate). The 10 µL portions of the 8.7 mM solution were
separated at different pHs (7.49, 7.61, 7.71, 7.91, and 8.91; TRIS
(11) Hugi, A. D.; Helm, L.; Merbach, A. E. HelV. Chim. Acta 1985, 68,
508.
(12) Vold, R. V.; Waugh, J. S.; Klein, M. P.; Phelps, D. E. J. Chem. Phys.
1968, 48, 3831.
(13) Meiboom, S.; Gill, D. ReV. Sci. Instrum. 1958, 29, 688.
Inorganic Chemistry, Vol. 44, No. 10, 2005 3563

Burai et al.
Figure 2. Percentage of the major (A) isomer of [Ln(DTPA-bz-NH2)-
(H2O)]2- (white column) and [Ln(EPTPA-bz-NH2)(H2O)]2- (dark column)
complexes.
Figure 1. HPLC separation of [Gd(DTPA-bz-NH2)(H2O)]2- diastereomers.
In the analysis of all chromatograms, as recorded by UV-
vis detection, we assume that the two diastereomers have
identical molar absorption coefficients.7 This was previously
proved for the Eu(EPTPA-bz-NH2)2- diastereomers: here
the transformation of B to A was also monitored by recording
1H NMR spectra as a function of time following the mixing
of the ligand with Eu3+.8 In parallel, HPLC injections were
done after the same periods of time. In the NMR spectrum
two series of signals appear that we assign to the two
diastereomers. At any time point, the integral of their NMR
peaks corresponds to the integral of the HPLC peaks. On
the other hand, the A:B ratio of [Eu(DTPA-bz-NH2)(H2O)]2-
determined from 1H NMR does not change with temperature
(<1% between 0 and 50 °C) or pH (5 < pH < 9).
Results and Discussion
The two diastereomeric forms, A and B, of [Ln(DTPA-
bz-NH2)(H2O)]2- complexes were separated by HPLC for
the whole series of lanthanides (following the notation
introduced by Schmitt-Willich et al. for [Gd(S-EOB-DTPA)-
(H2O)]2-,7 we call the two components A and B according
to the order of elution). For each lanthanide complex, one
major (A) and one minor peak (B) appear in the chromato-
gram with retention times ranging between 5.54 and 6.01
min and 7.47-8.44 min, for A and B, respectively (Figure
1). The ratio of the two isomers has been determined from
the peak areas and found to be practically constant (around
60:40 for A:B; see Supporting Information Table S1) along
the lanthanide series. The ratio of 62:38 obtained by us for
[Y(DTPA-bz-NH2)(H2O)]2- is in good agreement with the
60:40 ratio published by Cummins et al.6 Similar and metal-
independent ratios were reported for EOB-DTPA complexes
(65:35, 60:40, and 57:43 for Gd, La, and Eu complexes,
respectively).7,14
The lanthanide-independent diastereomeric ratios for the
[Ln(DTPA-bz-NH2)(H2O)]2- and other DTPA-derivative
complexes contrast strikingly with those communicated for
[Ln(EPTPA-bz-NH2)(H2O)]2- and [Ln(EPTPA-bz-NO2)-
(H2O)]2-, where there is a large variation in the ratio along
the lanthanide series.8 In comparison to DTPA,5- the ligand
EPTPA5- possesses one additional CH2 group in the amine
backbone. Figure 2 compares the percentage of the major
(A) isomer for [Ln(DTPA-bz-NH2)(H2O)]2- and [Ln(EPTPA-
bz-NH2)(H2O)]2- complexes. In the case of [Ln(EPTPA-bz-
NH2)(H2O)]2-, the fraction of the A isomer shows a
maximum in the middle of the Ln series. Noteworthy, on
mixing of buffered EPTPA and lanthanide solutions, the
initial diastereomer ratio is around 50:50 for any lanthanide.
Afterward, a slow interconversion of B to A occurs until
the equilibrium values as indicated in Figure 2 are reached.
The explanation of this phenomenon is still an open question.
In contrast, such interconversion following the complex
formation is not observed for DTPA complexes; the A:B
ratio measured immediately after mixing the ligand and the
metal remains constant even after several months.
It has also to be noted that the assignment of the
chromatographic peaks to the corresponding diastereomeric
structure is still not clarified. The commercially available
DTPA-bz-NH2 ligand is in a racemic form (4R and 4S); thus,
the two chromatographic peaks in Figure 1 represent most
likely the coelution of 4S6S/4R6R and 4S6R/4R6S diaster-
1
eomers. A H and 13C NMR study has been performed on
the diamagnetic La3+ complex of the optically pure S-EOB-
DTPA.7 It was concluded that the two isomers differ in the
configuration of the central nitrogen (6th position). Most
presumably, the major isomer (A) has 4S6R, whereas the
minor (B) isomer has 4S6S configuration. Additional NMR
and chiroptical experiments on [Eu(S-EOB-DTPA)(H2O)]2-
support this assumption and prove the Λ helicity around the
metal center.14
Isomerization Rate. In the preliminary communication
on [Gd(EPTPA-bz-NH2)(H2O)]2- and [Gd(EPTPA-bz-NO2)-
(H2O)]2-, we reported the relaxation properties of the two
individual isomers; however, detailed kinetic measurements
regarding the interconversion of the isomers were not
performed.8 We qualitatively observed that the rate of
interconversion was relatively fast, and this prevented us from
measuring the properties of A and B in HPLC-separated
samples containing exclusively one isomer.
In contrast to EPTPA, isomer transformation was found
to be slow for EOB-DTPA complexes.7 A variable-pH and
-temperature kinetic study on [Gd(EOB-DTPA)(H2O)]2-
showed that the transformation is a slow, acid-catalyzed
process, with t1/2 ) 150.9 h at pH 7.0 and 25 °C.7
(14) Thompson, N. C.; Parker, D.; Schmitt-Willich, H.; Su¨lzle, D.; Muller,
G.; Riehl, J. P. J. Chem. Soc., Dalton Trans. 2004, 1892.
3564 Inorganic Chemistry, Vol. 44, No. 10, 2005

[Gd(DTPA-bz-NH2)(H2O)]2- Diastereomers
Figure 3. kobs rate constants as a function of pH for the isomerization of
[Gd(DTPA-bz-NH2)(H2O)]2- diastereomers and the result of the fit. Squares
and the solid line represent the reaction from A to B, and circles and the
dashed line, the reaction from B to A.
Figure 4. kobs rate constants as a function of pH for the isomerization of
[Gd(EPTPA-bz-NH2)(H2O)]2- diastereomers.
of the diastereomers occurs via the breaking and re-formation
of the coordination bond between the metal and the donor
atoms, the pending arm on the ligand has only limited
influence on the interconversion rate. The faster isomerization
of [Gd(DTPA-bz-NH2)(H2O)]2- may be the consequence of
the amino group on the benzyl arm which can protonate and
assist the acid catalysis. (The protonation constant has not
been determined for this amine group; however, it is known
for aniline, log K ) 4.7). Presumably, a spontaneous
isomerization pathway without participation of protons,
described by the k0 rate constant, cannot be excluded either.
Dissociation and metal exchange kinetic studies on [Ln(DT-
PA)(H2O)]2- derivatives showed that spontaneous dissocia-
tion also occurs in addition to the proton-catalyzed process.15
However, the k0 rate constant was obtained with a large error;
thus, at least in this pH range, no reliable value of k0 can be
calculated. Most likely k0 is on the order of 10-8-10-9 s-1.
For a direct comparison, the isomerization rate of [Gd-
(EPTPA-bz-NH2)(H2O)]2- diastereomers has also been mea-
sured. Only the transformation of the minor isomer of
[Gd(EPTPA-bz-NH2)(H2O)]2- was monitored after separa-
tion. The pH range of these experiments was higher than
that of used for [Gd(DTPA-bz-NH2)(H2O)]2-, as below pH
7.5 the transformation of the [Gd(EPTPA-bz-NH2)(H2O)]2-
diastereomers is too fast to be followed by HPLC. The kobs
rate constants measured are presented in Figure 4. The k1
second-order rate constant calculated is (3.03 ( 0.07)×104
s-1 M-1, around 2500 times higher than that for [Gd(DTPA-
bz-NH2)(H2O)]2-. The faster isomerization of [Gd(EPTPA-
bz-NH2)(H2O)]2- can be most likely explained in terms of a
faster dissociation of the complex to form free ligand and
metal as compared to [Gd(DTPA-bz-NH2)(H2O)]2-. Al-
though a detailed kinetic study of the dissociation and metal
exchange reaction of [Gd(EPTPA)(H2O)]2- is not yet avail-
able, it is very plausible that, due to its elongated amine
backbone, this complex is kinetically less inert than
We have studied the isomerization kinetics of [Gd(DTPA-
bz-NH2)(H2O)]2- diastereomers at variable pH. A concen-
trated stock solution of the complex was separated on an
analytical column to obtain samples containing >99% pure
A and B isomers (HPLC method 2; see Experimental
Section). After the appropriate pH was set, the transformation
of the isomers occurring in a thermostated solution (25 °C)
was monitored in time by using HPLC method 2. The
isomerization of both A and B correspond to a first-order
kinetic reaction.
kBA
A { } B
(1)
kAB
The kinetic data obtained, i.e., the A and B molar fractions
at times t, were fitted to a single-exponential curve as follows.
from A to B:
A ) (A0 - Af)e-k t + Af
(2)
AB
from B to A:
B ) (B0 - Bf)e-k t + Bf
(3)
BA
Here A0 and B0 are the initial, whereas Af and Bf are the
final (equilibrium), molar fractions of A and B isomers in
the solution. The observed rate constants, kAB and kBA
(hereafter kobs), are identical within the experimental error
at any pH. The experimental kobs values are presented in
Figure 3 as a function of pH. According to proton catalysis,
k
obs increases with decreasing pH and is linearly proportional
to the H+ concentration:
kobs ) k0 + k1[H+]
(4)
When the experimental kobs values were fitted, k1 ) 11.7
( 0.5 and 11.4 ( 0.7 s-1 M-1 second-order rate constants
were calculated for the transformation from A to B and from
B to A, respectively. The k1 value for [Gd(DTPA-bz-NH2)-
(H2O)]2- is three times higher than that of obtained for [Gd-
(EOB-DTPA)(H2O)]2-, 3.92 s-1 M-1.7 As the transformation
(15) (a) Bru¨cher, E.; Laurenczy, G. J. Nucl. Chem. 1981, 41, 2089. (b)
Sarka, L.; Burai, L.; Bru¨cher, E. Chem.sEur. J. 2000, 6, 719. (c)
Burai, L.; Bru¨cher, E.; Kira´ly, R.; Solymosi, P.; V´ıg, T. Acta Pharm.
Hung. 2000, 70, 89.
Inorganic Chemistry, Vol. 44, No. 10, 2005 3565

Burai et al.
[Gd(DTPA)(H2O)]2-. In conclusion, [Ln(DTPA-bz-NH2)-
(H2O)]2- diastereomers transform slowly enough to maintain
sufficient isomer purity over a long period after separation.
Hence, [Ln(DTPA-bz-NH2)(H2O)]2- and other DTPA de-
rivatives are appropriate for diverse investigations of the
individual diastereomers.
Water Exchange Rate. For Gd-chelate complexes used
as MRI contrast agents it is a basic requirement to have at
least one water molecule in the inner coordination sphere.
The inner sphere water molecule is in exchange with the
bulk so that the paramagnetic relaxation enhancement effect
of Gd3+ is transferred to the environment. The rate of water
exchange from the inner sphere to the bulk can be directly
determined by variable-temperature 17O NMR relaxation
measurements.
Despite the relatively fast isomerization of [Gd(EPTPA-
bz-NH2)(H2O)]2-, the individual water exchange rate of the
two isomers could be measured by taking advantage of their
approximately equal ratio on complex formation and their
subsequent transformation to the equilibrium state (93:7
A:B), as reported previously.8 Concomitant to the transfor-
mation of B to A, the 1/T2 relaxation rates increase with
time for both [Gd(EPTPA-bz-NH2)(H2O)]2- and [Gd(EPTPA-
bz-NO2)(H2O)]2- complexes showing that the two diaster-
eomers have different water exchange rates. The evolution
of the 17O 1/T2 relaxation rates was monitored after mixing
buffered Gd3+ and EPTPA-bz-NH2 or EPTPA-bz-NO2 solu-
tions, and parallel HPLC injections from the same samples
were done to measure the diastereomer ratio in the given
time period. From the variation of the 1/T2 values over time
and by using the A:B ratios determined by HPLC for each
time point, the water exchange rates were calculated. We
assumed that the electronic relaxation parameters and the
hyperfine coupling constant were identical for the two
isomers, and we used the values previously determined for
[Gd(EPTPA-bz-NO2)(H2O)]2- (equilibrium mixture of A and
B).9 Although not spectacular, the difference in the water
exchange rate was well distinct between the two [Gd(EPTPA-
bz-NH2)(H2O)]2- and [Gd(EPTPA-bz-NO2)(H2O)]2- isomers.
In both cases, the kex values roughly doubled from the A
(93%) to the B (7%) isomer (kexi298 is (1.4 ( 0.1) × 108 and
(2.6 ( 0.2) × 108 s-1 for A and B of [Gd(EPTPA-bz-NO2)-
(H2O)]2- and (1.8 ( 0.1) × 108 and (4.4 ( 0.2) × 108 s-1
for A and B of [Gd(EPTPA-bz-NH2)(H2O)]2-). This implies
that, e.g. for [Gd(EPTPA-bz-NH2)(H2O)]2-, the B isomer,
present in 7%, is responsible for 16% of the overall water
exchange.
Figure 5. Reduced transverse (top) and longitudinal (middle) 17O
relaxation rates and reduced 17O chemical shifts (bottom) as a function of
the inverse temperature measured in separated samples of A (red circles)
and B (blue triangles) diastereomers of [Gd(DTPA-bz-NH2)(H2O)]2- and
in an equilibrium mixture of A and B (green squares). The lines correspond
to the fitted curves using the Solomon-Bloembergen-Morgan theory of
paramagnetic relaxation. The reduced transverse (1/T2r) and longitudinal
relaxation rates (1/T1r) and chemical shifts (∆ωr) are defined as follows,
where T1,2 and ω are the longitudinal and transverse relaxation rates and
chemical shifts measured in the GdIII complex solution, whereas T1,2A and
ωA are those measured in a diamagnetic reference solution and Pm is the
molar fraction of coordinated water: 1/T1,2r ) (1/Pm)[1/T1,2 -1/T1,2A]; ∆ωr
) (1/Pm)(ω - ωA).
relaxation rates and chemical shift of the samples were
measured in the temperature range 0-70 °C referred to a
buffered aqueous sample at identical pH. At the end of the
experiments, the isomer purity of solutions A and B was
checked and found >99%. The results of the 17O NMR
experiments are presented in Figure 5.
In contrast to the EPTPA-derivative complexes, for [Gd-
(DTPA-bz-NH2)(H2O)]2- the water exchange rate could be
measured on the separated, pure diastereomers. After HPLC
separation, their isomerization rate is slow enough to
maintain the isomer purity for the time of the entire 17O NMR
experiment even at higher temperatures; thus, the classical
technique was used for water exchange rate determination.
We had three samples containing (1) the pure major isomer
(A), (2) the pure minor isomer (B), and (3) a nonseparated
[Gd(DTPA-bz-NH2)(H2O)]2- solution in isomer equilibrium
state (AB; 60% A and 40% B). The oxygen-17 1/T1, 1/T2
The reduced transverse 17O relaxation rates, 1/T2r, are in
all three cases in the fast exchange region at higher and in
the slow exchange region at lower temperatures with a
changeover around 300 K for B and 320 K for A, the AB
sample being in between. The reduced chemical shifts at
higher temperatures are very similar for the three samples
showing that both diastereomers have the same number of
inner sphere water molecule (q). For analogous [Gd(DTPA)-
(H2O)]2- derivatives one inner sphere water molecule was
reported; thus, we assume q ) 1 for [Gd(DTPA-bz-NH2)-
3566 Inorganic Chemistry, Vol. 44, No. 10, 2005

[Gd(DTPA-bz-NH2)(H2O)]2- Diastereomers
Table 1. Parameters Characterizing Water Exchange, Rotation, and
Electron Spin Relaxation As Obtained from 17O NMR Data for A and B
Diastereomers of [Gd(DTPA-bz-NH2)(H2O)]2- and for the Equilibrium
Solution AB (60% A and 40% B)a,b
eomers display a different orientation of the two square
planes formed by the four cyclen nitrogens and the four
binding oxygens, making an angle of ca. 40° in one of the
structures (M-type) whereas it is reversed and reduced to
ca. 20° in the m-type structures. In solution, the two isomers,
m and M, may exist in equilibrium.18 In solid state for [Ln-
(DOTA)(H2O)]3+ complexes, m-type structures have been
observed for La and Ce and M-type structures for Eu, Gd,
Dy, Ho, Lu, and Y.19 These diastereomers cannot be
separated by HPLC techniques as they interchange very
quickly (for the m f M process a rate of k ) 511 s-1 was
reported at 306 K).18 However, they are well distinguishable
[Gd(DTPA)-
A
B
AB
(H2O)]2- c
kex298/106 s-1
∆Hq/kJ mol-1
5.7 ( 0.2
3.1 ( 0.1
4.7 ( 0.1
4.1
52.0
+56.2
-3.8
103
18
0.25
1.6
54.4 ( 1.1 53.5 ( 1.1 51.9 ( 0.7
∆Sq/J mol-1 K-1 +66.8 ( 3.4 +58.6 ( 3.7 +57.0 ( 3.2
(A/p)/106 rad s-1 -3.7 ( 0.1 -3.6 ( 0.1 -3.5 ( 0.1
τR298/ps
246 ( 5
223 ( 5
256 ( 4
ER/kJ mol-1
τv298/ps
22.3 ( 0.7 20.4 ( 0.7 19.7 ( 0.4
19 ( 2
1
2.5 ( 0.2
0.1
20 ( 2
1
2.4 ( 0.2
0.1
21 ( 1
1
3.0 ( 0.2
0.1
Ev/kJ mol-1
∆2/1020 s-2
Cos
0.15
0.2
1
by H or 13C NMR spectroscopy. The water exchange rate
has been directly assessed on the two diastereomers of the
tetraamide derivative [Eu(DOTAM)(H2O)]3+ in acetonitrile-
water solvent.20 The water exchange on m is about 50 times
faster than on M, and even though the equilibrium constant
K ) [M]/[m] equals 4.5, the contribution of m to the overall
exchange rate represents 90%.
The difference observed in the water exchange rate of the
diastereomers of the linear DTPA- or EPTPA-GdIII com-
plexes is of no weight compared to this enormous variation
for the macrocyclic complexes. It is clear that in the two
cases the origin of the presence of diastereomers is funda-
mentally different. Clearly, the structural difference between
the m and M isomers for DOTA-type complexes acts much
more upon the inner coordination sphere and consequently
modifies more strongly the water exchange rate than for the
linear chelates. In this latter case, the differing configuration
of the carbon in position-4 or that of the central nitrogen in
the ligand will not essentially influence the inner coordination
sphere of the complex, which could in turn lead to signifi-
cantly different water exchange rates.
When the 17O NMR data are fitted to the Solomon-
Bloembergen-Morgan theory, the parameters describing
electron spin relaxation are also calculated. The recent
development in the field of electronic relaxation of GdIII
complexes showed that this model is not really adequate;
however, due to its relative simplicity, it remains widely
used.21 Therefore, we do not wish to interpret in details the
parameters obtained in our fit. Nevertheless, they show that
electronic relaxation does not differ basically for the two
diastereomers A and B and the mixture AB. This is also
supported by EPR measurements performed at different fields
on separated samples of A, B, and the equilibrium solution
a Italicized values were fixed in the fit. b The equations used in the
analysis of the 17O NMR data are listed in the Supporting Information.
The following parameters are calculated: kex298, water exchange rate at 298
K; ∆Hq, activation enthalpy of water exchange; ∆Sq, activation entropy
of water exchange; A/p, hyperfine coupling constant, between the water
oxygen and Gd3+; τR298, rotational correlation time of the complex at 298
K; ER, activation energy of rotation; τv298, electronic zero-field-splitting
modulation time at 298 K; Ev, activation energy of electronic zero-field-
splitting; ∆2, trace of the square of the transient zero-field-splitting tensor;
Cos, outer-sphere contribution to the chemical shift. c Reference 16.
(H2O)]2-.This is confirmed by the values of the hyperfine
coupling constant, A/p, obtained from the analysis of the
chemical shift data (see below). The experimental data have
been fit to the Solomon-Bloembergen-Morgan theory
which was successfully applied for many Gd-poly(amino
carboxylate) complexes.16 The parameters describing water
exchange, rotation, and electronic relaxation, as obtained
from the fit, are compared in Table 1 for A, B, and AB
samples and for [Gd(DTPA)(H2O)]2-.
The water exchange is faster, though not dramatically, for
the A than for the B isomer of [Gd(DTPA-bz-NH2)(H2O)]2-
(kex
is 5.7 × 106 and 3.1 × 106 s-1 for A and B,
298
respectively). In the equilibrium solution (AB) with 60% A
and 40% B, the water exchange rate measured, kex298 ) 4.7
× 106 s-1, agrees well with that calculated from values for
A and B, 4.66 × 106 s-1. Interestingly, the ratio of kex298 for
A and B is close to 2, similar to that found for the
diastereomers of Gd(EPTPA)2- derivatives. However, on the
basis of only two examples, it is difficult to state any general
rule in this respect. In comparison to other [Gd(DTPA)-
(H2O)]2- derivatives, [Gd(DTPA-bz-NH2)(H2O)]2- has a
298
similar water exchange rate (kex ) 4.1 × 106 and 3.6 ×
106 s-1 for [Gd(DTPA)(H2O)]2- and [Gd(EOB-DTPA)-
(H2O)]2-, respectively).16,17 This let us conclude that the
aminobenzyl pending arm on the DTPA has only minor
influence on the water exchange.
(18) Aime, S.; Botta, M.; Fasano, M.; Marques, M. P. M.; Geraldes, C. F.
G. C.; Pubanz, D.; Merbach, A. E. Inorg. Chem. 1997, 36, 259.
(19) (a) Aime, S.; Barge, A.; Botta, M.; Fasano, M.; Ayala, J. D.; Bombieri,
G. Inorg. Chim. Acta 1996, 246, 423. (b) Parker, D.; Pulukkody, K.;
Smith, F. C.; Batsanov, A.; Howard, J. A. K. J. Chem. Soc., Dalton
Trans. 1994, 689. (c) Benetollo, F.; Bombieri, G.; Aime, S.; Botta,
M. Acta Crystallogr., Sect. C: Cryst. Struct. Commun. 1999, 55, 353.
(d) Dubost, J. P.; Leger, J. M.; Langlois, M. H.; Meyer, D.; Schaefer,
M. C. R. Acad. Sci. Paris 1991, 312, 349.
(20) (a) Aime, S.; Barge, A.; Bruce, J. I.; Botta, M.; Howard, J. A. K.;
Moloney, J. M.; Parker, D.; De Sousa, A. S.; Woods, M. J. Am. Chem.
Soc. 1999, 121, 5762. (b) Aime, S.; Barge, A.; Botta, M.; De Sousa,
A. S.; Parker, D. Angew. Chem., Int. Ed. 1998, 37, 2673. (c) Dunand,
A. F.; Aime, S.; Merbach, A. E. J. Am. Chem. Soc. 2000, 122, 1506.
(21) Borel, A.; Helm, L.; Merbach, A. E. In Very High Frequency (VHF)
ESR/EPR; Vol. 22 of Biological Magnetic Resonance; Plenum Press:
New York, 2004; pp 219-260.
For the macrocyclic LnDOTA- complexes, often paral-
leled with the acyclic DTPA family, the existence of
diastereomers is a well-described phenomenon. It is the
consequence of the different orientation of the acetate groups
and the different dihedral angles of the ethylene bridges in
the tetraazacyclododecane ring. Basically, the two diaster-
(16) Micskei, K.; Helm, L.; Bru¨cher, E.; Merbach, A. E. Inorg. Chem. 1993,
32, 3844.
(17) To´th, EÄ .; Burai, L.; Bru¨cher, E.; Merbach, A. E. J. Chem. Soc., Dalton
Trans. 1997, 1587.
Inorganic Chemistry, Vol. 44, No. 10, 2005 3567

Burai et al.
lanthanide series, in contrast to the previously described LnIII
EPTPA derivatives, where it varied as a function of the cation
size with a maximum for the middle lanthanides. The
interconversion between the two diastereomers of [Gd-
(DTPA-bz-NH2)(H2O)]2- is a proton-catalyzed process,
which is slow enough to allow for investigating the individual
isomers in pure form after HPLC separation. The water
exchange rate has been determined for both isomers and,
similarly to LnIII EPTPA derivatives, was found to differ by
a factor of 2. This variance in the exchange rate has no
consequence on the water proton relaxivity of the two
diastereomers, since it is solely limited by fast rotation.
However, such difference in kex will considerably affect
proton relaxivity when these diastereomers are linked to a
slowly rotating macromolecule. Once the rotation is opti-
mized, it is the slow water exchange which will exclusively
or partly limit relaxivity; in this case a factor of 2 in the
water exchange rate can lead to a remarkably different
relaxivity for the two macromolecular diastereomer com-
plexes. Consequently, these results have important implica-
tions for future development of GdIII-based, macromolecular
MRI contrast agents, since, due to the use of bifunctional
ligands, their synthesis inevitably involves the presence of
diastereomers.
Figure 6. Water proton relaxivities measured in separated samples of A
(red circles) and B (blue triangles) diastereomers of [Gd(DTPA-bz-NH2)-
(H2O)]2-, as well as in an equilibrium mixture of A and B (green squares).
AB. For the three samples, the peak-to-peak line widths were
equal within experimental error (Supporting Information).
Similarly, after mixing EPTPA-NO25- or EPTPA-NH25- with
Gd3+ we detected no change in the X-band EPR peak-to peak
line widths as a function of time. This means that the
isomerization process, which occurs from the formation of
the complex with 1:1 A:B ratio until equilibrium, does not
affect the electronic relaxation; thus, it must be the identical
for A and B.
Water proton relaxivities have been also measured for the
two [Gd(DTPA-bz-NH2)(H2O)]2- diastereomers. As it is
presented in Figure 6, the relaxivity profiles for A and B
isomers as well as for the equilibrium state (AB) of [Gd-
(DTPA-bz-NH2)(H2O)]2- are the same within the experi-
mental error. For the low molecular weight complexes the
relaxivity is limited by rotation rather than water exchange.
As a consequence, the different water exchange rate on the
two diastereomers has practically no influence on the
relaxivity which is determined by the fast rotation. Due to
its larger size, [Gd(DTPA-bz-NH2)(H2O)]2- has a longer
Acknowledgment. This paper is dedicated to Professor
Erno˜ Bru¨cher on the occasion of his 70th birthday. The
authors are grateful to Meriem Benmelouka for the EPR
measurements. This research was financially supported by
the Swiss National Science Foundation and the Office for
Education and Science (OFES). This work was carried out
in the frame of the EC COST Action D18.
Supporting Information Available: Tables of the diastereo-
meric ratios for [Ln(DTPA-bz-NH2)(H2O)]2- and [Ln(EPTPA-bz-
NH2)(H2O)]2-, the experimental and calculated rate constants for
the isomerization reaction, results of the variable-temperature 17O
NMR measurements, transverse electron spin relaxation rates, and
the relaxivity values, as well as equations of the Solomon-
Bloembergen-Morgan theory used for the analysis of 17O NMR
data. This material is available free of charge via the Internet at
http://pubs.acs.org.
298
rotational correlation time than [Gd(DTPA)(H2O)]2- (τR
) 256 ps vs 103 ps, respectively) which is reflected in its
higher relaxivity; r1 ) 7.6 mM-1 s-1 in comparison to 5.6
mM-1 s-1 for [Gd(DTPA)(H2O)]2- at 10 MHz and 25 °C.
Conclusion
An HPLC analysis showed that the diastereomer ratio for
LnIII DTPA derivatives remains constant throughout the
IC048645D
3568 Inorganic Chemistry, Vol. 44, No. 10, 2005