See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/279935672

17O NMR study of water exchange on [Gd(DTPA)(H2O)]2- and [Gd(DOTA)

(H2O)]- related to NMR imaging

Article  in  Inorganic Chemistry · January 1993

CITATIONS READS

55 315

4 authors, including:

Lothar Helm Erno Brücher

École Polytechnique Fédérale de Lausanne University of Debrecen
275 PUBLICATIONS   12,138 CITATIONS    131 PUBLICATIONS   3,359 CITATIONS   

SEE PROFILE SEE PROFILE

André E Merbach
École Polytechnique Fédérale de Lausanne
525 PUBLICATIONS   19,590 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Contrast agents for medical imaging View project

Reaction mechanisms in coordination chemistry View project

All content following this page was uploaded by Lothar Helm on 13 July 2015.

The user has requested enhancement of the downloaded file.

3844 Inorg. Chem. 1993, 32, 38443850

NMR Study of Water Exchange on [Gd(DTPA)(H2O)lZ and [Gd(DOTA)(HzO)]- Related to

170
NMR Imaging's

KBroly Micskei,t?*Lothar Helm,? Em0 Briicher,'** and Andre E. Merbach'J

Institut de Chimie Minerale et Analytique, Place du Chlteau 3, CH-1005 Lausanne, Switzerland, and
Institute of Inorganic and Analytical Chemistry, Kossuth University, H-4010 Debrecen, Hungary
Received April 6, 1993

The water-exchangerate constants for the [Gd(DTPA)(H20)lz- and [Gd(DOTA)(HzO)]- complexes as determined
by variable-temperature I7O NMR are respectively kexZ9*= (4.1 f 0.3) X lo6 and (4.8 f 0.4) X 106 s-1. The
activation volumes (AV), measured up to 200 MPa, are 12.5 f 0.2 and 10.5 f 0.2 cm3mol-', indicating an extreme
dissociative activation mode for the water-exchange mechanism. The mechanism (D) is further supported by the
large activation enthalpies (AH*= +52.0 f 1.4 and +48.8 f 1.6 kJ mol-') and positive entropies (AS*= +56.2
f 5 and +46.6 f 6 kJ mol-' K-') obtained for the [Gd(DTPA)(H20)I2- and [Gd(DOTA)(HtO)]- complexes,
respectively. In the first coordination sphere of these two monoaqua complexes there is only space for one water
molecule, and thus the bond breaking of the coordinated water should be the rate-determining step. The I7O
relaxation contribution of the second coordination sphere was estimated by investigating [Gd(TETA)]-, which has
no water in the first coordination sphere. All these considerationslead to the conclusion that the effectiveness of
[G~(DTPA)(HzO)]~- and [Gd(DOTA)(HzO)]- as contrast agents in MRI is not limited by the relatively low
water-exchange rates but by TI,, the longitudinal relaxation time of water protons in the first coordination sphere.

Introduction
In recent years magnetic resonance imaging (MRI) became
an important clinical modality in diagnosticmedi~ine.~ The MRI
signal mainly comes from the protons of the water molecules in
the body fluids and tissues. The signal intensities, and therefore
the image contrast, depend on the longitudinal (l/TI) and
transverse (l/TZ) relaxation rates of the water protons. It is
common to increase the contrast of the images by increasing the
relaxationrates by addition of paramagnetic substances,so-called
MRI contrast agents. An acceleration of relaxation can be
achieved by adding paramagnetic ions like Gd3+,Fe3+,and MnZ+,
0
* 0
a
-0

which are in an electron spin S ground state at the oxidation

states indicated. Unfortunately these ions are very toxic in a free
state when they are injected into the body in concentrations DOTA4.
necessary for their use as relaxation agents. Therefore only very
stable and inert complexes of these ions are employed. Besides
the strong complexation, the ligands used have to allow water
molecules to enter into the first coordination sphere of the metal
ion and by rapid exchange with water in the outer coordination
shell ensure the desired effect of relaxation enhancement.
The most widely used contrast agents nowadays are complexes 0-
of gadolinium(II1) such as [Gd(DTPA)(Hz0)l2- and -O\ /

[Gd(DOTA)(H20)]- (for the structures of the ligandssee Figure

l).' These complexes are extremelyinert and contain one water TETA4.
molecule in the first coordination shell, according to X-ray
diffraction data on [Gd(DTPA)(HzO)]:z- [Nd(DTPA)(HzO)]2-,8
and [Eu(DOTA)(H~O)]-~ and 'H and I3C NMR as well as \
-0
t Institut de Chimie Minerale et Analytique.
8 Kossuth University.
YNwNQ-
0 0
( 1 ) High-pressure NMR Kinetics. 58. For part 57, see ref 2.
(2) Abou-Hamdan, A.; Burki,N.; Lincoln, S.F.;Merbach, A. E.; Vincent, Figure 1. Schematic drawing of the ligands used in this study (DTPAS-
S. Inorg. Chim. Acta 1993, 207, 27. = diethylenetriamine-N,N',N'',N'"-pentaacetate; DOTAC = 1,4,7,10-
(3) Biinzli, J.-C. G.; Choppin, G . R. Lanthanide Probes in Life, Chemical tetraazacyclododecane-N,N',N'',N"'-tetraactate; TETA' = 1,4,8,11-
and Earth Sciences; Elsevier: Amsterdam, 1989; Chapter 5. tetraazacyclotetradeane-N,N',N'',N"'-tetraacetate).
(4) Lauffer, R. B. Chem. Rev. 1987, 87, 901.
(5) Goldstein, H.; Lumma, W.; Rudzik, A. Annu. Rep. Med. Chem. 1989,
24, 265.
(6) Desreux, J. F.; Barthelemy, P. P. Nucl. Med. Biol. 1988, 15, 9.
luminescence studies on [ Ln(DTPA) ( H 2 0 ) ] *- and
(7) Gries, H.; Miklautz, H. Physiol. Chem. Phys. Med. N M R 1984, 16, [L~(DOTA)(HZO)]-.'&'~ These studies also show that all donor
105. atoms of DTPA and DOTA are actually coordinated to the ion.
(8) Stezowski, J. J.; Hoard, J. L. Isr. J. Chem. 1984, 24, 323.
(9) Spirlet, M. R.; Rebizant, J.; Desreux J. F.;Loncin, F. Inorg. Chem. Similar studies using the TETA ligandl2J3(see Figure 1 for
1984, 23, 359. structure) show that in this case no water molecules enter into
0020-1669/93/1332-3844$04.00/0 0 1993 American Chemical Society
170NMR Study of Water Exchange Inorganic Chemistry, Vol. 32, No. 18, 1993 3845
the first coordination shell, making this complex much less Table I. Compositions of Samples Used for Variable-Temperature
attractive as a contrast agent. (Nos. 1-4) and Variable-Pressure (Nos. 5 and 6) 170NMR Studies
The relaxation enhancement of water protons caused by these molality/ ligand
substances is measured by 'H NMR and is normally interpreted no. aqueous soha molkg-' excess/% 1O'Z" PH
using the Solomon-Bloembergen-Morgan (SBM) theory,4as well 1 HClO4-acidified water 2.0
as the equations given by Swift and Connick.14 One important 2 [G~(DTPA)(HZO)]~- 0.0500 2.5 9.00 5.3
factor in this theory is the rateconstant ofwater exchange between 3 [Gd(DOTA)(H20)]- 0.0504 2.3 9.07 5.1
the first coordination sphere of the metal ion and the outer 4 [Gd(TETA)]- 0.0484 6.1 7.7
coordination shell, which transfers the fast relaxation of a nuclear 5 [Gd(DTPA)(H20)]'- 0.0495 3.5 8.92 5.1
spin when coupled to an electron spin into the bulk of the water 6 [Gd(DOTA)(H20)]- 0.0495 3.9 8.92 5.0
solution. To obtain a maximum effect, the residence time, T,,,, 5.5% I7O enriched water.
of a water molecule in the bound state, which is the reciprocal
of the exchange rate constant, k,,, has to be of the right order to the solution and left it for a day for complete complexation. Then the
of magnitude, lV-10-9s, depending on different factors such pH of all solutions was adjusted using 1 M NaOH and/or HC104. For
as the electron spin relaxation rate and the molecular tumbling NMR measurements we evaporated the water of the samples under
time of the ~ o m p l e x . ~ vacuum and added an identical weighed amount of 5.5% I7O enriched
Water-exchange rates for triply charged cations in general water (Yeda). The pHs of these solutions were then checked again. The
extend over more than 18 orders of magnitude.15 The lanthanide sample compositions are given in Table I.
aqua ions are all situated a t the very fast exchange side of the UV-Visible Measurements. Variable-temperature UV-visible mca-
surements were performed (Perkin-Elmer Lambda 5 spectrometer) to
whole exchange rate spectrum, with rate constants in the range check the possibility of the presence of coordination equilibria of the
107-109s-I.l6 Until now, very little information is available on complexesstudied. For this purpose weused the europium(II1)complexes,
water exchange rates of cations with multidentate ligands used which have very sensitive electronic absorption in the ranges 390 nm <
as contrast agents. Southwood-Joneset al. published17exchange X < 400 nm and 578 nm < X < 581 nm. The temperature dependenct
rates for [Gd(PDTA)(H20)2]- which are about 3 times slower of these transitions was measured id the 5 'C < T < 85 ' C range. All
than those for the neat aqua complex. This is in contrast to experimental details and results are collected in the supplementary
observations made on first-row transition metals, where coor- material.
dination of anionic ligands labilizes the metal-water bond1*and NMR Measurements. Varidble-temperature1 7 0NMR measurements
leads therefore to faster exchange rates. were performed at two magnetic fields using Bruker AM-400 (9.4 T,
54.2 MHz) and AC-200 (4.7 T, 27.1 MHz) spectrometers. Bruker VT-
The water-exchange rate, which is the exchange rate of the
1000 temperature control units were used to stabilize the temperature
oxygen atom, represents a lower limit for the exchange rate of in the range 273-373 K. The temperature was measured before and
the water protons. In these fast exchanging systems it may be after each measurement by substituting the sample with an NMR tube
expected to be very close to the proton-exchange rate, and so it containing a platinum resistance therm~meter.'~ The inversion recovery
is important for a better understanding of the mode of operation method20was applied to measure longitudinalrelaxation rates, 1/ TI,and
of the relaxation agents based on these gadolinium complexes. the Carr-Purcell-Meiboom-Gillspin-echotechniqu6Iwas used to obtain
Furthermore, it has been shown that variable-temperature and transverse relaxation rates, 1/T2.
variable-pressure 170NMR can provide information on the All solutions were introduced into spherical glass containers, fitting
exchange reaction mechanism. We found it worthwhile, therefore, into ordinary 10-mm NMR tubes, in order to eliminate magnetic
to perform a complete 1 7 0 NMR study on DTPAS- and DOTAe susceptibilitycorrections to chemical shifts.22 This technique also leads
to a small gas volume above the solutions and reduces therefore problems
complexes of Gd3+. The [Gd(TETA)]-complex, having no water due to distillation of water at the highest temperatures.
in the first coordination sphere, was included because it provides The eliminationof magnetic susceptibilityeffects by the use of spherical
us for the first time with information on the possible contribution cells was checked in the following way: The chemical shifts of a 1.00 X
of second-coordination-sphere effects on 170relaxation on lo-* m sampleof [Gd(H20)#+ sitting in the spherical cell was measured
lanthanide complexes. at 25.0 OC relative to an acidified water external standard using first a
superconducting NMR magnet (4.7 T) where the static magnetic field,
Experimental Section BO,is parallel to the cylindrical NMR tubes and then an electromagnet
(1.4 T, 8.13 MHz, 170) where Bo is perpendicular to the sample tube.
Chemicals and S o l ~ t i Gd(C104)p
~~. and Eu(CIO4)3 stock solutions
If there were any distortion from spherical symmetry of the sample, for
were prepared by dissolving excess Gd203 and EuzO3 (NUCOR Corp.,
example due to solution entering the cylindrical part of the container,
Phoenix, AZ; 99.99%) in perchloric acid (Merck, p.a., 70%). The pHs
there would be a differencein the observed chemical shifts. The equations
of the solutions were regulated after filtration to 4. The lanthanide
describing the corrections to observed chemical shifts for an infinite
concentration was determined by chelatometric titration with NazH2-
cylindrical shaped sample are given (eq 1 for superconductingmagnets
EDTA solution using xylenol orange as indicator and urotropin for pH
and eq 2 for electromagnets).
regulation.
HsDTPA was purchased from Fluka (puriss p.a., >99%) and H4-
DOTA was used as obtained from Schering (Berlin). H4TETA was aldh = 6 + ( 4 ~ / 3 ) ( ~ r e rX- - ~ J (1)
provided by Prof. J. F. Desreux (Lihge). The purity of the ligands was
checked using pH-metry, chelatometry, and IH NMR. All solutions
were prepared by weight. To the adequate amount of solid ligand we
added distilled water and decarbonated 1 M NaOH for stoichiometric
deprotonation. We added a weighed amount of Gd(C104)3 or Eu(C104)3 The chemical shifts measured in the different magnets were self-
consistent so the correction due to deviation from sphericalgeometry was
(10) Jenkins, B. G.; Lauffer, R. B . Inorg. Chem. 1988, 27, 4730. negligible.
(11) Peters, J. A. Inorg. Chem. 1988, 27, 4686. Variable-pressureNMR measurements were made on a Bruker AM-
(12) Bryden, C. C.; Reilley, C. N. Anal. Chem. 1982, 54, 610. 400 spectrometer equipped with a previously described home built
(13) Spirlet, M. R.; Rebizant, J.; Loncin, M. F.; Desreux, J. F. Inorg. Chem. pr~behead.~'The temperature was measured with an accuracy of A0.5
1984, 23,4278.
(14) Swift, T. J.; Connick, R. E. J. Chem. Phys. 1962, 37, 307. K after all corrections using a built-in Pt resistor. The r / 2 pulse length
(15) Merbach, A. E. Pure Appl. Chem. 1987, 59, 161.
(16) Cossy, C.; Helm, L.; Merbach, A. E. Inorg. Chem. 1988, 27, 1973. (19) Amman, C.; Meyer, P.; Merbach, A. E. J. Magn. Reson. 1982 46,319.
(17) Southwood-Jones,R. V.; Earl, W. L.; Newman K. E.; Merbach, A. E. (20) Vold, R. V.; Waugh, J. S.; Klein, M. P.; Phelps, D. E.J. Chem. Phys.
J . Chem. Phys. 1980. 73, 5909. 1968.48, 3831.
(18) Margerum, D. W.; Cayley, G.R.; Weatherburn, D. C.; Pagenkopf, G. (21) Meiboom, S.; Gill, D. Reu. Sci. Insfrum. 1958, 29, 688.
K. In Coordinotion Chemistry; Martell, A. E., Ed.; American Chemical (22) Hugi, A. D.; Helm, L.; Merbach, A. E. Helu. Chim. Acru 1985,68,508.
Society: Washington, DC, 1978; Vol. 2. (23) Frey, U.; Helm, L.; Merbach, A. E. High Pressure Res. 1990, 2, 237.
3846 Inorganic Chemistry,Vol. 32, No. 18, 1993 Micskei et al.
was approximately 15 H, depending slightly on sample concentration
and temperature.

ReSUltS
UV-Visible Spectroscopy. A single absorption band for the
solutions of [Eu( DTPA) (HzO)]2- and [Eu(DOTA)(HzO)]-
complexes was observed in the 392 nm C X C 397 nm range, while
in the case of [Eu(TETA)]- the band detected in the same range
was split into two peaks at 393.8 and 394.6 nm. The peak
intensities ( E ) and positions of the maxima (A) decrease on 0 A- ,M , +, ,%. ,” ,
increasing the temperature from 278 to 358 K. The relative 2.70 2.90 3.10 3.30 3.50 3.70
intensities of the two peaks of [Eu(TETA)]- do not change, but 1OOOTT IK”
this could be the result of the equilibriumof two different species,
with negligible enthalpy of reaction. G r a e ~ p measured
i~~ the Figure 2. Temperature dependence of I7ONMR transverse relaxation
rate increase due to the presence of [Gd(DTPA)(HzO)]% (0, 0).
7Fp5Dotransition of [Eu(TETA)]-in the 578 nm C X C 581 nm [Gd(DOTA)(HzO)]- (m, O ) , and [Gd(TETA)]- (A, A). Filled symbols
range at different temperatures and observed a single absorption correspond to 9.4-T and open symbols to 4.7-T results, respectively.
band. Thisverysensitive transitionexcludesreliablytheformation
of different species in our studied temperature range. Taking further coordination shells. Aw, and l/Tl- give the sum of all
into account theresultspublishedfor [Eu(EDTA)(H20),]-, where outer-sphere contributions to Awr and l/TIr. Pm is the partial
solvation equilibria between two species was pr0ved,2~only single molar fraction of solvent in the first coordination sphere of the
species could be detected in the solutions of the complexes we cation, and 7, is the mean lifetime of a water molecule in it.
studied.26
The differences of the transverse relaxation rates measured
Variable-Tempernhue NMR. We measured the temperature
with and without paramagnetic complexes are shown in Figure
dependence of the 1 7 0 NMR longitudinal and transverse
2. The DTPAS-and the DOTA’complexes, which both contain
relaxation rates as well as the chemical shifts of aqueoussolutions
one water molecule in the first coordinationsphere, show typical
of [Gd(DTPA) (HzO)]z-, [Gd(DOTA) (HzO)]-, and [Gd-
Swift-Connick temperature behavior, the negative slope at low
(TETA)]-. The observed acceleration of the relaxation rates
temperature corresponding to the kinetic region. However for
and the observed shift of the resonance frequencies can be
described by the formulas developed by Swift and Connick14for the TETA& complex we observe only a small increase in the
1/ TZ(eq 3) and Aw (eq 4), the chemical shift difference, and by transverse relaxation rate. The phenomenon can be explained in
Zimmermann and Brittinz9for 1/T1 (eq 5 ) . As an extension to three different ways.
earlier formulations, we show outer-sphere contributions to First, a small part of the TETA” complex may be present as
relaxation and chemical shift explicitly. a monoaqua species, leading to a very small value of Pm. We
reject this possibility on the basis of our UV-visible spectroscopy
measurements. Second, a small amount of octaaqua ion may be
present. This supposition was checked by measuring samples at
different ligand excess (650%) and pH (7.7-8.5). These data
(see supplementary material) are close to the values shown in
Figure 2, excluding this phenomenon. Finally, it is possible that
no water molecule is found in the first coordinationsphere of the
gadolinium(II1)ion in this complex, which means P m is zero, and
the entire increase in relaxation can be attributed to a second-
sphere effect.
To be able to describethe temperature dependence of the 1/ Tzr
data, we have to know the temperature variationsof the different
NMR and kinetic parameters. The mean lifetime Tm is related
to the pseudo-first-order rate constant for the exchange of a
particular solvent molecule, k,,, whose temperature dependence
Symbolswithout index (1/T2,1 /TI, and w ) apply to quantities is given by the Eyring equation (6),where AS*and i W * are the
measured in systems containingthe paramagneticcations. 1/ TZA activation entropy and the activation enthalpy, respectively.
and l / T I are
~ the relaxation rates in acidified water, and l/Tzm
and 1ITlmare those from water directly coordinated to the Gd3+ 1 kBT
ion. WA and wm are the 1 7 0 NMR chemical shifts from acidified
- = k,, = exp(M*/R - AH*/RT) (6)
Tm
water and from water bound to the paramagnetic cation expressed
in rad s-I. 1/T2, and 1/Tlr refer to 1 7 0 reduced relaxation rates,
and Awr refers to the reduced chemical shift observed for the The temperaturedependence of Awm,the chemicalshift between
NMR signal of bulk water caused by the presence of the Gd3+ the bound and the free solvent, has been successfully described
ions. The first term in eq 3 describes the first-coordination-shell by a power series of the inverse t e m p e r a t ~ r e . ~In~the present
effect, and the second term, 1/ T2-, includes the influences of all case, however, only the first term due to hyperfine interaction
between the unpaired electrons of the Gd3+ion and the 1 7 0nucleus
(24) Graeppi, N. Personal communication, University of Lausanne. of bound water will be considered (eq 7), where is the Land&
(25) Geier, G.; Karlen, U.; von Zelewsky, A. Helu. Chim.Aero 1969, 52,
1967.
(26) For [Ln(DOTA)(H20)]- (Ln = Eu,Tb) the low-temperaturelimiting
’Hand I3C spectra support the presence of two isomeric forms. The (7)
structures of the two isomers differ in the layout of the acetate arm,
the cyclododecane ring conformation being identical in both species.2’-**
(27) Brittain, H. G.; Desreux, F. F. Inorg. Chem. 1984, 23, 4459. factor, ~g is the Bohr magneton, 71is the gyromagnetic ratio of
(28) Aime, S.; Botta, M.; Ermondi, G. Inorg. Chem. 1992, 31,4291.
(29) Zimmermann, J. R.; Brittin, W. E. J. Phys. Chem. 1957, 61, 1328. the nucleus I (which is 1 7 0 in this case), S is the electron spin
(30) Bloembergen, N. J. Chem. Phys. 1957, 27, 595. quantumnumber of Gd3+(which is 7/2), andA/h is the hyperfine
170NMR Study of Water Exchange Inorganic Chemistry, Vol. 32, No. 18, 1993 3841

11). The electron spin relaxation rates for metal ions in solution
0 r1"O
-1= - 1.- 1
j = 1,2 (1 1)
7sj 7m Tje
with S 1 l / 2 are normally governed by a transient zero-field
splitting (ZFS), induced by distortionsof the complex. Analytical
~ limit of o,sv << 1
expressions are given by M ~ L a c h l a nin~the
(eqs 12-14).33 In these equations A* is the trace of the square

0
2.75 2.95 3.15 3.35 3.55 3.75
lOOO/T IK"
-
1 1
-A27,[4S(S + 1) - 3](J0 + 5J1 + 2J2) (13)
300 100 T2e 50

Jj = [ l +(~W~T~)~]-' (14)
of the zero-field-splitting (ZFS)tensor and rVis the correlation
time for the modulation of the ZFS. It is generally assumed that
the temperature dependence of 7, follows an Arrhenius behavior
(eq 15). The electron spin relaxation times Tle and Tz, of the

0 0
2.75 2.95 3.15 3.35 3.55 3.75 Gd3+ion are generally very long (of the order of 10-7-10-9 s) and
100OfT IK" allow therefore to a very good approximation the use of eq 16
Figure 3. Longitudinal I7O NMR relaxation rates data for (a) 0.05 M instead of eq 10.
[Gd(DTPA)(H20)I2-and (b) 0.05 m [Gd(DOTA)(H20)]- in water as
a function of temperature and magnetic field ( 0 )9.4 T; (0)4.7 T. The
thin curve shows the relaxation rates, l / T l ~of, acidified water as a
function of temperature. Filled symbols represent the relaxation increase
caused by the paramagnetic complex, and the thick curve represents the
best fit through these points. In the case of the longitudinal relaxation, the important
contributions come from the dipolar relaxation 1/?'1dd and from
coupling constant. Theouter-spherecontribution to thechemical thequadrupolerelaxation 1 /TIq. These relaxation rates are both
shift is assumed to be linearly related to Aw, by a constant C
, mostly governed by thecorrelation time ?,describing the rotational
(eq 8). motion of the whole complex (eqs 17-19). The oxygen-metal
Awm = CmAwm (8)
The relaxation of an 170 nucleus in the first coordinationsphere
of a paramagnetic ion can be expressedas a sum of three relaxation
mechanisms: (i) the relaxation due to interaction of the electric -1= - +1- + - 1 1
quadrupole moment with electric field gradients, (ii) the scalar j = 1,2
?dj Tje 7m 7c
interaction of the nuclear spin with the electron spin, and (iii) the
dipoledipole interaction of the nuclear spin with the electron
spin (eq 9).
-=-+-+-
1 1 1 1
j = 1,2 (9) distance, r, was f i e d to 2.5 A.34 The product x(1 + v2/3)lI2,
Tim TI¶ qsc Tjdd
which contains the quadrupole coupling constant x and the
In the case of the transverse relaxation, the scalar contribution asymmetry parameter tl, was set to the value of pure water, that
1 / T h i s by far the most important and we neglect therefore is 7.58 MHz. The temperature variation of the rotational
the other two (eq 10). In this last equation ws is the Lamor correlation time, T,, can be expressed by eq 20.

The temperature dependence of the longitudinal relaxation

frequency of the metal electron spin and 1/T8j is the sum of the rates of the [G~(DTPA)(HzO)]~- and [Gd(DOTA)(HzO)]-
exchange rate constant and the electron spin relaxation time (eq solutions and of acidified water are shown in Figure 3. The
(31) 1/T% is identical to 1/Tlq as expressed by eq 19, and 1/Tm is given
increase in relaxation due to the interaction with the paramagnetic
by center representsonly a few percent of the total observed relaxation
rates and so is attributed only to the inner-sphere effect (l/Tlm
~~

(32) McLachlan, A. D. Proc. R. Soc. London, A 1964, 280, 271.

(33) Powell, D. H.; Briicher, E.; Gonzalez, G.; Grinberg, 0.Y.;KBhler, K.;
The validity of the approximation was verified by using rotational Lebedcv,Y.S.; Merbach, A. E.; Micskei, K.; Ottaviani, F.; von Zelewsky,
correlation times as obtained from 1/ TI,;
the sum of these two relaxation A. Helv. Chim. Acta, in press.
rates calculated is at most 5% of the scalar relaxation rate; the (34) Helm, L.; Merbach, A. E. Eur. J. Solid SfufeInorg. Chem. 1991,28,
approximation introduced above (eq 10) is fully justified. 245.
3848 Inorganic Chemistry,Vol. 32, No. 18, 1993 Micskei et al.

I I 1 1

l5 t
14 I

l2 t i
r

'9
x -0.50

2.75
I

3.00
0
I

3.25 3.50
-0* I
2.75
I

3.00 3.25
I 1

3.50
1

IOOOIT IK -' IOOO/l IK -'

Figure 4. Variable-temperature 1 7 0 NMR data for 0.05 m [Gd- Figure 5. Variable-temperature 170 NMR data for 0.05 m [Gd-
(DTPA)(H20)J2- in water showing ln(l/Tzr) and Awr vs inverse (DOTA)(H20)]- in water showing ln(l/T2,) and Aw, vs inverse
temperature: (13)9.4-T data; (0)4.7-T data. The curvca are the rcault temperature: ( 0 )9.4-T data; (0)4.7-T data. The curves are the result
of a simultaneous fit of all data. The short dashed curve shows the of a simultaneous fit of all data. The short dashed curve shows the
contribution of 1/~,,,,and the long dashed curves show the contribution contribution of l/rm,and the long dashed curves show the contribution
of 1 / T b to 1/T2,. Of 1/T& to l/Tzr.

= 0). A least-squares fit of these data allows a rough estimation of acti~ati0n.l~ Very often the deviation from linear behavior is
of the rotational correlation times and their activation energies well within experimental error, and 46' is thereforeset to
(eqs 5 and 17-20) for DTPA (DOTA): T~~~~= 103 f 10 ps (90 The evaluation of the pressure dependence of 1 / Tzmis more
f 15 ps) and Ec = 18 f 2 kJ mol-] (17 f 3 W mol-'). difficult (eq 16). In earlier studies is was found that the scalar
coupling constant A / h is nearly pressure in depend en^^' We
For the analysis of the transverse relaxation rates and the fixed it therefore to its value at ambient pressure. The pressure
chemical shifts, a combined nonlinear least-squarestreatment of dependence of the scalar correlation time ~~1(eq 11) is given by
all corresponding data was performed using eqs 3, 4, 6-8, and the pressuredependenceof the electronspin relaxationrate 1 / Tle
1 1-16.35 In a first approximation we neglected the contribution and the pressure dependence of the exchange rate kex. At the
from the outer-sphere interaction with the Gd3+. The adjusted higher temperatures (337.4 and 339.8 K) ~~1 is about 3040%
parameters are AH*, AS*(or kcx298), A / & ,C,, T ~E,, and ~ A2
~ ~ governed
, by 1/Tle and 7040%by kex. The results obtained at
(1/T2, = 0). The last three parameters defining the electronic both temperatures were fitted simultaneously to q s 3, 11, 16,
relaxation rate 1/ Tleare stronglycorrelated, leadingto relatively and 21 with variable parameters AV and (kcx)oT,and l/Tle was
large statistical errors. The results are given in Table 11, and the kept pressure independent. 1 /TI, is related to T, according to eq
calculated curves for DTPA and DOTA are shown in Figures 4 12, and reasonable pressurevariation of 7, (lAVV*II 5 cm3mol-')
and 5 , respectively. has no effect on the kinetic A V within the reported experimental
uncertainty.
VariabbpressUreNMR. the variable-pressure measurements
were performed at 285.6 K/337.4 K (DTPA) and 285.0 K/339.8 Discuseion
K (DOTA) and the results are given in Figure 6. At the lower The rate of water exchange from the inner sphere of a metal
temperatures 1/T2, depends mainly on k, = l/Tmwhereas at the ion is affected strongly by the substitution of one or more water
higher ones l/Tzmgives the major contribution (see Figures 4 molecules by a coordinated ligand. For the complexes of the
and 5). The pressure dependence of the exchange rate constant first-row transition elements there are a lot of experimentaldata
k,, is usually assumed to be a quadratic function of the pressure showing that multidentate ligands increase significantly the rate
P (eq 21). In this equation (kex)oris the rate constant at zero of exchange of the remaining water molecules.18 Contrary to
these results, the water-exchange rate was found to be lower for
ln(kex) = ln(k,)Z- PAV,*/RT+ pAfi'/2RT (21) the [Gd(PDTA)(H20)2]- complex (kcx298= 3.3 X lo8 s-l) than
(36) This approximation waa further checked by using A@* as a variable in
pressure and temperature T , AVO*is the activation volume of the the fitting procedure, leading to the following values for DTPA
(DOTA): AV/cm-'mol-l = +14.4&0.5 (+12.6&0.3), lO-*A@*/cm-~
reaction at zero pressure, and A@*is the compressibilitycoefficient *
mol-I MPa-I = +1.9 & 0.5 (+2.3 0.3). Using this procedure, the
goodness of the fit is not significantly improved and the difference in
AV values is within the error accepted for such determinations (a1
(35) During the fitting procedure, a higher weight was given to relaxation ~ m mol-',
- ~ but at least 10%).
data as comparedto chemicalshift data; due to relatively small observed (37) Cosay, C.; Helm, L.; Merbach, A. E.Inors. Chem. 1989, 28, 2699.
shifts (20-120 Hz)compared to observed line widths of 1W500 Hz, (38) Ducommun, Y.;Newman, K.E.;Merbach, A. E.Inorg. Chem. 1980,
the uncertainty in Aur is high. 19, 3696.
170NMR Study of Water Exchange Inorganic Chemistry, Vol. 32, No. 18, 1993 3849
Table II. Kinetic and NMR Parameters Derived from Relaxation
14.00 and Chemical Shift Data as a Function of Temperature and Pressure

13.75
4.1 A 0.3 4.8A 0.4
h
& 52.0A 1.4 *
48.8 1.6
--
% 13.50 +56.2 f 5
+12.5 f 0.2
+46.6 6
+lo3 f 0.2
13.25
-3.8 f 0.2 -3.4 f 0.3
0.13A 0.06 0.25 0.08
13.00 1.5 f 0.2 1.2 f 0.3
6.3f 0.8 3.8 A 0.4
7f4 6k4
a The fitted exchange rate constants at zero pressure are as follows.
13.75 DTPA system: ( k a x ) ~ =~ 1~.3. 7~X lo6 s-I, (k,)0337.4 = 5.70 x 107 s-1.
DOTA system: (&,,)0285.0 = 1.83 X 106 s-I, (&u)~339,8 = 6.24 X lo's-I.
A
&
5 13.50 Table III. Comparison of Relevant ESR Parameters for Different
-c
Y
Gd3+ Complexes
complexa 1V'9Az/s-2 10+13~v298/sE+ mol-'
13.25
[Gd(HzO)el'+ 9.3 72 15.4
[Gd(PDTA)(HzO)z]- * 8.0 480 10.2
13.001 ' I I I [Gd(DTPA-BMA)(HzO)] 3.8 1400 17.6
0 50 100 150 200 [Gd(DTPA)(HzO)]2- 0 1.5 6.3 7
[Gd(DOTA)(H20)]- 0 1.2 3.8 6
? /M?a
Figure 6. Variable-pressure I7O N M R data for 0.05 m [Gd- a PDTA' = 1,3-propylenediaminetetraacetate;DTPA-BMAI- =
(DTPA)(H2O)l2- and [Gd(DOTA)(H20)]- in water. The curves diethylenetriaminepentaacetatebis(methylamide1. From ref 33. This
represent the result of a simultaneous tit of the data, measured at two work.
temperatures, for DTPA and DOTA, respectively.
by the functional groups of DTPA5- or DOTA' and there is only
for [Gd(HzO)8I3+ (keXzg8 = 1.19 X lo9 S-I).~~ Unfortunately, enough space for the coordination of one water molecule. As a
there are no other data known for the lanthanide complexes, but consequence, during the water exchange there is no possibility
our results obtained for [Gd(DTPA)(Hz0)lz- and [Gd- for the increase of the coordination number to 10. The entering
(DOTA)(H20)]- (Table 11) show thesame trend, since the water- HzO molecule can be coordinated only on the site of the leaving
exchange rates for these complexes are more than 2 orders of HzO and cannot participate in the bond-breaking process, 80 that
magnitude lower than that for [Gd(H2O)8I3+. As a consequence the exchange is relatively slow. In the exchange reaction of the
of the unexpectedly slow water exchange, at lower temperatures Gd3+aqua ion, the species [Gd(Hz0)8l3+can easily coordinate
the transverse relaxation rate of H2I7O is controlled by the a ninth HzO molecule, forming a [Gd(Hz0)9I3+intermediate,
chemical exchange (Figure 2.). The explanation of the different which to a small extent participates in an aquation equilibrium,
behaviors of the complexes of the first-row transition elements as shown from neutron-scattering studies on aqueous solutions
and the lanthanides is not easy. We can take into consideration of Ln3+ ions.34 As a result of this equilibrium, the coordination
that the chemical bonds in the lanthanide complexes are of the ninth molecule is only weakly hindered. It helps in the
predominantly ionic, while in the complexes of transition metals, bond breaking of the leaving HzO, and the process can be faster
the bonds are partly covalent, and as was discussed by Margerum than the dissociation of HzO from [Gd(DTPA)(Hz0)lz- or
et al., in the case of the Ni2+complexes18the increasing a-donating [Gd(DOTA)(HzO)]- in spite of the somewhat longer Ln3+-OH2
strength of the coordinated ligand leads to the increase in the bonds in the c o m p l e ~ e s . ~This
. ~ ~dissociative activation mode is
water-exchange rate.18 The acetate groups are weak u-donors, further supported by the large activation enthalpies AH* and
and the labilizing effect of the polyacetate ligands is relatively positive activation entropies AS' (Table 11).
low.4-18 In the discussion of the results obtained for the [Gd-
For interpretation of the low water-exchange rate of the (DTPA)(Hz0)l2- and [Gd(DOTA)(HzO)]- complexes, we can
complexes of DTPAS- and DOTAG, the obtained activation distinguish between parameters giving information on the electron
volume data are also useful (Table 11). The activation volume spin relaxation and kinetic parameters linked to the water
for the water exchangeof [Gd(Hz0)8l3+was not determined, but exchange from the first coordination sphere. Until now, infor-
the data obtainedfor the Tb3+-Tm3+series (AV=-6 cm3m0l-I)3~ mation on the electron spin relaxation of these complexes has not
clearly indicate a concerted associative I, mechanism for these been available. We can compare our results to ESR and NMR
elements,which can be assumed to bevalid for the next neighbor data obtained on the aqua ion and on similar Gd3+ complexes.
in the series, the Gd3+aqua ion. However, the activationvolumes From Table I11 we see that the trace of the square of the ZFS
obtained for the water exchange of [Gd(DTPA)(Hz0)Jz- and tensor, Az, is smaller for [Gd(DTPA)(Hz0)l2- and [Gd-
[Gd(DOTA)(HzO)]- are large and positive, indicating a mech- (DOTA)(HzO)]- than for the other systems and the correlation
anism close to a limitingdissociative (D) mechanism for the water time, 7v298,for the modulation of the ZFS is much shorter even
exchange. These activation volumes are the largest values than in the water complex. The distortion of the ZFS may be
compared with those for other ions: e.g., a value of +7.2 cm3 caused by impact of the complex with solvent water m o l e c ~ l e s . ~ ~
mol-' has been reported for N i ( H ~ 0 ) 6 ~ The
+ . ~significant
~ change In our case, where a relatively large and rigid ligand is bound to
in the mechanism of the water exchange is probably the result the paramagnetic ion, such a process may be less effective.43
of steric hindrance. In the complex, the Gd3+ion is surrounded However, the binding of a ligand can lead to an anisotropy of the
g tensor that can be the source of other relaxation mechanisms
(39) The rate constant 1.06 X lo9 s-I for the water exchange of the Gd3+ ion
was obtained by assuming the prtscnce of [Gd(H20)9l3+ions.17 The (40) Stezowsky, J. J.; Howard, J. L. Isr. J. Chem. 1984,24, 323.
results of neutron diffraction studies proved the predominance of [Gd- (41) Sinha, S. P.Srruct. Bonding (Berlin) 1976,25, 69.
(H20)sl3+ions, and the recalculated value of the rate constant is kWz9* (42) Bloembergen, N.; Morgan, L. 0.J. Chem. Phys. 1961, 34, 842.
= (918) x 1.06 x 109 S-1 = 1.19 x 109 S-1. (43) Dwek, R.A.Adu. Mol. Relax. Processes 1972,4 , 1 .
 3850 Inorganic Chemistry, Vol. 32, No. 18, 1993 Micskei et al.
for the electron spin, such as spin rotation interaction, so that the proton has a much higher contribution from dipolbdipole
values we obtain may not be very accurate. interaction with the electron spin (about 55 times for the same
The exchange of water molecules between the inner sphere of distance due to the difference in the gyromagnetic ratio). If we
Gd3+and the bulk water is of high importance in MRI, since the consider now a water molecule in the second coordination sphere,
paramagnetic effect on bound water is transferred to the bulk where scalar interaction is unimportant, we find a much higher
water (and to the tissue) by this process. The optimal value of interaction between water protons and the paramagnetic ion than
the residence lifetime, T,, for obtaining high relaxivities in MRI between the oxygen-17 nucleus and the ion.
depends mainly on the rotational correlation time, T ~ of , the All these considerations lead to the conclusion that the
complex and on the MRI frequency. In freely rotating complexes, effectivenessof [Gd(DTPA) (HzO)] 2- and [Gd(DOTA)(H2O)l-
the longitudinal relaxation time of water protons in the bound as contrast agents in MRI is limited not by the relatively low
state, TlmProton, is dominated by the rotational correlation time water-exchange rates but by Tim, the longitudinal relaxation time
rC and is of the order of 10-4-1V s. Because TlmProton >> T,, the of water protons in the first coordination sphere.
influence of the proton residence time, for which the water
Acknowledgment. We thank Prof. J. F. Desreux for providing
residence time can be taken as an upper limit, on the relaxivity
the TETA ligand used in this study as well as Dr. H. Powell for
is small in this case. If, however, the complex is immobilized, T~
useful discussions in the interpretation of electron relaxation
becomes longer and can have the same magnitude as T I ,and T ~ .
problems. This work was financially supported by the Swiss
According to calculations carried out by Lauffer4 for such
National Science Foundation (Grants 20-32703.9 1 and 70UP-
immobilized molecules, the optimal T, values are between
029558) and the Hungarian Scientific Research Foundation
and lo-* s at a magnetic field strength of 0.47 T. The residence
(Grants 1643/91 and 1724/91).
lifetimes of water molecules as obtained by I7O NMR for
[Gd(DTPA) (HzO)]2- and [Gd(DOTA) (HzO)]- represent an Sopplemeatary MaterialAvailable: Variable-temperature UV-visible
upper limit for proton residence times and are close to this interval spectra of [Eu(DTPA)(HzO)]z-, [ Eu(DOTA)(HzO)]2-, and [Eu-
(about s at 37 "C). (TETA)I2-(Figures S l S 4 ) , variable-temperatureI7O relaxation rates
The low oxygen- 17 relaxivity observed for [Gd(TETA)]- of acidified water (Table Sl), variable-temperatureI7Orelaxation rates
and chemicalshiftsof [Gd(DTPA)(HzO)]2- (TableSZ),variablspreasure
(Figure 2) is in contrast with its relatively high proton relaxivity4 I7Orelaxation ratesof [Gd(DTPA)(H20)]z- (9.4 T) (TableS3),variable-
(approximately 3040% of that of [Gd(DTPA)(HzO)I2- or temperature 170relaxation rates and chemical shifts of [Gd(DOTA)-
[Gd(DOTA)(HzO)]-). This finds its explanationin the difference (H20)]- (Table S4), variable-pressure I7O relaxation rates of [Gd-
of relaxation mechanisms between proton and oxygen-17. While (DOTA)(HzO)]- (9.4 T) (Table S5), and variable-temperature 170
the latter in the first coordination sphere interacts mainly via relaxation rates andchemical shiftsof [Gd(TETA)]-(TableS6)(9 pages).
scalar interaction with the electron spin of the Gd3+ ion, the Ordering information is given on any current masthead page.

View publication stats