Thermodynamic stability and physicochemical characterization of ligand (4S)-4-benzyl-3,6,10-tris(carboxymethyl)-3,6,10-triazadodecanedioic acid (H5[(S)-4-Bz-ttda]) and its complexes formed with lanthanides, calcium(II), zinc(II), and copper(II)

Article abstract of DOI:10.1002/1522-2675(200204)85:4<1033::AID-HLCA1033>3.0.CO;2-N

A derivative of H₅ttda (=3,6,10-tris (carboxymethy -3,6,10-triazadodecanedioic acid = N-{2-[bis(carboxymethyl)amino]ethyl}-N- {(3-[bis(carboxymethyl)amino]propyl}glycine), H₅[(S)-4-Bz-ttda] (=(4S)-4-benzyl-3,6,10-tris(carboxymethyl)-

Full text of DOI:10.1002/1522-2675(200204)85:4<1033::AID-HLCA1033>3.0.CO;2-N

Helvetica Chimica Acta Vol. 85 (2002)
1033
Thermodynamic Stability and Physicochemical Characterization of
Ligand (4S)-4-Benzyl-3,6,10-tris(carboxymethyl)-3,6,10-triazadodecanedioic
Acid (H5[(S)-4-Bz-ttda]) and Its Complexes Formed with Lanthanides,
Calcium(II), Zinc(II), and Copper(II) Ions
by Tsan-Hwang Chenga), Tzu-Ming Leea), Ming-Hung Oua), Cha-Ru Lia), Gin-Chung Liub), and
Yun-Ming Wang*a)
a) School of Chemistry, Kaohsiung Medical University, 100 Shih-Chuan 1st Road, Kaohsiung, Taiwan 807,
Republic of China
b) Department of Radiology, Kaohsiung Medical University, 100 Shih-Chuan 1st Road, Kaohsiung, Taiwan 807,
Republic of China
A derivative of H5ttda (ˆ 3,6,10-tris(carboxymethyl)-3,6,10-triazadodecanedioic acid ˆ N-{2-[bis(carboxy-
methyl)amino]ethyl}-N-{3-[bis(carboxymethyl)amino]propyl}glycine), H5[(S)-4-Bz-ttda] (ˆ(4S)-4-benzyl-
3,6,10-tris(carboxymethyl)-3,6,10-triazadodecanedioic acid ˆ N-{(2S)-2-[bis(carboxymethyl)amino]-3-phenyl-
propyl}-N-{3-[bis(carboxymethyl)amino]propyl}glycine; 1) carrying a benzyl group was synthesized and
characterized. The stability constants of the complexes formed with Ca2‡, Zn2‡, Cu2‡, and Gd3‡ were
determined by potentiometric methods at 25.0 Æ 0.18 and 0.1m ionic strength in Me4NNO3. The observed water
proton relaxivity value of [Gd{(S)-4-Bz-ttda}]2À was constant with respect to pH changes over the range
pH 4.5 12.0. From the 17O-NMR chemical shift of H2O induced by [Dy{(S)-4-Bz-ttda}]2À at pH 6.80, the
presence of 0.9 inner-sphere water molecules was deduced. The water proton spin-lattice relaxation rate for
[Gd{(S)-4-Bz-ttda}]2À at 37.0 Æ 0.18 and 20 MHz was 4.90 Æ 0.05 mmÀ1 sÀ1. The EPR transverse electronic
relaxation rate and 17O-NMR transverse-relaxation time for the exchange lifetime of the coordinated H2O
molecule (tM), and 2H-NMR longitudinal-relaxation rate of the deuterated diamagnetic lanthanum complex for
the rotational correlation time (tR) were thoroughly investigated, and the results were compared with those
previously reported for the other lanthanide(III) complexes. The exchange lifetime (tM) for [Gd{(S)-4-Bz-
ttda}]2À (2.3 Æ 1.3 ns) was significantly shorter than that of the [Gd(dtpa)(H2O)]2À complex (dtpa ˆ diethyl-
enetriaminepentaacetic acid). The rotational correlation time tR for [Gd{(S)-4-Bz-ttda}]2À (70Æ 6 ps) was slightly
longer than that of the [Gd(dtpa)(H2O)]2À complex. The marked increase of relaxivity of [Gd{(S)-4-Bz-ttda}]2À
mainly resulted from its longer rotational time rather than from its fast water-exchange rate. The noncovalent
interaction between human serum albumin (HSA) and the [Gd{(S)-4-Bz-ttda}]2À complex containing the
hydrophobic substituent was investigated by measuring the solvent proton relaxation rate of the aqueous
solutions. The association constant (KA) was less than 100 mÀ1, indicating a weaker interaction of [Gd{(S)-4-Bz-
ttda}]2À with HSA.
Introduction. The use of paramagnetic metal complexes as contrast agents for
magnetic resonance imaging (MRI) has been under intense scrutiny [1 3]. A major
class of contrast agents are the paramagnetic substances that accelerate magnetic
relaxation processes of H2O protons. Among the paramagnetic contrast agents, GdIII
complexes are of particular interest. The search for analogous complexes that would
enter hepatocytes or provide a strong noncovalent bonding interaction with the blood
plasma protein human serum albumin (HSA) led to a series of GdIII complexes. They
are prepared from poly(aminocarboxylic) chelating ligands with various lipophilic
substituents. Representative examples are MS-325 (ˆ [(4R)-4-({[(4,4-diphenylcyclo-

1034
Helvetica Chimica Acta Vol. 85 (2002)
hexyl)phosphono]oxy}methyl)-3,6,9-tris(carboxymethyl)-3,6,9-triazaundecanedioa-
to(6À)]gadolinate(3À)) [4], [Gd(bopta)]2À (ˆ [4-carboxy-5,8,11-tris(carboxymethyl)-
1-phenyl-2-oxa-5,8,11-triazatridecan-13-oato(5À)]gadolinate(2À)) [5], [Gd(eob-
dtpy)]2À (ˆ [3,6,9-tris(carboxymethyl)-4-(4-ethoxybenzyl)-3,6,9-triazaundecanedioato-
(5À)]gadolinate(2À)) [6], and MP-2269 (ˆ(N2,N2-bis{2-[bis(carboxymethyl)amino]-
ethyl}-N6-{N-[(4-pentylbicyclo[2.2.2]oct-1-yl)carbonyl]-l-a-aspartyl-l-a-aspartyl}-l-
lysinato(7À))gadolinate(4À) [7].
The presence of an aromatic group in the ligand seems to lead to some tissue
specificity of the complex as well as interaction through noncovalent binding with
serum albumin [5 7]. To extend these observations, we synthesized H5[(S)-4-Bz-ttda]
(ˆ(4S)-4-benzyl-3,6,10-tricarboxymethyl-3,6,10-triazadodecanoic
acid ˆ N-{(2S)-2-
[bis(carboxymethyl)amino]-3-phenylpropyl-N-{3-[bis(carboxymethyl)amino]propyl}-
glycine; 1), a derivative of H5ttda [8] bearing a benzyl group (Scheme). The
protonation, thermodynamic, and conditional stability constants of its complexes with
Gd3‡, Cu2‡, Zn2‡, and Ca2‡ were also determined. Furthermore, its selectivity for Gd3‡
1
over endogenously available metal ions was determined. The H-NMR for complex
[La{(S)-4-Bz-ttda}]2À in solution was also studied. The number of inner-sphere water
molecules of [Dy{(S)-4-Bz-ttda}]2À was determined by the 17O-NMR chemical shift of
the water as a function of the Dy3‡ concentration. The water proton spin-lattice
relaxivity r1 of the [Gd{(S)-4-Bz-ttda}]2À complex vs. temperature and pH was
Scheme. Synthesis of (S)-4-Bz-TTDA (1)

Helvetica Chimica Acta Vol. 85 (2002)
1035
determined. The EPR and 17O-NMR transverse-relaxation-rate data were analyzed
together in a simultaneous multiple-parameter least-squares fitting procedure to
2
determine the water residence lifetime. The H-NMR was used to determine the
rotational correlation time. In the experiment of albumin binding ability, the value of
KA for ([Gd{(S)-4-Bz-ttda}]2À) ¥ HSA was also determined.
Results and Discussion. Protonation Constants. Table 1 summarizes the proto-
nation constants of H5[(S)-4-Bz-ttda], H5[(S)-4-eob-dtpa], the ligand of MS-325,
H5dtpa, and H5ttda [4][8 10]. The titration curve of H5[(S)-4-Bz-ttda] shows an
increase at pH 8.7 5.7 at a ˆ 3 (a ˆ mol base/mol ligand present). This is due to the
large difference between the second (log K2) and third (log K3) protonation constants,
i.e., 8.91 and 5.10. The protonation constant of H5[(S)-4-Bz-ttda] is significantly higher
than those of H5[(S)-4-eob-dtpa] and H5dtpa and is similar to those of H5ttda and the
ligand of MS-325. This is due to higher basicity in the skeleton of H5ttda as compared to
H5dtpa. However, the benzyl-substituent at the ethylene backbone in H5[(S)-4-Bz-
ttda] does not alter the protonation constant significantly as compared to H5ttda. The
difference of the protonation constant of H5[(S)-4-Bz-ttda] and H5ttda is similar to that
of H5[(S)-4-eob-dtpa] and H5dtpa.
Thermodynamic Constants. For the potentiometric titration curves of the com-
plexes of Ca2‡, Cu2‡, and Zn2‡ with H5[(S)-4-Bz-ttda], see Appendix (Fig. 1S). The
thermodynamic stability constants of metal chelates with [(S)-4-Bz-ttda], [(S)-4-eob-
dtpa]5À, the ligand of MS-325, dtpa5À, and ttda5À are shown in Table 2 [4][8 10]. The
stability constant of the Gd3‡ complex with [(S)-4-Bz-ttda]5À is similar to those with
[(S)-4-eob-dtba]5À and the ligand of MS-325 but slightly higher than those with ttba5À
and dtpa5À. This interesting result might be attributed to the lipophilic substituent at the
ethylene backbone of [(S)-4-Bz-ttda]5À, [(S)-4-eob-dtpa]5À, and the ligand of MS-325.
It has been reported that Gd3‡ complexes with poly(aminocarboxylates) display a
hydrophilic region (the metal ion and the carboxylates) and a hydrophobic region (the
ligand backbone) [11]. With the lipophilic substituent at the ethylene backbone, the
ligands tend to separate more compared to those without any substituent or with a
hydrophilic substituent, resulting in less steric hindrance as they coordinate to the Gd3‡
ion. The order of the conditional stability constants of Gd3‡ complexes (log KGdL) at
pH 7.4 is [Gd{(S)-4-Bz-ttda}]2À ꢀ [Gd{(S)-4-eob-dtpa}]2À > MS-325 ꢀ [Gd(dtpa)]2À
ꢀ
[Gd(ttda)]2À, as shown in Table 2. The results of pH dependence of the conditional
stability constants for [Gd{(S)-4-Bz-ttda}]2À and [Gd(ttda)]2À are very similar (see
Table 1. Protonation Constants of the Ligands at 25.0 Æ 0.18a)
H5[(S)-4-Bz-ttda] H5[(S)-4-eob-dtba]b) Ligand of MS-325c) H5dtpad) H5ttdae)
[HL]/[L][H]
[H2L]/[HL][H]
[H3L]/[H2L][H]
[H4L]/[H3L][H]
SpKa
10.70(0.02)
8.91(0.06)
5.10(0.04)
3.37(0.03)
28.08
10.91
8.63
4.26
2.73
26.53
11.15
8.62
4.51
2.96
27.24
10.49
8.60
4.28
2.64
26.01
10.60
8.92
5.12
2.80
27.44
a) All data were obtained in aqueous Me4NNO3 solution, I ˆ 0.1m, except for b) and c). b) From [10] (KCl, I ˆ
0.1m). c) From [4] (Me4NCl, I ˆ 0.1m). d) From [9]. e) From [8].

1036
Helvetica Chimica Acta Vol. 85 (2002)
Table 2. Stability Constants, Selectivity Constants, and Modified Selectivity Constants for Gd3‡, Zn2‡, Ca2‡, and
Cu2‡ Complexes at 25.0 Æ 0.18a)
b
d
e
[(S)-4-Bz-ttda]5À [(S)-4-eob-dtba]5À
)
Ligand of MS-325c) dtpa5À
)
ttda5À
)
[GdL]/[Gd][L]
log KGdL' (pH 7.4)
log ([CaL]/[Ca][L])
23.79(0.04)
18.97
15.95(0.03)
5.24(0.04)
11.13
19.70(0.01)
3.89(0.02)
14.88
18.61 (0.02)
8.63(0.02)
13.79
23.46
18.70
11.74
5.75
23.2
18.20
10.45
5.66
5.45
21.3
5.16
16.30
17.82
5.60
12.82
5.38
12.75
1.9
22.46
18.14
10.75
6.11
6.43
21.38
4.81
17.0
18.70
5.60
14.38
3.76
11.71
1.08
22.77
18.04
14.45
log b
CaHL
log KCaL' (pH 7.4)
log [CuL]/[Cu][L]
6.98
9.72
19.31
log b
CuHL
log KCuL' (pH 7.4)
log [ZnL]/[Zn][L]
14.58
18.59
log b
ZnHL
log KZnL' (pH 7.4)
13.86
4.18
8.32
3.46
8.44
selectivity[log K(Gd/Zn)] 5.18
selectivity[log K(Gd/Ca)] 7.84
selectivity[log K(Gd/Cu)] 4.09
11.72
log Ksel'
9.35
7.89
7.06
a) All data were obtained in aqueous Me4NNO3, I ˆ 0.1m, except for b) and c). b) From [10] (KCl, I ˆ 0.1m).
c) From [4] (Me4NCl, I ˆ 0.1m). d) From [9]. e) From [8].
Fig. 2S) in the Appendix. The conditional stability constants at pH > 11 for [Gd{(S)-4-
Bz-ttda}]2À and [Gd(ttda)]2À differ by a factor of 10.5, similar to the value obtained at
pH 7.4 (8.5). Thus the stability constant of [Gd{(S)-4-Bz-ttda}]2À is slightly higher than
that of [Gd(ttda)]2À at all pH values.
The species-distribution curve of [Gd{(S)-4-Bz-ttda}]2À (see Fig. 3S in the
Appendix) indicates that there is still some free Gd3‡ at pH 3 but the complex is
fully formed above pH 4.53. The pM values [12] for Gd3‡, Zn2‡, Ca2‡, and Cu2‡
complexes with [(S)-4-Bz-ttda]5À, [(S)-4-eob-dtpa]5À, the ligand of MS-325, dtpa5À, and
ttda5À are given in the Appendix (Table 1S). The pGd value of [Gd{(S)-4-Bz-ttda}]2À
(17.97) resembles that of [Gd{(S)-4-eob-dtpa}]2À (17.70) and is slightly larger than those
of [Gd(ttda)]2À (17.03) and [Gd(dtpa)]2À (17.14). The stability constant of [Gd{(S)-4-
Bz-ttda}]2À is slightly larger than those of [Gd(dtpa)]2À and [Gd(ttda)]2À, resulting in an
increased pGd value for [Gd{(S)-4-Bz-ttda}]2À. However, the pGd value for [Gd{(S)-4-
Bz-ttda}]2À is closed to that of [Gd{(S)-4-eob-dtpa}]2À due to their similar stability. The
pM value of [Gd{(S)-4-Bz-ttda}]2À is larger than those of [Zn{(S)-4-Bz-ttda}]2À,
[Cu{(S)-4-Bz-ttda}]2À, and [Ca{(S)-4-Bz-ttda}]2À, indicating that Zn2‡, Ca2‡, and Cu2‡
cannot replace Gd3‡ in the [Gd{(S)-4-Bz-ttda}]2À complex at physiological pH.
The logarithmic selectivity constants for [(S)-4-Bz-ttda]5À, [(S)-4-eob-dtpa]5À, the
ligand of MS-325, dtpa5À, and ttda5À are also shown in Table 2. The log K(Gd/Ca) value
of [(S)-4-Bz-ttda]5À (7.84) is similar to that of ttda5À (8.32) but lower than those of [(S)-
4-eob-dtpa]5À, the ligand of MS-325, and dtpa5À. This result shows that ligands of type
ttda5À form relatively more-stable complexes with Ca2‡ than ligands of type dtpa5À. This
trend is also similar to that previously established for N-substituted [Ca(dtpa-py)]2À
(ˆ [3,9-bis(carboxymethyl)-6-(pyridylmethyl)-3,6,9-triazadodecanedioato(4À)]calciate-
(2À)) (log K ˆ 8.35) [13] and [Ca(ttda-py)]2À (ˆ [3,10-bis(carboxymethyl)-6-(pyridyl-
methyl)-3,6,10-triazadodecanedioato(4À)calciate(2À)) (log K ˆ 12.02) [14]. The log

Helvetica Chimica Acta Vol. 85 (2002)
1037
K(Gd/Cu) and log K(Gd/Zn) values of [(S)-4-Bz-ttda]5À are significantly higher than
those of dtpa5À and ttda5À, indicating that it exerts a more favorable selectivity toward
Gd3‡ over Zn2‡ and Cu2‡. The modified selectivity constants (Ksel') [9] values of [(S)-4-
Bz-ttda]5À, the ligand of MS-325, dtpa5À, and ttda5À at pH 7.4 were also determined
(Table 2). The log Ksel' of [(S)-4-Bz-ttda]5À is higher than those of the ligand of MS-325,
dtpa5À, and ttda5À. Thus, [(S)-4-Bz-ttda]5À forms a Gd3‡ complex that is more stable
than MS-325, [Gd(dtpa)]2À, and [Gd(ttda)]2À toward transmetallation with Ca2‡, Zn2‡,
and Cu2‡ metal ions at pH 7.4. This implies that the toxicity of [Gd{(S)-4-Bz-ttda}]2À
may be less than that of MS-325, [Gd(dtpa)]2À, or [Gd(ttda)]2À.
Dysprosium(III)-Induced Water 17O-NMR Shifts. Fig. 1 shows the dysprosium(III)-
induced water 17O-NMR shifts against Dy3‡ chelate concentration for solutions of
DyCl3 and [Dy{(S)-4-Bz-ttda}]2À in D2O at 258. The slope obtained for [Dy{(S)-4-Bz-
ttda}]2À at pH 7.00 is À44.1 ppm mmÀ1 (r2 ˆ 0.998). On the other hand, the slope for
DyCl3 is À382.8 ppm mmÀ1 (r2 ˆ 0.999), and the hydration number eight has been
proposed for the Dy3‡ ion [15 17]. Therefore, the [Dy{(S)-4-Bz-ttda}]2À complex at
pH 6.8 contains 0.9 inner-sphere H2O molecules per Dy3‡ ion. The number of Ln3‡-
bound H2O molecules in the metal complex provides information on the coordination
mode of the ligand. The coordination sites of the Gd3‡ ion in [Gd{(S)-4-Bz-ttda)]2À are
Fig. 1. Dysprosium(III)-induced water 17O-NMR shift vs. Dy3‡ chelate concentration in D2O at 25.0 Æ 0.18

1038
Helvetica Chimica Acta Vol. 85 (2002)
occupied by one H2O molecule while eight sites are available for the [(S)-4-Bz-ttda]5À
ligand.
Solution NMR of Lanthanum(III) Chelate. The solution 1H-NMR spectrum of 0.1m
[La{(S)-4-Bz-ttda}]2À in D2O at pD 7.0 and 25 Æ 0.18 is shown in Fig. 2. The peak
assignments were also confirmed by a 2D COSY spectrum. The CH2 protons of the
terminal acetate groups (Ha) give four AB patterns at d 3.58 (J ˆ 40 Hz), 3.56 (J ˆ
41 Hz), 3.25 (J ˆ 41 Hz), and 3.04 (J ˆ 40 Hz), respectively. The lack of the fifth AB
pattern indicates that there are only four carboxylate groups bound to the La3‡ ion. This
result is similar to that obtained from the solution NMR study of [La(dtpa)]2À [18][19].
In the case of [La(dtpa)]2À, there are two different AB patterns for the four terminal
acetate groups and a s for the central acetate, indicating that the central carboxylate is
not coordinate in the La3‡ complex.
Fig. 2. 1H-NMR Spectrum of [La{(S)-4-Bz-ttda}]2À at 25 Æ 0.18
Relaxometric Studies of the Gadolinium(III) Complexes. The relaxivity r1 values for
[Gd{(S)-4-Bz-ttda}]2À, measured at various pH values (Fig. 3), exhibit no pH
dependence over the pH range 4.5 12.0, indicating that no free Gd3‡ ion dissociates
from the [Gd{(S)-4-Bz-ttda}]2À complex in this pH range of the solution, consistent

Helvetica Chimica Acta Vol. 85 (2002)
1039
with the species distribution curve of [Gd{(S)-4-Bz-ttda}]2À (see Appendix, Fig. 3S).
Thus, the hydration number of [Gd{(S)-4-Bz-ttda}]2À remains constant in this pH range.
When the pH is below4.5, the [Gd{( S)-4-Bz-ttda}]2À complex may undergo partial
dissociation and result in a higher relaxivity. The longitudinal relaxivity r1 value of
[Gd{(S)-4-Bz-ttda}(H2O)]2À is 4.90 Æ 0.05 mmÀ1 sÀ1 at pH 6.8, 37.0 Æ 0.18, and 20 MHz.
The value is significantly higher than those of [Gd(ttda)(H2O)]2À (3.85 mmÀ1 sÀ1) [8]
and [Gd(dtpa)(H2O)]2À (3.89 mmÀ1 sÀ1) [9].
Fig. 3. pH Dependence of the relaxivity r1 for [Gd{(S)-4-Bz-ttda}]2À at 37 Æ 0.18 and 20 MHz
The paramagnetic contribution of the solvent longitudinal relaxivity is obtained by
Eqn. 1 [20], where C is the molar concentration of the gadolinium(III) complex, q the
number of water molecules bound to metal ion, T1M the longitudinal relaxation time of
the bound water protons, and tM the residence lifetime of the bound water. Because of
the opposite temperature dependencies of T1M and tM, two cases can be considered: 1)
is
fast water exchange (T1M ꢁ tM), r increases while temperature decreases; 2) slow
1p
is
water exchange (T1M ꢂ tM), r decreases while temperature decreases. Fig. 4 shows
1p
the temperature dependence of the relaxivity (r1) for the [Gd{(S)-4-Bz-ttda}]2À
complex at pH 6.8 and 20 MHz in the temperature range 278 343 K, i.e. a
monoexponential decrease of r1 while temperature increases. This is characteristic of
the fast chemical exchange behavior occurring when the tM of the coordinated H2O
molecule is much shorter than T1M of the bound water proton.
is
r
ˆ Cq/[55.6 (T1M ‡ tM)]
(1)
1p

1040
Helvetica Chimica Acta Vol. 85 (2002)
Fig. 4. Temperature dependence of the relaxivity r1 for the [Gd{(S)-4-Bz-ttda})]2À at pH 7.0 and 20 MHz
In fact, T1M can be expressed by Eqn. 2, where S is the electron-spin quantum
number (7/2 for Gd3‡), gH the proton nuclear magnetogyric ratio, b the Bohr
¬
magneton, g the Lande factor for the free electron, and rH the distance between the
metal ion and the bound water protons; wH and wS are the proton and electron Larmor
frequencies, respectively, and tCi (i ˆ 1, 2) is the correlation time of the modulation of
the dipolar electron-proton coupling. The overall correlation time tCi consists of
contributions from tM, tR, and tS (the electronic relaxation time of the metal ion)
(Eqn. 3). To realize how tM and tR have influences on r1 of [Gd{(S)-4-Bz-
2
ttda}(H2O)]2À, 17O- and H-NMR were used to determine the values of tM and tR,
respectively.
2 ꢀ
ꢁ
1
2 g2Hg2SꢃS ‡ 1†b
3tC1
7tC2
ˆ
‡
ꢃ2†
ꢃ3†
rH6
1 ‡ w2Ht2C1 1 ‡ w2St2C2
T1M 15
1
tCi tR tM tS
1
1
1
ˆ
‡
‡
Water-Exchange-Lifetime Studies of Gadolinium(III) Complexes. The measured
peak-to-peak line widths DHpp of the derivative spectrum can be related to the overall
¬
transverse-electronic-relaxation rate 1/T2e via Eqn. 4, where gL is the isotropic Lande g
factor (gL ˆ 2.0 for Gd3‡) [21].

Vol. 85 (2002)
1041
Helvetica Chimica Acpta
1
T2e
gLmBp 3
ˆ
DHpp
ꢃ4†
h
The temperature dependence of transverse-electronic-relaxation rates at the X-
band (0.34 T) at pH 7 for a 50 mm solution of [Gd{(S)-4-Bz-ttda}(H2O)]2À is given in
the Appendix (Fig. 4S). The data were fitted simultaneously with the following results
of 17O-NMR measurements. Analysis of the temperature dependence of the transverse
relaxation rate for the 17O water nuclei is the most accurate method for evaluating the
exchange lifetime of the water molecules directly coordinated to the metal in a
paramagnetic Gd3‡ chelate [22]. According to the theory of Swift and Connick [20], the
paramagnetic contribution (RO2p ) to the observed transverse relaxation rate is given by
Eqn. 5, where RO2M represents the 17O transverse relaxation rate of the coordinated
water molecule and DwOM the chemical-shift difference between the coordinated and
bulk water 17O-NMR resonances. R2OM is expressed by Eqns. 6 and 7, where S is the
electronic spin quantum number (7/2 for Gd(III)) and A/ꢀh the Gd,17O scalar coupling
constant; tei (i ˆ 1, 2) 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 mean residence
lifetime (tOM) of the water molecule at the paramagnetic site.
Cq
55:6
À1 RO2M ‡ ꢃtMO †À1R2OM ‡ DwOM2
RO2p
ˆ
ꢃtOM†
ꢃ5†
À1
ꢃR2OM ‡ ꢃtMO † †2 ‡ DOM2
ꢂ ꢃ2
ꢂ
ꢃ
1
3
A
ꢀh
te2
1 ‡ w2s t2e2
RO2M
ˆ
SꢃS ‡ 1† te1
‡
ꢃ6†
ꢃ7†
À1
À1
ei
t
ˆ tOMÀ1 ‡ T
ei
The temperature dependence of R2OM is determined by the temperature effect on tOM,
tV, and DwOM according to Eqns. 8 and 9 where the subscript j refers to the different
correlation times, DHj is the activation enthalpy for the corresponding dynamic process,
B the applied magnetic field strength, and kB the Boltzmann constant.
ꢀ
ꢂ
ꢃꢁ
ꢃtÀj 1
†
298:15T
298:15
1
1
T
DHj
R
À1
ꢃtj†
ˆ
exp
À
ꢃ8†
ꢃ9†
T
298:15
gLmBSꢃS ‡ 1†B A
DwOM
ˆ
3kBT
ꢀh
The water-exchange rate for [Gd{(S)-4-Bz-ttda)(H2O)]2À was obtained by
measuring the 17O-NMR transverse-relaxation rate (R20p ) as a function of temperature
[21][23]. The data and its best simulation according to Eqns. 4 9 is shown in Fig. 5. As
there are a large number of parameters to be determined in the quantitative analysis of
the 17O-NMR transverse-relaxation rate (R2Op ) vs. T profiles, it is convenient to fix some
of them. On this basis, besides the value of q and A/ꢀh (À 3.8 ¥ 106 rad sÀ1), the value of

1042
Helvetica Chimica Acta Vol. 85 (2002)
Fig. 5. Temperature dependence of the 17O-NMR water transverse-relaxation rate at 7.05 T and pH 7 for a 50 mm
solution of [Gd{(S)-4-Bz-ttda}(H2O)]2À. The line represents the simultaneous least-squares fit to all data points
as described in the test.
Table 3. Kinetic and NMR Parameters Obtained from the Simultaneous Fit of 17O-NMR and EPR Data
D2 [sÀ2 ¥ 1019
15 Æ 1.2
]
t2v98[ps]
t3M10[ns]
t3R10[ps]a)
DHv[kJ molÀ1
]
[Gd{(S)-4-Bz-ttda}(H2O)]2À
[Gd{(S)-4-Bz-dtpa}(H2O)]2Àb
3.2 Æ 0.3
23 Æ 1
2.3 Æ 1.3
70 Æ 6
65
64
8.8 Æ 2.7
8.6 Æ 8.0
1.7 Æ 0.9
1.6 Æ 1.0
1.6 Æ 1.8
)
87 Æ 25
82 Æ 21
[Gd{(S)-4-eob-dtpa}(H2O)]2Àb
)
23 Æ 0.1
4.6 Æ 0.2
4.6 Æ 0.2
4.0 Æ 0.2
25 Æ 1
[Gd(ttda)(H2O)]2Àc
)
6.3 Æ 2.5
143 Æ 26
57 Æ 4
[Gd(dtpa)(H2O)]2Àb
)
25 Æ 1
58
a
) The values were obtained by H-NMR. b) From [33]. c) From [16].
2
DHM is fixed at 30 kJ/mol [24]. The parameters providing the best fit of the data for
[Gd{(S)-4-Bz-ttda}(H2O)]2À are listed in Table 3. The obtained tOM value of 2.3 ns at
310 K is very close to the optimal range for the attainment of the high relaxivity
expected when the molecular reorientational time of the complex is lengthened to the
nanosecond range [25]. By varying the temperature over a wide range, RO2p is
dominated by 1/tM in the slowkinetic region at lowtemperature and is dominated by 1/
tei in the fast kinetic region at high temperature. The maximum in the profile of 17O-

Helvetica Chimica Acta Vol. 85 (2002)
1043
NMR transverse-relaxation rate (RO2p ) vs. T corresponds to the translation from the
slowto the fast kinetic regions and was not found for [Gd{( S)-4-Bz-ttda}(H2O)]2À,
indicating the fast water-exchange rate for the complex.
As shown in Table 3, the water-exchange lifetime of [Gd{(S)-4-Bz-ttda}(H2O)]2À
(2.3 Æ 1.3 ns) is similar to that of [Gd(ttda)(H2O)]2À (6.3 Æ 2.5 ns) but lower than those
of [Gd{(S)-4-Bz-dtpa}(H2O)]2À (87 Æ 25 ns), [Gd{(S)-4-eob-dtpa)(H2O)]2À (82 Æ
21 ns), and [Gd(dtpa)(H2O)]2À (143 Æ 26 ns) [26]. The smaller water-exchange
lifetime for [Gd{(S)-4-Bz-ttda}(H2O)]2À and [Gd(ttda)(H2O)]2À is perhaps due to
the longer backbone in the ttda5À ligand, which might be pulled tightly into the first
coordination sphere and lead to the highly steric constraint on the water binding site
[27]. Analogous considerations can be made for [Gd(egta)(H2O)]À (ˆaqua[3,12-
bis(carboxymethyl)-6,9-dioxa-3,12-diazatetradecanedioato-(4À)]gadolinate(1À)) [28].
Two coordinating O-atoms in this complex are linked by an ethylenic group, which
induces severe constraints on the atoms around the site occupied by the water
molecule, thus favoring the water-exchange process.
Since the tM value of [Gd{(S)-4-Bz-ttda}(H2O)]2À is similar to that of
[Gd(ttda)(H2O)]2À, the relaxivity r1 of [Gd{(S)-4-Bz-ttda)(H2O)]2À should be similar
to that of [Gd(ttda)(H2O)]2À. However, the r1 of [Gd{(S)-4-Bz-ttda)(H2O)]2À is
significantly higher than that of [Gd(ttda)(H2O)]2À. On the other hand, the tM value of
[Gd(ttda)(H2O)]2À is significantly lower than that of [Gd(dtpa)(H2O)]2À, but both
have similar r1-values. It can be concluded that for small Gd3‡ chelates, the mean
residence lifetime tM does not contribute significantly to the overall correlation time for
the dipolar interaction, except for Gd3‡ complexes of dtpa5À bis(amide).
2
Determination of tR by H-NMR. In diamagnetic molecules, the relaxation rate of
2
the H-nucleus is predominantly determined by the quadrupolar mechanism [29],
which is strictly intramolecular and solely modulated by rotation of the molecule. For
fast-tumbling systems, the relaxation rate R1 is thus directly related to the rotational
correlation time tR (Eqn. 10), where the quadrupolar coupling constant (e2qQ/h)
depends on the hybridization state of the C-atom carrying the 2H-atom; its value is ca.
170 kHz in the case of an sp3 C-atom.
ꢂ
ꢃ2
1
T1
3
8
e2qQ
h
R1 ˆ
ˆ
tR
ꢃ10†
The measurement of tR was performed on the diamagnetic lanthanum(III)
complexes deuterated in the a-position relative to the carboxylate groups. At 310 K,
the values of tR for La3‡ complexes with [(S)-4-Bz-ttda]5À, [(S)-4-Bz-dtpa]5À, [(S)-4-
eob-dtpa]5À, ttda5À, and dtpa5À are shown in Table 3. The tR value of [La{(S)-4-Bz-
(2H10)ttda}]2À (70 Æ 6 ps) is similar to that of [La{(S)-4-Bz-(2H10)dtpa}]2À (65 ps) [26].
Thus, the change of the C-backbone length does not alter tR significantly. On the
other hand, the higher tR values for [La{(S)-4-Bz-(2H10)ttda}]2À, [La{(S)-4-eob-
(2H10)dtpa}]2À, and [La{(S)-4-Bz-(2H10)dtpa}]2À compared to those of [La{(2H10)dtpa}]2À
and [La{(2H10)ttda}]2À indicate that a benzyl or an ethoxybenzyl group at the C-
backbone increases the tR value and causes the higher relaxivity of the Gd3‡ complexes.

1044
Helvetica Chimica Acta Vol. 85 (2002)
Fig. 6. Plot of the water proton relaxation rate of a 0.07 mm [Gd{(S)-4-Bz-ttda}(H2O)]2À solution as a function
of the HSA concentration. pH 7.5, 50 mm phosphate buffer, 20 MHz, 25 Æ 0.18.
Binding of Gadolinium(III) Chelate to Human Serum Albumin (HSA). Binding of
a contrast agent to HSA is highly interesting because the formed macromolecular
adduct compared to the free complex has a higher relaxivity and a longer retention time
in blood, which would qualify it as a contrast agent for angiographic applications [30].
The characterization of the binding parameters was carried out by measuring the
relaxation rates of a [Gd{(S)-4-Bz-ttda)(H2O)]2À solution in the presence of different
concentrations of HSA at 20 MHz and 258 [31], as shown in Fig. 6. Assuming that there
are two binding sites with similar affinity in albumin, as suggested by previous studies,
the value of KA < 1 ¥ 102 mÀ1 for the adduct ([Gd{(S)-4-Bz-ttda](H2O)]2À) ¥ HSA w as
obtained, which is smaller than those of ([Gd{(S)-4-eob-dtpa}(H2O)]2À) ¥ HSA (772 Æ
195mÀ1) [6] and ([Gd(bopta)(H2O)]2À) ¥ HSA (290 Æ 60mÀ1) [5] bearing hydrophobic
substituents in the ligands. However, the value of KA for adducts ([Gd{(S)-4-Bz-
ttda}(H2O)]2À) ¥ HSA, ([Gd{(S)-4-eob-dtpa}(H2O)]2À) ¥ HSA, and ([Gd(bopta)-
(H2O)]2À) ¥ HSA are significantly lower than that of (MS-325) ¥ HSA (3.0 Æ 0.2 ¥
104 mÀ1) [32], which possesses one diphenylcyclohexyl substituent. Thus, only less-
hydrophobic substituents at the surface of the ligand can lead to a smaller association
constant KA.
In conclusion, the linear poly(aminocarboxylate) ligand [(S)-4-Bz-ttda]5À forms a
thermodynamically stable complex with the trivalent gadolinium cation. This complex

Helvetica Chimica Acta Vol. 85 (2002)
1045
does not dissociate under physiological conditions (pH 7.4) and does not exchange with
Ca2‡, Cu2‡, or Zn2‡ to an appreciable extent. From analysis of the 17O-NMR
relaxometric properties, the replacement of the ethane-1,2-diyl backbone with a
propane-1,3-diyl backbone moiety and the introduction of a hydrophobic substituent at
the backbone increase the steric constraints at the water binding site and, therefore,
enhance the water-exchange rate. For small gadolinium complexes like those studied in
this work, tR is the determinant correlation time at high fields and physiological
temperature. An increased tR results, thus, in a higher relaxivity. [Gd{(S)-4-Bz-
ttda}(H2O)]2À possesses a higher relaxivity, a higher thermodynamic stability constant,
an optimal water-exchange rate, and a longer rotational correlation time, which might
result in a novel type of contrast agent for MRI. However, the relatively weak binding
between [Gd{(S)-4-Bz-ttda}(H2O)]2À and HSA reduces its suitability for potential use
as a contrast agent in MR angiography. With the advantages of a short tM and long tR of
[Gd{(S)-4-Bz-ttda}]2À, other ttda5À derivatives and their Gd3‡ complexes are under
study in our laboratory. We expect to obtain a novel contrast agent for MR angiography
by tuning the lipophilicity of complexes with different pendant hydrophobic
substituents at the backbone of the ttda5À ligand.
Experimental Part
1. General. GdCl3 ¥ 6 H2O (99.9%), DyCl3 ¥ 6 H2O (99.9%), and LaCl3 ¥ 7 H2O (99.9%) were obtained from
Aldrich and used without further purification. The concentration of Gd3‡, Dy3‡, and La3‡ were determined by
chelatrometric titration with H4edta, and xylenol orange was used as an indicator. All other reagents used for the
synthesis of the ligand were purchased from commercial sources unless otherwise noted. 1H- and 13C-NMR
spectra and elemental analyses were used to confirm the composition of the products. 17O-Enriched water
(10.5%) was purchased from Isotec Inc. HSA was purchased from Sigma (St. Louis, Mo., USA) and used
without any further purification; the molecular mass was assumed to be 69 kDa [33]. The concentration of the
aq. solns. of the Gd3‡ complexes was obtained from the known millimolar relaxivity of the paramagnetic
complex (RF1p ) by measuring their longitudinal water proton relaxation rates (R1obs) at 20 MHz and 258 (Eqn. 11)
[34], where 0.38 sÀ1 is the relaxation rate of pure water at this temperature and magnetic field strength.
R1obs ˆ RF1p[GdL] ‡ 0.38
(18)
2. Synthesis. l-Phenylalanine Methyl Ester Hydrochloride (3 ¥ HCl). To l-phenylalanine (2; 10.0 g,
60.62 mmol) and anh. MeOH (50 ml), thionyl chloride (7.57 g, 63.61 mmol) was added below 58 (ice bath).
After 30 min, the mixture was brought to r.t. for another 19 h. The soln. was evaporated and the white
precipitated product collected in a B¸chner funnel and dried in vacuo: 3 ¥ HCl (9.3 g, 85.6%). 1H-NMR (D2O):
7.33 7.16 (m, 5 arom. H); 4.31 (t, HÀC(2)); 3.72 (s, MeO); 3.24 3.13 (m, 2 HÀC(3)). 13C-NMR (D2O): 173.1;
136.8; 132.5; 132.3; 131.2; 57.0; 56.5; 38.4. Anal. calc. for C10H13NO2 ¥ HCl (215.68): C 55.69, H 6.45, N 6.49;
found: C 55.32, H 6.54, N 6.35.
N'-(3-Aminopropyl)-l-phenylalaninamide Trihydrochloride (4 ¥ 3 HCl). A soln. of 3 (8 g, 44.63 mmol) in
anh. MeOH (50 ml) was added to propane-1,3-diamine (25 ml) at r.t.. After 15 h at r.t., the mixture was
evaporated. The pale crude yellowoil was purified by cation exchange ( AG 50W Â 8 column (200 400 mesh,
‡
H
form, 100 ml of resin), 4.2-cm column diameter; elution with aq. HCl soln.). The product was eluted with
4.5 5m HCl, while the earlier fractions contained inorg. salts (NaCl) and impurities: 4 ¥ 3 HCl (3.44g, 34.8%).
Amorphous, hygroscopic, yellowpowder. 1H-NMR (D2O): 7.42 7.27 (m, 5 arom. H); 4.18 (t, HÀC(2)); 3.29
3.04 (m, NCH2CH2CH2N); 2.68 (t, 2 HÀC(3)); 1.68 (quint., NCH2CH2CH2NH). 13C-NMR (D2O): 172.2; 137.1;
132.5; 132.2; 131.0; 57.4; 39.8; 39.7; 39.2; 29.0. Anal. calc. for C12H14N3O ¥ 3 HCl (330.68): C 66.65, H 6.53,
N 19.43; found: C 66.41, H 6.40, N 19.25.
N-[(2S)-2-Amino-3-phenylpropyl]propane-1,3-diamine Trihydrochloride (5 ¥ 3 HCl). To
4 (3 g,
13.55 mmol) under N2, anh. THF (50 ml) was added by syringe. The amide only partially dissolved in the

1046
Helvetica Chimica Acta Vol. 85 (2002)
THF. The mixture was cooled to À5 to 08, 1m BH3 ¥ THF (50 ml) added by syringe, and then the mixture
gradually warmed up and brought to reflux for 18 h. Then the soln. was evaporated, the residue dissolved in
EtOH (100 ml) and 6m HCl (10 ml), and the resulting soln. refluxed for 12 h. The soln. was evaporated, the
residue dissolved in dist. H2O, and the pH of the soln. adjusted to 1.5 2.0 with 6m HCl. The crude product was
‡
purified by cation exchange (AG 50WÂ 8 column (200 400 mesh, H form, 100 ml of resin); 4.2-cm column
diameter; elution with aq. HCl soln.). The product was eluted with 5.5 6m HCl: 5 ¥ 3 HCl (2.63 g, 79.9%).
Amorphous, hygroscopic, yellowpowder. 1H-NMR (D2O): 7.43 7.32 (m, 5 arom. H); 3.95 (quint., HÀC(2'));
3.48, 3.35 (m, 2 HÀC(1')); 3.17 (t, HNCH2CH2CH2NH2); 3.06 (t, HNCH2CH2CH2NH2); 3.19 2.98
(m, 2 HÀC(3')); 2.10 2.06 (quint., HNCH2CH2CH2NH2). 13C-NMR (D2O): 133.9; 129.6; 129.5; 128.4; 50.2;
48.9; 45.8; 36.6; 36.5; 23.9. Anal. calc. for C12H16N3 ¥ 3 HCl (316.70): C 71.25, H 7.97, N 20.77; found: C 71.12,
H 7.85, N 20.55.
(4S)-4-Benzyl-3,6,10-tris(carboxymethyl)-3,6,10-triazadodecanedioic Acid (ˆ N-{(2S)-2-[Bis(carboxy-
methyl)amino]-3-phenylpropyl}-N-{3-[bis(carboxymethyl)amino]propyl}glycine; H5[(S)-4-Bz-ttda]; 1). To a
soln. of 5 (2.5 g, 12.05 mmol) and K2CO3 (15 g, 108.5 mmol) in MeCN (250 ml), tert-butyl bromoacetate (8.9 ml,
60.29 mmol) was added. The mixture was heated and refluxed for 24 h. The K2CO3 was removed by filtration
through a B¸chner funnel and washed with MeCN (40 ml). The filtrate was evaporated and the residue taken up
in H2O and extracted with CHCl3 (3 Â 50 ml). The extract was evaporated, the residue dissolved in 2m HCl
(50 ml), and the soln. stirred for 4 h at r.t. and then evaporated. The acid treatment was repeated three times
(1H-NMR: complete removal of tert-butyl group). The residue was dissolved in dist. H2O (50 ml) and the soln.
alkalinized with aq. ammonia to pH 11.0 and submitted to anion exchange (AG1 Â 8 column (200 400 mesh,
HCOÀ2 form, 60 ml of resin); 3.0-cm column diameter; elution with aq. HCO2H soln.). The product was eluted
with 0.7m HCO2H. After evaporation, the residue was coevaporated 5 times with H2O (200 ml) to remove
HCO2H. The residue was dried in vacuo: 1 (2.27 g, 37.89%). Pale yellowhygroscopic powder. 1H-NMR (D2O):
7.33 7.25 (m, 5 arom. H); 3.82 (s, 8 H, terminal NCH2COOH); 3.48 (s, central NCH2COOH); 3.33 3.01
(m, NCH2CH2CH2N, NCHCH2N); 3.05, 2.62 (2q, PhCH2); 2.03 (quint., NCH2CH2CH2). 13C-NMR (D2O):
174.6; 173.0; 172.6; 136.7; 129.1; 128.4; 126.9; 63.0; 59.3; 56.9; 56.4; 54.7; 55.9; 54.9; 38.9; 26.1. Anal. calc. for
C22H31N3O10 (497.50): C 53.11, H 6.28, N 8.44; found: C 52.93, H 6.14, N 8.29.
3. Deuteration. Deuteration of ligands (or of their lanthanum complexes) at the a-position with respect to
the carboxylate groups were performed by the procedure described in [35].
4. Complexation. The La3‡, Dy3‡, and Gd3‡ complexes were prepared by mixing equimolar amounts of
hydrated LnCl3 and ligand. The pH was maintained at 7.5 with 1m NaOH. At r.t., complex formation was
instantaneous. The pH of the soln. was then increased to 8 9 by adding 1m NaOH to precipitate the excess of
uncomplexed Ln3‡ ions. The soln. was then evaporated and the residue dried overnight at 608. The synthesis of
the Gd3‡ complex for the HSA experiment was carried out by the procedure described in [34].
5. Potentiometric Measurements. pH Titration measurements of H5[(S)-Bz-ttda] in the absence and
presence of Gd3‡, Ca2‡, Zn2‡, and Cu2‡ were performed with a Metrohm pH meter equipped with a PT-100
combination pH electrode. The pH electrode was calibrated with three standard buffer solns., and all
calibrations and titrations were carried out under CO2-free N2 (to avoid any contact with CO2) in a glass-
jacketed vessel (20 ml) thermostatted at 25.0 Æ 0.18, and 0.1m ionic strength in Me4NNO3. The autotitration
system consisted of a 702-SM titroprocessor and a 728 stirrer. CO2-Free 0.10m KOH was used as the titrant to
minimize ionic strength change during the titration. N2 was bubbled through the solns. to exclude CO2. Each
titration was performed at least 3 times.
For the potentiometric equilibrium studies, solns. of [(S)-4-Bz-ttda]5À in the absence of metal ion and then
in the presence of each metal ion were used in which the M/L ratio was 1:1. The d values were obtained after
addition of 0.005-ml increments of standard 0.100 mm KOH, and after stabilization from this side of the pH
scale equilibrium was then approached from the other side by adding standard 0.10m acid soln. To determine the
stability constant for [Gd{(S)-4-Bz-ttda}]2À, competition titration with H4edta was performed. A ratio Gd3‡
[(S)-4-Bz-ttda]5À/edta4À of 1:1:1 was employed. Under these conditions, the metal ion was partitioned between
[(S)-4-Bz-ttda]5À and edta4À, and the rate of transmetallation was fast enough between pH 2 4. However,
complexation was usually rapid (3 5 min per point to give a stable pH reading) in the Cu2‡, Ca2‡, and Zn2‡
studies. The same values of the stability constants were obtained either by direct or back titration.
6. Computational Method. The ligand soln. (1 mm) was titrated with KOH over a pH range from 2 to 11. The
titration data was fit to a Fortran computer program PKAS [36] written for polyprotonic weak acid equilibrium.
Equimolar metal/ligand solns. were titrated with KOH over the pH range 2 11, collecting ca. 110 data points
per titration, and the stability constants were determined by analysis of the titration curve with the Fortran
‡
computer program BEST [36]. The average difference between observed and calculated À log [H ] w as<0.04

Helvetica Chimica Acta Vol. 85 (2002)
1047
throughout all titrations. A value of 13.78 was employed for pKw at 258. The species-distribution diagrams were
calculated with the Fortran program SPE and SPEPLOT [36].
7. Relaxation Measurement. Combining equal molar amounts of the stock GdCl3 and the [(S)-4-Bz-ttda]5À
ligand soln. gave the Gd3‡ chelate soln. A slight excess (3%) of the ligand was used, and the soln. was left for at
least 2 h at r.t. to ensure completion of the complexation. The Gd3‡ chelate solns. at various pH values were
prepared by combining the buffer soln. with an approximately diluted Gd3‡ complex soln. in a 1 :1 (v/v) ratio.
The following buffer systems (all 0.10m) were used: chloroacetic acid/NaOH (pH 2 and 3), acetic acid/NaOH
(pH 4 and 5), PIPES (ˆ piperazine-1,4-diethanesulfonic acid)/NaOH (pH 6.0 and 6.8), and ammonia/HCl
(pH 9 and 10) [37]. These buffer solns. were used to maintain constant ionic strength (i.e., 0.10m). The 0.10m
buffers were sufficient to maintain the soln. pH within the desired range. The buffered Gd3‡ chelate solns. were
all allowed to equilibrate for at least 2 h. The pH of these solns. was determined immediately before relaxation
time T1 measurements.
Relaxation times T1 of aq. solns. of [Gd{(S)-4-Bz-ttda}] were measured to determine relaxivity r1. All
measurements were made with a Bruker Minispec-NMS-120 NMR spectrometer operating at 20 MHz as a
function of temp. Before each measurement, the spectrometer was tuned and calibrated. The values of T1 were
measured from eight data points generated by an inversion-recovery pulse sequence. The slopes of plots of 1/T1
vs. the concentration of Gd3‡ complex give r1 in mmÀ1 sÀ1
.
8. EPR Measurements. The EPR spectra were recorded at the X-band (0.34 T) with a Bruker ER-200D-
SRC spectrometer operated in continuous-wave mode, the samples being in 1-mm glass tubes. The cavity temp.
was stabilized by electronic temp. control of the gas flowing through the cavity and verified by substituting a
thermometer for the sample tube. Measurements were made from 273 up to 363 K. The peak-to-peak linewidth
was measured from the recorded spectra by using the instrument software.
9. NMR Measurements. 17O-NMR. The hydration number of [Dy{(S)-4-Bz-ttda}]2À was determined by the
method of Alpoim et al. [38]. The 17O-NMR spectra were recorded by a Varian-Gemini-400 spectrometer at 258.
The induced 17O shift (d.i.s.) measurements were determined with D2O as external standard. An equimolar soln.
of Dy3‡ and ligand was prepared, and a stoichiometric amount of standardized NaOH was added so that the
complex was fully formed. Six solns. of varying Dy3‡ concentrations were prepared by serial dilution of the stock
soln.
For the measurement of the 17O-NMR transverse-relaxation rate, a Varian Gemini-300 (7.05 T, 40.65 MHz)
spectrometer was used, equipped with a 5-mm probe (external D2O lock). Experimental settings were: spectral
width 10000 Hz, pulse width 7 ms, acquisition time 10 ms, no sample spinning. A Varian VT-J103 temp. control
unit was used to stabilize the temp. The value of the transverse relaxation rate was obtained by evaluating the
linewidth at half-height (Dn1/2) of the water 17O signal (R2 ˆ p Dn1/2). Solns. containing 2.6% of the 17O isotope
were used.
2H-NMR. The measurement was carried out in a 10-mm (o.d.) tube on a Varian Gemini-400 (9.4 T)
spectrometer equipped with a broadband probe. No field frequency lock was used.
1
Solution NMR. H-NMR Spectra were measured with a Varian Gemini-400 spectrometer in 5-mm (o.d.)
tubes at 258. The La3‡ complex for 1H-NMR study was prepared by mixing an aliquot of the ligand soln. with an
amount of the appropriate La3‡ soln. to have a 1 :1 molar ratio at pD 7.0 (by using NaOD). The final
concentration of [La{(S)-4-Bz-ttda}]2À was ca. 30 mm. The chemical shifts were referenced to D2O (4.72 ppm).
Data Analysis. The simultaneous least-squares fitting and the binding constant KA were determined by
¾
fitting the experimental data with the program Scientist for WindowsTM by Micromath¾, version 2.0.
Appendix:
Table 1S. pM Values for the Gd2‡, Zn2‡, Ca2‡, and Cu2‡ Complexes at pH 7.4a)
b
d
e
[(S)-4-Bz-ttda]5À
[(S)-4-eob-dtpa]5À
)
Ligand of MS-325 c)
dtpa5À
)
ttda5À
)
pGd
pCu
pCa
pZn
17.97
13.88
10.13
14.04
17.70
17.20
15.31
4.46
17.14
16.06
5.45
17.03
13.56
9.73
5.99
11.83
13.39
13.22
a) All data were obtained in aqueous Me4NNO3 solution, I ˆ 0.1m, except for b) and c). b) From [10] (KCl, I ˆ
0.1m). c) From [4] (Me4NCl, I ˆ 0.1m). d) From [9]. e) From [8].

1048
Helvetica Chimica Acta Vol. 85 (2002)
Fig. 1S. Titration curves for 1:1 ratios of metal ion to [( S)-4-Bz-ttda]5À (ˆ L). m ˆ 0.1m Me4NNO3; Tˆ 25.0 Æ
0.18; concentrations of ligand and metal ion ˆ 1.0 ¥ 10À3 m. a ˆ mol base/mol ligand.
Fig. 2S. pH Dependencies of conditional stability constants for [Gd{(S)-4-Bz-ttda}]3À and [Gd(ttda)]2À

Helvetica Chimica Acta Vol. 85 (2002)
1049
Fig. 3S. Species distribution curves for the 1.0 ¥ 10À3 m [Gd{(S)-4-Bz-ttda}]2À system containing a 1:1molar ratio
of Gd37 to ligand. Tˆ 25.0 Æ 0.18; m ˆ 0.10m Me4NNO3; % ˆ percent relative to 1.0 ¥ 10À3 m total [(S)-4-Bz-
ttda]5À species (ˆ100%).
Fig. 4S. Temperature dependence of transverse electronic relaxation rates at the X-band (0.34 T) and pH 7 for
50 mm [Gd(S)-4-Bz-ttda)(H2O)]2À. The line represents the simultaneous least-squares fit to all data points as
described in the text.

1050
Helvetica Chimica Acta Vol. 85 (2002)
We are grateful to the National Science Council of the Republic of China for financial support under
Contract No. NSC 88-2113-M037-016.
REFERENCES
[1] P. Caravan, J. J. Ellison, T. J. McMurry, R. B. Lauffer, Chem. Rev. 1999, 99, 2293.
[2] S. Aime, M. Botta, M. Fasano, S. G. Crich, E. Terreno, Coord. Chem. Rev. 1999, 185, 321.
[3] M. G. Duarte, M. H. Gil, J. A. Peters, J. M. Colet, L. Vander Elst, R. N. Muller, C. F. G. C. Geraldes,
Bioconjugate Chem. 2001, 12, 170.
[4] P. Caravan, C. Comuzzi, W. Crooks, T. J. McMurry, G. R. Choppin, S. R. Woulfe, Inorg. Chem. 2001, 40,
2170.
[5] F. Uggeri, S. Aime, P. L. Anelli, M. Botta, M. Brocchetta, C. de Hae»n, G. Ermondi, M. Grondi, P. Paoli,
Inorg. Chem. 1995, 34, 633.
[6] L. Vander Elst, F. Chapelle, S. Laurent, R. N. Muller, J. Biol. Inorg. Chem. 2001, 6, 196.
¬
[7] K. Adzamli, E. Toth, M. P. Periasamy, S. H. Koenig, A. E. Merbach, M. D. Adams, Magn. Reson. Mater.
Phys., Biol. Med. 1999, 8, 163.
[8] Y. M. Wang, C. H. Lee, G. C. Liu, R. S. Sheu, J. Chem. Soc., Dalton Trans. 1998, 4113.
[9] W. P. Cacheris, S. C. Quay, S. M. Rocklage, Magn. Reson. Imaging 1990, 8, 467.
[10] H. S. Willich, M. Brehm, C. L. J. Ewers, G. Michel, A. M. Fahrnow, O. Petrov, J. Platzek, B. Rad¸chel, D.
S¸lzle, Inorg. Chem. 1999, 38, 1134.
¬
[11] E. Toth, L. Burai, A. E. Merbach, Coord. Chem. Rev. 2001, 216, 363.
[12] W. R. Harris, K. N. Raymond, F. L. Weitl, J. Am. Chem. Soc. 1981, 103, 1316.
[13] T. H. Cheng, Y. M. Wang, W. T. Lee, G. C. Liu, Polyhedron 2000, 19, 2027.
[14] T. H. Cheng, Y. M. Wang, K. T. Lin, G. C. Liu, J. Chem. Soc., Dalton Trans. 2001, 3357.
[15] T. Kowall, F. Foglia, L. Helm, A. E. Merbach, J. Am. Chem. Soc. 1995, 117, 3790.
[16] C. Cossy, L. Helm, D. H. Powell, A. E. Merbach, New J. Chem. 1995, 19, 27.
[17] C. Cossy, A. C. Barnes, J. E. Enderby, A. E. Merbach, J. Chem. Phys. 1989, 90, 3254.
[18] G. R. Choppin, P. A. Baisden, S. A. Khan, Inorg. Chem. 1979, 18, 1330.
[19] J. A. Peters, Inorg. Chem. 1988, 27, 4686.
[20] T. J. Swift, R. E. Connick, J. Chem. Phys. 1962, 37, 307.
[21] J. Reuben, J. Phys. Chem. 1971, 75, 3164.
[22] S. Aime, M. Botta, S. G. Crich, G. Giovenzana, R. Pagliarin, M. Sisti, E. Terreno, Magn. Reson. Chem. 1998,
36, S200.
[23] S. Aime, M. Botta, M. Fasano, E. Terreno, Acc. Chem. Res. 1999, 32, 941.
[24] S. Aime, E. Gianolio, E. Terreno, G. B. Giovenzana, R. Pagliarin, M. Sisti, G. Palmisano, M. Botta, M. P.
Lowe, D. Parker, J. Biol. Inorg. Chem. 2000, 5, 488.
[25] S. Aime, M. Botta, M. Fasano, E. Terreno, Chem. Soc. Rev. 1998, 27, 19.
¬
¬
[26] S. Laurent, L. Vander Elst, S. Houze, N. Guerit, R. N. Muller, Helv. Chim. Acta 2000, 83, 394.
[27] K. Micskei, D. H. Powell, L. Helm, E. Br¸cher, A. E. Merbach, Magn. Reson. Chem. 1993, 31, 1011.
[28] S. Aime, A. Barge, A. Borel, M. Botta, S. Chemerisov, A. E. Merbach, U. M¸ller, D. Pubanz, Inorg. Chem.
1997, 36, 5104.
[29] T. K. Hitchens, R. G. Bryant, J. Phys. Chem. 1995, 99, 5612.
[30] S. Aime, M. Botta, L. Frullano, S. G. Crich, G. B. Giovenzana, R. Pagliarin, G. Palmisano, M. Sisti, Chem.
Eur. J. 1999, 5, 1253.
[31] S. Aime, M. Botta, M. Fasano, S. G. Crich, E. Terreno, J. Biol. Inorg. Chem. 1996, 1, 312.
[32] S. Aime, M. Chiaussa, G. Digilio, E. Gianolio, E. Terreno, J. Biol. Inorg. Chem. 1999, 4, 766.
[33] Jr. T. Peters, Adv. Protein Chem. 1986, 37, 161.
[34] S. Aime, M. Botta, G. Ermondi, F. Fedeli, F. Uggeri, Inorg. Chem. 1992, 31, 1100.
[35] W. D. Wheeler, J. J. Legg, Inorg. Chem. 1985, 24, 1292.
[36] A. E. Martell, R. J. Motekaitis, −Determination and Use of Stability Constants×, 2nd edn., VCH, NewYork,
1992.
[37] C. A. Chang, H. G. Brittain, J. Telser, M. F. Tweedle, Inorg. Chem. 1990, 29, 4468.
[38] M. C. Alpoim, A. M. Urbano, C. F. G. C. Geraldes, J. A. Peters, J. Chem. Soc., Dalton Trans. 1992, 463.
Received November 12, 2001