Synthesis and relaxometric properties of gadolinium(III) complexes of new triazine-based polydentate ligands

Article abstract of DOI:10.1002/hlca.200900189

Two new derivatives based on an s-triazine structural motif were synthesized by attaching two 2,2′-hydrazinylidenebis[acetic acid] moieties to the triazine ring to reach an overall heptadenticity for the complexation of lanthanide(III) cations. The remain

Full text of DOI:10.1002/hlca.200900189

2414
Helvetica Chimica Acta – Vol. 92 (2009)
Synthesis and Relaxometric Properties of Gadolinium(III) Complexes of New
Triazine-Based Polydentate Ligands
by Lorenzo Teia), Marina Benzia), Filip Kielara), Mauro Botta*a), Camilla Cavallottib),
Giovanni Battista Giovenzanac), and Silvio Aime*d)
a
`
) Dipartimento di Scienze dellꢀAmbiente e della Vita, Universita degli Studi del Piemonte Orientale
ꢁA. Avogadroꢀ, Viale T. Michel 11, I-15100 Alessandria (phone: þ 39-0131360253;
fax: þ 39-0131360250; e-mail: mauro.botta@mfn.unipmn.it)
b) CAGE Chemicals srl, Via Quarello 11/a, I-10135 Torino
c
`
) DiSCAFF & DFB Center, Universita degli Studi del Piemonte Orientale ꢁA. Avogadroꢀ, Via Bovio 6,
I-28100 Novara
d
`
) Dipartimento di Chimica IFM and Center of Molecular Imaging, Universita degli Studi di Torino,
Via Nizza 52, I-10126 Torino (phone: þ 39-0116706451; fax: þ 39-0116707855;
e-mail: silvio.aime@unito.it)
Dedicated to Professor Jean-Claude Bꢀnzli on the occasion of his 65th birthday
Two new derivatives based on an s-triazine structural motif were synthesized by attaching two 2,2’-
hydrazinylidenebis[acetic acid] moieties to the triazine ring to reach an overall heptadenticity for the
complexation of lanthanide(III) cations. The remaining reactive site was exploited for the substitution
with a functionizable amino group (see H4L1) and a lipophilic moiety (see H4L2). Luminescence-lifetime
determinations revealed the presence of a single H2O molecule coordinated for [Eu(L1)]. A complete
1H-NMR relaxometric study was carried out for the octacoordinated [Gd(L1)] and [Gd(L2)] complexes.
A remarkably long H2O residence lifetime (298tm ¼ 5.2 ms) was found by 17O-NMR in the case of
[Gd(L1)]. Micelle formation of the lipophilic complex [Gd(L2)] was evidenced, the critical micellization
concentration (cmc) determined, and relaxometric properties of the system investigated.
Introduction. – Chemical research in the field of contrast agents (CAs) has
paralleled the rapid success of magnetic resonance imaging (MRI) as the predominant
diagnostic technique in modern medicine [1 – 3]. To enhance the information content of
an image and to shorten the time of acquisition, the use of CAs is often needed; in
particular, the efforts have focused on GdIII complexes because of the metal-related
favorable features, among which i) the seven unpaired electrons, ii) the long electronic
relaxation time, and iii) the high magnetic moment [4]. The evaluation of a CA is first
carried out ꢁin vitroꢀ by assessing its relaxivity (r1p), defined as the relaxation
enhancement of the H2O H-atoms in a solution containing the paramagnetic agent at a
1 mm concentration. Current commercial CAs consist of low-molecular-mass GdIII
complexes with polyaminopolycarboxylate chelates characterized by relaxivities of 4 –
5 mmꢀ1 sꢀ1 (20 MHz and 298 K) and a single coordinated H2O molecule (q ¼ 1).
According to theory, the increase of the hydration state and/or H2O exchange rate and
slowing the tumbling of the complex can lead to high values of relaxivity. In fact, much
ꢂ 2009 Verlag Helvetica Chimica Acta AG, Zꢃrich

Helvetica Chimica Acta – Vol. 92 (2009)
2415
work has been done in the last decades to pursue such goals and thus to obtain optimal
CAs [5].
The synthesis of ligands capable of forming GdIII complexes with two (or more)
H2O molecules bound in the first (inner) coordination sphere has been an intriguing
challenge as testified by the number of different approaches presented in the literature
[6 – 11]. For this purpose, both acyclic and macrocyclic polyaminocarboxylate ligands
have been thoroughly monitored to obtain GdIII complexes with q ¼ 2 without
compromising thermodynamic and kinetic stabilities.
In this context, we designed a new series of ligands based on the common s-triazine
(¼1,3,5-triazine) heterocyclic ring which is a well-known and tough structural motif for
the construction of a large variety of interesting molecules, showing an outstanding
ability in intermolecular interactions [12]. Cyanuric chloride (i.e., 2,4,6-trichloro-1,3,5-
triazine) is the cheap and versatile starting material for the straightforward preparation
of highly structured molecules, as sequential reactivity of the three halogen atoms is
easily controlled by the experimental conditions employed. Although both acyclic and
cyclic pyridine-containing ligands have been reported for lanthanide ions [10][11], no
examples with other heterocyclic rings involved in the coordination of the LnIII ions are
reported. We were then prompted to explore the possibility to employ the triazine ring
as coordinating unit, relying on its structural versatility to implant auxiliary
coordinating groups and tailor-made recognition moieties. The introduction of
additional remote functional groups would allow conjugation of the corresponding
complexes to macromolecules or biological targeting vectors. Two 2,2’-hydrazinylide-
nebis[acetic acid] moieties complete the overall heptadenticity of the ligand, to obtain a
stable coordinating environment for lanthanide ions. Finally, the tritopic nature of s-
triazine offers a third additional position where it is possible to locate a suitably
designed R group; in the ligands H4L1 and H4L2 reported here, the variable R group is
an amino group and a lipophilic amino group, respectively.
Results and Discussion. – Synthesis. The first step of the synthesis of H4L1 and H4L2
was accomplished, starting from cyanuric chloride (1), by reaction with tritylamine (2a)
and octadecylamine(2b), respectively (Scheme). Hydrazine was then attached in both
4 and 6 positions of the triazinamines 3 (!4) followed by tetraalkylation with tert-butyl
bromoacetate (! 5). Deprotection of the tert-butyl esters (and trityl group for H4L1)
by acid hydrolysis with CF3COOH gave rise to H4L1 and H4L2 in 24 and 32% overall
yield respectively, starting from cyanuric chloride.

2416
Helvetica Chimica Acta – Vol. 92 (2009)
Scheme
i
i) Na2CO3, Et2O, 0 – 58, then r.t. ii) NH2NH2 · H2O, MeOH, r.t. or reflux. iii) BrCH2COOtBu, Pr2NEt,
MeCN/toluene, 70 – 758. iv) BrCH2COOtBu, K2CO3, MeCN, reflux. v) CF3COOH/PhOMe, r.t.
Hydration-State Determination. The prepared novel free ligands H4L1 and H4L2 are
heptadentate, and it was expected that they will give rise to lanthanide complexes with
two inner-sphere H2O molecules (q ¼ 2). The most widely used method of assessing the
hydration number of lanthanide complexes for a given ligand is based on the lifetime
measurement of its respective Eu complex. Eu-Complex [Eu(L1)] has, therefore, been
prepared to determine the hydration number q for the complexes of L1. The
luminescence lifetimes measured in H2O and D2O are 0.67 and 1.92 ms, respectively.
The hydration number qEu for an Eu complex is calculated with Eqn. 1, where tH2O and
tD2O are the luminescence lifetimes in H2O and D2O, respectively, and the 0.25 term is
the contribution of the second-sphere H2O molecules [13]. The latter term is an
estimate from previous empirical measurements [13]. Substituting the measured
lifetime values in Eqn. 1 gives a q value of 0.87 ꢁ 0.06 for [Eu(L1)]. Based on this
observation it should be assumed that the complex [Gd(L1)] is a monohydrated (q ¼ 1)
octadentate species, which is an unexpected result for this heptadentate ligand.
qEu ¼ 1.2 (1/tH O ꢀ 1/tD O ꢀ 0.25)
(1)
2
2
Relaxometric Characterization. [Gd(L1)]. At 20 MHz and 298 K, the value of the
water-proton relaxivity r1p of [Gd(L1)] was measured to be 3.7 mmꢀ1 sꢀ1, a value well
below the expected range (4.5 – 6 mmꢀ1 sꢀ1) for a Gd-complex endowed with one
coordinated H2O molecules (q ¼ 1) [2][3]. This suggests the occurrence of a limitation
of the relaxivity by a slow rate of H2O exchange. This hypothesis is corroborated by the
fact that the relaxivity at 310 K (3.9 mmꢀ1 sꢀ1) is higher than at 298 K. The pH
dependency of r1p is shown in Fig. 1 and provides some further useful information. The
relaxivity is constant in the pH range 4 to 9, whereas it increases both at low pH and in
the basic region. The observed behavior is reminiscent of a similar trend found in the

Helvetica Chimica Acta – Vol. 92 (2009)
2417
case of cationic GdIII complexes with tetraamide derivatives of H4DOTA (¼1,4,7,10-
tetraazacyclododecane-1,4,7-tetraacetic acid) which was accounted for in terms of a
contribution of an acid- and base-catalyzed prototropic exchange to the overall H2O
exchange [14]. The prototropic contribution increases the effective H2O exchange rate,
kex, and then increases the relaxivity. Based on these observations, we can predict that
[Gd(L1)] is characterized by a tm value (tm ¼ 1/kex is the mean residence lifetime of the
coordinated H2O molecule) of the order of microseconds.
*
Fig. 1. Plot of H-atom relaxivity r1p at 20 MHz and 298 K vs. pH for [Gd(L1)] ( ) and [Gd(L2)] (
)
$
The above qualitative considerations can find support in the determination and
analysis of the magnetic-field dependence of the relaxivity, the so-called nuclear
magnetic relaxation dispersion (NMRD) profile and in the temperature dependence of
the 17O-NMR transverse relaxation rate that provides direct information on the kinetic
parameters of the H2O exchange [15]. The NMRD profiles, measured at 298 and 310 K
and at neutral pH are characterized by the typical shape and amplitude of monoaquo,
low-molecular-mass complexes, featuring a single dispersion around 4 – 6 MHz (Fig. 2).
The relaxivity is particularly low in the constant region at low fields, and it is higher at
310 than at 298 K over the entire range of the H-atom Larmor frequencies (0.01 –
70 MHz) investigated. Both these features are consistent with a slow rate of H2O
exchange. It is well known that in general, the NMRD profiles do not allow extracting
accurate information on the parameters that characterize the H2O exchange. These are
conveniently evaluated through the association of variable-temperature 17O-NMR
study in a fairly concentrated solution of the paramagnetic complex. The R2p data,
measured at 9.4 T in an 10.9 mm aqueous solution of [Gd(L1)], show an increase when
the temperature increases in the range 298 – 340 K (Fig. 3). The 17O-NMR data were
analyzed in terms of the Swift – Connick equations [15][16], whereas the NMRD

2418
Helvetica Chimica Acta – Vol. 92 (2009)
Fig. 2. 1H-NMRD Relaxation data (relaxivity r1p vs. 1H Larmor frequency) for [Gd(L1)] at pH 7.2 and
~
*
298 K ( ) and 310 K ( ). The solid lines represent the best results of the fitting to the experimental
points (see Table).
Fig. 3. Plot of the paramagnetic contribution to the 17O transverse relaxation rate R2p as a function of
temperature at 9.4 T (12.2 MHz) and neutral pH for a 10.9 mm aqueous solution of [Gd(L1)]

Helvetica Chimica Acta – Vol. 92 (2009)
2419
profiles were fitted to the standard equations for the inner- (IS) and outer-sphere (OS)-
relaxation contributions by fixing the values of the following parameters: the hydration
number (q ¼ 1), the Gd – Hw distance (r¼ 3.0 ꢄ), the distance of closest approach of
the bulk H2O molecules (a ¼ 4.0 ꢄ) to the metal ion, and the relative diffusion
coefficient (D ¼ 2.4 · 10ꢀ5 cm2 sꢀ1) for solute and solvent [1 – 3]. The best fit parameters
are listed in the Table and clearly confirm the occurrence of a remarkably long
residence lifetime of the bound H2O molecule that represents a limitation to the
relaxation efficacy of the complex: 298tm ¼ 5.2 ms (298kex ¼ (1.9 ꢁ 0.1) · 105 sꢀ1). Whereas
the other relaxation parameters are in the range of values typical of acyclic, low-
molecular-mass GdIII complexes, the H2O-exchange rate of [Gd(L1)] is surprisingly
slow, assuming a value which is observed for the first time in the case of an anionic,
octacoordinate GdIII complex for which an associative mechanism of exchange should
be expected. Clearly, to find an explanation for this result, a careful elucidation of the
exchange mechanism and structural information are needed. Based on the available
data, we can only make the hypothesis of the presence of H-bonding interactions
involving the bound H2O molecule and the proximate NH groups, mediated by second-
sphere H2O molecules. This interaction could stabilize the bound H2O molecule (as
indicated by the high value of the enthalpy of activation, DH# ¼ 63.6 kJ/mol), and
render more difficult the approach of a new H2O molecule.
Table. Selected Relaxation Parameters for [Gd(L1)] and [Gd(L2)] Derived from the Analysis of 17O-
NMR Dataa), 1H-Relaxivity as a Function of Temperature (20 MHz), and Magnetic-Field Strength
(NMRD)b)
Parameter
[Gd(L1)]
[Gd(L2)]
310
r
k
[mmꢀ1 sꢀ1
]
4.0
44.8
1p
298
[106; sꢀ1
]
0.19 ꢁ 0.01
63.6 ꢁ 2.8
7.8 ꢁ 0.3
15 ꢁ 1
74.6 ꢁ 1.4
17 ꢁ 1
–
3.4 ꢁ 0.3
35.0 ꢁ 3.1
0.64 ꢁ 0.03
45 ꢁ 3
ex
DHm [kJ molꢀ1
]
D2 [1019; sꢀ2
]
298tV [ps]
298
t
[ps]
8200 ꢁ 365
18 ꢁ 1
Rg
DHRg [kJ molꢀ1
]
298
t
[ps]
1334 ꢁ 108
16 ꢁ 1
Rl
DHRl [kJ molꢀ1
S2
]
–
–
0.49 ꢁ 0.01
b
a) [Gd(L1)]. ) For the parameters q, r, a, DHV, and 298D, the values of 1, 3.0 ꢄ, 4.0 ꢄ, 2.0 kJ molꢀ1, and
2.24 · 10ꢀ5 cm2 sꢀ1 were used, respectively.
The temperature dependence of the H-atom relaxivity was also measured at
20 MHz, and the data are shown in Fig. 4. The curve through the experimental data is
the result of the analysis carried out as detailed in [17] and by using the parameters of
the Table. The r1p increases by lowering T from 335 to ca. 315 K, then it flattens and
begins to decrease below ca. 300 K. A minimum is reached around 285 K, and then the
relaxivity increases again with decreasing temperature. The overall behavior reflects
the slow rate of exchange of the bound H2O molecule and is the consequence of the
opposite temperature dependence of the IS and OS contributions to the r1p of
[Gd(L1)], as shown in Fig. 4 by the curves calculated with the parameters of the Table.

2420
Helvetica Chimica Acta – Vol. 92 (2009)
Fig. 4. Temperature dependence of the 1H relaxivity r1p for [Gd(L1)] at 20 MHz and pH 7.2. The solid line
&
through the data points ( ) represents the best result of the fitting to the experimental data (see Table);
the calculated inner- and outer sphere contributions to the relaxivity are also represented.
[Gd(L2)]. The ability of the lipophilic analogue [Gd(L2)] to self-assemble into
micelles was assessed by measuring the observed relaxation rate vs. concentration at
20 MHz and 298 K. Above the critical micellar concentration (cmc), the relaxation rate
is the sum of two contributions, one arising from the monomeric complex (present at
the concentration given by the cmc) and the second due to the aggregated form, i.e., the
micelles. Consequently, the H2O H-atom relaxation rate of this paramagnetic system
can be expressed as in Eqn. 2 [18], where rn1:a: and r1a are the relaxivities of the
nonaggregated and aggregated forms, respectively, Rd1 is the relaxation rate of pure
H2O (0.38 sꢀ1 at 20 MHz and 298 K), and C is the analytical concentration of GdIII.
Clearly, the relaxivity of the free complex, rn1:a:, is simply expressed by Eqn. 3. By
plotting Ro1bs ꢀ R1d as a function of the parameter C, r1n:a:, and r1a can be obtained by
fitting Eqns. 2 and 3, as shown in Fig. 5.
Ro1bs ꢀ Rd1 ¼ (r1n:a: ꢀ r1a)cmc þ r1a C
(2)
(3)
Ro1bs ꢀ Rd1 ¼ rn1:a:
C
At 20 MHz and 298 K, the relaxivity of [Gd(L2)] is 20.2 ꢁ 0.9 mmꢀ1 sꢀ1 below the
cmc, whereas the relaxivity of the aggregated form is 36.7 ꢁ 0.3 mmꢀ1 sꢀ1, and the cmc is
0.056 ꢁ 0.002 mm. A scrupulous study of the micellization and relaxometric properties
was previously reported for a series of [Gd(DOTA)] derivatives substituted with
aliphatic chains of increasing length [19]. As expected, the cmc decreases following the
lengthening of the chain from C10 to C18, and for the latter, a cmc of 0.06 mm was found,

Helvetica Chimica Acta – Vol. 92 (2009)
2421
Fig. 5. Plot of the H2O 1H longitudinal relaxation rate at 20 MHz and 298 K as a function the total GdIII
concentration for [Gd(L2)] and least-squares fit according to Eqn. 2
in perfect agreement with the value found for [Gd(L2)], a complex functionalized with
the same pendant aliphatic chain. On the other hand, the ra1 value for [Gd(L2)] is
sensibly higher than those found in the case of other (monoaquo) gadolinium(III)
complexes bearing a ligand with an aliphatic chain of comparable length (ra1p ꢂ 17 –
22 mmꢀ1 sꢀ1). To investigate this and other aspects, NMRD profiles were recorded in
the frequency range 0.01 to 70 MHz at 298 (Fig. 6), 310, and 318 K. The profiles present
small differences in their amplitudes but share a general shape with a broad hump
centered around 20 MHz, typical of slowly tumbling macromolecular systems.
Furthermore, the relaxivity values increase on passing from 298 to 318 K, indicating
the occurrence of a relatively slow rate of H2O exchange that represent a limiting factor
to the relaxivity.
The analysis of the NMRD profiles requires a different approach since the presence
of a relatively fast local rotation of the complex about the main axis of the aliphatic
chain superimposed to the global motion of the micellar aggregate needs to be taken
into account. This is possible by incorporating the description of the rotational
dynamics according to the model-free Lipari – Szabo approach into the Solomon –
Bloembergen – Morgan equations for the IS relaxation mechanism [20]. This model
allows separating the contribution of the overall global rotation of the paramagnetic
micelle (tRg) from the contribution of a faster local motion (tRl) due to the free rotation
of the coordination cage about the aliphatic chain. The correlation of the two types of
motions is described by the parameter S2 whose value is comprised between zero
(completely independent motions) and one (totally correlated motions). A least-
square fit of the data was performed at 298 K (Fig. 6) by fixing the values of q, r, a, and

2422
Helvetica Chimica Acta – Vol. 92 (2009)
Fig. 6. 1H-NMRD profile (relaxivity r1p vs. 1H Larmor frequency) of [Gd(L2)] at pH 7.2 and 298 K at a
concentration above the cmc (0.3 mm). The solid curve through the data points is calculated with the
parameters reported in the Table.
D and treating as variable parameters D2, tV, tm , tRg, tRl, and S2. The best-fit parameters
are listed in the Table and indicate that a significantly lower rotational dynamics of the
micellar system together with a faster rate of H2O exchange are responsible of the high
relaxivity values. In fact, the global reorientational correlation time is from three to
four times longer than for corresponding complexes in the aggregated form based on
the DOTA chelating unit, and this suggests a larger dimension of the micelles formed
by [Gd(L2)]. On the other hand, the value of the order parameter, S2 ¼ 0.49, is rather
close to those of the related systems, indicating a similar degree of rotational flexibility
about the aliphatic chain.
Since the residence lifetime of the coordinated H2O molecule is only roughly
estimated from the analysis of the NMRD profiles, we decided to obtain an
independent evaluation of tm from the temperature dependence of the H-atom
relaxivity. The data were obtained in the temperature interval 274 – 333 K and are
reported in Fig. 7. The relaxivity increases with temperature up to ca. 320 K where a
broad hump is formed, and then it starts to decrease above 330 K. Clearly, this behavior
appears to be dictated by the contribution of the IS relaxivity that is severely limited by
the slow H2O-exchange rate and then reproduces its temperature dependence. In the
fitting procedure, the relaxation parameters were fixed to the value obtained from the
refinement of the NMRD profiles, with the exception of their enthalpy of activation
and 298tm that were let to vary. A good fit of the data is obtained with a 298tm value of
0.3 ms. This value is in very good agreement with that estimated from NMRD data, and
it is significantly shorter (one order of magnitude) than that found for [Gd(L1)].

Helvetica Chimica Acta – Vol. 92 (2009)
2423
Finally, it is worth observing that this result is also in agreement with the pH
dependency of the relaxivity (Fig. 1). Unlike the behavior observed for [Gd(L1)], in
the case of [Gd(L2)], the increase of the relaxation rate with pH in the basic region is
sensibly attenuated, as a result of a shorter value of the H2O residence lifetime.
1
Fig. 7. Variable-temperature H-NMR relaxation data (relaxivity r1p) for [Gd(L2)] at 20 MHz and at a
concentration above the cmc (0.25 mm), normalized to 1 mm concentration of Gd. The solid line through
*
the data points ( ) represents the best result of the fitting to the experimental data (see the Table); the
calculated inner- and outer-sphere contributions to the relaxivity are also represented.
Conclusions. – Two novel, potentially heptadentate triazine-containing amino-
carboxylic ligands were synthesized and characterized. Unexpectedly, the correspond-
ing GdIII complexes possess a single H2O molecule in their inner coordination sphere.
Thus, [Gd(L1)] and [Gd(L2)] represent rare examples of monoaquo octacoordinate
GdIII complexes in aqueous media. The exchange rate of the coordinated H2O molecule
is remarkably low, another unusual property for GdIII chelates characterized by an
octacoordinate ground state. Presumably, both electronic and structural effects are
responsible for these findings that need further studies for their elucidation. In
addition, for [Gd(L2)], only differing from [Gd(L1)] by the presence of a C18 aliphatic
chain in place of an amino H-atom, the H2O residence lifetime is significantly reduced.
This lipophilic complex forms large micellar aggregates at low concentration which
show high relaxivity values in spite of the presence of a high degree of flexibility due to
local rotation about the aliphatic chain.
The authors thank Dr. Claudio Cassino (Alessandria) for help with the 17O-NMR-data collection.
Financial support from Regione Piemonte (ꢁConverging Technologiesꢀ 2007, NanoIGT project) is
gratefully acknowledged. This work was carried out under the auspices of CIRCMSB and COST Action
D38.

2424
Helvetica Chimica Acta – Vol. 92 (2009)
Experimental Part
General. Unless otherwise noted, starting materials were obtained from commercial suppliers and
used without further purification. CC ¼ Column chromatography.
4,6-Dichloro-N-(triphenylmethyl)-1,3,5-triazin-2-amine (3a). The 2,4,6-trichloro-1,3,5-triazine
(2.84 g, 15.4 mmol) and Na2CO3 (1.96 g, 18.5 mmol) were suspended in Et2O (27 ml). To the stirred
and cooled (ice-bath) suspension, tritylamine (4.00 g, 15.4 mmol) was added portionwise, then the ice-
bath was removed, and the mixture was stirred at r.t. for 20 h. The mixture was filtered and the solid
washed with CH2Cl2. The combined filtrate and washings were concentrated, and the residue was
recrystallized from AcOEt/petroleum ether: pure 3a (4.00 g, 64%). White solid. M.p. 293 – 2948.
1H-NMR (300 MHz, CDCl3, 298 K): 7.34 – 7.22 (m, 15 H); 7.13 (br. s, 1 H). 13C-NMR (75.4 MHz, CDCl3,
298 K): 170.0 (s); 169.6 (s); 165.7 (s); 143.2 (s); 128.8 (d); 128.2 (d); 127.5 (d); 71.9 (s). ESI-MS: 417.2
([M þ Na]þ), 419.2 ([M þ H]þ); calc. for C22H16Cl2N4, 416.1 and 418.1.
4,6-Dihydrazinyl-N-(triphenylmethyl)-1,3,5-triazin-2-amine (4a). A soln. of 3a (2.50 g, 6.14 mmol) in
NH2NH2 · H2O (3.88 ml, 80 mmol) and MeOH (15 ml) was stirred at r.t. for 48 h. The mixture was then
cooled and filtered, and the solid residue was washed thoroughly with aq. MeOH and subsequently dried
1
in vacuo: 4a (2.13 g, 87%). White solid. M.p. 252 – 2548. H-NMR (300 MHz, (D6)DMSO, 298 K): 7.52
(br. s, 2 H); 7.38 – 7.12 (m, 15 H); 6.88 (br. s, 1 H); 3.18 (br. s, 4 H). 13C-NMR (75.4 MHz, (D6)DMSO,
363 K): 167.7 (s); 165.9 (s); 146.3 (s); 129.2 (d); 128.0 (d); 126.8 (d); 70.4 (s). ESI-MS: 399.3 ([M þ H]þ),
243.5 (Ph3Cþ); calc. for C22H22N8, 398.2.
Tetra(tert-butyl) 2,2’,2’’,2’’’-{{6-[(Triphenylmethyl)amino]-1,3,5-triazine-2,4-diyl}dihydrazin-2-yl-1-
ylidene}tetrakis[acetate] (5a). Compound 4a (2.00 g, 5.03 mmol) and powdered anh. K2CO3 (4.16 g,
30.2 mmol) were suspended in MeCN (15 ml). tert-Butyl bromoacetate (3.70 ml, 25.1 mmol) was added
dropwise, and then the suspension was refluxed for 6 h. Inorg. salts were removed by filtration, and the
filtrate was concentrated. The residue was then purified by CC (AcOEt/petroleum ether 3 :7 ! 5 :5):
1
pure 5a (2.40 g, 57%). Off-white solid. M.p. 203 – 2058. H-NMR (300 MHz, CDCl3, 298 K): 7.32 – 7.12
(m, 17 H); 3.72 (br. s, 8 H); 1.42 (s, 36 H). 13C-NMR (75.4 MHz, CDCl3, 298 K): 169.6 (s); 166.4 (s);
165.9 (s); 145.5 (s); 129.1 (d); 127.7 (d); 126.8 (d); 81.7 (s); 70.5 (s); 57.8 (t); 28.2 (q). ESI-MS: 855.3
t
([M þ H]þ), 634.2 ([M þ Na ꢀ Ph3C]þ), 612.4 ([M þ H ꢀ Ph3C]þ), 578.5 ([M þ Na ꢀ Ph3C ꢀ Bu]þ),
243.5 (Ph3Cþ); calc. for C46H62N8O8, 854.5.
2,2’,2’’,2’’’-[(6-Amino-1,3,5-triazine-2,4-diyl)dihydrazin-2-yl-1-ylidene]tetrakis[acetic Acid] (H4L1).
The ester 5a (500 mg, 0.585 mmol) was dissolved in a mixture of CF3COOH (20 ml) and anisole (5 ml)
and stirred at r.t. overnight. The waxy residue obtained after vacuum evaporation was dissolved in
MeOH (2 ml), and excess Et2O was added dropwise leading to the precipitation of an off-white solid. The
latter was washed thoroughly with Et2O, then dried in vacuo to constant weight: H4L1 (172 mg, 76%).
Amorphous white solid. 1H-NMR (300 MHz, D2O þ K2CO3, 298 K): 3.53 (br. s, 8 H). 13C-NMR
(75.4 MHz, D2O þ K2CO3, 298 K): 177.9 (s); 166.0 (s); 165.1 (s); 59.2 (t). ESI-MS: 411.2 ([M þ Na]þ),
389.2 ([M þ H]þ); calc. for C11H16N8O8, 388.1.
4,6-Dichloro-N-octadecyl-1,3,5-triazin-2-amine (3b). As described for 3a, with 2,4,6-trichloro-1,3,5-
triazine (2.84 g, 15.4 mmol), Na2CO3 (1.96 g, 18.5 mmol), Et2O (50 ml), and octadecylamine (4.16 g,
15.4 mmol; addition within 30 min); at r.t. overnight. After washing with hot AcOEt and evaporation, the
residue was submitted to flash chromatography (AcOEt/petroleum ether 5 :95): 3b (5.60 g, 87%). White
solid. M.p. 293 – 2948. 1H-NMR (300 MHz, CDCl3, 298 K): 6.36 (t, J ¼ 5.3, 1 H); 3.47 (q, J ¼ 6.7, 2 H); 1.59
(quint., J ¼ 7.1, 2 H); 1.30 – 1.18 (m, 30 H); 0.86 (br. t, J ¼ 6.6, 3 H). 13C-NMR (75.4 MHz, CDCl3, 298 K):
171.2 (s); 169.7 (s); 165.9 (s); 41.7 (t); 32.0 (t); 29.8 (8t); 29.7 (t); 29.6 (t); 29.5 (t); 29.4 (t); 29.2 (t); 29.1
(t); 26.7 (t); 22.8 (t); 14.2 (q). ESI-MS: 429.1 and 431.1 ([M þ Na]þ), 417.3 and 419.3 ([M þ H]þ); calc.
for C21H38Cl2N4, 416.2 and 418.3.
4,6-Dihydrazinyl-N-octadecyl-1,3,5-triazin-2-amine (4b). As described for 4a, with 3b (1.10 g,
2.64 mmol), NH2NH2 · H2O (6.0 ml, 124 mmol) and MeOH (10 ml); at 808 for 4 h. Washing with H2O: 4b
1
(770 mg, 72%). White solid. M.p. 152 – 1548. H-NMR (300 MHz, (D6)DMSO, 298 K): 6.32 – 6.19 (m,
2 H); 3.38 (m, 4 H); 2.96 (br. t, J ¼ 7.5, 2 H); 1.46 (br. quint., 2 H); 1.32 – 1.15 (m, 30 H); 0.84 (br. t, J ¼
6.5, 3 H). 13C-NMR (75.4 MHz, (D6)DMSO, 298 K): 168.9 (s); 165.8 (s); 46.2 (t); 31.8 (t); 29.7 (9t); 29.8
(2t); 29.6 (t); 28.2 (t); 27.2 (t); 22.8 (t); 13.9 (q). ESI-MS: 409.5 ([M þ H]þ); calc. for C21H44N8, 408.4.

Helvetica Chimica Acta – Vol. 92 (2009)
2425
Tetra(tert-butyl) 2,2’,2’’,2’’’-{[6-(Octadecylamino)-1,3,5-triazine-2,4-diyl]dihydrazin-2-yl-1-ylidene}-
tetrakis[acetate] (5b). To a suspension of 4b (1.50 g, 3.68 mmol) in a mixture of MeCN (10 ml) and
i
toluene (30 ml), Pr2NEt (4.45 ml, 25.7 mmol) was added, followed by dropwise addition of tert-butyl
bromoacetate (2.71 ml, 18.40 mmol). The milky suspension was stirred and heated to 70 – 758 for 3.5 h.
Volatiles were evaporated, then the residue was partitioned between CH2Cl2 and 10% aq. Na2CO3 soln.
The org. extract was dried (Na2SO4) and evaporated and the residue purified by CC (AcOEt/petroleum
ether 2 :8 ! 3 :7): pure 5b (1.96 g, 62%). Light yellow viscous oil. 1H-NMR (300 MHz, CDCl3, 298 K):
6.94 (br. s, 2 H); 3.71 (s, 8 H); 3.28 (m, 2 H); 1.43 (m, 2 H); 1.39 (s, 36 H); 1.21 (m, 30 H); 0.83 (br. t, J ¼
7.1, 3 H ) . 13C-NMR (75.4 MHz, CDCl3, 298 K): 169.1 (s); 166.6 (s); 166.2 (s); 81.4 (s); 60.9 (t); 57.9 (t);
40.8 (t); 31.9 (t); 29.6 (10t); 29.4 (t); 29.3 (t); 28.1 (t); 27.6 (t); 26.9 (t); 22.6 (t); 14.0 (q). ESI-MS: 887.3
([M þ Na]þ), 865.5 ([M þ H]þ); calc. for C45H84N8O8, 864.6.
2,2’,2’,2’’-{[6-(Octadecylamino)-1,3,5-triazine-2,4-diyl]dihydrazin-2-yl-1-ylidene}tetrakis[acetic acid]
(H4L2). As described for H4L1, with 5b (610 mg, 0.705 mmol), CF3COOH (10 ml), and anisole (2 ml):
H4L2 (371 mg, 82%). Off-white amorphous solid. 1H-NMR (300 MHz, CD3OD, 323 K): 3.90 (m, 8 H);
3.34 (t, J ¼ 7.1, 2 H); 1.54 (br. quint., 2 H); 1.36 – 1.22 (m, 30 H); 0.89 (t, J ¼ 6.7, 3 H). 13C-NMR
(75.4 MHz, CD3OD, 323 K): 171.6 (s); 164.1 (s); 157.3 (s); 156.6 (s); 57.2 (t); 41.0 (t); 31.8 (t); 29.4 (12t);
29.1 (t); 29.0 (t); 26.6 (t); 22.4 (t); 13.1 (q). ESI-MS: 663.3 ([M þ Na]þ), 641.3 ([M þ H]þ); calc. for
C29H52N8O8, 640.4.
GdIII Complexes. The GdIII complexes were prepared by mixing stoichiometric amounts of the
chloride salt to an aq. soln. of the corresponding ligand and monitoring the formation of the complex by
measurement of the solvent H-atom relaxation rate (1/T1) while keeping a constant neutral pH with 1m
NaOH. The final solns. were checked with the orange xylenol test [21] and accepted for characterization
when free [GdIII] < 1%.
EuIII Complexes. The EuIII complexes were prepared by mixing 10% excess of the chloride salt to an
aq. soln. of the corresponding ligand, keeping a constant neutral pH with 1m NaOH. The excess of metal
ions was removed by precipitation as hydroxide at basic pH followed by centrifugation.
Luminescence Lifetime Measurements. The luminescence lifetime measurements were carried out on
a Fluorolog fluorometer with a photomultiplier tube, electronics, and software supplied by Horiba Jobin
Yvon. Horiba Jobin Yvon Spectraled source was used to directly excite the EuIII ion at 395 nm, as spectral
measurements have not shown any significant sensitization of the Eu emission by the triazine ring. The
lifetime measurement was carried out with the DataStation v2.5 software. The subsequent data analysis
was performed with the DAS6 v6.4 software.
1H-NMR Relaxometric Measurements. The H2O H-atom longitudinal relaxation rates of dil. aq. solns.
of complexes [Gd(L1)] and [Gd(L2)] were measured with a Stelar-Spinmaster spectrometer (Mede,
Italy) operating at 0.5 T and 298 K. For the measurement of the relaxation rates, the standard inversion-
recovery method was employed (16 experiments, 4 scans) with a typical 908 pulse width of 3.5 ms, and the
reproducibility of the T1 data was ꢁ 0.5%. The temp. was controlled with a Stelar-VTC-91 air-flow heater
equipped with a copper-constantan thermocouple (uncertainty ꢁ 0.18). The H-atom 1/T1 NMRD
profiles were measured on a fast field-cycling Stelar-Spinmaster-FFC relaxometer over a continuum of
magnetic-field strengths from 0.00024 to 0.5 T (corresponding to 0.01 – 20 MHz H-atom Larmor
frequencies). The relaxometer operates under computer control with an absolute uncertainty in 1/T1 of
ꢁ 1%. Additional data points at 20 – 70 MHz were obtained on a Stelar relaxometer.
17O-NMR. Variable-temp. 17O-NMR measurements were recorded on a Jeol-ECP-400 (9.4 T)
spectrometer equipped with a 5 mm probe and a standard temp.-control unit. Aq. solns. of the complexes
containing 2.8% of the 17O isotope (Cambridge Isotope) were used. The observed transverse relaxation
rates (Ro2bs ) were calculated from the signal width at half-height (Dn1/2): Ro2bs ¼ p ꢃ Dn1/2
.
REFERENCES
[1] ꢁThe Chemistry of Contrast Agents in Medical Magnetic Resonance Imagingꢀ, Eds. E. Toth and
A. E. Merbach, Wiley, New York, 2001.
[2] S. Aime, M. Botta, E. Terreno, Adv. Inorg. Chem. 2005, 57, 173.

2426
Helvetica Chimica Acta – Vol. 92 (2009)
[3] P. Caravan, J. J. Ellison, T. J. McMurry, R. B. Lauffer, Chem. Rev. 1999, 99, 2293.
[4] H. J. Weinmann, A. Mꢃhler, B. Radꢃchel, ꢁBiomedical Magnetic Resonance Imaging and
Spectroscopyꢀ, Ed. I. R. Young, Wiley, Chichester, 2001, p. 705.
[5] P. Caravan, Chem. Soc. Rev. 2006, 35, 512.
[6] J. Xu, S. J. Franklin Jr., D. W. Whisenhunt, K. N. Raymond, J. Am. Chem. Soc. 1995, 117, 7245; S.
Hajela, M. Botta, S. Giraudo, J. Xu, K. N. Raymond, S. Aime, J. Am. Chem. Soc. 2000, 122, 11228;
D. M. J. Doble, M. Botta, J. Wang, S. Aime, A. Barge, K. N. Raymond, J. Am. Chem. Soc. 2001, 123,
10758; E. J. Werner, A. Datta, C. J. Jocher, K. N. Raymond, Angew. Chem., Int. Ed. 2008, 47, 8568.
[7] D. Messeri, M. P. Lowe, D. Parker, M. Botta, Chem. Commun. 2001, 2742; S. Aime, L. Calabi, C.
Cavallotti, E. Gianolio, G. B. Giovenzana, P. Losi, A. Maiocchi, G. Palmisano, M. Sisti, Inorg. Chem.
2004, 43, 7588; G. B. Giovenzana, G. Palmisano, M. Sisti, C. Cavallotti, S. Aime, L. Calabi (to Bracco
Imaging S.p.A.), PCT Int. Appl. 2003, WO0308390; Z. Baranyai, F. Uggeri, G. B. Giovenzana, A.
Benyei, E. Brucher, S. Aime, Chem. – Eur. J. 2009, 15, 1696.
[8] R. Ruloff, R. N. Muller, D. Pubanz, A. E. Merbach, Inorg. Chim. Acta 1998, 275 – 276, 15.
[9] J. Costa, E. Toth, L. Helm, A. E. Merbach, Inorg. Chem. 2005, 44, 4747; J. B. Livramento, L. Helm,
A. Sour, C. OꢀNeil, A. E. Merbach, E. Toth, Dalton Trans. 2008, 1195; R. Ruloff, G. van Koten, A. E.
Merbach, Chem. Commun. 2004, 842.
[10] V.-M. Mukkala, C. Sund, M. Kwiatkowski, P. Pasanen, M. Hogberg, J. Kankare, H. Takalo, Helv.
Chim. Acta 1992, 75, 1621; L. Pellegatti, J. Zhang, B. Drahos, S. Villette, F. Suzenet, G. Guillaumet, S.
Petoud, E. Toth, Chem. Commun. 2008, 6591; S. Laurent, L. Vander Elst, M. Wautier, C. Galaup,
R. N. Muller, C. Picard, Bioorg. Med. Chem. Lett. 2007, 17, 6230; J. Ketola, J. Katajisto, H. Hakala, J.
Hovinen, Helv. Chim. Acta 2007, 90, 607.
[11] S. Aime, M. Botta, S. Geninatti Crich, G. B. Giovenzana, G. Jommi, R. Pagliarin, M. Sisti, Inorg.
Chem. 1997, 36, 2992; S. Aime, E. Gianolio, D. Corpillo, C. Cavallotti, G. Palmisano, M. Sisti, G. B.
Giovenzana, R. Pagliarin, Helv. Chim. Acta 2003, 86, 615; S. Aime, C. Cavallotti, E. Gianolio, G. B.
Giovenzana, G. Palmisano, M. Sisti, Org. Lett. 2004, 6, 1201; S. Aime, M. Botta, L. Frullano, S.
Geninatti Crich, G. B. Giovenzana, R. Pagliarin, G. Palmisano, F. Riccardi Sirtori, M. Sisti, J. Med.
Chem. 2000, 43, 4017.
[12] T. J. Mooibroek, P. Gamez, Inorg. Chim. Acta. 2007, 360, 381.
[13] A. Beeby, I. M. Clarkson, R. S. Dickins, S. Faulkner, D. Parker, L. Royle, A. S. de Sousa, J. A. G.
Williams, M. Woods, J. Chem. Soc., Perkin Trans. 1999, 2, 493.
[14] S. Aime, A. Barge, M. Botta, D. Parker, A. S. de Sousa, J. Am. Chem. Soc. 1997, 119, 4767; S. Aime,
A. Barge, J. I. Bruce, M. Botta, J. A. K. Howard, J. M. Moloney, D. Parker, A. S. de Sousa, M.
Woods, J. Am. Chem. Soc. 1999, 121, 5762
[15] S. Aime, M. Botta, M. Fasano, E. Terreno, Acc. Chem. Res. 1999, 32, 941; L. Helm, G. M. Nicolle,
A. E. Merbach, Adv. Inorg. Chem. 2005, 57, 327.
[16] T. J. Swift, R. E. Connick, J. Chem. Phys. 1962, 37, 307.
[17] L. Thompson, D. Parker, D. A. Fulton, J. A. K. Howard, S. U. Pandya, H. Puschmann, K.
Senanayake, P. A. Stenson, A. Badari, M. Botta, S. Avedano, S. Aime, Dalton Trans. 2006, 5605.
[18] G. M. Nicolle, E. Toth, K.-P. Eisenwiener, H. R. Macke, A. E. Merbach, J. Biol. Inorg. Chem. 2002, 7,
757.
´
´
´
[19] S. Torres, J. A. Martins, J. P. Andre, C. F. G. C. Geraldes, A. E. Merbach, E. Toth, Chem. – Eur. J.
2006, 12, 940.
[20] a) G. Lipari, A. Szabo, J. Am. Chem. Soc. 1982, 104, 4546 – 4559; b) G. Lipari, A. Szabo, J. Am.
Chem. Soc. 1982, 104, 4559.
[21] A. Barge, G. Cravotto, E. Gianolio, F. Fedeli, Contrast Media Mol. Imaging 2006, 1, 184.
Received May 20, 2009