Designing novel contrast agents for magnetic resonance imaging. Synthesis and relaxometric characterization of three gadolinium(III) complexes based on functionalized pyridine-containing macrocyclic ligands

Article abstract of DOI:10.1002/hlca.200390061

The three novel pyridine-containing 12-membered macrocyclic ligands 1-3 were synthesized. The coordinating arms are represented by three acetate moieties in 1 and 3 and by one acetate and two phosphonate moieties in 2. In all three ligands, the acetate arm in the distal position is substituted, with a benzyl group in 1 and 2 and with an arylmethyl moiety in 3. The relaxivities r_1p (20 MHz, 25°) of Gd^III complexes are: GD · 1, r_1p = 8.3 mm^-1 s^-1; GD · 2, r_1p = 8.1 mM^-1 s^-1; Gd · 3, r_1p = 10.5 mM^-1 s^-1. ¹H-NMRD and ¹⁷O-NMR T₂ data show that Gd · 1 and Gd · 3 contain two H₂O molecules in the inner sphere, whereas the presence of two phosphonate arms allows the coordination of only one H₂O molecule in Gd · 2. Interestingly, the exchange lifetime of coordinated H₂O in the three complexes is similar in spite of the difference in the coordination number of the Gd^III ion (i.e., 9 in Gd · 1 and Gd · 3, and 8 in Gd · 2). ¹H-Relaxometric measurements at different pH and in the presence of lactate and oxalate were carried out to get some insight into the formation of ternary complexes from Gd · 1 and Gd · 3. Finally, it was found that binding to human-serum albumin (HSA) of Gd · 1 and Gd · 2, though weak, yields limited relaxivity enhancements, likely as a consequence of effects on the hydration sphere caused by donor atoms on the surface of the protein.

Full text of DOI:10.1002/hlca.200390061

Helvetica Chimica Acta Vol. 86 (2003)
615
Designing Novel Contrast Agents for Magnetic Resonance Imaging.
Synthesis and Relaxometric Characterization of three Gadolinium(III)
Complexes Based on Functionalized Pyridine-Containing Macrocyclic
Ligands
by Silvio Aime*, Eliana Gianolio, and Davide Corpillo
¡
Dipartimento di Chimica I.F.M., Universita degli Studi di Torino, Via P. Giuria 7, I-10125 Torino
and
Camilla Cavallotti, Giovanni Palmisano*, and Massimo Sisti
¡
Dipartimento di Scienze Chimiche Fisiche e Matematiche, Universita degli Studi dell×Insubria, Via Valleggio 11,
I-22100 Como
and
Giovanni B. Giovenzana
¡
Dipartimento di Scienze Chimiche Alimentari Farmaceutiche e Farmacologiche, Universita degli Studi del
Piemonte Orientale ™A. Avogadro∫, Via Bovio 6, I-28100 Novara
and
Roberto Pagliarin
¡
Dipartimento di Chimica Organica e Industriale, Universita degli Studi di Milano, Via Venezian 21, I-20133 Milano
The three novel pyridine-containing 12-membered macrocyclic ligands 1 3 were synthesized. The
coordinating arms are represented by three acetate moieties in 1 and 3 and by one acetate and two phosphonate
moieties in 2. In all three ligands, the acetate arm in the distal position is substituted, with a benzyl group in 1 and
2 and with an arylmethyl moiety in 3. The relaxivities r1p (20 MHz, 258) of GdIII complexes are: GD ¥ 1, r1p
ˆ
8.3 mmÀ1 sÀ1; GD ¥ 2, r1p ˆ 8.1 mmÀ1 sÀ1; Gd ¥ 3, r1p ˆ 10.5 mmÀ1 sÀ1. H-NMRD and 17O-NMR T2 data show that
Gd ¥ 1 and Gd ¥ 3 contain two H2O molecules in the inner sphere, whereas the presence of two phosphonate arms
allows the coordination of only one H2O molecule in Gd ¥ 2. Interestingly, the exchange lifetime of coordinated
H2O in the three complexes is similar in spite of the difference in the coordination number of the GdIII ion (i.e., 9
in Gd ¥ 1 and Gd ¥ 3, and 8 in Gd ¥ 2). 1H-Relaxometric measurements at different pH and in the presence of
lactate and oxalate were carried out to get some insight into the formation of ternary complexes from Gd ¥ 1 and
Gd ¥ 3. Finally, it was found that binding to human-serum albumin (HSA) of Gd ¥ 1 and Gd ¥ 2, though weak,
yields limited relaxivity enhancements, likely as a consequence of effects on the hydration sphere caused by
donor atoms on the surface of the protein.
1
Introduction.
It is well established that the intrinsic contrast in magnetic-
resonance images (MRI) can be augmented by the use of contrast agents, which are
chemicals able to markedly enhance H2O protons relaxation rates [1][2]. Most
applications of contrast agents has dealt with the use of GdIII complexes because this
metal ion couples a large magnetic moment with a long electronic-spin relaxation time.
This metal does not possess any physiological function in mammals, and its
administration as a free ion is extremely toxic even at low doses (10 20 mmol/kg).
For this reason, it is necessary to use ligands that form very stable chelates with GdIII
ions [3][4].

616
Helvetica Chimica Acta Vol. 86 (2003)
The efficacy of a GdIII chelate as MRI contrast agent is usually assessed in vitro by
measuring its relaxivity, which represents the relaxation enhancement of H2O protons
promoted by the paramagnetic complex at 1 mm concentration. At the magnetic-field
strength of clinical tomographs, high relaxivities can be obtained when the GdIII chelate
is part of a slowly moving macromolecular system [4]. However, to fully exploit the
relaxation enhancement brought about by the long re-orientation time tR, the GdIII
complex must display a short exchange lifetime tM of the coordinated H2O molecule(s)
[4 6]. Often, tM is too long and acts as a limiting factor for the attainment of the high
relaxivities expected for slowly moving systems. Thus, the first step towards the search
for high relaxivities is the identification of a GdIII chelate displaying the best values for
the hydration number q and for tM, because the amplification effect associated with the
formation of a macromolecular adduct strictly depends on the relaxometric properties
of the free complex.
GdIII Chelates with q ˆ 2 have often been addressed, as these may represent a good
compromise between high hydration and sufficiently high thermodynamic stability.
Moreover, the presence of two H2O molecules in the inner coordination sphere is
expected to endow such systems with a faster H2O exchange than that found with
complexes of octacoordinating ligands [7][8]. With regard to this, [Gd(DO3A)]
(Fig. 1) and several of its functionalized derivatives have been investigated in depth [9].
The main drawback found with this structural type lies in a relaxivity −quenching× effect
upon interacting with proteins, as donor atoms of Asp or Glu residues may replace the
coordinated H2O molecules. It has been suggested that the occurrence of such inter-
actions can be anticipated by investigating the formation of ternary complexes between
a given complex and substrates like lactate, malonate, or, simply, carbonate [10].
Fig. 1. Three classes of heptadentate ligands: DO3A, HOPO, and PC ligands
Whereas [Gd(DO3A)]-like complexes show quite strong interactions with the latter
substrates, it has been found that two other classes of heptacoordinated GdIII chelates
(q ˆ 2) do not. The first class is represented by [Gd(HOPO)] complexes developed by
Raymond and co-workers [11 13]. The second class is represented by GdIII complexes
with ligands based on a 12-membered N4 pyridine-containing macrocycle (PC-type ligands)
in which the three amino N-atoms bear acetic or phosphonic arms (Fig. 1) [14 16].
Among the GdIII complexes with PC-type ligands, we have recently identified
[Gd(PCP2A)] (PCP2A ˆ pyridine-containing macrocycle (PC) bearing two acetate
(A) and one methylenephosphonate (P) arms) as a very promising contrast agent for
improved applications in MRI [17]. It shows a very high stability constant (logKF ˆ
23.4) and a relaxivity ca. two times higher than the values reported for contrast agents

Helvetica Chimica Acta Vol. 86 (2003)
617
currently used in clinical practice. Moreover, [Gd(PCP2A)] displays a fast exchange of
the coordinated H2O molecules.
The results now presented deal with GdIII complexes of modified PC-type ligands
and are aimed at assessing the effects on the observed relaxivity arising from the
presence of an aromatic substituent at the distal arm and from the replacement of
acetate for phosphonate arms at the two equivalent N-atoms. The structures of the
three novel PC-type ligands 1 3 synthesized in this work are shown in Fig. 2.
Fig. 2. PC-Type ligands 1
3
Results and Discussion. Synthesis of Ligands 1 3 (Scheme 1). The strategy we
envisioned for the synthesis of 1 3 was essentially based on the nucleophilic ring
opening of 1-tosylaziridines by the appropriate a-amino esters to afford centrally N-
functionalized diethylenetriamines. This approach proved to be highly efficient in the
synthesis of diethylenetriamine derivatives, and it is well documented in the literature
[17]. As already reported by us [15], these compounds, where the tosyl moieties act as
protecting-activating groups, are reactive nucleophiles allowing ring cyclization with
2,6-bis(halomethyl)pyridine under appropriate experimental conditions.
Thus, the commercially available a-amino acids l-phenylalanine (4) and l-tyrosine
(5) were transformed into the corresponding methyl esters by treatment with thionyl
chloride in MeOH at 08. The reactivity of the two ester derivatives 6 and 7 towards 1-
tosylaziridine (obtained by standard methods from −N,O-ditosylethanolamine× ˆ 2-
{[(4-methylphenyl)sulfonyl]amino}ethyl 4-methylbenzenesulfonate) proved to be
similar: the ring opening of 1-tosylaziridine (2.2 equiv./equiv. ester) took place in
refluxing toluene affording the two functionalized diethylenetriamines 8 and 9 in
excellent yields (90%) after silica-gel chromatography.
The evaluation of these centrally N-functionalized derivatives in macrocyclization
reactions was first performed with compound 8. By reacting 8 with 2,6-bis(chlorome-
thyl)pyridine in refluxing anhydrous MeCN in the presence of anhydrous K2CO3, the
macrocycle 12 was obtained in 65% yield after crystallization. The experimental
conditions we adopted here represent a modification of the RichmanÀAtkins protocol.
We already observed [15] that higher yield in PCTA (pyridine-containing macrocycle
triacetate) derivatives may be obtained by performing the macrocyclization step under
heterogeneous conditions as reported above. Indeed, under RichmanÀAtkins conditions

618
Helvetica Chimica Acta Vol. 86 (2003)
Scheme. Synthesis of Ligands 1
3
i) MeOH, SOCl2, 08. ii) 1-Tosylaziridine, toluene, reflux. iii) TsCl, Py, 08. iv). tBuMe2SiCl, iPr2EtN, MeCN, 08. v)
2,6-Bis(chloromethyl)pyridine, anh. K2CO3, MeCN, reflux. vi) 48% aq. HBr, PhOH, AcOH, reflux. vii)
ClCH2COOH, KOH, H2O, 808. viii) 30% HBr in AcOH, 608. ix) H3PO3, (CH2O)n, 6m aq. HCl, reflux. x)
BrCH2COOMe, Ag2CO3, MeCN, r.t. xi) KOH, MeOH, 608

Helvetica Chimica Acta Vol. 86 (2003)
619
(NaH/DMF as base/solvent system), compound 12 was isolated in less than 30% yield.
Removal of the protecting-activating tosyl groups and hydrolysis of the methyl ester
moiety to give the macrocyclic polyamino acid 15 was accomplished by refluxing 12
with 48% aqueous HBr in AcOH in the presence of phenol as tosyl scavenger.
The macrocyclic amino acid 15 represents the key intermediate for the preparation
of ligands 1 and 2. Thus, alkylation of 15 with chloroacetic acid in the presence of KOH
at 808 at pH ca. 10, followed by treatment with conc. HCl solution at room temperature
(pH 0) and chromatography (XAD 1600) gave pure 1 in 68% yield. The introduction of
the two phosphonomethyl substituents at the two NH moieties of 15 was accomplished
by reaction with H3PO3 and CH2O in 6m HCl under reflux, affording pure 2 in 61%
yield after precipitation by EtOH.
The reaction of diethylenetriamino 9 with 2,6-bis(chloromethyl)pyridine in
refluxing MeCN in the presence of K2CO3 afforded polymeric materials as the only
products. Likely, the lability of the phenolic OH group towards oxidation is responsible
for the failure of the macrocyclization step. However, upon protecting the OH group of
9 as tosyl ( ! 10) or (tert-butyl)dimethylsilyl ether derivatives ( ! 11), we were able to
observe the formation of the desired macrocycles 13 (60%) and 14 (62%), respectively,
in the reaction with 2,6-bis(chloromethyl)pyridine under heterogeneous conditions.
Treatment of 13 with aqueous HBr solution, AcOH, and phenol removed the N-tosyl
groups while the phenolic tosylate moiety remained unaffected ( ! 16). However,
when compound 16 was subjected to alkaline hydrolysis (KOH in H2O, MeOH, glyme,
PEG), no reaction was observed at room temperature, and, at higher temperature,
extensive decomposition took place. The removal of the protecting groups of 14 proved
more successful. By reacting 14 with anhydrous 33% HBr in AcOH, compound 17 was
obtained in almost quantitative yield. Surprisingly, the methyl ester moiety remained
unaffected under these conditions. The use of 48% aqueous HBr, which worked well
with 12, afforded lower yields of 17. Attempted alkylation of 17 with chloroacetic acid
in H2O (pH 10) produced only a complex reaction mixture in which the desired product
could not be identified. The ease of oxidation of the phenolic OH group could be
responsible for the failure of this reaction. No success was met with also for the reaction
of 17 with tert-butyl bromoacetate in various solvents (MeCN, DMF) and in the
i
presence of different bases (K2CO3, Et3N, Pr2EtN). However, by reacting 17 with
methyl bromoacetate in the presence of Ag2CO3 and in the dark, compound 18 was
obtained in 58% yield. NMR and MS analyses confirmed the structure of 18, which
likely derives from a Kolbe-like carboxylation side reaction, followed by normal
peralkylation. Finally, alkaline hydrolysis of 18 (KOH/MeOH, reflux) afforded ligand 3.
Synthesis and Relaxometric Characterization of Gadolinium(III) Complexes with
Ligands 1 3. The GdIII complexes of ligands 1 3 were synthesized by adding
stoichiometric amounts of GdCl3 to aqueous solutions of the ligands while maintaining
neutral pH with 1m NaOH. Formation of the complex was followed by measuring the
H2O proton relaxation rate on a relaxometer operating at 20 MHz.
The relaxivities r1p of Gd ¥ 1, Gd ¥ 2, and Gd ¥ 3, measured at 20 MHz and 258, are 8.3,
8.1, and 10.5 mmÀ1 sÀ1 respectively. In Fig. 3, a plot of the relaxivities of these complexes
and of previously reported ones based on heptadentate ligands, as a function of the
molecular mass is reported. The relaxivities of the complexes Gd ¥ 1, Gd ¥ 2, and Gd ¥ 3
are in the range of expected values for similarly sized complexes with q ˆ 2.

620
Helvetica Chimica Acta Vol. 86 (2003)
Fig. 3. Relaxivities (measured at 0.47 T, pH 7, and 258) of some GdIII complexes with q ˆ 2 as a function of their
molecular mass
More insight into the parameters involved in paramagnetic relaxation pathways and
into the effective hydration state were obtained by measuring the 1/T1 NMRD profiles
over an extended observation frequencies range at 258 (Fig. 4,a c). Experimental data
were fitted considering that the observed relaxivity is given by the sum of four
contributions shown in Eqn. 1 [18], where R1d represents the diamagnetic contribution
is
of pure H2O protons (0.38 sÀ1), R is given by the contribution of the inner-sphere H2O
1p
2nd
molecule(s) directly coordinated to the metal ion, R arises from the second-sphere
1p
os
H2O molecules, and R deals with the contribution of H2O molecules that diffuse in the
1p
proximity of the paramagnetic complex. The latter contribution, for small GdIII
complexes, may be considered to be a constant value, i.e., ca. 2.3 2.5 mmÀ1 sÀ1 at 258
and 20 MHz [5].
Ro1bs ˆ R1d ‡ R ‡ R
‡ R
(1)
is
2nd
1p
os
1p
1p
The inner- and the second-sphere relaxivities may be evaluated on the basis of the set
of equations derived from the Solomon Bloembergen Morgan (SBM) theory
(Eqns. 2 4) [19][20], where q is the number of inner- or second-sphere H2O
molecules, T1M is the longitudinal relaxation time of their protons, r is the GdÀH
distance, tR is the molecular re-orientation time, tM is the coordinated-H2O exchange
lifetime, and tS is the electronic relaxation time.
q
R1p ˆ
(2)
55:5ꢀT1M ‡ tM†
K
r6
À1
(T1M
)
1
f(tC)
(3)
(4)
(tC)À1 ˆ (tS)À1 ‡ (tM)À1 ‡ (tR)À1

Helvetica Chimica Acta Vol. 86 (2003)
621
Fig. 4. 1/T1 NMRD profiles of a 1 mm solution of a) Gd ¥ 1, b) Gd ¥ 2, and c) Gd ¥ 3 at pH 7 and 258. The solid
curves through the data points were calculated by means of the parameters reported in the Table.

622
Helvetica Chimica Acta Vol. 86 (2003)
The best fit of the experimental data was obtained by considering two coordinated
H2O molecules for Gd ¥ 1 and Gd ¥ 3 and only one inner-sphere H2O molecule for Gd ¥ 2.
The relatively high relaxivity featuring Gd ¥ 2 is due, besides the inner-sphere
contribution, to the presence of two second-sphere H2O molecules at a distance of
3.8 ä from the GdIII ion. The occurrence of two second-sphere H2O molecules was
surmised on the basis of a comparison with the previously reported results on the
parent chelate [Gd(PCP2A)(H2O)2] [17]. In the latter complex, the observed R1p value
(8.3 mmÀ1 sÀ1) was justified on the basis of the presence of two inner-sphere and one
second-sphere H2O molecules. Passing from [Gd(PCP2A)(H2O)2] to Gd ¥ 2, one may
expect that the presence of two phosphonate groups yields either a decrease of the
number of directly coordinated H2O molecules (from two to one) and an increase of
the number of H2O molecules in the second coordination sphere (from one to two).
The results from the best-fit procedure were fully consistent with this expectation. The
relaxometric parameters determined by fitting the NMRD experimental data to
Eqns. 1 4 for the three complexes (see Table) are similar to those reported for related
complexes derived from PCTA-like ligands.
Table. Best-Fit Parameters Obtained from the Analysis of the NMRD and 17O-NMR Profiles Considering a
Metal H Distance of 3.1 ä for the Inner-Sphere H2O Molecules and 3.8 ä for the Second-Sphere Water, and a
Metal O Distance of 2.5 ä. D2 ˆ mean-square zero-field energy, tv ˆ correlation time for the modulation of the
zero-field-splitting interaction, tR ˆ molecular re-orientation time; tM ˆ exchange lifetime of coordinated H2O;
q ˆ hydration number of inner sphere; q2nd ˆ hydration number of second sphere.
D2 [1019 sÀ2
]
tV [ps]
tR [ps]
tM [ns]
q
q2nd
Gd ¥ 1
Gd ¥ 2
Gd ¥ 3
Gd ¥ 3 (pH 10)
1.98
3.7
2.2
36
27
27
39
92
120
134
130
58
75
79
2
1
2
1
0
2
0
0
1.02
180
It is interesting to note that while the substitution of one coordinating acetate arm
(A) with a phosphonate one (P), as in the case of [Gd(PCP2A)] [17], does not cause
reduction of the hydration state, the hindrance caused by the introduction of two
phosphonate groups (see Gd ¥ 2) leads to the formation of a complex with q ˆ 1 as in the
case of [Gd(PCTP)] [12] which contains three phosphonate moieties [9]. Thus, the
progressive replacement of the acetate arms with phosphonate ones causes a reduction
in hydration state of the metal ion analogously to what was observed on going from
DOTA (ˆ DO3A; see Fig. 1) to DOTP LnIII complexes [21].
To determine the exchange lifetime of the coordinated H2O molecules in the three
GdIII complexes, we measured their 17O-NMR transverse relaxation time as a function
of temperature (Fig. 5,a c) [22]. According to the Swift-Connick theory [23], the
paramagnetic contribution (RO2p ) to the observed transversal relaxation rate is given by
Eqn. 5, and the temperature dependence of DwOM is described by Eqn. 6, where B0 is the
magnetic-field strength, and A/ꢀh is the Gd,17O scalar coupling constant (whose value
for (polyaminocarboxylato)gadolinium(III) complexes has been fixed to À3.8 ¥ 106
rad sÀ1) [18]. For relatively small-sized GdIII chelates and at 2.1 T, R2OM is essentially
dominated by the electron-nucleus scalar interaction (see Eqns. 7 and 8), and, finally,
the temperature dependence of RO2p is expressed in terms of the Eyring relationship for

Helvetica Chimica Acta Vol. 86 (2003)
623
tM and tn according to Eqn. 9, where j refers to the two different dynamic processes
involved (j ˆ n, M) and DHj is the corresponding activation enthalpy.
O2
2M
O
2
À1
qC
R
‡ tM
R
‡ DwOM
2M
RO2p
ˆ
tÀM1
(5)
(6)
À
Á2
2
RO2M ‡ tÀM1 ‡DwOM
55:6
gemBSꢀS ‡ 1†B0 A
DwOM
ˆ
3kBT
ꢀ
ꢀh
ꢀ ꢁ2
ꢁ
1
3
A
tE2
1‡w2s t2E2
RO2M
ˆ
SꢀS ‡ 1† tE1
‡
(7)
(8)
ꢀh
ˆ TÀ1 ‡ tMÀ1
À1
t
Ei
iE
À
Á
ꢂ
ꢀ
ꢁꢃ
tjÀ1 298:15T
298:15
À ÁÀ1
DHj
R
1
1
T
tj
ˆ
exp
À
(9)
T
298:15
In Fig. 5,a c, the variable-temperature 17O-NMR profiles of the three complexes
are reported; all of them show the typical behavior of fast-exchange conditions, since
R2p steadily decreases as the temperature increases. The exchange lifetimes tM
determined at 298 K (Table) are very similar. This is a surprising result, as we would
have expected different tM values for Gd ¥ 1 and Gd ¥ 2 because the difference in their
coordination number might suggest that two different exchange mechanisms are
operating. In fact, the GdIII ion in complex 1 has a coordination number 9; therefore,
the mechanism operating the exchange of the coordinated H2O molecules should be a
dissociative one. On the other hand, the coordination number of the GdIII ion in
complex 2 is 8, and we expect the occurrence of an associative mechanism. Although it
has been assessed that the H2O-exchange rate may depend upon several factors, it
was early established that the very fast exchange observed for octacoordinate
complexes (such as the aqua ion [Gd(H2O)8]3‡ or [Gd(PDTA)] occurs by an
associative mechanism involving a nonacoordinate transition state [24][25][9]. Likely,
the enthalpy of the GdÀOH2 bond is the rate-determining parameter for both
complexes.
The coordination geometry in Gd ¥ 2 is a square antiprism whose eight coordination
sites are taken by the four N-atoms (lower N4 plane) and by two phosphonate O-atoms,
one acetate O-atom, and one H2O O-atom (upper O4 plane) [3]. Each phosphonate
group keeps one H2O molecule in the second coordination sphere. One may envisage a
possible exchange pathway that, after the dissociation of the coordinated H2O, delivers
the second-sphere H2O into the inner coordination sphere. In principle, such a
mechanism might have some analogy with that operating in Gd ¥ 1. One may envisage
that the rate-determining step in the exchange path occurring in the latter complex is
the dissociation of the H2O molecule in the O4 plane, whose vacancy is then occupied
by the H2O that is in the capping position.
Formation of Ternary Complexes. The study of the pH effect on the relaxivity of a
GdIII complex may provide several relevant information. In the acid limb, increase of r1
may report the protonation of ligand functionalities, which may change the hydration

624
Helvetica Chimica Acta Vol. 86 (2003)
Fig. 5. Temperature dependence of the transverse water relaxation rate (R2p) for a 20 mm solution of a) Gd ¥ 1
(2.1 T, pH 7), b) Gd ¥ 2 (7 T, pH 7), and c) Gd ¥ 3 (2.1 T, pH 7)

Helvetica Chimica Acta Vol. 86 (2003)
625
state or even may cause the release of GdIII ions. In the basic limb, a decrease in
relaxivity may be an indication of the formation of ternary complexes with the
carbonate anion [26][9]. The latter condition is often met with heptacoordinate
complexes, and it is a clue to the possible formation of ternary complexes in the
presence of other O-donor substrates [27].
In Fig. 6, the plots of r1 values as function of the solution pH for the three complexes
are reported. All three complexes are very stable at acid pH, as expected on the basis of
very high Kf values previously reported for [Gd(PCTA)] [16] and [Gd(PCP2A)] [17].
At the basic limb, only Gd ¥ 2 maintains a constant r1p, whereas Gd ¥ 1 and Gd ¥ 3 show
similar decreases in r1, characteristic of the formation of an adduct with carbonate ions.
To get more insight into the structural and dynamic properties of the ternary complex
with carbonate anion, the 1/T1 NMRD profile of Gd ¥ 3 at pH 10 was recorded (Fig. 7).
The fit of the experimental data to the values calculated on the basis of SBM theory are
reported in the Table. From this analysis, it was found that, in the ternary complex, the
carbonate ion replaces only one H2O molecule. The remaining inner-sphere H2O
molecule displays a tM value of 180 ns as assessed by 17O-NMR R2 measurements at
variable temperature. This value is in accord with other macrocyclic-based systems with
q ˆ 1 derived from DOTA (ˆ DO3A; see Fig. 1). On this basis, we may tentatively
suggest that there is a direct correspondence between the two structures drawing the
conclusion that the carbonate O-atom coordinates the GdIII ion at a corner position of
the upper O4 plane. Therefore, the H2O molecule in the Gd ¥ 3/carbonate system
occupies an apical position. The observed elongation of tM in the ternary complex
supports the view forwarded in the previous section that a H2O molecule coordinated in
the O4 plane is responsible for the faster exchange detected for Gd ¥ 1 and Gd ¥ 3.
&
&
Fig. 6. pH Dependence of the proton relaxivity (r1p) of Gd ¥ 1 (.), Gd ¥ 2 ( ), and Gd ¥ 3 ( ) measured at 0.47 T
and 258

626
Helvetica Chimica Acta Vol. 86 (2003)
*
Fig. 7. 1/T1 NMRD Profiles of a 1 mm solution of Gd ¥ 3 at pH 7 (.) and pH 10 ( ) and 258. The solid curves
through the data points were calculated with the parameters reported in the Table.
The pH dependence anticipates the results obtained by monitoring the changes in r1
upon addition of increasing amounts of O-donor substrates like lactate, oxalate, and
tartrate. In fact, only Gd ¥ 1 and Gd ¥ 3 show a decrease of r1 upon addition of the
substrates, whereas r1 of Gd ¥ 2 is unaffected. The different behavior clearly reflects the
hydration state of the three complexes: when q ˆ 1, no formation of ternary adduct is
detected, whereas, when q ˆ 2, a bidentate ligand can replace the two inner-sphere H2O
molecules. The binding affinity towards lactate is rather low for Gd ¥ 1 (Ka ˆ 16 mÀ1) and
similar to the value found for the parent [Gd(PCTA)] complex (Ka ˆ 22 mÀ1), while it
is higher for Gd ¥ 3 (Ka ˆ 188 mÀ1). The presence of two negative charges in Gd ¥ 3 does
not hamper lactate binding. Actually, the increased affinity of Gd ¥ 3 for this ligand may
reflect minor structural changes on the inner coordination cage caused by the outer
substituent.
A much stronger binding affinity was observed for the oxalate ligand (Ka ˆ 382 mÀ1
for Gd ¥ 1 and Ka ˆ 2.4 ¥ 105 mÀ1 for Gd ¥ 3). Again Gd ¥ 3, in spite of the residual negative
charge, shows a stronger affinity for the entering substrate.
Binding to Macromolecules. As anticipated in the introduction, the attainment of
high relaxivities is pursued through the binding of the GdIII chelate to a macro-
molecular system. An approach that has met with a large success consists of forming
noncovalent adducts between a suitably functionalized GdIII complex and human-
serum albumin (HSA) [5][28][29]. Although the substituents present on the surface of
the complexes Gd ¥ 1, Gd ¥ 2, and Gd ¥ 3 were not designed for this specific purpose, we
investigated their affinity for HSA.
The assessment of the HSA-binding capability of a given GdIII complex is
conveniently carried out in vitro by means of the proton relaxation enhancement
(PRE) method. In this method, by titrating a solution of the paramagnetic complex

Helvetica Chimica Acta Vol. 86 (2003)
627
with a concentrated solution of HSA, it is possible to determine the thermodynamic
b
binding constant (Ka) and the relaxivity of the macromolecular adduct (r1 ) [27]. This
method is based on the measurement of the increase in the H2O proton longitudinal
relaxation rate that take place when a paramagnetic complex binds to a slowly tumbling
macromolecule as a consequence of the elongation of its tR value. Gd ¥ 1 binds very
weakly (Ka ˆ 200 mÀ1), yielding an estimated value for the relaxivity of the all-bound
adduct of ca. 40 mmÀ1sÀ1. Gd ¥ 2 binds more tightly (Ka ˆ 3300 mÀ1), but the attainable
b
relaxivity for the paramagnetic macromolecular adduct is significantly smaller (r1 ˆ
24.4 mmÀ1sÀ1). Gd ¥ 3 does not interact at all with HSA. The observed behavior can be
accounted for in terms of the following assumptions: i) The phosphonate groups
contribute to the binding (as previously established for [Gd(DOTP)] and [Gd(PCTP)]
[30]), but the geometry at the interacting site is such that either the inner-sphere H2O
molecule exchange is made much more difficult, or the H2O is excluded by direct
coordination of protein donor atoms to the paramagnetic center. The relaxivity found
for the all-bound Gd ¥ 2 adduct is similar to that observed for the [Gd(DOTP)]/HSA
adduct where only second-sphere and outer-sphere contributions are operating. ii) On
comparing the binding affinities of Gd ¥ 1 and Gd ¥ 2, one draws the conclusion that the
benzyl moiety is definitively inadequate to promote hydrophobic binding. iii) The
presence of two negative charges at the aromatic substituent in Gd ¥ 3 further reduces
the binding towards HSA, i.e., the system becomes too hydrophilic for yielding any
measurable binding to the protein.
The estimate of a relaxivity of 40 mmÀ1sÀ1 for the Gd ¥ 1/HSA adduct appears lower
than expected for the characteristics of this complex. Likely, decreased hydration takes
place upon binding to the protein. Thus, the binding through the benzylic moiety at the
distal arm appears unable to avoid the interference of the protein with the H2O
coordinated to the GdIII center. Suitable functionalization should involve large-sized
groups able to keep the chelate moiety sufficiently far apart from the protein surface.
To get some support for this hypothesis, we carried out analogous measurements in
the presence of poly-b-cyclodextrin (poly-b-CD, polymerization degree ˆ 12), which
has been previously shown to be a good model for the formation of macromolecular
adducts based on simple hydrophobic interactions [31]. The binding of Gd ¥ 1 and Gd ¥ 2
with poly-b-CD yielded Ka values of 300 and 95 mÀ1, respectively. Interestingly, the
relaxation enhancements measured for the two adducts are significantly higher than
those found in the case of the corresponding adducts with HSA, being ca. 57 and
46 mmÀ1sÀ1 for Gd ¥ 1 and Gd ¥ 2, respectively. The binding affinity of Gd ¥ 3 towards
poly-b-CD is very small (Ka ˆ 15 mÀ1), and the calculation of the relaxivity of the all-
bound system is more subject to error. Within these limits, we estimate a value for the
macromolecular adduct of 68 mmÀ1sÀ1. Thus, the results from the analysis of the
relaxometric titration with poly-b-CD clearly indicate that very high relaxivities can be
obtained upon the formation of macromolecular adducts provided that the macro-
molecule itself has no −quenching× effect on the hydration properties of the para-
magnetic agent.
Conclusions. In summary, it is worth to highlight the main conclusions we draw
from the results obtained in this work: i) In agreement with previous observations
[9][17], the presence of two coordinating phosphonate groups causes a decrease in the

628
Helvetica Chimica Acta Vol. 86 (2003)
number of the inner-sphere H2O molecules. The effect on the observed relaxivity is
compensated by the ability of phosphonate moieties to increase the second hydration
sphere of the complex.
ii) The rule that GdIII complexes with coordination number 9 should show a slower
exchange rate of the coordinated H2O with respect to systems endowed with
coordination number 8 is not confirmed here. Indeed, 17O-NMR T2 data unambigu-
ously show that tM of the coordinated H2O in Gd ¥ 1 and Gd ¥ 2 are very similar in spite
of the change in coordination number.
iii) The active role of HSA on the hydration properties of Gd ¥ 1 and Gd ¥ 2 indicates
that the benzyl moiety, although bonded to the distal acetate arm, is unable to keep the
coordination cage sufficiently inert to attack of donor atoms present on the surface of
the protein. Likely, larger-sized groups have to be used to avoid such behavior, which
largely −quenches× the attainable relaxivity enhancement.
Experimental Part
1. General. All chemicals were purchased from Aldrich and used without further purification. Anal. TLC:
silica gel 60F254 (Macherey-Nagel), detection by spraying with Dragendorff reagent. M.p.: uncorrected; B¸chi-
510 apparatus. 1H- and 13C-NMR Spectra: at 200 and 50.3 MHz, resp.; Bruker AC-200 spectrometer; CDCl3
soln., unless stated otherwise; chemical shifts d in ppm downfield from internal SiMe4 (ˆ0.00 ppm), J in Hz.
Mass spectra: ESI-ion trap (Bruker Esquire 3000 Plus) or MALDI-TOF (Bruker Reflex III spectrometer) in m/
z. Elemental analyses: Perkin-Elmer 240 instrument.
2. Esterification: General Procedure. To a suspension of l-amino acid 4 or 5 (60.0 mmol) in MeOH (80 ml)
at 08, SOCl2 (9.0 ml) is slowly added. The obtained soln. is stirred at r.t. for ca. 2.5 h and is then heated at 408 for
1 h. The solvent is evaporated. To the oily residue, 10% aq. Na2CO3 soln. is added and the suspension extracted
with AcOEt (3 Â 30 ml). The org. phase is dried (Na2SO4) and evaporated to give pure 6 and 7, resp.
l-Phenylalanine Methyl Ester (6): Colorless oil (8.6 g, 80%). 1H-NMR: 7.23 7.10 (m, 5 H); 3.62 3.54
(m, 4 H); 3.00 (m, 1 H); 2.80 (dd, J ˆ 12.4, 7.2, 1 H); 1.42 (br. s, 2 H). 13C-NMR: 172.2 (C); 137.1 (C); 129.1
(CH); 128.4 (CH); 126.6 (CH); 55.6 (CH); 51.7 (Me); 41.0 (CH2).
l-Tyrosine Methyl Ester (7): Light yellow solid (9.4 g, 80%). M.p. 134 1358. 1H-NMR: 6.87, 6.65
(AA'BB', J ˆ 8.1, 4 H); 3.59 3.53 (m, 4 H); 2.87 (dd, J ˆ 13.6, 4.7, 1 H); 2.62 (br. s, 3 H); 2.67 (dd, J ˆ 13.6, 7.6,
1 H). 13C-NMR: 175.3 (C); 156.1 (C); 130.1 (CH); 127.4 (C); 115.5 (CH); 55.8 (CH); 51.8 (Me); 40.1 (CH2).
3. Reaction with 1-Tosylaziridine: General Procedure. The appropriate amino acid methyl ester 6 or 7
(20 mmol) and 1-tosylaziridine (44 mmol) are dissolved in toluene (20 ml) and the soln. refluxed for 6 h. The
solvent is then evaporated and the waxy solid purified by FC (silica gel; for 8, hexane/iPrOH 9 :1; for 9, hexane/
iPrOH/CH2Cl2 9 :0.5 :0.5).
N,N-Bis{2-{[(4-methylphenyl)sulfonyl]amino}ethyl}-l-phenylalanine Methyl Ester (8). Yellow waxy solid
(10.3 g, 90%). 1H-NMR: 7.71, 7.27 (AA'BB', J ˆ 8.2, 8 H); 7.23 7.12 (m, 5 H); 5.11 (br. s, 2 H); 3.59 (s, 3 H);
3.43 (t, J ˆ 7.4, 1 H); 2.97 (dd, J ˆ 14.1, 7.4, 1 H); 2.85 2.60 (m, 9 H); 2.39 (s, 6 H). 13C-NMR: 172.3 (C); 143.2
(C); 138.0 (C); 136.7 (C); 129.6 (CH); 129.1 (CH); 128.6 (CH); 127.0 (CH); 126.6 (CH); 64.4 (CH); 51.4 (Me);
‡
‡
50.5 (CH2); 40.8 (CH2); 35.6 (CH2); 21.4 (Me). ESI-MS: 574 ([M ‡ H] ), 1147 ([2M ‡ H] ). Anal. calc. for
C28H35N3O6S2 (573.7): C 58.62, H 6.15, N 7.32; found: C 58.44, H 6.22, N 7.08.
N,N-Bis{2-{[(4-methylphenyl)sulfonyl]amino}ethyl}l-tyrosine Methyl Ester (9): Yellow waxy solid (10.6 g,
90%). 1H-NMR: 7.71, 7.30 (AA'BB', J ˆ 8.1, 8 H); 7.08, 6.85 (AA'BB', J ˆ 8.4, 4 H); 6.22 (br. s, 1 H); 4.66 (br. s,
2 H); 3.67 (s, 3 H); 3.39 (dd, J ˆ 9.9, 5.4, 1 H); 2.95 (dd, J ˆ 14.3, 5.4, 1 H); 2.86 2.57 (m, 9 H); 2.41 (s, 6 H).
13C-NMR: 172.6 (C); 154.8 (C); 143.4 (C); 136.5 (C); 130.2 (CH); 129.7 (CH); 129.6 (C); 127.0 (CH); 115.8
‡
(CH); 64.4 (CH); 51.5 (Me); 50.1 (CH2); 40.7 (CH2); 34.4 (CH2); 21.4 (Me). ESI-MS: 590 ([M ‡ H] ), 1179
‡
([2M ‡ H] ). Anal. calc. for C28H35N3O7S2 (589.7): C 57.03, H 5.98, N 7.13; found: C 56.90, H 6.09, N 6.99.
4. O-[(4-Methylphenyl)sulfonyl]-N,N-bis{2-{[(4-methylphenyl)sulfonyl]amino}ethyl}-l-tyrosine Methyl
Ester (10). A soln. of 9 (5.30 g, 9.0 mmol) and TsCl (3.20 g, 17 mmol) in pyridine (15 ml) is stirred at r.t. for
1 day. Then more TsCl is added (1.7 g, 9.0 mmol) and the mixture kept stirring for one more day. After addition
of 10% aq. Na2CO3 soln. (20 ml), the mixture is extracted with AcOEt (3 Â 10 ml) and the org. phase dried

Helvetica Chimica Acta Vol. 86 (2003)
629
(Na2SO4), and evaporated: 10 (4.8 g, 72%). Yellow waxy solid, pure enough to be used directly for the following
step. 1H-NMR: 7.72, 7.30 (2 AA'BB', J ˆ 8.1, 12 H); 7.14, 6.91 (AA'BB', J ˆ 8.5, 4 H); 4.96 (br. s, 2 H); 3.64
(s, 3 H); 3.44 (t, J ˆ 7.2, 1 H); 3.02 2.60 (m, 10 H); 2.46 (s, 6 H); 2.42 (s, 3 H). 13C-NMR: 172.4 (C); 147.4 (C);
143.3 (C); 142.4 (C); 138.0 (C); 137.3 (C); 132.2 (C); 130.4 (CH); 129.7 (2 Â CH); 128.4 (CH); 127.0 (CH); 122.5
‡
(CH); 64.7 (CH); 51.6 (Me); 50.7 (CH2); 40.9 (CH2); 34.8 (CH2); 21.4 (2 Me). ESI-MS: 744 ([M ‡ H] ). Anal.
calc. for C35H41N3O9S3 (743.9): C 56.51, H 5.56, N 5.65; found: C 56.60, H 5.71, N 5.49.
5. O-[(tert-Butyl)dimethylsilyl]-N,N-bis{2-{[(4-methylphenyl)sulfonyl]amino}ethyl}-l-tyrosine Methyl Es-
ter (11). To a soln. of 9 (15 g, 26.0 mmol) and (tert-butyl)dimethylsilyl chloride (5.8 g, 38.0 mmol) in MeCN
(100 ml) containing 4-ä-molecular sieves, iPr2EtN (9.86 g, 76.4 mmol) is slowly added and the mixture refluxed
for 1 h under N2. The mixture is filtered, the filtrate evaporated, and the residue rinsed with H2O and extracted
with AcOEt (3 Â 30 ml). The org. phase is dried (Na2SO4) and evaporated: 11 (17.0 g, 93%). Yellow waxy solid,
pure enough to be used for the following step. 1H-NMR: 7.74, 7.28 (AA'BB', J ˆ 8.2, 8 H); 7.02, 6.77
(AA'BB', J ˆ 8.3, 4 H); 5.09 (br. s, 2 H); 3.60 (s, 3 H); 3.38 (t, J ˆ 7.7, 1 H); 2.89 2.59 (m, 10 H); 2.41 (s, 6 H);
1.00 (s, 9 H); 0.09 (s, 6 H). 13C-NMR: 172.9 (C); 154.6 (C); 143.6 (C); 137.3 (C); 131.0 (C); 130.5 (CH); 130.1
(CH); 127.5 (CH); 120.5 (CH); 65.2 (CH); 51.8 (Me); 51.1 (CH2); 41.4 (CH2); 35.5 (CH2); 26.1 (Me); 21.9
‡
(Me); 18.2 (C); À3.2 (Me). ESI-MS: 704 ([M ‡ H] ). Anal. calc. for C34H49N3O7S2Si (704.0): C 58.01, H 7.02,
N 5.97; found: C 57.78, H 7.05, N 6.02.
6. Macrocyclization Reaction with 2,6-Bis(chloromethyl)pyridine: General Procedure. The functionalized
diethylenetriamine 8, 10, or 11 (10.0 mmol) and 2,6-bis(chloromethyl)pyridine (10.0 mmol) are dissolved in
MeCN (60 ml), and anh. K2CO3 (40.0 mmol) is added. The mixture is vigorously stirred and refluxed for 6 h.
The solvent is evaporated, the residue washed with H2O and extracted with AcOEt (3 Â 30 ml), the org. phase
dried (Na2SO4) and evaporated, and the yellow waxy solid crystallized from AcOEt/hexane.
3,9-Bis[(4-Methylphenyl)sulfonyl)]-a-(phenylmethyl)-3,6,9,15-tetraazabicyclo[9.3.1]pentadeca-1(15),11,13-
triene-6-acetic Acid Methyl Ester (12): White solid (4.4 g, 65%). M.p. 152 1548 (AcOEt/hexane). 1H-NMR: 7.76
(t, J ˆ 7.7, 1 H); 7.70 (1/2 AA'BB', J ˆ 8.2, 4 H); 7.39 7.13 (m, 11 H); 4.28 (d, J ˆ 12.1, 2 H); 4.23 (d, J ˆ 12.1,
2 H); 3.58 (s, 3 H); 3.48 (dd, J ˆ 8.8, 6.7, 1 H); 3.08 2.87 (m, 5 H); 2.72 (dd, J ˆ 14.1, 8.8, 1 H); 2.47 (s, 6 H);
2.37 2.22 (m, 4 H). 13C-NMR: 174.2 (C); 155.3 (C); 144.0 (C); 139.3 (CH); 138.6 (C); 135.9 (C); 130.2 (CH);
129.4 (CH); 128.6 (CH); 127.6 (CH); 126.7 (CH); 124.8 (CH); 68.0 (CH); 55.0 (CH2); 53.1 (CH2); 51.9 (Me);
‡
46.2 (CH2); 37.3 (CH2); 22.0 (Me). ESI-MS: 677 ([M ‡ H] ). Anal. calc. for C35H40N4O6S2 (676.8): C 62.11,
H 5.96, N 8.28; found: C 61.97, H 5.99, N 8.28.
3,9-Bis[(4-methylphenyl)sulfonyl]-a-{{4-{[(4-methylphenyl)sulfonyl]oxy}phenyl}methyl}3,6,9,15-tetraaza-
bicyclo[9.3.1]pentadeca-1(15),11,13-triene-6-acetic Acid Methyl Ester (13): White solid (5.1 g, 60%). M.p. 165
1
1678 (AcOEt/hexane). H-NMR: 7.71 (t, J ˆ 7.8, 1 H); 7.66, 7.32 (AA'BB', J ˆ 7.8, 12 H); 7.27 (d, J ˆ 7.8, 2 H);
7.03, 6.84 (AA'BB', J ˆ 8.4, 4 H); 4.26 (d, J ˆ 12.4, 2 H); 4.19 (d, J ˆ 12.4, 2 H); 3.53 (s, 3 H); 3.37 (dd, J ˆ 8.4,
6.6, 1 H); 3.07 2.83 (m, 5 H); 2.67 (dd, J ˆ 14.6, 8.4, 1 H); 2.42 (s, 9 H); 2.40 2.21 (m, 4 H). 13C-NMR: 173.4
(C); 154.9 (C); 148.1 (C); 145.3 (C); 143.6 (C); 138.8 (CH); 137.3 (C); 135.6 (C); 132.4 (C); 130.1 (CH); 129.9
(CH); 129.8 (CH); 128.5 (CH); 127.1 (CH); 124.2 (CH); 122.1 (CH); 67.5 (CH); 54.6 (CH2); 52.7 (CH2); 51.5
‡
(CH2); 45.9 (CH2); 36.2 (CH2); 21.7 (Me); 21.5 (Me). ESI-MS: 847 ([M ‡ H] ). Anal. calc. for C42H46N4O9S3
(847.0): C 59.56, H 5.47, N 6.61; found: C 59.65, H 5.61, N 6.48.
a-{{4-{[(tert-Butyl)dimethylsilyl]oxy}phenyl}methyl}-3,9-bis[(4-methylphenyl)sulfonyl]-3,6,9,15-tetraaza-
bicyclo[9.3.1]pentadeca-1(15),11,13-triene-6-acetic Acid Methyl Ester (14): White solid (5.0 g, 62%). M.p. 200
2028 (AcOEt/hexane). 1H-NMR: 7.77 (t, J ˆ 7.7, 1 H); 7.70, 7.35 (AA'BB', J ˆ 8.1, 8 H); 7.38 (d, J ˆ 7.7, 2 H); 6.99,
6.73 (AA'BB', J ˆ 8.3, 4 H); 4.28 (d, J ˆ 12.0, 2 H); 4.21 (d, J ˆ 12.0, 2 H); 3.56 (s, 3 H); 3.41 (dd, J ˆ 8.5, 7.1,
1 H); 3.02 2.89 (m, 5 H); 2.97 (dd, J ˆ 16.3, 7.1, 1 H); 2.45 (s, 6 H); 2.38 2.30 (m, 4 H); 0.99 (s, 9 H); 0.19
(s, 6 H). 13C-NMR: 174.3 (C); 155.3 (C); 154.4 (C); 143.9 (C); 139.2 (CH); 135.9 (C); 131.2 (C); 130.3 (CH);
130.2 (CH); 127.6 (CH); 124.8 (CH); 120.3 (CH); 68.1 (CH); 55.0 (CH2); 53.0 (CH2); 51.8 (Me); 46.3 (CH2);
‡
36.6 (CH2); 26.1 (Me); 22.0 (Me); 18.6 (C); À4.0 (Me). ESI-MS: 807 ([M ‡ H] ). Anal. calc. for
C41H54N4O7S2Si (807.1): C 61.01, H 6.74, N 6.94; found: C 60.88, H 6.91, N 6.90.
7. a-(Phenylmethyl)-3,6,9,15-tetraazabicyclo[9.3.1]pentadeca-1(15),11,13-triene-6-acetic Acid (15). A mix-
ture of 12 (3.96 g, 5.87 mmol), 48% aq. HBr soln. (39.6 ml, 352 mmol), AcOH (10.0 ml, 175 mmol), and phenol
(3.31 g, 35.2 mmol) is refluxed for 12 h. The soln. is then diluted with H2O and extracted with CH2Cl2 (3 Â
20 ml). The aq. phase is evaporated, and the same quantity of reactants is added to the residue; the resulting
soln. is refluxed for additional 12 h. The soln. is then extracted with CH2Cl2 (3 Â 20 ml), the aq. phase
concentrated to ca. 2 ml, and Et2O (15 ml) added. The white-pink crystals obtained are isolated by filtration and
washed with acetone: 15 ¥ 2 HBr¥ 2 H2O (2.43 g, 75%). White solid. M.p. 125 1308. 1H-NMR (D2O): 7.76 (t, J ˆ
7.8, 1 H); 7.27 7.20 (m, 6 H); 7.11 (tt, J ˆ 7.1, 1.4, 1 H); 4.44 (s, 4 H); 3.85 (t, J ˆ 7.6, 1 H); 3.15 (dd, J ˆ 14.2, 7.1,

630
Helvetica Chimica Acta Vol. 86 (2003)
1 H); 3.01 (dd, J ˆ 14.2, 8.0, 1 H); 2.96 (m, 8 H). 13C-NMR (D2O): 176.1 (C); 148.9 (C); 140.0 (CH); 138.2 (C);
129.6 (CH); 129.5 (CH); 127.6 (CH); 122.6 (CH); 69.2 (CH); 51.8 (CH2); 49.9 (CH2); 46.9 (CH2); 35.0 (CH2).
‡
‡
ESI-MS: 355 ([M ‡ H] ), C20H26N4O2 ; calc. 354.4). Anal. calc. for C20H26N4O2 ¥ 2 HBr ¥ 2 H2O (552.3): C 43.49,
H 5.84, N 10.14; found: C 43.37, H 5.85, N 9.96.
a-{{4-{[(4-Methylphenyl)sulfonyl]oxy}phenyl}methyl}-3,6,9,15-tetraazabicyclo[9.3.1]pentadeca-1(15),11,13-
triene-6-acetic Acid (16). As described for 15, 13 (1.69 g, 2.00 mmol): 16 ¥ 2 HBr (0.95 g, 69%). Amorphous
powder. 1H-NMR (D2O): 7.73 (t, J ˆ 7.8, 1 H); 7.42 (1/2 AA'BB', J ˆ 7.8, 2 H); 7.23 7.05 (m, 6 H); 6.65
(1/2 AA'BB', J ˆ 8.2, 2 H); 4.38 (s, 4 H); 3.72 (t, J ˆ 7.6, 1 H); 3.05 (dd, J ˆ 13.7, 6.7, 1 H); 3.03 2.75 (m, 9 H);
‡
‡
2.16 (s, 3 H). ESI-MS: 525 ([M ‡ H] ), C27H32N4O5S ; calc. 524.6). Anal. calc. for C27H32N4O5S ¥ 2 HBr (686.5):
C 47.24, H 4.99, N 8.16; found: C 46.98, H 5.09, N 8.03.
8. a-[(4-Hydroxyphenyl)methyl]-3,6,9,15-tetraazabicyclo[9.3.1]pentadeca-1(15),11,13-triene-6-acetic Acid
Methyl Ester (17). A soln. of 14 (2.20 g, 2.73 mmol), 33% HBr in AcOH (30 ml), and phenol (1.54 g, 16.4 mmol)
is heated at 608 for 18 h. H2O is then added (50 ml), and the mixture is extracted with CH2Cl2 (3 Â 20 ml). The
aq. phase is evaporated: 17 ¥ 2 HBr (1.48 g, 99%). Amorphous solid, which is used for the next reaction without
further purification. 1H-NMR (D2O): 7.52 (t, J ˆ 7.8, 1 H); 7.01 (d, J ˆ 7.8, 2 H); 6.84, 6.41 (AA'BB', J ˆ 8.4,
4 H); 4.19 (s, 4 H); 3.58 (t, J ˆ 7.6, 1 H); 3.29 (s, 3 H); 2.77 2.60 (m, 10 H). 13C-NMR (D2O): 170.5 (C); 149.7
(C); 149.5 (C); 140.7 (CH); 131.3 (CH); 131.0 (C); 123.3 (CH); 116.8 (CH); 69.8 (CH); 53.4 (Me); 52.3 (CH2);
‡
‡
49.6 (CH2); 47.5 (CH2); 34.9 (CH2). ESI-MS: 385 ([M ‡ H] , C21H28N4O3 ; calc. 384.5). Anal. calc. for
C21H28N4O3 ¥ 2 HBr (546.3): C 46.17, H 5.54, N 10.26; found: C 45.89, H 5.69, N 10.12.
9. a6-(Phenylmethyl)-3,6,9,15-tetraazabicyclo[9.3.1]pentadeca-1(15),11,13-triene-3,6,9-triacetic Acid (1). To
a soln. of 15 (0.50 g, 0.91 mmol) and chloroacetic acid (0.215 g, 2.28 mmol) in H2O (1 ml), 6m aq. KOH (0.5 ml)
and 2 drops of a phenolphthalein soln. (0.2% in MeOH) are added. The soln. is heated at 808 for 6 h,
maintaining pH 10 (persistent pink color of phenolphthalein) by slow addition of KOH soln. The soln. is then
acidified with aq. 2n HCl and concentrated to a small volume. The residue is purified by column
chromatography (XAD 1600, H2O ! H2O/MeOH 7 :3): pure 1 (0.29 g, 68%). White amorphous solid.
1H-NMR (D2O): 7.76 (t, J ˆ 7.8, 1 H); 7.47 (d, J ˆ 7.8, 2 H); 7.26 7.06 (m, 5 H); 4.52 (s, 4 H); 3.77 (m, 5 H);
3.37 3.13 (m, 4 H); 2.99 2.70 (m, 6 H). 13C-NMR (D2O): 177.8 (C); 176.2 (C); 150.6 (C); 141.3 (CH); 138.6
(C); 130.5 (CH); 130.1 (CH); 128.2 (CH); 124.2 (CH); 67.4 (CH), 60.6 (CH2); 57.6 (CH2); 54.6 (CH2); 49.1
‡
‡
‡
(CH2); 35.2 (CH2). MALDI-MS: 471 ([M ‡ H] ), 493 ([M ‡ Na] ), 509 ([M ‡ K] ). Anal. calc. for
C24H30N4O6 (470.5): C 61.26, H 6.43, N 11.91; found: C 61.25, H 6.60, N 12.01.
10. a-(Phenylmethyl)-3,6-bis(phosphonomethyl)-3,6,9,15-tetraazabicyclo[9.3.1]pentadeca-1(15),11,13-tri-
ene-6-acetic Acid (2). A soln. of 15 (0.15 g, 0.27 mmol) in 6m HCl (5 ml) and phosphorous acid (0.114 g,
1.39 mmol) is brought to reflux, and paraformaldehyde (0.1 g, 3.30 mmol) is portionwise added within ca. 1 h. At
the end of the addition, the soln. is refluxed for further 8 h. The mixture is concentrated to ca. 1 ml, and EtOH is
added to precipitate 2, which is filtered, washed with EtOH and then with Et2O and finally dried: 2 (0.096 g,
1
61%). White amorphous solid. H-NMR (D2O): 7.77 (t, J ˆ 7.8, 1 H); 7.27 (d, J ˆ 7.8, 2 H); 7.16 (m, 4 H); 7.10
(m, 1 H); 4.77 (m, 4 H); 3.75 (t, J ˆ 8.1, 1 H); 3.59 3.40 (m, 6 H); 3.42 (d, J ˆ 11.8, 4 H); 3.13 2.94 (m, 4 H).
13C-NMR (D2O): 174.4 (C); 148.5 (CH); 139.4 (CH); 136.7 (C); 128.6 (CH); 128.2 (CH); 126.3 (C); 122.1
(CH), 65.0 (CH); 58.7 (CH2); 54.1 (CH2); 51.0 (d, J(C,P) ˆ 138, CH2); 46.4 (CH2); 33.0 (CH2). MALDI-MS:
‡
‡
‡
‡
581 ([M ‡ K] , 565 ([M ‡ Na] , 543 ([M ‡ H] , C22H32N4O8P2 ; calc. 542.5). Anal. calc. for C22H32N4O8P2 ¥
2 H2O (578.5): C 45.68, H 6.27, N 9.68; found: C 45.39, H 6.39, N 9.50.
11. a6-{[3-Carboxy-4-(2-methoxy-2-oxoethoxy)phenyl]methyl}-3,6,9,15-tetraazabicyclo[9.3.1]pentadeca-
1(15),11,13-triene-3,6,9-triacetic Acid Trimethyl Ester (18). A mixture of 17 (0.250 g, 0.46 mmol), silver
carbonate (0.758 g, 2.75 mmol), and methyl bromoacetate (260 ml, 2.75 mmol) in MeCN (2 ml) is vigorously
stirred for 2 h in the dark. The mixture is then filtered and the soln. obtained is evaporated. The crude product is
purified by FC (silica gel, CH2Cl2/MeOH/NH3 95 :5 :0.1): 18 (0.172 g, 58%). 1H-NMR: 7.71 (m, 2 H); 7.37 7.27
(m, 2 H); 7.00 (m, 1 H); 6.75 (m, 1 H); 4.37 (d, J ˆ 12.0, 2 H); 4.23 (d, 12.0, 2 H); 3.96 (s, 2 H); 3.75 (s, 6 H); 3.70
(s, 3 H); 3.55 (s, 4 H); 3.53 (s, 3 H); 3.21 (t, J ˆ 6.4, 1 H); 2.92 (dd, J ˆ 14.5, 6.4, 1 H); 2.75 2.29 (m, 9 H).
13C-NMR: 173.4 (C); 169.7 (C); 169.0 (C); 168.1 (C); 156.2 (C); 155.1 (C); 138.2 (CH); 130.5 (C); 130.2 (CH);
130.1 (CH); 123.0 (CH); 114.5 (CH); 113.9 (C); 70.4 (CH2); 65.2 (CH2); 61.5 (CH2); 52.1 (Me); 52.0 (Me); 51.8
‡
‡
(CH); 51.3 (Me); 49.2 (CH2); 47.5 (CH2); 29.6 (CH2). MALDI-MS: 645 ([M ‡ H] ), 667 ([M ‡ Na] ), 683
‡
([M ‡ K] ). Anal. calc. for C31H40N4O11 (644.7): C 57.76, H 6.25, N 8.69; found: C 57.53, H 6.41, N 8.77.
12. a6-{[3-Carboxy-4-(carboxymethoxy)phenyl]methyl}-3,6,9,15-tetraazabicyclo[9.3.1]pentadeca-1(15),11,13-
triene-3,6,9-triacetic Acid (3). A soln. of 18 (0.110 g, 0.171 mmol) and KOH (0.400 g, 0.700 mmol) in MeOH
(5 ml) is vigorously stirred and heated at 608 for 4 h. The solvent is then evaporated and the solid obtained is
washed thoroughly with MeOH/Et2O: 3 (0.059 g). White amorphous solid. 1H-NMR (D2O): 7.69 (t, J ˆ 7.5,

Helvetica Chimica Acta Vol. 86 (2003)
631
1 H); 7.23 (m, 2 H); 6.95 (m, 2 H); 6.65 (d, J ˆ 8.0, 1 H); 4.28 4.12 (m, 6 H); 3.81 (m, 1 H); 3.49 (s, 4 H); 3.25
2.95 (m, 6 H); 2.79 2.52 (m, 4 H). 13C-NMR (D2O): 177.6 (C); 176.8 (C); 176.7 (C); 173.4 (C); 157.2 (C); 156.9
(C); 139.7 (CH); 131.9 (C); 130.8 (CH); 130.7 (2 CH); 115.4 (CH); 115.2 (C); 81.6 (CH); 70.4 (CH2); 67.8
‡
‡
(CH2); 65.1 (CH2); 59.2 (CH2); 54.5 (CH2); 51.0 (CH2). MALDI-MS: 589 ([M ‡ H] , C27H32N4O11 ; calc. 588.6),
‡
‡
611 [M ‡ Na] ), 627 ([M ‡ K] ).
13. Water Proton Relaxivity Measurements. The longitudinal H2O proton relaxation rate was measured with
a Stelar-Spinmaster spectrometer (Stelar, Mede, Pavia, Italy) operating at 20 MHz, by means of the standard
inversion-recovery technique (16 experiments, 2 scans). A typical 908 pulse width was 3.5 ms and the
reproducibility of the T1 data was Æ0.5%. The temp. was controlled with a Stelar-VTC-91 air-flow heater
equipped with a copper-constantan thermocouple (uncertainty Æ 0.18). The proton 1/T1 NMRD profiles were
measured over a continuum of magnetic-field strength from 0.00024 to 0.28 T (corresponding to 0.01 12 MHz
proton Larmor frequency) on a Stelar-Fast-Field-Cycling relaxometer. This relaxometer works under complete
computer control with an absolute uncertainty in 1/T1 of Æ 1%. Data points at 20 MHz, were added to the
experimental NMRD profiles and were recorded on the Stelar Spinmaster (20 MHz). The concentration of the
GdIII complexes in the solutions used for the relaxometric determinations was obtained by means of ICP
analysis.
14. 17O-NMR Measurements. Variable-temperature 17O-NMR measurements were recorded at 2.1 T by
means of the Jeol EX-90 and at 7.05 T by means of the Bruker Avance-300 spectrometer, equipped with a 5-mm
probe, by using a D2O external lock. Experimental settings were: spectral width 10000 Hz, 908 pulse (7 ms),
acquisition time 10 ms, 1000 scans, and without sample spinning. Aq. solns. containing 2.6% of 17O isotope
(Yeda, Israel) were used. The observed transverse relaxation rates (R02obs ) were calculated from the signal width
at half-height (Dn1/2): R02obs ˆ p Dn1/2
.
Financial support from Bracco Imaging S.p.a., Milan, and from Italian CNR (Target Project on
Biotechnology) is gratefully acknowledged. Two of the authors (C. C. and M. S.) are thankful to the
¡
Dipartimento di Chimica Organica e Industriale, Universita degli Studi, Milan, Italy, for kind hospitality in their
laboratories.
REFERENCES
[1] B. L. Engelstad, G. L. Wolf, in −Magnetic Resonance Imaging×, Eds. D. D. Stark and W. G. Bradley Jr.,
Mosby, St Louis, pp. 161 181, 1988.
[2] P. A. Rinck, −Magnetic Resonance in Medicine×, Blackwell Scientific Publications, Oxford, 1993.
[3] P. Caravan, J. J. Ellison, T. J. McMurry, R. B. Lauffer, Chem. Rev. 1999, 99, 2293.
[4] −The Chemistry of Contrast Agents in Medical Magnetic Resonance Imaging×, Eds. A. E. Merbach and E.
Toth, John Wiley & Sons, Ltd., Chichester, 2001.
[5] R. B. Lauffer, Chem. Rev. 1987, 87, 901.
[6] S. Aime, M. Botta, M. Fasano, E. Terreno, Chem. Soc. Rev. 1998, 27, 19.
[7] E. Toth, O. M. Ni Dhubhghaill, G. Besson, L. Helm, A. E. Merbach, Magn. Reson. Chem. 1999, 37, 701.
[8] D. H. Powell, O. M. Ni Dhubhghaill, D. Pubanz, L. Helm, Y. S. Lebedev, W. Schlaepfer, A. E. Merbach, J.
Am. Chem. Soc. 1996, 118, 9333.
[9] S. Aime, M. Botta, S. Geninatti Crich, G. B. Giovenzana, R. Pagliarin, M. Sisti, E. Terreno, Magn. Reson.
Chem. 1998, 36, S200.
[10] 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.
[11] J. Xu, S. J. Franklin Jr., D. W. Whisenhunt, K. N. Raymond, J. Am. Chem. Soc. 1995, 117, 7245.
[12] S. Hajela, M. Botta, S. Giraudo, J. Xu, K. N. Raymond, S. Aime, J. Am. Chem. Soc. 2000, 122, 11228.
[13] D. M. J. Doble, M. Botta, J. Wang, S. Aime, A. Barge, K. N. Raymond, J. Am. Chem. Soc. 2001, 123, 10758.
[14] W. D. Kim, G. E. Kiefer, F. Maton, K. McMillan, R. N. Muller, A. D. Sherry, Inorg. Chem. 1995, 34, 2233.
[15] S. Aime, M. Botta, S. Geninatti Crich, G. B. Giovenzana, G. Jommi, R. Pagliarin, M. Sisti, Inorg. Chem.
1997, 36, 2992.
[16] S. Aime, M. Botta, S. Geninatti Crich, G. B. Giovenzana, G. Jommi, R. Pagliarin, M. Sisti, J. Chem. Soc.,
Chem. Commun. 1995, 1885.
[17] 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.
[18] L. Banci, I. Bertini, C. Luchinat, −Nuclear and Electron Relaxation×, VCH, Weinheim, 1991.

632
Helvetica Chimica Acta Vol. 86 (2003)
[19] I. Solomon, N. Bloembergen, J. Chem. Phys. 1956, 25, 261.
[20] N. Bloembergen, L. O. Morgan, J. Chem. Phys. 1961, 34, 842.
[21] M. Botta, Eur. J. Inorg. Chem. 2000, 399.
[22] K. Mickei, L. Helm, E. Brucher, A. E. Merbach, Inorg. Chem. 1993, 32, 3844.
[23] T. J. Swift, R. E. J. Connick, J. Chem. Phys. 1962, 37, 307.
[24] D. H. Powell, M. Favre, N. Graeppi, O. M. Ni Dhubhghaill, D. Pubanz, A. E. Merbach, J. Alloys Compd.
1995, 225, 246.
[25] G. Gonzalez, D. H. Powell, V. Tissieres, A. E. Merbach, J. Chem. Phys. 1994, 98, 53.
[26] C. A. Chang, H. G. Brittain, J. Telser, M. F. Tweedle, Inorg. Chem. 1990, 29, 4468.
[27] L. Burai, V. Hietapelto, R. Kiraly, E. Toth, E. Brucher, Magn. Reson. Med. 1997, 38, 146.
[28] S. Aime, M. Botta, M. Fasano, S. Geninatti Crich, E. Terreno, J. Biol. Inorg. Chem. 1996, 1, 312.
[29] S. Aime, M. Chiaussa, G. Digilio, E. Gianolio, E. Terreno, J. Biol. Inorg. Chem. 1999, 4, 766.
[30] S. Aime, M. Botta, S. Geninatti Crich, G. B. Giovenzana, R. Pagliarin, R. Piccinini, M. Sisti, E. Terreno, J.
Biol. Inorg. Chem. 1997, 2, 470.
[31] S. Aime, M. Botta, F. Fedeli, E. Gianolio, E. Terreno, P. L. Anelli, Chem. Eur. J. 2001, 7, 5261.
Received July 16, 2002