Redox-active magnetic resonance imaging contrast agents: Studies with thiol-bearing 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetracetic acid derivatives

Article abstract of DOI:10.1021/jm300736f

The synthesis and structure-activity relationships of a homologous series of 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid gadolinium(III) complexes bearing thiol-terminated alkyl side chains from three to nine carbons in length are reported. The observed binding with human serum albumin (HSA) of the compounds having C-3 through C-7 side chain lengths was inhibited by homocysteine in a manner consistent with single-site binding. The observed binding with HSA of the compounds having C-8 and C-9 side chain lengths was only partly inhibited by homocysteine, consistent with multisite binding. The binding affinity of the C-7 compound could be related to the HSA oxidation state. 2D ¹H-¹H NMR TOCSY provided evidence of covalent binding of the europium analog of the C-6 compound to HSA-Cys³⁴. The longitudinal water-proton MRI relaxivities of the gadolinium complexes at 7 T increased upon binding to HSA. On the basis of these results, the C-6 and C-7 compounds were identified as promising redox-sensitive MRI contrast agents.

Full text of DOI:10.1021/jm300736f

Article
pubs.acs.org/jmc
Redox-active Magnetic Resonance Imaging Contrast Agents: Studies
with Thiol-bearing 1,4,7,10-Tetraazacyclododecane-1,4,7,10-
tetracetic Acid Derivatives
Bhumasamudram Jagadish,† Gerald P. Guntle,‡ Dezheng Zhao,‡ Vijay Gokhale,§ Tarik J. Ozumerzifon,†
Ali M. Ahad,† Eugene A. Mash,† and Natarajan Raghunand*,‡,∥
†Department of Chemistry and Biochemistry, University of Arizona, Tucson, Arizona 85721-0041, United States
‡Arizona Cancer Center, University of Arizona, Tucson, Arizona 85724-5024, United States
§Department of Pharmacology and Toxicology, University of Arizona, Tucson, Arizona 85721-0240, United States
∥Department of Radiology, University of Arizona, Tucson, Arizona 85724-5067, United States
S
* Supporting Information
ABSTRACT: The synthesis and structure−activity relationships of a homologous
series of 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid gadolinium(III)
complexes bearing thiol-terminated alkyl side chains from three to nine carbons in
length are reported. The observed binding with human serum albumin (HSA) of the
compounds having C-3 through C-7 side chain lengths was inhibited by
homocysteine in a manner consistent with single-site binding. The observed
binding with HSA of the compounds having C-8 and C-9 side chain lengths was
only partly inhibited by homocysteine, consistent with multisite binding. The
binding affinity of the C-7 compound could be related to the HSA oxidation state. 2D 1H−1H NMR TOCSY provided evidence
of covalent binding of the europium analog of the C-6 compound to HSA-Cys34. The longitudinal water-proton MRI relaxivities
of the gadolinium complexes at 7 T increased upon binding to HSA. On the basis of these results, the C-6 and C-7 compounds
were identified as promising redox-sensitive MRI contrast agents.
drug molecules.4 Elevated expression of the gene targets of
INTRODUCTION
■
NRF2 has been implicated in chemoresistance in ovarian cancer
In vivo results indicate that the redox state of thiol/disulfide
cells in vitro.5 Elevated cellular thiol concentrations have also
pools in human plasma varies little between healthy individuals,
been reported to diminish radiation response by scavenging
and in vitro both cultured cells and perfused tissues are able to
intracellular ROS produced by ionizing radiation.6 Contrarily,
reduced glutathione and N-acetylcysteine have been shown to
inhibit the activation and function of MMP-9, a matrix
metalloproteinase implicated in the malignant progression of
precancerous lesions, by preserving the cysteine sulfur blockage
of the active site in the pro-peptide.7 The physiological
nonprotein thiols (NPSH) cysteine, homocysteine, and
glutathione (GSH) have been reported to cause inactivation
of transforming growth factor-β, which can induce epithelial-to-
mesenchymal transition of cells in late-stage tumors and
promote tumor growth and metastasis.8 Satoh, et al. have
revealed a different mechanism by which NRF2 can play a role
in the prevention of cancer metastasis, through its ability to
preserve the redox balance in the hematopoietic and immune
systems.9 Thus, high levels of tumor NPSH may influence both
cancer progression and tumor response to therapies, but toward
opposing clinical outcomes. This highlights the need for
noninvasive techniques for measuring tumor redox in vivo.
regulate the extracellular redox state.1 In mammalian cells,
response to oxidative stress is brought about by the nuclear
factor erythroid 2-related factor 2 (NRF2) transcription factor,
which is normally sequestered in the cytoplasm by Kelch-like
ECH-associated protein 1 (KEAP1). In response to oxidative
stress, NRF2 dissociates from KEAP1 and activates tran-
scription of many protective genes, including genes for
glutathione synthesis and genes that encode for proteins such
as glutathione reductase, thioredoxin, thioredoxin reductase,
peroxiredoxin, and sulfiredoxin that restore oxidized intra-
cellular thiols to their reduced states. NRF2 also controls the
expression of the multidrug resistance protein (MRP), a
putative glutathione transporter, thereby providing cells with
the ability to control both intracellular and extracellular redox
states in response to oxidative stress.2 In vivo, plasma albumin
exerts a buffering action on redox states of plasma thiols.
Perturbation of thiol/disulfide balance is a feature of many
pathological states, especially inflammatory diseases such as
multiple sclerosis.3
Reduced thiols can negatively impact the antitumor activities
of platinum drugs, nitrogen mustards, doxorubicin, and
nitrosoureas by conjugation to electrophilic groups on the
Received: May 26, 2012
Published: November 13, 2012
© 2012 American Chemical Society
10378
dx.doi.org/10.1021/jm300736f | J. Med. Chem. 2012, 55, 10378−10386

Journal of Medicinal Chemistry
Article
Previously we reported the synthesis of 1,4,7,10-tetraazacy-
clododecane-1,4,7,10-tetraacetic acid (DOTA) monoamide
gadolinium(III) complexes 1 and 2 which bear propyl and
hexyl side chains that terminate with a thiol group.10
Complexes 1 and 2 bound to HSA in a demonstrably redox-
sensitive manner. Recently the utility of 2 as a Magnetic
Resonance Imaging (MRI) biomarker of tumor response to
redox-active drugs in mice was demonstrated.11 We also
prepared DO3A gadolinium(III) complexes (not shown)
possessing different ring nitrogen-attached side chains that
also terminate with a thiol group.12 Only one DO3A complex,
bearing a 9-mercapto-3,6-dioxanonyl chain, bound in a redox-
sensitive fashion. Other DO3A complexes additionally bound,
we believe, to surface carboxylates of HSA. The DO3A
complexes exhibited smaller longitudinal water proton
relaxivities (r1) in both free and bound states than did the
DOTA-monoamide complexes. This is presumably due to the
displacement of water molecules from the DO3A complexes by
acetate in the buffer in the free state and by surface carboxylates
in the bound state. In a continuation of our work with Gd(III)-
DOTA-monoamide complexes bearing thiol-terminated side
chains, we have synthesized 3a−3e that form a homologous
series with 1 and 2, studied the relationship of side chain length
and binding to HSA, evaluated the relaxivity of these complexes
in free and bound states, and used the europium analog of 2 in
2D NMR studies to probe binding of the complex to HSA at
the Cys34 site.
structures of all compounds, except 3a−3e, 8a−8e, 10, and 11,
were characterized by chromatographic mobility, 1H NMR, 13
C
NMR, and HRMS. Compounds 3a−3e, 8a−8e, and 10 were
characterized by chromatographic mobility and showed
appropriate masses and isotopic patterns in HRMS analysis.
Compound 11 was characterized by chromatographic mobility,
1H NMR, and HRMS.
Effect of Oxidation State of HSA on Binding Affinity
of Complex 3c to Protein. The binding of 3c to protein was
measured in phosphate-buffered saline (PBS) solutions that
contained 0.6 mM recombinant HSA (rHSA, Invitria),
recombinant lipid-reduced HSA (rHSA-LR, Invitria), HSA
(Sigma), or mercapt-HSA (mHSA) prepared from HSA by the
method of Jacobsen.16 The thiol content of all four HSA
preparations was assayed colorimetrically using Ellman’s
reagent (Figure 1).17 While Ellman’s reagent is known to
react slowly and/or incompletely with thiols in undenatured
proteins,18 our results point to significant relative differences in
thiol oxidation states between the four types of HSA.
The oxidation status of the Cys34 residue in HSA would be
expected to affect the binding of thiol-terminated Gd-DOTA-
monoamide chelates. Figure 2 depicts the binding of 3c to the
four types of HSA in PBS as determined by the method
described previously.12 The calculated association constant
(KA) for each HSA preparation correlates with the free thiol
content of the protein. In addition, the bound fraction was
below the theoretical maximum of one Gd per protein that
would be expected for single-site binding. These findings are
consistent with binding of 3c to HSA at the Cys34 residue.
2D 1H−1H NMR TOCSY Study of the Binding of Thiol-
bearing Complex 11 to HSA. Sadler and colleagues
developed a model for structural changes in HSA upon binding
of drugs to Cys34, whereby the Cys34 moves from a buried to an
exposed environment. They proposed that movement of Cys34
is accompanied by movement of His3, which is part of the N-
terminal segment of the protein, and have reported NMR
methods for monitoring changes at His3 upon binding of a drug
to Cys34.19,20 We used the europium-bearing complex 11 in 2D
NMR studies to study binding of the complex to HSA.
Compound 11 was incubated at 37 °C overnight with either
HSA or mHSA in phosphate-buffered D2O saline, followed by
freeze-drying and reconstitution in D2O for NMR spectrosco-
py. Deuterated solutions of HSA and mHSA without 11 were
prepared analogously. All solutions contained sodium formate
as the chemical shift reference (8.5 ppm). Consistent with
Sadler’s results for binding of the drug Auranofin at Cys34,19 we
observed changes in the splitting of the Hεl resonance of the
His3 residue of HSA at ∼8.2 ppm in 2D 1H−1H NMR TOCSY
spectra that are indicative of the binding of 11 to HSA at the
Cys34 residue (Figure 3). Specifically, the Cys34 residue of HSA
(left panel) is partially disulfide bonded to endogenous NPSH,
resulting in the splitting of the Hεl resonance of the His3
residue at ∼8.2 ppm (left panel expanded inset box). Chemical
reduction of HSA to mHSA (Cys34 in thiol/thiolate form)
results in disappearance of the splitting of this resonance
(center panel expanded inset box). Incubation of 11 with
mHSA (1:1) restores the split resonance (right panel expanded
inset box), indicating that disulfide bond formation has
occurred between 11 and the Cys34 residue.
RESULTS
■
Synthesis. The synthesis of gadolinium complexes 3a−3e
was accomplished by modification of our literature procedure
(Scheme 1).10 Phthalimides 5a−5e were prepared in 81−93%
yields by reaction of the known bromides 4a−4e13−15 with
trityl mercaptan and sodium hydride in anhydrous DMF.
Removal of the phthalimido groups in 5a−5e with hydrazine
hydrate in refluxing ethanol gave the amines 6a−6e in 80−92%
yields. Treatment of DOTA with isobutyl chloroformate,
followed by addition of amines 6a−6e, gave in 70−75% yields
the corresponding tricarboxylic acids 7a−7e. Reaction of 7a−
7e with Gd(OAc)3 in refluxing water gave the Gd(III)
complexes 8a−8e in 85−94% yields. Finally, the trityl groups
were removed with TFA/CH2Cl2 (4% v/v) in the presence of
Et3SiH, affording in 65−91% yields 3a−3e.
The europium-loaded analog of 2 was synthesized starting
from 910 in two steps (Scheme 2). Reaction of 9 with
Eu(OAc)3 in refluxing water gave the Eu(III) complex 10 in
79% yield. Deprotection of the trityl group using TFA/CH2Cl2
(4% v/v) in the presence of Et3SiH gave 11 in 79% yield. The
Additionally, incubation of 11 with HSA results in the
appearance of new AMX systems at ∼4.2 ppm, which are
consistent with the αH-βH2 resonances of free cystine (Figure
4, center panel, solid lines) and the slightly more upfield
10379
dx.doi.org/10.1021/jm300736f | J. Med. Chem. 2012, 55, 10378−10386

Journal of Medicinal Chemistry
Article
Scheme 1. Synthesis of Gd-DOTA-monoamide Complexes 3a−e
Scheme 2. Synthesis of Eu-DOTA-monoamide Complex 11
crosspeaks due either to free cysteine or a mixed disulfide of
free cysteine with 11 (Figure 4, center panel, dashed lines).
These new AMX systems are not present in spectra of HSA by
itself (Figure 4, left panel) and do not appear in spectra of
mHSA incubated with 11 (Figure 4, right panel).
The results in Figure 4 suggest that this batch of HSA
purified from donor plasma (Sigma) carries cysteine bound to
the Cys34 residue, and that this cysteine is replaced and released
by the competitive binding of 11, followed by oxidation of the
free cysteine to cystine or to the mixed disulfide. Any such
small molecule thiol would have been removed during the
preparation of mHSA, explaining the absence of these cross
peaks in the right panel. These resonances are invisible in
spectra of HSA by itself (left panel), most probably due to
broadening arising from attachment to the macromolecule.
Upon release into solution, the sharper crosspeaks of the small
molecule become visible. Taken together, the NMR results in
Figures 3 and 4 provide evidence for the binding of 11 to HSA
at the targeted Cys34 residue.
Figure 1. Ellman’s assay of the thiol content of four types of human
serum albumin: recombinant HSA (rHSA), lipid-reduced recombinant
HSA (rHSA-LR), human serum albumin purified from donor plasma
(HSA), and mercapt-HSA (mHSA) prepared from commercial HSA
by the method of Jacobsen.
Binding and Relaxometry. The binding of compounds 1,
2, and 3a−e to mHSA prepared from commercially obtained
HSA was measured at 37 °C in PBS. The measured KA values
10380
dx.doi.org/10.1021/jm300736f | J. Med. Chem. 2012, 55, 10378−10386

Journal of Medicinal Chemistry
Article
complexes 3d and 3e, indicating that some fraction of the total
binding of these two complexes to mHSA is thiol-sensitive.
The longitudinal water-proton relaxivities (r1) of compounds
1, 2, and 3a−3e were measured at 7 T at 37 °C in PBS in the
absence mHSA to give r1,free and in the presence of mHSA to
give r1,bound (Table 2). As expected, r1,bound was observed to be
higher than r1,free for all complexes, consistent with an increase
in molecular reorientation time upon binding of the Gd
complex to the macromolecular protein. The increase in r1,bound
relative to r1,free was comparable for all complexes.
DISCUSSION
■
There is continuing interest in the development of NMR and
MRI biomarkers of tumor redox status. Recently Brindle et al.
have detected reduction of hyperpolarized [1-13C]-dehydroas-
corbic acid in lymphoma tumors in vivo but no detectable
oxidation of [1-13C]-ascorbic acid in the same tumors,
indicating that these tumors maintain a reduced microenviron-
ment.22 Sherry and co-workers report that a DOTA-tetraamide
europium complex with two quinolinium moieties is silent on
Chemical Exchange Saturation Transfer (CEST) MRI images
in the oxidized form but is activated upon reduction by β-
NADH.23 We have previously demonstrated by dynamic
contrast-enhanced MRI that 1 and 2 are retained in vivo
longer than the comparably sized Gd-DTPA. The in vivo
washout of 1 and 2 is speeded up by an intravenous chase bolus
of homocysteine, consistent with the spontaneous binding of
these complexes to plasma albumin in a reversible redox-
sensitive manner following intravenous administration.10 Aime
and co-workers have reported success in labeling tumor cells in
vitro with a Gd-DO3A-based disulfide complex designed to
bind to protein thiols on the cell surface.24 They have also
demonstrated uptake into cells in xenograft tumors in vivo after
direct injection into the tumor but significantly not following
intravenous injection, possibly due to reaction of the disulfide
with thiols in plasma and on the surface of blood vessels before
it can reach the tumor.25 We recently utilized 2 to provide a
contrast-enhanced MRI biomarker of the response of tumor
xenografts to the thiol-modulating drugs buthionine sulfox-
imine and Imexon.11 In the present work we have investigated
Figure 2. Binding of 3c to rHSA, rHSA-LR, HSA, and mHSA,
measured at 37 °C in PBS. Protein concentration was 0.6 mM, and
data were fitted to a single-site binding model.
for these compounds showed a steady increase in magnitude
with increase in linker length (Table 1). The binding of
complexes 1, 2, and 3a−3c to mHSA was well described by a
single-site binding model. However, the binding of complexes
3d and 3e was not adequately described by a single-site binding
model, and for both of these complexes the Gd:mHSA binding
stoichiometry exceeded 1:1 at complex concentrations of 1
mM. It is possible that the longer hydrophobic linkers in 3d
and 3e allow these complexes to bind additionally to lipid-
binding sites on HSA.21 The binding of compounds 1, 2, and
3a−3c to mHSA could be competitively inhibited by addition
of homocysteine (Table 1). This supports the hypothesis that
disulfide formation between the Gd complex and HSA is an
essential component of binding. A decrease in binding with
increasing homocysteine concentration was also observed for
Figure 3. Downfield regions of the 2D 1H−1H NMR TOCSY spectra of human serum albumin in deuterated buffer at pD ∼7.4. The Cys34 residue
of HSA (left panel) is partially in a disulfide form, resulting in the splitting of the Hεl resonance of the His3 residue at ∼8.2 ppm (expanded inset
box). Chemical reduction of HSA to mHSA (Cys34 in thiol/thiolate form) results in disappearance of the splitting of this resonance (center panel
expanded inset box). Incubation of 11 with mHSA (1:1) restores the split resonance (right panel expanded inset box), indicating disulfide bond
formation between 11 and the Cys34 residue.
10381
dx.doi.org/10.1021/jm300736f | J. Med. Chem. 2012, 55, 10378−10386

Journal of Medicinal Chemistry
Article
Figure 4. Upfield regions of the 2D 1H−1H NMR TOCSY spectra of human serum albumin in deuterated buffer at pD ∼7.4. Incubation of 11 with
HSA (1:1) (center panel) results in the appearance of new AMX systems at 3.2−4.2 ppm (inset box) which are consistent with the αH-βH2
resonances of free cystine (solid lines) as well as either free cysteine or the mixed disulfide of free cysteine and 11 (dashed lines). These new
crosspeaks are not found in the corresponding region of the spectra of HSA by itself (left panel) or when 11 is incubated with mHSA (1:1) (right
panel). This suggests the release of cysteine bound to HSA at Cys34 by competitive binding of 11 to that site, followed by oxidation of the free
cysteine to cystine or the mixed disulfide.
a
Table 1. Inhibition of Binding of Compounds 1, 2, and 3a−3e to 0.6 mM mHSA
homocysteine concentration (mM)
compound
0.0
0.5
1.0
2.0
KA (mM−1
)
1
6.0 1.5
11.3 1.5
16.8 8.3
32.1 17.9
69.8 7.2
2.8 1.2
5.1 1.5
1.2 0.4
1.4 0.4
6.9 2.4
8.5 3.0
15.8 3.3
0.7 0.2
0.4 0.4
4.4 2.1
4.8 1.3
6.6 1.4
3a
3b
2
11.7 5.0
14.7 3.3
32.9 3.5
3c
3d
b
b
b
b
83.4 17.0
b
42.6 23.9
b
26.0 14.5
b
12.8 7.2
b
3e
a
b
Measured at 37 °C. Poor fit to a single-site binding equation, and the Gd:HSA binding stoichiometry exceeds 1:1 at 1 mM [Gd], but binding
decreased as the concentration of homocysteine was increased.
1
Table 2. Longitudinal MRI Relaxivities of Compounds 1, 2,
preparations. 2D H−1H NMR TOCSY provided evidence
and 3a−3e in PBS (r1,free) and when Bound to mHSA
for the covalent binding of 11, the europium-bearing analog of
2, to Cys34 of mHSA. This is taken as evidence for similar
binding behaviors for 1, 2, and 3a−3e with respect to covalent
attachment to Cys34.
a
(r1,bound
)
compound
r1,free (1/mM s)
r1,bound (1/mM s)
1
3.48 0.15
3.01 0.11
3.13 0.04
3.24 0.09
3.75 0.21
3.31 0.12
3.00 0.15
4.62 0.14
4.60 0.55
4.64 0.31
4.78 0.19
5.68 0.57
5.85 0.34
4.73 0.19
Finally, the longitudinal MRI relaxivities of complexes 1, 2,
and 3a−3e are higher when bound to HSA than when free. The
relaxivity of a gadolinium agent depends on a number of
factors: the applied field (B0), the electronic relaxation time
(T1e), the correlation time for rotational motion (τR), the inner-
sphere hydration number (q), the rate of water (or water-
proton) exchange at the metal ion (i.e., kex = 1/τm for the inner-
sphere), the distance between the Gd(III) ion and the water-
proton (rGd−H), and corresponding parameters for second-
sphere and outer-sphere relaxation.26,27 The observed higher
values of r1,bound of our complexes relative to the corresponding
values of r1,free are consistent with a slowing down of the τR of
the Gd−H axis upon binding to the macromolecule.
3a
3b
2
3c
3d
3e
a
Measured at 37 °C and 7 T.
the influence of alkyl linker length on the HSA binding
properties of the set of C3−C9 homologues, including 1 and 2.
The binding of complexes 1, 2, and 3a−3c to HSA was well-
described by single-site binding equations, while the KA values
for complexes 3d and 3e with C-8 and C-9 linkers, respectively,
were higher than predicted and indicative of multisite binding.
The binding of all complexes to HSA was inhibited by
homocysteine, suggesting that covalent attachment by for-
mation of a disulfide link with Cys34 is an important component
of binding.
CONCLUSIONS
■
The difference between r1,bound and r1,free of DOTA-monoamide
thiol complexes of gadolinium can be exploited to produce
enhancement in MRI images that is dependent on the local
redox status. We have recently demonstrated that the
shortening of MRI longitudinal relaxation times (T1) of
xenograft tumors after injection of 2 was greater in mice that
were treated with the thiol-oxidizing drug Imexon relative to
The measured KA for binding of 3c to four different HSA
preparations was not identical, and this could be explained by
differences in the Cys34 thiol oxidation states of the
10382
dx.doi.org/10.1021/jm300736f | J. Med. Chem. 2012, 55, 10378−10386

Journal of Medicinal Chemistry
Article
saline treatment.11 High tumor levels of thiols can decrease
tumor sensitivity to certain drugs and to radiotherapy. An MRI
method for post-therapy assessment of tumor response to
redox-modulating therapies would assist in the development of
novel redox-modulating drugs for preadministration or
coadministration with chemotherapy or radiotherapy. The
results reported here indicate that 3c would perform better
than 2 as an MRI biomarker of tumor response to redox-active
therapies due to its superior thiol-sensitive binding affinity for
HSA. To improve upon 2 and 3c, we are pursuing strategies to
increase r1,bound by speeding up τm at the HSA-bound
gadolinium. We are also exploring strategies to quench r1,bound
by increasing rGd−H when the complex is bound to HSA, with a
view to creating an MRI contrast agent that will highlight only
highly reducing tissue microenvironments.
2-(5-(Tritylthio)pentyl)isoindoline-1,3-dione (5b). Yield 81%,
white solid, mp 124−125 °C, Rf 0.57 (EtOAc/hexanes 3:7). 1H NMR
(500 MHz, CDCl3) δ 1.28 (2, quintet, J = 7.0 Hz), 1.41 (2, quintet, J =
7.0 Hz), 1.54 (2, quintet, J = 7.0 Hz), 2.13 (2, t, J = 7.0 Hz), 3.59 (2, t,
J = 7.0 Hz), 7.15−7.20 (3, m), 7.22−7.28 (6, m), 7.35−7.40 (6, m),
7.68−7.70 (2, dd, J = 5.0 Hz, 3 Hz), 7.81−7.83 (2, dd, J = 5.5 Hz, 3
Hz); 13C NMR (125 MHz, CDCl3) δ 26.3, 28.2, 28.3, 31.7, 37.8, 66.4,
123.1, 126.4, 127.7, 129.5, 132.1, 133.8, 144.9, 168.2; HRMS (ESI+,
m/z) calculated for C32H29NNaO2S 514.1811, found 514.1807 (M +
Na)+.
2-(8-(Tritylthio)octyl)isoindoline-1,3-dione (5d). Yield 91%,
tan solid, mp 68−71 °C, Rf 0.58 (EtOAc/hexanes 3:7). 1H NMR (500
MHz, CDCl3) δ 1.20−1.28 (8, m), 1.32−1.38 (2, m), 1.60−1.65 (2,
m), 2.11 (2, t, J = 7.0 Hz), 3.65 (2, t, J = 7.0 Hz), 7.17−7.21 (3, m),
7.24−7.28 (6, m), 7.37−7.41 (6, m), 7.68−7.71 (2, m), 7.81−7.84 (2,
m); 13C NMR (125 MHz, CDCl3) δ 26.7, 28.5, 28.8, 28.9, 31.9, 38.0,
66.3, 123.0, 126.3, 127.6, 129.5, 132.1, 133.7, 145.0, 168.2; HRMS
(ESI+, m/z) calculated for C35H35NNaO2S 556.2281, found 556.2275
(M + Na)+.
EXPERIMENTAL SECTION
■
2-(9-(Tritylthio)nonyl)isoindoline-1,3-dione (5e). Yield 89%,
1
Synthesis. Reactions were conducted with flame-dried glassware
under a positive pressure of argon. Hydroscopic solvents were
transferred via an oven-dried syringe or cannula. All reagents and
solvents were commercially available and were used as received.
Solutions were concentrated in vacuo using a rotary evaporator.
Analytical thin-layer chromatography (TLC) was performed on
precoated silica gel 60 F-254 glass plates. TLC visualization required
using UV light and/or staining. Anisaldehyde stain (100 mL
anisaldehyde, 50 mL glacial AcOH, 100 mL conc H2SO4, 1 L 95%
EtOH) and PMA stain (5 g phosphomolybdic acid, 100 mL 95%
EtOH) were the most commonly used TLC stains. Flash and gravity
chromatography were performed using silica gel 60 (230−400 mesh).
Melting points are uncorrected. Nuclear magnetic resonance (NMR)
experiments were performed on a 500 MHz spectrometer. NMR
spectra were referenced to TMS (0.00 ppm), CDCl3 (7.24 ppm, 77.0
ppm), CD3OD (3.31 ppm), or CH3CN in D2O (119.68). Mass
spectrometry was conducted using ESI. Free gadolinium content in the
final compounds was analyzed by means of the xylenol orange test28
and found to be negative. Purities of the final compounds were >95%
and were determined by RP-HPLC analysis (see the Supporting
Information).
white solid, mp 59−61 °C, Rf 0.56 (EtOAc/hexanes 3:7). H NMR
(500 MHz, CDCl3) δ 1.10−1.38 (12, m), 1.64 (2, m), 2.11 (2, t, J =
7.0 Hz), 3.65 (2, t, J = 7.0 Hz), 7.18−7.21 (3, m), 7.25−7.30 (6, m),
7.39−7.43 (6, m), 7.69 (2, m), 7.83 (2, m); 13C NMR (125 MHz,
CDCl3) δ 26.8, 28.5, 28.9, 29.0, 29.2, 31.9, 38.0, 66.3, 123.1, 126.4,
127.7, 127.8, 129.5, 133.7, 145.0, 168.4; HRMS (ESI+, m/z) calculated
for C36H37NNaO2S 570.2429, found 570.2437 (M + Na)+.
Procedure for Synthesis of Tritylthioalkylamines 6a−6e: 7-
(Tritylthio)heptan-1-amine (6c). To a suspension of 5c (6.0 g, 11.5
mmol) in absolute ethanol (75 mL) was added hydrazine hydrate
(1.84 g, 57.8 mmol, 1.8 mL) and the mixture was heated to reflux.
After 5 h, the mixture was filtered and the filtrate concentrated under
reduced pressure. CHCl3 (100 mL) was added to the residue, the
mixture stirred for 0.5 h, and filtered. The precipitate was washed with
CHCl3 (100 mL) and the combined organic fractions were washed
with water (3 × 100 mL), dried (MgSO4), filtered, and concentrated.
The crude was purified by flash chromatography on silica gel eluted
with CH2Cl2/MeOH/aq NH4OH (5:3:0.1) to afford 6c (3.6 g, 9.2
mmol, 80%) as a white solid, mp 74−76 °C, Rf 0.36 (CH2Cl2/MeOH/
1
aq NH4OH 5:3:0.1). H NMR (500 MHz, CDCl3) δ 1.13−1.27 (6,
m), 1.37 (4, m), 2.12 (2, t, J = 7.5 Hz), 2.63 (2, t, J = 7.0 Hz), 7.17−
7.20 (3, m), 7.25−7.28 (6, m), 7.40 (6, m); 13C NMR (125 MHz,
CDCl3) δ 26.6, 28.5, 28.9, 29.0, 31.9, 33.7, 42.2, 66.3, 126.4, 127.7,
129.5, 145.0; HRMS (ESI+, m/z) calculated for C26H32NS 390.2250,
found 390.2249 (M + H)+.
Procedure for Synthesis of Tritylthioalkylphthalimides 5a−
5e: 2-(7-(Tritylthio)heptyl)isoindoline-1,3-dione (5c). To a
suspension of NaH (0.59 g, 24.7 mmol) in dry DMF (25 mL) at 0
°C under argon was added triphenylmethanethiol (6.8 g, 24.7 mmol)
portionwise with cooling and under an argon atmosphere. After 0.5 h,
a solution of bromide 4c (8.0 g, 24.7 mmol) in DMF (25 mL) was
added dropwise with stirring. The reaction mixture was stirred for
another 5 min, brought to rt, and was stirred overnight. Ethyl acetate
(200 mL) was added to the reaction mixture and the organic phase
was washed with water (3 × 100 mL), brine (50 mL), dried (MgSO4),
and filtered. Concentration under reduced pressure gave 5c (10.0 g,
19.3 mmol, 83%) as a white solid, mp 72−73 °C, Rf 0.62 (EtOAc/
4-(Tritylthio)butan-1-amine (6a). Yield 86%, white solid, mp
1
88−92 °C, Rf 0.46 (CH2Cl2/MeOH/aq NH4OH 5:3:0.1). H NMR
(500 MHz, CDCl3) δ 1.20−1.48 (4, m), 2.16 (2, t, J = 7.0 Hz), 2.53
(2, t, J = 7.0 Hz), 7.17−7.21 (3, m), 7.24−7.28 (6, m), 7.38−7.43 (6,
m); 13C NMR (125 MHz, CDCl3) δ 31.6, 32.8, 41.4, 66.2, 126.3,
127.6, 129.3, 144.7; HRMS (ESI+, m/z) calculated for C23H25NNaS
370.1600, found 370.1604 (M + Na)+.
5-(Tritylthio)pentan-1-amine (6b). Yield 86%, white solid, mp
1
1
hexanes 3:7). H NMR (500 MHz, CDCl3) δ 1.15−1.28 (8, m), 1.36
85−89 °C, Rf 0.32 (CH2Cl2/MeOH/aq NH4OH 5:3:0.1); H NMR
(2, m), 1.61 (2, quintet, J = 7.5 Hz), 2.11 (2, t, J = 7.5 Hz), 3.63 (2, t, J
= 7.5 Hz), 7.17−7.21 (3, m), 7.24−7.28 (6, m), 7.38−7.41 (6, m),
7.69 (2, dd, J = 5.4 Hz, 3.0 Hz), 7.83 (2, dd, J = 5.4 Hz, 3.0 Hz); 13C
NMR (125 MHz, CDCl3) δ 26.6, 28.4, 28.5, 28.7, 28.8, 31.9, 37.7,
40.2, 66.3, 123.1, 126.4, 127.7, 129.5, 132.1, 133.7, 145.0, 168.3;
HRMS (ESI+, m/z) calculated for C34H33NNaO2S 542.2124, found
542.2129 (M + Na)+.
2-(4-(Tritylthio)butyl)isoindoline-1,3-dione (5a). Yield 93%,
white solid, mp 126−127 °C, Rf 0.55 (EtOAc/hexanes 3:7). 1H NMR
(500 MHz, CDCl3) δ 1.40 (2, quintet, J = 7.0 Hz), 1.62 (2, quintet, J =
7.0 Hz), 2.17 (2, t, J = 7.0 Hz), 3.55 (2, t, J = 7.0 Hz), 7.00−7.18 (3,
m), 7.22−7.27 (6, m), 7.35−7.42 (6, m), 7.65−7.72 (2, dd, J = 5.5 Hz,
3 Hz), 7.79−7.83 (2, dd, J = 5.5 Hz, 3 Hz); 13C NMR (125 MHz,
CDCl3) δ 25.9, 28.0, 31.4, 37.5, 66.4, 123.1, 126.4, 127.7, 129.3, 129.5,
132.0, 133.8, 144.8, 168.1; HRMS (ESI+, m/z) calculated for
C31H27NNaO2S 500.1655, found 500.1648 (M + Na)+.
(500 MHz, CDCl3) δ 1.21−1.32 (2, m), 1.35−1.48 (4, m), 2.15 (2, t, J
= 7.0 Hz), 2.58 (2, t, J = 7.0 Hz), 7.16−7.21 (3, m), 7.24−7.28 (6, m),
7.38−7.42 (6, m); 13C NMR (125 MHz, CDCl3) δ 26.3, 28.4, 31.9,
33.2, 41.9, 66.4, 126.4, 127.7, 129.5, 144.9; HRMS (ESI+, m/z)
calculated for C24H27NNaS 384.1762, found 384.1760 (M + Na)+.
8-(Tritylthio)octan-1-amine (6d). Yield 92%, Light yellow solid,
mp 45−46 °C, Rf 0.33 (CH2Cl2/MeOH/aq NH4OH 5:3:0.1). 1H
NMR (500 MHz, CDCl3) δ 1.10−1.30 (10, m), 1.34−1.44 (2, m),
2.13 (2, t, J = 7.0 Hz), 2.65 (2, t, J = 7.0 Hz), 7.17−7.22 (3, m), 7.24−
7.29 (6, m), 7.38−7.43 (6, m); 13C NMR (125 MHz, CDCl3) δ 28.6,
28.9, 29.1, 29.2, 32.0, 33.6, 42.2, 66.3, 126.4, 127.7, 129.5, 145.0;
HRMS (ESI+, m/z) calculated for C27H34NS 404.2406, found
404.2406 (M + H)+.
9-(Tritylthio)nonan-1-amine (6e). Yield 81%, viscous oil, Rf 0.29
1
(CH2Cl2/MeOH/aq NH4OH 5:3:0.1). H NMR (500 MHz, CDCl3)
δ 1.10−1.30 (10, m), 1.35−1.45 (4, m), 2.12 (2, t, J = 7.5 Hz), 2.66 (2,
10383
dx.doi.org/10.1021/jm300736f | J. Med. Chem. 2012, 55, 10378−10386

Journal of Medicinal Chemistry
Article
t, J = 7.0 Hz), 7.17−7.22 (3, m), 7.24−7.29 (6, m), 7.38−7.43 (6, m);
13C NMR (125 MHz, CDCl3) δ 28.8. 28.5, 29.9, 29.1, 29.3, 32.0, 33.7,
concentrated in vacuo, and the residue loaded onto a flash silica gel
column. Elution with CH2Cl2/MeOH/aq NH4OH (5:3:0.5) followed
by removal of solvents gave a viscous oil which was dissolved in a
minimum amount of water and freeze-dried. By this procedure 8c
(0.80 g, 0.86 mmol, 94%) was obtained as a white fluffy solid, 256−
260 °C (dec), Rf 0.54 (CH2Cl2/MeOH/aq NH4OH 5:3:1). HRMS
(ESI+, m/z) calculated for C42H55GdN5O7S 931.3066, found 931.3067
(M + H)+.
{1,4,7-Tris(carboxymethyl)-10-[N-(4-trityl-thiobutyl)-
carbamoyl]-1,4,7,10-tetraazacyclododecanato}gadolinium
(8a). Yield 85%, a white fluffy solid, 264−268 °C (dec), Rf 0.55
(CH2Cl2/MeOH/aq NH4OH 5:3:1). HRMS (ESI+, m/z) calculated
for C39H49GdN5O7S 889.2588, found 889.2581 (M + H)+.
{1,4,7-Tris(carboxymethyl)-10-[N-(5-trityl-thiopentyl)-
carbamoyl]-1,4,7,10-tetraazacyclododecanato}gadolinium
(8b). Yield 91%, a white fluffy solid, 247−250 °C (dec), Rf 0.54
(CH2Cl2/MeOH/aq NH4OH 5:3:1). HRMS (ESI+, m/z) calculated
for C40H51GdN5O7S 903.2752, found 903.2740.
{1,4,7-Tris(carboxymethyl)-10-[N-(8-trityl-thiooctyl)-
carbamoyl]-1,4,7,10-tetraazacyclododecanato}gadolinium
(8d). Yield 89%, a white fluffy solid, 230−234 °C (dec), Rf 0.53
(CH2Cl2/MeOH/aq NH4OH 5:3:1). HRMS (ESI+, m/z) calculated
for C43H57GdN5O7S 945.3214, found 942.3217 (M + H)+.
42.2, 66.3, 126.4, 127.7, 129.5, 145.0; HRMS (ESI+, m/z) calculated
for C28H36NS 418.2563, found 418.2557 (M + H)+.
Procedure for the Synthesis of DOTA-based Tricarboxylic
Acids 7a−7e: [1,4,7-Tris(carboxymethyl)-10-[N-(7-trityl-
thioheptyl)carbamoyl]-1,4,7,10-tetraazacyclododecane (7c).
1,4,7,10-Tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA,
2.62 g, 6.0 mmol) was dissolved in water (3 mL), lyophilized, and
dissolved in DMF/TEA (4:1, v/v, 150 mL) and stirred at rt for 24 h.
Isobutyl chloroformate (0.41 g, 3 mmol, 0.39 mL) was then added to
the reaction mixture at 0 °C. After 1 h, amine 6c (0.78 g, 2.0 mmol)
was added and the reaction mixture stirred. After 48 h, the solvent was
removed under reduced pressure and the crude solid was loaded onto
a gravity silica gel column. Elution with CH2Cl2/MeOH/aq NH4OH
(5:3:0.3) followed by removal of solvents gave a viscous oil which was
dissolved in a minimal amount of water and freeze-dried to afford 7c
(1.08 g, 1.39 mmol, 70%), which was obtained as a white fluffy solid,
mp 200−202 °C (dec), Rf 0.57 (CH2Cl2/MeOH/aq NH4OH 5:3:1).
1H NMR (500 MHz, CD3OD) δ 1.15−1.28 (6, m), 1.35 (2, quintet, J
= 7.5 Hz), 1.45 (2, m), 2.12 (4, m), 2.29 (4, m), 2.85 (8, m), 3.12 (8,
m), 3.5−3.8 (4, m), 7.18−7.23 (3, m), 7.24−7.29 (6, m), 7.36−7.39
(6, m); 13C NMR (125 MHz, D2O) δ 27.0, 29.3, 32.4, 40.1, 48.5, 59.5,
61.6, 66.8, 127.1, 128.4, 129.0, 145.4, 174.0, 180.2; HRMS (ESI+, m/z)
calculated for C42H58N5O7S 776.4051, found 776.4052 (M + H)+.
[1,4,7-Tris(carboxymethyl)-10-[N-(4-trityl-thiobutyl)-
carbamoyl]-1,4,7,10-tetraazacyclododecane (7a). Yield 74%,
white fluffy solid, 210−214 °C (dec), Rf 0.56 (CH2Cl2/MeOH/aq
NH4OH 5:3:1). 1H NMR (500 MHz, CD3OD) δ 1.35 (2, quintet, J =
7.5 Hz), 1.43 (2, m), 2.17 (4, t, J = 7.0 Hz), 2.28 (4, m), 2.84 (8, m),
3.2 (8, m), 3.50−3.75 (4, m), 7.19−7.23 (3, m), 7.26−7.30 (6, m),
7.36−7.39 (6, m); 13C NMR (125 MHz, D2O) δ 26.4, 28.5, 39.8, 48.8,
59.3, 61.5, 66.9, 127.2, 128.4, 129.9, 145.3, 174.1, 180.3. HRMS (ESI+,
m/z) calculated for C39H52N5O7S 734.3582, found 734.3586 (M +
H)+.
{1,4,7-Tris(carboxymethyl)-10-[N-(9-trityl-thiononyl)-
carbamoyl]-1,4,7,10-tetraazacyclododecanato}gadolinium
(8e). Yield 89%, a white fluffy solid, 218 °C (dec), Rf 0.53 (CH2Cl2/
MeOH/aq NH4OH 5:3:1). HRMS (ESI+, m/z) calculated for
C44H59GdN5O7S 959.3379, found 959.3396 (M + H)+.
Procedure for Synthesis of Gadolinium Complexes 3a−3e:
{1,4,7-Tris(carboxymethyl)-10-[N-(7-mercaptoheptyl)-
carbamoyl]-1,4,7,10-tetraazacyclododecanato}gadolinium
(3c). To a solution of 8c (0.82 g, 0.88 mmol) and Et3SiH (0.12 g, 1.01
mmol, 0.16 mL) in CH2Cl2 (30 mL) was added TFA (1.2 mL, 4% v/
v) and the mixture was stirred at room temperature. After 10 min,
volatiles were removed in vacuo, the residue was dissolved in EtOAc
(20 mL), and the solution was extracted with water (2 × 10 mL). The
organic layer was discarded and the aqueous layers were lyophilized.
Flash chromatography of the residue on reverse phase silica gel (H2O/
CH3CN 6:4), followed by lyophilization, gave 0.55 g (0.80 mmol,
91%) of the free thiol 3c as a white fluffy solid, 310 °C (dec), Rf 0.51
(H2O/CH3CN 6:4). HRMS (ESI+, m/z) calculated for
C23H41GdN5O7S 689.1965, found 689.1956 (M + H)+.
{1,4,7-Tris(carboxymethyl)-10-[N-(4-mercaptobutyl)-
carbamoyl]-1,4,7,10-tetraazacyclododecanato}gadolinium
(3a). Yield 82%, a white fluffy solid, 230 °C (dec), Rf 0.68 (H2O/
CH3CN 6:4). HRMS (ESI+, m/z) calculated for C21H35GdN5O7S
647.1495, found 647.1491 (M + H)+.
{1,4,7-Tris(carboxymethyl)-10-[N-(5-mercaptopentyl)-
carbamoyl]-1,4,7,10-tetraazacyclododecanato}gadolinium
(3b). Yield 78%, a white fluffy solid, 200 °C (dec), Rf 0.63 (H2O/
CH3CN 6:4). HRMS (ESI+, m/z) calculated for C21H35GdN5O7S
661.1652, found 661.1644 (M + H)+.
{1,4,7-Tris(carboxymethyl)-10-[N-(8-mercaptooctyl)-
carbamoyl]-1,4,7,10-tetraazacyclododecanato}gadolinium
(3d). Yield 65%, a white fluffy solid, 303 °C (dec), Rf 0.42 (H2O/
CH3CN 6:4). HRMS (ESI+, m/z) calculated for C24H43GdN5O7S
703.2122, found 703.2126 (M + H)+.
{1,4,7-Tris(carboxymethyl)-10-[N-(8-mercaptooctyl)-
carbamoyl]-1,4,7,10-tetraazacyclododecanato}gadolinium
(3e). Yield 89%, a white fluffy solid, 190 °C (dec), Rf 0.39 (H2O/
CH3CN 6:4). HRMS (ESI+, m/z) calculated for C25H45GdN5O7S
717.2278, found 717.2273 (M + H)+.
{1,4,7-Tris(carboxymethyl)-10-[N-(6-trityl-thiohexyl)-
carbamoyl]-1,4,7,10-tetraazacyclododecanato}europium (10).
To a solution of 910 (0.58 g, 0.76 mmol) in water (10 mL) was added
Eu(OAc)3 (0.25 g, 0.76 mmol) and the mixture heated to reflux. After
2 h, the reaction mixture was cooled, the volatiles were removed in
vacuo, and the residue was loaded on a flash silica gel column. Elution
with CH2Cl2/MeOH/aq NH4OH (5:3:0.5) followed by removal of
solvents gave a viscous oil which was dissolved in a minimum amount
of water and freeze-dried. By this procedure, 10 (0.54 g, 0.60 mmol,
79%) was obtained as a white fluffy solid, 254 °C (dec), Rf 0.54
[1,4,7-Tris(carboxymethyl)-10-[N-(5-trityl-thiopentyl)-
carbamoyl]-1,4,7,10-tetraazacyclododecane (7b). Yield 74%,
white fluffy solid, 204−208 °C (dec), Rf 0.54 (CH2Cl2/MeOH/aq
1
NH4OH 5:3:1). H NMR (500 MHz, CD3OD) δ 1.26 (2, m), 1.33−
1.42 (4, m), 2.16 (4, m), 2.28 (4, m), 2.85 (8, m), 3.26 (8, m), 3.50−
3.75 (4, m), 7.19−7.23 (3, m), 7.26−7.30 (6, m), 7.36−7.39 (6, m);
13C NMR (125 MHz, D2O) δ 26.38, 28.6, 32.4, 39.7, 48.4, 57.5, 59.4,
66.7, 127.1, 128.3, 129.8, 145.3, 173.9, 180.1; HRMS (ESI+, m/z)
calculated for C40H54N5O7S 748.3738, found 748.3743 (M + H)+.
[1,4,7-Tris(carboxymethyl)-10-[N-(8-trityl-thiooctyl)-
carbamoyl]-1,4,7,10-tetraazacyclododecane (7d). Yield 75%,
white fluffy solid, 174 °C (dec), Rf 0.5 (CH2Cl2/MeOH/aq
1
NH4OH 5:3:0.5). H NMR (500 MHz, CD3OD) δ 1.14−1.28 (10,
m), 1.34 (2, m), 1.47 (2, m), 2.12 (2, m), 2.31 (4, m), 2.85 (8, m),
3.10−3.16 (8, m), 3.50−3.80 (4, m), 7.18−7.31 (9, m), 7.36−7.41 (6,
m); 13C NMR (125 MHz, D2O) δ 23.6, 27.3, 28.9, 29.4, 32.3, 39.6,
48.4, 48.7, 51.7, 52.3, 54.7, 56.1, 57.3, 61.7, 66.7, 127.0, 128.3, 129.8,
145.3, 170.2, 171.0, 177.8; HRMS (ESI+, m/z) calculated for
C43H60N5O7S 790.4208, found 790.4195 (M + H)+.
[1,4,7-Tris(carboxymethyl)-10-[N-(9-trityl-thiononyl)-
carbamoyl]-1,4,7,10-tetraazacyclododecane (7e). Yield 72%,
white fluffy solid, 176−180 °C (dec), Rf 0.51 (CH2Cl2/MeOH/aq
1
NH4OH 5:3:0.5). H NMR (500 MHz, CD3OD) δ 1.12−1.37 (14,
m), 1.49 (2, m), 2.11 (2, t, J = 7.5 Hz), 3.00 (8, m), 3.17 (3, m), 3.25−
3.50 (9, m), 3.66 (4, m), 7.28−7.21 (3, m), 7.24−7.28 (6, m), 7.37−
7.40 (6, m); 13C NMR (125 MHz, D2O) δ 27.4, 29.0, 29.5, 29.8, 32.4,
39.7, 48.8, 51.8, 52.3, 57.4, 66.8, 127.1, 128.3, 129.8, 145.4, 170.2,
171.0, 178.0; HRMS (ESI+, m/z) calculated for C44H62N5O7S
804.4364, found 804.4364 (M + H)+.
Procedure for Synthesis of Gadolinium Complexes 8a−8e:
{1,4,7-Tris(carboxymethyl)-10-[N-(7-trityl-thioheptyl)-
carbamoyl]-1,4,7,10-tetraazacyclododecanato}gadolinium
(8c). To a solution of 7c (0.75 g, 0.96 mmol) in water (10 mL) was
added Gd(OAc)3 (0.32 g, 0.96 mmol) and the mixture was heated to
reflux. After 2 h, the reaction mixture was cooled, the solvent
10384
dx.doi.org/10.1021/jm300736f | J. Med. Chem. 2012, 55, 10378−10386

Journal of Medicinal Chemistry
Article
(CH2Cl2/MeOH/aq NH4OH 5:3:1). HRMS (ESI+, m/z) calculated
for C41H53EuN5O7S 912.2872, found 912.2863 (M + H)+.
braking time. It was assumed that gadolinium bound to HSA (Gdbound)
would not pass through the membrane and that gadolinium in the
filtrate (Gdfree) accurately represented the unbound gadolinium in
each sample. Gadolinium concentrations in the filtrates were
determined by MRI relaxometry. The apparent equilibrium binding
constant (KA) for each complex was calculated from the Law of Mass
Action as described previously.12
{1,4,7-Tris(carboxymethyl)-10-[N-(6-mercaptohexyl)-
carbamoyl]-1,4,7,10-tetraazacyclododecanato}europium (11).
To a solution of 10 (0.33 g, 0.36 mmol) and Et3SiH (48.0 μg, 0.41
mmol, 65 μL) in CH2Cl2 (15 mL) was added TFA (0.6 mL, 4% v/v)
and the mixture was stirred at room temperature. After 10 min,
volatiles were removed in vacuo, the residue was dissolved in EtOAc
(20 mL), and the solution was extracted with water (2 × 10 mL). The
organic layer was discarded and the aqueous layers were lyophilized.
Flash chromatography of the residue on reverse phase silica gel (H2O/
CH3CN 6:4), followed by lyophilization, gave 0.19 g (0.28 mmol,
79%) of the free thiol 11 as a white fluffy solid, 205 °C (dec), Rf 0.51
(H2O/CH3CN 6:4). HRMS (ESI+, m/z) calculated for
C22H39EuN5O7S 670.1770, found 670.1777 (M + H)+.
Preparation of Mercapt-HSA (mHSA). To a 1.0 mM solution of
HSA (5.2 g, 0.08 mmol) in 0.1 M sodium phosphate buffer (80 mL,
pH 6.86) containing 0.3 M NaCl (1.4 g, 2.38 mmol) was added
dithiothreitol (DTT, 62 mg, 0.4 mmol) and the mixture was stirred at
rt. After 45 min, the mixture was loaded into a dialysis membrane and
dialyzed at 4 °C against 0.05 M TRIS buffer (pH 7.3) for 10 days and
against distilled water for 5 days. Lyophilization gave mHSA as a fluffy
solid.
Ellman’s Assay of HSA. The four types of HSA were made up as
0.67 mM solutions in PBS buffer adjusted to pH 8.0 and kept on ice
until use. Ellman’s reagent (DTNB) was prepared as a 4 mM stock
solution in PBS buffer at pH 8.0 and kept on ice wrapped in aluminum
foil until use. A 10 mM stock solution of L-cysteine (Sigma) was used
as a standard and was prepared in PBS buffer at pH 8.0 and kept on ice
until use. A 96-well plate was kept on ice and loaded with the various
types of HSA (0.57 mM final concentration) or L-cysteine (0−0.3 mM
final concentration) and DTNB (0.57 mM final concentration). The
96-well plate was then transferred to an optical absorbance plate reader
and the optical absorbance at 412 nm measured at room temperature.
Quadruplicate data points were determined and averaged.
Relaxivity of Gd Complexes in the Absence of HSA.
Measurements of the longitudinal water-proton relaxivities of Gd
complexes in buffered saline were made at 37 °C on a 7 T Bruker
Biospec MR Instrument (Bruker Biospin, Billerica, MA). 96-Well
tissue culture plates (Falcon) were cut down to 6 × 8 wells, creating a
sample tray which fit inside a 72 mm ID birdcage radio frequency
transmitter-receiver coil (Bruker). Aliquots (250 μL) of the
gadolinium solutions were loaded into these wells. Spaces between
the wells were filled with water in order to reduce air−water
susceptibility artifacts in the images. The sample tray was maintained
at 37 °C during imaging by flowing heated air over the sample tray.
Sample temperature was continuously monitored using a fluoroptic
temperature probe (Luxtron Corporation, Santa Clara, CA). Spin−
echo MR images of cross sections of the wells were acquired with
recycle times (TR) ranging between 35 and 8000 ms and an echo time
(TE) of 4 ms. Signal intensity S in each well was fitted to eq 1 to
extract the T1 at each solution condition. The factor c was between
0.99 and 1.0 in all regressions.
⎛
⎝
⎞
⎠
⎡
⎤
−T
R
⎜
⎟
⎟
S = S 1 − c × exp⎢
⎥
⎜
0
T
⎣
⎦
1
(1)
The T1 relaxation times of the solutions of gadolinium in saline
containing 0−2.0 mM homocysteine thus calculated were then fit to
eq 2, and a longitudinal relaxivity r1 obtained for each solution
condition.
1
1
T
10
=
+ r × [Gd]
1
NMR Studies of Interactions between Compound 11 and
HSA. The 1D and 2D NMR experiments were carried out using a
Bruker DRX600 spectrometer at 310 K. The 1D 1H spectra with 9000
Hz spectral width and 8 k time domain data points were acquired
using presaturation for residual water suppression. The data were zero-
filled to 32 k, apodized with an optimized combination of squared-
sinebell and Gaussian functions for resolution enhancement, and
T
(2)
1
Here 1/T10 is the relaxation rate in the absence of contrast agent, and
[Gd] is the concentration of gadolinium in the solution.
Relaxivity of Gd Complexes in the Presence of HSA. The
relaxivity of gadolinium in HSA-containing solutions was approxi-
mated to have only two components: “free” (monomer, homodimer,
heterodimer), and “bound” (to HSA). The T1 times of the solutions of
gadolinium complex in saline + HSA were inserted into eq 3 and fitted
for the relaxivity of bound gadolinium, r1,bound
1
Fourier transformed. The 2D H−1H spectra were usually acquired
with a spectral width of 9000 Hz, 512 t1 points, 2K data points in t2,
and 96 or 112 transients for each t1 point, and all 2D spectra were
recorded in pure absorption with time-proportional phase incremen-
tation (TPPI). The residual solvent resonance in D2O was suppressed
by selective irradiation during the preparation period. The 2D 1H−1H
TOCSY experiments used a mixing time of 65 ms; a delay, 2.4 times
the 90° pulse, was inserted between each pulse of each block of the
MLEV-17 spin lock period. Conventional pulse sequences were used
for DQF-COSY spectra. All 1D and 2D spectra were processed on a
Unix computer using TopSpin 2.0 (Bruker). Prior to 2D Fourier
transformation, the data were multiplied by squared sine bells in both
t1 and t2 and zero-filled to yield a 2048 × 2048 matrix of real data
points.
1
1
T
10
=
+ r
× [Gdfree] + r
× [Gdtotal − Gdfree]
1,free
1,bound
T
1
(3)
where Gdfree was obtained as described previously10 with KA
constrained to equal the binding constant calculated for the respective
HSA-containing solution. In these regressions, r1,free was constrained to
equal the relaxivity measured in the corresponding HSA-free solution.
ASSOCIATED CONTENT
■
Binding of Gd-thiols to HSA. Solutions of Gd complexes 1, 2,
and 3a−3e were made in PBS, and in corresponding solutions that
contained 0.6 mM human serum albumin from donor plasma (HSA,
Sigma), recombinant HSA (rHSA, Invitria), recombinant lipid-reduced
HSA (rHSA-LR, Invitria), or mercapt-HSA (mHSA) prepared from
HSA obtained from Sigma, as well as 0−2.0 mM homocysteine
(Sigma), to final gadolinium concentrations of 0−1.0 mM. All
solutions also contained 10 mM sodium azide, and the final pH of all
S
* Supporting Information
1H and 13C NMR spectra for compounds 5a−5e, 6a−6e, 7a−
7e, 11, full size 2D 1H−1H NMR TOCSY spectra, and
analytical RP-HPLC chromatograms for 3a−3e. This material
is available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
solutions was 7.40
0.05 at room temperature. Solutions were
■
allowed to equilibrate overnight at 37 °C under air atmosphere prior
to measurements. Aliquots (500 μL) of each solution containing HSA
were placed in prewarmed ultrafiltration units (Amicon UItra-4
Centrifugal Filter Units, 30000 MW cutoff, Millipore Corporation)
and immediately centrifuged at 7450× g for 10.7 min, inclusive of
Corresponding Author
*Phone (520) 626-6041; e-mail raghunan@email.arizona.edu.
Notes
The authors declare no competing financial interest.
10385
dx.doi.org/10.1021/jm300736f | J. Med. Chem. 2012, 55, 10378−10386

Journal of Medicinal Chemistry
Article
(12) Raghunand, N.; Guntle, G. P.; Gokhale, V.; Nichol, G. S.; Mash,
E. A.; Jagadish, B. Design, synthesis, and evaluation of 1,4,7,10-
tetraazacyclododecane-1,4,7-triacetic acid derived. redox-sensitive
contrast agents for magnetic resonance imaging. J. Med. Chem. 2010,
53, 6747−6757.
ACKNOWLEDGMENTS
■
This work was supported by grants R01-CA118359, P01-
CA017094, and P30-CA023074 from the National Institutes of
Health.
(13) Ke, C.; Yang, C.; Mori, T.; Wada, T.; Liu, Y.; Inoue, Y. Catalytic
enantiodifferentiating photocyclodimerization of 2-anthracenecarbox-
ylic acid mediated by a non-sensitizing chiral metallosupramolecular
host. Angew. Chem., Int. Ed. 2009, 48, 6675−6677.
(14) Choi, J. Y.; Seo, H. N.; Lee, M. J.; Park, S. J.; Park, S. J.; Jeon, J.
Y.; Kang, J. H.; Pae, A. N.; Rhim, H.; Lee, J. Y. Synthesis and biological
evaluation of novel T-type calcium channel blockers. Bioorg. Med.
Chem. Lett. 2007, 17, 471−475.
(15) Alisi, M. A.; Brufani, M.; Filocamo, L.; Gostoli, G. Synthesis and
structure-activity relationships of new acetylcholinesterase inhibitors:
morpholinoalkyl carbamoyloxyseroline derivatives. Bioorg. Med. Chem.
Lett. 1995, 5, 2077−2080.
(16) Sengupta, S.; Chen, H.; Togawa, T.; DiBello, P. M.; Majors, A.
K.; Budy, B.; Ketterer, M. E.; Jacobsen, D. W. Albumin thiolate anion
is an intermediate in the formation of albumin-S-S-homocysteine. J.
Biol. Chem. 2001, 276, 30111−30117.
(17) Ellman, G. L. Tissue sulfhydryl groups. Arch. Biochem. Biophys.
1959, 82, 70−77.
ABBREVIATIONS USED
■
ARE, antioxidant response element; CAs, contrast agents;
DCM, dichloromethane; DO3A, 1,4,7,10-tetraazacyclodode-
cane-1,4,7-triacetic acid; DOTA, 1,4,7,10-tetraazacyclodode-
cane-1,4,7,10-tetraacetic acid; GSH, reduced glutathione;
HSA, human serum albumin; GGT, γ-glutamyltransferase;
KEAP1, Kelch-like ECH-associated protein 1; mHSA, mercapt-
human serum albumin; MRI, magnetic resonance imaging;
MRP, multidrug resistance-associated protein; NPSH, non-
protein thiols; NRF2, nuclear factor erythroid 2-related factor
2; PBS, phosphate-buffered saline; PEG, polyethylene glycol;
rHSA, recombinant human serum albumin; rHSA-LR,
recombinant lipid-reduced human serum albumin; TBAI,
tetrabutylammonium iodide; TOCSY, total correlation spec-
troscopy
(18) Takeda, K.; Shigemura, A.; Hamada, S.; Gu, W.; Fang, D.; Sasa,
K.; Hachiya, K. Dependence of reaction rate of 5,5′-dithiobis-(2-
nitrobenzoic acid) to free sulfhydryl groups of bovine serum albumin
and ovalbumin on the protein conformations. J. Protein Chem. 1992,
11, 187−192.
REFERENCES
■
(1) Moriarty-Craige, S. E.; Jones, D. P. Extracellular thiols and thiol/
disulfide redox in metabolism. Annu. Rev. Nutr. 2004, 24, 481−509.
(2) Hayes, J. D.; McMahon, M. NRF2 and KEAP1 mutations:
permanent activation of an adaptive response in cancer. Trends
Biochem. Sci. 2009, 34, 176−188.
(3) Di Giuseppe, D.; Ulivelli, M.; Bartalini, S.; Battistini, S.; Cerase,
A.; Passero, S.; Summa, D.; Frosali, S.; Priora, R.; Margaritis, A.; Di
Simplicio, P. Regulation of redox forms of plasma thiols by albumin in
multiple sclerosis after fasting and methionine loading test. Amino
Acids 2010, 38, 1461−1471.
(4) Chen, X.; Carystinos, G. D.; Batist, G. Potential for selective
modulation of glutathione in cancer chemotherapy. Chemico-Biol.
Interactions 1998, 111−112, 263−275.
(5) Ong, P. S.; Chan, S. Y.; Ho, P. C. Microarray analysis revealed
dysregulation of multiple genes associated with chemoresistance to
As2O3 and increased tumor aggressiveness in a newly established
arsenic-resistant ovarian cancer cell line, OVCAR-3/AsR. Eur. J Pharm.
Sci. 2012, 45, 367−378.
(6) Tuttle, S.; Horan, A.; Koch, C. J.; Held, K.; Manevich, Y.;
Biaglow, J. Radiation-sensitive tyrosine phosphorylation of cellular
proteins: sensitive to changes in GSH content induced by pretreat-
ment with N-acetyl-L-cysteine or L-buthionine-S,R-sulfoximine. Int. J
Radiat. Oncol. Biol. Phys. 1998, 42, 833−838.
(7) Pei, P.; Horan, M. P.; Hille, R.; Hemann, C. F.; Schwendeman, S.
P.; Mallery, S. R. Reduced nonprotein thiols inhibit activation and
function of MMP-9: implications for chemoprevention. Free Radic.
Biol. Med. 2006, 41, 1315−1324.
(8) Blakytny, R.; Erkell, L. J.; Brunner, G. Inactivation of active and
latent transforming growth factor beta by free thiols: potential redox
regulation of biological action. Int. J Biochem. Cell. Biol. 2006, 38,
1363−1373.
1
(19) Christodoulou, J.; Sadler, P. J.; Tucker, A. H NMR of albumin
in human blood plasma: drug binding and redox reactions at Cys34.
FEBS Lett. 1995, 376, 1−5.
(20) Stewart, A. J.; Blindauer, C. A.; Berezenko, S.; Sleep, D.; Tooth,
D.; Sadler, P. J. Role of Tyr84 in controlling the reactivity of Cys34 of
human albumin. FEBS J. 2005, 272, 353−362.
(21) Kragh-Hansen, U.; Chuang, V. T. C.; Otagiri, M. Practical
aspects of the ligand-binding and enzymatic properties of human
serum albumin. Biol. Pharm. Bull. 2002, 25, 695−704.
(22) Bohndiek, S. E.; Kettunen, M. I.; Hu, D. E.; Kennedy, B. W.;
Boren, J.; Gallagher, F. A.; Brindle, K. M. Hyperpolarized [1−13C]-
ascorbic and dehydroascorbic acid: vitamin C as a probe for imaging
redox status in vivo. J. Am. Chem. Soc. 2011, 133, 11795−11801.
(23) Ratnakar, S. J.; Viswanathan, S.; Kovacs, Z.; Jindal, A. K.; Green,
K. N.; Sherry, A. D. Europium(III) DOTA-tetraamide Complexes as
Redox-Active MRI Sensors. J. Am. Chem. Soc. 2012, 134, 5798−5800.
(24) Digilio, G.; Menchise, V.; Gianolio, E.; Catanzaro, V.; Carrera,
C.; Napolitano, R.; Fedeli, F.; Aime, S. Exofacial protein thiols as a
route for the internalization of Gd(III)-based complexes for MRI cell
labelling. J. Med. Chem. 2010, 53, 4877−4890.
(25) Menchise, V.; Digilio, G.; Gianolio, E.; Cittadino, E.; Catanzaro,
V.; Carrera, C.; Aime, S. In vivo labeling of B16 melanoma tumor
xenograft with a thiol-reactive gadolinium based MRI contrast agent.
Mol. Pharmaceutics 2011, 8, 1750−1756.
(26) Inner-sphere relaxation refers to relaxation enhancement of a
water molecule directly coordinated to the gadolinium. Second-sphere
relaxation occurs when water molecules in the second coordination
sphere (H-bonded to atoms on the complex) are relaxed via a dipolar
mechanism. Outer-sphere relaxation arises from the translational
diffusion of water molecules near the Gd(III) complex.
(9) Satoh, H.; Moriguchi, T.; Taguchi, K.; Takai, J.; Maher, J. M.;
Suzuki, T.; Winnard, P. T., Jr.; Raman, V.; Ebina, M.; Nukiwa, T.;
Yamamoto, M. Nrf2-deficiency creates a responsive microenvironment
for metastasis to the lung. Carcinogenesis 2010, 31, 1833−1843.
(10) Raghunand, N.; Jagadish, B.; Trouard, T. P.; Galons, J. P.;
Gillies, R. J.; Mash, E. A. Redox-sensitive contrast agents for MRI
based on reversible binding of thiols to serum albumin. Magn. Reson.
Med. 2006, 55, 1272−1280.
(27) Caravan, P.; Ellison, J. J.; McMurry, T. J.; Lauffer, R. B.
Gadolinium(III) chelates as MRI contrast agents: structure, dynamics
and applications. Chem. Rev. 1999, 99, 2293−2352.
(28) Barge, A.; Cravotto, G.; Gianolio, E.; Fedeli, F. How to
determine free Gd and free ligand in solution of Gd chelates. A
technical note. Contrast Media Mol. Imaging 2006, 1, 184−188.
(11) Guntle, G. P.; Jagadish, B.; Mash, E. A.; Powis, G.; Dorr, R. T.;
Raghunand, N. Tumor xenograft response to redox-active therapies
assessed by Magnetic Resonance Imaging using a thiol-bearing DOTA
complex of gadolinium. Trans. Oncol. 2012, 5, 190−199.
10386
dx.doi.org/10.1021/jm300736f | J. Med. Chem. 2012, 55, 10378−10386